Lectin Purification and Analysis: Methods and Protocols (Methods in Molecular Biology, 2132) 1071604295, 9781071604298

This volume presents a series of experimental protocols for glycan analysis using lectins and the functional analysis of

102 44 26MB

English Pages 717 [691] Year 2020

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Preface
Contents
Contributors
Chapter 1: Structural Database for Lectins and the UniLectin Web Platform
1 Introduction
2 Existing Lectin Databases
3 UniLectin3D, Database of Curated Lectin 3D Structure, and Their Interacting Ligands
3.1 Searching by Keywords
3.2 Searching by Kingdom Order, Carbohydrate-Binding Site Class, and Species Family
3.3 Searching by Monosaccharide and Associate IUPAC Sequence
3.4 Searching by Fold
3.5 Advanced Search: Searching by Multiple Criteria
3.6 Lectin Sequence Detailed Interface
3.7 X-Ray Structure Detailed Interface
4 PropLec: Database of Structure-Based Predicted β-Propellers
4.1 Searching by Keywords and by Family
4.2 Searching by Number of Blades in the Propeller
4.3 Searching by Phylum
4.4 Advanced Search
4.5 Detailed Results
4.6 Searching by Other Pfam Functional Domains
5 Conclusion and Discussion
References
Chapter 2: Purification of Sugar-Binding Peptides from L-Type Lectins
1 Introduction
2 Materials
2.1 Preparation of Sugar-Immobilized Sepharose
2.2 Purification of L-Type Lectins
2.3 Purification and Identification of the Sugar-Binding Peptide of L-Type Lectins
3 Methods
3.1 Preparation of Sugar-Immobilized Sepharose
3.2 Purification of L-Type Lectins
3.3 Purification and Identification of the Sugar-Binding Peptides of L-Type Lectins
4 Notes
References
Chapter 3: Recombinant Expression and Purification of Animal Intracellular L-Type Lectins
1 Introduction
2 Materials
2.1 Construction of Expression Vectors
2.2 Expression and Purification of Animal L-Type Lectins
3 Methods
3.1 Soluble Expression and Purification of Human VIP36-CRD
3.2 Refolding and Purification of ERGIC-53-CRD
3.3 Refolding and Purification of VIPL-CRD
4 Notes
References
Chapter 4: Frontal Affinity Chromatography: A Highly Suitable Retardation Phenomenon-Based Research Tool for Analyzing Weak In...
1 Introduction
2 Peculiar Characteristics of Lectin-Glycan Interactions
3 Significance of Fuzzy and Weak Interactions in Biological Systems
4 FAC: A Powerful Analytical Tool for Weak Interactions
5 FAC Viewed as Retardation Analysis of Bioaffinity
6 Two Diverse Aspects of FAC
7 Concluding Remarks
References
Chapter 5: Metazoan Soluble β-Galactoside-Binding Lectins, Galectins: Methods for Purification, Characterization of Their Carb...
1 Introduction
2 Materials
2.1 Purification of Recombinant Galectins
2.2 Carboxymethylation of Galectin-1
2.3 Hemagglutinin Assay
2.4 Biotinylation Using Primary Amine Residues of Galectin
2.5 Biotinylation Using Cysteine Residues of Galectin-1
2.6 Enzyme-Linked Lectin Sorbent Assay
2.7 Surface Plasmon Resonance (SPR) Assay
2.8 Staining with Galectin
2.9 Immunostaining of Galectins
2.10 Special Equipment
3 Methods
3.1 Purification of a Galectin
3.2 Carboxymethylation of Galectin-1
3.3 Quality Control of Purified Galectin by Hemagglutination Assay
3.3.1 Preparation of Red Blood Cells (RBC)
3.3.2 Calibration of RBC
3.3.3 Hemagglutination Assay
3.4 Biotinylation of Galectins Using Primary Amine Residues of a Galectin
3.5 Biotinylation of Galectin-1 Using Cysteine Residues
3.6 Carbohydrate-Binding Specificity Analysis of Galectin: Enzyme-Linked Lectin Sorbent Assay
3.7 Carbohydrate-Binding Specificity Analysis of Galectin: SPR Assay
3.8 Visualization of Ligands of a Galectin Expressed in Cells (Lectin Staining)
3.9 Detection of a Galectin Expressed in Cells Using an Antibody against Galectin
4 Notes
References
Chapter 6: Expression, S-Nitrosylation, and Measurement of S-Nitrosylation Ratio of Recombinant Galectin-2
1 Introduction
2 Materials
2.1 Preparation of Recombinant Gal-2
2.2 S-Nitrosylation of Recombinant Gal-2
2.3 Measuring S-Nitrosylation by Saville-Griess Assay
3 Methods
3.1 Expression of Recombinant Gal-2 in Escherichia coli
3.2 Purification of Recombinant Gal-2 by Affinity Chromatography
3.3 S-Nitrosylation of Recombinant Gal-2
3.4 Measuring S-Nitrosylation of Gal-2 by Saville-Griess Assay
4 Notes
References
Chapter 7: Expression and Purification of Full-Length and Domain-Fragment Recombinant Pentraxin 3 (PTX3) Proteins from Mammali...
1 Introduction
2 Materials
2.1 Plasmids for Recombinant PTX3 Protein Expression
2.2 Cell Culture for the Mammalian Expression System
2.3 Expression of the N-Terminal Domain of PTX3 by Bacterial Cells
2.4 Purification of Recombinant PTX3 Proteins from Mammalian Cells with a 6xHis Tag Affinity Column
2.5 Purification of Recombinant N-Terminal Domain PTX3 Proteins from Bacterial Cells with a 6xHis Tag Affinity Column
3 Methods
3.1 Establishment of Recombinant PTX3 Expressing Stable Mammalian Cell Lines
3.2 Large-Scale Culture of PTX3-Expressing Stable Cell Lines
3.3 Expression of the N-Terminal Domain of PTX3 by Bacterial Cells
3.4 Purification of Recombinant PTX3 Proteins from the Mammalian Cell Culture Supernatant with a 6xHis Tag Affinity Column
3.5 Purification of Recombinant N-Terminal Domain PTX3 Proteins from Bacterial Cells with a 6xHis Tag Affinity Column
4 Notes
References
Chapter 8: Identification of Siglec Cis-Ligands by Proximity Labeling
1 Introduction
2 Materials
2.1 Proximity Labeling Using Tyramide
2.2 Flowcytometry Analysis
2.3 Immunoprecipitation, SDS-PAGE, and Western Blotting
3 Methods
3.1 Proximity Labeling
3.2 Flow Cytometry Analysis of Labeling Efficacy of Cell Surface Molecules
3.3 SDS-PAGE and Western Blotting for Biotinylated Proteins
3.4 Immunoprecipitation of Biotinylated Proteins, SDS-PAGE, and Western blotting to Identify Cis-Ligands
4 Notes
References
Chapter 9: Preparation of Recombinant Siglecs and Identification of Their Ligands
1 Introduction
2 Materials
2.1 Preparation of Recombinant Siglec-Fc
2.1.1 Equipment
2.1.2 Reagents and Consumables
2.2 Flow Cytometry Using Siglec-Fc as a Probe
2.2.1 Equipment
2.2.2 Reagents and Consumables
2.3 Proximity Labeling Using Siglec-Fc as a Probe
2.3.1 Equipment
2.3.2 Reagents and Consumables
3 Methods
3.1 Preparation of Recombinant Siglec-Fc
3.1.1 Maintenance of Expi293F Cells
3.1.2 Transfection
3.1.3 Purification
3.2 Flow Cytometry Using Siglec-Fc as a Probe
3.3 Proximity Labeling with Siglec-Fc
3.3.1 Biotin-Tyramide Labeling of Ligand Proteins
3.3.2 Affinity Purification of Biotinylated Proteins
3.3.3 Evaluation of Biotin Labeling
4 Notes
References
Chapter 10: Purification, Quantification, and Functional Analysis of Collectins
1 Introduction
2 Materials
2.1 Buffers and Assay Solutions
3 Methods
3.1 Purification of Native and Recombinant Human MBL
3.1.1 In the Case of Native Human MBL
3.1.2 In the Case of Recombinant Human MBL from a Mammalian Expression System
3.1.3 Purification Step
3.2 Preparation of the CL-K1-Rich Fraction from Human Serum or Plasma
3.3 Quantification of Human MBL by Sandwich ELISA
3.4 Saccharide-Binding Assay
3.4.1 To Assay Sugar Selectivity
3.5 Assays of Biological Function
3.5.1 Complement-Dependent Passive Hemolysis
3.5.2 Hemagglutination Inhibition (HI) Test
3.5.3 Neutralization (NT) Test
3.5.4 The Immunoplaque Assay for Viral Growth Inhibition
4 Notes
References
Chapter 11: Selectin-Binding Assay by Flow Cytometry
1 Introduction
2 Materials
2.1 Cell Culture
2.2 Flow Cytometry
3 Methods
3.1 Cell Preparation
3.2 Flow Cytometry
3.3 Neutralization of sLex/a Using Monoclonal Antibodies
4 Notes
References
Chapter 12: Direct Binding Analysis Between C-Type Lectins and Glycans Using Immunoglobulin Receptor Fusion Proteins
1 Introduction
2 Materials
2.1 Transfection
2.2 Evaluation of Concentration
2.3 Molecular Weight Analysis by Western Blotting
2.4 Molecular Weight Analysis Using Coomassie Brilliant Blue
2.5 Binding Assay
2.6 Staining
3 Methods
3.1 Transfection
3.2 Evaluation of Concentration
3.3 Molecular Weight Analysis by Western Blotting
3.4 Molecular Weight Analysis Using Coomassie Brilliant Blue
3.5 Binding Assay
3.6 Staining
4 Notes
References
Chapter 13: Expression and Characterization of Hyaluronan-Binding Protein Involved in Hyaluronan Depolymerization: HYBID, Alia...
1 Introduction
2 Materials
2.1 Cell Culture
2.2 Transfection of Expression Vector and Positive Clones Selection
2.3 HA Labeling
2.4 Binding Assay
3 Methods
3.1 Preparation of HYBID Stable Transfectants (HYBID/HEK293 Cells)
3.2 Preparation of [3H]-Labeled High-Molecular-Weight HA ([3H]HA)
3.3 Depolymerization of [3H]-Labeled HA ([3H]HA) and FA-HA H1
3.3.1 Depolymerization of [3H]HA
3.3.2 Depolymerization of FA-HA H1
3.4 Binding Assay for HYBID to Glycosaminoglycans (GAGs) by Coprecipitation with CPC
4 Notes
References
Chapter 14: Paracoccin: Purification and Validation of Its Lectin and Enzymatic Properties
1 Introduction
2 Materials
2.1 Crude Extract of P. brasiliensis Yeasts
2.2 Protein Expression by E. coli
2.3 Protein Expression by P. pastoris
2.4 PCN Purification by Affinity to GlcNAc
2.5 PCN Purification by Affinity to Particulate Chitin
2.6 Assay for PCN Binding to Laminin
2.7 N-Acetylglucosaminidase (NAGase) Activity Assay
3 Methods
3.1 Preparation of the Crude Extract of P. brasiliensis
3.2 PCN Expression in E. coli and Purification Process
3.3 PCN Expression in P. pastoris and Purification Guidelines
3.4 Purification of PCN by Affinity to Immobilized GlcNAc
3.5 Purification of PCN by Affinity to Particulate Chitin
3.6 Lectin Activity: Assay for PCN Binding to Laminin
3.7 Enzymatic Properties: NAGase Activity Assay
4 Notes
References
Chapter 15: In Vitro Mannosidase Assay of EDEMs: ER Degradation-Enhancing α-Mannosidase-Like Proteins
1 Introduction
2 Materials
2.1 Preparation of Buffers and Reagents
2.2 Preparation of Plasmids and Poly-l-Lysine-Coated Dishes
3 Methods
3.1 Cell Culture and Transfection
3.2 Preparation of Recombinant Proteins from Cultured Cells
3.3 In Vitro Mannosidase Assay
4 Notes
References
Chapter 16: Galactose-Specific, Hemolytic Lectin CEL-III from Cucumaria echinata
1 Introduction
2 Materials
2.1 Reagents/Solutions/Materials
2.2 Preparation of the Carbohydrate-Immobilized Microplate
2.3 Preparation of the Colloidal Gold Solution
3 Methods
3.1 Purification of CEL-III
3.2 Microplate Assay for the Binding of CEL-III to Lactose
4 Notes
References
Chapter 17: Simple and Rapid Detection of Glycoforms by ``Lectin Inhibition´´ Assay
1 Introduction
2 Materials
2.1 Periodate Treatment of Antibody
2.2 Preparation of Tf Glycoforms
2.3 SSA Inhibition in ELISA
3 Methods
3.1 Periodate Treatment of Antibody
3.2 Preparation of Tf Glycoforms
3.3 SSA Inhibition in ELISA
4 Notes
References
Chapter 18: Glycoform-Specific Visualization in Immunohistochemistry by ``Lectin Inhibition´´
1 Introduction
2 Materials
2.1 Periodate Treatment of Antibody
2.2 SSA Inhibition with Tissue Sections
2.3 SSA Staining with Sialidase-Treated Sections
2.4 SSA Inhibition with Sialidase-Treated Sections
3 Methods
3.1 Periodate Treatment of Antibody
3.2 SSA Inhibition with Tissue Sections
3.3 SSA Staining with Sialidase-Treated Sections
3.4 SSA Inhibition with Sialidase-Treated Sections
4 Notes
References
Chapter 19: Botulinum Hemagglutinin: Critical Protein for Adhesion and Absorption of Neurotoxin Complex in Host Intestine
1 Introduction
2 Structure of Botulinum HA
3 Carbohydrate Binding of HA1 and HA3
4 Differences in Lectin Activities Between Serotypes
5 L-PTC/A Is Absorbed from M-Cells in the Mouse Intestine by the Carbohydrate-Binding Activity of HA
6 Apical Cell Surface Binding of HA of L-PTC/A-D
7 Basolateral Cell Surface Binding of HA/[A, B]
8 Multivalency Effect of HA
9 Conclusion
References
Chapter 20: Functional Analysis of Botulinum Hemagglutinin (HA)
1 Introduction
2 Materials
2.1 Generation of Recombinant HA
2.2 Cell ELISA
2.3 HA Pull-Down Assay
2.4 In Situ Loop Assay
2.5 Immunofluorescence Staining
3 Methods
3.1 Generation of Recombinant HA
3.2 Cell ELISA: Binding Assay Using Epithelial Cells
3.3 HA Pull-Down Assay: Binding Assay to Human Fc-Tagged Glycoprotein
3.4 Localization of HA in the Mouse Intestine
3.4.1 Mouse In Situ Loop Assay
3.4.2 Immunofluorescence Staining
4 Notes
References
Chapter 21: Purification and Functional Characterization of the Effects on Cell Signaling of Mytilectin: A Novel β-Trefoil Lec...
1 Introduction
2 Materials
2.1 Mussels
2.2 Lectin Purification
2.3 Cell Culture and Cell Growth Assay
2.4 Antibodies and Western Blotting
3 Methods
3.1 Purification of MytiLec-1
3.2 Cell Growth Quantification for MytiLec-1 Treated Cells
3.3 Lectin-Dependent Signal Transduction
4 Notes
References
Chapter 22: Lectin-Type Ubiquitin Ligase Subunits: Fbs Proteins and Their Applications for Use
1 Introduction
2 Materials
2.1 N-Glycopeptide Enrichment Method
2.2 Detecting Damaged Organelles in Cells Using Fbs3
3 Methods
3.1 N-Glycopeptide Enrichment Method
3.1.1 Preparation of Recombinant Fbs1 GYR Protein
3.1.2 Preparation of Sample Peptide from Cultured Cells
3.1.3 Enrichment of N-Glycopeptides by N-Glyco-FASP Method
3.2 Detection of Damaged Organelles in Cells by Using Fbs3
3.2.1 Plasmid Construction and Transfection
3.2.2 Detecting Organelle Damage by Time-Lapse Imaging
4 Notes
References
Chapter 23: F-Type Lectins: Structure, Function, and Evolution
Abbreviations
1 Introduction
2 Structural Features of FTLs
2.1 The FTL Structural Fold
2.2 Carbohydrate Recognition by the FTLD
2.3 Domain and Subunit Organization of FTLs
2.3.1 FTLs with Tandemly Arrayed FTLDs
2.3.2 FTL Isoforms and Diversity in Ligand Recognition
2.3.3 Oligomeric Organization of FTL Subunits
2.3.4 Structural Diversification of FTLs
3 Taxonomic Distribution and Evolution of the FTLD
4 Biological Roles of FTLs
5 Conclusions
References
Chapter 24: Purification and Biochemical Characterization of Selected F-Type Lectins
Abbreviations
1 Introduction
2 Materials
2.1 Purification of FTLs
2.2 Biochemical Characterization of FTLs
3 Methods
3.1 Preparation of Liver Tissue Extracts
3.2 Preparation of Serum
3.3 Expression of Recombinant FTL Proteins
3.4 FTL Isolation by Affinity Chromatography
3.5 FTL Purification by Size Exclusion Chromatography (SEC)
3.6 Edman Sequence Analysis
3.7 Assessment of the Native Molecular Size by SEC
3.8 ELISA for Analysis of Carbohydrate-Binding Activity and Specificity
3.9 Agglutination Assay
3.10 Agglutination-Inhibition Assay
3.11 Glycan Array
4 Notes
References
Chapter 25: LecA (PA-IL): A Galactose-Binding Lectin from Pseudomonas aeruginosa
1 Introduction
2 Materials
2.1 Gene Sequence
2.2 Protein Production in Escherichia coli
2.3 Protein Purification
2.4 Surface Plasmon Resonance
2.5 Protein Crystallization
3 Methods
3.1 LecA Production
3.1.1 Transformation and Preculture
3.1.2 Culture
3.1.3 LecA Purification by Affinity Chromatography
3.2 Surface Plasmon Resonance (SPR)
3.2.1 Sample Preparation
3.2.2 LecA Immobilization on SPR Chip Using Amine Coupling Method
3.3 LecA Crystallization
3.3.1 Sample Preparation
3.3.2 Hanging Drop Vapour Diffusion
4 Notes
References
Chapter 26: Glycan Recognition and Application of P-Type Lectins
1 Introduction
2 Materials
2.1 Preparation of Recombinant Proteins
2.2 SDS-PAGE
2.3 Lectin Blotting
3 Methods
3.1 Expression and Purification of Recombinant Human Dom9-his by Methylotrophic Yeast
3.2 Lectin Blotting with the Dom9-his as a Probe
4 Notes
References
Chapter 27: Purification and Assays of Tachylectin-5
1 Introduction
2 Materials
2.1 Preparation of Hemolymph Plasma
2.2 Purification of Tachylectin-5
2.2.1 Preparation of Acetyl Group-Immobilized Resin (Acetyl-Toyopearl)
2.2.2 Acetyl-Toyopearl Column Chromatography
2.3 Hemagglutination Assay
2.4 Bacterial Agglutination Assay
2.5 A Synergistic Effect of Tachylectin-5A or Tachylectin-5B on Antimicrobial Activity of Big Defensin
3 Methods
3.1 Preparation of Hemolymph Plasma
3.2 Purification of Tachylectin-5
3.2.1 Preparation of Acetyl-Toyopearl
3.2.2 Acetyl-Toyopearl Column Chromatography
3.3 Hemagglutination Assay
3.4 Bacterial Agglutination Assay
3.5 A Synergistic Effect of Tachylectin-5A or Tachylectin-5B on Antimicrobial Activity of Big Defensin
4 Notes
References
Chapter 28: Preparation of Soluble Malectin and Its Tetramer
1 Introduction
2 Materials
2.1 Construction of Plasmids for sMAL Expression in Escherichia coli
2.2 Expression and Purification of sMAL
2.3 Biotinylation of sMAL and Tetramer Preparation for Cell Binding Assay
3 Methods
3.1 Construction of Plasmids for sMAL Expression in E. coli
3.2 Expression and Purification of sMAL
3.3 Biotinylation of sMAL and Tetramer Preparation for Cell Binding Assay
4 Notes
References
Chapter 29: Calnexin/Calreticulin and Assays Related to N-Glycoprotein Folding In Vitro
1 Introduction
2 Materials
2.1 Expression of CRT in Bacteria
2.2 Purification of CRT
2.2.1 Glutathione Sepharose Column Chromatography
2.2.2 Cleavage of GST-CRT Fusion Protein
2.2.3 Mono Q Anion-Exchange Liquid Chromatography
2.3 Aggregation Suppression Assay
2.4 Heat-Induced Inactivation of Jack Bean α-Mannosidase
2.5 The Enzyme Assay for α-Mannosidase
3 Methods
3.1 Expression of Mammalian CRT in Bacteria
3.2 Purification of Mammalian CRT Expressed in Bacteria
3.2.1 Glutathione Sepharose Column Chromatography
3.2.2 Cleavage of GST-CRT Fusion Protein
3.2.3 Mono Q Anion-Exchange Chromatography
3.3 Aggregation Suppression Assay
3.4 Heat-Induced Inactivation of Jack Bean α-Mannosidase
3.5 The Enzyme Assay for α-Mannosidase
4 Notes
References
Chapter 30: Purification and Assays of Tachylectin-2
1 Introduction
2 Materials
2.1 Preparation of Hemocyte Extract
2.2 Purification of Tachylectin-2
2.2.1 Preparation of Dextran Sulfate-Sepharose CL-6B
2.2.2 Dextran Sulfate-Sepharose CL-6B Column Chromatography
2.2.3 CM-Sepharose CL-6B Column Chromatography
2.2.4 Mono S Column Chromatography
2.3 Hemagglutination Assay
2.4 Bacterial Agglutination Assay
2.5 Determination of Association Constants
2.6 Analytical Ultracentrifugation
3 Methods
3.1 Preparation of Hemocyte Extract
3.2 Purification of Tachylectin-2
3.2.1 Preparation of Dextran Sulfate-Sepharose CL-6B
3.2.2 Dextran Sulfate-Sepharose CL-6B Column Chromatography
3.2.3 CM-Sepharose CL-6B Column Chromatography
3.2.4 Mono S Column Chromatography
3.3 Hemagglutination Assay
3.4 Bacterial Agglutination Assay
3.5 Determination of Association Constants
3.6 Analytical Ultracentrifugation
4 Notes
References
Chapter 31: Purification and Assays of Tachycitin
1 Introduction
2 Material
2.1 Preparation of Hemocyte Debris
2.2 Purification of Tachycitin
2.2.1 Preparation of Protein Extract from Hemocyte Debris
2.2.2 Sephadex G-50 Column Chromatography
2.2.3 S-Sepharose FF Column Chromatography
2.3 Antimicrobial Activity
2.4 Bacterial Agglutination Assay
2.5 Analytical Ultracentrifugation
3 Methods
3.1 Preparation of Hemocyte Debris
3.2 Purification of Tachycitin
3.2.1 Preparation of Protein Extract from Hemocyte Debris
3.2.2 Sephadex G-50 Column Chromatography
3.2.3 S-Sepharose FF Column Chromatography
3.3 Antimicrobial Activity
3.4 Bacterial Agglutination Assay
3.5 Analytical Ultracentrifugation
4 Notes
References
Chapter 32: Methods for Purifying Datura stramonium Agglutinin and Producing Recombinant Agglutinin Protein in a Heterologous ...
1 Introduction
2 Materials
2.1 Plant Materials
2.2 Bacterial Strains
2.3 cDNA Material
2.4 Hemagglutination Assay
2.5 Preparation of Chitin-Gel Via Chitosan Acetylation
2.6 Buffers and Materials for Purifications of Native and Recombinant DSA
2.7 Buffers and Media for Transformation of Arabidopsis thaliana
2.8 Detection of Transgene and Recombinant Protein
3 Methods
3.1 Hemagglutination Assay
3.2 Preparation of Chitin-Gel Via Chitosan Acetylation
3.3 Purification of DSA Isolectins from Datura Seeds Using Affinity Chromatography
3.4 Separation of DSA Isolectins Using Hydrophobic Interaction Chromatography
3.5 Generation of DSA-B Transgenic Arabidopsis Plants
3.6 Purification of DSA Recombinant BB-Isolectin from Transgenic Arabidopsis Plants
4 Notes
References
Chapter 33: ZG16p, an Animal Homologue of Plant β-Prism Fold Lectins: Purification Methods of Natural and Recombinant ZG16p an...
1 Introduction
2 Materials
2.1 Purification of ZG16p from Rat Pancreas (Rat ZG16p)
2.2 Expression and Purification of Recombinant ZG16p
2.3 Inhibition of Cancer Cell Growth by ZG16p
2.3.1 Cell Culture
2.3.2 ZG16p Solution
2.3.3 Cell Proliferation Assay
3 Methods
3.1 Purification of Rat ZG16p
3.1.1 Preparation of Zymogen Granule Extracts
3.1.2 Purification of Rat ZG16p Using Heparin Column Chromatography
3.2 Expression and Purification of Recombinant ZG16p
3.3 Inhibition of Cancer Cell Growth by ZG16p
4 Notes
References
Chapter 34: ArtinM: Purification and Evaluation of Biological Activities
1 Introduction
2 Materials
2.1 Purification of ArtinM
2.2 Analysis of the Purity of ArtinM
2.3 Endotoxin Removal from ArtinM Solution
2.4 Biological Activity of ArtinM
3 Methods
3.1 Preparation of Jackfruit Crude Seed Extract
3.2 Purification of ArtinM
3.2.1 Jacalin Removal from Jackfruit Crude Seed Extract
3.2.2 Purification of ArtinM from Crude Seed Extract Depleted of Jacalin
3.3 Endotoxin Removal from ArtinM Solution
3.4 Biological Activity of ArtinM
4 Notes
References
Chapter 35: Bacterial Expression of Rhamnose-Binding Lectin from Catfish Eggs
1 Introduction
2 Materials
2.1 Protein Expression in KRX and Isolation of Inclusion Bodies
2.2 Denaturation, Refolding, and Purification
2.2.1 Buffers
2.2.2 Affinity Resins and the Others
2.3 Protein Detection
2.4 Hemagglutination and Inhibition Assays
3 Methods
3.1 Protein Expression in KRX and Isolation of Inclusion Bodies
3.2 Denaturation, Refolding, and Purification of Recombinants Except for D3DeltaC and D3DeltaN (See Note 2)
3.3 Denaturation, Purification, and Refolding of D3DeltaC and D3DeltaN
3.4 Protein Detection
3.5 Hemagglutination and Inhibition Assays
4 Notes
References
Chapter 36: A Bioassay for Determining Symbiotic Zooxanthellae Shape Control Using Lectin SLL-2 from the Octocoral Sinularia l...
1 Introduction
2 Materials
2.1 Selecting Zooxanthellae
2.2 Zooxanthellae
2.3 Medium
2.4 Cultivating and Observing Zooxanthellae
2.5 Preparation of Carbohydrate-Fixed Resin
2.6 Lectin Preparation
3 Methods
3.1 Selecting Zooxanthellae
3.2 Preparation of Sterilised Medium
3.3 Cultivating Zooxanthellae and Plotting Growth Curve
3.4 Measuring Diurnal Change in Zooxanthella Form
3.5 Counting the Number of Zooxanthellae Using Fluorescence
3.6 Preparation of Galactose-Fixed Resin
3.7 Preparing a Lectin SLL-2 from the Coral Sinularia Lochmodes
3.8 Bioassay for Symbiosis-Related Factor
4 Notes
References
Chapter 37: MIC4 from Toxoplasma gondii: A Lectin Acting as a Toll-Like Receptor Agonist
1 Introduction
2 Material
2.1 Plasmid Constructs
2.2 Cells
2.3 Culture Medium
2.4 Buffers and Solutions
2.5 Animal
2.6 Agonists of TLR
2.7 Others
3 Methods
3.1 Expression of Recombinant MIC4 (rMIC4) and rMIC4(K469M)
3.2 rMIC4 and rMIC4(K469M) SDS-PAGE
3.3 Isolation of Inclusion Bodies Through Centrifugation
3.4 Inclusion Bodies Solubilization
3.5 Ni-Sepharose Chromatography
3.6 Protein Refolding by Dialysis Method
3.7 Obtention, Culture, and Phenotyping of Bone Marrow-Derived Macrophage (BMDM)
3.8 The rMIC4 and rMIC4(K469M) Activity on Macrophages
3.9 The rMIC4 Activity on TLR2-Transfected HEK293T Cell Line: A Gene Reporter Assay
4 Notes
References
Chapter 38: Production and Characterization of MIC1: A Lectin from Toxoplasma gondii
1 Introduction
2 Materials
2.1 The Expression Plasmid
2.2 Competent Cells for Chemical Transformation
2.3 Culture Medium
2.4 Buffers and Solutions
2.5 Antibodies
2.6 Others
3 Methods
3.1 Expression of Recombinant MIC1 (rMIC1) and rMIC1 (T126A/T220A)
3.2 rMIC1 and rMIC1(T126A/T220A) SDS-PAGE
3.3 Isolation of Inclusion Bodies Through Centrifugation
3.4 Inclusion Body Solubilization
3.5 Ni-Sepharose Chromatography
3.6 Protein Refolding by Dialysis Method
3.7 Sugar-Binding Assay
4 Notes
References
Chapter 39: Affinity Labeling and Purification of Plant Chitin-Binding LysM Receptor with Chitin Octasaccharide Derivatives
1 Introduction
2 Materials
2.1 Preparation of Microsomal Membrane Fraction
2.2 Preparation of Biotinylated Chitooligosaccharide
2.3 Affinity Biotinylation
2.4 Purification of CEBiP
2.5 Equipment
3 Methods
3.1 Preparation of Microsomal Membrane Fraction from Rice Cultured Cells
3.2 Preparation of a Biotinylated Chitooligosaccharide
3.3 Affinity Labeling of a LysM Protein CEBiP in Rice Microsomal Membrane Fraction
3.4 Preparation of Affinity Columns for the Purification
3.5 Purification of CEBiP from Rice Plasma Membrane
4 Notes
References
Chapter 40: Purification of GNA-Related Lectins from Natural Sources
Abbreviations
1 Introduction
2 Materials
3 Methods
3.1 Lectin Extraction
3.2 Lectin Purification
3.2.1 Ion Exchange Chromatography
3.2.2 Affinity Chromatography on Mannose-Sepharose 4B
4 Notes
References
Chapter 41: Expression, Purification, and Applications of the Recombinant Lectin PVL from Psathyrella velutina Specific for Te...
1 Introduction
2 Materials
2.1 Production of rPVL
2.2 Purification of rPVL
2.3 Hemagglutination Assays (HA)
2.4 Affinity Measurements by Surface Plasmon Resonance (SPR)
2.5 rPVL Labelling
2.6 Enzyme-Linked Lectin Assays (ELLA)
2.7 Western Blotting of O-GlcNAcylated Proteins
2.8 Immunofluorescence Analysis on Whole Cells
2.9 Immunohistochemistry
3 Methods
3.1 Production of rPVL
3.2 Purification of rPVL
3.3 Hemagglutination Assays (HA)
3.4 Affinity Measurements by SPR
3.5 rPVL Labelling
3.5.1 Biotinylation
3.5.2 Horseradish Peroxidase Coupling
3.5.3 Alexa Fluor 488 Coupling
3.6 Enzyme-Linked Lectin Assay
3.7 Western Blotting of O-GlcNAcylated Proteins
3.8 Immunofluorescence Analysis on Whole Cells
3.9 Immunohistochemistry
4 Notes
References
Chapter 42: Yeast Flocculin: Methods for Quantitative Analysis of Flocculation in Yeast Cells
1 Introduction
2 Materials
2.1 Stock Solutions
2.2 Flocculation Assay
3 Method
3.1 Flocculation Assay
4 Notes
References
Chapter 43: Pleurotus cornucopiae Mycelial Lectin (PCL-M): Purification and Detection of the Activity on Mycelial Surface
1 Introduction
2 Materials
2.1 Organism
2.2 Culture Medium
2.3 Hemagglutination Assay
2.4 Purification
3 Methods
3.1 Culturing of the Mycelia
3.2 Assay Procedure
3.3 Purification of PCL-M
3.4 Observation of Adhesion of Red Blood Cells to Mycelia Mediated by PCL-M
4 Notes
References
Chapter 44: Expression and Purification of a Human Pluripotent Stem Cell-Specific Lectin Probe, rBC2LCN
1 Introduction
2 Materials
2.1 Reagents
2.2 Instruments
3 Methods
3.1 Construction of rBC2LCN-pET27b Plasmid
3.2 Expression of rBC2LCN-pET27b Plasmid in E. coli
3.3 Preparation of l-Fucose-Sepharose for Affinity Purification of rBC2LCN
3.4 Affinity Purification
3.5 Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis (SDS-PAGE) of Eluted Fractions
3.6 Dialysis of Eluted Fractions
3.7 Protein Quantification and Preservation
3.8 Cy3 Labeling of rBC2LCN
3.9 Live Staining
4 Notes
References
Chapter 45: Preparation of Fluorescent Recombinant Shiga Toxin B Subunit and Its Application to Flow Cytometry
1 Introduction
2 Materials
2.1 Expression of Stx1B in E. coli
2.2 Purification of Stx1B
2.3 Fluorescent Labeling
2.4 Flow Cytometry
3 Methods
3.1 Purification of Recombinant Stx1B
3.2 Fluorescence Conjugation of Recombinant Stx1B
3.3 Application of Fluorescent Stx1B to Flow Cytometry
4 Notes
References
Chapter 46: LecB, a High Affinity Soluble Fucose-Binding Lectin from Pseudomonas aeruginosa
1 Introduction
2 Materials
2.1 LecB Production
2.2 LecB Purification
2.3 Quality Control by Evaluation of Affinity
3 Methods
3.1 LecB Production
3.1.1 Transformation and Preculture
3.1.2 Culture
3.2 LecB Purification
3.3 Quality Control by Evaluation of Affinity
3.3.1 Sample Preparation
3.3.2 ITC Experiment
4 Notes
References
Chapter 47: Sialoglycovirology of Lectins: Sialyl Glycan Binding of Enveloped and Non-enveloped Viruses
1 Introduction
2 Determination of Sia-Binding Specificities for Viruses
3 Sialylglycoconjugates as Host Receptor Determinants of Viral Spike Glycoproteins for Enveloped Viral Infection
3.1 Orthomyxoviridae
3.2 Paramyxoviridae
3.3 Coronaviridae
4 Sialylglycoconjugates as Host Receptor Determinants of Viral Capsid Proteins for Non-enveloped Viral Infection
4.1 Picornaviridae
4.2 Caliciviridae
4.3 Reoviridae
4.4 Adenoviridae
4.5 Parvoviridae
4.6 Polyomaviridae
5 Molecular and Structural Basis of Viral Lectin-Sialyl Glycan Interactions
6 Conclusions, Perspectives, and Future Directions
References
Chapter 48: Hemagglutinin Inhibitors are Potential Future Anti-Influenza Drugs for Mono- and Combination Therapies
1 Introduction
2 Hemagglutinin Inhibitors (HAIs) of Influenza A Viruses
2.1 Mode of Terminal Sialyl-Galactose Linkage of Sialyl Sugar Chain Receptors That Bind to Influenza A Virus HA
2.2 Molecular Length and Structure of Sialyl Glycan Receptors
2.3 Multivalent Sialyl Sugar Chain Dimension That Can Reach the Binding Site of HA
2.4 Clustering of Sialyl Sugar Chains on Macromolecular Scaffolds
2.5 Modification of the C-3 Position of Terminal Sialic Acid in Sialyl Glycan Inhibits Catalytic Hydrolysis of Influenza Virus...
3 Synergizing the HAI Activity by Influenza Virus Neuraminidase-Targeting Drugs
4 Conclusion
References
Chapter 49: Preparation and Detection of Glycan-Binding Activity of Influenza Virus
1 Introduction
2 Materials
2.1 Propagation of Influenza Viruses
2.1.1 General Materials
2.1.2 Amplification of Influenza Virus in Embryonated Chicken Eggs
2.1.3 Amplification of Swine and Human Influenza Virus in MDCK Cells
2.2 Determination of Their Hemagglutination Units
2.3 Preparation of Glycan Array for Influenza Virus Detection
2.4 Detection of Influenza Virus Using Glycan Array
3 Methods
3.1 Propagation of Influenza Viruses
3.1.1 Amplification of Avian Influenza Virus in Embryonated Chicken Eggs
3.1.2 Amplification of Classical Swine and Human Influenza Viruses in MDCK Cells
3.2 Relative Quantities of Virus Particles by Hemagglutination Assay
3.2.1 Preparation of 0.5% Erythrocyte Suspension in PBS
3.2.2 Determination of Hemagglutination Units
3.3 Preparation of Glycan Array for Influenza Virus Detection
3.3.1 Conversion of SGP to α2,3SGP
3.3.2 Preparation of the Neoglycoprotein
3.4 Binding Assay Using the Neoglycoprotein
4 Notes
References
Chapter 50: Screening for Components/Compounds with Anti-Rotavirus Activity: Detection of Interaction Between Viral Spike Prot...
1 Introduction
2 Materials
2.1 MA104 Cell Culture
2.2 Virus Purification
2.3 Screening of Rotavirus-Binding Glycoproteins
2.4 Infection Assay and Immunofluorescence Detection
3 Methods
3.1 Rotavirus Purification
3.2 Screening of Rotavirus-Binding Glycoproteins by Using Evanescent Field-Based Fluorescence-Assisted Detection
3.2.1 Trypsin Activation
3.2.2 Rotavirus-Glycoprotein-Binding Assay
3.3 Rotavirus Infection Assay Using MA104 Cells
3.3.1 MA104 Cell Maintenance
3.3.2 Cell Preparation for Infection Assay
3.3.3 Neutralization Assay
3.3.4 Pretreatment Assay
3.3.5 Indirect Immunofluorescence Detection
4 Notes
References
Chapter 51: ELISA-Based Methods to Detect and Quantify Norovirus Virus-Like Particle Attachment to Histo-Blood Group Antigens
1 Introduction
2 Materials
2.1 Expression and Purification of Recombinant VLPs
2.2 Hemagglutination Inhibition Assay
2.3 Saliva-VLP Binding Assay
2.4 Synthetic Carbohydrate-VLP Binding Assay
3 Methods
3.1 Expression and Purification of Recombinant VLPs
3.2 Detection of Soluble ABH Antigens in Saliva by Hemagglutination Inhibition Assay
3.3 ELISA-Based Binding Assay 1: Saliva-VLP Binding Assay
3.4 ELISA-Based Binding Assay 2: Synthetic Carbohydrate-VLP Binding Assay
4 Notes
References
Chapter 52: FAM3B/PANDER-Like Carbohydrate-Binding Domain in a Glycosyltransferase, POMGNT1
1 Introduction
2 Materials
2.1 Lectin-Like Activity of POMGNT1 Stem Domain
2.1.1 Preparation of Recombinant Stem Domain
2.1.2 Carbohydrate-Binding Assay of Stem Domain
2.2 Functional Assay of Stem Domain
2.2.1 Cell Culture, Transfection, and Preparation of Cell Membrane Fraction
2.2.2 Enrichment of α-DG and Western Blotting
2.2.3 Glycosyltransferase Assay
3 Methods
3.1 Lectin-Like Activity of POMGNT1 Stem Domain
3.1.1 Preparation of Recombinant POMGNT1stem
3.1.2 Carbohydrate-Binding Assay of POMGNT1stem
3.2 Functional Assay of Stem Domain by Rescue Experiment in IIH6 Epitope Synthesis
3.2.1 Cell Culture, Transfection, and Preparation of Cell Membrane Fraction
3.2.2 Enrichment of α-DG and Western Blotting
3.2.3 Glycosyltransferase Assay of POMGNT1
4 Notes
References
Chapter 53: Mannose-Specific Oyster Lectin CGL1
1 Introduction
2 Materials
2.1 Mannose-Immobilized Resin
2.1.1 Reagents/Solutions
2.1.2 Equipment
2.2 Purification of CGL1
2.2.1 Reagents/Solutions/Materials
2.2.2 Equipment
3 Methods
3.1 Preparation of the Mannose-Immobilized Resin
3.2 Purification of CGL1
4 Notes
References
Chapter 54: Receptor-Binding Assays of Enterovirus D68
1 Introduction
2 Materials
2.1 Regents
2.2 Antibodies
2.3 Cells
2.4 Experimental Animals
2.5 Equipment
3 Methods
3.1 Clinical Sample Handling
3.2 Virus Isolation
3.3 Preparation of Polyclonal Antibody
3.4 Preparation of Slides for Glycan Array Analysis
3.5 Modification of Erythrocytes
3.6 Glycan Array Analysis
3.7 Hemagglutination Test Using Modified Erythrocytes
4 Notes
References
Chapter 55: Large-Scale Expression and Purification of Mumps Virus Hemagglutinin-Neuraminidase for Structural Analyses and Gly...
1 Introduction
2 Materials
2.1 Large-Scale MuV-HN Protein Expression
2.2 Purification and Modification of His-Tagged MuV-HN Proteins
2.3 Crystallization of the MuV-HN Protein
3 Methods
3.1 Large-Scale MuV-HN Protein Expression
3.1.1 Preparation of Expression Vectors Encoding MuV-HN Proteins
3.1.2 Large-Scale Cell Culture
3.1.3 Transient Transfection of HEK293S GnTI(-) Cells with Expression Vectors
3.2 Purification and Modification of the His-Tagged MuV-HN Protein
3.2.1 Affinity Chromatography
3.2.2 Methylation of Lysine Residues
3.2.3 Size Exclusion Chromatography (SEC)
3.3 Crystallization of the MuV-HN Protein
4 Notes
References
Chapter 56: Purification of AJLec: A Novel Galactose-Specific Lectin from the Sea Anemone Anthopleura japonica
1 Introduction
2 Materials
2.1 Lactose-Immobilized Resin
2.1.1 Reagents/Solutions
2.1.2 Equipment
2.2 Purification of AJLec
2.2.1 Reagents/Solutions/Materials
2.2.2 Equipment
3 Methods
3.1 Preparation of the Lactose Resin
3.2 Purification of AJLec
4 Notes
References
Chapter 57: Annexin Lectins: Ca2+-Dependent Heparin-Binding Activity, Phosphatidylserine-Binding Activity, and Anticoagulant A...
1 Introduction
2 Materials
2.1 Expression and Purification of GST (Glutathione S-Transferase)-Tagged Annexins (GST-ANXs)
2.2 Solid-Phase Heparin-Binding Assay
2.3 Solid-Phase PS-Binding Assay
2.4 Inhibition Assay of Plasma Coagulation Reaction
3 Methods
3.1 Expression and Purification of GST-ANXs
3.2 Solid-Phase Heparin-Binding Assay
3.2.1 Preparation of BSA-Heparin
3.2.2 Binding Assay
3.3 Solid-Phase PS-Binding Assay
3.4 Inhibition Assay of Plasma Coagulation Reaction
4 Notes
References
Chapter 58: Expression, Purification, and Functional Characterization of Tectonin 2 from Laccaria bicolor: A Six-Bladed Beta-P...
1 Introduction
2 Materials
2.1 Production of Recombinant Lb-Tec2
2.2 Purification of Recombinant Lb-Tec2
2.3 Isothermal Titration Microcalorimetry
2.4 Surface Plasmon Resonance
2.5 Caenorhabditis elegans Biotoxicity Assay
2.6 Localization of Lb-Tec2-Binding in C. elegans
2.7 Lectin Blot
2.8 Bacterial Agglutination Assay
3 Methods
3.1 Production of Recombinant Lb-Tec2
3.2 Purification of Recombinant Lb-Tec2
3.3 Isothermal Titration Microcalorimetry
3.4 Surface Plasmon Resonance
3.5 C. elegans Biotoxicity Assay
3.6 Localization of Lb-Tec2-binding in C. elegans
3.7 Lectin Blot
3.8 Bacterial Agglutination Assay
3.8.1 Assay with Purified Protein
3.8.2 Assay with Soluble Protein Extract
4 Notes
References
Chapter 59: The OAAH Family: Anti-Influenza Virus Lectins
1 Introduction
2 Materials
2.1 Cyanobacterium
2.2 Chromatography
2.3 Virus
2.4 Host Cell
2.5 Amide Black Assay
2.6 Neutral Red Dye Uptake Assay
2.7 Immunofluorescence Microscopy
2.8 Surface Plasmon Resonance Analysis
2.9 Enzyme-Linked Immunosorbent Assay (ELISA)
3 Methods
3.1 Purification of the Lectin (OAA)
3.2 Assay for Anti-Influenza Activity of Lectins
3.2.1 Amide Black Assay
3.2.2 Neutral Red Dye Uptake Assay
3.2.3 Direct Immunofluorescence Microscopy
3.3 Assay for Interaction Between Lectin and Influenza Viral Hemagglutinin (HA)
3.3.1 Surface Plasmon Resonance (SPR) Analysis with a Biacore X100
3.3.2 ELISA Assay
4 Notes
References
Correction to: Hemagglutinin Inhibitors are Potential Future Anti-Influenza Drugs for Mono- and Combination Therapies
Index
Recommend Papers

Lectin Purification and Analysis: Methods and Protocols (Methods in Molecular Biology, 2132)
 1071604295, 9781071604298

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

Methods in Molecular Biology 2132

Jun Hirabayashi Editor

Lectin Purification and Analysis Methods and Protocols

METHODS

IN

MOLECULAR BIOLOGY

Series Editor John M. Walker School of Life and Medical Sciences University of Hertfordshire Hatfield, Hertfordshire, UK

For further volumes: http://www.springer.com/series/7651

For over 35 years, biological scientists have come to rely on the research protocols and methodologies in the critically acclaimed Methods in Molecular Biology series. The series was the first to introduce the step-by-step protocols approach that has become the standard in all biomedical protocol publishing. Each protocol is provided in readily-reproducible step-bystep fashion, opening with an introductory overview, a list of the materials and reagents needed to complete the experiment, and followed by a detailed procedure that is supported with a helpful notes section offering tips and tricks of the trade as well as troubleshooting advice. These hallmark features were introduced by series editor Dr. John Walker and constitute the key ingredient in each and every volume of the Methods in Molecular Biology series. Tested and trusted, comprehensive and reliable, all protocols from the series are indexed in PubMed.

Lectin Purification and Analysis Methods and Protocols

Edited by

Jun Hirabayashi Cellular and Molecular Biotechnology Research Institute, National Institute of Advanced Industrial Science and Technology, Tsukuba, Ibaraki, Japan

Editor Jun Hirabayashi Cellular and Molecular Biotechnology Research Institute National Institute of Advanced Industrial Science and Technology Tsukuba, Ibaraki, Japan

ISSN 1064-3745 ISSN 1940-6029 (electronic) Methods in Molecular Biology ISBN 978-1-0716-0429-8 ISBN 978-1-0716-0430-4 (eBook) https://doi.org/10.1007/978-1-0716-0430-4 © Springer Science+Business Media, LLC, part of Springer Nature 2020, corrected publication 2020 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. The 3-D structure of the artificial lectin “Mitsuba-1” consisting of triple identical repeat subdomains of the β-trefoil fold by protein engineering (See Chapter 21). This Humana imprint is published by the registered company Springer Science+Business Media, LLC, part of Springer Nature. The registered company address is: 1 New York Plaza, New York, NY 10004, U.S.A.

Preface It is a great pleasure to me that the present collection of lectin protocols is published in the Methods in Molecular Biology series, especially because this edition substantially follows the previous one entitled Lectins, which was edited by myself and published in 2014 (Methods Mol Biol, 1200:1–613). This volume collected a series of experimental protocols for glycan analysis using lectins and the functional analysis of lectins. The most distinguished feature of this edition is the inclusion of two comprehensive overview chapters, which are “Comprehensive List of Lectins: Origins, Natures, and Carbohydrate Specificities” (Kobayashi et al., Methods Mol Biol, 1200:555–577, 2014) and “Lectin Structures: Classification Based on the 3-D Structures” (Fujimoto et al., Methods Mol Biol, 1200:579–606, 2014). An apparent indication by these comprehensive data is that a group of lectins are much more widely spread than had ever been thought of by lectin researchers. In fact, it is now evident that there are more than 50 lectin scaffolds in terms of Pfam protein family, for which 3D structures and sugar-binding functions are known, but who could expect such a situation in the twentieth century. Another important indication of the overview is that many of the newly discovered lectins (and carbohydrate-binding domains) are rather rare member(s) of the protein families to which they belong. Such examples include the POMGnT1 stem region (Kuwabara et al., Proc Natl Acad Sci U S A, 113:9280–9285. 2016). Hence, these observations led me to a hypothesis that a sugar-binding domain can be generated from any protein scaffold either by nature or artificial engineering, although this has never been achieved (Hirabayashi and Arai, Interface Focus, 9: 201800682019, 2019). To know the answer to the above question, it is essential to understand both common and unique properties of all the known lectin proteins as well as lectin domains, parts of functional proteins to support them. Such lectin proteins and domains include conventional plant lectins, endogenous animal lectins, and lectins from diverse microorganisms. Also, an important fact is that most viruses are equipped with carbohydrate-binding domains for host cell targeting, or hemagglutinin in case of influenza virus. Unfortunately, many of these virus domains have not been understood as lectin. Without doubt, it is a time of learning once more about lectins from a more systematic viewpoint to integrate our intellect. Apparent basis for this achievement is on how to purify and analyze lectins from all kinds of the living organisms. The present edition includes 53 standard protocol chapters and 6 overview chapters, which cover a wide range of lectins from mammals (18), nonmammalian animals (12), plants (7), algae (1), fungi (6), protists (2), bacteria (6), and viruses (7; figures in parentheses are the numbers of relevant chapters). I hope the book will become an essential resource for scientists who wish to learn more about life systems, because they are composed of cells with diverse glycoconjugates, of which biological meaning is decoded by a series of lectins in time. Takamatsu, Ibaraki, Japan

Jun Hirabayashi

v

Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v xiii

1 Structural Database for Lectins and the UniLectin Web Platform . . . . . . . . . . . . . 1 Franc¸ois Bonnardel, Serge Perez, Fre´de´rique Lisacek, and Anne Imberty 2 Purification of Sugar-Binding Peptides from L-Type Lectins . . . . . . . . . . . . . . . . . 15 Kazuo Yamamoto 3 Recombinant Expression and Purification of Animal Intracellular L-Type Lectins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Tadashi Satoh and Koichi Kato 4 Frontal Affinity Chromatography: A Highly Suitable Retardation Phenomenon-Based Research Tool for Analyzing Weak Interactions Between Biomolecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Kenichi Kasai 5 Metazoan Soluble β-Galactoside-Binding Lectins, Galectins: Methods for Purification, Characterization of Their Carbohydrate-Binding Specificity, and Probing Their Ligands . . . . . . . . . . . . . . . . 39 Guillaume St-Pierre, Ann Rancourt, and Sachiko Sato 6 Expression, S-Nitrosylation, and Measurement of S-Nitrosylation Ratio of Recombinant Galectin-2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 Mayumi Tamura and Yoichiro Arata 7 Expression and Purification of Full-Length and Domain-Fragment Recombinant Pentraxin 3 (PTX3) Proteins from Mammalian and Bacterial Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 Kenji Daigo and Takao Hamakubo 8 Identification of Siglec Cis-Ligands by Proximity Labeling . . . . . . . . . . . . . . . . . . . 75 Amin Alborzian Deh Sheikh, Chizuru Akatsu, and Takeshi Tsubata 9 Preparation of Recombinant Siglecs and Identification of Their Ligands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Lan-Yi Chang, Penk Yeir Low, Deepa Sridharan Kaia Gerlovin, and Takashi Angata 10 Purification, Quantification, and Functional Analysis of Collectins . . . . . . . . . . . . 99 Katsuki Ohtani and Nobutaka Wakamiya 11 Selectin-Binding Assay by Flow Cytometry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Keiichiro Sakuma and Reiji Kannagi 12 Direct Binding Analysis Between C-Type Lectins and Glycans Using Immunoglobulin Receptor Fusion Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . 119 Miyuki Watanabe, Zakaria Omahdi, and Sho Yamasaki

vii

viii

13

14

15

16

17

18

19

20 21

22

23 24

25

26

Contents

Expression and Characterization of Hyaluronan-Binding Protein Involved in Hyaluronan Depolymerization: HYBID, Alias KIAA1199 and CEMIP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Susana de Vega, Hiroyuki Yoshida, and Yasunori Okada Paracoccin: Purification and Validation of Its Lectin and Enzymatic Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nayla de Souza Pitangui, Fabrı´cio Freitas Fernandes Relber Aguiar Gonc¸ales, and Maria Cristina Roque-Barreira In Vitro Mannosidase Assay of EDEMs: ER Degradation-Enhancing α-Mannosidase-Like Proteins. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nobuko Hosokawa Galactose-Specific, Hemolytic Lectin CEL-III from Cucumaria echinata . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tomomitsu Hatakeyama Simple and Rapid Detection of Glycoforms by “Lectin Inhibition” Assay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Kyoka Hoshi, Yasuhiro Hashimoto, and Hiromi Ito Glycoform-Specific Visualization in Immunohistochemistry by “Lectin Inhibition”. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hiromi Ito, Kyoka Hoshi, Yasuhiro Hashimoto, and Takashi Honda Botulinum Hemagglutinin: Critical Protein for Adhesion and Absorption of Neurotoxin Complex in Host Intestine . . . . . . . . . . . . . . . . . . . Sho Amatsu and Yukako Fujinaga Functional Analysis of Botulinum Hemagglutinin (HA) . . . . . . . . . . . . . . . . . . . . . Takuhiro Matsumura and Yukako Fujinaga Purification and Functional Characterization of the Effects on Cell Signaling of Mytilectin: A Novel β-Trefoil Lectin from Marine Mussels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yuki Fujii, S. M. Abe Kawsar, Imtiaj Hasan, Hideaki Fujita Marco Gerdol, and Yasuhiro Ozeki Lectin-Type Ubiquitin Ligase Subunits: Fbs Proteins and Their Applications for Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yukiko Yoshida F-Type Lectins: Structure, Function, and Evolution. . . . . . . . . . . . . . . . . . . . . . . . . Gerardo R. Vasta and Chiguang Feng Purification and Biochemical Characterization of Selected F-Type Lectins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chiguang Feng and Gerardo R. Vasta LecA (PA-IL): A Galactose-Binding Lectin from Pseudomonas aeruginosa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sakonwan Kuhaudomlarp, Emilie Gillon, Annabelle Varrot, and Anne Imberty Glycan Recognition and Application of P-Type Lectins . . . . . . . . . . . . . . . . . . . . . . Kei Kiriyama and Kohji Itoh

129

139

151

159

165

173

183 191

201

215 225

241

257

267

Contents

27 28 29

30 31 32

33

34

35

36

37

38

39

40

Purification and Assays of Tachylectin-5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Shun-ichiro Kawabata and Toshio Shibata Preparation of Soluble Malectin and Its Tetramer . . . . . . . . . . . . . . . . . . . . . . . . . . . Sheng-Ying Qin, Dan Hu, and Kazuo Yamamoto Calnexin/Calreticulin and Assays Related to N-Glycoprotein Folding In Vitro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yoshito Ihara, Midori Ikezaki, Maki Takatani, and Yukishige Ito Purification and Assays of Tachylectin-2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Shun-ichiro Kawabata and Toshio Shibata Purification and Assays of Tachycitin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Shun-ichiro Kawabata and Toshio Shibata Methods for Purifying Datura stramonium Agglutinin and Producing Recombinant Agglutinin Protein in a Heterologous Plant Host . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Suguru Oguri ZG16p, an Animal Homologue of Plant β-Prism Fold Lectins: Purification Methods of Natural and Recombinant ZG16p and Inhibition Assay of Cancer Cell Growth Using ZG16p . . . . . . . . . . . . . . . . . . Akiko Mito, Kaori Kumazawa-Inoue, and Kyoko Kojima-Aikawa ArtinM: Purification and Evaluation of Biological Activities . . . . . . . . . . . . . . . . . . Thiago Aparecido da Silva, Patrı´cia Kellen Martins Oliveira-Brito, Sandra Maria de Oliveira Thomaz, and Maria Cristina Roque-Barreira Bacterial Expression of Rhamnose-Binding Lectin from Catfish Eggs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Shigeki Sugawara, Takeo Tatsuta, and Masahiro Hosono A Bioassay for Determining Symbiotic Zooxanthellae Shape Control Using Lectin SLL-2 from the Octocoral Sinularia lochmodes . . . . . . . . . Mitsuru Jimbo, Ryota Takeuchi, and Mayu Yoshino MIC4 from Toxoplasma gondii: A Lectin Acting as a Toll-Like Receptor Agonist. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fla´via Costa Mendonc¸a-Natividade, Rafael Ricci-Azevedo, and Maria Cristina Roque-Barreira Production and Characterization of MIC1: A Lectin from Toxoplasma gondii. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fla´via Costa Mendonc¸a-Natividade, Rafael Ricci-Azevedo Sandra Maria de Oliveira Thomaz, and Maria Cristina Roque-Barreira Affinity Labeling and Purification of Plant Chitin-Binding LysM Receptor with Chitin Octasaccharide Derivatives. . . . . . . . . . . . . . . . . . . . . . Tomonori Shinya, Naoto Shibuya, and Hanae Kaku Purification of GNA-Related Lectins from Natural Sources . . . . . . . . . . . . . . . . . . Els J. M. Van Damme

ix

277 285

295

309 317

325

339 349

359

369

379

391

401 413

x

41

42

43

44

45

46

47

48

49

50

51

52

53 54

Contents

Expression, Purification, and Applications of the Recombinant Lectin PVL from Psathyrella velutina Specific for Terminal N-Acetyl-Glucosamine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Oriane Machon and Annabelle Varrot Yeast Flocculin: Methods for Quantitative Analysis of Flocculation in Yeast Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hiromi Maekawa and Kaoru Takegawa Pleurotus cornucopiae Mycelial Lectin (PCL-M): Purification and Detection of the Activity on Mycelial Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . Suguru Oguri Expression and Purification of a Human Pluripotent Stem Cell-Specific Lectin Probe, rBC2LCN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hiroaki Tateno Preparation of Fluorescent Recombinant Shiga Toxin B Subunit and Its Application to Flow Cytometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Toshiyuki Yamaji LecB, a High Affinity Soluble Fucose-Binding Lectin from Pseudomonas aeruginosa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Emilie Gillon, Annabelle Varrot, and Anne Imberty Sialoglycovirology of Lectins: Sialyl Glycan Binding of Enveloped and Non-enveloped Viruses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nongluk Sriwilaijaroen and Yasuo Suzuki Hemagglutinin Inhibitors are Potential Future Anti-Influenza Drugs for Mono- and Combination Therapies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nongluk Sriwilaijaroen and Yasuo Suzuki Preparation and Detection of Glycan-Binding Activity of Influenza Virus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Shin-ichi Nakakita, Nongluk Sriwilaijaroen, Yasuo Suzuki, and Jun Hirabayashi Screening for Components/Compounds with Anti-Rotavirus Activity: Detection of Interaction Between Viral Spike Proteins and Glycans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Keita Yamada, Junko Nio-Kobayashi, and Mizuho Inagaki ELISA-Based Methods to Detect and Quantify Norovirus Virus-Like Particle Attachment to Histo-Blood Group Antigens. . . . . . . . . . . . . . Haruko Shirato FAM3B/PANDER-Like Carbohydrate-Binding Domain in a Glycosyltransferase, POMGNT1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hiroshi Manya, Naoyuki Kuwabara, Ryuichi Kato, and Tamao Endo Mannose-Specific Oyster Lectin CGL1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hideaki Unno and Tomomitsu Hatakeyama Receptor-Binding Assays of Enterovirus D68. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tadatsugu Imamura, Michiko Okamoto, and Hitoshi Oshitani

421

437

445

453

463

475

483

547

567

585

597

609 621 629

Contents

Large-Scale Expression and Purification of Mumps Virus Hemagglutinin-Neuraminidase for Structural Analyses and Glycan-Binding Assays. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Marie Kubota and Takao Hashiguchi 56 Purification of AJLec: A Novel Galactose-Specific Lectin from the Sea Anemone Anthopleura japonica. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hideaki Unno 57 Annexin Lectins: Ca2+-Dependent Heparin-Binding Activity, Phosphatidylserine-Binding Activity, and Anticoagulant Activity . . . . . . . . . . . . . . Moeka Nakayama, Miyuki Tsunooka-Ohta, and Kyoko Kojima-Aikawa 58 Expression, Purification, and Functional Characterization of Tectonin 2 from Laccaria bicolor: A Six-Bladed Beta-Propeller Lectin Specific for O-Methylated Glycans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ¨ nzler, Therese Wohlschlager, Alexander Titz, Markus Ku and Annabelle Varrot 59 The OAAH Family: Anti-Influenza Virus Lectins . . . . . . . . . . . . . . . . . . . . . . . . . . . Yuichiro Sato, Makoto Hirayama, Kinjiro Morimoto, and Kanji Hori Correction to: Hemagglutinin Inhibitors are Potential Future Anti-Influenza Drugs for Mono- and Combination Therapies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xi

55

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

641

653

661

669

683

C1 695

Contributors CHIZURU AKATSU • Department of Immunology, Medical Research Institute, Tokyo Medical and Dental University, Tokyo, Japan AMIN ALBORZIAN DEH SHEIKH • Department of Immunology, Medical Research Institute, Tokyo Medical and Dental University, Tokyo, Japan SHO AMATSU • Department of Bacteriology, Graduate School of Medical Sciences, Kanazawa University, Ishikawa, Japan; Department of Forensic Medicine and Pathology, Graduate School of Medical Sciences, Kanazawa University, Ishikawa, Japan TAKASHI ANGATA • Institute of Biological Chemistry, Academia Sinica, Taipei, Taiwan YOICHIRO ARATA • Faculty of Pharma-Science, Teikyo University, Tokyo, Japan FRANC¸OIS BONNARDEL • Univ. Grenoble Alpes, CNRS, CERMAV, Grenoble, France; Swiss Institute of Bioinformatics, Geneva, Switzerland; Computer Science Department, UniGe, Geneva, Switzerland LAN-YI CHANG • Institute of Biological Chemistry, Academia Sinica, Taipei, Taiwan KENJI DAIGO • Department of Protein-Protein Interaction Research, Institute for Advanced Medical Sciences, Nippon Medical School, Kawasaki, Kanagawa, Japan THIAGO APARECIDO DA SILVA • Departamento de Biologia Celular e Molecular e Bioagentes Patogeˆnicos, Faculdade de Medicina de Ribeira˜o Preto, Universidade de Sa˜o Paulo, Ribeira˜o Preto, SP, Brazil SANDRA MARIA DE OLIVEIRA THOMAZ • Departamento de Biologia Celular e Molecular e Bioagentes Patogeˆnicos, Faculdade de Medicina de Ribeira˜o Preto, Universidade de Sa˜o Paulo, Ribeira˜o Preto, SP, Brazil NAYLA DE SOUZA PITANGUI • Department of Cellular and Molecular Biology, Ribeira˜o Preto School of Medicine, University of Sa˜o Paulo, Ribeira˜o Preto, Sa˜o Paulo, Brazil SUSANA DE VEGA • Department of Pathophysiology for Locomotive and Neoplastic Diseases, Juntendo University Graduate School of Medicine, Bunkyo-ku, Tokyo, Japan TAMAO ENDO • Molecular Glycobiology, Research Team for Mechanism of Aging, Tokyo Metropolitan Geriatric Hospital and Institute of Gerontology, Itabashi-ku, Tokyo, Japan CHIGUANG FENG • Department of Microbiology and Immunology, University of Maryland School of Medicine, UMB, and Institute of Marine and Environmental Technology, Columbus Center, Baltimore, MD, USA FABRI´CIO FREITAS FERNANDES • Department of Cellular and Molecular Biology, Ribeira˜o Preto School of Medicine, University of Sa˜o Paulo, Ribeira˜o Preto, Sa˜o Paulo, Brazil YUKI FUJII • Graduate School of Pharmaceutical Sciences, Nagasaki International University, Sasebo, Japan YUKAKO FUJINAGA • Department of Bacteriology, Graduate School of Medical Sciences, Kanazawa University, Kanazawa, Ishikawa, Japan HIDEAKI FUJITA • Graduate School of Pharmaceutical Sciences, Nagasaki International University, Sasebo, Japan MARCO GERDOL • Department of Life Sciences, University of Trieste, Trieste, Italy KAIA GERLOVIN • Institute of Biological Chemistry, Academia Sinica, Taipei, Taiwan EMILIE GILLON • Univ. Grenoble Alpes, CNRS, CERMAV, Grenoble, Alpes, France RELBER AGUIAR GONC¸ALES • Department of Cellular and Molecular Biology, Ribeira˜o Preto School of Medicine, University of Sa˜o Paulo, Ribeira˜o Preto, Sa˜o Paulo, Brazil

xiii

xiv

Contributors

TAKAO HAMAKUBO • Department of Protein-Protein Interaction Research, Institute for Advanced Medical Sciences, Nippon Medical School, Kawasaki, Kanagawa, Japan IMTIAJ HASAN • Department of Biochemistry and Molecular Biology, Faculty of Science, University of Rajshahi, Rajshahi, Bangladesh TAKAO HASHIGUCHI • Department of Virology, Faculty of Medicine, Kyushu University, Fukuoka, Japan YASUHIRO HASHIMOTO • Department of Human Life Science, Fukushima Medical University School of Nursing, Fukushima, Japan TOMOMITSU HATAKEYAMA • Graduate School of Engineering, Nagasaki University, Nagasaki, Japan JUN HIRABAYASHI • Cellular and Molecular Biotechnology Research Institute, National Institute of Advanced Industrial Science and Technology, Tsukuba, Ibaraki, Japan MAKOTO HIRAYAMA • Graduate School of Integrated Sciences for Life, Hiroshima University, Higashi-Hiroshima, Japan TAKASHI HONDA • Department of Human Life Science, Fukushima Medical University School of Nursing, Fukushima, Japan KANJI HORI • Graduate School of Integrated Sciences for Life, Hiroshima University, Higashi-Hiroshima, Japan KYOKA HOSHI • Department of Biochemistry, Fukushima Medical University, Fukushima, Japan NOBUKO HOSOKAWA • Laboratory of Molecular and Cellular Biology, Institute for Frontier Life and Medical Sciences, Kyoto University, Sakyo-ku, Kyoto, Japan MASAHIRO HOSONO • Division of Cell Recognition Study, Institute of Molecular Biomembrane and Glycobiology, Tohoku Medical and Pharmaceutical University, Sendai, Japan DAN HU • Institute of Traditional Chinese Medicine & Natural Products, College of Pharmacy, Jinan University, Guangzhou, China YOSHITO IHARA • Department of Biochemistry, Wakayama Medical University, Wakayama, Japan MIDORI IKEZAKI • Department of Biochemistry, Wakayama Medical University, Wakayama, Japan TADATSUGU IMAMURA • Center for Postgraduate Education and Training, National Center for Child Health and Development, Tokyo, Japan ANNE IMBERTY • Univ. Grenoble Alpes, CNRS, CERMAV, Grenoble, Alpes, France MIZUHO INAGAKI • Faculty of Applied Biological Sciences, Gifu University, Gifu, Japan HIROMI ITO • Department of Biochemistry, Fukushima Medical University, Fukushima, Japan KOHJI ITOH • Department of Medicinal Biotechnology, Institute for Medicinal Research, Graduate School of Pharmaceutical Sciences, Tokushima University, Tokushima, Japan YUKISHIGE ITO • Synthetic Cellular Chemistry Laboratory, RIKEN, Saitama, Japan MITSURU JIMBO • Department of Marine Biosciences, Kitasato University, Sagamihara, Kanagawa, Japan HANAE KAKU • Department of Life Sciences, School of Agriculture, Meiji University, Kawasaki, Kanagawa, Japan REIJI KANNAGI • The Institute of Biomedical Sciences (IBMS), Academia Sinica, Taipei, Taiwan KENICHI KASAI • Faculty of Pharmaceutical Sciences, Teikyo University, Itabashi City, Tokyo, Japan

Contributors

xv

KOICHI KATO • Graduate School of Pharmaceutical Sciences, Nagoya City University, Nagoya, Japan; Exploratory Research Center on Life and Living Systems (ExCELLS), National Institutes of Natural Sciences, Okazaki, Japan; Institute for Molecular Science, National Institutes of Natural Sciences, Okazaki, Japan RYUICHI KATO • Structural Biology Research Center, Institute of Materials Structure Science, High Energy Accelerator Research Organization, Tsukuba, Ibaraki, Japan SHUN-ICHIRO KAWABATA • Department of Biology, Faculty of Science, Kyushu University, Fukuoka, Japan S. M. ABE KAWSAR • Department of Chemistry, Faculty of Science, University of Chittagong, Chittagong, Bangladesh KEI KIRIYAMA • Department of Medicinal Biotechnology, Institute for Medicinal Research, Graduate School of Pharmaceutical Sciences, Tokushima University, Tokushima, Japan KYOKO KOJIMA-AIKAWA • Natural Science Division, Faculty of Science, Ochanomizu University, Tokyo, Japan; Natural Science Division, Faculty of Core Research, Ochanomizu University, Tokyo, Japan MARIE KUBOTA • Department of Virology, Faculty of Medicine, Kyushu University, Fukuoka, Japan SAKONWAN KUHAUDOMLARP • Univ. Grenoble Alpes, CNRS, CERMAV, Grenoble, Alpes, France KAORI KUMAZAWA-INOUE • Natural Science Division, Faculty of Science, Ochanomizu University, Tokyo, Japan MARKUS KU¨NZLER • Department of Biology, Institute of Microbiology, Swiss Federal Institute of Technology (ETH), Zu¨rich, Switzerland NAOYUKI KUWABARA • Structural Biology Research Center, Institute of Materials Structure Science, High Energy Accelerator Research Organization, Tsukuba, Ibaraki, Japan FRE´DE´RIQUE LISACEK • Swiss Institute of Bioinformatics, Geneva, Switzerland; Computer Science Department, UniGe, Geneva, Switzerland; Section of Biology, UniGe, Geneva, Switzerland PENK YEIR LOW • Institute of Biological Chemistry, Academia Sinica, Taipei, Taiwan ORIANE MACHON • Univ. Grenoble Alpes, CNRS, CERMAV, Grenoble, France HIROMI MAEKAWA • Center for Promotion of International Education and Research, Faculty of Agriculture, Kyushu University, Fukuoka, Japan HIROSHI MANYA • Molecular Glycobiology, Research Team for Mechanism of Aging, Tokyo Metropolitan Geriatric Hospital and Institute of Gerontology, Itabashi-ku, Tokyo, Japan TAKUHIRO MATSUMURA • Department of Bacteriology, Graduate School of Medical Sciences, Kanazawa University, Kanazawa, Ishikawa, Japan FLA´VIA COSTA MENDONC¸A-NATIVIDADE • Laboratory of Immunochemistry and Glycobiology, Department of Cell and Molecular Biology and Pathogenic Bioagents, Ribeira˜o Preto Medical School, University of Sa˜o Paulo (FMRP/USP), Ribeira˜o Preto, SP, Brazil AKIKO MITO • Natural Science Division, Faculty of Science, Ochanomizu University, Tokyo, Japan KINJIRO MORIMOTO • Faculty of Pharmacy, Yasuda Women’s University, Hiroshima, Japan SHIN-ICHI NAKAKITA • Department of Functional Glycomics, Life Science Research Center, Kagawa University, Kita-gun, Kagawa, Japan MOEKA NAKAYAMA • Graduate School of Humanities and Sciences, Ochanomizu University, Tokyo, Japan JUNKO NIO-KOBAYASHI • Faculty of Medicine and Graduate School of Medicine, Hokkaido University, Sapporo, Japan

xvi

Contributors

SUGURU OGURI • Department of Northern Biosphere Agriculture, Faculty of Bioindustry, Tokyo University of Agriculture, Abashiri, Hokkaido, Japan KATSUKI OHTANI • Department of Clinical Nutrition, Rakuno Gakuen University, Ebetsu, Hokkaido, Japan YASUNORI OKADA • Department of Pathophysiology for Locomotive and Neoplastic Diseases, Juntendo University Graduate School of Medicine, Bunkyo-ku, Tokyo, Japan MICHIKO OKAMOTO • Department of Virology, Tohoku University Graduate School of Medicine, Sendai, Miyagi, Japan PATRI´CIA KELLEN MARTINS OLIVEIRA-BRITO • Departamento de Biologia Celular e Molecular e Bioagentes Patogeˆnicos, Faculdade de Medicina de Ribeira˜o Preto, Universidade de Sa˜o Paulo, Ribeira˜o Preto, SP, Brazil ZAKARIA OMAHDI • Department of Molecular Immunology, Research Institute for Microbial Diseases, Osaka University, Osaka, Japan; Laboratory of Molecular Immunology, Immunology Frontier Research Center (WPI-IFReC), Osaka University, Osaka, Japan HITOSHI OSHITANI • Department of Virology, Tohoku University Graduate School of Medicine, Sendai, Miyagi, Japan YASUHIRO OZEKI • Department of Life and Environmental System Science, School of Sciences, Yokohama City University, Yokohama, Japan SERGE PEREZ • Univ. Grenoble Alpes, CNRS, CERMAV, Grenoble, France SHENG-YING QIN • Clinical Experimental Center, First Affiliated Hospital of Jinan University, Guangzhou, China; Clinical Medical Research Institute, Jinan University, Guangzhou, China ANN RANCOURT • Glycobiology and Bioimaging Laboratory, Research Centre for Infectious Diseases, Faculty of Medicine, Laval University and GlycoNET, Quebec, Canada RAFAEL RICCI-AZEVEDO • Laboratory of Immunochemistry and Glycobiology, Department of Cell and Molecular Biology and Pathogenic Bioagents, Ribeira˜o Preto Medical School, University of Sa˜o Paulo (FMRP/USP), Ribeira˜o Preto, SP, Brazil MARIA CRISTINA ROQUE-BARREIRA • Laboratory of Immunochemistry and Glycobiology, Department of Cell and Molecular Biology and Pathogenic Bioagents, Ribeira˜o Preto School of Medicine, University of Sa˜o Paulo (FMRP/USP), Ribeira˜o Preto, SP, Brazil KEIICHIRO SAKUMA • Division of Pathophysiology, Aichi Cancer Center Research Institute, Nagoya, Japan SACHIKO SATO • Glycobiology and Bioimaging Laboratory, Research Centre for Infectious Diseases, Faculty of Medicine, Laval University and GlycoNET, Quebec, Canada YUICHIRO SATO • Faculty of Pharmacy, Yasuda Women’s University, Hiroshima, Japan TADASHI SATOH • Graduate School of Pharmaceutical Sciences, Nagoya City University, Nagoya, Japan TOSHIO SHIBATA • Department of Biology, Faculty of Science, Kyushu University, Fukuoka, Japan NAOTO SHIBUYA • Department of Life Sciences, School of Agriculture, Meiji University, Kawasaki, Kanagawa, Japan TOMONORI SHINYA • Institute of Plant Science and Resources, Okayama University, Kurashiki, Japan HARUKO SHIRATO • Department of Virology II, National Institute of Infectious Diseases, Musashi-Murayama, Tokyo, Japan DEEPA SRIDHARAN • Institute of Biological Chemistry, Academia Sinica, Taipei, Taiwan

Contributors

xvii

NONGLUK SRIWILAIJAROEN • Department of Preclinical Sciences, Faculty of Medicine, Thammasat University, Pathumthani, Thailand; College of Life and Health Sciences, Chubu University, Kasugai, Aichi, Japan GUILLAUME ST-PIERRE • Glycobiology and Bioimaging Laboratory, Research Centre for Infectious Diseases, Faculty of Medicine, Laval University and GlycoNET, Quebec, Canada SHIGEKI SUGAWARA • Division of Cell Recognition Study, Institute of Molecular Biomembrane and Glycobiology, Tohoku Medical and Pharmaceutical University, Sendai, Japan YASUO SUZUKI • College of Life and Health Sciences, Chubu University, Kasugai, Aichi, Japan MAKI TAKATANI • Synthetic Cellular Chemistry Laboratory, RIKEN, Saitama, Japan KAORU TAKEGAWA • Department of Bioscience & Biotechnology, Faculty of Agriculture, Kyushu University, Fukuoka, Japan RYOTA TAKEUCHI • Department of Marine Biosciences, Kitasato University, Sagamihara, Kanagawa, Japan; AIST-Osaka University Advanced Photonics and Biosensing, Open Innovation Laboratory, AIST, Photonics Center Osaka University P3, Suita, Osaka, Japan MAYUMI TAMURA • Faculty of Pharma-Science, Teikyo University, Tokyo, Japan HIROAKI TATENO • Biotechnology Research Institute for Drug Discovery (BRD), National Institute of Advanced Industrial Science and Technology, Tsukuba, Ibaraki, Japan; Life Science Center for Survival Dynamics, Tsukuba Advanced Research Alliance, University of Tsukuba, Tsukuba, Ibaraki, Japan; JST PRESTO, Tsukuba, Ibaraki, Japan TAKEO TATSUTA • Division of Cell Recognition Study, Institute of Molecular Biomembrane and Glycobiology, Tohoku Medical and Pharmaceutical University, Sendai, Japan ALEXANDER TITZ • Chemical Biology of Carbohydrates, Helmholtz Institute for Pharmaceutical Research Saarland (HIPS), Helmholtz Centre for Infection Research, Saarbru¨cken, Germany; Deutsches Zentrum fu¨r Infektionsforschung (DZIF), HannoverBraunschweig, Germany; Department of Pharmacy, Saarland University, Saarbru¨cken, Germany TAKESHI TSUBATA • Department of Immunology, Medical Research Institute, Tokyo Medical and Dental University, Tokyo, Japan MIYUKI TSUNOOKA-OHTA • Graduate School of Humanities and Sciences, Ochanomizu University, Tokyo, Japan HIDEAKI UNNO • Graduate School of Engineering, Nagasaki University, Nagasaki, Japan; Organization for Marine Science and Technology, Nagasaki University, Nagasaki, Japan ELS J. M. VAN DAMME • Laboratory of Glycobiology and Biochemistry, Department of Biotechnology, Faculty of Bioscience Engineering, Ghent University, Gent, Belgium ANNABELLE VARROT • Univ. Grenoble Alpes, CNRS, CERMAV, Grenoble, Alpes, France GERARDO R. VASTA • Department of Microbiology and Immunology, University of Maryland School of Medicine, UMB, Institute of Marine and Environmental Technology, Columbus Center, Baltimore, MD, USA NOBUTAKA WAKAMIYA • Department of Medicine and Physiology, Rakuno Gakuen University, Hokkaido, Japan MIYUKI WATANABE • Department of Molecular Immunology, Research Institute for Microbial Diseases, Osaka University, Osaka, Japan; Laboratory of Molecular Immunology, Immunology Frontier Research Center (WPI-IFReC), Osaka University, Osaka, Japan THERESE WOHLSCHLAGER • Department of Biology, Institute of Microbiology, Swiss Federal Institute of Technology (ETH), Zu¨rich, Switzerland; Department of Biosciences, Bioanalytical Research Labs, University of Salzburg, Salzburg, Austria KEITA YAMADA • Faculty of Pharmacy, Osaka Ohtani University, Tondabayashi, Japan

xviii

Contributors

TOSHIYUKI YAMAJI • Department of Biochemistry and Cell Biology, National Institute of Infectious Diseases, Tokyo, Japan KAZUO YAMAMOTO • Department of Integrated Biosciences, Graduate School of Frontier Sciences, University of Tokyo, Kashiwa, Chiba, Japan SHO YAMASAKI • Department of Molecular Immunology, Research Institute for Microbial Diseases, Osaka University, Osaka, Japan; Laboratory of Molecular Immunology, Immunology Frontier Research Center (WPI-IFReC), Osaka University, Osaka, Japan HIROYUKI YOSHIDA • Department of Biological Science Research, Kao Corporation, Chu o, ¯¯ Tokyo, Japan YUKIKO YOSHIDA • Ubiquitin Project, Tokyo Metropolitan Institute of Medical Science, Tokyo, Japan MAYU YOSHINO • Department of Marine Biosciences, Kitasato University, Sagamihara, Kanagawa, Japan

Chapter 1 Structural Database for Lectins and the UniLectin Web Platform Franc¸ois Bonnardel, Serge Perez, Fre´de´rique Lisacek, and Anne Imberty Abstract The search for new biomolecules requires a clear understanding of biosynthesis and degradation pathways. This view applies to most metabolites as well as other molecule types such as glycans whose repertoire is still poorly characterized. Lectins are proteins that recognize specifically and interact noncovalently with glycans. This particular class of proteins is considered as playing a major role in biology. Glycan-binding is based on multivalence, which gives lectins a unique capacity to interact with surface glycans and significantly contribute to cell–cell recognition and interactions. Lectins have been studied for many years using multiple technologies and part of the resulting information is available online in databases. Unfortunately, the connectivity of these databases with the most popular omics databases (genomics, proteomics, and glycomics) remains limited. Moreover, lectin diversity is extended and requires setting out a flexible classification that remains compatible with new sequences and 3D structures that are continuously released. We have designed UniLectin as a new insight into the knowledge of lectins, their classification, and their biological role. This platform encompasses UniLectin3D, a curated database of lectin 3D structures that follows a periodically updated classification, a set of comparative and visualizing tools and gradually released modules dedicated to specific lectins predicted in sequence databases. The second module is PropLec, focused on β-propeller lectin prediction in all species based on five distinct family profiles. This chapter describes how UniLectin can be used to explore the diversity of lectins, their 3D structures, and associated functional information as well as to perform reliable predictions of β-propeller lectins. Key words Lectin, Carbohydrate-binding protein, Database, Classification, Sequence, 3D structure, Profile-based prediction

1

Introduction Lectins are oligomeric proteins that bind mono- and oligosaccharides reversibly and specifically while displaying no catalytic or immunological activity [1]. Those complex carbohydrates (also referred to as glycans) occur in the form of single molecules, or as part of glycoconjugates (glycoproteins and glycolipids). They constitute the most abundant class of biomolecules on Earth. Complex carbohydrates are built for high-density bio-coding, the information being carried and encoded in their 3D-structures and

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_1, © Springer Science+Business Media, LLC, part of Springer Nature 2020

1

2

Franc¸ois Bonnardel et al.

sometimes in their dynamics. Lectins are powerful macromolecular tools to decipher the high-density bio-encoding of complex carbohydrates. Along with a high specificity, lectins exhibit diverse architectures and modes of multivalence relating to their function. Some lectins exhibit “architectural multivalence,” consisting of one macromolecular structure with several equivalent carbohydrate recognition domains. Other lectins, such as adhesins, are membranebound and include only one carbohydrate recognition domain, attached to fimbriae or flagella tethered to the cell surface. Several fimbrial structures clustered together at the cell surface also give a multivalent presentation of the lectins in the extracellular environment. All of these recognition processes play important roles in fertilization, embryogenesis, inflammation, metastasis, and parasite–symbiote recognition in microbes, invertebrates, plants, and vertebrates. In 1984, Gallagher [2] made a first attempt to classify and establish a nomenclature of lectins. Then, an overview of seven families of plant lectins [3] and a classification of animal lectins [4] were proposed. Historically, these classifications were based on the discovery of new lectins supported by the resolution of their 3D structures. The accumulation of data revealed the occurrence of similar or closely similar protein motifs across several kingdoms. Consequently, a species-independent classification based on lectin sequence and 3D structure appeared more relevant. Such an approach was proposed through the association of 3D features and Pfam domains [5]. Unfortunately, the approximate definition of Pfam domains [6] that are rarely specific to lectin sequences not only reduces the accuracy of classes but also precludes the classification of newly discovered lectins. In the same vein, a classification of fucose-binding lectins based on Pfam domains was suggested by H. Makyio and R. Kato in 2016 [7]. The following examples illustrate the possible sources of ambiguity. Using “lectin” and “glycan binding” as keywords to search the UniProtKB database [8] returns roughly 159,000 and 230,000 distinct proteins, respectively (UniProt 2019-03 release). Searching the Protein Data Bank (PDB) through PDBe [9] with the same keywords returns 3634 and 121 structures, respectively (PDB 2019-04-24 release). A close examination of the results indicates that some of these proteins are in fact glycoproteins that bind glycan covalently, and others are enzymes that recognize and modify glycans. This highlights the need for a robust classification of lectins and their ligands based on a large panel of finely curated or filtered information extracted from genomic data as well as 3D structures and their inter-atomic features.

UniLectin Structural Database

2

3

Existing Lectin Databases Due to the crucial role of lectins in cell recognition and interactions, a large amount of experimental information has been published. They encompass peptide sequences of the lectin domain as well as full protein sequences, 3D structures derived from diffraction studies that possibly contain the interacting carbohydrate, and glycan arrays that reflect the specificity for different glycans and their target affinity. Such a large body of information needs to be integrated and organized in a public database that is interoperable with other relevant omics databases. Several initiatives that we now briefly summarize have been launched to address this question. Glyco3D [10] includes a family of databases covering the 3D features of monosaccharides, disaccharides, oligosaccharides, polysaccharides, glycosyltransferases, lectins, monoclonal antibodies, and glycosaminoglycan-binding proteins that have been developed with nonproprietary software and are freely available to the scientific community (http://glyco3d.cermav.cnrs.fr). The Lectin Frontier Database (LfDB: https://acgg.asia/lfdb2/) [11] contains 400 lectins with curated information including lectin name, Pfam family, name of the interacting glycan, species, fold, PDB 3D structure, protein sequence, and reference. It is now integrated in GlyCosmos, a newly developed portal (https://glycosmos.org/ lectins/) for accessing and exploring knowledge in glycobiology. LectinDB (http://proline.physics.iisc.ernet.in/lectindb/) [12] is an annotated database mainly of plant lectins. It can be searched based on their species, accession number, lectin domain, fold, PDB code, and interacting glycan; a protein sequence can be compared to the lectins in lectinDB; an overview of lectin available in each species is also available; each lectin is associated with ataxon and a UniProt entry. SugarBindDB (http://sugarbind.expasy.org/) [13] provides integrated information on pathogen lectins and their corresponding glycan targets. This curated database describes either bacterial or viral proteins and details their binding specificity. Each ligand can be matched to full glycan structures of GlyConnect [14] and their associated glycoproteins to reveal potential glycanmediated interactions between pathogen lectins and host glycoproteins. When 3D structures are available from PDB, they can be visualized with the LiteMol software [15]. It also includes affinity data when available. Despite this high level of data and tool integration, SugarBindDB remains focused on pathogenic virus and bacterial species. The databases of the Consortium for Functional Glycomics (CFG resources: http://www.functionalglycomics.org/static/con sortium/resources.shtml) [16] include a repository of glycan array data used to measure the specificity of a lectin to multiple glycans.

4

Franc¸ois Bonnardel et al.

These databases are unfortunately no longer maintained. The latest related resource is GLAD (https://glycotoolkit.com/GLAD/) [17], a web-based tool designed to visualize and analyze glycan microarray data, providing a list of available glycan array for lectins. Most of these lectin-dedicated databases lack interaction details with proteins and protein families [8,18,19], protein structure classification databases [20, 21], glycans or protein glycan interaction analysis tools [22, 23], and do not necessarily comply with glycan textual and visual representations now commonly accepted in the glycoscience community [24]. Such spread out information led us to design and implement the UniLectin platform (https:// unilectin.eu) in order to address the data integration and classification issues. The present chapter describes the current content and the possible use of the UniLectin comparative, predictive, and visualizing tools. The platform consists of modules; the main one being UniLectin3D a curated database of three-dimensional structures of lectins, as established mainly from crystallographic methods. It includes a classification of lectins, along with the knowledge of interacting glycans. Based on this information, we are gradually populating other modules dedicated to specific lectins predicted in sequence databases.The second module, PropLec, is focused on families of lectins that display a β-propeller structure.

3 UniLectin3D, Database of Curated Lectin 3D Structure, and Their Interacting Ligands The UniLectin3D database includes structural information on lectins along with their interactions with carbohydrate ligands. A curated classification is proposed based on origin and fold in association with the literature and functional data such as known specificity. The content of UniLectin3D is centered on threedimensional data, using PDB information, with an appropriate curation of the glycan topology. It provides a family-based classification and cross-links to specialized glyco-related databases. Finally, the 3D visualization of contacts between the lectin and the ligand is done via the Protein-Ligand Interaction Profiler (PLIP) application. The introduction of such a feature is likely to meet the expectations of lectin specialists. The three-dimensional structures reported in UniLectin3D are those of lectins crystallized with or without their carbohydrate (glycans) ligands and non-carbohydrate ligands (see background for detailed information). The current 2117 lectin structures were all manually curated; this corresponds to 534 different lectins (as of 2019-04-17). Bibliographic entries cover 860 published articles describing at least one structure. The first classification level, referred to as “origins,” separates the lectins into seven different classes reflecting the main orders of the living

UniLectin Structural Database

5

kingdom. The second level orders the lectins according to the protein fold into 75 classes. The third level separates the lectins according to their species in 309 families. Among the 2117 3D structures, 1338 occur as complexed with glycans. The most commonly observed monosaccharides are as follows: galactose (Gal) 31% (730), N-acetyl glucosamine (GlcNAc) 16% (377), glucose (Glc) 15% (345), mannose (Man) 12% (283), fucose (Fuc) 9% (211), sialic acid (Neu5Ac) 9% (211), and N-acetyl galactosamine (GalNAc) 7% (170). Rarer sugars are also observed in complexes with lectins (Rhamnose, Arabinose. . .). The ligands occur as monosaccharides, and also as oligosaccharides or glycoconjugates. The set of distinct glycan ligands amounts to 222. The following options are available for searching: (1) keywords, (2) kingdom order, (3) historical classification, (4) monosaccharide and associate IUPAC sequence, (5) fold of the binding site, and (6) multiple criteria. Once selected, lectins can be explored (and their features pictured and downloaded) by sequence (with the UniProt AC) and structure (with the PDB ID). For each lectin, a detailed page is available with 3D visualization, interactions, and links to external databases. 3.1 Searching by Keywords

The search can be performed by entering keywords as textual input: i.e. human, PDB code, UniProt accession number, lectin name, type of domains, fragment of glycan sequence, or textual fragments of the title of a publication: .e: human, propeller, 1TL2 (PDB), Q47200 (UniProt), GalNAc, Lewis.

3.2 Searching by Kingdom Order, Carbohydrate-Binding Site Class, and Species Family

Different modes are available for browsing the database and visualizing data coverage with respect to taxonomy. Lectin structures can be explored using two interactive graphical representations: a sunburst and a tree. The inner circle of the sunburst corresponds to the kingdom orders along with their respective percentage of occurrence. For a given kingdom order, (e.g., animal) the class of the carbohydrate-binding site is displayed in the central section (e.g., galectins), whereas the species families can be browsed on the outer section (e.g., galectin 3). The hierarchical taxonomic tree (Fig. 1a) can be used to explore the classification with the tree leaves expanding by clicking on the blue node and at each level. An advanced search can be launched by clicking on the label of the corresponding level.

3.3 Searching by Monosaccharide and Associate IUPAC Sequence

Glycans are described following the encoded IUPAC condensed nomenclature (http://www.sbcs.qmul.ac.uk/iupac/2carb/38. html), for example, Gal(b1-4)GlcNAc(b1-2)Man(a1-3)[GlcNAc (b1-2)Man(a1-6)]Man.

6

Franc¸ois Bonnardel et al.

Fig. 1 Interfaces for the exploration of the lectin classification and associated carbohydrate ligands. (a) Treeview of the classification highlighting the option of opening each node to display branching at the lower level. (b) Sunburst as an overview of the glycans distribution. (c) Dedicated page for selecting monosaccharides to be searched in all structures. (d) Example of the oligosaccharides obtained by selecting D-Neup5Gc in the monosaccharide panel

The nature of the interacting glycans can be visualized on a sunburst (Fig. 1b). The inner circle corresponds to the monosaccharide that composes the carbohydrate chains interacting with the lectins, together with their respective percentage of occurrence. The outer circle corresponds to the carbohydrate chains. For a given monosaccharide (e.g., D-Galp), the number of occurrences of glycans containing the selected monosaccharide is displayed (e.g., 29 in the case of Gal(b1-3)GalNAc). Clicking on a selected glycan (e.g., Gal(b1-3)GalNAc) lists the 3D structures of lectins complexed with this carbohydrate.

UniLectin Structural Database

7

A distinct interface is accessible (Fig. 1c) by clicking the “glycan search” button. It allows searching for glycans through a combination of monosaccharides. Only the carbohydrates present in at least one lectin 3D structure and containing the selected monosaccharides are displayed in the SNFG format (Fig. 1d). By clicking on a carbohydrate name, an advanced search is launched based on it. 3.4 Searching by Fold

The lectin fold is defined by the relative spatial arrangement of secondary structure elements. It characterizes the multiplicity of glycan binding, and it is directly linked to the ability of lectins to bind and cross-link glycan containing molecules in a multivalent fashion. A sunburst representation indicates the occurrence of the main fold of lectins. Clicking on a selected section of the circle displays the nature and the occurrence of the folds as identified in the crystal structures. Upon a click, buttons prompt the information pertaining to the lectin displaying the particular fold.

3.5 Advanced Search: Searching by Multiple Criteria

This advanced search option offers a range of criteria to be selected in a combined fashion in order to search the whole database for specific lectins or structures. Practically, lectins can be searched by 3D structure or by sequence family with the support of drop-down lists (Fig. 2a), and they can be filtered based on a large number of features (Fig. 2b): (1) The classification of lectins (Origin, Class, and Family); (2) The nature of the fold and taxonomic details of the lectin; (3) Keywords from the title of a reference article; (4) A unique feature is the search of fragments of glycan ligands, also called oligosaccharide motifs, that interact with the lectin ; (5) A cut˚ ) of the X-ray structure can be used as a filter off on the resolution (A for selecting of high-quality data. UniLectin3D allows precise taxonomic search (6) for all lectins that have been structurally characterized in a given organism. The resolution criteria relate to the quality of the structural determination (high numeric values of resolution, such as 4 A˚, indicate poor resolution, while low numeric values, such as 1.5 A˚, indicate a good resolution (the median resolution for X-ray crystallographic results in the Protein Data Bank is 2.05 A˚). Whenever the resolution is set to 0, the structures are not filtered.

3.6 Lectin Sequence Detailed Interface

The accession number (AC) that is assigned to each lectin sequence upon inclusion into UniProtKB can be used for searching. As one lectin sequence can be related to multiple PDB structures these are displayed in the results together with the ligand(s) shown in the SNFG representation. Some structures may have been published in multiple articles; these are listed together with the corresponding link to PubMed (http://www.ncbi.nlm.nih.gov/pubmed). When available, the structures are displayed on the main protein sequence in a 2D viewer provided by PDBe [9].

8

Franc¸ois Bonnardel et al.

Fig. 2 Detailed interface of lectin search with multiple criteria showing stored information and 3D-visualization. (a) PDB code for selecting one lectin. (b) Multicriterion window for advanced searc h. (c) Example of results obtained with entry 4POT. (d) Detailed graphical information provided for the entry shown in (a) 3.7 X-Ray Structure Detailed Interface

The PDB code that is assigned to each lectin structure is used to list available lectin structures and to display more detailed information (Fig. 2c). Each structure is related to a protein with a UniProt AC. Other related PDB structures are listed together with their interacting glycans, if any. The 3D X-ray structure of the lectin is visualized directly and interactively with the integration of the LiteMol [15] and NGL [25] viewers (Fig. 2d). Information about the interaction occurring between the glycan and the combining sites of the lectin can be obtained using the Protein-Ligand Interaction Profiler (PLIP) server [23]. The

UniLectin Structural Database

9

description of the glycan complies with representations and numerical descriptors that allow for cross-linking to other databases in glycoscience. The architecture and navigation tools are designed to extend the search to all organisms, as well as to search for all glycan epitopes complexed within specified binding sites. The NGL viewer adapted to SwissModel [26] displays the interactions resulting from the PLIP application. This offers a detailed 3D visualization of the specific features of the interactions between the glycans and the surrounding amino acid residues and possible metal ions. A complementary description of the 3D interactions between the lectin and glycan is given by the domain viewer of the PDBe.

4

PropLec: Database of Structure-Based Predicted β-Propellers The PropLec database includes predicted β-propeller lectins along with their features and conserved regions. In structural biology, a β-propeller is a particular type of beta-sheet protein architecture characterized by four to eight highly symmetrical blade-shaped β-sheets arranged toroidally around a central axis. Together the β-sheets form a funnel-like active site. The blade consists of a small domain of less than 50 amino acids. The repeated blades hamper the identification of similar lectins when using common software based on pairwise sequence alignment such as BLAST [27]. However, the multiple alignment of blades manually adjusted with knowledge of 3D structures produces a unique conserved domain. This blade domain can then be used to compare all known β-propeller lectins and this systematic comparison led to the definition of five distinct families. To simplify the nomenclature, each family is named based on the number of blades, e.g., PropLec5A (Tachylectin 2 like), PropLec6A (RSL and AAL like), PropLec6B (Tectonin like), PropLec7A (PLL like), and PropLec7B (PVL like) (Fig. 3a). The specific signature of each family has been used to predict with HMMER [28] the possible β-propeller lectins from the UniProt sequence database [8] and are associated with RefSeq if available [29]. The results of this prediction are stored and can be searched in the PropLec module of UniLectin, based on their family, species, taxonomy, number of blades, associated enzymatic functions, and other additional features.

4.1 Searching by Keywords and by Family

The features of the predicted β-propeller lectins can be explored using multiple criteria from the homepage. A quick search can be performed by keywords: accession number, species name, or protein name. The five families are then made accessible through buttons along with a pie chart depiction of the distribution of the number of predicted proteins in each family. Based on the two distinct sets, Animal and Fungi, the distribution of the number of

10

Franc¸ois Bonnardel et al.

Fig. 3 β-Propeller families. (a) The five families of β-propellers used to build the database. (b) Pie charts representing family distribution in animal (metazoan) and fungal lectins

predicted β-propellers is represented (Fig. 3b). As expected, a majority of the predicted Animal β-propellers are PropLec6B and a majority of the predicted Fungi β-propellers are PropLec6A. Surprisingly, ProLec7A from Photorhabdus luminescens is predicted in both Animals and Fungi, and PropLec 7B from Psathyrella velutina is predicted in Animals. 4.2 Searching by Number of Blades in the Propeller

The number of blades in the predicted β-propeller lectins can be used to search for particular structures. The result is shown as a histogram representing the distribution of the predicted lectins relative to the number of blades identified in sequences. Clicking on any of the bars of the histogram prompts the details of the corresponding predicted lectins.

4.3 Searching by Phylum

The search can be performed by selecting in the Taxonomy sunburst either a superkingdom, a phylum, or a species. The distribution of predicted lectins across species can be explored at each level. As the sunburst is built as concentric circles, the inner circle represents the superkingdom, the next circle the kingdom, the third circle the phylum, the fourth circle the species group, and the outer circle the species. Clicking on any of the circle sections prompts the details of the predicted lectins found in the selected taxonomic group.

UniLectin Structural Database

11

4.4

Advanced Search

Predicted propeller lectins can be selected using a combination of criteria, and the corresponding lectins are then ordered by scores (highest to lowest). Possible combinations involve the following: (1) the prediction score threshold, which is set 0.25 by default; (2) the identified propeller family; (3) the number of blades identified; (4) the maximum distance between blades; (5) keywords that exclude proteins based on their description, set by default to “partial, synthetic, and undefined,” (6) taxonomy, (7) Pfam domains, (8) RefSeq AC, (9) protein name and description; and (10) UniProt AC. The “checkbox pathogen” button offers the possibility to restrict the search to a particular pathogen species, based on the NIH predefined list of pathogen species. A graphic overview of the properties of predicted propeller lectins resulting from searching with a single criterion or a combination of criteria, is generated. It shows (1) the distribution of the PropLec families; (2) the distribution of the number of blades; and (3) a sunburst and a tree representation, as in the homepage. Clicking on the graphic sections displays further details. The predicted lectins matching the criteria are ordered by score with 20 items per page. For each predicted lectin, the following features are displayed: (1) protein name, (2) UniProt AC and RefSeq AC, which can be clicked to access the corresponding pages in the respective sequence databases, (3) protein length, (4) species, (5) domain identified in the Proplec families, (6) number of blades, (7) similarity score of the predicted protein blades to the reference blade, and (8) protein-coding gene list including chromosome number(s) and location. For each predicted lectin, an in-house 2D sequence viewer indicates the localization of the predicted blades and possible Pfam domains. Zooming in on the sequence is performed via a “drag and drop” button. Further details are available by clicking the “more information” button.

4.5

Detailed Results

For each lectin, a detailed panel and page with the NCBI gene viewer and a representation of the blade conservation compared to the reference are available. The protein features were described in the previous section. The protein-coding gene and its location on a chromosome are represented by the NCBI viewer, when the information is available [30]. The view can be scrolled back and moved by “drag and drop” to check the surrounding genes. To compute the score, the predicted lectin blades have been aligned against the reference blades by the MUltiple Sequence Comparison by Log-Expectation (MUSCLE) [31]. To provide a more detailed view of the blade conservation, the resulting alignment is represented in two distinct bar charts. The first bar chart, on top, contains the amino acid conservation of the predicted blades. The second represents the amino acid conservation of the reference blade. As all blades are aligned, the bar chart position facilitates

12

Franc¸ois Bonnardel et al.

the comparison between sequences. The binding sites are represented, along with the amino acids known to interact with glycans either by a hydrogen bond or by hydrophobic interactions. 4.6 Searching by Other Pfam Functional Domains

5

Lectin domains can either be a whole lectin protein or only one part of a larger multifunctional protein. Proteins with a predicted lectin β-propeller domain(s) and other functional Pfam domains can be explored to evaluate possible combination(s). Such complex architecture is of particular interest as it highlights the specificity of lectins toward a definite carbohydrate in combination with other functional domain(s) (i.e., a glycan transferase). For each component of the protein architecture, a button prompts the list of predicted lectins displaying a similar pattern.

Conclusion and Discussion The development of high-quality glycomics databases counteracts the lack of precision reflected in the abundance of unreviewed and incorrect information regarding both glycoconjugates and glycanbinding proteins in genome and protein databases. Here, we reviewed databases with information on lectins and their interacting glycans (mono, oligo, and polysaccharides). The recently released UniLectin platform provides a curated classification of lectins along with their reviewed interactions with glycans. Tools that facilitate lectin knowledge exploration were implemented. UniLectin has recently exceeded 2000 lectin structures. The platform also includes a tutorial that describes step by step the usage of simple and advanced search of the UniLectin databases covering lectin 3D structures and predicted β -propeller lectins . We strive to ensure content accuracy and regular updates of the UniLectin platform as well as to provide a user-friendly tool collection (search, visualization, etc.). Currently, UniLectin has a growing community whose feedback is key to driving further development. Based on UniLectin3D curated information and lectin classes, a high-quality prediction of lectins in genomes is a reachable short-term goal, but it still requires the revision of the current classification criteria, including amino acid sequence patterns.

Acknowledgments The authors acknowledge support by the ANR PIA Glyco@Alps (ANR-15-IDEX-02) and the Alliance Campus Rhodanien Co-funds (http://campusrhodanien.unige-cofunds.ch).

UniLectin Structural Database

13

References 1. Lis H, Sharon N (2002) Lectins: carbohydratespecific proteins that mediate cellular recognition. Chem Rev 98:637–674 2. Gallagher JT (1984) Carbohydrate-binding properties of lectins: a possible approach to lectin nomenclature and classification. Biosci Rep 4:621–632 3. Peumans WJ, Van Damme EJ, Barre A et al (2001) Classification of plant lectins in families of structurally and evolutionary related proteins. Adv Exp Med Biol 491:27–54 4. Kaltner H, Gabius H-J (2011) In: Wu AM (ed) Animal lectins: from initial description to elaborated structural and functional classification. The molecular immunology of complex carbohydrates—2 advances in experimental medicine and biology, vol 491. Springer, Boston, MA, pp 79–94 5. Fujimoto Z, Tateno H, Hirabayashi J (2014) Lectin structures: classification based on the 3-D structures. Methods Mol Biol 1200:579–606 6. Finn RD, Bateman A, Clements J et al (2014) Pfam: the protein families database. Nucleic Acids Res 42:D222–D2230 7. Makyio H, Kato R (2016) Classification and comparison of fucose-binding lectins based on their structures. Trends Glycosci Glycotechnol 28:E25–E37 8. The UniProt Consortium (2019) UniProt: a worldwide hub of protein knowledge. Nucleic Acids Res 47:D506-D515. 9. Mir S, Alhroub Y, Anyango S et al (2018) PDBe: towards reusable data delivery infrastructure at protein data bank in Europe. Nucleic Acids Res 46:D486–D492 10. Pe´rez S, Sarkar A, Rivet A et al (2015) Glyco3d: a portal for structural glycosciences. Methods Mol Biol 1273:241–258 11. Hirabayashi J, Tateno H, Shikanai T et al (2015) The lectin frontier database (LfDB), and data generation based on frontal affinity chromatography. Molecules 20:951–973 12. Chandra NR, Kumar N, Jeyakani J et al (2006) Lectindb: a plant lectin database. Glycobiology 16:938–946 13. Mariethoz J, Khatib K, Alocci D et al (2016) SugarBindDB, a resource of glycan-mediated host-pathogen interactions. Nucleic Acids Res 44:D1243–D1250 14. Alocci D, Mariethoz J, Gastaldello A et al (2019) GlyConnect: glycoproteomics goes visual, interactive, and analytical. J Proteome Res 18:664–677

15. Sehnal D, Deshpande M, Varˇekova´ RS et al (2017) LiteMol suite: interactive web-based visualization of large-scale macromolecular structure data. Nat Methods 14:1121–1122 16. Raman R, Venkataraman M, Ramakrishnan S et al (2006) Advancing glycomics: implementation strategies at the consortium for functional glycomics. Glycobiology 16:82R–90R 17. Mehta AY, Cummings RD (2019) GLAD: GLycan Array Dashboard, a visual analytics tool for glycan microarrays. Bioinformatics 35 (18):3536–3537 18. Marchler-Bauer A, Derbyshire MK, Gonzales NR et al (2015) CDD: NCBI’s conserved domain database. Nucleic Acids Res 43: D222–226 19. Mitchell AL, Attwood TK, Babbitt PC et al (2019) InterPro in 2019: improving coverage, classification and access to protein sequence annotations. Nucleic Acids Res 47: D351–D360 20. Chandonia JM, Fox NK, Brenner SE (2019) SCOPe: classification of largemacromolecular structures in the structural classification of proteins-extendeddatabase. Nucleic Acids Res 47: D475–D48 21. Sillitoe I, Dawson N, Lewis TE et al (2019) CATH: expanding the horizons of structurebased functional annotations for genome sequences. Nucleic Acids Res 47:D280–D284 22. Lu¨tteke T, von der Lieth CW (2004) PDB-care (PDB CArbohydrate REsidue check): a program to support annotation of complex carbohydrate structures in PDB files. BMC Bioinformatics 5:69 23. Salentin S, Schreiber S, Haupt VJ et al (2015) PLIP: fully automated protein-ligand interaction profiler. Nucleic Acids Res 43: W443–W447 24. Varki A, Cummings RD, Aebi M et al (2015) Symbol nomenclature for graphical representation of glycans. Glycobiology 25:1323–1324 25. Rose AS, Bradley AR, Valasatava Y et al (2018) NGL viewer: web-based molecular graphics for large complexes. Bioinformatics 34:3755–3758 26. Bienert S, Waterhouse A, De Beer TAP et al (2017) The SWISS-MODEL repository-new features and functionality. Nucleic Acids Res 45:D313–D319 27. Camacho C, Coulouris G, Avagyan V et al (2009) BLAST+: architecture and applications. BMC bioinformatics 10:421

14

Franc¸ois Bonnardel et al.

28. Finn RD, Clements J, Arndt W et al (2015) HMMER web server: 2015 Update. Nucleic Acids Res 43:W30–W38 29. O’Leary NA, Wright MW, Brister JR et al (2016) Reference sequence (RefSeq) database at NCBI: current status, taxonomic expansion, and functional annotation. Nucleic Acids Res 44:D733–D745

30. Brown GR, Hem V, Katz KS et al (2015) Gene: a gene-centered information resource at NCBI. Nucleic Acids Res 43:D36–D42 31. Edgar RC (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 32:1792–1797

Chapter 2 Purification of Sugar-Binding Peptides from L-Type Lectins Kazuo Yamamoto Abstract L-type lectin is the most famous plant lectin family, and purification of many kinds of L-type lectins were reported previously. In general, this type of lectin is a major component of the seed proteins, and it is purified from the seeds. Recently, affinity carriers immobilized with haptenic sugars are generally used for purification. Key words L-type lectin, Leguminous lectin, Metal ions, Sugar-binding peptide, Tetramer

1

Introduction Leguminous lectin, which is also called as “L-type lectin,” is one of the largest and most popular of the plant lectin families. It consists of a subunit with a molecular weight of approximately 30 kDa and forms dimer or tetramer. Each subunit has a sugar-binding site, and its binding to sugar ligand requires both Ca2+ and Mn2+ ions to keep the conformation of the sugar-binding site stable and to augment the binding between sugar ligand and lectin. The amino acid sequence homology among L-type lectins is approximately 20–40%, and their tertiary structures are also quite similar. However, the sugar-binding specificity of L-type lectins shows a wide variation compared to those of other lectin families. The variations in the sugar-binding specificities of L-type lectins depend on one of the sugar-binding loops. The sugar-binding site of an L-type lectin consists of four loops (loop A, B, C, and D from the N-terminus, Fig. 1) [1]. Among these loops, the longest loop C shows variations in both amino acid sequence and structure. Loop C is located in the center of metal ion-binding region, and the metal ion (Ca2+ and Mn2+) binding to this region keeps the conformation of loop C stable, which explains that Ca2+ and Mn2+ ions are required for L-type lectin sugar-binding activity. Loop C contains the amino acid residue Asn, which is involved in Ca2+-binding, and this residue is strictly conserved in L-type lectins. This residue plays an

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_2, © Springer Science+Business Media, LLC, part of Springer Nature 2020

15

16

Kazuo Yamamoto

Fig. 1 Sugar-binding region of leguminous lectins consists of four loops, especially loop C being largely involved in sugar-binding specificity

important role in both stable loop C conformation and enhancement of sugar binding via coordinate bond to Ca2+. Thus, the upstream and downstream amino acid sequences of the conserved Asn in loop C show wide variations in both sequence and length. In our series of studies we found that the loop C of L-type lectins seems to be a “variable binding loop,” which is just similar to those of the antigen binding sites of immunoglobulins [2–6]. Actually, we established several mutated L-type lectins with unique sugarbinding specificities by introducing substitutions in the loop C coding region and subsequent screening of expressed recombinant lectins using cell surface-display method [7, 8]. In this chapter, we describe the general purification method of L-type lectins and their sugar-binding peptide.

2

Materials

2.1 Preparation of Sugar-Immobilized Sepharose

1. Sepharose 4B. 2. Sodium oxide. 3. Epichlorohydrin. 4. Hydrazine hydrate. 5. Sodium borohydride. 6. Sodium cyanoborohydride. 7. Dipotassium hydrogenphosphate. 8. Lactose or maltose. 9. Shaker at 40  C. 10. Glass filter.

L-Type Lectins

2.2 Purification of L-Type Lectins

17

1. Leguminous seeds. 2. Mixer. 3. PBS (10 mM sodium phosphate buffer, pH 7.4 containing 0.15 M NaCl). 4. Centrifuge. 5. Ammonium sulfate. 6. CaCl2 and MnCl2. 7. Lactose-Sepharose, maltose-Sepharose, glycopeptide-coupled Sepharose, etc.

2.3 Purification and Identification of the Sugar-Binding Peptide of L-Type Lectins

1. Sugar-immobilized Sepharose column (see Subheading 2.1). 2. Endoproteinases such as trypsin, chymotrypsin, endoproteinases GluC, AspN, or LysC. 3. Reversed-phase high-performance liquid chromatography (HPLC). 4. C18 column for reversed-phase HPLC. 5. Protein sequencer or mass spectrometry.

3

Methods

3.1 Preparation of Sugar-Immobilized Sepharose

1. One hundred grams of Sepharose 4B beads is washed with H2O on a glass filter. 2. Add 65 mL of 2 M NaOH containing 0.2% NaBH4, 15 mL epichlorohydrin, and 150 mL H2O. 3. Shake at 40  C for 4 h. 4. After shaking, Sepharose beads are washed with H2O. 5. Epoxylated Sepharose is then mixed with 75 mL of hydrazine hydrate and 1.5 g NaBH4 and shaken at 40  C for 4 h. 6. After shaking, Sepharose beads are washed with H2O. 7. Hydrazide-Sepharose is then incubated with 75 mL of 0.2 M K2HPO4 containing 0.2 M lactose or maltose. 8. Add 1.2 g of NaCNBH4 and shake at 40  C for more than 24 h (see Note 1).

3.2 Purification of L-Type Lectins

L-type lectins are usually purified from the seeds of leguminous plants. In most cases, lectins are a major component of seeds, thus seeds are used as a starting material. Before the purification of lectins from the seeds, the defat process enhances the recovery of proteins including lectins. The general procedure for the purification of L-type lectins is as follows:

18

Kazuo Yamamoto

1. Proteins are extracted from the seeds by grinding the seeds in PBS using a mixer. The ground powder of the seeds is suspended in PBS and allowed to stand at 4  C and stirred overnight. 2. After the centrifugation of the extracted seeds in PBS for 60 min at 18,600  g, the supernatant is collected. 3. Add ammonium sulfate gradually up to 80% saturation for salting-out proteins (see Note 2). 4. The ammonium sulfate precipitate is collected by centrifugation at 18,600  g for 60 min. 5. The precipitate is then suspended in PBS and dialyzed thoroughly against PBS. 6. After the centrifugation of the dialyzed extract to remove insoluble materials, the supernatant is used as a crude extract of leguminous seeds. 7. For the purification of the lectin from the crude extract, sugarimmobilized Sepharose column is usually used. In the case of galactose-binding lectins, the lactose-Sepharose column is helpful, since the preparation of lactose-Sepharose is easy (see Subheading 3.1). Mannose-binding lectins are usually purified by using maltose-Sepharose column. Other lectins, which are specific for sialic acid, fucose, N-acetylgalactosamine, or other sugars, glycopeptides prepared from appropriate glycoproteins are immobilized to NHS-activated Sepharose beads and used for affinity column (see Note 3). 8. Crude extract of the lectins from the leguminous seeds is applied to the haptenic sugar-immobilized Sepharose column. 9. After washing the column with PBS, the adsorbed lectins are eluted from the column by haptenic sugar-containing PBS. In the case of lactose- or maltose-Sepharose columns, 0.2–0.5 M sugars in PBS are used, since lectin is sharply eluted from the column in smaller amount of elution buffer. 10. If large amounts of sugars are not available, 50 mM glycineHCl, pH 3.0, is used for the inactivation and elution of lectins adsorbed to the column. 11. When lectins are eluted by 50 mM glycine-HCl, pH 3.0, the neutralization of the eluate should be performed just after the elution by adding a small amount of 1 M Tris–HCl, pH 8.0, containing 2 mM CaCl2 and 2 mM MnCl2 (see Note 4). 12. The eluate is thoroughly dialyzed against H2O and lyophilized. Lyophilized lectins are stable at 4  C for many years.

L-Type Lectins

3.3 Purification and Identification of the Sugar-Binding Peptides of L-Type Lectins

19

One of the characteristics of L-type lectins is their variety of sugarbinding specificities. Such biochemical properties are based on their structures, especially that of sugar-binding site. Previously, we reported identification and isolation of sugar-binding peptides from several L-type lectins with different specificities [2]. Additional studies showed that isolated sugar-binding peptides from L-type lectins had homology to each other and interestingly each of them was derived from its metal-binding region [3]. 1. One milligram of L-type lectin is subjected to endoproteinase digestion at 37  C for 18 h. 2. After centrifugation of the digested material for 10 min at 20,000  g, the supernatant is applied on to the haptenic sugar-immobilized Sepharose column (5 mL) and then the column is washed with PBS at a slow flow rate such as 20 μL/min (see Notes 5 and 6). 3. Collect eluate in each fraction of 0.5 mL. 4. Elute bound material with haptenic sugar-containing PBS such as 0.5 M lactose. 5. To monitor the digested peptides in each fraction, a small part of the fraction is analyzed on C18 reversed-phase HPLC under appropriate conditions. 6. Sugar-binding peptide(s) could be eluted later than other peptides (Fig. 2).

Fig. 2 Elution profiles of each fraction of Asp-N digest of Bauhinia purpurea agglutinin (BPA) from a column of lactose-Sepharose. Sugar-binding peptide of BPA is indicated by arrows in fractions 7 and 8 [2]

20

Kazuo Yamamoto

7. To identify the sugar-binding peptide of the L-type lectin, amino acid sequencing of the peptide is carried out by using the amino acid sequencer of mass spectrometry.

4

Notes 1. NaCNBH4 is very toxic, so handle it carefully in a draft chamber. 2. If the seeds are rich in lipids, delipidation is performed by grinding the seeds in acetone before proteins are extracted in PBS. 3. Glycopeptides are prepared from glycoproteins by pronase digestion. After fractionation of glycopeptides on a gel filtration column such as Sephadex G25, glycopeptides are coupled with NHS-activated Sepharose 4B. 4. Ca2+ and Mn2+ ions are usually released when L-type lectins are allowed to stand under pH below 4.0. 5. Metal ions bound to L-type lectins seem to keep the conformation of the sugar-binding peptide in that of native protein during endoproteinase digestion, which allows us to purify the peptide on a sugar-immobilized column. Thus, the denaturation of the digested material may cause loss of its affinity to sugars. 6. Undigested lectin binds to the sugar-immobilized column and is eluted by haptenic sugar-containing PBS. However, sugarbinding peptides could not bind to the column, since the affinity of the peptide to haptenic sugar is weak.

References 1. Sharon N, Lis H (2003) Lectins, 2nd edn. Kluwer Academic Publishers, Dordrecht 2. Yamamoto K, Konami Y, Osawa T (1991) Purification and characterization of a carbohydratebinding peptide from Bauhinia purpurea lectin. FEBS Lett 281:258–262 3. Yamamoto K, Konami Y, Osawa T, Irimura T (1992) Carbohydrate-binding peptides from several anti-H(O) lectins. J Biochem 111:436–439 4. Yamamoto K, Konami Y, Osawa T, Irimura T (1992) Alteration of carbohydrate-binding specificity of Bauhinia purpurea lectin through the construction of chimeric lectin. J Biochem 111:87–90

5. Yamamoto K, Konami Y, Osawa T (2000) Bauhinia purpurea lectin and Lens culinaris lectin recognizes unique carbohydrate structure. J Biochem 127:129–135 6. Yamamoto K, Maruyama IN, Osawa T (2000) Cyborg lectins: Novel leguminous lectins having unique specificities. J Biochem 127:137–142 7. Soga K, Abo H, Qin SY et al (2015) Mammalian cell surface display as a novel method for developing engineered lectins with novel characteristics. Biomol Ther 5:1540–1562 8. Abo H, Soga K, Tanaka H et al (2015) Mutated leguminous lectin containing a heparin-binding like motif in a carbohydrate-binding loop specifically binds to heparin. PLoS One 10:e0145834

Chapter 3 Recombinant Expression and Purification of Animal Intracellular L-Type Lectins Tadashi Satoh and Koichi Kato Abstract Animal leguminous-type (L-type) lectins, including ERGIC-53 and VIP36 are responsible for intracellular transport and quality control of N-linked glycoproteins in the early secretory pathway. These lectins possess the carbohydrate recognition domain (CRD), which recognizes high-mannose-type glycans in a Ca2+dependent manner. Here we describe the procedures involved in bacterial overproduction and purification of the CRDs of the animal L-type lectins. Key words Cargo receptor, ERGIC-53, Low-temperature expression, Refolding, VIP36, VIPL

1

Introduction ER–Golgi intermediate compartment protein of 53 kDa (ERGIC53) and 36 kDa vesicular-integral membrane protein (VIP36) are categorized as leguminous-type (L-type) lectins, and they operate as cargo receptors for the intracellular trafficking of certain N-linked glycoproteins in the early secretory pathway of animal cells [1–3]. In particular, ERGIC-53 forms a cargo receptor complex with an EF-hand protein called multiple coagulation factor deficiency protein 2 (MCFD2), which transports blood coagulation factors V and VIII (FV and FVIII) in the early secretory pathway [4]. The ER-resident VIP36-like protein, also known as VIPL, is supposed to be involved in the ER quality control system. Several research groups, including ours, have established procedures for the expression and purification of these animal L-type lectins using bacterial, insect, and mammalian cell expression systems [5–14]. These methods have facilitated structural and functional studies, thereby contributing to our understanding of the biological roles of L-type lectins. Here, we summarize the procedures involved in the recombinant overexpression and purification of carbohydrate recognition

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_3, © Springer Science+Business Media, LLC, part of Springer Nature 2020

21

22

Tadashi Satoh and Koichi Kato

domain (CRD) of animal L-type lectins, i.e., ERGIC-53, VIP36, and VIPL, using easy-to-use bacterial expression systems. The CRD of VIP36 can be produced using a low-temperature expression approach, whereas those of ERGIC-53 and VIPL can be obtained using an oxidative refolding method.

2

Materials

2.1 Construction of Expression Vectors

1. Genes: cDNA encoding human ERGIC-53 (LMAN1, UniProt ID: P49257), human VIP36 (LMAN2, Q12907), and human VIPL (LMAN2, Q9H0V9). 2. The expression vectors: VIP36; modified pET3c (Novagen) containing a hexahistidine tag and PreScission protease cleavage recognition site (LEVLFQGP) after the BamHI site; ERGIC-53; pCold-III (TaKaRa Bio Inc.), VIPL; pET22b (Novagen).

2.2 Expression and Purification of Animal L-Type Lectins

1. Escherichia coli strains: BL21(DE3) Codon-Plus (Agilent Technologies) and BL21(DE3)pLysS (Novagen). 2. Culture regents: Luria Bertani (LB) medium, 0.5 M isopropyl β-D-thiogalactoside (IPTG), 50 mg/mL ampicillin. 3. Protein purification reagents: 1.0 M Tris–HCl (pH 7.5), 1.0 M Tris–HCl (pH 8.0), 5.0 M NaCl, 1.0 M CaCl2, 0.5 M ethylenediaminetetraacetic acid (EDTA), 1.0 M dithiothreitol (DTT), 4.0 M imidazole, 1.0 M 4-(2-aminoethyl)benzenesulfonyl fluoride (AEBSF), Triton X-100, and PreScission protease (GE Healthcare). 4. Denaturation buffer: 6.0 M guanidine–HCl, 50 mM Tris–HCl (pH 8.0), and 0.5 mM DTT. 5. Refolding buffer: 50 mM Tris–HCl (pH 8.0), 0.4 M L-arginine, 5 mM reduced glutathione, and 0.5 mM oxidized glutathione. 6. Columns: Chelating Sepharose, Resource Q, Superdex 75, and Superose 12 (GE Healthcare). 7. Others: Centriprep YM-10 (Merck Millipore) and QuixStand System (GE Healthcare).

3

Methods

3.1 Soluble Expression and Purification of Human VIP36-CRD

1. Amplify the gene encoding the CRD of human VIP36 (residues 45–296) by PCR using a high-fidelity DNA polymerase. 2. Carry out the subcloning of the amplified DNA fragment into the NdeI and BamHI sites of a modified pET3c vector using conventional molecular biology techniques.

Purification of Animal L-Type Lectins

23

3. Transform an E. coli BL21(DE3)pLysS using the abovementioned plasmid. 4. Preculture the E. coli cells with 5 mL of Luria Bertani (LB) medium containing 50 μg/mL ampicillin at 37  C for 12 h. 5. Inoculate the precultured cells into the 1 L medium and incubate them until OD600 reaches an approximate value of 0.8 at 18  C or 37  C. 6. Add 0.08 mM IPTG to induce protein overexpression, and incubate for 8 h at 18  C (see Note 1). 7. Harvest the E. coli cells by centrifugation, and resuspend them using 30 mL of a lysis buffer containing 10 mM Tris–HCl (pH 8.0) and 1 mM CaCl2 (see Note 2). 8. Lyse the E. coli cells by sonication on ice, and collect the supernatant after centrifugation (26,740  g, 30 min, 4  C). 9. Load the soluble fraction onto 5 mL of Ni2+-charged Chelating Sepharose equilibrated with the lysis buffer. 10. Wash the resin extensively with 50 mL of the lysis buffer, and eluate the VIP36-CRD protein using a lysis buffer containing 500 mM imidazole. 11. Dialyze the sample against 25 mM Tris–HCl (pH 7.5), 150 mM NaCl, and 1 mM DTT. 12. Concentrate the dialyzed sample up to 1 mg/mL using Centriprep YM-10. 13. Add a PreScission protease into the 1 mg/mL VIP36-CRD sample (3 units per 1 mg protein). 14. Confirm the proteolysis by SDS-PAGE, and quench the enzymatic reaction by addition of 10 mM 4-(2-aminoethyl)benzenesulfonyl fluoride (AEBSF). 15. Dialyze the sample against the lysis buffer. 16. Further purify the sample using the size exclusion column of Superose 12 equilibrated with the lysis buffer. 17. Store the purified VIP36-CRD at Note 3). 3.2 Refolding and Purification of ERGIC-53-CRD

80



C until use (see

1. Carry out the subcloning of CRD of human ERGIC-53 (residues 31–269) into the NdeI and BamHI sites of a pCold-III vector (see Note 4). 2. Use E. coli BL21(DE3)Codonplus cells for the overproduction of proteins. 3. Preculture the E. coli cells using the same protocol as that for VIP36-CRD. Inoculate the precultured cells into the 1 L medium and incubate them until OD600 reaches the values of 0.5–0.6 at 37  C.

24

Tadashi Satoh and Koichi Kato

4. Add 0.5 mM IPTG to induce protein overexpression, and incubate for 12–24 h at 15  C (see Note 5). 5. Harvest the E. coli cells by centrifugation, and resuspend them using 30 mL of a lysis buffer containing 10 mM Tris–HCl (pH 8.0), 2 mM CaCl2, 150 mM NaCl, and 1 mM EDTA. 6. Lyse the E. coli cells by sonication on ice, and collect the precipitant after centrifugation (26,740  g, 30 min, 4  C). 7. Wash the precipitant with a buffer containing 50 mM Tris–HCl (pH 8.0), 2% Triton X-100, and 1 mM EDTA (three times). 8. Dissolve the inclusion body in 3–5 mg/mL protein solution with a denaturation buffer containing 6 M guanidine–HCl, 50 mM Tris–HCl (pH 8.0), and 0.5–1.0 mM DTT, and incubate them for 1 h at room temperature (see Note 6). 9. Perform the refolding of ERGIC-53-CRD using a dilution method: Dilute the denatured sample to 0.03–0.05 mg/mL using a refolding buffer containing 50 mM Tris–HCl (pH 8.0), 0.4 M L-arginine, 5 mM reduced glutathione, and 0.5 mM oxidized glutathione, and incubate the sample for 12 h at 16  C (see Note 6). 10. Dialyze the sample against a size exclusion chromatography (SEC) buffer containing 20 mM Tris–HCl (pH 7.5), 2 mM CaCl2, 150 mM NaCl, and 1 mM EDTA (see Note 7). 11. Concentrate the dialyzed sample using a proper concentrator such as QuixStand System and Centriprep YM-10. 12. Further purify the sample using the size exclusion column of Superose 12, equilibrated with the SEC buffer. 13. Store the purified ERGIC-53-CRD at Note 8). 3.3 Refolding and Purification of VIPL-CRD

80  C until use (see

1. Carry out the subcloning of human VIPL-CRD (residues 48–311) into the NdeI and EcoRI sites of a pET22b vector (see Note 4). 2. Use E. coli BL21(DE3)Codonplus cells for the overproduction of protein. 3. Preculture the E. coli cells using the same protocol as that for VIP36-CRD. Inoculate the precultured cells into the 1 L medium and incubate them until OD600 reaches values of 0.7–0.8 at 37  C. 4. Add 0.5 mM IPTG to induce protein overexpression, and incubate for 4 h at 37  C. 5. Perform the refolding of VIPL-CRD using the same protocol as that for ERGIC-CRD. 6. Dialyze the refolded and concentrated VIPL-CRD against 10 mM Tris–HCl (pH 8.0).

Purification of Animal L-Type Lectins

25

7. Purify the sample using an anion-exchange column, Resource Q in 10 mM Tris–HCl (pH 8.0) developed with a 0–1.0 M NaCl linear gradient. 8. Further purify the sample using a size exclusion column of Superdex 75, equilibrated with 10 mM Tris–HCl (pH 8.0), 1 mM CaCl2, and 150 mM NaCl. 9. Store the purified VIPL-CRD at 80  C until use (see Note 9).

4

Notes 1. The induction of protein expression by low IPTG concentration (0.08 mM) and the cultivation at low temperature (18  C) are essential for obtaining VIP36-CRD in a soluble form. Glutathione S-transferase (GST)-fused canine VIP36-CRD (residues 51–301) can be also expressed by a similar approach [8]. In the case of the canine protein, the GST-tagging is essential for the soluble expression. The canine VIP36-CRD is crystallizable [8] . 2. The CRD of VIP36 contains one Ca2+ ion, and the metal binding is indispensable for the interaction of high-mannosetype oligosaccharides (Fig. 1a) [6, 8, 14]. Therefore, Ca2+ ion helps in the protein stability of VIP36-CRD. 3. The purified VIP36-CRD can be used for interaction studies, such as frontal affinity chromatography. This L-type lectin recognizes high-mannose-type oligosaccharides containing the non-glucosylated D1 mannosyl branch Man-α1,2-Man-α 1,2-Man (Fig. 1b) and exhibits a bell-shaped pH dependence of the sugar binding with an optimal pH value of 6.5 [5, 6]. 4. The procedures for the production of CRDs of human ERGIC-53 and VIPL using insect and mammalian cells have also been established [5, 9, 15]. However, the bacterial expression system is currently recommended more in terms of the yield, cost, and time. 5. The induction of protein expression at low temperature (15  C) is mandatory for the pCold expression system (see manufacturer’s manual, TaKaRa Bio Inc.). 6. Animal L-type lectins exhibit β-sandwich fold and contain one disulfide bond connecting two β-sheets (Figs. 1a and 2a) [8, 9]. Therefore, the complete cleavage of the incorrect disulfide bonds of the inclusion body and addition of the redox reagents (reduced and oxidized glutathiones) are important manipulations for obtaining large quantities of ERGIC-53 and VIPL CRDs in the refolding experiments.

26

a

Tadashi Satoh and Koichi Kato

b

Man Man Ca2+

L262

Y164 D131

G260

D162 Ca2+

D261 S96

D193 N166 α1,2

Man Man

H190

Fig. 1 Crystal structure of canine VIP36 CRD. (a) Overall structure of VIP36 CRD in complex with Ca2+ and Man-α1,2-Man (PDB code: 2DUR) [8]. The bound dimannosyl ligand and Ca2+ ion are indicated as green stick and magenta sphere models, respectively. Residues involved in disulfide bridge are shown as sphere models. (b) Close-up view of the sugar- and Ca2+-binding site of VIP36. Residues involved in sugar binding and Ca2+ coordination are shown as stick models. Hydrogen bonds and Ca2+-coordinating bonds are indicated as dotted and solid lines, respectively

7. The CRD of ERGIC-53 possesses two Ca2+ ions, and it binds and high-mannose-type oligosaccharides in Ca2+pH-dependent manners (Fig. 2b) [5, 15, 16]. The Ca2+ ion contributes to the protein stability of CRD of ERGIC-53. This L-type lectin effectively interacts with the oligosaccharides under neutral conditions, but seldom at lower pH. 8. The purified human ERGIC-53-CRD can be utilized for the interaction studies [5], and is crystallizable in the presence and absence of MCFD2 (Fig. 2a) [7, 9, 15, 17–19]. Structural comparison between the complexed and uncomplexed forms of MCFD2 indicates that its FV/FVIII-binding site is allosterically activated by the CRD of ERGIC-53 [7, 9, 15, 17–19] (Fig. 2c). 9. The purified human VIPL-CRD can be used for the interaction studies. Like VIP36, this L-type lectin recognizes highmannose-type oligosaccharides containing the non-glucosylated D1 mannosyl branch [5, 13]. However, the pH dependence of VIPL-CRD is similar to that of ERGIC-53, but seldom to VIP36.

Purification of Animal L-Type Lectins

a

27

b

Man Man

L253 G251 D121

Ca2+

N162 F154 D152

S88

N161 D155

G252

Ca2+

α1,2

MCFD2

Ca2+

N156

Man

ERGIC-53 CRD

Man

H178

D181 D157

c MCFD2 (inactive)

N-glycan MCFD2 (active)

FV/FVIII

ERGIC-53

MCFD2

ERGIC-53 CRD

Fig. 2 Crystal structure of human ERGIC-53 CRD complexed with MCFD2. (a) Overall structure of ERGIC-53 CRD (marine blue) in complex with MCFD2 (pink) in the presence of Ca2+ and Man-α1,2-Man-α1,2-Man (PDB code: 3WNX) [19]. In the trimannosyl ligand, the two mannose residues showing unambiguous electron densities were indicated. (b) Close-up view of the sugar- and Ca2+-binding site of ERGIC-53 CRD. (c) Left: Superimposition of NMR-derived solution structure of uncomplexed MCFD2 (orange, PDB code: 2VRG) and ERGIC-53/MCFD2 complex (gray and pink). Right: The working mechanism of functional interplay between ERGIC-53 and MCFD2. MCFD2 adopts an inactive conformation in its uncomplexed state, and becomes allosterically activated through the complex formation with ERGIC-53 so as to capture FV and FVIII. In the cargo receptor/glycoprotein complex, ERGIC-53 interacts with their N-glycans, whereas MCFD2 binds polypeptide segments of the coagulation factors ((c) adapted from Kamiya et al. [1] with the permission of Elsevier B.V.)

Acknowledgements We thank Dr. Yukiko Kamiya (Nagoya University, Japan) for useful discussion. This work was supported in part by JSPS KAKENHI (Grant Numbers JP19H03361 to T.S.). We thank Drs. Kazuo Yamamoto (The University of Tokyo, Japan) and Hans-Peter Hauri (University of Basel, Switzerland) for providing protein expression systems.

28

Tadashi Satoh and Koichi Kato

References 1. Kamiya Y, Satoh T, Kato K (2012) Molecular and structural basis for N-glycan-dependent determination of glycoprotein fates in cells. Biochim Biophys Acta 1820:1327–1337 2. Satoh T, Yamaguchi T, Kato K (2015) Emerging structural insights into glycoprotein quality control coupled with N-glycan processing in the endoplasmic reticulum. Molecules 20:2475–2491 3. Lederkremer GZ (2009) Glycoprotein folding, quality control and ER-associated degradation. Curr Opin Struct Biol 19:515–523 4. Zhang B, McGee B, Yamaoka JS, Guglielmone H, Downes KA, Minoldo S, Jarchum G, Peyvandi F, de Bosch NB, RuizSaez A, Chatelain B, Olpinski M, Bockenstedt P, Sperl W, Kaufman RJ, Nichols WC, Tuddenham EG, Ginsburg D (2006) Combined deficiency of factor V and factor VIII is due to mutations in either LMAN1 or MCFD2. Blood 107:1903–1907 5. Kamiya Y, Kamiya D, Yamamoto K, Nyfeler B, Hauri HP, Kato K (2008) Molecular basis of sugar recognition by the human L-type lectins ERGIC-53, VIPL, and VIP36. J Biol Chem 283:1857–1861 6. Kamiya Y, Yamaguchi Y, Takahashi N, Arata Y, Kasai KI, Ihara Y, Matsuo I, Ito Y, Yamamoto K, Kato K (2005) Sugar-binding properties of VIP36, an intracellular animal lectin operating as a cargo receptor. J Biol Chem 280:37178–37182 7. Nishio M, Kamiya Y, Mizushima T, Wakatsuki S, Sasakawa H, Yamamoto K, Uchiyama S, Noda M, McKay AR, Fukui K, Hauri HP, Kato K (2010) Structural basis for the cooperative interplay between the two causative gene products of combined factor V and factor VIII deficiency. Proc Natl Acad Sci U S A 107:4034–4039 8. Satoh T, Cowieson NP, Hakamata W, Ideo H, Fukushima K, Kurihara M, Kato R, Yamashita K, Wakatsuki S (2007) Structural basis for recognition of high mannose type glycoproteins by mammalian transport lectin VIP36. J Biol Chem 282:28246–28255 9. Velloso LM, Svensson K, Schneider G, Pettersson RF, Lindqvist Y (2002) Crystal structure of the carbohydrate recognition domain of p58/ ERGIC-53, a protein involved in glycoprotein export from the endoplasmic reticulum. J Biol Chem 277:15979–15984

10. Appenzeller C, Andersson H, Kappeler F, Hauri HP (1999) The lectin ERGIC-53 is a cargo transport receptor for glycoproteins. Nat Cell Biol 1:330–334 11. Fiedler K, Parton RG, Kellner R, Etzold T, Simons K (1994) VIP36, a novel component of glycolipid rafts and exocytic carrier vesicles in epithelial cells. EMBO J 13:1729–1740 12. Hauri HP, Kappeler F, Andersson H, Appenzeller C (2000) ERGIC-53 and traffic in the secretory pathway. J Cell Sci 113 (Pt 4):587–596 13. Yamaguchi D, Kawasaki N, Matsuo I, Totani K, Tozawa H, Matsumoto N, Ito Y, Yamamoto K (2007) VIPL has sugar-binding activity specific for high-mannose-type N-glycans, and glucosylation of the alpha1,2 mannotriosyl branch blocks its binding. Glycobiology 17:1061–1069 14. Hara-Kuge S, Ohkura T, Seko A, Yamashita K (1999) Vesicular-integral membrane protein, VIP36, recognizes high-mannose type glycans containing alpha1-->2 mannosyl residues in MDCK cells. Glycobiology 9:833–839 15. Velloso LM, Svensson K, Pettersson RF, Lindqvist Y (2003) The crystal structure of the carbohydrate-recognition domain of the glycoprotein sorting receptor p58/ERGIC53 reveals an unpredicted metal-binding site and conformational changes associated with calcium ion binding. J Mol Biol 334:845–851 16. Appenzeller-Herzog C, Roche AC, Nufer O, Hauri HP (2004) pH-induced conversion of the transport lectin ERGIC-53 triggers glycoprotein release. J Biol Chem 279:12943–12950 17. Wigren E, Bourhis JM, Kursula I, Guy JE, Lindqvist Y (2010) Crystal structure of the LMAN1-CRD/MCFD2 transport receptor complex provides insight into combined deficiency of factor V and factor VIII. FEBS Lett 584:878–882 18. Zheng C, Page RC, Das V, Nix JC, Wigren E, Misra S, Zhang B (2013) Structural characterization of carbohydrate binding by LMAN1 protein provides new insight into the endoplasmic reticulum export of factors V (FV) and VIII (FVIII). J Biol Chem 288:20499–20509 19. Satoh T, Suzuki K, Yamaguchi T, Kato K (2014) Structural basis for disparate sugarbinding specificities in the homologous cargo receptors ERGIC-53 and VIP36. PLoS One 9: e87963

Chapter 4 Frontal Affinity Chromatography: A Highly Suitable Retardation Phenomenon-Based Research Tool for Analyzing Weak Interactions Between Biomolecules Kenichi Kasai Abstract The greatest advantage of frontal affinity chromatography (FAC) is that the analyte concentration does not need to be taken into consideration, and this renders FAC an extremely favorable analytical tool for weak interactions. In this short review, we propose a straightforward explanation of the underlying mechanism. When FAC is performed using analyte solutions at relatively high concentrations, concentration-dependent retardation is observed due to competition among analyte molecules, and the elution volume changes depending on the degree of saturation of the immobilized ligand. However, when the analyte concentration is very low, no competition occurs among the analytes, and the elution volume reaches a constant value, which reflects the proportion of bound state to free state of a single analyte molecule. Therefore, the binding strength can be determined using a minimum analyte concentration. Key words Frontal affinity chromatography, FAC, Lectin profiling, Glycobiology, Weak interaction, Retardation analysis

1

Introduction We have previously described the principle of frontal affinity chromatography (FAC) in the course of a study focusing on the substrate recognition mechanism of enzymes [1]; however, we also recognized the utility of FAC as an analytical tool for investigating the interactions between non-enzyme proteins and their ligands, e.g., those between lectins and glycans [2, 3]. FAC quickly proved to be particularly useful for the analysis of weak interactions [4–6], which had been impeded by the lack of appropriate methodologies. In FAC, a relatively large volume of dilute analyte solution is continuously applied to a column packed with a weak affinity adsorbent; subsequently, the binding strength between the analyte and the immobilized counterpart can be determined from the retardation extent of the former [7, 8]. The greatest advantage of

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_4, © Springer Science+Business Media, LLC, part of Springer Nature 2020

29

30

Kenichi Kasai

FAC is that the analyte concentration does not need to be taken into consideration, and this facilitates the marked reduction of analyte concentration. Therefore, the consumption of rare analytes, such as glycans, can be maintained at a minimum. This renders FAC an extremely favorable tool for lectin research [9–13]. In addition, an automated FAC apparatus, designed for lectin profiling, is now available [14, 15]. Furthermore, data on the detailed binding properties of many lectins have already been available in both academic papers [16] and the public database LfDB [17] (https://acgg.asia/lfdb2/). However, despite the clear benefits of maintaining analyte concentrations at a minimum level, the mechanism underlying this reduction has not yet been elucidated. Recently, a straightforward explanation for this mechanism has been proposed, as described below.

2

Peculiar Characteristics of Lectin–Glycan Interactions The following two fundamental unresolved questions are associated with the understanding of lectin–glycan interactions: (1) Why do multiple partner glycans exist for each lectin? (2) Why is the binding strength of lectins toward their partner glycans usually low? The behavior of lectins toward their binding partners is interesting. Generally, proteins recognize only one or a very limited number of partner molecules, and bind to them strongly. Strict specificity and strong binding are requisite qualities for most functional proteins including enzymes, receptors, and antibodies. These properties ascertain the consistency and reproducibility of biological processes in living organisms. However, this is not always the case for lectins. A lectin binds multiple glycans with varying affinities; furthermore, the binding is comparatively weak, and the lifetime of the complex is relatively short. Lectins look “unfaithful,” “less affectionate,” and “unsustainable” toward their binding partners. Nevertheless, for such apparently defective systems to continue existing, they should have presented certain evolutionary advantages. To understand the significance of the glycan recognition mechanism in nature, collecting data on the detailed glycan-binding characteristics of lectins is becoming increasingly important. The unusual properties of lectin can partially be attributed to the microheterogeneity of glycoconjugates. Thus, in the case of glycoproteins, a group of homogeneous proteins is rendered almost completely heterogeneous after glycan modifications. Each protein molecule, coded by the same gene and synthesized as a clone, is equipped with different glycan chains and acquires structural

Frontal Affinity Chromatography

31

individuality. The glycan part of each protein will interact with lectins in a different manner and with varying intensity, leading to diversification of the protein fate and flexibility in the biological processes of life systems. Furthermore, cell surface glycans may contribute to cellular activity in a similar manner. Glycans are synthesized without structural information; therefore, the production of a glycan with fixed structure is not guaranteed. Should a lectin exhibit too rigorous specificity, the probability of encountering its partner will be too low, and the lectin may fail to fulfill its mission. The loose specificity of lectins might present a compromise for the reduction of such risk.

3

Significance of Fuzzy and Weak Interactions in Biological Systems Lectin binding to glycoconjugates often triggers the initiation of biological processes. However, after fulfilling the biological function of the process, it should be interrupted by disrupting lectin– glycan interactions. Strong interactions would be unfavorable under such circumstances, because the formed complex would dissociate very slowly. To adapt to the rapidly changing circumstances, low affinity would be advantageous. Weakly binding lectins appear to have been preserved during evolution due to their capability for rapid switch-off. Therefore, low affinity toward glycans is also a quality requisite for lectins. The binding properties of lectins, e.g., loose specificity and low affinity, apparently oppose the generally accepted merits for most proteins. Nevertheless, their importance must have increased after the adoption of glycans as third media for biological information systems. Lectins bring fuzziness to almost all biological phenomena. Due to the existence of multiple partner glycans, exhibiting similar affinities, the final partner molecule cannot be predicted. Furthermore, the binding probability depends not only on the affinity but also on the quantity of the binding partner. Even a low-affinity glycan species will be preferentially recognized by a lectin compared to a high-affinity one, if the former is more abundant. Consequently, glycoproteins or cells, carrying low-affinity glycan species, might be stimulated to a higher extent and have greater influence on the activity of the organism. Lectin–glycan interactions cause deviations from fixed programs by introducing uncertainty into biological systems. The incorporation of fuzziness may have contributed to the diversity and survival ability of living organisms in a drastically changing environment.

32

4

Kenichi Kasai

FAC: A Powerful Analytical Tool for Weak Interactions Understanding the detailed binding properties of lectins is of high significance for the life sciences. Nevertheless, there is an inherent difficulty associated with the analysis of weak interactions. Generally, the binding strength is measured with the complex fraction formed by the interaction. However, in the case of weak interactions, the fraction of the formed complex is very small. To raise the degree of saturation, a large quantity of the partner molecules must be added. However, this is almost impossible in the case of lectin– glycan interactions due to the costly and tedious procedure for glycan preparation. Currently, FAC is the only procedure that can overcome such difficulties. It enables a reduction of glycan consumption to a minimum level, because it does not depend on the measurement of the formed complex. Instead, binding strength is determined by the extent of the analyte glycan retardation during the passage through an immobilized lectin column. FAC is based on a simple and straightforward principle. Even for a weakly interacting couple, analyte concentration can be maintained at the lowest level that allows drawing its elution profile. Moreover, it is not necessary to know the exact concentration of analyte glycan [7, 8]. These advantages render FAC as the most suitable analytical tool for evaluating weak interactions. Furthermore, the use of a small immobilized lectin column and a variety of fluorescent oligosaccharides, e.g., pyridylaminated (PA)-oligosaccharides, has turned FAC into an extremely effective and high-throughput analytical tool for profiling lectins (including their binding specificity and binding strength).

5

FAC Viewed as Retardation Analysis of Bioaffinity In FAC, the analyte solution volume larger than the one in the affinity column is applied. An elution curve, comprising a front area and plateau region, is obtained (Fig. 1, curve II). Retardation of the analyte (A) with affinity for the immobilized ligand (B) occurs due to their interaction during the passage through the column; additionally, the effluent volume, at which the analyte begins to leak (elution volume, V), is measured. V can be considered the effluent volume, at which a hypothetical boundary of the analyte solution would appear, if the boundary was undisturbed in any way; its accurate value can be easily deduced, as described previously [7, 8]. The amount of retarded A is equal to the area surrounded by the two elution curves in Fig. 1: the right (II) and left curves (I) are representative of the analyte and of a reference substance having no affinity for the adsorbent, respectively. This area corresponds to the

Frontal Affinity Chromatography

33

[A] [A]0

[A]0(V – V0) I 0

II Vo

V

Effluent volume

Fig. 1 Elution profiles in frontal affinity chromatography. Curve II is the elution pattern of the analyte, and curve I is that of a reference substance that has no affinity for the adsorbent. [A]0 is the concentration of the analyte solution

amount of analyte forming complexes with the immobilized ligand in the column, and is equal to the rectangle, [A]0(V  V0), where [A]0 is the initial concentration of A, V is the elution volume of A, and V0 is that of the reference substance that passes through the column without retardation. V0 can be considered as the volume of the space in the column that is not occupied by the matrix of packing material. This value is very close to the column volume, if agarose gel is utilized as the supporting material of the adsorbent (absence of nonspecific interaction of the reference substance with the matrix is postulated). [A]0(V  V0) is equal to the “specifically” adsorbed A, from which the degree of saturation of B can be deduced. The relationship between the dissociation constant (Kd) and the chromatographic parameters is given as follows (Eq. 1):  ½A ½B ½A 0 ½B0  ½A0 ðV  V 0 Þ=V 0 Kd ¼ ¼ ½AB ½A0 ðV  V 0 Þ=V 0 ¼

Bt  ½A 0 V V0

ð1Þ

here: [A], [B], and [AB] are the concentrations of A, B, and AB, respectively; [B]0 is the concentration of immobilized ligand; Bt is the total amount of the immobilized ligand in the column; i.e., V0[B]0. This equation can be rearranged into the following form (Eq. 2): ½A 0 ðV  V 0 Þ ¼

B t ½A 0 ½A 0 þ K d

ð2Þ

This equation is equivalent to the Michaelis–Menten equation of enzyme kinetics, and in principle, to the Langmuir’s adsorption isotherm. Equation (2) gives a hyperbolic curve depicted in Fig. 2, indicating that the column becomes saturated at the infinite concentration of A. Bt and Kd correspond to the maximum velocity and the Michaelis constant, respectively.

Kenichi Kasai

Bt

[A]0(V – V0)

34

Bt /2

0

Kd

[A]0

Fig. 2 [A]0(V  V0) vs. [A]0 plot (Eq. 2). This plot is analogous to the Michaelis–Menten plot. Values of Bt and Kd can be calculated from the coordinates of the two asymptotes of the hyperbola V

Vm Bt /2 Vm + Vo 2

[A]0(V – V0)

V0

0

Kd

[A]0

Fig. 3 V vs. [A]0 plot (Eq. 3). Dependency of elution volume on concentration of A. Both elution volume and amount of adsorbed analyte can be predicted for a given [A]0, provided that the values of Kd and Bt are known

Equation (3) can be derived from Eq. (2) as follows: V ¼ V0 þ

Bt ½A 0 þ K d

ð3Þ

This equation is useful to understand the concept of FAC. A plot of V vs. [A]0 is also a hyperbola and the two asymptotes correspond to –Kd and V0, respectively (Fig. 3). When [A]0 increases, V decreases; however, V cannot be lower than V0. Therefore, V0 is considered the limit of V when [A]0 approaches infinity, i.e., the immobilized ligand becomes saturated. In contrast, as the concentration of the analyte decreases, the extent of retardation increases. When [A]0 approaches zero, V approaches the maximum value, Vm:

Frontal Affinity Chromatography

Vm ¼ V0 þ

Bt Kd

35

ð4Þ

Therefore, the elution volume varies from Vm to V0 depending on [A]0, and the amount of adsorbed A varies from [A]0(V  V0) to Bt. From Eqs. (4) and (5) is derived: Kd ¼

Bt ðV m V 0 Þ

ð5Þ

This indicates that Kd is proportional to the reciprocal value of the extent of retardation. In other words, the association constant is proportional to the extent of retardation. Once the Bt value of a given affinity column is obtained using concentration dependence analysis for an appropriate analyte, Kd values for other analytes can be determined using single chromatographic runs, provided that their concentrations are adequately low (e.g., lower than 1% of Kd). This procedure provides a great advantage from an experimental viewpoint. It is not necessary to know the exact concentration of the analyte. Moreover, even for a weakly interacting analyte (one having a large Kd), raising the concentration of the analyte is not required. To elucidate the reason for the independence of the experimental results from the analyte concentration, Eq. (5) is rearranged to Eq. (6) as follows: ðV m V 0 Þ ¼

Bt Kd

ð6Þ

At a low A concentration, [B]0 can be considered as [B]. Accordingly, the following equation is derived (Eq. 7): ðV m V 0 Þ ½B0 ½AB½B ½AB ¼ ¼  Kd V0 ½A½B ½A

ð7Þ

This indicates that the proportion of the extent of analyte retardation to the elution volume of the reference substance is equal to the proportion of the bound analytes to the free ones. However, if considered from the viewpoint of a single A molecule, the right term indicates the proportion of bound state to free state. If [AB]/[A] ¼ 1, namely [B]0 ¼ Kd, the proportion of the adsorbed state to the free state of an A molecule becomes 1:1. Since A can migrate with the effluent only when it is free, the chance of migration decreases to 50%. Therefore, the time required for A to pass through the column increases twofold, compared to the standard molecule having no affinity to the column; (Vm  V0) being equal to V0. When [AB]/[A] ¼ 10, namely [B]0 is ten times Kd, the bound state of A is ten times that of the free state. Therefore, before it gets a chance to migrate once, it must wait ten times.

36

Kenichi Kasai

Consequently, the time required for passage through the column becomes 11 times that for the standard molecule, and (Vm  V0) becomes 10V0.

6

Two Diverse Aspects of FAC FAC is characterized by two diverse aspects in accordance with the concentration of the analyte, e.g., a saturation phenomenon and simple retardation. When FAC is performed using analyte solutions at relatively high concentrations, concentration-dependent retardation is observed. Competition among analyte molecules reduces the chance of interaction with the immobilized ligand, and the elution volume changes depending on the degree of saturation of the immobilized ligand. However, when the analyte concentration is very low, the degree of saturation of the immobilized ligand decreases to almost zero. Under such conditions, no competition occurs among the analytes. The elution volume reaches Vm, which reflects the proportion of bound state to free state of a single analyte molecule. Therefore, the binding strength can be determined using a minimum analyte concentration, and knowing the exact analyte concentration is not necessary. Notably, Bt, the total amount of immobilized ligands in the column retaining binding ability, must be evaluated in advance for calculating the dissociation constants. This value can be obtained by analyzing the concentration dependency according to Eq. (2). Therefore, analyzing the saturation phenomenon is essential for the estimation of the column capacity. For this purpose, one of abundantly available oligosaccharides such as p-nitrophenyl oligosaccharide is applied, and elution profiles are detected by a spectrophotometer. Once the Bt is determined, it is valid for the analysis of all other glycans of interest.

7

Concluding Remarks FAC is based on a concept completely different from almost all analytical tools for molecular interactions. Assessing the retardation caused by binding, instead of quantifying the formed complex, presents the most suitable analytical tool for weak interactions. Successive measurements of the retardation extent of PA-oligosaccharides through an immobilized lectin column deliver an accurate glycan-binding profile of the lectin. Analytical methods, based on the migration of biomolecules, such as size-exclusion chromatography and SDS-PAGE, are routinely utilized in biological sciences. However, their theoretical base is not solid; the molecular weight of proteins is estimated empirically using protein standards. Contrarily, FAC has a firm theoretical

Frontal Affinity Chromatography

37

foundation. In principle, it is applicable even for single molecule analysis. The development of appropriate technologies will enable the determination of binding constants by single molecule analysis. Utilization of FAC in the field of biological sciences is still limited, despite its definitive merits. This may be attributed to the limited number of studies conducted to investigate the role of fuzzy and weak interactions in living organisms. However, many systems, operating in a manner similar to the one in lectins exist in nature, including olfactory and taste receptors. Therefore, the scientific interest in fuzzy and weak interactions is expected to substantially increase in the near future, and FAC will serve as an essential tool for addressing these research questions. References 1. Kasai K, Ishii S (1975) Quantitative analysis of affinity chromatography of trypsin. A new technique for investigation of protein-ligand interaction. J Biochem 77:262–264 2. Oda Y, Kasai K, Ishii S (1981) Studies on the specific interaction of concanavalin A and saccharides by affinity chromatography. Application of quantitative affinity chromatography to a multivalent system. J Biochem 89:285–296 3. Ohyama Y, Kasai K, Nomoto H et al (1985) Frontal affinity chromatography of ovalbumin glycoasparagines on a concanavalin A-Sepharose column. A quantitative study of the binding specificity of the lectin. J Biol Chem 260:6882–6887 4. Nakano NI, Oshio T, Fujimoto Y et al (1978) Study of drug-protein interaction of bovine serum albumin and salicylic acid. J Pharm Sci 67:1005–1008 5. Sato C, Yamakawa N, Kitajima K (2010) Measurement of glycan-based interaction by frontal affinity chromatography and surface plasmon resonance. Methods Enzymol 478:219–232 6. Zhang B, Palcic MM, Mo H et al (2001) Rapid determination of the binding affinity and specificity of the mushroom Polyporus squamosus lectin using frontal affinity chromatography coupled to electrospray mass spectrometry. Glycobiology 11:141–147 7. Kasai K, Oda Y, Nishikata M et al (1986) Frontal affinity chromatography. Theory for its application to studies on specific interaction of biomolecules. J Chromatogr 376:33–47 8. Hirabayashi J, Arata Y, Kasai K (2003) Frontal affinity chromatography as a tool for elucidation of sugar recognition properties of Lectins. Methods Enzymol 362:353–368

9. Kasai K (2014) Frontal affinity chromatography (FAC): theory and basic aspects. Methods Mol Biol 200:243–256 10. Kasai K (2014) Frontal affinity chromatography: a unique research tool for biospecific interaction that promotes glycobiology. Proc Jpn Acad Ser B Phys Biol Sci 90:215–234 11. Hirabayashi J, Arata Y, Kasai K (2000) Reinforcement of frontal affinity chromatography for effective analysis of lectin-oligosaccharide interactions. J Chromatogr 890:262–272 12. Arata Y, Hirabayashi J, Kasai K (2001) Sugarbinding properties of the two lectin domains of the tandem repeat-type galectin LEC-1 (N32) of Caenorhabditis elegans. Detailed analysis by an improved frontal affinity chromatography method. J Biol Chem 276:3068–3077 13. Hirabayashi J, Hashidate T, Arata Y et al (2002) Oligosaccharide specificity of galectins: a search by frontal affinity chromatography. Biochim Biophys Acta 1572:232–254 14. Nakamura-Tsuruta S, Uchiyama N, Hirabayashi J (2006) High-throughput analysis of lectin-oligosaccharide interactions by automated frontal affinity chromatography. Methods Enzymol 415:311–325 15. Tateno H, Nakamura-Tsuruta S, Hirabayashi J (2007) Frontal affinity chromatography: sugarprotein interactions. Nat Protoc 2:2529–2537 16. Iwaki J, Hirabayashi J (2018) Carbohydratebinding specificity of human Galectins: an overview by frontal affinity chromatography. Trends Glycosci Glycotechnol 30: SE137–SE153 17. Hirabayashi J, Tateno H, Shikanai T, AokiKinoshita KF, Narimatsu H (2015) The Lectin frontier database (LfDB), and data generation based on frontal affinity chromatography. Molecules 20:951–973

Chapter 5 Metazoan Soluble β-Galactoside-Binding Lectins, Galectins: Methods for Purification, Characterization of Their Carbohydrate-Binding Specificity, and Probing Their Ligands Guillaume St-Pierre, Ann Rancourt, and Sachiko Sato Abstract Galectins are a family of soluble β-galactoside-binding proteins that share conserved carbohydrate recognition domain. Galectins are found in most multicellular organisms and exert various biological functions by binding to the surface glycoconjugates as lectins. In this chapter, we describe the general methods of purification of galectins, quality control of purified galectins, some example methods of evaluating their carbohydrate-binding activity, and use of galectin to search or detect galectin ligands as well as a series of precautions for the usage of galectins. Key words Galectin, Affinity chromatography, Enzyme-linked lectin sorbent assay, Surface plasmon resonance, Reducing agent, Isotonicity, Lectin staining

1

Introduction In mammals, there are several lectin families, including C-type lectins that require Ca2+ for their binding to glycans, Siglec, and R-type lectins. As all those lectins are synthesized through the secretory pathway, they are glycosylated proteins and delivered to some organelle, such as lysosome, secreted to the extracellular space or transported to the membrane where those lectins remain as membrane proteins. In contrast to those mammalian lectins, galectins are unique as their synthesis occurs in the cytosol and their original localization is in the cytosol [1, 2]. Currently, 15 members of the galectin family (Galectin-1 to -15) have been identified in mammals. Galectins are classified into three groups based on their quaternary structure [3]. Prototype galectins exist as either monomer or dimer of carbohydrate recognition domain (CRD), while tandem-repeat type galectins have two CRDs that are linked by a short peptide domain. Galectin-3 uniquely represents a chimera

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_5, © Springer Science+Business Media, LLC, part of Springer Nature 2020

39

40

Guillaume St-Pierre et al.

type that contains one CRD at the C-terminus and an additional non-CRD at the N-terminal domain rich in proline and lysine residues. Since the cytosol is a highly reduced environment, galectins do not contain any disulfide bonds in natural forms even though they contain some cysteine residues. Because of this unique nature, fully active recombinant galectins can be produced using Escherichia coli, which lacks the capacity to form disulfide bonds. Due to the histological reason of the original discovery of the first member of galectins (a galectin of which structure is relatively close to galectin-1) four decades ago, it was initially believed that all members of galectin family require sulfhydryl residues to maintain their carbohydrate-binding activity and are sensitive to oxidation. Indeed, they were originally referred to as S-type lectins [1, 2]. It is now well established that most of the galectins with a notable exception of galectin-1 can maintain their glycan-binding activity in the absence of reducing agents. However, it should be noted that some recombinant galectins, which can be purchased from some suppliers, contain dithiothreitol (DTT) or 2-mercaptoethanol as an additive. Since many biological reactions are highly sensitive to reducing agents [4, 5], one has to be cautious when such galectin preparations are used. Galectins play roles in many biological phenomena both inside and outside of cells (Fig. 1) [2, 6–11]. Most of them depend on their lectin activities, while some are mediated by protein–protein interactions. In many physiological conditions, such as infections, inflammation, and injury-initiated tissue regeneration, such as muscle differentiation, galectins are released or secreted from tissues, and high levels of galectins are found to be in the extracellular milieu, although the molecular mechanism of this nonclassical leaderless secretory pathway remains elusive [12]. In this chapter, methods related to mammalian galectins are described, while most of those techniques and notes could also be applied to invertebrate galectins.

2

Materials

2.1 Purification of Recombinant Galectins

1. E. coli strain BL21 (DE3) transformed with an expression plasmid of a galectin. 2. Isopropyl β-D-1-thiogalactopyranoside (IPTG) (Ambion). 3. Modified Terrific Broth: In 900 mL of Mili Q water, add 24 g yeast extract, 20 g tryptone, 4 mL glycerol, and stir until all solutes are dissolved and then sterilize by autoclaving for 20 min. After the solution becomes cool, add 100 mL of sterile phosphate buffer (0.17 M KH2PO4 and 0.72 M K2HPO4) and 1 mL of sterile MgSO4 (1 M).

Galectin Purification, Characterization and Ligand Binding

41

Fig. 1 Atlas of galectin. Galectins are primarily found in the cytosol, while galectins are either passively released or actively secreted through a leadless secretory pathway [12]. Some functions of galectins related to both host–pathogen interactions and cell biology are illustrated. Locations of galectins, host-derived galectin oligosaccharide ligands, and microbe-derived galectin glycan ligands are indicated in light blue, pink, and light green, respectively

4. Ice-cold lysis-sonication buffer (22 mM Tris–HCl, pH 7.5, 5 mM EDTA, 1 mM DTT). A protease inhibitor cocktail (Sigma) (1:100 dilution) is added just before usage. 5. α-lactose-agarose column (Sigma). 6. PBS (). 7. 150 mM lactose-containing PBS (). 8. Dialysis tube (Spectra/Por®, MWCO3,500, Spectrum Laboratories). Boil for 5 min and then wash using Milli-Q water. 9. HiPrep 26/10 desalting column (GE HealthCare) coupled with a peristaltic pump P-1 (GE Healthcare). 10. Acticlean Etox column (Sterogene). 11. Syringe filter for sterilization (0.22 μm pore size, Millex-GV low protein-binding, Durapore® PDVF, Millipore). 12. Bradford protein assay kit for protein detection and estimation of protein concentration (Bio-Rad).

42

Guillaume St-Pierre et al.

2.2 Carboxymethylation of Galectin-1

1. 1 M iodoacetamide (Sigma) for stabilizing galectin-1: freshly prepared.

2.3 Hemagglutinin Assay

1. Human blood. 2. Heparinized tube. 3. 3% glutaraldehyde in PBS. 4. PBS-0.0025% sodium azide. 5. Round-bottom 96-well plate. 6. Limulus amebocyte lysate (LAL) assay for quantification of endotoxin contamination: QCL-1000™ assay (Lonza).

2.4 Biotinylation Using Primary Amine Residues of Galectin

1. EZ-link™ Sulfo-NHS-LC-Biotin (PIERCE).

2.5 Biotinylation Using Cysteine Residues of Galectin-1

1. EZ-link™ Maleimide-PEG2-biotin non-carboxymethylated galectin-1).

2.6 Enzyme-Linked Lectin Sorbent Assay

1. ELISA plate (Cat #442404, NUNC-Thermo Scientific).

2. PD-10 desalting column (GE Healthcare). 3. 50 mM Tris–HCl, pH 7.5. (PIERCE)

(for

2. PD-10 desalting column (GE Healthcare).

2. 5 μg/mL asialofetuin (Sigma) in PBS (). 3. 10 mg/mL bovine serum albumin (Cat# A7906, Sigma). 4. Washing buffer: 20 mM Tris–HCl pH 7.5–0.05% Triton X-100. 5. Streptavidin-horse radish peroxidase (HRP) (Cat#65R-S104, Fitzgerald). 6. 3,30 ,5,50 -Tetramethylbenzidine (TMB) substrate solution (Cat#42R-TB102, Fitzgerald). 7. Stop solution: 0.18 M H2SO4.

2.7 Surface Plasmon Resonance (SPR) Assay

1. SPR chip: For Bio-Rad ProteOn™, ProteOn™ NLC sensor chip (#1765021), and for BiaCore, Sensor chip CM5 (Carboxymethylated dextran matrix, #BR100399) or Sensor Chip PEG (polyethylene glycol, #29245706). 2. 20 mM Tris–HCl pH 7.5–0.15 M NaCl-0.05% Triton X-100.

2.8 Staining with Galectin

1. 3.7% paraformaldehyde (PFA) solution. Prepare fresh daily by the following procedure: (a) Add 370 mg PFA powder and a drop (~5 μL) of 1 N NaOH to 8 mL of water. (b) Heat to 60  C until it dissolves.

Galectin Purification, Characterization and Ligand Binding

43

(c) Add 1 mL of 10 PBS () and a drop of 1 N HCl and check pH by using a pH indicator paper (Whatman) and adjust the pH to be neutral. (d) Add more water to make 10 mL of 3.7% PFA-PBS. 2. 3% BSA (Sigma, A7030, essentially globulin-free) in PBS (). 3. Biotinylated galectin: see Subheadings 2.4 and 2.5 (also see Note 1). 4. Streptavidin-Alexa (molecular probe). 2.9 Immunostaining of Galectins

1. 3.7% PFA solution: see Subheading 2.8. 2. Carbo-free blocking solution (Vector laboratories). 3. Biotinylated antibody against a galectin. 4. 0.25% Triton X-100 in PBS (). 5. Streptavidin-Alexa (molecular probe).

2.10 Special Equipment

1. T70.1 rotor (Beckman Coulter). 2. Shaker. 3. SPR ProteOn (Bio-Rad) or Biacore X100. 4. ELISA plate reader. 5. Epifluorescence or confocal microscope.

3

Methods

3.1 Purification of a Galectin

1. Inoculate E. coli strain BL21 (DE3) transformed with an expression plasmid of a galectin in modified Terrific broth and cultured overnight at 37  C at 250 rpm. For 2 L of bacterial culture, 3 flasks (2 L capacity) are used. 2. Add IPTG (final concentration 1 mM) for 3 h at 37  C at 250 rpm to induce the expression of recombinant galectin. 3. Collect bacteria by centrifugation at 7500  g for 20 min at 4  C. 4. Resuspend the bacterial pellet obtained from 2 L culture suspension in 50 mL of the ice-cold lysis-sonication buffer. 5. Sonicate the suspension solution (at 120 W for 30 s) in ice. Repeat sonication eight times with 1 min interval between the sonication. 6. Centrifuge the lysate at 112,500  g for 30 min at 4  C using T70.1 rotor. 7. Apply the supernatant to lactose-agarose column (1 mL bed volume) (see Notes 2 and 3). Galectin’s affinity for β-galactoside is used to purify galectin. In this way, nearly 100% of purified galectin is active as a lectin.

44

Guillaume St-Pierre et al.

8. Wash the column thoroughly with 20 mL PBS () (see Note 4). 9. Elute galectin that is bound to lactose-agarose in 1 mL fractions by using PBS containing 150 mM lactose. A total of 10–15 fractions are collected. 10. Detect the fractions that contain galectin by using Bradford protein assay kit and collect the blue positive fractions. In this condition, 3–5 fractions contain galectin. 11. Remove lactose. There are two options for the removal (see Note 5). (a) Gel filtration (see Note 6): Apply the pooled fractions (maximum volume to be applied is 15 mL) to a HiPrep 26/10 desalting column (bed volume 53 mL). Collect approximately 30 fractions of 1 mL by using PBS () at a flow rate of 100 mL/h. Detect galectin-containing fractions by using Bradford protein assay and pool the protein-positive fractions. (b) Dialysis: Lactose can be removed by dialysis. Transfer the pooled fractions to dialysis tube. Tightly seal the tube with dialysis clips. Submerge the dialysis tube in a beaker containing PBS () and a stirring magnet at 4  C. Change PBS () every 4–6 h four times (some can be done overnight). For example, for 5 mL galectin fractions that contain 150 mM lactose, 2 L of PBS is used for dialysis (1:400 dilution, 0.375 mM) and the dialysis tube is transferred to a new 2 L PBS every 4 to 6 h four times (total dilution is 1: 2.56  1010), final estimated lactose concentration would be 6 pM. 12. To remove endotoxin (see Note 7), pass the purified galectin on an Acticlean Etox column. Endotoxin activity is assessed by LAL assay by following the manufacturer’s instruction. 13. Finally, filter-sterilize the galectin-containing solution using syringe filter (0.22 μm). 14. Sterile galectin can be stored in 80  C with a notable exception of some galectins (see Notes 8–11). Some galectin can be kept in sterile condition at 4  C for a couple of weeks. 3.2 Carboxymethylation of Galectin-1

Galectin-1 can be stabilized by carboxymethylating cysteine residues, while it is important to verify whether carboxymethylation does not alter its biological activity of interest first by comparing the activity using both nonmodified galectin-1 and carboxymethylated galectin-1. When biotinylated galectin-1 is necessary, there are two options. Use carboxymethylated galectin-1 and use lysine residues for biotinylation (see Subheading 3.4). Another option is to use the

Galectin Purification, Characterization and Ligand Binding

45

cysteine residues of galectin-1 (see Subheading 3.5). In this way, galectin-1 is both biotinylated and relatively resistant against oxidation. Galectin-1 that is freshly eluted from lactose-agarose column using 150 mM lactose can be used for carboxymethylation (see step 10 of Subheading 3.1). Otherwise, lactose (final concentration is minimum 40 mM) should be added to protect its CRD. 1. Add iodoacetamide (final concentration 100 mM) to lactosecontaining galectin-1 and incubate overnight at 4  C. 2. Remove free iodoacetamide and lactose by either dialysis or gel filtration (see step 11 of Subheading 3.1). 3. Pass carboxymethlayted galectin-1 through Acticlean Etox column and filter-sterilize. 3.3 Quality Control of Purified Galectin by Hemagglutination Assay 3.3.1 Preparation of Red Blood Cells (RBC)

Quality control of purified galectins is critical to obtain reproducible results (see Note 11).

1. Collect 10 mL of blood in a heparinized tube. 2. Spin at 3500 rpm (2000  g) for 5 min and remove buffy coat as much as possible. 3. Wash RBC three times with 50 mL PBS (). 4. Dilute RBC at glutaraldehyde.

8%

(volume/volume)

in

PBS-3%

5. Rotate the solution slowly for 1 h at room temperature. 6. Wash fixed RBC five times with PBS-0.025% NaN3. 7. Fixed RBC can be stored at 4  C (see Note 12). 3.3.2 Calibration of RBC

It is important to predetermine the minimum concentration of RBC required for hemagglutination assay to avoid false-positive and negative results. 1. Make a series of RBC dilutions (2–20 fold). For example, add first PBS () (from 95 μL to 60 μL) and then 8% RBC (from 5 μL to 40 μL) to a well of a U-shaped 96-well plate. 2. Gently mix and incubate at 37  C for 30 min, followed by incubation at 4  C for 2 h. 3. Determine the minimum concentration of RBC that forms a tight button-like precipitation at the bottom of the well.

3.3.3 Hemagglutination Assay

1. For quality control of a recombinant galectin, prepare serial dilutions of the galectin, ranging from 0 to 10 μM, and mix with the appropriate quantity of RBC (see Note 13).

46

Guillaume St-Pierre et al.

2. Incubate at 37  C for 30 min, followed by incubation at 4  C for 2 h. 3. Evaluate hemagglutination. When RBC do not form tight button precipitation, hemagglutination is positive (galectin molecules cross-link RBC, thereby preventing the precipitation). When RBC form tight button precipitation as seen in the calibration of RBC, hemagglutination is negative. 3.4 Biotinylation of Galectins Using Primary Amine Residues of a Galectin

Biotinylated galectin can be used to detect its ligands on an immobilized material such as a membrane of western blot or a well by using a streptavidin-coupled detection system (see Subheading 3.6). It can be used to visualize galectin ligands in a cell or a tissue (see Subheading 3.8). For this method, there is an important point to be considered for sample preparation (see Note 14). 1. Add lactose (minimum final concentration 40 mM) to a purified galectin. 2. Option: Add Triton X-100 (final concentration 0.05%) to minimize the precipitation during the biotinylation step (see Note 15). 3. Add EZ-link sulfo-NHS-LC-biotin (tenfold excess) to 0.5 mg/mL galectin. EZ-link sulfo-NHS-LC-biotin solution should be prepared just before its usage as suggested in the manufacturer’s instruction. 4. Incubate for 2 h at 4  C. 5. Add 1 M Tris–HCl (pH 7.5) to obtain a final concentration of 50 mM Tris to terminate the reaction. 6. Remove lactose by using either dialysis or gel filtration (PD-10 or HiPrep 26/10 desalting column) following the manufacturer’s instruction (see Note 6). 7. Repurify biotinylated galectin by using lactose-agarose column to obtain active galectin and remove lactose (see Note 16).

3.5 Biotinylation of Galectin-1 Using Cysteine Residues

1. Add lactose (minimum final concentration 40 mM) to a purified galectin. 2. Add EZ-link™ Maleimide-PEG2-biotin (tenfold excess) to 0.5 mg/mL galectin-1. 3. Incubate for 2 h at 4  C. 4. Remove lactose by using either dialysis or gel filtration (PD-10 or HiPrep26/10 desalting column, following the manufacturer’s instruction). 5. Repurify biotinylated galectin by using lactose-agarose column to obtain active galectin-1 and remove lactose (see Note 16).

Galectin Purification, Characterization and Ligand Binding

3.6 Carbohydrate-Binding Specificity Analysis of Galectin: Enzyme-Linked Lectin Sorbent Assay

47

There are many ways to study the carbohydrate-binding specificity of galectins. Some of the examples are the frontal affinity chromatography, isothermal titration microcalorimetry, glycan microarrays, surface plasmon resonance, and enzyme-linked lectin sorbent assay [13]. Here, we explain an enzyme-linked lectin sorbent assay, one of the simplest ways, which does not require any specialized equipment (see Notes 17 and 18). In this assay, we test the capacity of a compound to inhibit the binding of galectin to immobilized asialofetuin, which is a glycoprotein carrying many β-galactoside residues. The optimal concentration of a galectin for the inhibition assay should be obtained first. We usually use the concentration that gives 50% of maximum binding. Then using the concentration, various concentrations of a compound of interest are added to estimate the inhibitory potency of the compound. 1. Incubate a well of an ELISA plate with 50 μL of asialofetuin (5 μg/mL in PBS) overnight at 4  C to coat the well with asialofetuin (see Note 19). 2. Rinse the well three times with washing buffer (20 mM Tris– HCl pH 7.5–0.05% Triton X-100). 3. Incubate the well with 200 μL of blocking buffer (10 mg/mL BSA (see Note 20) in washing buffer) for 2 h at 4  C. 4. Wash the well three times with ice-cold washing buffer. 5. Add 50 μL of biotinylated galectin in the presence or absence of a compound of interest and incubate for 2 h at 4  C (see Note 21). Various negative controls are necessary, including wells without either asialofetuin, galectin, or streptavidin-HRP. It is recommended to include lactose or thiodigalactoside as a general galectin inhibitor when optimizing the assay. 6. Wash the well three times with ice-cold washing buffer to remove unbound galectin (see Note 22). 7. Incubate the well with 50 μL of Streptavidin-HRP (1:15,000 in blocking buffer) for 1 h at 4  C. 8. Wash the well four times with ice-cold washing buffer. 9. Add 100 μL of TMB substrate solution and incubate at room temperature in dark. 10. Stop the reaction by adding 50 μL of stop solution (0.18 M H2SO4) and immediately measure the absorbance at 450 nm.

3.7 Carbohydrate-Binding Specificity Analysis of Galectin: SPR Assay

Another approach for the study of the carbohydrate-binding specificity of galectin is a surface plasmon resonance assay, where the interaction between a galectin and a glycan or compound can be measured in real time. The high sensitivity and low sample consumption are an advantage of this assay. One molecule (for example, either a galectin or a molecule of interest, which is expected to have affinity for galectin) is immobilized to the surface of a sensor

48

Guillaume St-Pierre et al.

chip and the other referred as an analyte (for example, either a potential ligand for galectin or a galectin, respectively) in free flow. Here, we briefly explain one approach in which an inhibitory potency of a compound is measured by a competitive assay. Thus, asialofetuin, which carries glycans that have an affinity for galectins, is coupled to the sensor chip, and two analytes, a galectin and a compound of interest, are in free flow. In this approach, various compounds can be tested using one SPR chip. This SPR condition also could create an environment relatively similar to a biological condition where a soluble galectin is interacting or sampling ligands. 1. Immobilize asialofetuin to a sensor chip using its primary amine by following the manufacturer’s instruction. Alternatively, biotinylated asialofetuin can be coupled to avidincoated chip. 2. Pass different concentrations of a galectin to obtain a dose response curve to choose an optimal concentration of galectin for the inhibition assay. 3. Prepare the solutions that contain both a fixed concentration of galectin and various concentrations of a compound of interest and another set of solutions that contain only the compound. 4. First, verify whether any concentrations of a compound of interest would not have any impact on the baseline before testing the interaction between galectin and the compound (see Note 23). 5. Then inject the mixture of galectin and the compound to test the inhibitory potency of a compound. 6. Plot the observed signal (RU) on the y-axis and the concentration in log on the x-axis to be able to calculate the concentration at which the compound inhibits 50% of the signal (IC50). The affinity of each compound for the galectin can be determined by comparing the IC50s of the compounds. A lower IC50 means higher affinity and vice versa. Please note that those IC50s are relative to the conditions used in the experiment (flow speed, galectin concentration, temperature, etc.). 3.8 Visualization of Ligands of a Galectin Expressed in Cells (Lectin Staining)

With a biotinylated galectin, galectin ligands expressed on the surface of cells can be detected by using fluorochrome-labeled streptavidin (see Note 24). Here a method of visualization of ligands of galectin-3 expressed on the surface of adherent cells is explained. There are a few essential points on the visualization of galectin ligands or treating cells with galectins to be considered (see Notes 25 and 26). 1. Fix adherent cells in freshly prepared 3.7% PFA solution for 15 min at room temperature.

Galectin Purification, Characterization and Ligand Binding

49

2. Incubate cells with 3% BSA (see Note 20) in PBS for 1 h at room temperature. 3. Wash gently with PBS. 4. Add biotinylated galectin-3 (60–120 μg/mL in PBS-1% BSA) for 30 min at room temperature. 5. Wash cells with PBS three times. 6. Incubate with Streptavidin-Alexa (1:1000 in 1% BSA-PBS) for 30–45 min at room temperature. 7. Wash with PBS three times. Cells are ready to be observed by a microscope. 3.9 Detection of a Galectin Expressed in Cells Using an Antibody against Galectin

Expression of a galectin in cells can be detected by using an antibody against the galectin (see Note 27). 1. Fix cells in freshly prepared 3.7% PFA solution for 15 min at room temperature. 2. (Option) Permeabilize cells by incubating 0.25% of Triton X-100-PBS for 5 min and then wash cells with PBS three times. 3. Incubate with 1 Carbo-free blocking solution for 1 h at room temperature. 4. Wash cells and incubate with anti-galectin antibody in 1% BSA-PBS for 1 h at room temperature. In the case of galectin-3, we normally use 2.5 μg/mL of biotinylated antiMac-2 antibody that binds to the N-terminal domain of galectin-3. 5. Wash cells with PBS three times. 6. Incubate with Streptavidin-Alexa (1:1000 in 1% BSA-PBS) for 30–45 min at room temperature. 7. Wash with PBS three times. Cells are ready to be observed by a microscope.

4

Notes 1. It is important to give a quick spin (13,000  g) with a tabletop microcentrifuge for 15 min at 4  C to remove any aggregates that may be formed during the storage. 2. It is possible to purify galectin by using a tag-based approach, such as GST (glutathione S-transferase) tag. It is still recommended to repurify those galectins through an affinity chromatography using a lactose-agarose column. In this way, all purified recombinant galectin has its lectin activity, while it remains elusive how much portion of galectin carries its lectin activity when only tag-based purification is used.

50

Guillaume St-Pierre et al.

3. We found that galectin-1 and 3 maintain their biding capacity to lactose and other glycan ligands in the presence of up to 1% Triton X-100. Thus, if necessary, Triton X-100 can be added to the lysis buffer as long as it is completely removed from final purified galectin, especially when purified galectin is used for cell biological assays. 4. It is important to completely remove EDTA and reducing agents that are used during the initial stage of purification at the step of lactose-agarose affinity chromatography, since any trace of reducing agents such as DTT (as low as 1 μM) inhibits following biological assays (but not biochemical assays, such as carbohydrate-binding assay) or introduce some artifact activities that are observed only when both galectin and a reducing agent are present [4]. 5. Lactose used for the elution of galectin from the column must be removed completely. 6. For gel filtration, it is essential to follow the instruction provided by GE, as there is a limit of volume that can be applied to a column. 7. In some cases, it is important to remove endotoxin from the purified galectins; thus, purified galectins are passed through Acticlean ETOX endotoxin-removing gels following manufacturer recommendations (Sterogene, Carlsbad, California) or any other type of column as long as the flow rate is well controlled to enable maximal interaction of samples with Acticlean. The endotoxin-binding capacity of this matrix is 20,000 EU/mL. Protein preparations that exceed 10 EU/ mg of protein are rejected or passed again through the ActiClean Etox endotoxin-removing gel. 8. It should be noted that galectins are not necessarily stable at 4  C. Thus it is recommended to freeze purified galectin if it is not planned for use within several days. Nonmodified galectin1 is extremely sensitive to oxidation. When nonmodified galectin-1 is required, it is recommended to be purified on the day of the experiment. To avoid oxidation, it is advisable to add a reducing agent to the purified galectin-1. However, it should be noted that the presence of a reducing agent may create some artifact phenomena in some biological assays. To avoid this complication, cysteine residues of galectin-1 can be carboxymethylated using iodoacetamide (see Subheading 3.2). 9. When a new aliquot of galectin is thawed, it is recommended to do quality control by using hemagglutination assay (see Subheading 3.3) and SDS-PAGE analysis. 10. In principle, purified galectin can be stable at 80  C for a relatively long time. One has to be cautious about the concentration of stored galectin, as some galectin (like tandem-repeat

Galectin Purification, Characterization and Ligand Binding

51

galectin) can be precipitated when the concentration is too high. For example, higher than 0.8 mg/mL of human galectin-9 (with a short linker) tends to precipitate. Modified galectins (such as biotinylated) are relatively prone to precipitation during storage. Repeated freezing and thawing should be avoided. Thus, it is recommended to prepare aliquots of a small amount of galectin for freezing. Nonetheless, it is highly recommended to check the quality of galectin each time when it is thawed. 11. It is also important to verify the molecular weight of purified galectin and the presence or absence of contaminant by using SDS-PAGE analysis. Thus, it is recommended to use 15% acrylamide gel for prototype galectins, such as galectin-1 and 12% for a chimera (galectin-3) and tandem-repeat type galectins, such as galectin-9 so that any smaller protein fragments below their molecular weights (their degraded products or contaminants) can be detected. In the case of galectin-3, it is important to check whether any protease-digested bands (between 15 kDa and 30–35 kDa) are absent, as the N-terminal domain can be cleaved by bacterial proteases, such as collagenases. This truncated CRD of galectin-3 cannot be removed by using a lactose-agarose. 12. Fixed RBC are stable at least for 3 months. If required, recalibration of RBC concentration is recommended. 13. We normally use 5 μL of 4% RBC in a 100 μL/well. Galectin-1 and 3 normally induce hemagglutination around 1 μM. Galectin-induced hemagglutination is blocked by 150 μM lactose. 14. Since the primary amine residues of galectin used to be biotinylated, it is essential to avoid any contamination of free primary amine residues, such as Tris buffer. Use either PBS () or HEPES buffer (pH 7.2–7.8) for the process. 15. In some cases, precipitation of galectin is observed during the biotinylation process. In this case, it is recommended to add Triton X-100 (final concentration 0.05%). After the biotinylation, the detergent is removed by repurifying galectin by using lactose-agarose column after removing lactose in the biotinylation solution. 16. Repurification of galectin after modification is highly recommended, as those modifications often inactivate some proportion of galectin. 17. In this assay design, a galectin ligand, asialofetuin, is immobilized onto the well. It is not recommended to directly coat wells with galectin, as galectin appears to be highly labile for the direct immobilization when its hydrophobicity is used for coating (unpublished observations), giving false-positive or

52

Guillaume St-Pierre et al.

negative results. If galectin is required to be immobilized, it is recommended to use either biotin–avidin interaction (biotinylated galectin and immobilized onto an avidin/streptavidincoated well) or the lysine residues of galectin to covalently linked to support instead of passive absorption of galectin. In any case, when galectin is required to be immobilized to a well, plate, array, or gel, it is highly recommended to perform quality control of the immobilized galectin, using established galectin ligands. 18. The style of enzyme-linked lectin sorbent assay can be changed according to the experimental objective. For example, it is possible to coat a compound of interest in the study of interaction with a galectin. When a compound is immobilized onto the well, it is important to verify whether the compound is indeed efficiently immobilized by using some plant lectins or antibodies. It is also important to test whether the binding of galectin to the compound is inhibited by lactose or not to verify whether the interaction is mediated by CRD of galectin. In some cases, the tertiary glycoprotein structure, which can present some glycans in a specific manner, may determine the interaction with a galectin. In this case, direct coating of the protein by using hydrophobicity of the material (a well, plate or array) is not recommended. For example, when the interaction of galectins with gp120 of HIV-1 is studied, such caution should be paid [14]. 19. Similar to normal ELISA, it is important to distribute the solution evenly on the well by gently tapping the plate. The plate should be sealed using an adhesive plate seal to inhibit any evaporation and kept in a moisture box. 20. While albumin is a non-glycosylated protein, some commercially available BSA contain other serum glycoproteins. When asialofetuin is used as the ligand of galectin on the well for the competitive assay, this contamination may not affect the assay, as fetuin may be one of the major contaminants from BSA preparation as long as the same lot of BSA is used. However, when a new glycan, such as glycan polymer, is used for coating to investigate whether a galectin binds to a glycan polymer, it is strongly recommended to test if BSA that is used for blocking the well is free of galectin ligands. 21. When biotinylated galectin is used for carbohydrate-binding assays or detection of galectin ligands, it is recommended to perform the assays in a sequential manner; first incubate with a biotinylated galectin, followed by streptavidin-coupled detection system, especially when galectin-3 is used. It appears that premixing biotinylated galectin-3 and streptavidin complex

Galectin Purification, Characterization and Ligand Binding

53

perturbs the formation of a stable interaction between galectin3 and its ligands (unpublished observation). 22. In the steps after this point, it is important not to leave the wells without liquid for more than a few minutes to avoid drying out. 23. It is extremely important to test all the concentrations of the compound, as the injection of some compounds induces background shift. For example, certain oligosaccharides that contain primary amine residues alter the background, and this phenomenon appears to be related to coating polysaccharides (such as carboxymethylated dextran and alginate polymer) (unpublished observation). 24. It is not recommended to co-stain cells or tissues with a plant lectin that binds to galectin’s ligand glycans, such as Lycopersicon esculentum lectin and Wheat germ agglutinin, since both would compete for the same ligands. In our experience, removal of galectin occurs even when the plant lectin staining was performed after galectin staining. 25. Caution should be taken when sugar antagonists are used to inhibit galectin’s lectin activity: Being β-galactoside-binding protein, the lectin activity can be readily inhibited by small β-galactose-containing oligosaccharides, such as lactose. Lactose is one of the most used binding inhibitors for galectins, although the affinity for lactose is often low and thereby high concentrations of lactose, such as 50–150 mM, are necessary to inhibit the lectin activity. When such a high concentration of lactose is applied to cells, it is critical to adjust the isotonicity of the buffer or medium containing lactose, as cells exposed to hypertonic solutions would become fragile to any successive treatment. Thus, the concentration of NaCl in the corresponding buffer or medium can be reduced for the accommodation of high doses of lactose, so that the lactosecontaining PBS has the same osmolality as saline (i.e., 317 mOsm/L). For example, to accommodate 200 mM lactose for PBS solution, it is necessary to remove 100 mmol NaCl (200 mOsmol) from 1 L of PBS. In other words, NaCl concentration is reduced to 2.15 g/L from 8 g NaCl/L, which is the concentration used for normal PBS. 26. As the majority of galectins are either divalent or multivalent in their glycan binding, galectin could aggregate cells when cells are used in suspension. Accordingly, one should be cautious when galectin-treated cells in suspension are subject to flow cytometry, since those aggregated cells would either not be subject to the analysis or be lost or damaged during the washing and resuspending steps.

54

Guillaume St-Pierre et al.

27. Since galectins are synthesized in the cytosol, their primary location is the cytosol. When the localization of galectins is not in the subject of interest, it is recommended to permeabilize the cells before staining with an antibody against galectin. If it is important to distinguish between the galectin bound on the surface of the cell membrane and the one present in the cytosol, do not permeabilize cells and directly label cells with the antibody. In this case, it should also be verified whether the fixation process (PFA fixation) would not compromise the membrane integrity, as some cells are sensitive to a high concentration of PFA. In this case, optimization for fixing condition is essential such as lowering the concentration of PFA. Please note that commercially available formalin solution, which is a stabilized form of formaldehyde solution, contains 10–12% methanol that can permeabilize cells. Thus, when the location of galectin is studied, avoid using formalin solution. References 1. Cummings RD, Liu F-T, Vasta GR (2015) Galectins in essentials of glycobiology, 3rd edn. Cold Spring Harbor Laboratory Press, New York. https://doi.org/10.1101/gly cobiology.3e.036 2. Hirabayashi J, Sato S, Leffler H (2018) Galectins updated: new discoveries, revisions and rebuttals. Trends Glycosci Glycotech 30: SE1–SE225 3. Hirabayashi J, Kasai K (1993) The family of metazoan metal-independent beta-galactosidebinding lectins: structure, function and molecular evolution. Glycobiology 3:297–304 4. Stowell SR et al (2007) Human galectin-1, -2, and -4 induce surface exposure of phosphatidylserine in activated human neutrophils but not in activated T cells. Blood 109:219–227 5. Ouellet M, St-Pierre C, Tremblay MJ, Sato S (2015) Effect of galectins on viral transmission. Methods Mol Biol 1207:397–420 6. Sato S, St-Pierre C, Bhaumik P, Nieminen J (2009) Galectins in innate immunity: dual functions of host soluble beta-galactosidebinding lectins as damage-associated molecular patterns (DAMPs) and as receptors for pathogen-associated molecular patterns (PAMPs). Immunol Rev 230:172–187 7. Chen H-Y, Weng I-C, Hong M-H, Liu F-T (2014) Galectins as bacterial sensors in the

host innate response. Curr Opin Microbiol 17:75–81 8. Nabi IR, Shankar J, Dennis JW (2015) The galectin lattice at a glance. J Cell Sci 128:2213–2219 9. Mendez-Huergo SP, Blidner AG, Rabinovich GA (2017) Galectins: emerging regulatory checkpoints linking tumor immunity and angiogenesis. Curr Opin Immunol 45:8–15 10. Dennis JW, Nabi IR, Demetriou M (2009) Metabolism, cell surface organization, and disease. Cell 139:1229–1241 11. Kamili NA et al (2016) Key regulators of galectin-glycan interactions. Proteomics 16:3111–3125 12. Sato S (2018) Cytosolic galectins and their release and roles as carbohydrate-binding proteins in host–pathogen interaction. Trends Glycosci Glycotech 30:SE199–SE209 13. Iwaki J, Hirabayashi J (2018) Carbohydratebinding specificity of human Galectins: an overview by frontal affinity chromatography. Trends Glycosci Glycotech 30:SE137–SE153 14. St-Pierre C et al (2011) Host-soluble galectin promotes HIV-1 replication through a direct interaction with Glycans of viral gp120 and host CD4. J Virol 85:11742–11751

Chapter 6 Expression, S-Nitrosylation, and Measurement of S-Nitrosylation Ratio of Recombinant Galectin-2 Mayumi Tamura and Yoichiro Arata Abstract S-nitrosylation, which involves the coupling of an NO group to the reactive thiol of Cys residue(s) in a polypeptide, is an important posttranslational modification detected in a variety of proteins. Here, we present the S-nitrosylation of recombinant galectin-2 (Gal-2) using S-nitrosocysteine and the measurement of the molecular ratio of S-nitrosylation of Cys residues in the Gal-2 protein. Key words Galectin, Galectin-2, Gal-2, S-nitrosylation, Cysteine, Saville-Griess assay

1

Introduction Gal-2 is a member of the galectin family that was originally found to be highly expressed in gastric cells and predominantly in the epithelial cells of the rat stomach [1]. Gastric mucous cells also showed strong Gal-2 immunoreactivity, with expression observed in the small intestine [2]. Furthermore, Gal-2 mRNA was detected in the human stomach [3]. Gal-2 contains two highly conserved cysteine residues in its polypeptide chain, and Gal-2 was identified by screening mouse gastric mucosal proteins that are uniquely sensitive to S-nitrosylation [4], which is an important posttranslational modification at the Cys residue(s) of a variety of proteins [5]. In human and mouse Gal-2, two cysteines exist at positions 57 and 75 (Fig. 1). We have shown that mouse Gal-2 loses its sugarbinding and hemagglutination abilities through oxidative inactivation after H2O2 treatment, which can be prevented by CysNO pretreatment [6]. Furthermore, experiments using the site-directed mutagenesis of proteins at these two positions suggested that two cysteine residues are S-nitrosylated by CysNO treatment [6]. We also identified the Cys residue in Gal-2, which is responsible for oxidative inactivation [7], and the structural mechanism underlying

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_6, © Springer Science+Business Media, LLC, part of Springer Nature 2020

55

56

Mayumi Tamura and Yoichiro Arata

Subunit 2 Subunit 1

Cys75

Cys75 Cys57

Cys57 Lactose

Lactose

Fig. 1 Location of two cysteine residues in the human Gal-2 dimer. The two peptide chains of the Gal-2 dimer are drawn as ribbons (PDB code 1HLC). Side chains of the two cysteine residues, Cys57 and Cys75, are shown in the space filling models and are indicated by arrows. The lactose molecules in complex with Gal-2 and the side chains of the amino acid residues in Gal-2 that are important for interaction with lactose are indicated in the ball–and-stick model

the S-nitrosylation-derived protection of Gal-2 against the oxidative inactivation by NMR [8]. These results suggested that Gal2 may have a protective role in the stomach due to NO-mediated S-nitrosylation. In this chapter, we present a detailed procedure for the expression and purification of recombinant Gal-2, S-nitrosylation of the recombinant proteins, and detection of this S-nitrosylation. S-nitrosylation was detected using a method based on the classical Saville-Griess assay [9, 10], wherein the S-nitrosylated protein produces nitrite ions in the presence Cu+ ions, which then react with sulfanilamide to be converted into diazonium salt [11, 12]. This is coupled to N-1-naphthylethylenediaminde dihydrochloride, resulting in azo dye formation (Fig. 2), which can be spectrophotometrically quantified.

2

Materials Prepare all solutions using ultrapure water and analytical-grade reagents. All reagents should be stored at 4  C (unless indicated otherwise).

2.1 Preparation of Recombinant Gal-2

1. Luria Broth (LB)/ampicillin (Amp) medium: LB containing 125 μg/mL ampicillin. Use a 500 mL shaking Erlenmeyer flask with baffles for 250 mL of LB media. Sterilize by autoclaving (at 120  C for 15 min). After cooling, add 250 μL of 125 mg/ mL Amp sterilized by filtration (see Note 1).

S-Nitrosylation of Galectin-2

Protein-Cys-SNO

Cu+

Protein-Cys-S- + NO

2NO + O2

2NO2

NO2 + NO

N2O3

NO2- + H2N

57

SO2NH2

H2O

2NO2- + H+ H+

N N+

SO2NH2

Sulfanilamide SO2NH2

HN N

N+

SO2NH2 + H2N N-1-Naphtylethylenediamine HN

H2N

N

N

SO2NH2

Azo dye

Fig. 2 Scheme for quantitation of S-nitrosylation in the protein using the Saville-Griess reaction. The S-nitrosylated protein (Protein-Cys-SNO in this figure) produces nitrite ions in the presence Cu+ ions, which then react with sulfanilamide to be converted into diazonium salt. This is coupled to N-1-naphthylethylenediaminde dihydrochloride, resulting in azo dye formation, which can be quantified spectrophotometrically

2. 0.1 M IPTG: Dissolve 0.24 g of isopropyl β-D-1-thiogalactopyranoside (IPTG) in water, to a total volume of 10 mL. Sterilize by filtration using a 0.2 μm filter unit (Advantec Toyo Kaisha, Ltd., Tokyo, Japan). Dispense 1 mL each in a 1.5 mL sample tube and store at 20  C until use. 3. 125 mg/mL ampicillin (Amp): Dissolve 1.25 mg of ampicillin sodium salt in water, up to total volume of 10 mL. Sterilize by filtration using a 0.2 μm filter unit. Dispense 0.5 mL each in a 1.5 mL sample tube and store at 20  C until use. 4. 20 EDTA-PBS (400 mM NaH2PO4, 3 M NaCl, 20 mM EDTA, pH 7.2): Dissolve 62.4 g of NaH2PO4·2H2O and 175.3 g of NaCl in approximately 700 mL of water, and add 40 mL of 0.5 M EDTA, pH 8.0. Adjust the pH of this solution to 7.2 using NaOH (see Note 2). Make up the solution to a total volume of 1 L with water. 5. EDTA-PBS: Dilute 20 EDTA-PBS 20-fold with water.

58

Mayumi Tamura and Yoichiro Arata

6. EDTA-ME-PBS: Dilute 20 EDTA-PBS 20-fold with water and add 2-mercaptoethanol to achieve a final concentration of 4 mM. Keep at 4  C until use. 7. EDTA-ME-PBS + 0.1 M lactose: Add 5 mL of 20 EDTAPBS and 3.6 g of lactose monohydrate to water and make up to a final volume of 100 mL. Add 2-mercaptoethanol to achieve a final concentration of 4 mM. Keep at 4  C until use. 8. Lactose-immobilized agarose column: Pack lactose agarose (can be purchased from Sigma) in an open column (bed volume; 5 mL). 9. Centrifugal filter unit: Amicon Ultra-15 (cut-off molecular weight 10,000, Merck). 2.2 S-Nitrosylation of Recombinant Gal-2

1. 100 mM HCl: Dilute 5 mol/L HCl, when needed. 2. 100 mM NaOH: Dilute 5 mol/L NaOH, when needed. 3. 200 mM L-cysteine: Transfer 12.1 mg of L-cysteine in a 1.5 mL sample tube and dissolve with 500 μL of 100 mM HCl. Prepare immediately before use. 4. 210 mM NaNO2: Transfer 7.2 mg of NaNO2 in a 1.5 mL sample tube and dissolve with 500 μL of 100 mM HCl. Prepare immediately before use. 5. EDTA-PBS: Dilute 20 EDTA-PBS 20-fold with water. 6. 1 mg/mL Gal-2: Dilute recombinant Gal-2 solution in EDTAPBS with EDTA-PBS to obtain a final concentration of 1 mg/ mL. 7. 25 mM CysNO: Add 100 μL of 200 mM L-cysteine and 100 μL of 210 mM NaNO2 to a 1.5-mL sample tube and incubate for 10 min by rotational mixing at room temperature in the dark. Add 200 μL of 100 mM NaOH and 400 μL of EDTA-PBS for neutralization and dilution. Measure the absorbance at 334 nm after diluting 25-fold with water to measure the CysNO concentration (ε ¼ 0.87 mM1 cm1) [13] (see Note 3). 8. Centrifugal filter unit: Amicon Ultra-0.5 (cut-off molecular weight 10,000, Merck).

2.3 Measuring SNitrosylation by Saville-Griess Assay

1. 0.5 mM S-nitrosoglutathione (GSNO): Transfer 4.2 mg of GSNO in a 1.5 mL sample tube and dissolve it in 500 μL of EDTA-PBS to make 25 mM GSNO. Dilute the solution 50 times to generate a 0.5 mM GSNO solution. Prepare immediately before use. GSNO will be used as a standard for the S-NO induction of proteins. 2. 10% sulfanilamide: Dissolve 1.0 g of sulfanilamide in 2.0 mL of 12 N HCl and make up to 10 mL by adding water (see Note 4). Keep the solution at 4  C under dark condition until use.

S-Nitrosylation of Galectin-2

59

3. 1% N-1-naphthylethylenediamine dihydrochloride: Dissolve 100 mg of N-1-naphthylethylenediamine dihydrochloride in water, making up to a final volume of 10 mL. Keep the solution at 4  C under dark condition until use. 4. 0.25% N-1-naphthylethylenediamine dihydrochloride: Dilute 1% N-1-naphthylethylenediamine dihydrochloride solution fourfold in water. Prepare immediately before use. 5. 10 mM CuCl2, 2.5% sulfanilamide: Measure 17.05 mg of CuCl2·2H2O and add to 2.5 mL of 10% sulfanilamide and 7.5 mL of water. Prepare immediately before use.

3

Methods

3.1 Expression of Recombinant Gal-2 in Escherichia coli

1. Select a single colony of E. coli BL21 (DE3) transformed with the Gal-2-expression plasmid (pET21a-Gal-2) (see Note 5) and inoculate it in a disposable 50-mL conical tube containing 10 mL LB/Amp medium. 2. Incubate overnight at 37  C. Prepare two tubes. 3. After the overnight incubation, inoculate the cultured media in two 500 mL shaking Erlenmeyer flasks with baffles containing 250 mL of LB/Amp medium. Shake the cultures at 37  C until their OD600 reaches between 0.6 and 0.8 (approximately 2–5 h of incubation). 4. Cool the flasks with iced water. 5. Add 1 mL of 0.1 M IPTG (to make a final concentration of 0.4 mM) and 0.25 mL of 125 mg/mL Amp to the culture media. Shake culture at 20  C overnight. 6. Recover the culture media into 50-mL conical tubes and centrifuge at 1580  g at 4  C for 20 min. 7. Remove the supernatant and suspend the remaining E. coli pellet in 5 mL of EDTA-ME-PBS in each tube. 8. Combine all suspensions in one tube and centrifuge again at 1580  g at 4  C for 15 min. 9. Remove the supernatant. 10. Suspend the E. coli pellet in 10 mL of EDTA-ME-PBS and disrupt the E. coli cells by sonication (see Note 6) on ice. Set the duty cycle to 50% and output control to 2.5. Sonicate for 5 min and cool down by incubating on ice for 3 min. 11. Repeat this sonication–cool down process three times. 12. After cell disruption, centrifuge (9400  g) for 45 min at 4  C. 13. Collect the supernatant fraction.

60

Mayumi Tamura and Yoichiro Arata

Protein conc. (mg/mL)

Lac 10 8 6 4 2 0

0

5

10

15

Fraction number

Fig. 3 Purification of recombinant Gal-2 by affinity chromatography on a lactoseimmobilized agarose column. Recombinant Gal-2 expressed in E. coli was applied to a lactose-immobilized agarose column. After the column was washed with EDTA-ME-PBS, specifically adsorbed Gal-2 was eluted with EDTA-ME-PBS containing 0.1 M lactose 3.2 Purification of Recombinant Gal-2 by Affinity Chromatography

1. Add the supernatant to a lactose-immobilized column (bed volume; 5 mL) equilibrated with 50 mL of EDTA-ME-PBS. The flow speed of the solution should be at 0.3–0.4 mL/min. 2. Collect the eluted fraction (5 mL each) in different tubes. 3. Wash the column with 6–8 times of the column’s bed volume with EDTA-ME-PBS. 4. Elute the adsorbed recombinant Gal-2 with 20 mL of EDTAME-PBS containing 0.1 M lactose (Fig. 3) (see Note 7). 5. Check the purification of recombinant Gal-2 by subjecting each eluted fraction to conventional SDS-polyacrylamide gel electrophoresis (see Note 8) and Coomassie Brilliant Blue staining. Collect the fractions containing lactose-eluted Gal2 and exchange the buffer to EDTA-PBS using Amicon Ultra15 to remove lactose and 2-mercaptoethanol (see Notes 9 and 10). 6. Add the lactose-eluted fraction containing Gal-2 to Amicon Ultra-15 and centrifuge until the volume in the filter device is reduced to about 1 mL. 7. Add an additional 10 mL of EDTA-PBS to the filter device and recentrifuge in the Amicon Ultra-15 device. 8. Repeat this process five times for buffer exchange to EDTAPBS (see Note 11).

3.3 S-Nitrosylation of Recombinant Gal-2

1. Add 25 μL of 25 mM CysNO to 500 μL of 1 mg/mL Gal-2 in EDTA-PBS. Prepare this mixture immediately before starting S-nitrosylation. 2. Incubate at room temperature for 60 min in the dark by rotary mixing (see Note 12).

S-Nitrosylation of Galectin-2

61

3. Remove the excess CysNO in the solution by buffer exchange. Transfer the reaction solution to the filter device of Amicon Ultra-0.5 and centrifuge the device until the volume of the solution in the filter device is reduced to less than 50 μL. 4. Add 450 μL of EDTA-PBS and centrifuge again. 5. Repeat this process three times (see Note 13). 6. Check the protein concentration of the final solution. 3.4 Measuring SNitrosylation of Gal2 by Saville-Griess Assay

1. Add 40 μL each of 0, 0.1, 0.2, 0.3, 0.4, and 0.5 mM GSNO solution to the wells of a flat-bottom 96-well plate. These wells are for generating a calibration curve. 2. Transfer 40 μL of 1 mg/mL Gal-2 solution to a different well after S-nitrosylation and removal of excess CysNO. 3. Add 80 μL of 10 mM CuCl2, 2.5% sulfanilamide solution, and 80 μL of 0.25% N-1-naphthylethylenediamine dihydrochloride solution to each well. 4. Incubate the 96-well plate for 20 min at 37  C. 5. Transfer the plate to a 4  C chamber to stop the reaction. 6. Measure the absorbance of each well at 540 nm using a microplate reader. 7. Generate a calibration curve and calculate the amount of NO in the S-nitrosylated Gal-2 solution. 8. Divide the molar concentration of NO in the reaction mixture (calculated in step 5) by the molar concentration of Gal-2 in the solution to obtain the molar ratio of S-nitrosylation for Gal-2 (number of NOs incorporated in one molecule of Gal-2) (see Note 14).

4

Notes 1. LB-Medium capsules (MP Biomedicals) are easy to handle and would make media preparation easier than weighing and dissolving media powder. 2. Adding approximately 12 g of NaOH pellets first and then adding NaOH solution slowly is recommended, since using an NaOH solution directly might not be enough for pH adjustment. Therefore, the total volume may exceed the targeted volume of 1 L. 3. Since the projected concentration of CysNO produced in the solution is much more than needed for the S-nitrosylation reaction, the solution can be used for S-nitrosylation without adjusting CysNO concentration for the following reaction, if the measured CysNO in the solution is in the range of 0.8–1.0 mM.

62

Mayumi Tamura and Yoichiro Arata

4. To dissolve sulfanilamide, use HCl at a concentration of 12 N. It is difficult to dissolve the sulfanilamide powder by HCl at a concentration around 5 N and often results in precipitants in the tube. Do not use the solution if precipitants occur. 5. Occasionally, for unknown reasons, the expression of the recombinant protein or the solubility of the expressed recombinant protein becomes substantially low. In these cases, obtaining a new expression clone by redoing the E. coli transformation using the stocked expression plasmid sometimes helps. 6. We use the TOMY, UD-211 system for sonication. 7. If you do not have enough time for the following buffer exchange procedure, we recommend leaving the eluted recombinant Gal-2 in the EDTA-ME-PBS solution containing 0.1 M lactose at 4  C rather than stopping during the buffer exchange process. Gal-2 is quite stable in the presence of lactose and 2-mercaptoethanol and can be stored for 3 days at 4  C. 8. We use a 15% gel. 9. Since galectins contain cysteine residues, which are important for their lectin activity, 2-mercaptoethanol is usually important. However, for the purpose of S-nitrosylation, 2-mercapthoethanol should be removed, since it inhibits the reaction. 10. We do not recommend buffer exchange by dialysis of a highconcentration Gal-2 solution, because this often results in precipitation of the Gal-2 protein. 11. Gal-2 in EDTA-PBS can be stored at 80  C in aliquots up to 12 months, without losing its activity. Avoid repeated freezing and thawing or storage at high concentration (higher than approximately 10 mg/mL), since these could result in insoluble precipitates of Gal-2. 12. Since the S-NO created by the –SH on the cysteine residue of the protein is sensitive to light, the tube must be covered by aluminum foil to avoid light exposure during the experimental process. 13. S-nitrosylated protein solution can be stored at 80  C for a few days with little to no release of NO from the cysteine residue under tight coverage of aluminum foil to avoid exposure to light. 14. When the obtained results of molar ratio of S-nitrosylation are substantially lower than expected, light exposure during the experimental process should be assessed by checking whether aluminum foil coverage was proper. Sometimes, unexpected light exposure during the experimental process leads to lower S-nitrosylation efficiency.

S-Nitrosylation of Galectin-2

63

References 1. Oka T, Murakami S, Arata Y et al (1999) Identification and cloning of rat galectin-2: expression is predominantly in epithelial cells of the stomach. Arch Biochem Biophys 361:195–201 2. Nio-Kobayashi J, Takahashi-Iwanaga H, Iwanaga T (2009) Immunohistochemical localization of six galectin subtypes in the mouse digestive tract. J Histochem Cytochem 57:41–50 3. Saal I, Lensch M, Lohr M et al (2005) Human galectin-2: expression profiling by RT-PCR/ immunohistochemistry and its introduction as a histochemical tool for ligand localization. Histol Histopathol 20:1191–1208 4. Ohtake K, Shimada N, Uchida H et al (2009) Proteomic approach for identification of protein S-nitrosation in mouse gastric mucosa treated with S-nitrosoglutathione. J Proteome 72:750–760 5. Hess DT, Matsumoto A, Kim SO et al (2005) Protein S-nitrosylation: purview and parameters. Nat Rev Mol Cell Biol 6:150–166 6. Tamura M, Saito M, Yamamoto K et al (2015) S-nitrosylation of mouse galectin-2 prevents oxidative inactivation by hydrogen peroxide. Biochem Biophys Res Commun 457:712–717

7. Tamura M, Sasai A, Ozawa R et al (2016) Identification of the cysteine residue responsible for oxidative inactivation of mouse galectin2. J Biochem 160:233–241 8. Sakakura M, Tamura M, Fujii N et al (2018) Structural mechanisms for the S-nitrosylationderived protection of mouse galectin-2 from oxidation-induced inactivation revealed by NMR. FEBS J 285:1129–1145 9. Saville BB (1958) A scheme for the colorimetric determination of microgram amounts of thiols. Analyst 83:670–672 10. Griess P (1879) Amidos€auren mit Alkoholradikalen. Chem Ber 12:426–428 11. Williams DL (1996) S-Nitrosothiols and role of metal ions in decomposition to nitric oxide. Methods Enzymol 268:299–308 12. Hoffmann J, Dimmeler S, Haendeler J (2003) Shear stress increases the amount of S-nitrosylated molecules in endothelial cells: important role for signal transduction. FEBS Lett 551:153–158 13. Gordge MP, Meyer DJ, Hothersall J et al (1995) Copper chelation-induced reduction of the biological activity of S-nitrosothiols. Br J Pharmacol 114:1083–1089

Chapter 7 Expression and Purification of Full-Length and Domain-Fragment Recombinant Pentraxin 3 (PTX3) Proteins from Mammalian and Bacterial Cells Kenji Daigo and Takao Hamakubo Abstract Although cell-based protein expression systems enable us a certain amount of protein suitable for subsequent biological experiments to be obtained, aggregates of the protein of interest are sometimes encountered during the purification procedure. Pentraxin 3 (PTX3), a member of the pentraxin family that is classified as a carbohydrate-binding protein based on its structure, comprises one of the humoral arms of the pattern recognition receptors that play an important role in the innate immune response. PTX3 comprises two domains; an N-terminal domain and a C-terminal domain. The C-terminal domain containing pentraxin signature has similar biological functions as other pentraxins such as C-reactive protein (CRP) and serum amyloid-P component (SAP). On the other side, the N-terminal domain is specific to PTX3. A supply of the PTX3 protein in full length or partial fragments is thus essential for the elucidation of its biological functions. Here we describe the expression and purification of recombinant PTX3. An argininecontaining buffer is essential for the elution of bacterially expressed PTX3 N-terminal domain to minimize aggregation. This method allows high-yield purification of full-length or domain-fragment recombinant PTX3 proteins for biological study. Key words Pentraxin 3, PTX3, Aggregation, Stable cell line, Protein expression, Protein purification, Arginine

1

Introduction Pentraxin 3 (PTX3) is the first-identified long pentraxin that is comprised of an approximate 200-amino acid N-terminal domain and a C-terminal pentraxin domain that contains the pentraxin signature (HxCxS/TWxS, where x is any amino acid), which is highly conserved among the pentraxin family members [1– 3]. Although most of the pentraxins assemble as a radial symmetric multimer, PTX3 forms an asymmetric octamer assisted by intermolecular disulfide bonds [4]. The non-redundant role of PTX3 against pathogens has been well studied [2–5]. Certain microbes are recognized by PTX3, and PTX3 exerts opsonic activity with the

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_7, © Springer Science+Business Media, LLC, part of Springer Nature 2020

65

66

Kenji Daigo and Takao Hamakubo

assistance of complement as well as the Fc-gamma and CR3 receptor. PTX3 contributes not only to pathogen resistance but also to the resolution of inflammation [6], female fertility [4, 7], extracellular matrix assembly [8], and oncosuppression [9]. Nevertheless, the pentraxins are categorized as a carbohydratebinding protein family [10], no binding to carbohydrate has been reported for PTX3 thus far. A supply of the recombinant PTX3 protein, including its specific domain fragment, is needed to determine the functions of PTX3 or the domains responsible, both in vitro and in vivo. To investigate the importance of multimer formation of PTX3, a cysteine residues mutated form should be also requierd [4]. Here we describe the production and purification of PTX3 recombinant proteins from mammalian and bacterial cells. These recombinant proteins have been successfully applied in both in vitro and in vivo assays [11, 12]. In the mammalian expression system, we were able to obtain either full-length or PTX3 domain proteins (we call domainfragment hereafter). Since the C-terminal PTX3 domain is glycosylated [13], the bacterial expression system was used only for N-terminal domain. The N-terminal domain of monomeric PTX3 mutein was also obtained. We were confronted with complete loss of the bacterially-expressed PTX3 proteins during the elution step and storage due to aggregation. The addition of arginine in the elution and storage buffers [14, 15] is effective for keeping the samples free of aggregation.

2

Materials

2.1 Plasmids for Recombinant PTX3 Protein Expression

1. Plasmid pEF4/Myc-His B vector: this is the expression vector for mammalian cells. For full-length and domain-fragment PTX3 expression, the coding sequences of human PTX3 (residues 1-381), the N-terminal domain of PTX3 (residues 1-178), and the C-terminal domain of PTX3 (179-381 amino acids) with an adaptation of the N-terminal signal sequence of PTX3 (residues 1-17) are cloned into the NotI/ XbaI site of the vector. A schematic representation of the PTX3 expression vectors for mammalian cells is shown in Fig. 1a. The PTX3 recombinant proteins obtained are provided in Fig. 1b. Secreted recombinant PTX3 proteins lack a signal sequence and have a Myc-6xHis amino acid sequence at their C-terminal site, which enables Ni-column purification. The vectors were supplied by Perseus Proteomics Inc. 2. Plasmid pCold II vector: this is the expression vector for bacterial cells. For N-terminal domain-fragment expression, the coding sequence of the N-terminal domain of PTX3 (N-term. Wild) that lacks an N-terminal signal sequence (residues 18–178, see Note 1), with an adaptation of the Tobacco etch

Expression and Purification of Recombinant Pentraxin 3 (PTX3)

67

a PTX3_N

pEF4/Myc-His B

PTX3_C

pEF4/Myc-His B

PTX3_Full pEF4/Myc-His B

b PTX3_N

NH3+

Myc-6xHis COONH3+

PTX3_C

Myc-6xHis COO-

PTX3_Full NH3+

Myc-6xHis COO-

c N-term. pCold II Wild

TEV

N-term. pCold II Mutant

TEV C49S C47S

C103S

d N-term. NH3+-6xHis TEV Wild N-term. NH3+-6xHis TEV Mutant

-COO-COOC49S C47S

C103S

Signal sequence

N-terminal domain

C-terminal domain

(1~17 amino acid)

(18~178 amino acid)

(179~381 amino acid)

Fig. 1 Schematic representation of the recombinant PTX3 expression vectors and obtained proteins. (a) Expression vectors for mammalian cells. The coding sequences of the N-terminal domain of PTX3 (residues 1–178) and C-terminal domain of PTX3 (residues 179–381) with an adaptation of the N-terminal signal sequence of PTX3 (residues 1–17) and full-length PTX3 (residues 1–381) are cloned into the mammalian expression vector pEF4/Myc-His B. (b) Composition of PTX3 proteins expressed from the vectors for mammalian cells. Note that they lack a signal sequence, and Myc and 6xHis are attached to their C-terminal sites. (c) Expression vectors for bacterial cells. For N-terminal domain PTX3 fragment expression, the coding sequence of the N-terminal domain of PTX3 that lacks N-terminal signal sequence (residues 18–178), with an adaptation of the Tobacco etch virus (TEV) protease sequence (ENLYFQG) in the N-terminal, is cloned into the NdeI/XhoI site of the plasmid pCold II vector. (d) Composition of PTX3 the proteins expressed from the vectors for the bacterial cells. Note that 6xHis and the TEV sequence are attached to their N-terminal sites

virus (TEV) protease sequence (ENLYFQG) in the N-terminal, as well as a mutant sequence in which the cysteine residues (47, 49, and 103) are changed into serine residues (N-term. Mutant), are respectively cloned into the NdeI/XhoI site of the vector. A schematic representation of the PTX3 expression vector for bacterial cells is shown in Fig. 1c. The PTX3 recombinant proteins obtained are shown in Fig. 1d. The expressed recombinant N-terminal protein has a 6xHis

68

Kenji Daigo and Takao Hamakubo

amino acid sequence at its N-terminal site, which enables Ni-column purification (Fig. 2). 1. FreeStyle™ CHO-S Cells (Thermo Fisher Scientific) (see Note 2). 2. Culture medium: FreeStyle™ CHO Expression Medium supplemented with 8 mM L-Gultamate, 0.5% Pluronic F-68 and antibiotics. 3. FreeStyle™ MAX Reagent (Thermo Fisher Scientific).

2

3

4

5

6

(kDa) 250 150 100 75 50

50 37

37

25 20

25 20

15

15

b

PTX3_N

1

250 150 100 75

(kDa) 1 150 100 75 50 37

2

3

4

5

(kDa)

N-term. Mutant

(kDa)

PTX3_Full

a

PTX3_C

4. Zeocin™, 100 mg/mL (Thermo Fisher Scientific).

N-term. Wild

2.2 Cell Culture for the Mammalian Expression System

250 150 100 75 50 37

25 20 15

25 20

Fig. 2 Results of recombinant full-length or domain-fragment PTX3 protein purification. (a) Left panel: Protein staining of recombinant full length of PTX3 protein purification from mammalian cell culture supernatant. Fractions derived from Subheading 3.4 were subjected to protein staining. Input (lane 1): culture supernatant. Post_column (lane 2): Ni-NTA flow-through fraction. Lane 3: wash fraction. Lane 4: eluate obtained with elution buffer 1. Lane 5: eluate obtained with elution buffer 2. Lane 6: eluate obtained with elution buffer 3. Right panel: Protein staining of purified recombinant PTX3 proteins from mammalian cell culture supernatant. From left to right, the full-length (PTX3_Full), the C-terminal domain (PTX3_C), and the N-terminal domain (PTX3_N) of human PTX3 along with Myc-Hisx6-tagged proteins expressed by mammalian cells. (b) Left panel: Protein staining of the recombinant N-terminal domain of PTX3 protein

Expression and Purification of Recombinant Pentraxin 3 (PTX3)

69

5. Erlenmeyer flask, 125 mL and 500 mL polycarbonate with a vented cap (Corning). 6. An orbital shaker and CO2 incubator. 7. ProClin™ 300 (Sigma-Aldrich). 1. BL21-CodonPlus-RIL cells (Agilent Technologies).

2.3 Expression of the N-Terminal Domain of PTX3 by Bacterial Cells

2. An LB agar plate containing 50 μg/mL of ampicillin. 3. LB medium containing 50 μg/mL of ampicillin. 4. Two shakers with temperature controls, one to be set at 37  C and the other at 15  C. 5. 1 M isopropyl β-D-1-thiogalactopyranoside (IPTG).

2.4 Purification of Recombinant PTX3 Proteins from Mammalian Cells with a 6xHis Tag Affinity Column

1. HisTrap HP 1 mL column (GE Healthcare).

2.5 Purification of Recombinant N-Terminal Domain PTX3 Proteins from Bacterial Cells with a 6xHis Tag Affinity Column

1. BugBuster master mix (Novagen).

2. Binding and wash buffer: 20 mM sodium phosphate, 500 mM NaCl, 20 mM imidazole, pH 7.4 (Table 1 for PTX3_Ful, PTX3_C and PTX3_N). 3. Elution buffer: 20 mM sodium phosphate, 500 mM NaCl with various concentrations of imidazole, pH 7.4 (Table 1 for PTX3_Ful, PTX3_C and PTX3_N).

2. Complete™, Mini, EDTA-free protease inhibitor cocktail (Roche). 3. HisTrap HP 1 mL column (GE Healthcare). 4. Binding and wash buffer: 20 mM sodium phosphate, 500 mM NaCl, 20 mM imidazole, pH 7.4 (Table 1 for bacterially expressed PTX3_N_wild and PTX3_N_mut). ä Fig. 2 (continued) purification from a bacterial cell expression system. The fractions derived from Subheading 3.5 were subjected to protein staining. Input (lane 1): bacterial cell lysate. Post_column (lane 2): Ni-NTA flow-through fraction. Lane 3: wash fraction. Lane 4: eluate obtained with elution buffer 1. Lane 5: eluate obtained with elution buffer 2. Right panel: Protein staining of purified recombinant N-terminal domain of PTX3 protein from a bacterial cell expression system. From left to right, the PTX3 N-terminal domain (N-term. Wild) and the N-terminal domain with all of the cysteine residues mutated (N-term. Mutant) along with Hisx6-tag and tobacco etch virus (TEV) protease sequence tagged proteins. (Right panels of (a and b) were originally published in Molecular & Cellular Proteomics. Daigo K, Yamaguchi N, Kawamura T, Matsubara K, Jiang S, Ohashi R, Sudou Y, Kodama T, Naito M, Inoue K, Hamakubo T. The proteomic profile of circulating pentraxin 3 (PTX3) complex in sepsis demonstrates the interaction with azurocidin 1 and other components of neutrophil extracellular traps. Mol Cell Proteomics. 2012 Jun;11(6):M111.015073. © The American Society for Biochemistry and Molecular Biology, Inc.)

70

Kenji Daigo and Takao Hamakubo

Table 1 Buffer compositions for the purification of recombinant PTX3 proteins Recombinant PTX3 protein

Washing buffer

PTX3_Full

Elution buffer 1

Elution buffer 2

Elution buffer 3

Basic buffer + 20 mM imidazole

Basic buffer + 50 mM imidazole

Basic buffer + 150 mM imidazole

Basic buffer + 500 mM imidazole

PTX3_C

Basic buffer + 5 mM imidazole

Basic buffer +50 mM imidazole

Basic buffer + 150 mM imidazole

Basic buffer + 500 mM imidazole

PTX3_N

Basic buffer + 20 mM imidazole

Basic buffer + 50 mM imidazole

Basic buffer + 150 mM imidazole

Basic buffer + 500 mM imidazole

Bacterially Basic buffer + expressed 20 mM PTX3_N_wild imidazole

Basic buffer + 250 mM imidazole, 0.2 M arginine

Basic buffer + 500 mM imidazole, 0.2 M arginine



Bacterially Basic buffer + expressed 20 mM PTX3_N_mut imidazole

Basic buffer + 250 mM imidazole, 0.2 M arginine

Basic buffer + 500 mM imidazole, 0.2 M arginine



Basic buffer: 20 mM sodium phosphate, 500 mM NaCl, pH 7.4

4. Elution buffer: 20 mM sodium phosphate, 500 mM NaCl, 200 mM arginine with various concentrations of imidazole, pH 7.4 (Table 1 for bacterially expressed PTX3_N_wild and PTX3_N_mut).

3

Methods

3.1 Establishment of Recombinant PTX3 Expressing Stable Mammalian Cell Lines

1. Transfect CHO-S cells with the vectors described in step 1 of Subheading 2.1 using FreeStyle™ MAX Reagent following the manufacturer’s instructions. 2. Maintain transfected cells in culture medium containing 0.5 mg/mL Zeocin starting 2 days after transfection. 3. Start the selection of PTX3-expressing cells by limited dilution 7 days after Zeocin treatment (see Note 3). 4. Freeze selected instructions.

3.2 Large-Scale Culture of PTX3-Expressing Stable Cell Lines

clones

following

the

manufacturer’s

1. Thaw the Cryovial of stable cells quickly in a 37  C water bath. 2. Suspend the thawed cells in culture medium and start to culture cells in the Erlenmeyer flask on an orbital shaker at 125 rpm. 3. Monitor cell viability and growth every day after thawing. If the cells have evidently been growing for several days, add 0.5 mg/

Expression and Purification of Recombinant Pentraxin 3 (PTX3)

71

mL Zeocin in the culture medium and keep growing the cells with culture media plus Zeocin (see Note 4). 4. Subculture cells to a density of 0.4  106 cells/mL every 2–4 days. 5. When the total cell number reaches 1  108, subculture cells to a density of 0.4  106 cells/mL in 250 mL culture medium with Zeocin. 6. When the cell density becomes more than 2  106 cells/mL, subculture cells to a density of 1  106 cells/mL in 500 mL culture medium with Zeocin and culture for 7 days. 7. After 7 days of culture, collect the culture supernatant and centrifuge at 2000  g for 30 min. Collect the supernatant and add the biocide reagent 0.1% ProClin ™ 300 and store at 20  C. 3.3 Expression of the N-Terminal Domain of PTX3 by Bacterial Cells

1. Transform BL21-CodonPlus-RIL cells with the vectors for bacterial expression described in step 2 of Subheading 2.1, and select transformed cells on an LB agar plate containing ampicillin. 2. Inoculate with 3 mL of LB medium containing ampicillin for ampicillin-resistant colonies, and then incubate at 37  C with shaking for 6 h. 3. Transfer 1 mL of the culture medium into 20 mL of LB medium containing ampicillin (see Note 5), then incubate at 37  C with shaking for 16 h. 4. Add all culture medium into 500 mL of the LB medium containing ampicillin, and then incubate at 37  C with shaking. 5. When the OD600 of the culture medium reaches 0.4–0.5 (see Note 6), transfer the culture to an incubator precooled to 15  C and let stand for 30 min (see Note 7). 6. Add 500 μL of 1 M IPTG (final 1 mM), and start culturing at 15  C with shaking for 24 h. 8. Collect the cells by centrifugation at 6000  g for 20 min. Store the cell pellet at 20  C.

3.4 Purification of Recombinant PTX3 Proteins from the Mammalian Cell Culture Supernatant with a 6xHis Tag Affinity Column

1. Thaw the frozen culture supernatants obtained from step 7 of Subheading 3.2. 2. Centrifuge the supernatants at 8000  g for 30 min, and then pass them through a 0.45 μm filter membrane. 3. Equilibrate a HisTrap HP 1 mL column with 10 mL of binding buffer (see Note 8).

72

Kenji Daigo and Takao Hamakubo

4. Apply the culture supernatants with the maximum flow rate of 2 mL/min. Collect the flow-through. 5. Wash the column with 400 mL of the binding buffer with the maximum flow rate of 2 mL/min. 6. Elute with 5 mL each of the elution buffers 1–3 with the maximum flow rate of 2 mL/min. Collect each eluate in 5 mL fractions. Typical purification result is shown in Fig. 2a. 3.5 Purification of Recombinant N-Terminal Domain PTX3 Proteins from Bacterial Cells with a 6xHis Tag Affinity Column

1. Add protease inhibitor cocktail into the BugBuster master mix prior to use. 2. Resuspend the bacterial cell pellets obtained in step 8 of Subheading 3.3 with the BugBuster master mix plus protease inhibitor cocktail (10 mL/500 mL culture pellet) by gentle pipetting or rotating at room temperature. 3. Incubate the cell suspension by rotation or gentle shaking for 20 min at room temperature. 4. Remove the insoluble fraction by centrifugation at 16,000  g for 20 min at 4  C, and then pass the supernatant through a 0.45 μm filter membrane. 5. Equilibrate the HisTrap HP 1 mL column with 10 mL of binding buffer (see Note 8). 6. Apply the lysed supernatants with the maximum flow rate of 2 mL/min. 7. Wash the column with 100 mL of binding buffer with a maximum flow rate of 2 mL/min. 8. Elute with 5 mL each of the elution buffers 1–2 containing arginine with a maximum flow rate of 2 mL/min. Collect each fraction in 5 mL quantities. Typical purification result is shown in Fig. 2b. 9. Measure the absorbance at 280 nm (A280) for each of the elution fractions immediately after the elution process. If the A280 is greater than 0.5, dilute the eluents with the same component of the elution buffer to A280 ¼ 0.5 to prevent aggregation of recombinant PTX3 during storage (see Note 9).

4

Notes 1. The codon optimization that some companies offer can be employed to obtain maximum protein expression using bacterial cells. 2. CHO-S cells are suspension cells and can be cultured in serumfree medium. Use a shaker flask, spinner flask or bioreactor with appropriate apparatus (e.g. an orbital shaker or low profile

Expression and Purification of Recombinant Pentraxin 3 (PTX3)

73

roller) as desired. Note that the CHO-S cells are maintained under 8% CO2. 3. For high-throughput selection, we recommend sandwich ELISA using a polyclonal anti-PTX3 antibody for the first antibody and HRP-conjugated anti-myc or 6xHis antibody for the second antibody. Western blotting may be employed when there is only a limited number of clones. 4. Continue culturing the cells without Zeocin for at least 3 or 4 days after the thawing process. Immediate Zeocin addition may result in a loss of cell viability. When the stable cells start to grow, they reach a density of 1–2  106 cells/mL after 2–4 days of seeding at 0.4  106 cells/mL. 5. Glycerol stock may be prepared from the rest of the 3 mL culture medium. It can be used for the next expression experiment by scooping a tiny portion of stock cells into 3 mL of LB medium containing ampicillin, then start from step 2 of Subheading 3.3. 6. Typically, it takes 2 h to reach OD600 ¼ 0.4–0.5. 7. Shaking in not necessary during 15  C incubation. 8. Pass the all buffers through a 0.45 μm filter membrane before purification procedure. Use chilled buffers during the purification procedure. Retain a small amount of each fraction (lysate, flow-through, column wash, elution) so that protein staining and Western blotting against PTX3 can be performed. 9. You will not see any aggregate formation after thawing from 20  C storage if its A280 is less than 0.5. When you see any precipitation after thawing, centrifuge briefly and measure the A280 of the supernatant as protein concentration.

Acknowledgments This work was supported by the research program of social collaboration course at Nippon Medical School funded by FUJIFILM Corporation and JSPS KAKENHI Grant Number 18K07104. We thank Dr. Kevin Boru of Pacific Edit for reviewing the manuscript. References 1. Bottazzi B, Doni A, Garlanda C et al (2010) An integrated view of humoral innate immunity: pentraxins as a paradigm. Annu Rev Immunol 28:157–183 2. Daigo K, Inforzato A, Barajon I et al (2016) Pentraxins in the activation and regulation of innate immunity. Immunol Rev 274:202–217

3. Daigo K, Mantovani A, Bottazzi B (2014) The yin-yang of long pentraxin PTX3 in inflammation and immunity. Immunol Lett 161:38–43 4. Inforzato A, Rivieccio V, Morreale AP et al (2008) Structural characterization of PTX3 disulfide bond network and its multimeric status in cumulus matrix organization. J Biol Chem 283:10147–10161

74

Kenji Daigo and Takao Hamakubo

5. Garlanda C, Hirsch E, Bozza S et al (2002) Non-redundant role of the long pentraxin PTX3 in anti-fungal innate immune response. Nature 420:182–186 6. Deban L, Russo RC, Sironi M et al (2010) Regulation of leukocyte recruitment by the long pentraxin PTX3. Nat Immunol 11:328–334 7. Scarchilli L, Camaioni A, Bottazzi B et al (2007) PTX3 interacts with inter-alpha-trypsin inhibitor: implications for hyaluronan organization and cumulus oophorus expansion. J Biol Chem 282:30161–30170 8. Doni A, Musso T, Morone D et al (2015) An acidic microenvironment sets the humoral pattern recognition molecule PTX3 in a tissue repair mode. J Exp Med 212:905–925 9. Bonavita E, Gentile S, Rubino M et al (2015) PTX3 is an extrinsic oncosuppressor regulating complement-dependent inflammation in cancer. Cell 160:700–714 10. Fujimoto Z, Tateno H, Hirabayashi J (2014) Lectin structures: classification based on the 3-D structures. Methods Mol Biol 1200:579–606

11. Daigo K, Nakakido M, Ohashi R et al (2014) Protective effect of the long pentraxin PTX3 against histone-mediated endothelial cell cytotoxicity in sepsis. Sci Signal 7:ra88 12. Daigo K, Yamaguchi N, Kawamura T et al (2012) The proteomic profile of circulating pentraxin 3 (PTX3) complex in sepsis demonstrates the interaction with azurocidin 1 and other components of neutrophil extracellular traps. Mol Cell Proteomics 11:M111.015073 13. Inforzato A, Reading PC, Barbati E et al (2012) The “sweet” side of a long pentraxin: how glycosylation affects PTX3 functions in innate immunity and inflammation. Front Immunol 3:407 14. Arakawa T, Ejima D, Tsumoto K et al (2007) Suppression of protein interactions by arginine: a proposed mechanism of the arginine effects. Biophys Chem 127:1–8 15. Tsumoto K, Umetsu M, Kumagai I et al (2004) Role of arginine in protein refolding, solubilization, and purification. Biotechnol Prog 20:1301–1308

Chapter 8 Identification of Siglec Cis-Ligands by Proximity Labeling Amin Alborzian Deh Sheikh, Chizuru Akatsu, and Takeshi Tsubata Abstract Siglecs are known to be bound and regulated by membrane molecules that display specific sialic acidcontaining ligands and are present on the same cell (cis-ligands). Because of the low-affinity binding of Siglecs to the glycan ligands, conventional methods such as immunoprecipitation are not suitable for identification of Siglec cis-ligands. Here we describe efficient and specific labeling of cis-ligands of CD22 (also known as Siglec-2) on B lymphocytes by proximity labeling using tyramide. This method may also be applicable to labeling of cis-ligands of other Siglecs. Key words Proximity labeling using tyramide, Tyramide signal amplification, Siglec, Cis-ligands

1

Introduction Sialic acid-binding immunoglobulin-like lectins (Siglces) are a family of type I membrane proteins expressed in various cell types, especially those of immune cells [1]. There are around 10 members in this family in both humans and mice. Most of them carry signaling function. Each member of the Siglec family is expressed in specific cell types and shows a specificity for the type and linkage of sialic acid. Siglecs are known to be bound and regulated by membrane molecules that display specific sialic acid-containing ligands and are present on the same cell (cis-ligands). Because of the low-affinity binding of Siglecs to the glycan ligands, conventional methods such as immunoprecipitation are not suitable for identification of Siglec cis-ligands. Tyramide signal amplification (TSA) (also known as catalyzed reporter deposition (CARD)) has been developed to amplify signals in various assays such as immunohistochemistry, immunoelectron microscopy, fluorescent in situ hybridization (FISH), and ELISA [2–7]. In TSA, horseradish peroxidase (HRP) bound to target molecules catalyzes oxidation of tyramide conjugated with biotin or fluorescein into an ultrashort-lived free radical that covalently binds to the nearby tyrosine residues within 20 nm [8], resulting in

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_8, © Springer Science+Business Media, LLC, part of Springer Nature 2020

75

76

Amin Alborzian Deh Sheikh et al.

CD22

Biotin-tyramide

HRP-conjugated anti-CD22 Ab

Radical Biotintyramide

Biotin

Fig. 1 Schematic representation of proximity labeling using biotin-tyramide. Spleen B cells are first treated with HRP-conjugated anti-CD22 antibody and then treated with H2O2 and biotin-tyramide. Ultrashort-lived tyramide radical generated by HRP covalently binds to the tyrosine residues in the nearby proteins. (Reprinted from Biochem. Biophys. Res. Commun. vol. 495, Alborzian Deh Sheikh, A. et al., Proximity labeling of cis-ligands of CD22/Siglec-2 reveals stepwise α2,6 sialic acid-dependent and -independent interactions, 854–859, Copyright 2018, with permission from Elsevier)

signal amplification. As the nearby proteins are efficiently labeled by TSA, this method is also used for proximity labeling to identify the molecules in the close vicinity of the target molecules on the cell surface (Fig. 1) [9–12]. Recently, we demonstrated that cis-ligands of CD22 (also known as Siglec-2) can be easily identified by this proximity labeling using tyramide [12]. CD22, an inhibitory receptor expressed in B lymphocytes (B cells), specifically recognizes α2,6 sialic acid [1]. Proximity labeling using HRP-conjugated anti-CD22 antibody and biotinconjugated tyramide induces biotinylation of various proteins including CD22, CD45, and IgM [12], indicating that these molecules are located in the close vicinity of CD22. Treatment with GSC718, a synthetic sialoside that inhibits binding of CD22 with sialic acid, abolished biotinylation of various proteins in this reaction. Moreover, proximity labeling failed to label proteins in B cells deficient in the sialyl transferase ST6GalI required for the generation of α2,6 sialic acid. These results clearly demonstrate that proximity labeling of the proteins including CD22, CD45, and IgM depends on the recognition of α2,6 sialic acid by CD22. Thus, proximity labeling using tyramide can efficiently label CD22 cis-ligands. Here we describe the detailed procedures to identify CD22 cisligands in primary spleen B cells by proximity labeling using tyramide. One of the features of this method is to identify cis-ligands

Identification of Siglec Cis-Ligands

77

on alive cells. This method may be applicable for the identification of the cis-ligands of other Siglecs. To confirm that the labeled proteins are Siglec cis-ligands, the lack of labeling in the absence of the interaction between Siglecs and sialic acid must be shown. For this purpose, cells deficient in synthesis of the ligand such as ST6GalI-deficient cells that lack the specific ligand of CD22, cells treated with neuraminidase, cells expressing the mutant Siglecs that lack ligand binding, or chemical compounds that inhibit ligand binding of Siglecs are useful.

2

Materials

2.1 Proximity Labeling Using Tyramide

1. Biotin-tyramide (Thermo): One vial of biotin-tyramide is dissolved in 150 μL DMSO and stored at 20  C. 2. Fluorescein-tyramide (see Note 1). 3. Hydrogen peroxide (H2O2, 30%), 4. Anti-low-affinity Fc receptor for IgG (FcγRII/III) monoclonal antibody (mAb) (2.4G2, ATCC HB-197), 5. Horse radish peroxidase (HRP)-conjugated anti-CD22 mAb. 6. Phosphate-buffered saline (PBS). 7. FACS buffer (PBS containing 2% FCS and 0.05% sodium azide). 8. Spleen B cells from wild-type or ST6GalI Note 2).

/

mice [13] (see

9. Mouse B cell line BAL17 [14]. 10. GSC718 [12, 15] dissolved in PBS. 2.2 Flowcytometry Analysis

1. FACS buffer (PBS containing 4% FCS and 0.05% sodium azide). 2. Flow cytometer (FACS Verse; BD Biosciences).

2.3 Immunoprecipitation, SDS-PAGE, and Western Blotting

1. Lysis buffer (20 mM Tris–HCl pH 8.0, 1% TritonX-100, 10% glycerol, 150 mM NaCl, 2 mM EDTA, 0.02% NaN3, 10 μg/ mL PMSF, and 1 mM Na3VO4). 2. Laemmli 2 SDS-PAGE sample buffer: 100 mM Tris–HCl (pH 6.8), 4% (w/v) SDS, 2% glycerol, 25 mM EDTA, 0.2% bromophenol blue, 200 mM β-mercaptoethanol. 3. Blocking buffer: TBS-T containing 5% bovine serum albumin (BSA). 4. TBS-T: 20 mM Tris–HCl (pH 7.4), 150 mM NaCl, 0.1% Tween-20. 5. Can Get (Toyobo).

Signal®

Immunoreaction

Enhancer

Solution

78

Amin Alborzian Deh Sheikh et al.

6. Monomeric streptavidin beads (SoftLink) (see Note 3). 7. HRP-conjugated streptavidin (Amersham). 8. HRP-conjugated anti-CD45 mAb (BD Pharmingen, clone RA3-6B2). 9. Goat anti-mouse CD22 antibody (Ab) (Santa Cruz). 10. Goat anti-mouse-IgM Ab (Southern Biotech). 11. HRP-conjugated anti-goat IgG Ab (Santa Cruz). 12. Enhanced chemiluminescence kit (Chemi-Lumi One L; Nacalai Tesque). 13. Digital imaging system (ImageQuant™ LAS4000 mini; GE Healthcare).

3

Methods

3.1 Proximity Labeling

1. Incubate 107 spleen B cells or 106 BAL17 cells (per sample) in 200 μL FACS buffer with anti-FcγRII/III mAb on ice for 20 min to block Fcγ receptor-mediated nonspecific binding of the HRP-conjugated antibodies to the cells. 2. Wash cells with cold PBS twice (see Note 4). 3. Incubate cells in 200 μL FACS buffer with 10 μg/mL HRP-conjugated anti-CD22 mAb on ice for 30 min. 4. Wash cells with cold PBS three times. 5. Resuspend cell pellets in 200 μL (per sample) prewarmed (37  C) PBS containing 1 mM H2O2 and biotin-tyramide or fluorescein-tyramide (1:50 dilution) with or without 100 μM GSC718, and incubate at 37  C for 10 min. 6. Wash cells with cold PBS three times (see Note 5).

3.2 Flow Cytometry Analysis of Labeling Efficacy of Cell Surface Molecules

1. Incubate biotin-labeled cells in 100 μL FACS buffer with fluorescein-conjugated streptavidin on ice for 30 min (Skip steps 1 and 2 for fluorescein-labeled cells). 2. Wash cells with cold FACS buffer twice. 3. Suspend cells in 500 μL cold FACS buffer. 4. Analyze cells using a flow cytometer (Fig. 2).

3.3 SDS-PAGE and Western Blotting for Biotinylated Proteins

1. Lyse biotin-labeled cells in 30 μL lysis buffer on ice for 20 min. 2. Add 30 μL 2 Laemmli SDS-PAGE sample buffer to the cell lysates and boil the samples for 5 min. 3. Sonicate samples for 5 min. 4. Spin down samples for 10 s and collect supernatants. 5. Separate proteins by SDS-PAGE using 6–12% gradient gels, and blot the proteins to PVDF membranes.

Identification of Siglec Cis-Ligands

79

Proximity labeling:

A _

4

x10 25

+

20 SSC

15 10 5 0 0

69.53% 5

10

15

20

60.45%

25 0 x104

5

10

15

20

25 x104

FSC Proximity labeling: _

B

+

Cell Count

600 500 400 300 200 100 0 0 102

103

104

105

0 102

103 104

105

Fluorescein

Fig. 2 Flow cytometry analysis of B cells labeled by proximity labeling using tyramide. Bal17 cells were treated with HRP-conjugated anti-CD22 antibody and then labeled using fluorescein-tyramide. (a) Forward (FSC) vs. side scatter (SSC) plots. Alive cells are gated in unlabeled (left) and proximity-labeled cells (right), and percentages of alive cells are indicated. (b) Histograms for the fluorescein levels of unlabeled (left) and proximity-labeled cells (right)

6. Incubate the membrane with blocking buffer with gentle shaking at RT for 1 h. 7. Incubate the membrane in Can Get Signal® Immunoreaction Enhancer Solution (see Note 6) containing HRP-conjugated streptavidin with gentle shaking at 4  C overnight. 8. Wash the membrane with TBS-T with gentle shaking at RT for 10 min three times. 9. Visualize proteins using an enhanced chemiluminescence kit and detect signals using a digital imaging system (Fig. 3).

80

Amin Alborzian Deh Sheikh et al.

A

WT B cells

GSC718

B

(-) (-) (+)

Proximity (-) (+) (+) labeling (kDa) 197.0

112.0 84.0

60.0 47.0 37.0 28.0 Blot: Streptavidin

Proximity (-) (+) (+) labeling (kDa) 197.0

112.0

84.0

60.0 47.0 37.0 28.0 Blot: Streptavidin

Fig. 3 Western blotting for proximity-labeled CD22 cis-ligands in mouse spleen B cells. Spleen B cells isolated from wild-type C57BL/6 mice (WT) (a and b) or ST6GalI / mice (b) deficient in the CD22 ligand α2,6 sialic acid were labeled with or without HRP-conjugated anti-CD22 and biotin-tyramide. In some samples, ligand binding of CD22 was inhibited by the synthetic sialoside GSC718 (a). Total cell lysates were analyzed by Western blotting using HRP-conjugated streptavidin to detect biotinylated proteins. Endogenously biotinylated proteins are indicated by arrow heads. (Reprinted from Biochem. Biophys. Res. Commun. vol. 495, Alborzian Deh Sheikh, A. et al., Proximity labeling of cis-ligands of CD22/Siglec-2 reveals stepwise α2,6 sialic aciddependent and -independent interactions, 854–859, Copyright 2018, with permission from Elsevier) 3.4 Immunoprecipitation of Biotinylated Proteins, SDS-PAGE, and Western blotting to Identify Cis-Ligands

1. Lyse cells in 100 μL lysis buffer on ice for 20 min. 2. Equilibrate monomeric streptavidin beads by washing the beads with lysis buffer three times, and incubate cell lysates with 5 μL monomeric streptavidin beads at 4  C for 30 min to 1 h (see Note 7). 3. Wash the beads with cold PBS three times. 4. Suspend the beads in 20 μL 2 Laemmli SDS-PAGE sample buffer, boil for 10 min and sonicate for 5 min (see Note 8). 5. Spin down samples for 10 s and collect supernatants. 6. Separate proteins by SDS-PAGE using 6–12% gradient gels, and blot the proteins to PVDF membranes. 7. Incubate the membrane with blocking buffer at RT for 1 h with gentle shaking.

Identification of Siglec Cis-Ligands

B

A

Proximity labeling (-)

(-) (+)

C

Proximity labeling (-)

Proximity labeling

(-) (+)

(kDa) 197.0

(kDa) 197.0

81

(-)

(-)

(+)

(kDa) 197.0

CD45

CD22 112.0

112.0

112.0

84.0

84.0

84.0 IgM

60.0

60.0

60.0

47.0

47.0

47.0

Blot: anti-CD22

Blot: anti-IgM

Blot: anti-CD45

Fig. 4 Identification of CD22, CD45, and IgM as CD22 cis-ligands. Spleen B cells from wild-type C57BL/6 mice were proximity-labeled with or without HRP-conjugated anti-CD22 and biotin-tyramide. Biotinylated proteins were immunoprecipitated using monomeric streptavidin beads. Total cell lysates and immunoprecipitates were analyzed by Western blotting for CD22 (a), Ig μ heavy chain (IgM) (b), and CD45 (c). (Reprinted from Biochem. Biophys. Res. Commun. vol. 495, Alborzian Deh Sheikh, A. et al., Proximity labeling of cis-ligands of CD22/Siglec-2 reveals stepwise α2,6 sialic acid-dependent and -independent interactions, 854–859, Copyright 2018, with permission from Elsevier)

8. Incubate the membrane in Can Get Signal® Immunoreaction Enhancer Solution (see Note 8) containing HRP-conjugated anti-CD45 mAb, goat anti-mouse CD22 Ab or goat antimouse-IgM Ab with gentle shaking at 4  C overnight. 9. Wash the membrane with TBS-T with gentle shaking at RT for 10 min three times. 10. The membranes incubated with goat anti-mouse CD22 Ab or goat anti-mouse-IgM Ab at step 8 are incubated with HRP-conjugated anti-goat IgG Ab in Can Get Signal® Immunoreaction Enhancer Solution (see Note 6) with gentle shaking at 4  C overnight, followed by washing with TBS-T three times. 11. Visualize proteins using an enhanced chemiluminescence kit and detect signals using a digital imaging system (Fig. 4).

82

4

Amin Alborzian Deh Sheikh et al.

Notes 1. Fluorescein-tyramide can be synthesized using NHS-fluorescein (Thermo Scientific) and tyramine (Sigma) as described previously [16]. 2. Spleen B cells are prepared by depleting T cells using anti-CD4 and anti-CD8 antibodies together with rabbit complement from mouse splenocytes, followed by percoll gradient centrifugation as described previously [17]. Around 4  107 B cells are obtained from one mouse spleen. Alternatively, mouse spleen B cells are obtained using kits. 3. Regular (tetrameric) streptavidin beads should not be used because of the difficulty in recovering the biotinylated proteins from the beads. 4. All procedures are done in 1.5 mL microfuge tubes. Cells are therefore washed by adding 1 mL of cold PBS or FACS buffer, followed by centrifugation at 4  C. To minimize cell loss during washing, centrifuge at 400  g with the lowest deceleration (or soft brake) using a swing rotor. Samples for flow cytometry can be prepared in FACS tubes. 5. In some protocols, ascorbic acid is used to stop the labeling, but we found that washing the cells by cold PBS is sufficient to stop the labeling reaction. 6. TBS-T containing 1% BSA can be used instead of Can Get Signal® Immunoreaction Enhancer Solution. 7. Use anti-fluorescein Ab-conjugated-beads for fluoresceinlabeled samples. 8. Dissociation of streptavidin–biotin is improved by boiling longer (10 min) together with sonication after boiling for 5 min.

Acknowledgments This work was funded by JPSP Grant-in-Aid for Scientific Research 26293062 (T.T.) and the Joint Usage/Research Program of Medical Research Institute, Tokyo Medical and Dental University (T.T.). References 1. Macauley MS, Crocker PR, Paulson JC (2014) Siglec-mediated regulation of immune cell function in disease. Nat Rev Immunol 14:653–666 2. Bobrow MN, Harris TD, Shaughnessy KJ et al (1989) Catalyzed reporter deposition, a novel method of signal amplification. Application to immunoassays. J Immunol Methods 125:279–285

3. Adams JC (1992) Biotin amplification of biotin and horseradish peroxidase signals in histochemical stains. J Histochem Cytochem 40:1457–1463 4. Kerstens HM, Poddighe PJ, Hanselaar AG (1995) A novel in situ hybridization signal amplification method based on the deposition of biotinylated tyramine. J Histochem Cytochem 43:347–352

Identification of Siglec Cis-Ligands 5. Mayer G, Bendayan M (1997) Biotinyltyramide: a novel approach for electron microscopic immunocytochemistry. J Histochem Cytochem 45:1449–1454 6. Raap AK, van de Corput MP, Vervenne RA et al (1995) Ultra-sensitive FISH using peroxidasemediated deposition of biotin- or fluorochrome tyramides. Hum Mol Genet 4:529–534 7. Schofer C, Weipoltshammer K, Almeder M et al (1997) Signal amplification at the ultrastructural level using biotinylated tyramides and immunogold detection. Histochem Cell Biol 108:313–319 8. Bendayan M (2001) Tech.Sight. Worth its weight in gold. Science 291:1363–1365 9. Li XW, Rees JS, Xue P et al (2014) New insights into the DT40 B cell receptor cluster using a proteomic proximity labeling assay. J Biol Chem 289:14434–14447 10. Rees JS, Li XW, Perrett S et al (2015) Protein neighbors and proximity proteomics. Mol Cell Proteomics 14:2848–2856 11. Chang L, Chen YJ, Fan CY et al (2017) Identification of Siglec ligands using a proximity labeling method. J Proteome Res 16:3929–3941 12. Alborzian Deh Sheikh A, Akatsu C, Imamura A et al (2018) Proximity labeling of cis-ligands of

83

CD22/Siglec-2 reveals stepwise alpha2,6 sialic acid-dependent and -independent interactions. Biochem Biophys Res Commun 495:854–859 13. Hennet T, Chui D, Paulson JC et al (1998) Immune regulation by the ST6Gal sialyltransferase. Proc Natl Acad Sci U S A 95:4504–4509 14. Stoddart A, Ray RJ, Paige CJ (1997) Analysis of murine CD22 during B cell development: CD22 is expressed on B cell progenitors prior to IgM. Int Immunol 9:1571–1579 15. Abdu-Allah HH, Watanabe K, Completo GC et al (2011) CD22-antagonists with nanomolar potency: the synergistic effect of hydrophobic groups at C-2 and C-9 of sialic acid scaffold. Bioorg Med Chem 19:1966–1971 16. Jiang S, Kotani N, Ohnishi T et al (2012) A proteomics approach to the cell-surface interactome using the enzyme-mediated activation of radical sources reaction. Proteomics 12:54–62 17. Nomura T, Han H, Howard MC et al (1996) Antigen receptor-mediated B cell death is blocked by signaling via CD72 or treatment with dextran sulfate and is defective in autoimmunity-prone mice. Int Immunol 8:867–875

Chapter 9 Preparation of Recombinant Siglecs and Identification of Their Ligands Lan-Yi Chang, Penk Yeir Low, Deepa Sridharan, Kaia Gerlovin, and Takashi Angata Abstract Siglecs are transmembrane receptor-like vertebrate lectins that recognize glycans containing sialic acid. Most Siglecs also interact with intracellular signal transduction molecules, and modulate immune responses. Recombinant soluble Siglecs fused with the fragment crystallizable (Fc) region of immunoglobulin G (Siglec-Fc) are a versatile tool for the investigation of Siglec functions. We describe protocols for the production of recombinant Siglec-Fc, the analysis of expression of Siglec ligands by flow cytometry, and the identification of the Siglec ligand candidates based on proximity labeling. Key words Siglec, Sialic acids, Ligands, Flow cytometry, Proximity labeling

1

Introduction Siglecs are a family of vertebrate lectins that belong to the immunoglobulin superfamily and recognize oligosaccharides containing sialic acid [1–3]. Lectins that belong to the immunoglobulin superfamily are collectively called I-type lectins (“I” stands for immunoglobulin), and Siglecs account for a majority of I-type lectins known to date [4, 5]. Apparently, several immunoglobulin superfamily proteins have independently acquired glycan recognition capability through convergent evolution, judging from the presence of other I-type lectins, such as Platelet and Endothelial Cell Adhesion Molecule 1 (PECAM1) [6, 7], that do not appear to share an immediate molecular ancestor with Siglecs. Most Siglecs are expressed on leukocytes (although the expression of Siglecs may not be restricted to leukocytes [8]), and are involved in various interactions between leukocytes and endogenous or exogenous ligands. Most Siglecs also interact with intracellular or transmembrane signal transduction molecules, such as tyrosine phosphatase SHP-1 (encoded by PTPN6) or tyrosine

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_9, © Springer Science+Business Media, LLC, part of Springer Nature 2020

85

86

Lan-Yi Chang et al.

kinase Syk (encoded by SYK) via adapter protein DAP12 (encoded by TYROBP), and modulate leukocyte responses. Siglecs may be classified into “inhibitory” or “activating” subsets, based on the signal transduction molecules they interact with (SHP-1 and DAP12, respectively). However, the functional attribution based on the interacting signaling molecule may be somewhat simplistic: some Siglecs that associate with SHP-1 (e.g., Siglec-8 [9] and Siglec-9 [10]) may transduce cell-activating signal, whereas those that associate with DAP12 (e.g., mouse Siglec-H [11]) may transduce suppressive signal. Siglecs are highly relevant to human health and diseases [8, 12], as they are involved in the development and homeostatic regulation of leukocytes [13–18], in leukocyte– microbe interactions [19, 20] and in leukocyte–cancer interactions [21, 22]. A few therapeutics targeting Siglecs are approved for clinical use, and many more are being developed [23]. As such, characterization of specific Siglec interactions and the identification of their ligands is necessary for the understanding of numerous diseases in which Siglecs are involved, and ultimately the development of therapeutics targeting such interactions. To investigate the functions of Siglecs, in particular to analyze the glycan recognition specificities and to probe for the presence of Siglec ligands on cells/tissues, various formats of recombinant soluble Siglecs, such as those fused with the Fc region of immunoglobulin G [24–26], biotinylation tag [27], and pentamerizing domain of cartilage oligomeric matrix protein [28], have been used. The most widely used format is the fusion proteins consisting of the extracellular domain of Siglec and the Fc part of immunoglobulin G (i.e., Siglec-Fc). In this chapter, we describe the protocol we are currently using for the production of recombinant Siglec-Fc by transient transfection of Expi293F cell line. A sample protocol for the staining of cells with Siglec-Fc for flow cytometry analysis is also provided. Identification of glycoconjugates that serve as ligands for Siglecs is important for mechanistic understanding of Siglec functions, but it is challenging because the interactions between Siglecs and ligands are inherently weak. In addition, Siglec ligands in the context of cell–cell interaction are integral membrane glycoproteins or glycolipids. As the classical affinity purification method usually requires the solubilization of membrane-associated ligands with detergent, which inevitably disrupts their higher-order assembly and reduces the ligand avidity, isolation of ligand by affinity capture may not be an ideal approach. Introduction of a chemical “handle” (a small reactive group, often photoactivatable) into sialic acids on glycoconjugates in situ and covalent conjugation with Siglec probe, followed by purification and protein identification [29], is one way to overcome this problem. However, this method requires prior knowledge regarding the compatibility of the chemical handle introduced into sialic acids and the target Siglec. Thus, a simple and versatile method that enables the identification of Siglec ligands

Ligand Identification with Recombinant Siglec

87

is still needed. To address this need, we developed a method to introduce biotin label into the Siglec ligand candidates by proximity labeling [30]. Proximity labeling is a generic term for the methods that use an enzyme-coupled “bait” protein to generate a short-lived reactive intermediate (typically a biotin derivative) to introduce a chemical label into the proteins in the vicinity of the bait [31]. Proteins labeled and purified may be used for identification by mass spectrometry or by antibody. Other laboratories have independently developed similar proximity labeling methods based on the same principle [32, 33]. These methods in principle can be applied to other lectins as well.

2

Materials

2.1 Preparation of Recombinant Siglec-Fc 2.1.1 Equipment

1. CO2 incubator, set at 37  C, 8% CO2. 2. Platform shaker (placed in the CO2 incubator). 3. Biosafety cabinet. 4. Centrifuge. 5. Rotator.

2.1.2 Reagents and Consumables

1. Mammalian expression vectors for recombinant Siglec-Fc. 2. Expi293F cell line (ThermoFisher Scientific A14527). 3. Expi293 Expression A1435101).

Medium

(ThermoFisher

Scientific

4. Erlenmeyer flask, 125 mL (Corning 431405). 5. Spin-X 0.22 μm microfilter (Corning 8160). 6. ExpiFectamine 293 Transfection Kit (ThermoFisher Scientific A14524). 7. Opti-MEM (ThermoFisher Scientific 31985070). 8. Dulbecco’s phosphate-buffered saline (D-PBS; e.g., Sigma D5652). 9. 20 mM HEPES-NaOH buffer, pH 7.0. 10. Arthrobacter ureafaciens sialidase (Nacalai 24229-74). 11. 0.1 M citrate-NaOH buffer, pH 3.0. 12. 1 M Na2CO3. 13. rProtein A Sepharose Fast Flow (GE Healthcare 17-1279-02). 14. Disposable plastic column (e.g., Poly-Prep Chromatography Columns, Bio-Rad 7311550). 15. Centrifugal ultrafiltration device (e.g., 10 or 30 kDa cutoff, Amicon Ultra, Millipore). 16. 6 M guanidine HCl. 17. BCA protein assay kit (ThermoFisher Scientific 23225).

88

Lan-Yi Chang et al.

2.2 Flow Cytometry Using Siglec-Fc as a Probe

1. Centrifuge. 2. Flow cytometer.

2.2.1 Equipment 2.2.2 Reagents and Consumables

1. Cell line of interest. 2. Recombinant Siglec-Fc (prepared as described in the protocol). 3. Fluorescein-conjugated goat anti-human IgG, Fcγ fragmentspecific (Jackson ImmunoResearch 109-095-098). 4. Staining buffer: 1% bovine serum albumin (BSA), 0.02% NaN3 in D-PBS. 5. Cell strainer, 40 μm Nylon (Falcon 352340).

2.3 Proximity Labeling Using Siglec-Fc as a Probe

1. Centrifuge.

2.3.1 Equipment

4. Magnetic stand (e.g., DynaMag-2 Magnet, ThermoFisher Scientific 12321D).

2. Water bath incubator. 3. Shaker incubator (e.g., Taitec M-BR-022UP).

5. Power supply for SDS-PAGE. 6. Electrophoresis equipment for SDS-PAGE. 7. Semi-dry transfer equipment (e.g., Trans-Blot SD Semi-Dry Transfer Cell, Bio-Rad 170-3940). 8. Platform shaker. 2.3.2 Reagents and Consumables

1. Cell line of interest. 2. Recombinant Siglec-Fc (prepared as described in the protocol). 3. ANTI-FLAG M2-Peroxidase (HRP) antibody (Sigma B8592). 4. Biotin-tyramide (prepared in house or purchased from commercial source, e.g., Sigma-Aldrich, PerkinElmer, ThermoFisher Scientific, etc.). 5. D-PBS. 6. 1% BSA in D-PBS (no azide). 7. Tris-buffered saline (TBS): 20 mM Tris–HCl, pH 7.5, 140 mM NaCl. 8. Labeling solution: TBS + 10 mM H2O2 + 10 μM biotintyramide (prepare immediately before use). 9. Quenching solution: TBS + 100 μM L-ascorbic acid. 10. RIPA buffer: 50 mM Tris–HCl, pH 8.0, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS with protease inhibitor cocktail (e.g., cOmplete Ultra Tablets, Roche 05892791001).

Ligand Identification with Recombinant Siglec

89

11. Dynabeads MyOne Streptavidin C1 (ThermoFisher Scientific, 65001). 12. PBSS: D-PBS + 0.1% SDS. 13. 2 SDS-PAGE Sample Buffer (Bio-Rad 161-0737; add DTT powder to 200 mM). 14. Polyacrylamide gel for SDS-PAGE. 15. SDS-PAGE running buffer: 25 mM Tris, 192 mM glycine, 0.1% SDS. 16. Polyvinylidene difluoride (PVDF) membrane (e.g., Polyscreen PVDF transfer membrane, PerkinElmer NEF1002001PK). 17. Towbin’s buffer: 25 mM Tris, 192 mM glycine, 10% methanol. 18. TBST: TBS + 0.1% Tween 20. 19. Blocking reagent: 3% BSA in TBST. 20. Streptavidin-HRP (Jackson ImmunoResearch 016-030-084). 21. Chemiluminescence reagent (e.g., Western Lightning ECL Pro, PerkinElmer NEL120001EA).

3

Methods

3.1 Preparation of Recombinant Siglec-Fc

We routinely obtain recombinant Siglec-Fc by transient transfection of Expi293F cell line. The construct we use contains AviTag (for enzymatic biotinylation; [34]) and FLAG tag (for detection) between extracellular domain of Siglec (typically three immunoglobulin-like domains) and hinge-Fc region of human IgG1. (Fig. 1; see Note 1).

3.1.1 Maintenance of Expi293F Cells

Maintain Expi293F cells (0.3–5  106 cells/mL) in Expi293 Expression Medium using an Erlenmeyer flask, on platform shaker (125 rpm) in 8% CO2 incubator, in accordance with the instructions provided by the manufacturer.

3.1.2 Transfection

1. Day 1: Seed 6  107 Expi293F cells in 30 mL of Expi293 Expression Medium and culture overnight. 2. Day 2: Adjust the cell density at 7.5  107 cells in 25.5 mL of Expi293 Expression Medium (see Note 2). Using Spin-X, filter-sterilize 30 μg of plasmid suspended in 500 μL of OptiMEM (12,000  g, 5 min). Meanwhile, in a sterile 15 mL conical tube, mix 2.5 mL of Opti-MEM and 81 μL of ExpiFectamine 293, and leave for 5 min. Add the filter-sterilized plasmid into the tube and mix well. Leave it at room temperature for 20 min. Avoid exposure to UV light. Add the DNA-ExpiFectamine complex to Expi293F cells in Erlenmeyer flask. Place the flask back into the incubator and incubate overnight.

90

Lan-Yi Chang et al.

Fig. 1 Overall design of Siglec-Fc. The extracellular domain of Siglec is fused with the hinge–Fc region of human IgG1 to obtain Siglec-Fc. Our expression constructs contain two tags (i.e., AviTag for enzymatic biotinylation and FLAG tag for detection/functionalization with anti-FLAG antibody) between the Siglec and the Fc region. C-terminus of FLAG tag (DYKDDDDK) is a recognition site for enterokinase, which can be used to remove hinge-Fc region if necessary. TEV protease cleavage site is introduced in a new expression vector, along with some mutations in hinge-Fc region to improve functionality (unpublished)

3. Day 3: Twenty hours after the transfection, add 150 μL of Transfection Enhancer 1 and 1.5 mL of Transfection Enhancer 2 (included in the Transfection Kit). Place the flask back into the incubator and culture for 3 days in CO2 incubator. 4. Day 6: Transfer the content of the flask into 50 mL conical tube and centrifuge at 300  g for 3 min. Transfer the supernatant to another 50 mL tube (to be further centrifuged later). Resuspend the cell pellet in 30 mL of fresh Expi293 Expression Medium, place back in the Erlenmeyer flask, and culture for 3 more days. The supernatant should be cleared further by centrifugation at 2000  g for 10 min. Collect supernatant and store at 4  C (or frozen). 5. Day 9: Transfer the content of the flask into 50 mL conical tube and centrifuge at 300  g for 3 min. Transfer the supernatant to another 50 mL tube and centrifuge again at 2000  g for 10 min to remove cell debris. Collect supernatant and store at 4  C (or frozen).

Ligand Identification with Recombinant Siglec 3.1.3 Purification

91

1. Add 0.2 mL of Protein A-Sepharose suspension (as 50% slurry; packed gel volume is ~0.1 mL) per 30 mL of culture sup of the transfected cells. Incubate 2–3 overnights at 4  C on a rotator (1–3 rpm) in cold room. 2. Pour the culture sup + Protein A-Sepharose into a disposable column. Wait until all culture sup pass through the column (see Note 3). 3. Add 20 gel volume (4 mL/0.2 mL Protein A-Sepharose) of D-PBS to the column and drain (¼“wash the gel”). Wash the gel two more times with 20 gel volume of D-PBS (total three washes). 4. Wash the gel once with 10 gel volume (2 mL/0.2 mL Protein A-Sepharose) of 20 mM HEPES-NaOH buffer. Cap the bottom of the column (make sure that the column will not leak), and add 5 gel volume of 100 mU/mL A. ufeafacience sialidase in 20 mM HEPES-NaOH buffer. Incubate at 37  C for 30 min. Remove the cap from the bottom of the column and drain. Wash the gel three times with 20 gel volume (4 mL/ 0.2 mL Protein A-Sepharose) of D-PBS to remove sialidase (see Note 4). 5. Elute Siglec-Fc fusion protein from the column three times with 10 gel volume (2 mL/0.2 mL Protein A-Sepharose) of 0.1 M citrate-NaOH buffer, pH 3.0, and collect the eluate. Add 1/10 volume (0.6 mL/6 mL eluate) of 1 M Na2CO3 to the eluate to neutralize. 6. Transfer the neutralized eluate to a centrifugal ultrafiltration device and centrifuge at 2000  g for 30 min. Check after 30 min to see if the solution is concentrated to 3wmv, http://www.ebi.ac.uk/pdbe/entry/pdb/ 3WMV 10. Terada D, Kawai F, Noguchi H et al (2016, 6) Crystal structure of MytiLec, a galactosebinding lectin from the mussel Mytilus galloprovincialis with cytotoxicity against certain cancer cell types. Sci Rep:28344 11. Fujimoto Z (2013) Structure and function of carbohydrate-binding module families 13 and 42 of glycoside hydrolases, comprising a β-trefoil fold. Biosci Biotechnol Biochem 77:1363–1371 12. Hasan I, Sugawara S, Fujii Y et al (2015) MytiLec, a mussel R-type lectin, interacts with surface glycan Gb3 on Burkitt’s lymphoma cells to trigger apoptosis through multiple pathways. Mar Drugs 13:7377–7389 13. Matsushima-Hibiya Y, Watanabe M, Hidari KI et al (2003) Identification of glycosphingolipid receptors for pierisin-1, a guanine-specific

Cell Regulation by a β-Trefoil Lectin ADP-ribosylating toxin from the cabbage butterfly. J Biol Chem 278:9972–9978 14. Chernikov OV, Wong WT, Li LH et al (2017) A GalNAc/gal-specific lectin from the sea mussel Crenomytilus grayanus modulates immune response in macrophages and in mice. Sci Rep 7:6315 15. Pohleven J, Obermajer N, Sabotic J et al (2009) Purification, characterization and cloning of a ricin B-like lectin from mushroom Clitocybe nebularis with antiproliferative activity against human leukemic T cells. Biochim Biophys Acta 1790:173–181 16. Bleuler-Martinez S, Stutz K, Sieber R et al (2017) Dimerization of the fungal defense lectin CCL2 is essential for its toxicity against nematodes. Glycobiology 27:486–500 17. Bovi M, Cenci L, Perduca M et al (2013) BEL β-trefoil: a novel lectin with antineoplastic properties in king bolete (Boletus edulis) mushrooms. Glycobiology 23:578–592 18. Liao JH, Chien CT, Wu HY et al (2016) A multivalent marine lectin from Crenomytilus grayanus possesses anti-cancer activity through recognizing globotriose Gb3. J Am Chem Soc 138:4787–4795 19. Singh RS, Kaur HP, Singh J (2014) Purification and characterization of a mucin specific mycelial lectin from Aspergillus gorakhpurensis:

213

application for mitogenic and antimicrobial activity. PLoS One 9:e109265 20. Lavanya V, Ahmed N, Khan MK et al (2016) Sustained mitogenic effect on K562 human chronic myelogenous leukemia cells by dietary lectin, jacalin. Glycoconj J 33:877–886 21. Smith PK, Krohn RI, Hermanson GT et al (1985) Measurement of protein using bicinchoninic acid. Anal Biochem 150:76–85 22. Wiechelman KJ, Braun RD, Fitzpatrick JD (1988) Investigation of the bicinchoninic acid protein assay: identification of the groups responsible for color formation. Anal Biochem 175:231–237 23. Gerdol M, Fujii Y, Hasan I et al (2017) The purplish bifurcate mussel Mytilisepta virgata gene expression atlas reveals a remarkable tissue functional specialization. BMC Genomics 18:590 24. Mangeney M, Lingwood CA, Taga S (1993) Apoptosis induced in Burkitt’s lymphoma cells via Gb3/CD77, a glycolipid antigen. Cancer Res 53:5314–5319 25. Okayama A, Mikasa K, Matsui N et al (2004) An interventional approach to block brain damage caused by Shiga toxin-producing Escherichia coli infection, by use of a combination of phosphodiesterase inhibitors. J Infect Dis 190:2129–2136

Chapter 22 Lectin-Type Ubiquitin Ligase Subunits: Fbs Proteins and Their Applications for Use Yukiko Yoshida Abstract Three lectin-type F-box proteins called Fbs (F-box protein-recognizing sugar chains) are found in mammals, and function as substrate-binding subunits in the SCF (Skp1/Cullin1/F-box protein) complex ubiquitin ligases. The SCFFbs recognizes cytosolic N-glycans as a signal for an adverse cellular state, and ubiquitinates glycoproteins which appear in the cytosol to remove them from cells. Although Fbs proteins recognize innermost Man3GlcNAc2 structure that is commonly found in most N-glycan structures, they preferentially bind high-mannose-type glycans. Recently, the recombinant Fbs1 derivative protein has been developed as a tool for comprehensive enrichment of N-glycopeptides. The labeled Fbs3 is also available as a tool for detecting organelle damage in cells as it has characteristic properties which cause it to quickly accumulate in damaged organelles. In this chapter, we introduce two applications of use for Fbs proteins: the unbiased N-glycopeptide capture method and the detection of damaged organelles in living cells. Key words F-box protein, lectin, N-glycan, N-glycopeptide, organelle, ubiquitin ligase

1

Introduction The SCF complex is one of well-studied families of complex-type E3 ubiquitin ligases, and is found in all eukaryotes. It consists of three invariable components, that is, a scaffold protein Cullin1, a RING protein Rbx1, and an adaptor protein Skp1. As well, there is a variable component of an F-box protein which binds to specific subunits [1]. Most F-box proteins recognize either structural changes or various post-translational modifications in order to ubiquitinate substrates for proteasomal degradation in cells in an accurate and timely manner [2]. Among more than 70 human F-box proteins, three F-box proteins called Fbs1–3 (F-box protein-recognizing sugar chains) recognize N-glycans of glycoproteins as a tag for ubiquitination, although Fbs1–3 are present in the cytosol where glycoproteins are not normally found [3]. Fbs1 is the first identified lectin-type F-box protein that was purified. It was derived from mouse brain extracts by using

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_22, © Springer Science+Business Media, LLC, part of Springer Nature 2020

215

216

Yukiko Yoshida

GlcNAc-terminated fetuin (GTF)-immobilized beads. N, N0 diacetylchitobiose (GlcNAc2) selectively dissociated the interaction between Fbs1 and GTF, and is actually able to bind to the innermost Man3GlcNAc2 in N-glycans through a small hydrophobic pocket in the loops located at the top of the β-sandwich [4, 5]. Although Man3GlcNAc2 is the common core motif of Nglycans, both Fbs1 and Fbs2 preferentially bind high-mannosetype glycans [4, 6, 7]. The innermost moiety of N-glycans is generally embedded in amino acid residues in native glycoproteins, and Fbs1 and Fbs2 prefer to interact with denatured glycoproteins [8]. Accordingly, SCFFbs1 and SCFFbs2 E3 ubiquitin ligases act on misfolded glycoproteins in the endoplasmic reticulum-associated degradation (ERAD) pathway [3, 4, 6]. Fbs3 also binds to glycoproteins modified with high-mannose glycans but can interact with glycoproteins found in lysosomes and endosomes, which are modified with complex-type glycan. Unlike Fbs1 and Fbs2, Fbs3 localizes to membranes of organelles via its Nmyristoylation in the steady state. Fbs3 is effectively recruited to damaged organelles through its membrane-targeting and glycoprotein-binding characteristics. Consequently, SCFFbs3 ubiquitinates exposed glycoproteins in lysosomes following lysosomal damage, which results in accelerated recruitment of autophagic machinery [9]. Fbs1 is an abundant protein in the brain [10]. The recombinant Fbs1 is also quite stable and its structural information is known [5, 11]. By using the inherent properties of Fbs1 which cause it to bind to the common core Man3GlcNAc2 of N-glycans, a recombinant mutant Fbs1 has recently been engineered which has been modified to bind to diverse types of N-glycans in order to enrich for N-glycopeptides [12]. This unbiased and efficient N-glycopeptide enrichment method enables the simultaneous determination of Nglycan composition and glycosylation sites. In this chapter, I introduce two applications using the Fbs proteins as lectins other than ubiquitin ligases. One is the highly efficient N-glycopeptide enrichment method which uses an Fbs1 recombinant protein, and the other is a method for detecting damaged organelles in the secretory pathway using Fbs3.

2

Materials

2.1 N-Glycopeptide Enrichment Method

1. Synthetic DNA for human Fbs1 substrate-binding domain (residues 92–296 with mutations of S155G, F173Y, and E174R), variant (hereafter, Fbs1 GYR) with linker, and BamHI/XhoI restriction enzyme sites (see Fig. 1 and Note 1). 2. Restriction enzymes (BamHI and XhoI) and T4 DNA ligase. 3. Plasmid pGEX-6P-2, 1 ml GSTrapFF, and PreScission protease (GE Healthcare).

F-box Protein-Recognizing Sugar Chains: Fbs

217

Fig. 1 Synthetic DNA sequence of Fbs1 GYR in this study (ref. [11]). DNA sequence encoding human Fbs1 substrate-binding domain (residues 92–296, shown below DNA sequence) with mutations of S155G, F173Y, and E174R (shown in red text), linker, and BamHI/XhoI restriction enzyme sites (underlined)

4. Escherichia coli DH5α and BL21(DE3) strains. 5. LB broth and LB Agar. 6. PreScission buffer: 50 mM Tris–HCl pH 8.0, 100 mM NaCl, 1 mM EDTA, and 1 mM dithiothreitol (DTT). 7. Lysis buffer: 100 mM Tris–HCl pH 7.5, 4% SDS, and 0.1 M DTT. 8. Urea buffer: 8 M urea in 100 mM Tris–HCl pH 8.5. 9. Microcon 10 K and 30 K devices (Millipore). 10. 1 mg ml1 Trypsin Gold (Promega) in 1 mM HCl.

218

Yukiko Yoshida

2.2 Detecting Damaged Organelles in Cells Using Fbs3

1. Restriction enzymes (SgfI, PmeI, and EcoICRI) and T4 DNA ligase (see Note 2). 2. Fbs3 clone (Flexi clone: FXC26619 or N-terminally Halo Tag clone: FHC26619) (see Note 3). 3. C-terminal Halo Tag vector: pFC14K (Promega). 4. Cell culture medium (e.g., Dulbecco Modified Eagle Medium [DMEM]) supplemented with 10% [vol vol1] fetal calf serum [FCS] and transfection reagents (polyethylenimine [Polysciences], Lipofectamine 2000 [Thermo Fisher Scientific], etc.) for appropriate cell lines. 5. Halo Tag ligand (TMR Ligand [Ex 552 nm/Em 578 nm], Oregon Green Ligand [Ex 492 nm/Em 520 nm], diAc FAM Ligand [Ex 499 nm/Em 518 nm], or Coumarin Ligand [Ex 362 nm/Em 450 nm].

3

Methods

3.1 N-Glycopeptide Enrichment Method 3.1.1 Preparation of Recombinant Fbs1 GYR Protein

1. For generating GST-tagged Fbs1 GYR expression vector, subclone the digested synthetic DNA fragment into the BamHI and XhoI restriction sites of pGEX-6P-2. Transform E. coli DH5α with resultant ligase mixture and plate on LB agar containing 200 μg ml1 ampicillin. Allow colony formation at 37  C. 2. Grow overnight culture by inoculating 3 ml of LB containing 200 μg ml1 ampicillin with single colonies, prepare the plasmids, and confirm single insertion in the correct orientation by digestion with BamHI and XhoI followed by agarose gel electrophoresis. 3. Transform E. coli BL21(DE3) with pGEX-Fbs1 GYR and plate on LB agar containing 200 μg ml1 ampicillin. Allow colony formation at 37  C. 4. Grow overnight culture by inoculating 5 ml of LB containing 200 μg ml1 ampicillin with a single colony. 5. Inoculate 1 liter of LB containing 200 μg ml1 ampicillin. Grow to OD600 of 1.0 at 30  C with 200 rpm shaking and 1 mM isopropyl-β-D-thiogalactoside (IPTG) for 3 h. 6. Pellet cells and wash once with phosphate-buffered saline (PBS). 7. Resuspend pellet in 50 ml PBS with 5 mM DTT using sonication. 8. Clear the lysate by centrifugation at 70,000  g for 20 min at 4  C and centrifuge the resultant supernatant at 300,000  g for 60 min at 4  C. 9. Apply the supernatant to GSTrap FF at flow rate of 1 ml min1 and wash the conjugated beads with more than 50 ml PBS.

F-box Protein-Recognizing Sugar Chains: Fbs

219

10. Equilibrate the beads with PreScission buffer, apply 20 units PreScission into the conjugated beads, and incubate the closed the column for 16 h at 4  C. 11. Elute the cleaved Fbs1 GYR protein with PreScission buffer. Confirm the eluted protein is almost pure by SDS-PAGE detected by CBB staining. Measure the concentration of Fbs1 GYR by a general method such as the BCA method or A280 (see Note 4). 3.1.2 Preparation of Sample Peptide from Cultured Cells

Tryptic peptides were prepared according to the filter-aided sample preparation (FASP) protocol (Fig. 2, upper) [13]. 1. Grow cells for each experimental condition. 2. Harvest cells by centrifugation at 200  g for 5 min, wash with cold PBS, and remove PBS. 3. Add lysis buffer to cell pellet, sonicate briefly, and incubate at 95  C for 5 min (see Note 5). 4. Clear the lysate by centrifugation at 20,000  g for 15 min at room temperature and transfer the supernatant into a new tube. Measure the protein concentration.

Fig. 2 Overview of the procedure for sample peptide preparation and enrichment of N-glycopeptides. Filter-aided sample preparations (FASP) for the proteomic analysis method are pictured in the upper figures. N-glycopeptides are enriched by Fbs1 GYR on the filter unit and unbound peptides are removed by centrifugation. The bound glycopeptides are eluted by formic acid (N-glyco-FASP; lower figures)

220

Yukiko Yoshida

5. Dilute 0.2–0.4 mg protein lysate with 200 μl of urea buffer and transfer it to a Microcon 30 K. 6. Centrifuge at 14,000  g at 18  C for 15 min and discard flow through solution in each centrifugation. Add 200 μl of urea buffer to the filter and centrifuge under the same conditions. Repeat this step. 7. Add 100 μl of urea buffer with 0.05 M iodoacetamide and incubate the samples for 20 min in darkness. 8. Centrifuge for 10 min under the above conditions and wash three times with 100 μl of urea buffer. 9. Add 100 μl of 40 mM NH4HCO3 to the filter and centrifuge for 10 min. Repeat this wash step twice. 10. Add 4 μg trypsin in 40 μl of 40 mM NH4HCO3 to the filter and incubate overnight at 37  C. 11. Transfer the filter to new tube, collect the digested peptide by centrifugation, and rinse the filter twice with 40 μl of 20 mM Tris–HCl pH 7.5 and 150 mM NaCl. 3.1.3 Enrichment of N-Glycopeptides by N-Glyco-FASP Method

This method is based on the N-glyco-FASP method which operates on the principle that N-glycosylated peptides are bound by lectin and retained on the filter while nonglycosylated peptides are washed through the filter (Fig. 2, lower) [12, 14]. 1. Add 100 μl of 5 mg ml1 Fbs1 GYR protein to the above 120 μl peptide solution; mix and transfer to new Microcon 10 K. 2. Incubate the mixture for 1 h at room temperature and centrifuge at 14,000  g at 18  C for 10 min (see Note 6). 3. Wash the N-glycopeptides captured by the Fbs1 GYR protein with 200 μl of 40 mM NH4HCO3 four times and elute the captured peptides with 100 μl of 50% formic acid three times. 4. Lyophilize the eluted peptide to remove formic acid and NH4HCO3 (see Note 7).

3.2 Detection of Damaged Organelles in Cells by Using Fbs3 3.2.1 Plasmid Construction and Transfection

Fbs3 protein is subject to N-myristoylation, which is necessary for membrane localization and allows it to accumulate rapidly around damaged organelles. Short C-terminal tags (HA, FLAG, and Myc) are acceptable for observing the recruitment of damaged organelles, but some fluorescence tags such as GFP suppress its recruitment due to artificial retention in mitochondria. Therefore, we use C-terminal Halo-tagged Fbs3 for live imaging. The Flexi Vector System is an easy method for transfer into a C-terminal vector, as the Flexi vector carries the lethal barnase gene, which is replaced by the DNA fragment of interest.

F-box Protein-Recognizing Sugar Chains: Fbs

221

1. Digest the Fbs3 Flexi clone or Halo-Fbs3 clone with SgfI and PmeI and check for the presence of the 0.85 kb Fbs3 cDNA fragment by agarose electrophoresis. These plasmids are ampicillin resistant but the C-terminal Halo-tag vector for subcloning is kanamycin resistant. Therefore, you do not have to purify the fragment. 2. Digest the pFC14K vector with SgfI and EcoICRI, mix with the digested DNA of Fbs3 clone prepared in 1, and incubate with T4 DNA ligase. 3. Transform E. coli DH5α with the resultant ligase mixture and plate on LB agar containing 25 μg ml1 kanamycin. Allow colony formation at 37  C. 4. Grow the culture overnight by inoculating 3 ml of LB containing 25 μg ml1 kanamycin with single colonies, prepare the plasmids, and confirm the insertion of cDNA sequencing. 5. In the case of the HeLa cells, seed 1  105 cells in 2 ml of DMEM medium on each glass-bottomed 35-mm dish in a 5% CO2 incubator at 37  C. Use a culture medium appropriate for these cell lines. 6. The next day, transfect HeLa cells with the plasmid encoding C-terminal Halo-tagged Fbs3 (Fbs3-Halo). Use 1 μg DNA and 9 μl of Lipofectamine 2000 (Thermo Fisher) reagent per dish. 3.2.2 Detecting Organelle Damage by Time-Lapse Imaging

1. Prepare a 1:200 dilution of Halo Tag Ligand in warm culture medium. 2. On the day following transfection, label cells expressing Fbs3Halo by adding 0.5 ml of the above diluted Halo Tag Ligand solution, and mix gently. 3. Incubate for 15 min at 37  C in a CO2 incubator. 4. Replace the ligand-containing medium with 2 ml of warm fresh medium twice (see Note 8). 5. Transfer to microscope, add reagent, and capture live-cell imaging by time-lapse mode (see Note 9).

4

Notes 1. For construction of the recombinant GST-Fbs1 GYR expression vector, use of synthetic DNA is easier and more efficient for the expression in E. coli because of codon optimization. However, you can perform the construction by general cloning and mutagenesis methods. The sequence information of human Fbs1 is available in NIBI reference sequence NM_012168.5.

222

Yukiko Yoshida

2. Although Flexi Enzyme Blends (SgfI & PmeI and SgfI & EcoICRI) are available from Promega, individual enzyme isoschizomers AsiSI (SgfI), MssI (PmeI), and Ecl36II (EcoICRI) are also available. 3. These clones are available from Promega. They are easy to clone into a Halo-tag vector, but Fbs3 cDNA fragment may be prepared by PCR amplification from cDNA derived from some cell lines expressing Fbs3 (PANC-1, MKN-7, A-359, etc.) with 50 - and 30 -primers containing the SgfI and PmeI sites, respectively (5’-CCTGGCGCGATCGCACCATGGGGCGCC TCGGTCTCCAG-30 and 5’-AGTGCTGTTTAAACGGAC AGACGGACTCGCACGATC-30 ) using KOD FX Neo (Toyobo). The N-terminal sequence of Fbs3 is GC-rich and with other polymerases, it may be difficult to amplify the Fbs3 cDNA. The sequence information of human Fbs3 is available in NIBI reference sequence NM_178820.4. 4. The eluted Fbs1 GYR protein is detected at the 23 kDa band in SDS-PAGE. 5. If you prepare sample peptide from some tissues, the lysis of solid tissue sample by homogenization using Polytoron, use appropriate scale of lysis buffer. 6. N-glycopeptides bound to Fbs1 GYR remain on the filter unit and unbound peptides are removed by centrifugation. 7. The eluted peptides are used for N-glycoside identification and N-glycan profiling. Details for both analyses are described in an original paper [12]. Briefly, for N-glycoside identification, lyophilized peptides were dissolved in 50 μl of 40 mM NH4HCO3 in H218O and incubated with PNGase F, and samples were immediately subjected to LC-MS/MS analysis to reduce the chemical deamidation in H218O. For N-glycan profiling, N-glycans released from N-glycopeptides using PNGase F were purified using Ultralink hydrazide resin, and then labeled with 2-AB. N-glycans were separated by ultra-performance liquid chromatography. 8. For time-lapse imaging, use of a medium without Phenol red, such as FluoroBrite DMEM (Thermo Fisher) is recommended. 9. Halo Tag ligand binds to Halo-tagged protein through covalent bonds during live-cell labeling and are available for fixing cells with paraformaldehyde. Therefore, antibodies can be stained with organelle marker proteins. The time lapse imaging of TMR ligand-labeled Hela cells treated with lysomotropic compound LLOMe, and the fixing cell images of cells labeled with diAc FAM ligand or TMR ligand, which are treated with ionophores, are shown in Fig. 3.

F-box Protein-Recognizing Sugar Chains: Fbs

223

Fig. 3 Detecting damaged organelles by using Fbs3-Halo. (a) Live cell imaging of TMR ligand-labeled HeLa cells expressing Fbs3-Halo. Ligand-labeled cells were treated with 1 mM l-leucyl-l-leucin methyl ester (LLOMe), a lysosomotropic compound, for the indicated time. (b) Halo Tag labeling using diAc FAM or TMR ligand with antibodies. HeLa cells were labeled with Halo Tag ligand and treated with Monensin or Nigericin for 6 h. Cells were then fixed with PFA and stained with the indicated antibodies

224

Yukiko Yoshida

References 1. Cardozo T, Pagano M (2004) The SCF ubiquitin ligase: insights into a molecular machine. Nat Rev Mol Cell Biol 5:739–751 2. Ravid T, Hochstrasser M (2008) Diversity of degradation signals in the ubiquitinproteasome system. Nat Rev Mol Cell Biol 9:679–690 3. Yoshida Y, Tanaka K (2018) Cytosolic N-glycans: Triggers for ubiquitination directing proteasomal and autophagic degradation: Molecular systems for monitoring cytosolic N-glycans as signals for unwanted proteins and organelles. Bioessays 40:1700215 4. Yoshida Y, Chiba T, Tokunaga F et al (2002) E3 ubiquitin ligase that recognizes sugar chains. Nature 418:438–442 5. Mizushima T, Hirao T, Yoshida Y et al (2004) Structural basis of sugar-recognizing ubiquitin ligase. Nat Struct Mol Biol 11:365–370 6. Yoshida Y, Tokunaga F, Chiba T et al (2003) Fbs2 is a new member of the E3 ubiquitin ligase family that recognizes sugar chains. J Biol Chem 278:43877–43884 7. Glenn KA, Nelson RF, Wen HM et al (2008) Diversity in tissue expression, substrate binding, and SCF complex formation for a lectin family of ubiquitin ligases. J Biol Chem 283:12717–12729 8. Yoshida Y, Adachi E, Fukiya K et al (2005) Glycoprotein-specific ubiquitin ligases

recognize N-glycans in unfolded substrates. EMBO Rep 6:239–244 9. Yoshida Y, Yasuda S, Fujita T et al (2017) Ubiquitination of exposed glycoproteins by SCFFBXO27 directs damaged lysosomes for autophagy. Proc Natl Acad Sci U S A 114:8574–8579 10. Erhardt JA, Hynicka W, Dibenedetto A et al (1998) A novel F box protein, NFB42, is highly enriched in neurons and induces growth arrest. J Biol Chem 273:35222–35227 11. Mizushima T, Yoshida Y, Kumanomidou T et al (2007) Structural basis for the selection of glycosylated substrates by SCF(Fbs1) ubiquitin ligase. Proc Natl Acad Sci U S A 104:5777–5781 12. Chen M, Shi X, Duke RM et al (2017) An engineered high affinity Fbs1 carbohydrate binding protein for selective capture of N-glycans and N-glycopeptides. Nat Commun 8:15487 13. Wisniewski JR, Zougman A, Nagaraj N et al (2009) Universal sample preparation method for proteome analysis. Nat Methods 6:359–362 14. Zielinska DF, Gnad F, Wisniewski JR et al (2010) Precision mapping of an in vivo N-glycoproteome reveals rigid topological and sequence constraints. Cell 141:897–907

Chapter 23 F-Type Lectins: Structure, Function, and Evolution Gerardo R. Vasta and Chiguang Feng Abstract F-type lectins (FTLs) are characterized by a fucose recognition domain (F-type lectin domain; FTLD) that displays a novel jellyroll fold (“F-type” fold) and unique carbohydrate- and calcium-binding sequence motifs. This novel lectin family comprises widely distributed proteins exhibiting single, double, or greater multiples of the FTLD, either tandemly arrayed or combined with other structurally and functionally distinct domains. Further, differences in carbohydrate specificity among tandemly arrayed FTLDs present in any FTL polypeptide subunit, together with the expression of multiple FTL isoforms in a single individual supports a striking diversity in ligand recognition. Functions of FTLs in self/nonself recognition include innate immunity, fertilization, microbial adhesion, and pathogenesis, among others, revealing an extensive structural/functional diversification. The taxonomic distribution of FTLDs is surprisingly discontinuous, suggesting that this lectin family has been subject to secondary loss, lateral transfer, and functional co-option along evolutionary lineages. Key words Fucose-binding, F-type lectin family, Fucolectin, Eel agglutinin, Cytolysin, Bindin, Structural fold, Self/nonself recognition

Abbreviations AAA CCP CDC C-FTLD CRD CTL CTLD DiscI and DiscII DlFBL FTL FTLD JspFL αL-Fuc MsaFBP32 N-FTLD PmF-lectin

Anguilla anguilla (European eel) agglutinin Complement control module Cholesterol-dependent cytolysin (lectinolysin) from Streptococcus mitis C-terminal FTLD Carbohydrate recognition domain C-type lectin C-type lectin domain Discoidins I and II from the slime mold Dyctiostelium discoideum FTL from sea bass Dicentrarchus labrax F-type lectin F-type lectin domain FTL from Japanese sea perch Lateolabrax japonicus L-fucose Morone saxatilis fucose-binding protein 32 kDa N-terminal FTLD FTL from the pearl oyster Pinctada martensii

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_23, © Springer Science+Business Media, LLC, part of Springer Nature 2020

225

226

Gerardo R. Vasta and Chiguang Feng

RbFTL-3 SauFBL VHSV

1

FTL from rock bream Oplegnathus fasciatus FTL from gilt head bream Sparus aurata Viral hemorrhagic septicemia virus

Introduction F-type lectins (FTLs), the most recently identified lectin family [1, 2], are fucose-binding proteins characterized by a fucose recognition domain (F-type lectin domain; FTLD) that displays a novel structural fold (the “F-type” fold) and unique fucose- and calciumbinding sequence motifs [3]. While FTLs can display a single FTLD, frequently associated with one or more structurally and functionally distinct domains in a single polypeptide, multiple tandemly arrayed FTLDs are also a common feature [1, 2, 4]. FTLs are widely distributed in Nature, from viruses to vertebrates [1, 2] and while some FTLs mediate immune recognition [1–8], others are involved in microbial pathogenesis [9–11], fertilization [12–14], and other functions. The identification of the FTL family resulted from the isolation of a 32 kDa fucose-binding lectin from serum and liver extracts from the striped bass (Morone saxatilis) which we designated MsaFBP32 [1, 2]. Protein sequencing, followed by cDNA and genomic cloning, enabled the identification of two similar, albeit not identical, tandemly arrayed fucose-binding domains. Analysis of the two 140 amino acid-long domain sequences failed to reveal a signature motif that would identify MsaFBP32 as a member of any known lectin family described at the time. A further search of sequence databases, however, revealed a stretch of N-terminus sequence from PXN1-XENLA, a “long pentraxin” cloned from the liver of the frog Xenopus laevis [15, 16], that appeared significantly similar to the MsaFBP32 lectin domain (see Fig. 2, Chapter 24). This information facilitated the cloning of similar lectins in several fish species, and subsequently, the in silico identification of related sequences in expressed sequence tag and genomic databases for multiple invertebrate and vertebrate species, as well as Streptococcus pneumoniae [1, 2] (Fig. 1). As a whole, this experimental and in silico effort led to the identification of the novel lectin family (FTL family) characterized by proteins present in both prokaryotes and eukaryotes, that displayed the newly identified lectin domain (FTLD), either tandemly arrayed or in mosaic combinations with other structurally and functionally distinct domains [1, 2] (Fig. 2). The functional characterization of FTLs from fish revealed their roles as opsonins for potential bacterial pathogens in innate immunity [7, 8], the Streptococcus FTLs as virulence factors (lectinolysins) [9–11], and the sperm acrosomal proteins (bindins) from the Pacific oyster (Crassostrea gigas) as extremely diversified FTLs with roles in

F-Type Lectins

227

Fig. 1 Multiple alignment of individual CRDs from selected members of the F-type lectin family. F-type lectins with multiple CRDs are parsed into individual domains. Sequence names are composed of species name, protein name, if given, followed by the domain number from N to C terminus in the case of tandem F-type lectins. Invariant residues are shaded in black; conserved residues in 80 of sequences are shaded in gray with white lettering; conservatively substituted residues in 60% of sequences are shaded in grey with black lettering. Sequence numbering reflects that mature polypeptide as determined by N-terminus sequencing of protein or from predicted cleavage site [12] of the polypeptide sequence derived from the cDNA’s open reading frame (ORF). Above the alignment is illustrated the secondary structure as determined from the Anguilla anguilla agglutin (AAA) crystallographic structure. Cylinders represent 310 helices and arrows represent beta strands. Residues that interact with the bound metal through either the backbone oxygen (B), side chain oxygen (S), or both (B/S) are highlighted in blue above the alignment. Cysteines involved in disulfide bridging are highlighted in red and partners are indicated by red italic numbers above the alignment. Residues involved in hydrogen bonding to αL-fucose are highlighted in yellow. Below the alignment are illustrated the degenerate primers (green arrows) used in cloning the zebrafish homolog. Msa, Morone saxatilis; Xla, Xenopus laevis; Omy, Oncorhyncus mykiss; Dre, Danio rerio; Cca, Cyprinus carpio; Fru, Fugu rubripes; Aja, Anguilla japonica; Aan, Anguilla anguilla (AAA of crystallographic structure); Ttr, Tachypleus tridentatus; Spn, Streptococcus pneumoniae; and CG9095 is a Drosophila melanogaster receptor homologous to furrowed (fw). The alignment was produced with ClustalX v.1.81 [13] and similarity shaded with GeneDoc v.2.6.002 [14]

fertilization [12–14]. In recent years, the growing number of sequenced genomes has enabled the identification of FTLs in additional taxa, including viruses, thereby revealing the strikingly wide distribution of the FTLD in Nature [17]. In the following sections, the most relevant structural, functional, and evolutionary aspects of the FTL family are discussed.

2

Structural Features of FTLs

2.1 The FTL Structural Fold

The resolution of the structure of the AAA/α-L-fucose (αL-Fuc) complex revealed a new lectin fold (FTL fold) and identified the amino acid residues that interact with the nonreducing terminal fucose and coordinate the divalent cation, and enabled modeling

228

Gerardo R. Vasta and Chiguang Feng

A

B

AAA TachylecƟn-4,

FTLD HTLD

MsaFBP32 D. rerio FBP38

CTLD

Discoidin II

PXN

X. tropicalis REG TMB

S. pneumoniae TIGR4

CCP

O. mikiss FBP77, X. laevis-FBL2 X. laevis PXN120 C. gigas bindin

(n)

Drosophila furrowed-like receptor

Fig. 2 Domain organization of F-type lectins (FTLs) from prokaryotes, invertebrates, and vertebrates: (a) schematic illustration of selected domain types found in FTLs: FTLD F-type lectin domain, HTLD H-type lectin domain, CTLD C-type lectin domain, PXN pentraxin, REG regulatory, TMB transmembrane domain, and CCP complement control protein domain. (b) Schematic illustration of the domain organization in selected examples of FTLs (1–5 FTLDs and chimeric molecular species) described in prokaryote, invertebrate, and vertebrate species. The subscript “n” indicates the extended number of CCP domains present in furrowed. Adapted from [1, 2]

those that interact with the subterminal sugar units of an oligosaccharide ligand [3]. Further, this structural information led to the rigorous identification of the FTL fucose- and calcium-binding sequence motifs [3]. The F-type lectin fold (Fig. 3a) consists of a β-barrel with jelly roll topology comprising two β-sheets of three and five antiparallel β-strands, respectively, placed against each other (Fig. 3a). On the “top” face of the barrel, the connecting β-strands from the opposite β-sheets form five loops (CDRs 1–5) that surround the heavily positively charged cleft that binds the αL-Fuc. Two short antiparallel strands close the “bottom” of the barrel [3]. At the side of the barrel, a substructure containing three 310 helices tightly coordinates a cation, most likely calcium, that contributes to stabilize the fold, but unlike C-type lectins (CTLs), does not directly interact with the carbohydrate [3]. The AAA subunits can form chlorideinduced trimers that contain one cation (Ca2+) per domain and several Cl placed on the threefold axis, and two trimers can form hexamers with opposing carbohydrate-binding surfaces [3] (Fig. 3b).

F-Type Lectins

A

B

C

D

229

Fig. 3 Structure of A. anguilla agglutinin (AAA) and quaternary structure of AAA oligomers. (a) Ribbon diagram of AAA showing the two β-sheets, the loops (CDRs) encircling the binding site, and 310 helices. Bound αL-Fuc is shown as a stick model above the lectin in yellow. Calcium is shown as a blue sphere. (b) Quaternary structure of the AAA hexamer: In each trimer, a single chlorine ion coordinated by Lys16 from each subunit determines the three-fold axis of rotation, and the hexamer is formed by two stacked trimers with opposing carbohydrate-binding faces. (c) Primary binding site: Interactions of the AAA binding site with αL-Fuc: The three basic amino acid residues that interact with the axial OH on C4 are indicated with the red boxes. The interaction of the disulfide bond (Cys82-Cys83) with the C1-C2 bond of the αL-Fuc is indicated with a circle. (d) Extended binding site: Model of the interactions between AAA and a terminally fucosylated Lea trisaccharide. Interactions on the protein with the αL-Fuc are indicated as in (c) above. Subterminal GlcNAc and Gal are indicated by purple circles, and the interacting amino acid residues are labeled. See text for details. Adapted from [3] 2.2 Carbohydrate Recognition by the FTLD

The AAA/αL-Fuc structure revealed that the protein binds to αL-Fuc through hydrogen bonds established between the side chains of three basic amino acid residues (His52, Arg79 and Arg86)

230

Gerardo R. Vasta and Chiguang Feng

situated in a shallow cleft, the “primary binding site,” and the axial 4-OH of the sugar. Interactions are also established between this basic triad and the ring O5 and equatorial 3-OH of the sugar [3] (Fig. 3c). A unique disulfide bridge formed by contiguous cysteines (Cys82 and Cys83) establishes a van der Waals contact with the bond between C1 and C2 of the αL-Fuc ring, and the C6, which fits into a hydrophobic pocket formed by His27 and Phe45, together with Leu23 and Tyr46 [3]. The AAA/αL-Fuc structure also enabled the modeling of potential interactions between the protein and αLFuc-containing oligosaccharides such as H and Lewis moieties, that are specifically recognized through interactions with amino acid residues located in the so called “extended binding site” [3] (Fig. 3d). AAA recognizes blood group H type 1 (Fucα1–2 Galβ1–3GlcNAcβ1–3Galβ1–4Glc) and Lea (Galβ1–3[Fucα1–4] GlcNAcβ1–3Galβ1–4Glc) oligosaccharides via additional interactions established between amino acid residues in CDRs 1–5, that encircle the binding cleft, with the subterminal units of the H1 and Lea trisaccharides [3]. Alignment of the AAA and MsaFBP32 sequences with those from multiple species [1, 2], and the structural analysis of the primary and extended binding sites of FTLDs of AAA and MsaFBP32 discussed above [3, 4], revealed highly conserved sequence motifs for carbohydrate and calcium binding. The fucose-binding sequence motif was defined as follows: His followed 24 residues down-stream by a segment of sequence that starts with an arginine followed one residue apart by a negatively charged residue, which salt-bridges the preceding arginine, and ends with a basic residue [HX24RXDX4 (R or K); where X indicates any amino acid residue]. The cation-binding sequence motif was defined as follows: h2DGx, where h indicates a small hydrophobic amino acid residue (i.e., V, A, or I) and x stands for a small hydrophilic residue (i.e., N, D, or S) [3]. Some FTLDs, however, deviate from the fucose-binding sequence motif, and the changes may suggest either different specificities or loss of sugar-recognition activity. For example, in Drosophila CG9095, two amino acid residues of the basic triad are replaced by aliphatic residues, which are unlikely to establish the hydrogen-bonds typical of the canonical FTLD [1, 2]. Similarly, replacements of metal-coordinating residues are frequent [1, 2]. 2.3 Domain and Subunit Organization of FTLs 2.3.1 FTLs with Tandemly Arrayed FTLDs

The amino acid sequences of the N-terminal FTLD (N-FTLD) and C-terminal FTLD (C-FTLD) of MsaFBP32 are similar but not identical, suggesting that they display different carbohydrate specificity [1, 2] (Fig. 4). This is supported by differences in the topology and surface potential of the primary and extended binding site of the N- and C-FTLDs [4], suggesting that the N-FTLD binding site recognizes more complex fucosylated oligosaccharides and with a relatively higher avidity than the C-FTLD [4].

F-Type Lectins

231

Fig. 4 Structure of the binary FTL MsaFBP32. View of the MsaFBP32 isolated monomeric subunit (left) and the asymmetric subunit of the crystal trimer (right). In the trimer, the subunits are arranged as “head-to-head” and “tail-to-tail” (in the image, the N-FTLDs are at the top and the N-terminal FTLDs are at the bottom), resulting in a cylindrical structure with two opposite binding “faces” with different carbohydrate specificity. Adapted from [3] 2.3.2 FTL Isoforms and Diversity in Ligand Recognition

In single individuals, the presence of multiple FTL isoforms (isolectins) that display amino acid replacements at positions that are critical for sugar recognition strongly suggests diversity in carbohydrate specificity [2–4], a feature that is key not only for proteins involved in innate immunity such as in the eel FTLs [5] but also for those that recognize heterogeneous endogenous glycan ligands, as the oyster bindins [12–14]. Nevertheless, the regulation of expression of FTLs in both immune and developmental processes and the detailed recognition mechanisms involved remain poorly understood.

2.3.3 Oligomeric Organization of FTL Subunits

FTLs that carry a single FTLD, such as AAA can become multivalent by forming oligomeric structures [3]. For those FTLs that carry two or more FTLDs in each polypeptide, such as MsaFBP32, their intrinsic multivalency and diversity in glycan recognition is further expanded by their association into oligomeric species [1, 2, 4] (Fig. 4). For example, although the N- and C-FTLDs of

232

Gerardo R. Vasta and Chiguang Feng

MsaFBP32 are structurally similar, important sequence differences suggest that the N-FTLD recognizes fucosylated oligosaccharides of higher complexity and with a relatively higher avidity than the C-FTLD [4]. The physiological structures of AAA are homotrimers with a threefold cyclic symmetry, which in turn dimerize to form a hexamer with three FTLD binding sites on opposite faces of the structure that optimizes their orientation and spacing (26 A˚) for binding to glycan moieties on microbial surfaces. In contrast, the MsaFBP32 subunit carries two tandemly arrayed FTLDs, that in the native trimer are arranged in a “tail-to-tail” manner, resulting in a virtual cylindrical structure with opposing globular structures (one with the three N-FTLDs and the other with the three C-FTLDs) connected by the linker peptides, that can potentially cross-link different glycans [4] (Fig. 4). 2.3.4 Structural Diversification of FTLs

The mining of currently available transcriptomic and genomic databases has revealed a striking variety of sequences encoding for multiple FTLDs that revealed not only the frequency of domain duplication but also domain shuffling within the FTL family [1, 2] (Figs. 1 and 2). Further, the presence in both prokaryotes and eukaryotes of FTLDs in various combinations with other structurally and functionally distinct domains that result in mosaic FTLs, suggest its extensive functional diversification in the evolution of the FTL family [1, 2]. For example, while prokaryotic FTLs mostly carry single FTLDs in combination with diverse domains, eukaryotes more frequently display tandemly arrayed FTLDs, also in combination with other domains such as carbohydrate-binding domains from other lectin families [CTL domains (CTLDs), and pentraxins], complement control modules (CCP), transmembrane domains, and FA58C domains [1, 2, 17]. For example, in Drosophila, these domains include a CCP, a CTLD, and a predicted transmembrane domain [1, 2, 18]. It is noteworthy that in CG9095 the CTLD is unlikely to bind carbohydrate because the canonical residues of the CRD are missing [1, 2] (Figs. 1 and 2). The considerable diversity evident from these topologies, in which the binding site motif is strictly conserved, suggests a diverse spectrum of functions fulfilled by specific recognition of αL-Fuc in various environments [1, 2]. Overall, while among closely related organisms the domain organization of FTLs is consistent with their taxonomic placement, unusual domain associations occur frequently as well as a strikingly discontinuous distribution among taxa [1, 2, 17]. This observation can be interpreted not only as a unique evolutionary and ecological adaptability of this lectin family but also as frequent lateral transfer along viral, prokaryotic, and eukaryotic lineages [1, 2, 17]. Finally, it is noteworthy that although the FTL structural fold is distinctive of FTLs from in viruses to vertebrates, a structure-based search (DALI database

F-Type Lectins

233

[19]) identified several proteins that share the jellyroll FTL fold with AAA, although with negligible sequence similarity (2–14% sequence identity with AAA) [3]. These proteins include, among others, the C1 and C2 repeats of human blood coagulation factor V (FVa-C1 and -C2) [20], the C-terminal domain of a bacterial sialidase [21], and the NH2-terminal domain of a fungal galactose oxidase [22, 23]. These observations are intriguing concerning the evolutionary history of FTLs as emerging from carbohydratebinding domains in glycoenzymes, or alternatively, suggest that the recognition properties of the FTLs have been drastically modified or co-opted to carry out distinct functions [1–3].

3

Taxonomic Distribution and Evolution of the FTLD The recognition of FTLs as a novel lectin family resulted from initial identification and characterization of the FTLD sequence motif and the F-type structural fold in taxa ranging from prokaryotes to amphibians [1–3]. FTLs were found in lophotrochozoan (molluscs and planaria) and ecdysozoan protostomes (horseshoe crabs and insects), deuterostome invertebrates (sea urchin), and coldblooded vertebrates such as elasmobranchs (skate), lobe- and ray-finned teleost fish, and amphibians (Xenopus spp. and salamander) [1, 2]. The organizational topologies of FTLDs in fish and amphibians appear to be lineage-related [1, 2]. For example, most teleost FTLs contain either two, or four tandemly arrayed FTLDs, whereas in Xenopus spp. FTLs are organized from single FTLDs to combinations of two, three, or four FTLDs, and as chimeric proteins containing five tandemly arrayed FTLDs adjacent to a pentraxin domain [1, 2]. Thus, the binary FTLs have diversified through lineage-dependent gene duplications that are unique to teleosts and amphibians [1, 2]. Intriguing features of FTLs, however, are their discontinuous taxonomic distribution and the diversified domain architecture, with the FTLD frequently found in combination with other structurally distinct domains. The absence of FTLDs in archaea, protozoa, urochordates, and higher vertebrates [1, 2, 17] suggests a functionally plastic FTLD, which may have either enhanced or lost its fitness value in some taxa [1, 2]. In the latter case, FLTs may have been selectively lost even in relatively closely related lineages [1, 2]. Recently, an exhaustive computational study on publically available databases contributed further insight into the occurrence of the FTLD [17]. In addition to the taxa already identified as expressing FTLDs [1, 2], these were identified in viruses, fungi, reptiles, birds, and prototherian mammals, including the monotremes (platypus) and didelphid marsupials (opossum) [17]. Furthermore, this study confirmed the FTL diversity observed in mollusks [14], the FTLD occurrence and

234

Gerardo R. Vasta and Chiguang Feng

domain organization diversity in hemichordates and cephalochordates, and the absence in archaea, protozoans, urochordates, and eutherian (placental) mammals, reported in earlier publications [1, 2].

4

Biological Roles of FTLs The biological roles of FTLs have only been experimentally investigated in only a few examples, implementing biochemical, molecular, immunological, and structural approaches. For most FTLs, however, their potential functions have been hypothesized based on their binding specificity for either endogenous or exogenous (viral, microbial) glycosylated ligands, cell or tissue localization, and the cues that modulate their gene expression such as immune challenge or environmental stressors. For example, in both the AAA and MsaFBP32, the multivalent trimeric arrangement of the FTLDs and the opposite orientation of the binding surfaces strongly suggest that these lectins can cross-link fucosylated glycans exposed on the surface of neighboring cells or microbial pathogens to host cells [1–4]. Modeling of the MsaFBP32 recognition of fucosylated oligosaccharides from prokaryotes and eukaryotes supports the observation that FTLs with binary tandem CRDs can function as opsonins that promote phagocytosis of microbial pathogens [1, 2, 4], as confirmed for FTLs from sea bass (DlFBL; Dicentrarchus labrax) and gilt head bream (SauFBL; Sparus aurata) [7, 8]. Expression of fish FTLs primarily takes place in liver [1, 2, 7, 8], the typical source of acute phase reactants, but also in gills [5, 7, 8], intestine [7, 8], and plasma [1, 2, 24], which are organs and body fluids continuously exposed to infectious challenge, suggesting a role(s) of FTLs in innate immune defense. However, upregulation of FTL expression by immune challenge has not been the general rule for the species examined. While LPS challenge significantly upregulated expression and increased secretion of FTLs in liver and gill tissue from A. japonica [5], protein levels of DlFBL were only modestly enhanced by Vibrio alginolyticus infectious challenge [8]. Similarly, an inflammatory challenge only increased the liver MsaFBP32 transcript levels in about threefold over the relatively high basal expression levels [1, 2]. In the Japanese sea perch (Lateolabrax japonicus), the FTL JspFL was only upregulated in spleen, while it was also constitutively expressed in liver and gills [25]. A recent in vitro study showed that RbFTL-3, an FTL that is highly expressed in the intestine of rock bream (Oplegnathus fasciatus), controls budding of the viral hemorrhagic septicemia virus (VHSV), increasing the viability of the VHSV infected cells and limiting hemorrhage in fish tissues [26].

F-Type Lectins

235

The roles of FTLs have been also studied in a few wellestablished genetically tractable systems, combining genetic approaches with the detailed analysis of the FTL structures, but some results have been controversial. For example, the slime mold Dyctiostelium discoideum, a model system for developmental and cell biology studies, was used to investigate the biological role(s) of discoidins I and II (DiscI and DiscII) [27]. Both discoidins had been reported as secreted lectins involved in cell-substratum adhesion and spore coat formation [28, 29], but in later studies no evidence of secretion was found. A recent study, however, concluded that DiscI mediates cell–substratum adhesion and streaming [30], although the mechanistic aspects are not fully understood. Similarly, while the role of the Drosphila furrowed gene in cell adhesion was firmly established, any participation of the FTLD in this process remains to be demonstrated [1, 2, 18]. As discussed above, FTLs are also present in viral, prokaryotic, and multicellular pathogens and parasites. It is firmly established that microbial pathogens and parasites interact with host cell glycans via carbohydrate-binding domains in their surface proteins [31], suggesting that FTLs, particularly lectinolysins, could participate in microbial virulence [9–11]. The Gram-positive bacterium Streptococcus mitis produce a cholesterol-dependent cytolysin (CDC, lectinolysin) that uses an FTLD to recognize the host’s fucosylated moieties [Lewis y (Ley) and Lewis b (Leb)], and significantly enhance their virulent pore-forming properties: upon binding to the host surface glycans, monomeric CDCs self-assemble to form large β barrel pores that lead to cell lysis [32]. Unlike the roles of FTLDs in innate immune host defense and the bacterial virulence functions discussed above, the acrosomal “bindins” from sperm of the Pacific oyster (C. gigas) are highly polymorphic proteins that bind to the egg perivitelline envelope during fertilization [12]. A single copy bindin gene can produce highly diversified transcripts by positive selection, recombination, and alternative splicing, generating an unusually polymorphic bindin repertory. It is noteworthy that an individual male oyster only translates one or two polymorphic bindins with one to five tandemly arrayed FTLDs [13]. The diversity of bindin FTLDs has been hypothesized as a sperm-oocyte co-evolution mechanism driven by the high diversification of egg surface receptors, and aimed at avoiding polyspermia [14]. Interestingly, PmF-lectin, an FTL from the pearl oyster (Pinctada martensii) is significantly upregulated by infectious challenge (Vibrio alginolyticus), suggesting that it is involved in the innate immune response [33]. In contrast, FTLs have been identified in both the mantle tissue and shell matrix of the clam Mya truncata, suggesting that during the shell biomineralization process, FTLs secreted by the mantle with immune functions, are later incorporated into the shell matrix [34].

236

A

Gerardo R. Vasta and Chiguang Feng

B

Opsonization/phagocytosis

Microbial glycans

Fertilization (?)

Oocyte glycans

FTLD (AAA)

C-FTLD

C-FTLD

(MsaFBP32) N-FTLD

(Bindin) N-FTLD

Anguilla sp., Morone sp. Dicentrarchus sp.

Cell adhesion (?)

Amoeba glycans C-HTLD (DiscI) N-FTLD Substratum glycans

Sperm glycans

Host glycans

C

Crassostrea sp.

Dictyostelium sp.

Fig. 5 Schematic illustration of self- and nonself recognition by FTLs in immunity, fertilization, and cell adhesion: The cartoon illustrates the potential binding activities of FTLDs identified in fish (AAA and MsaFBP32), oyster (bindins), and slime mold (DiscII) discussed in the text: (a) Both AAA and MsaFBP32 oligomers are multivalent and can cross-link glycans on the surface of potential microbial pathogens to the surface of the macrophages, leading to opsonization, phagocytosis, and intracellular killing of the infectious agent. (b) The highly diversified oyster (Crassostrea gigas) bindins, carrying up to five FTLDs, may selectively cross-link sperm or acrosomal glycans to the egg perivitelline envelope enabling only fertilization by sperm that matches the egg glycans and prevent polyspermia. (c) Discoidins secreted by the slime mold (Dyctiostelium discoideum) amoeba may cross-link surface glycans to substratum components, enabling cell–substratum adhesion and streaming

5

Conclusions FTLs are a family of lectins that is structurally highly diversified, of broad taxonomic distribution, and essentially pleiotropic, as they participate in a vast array of functions based on “self” and “nonself” recognition that encompass not only innate immunity but also fertilization, cell adhesion, microbial virulence, among others yet to be unraveled (Fig. 5). Further, increasing evidence has accumulated to support the notion that along their evolution, selected FTLs were co-opted to carry out different functions that may not rely on active carbohydrate-binding sites, and therefore, this property, which is inherent to their definition as lectins, may have been lost in the process. For example, despite the FTL diversity evident in amphibians, reptiles, birds, and prototherian mammals, no bona fide FTL homologs are detectable in genomes of eutherian mammals [1, 2, 17]. Therefore, above the level of the prototherian

F-Type Lectins

237

mammals this lectin family may have been lost as such, either by becoming truly extinct or by being co-opted into other functions, as proposed for the C-1 and C-2 domains of the mammalian clotting factors V and VIII [1–3]. In contrast, a substantial expansion in both FTLD occurrence and domain organization diversity has been observed in mollusks, hemichordates, and cephalochordates, while FTLDs being absent in the urochordates [17], illustrating the paucity in distribution of FTLs among and within evolutionary lineages. It is tempting to speculate that the increased FTLD diversification observed in some taxa, together with the lack of FTLDs in others could be rather attributed to compensatory effects among multiple lectin families, or other recognition molecules, depending on selective advantage(s) that each can provide to any given taxa as more or less effective pattern recognition receptors in innate immunity. In support of this view, it is noteworthy that the urochordate ascidian Clavelina picta, which lacks FTLs, expresses a highly diversified repertoire of fucose-binding CTLs, suggesting that the expansion of the CTL repertoire probably reflects the selective advantage that fucose-binding CTLs provide over FTLs to the ascidian’s innate immune responses [35, 36]. In addition to functions carried out by FTLs, such as pathogen recognition, immobilization, and opsonization, CTLs can also initiate complement activation, an ancient enzyme-driven mechanism that can rapidly amplify opsonization and effect direct killing of the potential pathogen via the membrane attack complex [35, 36]. Therefore, it is possible that these and other functional advantages offered by CTLs led to their expansion as innate immune defense mechanisms in higher mammals, simultaneously with the contraction, co-option, or loss altogether of the FTL family members. Finally, the discontinuous taxonomic distribution of viral and prokaryotic FTLDs suggests the eukaryotic origin for the FTLD, that throughout evolution has been subjected to extensive duplication, mutation, and lateral transfer [1, 2, 17]. The recent increase in the availability of public genome, transcriptome, and proteome databases for mammalian and various nonmammalian model organisms, such as Drosophila, Caenorhabditis elegans, and zebrafish (Danio rerio), coupled with structural studies and innovative forward and reverse genetic approaches for functional analyses has the potential to uncover novel structural, functional, and evolutionary features of members of the FTL family. Furthermore, given the subtle differences observed in the carbohydrate specificity of FTLs for fucosylated ligands, it is expected that their applications in glycosciences as useful reagents for glycan separation or lectin histochemistry, among others, will quickly expand. Finally, FTLs should be promising diagnostic reagents for the identification of the diverse fucosylated moieties that are prevalent during neoplastic transformation and potentially useful therapeutic tools in cancer [37, 38].

238

Gerardo R. Vasta and Chiguang Feng

Acknowledgments The author’s research reviewed herein was supported by Grants IOS 1050518, IOB-0618409, MCB 0077928, and IOS-0822257 from the National Science Foundation, and Grant R01GM070589 from the National Institutes of Health to G.R.V. References 1. Odom EW, Vasta GR (2006) Characterization of a binary tandem domain F-type lectin from striped bass (Morone saxatilis). J Biol Chem 281:1698–1713 2. Odom EW (2004) F-type lectins: biochemical, genetic and structural characterization of a novel lectin family in lower vertebrates. PhD thesis, MEES program, University of Maryland. 3. Bianchet MA, Odom EW, Vasta GR, Amzel LM (2002) A novel fucose recognition fold involved in innate immunity. Nat Struct Biol 9:628–634 4. Bianchet MA, Odom EW, Vasta GR, Amzel LM (2010) Structure and specificity of a binary tandem domain F-lectin from striped bass (Morone saxatilis). J Mol Biol 401:239–252 5. Honda S, Kashiwagi M, Miyamoto K, Takei Y, Hirose S (2000) Multiplicity, structures, and endocrine and exocrine natures of eel fucosebinding lectins. J Biol Chem 275:33151–33157 6. Saito T, Hatada M, Iwanaga S, Kawabata S (1997) A newly identified horseshoe crab lectin with binding specificity to O-antigen of bacterial lipopolysaccharides. J Biol Chem 272:30703–30708 7. Cammarata M, Benenati G, Odom EW, Salerno G, Vizzini A, Vasta GR, Parrinello N (2007) Isolation and characterization of a fish F-type lectin from gilt head bream (Sparus aurata) serum. Biochim Biophys Acta 1770:150–155 8. Salerno G, Parisi MG, Parrinello D, Benenati G, Vizzini A, Vazzana M, Vasta GR, Cammarata M (2009) F-type lectin from the sea bass (Dicentrarchus labrax): purification, cDNA cloning, tissue expression and localization, and opsonic activity. Fish Shellfish Immunol 27:143–153 9. Ohkuni H, Todome Y, Okibayashi F, Watanabe Y, Ohtani N, Ishikawa T, Asano G, Kotani S (1997) Purification and partial characterization of a novel human platelet aggregation factor in the extracellular products of

Streptococcus mitis, strain Nm-65. FEMS Immunol Med Microbiol 17:121–129 10. Tweten RK (2005) The cholesterol-dependent cytolysins: a family of versatile pore-forming toxins. Infect Immun 73:6199–6209 11. Feil SC, Lawrence S, Mulhern TD, Holien JK, Hotze EM, Farrand S, Tweten RK, Parker MW (2012) Structure of the lectin regulatory domain of the cholesterol-dependent cytolysin lectinolysin reveals the basis for its lewis antigen specificity. Structure 20(2):248–258 12. Moy GW et al (2008) Extraordinary intraspecific diversity in oyster sperm bindin. Proc Natl Acad Sci U S A 105:1993–1998 13. Moy GW, Vacquier VD (2008) Bindin genes of the Pacific oyster Crassostrea gigas. Gene 423:215–220 14. Springer SA et al (2008) Oyster sperm bindin is a combinatorial fucose lectin with remarkable intra-species diversity. Int J Dev Biol 52:759–768 15. Seery LT, Schoenberg DR, Barbaux S, Sharp PM, Whitehead AS (1993) Identification of a novel member of the pentraxin family in Xenopus laevis. Proc Biol Sci 253(1338):263–270 16. Tennent GA, Pepys MB (1994) Glycobiology of the pentraxins. Biochem Soc Trans 22 (1):74–79 17. Bishnoi R, Khatri I, Subramanian S, Ramya TNC (2015) Prevalence of the F-type lectin domain. Glycobiology 25(8):888–901 18. Leshko-Lindsay L, Corces VG (1997) The role of selectins in Drosophila eye and bristle development. Development 124:169–180 19. Holm L, Sander C (1993) Protein structure comparison by alignment of distance matrices. J Mol Biol 233:123–138 20. Macedo-Ribeiro S, Bode W, Huber R, QuinnAllen MA, Kim SW, Ortel TL, Bourenkov GP, Bartunik HD, Stubbs MT, Kane WH, FuentesPrior P (1999) Crystal structures of the membrane-binding C2 domain of human coagulation factor V. Nature 402 (6760):434–439

F-Type Lectins 21. Gaskell A, Crennell S et al (1995) The three domains of a bacterial sialidase: a betapropeller, an immunoglobulin module and a galactose-binding jelly-roll. Structure 3:1197–1205 22. Ito N, Phillips SE et al (1991) Novel thioether bond revealed by a 1.7 a crystal structure of galactose oxidase. Nature 350:87–90 23. Firbank SJ, Rogers MS et al (2001) From the cover: crystal structure of the precursor of galactose oxidase: an unusual self-processing enzyme. Proc Natl Acad Sci U S A 98:12932–12937 24. Parisi MG, Cammarata M, Benenati G, Salerno G, Mangano V, Vizzini A, Parrinello N (2010) A serum fucose-binding lectin (DlFBL) from adult Dicentrarchus labrax is expressed in larva and juvenile tissues and contained in eggs. Cell Tissue Res 341:279–288 25. Qiu L, Lin L, Yang K, Zhang H, Li J, Zou F, Jiang S (2011) Molecular cloning and expression analysis of a F-type lectin gene from Japanese sea perch (Lateolabrax japonicus). Mol Biol Rep 38(6):3751–3756 26. Cho SY, Kwon J, Vaidya B, Kim JO, Lee S, Jeong EH, Baik KS, Choi JS, Bae HJ, Oh MJ, Kim D (2014) Modulation of proteome expression by F-type lectin during viral hemorrhagic septicemia virus infection in fathead minnow cells. Fish Shellfish Immunol 39(2):464–474 27. Mathieu SV, Araga˜o KS, Imberty A, Varrot A (2010) Discoidin I from Dictyostelium discoideum and Interactions with oligosaccharides: specificity, affinity, crystal structures, and comparison with discoidin II. J Mol Biol 400 (3):540–554 28. Barondes SH, Cooper DN, Haywood-Reid PL (1983) Discoidin I and discoidin II are localized differently in developing Dictyostelium discoideum. J Cell Biol 96:291–296 29. Barondes SH, Haywood-Reid PL, Cooper DN (1985) Discoidin I, an endogenous lectin, is externalized from Dictyostelium discoideum in multilamellar bodies. J Cell Biol 100:1825–1833

239

´ lvarez-Gonza´lez B, del A ´ lamo 30. Bastounis E, A JC, Lasheras JC, Firtel RA (2016) Cooperative cell motility during tandem locomotion of amoeboid cells. Mol Biol Cell 27 (8):1262–1271 31. Imberty A, Varrot A (2008) Microbial recognition of human cell surface glycoconjugates. Curr Opin Struct Biol 18(5):567–576 32. Bouyain S, Geisbrecht BV (2012) Host glycan recognition by a pore forming toxin. Structure 20(2):197–198 33. Chen J, Xiao S, Yu Z (2011) F-type lectin involved in defense against bacterial infection in the pearl oyster (Pinctada martensii). Fish Shellfish Immunol 30(2):750–754 34. Arivalagan J, Marie B, Sleight VA, Clark MS, Berland S, Marie A (2016) Shell matrix proteins of the clam, Mya truncata: roles beyond shell formation through proteomic study. Mar Genomics 27:69–74 35. Quesenberry MS, O’leary N, Ahmed H, Bianchet M, Amzel M, Marsh A, Vasta GR (1998) The protochordate Clavelina picta has key components of innate immunity in mammals MBP-like, MASP-like, and complementlike molecules. FASEB J 12(8):A1351, Il 24 36. Quesenberry MS, Ahmed H, Elola MT, O’Leary N, Vasta GR (2003) Diverse lectin repertoires in tunicates mediate broad recognition and effector innate immune responses. Integr Comp Biol 43(2):323–330 37. Wu L, Yang X, Duan X, Cui L, Li G (2014) Exogenous expression of marine lectins DlFBL and SpRBL induces cancer cell apoptosis possibly through PRMT5-E2F-1 pathway. Sci Rep 4:4505 38. Li G, Gao Y, Cui L, Wu L, Yang X, Chen J (2016) Anguilla japonica lectin 1 delivery through adenovirus vector induces apoptotic cancer cell death through interaction with PRMT5. J Gene Med 18(4–6):65–74 39. Cammarata M, Vazzana M, Chinnici C, Parrinello N (2001) A serum fucolectin isolated and characterized from sea bass Dicentrarchus labrax. Biochim Biophys Acta 1528:196–202

Chapter 24 Purification and Biochemical Characterization of Selected F-Type Lectins Chiguang Feng and Gerardo R. Vasta Abstract The purification of fucose-binding lectins from the liver of striped bass (Morone saxatilis), a teleost fish, and the identification of a novel lectin sequence motif led to the identification of a new family of lectins, the F-type lectins (FTLs) (see overview of the FTL family in Chapter 23). Isolation and purification of these proteins from liver extracts of striped bass was accomplished by affinity chromatography and size exclusion, and their identification as FTLs, by direct Edman sequencing, and protein, transcript, and gene sequence analysis. The development of specific antibodies against the M. saxatilis FTL provided an additional tool for the identification of FTLs. These methods have been successfully used for the purification of the FTL family members from tissues and body fluids of various animal species. Production and characterization of FTLs has been facilitated by the expression of the recombinant proteins. In this chapter, the biochemical characterization of FTLs is focused on the analysis of their carbohydrate specificity. Key words Affinity and size-exclusion chromatography, Sequence analysis, F-type lectin sequence motif, Recombinant F-type lectin, Carbohydrate specificity

Abbreviations BSA BSM CBB ELISA FBD FTL Igs IPTG MsaFBP32 NCFG OD OSM PBS PMSF PSM

Bovine serum albumin Bovine submaxillary mucin Coomassie Brilliant Blue Enzyme-linked immunosorbent assay Fucose-binding domains F-type lectin Immunoglobulins Isopropyl β-D-thiogalactoside M. saxatilis fucose-binding protein 32 kDa National Center for Functional Glycomics Optical density Ovine submaxillary mucin Phosphate-buffered saline Phenylmethylsulfonyl fluoride Porcine stomach mucin

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_24, © Springer Science+Business Media, LLC, part of Springer Nature 2020

241

242

Chiguang Feng and Gerardo R. Vasta

RBC RP-HPLC SEC TFA

1

Red blood cells Reverse phase-HPLC Size exclusion chromatography Trifluoroacetic acid

Introduction Taking advantage of their carbohydrate-binding properties, affinity chromatography on selected sugars and glycans conjugated to various resins has become the most widely used method for the initial isolation of lectins from animal [1–8] or plant tissues [9, 10]. Subsequently, various chromatographic, electrophoretic, or centrifugation methods based on additional molecular properties (molecular size, charge, hydrophobicity, and general hydrodynamic behavior) are routinely applied for the rigorous purification of the isolated proteins. Through affinity chromatography on L-fucose-Sepharose, the first member of the FTL family was identified from the striped bass Morone saxatilis [7, 11–13]. In contrast to many known C-type lectins, the 32 kDa protein designated as MsaFBP32 (M. saxatilis fucose-binding protein 32 kDa) did not require calcium or other divalent cations for binding to cells (Fig. 1). Partial Edman sequencing of the protein (Table 1) followed by cDNA and genomic cloning revealed two 140-amino acid tandemly arrayed domains in MsaFBP32. Initial BLAST search revealed that the MsaFBP32 N- and C-terminal fucose-binding domains (FBDs) shared significant similarity to the N-terminal region of the long pentraxin or PXN-fusion protein from the African clawed frog Xenopus laevis with a highly conserved set of amino acid residues that defined a tentative fucose-binding sequence motif (Fig. 2) [14]. Further BLAST search of EST and genomic databases revealed multiple sequences from X. laevis and X. tropicalis, as well as many sequences from invertebrate and vertebrate species that shared the FBD sequence motif [7, 11]. Recent advances in transcript and gene sequencing methodologies (Next-Generation Sequencing; NGS) have facilitated the identification of proteins that carry a particular feature of interest, such as the FTL sequence motif. We recently used this strategy to identify FTLs from the zebrafish Danio rerio, a genetically tractable model system (Odom, Feng and Vasta, unpublished) (Fig. 3). In the following sections, we summarize the methodology for the purification of authentic or recombinant FTL family members as routinely carried out in our laboratory, and the characterization of their biochemical properties, focusing on their binding selectivity for glycosylated ligands as tested by enzyme-linked immunosorbent assay (ELISA), microhemagglutination, and glycan array analysis.

Purification and Characterization of F-Type Lectins

243

Fig. 1 SDS-PAGE analysis of the purification steps for the bass serum lectin. Coomassie stain reflects relative band intensities while silver stain is for maximum sensitivity. Lanes are labeled according to the purification steps. MW: Molecular weight standards, Serum: Striped bass serum applied to the affinity chromatography column, Aff: Eluate from the affinity chromatography column, SEC: Peak eluate from the size-exclusion column

Table 1 Edman sequencing of peptides isolated from striped bass hepatic lectin: Residues in parentheses represent the most abundant amino acid when multiple peaks were produced during a degradation cycle. When an amino acid could not be assigned with certainty, the residue was designated with an X. NA indicates not applicable Peptide name

Digestion enzyme

Peptide sequence

NH2-terminal



YNYKNYAL(R)GKATQXA(R)YLX(T)(S)

3081

Trypsin

NSDFEAGSCTHTIEQTNPX(S)

3082

Trypsin

YVTVLLPGTNK

3083

Trypsin

VDLIEPYITSITIT(N)(R)

V437

Trypsin

YVNIVIPGREEYLTLCEVEVYGSVL(L)

V438

Trypsin

ATQSSLFESGIAYNAIDGNQAN(NNWEMASETH)

Lys-C

(K)NTMNP

Residues in parenthesis are tentatively assigned

244

Chiguang Feng and Gerardo R. Vasta

Fig. 2 Similarity of the Xenopus laevis pentraxin-fusion protein to the tandem domains of the striped bass lectin. (a) Schematic domain organization of the X. laevis pentraxin-fusion protein (the oval represents the pentraxin domain) and the 32 kDa striped bass (a hybrid from M. saxatilis x M. chrysops; MsxMc) fucosebinding protein (the rectangles represent the N- and C-terminal fucose-binding domains). (b) Multiple amino acid alignment of the N- and C-terminal domains of the striped bass FBP32 and the N-terminal domain of the pentraxin fusion protein (rectangles illustrated in panel a). The alignment is marked every ten residues with an asterisk and residue numbering follows from the mature polypeptide sequence. Note the consensus of Cys residues. (c) The FTL canonical fucose-binding sequence motif: HX24RXDX4 (R or K) (X indicates any amino acid residue]: His followed 24 residues downstream by a segment of sequence that starts with an arginine followed one residue apart by a negatively charged residue, which salt-bridges the preceding arginine and ends with a basic residue [6]. (d) The FTL cation-binding sequence motif: h2DGx, where h indicates a small hydrophobic amino acid residue (i.e., V, A, or I) and x stands for a small hydrophilic residue (i.e., N, D, or S) [6]

2

Materials

2.1 Purification of FTLs

1. 0.1 M PMSF: dissolve 17.4 g of phenylmethylsulfonyl fluoride (PMSF) in 1 L of ethanol and pass throw 0.22 μm sterile filter. Keep at 4  C. 2. Extraction buffer: prepare 10 mM Tris–HCl at pH 7.8, with 100 mM lactose, 25 mM KCl, 0.1 mM PMSF, 2 μg/mL aprotinin, 1 μg/mL leupeptin, and 1 μg/mL pepstatin. Keep at 4  C. 3. TBS-Ca: prepare 50 mM Tris–HCl at pH 7.5, with 100 mM NaCl and 10 mM CaCl2. Keep at 4  C.

Purification and Characterization of F-Type Lectins

245

Fig. 3 Multiple alignment of two-CRD F-type lectins from zebrafish (Danio rerio), pufferfish (Fugu rubripes and Tetraodon nigroviridis), and striped bass (Morone saxatilis). Putative mature peptides were determined from cleavage sites predicted by the SignalP algorithm [22]. Only the two full sequence paralogues of the three identified in zebrafish are included. The longer ORF of DreFBPL1 is illustrated by the unmatched residue stretch toward the carboxy terminal side. The ORF of DreFBPL3, identified from dbEST was confirmed by PCR. Orthologues and paralogues are paired between the pufferfish species. The pufferfish ORFs were reconstructed from their gene sequence using the known gene organization from the striped bass. Cysteines involved in disulfide bridging are highlighted in red and partners are indicated by red italic numbers above the alignment. Residues involved in hydrogen bonding to L-fucose are highlighted in yellow. Abbreviations: Dre, Danio rerio; Fru, Fugu rubripes; Tni, Tetraodon nigroviridis; and Msa, Morone saxatilis

4. PBS: dilute 100 mL of PBS (10), pH 7.4 into 900 mL of distilled water. 5. LB medium: dissolve 21.5 g of Luria Broth power (RPI Research Products) in distilled water, autoclave for 20 min at 121  C. Supply appropriated antibiotics (e.g., 50 μg/mL of Kanamycin) before use. 6. 0.1 M IPTG: dissolve 23.8 g of isopropyl β-D-thiogalactoside (IPTG) in distilled water and pass throw 0.22 μm sterile filter. 7. BugBuster lysis buffer: dilute 10 BugBuster (Novagen) with distilled water, supplied with 1 mM PMSF, 1 protease inhibitor (Calbiochem), Benzonase Nuclease (EMD Millipore), and lysozyme (Sigma). 8. L-fucose-Sepharose-6B column: covalently bound L-fucose to the chromatography resin Sepharose-6B-CL (Sigma, St. Louis, MO, USA) by the divinylsulfone method [15]. 9. 200 mM L-fucose in TBS-Ca: dissolve 32.8 g of L-fucose in 1 L of TBS-Ca and stir until complete dissolution.

246

Chiguang Feng and Gerardo R. Vasta

10. UviCord II detector (Amersham Biosciences). 11. Centricon 10 (Millipore: Billerica, MA). 12. Superose 12 10/30 HR column (Amersham Biosciences). 2.2 Biochemical Characterization of FTLs

1. PBST: dissolve 1 mL of Tween 20 in 1 L of 1 PBS and stir until complete dissolution. 2. 3% BSA in PBS: dissolve 3 g of bovine serum albumin (BSA) in 100 mL of 1 PBS and stir until complete dissolution. 3. 2% formaldehyde in PBS: mix same volume of 4% formaldehyde (Fisher Scientific) with 1 PBS to prepare 2% solution. 4. 100 mM glycine in PBS: dissolve 7.5 g of glycine in 1 L of 1 PBS and stir until complete dissolution. 5. Low-molecular size SEC standards of known hydrodynamic radius (Amersham Biosciences). Prepare two sets of calibration protein mixes in water: ˜ €a, Set I consisting of bovine serum albumin (BSA; 35.5 AA ˜ €a, 67 kDa; 7 mg/mL) and chymotrypsinogen 27 A (20.9 AA 25 kDa; 3 mg/mL). ˜ €a, 43 kDa; Set II consisting of ovalbumin (30.5 AA ˜ 7 mg/mL) and RNase A (16.4 AA€a, 13.7 kDa; 10 mg/mL) (see Note 1). 6. Dextran blue solution: dissolve dextran blue 2000 (Amersham Biosciences) to 2 mg/mL in a solution of 5 mg/mL chromatography-grade acetone (0.792 g/cm3) (Fisher Scientific) in water. 7. TMB microwell Peroxidase Substrate (SeraCare Life Sciences Inc.). 8. 0.1% Pronase: dissolve Pronase (Calbiochem-Novabiochem; La Jolla, CA, USA) in physiological saline. 9. Applied Biosystems model 140A pumps. 10. Model 1000S diode-array detector (2.3 mL flow cell, 0.0025 inch i.d. tubing; Applied Biosystems; Foster City, CA, USA). 11. Zorbax-SB C-18 silica column (0.1  15 cm, dp~5 mm, 300 A˚ pore size; Microtech Scientific; Saratoga, CA, USA). 12. Aquapore ODS-300 C-18 silica column (0.1  25 cm, dp~7 mm, 300 A˚ pore size, Applied Biosystems). 13. Model 477A/120A pulsed-liquid sequencing system (Applied Biosystems). 14. SpectraMax340 Plate Reader (Molecular Devices). 15. 96-well Terasaki plates (Robbins Scientific: Mountain View, CA). 16. Prism 6 (Graphpad Software Inc).

Purification and Characterization of F-Type Lectins

3

247

Methods

3.1 Preparation of Liver Tissue Extracts

1. Mince the fresh tissue on a pan chilled on ice and liquefy with 200 mL of chilled extraction buffer in a chilled Waring blender for 1 min. 2. Centrifuge the liquefied mixture at 12,000  g for 15 min at 4  C and collect supernatant. 3. Pass the supernatant through cheesecloth and stored on wet ice. 4. Resuspend the pellet with 200 mL of chilled extraction buffer, repeat twice and pool with the first supernatant. 5. Keep the homogenate on wet ice for 1 h. 6. Remove the floating coalesced lipid droplets with a pipette. 7. Centrifuge the homogenate at 14,000  g for 1 h at 4  C.

3.2 Preparation of Serum

1. Collect peripheral blood (~10 mL/kg) from the caudal vein using an 18-gauge hypodermic needle and allow to retract overnight at 4  C. 2. Centrifuge to separate serum from the clot. 3. Dialyze the serum samples against 1000 volumes of TBS-Ca with four changes in 24 h at 4  C. 4. Dilute the dialyzed serum 1:1 in TBS-Ca for purification.

3.3 Expression of Recombinant FTL Proteins

Recombinant protein expression is carried out as reported earlier [16, 17]. 1. Grow the FTL-construct-transformed Escherichia coli BL21 (DE3) overnight at 37  C in LB medium containing 30 μg/mL kanamycin. 2. Dilute the bacteria in 3 L of fresh LB medium containing 30 μg/mL kanamycin to OD650 ¼ 0.1 and incubate at 37  C for 2–3 h until reaching OD650 ¼ 0.4–0.6. 3. Apply the 0.1 M IPTG to reach final concentration of 0.2 to 2 mM to induce the recombinant protein expression at 37  C for 3 h. 4. Collect the bacteria culture and spin for 20 min. 5. Remove the culture supernatant and resuspend the bacteria pellet in 1 BugBuster lysis buffer. 6. Sonicate the suspension on ice for 15 s three times. 7. Centrifuge 12,000  g for 20 min at 4  C. 8. Collect the clear supernatant, which contains most of the recombinant proteins, for purification.

248

Chiguang Feng and Gerardo R. Vasta

3.4 FTL Isolation by Affinity Chromatography

All steps are performed at 4  C as described previously [7, 11]. 1. Pass the tissue homogenate, serum sample, or bacteria clear lysate, over a divinyl sulfone-conjugated L-fucose-Sepharose6B column [15] at 0.4 mL/min and monitor at 280 nm with an UviCord II detector. 2. Wash the column to base line with TBS-Ca. 3. Elute the bound FTLs with TBS, 30 mM EDTA, until to base line again. 4. Finally elute with 200 mM L-fucose in TBS-Ca. 5. Pool fractions (0.5 mL) containing protein peaks. 6. Concentrate by ultrafiltration in a Centricon 10 to a final concentration of ~10 mg/mL. 7. Analyze eluted protein(s) in SDS-PAGE with Coomassie Brilliant Blue (CBB) stain (Fig. 1a).

3.5 FTL Purification by Size Exclusion Chromatography (SEC)

1. Connect a Superose 12 10/30 HR column to an HPLC system, or appropriate SEC column to an FPLC system. 2. Pre-equilibrate with TBS-Ca. 3. Load the affinity eluate to HPLC at room temperature at a flow rate of 0.4 mL/min. 4. Monitor the elution at 280 nm. 5. Analyze in SDS-PAGE for FTL purity with CBB and AgCl stains [18] (Fig. 1). 6. For natural protein, perform Edman sequence analysis or mass spec to determine partial peptide sequence [7, 11] (see Subheading 3.6). Verify the identity of FTL by BLAST over GenBank. 7. Store the purified lectin at 4  C (see Note 2).

3.6 Edman Sequence Analysis

1. Digest purified lectin (50 μg) with trypsin and lysyl endopeptidase in 0.05 M Tris–HCl (pH 8.5), 1 M guanidine hydrochloride (E/S ¼ 1:50, w/w, at 30  C, 10–20 h). 2. Acidify the digests with 10% TFA to pH~2. 3. Set up the microbore RP-HPLC system with Applied Biosystems model 140A pumps and model 1000S diode-array detector (2.3 mL flow cell, 0.0025 inch i.d. tubing), along with either a Zorbax-SB C-18 silica column (0.1  15 cm, ˚ pore size), or an Applied Biosystems Aquadp~5 mm, 300 A pore ODS-300 C-18 silica column (0.1  25 cm, dp~7 mm, ˚ pore size). 300 A 4. Equilibrate the system at room temperature in 0.1% aqueous TFA.

Purification and Characterization of F-Type Lectins

249

5. Apply the digested peptides and eluted at flow rates of 50–80 μL/min using linear gradients of acetonitrile/water/ TFA (80:20:0.1, v/v). 6. Monitor the column effluent at 215 nm, the UV absorption spectra of the absorbing material was determined. 7. Collect the eluate and stored at 20  C prior to further analysis. 8. If required, repeat chromatography of some fractions in the second RP-HPLC elution solvent system consisting of 2-propanol/acetonitrile/water/TFA (70:20:10:0.1, v/v) prior to sequencing. 9. Perform automated Edman degradation [19] of the separated peptides on an Applied Biosystems model 477A/120A pulsedliquid sequencing system. The PTH-amino acids are separated and identified as described previously [20]. 10. Perform the NH2-terminus sequencing with a sequencer, for example, Beckman LG3000 gas-phase sequencer (the Bioanalytical Laboratory of the Institute of Marine and Environmental Technology (IMET), Baltimore, Maryland) (see Note 3). 11. Repeat seven cycles to complete sequence of seven residues (Table 1). 12. Verify the identity of FTL by BLAST with annotated sequences at GenBank. 3.7 Assessment of the Native Molecular Size by SEC

Calibrate the Superose 12 10/30 HR column used in Subheading 3.5 for final purification, with low-molecular size SEC standards of known hydrodynamic radius following the procedure described [21] while using the same HPLC or FPLC system [2–4, 7]. 1. Connect a Superose 12 10/30 HR column to an HPLC system, or appropriate SEC column to an FPLC system. 2. Pre-equilibrate with TBS-Ca. 3. Inject 200 μL of each set separately. 4. Allow complete peak separation for each of the marker proteins. 5. Determine the column void volume (V0) and total liquid column volume (Vt) by injecting 200 μL of dextran blue 2000 solution (2 mg/mL). 6. Calculate a distribution coefficient (KD) as: KD ¼ (Vr-V0)/ (Vt-V0), where Vr is the elution volume of a peak. 7. Prepare a plot of KD vs log10MW. 8. Make a third-order polynomial curve fit with Prism 6. 9. Interpolate molecular weight for the experimental sample (see Note 4).

250

Chiguang Feng and Gerardo R. Vasta

3.8 ELISA for Analysis of Carbohydrate-Binding Activity and Specificity

Solid phase ELISA binding or inhibition assay are carried out as reported earlier [16, 17]. 1. Dilute neoglycoproteins, authentic glycoproteins (untreated or glycosidase-treated), or cellular extracts as binding ligands to appropriated concentration (1–50 μg/mL) in PBS. 2. Apply 100 μL of solution into 96-well microtiter plates each well in triplicates. 3. Seal and incubate the plates for 3 h at 37  C. 4. Wash the plates with 0.1% Tween in PBS (PBST) three times. 5. Apply 200 μL of 3% BSA in PBS, seal and block overnight at 4  C. 6. Wash the plates three times with cold PBS and store at 4  C until use. 7. FTL binding activity and carbohydrate specificity are assessed by either of the following procedures: (a) For FTL binding activity assays, serially dilute the natural or recombinant FTL proteins, dispense into the glycancoated plates 100 μL per well, and incubate for 1 h at 4  C. (b) For FTL binding-inhibition assays, preincubate the natural or recombinant FTL proteins in PBS with serial dilutions of test sugars (mono-, di-, and oligosaccharides; polysaccharides), glycoproteins (PSM, asialo-PSM, OSM, asialo-OSM, BSM, or asialo-BSM), or neoglycoproteins for 1 h at 4  C. Then deliver the mixtures into the coated plates and incubate for 1 h at 4  C. 8. Wash the plates three times with PBS (see Note 5). 9. Apply purified immunoglobulins (see Note 6) from rabbit antiFTL antiserum in 1% BSA/PBS and incubate for 1 h at room temperature. 10. Wash the plates four times with PBST, apply HRP-conjugated anti-rabbit antibodies, and incubate for 1 h at room temperature. 11. Wash the plate six times with PBST, apply 100 μL of TMB microwell Peroxidase Substrate, and incubate for 5–10 min. 12. Add 100 μL of 1 M HCl each well to stop the reaction. 13. Measure the optical densities (OD) at 450 nm on SpectraMax340 Plate Reader controlled by SoftmaxPro software, version 1. 14. Export the data and analyze at Excel or Prism GraphPad.

Purification and Characterization of F-Type Lectins

3.9 Agglutination Assay

251

Microhemagglutination tests are carried out in BSA-blocked, 96well Terasaki plates as reported earlier [2–4, 7]. 1. Block the plates overnight at 4  C in 1% (w/v) BSA in physiological saline (0.85% NaCl, 0.2% sodium azide). 2. Wash 50 μL of packed RBCs four times at 800  g for 2 min in physiological saline. 3. Resuspend in 50 μL of 0.1% Pronase and incubate for 20 min at 37  C. 4. Wash the cells four times with physiological saline and once with TBS-Ca. 5. Resuspend the treated RBC (Pr-RBC) in assay buffer, counted with a hemocytometer, and resuspended at 5  106 cells/mL in assay buffer. 6. Mix 5 μl of cell suspension per well with equal volumes of two-fold serial dilutions of MsaFBP32 in TBS-Ca in 96-well Terasaki plates. 7. Gently vortex-mix the plates for 10 s and incubate at room temperature for 1 h. 8. Read agglutination under a microscope, and score from 0 (negative) to 4 (all RBC in a single or a few large clumps). 9. Record the reciprocal of the highest dilution of lectin showing an agglutination score of +1/2 as the titer. The lectin specific activity is defined as titer/mg of protein/mL.

3.10 AgglutinationInhibition Assay

Carbohydrate binding specificity is analyzed with the use of the microhemagglutination assay described above [2–4, 7]. 1. Prepare the lectin solution with a titer of two agglutination units in TBS-Ca. 2. Dissolve the carbohydrates to be tested as inhibitors in TBS at concentrations up to 200 mM for mono- and oligosaccharides and 10 mg/mL for polysaccharides and glycoproteins and brought to pH 7.6 with concentrated NaOH. 3. Prepare the two-fold serial dilutions of the inhibitors in the same buffer (see Note 7). 4. Mix 5 μL of ligand dilution with 5 μL of lectin in 96-well Terasaki plates. 5. Incubate the mixtures for 45 min. 6. Apply 2 μL of the Pr-RBCs suspension (107 cells/mL) to each well (see Note 8). 7. Vortex-mix for 10 s and incubate at room temperature for 1 h. 8. Read agglutination under a microscope, and score from 0 (negative) to 4.

252

Chiguang Feng and Gerardo R. Vasta

Table 2 Hemagglutination-inhibition profile of MsaFBP32 by monosaccharides, oligosaccharides, and glycoproteins 50% inhibition value

Inhibitora

MsaFBP32 specificity AAA specificity factorb factorc

μM Fucα6GlcNAc

0.5

2.60

Methyl α-L-Fuc

0.7

1.86

4

L-Fuc

1.3

1.00

1

2 -Fucosyllactose (Fucα2Galβ4Glc)

2.5

0.52

2

3-Fucosyllactose (Fucα3[Galβ4]Glc)

3

0.43

0.53

H-disaccharide (Fucα2Gal)

3.6

0.36

80

ρ-Nitrophenyl-α-L-Fuc

3.6

0.36

ND

L-Gal

4

0.33

0.064

D-Fucosylamine

4

0.33

ND

D-Fuc

5

0.26

NI

ρ-Nitrophenyl-β-L-Fuc

8

0.16

ND

Methyl β-L-Fuc

9

0.14

0.25

ρ-Nitrophenyl-β-D-Fuc

16

0.08

ND

D-Man

20

0.07

D-Gal

25

0.05

35

0.04

D-glucosamine

100

0.01

D-arabinose

ND

Inhibitora

50% inhibition value

0

D-mannosamine

HCl

0.032

MsaFBP32 specificity AAA specificity factorb factorc

mg/mL

a

Porcine stomach mucin

0.08

1

NA

asialo-PSM

0.08

1

NA

asialo-BSM

10

0.0080

NA

Bovine submaxillary mucin

40

0.0020

NA

Bovine lung galactan

55

0.0015

NA

Hyaluronic acid

160

0.0005

NA

Helix pomatia galactan

400

0.0002

NA

Concentration of carbohydrates and glycoproteins required to inhibit 50% of the agglutinating activity as scored in assay are shown. Pronase-treated human A red blood cells were used as test cells b Calculated in relation to L-fucose and porcine stomach mucin, respectively c Data are from Ref. [23]

Purification and Characterization of F-Type Lectins

253

9. Interpolate 50% inhibition values from curve fits of plots of percent inhibition versus inhibitor concentration (Table 2; Fig. 4). 3.11

Glycan Array

Glycan array analysis is carried out at the National Center for Functional Glycomics (NCFG) at Harvard University (https:// ncfg.hms.harvard.edu/), on version 5.0 of the array printed with 611 glycans in replicates of six [16, 17]. 1. Dialyze the natural or recombinant protein against TBS containing 10 mM Ca, with three buffer exchanges. 2. FTL samples are submitted to NCFG for glycan array analysis.

Fig. 4 Hemagglutination-inhibition curves for (a) monosaccharides, (b) polysaccharides, and (c) glycoproteins

254

Chiguang Feng and Gerardo R. Vasta

3. Results from three different protein concentrations (usually, 2, 10, 50, or 150 μg/mL) are processed to reveal the recognition pattern of the FTL of interest.

4

Notes 1. To yield peaks of similar absorbance at 280 nm. 2. Purified proteins are further prepared for crystallization for structure determination, and computational modeling and docking experiments with various carbohydrate ligands. 3. Any equivalent systems can be used. 4. The estimate of molecular weight assumes that the experimental protein is globular and behaves hydrodynamically similar to the calibration markers. 5. If the FTL has relatively low affinity/avidity for the immobilized glycoprotein ligand, it may be useful to fix the bound FTL after washing off the unbound FTL, following the additional steps, before adding the specific anti-FTL antibody: (a) Apply 100 μL of 2% formaldehyde in PBS each well and incubate for 30 min at room temperature. (b) Apply each well with 100 mM glycine in PBS and incubate for 5 min. (c) Repeat one more time with glycine incubation. (d) Wash the plate three times with PBST. 6. Antiserum production was carried out by immunizing rabbit with authentic or recombinant protein and immunoglobulin G was purified with Protein A agarose column as described previously [7, 16]. 7. Inhibition by mono- and oligosaccharides is tested at concentrations from 0.1 to 200 mM. Inhibition by glycoproteins, asialoglycoproteins, and polysaccharides is tested at concentrations up to 10 mg/mL (Table 2). 8. Use the substitutions of the inhibitor solution by TBS and substitution of purified lectin by TBS-Ca as negative controls.

Acknowledgments The author’s research reviewed herein was supported by Grants IOS 1050518, IOB-0618409, MCB 0077928, and IOS-0822257 from the National Science Foundation, and Grant R01GM070589 from the National Institutes of Health to G.R.V.

Purification and Characterization of F-Type Lectins

255

References 1. Vasta GR, Ahmed H (2008) Animal Lectins: a functional view. Taylor & Francis, CRC Press, Boca Raton, p 558 2. Vasta GR, Hunt JC, Marchalonis JJ et al (1986) Galactosyl-binding lectins from the tunicate Didemnum candidum. Purification and physicochemical characterization. J Biol Chem 261:9174–9181 3. Vasta GR, Marchalonis JJ (1986) Galactosylbinding lectins from the tunicate Didemnum candidum. Carbohydrate specificity and characterization of the combining site. J Biol Chem 261:9182–9186 4. Ahmed H, Fink NE, Pohl J et al (1996) Galectin-1 from bovine spleen: biochemical characterization, carbohydrate specificity and tissue-specific isoform profiles. J Biol Chem 120:1007–1019 5. Ahmed H, Du SJ, O’leary N et al (2004) Biochemical and molecular characterization of galectins from zebrafish (Danio rerio): notochord-specific expression of a prototype galectin during early embryogenesis. Glycobiology 14:219–232 6. Bianchet MA, Odom EW, Vasta GR et al (2002) A novel fucose recognition fold involved in innate immunity. Nat Struct Biol 9:628–634 7. Odom EW, Vasta GR (2006) Characterization of a binary tandem domain F-type lectin from striped bass (Morone saxatilis). J Biol Chem 281:1698–1713 8. Tasumi S, Vasta GR (2007) A galectin of unique domain organization from hemocytes of the eastern oyster (Crassostrea virginica) is a receptor for the protistan parasite Perkinsus marinus. J Immunol 179:3086–3098 9. Latha VL, Rao RN, Nadimpalli SK (2006) Affinity purification, physicochemical and immunological characterization of a galactosespecific lectin from the seeds of Dolichos lablab (Indian lablab beans). Protein Expr Purif 45:296–306 10. Mo H, Meah Y, Moore JG et al (1999) Purification and characterization of Dolichos lablab lectin. Glycobiology 9:173–179 11. Odom EW (2004). F-type Lectins: biochemical, genetic and structural characterization of a novel Lectin family in lower vertebrates. Ph.D.

thesis. Baltimore, MD: University of Maryland, MEES Program 12. Vasta GR, Ahmed H, Bianchet MA et al (2012) Diversity in recognition of glycans by F-type lectins and galectins: molecular, structural, and biophysical aspects. Ann N Y Acad Sci 1253:E14–E26 13. Vasta GR, Amzel LM, Bianchet MA et al (2017) F-type Lectins: a highly diversified family of Fucose-binding proteins with a unique sequence motif and structural fold, involved in self/non-self-recognition. Front Immunol 8:1648 14. Tennent GA, Pepys MB (1994) Glycobiology of the pentraxins. Biochem Soc Trans 22:74–79 15. Hermanson GT, Mallia AK, Smith PK (1992) Immobilized affinity ligand techniques. Academic Press, San Diego 16. Feng C, Ghosh A, Amin MN et al (2015) Galectin CvGal2 from the eastern oyster (Crassostrea virginica) displays unique specificity for ABH blood group oligosaccharides and differentially recognizes sympatric Perkinsus species. Biochemistry 54:4711–4730 17. Feng C, Ghosh A, Amin MN et al (2013) The galectin CvGal1 from the eastern oyster (Crassostrea virginica) binds to blood group a oligosaccharides on the hemocyte surface. J Biol Chem 288:24394–24409 18. Chevallet M, Luche S, Rabilloud T (2006) Silver staining of proteins in polyacrylamide gels. Nat Protoc 1:1852–1858 19. Hewick RM, Hunkapiller MW, Hood LE et al (1981) A gas-liquid solid phase peptide and protein sequenator. J Biol Chem 256:7990–7997 20. Pohl J (1994) Sequence analysis of peptide resins from Boc/benzyl solid-phase synthesis. Methods Mol Biol 36:107–129 21. Coligan JE (1995) Current protocols in protein science. Wiley, Brooklyn, NY 22. Nielsen H, Engelbrecht J, Brunak S et al (1997) Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Eng 10:1–6 23. Watkins WM, Morgan WT (1952) Neutralization of the anti-H agglutinin in eel serum by simple sugars. Nature 169:825–826

Chapter 25 LecA (PA-IL): A Galactose-Binding Lectin from Pseudomonas aeruginosa Sakonwan Kuhaudomlarp, Emilie Gillon, Annabelle Varrot, and Anne Imberty Abstract LecA/PA-IL (Pfam PF07828) is a soluble galactose-binding lectin from bacterium Pseudomonas aeruginosa. The lectin is specific for α-galactose present on glycosphingolipids of the globoside family and has therefore been proposed to play a role in cell adhesion and in internalization of bacteria in epithelial cells. The lectin has also direct toxic activity. Search for high-affinity inhibitors can be performed on the recombinant lectin, with use of surface plasmon resonance assays and structural studies. Key words Lectins, Protein-Carbohydrate Interactions, X-ray crystallography, Surface Plasmon Resonance, Pseudomonas aeruginosa

1

Introduction Pseudomonas aeruginosa is a Gram-negative bacterium and a leading pathogen responsible for acute as well as chronic infections of immune-compromised patients and patients suffering from cystic fibrosis [1]. Among other virulent factors, P. aeruginosa produces LecA, a calcium-dependent galactose-specific soluble lectin, also referred to as PA-IL [2], the production of which is regulated by quorum sensing [3]. LecA was initially identified and characterized in the cytoplasm of P. aeruginosa but large quantities are present on the outer membrane of the bacterium, suggesting it may play a role in adhesion [4]. LecA is involved in biofilm development [5]. It also has strong cytotoxic effect and has been demonstrated to induce permeability effect in epithelial cells, resulting in injury in lungs [6] and intestines [7]. The lectin is specific for globosides, that is, glycosphingolipids terminated by an α-galactose (α-Gal) residue. The interaction with cell membrane glycolipids results in signaling through protein phosphorylation [8] and bacteria intracellular invasion through membrane invagination [9].

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_25, © Springer Science+Business Media, LLC, part of Springer Nature 2020

257

258

Sakonwan Kuhaudomlarp et al.

Fig. 1 Crystal structure of LecA in complex with Gal. (a) The overall structure of LecA tetramer in complex with Gal. The protein backbone of a tetrameric LecA is shown in cartoon representation (yellow), Gal is shown as gray cylinders and calcium ion as green spheres. (b) The Gal binding site in a LecA monomeric subunit. Hydrogen bonds are shown in black dash lines and contacts with calcium ion in red dash lines. Calcium is represented as a light green sphere and Gal in gray cylinder

The lectin LecA (51 kDa) is composed of four subunits of 121 amino acids. It binds to α-Gal monosaccharide and terminal α-Gal on oligosaccharides with moderate μM affinity [10]. The presence of aromatic group on the sugar anomeric position, either in α or β configurations, yields to higher affinity, with KD of 5 μM for β-galactosides through CH-π interaction between His50 and the aromatic ring of the galactoside aglycone [11]. The first X-ray structure of LecA in complex with α-Gal and calcium was reported in 2003 (Fig. 1) [12], which has since provided a solid framework for the successive development of inhibitors against LecA and subsequently P. aeruginosa infections, in particular, the development of aromatic glycoside inhibitors [11, 13], multivalent galactoside inhibitors [14–19], and a covalent inhibitor targeting the carbohydrate-binding site in LecA [20]. Moreover, a structure of LecA in complex with melibiose (αGal1-6Glc) unraveled a secondary binding site that could be used as an additional target for inhibitor design [21].

2 2.1

Materials Gene Sequence

2.2 Protein Production in Escherichia coli

GenBank ID:NP_251260.1|PA-I [P. aeruginosa PAO1].

galactophilic

lectin

1. Escherichia coli BL21 (DE3) competent cells. 2. Lysogeny broth (LB) (Invitrogen) and LB agar containing 100 μg/mL ampicillin. 3. Isopropyl β-D-1-thiogalactopyranoside (IPTG).

LecA (PA-IL): A Galactose-Binding Lectin from Pseudomonas aeruginosa

2.3 Protein Purification

259

1. Pierce™ D-galactose agarose column (ThermoFisher). 2. NGC™ Chromatography System (BIO-RAD) with ChromLab software. 3. One Shot Cell disruptor (Constant Systems Ltd., UK). 4. Buffer A (20 mM Tris–HCl pH 7.5, 100 mM NaCl, and 100 μM CaCl2) (see Note 1). 5. Buffer B (20 mM Tris–HCl pH 7.5, 100 mM NaCl, and 100 mM EDTA).

2.4 Surface Plasmon Resonance

1. BIACORE™ X100 SPR system (GE Healthcare). 2. BIACORE sensor chip CM7 (GE Healthcare). 3. 1 PBS-P+ (10 mM phosphate buffer, 2.7 mM KCl, 0.137 M NaCl, 0.05% Tween-20, 0.1 mM CaCl2). 4. 10 mM acetate buffer pH 4.5. 5. 1 PBS buffer containing 100 μM CaCl2. 6. 4-nitrophenyl β-D-galactopyranoside 301.25, Sigma-Aldrich).

(β-pNP-Gal,

MW

7. Amine Coupling Kit (GE Healthcare) containing 1-ethyl-3(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC) (750 mg), N-hydroxysucci nimide (NHS) (115 mg), and ethanolamine-HCl pH 8.5 (1 M, 10.5 ml) in separate vials. 2.5 Protein Crystallization

1. 24 or 48 wells VDX Plate with sealant 24 well (Hampton Research) or SuperClear™, pregreased 24 well plates (Jena Bioscience). 2. Siliconized glass circle cover slides (22 mm, Hampton research). 3. Crystallization solution for LecA in complex with Gal (1.5 M ammonium sulfate, 100 mM sodium acetate pH 4.7, 5% 2-methyl-2,4-pentanediol (MPD), and 2% glycerol) (see Note 2). 4. D-galactose (Gal, MW 180.16, Sigma-Aldrich).

3 3.1

Methods LecA Production

3.1.1 Transformation and Preculture

1. The gene encoding LecA was obtained from template DNA from Pseudomonas aeruginosa ATCC 15692. The gene was successfully ligated into an expression vector pET25b to create a recombinant plasmid called pET25b-PAIL, which is available upon request. 2. 1 μL of the pET25b-PAIL plasmid was mixed with 100 μL of competent BL21(DE3) cells, followed by incubation on ice for 10 min.

260

Sakonwan Kuhaudomlarp et al.

3. The plasmid was transformed into the competent cells by heat shock at 42  for 45 s, then rest on ice for 3 min. 4. The transformed cells were inoculated in 900 μL liquid LB and incubated for 1 h at 37  C with agitation at 180 rpm. 5. 100 μL of the liquid culture was spread on LB agar containing 100 μg/mL ampicillin and incubated at 37  C overnight to select for cells containing the recombinant plasmid. 6. In order to prepare precultures, cell colonies containing the recombinant plasmid were inoculated in 10 mL liquid LB containing 100 μg/mL ampicillin and incubated at 37  C overnight with agitation at 180 rpm. 3.1.2 Culture

1. 10 mL of the preculture was transferred into 1 L of liquid LB containing 100 μg/mL ampicillin. 2. The 1-L culture was incubated at 37  C with agitation at 180 rpm until OD600 of 0.6 was reached. 3. The expression of LecA was induced by adding 0.5 mM IPTG to the culture, then further incubated at 30  C for 3 h with agitation at 180 rpm. 4. The cells were harvested by centrifugation (6000  g, 4  C, 10 min), then resuspended in liquid LB and centrifuged again in 50 mL falcons at 6000  g for 10 min. 5. The pellets were stored at 20  C until required.

3.1.3 LecA Purification by Affinity Chromatography

1. The cell pellet from 1 L of culture was resuspended in 30 mL of Buffer A supplemented with 1 μL of Denarase 100KU and incubated at room temperature for 10 min. 2. The cells were lysed by a cell disruptor at 1.9 kbar. 3. The lysate was centrifuged (24,000  g, 30 min, 4  C). The supernatant was filtered with 0.45 μm syringe filter. 4. D-Gal agarose 5-mL column was pre-equilibrated with Buffer A (3 column volumes, 3 mL/min). 5. The filtered supernatant containing LecA was injected onto the pre-equilibrated column at low flow to promote interaction of LecA with Gal on the column. 6. The column was washed with Buffer A to remove unbound proteins (approximately 3 CVs or until Abs280 reached 0). 7. Bound LecA protein was eluted with 100% Buffer B and collected as 5-mL fractions. 8. The column was rinsed with milliQ water +0.02% sodium azide and stored in 20% ethanol.

LecA (PA-IL): A Galactose-Binding Lectin from Pseudomonas aeruginosa

261

9. The collected protein fractions were combined and dialyzed in 5 L of milliQ water for 3 days at 4  C while changing the water twice per day. The dialyzed protein was lyophilized and stored at 20  C until required (see Note 3). 3.2 Surface Plasmon Resonance (SPR)

3.2.1 Sample Preparation

SPR is a label-free analytical technique for detecting intermolecular interactions between an analyte in solution and an immobilized ligand on the SPR sensor surface. The method is based on the change in the refractive index of the biosensor surface, which occurs when the analyte interacts with the immobilized ligand. SPR has a broad range of sensitivity detection (nM-mM) and requires relatively small quantity of samples per analysis. SPR has been used for analysis of various protein-carbohydrate interactions [22]. SPR experiments enable thermodynamics (i.e., dissociation constant, KD) and/or kinetics parameters (kon and koff) to be determined. We have previously demonstrated the use of SPR for characterization of LecA and their prospective glycomimetics compounds [16, 17, 23]. Herein, we provided a methodology for preparation of immobilized LecA biosensor surface and an example of SPR analysis of an interaction between immobilized LecA and a glycomimetic compound. 1. Prepare LecA solution from the lyophilized protein powder (approx. 2 mg) from Subheading 3.1.3 in 500 μL of 1 PBS buffer containing 100 μM CaCl2. 2. Centrifuge the solution at 7000  g (5 min) to remove precipitated protein. 3. Determine the protein concentration in the supernatant in mg/mL using absorbance at 280 nm (absorbance coefficient of LecA is 2.191). The concentration should be at least 1 mg/ mL. From this stock solution, prepare 500 μL of 100 μg/mL LecA in 10 mM acetate buffer pH 4.5 (see Note 4). 4. Prepare 100 μL of 0.1 mM solution of β-pNP-Gal in 1 PBS-P + (see Note 5). 5. Add 10 mL filtered milliQ water into EDC and NHS vials from GE Healthcare to dissolve the compounds. Store them separately at 20  C in Eppendorf tubes as 100 μL aliquots until required.

3.2.2 LecA Immobilization on SPR Chip Using Amine Coupling Method

1. Pre-equilibrate BIACORE X100 SPR system with 1 PBS-P+ and dock a new CM7 chip. Set up a manual run on BIACORE X100 Control software and the temperature of the system at 25  C. 1 PBS-P+ is used as a running buffer for chip activation, LecA immobilization, and subsequent analyses. 2. Mix thoroughly an equal volume of EDC and NHS in an SPR tube (a minimum volume required for one injection is 130 μL). Ensure that there is no air bubble formed in the solution (see Note 6).

262

Sakonwan Kuhaudomlarp et al.

3. Inject the EDC/NHS mixture at 10 μL/min with a contact time of 540 s to activate the docked CM7 chip (both on flow cell 1 and 2). Follow the change in the response on sensorgram during the injection. Repeat this step until the response relative to the preinjection baseline is 100–400 resonance units (RU). 4. Fix LecA on the activated surface on flow cell 2 only by injecting 100 μg/mL LecA in 10 mM acetate buffer pH 4.5 at 10 μL/min with a contact time of 540 s. Repeat this step until the response relative to the baseline reaches ~10,000 RU. 5. Inject 1 M ethanolamine-HCl pH 8.5 at 10 μL/min with a contact time of 540 s on flow cell 1 and 2 to deactivate remaining active groups and remove noncovalently bound LecA. Flow cell 1 is referred to as a reference flow cell (no LecA) and flow cell 2 as an active flow cell (containing LecA). 6. β-pNP-Gal was used as a positive control to monitor the activity of immobilized LecA on the CM5 chip prior to analysis of analytes of interest. KD of β-pNP-Gal was previously reported as 26.1 μM [13], which is sufficiently good for its use as a positive control. Figure 2 shows a typical sensorgram response after an injection of 0.1 mM β-pNP-Gal at 30 μL/min flow rate with a contact time of 30 s and dissociation time of 60 s.

Fig. 2 A typical sensorgram for β-pNP-Gal. The analyte was injected at the flow rate of 30 μl/min with a contact time of 30 s and dissociation time of 60 s. The report point for the binding response was automatically detected by BIACORE X100 control software when the equilibrium is reached (in this case, at 88 s as indicated by a cross). The spikes indicate the buffer mismatches caused by the analyte

LecA (PA-IL): A Galactose-Binding Lectin from Pseudomonas aeruginosa

263

Fig. 3 A multicycle kinetic experiment for compound 1 for determination of KD. (a) Sensorgrams of multicycle injections of a range of concentrations of compound 1. (b) The binding responses that were determined from the sensorgrams in (a) were plotted against the compound concentrations. KD is determined as the concentration at which half of the maximum response is reached

7. For determination of kinetic or affinity parameters, multiple injections of a range of concentrations of an analyte of interest are required. Figure 3a represents an example of multiple cycle kinetic experiment of compound 1 (60 s contact time and 60 s dissociation time, flow rate ¼ 30 μL/min). The sensorgram for compound 1 has a square-like shape, indicating very fast association and dissociation of the compound from immobilized LecA, which hindered the determination of the kinetic parameters. However, the binding response and KD could be determined by plotting the binding responses at equilibrium against the corresponding analyte concentrations (Fig. 3b) (see Note 7). 3.3 LecA Crystallization

3.3.1 Sample Preparation

We demonstrated below a detailed methodology for co-crystallization of LecA with Gal following our previously published protocol from Cioci et al. [12], which reported the first structure of LecA in complex with Gal (PDB code 1OKO) (Fig. 1). 1. Prepare LecA solution from the lyophilized protein powder (approx. 2 mg) from Subheading 3.1.3 in milliQ water supplemented with 0.1 mM CaCl2 (see Note 8). 2. Centrifuge the solution at 7000  g (5 min) to remove precipitated protein. 3. Determine the protein concentration in the supernatant in mg/mL using absorbance at 280 nm (absorbance coefficient of LecA is 2.191). The final concentration should be around 10 mg/mL.

264

Sakonwan Kuhaudomlarp et al.

4. Prepare a 50 mM solution of Galactose in milliQ water. 5. Prepare crystallization solution containing 1.5 M ammonium sulphate, 100 mM sodium acetate pH 4.7, 5% 2-methyl-2,4pentanediol (MPD), and 2% glycerol (1 mL per well). 3.3.2 Hanging Drop Vapour Diffusion

1. Add 1 μL of 50 mM galactose to 24 μL of LecA at 10 mg/mL and incubate for 1 h at room temperature (see Note 9). 2. Aliquots 1 mL of the crystallization solution into a 24 wells crystallization plate. 3. Mix 1 μL of protein-ligand solution (1 in Subheading 3.3.2) with 1 μL of the crystallization solution (5 in Subheading 3.3.1) on a 22-mm cover slip to create a crystallization drop. 4. Close the well containing crystallization solution with the cover slip containing the LecA-Gal drop. Ensure that the drop is hanging inside the well but not in contact with the grease. Press gently on the cover slip to ensure good contact between the cover slip and the grease on top of the well. 5. Close the plate lid and store the plate at 19  C. Monitor the formation of protein crystals every 2–3 days until the crystals are formed.

4

Notes 1. LecA is a calcium-dependent lectin and it is advised to maintain a CaCl2 concentration of 100 μM throughout the purification and the SPR experiment. If needed, this concentration can be decreased to 10 μM. 2. LecA can also be crystallized using PEG as a precipitant. For example, the crystallization conditions for LecA in complex with melibiose (αGal1-6Glc) is 15% PEG 5000 MME, 100 mM KSCN, and 100 mM sodium acetate pH 4.6 [21]. The favorite crystallization pH is 4.6 but some crystals can also be obtained at basic pH with 100 mM Tris 8.5 [18]. 3. LecA is a very robust protein that can be lyophilized and stored at 20  C for months or years. It is only recommended to aliquot in small quantities to avoid defreezing and refreezing. 4. Preparation of LecA in acetate buffer pH 4.5 is required to ensure that the protein is positively charged (theoretical pI of LecA is 5.04), allowing electrostatic interaction with the negatively charge on the chip surface, which provides efficient means for concentrating positively charged protein on the surface. However, LecA has limited stability in the acetate buffer and therefore the protein solution should be prepared as shortly as possible before injection.

LecA (PA-IL): A Galactose-Binding Lectin from Pseudomonas aeruginosa

265

5. For a perfect SPR measurement, the main recommendation is to avoid buffer mismatch by using the running buffer to prepare the analyte solution. It is strongly advised not to use organic buffering agent such as HEPES during chip activation and LecA fixation because the organic agent could react with the activated surface. 6. NHS/EDC mixture should be mixed shortly before injection. 7. In order to compensate for bulk refractive index differences between flow buffer and analyte sample as well as some non-specific binding to the sensorchip surface, the BIACORE system automatically performs reference subtraction, where the absolute response of the active flow cell (flow cell containing immobilized LecA) is subtracted by that recorded on the reference flow cell (the flow cell without LecA). During data analysis such as kinetic and affinity determination, double referencing is also performed, where reference subtraction, as well as the subtraction of blank injections (zero analyte concentrations) are performed. 8. LecA can also be prepared in 20 mM Tris–HCl or HEPESNaOH pH 7.5 supplemented with 0.1 mM CaCl2 for ligands that to be in basic solution. 9. LecA ligand can also be added to the crystallization solutions instead of the protein solution but it requires much more ligand and is not recommended for precious ligand.

Acknowledgments The authors acknowledge support by the ANR PIA Glyco@Alps (ANR-15-IDEX-02), Labex ARCANE and CBH-EUR-GS (ANR-17-EURE-0003), and the French Cystic Fibrosis Association Vaincre la Mucoviscidose. References 1. Lyczak JB, Cannon L, Pier GB (2002) Lung infection associated cystic fibrosis. Clin Microbiol Rev 15:194–222 2. Gilboa-Garber N (1982) Pseudomonas aeruginosa Lectins. Methods Enzymol 83:378–385 3. Winzer K, Falconer C, Garber N et al (2000) The Pseudomonas aeruginosa Lectins PA-IL and PA-IIL are controlled by quorum sensing and by RpoS. J Bacteriol 182:6401–6411 4. Glick J, Garber N (2009) The intracellularl localization of Pseudomonas aeruginosa lectins. Microbiology 129:3085–3090

5. Diggle SP, Stacey RE, Dodd C et al (2006) The galactophilic lectin, LecA, contributes to biofilm development in Pseudomonas aeruginosa. Environ Microbiol 8:1095–1104 6. Chemani C, Imberty A, De Bentzmann S et al (2009) Role of LecA and LecB lectins in Pseudomonas aeruginosa-induced lung injury and effect of carbohydrate ligands. Infect Immun 77:2065–2075 7. Laughlin R, Musch M, Hollbrook C et al (2000) The key role of Pseudomonas aeruginosa PA-I lectin on experimental gut-derived sepsis. Ann Surg 232:133–142

266

Sakonwan Kuhaudomlarp et al.

8. Zheng S, Eierhoff T, Aigal S et al (2017) The Pseudomonas aeruginosa lectin LecA triggers host cell signalling by glycosphingolipiddependent phosphorylation of the adaptor protein CrkII. Biochim Biophys Acta, Mol Cell Res 1864:1236–1245 9. Eierhoff T, Bastian B, Thuenauer R et al (2014) A lipid zipper triggers bacterial invasion. Proc Natl Acad Sci 111:12895–12900 10. Blanchard B, Nurisso A, Hollville E et al (2008) Structural basis of the preferential binding for globo-series glycosphingolipids displayed by Pseudomonas aeruginosa lectin I. J Mol Biol 383:837–853 11. Kadam R, Garg D, Schwartz J et al (2013) CHπ “T-shape” interaction with Histidine explains binding of aromatic Galactosides to Pseudomonas aeruginosa Lectin LecA. ACS Chem Biol 8:1925–1930 12. Cioci G, Mitchell E, Gautier C et al (2003) Structural basis of calcium and galactose recognition by the lectin PA-IL of Pseudomonas aeruginosa. FEBS Lett 555:297–301 13. Rodrigue J, Ganne G, Blanchard B et al (2013) Aromatic thioglycoside inhibitors against the virulence factor LecA from Pseudomonas aeruginosa. Org Biomol Chem 11:6906 14. Kadam R, Bergmann M, Hurley M et al (2011) A glycopeptide dendrimer inhibitor of the galactose-specific lectin LecA and of Pseudomonas aeruginosa biofilms. Angew Chem Int Ed Engl 50:10631–10635 15. Visini R, Jin X, Bergmann M et al (2015) Structural insight into multivalent falactoside binding to Pseudomonas aeruginosa lectin LecA. ACS Chem Biol 10:2455–2462

16. Cecioni S, Lalor R, Blanchard B et al (2009) Achieving high affinity towards a bacterial Lectin through multivalent topological isomers of calix[4]arene Glycoconjugates. Chemistry 15:13232–13240 17. Cecioni S, Praly JP, Matthews S et al (2012) Rational design and synthesis of optimized Glycoclusters for multivalent Lectincarbohydrate interactions: influence of the linker arm. Chemistry 18:6250–6263 18. Novoa A, Eierhoff T, Topin J et al (2014) A LecA ligand identified from a galactosideconjugate array inhibits host cell invasion by Pseudomonas aeruginosa. Angew Chem Int Ed Engl 53:8885–8889 19. Bergmann M, Michaud G, Visini R et al (2015) Multivalency effects on Pseudomonas aeruginosa biofilm inhibition and dispersal by glycopeptide dendrimers targeting lectin LecA. Org Biomol Chem 14:138–148 20. Wagner S, Hauck D, Hoffmann M et al (2017) Covalent Lectin inhibition and application in bacterial biofilm imaging. Angew Chem Int Ed Engl 56:16559–16564 21. Blanchard B, Imberty A, Varrot A (2014) Secondary sugar binding site identified for LecA lectin from Pseudomonas aeruginosa. Proteins 82:1060–1065 22. Duverger E, Frison N, Roche AC et al (2003) Carbohydrate-lectin interactions assessed by surface plasmon resonance. Biochimie 85:167–179 23. Cecioni S, Faure S, Darbost U et al (2011) Selectivity among two lectins: probing the effect of topology, multivalency and flexibility of “clicked” multivalent glycoclusters. Chemistry 17:2146–2159

Chapter 26 Glycan Recognition and Application of P-Type Lectins Kei Kiriyama and Kohji Itoh Abstract Cation-dependent mannose 6-phosphate receptor (CD-MPR) and cation-independent MPR (CI-MPR) belong to the P-type lectin family. Both intracellular and cell surface MPRs can recognize and bind with the terminal mannose 6-phospahte (M6P) residues of N-glycans attached to the mammalian lysosomal enzymes and the related co-factors. Domain9 (Dom9), which is one of the extracytoplasmic region of the CI-MPR, has relatively higher affinity for M6P residues. Here we describe the production of recombinant Dom9-His protein by Pichia pastris, purification, and application as a probe for lectin blotting. Key words CI-MPR, M6P residue, Dom9, Yeast Pichia pastris, His-tag, Lectin blotting

1

Introduction There are two kinds of the P-type family of lectins: the 300-kDa cation-independent mannnose 6-phosphate receptor (CI-MPR) and the 46-kDa cation-dependent MPR (CD-MPR). Both MPRs have the physiological functions to deliver a class of lysosomal enzymes through binding with the terminal mannose 6-phosphate (M6P) residues-carrying N-glycans attached to the lysosomal enzymes (see Fig. 1) [1–3]. The MPRs are typeItransmembrane glycoproteins which consist of four regions: (1) an amino-terminal signal sequence, (2) an extracytoplasmic region which contains the ligand binding sites, (3) a transmembrane region, and (4) a carboxyl-terminal cytoplasmic region. The extracytoplasmic region of the CI-MPR consists of 15 tandemly repeating domains with an average size of 147 amino acids, whereas that of the CD-MPR consists of single domain [4– 6]. There is high sequence homology (16–38%) among the 15 repeating domains of the CI-MPR, and the homology is 14–– 28% between the 154-residue-containing extracytoplasmic domain of the CD-MPR and each of the 15 repeating domains of the CI-MPR [7].

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_26, © Springer Science+Business Media, LLC, part of Springer Nature 2020

267

268

Kei Kiriyama and Kohji Itoh

Rough ER

Lysosome

Golgi apparatus

Endosome P

P

P

P

P

P

P

P

Plasma membrane Newly synthesized lysosomal enzyme

Cell surface MPRs

Extracellular lysosomal enzyme

Intracellular MPRs

High-mannose-type oligosaccharide P

Mannose-6-phosphate

Fig. 1 The physiological functions of MPRs in target cells. Intracellular MPRs have the roles sorting newly synthesized lysosomal enzymes as shuttles to lysosomes while cell surface MPRs including the CI-MPRs take part in the endocytosis of extracellular lysosomal enzymes through binding with the terminal M6P residues-carrying N-glycans attached to the lysosomal enzymes. Both MPRs cycle between intracellular vesicles while only the CI-MPRs contribute to take up the extracellular lysosomal enzymes

In the extracytoplasmic region of the CI-MPR, the Domain1–3 (Dom1–3), Domain5 (Dom5), and Domain9 (Dom9) have relatively higher affinity for the M6P residues. The essential amino acid residues of the CD-MPR responsible for recognizing and binding with the M6P residues are four species (Gln-66, Arg-111, Glu-133 and Tyr-143) (see Fig. 2) [8], and the three domains (Dom1–3, Dom5 and Dom9) of the CI-MPR also have similar sequences corresponding to that of CD-MPR. In the Dom1–3 and Dom9, but not in the Dom5, there are two cysteine residues necessary for a disulfide bond formation in the M6P-binding pocket [5]. Thus, the binding affinity of Dom5 to the M6P residues is 300-fold lower than those of Dom1–3 and Dom9 [9]. The Dom1–3 has the higher affinity for the M6P residues, in which the structure of Dom3 bound for the M6P is supported by the Dom1 and 2. In contrast, the affinity of Dom3 alone is 1000-fold lower. On the other hand, it is reported that the Dom9 itself has high selectivity and affinity for M6P residues [10]. Recently the recombinant human Dom9-His fusion protein consisting of the Dom9 and His-tag attached to the carboxyl-terminus of Dom9 has been produced by methylotrophic yeasts Pichia pastris and Ogataea minuta [10, 11].

Application of Mannose 6-Phosphate Receptor Domains

269

Fig. 2 Amino acid sequence alignment of the domains contributing to bind with the M6P residues among the human CD-MPR and Dom3/Dom9 of the human CI-MPR. Identical amino acid residues are colored gray. The amino acids indicated in bold style are the essential amino acid residues for recognizing and binding with the M6P residues. The homology is 28.8% or 24.7% between the 66–185 residues extracytoplasmic domain of the CD-MPR and the 39–185 residues of the Dom3 or Dom9

Moreover, the cell surface CI-MPR can take up the extracellular lysosomal enzymes through binding with the terminal M6P-carrying glycans attached to the lysosomal enzymes into the cell and deliver to lysosomes as a shuttle [12, 13]. Enzyme replacement therapy (ERT) has been clinically applied to lysosomal storage diseases by utilizing the intralysosomal delivery of the therapeutic lysosomal enzymes. The administrated therapeutic lysosomal enzymes are also taken up by the cells expressing the cell surface CI-MPR to be transported into lysosomes and degrade accumulated substrates [14, 15]. Lectin blotting with the Dom9-His as a probe described above is utilized to qualify the presence of the M6P residues attached to the recombinant lysosomal enzymes. The Dom9-His also can be used to quantify the M6P contents. Because the CI-MPRs dimerize themselves on the cell surface to bind 2 mol of M6P per one enzyme molecule, the efficiency of intracellular uptake is dependent on the M6P content per lysosomal enzyme. Affinity chromatography by utilizing the Dom9-His-conjugated beads is also applicable to purify the terminal M6P-carrying lysosomal enzymes [16]. This chapter describes the production of recombinant Dom9His by a trangenic P. pastris strain and purification as well as the application for lectin blotting as a probe.

2

Materials

2.1 Preparation of Recombinant Proteins

1. 1 M potassium phosphate buffer (pH 6.0): Mix 1 M KH2PO4 solution with 1 M K2HPO4 solution to adjust the pH to 6.0 (see Note 1). 2. 1.7% (w/v) YNB: Weigh 1.7 g of yeast nitrogen base not containing (NH4)2SO4, and transfer to a 100 mL graduated

270

Kei Kiriyama and Kohji Itoh

cylinder. Dissolve with water to adjust the volume to 100 mL, and then filtered through a filter. 3. BMGY: 2% (w/v) peptone, 1% (w/v) yeast extract, 0.15 M (NH4)2SO4, 1% (v/v) glycerol, 0.1 M potassium phosphate buffer, and 0.17% (w/v) YNB. Weigh 10 g peptone, 5 g yeast extract, and 10 g (NH4)2SO4, and transfer to a 500 mL graduated cylinder. Add water and 5 mL of glycerol to adjust the volume to 400 mL, and then add each 50 mL of 1 M potassium phosphate buffer (pH 6.0) and 1.7% (w/v) YNB to be autoclaved. 4. BMMY: 2% (w/v) peptone, 1% (w/v) yeast extract, 0.15 M (NH4)2SO4, 0.1 M potassium phosphate buffer (pH 6.0), 0.17% (w/v) YNB, and 0.5% (v/v) MeOH. Weigh 10 g peptone, 5 g yeast extract, and 10 g (NH4)2SO4, and transfer to a 500 mL graduated cylinder. Dissolve with water to adjust the volume of 400 mL, and add each 50 mL of 1 M potassium phosphate buffer (pH 6.0) and 1.7% (w/v) YNB, and then add 2.5 mL of MeOH after autoclaving. 5. 20% (v/v) EtOH: Add 100 mL of EtOH to adjust the volume to 500 mL with water, and then filtered through a filter. 6. 100 mM sodium phosphate buffer (pH 7.5): Mix 100 mM NaH2PO4 solution and 100 mM Na2HPO4 solution to adjust the pH to 7.5 (see Note 2). 7. 20 mM sodium phosphate buffer (pH 7.5): Add 100 mL of 100 mM sodium phosphate buffer to a 500 mL graduated cylinder to adjust the volume to 500 mL with water. 8. 20 mM sodium phosphate buffer (pH 7.5)/20 mM imidazole: Add 100 mL of 100 mM sodium phosphate buffer and 300 mL water to a 500 mL glass beaker. Weigh 0.68 g imidazole and transfer to the glass beaker. After dissolving, adjust the pH to 7.5 and the volume to 500 mL with water. 9. 20 mM sodium phosphate buffer (pH 7.5)/500 mM imidazole: Add 100 mL of 100 mM sodium phosphate buffer and 300 mL water to a 500 mL glass beaker. Weigh 17 g imidazole and transfer to the glass beaker. Adjust the pH to 7.5 and the volume to 500 mL with water. 10. Cells: yeast P. pastris overexpressing the recombinant human Dom9-His. 11. Vector: pPIC9 vector for the expression in P. pastris strain. 12. Column: open mini-column 5  50 mm. 13. Carrier: Ni-Sepharose 6 Fast Flow(GE Healthcare). 14. Concentrator: Amicon Ultra-10 (Merck Millipore).

Application of Mannose 6-Phosphate Receptor Domains

2.2

SDS-PAGE

271

1. 30% acrylamide solution [29.2% (w/v) Acrylamide/0.8% (w/v) Bis] for running gel preparation: Weigh 146 g of acrylamide monomer and 4 g Bis (cross-linker), and transfer to a 500 mL graduated cylinder containing about 100 mL of water. Then dissolve to adjust the volume to 500 mL with water and stored at 4  C. 2. Running gel buffer: 1.5 M Tris–HCl (pH 8.8). Add about 100 mL water to a 500 mL graduated cylinder or a glass beaker. Weigh 90.855 g Tris and transfer to the cylinder. Dissolve in water and adjust the pH with HCl and the volume to 500 mL with water. Stored at 4  C. 3. Stacking gel buffer: 0.5 M Tris–HCl (pH 6.8). Weigh 30.3 g Tris and prepare a 500 mL solution as described in step 2. Stored at 4  C. 4. 10% SDS: 10% (w/v) sodium dodecyl sulfate (SDS) solution in water. 5. 10% APS: 10% (w/v) ammonium persulfate (APS): solution in water. Stored at 20  C. 6. N,N,N0 ,N0 -Tetramethylethylenediamine. 7. SDS-PAGE electrophoresis buffer: 0.025 M Tris–HCl (pH 8.3), 0.192 M glycine, and 0.1% SDS. 8. SDS sample buffer (6): 0.3 M Tris–HCl (pH 6.8), 36% (v/v) glycerol, 1.2% (v/v) 2-mercaptoethanol, 24% (w/v) SDS, and 0.012% Bromophenol blue. Stored at 20  C.

2.3

Lectin Blotting

1. Polyvinylidene Millipore).

difluoride

(PVDF)

membrane

(Merck

2. Filter paper. 3. Western blot transfer buffer: 0.048 M Tris–HCl (pH 8.5), 0039 M glycine, and 20% (v/v) methanol. 4. Tris-buffered saline (TBS; 10): 1.37 M NaCl, 0.027 mM KCl, and 0.25 M Tris–HCl (pH 7.4). 5. TBS containing 0.1% (v/v) Tween20 (TBST). 6. Blocking solution: Mix Blocking One (Nacalai Tesque) and TBST at a volume ratio of 1:1. Stored at 4  C. 7. Sample: purified modified and protease-resistant HexB (mod2B) (see Ref. [17]). 8. PNGase F (New England Biolabs). 9. Dom9-His solution: 5 μg/mL recombinant Dom9-His in Blocking solution. Prepare this solution immediately before use. 10. Anti-His solution: Purified mouse monoclonal Penta-His antibody (QIAGEN) (1:500 dilution). Confirm the optimal concentration under each experimental condition.

272

Kei Kiriyama and Kohji Itoh

11. Anti-mouse IgG horseradish peroxidase (HRP)-conjugated solution from (Cell Signaling technology). 12. Coloring reagent: Western Lightning Chemiluminescence Reagent Ultra (Perkin Elmer).

3

Methods

3.1 Expression and Purification of Recombinant Human Dom9-his by Methylotrophic Yeast

1. Add the yeast P. pastris strain overexpressing the recombinant Dom9-His to a flask containing 50 mL of BMGY (see Note 3). 2. Incubate at 30  C for 2 days with shaking at 150 rpm. 3. Transfer to a 50 mL tube. Centrifuge at 3500  g for 10 min at 4  C. 4. After removing the supernatant, the resultant cell pellets were resuspended in 50 mL of BMGY and transfer to a 200 mL Erlenmeyer flask. 5. Incubate at 30  C for 7 days with shaking at 150 rpm (see Note 4). 6. Centrifuge as described in step 3. 7. Collect the supernatant and then filtered with a 0.45 μm filter (see Note 5). 8. Exchange the buffer composition of the solution to 20 mM sodium phosphate buffer (pH 7.5) after concentrating with 10-kDa cut concentrator (see Note 6). 9. Pack 100 μL gel of Ni-Sepharose 6 Fast Flow into the open column (Item 12 of Subheading 2.1) and wash the gel with 20% (v/v) EtOH. 10. Add 20 mM sodium phosphate buffer (pH 7.5) to the column for equilibration. 11. Add the supernatant containing the Dom9-His prepared by step 8 to the column at 4  C (see Note 7). 12. Add 1 mL of 20 mM sodium phosphate buffer (pH 7.5)/ 20 mM imidazole to the column twice for washing at 4  C. 13. Add total 3 mL of 20 mM sodium phosphate buffer (pH 7.5)/ 500 mM imidazole to the column at 4  C to elute the Dom9His and collect the eluate. 14. Transfer the eluate to the Amicon Ultra-10 and centrifuge for concentration at 2300  g for 20 min at 4  C. 15. Add 20 mM sodium phosphate buffer (pH 7.5) to the Amicon and centrifuge at 2300  g for 20 min at 4  C. Repeat this step total three times to remove imidazole and concentrate to less than 200 μL (see Note 8).

Application of Mannose 6-Phosphate Receptor Domains

3.2 Lectin Blotting with the Dom9-his as a Probe

273

1. Heat the samples mixed with 1/6 volume of SDS sample dye at 100  C for 3 min. Apply the samples to SDS-PAGE (see Note 9). 2. The gel after the electrophoresis was transblotted to the PVDF membrane at 15 V for 1 h. 3. The blotted membrane was treated with blocking solution (Item 6 of Subheading 2.3) and shaking for 1.5 h. 4. Add the Dom9-His solution (Item 9 of Subheading 2.3) to the blocked membrane and incubate for 1 h. 5. Rinse the with TBST.

Dom9-His-treated

membrane

three

times

6. Add anti-His solution to the rinsed membrane and incubate for 1 h (see Note 10). 7. Wash the anti-His-treated membrane as described in step 5. 8. Add HRP-conjugated anti-mouse IgG solution and incubate for 1 h. 9. Rinse the HRP-conjugated anti-mouse IgG-treated membrane three times with TBST and once with TBS. 10. Add the chemiluminescence reagent and incubate for 1 min. 11. Detect the terminal M6P-carrying enzyme species on the treated membrane by chemiluminescence detection device (see Fig. 3).

a Western blotting PNGase F (–) KDa

(+)

Purified mod2B

b CBB staining PNGase F (–) KDa

100

100

75

75

50

50

37

37

(+)

Purified mod2B

Fig. 3 Lection blotting with Dom9-His solution as a probe. One microgram of purified mod2B (see Ref. [17]) with terminal M6P-carrying N-glycans per lane was applied. (a) The band of purified mod2B was detected by lectin blotting with Dom9-His. However, the band was not detected after PNGase F treatment. (b) CBB staining showed the presence of the mod2B protein bands before and after PNGase F treatment. These show that recombinant Dom9-His can be utilized for detection of glycoprotein containing M6P residues as a probe

274

4

Kei Kiriyama and Kohji Itoh

Notes 1. Weigh 68.05 g KH2PO4 and transfer to the cylinder. Add water to adjust the volume to 500 mL (1 M KH2PO4 solution). Weigh 87.1 g K2HPO4 and dissolve in water to adjust the volume to 500 mL (1 M K2HPO4 solution). 2. Weigh 6.0 g NaH2PO4 and 7.1 g Na2HPO4. Add water to prepare each 500 mL solution (100 mM NaH2PO4 and 100 mM Na2HPO4 solution) as described in the step 1. 3. We produced the recombinant human Dom9-His cDNA construct and expressed in yeast P. pastris according to the method described by Akeboshi et al. (2007) and Hancock et al. (2002) (see Ref. [10, 11]). 4. Expression of the two gene in yeast P. pastris encoding alcohol oxidases AOX1 and AOX2 is induced by the addition of methanol. The pPIC9 vector containing AOX1 promoter sequence is utilized to express the Dom9-His proteins at high levels. Methanol (each 250 μL) is added to 500 mL of the culture medium every day. 5. Filtration of the supernatant with a filter membrane is performed to remove the yeast cells completely from the supernatant. 6. Adjust the pH of the supernatant from 6.5 to 7.6–7.8 before the purification with Ni-Sepharose column chromatography. The buffer composition of the eluate containing the Dom9His is exchanged for 20 mM sodium phosphate buffer (pH 7.5) by centrifugation at 2300  g for 20 min at 4  C with 10-kDa Amicon Ultra-10 because the molecular mass of the Dom9-His is about 20-kDa. The pH in the supernatant was adjusted by adding basic salts directly to the supernatant. 7. Filtration of the supernatant can prevent clogging of the Ni-Sepharose column. 8. The concentration of the Dom9-His as a probe for lectin blotting is recommended to be 2–5 μg/mL, which is achieved by concentrating the eluate from Ni-Sepharose column to less than 200 μL. The concentration of the purified Dom9-His was determined by Western blotting and CBB staining (see Fig. 4). The purified Dom9-His fractions were stored at 30  C. 9. The samples of few volume may not be detected for lectin blotting with Dom9-His as a probe, depending on the various types of coloring reagents, chemiluminescence detection devices, and the M6P content of lysosomal enzymes. The volume of the purified terminal M6P-carrying lysosomal enzymes is recommended to be over 1 μg.

Application of Mannose 6-Phosphate Receptor Domains

a

275

b

Elu tio n

KDa

KDa 50 40 30 20

Su pe rna tan Elu t tio nf rac tio n

CBB staining

fra ctio n

Western blotting

100 70 45

Recombinant Dom9-His 23.5-KDa

30 20 14

Fig. 4 Purity and yield of the recombinant Dom9-His. (a) Single band of the Dom-His (23.5-kDa) detected by Western blotting with anti-His antibody as a probe. (b) Major single band due to the Dom9-His in Ni-Sepharose elution fraction stained with Coomassie Brilliant Blue (CBB) R-350. Approximately 0.45 mg of the recombinant Dom9-His from 1 L of the conditioned medium of the methyltrophic yeast P. pastoris

10. The condition under which the blotted membrane is treated with the Dom9-His is also available for incubating overnight at 4  C. References 1. von Figura K, Hasilik A (1986) Lysosomal enzymes and their receptors. Annu Rev Biochem 55:167–193 2. Dahms NM, Lobel P, Kornfeld S (1989) Mannose 6-phosphate receptors and lysosomal enzyme targeting. J Biol Chem 264:12115–12118 3. Munier-Lehmann H, Mauxion F, Hoflack B (1996) Function of the two mannose 6-phosphate receptors in lysosomal enzyme transport. Biochem Soc Trans 24:133–136 4. Dahms NM, Kornfeld S (1989) The cationdependent mannose 6-phosphate receptor. Structural requirements for mannose 6-phosphate binding and oligomerization. J Biol Chem 264:11458–11467 5. Roberts DL, Weix DJ, Dahms NM et al (1998) Molecular basis of lysosomal enzyme recognition: three-dimensional structure of the cationdependent mannose 6-phosphate receptor. Cell 93:639–648 6. Marron-Terada PG, Bollinger KE, Dahms NM (1998) Characterization of truncated and glycosylation-deficient forms of the cationdependent mannose 6-phosphate receptor

expressed in baculovirus-infected insect cells. Biochemistry 37:17223–17229 7. Lobel P, Dahms NM, Kornfeld S (1988) Cloning and sequence analysis of the cationindependent mannose 6-phosphate receptor. J Biol Chem 263:2563–2570 8. Sun G, Zhao H, Kalyanaraman B et al (2005) Identification of residues essential for carbohydrate recognition and cation dependence of the 46-kDa mannose 6-phosphate receptor. Glycobiology 15:1136–1149 9. Reddy ST, Chai W, Childs RA et al (2004) Identification of a low affinity mannose 6-phosphate-binding site in domain 5 of the cationindependent mannose 6-phosphate receptor. J Biol Chem 279:38658–38667 10. Hancock MK, Yammani RD, Dahms NM (2002) Localization of the carbohydrate recognition sites of the insulin-like growth factor II/mannose 6-phosphate receptor to domains 3 and 9 of the extracytoplasmic region. J Biol Chem 277:47205–47212 11. Akeboshi H, Chiba Y, Kasahara Y et al (2007) Production of recombinant β-Hexosaminidase a, a potential enzyme for replacement therapy

276

Kei Kiriyama and Kohji Itoh

for Tay-Sachs and Sandhoff diseases in the Methylotrophic yeast Ogataea minuta. Appl Environ Microbiol 73:4805–4812 12. Nolan CM, Creek KE, Grubb JH et al (1987) Antibody to the phosphomannosyl receptor inhibits recycling of receptor in fibroblasts. J Cell Biochem 35:137–151 13. Stein M, Zijderhand-Bleekemolen JE, Geuze H et al (1987) Mr 46,000 mannose 6-phosphate specific receptor: its role in targeting of lysosomal enzymes. EMBO J 6:2677–2681 14. Hawkes C, Kar S (2003) Insulin-like growth factor-II/mannose-6-phosphate receptor: widespread distribution in neurons of the central nervous system including those expressing

cholinergic phenotype. J Comp Neurol 458:113–127 15. Barton NW, Brady RO, Dambrosia JM et al (1991) Replacement therapy for inherited enzyme deficiency  macrophage-targeted glucocerebrosidase for Gaucher’s disease. N Engl J Med 324:1464–1470 16. Tong PY, Gregory W, Kornfeld S (1989) Ligand interactions of the cation-independent mannose 6-phosphate receptor. The stoichiometry of mannose 6-phosphate binding. J Biol Chem 264:7962–7969 17. Kitakaze K, Mizutani Y, Sugiyama E et al (2016) Protease-resistant modified human β-hexosaminidase B ameliorates symptoms in GM2 gangliosidosis model. J Clin Invest 126:1691–1703

Chapter 27 Purification and Assays of Tachylectin-5 Shun-ichiro Kawabata and Toshio Shibata Abstract Tachylectin-5, a 41-kDa protein with a common fold of the C-terminal globular domain of the γ-chain of fibrinogen, is purified from horseshoe crab hemolymph plasma by affinity column chromatography, using acetyl-group-immobilized resin. Two types of isolectins, tachylectin-5A and tachylectin-5B, are obtained by stepwise elution with GlcNAc at 25 and 250 mM, respectively. Tachylectins-5A and -5B exhibit extraordinarily strong hemagglutinating activity against all types of human erythrocytes (the minimum agglutinating concentration of 0.004–0.008 μg/mL for tachylectin-5A and 0.077–0.27 μg/mL for tachylectin-5B). Their hemagglutinating activities are inhibited by acetyl group-containing sugars and noncarbohydrates such as sodium acetate, acetylcholine, and acetyl CoA (the minimum inhibitory concentrations of 1.3–1.6 mM), indicating that the acetyl group is required and sufficient for recognition by tachylectins5A and -5B. EDTA inhibits their hemagglutinating activity, whereas the inhibition is overcome by adding an excess amount of Ca2+. Tachylectins-5A and -5B also exhibit bacterial agglutinating activity against both Gram-negative bacteria (the minimum agglutinating concentrations of 0.04–0.08 μg/mL for tachylectin5A and 0.05–0.11 μg/mL for tachylectin-5B) and Gram-positive bacteria (the minimum agglutinating concentrations of 0.3–2.4 μg/mL for tachylectin-5A and 15.1–26.8 μg/mL for tachylectin-5B). Interestingly, tachylectins-5A and -5B enhance the antimicrobial activity of a hemocyte-derived peptide, big defensin. Key words Fibrinogen-related proteins, Hemolymph plasma, Hemagglutination, Acetyl-containing substances, Tachylectin

1

Introduction Tachylectins-5A and 5B are the most abundant lectins in hemolymph plasma of the Japanese horseshoe crab Tachypleus tridentatus [1], whereas hemolymph plasma also contains isoforms of tachylectins-1 and -3 and several types of C-reactive proteins [2– 4]. We had a plan to purify these plasma-derived lectins by affinity column chromatography, using sugar-immobilized Toyopearl AF-amino-650 M. For preparing control resin without sugar ligands, the functional group of this resin was blocked by acetic anhydride, resulting in the immobilization of N-acetyl group on the surface of the resin (acetyl-Toyopearl). When hemolymph

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_27, © Springer Science+Business Media, LLC, part of Springer Nature 2020

277

278

Shun-ichiro Kawabata and Toshio Shibata

plasma was applied to the acetyl-Toyopearl column, about two-thirds of the total hemagglutinating activity in plasma against all types of human erythrocytes unexpectedly bound to the acetyl Toyopearl resin, and two types of lectins, tachylectin-5A and tachylectin-5B, were obtained [1]. The overall sequence identity between tachylectins-5A (269 amino acids) and -5B (289 amino acids) is 45%. Tachylectins-5A and -5B consist of a short N-terminal Cys-containing segment and a C-terminal fibrinogen-γ-chain-like domain with the highest sequence identity to that of mammalian ficolins (~50%). Interestingly, tachylectins-5A and -5B lack the N-terminal collagenous domain found in ficolins and collectins. The crystal structure of tachylectin-5A is superimposed onto that of the C-terminal globular domain of the γ-chain of mammalian fibrinogen, functioning as fibrin polymerization [5]. The binding pocket of the acetyl group of tyachylectin-5A strictly corresponds to the polymerization pocket of the γ-chain. Fibrinogen-related proteins have been identified not only from vertebrates but also from invertebrates [6–12]. Tachylectins-5A and -5B are present at the concentration of ~10 μg/mL in normal hemolymph plasma, a sufficient concentration for maintaining their agglutination activity, and therefore, they may play an important role in the recognition of invading pathogen at the forefront of the innate immune system in horseshoe crabs [1–4]. The formation of hexameric and octameric bouquet-like arrangement of tachylectin5A is observed by electron microscopy, and the Cys-6 and Cys-170 residues of tachylectin-5A are expected to be involved in the oligomerization through internal disulfide bridges [1]. However, crystallized tachylectin-5A is a monomer form in the asymmetric unit (Fig. 1), and these two cysteine residues are present in the free SH forms in native tachylectin-5A [5], indicating tachylectin-5A forms the unique oligomeric structure by noncovalent interaction. Here we describe a protocol of the purification and characterization of tachylectins-5A and -5B [1].

2

Materials

2.1 Preparation of Hemolymph Plasma

Glassware and metalware used for the preparation of hemocyte lysate and the dextran sulfate chromatography are sterilized by heating at 220  C for 3 h. All the buffer solutions used for these steps are made up with pyrogen-free distilled water and autoclaved for 30 min (see Note 1). 1. 70% (v/v) ethanol. 2. Pyrogen-free distilled water (Otsuka Pharmaceutical, Tokyo, Japan). 3. 20 mM Tris–HCl, pH 8.0, containing 50 mM NaCl.

Characteristics of Tachylectin-5

279

Fig. 1 The crystal structure of tachylectin-5A complexed with GlcNAc (PDB code 1JC9 [5]). A Ca2+ ion is represented by a yellow ball

4. 20 mM Tris–HCl, pH 8.0, containing 0.5 M NaCl. 5. 20 mM Tris–HCl, pH 8.0, containing 0.5 M NaCl and 0.1 M caffeine. 6. Disposable sterilized needle. 7. Disposable sterilized plastic container. 2.2 Purification of Tachylectin-5

1. Toyopearl AF-Amino-650 M (Tosoh, Tokyo, Japan).

2.2.1 Preparation of Acetyl Group-Immobilized Resin (Acetyl-Toyopearl)

3. Acetic anhydride.

2.2.2 Acetyl-Toyopearl Column Chromatography

1. 20 mM Tris–HCl, pH 8.0, containing 10 mM CaCl2.

2. 0.2 M sodium acetate. 4. 1 M NaOH. 5. 20 mM Tris–HCl, pH 8.0, containing 10 mM CaCl2.

2. 20 mM Tris–HCl, pH 8.0, containing 10 mM CaCl2 and 0.5 M NaCl. 3. 20 mM Tris–HCl, pH 8.0, containing 10 mM CaCl2 and 25 mM GlcNAc. 4. 20 mM Tris–HCl, pH 8.0, containing 10 mM CaCl2 and 250 mM GlcNAc.

2.3 Hemagglutination Assay

1. Out-dated human A, B, O-type concentrated erythrocytes (see Note 2). 2. 20 mM Tris–HCl, pH 7.5, containing 0.15 M NaCl and 5 mM CaCl2. 3. Round-bottomed microtiter plates.

280

Shun-ichiro Kawabata and Toshio Shibata

2.4 Bacterial Agglutination Assay

1. Bacteria: Staphylococcus aureus 209P, Staphylococcus epidermidis K3, Staphylococcus saprophyticus KD, Micrococcus luteus, Enterococcus hirae, and Escherichia coli strain B. 2. 3% Tryptosoy broth sterilized by autoclaving for 20 min. 3. 20 mM Tris–HCl, pH 7.5, containing 0.15 M NaCl and 5 mM CaCl2. 4. Round-bottomed microtiter plates.

2.5 A Synergistic Effect of Tachylectin-5A or Tachylectin-5B on Antimicrobial Activity of Big Defensin

1. Bacteria: Salmonella typhimuriumLT2 (S), Salmonella minnesota R595 (Re), Escherichia coli B, Klebsiella pneumoniae, Staphylococcus aureus, and Candida albicans (see Note 3). 2. 3% Tryptosoy broth sterilized by autoclaving for 20 min. 3. 20 mM Tris–HCl, pH 7.0. 4. 1% agar plates containing 3% Tryptosoy broth. 5. Guanofracin-Sabouraud agar plates for C. albicans (see Note 4). 6. Big defensin purified from hemocyte debris [13].

3

Methods

3.1 Preparation of Hemolymph Plasma

1. Bleed hemolymph by inserting a sterilized needle into the joint between the cephalothorax and the abdomen rinsed with 70% ethanol and then, with sterilized water. 2. Collect the hemolymph (~150 mL from each horseshoe crab) into a sterilized container containing one-tenth volume (~15 mL) of 20 mM Tris–HCl, pH 8.0, containing 0.5 M NaCl and 0.1 M caffeine chilled on ice. 3. Pool the hemolymph into sterilized metal centrifugation tubes (~300 mL each). 4. Centrifuge the hemolymph at 3000  g for 15 min at 4  C and separated into hemocytes and hemolymph plasma. 5. Store hemocytes and hemolymph plasma at 80  C until use, and both are stable at least for 2 years.

3.2 Purification of Tachylectin-5

1. Wash 10 g of Toyopearl AF-Amino-650 M resin with distilled water.

3.2.1 Preparation of Acetyl-Toyopearl

2. Mix the washed resin with 8 mL of 0.2 M sodium acetate and 4 mL of acetic anhydride. 3. Incubate the mixture on ice for 30 min at 25  C. 4. Added another 4 mL of acetic anhydride to the mixture and incubate for 30 min at 25  C.

Characteristics of Tachylectin-5

281

5. Wash the resin with several times with distilled water and 1 M NaOH reciprocally and then wash with 20 mM Tris–HCl, pH 8.0, containing 10 mM CaCl2. 6. Suspend the washed resin in the same buffer. 3.2.2 Acetyl-Toyopearl Column Chromatography

1. Equilibrate an acetyl-Toyopearl column (1  8 cm) with 20 mM Tris–HCl, pH 8.0, containing 10 mM CaCl2. 2. Apply 500 mL of hemolymph plasma and wash the column extensively (~200 mL) with the equilibration buffer containing 0.5 M NaCl. 3. Bound proteins were eluted with 25 and 250 mM GlcNAc in the equilibration buffer in a stepwise fashion (50 mL each). 4. Pool the 25-mM GlcNAc and 250-mM GlcNAc-eluted fractions, containing tachylectins-5A and -5B, respectively.

3.3 Hemagglutination Assay

1. Centrifuge human concentrated erythrocytes at 1000  g for 3 min and wash three times with 20 mM Tris–HCl, pH 7.5, containing 0.15 M NaCl and 5 mM CaCl2. 2. Prepare 2% suspension (v/v) of erythrocytes in the same buffer (see Note 5). 3. Mix 25 μL of the suspension of erythrocytes with 25 μL of a two-fold serial dilution of tachylectin-5A or tachylectin-5B in a round-bottomed microtiter plate and incubate for 1 h at 37  C. 4. The titer is defined as a reciprocal value of the endpoint dilution causing hemagglutination. 5. For screening of inhibitors for hemagglutination, premix 12.5 μL of a test sample in the same buffer and 12.5 μL of tachylectin-5A or tachylectin-5B (0.4 μg/mL) before adding 25 μL of the suspension of erythrocytes. Inhibition activity is expressed as the minimum inhibitory concentration of the test sample.

3.4 Bacterial Agglutination Assay

1. Culture bacteria in 5 mL of 3% Tryptosoy broth for 12 h at 37  C. 2. Collect bacteria by centrifugation at 4000  g for 2 min and wash with 20 mM Tris–HCl, pH 7.5, containing 0.15 M NaCl and 5 mM CaCl2. 3. Suspend the washed bacteria in ~1.0 mL of the same buffer to obtain an absorbance at 600 nm of ~10. 4. Mix 25 μL of the suspension of bacteria with 25 μL of a twofold serial dilution of tachylectin-5A or tachylectin-5B in a roundbottomed microtiter plate and incubate for 12 h at 25  C. 5. The bacterial agglutinating activity is expressed as the minimum agglutinating concentration of tachylectin-5A or tachylectin-5B.

282

Shun-ichiro Kawabata and Toshio Shibata

3.5 A Synergistic Effect of Tachylectin-5A or Tachylectin-5B on Antimicrobial Activity of Big Defensin

1. Culture bacteria in 3 mL of 3% Tryptosoy broth for 12 h at 37  C. 2. Collect bacteria by centrifugation at 4000  g for 2 min and wash them with 20 mM Tris–HCl buffer, pH 7.0. 3. Suspend the washed bacteria in 3 mL of 20 mM Tris–HCl buffer, pH 7.0 by mixing with vortex. 4. Dilute the bacterial suspension ~105-fold with the same buffer by using a sterile micropipette tip to transfer 10 μL to a fresh tube containing 1 mL of the same buffer repeatedly (see Note 6). 5. Mix 25 μL of big defensin (2.5 μg/mL) with 25 μL of tachylectin-5A or tachylectin-5B (1 μg/mL) and add the mixture to 450 μL of the diluted bacterial suspension. As a negative control experiment, mix 25 μL of the Tris–HCl buffer with 25 μL of big defensin (2.5 μg/mL). 6. Incubate the mixture for 1 h at 37  C. 7. Plate 100 μL of the reaction mixture onto 1% agar plates containing 3% Tryptosoy broth and incubate at 37  C for 24 h. For C. albicans, use guanofracin-Sabouraud agar plates. 8. Count the number of bacterial colonies on the agar plates and make the counted number ten times to obtain colony forming units (CFU, number of colonies/mL of test sample) (see Note 7).

4

Notes 1. If these procedures for sterilization are omitted, the coagulation factors in hemocytes will be converted to the active forms, leading to clot formation. 2. Out-dated human concentrated erythrocytes can be stored for at least 1 month at 4  C. 3. Bacteria can be stored for many years in media containing 15% glycerol at 80  C. 4. The guanofracin-Sabouraud agar plate for C. albicans is commercially available. 5. Volume of erythrocytes is measured in a centrifuge tube with scale. 6. Repeat the dilution step serially to obtain the diluted bacterial suspension of 5000–10,000 cells/mL. OD ¼ 1.0 (optical density at 600 nm) of the bacterial suspension is roughly equivalent to 8  108 cells/mL. 7. Use at least five plates for one test sample and calculate the average value of CFU.

Characteristics of Tachylectin-5

283

References 1. Gokudan S, Muta T, Tsuda R et al (1999) Horseshoe crab acetyl group-recognizing lectins involved in innate immunity are structurally related to fibrinogen. Proc Natl Acad Sci U S A 96:10086–10091 2. Kawabata S, Tsuda R (2002) Molecular basis of non-self recognition by the horseshoe crab tachylectins. Biochim Biophys Acta 1572:414–421 3. Iwanaga S (2002) The molecular basis of innate immunity in the horseshoe crab. Curr Opin Immunol 14:87–95 4. Kawabata S (2011) Immunocompetent molecules and their response network in horseshoe crabs. In: So¨derh€all K (ed) Invertebrate immunity. Springer Science + Business Media, New York, pp 122–136 5. Kairies N, Beisel H-G, Fuentes-Prior P et al (2001) The 2.0-A˚ crystal structure of tachylectin 5 provides evidence for the common origin of the innate immunity and the blood coagulation systems. Proc Natl Acad Sci U S A 98:13519–13524 6. Xu X, Doolittle RF (1990) Presence of vertebrate fibrinogen-like sequence in an echinoderm. Proc Natl Acad Sci U S A 87:2097–2101 7. Adema CM, Hertel LA, Miller RD, Loker ES (1997) A family of fibrinogen-related proteins that precipitates parasite-derived molecule is produced by an invertebrate after infection. Proc Natl Acad Sci U S A 94:8691–8696

8. Kurachi S, Song Z, Takagaki M et al (1998) Sialic-acid binding lectin from the slug Limax flavus: cloning, expression of the polypeptide, and tissue localization. Eur J Biochem 254:217–222 9. Dimopoulos G, Casavant TL, Chang SR et al (2000) Anopheles gambiae pilot gene discovery project: identification of mosquito innate immunity genes from expressed sequence tags generated from immune-competent cell lines. Proc Natl Acad Sci U S A 97:6619–6624 10. Kenjo A, Takahashi M, Matsushita M et al (2001) Cloning and characterization of novel ficolins from the solitary ascidian, Halocynthia roretzi. J Biol Chem 276:19959–19965 11. Leonard PM, Adema CM, Zhang S-MLoker ES (2001) Structure of two FREP genes that combine IgSF and fibrinogen domains, with comments on diversity of the FREP gene family in the snail. Biomphalaria glabrata. Gene 269:155–165 12. Angthong P, Roytrakul S, Jarayabhand P, Jiravanichpaisal P (2017) Characterization and function of a tachylectin 5-like immune molecule in Penaeus monodon. Dev Comp Immunol 76:120–131 13. Saito T, Kawabata S, Shigenaga T et al (1995) A novel big defensin identified in horseshoe crab hemocytes: isolation, amino acid sequence, and antibacterial activity. J Biochem 117:1131–1137

Chapter 28 Preparation of Soluble Malectin and Its Tetramer Sheng-Ying Qin, Dan Hu, and Kazuo Yamamoto Abstract Malectin is a membrane-anchored endoplasmic reticulum (ER)-resident lectin, which is first identified in Xenopus laevis in 2008 and highly conserved among animals. Malectin plays an important role in the quality control of glycoprotein in the ER through recognizing the Glc2Man9GlcNAc2 (G2M9) oligosaccharide chain on the newly synthesized glycoproteins. In this chapter, we will describe the preparation of recombinant soluble malectin and its tetramer, which might be developed as useful tools for detection of Glcα13Glc containing glycans on the cell surface. Key words Lectin, Glycan, Quality control of glycoprotein, Malectin, G2M9

1

Introduction Malectin is a type I membrane-anchored lectin localized in endoplasmic reticulum (ER), which is first identified in Xenopus laevis in 2008. This lectin is highly conserved among animals [1]. Previous studies have shown that malectin specifically recognizes Glc2Man9GlcNAc2 (G2M9), an oligosaccharide appearing in the early stage of N-glycan processing [1, 2], suggesting that malectin is likely to be involved in the quality control of glycoproteins. We and other groups have demonstrated that malectin is upregulated under ER stress conditions and preferentially associated with folding defective glycoproteins [2, 3]. We have further elucidated the mechanism by which malectin selectively retains misfolded proteins in the ER. We found that malectin forms a complex with ribophorin I, in which malectin recognizes G2M9 glycan portion while ribophorin I recognizes the misfolded protein backbone portion [4, 5]. Very recently, we observed that the subcellular distribution of malectin is also regulated by ribophorin I [6]. All the above results indicate that malectin might play an import role in the quality control of glycoproteins. In view of its unusual sugarbinding specificity and important physiological functions, it is very necessary to know about the preparation of recombinant soluble

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_28, © Springer Science+Business Media, LLC, part of Springer Nature 2020

285

286

Sheng-Ying Qin et al.

sMAL

sMAL

+biotin

+SA-PE bio

bio bio bio bio SA PE

sMAL

bio-sMAL

bio-sMAL-SA-PE complex

Fig. 1 Schematic representation of the preparation of sMAL-bio-SA-PE tetramer

malectin and its tetramer, which might be developed as useful tools for detection of Glcα1-3Glc containing glycans. Malectin is a type I membrane protein consisting of an Nterminal signal peptide (aa 1-26), a luminal domain (aa 27-268), a C-terminal transmembrane helix (aa 269-290) and a short cytoplasmic fraction (aa 291-292) [1]. Considering the weak interaction between sugar and lectin [7, 8], the soluble luminal region of malectin fused with an N-terminal His-tag and a C-terminal biotinylating tag (sMAL) was prepared and biotinylated (bio-sMAL), which is then used to prepare a tetrameric complex with R-phycoerythrin (PE)-labeled streptavidin (SA) (bio-sMAL-SA-PE) (Fig. 1).

2

Materials

2.1 Construction of Plasmids for sMAL Expression in Escherichia coli

1. cDNA library of human HeLaS3 cells. 2. T4 polynucleotide kinase kit (Toyobo, 10 T4 polynucleotide kinase buffer, 10 mM ATP, and 10 U/μL T4 polynucleotide kinase). 3. KOD-plus DNA polymerase kit (Toyobo, 10 KOD plus buffer, 2 mM dNTPs, 25 mM MgSO4, and 1 U/μL KOD plus polymerase). 4. QIAquick Gel Extraction kit, QIAGEN Plasmid Min kit (Qiagen). 5. DNA modifying enzymes: SmaI, NdeI, and EcoRI. 6. Agarose (Takara). 7. Sterile deionized water. 8. 1 Tris-acetate electrophoresis buffer (1 TAE buffer): Prepare a stock solution of 50 TAE and dilute it 1:50 with deionized water before use.

Preparation of Recombinant Malectin

287

9. 50 TAE buffer: 242 g/L Tris base, 57.1 mL/L glacial acetic acid, and 100 mL/L of 0.5 M EDTA. Dissolve 242 g of Tris base in a sufficient amount of deionized water, then add 57.1 mL of glacial acetic acid and 100 mL of 0.5 M EDTA, and finally made up to 1 L with deionized water. 10. 0.5 M ethylenediamine tetraacetic acid (EDTA):Weigh 186.1 g Na2EDTA·2H2O and 20 g NaOH, dissolve them with deionized water, and finally dilute to 1 L with deionized water, autoclave and store at room temperature (see Note 1). 11. Ethidium bromide (10 mg/mL). 2.2 Expression and Purification of sMAL

1. LB medium: Weigh 10 g of tryptone, 5 g of yeast extract, and 10 g of NaCl, dissolve them with 1 L deionized water, autoclave and store at 4  C. Add ampicillin to a final concentration of 100 μg/mL and 25 μg/mL chloramphenicol before use. 2. 100 mg/mL ampicillin (stock solution): Dissolve 10 g of ampicillin in about 40 mL of deionized water, finally made up to 100 mL and sterilized by filtration. Dispense into small portions for storage at 20  C. 3. 25 mg/mL chloramphenicol (stock solution): Dissolve 2.5 g of chloramphenicol in a sufficient amount of absolute ethanol, finally made up to 100 mL and sterilized by filtration. Dispense into small portions for storage at 20  C. 4. 1 M isopropyl-β-D-thio-galactoside (IPTG) (stock solution): Dissolve 2.38 g of IPTG in deionized water and adjust the final volume to 10 mL. Filter sterilize and dispense into small portions for storage at 20  C. 5. Phosphate buffer saline (PBS) (stock solution of 10 PBS): Weigh 80 g of NaCl, 2.0 g of KCl, 14.2 g of Na2HPO4, and 2.7 g of KH2PO4, dissolve them with about 800 mL of deionized water, then adjust the pH to 7.5 with HCl, and finally dilute to 1 L with deionized water, autoclave and store at room temperature (see Note 2). 6. 100 mM phenylmethylsulfonyl fluoride (PMSF) (stock solution): Dissolve 174.1 mg of PMSF in absolute ethanol and adjust the final volume to 10 mL. Dispense into small portions and store at 20  C (see Note 3). 7. 1 M imidazole (stock solution): Weigh 34.04 g of imidazole, dissolve them with about 450 mL of deionized water, then adjust the pH to 7.5 with HCl, and finally dilute to 500 mL with deionized water. 8. Lysozyme. 9. HisTrap HP column (5 mL; GE Healthcare, Uppsala, Sweden).

288

Sheng-Ying Qin et al.

10. SDS-PAGE (12.5% polyacrylamide gel). 11. 6 solubilizing buffer: 375 mM Tris–HCl, pH 6.8, containing 12% SDS, 0.6% bromophenol blue, and 60% glycerol (see Note 4). 12. 50 mM sodium phosphate buffer, pH 7.4: 50 mM Na2HPO4 solution and 50 mM KH2PO4 solution in a certain ratio to make pH to 7.4 (see Note 5). 2.3 Biotinylation of sMAL and Tetramer Preparation for Cell Binding Assay

1. BirA ligase (Avidity, Denver, CO). 2. Biotin. 3. HisTrap HP column. 4. Streptavidin (SA) (BD Biosystems, San Jose, CA). 5. PBS(): Weigh 8.01 g of NaCl and 0.20 g KCl, take 200 mL of mixture of 50 mM sodium phosphate buffer, pH 7.4, dissolve to a volume of 1 L with deionized water (see Note 6). 6. 6 solubilizing buffer: 375 mM Tris–HCl, pH 6.8, containing 12% SDS, 0.6% bromophenol blue, and 60% glycerol. 7. SDS-PAGE (12.5% polyacrylamide gel). 8. R-phycoerythrin-labeled streptavidin (BD Biosystems, San Jose, CA).

(PE-SA)

9. HeLaS3 cells are maintained in RPMI1640. 10. RPMI1640 medium. 11. 0.25% trypsin. 12. Phosphate buffer saline (PBS) (stock solution of 10 PBS): Weigh 80 g of NaCl, 2.0 g of KCl, 14.2 g of Na2HPO4, and 2.7 g of KH2PO4, dissolve them with about 800 mL of deionized water, then adjust the pH to 7.5 with HCl, and finally dilute to 1 L with deionized water, autoclave and store at room temperature. 13. Fetal calf serum. 14. Deoxynojirimycin (DNJ). 15. HEPES-buffered saline (HBS): 20 mM HEPES-NaOH, pH 7.4, 136 mM NaCl, 4.7 mM KCl, 0.2% BSA, 1 mM CaCl2, and 0.1% NaN3. Weigh 4.8 g of HEPES, 7.9 g of NaCl, 0.4 g of KCl, 2 g of BSA, 0.1 g of CaCl2, and 1 g of NaN3, dissolve them with about 900 mL of deionized water, then adjust the pH to 7.4 with NaOH, and finally dilute to 1 L with deionized water, autoclave and store at room temperature. 16. FACScalibur system equipped with CELLQuest software: BD Biosciences, San Jose, CA. 17. Nigerose (Wako).

Preparation of Recombinant Malectin

3

289

Methods

3.1 Construction of Plasmids for sMAL Expression in E. coli

1. Phosphorylatin of primers (Table 1) is carried out using T4 polynucleotide kinase kit. Set up a 10 μL reaction with 1 μL of 10  T4 polynucleotide kinase buffer, 1 μL of 10 mM ATP, 4 μL of 50 μM primer, 0.5 μL of 10 U/μL T4 polynucleotide kinase, and 3.5 μL H2O. Carry out the reaction at 37  C for 2 h and then denature at 70  C for 10 min. Store at 4  C before use. 2. The full-length cDNA of human malectin (NM_014730.2) is amplified by PCR from a cDNA library of human HeLaS3 cells using phosphorylated primers: Malectin-F 50 -CGTGGCGCTGTTTTTCTGAGTCC-30 and Malectin-R 50 -TTTCTTTCCCACACCCCTCCACC-30 . Set up a 50 μL volume with 5 μL of 10 KOD plus buffer, 5 μL of 2 mM dNTPs, 2 μL of 25 mM MgSO4, 1 μL of 1 U/μL KOD plus polymerase, 0.5 μL of cDNA, 1 μL of 20 μM phosphorylate Mal-F, 1 μL of 20 μM phosphorylate Mal-R, and 34.5 μL H2O. The conditions for cycling are: one cycle at 94  C for 2 min, followed by 30 cycles at 94  C for 15 s, 60  C for 30 s, and 68  C for 60 s, then one cycle at 68  C for 7 min. The amplified products are purified using QIAquick Gel Extraction kit and ligated into the SmaI site of pBluescript II SK (+) to generate pBlue-malectin using DNA Ligation kit. 3. Using pBlue-malectin as a template, the coding sequence for the luminal domain of malectin (residues 42-228) is amplified by PCR using the primers: malectin-domain-F 50 -GGGAATTCCATATGGCCGGGCTGCCCGAGAGC-30 (NdeI restriction site is shown in Italic) and phosphorylated malectin-domain-R 50 -CTCCAATCCCGGATGAGGCTG-30 . Set up a 50 μL volume with 5 μL of 10  KOD plus buffer, 5 μL of 2 mM dNTPs, 2 μL of 25 mM MgSO4, 1 μL of 1 U/μL KOD plus polymerase, 1 μL of 10 ng/μL pBlue-malectin, 1 μL of 50 μM F primer, 1 μL of 50 μM R primer, and 34 μL

Table 1 Oligonucleotides used to construct plasmids for sMAL-Bio expression in E. coli Primer name

Sequence

Malectin-F

50 -CGTGGCGCTGTTTTTCTGAGTCC-30

Malectin-R

50 -TTTCTTTCCCACACCCCTCCACC-30

Malectin domain-F

50 -GGGAATTCCATATGGCCGGGCTGCCCGAGAGC-30

Malectin domain-R

50 -CTCCAATCCCGGATGAGGCTG-30

AviT-EcoRI-R

50 -CGGAATTCTTATTCGTGCCATTCGATTTTC-30

290

Sheng-Ying Qin et al.

H2O. The conditions for cycling are: one cycle at 94  C for 2 min, followed by 30 cycles at 94  C for 15 s, 60  C for 30 s, and 68  C for 50 s, then one cycle at 68  C for 7 min. Amplified products are purified using QIAquick Gel Extraction kit followed by treatment with NdeI. The resulting product is then ligated into the NdeI and SmaI sites of pET3Cbio with a Cterminal biotinylation Avi-Tag to generate pET3CBio-sMAL. 4. Using pET3CBio-sMAL as a template, the coding sequence for both luminal domain of malectin and Avi-Tag is amplified by PCR using the primers: malectin-domain-F 50 -GGGAATTCCATATGGCCGGGCTGCCCGAGAGC-30 (NdeI restriction site is shown in Italic) and AviT-EcoRI-R 50 -CGGAATTCTTATTCGTGCCATTCGATTTTC-30 (EcoRI restriction site is shown in Italic). Set up a 50 μL volume with 5 μL of 10 KOD plus buffer, 5 μL of 2 mM dNTPs, 2 μL of 25 mM MgSO4, 1 μL of 1 U/μL KOD plus polymerase, 1 μL of 10 ng/μL pET3CBio-sMAL, 1 μL of 50 μM F primer, 1 μL of 50 μM R primer, and 34 μL H2O. The conditions for cycling are: one cycle at 94  C for 2 min, followed by 30 cycles at 94  C for 15 s, 60  C for 30 s, and 68  C for 60 s, then one cycle at 68  C for 7 min. The amplified products are purified using QIAquick Gel Extraction kit followed by digestion with NdeI and EcoRI and then inserted into the NdeI and EcoRI sites of pColdI to generate pColdI-sMAL-Bio (Fig. 2). HelaS3 cDNA library 1

26 42

228

269

290

Malectin

pBlue-Malectin

pET3cBio-Malectin

pColdI-Malectin-Bio

signal peptide

biotinylation sequence

lectin domain

Factor Xa site

transmembrane domain

His6 tag

Fig. 2 Schematic representation of plasmid construction for sMAL-bio expression in E. coli

Preparation of Recombinant Malectin

1. BL21 (DE3) pLysS E. coli transformed with pColdI-sMAL-Bio is grown at 37  C in LB medium containing 100 μg/mL ampicillin and 25 μg/mL chloramphenicol (see Note 7). 2. When the OD600 of the culture reached about 0.5, the cells are cooled at 15  C for 30 min and then add IPTG to a final concentration of 1 mM. 3. The culture is incubated for an additional 24 h at 15  C. After that, the cells are harvested by centrifugation at 5000  g for 15 min at 4  C. 4. The cell pellets are washed with PBS and then resuspended in PBS containing 1 mM PMSF and 10 mM imidazole. 5. The cell suspension is frozen at 80  C and then thawed in cool water. 6. Immediately after the sample is completely thawed, lysozyme is added to a final concentration of 100 μg/mL. 7. Cells are subjected to sonication (level 5; eight times for 20 s each) and then centrifuged at 8000  g for 15 min (see Note 8). 8. The supernatant is collected and applied to a HisTrap HP column equilibrated with PBS containing 10 mM imidazole. The column is washed with PBS containing 10 mM imidazole and then eluted with a linear gradient of imidazole from 10 mM to 500 mM in PBS (25 mL total). Elution is performed at a rate of 1 mL/min and fractions of 1.0 mL are collected.

94 kDa 66 kDa 42 kDa 30 kDa 20 kDa 14 kDa

Fig. 3 Purification of sMAL by HisTrap HP column

B4

B5

B6

B7

B8

B9

B10

B13

A15

FT

9. Fractions containing sMAL are confirmed by SDS-PAGE and pooled together (Fig. 3).

Input

3.2 Expression and Purification of sMAL

291

bio-sMAL+SA

sMAL+SA

bio-sMAL

sMAL

Sheng-Ying Qin et al.

SA

292

94 kDa 66 kDa 42 kDa

30 kDa 20 kDa 14 kDa

Fig. 4 Confirmation of biotinylation of sMAL effect. The samples SA, sMAL, and bio-sMAL were directly loaded to start electrophoresis analysis. The samples sMAL+SA and bio-sMAL+SA were respectively mixed sMAL and bio-sMAL with SA and placed at 4  C for 30 min, before loading to start electrophoresis analysis

10. To confirm the folding status of recombinant sMAL, it is subjected to CD spectra analysis using a Jasco J-725 spectropolarimeter at room temperature. Sample contained 0.1–0.2 mg/mL of protein is dissolved in 50 mM sodium phosphate buffer, pH 7.4. Spectrum is recorded as the average of 20 scans over a range of 190–250 nm. 3.3 Biotinylation of sMAL and Tetramer Preparation for Cell Binding Assay

1. Purified sMAL of 4.5 mg is biotinylated with biotin ligase BirA following the manufacturer’s protocol and free biotin is removed using a HisTrap HP column. 2. Biotinylation is validated with a gel shift assay in polyacrylamide gels using SA as follows. One microgram of sMAL is mixed with SA at a molar ratio of 1:2 at 4  C for 1 h in 10 μL of 10 mM sodium phosphate, pH 7.4, containing 137 mM NaCl and 2.68 mM KCl [PBS()]. The samples are added to 2 μL of 6 solubilizing buffer and electrophoresed in 12.5% polyacrylamide gels (Fig. 4). 3. After successful biotinylation of sMAL, 10 mg/mL bio-sMAL is mixed with PE-SA at a molar ratio of 4:1 and then incubated on ice for 30 min to form a bio-sMAL-SA-PE complex.

Preparation of Recombinant Malectin

293

4. HeLaS3 cells treated with the α-glucosidase inhibitor deoxynojirimycin (DNJ) for 20 h are harvested and then suspended in HEPES-buffered saline (HBS) at a concentration of 2  107 cells/mL. 5. An aliquot of the cell suspension (10 μL) is incubated with 10 μL of sMAL complexed with bio-sMAL-SA-PE at the indicated concentrations at room temperature for 30 min. 6. After washing with HBS, cells are resuspended in 200 μL of HBS containing 0.5% paraformaldehyde and then analyzed by flow cytometry using a FACScalibur system equipped with CELLQuest software. 7. For disaccharide inhibition assays, nigerose is added to the cells during incubation with bio-sMAL-SA-PE.

4

Notes 1. NaOH is used to adjust the pH value of solution to 8.0, so the exact amount of NaOH is determined according to the specific conditions. Generally, about 20 mg of NaOH is required to make 1 L 0.5 M of EDTA. 2. Dilute ten times to 1 PBS with deionized water and autoclave before use. 3. The PMSF stock solution taken from 20  C is not easy to dissolve. Be patient and wait until the precipitate is completely dissolved. 4. 1 M Tris–HCl is used to dilute to 375 mM with deionized water. 1 M Tris–HCl, pH 6.8 (stock solution): Weigh 121.14 g of Tris, dissolve it with about 800 mL of deionized water, then adjust the pH to 6.8 with HCl, and finally dilute to 1 L with deionized water, autoclave and store at room temperature. 5. 50 mM Na2HPO4 solution: Dissolve 7.1 g of Na2HPO4 in deionized water and adjust the final volume to 1 L. 50 mM KH2PO4 solution: Dissolve 6.8 g of KH2PO4 in deionized water and adjust the final volume to 1 L. 6. 50 mM sodium phosphate buffer, pH 7.4: 50 mM Na2HPO4 solution and 50 mM KH2PO4 solution in a certain ratio to make pH to 7.4. 7. Rosseta-Gami (DE3) E. coli can also be used, and the method and result are the same as using BL21 (DE3) pLysS E. coli. 8. Sonication needs to be repeated several times until the crude extract becomes clear and transparent.

294

Sheng-Ying Qin et al.

Acknowledgments This work was supported by National Natural Science Foundation of China. [81602436], Natural Science Foundation of Guangdong Province [2016A030313102], and the Fundamental Research Funds for the Central Universities [21617495]. References 1. Schallus T, Jaeckh C, Fehe´r K et al (2008) Malectin: a novel carbohydrate-binding protein of the endoplasmic reticulum and a candidate player in the early steps of protein N-glycosylation. Mol Biol Cell 19:3404–3414 2. Chen Y, Hu D, Yabe R et al (2011) Role of Malectin in Glc2Man9GlcNAc2-dependent quality control of α1-antitrypsin. Mol Biol Cell 22:3559–3570 3. Galli C, Bernasconi R, Solda` T et al (2011) Malectin participates in a backup glycoprotein quality control pathway in the mammalian ER. PLoS One 6:e16304 4. Qin SY, Hu D, Matsumoto K et al (2012) Malectin forms a complex with ribophorin I for enhanced association with misfolded glycoproteins. J Biol Chem 287:38080–38089

5. Takeda K, Qin SY, Matsumoto N, Yamamoto K (2014) Association of malectin with ribophorin I is crucial for attenuation of misfolded glycoprotein secretion. Biochem Biophys Res Commun 454:436–440 6. Yang QP, Fu MF, Gao H et al (2018) Subcellular distribution of endogenous malectin under rest and stress conditions is regulated by ribophorin I. Glycobiology 28(6):374–381 7. Kawasaki N, Matsuo I, Totani K et al (2007) Detection of weak sugar binding activity of VIP36 using VIP36-streptavidin complex and membrane-based sugar chains. J Biochem 141:221–229 8. Yamamoto K (2014) Intracellular lectins are involved in quality control of glycoproteins. Proc Jpn Acad Ser B Phys Biol Sci 90:67–82

Chapter 29 Calnexin/Calreticulin and Assays Related to N-Glycoprotein Folding In Vitro Yoshito Ihara, Midori Ikezaki, Maki Takatani, and Yukishige Ito Abstract Calnexin (CNX) and calreticulin (CRT) are ER-resident lectin-like molecular chaperones involved in the quality control of secretory or membrane glycoproteins. They can exert molecular chaperone functions via specific binding to the early processing intermediates of Glc1Man9GlcNAc2 oligosaccharides of N-glycoproteins. CNX and CRT have similar N-terminal luminal domains and share the same jelly roll tertiary structure as legume lectins. In addition to the lectin-like interactions, CNX and CRT also suppress the aggregation of non-glycosylated substrates through interaction with hydrophobic peptide parts, suggesting a general chaperone function in glycan-dependent and glycan-independent manners. This chapter describes the isolation and purification of CRT produced in a bacterial expression system. We also introduce in vitro assays to estimate the molecular chaperone functions of CRT via the interaction with monoglucosylated Nglycans using Jack bean α-mannosidase as a target substrate. These assays are valuable in assessing quality control events related to the CNX/CRT chaperone cycle and lectin functions. Key words Aggregation, Calnexin, Calreticulin, Chaperone, Endoplasmic reticulum, α-Mannosidase, N-glycan

1

Introduction N-glycans of glycoproteins are initially synthesized in the endoplasmic reticulum (ER) by the stepwise addition of sugars to the lipid dolichylphosphate to form the lipid-linked oligosaccharide Glc3Man9GlcNAc2-PP-dolichol [1]. The oligosaccharide is then transferred to specific asparagine residues at the consensus sites for N-glycosylation in newly synthesized polypeptides in the ER. After the transfer, α-glucosidases I and II mediate removal of the terminal glucose residues from the precursor oligosaccharides. Then, a specific structure of monoglucosylated oligosaccharide, Glc1Man5-9GlcNAc2, in glycoproteins is recognized by the

This work was supported by a Ministry of Education, Culture, Sports, Science and Technology of Japan Grant-inAid for Scientific Research (JP16H06290). Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_29, © Springer Science+Business Media, LLC, part of Springer Nature 2020

295

296

Yoshito Ihara et al.

chaperone cycle of calnexin (CNX) or calreticulin (CRT) (i.e., CNX/CRT cycle), by which the proper folding and quality of glycoproteins are controlled in the ER [2–4]. In the proper folding process, the terminal glucose is removed by α-glucosidase II, and deglucosylated proteins are transported from the ER to Golgi apparatus. If proper folding is not achieved, the deglucosylated high-mannose type N-glycans (e.g., Man9GlcNAc2) are again re-glucosylated by UDP-glucose: glycoprotein glucosyltransferase (UGGT), a sensor of immature glycoprotein substrates, to return them to the CNX/CRT cycle [5]. Nevertheless, if proper folding is still not completed, deglucosylated glycans of the substrates are further trimmed by ER α-mannosidases (i.e., ER mannosidase I or EDEMs). Consequently, the immature misfolded substrates are degraded by ER-associated degradation (ERAD) via the recognition of the high-mannose type N-glycans (e.g., Man7-5GlcNAc2) with other ERAD-related lectin-like molecules, such as OS9 and XTP3-B [3]. CNX is a type I membrane-bound lectin-like chaperone in the ER and also has various non-chaperone functions in the cell [2]. CRT is a soluble ER-resident protein paralogous to CNX [6]. Mammalian CNX and CRT share approximately 40% amino acid sequence identity in the structures composed of the aminoterminal globular domain (N-domain), proline-rich hairpin-like domain (P-domain), and carboxy-terminal domain (C-domain). CNX and CRT also share similar chaperone functions for newly synthesized N-glycoproteins in the ER, and a high-affinity binding capacity for the monoglucosylated N-glycans on the proteins is a common characteristic of both molecules [7–9]. In recent years, the structures of CNX and CRT have been extensively investigated, and the results revealed that the lectin sites, at which the monoglucosylated oligosaccharide is bound, are located in the globular N-domain, and the structures are thought to be similar between CNX and CRT [10–12]. The predicted structure of the lectin site consists of a jelly roll fold formed by a sandwich of one convex and one concave antiparallel β-sheet. The β-sandwich structure is characteristic of legume lectins, such as Con A and PHA-L, and they are called L-type lectins. Thus, CNX and CRT are also thought to be members of the L-type lectin family due to their similarity with legume lectins [13]. Peptide-binding functions of CNX and CRT were also reported in various models in vitro [14–16] and in vivo [17], leading to the proposal of dual carbohydrate-/peptide-binding properties of CNX and CRT [18]. Several studies indicated that the peptide-binding site is located close to the lectin site of the N-domain, and alternative usage of the site may be controlled in a substrate-dependent manner, modified by a variety of factors, such as Ca2+, Zn2+, and temperature [10, 11, 19–23]. Meanwhile, the P-domain constitutes an extended region stabilized by three or four

Chaperone Functions of Calnexin/Calreticulin

297

antiparallel β-strands, and it forms a long hairpin-like structure that interacts with other chaperone-related molecules, such as ERp57 and ERp29, and cyclophilin B [24–27]. The C-domain contains highly acidic amino acid sequences with high-capacity and low-affinity Ca2+-binding regions to exert the Ca2+-buffering functions of the protein. In CNX, the main functional part consisting of the globular N-plus P-domain is localized inside the ER, and the C-terminal transmembrane domain is anchored in the ER membrane. In contrast, CRT is localized in the ER as a soluble protein, by utilizing its C-terminal KDEL sequence for ER retrieval. In addition to the chaperone functions for N-glycoproteins, CRT is involved in a variety of biological processes including the regulation of Ca2+ homeostasis and intracellular signaling, cell adhesion, gene expression, nuclear transport, carcinogenesis, and cancer immunity [6, 28–30]. Therefore, it is interesting to know how the lectin-based interactions cooperate with other functions of CRT or CNX in the cell. In this chapter, we focused on assays to estimate the lectin-related functions of CRT. Thus, this chapter describes functional assays of CRT using bacterially expressed recombinant CRT. The assays assess the molecular chaperone functions of CRT against N-glycosylated or non-glycosylated substrate proteins in vitro.

2

Materials

2.1 Expression of CRT in Bacteria

1. Bacterial expression vector for glutathione-S-transferase fused with mouse CRT (pGEX-GST-mCRT) [31] (see Notes 1–5). 2. Luria-Bertani (LB) broth: per liter, 10 g tryptone (BD Difco), 5 g yeast extract (BD Difco), 5 g NaCl, pH to 7.5 with NaOH—use deionized water and sterilize by autoclaving. 3. LB broth supplemented with 0.1 mg/mL ampicillin. 4. LB agar: LB broth with 15 g/L Bacto agar (BD Difco). 5. LB agar plate (1.5% w/v) supplemented with 0.1 mg/mL ampicillin. 6. Isopropyl β-D-thiogalactopyranoside (IPTG, Nacalai Tesque Inc., Kyoto, Japan): Stock solution (100 mM) is prepared in water, sterilized by filtration (0.22 μm), and stored at 20  C. 7. Ampicillin sodium salt (Nacalai Tesque Inc.): Stock solution (50 mg/mL) is prepared in water, sterilized by filtration (0.22 μm), and stored at 20  C. 8. Syringe-driven filter (e.g., Millex HA filter unit, 0.45 μm, Merck Millipore). 9. Drying autoclave. 10. Incubator shaker (e.g., BIO-Shaker 40LF, TAITEC, Japan).

298

Yoshito Ihara et al.

11. Cell disruptor and homogenizer (e.g., Branson Sonifier). 12. High-speed refrigerated centrifuge. 13. Spectrophotometer. 2.2 Purification of CRT 2.2.1 Glutathione Sepharose Column Chromatography

1. Glutathione Sepharose 4 Fast Flow (GE Healthcare). 2. Tris-buffered saline (TBS, 20 mM Tris–HCl [pH 7.5] and 150 mM NaCl). 3. Nonidet P-40 (NP-40, Nacalai Tesque Inc.). 4. Reduced glutathione solution (10 mM) prepared in TBS just before use (see Note 6). 5. Pierce Coomassie (Bradford) protein assay kit (Thermo Fisher Scientific). 6. Econo-column chromatography column (1.5  15 cm, BioRad). 7. NAP-5 column (GE Healthcare). 8. Amicon Ultra-15 10 K device (Merck Millipore). 9. RediFrac fraction collector (GE Healthcare). 10. Systems for SDS-polyacrylamide gel electrophoresis and Coomassie staining. 11. High-speed refrigerated centrifuge. 12. Spectrophotometer.

2.2.2 Cleavage of GST-CRT Fusion Protein

1. PreScission protease (cat. no. 27484301, GE Healthcare) 2000 units/mL. 2. PreScission cleavage buffer: 50 mM Tris–HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, and 1 mM dithiothreitol (DTT). 3. Glutathione Sepharose 4 Fast Flow (GE Healthcare). 4. Tris-buffered saline (TBS, 20 mM Tris–HCl [pH 7.5] and 150 mM NaCl). 5. Amicon Ultra-15 10 K device (Merck Millipore). 6. RediFrac fraction collector (GE Healthcare). 7. Cold storage chamber or refrigerator. 8. Systems for SDS-polyacrylamide gel electrophoresis and Coomassie staining. 9. High-speed refrigerated centrifuge.

2.2.3 Mono Q Anion-Exchange Liquid Chromatography

1. Mono Q 4.6/100 PE (GE Healthcare). 2. MQ buffer A: 20 mM Tris–HCl (pH 7.5), 100 mM NaCl, and 0.5 mM CaCl2. 3. MQ buffer B: 20 mM Tris–HCl (pH 7.5), 1000 mM NaCl, and 0.5 mM CaCl2.

Chaperone Functions of Calnexin/Calreticulin

299

4. Pierce Coomassie (Bradford) protein assay kit (Thermo Fisher Scientific). 5. Amicon Ultra-4 10 K device (Merck Millipore). 6. AKTA pure 25 chromatography system (GE Healthcare). 7. Systems for SDS-polyacrylamide gel electrophoresis and Coomassie staining. 8. High-speed refrigerated centrifuge. 9. Spectrophotometer. 2.3 Aggregation Suppression Assay

1. Purified CRT. 2. Jack bean α-mannosidase (cat. no. M-7257, Sigma-Aldrich, Inc., MO, USA). 3. Endoglycosidase H (Endo H) (cat. no. P-0702S, New England Biolab.) 500,000 units/mL. 4. Oligosaccharides or their derivatives (e.g., Glc1Man4-biotin [Glc-α1,3-Man-α1,2-Man-α1,2-Man-α1,3-Man-biotin]) provided by Dr. Yukishige Ito (RIKEN, Japan). 5. Denaturation buffer: 100 mM Tris–HCl (pH 8.0) and 6 M guanidine hydrochloride. 6. Glycoprotein denaturation buffer: 0.5% SDS and 40 mM DTT. 7. 10 Endo H reaction buffer: 500 mM sodium phosphate (pH 6.0). 8. Aggregation buffer: 10 mM Hepes (pH 7.5), 150 mM NaCl, and 5 mM CaCl2. 9. NAP-5 column (GE Healthcare). 10. Systems for vacuum concentrator or freeze dryer (lyophilizer). 11. UV-VIS spectrophotometer equipped with thermal controller (e.g., UV-2600 plus TCC-100, Shimadzu, Japan).

2.4 Heat-Induced Inactivation of Jack Bean α-Mannosidase

1. Purified CRT. 2. Immunoglobulin G (Rabbit IgG, cat. no. 2729S, Cell Signaling Technology). 3. Jack bean α-mannosidase (cat. no. M-7257, Sigma-Aldrich). 4. 2 Inactivation buffer: 20 mM Hepes (pH 7.5), 100 mM NaCl, and 2 mM CaCl2. 5. Heating block (e.g., Dry ThermoUnit, DTU-1B, TAITEC, Japan).

2.5 The Enzyme Assay for α-Mannosidase

1. p-Nitrophenyl α-D-mannopyranoside (cat. no. N2127, SigmaAldrich) dissolved in dimethyl sulfoxide (DMSO) at 200 mM (i.e., 40 Substrate stock solution).

300

Yoshito Ihara et al.

2. 2 Reaction buffer: 100 mM citrate-phosphate buffer (CPB, pH 4.7). 3. 5 Stop solution: 2 M Tris–HCl (pH 8.8). 4. Spectrophotometer.

3

Methods

3.1 Expression of Mammalian CRT in Bacteria

1. Transform Escherichia coli (E. coli) cells (e.g., BL21 DE3) with pGEX-GST-mCRT (see Subheading 2.1, item 1). 2. Spread the transformed cells on LB agar plates supplemented with ampicillin (0.1 mg/mL), and then cultivate overnight (see Subheading 2.1, items 2–5). 3. Grow a single colony overnight in LB broth medium containing ampicillin (0.1 mg/mL) at 37  C on a shaker (see Subheading 2.1, item 10). 4. Inoculate fresh 500-mL culture with the cells. 5. Grow the cells in the medium at 37  C on a shaker until the absorbance at 600 nm reaches 0.4. 6. At this point, add IPTG (see Subheading 2.1, item 6) to the culture medium (final concentration of 0.5 mM) to induce the expression of GST-CRT fusion protein. 7. Keep on growing the cells by culturing for an additional 4 h at 37  C on the shaker. 8. Harvest the cells by centrifugation at 2500  g for 15 min at 4  C. 9. Suspend the cell pellet in 30 mL of TBS (pH 7.5) plus 1% NP-40. 10. Solubilize the cell suspension with a tip-type cell disruptor or sonicator (30 s  6) (see Subheading 2.1, item 11) on ice. 11. Recover the cell supernatant (i.e., bacterial extract) by centrifugation at 12,000  g for 20 min at 4  C, and then filter using a syringe filter unit (0.45 μm) (see Subheading 2.1, item 8).

3.2 Purification of Mammalian CRT Expressed in Bacteria 3.2.1 Glutathione Sepharose Column Chromatography

Recombinant GST-CRT protein is purified from the bacterial extract using a glutathione Sepharose column (10 mL) (see Subheading 2.2.1, item 1). 1. Load the sample onto the glutathione Sepharose in an Econocolumn chromatography column (1.5  15 cm, BioRad) pre-equilibrated with TBS (pH 7.5) plus 1% NP-40 (see Subheading 2.2.1, items 2 and 3). 2. Wash the column with 50 mL of TBS plus 1% NP-40, and then 100 mL of TBS at a flow rate of 0.5 mL/min.

Chaperone Functions of Calnexin/Calreticulin

301

3. Elute bound GST-CRT protein with 50 mL of 10 mM reduced glutathione in TBS (see Subheading 2.2.1, item 4 and Note 6) at a flow rate of 0.5 mL/min. Collect the eluted fractions using a fraction collector (1 mL/fraction). The proteins in the fractions are monitored by measuring the absorbance at 280 nm. 4. Confirm the purity of GST-CRT in the fractions by SDS-PAGE and Coomassie staining. 5. Collect the fractions containing GST-CRT protein and concentrate them to a volume of less than 0.5 mL using an Amicon Ultra-15 10 K device according to the manufacturer’s instructions. 6. To remove excess glutathione, concentrated GST-CRT is desalted with a NAP-5 column pre-equilibrated with TBS according to the manufacturer’s instructions. 7. The protein concentration is determined using a Bradford protein assay kit (see Subheading 2.2.1, item 5). 3.2.2 Cleavage of GST-CRT Fusion Protein

GST-CRT fusion protein is cleaved with PreScission protease according to the manufacturer’s instructions. 1. Mix an aliquot of 1 μL (2 units) of PreScission protease with 100 μg of tagged protein in 1 mL of the PreScission cleavage buffer (see Subheading 2.2.2, item 2) in a 1.5-mL sample tube. 2. Incubate at 4  C for 24 h or more. 3. Check the cleavage of GST-CRT with a small aliquot of the sample by SDS-PAGE and Coomassie blue staining. 4. Once digestion is complete, load the sample onto glutathione Sepharose (10 mL) (see Subheading 2.2.2, item 3) in an Econo-column pre-equilibrated with TBS (pH 7.5) (see Subheading 2.2.2, item 4), to remove the GST moiety of the tagged CRT and PreScission protease from the CRT protein. 5. Collect the pass-through fractions containing CRT protein using the fraction collector (1 mL/fraction). 6. Concentrate the fraction samples to a volume of less than 1 mL using an Amicon Ultra-15 10 K device according to the manufacturer’s instructions, to change the buffer composition to TBS.

3.2.3 Mono Q Anion-Exchange Chromatography

The CRT protein sample is further purified by a Mono Q anionexchange liquid chromatography column (5 mL) (see Subheading 2.2.3, item 1) in an AKTA pure 25 chromatography system under the control of UNICORN 6.4 software (GE Healthcare). 1. Load the sample (0.5–1 mL) onto the column pre-equilibrated with MQ buffer A (see Subheading 2.2.3, item 2).

302

Yoshito Ihara et al.

2. Wash the column with 50 mL of MQ buffer A at a flow rate of 1.0 mL/min. 3. Elute CRT using 8 mL of the buffer with a linear 100 to 600 mM NaCl gradient, and 8 mL of the buffer with 600 mM NaCl isocratic, then the NaCl concentration is increased to 1000 mM at a time, and 8 mL of the buffer with 1000 mM NaCl isocratic in 20 mM Tris–HCl (pH 7.5) and 0.5 mM CaCl2 (MQ buffer B) (see Subheading 2.2.3, item 3) at a flow rate of 0.5 mL/min. The proteins in the fractions are monitored by measuring the absorbance at 280 nm. 4. Confirm the purity of CRT in the fractions by SDS-PAGE and Coomassie staining. 5. Collect the fractions containing pure CRT and concentrate them to a volume of less than 1 mL using an Amicon Ultra-4 10 K device, and the protein concentration is determined using a Bradford protein assay kit, as described (see Notes 7 and 8). 3.3 Aggregation Suppression Assay

1. Jack bean α-mannosidase (3.4 mg/mL in ammonium sulfate suspension) (see Notes 9–11) is diluted 10 times with cold water and then desalted by the NAP-5 column pre-equilibrated with TBS. The desalted α-mannosidase sample (170 μg) is lyophilized by a vacuum concentrator or freeze dryer, dissolved at a concentration of 24 μM in the denaturation buffer containing 6 M guanidine hydrochloride (see Subheading 2.3, item 5), and then denatured at room temperature for 60 min. 2. Alternatively, the lyophilized α-mannosidase as above (85 μg) is dissolved in 45 μL of the glycoprotein denaturation buffer (see Subheading 2.3, item 6). After 5 min boiling, an aliquot of 5 μL of 10 Endo H reaction buffer (see Subheading 2.3, item 7) is added, and incubated with 2 units of Endo H (see Subheading 2.3, item 3) at 37  C overnight for the removal of monoglucosylated high-mannose type N-glycans. The Endo H-treated α-mannosidase is precipitated by adding 300 μL of cold ethanol. After 1 h of incubation on ice, the sample is centrifuged at 10,000  g for 15 min at 4  C. Then, the precipitated α-mannosidase sample is dried by a vacuum concentrator, dissolved at a concentration of 24 μM in the denaturation buffer containing 6 M guanidine hydrochloride, and denatured at room temperature for 60 min as above. 3. Add 6.25 μL of denatured α-mannosidase to the bottom of a crystal cuvette and dilute very rapidly with 493.75 μL of the aggregation buffer (see Subheading 2.3, item 8) containing various amounts of CRT or immunoglobulin G (IgG) (negative control). Alternatively, other agents such as oligosaccharides (e.g., Glc1Man4-biotin) may be included in the

Chaperone Functions of Calnexin/Calreticulin

303

aggregation assay mixture. The final concentration of denatured α-mannosidase is 0.3 μM, and molar ratios of α-mannosidase/CRT typically used are 1:1 to 1:8 (see Note 11). 4. Samples are prepared in duplicate and aggregation is induced by incubating at 30  C under the thermal controller and monitored by measuring light scattering at 360 nm at 1-min intervals over 60 min using a spectrophotometer (see Subheading 2.3, item 11). 5. This assay has been used to demonstrate the ability of CRT to suppress the aggregation of monoglucosylated or deglycosylated glycoproteins [15, 32]. Examples of the aggregation assay results are shown in Fig. 1. In Fig. 1c, Glc1Man4-biotin was added to the aggregation assay mixture to examine its effect on the lectin-mediated chaperone function of CRT against unfolded α-mannosidase (see Note 12). 3.4 Heat-Induced Inactivation of Jack Bean α-Mannosidase

1. Jack bean α-mannosidase (in ammonium sulfate suspension) is diluted 10 times with cold water, and then the concentration is adjusted to 0.2 μM in the inactivation buffer (final concentration: 10 mM Hepes (pH 7.5), 50 mM NaCl, and 1 mM CaCl2) (see Subheading 2.4, items 3, 4, and Note 11). 2. Mix an aliquot of 100 μL of the diluted α-mannosidase solution with 100 μL of the inactivation buffer containing various amounts of CRT or IgG (negative control) in a 1.5-mL sample tube. Alternatively, other agents such as oligosaccharides (e.g., Glc1Man4-biotin) may be included in the inactivation assay mixture. The final concentration of α-mannosidase is 0.1 μM, and molar ratios of α-mannosidase/CRT typically used are 1:1 to 1:8 (see Note 11). 3. Incubate the mixture solution at 60 inactivation.



C for 60 min for

4. An aliquot of 10 μL of the solution is taken and subjected to the α-mannosidase enzyme assay (see Subheading 3.5). 3.5 The Enzyme Assay for α-Mannosidase

The enzyme assay for α-mannosidase is performed according to the method of Li Y.-T. (1967) [33] with a slight modification. 1. An aliquot of 390 μL of the α-mannosidase assay buffer (final concentration: 50 mM CPB and 5 mM p-nitrophenyl α-Dmannopyranoside) is prepared by mixing 10 μL of 40 substrate stock solution (see Subheading 2.5, item 1), 200 μL of 2 reaction buffer (see Subheading 2.5, item 2), and 180 μL of water (see Notes 13 and 14). 2. Add 10 μL of the sample aliquot to the assay buffer. 3. Incubate the mixture at 25  C for 5 min.

304

Yoshito Ihara et al.

Fig. 1 Effects of CRT on aggregation or thermal inactivation of Jack bean α-mannosidase. (a) Chemically denatured Jack bean α-mannosidase (JBM) (0.3 μM) was mixed with different amounts of CRT at the ratios indicated in the figure, and then the mixture was incubated at 30  C for 60 min. Aggregation was monitored by measuring light scattering at 360 nm as described in Methods. CRT exerted a suppressive effect on the aggregation of JBM. (b) After treatment with Endo H, deglycosylated JBM (dJBM) (0.3 μM) was chemically denatured, mixed with CRT at the ratios indicated in the figure, and then the mixture was incubated at 30  C for 60 min. CRT showed a weaker effect on the aggregation of dJBM. (c) Chemically denatured JBM (0.3 μM) was mixed with CRT (0.6 μM) in the presence or absence of Glc1Man4-biotin (G1M4, 20 or 100 μM), and then the mixture was incubated at 30  C for 60 min. Monoglucosylated oligosaccharides of G1M4 interfered with the chaperone function of CRT. (d) Desalted JBM (0.1 μM) was mixed with CRT or IgG (0.8 μM) in the inactivation buffer. Alternatively, JBM (0.1 μM) was mixed with CRT (0.8 μM) plus G1M4 (100 μM). Then, the mixtures were incubated at 60  C for 60 min. The aliquots were subjected to the α-mannosidase assay as described in Methods. The activities are indicated as values relative to the initial activity of untreated JBM (100%). CRT protected JBM against heat-induced inactivation, but it was suppressed with G1M4, suggesting the lectin-dependent effect of CRT on the thermal inactivation of JBM

4. Add 100 μL of 5 stop solution (see Subheading 2.5, item 3). 5. Measure the absorbance at 405 nm using a spectrophotometer. 6. Examples of the inactivation assay results are shown in Fig. 1d (see Note 15).

Chaperone Functions of Calnexin/Calreticulin

4

305

Notes 1. All reagents and chemicals are of the highest grade available. 2. Ultrapure water is used for biochemical assays. 3. Bacterial expression vector for glutathione-S-transferase fused with mouse calreticulin (pGEX-GST-mCRT): CRT gene construct corresponding to the coding region lacking the N-terminal signal sequence was generated by PCR with PfuTurbo DNA polymerase (Agilent Technologies Inc., Japan) using mouse CRT cDNA (pCRII-mCRT) [34] as a template. To express the protein as a fusion protein with glutathione S-transferase (GST), the construct bearing EcoRI and XhoI restriction sequences at the 50 -terminus and 30 -terminus, respectively, was subcloned into the bacterial expression vector pGEX-6p-1 (GE Healthcare). The vector for the expression of GST-CRT fusion protein (i.e., pGEX-GST-mCRT) [31] can be obtained from the authors. 4. The bacterial expression plasmid coding mouse CRT [pmcsg7mCRT(WT), cat. no. #83505] is also available from Addgene (https://www.addgene.org/). 5. In the case of CNX, only the ER luminal segment of CNX is fused to GST, and the bacterial expression plasmids were constructed to adopt them for similar in vitro studies [14, 32]. 6. Reduced glutathione solution must be prepared just before use and should not be stored for more than 24 h in dilute solutions. 7. CRT was eluted at 0.5 M NaCl and judged to be >95% pure by SDS-PAGE. Approximately 0.5–1.0 mg of purified CRT was obtained from 1 L of bacterial culture. Purified CRT in solution can be stored at 20  C for at least 1 year and can be stored at 4  C for at least 2 weeks. 8. As alternative sources, CRT protein is commercially available from various companies: OriGene (cat. no. TP303222); Creative Biomart, New York, NY, USA (cat. no. CALR-2368H); Enzo Life Sciences, Ann Arbor, MI, USA (cat. no. ADI-SPP600-D); Abbexa, Cambridge, UK (cat. no. abx651473); and MyBioSource, San Diego, CA, USA (cat. no. MBS203510); abcam (cat. no. ab15729). GST-CRT fusion protein is also available from Abnova, Taipei, Taiwan (cat. no. H00000811P01). 9. It was reported that Jack bean α-mannosidase possesses Glc1Man9GlcNAc2 glycans [35, 36]. Therefore, α-mannosidase has been used as a substrate to monitor aggregation suppression by CNX or CRT [15, 32, 37]. The α-mannosidase is very useful because it is commercially available.

306

Yoshito Ihara et al.

10. The protein concentration of Jack bean α-mannosidase (Sigma-Aldrich) differs in each lot of the reagent. 11. To calculate the protein concentration (molar concentration) of samples, the molecular weights of CRT, α-mannosidase, and IgG are set as 46,000 (monomer), 110,000 (monomer), and 150,000 (monomer), respectively. 12. Chicken immunoglobulin Y (IgY), which possesses monoglucosylated N-glycans [38], is also available as a substrate to study the lectin-chaperone functions of CNX or CRT [15, 39]. 13. The α-mannosidase assay buffer must be prepared just before use and should not be stored. 14. The 40 stock solution of α-mannosidase substrate contains 200 mM p-nitrophenyl α-D-mannopyranoside in DMSO and can be stored at 20  C for at least 1 week. 15. An enzyme assay kit for α-mannosidase is also commercially available from BioAssay systems, Hayward, CA, USA (QuantiChrom™ α-Mannosidase Assay kit, cat. no. DAMA-100). References 1. Kornfeld R, Kornfeld S (1985) Assembly of asparagine-linked oligosaccharides. Annu Rev Biochem 54:631–664. https://doi.org/10. 1146/annurev.bi.54.070185.003215 2. Lamriben L, Graham JB, Adams BM, Hebert DN (2016) N-glycan-based ER molecular chaperone and protein quality control system: the calnexin binding cycle. Traffic 17:308–326. https://doi.org/10.1111/tra.12358 3. Roth J, Zuber C (2017) Quality control of glycoprotein folding and ERAD: the role of N-glycan handling, EDEM1 and OS-9. Histochem Cell Biol 147:269–284. https://doi. org/10.1007/s00418-016-1513-9 4. Tannous A, Pisoni GB, Hebert DN, Molinari M (2015) N-linked sugar-regulated protein folding and quality control in the ER. Semin Cell Dev Biol 41:79–89. https://doi.org/10. 1016/j.semcdb.2014.12.001 5. Caramelo JJ, Parodi AJ (2015) A sweet code for glycoprotein folding. FEBS Lett 589:3379–3387. https://doi.org/10.1016/j. febslet.2015.07.021 6. Michalak M, Groenendyk J, Szabo E, Gold LI, Opas M (2009) Calreticulin, a multi-process calcium-buffering chaperone of the endoplasmic reticulum. Biochem J 417:651–666. https://doi.org/10.1042/BJ20081847 7. Hammond C, Braakman I, Helenius A (1994) Role of N-linked oligosaccharide recognition, glucose trimming, and calnexin in glycoprotein

folding and quality control. Proc Natl Acad Sci U S A 91:913–917 8. Spiro RG, Zhu Q, Bhoyroo V, Soling HD (1996) Definition of the lectin-like properties of the molecular chaperone, calreticulin, and demonstration of its copurification with endomannosidase from rat liver Golgi. J Biol Chem 271:11588–11594 9. Ware FE, Vassilakos A, Peterson PA, Jackson MR, Lehrman MA, Williams DB (1995) The molecular chaperone calnexin binds Glc1Man9GlcNAc2 oligosaccharide as an initial step in recognizing unfolded glycoproteins. J Biol Chem 270:4697–4704 10. Chouquet A, Paidassi H, Ling WL, Frachet P, Houen G, Arlaud GJ, Gaboriaud C (2011) X-ray structure of the human calreticulin globular domain reveals a peptide-binding area and suggests a multi-molecular mechanism. PLoS One 6:e17886. https://doi.org/10.1371/ journal.pone.0017886 11. Kozlov G, Pocanschi CL, Rosenauer A, BastosAristizabal S, Gorelik A, Williams DB, Gehring K (2010) Structural basis of carbohydrate recognition by calreticulin. J Biol Chem 285:38612–38620. https://doi.org/10. 1074/jbc.M110.168294 12. Schrag JD, Bergeron JJ, Li Y, Borisova S, Hahn M, Thomas DY, Cygler M (2001) The structure of calnexin, an ER chaperone involved in quality control of protein folding. Mol Cell 8:633–644

Chaperone Functions of Calnexin/Calreticulin 13. Cummings RD, Etzler ME, Surolia A (2017) L-type lectins. In: Varki A, Cummings RD, Esko JD et al. (eds) Essentials of glycobiology [Internet], 3rd edn. Cold Spring Harbor (NY). https://doi.org/10.1101/glycobiology.3e. 032 14. Ihara Y, Cohen-Doyle MF, Saito Y, Williams DB (1999) Calnexin discriminates between protein conformational states and functions as a molecular chaperone in vitro. Mol Cell 4:331–341 15. Saito Y, Ihara Y, Leach MR, Cohen-Doyle MF, Williams DB (1999) Calreticulin functions in vitro as a molecular chaperone for both glycosylated and non-glycosylated proteins. EMBO J 18:6718–6729. https://doi.org/10. 1093/emboj/18.23.6718 16. Wiuff C, Houen G (1996) Cation-dependent interactions of calreticulin with denatured and native proteins. Acta Chem Scand 50:788–795 17. Danilczyk UG, Williams DB (2001) The lectin chaperone calnexin utilizes polypeptide-based interactions to associate with many of its substrates in vivo. J Biol Chem 276:25532–25540. https://doi.org/10. 1074/jbc.M100270200 18. Williams DB (2006) Beyond lectins: the calnexin/calreticulin chaperone system of the endoplasmic reticulum. J Cell Sci 119:615–623. https://doi.org/10.1242/jcs. 02856 19. Boelt SG, Norn C, Rasmussen MI, Andre I, Ciplys E, Slibinskas R, Houen G, Hojrup P (2016) Mapping the Ca(2+) induced structural change in calreticulin. J Proteome 142:138–148. https://doi.org/10.1016/j. jprot.2016.05.015 20. Lum R, Ahmad S, Hong SJ, Chapman DC, Kozlov G, Williams DB (2016) Contributions of the lectin and polypeptide binding sites of calreticulin to its chaperone functions in vitro and in cells. J Biol Chem 291:19631–19641. https://doi.org/10.1074/jbc.M116.746321 21. Moreau C, Cioci G, Iannello M, Laffly E, Chouquet A, Ferreira A, Thielens NM, Gaboriaud C (2016) Structures of parasite calreticulins provide insights into their flexibility and dual carbohydrate/peptide-binding properties. IUCrJ 3:408–419. https://doi.org/10. 1107/S2052252516012847 22. Tan Y, Chen M, Li Z, Mabuchi K, Bouvier M (2006) The calcium- and zinc-responsive regions of calreticulin reside strictly in the N-/C-domain. Biochim Biophys Acta 1760:745–753. https://doi.org/10.1016/j. bbagen.2006.02.003

307

23. Wijeyesakere SJ, Gafni AA, Raghavan M (2011) Calreticulin is a thermostable protein with distinct structural responses to different divalent cation environments. J Biol Chem 286:8771–8785. https://doi.org/10.1074/ jbc.M110.169193 24. Ellgaard L, Riek R, Herrmann T, Guntert P, Braun D, Helenius A, Wuthrich K (2001) NMR structure of the calreticulin P-domain. Proc Natl Acad Sci U S A 98:3133–3138. https://doi.org/10.1073/pnas.051630098 25. Kozlov G, Munoz-Escobar J, Castro K, Gehring K (2017) Mapping the ER Interactome: the P domains of calnexin and calreticulin as plurivalent adapters for foldases and chaperones. Structure 25:1415–1422.e3. https:// doi.org/10.1016/j.str.2017.07.010 26. Nakao H, Seko A, Ito Y, Sakono M (2017) PDI family protein ERp29 recognizes P-domain of molecular chaperone calnexin. Biochem Biophys Res Commun 487:763–767. https://doi.org/10.1016/j. bbrc.2017.04.139 27. Oliver JD, van der Wal FJ, Bulleid NJ, High S (1997) Interaction of the thiol-dependent reductase ERp57 with nascent glycoproteins. Science 275:86–88 28. Eggleton P, Bremer E, Dudek E, Michalak M (2016) Calreticulin, a therapeutic target? Expert Opin Ther Targets 20:1137–1147. https://doi.org/10.1517/14728222.2016. 1164695 29. Feng M, Marjon KD, Zhu F, WeissmanTsukamoto R, Levett A, Sullivan K, Kao KS, Markovic M, Bump PA, Jackson HM, Choi TS, Chen J, Banuelos AM, Liu J, Gip P, Cheng L, Wang D, Weissman IL (2018) Programmed cell removal by calreticulin in tissue homeostasis and cancer. Nat Commun 9:3194. https:// doi.org/10.1038/s41467-018-05211-7 30. Merlinsky TR, Levine RL, Pronier E (2019) Unfolding the role of calreticulin in myeloproliferative neoplasm pathogenesis. Clin Cancer Res 25:2956. https://doi.org/10.1158/ 1078-0432.CCR-18-3777 31. Totani K, Ihara Y, Matsuo I, Koshino H, Ito Y (2005) Synthetic substrates for an endoplasmic reticulum protein-folding sensor, UDP-glucose: glycoprotein glucosyltransferase. Angew Chem Int Ed 44:7950–7954. https://doi.org/10.1002/anie.200502723 32. Stronge VS, Saito Y, Ihara Y, Williams DB (2001) Relationship between calnexin and BiP in suppressing aggregation and promoting refolding of protein and glycoprotein

308

Yoshito Ihara et al.

substrates. J Biol Chem 276:39779–39787. https://doi.org/10.1074/jbc.M107091200 33. Li YT (1967) Studies on the glycosidases in jack bean meal. I. Isolation and properties of alpha-mannosidase. J Biol Chem 242:5474–5480 34. Kageyama K, Ihara Y, Goto S, Urata Y, Toda G, Yano K, Kondo T (2002) Overexpression of calreticulin modulates protein kinase B/Akt signaling to promote apoptosis during cardiac differentiation of cardiomyoblast H9c2 cells. J Biol Chem 277:19255–19264. https://doi. org/10.1074/jbc.M112377200 35. Gnanesh Kumar BS, Pohlentz G, Schulte M, Mormann M, Siva Kumar N (2014) Jack bean alpha-mannosidase: amino acid sequencing and N-glycosylation analysis of a valuable glycomics tool. Glycobiology 24:252–261. https://doi. org/10.1093/glycob/cwt106 36. Kimura Y, Hess D, Sturm A (1999) The N-glycans of jack bean alpha-mannosidase.

Structure, topology and function. Eur J Biochem 264:168–175 37. Thomson SP, Williams DB (2005) Delineation of the lectin site of the molecular chaperone calreticulin. Cell Stress Chaperones 10:242–251 38. Ohta M, Hamako J, Yamamoto S, Hatta H, Kim M, Yamamoto T, Oka S, Mizuochi T, Matsuura F (1991) Structures of asparaginelinked oligosaccharides from hen egg-yolk antibody (IgY). Occurrence of unusual glucosylated oligo-mannose type oligosaccharides in a mature glycoprotein. Glycoconj J 8:400–413 39. Patil AR, Thomas CJ, Surolia A (2000) Kinetics and the mechanism of interaction of the endoplasmic reticulum chaperone, calreticulin, with monoglucosylated (Glc1Man9GlcNAc2) substrate. J Biol Chem 275:24348–24356. https://doi.org/10.1074/jbc.M003102200

Chapter 30 Purification and Assays of Tachylectin-2 Shun-ichiro Kawabata and Toshio Shibata Abstract Tachylectin-2, a 27-kDa protein consisting of a five-bladed β-propeller structure, is purified by three steps of chromatography, including dextran sulfate-Sepharose CL-6B, CM-Sepharose CL-6B, and Mono S. Three isolectins of tachylectin-2 including tachylectin-2a, -2b, and -2c are purified. These isolectins exhibit hemagglutinating activity against human A-type erythrocytes in a Ca2+-independent manner with tachylectin-2b showing the highest activity. Tachylectin-2b specifically agglutinates Staphylococcus saprophyticus KD. The tachylectin-2b-mediated hemagglutination is inhibited in the presence of GlcNAc and GalNAc. The association constants for GlcNAc and GalNAc are Ka ¼ 1.95  104 M1 and Ka ¼ 1.11  103 M1, respectively. Ultracentrifugation analysis shows that tachylectin-2b is present in monomer form in solution. Key words β-propeller structure, Hemagglutination, GlcNAc

1

Introduction In the innate immune immunity in the Japanese horseshoe crab Tachypleus tridentatus, unique microbial polysaccharides including lipopolysaccharides (LPS) of Gram-negative bacteria are recognized through immunocompetent molecules such as many types of lectins and antimicrobial peptides derived from hemocytes and hemolymph plasma [1–3]. Horseshoe crab hemolymph contains one type of granular hemocytes and the hemocyte stores four types of lectins including tachylectins-1 [4], tachylectins-2 [5], tachylectins-3 [6], and tachylectins-4 [7] in its large granules. These tachylectins are secreted from hemocytes in response to LPS. Horseshoe crab hemolymph plasma also contains several types of lectins, such as tachylectins-5A and tachylectins-5B [8], isoforms of tachylectins-1 and tachylectins-3 [9–11], and three types of C-reactive proteins (CRPs) [12]. These lections exhibit unique binding specificity for monosaccharides and polysaccharides on microbial cell walls. Tachylectin-1 bonds to 2-keto-3-deoxyoctonate on Gram-negative bacteria [4], whereas tachylectin-2 binds

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_30, © Springer Science+Business Media, LLC, part of Springer Nature 2020

309

310

Shun-ichiro Kawabata and Toshio Shibata

to GlcNAc or GalNAc and lipoteichoic acids of Gram-positive bacteria [5]. Tachylectin-3 recognizes a certain sugar chain on O-antigens of LPS, and tachylectin-4 recognizes colitose (3-deoxy-L-fucose) and an O-antigen part of LPS of Escherichia coli O111:B4 [6]. In contrast, tachylectins-5A and tachylectins-5B, acetyl-group-recognizing lectins, exhibit extraordinarily strong hemagglutinating activity against all types of human erythrocytes and bacterial agglutinating activity [8]. Horseshoe crab CRPs exhibit LPS-binding activity and their expression is upregulated by the bacterial infection. The crystal structural analyses of tachylectin-2 [13] and tachylectin-5A [14] indicate the importance of its multivalency to achieve high specificity and high affinity against pathogenic ligands. For example, tachylectin2 (236 amino acids) is composed of five tandem WD-like repeats and builds up a five-bladed β-propeller structure with an equivalent GlcNAc/GalNAc-binding site at each peak of the β-propeller blades (Fig. 1) [13]. On the other hands, plasma-derived lectins including galactose-binding protein, the isoform of tachylectin-1, tachylectins-5A and tachylectins-5B, carcinolectin-5, and CRPs are involved in the complement-dependent clearance system of invading microbes [3, 15–17]. Here we describe a protocol of the purification and characterization of tachylectin-2 [5].

Fig. 1 The crystal structure of tachylectin-2 complexed with GlcNAc (PDB code 1TL2 [13])

Characteristics of Tachylectin-2

2

311

Materials

2.1 Preparation of Hemocyte Extract

Glassware and metalware used for the preparation of hemocyte lysate and the dextran sulfate chromatography are sterilized by heating at 220  C for 3 h. All the buffer solutions used for these steps are made up with pyrogen-free distilled water and autoclaved for 30 min (see Note 1). 1. 70% (v/v) ethanol. 2. Pyrogen-free distilled water (Otsuka Pharmaceutical, Tokyo, Japan). 3. 20 mM Tris–HCl, pH 8.0, containing 50 mM NaCl. 4. 20 mM Tris–HCl, pH 8.0, containing 0.5 M NaCl. 5. 20 mM Tris–HCl, pH 8.0, containing 0.5 M NaCl and 0.1 M caffeine. 6. Disposable sterilized needle. 7. Disposable sterilized plastic container.

2.2 Purification of Tachylectin-2

1. Sepharose England).

2.2.1 Preparation of Dextran Sulfate-Sepharose CL-6B

2. Dextran sulfate (GE Healthcare).

CL-6B

(GE

Healthcare,

Buckinghamshire,

3. CNBr. 4. Acetonitrile. 5. 10 M NaOH. 6. 20 mM Tris–HCl, pH 8.0.

2.2.2 Dextran Sulfate-Sepharose CL-6B Column Chromatography

1. 20 mM Tris–HCl, pH 8.0, containing 50 mM NaCl. 2. 20 mM Tris–HCl, pH 8.0, containing 0.3 M NaCl. 3. 20 mM Tris–HCl, pH 8.0, containing 0.5 M NaCl. 4. 20 mM Tris–HCl, pH 8.0, containing 2.0 M NaCl.

2.2.3 CM-Sepharose CL-6B Column Chromatography

1. CM-Sepharose CL-6B (GE healthcare).

2.2.4 Mono S Column Chromatography

1. Mono S HR 5/5 (GE Healthcare).

2. 20 mM Tris–HCl, pH 7.5. 3. 20 mM Tris–HCl, pH 7.5, containing 0.15 M NaCl.

2. 50 mM sodium acetate, pH 5.5. 3. 50 mM sodium acetate, pH 5.5, containing 0.2 M NaCl.

2.3 Hemagglutination Assay

1. Outdated human A-, B-, and O-type concentrated erythrocytes (see Note 2). 2. 20 mM Tris–HCl, pH 7.5, containing 0.15 M NaCl. 3. Round-bottomed microtiter plates.

312

Shun-ichiro Kawabata and Toshio Shibata

2.4 Bacterial Agglutination Assay

1. Bacteria: Staphylococcus aureus 209P, Staphylococcus epidermidis K3, Staphylococcus saprophyticus KD, Micrococcus luteus, Enterococcus hirae, and Escherichia coli strain B. 2. 3% Tryptosoy broth sterilized by autoclaving for 20 min. 3. 20 mM Tris–HCl, pH 7.5, containing 0.15 M NaCl. 4. Round-bottomed microtiter plates.

2.5 Determination of Association Constants

1. 20 mM Tris–HCl, pH 7.5, containing 0.15 M NaCl

2.6 Analytical Ultracentrifugation

1. 20 mM Tris–HCl, pH 7.5, containing 0.15 M NaCl.

3

2. 100 mM GlcNAc and GalNAc as stock solutions.

Methods

3.1 Preparation of Hemocyte Extract

1. Bleed hemolymph by inserting a sterilized needle into the joint between the cephalothorax and the abdomen rinsed with 70% ethanol and then with sterilized water. 2. Collect the hemolymph (~150 mL from each horseshoe crab) into a sterilized container containing one-tenth volume (~15 mL) of 20 mM Tris–HCl, pH 8.0, containing 0.5 M NaCl and 0.1 M caffeine chilled on ice. 3. Pool the hemolymph into sterilized metal centrifugation tubes (~300 mL each). 4. Centrifuge the hemolymph at 3000  g for 15 min at 4  C and separated into hemocytes and hemolymph plasma. 5. Store hemocytes and hemolymph plasma at 80  C until use, and both are stable at least for 2 years. 6. Thaw frozen hemocytes (50–100 g) and suspend the hemocytes in 200 mL of 20 mM Tris–HCl, pH 8.0, containing 50 mM NaCl in the sterilized metal centrifugation tube. 7. Homogenize the hemocyte suspension with a Polytron homogenizer (Model PT3100, Kinematica AG, Lucerne, Switzerland) for 3 min (see Note 3), and the lysate is chilled on ice before centrifugation. 8. Centrifuge the lysate at 12,000  g for 30 min at 4  C and transfer the extract into a sterilized container. 9. Re-extract the pellet three times and pool them as hemocyte extract.

Characteristics of Tachylectin-2

3.2 Purification of Tachylectin-2 3.2.1 Preparation of Dextran Sulfate-Sepharose CL-6B

313

1. Add 500 mL of 50 mg/mL dextran sulfate chilled on ice to Sepharose CL-6B (250 mL) washed with distilled water and mix well by stirring. 2. Add 120 mL of 0.5 g/mL CNBr in acetonitrile and incubate the mixture on ice for 45 min and maintain the pH at ~10.5 and the temperature at ~10  C by dropping 10 M NaOH and ice made from sterilized water. 3. Wash the dextran sulfate-coupled Sepharose CL6B with sterilized water and suspend the resin in 20 mM Tris–HCl, pH 8.0.

3.2.2 Dextran Sulfate-Sepharose CL-6B Column Chromatography

1. Equilibrate a dextran sulfate-Sepharose CL6B column (5  20 cm) with 20 mM Tris–HCl, pH 8.0, containing 50 mM NaCl. 2. Apply hemocyte extract prepared from ~50 g of hemocytes and start to collect the eluate by a fraction collector. 3. Wash the column with three column volumes of the equilibration buffer, and bound proteins were eluted with the same buffer containing 0.3, 0.5, and 2.0 M NaCl in a stepwise fashion (200 mL each). 4. Pool fractions containing the 27-kDa protein in the flowthrough fraction, judged by SDS-PAGE in 12% gel.

3.2.3 CM-Sepharose CL-6B Column Chromatography

1. Apply the pooled fraction to a CM-Sepharose CL-6B column (2  12 cm) equilibrated with 20 mM Tris–HCl, pH 7.5, and wash the column with five column volumes of the equilibration buffer. 2. Elute proteins with a linear gradient (400 mL) of 0–0.15 M NaCl in the same buffer and pool the eluted fractions containing the 27-kDa protein, judged by SDS-PAGE. 3. Dialyze the pooled sample against 50 mM sodium acetate, pH 5.5.

3.2.4 Mono S Column Chromatography

1. Apply the dialyzed sample to a Mono S HR 5/5 column equilibrated with 50 mM sodium acetate, pH 5.5, and washed the column with the equilibration buffer (~10 mL). 2. Elute tachylectin-2 isoforms in order, tachylectin-2a, tachylectin-2b, and tachylectin-2c, with a linear gradient (~20 mL) of 0–0.2 M NaCl in the same buffer.

3.3 Hemagglutination Assay

1. Centrifuge human concentrated erythrocytes at 1000  g for 3 min and wash three times with 20 mM Tris–HCl, pH 7.5, containing 0.15 M NaCl. 2. Prepare 2% suspension (v/v) of erythrocytes in the same buffer (see Note 4).

314

Shun-ichiro Kawabata and Toshio Shibata

3. Mix 25 μL of the suspension of erythrocytes with 25 μL of a twofold serial dilution of tachylectin-2 in a round-bottomed microtiter plate and incubate for 1 h at 37  C. 4. The titer is defined as a reciprocal value of the endpoint dilution causing hemagglutination. 5. For screening of inhibitors for hemagglutination, premix 12.5 μL of a test sample in the same buffer and 12.5 μL of tachylectin-2 (10 μg/mL) before adding 25 μL of the suspension of erythrocytes. Inhibition activity is expressed as the minimum inhibitory concentration of the test sample. 3.4 Bacterial Agglutination Assay

1. Culture bacteria in 5 mL of 3% Tryptosoy broth for 12 h at 37  C. 2. Collect bacteria by centrifugation at 4000  g for 2 min and wash with 20 mM Tris–HCl, pH 7.5, containing 0.15 M NaCl. 3. Suspend the washed bacteria in ~1.0 mL of the same buffer to obtain an absorbance at 600 nm of ~10. 4. Mix 25 μL of the suspension of bacteria with 25 μL of a twofold serial dilution of tachylectin-2 in a round-bottomed microtiter plate and incubate for 12 h at 25  C. 5. The bacterial agglutinating activity is expressed as the minimum agglutinating concentration of tachylectin-2.

3.5 Determination of Association Constants

1. Measure emission spectra of ~40 μg/mL tachylectin in 20 mM Tris–HCl, pH 7.5, containing 0.15 M NaCl in the absence or presence of a ligand, induced by excitation at 280 nm with a Hitachi F-4000 fluorescence spectrophotometer at 25  C. 2. Calculate the association constants (Ka) according to the equation [18], log (F  F0)/(F1  F) ¼ log [S] + log Ka: F0 ¼ the fluorescence intensity in the absence of a ligand, F ¼ the fluorescence intensity in the presence of a ligand, F1 ¼ the fluorescence intensity saturated with a ligand, and [S] ¼ the concentration of a ligand.

3.6 Analytical Ultracentrifugation

1. Dilute tachylectin-2 in 20 mM Tris–HCl, pH 7.5, containing 0.15 M NaCl by the same buffer to give absorbance at 280 nm of 1.16, 0.81, and 0.46. 2. Perform sedimentation equilibrium run in a Beckman Optima XL-A analytical ultracentrifuge at 12,000  g at 4  C for 20 h. 3. Determine the concentration gradients in the cells spectrophotometrically at 280 nm. The partial specific volume of tachylectin-2 is calculated to be 0.73 mL/g, based on the amino acid composition.

Characteristics of Tachylectin-2

4

315

Notes 1. If these procedures for sterilization are omitted, the coagulation factors in hemocytes will be converted to the active forms, leading to clot formation. 2. Outdated human concentrated erythrocytes can be stored for at least 1 month at 4  C. 3. The blade of Polytron is sterilized overnight by soaking in 95% ethanol containing 0.2 M NaOH. 4. Volume of erythrocytes is measured in a centrifuge tube with scale.

References 1. Kawabata S, Tsuda R (2002) Molecular basis of non-self recognition by the horseshoe crab tachylectins. Biochim Biophys Acta 1572:414–421 2. Iwanaga S (2002) The molecular basis of innate immunity in the horseshoe crab. Curr Opin Immunol 14:87–95 3. Kawabata S (2011) Immunocompetent molecules and their response network in horseshoe crabs. In: So¨derh€all K (ed) Invertebrate immunity. Springer Science + Business Media, New York, pp 122–136 4. Saito T, Kawabata S, Hirata M, Iwanaga S (1995) A novel type of Limulus lectin-L6. J Biol Chem 270:14493–14499 5. Okino N, Kawabata S, Saito T et al (1995) Purification, characterization, and cDNA cloning of a 27-kDa lectin (L10) from horseshoe crab hemocytes. J Biol Chem 270:31008–31015 6. Inamori K, Saito T, Iwaki D et al (1999) A newly identified horseshoe crab lectin with specificity for blood group A antigen recognizes specific O-antigens of bacterial lipopolysaccharides. J Biol Chem 274:3272–3278 7. Saito T, Hatada M, Iwanaga S, Kawabata S (1997) A newly identified horseshoe crab lectin with binding specificity to O-antigen of bacterial lipopolysaccharides. J Biol Chem 272:30703–30708 8. Gokudan S, Muta T, Tsuda R et al (1999) Horseshoe crab acetyl group-recognizing lectins involved in innate immunity are structurally related to fibrinogen. Proc Natl Acad Sci U S A 96:10086–10091 9. Nagai T, Kawabata S, Shishikura F, Sugita H (1999) Purification, characterization, and amino acid sequence of an embryonic lectin in

perivitelline fluid of the horseshoe crab. J Biol Chem 274:37673–37678 10. Chiou S-T, Chen Y-W, Chen S-C, Chao C-F, Liu T-Y (2000) Isolation and characterization of protein that bind to galactose, lipopolysaccharide of Escherichia coli, and protein A of Staphylococcus aureus from the Hemolymph of Tachypleus tridentatus. J Biol Chem 275:1630–1634 11. Chen S-C, Yen C-H, Yeh M-S, Huang C-J, Liu T-Y (2001) Biochemical properties and cDNA cloning of two new lectins from the plasma of Tachypleus tridentatus. J Biol Chem 276:9631–9639 12. Iwaki D, Osaki T, Yoshimitsu M et al (1999) Functional and structural diversities of C-reactive proteins present in horseshoe crab hemolymph plasma. Eur J Biochem 264:314–326 13. Beisel H-G, Kawabata S, Iwanaga S, Huber R, Bode W (1999) Tachylectin-2: crystal structure of a specific GlcNAc/GalNAc-binding lectin involved in the innate immunity host defense of the Japanese horseshoe crab Tachypleus tridentatus. EMBO J 18:2313–2322 14. Kairies N, Beisel H-G, Fuentes-Prior P et al (2001) The 2.0-A˚ crystal structure of tachylectin 5 provides evidence for the common origin of the innate immunity and the blood coagulation systems. Proc Natl Acad Sci U S A 98:13519–13524 15. Ng PM, Le Saux A, Lee CM et al (2007) C-reactive protein collaborates with plasma lectins to boost immune response against bacteria. EMBO J 26:3431–3440 16. Saux AL, Ng PML, Kho JJY et al (2008) The macromolecular assembly of pathogenrecognition receptors is impelled by serine

316

Shun-ichiro Kawabata and Toshio Shibata

protease, via their complement control protein modules. J Mol Biol 377:902–913 17. Tagawa K, Yoshihara T, Shibata T et al (2012) Microbe-specific C3b deposition in the horseshoe crab complement system in a C2/factor B-dependent or -independent manner. PLoS One 7:e3678

18. Matusmoto I, Yamaguchi H, Seno N, Shibata Y, Okuyama T (1982) In: Hirano S, Tokura S (eds) Proceeding of the second international conference and chitin and chitosan, Sapporo, 12–14 July 1982, Japan. The Japanese Society of Chitin and Chitosan, Tottori, pp 165–170

Chapter 31 Purification and Assays of Tachycitin Shun-ichiro Kawabata and Toshio Shibata Abstract An antimicrobial peptide tachycitin (73 amino acids) is purified by steps of chromatography, including Sephadex G-50 and S Sepharose FF, from the acid extract of hemocyte debris of horseshoe crabs. Tachycitin is present in monomer form in solution, revealed by ultracentrifugation analysis. Tachycitin exhibits bacterial agglutination activity and inhibits the growth of both Gram-negative bacteria, Gram-positive bacteria, and fungus Candida albicans. Interestingly, tachycitin shows synergistic antimicrobial activity in corporation with another antimicrobial peptide, big defensin. Tachycitin shows a specific binding activity to chitin but not to cellulose, mannan, xylan, and laminarin. Tachycitin is composed of the N-terminal threestranded β-sheet and the C-terminal two-stranded β-sheet following a short helical turn, and the C-terminal structural motif shares a significant structural similarity with the chitin-binding domain derived from a plant chitin-binding protein, hevein. Key words Antimicrobial peptide, Bacterial agglutination, Chitin-binding protein, Tachycitin

1

Introduction Tachycitin is a small granular component of hemocytes derived from the Japanese horseshoe crab Tachypleus tridentatus [1]. Other antimicrobial peptides identified from T. tridentatus, including tachyplesin [2], big defensin [3], and tachystatins [4] are also localized in the small granules in hemocytes and secreted into extracellular fluid in response to bacterial lipopolysaccharides [5– 7]. All of these antimicrobial peptides have not only antimicrobial activity but also binding activity to chitin [1, 6]. Chitin, a polymer of GlcNAc through the β-1,4-linkage, is a component of the cell wall of fungi and is considered to be one of target substances recognized by the innate immune system of horseshoe crabs. The solution structures of these antimicrobial peptides have determined by the NMR structural analyses [8–12]. Tachycitin (73 amino acid residues) is divided into the N-terminal and C-terminal domains, and the C-terminal forming a hairpin loop of a two-stranded β-sheet superimposes the structure of the chitin-binding region of hevein, an antifungal peptide derived from the rubber tree Hevea

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_31, © Springer Science+Business Media, LLC, part of Springer Nature 2020

317

318

Shun-ichiro Kawabata and Toshio Shibata

Fig. 1 Comparison of the NMR structures of tachycitin and hevein. (a) The NMR structure of tachycitin (PDB code 1DQC [9]). (b) The NMR structure of hevein (PDB code 1Q9B [13, 14]). The homologous β-hairpin loops between tachycitin and hevein are indicated by dotted ellipses

Fig. 2 The NMR structure of tachyplesin (PDB code 2RTV [8])

brasiliensis [9, 13, 14] (Fig. 1). Tachyplesin (17 amino acid residues) is also composed of a homologous β-hairpin loop with that of tachycitin [8] (Fig. 2), and hydrophobic residues clustered on the one face of their β-hairpin loops may function as chitin-binding motif [6, 8]. A major component of exoskeletons of arthropods including horseshoe carbs is chitin, and these antimicrobial peptides released from hemocytes may bind to chitin exposed at the sites of injury to enhance wound healing. Interestingly, one of the antimicrobial peptides, tachyplesin, acts as a secondary secretagogue to enhance exocytosis of granular components of hemocytes in the absence of lipopolysaccharides, indicating an endogenous amplification system for hemocyte exocytosis [15]. Here, we describe a protocol of the purification and characterization of tachycitin [1].

Characteristics of Tachycitin

2

319

Material

2.1 Preparation of Hemocyte Debris

Glassware and metalware used for the preparation of hemocyte lysate and the dextran sulfate chromatography are sterilized by heating at 220  C for 3 h. All the buffer solutions used for these steps are made up with pyrogen-free distilled water and autoclaved for 30 min (see Note 1). 1. 70% (v/v) ethanol 2. Pyrogen-free distilled water (Otsuka Pharmaceutical, Tokyo, Japan). 3. 20 mM Tris–HCl, pH 8.0, containing 50 mM NaCl 4. 20 mM Tris–HCl, pH 8.0, containing 0.5 M NaCl 5. 20 mM Tris–HCl, pH 8.0, containing 0.5 M NaCl and 0.1 M caffeine 6. Disposable sterilized needle. 7. Disposable sterilized plastic container.

2.2 Purification of Tachycitin 2.2.1 Preparation of Protein Extract from Hemocyte Debris

1. 10% and 30% acetic acid.

2.2.2 Sephadex G-50 Column Chromatography

1. Sephadex G-50 (GE Healthcare, Buckinghamshire, England).

2.2.3 S-Sepharose FF Column Chromatography

1. S-Sepharose FF (GE Healthcare).

2. 10% acetic acid.

2. 20 mM Tris–HCl, pH 8.0, containing 0.1 M NaCl. 3. 20 mM Tris–HCl, pH 8.0, containing 0.4 M NaCl. 4. Sephadex G-25 (GE Healthcare). 5. 10% acetic acid.

2.3 Antimicrobial Activity

1. Bacteria: Salmonella typhimuriumLT2 (S), Salmonella minnesota R595 (Re), Escherichia coli B, Klebsiella pneumoniae, Staphylococcus aureus, and Candida albicans (see Note 2). 2. 3% Tryptosoy broth sterilized by autoclaving for 20 min 3. 10 mM sodium phosphate, pH 7.0 4. 1% agar plates containing 3% Tryptosoy broth

320

Shun-ichiro Kawabata and Toshio Shibata

5. Guanofracin-Sabouraud agar plates for C. albicans (see Note 3). 6. Big defensin purified from hemocyte debris [3]. 2.4 Bacterial Agglutination Assay

1. Bacteria: Staphylococcus aureus 209P, Staphylococcus epidermidis K3, Staphylococcus saprophyticus KD, Micrococcus luteus, Enterococcus hirae, and Escherichia coli strain B. 2. 3% Tryptosoy broth sterilized by autoclaving for 20 min 3. 20 mM Tris–HCl, pH 7.5, containing 0.15 M NaCl 4. Round-bottomed microtiter plates.

2.5 Analytical Ultracentrifugation

3

1. 20 mM Tris–HCl, pH 7.5, containing 0.05 M NaCl.

Methods

3.1 Preparation of Hemocyte Debris

1. Bleed hemolymph by inserting a sterilized needle into the joint between the cephalothorax and the abdomen rinsed with 70% ethanol and then with sterilized water. 2. Collect the hemolymph (~150 mL from each horseshoe crab) into a sterilized container containing one-tenth volume (~15 mL) of 20 mM Tris–HCl, pH 8.0, containing 0.5 M NaCl and 0.1 M caffeine chilled on ice. 3. Pool the hemolymph into sterilized metal centrifugation tubes (~300 mL each). 4. Centrifuge the hemolymph at 3000  g for 15 min at 4  C and separated into hemocytes and hemolymph plasma. 5. Store hemocytes and hemolymph plasma at 80  C until use, and both are stable at least for 2 years. 6. Thaw frozen hemocytes (50–100 g) and suspend the hemocytes in 200 mL of 20 mM Tris–HCl, pH 8.0, containing 50 mM NaCl in the sterilized metal centrifugation tube. 7. Homogenize the hemocyte suspension with a Polytron homogenizer (Model PT3100, Kinematica AG, Lucerne, Switzerland) for 3 min (see Note 4), and the lysate is chilled on ice before centrifugation. 8. Centrifuge the lysate at 12,000  g for 30 min at 4  C and transfer the extract into a sterilized container. 9. Re-extract the pellet three times and pool them as hemocyte extract. The remaining pellet is stored as hemocyte debris at 80  C.

Characteristics of Tachycitin

3.2 Purification of Tachycitin 3.2.1 Preparation of Protein Extract from Hemocyte Debris

321

1. Extract proteins twice from the hemocyte debris (30 g) dissolved in 200 mL of 30% acetic acid by homogenizing with a Polytron homogenizer (Model PT3100, Kinematica AG, Lucerne, Switzerland) for 3 min at room temperature. 2. Centrifuge the extract at 14,000  g for 15 min at room temperature and lyophilize the supernatant. 3. Dissolve the dried material in 50 mL of 10% acetic acid.

3.2.2 Sephadex G-50 Column Chromatography

1. Apply the dissolved material to a Sephadex G-50 column (3.6  110) equilibrated with 10% acetic acid. 2. Pool and lyophilize fractions containing the 8-kDa protein, judged by SDS-PAGE in 15% gel.

3.2.3 S-Sepharose FF Column Chromatography

1. Dissolve the dried material in 20 mM Tris–HCl, pH 8.0, containing 0.1 M NaCl. 2. Apply the to an S-Sepharose FF column (2  32 cm) equilibrated with the same buffer and wash the column with two column volumes of the equilibration buffer. 3. Elute proteins with a linear gradient (600 mL) of 0.1–0.4 M NaCl in the same buffer and two peaks are separated, both of which contains the 8-kDa protein (see Note 5). 4. Pool the first peak containing tachycitin. 5. Apply the pooled fraction to Sephadex G-25 equilibrated with 10% acetic acid to remove salt. 6. Lyophilize the desalted fraction.

3.3 Antimicrobial Activity

1. Culture bacteria in 3 mL of 3% Tryptosoy broth for 12 h at 37  C. 2. Collect bacteria by centrifugation at 4000  g for 2 min and wash them with 10 mM sodium phosphate buffer, pH 7.0. 3. Suspend the washed bacteria in 3 mL of 10 mM sodium phosphate buffer, pH 7.0 by mixing with vortex. 4. Dilute the bacterial suspension ~105-fold with the same buffer by using a sterile micropipette tip to transfer 10 μL to a fresh tube containing 1 mL of the same buffer repeatedly (see Note 6). 5. Add 50 μL of tachycitin to 450 μL of the diluted bacterial suspension. As a negative control experiment, add 50 μL of the phosphate buffer to 450 μL of the diluted bacterial suspension. In the case of determination for a synergistic effect between tachycitin and big defensin on antimicrobial activity, add 50 μL of the mixture of big defensin and tachycitin to 450 μL of the diluted bacterial suspension. 6. Incubate the mixture for 1 h at 37  C.

322

Shun-ichiro Kawabata and Toshio Shibata

7. Plate 100 μL of the reaction mixture onto 1% agar plates containing 3% Tryptosoy broth and incubate at 37  C for 24 h. For C. albicans, use guanofracin-Sabouraud agar plates. 8. Count the number of bacterial colonies on the agar plates and make the counted number 10 times to obtain colony forming units (CFU, number of colonies/mL of test sample). 9. Antimicrobial activity is expressed as 50% inhibitory concentration of tachycitin (IC50) (see Note 7). 3.4 Bacterial Agglutination Assay

1. Culture bacteria in 5 mL of 3% Tryptosoy broth for 12 h at 37  C. 2. Collect bacteria by centrifugation at 4000  g for 2 min and wash with 20 mM Tris–HCl, pH 7.5, containing 0.15 M NaCl. 3. Suspend the washed bacteria in ~1.0 mL of the same buffer to obtain an absorbance at 600 nm of ~10. 4. Mix 25 μL of the suspension of bacteria with 25 μL of a twofold serial dilution of tachycitin in a round-bottomed microtiter plate and incubate for 12 h at 25  C. 5. The bacterial agglutinating activity is expressed as the minimum agglutinating concentration of tachycitin.

3.5 Analytical Ultracentrifugation

1. Dilute tachycitin in 20 mM Tris–HCl, pH 7.5, containing 0.05 M NaCl by the same buffer. 2. Perform sedimentation equilibrium run in a Beckman Optima XL-A analytical ultracentrifuge at 37,000  g at 20  C for 20 h. 3. Determine the concentration gradients in the cells spectrophotometrically at 280 nm. The partial specific volume of tachycitin is calculated to be 0.717 mL/g, based on the amino acid composition.

4

Notes 1. If these procedures for sterilization are omitted, the coagulation factors in hemocytes will be converted to the active forms, leading to clot formation. 2. Bacteria can be stored for many years in media containing 15% glycerol at 80  C. 3. The guanofracin-Sabouraud agar plate for C. albicans is commercially available. 4. The blade of Polytron is sterilized overnight by soaking in 95% ethanol containing 0.2 M NaOH. 5. The first peak contains tachycitin and the second peak contains another antimicrobial peptide, big defensin.

Characteristics of Tachycitin

323

6. Repeat the dilution step serially to obtain the diluted bacterial suspension of 5000–10,000 cells/mL. OD ¼ 1.0 (optical density at 600 nm) of the bacterial suspension is roughly equivalent to 8  108 cells/mL. 7. Use at least five plates for one test sample and calculate the average value of CFU. References 1. Kawabata S, Nagayama R, Hirata M et al (1996) Tachycitin, a small granular component in horseshoe crab hemocytes, is an antimicrobial protein with chitin-binding activity. J Biochem 120:1253–1260 2. Nakamura T, Furunaka R, Miyata T et al (1988) Tachyplesin, a class of antimicrobial peptide from the hemocyte of the horseshoe crab Tachypleus tridentatus: isolation and chemical structure. J Biol Chem 263:16709–16713 3. Saito T, Kawabata S, Shigenaga T et al (1995) A novel big defensin identified in horseshoe crab hemocytes: isolation, amino acid sequence and antibacterial activity. J Biochem 117:1131–1137 4. Osaki T, Omotezako M, Nagayama R et al (1999) Horseshoe crab hemocyte-derived antimicrobial polypeptides, tachystatins, with sequence similarity to spider neurotoxins. J Biol Chem 274:26172–26178 5. Iwanaga S (2002) The molecular basis of innate immunity in the horseshoe crab. Curr Opin Immunol 14:87–95 6. Kawabata S, Osaki T, Iwanaga S (2002) Innate immunity in the horseshoe crab. In: Ezekowitz RAB, Hoffman JA (eds) Infectious disease: innate immunity. Humana Press Inc, Totowa, NJ, pp 109–125 7. Kawabata S (2011) Immunocompetent molecules and their response network in horseshoe crabs. In: So¨derh€all K (ed) Invertebrate immunity. Springer Science + Business Media, New York, pp 122–136 8. Kawano K, Yoneya T, Miyata T et al (1990) Antimicrobial peptide Tachyplesin I, isolated

from hemocytes of the horseshoe carb Tachypleus tridentatus. J Biol Chem 265:15365–15367 9. Suetake T, Tsuda S, Kawabata S et al (2000) Chitin-binding proteins in invertebrates and plants comprise a common chitin-binding structural motif. J Biol Chem 275:17929–17932 10. Fujitani N, Kawabata S, Osaki T et al (2002) Structure of the antimicrobial peptide tachystatin A. J Biol Chem 277:23651–23657 11. Fujitani N, Kouno T, Nakahara T et al (2007) The solution structure of horseshoe crab antimicrobial peptide tachystatin B with an inhibitory cysteine-knot motif. J Pept Sci 13:269–279 12. Kouno T, Fujitani N, Mizuguchi M et al (2008) A novel beta-defensin structure: a potential strategy of big defensin to overcome resistance by Gram-positive bacteria. Biochemistry 47:10611–10619 13. Asensio JL, Canada FJ, Briux M et al (1995) The interaction of hevein with Nacetylglucosamine-containing oligosaccharides: solution structure of hevein complex to chitobiose. Eur J Biochem 230:621–633 14. Asensio JL, Canada FJ, Briux M et al (1998) NMR investigation of protein-carbohydrate interactionsrefined three-dimensional structure of the complex between hevein and methyl beta-chitobioside. Glycobiology 8:569–577 15. Ozaki A, Ariki S, Kawabata S (2005) An antimicrobial peptide tachyplesin acts as a secondary secretagogue and amplifies lipopolysaccharide-induced hemocyte exocytosis. FEBS J 272:3863–3871

Chapter 32 Methods for Purifying Datura stramonium Agglutinin and Producing Recombinant Agglutinin Protein in a Heterologous Plant Host Suguru Oguri Abstract Datura stramonium seeds contain at least three chitin-binding isolectins as homo- or heterodimers of A and B subunits. This lectin has been used for the detection and isolation of sugar chains with N-acetyllactosaminyl structures on highly branched N-glycans. In terms of future diagnostic use, the development of a recombinant lectin will be the most effective approach for producing homogeneous lectin preparations. This chapter presents details of the procedure used for lectin purification and also describes a method that can be used for producing active recombinant homodimeric BB-isolectin in Arabidopsis plants. Key words Datura stramonium, Seed, Chitin binding, Recombinant lectin, Solanaceae, DSA

1

Introduction Datura stramonium agglutinin (DSA) is a chitin-binding lectin that has been extracted and purified from D. stramonium seeds [1, 2] (Fig. 1). DSA was originally proposed to be a dimeric glycoprotein composed of two nonidentical subunits that are linked by disulfide bonds. Using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), the sizes of the DSA subunits were estimated to be 40 and 45 kDa. Subsequently, using hydrophobic interaction chromatography, it was revealed that DSA occurs as a mixture of three different isoforms, referred to as isolectins, which correspond to the AA, AB, and BB combinations of the two subunits (A and B). The A- and B-subunits have similar, although nonidentical, molecular weights [3]. A cDNA-encoding DSA subunit B (DSA-B, DDBJ/EMBL/GenBank accession No. AB618634) was isolated by our group and we subsequently analyzed its molecular structure [4]. On the basis of MALDI-TOF MS analysis, we estimated that the molecular mass of the homodimeric BB-isolectin is 68,821 Da, whereas the molecular mass of

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_32, © Springer Science+Business Media, LLC, part of Springer Nature 2020

325

Suguru Oguri 1000

Activity (Titer/mg protein)

326

800 600 400 200 0

I

II

III

IV

Stages of fruits development

Fig. 1 Changes in hemagglutinating activity during fruit ripening in Datura stramonium. Fruit development of D. stramonium was divided into four stages: stage I, 1–7 days after flowering (DAF); stage II, 8–14 DAF; stage III, 15–21 DAF, and stage IV, 22–28 DAF. Fruits were harvested during each stage and the seeds were collected. The hemagglutinating activities of the seed extracts were measured. Whole fruits were used for stage I. Inset: (left) a flower, (right) a mature fruit and seeds

the B-subunit is estimated to be 37,748 Da, and that of the deglycosylated B-subunit, as an S-pyridylethylated derivative, is 26,491 Da [4]. Furthermore, we estimate that the carbohydrate content of the BB-isolectin is 32.7%. Owing to its high carbohydrate content and unusual amino acid composition, DSA-B is resolved as an SDS-PAGE band with a molecular weight of approximately 40 kDa. The cDNA of DSA-B encodes 279 amino acids, and Fig. 2 shows the proposed molecular structure of the encoded protein. DSA-B is a chimeric protein comprising a combination of two protein motifs, namely, the chitin-binding domains (CBDs) and extensin-like domains (ELDs). The CBD is a common structural motif comprising 30–43 amino acids that is similar to hevein, a small chitin-binding protein found in rubber tree latex. This motif (pfam, PF00187) includes eight Cys residues at conserved positions that are responsible for the formation of four intra-domain disulfide bridges. Many plant chitin-binding lectins belonging to the hevein family are known to consist of multiple CBDs. For example, the agglutinin obtained from Urtica dioica and wheat germ agglutinin comprise two and four tandem arrayed CBDs, respectively. DSA-B consists of four CBDs, each of 40 residues, that are separated by the insertion of an ELD of 37 residues containing four SerPro4-6 motifs (Fig. 2). Although each CBD contains eight conserved Cys residues, the second of these domains also

Production of Recombinant Datura stramonium Agglutinin in a. . .

H2N- Sp

Pp

CBD1

CBD2

CBD4

CBD3

327

-COOH

SH

Sp

:Signal pepƟde

Pp

:PropepƟde

O-(oligo)-galactoside : Extensin-like domain

:Start site of mature protein

O-polyarabinoside

: ChiƟn-binding domain

:Proposed C-terminal processing

Fig. 2 Domain structure of DSA-B. The N-terminus of the mature peptide starts at residue 50. CBD1 to CBD4 each consist of 40 residues, which share 62–82% identities. CBD2 contains an additional Cys residue (-SH), which may participate in dimerization through intra-disulfide bridge formation. Note that the number of galactose residues and sites of polyarabinosylation in the extensin-like domain (ELD) are not exact

contains an additional Cys residue, which may participate in dimerization through inter-disulfide bridge formation (Fig. 2). The ELD is a hydroxyproline-rich glycoprotein domain that resembles extensin, a cell wall glycoprotein. The composition of this domain is similar to that of the chitin-binding lectins identified in plants of the Solanaceae family, such as potato and tomato [5, 6]. The hemagglutinating activity of DSA has been found to be inhibited by β1-4-linked N-acetylglucosamine (GlcNAc) oligomers and N-acetyllactosamine (Galβ1-4GlcNAc, LacNAc) [7] and shows the highest affinity for highly branched N-glycans consisting of LacNAc [8]. On the basis of these specificities, DSA has been used to analyze structural changes in sugar chains during oncogenesis [9]. More recently, DSA has been used in a lectin microarray for glycome profiling and has also been applied in glycoproteomics research in combination with mass spectrometry [10–12]. Furthermore, Sasaki et al. showed that DSA inhibits the proliferation of rat C6 glioma cells by binding specifically to cell surface glycans on these cells [13]. Accordingly, DSA can be considered a valuable tool in glycobiological studies. In such studies, the heterogeneity of lectin preparations may be an important factor affecting results. For use in biomedical research, lectins extracted from their natural sources have several drawbacks, including low yields, contamination with other nontarget lectins, and batch-to-batch variation in the lectin source that results in heterogeneity of lectin-binding properties. However, these potential problems can be overcome to varying extents by recombinant expression and production of lectins in heterologous

328

Suguru Oguri

expression systems [14]. To date, commercially available DSA preparations have been purified from D. stramonium seeds and have been distributed as mixtures of several isoforms. We, however, have generated DSA-B-overexpressing transgenic Arabidopsis plants, from which a recombinant DSA-B (rDSA-B) was produced as a homodimeric glycoprotein with a molecular mass similar to that of the native form. Moreover, we found that the carbohydratebinding specificity of the recombinant DSA-B was the same as that of the native BB-isolectin [4]. In terms of future diagnostic use, the production of recombinant DSA will undoubtedly have distinct advantages. Nevertheless, although we were successful in producing a homogeneous DSA in a heterologous plant host, the yield of rDSA-B obtained was not sufficient for potential commercial use. Compared to the content of native BB-isolectin in Datura seed, the amount of rDSA-B that accumulated in transgenic Arabidopsis plants was less than one-tenth (44 μg/g fresh weight). Therefore, further work is needed to enhance the production yield of rDSA-B in heterologous plant hosts. In this chapter, we describe a procedure used for isolation of the native BB-isolectin from Datura seeds and also describe a method that can be used for producing recombinant lectins in Arabidopsis plants.

2 2.1

Materials Plant Materials

1. Datura stramonium L. 12441-87HK, obtained from the National Institute of Biomedical Innovation, Japan (NIBIO). 2. Arabidopsis thaliana L. ecotype Col-0.

2.2

Bacterial Strains

1. Agrobacterium tumefaciens strain C58C1Rif harboring p35dsa-b [4], stored as a glycerol stock at 70  C.

2.3

cDNA Material

1. p35dsa-b. The nucleotide sequence of DSA-B cDNA (AB618634), encoding residues 1–279, is amplified by PCR. To facilitate cloning into the binary vector pRI101-AN (Takara Bio, Shiga, Japan) containing the CaMV 35S promoter, an NdeI restriction site is incorporated into the sense primer, 50 -CACTGTTGATACATATGATGAGAATGAGACATACC30 , and an EcoRI site is incorporated into the antisense primer, 50 -TGTTGATTCAGAATTCTAGATAGCATTAAGCAAG-30 . The amplified fragment is subcloned and the resulting NdeI– EcoRI fragment (842 bp) is inserted into the corresponding site of the pRI101-AN vector (see Notes 1 and 2).

Production of Recombinant Datura stramonium Agglutinin in a. . .

2.4 Hemagglutination Assay

329

1. Phosphate-buffered saline (PBS): 137 mM NaCl, 8.1 mM Na2HPO4, 2.68 mM KCl, and 1.47 mM KH2PO4, pH 7.4. Autoclaved and stored at 4  C. 2. A 2% suspension of rabbit erythrocytes in PBS. Stored at 4  C (see Note 3). 3. 96-Well U-bottom plate.

2.5 Preparation of Chitin-Gel Via Chitosan Acetylation

1. Chitosan 10 (Deacetylation rate min. 80.0 mol %, FUJIFILM Wako Pure Chemicals, Osaka, Japan Cat. No. 034-22922) (see Note 4). 2. 10% acetic acid: 10% (v/v) acetic acid in water. 3. Acetic anhydride.

2.6 Buffers and Materials for Purifications of Native and Recombinant DSA

1. Dialysis tubing (MWCO 10,000). 2. Phenyl-Sepharose HP (GE Healthcare UK Ltd., Amersham Place, England). 3. Toyopearl HW-55F (Tosoh Corp., Tokyo, Japan). 4. 1 M NaCl in PBS. 5. 4.2% DAP: 4.2% (v/v) 1,3-diaminopropane (DAP) in water. Add 840 μL of DAP to 20 mL of pure water (see Note 5). 6. 80% ammonium sulfate-saturated PBS. Add 561 g of ammonium sulfate to 1 L of PBS. Store at 4  C. 7. 200 mM (4-amidinophenyl)-methane-sulfonyl (APMSF) in water. Store at 20  C.

fluoride

8. 50 mM ammonium bicarbonate. 2.7 Buffers and Media for Transformation of Arabidopsis thaliana

All media and buffers are made up as aqueous preparations. The media used for bacterial and tissue culture are autoclaved prior to use. 1. LB medium: 1% (w/v) Bacto-Tryptone, 0.5% (w/v) BactoYeast extract, 1% (w/v) NaCl, pH 7.0. 2. LB agar medium: LB medium supplemented with 1.5% (w/v) agar. 3. MS medium: 4.3 g of MS basal salt mixture, 2% (w/v) sucrose, 100 mg/L myo-inositol, 10 mg/L thiamine-HCl, 1 mg/L nicotinic acid, and 1 mg/L pyridoxine hydrochloride, pH 5.6. Adjust pH with KOH. 4. MS agar medium: MS medium supplemented with 0.8% (w/v) agar. For selection of transformants, add 50 mg/L kanamycin and 200 mg/L vancomycin. 5. Floral dip inoculation medium: 5% sucrose, 0.05% (v/v) Silwet L-77 in water. Dissolve 0.5 g of sucrose in 10 mL of distilled water. Add 5 μL of Silwet L-77. Prepare immediately prior to use without autoclaving.

330

Suguru Oguri

2.8 Detection of Transgene and Recombinant Protein

3

1. Primer 35S: 50 -GATGTGATATCTCCACTGAC-30 . 2. Primer M13RV: 50 -GGAAACAGCTATGACCATGATTAC-30 . 3. Rabbit anti-tomato lectin antiserum (see Note 6). 4. Goat anti-rabbit IgG horseradish peroxidase conjugate.

Methods

3.1 Hemagglutination Assay

1. Prepare a 2% rabbit erythrocyte suspension in PBS.

3.2 Preparation of Chitin-Gel Via Chitosan Acetylation

1. Weigh 2.5 g of chitosan into a mortar and grind to a powder.

2. Dilute the sample (20 μL) with PBS in a 96-well U-bottom plate and mix with an equal volume of 2% rabbit erythrocyte suspension. Hemagglutination units (titer) are calculated to be the reciprocal of the dilution multiple that yields a positive reaction after 1 h at room temperature.

2. Add 50 mL of 10% (v/v) acetic acid solution to the chitosan powder and grind well with a pestle. 3. Cover the mortar with parafilm and store overnight at room temperature to facilitate gelatinization of the chitosan. 4. Slowly add the chitosan gel to 225 mL of methanol on a magnetic stirrer. 5. Add 3.75 mL of acetic anhydride and stir the solution for 1 min with a glass rod. The solution will solidify. Allow the gel to stand for 30 min at room temperature. 6. Crush the gel into small pieces in methanol using a grinder. 7. Thoroughly wash the gel with distilled water through a glass filter. 8. Finally, add sodium azide to the gel slurry to a final concentration of 0.05% and store at 4  C.

3.3 Purification of DSA Isolectins from Datura Seeds Using Affinity Chromatography

The procedure described here essentially follows that originally reported by Kilpatrick and Yeoman [1], with the exception of the use of a chitin-gel column in affinity chromatography [4]. 1. Fill seedling pots with soil for plant culture. Sow seeds of D. stramonium L. in the pots. Grow the resulting seedlings in a greenhouse for 1 month (see Note 7). Harvest ripe fruits at 25–28 days after pollination, before the fruits become dry and split. Cut fruits and collect mature seeds (Fig. 1) (see Note 8). Approximately 2 g of seeds are obtained from a mature green fruit. Store at 30  C until further use. 2. Homogenize 5 g of frozen seeds with 100 mL of PBS using a motor and pestle for 10 min.

Production of Recombinant Datura stramonium Agglutinin in a. . .

331

3. Transfer the homogenate to a glass beaker, place on a magnetic stirrer, and stir using a stir bar at 4  C overnight. 4. Pass the homogenate through cheesecloth to remove seeds and debris. Adjust the volume to 100 mL with PBS. 5. Remove the debris by centrifuging the homogenate at 12,000  g for 20 min at 4  C. 6. For heat treatment, transfer the supernatant to glass flasks and incubate the flasks at 70  C for 15 min in a water bath. 7. Remove the debris by centrifuging the seed extract at 12,000  g for 20 min at 4  C. 8. Add three volumes of ice-cold acetone to the seed extract and mix well. Incubate overnight at 30  C. 9. Obtain proteins by centrifuging the solution at 12,000  g for 20 min at 4  C. Decant supernatant and air-dry the pellet for 5–20 min. Resuspend pellets in 50 mL of PBS. 10. Load the crude protein solution into dialysis tubing (MWCO 10,000 Da) and dialyze against 3 L of PBS for 4 h at 4  C. Change the dialysis buffer twice. 11. Centrifuge the homogenate at 10,000  g for 20 min at 4  C. 12. Remove debris by centrifuging the solution at 12,000  g for 20 min. Transfer the clear supernatant to a conical tube. 13. Add 5 g of chitin-gel to the solution. 14. Gently mix the gel slurry overnight at 4  C by inverting on a rotator. 15. Transfer the supernatant to a fresh tube and pack the precipitate into a column (see Note 9). 16. Load the supernatant onto the column. 17. Wash the column with PBS until the change in A280 of the column effluent is less than 0.05 absorbance unit. 18. Wash the column with 30 mL of PBS containing 1 M NaCl. 19. Wash the column with 15 mL of distilled water. 20. Elute the adsorbed proteins from the column using 20 mL of 4.2% (v/v) DAP at a flow rate of 1 drop/10 s. Collect 1 mL of column effluent per tube. 21. Measure the A280 in each fraction. 22. Collect the protein peak and dialyze it against 3 L of PBS for 4 h at 4  C. Change the dialysis buffer twice. This fraction is composed of DSA isolectins and unidentified chitin-binding proteins (Fig. 3a, lane 5). Although gel filtration chromatography using Toyopearl HW-55F column is effective in

332

Suguru Oguri

b

a M kDa 97.2 66.4 45.0

1

2

3

4

5

1

6

2

3

4

5 M kDa

subunit

A B

29.0

20.1 14.3

97.2 66.4 45.0 29.0 20.1 14.3

Fig. 3 Purification of native and recombinant DSA-B. A crude extract prepared from Datura seeds (a) and dsa-b transgenic Arabidopsis T2 plants (b) were subjected to chromatography using a chitin-gel column and subsequent HW-55F gel filtration column. Aliquots from each step of the purification were analyzed by 12.5% SDS-PAGE under reducing conditions and stained with Coomassie brilliant blue. (a) Lane:1, crude extract; 2, acetone precipitation; 3, flow-through; 4, fraction eluted with PBS containing 1 M NaCl; 5, fraction eluted with 4.2% DAP; 6, HW-55F gel filtration. Locations of DSA-A and -B subunits are indicated. (b) Lane:1, crude extract; 2, flow-through; 3, fraction eluted with 4.2% DAP. The rDSA-B (lane 4) and 30-kDa protein (lane 5) were separated by HW-55F gel filtration. Lane M shows size standards

separating DSA and the chitin-binding protein of 25 kDa, separation of the individual DSA dimers (AA, AB, and BB) is not achieved (Fig. 3, lane 6). The procedure used for the purification of isolectins is described below. 3.4 Separation of DSA Isolectins Using Hydrophobic Interaction Chromatography

1. Measure the sample volume after dialysis (see step 22 of Subheading 3.3). 2. Add PBS containing ammonium sulfate to 80% saturation to the dialyzed sample to prepare a 15% saturated solution. 3. Centrifuge the solution at 12,000  g for 20 min at 4  C. 4. Load the clear supernatant onto a Phenyl-Sepharose HP column (1.6  14.5 cm) equilibrated with column loading buffer (PBS containing ammonium sulfate to 15% saturation). 5. Wash the column with 150 mL of the loading buffer. 6. Elute the adsorbed proteins from the column using a linear gradient of ammonium sulfate (15%–0% saturation, total 800 mL), at a flow rate of 1 mL/min. Collect 8 mL of column effluent in each tube. 7. Measure the hemagglutinating activity in each fraction. 8. Check the purity of active fractions by SDS-PAGE, using 12.5% polyacrylamide gel.

Production of Recombinant Datura stramonium Agglutinin in a. . .

333

9. Pool the fractions showing hemagglutinating activity into three fractions (I–III) (see Note 10). 10. Purify further each fraction by performing an identical chromatography on the same column. 11. Check the purity of each by SDS-PAGE, using 12.5% polyacrylamide gel (see Note 11). 3.5 Generation of DSA-B Transgenic Arabidopsis Plants

The basic procedure used for Arabidopsis transformation is described in refs. 15, 16. 1. Grow 15 Arabidopsis plants in pots containing moistened soil under continuous light at 22  C. Plant three seedlings per 7.5 cm pot. 2. Prepare Agrobacterium tumefaciens strain C58C1Rif carrying the binary plasmid p35dsa-b. Streak a glycerol stock of the strain on LB agar supplemented with 50 μg/mL kanamycin and 20 μg/mL rifampicin. Incubate the plate at 28  C for 2 days (see Note 12). 3. For pre-culture, inoculate a single colony into 1 mL of liquid LB medium containing 50 μg/mL kanamycin and 50 μg/mL rifampicin. Incubate at 28  C for 20 h with shaking (160 rpm). Inoculate 3 mL of liquid LB medium containing the same antibiotics with 30 μL of the pre-culture and incubate at 28  C for 12 h with shaking (160 rpm). Measure the OD600 of the bacterial culture. 4. Collect the Agrobacterium cells by centrifugation at 8000  g for 5 min at room temperature. 5. Decant the supernatant and resuspend the pellet to OD600 ¼ 0.8 in the floral dip inoculation medium. For transformation of 15 Arabidopsis plants, 4 mL of the Agrobacterium cell suspension (inoculum) is needed. 6. Place one drop of the inoculum onto each flower bud using a pipette. 7. Cover the inoculated plant with Saran wrap and place under continuous light at 22  C for 2 days. Stop watering. 8. Remove Saran wrap from plants and resume watering. 9. Grow the plants for a further month until the siliques turn brown. 10. Stop watering to allow the seeds to dry. 11. Harvest and air-dry seeds. 12. Select positive transformants on MS agar medium supplemented with 50 mg/L kanamycin and 200 mg/L vancomycin (see Notes 13 and 14).

334

Suguru Oguri

13. DNA insertion in T1 plants is confirmed by PCR. Use the primer pair 35S and M13RV for PCR, performed with an annealing temperature of 55  C. 14. Select transformants harboring dsa-b. 15. Grow seedlings of transformants independently in pots containing moistened soil under continuous light at 22  C. 16. Remove a leaf and place in a microtube on ice. 17. Homogenize the leaf with three volumes of PBS using a small pestle. 18. Centrifuge the homogenate at 10,000  g for 10 min at 4  C. 19. Collect supernatants and measure the hemagglutination activities in these extracts. 20. Separate proteins in aliquots of T1 leaf extracts by SDS-PAGE and detect rDSA-B by western blotting using anti-tomato lectin antiserum (Fig. 4) (see Note 6). 21. On western blotting analysis, rDSA-B shows a prominent band of 40 kDa, which is accompanied by a faint band of 30 kDa in all analyzed transformants (Fig. 4a) (see Note 15). 22. T2 generation plants are obtained from the self-pollination of T1 plants. 3.6 Purification of DSA Recombinant BB-Isolectin from Transgenic Arabidopsis Plants

1. Grow Arabidopsis T2 plants in soil for 1 month. 2. Homogenize 5 g of whole plants with 10 mL of PBS using a motor and pestle for 5 min. 3. Centrifuge the homogenate at 10,000  g for 20 min at 4  C and collect the extract as the supernatant. 4. Homogenize the residual pellet in 10 mL of PBS and centrifuge at 10,000  g for 20 min at 4  C. 5. Combine the extracts. 6. Add 200 mM APMSF to the supernatant to give a 0.2 mM solution. 7. Add 2 g of chitin-gel to the solution. Gently mix the gel slurry overnight at 4  C by inverting on a rotator. 8. After overnight incubation, the slurry is packed into a column (see Note 9). 9. Load the supernatant onto the column. 10. Wash the column with 10 mL of PBS. 11. Wash the column with 10 mL of PBS containing 1 M NaCl. 12. Wash the column with 10 mL of distilled water. 13. Elute the adsorbed proteins from the column using 8 mL of 4.2% (v/v) DAP at a flow rate of 1 drop/10 s. Collect 1 mL of column effluent per tube. Measure the A280 of each fraction.

Production of Recombinant Datura stramonium Agglutinin in a. . .

335

dsa-b transgenic line (T2) M S B W #1 #2 #3 #6 #12#15

a kDa 97.2 66.4 45.0 29.0 20.1 14.3

b kDa 97.2 66.4 45.0 29.0 20.1 14.3

Fig. 4 Detection of DSA-B produced in dsa-b transgenic Arabidopsis T2 plants. Crude extracts (100 μg of protein equivalent) prepared from dsa-b transgenic Arabidopsis T2 plants (lanes #1–#15) and wild type (lane W) were resolved by 12.5% SDS-PAGE under reducing condition, and proteins were transferred onto a nitrocellulose membrane. After proteins had been stained with Ponceau S (a), DSA-B was detected by immunoblotting using anti-tomato lectin antiserum (b). Lane: S, 5 μg of Datura seed protein; B, 0.2 μg of purified native BB-isolectin. Lane M: size standards

14. Collect the protein peak and dialyze it against 3 L of 50 mM ammonium bicarbonate for 4 h at 4  C. Change the dialysis buffer twice. 15. Lyophilize the dialyzed eluate and dissolve the lyophilized powder in 500 μL of PBS. 16. Apply the sample solution to a Toyopearl HW-55F gel filtration column (1.6  95 cm) equilibrated with PBS. Elute rDSAB at a flow rate of 1 mL/min. Collect 2 mL of column effluent in each tube. 17. Measure the hemagglutinating activity in each fraction.

336

Suguru Oguri

18. Check the purities of active fractions by SDS-PAGE, using 12.5% polyacrylamide gel (see Note 16). 19. Pool the fractions containing hemagglutinating activity (see Note 17 and 18).

4

Notes 1. We use the KOD-Plus polymerase (Toyobo, Osaka, Japan) as the PCR enzyme to amplify DSA-B gene. Add dimethyl sulfoxide to the reaction mixture to a final concentration of 5%. The PCR is performed with an annealing temperature of 59  C. 2. We use E. coli DH5α strain for subcloning of DSA-B gene. It may be difficult to get a positive clone transformed by plasmid vector containing DSA-B gene. After the transformation, grow transformants on the selective plate at 30  C and pick up small colonies growing slowly. 3. We purchased 50% rabbit erythrocytes in Alsever’s solution from a vendor. Human erythrocytes can also be used for the hemagglutinating assay. 4. Various grades of chitosan from different sources are provided by vendors. Some of these do not solidify during the acetylation step used in this procedure. 5. 1,3-diaminopropane (DAP) is a harmful alkaline reagent. Accordingly, when working with DAP, use gloves and eye protection. 6. This antiserum was prepared by immunizing rabbit with the purified tomato lectin [17]. The anti-tomato lectin antiserum cross reacts with DSA. 7. Datura stramonium is a self-pollinating plant. To promote pollination, shake the flower gently by hand. 8. Hemagglutinating activity is highest in the mature seeds (Fig. 1). 9. We use a 10-mL disposable syringe. 10. The fractions I and II yield the AB- and the BB-isolectin, respectively. The fraction III produces the isolectin C with an apparent molecular mass of 32 kDa (see ref. 4). The proportions of the hemagglutinating activities of fractions I, II and III were estimated at 53.3, 16.6, and 4.1% of total lectin activity in the seed extract, respectively [4]. 11. Purified BB-isolectin appears as a single 40-kDa band after performing SDS-PAGE under reducing conditions. In our experiment, 0.3 mg of BB-isolectin was purified from 5 g of Datura seeds. The yield obtainable using this purification process was approximately 14%. The BB-isolectin content in the seed amounted to 460 μg/g fresh weight.

Production of Recombinant Datura stramonium Agglutinin in a. . .

337

12. Agrobacterium tumefaciens strain LBA4404 may be used as the host bacterium instead of strain C58C1Rif. Electrocompetent LBA4404 cells were obtained from a vendor. 13. Vancomycin is used to eliminate Agrobacterium. It is not included in the medium for selection of the T2 generation. 14. The seedlings of transformants show green leaves and wellestablished roots, whereas the cotyledons of non-transformants turn white. 15. The 30-kDa protein has chitin-binding activity. The N-terminal sequence of the 30-kD protein begin at Leu160 of the DSA-B sequence, indicating that the latter part of rDSA-B is generated by endogenous proteolytic digestion (see ref. 4). 16. Proteins are eluted as wide single peak from the column. The recombinant BB-isolectin is eluted at the former part of the peak accompanying hemagglutination activity. The 30-kDa protein is yielded at latter part of the peak and it does not show hemagglutinating activity. 17. In our experiment, 0.0645 mg of rDSA-B was purified from 5 g of leaves from transgenic T2 plant. The yield obtainable using this purification process was approximately 29%. The recombinant lectin content in the transgenic Arabidopsis plant amounted to 44 μg/g fresh weight, corresponding to 0.47% of the total soluble proteins [4]. 18. The molecular mass of the native rDSA-B was determined to be 64,296 Da by MALDI-TOF MS, which is 4525 Da smaller than the native homodimeric BB-isolectin [4]. The difference in molecular mass between the native and recombinant DSA-B may results from host plant-mediated glycosylation. References 1. Kilpatrick DC, Yeoman MM (1978) Purification of the lectin from Datura stramonium. Biochem J 175(3):1151–1153 2. Crowley JF, Goldstein IJ (1981) Datura stramonium lectin: isolation and characterization of the homogeneous lectin. FEBS Lett 130 (1):149–152 3. Broekaert WF, Allen AK, Peumans WJ (1987) Separation and partial characterization of isolectins with different subunit compositions from Datura stramonium seeds. FEBS Lett 220(1):116–120 4. Nishimoto K, Tanaka K, Murakami T, Nakashita H, Sakamoto H, Oguri S (2015) Datura stramonium agglutinin: cloning, molecular characterization and recombinant production in Arabidopsis thaliana. Glycobiology 25(2):157–169

5. Van Damme EJ, Barre A, Rouge P, Peumans WJ (2004) Potato lectin: an updated model of a unique chimeric plant protein. Plant J 37 (1):34–45 6. Oguri S, Amano K, Nakashita H, Nagata Y, Momonoki YS (2008) Molecular structure and properties of lectin from tomato fruit. Biosci Biotechnol Biochem 72 (10):2640–2650 7. Crowley JF, Goldstein IJ, Arnarp J, Lonngren J (1984) Carbohydrate binding studies on the lectin from Datura stramonium seeds. Arch Biochem Biophys 231(2):524–533 8. Itakura Y, Nakamura-Tsuruta S, Kominami J, Tateno H, Hirabayashi J (2017) Sugar-binding profiles of chitin-binding lectins from the hevein family: a comprehensive study. Int J

338

Suguru Oguri

Mol Sci 18(6). https://doi.org/10.3390/ ijms18061160 9. Saitoh O, Wang WC, Lotan R, Fukuda M (1992) Differential glycosylation and cell surface expression of lysosomal membrane glycoproteins in sublines of a human colon cancer exhibiting distinct metastatic potentials. J Biol Chem 267(8):5700–5711 10. Sun Q, Kang X, Zhang Y, Zhou H, Dai Z, Lu W, Zhou X, Liu X, Yang P, Liu Y (2009) DSA affinity glycoproteome of human liver tissue. Arch Biochem Biophys 484(1):24–29 11. Kinoshita M, Mitsui Y, Kakoi N, Yamada K, Hayakawa T, Kakehi K (2014) Common glycoproteins expressing polylactosamine-type glycans on matched patient primary and metastatic melanoma cells show different glycan profiles. J Proteome Res 13(2):1021–1033 12. Kuno A, Uchiyama N, Koseki-Kuno S, Ebe Y, Takashima S, Yamada M, Hirabayashi J (2005) Evanescent-field fluorescence-assisted lectin microarray: a new strategy for glycan profiling. Nat Methods 2(11):851–856

13. Sasaki T, Yamazaki K, Yamori T, Endo T (2002) Inhibition of proliferation and induction of differentiation of glioma cells with Datura stramonium agglutinin. Br J Cancer 87(8):918–923 14. Oliveira C, Teixeira JA, Domingues L (2013) Recombinant lectins: an array of tailor-made glycan-interaction biosynthetic tools. Crit Rev Biotechnol 33(1):66–80 15. Clough SJ, Bent AF (1998) Floral dip: a simplified method for agrobacterium-mediated transformation of Arabidopsis thaliana. Plant J 16(6):735–743 16. Narusaka M, Shiraishi T, Iwabuchi M, Narusaka Y (2010) The floral inoculating protocol: a simplified Arabidopsis thaliana transformation method modified from floral dipping. Plant Biotechnol 27(4):349–351 17. Naito Y, Minamihara T, Ando A, Marutani T, Oguri S, Nagata Y (2001) Domain construction of cherry-tomato lectin: relation to newly found 42-kDa protein. Biosci Biotechnol Biochem 65(1):86–93

Chapter 33 ZG16p, an Animal Homologue of Plant β-Prism Fold Lectins: Purification Methods of Natural and Recombinant ZG16p and Inhibition Assay of Cancer Cell Growth Using ZG16p Akiko Mito, Kaori Kumazawa-Inoue, and Kyoko Kojima-Aikawa Abstract ZG16p is a soluble 16-kDa protein abundantly expressed in the pancreas and gut, and has a β-prism fold structure similar to that of mannose-binding Jacalin-related lectins (mJRLs) such as BanLec, Heltuba, and Artocarpin. ZG16p binds to mannose via the well-conserved GXXXD loop among mJRLs and sulfated glycosaminoglycans (e.g., heparin and heparan sulfate) via the basic patch of molecular surface. In addition to the above binding activities, ZG16p has inhibitory activity against proliferation of colon cancer cells. This manuscript describes purification of rat pancreatic ZG16p and recombinant ZG16p expressed in Escherichia coli expression system, and cell growth inhibition assay using ZG16p as an inhibitor. Key words Jacalin-related lectins, β-prism fold lectins, ZG16p, Colon cancer, Cell growth inhibition, Animal lectins

1

Introduction Zymogen granule protein 16 (ZG16p) is a soluble protein originally identified in pancreatic zymogen granules [1]. ZG16p binds to highly sulfated heparan sulfate glycans localized in the luminal surface of the pancreatic zymogen granule membrane [2] and, thus, is implicated in modulation of condensation-sorting (process of selective packing and sorting of pancreatic enzymes to the zymogen granule membrane) [3, 4]. ZG16p was also detected in goblet cells of the intestine, serosanguineous acinar cells of the parotid gland, and serum [5]. The crystal structure of ZG16p revealed a β-prismfold structure similar to that of mannose-binding Jacalin-related lectins (mJRLs) [6]. BanLec, one of mJRLs, binds to mannose via two loops (the GG loop and the GXXXD loop) [7], and ZG16p also binds to mannose through the GXXXD loop [5, 6]. Lys and Arg residues are also involved in the interaction of ZG16p with sulfated glycosaminoglycans [2, 6, 8] (Fig. 1).

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_33, © Springer Science+Business Media, LLC, part of Springer Nature 2020

339

340

A

Akiko Mito et al.

R79 R55

R53

D151 K36

R37

D151

R55 R53

R37 R79

K36

B R79 R55 R53

D151 K36 R37

R55

D151

R53 R37 R79

K36

Fig. 1 Structural characteristics of human ZG16p. (a) Crystal structure and amino acid sequence of human ZG16p. Positions of a mannose-binding residue (Asp151, blue) and basic amino acid residues responsible for binding to heparin (Lys36, Arg37, Arg53, Arg55, and Arg79, green) are indicated on the crystal structure (PDB ID: 3APA) and the amino acid sequence. The α-helices and β-strands are shown in red and yellow, respectively. (b) The electrostatic surface potential of ZG16p. The surface models of human ZG16p (PDB: 3APA) is colored according to the electrostatic surface potential. Blue, positive; red, negative; scale from 35 to +35 kcal/mol

Purification of ZG16p and Inhibition of Cancer Cell Growth by ZG16p

341

Several plant lectins have been reported to regulate the growth of colorectal cancer cell lines. Peanut agglutinin lectin that binds to the Thomsen-Friedenreich antigen (Galβ1-3GalNAcα) promotes the proliferation of HT29 cells [9]. Mulberry leaf lectin that binds to galactose and GalNAc induces apoptosis of HCT15 cells [10]. Some animal lectins that bind to galactoside also affect colorectal cancer cell growth. For example, Galectin-3 [11] promotes and Galectin-4 [12] inhibits growth of some colorectal cancer cell lines, and ZG16p inhibits growth of several colorectal cell lines and colorectal cancer organoid [13]. These observations indicate possible regulation of colorectal cancer progression by endogenous lectins via interaction with cell surface carbohydrate chains in vivo. Here, we describe protocols for purification of natural and recombinant ZG16p, and for assay of growth inhibition of colon cancer cells by ZG16p.

2

Materials

2.1 Purification of ZG16p from Rat Pancreas (Rat ZG16p)

1. Rat pancreas. 2. Heparin-Sepharose CL-6B (4 mL, GE Healthcare). 3. Extraction buffer: 5 mM MOPS (pH 7.0) containing 0.25 M sucrose, 0.1 mM magnesium sulfate, and 0.1 mM phenylmethylsulfonyl fluoride (PMSF). 4. Percoll solution: 50 mM MES (pH 5.5), containing 50% Percoll, 0.25 M sucrose, 2 mM EGTA, 0.1 mM MgSO4, and 0.1 mM PMSF. 5. Granule buffer: 20 mM MES (pH 5.5) containing 0.3 M sucrose, 1 mM EGTA, 0.1 mM MgSO4, and 0.1 mM PMSF. 6. Rupture buffer: 10 mM MOPS (pH 6.5) containing 150 mM sodium acetate and 0.1 mM PMSF. 7. Spectrophotometer. 8. Fraction collector. 9. Refrigerated centrifuge and refrigerated ultracentrifuge.

2.2 Expression and Purification of Recombinant ZG16p

1. Expression plasmid harboring cDNA encoding human ZG16p in E. coli: We typically use pASK-IBA5plus-ZG16p, a Streptagged expression construct. 2. DH5α Escherichia coli cells. 3. Anhydrotetracycline (200 μg/L at final conc.). 4. LB/Amp medium: LB broth containing 0.1 mg/mL ampicillin. 5. Binding buffer: 100 mM Tris–HCl pH 8.0, 150 mM NaCl. 6. 10 chaperone dissociation solution: 20 mM ATP, 100 mM MgSO4, and 1.5 M KCl.

342

Akiko Mito et al.

7. StrepTrap HP column (1 mL, GE Healthcare). 8. Dissociation buffer: 2.5 mM desthiobiotin in Binding buffer. 9. PBS: 10 mM Na2HPO4, 2 mM KH2PO4, 137 mM NaCl, and 2.7 mM KCl, pH 7.4. 10. Shaking culture chamber. 11. Spectrophotometer. 12. Refrigerated centrifuge. 2.3 Inhibition of Cancer Cell Growth by ZG16p 2.3.1 Cell Culture

1. Human colorectal cancer cell line, HCT116 (RIKEN). 2. Cell culture medium: Dulbecco’s modified Eagle’s medium (DMEM)/Ham’s F12 supplemented with 10% heatinactivated fetal bovine serum (FBS). 3. Serum-free medium: We typically use ASF104 (Ajinomoto). 4. Phosphate-buffered saline (PBS) pH 7.4 (Thermo Fisher). 5. TrypLE Express (Thermo Fisher). 6. Antibiotics: penicillin/streptomycin. 7. 100-mm culture dish. 8. CO2 incubator.

2.3.2 ZG16p Solution

1. Recombinant ZG16p (Strep-tagII-ZG16p): Purify StreptagII-ZG16p by affinity chromatography (see Note 1) [13]. Prepare ZG16p solutions (0, 35, and 70 μg/mL) in serum free medium supplemented with penicillin and streptomycin. Adjust the volume of PBS in ZG16p solutions. 2. PBS (sterilized): 10 mM Na2HPO4, 2 mM KH2PO4, 137 mM NaCl, 2.7 mM KCl, pH 7.4. Pass through a 0.22-μm filter unit. 3. 0.22 μm membrane filter unit (Advantec).

2.3.3 Cell Proliferation Assay

1. Cell counting Kit-8 (Dojindo). 2. 48-well culture plate 3. Microplate reader.

3

Methods

3.1 Purification of Rat ZG16p 3.1.1 Preparation of Zymogen Granule Extracts

The preparations of rat pancreatic zymogen granules, granule membranes, and granule contents are based on previously reported methods [2]. All procedures are performed at 4  C. 1. Remove pancreata from female rats (8 week old). 2. Homogenize pancreata on ice in extraction buffer. 3. Centrifuge at 200  g for 5 min.

Purification of ZG16p and Inhibition of Cancer Cell Growth by ZG16p

343

4. Transfer the supernatant (post-nuclear fraction) to another centrifugation tube. 5. Centrifuge again at 2,000  g for 15 min. 6. Discard the top mitochondrial layer and resuspend the crude zymogen granule pellet obtained in ice-cold Percol solution. 7. Centrifuge at 25,000  g for 90 min. 8. Discard upper layer and carefully transfer a dense white band (zymogen granule layer) near the bottom of the tube to another centrifugation tube. 9. Add Granule buffer and suspend gently. 10. Centrifuge at 25,000  g for 15 min. 11. Discard the supernatant. 12. Add rupture buffer to the pellet and incubate on ice for 15 min. 13. Centrifuged at 100,000  g for 15 min. 14. Collect the supernatant (zymogen granule extracts). 3.1.2 Purification of Rat ZG16p Using Heparin Column Chromatography

1. Apply zymogen granule extracts onto a Heparin column. 2. Wash the column with rupture buffer to remove unbound proteins. 3. Elute step-wisely with rupture buffer containing 0.15, 0.3, and 0.5 mM NaCl. 4. Collect fractions containing ZG16p (Fig. 2).

3.2 Expression and Purification of Recombinant ZG16p

1. Expression vector preparation (see Note 2): The cDNA encoding human ZG16p is amplified by PCR using a human pancreas cDNA library, and the fragments obtained are ligated into pASK-IBA5plus plasmid at suitable restriction sites. 2. Transform the plasmid DNA into E. coli DH5α competent cells. 3. Plate aliquots of the transformation reactions on LB agar containing 100 μg/mL ampicillin and incubate overnight at 37  C. 4. Inoculate 10 mL of LB medium containing 100 μg/mL ampicillin with a single colony and incubate the cultures overnight at 37  C in a shaking incubator. 5. Inoculate 350 mL of LB medium containing 100 μg/mL ampicillin with 3.5 mL of overnight culture. 6. Incubate the culture for 2 h, or until cells reach mid-log growth (A550 of 0.5–1.0) at 37  C in a shaking incubator. 7. Induce the culture by adding anhydrotetracycline to a final concentration of 200 μg/L and continue incubation for 4 h at 30  C.

Akiko Mito et al. Fraction No.

A 0 0.5

5

10

15

20

25

30

35

40

0.4

A280

344

0.3 0.2 0.1

B

0

0 M

kDa

0.15 M

0.3 M

0.5 M

66 45 31 21 14

0 M

0.15 M

0.3 M

0.5 M

Fig. 2 Affinity chromatography of rat zymogen granule contents on a heparinSepharose column. Rat zymogen granules were osmotically ruptured by 10 mM MOPS. After centrifugation, the resulting supernatant (ZGC) was applied to a heparin-Sepharose column (0.5  5 cm), and the adsorbed proteins were eluted stepwise with 0.15, 0.3, and 0.5 M NaCl to collect successive 0.5 mL fractions. (a) The elution profile. (b) SDS-PAGE of the proteins in the fractions. Arrows and an arrowhead indicate the positions of protein markers and ZG16p, respectively

8. Centrifuge the culture to collect the E. coli cells as a pellet. 9. Remove the supernatants by aspiration and resuspend the pellet in 15 mL of Binding buffer and transfer in a 50-mL tube. 10. Disrupt E. coli cells by sonication in appropriate conditions. 11. Centrifuge the cell lysate, transfer the supernatant in a 50-mL plastic tube, add 1/10 volume of 10 chaperon dissociation solution, and incubate at 37  C for 30 min. 12. Centrifuge and inject the supernatant into a StrepTrap HP column using a plastic syringe. 13. Wash the column with 5 mL binding buffer. 14. Elute Strep-tagII-ZG16p with dissociation buffer. For purity and quality evaluation of obtained recombinant proteins, SDS-polyacrylamide gel electrophoresis and circular dichroism (CD) spectrometry analysis are recommended (see Note 3).

Purification of ZG16p and Inhibition of Cancer Cell Growth by ZG16p

1. Culture HCT116 cells to 80–90% confluence in a 100-mm culture dish. 2. Prepare a cell suspension using a culture medium and determine the cell number. Adjust the cell number to 1.7  104 cells/mL (see Note 4). 3. Add 300 μL of the cell suspension (5.0  103 cells) using wideorifice pipette tips to each well of a 48-well plate (n ¼ 3 for each assay condition) (see Note 5). 4. Culture cells for 24 h. 5. Wash cells carefully with ASF104 supplemented with penicillin and streptomycin using a micropipette (see Note 6). 6. Add 300 μL of Strep-tagII-ZG16p/ASF104 at different concentrations (0, 35, and 70 μg/mL) (see Note 7). 7. Culture cells for another 72 h without changing culture medium. 8. Add 30 μL of Cell Counting Kit-8 (one-tenth of the medium) to each well, and mix well by gently tapping the plate. 9. Incubate the plate at 37  C for 2 h. 10. Measure absorbance at 450 nm using a microplate reader. 11. Analyze data using BellCurve for Excel (Social Survey Research Information Co., Ltd.). Perform multiple group comparison by Dunnett’s test (Fig. 3). Cell images are shown in Fig. 4.

1.4 1.2

Abs (450 nm)

3.3 Inhibition of Cancer Cell Growth by ZG16p

345

1.0

***

0.8 0.6

***

0.4 0.2 0.0

0

35

70

ZG16p (μg/mL) (***p250

D3ΔN

>250

a

Hemagglutination activity was represented as a minimum protein concentration (μg/ mL) required for complete hemagglutination of rabbit erythrocytes

50TBS9. Resulted protein is adsorbed onto both Ni and galactose-Sepharose columns, indicating that recombinant lectin is successfully obtained as an active form. In fact, the recombinant lectin also possesses hemagglutination activity toward rabbit erythrocytes (Table 1). Approximately 1 mg of recombinant protein is obtained from 100 mg of inclusion bodies. 6. As shown in Table 1, D3ΔC and D3ΔN show no hemagglutination activity, indicating that the two RBL motifs are essential for this lectin. The D3 molecule is capable of binding to galactose-sepharose as it is not found in the flow-through fraction of this column, but it partially behaves like a divalent molecule or with non-covalent association. However, it does not form a huge aggregate unlike the recombinants with more than three domains (Fig.2). D23, D123, and D2323 show comparable hemagglutination activity, suggesting that the number of domains does not affect the activity. Meanwhile, the activity of D123ΔH is slightly lower compared with the His-tagged recombinants, but all of the recombinants are less active than SAL (Table 1). This may be attributed either to the additional amino acid sequence from the plasmid or to partial misfolding. Optimizing for more suitable refolding conditions and understanding the underlying mechanisms are required to overcome these problems.

Bacterial Expression of Rhamnose-Binding Lectin

367

References 1. Ogawa T, Watanabe M, Naganuma T et al (2011) Diversified carbohydrate-binding kectins from marine resources. J Amino Acids. https://doi.org/10.4061/2011/838914 2. Sakakibara F, Takayanagi G, Kawauchi H (1981) An L-rhamnose-binding lectin in the eggs of Misgurnus anguillicaudatus. Yakugaku Zasshi 101(10):918–925 3. Sugawara S, Hosono M, Ogawa Y et al (2005) Catfish egg lectin causes rapid activation of multidrug resistance 1 P-glycoprotein as a lipid translocase. Biol Pharm Bull 28 (3):434–441 4. Watanabe Y, Tateno H, Nakamura-Tsuruta S et al (2009) The function of rhamnose-binding lectin in innate immunity by restricted binding to Gb3. Dev Comp Immunol 33(2):187–197 5. Ozeki Y, Matsui T, Suzuki M et al (1991) Amino acid sequence and molecular characterization of a D-galactoside-specific lectin purified from sea urchin (Anthocidaris crassispina) eggs. Biochemistry 30(9):2391–2394 6. Hosono M, Ishikawa K, Mineki R et al (1999) Tandem repeat structure of rhamnose-binding lectin from catfish (Silurus asotus) eggs. Biochim Biophys Acta 1472:668–675 7. Okamoto M, Tsutsui S, Tasumi H et al (2005) Tandem repeat L-rhamnose-binding lectin from the skin mucus of ponyfish, Leiognathus nuchalis. Biochem Biophys Res Commum 333 (2):463–469 8. Cammarata M, Parisi MG, Benenati G (2014) A rhamnose-binding lectin from sea bass

(Dicentrarchus labrax) plasma agglutinates and opsonizes pathogenic bacteria. Dev Comp Immunol 44(2):332–340 9. Hatakeyama T, Ichise A, Yonekura T et al (2015) cDNA cloning and characterization of a rhamnose-binding lectin SUL-I from the toxopneustid sea urchin Toxopneustes pileolus venom. Toxicon 94:8–15 10. Sugawara S, Im C, Kawano T et al (2017) Catfish rhamnose-binding lectin induces G0/1 cell cycle arrest in Burkitt’s lymphoma cells via membrane surface Gb3. Glycoconj J 34:127–138 11. Sugawara S, Sasaki S, Ogawa Y et al (2005) Catfish (Silurus asotus) lectin enhances the cytotoxic effects of doxorubicin. Yakugaku Zasshi 125(3):327–334 12. Hatakeyama T, Ichise A, Unno H et al (2017) Carbohydrate recognition by the rhamnosebinding lectin SUL-I with a novel threedomain structure isolated from the venom of globiferous pedicellariae of the flower sea urchin Toxopneustes pileolus. Protein Sci 26:1547–1583 13. Fornstedt N, Porath (1975) Characterization studies on a new lectin found in seeds of Vicia ervilia. FEBS Lett 57:187–191 14. Sano K, Ogawa H (2014) Hemagglutination (inhibition) assay. In: Hirabayashi J (ed) Lectins. Methods in molecular biology, vol 1200. Springer, Heidelberg, pp 47–52

Chapter 36 A Bioassay for Determining Symbiotic Zooxanthellae Shape Control Using Lectin SLL-2 from the Octocoral Sinularia lochmodes Mitsuru Jimbo, Ryota Takeuchi, and Mayu Yoshino Abstract Symbiosis with zooxanthellae is essential for survival of corals. Using a bioassay, we report the H-type lectin SLL-2 purified from the octocoral Sinularia lochmodes to restrict zooxanthellae form to spherical cells. However, the factor for initiating or maintaining a symbiotic relationship between a host and zooxanthellae has not been found in many corals. This bioassay is useful for evaluating the role of a lectin as a symbiosisrelated factor. Key words Zooxanthella, SLL-2, Transformation, Bioassay, Symbiosis

1

Introduction Symbiosis is an interaction between two different organisms living in close physical association, in which one, the host, often benefits. Some invertebrates have symbiotic zooxanthella microalgae in their bodies, which supply a photosynthate to the host as a nutrient [1]. The host prevents the alga from leaving, while the zooxanthella seems to inhibit phagocytosis and remain within the host’s cells in a compartment known as a symbiosome. How a host initiates or maintains a symbiotic relationship with zooxanthellae is not well understood. Corals are well-known symbiotic organisms and harbour zooxanthellae in their cells. More than half of the photosynthate produced by zooxanthellae is excreted by the algae and moves to the host cells. Free-living zooxanthellae do not, however, excrete photosynthate. Symbiosis also results in a change in the shape of zooxanthellae [2]. Free-living zooxanthellae are spherical and unable to move during the night; during the day, however, they are snowman-like in shape, have two flagellae and can swim. When within a host’s cells, zooxanthellae maintain a spherical shape.

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_36, © Springer Science+Business Media, LLC, part of Springer Nature 2020

369

370

Mitsuru Jimbo et al.

These phenomena suggest some factor(s) are controlling zooxanthellal shape. We purified a lectin, SLL-2, from the octocoral Sinularia lochmodes. This lectin’s amino acid sequence and structure are similar to H-type lectin from Helix pomatia [3, 4]. SLL-2 can cause zooxanthellae to transform to a coccoid form and to maintain this shape for 1 week. SLL-2 is distributed around the zooxanthellae in the host coral, suggesting it prevents zooxanthellae from leaving their host. Another lectin, CecL, purified from the solitary coral Ctenactis echinata, has a similar function with inhibitory activity of cell division [5]. Lectins of other corals have been found distributed around the zooxanthellae within host cells [6, 7], suggesting that a lectin somehow controls the shapes and motility of zooxanthellae. Zooxanthellae are diverse [8]. Each coral harbours a different genus, and the lectin around them within a host cell also differs. Quantifying the effect of a lectin has on zooxanthellae would aid identification of the factor(s) that control zooxanthellal shape and motility. A method for quantifying the effect of an H-type lectin such as SLL-2 is presented here.

2

Materials All solutions use ultrapure water and guaranteed reagents.

2.1 Selecting Zooxanthellae

1. A bead crusher or similar apparatus such as SK mill SK-100 (Tokken Inc., Chiba, Japan). 2. TE buffer: 1 mM EDTA and 10 mM Tris–HCl (pH 8.0). 3. PCR tubes. 4. 0.5 mL and 1.5 mL microtubes. 5. Thermostable DNA polymerase (e.g. Taq polymerase). 6. PCR primer for identifying clade: ‘itsD’ (50 -GTGAATTGCA GAACTCCGTG-30 ) and ‘ITS-2rev2’ (50 -CCTCCGCTTA CTTATATGCTT-30 ) primers [9]. 7. Sephacryl S-300 (GE Healthcare, Chicago, IL, USA). 8. BigDye Terminator v 3.1 Cycle Sequencing Kit (Thermo Fisher Scientific, Waltham, MA, USA).

2.2

Zooxanthellae

Zooxanthellae can be obtained from The National Centre for Marine Algae and Microbiota (NCMA, USA), the Commonwealth Scientific and Industrial Research Organisation (CSIRO, Australia), or the National Institute of Technology and Evaluation (NITE, Japan). NCMA has many strains of zooxanthellae available, each of which, when sent, is accompanied by culture information (e.g. culture conditions and media). However, the genetic clade of

Bioassay for Symbiosis-Related Factor

371

zooxanthellae distributed from each organisation is determined according to different papers. It is recommended that their genetic clade is checked [8, 10]. 2.3

Medium

Almost all zooxanthellae can be cultured in Daigo’s IMK medium (commonly used for marine microalgae culture), but the distributor of a zooxanthellae strain may recommend a different medium. For consistency, a medium is usually made from artificial sea water (see Note 1). For our method, the IMK medium is prepared. 1. 100 mL Erlenmeyer flask 2. SILICOSEN (Shin-Etsu Polymer, Tokyo, Japan). 3. A dry heat steriliser. 4. Daigo’s artificial sea water SP for marine microalgae medium (Fuji Film Wako Pure Chemical, Osaka, Japan). 5. Daigo’s IMK medium (Fuji Film Wako Pure Chemical). 6. A 0.22-μm sterilised filter LaboDisk (0.2 μm, Advantec Toyo Kaisha, Tokyo, Japan). 7. A peristaltic pump.

2.4 Cultivating and Observing Zooxanthellae

1. Clean bench. 2. A gridded microscope slide. 3. Cover glass. 4. Fluorescence microscope equipped with blue excitation and digital camera. 5. A low-protein-binding 24-well titreplate (e.g. EZ-BindShut II; 4820-800SP AGC Techno Glass, Shizuoka, Japan). 6. Incubator (for culture) capable of maintaining 25–27  C, with LED light source. 7. Fluorescence plate reader.

2.5 Preparation of Carbohydrate-Fixed Resin

1. 100 mL beaker 2. Epoxy-activated Sepharose 6B (GE Healthcare Bioscience, cat. no. 17048001). 3. 5 mg/mL galactose in water (pH 13) 4. 1 M 2-aminoethanol 5. Acid solution: 0.1 M acetic acid, 0.1 M NaCl (pH 4.0). 6. Neutralise solution: 0.1 M Tris–HCl, 0.1 M NaCl (pH 8.0). 7. Equilibration buffer: 150 mM NaCl, 50 mM Tris–HCl (pH 8.0). 8. 200 mL Erlenmeyer flask 9. A glass filter.

372

Mitsuru Jimbo et al.

10. An aspirator. 11. A Bu¨chner flask. 12. An empty column. 2.6 Lectin Preparation

1. An octocoral Sinularia lochmodes. Similar lectins may be purified from other Sinularia species. 2. Equilibration buffer: 150 mM NaCl, 50 mM Tris-HCl (pH 8.0). 3. Elution buffer: 0.2 M Tris–HCl (pH 8.0).

D-galactose,

150 mM NaCl, 50 mM

4. UV spectrometer a 0.22-μm sterilised filter DISMIC (Advantec Toyo Kaisha). 5. BCA protein assay kit (Thermo Fisher Scientific, cat. no. 23225). 6. Another H-type lectin. A Helix pomatia lectin HPA can be purchased from Merck KGaA, Darmstadt, Germany (L33821MG). Dissolve 1 mg of this lectin in IMK medium, and sterilise it with a 0.22-μm sterilised filter DISMIC (Advantec Toyo Kaisha). 7. D-galactose agarose (Thermo Fisher Scientific, Waltham, MA, USA, cat. no. 20372).

3

Methods Handle zooxanthella strains aseptically.

3.1 Selecting Zooxanthellae

1. Collect coral and crush about 30 mg using a bead crusher or similar apparatus. 2. Add 1000 μL TE buffer and mix well. Freeze the solution overnight at 20  C. 3. Thaw at room temperature, then boil for 5 min. 4. Add 1 μL of the mixture to PCR mix with ‘itsD’ and ‘ITS2rev2’ primers, and perform PCR (cycle: 94  C, 45 s at 51  C, 45 s at 72  C, 60 s  25 cycles) according the protocol in Koike and Yamashita [11]. 5. Purify the PCR product with Sephacryl S-300 as follows: (a) Slit the bottom of a 0.5-mL microtube for use as a minicolumn. (b) Add 200 μL Sephacryl S-300 to the 0.5-mL microtube. (c) Set the minicolumn on the 2-mL tube, and centrifuge at 770  g for 3 min. (d) Transfer the minicolumn to a new 1.5-mL microtube. (e) Add 10 μL water and then 10 μL PCR product to the minicolumn, and centrifuge at 770  g for 8 min.

Bioassay for Symbiosis-Related Factor

373

6. Add 1–2 μL eluate to a new PCR tube and perform cycle sequencing using BigDye Terminator v 3.1 Cycle Sequencing Kit. 7. Perform similarity search against gene database, and determine zooxanthella clade. 8. Purchase zooxanthellae corresponding to the determined clade. 3.2 Preparation of Sterilised Medium

1. Close a 100 mL Erlenmeyer flask with SILICOSEN, and sterilise using dry heat at 180  C for 6 h. 2. Dissolve 36.0 g Daigo’s artificial sea water SP in 1 L distilled water. 3. Add 285 mg Daigo’s IMK medium and dissolve it in the solution. 4. Sterilise the prepared IMK medium with a 0.22-μm sterilised filter LaboDisk using a peristaltic pump (see Note 2). 5. Add about 80 mL filtered IMK medium to the sterile Erlenmeyer flask.

3.3 Cultivating Zooxanthellae and Plotting Growth Curve

1. Add 1–2 mL zooxanthellae culture to sterilised IMK medium in the Erlenmeyer flask prepared in Subheading 3.2 (see Note 3). 2. Incubate flask at 25  C in a 12 h L:D cycle (see Note 4). 3. When the culture reaches the stationary phase, transfer 1–2 mL (drawn by Pasteur pipette) to a new Erlenmeyer flask containing IMK medium to maintain the strain. 4. To measure cell density, recover a 100 μL aliquot of the medium every 2 or 3 days; put 2 μL on a gridded microscope slide, then cover with a cover glass. 5. Count the number of zooxanthellae. Cell density (cells/mL) is calculated by multiplying the number of zooxanthellae in 2 μL by 500. 6. Continue measuring cell density until the stationary phase. 7. Plot cell density against time, from the moment zooxanthellae were added, and determine the cell density for mid- to late-log phase (Fig. 1).

3.4 Measuring Diurnal Change in Zooxanthella Form

1. Add 500 μL IMK medium to three wells of a low-proteinbinding 24-well titreplate. 2. Add 5000 cells of zooxanthellae to each well (see Note 5). 3. Add IMK medium into each well to 1 mL. 4. Culture the plate at 25  C in a 12-h L:D cycle for the next 1–2 days (see Note 6).

Mitsuru Jimbo et al. 106

Zooxanthellae (cells/ml)

374

105 104 103 102 0

5

10

15

20

Time (days)

Fig. 1 Growth curve for zooxanthella NBRC102920 purchased from the National Institute of Technology and Evaluation

Fig. 2 Cultured zooxanthellae

5. Confirm the number of zooxanthellae clumps is low (less than 10% of the total number of zooxanthellae). 6. The next day, after the light is turned on, recover 2 μL of the medium from the upper part of the well every 2 h, place on a gridded microscope slide, then cover with a cover glass. 7. Count the swimming and coccoid cells. The zooxanthellae cell density (cells/mL) is calculated by multiplying the number of zooxanthellae in 2 μL by 500. 8. Alternatively, take six photographs per well without overlap every 2 h (see Note 7). 9. Count the numbers of swimming and coccoid cells. Swimming cells will be blurred while coccoid cells will appear as clear images (Fig. 2).

Bioassay for Symbiosis-Related Factor

375

swiming zooxanthellae (cells)

300

200

100

0 12

18

Time

Fig. 3 Numbers of motile zooxanthellae when the light is on. This shows a transition in the number of motile cells of a strain NBRC102920 isolated from a giant clam Tridacna maxima from Palau

Sum the numbers of swimming and coccoid cells and calculate the percentage of motile cells using (motile cell %) ¼ (motile cells)/(motile cells) + (coccoid cells)  100. 10. Calculate mean  standard deviation from three replicates. 11. Determine a suitable time to perform a bioassay for symbiosisrelated factor. In the case of strain NBRC102920, it is between 12:00 and 14:00 (Fig. 3). 3.5 Counting the Number of Zooxanthellae Using Fluorescence

1. Calculate the cell density of zooxanthellae using the method in Subheading 3.2. 2. Dilute the zooxanthellae culture to 10, 100, 1000, 10,000 and 100,000 cells/mL using tenfold serial dilution. Add 1 mL diluent to a 24-well titreplate. 3. Measure in vivo chlorophyll a fluorescence (Ex. 485  40 nm, Em. 645  40 nm) using fluorescence plate reader. 4. Make standard curve with cell density along X axis and fluorescent intensity along Y axis.

3.6 Preparation of Galactose-Fixed Resin

1. Add 5 g of Epoxy-activated Sepharose 6B to 100 mL ultrapure water in a 100-mL beaker and stand for 60 min. 2. Remove water by aspiration, but take care not to aspirate resins. 3. Add 200 mL of water and resuspend, and collect the resin by vacuum filtration. Repeat this step four times. 4. Resuspend the resin with 50 mL of 5 mg/mL galactose (pH 13), transfer the slurry to a 200-mL Erlenmeyer flask, and incubate at 40  C overnight with shaking.

376

Mitsuru Jimbo et al.

5. Transfer the resin to a glass filter to remove solution using vacuum filtration. 6. Wash the resin three times with 200 mL of 1 M 2-aminoethanol by vacuum filtration. 7. Resuspend the resin with 200 mL of 1 M 2-aminoethanol and transfer to a 200-mL Erlenmeyer flask. 8. Incubate at 40  C overnight with shaking. 9. Transfer the resin to a glass filter to remove the solution using vacuum filtration. 10. Wash the resin four times with 200 mL of acid solution using vacuum filtration, followed by equal volume of neutralise solution. 11. Wash the resin three times with 200 mL of equilibration buffer. 12. Resuspend the resin with equal volume of equilibration buffer. 13. Pour the resin into an empty column and equilibrate by five volumes of equilibration buffer to build the column. 3.7 Preparing a Lectin SLL-2 from the Coral Sinularia Lochmodes

The H-type lectin SLL-2 can be purified by galactose-Sepharose 6B prepared according to Jimbo’s protocol [12]: 1. Collect a sample and store at 80  C before use. 2. Take a piece of sample (~10 g) and weigh it. Crush using homogeniser, beads crusher or a blender. For large-scale preparation (~100 g) use a stamp mill. 3. Add three volumes of equilibration buffer to the crushed sample, and mix. 4. Transfer the solution to 50 mL conical tube and centrifuge at 4  C and 10,000  g for 20 min. 5. Collect supernatant. 6. Apply the supernatant to a column of galactose-fixed resin previously equilibrated with equilibration buffer. 7. Wash the column with five volumes of the same buffer. 8. Elute the bound lectin with elution buffer, and collect eluate with several fractions (5–10 mL each). 9. Measure OD280 to determine SLL-2-containing fractions. 10. Dialyse SLL-2-containing fractions against more than 100 volumes of IMK medium three times (see Note 8). 11. Sterilise the dialysed solution with a 0.22-μm sterilised filter. 12. Quantify protein concentration, for example, with a BCA protein assay kit.

Bioassay for Symbiosis-Related Factor

377

Control DBL HPA SLL 15

10

5

0

cell density (x 104 cells / ml)

0

20

40

60

80

The percentage of motile cells (%)

Fig. 4 Percentage of motile cells and cell density. This shows the effect of lectins on a zooxanthella strain CS-156. SLL-2, HPA and a seed lectin from the herb Dolicos biflorus (DBL) all bind to a sugar chain Forssman antigen 3.8 Bioassay for Symbiosis-Related Factor

Each experiment is replicated three times. For a control, zooxanthellae are cultivated in three wells without additives. 1. Count the number of zooxanthellae in the culture in which cell density and motile percentage are high. Zooxanthellae in the culture should be in mid- to late-log phase. 2. Add 0.5 mL IMK medium to 24-well microtitre plate. 3. Add 1000 zooxanthellal cells to each well. 4. Add IMK medium to each well to 1 mL. 5. Incubate the plate at 25  C in a 12 h L:D cycle for 24 h. 6. Add 100 μL of lectin SLL-2 at 100, 300 and 1000 μg/mL to each well. 7. Measure cell density and motile percentage using the methods in Subheadings 3.3 and 3.4 (Fig. 4).

4

Notes 1. Some zooxanthellae do not grow, or only grow slowly in artificial sea water. Should this occur, natural sea water can be used instead. The condition of the zooxanthellae is important for obtaining consistent results. In this experiment, the algae should be collected at late-log growth phase, during which time cell density and the proportion of motile cells will be high. 2. An autoclave is not suitable for sterilisation when IMK medium is made, as some components are heat labile, and precipitation often results from autoclaving. 3. Zooxanthellae grow slowly when their cell density within a culture medium is low. If cell density is low, add a greater volume of zooxanthellae culture. Half of the culture can be used immediately upon receipt of zooxanthellae from a distributor.

378

Mitsuru Jimbo et al.

4. Zooxanthellal growth varies according to the light used. LED light may be too strong; 50 μmol m2 s1 appears to be satisfactory. 5. A culture should be used before it reaches the stationary phase. This phase affects the number of swimming zooxanthellae; when clumps of zooxanthellae increase, the motile cell percentage decreases. 6. Transfer may lead to zooxanthellae changing to a coccoid form. Several days may pass before reversion to normal diurnal change. 7. Zooxanthellae do not distribute evenly in a well: a representative region in the well should be selected. 8. Some nitrogen-containing compounds affect zooxanthella behaviour [13]. Extracts should be dialysed to remove them.

Acknowledgements We wish to thank Steve O’Shea, PhD, from Edanz Group (www. edanzediting.com/ac), for editing an earlier draft of this manuscript. References 1. Davies PS (1984) The role of zooxanthellae in the nutritional energy requirements of Pocillopora eydouxi. Coral Reefs 2:181–186 2. Koike K, Jimbo M, Sakai R et al (2004) Octocoral chemical signaling selects and controls dinoflagellate symbionts. Biol Bull 207:80–86 3. Jimbo M, Koike K, Sakai R et al (2005) Cloning and characterization of a lectin from the octocoral Sinularia lochmodes. Biochem Biophys Res Commun 330:157–162 4. Kita A, Jimbo M, Sakai R et al (2015) Crystal structure of a symbiosis-related lectin from octocoral. Glycobiology 25:1016–1023 5. Jimbo M, Yamashita H, Koike K et al (2010) Effects of lectin in the scleractinian coral Ctenactis echinata on symbiotic zooxanthellae. Fish Sci 76:355–363 6. Kvennefors ECE, Leggat W, Kerr CC et al (2010) Analysis of evolutionarily conserved innate immune components in coral links immunity and symbiosis. Dev Comp Immunol 34:1219–1229 7. Vidal-Dupiol J, Adjeroud M, Roger E et al (2009) Coral bleaching under thermal stress: putative involvement of host/symbiont recognition mechanisms. BMC Physiol 9:14

8. LaJeunesse TC, Parkinson JE, Gabrielson PW et al (2018) Systematic revision of Symbiodiniaceae highlights the antiquity and diversity of coral endosymbionts. Curr Biol 28:2570–2580.e6 9. Pochon X, Stat M, Takabayashi M et al (2010) Comparison of endosymbiotic and free-living Symbiodinium (dinophyceae) diversity in a hawaiian reef environment. J Phycol 46:53–65 10. Pochon X, Gates RD (2010) A new Symbiodinium clade (Dinophyceae) from soritid foraminifera in Hawai. Mol Phylogenet Evol 56:492–497 11. Koike K, Yamashita H, Oh-Uchi A et al (2007) A quantitative real-time PCR method for monitoring Symbiodinium in the water column. Galaxea 9:1–12 12. Jimbo M, Yanohara T, Koike K et al (2000) The D-galactose-binding lectin of the octocoral Sinularia lochmodes: characterization and possible relationship to the symbiotic dinoflagellates. Comp Biochem Physiol 125:227–236 13. Sjoblad RD, Chet I, Mitchell R (1978) Quantitative assay for algal chemotaxis. Appl Environ Microbiol 36:847–850

Chapter 37 MIC4 from Toxoplasma gondii: A Lectin Acting as a Toll-Like Receptor Agonist Fla´via Costa Mendonc¸a-Natividade, Rafael Ricci-Azevedo, and Maria Cristina Roque-Barreira Abstract Tachyzoites, which are infective forms of Toxoplasma gondii, use their actinomyosin system to move over surfaces and invade host cells. Central to this process is the regulated release of micronemes organelles contents. The microneme protein 4 (MIC4) has the property to recognize galactosides residues linked to glycoproteins on the host cell surface. This property allows that MIC4 binds to TLR2- and TLR4 N-linked glycans and promote the activation of cell innate immune cells and secretion of inflammatory cytokines, acting on resistance against the parasite. Obtention of MIC4 from T. gondii requires several purification steps, is time-consuming and provides low yield. Therefore, this section details the protocol for prokaryotic expression, production, and purification of recombinant MIC4 (rMIC4) and for experimental assays to confirm its biological activity. Key words Toxoplasma gondii, Microneme proteins, Lectins of pathogens, Carbohydrate recognition domain, Production of recombinant protein, Assay of lectin activity on macrophages, Assay of lectin activity on TLR-expressing cells

1

Introduction T. gondii is an intracellular obligate parasite and the causative agent of toxoplasmosis. The fact that there is no warm-blooded species resistant to T. gondii infection may clarify why toxoplasmosis is a widespread disease in the world [1]. The tachyzoite, a fast replicating form of the parasite, actively force its entry into host cells. The process is initiated by contact with the host cell plasma membrane, followed by reorientation and penetration, which is managed by the glideosome, composed by an actomyosin system that underlies the plasma membrane, leads to gliding motility allowing the parasite migration across biological barriers and promotes an active host cell entry and formation of a parasitophorous vacuole [2, 3]. The transition through the mentioned stages requires a highly controlled release of proteins from tachyzoites apical organelles, that is,

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_37, © Springer Science+Business Media, LLC, part of Springer Nature 2020

379

380

Fla´via Costa Mendonc¸a-Natividade et al.

micronemes, rhoptries, and dense granules [4]. The discarded content of micronemes promotes the parasites adhesion to the host cell surface [4]. Among the released microneme proteins, MIC6, MIC1, and MIC4 form a complex, which is anchored through MIC6 into the parasite membrane. This protein complex interacts with host cells through the MIC1 and MIC4 carbohydrate recognition domains (CRD), which bind to glycans with sialic acid and β-galactosides, respectively. Despite found in a complex that is MIC6 anchored to the parasite membrane, soluble MIC4 or MIC1 alone preserve the ability of binding to the same glycoconjugates on the host cell and induce cell activation. MIC4, by recognizing Nglycans linked to TLR2 and TLR4, expressed on innate immune cells, induces NF-kB translocation into the nucleus, which results in activation of proinflammatory genes, including the IL-12 one, and triggering of a protective immune response [5]. MIC4, a 61-kDa protein, has six apple domains. The fifth accounts for the galactosebinding activity, which features MIC4 as a lectin [6]. This indication is reinforced by concomitantly assaying a WT MIC4 and a mutated MIC4 in its CRD, expected to be unable to bind galactose or activate innate immune cells. MIC4, like few other lectins, interacts with TLR N-glycans and induces activation of innate immune cells [7]. The emerging interest in using TLRs agonists as therapeutic tools motivate the protocols description, in this section, to evaluate the functionality of rMIC4.

2

Material

2.1 Plasmid Constructs

1. pEXP17-MIC4 and pET28a-MIC4(K469M) expression constructs producing fusion proteins with N-terminal 6-histidine (6 His) tag (see Note 1). 2. The TLR system of mammalian cells expression composed by: pDisplay-TLR2, pDisplay-TLR1, pDisplay (empty vector), pcDNA3.1- CD14, pcDNA3.1- CD36, pGL2-NF-κB-dependent pELAM-luciferase and pRL-Renilla (see Note 2).

2.2

Cells

1. Rubidium chloride competent E. coli BL21 (DE3) prepared as described by Green and Rogers [8]. 2. Human embryonic kidney 293T (HEK293T) cells, originally acquired from American Tissue Culture Collection (ATCC, Rockville, MD, USA).

2.3

Culture Medium

1. Luria-Bertani (LB) broth supplemented with ampicillin (100 μg/mL), in distilled water. 2. LB agar containing 1.5% of agar and ampicillin (100 μg/mL) and chloramphenicol (34 μg/mL) in distilled water. 3. DMEM medium supplemented with 10% FBS, pH 7.4.

MIC4 is a TLR2 Agonist

381

4. RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS), 10 U/mL penicillin, 10 μg/mL streptomycin, 2 mM L-glutamine, and 25 mM HEPES pH 7.2. 5. RPMI 20/30 medium: RPMI 1640 medium supplemented with 20% FBS, 30% of L-Cell Conditioned Media (from L929 cells), 10 U/mL penicillin, and 10 μg/mL streptomycin, 2 mM L-glutamine, 25 mM HEPES pH 7.2. 2.4 Buffers and Solutions

1. Sample buffer: 250 mM Tris–HCl pH 6.8, 10% (w/v) SDS (sodium dodecyl sulphate), 30% (v/v) glycerol, 0.05%, 4% (v/v) β-mercaptoethanol, 0.05% (w/v) bromophenol blue. 2. Coomassie blue protein stain: 3 g/L Coomassie Brilliant Blue G-250, 45% methanol, 10% glacial acetic acid, and 45% distilled water. 3. Destaining solution: 50% methanol, 12% acetic acid, and 38% distilled water. 4. Disruption Buffer: 20 mM Tris–HCl, 500 mM NaCl, 5 mM EDTA (ethylenediamine tetraacetic acid), 0.1 mM PMSF (phenylmethylsulfonyl fluoride), 0.2 mg/mL lysozyme, 20 μg/mL DNase, 1 mM MgCl2, pH 8.0, in distilled water. 5. Washing buffer I: 20 mM Tris–HCl, 500 mM NaCl, 0.1 mM PMSF and 0.5% Triton X-100, pH 8.0, in distilled water. 6. Washing buffer II: 20 mM Tris–HCl, 500 mM NaCl, 0.1 mM PMSF, pH 8.0, in distilled water. 7. Binding buffer I: 8 M urea, 20 mM Tris–HCl, 500 mM NaCl, 30 mM imidazole, 10 mM β-mercaptoethanol, pH 7.4, in distilled water. 8. Binding buffer II: 8 M urea, 20 mM Tris–HCl, 500 mM NaCl, 5 mM imidazole, 10 mM β-mercaptoethanol, pH 7.4, in distilled water. 9. Elution buffer: 8 M urea, 20 mM Tris–HCl, 500 mM NaCl, 500 mM imidazole, 10 mM β-mercaptoethanol, pH 7.4, in distilled water. 10. Sodium chloride solution: 1.5 M NaCl, in distilled water. 11. Dialysis buffer: 20 mM Tris–HCl, 500 mM NaCl, 50 mM glycine, 0.5 M L-arginine, 5 mM L-cysteine, 1 mM L-cystine, pH 8.0, in distilled water (see Note 3). 12. Phosphate-buffered saline (PBS): 137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KH2PO4, pH 7.4, in distilled water. 13. Mild trypsin: 0.025% trypsin and 0.01% EDTA in sterile PBS.

2.5

Animal

1. A C57/BL6 mouse with 6–12 weeks old.

382

2.6

Fla´via Costa Mendonc¸a-Natividade et al.

Agonists of TLR

1. LPS Ultrapure (from E. coli 0111: B4) (Sigma-Aldrich, St. Louis, MO, USA). 2. Pam3CSK4 (P3C) (EMC Microcollections, Tu¨bingen, DE).

2.7

Others

1. 12- and 96-well plates, clear polystyrene plate, flat bottom, sterile with lid. 2. 15-mL conical sterile polypropylene centrifuge tubes and Pasteur pipette. 3. 70% ethanol, 30% isopropyl alcohol, and 20% ethanol. 4. 90  15 mm Petri dish. 5. Affi-Prep Polymyxin Resin (Bio-Rad, Hercules, CA, USA). 6. Chromatography column Ni Sepharose™ High Performance (GE Healthcare, IL, USA). 7. Dialysis membrane—10 K molecular weight cutoff (Slide-ALyzer™ Dialysis Cassettes, Thermo Scientific, Leics, UK). 8. Isopropyl-β-D-thiogalactopyranoside (IPTG) (Sigma-Aldrich). 9. Kit for endotoxin quantification, Pierce™ LAL Chromogenic Endotoxin Quantitation Kit (Thermo Scientific). 10. Polymyxin B sulphate salt (Sigma-Aldrich). 11. Gauze, scissors, and forceps for small animal surgeries. 12. Tissue culture nontreated Petri dish of 100 mm  20 mm.

3

Methods

3.1 Expression of Recombinant MIC4 (rMIC4) and rMIC4(K469M)



The rMIC4 expression and purification follows the same methodology described for rMIC1, which is the subject of other section in this book. Herein the methods are equally described, however using the specific plasmid-encoding MIC4 and the CRD-truncated form of MIC4. 1. Take rubidium chloride competent E. coli BL21 (DE3) out of 80  C and thaw on ice. 2. Mix gently: 1–5 μL of pEXP17-MIC4 or pET28a-MIC4 (K469M) (10 pg–100 ng) into 50–100 μL of competent cells in a microcentrifuge tube. pEXP17 or pET28a plasmid, and E. coli BL21 (DE3) may be used as positive control and negative control, respectively (see Note 1). 3. Proceed the bacterial transformation according the heat shock transformation method [8, 9]. 4. Plate 50–100 μL of cells in each Petri dish containing LB agar and incubate plates at 37  C overnight.

MIC4 is a TLR2 Agonist

383

5. Pick single colonies from E. coli pEXP17-MIC4 or pET28aMIC4(K469M) plates and inoculate a 5-mL LB broth and grow the cultures overnight at 37  C with vigorous shaking. 6. Transfer the content of the culture and inoculate to 500 mL of fresh LB broth and grow the culture for 2–3 h until reached the optical density (OD600nm) 0.6–0.8. 7. Harvest 1 mL aliquot of rMIC4 and rMIC4(K469M) culture that correspond to sample “time zero” (T0), maintained at 20  C. 8. Add 0.5 mM IPTG, and the culture should grow for additional 4 h. At this point harvest two 1 mL-aliquots from rMIC4 and rMIC4(K469M) that will correspond to sample “time four” (T4), maintained at 20  C. 9. Harvest the culture and the T0- and T4-rMIC4 and –rMIC4 (K469M) aliquots by centrifugation at 4500  g for 20 min at 4  C. If necessary, the cell pellet can be stored at 80  C for a few weeks. 3.2 rMIC4 and rMIC4(K469M) SDS-PAGE

1. Prepare the samples by dissolving the rMIC4 or rMIC4 (K469M) expressing bacteria aliquots (T0 and T4) in 50 μL of distilled water. Mix the sample with sample buffer in the 5:1 proportion. Heat the samples at 100  C for 10 min. 2. Prepare a 1.0-mm thick 12% SDS-PAGE and submit the samples to electrophoresis at 80 V for the initial 15 min and then maintain a constant voltage of 120 V (see Note 4). Stop the electrophoresis when sample buffer has migrated to the end of the gel (approximately 2 h). 3. Stain gel with Coomassie blue, for 2–4 h, until the gel is a uniform blue color. 4. Incubate the gel under slowly stirring with destaining solution until background is clear. The gel can be stored in distilled water, for a few days. 5. Before proceed protocol confirm the increase in the thickness of a band approximately to 60 kDa on T4 lane compared to T0 lane—of rMIC4 and rMIC4(K469M).

3.3 Isolation of Inclusion Bodies Through Centrifugation

1. Resuspend the cell pellets in 20-mL disruption buffer. 2. Stirrer for 30 min at room temperature. 3. Perform the mechanical lysis as follows: (a) Disrupt cell samples, by ultrasonication for 2 min and 40 s, three times, on ice (alternate pulse on 40 s, pulse off 40 s, temperature of probe 4  C, and 70% amplitude). (b) Harvest the samples by centrifugation at 15,200  g for 15 min at 4  C. Remove the supernatants.

384

Fla´via Costa Mendonc¸a-Natividade et al.

(c) Repeat steps 3(a) and (b) (sonication/centrifugation) for three times using 20-mL disruption buffer to resuspend the pellets before sonication. 4. Resuspend the pellets containing the inclusion bodies in 5 mL washing buffer I. Centrifuge the mixtures at 15,200  g for 15 min at 4  C, discard supernatants; repeat wash twice; in the last wash resuspend the samples using washing buffer II. 5. Harvest the samples by centrifugation at 15,200  g for 15 min at 4  C. Remove the supernatants. 3.4 Inclusion Bodies Solubilization

1. Resuspend the pellets in binding buffer I and solubilize the samples, under stirring, for 1 h at room temperature or overnight at 4  C. 2. Harvest the mixtures by centrifugation at 15,200  g for 15 min at 4  C and collect the supernatants.

3.5 Ni-Sepharose Chromatography

1. Wash the Ni-Sepharose columns with 5 volumes of distilled water to remove ethanol included for the maintenance of the columns. 2. Equilibrate the columns with 5–10 volumes of column with binding buffer I. 3. Apply each pretreated sample separately, under slowly stirring, for 1 h at room temperature. 4. Wash with 5 volumes of column using binding buffer II. 5. Elute with 5 volumes of column using elution buffer and store the eluate at 20  C. 6. Analyze preparations purity by SDS-PAGE. 7. Regenerate the Ni-Sepharose columns by washing with 5–10 volumes of binding buffer I. 8. Before store Ni-Sepharose columns: wash with 10 volumes of column using sodium chloride solution. After, 10 volumes of distilled water; 10 volumes of 30% isopropyl alcohol for 20 min and immediately, 10 volumes of distilled water. Store the column at 4  C with 20% ethanol.

3.6 Protein Refolding by Dialysis Method

1. Dialyze the solution of denatured rMIC4 and rMIC4(K469M) against 1 L of freshly made 6, 5, 4, 3, 2, 1, 0.5, and 0 M urea, respectively, with dialysis buffer, pH 8.0, using a dialysis membrane. The proteins should be dialyzed against each one of the urea concentrations, during 24 h at 4  C, totalizing a duration of 8 days. 2. To remove remaining LPS, the Affi-Prep Polymyxin Resin should be used.

MIC4 is a TLR2 Agonist

385

3. The endotoxin contamination may be verified using LAL Chromogenic Endotoxin Quantitation Kit. 4. Prior to use in cell-stimulation experiments, rMIC4 and rMIC4(K469M) preparations should be incubated with 50 μg/mL polymyxin B sulphate salt for 30 min at 37  C to neutralize any residual LPS. 3.7 Obtention, Culture, and Phenotyping of Bone Marrow-Derived Macrophage (BMDM)

1. Euthanasia, by carbon dioxide inhalation, of one mouse that was sprayed it with 70% ethanol. 2. Use scissors to obtain femurs, cutting through the tibia below the knee joints as well as through the pelvic bone, close to the hip joint, and remove muscles connected to the bone using gauze. 3. Place the femurs into a 15-mL polypropylene tube containing sterile PBS on ice. 4. In a tissue culture hood, place the bones in 70% ethanol for 1 min and wash it in sterile RPMI 1640 medium. 5. Cut both epiphyses using sterile scissors and forceps in sterile environment. 6. Flush the bones with a syringe filled with RPMI 1640 to extrude bone marrow into a 15-mL sterile polypropylene tube. 7. Homogenize gently the cell suspension with a Pasteur pipette. 8. Harvest cells by centrifugation at 300  g for 10 min and resuspend the cell pellet in 40 mL RPMI 20/30. Cells can be frozen in this step if needed (see Note 5). 9. Seed the cells in tissue culture nontreated Petri dish and incubated at 37  C in a 5% CO2 humidified atmosphere. 10. On day 4 after the seeding, add an extra 10 mL of fresh RPMI 20/30 medium and incubate for more 3 days. 11. On day 7, remove the supernatant and wash the cells with 10 mL of sterile PBS at 37  C. 12. Add 10 mL ice-cold PBS to each plate and incubate at 4  C for 10 min and then incubate the dishes at 37  C for 2 min. 13. Detach the cells by gently pipetting the PBS around the plate. 14. Harvest the cells by centrifugation at 300  g for 10 min and resuspend in 10 mL of RPMI 1640 medium. 15. Analyze the harvested cells by flow cytometry to determine the proportion of cells with macrophage phenotype (see Note 6).

3.8 The rMIC4 and rMIC4(K469M) Activity on Macrophages

1. Seed and cultivate 1  106/mL BMDMs in 96-well tissue culture microplate, at 37  C in a 5% CO2 humidified atmosphere for 12 h.

386

Fla´via Costa Mendonc¸a-Natividade et al.

Fig. 1 IL-12 Production induced by rMIC4 depends on its carbohydrate recognition domain. Bone marrow-derived macrophages were stimulated with 5 μg/mL rMIC4 or same concentration of its CRD mutated form (rMIC4-K469M), for 48 h. LPS (100 ng/mL) and medium alone were used as positive and negative controls, respectively. IL-12(p40) level was measured by ELISA. ∗p < 0.05 by one-way ANOVA, followed by Bonferroni’s posttest [5]

2. Stimulate the BMDMs with 5 μg/mL rMIC4 or rMIC4 (K469M), using 100 ng/mL LPS Ultrapure and medium as positive and negative controls, respectively. 3. Incubate the cells at 37  C in a 5% CO2 humidified atmosphere for 48 h. 4. Assay the IL-12(p40) concentration in the supernatants of cell cultures, through standard ELISA (see Note 7) (Fig. 1). 3.9 The rMIC4 Activity on TLR2-Transfected HEK293T Cell Line: A Gene Reporter Assay

1. Cultivate the human embryonic kidney (HEK) 293T cells in DMEM, without antibiotics, at 37  C in a 5% CO2 humidified atmosphere. 2. The day before transfection of TLR2, seed 5  105 cells/well on 12-wells plates. 3. At the day of transfection, the HEK293T must be with 70–80% confluence. 4. Proceed the co-transfection with a combination of (1) 0.22 μg TLR2 and 0.22 μg TLR1, and 0.22 μg CD14 and 0.22 μg CD36, 0.22 μg NF-κB-dependent pELAM-luciferase reporter gene construct and 7.4 ng Renilla plasmid, or (2) 0.22 μg NF-κB-dependent pELAM-luciferase construct and 7.4 ng Renilla plasmid, for experimental control. The total amount of DNA in each transfection may be normalized to 2 μg by adding empty vector. Choose the most appropriate transfection reagent and protocol for your lab (see Note 8). 5. Grow the cells for 24 h at 37  C in a 5% CO2 humidified atmosphere.

MIC4 is a TLR2 Agonist

387

Fig. 2 In vitro cell activation induced by rMIC4 is mediated by TLR2. HEK293T cells expressing TLR2/1 were stimulated with 50 nM rMIC4 for detection of either (a) NF-κB activation and (b) IL-8 production. As positive and negative controls, 1 nM P3C and Medium were used, respectively. ∗p < 0.05 by one-way ANOVA, followed by Bonferroni’s post-test

6. Detach the cells from the plate by mild trypsin treatment and harvest them by centrifugation at 300  g, 4  C for 10 min. 7. Seed 4  104 cells per well on 96-well plates and incubate for 24 h at 37  C in a 5% CO2 humidified atmosphere. Remove the supernatant. 8. Stimulate the cells expressing TLR with 50 nM rMIC4, diluted in 100 μL DMEM medium. As positive and negative controls use 1 nM P3C and medium, respectively. Also, stimulate cells of experimental control with either medium or 1 nM P3C. 9. Incubate for 4 h, at 37  C in a 5% CO2 humidified atmosphere, for the luciferase reporter assay (see Notes 9 and 10) (Fig. 2a). 10. Stimulate the cells expressing TLR with 50 nM rMIC4, rMIC4 (K469M), or 1 nM P3C, or medium diluted 150 μL DMEM. Also, stimulate cells of experimental control with either medium or 1 nM P3C. 11. Incubate for 24 h at 37  C in a 5% CO2 humidified atmosphere, and measure the IL-8 concentration in the culture supernatant by ELISA (see Notes 7 and 10) (Fig. 2b).

4

Notes 1. The MIC4 gene can be amplified from cDNA of the T. gondii strain ME49, based on the sequence published in the GenBank (Accession number AF143487.2) with 6 His tag added on the N-terminal and cloned into pEXP17-MIC4. The construct pET28a-MIC4(K469M) with alteration in MIC4’s CRD can be synthesized by GenScript (NJ, USA).

388

Fla´via Costa Mendonc¸a-Natividade et al.

2. Details about the TLR system of mammalian cells expression can be found in Carneiro et al. [10]. 3. To prepare dialysis buffer, get warm the solution and use acetic acid to adjust pH. Mixture will be completely dissolved when it reaches the final pH ¼ 8.0. The dialysis buffer helps the refolding process due the L-cysteine and L-cystine which promote reshuffling of disulphide bonds and arginine that helps to suppress protein aggregation [11, 12]. 4. It is recommended to apply the same concentration of cells to the SDS-PAGE gel according to the optical density (OD) at 600 nm of T0- and T4-recombinant proteins. 5. As described by Marim et al. [13], bone marrow cells can be cryopreserved without any loss on differentiation capacity and viability. 6. Over 92% of harvested cells express high levels of F4/80 antigen. 7. There are many ELISA kits, such as Mouse IL-12(p40) and Human IL-8 ELISA set (OptEIA set, BD Biosciences, San Diego, CA, USA). 8. We recommend the Lipofectamine 2000 Transfection Reagent (Invitrogen, Carlsbad, CA, USA). 9. The Dual Luciferase Reporter Assay System (Promega Cor, Madison, WI) is the most recommended for this assay. 10. The period of stimulation of HEK293 T cells should be of 4 h when the response will be detected through the luciferase reporter assay, and of 24 h when activation will be evaluated through IL-8 measurement [14–16]. References 1. Hill DE, Dubey JP (2018) Toxoplasma gondii. In: Ortega YR, Sterling CR (eds) Foodborne parasites. Springer International Publishing, Cham, pp 119–138 2. Carruthers V, Boothroyd JC (2007) Pulling together: an integrated model of Toxoplasma cell invasion. Curr Opin Microbiol 10:83–89 3. Fre´nal K, Dubremetz JF, Lebrun M, SoldatiFavre D (2017) Gliding motility powers invasion and egress in Apicomplexa. Nat Rev Microbiol 15:645–660 4. Carruthers VB, Sibley LD (1997) Sequential protein secretion from three distinct organelles of Toxoplasma gondii accompanies invasion of human fibroblasts. Eur J Cell Biol 73:114–123 5. Sardinha-Silva A, Mendonc¸a-Natividade FC, Pinzan CF et al (2019) The lectin-specific activity of Toxoplasma gondii microneme

proteins 1 and 4 binds Toll-like receptor 2 and 4 N-glycans to regulate innate immune priming. PLoS Pathog 15:e1007871 6. Marchant J, Cowper B, Liu Y et al (2012) Galactose recognition by the apicomplexan parasite Toxoplasma gondii. J Biol Chem 287:16720–16733 7. Ricci-Azevedo R, Roque-Barreira M-C, Gay NJ (2017) Targeting and recognition of Tolllike receptors by plant and pathogen lectins. Front Immunol 8:1820 8. Green R, Rogers EJ (2013) Transformation of chemically competent E. coli. Methods Enzymol 529:329–336 9. Froger A, Hall JE (2007) Transformation of plasmid DNA into E. coli using the heat shock method. J Vis Exp (6):253. https://doi.org/ 10.3791/253

MIC4 is a TLR2 Agonist 10. Carneiro AB, Iaciura BMF, Nohara LL et al (2013) Lysophosphatidylcholine triggers TLR2-and TLR4-mediated signaling pathways but counteracts LPS-induced NO synthesis in peritoneal macrophages by inhibiting NF-κB translocation and MAPK/ERK phosphorylation. PLoS One 8:e76233 11. Tsumoto K, Umetsu M, Kumagai I et al (2004) Role of arginine in protein refolding, solubilization, and purification. Biotechnol Prog 20:1301–1308 12. Moghadam M, Ganji A, Varasteh A et al (2015) Refolding process of cysteine-rich proteins: Chitinase as a model. Rep Biochem Mol Biol 4:19–24 13. Marim FM, Silveira TN, Lima DS Jr, Zamboni DS (2010) A method for generation of bone marrow-derived macrophages from

389

cryopreserved mouse bone marrow cells. PLoS One 5:e15263 14. Kurt-Jones EA, Popova L, Kwinn L et al (2000) Pattern recognition receptors TLR4 and CD14 mediate response to respiratory syncytial virus. Nat Immunol 1:398–401 15. Lien E, Sellati TJ, Yoshimura A et al (1999) Toll-like receptor 2 functions as a pattern recognition receptor for diverse bacterial products. J Biol Chem 274:33419–33425 16. Medvedev AE, Kopydlowski KM, Vogel SN (2000) Inhibition of lipopolysaccharideinduced signal transduction in endotoxintolerized mouse macrophages: dysregulation of cytokine, chemokine, and toll-like receptor 2 and 4 gene expression. J Immunol 164:5564–5574

Chapter 38 Production and Characterization of MIC1: A Lectin from Toxoplasma gondii Fla´via Costa Mendonc¸a-Natividade, Rafael Ricci-Azevedo, Sandra Maria de Oliveira Thomaz, and Maria Cristina Roque-Barreira Abstract Some lectins of pathogens interact with host cells through the recognition of specific carbohydrates displayed on the mammals’ cell surface. The microneme protein 1 (MIC1) from Toxoplasma gondii has a lectin domain that specifically binds sialic acid residues, often found in the terminal positions of N-glycans of mammalian cells. The necessary studies on the MIC1 biological roles have been limited initially by the laborious purification of the protein from T. gondii tachyzoites and the low yields verified. Then Escherichia coli has been transformed with a construct containing the MIC1 gene, and the obtained recombinant MIC1 (rMIC1) has been purified from the inclusion bodies. Herein, we detail the methodology of heterologous production and purification of rMIC1 and protocols to assay the rMIC1 lectin ability. Key words Toxoplasma gondii, Microneme proteins, Lectins of pathogens, Carbohydrate recognition domain, Production of recombinant protein, Inclusion bodies, Refolding of recombinant proteins, Sugar-binding assay

1

Introduction Most Apicomplexa parasites ensured a successful host infection by developing a microneme arsenal. The more microneme proteins (MICs) an Apicomplexan parasite has, more efficiently it performs the processes of motility, gliding, and active invasion of the host cells [1]. Toxoplasma gondii is a successful Apicomplexa parasite that infects around 30% of the human population [2, 3] and possesses about 16 different MICs. In addition to microneme proteins, specialized organelles into the parasites’ apical pole have rhoptry neck proteins (RONs), which also contribute to the active entry in the host cell [4]. Some MICs act as adhesins [5, 6], which interact tightly with glycoproteins on the host cell membrane, contributing to the virulence of the parasite [7, 8]. The secreted MICs usually form complexes, such as MIC6/MIC1/MIC4, where MIC6 anchors the complex in the parasite’s membrane, whereas MIC1

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_38, © Springer Science+Business Media, LLC, part of Springer Nature 2020

391

392

Fla´via Costa Mendonc¸a-Natividade et al.

and MIC4 are exposed and interact with host cells through its carbohydrate recognition domains (CRD) [6, 9, 10]. A lactosebinding fraction (Lac+), isolated from the soluble tachyzoite antigen of T. gondii RH strain, is constituted of MIC1/MIC4 [11], which is a proper preparation to assay the interactions established by MICs subcomplex and their biological relevance. Nevertheless, Lac+ isolation is a difficult process with many purification steps, which provides only low protein yields. Hence, recombinant MIC1 (rMIC1) and MIC4 (rMIC4) proteins were generated, and they reproduced the biological properties verified for the native Lac+ fraction [10, 12]. MIC1 is a lectin that recognizes oligosaccharides with terminal α(2–3)-sialyl residue linked to β-galactosides [7, 10, 13]. MIC1 structural studies revealed in the N-terminal region two MAR (microneme adhesive repeat) domains—MAR1 and MAR2— which have high affinity and specificity to sialic acid on the surface of mammalian cells. Besides, MIC1 also interacts with MIC4, contributing to the complex formation [7, 9]. In addition to the N-terminal MAR domains, MIC1 has a nonfunctional galectinlike domain that does not interact with the host cells’ surface. Nonetheless, the galectin-like domain is essential for MIC1 association with MIC6, anchored on the parasite membrane [14]. This section details the heterologous production and purification of rMIC1 in Escherichia coli, as well the sugar-binding assay to functionally validate rMIC1 lectinic property.

2

Materials

2.1 The Expression Plasmid

1. pEXP17-MIC1 and pET28a-MIC1(T126A/T220A) expression constructs producing fusion proteins with N-terminal 6-histidine (6 His) tag (see Note 1).

2.2 Competent Cells for Chemical Transformation

1. Rubidium chloride-competent E. coli BL21 (DE3) prepared as described by Green and Rogers [15].

2.3

1. Luria-Bertani (LB) broth supplemented with ampicillin (100 μg/mL), in distilled water.

Culture Medium

2. LB agar containing 1.5% of agar and ampicillin (100 μg/mL) and chloramphenicol (34 μg/mL), in distilled water. 2.4 Buffers and Solutions

1. Sample buffer: 250 mM Tris–HCl pH 6.8, 10% (w/v) SDS (sodium dodecyl sulfate), 30% (v/v) glycerol, 0.05%, 4% (v/v) β-mercaptoethanol, and 0.05% (w/v) bromophenol blue. 2. Coomassie blue protein stain: 3 g/L Coomassie Brilliant Blue G-250, 45% methanol, 10% glacial acetic acid, and 45% distilled water.

The Lectin MIC1 from Toxoplasma gondii

393

3. Destaining solution: 50% methanol, 12% acetic acid, and 38% distilled water. 4. Disruption buffer: 20 mM Tris–HCl, 500 mM NaCl, 5 mM EDTA (ethylenediaminetetraacetic acid), 0.1 mM PMSF (phenylmethylsulfonyl fluoride), 0.2 mg/mL lysozyme, 20 μg/mL DNase, and 1 mM MgCl2, pH 8.0, in distilled water. 5. Washing buffer I: 20 mM Tris–HCl, 500 mM NaCl, 0.1 mM PMSF, and 0.5% Triton X-100, pH 8.0, in distilled water. 6. Washing buffer II: 20 mM Tris–HCl, 500 mM NaCl, and 0.1 mM PMSF, pH 8.0, in distilled water. 7. Binding buffer I: 8 M urea, 20 mM Tris–HCl, 500 mM NaCl, 30 mM imidazole, and 10 mM β-mercaptoethanol, pH 7.4, in distilled water. 8. Binding buffer II: 8 M urea, 20 mM Tris–HCl, 500 mM NaCl, 5 mM imidazole, and 10 mM β-mercaptoethanol, pH 7.4, in distilled water. 9. Elution buffer: 8 M urea, 20 mM Tris–HCl, 500 mM NaCl, 500 mM imidazole, and 10 mM β-mercaptoethanol, pH 7.4, in distilled water. 10. Dialysis buffer: 20 mM Tris–HCl, 500 mM NaCl, 50 mM glycine, 0.5 M L-arginine, 5 mM L-cysteine, and 1 mM Lcystine, pH 8.0, in distilled water (see Note 2). 11. Sodium chloride solution: 1.5 M NaCl, in distilled water. 12. Sodium carbonate buffer: 7.13 g NaHCO3 and 1.59 g Na2CO3, pH 9.5, in 1 L distilled water. 13. Phosphate-buffered saline (PBS): 137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, and 1.8 mM KH2PO4, pH 7.4, in distilled water. 14. PBS-T: 0.05% (v/v) Tween 20 in PBS. 15. PBS-T—3% gelatin: 3% (w/v) gelatin in PBS-T. 16. PBS-T—1% gelatin: 1% (w/v) gelatin in PBS-T. 17. Sodium hydroxide stop solution: 1 M NaOH in distilled water. 18. p-NPP solution: 1 mg/mL of p-NPP ( p-nitrophenyl phosphate) disodium salt may be dissolved in Tris 1 M, pH 9.8, containing 0.3 mM MgCl2. 2.5

Antibodies

1. Anti-Chicken IgY (IgG) (whole molecule)—Alkaline Phosphatase antibody produced in rabbit (Sigma-Aldrich, St. Louis, MO, USA). 2. Polyclonal IgY anti-rMIC1—obtained from the egg yolk of immunized chickens with rMIC1, as described by Akita and Nakai [16].

394

2.6

Fla´via Costa Mendonc¸a-Natividade et al.

Others

1. 90  15 mm Petri dish. 2. 30% isopropyl alcohol and 20% ethanol. 3. 96-Well enzyme immunoassay/radioimmunoassay (EIA/RIA) microplate. 4. Affi-Prep Polymyxin Resin (Bio-Rad, Hercules, CA, USA). 5. Chromatography column Ni Sepharose™ High Performance (GE Healthcare, IL, USA). 6. Dialysis membrane—10 K molecular weight cutoff (Slide-ALyzer™ Dialysis Cassettes, Thermo Scientific, Leics, UK). 7. Fetuin glycoprotein (F2379—Sigma-Aldrich). 8. Gelatin (Sigma-Aldrich). 9. Isopropyl-β-D-thiogalactopyranoside (IPTG) (Sigma-Aldrich). 10. Kit for endotoxin quantification, Pierce™ LAL Chromogenic Endotoxin Quantitation Kit (Thermo Scientific). 11. Polymyxin B sulfate salt (Sigma-Aldrich). 12. Tween 20 (Sigma-Aldrich).

3

Methods

3.1 Expression of Recombinant MIC1 (rMIC1) and rMIC1 (T126A/T220A)

1. Take rubidium chloride-competent E. coli BL21 (DE3) out of 80  C and thaw on ice. 2. Mix gently: 1–5 μL (10 pg–100 ng) of pEXP17-MIC1 and pET28a-MIC1 (T126A/T220A) into 50–100 μL of competent cells in the, respectively, identified microcentrifuge tubes. pEXP17 or pET28a plasmid and E. coli BL21 (DE3) may be used as positive control and negative control, respectively (see Note 1). 3. Incubate the competent cells/DNA mixture on ice for 15 min. 4. Heat shock each transformation tube by placing the tube into a 42  C for 60 s. 5. Put the tubes back on ice for 5 min. 6. Add 800 μL LB media (without antibiotic) to the cells and grow in 37  C shaking incubator for 60 min. 7. Harvest cells by centrifugation at 1000  g for 1 min and resuspend the cell pellet in 200 μL of LB media. 8. Plate 50–100 μL of cells in each Petri dish containing LB agar and incubate plates at 37  C overnight. 9. Pick a single colony from E. coli pEXP17-MIC1 and pet28arMIC1(T126A/T220A) plates, and inoculate a 5 mL LB broth and grow the cultures overnight at 37  C with vigorous shaking.

The Lectin MIC1 from Toxoplasma gondii

395

10. Transfer each culture separately into 500 mL of fresh LB broth and grow the culture for 2–3 h until reaching the optical density (OD600nm) 0.6–0.8. 11. Harvest 1 mL aliquot of rMIC1- and rMIC1(T126A/T220A) culture that correspond to sample “time zero” (T0), maintained at 20  C. 12. Add 0.5 mM IPTG, to induce the protein expression, and the culture should grow for additional 4 h. At this point harvest two 1 mL aliquots from rMIC1 and rMIC1(T126A/T220A) that will correspond to sample “time four” (T4), maintained at 20  C. 13. Harvest the culture and the T0- and T4-rMIC1 and -rMIC1 (T126A/T220A) aliquots by centrifugation at 4500  g for 20 min at 4  C. If necessary, the cell pellet can be stored at 80  C for a few weeks. 3.2 rMIC1 and rMIC1(T126A/ T220A) SDS-PAGE

1. Prepare the samples by dissolving the rMIC1 and rMIC1 (T126A/T220A) expressing bacteria aliquots (T0 and T4) in 50 μL of distilled water. Mix the sample with sample buffer in the 5:1 proportion. Heat the samples at 100  C for 10 min. 2. Prepare a 1.0 mm-thick 12% SDS-PAGE and submit the samples to electrophoresis at 80 V for the initial 15 min, and after maintain a constant voltage of 120 V (see Note 3). Stop the electrophoresis when sample buffer has migrated to the end of the gel (approximately 2 h) (Fig. 1). 3. Stain gel with Coomassie blue, for 2–4 h, until the gel is a uniform blue color. 4. Incubate the gel under slowly stirring with destaining solution until background is clear. 5. The gel can be stored in distilled water, for a few days.

3.3 Isolation of Inclusion Bodies Through Centrifugation

1. Resuspend the cell pellets in 20 mL disruption buffer. 2. Stir for 30 min at room temperature. 3. Perform the mechanical lysis as follows: (a) Disrupt cell samples, by ultrasonication for 2 min and 40 s, three times, on ice (alternate pulse on 40 s; pulse off 40 s; temperature of probe 4  C; and 70% amplitude). (b) Harvest the samples by centrifugation at 15,200  g for 15 min at 4  C. Remove the supernatants. (c) Repeat steps 3(a) and (b) (sonication/centrifugation) for three times using 20 mL disruption buffer to resuspend the pellets before sonication.

396

Fla´via Costa Mendonc¸a-Natividade et al.

Fig. 1 MIC1 heterologous production in E. coli detected by electrophoresis. SDS-PAGE of MIC1 expressing bacteria aliquots before (T0) and after (T4) IPTG addition to the culture. Gel was stained with Coomassie blue. After 4 h induction, a band can be visualized in a migration position correspondent to histidinetagged rMIC1

4. Resuspend the pellets containing the inclusion bodies in 5 mL washing buffer I. Centrifuge the mixtures at 15,200  g for 15 min at 4  C, and discard supernatants; repeat wash twice; in the last wash resuspend the samples using washing buffer II. 5. Harvest the samples by centrifugation at 15,200  g for 15 min at 4  C. Remove the supernatants. 3.4 Inclusion Body Solubilization

1. Resuspend the pellets in binding buffer I and solubilize the samples, under stirring, for 1 h at room temperature or overnight at 4  C. 2. Harvest the mixtures by centrifugation at 15,200  g for 15 min at 4  C and collect the supernatants.

3.5 Ni-Sepharose Chromatography

1. Wash the Ni-Sepharose columns with 5 volumes of distilled water to remove ethanol included for the maintenance of the columns. 2. Equilibrate the columns with 5–10 volumes of column with binding buffer I. 3. Apply each pretreated sample separately, under slow stirring, for 1 h at room temperature.

The Lectin MIC1 from Toxoplasma gondii

397

Fig. 2 rMIC1 homogeneity detected by electrophoresis. SDS-PAGE of rMIC1 and rMIC1(T126A/T220A) after purification steps. Gel was stained with Coomassie blue. An isolated band with approximately 70 kDa can be visualized and indicates the purity of the samples

4. Wash with 5 volumes of column using binding buffer II. 5. Elute with 5 volumes of column using elution buffer and store the eluate at 20  C. 6. Analyze preparations’ purity by SDS-PAGE (Fig. 2). 7. Regenerate the Ni-Sepharose columns by washing with 5–10 volumes of binding buffer I. 8. Before storing Ni-Sepharose columns wash with 10 volumes of column using sodium chloride solution. After, 10 volumes of distilled water; 10 volumes of 30% isopropyl alcohol for 20 min, and immediately 10 volumes of distilled water. Store the column at 4  C with 20% ethanol. 3.6 Protein Refolding by Dialysis Method

1. Dialyze the solution of denatured rMIC1 and rMIC1(T126A/ T220A) against 1 L of freshly made 6, 5, 4, 3, 2, 1, 0.5, and 0 M urea, respectively, with dialysis buffer, pH 8.0, using a dialysis membrane. The proteins should be dialyzed against each one of the urea concentration, during 24 h at 4  C, totalizing a duration of 8 days. 2. To remove remaining LPS, the Affi-Prep Polymyxin Resin should be used.

398

Fla´via Costa Mendonc¸a-Natividade et al.

3. The endotoxin contamination may be verified using LAL Chromogenic Endotoxin Quantitation Kit. 4. Prior to use in cell stimulation experiments, rMIC1 and rMIC1 (T126A/T220A) preparations should be incubated with 50 μg/mL polymyxin B sulfate salt for 30 min at 37  C to neutralize any residual LPS. 3.7 Sugar-Binding Assay

1. Coating 96-well microplates with 5 μg/well of fetuin glycoprotein diluted in 50 μL of sodium carbonate buffer per well. Incubate overnight at 4  C (see Note 4). 2. Aspirate and wash wells three times with 200 μL of PBS-T. 3. Block nonspecific binding by adding 100 μL of PBS-T-3% gelatin for 90 min (see Note 5). 4. Repeat step 2. 5. Add 50 μL rMIC1 or rMIC1(T126A/T220A) (20 μg/mL) diluted in PBS-T-1% gelatin. Incubate overnight at 4  C. 6. Repeat step 2. 7. Add 50 μL of the policlonal IgY anti-rMIC1 as primary antibody (1:5000), in PBS-T-1% gelatin (see Notes 6 and 7). Incubate 1 h at room temperature or overnight at 4  C. 8. Repeat step 2, but total five washes. 9. Add 50 μL of the secondary antibody anti-IgY conjugated to alkaline phosphatase (AP), following the manufacturer’s instructions. Incubate 1 h in the dark at room temperature. 10. Repeat step 2, but with a total of seven washes. 11. Add 50 μL of the p-NPP solution (see Note 8) and incubate in the dark until color reaches a suitable intensity (see Note 9). 12. Stop the reaction by adding 25 μL of sodium hydroxide stop solution. 13. Measure the absorbance at 450 nm in a microplate-scanning spectrophotometer (Fig. 3).

4

Notes 1. To generate the recombinant protein used in the binding assay, MIC1 gene can be amplified from cDNA of the T. gondii strain ME49, based on the sequences published in the GenBank (Accession number Z71786.1), with 6 His tag added to the N-terminal and cloned into pEXP17-MIC1. The plasmid with rMIC1(T126A/T220A), carrying mutations into the CRDs of MIC1 that compromise its lectin property [17], can be synthesized by GenScript (NJ, USA) using a pET28a vector.

The Lectin MIC1 from Toxoplasma gondii

399

Fig. 3 rMIC1 binding to fetuin depends on an intact CRD. The binding to fetuin denoted by OD increasing is done by rMIC1 but not rMIC1(T126A-T220A), thus confirming that only rMIC1 exerts lectin activity. ∗p < 0.05 by one-way ANOVA, followed by Bonferroni’s post-test

2. To prepare dialysis buffer, get warm the solution and use acetic acid to adjust pH. Mixture will be completely dissolved when it reaches the final pH ¼ 8.0. The dialysis buffer helps the refolding process due to the L-cysteine and L-cystine, which promote reshuffling of disulfide bonds and arginine that helps to suppress protein aggregation [18, 19]. 3. It is recommended to apply the same concentration of cells to the SDS-PAGE gel according to the optical density (OD) at 600 nm of T0- and T4-recombinant proteins. 4. Fetuin was used because it is a glycoprotein rich in N-glycans containing both α(2,3)- and α(2,6)-linked sialic acids in the terminal position [20, 21], being the terminal α(2,3)-sialic acid the main target for rMIC1 [6, 10]. 5. To avoid nonspecific interactions, the used blocking and washing buffer contains gelatine, instead albumin or defat milk. Gelatin is virtually free of sugars; as a consequence, it is not recognized by the CRD of the assayed lectin. Purchased albumin is usually contaminated with glycoproteins, while milk composition includes glycoproteins and oligosaccharides, which are potential targets for the assayed lectin CRD. 6. The polyclonal IgY anti-rMIC1 (primary antibody) is purified from egg yolk of chickens immunized with rMIC1. Immunization was performed according to the protocol of Akita and Nakai [16].

400

Fla´via Costa Mendonc¸a-Natividade et al.

7. Remind that some lectins may interact with IgY glycans or other immunoglobulins. In both cases, the occurrence of undesirable binding is revealed by appropriate negative controls. 8. Prepare p-NPP solution immediately before use. 9. Incubation time may vary from 10 to 30 min. It should not exceed 30 min. References 1. Carruthers VB, Tomley FM (2008) Microneme proteins in apicomplexans. Subcell Biochem 47:33–45 2. Robert-Gangneux F, Darde´ M-L (2012) Epidemiology of and diagnostic strategies for toxoplasmosis. Clin Microbiol Rev 25:264–296 3. Montoya JG, Liesenfeld O (2004) Toxoplasmosis. Lancet 363:1965–1976 4. Carruthers VB, Sibley LD (1997) Sequential protein secretion from three distinct organelles of Toxoplasma gondii accompanies invasion of human fibroblasts. Eur J Cell Biol 73:114–123 5. Reiss M, Viebig N, Brecht S et al (2001) Identification and characterization of an escorter for two secretory adhesins in toxoplasma gondii. J Cell Biol 153:563–578 6. Friedrich N, Santos JM, Liu Y et al (2010) Members of a novel protein family containing microneme adhesive repeat domains act as sialic acid-binding lectins during host cell invasion by apicomplexan parasites. J Biol Chem 285:2064–2076 7. Blumenschein TMA, Friedrich N, Childs RA et al (2007) Atomic resolution insight into host cell recognition by Toxoplasma gondii. EMBO J 26:2808–2820 8. Ce´re`de O, Dubremetz JF, Soeˆte M et al (2005) Synergistic role of micronemal proteins in Toxoplasma gondii virulence. J Exp Med 201:453–463 9. Marchant J, Cowper B, Liu Y et al (2012) Galactose recognition by the apicomplexan parasite Toxoplasma gondii. J Biol Chem 287:16720–16733 10. Sardinha-Silva A, Mendonc¸a-Natividade FC, Pinzan CF et al (2019) The lectin-specific activity of Toxoplasma gondii microneme proteins 1 and 4 binds Toll-like receptor 2 and 4 N-glycans to regulate innate immune priming. PLoS Pathog 15:e1007871 11. Lourenc¸o EV, Pereira SR, Fac¸a VM et al (2001) Toxoplasma gondii micronemal protein MIC1

is a lactose-binding lectin. Glycobiology 11:541–547 12. Pinzan CF, Sardinha-Silva A, Almeida F et al (2015) Vaccination with recombinant microneme proteins confers protection against experimental toxoplasmosis in mice. PLoS One 10:e0143087 13. Friedrich N, Matthews S, Soldati-Favre D (2010) Sialic acids: key determinants for invasion by the Apicomplexa. Int J Parasitol 40:1145–1154 14. Saouros S, Edwards-Jones B, Reiss M et al (2005) A novel galectin-like domain from Toxoplasma gondii micronemal protein 1 assists the folding, assembly, and transport of a cell adhesion complex. J Biol Chem 280:38583–38591 15. Green R, Rogers EJ (2013) Transformation of chemically competent E. coli. Methods Enzymol 529:329–336 16. Akita E, Nakai S (1992) Immunoglobulins from egg yolk: isolation and purification. J Food Sci 57:629–634 17. Hager KM, Carruthers VB (2008) MARveling at parasite invasion. Trends Parasitol 24 (2):51–54 18. Tsumoto K, Umetsu M, Kumagai I et al (2004) Role of arginine in protein refolding, solubilization, and purification. Biotechnol Prog 20:1301–1308 19. Moghadam M, Ganji A, Varasteh A et al (2015) Refolding process of cysteine-rich proteins: Chitinase as a model. Rep Biochem Mol Biol 4:19–24 20. Green ED, Adelt G, Baenzigerst JU et al (1988) The asparagine-linked oligosaccharides on bovine fetuin. J Biol Chem 263:18253–18268 21. Baenziger JU, Fiete D (1979) Structure of the complex oligosaccharides of fetuin. J Biol Chem 254:789–795

Chapter 39 Affinity Labeling and Purification of Plant Chitin-Binding LysM Receptor with Chitin Octasaccharide Derivatives Tomonori Shinya, Naoto Shibuya, and Hanae Kaku Abstract Lysin motif (LysM) is a carbohydrate-binding modules found in all kingdoms. LysM binds to N-acetylglucosamine-containing molecules such as peptidoglycan, chitin, Nod factor, and Myc factor and is found in peptidoglycan hydrolases, chitinases, and plant pathogen effectors and plant receptor/co-receptor for defense and symbiosis signaling. This chapter describes the synthesis of a nonradioactive chitin ligand, biotinylated chitin octasaccharide, (GlcNAc)8-Bio, and its application for the detection and characterization of chitin-binding LysM receptor CEBiP in the microsomal membrane fraction of rice suspensioncultured cells by affinity labeling. We also describe the purification of CEBiP from the plasma membrane of the rice cells by affinity chromatography with the synthesized (GlcNAc)8-APEA-CH-Sepharose as an affinity matrix. Key words Affinity labeling, Affinity ligand, Biotinylation, CEBiP, Chitin oligosaccharide, Chitin receptor, LysM, Purification

1

Introduction Lysin motif (LysM) consists of 44–65 amino acid residues and has βααβ secondary structure with two β-strands that form an antiparallel β-sheet [1]. LysM was originally identified in bacterial lytic enzymes [2] and is widely distributed among prokaryotes and eukaryotes as a module in peptidoglycan hydrolases, chitinases, esterases, reductases, amidases, and plant pathogen effectors and plant receptors/co-receptors for the defense or symbiosis signaling [1, 3–8]. Both plants and mammals have the ability to detect infection by recognizing a group of molecules characteristic of microorganisms called microbe-associated molecular patterns (MAMPs) or pathogen-associated molecular patterns (PAMPs) through the pattern recognition receptor (PRR) and to initiate a protective response. Many carbohydrate MAMPs/PAMPs derived from cell wall polysaccharides of microorganisms such as fungal chitin,

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_39, © Springer Science+Business Media, LLC, part of Springer Nature 2020

401

402

Tomonori Shinya et al.

β-glucan, and bacterial peptidoglycan and lipopolysaccharide (LPS) initiate immune responses in plants and mammals. Among these PRRs, LysM-containing PRRs have only been identified in plant defense signaling mediated by carbohydrate MAMPs/PAMPs and have not been found in mammals. Interestingly, LysM receptors also play important roles for plant symbiosis signaling [6, 8, 9]. Rice chitin elicitor-binding protein, CEBiP, is a GPI-anchored LysM receptor for chitin-triggered plant immunity [10]. CEBiP forms a receptor complex with a LysM receptor-like kinase, Oryza sativa chitin elicitor receptor kinase 1 (OsCERK1) for the initiation of chitin defense signaling [11]. This chitin-induced defense response is strictly dependent on the size and structure of N-acetylchitooligosaccharides, i.e., N-acetylchitoheptaose, (GlcNAc)7, and N-acetylchitooctaose, (GlcNAc)8, showed the highest activity, whereas the N-acetylchitooligosaccharides shorter than hexamer or deacetylated forms (chitosan oligosaccharides) showed much lower or no activity [12]. As OsCERK1 shows no detectable chitin binding activity, CEBiP plays a major role for chitin recognition in chitin signaling. Recently, it was indicated from NMR-based epitope mapping that two CEBiP molecules simultaneously bind to one chitin oligosaccharide from the opposite side and form a sandwichtype dimer for the initiation of defense signaling [13, 14]. This model also provided the molecular basis of strict requirement of Nacetyl groups and proper size of N-acetylchitooligosaccharides for the initiation of chitin signaling. Originally, the presence of CEBiP in the rice plasma membrane was detected by affinity labeling with 125I-labeled 2-(4-aminophenyl)ethylamino (APEA) conjugate of (GlcNAc)8, which was synthesized in our laboratory [15]. We also indicated that the chitin binding protein was widespread in plants such as wheat, barley, and carrot by the similar approach [16]. As the experiment with radioactive compounds requires a special facility and care for safety, we developed an alternative method to analyze the interaction of chitin oligosaccharides and corresponding receptors by using a nonradioactive ligand. In this chapter, we describe the synthesis of nonradioactive carbohydrate ligand, biotinylated chitin octasaccharide, (GlcNAc)8-Bio, and its application for affinity labeling to detect and characterize CEBiP in the microsomal membrane fraction (MF) of rice suspension-cultured cells. We also describe the purification of CEBiP from the plasma membrane of the rice cells by affinity chromatography with the synthesized (GlcNAc)8-APEACH-Sepharose as an affinity matrix.

Affinity Labeling of LysM-Receptor

2

403

Materials

2.1 Preparation of Microsomal Membrane Fraction

1. Homogenization buffer: 0.3 M sucrose, 5 mM EDTA, 5 mM EGTA, 50 mM MES (pH 7.6), 20 mM NaF, 1 mM DTT (dithiothreitol), 4 mM SHAM (salicylhydroxamic acid), 2 mM PMSF (phenylmethylsulfonyl fluoride), 2.5 mM Na2S2O5. 2 stock solution of 0.3 M sucrose, 5 mM EDTA, and 5 mM EGTA in 50 mM MES (pH 7.6) can be kept at 4  C for 2–3 weeks. The 20 mM NaF, 1 mM DTT, 4 mM SHAM, 2 mM PMSF, and 2.5 mM Na2S2O5 were added before use. 2. MF buffer: 0.25 M sucrose and 10 mM NaH2PO4 (pH 7.8).

2.2 Preparation of Biotinylated Chitooligosaccharide

1. 1 mg/mL (GlcNAc)8 (see Note 1). 2. 50 mg/mL biocytin hydrazide (EZ-Link™ HydrazideBiocytin, Thermo Fisher Scientific). 3. 50 mg/mL NaCNBH3 (see Note 2). 4. Bio-Gel P2 gel filtration column (1.0 cm ϕ  17.5 cm, Bio-Rad Laboratories). 5. HPLC solvent A, water; HPLC solvent B, methanol. 6. Inertsil ODS-3 (4.6 mm ϕ  250 mm, GL Sciences Inc.).

2.3 Affinity Biotinylation

1. 5 Binding buffer: 50 mM sodium phosphate and 4.5% NaCl (pH 7.2). 2. 3% EGS (ethylene glycol bis[succinimidylsuccinate]) in DMSO (dimethyl sulfoxide) (see Note 3). 3. 1 M Tris (base). 4. SDS-PAGE sample loading buffer: 250 mM Tris–HCl (pH 6.8), 5% SDS (sodium dodecyl sulfate), 5% β-mercaptoethanol, 50% glycerol, and 0.2% bromophenol blue. 5. SDS-PAGE running buffer: 25 mM Tris, 192 mM glycine, and 0.1% SDS. 6. PVDF membrane. 7. Transfer buffer (Towbin buffer): 25 mM Tris, 192 mM glycine, and 20% methanol. 8. Wash solution I (PBS): 20 mM phosphate buffer containing 0.15 M NaCl (pH 7.0). 9. Wash solution II (PBS-T): wash solution I containing 0.1% Tween-20 (pH 7.0). 10. Blocking solution: 5% skim milk in PBS-T. 11. Primary antibody solution: rabbit anti-biotin (1:2000) in 3% skim milk/PBS-T.

404

Tomonori Shinya et al.

12. Secondary antibody solution: HRP-conjugated secondary antibody (goat anti-rabbit IgG antibody 1:2000) in 3% skim milk/PBS-T. 13. Enhanced chemiluminescence (ECL) kit. 2.4 Purification of CEBiP

1. (GlcNAc)8-APEA-CH-Sepharose ϕ  15 cm) (see Note 4).

4B

column

(1

cm

2. Glycine-CH-Sepharose 4B column (1 cm ϕ  13 cm) (see Note 5). 3. Sephadex G-75 column (1 cm ϕ  19 cm). Sephadex G-75 is purchased from GE Healthcare. 4. 1% ovalbumin solution. 5. 0.17 M glycine-HCl (pH 2.3). 6. 0.005 T-TBS: 0.005% Triton X-100, 0.1 M NaCl, 2 mM DTT, 1 mM MgCl2, and 1 mM PMSF in 25 mM Tris–HCl buffer (pH 7.5). 7. 0.5 T-TBS: 0.5% Triton X-100, 0.1 M NaCl, 2 mM DTT, 1 mM MgCl2, and 1 mM PMSF in 25 mM Tris–HCl buffer (pH 7.5) (see Note 6). 8. Chitosan octamer (1 mg/mL) in 0.005 T-TBS (10 mL). 9. Cellohexose (1 mg/mL) in 0.005 T-TBS (10 mL). 10. 1 M Tris (base). 2.5

Equipment

1. Mortar and pestle. 2. Ultracentrifuge. 3. Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS). 4. Freeze dryer. 5. UV spectrophotometer. 6. Gel filtration chromatography system. 7. High-performance liquid chromatography (HPLC) system. 8. Apparatus for SDS-PAGE. 9. Semidry blotting apparatus and power supply. 10. Chemiluminescence imaging system.

3

Methods

3.1 Preparation of Microsomal Membrane Fraction from Rice Cultured Cells

1. Add 1.5–2 times volume of the homogenization buffer to rice suspension cultured cells. 2. Homogenize with a mortar and pestle under cooled condition. 3. Centrifuge the homogenate for 10 min at 12,000  g at 0  C.

Affinity Labeling of LysM-Receptor

405

4. Recentrifuge the supernatant under similar conditions again. 5. Ultracentrifuge the supernatant for 40 min at 100,000  g at 4  C. 6. Resuspend the precipitate in MF buffer. 7. Divide the microsomal membrane preparation (MF, 2–5 mg/ mL) into microtubes, and freeze immediately in liquid nitrogen. 8. Store MF at 80  C until use. The conjugates of biocytin hydrazide and chitooligosaccharides are prepared by reductive amination. A procedure for preparation of biotinylated N-acetylchitooctaose, (GlcNAc)8-Bio, is described below (Fig. 1, see Note 7).

3.2 Preparation of a Biotinylated Chitooligosaccharide

1. Dissolve 10 mg N-acetylchitooctaose (GlcNAc)8 to a concentration of 1 mg/mL in water at 4  C and leave overnight (see Note 8). 2. Add 1 mL of 50 mg/mL biocytin hydrazide to (GlcNAc)8 solution and mix the solution. 3. Add 250 μL of 50 mg/mL NaCNBH3.

(GlcNAc)8-Bio

O HN

HO HO

O NH

CH2OH O HO

COCH3

CH2OH

O

O HO

NH COCH3

6

NH

NH

O

OH CH

N NH

C

CH

(CH2)4

NH2

COCH3

NH

C

(CH2)4

O

CH2OH

S

Cross-link Ligand receptor interaction

NH2

CEBiP Plasma membrane Fig. 1 Schematic diagram of affinity labeling of CEBiP with (GlcNAc)8-Bio. Chemical cross-linker such as EGS forms conjugate of (GlcNAc)8-Bio and chitin binding protein by cross-linking the amino groups of (GlcNAc)8-Bio and the binding protein

406

Tomonori Shinya et al.

4. Heat at 80  C for 60 min and leave overnight at room temperature. 5. Analyze the reaction product by MALDI-TOF MS to confirm (GlcNAc)8-Bio synthesis. 6. Freeze-dry the reaction product. 7. Resuspend the dried sample in 2 mL pure water and divide the suspension into two microtubes. 8. Centrifuge the suspension for 5 min at 12,000  g at room temperature. Transfer the supernatant into a new tube (see Note 9). 9. Resuspend the precipitate in 1 mL water in each microtube, and centrifuge the suspension for 5 min at 12,000  g at room temperature. Transfer the supernatant into a new tube. Repeat this step four more times (see Note 9). 10. Confirm the presence of (GlcNAc)8-Bio in each supernatant fraction by MALDI-TOF MS. Collect (GlcNAc)8-Bio-rich fractions for further purification. 11. Apply the (GlcNAc)8-Bio-containing fraction onto a gel filtration column (e.g., 1 mL sample load to a 1.0 cm ϕ  17.5 cm column) equilibrated with water, and fractionate by eluting with water. 12. Collect the (GlcNAc)8-Bio-containing fractions by measuring the absorbance at 220 nm and by MALDI-TOF MS analysis. Freeze-dry the fractions and resuspend the dried sample in water which is used for further purification. 13. Purify (GlcNAc)8-Bio further by HPLC with an ODS column (see Note 10). After the separation, select the (GlcNAc)8-Biocontaining fraction by MALDI-TOF MS. 14. Freeze-dry the purified fraction and measure the yield (see Note 11). 15. Resuspend the purified (GlcNAc)8-Bio in water and store at 30  C (see Note 12). 3.3 Affinity Labeling of a LysM Protein CEBiP in Rice Microsomal Membrane Fraction

1. Set up reaction mixture with or without competitor such as (GlcNAc)8 (Table 1. see Notes 13 and 14). 2. Incubate the reaction mixture for 30 min at 4  C. 3. Add 4 μL of 6 ng/μL (GlcNAc)8-Bio (final concentration, 0.4 μM) and leave for 60 min on ice. The reaction mixture with competitor contains 40 μM non-labeled (GlcNAc)8. 4. Add 3 μL of 3% EGS (ethylene glycol bis[succinimidylsuccinate]), mix immediately by vortex, and incubate for 30 min at room temperature (see Notes 15 and 16).

Affinity Labeling of LysM-Receptor

407

Table 1 Preparation of reaction mixtures with/without competitor Competitor



+

Rice microsomal membrane (5 μg/μL)

5 μL

5 μL

5 binding buffer

6 μL

6 μL

0.2 μg/μL (GlcNAc)8

0 μL

10 μL

Water

15 μL

5 μL

5. In order to stop cross-link reaction, add 3 μL of 1 M Tris solution, mix by vortex, and leave at room temperature for 5 min (see Note 17). 6. Add 9 μL of 5 SDS sample buffer and heat at 95  C for 5 min. 7. Load the samples onto a SDS-PAGE gel and perform gel electrophoresis. 8. Transfer the proteins from the SDS-PAGE gel to a PVDF membrane by using a blotting apparatus at 15 V for 40 min. 9. Transfer the PVDF membrane into the blocking solution, and shake gently for 30 min on a shaker at room temperature. 10. Discard the blocking solution. 11. Wash the membrane in PBS-T for 5 min and in PBS for 5 min. 12. Transfer the PVDF membrane into the primary antibody solution, and shake gently for 1 h on a shaker at room temperature (see Note 18). 13. Discard the antibody solution. 14. Wash the membrane once in PBS-T for 5 min and two times in PBS for 5 min. 15. Transfer the PVDF membrane into the secondary antibody solution, and shake gently for 1 h on a shaker. 16. Wash the membrane once in PBS-T for 5 min and three times in PBS for 5 min. 17. Use ECL kit for the detection of HRP-conjugated antibody. 18. Visualize biotinylated protein(s) using a chemiluminescence imaging system (Fig. 2). 3.4 Preparation of Affinity Columns for the Purification

1. Two precolumns of Sephadex G-75 and glycine-CH-Sepharose 4B are connected with (GlcNAc)8-APEA-CH-Sepharose 4B column (Fig. 3, see Note 19). 2. The connected columns are successively prewashed with 1% ovalbumin solution and 0.17 M glycine-HCl (pH 2.3) and equilibrated with 0.005 T-TBS before use.

Tomonori Shinya et al.

(GlcNAc)8 kDa

-

+

150 100 75

CEBiP

50

Fig. 2 Example of Western blot after affinity biotinylation of rice microsomal membrane. High-affinity binding protein for chitin oligosaccharides, CEBiP, is visualized as a biotinylated protein. Biotinylation of CEBiP by (GlcNAc)8-Bio is completely inhibited with 100-fold excess of (GlcNAc)8

A Precolumns

408

B c

a

b kDa 97.4 66 46 30 21.5 14.3

c

(a) Sephadex G-75 (b) Glycine-CH-Sepharose (c) (GlcNAc)8-APEA-CH-Sepharose

1

2

Fig. 3 Purification of CEBiP from the plasma membrane of rice cell. (A) Affinity columns used for purification. Precolumns were packed with Sephadex G-75 (a) and glycine-CH-Sepharose 4B (b), respectively. Column (c), the major affinity column of (GlcNAc)8-APEA-CH-Sepharose 4B. (B) Silver staining of the SDS-PAGE gel. Lane 1, eluate from the column (c) with the elicitor-inactive sugar (cellohexaose and chitosan octasaccharide) solution, Lane 2, CEBiPcontaining fraction eluted from column (c) with 0.17 M glycine-HCl (pH 2.3). Arrowheads indicate CEBiP and its limited proteolysis product

Affinity Labeling of LysM-Receptor

3.5 Purification of CEBiP from Rice Plasma Membrane

409

1. Homogenize the plasma membrane fraction (PM) with glass homogenizer under cooled condition (see Note 20). 2. Add 0.5 T-TBS (0.6 mg PM/mL) to the homogenate and keep the solution for 60 min at 4  C. 3. Ultracentrifuge the solution for 60 min at 200,000  g at 4  C. 4. Apply the supernatant slowly on the connected columns (Fig. 3). 5. Wash with 0.005 T-TBS (30 mL) and subsequently wash with the elicitor-inactive sugar solution of chitosan octamer (1 mg/ mL) and cellohexose (1 mg/mL) in 0.005 T-TBS (10 mL), and then remove the precolumns (Fig. 3, see Note 21). 6. Elute the CEBiP-containing fraction with 0.17 M glycine-HCl (pH 2.3) from the (GlcNAc)8-APEA-CH-Sepharose 4B column, collect by using a fraction collector, and neutralize the eluate immediately with 1/10 volume of 1 M Tris solution to avoid inactivation (see Notes 22 and 23).

4

Notes 1. Chitooligosaccharides are commercially available from IsoSep and Elicityl. We usually prepare (GlcNAc)8 by acetylation of (GlcN)8 [15]. 2. Prepare the NaCNBH3 reagent immediately before use. 3. Prepare the cross-linker reagent immediately before use. 4. (GlcNAc)8-APEA can be prepared with (GlcNAc)8 and 2-(4-aminophenyl)ethylamine (APEA) by reductive amination [15]. (GlcNAc)8-APEA is coupled with the activated CH-Sepharose 4B (GE Healthcare) according to the manufacturer’s protocol. The remaining excess active groups in the beads are blocked with glycine. 5. Glycine-CH-Sepharose 4B can be prepared by incubating the activated CH-Sepharose 4B beads with glycine solution according to the manufacturer’s protocol. 6. Triton X-100 should be the highly purified grade (e.g., polyethylene glycol mono-p-isooctylphenyl ether from Nacalai Tesque, INC) [10]. 7. Refer also to the manufacturer’s protocol for EZ-Link™ Hydrazide-Biocytin. 8. Sonication is a way to increase the solubility of (GlcNAc)8. 9. (GlcNAc)8-Bio is more soluble in water than (GlcNAc)8; therefore the supernatants obtained by the first few extractions are rich in (GlcNAc)8-Bio.

410

Tomonori Shinya et al.

10. An example of HPLC condition for (GlcNAc)8-Bio separation is as follows: column, Inertsil ODS-3 (GL Sciences Inc., 4.6 mmϕ  250 mm); detection 210 nm and flow rate 0.5 mL/min; solvent A, water, and solvent B, methanol; and gradient condition: 0–40 min, 0–50% solvent B. Non-labeled (GlcNAc)8, biocytin hydrazide, and (GlcNAc)8-Bio were eluted in this order under this condition. 11. Typically the recovery of (GlcNAc)8-Bio is in the range 0.5–1 mg from 10 mg (GlcNAc)8. 12. (GlcNAc)8-Bio can be used for kinetic analysis of chitin binding protein by using surface plasmon resonance (SPR) system as well as affinity labeling. We demonstrated kinetic analysis of LysM proteins secreted from phytopathogens by Biacore with Sensor Chip SA on which (GlcNAc)8-Bio is conveniently immobilized [17, 18]. 13. It was possible to enhance interaction between ligand and binding protein by addition of a detergent, Triton X-100, at around 0.1–1.0% (w/v). 14. To determine the specific nature of the binding between ligand and binding protein, set up control reaction mixtures with excess (10–100 times) of non-labeled ligand. To ensure binding specificity, compounds with similar structure to a target ligand are used for competitor compounds. Shorter chitooligosaccharides such as (GlcNAc)5 and chitosan octasaccharide, (GlcN)8 which is deacetylated (GlcNAc)8, are examples for affinity labeling assay with (GlcNAc)8-Bio [5, 15]. 15. Glutaraldehyde (2.5%) and DTSSP (3,30 -dithiobis[sulfosuccinimidyl propionate]) (3%) can be used as chemical crosslinkers. Cross-linker reagents are prepared immediately before use. 16. As the cross-linker, e.g., EGS, reacts with primary amines, it is recommended to avoid buffers containing primary amine during cross-linking reaction. For the same reason, excess of Tris or other amine-containing buffer can be used for quenching EGS. Refer also to manufacturer’s protocol for EGS. 17. Biotinylated CEBiP obtained by this step can be purified by streptavidin beads [19]. DTSSP is a useful cross-linker to elute the affinity-bound protein from streptavidin beads since the disulfide bond in DTSSP spacer arm can be easily cleaved with reducing agents. Increase of (GlcNAc)8-Bio enhances recovery of CEBiP. 18. Alternatively, streptavidin-HRP can be used for detection of biotinylated protein. In the use of streptavidin-HRP, BSA (bovine serum albumin) is used for blocking reagent but not skim milk due to contamination of biotin.

Affinity Labeling of LysM-Receptor

411

19. The attachment of two precolumns is used to prevent the contamination of the nonspecifically bound proteins. 20. The plasma membrane is purified from the rice microsomal membrane fraction by aqueous two-phase partitioning (see Subheading 3.1) [20]. The use of purified plasma membrane can eliminate the background contamination. 21. The elution with the elicitor-inactive sugar solution and the attachment of two precolumns are used to prevent the contamination of the nonspecifically bound proteins in the final eluate. 22. The collecting microtubes should contain 1/10 volume of 1 M Tris solution in advance, so that the eluate can be neutralized immediately to recover active form of CEBiP. 23. If necessary, the protein in the eluate can be concentrated as follows. The 1/10 volume of 5 M NaCl solution and 4.5 times volume of methanol are added to each tube and left at 80  C overnight. The tubes are centrifuged for 2 h at 20,630  g at 4  C, and the concentrated protein is obtained as a precipitate.

Acknowledgments This article was supported by JSPS KAKENHI Grant Number JP18H02208 to H.K., JP18K05558 to T.S., and MEXTSupported Program for the Strategic Research Foundation at Private Universities 2014–2018 (S1411023) from MEXT, Japan, to H.K. References 1. Buist G, Steen A, Kok J et al (2008) LysM, a widely distributed protein motif for binding to (peptido)glycans. Mol Microbiol 68:838–847 2. Ponting CP, Aravind L, Schultz J et al (1999) Eukaryotic signalling domain homologues in archaea and bacteria. Ancient ancestry and horizontal gene transfer. J Mol Biol 289:729–745 3. Akcapinar GB, Kappel L, Sezerman OU et al (2015) Molecular diversity of LysM carbohydrate-binding motifs in fungi. Curr Genet 61:103–113 4. Zhang XC, Cannon SB, Stacey G (2009) Evolutionary genomics of LysM genes in land plants. BMC Evol Biol 9:183 5. Shinya T, Motoyama N, Ikeda A et al (2012) Functional characterization of CEBiP and CERK1 homologs in arabidopsis and rice reveals the presence of different chitin receptor systems in plants. Plant Cell Physiol 53:1696–1706

6. Shinya T, Nakagawa T, Kaku H et al (2015) Chitin-mediated plant-fungal interactions: catching, hiding and handshaking. Curr Opin Plant Biol 26:64–71 7. Desaki Y, Kohari M, Shibuya N et al (2019) MAMP-triggered plant immunity mediated by the LysM-receptor kinase CERK1. J Gen Plant Pathol 85:1–11 8. Miyata K, Kozaki T, Kouzai Y et al (2014) The bifunctional plant receptor, OsCERK1, regulates both chitin-triggered immunity and arbuscular mycorrhizal symbiosis in rice. Plant Cell Physiol 55:1864–1872 9. Kelly S, Radutoiu S, Stougaard J (2017) Legume LysM receptors mediate symbiotic and pathogenic signalling. Curr Opin Plant Biol 39:152–158 10. Kaku H, Nishizawa Y, Ishii-Minami N et al (2006) Plant cells recognize chitin fragments for defense signaling through a plasma

412

Tomonori Shinya et al.

membrane receptor. Proc Natl Acad Sci U S A 103:11086–11091 11. Shimizu T, Nakano T, Takamizawa D et al (2010) Two LysM receptor molecules, CEBiP and OsCERK1, cooperatively regulate chitin elicitor signaling in rice. Plant J 64:204–214 12. Yamada A, Shibuya N, Kodama O et al (1993) Induction of phytoalexin formation in suspension-cultured rice cells by N-acetyl-chitooligosaccharides. Biosci Biotechnol Biochem 57:405–409 13. Hayafune M, Berisio R, Marchetti R et al (2014) Chitin-induced activation of immune signaling by the rice receptor CEBiP relies on a unique sandwich-type dimerization. Proc Natl Acad Sci U S A 111:E404–E413 14. Squeglia F, Berisio R, Shibuya N et al (2017) Defense against pathogens: structural insights into the mechanism of chitin induced activation of innate immunity. Curr Med Chem 24:3980–3986 15. Ito Y, Kaku H, Shibuya N (1997) Identification of a high-affinity binding protein for N-acetylchitooligosaccharide elicitor in the plasma membrane of suspension-cultured rice cells by affinity labeling. Plant J 12:347–356

16. Okada M, Matsumura M, Ito Y et al (2002) High-affinity binding proteins for N-acetylchitooligosaccharide elicitor in the plasma membranes from wheat, barley and carrot cells: conserved presence and correlation with the responsiveness to the elicitor. Plant Cell Physiol 43:505–512 17. Mentlak TA, Kombrink A, Shinya T et al (2012) Effector-mediated suppression of chitin-triggered immunity by Magnaporthe oryzae is necessary for RICE blast disease. Plant Cell 24:322–335 18. Takahara H, Hacquard S, Kombrink A et al (2016) Colletotrichum higginsianum extracellular LysM proteins play dual roles in appressorial function and suppression of chitintriggered plant immunity. New Phytol 211:1323–1337 19. Shinya T, Osada T, Desaki Y et al (2010) Characterization of receptor proteins using affinity cross-linking with biotinylated ligands. Plant Cell Physiol 51:262–270 20. Yoshida S, Uemura M, Niki T et al (1983) Partition of membrane-particles in aqueous 2-polymer phase system and its practical use for purification of plasma-membranes from plants. Plant Physiol 72:105–114

Chapter 40 Purification of GNA-Related Lectins from Natural Sources Els J. M. Van Damme Abstract The Galanthus nivalis lectin, abbreviated as GNA, is the model protein for a large group of mannosebinding lectins. Here, we describe the purification of GNA starting from dry bulbs. Using a combination of ion exchange chromatography and affinity chromatography on mannose-Sepharose, a highly pure preparation of GNA can be obtained. Key words Agglutination activity, Glycan binding, GNA, Plant lectin, Snowdrop

Abbreviations GNA

1

Galanthus nivalis agglutinin

Introduction In 1987 a new type of mannose-binding lectin was first isolated and characterized from the bulbs of snowdrop (Galanthus nivalis), further referred to as the G. nivalis agglutinin (or abbreviated as GNA) [1]. GNA is a 50 kDa tetrameric protein, composed of non-covalently linked 12.5 kDa subunits. Sequence analysis and X-ray diffraction analysis revealed that GNA represented a novel family of carbohydrate binding proteins with a unique threedimensional fold. The so-called GNA domain consists of approximately 110 amino acids and forms a typical β-prism fold, consisting of three four-stranded β-sheets organized around a pseudosymmetry axis [2]. The three β-sheets each harbor a mannose-binding site, resulting in three mannose-binding sites for the tetrameric GNA. Since 1987 GNA-related lectins have been identified in numerous plant species. Many GNA-related lectins have originally been purified from bulbs of monocot species, such as Narcissus sp. and Allium sp.. Therefore, the terms “monocot mannose-binding

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_40, © Springer Science+Business Media, LLC, part of Springer Nature 2020

413

414

Els J. M. Van Damme

lectins” and “bulb lectin” were introduced as a collective name [3– 7]. However, both terms are confusing since GNA-related lectins are not confined to the taxa of the monocotyledons (Liliopsida) [8] and have also been identified and purified from vegetative tissues (e.g., Allium sativum leaf lectin [9]). A recent study focusing on the evolutionary relationships of lectin domains revealed that proteins with GNA domains occur in a wide range of species from lower plants (gymnosperms, ferns, chlorophytes) to monocots and dicots. The family of GNA-related lectins represents one of the largest lectin families in plants [8, 10, 11]. Furthermore the GNA domain has also been reported in bacteria [12], fungi [13], and animals [14]. Therefore, it is advisable to name this lectin family after the first identified member, in casu the G. nivalis agglutinin or GNA [15]. The carbohydrate binding domain equivalent to the snowdrop lectin is referred to as the “GNA domain,” corresponding to PF01453 in the Pfam database. Based on the data obtained with hapten inhibition assays, GNA was originally considered a strictly mannose-specific lectin [1]. In addition, GNA binds to terminal mannose residues in N-glycans [16]. Glycan array analyses demonstrated that GNA (and other related lectins) interacts only weakly with the monosaccharide mannose and shows a stronger binding toward oligomannosides and high-mannose N-glycans [17]. Furthermore some two-domain lectins (e.g., the tulip lectin TxLC-I [18]) interact also with complex N-glycans [19]. In some lectins one or two carbohydrate sites of the GNA domain are inactive [19, 20]. For instance, in the sweet protein curculin from Curculigo latifolia fruits, all three sites of the GNA domain are inactive, resulting in a protein that is completely devoid of sugar binding activity [21]. Most GNA-related lectins are hololectins composed of 12–14 kDa subunits and exist either as monomeric, dimeric, or tetrameric proteins. However, sequences in which the GNA domain is linked to one or more other protein domains have also been reported and thus encode chimeric proteins [6, 19]. Examples are the S-locus glycoproteins from Brassicaceae species composed of a GNA domain, an S-locus glycoprotein domain, and a C-terminal plant PAN/APPLE-like (PAN-A) domain. S-locus receptor kinases contain a domain combination similar to S-locus glycoproteins with an additional transmembrane domain and protein kinase domain. Though there is clear sequence similarity with the GNA domain the carbohydrate binding activity of most of these chimeric proteins remains to be demonstrated. Because of its unique carbohydrate binding properties toward terminal mannose residues, GNA (immobilized on a column) is commonly used for the purification of glycoproteins or for the detection of mannose in samples using histochemistry

GNA-Related Lectins

415

[22, 23]. In planta GNA and GNA-related lectins are part of the plant defense system, e.g., toward herbivores [24–26]. Recently the GNA domain was also shown to be involved in defense signaling [27]. This chapter describes the purification scheme to obtain reasonable quantities of a highly purified lectin.

2

Materials 1. Bulbs of G. nivalis. 2. Waring blender. 3. Centrifuge with cooling. 4. 20 mM 1,3-propanediamine. 5. 0.1 M Tris–HCl pH 8.7 containing 0.5 NaCl. 6. 1 M NaCl. 7. 2 M ammonium sulfate. 8. 1 M HCl. 9. SP Sepharose Fast Flow (GE Healthcare) matrix. 10. Mannose-Sepharose (Sigma-Aldrich) matrix.

3

Methods Carry out all procedures at room temperature unless otherwise specified.

3.1

Lectin Extraction

1. Clean the bulbs, and remove all layers with dead cells (see Note 1) 2. Put the clean bulbs in 20 mM 1,3-propanediamine buffer containing some thiourea (see Note 2). 3. Whole bulbs (50 g) are homogenized in 500 mL buffer with a blender. 4. Pass the obtained extract through a coarse sieve (see Note 3). 5. Centrifuge the extract at 10,000  g for 15 min at 4  C. 6. Add CaCl2 to the supernatant in a concentration of 1 g/L, set the pH at 10–10.5, and store the extract in the cold room at 4  C, at least overnight (see Note 4). 7. Centrifuge the extract at 20,000  g for 10 min (with cooling) and collect the supernatant. 8. Set the pH of the supernatant at 3 using 4 M acetic acid (see Note 5). 9. Centrifuge the extract for 10 min at 20,000  g (with cooling) and collect the supernatant (see Note 6).

416

Els J. M. Van Damme

3.2 Lectin Purification

3.2.1 Ion Exchange Chromatography

G. nivalis (snowdrop) bulbs contain very high concentrations of the lectin (approximately 10% of the total protein is lectin). The lectin is stable in a pH range of 3–12. To purify GNA, ion exchange chromatography is used, followed by specific binding of the lectin to a column with immobilized mannose. 1. Equilibrate the SP Sepharose Fast Flow column with 50 mM acetic acid (pH 4) (see Note 7). 2. Load the extract on the column (see Note 8). 3. Wash the column with 50 mM acetic acid till the OD280 of the flow-through is below 0.2. 4. Desorb the lectin from the column with a solution of 0.1 M Tris–HCl pH 8.7 containing 0.5 M NaCl. Fractions are collected and tested for lectin activity (see Notes 9–11). 5. Regenerate the matrix by washing with 1 M NaCl and distilled water.

3.2.2 Affinity Chromatography on Mannose-Sepharose 4B

1. Equilibrate the mannose column with 1 M ammonium sulfate. 2. Pool the fractions (obtained after ion exchange chromatography) with lectin activity (OD > 0.3), and bring the solution to 1 M ammonium sulfate ((NH4)2SO4) by adding an equal volume of 2 M ammonium sulfate. Set the pH of the lectin solution to 7, using 1 M HCl. 3. Apply the lectin solution on the mannose column (see Note 11). 4. After loading, wash the column with 1 M ammonium sulfate till the OD280 is below 0.2. 5. Elute the protein bound to the mannose column using 20 mM 1,3-diaminepropane-HCl (pH 10–11) (see Notes 12 and 13). 6. Regenerate the mannose column by washing with 20 mM 1,3-diaminepropane-HCl (pH 10–11) and water.

4

Notes 1. Bulbs can be purchased from a local shop; dry resting bulbs contain the highest lectin concentration. Lectin concentration in the bulbs is much higher than in the vegetative tissues. 2. A low concentration of thiourea is added to prevent oxidation reactions in the extract. 3. A household sieve (as for flour) can be used to remove large particles. 4. Incubation at 4  C results in the precipitation of several unwanted proteins and other compounds.

GNA-Related Lectins

417

5. At pH 3 conditions are created in which GNA is stable and remains soluble, while several other proteins denature and precipitate. 6. During the centrifugation the precipitate was completely spun down and a clear extract was obtained. If necessary, the extract can be filtered through glass wool or filter paper. 7. SP Sepharose Fast Flow is a sulfopropyl (SP) strong cation exchange chromatography resin for fast protein purification. 8. The lectin will be retained on the column. From time to time, test the flow-through for agglutination activity (see Note 9). This test will allow to check if the column gets saturated. Also measure the OD280 of all fluids running through the column (flow-through, washing solution, eluate). 9. For agglutination assays, mix 10 μL of crude protein extract with 10 μL of 1 M ammonium sulfate and 30 μL of a 2% solution of trypsin-treated rabbit erythrocytes (see Note 10). To obtain a negative control, mix 20 μL of 1 M ammonium sulfate and 30 μL of a 2% solution of trypsin-treated rabbit erythrocytes. To obtain a positive control, mix 10 μL of a purified lectin solution (1 μg/μL) with 10 μL of 1 M ammonium sulfate and 30 μL of a 2% solution of trypsin-treated rabbit erythrocytes. 10. Transfer 10 μL aliquots of the diluted extracts to glass tubes (0.5 cm diameter) or polystyrene 96 U-welled microtiter plates, and mix with 10 μL of 1 M ammonium sulfate and 30 μL of a 2% solution of trypsin-treated rabbit erythrocytes. Incubate the samples at room temperature. The lectin in the samples will bind to the carbohydrates present on the surface of the red blood cells and as such will agglutinate these cells. 11. The flow-through is regularly tested for agglutination activity. If the column is saturated, stop loading the protein. 12. Alternatively, the lectin can be eluted with mannose, but high concentrations of the sugar (0.2 M or higher) are required. 13. The purity of the lectin can be analyzed by SDS-PAGE (Fig. Fig. 1). Normally, a >95 % pure lectin solution will be obtained. If needed, an additional chromatography step on Q Sepharose Fast Flow (strong anion exchanger) can be performed to obtain 100% purity.

Acknowledgments This work was supported mainly by grants from Ghent University and the Fund for Scientific Research-Flanders.

418

Els J. M. Van Damme

180 130 100 70 55 40 35

25

15

10 Fig. 1 Electrophoresis pattern of purified GNA (15 μg) obtained after ion exchange chromatography and affinity chromatography on mannoseSepharose. Electrophoresis was performed on a 15% acrylamide gel. The molecular mass of the marker proteins is shown on the right in kDa References 1. Van Damme EJM, Allen AK, Peumans WJ (1987) Isolation and characterization of a lectin with exclusive specificity towards mannose from snowdrop (Galanthus nivalis) bulbs. FEBS Lett 215:140–144 2. Hester G, Kaku H, Goldstein IJ et al (1995) Structure of mannose-specific snowdrop (Galanthus nivalis) lectin is representative of a new plant lectin family. Nat Struct Mol Biol 2:472–479 3. Van Damme EJM, Smeets K, Peumans WJ (1995) The mannose-binding monocot lectins and their genes. In: Pusztai A, Bardocz S (eds) Lectins: biomedical perspectives. Taylor and Francis, London, pp 59–80 4. Van Damme EJM, Peumans WJ, Barre A et al (1998) Plant lectins: a composite of several distinct families of structurally and evolutionary related proteins with diverse biological roles. Crit Rev Plant Sci 17:575–692 5. Van Damme EJM, Peumans WJ, Pusztai A et al (1998) Handbook of plant lectins: properties and biomedical applications. John Wiley & Sons, Chichester, p 452

6. Van Damme EJM, Lannoo N, Peumans WJ (2008) Plant lectins. Adv Bot Res 48:107–209 7. Chandra NR, Kular N, Jeyakani J et al (2006) Lectindb: a plant lectin database. Glycobiology 16:938–946 8. Barre A, Bourne Y, Van Damme EJM et al (2019) Overview of the structure-function relationships of mannose-specific lectins from plants, algae and fungi. Int J Mol Sci 20:254 9. Smeets K, Van Damme EJM, Verhaert P et al (1997) Isolation, characterization and molecular cloning of the mannose-binding lectins from leaves and roots of garlic (Allium sativum L.). Plant Mol Biol 33:223–234 10. Van Holle S, De Schutter K, Eggermont L et al (2017) Comparative study of lectin domains in model species: new insights into evolutionary dynamics. Int J Mol Sci 18:1136 11. Van Holle S, Van Damme EJM (2019) Messages from the past: new insights in plant lectin evolution. Front Plant Sci 10:36 12. Parret AH, Schoofs G, Proost P et al (2003) Plant lectin-like bacteriocin from a

GNA-Related Lectins rhizosphere-colonizing Pseudomonas isolate. J Bacteriol 185:897–908 13. Fouquaert E, Peumans WJ, Gheysen G et al (2011) Identical homologs of the Galanthus nivalis agglutinin in Zea mays and Fusarium verticillioides. Plant Physiol Biochem 49:46–54 14. Tsutsui S, Tasumi S, Suetake H et al (2003) Lectins homologous to those of monocotyledonous plants in the skin mucus and intestine of pufferfish, Fugu rubripes. J Biol Chem 278:20882–20889 15. Van Damme EJM, Barre A, Rouge´ P et al (2004) Cytoplasmic/nuclear plant lectins: a new story. Trends Plant Sci 9:484–489 16. Shibuya N, Goldstein IJ, Van Damme EJM et al (1988) Binding properties of a mannosespecific lectin from the snowdrop (Galanthus nivalis) bulb. J Biol Chem 263:728–734 17. Van Damme EJM, Smith DF, Cummings R et al (2008) Glycan arrays to decipher the specificity of plant lectins. In: Wu AM (ed) The molecular immunology of complex carbohydrates. Kluwer Academic/Plenum Publishers, New York, pp 841–854 18. Van Damme EJM, Brike´ F, Winter HC et al (1996) Molecular cloning of two different mannose-binding lectins from tulip bulbs. Eur J Biochem 236:419–427 19. Van Damme EJM, Nakamura-Tsurata S, Smith DF et al (2007) Phylogenetic and specificity studies of two-domain GNA-related lectins: generation of multispecificity through domain duplication and divergent evolution. Biochem J 404:51–61

419

20. Barre A, Van Damme EJM, Peumans WJ et al (1996) Structure-function relationship of monocot mannose-binding lectins. Plant Physiol 112:1531–1540 21. Barre A, Van Damme EJM, Peumans WJ et al (1997) Curculin, a sweet-tasting and tastemodifying protein, is a non-functional mannose-binding lectin. Plant Mol Biol 33:691–698 22. Van Damme EJM (2011) Lectins as tools to select for glycosylated proteins. In: Gevaert K, Vandekerckhove J (eds) Methods in molecular biology – gel-free proteomics, vol 753. Springer, LLC, New York, pp 289–297 23. Go´mez-Santos L, Alonso E, Dı´az-Flores L et al (2017) Transdifferentiation of mucous neck cells into chief cells in fundic gastric glands shown by GNA lectin histochemistry. Tissue Cell 49:746–750 24. Hilder VA, Powell KS, Gatehouse AMR et al (1995) Expression of snowdrop lectin in transgenic tobacco plants results in added protection against aphids. Transg Res 4:18–25 25. Peumans WJ, Van Damme EJM (1995) Lectins as plant defense proteins. Plant Physiol 109:347–352 26. Vandenborre G, Smagghe G, Van Damme EJM (2011) Plant lectins as defense proteins against phytophagous insects. Phytochemistry 72:1538–1550 27. Kim NH, Lee DH, Choi DS et al (2015) The pepper GNA-related lectin and PAN domain protein gene, CaGLP1, is required for plant cell death and defense signaling during bacterial infection. Plant Sci 241:307–315

Chapter 41 Expression, Purification, and Applications of the Recombinant Lectin PVL from Psathyrella velutina Specific for Terminal N-Acetyl-Glucosamine Oriane Machon and Annabelle Varrot Abstract The lectin PVL from the mushroom Psathyrella velutina is the founding member of novel family of fungal lectins. It adopts a seven bladed β-propeller presenting six binding sites specific for the recognition of non-reducing terminal N-acetyl-glucosamine (GlcNAc). The latest can be mainly found in glycoconjugates presenting truncated glycans where aberrant β-GlcNAc terminated glycans represent tumor markers. It can also be found in O-GlcNAcylated proteins where disruption of the O-GlcNAcylation homeostasis is associated with many physiopathological states. The recombinant PVL lectin proved to be a very powerful tool for labelling terminal GlcNAc antigens displayed by extracellular glycoconjugates but also by OGlcNAcylated proteins found in the cytoplasm and nucleus. This chapter will describe how to produce and purify recombinant PVL and several applications for rPVL as probe for the detection of terminal OGlcNAc. Key words Terminal O-GlcNAc, Tumor markers, O-GlcNAcylation detection, Lectin probe, ELLA

1

Introduction Glycosylation is one of the most prominent and diverse forms of posttranslational modification (PTM) of proteins and lipids. It is performed by an incredibly complex and dynamic biosynthetic machinery affected by scores of genetic and environmental factors. It results in a non-templated glycome, where glycans have a key role on mediating or modulating the function of their carriers. The glycome carries cell signatures of health and disease as glycosylation defects or alterations are commonly connected with physiopathological states such as cancers, chronic inflammatory diseases, or infections [1]. N-Acetyl-glucosamine (GlcNAc) is one of the major glycan moieties found in glycoconjugates of the extracellular matrix where it is usually β-linked and rarely exposed in terminal position on the surface of healthy tissues except in gastric mucosa where

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_41, © Springer Science+Business Media, LLC, part of Springer Nature 2020

421

422

Oriane Machon and Annabelle Varrot

α1,4-GlcNAc-capped mucins, secreted by gland mucous cells, are exposed [2]. Truncated N- and O-glycans and glycosphingolipids displaying aberrant non-reducing terminal β-GlcNAc are considered as tumor markers: i.e., truncated core 3 O-glycans (GlcNAcβ13GalNAcα-O-Ser/Thr) in colorectal cancers [3]. A simple and distinct O-glycosylation called O-GlcNAcylation has been uncovered in the early 1980s [4]. It consists of the addition of a single GlcNAc to specific serine or threonine residues of cytoplasmic, nuclear, and mitochondrial proteins. O-GlcNAcylation is ubiquitous in metazoans and widespread in living beings, where it plays a key role in the regulation of multiple cellular processes from epigenetics and stress response to cell signaling. It is a highly dynamic modification interplaying extensively with phosphorylation. Disturbance of O-GlcNAcylation homeostasis can be associated with the pathogenesis of many diseases such as diabetes, cancers, or Alzheimer disease (reviewed in [5, 6]). GlcNAc is a small, non-charged, and labile moiety to be readily detectable by conventional mass spectrometry and gel electrophoresis. The detection of O-GlcNAcylation remains challenging, and the lack of efficient tools for this has hampered both the discovery and the study of this essential modification. Several fungal lectins highly specific for non-reducing terminal GlcNAc epitope have been isolated from mushrooms. The first one, PVL (Psathyrella velutina lectin), has been isolated from Lacrymaria velutina formally known as P. velutina [7]. It is also able to recognize non-reducing terminal N-acetylneuraminic acid (Neu5Ac) residues but with lesser affinity [8]. The resolution of PVL structure uncovered a novel family of lectins characterized by a seven bladed β-propeller fold [9]. Only three fungal members have been structurally characterized to date: PVL, AANL (Agrocybe aegerita lectin 2 formerly called AAL2), and PAL (P. asperospora lectin) [10–12]. Their structures in complex with GlcNAc gave the molecular basis for the recognition of terminal GlcNAc as well as those for terminal Neu5Ac in the case of PVL. Six binding sites in the form of a shallow pocket were identified at the interface between blades apart on the one involved in the Velcro closure of the propeller (Fig. 1). A signature motif was identified for this family (PropLec7B) that can now be used to predict new members to this family in translated genomes and sequence databases [13]. PVL and AANL can be produced in recombinant form without impairing their structure and recognition aptitudes [10, 11]. A clear avidity effect could be observed for rPVL where its affinity increases from 132 μM for single binding event as observed for GlcNAc on PVL-coated sensor chip by surface plasmon resonance or in solution by isothermal microcalorimetry [10] to 60 nM in the case of multivalent presentation of rPVL on GlcNAc-coated SPR chips [14]. Lectins from this family present therefore great potential in

Expression, Purification and Applications of the PVL Lectin

423

Fig. 1 Surface and cartoon representation of the seven-bladed β-propeller of recombinant PVL in complex with GlcNAc-β1,3Gal disaccharide depicted in balls and sticks (PDB 4UP4)

biomedical applications as reagent for the detection of terminal OGlcNAc and for glycoconjugates separation [8, 15]. rPVL can be conjugated with different molecules (fluorophore, biotin, or peroxidation) and proved to be a very powerful probe to label both truncated glycans on human cancer cells and tissues by FACS, immunofluorescence, or histochemistry [10] and GlcNAcylated proteins using western blot [14]. rPVL can also be used in enzyme-linked lectin assay (ELLA) to detect terminal GlcNAc in glycoconjugates. In this chapter, we will detail protocols used for the expression, purification of rPVL, as well as possible conjugation and its use in different detection methods.

2

Materials Prepare all solutions using analytical grade reagents and ultrapure water and filter on 0.22 μM membrane before usage. Follow waste disposal regulations.

2.1 Production of rPVL

1. pET25b-rPVL plasmid encoding for the PVL protein obtained according to previously published protocol [10]. 2. Escherichia coli Tuner™ (DE3) cells (Novagen-Merck). 3. Lennox L Broth Base (LB, Invitrogen). Dissolve 20 g in 1 L of distilled water and autoclave 20 min at 121  C. 4. 100 mg/mL of ampicillin (Amp, culture grade sodium salt) in water. 5. 1 M isopropyl-β-D-thiogalactopyranoside (IPTG).

424

Oriane Machon and Annabelle Varrot

6. 1 L polycarbonate centrifuge bottle (Nalgene) and 50 mL sterile conical centrifuge tubes. 7. Centrifuge with adapted rotors. 2.2 Purification of rPVL

1. 0.45 μm polyethersulfone (PES) syringe filter. 2. C10/10 column with flow adapter (GE Healthcare Life Sciences) or empty Econo-Pac® Chromatography Columns (Bio-Rad). 3. LNT2 affinity resin (GLY011-Gel, Elicityl, Crolles, France). 4. Endonuclease such as Benzonase or Denarase at 250 U/μL. 5. One Shot table top cell disruptor (Constant Systems Ltd). 6. Binding buffer: 20 mM Tris–HCl pH 7.5, 150 mM NaCl, and 100 μM CaCl2. 7. Washing buffer: 20 mM Tris–HCl pH 7.5, 1 M NaCl, and 100 μM CaCl2. 8. Elution buffer: 20 mM Tris–HCl pH 7.5, 150 mM NaCl, 100 μM CaCl2, and 500 mM GlcNAc (Carbosynth, UK). 9. 12% SDS-PAGE gel with vertical gel electrophoresis unit and universal power supply. 10. Precision Plus Protein™ Unstained Standards (Bio-Rad). 11. InstantBlue™ (Expedeon). 12. Snakeskin dialysis tubing 3500 MWCO (Thermo Fisher Scientific).

2.3 Hemagglutination Assays (HA)

1. U-shaped 96-well plates (Nunc, Rochester, NY).

2.4 Affinity Measurements by Surface Plasmon Resonance (SPR)

1. Biacore X100 (GE Healthcare Life Sciences).

2. 4% rabbit erythrocytes in 150 mM NaCl.

2. CM5 sensor chips (GE Healthcare Life Sciences). 3. Running buffer: PBS-Ca (phosphate-buffered saline:10 mM Na2HPO4, 1.76 mM KH2PO4, 2.7 mM KCl, 137 mM NaCl pH 7.4 supplemented with 100 μM CaCl2) as running buffer. Dissolve one tablet (Sigma-Aldrich, ref. P4417) in 200 mL of water. 4. Amine coupling kit (GE Healthcare Life Sciences). 5. Streptavidin from Streptomyces avidinii (Sigma-Aldrich) at 100 μg/mL in 10 mM sodium acetate pH 5.0 for chip immobilization. 6. 1 M ethanolamine for blocking the sensor surface. 7. β-D-GlcNAc-PAA-biotin and α-L-fucose-PAA-biotin (Lectinity, Moscow). 8. GlcNAc (Carbosynth, UK).

Expression, Purification and Applications of the PVL Lectin

2.5

rPVL Labelling

425

1. N,N-Dimethylformamide (DMF). 2. D-Salt polyacrylamide desalting column 6K (Pierce). 3. Biotinamidohexanoyl-6-aminohexanoic acid N-hydroxysuccinimide ester (Sigma-Aldrich) for biotinylation. 4. EZ-Link™ Plus Activated Peroxidase kit (Thermo Fisher Scientific) for coupling with horseradish peroxidase. 5. Alexa Fluor™ 488 NHS Ester (ref. A20000, Thermo Fisher Scientific) for coupling with fluorophore.

2.6 Enzyme-Linked Lectin Assays (ELLA)

1. 96-Well plates flat bottom Microlon® (Greiner Bio-One) or Nunc MaxiSorp™ (Thermo Fisher Scientific). 2. Biotinylated rPVL. 3. PBS buffer pH 7.4. 4. PBS-BSA: PBS supplemented with 0.1% (w/v) Grade V bovine serum albumin (BSA, Sigma-Aldrich, ref. A7030). 5. Tween 20 (Sigma-Aldrich). 6. PAA-β-N-acetylglucosamine (Lectinity, Moscow). 7. Streptavidin-peroxidase (Sigma-Aldrich, ref. S5512) and 3,30 ,5,50 -tetramethylbenzidine (TMB, Thermo Fisher Scientific) for detection of biotinylated lectins. 8. TECAN Spark M10 plate reader.

2.7 Western Blotting of O-GlcNAcylated Proteins

1. Cell lines are obtained according to previously published protocol [14]. 2. Lysis buffer: 10 mM Tris–HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% (v/v) Triton X-100, 0.5% (w/v) sodium deoxycholate, 0.1% (w/v) SDS, and protease inhibitors. 3. 10% SDS-PAGE gel with Mini-PROTEAN® Tetra electrophoresis system, Mini Trans-Blot® Cell, and universal power supply (Bio-Rad). 4. Nitrocellulose membrane (GE Healthcare Life Sciences). 5. Ponceau S (Sigma-Aldrich). 6. Grade V BSA for membrane saturation. 7. Tris-buffered saline (TBS)-Tween buffer: 15 mM Tris–HCl, 140 mM NaCl, and 0.05% (v/v) of Tween 20 (Sigma-Aldrich), pH 8.0. 8. HRP-labelled rPVL at 1 mg/mL in TBS. 9. Chemiluminescence imaging (Fusion Solo system, Vilber Lourmat Collegien).

426

Oriane Machon and Annabelle Varrot

2.8 Immunofluorescence Analysis on Whole Cells

1. Cells lines cultured according to recommended protocols. 2. Four-chamber culture slides. 3. Ice-cold PBS buffer pH 7.4. 4. Paraformaldehyde and BSA. 5. Alexa Fluor 488 labelled rPVL at 5 μg/mL in PBS-BSA 2%. 6. BX41 microscope equipped with a DP70 digital camera system (Olympus, Tokyo, Japan). 7. Pseudo-confocal microscope ApoTome equipped with AxioCam MRm (N/B).

2.9 Immunohistochemistry

1. Fixed tissue sections. 2. Biotinylated rPVL at 0.7 μg/mL. 3. 3% (v/v) hydrogen peroxide or 5% (w/v) bovine serum albumin in PBS buffer. 4. Streptavidin at 0.1 mg/mL and biotin at 0.5 mg/mL. 5. Streptavidin-peroxidase. 6. HRP detection kit (DAB, Ventana Medical Systems, or AEC, Vector Laboratories). 7. Mayer’s hematoxylin solution (Merck, Whitehouse Station, NJ). 8. Sialidase (New England acetylhexosaminidase.

Biolabs)

and

ß-D-N-

9. BX41 microscope equipped with a DP70 digital camera system (Olympus, Tokyo, Japan) or NanoZoomer slide scanner (Hamamatsu, Hamamatsu City).

3

Methods All measurements are done at room temperature (22  C) unless otherwise stated. The protein concentration was verified by measuring optical density at 280 nm by using a theoretical extinction coefficient of 65,890 M1/cm and a molecular weight of 42,974 Da and a NanoDrop 2000 (Thermo Fisher Scientific).

3.1 Production of rPVL

1. Inoculate 1 L of LB media supplemented with 100 μg/mL of ampicillin with 30 mL of an overnight preculture of Escherichia coli Tuner™ (DE3) cells harboring the pET25b-rPVL plasmid, and place in an incubation shaker at 160 rpm and 37  C. 2. Grow the cells until an OD600 of 0.6 and switch the incubator temperature to 16  C. 3. When the culture reached an OD600 of 0.8, induce rPVL expression by adding IPTG to a final concentration of 0.1 mM. 4. Culture the cells overnight for at least 16 h at 160 rpm.

Expression, Purification and Applications of the PVL Lectin

427

5. Harvest the cells by 10 min centrifugation at 5000  g. 6. Transfer the cells in a sterile 50 mL plastic tube and weight the cell pellet (see Note 1). 3.2 Purification of rPVL

1. Resuspend the cell pellet with cold binding buffer at the rate of 5 mL of binding buffer per gram of wet cells using a vortex. 2. Add 1 μL of endonuclease and incubate on a rotating wheel for 30 min. 3. Lyse the cells using the One Shot cell disruptor at a pressure of 1.9 kbars (see Note 2). 4. Clarify the sample by centrifugation at 24,000  g and 4  C for 30 min. 5. Filter the supernatant using a 0.45 μM PES syringe filter. 6. Load the filtered supernatant on 2 mL Toyopearl-LnT2 resin (see Note 3) equilibrated with binding buffer for purification by affinity chromatography, and apply binding buffer until OD280nm reaches a baseline. Collect fractions of 5 mL. 7. Wash unspecifically bound proteins with 10 column volumes (CVs) of washing buffer and then 5 CVs of binding buffer. 8. Elute with 10 CVs of elution buffer and collect fractions of 1 mL. 9. Analyze fractions of each step by gel electrophoresis on 12% SDS-PAGE gel after adding denaturing loading buffer to each fraction aliquots and heating at 100  C for 5 min. Load 10 μL of fraction as well as protein standards prior running the gel at 200 V in Tris-glycine buffer. Unpack the gel and rinse it with water prior coloration with 10 mL of InstantBlue™ for 1 h under low shaking. Rinse and store the gel in water (Fig. 2).

Fig. 2 12% SDS-PAGE gel electrophoresis of rPVL purification. I insoluble fraction, S soluble fraction, FT unbound fraction, W1 wash with equilibration buffer, W2 fraction for washing step with 1 M NaCl, L ladder, 1–10, eluted fractions. Ladder is 250, 150, 100, 75, 50, 37, 25, 20, and 15 kDa from top to bottom

428

Oriane Machon and Annabelle Varrot

10. Pool fractions containing >95% pure protein and dialyze extensively against PBS buffer supplemented with 100 μM CaCl2 for 7 days prior storage at 4  C (see Note 4). 3.3 Hemagglutination Assays (HA)

Hemagglutination permits to check rapidly the binding activity of rPVL and its ability to attach to surface glycoconjugates of the red blood cells. Non-agglutinating cells will sediment at the bottom of the well and form a red button, while agglutinating cells will form a diffuse network preventing the cells’ sedimentation. 1. Dilute rabbit erythrocytes (RBC) to a 4% solution in 150 mM NaCl. 2. Deposit 25 μL of 150 mM NaCl in a 96-well plate. 3. Deposit 25 μL of rPVL at 1 mg/mL in the first well and perform a twofold dilution series. 4. Add 50 μL of 4% RBCs and incubate 1 h. 5. Determine hemagglutination unit (HAU) as the minimum concentration of lectin required to observe hemagglutination (see Note 5). 6. Use 150 mM NaCl as negative control and lectin known to hemagglutinate as positive control.

3.4 Affinity Measurements by SPR

Surface plasmon resonance (SPR) experiments allow to monitor the interaction between two molecules in real time. In a flow channel, one partner (ligand) is immobilized on a biosensor chip, and the other (analyte) is passing over the ligand surface. All experiments were performed on a Biacore X100 at 25  C in PBS-Ca as running buffer at a flow rate of 30 μL/min. 1. Activate two channels by injecting 340 μL of a fresh mixture of EDC/NHS for 400 s using a classical amine coupling procedure. 2. Inject streptavidin for 500 s. 3. Inactivate remaining reacting species by injecting 240 μL of 1 M ethanolamine for 400 s. 4. Immobilize biotinylated PAA-β-GlcNAc at 200 μg/mL for 10 min on channel 2 (see Note 6). 5. Inject rPVL on both channels in series of twofold dilutions between 0 and 2 μM (140 μL, dissociation 400 s). 6. Regenerate the chip after each injection with 0.5 M GlcNAc in water for 30 s. 7. Evaluate the data by using the Biacore X100 evaluation software, and measure binding as resonance units over time after blank subtraction. Determine the dissociation constant by

Expression, Purification and Applications of the PVL Lectin

429

Fig. 3 Sensograms for the interaction of rPVL (various concentrations) injected on a CM5 chip coated with streptavidin/biotin-PAA-β-D-GlcNAc leading here to a Kd of 42 nM after steady-state analysis

plotting response at equilibrium against analyte concentration (Fig. 3). 8. Perform at least duplicates. 3.5

rPVL Labelling

3.5.1 Biotinylation

All reactions are done at room temperature using essentially manufacturer instructions. Unbound label is separated from labelled rPVL and buffer is exchanged to PBS + 100 μM CaCl2 using a D-salt polyacrylamide desalting column 6K (see Note 7). Fractions of 500 μL are collected and their optical density is measured at 280 nm. Labelled rPVL is stored at 20  C in aliquots of 250 μL. 1. Prepare a solution of biotinamidohexanoyl-6-aminohexanoic acid N-hydroxysuccinimide ester at 100 mM in DMF. 2. Add 6 mmol of biotin-NHS-ester to a solution of 10 mg/mL of rPVL in PBS buffer pH 7.4 (see Note 8). The volume added (μL) is calculated according to the following equation: V ¼

mprotein ðmgÞ  6 mmol 1, 000, 000  g  100 mM: M protein mol

3. Incubate 30 min under agitation. 3.5.2 Horseradish Peroxidase Coupling

1. Reconstitute 1 mg of lyophilized EZ-Link Plus Activated Peroxidase with 100 μL of ultrapure water to obtain a solution at 10 mg/mL. 2. Prepare rPVL at 1 mg/mL in PBS buffer pH 7.4 (see Note 9). 3. Incubate 1 mL of rPVL with 100 μL of HRP solution for 1 h. 4. Add 10 μL of sodium cyanoborohydride to the reaction in the fume hood, and incubate at room temperature for 15 min. 5. Add 20 μL of quenching buffer and incubate for 15 min.

430

Oriane Machon and Annabelle Varrot

3.5.3 Alexa Fluor 488 Coupling

1. Prepare a solution of Alexa Fluor 488 NHS Ester at 10 mg/mL in DMF (see Note 10). 2. Prepare a solution of rPVL at 10 mg/mL in 50 mM sodium bicarbonate buffer pH 9.7. 3. Add 100 μL of reactive compound to 1 mL of rPVL drop by drop while stirring. 4. Incubate for 1 h under constant stirring. 5. Calculate the degree of labelling (DOL) according to the following formula after correcting the protein concentration at 280 nm for the contribution of the dye (Aprot ¼ A280  A495  0.11): DOL ¼

3.6 Enzyme-Linked Lectin Assay

A 495nm  MW protein ½protein  71000

1. Prepare 96-well plates by incubating 100 μL of 5 μg/mL solution of β-GlcNAc-PAA diluted in PBS buffer pH 7.4 overnight at room temperature (see Note 11). 2. Wash thrice the plate with 200 μL/well of PBS. 3. Saturate wells by 200 μL/well of 2% (w/v) BSA in PBS for 1 h at 37  C. 4. Wash five times the plate with 200 μL/well PBS-Tween 20 0.05%, and washes are repeated between each step until the end of the assay. 5. Dilute biotinylated rPVL in PBS-BSA in a range from 0 to 1000 ng/well. 6. Deposit 100 μL/well and incubate it for 2 h at room temperature. 7. Add 100 μL/well of 0.25 μg/mL of streptavidin-peroxidase diluted in PBS-BSA. 8. Incubate for 25 min at room temperature. 9. Wash the plate and reveal 3,30 ,5,50 -tetramethylbenzidine.

with

100

μL/well

of

10. Stop the revelation by addition of 100 μL/well of 0.1 M of hydrogen chloride. 11. Read the optical density at 450 nm with a plate reader. 3.7 Western Blotting of O-GlcNAcylated Proteins

1. Lyse cell lines with lysis buffer. 2. Deposit equal amounts of protein on 10% SDS-PAGE under reducing conditions. 3. Electroblot the gel on nitrocellulose using Mini Trans-Blot® Cell for 1 h at 100 V in cold transfer buffer (see Note 12).

Expression, Purification and Applications of the PVL Lectin

431

Fig. 4 Merge microscopy images of cancer cells treated for 30 min at 37  C with 5 μg/mL rPVL labelled with Alexa Fluor 488. Blue channel shows nuclei labelled with DAPI staining, and green channel shows rPVL-Alexa Fluor 488. Human cancer cell lines: MDA-MB-231, derived from metastatic site of mammary gland/breast; H358, derived from bronchioalveolar carcinoma and non-small cell lung cancer; and H441, derived from papillary adenocarcinoma

4. Color the nitrocellulose with Ponceau red staining to verify efficiency of the transfer and equal loading (see Note 12). 5. Saturate the membrane for 45 min with 4% (w/v) Grade V BSA in TBS-T (see Note 12). 6. Deposit HRP-rPVL 1:5000 for 1 h at 4  C (see Note 12). 7. Wash the membrane three times with TBS-T for 10 min each. 8. Use chemiluminescence imaging to reveal the membrane. 3.8 Immunofluorescence Analysis on Whole Cells

1. Inoculate four-chamber culture slides with 4  104 cells per chamber for 24 h. 2. Wash the cells with ice-cold PBS. 3. Fix the cells with 2% paraformaldehyde for 10 min at 4  C. 4. Wash three times with cold PBS for 5 min each. 5. Saturate slides with 1% (w/v) BSA in PBS. 6. Wash again three times with cold PBS for 5 min each. 7. Deposit rPVL-Alexa 488 on the cells to stain them (Fig. 4, see Note 13). 8. Examine cells with BX41 microscope and a pseudo-confocal microscope ApoTome.

3.9 Immunohistochemistry

1. Deparaffinize tissue sections (see Note 14). 2. Block endogenous peroxidases by adding 3% (v/v) hydrogen peroxide in PBS for 5 min (see Note 15). 3. Saturate sections with 5% (w/v) BSA in PBS for 30 min. 4. Deposit biotinylated rPVL at 0.7–1 μg/mL for 1 h at room temperature for ethanol-fixed sections or 2 μg/mL for formalin-fixed sections, and wash it twice with PBS.

432

Oriane Machon and Annabelle Varrot

Fig. 5 Labelling of tumoral tissues with rPVL. Colon sections of tumoral or adjacent tissue obtained from tumor surgical resections before the law 88-138 of December 20, 1988 from patients # 5345 stained with 1 μg/mL rPVL-biot in the presence of 0.1 M fucose or 0.1 M GlcNAc followed by streptavidin-HRP. Slides were imaged using a NanoZoomer slide scanner with a 20 magnification. (Image partly reproduced from [10])

5. Add indirect biotin-streptavidin system, and use the corresponding detection kit according to the manufacturer’s instructions (see Note 16). 6. Wash the developed slides twice with PBS. 7. Perform counterstaining using hematoxylin. 8. Wash with water, dehydrate, and mount the sections. 9. Observe the sections with BX41 microscope or imaged with a NanoZoomer slide scanner (Fig. 5).

4

Notes 1. Cells can be resuspended in some LB prior transfer in the preweighted 50 mL sterile conical centrifuge tube and centrifugation for 10 min at 5000  g. Remove the supernatant and weight the cell pellet. The cell pellet can be stored at 20 or 80  C if the purification cannot be done immediately. 2. Cell lysis can also be performed with a French press or by sonication on ice. In the latter case, 0.5 pulses are recommended for six periods of 30 s at 250 W spaced with a rest time of 1 min on ice in between to avoid overheating and protein denaturation. 3. Pack resin in empty chromatography column, and use an automated purification system to improve purification yield by allowing a longer contact time with the matrix (flow rate of 0.5–1 mL/min). 9 mg of rPVL can be purified for 1 mL of resin with automated system compared to only 3 mg when gravity is used. If gravity is used, reapply the sample three

Expression, Purification and Applications of the PVL Lectin

433

times to the column prior elution. Most GlcNAc affinity matrix tested was not recognized by rPVL apart from the GlcNAc Gel (EY, USA, ref. CG-003-5), but it has been unavailable for many months. The yield of purification was however much lower, and we usually purified only 1 mg of rPVL per mL of gel. Chitin matrix has been originally used for the native protein [7], but binding of rPVL proved to be very poor on that matrix that led to very big and dilute sample. The LNT2 matrix has been developed especially for the purification of rPVL. 4. rPVL is stable in PBS + CaCl2 for several weeks at 4  C but it is recommended to centrifuge the sample prior use. For longterm storage, dialyze the protein against ultrapure water prior freeze-drying and storage at 20  C. To avoid precipitation due to the brutal change from buffer to water, rPVL should be dialyzed first against a buffer containing 50–100 mM NaCl and 100 μM CaCl2 for 3 days and then against ultrapure water for another 3 days. CaCl2 seems to have a structural role that is why it is added to all buffers, but it is not implicated in the binding of ligands. It should not be removed immediately from the dialysis buffer. 5. Hemagglutination inhibition assays (HIA) with GlcNAcylated ligands can also be performed. 25 μl of twofold diluted ligand in 150 mM NaCl was incubated for 60 min at room temperature (RT) with 25 μL of rPVL at a concentration equal to two HAU prior addition of 50 μL of 4% RBC. Reading was performed after 1 h at room temperature, and the HA inhibitory titer is defined as the concentration of inhibitor necessary for complete inhibition of hemagglutination. 6. Use a low-density chip to avoid mass transport. A low-density chip is obtained by immobilizing a mixture of 10% recognized carbohydrate (β-D-GlcNAc-PAA-biotin) and 90% non-bound carbohydrate (α-L-Fuc-PAA-biotin) to 270 resonance unit (RU). A reference surface with non-bound carbohydrate is present in flow channel 1, thus allowing for subtraction of bulk effects and nonspecific interactions with unbound carbohydrate (same as previously, α-L-Fuc-PAA-biotin). This chip will permit to measure avidity (multivalent binding) of rPVL. The chip can also be prepared with other oligosaccharides such as sialylated one, but a weaker biding will be observed [14]. To get monovalent affinity, a rPVL chip should be made where 3500 RU of rPVL (100 μg/mL in 10 mM sodium acetate pH 6.2) is immobilized using standard procedure for amine coupling on flow channel 2 and where flow channel 1 has only been activated and deactivated. In this case, twofold cascade

434

Oriane Machon and Annabelle Varrot

dilutions of oligosaccharides are then injected on both channels. Similar procedure has been used for PAL and AANL [11, 12]. Affinity can also be measured in solution by isothermal microcalorimetry as described in [10]. 7. The use of polyacrylamide-based desalting gel is favored to classical Sephadex G-25 such as PD-10 based on dextran (GE Healthcare Life Sciences). It will limit weak recognition or unspecific interactions with the sugar-based matrix as observed for some lectins. 8. For coupling using primary amine, avoid buffers containing primary amines such as Tris, glycine, and ammonium salts. 9. The use of carbonate-bicarbonate buffer pH 9.4 is usually recommended as it improves the efficacy of HRP conjugation, but the reaction worked well in PBS for rPVL. 10. Resuspend the Alexa Fluor 488 in DMF just before use, and use immediately since the compound in not stable in solution. A ratio of 1 mg of compound for 10 mg of protein has been used. Other fluorophores such as Alexa Fluor 680 can also be used. 11. 500 ng/well of antibodies or glycosylated protein to be analyzed can also be immobilized overnight. Use a commercial antibody first such as human IgG1 kappa (Sigma-Aldrich) to do a dilution range of rPVL from 0 to 1000 ng/well, and determine which concentration of rPVL to use for further experiments. The concentration giving half signal is chosen since it gives enough signal to differentiate the quantity of terminal GlcNAc on glycosylated proteins without saturation (65 ng/well has often been used to perform ELLA on antibodies). 12. Semidry transfer can also be performed. During wet transfer, keeping the transfer buffer cold is essential and could be done by adding a cooling unit or running it in a cold room. Preparing the SDS-PAGE gel the day before usually leads to better transfer. The use of a prestained protein ladder such as the Dual Color Standards (Bio-Rad) could replace the Ponceau red staining step. The membrane can be blocked either by BSA or milk. Biotinylated-rPVL (starting concentration of 0.6 mg/ mL) followed by an incubation with HRP-avidin (dilution of 1:10,000) can also be used, but the labelling is less specific due to the recognition of other biotinylated proteins. A mouse monoclonal anti-O-GlcNAc (R2, VWR, Fontenay-sous-Bois) can be used according to published protocol, but labelling with HRP-rPVL is more efficient [14]. 13. rPVL labelled with Alexa Fluor 488 or other fluorophores can also be used for flow cytometry experiments on cancer cell lines where the glycosylation is or not inhibited as described in [10].

Expression, Purification and Applications of the PVL Lectin

435

14. Follow ethics statements on tissue sections. Required consent or certification should be obtained as well as their use approval by a Research Ethics Committee when necessary. 15. At this point, tissues sections can also be treated with 50 U of sialidase (New England Biolab) for 2 h at 37  C to limit noise from recognition of sialylated oligosaccharides. 25 U of ß-D-Nacetylhexosaminidases can also be used in similar conditions to confirm that the labelling is dependent of GlcNAc recognition. Add then fresh enzymes prior overnight incubation at 37  C. Slides treated with the enzyme buffer were used as control. Inhibition of binding can be done by adding 100 mM GlcNAc at the same time than the labelled protein. 16. HRP-conjugated avidin D (Vector Laboratories, Burlingame, CA) diluted in 1% BSA in PBS can be used with detection, thanks to the addition of 3-amino-ethyl-carbazole (AEC kit; Vector Laboratories, Burlingame, CA) to the slides.

Acknowledgments This work has been supported by CDP GLYCO@ALPS (ANR-15IDEX-02). Thanks to Jacques Le Pendu for permission to use the histochemistry figures. References 1. Reily C, Stewart TJ, Renfrow MB, Novak J (2019) Glycosylation in health and disease. Nat Rev Nephrol 15:346 2. Ishihara K, Kurihara M, Goso Y, Urata T, Ota H, Katsuyama T, Hotta K (1996) Peripheral alpha-linked N-acetylglucosamine on the carbohydrate moiety of mucin derived from mammalian gastric gland mucous cells: epitope recognized by a newly characterized monoclonal antibody. Biochem J 318(Pt 2):409–416 3. Pedersen JW, Blixt O, Bennett EP, Tarp MA, Dar I, Mandel U, Poulsen SS, Pedersen AE, Rasmussen S, Jess P, Clausen H, Wandall HH (2011) Seromic profiling of colorectal cancer patients with novel glycopeptide microarray. Int J Cancer 128(8):1860–1871 4. Torres CR, Hart GW (1984) Topography and polypeptide distribution of terminal N-acetylglucosamine residues on the surfaces of intact lymphocytes. Evidence for O-linked GlcNAc. J Biol Chem 259(5):3308–3317 5. Yang X, Qian K (2017) Protein O-GlcNAcylation: emerging mechanisms and functions. Nat Rev Mol Cell Biol 18 (7):452–465

6. Zachara N, Akimoto Y, Hart GW (2017) The O-GlcNAc modification. In: Varki A, Cummings RD, Esko JD, Stanley P, Hart GW, Aebi M, Darvill AG, Kinoshita T, Packer NH et al (eds) Essentials of glycobiology. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp 239–251 7. Kochibe N, Matta KL (1989) Purification and properties of an N-acetylglucosamine-specific lectin from Psathyrella velutina mushroom. J Biol Chem 264(1):173–177 8. Ueda H, Kojima K, Saitoh T, Ogawa H (1999) Interaction of a lectin from Psathyrella velutina mushroom with N-acetylneuraminic acid. FEBS Lett 448(1):75–80 9. Cioci G, Mitchell EP, Chazalet V, Debray H, Oscarson S, Lahmann M, Gautier C, Breton C, Perez S, Imberty A (2006) Beta-propeller crystal structure of Psathyrella velutina lectin: an integrin-like fungal protein interacting with monosaccharides and calcium. J Mol Biol 357 (5):1575–1591 10. Audfray A, Beldjoudi M, Breiman A, Hurbin A, Boos I, Unverzagt C, Bouras M, Lantuejoul S, Coll JL, Varrot A, Le Pendu J, Busser B,

436

Oriane Machon and Annabelle Varrot

Imberty A (2015) A recombinant fungal lectin for labeling truncated glycans on human cancer cells. PLoS One 10(6):e0128190 11. Ren XM, Li DF, Jiang S, Lan XQ, Hu Y, Sun H, Wang DC (2015) Structural basis of specific recognition of non-reducing terminal N-acetylglucosamine by an Agrocybe aegerita lectin. PLoS One 10(6):e0129608 12. Ribeiro JP, Ali Abol Hassan M, Rouf R, Tiralongo E, May TW, Day CJ, Imberty A, Tiralongo J, Varrot A (2017) Biophysical characterization and structural determination of the potent cytotoxic Psathyrella asperospora lectin. Proteins 85(5):969–975

13. Bonnardel F, Kumar A, Wimmerova M, Lahmann M, Perez S, Varrot A, Lisacek F, Imberty A (2019) Architecture and evolution of blade assembly in beta-propeller lectins. Structure 27(5):764–775 14. Machon O, Baldini SF, Ribeiro JP, Steenackers A, Varrot A, Lefebvre T, Imberty A (2017) Recombinant fungal lectin as a new tool to investigate O-GlcNAcylation processes. Glycobiology 27(2):123–128 15. Liu W, Han G, Yin Y, Jiang S, Yu G, Yang Q, Yu W, Ye X, Su Y, Yang Y, Hart GW, Sun H (2018) AANL (Agrocybe aegerita lectin 2) is a new facile tool to probe for O-GlcNAcylation. Glycobiology 28(6):363–373

Chapter 42 Yeast Flocculin: Methods for Quantitative Analysis of Flocculation in Yeast Cells Hiromi Maekawa and Kaoru Takegawa Abstract Flocculation, the clump forming property of yeast, has long been appreciated in breweries and utilized as an off-cost method to enable the reuse of yeast cells. Members of the flocculin protein family were identified as the adherent proteins on the cell surface responsible for flocculation, and their properties have been investigated. Crystal structures of the adhesion domain of flocculins revealed their unique mode of ligand binding where a calcium ion is located in the middle of the interface between flocculin and the interacting sugar. Here we describe the most commonly used flocculation assay. The method is simple and easy, yet it is the most direct and reliable assay to evaluate the flocculation cellular phenotype. Key words Yeast flocculin, Cell-cell adhesion, Calcium-dependent aggregation, Brewing production, GPI-anchored protein, FLO genes

1

Introduction Cell-cell adhesion is involved in a variety of biological processes including fungal, viral, and bacterial infections and organizing multicellular structures such as biofilms, neural differentiation, embryonic development, cellular communication, and tumor metastasis. One such phenomenon is yeast flocculation. Flocculation is defined as “the asexual, homotypic, reversible and calciumdependent aggregation of yeast cells to form flocs containing thousands of cells that rapidly sediment to the bottom of the liquid growth substrate” [1, 2]. It is distinct from sexual aggregation, co-flocculation, and chain formation: sexual aggregation is specific to the mating conditions and appears only in cells that are directed to the mating process [3], co-flocculation is heterotypic aggregation between flocculent and non-flocculent cells [4], while chain formation occurs when cells fail to separate from each other after mitotic divisions, and the resulting aggregates can contain up to 30–50 cells [4].

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_42, © Springer Science+Business Media, LLC, part of Springer Nature 2020

437

438

Hiromi Maekawa and Kaoru Takegawa

Because yeast flocculation promotes adherence into clumps and subsequently sediment rapidly from the medium, flocculation behavior has been important traditionally in brewing production as means to separate cells from broth at low cost [5]. It has been noted that flocculation properties vary significantly from strain to strain. Identification of a protein family of cell wall glycoproteins called adhesins, which confers adhesion properties, and their molecular analyses have shed light on the mechanisms behind flocculation [5]. Flocculation characteristics are determined by the properties of cell wall components, which in Saccharomyses cerevisiae consist of β-glucan, chitin, and a mannoprotein-containing fibrillar outer layer. Flocculation ability has a positive correlation with the hydrophobicity of the cell surface, but not with surface charge [6–9]. Cell surface hydrophobicity is affected by the adhesion molecules in the cell wall [10–12]. In S. cerevisiae, two adhesin protein families, collectively called flocculins, can directly promote flocculation. One consists of Flo1, Flo5, Flo9, and Flo10, which show high amino acid sequence similarity and recognize mannose residues on the cell surface to promote flocculation [2, 13–17] (Fig. 1). The second family is represented by Flo11 protein, which is also involved in different types of adherent behavior such as the biofilm formation and filamentation [17–19]. Flocculins are redundant; therefore the loss of one or more flocculin genes can, in general, be substituted by higher expression of other flocculin genes [13]. Composition and sequences of flocculin genes, as well as their expression levels, vary from strain to strain, and in many cases their differences correlate with the degree of flocculation and other morphological phenotypes [20].

S. pombe wild type -

OP-gsf2+

Fig. 1 Flocculation phenotype in S. pombe. Wild-type cells carrying Pnmt41-gsf2+ plasmid are grown in minimal medium either with thiamine (repressed, left) or without thiamine (expressed, right)

Analysis of Flocculation in Yeast Cells

439

Flocculins contain a C-terminal transmembrane domain and/or a GPI attachment site at the C-terminus. Flocculins extend their N-terminus toward the outside of the cell surface [21, 22]. The central domain consists of highly repetitive sequences that have high serine and threonine residue content (46% in Flo1). Extensive O-glycosylation in these sequences assists in the formation of a long rod-like structure [23]. At the N-terminus, following the signal sequence, resides the N-terminal domain containing the so-called PA14 domain in Flo1 and its homologues, which is responsible for the interactions with sugars. Instead of a PA14 domain, Flo11 has a specific domain that is found only within Ascomycota [2]. The structural model of the N-terminal domain of Flo1, predicted from the crystal structure of the anthrax toxin protective antigen, a founding member of PA14 domain proteins [24, 25], suggested that the tryptophan at position 228 is involved in the specific recognition of mannose. The calcium ion is directly involved in the interaction of the Flo1 N-terminal domain with sugars. Based on the crystal structure of N-Flo5 together with Ca2 + /mannose, the calcium ion is located between the sugar and protein to promote the recognition of the C2-hydroxyl group. The side chain of Q98, which is located outside of the PA14 domain in the Flo5 subdomain, also plays a significant role in the interaction with the sugar. In contrast to S. cerevisiae, the fission yeast Schizosaccharomyces pombe does not exhibit a flocculation phenotype in wild-type cells. However, a dominant, nonsexual, and calcium-dependent aggregation phenotype was observed in the certain mutant strains including gsf1 mutant. Flocculation is inhibited in the presence of galactose, but not mannose or glucose [26]. Gsf2 protein, which is the only flocculin identified in S. pombe (Fig. 1), has a similar domain structure to that of S. cerevisiae Flo1 homologues, namely, an N-terminal globular domain, a stalk-like domain consisting of repetitive sequences, and a C-terminal transmembrane domain and GPI-anchor sequence, although there is no similarity in amino acid sequence [27]. As gsf2+ transcription is the crucial step in the regulation of flocculation in S. pombe, the flocculation phenotypes correlate well with the gsf2 RNA levels in flocculant mutants such as gsf1Δ, lkh1Δ, and tup12Δ [27]. An in vitro flocculation assay is difficult because purification of functional full-length flocculins from yeast is problematic due to their biochemical properties such as glycosylation and embedding into membranes [28] (Fig. 2). Only the carbohydrate binding domain of a flocculin has been expressed and purified from Pichia pastoris to use in an in vitro binding assay for quantitative assessment [29]. In contrast, a simple cell clump formation assay has long been used to evaluate the flocculation ability of yeast cells as well as the strength and the specificity of sugars. Here we describe the experimental procedure of this classical, yet still the most commonly used in vivo assay, using the S. pombe flocculin, Gsf2, as an example.

440

Hiromi Maekawa and Kaoru Takegawa

Fig. 2 Structure of S. pombe flocculin. (a) Schematic representation of the Gsf2 protein on the cell surface. C-terminus of Gsf2 is embedded into the plasma membrane by the GPI-anchor, and the N-terminal domain extends outside of the cell. (b) Hydropathy plot of the amino acid sequence of Gsf2 protein. 1. Signal sequence; 2. repeated sequence A, the putative galactose interacting domain; 3. repeated sequence B, the Ser-Thr rich sequence; 4. GPI-anchor domain

2

Materials Prepare all solutions using ultrapure water (except for medium) and Wako (Osaka, Japan) special grade reagents or equivalents. Sterilize all solutions by autoclaving or by filter sterilization. Prepare and store all reagents at room temperature.

2.1

Stock Solutions

1. 0.5 M EDTA, pH 8.0: Weigh 93.1 g of disodium EDTA∙2H2O in a glass beaker, and add water to a volume of ~400 mL. Adjust pH to 8.0 with NaOH pellets or 5 N NaOH solution while stirring vigorously on a magnetic stirrer (approx. 10 g NaOH is required) (see Note 1). Transfer to a graduated cylinder and adjust volume to 500 mL with water. 2. 0.1 M citrate, pH 4.0: Dissolve 2.10 g citric acid monohydrate in 80 mL water in a glass beaker. Transfer to a graduated cylinder and make up to 100 mL with water (0.1 M citric acid). Dissolve 2.94 g trisodium citrate dihydrate in 80 mL water in a glass beaker. Transfer to a graduated cylinder and make up to 100 mL (0.1 M trisodium citrate). To make 0.1 M citrate buffer, pH 4.0, mix 59 mL of 0.1 M citric acid and 41 mL of 0.1 M trisodium citrate.

2.2 Flocculation Assay

1. Yeast extract with supplements (YES) medium: 0.5% w/v Bacto™ Yeast Extract, 3% w/v glucose, supplements (225 mg/L adenine, histidine, leucine, uracil, and lysine hydrochloride), and 2% w/v Bacto™ Agar (for solid media only). 2. Deflocculation buffer: 50 mM citrate, pH 4.0, 10 mM EDTA.

Analysis of Flocculation in Yeast Cells

441

3. Citrate buffer: 50 mM citrate, pH 4.0. 4. 1.0 M EDTA, pH 8.0. 5. Flocculation buffer: 50 mM EDTA, pH 8.0. 6. 5 M CaCl2: Dissolve 73.5 g of calcium chloride dihydrate in 80 mL water in a glass beaker. Transfer to a graduated cylinder and make up to 100 mL with water. 7. Spectrophotometer.

3

Method

3.1 Flocculation Assay

1. Inoculate freshly formed yeast colonies in 5 mL YES (or other appropriate media) and grow until stationary phase. 2. Inoculate cells in 5 mL of the same medium to an initial A600 of 0.05. 3. Grow cells for 24 h until stationary phase. 4. Harvest cells and wash twice with 0.1 M EDTA, pH 8.0. 5. Suspend cells in deflocculation buffer and vortex vigorously until the flocs, if any, turn to a homogeneous suspension by visual inspection. 6. Dilute the cell suspension to A600 of 10 in citrate buffer (see Note 2). 7. Transfer 0.3 mL of the resulting cell suspension into a 1.0-cm cuvette containing 2.7 mL of flocculation buffer (see Note 3). 8. Seal the top of cuvettes with a piece of parafilm. 9. Vortex vigorously for >10 s and immediately measure A600 (see Note 4) (Measurement A). 10. Add 60 μL of 5 M CaCl2 to the cuvette to make a final concentration of 100 mM. Mix well by inverting the cuvette several times. 11. Immediately measure A600 at 20 s intervals up to 120 s (Measurement B0 s, B20 s, B40 s, . . .) (see Note 5). 12. Calculate the flocculation ability according to the following formula (see Note 6): Flocculation ability ð%Þ ¼ fðA  B Þ=A g  100 A typical result is shown in Fig. 3. Flocculation activity is evaluated in wild-type, gsf1Δ, lkh1Δ, tup11Δ, and tup12Δ cells. Flocculation phenotype varies from no aggregation in wild-type and tup11Δ cells, through a significant but moderate level of aggregation in tup12Δ cells, to strong aggregation in gsf1Δ and lkh1Δ cells.

442

Hiromi Maekawa and Kaoru Takegawa

Flocculation activity (%)

100

gsf1Δ

80

lkh1Δ 60

tup12Δ

40 20

tup11Δ

0 0

20

40

60 80 Time (sec)

100

WT 120

Fig. 3 Flocculation phenotype of S. pombe strains. Cells grown in the complete YES medium were subjected to the flocculation assay

WT

gsf1Δ gsf1Δ gsf2Δ lkh1Δ tup11Δ tup12Δ

gsf2+

rRNA

Fig. 4 RNA levels of S. pombe strains. Total RNAs were prepared from cells grown in the complete YES medium. Northern analysis was performed using gsf2-specific probe

Northern blot analysis revealed that the strength of flocculation phenotype in each S. pombe strain has a clear correlation with the expression level of the gsf2+ gene (Fig. 4).

4

Notes 1. Add a small amount of NaOH pellets at a time and stir until all the pellets are dissolved completely. Repeat until most of the disodium EDTA is dissolved. Then adjust pH to 8.0 with 5 N NaOH solution. The disodium salt of EDTA will not go into solution until the pH is ~8.0. 2. A600 is measured using dilutions so that the spectrophotometer reading is 100 kDa) through disulfide bonds [2]. The activity of PCL-M can be inhibited by 2-mercaptoethanol, indicating that the multimer is active. PCL-M is GalNAc specific and its activity is inhibited by galactose; however, the galactose concentration required for complete inhibition is ten times higher than the

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_43, © Springer Science+Business Media, LLC, part of Springer Nature 2020

445

446

Suguru Oguri

Fig. 1 Fruiting body and mycelia of Pleurotus cornucopiae. (a) Fruiting body. (b) Mycelia grown in the liquid YMG medium. (c, d) Mycelia grown on solid (sawdust and rice bran) medium. Growth at 5 days (c) and 23 days (d)

concentration of GalNAc required for the same inhibition. In addition, PCL-M activity can be reversibly inhibited by EDTA; CaCl2 restores activity of EDTA-inhibited PCL-M [2]. Two cDNAs encoding isoforms of PCL-M have been isolated (DDBJ/EMBL/GenBank accession nos. AB115424 and AB116253). Both the genes encode deduced proteins of 373 amino acid residues, which share 97% identity between them [3]. The recombinant PCL-M produced by Pichia pastoris exhibits hemagglutination [4]. The N-terminus of mature PCL-M starts at 34th residue of the deduced sequence; however, SignalP 5.0 program (http://www.cbs.dtu.dk/services/SignalP/) has predicted that the first 18 residues represent the N-terminal signal sequence, suggesting that proteolytic removal of an N-terminal propeptide may occur in the recombinant protein (Fig. 2) [3]. The mature PCL-M polypeptide contains at least three potential N-glycosylation sites and six Cys residues. PCL-M consists of two homologous domains, the N- and C-terminal domains, which are arrayed in tandem and have 41% identity with each other (Fig. 2). It has been reported that both the domains share sequence

Pleurotus cornucoiae Mycelial Lectin (PCL-M)

Sp Pp

1 18

C

CC

CCC

34 50

213 214 N-terminal Domain

Sp

:Signal pepƟde and propepƟde

C :Cys residue

447

373 C-terminal Domain

Pp

:PropepƟde : N-glycosylaƟon site

:Start site of mature protein

Fig. 2 Domain construction of PCL-M. N-terminus of the mature peptide starts at residue 34. Positions of six Cys residues and potential N-glycosylation sites in the mature PCL-M are indicated. N-glycan structures are not exact

similarities with the rhamnoside-binding domain (SaCBM67) of Streptomyces avermitilis α-L-rhamnosidase, which belongs to the glycoside hydrolase family 78 [5]. SaCBM67 is a non-catalytic carbohydrate binding module of 160 amino acid residues that displays a β-jellyroll fold consisting of 11 β-strands. SaCBM67 binds to L-rhamnose in a calcium-dependent manner, with a single calcium ion forming coordinate bonds with the sugar [5]. The four calcium coordinating residues in SaCBM67 are all conserved in C-terminal domain in PCL-M (Fig. 3), but only two are conserved in the N-terminal domain of PCL-M. As shown in the crystal structure of this enzyme [5], these conserved residues might take part in the binding of sugar by PCL-M through formation of calcium coordinate bonds. The PCL-M gene comprises five exons and five introns, and Southern blot analysis indicated multiple copies of the PCL-M gene in the P. cornucopiae genome [3]. After cloning of PCL-M, some candidate PCL-M orthologs were found in Basidiomycetes belonging to the Agaricales order. Here, PCL-M was purified from the dikaryotic mycelial colony grown on solid medium prior to formation of fruit bodies. Based on the observation that rabbit blood cells adhered to the surface of the mycelia grown on solid medium, PCL-M is thought to localize on the cell surface of the mycelia [2] (Fig. 4d). This lectin is not contained in vegetatively growing mycelia in liquid medium (Fig. 4c). Concomitantly, a sequence similar to the motif for Aspergillus oryzae solid-state specific expression has been reported in the 50 upstream of the PCL-M gene [3]. Furthermore, monokaryotic

448

Suguru Oguri

PCL-M_N PCL-M_C SaCBM67

50 AFLNISKYIWTGENTVPEANNPVGTRAFRKNITSACGKCATCATIVVAADDSSTVYVNGV 214 -PLAQSKWIWTSANAATTA--PAASNAFRKTID-DCTKVAVCATVLISADNHYKLYVNGQ 133 PSLEGSSWIWFPEGEPANS-APAATRWFRRTVD--LPDDITGATLAISADNVYAVSVDGA

PCL-M_N PCL-M_C SaCBM67

110 AIG------SGAGWTT-GQVYFAP-LNPSSNLFAIAGVNN-VARAALMATINIHYSDGTH 270 AVG------SGDSFGR-AEAYSIPKLHPTLNTFAIDAKND-EGPAGVIATIHITYRDGTN 190 EVARTDLEADNEGWRRPAVIDVLDHVHSGNNTLAVSASNASVGPAGWICVLVLTTASGE-

* * * PCL-M_N PCL-M_C SaCBM67

*

161 ETFITDESWKTVRGAAPQGFQLPATSDSTWTFAMLQGFPQNSFWG-NPALPPVL 322 QTIVTDGSWKASQ-TVPNGFQETFFDDSDWVTATVVGNYGIAPWGSAVAIPPA249 KKIFSDASWKSTDHEPADGWREPDFDDSGWPAAKVAAAWGAGPWG---RVAP--

** Fig. 3 Comparison of the amino acid sequences of the N- and C-terminal domains of PCL-M and SaCBM67. PCL-M_N and PCL-M_C indicate the N- and C-terminal domains of PCL-M, respectively (BAD16579); SaCBM67, rhamnoside-binding domain of Streptomyces avermitilis α-L-rhamnosidase (Protein Data Bank ID: 3W5M_A). Sequences were aligned using CLUSTAL W 2.1. Identical amino acids are represented by whiteon-black letters. Arrowheads indicate the four residues that coordinate calcium ions in SaCBM67 [5]. Asterisks indicate the residues participating in L-rhamnose binding in SaCBM67 [5]

mycelia grown on solid medium did not produce PCL-M (Fig. 4a, b). Due to its stage-specific expression and localization, it may be that PCL-M participates in the process of fruiting body formation in P. cornucopiae. PCL-M may stimulate the formation of primordia in P. cornucopiae by adhering hyphae to each other, although the existence of such ligands on the P. cornucopiae cell wall remains unknown.

2

Materials

2.1

Organism

1. P. cornucopiae strain KC-2. This strain is available from ATCC (ATCC 201046) (see Notes 1 and 2) [2].

2.2

Culture Medium

1. Liquid yeast extract-malt extract-glucose (YMG) medium: 0.4% Bacto-Yeast extract, 1% Bacto-Malt extract, 0.4% glucose, pH 7.0. Add about 900 mL water to a glass beaker. Weigh 4 g Bacto-Yeast extract, 10 g Bacto-Malt extract, and 4 g glucose to the beaker. Mix and adjust the pH to 7.0 with 5 M NaOH. Adjust volume to 1 L with water. Sterilize by autoclaving. 2. YMG agar medium: 1.5% agar in YMG liquid medium. Add 15 g agar to 1 L of YMG liquid medium and sterilize by autoclaving. Pour into sterile plates. Use 25–30 mL medium per standard 90 mm petri dish. Allow the medium to solidify and store inverted at 4  C. 3. Sawdust and rice bran medium (solid medium): Weigh 1 L sawdust (see Note 3) and 150 mL rice bran in a container. Add 700 mL of water and mix well. Divide the medium into wide-mouthed culture flasks and cover with aluminum foil. Sterilize by autoclaving.

a

b

c

d

449

Dikaryotic

Monokaryotic

Pleurotus cornucoiae Mycelial Lectin (PCL-M)

Fig. 4 Observation of adhesion of red blood cells to mycelia mediated by PCL-M. Monokaryotic mycelia with a different mating type were obtained from P. cornucopiae KC-2 strain and then cultured on solid medium for 30 days (a, b). Dikaryotic mycelia generated by mating of (a) and (b) were cultured in liquid YMG medium for 7 days (c) or solid medium for 23 days (d). Mycelia were mixed with rabbit red blood cells according to the methodology described in Subheading 3.4. Adhesion of red blood cells is observed in only dikaryotic mycelia grown on solid medium (d). Samples were examined under a microscope at a 100 magnification. Inset: 400 magnification. Bars are equivalent to 0.1 mm 2.3 Hemagglutination Assay

1. Tris-buffered saline (TBS): 0.15 M NaCl and 10 mM Tris– HCl, pH 7.4. 2. Diluent solution: 1 mg/mL bovine serum albumin (BSA; see Note 4) and 50 mM CaCl2 in TBS. 3. Rabbit erythrocytes: 2% suspension of rabbit erythrocytes in TBS.

2.4

Purification

1. PSM-Sepharose 4B. Porcine stomach mucin (PSM) was conjugated to CNBr-activated Sepharose 4B (GE Healthcare UK Ltd., Amersham Place, England) according to the manufacturer’s instructions. Use 7 mL of 3 mg/mL PSM in coupling buffer (0.5 M NaCl, 0.1 M NaHCO3, pH 8.3) for 1 g of CNBr-activated Sepharose 4B.

450

Suguru Oguri

2. Phenylmethylsulfonyl fluoride (PMSF) solution: 100 mM PMSF in 2-propanol. Dissolve 17.4 mg PMSF in 1 mL of 2-propanol. Store at 20  C.

3

Methods

3.1 Culturing of the Mycelia

1. Mycelia are maintained on YMG agar slants (see Note 5). 2. For pre-culture, mycelia are cultured in 100 mL of liquid YMG medium in a 500 mL flask at 28  C for 7 days with shaking (160 rpm) (Fig. 1b). 3. Inoculate the pre-cultured mycelia (3 g wet weight) onto 300 mL of sawdust and rice bran medium. Incubate in the dark at 25  C for 30 days. At this time, mycelia should be spreading all over the medium, and a thick colony should be covering the surface (Fig. 1c) (see Note 6). 4. Harvest the mycelia growing on the surface of the solid medium by scraping with forceps (see Note 7).

3.2

Assay Procedure

1. Prepare the 2% suspension of rabbit erythrocytes in TBS. 2. Dilute the sample (20 μL) with diluent solution using a 96-well U-bottom plate, and mix it with an equal volume of the 2% suspension of rabbit erythrocytes. Hemagglutination units (titer) are calculated as the reciprocal of the dilution multiple that yields a positive reaction after 1 h at room temperature.

3.3 Purification of PCL-M

1. Weigh 100 g of frozen mycelia into a blender. 2. Add 200 mL of TBS to the mycelia in a blender. 3. Add 2 mL of 100 mM PMSF to the blender to achieve a final concentration of 1 mM just before homogenizing. 4. Homogenize the mycelia in a blender for 10 min. 5. Centrifuge the homogenate at 10,000  g for 20 min at 4  C. 6. Add 150 mL of TBS to the precipitate and homogenize again using the blender. 7. Centrifuge the homogenate at 10,000  g for 20 min at 4  C. 8. Combine the extracts. 9. Slowly add solid ammonium sulfate to the supernatant while stirring gently to make an 80% saturated solution. Incubate at 4  C overnight. 10. Collect the proteins by centrifugation at 10,000  g for 20 min at 4  C. Decant the supernatant and resuspend the pellet in 30 mL of TBS. Don’t dialyze this solution (see Note 8). 11. Load the resuspended pellet onto the PSM-Sepharose 4B column (0.6  3.5 cm) equilibrated with TBS.

Pleurotus cornucoiae Mycelial Lectin (PCL-M)

451

12. Wash the column with TBS until the A280 is below 0.05. 13. Wash the column with TBS containing 1 M NaCl. 14. Elute PCL-M with 5 mM EDTA in TBS (see Note 9). Collect the column effluent in tubes (1 mL/tube). 15. Add 500 mM CaCl2 in TBS to the fractions to achieve a final concentration of 50 mM in order to restore hemagglutinating activity. Stand these fractions at 4  C overnight. 16. Measure hemagglutination in each fraction. 17. Yield active fractions and check the purity by 12.5% SDS-PAGE (see Notes 10 and 11). 3.4 Observation of Adhesion of Red Blood Cells to Mycelia Mediated by PCL-M

1. Culture P. cornucopiae mycelia using sawdust and rice bran medium or liquid YMG medium. 2. Collect mycelia using forceps and transfer it to a glass slide. 3. Mix with 1% suspension of rabbit erythrocytes. 4. Following a 5-min incubation, examine the samples under a microscope (Fig. 4). 5. Erythrocytes form aggregates on the surface of the dikaryotic mycelia grown on the solid medium (Fig. 4d) (see Note 12).

4

Notes 1. Strain KC-2 is PCL-F deficient. This strain was used for purification of PCL-M [2]. 2. We tested the hemagglutination activities of the solid-grown mycelia from ten strains of P. cornucopiae. Hemagglutinating activities were detected in all tested mycelia from these strains. 3. We used conifer sawdust exposed to air for over a half year. Sawdust from a broad-leaved tree should be suitable for the culture medium. 4. The addition of BSA prevents loss of activity in the diluted sample. 5. Long-term maintenance of mycelium on YMG agar results in loss of the ability of fruiting body formation. To avoid this problem, regeneration of mycelium from fruiting bodies is effective. 6. PCL-M was detected in mycelia after 16 days post-inoculation, when mycelia had spread all over the medium (see ref. 2). 7. After harvesting the mycelia, the remaining culture may be used for fruiting body formation, as per the following procedure. Pour ice-cold distilled water into the culture flask and incubate for 1 h. Discard the water and place the culture flask into a humid chamber in the dark at 20  C. Fruiting bodies will form after several days.

452

Suguru Oguri

8. PCL-M adheres to the materials for purification. Desalting with the cellulose membrane and Sephadex gel results in reduction in the yield of protein. Avoid using a reverse-phase column (both silica base and polymer base) for purification. Among tested materials, Toyopearl gel (Tosoh Corp., Tokyo, Japan) was suitable for the purification of PCL-M. Using a high-salt buffer (e.g., 0.5 M NaCl) may improve recovery of PCL-M. 9. 0.1 M acetic acid containing 1 M NaCl can be used as an alternate elution buffer. Addition of 1 M NaCl to the buffer is essential. PCL-M was not eluted from the column using a buffer containing 0.1 M acetic acid alone. Neutralize the eluate by addition of 1 M Tris. 10. From 100 g of mycelia, 0.75 mg of purified PCL-M was obtained [2]. 11. The purified PCL-M solution contains salt. We used the Toyopearl HW-40 column (Tosoh Corp., Japan) equilibrated with 10% (v/v) acetic acid for desalting of the purified PCL-M solution. 12. Adhesion of erythrocytes to the mycelia is inhibited by the addition of asialo-bovine submaxillary gland mucin, EDTA, or anti-PCL-M serum, the inhibitors of hemagglutination caused by PCL-M. References 1. Yoshida M, Kato S, Oguri S, Nagata Y (1994) Purification and properties of lectins from a mushroom, Pleurotus cornucopiae. Biosci Biotechnol Biochem 58(3):498–501 2. Oguri S, Ando A, Nagata Y (1996) A novel developmental stage-specific lectin of the basidiomycete Pleurotus cornucopiae. J Bacteriol 178 (19):5692–5698 3. Sumisa F, Ichijo N, Yamaguchi H, Nakatsumi H, Ando A, Iijima N, Oguri S, Uehara K, Nagata Y (2004) Molecular properties of mycelial aggregate-specific lectin of Pleurotus cornucopiae. J Biosci Bioeng 98(4):257–262

4. Sumisa F, Iijima N, Ando A, Shiga M, Kondo K, Amano K, Nagata Y (2004) Properties of mycelial aggregate-specific lectin of Pleurotus cornucopiae produced in Pichia pastoris. Biosci Biotechnol Biochem 68(4):959–960. https:// doi.org/10.1271/bbb.68.959 5. Fujimoto Z, Jackson A, Michikawa M, Maehara T, Momma M, Henrissat B, Gilbert HJ, Kaneko S (2013) The structure of a Streptomyces avermitilis alpha-L-rhamnosidase reveals a novel carbohydrate-binding module CBM67 within the six-domain arrangement. J Biol Chem 288(17):12376–12385. https://doi. org/10.1074/jbc.M113.460097

Chapter 44 Expression and Purification of a Human Pluripotent Stem Cell-Specific Lectin Probe, rBC2LCN Hiroaki Tateno Abstract rBC2LCN is a recombinant N-terminal domain of BC2L-C lectin that is derived from the gram-negative bacteria Burkholderia cenocepacia and specifically binds to Fucα1-2Galβ1-3GlcNAc/GalNAc. Glycome analysis using a high-density lectin microarray revealed that rBC2LCN specifically binds to human pluripotent stem cells (hPSCs) but not to non-hPSCs. The lectin can be added to the cell culture medium for the live staining of hPSCs without causing visible cytotoxicity. Moreover, it can be used in flow cytometric analysis and for the staining of fixed hPSCs. rBC2LCN is a single-chain molecule with a low molecular weight (trimer of 16 kDa), which can be produced in large quantities in Escherichia coli (0.1 g/L). Therefore, rBC2LCN may be a cost-effective probe for use in hPSCs, unlike other hPSC surface marker antibodies that require a mammalian cell expression system for production. In this study, we describe the protocols for the expression and purification of rBC2LCN from E. coli and live staining of hPSCs using this probe. Key words Human pluripotent stem cells, Purification, Expression, Escherichia coli, Lectin, rBC2LCN

1

Introduction Human pluripotent stem cells (hPSCs) such as human embryonic stem cells (hESCs) [1] and human induced pluripotent stem cells (hiPSCs) [2] have immense potential as the sources of cell-based therapies owing to their unlimited self-renewal and pluripotent differentiation capacities. Particularly, hiPSCs are highly promising not only for use in regenerative medicine but also for use in disease modeling and drug development as they can be generated from various adult somatic cells simply via the introduction of reprogramming factors. Enormous efforts have been undertaken to establish hPSC-based therapies for various degenerative diseases [3]. In 2014, the first-in-human clinical trial using hiPSC-derived retinal pigment epithelium was conducted by the RIKEN Center for Developmental Biology, Kobe, for treating wet age-related

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_44, © Springer Science+Business Media, LLC, part of Springer Nature 2020

453

454

Hiroaki Tateno

macular degeneration [4, 5]. However, although the clinical and industrial applications of hPSC-based cell therapies are becoming an increasingly realistic prospect, a major safety concern still exists: the formation of tumors by residual hPSCs in differentiated cell populations [6, 7]. Over the past years, a number of animal studies have highlighted the tumorigenic risk associated with hPSCs [8– 12]. Reportedly, as few as 100 hPSCs are sufficient for teratoma formation [8, 13]. Therefore, the complete removal of hPSCs from final cell products without compromising their viability, safety, efficacy, and functional properties is a prerequisite for the successful clinical application of hPSC-based cell therapies. The removal of residual hPSCs from hPSC-derived cells is also important for establishing disease models. To develop a novel probe that can be used for the identification and elimination of hPSCs, we previously performed comprehensive glycome analysis of a large set of hiPSCs and hESCs using a high-density lectin microarray [14]. We found that a recombinant N-terminal domain of lectin BC2L-C (rBC2LCN) isolated from Burkholderia cenocepacia demonstrated exclusive binding to all the tested undifferentiated hPSCs but not to the differentiated somatic cells [14, 15]. rBC2LCN is the trimeric form of a 16-kDa domain that comprises 11 β strands and a short αhelix [15]. We previously demonstrated that fluorescencelabeled rBC2LCN can effectively stain fixed hPSCs [16] and that live staining of hPSCs could be performed by simply adding rBC2LCN into the culture medium with little or no toxicity even at a high concentration (approximately 100 μg/mL) [16]. Podocalyxin, a hyperglycosylated type 1 transmembrane protein that is known to be highly expressed on hPSCs, was reported to be a predominant cell surface ligand of rBC2LCN [17]. rBC2LCN exhibits significant affinity to a mucin-type O-glycan containing an H type 3 epitope isolated from human 201B7 hiPSCs, indicating that H type 3 (Fucα1-2Galβ1-3GalNAc) is a novel hPSC surface marker [17, 18]. rBC2LCN is a small, single-chain protein that can be easily purified to homogeneity from the soluble fractions of Escherichia coli (>0.1 g/L) using single-step, sugar-immobilized affinity chromatography. However, hPSC marker antibodies are large and complex proteins (>150 kDa) that are produced using mammalian cells. For example, SSEA3 and TRA-1-60/81 are mouse/rat IgMs with a molecular weight of 900 kDa. Mouse IgG is another hPSC marker antibody with a molecular weight of 150–170 kDa (Table 1). rBC2LCN has a high potential for use as a novel hPSC marker. In this manuscript, we describe the protocols for the expression and purification of rBC2LCN from E. coli and live staining of hPSCs using rBC2LCN.

Expression and Purification of rBC2LCN

455

Table 1 Surface markers of hPSCs

Marker

Epitope structure

Detection probe

MW (kDa)

SSEA3

GalNAcβ1-3Gala1-4Galβ1- (Gb5)

Rat IgM (MC631)

900

SSEA4

Siaα2-3Galβ1-3GalNAcb1(sialylGb5)

Mouse IgG3 (MC813–70)

170

SSEA5

Fucα1-2Galβ1-3GlcNAc

Mouse IgG1 (8E11)

150

TRA-1-60

Galβ1-3GlcNAcβ1-3Galβ1-4GlcNAc

Mouse IgM (REA157)

900

TRA-1-81

Galβ1-3GlcNAcβ1-3Galβ1-4GlcNAc

Mouse IgM (TRA-1-81) 900

Low-sulfated keratan sulfate

Low-sulfated keratan sulfate

Mouse IgG1 (R10-G)

150

LNFP I

Fucα1-2Galβ1-3GlcNAcβ1-3Galβ14Glc

Mouse IgG1 (R17-F)

150

H type 3

Fucα1-2Galβ1-3GlcNAc/GalNAc

Lectin (rBC2LCN)

2 2.1

48

Materials Reagents

1. E. coli BL21 CodonPlus (DE3)-RIL. 2. Kanamycin. 3. Luria-Bertani (LB) broth. 4. LB agar medium: LB broth containing 1% agar powder (Nacalai Tesque, Kyoto, Japan; Cat. No. 01162-15) and 10 μg/mL Kanamycin. 5. SOC broth: 2% tryptone (Nacalai Tesque, Kyoto, Japan; Cat. No. 35640-95) containing 0.5% dried yeast extracts (Nacalai Tesque, Kyoto, Japan; Cat. No. 15838-45) and 10 mM NaCl (FUJIFILM Wako Chemical Co., Osaka, Japan; Cat. No. 191-01665). 6. IPTG. 7. Plasmid (rBC2LCN-pET27b). 8. PBS: 6-mM Na2HPO4·12H2O (FUJIFILM Wako Chemical Co., Osaka, Japan; Cat. No. 196-02835), 1.4-mM KH2PO4 (FUJIFILM Wako Chemical Co., Osaka, Japan; Cat. No. 169-04245), 140-mM NaCl (FUJIFILM Wako Chemical Co., Osaka, Japan; Cat. No. 191-01665), and 2.7-mM KCl (FUJIFILM Wako Chemical Co., Osaka, Japan; Cat. No. 163-03545), pH 7.0. 9. PBSE: PBS containing 1-mM EDTA.

456

Hiroaki Tateno

10. PBSET: PBSE containing 0.1% Triton X-100 (Nacalai Tesque, Kyoto, Japan; Cat. No. 35501-15). 11. Protease inhibitor cocktail. 12. Coomassie staining reagent. 13. NaOH. 14. CH3COONa∙3H2O. 15. Tris-HCl. 16. XV Pantera MP Gel. 17. L-Fucose. 18. Sepharose CL-4B. 2.2

Instruments

1. Bioshaker. 2. Centrifuge. 3. UV-vis spectrophotometer. 4. Autoclave. 5. Sonicator. 6. Large centrifuge tubes. 7. Small centrifuge tubes. 8. 3-L culture flask. 9. Thermo Minder SD mini. 10. Heat block. 11. Disposable Poly-Prep Chromatography Column. 12. Dialysis membrane, size 36. 13. Reagent reservoir. 14. Cy3 Mono-Reactive Dye.

3

Methods

3.1 Construction of rBC2LCN-pET27b Plasmid

1. rBC2LCN (1–156 amino acids) from B. cenocepacia was inserted between the NheI and XhoI restriction sites of the backbone pET27b expression vector (Fig. 1).

3.2 Expression of rBC2LCN-pET27b Plasmid in E. coli

1. Transform rBC2LCN-pET27b into the competent cells of E. coli BL21 CodonPlus (DE3)-RIL via heat shock followed by inoculation on LB agar plates containing 10 μg/mL of kanamycin, and then culture at 37  C overnight. 2. Preculture the rBC2LCN-pET27b-transformed E. coli in 5 mL of LB broth containing 10 μg/mL of kanamycin using a shaking incubator (140–160 rpm) at 37  C overnight.

Expression and Purification of rBC2LCN

457

Km

NdeI rBC2LCN-pET27b

rBC2LCN XhoI

rBC2LCN (156 aa): MPLLSASIVSAPVVTSETYVDIPG LYLDVAKAGIRDGKLQVILNVPTP YATGNNFPGIYFAIATNQGVVAD GCFTYSSKVPESTGRMPFTLVAT IDVGSGVTFVKGQWKSVRGSAM HIDSYASLSAIWGTAAPSSQGSG NQGAETGGTGAGNIGGG

Fig. 1 Schematic representation of rBC2LCN-pET27b

3. Add 5 mL of the preculture into 1 L of LB broth containing 10 μg/mL of kanamycin, and culture using a shaking incubator (140–160 rpm) at 37  C until cell growth reaches an OD600 of 0.4. 4. Culture the cell culture flask at 20  C and for additional 30 min. 5. Add IPTG (1 mM) into the culture medium, and culture the cells in a shaking incubator (140–160 rpm) at 20  C for 24 h (see Notes 1 and 2). 6. Harvest E. coli by centrifugation at 4330  g for 30 min (see Note 3). 7. Suspend the harvested E. coli pellet in 12 mL of PBSET containing a protease inhibitor cocktail followed by sonication on ice (see Note 4). 8. Pellet insoluble fractions by centrifugation at 24,910  g for 30 min. 9. Recover soluble fractions for further purification. 3.3 Preparation of L-Fucose-Sepharose for Affinity Purification of rBC2LCN

1. Activate 20 g of Sepharose CL-4B beads using 5% of epichlorohydrin in 0.5-M NaOH at 40  C for 2 h. 2. Wash activated Sepharose beads with 500 mL of water. 3. Dissolve 33 g of L-fucose in 80 mL of 1-M NaOH (pH 13), and incubate with the activated Sepharose beads at 40  C for 24 h. 4. Wash Sepharose beads with water, and suspend in 50 mL of 1-M monoethanolamine (pH 8.0) followed by incubation at 37  C overnight. 5. Finally, wash Sepharose beads with 0.1-M acetate buffer (pH 4.0)/0.5-M NaCl and 0.1-M Tris (pH 8.0)/0.5-M NaCl.

Elution 2

Elution 1

Wash 3

Wash 2

kDa

Wash 1

Hiroaki Tateno

Passed-through

458

140 100 70 50 40 30

20

15

Fig. 2 Results of SDS-PAGE analysis performed on each fraction obtained from the purification step 3.4 Affinity Purification

1. For purification, pack 2–2.5 mL of L-fucose-Sepharose into a Bio-Rad disposable column, and equilibrate using 8 mL of PBSE. Prepare four columns for the purification of rBC2LCN from a 1-L rBC2LCN-pET27b-transformed E. coli culture. 2. Apply E. coli soluble fractions onto these four columns containing L-fucose-Sepharose (see Note 5). 3. Wash the columns three times with 4 mL of PBSE. 4. Elute rBC2LCN with 4 mL of 0.2-M L-fucose in PBSE three times.

3.5 Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis (SDS-PAGE) of Eluted Fractions

1. Each fraction obtained from the purification steps was analyzed using SDS-PAGE (Fig. 2). The 16-kDa band corresponds to the purified rBC2LCN.

3.6 Dialysis of Eluted Fractions

1. Dialyze eluted fractions using the 10-kDa cut dialysis membrane against 5 L of 1/10 PBS at 4  C. After 3 and 12 h, exchange external solutions with new ones. After 18 h, recover internal solutions.

Expression and Purification of rBC2LCN

459

100 wm

Fig. 3 Live staining of hiPSC 201B7 using 0.1 μg/mL of Cy3-labeled rBC2LCN 3.7 Protein Quantification and Preservation

1. Quantify proteins using the BCA assay. From a 1-L culture, 80–100 mg of rBC2LCN is generally recovered.

3.8 Cy3 Labeling of rBC2LCN

1. Briefly, incubate 10 μL of 1-mg/mL rBC2LCN in PBS with Cy3-NHS (for 10 μg of labeled protein equivalent) at room temperature in the dark for 1 h. 2. Then add 90 μL of TBS. 3. Apply the solution through a miniature column containing 100 μL of Sephadex G-25 fine to remove unreacted Cy3. 4. Store Cy3-labeled rBC2LCN in the dark at 4  C or 20  C (see Note 6).

3.9

Live Staining

1. Add Cy3-labeled rBC2LCN (0.1–1 μg/mL) to the cell culture medium, and incubate for 1–2 h in a CO2 incubator. 2. Observe using fluorescence microscope at the end of the incubation period (Fig. 3).

4

Notes 1. IPTG has to be added at OD600 between 0.4 and 0.6. Induction by IPTG at higher than OD600 of 0.8 would decrease the induced protein yield. 2. The induction of protein expression by IPTG was performed at 20  C to increase the soluble and active forms of lectins. 3. The recovered E. coli can be stored at 80  C until purification.

460

Hiroaki Tateno

4. EDTA and protease inhibitor cocktail were added into PBSET for use to lyse E. coli to inhibit protease degradation. 5. The E. coli extract should be filtrated by 0.45 μm of PVDF membrane before the affinity purification. 6. The concentration of Cy3-labeled rBC2LCN was calculated according to the following equation: ½Cy3‐labeled rBC2LCN ¼ ½A280  ð0:08  A552Þ=13, 688 ð1Þ Molar extinction of rBC2LCN: 13688. Correction factor of Cy3: 0.08.

Acknowledgments We thank the RIKEN BioResource Center for providing the hiPS cell line 201B7 (HPS0063). References 1. Thomson JA, Itskovitz-Eldor J, Shapiro SS, Waknitz MA, Swiergiel JJ, Marshall VS, Jones JM (1998) Embryonic stem cell lines derived from human blastocysts. Science 282 (5391):1145–1147 2. Takahashi K, Tanabe K, Ohnuki M, Narita M, Ichisaka T, Tomoda K, Yamanaka S (2007) Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell 131 (5):861–872. https://doi.org/10.1016/j.cell. 2007.11.019. S0092-8674(07)01471-7 [pii] 3. Garber K (2013) Inducing translation. Nat Biotechnol 31(6):483–486. https://doi.org/ 10.1038/nbt.2602 4. Kamao H, Mandai M, Okamoto S, Sakai N, Suga A, Sugita S, Kiryu J, Takahashi M (2014) Characterization of human induced pluripotent stem cell-derived retinal pigment epithelium cell sheets aiming for clinical application. Stem Cell Rep 2(2):205–218. https:// doi.org/10.1016/j.stemcr.2013.12.007 5. Mandai M, Watanabe A, Kurimoto Y, Hirami Y, Morinaga C, Daimon T, Fujihara M, Akimaru H, Sakai N, Shibata Y, Terada M, Nomiya Y, Tanishima S, Nakamura M, Kamao H, Sugita S, Onishi A, Ito T, Fujita K, Kawamata S, Go MJ, Shinohara C, Hata KI, Sawada M, Yamamoto M, Ohta S, Ohara Y, Yoshida K, Kuwahara J, Kitano Y, Amano N, Umekage M, Kitaoka F, Tanaka A, Okada C, Takasu N, Ogawa S, Yamanaka S, Takahashi M

(2017) Autologous induced stem-cell-derived retinal cells for macular degeneration. N Engl J Med 376(11):1038–1046. https://doi.org/ 10.1056/NEJMoa1608368 6. Lee AS, Tang C, Rao MS, Weissman IL, Wu JC (2013) Tumorigenicity as a clinical hurdle for pluripotent stem cell therapies. Nat Med 19 (8):998–1004. https://doi.org/10.1038/nm. 3267 7. Ben-David U, Nudel N, Benvenisty N (2013) Immunologic and chemical targeting of the tight-junction protein Claudin-6 eliminates tumorigenic human pluripotent stem cells. Nat Commun 4:1992. https://doi.org/10. 1038/ncomms2992. ncomms2992 [pii] 8. Hentze H, Soong PL, Wang ST, Phillips BW, Putti TC, Dunn NR (2009) Teratoma formation by human embryonic stem cells: evaluation of essential parameters for future safety studies. Stem Cell Res 2(3):198–210. https:// doi.org/10.1016/j.scr.2009.02.002. S18735061(09)00018-X [pii] 9. Kawai H, Yamashita T, Ohta Y, Deguchi K, Nagotani S, Zhang X, Ikeda Y, Matsuura T, Abe K (2010) Tridermal tumorigenesis of induced pluripotent stem cells transplanted in ischemic brain. J Cereb Blood Flow Metab 30 (8):1487–1493. https://doi.org/10.1038/ jcbfm.2010.32 10. Lee AS, Tang C, Cao F, Xie X, van der Bogt K, Hwang A, Connolly AJ, Robbins RC, Wu JC (2009) Effects of cell number on teratoma

Expression and Purification of rBC2LCN formation by human embryonic stem cells. Cell Cycle 8(16):2608–2612 11. Roy NS, Cleren C, Singh SK, Yang L, Beal MF, Goldman SA (2006) Functional engraftment of human ES cell-derived dopaminergic neurons enriched by coculture with telomeraseimmortalized midbrain astrocytes. Nat Med 12(11):1259–1268. https://doi.org/10. 1038/nm1495 12. Yamashita T, Kawai H, Tian F, Ohta Y, Abe K (2011) Tumorigenic development of induced pluripotent stem cells in ischemic mouse brain. Cell Transplant 20(6):883–891. https://doi. org/10.3727/096368910X539092 13. Gropp M, Shilo V, Vainer G, Gov M, Gil Y, Khaner H, Matzrafi L, Idelson M, Kopolovic J, Zak NB, Reubinoff BE (2012) Standardization of the teratoma assay for analysis of pluripotency of human ES cells and biosafety of their differentiated progeny. PLoS One 7(9): e45532. https://doi.org/10.1371/journal. pone.0045532 14. Tateno H, Toyota M, Saito S, Onuma Y, Ito Y, Hiemori K, Fukumura M, Matsushima A, Nakanishi M, Ohnuma K, Akutsu H, Umezawa A, Horimoto K, Hirabayashi J, Asashima M (2011) Glycome diagnosis of human induced pluripotent stem cells using lectin microarray. J Biol Chem 286 (23):20345–20353. https://doi.org/10. 1074/jbc.M111.231274. M111.231274 [pii]

461

15. Sulak O, Cioci G, Delia M, Lahmann M, Varrot A, Imberty A, Wimmerova M (2010) A TNF-like trimeric lectin domain from Burkholderia cenocepacia with specificity for fucosylated human histo-blood group antigens. Structure 18(1):59–72. https://doi.org/10. 1016/j.str.2009.10.021. S0969-2126(09) 00470-5 [pii] 16. Onuma Y, Tateno H, Hirabayashi J, Ito Y, Asashima M (2013) rBC2LCN, a new probe for live cell imaging of human pluripotent stem cells. Biochem Biophys Res Commun 431 (3):524–529. https://doi.org/10.1016/j. bbrc.2013.01.025. S0006-291X(13)000612 [pii] 17. Tateno H, Matsushima A, Hiemori K, Onuma Y, Ito Y, Hasehira K, Nishimura K, Ohtaka M, Takayasu S, Nakanishi M, Ikehara Y, Ohnuma K, Chan T, Toyoda M, Akutsu H, Umezawa A, Asashima M, Hirabayashi J (2013) Podocalyxin is a glycoprotein ligand of the human pluripotent stem cellspecific probe rBC2LCN. Stem Cells Transl Med 2(4):265–273. https://doi.org/10. 5966/sctm.2012-0154. sctm.2012-0154 [pii] 18. Hasehira K, Tateno H, Onuma Y, Ito Y, Asashima M, Hirabayashi J (2012) Structural and quantitative evidence for dynamic glycome shift on production of induced pluripotent stem cells. Mol Cell Proteomics 11 (12):1913–1923. https://doi.org/10.1074/ mcp.M112.020586. M112.020586 [pii]

Chapter 45 Preparation of Fluorescent Recombinant Shiga Toxin B Subunit and Its Application to Flow Cytometry Toshiyuki Yamaji Abstract Shiga toxin (Stx) is a major virulence factor of enterohemorrhagic Escherichia coli (E. coli). Stx consists of one enzymatic A subunit and five B subunits (StxB) that are involved in binding. The StxB pentamer specifically recognizes a glycosphingolipid, globotriaosylceramide (Gb3), as a receptor; therefore, it can be used as a probe to detect Gb3. This chapter describes the preparation of recombinant Stx1B proteins using E. coli, their conjugation with fluorescent dyes, and their application for flow cytometry. The prepared fluorescent StxB proteins bound to cells of several lines, including the HeLa human cervix adenocarcinoma cell line and the THP-1 human monocytic leukemia cell line. Furthermore, the probe was useful for confirmation of several sphingolipid-deficient HeLa cell lines that were constructed using genome editing. Key words Shiga toxin, Gb3, Glycosphingolipid, Affinity purification, Fluorescent Stx1B, Flow cytometry

1

Introduction Shiga toxin (Stx) is a major virulence factor found in enterohemorrhagic Escherichia coli, as well as Shigella dysenteriae. These bacteria can cause bloody diarrhea, hemorrhagic colitis, and sometimes lifethreatening hemolytic-uremic syndrome and encephalopathy [1– 4]. Stx is classified as an AB5 toxin, which consists of one A subunit that has enzymatic activity and five B subunits involved in cellsurface binding. The A subunit of Stx inhibits protein synthesis by removing an adenine residue from the 28S RNA of 60S ribosomal subunits [5]. The five B subunits of Stx form a pentamer that binds to a specific glycosphingolipid (GSL), globotriaosylceramide (Gb3, Galα1,4Galβ1,4Glcβ1-ceramide) (Fig. 1) on the cell surface, and transports the toxin into the endoplasmic reticulum. Shigatoxigenic E. coli (STEC) strains (e.g., serotype O157:H7) can produce two types of Stx, designated Stx1 and Stx2. Both types of Stx can bind to terminal Galα1,4Gal-containing GSLs, including Gb3 and galabiosylceramide (Gb2, Galα1,4Galβ1-ceramide) [6–

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_45, © Springer Science+Business Media, LLC, part of Springer Nature 2020

463

464

Toshiyuki Yamaji

A

B

X O

HO

H N

O

Serine+Palmitoyl CoA

Glucose

Sialic acid

3-ketodihydrosphingosine

Galactose

Choline

Ac

Dihydrosphingosine

N-acetyl glucosamine

Dihydroceramide

Ac

N-acetyl galactosamine

Ceramide

SM

Phosphate Sulfate

Sphingosine1-phosphate

Sphingosine

Glycosphingolipids GalCer

GlcCer

ganglio series

LacCer Sulfatide Gb2

X=H : Ceramide

GM3

GD3

GM4 Ac

Ac

Ac

Ac

X=phospho choline : SM

Stx

Ac

Gb4 lactoseries

GD2

Ac

X=sugars : GSLs neolactoseries

GM2

GA2

Gb3

Lc3

Ac

Ac

Ac

GD1b Ctx

Ac

Forssman antigen globo series

Ac

GM1

GA1

Ac

GD1a

GT1b

asialoganglio series

Fig. 1 Structure and metabolism of glycosphingolipids. (a) Sphingolipid structure. Ceramide-containing C16:0 fatty acid is shown. (b) Glycosphingolipid biosynthesis. Cer ceramide, SM sphingomyelin, GlcCer glucosylceramide, LacCer lactosylceramide, Gb3 globotriaosylceramide, Gb4 globoside, Lc3 lactotriaosylceramide, GalCer galactosylceramide, Gb2 galabiosylceramide, GA2 asialoGM2, GA1 asialoGM1, GM4 sialyl GalCer. Binding of Stx to Gb3 as well as Gb2 and Ctx to GM1 is shown. This figure is a partially modified version of a previously published figure [21]

8]. A variant of Stx2 can also bind to other globo-series lipids, including globotetrasylceramide (Gb4, GalNAcβ1,3Galα1,4Galβ1,4Glcβ1ceramide) [9]. Mice in which the Gb3 synthase gene, also known as A4GalT, has been knocked out show complete resistance to both Stx1 and Stx2 [10]. Each B subunit monomer has 3 trisaccharide-binding pockets, so a total of 15 trisaccharides can bind to the B pentamer [11]. In addition to Stx, other AB5 toxins include cholera toxin (Ctx), produced by Vibrio cholerae; subtilase cytotoxin (SubAB), produced by locus of enterocyte effacement (LEE)-negative STEC; and ricin, produced in the seeds of Ricinus communis (the castor bean). These toxins also recognize glycans via their B pentamer. Ctx binds to a GSL, GM1 (Galβ1,3GalNAcβ1,4(Saα2,3)Galβ1,4Glcβ1ceramide) (Fig. 1) [12], whereas SubAB and ricin mainly bind to terminal sialic acid-containing and terminal galactose-containing

Shiga Toxin B-Subunit and Flow Cytometry

465

glycans on proteins, respectively [13, 14]. Therefore, AB5 toxins, including Stx, can be used as lectin-like glycan-binding probes. This chapter describes the preparation of fluorescence-conjugated Stx1B proteins and their application for flow cytometry.

2

Materials In this chapter, water indicates ultrapure water prepared by Milli-Q system (Merck Millipore, Darmstadt, Germany).

2.1 Expression of Stx1B in E. coli

1. Glycerol stock of E. coli BL21 (DE3) transformed with pET28Stx1-B subunit-His (pET-1BH) [15], which was a kind gift from Dr. Kiyotaka Nishikawa (Doshisha University, Japan). 2. Lysogeny broth (LB) liquid medium and agar plates containing 50 μg/mL kanamycin. 3. 1 M IPTG (isopropyl-β-D-1-thiogalactopyranoside): Dissolve 2.38 g IPTG in about 9 mL water and make up to 10 mL with water. Store at 20  C. 4. Thermostatic shaker that can shake a 2–3 L Erlenmeyer flask.

2.2 Purification of Stx1B

1. 5 equilibration buffer: 250 mM NaH2PO4 and 1.5 M NaCl, pH 7.0. Dissolve 39.0 g NaH2PO4·2H2O and 87.75 g NaCl in 950 mL water. Adjust to pH 7.0 using NaOH and make up to 1 L with water. Store at 4  C. 2. 1 equilibration buffer: To prepare 1 L, add 200 mL 5 equilibration buffer to 800 mL of water. Store at 4  C. 3. 50 protease inhibitor: Dissolve one tablet of cOmplete Protease Inhibitor Cocktail (Roche, Basel, Switzerland) in 1 mL water. Store at 20  C. 4. Filter: Millex-HV 0.45 μm (Merck Millipore). 5. TALON Metal Affinity Resin (Takara, Otsu, Japan) (see Note 1). 6. Disposable 5 mL polypropylene column (Thermo Fisher Scientific, MA, USA). 7. 1 M imidazole. 8. 10 mM imidazole/equilibration buffer: To prepare 200 mL, add 2 mL 1 M imidazole and 40 mL 5 equilibration buffer to 158 mL water. Store at 4  C. 9. 50 mM imidazole/equilibration buffer: To prepare 200 mL, add 10 mL 1 M imidazole and 40 mL 5 equilibration buffer to 150 mL water. Store at 4  C.

466

Toshiyuki Yamaji

10. 150 mM imidazole/equilibration buffer: To prepare 200 mL, add 30 mL 1 M imidazole and 40 mL 5 equilibration buffer to 130 mL water. Store at 4  C. 11. BCA Protein Assay Kit (Thermo Fisher Scientific). 12. Gel filtration column: NAP25 (bed volume, 2.5 mL) (GE Healthcare, Chicago, USA). 13. PBS (phosphate-buffered saline). 14. Ultrasonicator (e.g., W-225R H-1 probe (Heat SystemsUltrasonics, New York, USA)). 15. Spectrophotometer. 16. SDS-PAGE apparatus. 17. Coomassie Brilliant Blue (CBB) staining solution. 2.3 Fluorescent Labeling

1. Alexa Fluor 555 Protein Labeling Kit (Thermo Fisher Scientific). 2. Magnetic stirrer. 3. Column stand.

2.4

Flow Cytometry

1. Trypsin/EDTA. 2. Cell culture medium. 3. 1% bovine serum albumin (BSA)/PBS: Dissolve 0.5 g BSA in 50 mL PBS. 4. CO2 incubator. 5. Flow cytometer.

3

Methods Preparation of fluorescent Stx1B has been briefly described previously [16]. This chapter describes the method in detail.

3.1 Purification of Recombinant Stx1B

Recombinant Stx1B proteins tagged with six histidine residues at the carboxy termini (1BH) are prepared using E. coli and TALON resin as follows: 1. Scrape a small amount of pET-1BH-transformed BL21 (DE3) from the glycerol stock using a pipette tip or a toothpick (see Note 2), and add it to 50 mL LB medium with 50 μg/mL kanamycin (LB Kan). Incubate overnight at 37  C with vigorous shaking (see Note 3). 2. Add 20 mL of the amplified E. coli to 1 L LB Kan in a 2–3 L flask (see Note 4), and incubate further at 37  C with vigorous shaking. When the turbidity of the E. coli suspension reaches

Shiga Toxin B-Subunit and Flow Cytometry

467

0.5–1.0 OD at 600 nm, add 1 mL 1 M IPTG (final 1 mM), and incubate at 37  C for a further 6 h to allow expression of Stx1B proteins. 3. Transfer the E. coli suspension to two 500 mL centrifuge tubes. Centrifuge them at 3000 rpm (1700–2000  g) at 4  C for 15 min, and discard the supernatant. The E. coli pellets may be stored at 80  C until the following purification step (see Note 5). 4. Add 45 mL cold equilibration buffer and 1 mL 50 protease inhibitor to the E. coli pellets from the 500 mL culture and resuspend it well by pipetting (see Note 6). 5. Sonicate the suspension (about 50 mL) on ice, using an ultrasonicator (sonication for 10 s with an interval for 1 min, repeated four times). 6. Transfer the sonicated suspension to a 50 mL tube and centrifuge it for 20 min at 10,000 rpm (9100  g) at 4  C. Transfer the supernatant to a fresh tube (see Note 7). 7. Add 20 mL cold equilibration buffer and 0.4 mL 50 protease inhibitor to the precipitates, and repeat steps 5 and 6. Combine the supernatants (50 + 20 ¼ 70 mL) and divide between two 50 mL tubes (35 mL in each). 8. To remove debris further, filter the supernatant using a MillexHV 0.45 μm filter, and pour the filtrate into a fresh 50 mL tube. 9. Wash 4 mL TALON resin with equilibration buffer four times (see Note 8). 10. Mix 35 mL of the filtrate with 2 mL TALON resin and rotate it for 2 h at 4  C. 11. Centrifuge the mixture for 15 min at 3000 (1700–2000  g) at 4  C and discard the supernatant.

rpm

12. Wash the resin holding Stx1B-His proteins with 20 mL equilibration buffer three times at 4  C. (Perform the following operations in a cold room.) 13. Resuspend the resin in 10 mL equilibration buffer and pack the resin slurry into an open column. Discard the flowthrough. 14. To wash the resin further, gently add 5 mL equilibration buffer three times, 8 mL 10 mM imidazole buffer twice, and 2 mL 50 mM imidazole buffer five times, in that order (see Note 9). Discard the flow-through. 15. To elute the Stx1B-His proteins, sequentially add 1 mL 150 mM imidazole buffer eight times (see Note 10). Collect each fraction in a 1.5 mL tube.

Toshiyuki Yamaji

16. Measure the absorbance of each fraction at 280 nm (OD280) and then combine the fractions including proteins. Stx1B-His proteins should be mainly present in fractions 3 and 4 (total 2 mL). 17. Wash with 2 mL PBS in an NAP25 open column more than seven times for equilibration. 18. To remove imidazole and replace the buffer with PBS, apply 2 mL of the fraction containing the Stx1B-His proteins (step 16) to the NAP25 column, and then sequentially add 0.5 mL PBS ten times. Collect each fraction in a 1.5 mL tube. 19. Measure the absorbance of each fraction at 280 nm, and combine the fractions that include proteins. Measure the concentration of Stx1B-His using the BCA method (see Note 11). 20. Purified proteins should be confirmed using SDS-PAGE and CBB staining (Fig. 2).

(kDa) 100 75 50 37 25 20 15

Stx1B

10

Marker

468

Fig. 2 Recombinant Stx1B proteins. Purified recombinant proteins were run on a 13.5% acrylamide SDS-PAGE gel in a reduced condition, and the gel was stained with Coomassie Brilliant Blue

Shiga Toxin B-Subunit and Flow Cytometry

3.2 Fluorescence Conjugation of Recombinant Stx1B

469

To prepare fluorescence-conjugated Stx1B, an Alexa Fluor Protein Labeling Kit can be used (Thermo Fisher Scientific). Labeling should be performed according to the manufacturer’s protocol. A brief protocol for Alexa Fluor 555 labeling is described as follows: 1. Add 1/10 volume (80–100 μL) 1 M NaHCO3 (final 0.1 M) to 1.2 mg/mL Stx1B-His in PBS (about 720–900 μL) (see Note 12). 2. Transfer the protein solution to a vial of reactive fluorescent dye (Alexa Fluor 555), and invert several times to dissolve the dye. 3. Stir the reaction mixture using a magnetic stirrer for 1 h at room temperature. (The vial should contain a small magnetic stirrer bar.) (Perform the following operations in a cold room.) 4. While stirring the reaction mixture, load the size-exclusion purification resin into the column until the resin is about 3 cm from the top of the column, and allow excess PBS to drain into the column bed. 5. Carefully load the reaction mixture (step 3) onto the column (see Note 13). Continue adding PBS until the fluorescenceconjugated proteins have been eluted. The labeled proteins can be distinguished from the unincorporated dyes (see Note 14). Collect the first band containing the labeled proteins into a collecting tube. 6. Measure protein concentration using the BCA method and the absorbance at 555 nm in a cuvette with a 1 cm path length. The approximate molecular weight of the recombinant 1B pentamer is 60,000. Calculate the degree of labeling as follows: moles dye per mole protein ¼ A555/(150,000  protein concentration (M)).

3.3 Application of Fluorescent Stx1B to Flow Cytometry

1. Culture cells in a 6-well plate for 1–3 days (see Note 15). 2. Trypsinize cells (1–4  105 cells/well) if the cells are adherent. 3. Add culture medium to the trypsinized cells and transfer the cell suspension to a 1.5 mL tube. 4. Centrifuge the tube for 5 min at 2000 rpm (400  g) at 4  C. Discard the supernatant. 5. Add 1 mL cold 1% BSA/PBS to the cell pellets, divide the cell suspension into two 1.5 mL tubes, and centrifuge again. Discard the supernatant. 6. Add 100 μL 10 μg/mL Alexa Fluor 555-conjugated Stx1B in 1% BSA/PBS to one tube containing cells, whereas add 100 μL 1% BSA/PBS to the other tube as a negative control. Suspend the cells well and incubate on ice for 45–60 min.

470

Toshiyuki Yamaji

Alexa555-Stx1B Alexa555-CtxB

A375

BeWo

JEG-3

LNCaP

Alexa555-Stx1B

Alexa555-CtxB

Alexa555-Stx1B

FITC-CtxB

K562

THP-1

HL60

HAP1

Fig. 3 Surface binding of Alexa Fluor 555-Stx1B with various cell lines. Cells were stained with (black bold lines) or without (black dashed lines) Alexa Fluor 555-Stx1B or fluorescent CtxB (Alexa Fluor 555-CtxB or FITCCtxB) and analyzed using FACS

7. Add 1 mL 1% BSA/PBS and centrifuge the tube for 5 min at 2000 rpm (400  g) at 4  C. Discard the supernatant. 8. Add 600–800 μL 1% BSA/PBS and pass the suspension through a mesh to remove debris. 9. Analyze the binding of Stx1B to cell surfaces using a flow cytometer. As examples of use of the Alexa Fluor 555-conjugated Stx1B proteins, the binding of Alexa Fluor 555-Stx1B with various cell lines and sphingolipid-remodeled HeLa cell lines is shown in Figs. 3 and 4. In Fig. 3, Alexa Fluor 555-conjugated CtxB (Thermo Fisher Scientific) and FITC-conjugated CtxB (Sigma, St. Louis, USA) were also used to compare cell-surface binding between Stx1B and CtxB. Stx1B bound to several cell lines, including HeLa

Shiga Toxin B-Subunit and Flow Cytometry

A

471

B Serine+Palmitoyl CoA

Serine palmitoyltransferase

de novo pathway

(SPTLC1, SPTLC2, SPTSSA)

Sphingosine

HeLa

SPTSSA

Ceramide

salvage pathway

UGCG

UGCG

LAPTM4A

B4GalT5

TM9SF2

GlcCer B4GalT5 LacCer LAPTM4A

A4GalT

Stx

TM9SF2 Gb3

A4GalT

Alexa555-Stx1B

Fig. 4 Confirmation of loss of Gb3 on sphingolipid-remodeled HeLa cell lines. (a) Sphingolipid enzymes. (b) Surface binding of Alexa Fluor 555-Stx1B on sphingolipid-remodeled HeLa cell lines. These cells were constructed using genome editing technology, as previously described [17, 18]. Cells were stained with (black bold lines) or without (black dashed lines) Alexa Fluor 555-Stx1B. In ΔSPTSSA cells, the effect of adding sphingosine (7.5 μM) on Stx1B binding is indicated as a gray line. In ΔLAPTM4A and ΔTM9SF2 cells, the gray line indicates staining in LAPTM4A- and TM9SF2-restored cells, respectively

(human cervix adenocarcinoma cell line), JEG-5 (human placenta choriocarcinoma cell line), THP-1 (human monocytic leukemia cell line), and HAP1 (human near-haploid cell line derived from the KBM7 chronic myelogenous leukemia cell line) (although there was limited binding to HAP1), indicating that these cell lines are Gb3-positive. Conversely, BeWo (human placenta choriocarcinoma cell line), LNCaP (human prostate adenocarcinoma cell line), K562 (human myelogenous leukemia cell line), and HL60 (promyelocytic leukemia cell line) were Stx1B-binding-negative. CtxB bound to all cell lines shown in this figure, although the amount of CtxB binding varied among these cell lines. Previously, we established several sphingolipid-remodeled HeLa cells using genome editing technology [17, 18]. Disruption of serine palmitoyltransferase small subunit A (SPTSSA), glucosylceramide (GlcCer) synthase (UGCG), lactosylceramide (LacCer) synthase (B4GalT5), and Gb3 synthase (A4GalT) showed loss of Stx1B binding (Fig. 4a, b). In SPTSSA-deficient cells, the addition of sphingosine recovered Gb3 expression through a salvage pathway. LAPTM4A and TM9SF2 were identified as genes that conferred resistance to Stx1 via the reduction of Gb3 when these genes

472

Toshiyuki Yamaji

were disrupted [18–20]. We confirmed that disruption of the LAPTM4A and TM9SF2 genes reduced Stx1B binding and restoration of their cDNAs recovered Gb3 expression (Fig. 4b). In this way, Stx1B can easily be used to confirm the knockout of sphingolipid-related genes in cell lines that express Gb3.

4

Notes 1. TALON Metal Affinity Resin consists of a tetradentate chelator charged with cobalt (pink in color), which has a remarkable affinity and specificity for 6 histidine-tagged proteins. 2. Frozen glycerol stock does not need to be thawed. Colonies of transformed E. coli on an LB Kan plate may be directly used for the next step (amplification). 3. A rotation speed of more than 200 rpm is required for aeration in a shaker, which helps more proteins to be produced. 4. At this point, the OD600 is likely less than 0.1. Additional incubation for 2–3 h is required before adding IPTG. 5. The expression of Stx1B may be confirmed using SDS-PAGE and CBB staining at this time. 6. Pellets of E. coli should be completely dispersed. 7. Take care not to take precipitates together because this may reduce the purity of Stx1B. Although most Stx1B proteins are likely to be contained in the precipitates, the supernatant will also contain a certain amount of Stx1B proteins. 8. Wash TALON resin during step 6 in Subheading 3.1 (centrifugation of sonicated suspension). 9. Monomeric Stx1B, as well as other irrelevant proteins, may be eluted using 50 mM imidazole buffer, but most Stx1B pentamers remain attached to the resin in this condition due to the strong binding. 10. Apply gently so as not to disturb the top of the resin. 11. Before conjugating fluorescent dye to Stx1B proteins, it should be confirmed whether the proteins can bind to Gb3-expressing cell surfaces by using anti-His6 antibody, which can be purchased from MBL (Nagoya, Japan) or Nacalai (Kyoto, Japan). 12. If the protein concentration is not sufficient, use centrifugal filters (e.g., Centricon (Millipore)) to concentrate it further. 13. Apply gently so as not to disturb the column bed.

Shiga Toxin B-Subunit and Flow Cytometry

473

14. A faint, broad band, which contains fluorescence-conjugated Stx1B proteins, can be observed below the band containing free fluorescent dyes. Only the faint band should be collected. 15. Increased cell density may reduce Stx1B binding to cell surfaces.

Acknowledgments This work was supported by AMED (No. JP19ae0101068j0104, 19fm0208005j0103) and JSPS KAKENHI (No. JP17K07357). References 1. Riley LW, Remis RS, Helgerson SD, McGee HB, Wells JG, Davis BR, Hebert RJ, Olcott ES, Johnson LM, Hargrett NT, Blake PA, Cohen ML (1983) Hemorrhagic colitis associated with a rare Escherichia coli serotype. N Engl J Med 308:681–685 2. Karmali MA, Steele BT, Petric M, Lim C (1983) Sporadic cases of hemolytic uremic syndrome associated with fecal cytotoxin and cytotoxin-producing Escherichia coli. Lancet 1:619–620 3. O’Brien AD, Holmes RK (1987) Shiga and Shiga-like toxins. Microbiol Rev 51:206–220 4. Paton JC, Paton AW (1998) Pathogenesis and diagnosis of Shiga toxin-producing Escherichia coli infections. Clin Microbiol Rev 11:450–479 5. Obrig TG, Moran TP, Colinas RJ (1985) Ribonuclease activity associated with the 60S ribosome-inactivating proteins ricin A, phytolaccin and Shiga toxin. Biochem Biophys Res Commun 130:879–884 6. Jacewicz M, Clausen H, Nudelman E, Donohue-Rolfe A, Keusch GT (1986) Pathogenesis of shigella diarrhea. XI. Isolation of a shigella toxin-binding glycolipid from rabbit jejunum and HeLa cells and its identification as globotriaosylceramide. J Exp Med 163:1391–1404 7. Lindberg AA, Brown JE, Stro¨mberg N, Westling-Ryd M, Schultz JE, Karlsson KA (1987) Identification of the carbohydrate receptor for Shiga toxin produced by Shigella dysenteriae type 1. J Biol Chem 262:1779–1785 8. Lingwood CA, Law H, Richardson S, Petric M, Brunton JL, De Grandis S, Karmali M (1987) Glycolipid binding of purified and recombinant Escherichia coli produced verotoxin in vitro. J Biol Chem 262:8834–8839 9. Samuel JE, Perera LP, Ward S, O’Brien AD, Ginsburg V, Krivan HC (1990) Comparison of

the glycolipid receptor specificities of Shiga-like toxin type II and Shiga-like toxin type II variants. Infect Immun 58:611–618 10. Okuda T, Tokuda N, Numata S, Ito M, Ohta M, Kawamura K, Wiels J, Urano T, Tajima O, Furukawa K, Furukawa K (2006) Targeted disruption of Gb3/CD77 synthase gene resulted in the complete deletion of globo-series glycosphingolipids and loss of sensitivity to verotoxins. J Biol Chem 281:10230–10235 11. Ling H, Boodhoo A, Hazes B, Cummings MD, Armstrong GD, Brunton JL, Read RJ (1998) Structure of the Shiga-like toxin I B-pentamer complexed with an analogue of its receptor Gb3. Biochemistry 37:1777–1788 12. Holmgren J, Lo¨nnroth I, Svennerholm L (1973) Fixation and inactivation of cholera toxin by GM1 ganglioside. Scand J Infect Dis 5:77–78 13. Byres E, Paton AW, Paton JC, Lo¨fling JC, Smith DF, Wilce MC, Talbot UM, Chong DC, Yu H, Huang S, Chen X, Varki NM, Varki A, Rossjohn J, Beddoe T (2008) Incorporation of a non-human glycan mediates human susceptibility to a bacterial toxin. Nature 456:648–652 14. Baenziger J, Fiete D (1979) Structural determinants of Ricinus communis agglutinin and toxin specificity for oligosaccharides. J Biol Chem 254:9795–9799 15. Watanabe M, Matsuoka K, Kita E, Igai K, Higashi N, Miyagawa A, Watanabe T, Yanoshita R, Samejima Y, Terunuma D, Natori Y, Nishikawa K (2004) Oral therapeutic agents with highly clustered globotriose for treatment of Shiga toxigenic Escherichia coli infections. J Infect Dis 189:360–368 16. Yamaji T, Nishikawa K, Hanada K (2010) Transmembrane BAX inhibitor motif containing (TMBIM) family proteins perturbs a trans-

474

Toshiyuki Yamaji

Golgi network enzyme, Gb3 synthase, and reduces Gb3 biosynthesis. J Biol Chem 285:35505–35518 17. Yamaji T, Hanada K (2014) Establishment of HeLa cell mutants deficient in sphingolipidrelated genes using TALENs. PLoS One 9: e88124 18. Yamaji T, Sekizuka T, Tachida Y, Sakuma C, Morimoto K, Kuroda M, Hanada K (2019) A CRISPR screen identifies LAPTM4A and TM9SF proteins as glycolipid-regulating factors. iScience 11:409–424 19. Tian S, Muneeruddin K, Choi MY, Tao L, Bhuiyan RH, Ohmi Y, Furukawa K, Furukawa K, Boland S, Shaffer SA, Adam

RM, Dong M (2018) Genome-wide CRISPR screens for Shiga toxins and ricin reveal Golgi proteins critical for glycosylation. PLoS Biol 16:e2006951 20. Pacheco AR, Lazarus JE, Sit B, Schmieder S, Lencer WI, Blondel CJ, Doench JG, Davis BM, Waldor MK (2018) CRISPR screen reveals that EHEC’s T3SS and Shiga toxin rely on shared host factors for infection. MBio 9: e01003–e01018 21. Yamaji T, Hanada K (2015) Sphingolipid metabolism and interorganellar transport: localization of sphingolipid enzymes and lipid transfer proteins. Traffic 16:101–122

Chapter 46 LecB, a High Affinity Soluble Fucose-Binding Lectin from Pseudomonas aeruginosa Emilie Gillon, Annabelle Varrot, and Anne Imberty Abstract LecB/PA-IIL (Pfam PF07472) from bacterium Pseudomonas aeruginosa is a fucose-binding lectin with unusual high affinity for glycans. The occurrence of LecB and related proteins is limited to few opportunistic bacterial species, some of them being responsible for severe infections in immune-compromised patients. This lectin is therefore of interest as a target for the design of anti-infectious compounds, but can also be used for research and biotechnology. LecB is a small protein that can be produced in good quantity in recombinant system and purified by affinity chromatography. Key words Pseudomonas aeruginosa, Soluble lectin, Fucose, Blood-group oligosaccharide, Microcalorimetry

1

Introduction Pseudomonas aeruginosa is one of the main human opportunistic pathogens and is responsible for infections such as folliculitis, keratitis, otitis and also pneumopathies in cystic fibrosis patients resulting in morbidity and mortality [1, 2]. Among several virulence factors, P. aeruginosa produces two soluble lectins named LecA (or PA-IL) and LecB (or PA-IIL) that are specific for galactose and fucose, respectively [3]. LecB consists of four subunits of 11.73 kDa, each with a high affinity binding site for L-fucose and fucose-containing glycoconjuguates [4]. The lectin also binds to Dmannose but with lower affinity. LecB is produced in the cytoplasm of P. aeruginosa but has been shown to be located in the outer membrane and involved in biofilm formation [5]. The crystal structure of LecB in complex with fucose [6] and with mannose [7] uncovered the molecular basis of its ligand recognition. LecB forms homotetramer, each monomer presenting a fucose binding site with two calcium ions that are directly involved in carbohydrate coordination. LecB displays exceptionally high affinities for fucose, with Kd values in the micromolar range

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_46, © Springer Science+Business Media, LLC, part of Springer Nature 2020

475

476

Emilie Gillon et al.

[8]. It binds to a large range of fucosylated antigens (ABH, Lewis, P and I systems) present on human tissues and in human milk [9, 10] with highest reported affinity in the nM range for Lewis a epitope [9] and for biantennary N-glycans presenting two H-type epitopes at the extremities of branches [11]. LecB is a target of interest for anti-adhesion and anti-biofilm compounds, and the search for glyco-derived drugs with potential therapeutical application is very active [12, 13]. As other lectins with specificity toward human glycans, LecB is also of biotechnical interest for the purification of glycoproteins or the labeling of glycoconjugates on tissues. LecB is now included in the few lectin arrays proposed in different laboratories for the fine characterization of glycans. Due to its propensity to easily form crystals, LecB has been utilized as a template for structural determination of smaller biomolecules that are difficult to crystallize by themselves: after attachment of a fucose residue, modified DNA duplexes or short antibacterial peptides could be co-crystallized with LecB and structurally characterized [14, 15] (see Fig. 1). Most studies have been performed using LecB from P. aeruginosa PAO1 strain. However, recent works demonstrated sequence variations in LecB when analyzing the many isolates from clinical or environmental origin. Three different clusters of sequences were identified, represented by LecB from PAO1, PA7, and PA14 virulent strains [11, 16]. Some of the amino acid variations are located near the fucose binding site but affect only slightly the specificity and affinity of LecB. LecB-like lectins are also present in a limited number of bacteria from the Ralstonia, Chromobacterium and Burkholderia species. CV-IIL from Chromobacterium violaceum is very similar to LecB [17], while RS2L from Ralstonia solanacearum is specific for mannoside [18]. Such variation in specificity has been demonstrated to be due to one amino acid variation at position 22 [19]. Burkholderia cenocepacia hosts four different LecB-like proteins with dimeric association of lectin and extension by several other proteins [20, 21]. The details of production and characterization listed here can be therefore applied, with small adjustments to the whole family of LecB-like lectins.

2

Materials Prepare all solutions using analytical grade reagents and ultrapure water and then filter on 0.22 μM membrane. Follow waste disposal regulations.

2.1

LecB Production

1. Dry block heater at 42  C. 2. Incubation shaker with cooling option. 3. 250 mL and 3 L baffled Erlenmeyer flasks. 4. Competent bacteria: Escherichia coli BL21(DE3).

LecB/PA-IIL from Pseudomonas aeruginosa

477

Fig. 1 Alignment of amino acid sequences for LecB sequences and crystal structure of LecBPA14 in complex with αMeFuc (5A6X). Protein is represented by ribbon, monosaccharides by sticks, and calcium ions by spheres

5. Plasmid: pET25b-LecB. 6. Lennox L Broth Base. 7. 100 mg/mL ampicillin. 8. 1 M isopropyl-β-D-thiogalactopyranoside (IPTG). 9. 1 L polycarbonate centrifuge bottle. 10. 50 mL sterile conical centrifuge tubes. 11. Centrifuge with adapted rotors. 2.2

LecB Purification

1. Endonuclease such as Benzonase or Denarase at 250 U/μL. 2. Rotating wheel. 3. One Shot tabletop cell disruptor (Constant Systems Ltd). 4. 35 mL centrifuge tube and centrifuge with rotor going to 24,000  g. 5. 0.45 μm PES syringe filter. 6. C10/10 column with flow adapter (GE Healthcare Life Sciences). 7. Equilibration buffer: 20 mM Tris pH 7.5, 100 mM NaCl, and 100 μM CaCl2 (see Note 1).

478

Emilie Gillon et al.

8. Elution buffer: 20 mM Tris pH 7.5, 100 mM NaCl, and 100 mM D-Mannose. 9. Low- or Medium-Pressure Chromatography Systems such as AKTAprime (GE Healthcare Life Sciences) or NGC (Bio-Rad). 10. 5 mL collection tubes. 11. Snakeskin dialysis tubing 3500 MWCO (Thermo Fisher Scientific). 12. 20% ethanol. 2.3 Quality Control by Evaluation of Affinity

1. ITC200 microcalorimeter (Malvern Panalytical). 2. NanoDrop 2000 spectrophotometer (Thermo Scientific). 3. ITC buffer: 20 mM Tris pH 7.5, 100 mM NaCl, and 100 μM CaCl2. 4. Methyl-α-L-fucopyranoside (αMeFuc, Sigma 32198).

3 3.1

Methods LecB Production

3.1.1 Transformation and Preculture

1. The plasmid pET25b-LecB has been obtained from template DNA from Pseudomonas aeruginosa ATCC 33347 as previously described [8] and is available upon request. 2. Add 1 μL of pET25b-LecB plasmid to 100 μL of competent BL21(DE3) cells before 10 min incubation on ice. 3. Heat shock at 42  C for 45 s before a 3 min stay on ice. 4. Add 900 μL of liquid LB and incubate horizontally for 1 h at 37  C with shaking at 180 rpm. 5. Spread 100 μL on a petri dish with LB agar and 100 μg/mL ampicillin, and incubate at 37  C overnight. 6. Transfer one colony to 10 mL liquid LB containing 100 μg/ mL ampicillin, and incubate overnight with shaking at 180 rpm.

3.1.2 Culture

1. Inoculate 1 L of liquid LB supplemented with 100 μg/mL ampicillin in 3 L Erlenmeyer with the overnight preculture. 2. Incubate at 37  C with shaking at 180 rpm until OD600nm reached 0.6. 3. Switch the temperature to 30  C, induce expression by adding 0.25 mM IPTG, and culture for 3 h at 180 rpm. 4. Centrifuge at 6000  g, 4  C for 15 min. 5. Resuspend the pellet in 35 mL liquid LB, transfer in 50 mL sterile conical tube, and centrifuge 6000  g for 10 min. 6. Remove supernatant and store the pellet at 20  C until use.

LecB/PA-IIL from Pseudomonas aeruginosa

3.2

LecB Purification

479

1. Resuspend the pellet in 30 mL of equilibration buffer. 2. Break the cells using a One Shot tabletop cell disruptor at a pressure of 1.9 kbar. 3. Centrifuge 24,000  g, 30 min at 4  C to clarify the sample. 4. Filter the supernatant on 0.45 μm with syringe filter. 5. Equilibrate a 10 mL D-Mannose agarose resin packed in C10/10 column connected to automated chromatography system such as NGC system with 5 column volumes (CV) of equilibration buffer at a flow rate of 1 mL/min. 6. Load the filtered supernatant at a flow rate of 1 mL/min and collect the flow-through. 7. Rinse the column with the equilibration buffer to wash out contaminants until the baseline of OD280nm on the chromatogram stabilized to a low value (3 mL/min flow rate). 8. Elute the protein with the elution buffer using a flow rate of 3 mL/min and collect fractions of 5 mL. 9. Rinse the column with 3 CVs of ultrapure water at a flow rate of 3 mL/min and 2 CVs of 20% ethanol at flow rate of 1 mL/min prior storage at 4  C. 10. Collect the fractions of the eluted peak and dialyzed against ultrapure water for 1 week, changing the bath each morning and evening. 11. Lyophilize protein and store at 20  C (see Note 2).

3.3 Quality Control by Evaluation of Affinity

3.3.1 Sample Preparation

Isothermal titration calorimetry (ITC) is an analytical technique, which is considered as the gold standard for analyzing intermolecular interactions in solution. It is widely used for quality control in biotechnology since it gives access to quantitative analysis of the interaction of a protein with its ligand. A single experiment allows to determine stoichiometry (n), binding enthalpy (ΔH), and affinity (i.e., dissociation constant, Kd). Since the method is based on the heat released or absorbed in the binding process, it is very well suited for protein-carbohydrate interaction, where the large number of hydrogen bonds result in strong exothermic signal [22]. 1. Prepare 400 μL of LecB at 50 μM in ITC buffer from the lyophilized protein. Weight 1.5 excess amount of the necessary protein, dissolve in the corresponding volume of buffer, and centrifuge 7000  g for 5 min. 2. Recover the supernatant and determine the precise concentration by spectroscopy at 280 nm using a molecular weight of 11,732 Da per monomer and a theoretical molar extinction coefficient of 6990 corresponding to absorbance 0.1% (¼1 g/ L) of 0.596. 3. Adjust the concentration to 50 μM with the ITC buffer.

480

Emilie Gillon et al.

4. Prepare 70 μL of 0.5 mM of αMeFuc in the same buffer as the lectin (see Note 3). 5. Degas all samples. 3.3.2 ITC Experiment

1. Fill the measuring cell with 50 μM LecB solution and the syringe with 0.5 mM αMeFuc solution. Cell volume is 200 μL and syringe volume is 40 μL. 2. Set up the parameters to DP 8, injection number 20, injection volume of 1 μL for the first one, and 2 μL for all following ones with an injection delay of 120 s. 3. Use the Origin software of the microcalorimeter for data analysis (see Note 4). The default parameter with a “one-site” model is fully compatible with the present experiments. Data will be obtained as indicated in Fig. 2 when clicking on the “ITC data” button (see Note 5). 4. Quality control is performed by comparing the obtained thermodynamic of binding with literature [23]. LecB was reported to have a submicromolar affinity for aMeFuc (Kd ¼ 430 nM) with strong enthalpy (ΔH ¼ 41.3 kJ/mol) and a stoichiometry corresponding to approximately one binding site per monomer (n ¼ 0.77).

4

Notes 1. LecB is a calcium-dependent lectin, and it is advised to maintain a CaCl2 concentration of 100 μM through the purification process. If needed, this concentration can be decreased to 10 μM. 2. LecB is a very robust protein that can be lyophilized and stored at 20  C for months or years. It is only recommended to aliquot in small quantities to avoid repeating cycles of defreezing and refreezing. 3. For a perfect ITC measurement, the main recommendation is to avoid buffer mismatch. Use exactly the same buffer for the protein solution and the sugar solution. 4. Alternatives to Origin are available in the public domain. The suite of integrated software packages NITPIC, SEDPHAT, and GUSSI offer powerful treatment of data with automated shape analysis of the injection peaks [24]. It also allows for integrating results from different calorimetric titration experiments into a global analysis. 5. In the Origin software, the default units are calories. This could be changed to international units Watt and Joules using the Tab ITC and then Display Watt/Joules.

LecB/PA-IIL from Pseudomonas aeruginosa

481

Time (min) 0

10

20

30

40

50

0.00

µcal/sec

-0.50 -1.00 -1.50

Data: Data1_NDH Model: Onesites Chi 2/DoF = 6021 N 0.835 ±0.00154 Sites K 1.94E6 ±9.93E4 M1 ΔH -1.105E4 ±34.71 cal/mol ΔS -8.30 cal/mol/deg

-2.00 -2.50

kcal mol-1 of injectant

0.00 -2.00 -4.00 -6.00 -8.00 -10.00 -12.00 0.0

0.5

1.0

1.5

2.0

Molar Ratio

Fig. 2 Typical ITC experiment of titration of LecBPAO1 by αMeFuc with thermogram (top) and integrated data (bottom). The observed stoichiometry of 0.835 indicates here that 16.5% of the protein is inactive. This does not affect the determination of thermodynamics constant, since they are expressed as a function of the concentration of ligand. The enthalpy of binding is 11.05 Kcal/ mol (46 kJ/mole) with a weak unfavorable contribution of entropy ΔS of 8.3 cal/mol/deg. (i.e., TΔS ¼ 10.3 kJ/mol) which is in full agreement with previous data [23]

Acknowledgments The authors acknowledge support by the ANR PIA Glyco@Alps (ANR-15-IDEX-02), Labex ARCANE and CBH-EUR-GS (ANR-17-EURE-0003), and the French Cystic Fibrosis Association Vaincre La Mucoviscidose. References 1. Govan JR, Deretic V (1996) Microbial pathogenesis in cystic fibrosis: mucoid Pseudomonas aeruginosa and Burkholderia cepacia. Microbiol Rev 60(3):539–574

2. Mesaros N, Nordmann P, Plesiat P et al (2007) Pseudomonas aeruginosa: resistance and therapeutic options at the turn of the new millennium. Clin Microbiol Infect 13(6):560–578

482

Emilie Gillon et al.

3. Gilboa-Garber N (1982) Pseudomonas aeruginosa lectins. Methods Enzymol 83:378–385 4. Gilboa-Garber N, Katcoff DJ, Garber NC (2000) Identification and characterization of Pseudomonas aeruginosa PA-IIL lectin gene and protein compared to PA-IL. FEMS Immunol Med Microbiol 29(1):53–57 5. Tielker D, Hacker S, Loris R et al (2005) Pseudomonas aeruginosa lectin LecB is located in the outer membrane and is involved in biofilm formation. Microbiology 151 (Pt 5):1313–1323 6. Mitchell E, Houles C, Sudakevitz D et al (2002) Structural basis for oligosaccharidemediated adhesion of Pseudomonas aeruginosa in the lungs of cystic fibrosis patients. Nat Struct Mol Biol 9:918–921 7. Loris R, Tielker D, Jaeger K-E et al (2003) Structural basis of carbohydrate recognition by the lectin LecB from Pseudomonas aeruginosa. J Mol Biol 331:861–870 8. Mitchell EP, Sabin C, Sˇnajdrova´ L et al (2005) High affinity fucose binding of Pseudomonas aeruginosa lectin PA-IIL: 1.0 A˚ resolution crystal structure of the complex combined with thermodynamics and computational chemistry approaches. Proteins 58:735–748 9. Perret S, Sabin C, Dumon C et al (2005) Structural basis for the interaction between human milk oligosaccharides and the bacterial lectin PA-IIL of Pseudomonas aeruginosa. Biochem J 389:325–332 10. Wu AM, Wu JH, Singh T et al (2006) Interactions of the fucose-specific Pseudomonas aeruginosa lectin, PA-IIL, with mammalian glycoconjugates bearing polyvalent Lewis (a) and ABH blood group glycotopes. Biochimie 88(10):1479–1492 11. Sommer R, Wagner S, Varrot A et al (2016) The virulence factor LecB varies in clinical isolates: consequences for ligand binding and drug discovery. Chem Sci 7:4990–5001 12. Boukerb A, Rousset A, Galanos N et al (2014) Anti-adhesive properties of glycoclusters against Pseudomonas aeruginosa lung infection. J Med Chem 57:10275–10289 13. Wagner S, Sommer R, Hinsberger S et al (2016) Novel strategies for the treatment of Pseudomonas aeruginosa infections. J Med Chem 59(13):5929–5969 14. Baeriswyl S, Gan BH, Siriwardena TN et al (2019) X-ray crystal structures of short

antimicrobial peptides as Pseudomonas aeruginosa Lectin B complexes. ACS Chem Biol 14 (4):758–766 15. Roethlisberger P, Istrate A, Marcaida Lopez MJ et al (2016) X-ray structure of a lectinbound DNA duplex containing an unnatural phenanthrenyl pair. Chem Commun 52 (26):4749–4752 16. Boukerb AM, Decor A, Ribun S et al (2016) Genomic rearrangements and functional diversification of lecA and lecB lectincoding regions impacting the efficacy of glycomimetics directed against Pseudomonas aeruginosa. Frontiers Microb 7:811 17. Pokorna´ M, Cioci G, Perret S et al (2006) Unusual entropy driven affinity of Chromobacterium violaceum lectin CV-IIL towards fucose and mannose. Biochemistry 45:7501–7510 18. Sudakevitz D, Kostlanova N, Blatman-Jan G et al (2004) A new Ralstonia solanacearum high affinity mannose-binding lectin RS-IIL structurally resembling the Pseudomonas aeruginosa fucose-specific lectin PA-IIL. Mol Microbiol 52:691–700 19. Adam J, Pokorna M, Sabin C et al (2007) Engineering of PA-IIL lectin from Pseudomonas aeruginosa—unravelling the role of the specificity loop for sugar preference. BMC Struct Biol 7:36 20. Lameignere E, Malinovska´ L, Sla´vikova´ M et al (2008) Structural basis for mannose recognition by a lectin from opportunistic bacteria Burkholderia cenocepacia. Biochem J 411:307–318 21. Sˇula´k O, Cioci G, Lameigne`re E et al (2011) Burkholderia cenocepacia BC2L-C is a super lectin with dual specificity and proinflammatory activity. PLoS Pathog 7:e1002238 22. Dam TK, Brewer CF (2002) Thermodynamic studies of lectin-carbohydrate interactions by isothermal titration calorimetry. Chem Rev 102(2):387–429 23. Sabin C, Mitchell EP, Pokorna´ M et al (2006) Binding of different monosaccharides by lectin PA-IIL from Pseudomonas aeruginosa: thermodynamics data correlated with X-ray structures. FEBS Lett 580:982–987 24. Brautigam CA, Zhao H, Vargas C et al (2016) Integration and global analysis of isothermal titration calorimetry data for studying macromolecular interactions. Nat Protoc 11 (5):882–894

Chapter 47 Sialoglycovirology of Lectins: Sialyl Glycan Binding of Enveloped and Non-enveloped Viruses Nongluk Sriwilaijaroen and Yasuo Suzuki Abstract On the cell sur “face”, sialoglycoconjugates act as receptionists that have an important role in the first step of various cellular processes that bridge communication between the cell and its environment. Loss of Sia production can cause the developmental of defects and lethality in most animals; hence, animal cells are less prone to evolution of resistance to interactions by rapidly evolved Sia-binding viruses. Obligative intracellular viruses mostly have rapid evolution that allows escape from host immunity, leading to an epidemic variant, and that allows emergence of a novel strain, occasionally leading to pandemics that cause healthsocial-economic problems. Recently, much attention has been given to the mutual recognition systems via sialosugar chains between viruses and their host cells and there has been rapid growth of the research field “sialoglycovirology.” In this chapter, the structural diversity of sialoglycoconjugates is overviewed, and enveloped and non-enveloped viruses that bind to Sia are reviewed. Also, interactions of viral lectins-host Sia receptors, which determine viral transmission, host range, and pathogenesis, are presented. The future direction of new therapeutic routes targeting viral lectins, development of easy-to-use detection methods for diagnosis and monitoring changes in virus binding specificity, and challenges in the development of suitable viruses to use in virus-based therapies for genetic disorders and cancer are discussed. Key words Sialylglycoconjugates, Viral lectins, Receptor-binding specificity, Virus–sialic acid interactions, Host/tissue/cellular tropism

1

Introduction One of the greatest discoveries in medical sciences is the discovery of sialic acid in the mid-1930s to mid-1980s. The Swedish chemist Blix separated the polyhydroxy acid part from a disaccharide crystalline compound isolated in 1936 from a boiling water solution of bovine submaxillary mucin of salivary glands and proposed in 1952 that it be internationally called “sialic acid (Sia)” (the Greek root sialos ¼ saliva) [1]. The German biochemist Klenk purified brain glycolipid from ganglion cells, neurons in brain gray matter, and called it “ganglioside” [2], and he called an acidic crystal from methanolysis cleaving the glycosidic linkages of the ganglioside “neuraminic acid (Neu)” in 1941 [2]. The German biochemist

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_47, © Springer Science+Business Media, LLC, part of Springer Nature 2020

483

484

Nongluk Sriwilaijaroen and Yasuo Suzuki

Gottschalk found a water-soluble, nitrogen-containing compound released from ovomucin/human urinary mucoprotein after incubation with V. cholerae/influenza A or B virus and reported it to be “N-substituted isoglucosamine” in 1951 [3]. The pathogen enzyme that releases this monosaccharide was named receptordestroying enzyme (RDE) according to its activity by Burnet and Stone in 1947. Later, the name “sialidase” was proposed by Heimer and Meyer in 1956 [4], and the name “neuraminidase (NA)” was proposed by Blix, Gottschalk, and Klenk in 1957 [5]. Finally, it appeared that a characteristic building block of Sia, N-substituted isoglucosamine, and other compounds found in the 1950s, such as hemataminic acid from a hematoside glycolipid of equine blood stroma in 1951 [6] and lactaminic acid from cow colostrum in 1954 [7], is identical to Neu, which is 5-amino-3,5-dideoxy-Dglycero-D-galacto-non-2-ulosonic acid (9-carbon α-keto acids, C9H17N1O8). In 1986, 2-keto-3-deoxy-D-glycero-D-galacto-nononic acid (C9H16O9), which is not Neu but a deaminated Neu, was isolated from rainbow trout egg polysialoglycoprotein and named “ketodeoxynonulosonic (KDN)” [8]. Currently, Sia (Fig. 1a) is used as the generic name for a family of acidic sugars derived from 9-carbon backbone derivatives of Neu and KDN. As shown in Fig. 1b, Sia modifications of the carbon backbone at the C5 position give four core molecules: Neu (C5-NH2, not found in nature, formed due to a side effect of methanolysis provoking de-N-acylation of Sia), Neu5Ac (C5-Nacetyl, the most common derivative), Neu5Gc (C5-N-glycolyl), and KDN (C5-hydroxyl). These four core molecules can carry more substituents as described in the legend of Fig. 1b [9]. In nature, Sias exist predominantly as sialylglycoconjugates on Nand O-linked glycoproteins as well as gangliosides in the cell plasma membrane (Fig. 1c) and occasionally on glycosylphosphatidylinositol (GPI)-anchored proteins as well as secreted glycoproteins. Negatively charged Sia moieties are typically found at the outermost ends of glycoconjugates linked through (1) α2-3 or α2-6 to Gal or GalNAc found in glycoproteins and gangliosides [10–12], (2) α2-6 to GlcNAc or Glc found in glycoproteins and gangliosides [13, 14], (3) α2-8 to the second Sia found in glycoproteins and gangliosides [15] and less often through (4) α2-4 to Gal or GlcNAc found in glycoproteins [16, 17], (5) α2-9 to the second Sia found in glycoproteins [18], and (6) Neu5Gc oligomer ((!5-OglycolylNeu5Gcα2!)n, n ¼ 4 to more than 40 Neu5Gc residues) only found so far in jelly coat glycoproteins of sea urchin eggs [19]. Figure 1 shows only the well-known sialylglycoconjugates recognized by viruses. These terminal sialyl linkage antennae on glycoconjugates are remarkably potential recognition determinants for endogenous

Enveloped and Non-enveloped Sia-Binding Viruses

485

Fig. 1 Sialoglycan structures of virus receptors on host cell membranes. Complexity levels of sialoglycan structures can be divided into four levels: (1) Sia core and core modifications, (2) Sia linkages, (3) branches,

486

Nongluk Sriwilaijaroen and Yasuo Suzuki

sialyl glycan-binding proteins (lectins) that mediate various biological processes, such as cell signaling that is pivotal for the development and maintenance of life functions, and for exogenous lectins of various exogenous agents such as bacteria, toxins, protozoa, and viruses [20] that mediate symbiotic or pathogenic processes. The parasitic relationship between hosts and pathogens drives the coevolutionary arms race between hosts and pathogens, resulting in diversity of Sias. For example, humans have evolved a mechanism to inactivate the gene encoding CMP-Neu5Ac hydroxylase, which converts CMP-Neu5Ac into CMP-Neu5Gc in other mammalian cells. In normal adult human cells, Neu5Gc is found less than 0.1% of total Sias from dietary intake [21], and loss of Neu5Gc production appears to render human cells resistant to some pathogen infections, such as infections with enterotoxigenic Escherichia coli K99 [22] and the malaria parasite Plasmodium ä Fig. 1 (continued) and (4) classes [20]. (a) The Sia structure is a 9-carbon α-keto carboxylic acid skeleton with different substituents R4, R5, R7, R8, and R9 as indicated. (b) The parent molecule, Neu, contains a substituent R5 being an amino group at C5. N-Acylation of the 5-amino group gives 5-N-acetyl-Neu (Neu5Ac) and hydroxylation of the 5-N-acetyl group gives 5-N-glycolyl-Neu (Neu5Gc). Deamination of the amino group at C5 of Neu gives 2-keto-3-deoxynononic acid (KDN). These four cores differing at the C5-position can carry one or more O-substituents of the hydroxyl groups with acetyl group(s) at C4, C7, C8, and/or C9 and less often a lactyl group at C9, a sulfate group at C8, or a methyl group at C8. Also, the C1 carboxylate group can react with a hydroxyl group or with the C5 amino group forming a lactone or a lactam, respectively, and free Sias with unsaturated and anhydro forms can also be found. (c) The C2 of Sia can form several types of glycosidic linkages with the penultimate sugar. Only three major linkages, Siaα2-3Gal, Siaα2-6Gal, and Siaα2-8Sia, are shown here. (d) Sialoglycans can be linear or branched (antennae). (e) Based on biomolecules under sugars, surface sialoglycans used for virus infection are classified into sialoglycoproteins and gangliosides (one or more Sia-containing glycosphingolipids). Sialoglycoproteins can be further classified on the basis of their covalent linkage to a protein through an Asn or a Ser/Thr into N-glycans or O-glycans, respectively. For O-glycans, the major core 1–4 subtypes based on the second sugar(s) attached to GalNAc-Ser/Thr are shown. Abbreviations: Cer ceramide (sphingosine-fatty acid complex), GM monosialoganglioside, GD disialoganglioside, GT trisialoganglioside, GQ tetrasialoganglioside, GM3 Siaα2-3Galβ1-4Glcβ1-10 Cer, GD3 Siaα2-8Siaα23Galβ1-4Glcβ1-10 Cer, ganglio-series, gangliosides containing GalNAcβ1-4Galβ1-4Glcβ1-10 Cer, GM2 GalNAcβ1-4(Siaα2-3)Galβ1-4Glcβ1-10 Cer, GM1 (GM1a) Galβ1-3GalNAcβ1-4(Siaα2-3)Galβ1-4Glcβ1-10 Cer, GM1b Siaα2-3Galβ1-3GalNAcβ1-4Galβ1-4Glcβ1-10 Cer, GD1a Siaα2-3Galβ1-3GalNAcβ1-4(Siaα2-3)Galβ14Glcβ1-10 Cer, GT1a Siaα2-8Siaα2-3Galβ1-3GalNAcβ1-4(Siaα2-3)Galβ1-4Glcβ1-10 Cer, GD2 GalNAcβ1-4 (Siaα2-8Siaα2-3)Galβ1-4Glcβ1-10 Cer, GD1b Galβ1-3GalNAcβ1-4(Siaα2-8Siaα2-3)Galβ1-4Glcβ1-10 Cer, GT1b Siaα2-3Galβ1-3GalNAcβ1-4(Siaα2-8Siaα2-3)Galβ1-4Glcβ1-10 Cer, GQ1b Siaα2-8Siaα2-3Galβ1-3GalNAcβ1-4(Siaα2-8Siaα2-3)Galβ1-4Glcβ1-10 Cer, neolacto-series gangliosides containing Galβ1-4GlcNAcβ13Galβ1-4Glcβ1-10 Cer, sialylparagloboside Siaα2-3/2-6Galβ1-4GlcNAcβ1-3Galβ1-4Glcβ1-10 Cer, i-active ganglioside Siaα2-3Galβ1-4GlcNAcβ1-3Galβ1-4GlcNAβ1-3Galβ1-4Glcβ1-10 Cer, I-active ganglioside Siaα23Galβ1-4GlcNAcβ1-3(Galα1-3Galβ1-4GlcNAcβ1-6)Galβ1-4GlcNAcβ1-3Galβ1-4Glcβ1-10 Cer, terminal sialylglycoconjugates, sialyl-LacdiNAc, SLDN (Siaα2-6GalNAcβ1-4GlcNAcβ1-), Sda Siaα2-3(GalNAcβ1-4)Galβ14GlcNAcβ1-, sialyllactosamine sialyl-LacNAc, SLN (Siaα2-3/2-6Galβ1-4GlcNAcβ1-), (di)Sialyl-Lex/a sialylLewisa (Siaα2-3Galβ1-3(Fucα1-4)GlcNAcβ1-), sialyl-Lewisx (Siaα2-3Galβ1-4(Fucα1-3)GlcNAcβ1-), or disialyl-Lewisa (Siaα2-3Galβ1-3(Fucα1-4)(Siaα2-6)GlcNAcβ1-), sialyllactose SL (Siaα2-3/2-6Galβ1-4Glcβ1-)

Enveloped and Non-enveloped Sia-Binding Viruses

487

reichenowi [23], but not to infection by pathogens with rapid evolution such as influenza A viruses, which are able to switch binding specificity including avidity to specific Sia species on the host glycan surface by changing a few amino acids in the binding pockets of hemagglutinin (HA) lectins [24]. Besides Neu5Gc, evasion of pathogens by evolutionary suppression of the expression of the other major Sia, Neu5Ac, which is critical in several endogenous functions [20], seems to be unfavorable. Abolishing Neu5Ac production in mice by inactivation of the UDP-GlcNAc 2-epimerase (GNE), a key enzyme of Neu5Ac biosynthesis, results in early embryonic lethality [25]. Mutations of the GNE gene in humans can result in impaired sialyl (Neu5Ac) O-glycan formation in sarcolemmal glycoproteins, a mechanism explaining a muscular disease called distal myopathy with rimmed vacuoles [26]. Several viruses cause important diseases in humans and livestock that affect health and have large social and economic burdens, whereas some viruses are useful for therapeutic applications. Being obligate intracellular parasites, viruses must invade and take over the host cellular machinery for survival and multiplication. To be successful in life, viruses must pass through three important stages, (1) entry, (2) gene expression and genome replication, and (3) exit, as shown in Fig. 2. The entry stage is not only critical for survival and continuation of their life cycle but also contributes to virus transmission and pathogenesis. This stage requires specific interactions between viral capsid proteins of non-enveloped viruses or viral spike glycoproteins of enveloped viruses and specific receptor molecules on the host cell surface (viral attachment). This chapter focuses on Sia-binding viruses. The characteristics of Sia–viral lectin interactions are described and the design and development of viral detection/antiviral drugs/virus-based therapies are discussed.

2

Determination of Sia-Binding Specificities for Viruses The ability of a virus to bind to Sias can be determined via (1) a simple hemagglutination test by incubation of the virus and Sia-rich erythrocytes and observation of erythrocyte agglutination (hemagglutination) by the naked eye. To confirm that Sias or O-acetyl Sias are required for virus binding, (2) a virus-Sia-binding inhibition test can be performed by (2–1) the use of erythrocytes pretreated with an NA enzyme from a bacterium or a virus, which cleaves terminal Sia residues, or the use of Sia-deficient cell lines such as CHO-Lec2 cells. Specific binding of a virus to O-acetyl Sias can be checked by the use of erythrocytes pretreated with O-acetyl esterase from a virus such as bovine coronavirus or influenza C virus, which cleaves O-acetyl groups on Sia residues. (2–2) Hemagglutination inhibition can be performed by using a Sia-containing compound, such as fetuin, as a competitor. To reveal roles of Sia type, linkage

488

Nongluk Sriwilaijaroen and Yasuo Suzuki

Fig. 2 Simplified general scheme of animal and human viral life cycle. The viral life cycle can be divided into three stages: entry, gene expression and genome replication, and exit. (I) The entry stage involves attachment of a virus particle to host cell surface receptors, penetration (by fusion of the viral envelope with the host plasma membrane or with the host endosomal membrane and by permeabilization of the plasma membrane (still under debate [93])/by permeabilization/lysis of the endosomal membrane/by an endoplasmic reticulum (ER)-associated degradation (ERAD) pathway for non-enveloped viruses) and uncoating that releases the viral nucleocapsid into the cytosol. (II) The next stage is viral gene expression and genome replication. DNA viruses, except for poxviruses with a large genome having their own RNA and DNA polymerases, transcribe and replicate their genomes within the nucleus by using host polymerases. In contrast, RNA viruses, which have their own enzymes for transcription and replication, transcribe and replicate their genomes within the cytoplasm. Except for orthomyxoviruses, which need the host 50 cap for viral mRNA synthesis, and hepatitis D virus (HDV), which lacks both enzymes for transcription and replication, they transcribe and replicate their genomes within the nucleus. (III) Finally, the exit step occurs after viral components are assembled to form progeny viruses and ultimately released from the cell. The enveloped membrane of enveloped viruses is derived from budding, either through an internal compartment followed by exit via a secretory pathway or through the plasma membrane giving direct exit. For non-enveloped viruses, the primary mode of exit is cell lysis, and for nonlytic viral spread, the non-enveloped viruses are released from cells via extracellular vesicles [193, 194]

type, sequence, compositions, chain lengths, and architectures in binding specificity of viruses/viral lectins, (3) direct binding of viruses/viral lectins to a variety of defined structures of sialyl glycopolymers coated on microplates or printed on microarray slides can be determined by quantitative detection. The detection can be

Enveloped and Non-enveloped Sia-Binding Viruses

489

performed by several methods such as measurement of the intensity of fluorescence products from viral NA enzyme activity [27] and measurement of fluorescently labeled antiviral/antiviral lectin antibodies or fluorescently labeled viruses/viral lectins [28] that correlate with the number of glycan-bound viruses/viral lectins. To reveal viral lectin–sialyl glycan interactions and conformational diversity of sialyl ligands bound to the viral lectins, (4) X-ray crystallography is now the most widely used technique for determining the atomic arrangement in a three-dimensional space of a co-crystal obtained after evaporation of viral lectin and sialyl glycan in an appropriate solvent.

3 Sialylglycoconjugates as Host Receptor Determinants of Viral Spike Glycoproteins for Enveloped Viral Infection For enveloped viruses (Table 1), only single-stranded (ss) RNA viruses have so far been shown to encode spike glycoproteins, which function as lectins that read codes of sialyl sugar chains during infection. These ssRNA enveloped viruses belong to the families of Orthomyxoviridae, Paramyxoviridae, and Coronaviridae. 3.1 Orthomyxoviridae

The role of the hemagglutination activity of tick-borne Thogoto virus remains to be determined [29]. Only five genera in the family Orthomyxoviridae have been confirmed to internalize into the host cell via endocytosis upon binding of their glycoproteins to the target Sia receptors. Each genus contains only one species. Infectious salmon anemia (ISA) virus (ISAV) in the genus Isavirus causes a systemic and lethal disease in farmed and wild salmonids, whereas influenza A, B, C, and D viruses in the genera Influenzavirus A, B, C, and D, respectively, cause influenza, a contagious respiratory disease, with differences in severity and host range as shown in Table 1. ISAV, which separately encodes hemagglutinin-esterase (HE) and fusion (F) glycoproteins separately, initiates viral attachment to terminal 4-O-acetyl-sialyl glycans via HE-F complexes. This attachment mediates dissociation of HE and F, and the virus is endocytosed into the host cell [30]. Influenza C and D viruses encode hemagglutinin-esterase-fusion (HEF) glycoproteins responsible for receptor binding, receptor destroying, and membrane fusion [31]. Influenza A and B viruses have hemagglutinin (HA) glycoproteins mediating receptor binding and membrane fusion and have separate neuraminidase (NA) glycoproteins possessing a receptor-destroying function [32, 33]. While HEF glycoproteins of influenza C and D viruses attach to 9-O-acetyl-Neu5Accarrying sugar chains found in the respiratory tract of animals [31] as a receptor determinant for infection in cattle and pigs (only C virus has been detected in humans), HAs of influenza A (Fig. 3a) and B viruses recognize α2-6Neu5Ac-carrying sugar chains and

Genome

Virus (Acronym) Disease

Viral lectin

Cellular receptor

Segmented Influenza B 8 () virus (IBV) ssRNAs

Influenza

Target site

Entry pathway

Ref.

In general, primarily Receptor[24, 32, in the respiratory medicated 35–45] system. endocytosis Human infection: and in uncomplicated acidificationcases, dominant induced in the fusion with tracheobronchial the tree; in fetal endosomal cases, found in membrane the lung parenchyma Avian viruses infect the respiratory and intestinal tracts of avian species. Some avian viruses may infect human ocular cells HPIV: multiple systemic organs Endocytosis [33–35, 40] Mainly in HA, NA α2-6Sia > α2-3Sia Respiratory tract for wild type and fusion humans (two lineages; viruses: with the (epidemics), B/Yamagata B/Yamagata / endosomal occasionally /16/88-like 73 (HA with membrane found in and B/ Phe95 and a seals and Victoria/2/ glycosylation at pigs 87-like Asn194) and viruses) B/Victoria/ 504/2000 (HA with Gly141, Arg162, and Asp196)

Host

Orthomyxoviridae Segmented Influenza Avian/swine/human A variety of HA (H1–H16 Avian-type 8 () A virus (IAV) influenza. animals, and subtypes), receptor: α2ssRNAs Zoonotic infections in humans NA (N1–N9 3Sia Humanhumans can cause mild (outbreaks, subtypes) type receptor: upper respiratory epidemics HPIV: so far, α2-6Sia For infection to severe and some avian long-term pneumonia and even pandemics) H5 and H7 circulating death. In addition, viruses viruses in infections may cause possessing humans, they conjunctivitis, the cleavage prefer α2gastrointestinal site cleaved 6Neu5Ac long symptoms, encephalitis by host LacNAc and encephalopathy. ubiquitous repeating units Highly pathogenic influenza proteases HPIV: avian-type virus (HPIV): severe receptor systemic disease

RNA genome

Family

Table 1 List of Sia-binding enveloped viruses, host range, viral lectin, receptor specificity and tissue/cellular tropism

Mild influenza

Segmented Influenza D 7 () virus (IDV) ssRNAs

Paramyxoviridae Linear () ssRNA

Farmed and wild salmonids

Pigs and mainly in cattle

Mainly in infants and children, pigs, cattle

Human hPIV-1: Humans parainfluenza laryngotracheobronchitis virus types (croup); 1, 2 and 3 hPIV-2: croup; (hPIV-1, hPIV-3: bronchiolitis hPIV-2 and pneumonia. and hPIV-3)

Segmented Infectious Fatal systemic infections in 8 () salmon salmon ssRNAs anemia virus (ISAV)

Mild influenza

Segmented Influenza C 7 () virus (ICV) ssRNAs

HN

HE

HEF

HEF

hPIV-1: respiratory Direct fusion tract; hPIV-2: on the host respiratory tract; plasma hPIV-3: lower membrane respiratory tract

hPIV-1: α23Neu5Ac branched LacNAc Site I: sialyl-Lex Site II: sulfated sialyl-Lex, α28Sia hPIV-2: α23Neu5Ac, α26Neu5Ac hPIV-3: α23Neu5Ac, α26Neu5Ac, α23Neu5Gc Site I (receptorbinding and -destroying activities): Neu5Ac Site II (receptorbinding activity): Neu5Ac

4-O-acetyl sialic acid

9-O-acetyl sialic acid

Endocytosis and fusion with the endosomal membrane Respiratory tract Endocytosis and fusion with the endosomal membrane Endothelial cells Endocytosis lining blood and fusion vessels and with the macrophage-like endosomal cells membrane

Respiratory tract

9-O-acetyl sialic acid

(continued)

[48–50, 57, 58]

[30]

[31]

[31]

Family

Table 1 (continued) Host

Mainly in domestic poultry

Linear () Newcastle ssRNA disease virus (NDV)

Avian pneumoencephalitis: mild conjunctivitis and influenza-like symptoms in humans

Primarily infect mice and laboratory animals

Linear () Sendai virus Clinical or subclinical in ssRNA (hemagglutin mice, asymptomatic in ating other animals virus of Japan , HVJ)

Primarily epidemic parotitis, Humans less frequently, orchitis, meningitis and pneumonia, and rarely unilateral nerve deafness, myocarditis, pancreatitis and nephritis

Virus (Acronym) Disease

Linear () Mumps virus ssRNA

Genome

HN

HN

HN

Viral lectin

Sialoglycoprotein (erythroglycan II), I-active and i-active gangliosides, Neu5Acα23Galβ13GalNAc Direct fusion α2-3Neu5Ac/ In poultry: respiratory tract, on the host Neu5Gcnervous system plasma paragloboside and membrane gangliosides gastrointestinal containing tract; linear lactoIn humans: series conjuntiva and type 2 respiratory tract oligosaccharide, Ganglioside GM3 (NeuAc or NeuGc), Site I: α2-3>α2-6Sialyllactose Site II: α2-3/α2-6Sialyllactose (α2-3 4,9-di-O-acetylSia Usually cause common cold, Humans and Domain A of S proteins of Human airway Endocytosis [65–67, 75] pneumonia, bronchitis; animals spike BCoV/HCoV: epithelial cells; and fusion respiratory and enteric (S) protein, 9-O-acetyl Sia In animals, with the disease in animals HE HE proteins: epithelial cells endosomal (a shorter Equine/bovine/ lining the membrane spike-like human CoV, respiratory protein) murine CoV and/or intestinal DVIM: 9-Otracts acetyl Sia Rodent (rat) CoV:4-O-acetyl Sia Bovine CoV:7,9di-O-acetyl Sia

Torovirus: Gastroenteritis four genotypes (1) equine torovirus (EqToV), (2) bovine torovirus (BToV), (3) porcine torovirus (PToV) (4) human torovirus (HToV)

Linear (+) ssRNA

Virus (Acronym) Disease

Genus Severe acute respiratory Betacorona illness, including fever, virus clade C, cough, and shortness of Middle East breath respiratory syndrome coronavirus (MERSCoV)

Genome

Linear (+) ssRNA

Viral lectin B

Cellular receptor

Target site

S1 : DPP4 Respiratory tract Bats (possible Domain (CD26) ¼ a original A of S1 primary receptor sources), Subunit S1A: α2-3Sia dromedary (S1A) of spike (S) glycans ¼ an camels glycoproteins attachment (major factor reservoirs), humans

Host

Ref. Mainly plasma [68, 70] membrane fusion

Entry pathway

HA hemagglutinin, NA neuraminidase, HEF hemagglutinin-esterase-fusion, HE hemagglutinin-esterase, HN hemagglutinin-neuraminidase, DVIM diarrhoea virus of infant mice, DPP4 dipeptidyl peptidase 4, CD26 cluster of differentiation 26

Family

Table 1 (continued)

Enveloped and Non-enveloped Sia-Binding Viruses

495

Fig. 3 Binding of viral lectins to sialoglycans on the outer surface of host cell membranes. A schematic cross section through each virus particle with a diameter in nm is shown to locate viral lectins. The viral lectins of human influenza H3 HA trimer (a), mumps HN tetramer (b), porcine torovirus HE dimer (c),

496

Nongluk Sriwilaijaroen and Yasuo Suzuki

Fig. 3 (continued) CVA24v VP1 monomer (d), reovirus T1L σ1 trimer (e), reovirus T3D σ1 trimer (f), HAdV-D37 fiber trimer (g), AAV1 VP3 monomer (h),

Enveloped and Non-enveloped Sia-Binding Viruses

497

Fig. 3 (continued) and SV40 VP1 pentamer (i) are shown as surface diagrams using PYMOL, with monomers colored in salmon, sky blue, lime, light orange, and warm pink. The viral lectins are extended from the viral envelope (green) or viral capsid (pink) directly or via the pentameric λ2 protein for reoviruses and via the pentameric penton base for adenoviruses. The head/knob of each monomer of each viral lectin (except for T3D

498

Nongluk Sriwilaijaroen and Yasuo Suzuki

cause epidemics in humans as shown in Table 1 [24, 33, 34]. However, only influenza A viruses have a wide host range of animals and a variety of subtypes due to high mutation rates from both reassortment between their genome segments and point mutations that allow them to quickly adapt to a new environment [35, 36]. H1– H16 and N1–N9 subtypes have been reported to use specific Sia receptors as key determinants for host/tissue tropism and specificity. Avian influenza A viruses primarily recognize α2-3Sia, which is mainly found on bird intestinal epithelial cells [37] and embryonated chicken egg chorioallantoic cells [38]. Human-adapted (pandemic and seasonal) influenza A viruses can be detected in nasal or nasopharyngeal samples, a throat swab, and tracheal and bronchial aspirates [39–41], in which α2-6Neu5Ac is dominant [42]. Infection in humans with a nonhuman influenza A virus that has crossed the species barrier via or not via an intermediate host can lead to a pandemic. So far four pandemics have been reported. While the H1N1/1918 pandemic (pdm) has an unknown origin, biochemical and evolutional studies have indicated that the HA genes of H2N2/1957, H3N2/1968, and H1N1/2009 pandemics were derived from nonhuman viruses, providing major viral antigens that are new to humans [24, 32]. The nonhuman avian (av)H2 to pdmH2/1957 HA and the avH3 to pdmH3/1968 HA changed their binding preference from α2-3Sia to α2-6Sia receptors, which are predominant in most regions of the human respiratory system such as the tracheal epithelium [42] except for the human alveoli region, in which α2-3Neu5Ac receptors are dominant [24], for efficient transmission among humans. Classical swH1 to pdmH1/ 2009 HA has decreased binding to Neu5Gc, which is found in the porcine respiratory tract [43] but is rare in healthy human adult hosts [21]. While nonhuman-to-pandemic HA lectins acquire a Sia type/linkage shift in binding specificity, pandemic-to-seasonal ä Fig. 3 (continued) σ1 containing a Sia-binding site in its body) interacts with a sialoglycan on a glycoprotein/ glycolipid (a brown dash), which is anchored to the host cell membrane. Details of interactions between viral amino acids in the receptor-binding site (RBS) and a sialoglycan receptor are zoomed in on one monomer. Sugar residues are shown as sticks with red color for Sia, yellow for Gal and GalNAc, and blue for GlcNAc and Glc according to sugar symbol colors in Fig. 1 (except for sugar residues of α2-3-sialyllactose in a mumps HN pocket that are in red since each residue cannot be colored independently). Only amino acids in RBS that are found to have direct contact with sugar residues are shown in lines with colors according to elements and are summarized on the upper part of each viral lectin-sialoglycan complex. Bond length in angstroms between two bonded atoms is on the dashed lines. PDB accession no. of each complex used for analysis is indicated in red text and details are provided in Table 3. Abbreviations in each schematic of a virus particle are M for matrix protein; NEP for nuclear export protein; NA for neuraminidase; HA for hemagglutinin; HN for hemagglutininneuraminidase; F for fusion protein; E and M for viruses in the family Coronaviridae for envelope protein and membrane protein, respectively; S for spike protein; and HE for hemagglutinin-esterase. Each monomer of PToV HE consists of 3 domains, MP, E, and R, that stand for a membrane-proximal domain, an esterase domain, and a receptor-binding domain, respectively. Residues are abbreviated as “res.” VP and CP stand for viral protein and capsid protein, respectively

Enveloped and Non-enveloped Sia-Binding Viruses

499

long-term human-adapted HAs drift binding preference to long α2-6Neu5Ac-polyLacNAc glycans. The presence of up to 10 LacNAc units in human respiratory sialyl N-glycans [44] but only 2 LacNAc units in a Neu5Ac(LacNAc)2 structure responsible for 0.15% of total human alveolar N-glycans might be another factor explaining why seasonal viruses are rarely found in human alveoli [24]. While little is known about ISAV including its potential to spark a pandemic, extensive studies on influenza A viruses [45–47] have indicated the continual threat of nonhuman influenza A viruses, especially avian and swine influenza viruses, to human infections highlighting the need for surveillance of changes in receptor binding HA amino acids and receptor binding preference to the α2-6Neu5Ac human-type receptor. 3.2

Paramyxoviridae

In the family Paramyxoviridae, human parainfluenza virus type 1 (hPIV-1), type 2 (hPIV-2), and type 3 (hPIV-3) [48–50], mumps virus [51], Sendai virus [52–54], and Newcastle disease virus (NDV) [52, 55, 56] have been shown to be Sia-binding viruses. Unlike the viruses in Orthomyxoviridae that fuse their envelope with the host endosomal membrane, these viruses have a non-segmented ()ssRNA genome that is released into the host cytosol via direct fusion on the host plasma membrane [57]. These viruses have undissociated binding- and destroying-receptor molecules (hemagglutinin-neuraminidase, HN) that trigger fusion activity of F proteins upon binding to sialic acid-containing receptors. HNs of hPIV-1 and hPIV-3 have been demonstrated to carry two Sia-binding sites. Results obtained by computer modeling analysis and glycan array assays suggested that binding-catalytic site I of hPIV-1 prefers α2-3-sialyl-Lex and binding site II of hPIV-1 prefers a sulfated α2-3-sialyl-Lex and α2-8Sia [50], in agreement with the results of an earlier study showing that hPIV1 preferentially binds to α2-3Neu5Ac linked to branched LacNAc [48]. Crystallographic studies on the hPIV-3 HN complexed with Neu5Ac revealed that site I with both receptor binding and NA activities is located on the globular head HN, whereas site II with receptor binding activity is near the hPIV-3 HN dimer interface and is thought to be involved in promotion of virus fusion [58]. Direct receptor binding assays showed that hPIV-3 can bind to α2-3Neu5Ac, α2-6Neu5Ac, and α2-3Neu5Gc receptors [48], suggesting flexibility of the binding pockets of hPIV-3 HN. Similar to hPIV-3 binding preference, hPIV-2 was shown to bind and cleave both α2-3 and α2-6Neu5Ac ligands on a glycan array [49]. However, hPIV-1 (often) and hPIV-2 (less frequently detected) cause laryngotracheobronchitis (croup) in children, whereas hPIV-3 causes pneumonia and bronchiolitis in infants [48]. Studies on receptor binding and pathological observations have suggested that these viruses might target different cell types in the same and different regions of the human respiratory system. Thus, further analysis of virus cell type binding specificity and

500

Nongluk Sriwilaijaroen and Yasuo Suzuki

sialoglycoconjugates in the human respiratory system for each age group is needed. Unlike the targets of the above-described respiratory viruses, mumps virus primarily infects the parotid and other salivary glands and sometimes spreads to other tissues and organs including the pancreas, testis, ovary, mammary glands, and kidney. Co-crystal structural analysis (Fig. 3b) and glycan-binding assays revealed that mumps virus HN proteins prefer α2-3Sia linked to unbranched sugar chains [51]. Uncovering the glycan expression profile of infected tissues should expand our understanding of this virus tropism and control. Due to similarities to hPIV-1 in sequence, structure, and antigenicity, Sendai virus is a so-called murine parainfluenza virus that primarily infects the respiratory tract of mice and laboratory animals including rats, hamsters, and guinea pigs and occasionally pigs. Since it can agglutinate red blood cells, it was previously known as hemagglutinating virus of Japan (HVJ). It has been shown to bind to linear and branched sialosylpolylactosamine (sialyl-poly-LacNAc) sequences in erythroglycan II on erythrocyte membranes [53] and gangliosides including I-active and i-active gangliosides with terminal Siaα2-3Galβ1-4GlcNAc (Fig. 1) coated on asialoerythrocytes [52]. Resistance of V. cholerae sialidase-treated cells to Sendai virus infection can be restored fully by re-sialylation of the cells with CMP-Sia and β-galactoside α2-3-sialytransferase but not with β-galactoside α2-6-sialytransferase. This indicated that NeuAcα2-3Galβ1-3GalNAc, but not NeuAcα2-6Galβ1-4GlcNAc, is a receptor determinant of Sendai virus infection [54]. More research works to clarify binding preferences and interactions of Sendai virus and sialyl glycan receptors should lead to the generation of a recombinant Sendai virus that is specific to cell types and can improve the efficacy and specificity of virus entry and thus safety for therapeutic approaches such as virus-based vaccines against human respiratory viruses [59], virus vectors to reprogram cell genomes for regenerative medicine, and oncolytic virotherapy against cancer [60]. NDV is a minor zoonosis. Exposure to a large amount of the virus is necessary for human infection, which typically causes mild conjunctivitis and/or influenza-like symptoms. NDV mainly infects domestic poultry and causes a range of diseases from nonapparent to severe respiratory/gastrointestinal/nervous system diseases depending on the virus strain. With a neurovirulent strain infection, infected poultry would develop pneumonitis followed by encephalitis, leading to severe economic effects on poultry production; hence, NDV is sometimes called avian pneumoencephalitis [61, 62]. Based on laboratory tests of their virulence (mean death time) in chicken embryos after allantoic inoculation, NDVs are divided into three groups: velogenic (most virulent), mesogenic (mid-virulent), and lentogenic (non-virulent). While velogenic

Enveloped and Non-enveloped Sia-Binding Viruses

501

strains cause economic loss, lentogenic strains are used for vaccination of chickens and for oncolytic virotherapy [62]. Understanding NDV attachment to its receptors and the viral entry mechanism is critical to find a way to control virus infection: inhibition of chicken infection by velogenic strains but enhancement of entry of a lentogenic strain into tumor cells. An early study showed that NDV prefers α2-3Sia with either Neu5Ac or Neu5Gc-paragloboside gangliosides containing linear lacto-series type 2 oligosaccharide and ganglioside GM3 with either NeuAc or NeuGc [52]. Recent advances in the Sia-protein co-crystal technique [55] and hemadsorption/fusion deficiency of dimer interface mutants [63] have revealed that NDV HN carries two Sia-binding active sites. Site I possesses receptor binding mediating NA activity and triggers HN interaction with the F protein, while site II has only receptor binding activity at the dimer interface that has been proposed to maintain virus-host Sia attachment during fusion. Molecular modeling of NDV HN with small molecules and analysis of the binding free energy of interactions [56] indicated that site I prefers α2-3sialyllactose. Site II can interact with α2-3-sialyllactose similar to α2-6-sialyllactose, but the interaction is much weaker than the interaction of site I with α2-3-sialyllactose. It is interesting that binding of NDV to red blood cells was inhibited by α2-3-sialyllactose but was increased in the presence of zanamivir (an NA inhibitor) preferential to bind to site I [56]. This raises important questions about whether NA activity is required for NDV NH binding and/or fusion and whether site II works independently or must be activated by site I binding to some compound such as zanamivir but not α2-3-sialyllactose. These questions should be answered in order to efficiently control the functions of these two sites. 3.3

Coronaviridae

In the family Coronaviridae, non-segmented (+)ssRNA torovirus [64], some coronaviruses in the genus Betacoronavirus clade A group 2 [65–67] and clade C [68], and transmissible gastroenteritis coronavirus (TGEV) in the genus Alphacoronavirus [69] (TGEV will be discussed in the family Parvoviridae) have been reported to bind to Sia. These viruses use spike (S) glycoprotein as a multifunctional molecule: its S1 subunit binds to a receptor, and its S2 subunit contains the fusion peptide (FP) that fuses viral and host membranes after activation by proteolytic cleavage. Fusion of these viruses occurs at the endosomal membrane, but Middle East respiratory syndrome coronavirus (MERS-CoV) mainly fuses its viral membrane at the plasma membrane since its S proteins are sensitive to be activated by both the pH-dependent endosomal protease and host secreted or surface proteases [70]. Toroviruses cause gastroenteritis in vertebrates, mainly cattle, pigs, horses, and humans, especially children. Based on comparative sequence analysis, four genotypes can be classified into (1) equine torovirus (Berne virus) (EqToV), (2) bovine torovirus (Breda virus)

502

Nongluk Sriwilaijaroen and Yasuo Suzuki

(BToV), (3) porcine torovirus (Markelo virus) (PToV), and (4) human torovirus (HToV) [71]. Little is known about the receptor binding specificity of torovirus S proteins. There is evidence that EqToV, who lacks an intact hemagglutinin-esterase (HE) gene, can agglutinate human, rabbit, and guinea-pig erythrocytes, and these hemagglutinations were decreased in the presence of fetuin and gangliosides. This evidence suggested that glycoproteins/glycolipids may be receptors of torovirus S proteins [72]. Coronaviruses can cause a variety of illnesses with varying severity and they have a wide range of hosts. Their infection is mainly associated with respiratory diseases, especially common colds in humans, and infection often leads to diarrhea in animals. Coronaviruses have sparked global concern about their pandemic potentials since zoonotic transmission, most likely from bats, to humans of severe acute respiratory syndrome SARS-CoV (not a Sia-binding virus) via civet cats, an intermediate host, which emerged first in Asia in 2002–2003, and of severe viral respiratory MERS-CoV via camels as an intermediate host, which emerged first in a middle eastern country in 2012 [73]. Only some coronaviruses use Sia as a receptor/assistant receptor/attachment factor as shown in Table 1. The S proteins of bovine coronavirus (BCoV) [74] and human coronavirus (HCoV) OC43 and HKU1 [75] in the genus Betacoronavirus clade (or lineage) A group 2 have been reported to bind to 9-O-acetylated Sia in the receptor-binding site. Typically, these betacoronaviruses also carry HE spike proteins that bind to O-acetylated Sias as well. Why are HE proteins present in some but not all toroviruses and coronaviruses? Results of evolution studies have suggested that intact HE genes present in three genera of toroviruses (BToV, PToV, and HToV) and Betacoronavirus clade A group 2 with S proteins binding to O-acetylated sialic acids [75] were acquired via independent heterologous RNA recombination events from a yet unknown source, possibly during a mixed infection with another virus, such as influenza C virus [76, 77]. Variants with the HE gene, which have become circulating strains, could be explained by the finding that the HE protein increases the efficiency of production of infectious virus [78] possibly by acting as a lectin for enhancing viral binding and as an enzyme that destroys receptors by de-O-acetylation (esterase) for enhancing release of trapped virions from the host mucosa and of budding virions from infected cells. However, it should be clarified how S proteins and HE proteins act cooperatively for supporting virus infection, and it should be clarified whether the virus can still infect cells if receptor binding at one site is inhibited. An understanding of the mechanism of initiation of infection should lead to the development of a potential strategy to combat virus infection. HE proteins of viruses from different hosts have distinct receptor/substrate preferences. Esterases of porcine/bovine toroviruses and equine/bovine/human coronaviruses are specific for 5-N-

Enveloped and Non-enveloped Sia-Binding Viruses

503

acetyl-9-O-acetylneuraminic acid (Neu5,9Ac2) [64, 77]. An esterase-deficient PToV HE (receptor-binding domain) can bind to 4,9-di-O-acetyl-Neu5Ac- [64] as shown in Fig. 3c. Esterases of bovine toroviruses prefer the di-O-acetylated substrate (5-N-acetyl-7(8),9-di-O-acetylneuraminic acid) [77]. Bovine coronavirus HE also binds to 7,9-di-O-acetyl Sia [67]. Most rodent coronaviruses express sialate-4-O-acetylesterases and the rat coronavirus HE protein binds to 4-O-acetyl Sia. The HE proteins of murine coronavirus DVIM (diarrhea virus of infant mice) cleave 9-O-acetylated Sias [64, 67, 77]. Some viral HE proteins may have more flexible pockets that would allow for a greater range of receptors/ substrates to be in their pockets. The HE proteins of human toroviruses were shown to have 74% sequence identity with those of bovine toroviruses [79], but there is a lack of information regarding receptor/substrate preferences. These receptor-binding and substrate-cleaving specificities of HE proteins in a host indicated that if the virus jumps to infect different animal species, its HE proteins need to acquire mutations to bind/cleave O-acetylated Sias expressed on the host target. MERS-CoV in the genus Betacoronavirus clade C binds to dipeptidyl peptidase 4 (DPP4 and also called CD26) as a primary receptor via domain B of S1 Subunit (S1B) [80] and binds to α23Sia glycans as an attachment factor via domain A of S1 Subunit (S1A) of its spike (S) protein designated S1A through S1D from the N-terminal S protein [68]. These bindings trigger infection in nonciliated bronchial epithelial cells and type II pneumocytes [73] that mainly have α2-3Sia receptors [81]. Reduction in MERS-CoV infection of sialidase-treated Calu-3 human airway cells indicated that the host Sia receptor is involved in determining MERS-CoV host/tissue tropism and transmission [68]. It remains unclear how this virus emerged to infect humans. However, MERS cases have continued to be reported [82], indicating the possibility of a pandemic in the future. This highlights the need for continuous surveillance of changes in MERS-CoV including two receptorbinding preferences and finding a way to control MERS.

4 Sialylglycoconjugates as Host Receptor Determinants of Viral Capsid Proteins for Non-enveloped Viral Infection Various non-enveloped viruses with either an RNA or DNA genome have been reported to encode lectins for binding to sialyl glycans, which mediate entry into the host cells (Table 2). 4.1

Picornaviridae

Several strains listed in Table 2 of small (+)ssRNA-containing viruses, pico-rna-viruses, in the family Picornaviridae, including equine rhinitis A virus (ERAV) in the genus Aphthovirus;

Genome

Disease

Enterovirus Acute hemorrhagic Humans 70 (EV70), such as conjunctivitis (AHC) strain J 670/71 and rarely with neurological sequelae Coxsackievirus A24 Ocular (AHC) and Humans variant (CVA24v) respiratory diseases such as 110386110392 strains

Linear (+) ssRNA

Linear (+) ssRNA

Viral lectin

α2-3-Sialyllactose

Cellular receptor

Target site

Entry pathway

Respiratory tract Endocytosis: acidic pH-mediated uncoating and capsid conformational change VP1 and VP3 α2-6 and α2-3 Spreads locally in ReceptorSialyllactosamine the upper mediated and sialyllactose respiratory endocytosis and tract before low spreading pH-activated throughout protease/viral the body capsid including conformational motor nerve change cells resulting in disruption of the endosomal membrane Not α2-3Neu5Ac OOcular cells; Endocytosis determined glycosylated, conjunctival nonGPI-anchor and corneal glycoproteins on cells corneal cells. CNS Ocular cells; Endocytosis VP1 Essential receptor: corneal cells ICAM-1 interacts and with VP1, VP2 and conjunctival VP3. cells Attachment receptor: Respiratory tract Sia α2-6 and α23Neu5Ac O-linked to corneal cell surface proteins but not to lipids. Binding studies in silico DSLNT >60 SL>30 SL and 60 SL>LSTc

VP1

RNA genome

Humans

Horses

Host

Enterovirus D68 Mild to severe (EV-D68; formerly respiratory illness; named rhinovirus possibly cause a 87) such as strains polio-like disorder Fermon, 670, 2042 called acute flaccid and 2284 myelitis (AFM)

Equine rhinitis A virus Upper (cold-like (ERAV); symptoms) and Sia-binding residues lower respiratory are conserved across tract diseases all ERAV strains

Virus (Acronym)

Linear (+) ssRNA

Picornaviridae Linear (+) ssRNA

Family

Table 2 List of Sia-binding non-enveloped viruses, host range, viral lectin, receptor specificity and tissue/cellular tropism

[86, 90, 96, 98]

[85, 93, 97]

[84, 89, 95]

[83]

Ref.

Rhesus macaques

Wild rodents, laboratory mice

Humans

Upper respiratory tract Cats disease; oral ulceration, acute arthritis, jaundice and death

Feline calicivirus (FCV) (strain F9)

Linear (+) ssRNA

Acute gastroenteritis

Linear (+) ssRNA

Human noroviruses, genogroup GII.3 (strain Chron1), GII.4 (Dijon and MI001)

Murine noroviruses, No/mild GV (MNV1, WU11 gastroenteritis in and S99) immunocompetent mice but lethal in innate immunedefective mice Rhesus macaques GI.1 Gastroenteritis Tulane virus

Linear (+) ssRNA

Caliciviridae Linear (+) ssRNA

Low-neurovirulent Mild CNS damage; Mice (natural Theiler’s murine disseminated hosts), rats encephalomyelitis encephalomyelitis virus (TMEV) (a chronic persistent (Theiler’s original demyelinating (TO) subgroup: infection of the white such as strains DA, matter) BeAn)

Linear (+) ssRNA

Many animals: pigs (the most common and severe), rodents (natural hosts), humans

Encephalomyocarditis Myocarditis, virus (EMCV) such encephalitis, as wild-type neurological mengovirus, K2 diseases, strain, BHK cellreproductive adapted 1086C disorders and strain (group I) diabetes with Lys231 in VP1, BRL celladapted 1086C strain with Lys49Glu, Leu142Phe and Ile180Ala in VP1

Linear (+) ssRNA

[104, 108]

[102, 103]

[91, 150]

[87, 92, 99–101]

(continued)

Not determined [105] Not A type 3 and B histo- Intestinal epithelial cells determined blood group antigens and Neu5Ac, 60 SLN Not Receptor: fJAM-A Mainly Clathrin-mediated [109] determined Receptor: α2-6Sia Nrespiratory endocytosis linked epithelial cells glycoprotein

Dynamin II and cholesterolmediated endocytosis

Not determined

Mengovirus: Cardiomyocytes, Unclear glycophorin on CNS human The virus can be erythrocytes isolated from K2: sialyl glycans on all tissues by human and equine 2 days after but not bovine infection in erythrocytes; piglets. 70-KD sialoglycoprotein (s) on human nucleated cells BHK/BRL-adapted 1086C: sialylated BHK/BRL cell surface α2-3-Sialyllactose Neurons Endocytosis

Virus-like Leb, H type 1 chain Small intestine and B antigen, and particles sialylated type (VLPs), 2 chain including protruding sialyl-Lex, sialyl(P) dimer of diLex, and 30 SL VP1 glycoconjugates Not GD1a, but not GM1 Intestine, determined and GA1, on the macrophages macrophage and dendritic surface cells

Mainly amino acids in VP2 puff B

VP1

Reoviridae

Family

Table 2 (continued)

Porcine sapovirus (PSaV)

Linear (+) ssRNA

10 Mammalian segmented orthoreovirus dsRNAs (reovirus) serotypes (types) 1, 2 and 3, including the prototypic type 1 strain (clone) Lang (T1L/1953), type 2 strain Jones (T2J/1955) and type 3 strain Dearing (T3D/1955) and other reovirus field isolate strains

Virus (Acronym)

Genome Pigs, humans and other animals

Host

Cellular receptor

All types bind to the same protein receptor σ1: JAM-A Type-specific differences in Sia-binding specificity – Hemagglutination analysis Types 1 and 2: prefer human erythrocytes Type 3: prefer bovine erythrocytes – Crystal analysis T1L σ1: GalNAcβ1-4 (Neu5Acα2-3)(GM2) > α23Neu5Ac (GM3) T3D σ1: α2-3, α2-6 and α2,8-linked Sia (either Neu5Ac or Neu5Gc)

P dimer of VP1 α2-3, α2-6Sia Olinked glycoprotein

Viral lectin

Respiratory disease Mammals σ1 protein (type 3, including most asymptomatic; mice, calves, types 1 and chimpanzees 2, common cold); and humans Enteritis and febrile exanthema in childhood; CNS disease in infants (few); Neonatal biliary atresia or congenital hepatitis (some)

Acute gastroenteritis

Disease

Entry pathway

ClathrinMainly dependent respiratory endocytosis system and intestine. Cells in respiratory system; Types 1 and 3: type 1 alveolar epithelial cells Cells in intestine; Type 1: crypt epithelial cells of the ileum Type 3: goblet and absorptive cells Cells in CNS; Type 1, ependymal cells; Type 2, not known but can be isolated from cerebrospinal fluid of infants with meningitis same as types 1 and 3. Type 3, neurons

Cells in the Clathrin-, intestinal tract dynamin II-, actin- and cholesterolmediated endocytosis

Target site

[111, 113, 114, 116–119]

[107, 110]

Ref.

Adenoviridae Linear dsDNA

Species: Human Corneal inflammation adenovirus D and epidemic Type: 8, 26, 37, 53, keratoconjunctivitis 54 and 64 (formally known as 19a)

11 Rotaviruses Severe infantile segmented (1) Sialidase-sensitive gastroenteritis dsRNAs strains (bind to (diarrheal disease) terminal Sia) Group A with P genotypes 1, 2, 3, 7 and 23. For examples, P1 strain boNCDV P2 strain siSA11 P3 strain huHCR3a, siRRV, caCU-1 P7 strain poOSU, poCRW-8 P23 strain poPRG942 Group C, strain poAmC-1 (2) Sialidase-resistant strains (bind to internal Sia) P4, huKUN, huRV-5 P5, boUK P6, huRV-3, huS12/ 85 P8, huMO, huWa (3) Sialidase-sensitive strains (bind to both terminal and internal Sia) P7 strain poTFR-41

VP8∗ subunit of VP4

Humans

Trimeric fiber knob

DNA genome

Humans and animals

Different strains use different endocytic pathways

Human corneal Corneal Caveolin-1epithelial (HCE) epithelial cells mediated cells: O-linked and endocytosis glycoproteins that conjunctival into human carry glycans cells corneal cells mimic the GD1a glycan  α2-3/ α2-6-sialyllactose

Typically, human strains prefer Neu5Ac. Porcine and bovine strains containing Gly187 in VP8∗ favor Neu5Gc over Neu5Ac. Simian strain containing Lys187 prefers Neu5Ac over Neu5Gc

(1) and (3) Terminal Mature Sia enterocytes P7 poOSU: GM3 and that line the GM2 tips of small P7 poCRW-8: GD1a intestinal villi Bovine P1 and canine P3: α2-6Sia on gangliosides or Nlinked glycoproteins Porcine P7 and P23: α2-6Sia and α23Sia on gangliosides or Nlinked glycoproteins (2) and (3) Internal (branched) Sia All internal Sia-binding viruses: GM1a (¼GM1)

(continued)

[132–139]

[120–129]

Virulent form: Turkeys, hemorrhagic chickens, enteritis in turkeys, pheasants splenomegaly in chickens, marble spleen disease in pheasants Avirulent form: vaccine

Species: Turkey siadenovirus A Type: 3

Linear dsDNA

Adeno-associated virus No disease has been Humans and (AAV) such as associated with AAV animals AAV1, AAV4, infections AAV5, AAV6, bovine AAV (BAAV)

Humans

Gastroenteritis

Species: Human adenovirus G Type: 52

Linear dsDNA

Host

Disease

Virus (Acronym)

Genome

Parvoviridae Linear ssDNA

Family

Table 2 (continued) Cellular receptor

Target site

VP1/VP2/ AAV4: α2-3Sia-OVP3 all glycans contain AAV5: either α2-3 or Sia-binding α2-6Sia N-glycans, motif coreceptor PDGFR AAV1 and AAV6: either α2-3 or α26Sia-N-glycans AAV6: also binds HSPG, coreceptor EGFR BAAV: gangliosides

Endocytosis

Endocytosis

Entry pathway

[143]

[140, 141]

Ref.

A broad range of Clathrin-mediated [158–161] target tissues; endocytosis serotype dependence, AAV4: CNS and retina AAV5: smooth muscle (SM), CNS, lung, retina AAV1: SM, CNS, retina and pancreas AAV6: SM, heart and lung

Human conjunctival (Chang C) cells: 50-kDa membrane protein (CD46) in a calciumdependent manner Above data were obtained by using HAdV-D37 Trimeric short O-linked Gastrointestinal fiber knob glycoproteins tract that carry α2-8linked polySia  Neu5Acα23Galβ14GlcNAcβ1Trimeric fiber α2-3 > α2-6 Intestinal tract knob Sialyllactose and spleen

Viral lectin

Human John Cunningham polyomavirus (JCPyV)

Merkel cell polyomavirus (MCPyV)

Circular dsDNA

Circular dsDNA

Humans

Skin cancer called Merkel cell carcinoma (MCC)

Humans

Severe kidney and brain Humans disorders

Asymptomatic or mild respiratory symptoms. Later, asymptomatic latency in renal urothelium

Tumors in animals Monkeys and Primary brain and bone humans cancers, malignant mesothelioma, and lymphomas in laboratory animals

VP1

VP1

VP1

VP1 Multiple cell Caveolae/rafttypes, found in mediated several human endocytosis tumor/cancer cells, including bone cancer and brain tumor cells GD3, GD2, GD1b Primary infection Caveolaeand GT1b in the mediated gangliosides respiratory endocytosis containing α2-8tract. Alternative linked disialic acid It moves to and pathway: motif persists in the caveolae/ Alternative receptor: kidney and clathrinnon-sialylated urinary system independent GAGs endocytosis via primary human renal proximal tubule epithelial cells Latent in various Clathrin-mediated Receptor: α2-6Sia body organs endocytosis neolacto-series on including the N-linked reno-urinary glycoproteins tract and Enhancement lymphoid receptor: 5-HT2 serotonin receptor tissues, Alternative receptor: possibly latent non-sialylated in CNS or GAGs transport to CNS after reactivation Second receptor: Skin Caveolae/lipid GT1b ganglioside raft-mediated Primary receptor: endocytosis GAGs such as heparan sulfate

GM1 ganglioside Alternative receptor: GAGs when the preferred GM1 receptor is unavailable

[172]

[167, 169, 175]

[166, 171, 175]

[165, 170, 175]

Abbreviations: bo bovine, ca canine, hu human, po porcine, si simian, CNS central nervous system, VCAM-1 vascular cell adhesion molecule 1, GAGs glycosaminoglycans, fJAM-A feline junction adhesion molecules

BK polyomavirus (BKPyV)

Simian virus 40 (SV40)

Circular dsDNA

Polyomaviridae Circular dsDNA

510

Nongluk Sriwilaijaroen and Yasuo Suzuki

enterovirus D68 (EV-D68), enterovirus 70 (EV70), and coxsackievirus A24 variant (CVA24v) in the genus Enterovirus; and encephalomyocarditis virus (EMCV) and low-neurovirulent Theiler’s murine encephalomyelitis virus (TMEV) in the genus Cardiovirus, have been confirmed to use a Sia-containing receptor for host cell attachment. ERAV is a horse pathogen [83]; EV-D68, EV70, and CVA24v are human pathogens [84–86]; EMCV has a wide host range including humans [87]; and low-neurovirulent TMEV is a mouse pathogen that is able to cause a chronic demyelinating disease and is thus used for studies on human multiple sclerosis [88] as shown in Table 2. The viral capsid of the picornaviruses comprises three virion proteins (VP1, VP2, and VP3) that form the shell and one virion protein (VP4) lying on the inner surface of the virus particle as illustrated in Fig. 3d. The crystal structures of the viral capsid proteins, VP1, VP2, VP3, and VP4, in complex with a sialyl ligand have shown that a Sia-binding pocket is in VP1 of ERAV [83], in a pit between VP1 and VP3 of EV-D68 [89], in VP1 of CVA24v [90], and mainly in the VP2 puff B of low-neurovirulent TMEV [88, 91]. Analysis of EMCV mutants/ variants has indicated that amino acids in VP1 are responsible for interactions with Sia on the host cell surface [87, 92]. Typically, binding of the viral capsid proteins of picornaviruses to receptors triggers endocytosis. When the endosome becomes acidic, the viral capsids undergo conformational change and/or a protease is activated, resulting in channel formation that allows the viral genome to pass through the host cytosol [84, 93], except for EMCV and poliovirus for which it remains unclear whether their genomes can directly penetrate through the plasma membrane due to no requirement of low pH for infection [93, 94]. Information on which virus strains require Sia and which sialyl glycan structure is a determinant for viral attachment and infection is critical to understand virus tropism and pathogenesis for diagnosis and treatment, especially the design of a detection system and inhibitors targeting Sia-binding sites in viral lectins. ERAV has been shown to require Sia on the host cell surface for its infection: reduction of the virus infection into host cells was observed when sialidase-treated cells were used or when α2-3-sialyllactose was added to be a competitive receptor for binding to the virus. α2-3 Sialyllactose was found to bind in VP1 of the ERAV capsid by X-ray crystallography, and the residues including Gln65, Ala118, Gln120, and Arg129 that interact with Sia are conserved across all ERAV strains, implying that all strains of ERAV require Sia as a receptor for virus entry [83]. Investigation of the role of Sia in infections by six EV-D68 strains isolated from patients during the period from 2009 to 2010 in comparison with the prototype Fermon strain isolated more than 50 years ago showed that Sia-deficient cells and sialidase-treated cells were resistant to infections by strains 670 (clade A), 2042

Enveloped and Non-enveloped Sia-Binding Viruses

511

(clade B), and 2284 (clade C) as in the case of the Fermon strain [89] but were sensitive to infections by the other three strains, 947 (clade B), 1348 (clade A), and 742 (clade A), and can thus be classified as Sia-dependent and Sia-independent strains, respectively [95]. The use of knockout cell lines and gene reconstitution, glycan array screening, infection inhibition assays by receptor analogues, and co-crystal structure analysis have indicated that both α2-3Neu5Ac and α2-6Neu5Ac on either lactose or lactosamine can be receptors for virus infection [89, 95], possibly explaining why EV-D68 infection causes a wide spectrum of illnesses. Although both EV70 and CVA24v (but not CVA24) are still major causes of acute hemorrhagic conjunctivitis (AHC) epidemics worldwide due to ocular infections of both conjunctival and corneal cells and still have the potential to spark pandemics [96], there are no vaccines or antiviral drugs for AHC diseases caused by EV70 and CVA24v [86]. Knowledge of the receptor-binding specificity of viruses is needed for prevention and treatment of infections. Viral attachment studies using linkage-specific sialidase-treated/sialidase-untreated cells, sialidase-treated followed by α2-3or α2-6-sialyltransferase-treated/α2-6-sialyltransferase-untreated cells, and cells blocked with α2-3Sia- or α2-6Sia-binding lectins suggested that the EV70 prototype strain J670/71 binds specifically to the α2-3-sialyl linkage on the host cell surface. EV70 can bind to phosphatidylinositol-specific phospholipase C-treated cells and tunicamycin-treated cells but not to benzyl N-acetyl-α-D-galactosaminide (benzylGalNAc)-treated cells or to proteinase K-treated corneal cells, suggesting that EV70 prefers binding to α2-3Sia Oglycosylated, non-GPI-anchored glycoproteins on human corneal epithelial (HCE) cells [85, 97]. Viral binding and infection studies using sialidase-, PNGase F-, tunicamycin-, or benzylGalNAc-treated/sialidase-, PNGase F-, tunicamycin-, or benzylGalNAcuntreated cells and binding competition assays in the presence of sialyl-Lex, 30 SLN, or 30 sialyl-TF (Neu5Acα2-3Galβ1-3GalNAcα1, TF, Thomsen-Friedenreich) suggested that CVA24v binds to and infects HCE cells via α2-3Neu5Ac O-linked cell surface proteins [86]. This shared α2-3Neu5Ac-binding preference could be a factor explaining why these viruses have overlapping cellular tropism. Binding to and infection of conjunctival cells by both AHC-causing human picornaviruses, EV70 and CVA24v, must be further investigated. In contrast to EV70, analysis of the crystal structures of CVA24v in complex with a range of sialyl glycans indicated that CVA24v binds strongly to 60 SL and DSLNT (Fig. 3d); binds weakly to LSTc, sialyl-Lex, 30 SL, and 30 SLN (Fig. 1); and does not bind to GM1, GM2, GD1a, GD1b, and GD3, suggesting that CVA24v can bind to both α2-3 and α2-6Neu5Ac with preference for 60 SL over 30 SL and 60 SL over LSTc structures and confirming that it does not bind to gangliosides [90]. The ability of CVA24v to bind well to α2-6Neu5Ac could partially explain why

512

Nongluk Sriwilaijaroen and Yasuo Suzuki

CVA24v more commonly than EV70 infects human upper respiratory tract tissue rich in α2-6Neu5Ac. The terminal α2-6Neu5Ac is also rich in mucin-type O-glycans in tear films that could facilitate the spread of CVA24v [90]. While the emergence of pandemic EV70 in 1971 after its first recognition in 1969 and the second pandemic in 1980 are still mysteries, two pandemics caused by CVA24v in 1985 after the emergence of CVA24v/1970 causing an AHC outbreak and in 2002 have recently been investigated [98]. Both nonvariant and variant CVA24 viruses use ICAM-1 as an essential receptor, but only infection of the AHC-causing CVA24 variants into HCE cells significantly depends on a sialylated cell surface. A comparison of the amino acid sequences in the Sia-binding pockets of nonvariant and variant CVA24 viruses indicated that most nonvariant CVA24 viruses and the first variant CVA24v/1970 virus contain VP1 with Phe250 but that all variant viruses since 1985 including the CVA24/2002 pandemic virus possess Tyr250 in VP1. Binding to and infection of HCE cells by the wild-type CVA24v virus were more efficient than binding and infection of the cells by a constructed Tyr250Phe CVA24v mutant. These findings suggested that the CVA24 ! CVA24v/1970 virus causing AHC is still unknown, that enhanced Sia-binding engagement by a change of Phe250 to Tyr250 in VP1 contributed to the CVA24v/1970 ! CVA24v/1985 pandemic, and that the change responsible for the emergence of the CVA24v/2002 pandemic is still unknown [98]. The detailed sialyl structures that are used for EMCV attachment are still not known. Earlier studies showed that the wild-type mengovirus, but not avirulent mutants, agglutinates human erythrocytes via binding to glycophorin, which is the major sialoglycoprotein on human erythrocytes [99]. The EMCV K2 strain agglutinated human and equine erythrocytes dominant in Neu5Ac and Neu5Gc, respectively, but did not agglutinate bovine erythrocytes dominant in Neu5Gc; hence, this selective agglutination needs to be further investigated [100]. Investigation of an EMCV receptor on permissive human cells revealed that the K2 strain uses a 70-kDa sialoglycoprotein(s) on HeLa and K562 cells as a cell surface receptor for virus attachment [101]. For the rat strain 1086C, the parental virus is not a Sia-binding virus [87]. The viruses, which were cultured in baby hamster kidney-21 (BHK-21) cells, produced two groups of variants. Group I with Lys231 in VP1 that becomes Sia-dependent can bind to and infect primary human cardiomyocytes more efficiently than can group II with Arg231 in VP1, which is not a Sia-dependent group [92]. In addition, the rat strain 1086C appeared to acquire the ability to replicate effectively in buffalo rat liver (BRL) cells as a result of 3 amino acid mutations, Lys49Glu, Leu142Phe, and Ile180Ala, in VP1 that provided the ability to bind to the sialylated BRL cell

Enveloped and Non-enveloped Sia-Binding Viruses

513

surface after 29 passages [87]. These findings indicate that some EMCVs adapt to a new host by acquisition of the ability to bind to Sia on the host cell surface. The cytopathic effect (CPE) caused by infection of low-neurovirulent TMEV strains BeAn and DA in BHK-21 cells was reduced either by the use of sialidase-treated cells or by addition of α2-3-sialyllactose to the medium, indicating the importance of sialylated surface molecules for infection of these persistent TMEV strains [88]. Single amino acid substitutions, Gln2161Ala/Arg/Trp/Phe or Gly2174Trp/Phe, in VP2 puff B of BeAn virus cause loss or reduction of viral attachment to erythrocytes and loss or reduction of viral spread among BHK-21 cells. However, the Gln2161Ala mutant virus appeared to recover to the wild-type Gln2161 virus with potential for binding and infection after prolonged passage in BHK-21 cells, suggesting that this mutant can acquire rapid adaptation in the host environment [91]. 4.2

Caliciviridae

Members of the family Caliciviridae with cuplike depressions in the viral surface, which have been shown to be Sia-binding viruses, include human enteric GII.3 (Chron1), GII.4 (Dijon) [102], and GII.4 (MI001) [103] and murine enteric GV (MNV1, WU11, and S99) [104] noroviruses in the genus Norovirus (Norwalk-like viruses), rhesus macaques enteric GI.1 Tulane virus (TV) in the genus Recovirus [105], cat respiratory GI (F9) feline calicivirus (FCV) in the genus Vesivirus [106], and porcine enteric GIII (Cowden) sapovirus (PSaV) in the genus Sapovirus (Sapporo-like viruses) [107]. The capsid of these non-enveloped viruses is comprised of 90 U of a single major capsid protein (VP1) in a dimeric form. Each VP1 monomer has 2 domains: a shell (S) domain and a protrusion (P) domain. Binding of the P domains to host receptors triggers receptor-mediated endocytosis that depends on clathrin, dynamin II, and/or cholesterol [108–110] as shown in Table 2, and the endocytosis is followed by penetration of the viral genome into the host cytosol for multiplication. The use of neoglycoproteins for binding studies has revealed that virus-like particles (VLPs) of Chron1 (GII.3), Dijon (GII.4), and Norwalk (GI.1) strains bind to Leb and H type 1 chain glycoconjugates. Chron1 and Dijon strains also bind to sialylated type 2 chain glycoconjugates including sialyl-Lex and sialyl-diLex but not to Lex or sialyl-Lea, indicating virus binding specificity [102]. Later analysis of binding stoichiometry by native mass spectrometry provided evidences that there are four B antigens or two α2-3-sialyllactoses (GM3 trisaccharides) forming a complex with a recombinant P domain dimer from the GII.4 MI001 variant. Epitope mapping showed direct interaction of α2-3Sia with the P domain [103].

514

Nongluk Sriwilaijaroen and Yasuo Suzuki

Binding to and infection of primary murine macrophages by a murine norovirus MNV1 strain were reduced in the presence of either Sambucus nigra lectin (SNL), which is preferential to α26Sia, or Maackia amurensis lectin (MAL), which is preferential to α2-3Sia, or when sialidase-treated macrophages were used. An enzyme-linked immunosorbent assay showed that MNV-1 bound to the ganglioside GD1a but not to GM1 or asialo-GM1 (GA1). In addition, reduction of MNV-1 binding to and infection of murine macrophages by the depletion of gangliosides in primary murine macrophages can be restored by the addition of GD1a. Similarly, a role of GD1a was also observed during binding to and infection of the macrophages with murine norovirus strains WU11 and S99 [104]. In addition to interactions with histo-blood group antigens (HBGAs), Tulane virus has been shown to bind to synthetic sialoglycoconjugates: strongly to Neu5Ac and weakly to α2-6SLN but not to Neu5Gc, α2-3SL, or a type A disaccharide (GalNAc-Gal). Tulane virus infection in permissive LLC-MK2 cells was significantly reduced by treatment of the host cells with either NA or α2-6Sia-binding SNL [105]. However, further investigation is needed to determine how HBGAs and sialoglycoconjugates coordinate with each other to attach to Tulane virus for mediating infection. In addition to feline junctional adhesion molecule-A (fJAM-A) being a receptor for feline calicivirus (FCV), reduction of FCV binding and infection by V. cholerae NA treatment of host cells indicated that Sia is necessary for virus binding and infection. FCV binding and infection were also reduced by α2-6Sia-binding SNL but not by α2-3Sia-binding MAL, indicating the importance of α2-6 linkage for infection. Furthermore, FCV binding and infection were inhibited in the presence of both tunicamycin, which inhibits N-glycosylation, and PNGase F, which releases N-linked oligosaccharides, but not in the presence of benzylGalNAc, which inhibits O-glycosylation. These findings indicated that FCV uses α2-6Sia on N-linked glycoproteins as a primary receptor or co-receptor for infection [106]. Attachment and infection of PSaV Cowden strain were markedly inhibited by treatment of cells with V. cholerae NA and were partially inhibited by treatment of cells with α2-3Sia-cutting sialidase S, α2-3Sia-binding MAL, or α2-6Sia-binding SNL, suggesting that the virus can attach to both α2-3Sia and α2-6Sia for infection. Virus binding and infection can be reduced by treatment of cells with proteases or with benzylGalNAc but not by treatment with tunicamycin or with DL-threo-1-phenyl-2-decanoylamino-3morpholino-1-propanol (PDMP), a glucosylceramide synthase inhibitor. These findings suggest that PSaV Cowden strain uses α2-3Sia and α2-6Sia on O-linked glycoproteins that are present on porcine intestinal epithelial cells as receptors for infection

Enveloped and Non-enveloped Sia-Binding Viruses

515

[107]. Interestingly, this virus did not agglutinate pig, human, rat, chicken, or cow red blood cells and did not bind to synthetic HBGAs including A and H types, which are known to be expressed in pigs. The sialyl sugar chain structures that are not found on those red blood cell membranes but are recognized by the PSaV Cowden strain should be investigated in detail. Detailed information on sialyl glycans recognized by each virus, not only this PSaV, and how many receptors are necessary for infection should lead to the development of effective strategies for control of calicivirus infection. 4.3

Reoviridae

Non-enveloped viruses that contain discrete 10–12 segmented dsRNAs with multilayered capsids belong to the family Reoviridae. “Reo-” is derived from respiratory enteric orphan viruses according to the pathology of the first members found in respiratory and enteric tracts as orphans, which did not cause symptomatic disease at the time of discovery. These viruses are currently members of the genus Orthoreovirus (“true” reoviruses). One species of this genus, the mammalian orthoreoviruses (also called reoviruses), which are classified as nonfusogenic viruses because infectious virus particles, which are activated in the clathrin-dependent endosome by proteolysis of outer shell proteins resulting in capsid conformational rearrangement and exposure of a hydrophobic part of a viral membrane-penetration protein, can penetrate, without fusing with the endosomal membrane, into the host cytosol [111], contains three serotypes (types), 1, 2, and 3, that infect a variety of mammalian species including humans. In humans, reoviruses usually cause subclinical or mild respiratory illness, such as common cold and enteritis, but sometimes might lead to severe illnesses (Table 2) including a CNS disease in infants [112]. The spread to and infection of the CNS by reoviruses in newborn mice have been also shown to be serotype-specific; type 1 spreads hematogenously to the infection site, ependymal cells, producing nonlethal hydrocephalus, whereas type 3 spreads neutrally to infect neurons, resulting in lethal encephalitis [113]. Type 2 has not been studied in detail due to the difficulty in experimental propagation. This serotype-specific pattern of neurotropism is primarily determined by the viral attachment σ1 protein encoded by the viral S1 gene. Ten segmented dsRNAs of the reoviruses are divided into three classes according to their size: three segments are large (L1, L2, L3), three are medium sized (M1, M2, M3), and four are small (S1, S2, S3, S4). Each segment encodes one protein, except for the S1 gene encoding structural and nonstructural proteins, which are denoted by Greek letters corresponding to the L, M, and S segments that encode them: λ, μ, and σ proteins with numbering of the proteins that is not related to the segment numbers that encode them, for example, the S4 gene encodes a σ3 protein [111]. As shown in Fig. 3e, f, the outer shell of this double-shelled virus is

516

Nongluk Sriwilaijaroen and Yasuo Suzuki

composed mainly of the μ1 protein including its cleavage products and the σ3 protein and partly of the homotrimeric spike σ1 protein interacting with the homopentameric turret-like spike λ2 protein at each vertex of 12 vertices of the virus particle. The serotype-specific differences in routes of spread and in recognition of different cells in the CNS of newborn mice are thought to be due not to conserved binding located at the base of the σ1 head domain of these viruses to junctional adhesion molecule-A (JAM-A) but to different binding specificities of σ1 proteins of these viruses to sialyl glycans. Early binding specificity studies of reoviruses showed that reovirus hemagglutination is serotype-specific; types 1 and 2 preferentially agglutinate human erythrocytes, while type 3 favors bovine erythrocytes [114]. However, detailed structures of the sialyl sugar chains from these erythrocytes are still unknown. It is only known that there is only Neu5Ac in human erythrocytes [115] but that there are both Neu5Ac and Neu5Gc with a higher ratio of Neu5Gc in bovine erythrocytes [53]. Detailed information on Sia-binding specificity of reoviruses has been obtained by using glycan array screening and structural studies on the co-crystals of recombinant protein σ1 of prototypic type 1 strain Lang (T1L) [116] or prototypic type 3 strain Dearing (T3D) [117] and sialyl glycan. Glycan array analysis of T1L σ1 binding to gangliosides indicated that the GM2 binding signal is stronger than GM3, GM1, and GD1a binding signals [116]. GM2 appeared to specifically decrease type 1, but not type 3, infection of mouse embryonic fibroblasts. The crystal structure of T1L σ 1 in complex with GM2 showed that both terminal Neu5Ac and GalNAc moieties of GM2 make contact with protein σ 1 in a shallow groove in the globular head domain, while the crystal structure of the T1L σ 1-GM3 (lacking terminal GalNAc) complex (Fig. 3e) showed that only terminal Neu5Ac of the GM3 trisaccharide interacts with T1L σ1 [116]. In contrast, the crystal structure of T3D σ 1 in complex with α2-3-sialyllactose (a trisaccharide of GM3) showed that Neu5Ac makes extensive contact with each body domain of homotrimeric σ 1 protein and that the lactose (Gal-Glc) moieties participate in the contact in different directions in each body domain, presumably as a result of flexibility of the three binding sites [117]. Figure 3f shows that not only terminal Neu5Ac but also the third sugar Glc of α2-3-sialyllactose has direct hydrogen bonds with the binding site of T3D σ 1. The T3D σ1 binding site can also accommodate α2-6-sialyllactose and α2-8-disialyllactose with either Neu5Ac or Neu5G [116, 117]. It was proposed that σ1 proteins of a reovirus first selectively bind to specific sialyl glycans on the host cell surface with relatively low affinity. This is followed by binding of the viral σ1 proteins to JAM-As on the same host cell surface with high affinity. These firm attachments of the virus to sialyl glycans and JAM-As on the cell surface mediate viral internalization via clathrin-dependent endocytosis [117].

Enveloped and Non-enveloped Sia-Binding Viruses

517

Both type 1 and type 3 reoviruses also cause serotype-specific patterns of infection in the intestine of adult mice; type 1 infects crypt epithelial cells with pathology restricted to the ileum, while type 3 prefers goblet and absorptive cells, causing a wide pathology of the duodenitis, jejunitis, and ulcerative colitis [118]. Studies on reovirus infection in rat lungs indicated that both type 1 and type 3, which can bind to α2-3-sialyl linkage, can infect and replicate in type I alveolar epithelial cells, leading to pneumonia, but that type 1 produced higher titers in the lungs than did type 3, having a wide range of binding preferences to α2-3-, α2-6-, and α2-8-sialyl linkages [119]. Serotype specificity of reoviruses to cellular tropism in the respiratory tract including the upper respiratory tract may exist and more investigation is needed. An understanding of the mechanisms responsible for the differences between reovirus serotypes (types) in infection potency, cellular tropism, and pathogenesis might lead to the development of methods for diagnosis and treatment. Such an understanding might also be useful for improvements of reoviruses in clinical therapeutic applications against cancers having changes in glycan composition profiles. The other genus in the family Reoviridae for which members attach to Sia for efficient infection is the genus Rotavirus. Rotaviruses, which are a major cause of viral diarrhea in animals worldwide and cause death in about half a million infants and young children each year, possess 11 dsRNA genome segments enclosed in triple-shelled capsids. The outer shell consists of the glycoprotein VP7 layer and protruded VP4 dimeric spikes. In the presence of trypsin (a serine protease), VP4 is cleaved into VP5∗ and VP8∗. These outer shell proteins have been shown to interact with several cell surface molecules; VP7 binds to integrins αvβ3 and αxβ2, VP5∗ binds to integrin α2β1 and to heat shock cognate protein 70 (hsc70), and VP8∗ binds to Sia. While binding to hsc70 is required for all virus strains, binding to integrins and to Sia appears to be strain-dependent [120]. It has been proposed that VP8∗ at the top of the VP4 molecule initiates cell attachment via Sia interactions. Sequentially or alternatively, VP5* at the body of VP4 interacts with integrin α2β1. Then VP5* at the foot of VP4 interacts with hsc70 followed by interactions of VP7 at the outer shell layer with integrins αvβ3 and αxβ2. These multistep interactions finally trigger endocytosis via a mechanism depending on the virus strain [120]. These two outer shell viral proteins, nonglycosylated protease-sensitive VP4 proteins and glycosylated VP7 proteins, carrying type-specific epitopes are also used for classification of rotaviruses into P and G genotypes, respectively. Based on results of hemagglutination tests, hemagglutination inhibition tests, or infection inhibition tests with NA treatment/ blocking with Sia-carrying inhibitors [121, 122], inhibition of virus binding to MA-104 cells or enterocytes by GM3 and GM2 [123]

518

Nongluk Sriwilaijaroen and Yasuo Suzuki

and/or direct interactions between GD1a and VP8∗ [124], all tested strains of P genotypes 1, 2, 3, and 7 in group A rotaviruses [121, 123, 124] and strain AmC-1 in group C rotavirus [122] have been shown to be Sia-dependent viruses that require binding to terminal Sia of glycoconjugates on the host cell surface, which is sensitive to hydrolysis by an NA. Therefore, these Sia-dependent viruses are subgrouped as sialidase-sensitive strains. Two human strains, KUN in genotype P4 and MO in P8, have been tested and found to be Sia-dependent viruses that require binding to internal (branched) Sia of glycoconjugates, i.e., GM1a (¼GM1) (Fig. 1), but not binding to asialo GM1a (called GA1). The internal (branched) Sia of gangliosides appears to be insensitive to hydrolysis by an NA [125]. Therefore, these Sia-dependent viruses are subgrouped as sialidase-resistant strains. Later, infections of human strains Wa (G1P1A(8), 1A being a serotype that precedes P genotype in parenthesis), RV-3 (G3P2A(6)), RV-5 (G2P1B(4)), and S12/85 (G3P2A(6)) and bovine strain UK (G6P7(5)) were shown to be inhibited by the GM1-binding cholera toxin B (CTB), while infections are increased by the use of cells in which exogenous GM1 is incorporated or by the use of sialidase-treated cells having increased GM1 levels [126]. Thus, these human and bovine rotaviruses in P genotypes 4, 5, 6, and 8 are classified as Sia-dependent viruses in a subgroup of sialidase-resistant strains. Surprisingly, infection of porcine strain TFR-41 (G5P9(7)), which is an sialidase-sensitive strain, was reduced by CTB treatment [126]. Thus, this virus is classified as an sialidase-sensitive strain that can bind to both terminal Sia and internal Sia as shown in Table 2. The binding preferences of rotaviruses for Sia species have been determined directly and indirectly. Infection of the human sialidaseresistant strains Wa, RV-3, and RV-5 in sialidase-treated cells is reduced in the presence of Neu5Acα2Me [126], suggesting that Neu5Ac is the Sia species preference of these strains. Studies on inhibition of rotavirus binding to host cells by Neu5Gc-containing aceramido-GM3Gc and Neu5Ac-containing aceramido-GM3Ac [127]; inhibition of rotavirus infection by an anti-Neu5Gc antibody, Neu5Gcα2Me, and Neu5Acα2Me [126, 128]; and analysis of the crystallographic structure of VP8∗ in complex with Neu5Gcα2Me [128] revealed that VP8∗ proteins from the porcine sialidase-sensitive strains CRW-8, OSU, and YM and the bovine strains SA11 and NCDV, which carry small residue Gly187, have greater specificity for Neu5Gc holding an extra hydroxyl group over Neu5Ac, whereas the rhesus (simian) NA-sensitive strain RRV VP8∗, which carries Lys187, has a higher preference for Neu5Ac than Neu5Gc and the bovine sialidase-resistant strain UK with Lys187 favors Neu5Ac. It was shown that CRW-8 virus acquires a Pro157-to-Ser157 mutation in VP8∗ that decreases Neu5Gc-binding affinity during cultivation in MA104 cells [128], indicating virus adaptation to host cell surface receptors.

Enveloped and Non-enveloped Sia-Binding Viruses

519

Recent studies on sialyl linkage-specificity of sialidase-sensitive rotavirus strains have demonstrated that infection of bovine strain NCDV in genotype P1 and canine strain CU-1 in P3 was decreased by α2-6Sia-binding SNL but not by α2-3Sia-binding MAL. Infection of porcine strains PRG9121 in P7 and PRG942 in P23 was inhibited by both α2-6Sia-binding SNL and α2-3Sia-binding MAL. Pretreatment of MA104 cells with either PDMP, which inhibits glucosylceramide synthase, or tunicamycin, which inhibits N-glycosylation, inhibited infection of all viruses. In contrast, pretreatment with benzylGalNAc, which inhibits O-glycosylation, did not affect any virus infection. These results suggested that bovine P1 NCDV and canine P3 CU-1 strains have binding preference to α2-6Sia on gangliosides or N-linked glycoproteins, while both porcine P7 PRG9121 and P23 PRG942 strains can bind to both α2-6Sia and α2-3Sia on gangliosides or N-linked glycoproteins [129]. A rotavirus vaccine is now available, but there is still lack of antiviral drugs for treatment. The receptor-binding specificity of rotaviruses could be useful for designing effective inhibitors against VP8∗ attachment at the initial step of infection. 4.4

Adenoviridae

Human adenoviruses (HAdVs) belong to the genus Mastadenovirus (mammalian) in the family Adenoviridae. They are mediumsized (70–100 nm) non-enveloped viruses with a linear dsDNA genome. Type 90 (Ad90 or HAdV-90) has recently been identified, and there are currently 90 identified HAdV types [130] that are grouped into 7 species, A to G. Only some types in species D that are associated with human ocular diseases and type 52, the only current member in species G, that is associated with gastroenteritis are thought to use Sia as a primary receptor (Table 2). As shown in Fig. 3g, the viral capsid of HAdV is constructed from hexon trimers and penton pentamers that project long fiber trimers outward from each vertex. The fiber trimers carry receptor-binding sites at each monomeric distal globular end, the knob domain, for binding to primary receptors on the host cell surface. The primary binding allows the penton pentamers to make contact with secondary receptors on the host cell surface, leading to endocytosis typically via clathrin [131]. However, it has been shown that endocytosis of the human adenovirus type 37 in species D (HAdV-D37) into corneal cells uses caveolin-1 proteins clustered in lipid rafts [132]. Based on reduction of fiber knob/virus binding and infection of sialidase-treated cells, 6 types of ocular pathogens in HAdV species D including 8, 26, 37, 53 (a natural intertypic recombinant of HAdV types 8, 22 and 37), 54 (believed to be an HAdV-D8 variant strain), and 64 (formally known as 19a, a natural recombinant of HAdV types 19p (p ¼ prototype), 22 and 37) [133] have so far been shown to use Sia as a primary cellular receptor for binding that is mediated by their trimeric fiber knobs [134–

520

Nongluk Sriwilaijaroen and Yasuo Suzuki

136]. Pretreatment of cells with an NA reduced interactions of HAdV-D19p, which does not cause an eye infection, but did not affect its productive infection, suggesting that Sia may support the virus attachment but is not used as a functional receptor for virus infection [134]. X-ray crystallographic studies have shown that both HAdV-D37 and HAdV-D19p fiber knobs, which have only two different amino acids (Lys240 and Asn340 for HAdV-37 and Glu240 and Asp340 for HAdV-D19p), bind to both α2-3 and α26-sialyllactose, indicating that sialyllactoses are not determinants of the tropism of these viruses [137]. Glycan array screening showed that the HAdV-37 knob domain specifically binds to GD1a glycan, which is a disialyl branched hexasaccharide. Surface plasmon resonance analysis revealed that GD1a glycan binding to the HAdVD37 knob has about 260-fold higher affinity with Kd of 19 μM than binding of sialyllactose to the HAdV-D37 knob with Kd of 5 mM [137], presumably because two terminal Sias on a single GD1a glycan can directly interact with two of three binding sites in the trimeric knob (Fig. 3g) as indicated by molecular modeling, nuclear magnetic resonance, and X-ray crystallographic studies [138]. Several experiments have indicated that HAdV-D37 binds to cell surface O-linked glycoproteins rather than to ganglioside. For example, neither reduction of ganglioside biosynthesis by treatment with P4 compound [(1R,2R)-1-phenyl-2-hexadecanoylamino-3-pyrrolidino-1-propanol] nor removal of cell surface Nglycans by treatment with PNGase F affected binding of HAdVD37 to HCE cells, while reduction of O-linked glycan synthesis by treatment with benzylGalNAc efficiently inhibited binding of HAdV-D37 to HCE cells and infection of the cells. Thus, O-linked glycoproteins that carry glycans that mimic GD1a glycan are functional host receptors of HAdV-D37. Amino acid sequence analysis revealed that fiber knobs of HAdV-D37 and HAdV-D64 have identity of 100% [136], suggesting that HAdV-D64 possesses the same binding pocket with preferential binding to O-linked glycoproteins bearing GD1a glycan. Fiber knobs of HAdV-D8, HAdVD26, HAdV-D53, and HAdV-D54 carrying some amino acids that are different from those in the above fiber knobs [135, 136] should be further identified for determining positions of amino acids in 3D structures of the fiber knobs and screened for their receptorbinding specificity. Using virus overlay protein blot assays, HAdV-D37 (Ad37) has been shown to interact with a human conjunctival (Chang C) membrane protein with an approximate molecular weight of 45 kDa (CAR) in a calcium-independent manner, with a Chang C membrane protein with a molecular weight of 50 kDa (a membrane protein that was later identified as CD46) in a calcium-dependent manner and with a Chang C membrane protein with a molecular weight of 60 kDa (sialylated protein) in a calcium-dependent manner. Pretreatment of Chang C cells with anti-CAR antibodies had

Enveloped and Non-enveloped Sia-Binding Viruses

521

little effect on HAdV-D37 infection, indicating that HAdV-D37 infection does not use CAR to infect Chang C conjunctival cells. Pretreatment of Chang C cells with an NA led to abolishment of HAdV-D37 binding to the 60-kDa protein but did not affect HAdV-D37 binding to the 50-kDa protein and infection. These findings suggested that HAdV-D37 can bind to and infect conjunctival cells through the 50-kDa protein (CD46) independently of Sia [139]. Typically, HAdVs have one type of fiber trimers, but all enteric HAdVs including species F containing HAdV-F40 and HAdV-F41 and species G containing HAdV-G52 have two different types of fiber trimers, long and short. All of these three viruses use their longer fibers for attachment to the coxsackie and adenovirus receptor (CAR) on target cells. However, HAdV-F40 and -F41 short fibers do not bind to Sia, while HAdV-G52 short fibers have been shown to attach specifically to Sia. It appears that HAdV-G52 requires either CAR or Sia for virus infection. Removal of the host cell surface Sia by NA treatment abolished HAdV-G52 binding to and infection of Sia-expressing original Chinese hamster ovary (CHO) cells but did not abolish (only partly reduced) virus binding to CHO cells that also express CAR (CAR-expressing CHO cells). When compared to original CHO cells, HAdV-G52 showed a remarkable increase in binding to CAR-expressing CHO cells [140]. Thus, HAdV-G52 may use its long or short fiber knob or both knobs for attachment to host CAR and/or Sia as a primary receptor or co-primary receptors to initiate virus infection. However, there are several questions that need to be answered for understanding the mechanism of virus infection. For example, why do all three known enteric HAdVs have two spikes (fibers), while other HAdVs need only one primary receptor-binding spike? Why do short fibers of HAdV species F not bind to Sia like short fibers of species G do? Both proteins and carbohydrates that are possible receptors on target gastrointestinal epithelial cells, especially those on the luminal (apical) surface that is a site of adenovirus infection, should be identified. Sialyl glycan structures to which HAdV-G52 preferentially binds have been determined [140, 141]. Glycan microarray screening using a variety of α2-3- and α2-6-sialylated probes showed that several probes with one or two α2-3-sialyl linkages can be bound by the HAdV-G52 short fiber knob (52SFK) but that probes with α26-sialyl linkages cannot be significantly detected for their binding to 52SFK. The strongest binding observed was to the probe with a type II (Galβ1-4GlcNAc) backbone sequence, Neu5Acα2-3Galβ14GlcNAcβ1-3Galβ1-4Glcβ- [140]. However, the use of a broader range of probes including α2-8- and α2-9-sialyl linkages in glycan microarray screening of 52SFK binding specificity indicated that 52SFK interacts preferentially with linear α2-8-linked oligoSia (polySia, !8Neu5Acα2-), especially at a degree of polymerization

522

Nongluk Sriwilaijaroen and Yasuo Suzuki

of 5 to 9 that produced very strong binding signals that were much greater than those produced by α2-3-sialyl linkage [141]. Since polySia is abundant in brain and lung cancers, the molecular relationships between HAdV-G52-mediated gastroenteritis and 52SFK binding specificity to polySia receptors remain unknown [141]. Determination of the glycosylation type that is required for virus binding revealed that inhibition of O-linked glycosylation, but not inhibition of glycolipids or N-linked glycosylation, resulted in reduction of both virus particles and 52SFK binding to A549 cells, suggesting that O-linked glycans play a major role in 52SFK binding specificity [140]. Nonhuman adenoviruses that have been shown to interact with Sia are canine adenovirus type 2 (CAdV-2), which causes a respiratory disease in dogs and is grouped in the genus Mastadenovirus (mammals) [142], and turkey adenovirus type 3 (TAdV-3), which has a virulent form known as turkey hemorrhagic enteritis virus (THEV) that causes hemorrhagic enteritis in turkeys, splenomegaly in chickens, and marble spleen disease in pheasants and an avirulent form that is used as a vaccine. TAdV-3 was previously grouped in subgroup II avian adenoviruses but is currently grouped in the species turkey siadenovirus A in the genus Siadenovirus (amphibians, dinosauria, testudine species) [143]. Crystallographic studies showed that the CAdV-2 fiber knob can bind to both α2-3-sialyl-Dlactose and CAR D1 in different binding sites. However, binding between the CAdV-2 fiber knob and highly sialylated glycophorin cannot be detected, while hemagglutination of CAdV-2 correlates well with expression of CAR on erythrocytes: CAdV-2 can agglutinate rat and human erythrocytes, which have high CAR levels, but cannot agglutinate erythrocytes from dogs, mice, rabbits, lemurs, and monkeys, which have undetectable levels of CAR [142]. Thus, CAR on erythrocytes is likely to be an important factor for CAdV2 binding. By glycan microarray analysis, fiber knobs of both virulent and avirulent TAdV-3 strains have been shown to interact with both α23- and α2-6-sialyllactoses but not to bind to α2-3-sialyllactosamine. Isothermal titration calorimetry showed that the binding affinities of both fiber knobs (heads) to α2-3- and α2-6-sialyllactoses were in the mM range at pH 6.0 and less at pH 7.2 with a two- to fourfold higher affinity for α2-3-sialyllactose. Avirulent fiber knobs appeared to bind 1.5- to 3-fold more strongly to sialyllactoses than did virulent fiber knobs. Crystallographic studies showed that the Sia-binding region of TAdV-3 fiber heads comprises amino acids at positions of 392 and 419–423, while two amino acid differences between virulent and avirulent TAdV-3 fiber head domains are at the positions of 354 and 376; hydrophobic Ile354 and hydrophilic Thr376 in a virulent fiber head are substituted with hydrophobic Met354 and Met376 carrying a long side chain with a sulfur atom in an avirulent fiber head [143]. These two amino acids are on the

Enveloped and Non-enveloped Sia-Binding Viruses

523

protein surface in separate regions different from the Sia-binding region. The 354 and 376 regions may be binding regions for different ligands, such as CAR and CD46, possibly responsible for virulent and avirulent pathogenesis that need more investigation. The presence of multi-receptor-binding sites of the fiber knobs (heads) of adenoviruses (such as HAdV-D37 fiber knobs containing binding sites for CD46 [139], Sia [138], and CAR [142]) should be taken into consideration when designing antivirals. The roles of these multi-receptor-binding sites in the life cycle of adenoviruses also need to be understood not only for efficient control of pathogenic viruses but also for design of efficient vectors for gene therapy and for development of oncolytic virotherapy with cellspecific tropism. 4.5

Parvoviridae

Both dependoviruses and autonomous parvoviruses, in the Parvovirinae subfamily in a family of 20–26-nm diameter small non-enveloped viruses with a 4–6-kb linear ssDNA genome, Parvoviridae, have been found to bind to Sia. Probably due to the very small amount of genetic materials to code the necessary biochemical apparatus, as indicated by its name, efficient replication of 4–5kb ssDNA dependoviruses, except for duck and goose parvoviruses, depends on a larger helper virus such as adenovirus, herpesvirus, or papillomavirus. Thus, they are nicknamed replication-defective parvoviruses or adeno-associated viruses (AAVs). The other 5–6-kb ssDNA autonomous parvoviruses can autonomously replicate inside the cells but only during the DNA synthesis (S) phase of the host cell cycle due to the very simple viral genome [144]. Although replication-defective parvoviruses have not been reported to cause pathology, replication-autonomous parvoviruses have been found to have both nonpathogenic and pathogenic members [145]. Binding of a parvovirus to one or more plasma membrane receptors initiates virus infection typically via clathrin-mediated endocytosis that is possibly breached by lipolytic activity of phospholipase A2 located on the N-terminal “unique region” of VP1 (uVP1). This leads to release and transport of the capsid and/or viral DNA into the cytosol and the nucleus [146]. Receptor binding is a function of a viral capsid protein (CP) as shown in Fig. 3h. Depending on the virus strain and the maturation stage, each CP is comprised of two to four overlapping capsid proteins (called VP1, VP2, VP3, and VP4). Sixty copies (subunits) of each CP are assembled to build an icosahedral capsid surrounding the parvovirus [147]. VP1 and VP2 are formed by alternative splicing of the same messenger RNA (mRNA), and the entire sequence of VP2 is encoded within the VP1 gene. In some viruses, a third structural protein, VP3, is formed (only in DNA-containing capsids) by cleavage of a peptide from the amino terminus of VP2.

524

Nongluk Sriwilaijaroen and Yasuo Suzuki

As shown in Fig. 3h, each VP is an isoform of VP1 shortened at the N-terminus and each VP thus contains the same C-terminal amino acids (receptor-binding region). VP2 arisen from alternative splicing (cleavage) of the same mRNA as that of VP1 lacks the N-terminal “unique region” of VP1 (uVP1) that is critical for releasing the virus from the endosome. VP3 and VP4 are likely to be produced through posttranslational cleavage of the N-terminal amino acids of VP2 and VP3, respectively. VP1 is produced at a much lower copy number than that of VP2. However, the major capsid component seems to be the final product of posttranslational cleavage [147]. As shown in Fig. 3h, it is an adeno-associated viral capsid consisting of VP1 (full-length VP protein), VP2 (lacking uVP1), and VP3 (lacking nuclear localization signal, NLS), but not VP4, with a ratio of around 1:1:10 [148, 149]. Why do all VPs contain the same receptor-binding site? Are there different roles? Binding of VP1 to a receptor(s) seems to be involved in triggering conformational changes in the uVP1 region [148], leading to exposure of the phospholipase A2 domain from the capsid interior that is critical for viral infectivity as mentioned above. The different roles of the receptor-binding regions present in VP1 and uVP1-lacking VP2 and VP3 should be investigated in detail. Receptors on the host cell plasma membrane, both carbohydrates and proteins, have been investigated for many viruses in this family. Here we focus on Sia receptors recognized by viral capsid proteins, especially adeno-associated viruses, which have been extensively studied as shown in Table 2. For autonomous parvoviruses, bovine parvovirus (BPV) [150] and porcine parvovirus (PPV) [151] are known to bind to α2-3 O-linked and either Olinked or N-linked sialic acids, respectively, on the cell surface for attachment. Binding to both α2-3 and α2-6Sia of canine parvovirus (CPV) and feline panleukopenia virus (FPV) which is a canineadapted form that emerged in 1978 in cats does not seem to be for cell infection, though their binding preferences to different Sia types, CPV specific for Neu5Ac and FPV specific for Neu5Gc, seem to be determinants of host adaptation in dogs and cats, respectively [152]. During evolution, it appeared that binding of CPV to Sia drifted from neutral pH to acidic pH in association with an Asn375Asp mutation in VP2. This virus variant (CPV type 2a, CPV-2a) emerged in 1979 and replaced the CPV-2 strain globally. FPV (Asp375) and CPV-2a (Asp375), which are transmitted through an oro-nasal route, can cause preferential binding to Sia at pH below 6.5 and were assumed to allow attachment of virus progenies released from infected progenitor cells in crypts of the intestinal epithelium to Sias on materials, such as mucus, in the intestinal lumen, where its pH is around 5.5–6.6. Such binding was suggested to enhance virus shedding in stools and persistence in the environment for transmission to other hosts [152]. This is an example of Sia-binding viral lectins, VP2 of CPV and FPV, that

Enveloped and Non-enveloped Sia-Binding Viruses

525

evolved not for infection of target cells. Another Sia-binding viral lectin that is not used for virus infection is spike (S) protein of several strains of transmissible gastroenteritis virus (TGEV), which is a linear (+)ssRNA enveloped virus in the genus Alphacoronavirus in the family Coronaviridae that causes fatal diarrhea in newborn piglets and binds to specific α2-3Neu5Gc. The protein is thought to allow the virus to persist in and pass through unfavorable environments, such as environments with low pH, proteases, and bile salts, when traveling in the alimentary tract to the target site, the intestinal epithelium [69]. Sia appears to be important for infection of minute virus of mice (MVM) and H-1 parvovirus (H-1PV, isolated from a human tumor cell line transplanted in rats), which are autonomous rodent parvoviruses that infect, propagate in, and kill tumor cells but not non-transformed cells and are thus promising antitumor agents [153]. The Sia-binding sites play a determinant role in cellular tropism of these two viruses and are thus important for tumor selectivity (oncotropism). Some studies have shown that a change in amino acid residues of the Sia-binding site affects receptorbinding preference that probably lead to a change in host range, expanded/restricted/retargeted [147, 154]. For example, constructed recombinant MVMp (p, prototype strain) viruses with VP Ile362Ser and/or Lys368Arg substitution(s) near the Sia-positioned dimple according to substitutions in lethal MVM variants derived from infection of severe combined immunodeficient (SCID) mice by the apathogenic strain (MVMp) were reported to have a lower affinity for Sia receptors, produce a large-plaque phenotype in cell lines in vitro, and cause lethal disease in SCID mice [147]. That study not only indicated that lowering of the Sia receptor-binding affinity of the virus leads to extension of viral tropism and a dramatic increase in viral pathogenicity but also suggested that a virus variant can be generated via both genetic engineering and natural selection. This raises serious concerns about whether the use of oncolytic viruses, especially in immunosuppressed patients, is safe and whether virus variants causing severe disease may emerge during oncovirotherapy. Virotherapeutics with formulations guaranteeing genetic stability may be required. In contrast to autonomous parvoviruses, adeno-associated viruses (AAVs) have not been reported to cause any disease (generally low levels of immunogenicity and toxicity), and they can infect both dividing and nondividing cells with long persistence in sitespecific integration in the host cell genome and in an episomal state when the gene part required for chromosomal integration was removed from its linear ssDNA [155, 156]. Although AAVs are replication-defective parvoviruses, AAVs are now widely used viral vectors carrying therapeutic genes for gene therapy of various disorders by which helper genes from an adenovirus for mediating

526

Nongluk Sriwilaijaroen and Yasuo Suzuki

AAV replication may be inserted in the other plasmid or directly inserted in the AAV vector. However, an immune response to AAV transduction has sometimes been observed, leading to limitation of the use of AAV vectors. For example, unexpected liver toxicity occurred when AAV2 was transduced into hepatocytes during a clinical gene therapy trial for hemophilia B due to cytotoxic T-lymphocyte (CTL) activation by the AAV2 capsid heparin binding motif [155]. This indicated that in addition to developments for increasing the AAV genome capacity and enhancing gene expression, an understanding of AAV-human host interactions is needed for the development of highly efficient transduction specific to target cells, not immune and other normal cells. AAVs are ubiquitous; 13 serotypes (AAV1–AAV13) that differ in their capsid structure have so far been identified in human and nonhuman primate sources. Like other parvoviruses, the viral capsid protein (CP as shown in Fig. 3) is used for binding to cell surface receptors for mediating viral entry and is thus a prime determinant of the host range and host specificity [157]. AAV1, AAV4, AAV5, and AAV6 bind to the terminal Sia by the viral capsid Sia-binding motif found on all CP components, VP1, VP2, and VP3, but they differ in binding preference for the asialo portion, including glycosidic linkage types. As shown in Table 2, AAV4 requires α2-3Sia on O-glycans, whereas AAV5 preferentially binds to either α2-3 or α26Sia on N-glycans [158]. AAV1 and AAV6 can bind to either α2-3 or α2-6Sia-N-glycans [159]. Although AAV1 and AAV6 are closely related with only 6 residues from a total of 736 amino acid residues being different in their VPs, AAV6 appears to additionally bind to negatively charged heparan sulfate proteoglycans (HSPG), which is made possible by a single amino acid difference at the position of 531 (AAV6 Lys531, but AAV1 Glu531 as indicated in Fig. 3h). Some co-receptors have been found to be necessary for optimal AAV attachment and internalization: platelet-derived growth factor receptor (PDGFR) is a co-receptor for AAV5 and epidermal growth factor receptor (EGFR) is a co-receptor for AAV6 [160]. In addition to the primary receptor-binding region, the co-receptor-binding site should be identified for site-specific modification of both the primary and co-receptor-binding sites on the AAV capsid for therapeutic applications. Transduction efficiency of each natural AAV serotype in major tissues has been investigated and the results, summarized in Table 2, show differential tropism of AAV serotypes [160]. Bovine adeno-associated virus (BAAV) is a nonprimate AAV that uses Sia on glycosphingolipid (ganglioside) as a receptor for transduction (infection) [161]. It was shown that BAAV injected via canalostomy can efficiently deliver genes and control gene expression of connexin, leading to rescue gap junction coupling in cochlear non-sensory cells of the inner ear in adult mice with a nonsyndromic hearing loss and deafness (DFNB1) phenotype

Enveloped and Non-enveloped Sia-Binding Viruses

527

[162]. Although persistence of BAAV-mediated gene replacement in the cochlea remains limited, the results of that study indicated that BAAV can be further developed as a recombinant viral vector for gene therapy. Taken together, the results indicate that the receptor expression pattern on the host cell surface varies not only between different host cell membranes but also between different stages of cell differentiation and maturation. Further extensive analysis and identification of the host receptor expression pattern together with studies on AAV binding specificities to receptors on target tissue cells and nontarget immune cells could lead to successful rational modification of the AAV capsid to specific receptors found on cells in the target tissue but not on immune cells. The constructed AAV vectors could increase transduction efficacy and specificity in the desired tissue and overcome concerns about adverse side effects including induction of an immune response. 4.6

Polyomaviridae

Typically, infections by viruses in the family Polyomaviridae are asymptomatic. However, the viruses can cause diseases, especially in immunocompromised individuals, such as severe kidney and brain disorders caused by human John Cunningham polyomavirus (JCPyV) and skin cancer called Merkel cell carcinoma (MCC) that is caused by Merkel cell polyomavirus (MCPyV) (Table 2). Also, BK polyomavirus (BKPyV) can be reactivated to cause nephropathy in renal transplant recipients [163]. In addition, polyomaviruses (poly- ¼ many, oma ¼ tumors) can transform cells in cultures and in immunocompromised laboratory animals including newborn mice [164]. Surprisingly, simian virus 40 (SV40), which is a monkey virus that was accidentally introduced into humans with contaminated polio vaccine and is able to be transmitted among humans, has recently been detected in a variety of human cancers [165]. The increasing number of detected human and animal polyomaviruses raises concerns that polyomaviruses might acquire mutation and/or recombination with human polyomaviruses with the potential to cause diseases and cancers. To infect a host cell, a polyomavirus, which is a non-enveloped DNA virus, must bind to a receptor(s) on the host plasma membrane, triggering caveolae/raft-mediated endocytosis or clathrin-mediated endocytosis (for JCPyV), and navigate through the ER before delivery of its DNA genome into the nucleus for transcription and replication (Fig. 1). It should be noted that a caveolin/clathrin-independent pathway was observed for some polyomaviruses such as entry of BKPyV into primary human renal proximal tubule epithelial cells [166]. Viral capsid protein 1 (VP1) is responsible for engagement of the host cell receptor, an important step that determines cellular tropism and pathogenesis of the virus. Internalization of JCPyV into cells via clathrin-mediated endocytosis unlike other related polyomaviruses, including murine polyomavirus (mPy causing

528

Nongluk Sriwilaijaroen and Yasuo Suzuki

tumors in newborn mice, not shown in the table), SV40, BKPyV, and MCPyV, suggested that JCPyV has receptor-binding preference that is different from other related polyomaviruses. JCPyV prefers α2-6Sia neolacto-series on N-linked glycoproteins [167, 168] while its binding to the 5-HT2 serotonin receptor seems to facilitate this endocytosis pathway [169]. SV40, mPy, and BKPyV (either caveolin-dependent or caveolin-independent entry) share a common binding for ganglio-series gangliosides [166, 170–172] with differential binding preferences. SV40 prefers the branched α2-3Sia GM1 ganglioside [170], mPyV mainly attaches α2-3Sia on GD1a/GT1b/GT1a [173, 174], and BKPyV and MCPyV commonly prefer to bind to α2-8Sia b-series gangliosides including GD3/GD2/GD1b/GT1b for BKPyV [171] and GT1b in cooperative binding with a glycosaminoglycan (GAG) for MCPyV [172]. Comparison of the crystal structures of BKPyV VP1-GD3 and SV40 VP1-GM1 (Fig. 3i) and site-directed mutagenesis studies have provided an important evidence that the amino acid at position 68 is responsible for different α2-3/α2-8 linkage preferences of these two viruses and thus acts as a determinant of receptor specificity. The use of distinct receptors to initiate virus infection contributes to the differences in virus host/tissue/cell tropism and pathological consequences as shown in Table 2. Viruses can rapidly adapt to their environment. Although polyomaviruses are circular dsDNA viruses using host polymerases for transcription and replication, progressive multifocal leukoencephalopathy (PML)-mutant strains of JCPyV have been detected with VP1 mutations near the apical Sia-binding pocket, possibly resulting from positive selection during the PML development [175]. Complete blockage of the attachment of PML-mutant virus-like particles and partial reduction of binding of wild-type JCPyV genotypes 2 to either SFT cells (gliosarcoma cells with the SV40 large T antigen) or ART cells (ovarian cancer cells with the SV40 large T antigen) by heparin and HS20, which are GAGs, were observed. These findings together with results of other experiments including experiments on infectivity in the presence of various treatments that inhibit engagement of GAGs or sialyl receptors indicated that while the wild-type JCPyV strain can use either GAGs or sialylated glycans for infection, PML-VP1 mutant JCPyV strains have lost the ability to bind to sialylated glycans but use GAGs as alternative receptors for infection. Experiments with wild-type BKV and a BKV-VP1 Phe76Trp constructed mutant showed the same results. When the GM1 receptor is unavailable, GAGs are used as alternative receptors for SV40 infection in some cell types [175]. These findings provide a better understanding of virus evolution and a switch of receptor-binding specificity that trigger alteration of virus tissue/cell tropism and virus pathology, an understanding that is critical for diagnosis and treatment.

Enveloped and Non-enveloped Sia-Binding Viruses

5

529

Molecular and Structural Basis of Viral Lectin–Sialyl Glycan Interactions The role of sialic acid in viral attachment, which is a key first step in infection of many viruses, has sparked much interest in characterization of viral lectin-Sia complexed structures. Characterization of the structures will facilitate further studies on a single amino acid mutation or a few mutations in the viral lectin binding pocket and investigation of the biological effects of the mutation or mutations for better understanding of virus evolution and variation in receptor-binding specificity, which is important for development of a plan for viral control. Characterization of the complexed structure will also facilitate studies on modifications of the sialyl glycan structure binding to the viral lectin that would pave the way for design of antivirals against viral lectin–Sia interactions. However, only a limited number of viral lectin-sialyl glycans have been co-crystallized for structural studies of their interactions. In the family Caliciviridae, only the crystal structure of norovirus P lectin of a GII.9 strain (VA207) in complex with α2-3-sialyl-Lex tetrasaccharide (PDB, 3pvd) has been determined, but the Sia residue is far from the binding site and does not participate in binding, and VA207 virus was thus classified as a Lewis carbohydrate-binding virus, not a Sia-binding virus [176]. Thus, only interactions between a sialyl glycan with viral lectins of Orthomyxoviridae human influenza H3 HA trimer, Paramyxoviridae mumps HN tetramer, Coronaviridae porcine torovirus HE, Picornaviridae CVA24v VP1/VP2/VP3/VP4, Reoviridae reovirus T1L σ1 trimer, Reoviridae reovirus T3D σ1 trimer, Adenoviridae HAdVD37 fiber trimer, Parvoviridae AAV VP3 monomer, and Polyomaviridae SV40 VP1 pentamer have been analyzed using PYMOL and are illustrated in Fig. 3. Their PDB structural information is shown in Table 3. As shown in Fig. 3a, interactions of the receptor-binding site located on the head of human H3/2010 HA with pentasaccharide Neu5Acα2-6Galβ1-4GlcNAcβ1-3Galβ1-4Glc (LSTc), a human respiratory receptor analog [177], indicate that negatively charged Neu5Ac has direct hydrogen bond interactions with Tyr98, Tyr137, Ser136, Asn145, Ser227, and Ser228. A change of Tyr98 to nonhydroxyl Phe and a change of Ser136 to negatively charged Asp found on H17 and H18 hemagglutinins make them lack the ability to bind to negatively charged Sia. Ser228 is well known to be a critical determinant of the receptor-binding specificity of H3 HAs to the α2-6-sialyl linkage [24], different from H1 HAs that contain D190/D225 as critical determinants for α2-6sialyl linkage [32]. Gly225 in human H3/2010 HA binds directly to Gal-2. Asn193 and Tyr159 provide direct hydrogen bond contacts with Glc-5 of long α2-6Sia receptors that are found in the human bronchus [44] but are rare in human alveoli [24].

530

Nongluk Sriwilaijaroen and Yasuo Suzuki

Table 3 Crystal structures of viral lectins in complex with sialyl glycans that were used for analysis

Family

Viral lectin-sialyl glycan complex

Orthomyxoviridae Influenza A/Minnesota/11/2010 (H3N2) virus hemagglutinin in complex with LSTc

PDB no.

Ref.

5xrs [177]

Paramyxoviridae

Mumps virus hemagglutinin-neuraminidase in complex with α2-3sialyllactose

5b2d

[51]

Coronaviridae

Porcine torovirus hemagglutinin-esterase in complex with 4,9-di-Oacetyl-Neu5Ac

3i1l

[64]

Picornaviridae

Coxsackievirus A24v VP1 in complex with α2-6Neu5Ac of disialyllacto-N-tetraose (DSLNT)

4q4y

[90]

Reoviridae

Reovirus type 1 (strain Lang) σ1 protein in complex with α2-3sialyllactose Reovirus type 3 (strain Dearing) σ1 protein in complex with α2-3sialyllactose

4gu4 [116] 3s6x [117]

Adenoviridae

Adenovirus type 37 fiber knob in complex with GD1a oligosaccharide 3n0i [138]

Parvoviridae

Adeno-associated virus serotype 1 in complex with α 2-3Neu5Ac of α2-3-sialyl-LacdiNAc (3SLDN)

5egc [180]

Polyomaviridae

SV40 VP1 pentamer in complex with GM1 oligosaccharide

3bwr [170]

Crystal structures of mumps HN in complex with trisaccharide α2-3-sialyllactose can be determined, whereas the co-crystal structure with α2-6-sialyllactose cannot be detected, indicating that mumps HN cannot bind efficiently to the α2-6-sialyl linkage [51]. The difference in sialyl linkage binding preference of human mumps HNs from human influenza HAs indicates that further analysis of sialyl glycans expressed on mumps target sites including human parotid gland epithelial cells is needed for clarifying the different tissue tropism of these viruses. Neu5Ac of α2-3-sialyllactose is bound by an arginine triad (Arg180, Arg422, and Arg512), Glu264, and Tyr323, whereas Glc-3 of the compound forms a hydrogen bond with Val476 in a top pocket of the HN head domain. These interactions suggest that the third sugar of the sialyl glycan is an important part of a receptor determinant for mumps virus. Further studies on receptor-binding specificity of this virus using α2-3-sialyl glycans with variation in the asialo portion should provide an insight into key binding determinants required for mumps infection. Currently known interactions suggest that α2-3Sia but not an α2-6 analog may be developed as a mumps inhibitor.

Enveloped and Non-enveloped Sia-Binding Viruses

531

The structure of the esterase-deficient PToV HE (Ser46Ala mutant) strain Markelo in complex with the synthetic receptor 4,9-di-O-acetyl-Neu5Acα2Me [64] shows the receptor ligand in the HA lectin domain (R, receptor-binding site) on the top of an HE homodimer. 4,9-Di-O-acetyl-Neu5Ac- forms hydrogen bonds with four amino acid residues, Arg161, Tyr164, Glu220, and Ser222, in the R domain and with one residue, Tyr118, in the esterase (E) domain. Studies on substrate specificities of enzymatic de-O-acetylation of PToV and BToV HEs indicated that PToV HE prefers 9-mono-O-acetylated Sias (Neu5,9Ac2), whereas BToV can catalyze both mono- and di-O-acetylated Sias (Neu5,7,9Ac3) as substrates. It is assumed that esterase and HA lectin pockets coevolved, and the PToV HE receptor-binding site seems to bind 9-mono- and exclude 7,9-di-O-acetylated Sias. Observation of the synthetic receptor 4,9-di-O-acetyl-Neu5Acα2Me in the receptorbinding pocket of an esterase-deficient PToV HE indicated that the receptor-binding domain of PToV HE can bind to 4,9-di-O-acetylNeu5Ac- and possibly that an O-acetyl group at C7, but not at C4, is obstructed, assumedly by the side chains of Val166 and Tyr118 near C7 of the Neu5Ac glycerol side chain. However, the difference in PToV and BToV HE esterase-substrate specificities indicated that viral HE proteins have adaptation in their host [64]. Further analysis of glycan profiles on each of the host mucins and host target cells on the mid-jejunum to distal ileum together with HE amino acid sequencing and direct receptor-binding assays and esterasesubstrate specificity assays should lead to an understanding of host-driven viral HE evolution of interactions between hostspecific HE proteins of toroviruses and O-acetylated Sias. Crystallization of CVA24v with several commercially available sialyloligosaccharides that have differences in glycan composition and linkage, including 60 SL, 30 SL, 30 SLN, LSTc, GD1a, sialyl-Lex, GD1b, DSLNT, GM1, GM2, and GD3, revealed that the virus has preferential binding to α2-6-sialyllactose (60 SL) and disialyllactoN-tetraose (DSLNT, a hexasaccharide (Neu5Acα2-3Galβ1-3 (Neu5Acα2-6)GlcNAcβ1-3Galβ1-4Glc-) that carries Neu5Acα23Gal- and Neu5Acα2-6GlcNAc- terminals); very weak binding to α2-3-sialyl glycans in 30 SL, 30 SLN, and sialyl-Lex; and undetectable binding to α2-8, α2-3-disialyl glycan in GD1b and GD3 and branched α2-3-sialylated glycans in GM1, GM2, and GD1a [90]. Figure 3d shows the crystal structure of CVA24v soaked with DSLNT. Neu5Ac of DSLNT is found in a shallow binding site at the tip of the crown, viral capsid protein VP1 on the virus shell. The Neu5Ac forms hydrogen bonds with Ser147 and Tyr145. Tyr250 was recently shown to be responsible for emergence of the CVA24v/1985 pandemic [98], and it was also shown to provide hydrogen bond formation with Neu5Ac in its clockwise rotated (cw) form due to flexibility of the VP1 structure (not shown here) [90].

532

Nongluk Sriwilaijaroen and Yasuo Suzuki

Crystal structures of T1L σ1-GM3 complex [116] and T3D-α2-3-sialyllactose complex [117] are shown in Fig. 3e and Fig. 3f, respectively. Both crystals show an elongated trimeric fiber of the outer-capsid protein σ1 that carries the Sia-binding sites in different regions: on the head domain of T1L σ1 and on the body domain of T3D σ1. The structure of T1L σ1 in complex with the GM3 trisaccharide of Neu5Acα2-3Galβ1-4Glc- (α2-3-sialyllactose) indicated that only Neu5Ac sugar forms hydrogen bonds with Asn353, Thr355, Ser370, Gln371, and Thr373. The lactose moiety of the GM3 trisaccharide does not make contact with the binding pocket. The crystal structure of the T1L σ1-GM2 complex (not shown here) revealed that in addition to the terminal Neu5Ac moiety, the other terminal GalNAc moiety of GM2 provides additional optimal contact via van der Waals interactions in the binding pocket. This suggested that additional GalNAc at the terminal plays an important role in contact with the T1L σ1 binding pocket, explaining why T1Lσ1 has higher binding preference for GM2 than for GM3 in a glycan array binding assay [116]. This is different from interactions between T3D σ1 and α2-3-sialyllactose [117] in that the α2-3-sialyllactose structure in the T3D σ1 pocket adopts a topology different from the topology seen in the T1L σ1 pocket. There are two key sets of T3D σ1 residues having direct hydrogen bonds with α2-3-sialyllactose: (1) Ile201, Arg202, Leu203, and Gly205 anchor with Neu5Ac and (2) Ser195 and Gly196 make optimal contacts with the third Glc. These demonstrate different binding preferences of T1L σ1 and T3D σ1 to sialyl glycans that contribute to different tropism of these viruses in the CNS of newborn mice studied for reovirus pathogenesis; T1L virus is mainly found in ependymal cells, whereas T3D virus is commonly found in neurons [178]. Also, these findings may be useful for manipulation of these viruses to specifically infect target cells for therapeutic applications. The HAdV-D37 knob does not bind to immobilized gangliosides including GD1a, but a glycan part of GD1a can be docked into the HAdV-D37 knob, and it is thought that HAdV-D37 binds to the GD1a glycan part-like structure on glycoproteins rather than to the GD1a ganglioside. The crystal structure of the HAdV-D37 fiber knob in complex with GD1a oligosaccharide was then determined [138]. Figure 3g shows the specific interactions between GD1a oligosaccharide, Neu5Acα2-3Galβ1-3GalNAcβ1-4 (Neu5Acα2-3)Galβ1-4Glc-, and an HAdV-D37 trimeric fiber capsid protein that protrudes from another capsid protein, a penton base. The two terminal Neu5Ac moieties of the GD1a glycan were found in different protomers (shown in salmon and sky blue colors) in Sia-binding sites in the HAdV-D37 knob. The HAdV-D37 knob residues, Tyr312, Pro317, and Lys345, of each protomer provide direct hydrogen bond formation with each terminal Neu5Ac. Lys345 in one protomer (sky blue color) also provides indirect

Enveloped and Non-enveloped Sia-Binding Viruses

533

hydrogen bond formation with GalNAc-3 via a water molecule (not shown). In silico substitution of Lys345 to Ala suggested and an in vitro experiment confirmed a crucial role of Lys345 in Sia-binding specificity of the HAdV-D37 knob. The Lys345Ala mutation leads to almost complete abolishment of binding of the HAdV-D37 knob to cells [138], indicating that the ability of this protein to bind to Sia on the cell surface is important for HAdVD37 knob-cell binding. The importance of Sia for HAdV-D37 infection was verified when HAdV-D37 infection of human corneal epithelial cells was decreased in the presence of multivalent Sia conjugated to human serum albumin compared to that in the presence of a monovalent Sia-conjugated inhibitor [179]. This finding highlights the possibility of development of multivalent Sia-containing antiviral drugs for specific treatment of HAdVD37-infected conjunctival and corneal cells. The AAV1 crystal was soaked with an α2-3-sialyl-LacdiNAc trisaccharide, Neu5Acα2-3GalNAcβ1-4GlcNAc (30 SLDN), a top hit glycan bound with AAV1 determined by a glycan array. Only the negatively charged C1 carboxylate of the terminal Neu5Ac was found to form hydrogen bonds with Asn447 and Arg448 residues in a Sia-binding pocket at the base of the protrusions of VP3 monomers (Fig. 3h). Residues Ser268, Asp270, Asn271, Asn447, Arg448, Ser472, V473 Asn500, Thr502, and Trp503, which are near the Neu5Ac 4 A˚, were considered to provide potential Neu5Ac binding contacts [180]. The AAV1-Neu5Ac interactions were used as information for rational structural engineering of AAV1 and AAV6 vectors (only 6 of a total of 736 residues being different) in order to improve therapeutic efficacy. For example, site-directed mutagenesis substitution of some of these residues forming the Sia-binding pocket indicated that the S472R mutant can increase binding to Sia for both AAV1 and AAV6 [180]. One of the six residues that differ between AAV1 and AAV6, Glu-531 for AAV1 (shown as a purple stick in Fig. 3h) but Lys-531 for AAV6, appeared to be responsible for the binding ability of AAV6 to heparin sulfate proteoglycan (HSPG). This binding ability of AAV6, but not that of AAV1, is one factor contributing to the difference in tissue tropism (Table 2) of these two viruses [181]. From glycan array screening of 258 synthetic physiologically relevant oligosaccharides, GM1 oligosaccharide, Galβ13GalNAcβ1-4(Neu5Acα2-3)Galβ1-4Glc-, produced the highest binding signal with recombinantly produced SV40 capsid protein 1, SV40 VP1. The GM1 oligosaccharide is found in a shallow groove at the SV40 VP1 capsid outer surface in a crystal of VP1 pentamer-GM1 oligosaccharide (Fig. 3i). Clearly, there are two critical sets of VP1 amino acid residues, (1) Ser68 and Gln84, which directly interact with Gal- of terminal Galβ1-3GalNAcβ1via hydrogen bond formation, and (2) Gln62, Ser68, Asn272, Ser274, and Thr276, which directly form hydrogen bonds with

534

Nongluk Sriwilaijaroen and Yasuo Suzuki

terminal Neu5Acα2-, on the Galβ1-4Glc stem of GM1 oligosaccharide. SV40 VP1–GM1 interactions are used as a model for comparison with sialyl oligosaccharide–VP1 interactions of JCPyV and BKPyV, viruses in the same family that are closely related with 74% amino acid identity among VP1 proteins of these three viruses, and for studies on their cellular tropism [170]. For example, structural comparison of SV40 VP1-GM1(one Neu5Ac- terminal and one Gal- terminal) interactions with BKPyV VP1-GD3 (one Neu5Acα2-8Neu5Ac- terminal) interactions suggested that the amino acid at position 68 may regulate selective oligosaccharide affinity [171]. It appeared that a single point mutation at this site from Lys in BKPyV VP1 to Ser in SV40 VP1 enables BKPyV to change binding preference from GD3 to GM1 oligosaccharide, which was confirmed by an in vitro binding assay and by cell culture. This finding highlights the plasticity of the viral binding site leading to a change of receptor-binding specificity, which could trigger a change of cellular tropism and pathogenicity [171]. However, in the future, SV40 VP1–GM1 interactions may be used for designing antiviral drugs against SV40 infection due to an increase in reports of SV40 association with human cancers including human brain tumors, bone cancers, malignant mesothelioma, and non-Hodgkin’s lymphoma [165].

6

Conclusions, Perspectives, and Future Directions Sialoglycoproteins and/or gangliosides are important for infection, pathogenesis, and transmission of several enveloped and non-enveloped viruses that use their spike glycoproteins and viral capsid proteins, respectively, as viral Sia-binding lectins. While some viral Sia-binding lectins may bind to Sia-containing materials, such as mucus, in order to persist in and pass through environments for transmission, most grasp host cell surface Sias (like doorknobs) either as sole, primary, or co-receptors, leading to opening of the host plasma or endosomal membrane to allow the release of the viral genome into the host cell for multiplication. While Sia-binding therapeutic viruses are useful for prevention and/or treatment of animal and human diseases, Sia-binding pathogens continue to threaten the health and life of humans and/or economic animals as epidemics and/or pandemics (including influenza A viruses, EV70, and CVA24v). For the development of effective and specific Sia-binding therapeutic viruses and for effective control of Sia-binding pathogens, we reviewed the history up to recent data for sialyl glycans and binding of viruses to Sia, and we generated tables of Sia-binding viruses giving a roadmap that primarily displays links of viral lectins, Sia-binding preference, disease, host range, tissue tropism, and entry pathway in an attempt to identify similarities and differences of these viruses as well as what is still

Enveloped and Non-enveloped Sia-Binding Viruses

535

unknown and needs to further studies. It is remarkable that (1) most of the enveloped viruses in the families Orthomyxoviridae, Paramyxoviridae, and Coronaviridae have both Sia receptorbinding activity and Sia receptor-destroying enzyme (RDE) activity. (1–1) Why do some of these viruses have both functions on the same molecule, while others have the functions on separate molecules? (1–2) Why do Sia-binding enveloped MERS-CoV and TGEV and non-enveloped viruses (even though the genus Siadenovirus contains a putative sialidase homologue gene for which the function is not known [143]) not encode an RDE? How can these viruses be released from traps by decoy Sia receptors such as those on mucins? (1–3) While influenza A viruses must acquire the ability to bind to α2-6Sia human-type receptors for efficient transmission among humans, why do NAs of all influenza A viruses keep cleavage specificity for α2-3Sia? How can these viruses be released/spread from α2-6Sia-decoy mucins or infected cells? Further studies on receptor-binding and receptor-destroying structures, functions and evolution could lead to answers and consequently open avenues for efficient control of virus spread/transmission. In addition, we can summarize that (2) sialyl glycan-binding preferences of viruses are associated with glycans presented on tissues. For example, TGEV, which is a swine intestinal virus, prefers binding to Neu5Gc over Neu5Ac present in pigs, while human-adapted viruses, such as influenza A viruses, have reduced preference for binding to Neu5Gc, which is rarely detected in normal human tissues. (2–1) The targets of these Sia-binding viruses are most often eyes, respiratory system, intestine, and nervous system implying that these organs contain high levels of Sia. Thus, identification of sialyl glycan structures, being viral receptors, on these tissues of various animals should lead to an understanding of viral host and tissue tropism. (3) Viruses with a broad range of hosts, such as influenza A viruses, coronaviruses, EMCV, caliciviruses, and rotaviruses, have zoonotic potential, and surveillance of these viruses should be maintained and plans for responses to newly emerging zoonotic diseases should be made. (4) It is important to realize that some states of different virus infections may cause the same disease (such as red eyes possibly caused by influenza A viruses, EV70, CAV24v, and HAdV-D). (5) While the entry pathway of Sia-binding enveloped viruses is direct fusion on the host plasma membrane or fusion with the endosomal membrane, the entry pathway of Sia-binding non-enveloped viruses is endocytosis that is followed by endosomal permeabilization/lysis as a result of proteolysis and/or capsid conformational change (picornaviruses [84], caliciviruses [109], reoviruses [111], and adenoviruses [131])/lipolysis (parvoviruses [146, 148]) or that is followed by the ERAD pathway (polyomaviruses [166]), except for EMCV, which possibly has direct penetration of its genome through the plasma membrane. (6) Molecular and structural studies have indicated that Sia-binding sites are

536

Nongluk Sriwilaijaroen and Yasuo Suzuki

usually on the heads of viral lectins except for reovirus σ1 capsid protein of T3D strain carrying Sia-binding sites on its body. It remains unknown why the Sia-binding sites are on the T3D σ1 body and on the T1L σ1 head, whereas the JAM-A-binding sites of both viruses are on their σ1 heads. Integrated analyses of similarities and differences in Sia-binding viruses in the above fields and viral lectin–sialyl glycan interactions will provide useful data for the design and development of new tools for combating pathogen infections and improvement of therapeutic viruses for therapeutic applications as follows. (1) Many simple tests have been established for detection of change in viral receptor-binding preference of viruses with pandemic potential such as viral NA-based detection [27] and immunochromatographic-based detection [182]. These detections can indicate whether α2-3/α2-6Sia-binding viruses are present in clinical samples or not and their binding specificity to α2-3Sia or α2-6Sia receptors. Due to the possibility of co-infections of Sia-binding pathogens in the samples, especially poultry stool samples, further improvements in the methods for detection, such as development of tags for labeling expected individual viruses in the sample, are needed. A receptor-binding-based diagnostic method that is able to identify a virus at the level of family, genus, subtype, species, or strain will be more specific and informative than current detection methods for viral identification, surveillance in pandemic preparedness, and treatment. (2) Viral lectins are potential targets because they are required for the crucial first stage of the virus life cycle, and several antiviral lectins have been developed, but most of them act against common pathogens causing widespread severe and life-threatening diseases in humans such as sialylmimetics against rotavirus infection [183], trivalent sialic acid-based inhibitors to treat EV-D68 infections [184], and 60 SLN-lipo PGA (Neu5Acα26Galβ1-4GlcNAcβ1-eicosanoyl chain poly-α-L-glutamic acid) against influenza epidemics and pandemics [185]. Some viruses are Sia-independent viruses, such as rotavirus human strains K8, KU, MO, and Wa [121] and EV-D68 strains 947, 1348, and 742 [95]. Some Sia-binding pathogens, such as MERS-CoV, CVA24v, JCPyV, and MCPyV, use more than one receptor for infection. Also some viruses including SV40, BKPyV, and JCPyV in the family Polyomaviridae can use an alternative receptor for infection. These facts should be taken into consideration for drug design and treatment. In addition, sialyl glycans are important for both host and Sia-dependent viruses, and antiviral drug design should thus be selective to viral lectins compared to counterpart host lectins in order to reduce host toxicity in treatment. (2–1) The use of the antiviral lectin 60 SLN-lipo PGA, which inhibits influenza infection via inhibition of HAs (influenza lectins) in attachment to cell surface Sia receptors, appeared to synergize with either of the two FDA (Food and Drug Administration)-approved NA

Enveloped and Non-enveloped Sia-Binding Viruses

537

inhibitors (oseltamivir carboxylate and zanamivir) [185], suggesting that antiviral lectins may be used in combination with an antiviral receptor-destroying enzyme, anti-NA/anti-esterase, for powered up inhibitory activities, minimized toxicity, and delayed development of resistance for potential treatment. (2–2) Some compounds, such as mumefural and its derivative [186] and Neu5Ac3αF-DSPE (C-3-fluorinated sialyl distearoylphosphatidylethanolamine) [187], inhibit influenza virus infection via inhibition of both HA and NA functions, suggesting that a molecule with dual inhibitory functions of binding and releasing can be designed as a new antiviral chemotype. (2–3) Some different viruses cause infection of the same cell and share Sia-binding specificity. For example, EV70, CAV24v, and HAdV-D can bind to and infect corneal cells through α2-3Neu5Ac-containing glycans. Thus, it is possible to design and develop broad-spectrum Sia-based antiviral drugs against Sia-binding pathogens infecting the same cells. (3) There is a Chinese saying “use the enemy to kill the enemy.” Some Sia-binding viruses, such as NDV [188], reoviruses [189], adenovirus vectors pseudotyped with fibers from HAdV-D [190], and AAVs [191], have been investigated for treatment of human diseases. Based on their oncolytic properties that preferentially infect and lyse cancer cells (so-called oncolytic virotherapy), reovirus T3D has been developed to be pelareorep (Reolysin®), which was approved by the FDA in 2015 for treatment of malignant glioma and in 2017 for treatment of metastatic breast cancer. It has been continued to be investigated for treatment of other cancers and cell proliferative disorders including malignant melanoma [189]. The lack of pathogenicity of AAVs allows them to be extensively investigated as gene-therapy vectors (called viral gene therapy). Alipogene tiparvovec (Glybera®) was developed as an AAV1 vector and was approved in 2012 by the European Commission for treatment of familial lipoprotein lipase deficiency (LPLD) in adult patients who have severe pancreatitis despite a strict low-fat diet [192]. Normally, low abundance of viral receptors on abnormal cells that need to be treated is a key limitation of the use of oncolytic viruses and viral vectors to treat human diseases. Thus, modification of viruses to efficiently bind to receptors present on the target cells is critical for virus infection. Typically, Sia-binding therapeutic viruses including reoviruses, AAVs, and vectors pseudotyped with HAdV-D fiber have more than one binding site, binding to a Sia receptor and another receptor(s). The design of a therapeutic virus for efficient binding to at least two receptors may help to prevent a mutation of the host receptor for escape from viral transfection/ transduction. Also, a virus that efficiently binds to two receptors could efficiently infect the host cell and be transmitted specifically among the target host cells. There is evidence that in addition to using ICAM-1 as an essential receptor, CVA24v, which emerged from CVA24 in 1970, caused the AHC pandemic in 1985 after

538

Nongluk Sriwilaijaroen and Yasuo Suzuki

adaptation of its VP1 to efficiently bind to Sia, an attachment receptor supporting ICAM-1-mediated infection of CVA24v [98]. It has also been shown that Sia-dependent rotavirus strains grow more efficiently than Sia-independent rotavirus strains in Sia-containing cells [121]. Thus, further analysis of receptor structures on host cells that need to be treated, experiments on virus binding preferences to sialyl glycan structures found on the host cells, and X-ray crystallographic studies of virus–sialyl glycan interactions coupled with structure-guided mutagenesis are needed to provide blueprints for engineering viral proteins to bind efficiently to specific sialyl glycans on the target host cells. This increased virus-Sia receptor affinity in cooperation with other receptor binding on the target cell could confine the virus to a specific plasma membrane of target cells and increase viral entry efficiency. Finally, although most viruses have the ability to adapt to their environment, continued and sustainable improvements in surveillance/diagnostic tools and antivirals would lead to eradication of viral pathogens such as in the case of smallpox virus, which has been declared to be eradicated since 1980. With continued improvements in research technology, it will be possible to generate recombinant therapeutic viruses with more efficient binding to dual receptors for synergistic internalization into target cells. This could prevent changes in cellular proteins to resist virus entry, reduce undesired toxicity to normal cells, and increase target cell specificity and transfection/transduction efficiency, leading to improvement of these viruses for treatment of animal and human diseases. References 1. Schauer R, Keirn S, Reuter G et al (1995) Biochemistry and role of sialic acids. In: Rosenberg A (ed) Biology of the sialic acids. Plenum Press, New York, pp 7–67 2. Yamakawa T (1987) History of ganglioside research. In: Rahmann H (ed) Gangliosides and modulation of neuronal functions. Springer-Verlag, Berlin, Heidelberg, New York, pp 3–9 3. Gottschalk A (1951) N-Substituted isoglucosamine released from mucoproteins by the influenza virus enzyme. Nature 167:845–847 4. Heimer R, Meyer K (1956) Studies on sialic acid of submaxillary mucoid. Proc Natl Acad Sci U S A 42:728–734 5. Blix FG, Gottschalk A, Klenk E (1957) Proposed nomenclature in the field of neuraminic and sialic acids. Nature 179:1088 6. Yamakawa T, Suzuki S (1951) The chemistry of the lipids of posthemolytic residue or stroma of erythrocytes. J Biochem 38:199–212

7. Kuhn R, Brossmer R, Schulz W (1954) The prosthetic group of the mucoproteins of cow colostrum. Chem Ber 87:123–127 8. Nadano D, Iwasaki M, Endo S et al (1986) A naturally occurring deaminated neuraminic acid, 3-deoxy-D-glycero-D-galacto-nonulosonic acid (KDN). J Biol Chem 261:11550–11557 9. Varki A (1992) Diversity in the sialic acids. Glycobiology 2:25–40 10. Sriwilaijaroen N, Kondo S, Yagi H et al (2012) Bovine milk whey for preparation of natural N-glycans: structural and quantitative analysis. Open Glycosci 5:41–50 11. Yasukawa Z, Sato C, Kitajima K (2005) Inflammation-dependent changes in α2,3,α2,6-, and α2,8-sialic acid glycotopes on serum glycoproteins in mice. Glycobiology 15:827–837 12. Toivonen S, Aitio O, Renkonen O (2001) α2,3-Sialylation of terminal GalNAc β1-3Gal determinants by ST3Gal II reveals the

Enveloped and Non-enveloped Sia-Binding Viruses multifunctionality of the enzyme. J Biol Chem 276:37141–37148 13. Sumiyoshi W, Nakakita S, Miyanishi N et al (2012) Hypersialylated type-I lactosaminecontaining N-glycans found in Artiodactyla sera are potential xenoantigens. Glycobiology 22:1031–1041 14. Kubo H, Irie A, Inagaki F et al (1990) Gangliosides from the eggs of the sea urchin, Anthocidaris crassispina. J Biochem 108:185–192 15. Yoshida Y, Kojima N, Kurosawa N et al (1995) Molecular cloning of Siaα2,3Galβ 1,4GlcNAc α2,8-sialyltransferase from mouse brain. J Biol Chem 270:14628–14633 16. Slomiany A, Slomiany BL (1978) Structures of the acidic oligosaccharides isolated from rat sublingual glycoprotein. J Biol Chem 253:7301–7306 17. Takasaki S, Yamashita K, Suzuki K et al (1979) The sugar chains of cold-insoluble globulin. J Biol Chem 254:8548–8553 18. Fukuda MN, Dell A, Oates JE et al (1985) Embryonal lactosaminoglycan. J Biol Chem 260:6623–6631 19. Kitazume S, Kitajima K, Inoue S et al (1994) Identification of polysialic acid-containing glycoprotein in the jelly coat of sea urchin eggs. J Biol Chem 269:22712–22718 20. Varki A, Schnaar RL, Schauer R (2015) Sialic acids and other nonulosonic acids. In: Varki A, Cummings RD, Esko JD, Stanley P, Hart GW, Aebi M, Darvill AG, Kinoshita T, Packer NH, Prestegard JH, Schnaar RL, Seeberger PH (eds) Essentials of glycobiology. Cold Spring Harbor, New York, pp 179–195 21. Suzuki Y, Ito T, Suzuki T et al (2000) Sialic acid species as a determinant of the host range of influenza A viruses. J Virol 74:11825–11831 22. Irie A, Koyama S, Kozutsumi Y et al (1998) The molecular basis for the absence of Nglycolylneuraminic acid in humans. J Biol Chem 273:15866–15871 23. Martin MJ, Rayner JC, Gagneux P et al (2005) Evolution of human-chimpanzee differences in malaria susceptibility: relationship to human genetic loss of N-glycolylneuraminic acid. Proc Natl Acad Sci U S A 102:12819–12824 24. Sriwilaijaroen N, Nakakita SI, Kondo S et al (2018) N-Glycan structures of human alveoli provide insight into influenza A virus infection and pathogenesis. FEBS J 285:1611–1634 25. Schwarzkopf M, Knobeloch K-P, Rohde E et al (2002) Sialylation is essential for early

539

development in mice. Proc Natl Acad Sci U S A 99:5267–5270 26. Tajima Y, Uyama E, Go S et al (2005) Distal myopathy with rimmed vacuoles: impaired Oglycan formation in muscular glycoproteins. Am J Pathol 166:1121–1130 27. Sriwilaijaroen N, Suzuki Y (2014) A simple viral neuraminidase-based detection for highthroughput screening of viral hemagglutininhost receptor specificity. Methods Mol Biol 1200:107–120 28. Wang CC, Chen JR, Tseng YC et al (2009) Glycans on influenza hemagglutinin affect receptor binding and immune response. Proc Natl Acad Sci U S A 106:18137–18142 29. Portela A, Jones LD, Nuttall P (1992) Identification of viral structural polypeptides of Thogoto virus (a tick-borne orthomyxo-like virus) and functions associated with the glycoprotein. J Gen Virol 73(Pt 11):2823–2830 30. Cook JD, Sultana A, Lee JE (2017) Structure of the infectious salmon anemia virus receptor complex illustrates a unique binding strategy for attachment. Proc Natl Acad Sci U S A 114: E2929–E2936 31. Wasik BR, Barnard KN, Ossiboff RJ et al (2017) Distribution of O-acetylated sialic acids among target host tissues for influenza virus. mSphere 2:e00379-16 32. Sriwilaijaroen N, Suzuki Y (2012) Molecular basis of the structure and function of H1 hemagglutinin of influenza virus. Proc Jpn Acad Ser B Phys Biol Sci 88:226–249 33. Lugovtsev VY, Smith DF, Weir JP (2009) Changes of the receptor-binding properties of influenza B virus B/Victoria/504/2000 during adaptation in chicken eggs. Virology 394:218–226 34. Ni F, Mbawuike IN, Kondrashkina E et al (2014) The roles of hemagglutinin Phe-95 in receptor binding and pathogenicity of influenza B virus. Virology 450–451:71–83 35. Sriwilaijaroen N, Suzuki Y (2014) Molecular basis of a pandemic of avian-type influenza virus. Methods Mol Biol 1200:447–480 36. Suzuki Y (2005) Sialobiology of influenza: molecular mechanism of host range variation of influenza viruses. Biol Pharm Bull 28:399–408 37. Ito T, Suzuki Y, Suzuki T et al (2000) Recognition of N-glycolylneuraminic acid linked to galactose by the α2,3 linkage is associated with intestinal replication of influenza A virus in ducks. J Virol 74:9300–9305 38. Sriwilaijaroen N, Kondo S, Yagi H et al (2009) Analysis of N-glycans in embryonated chicken egg chorioallantoic and amniotic cells

540

Nongluk Sriwilaijaroen and Yasuo Suzuki

responsible for binding and adaptation of human and avian influenza viruses. Glycoconj J 26:433–443 39. WHO website (2009) Clinical management of human infection with pandemic (H1N1) 2009: revised guidance. https://www.who. int/csr/resources/publications/swineflu/ clinical_management_h1n1.pdf 40. WHO website (2018) Influenza (seasonal). https://www.who.int/news-room/factsheets/detail/influenza-(seasonal) 41. Kuiken T, Taubenberger JK (2008) Pathology of human influenza revisited. Vaccine 26:D59–D66 42. Shinya K, Ebina M, Yamada S et al (2006) Avian flu: influenza virus receptors in the human airway. Nature 440:435–436 43. Sriwilaijaroen N, Kondo S, Yagi H et al (2011) N-Glycans from porcine trachea and lung: predominant Neu5Acα2-6Gal could be a selective pressure for influenza variants in favor of human-type receptor. PLoS One 6: e16302 44. Walther T, Karamanska R, Chan RW et al (2013) Glycomic analysis of human respiratory tract tissues and correlation with influenza virus infection. PLoS Pathog 9: e1003223 45. Watanabe Y, Ibrahim MS, Ellakany HF et al (2011) Acquisition of human-type receptor binding specificity by new H5N1 influenza virus sublineages during their emergence in birds in Egypt. PLoS Pathog 7:e1002068 46. Wu P, Peng Z, Fang VJ et al (2016) Human infection with influenza A(H7N9) virus during 3 major epidemic waves, China, 2013–2015. Emerg Infect Dis 22:964–972 47. Bowman AS, Nelson SW, Page SL et al (2014) Swine-to-human transmission of influenza A (H3N2) virus at agricultural fairs, Ohio, USA, 2012. Emerg Infect Dis 20:1472–1480 48. Suzuki T, Portner A, Scroggs RA et al (2001) Receptor specificities of human respiroviruses. J Virol 75:4604–4613 49. Tappert MM, Smith DF, Air GM (2011) Fixation of oligosaccharides to a surface may increase the susceptibility to human parainfluenza virus 1, 2, or 3 hemagglutininneuraminidase. J Virol 85:12146–12159 50. Alymova IV, Portner A, Mishin VP et al (2012) Receptor-binding specificity of the human parainfluenza virus type 1 hemagglutinin-neuraminidase glycoprotein. Glycobiology 22:174–180 51. Kubota M, Takeuchi K, Watanabe S et al (2016) Trisaccharide containing α2,3-linked sialic acid is a receptor for mumps virus. Proc Natl Acad Sci U S A 113:11579–11584

52. Suzuki Y, Suzuki T, Matsunaga M et al (1985) Gangliosides as paramyxovirus receptor. Structural requirement of sialooligosaccharides in receptors for hemagglutinating virus of Japan (Sendai virus) and Newcastle disease virus. J Biochem 97:1189–1199 53. Suzuki Y, Hirabayashi Y, Suzuki T et al (1985) Occurrence of O-glycosidically peptide-linked oligosaccharides of poly-N-acetyllactosamine type (erythroglycan II) in the I-antigenically active Sendai virus receptor sialoglycoprotein GP-2. J Biochem 98:1653–1659 54. Markwell MA, Paulson JC (1980) Sendai virus utilizes specific sialyloligosaccharides as host cell receptor determinants. Proc Natl Acad Sci U S A 77:5693–5697 55. Zaitsev V, von Itzstein M, Groves D et al (2004) Second sialic acid binding site in Newcastle disease virus hemagglutininneuraminidase: implications for fusion. J Virol 78:3733–3741 56. Porotto M, Fornabaio M, Greengard O et al (2006) Paramyxovirus receptor-binding molecules: engagement of one site on the hemagglutinin-neuraminidase protein modulates activity at the second site. J Virol 80:1204–1213 57. Aguilar HC, Henderson BA, Zamora JL et al (2016) Paramyxovirus glycoproteins and the membrane fusion process. Curr Clin Microbiol Rep 3:142–154 58. Porotto M, Fornabaio M, Kellogg GE et al (2007) A second receptor binding site on human parainfluenza virus type 3 hemagglutinin-neuraminidase contributes to activation of the fusion mechanism. J Virol 81:3216–3228 59. Russell CJ, Hurwitz JL (2016) Sendai virus as a backbone for vaccines against RSV and other human paramyxoviruses. Expert Rev Vaccines 15:189–200 60. Nakanishi M, Otsu M (2012) Development of Sendai virus vectors and their potential applications in gene therapy and regenerative medicine. Curr Gene Ther 12:410–416 61. Brown VR, Bevins SN (2017) A review of virulent Newcastle disease viruses in the United States and the role of wild birds in viral persistence and spread. Vet Res 48:68 62. Tayeb S, Zakay-Rones Z, Panet A (2015) Therapeutic potential of oncolytic Newcastle disease virus: a critical review. Oncolytic Virother 4:49–62 63. Mahon PJ, Mirza AM, Iorio RM (2011) Role of the two sialic acid binding sites on the Newcastle disease virus HN protein in triggering the interaction with the F protein required

Enveloped and Non-enveloped Sia-Binding Viruses for the promotion of fusion. J Virol 85:12079–12082 64. Langereis MA, Zeng Q, Gerwig GJ et al (2009) Structural basis for ligand and substrate recognition by torovirus hemagglutinin esterases. Proc Natl Acad Sci U S A 106:15897–15902 65. Schwegmann-Wessels C, Herrler G (2006) Sialic acids as receptor determinants for coronaviruses. Glycoconj J 23:51–58 66. Zeng Q, Langereis MA, van Vliet AL et al (2008) Structure of coronavirus hemagglutinin-esterase offers insight into corona and influenza virus evolution. Proc Natl Acad Sci U S A 105:9065–9069 67. Bakkers MJ, Zeng Q, Feitsma LJ et al (2016) Coronavirus receptor switch explained from the stereochemistry of protein-carbohydrate interactions and a single mutation. Proc Natl Acad Sci U S A 113:E3111–E3119 68. Li W, Hulswit RJG, Widjaja I et al (2017) Identification of sialic acid-binding function for the Middle East respiratory syndrome coronavirus spike glycoprotein. Proc Natl Acad Sci U S A 114:E8508–E8517 69. Schultze B, Krempl C, Ballesteros ML et al (1996) Transmissible gastroenteritis coronavirus, but not the related porcine respiratory coronavirus, has a sialic acid (N-glycolylneuraminic acid) binding activity. J Virol 70:5634–5637 70. Liang R, Wang L, Zhang N et al (2018) Development of small-molecule MERS-CoV inhibitors. Viruses 10:721 71. Smits SL, Lavazza A, Matiz K et al (2003) Phylogenetic and evolutionary relationships among torovirus field variants: evidence for multiple intertypic recombination events. J Virol 77:9567–9577 72. Zanoni R, Weiss M, Peterhans E (1986) The haemagglutinating activity of Berne virus. J Gen Virol 67:2485–2488 73. Cui J, Li F, Shi ZL (2019) Origin and evolution of pathogenic coronaviruses. Nat Rev Microbiol 17:181–192 74. Schultze B, Gross HJ, Brossmer R et al (1991) The S protein of bovine coronavirus is a hemagglutinin recognizing 9-O-acetylated sialic acid as a receptor determinant. J Virol 65:6232–6237 75. Hulswit RJG, Lang Y, Bakkers MJG et al (2019) Human coronaviruses OC43 and HKU1 bind to 9-O-acetylated sialic acids via a conserved receptor-binding site in spike protein domain A. Proc Natl Acad Sci U S A 116:2681–2690

541

76. Cornelissen LA, Wierda CM, van der Meer FJ et al (1997) Hemagglutinin-esterase, a novel structural protein of torovirus. J Virol 71:5277–5286 77. Smits SL, Gerwig GJ, van Vliet AL et al (2005) Nidovirus sialate-O-acetylesterases: evolution and substrate specificity of coronaviral and toroviral receptor-destroying enzymes. J Biol Chem 280:6933–6941 78. Desforges M, Desjardins J, Zhang C et al (2013) The acetyl-esterase activity of the hemagglutinin-esterase protein of human coronavirus OC43 strongly enhances the production of infectious virus. J Virol 87:3097–3107 79. de Groot RJ (2006) Structure, function and evolution of the hemagglutinin-esterase proteins of corona- and toroviruses. Glycoconj J 23:59–72 80. Wang N, Shi X, Jiang L et al (2013) Structure of MERS-CoV spike receptor-binding domain complexed with human receptor DPP4. Cell Res 23:986–993 81. Zeng H, Goldsmith CS, Maines TR et al (2013) Tropism and infectivity of influenza virus, including highly pathogenic avian H5N1 virus, in ferret tracheal differentiated primary epithelial cell cultures. J Virol 87:2597–2607 82. WHO website (2019) Middle East respiratory syndrome coronavirus (MERS-CoV) – Oman. https://www.who.int/csr/don/04march-2019-mers-oman/en/ 83. Fry EE, Tuthill TJ, Harlos K et al (2010) Crystal structure of equine rhinitis A virus in complex with its sialic acid receptor. J Gen Virol 91:1971–1977 84. Wang D, Guo H, Chang J et al (2018) Andrographolide prevents EV-D68 replication by inhibiting the acidification of virus-containing endocytic vesicles. Front Microbiol 9:2407 85. Nokhbeh MR, Hazra S, Alexander DA et al (2005) Enterovirus 70 binds to different glycoconjugates containing α2,3-linked sialic acid on different cell lines. J Virol 79:7087–7094 86. Mistry N, Inoue H, Jamshidi F et al (2011) Coxsackievirus A24 variant uses sialic acidcontaining O-linked glycoconjugates as cellular receptors on human ocular cells. J Virol 85:11283–11290 87. Guy M, Chilmonczyk S, Cruciere C et al (2009) Efficient infection of buffalo rat liverresistant cells by encephalomyocarditis virus requires binding to cell surface sialic acids. J Gen Virol 90:187–196

542

Nongluk Sriwilaijaroen and Yasuo Suzuki

88. Zhou L, Luo Y, Wu Y et al (2000) Sialylation of the host receptor may modulate entry of demyelinating persistent Theiler’s virus. J Virol 74:1477–1485 89. Liu Y, Sheng J, Baggen J et al (2015) Sialic acid-dependent cell entry of human enterovirus D68. Nat Commun 6:8865 90. Zocher G, Mistry N, Frank M et al (2014) A sialic acid binding site in a human picornavirus. PLoS Pathog 10:e1004401 91. Kumar AS, Kallio P, Luo M et al (2003) Amino acid substitutions in VP2 residues contacting sialic acid in low-neurovirulence BeAn virus dramatically reduce viral binding and spread of infection. J Virol 77:2709–2716 92. Hammoumi S, Guy M, Eloit M et al (2012) Encephalomyocarditis virus may use different pathways to initiate infection of primary human cardiomyocytes. Arch Virol 157:43–52 93. Tuthill TJ, Groppelli E, Hogle JM et al (2010) Picornaviruses. Curr Top Microbiol Immunol 343:43–89 94. Carocci M, Bakkali-Kassimi L (2012) The encephalomyocarditis virus. Virulence 3:351–367 95. Baggen J, Thibaut HJ, Staring J et al (2016) Enterovirus D68 receptor requirements unveiled by haploid genetics. Proc Natl Acad Sci U S A 113:1399–1404 96. Nilsson EC, Jamshidi F, Johansson SM et al (2008) Sialic acid is a cellular receptor for coxsackievirus A24 variant, an emerging virus with pandemic potential. J Virol 82:3061–3068 97. Alexander DA, Dimock K (2002) Sialic acid functions in enterovirus 70 binding and infection. J Virol 76:11265–11272 98. Baggen J, Hurdiss DL, Zocher G et al (2018) Role of enhanced receptor engagement in the evolution of a pandemic acute hemorrhagic conjunctivitis virus. Proc Natl Acad Sci U S A 115:397–402 99. Anderson K, Bond CW (1987) Biological properties of mengovirus: characterization of avirulent, hemagglutination-defective mutants. Arch Virol 93:31–49 100. Tavakkol A, Burness AT (1990) Evidence for a direct role for sialic acid in the attachment of encephalomyocarditis virus to human erythrocytes. Biochemistry 29:10684–10690 101. Jin YM, Pardoe IU, Burness AT et al (1994) Identification and characterization of the cell surface 70-kilodalton sialoglycoprotein(s) as a candidate receptor for encephalomyocarditis virus on human nucleated cells. J Virol 68:7308–7319

102. Rydell GE, Nilsson J, Rodriguez-Diaz J et al (2009) Human noroviruses recognize sialyl Lewis x neoglycoprotein. Glycobiology 19:309–320 103. Wegener H, Mallagaray A, Schone T et al (2017) Human norovirus GII.4(MI001) P dimer binds fucosylated and sialylated carbohydrates. Glycobiology 27:1027–1037 104. Taube S, Perry JW, Yetming K et al (2009) Ganglioside-linked terminal sialic acid moieties on murine macrophages function as attachment receptors for murine noroviruses. J Virol 83:4092–4101 105. Tan M, Wei C, Huang P et al (2015) Tulane virus recognizes sialic acids as cellular receptors. Sci Rep 5:11784 106. Stuart AD, Brown TD (2007) α2,6-linked sialic acid acts as a receptor for Feline calicivirus. J Gen Virol 88:177–186 107. Kim DS, Hosmillo M, Alfajaro MM et al (2014) Both α2,3- and α2,6-linked sialic acids on O-linked glycoproteins act as functional receptors for porcine Sapovirus. PLoS Pathog 10:e1004172 108. Gerondopoulos A, Jackson T, Monaghan P et al (2010) Murine norovirus-1 cell entry is mediated through a non-clathrin-, non-caveolae-, dynamin- and cholesterol-dependent pathway. J Gen Virol 91:1428–1438 109. Stuart AD, Brown TD (2006) Entry of feline calicivirus is dependent on clathrin-mediated endocytosis and acidification in endosomes. J Virol 80:7500–7509 110. Soliman M, Kim DS, Kim C et al (2018) Porcine sapovirus Cowden strain enters LLC-PK cells via clathrin- and cholesteroldependent endocytosis with the requirement of dynamin II. Vet Res 49:92 111. Danthi P, Holm GH, Stehle T et al (2013) Reovirus receptors, cell entry, and proapoptotic signaling. Adv Exp Med Biol 790:42–71 112. Tyler KL, Barton ES, Ibach ML et al (2004) Isolation and molecular characterization of a novel type 3 reovirus from a child with meningitis. J Infect Dis 189:1664–1675 113. Tyler KL, McPhee DA, Fields BN (1986) Distinct pathways of viral spread in the host determined by reovirus S1 gene segment. Science 233:770–774 114. Eggers HJ, Gomatos PJ, Tamm I (1962) Agglutination of bovine erythrocytes: a general characteristic of reovirus type 3. Proc Soc Exp Biol Med 110:879–881 115. Bulai T, Bratosin D, Pons A et al (2003) Diversity of the human erythrocyte membrane sialic acids in relation with blood groups. FEBS Lett 534:185–189

Enveloped and Non-enveloped Sia-Binding Viruses 116. Reiss K, Stencel JE, Liu Y et al (2012) The GM2 glycan serves as a functional coreceptor for serotype 1 reovirus. PLoS Pathog 8: e1003078 117. Reiter DM, Frierson JM, Halvorson EE et al (2011) Crystal structure of reovirus attachment protein sigma1 in complex with sialylated oligosaccharides. PLoS Pathog 7: e1002166 118. Rubin DH, Eaton MA, Anderson AO (1986) Reovirus infection in adult mice: the virus hemagglutinin determines the site of intestinal disease. Microb Pathog 1:79–87 119. Morin MJ, Warner A, Fields BN (1996) Reovirus infection in rat lungs as a model to study the pathogenesis of viral pneumonia. J Virol 70:541–548 120. Diaz-Salinas MA, Romero P, Espinosa R et al (2013) The spike protein VP4 defines the endocytic pathway used by rotavirus to enter MA104 cells. J Virol 87:1658–1663 121. Ciarlet M, Ludert JE, Iturriza-Gomara M et al (2002) Initial interaction of rotavirus strains with N-acetylneuraminic (sialic) acid residues on the cell surface correlates with VP4 genotype, not species of origin. J Virol 76:4087–4095 122. Svensson L (1992) Group C rotavirus requires sialic acid for erythrocyte and cell receptor binding. J Virol 66:5582–5585 123. Rolsma MD, Gelberg HB, Kuhlenschmidt MS (1994) Assay for evaluation of rotaviruscell interactions: identification of an enterocyte ganglioside fraction that mediates group A porcine rotavirus recognition. J Virol 68:258–268 124. Haselhorst T, Fleming FE, Dyason JC et al (2009) Sialic acid dependence in rotavirus host cell invasion. Nat Chem Biol 5:91–93 125. Guo CT, Nakagomi O, Mochizuki M et al (1999) Ganglioside GM(1a) on the cell surface is involved in the infection by human rotavirus KUN and MO strains. J Biochem 126:683–688 126. Fleming FE, Bohm R, Dang VT et al (2014) Relative roles of GM1 ganglioside, N-acylneuraminic acids, and α2β1 integrin in mediating rotavirus infection. J Virol 88:4558–4571 127. Rolsma MD, Kuhlenschmidt TB, Gelberg HB et al (1998) Structure and function of a ganglioside receptor for porcine rotavirus. J Virol 72:9079–9091 128. Yu X, Dang VT, Fleming FE et al (2012) Structural basis of rotavirus strain preference toward N-acetyl- or N-glycolylneuraminic acid-containing receptors. J Virol 86:13456–13466

543

129. Kim JY, Kim DS, Seo JY et al (2017) Glycanspecificity of four neuraminidase-sensitive animal rotavirus strains. Vet Microbiol 207:159–163 130. Houldcroft CJ, Beale MA, Sayeed MA et al (2018) Identification of novel adenovirus genotype 90 in children from Bangladesh. Microb Genom 4:e000221 131. Meier O, Boucke K, Hammer SV et al (2002) Adenovirus triggers macropinocytosis and endosomal leakage together with its clathrinmediated uptake. J Cell Biol 158:1119–1131 132. Yousuf MA, Zhou X, Mukherjee S et al (2013) Caveolin-1 associated adenovirus entry into human corneal cells. PLoS One 8: e77462 133. Hiroi S, Morikawa S, Takahashi K et al (2013) Molecular epidemiology of human adenoviruses D associated with epidemic keratoconjunctivitis in Osaka, Japan, 2001–2010. Jpn J Infect Dis 66:436–438 134. Arnberg N, Kidd AH, Edlund K et al (2000) Initial interactions of subgenus D adenoviruses with A549 cellular receptors: sialic acid versus alpha(v) integrins. J Virol 74:7691–7693 135. Baker A, Mundy R, Davies J et al (2019) Adenovirus serotype 26 utilises sialic acid bearing glycans as a primary cell entry receptor. bioRxiv:580076 136. Chandra N, Frangsmyr L, Imhof S et al (2019) Sialic acid-containing glycans as cellular receptors for ocular human adenoviruses: implications for tropism and treatment. Viruses 11:395 137. Burmeister WP, Guilligay D, Cusack S et al (2004) Crystal structure of species D adenovirus fiber knobs and their sialic acid binding sites. J Virol 78:7727–7736 138. Nilsson EC, Storm RJ, Bauer J et al (2011) The GD1a glycan is a cellular receptor for adenoviruses causing epidemic keratoconjunctivitis. Nat Med 17:105–109 139. Wu E, Fernandez J, Fleck SK et al (2001) A 50-kDa membrane protein mediates sialic acid-independent binding and infection of conjunctival cells by adenovirus type 37. Virology 279:78–89 140. Lenman A, Liaci AM, Liu Y et al (2015) Human adenovirus 52 uses sialic acidcontaining glycoproteins and the coxsackie and adenovirus receptor for binding to target cells. PLoS Pathog 11:e1004657 141. Lenman A, Liaci AM, Liu Y et al (2018) Polysialic acid is a cellular receptor for human adenovirus 52. Proc Natl Acad Sci U S A 115:E4264–E4273

544

Nongluk Sriwilaijaroen and Yasuo Suzuki

142. Seiradake E, Henaff D, Wodrich H et al (2009) The cell adhesion molecule “CAR” and sialic acid on human erythrocytes influence adenovirus in vivo biodistribution. PLoS Pathog 5:e1000277 143. Singh AK, Berbis MA, Ballmann MZ et al (2015) Structure and sialyllactose binding of the carboxy-terminal head domain of the fibre from a siadenovirus, turkey adenovirus 3. PLoS One 10:e0139339 144. Luo Y, Kleiboeker S, Deng X et al (2013) Human parvovirus B19 infection causes cell cycle arrest of human erythroid progenitors at late S phase that favors viral DNA replication. J Virol 87:12766–12775 145. Huang LY, Halder S, Agbandje-McKenna M (2014) Parvovirus glycan interactions. Curr Opin Virol 7:108–118 146. Vihinen-Ranta M, Suikkanen S, Parrish CR (2004) Pathways of cell infection by parvoviruses and adeno-associated viruses. J Virol 78:6709–6714 147. Lo´pez-Bueno A, Rubio M-P, Bryant N et al (2006) Host-selected amino acid changes at the sialic acid binding pocket of the parvovirus capsid modulate cell binding affinity and determine virulence. J Virol 80:1563–1573 148. Venkatakrishnan B, Yarbrough J, Domsic J et al (2013) Structure and dynamics of adeno-associated virus serotype 1 VP1-unique N-terminal domain and its role in capsid trafficking. J Virol 87:4974–4984 149. Popa-Wagner R, Sonntag F, Schmidt K et al (2012) Nuclear translocation of adenoassociated virus type 2 capsid proteins for virion assembly. J Gen Virol 93:1887–1898 150. Blackburn SD, Cline SE, Hemming JP et al (2005) Attachment of bovine parvovirus to O-linked α2,3 neuraminic acid on glycophorin A. Arch Virol 150:1477–1484 151. Boisvert M, Fernandes S, Tijssen P (2010) Multiple pathways involved in porcine parvovirus cellular entry and trafficking toward the nucleus. J Virol 84:7782–7792 152. Lo¨fling J, Lyi S, Parrish C et al (2013) Canine and feline parvoviruses preferentially recognize the non-human cell surface sialic acid N-glycolylneuraminic acid. Virology 440:89–96 153. Nu¨esch JPF, Lacroix J, Marchini A et al (2012) Molecular pathways: rodent parvoviruses—mechanisms of oncolysis and prospects for clinical cancer treatment. Clin Cancer Res 18:3516–3523 154. Etingov I, Itah R, Mincberg M et al (2008) An extension of the Minute Virus of Mice tissue tropism. Virology 379:245–255

155. Daya S, Berns KI (2008) Gene therapy using adeno-associated virus vectors. Clin Microbiol Rev 21:583–593 156. Ehrhardt A, Xu H, Kay MA (2003) Episomal persistence of recombinant adenoviral vector genomes during the cell cycle in vivo. J Virol 77:7689–7695 157. Mietzsch M, Broecker F, Reinhardt A et al (2014) Differential adeno-associated virus serotype-specific interaction patterns with synthetic heparins and other glycans. J Virol 88:2991–3003 158. Kaludov N, Brown KE, Walters RW et al (2001) Adeno-associated virus serotype 4 (AAV4) and AAV5 both require sialic acid binding for hemagglutination and efficient transduction but differ in sialic acid linkage specificity. J Virol 75:6884–6893 159. Wu Z, Miller E, Agbandje-McKenna M et al (2006) α2,3 and α2,6 N-linked sialic acids facilitate efficient binding and transduction by adeno-associated virus types 1 and 6. J Virol 80:9093–9103 160. Nonnenmacher M, Weber T (2012) Intracellular transport of recombinant adenoassociated virus vectors. Gene Ther 19:649–658 161. Schmidt M, Chiorini JA (2006) Gangliosides are essential for bovine adeno-associated virus entry. J Virol 80:5516–5522 162. Crispino G, Galindo Ramirez F, Campioni M et al (2017) In vivo genetic manipulation of inner ear connexin expression by bovine adeno-associated viral vectors. Sci Rep 7:6567 163. Jamboti JS (2016) BK virus nephropathy in renal transplant recipients. Nephrology (Carlton) 21:647–654 164. Prado JCM, Monezi TA, Amorim AT et al (2018) Human polyomaviruses and cancer: an overview. Clinics (Sao Paulo) 73:e558s 165. Vilchez RA, Kozinetz CA, Arrington AS et al (2003) Simian virus 40 in human cancers. Am J Med 114:675–684 166. Zhao L, Marciano AT, Rivet CR et al (2016) Caveolin- and clathrin-independent entry of BKPyV into primary human proximal tubule epithelial cells. Virology 492:66–72 167. Liu CK, Wei G, Atwood WJ (1998) Infection of glial cells by the human polyomavirus JC is mediated by an N-linked glycoprotein containing terminal α(2-6)-linked sialic acids. J Virol 72:4643–4649 168. Neu U, Maginnis MS, Palma AS et al (2010) Structure-function analysis of the human JC polyomavirus establishes the LSTc pentasaccharide as a functional receptor motif. Cell Host Microbe 8:309–319

Enveloped and Non-enveloped Sia-Binding Viruses 169. Maginnis MS, Nelson CD, Atwood WJ (2015) JC polyomavirus attachment, entry, and trafficking: unlocking the keys to a fatal infection. J Neurovirol 21:601–613 170. Neu U, Woellner K, Gauglitz G et al (2008) Structural basis of GM1 ganglioside recognition by simian virus 40. Proc Natl Acad Sci U S A 105:5219–5224 171. Neu U, Allen SA, Blaum BS et al (2013) A structure-guided mutation in the major capsid protein retargets BK polyomavirus. PLoS Pathog 9:e1003688 172. Schowalter RM, Pastrana DV, Buck CB (2011) Glycosaminoglycans and sialylated glycans sequentially facilitate Merkel cell polyomavirus infectious entry. PLoS Pathog 7: e1002161 173. Buch MH, Liaci AM, O’Hara SD et al (2015) Structural and functional analysis of murine polyomavirus capsid proteins establish the determinants of ligand recognition and pathogenicity. PLoS Pathog 11:e1005104 174. Tsai B, Gilbert JM, Stehle T et al (2003) Gangliosides are receptors for murine polyoma virus and SV40. EMBO J 22:4346–4355 175. Geoghegan EM, Pastrana DV, Schowalter RM et al (2017) Infectious entry and neutralization of pathogenic JC polyomaviruses. Cell Rep 21:1169–1179 176. Chen Y, Tan M, Xia M et al (2011) Crystallography of a Lewis-binding norovirus, elucidation of strain-specificity to the polymorphic human histo-blood group antigens. PLoS Pathog 7:e1002152 177. Bangaru S, Zhang H, Gilchuk IM et al (2018) A multifunctional human monoclonal neutralizing antibody that targets a unique conserved epitope on influenza HA. Nat Commun 9:2669 178. Forrest JC, Dermody TS (2003) Reovirus receptors and pathogenesis. J Virol 77:9109–9115 179. Johansson SM, Nilsson EC, Elofsson M et al (2007) Multivalent sialic acid conjugates inhibit adenovirus type 37 from binding to and infecting human corneal epithelial cells. Antivir Res 73:92–100 180. Huang LY, Patel A, Ng R et al (2016) Characterization of the adeno-associated virus 1 and 6 sialic acid binding site. J Virol 90:5219–5230 181. Wu Z, Asokan A, Grieger JC et al (2006) Single amino acid changes can influence titer, heparin binding, and tissue tropism in different adeno-associated virus serotypes. J Virol 80:11393–11397

545

182. Watanabe Y, Ito T, Ibrahim MS et al (2015) A novel immunochromatographic system for easy-to-use detection of group 1 avian influenza viruses with acquired human-type receptor binding specificity. Biosens Bioelectron 65:211–219 183. Fazli A, Bradley SJ, Kiefel MJ et al (2001) Synthesis and biological evaluation of sialylmimetics as rotavirus inhibitors. J Med Chem 44:3292–3301 184. Spjut S, Qian W, Bauer J et al (2011) A potent trivalent sialic acid inhibitor of adenovirus type 37 infection of human corneal cells. Angew Chem Int Ed Engl 50:6519–6521 185. Sriwilaijaroen N, Suzuki K, Takashita E et al (2015) 6SLN-lipo PGA specifically catches (coats) human influenza virus and synergizes neuraminidase-targeting drugs for human influenza therapeutic potential. J Antimicrob Chemother 70:2797–2809 186. Sriwilaijaroen N, Kadowaki A, Onishi Y et al (2011) Mumefural and related HMF derivatives from Japanese apricot fruit juice concentrate show multiple inhibitory effects on pandemic influenza A (H1N1) virus. Food Chem 127:1–9 187. Guo CT, Sun XL, Kanie O et al (2002) An Oglycoside of sialic acid derivative that inhibits both hemagglutinin and sialidase activities of influenza viruses. Glycobiology 12:183–190 188. Sa´nchez D, Pelayo R, Medina LA et al (2016) Newcastle disease virus: potential therapeutic application for human and canine lymphoma. Viruses 8:3 189. Eissa IR, Bustos-Villalobos I, Ichinose T et al (2018) The current status and future prospects of oncolytic viruses in clinical trials against melanoma, glioma, pancreatic, and breast cancers. Cancers (Basel) 10:356 190. Coughlan L, Alba R, Parker AL et al (2010) Tropism-modification strategies for targeted gene delivery using adenoviral vectors. Viruses 2:2290–2355 191. Miesbach W, Meijer K, Coppens M et al (2018) Gene therapy with adeno-associated virus vector 5–human factor IX in adults with hemophilia B. Blood 131:1022–1031 192. Bu¨ning H (2013) Gene therapy enters the pharma market: the short story of a long journey. EMBO Mol Med 5:1–3 193. Bird SW, Kirkegaard K (2015) Escape of non-enveloped virus from intact cells. Virology 479-480:444–449 194. Yamauchi Y, Helenius A (2013) Virus entry at a glance. J Cell Sci 126:1289–1295

Chapter 48 Hemagglutinin Inhibitors are Potential Future Anti-Influenza Drugs for Mono- and Combination Therapies Nongluk Sriwilaijaroen and Yasuo Suzuki Abstract Infections by H1-H16 influenza A viruses require sufficient binding of viral hemagglutinins (HAs) to specific target receptors, glycoconjugates bearing sialyl sugar chains, on the host cell surface. Synthesized sialyl sugar chains targeting sialyl sugar-binding sites in HAs that are immutable as long as the virus does not switch to a different host species might therefore be highly effective candidate drugs for inhibition of the initial required step of virus entry. In this chapter, we describe the following aspects of updated sialyl sugar chains as influenza A virus HA inhibitors (HAIs): (1) mode of terminal sialyl-galactose linkage, (2) molecular length and structure of sialyl glycan receptors, (3) multivalent sialyl sugar chain dimension, (4) clustering of sialyl sugar chains on macromolecular scaffolds, and (5) enhancement of the stability of sialyl sugar chain HA inhibitors. We also discuss about the use of HAI-based combinations that should be considered for future influenza therapy. Key words Influenza, Hemagglutinin, Hemagglutinin inhibitor, Hemagglutinin-based drug development, Hemagglutinin inhibitor combination therapy, Neuraminidase, Sialyl sugar chain

1

Introduction Despite extensive surveillance, influenza caused by H1-H16 influenza A viruses remains one of the world’s most widely distributed zoonoses that can lead to a pandemic with a huge death toll. Also, despite the availability of an annual vaccine, influenza epidemics continue to occur worldwide with substantial morbidity and mortality rates that have a great socioeconomic burden. Two envelope spike glycoproteins of influenza A viruses, hemagglutinin (HA) and neuraminidase (NA), cooperatively work for successful infection and spread. Both spikes recognize sialyl sugar chains expressed on the host cell surface but are involved in different mechanisms. While HAs are responsible for the attachment of a viral particle to

The original version of this chapter was revised. The correction to this chapter is available at https://doi.org/ 10.1007/978-1-0716-0430-4_60 Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_48, © Springer Science+Business Media, LLC, part of Springer Nature 2020

547

548

Nongluk Sriwilaijaroen and Yasuo Suzuki

Fig. 1 Replication cycle of influenza A viruses and target steps of inhibitors that have been approved in Japan (laninamivir and favipiravir) or have been approved both in Japan and by the FDA (all of the remaining inhibitors listed in the figure). When an influenza virus enters the host body, (1) the virus must “penetrate” through a mucus layer in the respiratory system of poultry and humans, intestine of birds, or eyes of humans by viral NA spikes. The NA spikes not only cleave Sias from the decoy receptors on mucus but also on cilia and cellular glycoconjugates. This allows the virus to access functional epithelial cellular receptors (target receptors). Clinically approved NA inhibitors, zanamivir, oseltamivir carboxylate (OC), peramivir, and laninamivir, inhibit this process. (2) The virus binds to target receptors on the surface of an epithelial cell by the viral HA spikes. While the HA spikes of human influenza viruses have binding preference to α2-6Sia, the HA spikes of avian viruses preferentially bind to α2-3Sia (attachment). No HA inhibitors have been approved so far. (3) Binding of the virus to target receptors mediates virus entry by the process of “endocytosis.” (4) The endosome has gradually increasing protons (lowering pH) actively pumped from the cytosol by ATPase proton pumps. (4.1) At a low pH in the endosome, viral HA0 proteins are stable (noninfectious form), but their cleaved products, disulfide-linked HA1-HA2 proteins that are generated by a tissue-specific expressed host trypsinlike serine endoprotease for HA0 with a monobasic (Arg/Lys#Gly) cleavage site or by ubiquitously expressed host proteases for HA0 of HPAI H5 and H7 with a multibasic cleavage site (furin and proprotein convertase 5/6 (PC5/6) in the trans-Golgi network, which are subtilisin-like proteases, cleave an Arg-X-Arg/Lys-Arg#Gly cleavage site, while mosaic serine protease large form (MSPL) and its splice variant transmembrane protease serine 13 (TMPRSS13) in the plasma membrane, which are type II transmembrane serine proteases, can cleave an Arg/Lys-X-Arg/Lys-Arg#Gly cleavage site by which X ¼ a nonbasic residue) [48], are sensitive, causing HA conformational change. This conformational change results in exposure of their fusion peptides, which are immediately inserted into the hydrophobic endosomal membrane, leading to “fusion” of the viral and endosomal membranes and pore formation. (4.2) Concurrently, at a low pH in the endosome, the viral pH-gated proton M2 channels unidirectionally conduct protons from the endosome into the interior of the virus, resulting in release of the viral ribonucleoproteins (vRNPs) from M1 proteins (genome unpacking) into the host cytosol through the fusion pore (uncoating). Viral acidification, unpacking, and uncoating processes

Hemagglutinin Inhibitor - Drug Discovery

549

a host cell and entry into the cell through specific HA-receptor interactions [1], NAs catalyze the hydrolysis of such receptors to allow the virus to escape from host decoy receptors and from the infected cell (Figs. 1 and 2, I) [2]. Several influenza virus drugs that act at different steps of the viral life cycle have been developed and clinically approved as shown in Fig. 1. However, unfortunately, influenza virus strains that are resistant to the M2 channel inhibitor adamantanes [3], NA inhibitors including oseltamivir carboxylate (OC) and peramivir [3, 4], and/or the cap-dependent endonuclease inhibitor baloxavir marboxil [5] have been detected. On the other hand, HA plays a critical role in viral binding and entry processes, and the sialyl sugar chain receptor-binding sites in the HA molecule of seasonal and pandemic influenza viruses are highly conserved. Therefore, the receptor-binding site of HA is an excellent target for developing universal anti-influenza drugs that preclude virus entry at the initial step of the viral life cycle (Fig. 1). However, no HA inhibitors have been approved so far. The rapid rate of mutation of epidemic influenza A viruses by antigenic variation results in loss of the efficiency of a vaccine for protection, and production of an effective vaccine for the first wave of the rapid spread of a new pandemic virus remains a difficult challenge due to the difficulty in prediction of newly emerged ä Fig. 1 (continued) are inhibited by adamantanes, M2 channel inhibitors. (5) The released vRNPs, which are complexes of ()vRNA, NP, and trimeric polymerases (viral RNA polymerase basic protein 1 (PB1), basic protein 2 (PB2), and acidic protein (PA)), are quickly imported into the nucleus. The ()vRNAs are “transcribed” to viral (+)mRNAs by PB1s, which require 50 -capped RNA fragments as primers from host mRNAs snatched by the cap-recognizing and binding PB2s, and the cap-endonucleolytically cleaving PAs. In contrast, the ()vRNAs are “replicated” to (+) complementary RNAs (cRNAs) that are used as templates for further synthesis of ()vRNAs by PB1s, which do not require a primer. Baloxavir marboxil, a PA inhibitor, inhibits the cap snatching process required for viral transcription, while favipiravir, a PB1 inhibitor, inhibits viral transcription and replication. (6) The viral (+)mRNAs are exported for “synthesis of viral proteins on cytosolic ribosomes and viral membrane proteins on endoplasmic reticulum (ER)-bound ribosomes” by the host “translation” machinery. The early synthesized viral proteins, which are required for viral transcription and replication, are imported back to the nucleus. (7) The viral membrane proteins HA and NA are “glycosylated” (glycans on their structures have effects on their structures and functions), while the small membrane protein M2 may be “glycosylated” (a glycan on its structure does not affect its function) [49], during their journey by “the secretory pathway” (M2, critical for equilibration of pH between the acidic trans-Golgi network and the host cytosol) through the ER and the Golgi apparatus to the plasma membrane. (8) All viral components are trafficked to the assembly site on the apical plasma membrane (assembly). (9) It is possible that a lipid raft is a platform for virus budding. A lipid raft clusters (concentrates) viral glycoproteins, HA and NA spikes. Interactions of HA and NA proteins with M1 proteins beneath the membrane and interactions of M1 proteins with M1 proteins, with M2 proteins, and with NP of vRNPs cause concentration of viral components. M2 integral membrane proteins are localized and alter membrane curvature, leading to membrane “budding” and finally scission. (10) However, the HA spikes of progeny viruses remain attached to the host sialylglycoconjugates. The NA spikes are required for removal of Sias from the host cell membrane. The progeny viruses are “released” and spread. NA inhibitors, including zanamivir, oseltamivir carboxylate, peramivir, and laninamivir, inhibit this release process. When the HA spikes of progeny viruses attach to receptors on new cells, the viruses continue another cycle [2]

Fig. 2 Proposed cellular mechanisms of the single action of 6’-sialyllactosaminelipo polyglutamic acid (6SLN-lipo PGA) and single action of a neuraminidase inhibitor (NAI) (OC/zanamivir) and the synergistic action of 6SLN-lipo PGA (hemagglutinin inhibitor (HAI)) and an NAI against human influenza A virus infection [39]. (I) In the absence of any inhibitor, both HAs (blue) and NAs (yellow) are active. HAs bind to sialylated host receptors, either target receptors or decoy receptors, whereas NAs cleave sialic acid (Sia) residues of host receptors, preventing binding of HAs to the receptors. Balanced HA-NA activities (1 U of virus infection) are critical for releasing the virus from decoy receptors on a non-endocytosis site and for virus binding to target receptors on an endocytosis site to achieve successful infection. (II) When 6SLN-lipo PGAs (selfintramolecular aggregates in solution due to removal of acyl chains to the nonpolar interior, 6SLN (sugar symbols) with an acyl chain (wave line) on PGA (gray double-solid line) linked via a lysine (green rectangle)) are present, they arrest HAs (HA inactive, gray) and hinder virus binding to the host cell through potentially cooperative inhibitions of (1) 6SLNs competing with host receptors for binding to the viral HAs, (2) acyl chains stabilizing the 6SLN-lipo PGA compound-HA complex via hydrophobic interactions (intermolecular aggregates), and (3) PGAs stabilizing the viral membrane sterically from attachment to the negatively charged sialylated host cell membrane (steric stabilization). In this case, lower HA binding to host receptors than balanced HA-NA activities occurs (hence, less than 1 U of virus infection). (III) When NAIs (red circle-backslash) are present, they compete with sialylated host receptors for the active sites of NAs, block sialidase activity (NA inactive, gray), and consequently prevent release of HAs from their binding (freezing). In this case, lower virus release from host receptors including decoy receptors than balanced HA-NA activities occurs (hence, less than 1 U of virus infection). (IV) In the presence of both 6SLN-lipo PGAs and NAIs, 6SLN-lipo PGAs bind HAs (HA inactive, gray) and hinder virus attachment to host receptors (resulting in lower HA binding to host receptors). NAIs bind to NAs (NA inactive, gray) and thus prevent discharge of HAs from their binding to either 6SLN-lipo PGAs (resulting in lower virus release from HAIs) or sialylated host receptors (resulting in lower virus release from host receptors including decoy receptors). These lead to a greater reduction in virus entry and release than that in the case of summation of the effects of the individual inhibitors (hence, much less than 1 U of infection)

Hemagglutinin Inhibitor - Drug Discovery

551

viral antigens. Therefore, antiviral agents that can be applied to rapidly mutable seasonal and pandemic influenza viruses should be developed. Recently, the functional balance of influenza A virus HAs, sialic acid (Sia) receptor-binding lectins, and NAs, Sia receptordestroying enzymes, has been reported [6] to have critical importance for viral movement: rolling and penetrating the respiratory/ mucus layer carrying sialylated decoy receptors to infect underlying epithelial cells and releasing from infected cells and rolling over epithelial surfaces possibly being involved in cell-to-cell spread to the next appropriate susceptible target cells. Combination therapy with a receptor-binding HA inhibitor (HAI) and a receptordestroying NA inhibitor (NAI) would be effective for future influenza control because the acquisition of simultaneous drug-resistant mutations of the viral proteins of HA and NA would be rare. In this chapter, we describe the functional and structural aspects of sialyl sugar chain HA inhibitors and provide structural information that will facilitate the rational design of novel influenza virus entry inhibitors targeting influenza A virus HA. We also focus on synergetic combination therapy using an HAI with clinically approved NAI.

2

Hemagglutinin Inhibitors (HAIs) of Influenza A Viruses One of the most important roles of influenza virus HA is binding to specific sialyl sugar chain receptors on host cell membranes and facilitating entry of viruses into host cells by the process of HA-mediated membrane fusion. Thus, HA-targeted agents may be potential anti-influenza drugs. In recent years, it has been shown that the HAs of avian influenza virus and human influenza virus have different receptor-binding specificities for recognition. We have been investigating the molecular structure and function of sialyl sugar chains of the HA receptor, and we found that clustered multivalent sialyl sugar chains mimicking receptors on appropriate scaffolds inhibit HA-mediated influenza virus infection. Sialyl sugar chain HA inhibitors that we have developed in the past few decades are listed in Table 1. From the data shown in Table 1, we propose the following key points for the development of HAIs of currently circulating influenza A viruses.

2.1 Mode of Terminal Sialyl-Galactose Linkage of Sialyl Sugar Chain Receptors That Bind to Influenza A Virus HA

Human and avian influenza A virus HAs recognize different sialic acid-containing receptors, referred to as human type (Neu5Acα26Gal) and avian type (Neu5Acα2-3Gal), respectively. Thus, inhibitors should have terminal Neu5Acα2-6Gal sialyl sugar chains for human seasonal and pandemic influenza A viruses and terminal Neu5Acα2-3Gal sialyl sugar chains for avian influenza A viruses. It appears that HA inhibitors that carry human-type or avian-type

[28] The avian influenza virus bound strongly to either Neu5Acα2-3LacNAc- or Neu5Acα2-3Lac-carrying γPGA. The human virus directly bound to the glycopolypeptide with terminal Neu5Acα2-6LacNAc preferentially over Neu5Acα2-6Lac The glycopolypeptide with terminal Neu5Acα2-6LacNAc inhibited hemagglutination and infection by human influenza viruses with higher activities than those of a natural inhibitor, serum glycoprotein fetuin containing both α2-6Neu5Ac and α2-3Neu5Ac

Multivalent sialyl oligosaccharides (Neu5Acα2-6 (or α2-3) LacNAc) with a γ-polyglutamic acid (γPGA) backbone

[27]

[21]

[14]

[19]

Infection of human H1pdm and swine influenza virus was inhibited by either short or long Neu5Acα2-6(LacNAc)1,3αPGA. Infection of seasonal 29-year circulating H1 and 40-year circulating H3 viruses was efficiently inhibited by long Neu5Acα2-6(LacNAc)3‐PGA) A reassortant H5 HA/H1N1pdm virus comprising H5 HA with four mutations of N158D/N224K/Q226L/T318I preferentially recognized α2-6Neu5Ac human-type receptors and can caused efficient respiratory droplet transmission in ferrets having exhaled aerosol plumes similar to that in humans. Neu5Acα2-6Gal-containing inhibitors are potential candidate drugs against humanadapted HAs Infection of human and swine influenza viruses was strongly inhibited by α2-6Neu5Ac glycopolymers carrying LacNAc repeats, but infection of avian and equine viruses was not inhibited by any α2-6Neu5Ac glycopolymer The Neu5Acα2-6Galβ1-4GlcNAcβ- carrying αPGA bound to human influenza viruses and significantly inhibited virus infection

Sialyllactosamine-carrying polyglutamic acids Multivalent sialyl oligosaccharides (Neu5Acα2-6 (or α2-3) (LacNAc)1-3) with an αpolyglutamic acid (αPGA) backbone

Ref.

Remark

Sialyl glyco compound

Table 1 Sialyl glyco compounds targeting influenza A virus hemagglutinin 552 Nongluk Sriwilaijaroen and Yasuo Suzuki

Lyso-GM3-Lys(BODIPY) αPGA (as shown on the right side) [34] efficiently inhibited direct binding of egg-adapted influenza A/PR/8/34 (H1N1) virus to the GM3 ganglioside

GM3 oligosaccharides ((Neu5Acα2-3Galβ1-4Glc)-carrying polyglutamic acids Multivalent lysogangliosides (Neu5Acα2-3Galβ1-4Glcβ1-sphingosine), each linked to a BODIPY-C3 dye via a lysine residue with an α-polyglutamic acid backbone (lyso-GM3-Lys(BODIPY) αPGA)

(continued)

6SLN-lipo PGA inhibited human-adapted influenza viruses, [39] both H1N1pdm/2009 and seasonal H3N2/2004 viruses, and its inhibitory activity was synergized by NA-targeting drugs tested, both zanamivir (Relenza®) and oseltamivir carboxylate (active form of Tamiflu®)

Multivalent sialyl oligosaccharides (Neu5Acα2-6LacNAc), each linked to an eicosanoyl chain, with an α-polyglutamic acid backbone (6SLN-lipo PGA)

Hemagglutinin Inhibitor - Drug Discovery 553

Remark

Ref.

Sialyllactosamine-carrying chitosan The α2-6Neu5Ac oligosaccharides-carrying chitosan Multivalent Neu5Acα2-6LacNAc-containing glycopolymer with a chitosan backbone inhibited human influenza virus hemagglutination with higher activity than that of natural hen’s egg yolk sialoglycopeptides and serum glycoprotein fetuin

[29]

[26] The Neu5Acα2-6Lac- and Neu5Acα2-3Lac-bearing Sialyllactose-carrying polystyrene polystyrene inhibited hemagglutination of egg-adapted Multivalent sialyl oligosaccharides (Neu5Acα2-6 and Neu5Acα2-3Lac) linked via an influenza A/PR/8/34 (H1N1) virus with higher activity amide linkage with a polystyrene backbone than that of the sialyllactose itself. The clustering of sialyl sugar chains could enhance the sensitivity and stabilize the binding activity

Sialyl glyco compound

Table 1 (continued)

554 Nongluk Sriwilaijaroen and Yasuo Suzuki

[33] The Neu5Acα2-3Lac-carrying dumbbell(1)6-amide-type dendrimer (as shown on the left side) showed higher hemagglutination and infection inhibitory activities against egg-adapted influenza A/PR/8/34 (H1N1) virus than did the Neu5Acα2-3Lac units carried by other dendrimers including the dumbbell(1)6-type dendrimer without amide linkages and ball-type, fan-type and pent-type dendrimers with/without amide linkages

Sialyllactose-carrying carbosilane dendrimers

(continued)

[23] A cyclic peptide, cyclo(Ser‐Gly‐Gly‐Gln‐Ser‐His‐Asp)3, was suggested to be an excellent scaffold for carrying GM3 oligosaccharides. This GM3 oligosaccharides-carrying cyclic peptide efficiently inhibited hemagglutination of egg-adapted influenza A/PR/8/34 (H1N1) virus. The cyclic peptide containing tri-dentate carbohydrate units produced greater inhibitory activity than that of the cyclic peptides containing di- or mono-dentate carbohydrate units

GM3 oligosaccharides-carrying cyclic peptides

Hemagglutinin Inhibitor - Drug Discovery 555

F

CO2H

NO2

O

HO

P

O

O

O

O

O (CH2)16CH3 (CH2)16CH3

O

O

CO2H

NO2

Neu5Ac3aOH-para-nitrophenol (pNP)

O HO OH H

OH OH

HO AcHN

Abbreviations: HA hemagglutinin, LacNAc Galβ1-4GlcNAc, Lac lactose: Galβ1-4Glc, n degree of sugar substitution based on degree of polymerization of polyglutamic acids, polystyrene or polyglucosamine (chitosan), NA neuraminidase, pdm: pandemic, PGA polyglutamic acid

O O HO F H Neu5Ac3aF-para-nitrophenol (pNP)

HO AcHN

OH OH

HO

COOH H N O H

Neu5Ac3aF-distearoylphosphatidylethanolamine (DSPE)

HO AcHN

O

[30] The Neu5Ac3αF-DSPE, in which the C-3 axial hydrogen atom was replaced by a fluorine atom, inhibited both the catalytic hydrolysis of NA and the binding activity of HA of influenza A/Aichi/2/68 (H3N2) virus [32] The C-3 modified compounds Neu5Ac3αF-pNP and Neu5Ac3αOH-pNP were found to have potent and selective inhibitory activities against the H3 HA. They showed resistance to hydrolysis by both N1 and N2 NAs

C-3 modified sialic acid-carrying DSPE/pNP

OH OH

In comparison to all sialyl PE derivatives including Neu5Ac- [31] PE (amide) and Neu5Ac-PE (methyl), the (Neu5Ac)2-PE displayed the strongest binding to and inhibitory activities against hemagglutination and infection by H3N2 subtypes but not by the H1N1subtype

Sialic acid-carrying dipalmitoylphosphatidylethanolamine ((Neu5Ac)2-PE)

Ref.

Remark

Sialyl glyco compound

Table 1 (continued)

556 Nongluk Sriwilaijaroen and Yasuo Suzuki

Hemagglutinin Inhibitor - Drug Discovery

557

receptor sialyl sugar chains can also be used as diagnostic and surveillance molecules for detecting human adaptation of avian influenza virus receptor-binding specificity [7–11]. Human influenza viruses show preferential 60 -sialyl-N-lactosamine (Neu5Acα26Galβ1-4GlcNAc, 6SLN) over 60 -sialyl-lactose (Neu5Acα26Galβ1-4Glc), and the α2-6 linkage is disfavored by influenza virus NA cleavage activity and is thus more stable than the α2-3 linkage, whereas it is favored for binding by seasonal and pandemic influenza HAs [1]. Since the emergence of new avian-avian reassortant low pathogenic avian influenza (LPAI) H7N9 viruses in China in 2013, avian H7N9 viruses have become able to cross the species barrier to infect humans with high morbidity and mortality rates. Recently, some of these viruses have become highly pathogenic to poultry and therefore might cause much more severe infections and deaths in humans [12]. It has been reported that only three amino acid mutations (V186K, K193T, G228S) of the recombinant H7 HA can switch its avian-type receptor binding to human-type receptor specificity [13]. Thus, the control and eradication of H7N9 viruses are needed to prevent a possible pandemic. A reassortant H5/H1N1 virus [14], which contains a highly pathogenic avian influenza virus (HPAI) H5 HA with four substitutions, N158D, N224K, Q226L, and T318I, and the remaining 7 gene segments from a 2009 pandemic (pdm) H1N1 virus, exhibited a binding preference shift from avian-type Neu5Acα2-3Gal to human-type Neu5Acα2-6Gal receptors. Also, this reassortant virus appeared to be able to cause droplet transmission in ferrets, which express naturally human-type influenza A virus receptors [15] and possess exhaled aerosol plumes similar to humans. This work revealed four key residues that are responsible for the Neu5Acα23Gal to Neu5Acα2-6Gal receptor shift in HPAI H5 HA-binding specificity for prediction/surveillance/countermeasures of the pandemic potential of isolates. Taken together, the results indicate that Neu5Acα2-6Gal-containing HAIs could be potential candidate drugs against humanadapted, both seasonal and pandemic, influenza A virus HAs. 2.2 Molecular Length and Structure of Sialyl Glycan Receptors

The human respiratory system including the nasopharynx, bronchus, and lungs contains N- and O-linked α2-6 and α2-3 linked sialyl-N-lactosamine (sialyl-LacNAc, SLN) sugar chains with extended LacNAc up to 10 U. These α2-6 and α2-3 sialyl glycans have been reported to be receptors of human and avian influenza viruses, respectively [16–18]. Long-term circulating human H3N2 viruses have highly glycosylated HA heads [19] and receptor-binding amino acid changes, and they have evolved preference for Neu5Acα2-6 receptors with extended poly-N-acetyl-lactosamine (poly-LacNAc, by which LacNAc (LN) is Galβ1-4GlcNAc) chains [19–21]. Our study [19]

558

Nongluk Sriwilaijaroen and Yasuo Suzuki

using multivalent sialyl oligosaccharides (Neu5Acα2-6/α23LN1,3) with an α-polyglutamic acid backbone (Table 1) also showed that H1 pandemic and avian influenza viruses can recognize both short Neu5AcLN1 and long Neu5AcLN3 sugar chains. Therefore, inhibitors may be designed as follows: Neu5Acα26LNn, by which n is the number of LN repeats that should be 3 for effective inhibition against circulating seasonal viruses and next pandemic viruses and short α2-3 sugar chains, in the case of using as monovalent sialyl sugar chain, that could be lower cost than longer chains against avian influenza viruses. 2.3 Multivalent Sialyl Sugar Chain Dimension That Can Reach the Binding Site of HA

The HA spike glycoprotein on the influenza A virus envelope is a homo-trimeric protein glycosylated by host glycosylation machinery, and each subunit contains one sialyloligosaccharide receptorbinding site. X-ray crystallography studies [22] revealing the threedimensional structure of the HA-sialyllactose (Neu5Acα2-3Galβ14Glc) complex showed that the three receptor-binding sites are located on the top of each globular head domain, like an equilateral triangle from the top view, with the distances between binding sites being approximately 40–50 A˚. To achieve a rational molecular design for an HA inhibitor, focusing on the multivalent sialyl sugar chain dimension that can reach the binding sites of HA is necessary. Ohta et al. chemoenzymatically synthesized a series of cyclic peptides presenting three sialyllactose chains, and each terminal Sia stretched out at full length is separated by approximately 50–70 A˚ [23]. The trivalent sialyl ligand-carrying compounds appeared to simultaneously occupy the three shallow binding pockets of the HA trimer using molecular docking and molecular mechanics (MM) of SYBYL. Studies on the inhibitory effects of these compounds on hemagglutination of egg-adapted influenza A/PR/8/34(H1N1) virus showed that the inhibitory effects of the trivalent sialyl ligand-carrying compounds were stronger than the inhibitory effects of bivalent and monovalent ligand-carrying compounds. In addition, α2-3sialyllactose (α2-3SL)-modified three-way junction (3WJ) DNA (three self-assembled ssDNA strands) by which the distance from the core side of α2-3SL of one DNA arm is approximately 60 A˚ from another arm has been reported to effectively inhibit hemagglutination of influenza A/PR/8/34 (H1N1) virus [24]. Sialyl glycans-carrying 3WJ DNA has a possible application as an inhibitor of influenza infection. Recently, modeling data have shown that α2-6SLN2 in each arm of a biantennary N-glycan was too short for bivalent binding to two binding sites of an HA trimer, whereas longer glycans with α26SLN3, α2-6SLN4, and α2-6SLN5 in each arm of a biantennary Nglycan were capable of bivalent binding to an H1 Cal/04/09 HA trimer [20]. By using an influenza receptor glycan microarray with human airway glycans, the 2009 H1N1pdm strain and recent

Hemagglutinin Inhibitor - Drug Discovery

559

highly glycosylated seasonal H3N2 influenza viruses were found to have evolved binding preference for a subset of human-type receptors containing branched (bi- and tri-antennary) glycans with polyLacNAc chains [20]. Therefore, an HAI designed as a trivalent sialyl glycan mimetic that can reach the three receptor-binding sites of HA trimers of influenza viruses could provide for maximum binding to viral HA trimers and subsequently potent inhibition of viral attachment to host cells. 2.4 Clustering of Sialyl Sugar Chains on Macromolecular Scaffolds

Normally, the binding affinity between a sugar chain (e.g., sialyllactose) and influenza virus HA revealed by proton NMR studies is considerably low, in the range of mM [25]. As shown in Table 1, enhanced blocking effects by multivalent glycoligands were found by clustering the sialyloligosaccharides on scaffolds [26] such as poly-L-glutamic acid (PGA) [27, 28], polystyrene [26], chitosan [29], phospholipid (phosphatidylethanolamine) [30–32], and dendrimers [33]. PGA composed of α- (or γ)-L-glutamic acid residues linked together via amide bonds is currently one of the most popular scaffolds for sialyl sugar chains due to its properties including high aqueous solubility for enhancing compound solubility and conformational flexibility for coating the viral surface that prevents the coated virus from accessing the negatively charged sialyl sugar chains on a target host cell. Together with its low toxicity, low immunogenicity, and biodegradability, PGA is suitable as a scaffold for sialyl sugar chains for clinical use [34, 35]. Kamitakahara et al. [34] chose PGA for conjugating with lysoganglioside GM3 (lysoGM3, Neu5Acα2-3Galβ1-4Glcβ1-10 sphingosine), by which the amino functionality of the sphingosine is suitable for connection to the polymer PGA via a lysine residue (Table 1). The amphiphilic nature of sphingosine, which is directly attached to the oligosaccharide moiety, could help present the epitope unit to the surface of the polymer by the formation of either intramolecular (folded) or intermolecular aggregates. Lysoganglioside GM3-carrying PGA showed picomolar inhibition of egg-adapted influenza virus HA binding to the GM3 ganglioside. Thus, clustering of sialyloligosaccharides on an appropriate scaffold will be important to enhance the binding to influenza virus HA.

2.5 Modification of the C-3 Position of Terminal Sialic Acid in Sialyl Glycan Inhibits Catalytic Hydrolysis of Influenza Virus NA and Enhances the Stability of an HA Inhibitor of the Viruses

Sialic acids at the nonreducing terminals of sialyl sugar chains can be cleaved by influenza virus NAs. If sialic acid-carrying inhibitors are resistant to sialidase activity, their effectiveness will be increased. Functional groups of sialic acid residues such as the carboxyl group at C-1; the hydroxyl groups at C-4, C-8, and C-9; and the acetamide group at C-5 have been demonstrated to play an important role in the binding to influenza virus HA. As shown in Table 1, a sialic acid analogue with axial hydrogen at the C-3 position replaced by a fluoride atom or a hydroxyl group resists influenza virus NA and inhibits viral HA and NA functions [30, 32]. Neu5Ac3αF-

560

Nongluk Sriwilaijaroen and Yasuo Suzuki

distearoylphosphatidylethanolamine (DSPE) [30], Neu5Ac3αFpara-nitrophenol ( pNP), and Neu5Ac3αOH-pNP [32] exhibited the most potent and selective inhibition of hemagglutination of the H3 HA subtype (A/Aichi/2/68 (H3N2) virus) but not of hemagglutination of the H1 HA subtype (A/PR/8/34 (H1N1) virus). However, they displayed inhibitory activities toward all tested influenza virus NA subtypes including N1 and N2. Neu5Ac3αF-DSPE tested for inhibition of infection of influenza A/Aichi/2/68 (H3N2) virus in MDCK cells showed IC50 of 5.6 μM. These works indicated that the C-3-modified Neu5Ac-based compounds are a new class of bifunctional anti-influenza drug candidates against HA and NA functions of H3 virus subtypes and are NA inhibitors against all NA virus subtypes. Recently, multivalent S-sialoside protein conjugates [36] have been reported to inhibit hemagglutination and some have been shown to inhibit NA activity. Persistent development of anti-HA and anti-NA bifunctional compounds is crucial for discovery of a potent stable dual function compound against all influenza virus subtypes.

3

Synergizing the HAI Activity by Influenza Virus Neuraminidase-Targeting Drugs Due to the high mutation rates of influenza A viruses possessing 8 ()ssRNA genome segments and RNA polymerase without proofreading, reformulation of the vaccine is required each year and all influenza A viruses circulating in humans, both 1968derived H3N2 and 2009-derived H1N1 viruses, have resistance to clinically approved adamantanes targeting the M2 ion channel integral viral membrane protein. Some of the influenza viruses sampled had resistance to NA inhibitors including OC, peramivir [3] and/or zanamivir, though reports of zanamivir-resistant isolates are very rare [4], and some had resistance to the cap-dependent endonuclease inhibitor baloxavir marboxil [3]. Although resistance to the NAI laninamivir and resistance to the RNA polymerase basic protein 1 (PB1) inhibitor favipiravir have so far not been detected in clinical isolates, resistance to these drugs could emerged at short notice. Thus, there is a need for anti-influenza combination therapies. Although widely used in almost all pathogen infections including bacteria, parasite, and virus infections, combination therapies have not been used for influenza treatment according to influenza antiviral treatment recommendations [37, 38], possibly due to a lack of effective synergism between drugs that are currently used. The NAIs OC and zanamivir remain the main monotherapy drugs recommended for influenza-infected patients. For prevention and delay of both new emergence and spread of drug-resistant strains, there is an urgent need for a new anti-

Hemagglutinin Inhibitor - Drug Discovery

561

influenza drug that can synergize with other drugs, especially NAIs, in influenza combination therapy. We thus developed a new potential hemagglutinin inhibitor, multivalent α2-6SLN oligosaccharides with an acyl chain clustered on an αPGA scaffold via lysine linkages, namely, 6SLN-lipo PGA (see Table 1), for human influenza therapeutic potential [39]. A direct binding assay showed that it specifically binds to HAs of human influenza viruses, both 1968-derived H3N2/2004 and 2009 H1N1pdm viruses. A hemagglutination inhibition assay indicated that it prevents binding of those viruses to erythrocytes. An influenza virus-multiplication inhibition assay indicated that it is a potent inhibitor with IC50 around 300–500 nM against virus multiplication of both seasonal H3N2/2004 and pandemic H1N1/2009 influenza viruses. Isobolographic analysis, which is a reliable tool for assessing the combined effects of a mixture of antimicrobial drugs, showed that 6SLN-lipo PGA has strong synergy with an NAI drug, either OC or zanamivir, and enhances the effect of the NAI drug for complete clearance of the virus from the cell culture. Based on the chemical structure and experimental results, we proposed possible mechanisms of the single action of 6SLN-lipo PGA inhibiting viral attachment to receptors, either host target receptors or host decoy receptors, and single action of an NAI (OC/zanamivir) inhibiting viral discharge from the receptors and synergistic action of 6SLN-lipo PGA (HAI) and an NAI against human influenza A virus infection. As shown in Table 1 and Fig. 2, 6SLN-lipo PGA carries multiple trisaccharide 6SLN moieties acting as multiple lures that mimic human influenza receptors on the host cell surface. Each 6SLN is linked to an acyl chain acting as a hook to generate the best fishing line for fishing an influenza virus carrying about 500 molecules of a trimeric HA acting as mouths that snap the lures and about 100 molecules of a tetrameric NA acting as teeth that cut the lure to discharge the virus. In brief, once human influenza virus HA specifically binds to 6SLN, the 6SLN-lipo PGA molecule undergoes a conformational change leading to burying of its acyl chain in the hydrophobic interior of HA, which would lock the 6SLN-lipo PGA molecule to HA, whereas a flexible PGA coats the viral surface, thus impeding access of the bound virus to the negatively charged sialic acid on the host cell surface (Fig. 2, II). Both OC (shikimic acid analogue) and zanamivir (sialic acid analogue) are competitive inhibitors competing with sialyl receptors for binding to the active sites of NAs, resulting in blockage of sialidase (Sia-cleaving (cutting)) activity, and thus the virus is trapped (frozen) in its binding (Fig. 2, III). Balanced HA-NA activities (Fig. 2, I) are required for viral movement to the endocytosis site of susceptible host cells for invasion. Imbalanced HA-NA activities by inhibiting one of them (Figs. 2, II and III), HA with Sia-attachment activity or NA with Sia-releasing activity, result in reduction of viral invasion or viral spread freely to susceptible host

562

Nongluk Sriwilaijaroen and Yasuo Suzuki

cells, respectively, leading to reduction of viral multiplication. Imbalanced HA-NA activities by inhibiting viral attachment with 6SLN-lipo PGAs (containing specific human HA binding by 6SLNs, hydrophobic locking by acyl chains, and nonspecific viral coating by the PGA) and inhibiting viral release from both host target/decoy receptors and 6SLN-lipo PGAs by OC/zanamivir (Fig. 2, IV) result in a greater reduction in viral invasion and viral spread than the summation of each effect of each inhibitor (synergistic effect). Understanding how 6SLN-lipo PGA action with a unique manner to inhibit virus infection will contribute to further design and development of a safe, effective, and stable HAI for use as HAI-based monotherapy and combination therapy for treatment of human influenza caused by any human influenza A virus including the next pandemic necessary to switch binding preference to α2-6Sia human-type receptors for efficient human-to-human transmission. This work also provides information showing that such an HAI is suitable for use in combination with an NAI drug not only for preventing/delaying drug-resistant development due to different targets of drug actions but also for minimizing toxicity due to the synergistic effect (much lower dose for each inhibitor in combined drug treatment than in single drug treatment for reaching the IC50). Moreover, this synergistic HAI-NAI combination that causes complete clearance of the virus from culture has promising potential for eradication of human influenza.

4

Conclusion In this chapter, we described (1) several structural HAI aspects to develop new effective influenza A virus HA inhibitors carrying sialyl sugar chains and (2) the possibility of a new influenza therapy by the synergistic combination of an HA inhibitor and a lowmolecular-weight NA inhibitor. Based on our review in this chapter including our recent findings of receptor-binding specificity of avian and human influenza viruses [19], current HA inhibitors may be further developed for more efficient universal inhibition of human-adapted HAs of seasonal and pandemic viruses. The compound should carry multiple clusters of triantennary sialyl long LN N-glycans, i.e., Neu5Acα2-6LN3 sugar chains (disfavored by NA cleavage activity), or multiple clusters of trivalent sialyl long LN glycans on a cyclic/junction peptide in order for three sialyl long sugar chains on each cluster to be able to reach and simultaneously occupy three pockets of an HA homo-trimer, and all of the multiple clusters should be connected to an appropriate macromolecular scaffold to occupy multiple HA homo-trimers on the viral surface. Such a compound would efficiently bind to human influenza virus HAs, leading to inhibition of viral binding to cell surface

Hemagglutinin Inhibitor - Drug Discovery

563

receptors and blockage of influenza virus infection with effective antiviral activity against both seasonal and pandemic influenza A viruses. The combination of the HAI compound with an influenza NA inhibitor that has different but cooperative inhibitory mechanisms will enhance the antiviral efficacy synergistically. The abovedescribed properties of HA inhibitors may reduce the appearance of drug-resistant mutants and enable reduction of drug dose and increase safety for medical use. In order to exploit the clinically approved, safe, and effective anti-influenza HA inhibitor drugs that mimic receptor sialyl sugar chains, technology for low-cost sialyl sugar chain synthesis should be developed. It is expected that recent advances in sialyl-glyco-chemistry will make this possible. In addition to receptor sialyl sugar chain inhibitors, several non-sialylated compounds that inhibit HA function including a broadly neutralizing antibody [40, 41], a small-molecule HA fusion inhibitor [42], an HA peptide that binds to a sialyl sugar chain [43], multivalent peptide-nanoparticle conjugates [44], and small molecules that inhibit HA [45] have been reported. Mumefural, a small molecule isolated from Japanese apricot fruit juice [46] and the extract from guava tea [47], inhibits activities of both the HA and NA spikes of influenza A viruses. These molecules also might be further developed to become future candidate drugs for influenza virus entry inhibition. References 1. Sriwilaijaroen N, Suzuki Y (2012) Molecular basis of the structure and function of H1 hemagglutinin of influenza virus. Proc Jpn Acad Ser B Phys Biol Sci 88:226–249 2. McAuley JL, Gilbertson BP, Trifkovic S et al (2019) Influenza virus neuraminidase structure and functions. Front Microbiol 10(39):1–13 3. NIID website (2019) Antiviral resistance surveillance in Japan (as of 21 May 2019) https:// www.niid.go.jp/niid/en/influ-resist-e.html 4. Sriwilaijaroen N, Magesh S, Imamura A et al (2016) A novel potent and highly specific inhibitor against influenza viral N1-N9 neuraminidases: insight into neuraminidaseinhibitor interactions. J Med Chem 59:4563–4577 5. Takashita E, Kawakami C, Morita H et al (2018) Detection of influenza A(H3N2) viruses exhibiting reduced susceptibility to the novel cap-dependent endonuclease inhibitor baloxavir in Japan, December 2018. Euro Surveill 24(3):1–5 6. Guo H, Rabouw H, Slomp A et al (2018) Kinetic analysis of the influenza A virus HA/NA balance reveals contribution of NA to virus-receptor binding and NA-dependent

rolling on receptor-containing surfaces. PLoS Pathog 14:e1007233 7. Suzuki Y (2005) Sialobiology of influenzamolecular mechanism of host range variation of influenza viruses. Biol Pharm Bull 28:399–408 8. Stevens J, Blixt O, Paulson JC et al (2006) Glycan microarray technologies: tools to survey host specificity of influenza viruses. Nat Rev Microbiol 4:857–864 9. Sriwilaijaroen N, Kondo S, Yagi H et al (2009) Analysis of N-glycans in embryonated chicken egg chorioallantoic and amniotic cells responsible for binding and adaptation of human and avian influenza viruses. Glycoconj J 26:433–443 10. Watanabe Y, Ito T, Ibrahim MS et al (2015) A novel immunochromatographic system for easy-to-use detection of group 1 avian influenza viruses with acquired human-type receptor binding specificity. Biosens Bioelectron 65:211–219 11. Matrosovich M, Herrler G, Klenk HD (2015) Sialic acid receptors of viruses. Top Curr Chem 367:1–28

564

Nongluk Sriwilaijaroen and Yasuo Suzuki

12. Shi J, Deng G, Kong H et al (2017) H7N9 virulent mutations detected in chicken in China pose an increased threat to humans. Cell Res 12:1409–1421 13. de Vries RP, Peng W, Grant OC et al (2017) Three mutations switch H7N9 influenza to human-type receptor specificity. PLoS Pathog 13:e1006390 14. Imai M, Watanabe T, Hatta M et al (2012) Experimental adaptation of an influenza H5 haemagglutinin confers respiratory droplet transmission to a reassortant H5 HA/H1N1 virus in ferrets. Nature 486:420–428 15. Ng PS, Bo¨hm R, Hartley-Tassell LE et al (2014) Ferrets exclusively synthesize Neu5Ac and express naturally humanized influenza A virus receptors. Nat Commun 5:5750 16. Velkov T (2013) The specificity of the influenza B virus hemagglutinin receptor binding pocket: what does it bind to? J Mol Recognit 26:439–449 17. Sriwilaijaroen N, Suzuki Y (2014) Molecular basis of a pandemic of avian-type influenza virus. Methods Mol Biol 1200:447–480 18. Walther T, Karamanska R, Chan RW et al (2013) Glycomic analysis of human respiratory tract tissues and correlation with influenza virus infection. PLoS Pathog 9:e1003223 19. Sriwilaijaroen N, Nakakita S, Kondo S et al (2018) N-Glycan structures of human alveoli provide insight into influenza A virus infection and pathogenesis. FEBS J 285:1611–1634 20. Peng W, de Vries RP, Grant OC et al (2017) Recent H3N2 viruses have evolved specificity for extended, branched human-type receptors, conferring potential for increased avidity. Cell Host Microbe 21:23–34 21. Hidari KI, Murata T, Yoshida K et al (2008) Chemoenzymatic synthesis, characterization, and application of glycopolymers carrying lactosamine repeats as entry inhibitors against influenza virus infection. Glycobiology 18:779–788 22. Weis W, Brown JH, Cusack S et al (1988) Structure of the influenza virus haemagglutinin complexed with its receptor, sialic acid. Nature 333:426–431 23. Ohta T, Miura N, Fujitani N et al (2003) Glycotentacles: synthesis of cyclic glycopeptides, toward a tailored blocker of influenza virus hemagglutinin. Angew Chem Int Ed Engl 42:5186–5189 24. Yamabe M, Kaihatsu K, Ebara Y (2018) Sialyllactose-modified three-way junction DNA as binding inhibitor of influenza virus hemagglutinin. Bioconjug Chem 29:1490–1494

25. Hanson JE, Sauter NK, Skehel JJ et al (1992) Proton nuclear magnetic resonance studies of the binding of sialosides to intact influenza virus. Virology 189:525–533 26. Tsuchida A, Kobayashi K, Matsubara N et al (1998) Simple synthesis of sialyllactosecarrying polystyrene and its binding with influenza virus. Glycoconj J 15:1047–1054 27. Totani K, Kubota T, Kuroda T et al (2003) Chemoenzymatic synthesis and application of glycopolymers containing multivalent sialyloligosaccharides with a poly (L-glutamic acid) backbone for inhibition of infection by influenza viruses. Glycobiology 13:315–326 28. Ogata M, Murata T, Murakami K et al (2007) Chemoenzymatic synthesis of artificial glycopolypeptides containing multivalent sialyloligosaccharides with a gamma-polyglutamic acid backbone and their effect on inhibition of infection by influenza viruses. Bioorg Med Chem 15:1383–1393 29. Makimura Y, Watanabe S, Suzuki T et al (2006) Chemoenzymatic synthesis and application of a sialoglycopolymer with a chitosan backbone as a potent inhibitor of human influenza virus hemagglutination. Carbohydr Res 341:1803–1808 30. Guo CT, Sun XL, Kanie O et al (2002) An Oglycoside of sialic acid derivative that inhibits both hemagglutinin and sialidase activities of influenza viruses. Glycobiology 12:183–190 31. Guo CT, Wong CH, Kajimoto T et al (2002) Synthetic sialylphosphatidylethanolamine derivatives bind to human influenza A viruses and inhibit viral infection. Glycoconj J 15:1099–1108 32. Sun XL, Kanie Y, Guo CT et al (2000) Synthesis of C-3 modified sialylglycosides as selective inhibitors of influenza hemagglutinin and neuraminidase. Eur J Org Chem 2000:2643–2653 33. Oka H, Onaga T, Koyama T et al (2009) Syntheses and biological evaluations of carbosilane dendrimers uniformly functionalized with sialyl α(2-3) lactose moieties as inhibitors for human influenza viruses. Bioorg Med Chem 17:5465–5475 34. Kamitakahara H, Suzuki T, Nishigori N et al (1998) A lysoganglioside/poly-L-glutamic acid conjugate as a picomolar inhibitor of influenza hemagglutinin. Angew Chem Int Ed Engl 37:1524–1528 35. Sriwilaijaroen N, Suzuki Y (2014) A simple viral neuraminidase-based detection for highthroughput screening of viral hemagglutininhost receptor specificity. Methods Mol Biol 1200:107–120

Hemagglutinin Inhibitor - Drug Discovery 36. Yang Y, Liu HP, Yu Q et al (2016) Multivalent S-sialoside protein conjugates block influenza hemagglutinin and neuraminidase. Carbohydr Res 435:68–75 37. WHO website (2010) WHO guidelines for pharmacological management of pandemic influenza A(H1N1) 2009 and other influenza viruses. https://www.who.int/csr/resources/ publications/swineflu/h1n1_guidelines_phar maceutical_mngt.pdf 38. CDC website (2018) Influenza antiviral medications: summary for clinicians. https://www. cdc.gov/flu/professionals/antivirals/sum mary-clinicians.htm 39. Sriwilaijaroen N, Suzuki K, Takashita E et al (2015) 6SLN-lipo PGA specifically catches (coats) human influenza virus and synergizes neuraminidase-targeting drugs for human influenza therapeutic potential. J Antimicrob Chemother 70:2797–2809 40. Lin Q, Li T, Chen Y et al (2018) Structural basis for the broad, antibody-mediated neutralization of H5N1 influenza virus. J Virol 92 (17):e00547–e00518 41. Ekiert DC, Friesen RH, Bhabha G et al (2011) A highly conserved neutralizing epitope on group 2 influenza A viruses. Science 333:843–850 42. van Dongen MJP, Kadam RU, Juraszek J et al (2019) A small-molecule fusion inhibitor of influenza virus is orally active in mice. Science 363(6431):eaar6221

565

43. Chen Q, Guo Y (2016) Influenza viral hemagglutinin peptide inhibits influenza viral entry by shielding the host receptor. ACS Infect Dis 2:187–193 44. Lauster D, Glanz M, Bardua M et al (2017) A multivalent peptide-nanoparticle conjugates for influenza-virus inhibition. Angew Chem Int Ed Engl 56:5931–5936 45. Zeng LY, Yang J, Liu S (2017) Investigational hemagglutinin-targeted influenza virus inhibitors. Expert Opin Investig Drugs 26:63–73 46. Sriwilaijaroen N, Kadowaki A, Onishi Y et al (2011) Mumefural and related HMF derivatives from Japanese apricot fruit juice concentrate show multiple effects on pandemic influenza A (H1N1) virus. Food Chem 127:1–9 47. Sriwilaijaroen N, Fukumoto S, Kumagai K et al (2012) Antiviral effects of Psidium guajava Linn. (guava) tea on the growth of clinical isolated H1N1 viruses: its role in viral hemagglutination and neuraminidase inhibition. Antivir Res 94:139–146 48. Kido H, Okumura Y, Takahashi E et al (2009) Host envelope glycoprotein processing proteases are indispensable for entry into human cells by seasonal and highly pathogenic avian influenza viruses. J Mol Genet Med 3:167–175 49. Wua CY, Lina CW, Tsaia TI et al (2017) Influenza a surface glycosylation and vaccine design. Proc Natl Acad Sci U S A 114:280–285

Chapter 49 Preparation and Detection of Glycan-Binding Activity of Influenza Virus Shin-ichi Nakakita, Nongluk Sriwilaijaroen, Yasuo Suzuki, and Jun Hirabayashi Abstract We describe a method to detect influenza virus using an evanescent-field-activated fluorescence scanner type glycan array and ELISA system. Neoglycoprotein was prepared by combination of organic chemistry and biomaterial preparation. These ligands were spotted on a glass plate or plastic well to make a glycan array and ELISA plate. We detected cultured influenza virus using glycan array and ELISA. Then, we showed that the neoglycoprotein binds to Cy3-labeled hemagglutinins (H1 and H5), a NeuAcα2,6LacNAc or NeuAcα2,3LacNAc recognized protein, as detected. Key words Influenza virus, Hemagglutinin, Neoglycoprotein, Glycan array, Evanescent-field-activated fluorescence scanner

1

Introduction Influenza virus infects by recognizing and binding to host cell surface glycans. Surface antigens, hemagglutinin (HA), and neuraminidase (NA) are expressed on the cell surface of the virus; the subtype of the virus is determined by the kind of surface antigens. HA binds to sialic acids (the most common being N-acetylneuraminic acid) on the host cell surface glycans and NA binding cleaves sialic acid residues from glycans to facilitate the spread of virions [1]. Influenza viruses are usually classified into avian-type (α2,3 linked sialic acid) and human-type (α2,6 linked sialic acid) based on the binding specificity. If the virus changes the binding specificity by mutation, then a virus that recognizes the avian-type may come to recognize the human-type, and a pandemic can occur. From the above, analyzing the sugar-binding properties of the virus leads to surveillance of pandemic. In recent years, methods for measuring the glycan-binding activity of influenza virus have been developed and used by various researchers [2–5]. Among

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_49, © Springer Science+Business Media, LLC, part of Springer Nature 2020

567

568

Shin-ichi Nakakita et al. Stock influenza virus Dilute virus (1/2-1 HAU) Amplification of Influenza Virus in Embryonated Chicken Eggs (see Methods 3.1.1)

Amplification of Influenza Virus in MDCK Cells (see Methods 3.1.2)

Detection of hemagglutinin unit (see Methods 3.2.2)

Detection of binding specificity Binding Assay Using the Neoglycoprotein (see Methods 3.3)

Fig. 1 Flow diagram for virus propagation and binding assay of influenza virus

them, receptor research of influenza virus using a glycan array has achieved great results [6–10]. Consortium for Functional Glycomics (CFC) has been practical in using the glycan array, and supplied mainly to researchers in the United States. There are 600 types of immobilized glycans, and the glycans are linked to the substrate through the spacer. In addition, information on the binding specificity of influenza virus is also published on the website, as well as detailed protocols for their measurement [11]. Herein, we report (1) Propagation of influenza viruses and the determination of their hemagglutination units, (2) Preparation of glycan array for influenza virus detection, and (3) Detection of influenza virus using glycan array (see Fig. 1).

2

Materials

2.1 Propagation of Influenza Viruses

1. Class II type A/B3 biological safety cabinet (Class II Type A/B3 BSC) (see Note 1).

2.1.1 General Materials

2. 1 sterile phosphate-buffered saline, pH 7.2–7.4 (PBS). 3. Vortex mixer. 4. 15- and 50-mL conical sterile polypropylene centrifuge tubes. 5. 200- and 1000-μL micropipettes and pipette tips. 6. 5-, 10- and 25-mL sterile polystyrene disposable serological pipettes with magnifier stripe. 7. Pipetting controller. 8. Bench-top biohazard bag holder kit. 9. Electrostatic discourage (ESD)-safe powder-free nitrile exam gloves. 10. 70% ethanol to disinfect work surface areas in the biological safety cabinet and all surface equipment and containers prior to taking into the cabinet. 11. An autoclave for sterilization of experimental dry and liquid materials before use and an autoclave for biohazardous wastes before disposal. 12. 1 phosphate-buffered saline, pH 7.2–7.4 (PBS).

Preparation of Glycan Arrays

569

13. 10 PBS stock solution: dissolve 800 g NaCl, 20 g KCl, 115 g Na2HPO4 or 217 g Na2HPO4·7H2O, and 20 g KH2PO4 in 1 L distilled water. Adjust the final volume to 10 L with distilled water and mix thoroughly. Store at room temperature (RT). 14. 1 PBS working solution: Mix one part of 10 stock with nine parts of distilled water (a 1:10 (v/v) dilution). After mixing thoroughly, store it at RT. If applied for cell/virus cultivation experiments, it must be autoclaved prior to storing at RT or better at 4  C. 2.1.2 Amplification of Influenza Virus in Embryonated Chicken Eggs

1. 9–11 day old embryonated chicken eggs of specific pathogenfree chickens. 2. Egg candling light in a dark room to select viable embryos and indicate the inoculation site. 3. A pencil for making marks/labels on the egg shell. 4. Cotton wool and 70% (v/v) ethanol for disinfecting the egg shell. 5. Sterile forceps for piercing/cracking/removing the egg shell. 6. 1-mL syringe with a 25-gauge (G)  5/800 needle for influenza inoculation into the allantoic cavity and 10-mL syringe with 18-G  11/200 needle for harvesting allantoic fluid. 7. Color dot stickers for sealing the inoculation hole in the egg shell. 8. Egg incubator. 9. 2-mL sterile cryovials.

2.1.3 Amplification of Swine and Human Influenza Virus in MDCK Cells

1. Madin-Darby canine kidney (MDCK) cells (see Notes 2 and 3). 2. Minimum essential medium (MEM) with Earle’s salt, L-glutamine, and nonessential amino acids. 3. 10,000 U/mL penicillin and 10,000 μg/mL streptomycin (100 PS; Gibco-BRL, Cat. No. 15140-122). 4. Acetylated trypsin, type V-S from bovine pancreas (Sigma, Cat. No. T6763) (see Note 4). 5. 75 cm2 cell culture flasks (Corning, Cat. No. 3290). 6. Refrigerated high-speed centrifuge and centrifuge bottles. 7. 80  C freezer or liquid nitrogen freezer. 8. Humidified 37  C, 5% CO2 incubator. 9. 2-mL sterile cryovials.

2.2 Determination of Their Hemagglutination Units

1. Refrigerated low-speed centrifuge with 15- and 50-mL swinging bucket rotors. 2. Non-sterile PVC (flexible) U-bottom 96-well plates and Saran™ polyvinylidene chloride (PVDC) wrap. 3. Erythrocytes (see Note 5).

570

Shin-ichi Nakakita et al.

4. 100- and 300-μL eight-channel pipettes and pipette tips. 5. Liquid reagent reservoirs. 6. 4  C refrigerator or icebox. 2.3 Preparation of Glycan Array for Influenza Virus Detection

1. Screw-cap test tubes with a Teflon seal (10  100 mm: 7.85 mL). 2. Anthrone (FUJIFILM Wako Pure Chemical Corp., Osaka, Japan). 3. Cytidine 50 - monophospho-β-D-N-acetylneuraminic (CMP-NeuAc) (Toyobo, Osaka, Japan).

acid

4. α2,3 Sialyltransferase (Fushimi Pharmaceutical Co. Ltd., Kagawa, Japan). 5. SGP (α2,6 Sialoglycopeptide) Co. Ltd., Kagawa, Japan).

(Fushimi

Pharmaceutical

6. Neu5Acα2,3Galβ1-4GlcNAc-β-ethylamine (Tokyo Chemical Industry Co. Ltd., Tokyo, Japan). 7. Neu5Acα2,6Galβ1-4GlcNAc-β-ethylamine (Tokyo Chemical Industry Co., Ltd., Tokyo, Japan). 8. Evaporative Light Scattering Detector (ELSD) (Teledyne ISCO, Lincoln, NE). 9. Biogel P2 (Bio-Rad Laboratories, Hercules, CA). 10. Anionic exchange column Mono-Q HR 5/5 (0.5  5.0 cm, GE Healthcare, Pittsburgh, PA) Mono-Q elution system A: ammonia(aq), pH 9.0. Mono-Q elution system B: 500 mM ammonium acetate (adjusted to pH 9.0 with ammonia(aq)). 11. Bovine serum albumin (BSA) (Sigma-Aldrich Co. LLC., St. Louis, MO). 12. N-(m-Maleimidobenzoyloxy)succinimide (MBS) (Nacalai Tesque, Kyoto, Japan). 13. NaBH4 (FUJIFILM Wako Pure Chemical Corp., Osaka, Japan). 14. Silicone rubber sheet (7-chamber type) attached to an epoxycoated glass slide (Rexxam Co., Ltd., Osaka, Japan). 2.4 Detection of Influenza Virus Using Glycan Array

1. Binding buffer: 25 mM Tris–HCl, pH 7.4, 0.8% NaCl, 1% (v/v) Triton X-100, 1 mM MnCl2, 1 mM CaCl2. 2. Influenza A H1N1 (A/California/07/2009) Hemagglutinin/ H1 (Sino biological, Wayne, PA). Influenza A H5N1 (A/Vietnam/1203/2004) Hemagglutinin/H5 (Immune Technology Corp. New York, NY). 3. H1N1 influenza virus (A/Narita/2009 strain). 4. Anti-human Influenza A, B, Rabbit, Polyclonal antibody, (TAKARA BIO INC., Shiga, Japan).

Preparation of Glycan Arrays

571

5. Mouse anti-influenza-A hemagglutinin (clone IA-H5N1), Monoclonal (PromoCell GmbH Sickingenstr., Heidelberg, Germany). 6. Anti-rabbit antibody conjugated to Cy3 (Bosterbio, Pleasanton, CA). 7. Anti-mouse antibody conjugated to Cy3 (Proteintech Group, Inc. Rosemont, IL). 8. Evanescent-field-activated scanner (Bio-Rex Scan 200; Rexxam Co., Ltd. Osaka, Japan).

3

Methods

3.1 Propagation of Influenza Viruses

All procedures with influenza viruses should be conducted in a class II BSC but with highly pathogenic viruses in class III BSC under sterile conditions for prevention of cross-contamination between the samples and the environment (see Note 1). If virus populations in the original specimens are lower than 8 hemagglutination units (HAU; see Subheading 3.3), avian viruses should be amplified in embryonated chicken eggs while classical swine and human viruses should be propagated in MDCK cells (see Notes 2 and 3). Egg allantoic fluid/cell culture supernatant concentrated viruses purified virus or semi-purified virus can be used in the receptor-binding specificity assay.

3.1.1 Amplification of Avian Influenza Virus in Embryonated Chicken Eggs

1. Candle the eggs in a dark room. Use a pencil to locate the inoculation site, which is approximately 5 mm above the air sac area into the allantoic cavity. 2. Soak cotton wool in 70% (v/v) ethanol. Use it to wipe the inoculation site. 3. After ethanol air drying, pierce the egg shell at the marked inoculation site with sterile forceps. 4. Inoculate 200 μL of diluted influenza virus in PBS with 100 U/ mL penicillin and 100 μg/mL streptomycin (PS) (1/2–1 HAU) into the allantoic cavity of each egg using a 1-mL syringe with a 25-G  5/800 needle inserted entirely at a 90 angle. 5. Seal the holes with dot stickers to prevent contamination and evaporation. Place the inoculated eggs back, with the air sac pointed upward, into the egg incubator set at 34  C with 40–60% humidity for 2 days. Automatically rotate the eggs 45 from one direction to the other every hour or manually turn them at least once a day. Care should be taken not to leave the embryonated eggs outside more than 30 min. 6. Chill the eggs in a cold room or at 4  C overnight to constrict the blood vessels reducing blood contamination in the infected allantoic fluid during harvesting.

572

Shin-ichi Nakakita et al.

7. Transfer the chilled eggs to a biosafety cabinet. After removal of the stickers, crack and remove the egg shell above the air sac carefully to avoid destroying the chorioallantoic membrane with sterile forceps. 8. Use a 10-mL syringe with a 18-G  11/200 needle to pierce the egg sac and draw as much allantoic fluid as possible (usually 5–10 mL of a clear, slightly yellowish fluid/egg) into a centrifuge bottle on ice. If the fluid is red, brown or cloudy, discard it because it could be contaminated by bacteria/blood and/or excessive embryo deterioration. 9. Clarify the virus-containing allantoic fluid by centrifugation at 1750  g for 10 min at 4  C. 10. Transfer the clear virus-containing fluid into a new centrifuge bottle and pellet the virus by centrifugation at 18,270  g for 3 h. 11. Resuspend the virus pellet in PBS. Aliquot and store at 80  C until use. 3.1.2 Amplification of Classical Swine and Human Influenza Viruses in MDCK Cells

1. One day before influenza virus infection, split MDCK cells in a growth medium (MEM Nacalai supplemented with 5% heatinactivated fetal bovine serum (FBS) and antibiotics (1 PS)) into a 75 mL flask to yield approximately an 80–90% confluent monolayer ready for the virus infection on the next day. 2. On infection day, while thawing the frozen influenza virus at RT, wash the cell confluent monolayer two times with MEM-PS to remove serum, a source of HA inhibitors bearing sialyl glycans resulting in competitive blockage of virus attachment to the host cell surface. 3. After gently pipetting up and down to disrupt virus clumps in an influenza virus-containing solution, immediately dilute the virus solution in an infection medium (MEM supplemented with 1 PS and 5 μg/mL acetylated trypsin) to a final concentration of approximately 1/2 to 1 HAU (see Note 4). 4. Replace the second wash from the monolayer with 1 mL of the diluted virus solution enough to cover the monolayer cells on the 75 mL flask. 5. Place the flask in a 37  C incubator for virus adsorption onto the cell monolayer. Intermittent shaking is done every 15 min to prevent the cells from drying. 6. After virus adsorption for 1 h, add 14 mL of the infection medium to each flask. Gently rock and place the flasks in a humidified 5% CO2 incubator at 37  C until a cytopathogenic effect has appeared (approximately 3–5 days). 7. Collect the culture supernatants into centrifuge bottles and clarify it at 1750  g for 10 min at 4  C.

Preparation of Glycan Arrays

573

8. After removal of cell debris, precipitate the virus in a supernatant at 18,270  g for 3 h at 4  C. Resuspend the virus precipitate in cold sterile PBS and use a sterile disposable 1-mL syringe a 25-gauge (G)  5/800 needle to break up the virus clumps by gently drawing up and down several times. Prepare aliquots of the desired volume and keep them at 80  C until use. 3.2 Relative Quantities of Virus Particles by Hemagglutination Assay

3.2.1 Preparation of 0.5% Erythrocyte Suspension in PBS

Depending on what we want to study, one of the several techniques including the plaque forming unit (PFU), hemagglutination unit (HAU), neuraminidase enzyme unit, and quantitative reversetranscription polymerase chain reaction (qRT-PCR) may be used for the quantity (concentration) of viruses in samples. To determine the amount of the virus samples used in evaluation of the receptorbinding specificity of influenza viruses, determination of HAU that is the dilution factor of the last well forming lattice structure due to binding of the viruses (mainly by viral HAs) to receptors on erythrocytes seems better suited than others. 1. Pipette 500 μL of packed erythrocytes (see Note 5) from commercial erythrocytes in a preservative solution (Alsever’s solution) into 30 mL cold PBS in a 50-mL centrifuge tube. 2. After gently inverting the tube 4–5 times, centrifuge the erythrocyte suspension at 600  g for 5 min. 3. Replace PBS with the fresh one and gently invert the tube until the erythrocytes are resuspended. 4. Centrifuge at 600  g for 5 min. After removal of PBS, keep the packed erythrocytes in an ice box. 5. To prepare 0.5% erythrocyte suspension, pipette 1 volume of the packed erythrocytes into 199 volumes of fresh PBS in an appropriate tube (see Note 6).

3.2.2 Determination of Hemagglutination Units

1. Add 50 μL of PBS to each well of a 96-well U-bottom plate. 2. Place the plate on an ice box, and add 50 μL of each virus sample into each well of the first column (see Note 7). Mix the content in each first well of the first column and transfer each 50 μL to each well of the next column using a multichannel pipette. Continue this two-fold dilution until the second last column, and discard 50 μL of the excess mixture. Thus, each well in the last column contains only PBS (control wells). 3. Add 50 μL of 0.5% erythrocyte suspension into every well and mix by gently pipetting up and down (see Note 6). 4. If it is required to move the plate to a 4  C refrigerator, cover the plate with Saran wrap. 5. Allow the erythrocytes to settle for 2 h. 6. Determine HAU by visual reading; unbound erythrocytes in control wells sink to the bottom of the well as a halo or a circle

574

Shin-ichi Nakakita et al.

for enucleated erythrocytes (mammalian erythrocytes) and a button for nucleated erythrocytes (avian erythrocytes), whereas virus-bound erythrocytes form a lattice coating the well (called hemagglutination). Minimal concentration (the highest dilution) causing complete hemagglutination is considered 1 HAU (approximately 2  106 virus particles) in the 50 μL virus sample and thus, the number of HAU/50 μL is equal to the reciprocal of the highest dilution of the virus sample tested. 3.3 Preparation of Glycan Array for Influenza Virus Detection

1. Place a 5 mg of lyophilized SGP, at the bottom of a screw-cap test tube.

3.3.1 Conversion of SGP to α2,3SGP

4. Incubate the tube at 80  C, 1 h.

2. Add 2 mL of 50 mM HCl into the tube. 3. Close the screw-cap. 5. Remove HCl in vacuo using a rotary pump connected to a glass cold trap. 6. The sample is lyophilized several times. 7. Add 90 mg of CMP-NeuAc, 25 μL of 1 M ammonium acetate (pH 7.0), 25 μL of 5 M NaCl and 0.1 U of α2,3 sialyltransferase (total volume 0.5 mL) into the tube. 8. Incubate the tube at 40  C, 24 h. 9. Heat the tube at 100  C, 10 min. 10. Purify the α2,3SGP using gel filtration on Bio gel P2. Elute with water using gravitational flow and column temperature at 25  C. 11. Test the fractions with anthrone sulfate and pool the anthrone sulfate-positive fractions and lyophilize. 12. Dissolve the lyophilized sample in water and perform anionic exchange chromatography using a Mono-Q column. 13. Equilibrate the Mono-Q column with Solvent A at a flow rate of 1.0 mL/min, and column temperature at 25  C. Inject the PA-glycan sample and elute using 100% Solvent A (0% Solvent B). The sample detection was done using ELSD. 14. After injection of the sample for 5 min, increase the proportion of Solvent B with a linear gradient from 0% to 10% B at 10 min, to 30% B at 22 min, and then to 100% B at 27 min. 15. Collect the fraction that contains the disialoglycan. 16. Lyophilize and store the α2,3SGP (see Note 8).

3.3.2 Preparation of the Neoglycoprotein

The following procedure (see Fig. 2) is based on the findings in a previous publication [12–14]. 1. Place >1 mg of a lyophilized glycan (SGP, α2,3SGP, ethylamine conjugated-glycan) into a test tube (see Note 9).

Preparation of Glycan Arrays

575

Lys Val

Ala

Siaa2-3Galb1-4GlcNAcb1-2Mana1

6 Manb1-4GlcNAcb1-4GlcNAcb- Asn Siaa2-3Galb1-4GlcNAcb1-2Mana1 3 Lys

SGP

Thr

O

O

N-OOC

N

O

MBS SGP-OC

O

SGP-OC

N O

BSA (SH)n (Reduced BSA)

SGP-MSB

O

O N O

S BSA m

SGP-BSA (neoglycoprotein)

Fig. 2 Preparation of MSB-glycan, and its coupling to reduced BSA

2. Add 0.2 mL of 0.5 M sodium phosphate (pH 7.0) to the tube and mix well. 3. Add 50 mg of N-(m-maleimidobenzoyloxy)succinimide (MSB) which has been dissolved in 0.2 mL DMF to convert 1-amino-l-deoxy-glycan to MSB-glycan. 4. Heat at 30  C for 30 min. 5. To remove unreacted MSB, extract the mixture in a tube with 0.2 mL of CH2Cl2 three times. 6. Carefully collect the aqueous phase that contains the MSBglycan and place it into a new test tube. 7. The MSB-glycan will be coupled to BSA to produce the neoglycoprotein. BSA has been reduced to remove the disulfide linkage (see Note 10). 8. Add the MSB-glycan solution (1 μmol/0.4 mL) dropwise into the solution containing reduced BSA (100 nmol BSA/1.24 mL). 9. Mix and incubate at room temperature for 2 h. 10. Dialyze the mixture extensively against 50 mM Tris–HCl, pH 7.4, 0.15 M NaCl (see Note 11). 11. Concentrate the sample to produce the neoglycoprotein at an equivalent of 2 mg BSA/mL (see Note 12).

576

Shin-ichi Nakakita et al.

3.4 Binding Assay Using the Neoglycoprotein

The following procedure is based on a previous report [12–14]. 1. Attach the 7-chamber silicone rubber sheet on an epoxy-coated glass slide. 2. Individually, place a drop (6 nL) of the neoglycoprotein into the chambers on top of an epoxy-coated glass slide. 3. Place the slide into a humidity-controlled box at 25  C for 15 h. 4. Wash the slide with the binding buffer three times. 5. Apply 80 μL of lectin solution into each chamber. For lectin assay, use 1–5 mg/mL of hemagglutinin (see Note 13). For influenza virus assay, use 20 hemagglutinin units of H1N1 influenza virus. 6. Incubate at 25  C, 1 h. 7. Remove the lectin solution from the chamber (see Note 14). 8. For hemagglutinin, apply 1 ng/mL of anti-hemagglutinin antibody (80 μL) in the chamber on the slide, and incubate at 25  C for 1 h. 9. For H1N1 influenza virus, incubate with anti-influenza virus rabbit antibody at 25  C for 1 h, and then with anti-rabbit antibody conjugated to Cy3 at 25  C for 1 h. 10. Remove the solution from the chambers. Fluorescent images were immediately acquired using the evanescent-field-activated scanner. 11. The results of the lectin assay are shown in Fig. 3. 12. The data for the influenza virus assay is depicted in Fig. 4. The H1N1 influenza virus bound only to sialylα2,6LacNAc (see Note 15). H5N1 HA 5 mg/mL Sialyl-a2,3Sialyl-a2,6LacNAc-BSA LacNAc-BSA

H1N1 HA 5 mg/mL Sialyl-a2,3Sialyl-a2,6LacNAc-BSA LacNAc-BSA

(mg/mL)

300 100 30 10 3

Fig. 3 Assay for the binding of hemagglutinin (H1 and H5) on a glycan array immobilized with Neu5Acα2,3LacNAc BSA andNeu5Acα2,6LacNAc BSA. Varying concentrations of Neu5Acα2,3LacNAc BSA and Neu5Acα2,6LacNAc BSA (10–1000 μg/mL) were spotted on an epoxy-coated glass slide. Ligand coupling was measured using 5 mg/mL of hemagglutinin and first and second antibody. Fluorescent images were immediately acquired using the evanescent-fieldactivated scanner

Preparation of Glycan Arrays

H1N1 virus 20 HAU a2,3SGP-BSA

SGP-BSA

577

(mg/mL) 1000 300 100 30 10

Fig. 4 Binding assay of influenza virus (H1N1 Narita 2009 strain) against α2,3SGP-BSA and SGP-BSA. Various concentrations of α2,3SGP and SGP-BSA (10–1000 μg/ml) were spotted into the wells of an epoxy-coated glass slide (6 nL/1 spot). H1N1 influenza virus (Narita 2009 strain) in binding buffer (20 hemagglutinin units) was applied to each of the arrays and incubated at 25  C for 1 h. Then the arrays were incubated with anti-influenza virus rabbit antibody at 25  C for 1 h, and then with anti-rabbit antibody conjugated to Cy3 at 25  C for 1 h. Fluorescent images were immediately acquired using the evanescent-field-activated scanner. The H1N1 influenza virus bound only to SGP (Neu5Acα2,6LacNAc)

4

Notes 1. To avoid cross-contamination from samples to samples, to the user and to the environment, and vice versa, (1) all procedures with a cell culture and/or with an influenza A virus must be carried out in a Class II BSC and with a highly pathogenic virus in a Class III BSC. (2) All materials used with cell/virus cultures must be sterile and (3) proper aseptic technique must always be used. (4) Materials exposed to influenza virus or suspected of containing the virus should be properly cleaned, such as cleaning glassware with sodium hypochlorite solution, or disposed in an appropriate container, such as discard contaminated influenza virus materials including syringes and eggs in autoclavable biosafety bags for autoclaving before disposal [15]. It is confirmed that the pH is around 2–3 using pH test paper. 2. The original influenza virus isolates must be amplified for use in several studies such as antigenic variation analysis and drug susceptibility assay. However, amplification of influenza A viruses having rapid evolution caused by high mutation rate through their error-prone RNA polymerase and/or high reassortment rate through genetic reassortment of their genetic segments may generate a dominant variant in a viral population under selection pressure of the host environment [16]. To

578

Shin-ichi Nakakita et al.

prevent generation of influenza variants, (1) an appropriate cell should be selected as a host for virus amplification and (2) an original virus isolate should not be propagated more than 5 times. In our lab, the virus infection period is 48 h per time. 3. Two main factors must be considered in selection of suitable influenza virus hosts; (1) choose the host cells that can serve the virus for efficient replication and transmission among the host cells and (2) choose the host cells that do not exert selective pressures on the virus to mutate (adapt). Typically, avian isolates are amplified in the chorioallantoic cavity of embryonated chicken eggs. If not adapt, human isolates replicate less efficiently in a chorioallantoic cavity and thus they are grown in egg amniotic cavities [17]. The most well-known mutants in adaptation of the human viruses in eggs are in their HAs resulting in an increase in binding to Siaα2,3Gal receptors on the host cell surface, which is an important step to enter new host cells [18–20]. The change in the receptor-binding specificity of egg-adapted viruses can be explained by finding Siaα2,3Gal N-glycans mainly on allantoic cells driving natural selection of influenza virus evolution [21]. Classical swine and human influenza virus isolates can be cultured in mammalian cells and evidences that almost all MDCK-derived progeny virions carry HAsequences identical to that of the original isolates [22, 23]. MDCK cells have become the most generally used for propagation of swine and human influenza viruses. In fact, there are diverse MDCK strains including MDCK (NBL-2), MDCK.1, MDCK.2, MDCK.P3, [24] and α2,6sialyltransferase plasmid-transfected MDCK cells (SIAT1 [25] and AX4 [26]) that affect virus infectivity and may drive emergence of mutated variant viruses. However, (1) MDCK strains and (2) the MDCK passage number that should be concerned have rarely been informed in the publications. In our lab, mainly we use MDCK (NBL-2; (ATCC® CCL-34™)) and AX4 cells with a total passage number less than 25. 4. If wanted, 2 μg/mL TPCK-trypsin can be used instead of 5 μg/mL acetylated trypsin for proteolytic cleavage of a non-infectious viral HA0 form into HA1 and HA2 subunits still linked by a disulphide bond, a form ready to infect MDCK cells. In the case of embryonated chicken eggs that have endogenous proteases for the proteolytic cleavage of viral HA0, adding exogenous trypsin into virus inoculation is not required. 5. HAs from different influenza virus strains have different binding preferences to sialyl types and linkages, which vary on erythrocytes from different animal species. This results in variation in HAUs when different types of erythrocytes are used [23, 27]. New virus isolates may be tested with a few types of erythrocytes to find the appropriate erythrocytes giving the

Preparation of Glycan Arrays

579

highest HAU. It should be noted that the recent human influenza A (H3N2) virus isolates have obvious low capacity to agglutinate chicken erythrocytes. Much higher relative virus titers of the recent human H3N2 viruses can be observed using guinea pig erythrocytes. In general, guinea pig erythrocytes which can be agglutinated potentially by all influenza virus strains are preferred for the hemagglutination assay. 6. The most common error in the hemagglutination assay is concentration of erythrocytes. To obtain accurate and precise HAUs, the concentration of erythrocytes must be prepared correctly, and the erythrocytes must be thoroughly suspended during distribution into each well. Lower than the expected concentration of erythrocytes in the assay will give falsely high HAUs whereas a higher erythrocyte concentration will give falsely low HAUs. The use of nonhomogeneous distribution of erythrocytes into each well may lead to inability to interpret the results. 7. Whereas viral HAs reversibly bind to sialic acid receptors on erythrocytes forming a lattice (a network) called hemagglutination, viral NAs can cleave sialic acid from the erythrocyte receptors eventually causing dissociation of the lattice. To prevent the effect of cleavage activity of viral NA enzymes on the hemagglutination assay, all steps should be on ice or at 4  C. 8. α2,3SGP is sold at Fushimi Pharmaceutical Co. Ltd. (Kagawa, Japan). 9. The amount of glycan is adjusted so that the molar ratio of BSA to glycan is 1:10. 10. Before coupling with MSB-glycan, BSA is first reduced, into a screw cap tube (10  100 mm), and mix 6.9 mg of BSA 0.3 mL of 6 M urea in 0.01 M EDTA to completely dissolve the BSA. Add 12 mg of NaBH4 and incubate at room temperature for 10 min. Add 0.12 mL of 1-butanol, mix, and incubate at room temperature for 20 min. Finally, add 0.6 mL of 0.1 M NaH2PO4 and 0.24 mL of acetone. This reduced BSA sample has to be used immediately after preparation. 11. Dialysis should be performed at a low temperature (i.e., 10–20  C) to prevent precipitation of the neoglycoprotein. 12. 30 Sialyllactose-human serum albumin, 60 sialyllactose-human serum albumin, 30 sialyllactose-polyacrylamide polymer, 60 sialyllactose-polyacrylamide polymer, Neu5Acα2,3Galβ14GlcNAcβ-polyacrylamide polymer, and sialoglycopolyacrylamide polymer are sold by GlycoTech Corp. (Gaithersburg, MD). These polymers can be used to substitute NGP. 13. Binding activity of Hemagglutinin is very weak. Therefore a protein is added with high concentration. Commercially available HA can be used for quality control of glycan arrays. Commercially available HA is summarized in Table 1.

11053-V08H 11085-V08H 11713-V08H 11699-V08H 11717-V08H 11723-V08H

Influenza B virus (B/Florida/4/2006) Hemagglutinin/HA

Influenza A H1N1 (A/California/07/2009) Hemagglutinin/HA Protein (His Tag)

Influenza A H5N1 (A/Hong kong/213/2003) Hemagglutinin/HA Protein (28 Ser/Trp, His Tag)

Influenza A H5N2 (A/American green-winged teal/California/ HKWF609/2007) Hemagglutinin/HA Protein (His Tag)

Influenza A H5N8 (A/duck/NY/191255-59/2002) Hemagglutinin/HAProtein (His Tag)

Influenza A H6N1 (A/northern shoveler/California/HKWF115/ 2007) Hemagglutinin/HA Protein (His Tag)

11082-V08H1

11068-V08H

IT-003-B11p

HA1(B/Phuket/3073/2013)

Influenza A H1N1 (A/Brevig Mission/1/1918) Hemagglutinin/ HA, Influenza A virus

IT-003-B11ΔTMp

HA(B/Phuket/3073/2013)

11068-V08H1

IT-003-SW12p

HA1(A/California/07/2009) (H1N1)

Influenza A H1N1 (A/Brevig Mission/1/1918) Hemagglutinin Pro (HA1 Subunit), Influenza A virus

IT-003-00430ΔTMp

HA(H3N2) (A/Hong Kong/4801/2014)

Sino Biological (CHN)

IT-003-SW12ΔTMp

HA(ΔTM)(A/California/07/2009(H1N1))

Immune Technologya (USA)

a

ab190125

Recombinant influenza A virus hemagglutinin H5 protein

Abcam (UK)

Cat. No

Name

Company

Table 1 Commercially available influenza HA

http://www.sinobiological.com/

http://immune-tech.com/index.php

https://www.abcam.com

URL

580 Shin-ichi Nakakita et al.

HA1-V82E2 HA9-V5253 HA9-V5228

Biotinylated Influenza A [A/Guinea fowl/Hong Kong/WF10/99 (H9N2)] Hemagglutinin 1 (HA1)

Recombinant Influenza A [A/Shanghai/2/2013(H7N9)] Hemagglutinin (HA), Fc Tag

Recombinant Influenza A [A/Shanghai/2/2013(H7N9)] Hemagglutinin 1 (HA1)

COSMO BIO is a supplier in Japan

a

HA1-V82E1

Biotinylated Influenza A [A/Guinea fowl/Hong Kong/WF10/99 (H9N2)] Hemagglutinin (HA)

Acro Biosystems (CHN)

Z03181-100

11693-V08H

Influenza A H10N3 (A/duck/Hong Kong/786/1979) Hemagglutinin/HA Protein (His Tag)

H1N1 (A/California/04/2009), Hemagglutinin

11719-V08H

Influenza A H9N2 (A/Guinea fowl/Hong Kong/WF10/99) Hemagglutinin/HA Protein (His Tag)

GenScript (USA)

11722-V08H

Influenza A H8N4 (A/pintail duck/Alberta/114/1979) Hemagglutinin/HA Protein (His Tag)

Influenza A H7N7 (A/Netherlands/219/03) Hemagglutinin Protein (HA1 Subunit) (His Tag)

https://www.acrobiosystems.com/

https://www.genscript.com/

Preparation of Glycan Arrays 581

582

Shin-ichi Nakakita et al.

14. For the glycoprotein-lectin interaction assay, we usually use an evanescent-field-activated fluorescence scanner [12], which eliminates the washes outlined in step 9 and allows for the highly sensitive and reproducible detection of fluorescence under an equilibrium condition. 15. Detection of influenza virus using the neoglycoprotein array is 20 hemagglutinin units, whereas that for a conventional assay is 28–29 (~10 μg/mL) [27, 28]. References 1. Stevens J, Blixt O, Paulson JC, Wilson IA (2006) Glycan microarray technologies: tools to survey host specificity of influenza viruses. Nat Rev Microbiol 4:857–864 2. Carvalho SB, Moleirinho MG, Wheatley D et al (2017) Universal label-free in-process quantification of influenza virus-like particles. Biotechnol J 12:1700031 3. Guo H, Rabouw H, Slomp A et al (2018) Kinetic analysis of the influenza A virus HA/NA balance reveals contribution of NA to virus-receptor binding and NA-dependent rolling on receptor-containing surfaces. PLoS Pathog 14:e1007233 4. Benton DJ, Martin SR, Wharton SA et al (2015) Biophysical measurement of the balance of influenza a hemagglutinin and neuraminidase activities. J Biol Chem 290:6516–6521 5. Wright ZVF, Wu NC, Kadam RU et al (2017) Structure-based optimization and synthesis of antiviral drug Arbidol analogues with significantly improved affinity to influenza hemagglutinin. Bioorg Med Chem Lett 27:3744–3748 6. Rillahan CD, Paulson JC (2011) Glycan microarrays for decoding the glycome. Annu Rev Biochem 80:797–823 7. Oyelaran O, Gildersleeve JC (2009) Glycan array. Recent advances and future challenges. Curr Opin Chem Biol 13:406–413 8. Fukui S, Feizi T, Galustian C et al (2002) Oligoglycan microarrays for high-throughput detection and specificity assignments of carbohydrate-protein interaction. Nat Biotechnol 20:1011–1017 9. Wang D, Liu S, Trummer BJ et al (2002) Carbohydrate microarrays for the recognition of cross-reactive molecular markers of microtubes and host cells. Nat Biotechnol 20:275–281 10. Nemanichvili N, Tomris I, Turner HL et al (2019) Fluorescent trimeric hemagglutinins reveal multivalent receptor binding properties. J Mol Biol 431:842–856

11. Takahashi T, Kawakami T, Mizuno T et al (2013) Sensitive and direct detection of receptor binding specificity of highly pathogenic avian influenza A virus in clinical samples. PLoS One 8:e78125 12. Tateno H, Mori A, Uchiyama N et al (2008) Glycoconjugate microarray based on an evanescent-field fluorescence-assisted detection principle for investigation of glycanbinding proteins. Glycobiology 18:789–798 13. Nakakita S, Hirabayashi J (2016) Preparation of glycan arrays using pyridylaminated glycans. Methods Mol Biol 1368:225–235 14. Kuno A, Uchiyama N, Koseki-Kuno S et al (2005) Evanescent-field fluorescence-assisted lectin microarray: a new strategy for glycan profiling. Nat Methods 2:851–856 15. http://www.who.int/csr/resources/ publications/biosafety/WHO_CDS_CSR_ LYO_2004_11/en/ 16. Sriwilaijaroen N, Suzuki Y (2012) Molecular basis of the structure and function of H1 hemagglutinin of influenza virus. Proc Jpn Acad Ser B Phys Biol Sci 88:226–249 17. Burnet F, Bull DH (1943) Changes in influenza virus associated with adaptation to passage in chick embryos. Aust J Exp Biol Med Sci 21:55–69 18. Ito T, Suzuki Y, Takada A et al (1997) Differences in sialic acid-galactose linkages in the chicken egg amnion and allantois influence human influenza virus receptor specificity and variant selection. J Virol 71:3357–3362 19. Gambaryan AS, Robertson JS, Matrosovich MN (1999) Effects of egg-adaptation on the receptor-binding properties of human influenza A and B viruses. Virology 258:232–239 20. Widjaja L, Ilyushina N, Webster RG et al (2006) Molecular changes associated with adaptation of human influenza A virus in embryonated chicken eggs. Virology 350:137–145

Preparation of Glycan Arrays 21. Sriwilaijaroen N, Kondo S, Yagi H et al (2009) Analysis of N-glycans in embryonated chicken egg chorioallantoic and amniotic cells responsible for binding and adaptation of human and avian influenza viruses. Glycoconj J 26:433–443 22. Rocha EP, Xu X, Hall HE et al (1993) Comparison of 10 influenza A (H1N1 and H3N2) haemagglutinin sequences obtained directly from clinical specimens to those of MDCK cell- and egg-grown viruses. J Gen Virol 74:2513–2518 23. Takemae N, Ruttanapumma R, Parchariyanon S et al (2010) Alterations in receptor-binding properties of swine influenza viruses of the H1 subtype after isolation in embryonated chicken eggs. J Gen Virol 91:938–948 24. Dukes JD, Whitley P, Chalmers AD (2011) The MDCK variety pack: choosing the right strain. BMC Cell Biol 12:43

583

25. Matrosovich M, Matrosovich T, Carr J et al (2003) Overexpression of the alpha-2,6-sialyltransferase in MDCK cells increases influenza virus sensitivity to neuraminidase inhibitors. J Virol 77:8418–8425 26. Hatakeyama S, Sakai-Tagawa Y, Kiso M et al (2005) Enhanced expression of an alpha2,6linked sialic acid on MDCK cells improves isolation of human influenza viruses and evaluation of their sensitivity to a neuraminidase inhibitor. J Clin Microbiol 43:4139–4146 27. Ito T, Suzuki Y, Mitnaul L et al (1997) Receptor specificity of influenza A viruses correlates with the agglutination of erythrocytes from different animal species. Virology 227:493–499 28. Xu R, de Vries RP, Zhu X et al (2013) Preferential recognition of avian-like receptors in human influenza A H7N9 viruses. Science 342:1230–1235

Chapter 50 Screening for Components/Compounds with Anti-Rotavirus Activity: Detection of Interaction Between Viral Spike Proteins and Glycans Keita Yamada, Junko Nio-Kobayashi, and Mizuho Inagaki Abstract Rotaviruses are the major etiologic agents of acute gastroenteritis. Viral attachment to the cell surface is crucial to initiate infection. The VP8∗ domain, the trypsinized cleavage fragment of the outermost spike protein VP4 of rotavirus, has a galectin-like structure required for binding to the cell surface. We used the evanescent-field fluorescence-assisted assay to understand the complex mechanism underlying the virusglycan/glycoprotein interaction. Besides, we have described virus infection assays, neutralization assay, and pretreatment assay, using cell culture. These approaches using rotavirus particles will provide novel information that has been difficult to obtain from glycan microarray using recombinant VP8∗. Key words Rotavirus, Galectin-like structure, Virus-glycan/glycoprotein interaction, Evanescentfield fluorescence-assisted assay, Virus purification, Virus infection, Neutralization assay, Pretreatment assay

1

Introduction Rotaviruses, in the familyReoviridae, are the major etiologic agents of acute gastroenteritis in young mammals including humans [1, 2]. Initiation of infection of various viruses including rotavirus is mediated by the recognition of glycans on the cell surface of host cells. Rotavirus is a three-layered particle (TLP), and its outermost capsid comprises a coat glycoprotein, VP7, and a spike protein, VP4 [1, 2]. Gene sequences encoding VP7 and VP4 are used to classify rotavirus into G and P genotypes according to glycoprotein and protease-sensitive protein, respectively [2]. It is essential that the VP4 is cleaved by trypsin for rotavirus infection, which generates derived cleavage fragments VP5∗ and VP8∗. VP8∗ contains a galectin-like structure [3, 4], and it interacts with the cell surface glycans on the host cell, such as sialoglycans and histo-blood group antigens (HBGAs) [5], while VP5∗ and VP7 interact with viral receptor(s), such as integrins and heat shock cognate protein

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_50, © Springer Science+Business Media, LLC, part of Springer Nature 2020

585

586

Keita Yamada et al.

hsc70 for post-attachment [3]. Several studies have used glycan microarrays in order to elucidate the binding between VP8∗ and glycans through the expression analysis of VP8∗ protein in Escherichia coli [6, 7] and baculovirus [8]. However, group A rotavirus has been identified to have at least 37 P genotypes [9]. Furthermore, the glycan-binding site of VP8∗ varies depending on P genotypes [5]. Hence, we consider that the trypsin cleavage of the virus particle is unique and critical to unravel the nature of rotavirus-cell interaction. Actually, we have reported that O-glycans of bovine kappa-casein is involved in direct binding to HRV, using the following methods [10]. In this chapter, we first describe the protocol for purification of rotavirus, TLP [11], followed by the protocol for evanescent fieldbased fluorescence-assisted detection [10, 12]. This method allows direct visualization of the binding between TLP and glycans/glycoproteins, and is useful for screening anti-rotaviral glycoproteins that competitively inhibit the attachment of the virus to host cells (Fig. 1a). Furthermore, we also propose the infection assays, neutralization assay and pretreatment assay [10], using a rhesus monkey kidney MA104 cell line, a well-established model for rotavirus studies. Neutralization assay is a general method for evaluation of the inhibitory activity against virus replication, while pretreatment assay is the evaluation of blocking the virus receptor on the cell surface by glycans/glycoproteins, which focused on the virus attachment step (Fig. 1). Combination of these three assays will permit us to understand the inhibition mechanism of glycans/ glycoproteins on rotavirus infection.

2

Materials

2.1 MA104 Cell Culture

1. Cell-culture flasks (25 cm2/75 cm2/150 cm2). 2. Sterile pipettes (5 mL/10 mL). 3. Sterile Pasteur pipettes. 4. Safety cabinet and aspirator. 5. Eagle’s Minimal Essential Medium (Eagle’s MEM): Store at 4  C. 6. Amphotericin B (250 μg/mL): Store at

30  C.

7. Penicillin (10,000 U/mL) / Streptomycin (10 mg/mL) solution: Store at 30  C. 8. Fetal bovine serum (FBS): Place in a water bath at 56  C for 30 min, Store at 30  C. 9. Tryptose phosphate broth (TPB): Dissolve 2.95 g TPB in 100 mL H2O and sterilize by autoclaving. Store at 4  C.

Screening for Anti-Rotavirus Activity Components A

587

sample (e.g., glycans/glycoproteins)

TLP

B

rotavirus receptors nucleus DLP

ER viroplasm

Fig. 1 The rotavirus replication cycle and evaluation of the inhibitory activity. Rotavirus replication cycle is reviewed in Desselberger [1]. (a) Screening for anti-rotavirus components using evanescent field-based fluorescence-assisted detection. It can visualize the binding between the virus particle (TLP) and glycans/glycoproteins (see Fig. 2). (b) Pretreatment assay is the evaluation of the interaction between virus receptors and glycans/glycoproteins. Before the virus infection, glycans/glycoproteins were added to cells and incubated at 37  C, followed by washing the inoculum before inoculation of the rotavirus (see Fig. 4). If the glycans/glycoproteins interact with virus receptors on the cell surface, the counts of the infected cell are decreased compared with the control. Neutralization is the evaluation of the inhibitory activity against virus attachment to virus replication (see Fig. 3)

10. 7.5% NaHCO3: Dissolve 7.5 g NaHCO3 in 100 mL H2O and sterilize by autoclaving. Store at 4  C. 11. Sterile phosphate-buffered saline (PBS) solution: Sterilize by autoclaving and store at 4  C. 12. 0.4% trypsin: Dissolve 1 g trypsin (1:250) in 250 mL sterile PBS followed by overnight stirring at 4  C. Collect the supernatant after centrifugation through filtration using a 0.22 μm filter membrane. Store at 30  C. 13. EDTA-trypsin: Mix 6.3 mL of 0.4% trypsin and 2.5 mL of 1% sterile EDTA. Adjust the volume to 100 mL with sterile PBS. 14. Growth medium: Mix Eagle’s MEM (78%), FBS (10%), TPB (10%), amphotericin B (0.5%), penicillin/ streptomycin

588

Keita Yamada et al.

solution (1%), and 7.5% NaHCO3 (0.5%) in a safety cabinet. Store at 4  C. 15. Maintenance medium: Mix Eagle’s MEM (86%), FBS (2%), TPB (10%), amphotericin B (0.5%), penicillin/ streptomycin solution (1%), and 7.5% NaHCO3 (0.5%) in a safety cabinet. Store at 4  C. 2.2

Virus Purification

1. Ultracentrifuge with a fixed angle and swing rotors. 2. Ultracentrifuge tubes (clear type). 3. Sucrose solutions (70% and 30%, w/v): 70% and 30% sucrose in H2O, receptivity. Sterilize by autoclaving and store at room temperature. 4. Cesium chloride solutions (55% and 40%, w/v): 55% and 40% cesium chloride in H2O, receptivity. Sterilize by autoclaving and store at room temperature.

2.3 Screening of Rotavirus-Binding Glycoproteins

1. Evanescent-field activated fluorescence scanner (Rexxam, Osaka, Japan). 2. Microarray-grade epoxy-coated glass slide. 3. Incubation chamber. 4. Silicon rubber sheet. 5. Trypsin factor IX (1000 μg/mL): Store at

30  C.

6. Hank’s balanced salt solution (Hank’s solution): Store at 4  C. 7. Spotting solution for DNA microarray. 8. Washing solution: 25 mM Tris–HCl (pH 7.4) containing 0.8% NaCl, and 1% Triton-X 100. Store at 4  C. 9. Blocking solution: 25 mM Tris–HCl (pH 7.4) containing 0.8% NaCl, 1% Triton-X 100, and 4% BSA. Store at 4  C. 10. Binding solution: 25 mM Tris–HCl (pH 7.4) containing 0.8% NaCl, 1% Triton-X 100, 1 mM MnCl2, 1 mM CaCl2, and 1% BSA. Store at 4  C. 11. Anti-rotavirus outermost capsid antibody (e.g., PA1-85845, Thermo Fisher Scientific, Rockford): Store at 30  C. 12. Cyanine dye 3 (Cy3) conjugated secondary antibody: Store at 30  C. 2.4 Infection Assay and Immunofluorescence Detection

1. 24-Well teflon-coated slide (well diameter: 4 mm). 2. Cover glass. 3. Petri dish: Place the wet paper in a Petri dish and sterilize by autoclaving. 4. Trypsin factor IX (1000 μg/mL): Store at

30  C.

5. Hank’s solution: Sterilize by autoclaving and store at 4  C.

Screening for Anti-Rotavirus Activity Components

589

6. 4% paraformaldehyde (4% PFA): Store at 4  C. 7. Cold Methanol: Store at

30  C.

8. Anti-rotavirus double-layered VP6 antibody (e.g., ab181695, Abcam, Cambridge): Store at 20  C. 9. FITC-conjugated secondary antibody: Store at

30  C.

10. Fluorescence microscope.

3

Methods

3.1 Rotavirus Purification

1. Inoculate rotavirus to confluent monolayers of MA104 cells (e.g., four 150 cm2 flasks), and propagate in cells (see Note 1). Allow the virus to grow at 37  C until infected cells show cytopathic effects. This requires approximately 4–5 days. 2. Harvest infected culture (cells plus medium), followed by freezing and thawing it three times. 3. Spin the harvest (e.g., 200 mL) at 15,000  g for 15 min in a fixed angle rotor to recover the clarified supernatant. 4. Spin the clarified supernatant (e.g., 200 mL) at 100,000  g for 2 h in a fixed angle rotor to recover the pellets that are to be suspended in 15 mL of PBS. 5. Take 70% sucrose solution in two ultracentrifuge tubes and add 30% of sucrose solution to it forming two distinct layers. 6. Divide the virus suspension collected in step 4 to an equal volume and add it to the tubes. 7. Ultracentrifuge at 220,000  g for 2 h at 4  C in a swing rotor. 8. Extract the crude virus fraction that appears as a white band between 30% and 70% sucrose layer. 9. Collect the crude virus fraction by directly inserting a syringe into the white band. 10. Place 55% cesium chloride solution in two ultracentrifuge tubes. 11. Overlay a 40% cesium chloride solution above it. 12. Carefully overlay the crude virus solution on top of the cesium chloride layers. 13. Ultracentrifuge at 8,000,000  g for 16–20 h, at 4  C in a swing rotor. 14. Extract and retain both bluish bands that are formed at about one third from the bottom of the tube; the upper and lower bands are TLP and double-layered particle (DLP), respectively. 15. Collect each fraction using a syringe.

590

Keita Yamada et al.

3.2 Screening of Rotavirus-Binding Glycoproteins by Using Evanescent FieldBased FluorescenceAssisted Detection

1. Dilute trypsin factor IX with Hank’s solution at a concentration of 20 μg/mL. 2. Treat virus solution with equal volumes of diluted trypsin, and incubate at 37  C for 30 min.

3.2.1 Trypsin Activation 3.2.2 RotavirusGlycoprotein-Binding Assay

Please read Subheading 3.2.2 with Fig. 2. 1. Dissolve the glycoprotein samples in the 50% aqueous spotting solution to prepare a glycoprotein solution of 0.5 mg/mL (see Note 2). 2. Spot 1 μL of glycoprotein solution on an epoxy-coated glass slide attached with a silicon rubber sheet to print the glycoprotein spots with a diameter of approximately 1 mm. 3. Incubate the glass slide in the incubation chamber containing wet paper at 25–30  C for 3 h to complete immobilization. 4. Rinse the immobilized glycoprotein containing glass slide three times with washing solution to remove excess non-immobilized materials. 5. Add the blocking solution to the glass slide and incubate for 1.5 h with shaking to block the unreacted epoxy groups. 6. Remove the blocking solution, and rinse the glass slide three times with washing solution.

Fig. 2 Flow chart of rotavirus-glycoprotein-binding assay (Subheading 3.2.2)

Screening for Anti-Rotavirus Activity Components

591

7. Dilute trypsinized virus with the binding solution at a titer of 102–10 fluorescent cell focus forming unit (ffu) /μL, and add 80 μL of them to each well on the glass slide (see Note 3). 8. Incubate them at 25–30  C for 2 h with shaking. 9. Remove the virus solution, and rinse the glass slide once with washing solution. 10. Dilute the anti-rotavirus outermost capsid antibody with the binding solution to 10 μg/mL, and add 80 μL of them to each well on the glass slide (see Note 4). 11. Incubate them at 25–30  C for 2 h with shaking. 12. Remove the antibody solution, and rinse the glass slide once with washing solution. 13. Dilute the Cy3 conjugated secondary antibody with binding solution to 10 μg/mL, add 80 μL of them to each well on the glass slide (see Note 4). 14. Incubate them at 25–30  C for 2 h with shaking. 15. Remove the secondary antibody solution, and rinse the glass slide three times with washing solution. 16. Acquire the fluorescent image of the glass slide using evanescent-field activated fluorescence scanner. 3.3 Rotavirus Infection Assay Using MA104 Cells

1. Remove the medium from the confluent monolayer of MA104 cells.

3.3.1 MA104 Cell Maintenance

3. Add EDTA-trypsin (1:10 dilution) and incubate the cells at 37  C for 5 min, until the cells begin to detach from the bottle.

2. Rinse with sterile PBS and remove it.

4. Add the growth medium, and suspend the cells with a sterile pipette. 5. Transfer the cells into a fresh bottle and add the growth medium. 6. Maintain the culture at 37  C in a humidified atmosphere containing 5% CO2. 3.3.2 Cell Preparation for Infection Assay

1. After trypsin treatment, suspend MA104 cells and dilute them to a concentration of 1  105 cells/mL in the growth medium. 2. Set a 24-well Teflon-coated slide in a Petri dish. 3. Add 20 μL of the cell suspension to each well, and culture at 37  C in 5% CO2 until the cells were grown to confluence. 4. Remove the growth medium from the cell and replace with the maintenance medium at 24 h before virus inoculation.

592

Keita Yamada et al.

Fig. 3 Flow chart of neutralization assay (Subheading 3.3.3) 3.3.3 Neutralization Assay

Please read Subheading 3.3.3 with Fig. 3. 1. Activate rotavirus by trypsin (see Subheading 3.2.1). 2. Dilute trypsinized virus with the maintenance medium at a titer of 103 fluorescent cell ffu/100 μL (see Note 5). 3. Mix equal volumes of the sample and diluted virus, and incubate at 37  C for 1 h (see Note 6). 4. Remove the medium from cells, and add 20 μL of the virussample mixture (see Note 7). 5. Incubate it for 1 h at 37  C in 5% CO2. 6. Remove the inoculum from cells, and add 20 μL of maintenance medium. 7. Culture for 16 h at 37  C in 5% CO2 (see Note 8). 8. After incubation, rinse the infected cells with sterile PBS. 9. Fix them with 4% PFA for 20 min, and then wash with sterile PBS. 10. Add 20 μL of cold methanol to each well for 2 min, and then wash with sterile PBS. 11. Air dry the slide, and store it at indirect immunofluorescence.

3.3.4 Pretreatment Assay

30  C until detection of

Please read Subheading 3.3.4 with Fig. 4. 1. Activate rotavirus by trypsin (see Subheading 3.2.1). 2. Dilute trypsinized virus with the maintenance medium at a titer of 103 ffu/100 mL (see Note 5). 3. Remove the medium, and add 20 μL of the sample (see Note 7). Maintain the culture for 1 h at 37  C in 5% CO2.

Screening for Anti-Rotavirus Activity Components

593

Fig. 4 Flow chart of pretreatment assay (Subheading 3.3.4)

4. Remove the sample from cells, and wash with sterile PBS. 5. Add 20 μL of the diluted virus to the cell, and culture for 1 h at 37  C in 5% CO2. 6. After 1 h, remove the inoculum and replace 20 μL of maintenance medium. 7. Proceed to steps 7–11 of Subheading 3.3.3. 3.3.5 Indirect Immunofluorescence Detection

1. Cover the permeated cells with the anti-VP6 monoclonal antibody; incubate for 30 min at 37  C. 2. Wash the slide with PBS three times with gentle swing. 3. Cover the permeated cells with FITC-conjugated secondary antibody; incubate for 30 min at 37  C. 4. Wash the slide with PBS three times with gentle swing. 5. Air dry the slide and mount it onto a clean cover glass. 6. Count the foci (infected cell) using fluorescence microscopy (see Note 9).

4

Notes 1. It is preferable to use 1 L of infectious culture medium since the yield of purified virus particles is roughly proportional to the starting volume.

594

Keita Yamada et al.

2. Place the spotted slides immediately in the incubation chamber as the immobilization efficiency of glycoproteins is reduced by drying. 3. Prepare multiple dilutions for the assay because the binding ability to glycoproteins varies greatly depending on the type of rotavirus. 4. Add reagents quickly to avoid drying. 5. Optimal dilutions should be predetermined by preliminary infection test. 6. Prepare the control, solvent without the test sample, and proceed with the experiment. 7. Gently aspirate the supernatant and inoculum, not to peel the monolayer cells. 8. Virus replication is promoted during incubation at 37  C, and the replication time differs depending on the strain. Perform the preliminary test repeatedly, and standardize the optimal incubation time at which the single-infected foci can be visualized (see Fig. 3, insert photo). 9. The inhibitory activity is evaluated as the percentage reduction in the foci numbers of infected cells, as compared with infected cells without the sample (control condition).

Acknowledgments We appreciate Dr. Osamu Nakagomi, Dr. Toyoko Nakagomi, Dr. Makoto Sugiyama, and Dr. Naoto Ito for their contribution in describing the virus purification. This work was supported by the Japan Society for the Promotion of Science [JSPS, Tokyo, Japan; Grant-in-Aid for Scientific Research (C) no. 17K08398 (to K.Y.) and 18K05504 (to M.I.)]. References 1. Desselberger U (2014) Rotaviruses. Virus Res 190:75–96 2. Greenberg HB, Estes MK (2009) Rotaviruses: from pathogenesis to vaccination. Gastroenterology 136:1939–1951 3. Baker M, Prasad BVV (2010) Rotavirus cell entry. In: Johnson J (ed) Cell entry by non-enveloped viruses, Current topics in microbiology and immunology, vol 343. Springer, Heidelberg, pp 121–148 4. Dormitzer PR, Sun ZY, Wagner G et al (2002) The rhesus rotavirus VP4 sialic acid binding domain has a galectin fold with a novel carbohydrate binding site. EMBO J 21:885–897

5. Hu L, Crawford SE, Czako R et al (2012) Cell attachment protein VP8∗ of a human rotavirus specially interacts with A-type histo-blood group antigen. Nature 485:256–259 6. Pang LL, Wang MX, Sun XM et al (2018) Glycan binding patterns of human rotavirus P [10] VP8∗ protein. Virol J 15:161. https:// doi.org/10.1186/s12985-018-1065-9 7. Kim HS, Lee B, Han SY et al (2017) Expression of bovine rotavirus VP8 and preparation of IgY antibodies against recombinant VP8. Acta Virol 61:143–149 8. Dunn SJ, Fiore L, Werner RL et al (1995) Immunogenicity, antigenicity, and protection

Screening for Anti-Rotavirus Activity Components efficacy of baculovirus expressed VP4 trypsin cleavage products, VP5∗ and VP8∗ from rhesus rotavirus. Arch Virol 140:1969–1978 9. Matthijnssens J, Ciarlet M, McDonald SM et al (2011) Uniformity of rotavirus strain nomenclature proposed by the Rotavirus Classification Working Group (RCWG). Arch Virol 156:1397–1413 10. Inagaki M, Muranishi H, Yamada K et al (2014) Bovine k-casein inhibits human

595

rotavirus (HRV) infection via direct binding of glycans to HRV. J Dairy Sci 97:2653–2661 11. Nakagomi O (2016) Title of Fundamentals of ultracentrifugal virus purification. https://ls. beckmancoulter.co.jp/files/cases/Fun damentals_of_Ultracentrifugal_Virus_Purifica tion.pdf. Accessed 28 May 2019 12. Kuno A, Uchiyama N, Koseki-Kuno S et al (2005) Evanescent-field fluorescence-assisted lectin microarray: a new strategy for glycan profiling. Nat Methods 2:851–856

Chapter 51 ELISA-Based Methods to Detect and Quantify Norovirus Virus-Like Particle Attachment to Histo-Blood Group Antigens Haruko Shirato Abstract Histo-blood group antigen (HBGA) recognition by norovirus (NoV) has been studied using various techniques. Enzyme-linked immunosorbent assays (ELISAs) using virus-like particles (VLPs) have enabled us to visualize the last step of HBGAs-NoV binding with a total reaction time of approximately 8 h. Herein, we describe two ELISA-based methods to detect and quantify NoV VLP attachment to HBGAs: saliva-VLP binding assay and carbohydrate-VLP binding assay. Key words Norovirus, Histo-blood group antigen, Virus-like particle, Enzyme-linked immunosorbent assay

1

Introduction Human norovirus (NoVs) infection is the most common cause of viral gastroenteritis worldwide. They are divided into seven genogroups (GI–GVII), with each genogroup being further divided into several genotypes. Each genotype is antigenically different. NoV is a small, round, non-enveloped virus, with a diameter of 38 nm and contains a single-stranded positive-sense 7.6 kb RNA genome encoding three open reading frames (ORFs). ORF1 encodes a nonstructural polyprotein, and ORF2 and ORF3 encode the major capsid protein VP1 and the minor capsid protein VP2, respectively (Fig. 1). A single virus particle is composed of one copy of the genomic RNA, 180 copies of VP1, and a few copies of VP2 [1, 2]. VP1 has two major domains: a shell (S) domain, which forms the core of the icosahedral virus shell, and a protruding (P) domain, which forms arches extending from the shell and is responsible for virus–host interactions.

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_51, © Springer Science+Business Media, LLC, part of Springer Nature 2020

597

598

Haruko Shirato

Virus extraction from the feces of the patient Norovirus A member of the family Caliciviridae Sphere 38nm in diameter Many morphologically similar but antigenically distinct viruses

Capsid protein expression in baculovirus Helicase

VPg Protease

ORF3

RdRp ORF2

ORF1

VP1 7.5kb

5' 1

2

3

3' 4

5

6

7

Recombinant baculovirus

180 copies (90 dimers) Norovirus Virus-Like Particle Diagnosis Candidate vaccine Virus-host interaction Fig. 1 Expression of capsid protein. A representation of the preparation of recombinant NoV VLPs

When genes encoding VP1 or both VP1 and VP2 were expressed in insect or mammalian cells, capsid proteins of approximately 58 kDa were synthesized and self-assembled into virus-like particles (VLPs) (Fig. 1) [3, 4]. These VLPs, though artificial, are morphologically and antigenically similar to native virions [3, 5–7] (Fig. 1). VLP expression has allowed the identification of cellular molecules that are binding targets for NoV.

Histo-Blood Group Antigen Recognition by Norovirus

599

Recognition site 1 GI strains prefer A antigens. GII strains prefer B antigens. Recognition site 2 Both GI and GII strains bind tightly to type 1 carbohydrates.

Gal NAc

3

Gal

3/4

Glc NAc

2

4/3

Fuc

Fuc

Fig. 2 Recognition sites. We discovered the recognition sites 1 and 2. For recognition site 1, GI strains prefer A antigens. GII strains prefer B antigens. For recognition site 2, both GI and GII strains bind tightly to type 1 carbohydrates

The interaction of NoVs with HBGAs is thought to play a critical role in the infection process. Different NoV genotypes exhibit different patterns of HBGA binding [8–11]. Epidemiological studies showed that individuals with different ABH phenotypes are infected with NoV strains in a genotype-specific manner. In vitro, it was shown that almost all NoV VLPs, except for a few genotypes, bind to histo-blood group antigens (HBGAs), which are structurally related oligosaccharides and include ABH and Lewis antigens (Fig. 2). For example, the GII.4 genotype includes global epidemic strains and the VLPs of GII.4 strains bind to more HBGAs than the strains of the other genotypes in vitro (Table 1). This characteristic may be linked to the worldwide transmission of GII.4 strains, which explains why globally epidemic strains belong to the GII.4 genotype. HBGA recognition by NoV has been studied using various techniques. Enzyme-linked immunosorbent assays (ELISAs) using VLPs have enabled us to visualize the last step of HBGANoV VLP binding with a total reaction time of approximately 8 h. In contrast, surface plasmon resonance (SPR) analysis using NoV VLPs has enabled us to visualize each binding step within a reaction time frame of 120–1200 s [10]. Crystallographic studies using the P domain have revealed that P domain proteins are sufficient for HBGA-NoV binding [12, 13]. These in vitro studies showed that the binding properties of human NoV to HBGAs were variable and that both the terminal residues and the internal structures of HBGAs were important for NoV–HBGA interaction.

r754 r7k r445

r1025 r76 r47

5 6 6

7

12 14

r485 rHV r18-3 r336 r104

No binding No binding H type 3, A, B H tyep 3, A, B H type 1, 2, 3, A, B, Le-a, Le-b A, B H type 2, 3, B, Le-b H type 1/2, 3, B, Le-b H type 3, B, Le-a, Le-b B No binding

r124 H type 1, 2, 3, A, Le-b r258 H type 1, 3, A, Le-a r645 H type 2, A, Le-a rCV H type 1/2, A, Le-a, Le-b rW18 H type 1/2, A, B, Le-a, Le-b

1 1 3 3 4

8

3 4

2

1

Binding pattern (a)

N N

N

N N N

N N N N N

N

N N

N

N

R Q

K

R R R

R R R R R

R

R R

R

R

G G

G

G G G

G G G G G

G

G G

G

G

R R

E

K T T

R R T T D

K

S R

K

T

Q Q

E

E Q Q

Q Q Q Q N

N

N N

N

N

D D

D

D D D

D D D D D

D

D D

D

D

L L

L

V V V

I I V V I

D

D D

D

D

G G

G

G G G

G G G G G

H

H H

H

H

L L

A

L A A

T T A A L

N

E E

R

N

Q Q

Q

Q Q Q

Q Q Q Q Q

T

S S

S

T

R D

R

R R R

R R R R T

F

P K

K

Q

D N

N

N D D

N N N D T

Q

T I

T

F

H K

H

H H H

H H H H H

Q

K V

R

T

N N

H

N N N

N N N N G

D

N D

D

N

F F

F

F F F

F F F F F

L

L I

I

L

b

T T

T

T T T

T T T T T

G

G Q

E

S

P P

P

P P P

P P P P P

W

W W

W

W

G G

G

G G G

G G G G G

S

S S

S

S

267b 291c 292c 293c 300c 322b 327b 329b 331b 333b 334b 335c 341b 368c 373b 374b 375b 377b

GI- and GII-specific amino acid residues of putative HBGAs binding site

These residuls have been proven to be important for NoV binding to HBGAs by mutagenesis analyses [10, 11] These residues are located near the only P2 domain cavity or lie in the vicinity of the cavity and thereore infer to play roles in NoV-HBGAs binding [15] c These residuls were predicted to be important for NoV binding to HBGAs by X-ray crystallographic analysis [16]

a

GII

GI

Genogroup Genotype VLP

Table 1 Each genotype recognizes own HBGAs

600 Haruko Shirato

Non-secretor

type O

type B

type AB

2.0 1.5 1.0

0

Leb

0.5 Lea

Detection by ELISA

Binding of GI.1 VLP

Tri-B

Addition of VLPs

0

Tri-A

Coating of microplate with carbohydrates

0.4

H type 3

ELISA with carbohydrates

0.8

H type 2

Detection by ELISA

1.2

type A

Addition of VLPs

601

Binding of GI.1 VLP

H type 1

Coating of microplate with saliva

OD 492 nm / internal standard

ELISA with saliva

OD 492 nm / internal standard

Histo-Blood Group Antigen Recognition by Norovirus

Fig. 3 Two ELISA-based methods to detect and quantify NoV VLP attachment to HBGAs; saliva-VLP binding assay (upper) and synthetic carbohydrate-VLP binding assay (lower)

Herein, we describe two ELISA-based methods to detect and quantify NoV VLP attachment to HBGAs; the saliva-VLP binding assay and synthetic carbohydrate-VLP binding assay (Fig. 3). We found the sensitivity of the saliva-VLP binding assay to be better than that of the carbohydrate-VLP binding assay. However, the recognition of H type 3, Lea, and Leb antigens by NoV could be detected with the carbohydrate-VLP binding assay, but not with the saliva-VLP binding assay [10]. For this reason, we performed binding assays with two ELISA methods. Differences in the reactivities between saliva samples and synthetic carbohydrates may be due to structural differences between the synthetic products and authentic antigens, which are thought to be present on mucin or mucin-like molecules [9, 11].

2

Materials

2.1 Expression and Purification of Recombinant VLPs

1. Tn5 cells: an insect cell line from Trichoplusia ni (Invitrogen, San Diego, CA). 2. Ex-CELL 400 medium (JRH Biosciences, Lenexa, KS). 3. SW28 rotor (Beckman Instruments, Inc., Palo Alto, CA).

602

Haruko Shirato

4. SW50.1 rotor (Beckman). 5. Phosphate-buffered saline, pH 7.5 (PBS). 6. Bio-Rad protein assay kit (Bio-Rad Laboratories, Hercules, CA). 2.2 Hemagglutination Inhibition Assay

1. Anti-H lectin (Gamma Biologicals, Inc., Houston, TX). 2. Anti-A antibody (Gamma Biologicals, Inc.). 3. Anti-B antibody (Gamma Biologicals, Inc.). 4. Indicator O (Gamma Biologicals, Inc.). 5. Indicator A1 (Gamma Biologicals, Inc.). 6. Indicator B (Gamma Biologicals, Inc.).

2.3 Saliva-VLP Binding Assay

1. Supernatant of saliva (see step 1 of Subheading 3.2). 2. Convalescent-phase serum (see Note 1). 3. 96-Well polystyrene microplate for antigen-down immunoassays (see Note 2). 4. Coating buffer: 50 mM carbonate-bicarbonate buffer, pH 9.6 (Sigma, St. Louis, MO). 5. PBS containing 0.05% Tween 20 (PBS-T). 6. PBS containing 5% skim milk (SM/PBS). 7. Horseradish peroxidase-conjugated anti-rabbit immunoglobulin G (Zymed Laboratories Inc., San Francisco, CA). 8. o-Phenylenediamine.

2.4 Synthetic Carbohydrate-VLP Binding Assay

1. Synthetic multivalent carbohydrate-biotin reagents conjugated to polyacrylamide (CHO-PAA-biotin; Glycotech, Rockville, MD). 2. Streptavidin-precoated plates (Thermo Electron Corporation, Vantaa, Finland). 3. Coating buffer (see item 4 of Subheading 2.3). 4. PBS-T (see item 5 of Subheading 2.3). 5. SM/PBS (see item 6 of Subheading 2.3). 6. Horseradish peroxidase-conjugated anti-rabbit immunoglobulin G (Zymed Laboratories Inc., San Francisco, CA). 7. o-Phenylenediamine.

3

Methods

3.1 Expression and Purification of Recombinant VLPs

1. Prepare VLPs by infecting subconfluent Tn5 insect cells with the recombinant baculoviruses. Grow Tn5 cells at 27  C in Ex-CELL 400 medium.

Histo-Blood Group Antigen Recognition by Norovirus

603

2. Harvest the culture medium at 6 days postinfection. Centrifuge the medium at 1000  g for 10 min to remove cell debris and centrifuge further at 10,000  g for 30 min to remove baculoviruses. 3. Concentrate the VLPs in the supernatant by centrifugation at 100,000  g for 2 h at 4  C. Suspend the pellet in a solution containing CsCl (1.9 g/4.5 mL) and centrifuge at 120,000  g for 20 h at 10  C. Pool the peak fractions containing the VLPs, dilute with PBS, and centrifuge at 200,000  g for 2 h at 4  C. 4. Examine the purified VLPs by electron microscopy and confirm that the VLPs form particles similar to native NoVs. 5. Determine the protein concentration using the Bio-Rad protein assay kit with bovine serum albumin as the protein standard. 6. Store VLPs at 70  C for stable storage. Also, to avoid freezing and thawing, aliquot and store the VLPs (see Note 3). 3.2 Detection of Soluble ABH Antigens in Saliva by Hemagglutination Inhibition Assay

1. Collect saliva samples from healthy donors (see Notes 4–6). 2. Boil the samples for 10 min immediately after collection centrifuge for 5 min at 13,000  g. 3. Collect the clear supernatant and store at

30  C until use.

4. Assay the samples for the presence of H, A, and B antigens by hemagglutination inhibition as follows: Mix either 100 μL of anti-H lectin, 50 μL of anti-A antibody, or 50 μL anti-B antibody with an equal volume of undiluted supernatant of saliva and incubate for 10 or 20 min at 26  C (see Note 7). 5. Add 50 μL of 3–4% suspension of O, A1, or B red blood cell as an indicator. Leave the mixture at 26  C for 5 min and centrifuge at 125  g for 1 min. 6. Measure the amount of HBGAs in a semiquantitative manner [14] using serial twofold dilutions of the supernatant of saliva samples (1- to 256-fold dilutions) with the hemagglutination inhibition assay as described above (Fig. 4) (see Note 8).

3.3 ELISA-Based Binding Assay 1: Saliva-VLP Binding Assay

1. Add 100 μL of supernatants of saliva serially diluted twofold with coating buffer to each well of a 96-well microplate and incubate overnight at 37  C in a humid atmosphere. Use convalescent-phase serum from a patient infected with NoV as the internal standard. The wells with only coating buffer are used as the negative control. 2. Wash the wells three times with 300 μL of PBS-T and block with 200 μL of SM/PBS for 1 h at room temperature. 3. Wash the wells three times with 300 μL of PBS-T, add the VLPs (1 μg/mL) in 100 μL of 1% SM/PBS-T (see Note 9), and incubate for 1 h at 37  C.

604

Haruko Shirato

Serial 2-fold dilutions of the samples were prepared (1- to 256-fold dilution) Dilution ratio

1:1

100 µl Saliva 0 µl Saline solution

1:2

100 µl 100 µl

1:4

1:8

1:16

1:32

1:64

1:128

1:256

100 µl

100 µl

100 µl

100 µl

100 µl

100 µl

100 µl

100 µl

100 µl

100 µl

100 µl

100 µl

100 µl

100 µl

Discard 100 µl

Mix 100 µ

Add 50 µ

Dilution ratio 1:1 1:2 Secretor1 Secretor2 Nonsecretor No agglomerates ± Small aggregates of RBCs Large aggregates of RBCs

g for 1 min.

1:4

1:8

1:16

1:32

1:64

1:128

1:256

± ±

Fig. 4 Semi-quantitation of soluble H antigen

4. Wash the wells six times with 300 μL of PBS-T, add 100 μL of the rabbit anti-recombinant NoV VLP antiserum (1:2000) in 1% SM/PBS-T (see Note 9), and incubate for 1 h at 37  C. 5. Wash the wells six times with 300 μL of PBS-T, add 100 μL of horseradish peroxidase-conjugated anti-rabbit immunoglobulin G (1:4000) in 1% SM/PBS-T, and incubate for 1 h at 37  C. 6. Wash the wells six times with 300 μL of PBS-T, add 100 μL of o-phenylenediamine as a substrate, and incubate for 30 min at room temperature. 7. Add 50 μL of 4 N H2SO4 to stop the reaction and measure the optical density at 492 nm (see Note 10). Typical results of ELISA with saliva are shown in Fig. 3 (upper). 3.4 ELISA-Based Binding Assay 2: Synthetic Carbohydrate-VLP Binding Assay

1. Rehydrate synthetic multivalent carbohydrate-biotin reagents conjugated to polyacrylamide to 1 mg/mL with 0.3 M sodium phosphate buffer and dilute to 2.5 μg/mL with Tris-buffered saline. 2. Add the carbohydrate solution (100 μL per well) to streptavidin-precoated plates and incubate for 2 h at 37  C. 3. Block the plates with 300 μL/well of 5% SM/PBS by incubating overnight at 4  C. 4. Add VLPs (1 μg/mL in a 100-μL volume of 5% SM/PBS) and incubate for 4 h at 37  C.

Histo-Blood Group Antigen Recognition by Norovirus

605

5. Wash the wells six times with 300 μL of PBS-T, add 100 μL of the rabbit anti-recombinant NoV VLP antiserum (1:2000) in 5% SM/PBS, and incubate for 2 h at 37  C. 6. Wash the wells six times with 300 μL of PBS-T, add 100 μL of horseradish peroxidase-conjugated anti-rabbit immunoglobulin G in 5% SM/PBS, and incubate for 1 h at 37  C. 7. Wash the wells six times with 300 μL of PBS-T and detect binding using o-phenylenediamine (see Note 10). Typical results of ELISA with synthetic carbohydrate are shown in Fig. 3 (lower).

4

Notes 1. Patient sera should be obtained and managed according to the regulations and guidelines of the laboratory to which the researcher belongs. 2. No special plate is required for saliva coating. Antigen-coated plate used in ELISA may be used. 3. To maintain stable binding to HBGAs, it is important for VLPs to retain their structure. The temperature at which they are stored is an important factor affecting their stability. NoV VLPs stored at 70  C retain their original morphology for 12 months. 4. Before collecting saliva, since saliva contains personal information, it is necessary to take informed consent of the saliva provider. 5. Saliva collections from the saliva providers are performed under the same conditions. After lunch, brush the teeth and collect saliva 30 min later. For 30 min, do not eat, drink, or smoke. 6. Supernatant of saliva is highly reactive and suitable to detect the binding of NoV to HBGAs. 10 min of boiling should be done immediately after collection of saliva to inactivate the HBGA degrading enzyme. 7. In the protocol attached to the reagent, “reaction time 10 min” may be written, but there are some samples that do not show aggregation inhibition when the reaction time is short. Some samples cannot prevent aggregation when the reaction time is 10 min, and can prevent aggregation when the reaction time is extended to 20 min. For this reason, the sample used for salivaVLP binding assay should have a reaction time of 20 min to accurately determine the presence or absence of HBGAs. 8. The ABH and Lewis HBGAs are carbohydrate epitopes present in several tissues of the human body. Types 1 and 3 chain ABH

606

Haruko Shirato

HBGAs are present on mucosal epithelial cell surface and in salivary secretions, with variations in the carbohydrate milieu in different individuals based on their secretor status and blood type. Individuals with null FUT2 alleles cannot synthesize ABH antigens in secretions and are called nonsecretors, although ABH antigens can be expressed in erythrocytes via FUT1. The FUT2 alleles of Caucasian nonsecretors are completely inactivated by nonsense mutations, whereas those of Asian nonsecretors are incompletely inactivated by missense mutations. Thus, Asian nonsecretors are incomplete nonsecretors and produce a small amount of ABH HBGAs in secretions. 9. It is important to determine the appropriate concentration of VLPs and appropriate dilution ratio of supernatant of saliva and rabbit anti-VLPs in comparison experiments of the binding amount. 10. Although the virus strains belong to the same genotype, sometimes the results may not match due to amino acid mutation. Norovirus is prone to amino acid displacement. The amino acid residues 267N, 291R, 292G, 293D, 300N, 322D, 327D, 328W, 329H, 331N, 333T, 334Q, 335F, 339S, 341T, 364I, 368N, 373L, 374S, 375W, 377S, and 430A (NV/68 numbering) in the P2 domain are predicted to be important for HBGA binding (also see Table 1). The finding that a single amino acid change in the P domain resulted in a change in the pattern of HBGA binding can explain the inconsistency in the results. References 1. Jiang X, Wang M, Wang K et al (1993) Sequence and genomic organization of Norwalk virus. Virology 195:51–61 2. Glass PJ, White LJ, Ball JM et al (2000) Norwalk virus open reading frame 3 encodes a minor structural protein. J Virol 74:6581–6591 3. Prasad BV, Hardy ME, Dokland T et al (1999) X-ray crystallographic structure of the Norwalk virus capsid. Science 286:287–290 4. Xi JN, Graham DY, Wang KN et al (1990) Norwalk virus genome cloning and characterization. Science 250:1580–1583 5. Green KY, Lew JF, Jiang X et al (1993) Comparison of the reactivities of baculovirusexpressed recombinant Norwalk virus capsid antigen with those of the native Norwalk virus antigen in serologic assays and some epidemiologic observations. J Clin Microbiol 31:2185–2191

6. Jiang X, Wang M, Graham DY et al (1992) Expression, self-assembly, and antigenicity of the Norwalk virus capsid protein. J Virol 66:6527–6532 7. Prasad BV, Rothnagel R, Jiang X et al (1994) Three-dimensional structure of baculovirusexpressed Norwalk virus capsids. J Virol 68:5117–5125 8. Harrington PR, Lindesmith L, Yount B et al (2002) Binding of Norwalk virus-like particles to ABH histo-blood group antigens is blocked by antisera from infected human volunteers or experimentally vaccinated mice. J Virol 76:12335–12343 9. Huang P, Farkas T, Zhong W et al (2005) Norovirus and histo-blood group antigens: demonstration of a wide spectrum of strain specificities and classification of two major binding groups among multiple binding patterns. J Virol 79:6714–6722

Histo-Blood Group Antigen Recognition by Norovirus 10. Shirato H, Ogawa S, Ito H et al (2008) Noroviruses distinguish between type 1 and type 2 histo-blood group antigens for binding. J Virol 82:10756–10767 11. Shirato-Horikoshi H, Ogawa S, Wakita T et al (2007) Binding activity of norovirus and sapovirus to histo-blood group antigens. Arch Virol 152:457–461 12. Choi JM, Hutson AM, Estes MK et al (2008) Atomic resolution structural characterization of recognition of histo-blood group antigens by Norwalk virus. Proc Natl Acad Sci U S A 105:9175–9180 13. Kubota T, Kumagai A, Ito H et al (2012) Structural basis for the recognition of Lewis antigens by genogroup I norovirus. J Virol 86:11138–11150

607

14. (1981) Appendix 2: saliva testing for ABH and Lewis. In: Sidmann FK (ed) Technical manual of the American Association of Blood Banks, 8th edn. Lippincott JB, Philadelphia, PA, pp 122–123 15. Chakravarty S, Hutson AM, Estes MK et al (2005) Evolutionary trace residues in noroviruses: importance in receptor binding, antigenicity, virion assembly, and strain diversity. J Virol 79:554–568 16. Tan M, Huang P, Meller J et al (2003) Mutations within the P2 domain of norovirus capsid affect binding to human histo-blood group antigens: evidence for a binding pocket. J Virol 77:12562–12571

Chapter 52 FAM3B/PANDER-Like Carbohydrate-Binding Domain in a Glycosyltransferase, POMGNT1 Hiroshi Manya, Naoyuki Kuwabara, Ryuichi Kato, and Tamao Endo Abstract Protein O-mannose β1,2-N-acetylglucosaminyltransferase 1 (POMGNT1) is one of the gene products responsible for α-dystroglycanopathy, which is a type of congenital muscular dystrophy caused by Omannosyl glycan defects. The originally identified function of POMGNT1 was as a glycosyltransferase that catalyzes the formation of the GlcNAcβ1-2Man linkage of O-mannosyl glycan, but the enzyme function is not essential for α-dystroglycanopathy pathogenesis. Our recent study revealed that the stem domain of POMGNT1 has a carbohydrate-binding ability, which recognizes the GalNAcβ1-3GlcNAc structure. This carbohydrate-binding activity is required for the formation of the ribitol phosphate (RboP)-3GalNAcβ1-3GlcNAc structure by fukutin. This protocol describes methods to assess the carbohydrate-binding activity of the POMGNT1 stem domain. Key words α-Dystroglycan, α-Dystroglycanopathy, Carbohydrate-binding, Fukutin, FAM3B, Glycosyltransferase, PANDER, POMGNT1, O-Mannosyl glycan

1

Introduction Recent studies have revealed that there are various structures of Omannosyl glycans in mammals. Based on the GlcNAc linkage to core Man residues, the structures of O-mannosyl glycans can be classified into three groups: core M1, GlcNAcβ1-2Man; core M2, GlcNAcβ1-2(GlcNAcβ1-6)Man; and core M3, GalNAcβ13GlcNAcβ1-4Man [1]. α-Dystroglycan (α-DG) is a component of the dystrophin-glycoprotein complex that contributes to the attachment between the extracellular matrix and the intracellular cytoskeleton. The core M3 glycan of α-DG is required for its binding to extracellular matrix molecules, such as laminin, and the GlcA-Xyl repeat on the core M3 glycan is known to be the laminin-binding epitope (IIH6 epitope) [1, 2]. We previously reported the complete structure of core M3 glycan, [3GlcAβ13Xylα1]n-3GlcAβ1-4Xylβ1-4RboP-RboP-3GalNAcβ13GlcNAcβ1-4(phospho-6)Manα1 [1, 3, 4]. Currently, the core M3

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_52, © Springer Science+Business Media, LLC, part of Springer Nature 2020

609

Hiroshi Manya et al. R129

610

1

38 59

251 300 Stem

660 Catalytic

Fig. 1 Schematic representation of human POMGNT1. A cytoplasmic tail (residues 1–37, gray), a transmembrane domain (residues 38–58, black), a stem domain (residues 59–250, blue), a linker region (residues 251–299, yellow), and a catalytic domain (residues 300–660, green) [7, 8]. The numbers above the boxes indicate the amino acid residue numbers of human POMGNT1. The arrows indicate the amino acids required for carbohydrate-binding activity (R129 and R179) or catalytic activity (W473 and M477)

glycan has been identified only in α-DG. A defect in core M3 glycan on α-DG is the primary cause of α-dystroglycanopathy, a group of congenital muscular dystrophies with brain malformation [1, 5, 6]. Protein O-mannose β1,2-N-acetylglucosaminyltransferase 1 (POMGNT1) is a Golgi-localized type II transmembrane glycosyltransferase and is composed of the following four domains, from the N-terminal side: a cytoplasmic tail, a transmembrane domain, a stem domain, and a catalytic domain (Fig. 1) [7, 8]. POMGNT1 has a greater number of amino acids in the stem domain than other glycosyltransferases. The function of the catalytic domain has been identified as catalyzing the formation of the GlcNAcβ1-2Man linkage by transferring GlcNAc from UDP-GlcNAc in core M1 and core M2 glycan biosynthesis [1, 5]. Although POMGNT1 is one of the gene products responsible for α-dystroglycanopathy, it is unrelated directly to the core M3 glycan biosynthesis [1, 5]. However, a defect in POMGNT1 is known to abolish not only the core M1 and core M2 structures but also the core M3 structure [1, 6]. Recently, we found that the stem domain of POMGNT1 has carbohydratebiding ability [8] and recognizes the GalNAcβ1-3GlcNAc structure on core M3 glycan; it is required for the formation of the RboP3GalNAcβ1-3GlcNAc structure by fukutin (Fig. 2). We also reported that fukutin forms a complex with POMGNT1 [9]. Thus, the carbohydrate-binding activity of the POMGNT1 stem domain plays a role in RboP modification by recruiting fukutin to the phosphorylated core M3 structure [GalNAcβ13GlcNAcβ1-4(phosphate-6)Man]. We confirmed the importance of the carbohydrate-binding activity by a rescue experiment of IIH6 biosynthesis using POMGNT1-deficient cells [8]. The carbohydrate-binding specificity of the POMGNT1 stem domain was determined by frontal affinity chromatography (FAC) and surface plasmon resonance (SPR) [8]. The FAC analysis showed that the stem domain bounds only to β-linked monosaccharides but not to any N-glycans, glycolipid saccharides, or other complex carbohydrate structures. We then tested the binding of the stem domain to the core structure of O-mannosyl glycan using

Carbohydrate-Binding Activity of POMGNT1

611

b2

A

DG

UDP-

DG

Catalytic

POMGNT1

Stem

B Catalytic Fukutin

CDP-

Stem

POMGNT1

P

DG

Rbo

Rbo

P

3

GalNAc,

GlcNAc,

P6 b3

DG

b4 Man, P phosphate

Fig. 2 Function of stem and catalytic domains of POMGNT1. (a) The catalytic domain of POMGNT1 forms the GlcNAcβ1-2Man linkage in core M1 and core M2 glycans. (b) POMGNT1 and fukutin form a complex. The stem domain of POMGNT1 recognizes the core M3 structure [GalNAcβ1-3GlcNAcβ1-4 (phospho-6)Man]. Fukutin is recruited to core M3 by the fukutin-POMGNT1 complex and efficiently transfers RboP to the C3 position of GalNAc

SPR, and the results showed that the specific ligands are GlcNAcβ12Man on core M1 and GalNAcβ1-3GlcNAc on core M3. Furthermore, residues R129 and R179 in the stem domain are required for binding to these carbohydrates. Our X-ray crystal structural analysis demonstrated that the folding of the carbohydrate-biding domain of POMGNT1 resembles a family with the sequence similarity 3B/pancreatic-derived factor (FAM3B/PANDER) [8], which is a cytokine-like FAM3 family member protein. Although it has been suggested that FAM3B is a mediator involved in glucose homeostasis [10], the molecular mechanism remains uncertain. In this chapter, we describe the assay protocols to assess the carbohydrate-binding activity of the POMGNT1 stem domain using SPR and rescue experiments in POMGNT1-deficient cells.

612

2

Hiroshi Manya et al.

Materials

2.1 Lectin-Like Activity of POMGNT1 Stem Domain 2.1.1 Preparation of Recombinant Stem Domain

1. pGEX-GST-POMGNT1stem: The stem domain of human POMGNT1 (residues 92–288) is cloned into pGEX-6P1 (GE Healthcare, Buckinghamshire, England), which vector expresses the stam domain as a glutathione-S-transferase (GST)-fusion protein. 2. pGEX-GST-R129Astem and pGEX-GST-R179Astem: Two missense mutants of the stem domain (R129A and R179A) are made by PCR-based site-directed mutagenesis. R129 or R179 residues are required for the interaction with carbohydrate and are conserved within the FAM3 superfamily [8]. 3. LB broth (Becton Dickinson, Franklin Lakes, NJ) supplemented with 50 μg/mL ampicillin (as will be described in item 6). 4. LB agar plate [1.5% (w/v) agar] supplemented with 50 μg/mL ampicillin (as will be described in item 6). 5. Isopropyl-D-thiogalactopyranoside (IPTG) (Nacalai Tesque, Kyoto, Japan): Prepare a 1 M stock solution in water, sterilize by filtration, and store at 20  C. 6. Ampicillin sodium salt (Nacalai Tesque): Prepare a 50 mg/mL stock solution in water, sterilize by filtration, and store at 20  C. 7. Glutathione-Sepharose column (GSTrap, 1 mL, GE Healthcare). Store at 4  C. 8. GST-binding buffer [phosphate-buffered saline (PBS)]: Prepare a 10 stock with 1.37 M NaCl, 27 mM KCl, 100 mM Na2HPO4, and 18 mM KH2PO4 (adjust to pH 7.4 with HCl if necessary) and store at room temperature. Prepare a working solution by diluting one part stock with nine parts water and store at 4  C. 9. PreScission buffer: 50 mM Tris–HCl, 100 mM NaCl, 1 mM EDTA, 1 mM DTT, pH 8.0. Store at 4  C. 10. PreScission protease (GE Healthcare): The required amount of enzyme (>2 U enzyme for 100 μg of GST-fusion protein) is diluted in 1 mL of PreScission buffer. Prepare fresh for each use. 11. Storage buffer: 20 mM Tris–HCl, 150 mM NaCl, 1 mM DTT, pH 7.5. Store at 4  C. 12. Centrifugal ultrafiltration device: Amicon ultra centrifugal filter unit (Merck Millipore, Burlington, MA).

2.1.2 Carbohydrate-Binding Assay of Stem Domain

1. Biacore T-200 (GE Healthcare). 2. CM5 sensor chip (GE Healthcare).

Carbohydrate-Binding Activity of POMGNT1

613

3. Running buffer: HBS-P buffer [10 mM HEPES, 150 mM NaCl, 0.005% (v/v) Surfactant P20, pH 7.2]. 4. The recombinant stem domains (POMGNT1stem, R129Astem, and R179Astem) are used as ligands and are immobilized onto the sensor chip. 5. p-Nitrophenol (pNP)-sugar derivatives are used as analytes: Man-α-pNP, Man-β-pNP, GlcNAc-α-pNP, GlcNAc-β-pNP, GlcNAc-β1,2-Man-pNP (core M1 type), GalNAc-β1,3GlcNAc-β-pNP (core M3 type), and GalNAc-β1,4GlcNAc-β-pNP (LacdiNAc type) are purchased from SigmaAldrich (Merck) and Tokyo Chemical Industry (Tokyo, Japan). 2.2 Functional Assay of Stem Domain 2.2.1 Cell Culture, Transfection, and Preparation of Cell Membrane Fraction

1. POMGNT1-deficient cells: IIH6 epitope-negative HEK293 cells are prepared by knocking out POMGNT1 using the CRISPR/Cas9 system [8]. 2. pcDNA3.1-POMGNT1 expression plasmid: Human cDNA encoding full-length POMGNT1 is inserted into the mammalian expression vectors, pcDNA3.1/Zeo and pSecTag2/Hygro (Thermo Fisher Scientific, Waltham, MA). 3. Expression vectors of mutant proteins (R129A, R179A, or W473A/M477A) are made by PCR-based site-directed mutagenesis using pcDNA3.1-POMGNT1 as a template. R129 and R179 residues are required for carbohydrate binding, and W473 and M477 residues are required for glycosyltransferase activity of POMGNT1 [8]. 4. Dulbecco’s modified Eagle’s medium (DMEM), fetal bovine serum (FBS), 100 penicillin-streptomycin-glutamine liquid (PC-SM-Gln, 10,000 U/mL penicillin, 10,000 μg/mL streptomycin, 29.2 mg/mL glutamine), and Lipofectamine 3000 transfection reagent (Thermo Fisher Scientific). 5. PBS is described in item 8 of Subheading 2.1.1. 6. Silicone blade cell scraper (Sumilon, Sumitomo Bakelite Co., Tokyo, Japan). 7. Homogenization buffer: 10 mM Tris–HCl, pH 7.4, 1 mM EDTA, 250 mM sucrose (SET buffer) with protease inhibitor cocktail (EDTA-free, Nakalai Tesque). Store at 4  C. Add the protease inhibitor cocktail directly before use.

2.2.2 Enrichment of α-DG and Western Blotting

Western blotting is carried out in accordance with standard methods. Please refer to experimental textbooks. Some specific reagents are described below. 1. Solubilization buffer: 50 mM Tris–HCl, pH 7.5, 1% (v/v) Triton X-100, 500 mM NaCl, and protease inhibitor cocktail. Store at 4  C. Add the protease inhibitor cocktail directly before use.

614

Hiroshi Manya et al.

2. 25% wheat germ agglutinin (WGA) agarose conjugate (WGA-agarose, J-Oil Mills, Tokyo, Japan): WGA-agarose beads are suspended in 50 mM Tris–HCl, pH 7.5, 500 mM NaCl, 1% (v/v) Triton X-100 at a 25% (v/v) slurry concentration. Store at 4  C. 3. WGA binding buffer: 50 mM Tris–HCl, pH 7.5, 500 mM NaCl, 1% (v/v) Triton X-100, 2 mM MnCl2, 2 mM CaCl2, and 2 mM MgCl2. Prepare fresh for use. 4. 4 loading buffer (Modified Laemmli [11] buffer): 250 mM Tris (do not adjust pH), 8% (w/v) SDS, 40% (w/v) glycerol, 2.84 mM 2-mercaptoethanol, 0.005% (w/v) bromophenol blue. Store at 20  C. 5. Primary antibodies: Mouse monoclonal antibody against the IIH6 epitope of α-DG (IIH6, Merck); rat monoclonal antibody against α-DG core protein (3D7) is produced by using human α-DG-Fc fusion protein [12]; rabbit antisera specific to POMGNT1 C-terminus is produced by using synthetic peptides corresponding to residues 649–660 (KEEGAPGAPEQT) of human POMGNT1 [13]. 6. Secondary antibodies: Anti-mouse, anti-Rat, and anti-rabbit IgG conjugated with horseradish peroxidase (HRP, GE Healthcare). 7. PBS containing 0.5% (v/v) Tween 20 (TPBS): Add 2.5 mL of 20% Tween 20 to 1000 mL of PBS and store at 4  C. 8. Blocking buffer: TPBS containing 5% (w/v) skim milk. Prepare fresh for use. 9. Enhanced chemiluminescent reagent kit: Any commercially available reagent such as ECL Western Blotting Substrate or ECL Plus Western Blotting Substrate (Thermo Fisher Scientific). 2.2.3 Glycosyltransferase Assay

The POMGNT1 glycosyltransferase activity is based on the amount of [3H]GlcNAc transferred from UDP-GlcNAc to mannosylpeptide (Ac-AAPT (Man) PVAAP-NH2) [14]. The reaction product is purified via reversed-phase HPLC, and the radioactivity is measured by a liquid scintillation counter (see Note 1). 1. UDP-N-acetyl-D-glucosamine (UDP-GlcNAc, SigmaAldrich): Prepare a 1 mM stock solution in water and store at 20  C. 2. UDP-GlcNAc [glucosamine-6-3H(N)] (UDP-[3H]-GlcNAc, 0.74–1.66 TBq/mmol, PerkinElmer, Inc., Wellesley, MA): Store at 20  C. 3. Mannosylpeptide (Ac-AAPT (Man) PVAAP-NH2, see Note 2): Prepare a 2 mM stock solution in water and store at 20  C.

Carbohydrate-Binding Activity of POMGNT1

615

4. POMGNT reaction buffer: 140 mM MES (adjust to pH 7.0 with NaOH), 2% (v/v) Triton X-100, 5 mM AMP, 200 mM GlcNAc, 10% (w/v) glycerol, 10 mM MnCl2. Store at 20  C without MnCl2. MnCl2 is added directly before use. 5. Reversed-phase column for HPLC: Wakopak 5C18-200 column (4.6  250 mm, Fujifilm Wako Pure Chem. Corp., Osaka, Japan). 6. 0.085% (v/v) trifluoroacetic acid (TFA) in water (solvent A): Add 0.85 mL of TFA to 1000 mL of HPLC-grade water and degas with an aspirator before use. 7. 0.085% (v/v) TFA in acetonitrile (solvent B): Add 0.85 mL of TFA to 1000 mL of HPLC-grade acetonitrile and degas by sonication before use. 8. Liquid scintillation cocktail: Any commercially available reagent such as Clear-sol II (Nacalai Tesque).

3

Methods

3.1 Lectin-Like Activity of POMGNT1 Stem Domain 3.1.1 Preparation of Recombinant POMGNT1stem

1. BL21(DE3) Escherichia coli cells are transformed with expression plasmid (pGEX-GST-POMGNT1stem, pGEX-GSTR129Astem, or pGEX-GST-R179Astem). Cultures are prepared by growing a single colony overnight in LB broth at 37  C. The overnight culture is then used to inoculate a fresh 50-mL culture, which is grown at 37  C to A600 ¼ 0.5. At this point, 1 mM IPTG is added to the culture to induce protein expression. The induced cells are grown in parallel for an additional 4 h at 37  C and then harvested by centrifugation at 6000  g for 15 min at 4  C. 2. The cell pellet is suspended in 10 mL of PBS, pH 7.4, and the cells are broken with a tip-type sonicator (see Note 3). The cell supernatant is recovered by ultracentrifugation at 100,000  g for 1 h. 3. Recombinant stem domains are purified from the supernatant with a GSTrap column as follows. Pre-equilibrate the GSTrap column with 10 mL of PBS. Load the supernatant onto the column and wash with >20 mL of PBS. To remove the GST-tag by on-column digestion, the buffer is replaced with PreScission buffer, and 1 mL of PreScission protease is injected onto the column. Following injection, the column is closed and incubated at 4  C for 16 h. Before elution, a new GSTrap column is connected downstream to the proteolytic reaction column in order to bind an excess of PreScission protease. The cleaved stem domains are eluted with >2 mL of PreScission buffer (see Note 4).

616

Hiroshi Manya et al.

4. The buffer containing the purified stem domains is replaced with storage buffer. 5. The purified stem domains are concentrated to 15–30 mg/mL by ultrafiltration, frozen in liquid nitrogen, and stored at 80  C. 3.1.2 Carbohydrate-Binding Assay of POMGNT1stem

The carbohydrate-binding activity is measured using SPR spectroscopy (Biacore T-200) [8]. 1. Ligand (POMGNT1stem, R129Astem, or R179Astem) is directly immobilized on a CM5 sensor chip by amine coupling. 50 μg/ mL ligand in HBS-P buffer is injected at a flow rate of 30 μL/ min, resulting in immobilization of the ligand corresponding to approximately 15,000–20,000 resonance units (RU). 2. For measuring the interaction of the analyte with the ligand, inject the 0–12.5 mM analyte with running buffer. The net response is calculated by subtracting the background response from the binding response. 3. The results are analyzed using the Biacore T-200 evaluation software. The Kd values are calculated by the steady-state analysis.

3.2 Functional Assay of Stem Domain by Rescue Experiment in IIH6 Epitope Synthesis 3.2.1 Cell Culture, Transfection, and Preparation of Cell Membrane Fraction

α-DG is enriched from the solubilized membrane fraction using WGA-agarose. IIH6 epitope synthesis is assayed by the western blot analysis using anti-IIH6 antibody and anti-α-DG core protein antibody (3D7), as previously described [3, 12].

1. Wild-type and POMGNT1-deficient HEK293 cells are maintained in DMEM supplemented with 10% FBS, 2 mM L-glutamine, and 100 U/mL penicillin, and 50 μg/mL streptomycin at 37  C with 5% CO2. 2. When the cells reach 90% confluence, the expression plasmid for POMGNT1 (wild type, R129A, R179A, or W473A/ M477A) is transfected using Lipofectamine 3000 reagent according to the manufacturer’s instructions (see Note 5). 3. At 24 h after transfection, the cells are subcultured to reduce their density to 70% confluence to prevent overgrowth (see Note 5). 4. After 4 days of culturing, the transfectants are collected. The culture supernatants are removed by aspiration, and the cells are rinsed gently with cold PBS. Next, 5 mL of cold PBS is added, and the cells are scraped into centrifugal tubes and washed with 10 mL of cold PBS. The cells are collected by centrifugation at 1000  g for 10 min at 4  C (see Note 6). 5. The cell pellet is homogenized with a tip sonicator in 1 mL of homogenization buffer (see Note 7). After centrifugation at

Carbohydrate-Binding Activity of POMGNT1

617

900  g for 10 min, the supernatant is dispensed in two tubes and subjected to ultracentrifugation at 100,000  g for 1 h. The precipitates are used as the membrane fraction (see Note 8). The sample in one tube is used to determine the protein concentration and to assess the IIH6 epitope expression by western blotting; the remaining sample in another tube is used to measure the POMGNT1 activity. 3.2.2 Enrichment of α-DG and Western Blotting

1. A 200 μL aliquot of solubilization buffer is added to 200 μg of the membrane fraction pellet (100,000  g precipitate). After standing at 30 min on ice, the membrane fraction is solubilized with moderate pipetting until it became transparent (the final concentration of total protein is 1 mg/mL). 2. Add 80 μL of 25% WGA-agarose beads and 220 μL lectin binding buffer to 200 μL of the solubilized fraction and incubate at 4  C for 2 h (i.e., the final concentration of the mixture is 0.4 mg/mL membrane fraction, 20 μL WGA-agarose beads, 1 mM MnCl2, 1 mM CaCl2 and 1 mM MgCl2 in a total volume of 500 μL). 3. After incubation, the beads are washed three times with 1 mL of WGA-binding buffer and boiled for 3 min in 20 μL of 2 SDS-PAGE loading buffer. The supernatants are collected for western blotting to detect IIH6 epitope and α-DG core protein. 4. To confirm expression of wild-type and mutant POMGNT1s, the membrane fractions are solubilized and boiled in 1 SDS-PAGE loading buffer. 5. The membrane fractions (20 μg) or enriched proteins (10 μL) are separated by SDS-PAGE (7.5% gel), and proteins are transferred to a polyvinylidene difluoride (PVDF) membrane. 6. The membranes are blocked with blocking buffer for 1 h at room temperature. The membranes are then incubated with primary antibody in blocking buffer for 2 h at room temperature. After washing with TPBS (3 times  5 min), the membranes are incubated with HRP-conjugated secondary antibodies in blocking buffer for 1 h at room temperature. After washing with TPBS (3 times  5 min), the proteins that bound to the antibody are visualized with ECL.

3.2.3 Glycosyltransferase Assay of POMGNT1

1. A 10 μL aliquot of 1 mM UDP-GlcNAc, 10 μL of UDP-[3H] GlcNAc (100,000 dpm/nmol), and 10 μL of 2 mM mannosylpeptide are mixed in microcentrifuge tube and dried with a centrifugal evaporator. 2. A 200 μg sample of membrane fraction is solubilized with 100 μL of POMGNT1 reaction buffer in a bath sonicator on ice, with moderate pipetting until it became transparent. 20 μL

618

Hiroshi Manya et al.

(50 μg) of the solubilized membrane fraction is added to the dried substrate (prepared in step 1), vortexed gently, and then incubated at 37  C for 2 h. The reaction is stopped by boiling at 100  C for 3 min. Water (180 μL) is added to the reaction mixture, which is then filtered with a centrifugal filter device. 3. The filtrate is analyzed by reversed-phase HPLC with the following conditions: The gradient solvents are aqueous 0.085% TFA (solvent A) and acetonitrile containing 0.085% TFA (solvent B). The mobile phase consists of only solvent A for 5 min, and a linear gradient to 25% of solvent B for 20 min. The peptide separation is monitored by measuring the absorbance at 214 nm, and the radioactivity of each fraction (1 mL) is measured by liquid scintillation counting.

4

Notes 1. Because the GlcNAc-mannosylpeptide, which is a product of the POMGNT1 reaction, is not separated from mannosylpeptide by HPLC, UDP-[3H]-GlcNAc is used to quantitatively analyze the product. 2. The mannosylpeptide is not commercially available, but it is possible to use benzyl-α-D-mannopyranoside (Sigma-Aldrich) as a substitute, which is commercially available [15]. 3. Semitranslucent cell suspensions are obtained by 3-s sonication with 3-s interval for 5–10 min. 4. If necessary, the sample is further purified by anion-exchange chromatography (Mono Q column) and gel filtration chromatography (Superdex200 column) (GE Healthcare) [13]. 5. Use a high concentration of cells for transfection because a high transfection efficiency is necessary to express the detectable levels of the IIH6 epitope. Furthermore, avoid overgrowth of the cells after transfection. 6. The cell pellet can be stored at 80  C after removal of PBS. 7. Typical sonication conditions to reach semitranslucent cell suspensions are: 10 cycles of 0.6-s pulses with 0.4-s interval; this procedure is repeated again. 8. The precipitate can be stored at 80  C after removing the supernatant.

Carbohydrate-Binding Activity of POMGNT1

619

Acknowledgments This work was partially supported by the Japan Society for the Promotion of Science Grants (JSPS) JP17H03987 to H.M., the National Center of Neurology and Psychiatry (NCNP) Intramural Research Grant 29-4 to T.E., and the Platform Project for Supporting Drug Discovery and Life Science Research (Basis for Supporting Innovative Drug Discovery and Life Science Research (BINDS)) from AMED JP18am0101083 and JP18am0101071 to R.K. References 1. Manya H, Endo T (2017) Glycosylation with ribitol-phosphate in mammals: new insights into the O-mannosyl glycan. Biochim Biophys Acta 1861:2462–2472 2. Yoshida-Moriguchi T, Yu L, Stalnaker SH et al (2010) O-Mannosyl phosphorylation of alphadystroglycan is required for laminin binding. Science 327:88–92 3. Kanagawa M, Kobayashi K, Tajiri M et al (2016) Identification of a post-translational modification with ribitol-phosphate and its defect in muscular dystrophy. Cell Rep 14:2209–2223 4. Manya H, Yamaguchi Y, Kanagawa M et al (2016) The muscular dystrophy gene TMEM5 encodes a ribitol β1,4-xylosyltransferase required for the functional glycosylation of dystroglycan. J Biol Chem 291:24618–24627 5. Yoshida A, Kobayashi K, Manya H et al (2001) Muscular dystrophy and neuronal migration disorder caused by mutations in a glycosyltransferase, POMGnT1. Dev Cell 1:717–724 6. Michele DE, Barresi R, Kanagawa M et al (2002) Post-translational disruption of dystroglycan-ligand interactions in congenital muscular dystrophies. Nature 418:417–422 7. Akasaka-Manya K, Manya H, Kobayashi K et al (2004) Structure-function analysis of human protein O-linked mannose β1,2-N-acetylglucosaminyltransferase 1, POMGnT1. Biochem Biophys Res Commun 320:39–44 8. Kuwabara N, Manya H, Yamada T et al (2016) Carbohydrate-binding domain of the POMGnT1 stem region modulates O-

mannosylation sites of alpha-dystroglycan. Proc Natl Acad Sci U S A 113:9280–9285 9. Xiong H, Kobayashi K, Tachikawa M et al (2006) Molecular interaction between fukutin and POMGnT1 in the glycosylation pathway of alpha-dystroglycan. Biochem Biophys Res Commun 350:935–941 10. Wilson CG, Robert-Cooperman CE, Burkhardt BR (2011) PANcreatic-DERived factor: novel hormone PANDERing to glucose regulation. FEBS Lett 585:2137–2143 11. Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680–685 12. Ohtsuka Y, Kanagawa M, Yu CC et al (2015) Fukutin is prerequisite to ameliorate muscular dystrophic phenotype by myofiber-selective LARGE expression. Sci Rep 5:8316 13. Xin X, Akasaka-Manya K, Manya H et al (2015) POMGNT1 is glycosylated by mucin-type Oglycans. Biol Pharm Bull 38:1389–1394 14. Takahashi S, Sasaki T, Manya H et al (2001) A new β-1,2-N-acetylglucosaminyltransferase that may play a role in the biosynthesis of mammalian O-mannosyl glycans. Glycobiology 11:37–45 15. Akasaka-Manya K, Manya H, Mizuno M et al (2011) Effects of length and amino acid sequence of O-mannosyl peptides on substrate specificity of protein O-linked mannose β1,2N-acetylglucosaminyltransferase 1 (POMGnT1). Biochem Biophys Res Commun 410:632–636

Chapter 53 Mannose-Specific Oyster Lectin CGL1 Hideaki Unno and Tomomitsu Hatakeyama Abstract A novel mannose-specific lectin, named CGL1 (15.5 kDa), was isolated from the oyster Crassostrea gigas. Characterization of CGL1 revealed that it has strict specificity for the mannose monomer and for high mannose-type N-glycans (HMTGs). The primary and crystal structure of CGL1 did not show any homology with known lectins. These characteristics of CGL1 may be helpful as a research tool and for clinical applications. We show a purification protocol of CGL1 from the Pacific oyster. Key words Mannose-specific, CGL1, Oyster, Crassostrea gigas

1

Introduction As with many invertebrates, resistance against infectious diseases in shell fish is provided by an innate immune system. In the innate immune system of bivalves such as oysters, clams and mussels, pattern recognition receptors are commonly associated with host defense responses against infection [1]. These pattern recognition receptors include lectins, which are thought to play a crucial role in the innate immune system through specific binding to polysaccharide-coated pathogenic bacteria. Several lectins have been identified in bivalves and they most frequently belong to the C-type lectin and galectin families, two of the major animal lectin groups with conserved folds [2, 3]. Lectins are ubiquitous in living organisms (from microorganisms and viruses to humans). Some of the lectin families are outspread in distant species and include lectins with a variety of specificities for carbohydrates. Lectins capable of binding with mannose have been confirmed in a number of lectin families where most of these do not have strict specificity for mannose alone. Some lectins have demonstrated binding specificity only for high mannose-type N-glycans (HMTGs) [2, 3]. While there are various folding patterns of these lectins and binding regions on the HMTGs, these lectins also exhibit anti-HIV activity by binding to the HMTGs on gp120, an envelope protein of HIV

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_53, © Springer Science+Business Media, LLC, part of Springer Nature 2020

621

622

Hideaki Unno and Tomomitsu Hatakeyama

[4]. Therefore, lectins with binding specificities for HMTGs can be used as potential microbicides in preventing HIV transmission. We isolated a novel mannose-specific lectin, named CGL1, from the Pacific oyster Crassostrea gigas. Characterization of CGL1 revealed a strict specificity for the mannose monomer and HMTGs [5]. Primary and tertiary structures do not have any similarity with those of the known lectins but do have similarity with the primary structures of the proteins in the natterin family whose function appears to involve toxicity [6]. The crystal structure of the CGL1 revealed a unique homodimer in which each protomer was composed of two domains related by a pseudo twofold axis (Fig. 1). Complex structures of CGL1 with mannose molecules showed that residues have eight hydrogen bond interactions with O1, O2, O3, O4, and O5 hydroxyl groups of mannose (Fig. 2). The complex interactions that are not observed with other mannose-binding lectins revealed the structural basis for the strict specificity for mannose. These characteristics of CGL1 may be helpful as a research tool and for clinical applications. An important step required before analysis is the purification of CGL1. Native CGL1 can be relatively easily purified from the oyster. Although we also succeeded in expressing recombinant CGL1 using Escherichia coli cells, the binding affinity of native CGL1 was found to be higher than that of the recombinant lectin. This chapter, therefore, describes the purification of CGL1 from C. gigas.

2

Materials CGL1 is purified from C. gigas by affinity chromatography using a mannose-immobilized resin and gel-filtration chromatography. The preparative procedure of the mannose-immobilized resin is based on a method reported previously [7]. All solutions should be prepared using deionized water. Carbohydrate solutions are prepared immediately before use. In this protocol, the suction– filtration system is used for washing the resin and is separated from the solution. “Room temperature” refers to temperatures between 15 and 25  C and “overnight” refers to time duration between 12 and 24 h. 1. Carbonate buffer: 50 mM NaHCO3, 0.5 M Na2CO3.

2.1 Mannose-Immobilized Resin

2. Mannose.

2.1.1 Reagents/Solutions

4. Divinyl sulfone.

3. Cellulose resin (e.g., Cellufine GCL-2000, JNC Corporation). 5. EDTA-TBS: 10 mM Tris–HCl, pH 7.6; 150 mM NaCl; and 20 mM EDTA.

Purification of CGL1

a

623

b

c

13

d

6

12

Domain B

5 3

8

11 10 14 9

4 7 2

1

Domain A

e

Domain B Domain A

Domain A Domain B

Fig. 1 Crystal structure of CGL1. Two types of crystal structures are shown in ribbon mode. (a) One mannose molecule per protomer (mannose in site A). (b) Two mannose molecules per protomer (mannoses in site A and B). Each protomer is shown in a different color. Mannose molecules that are bound to CGL1 are shown as orange stick figures. Each protomer of the dimer structure is related by twofold crystallographic symmetry. Side (c) and top (d) views of CGL1 domains are shown as a ribbon diagram. Mannose molecules bound to sites A and B are shown as red and blue stick figures, respectively. The numbering is an order from N-terminal. (e) Superposed structures of domains A and B, which are related according to pseudo-twofold symmetry

624

Hideaki Unno and Tomomitsu Hatakeyama

Ala127 Phe126

Mannose

4

5

Ala127 Phe126

Mannose

4

5

1

1 3 Asp22

2

His52

Glu59

2

Glu59

Lys43

3

Asp22

His52

Lys43

Fig. 2 Illustration of the interactions between mannose and residues at sites A of CGL1 (stereo view). Mannose and residues are shown as stick figures. The dotted lines denote hydrogen bonds 2.1.2 Equipment

1. Suction-filtration system: The system is typically composed of a glass filter, Bu¨chner flask, silicon tube, and an aspirator. 2. Mixing equipment: magnetic stirrer and stirring bar, laboratory rocker.

2.2 Purification of CGL1

1. Oysters (C. gigas) purchased from a local dealer. Remove the shells, and store the bodies at 20  C until use.

2.2.1 Reagents/ Solutions/Materials

2. Cleaning tissue, commonly used in laboratories. 3. Strainer, used as a kitchen tool. 4. Mannose-immobilized resin (method for preparation, as described in Subheading 3.1) and stored at 4  C. 5. Tris-buffered saline (TBS, 10): 1.5 M NaCl, 0.1 M Tris–HCl at pH 7.6. 6. TBS: 10 mM Tris–HCl, pH 7.6; 150 mM NaCl. Stored at 4  C. 7. 100 mM mannose–TBS: 1.8 g of mannose, 100 mL of TBS. Stored at 4  C. 8. Open column system: open column (10 mL size), silicon tubes, tubing clamps. 9. 1 and 2 L beakers. 10. Dialysis tube.

2.2.2 Equipment

1. Blender, used as a kitchen appliance. 2. Suction–filtration system (mentioned in Subheading 2.1). 3. UV absorptiometer.

Purification of CGL1

625

4. Sample concentrators, molecular weight cutoff (MWCO): 10 K. 5. Mixing equipment: stirrers and stirring bars. 6. Size-exclusion chromatography (SEC) system: HiLoad 26/60 ¨ KTASuperdex 200 prep grade column (GE Healthcare) and A prime Plus apparatus (GE Healthcare).

3

Methods

3.1 Preparation of the MannoseImmobilized Resin

1. Wash 20 mL of the cellulose resin three times with the carbonate buffer. Mix the resin with 20 mL of 10% (v/v) vinyl sulfone in the carbonate buffer (see Note 1). Stir the resin using a stirring bar for 1 h at room temperature (see Note 2). 2. Wash the resin three times with each, deionized water and then the carbonate buffer. 3. Covalent reaction of divinyl sulfone and carbohydrates: Mix 2 g of mannose with 20 mL of the carbonate buffer, and add this solution to the washed resin. Keep the suspension overnight at room temperature with stirring (see Note 3). 4. Wash the resin three times with deionized water and the carbonate buffer, each, followed by a wash with EDTA-TBS. Transfer the washed resin into 50 mL tube and add 10 mL of EDTA-TBS to the tube. The resin is stored at 4  C (see Note 4).

3.2 Purification of CGL1

1. Wash 5 mL of the mannose-immobilized resin three times with cold TBS using the suction–filtration system. 2. Crush the defrosted C. gigas (500 g) in a 500 mL cold TBS less (see Note 5), and remove (9500  g for 30 min at 4  C) tissue (see Note 6).

(the Pacific oyster) samples using a blender for 1 min or the debris by centrifugation and filtration with a cleaning

3. Add the mannose-immobilized resin (5 mL) to the supernatant and stir mildly for 1 h at 4  C. 4. Remove the resin by centrifugation (4000  g for 5 min at 4  C). 5. Add 200 mL of cold TBS to the CGL1-bound resin, and wash the resin three times with cold TBS using the suction–filtration system (see Note 7). 6. Pour the resin suspension in TBS into an open column and wash with cold TBS. The fractions are collected in test tubes (10 mL/tube) and their absorbance at 280 nm is checked.

626

Hideaki Unno and Tomomitsu Hatakeyama

a

b 100 mM mannose-TBS

ABS at 280 nm

3

2

1

0

0

R kDa

M

C

NR L

C

L

66 45 29

12.5

2

4

6

8

10

Fraction number (5 mL)

Fig. 3 Purification of CGL1. (a) TBS extract from the Pacific oyster (C. gigas) was applied to a mannoseconjugated cellulose column (1.4  3.5 cm) equilibrated with TBS. CGL1 that was bound to the column was eluted with TBS containing 100 mM mannose (arrow). ABS absorbance. (b) An SDS-PAGE pattern under reducing (R) and nonreducing (NR) conditions. Numbers on the left indicate the molecular weights of marker proteins as follows: BSA (66 kDa), ovalbumin (45 kDa), carbonic anhydrase (29 kDa), and cytochrome C (12.5 kDa). M molecular marker, C crude extract, L lectin

7. Elute CGL1 with 100 mM mannose-TBS (Fig. 3a). 8. After concentrating the eluted protein solution using a sample concentrator (MWCO: 10 K), apply the protein solution to the Superdex 200 column equilibrated with TBS (see Note 8). 9. Dialyze the eluted CGL1 against TBS. 10. Check the purity of the protein by SDS-PAGE (Fig. 3b) (see Note 9).

4

Notes 1. Divinyl sulfone is a toxic reagent. Protective equipment must be worn during its handling. 2. Use of a small stirring bar with low rotative speed is recommended. High-magnetic force and high-speed rotation would lead to grinding of the resin due to the stirring bar. 3. Mild agitation is recommended for preventing grinding of the resin as well as the effects in Note 2. We used a 50 mL tube and laboratory rocker as the container and mixing machine, respectively, at this step. Minimal agitation would be adequate for the mixing step and the precipitation of resin at a bottom of the container would be prevented.

Purification of CGL1

627

4. The sugar-immobilized resin should be stored in EDTA-TBS at 4  C to prevent degradation of the carbohydrates by microorganisms. The mannose-immobilized resin can be reused for the purification of CGL1. However, the period of reuse should be limited to 1 year because the divinyl sulfone cross-links would get hydrolyzed under normal conditions. 5. We usually use zipper storage bags as containers for the frozen oyster samples. During the unfreezing step, storage bags containing the oyster are immersed in water for thawing. A fluid derived from the oyster is found to accumulate at the bottom of the bag after thawing. During the transfer of thawed oysters into the blender, care is taken to transfer the fluid in it as well, since it is anticipated that the fluid might also contain some amounts of CGL1. 6. During extraction of the supernatant after centrifugation, care is taken to prevent contamination with the precipitation components that might block the mannose-immobilized resin column. For preventing the contamination, a simple filtration step using a handmade filtration tool is carried out. The filtration tool is made easily using a beaker, strainer used as a kitchen tool, and 2-ply cleaning tissues commonly used in laboratories. Gravitational filtration of the supernatant is conducted using the tool to eliminate solid contaminants. 7. At the washing step using suction–filtration, it is recommended to release the decompression early to prevent drying of the CGL1-bound resin, since drying of the resin would result in denaturation of the proteins. 8. Approximately 10 mg of the purified CGL1 is obtained from 500 g of the Pacific oyster. Affinity chromatography using the mannose-immobilized resin can be used to successfully obtain CGL1 with sufficient purity (over 95%) for general use (Fig. 3). However, further purification is achieved using size-exclusion chromatography (SEC), in which high-quality crystals of CGL1 are obtained. Depending on its application, the SEC step can be abbreviated. 9. The molecular mass of CGL1 is 15.5 kDa/monomer, and the estimated absorption at 280 nm at a concentration of 1 mg/ mL is 1.48. References 1. Allam B, Raftos D (2015) Immune responses to infectious diseases in bivalves. J Invertebr Pathol 131:121–136 2. Tanaka H, Chiba H, Inokoshi J et al (2009) Mechanism by which the lectin actinohivin

blocks HIV infection of target cells. Proc Natl Acad Sci U S A 106:15633–15638 3. Bewley CA, Gustafson KR, Boyd MR et al (1998) Solution structure of cyanovirin-N, a potent HIV-inactivating protein. Nat Struct Biol 5:571–578

628

Hideaki Unno and Tomomitsu Hatakeyama

4. Koharudin LMI, Gronenborn AM (2014) Antiviral lectins as potential HIV microbicides. Curr Opin Virol 7:95–100 5. Unno H, Matsuyama K, Tsuji Y et al (2016) Identification, characterization, and X-ray crystallographic analysis of a novel type of mannosespecific lectin CGL1 from the Pacific oyster Crassostrea gigas. Sci Rep 6:29135

6. Magalha˜es GS, Lopes-Ferreira M, Junqueira-deAzevedo ILM et al (2005) Natterins, a new class of proteins with kininogenase activity characterized from Thalassophryne nattereri fish venom. Biochimie 87:687–699 7. Porath J (1974) General methods and coupling procedures. Methods Enzymol 34:13–30

Chapter 54 Receptor-Binding Assays of Enterovirus D68 Tadatsugu Imamura, Michiko Okamoto, and Hitoshi Oshitani Abstract Human enterovirus D68 (EV-D68) is a causative agent for acute respiratory infections and potentially central nervous system illnesses with increasing epidemiological significance. Recent studies have highlighted the role of sialic acids as a functional receptor for EV-D68 in vitro. However, further investigations are required to reveal its significance in actual infections in human. Key words Enterovirus D68, Picornaviruses, Sialic acids, Respiratory tract infections, Acute flaccid paralysis

1

Introduction Human enterovirus D68 (EV-D68) is a member of species enterovirus D, the genus Enterovirus, and the family of Picornaviridae. EV-D68 was first isolated from four pediatric patients hospitalized with lower respiratory tract infections in CA, United States, in 1962 [1]. Respiratory samples collected from those four patients were inoculated onto primary rhesus monkey kidney (MK) cells, leading to the isolation of four strains with typical enterovirus-like cytopathic effect: Fermon, Franklin, Robinson, and Rhyne. Those four isolates were tested for serological identification by using neutralization tests against immune sera specific for panels of viruses [1]. The test revealed that the isolates were antigenically distinct from any of the known enteroviruses. Therefore, the four isolates were proposed to be members of a new serotype of the Picornaviridae family. The Fermon strain was selected as the representative strain of this new serotype, considering the identical antigenic properties among those four strains [1]. After the initial identification in 1962, detection of this virus was only rarely reported until the early 2000s [2]. However, increased detection of EV-D68 from patients with acute respiratory infections was reported from a number of similar reports from different parts of the world [3, 4]. EV-D68 strains detected in

Jun Hirabayashi (ed.), Lectin Purification and Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 2132, https://doi.org/10.1007/978-1-0716-0430-4_54, © Springer Science+Business Media, LLC, part of Springer Nature 2020

629

630

Tadatsugu Imamura et al.

Fig. 1 Genome structure of EV-D68. Numbers below the bars indicate the location of each genome regions in the nucleotide sequences of the Fermon strain

recent years are comprised of three genetic groups, clades A–C, and emergence of genetically divergent viruses is regarded as one of the possible underlying mechanisms for such increased detections (Fig. 1) [4]. Disease severity ranged from mild upper respiratory infections, in most of the cases, to severe lower respiratory infections, including a considerable number of fatal cases [4]. Notably, more than 1000 cases of acute respiratory infections were reported in the United States in 2014 [5, 6]. And during this nationwide outbreak, a surge of acute flaccid myelitis (AFM) was observed with more than 100 reported cases in 2014 alone [7]. Based on the strong temporal and geographical association between the surge of AFM and EV-D68 outbreak, it was suggested that EV-D68 might be the causative agent for the central nervous system illnesses [8]. However, EV-D68 was detected from only a limited number of those AFM patients, mostly in respiratory samples, not cerebrospinal fluid or serum [8, 9]. Therefore, the causal relationships between EV-D68 infection and AFM are still not well established. Receptor-binding properties of EV-D68 have been extensively studies in recent years. In 1962, it was reported that EV-D68 strongly agglutinates guinea pig erythrocytes [1]. In 1991, sialidase treatment of HeLa cell monolayers was shown to cause an approximately 90% reduction of EV-D68 attachment and replication,

Protocols for Sialic Acid Binding Assays of Enterovirus D68

631

which suggested that sialic acids (SAs) were possible receptors for EV-D68 [10]. We further investigated the SA-binding specificities of EV-D68 by using glycan array analysis as well as enzymatically modified erythrocytes and revealed that EV-D68 had stronger affinity with α2,6 linked SAs (Neu5Acα2,6Gal) than α2,3 linked SAs (Neu5Acα2,3Gal) [11]. It is known that α2,6 linked SAs are dominantly expressed in the upper respiratory tract while α2,3 linked SAs are in the lower respiratory tract of the human airway [12]. These results suggest that EV-D68 may have an affinity for the upper respiratory tract compared with the lower respiratory tract. Additional studies using knockout cells lacking SAs on the cell surface have demonstrated that α2,6 linked SAs were associated with efficient growth of EV-D68, which suggested that α2,6 linked SAs is the functional receptor for EV-D68 [13]. In the study, EV-D68 also propagated on cell cultures without α2,6 linked SAs as well as those only with α2,3 linked SAs on their surface [13]. Those results suggested that α2,3 linked SAs might take part in the propagation of EV-D68 on cell cultures; however, the significance of α2,3 linked SAs in the actual infections in human still remains unknown. The receptor-binding sites and the conformational changes of EV-D68 following receptor bindings have been intensively studied in recent years. Based on these studies, it is expected that sialic acid receptors bind to the depression part of the virion surface called canyon, which causes ejection of pocket factor and induces destabilization of the virus which eventually leads to release of the virus genome [14, 15]. We previously investigated on the sialidase activity of EV-D68, using the 20 -(4-methylumbelliferyl)-D-N-acetylneuraminic acid sodium salt hydrate [11]. As a result, sialidase activity was not detected for any of the tested EV-D68 strains, including the Fermon strain and the strains isolated in Japan in 2010. It is widely known that influenza viruses have sialidase activity by which the newly synthesized viral proteins are removed from SAs and the virus progenies are released from the host cell surface [16]. The reasons for the lack of sialidase activity in EV-D68 strains are still not fully understood. It was previously shown that, in contrast to enveloped viruses such as influenza viruses, nonenveloped viruses such as EV-D68 are released from the host cells by using mechanisms called cell lysis [17]. The underlying mechanisms for the absence of sialidase activity in EV-D68 strains might reside in those differences in the virus release mechanisms. Recent studies have shown that EV-D68 might potentially cause central nervous system illnesses. However, the entry mechanisms of EV-D68 into neuron cells remained unknown until recently. In 2016, it was reported that neuron-specific intercellular adhesion molecule (ICAM-5) was associated with infection and replication of EV-D68 in various types of cells, which suggested that ICAM-5 might be one of the cellular receptors of EV-D68

632

Tadatsugu Imamura et al.

[18]. Since ICAM-5 is extensively expressed on the telencephalon region of the human brain, ICAM-5 is suspected as a receptor that is associated with the neurological pathogenesis of the virus [18]. Despite the recent years of progress on the underlying mechanisms of neurological illnesses associated with EV-D68 infections, the roles of ICAM-5 in EV-D68 infections were still only analyzed on the cell culture models. Therefore, further studies using animal models are expected.

2 2.1

Materials Regents

1. Cell culture medium: Eagle’s Minimum Essential Medium (E-MEM) containing 2% of calf serum and 1.7% of glucose. 2. 8% (w/v) polyethylene glycol (PEG). 3. 40% (w/v) sucrose. 4. Incomplete Freund’s adjuvant. 5. N-(3-Maleimidobenzoyloxy)succinimide. 6. N,N-Dimethylformamide. 7. Phosphate buffered salts (PBS). 8. 10% (v/v) formalin. 9. Bovine serum albumin. 10. Arthrobacter ureafaciens sialidase. 11. α2-3/α2-6 sialyltransferase. 12. Cytidine-50 -monophospho-N-acetylneuraminic sodium salt.

acid

2.2

Antibodies

1. Anti-rabbit IgG sera conjugated with Cy3.

2.3

Cells

1. Monolayer cell culture of rhabdomyosarcoma cells (RD-18S). 2. Guinea pig erythrocytes (GPE).

2.4 Experimental Animals

1. Specific pathogen-free female Hartley guinea pigs (6 weeks of age).

2.5

1. Centrifuge.

Equipment

2. CO2 incubator. 3. Ultracentrifuge. 4. Microarray-grade epoxy-coated glass slide. 5. Bio-REX SCAN 200.

Protocols for Sialic Acid Binding Assays of Enterovirus D68

3

633

Methods

3.1 Clinical Sample Handling

1. Centrifuge the clinical samples (e.g., nasopharyngeal swabs in E-MEM containing 2% of calf serum and 1.7% of glucose, nasal aspirates, or cerebrospinal fluid) at 1700  g for 15 min at 4  C. 2. Use the supernatants immediately for further virus isolation, or store them at 80  C until use.

3.2

Virus Isolation

1. Inoculate the clinical samples on monolayer cell culture (see Note 1) in E-MEM containing 2% of calf serum and 1.7% of glucose. 2. Centrifuge the plate at 447  g for 30 min at room temperature. 3. Incubate the inoculated cells at 34  C in a 5% CO2 incubator. 4. After the incubation, harvest cultured fluid. Successful growth of EV-D68 can be confirmed by the appearance of cytopathic effect. The duration before harvest depends on virus titers in the clinical samples, and it may occasionally require incubation longer than a week. 5. Store the harvested fluid at 80  C until use.

3.3 Preparation of Polyclonal Antibody

1. Prepare virus-cultured fluids as described in Subheading 3.2 (see Note 2). 2. Purify the culture fluids by precipitation with 8% PEG. 3. Centrifuged the cultured fluids after PEG precipitation on a 40% sucrose gradient at 274,400  g for 3 h. 4. Inactivate the virus, by mixing 1.0 mL of purified cultured fluids with 10 μL of 10% formalin, and incubate them at room temperature for 2 h. 5. Mix 1.0 mL of formalin-inactivated virus with1.0 mL of incomplete Freund’s adjuvant. 6. Inject 1.5 mL of formalin-inactivated virus mixed with adjuvant to specific pathogen-free female Hartley guinea pigs (6 weeks of age) intraperitoneally (see Note 3). 7. Inject the guinea pigs with the mixture of antigens and adjuvant for three times in total with intervals of 1 and 2 weeks sequentially. 8. Collect the serum 1 week after the third injection if adequate elevation of antibody titer is confirmed. 9. Process the collected serum by heat inactivation at 56  C for 30 min. 10. Store the serum at 80  C until use.

634

Tadatsugu Imamura et al.

3.4 Preparation of Slides for Glycan Array Analysis

1. Prepare oligosaccharides from various biomaterials as described previously. 2. After the catalytic reduction, mix 2.5 μmol of each oligosaccharide with 160 μmol of N-(3-maleimidobenzoyloxy)succinimide dissolved in N,N-dimethylformamide. 3. Incubate at 30  C for 30 min. 4. After chloroform extraction, mix the products with 100 nmol of reduced bovine serum albumin. 5. Incubated at room temperature for 2 h. 6. Purify the products by dialysis with PBS. 7. Add 5 ng of synthesized glycoproteins on a microarray-grade epoxy-coated glass slide.

3.5 Modification of Erythrocytes

1. Incubate 100 μL of 10% GPE with 5 mU of A. ureafaciens sialidase at 37  C for 3 h (see Note 4). 2. Confirm the removal of sialic acids by complete loss of hemagglutination (HA) titer for the control viruses (see Note 5). 3. After the sialidase treatment, wash GPE with PBS twice. 4. Incubate 100 μL of 10% sialidase treated-GPE with 1.25~10 mU of α2-3 or α2-6 sialyltransferase and 1.0 mM of cytidine-50 -monophospho-N-acetylneuraminic acid sodium salt at 37  C for 4 h. 5. Confirm resialylation by complete recovery of HA titer for the control viruses (see Note 5).

3.6 Glycan Array Analysis

1. Add 5 ng of synthesized glycoproteins on a microarray-grade epoxy-coated glass slide. 2. Dilute isolated viruses to two hemagglutination units (HAU) in binding buffer, and add 80 μL of each diluent to the glass chambers. 3. Incubate the glass chambers at room temperature for 1 h being gently shaken. 4. After the incubation, remove the diluent. 5. Add 80 μL of polyclonal antibody for EV-D68 strains to the chambers. 6. Incubate the chambers at room temperature for 1 h being gently shaken. 7. After the incubation, remove the polyclonal antibodies. 8. Add 80 μL of anti-rabbit IgG sera conjugated with Cy3, which was diluted to 1:1000 with binding buffer, to each chamber. 9. Incubate the chambers at room temperature for 1 h being gently shaken.

Protocols for Sialic Acid Binding Assays of Enterovirus D68

635

10. After the incubation, measure the fluorescence of the glass chambers. In the previous study, we used Bio-REX SCAN 200 [11]. An example of results is shown in Fig. 2. In this analysis, binding capacities of EV-D68 strains isolated in Japan and the Philippines to glycol-oligosaccharides with α2-3 or α2-6 sialic acid terminals were tested. The result indicates that the strains had higher affinity to glycol-oligosaccharides with α2-6 sialic acid terminals compared to those with α2-3 sialic acid terminals [11]. 3.7 Hemagglutination Test Using Modified Erythrocytes

1. Prepare a V-shaped bottom plate containing 50 μL of virus twofold serially diluted in PBS (see Note 4). 2. Add 50 μL of 0.75% guinea pig erythrocyte suspension to each well of the V-shaped bottom plate. 3. After mixing, incubate the plate at 4  C for 2 h. 4. Determine the HAU of the virus as the reciprocal value of the last dilution exhibiting complete hemagglutination. An example of results is shown in Fig. 3. In this analysis, HAU of EV-D68 strains isolated in Japan and the Philippines were measured. The result indicates that the strains had higher HAU for erythrocytes treated with α2-6 sialyltransferase; however, only limited HAU for those treated with α2-3 sialyltransferase [11].

4

Notes 1. Several types of cells have been shown to be susceptible for EV-D68 [10, 11, 19]. Types of cell lines that were used for isolation or propagation of EV-D68 in previous studies are listed in Table 1. We previously used rhabdomyosarcoma (RD) cells 18S (RD18S) and A (RD-A) for virus isolation, propagation, and neutralization tests. In our studies, some of the isolated viruses grew efficiently on RD-18S but less on RD-A (data not shown). If the virus does not grow efficiently on the cell line that you selected, using other types of cell lines should be considered. 2. The optimal virus titer of EV-D68 cultured fluids for producing polyclonal antibodies in experimental animals has not been documented. In the previous study, we used 500 ml cultured fluids with virus titer of 106 TCID50/mL, which was eventually adequate for producing polyclonal antibodies in guinea pigs [11]. 3. In addition to guinea pigs, we previously used specificpathogen-free female New Zealand White rabbit (12 weeks of age) to produce polyclonal antibodies following the protocol described in Subheading 3.3 by which efficient production of polyclonal antibodies was confirmed [11].

636

Tadatsugu Imamura et al.

Fig. 2 (a) Sialyloligosaccharides used in the study. Six types of sialyloligosaccharides were used in the study. (b) Receptor-binding specificities of EV-D68. The receptor-binding specificities of the prototype strain Fermon, and four strains detected in recent years, including two strains from Japan and two strains from the Philippines, were tested

Protocols for Sialic Acid Binding Assays of Enterovirus D68

637

Fig. 3 Hemagglutination activity of EV-D68 to modified erythrocytes. Hemagglutination activities of three strains of EV-D68: the prototype strain Fermon, a strain isolated from Japan, and a strain from the Philippines are shown. Influenza A(H1N1) 2009 pandemic (H1N1pdm) and human parainfluenza virus 1 (PIV1) were used as control viruses. Bars with HAU no higher than 1 was indicated with “