PDZ Mediated Interactions: Methods and Protocols (Methods in Molecular Biology, 2256) 1071611658, 9781071611654

Citation preview

Methods in Molecular Biology 2256

Jean-Paul Borg Editor

PDZ Mediated Interactions 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-by step 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.

PDZ Mediated Interactions Methods and Protocols

Edited by

Jean-Paul Borg Centre de Recherche en Cancérologie de Marseille, Aix-Marseille University, Inserm, CNRS, Institut Paoli-Calmettes, Marseille, France; Institut Universitaire de France (IUF), Paris, France

Editor Jean-Paul Borg Centre de Recherche en Cance´rologie de Marseille Aix-Marseille University, Inserm, CNRS, Institut Paoli-Calmettes Marseille, France Institut Universitaire de France (IUF) Paris, France

ISSN 1064-3745 ISSN 1940-6029 (electronic) Methods in Molecular Biology ISBN 978-1-0716-1165-4 ISBN 978-1-0716-1166-1 (eBook) https://doi.org/10.1007/978-1-0716-1166-1 © Springer Science+Business Media, LLC, part of Springer Nature 2021 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. 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 All biological functions are regulated by protein networks whose organization relies on finely tuned protein-protein and protein-lipid interactions. PDZ domains represent one of the most widely distributed protein-protein interaction domains and contribute to a large number of biological processes, from the plasma membrane to the nucleus, especially in cellcell communication and cell polarity. Their importance in physiology and pathologies such as cancer, neurodegenerative and infectious diseases being now well established since their discovery in the 1990s, they have brought the interest of many laboratories, which has led to the development of dedicated techniques able to predict and identify their ligands, characterize their functions in normal and pathological conditions, and, more recently, conduct the design of peptide or chemical inhibitors. This volume provides a comprehensive overview of the techniques currently applied to identify and characterize PDZ-mediated interactions and opens the discussion on priority topics emerging in this area of investigation (promiscuity, multimodularity, regulation, and viral recognition by PDZ domains). Marseille, France

Jean-Paul Borg

v

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

v ix

1 Identification of PDZ Interactions by Yeast Two-Hybrid Technique . . . . . . . . . . Monica Castro-Cruz, Marta Monserrat-Gomez, Jean-Paul Borg, Pascale Zimmermann, and Eric Bailly 2 Identification of PDZ Interactions by Affinity Purification and Mass Spectrometry Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Avais M. Daulat, Ste´phane Audebert, Moˆnica Wagner, Luc Camoin, and Jean-Paul Borg 3 Identification of PDZ Interactions by Proteomic Peptide Phage Display . . . . . . . ¨ chow, Gustav N. Sundell, and Ylva Ivarsson Susanne Lu 4 A Computational Protocol to Analyze PDZ/PBM Affinity Data Obtained by High-Throughput Holdup Assay. . . . . . . . . . . . . . . . . . . . . . . . . Pau Jane´, Lionel Chiron, Goran Bich, Gilles Trave´, and Yves Nomine´ 5 Study of PDZ–Peptide and PDZ–Lipid Interactions by Surface Plasmon Resonance/BIAcore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pascale Zimmermann and Antonio Luis Egea-Jimenez 6 PDZ Sample Quality Assessment by Biochemical and Biophysical Characterizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ce´lia Caillet-Saguy, Se´bastien Bruˆle´, Nicolas Wolff, and Bertrand Raynal 7 Crystallographic Studies of PDZ Domain–Peptide Interactions of the Scribble Polarity Module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Janesha C. Maddumage, Bryce Z. Stewart, Patrick O. Humbert, and Marc Kvansakul 8 A Fluorescence-Based Assay to Determine PDZ–Ligand Binding Thermodynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Young Joo Sun and Ernesto J. Fuentes 9 Unveiling the Folding Mechanism of PDZ Domains . . . . . . . . . . . . . . . . . . . . . . . . Candice Gautier and Stefano Gianni 10 Development of Peptide-Based PDZ Domain Inhibitors. . . . . . . . . . . . . . . . . . . . . Dominik J. Essig, Javier R. Balboa, and Kristian Strømgaard 11 Dynamic Control of Signaling by Phosphorylation of PDZ Binding Motifs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ma´rton A. Simon and La´szlo Nyitray 12 Chemical Synthesis of PDZ Domains. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Christin Kossmann, Sana Ma, Louise S. Clemmensen, and Kristian Strømgaard

1

vii

17

41

61

75

89

125

137 149 157

179 193

viii

13

14

15 16

Contents

Viral PDZ Binding Motifs Influence Cell Behavior Through the Interaction with Cellular Proteins Containing PDZ Domains . . . . . . . . . . . . . ˜ o-Rodriguez, Jose M. Honrubia, Carlos Castan ´ lvarez, Isabel Sola, and Luis Enjuanes Javier Gutie´rrez-A Computational Design of PDZ-Peptide Binding. . . . . . . . . . . . . . . . . . . . . . . . . . . . Nicolas Panel, Francesco Villa, Vaitea Opuu, David Mignon, and Thomas Simonson Mechanoregulation of PDZ Proteins, An Emerging Function . . . . . . . . . . . . . . . . Elsa Bazellie`res and Andre´ Le Bivic Rational Design of PDZ Domain Inhibitors: Discovery of Small Organic Compounds Targeting PDZ Domains . . . . . . . . . . . . . . . . . . . . . Laurent Hoffer, Philippe Roche, and Xavier Morelli

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

217

237

257

277 291

Contributors STE´PHANE AUDEBERT • Aix Marseille Univ, CNRS, INSERM, Institut Paoli-Calmettes, CRCM, Marseille Prote´omique, Marseille, France ERIC BAILLY • Centre de Recherche en Cance´rologie de Marseille (CRCM), Aix-Marseille Universite´, Inserm, CNRS, Institut Paoli-Calmettes, Marseille, France JAVIER R. BALBOA • Department of Drug Design and Pharmacology, Center for Biopharmaceuticals, University of Copenhagen, Copenhagen, Denmark; Novo Nordisk A/ S, Research Chemistry 3, Ma˚løv, Denmark ELSA BAZELLIE`RES • Aix Marseille Universite´, CNRS, IBDM–UMR7288, Turing Centre for Living Systems, Marseille, France GORAN BICH • (Equipe labelise´e Ligue, 2015) Institut de Ge´ne´tique et de Biologie Mole´culaire et Cellulaire (IGBMC), INSERM U1258/CNRS UMR 7104/Universite´ de Strasbourg, Illkirch, France JEAN-PAUL BORG • Centre de Recherche en Cance´rologie de Marseille, Aix-Marseille University, Inserm, CNRS, Institut Paoli-Calmettes, Marseille, France; Institut Universitaire de France (IUF), Paris, France SE´BASTIEN BRUˆLE´ • Institut Pasteur, Plate-forme de Biophysique Mole´culaire, CNRS UMR 3528, Paris, France CE´LIA CAILLET-SAGUY • Institut Pasteur, Unite´ Re´cepteurs-Canaux, CNRS UMR 3571, Paris, France LUC CAMOIN • Aix Marseille Univ, CNRS, INSERM, Institut Paoli-Calmettes, CRCM, Marseille Prote´omique, Marseille, France CARLOS CASTAN˜O-RODRIGUEZ • Department of Molecular and Cell Biology, Centro Nacional de Biotecnologı´a (CNB-CSIC), Madrid, Spain MONICA CASTRO-CRUZ • Centre de Recherche en Cancerologie de Marseille (CRCM), Equipe Zimmermann labellise´e Ligue 2018 – 2019, Aix-Marseille Universite´, Inserm, CNRS, Institut Paoli-Calmettes, Marseille, France; Department of Human Genetics, K. U. Leuven, Leuven, Belgium LIONEL CHIRON • CASC4DE, Strasbourg, France LOUISE S. CLEMMENSEN • Department of Drug Design and Pharmacology, Center for Biopharmaceuticals, University of Copenhagen, Copenhagen, Denmark AVAIS M. DAULAT • Aix Marseille Univ, CNRS, INSERM, Institut Paoli-Calmettes, CRCM, Equipe labellise´e Ligue ‘Cell polarity, cell signaling and cancer’, Marseille, France ANTONIO LUIS EGEA-JIMENEZ • Centre de Recherche en Cance´rologie de Marseille (CRCM), Equipe Zimmermann labellise´e Ligue 2018, Aix-Marseille Universite´, Inserm, CNRS and Institut Paoli-Calmettes, Marseille, France LUIS ENJUANES • Department of Molecular and Cell Biology, Centro Nacional de Biotecnologı´a (CNB-CSIC), Madrid, Spain DOMINIK J. ESSIG • Department of Drug Design and Pharmacology, Center for Biopharmaceuticals, University of Copenhagen, Copenhagen, Denmark; Novo Nordisk A/ S, Research Chemistry 3, Ma˚løv, Denmark ERNESTO J. FUENTES • Department of Biochemistry, University of Iowa, Iowa City, IA, USA; Holden Comprehensive Cancer Center, University of Iowa, Iowa City, IA, USA

ix

x

Contributors

CANDICE GAUTIER • Istituto Pasteur-Fondazione Cenci Bolognetti and Istituto di Biologia e Patologia Molecolari del CNR, Dipartimento di Scienze Biochimiche “A. Rossi Fanelli”, ` di Roma, Rome, Italy Sapienza Universita STEFANO GIANNI • Istituto Pasteur-Fondazione Cenci Bolognetti and Istituto di Biologia e Patologia Molecolari del CNR, Dipartimento di Scienze Biochimiche “A. Rossi Fanelli”, ` di Roma, Rome, Italy Sapienza Universita ´ ´ JAVIER GUTIERREZ-ALVAREZ • Department of Molecular and Cell Biology, Centro Nacional de Biotecnologı´a (CNB-CSIC), Madrid, Spain LAURENT HOFFER • Centre de Recherche en Cance´rologie de Marseille (CRCM), AixMarseille Universite´, Inserm, CNRS and Institut Paoli-Calmettes, Marseille, France JOSE M. HONRUBIA • Department of Molecular and Cell Biology, Centro Nacional de Biotecnologı´a (CNB-CSIC), Madrid, Spain PATRICK O. HUMBERT • Department of Biochemistry and Genetics, La Trobe University, Bundoora, VIC, Australia; La Trobe Institute for Molecular Science, La Trobe University, Bundoora, VIC, Australia YLVA IVARSSON • Department of Chemistry, BMC, Uppsala University, Uppsala, Sweden PAU JANE´ • (Equipe labelise´e Ligue, 2015) Institut de Ge´ne´tique et de Biologie Mole´culaire et Cellulaire (IGBMC), INSERM U1258/CNRS UMR 7104/Universite´ de Strasbourg, Illkirch, France CHRISTIN KOSSMANN • Department of Drug Design and Pharmacology, Center for Biopharmaceuticals, University of Copenhagen, Copenhagen, Denmark MARC KVANSAKUL • Department of Biochemistry and Genetics, La Trobe University, Bundoora, VIC, Australia; La Trobe Institute for Molecular Science, La Trobe University, Bundoora, VIC, Australia ANDRE´ LE BIVIC • Aix Marseille Universite´, CNRS, IBDM–UMR7288, Turing Centre for Living Systems, Marseille, France SUSANNE LU¨CHOW • Department of Chemistry, BMC, Uppsala University, Uppsala, Sweden SANA MA • Department of Drug Design and Pharmacology, Center for Biopharmaceuticals, University of Copenhagen, Copenhagen, Denmark JANESHA C. MADDUMAGE • Department of Biochemistry and Genetics, La Trobe University, Bundoora, VIC, Australia; La Trobe Institute for Molecular Science, La Trobe University, Bundoora, VIC, Australia DAVID MIGNON • Laboratoire de Biologie Structurale de la Cellule (CNRS UMR7654), Ecole Polytechnique, Palaiseau, France MARTA MONSERRAT-GOMEZ • Centre de Recherche en Cance´rologie de Marseille (CRCM), JPB team is Equipe labellise´e Ligue 2018 – 2019, Aix-Marseille Universite´, Inserm, CNRS, Institut Paoli-Calmettes, Marseille, France XAVIER MORELLI • Centre de Recherche en Cance´rologie de Marseille (CRCM), Aix-Marseille Universite´, Inserm, CNRS and Institut Paoli-Calmettes, Marseille, France YVES NOMINE´ • (Equipe labelise´e Ligue, 2015) Institut de Ge´ne´tique et de Biologie Mole´ culaire et Cellulaire (IGBMC), INSERM U1258/CNRS UMR 7104/Universite´ de Strasbourg, Illkirch, France LA´SZLO´ NYITRAY • Department of Biochemistry, ELTE Eo¨tvo¨s Lora´nd University, Budapest, Hungary VAITEA OPUU • Laboratoire de Biologie Structurale de la Cellule (CNRS UMR7654), Ecole Polytechnique, Palaiseau, France NICOLAS PANEL • Laboratoire de Biologie Structurale de la Cellule (CNRS UMR7654), Ecole Polytechnique, Palaiseau, France

Contributors

xi

BERTRAND RAYNAL • Institut Pasteur, Plate-forme de Biophysique Mole´culaire, CNRS UMR 3528, Paris, France PHILIPPE ROCHE • Centre de Recherche en Cance´rologie de Marseille (CRCM), Aix-Marseille Universite´, Inserm, CNRS and Institut Paoli-Calmettes, Marseille, France MA´RTON A. SIMON • Department of Biochemistry, ELTE Eo¨tvo¨s Lora´nd University, Budapest, Hungary THOMAS SIMONSON • Laboratoire de Biologie Structurale de la Cellule (CNRS UMR7654), Ecole Polytechnique, Palaiseau, France ISABEL SOLA • Department of Molecular and Cell Biology, Centro Nacional de Biotecnologı´a (CNB-CSIC), Madrid, Spain BRYCE Z. STEWART • Department of Biochemistry and Genetics, La Trobe University, Bundoora, VIC, Australia; La Trobe Institute for Molecular Science, La Trobe University, Bundoora, VIC, Australia KRISTIAN STRØMGAARD • Department of Drug Design and Pharmacology, Center for Biopharmaceuticals, University of Copenhagen, Copenhagen, Denmark YOUNG JOO SUN • Department of Biochemistry, University of Iowa, Iowa City, IA, USA GUSTAV N. SUNDELL • Department of Chemistry, BMC, Uppsala University, Uppsala, Sweden; Shanghai Institute for Advanced Immunochemical Studies, ShanghaiTech University, Shanghai, China GILLES TRAVE´ • (Equipe labelise´e Ligue, 2015) Institut de Ge´ne´tique et de Biologie Mole´ culaire et Cellulaire (IGBMC), INSERM U1258/CNRS UMR 7104/Universite´ de Strasbourg, Illkirch, France FRANCESCO VILLA • Laboratoire de Biologie Structurale de la Cellule (CNRS UMR7654), Ecole Polytechnique, Palaiseau, France ˆ MONICA WAGNER • Aix Marseille Univ, CNRS, INSERM, Institut Paoli-Calmettes, CRCM, Marseille Prote´omique, Marseille, France NICOLAS WOLFF • Institut Pasteur, Unite´ Re´cepteurs-Canaux, CNRS UMR 3571, Paris, France PASCALE ZIMMERMANN • Centre de Recherche en Cancerologie de Marseille (CRCM), Equipe Zimmermann labellise´e Ligue 2018 – 2019, Aix-Marseille Universite´, Inserm, CNRS, Institut Paoli-Calmettes, Marseille, France; Department of Human Genetics, K. U. Leuven, Leuven, Belgium

Chapter 1 Identification of PDZ Interactions by Yeast Two-Hybrid Technique Monica Castro-Cruz, Marta Monserrat-Gomez, Jean-Paul Borg, Pascale Zimmermann, and Eric Bailly Abstract The yeast two-hybrid technique is a powerful method to detect direct protein–protein interactions. Due to its accessibility, speed, and versatility, this technique is easy to set up in any laboratory and suitable for small and large scale screenings. Here we describe the implementation of an array-based screening that allows for the probing of the entire human PDZ ORFeome (or hPDZome) by yeast two-hybrid technique. With this approach, one can rapidly identify the PDZ domains that are able to interact (up to KD in the high μmolar range) with any candidate protein among a panel of 266 individual clones, thereby comprehensively identifying its PDZ interactome. Key words Yeast two-hybrid, Two-hybrid array, Protein–protein interaction, GAL4, Yeast strain, PDZ interactions

1

Introduction The yeast two-hybrid (Y2H) method was developed in the late 1980s and remains a straightforward genetic approach to detect direct protein–protein interactions (PPIs) [1, 2]. It takes advantage of the modular nature of transcription factors (TFs) like GAL4 to render the expression of reporter genes dependent on the binding properties of two partner proteins (Fig. 1). GAL4 comprises two functionally independent domains, a DNA binding domain (GAL4-BD) and a transcription activation domain (GAL4-AD) that can be separately fused to the proteins of interest. If the resulting hybrid components can interact through the two proteins of interest, then a functional GAL4 is reconstituted allowing for the transcription of the reporter/survival genes. In this setting, GAL4BD is commonly appended to one partner that serves as “bait” (X),

Monica Castro-Cruz and Marta Monserrat-Gomez contributed equally to this work. Jean-Paul Borg (ed.), PDZ Mediated Interactions: Methods and Protocols, Methods in Molecular Biology, vol. 2256, https://doi.org/10.1007/978-1-0716-1166-1_1, © Springer Science+Business Media, LLC, part of Springer Nature 2021

1

2

Monica Castro-Cruz et al.

Fig. 1 Principle/Overview of the yeast two-hybrid technique. (a) The protein of interest (X), corresponding to the bait is fused to the DNA binding domain of GAL4 (GAL4-BD-X), while the putative binder (Y) is fused to the activation domain (GAL4-AD-Y). (b) GAL4-BD-X binds the upstream activator sequence (UAS) of the promoter. Upon interaction between the bait and the prey, a functional GAL4 is reconstituted, thereby promoting the recruitment of RNA polymerase II and the transcription of the reporter gene

whereas Gal4-AD is fused to potential interactors called “prey” (Y) (Fig. 1). The bait and prey constructs are separately expressed in haploid yeast strains of opposite mating type (MATa and MATα). The auxotrophic markers HIS3 and URA3 together with the lacZ gene are the three most frequently used reporter/survival genes, the latter enabling a colorimetric readout of the tested PPI [3, 4]. Even though the Y2H approach is a powerful screening approach, it should be kept in mind that it is prone to false-positive and false-negative results [5]. False positives may result from spurious activation of the reporter gene in the absence of an interacting partner. False negatives, on the other hand, may stem from the requirement of posttranslational modifications, potential steric constrains imposed by the cloning or from the failure of the interacting partners to reach the nuclear compartment. It is always recommended to add negative and positive controls in addition to the proteins of interest. It is also crucial to cross-validate the results of a Y2H screen by alternative approaches (coimmunoprecipitation, mass spectrometry, and surface plasmon resonance, among others).

Y2H Screening of PDZ Interactions

3

Since its development, the Y2H approach has now been applied to an impressive variety of organisms and in various scale settings, from pairwise to genomic level screens [4]. In particular, Y2H data obtained with random cDNA libraries have opened the way to the establishment of highly elaborate interactome maps within the proteome of numerous species. However, in recent years the use of smaller and arrayed libraries has been increasingly exploited. These arrayed Y2H libraries can assemble proteins or domains belonging to the same structural family or to diverse subcellular entities. Their use can lead to a more comprehensive picture of the interactions mediated by the array members [6–9]. Moreover, a major advantage of arrays over random cDNA libraries when screening by Y2H is their suitability for high-throughput technologies, making their implementation faster, more accessible, and comprehensive. Here, we describe the implementation of a Y2H screen that focuses on PDZ-mediated interactions using an array-based human PDZ ORFeome (thereafter named PDZome) generated in our laboratory [10]. The PDZome holds 266 PDZ domains based on the prediction and manual annotation of all the PDZ sequences. New boundaries at the N- and C-termini have been established, contributing to the folding, solubility, and affinity of the PDZ domains [11]. As depicted in Fig. 2, candidate proteins known or suspected to contain PDZ-binding motifs (PBMs) are used as baits to screen the individual Y2H prey clones of the hPDZome array.

2

Materials All solutions should be prepared using ultrapure water and autoclaved or sterile filtered. Store all reagents at room temperature (unless indicated otherwise).

2.1 Design of Primers for Bait Constructs

The choice of a bait sequence to screen the PDZome library may be guided by preliminary evidence of PDZ binding properties of a candidate protein. When no information is available and because most PDZ domains interact with the C-terminal end of their protein targets, a standard approach is to fuse the last fifteen residues of a candidate protein in frame with the C-terminus of the Gal4-BD moiety. In this case, generating the same bait but truncated of its last three amino acids provides a convenient negative control as this truncation is known to disrupt most PBM/PDZ interactions [12]. Of note, the use of larger cytosolic fragments may also prove relevant when suspecting the occurrence of an internal noncanonical PBM within a given PDZ binding protein. Below are guidelines for the design of primers for the cloning of bait constructs using the Gateway® cloning system. This cloning strategy first requires the addition of attB1 and attB2

4

Monica Castro-Cruz et al.

Fig. 2 Construction of the hPDZome array for yeast two-hybrid screening. (a) The human PDZome was built using the Gateway ® system. The 266 entry clones in the pZeo vector were provided in 96-well plates by NZYTech. Each entry clone corresponds to one ORF of the hPDZome. (b) The 266 PDZ ORFs were subcloned into the pACT2-AD vector using Gateway ® LR clonase and verified by sequencing. (c) Once validated, each pACT2-AD PDZ prey was transformed into the Y187 prey yeast strain and transformants were selected by plating on synthetic complete agar plates without leucine (SC agar -Leu). The final array ready for mating consists of three 96-well plates with yeast clones expressing the hPDZome fused to the Gal4 activation domain (AD)

recombination sites flanking the target gene sequence. Introduction of these sites is easily achieved by PCR using oligonucleotides designed as follows. 1. Design a For-attB1-ORF forward primer based on the sequence 50 -GGGG ACA AGT TTG TAC AAA AAA GCA GGC TNN NNN-30 where the attB1 recombination site is underlined and N letters correspond to the 50 -end of your

Y2H Screening of PDZ Interactions

5

bait gene. The codon sequence shown here within the attB1 site is in frame with the coding sequence of the upstream GAL4BD region. It is therefore critical to make sure that the sequence of the bait is in frame with the indicated attB1 reading sequence as it is the case here with the TNN fusion codon. 2. Design a Rev-attB2-ORFwt reverse primer based on the sequence 50 -GGGG AC CAC TTT GTA CAA GAA AGC TGG GTC TTA NNN-30 where the attB2 recombination site is underlined and N letters correspond to the reverse and complementary 30 -end wild-type sequence of the bait construct. Pay attention to the fact that a stop codon has to be placed between the 30 -end of the target gene sequence and the attB2 site as additional amino acid residues at the C-terminus of the bait are likely to disrupt its PBM properties. In the above Rev-attB2-ORFwt sequence, a TAA stop codon (denoted in bold letters) has been introduced but the endogenous stop codon of the bait target can be used as well. 3. Following the same guidelines as described in item 2, design a second Rev-attB2-ORFmut reverse primer in which the sequence of the last three C-terminal codons of the bait is lacking. This primer will serve for the cloning of the PBM mutant bait construct. 2.2 PCR Amplification and Purification of the Bait Constructs

1. 5 mM dNTPs. 2. 10 μM forward and reverse primers. 3. Bait DNA to be used as a template for the PCR reaction. 4. High-fidelity DNA polymerase. 5. Concentrated stocks of reaction buffer and magnesium chloride provided with the DNA polymerase. 6. TAE buffer: 40 mM Tris, 40 mM acetate, 1 mM EDTA, pH 8.3. 7. 1% or 2% agarose gels prepared in TAE buffer. 8. Gel DNA extraction kit of your choice. Be aware that some kits may not be optimized for the recovery of small DNA fragments.

2.3 Cloning of the Bait Constructs

1. pDONr/pZeo® plasmid DNA. 2. pGBT9-GW plasmid DNA. 3. BP recombinase®. 4. LR recombinase®. 5. TE buffer: 10 mM Tris–HCl, pH 8; and 1 mM EDTA, pH 8. 6. Chemically competent DH5α bacteria.

6

Monica Castro-Cruz et al.

7. LB liquid medium: 10 g/L tryptone, 10 g/L NaCl, 5 g/L yeast extract. 8. LB solid medium: 10 g/L tryptone, 10 g/L NaCl, 5 g/L yeast extract, 20 g/L agar. 9. 1000 Zeocin stock solution: 100 mg/mL Zeocin 10. 1000 Ampicillin stock solution: 100 mg/mL Ampicillin 11. DNA miniprep extraction kit. 12. BsrGI restriction enzyme with its reaction buffer. 1. Y187: MATα, ura3-52, his3-200, ade2-101, trp1-901, leu2-3, 112, gal4Δ, met-, gal80Δ, MEL1, URA3:: GAL1UAS -GAL1TATA-lacZ.

2.4 Y2H Yeast Strains

2. AH109: MATa, trp1-901, leu2-3, 112, ura3-52, his3-200, gal4Δ, gal80Δ, LYS2:: GAL1UAS-GAL1TATA-HIS3, GAL2UAS-GAL2TATA-ADE2, URA3:: MEL1UAS-MEL1 TATA-lacZ. 1. Yeast extract–peptone–dextrose (YPD) medium: 10 g/L yeast extract 10 g/L, 20 g/L Bacto peptone 20 g/L, 20 g/L glucose, 100 mg/L adenine hemisulfate.

2.5 Yeast Culture Media

2. YPD solid medium: YPD medium plus 20 g/L Bacto agar. 3. Amino acid powder mix: Mix 6 g of each of the amino acids listed in Table 1, supplement the mix with 6 g of adenine hemisulfate. 4. 125 supplement solutions: 20 mM uracil; 100 mM leucine; 100 mM histidine; 40 mM tryptophan. Histidine and tryptophan 125X solutions should be stored at 4  C and protected from light 5. Synthetic Complete (SC) liquid Medium: 1.7 g/L yeast nitrogen base without amino acids, 5 g/L ammonium sulfate, 20 g/ L glucose, 1.3 g/L amino acid powder mix. Add 8 mL per liter of the required 125 supplement solutions to achieve the desired selection condition as detailed in Table 2. 6. SC solid medium: SC medium plus 20 g/L Bacto agar.

Table 1 Amino acid list to supplement synthetic complete media Alanine

Asparagine

Glutamine

Lysine

Proline

Tyrosine

Arginine

Cysteine

Glycine

Methionine

Serine

Valine

Aspartic acid

Glutamic acid

Isoleucine

Phenylalanine

Threonine

Y2H Screening of PDZ Interactions

7

Table 2 Amino acid dropout supplements and selection purposes Selection purpose

Selective SC medium

Required supplement (8 mL/L)

Prey selection

SC -Leu

Trp, His, Ura

Bait selection

SC -Trp

Leu, His, Ura

Bait autoactivation readout

SC -Trp -His

Leu, Ura

Mating efficiency readout

-Leu -Trp

His, Ura

Y2H interaction readout

-Leu -Trp -His

Ura

2.6 Yeast Transformation

1. Wild-type and PBM mutant Y2H bait vectors (pGBT9-GW). 2. YPD liquid medium. 3. Sterile ddH2O. 4. SC –Trp plates (see Table 2). 5. 10 TE buffer: 100 mM Tris–HCl, pH 8; and 10 mM EDTA, pH 8 (autoclave). 6. 10 lithium acetate (LiAc): 1 M LiAc. Adjust pH to 7.5 with dilute acetic acid, filter-sterilize on 0.22 μm membrane, and store at 4  C. 7. 10 mg/mL single-stranded carrier DNA from salmon sperm. 8. PEG–TE–LiAC solution: 50% (w/v) polyethylene glycol 4000 (PEG 4000) in 1 TE–LiAc solution (autoclave). 9. Dimethyl sulfoxide (DMSO).

2.7 Screening of the Human PDZome Library

1. SC -Leu plates. 2. Y2H PDZome array (frozen stock). 3. Wild-type and PBM mutant bait-containing AH109 strains. 4. SC -Leu liquid medium. 5. SC -Trp liquid medium. 6. YPD liquid medium. 7. SC -Leu, -Trp plates. 8. SC -Leu, -Trp, -His plates. 9. 96w polypropylene deep-well plates 10. 96-well U-bottom microplates.

8

3

Monica Castro-Cruz et al.

Methods

3.1 Amplification and Purification of the Y2H Bait Constructs

1. Set up two 50 μL PCR reactions to amplify the wild type and PBM mutant versions of the bait of interest using the forward and reverse primers described in Subheading 2.1. 2. Start by mixing 1 μL of dNTPs, 2 μL of Forward and Reverse primers, 50–100 ng of template bait DNA, the volume of PCR reaction buffer and magnesium chloride recommended by the DNA polymerase manufacturer, complete with sterile ddH2O water up to 50 μL and finish with the addition of DNA polymerase, as recommended by the manufacturer. 3. Run both PCR reactions according to the DNA polymerase manufacturer recommendations. A rather low number (typically 18 to 20) of PCR cycles is generally enough when amplifying short DNA fragments. 4. Check the size and abundance of the PCR product by electrophoresis, using a 1–2% agarose gel (depending on the size of the amplified DNA fragment) and loading 2 μL of the PCR reaction. 5. Gel-purify 10–20 μL of the PCR products on a new agarose gel to remove any unreacted primers that may interfere with the recombination step. To accomplish this step use the reagents provided with the gel DNA extraction kit and following the manufacturer instructions. Be aware that small PCR products (0.20 represents a high-confidence binding event. Below this threshold, 0.10 < BI 5000 Daltons, such as proteins, form well-defined structures that undergo thermally induced conformational changes [34]. These structural rearrangements result in the absorption of heat caused by the redistribution of noncovalent bonds. Differential scanning calorimeters measure this heat uptake. Concerning NanoDSF, this technique records the intrinsic fluorescence of Tryptophan (Trp) and Tyrosine (Tyr) residues, which are very sensitive to changes in their local environment. Thermal unfolding is measured by monitoring the intrinsic Trp and Tyr fluorescence intensity, and the position of the emission maximum as a function of temperature. The fluorescence intensity ratio between 330 and 350 nm is defined as an empirical parameter to monitor the evolution of the microenvironment of the aromatic residues during protein denaturation throughout the temperature increases. This ratio sharply increases/decreases during thermal unfolding, allowing to determine a Tm value [35]. The applicability of nanoDSF is highly dependent on the presence of Trp and Tyr in the folded core of the PDZ that are exposed upon unfolding. Moreover, it is necessary to exclude that the observed signal changes are caused by aggregation, as this will also lead to variations in the fluorophore environment. A back-scattering measurement can also be performed to determine if aggregation occurs (before or concomitantly with the denaturation). The characteristics of the temperature gradient are essential as it is related to the activation energy via the Arrhenius equation [6]. Typically, a heating rate of 1  C/min is applied. Overall, comparing the melting temperatures (ΔTm) in different buffer compositions allows researchers to define the optimal buffer condition as an increase in Tm corresponds to a better thermal stability and to a reduced conformational flexibility. 1.4.3 Nuclear Magnetic Resonance (NMR)

NMR is based on the measure of the absorption of radiofrequency (RF) radiation by an atomic nucleus located in a strong magnetic field. The principle of NMR is that atomic nuclei, with an odd number of protons (1H, 13C, 15N, 31P, . . .), neutrons, or both, have an intrinsic nuclear spin. When an atomic nucleus with a nonzero spin is placed in a magnetic field, the nuclear spin aligned in the same direction or in the opposite direction to the field. Different energies characterize these two types of nuclear spin alignment, and the application of a magnetic field facilitates the degeneration of nuclear spins. An atomic nucleus whose spin is aligned with the field will have less energy than when its spin is aligned in the opposite direction of the field. The energy of an NMR transition depends on the magnetic field strength as well as on the proportionality factor applied to each nucleus called the gyromagnetic ratio. The local environment

PDZ Sample Quality Assessment

101

around a given nucleus in a molecule tends to slightly disturb the local magnetic field exerted on that nucleus and to affect its transition energy. This dependence of the transition energy on the position of a particular atom in a molecule makes NMR extremely useful for determining the structure of molecules. The sample placed in an intense magnetic field will be disturbed by radiofrequency pulses. The recording of the return to the equilibrium of the spins makes it possible to have access to the chemical environment of the atoms. This information offers the possibility of structural and dynamic analyzes as well as the study of interactions involving biological macromolecules. NMR can be used to characterize the degree of folding of a protein by observing the dispersion of the resonance peaks. Indeed, in the one-dimensional (1D) 1H spectrum or the two-dimensional (2D) 1H-15N correlation spectra if the protein is 15N-labeled, the peaks for a well-folded protein are narrow and sharp and distributed over a large range of chemical shifts (good signal dispersion) (Fig. 6). 1H resonances can be especially found at values 8.5 ppm (corresponding to down field-shifted amide protons). In contrast, the peaks can be broader and not as widely dispersed in the spectrum of an unfolded or partially folded protein. Moreover, observed linewidths of peaks are related to the molecular weight of the protein, and then may be indicative of autoassociation or aggregation. In addition to the evaluation of the folding and the stability of a protein, a 1D 1H spectrum provides information about purity. Indeed, impurities with low molecular weight and observable nuclei give rise to sharp signals amongst the broader envelope of the protein resonances. The most common 2D spectrum to obtain structural information about a protein is the 1H-15N heteronuclear single quantum coherence (HSQC) spectrum. It correlates the nitrogen atom of an amide group with the directly attached amide proton. Since there is only one backbone HN per amino acid, except for Pro, each HSQC signal represents one single amino acid. The HSQC also contains signals from the NH2 groups of the side chains of asparagine and glutamine (also lysine and arginine depending on the pH values) and of the aromatic HN protons of Trp and Histidine. The signals may cover a spectral range from 6.0 to 12 ppm. If a protein is folded, its signals are distributed over the complete spectral range; if not, signals are located between 7.5 and 8.5 ppm in the proton dimension. Signals outside these regions indicate that the protein is folded and in a defined three-dimensional (3D) state. The chemical shift is very sensitive to the overall structure so that even slight conformational changes will affect the signals in a 1H-15N HSQC spectrum, making it a very efficient tool to check conformational stability and folding states.

Ce´lia Caillet-Saguy et al.

A

Downfield backbone NH

11

10

9

Downfield Hα

8

7

6

5

1H

Upfield CH3

4

3

2

1

0

-1

-2

ppm

(ppm)

B 105

(ppm)

110

115

15N

102

120

125

130

10.0

9.5

9.0

8.5 8.0 1H (ppm)

7.5

7.0

Fig. 6 NMR spectra of MAST2-PDZ. (a) 1H 1D spectrum of 150 μM MAST2-PDZ in complex with a viral peptide at 600 MHz 1H frequency. (b) 1H-15N 2D HSQC spectra of 150 μM MAST2-PDZ alone (in black) and in complex with a viral peptide (in red) at 600 MHz 1H frequency

PDZ Sample Quality Assessment

2 2.1

103

Materials Purity

2.1.1 Basic Reagents for SDS-PAGE Electrophoresis

1. Electrophoresis tank. 2. Power supply. 3. Polyacrylamide gels. Novex™ 4–12% Tris-Glycine Mini Gels offers a broad range of molecular weight resolution from 6 to 400 kDa. Standard 10–12% polyacrylamide gels can also be used to resolve PDZ domains. 4. 2 Sample buffer: 100 mM Tris–HCl pH 6.8, 0.2% Bromophenol Blue (BBP), 4% SDS, 20% glycerol, 200 mM dithiothreitol (DTT) (see Note 2). Add 1 ml of 1 M Tris pH 6.8, a pinch of BromoPhenol Blue (BPB) and vortex, 4 ml of 10% SDS, 2 ml of 100% glycerol, and 2 ml of 1 M DTT. 5. Molecular-weight markers such as PageRuler™ Prestained Protein Ladders from 10 to 180 kDa. 6. Migration buffer such as NuPAGE MES SDS Running Buffer. Add 50 ml of the NuPAGE™ MES SDS Running Buffer (20) and complete to 1 l with pyrolysis water. 7. Home-made or commercial protein staining solution. InstantBlue is a ready to use Coomassie protein stain for polyacrylamide gels.

2.1.2 Capillary Electrophoresis (CE)

1. Capillary electrophoresis system in this case a LabChip GX II (PerkinElmer). 2. LabChip HT Protein Express Chip (PerkinElmer, 760499). 3. Protein Express CLS960008).

Assay

Reagent

Kit

(PerkinElmer,

4. HT Protein 200 Sample Buffer (PerkinElmer 760518). 5. 384-well PCR plate (Greiner Bio-One) (see Note 3) 2.1.3 UV Spectroscopy

1. UV-Vis spectrometer. 2. Quartz cuvette suitable for UV (190–400 nm). 3. Sample buffer as described in Subheading 2.2, item 4. 4. Special wipe such as Kimwipes to clean the cuvette before measuring.

2.2 Integrity Measurements by MALDI-TOF 2.2.1 Total Mass

1. MALDI steel plate. A freshly prepared matrix solution: 25 mg/ml α-cyano-4hydroxycinnamic acid (HCCA) in 50% (v/v) acetonitrile, 0.1% (v/v) trifluoroacetic acid (TFA) in HPLC water. 2. ZipTip with C4 resin (Merck Millipore). 3. High quality protein standard from Bruker, Laserbio Labs, or Sciex.

104

Ce´lia Caillet-Saguy et al.

2.2.2 ISD and T3 Sequencing

1. MALDI steel plate. 2. Two matrix solutions: 25 mg/ml 1,5-diaminonaphthalene (DAN) and 25 mg/ml of a mixture composed of 2,5-dihydroxybenzoic acid and 2-hydroxy-5-methoxybenzoic acid (super-DHB). Both matrices are dissolved in 50% Acetonitrile (v/v), 0.1% (v/v) TFA in HPLC water. 3. ZipTip with C4 resin (Merck Millipore). 4. 1 mg/ml BSA in PBS as ISD standard.

2.3

Homogeneity

2.3.1 Dynamic Light Scattering (DLS)

1. 0.22 μM Syringe filter. 2. Reusable quartz cuvettes (Hellma) or plates depending on the model. 3. Detergent solution: 2% (v/v) Hellmanex. 4. Buffer used for PDZ sample: 1 ml. 5. PDZ sample.

2.3.2 Size Exclusion Chromatography (SEC) with Multi Angle Light Scattering (SEC-MALS)

1. Size exclusion column (e.g., Superdex 75 increase from GE healthcare). 2. PBS (137 mM [NaCl], 2.7 mM [KCl], 10 mM [Na2HPO4], 1.8 mM [KH2PO4]) (1 l) or other buffer (see Note 4). 3. PDZ sample. Prepare your sample in the chosen buffer with a protein concentration at least of 1 mg/ml. This is particularly true for PDZ domain if you want sufficient signal on the light scattering detector as the signal is proportional to the molecular mass of the protein.

2.4 Conformational Stability/Folding State

1. Aviv 215 spectropolarimeter.

2.4.1 Circular Dichroism (CD)

3. PDZ sample diluted at 0.2 mg/ml with PDZ sample buffer.

2. 0.2 mm path-length cylindrical cell (Hellma) (see Note 5). 4. PDZ sample buffer. 5. 70% ethanol. 6. 2% (v/v) Hellmanex.

2.4.2 Differential Scanning Fluorimetry (Nano DSF)

1. Prometheus NT.48. 2. 70% ethanol to clean the lens. 3. Quartz capillary. 4. Buffer used for PDZ domain preparation and storage. 5. PDZ sample diluted at 0.2 mg/ml with PDZ sample buffer. 6. Peptide solubilized in PDZ sample buffer (tenfold molar excess relative to PDZ sample).

PDZ Sample Quality Assessment

105

1. 600 MHz (or higher fields) NMR spectrometer equipped with triple resonance 1H{13C/15N} PFG probe (for heteronuclear NMR detection).

2.4.3 NMR

2. 4 mm NMR tube (Shigemi INC). 3. 100 μM PDZ sample (250 μl). 4. 100% D2O to add to the PDZ sample at 1–5% final concentration.

3

Methods We used the PDZ domain of the human kinase MAST2 (UniProtKB Q6P0Q8) as a practical model. This construct of 96 residues (MAST2-PDZ) (Fig. 7a) has a molecular weight of 10435.9 Da.

3.1

The objective is to quantify the PDZ construct of interest and assess the presence (and level) of contaminants using techniques that are available to most laboratories.

Purity

A simple protocol to run an SDS-PAGE is given below:

3.1.1 SDS-Page

1. Prepare 20 μl of sample by adding 10 μl of 2 denaturing loading buffer to 10 μl of protein and heat sample at 95  C for 5 min (see Note 6). 2. Prepare the precasted gel by removing the comb and the white tape near the bottom of the gel cassettes and place the gel in the mini gel tank (see Note 7). A

B

MAST2-PDZ NH2GGSMRPPIIIHRAGKKYGFTLRAIRVYMGDSDVYTVHHMVWHVEDGGPASEAGLRQ GDLITHVNGEP VHGLVHTEVVELILKSGNKVAISTTPLENCOOH

MW

C

MW

98 kDa 62 kDa 49 kDa 38 kDa 28 kDa 17 kDa 14 kDa 6 kDa

Fig. 7 Sequence and gel SDS-PAGE of MAST2-PDZ. (a) Primary sequence of MAST2-PDZ. (b) The SDS-PAGE gels of a GSTrap chromatography (left panel) and a size-exclusion chromatography (right panel) of the eluted fractions containing MAST2-PDZ

106

Ce´lia Caillet-Saguy et al.

3. Fill the chambers with running buffer. 4. Load samples and ladders in the appropriate wells. 5. Run the gel following the manufacturer’s instructions (see Note 8). 6. After electrophoresis, remove the gel from the tank and transfer directly into the InstantBlue staining solution. Be sure that the gel moves freely in stain to facilitate diffusion. Typically, about 20 ml is needed to cover the gel. 7. Colored proteins bands will start to develop immediately and a suitable intensity is typically achieved after 15 min incubation at room temperature with gentle shaking. The protein sample is considered pure when no other band other than the expected one for your protein of interest is detected by SDS-PAGE using sensitive staining [20]. In the case of MAST2-PDZ, the protein is purified as previously described [36]. The 2-step purification process improved drastically the purity of the sample as assessed by SDS-PAGE in Fig. 7b. The GSTrap column allowed to remove a high content of contaminants and to concentrate the MAST2-PDZ construct. The SDS-PAGE gel of the eluted fractions showed a major band of the protein and only a few contaminants extra-bands (Fig. 7b left panel). Fractions containing MAST2-PDZ are pooled and then purified by size-exclusion chromatography and the only band remaining on the corresponding gel SDS-PAGE is the MAST2PDZ construct. A very weak band around 30 kDa slightly overlap with the first fractions of MAST2-PDZ. The pool of fractions of gel filtration is then concentrated and the gel SDS-PAGE revealed one unique band indicative of the high degree of purity by SDS-PAGE technique. This construct has a migration that does not match with its calculated molecular weight as already observed. This might be due to the effect of SDS binding [37] and/or the hydrophobicity of the construct [38]. The limits of this technique are that the detection is restricted to protein contaminants, and the detection is proportional to the sample loading and gel staining and should be confirmed by mass spectroscopy. 3.1.2 Capillary Gel Electrophoresis (CGE)

Capillary gel electrophoresis (CGE) separates proteins according to their molecular mass similarly to SDS-PAGE, and this methodology has been previously exploited to validate bacterial expression of a library of 266 known human PDZ constructs developed for the high-throughput “holdup” chromatographic assay for the determination of PDZ-PBM affinities [39]. Below we outline the main steps of this high-throughput protocol. 1. Dilute to 1:8 (v:v), 1:16 (v:v), and 1:32 (v:v) each PDZ supernatant to be quantified.

PDZ Sample Quality Assessment

107

2. Dispense 5.6 μl/well of HT Protein 200 Sample buffer in as many wells of a 384 well plate as required. 3. Add 4 μl/well of the PDZ dilutions to each well. 4. Boil the samples at 95  C for 5 min in a dry bath. 5. Add 20.6 μl/well of distilled H2O. 6. PDZ samples within the 384 well plate is then analyzed with the High Sensitivity HT Protein Express protocol (10–100 kDa program) with the LabChip GXII device (PerkinElmer) following supplier’s instructions. This technology allows to electrophoretically separate, stain, destain, detect by laser-induced fluorescence and analyze the protein samples. The data analysis provides protein concentration, molecular weight sizing, and purity evaluation using ladder and marker calibration standards. With sample acquisition time of about 40 s, the instrument takes approximately 4 h to analyze 384 protein samples. 7. The LabChip GX software is then used for data analysis. This allows to visualize results via an electropherogram or virtual gel view (Fig. 8) or in a tabular form to export into a spreadsheet format. At the end of the run, the concentrations of the initial cultures of the soluble PDZ are calculated from the concentrations determined per PDZ in each serial dilution. The PDZ lysate concentrations range from 10 μM for the least concentrated and up to 100 μM or higher. The concentration of lysozyme is constant and used as an internal reference for quantification. A limitation to the quantification is reached when the construct is poorly expressed and falls in the background of the Escherichia coli proteins in the case of a lysate. On the contrary, the electropherogram allows to easily detect contaminants (Fig. 8). This CGE approach is fast, efficient and allows high resolving separations with low solvent consumption and minimal operating cost considering the high number of samples. However, the LabChip GXII system is not common in most laboratories. Its running cost, the short-use warranty of ships and kits, can constitute a limitation. During the 4 h run, a diminution in data quality may happen due to samples drying into the 384 well plate. 3.1.3 UV Spectroscopy

UV spectroscopy between 220–240 nm and 340 nm is a good quality test to determine protein concentration (using A280nm) and to detect aggregation and molecule contamination. Indeed, UV-visible spectroscopy can detect the presence of large particles such as aggregates (radius higher than 200 nm) in a protein preparation by monitoring the absorbance signal above 320 nm, where aggregate-free protein samples are not supposed to absorb light. If the signal increases as the wavelength diminishes between 340 nm

108

Ce´lia Caillet-Saguy et al.

Fig. 8 Interface of the data analysis software LabChip GX reviewer. Electropherogram (top) and virtual gel view (bottom) of the serial dilution of the supernatant of lysed cell extract of MBP-tagged MAST2-PDZ construct of the PDZome library. The concentration of lysozyme is constant and used as an internal reference for quantification. The peak corresponding to the PDZ construct is decreasing upon increasing dilution factor

λ (nm)

λ (nm)

Fig. 9 UV spectra of two MAST2-PDZ samples after purification

and 320 nm, it can be attributed to the scattering of light by large aggregates present in the sample. In Fig. 9, the two UV spectra of a sample of MAST2-PDZ show a difference in the absorbance above 320 nm, indicative of a difference in the presence of aggregates. The aggregation index (AI) can be calculated: AI ¼ 100  A340nm/ (A280nm  A340nm) with A280nm and A340nm, the absorbances at 280 and 340 nm, respectively. AI 0.5

>0.5

pH 8.0

220

0.02

0.13

0.24

>0.5

HEPES

pH 7.5

230

0.37

0.5

>0.5

>0.5

PIPES

pH 7.0

230

0.2

0.49

0.29

>0.5

MOPS

pH 7.0

230

0.1

0.34

0.28

>0.5

MES

pH 6.0

230

0.07

0.29

0.29

0.15

Cacodylate

pH 6.0

210

0.01

0.20

0.22

NaClO4

170

0

0

0

0

NaF, KF

170

0

0

0

0

Boric Acid

180

0

0

0

0

NaCl

205

0

0.02

>0.5

>0.5

Na2HPO4

210

0

0.05

0.3

>0.5

NaH2PO4

195

0

0

0.01

0.15

Na Acetate

220

0.03

0.17

>0.5

>0.5

Glycine

220

0.03

0.1

>0.5

>0.5

Diethylamine

240

0.4

>0.5

>0.5

>0.5

NaOH

pH 12

230

>0.5

>2

>2

>2

Boric Acid, NaOH

pH 9.1

200

0

0

0.09

0.3

absorb in the far UV range, samples for CD should not contain NaCl and should be non-HCl buffered. Other buffer components such as DTT, imidazole, DMSO, and glycerol should also be removed via dialysis or buffer exchange before proceeding with the experiment. The protocol below describes the sample preparation of cell polarity proteins for CD spectroscopy. It can be used to estimate the secondary structure components and their proper folding patterns, before obtaining higher-resolution structural information by X-ray crystallography. 1. Wash 15 mL centrifugal concentrator with 15 mL of phosphate buffer by centrifugation. 2. Add 0.1 mg of cell polarity protein with the final concentration of 0.3 mg/mL to the concentrator.

Biophysics of Scribble Module PDZ Interactions

131

3. Top up the concentrator with phosphate buffer and centrifuge for 20 min at 4000  g. 4. After the first centrifugation, discard the flow through and top up the concentrator again with fresh phosphate buffer and spin again for up to 20 min at 4000  g. 5. Repeat step 4 for another 4 times. 6. Concentrate the protein sample (now in phosphate buffer) to a final concentration of 0.3 mg/mL and use approximately 300 μL for CD experiment (see Note 1). 3.2 Interactions of Polarity Protein PDZ Domains with Ligands

Techniques used to analyze protein interactions such as surface plasmon resonance (SPR) can be hampered by the requirement for immobilization of proteins before the interaction can be observed. Isothermal titration calorimetry (ITC) provides a quantitative in-solution method for measuring protein interactions without immobilization. In a single measurement, ITC allows for accurate measurements of the binding constant (KD), reaction stoichiometry (n), enthalpy (ΔH), and entropy (ΔS) of a given biomolecular binding event whereby heat is either released or absorbed [23]. ITC is highly sensitive to small differences in concentration of buffer components and pH, thus making control measurements to establish buffer effects advisable. In addition to buffer control measurements, control measurements to gauge PDZ domain protein quality interactions are useful. For this purpose, we use a synthetic peptide that binds to a vast range of PDZ domains with high affinity termed superpeptide (RSWFETWV), which was identified via phage display experiments [24]. 1. Ensure reference cell contains water. 2. Rinse sample cell and syringe with dH2O and storage buffer three times each (see Note 4). 3. Titrations are performed at 25  C with a stirring speed of 750 rpm. 4. A total of 20 injections with 2 μL of the peptide solution each and a spacing of 180 s were titrated into the 200 μL protein sample, except for the first injection which was only 0.4 μL (see Note 5). 5. Process the raw thermograms with Origin® to obtain the binding parameters of each interaction. Typical examples are shown in Fig. 1.

3.3 Preparation of Cell-Polarity Protein–Peptide Complexes for Crystallization

Scribble and Dlg PDZ domains interact with a diverse range of known partners, and preparation of complexes of these domains with the interacting partners is a key step to understand their multifunctionality and promiscuous behavior, and ultimately their role in disease mechanisms. In this example we demonstrate how to

132

Janesha C. Maddumage et al.

Fig. 1 Biophysical characterization of Scribble PDZ domain–ligand complexes. (a) Cartoon structure of D. melanogaster Scribble PDZ1 domain bound to Gukh with associated interaction measured by ITC and CD spectrum of Scribble PDZ1 domain. (b) Cartoon structure of human Scribble PDZ1 domain bound to phosphoMCC with associated interaction measured by ITC and CD spectrum of Scribble PDZ1 domain. (c) Cartoon structure of human Scribble PDZ3 domain bound to β-Pix with associated interaction measured by ITC and CD spectrum of Scribble PDZ3 domain

prepare a complex of an individual PDZ domain with a peptide corresponding to the C-terminal PDZ binding motif (PBM) of the interacting partner that has been identified using ITC. In general PDZ domains typically harbor micromolar affinities toward their binding partners, and the outlined method has been successfully used to prepare complexes for crystallization trials of Scribble/Dlg PDZ domains with peptides [25, 26] and led to the determination of a number of crystal structures (see Table 1 and Fig. 1).

Biophysics of Scribble Module PDZ Interactions

133

1. Wash 1 mL centrifuge concentrator with PDZ storage buffer. 2. Add 1 mg of PDZ domain in final sample buffer. 3. Slowly add at least 2–5 molar excess of the target peptide to centrifugal concentrator while stirring with pipette to avoid local precipitation of sample (see Note 6). 4. Concentrate the complex to a final concentration of 5 mg/mL (see Note 7). 3.4 Crystallization of Tandem Scribble PDZ Domains Bound to Interacting Peptides

4

As Scribble and Dlg consist of multi-PDZ domains, two or more of these PDZ domains can be connected to each other in tandem to form PDZ supramodules, which may give rise to unique binding properties with their interactors that are different from when they exist as individual domains [27]. Since these tandem PDZ domains are often connected to each other via very short linkers, these domains act as a one structural and functional unit [27]. Preparation of complexes of tandem domains with a target peptide/peptide can be also achieved using a similar method as mentioned above. The protein concentration and the protein–peptide ratio may vary. A successful crystallization example is the structure determination of Scribble PDZ34 (8 mg/mL) in complex with a commercially synthesized peptide (S-W-F-Q-T-D-L) in a 5:1 peptide– protein molar ratio (PDB Id:4WYU) [28].

Notes 1. Phosphate buffer is prepared by mixing Na2HPO4 and NaH2PO4 to achieve the desired pH in order to avoid the use of HCl. 2. Suitable peptide length for ITC is eight or more amino acids. However, peptides of four amino acids in length have been successfully characterized. 3. The synthetic peptide stocks are stored in dH2O at a concentration of 5 mM. 4. It is crucial to have an identical pH for sample and peptide buffers when performing ITC to prevent spurious heat peaks due to the heat of dilution of protons. 5. Usually PDZ domain proteins yield high-quality thermograms at concentrations of 75 μM for protein and 900 μM for peptide; however, some concentration optimization may be required. 6. Peptide concentration in crystallization trials should be at least at a twofold molar excess. 7. Final concentration of the sample for crystallization experiment may vary with each sample and should be confirmed using a spectrophotometer to measure UV absorbance at 280 nm wavelength.

134

Janesha C. Maddumage et al.

References 1. Nelson WJ (2003) Adaptation of core mechanisms to generate cell polarity. Nature 422:766–774 2. Halaoui R, McCaffrey L (2015) Rewiring cell polarity signaling in cancer. Oncogene 34:939–950 3. Stephens R, Lim K, Portela M, Kvansakul M, Humbert PO, Richardson HE (2018) The scribble cell polarity module in the regulation of cell signaling in tissue development and tumorigenesis. J Mol Biol 430:3585–3612 4. Subbaiah VK, Kranjec C, Thomas M, Banks L (2011) PDZ domains: the building blocks regulating tumorigenesis. Biochem J 439:195–205 5. Ye F, Zhang M (2013) Structures and target recognition modes of PDZ domains: recurring themes and emerging pictures. Biochem J 455:1–14 6. Alie´ A, Manuel M (2010) The backbone of the post-synaptic density originated in a unicellular ancestor of choanoflagellates and metazoans. BMC Evol Biol 10:34 7. Belahbib H, Renard E, Santini S, Jourda C, Claverie JM, Borchiellini C, Le Bivic A (2018) New genomic data and analyses challenge the traditional vision of animal epithelium evolution. BMC Genomics 19:393 8. Daulat AM, Puvirajesinghe TM, Camoin L, Borg JP (2018) Mapping cellular polarity networks using mass spectrometry-based strategies. J Mol Biol 430:3545–3564 9. Pires HR, Boxem M (2018) Mapping the polarity interactome. J Mol Biol 430:3521–3544 10. Wen W, Zhang M (2018) Protein complex assemblies in epithelial cell polarity and asymmetric cell division. J Mol Biol 430:3504–3520 11. Songyang Z, Fanning AS, Fu C, Xu J, Marfatia SM, Chishti AH, Crompton A, Chan AC, Anderson JM, Cantley LC (1997) Recognition of unique carboxyl-terminal motifs by distinct PDZ domains. Science 275:73–77 12. Hillier BJ, Christopherson KS, Prehoda KE, Bredt DS, Lim WA (1999) Unexpected modes of PDZ domain scaffolding revealed by structure of nNOS-syntrophin complex. Science 284:812–815 13. Awadia S, Huq F, Arnold TR, Goicoechea SM, Sun YJ, Hou T, Kreider-Letterman G, Massimi P, Banks L, Fuentes EJ, Miller AL, Garcia-Mata R (2019) SGEF forms a complex with scribble and Dlg1 and regulates epithelial junctions and contractility. J Cell Biol 218:2699–2525

14. Penkert RR, DiVittorio HM, Prehoda KE (2004) Internal recognition through PDZ domain plasticity in the Par-6-Pals1 complex. Nat Struct Mol Biol 11:1122–1127 15. Chetkovich DM, Chen L, Stocker TJ, Nicoll RA, Bredt DS (2002) Phosphorylation of the postsynaptic density-95 (PSD-95)/discs large/zona occludens-1 binding site of stargazin regulates binding to PSD-95 and synaptic targeting of AMPA receptors. J Neurosci 22:5791–5796 16. Clairfeuille T, Mas C, Chan AS, Yang Z, TelloLafoz M, Chandra M, Widagdo J, Kerr MC, Paul B, Me´rida I, Teasdale RD, Pavlos NJ, Anggono V, Collins BM (2016) A molecular code for endosomal recycling of phosphorylated cargos by the SNX27-retromer complex. Nat Struct Mol Biol 23:921–932 17. Sundell GN, Arnold R, Ali M, Naksukpaiboon P, Orts J, Gu¨ntert P, Chi CN, Ivarsson Y (2018) Proteome-wide analysis of phospho-regulated PDZ domain interactions. Mol Syst Biol 14:e8129 18. Grant SGN (2019) Synapse diversity and synaptome architecture in human genetic disorders. Hum Mol Genet 21:219–225 19. Thomas M, Banks L (2018) Upsetting the balance: when viruses manipulate cell polarity control. J Mol Biol 430:3481–3503 20. Franck B (1965) Optical circular dichroism. Principles, measurements, and applications. Von L. Velluz, M. Legrand und M. Grosjean. Angew Chem 77:875–875 21. Greenfield NJ (2006) Using circular dichroism spectra to estimate protein secondary structure. Nat Protoc 1:2876–2890 22. Aviv Biomedical, Inc. Circular dichroism spectrophotometer. Instrument manual, 2009 23. Saponaro A (2018) Isothermal titration calorimetry: a biophysical method to characterize the interaction between label-free biomolecules in solution. Bio-protocol 8:e2957 24. Zhang Y, Yeh S, Appleton BA, Held HA, Kausalya PJ, Phua DC, Wong WL, Lasky LA, Wiesmann C, Hunziker W, Sidhu SS (2006) Convergent and divergent ligand specificity among PDZ domains of the LAP and zonula occludens (ZO) families. J Biol Chem 281:22299–22311 25. Lim KYB, Go¨dde NJ, Humbert PO, Kvansakul M (2017) Structural basis for the differential interaction of scribble PDZ domains with the guanine nucleotide exchange factor beta-PIX. J Biol Chem 292:20425–20436

Biophysics of Scribble Module PDZ Interactions 26. Caria S, Stewart BZ, Jin R, Smith BJ, Humbert PO, Kvansakul M (2019) Structural analysis of phosphorylation-associated interactions of human MCC with scribble PDZ domains. FEBS J 286:4910–4925 27. Feng W, Zhang M (2009) Organization and dynamics of PDZ-domain-related supramodules in the postsynaptic density. Nat Rev Neurosci 10:87–99 28. Ren J, Feng L, Bai Y, Pei H, Yuan Z, Feng W (2015) Interdomain interface-mediated target recognition by the scribble PDZ34 supramodule. Biochem J 468:133–144 29. Caria S, Magtoto CM, Samiei T, Portela M, Lim KYB, How JY, Stewart BZ, Humbert PO, Richardson HE, Kvansakul M (2018) Drosophila melanogaster Guk-holder interacts with the scribbled PDZ1 domain and regulates epithelial development with scribbled and discs large. J Biol Chem 293(12):4519–4531 30. How JY, Caria S, Humbert PO, Kvansakul M (2019) Crystal structure of the human scribble PDZ1 domain bound to the PDZ-binding motif of APC. FEBS Lett 593:533–542 31. Zhang Z, Li H, Chen L, Lu X, Zhang J, Xu P, Lin K, Wu G (2011) Molecular basis for the recognition of adenomatous polyposis coli by the discs large 1 protein. PLoS One 6:e23507 32. Slep KC (2012) Structure of the human discs large 1 PDZ2- adenomatous polyposis coli cytoskeletal polarity complex: insight into peptide engagement and PDZ clustering. PLoS One 7:e50097 33. Mischo A, Ohlenschl€ager O, Hortschansky P, Ramachandran R, Go¨rlach M (2013)

135

Structural insights into a wildtype domain of the oncoprotein E6 and its interaction with a PDZ domain. PLoS One 8:e62584 34. Liu Y, Henry GD, Hegde RS, Baleja JD (2007) Solution structure of the hDlg/SAP97 PDZ2 domain and its mechanism of interaction with HPV-18 papillomavirus E6 protein. Biochemistry 46:10864–10874 35. Zhang Y, Dasgupta J, Ma RZ, Banks L, Thomas M, Chen XS (2007) Structures of a human papillomavirus (HPV) E6 polypeptide bound to MAGUK proteins: mechanisms of targeting tumor suppressors by a high-risk HPV oncoprotein. J Virol 81:3618–3626 36. Sainlos M, Iskenderian-Epps WS, Olivier NB, Choquet D, Imperiali B (2013) Caged monoand divalent ligands for light-assisted disruption of PDZ domain-mediated interactions. J Am Chem Soc 135:4580–4583 37. Wang W, Weng J, Zhang X, Liu M, Zhang M (2009) Creating conformational entropy by increasing interdomain mobility in ligand binding regulation: a revisit to N-terminal tandem PDZ domains of PSD-95. J Am Chem Soc 131:787–796 38. Zeng M, Shang Y, Araki Y, Guo T, Huganir RL, Zhang M (2016) Phase transition in postsynaptic densities underlies formation of synaptic complexes and synaptic plasticity. Cell 166:1163–1175 39. Raman AS, White KI, Ranganathan R (2016) Origins of allostery and evolvability in proteins: a case study. Cell 166:468–480

Chapter 8 A Fluorescence-Based Assay to Determine PDZ–Ligand Binding Thermodynamics Young Joo Sun and Ernesto J. Fuentes Abstract Postsynaptic density-95, disks-large, and zonula occludens-1 (PDZ) domain interactions with cognate linear binding motifs (i.e., PDZ-binding motifs or PBMs) are important for many biological processes and can be pathological when disrupted. There are hundreds of PDZ–PBM interactions reported but few have been quantitatively determined. Moreover, PDZ–PBM interactions have been identified as potential therapeutic targets. To thoroughly understand PDZ–PBM binding energetics and their specificity, we have developed a sensitive and quantitative equilibrium binding assay. Here, we describe a protocol for determining PDZ–PBM binding energetics using fluorescence anisotropy-based methodology. Key words PDZ domain, PDZ-binding motif, Fluorescence anisotropy, Protein–protein binding, CASK, Scribble, SGEF

1

Introduction Postsynaptic density-95, Disks-large, and Zonula occludens-1 (PDZ) proteins are ubiquitously found in many types of mammalian cells and regulate the spatial and temporal function of a diverse set of signaling pathways. A distinguishing feature of these proteins is the small (~90 amino acids, ~10 kDa), structurally conserved protein–protein interaction module known as a PDZ domain that selectively interacts with linear C-terminal and internal peptide motifs (i.e., PDZ-binding-motifs or PBMs) [1]. PDZ–PBM interactions and their specificity are critical for many biological processes including the maintenance of cell polarity, neuronal development, and signal transduction. Thus, it is not surprising that genetic mutations in PDZ proteins or perturbation of PDZ–PBM interactions can contribute to pathologies such as neuronal disorders and complications form brain injury, cancer, cystic fibrosis, and viral infections (reviewed in [1]).

Jean-Paul Borg (ed.), PDZ Mediated Interactions: Methods and Protocols, Methods in Molecular Biology, vol. 2256, https://doi.org/10.1007/978-1-0716-1166-1_8, © Springer Science+Business Media, LLC, part of Springer Nature 2021

137

138

Young Joo Sun and Ernesto J. Fuentes

Although PDZ–PBM interactions have been extensively characterized, there remains inadequate understanding of the general molecular mechanisms that determine PDZ–PBM specificity, particularly for internal PBMs. This is the result of the low sequence identity among PDZ domain homologs, promiscuous binding profiles, and context-dependent interaction mechanisms. Physiological PDZ–PBM interactions have relatively weak binding affinities, with a dissociation constant (Kd) ranging from μM to low mM [2–4]. To thoroughly characterize PDZ–PBM interactions it is necessary to determine the binding energetics (i.e., ΔGb, Gibbs free energy of binding) of PDZ–PBM interactions. The binding energetics coupled with high-resolution structural information and mutagenesis can provide deep insights into the binding mechanism and specificity of PDZ–PBM interactions [5–14]. Importantly, this information can be used to design potential PDZ–PBM protein– ligand inhibitors. Indeed, over the past ~10 years PDZ–PBM interactions have been identified as potential therapeutic targets (reviewed in [1]). Here, we describe a general protocol for determining the binding energetics of PDZ–PBM interactions using a robust and simple fluorescence anisotropy-based assay sensitive to interactions with dissociation constants in the 1 to ~500 μM range [15–17].

2 2.1

Materials Equipment

1. A spectrofluorometer equipped with excitation and emission polarizers and a magnetic stirrer is used to collect fluorescence anisotropy data [15]. Here, we use a Fluorolog-3 (Jobin Yvon, Horiba, NJ) controlled by the FluorEssence V3.8 software program (Jobin Yvon, Horiba, NJ). The spectrofluorometer is set to an excitation wavelength at 340 nm and an emission wavelength at 550 nm, specific for the dansyl [5-(dimethyl amino)naphthalene-1-sulfonyl] chloride fluorophore (see Note 1), with constant stirring at 25
C. The instrument light slit widths are adjusted in the range of 3–9 nm to optimize the signal-to-noise ratio and maximum output intensity—aiming for ~one million counts per second on the detector (see Note 2). 2. A quartz cuvette containing 4 polished windows, compatible with a magnetic stirring platform is used. We use a 2 mL,10 mm length path cuvette equipped with a stopper and stir bar (Hellma, NY; catalog #119F-10-40).

2.2 Constructs, Medium, and Reagents for PDZ Domain Purification

1. PDZ domains cloned into bacterial expression plasmids are used. Here we use the CASK PDZ domain cloned into pET28a (Novagen) and the Scribble PDZ1 cloned into a modified pET21a (Novagen) [18].

PDZ Domain Binding Thermodynamics

139

2. Luria–Bertani (LB) medium: 10 g/L tryptone, 5 g/L yeast extract, and 10 g/L sodium chloride. 3. 100 mg/mL ampicillin (pET21a) or 50 mg/mL kanamycin (pET28a) stock solution. 4. E. coli bacterial strain BL21(DE3) (Novagen). 5. 1 M isopropyl 1-thio-β-d-galactopyranoside (IPTG) stock solution. 2.3 Reagents and Solutions

All buffers should be of the highest purity available. 1. Binding buffer: 20 mM sodium phosphate pH 6.8, 50 mM sodium chloride, and 0.5 mM ethylenediaminetetraacetic acid (EDTA). 2. Peptides containing C-terminal PBMs are typically commercially synthesized. The peptides used here were chemically synthesized by GenScript Inc. (Piscataway, NJ) and used at >95% purity. Peptides corresponding to the C-terminus of partner proteins were 8 amino acids long, N-terminally dansylated and contained a free carboxylate at the C-terminus. An internal peptide from SH3-containing guanine nucleotide exchange factor (SGEF) contained 14 amino acids and was N-terminally dansylated. In addition, the C-terminal carboxylate group was amidated (see Note 3). 3. The amino acid sequence of the C-termini of the following human proteins was used in the development of this protocol: Neurexin-1 (residues 1470-1477: NKDKEYYV), Caspr4 (residues 1301-1308: ENQKEYFF), and Syndecan-1 (residues 303–310: TKQEEFYA). The internal SGEF peptide was derived from residues 42–55: KPNGLLITDFPVED [18]. 4. Resuspend ~2 mg aliquot of lyophilized peptide in 1 mL of binding buffer and adjust the pH to 6.8 to obtain a highly concentrated (1–2 mM) master stock solution. 5. Prepare a working stock solution of 0.130 mM stock dansylated-PBM by diluting the master stock with binding buffer (see Notes 4 and 5). 6. Store all PBM peptide stock solutions at 20
C in the dark (see Note 6).

3

Methods

3.1 Purified CASK and Scribble PDZ Domains

A protocol for the expression of CASK and Scribble PDZ domains is described below. Purification of these recombinant PDZ domains is beyond the scope of this chapter but can be achieved using similar methods and buffers as previously published [7, 8, 11]. In brief, CASK PDZ was purified using cation exchange and size-exclusion

140

Young Joo Sun and Ernesto J. Fuentes

chromatography (Superdex 75, GE Healthcare Life Sciences) while purification of the first PDZ domain of Scribble was carried out using nickel-chelate affinity chromatography. The N-terminal 6  His affinity tag of Scribble PDZ1 was removed by proteolysis with recombinant tobacco etch virus (rTEV) protease for 36 h at 4
C. Undigested protein, cleaved 6  His tag, and His-tagged rTEV were separated from the Scribble PDZ1 domain by nickelchelate chromatography. The digested PDZ1 protein was further purified using Superdex 75 size-exclusion chromatography. All proteins were concentrated to ~0.5–2.0 mM in binding buffer (see Note 5). The prepared samples were generally used immediately (see Note 7). 1. Supplement LB medium with the appropriate antibiotic for the desired PDZ domain (final concentration of 100 μg/mL ampicillin for the mpET21a CASK PDZ and 50 μg/mL kanamycin for pET28a Scribble PDZ1). 2. Grow bacterial cells transformed with either CASK or Scribble PDZ constructs in LB medium supplemented with antibiotic at 37
C under vigorous agitation to an optical density of a 0.6–0.8 measured at 600 nm wavelength. 3. Cool cultures to 18
C. 4. Induce protein expression by adding IPTG to 1 mM final concentration. 5. Incubate for an additional 16–18 hrs at 18
C. 6. Harvest bacteria by centrifugation. 7. Proceed with the purification of proteins and concentrate to ~0.5 to 2.0 mM in binding buffer (see Note 5). The prepared samples are generally used immediately (see Note 7). 3.2 Experimental Sample Preparation

3.2.1 Cuvette and PBM Peptide Preparation

The experimental design calls for using serial dilutions of the PDZ domain protein to cover the concentration range of 1–400 μM in discrete steps. Thus, a 0.5 mM stock provides reliable quantification ranging from 1 to 100 μM Kd, while a 2.0 mM stock provides reliable quantification ranging from 20 to 200 μM Kd. The PDZ domain concentration range and number of titration points can be adjusted in subsequent experiments to optimize the titration and obtain a more reliable Kd determination. 1. Rinse the cuvette with distilled and deionized H2O (ddH2O) and ethanol using a vacuum cuvette washer (see Note 8). 2. Air dry the cuvette after rinsing and/or cleaning. 3. Add stir bar and 1290 μL of binding buffer. 4. Add 10 μL of dansylated-PBM peptide stock solution to the cuvette to obtain a 1.0 μM PBM peptide concentration (see Note 2).

PDZ Domain Binding Thermodynamics

141

5. Gently mix the sample using a pipette being careful to avoid introducing bubbles (see Note 9). 6. Cover the cuvette to prevent introducing dust particles and to minimize the dansyl-fluorophore light exposure (see Note 10). 3.2.2 Preparation of PDZ Domain Dilution Stock Solutions

All the PDZ domain diluted solutions described below should be prepared in advance and on ice. 1. Prepare 26 μL of a 100-fold diluted PDZ domain stock per experiment (concentration 5–20 μM). 2. Prepare 10 μL of a tenfold diluted PDZ domain stock per experiment (concentration 50–200 μM). 3. Prepare 656 μL PDZ domain stock per experiment (concentration 0.5–2 mM).

3.3 Binding Assay Parameters and Data Collection

Below are the experimental parameters defined in the FluorEssence V3.8 software. Anisotropy data is collected after each addition of PDZ protein. 1. Experiment type: anisotropy. 2. Temperature: 25
C (or other desired temperature). 3. Number of data points is ~29 (see Note 11). 4. After adding the appropriate volume of PDZ domain, gently mix the sample using a micropipette to avoid out-gassing followed by the measurement of fluorescence anisotropy (three measurements are taken and averaged). A typical titration experiment uses the following PDZ domain concentrations and volumes per titration step. 5. Data point 1: the initial background measurement without any PDZ domain. This serves as the baseline control. 6. Data point 2 and 3: add 3 μL of 100-fold diluted stock PDZ domain (concentration 5–20 μM) at each step. 7. Data point 4 and 5: add 5 μL of 100-fold diluted stock PDZ domain (concentration 520 μM) at each step. 8. Data point 6: add 10 μL of 100-fold diluted stock PDZ domain (concentration 5–20 μM). 9. Data point 7 and 8: add 5 μL of tenfold diluted stock PDZ domain (concentration 50–200 μM) at each step. 10. Data point 9 and 10: add 3 μL of PDZ domain stock (concentration 0.5–2 mM) at each step. 11. Data point 11 and 12: add 5 μL of PDZ domain stock (concentration 0.5–2 mM) at each step. 12. Data point 13 and 14: add 10 μL of PDZ domain stock (concentration 0.5–2 mM) at each step.

142

Young Joo Sun and Ernesto J. Fuentes PDZ Vstock ( L) 0 3 3 5 5 10 5 5 3 3 5 5 10 10 20 20 20 20 20 20 20 60 60 60 60 60 60 60 60 60

added

Anisotropy

Trials

StdErrAniso (%)

0.0283667495155055 0.0274686828061678 0.029768683937325 0.0286828152753334 0.0289592691855773 0.0279688320612641 0.0302485839847933 0.0282586258847039 0.0288619028287986 0.0306373436606708 0.0338288561137255 0.0348562383472131 0.0373794594208286 0.0405176201406837 0.0473504993123733 0.0502904025786218 0.0522171120090482 0.0532442012233099 0.0540631222118147 0.057123332166628 0.0589739543476869 0.0641761287443581 0.0657427875852492 0.0689649065572496 0.0693739146349071 0.0709325840527179 0.0716128324888053 0.0738748579212796 0.0735578220264764 0.0757814104680961

3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3

4.6773 3.5528 1.898 5.5035 1.7806 4.9895 3.0321 1.3758 1.7269 3.6168 3.1001 1.9562 2.08 1.0867 2.422 1.5704 1.3038 1.3866 3.3913 4.6715 2.042 5.0507 4.1818 3.3367 2.9321 1.7149 0.6701 4.1681 2.4364 1.4858

Fig. 1 Sample data output of a PDZ–ligand PBM binding assay. The “PDZ added Vstock” column shows the volume of stock solution for each data point of the titration. The “Anisotropy” column shows the average anisotropy value measured for each titration data point. The “Trials” column shows the number of measurements used for calculating the average anisotropy. The “StdErrAniso” column shows the standard deviation of the anisotropy

13. Data point 15 to 21: add 20 μL of PDZ domain stock (concentration 0.5–2 mM) at each step. 14. Data point 22 to 29: add 60 μL of PDZ domain stock (concentration 0.5–2 mM) at each step. 15. Each PDZ–PBM binding assay is collected in triplicate (using either biological or technical replicates). Figure 1 shows an example of the output from a typical titration dataset. 3.4

Data Processing

Data processing requires the calculation of PDZ concentration for each titration step, [PDZ]n, where n is the individual data point collected. 1. The PDZ concentration of the first data point is 0: [PDZ]1 ¼ 0.

PDZ Domain Binding Thermodynamics PDZ CASK PDZ

PBM SDC1

143

Dataset # 2

Protein Concentration (µM) 1229 Stock Conc. (µM) 0.00 12.29 12.29 12.29 12.29 12.29 122.9 122.9 1229 1229 1229 1229 1229 1229 1229 1229 1229 1229 1229 1229 1229 1229 1229 1229 1229 1229 1229 1229 1229 1229

Added Vol. (µL) 0 3 3 5 5 10 5 5 3 3 5 5 10 10 20 20 20 20 20 20 20 60 60 60 60 60 60 60 60 60

Total Vol. (µL) 1300 1303 1306 1311 1316 1326 1331 1336 1339 1342 1347 1352 1362 1372 1392 1412 1432 1452 1472 1492 1512 1572 1632 1692 1752 1812 1872 1932 1992 2052

Conc. (µM) 0.00 0.03 0.06 0.10 0.15 0.24 0.70 1.16 3.91 6.65 11.19 15.69 24.60 33.38 50.55 67.25 83.47 99.25 114.60 129.54 144.08 185.49 223.86 259.50 292.70 323.70 352.72 379.93 405.51 429.59

Anisotropy 0.02798 0.03079 0.02941 0.03041 0.02941 0.03037 0.02993 0.02853 0.02696 0.03148 0.03043 0.03221 0.03707 0.03854 0.04198 0.04385 0.04766 0.05292 0.05184 0.05391 0.05279 0.05895 0.05972 0.06396 0.06525 0.06689 0.06948 0.07127 0.07036 0.07239

Base line corr. 0.00000 0.00281 0.00143 0.00243 0.00143 0.00239 0.00195 0.00055 -0.00102 0.00350 0.00245 0.00423 0.00909 0.01056 0.01400 0.01587 0.01969 0.02494 0.02386 0.02593 0.02481 0.03097 0.03174 0.03598 0.03727 0.03891 0.04150 0.04329 0.04238 0.04441

Fig. 2 Processed data of the CASK PDZ–SDC1 PBM binding assay. The concentration of CASK PDZ domain is indicated. The column labeled “Stock Conc.” shows the concentration of the PDZ stock used for the collection of each data point. The column labeled “Conc.” indicates the concentration of PDZ domain sample in the cuvette for each data point. The “Base line corr.” column is the anisotropy value after baseline correction (corrA)

2. The equation for calculating the PDZ concentration at [PDZ]n data point is.


½PDZn ¼ ½PDZn1  tot V n1 þ ½PDZstock  added V stock =tot V n ð1Þ where totVn is the total volume added over n titration steps, tot Vn-1 is total volume of PDZ protein added at the n-1 titration point, and addedVstock is the volume added of stock PDZ

144

Young Joo Sun and Ernesto J. Fuentes

domain ([PDZ]stock) (i.e., totVn ¼ totVn-1 + addedVstock summed over n titration steps). This calculation can be conveniently performed in spreadsheet software (e.g., Microsoft Excel). 3. Baseline correction: The anisotropy value of the first data point (A1 at [PDZ]1 ¼ 0) is subtracted from the anisotropy value of each subsequent data point (An at [PDZ]n) to obtain the corrected anisotropy at each PDZ concentration (corrAn ¼ An - A1). Figure 2 shows an example of processed data computed in a spreadsheet program. 3.5 Data Analysis and Binding Curve Presentation

1. The binding curves are fit to a standard hyperbolic binding model:   B max ½PDZ corr A¼ ð2Þ K d þ ½PDZ where corrA is the corrected anisotropy at each titration step, Bmax is the maximum anisotropy at PDZ domain saturation, Kd is the dissociation constant, and [PDZ] is the total concentration of the PDZ domain in solution. SigmaPlot (Systat Software Inc., CA) was used to determine Bmax and Kd by fitting the binding data to Eq. 2 using nonlinear regression analysis (see Note 12) [6, 15]. The Kd of each PDZ–PBM pair is measured in triplicate and reported as the mean and standard error of the mean. 2. Each data point is normalized to the fitted Bmax for graphical presentation of multiple binding curves in a single plot. Figure 3 shows the presentation of binding curves for CASK PDZ– and Scribble PDZ1–ligand binding reactions. 3. The Gibbs free energy of binding (ΔGb) is calculated by ΔG b ¼ RT∗ lnðKd Þ

ð3Þ

where R is the universal gas constant and T is the given experimental temperature. The error in free energy can be obtained by propagation of the error in Kd.

4

Notes 1. The dansyl fluorophore can affect binding affinity by directly interacting with PDZ domains [7, 8]. For relative affinity measurements this may not be an issue. PDZ domain–fluorophore interactions can be minimized by using other fluorophores or longer peptides [15]. 2. Both the peptide concentration and slit-width can be adjusted to optimize signal-to-noise ratio. However, the concentration

PDZ Domain Binding Thermodynamics

145

A

SGEF NRXN1 SCD1 Caspr4

PDZ [ M]

B PDZ/ligand

Kd ( m)

∆Gb (kcal/mol)

Scribble PDZ1/SGEF

7.3 ± 1.5

-7.01 ± 0.12

CASK PDZ/NRXN1

32.2 ± 4.6

-6.13 ± 0.08

CASK PDZ/SCD1

121 ± 19

-5.34 ± 0.10

CASK PDZ/Caspr4

N. B.

N. B.

Fig. 3 PDZ–PBM binding curves and energetics. (a) Representative binding curves for PDZ–PBM interactions. The Caspr4 binding curve is an example of negative control C-terminal peptide that does not bind (i.e., N.B., no binding) the CASK PDZ domain. The CASK PDZ domain binds C-terminal NRXN1 and SDC1 peptides. The Scribble PDZ1 domain binds an internal peptide derived from SGEF. (b) Dissociation constants and Gibbs free energy of binding for several PDZ–PBM interactions. The reported dissociation constants are the average and standard error derived from at least three independent experiments

of peptide should be kept ~tenfold lower than the Kd to obtain reliable fits of the data. 3. The peptide ligand can be either a C-terminal or internal ligand derived from the target full-length protein. For C-terminal ligands, generally 6–8 terminal residues are used but additional residues may be involved, and this should be determined empirically. Internal PBM sequences can also be used [16]. Again, the exact residues will vary for each interaction and should be determined empirically. Internal ligands also

146

Young Joo Sun and Ernesto J. Fuentes

should have their N-terminus acetylated to mimic an internal protein sequence by avoiding an electrostatic dipole contribution (see [15] for additional details). In addition, the C-terminus should be amidated to neutralize the free carboxylate group. 4. Peptides are typically purified via high-performance liquid chromatography purification using acidic buffers (e.g., trifluoracetic acid) prior to lyophilization. Thus, care should be taken to adjust the pH of the buffer solution upon solubilizing the peptide. We use a benchtop pH meter equipped with microelectrode for small volume samples. 5. The concentration of peptide and protein in solution can be measured by UV absorbance at 280 nm wavelength using a spectrophotometer. The extinction coefficient can be calculated from the amino acid sequence (e.g., ExPASy—ProtParam) [19, 20]. A fluorophore can contribute to the 280 nm wavelength absorbance. However, the excitation and emission wavelengths of the dansyl fluorophore are more than 10 nm away from 280 nm (λex ¼ 340, λem ¼ 550); thus, the fluorophore should not significantly affect the peptide concentration determination at the 280 nm wavelength [15]. If the peptide lacks amino acids with chromophores, one can use the extinction coefficient of dansyl chloride (ε ~4350 M1 l cm1) [21] to estimate the peptide concentration. Alternatively, one can use a color-based protein assay (e.g., bicinchoninic acid or Bradford assay, Thermo Scientific). 6. Dansylated peptide solutions should be kept away from light by either covering them with aluminum foil or using tinted (light free) microfuge tubes. Stock dansylated peptides are generally stored in small aliquots to avoid repetitive freeze and thaw cycles. 7. Using the PDZ protein sample immediately after purification is highly recommended. Long-term storage of proteins at 4
C or at 20
C can significantly affect protein stability. If storage is required, the integrity of the PDZ sample should be tested periodically with a known reference peptide. 8. The Hellmanex™ III (Hellma, NY) cleaning concentrate can be used periodically to remove biological material from the surface of cuvettes followed by thorough rinsing with ddH2O and ethanol using a vacuum cuvette washer. 9. When preparing buffers for sample preparation, filtering and degassing are highly recommended. Particulate matter and air bubbles scatter light and disrupt the spectroscopic measurements. We filter buffers with a 0.45 μm membrane using a vacuum filter unit attached to a dry vacuum system (Welch) followed by continuous stirring under vacuum for 15–30 min.

PDZ Domain Binding Thermodynamics

147

10. To prevent fluorophore bleaching by light exposure, it is highly recommended to cover the cuvette (containing the fluorophore-peptide) in the spectrofluorometer with the lid closed. The sample is now ready for the data acquisition. If necessary, the sample can be equilibrated for several minutes (typically 1–5 min) or until the measured anisotropy value is stable over time. 11. The number of data points can vary depending on the sample and affinity of the PDZ–ligand interaction. Data collection is complete when the anisotropy value plateaus (i.e., three consecutive data points have similar anisotropy values). However, we typically collect two or three additional data points after approaching the anisotropy plateau to ensure binding saturation has been reached. 12. The following assumptions are made when fitting the binding data. First, the PDZ–PBM binding stoichiometry is 1:1, which is true for all known PDZ–PBM interactions. Second, the concentration of free PDZ domain is on the order of the total PDZ domain concentration. Third, there is no significant change in fluorescence intensity (50 mP). 7. For a competitive assay, a direct FP measurement is also required to be analyzed. Furthermore, to avoid experimental

190

Ma´rton A. Simon and La´szlo´ Nyitray

artifacts, the deviation between the experimental windows of the direct and competitive measurements should not exceed 20%. 8. During dialysis samples are diluted to some extent. Never use DTT as a reducing agent during ITC measurement. The amount of samples needed for the measurements depends on the type of instrument. 9. Be aware of avoiding air bubbles in the sample cell and in the syringe. Note that larger volumes of samples are needed to fill the sample cell and the syringe than the actual volumes of the cell and syringe. 10. If the change of heat is small, measure the interaction of interest on a different temperature, as ΔH is temperaturedependent. 11. You should use only the inner 60 cells (minimizing the potential for thermal gradient and also for evaporation at the edges). At least 4–5 parallel measurements is recommended for individual interactions. 12. Pipet FuGene HD reagent directly to the solution, not to the plastic wall of the tube. 13. Regarding the small cell number (20,000/well), try to avoid to disturbing the cell monolayer. Remove the medium slowly with extra care to maintain it. 14. For higher expression level of the protein(s) of interest, incubate the cells for 36–42 h before the measurement. 15. Always prepare Nano-Glo Live Cell reagent freshly. The furimazine component will decay to some extent. Do not freezethaw the Nano-Glo Live Cell Substrate more than 4–5 times. 16. The luminescence signal first increases, reaches a plateau, then begins to decrease. This usually takes 15–30 min, depending on the number of freeze–thaw cycles. 17. Always add water prior to EGF (or any other compound). Be as fast as possible!

Acknowledgments ˝ Go´gl for reading the manuscript. We also We thank Dr. Gergo thank Vikto´ria Bilics for contributing in the FP measurement. This work was supported by the National Research, Development and Innovation Office (NKFIH) grants K119359 (to LN). MAS was supported through the New National Excellence Program of the Hungarian Ministry of Human Capacities. Project no. 20181.2.1-NKP-2018-00005 has been implemented with the support provided from the National Research, Development and

Regulation of RSK1-PDZ Domain Interactions

191

Innovation Fund of Hungary, financed under the 2018-1.2.1-NKP funding scheme. This work was completed in the ELTE Thematic Excellence Programm 2020 Supperted by the Nation Research, Development and Innovation Office (TKP2020-IKA-05) References 1. Loregian A, Palu` G (2005) Disruption of protein-protein interactions: towards new targets for chemotherapy. J Cell Physiol 204:750–762. https://doi.org/10.1002/jcp. 20356 2. Vidal M, Cusick ME, Baraba´si AL (2011) Interactome networks and human disease. Cell 144:986–998. https://doi.org/10. 1016/j.cell.2011.02.016 3. Nishi H, Hashimoto K, Panchenko AR (2011) Phosphorylation in protein-protein binding: effect on stability and function. Structure 19:1807–1815. https://doi.org/10.1016/j. str.2011.09.021 4. Landry CR, Freschi L, Zarin T, Moses AM (2014) Turnover of protein phosphorylation evolving under stabilizing selection. Front Genet 5:1–6. https://doi.org/10.3389/ fgene.2014.00245 5. Van Roey K, Uyar B, Weatheritt RJ, Dinkel H, Seiler M, Budd A, Gibson TJ, Davey NE (2014) Short linear motifs: ubiquitous and functionally diverse protein interaction modules directing cell regulation. Chem Rev 114:6733–6778. https://doi.org/10.1021/ cr400585q 6. Lee HJ, Zheng JJ (2010) PDZ domains and their binding partners: structure, specificity, and modification. Cell Commun Signal 8:1–18. https://doi.org/10.1186/1478811X-8-8 7. Luck K, Charbonnier S, Trave´ G (2012) The emerging contribution of sequence context to the specificity of protein interactions mediated by PDZ domains. FEBS Lett 586:2648–2661. https://doi.org/10.1016/j.febslet.2012.03. 056 8. Nourry C, Grant SGN, Borg JP (2003) PDZ domain proteins: plug and play! Sci STKE 2003:1–13 ´ L, Vada´szi H, 9. Go´gl G, Biri-Kova´cs B, Po´ti A ´ cs A, Szeder B, Bodor A, Schlosser G, A Turia´k L, Buday L, Alexa A, Nyitray L, Reme´nyi A (2018) Dynamic control of RSK complexes by phosphoswitch-based regulation. FEBS J 285:46–71. https://doi.org/10. 1111/febs.14311 10. Sundell GN, Arnold R, Ali M, Naksukpaiboon P, Orts J, Gu¨ntert P, Chi CN,

Ivarsson Y (2018) Proteome-wide analysis of phosphor-regulated PDZ domain interactions. Mol Syst Biol 14:1–22. https://doi.org/10. 15252/msb.20178129 11. Nishi H, Shaytan A, Panchenko AR (2014) Physicochemical mechanisms of protein regulation by phosphorylation. Front Genet 5:1–10. https://doi.org/10.3389/fgene. 2014.00270 12. Pawson T (2004) Specificity in signal transduction: from phosphotyrosine-SH2 domain interactions to complex cellular systems. Cell 116:191–203. https://doi.org/10.1016/ S0092-8674(03)01077-8 13. Pedersen SW, Albertsen L, Moran GE, Levesque B, Pedersen SB, Bartels L, Wapenaar H, Ye F, Zhang M, Bowen ME, Strømgaard K (2017) Site-specific phosphorylation of PSD-95 PDZ domains reveals finetuned regulation of protein-protein interactions. ACS Chem Biol 12:2313–2323. https://doi.org/10.1021/acschembio. 7b00361 14. Go´gl G, Biri-Kova´cs B, Durbesson F, Jane P, Nomine Y, Kostmann C, Bilics V, Simon M, Reme´nyi A, Vincentelli R, Trave G, Nyitray L (2019) Rewiring of RSK–PDZ interactome by linear motif phosphorylation. J Mol Biol 431:1234–1249. https://doi.org/10.1016/j. jmb.2019.01.038 15. Vincentelli R, Luck K, Poirson J et al (2015) Quantifying domain-ligand affinities and specificities by high-throughput holdup assay. Nat Methods 12:787–793. https://doi.org/10. 1038/nmeth.3438 16. Mortier E, Wuytens G, Leenaerts I, Hannes F, Heung MY, Degeest G, David G, Zimmermann P (2005) Nuclear speckles and nucleoli targeting by PIP2-PDZ domain interactions. EMBO J 24:2556–2565. https://doi.org/10. 1038/sj.emboj.7600722 17. Gianni S, Walma T, Arcovito A, Calosci N, Bellelli A, Engstro¨m A, Travaglini-Allocatelli C, Brunori M, Jemth P, Vuister GW (2006) Demonstration of long-range interactions in a PDZ domain by NMR, kinetics, and protein engineering. Structure 14:1801–1809. https:// doi.org/10.1016/j.str.2006.10.010

192

Ma´rton A. Simon and La´szlo´ Nyitray

18. Thomas GM, Rumbaugh GR, Harrar DB, Huganir RL (2005) Ribosomal S6 kinase 2 interacts with and phosphorylates PDZ domain-containing proteins and regulates AMPA receptor transmission. Proc Natl Acad Sci U S A 102:15006–15011. https://doi. org/10.1073/pnas.0507476102 19. Shi GX, Yang WS, Jin L et al (2017) RSK2 drives cell motility by serine phosphorylation of LARG and activation of rho GTPases. Proc Natl Acad Sci U S A 115:E190–E199. https:// doi.org/10.1073/pnas.1708584115 20. Park S-H, Raines RT (2015) Fluorescence polarization assay to quantify protein-protein interactions. Methods Mol Biol 261:161–165. https://doi.org/10.1385/159259-762-9:161 21. Roehrl MHA, Wang JY, Wagner G (2004) Discovery of small-molecule inhibitors of the NFAT-calcineurin interaction by competitive high-throughput fluorescence polarization screening. Biochemistry 43:16067–16075. https://doi.org/10.1021/bi048232o 22. Hall MD, Yasgar A, Peryea T et al (2016) Fluorescence polarization assays in highthroughput screening and drug discovery: a review. Methods Appl Fluoresc 4:022001.

https://doi.org/10.1088/2050-6120/4/2/ 022001 23. Roehrl MHA, Wang JY, Wagner G (2004) A general framework for development and data analysis of competitive high-throughput screens for small-molecule inhibitors of protein-protein interactions by fluorescence polarization. Biochemistry 43:16056–16066. https://doi.org/10.1021/bi048233g 24. Macek P, Cliff MJ, Embrey KJ et al (2018) Myc phosphorylation in its basic helix-loop-helix region destabilizes transient-helical structures, disrupting max and DNA binding. J Biol Chem 293:9301–9310. https://doi.org/10.1074/ jbc.RA118.002709 25. Dixon AS, Schwinn MK, Hall MP et al (2016) NanoLuc complementation reporter optimized for accurate measurement of protein interactions in cells. ACS Chem Biol 11:400–408. https://doi.org/10.1021/ acschembio.5b00753 26. Simon M, Go´gl G, Ecse´di P, et al (2019) High throughput competitive fluorescence polarization assay reveals functional redundancy in the S100 protein family. bioRxiv. https://doi.org/ 10.1101/718155

Chapter 12 Chemical Synthesis of PDZ Domains Christin Kossmann, Sana Ma, Louise S. Clemmensen, and Kristian Strømgaard Abstract Developments in chemical protein synthesis have enabled the generation of tailor-made proteins including incorporation of many types of modifications into proteins, enhancing our ability to control site-specificity of protein posttranslational modifications (PTMs), modify protein backbones and introduce photocrosslinking probes. For PDZ (postsynaptic density protein, disks large, zonula occludens) protein domains, expressed protein ligation (EPL) has been employed to introduce analogs of cognate amino acids, amideto-ester bond mutations, and phosphorylations in the study of PDZ domain-mediated protein-protein interactions (PPIs). Here, we present protocols for EPL of PDZ domains focusing on phosphorylation and amide-to-ester modifications in the PDZ domain proteins. Key words Expressed protein ligation, Protein modification, Phosphorylation, Amide-to-ester mutation, Solid-phase peptide synthesis, Native chemical ligation

1

Introduction The introduction of chemical modifications into proteins has broad applicability for not only biotechnological and biomedical purposes but also in studies of fundamental biological functions such as cell signaling, protein-protein interactions (PPIs), and posttranslational modifications (PTMs). Chemical synthesis of proteins via solidphase peptide synthesis (SPPS) enables veritably any modification at any desired position [1, 2] and allows both the introduction of noncanonical amino acids as well as peptide or protein backbone modifications. The seemingly limitless scope of chemical space that can be leveraged in SPPS is a major advantage over recombinant methods to introduce protein modifications. However, the constraints in the coupling efficiency of longer peptide chains severely restrict the size of peptides or proteins that can be generated by SPPS. Undeniably, improvements in 9-fluorenylmethoxycarbonyl (Fmoc)/tert-butyl (t-Bu)-SPPS and tert-butyloxycarbonyl (Boc)/ benzyl (Bzl)-SPPS, such as building blocks to minimize

Jean-Paul Borg (ed.), PDZ Mediated Interactions: Methods and Protocols, Methods in Molecular Biology, vol. 2256, https://doi.org/10.1007/978-1-0716-1166-1_12, © Springer Science+Business Media, LLC, part of Springer Nature 2021

193

194

Christin Kossmann et al.

aggregation, high-swelling resins, technological advancements in automated SPPS, and increased purity of the protected amino acids have advanced the previously set limit for peptides of 50–70 amino acids, depending on the sequence [3–6]. Still, the synthesis of proteins, such as PDZ domains that comprise typically 80–100 amino acids, is generally not feasible by SPPS. In conjunction with SPPS, the development of native chemical ligation (NCL), enabled by the pioneering work of Kent and colleagues in 1994, expanded the range of protein size that can be synthesized. This reaction involves a chemoselective conjugation of one peptide bearing a C-terminal α-thioester with a second peptide bearing an N-terminal Cys, resulting in a native peptide bond [7]. Proceeding initially through transthioesterification driven by nucleophilic attack of the N-terminal Cys thiol moiety to the carbonyl carbon of the thioester, a new intermolecular thioester intermediate is formed. Subsequently, this thioester intermediate rearranges through an S-to-N acyl shift by the α-amine of the Cys. This leads to the formation of a native amide bond (see Fig. 1) [8, 9]. For the synthesis of larger proteins, ligation of several peptide fragments is possible, either by stepwise ligation from the N- to the C-terminus or by one-pot ligations with Cys protection groups [10–13]. By applying NCL, several small proteins have been successfully synthesized, such as the HIV-1 protease covalent dimer, encompassing 203 amino acids, generated from four synthetic peptide fragments [14]. However, the synthesis of larger proteins (>20 kDa) by NCL remains challenging, because the multiple ligation steps impact the overall yield, and certainly cannot match the yields of recombinant systems. Thus, to address these concerns, expressed protein ligation (EPL), a derivative of NCL, was developed to leverage the advantages of both recombinant protein expression and SPPS. EPL involves a reaction between a recombinant protein fragment with a synthetic peptide fragment containing the desired chemical modification [8, 15]. Multistep ligations involving the joining of more than two fragments can be performed, but the reaction always requires a combination of synthetic and recombinant components. Generally, there are two types of EPL strategies that can be used for targeting the insertion of modifications at different positions of the protein: (1) modifications in the N-terminal region, where a synthetic peptide is ligated to a recombinant fragment with an N-terminal Cys, and (2) modifications in the C-terminal region where an N-terminal recombinant thioester protein is joined with a synthetic peptide. In the first strategy, a recombinant protein fragment with an N-terminal Cys residue is generated typically by introduction of a cleavage site, such as a protease recognition site, immediately adjacent to the N-terminal Cys. Then, proteases such as Factor Xa or enterokinase cleave and reveal the N-terminal Cys residue (see Fig. 2). The peptide thioester can be generated by Fmoc- as well

Synthetic PDZ Domains

195

Fig. 1 Mechanism of NCL. Initially, the C-terminal thioester of one peptide fragment undergoes a transthioesterification with the N-terminal Cys of a second peptide fragment. Secondly, an S-to-N acyl shift leads to the formation of a native peptide bond between the two fragments

Fig. 2 Expressed protein ligation. Left: Strategy for N-terminal ligation. The N-terminal peptide fragment is synthesized as a thioester and ligated to a recombinantly expressed protein, which is prior cleaved by Factor Xa to result in an unprotected cysteine. Right: Strategy for C-terminal modification. The C-terminal peptide fragment is synthesized and ligated to a recombinantly expressed protein which was designed with an intein (fused to a chitin-binding domain (CBD)) to obtain a thioester after MESNa treatment

196

Christin Kossmann et al.

Fig. 3 Strategies for the generation of peptide thioesters. (a) A thiol-linked resin is used for Boc-SPPS, which reveals the functional thioester after HF cleavage. (b) In Fmoc-SPPS hydrazine resins can be used. The peptide is elongated on the hydrazine linked resin. After TFA cleavage, a C-terminal hydrazide is generated, which can be oxidized to an azide. By adding thiol additives (e.g., MPAA, MESNa, TFET), thiolysis of the azide leads to the functional thioester

as Boc-synthesis. For Boc-synthesis, a resin with a thioester linker is used, which yields a functional thioester after the hydrofluoric acid (HF) cleavage. Due to instability of the thioester moiety in basic conditions during the deprotection steps in Fmoc-synthesis, peptide thioesters need to be generated only after peptide chain elongation. The most common methods employ the use of N-acylurea [16] or hydrazine linkers [17], both resulting in a thioester moiety after activation of the trifluoroacetic acid (TFA)-cleaved peptide fragment (see Fig. 3). For the second strategy that generates C-terminally modified proteins, a method was established in 1998 by Muir et al. that adapted biologically occurring protein splicing reactions to create recombinant proteins with a reactive α-thioester moiety [18]. Critical residues that are nucleophiles of the C-extein are mutated to unreactive amino acids such as Ala. This arrests the reaction at the initial thioester formation step upon intein cleavage with the addition of a nucleophilic thiol, such as the sodium salt of mercaptoethane sulfonate (MESNa) [8, 18]. This protein thioester is then isolated and ligated to a synthetic peptide with an N-terminal Cys, which is typically synthesized using Fmoc-based SPPS. With these EPL strategies, larger protein fragments can be generated efficiently through recombinant techniques while allowing the introduction of chemical modifications through the synthetic peptide with site-specificity. There are several considerations when planning protein semisynthesis by EPL strategies. Firstly, semisynthetic proteins have to be refolded after ligation as EPL is typically conducted under denaturing conditions, which can be a challenge. Another consideration is the length of the synthetic peptide(s) and EPL is best applied to the generation of proteins with modifications in the Nor C-terminal regions, to avoid ligation of multiple fragments. The generation of thioesters is also only limited to certain residues because some residues are either incompatible (Asp, Glu, Asn, Pro, and Gln) or result in unfavorable EPL reaction kinetics (Ile, Lys, Leu, Thr, and Val). Additionally, the requirement of a Cys

Synthetic PDZ Domains

197

residue at the ligation junction can be a limitation if an appropriate ligation site with a native Cys residue cannot be identified. In this case, a Cys mutation has to be introduced at the desired ligation site, which may not be possible due to a consequently building of sulfide bridges/heteromerization and eventually loss of protein function. However, a number of approaches involving ligation auxiliaries, Cys surrogates, and desulfurization methods, have been developed to allow more flexibility in the selection of ligation sites besides Cys [11, 19–23]. With these improvements, EPL has become applicable for the generation of a wider spectrum of proteins. Several types of chemical modifications have been incorporated into PDZ domains using both strategies of EPL to investigate sidechain interactions [24], backbone interactions [25], and phosphorylations [26]. Studies of backbone H-bond interactions are especially crucial for PDZ binding because the binding of C-terminal peptide ligands to the PDZ protein domain is facilitated by a conserved carboxylate-binding site via backbone hydrogen amide bonds [25]. This was deciphered by the introduction of amide-toester mutations to these and other regions of PDZ domains through EPL. Peptide fragments containing backbone amide-toester mutations (depsipeptides) were ligated to a recombinantly expressed protein fragment, enabling the investigation of specific backbone hydrogen bonds in the PDZ domain protein [25]. EPL also enabled the generation of semisynthetic phosphorylated PDZ domains, specifically of PDZ domains of the postsynaptic density protein 95 (PSD-95), which allowed its phosphoregulated PPIs to be studied [26, 27]. A riveting feature of this work is the direct comparison of the effects of introducing phosphomimetics versus a semisynthetic protein with a ‘true’ phosphorylation. Although phosphomimetics, Glu and Asp, are used often to study this important PTM, structural and chemical differences of the mimetics could render them as inadequate substitutions, especially in cases involving phosphotyrosine. In contrast, in vitro phosphorylation uses kinases to install phosphate-groups onto proteins, but this would not be feasible in certain cases where the enzyme is unknown, or site-specificity is desired, but the enzyme modifies multiple sites in the same protein. Recently, an alternative method leverages a genetic code expansion technique to incorporate more than 200 nonproteinogenic amino acids, including phosphorylated amino acids [28]. However, an enigmatic challenge with this technology is that not all protein sites are amenable to modifications, and the reason for this is still in contention. In these situations, EPL is an attractive option to introduce phosphorylation site-specifically to proteins in a reliable manner. Herein, we provide protocols that have successfully generated PDZ domains with modifications in either the N- or C-terminal regions. Specifically, synthesis of peptide thioesters by Boc- and

198

Christin Kossmann et al.

Fmoc-SPPS via the hydrazine method are described along with the expression of C-terminal thioesters via intein fusion proteins and N-terminal Cys containing proteins. PDZ domains inconveniently lack a Cys residue where we desire to place a ligation junction, but we circumvent this by employing radical desulfurization, which will also be described in these protocols.

2

Materials

2.1 Plasmid Construction

1. Isopropyl β-D-1-thiogalactopyranoside (IPTG) inducible bacterial expression vector with ampicillin resistance that encodes the codon-optimized DNA construct of the target protein fused to a polyhistidine purification tag, for example pRSET (PDZ) (see Note 1). 2. Intein-encoding plasmid with ampicillin resistance for IPTG inducible bacterial expression, for example pTWIN1. 3. 10 μM oligonucleotide primer stock solutions in nuclease-free water (see Table 1). 4. Site-directed mutagenesis kit based on a high-fidelity DNA polymerase. 5. Nuclease-free water. 6. Chemically competent E. coli cloning strain (TOP10). 7. S.O.C medium. 8. DNA miniprep kit. 9. 50 mg/mL ampicillin stock solution in ultrapure water. 10. Sterile Luria Broth (LB) agar plates: mix LB agar into ultrapure water. Autoclave at 121
C for 30 min and cool down to 50
C before adding ampicillin (100 μg/mL) and distributing on 92 mm x 16 mm sterile clear petri dishes. Store plates at 4
C. 11. Sterile Luria Broth (LB) media. 12. High-fidelity DNA polymerase and appropriate buffers. 13. NdeI and SapI restriction enzymes. 14. TAE buffer: 1 M ethylenediamine tetraacetic acid, 0.04 M Trisbase, 0.01 M acetic acid. 15. 1.5% agarose gel, prepared by melting solid agarose (ultrapure agarose) suspended in TAE buffer using a microwave oven. 16. DNA gel purification kit. 17. T4 DNA ligase and appropriate buffers.

2.2 Protein Expression

1. Chemically competent E. coli expression strain (BL21 dE3 pLys). 2. Sterile LB agar plates (see Subheading 2.1).

Synthetic PDZ Domains

199

Table 1 Primers used for the construction of plasmids encoding recombinant EPL fragments Forward 1. Primers for introduction of factor Xa site placed adjacent to the N-terminal Cysa,b

0

5 -(NNN)5 ATC GAG GGA AGG TGC (NNN)5-3´

2. Primers for PDZ cloning into pTWIN1c,d 50 -GGA ATT CAT ATG (NNN)10-3´

Reverse 50 -(NNN)5 GCA CCT TCC CTC GAT (NNN)5-3´ 50 -GGT GGT TGC TCT TCC (NNN)6-3´

a

Bold letters highlight the gene encoding factor Xa site and a Cys residue (NNN)x corresponds to flanking codons from the target gene, with x being the number of codons c NdeI site is underlined in the sequence d SapI site is italicized in the sequence b

3. Sterilized Luria Broth (pH 7.2): dissolve premixed LB broth in 1 L of ultrapure water according to manufacturer’s instructions and autoclave at 121
C for 30 min. 4. Sterilized 300 mL and 5 L culture flasks, autoclaved at 121
C for 30 min. 5. 50 mg/mL ampicillin stock solution in ultrapure water. 6. 1 M IPTG stock solution in ultrapure water. 2.3 Protein Purification

1. Elution buffer (50 mL): 50 mM sodium phosphate pH 7.4, 250 mM imidazole. 2. HisTrap Fast Flow (FF) columns, 5 mL. 3. Lysis Buffer (50 mL): 50 mM sodium phosphate pH 7.4, 10 mM MgCl2, 25 μg/mL DNase, complete protease inhibitor (1 tablet/50 mL). Add DNase and protease inhibitor right before use. 4. Wash buffer (200 mL): 50 mM sodium phosphate pH 7.4, 25 mM imidazole.

2.4 Thioester Generation

1. Thiolysis buffer: 1–2 M urea, 50 mM sodium phosphate pH 6.8, 150 mM NaCl. 2. Dialysis membrane or device, with molecular weight cutoff (MWCO) of 10 kDa. 3. Sodium 2-mercaptoethanesulfonate (MESNa).

2.5 Factor Xa Cleavage

1. Cleavage buffer: 50 mM Tris–HCl pH 8, 100 mM NaCl, 6 mM calcium chloride. 2. Factor Xa protease. 3. Dialysis membrane or device, with a MWCO of 10 kDa.

200

Christin Kossmann et al.

2.6 Solid-Phase Peptide Synthesis

1. 2-(tritylmercapto)acetyl-L-leucinyl-PAM (Trt-S-Ac-L-Leu-PAM) resin for Boc-based synthesis of peptide thioesters. 2. 2-clorotrityl chloride (2-CTC) resin on polystyrene for Fmocbased chemical synthesis of peptide hydrazides. 3. Preloaded Wang resin for Fmoc-based synthesis. 4. Fmoc-PAL-PEG-PS™ for Fmoc-based chemical synthesis of long peptides. 5. Phenylacetamidomethyl (PAM) resin for Boc-based synthesis. 6. Fmoc/Boc-protected canonical amino acids. 7. Fmoc-Ser(PO(OBzl)OH)-OH, Fmoc-Thr(PO(OBzl)OH)OH, Fmoc-Tyr(PO(OBzl)OH)-OH for phosphopeptides. 8. α-hydroxy amino acids (see Note 2) for depsipeptides. 9. 2-(1-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (HBTU) prepared as a 0.5 M stock solution in DMF in a light-protected bottle. 10. Diisopropylamine (DIEA). 11. 20% (v/v) piperidine in DMF. 12. Fmoc-cleavage mix: 82.5% (v/v) TFA, 5% (v/v) phenol, 5% (v/v) H2O, 5% (v/v) thioanisole, 2.5% (v/v) 1,2-ethanedithiol (EDT). Prepare fresh. 13. Peptide dissolution buffer: 20% MeCN, 0.1% TFA. 14. Ninhydrin test/Kaiser test kit. 15. Disposable Reaction Tube (10 mL) (SiliCycle Inc.).

2.7 Peptide Purification

1. High-pressure liquid chromatography (HPLC) buffer A: 5% (v/v) acetonitrile (MeCN), 0.1% (v/v) TFA in ultrapure water, filtered with 0.2 μm filters. 2. HPLC Buffer B: 95% MeCN, 0.1% TFA in ultrapure water, filtered with 0.2 μm filters.

2.8 Expressed Protein Ligation

1. NaNO2 stock solution: 0.2 M in ultrapure water. Prepare fresh. 2. Dithiothreitol (DTT) stock solution: 1 M in ultrapure water. Can be stored at 20
C. 3. 300 mM Tris(2-carboxyethyl)phosphine (TCEP) stock solution in ultrapure water.

hydrochloride

4. Ligation buffer 1: 0.2 M sodium phosphate pH 3.0, 6 M guanidinium hydrochloride (Gu·HCl). 5. Ligation buffer 2: 0.2 M 4-mercaptophenylacetic acid (MPAA) in 6 M Gu·HCl, 0.2 M sodium phosphate pH 7.0. Prepare fresh.

Synthetic PDZ Domains

2.9

Desulfurization

201

1. Solubilizing buffer: 0.2 M sodium phosphate buffer pH 7, 6 M Gu·HCl. 2. Radical initiator VA-044. 3. Reduced glutathione. 4. 0.5 M TCEP stock solution in 500 μl 6 M Gu·HCl, 0.2 M sodium phosphate pH 7.

2.10

Equipment

1. Mass Spectrometer with electron spray ionization (ESI) coupled to an HPLC system (ESI-LC/MS), equipped with a C18 reversed-phase column for the analysis of peptides and a C8 reversed-phase column for the characterization of proteins, using a linear gradient of a binary buffer system containing H2O/MeCN/TFA (A: 95/5/0.1; B: 5/95/0.1) at a flow rate of 1 mL/min for proteins and H2O/MeCN/formic acid (A: 95/5/0.1; B: 5/95/0.1) at a flow rate of 1 mL/min for peptides. 2. Fast protein liquid chromatography (FPLC) system, equipped with a gel filtration column for size-exclusion chromatography (SEC), operating at a flow rate of 1 mL/min. 3. Cell disruptor system. 4. 4–15 L Freeze Dryer. 5. MiniBlock Shaker. 6. Benchtop pH Meter. 7. Analytical reversed-phase ultra-high performance liquid chromatography (RP-UPLC) system with a C18 reversed-phase column for the analysis of peptides and a C8 reversed-phase column for the analysis of proteins, using a binary buffer system consisting of H2O/MeCN/TFA (A: 95/5/0.1; B: 5/95/0.1) at 0.45 mL/min. 8. Preparative reversed-phase high-performance liquid chromatography (RP-HPLC) system with a C18 reversed-phase column for the purification of peptides and a reversed-phase C4 column for the purification of proteins, using a linear gradient of a binary solvent system of H2O/MeCN/TFA (A: 95/5/ 0.1; B: 5/95/0.1) at a flow rate of 20 mL/min.

3

Methods

3.1 Plasmid Construction for Expression of Recombinant Fragments

Prior to moving forward with the protocols below, a bacterial expression vector with a gene encoding the protein-of-interest has to be procured or cloned. Our PDZ domains are encoded in a pRSET vector, so the following steps will refer to this plasmid, pRSET(PDZ), but other constructs can be used as well.

202

Christin Kossmann et al.

3.1.1 Cloning of Plasmid Encoding C-Terminal Fragment with N-Terminal Cys

For the generation of recombinant C-terminal fragment, the protein is expressed with a hexa-histidine purification tag and a recognition site for factor Xa, which upon cleavage after protein isolation reveals the N-terminal Cys. The following protocol details the cloning strategy, utilizing site-directed mutagenesis, to obtain the plasmid encoding the C-fragment, pRSET(ΔCPDZ), starting with the construct that encodes the full-length protein, pRSET(PDZ). 1. In a PCR tube, combine factor Xa primer pairs from Table 1, the plasmid encoding full-length protein pRSET(PDZ), and appropriate reagents, including a high-fidelity DNA polymerase from a site-directed mutagenesis kit. 2. Load PCR samples onto a thermocycler and execute temperature cycles according to manufacturer’s instructions. 3. To digest the DNA template, add 1 μL of DpnI restriction enzyme to PCR mixture and let incubate at 37
C overnight, or at least 8 h. 4. Transform 1 μL of PCR digest into 10 μL chemically competent TOP10 E. coli cells. After a heat-shock treatment of 45 s at 42
C followed by 2 min incubation on ice, grow cells in 250 μL of S.O.C medium at 37
C and 200 rpm for at least 45 min. 5. Transfer cells to sterile LB-ampicillin agar plates and incubate overnight at 37
C. 6. Inoculate one colony into 5 mL of sterile LB-ampicillin media and incubate overnight at 37
C and 200 rpm. 7. Isolate plasmid with DNA miniprep kit using manufacturer’s instructions. 8. Confirm plasmid sequence with DNA sequencing to verify that the mutations were inserted correctly into the plasmid and store DNA at 20
C (see Note 3).

3.1.2 Gene Insertion into an Intein-Encoding Plasmid for the Generation of Protein Thioesters

Below, is a protocol to prepare pTWIN1-His7(ΔNPDZ) by cloning the PDZ domain fragment into a Mxe GyrA-encoding pTWIN1His7 vector, which has been modified to contain a heptahistidine tag. The original pTWIN1 vector with a chitin-binding domain (CBD) that leverages the IMPACT protein purification system can be used as well. 1. Using primer pairs from Table 1, amplify the gene encoding the PDZ domain fragment with a high-fidelity DNA polymerase, following manufacturer’s instructions for reagent amounts and thermal cycles. 2. Digest PCR fragment and pTWIN1-His7 vector with 1 μL each of NdeI and SapI restriction enzymes at 37
C for 2 h. Isolate digestion products with DNA gel electrophoresis. Extract fragment from agarose gel with DNA gel purification kit.

Synthetic PDZ Domains

203

3. Ligate the PCR fragment into pTWIN1 vector with T4 DNA ligase according to the manufacturer’s instructions. This forms the plasmid encoding the PDZ domain fragment fused to the intein, pTWIN1-His7(ΔNPDZ). 4. Transform 1 μL of ligation mixture into 10 μL of chemically competent E. coli cloning strain, TOP10. After growing the culture in S.O.C. media for 45 min, apply 50 μL of cells onto an LB agar plate with 100 μg/mL ampicillin. Allow colonies to form overnight in an incubator at 37
C. 5. Prepare 5 mL of overnight culture in LB medium with 100 μg/mL ampicillin by inoculating one colony from the LB-ampicillin plate. Allow the culture to grow overnight in a shaking incubator at 37
C at 200 rpm. 6. Isolate plasmid, pTWIN1-His7(ΔNPDZ), from overnight culture with a DNA miniprep kit, according to manufacturer’s instructions. Verify plasmid with DNA sequencing and store at 20
C. 3.2 Protein Expression

The following steps can be applied to the expression of both the N- and C-terminal recombinant fragments. 1. Transform 1 μL DNA (obtained in Subheading 3.1) into 10 μL of chemically competent E. coli expression strain, BL21 dE3 pLys. After growing in 100 μL of S.O.C. media for 45 min at 37
C and shaking at 200 rpm, spread 10 μL of cells onto an LB agar plate containing 100 μg/mL ampicillin. Allow colonies to form overnight in an incubator at 37
C. 2. Prepare the starting culture by inoculating one colony from the LB-ampicillin plate into a 300-mL flask containing 100 mL of sterile LB broth and allow the culture to grow overnight in a shaking incubator at 37
C and shaking at 200 rpm. 3. Measure the optical density of the starting culture at the wavelength of 600 nm (OD600) with a UV spectrophotometer. Add starting culture to prewarmed 5 L flasks containing 1 L LB broth with 100 μg/mL ampicillin to target a starting OD600 value of 0.1. 4. Allow culture to grow in a shaking incubator at 37
C and 200 rpm until the OD600 value reaches between 0.4 and 0.8. Induce expression by adding IPTG stock to target a final concentration of 1 mM to the culture. Then allow culture to grow for another 4 h at 37
C and shaking at 200 rpm (see Note 4). 5. Harvest cells by centrifuging the culture at 10,000  g for 10 min at 4
C. Discard supernatant and collect pellet. Store cell pellets at 20
C for later processing or proceed directly into the next protein purification steps.

204

Christin Kossmann et al.

3.3 Protein Purification from Bacterial Lysates 3.3.1 Cell Lysis

The protocols below are applicable to the purification of N- and C-terminal recombinant fragments.

1. Thaw cell pellet on ice and add 5–10 mL ice-cold lysis buffer per g pellet to the thawed cell pellet. 2. Solubilize the pellet until a homogeneous solution is reached. A magnetic stirrer may be used for effective solubilization. 3. Lyse cells by using a cell disrupter at 4
C, applying a pressure of 26 kpsi, and repeat this step once. 4. Pellet cell debris by centrifugation for 30 min, at 4
C and 30,000  g (see Note 5).

3.3.2 His-Trap Purification

1. Filter cell lysate using a syringe and a 0.45 μm syringe filter. Subsequently, dilute it 2/1 in wash buffer. Maintain the lysate sample at 0–4
C during the entire purification process by performing the next steps in a refrigerator and keeping samples on ice. 2. Measure the pH of the sample and adjust it to pH 7.4 with 1 M HCl or NaOH (see Note 6). 3. Prepare a 5-mL His-trap FF column by washing it with 3 column volumes (CV) deionized water and equilibrating it with 5 CV wash buffer. This can be performed with either a peristaltic pump with a flow rate of 3.5 mL/min or manually with a syringe. 4. Load the filtered and diluted lysate sample onto the column and remove nonbinding residues by washing it with 20 CV wash buffer. 5. Elute the His-tag protein with 5 CV elution buffer and collect the flow-through in 1.5 mL fractions in 2 mL reaction tubes. 6. Measure the protein concentration of each fraction on a spectrophotometer at 280 nm. Then verify purity and molecular weight of the protein with UPLC and LCMS, respectively. Pool pure fractions to target a purity of >85–90% from UPLC (see Note 7). 7. Concentrate the combined sample using centrifugal filters with a MWCO below the molecular weight of the respective protein.

3.3.3 FPLC Purification

If the protein sample does not meet target purity of >85–90% after the His-trap purification, a size-exclusion (SEC) purification step can be performed using an FPLC. A flow rate of 1 mL/min at 4
C is recommended. 1. Equilibrate the column with 5 CV wash buffer.

Synthetic PDZ Domains

205

2. Load the sample onto the column using a syringe and wash the column with 8–10 CV of wash buffer. 3. Elute the protein with 5 CV wash buffer and collect 1.5 mL fractions in 2 mL reaction tubes. 4. Measure the protein concentration of each fraction on a spectrophotometer at 280 nm. Then verify purity and molecular weight of the protein with UPLC and LCMS, respectively. Pool pure fractions to target a purity of >85–90% from UPLC. 5. Concentrate the combined sample using Amicon® Ultra Centrifugal Filters with the suitable MWCO and measure the final sample on LC-MS and UPLC. 3.4 Protein Thioester ΔNPDZ Formation and Isolation

1. Dialyze recombinant protein from pTWIN1-His7(ΔNPDZ) against thiolysis buffer in a 1 L beaker. Replace dialysis buffer with fresh thiolysis buffer every 2 h for at least 2 times. After the last buffer exchange, allow the sample to dialyze overnight, or about 8–12 h. 2. Transfer dialyzed protein sample into a 50 mL-conical tube and add MESNa powder directly into the tube to target 100 mM concentration. Adjust pH to 6.8 with 1 M HCl or NaOH, if necessary. 3. Let thiolysis reaction proceed to form the protein thioester on a rocking table at 4
C, for 0.5–3 days, as reaction time is highly dependent on the intein and the amino acid at the cleavage junction (see Note 8). Monitor reaction by removing 10 μL aliquots every hour and analyze on LC-MS. Meanwhile, equilibrate a 5-mL His-trap FF column according to Subheading 3.3.2, step 3 with thiolysis buffer in preparation for protein thioester purification. 4. Purify reaction mixture immediately upon completion (see Note 9), by flowing the reaction mix through a His-trap FF column, which sequesters the cleaved, histidine-tagged intein and any uncleaved fusion protein. Collect eluent, which contains the protein thioester. Continue to elute with 5 CV thiolysis buffer and collect 1 CV per fraction. 5. Analyze fractions with LC-MS and pool the cleanest fractions. If the final sample is below the target purity of 80% after this step, the protein thioester can be further purified using SEC (see Subheading 3.3.3). 6. Dialyze protein thioester against ultrapure water to remove salts and buffer components. Freeze dry and lyophilize sample to produce white solids. Characterize the final product with UPLC and LC-MS. Store at 20
C until they are ready to be used for processing.

206

Christin Kossmann et al.

3.5 Factor Xa Cleavage to Generate N-Terminal Cys Protein Fragment ΔCPDZ

1. Dialyze recombinant protein from pRSET(ΔCPDZ) against 1 L of factor Xa buffer at 4
C. Exchange dialysis buffer with fresh buffer every 2 h for 3 times. Allow sample to dialyze overnight after the last buffer exchange. 2. Transfer protein sample to a fresh 50 mL-conical tube. Add factor Xa according to manufacturer’s instructions and allow enzymatic cleavage to progress at 4
C on a rotating or rocking table for 12–24 h. Remove 10 μL aliquots every hour and monitor reaction with LC-MS (see Note 10). 3. Prepare a 5-mL His-trap FF column for purification by following Subheading 3.3.2, step 3 with cleavage buffer. Purify cleavage mixture immediately after reaction is complete to avoid nonspecific cleavage, using steps 4 and 5 in Subheading 3.4 with factor Xa cleavage buffer as the eluent. 4. Dialyze the final protein fragment against ultrapure water to remove salt and buffer components, and lyophilize protein to form white amorphous solids. Analyze compound with UPLC and LC-MS, and store at 20
C until they are ready to be used for processing.

3.6 Solid-Phase Peptide Synthesis

SPPS can be performed with two protection techniques: Boc/Bzl or Fmoc/t-Bu. The latter has the advantage of an orthogonal protection system, which allows milder conditions and a broader range in the pH-dependent reaction conditions for selective cleavage of side-chain protection groups and the Fmoc-group [3]. The Fmoc-SPPS method is preferred over Boc-SPPS in regard to phosphorylated peptides because phosphate groups are not stable during HF cleavage. Several automated synthesizers are available, enabling fast and convenient SPPS. For the inclusion of expensive building blocks, such as phospho-amino acids, manual synthesis is often preferred, because less material is needed and easy direct monitoring of the reaction efficiency can be performed via Kaiser test or test cleavage (see Note 11). To insert amide-to ester mutations, Boc-SPPS is the method of choice over Fmoc-SPPS, because the basic Nα-deprotection step in Fmoc-SPPS can lead to hydrolysis of the ester group. Different resins are used for the synthesis of N- and C-terminal fragments. C-terminal fragments do not need special requirements for the resin and can be generated on commercially available resins (see Note 12). For N-terminal fragments, the generation of a C-terminal thioester is essential for the ligation reaction. Therefore, functional resins are used, for example, comprising a hydrazine linker in Fmoc-SPPS or thioester generating resins in Boc-SPPS.

3.7 Fmoc/t-Bu-SPPS for the Synthesis of Phosphopeptides

The protocol below, for Fmoc-SPPS is suitable for the synthesis of peptides with canonical amino acids and for inserting modifications like glycosylations and phosphorylations. We provide a standard protocol for Fmoc-SPPS, including the example of the introduction of phospho-amino acid building blocks.

Synthetic PDZ Domains

207

Fig. 4 Fmoc-hydrazine coupling to 2-CTC resin. (a) Structural formula and reaction conditions, (b) Experimental setup 3.7.1 Resin Loading

For synthetic N-terminal peptide fragments, a functional hydrazine linker is essential for the ligation reaction. Fmoc-hydrazono-pyruvyl-aminomethyl polystyrene resins are commercially available, with the first amino acid precoupled for all standard L-amino acids. If a noncanonical first amino acid is needed, manual preloading of 2-chlorotrityl chloride (2-CTC) resin applying the below protocol can be performed (see Fig. 4). 1. Swell 2-CTC resin (2 g) in DCM (16 mL) at 0
C in a round bottom flask in an ice-bath for at least 30 min. 2. Dissolve Fmoc-hydrazine hydrochloride (4 eq) and DIEA (10 eq) in DMF (20 mL) and CH2Cl2 (4 mL). 3. Add the solution from step 2 dropwise to the resin slurry at 0
C and stir continuously and gently overnight from 0
C to room temperature (RT). 4. Add 80 eq methanol (0.32 mL) to quench the nonreacted 2-CTC resin. Subsequently, wash the resin three times with each: 5 mL DMF, 5 mL H2O, 5 mL methanol, 5 mL ethyl ether and keep under high vacuum for at least 1 h for complete drying and store at 4
C [17]. 5. Test the resin loading via Fmoc-quantification by first weighing out ~10 mg dry resin per sample into a 1.5 mL reaction tube. Add 1 mL of 20% (v/v) piperidine in DMF and incubate for 10 min while shaking at 500 rpm at 25
C to completely

208

Christin Kossmann et al.

remove all Fmoc-groups. Filter the solution with a 0.2 μm syringe filter and dilute the filtered liquid 1/50 and 1/100 in methanol to a total volume of 1 mL. Additionally, prepare blanks for each dilution with 20% (v/v) piperidine. Measure the absorbance at 298.8 nm in a silica cuvette and calculate the resin loading with the equation [29]:   A ½nmV sample ½Ll ½cmD106 loading mmol  g1 ¼ 298:8 ε298:8 ½Lmol1 cm1 mresin ½mg where: A298.8 ¼ Absorption of the sample at 298.8 nm. Vsample ¼ Sample volume [L]. L ¼ Optical path length of the cell (e.g., 1 cm) [cm]. D ¼ Dilution factor. 106 ¼ Conversion factor of mol to mmol and mg1 to g1. ε298.8 ¼ Molar absorption coefficient of Fmoc at 298.8 ¼ 6089 [L  mol1  cm1]. 3.7.2 Fmoc-SPPS Coupling Cycle

A fritted reaction vessel with a cap and a septum, combined with a MiniBlock system (Mettler Toledo, Columbia MD, USA), is convenient equipment for peptide synthesis. With this equipment setup, multiple peptides can be synthesized in parallel, and excess solvents can be flushed out into a collection flask with a flow of air or nitrogen gas (see Note 13). We use a 0.1 mmol scale as a standard scale, thus the amounts mentioned in the following protocol refer to a 0.1 mmol scale. Amino acid coupling can be performed with several coupling reagents. The following steps describe the use of HBTU and DIEA. The ratio of amino acid/HBTU/DIEA is 4/4/8 equivalents relative to the resin loading. 1. Place the resin in a fritted reaction vessel equipped with a bottom cap and a septum. Close the vessel with the bottom cap and add 2 mL DMF. The resin must be completely covered. Let the resin swell for at least 20 min. 2. If a precoupled Fmoc-protected resin is used, the synthesis starts with the deprotection of the precoupled amino acid. If no Fmoc-protection group is attached, proceed directly with the coupling step (see step 6). 3. Add 2 mL 20% piperidine to the resin-bound peptide and incubate, while shaking for 2 min on a MiniBlock system with 450 rpm. Subsequently, drain the solvents by flushing it with air through the filter. 4. Repeat step 3 to ensure complete Fmoc deprotection.

Synthetic PDZ Domains

209

5. Flow-wash with DMF for 1 min before continuing with the coupling of the next amino acid. 6. Dissolve 4 eq Fmoc-protected amino acid in 4 eq HBTU (0.8 mL) from the stock solution in a 10 mL glass vial. After complete dissolution, add 8 eq. (0.14 mL) DIEA into the mixture to preactivate the amino acid for 2 min (see Note 14). Transfer the preactivated amino acid mixture to the reaction vessel containing the deprotected resin-bound peptide (see Note 15). 7. For the introduction of the phosphorylated amino acid perform step 4, but with 4 eq Fmoc-protected phosphorylated building block and dissolve them with 4 eq. 1-[Bis (dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxid hexafluorophosphate (HATU) (152 mg) in 0.8 mL DMF. After complete dissolution, add 8 eq. (0.14 mL) DIEA, mix and preactivate for 2 min. Transfer the preactivated amino acid mixture to the reaction vessel containing the deprotected resin-bound peptide. 8. Close the reaction vessel with a septum and incubate under shaking at RT for 1 h (see Note 16). 9. Flow-wash with DMF for 30 s to remove excess coupling mixture and by-products (see Note 17). 10. Coupling efficiency can be tested via a Kaiser test (see Note 18). If coupling is incomplete, indicated by a blue color of the resin beads, repeat coupling. If the Kaiser test is negative (no blue color change), continue with deprotection and coupling of the next amino acid (see Note 19). 3.7.3 Cleavage and Global Deprotection

1. Prepare 5 mL cleavage mix (TFA/phenol/H2O/thioanisole/ EDT 82.5/5/5/5/2.5) per 0.1 mmol scale synthesis (see Note 20) [30]. 2. Add 5 mL cleavage mixture to the deprotected peptide and incubate for 1–2.5 h while shaking (see Note 21). 3. Transfer the cleavage solution into a 50 mL-conical tube by filtering it through a reaction vessel, equipped with a filter. To achieve full recovery, flush subsequently 2 with TFA/DCM (0.3–1 mL in total). 4. Precipitate the peptide with 30–40 mL ice-cold diethyl ether and spin it down by centrifuging 7 min with 3500  g at 4
C (see Note 22). Remove the clear supernatant by decanting and dissolve the peptide pellet in a solution of 25% acetonitrile, 0.1% TFA in ultrapure water. Lyophilize the peptide to obtain white solids, which can be stored long-term at 20
C (see Note 23) until further use.

210

Christin Kossmann et al.

5. Dissolve the pellet in 20% MeCN, 0.1% TFA in water, filter with a 0.2 μm syringe filter, and purify with preparative RP-HPLC, applying a linear gradient of a binary solvent system of H2O–MeCN–TFA (A: 95:5:0.1; B: 5:95:0.1) at a flow rate of 20 mL/min. 6. Lyophilize the purified sample to obtain white solids, which can be stored long-term at 20
C until further use. 3.8 Boc/Bzl-SPPS for the Insertion of Amide-to Ester Mutations

3.8.1 Synthesis of α-Hydroxy Acids

The insertion of an ester bond into a peptide is usually performed by coupling the α-hydroxy acid to the N-terminus of the growing peptide. Subsequently, an ester bond is formed with the following α-amino acid. Another option to incorporate amide-to-ester substitution is the use of a preformed dipeptide building block [Boc-amino acid-(CO)O-amino acid-OH] in standard Boc-SPPS. These building blocks are produced by solution-phase synthesis. The protocols below detail the synthesis of α-hydroxy acids. 1. Treat 1 mmol Boc-protected amino acid with 3 mL TFA for 15 min to remove the Boc protection group. 2. Evaporate TFA and dissolve the residue in a 4 mL mixture of 1:1 (v/v) dioxane and water and cool it to 0
C in an ice bath. Subsequently, add 2.0 mmol tert-butylnitrite and incubate at 20
C for 1 h, constantly stirring. 3. Upon completion, remove the solvent in vacuo and purify the resulting residue by silica gel chromatography (DCM/MeOH 10/1 as a standard method).

3.8.2 Boc-SPPS Coupling Cycle

Merrifield (chloromethyl) or PAM (phenylacetamidomethyl) resins are the most commonly used in the standard Boc-SPPS protocol and suitable for N-terminal fragments [31]. For thioester peptides, Trt-S-Ac-L-Leu-PAM resin is used, which results in a functional thioester after HF cleavage. In the protocol below, we use HBTU and DIEA for coupling in a 0.1 mmol scale. The stoichiometric ratio of the reagents is 4/4/8 (AA/HBTU/DIEA) throughout the whole protocol (see Note 24). 1. Weigh out the desired resin in an adequate amount and transfer it to a fritted reaction vessel with a stopcock and a Teflon screw cap. After closing the vessel with the stopcock, fill in DMF until the resin is completely covered. Let the resin swell for at least 30 min. If the resin is precoupled with the first Boc-protected amino acid, proceed to the deprotection step. If no Boc-protection group is attached, proceed directly with the coupling step (step 6). 2. Flow wash the resin with DMF for 30 s (see Note 25).

Synthetic PDZ Domains

211

3. Close the reaction vessel with the stopcock and add 2 mL neat TFA. Close the reaction vessel with the Teflon screw cap and allow deprotection for 1 min. 4. Repeat the deprotection (step 2) to ensure complete Boc-deprotection. 5. Flow wash the resin with DMF for 30 s. 6. Weigh 4 eq of Boc-protected amino acid into a 10 mL glass vial and add 4 eq of HBTU (0.8 mL) from the 0.5 M stock solution. Dissolve the amino acid by shaking. 7. After complete dissolution, add 8 eq DIEA (0.14 mL) to the amino acid–HBTU mixture. Allow preactivation for 2 min and transfer the solution to the resin. 8. Close the reaction vessel with the Teflon screw cap and shake the mixture in a wrist arm shaker. An incubation time of 10–20 min is needed for complete coupling efficiency. After 10–20 min, the coupling efficiency can be tested with a Kaiser test (see Note 18). If the coupling is shown to be incomplete by a blue color change of the resin beads, repeat the coupling step. In case of complete coupling (no blue color change), continue with the deprotection step and the coupling of the next amino acid. 3.8.3 Cleavage and Global Deprotection

Cleavage and deprotection of peptides synthesized by the Boc/Bzl is most efficiently done by the anhydrous hydrogen fluoride, a strong acid that is volatile and toxic and must be handled in specialized equipment. The protocol for HF cleavage is beyond the scope of this chapter, but extensively described elsewhere [32].

3.9 Expressed Protein Ligation

For both, C- and N-terminal modifications, the same ligation protocol is used. If a modification is implemented in the N-terminal peptide fragment, this fragment bears a thioester in case of Boc-SPPS, or a hydrazine linker in case of Fmoc-SPPS. For the latter option, activate the hydrazide via oxidation preliminary to the ligation step (see Subheading 3.9.1).

3.9.1 Oxidation of Hydrazide Peptide to Generate an Active Thioester

1. Weight out the hydrazide peptide in a reaction tube and dissolve it to a final concentration of 4 mM in ligation buffer 1.

3.9.2 Ligation

1. Weigh out the fragments with a proportion of 1/1.2 of C-terminal fragment to the N-terminal fragment.

2. Add 5 eq of the 0.2 M NaNO2 solution to the hydrazide peptide and cool the reaction mixture to 0
C, while stirring (see Note 26). Complete oxidation takes 20 min (see Note 27).

2. Dissolve the N-terminal thioester peptide to a final concentration of 2 mM in ligation buffer 2 and adjust pH to 6.0 with NaOH (see Note 28).

212

Christin Kossmann et al.

3. Add the C-terminal peptide to the reaction mixture and adjust pH carefully to 6.8–7.0 with NaOH or HCl (see Note 29). 4. Incubate the reaction for 2–8 h at 25
C on a shaker at 500 rpm. If overnight reactions are necessary, incubate the solution at 4
C while constant shaking. 5. Quench the reaction by adding 60 eq TCEP from the TCEP stock solution. 6. The ligation product can be purified directly, either via sizeexclusion chromatography (see Subheading 3.3.3) or via preparative HPLC. 3.10

Desulfurization

Since the native chemical ligation reaction often requires the mutation of Ala to Cys, a conversion of the Cys to Ala might be necessary to remove EPL artifacts. This step can be performed by metal-free desulfurization with TCEP, a radical initiator (VA-044) and reduced glutathione. 1. Place the ligation product in a 1.5 mL reaction vessel and dissolve it in solubilizing buffer to a final concentration of 2.5 mM. 2. Add 0.5 M TCEP solution to the peptide to achieve a final TCEP concentration of 0.2 M. Mix by vortexing. 3. Add reduced glutathione (solid) to a final concentration of 0.04 M. 4. Add VA-044 (solid) to a final concentration of 0.02 M. 5. The reaction takes 5–20 h at RT, approximately 20
C, on a shaker at 500 rpm (see Note 30). 6. Test the completion of desulfurization periodically by removing 10 μL sample and measure it by LC-MS and UPLC.

4

Notes 1. Procuring or constructing a plasmid that encodes the fulllength protein-of-interest will give flexibility in the cloning of all recombinant fragments for EPL. For PDZ domains, we purchased the full-length sequence of the PDZ domain and cloned it into pRSET vector. This construct can be used either in mutagenesis to generate a plasmid encoding C-fragment, or as the DNA template in PCR to generate the fragment that would be cloned into an intein-encoding plasmid. 2. Six (Ala, Gly, Phe, Ile, Leu, Met, Asn, Gln, and Val) of the nineteen α-hydroxy acids are commercially available.

Synthetic PDZ Domains

213

3. We use T7-based expression vectors for PDZ domains, and therefore T7 promoter forward (5- TAATACGACTCACTA TAGGG -3) and the T7 terminator reverse (5- GCTAGT TATTGCTCAGCGG -3) sequencing primers are typically used to verify our plasmids. 4. Certain PDZ fragments when fused to an intein are expressed as insoluble protein. Optimization of expression conditions such as lowering temperature could allow expression of soluble protein. Otherwise, it is possible to isolate and solubilize the inclusion bodies with high concentration of denaturant (6 M Gu·HCl or urea) prior to protein purification. 5. If the recombinant protein is expressed in inclusion bodies, collect the protein pellet rather than the supernatant after centrifugation. Then dissolve the pellet in lysis buffer supplemented with high concentrations of denaturant (e.g., 6 M Gu·HCl). Incubate for 1 h until all particles are visually solubilized. Centrifuge the lysate at 50,000  g for 60 min and collect supernatant. The lysate can then be purified using the same IMAC steps but with elution buffers containing the same concentration of denaturant in lysis buffer. 6. The pH should be adjusted to a pH above the pKa of His. pH 8 is attempted for enhancing the fraction of His being deprotonated and thus able to chelate to the Ni column. 7. The highest possible purity is desired to ensure a clean ligation reaction with few site products and a better estimation of equivalents used during NCL. 8. Reaction conditions may need to be optimized for different proteins and inteins in order to maximize cleavage efficiency while minimizing thioester hydrolysis. As a general rule, higher thiolysis temperature, MESNa concentration, and reaction pH would increase intein cleavage rate. However, if thioester hydrolysis persists, adding urea to target a concentration between 1–2 mM, as well as lowering reaction temperature and pH can minimize the formation of the hydrolyzed side product. For any pH adjustments performed during thiolysis, it is also critical that the NaOH or HCl is added carefully so that the reaction does not reach extreme ranges where the thioester can hydrolyze. 9. If a pH adjustment is necessary before purification with a nickel column, carefully apply 1 M NaOH or HCl to prevent the sample from reaching extreme pH ranges, where the protein thioester can hydrolyze. 10. Although performing the reaction at 4
C overnight, or 8–12 h, can prevent nonspecific cleavage of proteins, which is known to occur in some PDZ domains, they can also result in incomplete cleavage. To improve efficiency, more enzyme can

214

Christin Kossmann et al.

be used or the reaction can also be conducted at room temperature or ~20
C, but this would only be feasible for proteins that are not as susceptible to nonspecific cleavage. Extended reaction time and buffer optimization can also improve factor Xa activity, or an alternative cleavage enzyme can be used instead, such as SUMO protease. 11. A small-scale cleavage of 1–3 mg of resin is a helpful tool to test the synthesis success. Truncations can be detected and annotated. 12. Polystyrene resin is not suitable for long peptides. For long peptides high swelling resins, like TG-Wang or ChemMatrix® resins are recommended. 13. Alternatively, solvents can also be removed by applying vacuum. 14. Insufficient preactivation or an excess of HBTU can lead to a tetramethylguanidinium termination adduct on amine (+98 Da). 15. The volume of the solution should cover the resin sufficiently so that the resin flows freely during shaking. If the volume is not sufficient, add DMF and prolong coupling time if necessary (test by Kaiser-test). 16. Amino acid coupling can also be performed with heating. This decreases the coupling time. 90
C enables complete coupling within 1 min, but fragile peptides may be degraded. With heating at 70
C, 5 min coupling time is recommended. 17. For peptides containing oxygen-sensitive moieties (e.g., Met or Cys) it is important not to flush the peptide with excess air, to avoid oxidation. Flushing with nitrogen can be used as an alternative. 18. To perform a Kaiser test, wash 1–2 mg resin with DCM (3  1 mL) and proceed according to the manufacturer’s instructions. 19. The synthesis can be stopped after any coupling step. In this case, wash the resin 1 min with DMF and 1 min with DCM and transfer the vessel to a desiccator. 20. Depending on the peptide properties, different cleavage mixtures may be needed [30]. 21. For short peptides, a cleavage time of 1 h is sufficient. For peptides >20 amino acids or Arg containing peptides, an increased incubation time should be allowed. Increased temperature may also be used to shorten cleavage time. 22. Short peptides or very hydrophilic peptides may not fully precipitate directly. In that case, let the solution rest on dry ice for 1 h before continuing.

Synthetic PDZ Domains

215

23. A sample for quality measurements should be taken before freezing the peptide. We run LC-MS and UPLC measurements as a standard for all synthesized peptides after cleavage. 24. Boc-Asn is activated with HBTU and HOBt, Boc-Asn(Xan) is activated with HBTU only. 25. For Gln couplings, DCM should be used for washing to avoid Pca formation. 26. If high amounts of by-products appear, cool to 15
C by using NaCl and ice. 27. The reaction can be monitored by converting the azide to a thioester by adding DTT or MPAA and measuring it in LC-MS and UPLC. 28. If an N-terminal Fmoc-synthesized hydrazine peptide is used after activation via oxidation, add ligation buffer 2 to a final concentration of 2 mM and remove cooling. 29. If the pH climbs very high above 7.0, the peptide thioester will hydrolyze. 30. Depending on the protein stability it can also be incubated at 4
C or 37
C with increased or decreased reaction times, respectively. References 1. Merrifield RB (1963) Solid-phase peptide synthesis. I. The synthesis of a tetrapeptide. J Am Chem Soc 85:2149–2154 2. Palomo JM (2014) Solid-phase peptide synthesis: an overview focused on the preparation of biologically relevant peptides. RSC Adv 4:32658–32672 3. Behrendt R, White P, Offer J (2016) Advances in Fmoc solid-phase peptide synthesis. J Pept Sci 22:4–27 4. Wo¨hr T, Mutter M (1995) Pseudo-prolines in peptide synthesis: direct insertion of serine and threonine derived oxazolidines in dipeptides. Tetrahedron Lett 36:3847–3848 5. Spare LK, Laude V, Harman DG, AldrichWright JR, Gordon CP (2018) An optimised approach for continuous-flow solid-phase peptide synthesis utilising a rudimentary flow reactor. React Chem Eng 3:875–882 6. Varela Y, Vanegas Murcia M, Patarroyo M (2018) Synthetic evaluation of standard and microwave-assisted solid-phase peptide synthesis of a long chimeric peptide derived from four plasmodium falciparum proteins. Molecules 23:2877 7. Dawson PE, Kent SB (2000) Synthesis of native proteins by chemical ligation. Annu Rev Biochem 69:923–960

8. Conibear AC, Watson EE, Payne RJ, Becker CF (2018) Native chemical ligation in protein synthesis and semi-synthesis. Chem Soc Rev 47:9046–9068 9. Johnson EC, Kent SB (2006) Insights into the mechanism and catalysis of the native chemical ligation reaction. J Am Chem Soc 128:6640–6646 10. Li J, Li Y, He Q, Li Y, Li H, Liu L (2014) One-pot native chemical ligation of peptide hydrazides enables total synthesis of modified histones. Org Biomol Chem 12:5435–5441 11. Thompson RE, Liu X, Alonso-Garcı´a N, Pereira PJB, Jolliffe KA, Payne RJ (2014) Trifluoroethanethiol: an additive for efficient one-pot peptide ligation- desulfurization chemistry. J Am Chem Soc 136:8161–8164 12. Huang YC, Chen CC, Gao S, Wang YH, Xiao H, Wang F, Li YM (2016) Synthesis of L-and D-ubiquitin by one-pot ligation and metal-free desulfurization. Chem Eur J 22:7623–7628 13. Ghassemian A, Wang CIA, Yau MK, Reid RC, Lewis RJ, Fairlie DP, Durek T (2013) Efficient chemical synthesis of human complement protein C3a. Chem Commun 49:2356–2358 14. Torbeev VY, Kent SB (2007) Convergent chemical synthesis and crystal structure of a

216

Christin Kossmann et al.

203 amino acid “covalent dimer” HIV-1 protease enzyme molecule. Angew Chem Int Ed 46:1667–1670 15. Durek T, Becker CF (2005) Protein semisynthesis: new proteins for functional and structural studies. J Biomed Eng 22:153–172 16. Blanco-Canosa JB, Nardone B, Albericio F, Dawson PE (2015) Chemical protein synthesis using a second-generation N-acylurea linker for the preparation of peptide-thioester precursors. J Am Chem Soc 13:7197–7209 17. Huang YC, Chen CC, Li SJ, Gao S, Shi J, Li YM (2014) Facile synthesis of C-terminal peptide hydrazide and thioester of NY-ESO-1 (A39-A68) from an Fmoc-hydrazine 2-chlorotrityl chloride resin. Tetrahedron Lett 70:2951–2955 18. Muir TW, Sondhi D, Cole PA (1998) Expressed protein ligation: a general method for protein engineering. Proc Natl Acad Sci U S A 95:6705–6710 19. Haase C, Rohde H, Seitz O (2008) Native chemical ligation at valine. Angew Chem Int Ed 47:6807–6810 20. Sato K, Kitakaze K, Nakamura T, Naruse N, Aihara K, Shigenaga A, Otaka A (2015) The total chemical synthesis of the monoglycosylated GM2 ganglioside activator using a novel cysteine surrogate. Chem Commun 51:9946–9948 21. Harpaz Z, Siman P, Kumar KA, Brik A (2010) Protein synthesis assisted by native chemical ligation at leucine. Chembiochem 11:1232–1235 22. Han J, Luby-Phelps K, Das B, Shu X, Xia Y, Mosteller RD, Broek D (1998) Role of substrates and products of PI 3-kinase in regulating activation of Rac-related guanosine triphosphatases by Vav. Science 279:558–560 23. Wan Q, Danishefsky SJ (2007) Free-radicalbased, specific desulfurization of cysteine: a powerful advance in the synthesis of

polypeptides and glycopolypeptides. Angew Chem Int Ed 46:9248–9252 24. Pedersen SW, Moran GE, Sereikaite˙ V, Haugaard-Kedstro¨m LM, Strømgaard K (2016) Importance of a conserved Lys/Arg residue for ligand/PDZ domain interactions as examined by protein semisynthesis. Chembiochem 17:1936–1944 25. Eildal JN, Hultqvist G, Balle T, Stuhr-HansenN, Padrah S, Gianni S, Jemth P (2013) Probing the role of backbone hydrogen bonds in protein–peptide interactions by amide-to-ester mutations. J Am Chem Soc 135:12998–13007 26. Pedersen SW, Albertsen L, Moran GE, Levesque B, Pedersen SB, Bartels L, Strømgaard K (2017) Site-specific phosphorylation of PSD-95 PDZ domains reveals fine-tuned regulation of protein-protein interactions. ACS Chem Biol 12:2313–2323 27. Haj-Yahya M, Lashuel HA (2018) Protein semisynthesis provides access to tau diseaseassociated post-translational modifications (PTMs) and paves the way to deciphering the tau PTM code in health and diseased states. J Am Chem Soc 140:6611–6621 28. Chin JW, Cropp TA, Anderson JC, Mukherji M, Zhang Z, Schultz PG (2003) An expanded eukaryotic genetic code. Science 301:964–967 29. Eissler S, Kley M, B€achle D, Loidl G, Meier T, Samson D (2017) Substitution determination of Fmoc-substituted resins at different wavelengths. J Pept Sci 23:757–762 30. Applied Biosystems (1998) Cleavage, deprotection, and isolation of peptides after Fmoc synthesis. Tech Bull 1–12 31. Schno¨lzer M, Alewood P, Jones A, Alewood D, Kent SB (2007) In situ neutralization in Boc-chemistry solid-phase peptide synthesis. Int J Pept Res Ther 13:31–44 32. Jensen KJ (2013) Solid-phase peptide synthesis: an introduction. Peptide synthesis and applications. Methods Mol Biol 1047:1–21

Chapter 13 Viral PDZ Binding Motifs Influence Cell Behavior Through the Interaction with Cellular Proteins Containing PDZ Domains Carlos Castan˜o-Rodriguez, Jose M. Honrubia, Javier Gutie´rrez-A´lvarez, Isabel Sola, and Luis Enjuanes

Abstract Viruses have evolved to interact with their hosts. Some viruses such as human papilloma virus, dengue virus, SARS-CoV, or influenza virus encode proteins including a PBM that interact with cellular proteins containing PDZ domains. There are more than 400 cellular protein isoforms with these domains in the human genome, indicating that viral PBMs have a high potential to influence the behavior of the cell. In this review we analyze the most relevant cellular processes known to be affected by viral PBM–cellular PDZ interactions including the establishment of cell–cell interactions and cell polarity, the regulation of cell survival and apoptosis and the activation of the immune system. Special attention has been provided to coronavirus PBM conservation throughout evolution and to the role of the PBMs of human coronaviruses SARS-CoV and MERS-CoV in pathogenesis. Key words PDZ, PBM, Virus, Pathogenesis, Replication

1

Introduction PDZ domains are protein–protein interaction sequences consisting of 80–90 amino acids. They form a structure composed of an antiparallel β barrel made of six β strands and two α helixes (Fig. 1) [1]. PDZ is an acronym formed as the abbreviation of the names of the three first proteins where this domain was identified: postsynaptic density-95 (PSD-95), the drosophila tumor suppressor protein Dlg1 (discs large 1), and the tight junction protein ZO-1 (zonula occludens 1) [2]. These domains can be found in hundreds of proteins, both in eukaryotic cells and in bacteria [3]. Up to 266 PDZ domains which are part of more than 400 different protein isoforms have been described in the human genome [4, 5]. These proteins may include from one to thirteen PDZ

Jean-Paul Borg (ed.), PDZ Mediated Interactions: Methods and Protocols, Methods in Molecular Biology, vol. 2256, https://doi.org/10.1007/978-1-0716-1166-1_13, © Springer Science+Business Media, LLC, part of Springer Nature 2021

217

218

Carlos Castan˜o-Rodriguez et al.

Fig. 1 Structure of a PDZ domain: The secondary structure of a PDZ domain is shown. Two alpha helixes (αA and αB) are shown in red and six beta sheets (βA, βB, βC, βD, βE, and βF) are shown in yellow. (Figure modified from [1])

domains as is the case of PICK1 and MUPP1, respectively. In some cases, two PDZ domains can be located in close proximity to each other inside the same protein, such that they are arranged into a PDZ tandem that may act as a single functional unit [4, 6]. PDZ domains are normally found together with other protein– protein interaction domains in the same protein. For example, proteins from the membrane-associated guanylate kinase (MAGUK) include an SH3 module, multiple PDZ domains, and a guanylate kinase domain which is inactive. The presence of different protein–protein interacting domains allows proteins containing PDZ domains to bind many proteins at the same time and act as a scaffold. This is why they participate in a wide variety of biological processes such as cell polarity regulation, cell–cell interactions, cell migration, proliferation and survival, intracellular transport, signal transduction, or protein arrangement [5, 7]. PDZ domains mainly interact with a short stretch of amino acid residues arranged in a specific manner called PDZ-binding-motifs (PBMs). However, they may also interact with other PDZ domains or even phospholipids [8, 9]. PBMs are specific sequences usually located in the last amino acids of the protein. Although controversial, PDZ domains can be classified in three different groups, according to the PBM sequence they interact with: Class I PDZ domains recognize the motif X-S/T-X-Φ-COOH (where X could be any amino acid and Φ is a hydrophobic amino acid (normally V, I, or L), class II PDZs bind the sequence X-Φ-S-Φ-COOH, and

PBM-PDZ Interactions and Viral Pathogenesis

219

Fig. 2 PBM–PDZ binding structure: The binding of AF6 protein PDZ domain (yellow) with Bcr PBM (blue) is shown. Dashed line (green) shows the interactions between the PDZ and the PBM amino acids (Adapted from [88])

class III PDZs bind X-D/E-X-Φ-COOH. However, there are other PDZ domains that do not fit into any of these groups [10]. Although PBMs are located in the carboxy-terminus end of proteins [1, 11, 12], internal PBMs have also been described and could be more common than originally thought [13]. Interestingly, some cellular proteins such as NHERF1 have a PDZ domain and a PBM in the same protein, which contributes to regulate their availability to interact with other proteins [14]. Structurally, PBMs bind PDZ domains in a process called β-augmentation [15], in which the PBM acts as a new β chain that binds βB chain from the PDZ domains (Fig. 2). Carboxy moiety of the last amino acid of the PBM interacts through hydrogen bridges with the amide side-chain of the residues in loop βA-βB of the PDZ, which is why they have great influence over the specificity of the PDZ for the PBM. However, an increasing amount of evidence indicates that PBM–PDZ interactions require more residues than the four amino acids of the PBM and those of loop βA-βB of the PDZ, as some PDZ domains require additional sequences beyond the domain in order to be functional. These PDZs are termed “extended PDZ domains” [6]. In fact, given the high variability of PDZ sequences both in length and in sequence, PBM–PDZ interactions can be highly diverse, which is why predictions of which PBM interact with a given PDZ have failed to be accurate [4]. Viruses have developed several mechanisms to interact with the host and use its machinery for their own benefit. One of these mechanisms includes the interaction with cellular PDZ proteins

220

Carlos Castan˜o-Rodriguez et al.

Table 1 Viral proteins with PBMs and the PDZ domains they interact with FAMILY

VIRUS

VIRAL PROTEIN

CELLULAR TARGET

PBM SEQUENCE

Adenoviridae

Human adenovirus (Ad9)

E4-ORF1

Dlg1, MAGI-1, MUPP1, PATJ, ZO-2 (77)

…FPSVKIATLV

HBV

Core Protein

GIPC (78), PTPN3 (79)

…RRSQSRESQC (non-canonical)

Hepadnaviridae HCV

…WISSECTTPC

NS4b

Scribble (20)

NS1

Dlg1, MAGI-1, MAGI-2, MAGI-3, Scribble (54) PDLIM2 (80)

…KMARTIESEV

HPV

E6

GAL/GOPC, Dlg1, Dlg4, MAGI-1, MAGI-2, MAIG-3, MUPP1, PATJ, PTPN3, PTPN13, Scribble, TIP-1, TIP-2/GIPC (16)

…SSRTRRETQL

RhPV

E7

(non-canonical) Orthomyxoviridae

Papillomaviridae

Retroviridae

IAV

HTLV1

Tax

Env

TBEV

NS5

Par3 (81) Dlg1 (82), Scribble (48), Pro-IL-16 (83) β1-Syntrophin, Dlg4, Lin-7, TIP-1, TIP2/GIPC, TIP-40 (19) MAGI-3 (84), Erbin (85), MAGI-1 (86) Dlg1(87) Scribble (61) RIMS2, ZO-1 (38) CASK, GIPC, ZO-2, GRIP2, Pro-IL-16 (21) ZO-1 (38)

…DIVCPSCASRV

…SEKHFRETEV …YSLINPESSL …STHEMYYSTA… (internal) …WELRLESSII …NESDPEGALW

DENV

NS5

WNV

NS5

HtrA2, OMP25, CLIM1, ZO-2, PTPN13, PDLIM4, PDZD2, GRIP2, Scribble, and others (21)

…DTIVVEDTVL

Rhabdoviridae

RABV

G

Dlg2, MAST2, MUPP1, PTPN4 (55)

…SHKSGGETRL

Poxviridae

Vaccinia virus

F11

F11 (24)

E 3a 7b E 5 M

PALS1 (50), Syntenin (23) GRIP11, APBA11 ND Syntenin1 ND ND

E

ND

3a 8

ND ND

…LSLSNLDFRL …SSEGVPDLLV …EPTTTTSVPL …QDLEEPCTKV …KPPLPPDEWF …HIIAPSSLIV …ADIELALLRA …SSEGVPDLLV …EPTTTTSVPL …HDVRVVLDFI

Flaviviridae

SARS-CoV

Coronaviridae

MERS-CoV

SARS-CoV-2

ND stands for nondetermined a ˜o-Rodriguez, E Bailly, P. Zimmermann, JP Borg, L. Enjuanes 2020, unpublished results C. Castan

PBM-PDZ Interactions and Viral Pathogenesis

221

through viral proteins including a PBM (Table 1). Given the functional versatility of proteins including PDZ domains, viruses deregulate a wide range of cellular functions through their PBMs modulating viral replication and dissemination, and contributing to viral pathogenesis. In some cases, viral PBMs interact with cellular PDZs contributing to their degradation or inactivation and in others, they cause their activation or change their function by subcellular relocalization [16, 17]. The first viral PBMs were identified two decades ago in viral oncoproteins, like human papillomavirus (HPV) E6 protein, or human T-cell lymphotropic virus type I (HTLV-1) tax protein [18, 19] or, more recently, Hepatitis C virus (HCV) NS4b protein [20]. Furthermore, viral PBMs have also been described in proteins from nononcogenic viruses, such as proteins NS1 from influenza virus, NS5 from Tick-borne Encephalitis Virus (TBEV), or SARS-CoV E protein [21–23]. The diversity among the viral PBMs could be similar to that of the cellular ones as some internal viral PBMs or viral proteins containing both a PDZ domain and a PBM have been identified. As an example, TBEV NS5 protein has two PBMs: one located in its carboxy terminus end, similarly to its homolog proteins from other flaviviruses such as Dengue Virus (DENV) or West Nile Virus (WNV), and an internal PBM in its MTase domain [21]. Also, F11 protein of vaccinia virus has both a PBM and a PDZ domain, and the two of them are effectively coordinated to promote viral dissemination [24]. In this chapter we will review how viral PBMs target relevant cellular processes governed by cellular proteins including PDZ domains: cell–cell junctions, polarity, and survival/apoptosis. Also, the influence of these viral PBMs on the host immune system contributing to viral pathogenicity will be discussed. In the last part of the review the focus will be set on the PBMs in CoVs proteins, their conservation through evolution and how they influence viral replication and pathogenicity.

2

Cellular Processes Targeted by Viral PBMs Viruses have adopted many strategies throughout evolution to use the cellular machinery for their own biological processes as well as to counteract host defenses. The best-known cellular processes affected by viral PBMs are cell–cell junction formation, cell polarity establishment, the regulation of cellular proliferation/apoptosis and of the immune system. For a clearer understanding of how viral PBMs work, it is worthy to focus on each of these processes one by one.

222

Carlos Castan˜o-Rodriguez et al.

2.1 Cell–Cell Junctions

Many cellular proteins including PDZ domains regulate the formation of membrane junctions which are located at points of cell–cell contact. There are three types of cell–cell junctions: tight junctions (TJ), adherens junctions (AJ), and desmosomes. Of the three, TJs are the most targeted by viral PBMs and they have an important role maintaining epithelial integrity by creating a barrier to diffusion of solutes across cellular membranes in order to maintain homeostasis in tissues and organs. Structurally, they are formed by transmembrane proteins which establish the contact between cells through their extracellular domains also called tight junction proteins, and by the junctional plaque, which is the complex of intracellular proteins that act as adaptors mediating the interaction of the cytoplasmic domains of the TJ proteins with the cytoskeleton, leading to a “tight” interaction between both cells [25]. Cellular proteins including PDZ domains such as MAGI-I, PATJ, or MUPP1 are some of these adaptor proteins that are targeted by the PBMs of different kinds of viruses [25]. These are some of the most relevant examples. Adenovirus type 9 (Ad9) is a human virus associated with benign eye infections that may also cause mammary tumors in experimental rats [26, 27]. Most human adenovirus infections lead to carcinogenesis by the products of their genes E1A and E1B, however Ad9 solely depends on the presence of a PBM in protein E4-ORF1 to promote cell transformation [28]. This protein has a type I PBM in its carboxy terminal (-ATLVCOOH ) that binds four cellular proteins including PDZ domains that are involved in establishing TJs: MUPP1, ZO-2, MAGI-1, and PATJ [29–32]. These interactions lead to the sequestration and consequent inactivation of the four proteins, contributing to the disruption of TJs, one of the hallmarks of carcinogenesis. HPV E6 protein has a class I PBM (-ETQLCOOH). This protein forms complexes with the cellular proteins including PDZs and E6AP ubiquitin ligase promoting the proteasome-mediated degradation of most of the cellular PDZ proteins interacting with E6 PBM. This viral protein also targets many cellular proteins including PDZ domains that contribute to TJ formation as PATJ, MUPP1, MAGI-1 [29, 32, 33] [34]. However, instead of relocalizing the cellular proteins as Ad9 E4-ORF1 did, HPV E6 promotes TJ disruption by targeting these proteins for degradation leading to tumor formation. Influenza A virus (IAV) is a respiratory virus that affects birds and mammals. It encodes protein NS1, which is a virulence factor, whose main function is to counteract the innate immune system antiviral mechanisms. This protein has a class I PBM in its carboxy terminus with a sequence that changes in isolates from different species: in highly pathogenic avian viruses the consensus sequence is -ESEVCOOH , while in less pathogenic human viruses PBM sequence is -RSKVCOOH . Each of these sequences have different

PBM-PDZ Interactions and Viral Pathogenesis

223

PDZ binding properties [22]. The avian virulent PBM binds PDZ proteins Scribble and Dlg1 [22]. Both proteins promote the assembly and stability of TJs through their interaction with proteins that are part of the junctional plaque [35, 36]. It was observed that in the context of the infection of an IAV including an NS1 protein with the -ESEV PBM, the viral PBM sequestered Dlg1 and Scribble to perinuclear puncta through a PBM–PDZ interaction disrupting TJs structurally and functionally [37]. However, its effect on virus pathogenesis is still to be described. Flaviviruses Dengue Virus (DV), West Nile Virus (WNV) and Tick-Borne Encephalitis Virus (TBEV) interact with ZO-1 and ZO-2 through their carboxy terminus PBM with an unknown effect [21, 38]. 2.2

Cell Polarity

Cell polarity is a phenomenon characterized by an asymmetrical distribution of biomolecules within the cells such as lipids, proteins, or an asymmetrical distribution of the cell itself by forming specific membrane domains, enriching organelles or the cytoskeleton at specific sites [39]. Cell polarity can be apicobasal, as in epithelial cells located in a multicellular sheet, including an apical membrane and a basolateral membrane; anterior–posterior, as in the case of migrating cells and planar cell polarity, which is developed within the plane of a given tissue. This phenomenon is required for the development of the organism and to maintain homeostasis, which is why a deregulation of cell polarity may lead to tumorigenesis, several birth defects [40] or diseases [41, 42]. Many cellular proteins including PDZ domains are relevant in regulating cell polarity [5]. Apical-basal cell polarity is disrupted by viral PBMs through the interaction with at least one of this three conserved protein complexes involving proteins with a PDZ domain: the Apical Crumbs complex involving PATJ-PALS1CRUMBS [43]; the TJ Par Complex involving PAR3-PAR6aPKC [44] and the Lateral Scribble Complex located at more basal location including Dlg1-Lgl-Scribble [45]. These are some of the most relevant examples of viral PBMs targeting apical-basal cell polarity. Oncovirus such as adenovirus E4-ORF1 and HPV-16 E6 protein PBMs interact and relocalize or eliminate PATJ disrupting the Apical Crumbs complex leading to the loss of apical-basal polarity, which promotes cell transformation [32]. HPV E6 PBM also targets the Lateral Scribble Complex such as Scribble and Dlg1 and promotes their degradation [18, 46]. The consequence of this degradation in cell polarity has not been fully studied but it could contribute to tumorigenesis, as there are cervical cancers in which Dlg1 and Scribble show a reduced expression or are absent [47]. Human T Cell Leukemia Virus 1 (HTLV-1) Tax protein includes a PBM in its carboxy terminus (-ETEVCOOH ) which is relevant for virus-induced leukemia. This PBM binds Scribble

224

Carlos Castan˜o-Rodriguez et al.

leading to perturbations in the cytoskeleton and the loss of T-cell polarization contributing to tumorigenesis [48, 49]. Cell polarity is also targeted by respiratory viruses such as IAV and SARS-CoV. In the case of coronaviruses (CoVs), SARS-CoV E protein, includes a PBM at its carboxy terminus (-DLLVCOOH) that interacts with the PDZ domain of PALS1. It was observed that this PBM relocalized PALS1 to the ERGIC in SARS-CoV infected cells, altering its original localization and disrupting the Apical Crumbs complex. This activity delayed the formation of TJs between epithelial cells and disrupted cell polarity [50]. IAV NS1 PBM targets Dlg1 and Scribble as described above, causing the disruption of the lateral Scribble complex although the physiological consequences of this disruption have not been clearly established. Both SARSCoV and IAV cause an exacerbated immune response that leads to lung edema and death in the most severe cases. Edema clearance in lung epithelia requires proper apico-basal polarity. Therefore, it has been hypothesized that the disruption of cell polarity of lung epithelia caused by these viral PBMs could hinder edema clearance which would lead to edema accumulation and the death of the host. However, this is yet to be confirmed. 2.3 Cell Survival and Apoptosis

Viruses inhibit cell apoptosis and enhance cell survival in order to have a proper environment to replicate. In the case of oncoviruses, altering cell–cell interactions and cell polarity is part of cell tumorigenesis. However, this is also achieved by their PBMs directly inhibiting signaling pathways that induce apoptosis and enhancing those that promote cell survival. Dlg1 and Scribble are well known for their tumor suppressor activity. Both proteins induce an inhibitory effect on the PI3K-Akt pathway, which upon activation, promotes the phosphorylation of Akt that leads to the activation of several cellular signaling cascades causing cell proliferation and survival. Dlg1 interacts with the PBM of cellular protein PTEN contributing to its activation. PTEN is a phosphatase that acts on PIP3 leading to the inhibition of the PI3K-Akt pathway. Similarly, Scribble inhibits this pathway by its interaction with the PBM of the cellular phosphatase PHLPP [51], localizing PHLPP to the membrane where it exerts its inhibitory effect over the PI3K-Akt pathway. HTLV-1 Tax protein PBM interacts with both Dlg1 and Scribble competing with PTEN and PHLPP, respectively, contributing to the activation of the PI3K-Akt pathway [52]. Similarly, HPV E6 protein interacts and promotes the degradation of MAGI-2 and MAGI-3 [53], two PDZ proteins which bind PTEN through its own PBM. It has been hypothesized that HPV would induce the degradation of MAGI-2 and MAGI-3 leading to the deregulation of PTEN causing cell malignancy. However, this effect is yet to be proven experimentally. There are many other signaling pathways involving cellular PDZ proteins. In this cases, E6 PBM affects cell proliferation and survival, as reviewed [16].

PBM-PDZ Interactions and Viral Pathogenesis

225

On the other hand, viruses that are highly pathogenic but do not generate tumors, like Rabies (RABV) or IAV, also interfere with cellular PDZs affecting cell survival and apoptosis, but with a different effect. For instance, the interaction of highly pathogenic avian IAV H7N1 NS1 protein PBM (-ESEVCOOH ) with Scribble led to an increase in viral titers that was not observed in a virus variant with a mutant PBM (-ESEACOOH ), which did not bind Scribble. This binding blocked Scribble antiapoptotic function relocalizing the protein into cytoplasmic puncta, suggesting that the inhibition of apoptosis by this PBM was responsible for an increase in virus titers [54]. RABV virus, a neurotropic virus that causes severe encephalitis in mammals, includes a PBM in the carboxy terminus end of its G protein. Infections by RABV variants that promote neuronal death by apoptosis lead to the survival of the host in contrast to those that promote cell survival, which are highly virulent [55]. This difference is due to the changes in the sequence of G protein PBM. Variants that include the PBM -QTRLCOOH bind the PDZ domain of cellular Ser/Thr kinase MAST2 with similar affinity as its natural binder, PTEN [56]. This leads to the activation of Akt and therefore, cell survival. In contrast, attenuated variants include the sequence -ETRLCOOH , showing an increased PDZ binding ability, binding to MAST2 and Dlg2, MUPP1 and PTPN4, a protein that upon activation, promotes neural apoptosis overcoming the effect of the binding to MAST2 [55]. Furthermore, a recent analysis of the sequence of the G protein of high morbidity and low morbidity RABV variants has revealed that low morbidity viruses lack a PBM in its carboxy terminus domain, further supporting the implications on pathogenesis on RABV G protein PBM [57, 58]. 2.4 Disruption of the Immune System

Viral PBMs also subvert the host immune system by triggering or inhibiting signaling pathways contributing to viral pathogenesis either by suppressing the immune system or by triggering an exacerbated immune response which is deleterious to the host. T cells are responsible for adaptative immune response of the host. The interaction with antigen presenting cells (APC) through their membrane T-cell receptor (TCR) triggers signaling pathways that lead to T-cell activation. These pathways are regulated in both APCs and T-cells by several factors including Dlg1 and Scribble [59]. One of them, is the Akt pathway, targeted by HTLV-1 Tax protein, discussed above, as it is also involved in the regulation of cell proliferation and apoptosis. The interaction of Tax protein PBM with these proteins inhibits T cell activation, contributing to the depletion of the adaptative immune response and the survival of the virus within T-cells. Also, Flaviviruses and CoVs influence the immune system through the interaction of their PBMs with cellular PDZs. IFN is one of the main determinants of host anti-viral response [60] and is targeted by proteins NS5 from the Flaviviridae family, inhibiting

226

Carlos Castan˜o-Rodriguez et al.

interferon (IFN) signaling in infected cells. TBEV NS5 protein includes a PBM at its carboxy-terminus and an internal PBM that binds Scribble [61]. This interaction does not relocalize or inhibit Scribble’s activity, it just targets NS5 protein to the membrane, where Scribble is generally located, and contributes to the inhibition of interferon mediated JAK-STAT signaling [61]. However, the exact mechanism governing this inhibition or the relevance of this interaction in the context of the viral infection is still to be elucidated. Similarly, the PBM–PDZ dependent interaction of IAV NS1 protein with both Dlg1 and Scribble has recently been reported in APCs [59]. Although the consequence of this interaction in vivo is still to be elucidated, it suggests a role of IAV NS1 PBM in modifying the host immune response. In the case of CoVs, our laboratory generated a collection of recombinant SARS-CoVs with mutations affecting E protein PBM. Their analysis in vivo showed that SARS-CoV variants missing E protein PBM were attenuated, indicating that E protein PBM was a virulence determinant. Then, we described that the interaction of this viral PBM with the PDZ domains of cellular protein syntenin during viral infection activated p38 MAPK. This interaction promoted the over expression of proinflammatory cytokines, causing the death of the host due to a pathogenic immune response [23]. These data indicated a clear role of PBM–PDZ interaction in the lung pathology observed in SARS-CoV infected patients. Both CoV and Flaviviruses infect cells of the immune system, although in some cases the infection is nonproductive [62– 65]. How the pathology of these viruses is influenced by the interaction of these viral PBMs with cellular PDZ proteins of immune cells still needs to be addressed.

3

Relevance of CoVs Proteins Including a PBM CoVs have several proteins including PBMs binding a set of cellular PDZs with the potential of influencing cell behavior. Some of these interactions are reviewed next. The relevance of SARS-CoV E protein PBM was further supported by the observation that a SARS-CoV attenuated deletion mutant lacking E protein (SARS-CoV-ΔE), after several passages in cell culture and in vivo, reverted to a virulent phenotype by incorporating chimeric proteins including new PBMs, thus compensating for the loss of E protein PBM and reinforcing its relevant role in virus virulence [66]. Also, SARS-CoV 3a protein has a PBM motif in its carboxy terminus (-SVPLCOOH ). However, this PBM is not involved in virus replication and pathogenesis, in contrast to E protein PBM, when located in its native protein, implying that either there is a hierarchy between both PBMs when they are located in their native

PBM-PDZ Interactions and Viral Pathogenesis

227

contexts [67] or that the PBM works in collaboration with an additional motif present in other viral proteins. Interestingly, when a SARS-CoV variant lacking proteins E and 3a simultaneously was generated, the virus was not viable. In fact, we described that either E or 3a protein PBM was essential for virus viability when one of the two proteins was deleted, indicating that there was a complementation between both viral PBMs [67]. Given the relevance of SARS-CoV E and 3a protein PBMs, the presence of PBM motifs in proteins of other CoVs species was reviewed (Table 2). To date, the PBMs of SARS-CoV E and 3a proteins are the only CoV PBMs that have been studied in detail. The relevance of PBMs in other CoV proteins in virus pathogenicity and replications has been determined with limited extent. Recently in our laboratory, we have shown that MERS-CoV E protein PBM is also a virulence factor, suggesting that this virus could trigger a similar virulence mechanism than SARS-CoV ´ lvarez, Enjuanes, 2020, unpublished results). (FJ Gutie´rrez-A Also, the relevance in virus replication and pathogenesis of human OC43-CoV E protein PBM has been recently proposed [68]. Another example is FIPV 7b protein, that includes a PBM that when modified, changes the subcellular localization of the protein [69]. A new human highly pathogenic coronavirus, named SARS-CoV-2 [70] emerged in Wuhan in December 2019 and to date it has spread to more than 200 countries with more than 51.000.000 confirmed cases and causing 1.280.000 deaths (data from WHO as to 13, November, 2020) leading to an unparalleled global crisis. This new pathogen was analyzed showing 79% sequence identity to SARS-CoV [71]. Interestingly, this virus has a 3a and E proteins in which the PBM sequence is identical to that of SARS-CoV, suggesting that these PBMs could be involved in virus replication and pathogenesis triggering the same cellular pathways that are activated by SARS-CoV E and 3a protein PBMs. Although functional studies are required to confirm this hypothesis, SARSCoV-2 PBMs could be a promising target for potential antivirals. Current theories state that CoVs were originated in bats and then transmitted to birds generating two branches of CoVs that were classified in four CoVs genera: Alphacoronavirus and Betacoronavirus, which are derived from bat CoVs and Gammacoronavirus and Deltacoronavirus, which are derived from bird CoVs [72]. Forty-three potential PBMs have been identified in 33 CoVs species. It has been observed that there are more PBMs in CoVs from genera derived from bat CoVs (Alphacoronavirus and Betacoronavirus) than in those derived from bird CoVs (Gammacoronavirus and Deltacoronavirus). This suggests that, during CoV evolution, PBMs were incorporated in CoV genomes earlier in genera Alphacoronavirus and Betacoronavirus. Furthermore, a PBM was identified in the carboxy terminus of the E protein of every analyzed CoV

Table 2 CoV proteins including a PBM GENERA

β

γ

Human

E

…APVPAEVLNV

E strain FS772/70 E strain Purdue 3a strain FS772/70 3a strain Purdue E E E 3a 7a strain BGF10, Insavc-1 7a strain K378 7b E 3a 7b E 5 M E 3a 8 E N E N E N E M N E N HE (only in strain F15) N E 5a E S 5a E 5c M ORF9 N E 5c M NS9 N

…AYNHDGALLV …AYNPDGALLA …AYAKLGLSTI …IEEVNSHIVV …DPLPSTVIDV …AYNPDGALLV …AYNPDEALLV …IEEVNSHTVV …CCHRLLVTLF …CCYRLLVTLI …KISQYQKSEL …SSEGVPDLLV …EPTTTTSVPL …QDLEEPCTKV …KPPLPPDEWF …HIIAPSSLIV …ADIELALLRA …SSEGVPDLLV …EPTTTTSVPL …HDVRVVLDFI …VIPSTLDDLI …DDPYVEDSVA …KPPVLDVDDV …EPYTEDTSEI …KPPVLDVDDV …EPYTEDTSEI …RLPLLEVDDI …SGADTALLRI …PDGLEDDSNV …KPPVLDVDDV …EPYTEDTSEI …YFMVENGTRL …PDGLEDDSNV …LAYTPTQSLV …FSNSVNSSLV …LVCTPTQSLV …PRNSKDGEYV …CIGNDAYLGV …KTETEKLYSV …NRKAYGSDEV …GTGDLEWSEA …PRNSKDGEYV …CIGNDAYLGV …KTETEKLYSV …NRKAYGSDEV …GTGDLEWSEA

HOST

FIPV

Cat

HCoV-229E NL63

α

Human

E 3a 7b E

CARBOXY TERMINUS SEQUENCE …AYNPDEAFLV …IEEVNSHTVV …KINQHHKTEL …DPFPKRVIDF

VIRUS

TGEV

Swine

PEDV PRCV

Swine Swine

CCoV

Dog

SARS-CoV

Human

MERS-CoV

Human

SARS-CoV-2

Human

HKU1

Human

OC-43

Human

HEV

Swine

MHV

Mouse

HCoV-4408

Bovine

BCV

Bovine

IBV

Bird

TCoV

Bird

SW1

Beluga whale

HKU22

Bottlenose dolphin

VIRAL PROTEIN

PBM-PDZ Interactions and Viral Pathogenesis

229

Table 2 (Continued) GENERA

δ

VIRUS

HOST

HKU15

Swine

ALCoV/GX/ F230/06

Leopard

HKU16

Bird

HKU17

Bird

HKU18

Bird

HKU19 HKU20

Bird Bird

HKU21

Bird

VIRAL PROTEIN 5c M NS9 N NS7 NS6 N E N E N E NS7a NS7b

CARBOXY TERMINUS SEQUENCE …CIGNDAYLGV …KTETEKLYSV …NRKAYGSDEV …AFEIKQESAA …RVWLILASWL …SLQVILEEEI …EIKRDEESTA …AFEIKQESAA …NAFEFKSSDA …HQFPRNSFSV …VKRKSLIDSA …SDADISSDDA

from genera Alphacoronavirus and Betacoronavirus. However, this was not the case in CoVs evolved from birds with the exception of E protein from bottlenose dolphin CoV HKU22. Interestingly, SARS-CoV, MERS-CoV, and SARS-CoV-2, the three human CoVs that are highly pathogenic, have three proteins including a PBM in their carboxy-terminus: proteins E, 3a and 7b in SARSCoV, proteins E, 5 and M in MERS-CoV and proteins E, 3a and 8 in SARS-CoV-2. In addition, MERS-CoV is the only CoV in which a PBM has been described in M protein, a structural CoV protein essential for viral assembly [73]. Remarkably, MERS-CoV proteins E and 5 are homologs of SARS-CoV and SARS-CoV2 proteins E and 3a, respectively, suggesting that they may play a similar role in MERS-CoV. Furthermore, SARS-CoV and MERSCoV, as most Betacoronaviruses, were originated in bats and then transmitted to humans through intermediate hosts: civet cats in the case of SARS-CoV, and camels in the case of MERS-CoV. GenBank data from genomes of hundreds of SARS-CoVs and MERS-CoVs variants isolated from bats, civets, camels and humans were analyzed showing that the PBMs from SARS-CoV E and 3a proteins were mostly conserved in bats, civets and humans similarly to the PBMs of MERS-CoV proteins E, 5 that were also mostly conserved in bats, camels and humans (Table 3, adapted from [67]). Interestingly, the mutations that affected the PBM core sequences introduced a different PBM, likely functional in the corresponding animal context. This phylogenetic conservation reinforces the relevance of viral PBMs in CoVs opening the possibility that they could be involved in CoV adaptation to the host.

230

Carlos Castan˜o-Rodriguez et al.

Table 3 PBMs conserved in MERS-CoV and MERS-CoV isolated from animals VIRAL PROTEIN

SARS-CoV E PROTEIN

MERS-CoV E PROTEIN

SARS-CoV 3a PROTEIN

MERS-CoV 5 PROTEIN

VIRUS

HOST

Nº ISOLATES

PBM

SARS-CoV SARS-CoV SARS-CoV Like

>100 >20 1

…NLNSSEGVPDDLV …NLNSSEGVPDDLV …NLNSSVGVPDDLV

1

…NLNSSDCVPDDLV

SARS-CoV Like MERS-CoV MERS-CoV MERS-CoV-Like MERS-CoV-Like MERS-CoV-Like HKU5 MERS-CoV-Like-HKU4 SARS-CoV SARS-CoV SARS-CoV SARS-CoV-Like Zaria Bat-CoV MERS-CoV

Human Civet Cats Bats R. pusillus Bats R. ferrumequinum Bats R. blasii Human Dromedary Bats N. capensis Bats V. superans Bats Pipistrellus Bats T. pachypus Human Human Civet Cats Bats R. sinicus H. commersoni Human

1 98 17 1 1 6 7 166 3 4 1 1 97

…SLNSSQEVPEFLV …QDSKPPLPPDEWV …QDSKPPLPPDEWV …QESKPPLPPEEWV …QESKPPLPPDEWV …QESHPPYPPEDWV …QENRPPFPPEDWV …IYDEPTTTTSVPL …IYDEPMTTTSVPL …IYDEPTTTTSVPL …IYDEPMTTTSVPL …IYDEPPTTTSVPL …VPLHIIAPSLIV

MERS-CoV

Dromedary

17

…VPLHIIAPSLIV

MERS-CoV-Like MERS-CoV-Like MERS-CoV-Like HKU5 MERS-CoV-Like-HKU4

Bats N. capensis Bats V. superans Bats Pipistrellus Bats T. pachypus

1 1 6 7

…VPLHIIAPVLSV …VPLHIIAPVLSV …VPLHIIAPVLTV …VPLHIIAPKLYV

SARS-CoV Like

SARS-CoV 7b protein does not seem to have any role in virus pathogenesis or replication as a deletion mutant simultaneously missing genes 6, 7a, 7b, 8a, 8b and 9b (SARS-CoV Δ6-9b) showed similar virus replication rates and caused pathogenesis similar to the wild type virus [74]. However, SARS-CoV 7b protein PBM is not present in SARS-CoV variants isolated from bats, in contrast to viruses isolated from civet cats and humans, indicating that 7b protein PBM could be involved in virus adaptation, similarly to the PBMs of E and 3a proteins [75]. To further explore the relevance of CoV PBMs and their involvement in virus replication and pathogenesis, the cellular PDZ binding partners of these viral PBMs need to be identified. To this end, a yeast-two hybrid assay was performed in which the highly pathogenic human CoV, SARS-CoV and MERS-CoV, PBMs were used as bait and the 266 human cellular PDZ domains were used as prey (Fig. 3). It was observed that SARS-CoV and MERSCoV E protein PBMs bind cellular PDZ proteins involved in the regulation of NfκB pathway, one of the hallmarks of SARS-CoV pathogenesis [76], potentially contributing to virus virulence

PBM-PDZ Interactions and Viral Pathogenesis

231

Fig. 3 Y2H assay to study the interaction between viral PBMs and cellular PDZ domains: Yeast from strain AH109 were transformed with a pGBT9 plasmid expressing GAL4 DNA binding domain fused to the last 15 amino acids of the proteins E and 3a of SARS-CoV and E and 5 of MERS-CoV. Yeasts from the strain Y187 were transformed by a library of every human PDZ domain fused to GAL4 activating domain [89]. Both strains were mated. If an interaction between a viral PBM and a cellular PDZ takes place, then GAL4 is reconstituted, activating a reporter gene, and therefore allowing for the identification of the partners involved in the interaction

˜ o-Rodriguez, P. Zimmermann, JP Borg, L. Enjuanes (C. Castan 2020, unpublished results).

4

Concluding Remarks Viruses have adapted throughout evolution to interact with their hosts. The introduction of PBMs that interact with cellular PDZ proteins is one of these modifications. Decades of scientific studies have revealed that there are many cellular processes affected by these PBMs. It has also been observed that different viral PBM sequences interact with a different set of cellular PDZ proteins even if, in some cases, they share some specific PDZ targets (Table 1). To date, the mechanisms used by viral oncoproteins including a PBM influencing cell malignancy have been the most studied as reviewed [16]. However, it is worth noting that both oncovirus and nononcovirus PBMs target identical cellular PDZ proteins that are involved in the same cellular processes, but the effect of the interaction is quite different. For example, the interaction of oncovirusPBMs with Dlg1 or Scribble proteins trigger signaling pathways

232

Carlos Castan˜o-Rodriguez et al.

leading to tumorigenesis and when they are targeted by Flavivirus or IAV protein PBMs they modulate the same signaling pathways affecting the same cellular processes but the result of the infection is quite different [52, 61]. This is probably due to the specific cellular context of each infection, as each of these viruses infects different cell types, and the effect in each viral infection is unique. Interestingly, one of the differences between viral and cellular PBMs is that as viruses evolve faster than the cellular genomes, viral PBMs may change throughout evolution to adapt to new hosts or contexts, as has been observed by the incorporation of new PBMs in SARS-CoV genome when E protein PBM was deleted [66] or with the changes in PBM core sequences in some SARS-CoV and MERS-CoV variants isolated from bats [67]. In fact, more than 40 different PBMs were identified in more than 33 CoV genomes isolated from different hosts, reinforcing the relevance of viral PBMs in virus adaptation and suggesting that CoVs could be a good model to study the relevance of viral PBMs in virus evolution. Furthermore, viral PBMs seem promising targets for antiviral therapy, as small peptides blocking the interaction of viral PBMs with the cellular PDZs could be used as antivirals to block virus pathogenicity. There is still much to be learned about how viral PBMs work. The field of nononcovirus PBMs is essentially unexplored with the exception of CoVs. To gain a better understanding on how viral PBMs influence viral infections, the mechanisms of replication and pathogenesis induced by the PBMs of nononcogenic viruses such as CoV, IAV, DENV or RABV need to be further studied. For the moment, we have shown that the study of the cellular mechanisms disrupted by human CoV PBMs are highly relevant for a better understanding of virus–host interaction. Hopefully, this will open the way to study viral PBMs of other highly pathogenic viruses in more detail. References 1. Gerek ZN, Keskin O, Ozkan SB (2009) Identification of specificity and promiscuity of PDZ domain interactions through their dynamic behavior. Proteins 77:796–811 2. Kennedy MB (1995) Origin of PDZ (DHR, GLGF) domains. Trends Biochem Sci 20:350 3. Ponting CP (1997) Evidence for PDZ domains in bacteria, yeast, and plants. Protein Sci 6:464–468 4. Luck K, Charbonnier S, Trave G (2012) The emerging contribution of sequence context to the specificity of protein interactions mediated by PDZ domains. FEBS Lett 586:2648–2661

5. Nourry C, Grant SG, Borg JP (2003) PDZ domain proteins: plug and play! Sci STKE 2003:RE7 6. Ye F, Zhang M (2013) Structures and target recognition modes of PDZ domains: recurring themes and emerging pictures. Biochem J 455:1–14 7. Subbaiah VK, Kranjec C, Thomas M, Banks L (2011) PDZ domains: the building blocks regulating tumorigenesis. Biochem J 439:195–205 8. Gallardo R, Ivarsson Y, Schymkowitz J, Rousseau F, Zimmermann P (2010) Structural diversity of PDZ-lipid interactions. Chembiochem 11:456–467

PBM-PDZ Interactions and Viral Pathogenesis 9. Hillier BJ, Christopherson KS, Prehoda KE, Bredt DS, Lim WA (1999) Unexpected modes of PDZ domain scaffolding revealed by structure of nNOS-syntrophin complex. Science 284:812–815 10. Harris BZ, Lim WA (2001) Mechanism and role of PDZ domains in signaling complex assembly. J Cell Sci 114:3219–3231 11. Hung AY, Sheng M (2002) PDZ domains: structural modules for protein complex assembly. J Biol Chem 277:5699–5702 12. Munz M, Hein J, Biggin PC (2012) The role of flexibility and conformational selection in the binding promiscuity of PDZ domains. PLoS Comput Biol 8:e1002749 13. Lenfant N, Polanowska J, Bamps S, Omi S, Borg JP, Reboul J (2010) A genome-wide study of PDZ-domain interactions in C. elegans reveals a high frequency of non-canonical binding. BMC Genomics 11:671 14. Bhattacharya S, Dai Z, Li J, Baxter S, Callaway DJ, Cowburn D, Bu Z (2010) A conformational switch in the scaffolding protein NHERF1 controls autoinhibition and complex formation. J Biol Chem 285:9981–9994 15. Remaut H, Waksman G (2006) Proteinprotein interaction through beta-strand addition. Trends Biochem Sci 31:436–444 16. James CD, Roberts S (2016) Viral interactions with PDZ domain-containing proteins-an oncogenic trait? Pathogens 5:8 17. Javier RT, Rice AP (2011) Emerging theme: cellular PDZ proteins as common targets of pathogenic viruses. J Virol 85:11544–11556 18. Kiyono T, Hiraiwa A, Fujita M, Hayashi Y, Akiyama T, Ishibashi M (1997) Binding of high-risk human papillomavirus E6 oncoproteins to the human homologue of the Drosophila discs large tumor suppressor protein. Proc Natl Acad Sci U S A 94:11612–11616 19. Rousset R, Fabre S, Desbois C, Bantignies F, Jalinot P (1998) The C-terminus of the HTLV-1 tax oncoprotein mediates interaction with the PDZ domain of cellular proteins. Oncogene 16:643–654 20. Hu B, Li S, Zhang Z, Xie S, Hu Y, Huang X, Zheng Y (2016) HCV NS4B targets scribble for proteasome-mediated degradation to facilitate cell transformation. Tumour Biol 37:12387–12396 21. Melik W, Ellencrona K, Wigerius M, Hedstrom C, Elvang A, Johansson M (2012) Two PDZ binding motifs within NS5 have roles in Tick-borne encephalitis virus replication. Virus Res 169:54–62

233

22. Obenauer JC, Denson J, Mehta PK, Su X, Mukatira S, Finkelstein DB, Xu X, Wang J, Ma J, Fan Y, Rakestraw KM, Webster RG, Hoffmann E, Krauss S, Zheng J, Zhang Z, Naeve CW (2006) Large-scale sequence analysis of avian influenza isolates. Science 311:1576–1580 ˜o JM, Nieto-Torres JL, 23. Jimenez-Guarden DeDiego ML, Regla-Nava JA, Fernandez˜o-Rodriguez C, Enjuanes Delgado R, Castan L (2014) The PDZ-binding motif of severe acute respiratory syndrome coronavirus envelope protein is a determinant of viral pathogenesis. PLoS Pathog 10:e1004320 24. Handa Y, Durkin CH, Dodding MP, Way M (2013) Vaccinia virus F11 promotes viral spread by acting as a PDZ-containing scaffolding protein to bind myosin-9A and inhibit RhoA signaling. Cell Host Microbe 14:51–62 25. Zihni C, Mills C, Matter K, Balda MS (2016) Tight junctions: from simple barriers to multifunctional molecular gates. Nat Rev Mol Cell Biol 17:564–580 26. Javier R, Raska K Jr, Macdonald GJ, Shenk T (1991) Human adenovirus type 9-induced rat mammary tumors. J Virol 65:3192–3202 27. Thomas DL, Schaack J, Vogel H, Javier R (2001) Several E4 region functions influence mammary tumorigenesis by human adenovirus type 9. J Virol 75:557–568 28. Javier R, Raska K Jr, Shenk T (1992) Requirement for the adenovirus type 9 E4 region in production of mammary tumors. Science 257:1267–1271 29. Lee SS, Glaunsinger B, Mantovani F, Banks L, Javier RT (2000) Multi-PDZ domain protein MUPP1 is a cellular target for both adenovirus E4-ORF1 and high-risk papillomavirus type 18 E6 oncoproteins. J Virol 74:9680–9693 30. Glaunsinger BA, Weiss RS, Lee SS, Javier R (2001) Link of the unique oncogenic properties of adenovirus type 9 E4-ORF1 to a select interaction with the candidate tumor suppressor protein ZO-2. EMBO J 20:5578–5586 31. Glaunsinger BA, Lee SS, Thomas M, Banks L, Javier R (2000) Interactions of the PDZ-protein MAGI-1 with adenovirus E4-ORF1 and high-risk papillomavirus E6 oncoproteins. Oncogene 19:5270–5280 32. Latorre IJ, Roh MH, Frese KK, Weiss RS, Margolis B, Javier RT (2005) Viral oncoprotein-induced mislocalization of select PDZ proteins disrupts tight junctions and causes polarity defects in epithelial cells. J Cell Sci 118:4283–4293 33. Storrs CH, Silverstein SJ (2007) PATJ, a tight junction-associated PDZ protein, is a novel

234

Carlos Castan˜o-Rodriguez et al.

degradation target of high-risk human papillomavirus E6 and the alternatively spliced isoform 18 E6. J Virol 81:4080–4090 34. Kranjec C, Banks L (2011) A systematic analysis of human papillomavirus (HPV) E6 PDZ substrates identifies MAGI-1 as a major target of HPV type 16 (HPV-16) and HPV-18 whose loss accompanies disruption of tight junctions. J Virol 85:1757–1764 35. Ivanov AI, Young C, Den Beste K, Capaldo CT, Humbert PO, Brennwald P, Parkos CA, Nusrat A (2010) Tumor suppressor scribble regulates assembly of tight junctions in the intestinal epithelium. Am J Pathol 176:134–145 36. Stucke VM, Timmerman E, Vandekerckhove J, Gevaert K, Hall A (2007) The MAGUK protein MPP7 binds to the polarity protein hDlg1 and facilitates epithelial tight junction formation. Mol Biol Cell 18:1744–1755 37. Golebiewski L, Liu H, Javier RT, Rice AP (2011) The avian influenza virus NS1 ESEV PDZ binding motif associates with Dlg1 and scribble to disrupt cellular tight junctions. J Virol 85:10639–10648 38. Ellencrona K, Syed A, Johansson M (2009) Flavivirus NS5 associates with host-cell proteins zonula occludens-1 (ZO-1) and regulating synaptic membrane exocytosis-2 (RIMS2) via an internal PDZ binding mechanism. Biol Chem 390:319–323 39. Ebnet K, Kummer D, Steinbacher T, Singh A, Nakayama M, Matis M (2018) Regulation of cell polarity by cell adhesion receptors. Semin Cell Dev Biol 81:2–12 40. Butler MT, Wallingford JB (2017) Planar cell polarity in development and disease. Nat Rev Mol Cell Biol 18:375–388 41. Schneeberger K, Roth S, Nieuwenhuis EES, Middendorp S (2018) Intestinal epithelial cell polarity defects in disease: lessons from microvillus inclusion disease. Dis Model Mech 11: dmm031088. https://doi.org/10.1242/ dmm.031088 42. Fischer E, Legue E, Doyen A, Nato F, Nicolas JF, Torres V, Yaniv M, Pontoglio M (2006) Defective planar cell polarity in polycystic kidney disease. Nat Genet 38:21–23 43. Medina E, Lemmers C, Lane-Guermonprez L, Le Bivic A (2002) Role of the Crumbs complex in the regulation of junction formation in Drosophila and mammalian epithelial cells. Biol Cell 94:305–313 44. Yamanaka T, Horikoshi Y, Suzuki A, Sugiyama Y, Kitamura K, Maniwa R, Nagai Y, Yamashita A, Hirose T, Ishikawa H, Ohno S (2001) PAR-6 regulates aPKC activity in a

novel way and mediates cell-cell contactinduced formation of the epithelial junctional complex. Genes Cells 6:721–731 45. Yamanaka T, Ohno S (2008) Role of Lgl/Dlg/ scribble in the regulation of epithelial junction, polarity and growth. Front Biosci 13:6693–6707 46. Nakagawa S, Huibregtse JM (2000) Human scribble (Vartul) is targeted for ubiquitinmediated degradation by the high-risk papillomavirus E6 proteins and the E6AP ubiquitinprotein ligase. Mol Cell Biol 20:8244–8253 47. Stephens R, Lim K, Portela M, Kvansakul M, Humbert PO, Richardson HE (2018) The scribble cell polarity module in the regulation of cell signaling in tissue development and tumorigenesis. J Mol Biol 430:3585–3612 48. Arpin-Andre C, Mesnard JM (2007) The PDZ domain-binding motif of the human T cell leukemia virus type 1 tax protein induces mislocalization of the tumor suppressor hScrib in T cells. J Biol Chem 282:33132–33141 49. Peres E, Blin J, Ricci EP, Artesi M, Hahaut V, Van den Broeke A, Corbin A, Gazzolo L, Ratner L, Jalinot P, Duc Dodon M (2018) PDZ domain-binding motif of tax sustains T-cell proliferation in HTLV-1-infected humanized mice. PLoS Pathog 14:e1006933 50. Teoh KT, Siu YL, Chan WL, Schluter MA, Liu CJ, Peiris JS, Bruzzone R, Margolis B, Nal B (2010) The SARS coronavirus E protein interacts with PALS1 and alters tight junction formation and epithelial morphogenesis. Mol Biol Cell 21:3838–3852 51. Li X, Yang H, Liu J, Schmidt MD, Gao T (2011) Scribble-mediated membrane targeting of PHLPP1 is required for its negative regulation of Akt. EMBO Rep 12:818–824 52. Cherian MA, Baydoun HH, Al-Saleem J, Shkriabai N, Kvaratskhelia M, Green P, Ratner L (2015) Akt pathway activation by human T-cell leukemia virus type 1 tax oncoprotein. J Biol Chem 290:26270–26281 53. Thomas M, Laura R, Hepner K, Guccione E, Sawyers C, Lasky L, Banks L (2002) Oncogenic human papillomavirus E6 proteins target the MAGI-2 and MAGI-3 proteins for degradation. Oncogene 21:5088–5096 54. Liu H, Golebiewski L, Dow EC, Krug RM, Javier RT, Rice AP (2010) The ESEV PDZ-binding motif of the avian influenza A virus NS1 protein protects infected cells from apoptosis by directly targeting scribble. J Virol 84:11164–11174 55. Prehaud C, Wolff N, Terrien E, Lafage M, Megret F, Babault N, Cordier F, Tan GS, Maitrepierre E, Menager P, Chopy D,

PBM-PDZ Interactions and Viral Pathogenesis Hoos S, England P, Delepierre M, Schnell MJ, Buc H, Lafon M (2010) Attenuation of rabies virulence: takeover by the cytoplasmic domain of its envelope protein. Sci Signal 3:ra5 56. Terrien E, Chaffotte A, Lafage M, Khan Z, Prehaud C, Cordier F, Simenel C, Delepierre M, Buc H, Lafon M, Wolff N (2012) Interference with the PTEN-MAST2 interaction by a viral protein leads to cellular relocalization of PTEN. Sci Signal 5:ra58 57. Seo W, Prehaud C, Khan Z, Sabeta C, Lafon M (2017) Investigation of rabies virus glycoprotein carboxyl terminus as an in vitro predictive tool of neurovirulence. A 3R approach. Microbes Infect 19:476–484 58. Seo W, Servat A, Cliquet F, Akinbowale J, Prehaud C, Lafon M, Sabeta C (2017) Comparison of G protein sequences of South African street rabies viruses showing distinct progression of the disease in a mouse model of experimental rabies. Microbes Infect 19:485–491 59. Barreda D, Sanchez-Galindo M, Lopez-FloresJ, Nava-Castro KE, Bobadilla K, SalgadoAguayo A, Santos-Mendoza T (2018) PDZ proteins are expressed and regulated in antigen-presenting cells and are targets of influenza A virus. J Leukoc Biol 103:731–738 60. Platanias LC (2005) Mechanisms of type-Iand type-II-interferon-mediated signalling. Nat Rev Immunol 5:375–386 61. Werme K, Wigerius M, Johansson M (2008) Tick-borne encephalitis virus NS5 associates with membrane protein scribble and impairs interferon-stimulated JAK-STAT signalling. Cell Microbiol 10:696–712 62. King NJ, Getts DR, Getts MT, Rana S, Shrestha B, Kesson AM (2007) Immunopathology of flavivirus infections. Immunol Cell Biol 85:33–42 63. Fernandez-Garcia MD, Mazzon M, Jacobs M, Amara A (2009) Pathogenesis of flavivirus infections: using and abusing the host cell. Cell Host Microbe 5:318–328 64. Yilla M, Harcourt BH, Hickman CJ, McGrew M, Tamin A, Goldsmith CS, Bellini WJ, Anderson LJ (2005) SARS-coronavirus replication in human peripheral monocytes/ macrophages. Virus Res 107:93–101 65. Zhou J, Chu H, Li C, Wong BH, Cheng ZS, Poon VK, Sun T, Lau CC, Wong KK, Chan JY, Chan JF, To KK, Chan KH, Zheng BJ, Yuen KY (2014) Active replication of Middle East respiratory syndrome coronavirus and aberrant induction of inflammatory cytokines and chemokines in human macrophages: implications for pathogenesis. J Infect Dis 209:1331–1342

235

66. Jimenez-Guardeno JM, Regla-Nava JA, NietoTorres JL, DeDiego ML, Castano-RodriguezC, Fernandez-Delgado R, Perlman S, Enjuanes L (2015) Identification of the mechanisms causing reversion to virulence in an attenuated SARS-CoV for the design of a genetically stable vaccine. PLoS Pathog 11:e1005215 ˜ o-Rodriguez C, Honrubia JM, Gutie´r67. Castan ´ lvarez J, DeDiego ML, Nieto-Torres JL, rez-A ˜o JM, Regla-Nava JA, Jimenez-Guarden Fernandez-Delgado R, Verdia-Ba´guena C, Queralt-Martı´n M, Kochan G, Perlman S, Aguilella VM, Sola I, Enjuanes L (2018) Role of severe acute respiratory syndrome coronavirus viroporins E, 3a, and 8a in replication and pathogenesis. mBio 9:e2325–e2317 68. Stodola JK, Dubois G, Le Coupanec A, Desforges M, Talbot PJ (2018) The OC43 human coronavirus envelope protein is critical for infectious virus production and propagation in neuronal cells and is a determinant of neurovirulence and CNS pathology. Virology 515:134–149 69. Florek D, Ehmann R, Kristen-Burmann C, Lemmermeyer T, Lochnit G, Ziebuhr J, Thiel HJ, Tekes G (2017) Identification and characterization of a Golgi retention signal in feline coronavirus accessory protein 7b. J Gen Virol 98:2017–2029 70. Gorbalenya AE, Baker SC, Baric RS, de Groot RJ, Drosten C, Gulyaeva AA, Haagmans BL, Lauber C, Leontovich AM, Neuman BW, Penzar D, Perlman S, LLM P, Samborskiy DV, Sidorov IA, Sola I, Ziebuhr J, Coronaviridae Study Group of the International Committee on Taxonomy of V (2020) The species severe acute respiratory syndrome-related coronavirus: classifying 2019-nCoV and naming it SARS-CoV-2. Nat Microbiol 5:536–544 71. Lu R, Zhao X, Li J, Niu P, Yang B, Wu H, Wang W, Song H, Huang B, Zhu N, Bi Y, Ma X, Zhan F, Wang L, Hu T, Zhou H, Hu Z, Zhou W, Zhao L, Chen J, Meng Y, Wang J, Lin Y, Yuan J, Xie Z, Ma J, Liu WJ, Wang D, Xu W, Holmes EC, Gao GF, Wu G, Chen W, Shi W, Tan W (2020) Genomic characterisation and epidemiology of 2019 novel coronavirus: implications for virus origins and receptor binding. Lancet 395:565–574 72. Lau SK, Woo PC, Li KS, Tsang AK, Fan RY, Luk HK, Cai JP, Chan KH, Zheng BJ, Wang M, Yuen KY (2015) Discovery of a novel coronavirus, China Rattus coronavirus HKU24, from Norway rats supports the murine origin of Betacoronavirus 1 and has implications for the ancestor of Betacoronavirus lineage A. J Virol 89:3076–3092

236

Carlos Castan˜o-Rodriguez et al.

73. Neuman BW, Kiss G, Kunding AH, Bhella D, Baksh MF, Connelly S, Droese B, Klaus JP, Makino S, Sawicki SG, Siddell SG, Stamou DG, Wilson IA, Kuhn P, Buchmeier MJ (2011) A structural analysis of M protein in coronavirus assembly and morphology. J Struct Biol 174:11–22 74. DeDiego ML, Pewe L, Alvarez E, Rejas MT, Perlman S, Enjuanes L (2008) Pathogenicity of severe acute respiratory coronavirus deletion mutants in hACE-2 transgenic mice. Virology 376:379–389 75. Pfefferle S, Krahling V, Ditt V, Grywna K, Muhlberger E, Drosten C (2009) Reverse genetic characterization of the natural genomic deletion in SARS-coronavirus strain Frankfurt1 open reading frame 7b reveals an attenuating function of the 7b protein in-vitro and in-vivo. Virol J 6:131 76. DeDiego ML, Nieto-Torres JL, Jimenez-Guar˜ o JM, Regla-Nava JA, Castan ˜ oden Rodriguez C, Fernandez-Delgado R, Usera F, Enjuanes L (2014) Coronavirus virulence genes with main focus on SARS-CoV envelope gene. Virus Res 194:124–137 77. Chung SH, Frese KK, Weiss RS, Prasad BV, Javier RT (2007) A new crucial protein interaction element that targets the adenovirus E4-ORF1 oncoprotein to membrane vesicles. J Virol 81:4787–4797 78. Razanskas R, Sasnauskas K (2010) Interaction of hepatitis B virus core protein with human GIPC1. Arch Virol 155:247–250 79. Hsu EC, Lin YC, Hung CS, Huang CJ, Lee MY, Yang SC, Ting LP (2007) Suppression of hepatitis B viral gene expression by proteintyrosine phosphatase PTPN3. J Biomed Sci 14:731–744 80. Yu J, Li X, Wang Y, Li B, Li H, Li Y, Zhou W, Zhang C, Wang Y, Rao Z, Bartlam M, Cao Y (2011) PDlim2 selectively interacts with the PDZ binding motif of highly pathogenic avian H5N1 influenza A virus NS1. PLoS One 6: e19511 81. Tomaic V, Gardiol D, Massimi P, Ozbun M, Myers M, Banks L (2009) Human and primate tumour viruses use PDZ binding as an evolutionarily conserved mechanism of targeting cell polarity regulators. Oncogene 28:1–8

82. Suzuki T, Uchida-Toita M, Yoshida M (1999) Tax protein of HTLV-1 inhibits CBP/p300mediated transcription by interfering with recruitment of CBP/p300 onto DNA element of E-box or p53 binding site. Oncogene 18:4137–4143 83. Wilson KC, Center DM, Cruikshank WW, Zhang Y (2003) Binding of HTLV-1 tax oncoprotein to the precursor of interleukin-16, a T cell PDZ domain-containing protein. Virology 306:60–67 84. Ohashi M, Sakurai M, Higuchi M, Mori N, Fukushi M, Oie M, Coffey RJ, Yoshiura K, Tanaka Y, Uchiyama M, Hatanaka M, Fujii M (2004) Human T-cell leukemia virus type 1 tax oncoprotein induces and interacts with a multiPDZ domain protein, MAGI-3. Virology 320:52–62 85. Song C, Wang W, Li M, Liu Y, Zheng D (2009) Tax1 enhances cancer cell proliferation via Ras-Raf-MEK-ERK signaling pathway. IUBMB Life 61:685–692 86. Makokha GN, Takahashi M, Higuchi M, Saito S, Tanaka Y, Fujii M (2013) Human T-cell leukemia virus type 1 tax protein interacts with and mislocalizes the PDZ domain protein MAGI-1. Cancer Sci 104:313–320 87. Blot V, Delamarre L, Perugi F, Pham D, Benichou S, Benarous R, Hanada T, Chishti AH, Dokhelar MC, Pique C (2004) Human Dlg protein binds to the envelope glycoproteins of human T-cell leukemia virus type 1 and regulates envelope mediated cell-cell fusion in T lymphocytes. J Cell Sci 117:3983–3993 88. Chen Q, Niu X, Xu Y, Wu J, Shi Y (2007) Solution structure and backbone dynamics of the AF-6 PDZ domain/Bcr peptide complex. Protein Sci 16:1053–1062 89. Belotti E, Polanowska J, Daulat AM, Audebert S, Thome V, Lissitzky JC, Lembo F, Blibek K, Omi S, Lenfant N, Gangar A, Montcouquiol M, Santoni MJ, Sebbagh M, Aurrand-Lions M, Angers S, Kodjabachian L, Reboul J, Borg JP (2013) The human PDZome: a gateway to PSD95-disc largezonula occludens (PDZ)-mediated functions. Mol Cell Proteomics 12:2587–2603

Chapter 14 Computational Design of PDZ-Peptide Binding Nicolas Panel, Francesco Villa, Vaitea Opuu, David Mignon, and Thomas Simonson Abstract This chapter describes two computational methods for PDZ–peptide binding: high-throughput computational protein design (CPD) and a medium-throughput approach combining molecular dynamics for conformational sampling with a Poisson–Boltzmann (PB) Linear Interaction Energy for scoring. A new CPD method is outlined, which uses adaptive Monte Carlo simulations to efficiently sample peptide variants that tightly bind a PDZ domain, and provides at the same time precise estimates of their relative binding free energies. A detailed protocol is described based on the Proteus CPD software. The mediumthroughput approach can be performed with standard MD and PB software, such as NAMD and Charmm. For 40 complexes between Tiam1 and peptide ligands, it gave high a2ccuracy, with mean errors of around 0.5 kcal/mol for relative binding free energies and no large errors. It requires a moderate amount of parameter fitting before it can be applied, and its transferability to other protein families is still untested. Key words Protein design, Ligand binding, MC simulation, Proteus program, Molecular mechanics, Implicit solvent

1

Introduction We focus here on the design of PDZ–peptide binding with computational approaches. One goal is to discover peptide ligands that could inhibit or modulate the activity of a given PDZ protein. For this, one should explore a space of peptide variants, perhaps allowing noncanonical amino acids (ncAAs) at selected positions, for binding or stability. Another goal is to redesign the PDZ domain itself, to alter its target binding and manipulate protein interaction networks in vitro or in cells. For this, one would explore a space of protein variants, where a few positions close to the peptide are allowed to mutate. Both applications should potentially consider a large space of sequences, so that medium or high-throughput approaches are desirable. Both applications involve designing a polypeptide, and so both are amenable to high-throughput, structure-based, computational protein design (CPD). CPD aims to

Jean-Paul Borg (ed.), PDZ Mediated Interactions: Methods and Protocols, Methods in Molecular Biology, vol. 2256, https://doi.org/10.1007/978-1-0716-1166-1_14, © Springer Science+Business Media, LLC, part of Springer Nature 2021

237

238

Nicolas Panel et al.

engineer proteins (and ligands) and optimize molecular properties such as binding affinity and binding specificity [1–4]. It can explore thousands of polypeptide sequences in 1–2 days using a desktop computer. PDZ complexes can also be studied with mediumthroughput approaches that combine molecular dynamics (MD) for conformational exploration with a simplified free energy function for binding affinities. These approaches [5–15] can explore a few dozen sequences in a few days using a few GPU computers or a small Linux cluster. Here, we outline both approaches and illustrate them by our recent work on the Tiam1 PDZ domain. In a typical design project, one might run a high-throughput calculation first, then characterize the top sequences with the medium-throughput approach. The latter step might lead to additional high-throughput calculations, perhaps involving additional mutating positions, and so on, in a series of design cycles that would hopefully include experimental steps. The illustrations below include high-throughput CPD protocols that can be run with our Proteus software [16, 17]. Our medium-throughput approach uses a free energy function that includes a Poisson–Boltzmann (PB) contribution, a surface area (SA) contribution, and a van der Waals (vdW) contribution. It can be run with standard MD and PB software. 1.1

General Issues

When engineering several amino acid positions, the mutation space grows exponentially with the number of positions. If just three positions are explored combinatorially, there are 8000 possible sequences (more, if ncAAs are allowed). Each sequence variant can occupy a vast number of conformations. Even if we use a model where the protein backbone is held fixed, there are thousands of rotamer combinations for the side chains at the three mutating side positions, and millions if we consider the whole binding interface. Another difficulty is that to engineer PDZ–peptide binding, we should consider both the bound and unbound states of each partner. If we engineer the protein, its stability should be maintained, so that the unfolded state will also play a role. If we care about specificity, we may have to design against off-target binding, which would involve additional proteins and states. To address all these issues, major approximations are needed for both conformational exploration and scoring. The first common approximation is to use a molecular mechanics description of the protein and peptide. Although some CPD tools like Rosetta use knowledge-based energy functions [18, 19], many studies have shown that molecular mechanics is a good approximation for many design problems. The second common approximation is to model solvent implicitly, as in continuum electrostatics. Formally, this can be viewed as an averaging operation over solvent configurations [20]. This leads to the concept of a potential of mean force

Computational Design of Binding

239

[9, 21]. Implicit solvent models have been actively developed for many years, both for medium and high-throughput applications. They usually involve a continuum electrostatic component, such as a Poisson–Boltzmann (PB) or Generalized Born (GB) energy term. For high-throughput CPD, additional approximations are usually necessary, outlined below. 1.2 PDZ–Peptide Issues

PDZ–peptide binding presents specific difficulties. First, the unbound peptide is quite flexible, and it is challenging to explore its motions and quantify the effect of mutations on its flexibility and entropy. Second, the binding interface is large and the affinity arises from many small contributions, which should be accurately captured. Third, with such a large interface, many residues undergo extensive burial upon binding, changing from a solvent-rich to a solvent-poor environment. This change often leads to electronic polarization, which is still a difficulty for molecular mechanics models. Fourth, peptide phosphorylation regulates binding in some cases, and phosphate–protein binding is also challenging to model. Fifth, the use of ncAAs to enhance binding or peptide stability means they must be part of the molecular mechanics model. This often means a specific extension of the model is needed.

1.3 Chapter Overview

We first outline our high-throughput CPD approach. We describe only briefly the technical details, which are available in published articles [16, 22]. We describe a protocol that allows us to select peptide variants based on their relative binding free energies, which is of great interest. The approach is based on an adaptive Monte Carlo method [22]. We present illustrative results for the Tiam1 PDZ domain. Next, we describe our medium-throughput approach [23], which is readily applied to dozens of peptide or protein variants (or a few hundred with more resources; roughly one variant per day and per GPU card). The complex is simulated with MD and explicit solvent. Then the binding free energy is computed with a free energy function that combines a Poisson– Boltzmann description of solutes and solvent, along with additional nonpolar SA and vdW free energy terms. We include selected results for the Tiam1 PDZ domain binding to a collection of peptides [23], some of which were taken from CPD predictions. Figure 1 shows Tiam1 bound to the Syndecan-1 peptide (Sdc1), which corresponds to the C-terminus of its natural target protein.

2

High-Throughput Design of PDZ–Peptide Binding Many successful CPD examples have been reported in recent years [3, 24–30]. Many were obtained with energy functions that included knowledge-based terms, such as those in the Rosetta energy function. Proteus, on the other hand, relies on “physics-

240

Nicolas Panel et al.

Fig. 1 The complex between the Tiam1 PDZ domain and the Sdc1 peptide (cross-eyed stereo). The peptide is yellow; its residues are labeled with their type. The Sdc1 sequence is 7TKQEEFYA0

based” CPD. One of its characteristics is the energy function drawn from molecular mechanics. The other is its sampling approach, which uses adaptive Monte Carlo to target the binding free energy, as opposed to simpler properties like the bound-state energy. The CPD methods described below are available (and further documented) in the recently released Proteus 3.0 package (https://proteus. polytechnique.fr) [17]. 2.1 Model Ingredients and System Setup

We use a molecular mechanics energy function along with an implicit solvent model that contains a Generalized Born (GB) term and a nonpolar term:

2.1.1 Energy Model

E ¼ E bond þ E angle þ E dihedral þ E improper þ E vdW þ E Coulomb þ E GB þ E NP ð1Þ The first six terms describe the internal and nonbonded contributions to the potential energy of the protein or peptide, and are borrowed from the Amber ff99SB molecular mechanics energy function [31]. The next two terms capture solvent effects via a GB approximation for electrostatic effects and a nonpolar term. This can be either an accessible surface area (SA) term or a Lazaridis–Karplus (LK) term [32].

2.1.2 Structures

We start from an X-ray complex between the Tiam1 PDZ domain (called “Tiam1”) and the Syndecan-1 octapeptide, Sdc1 [33]. Hydrogen positions are added and a slight energy minimization is done (200 steps) to remove poor steric contacts. To model the unbound peptide, the protein atoms are removed. We assume a few amino acid positions on the peptide are to be redesigned. We refer to these as “active” positions. All other peptide and protein positions are “inactive”: they will explore rotamers but not mutate.

Computational Design of Binding

241

At all inactive positions, rotamers from a library are positioned. We use the Tuffery rotamer library, slightly extended [34, 35]. It has about ten rotamers per amino acid type. At the active positions, we position all the rotamers for all the allowed amino acid types, around 200 if all types are allowed. In fact, we do not mutate into Gly or Pro, so there are 18 possible types. With four active positions, there are 104,976 possible sequences. At this point, each residue in the system has around 10 rotamer conformations in place, or 200 if it is active. All these are recapitulated in a single PDB file. The protein and peptide backbone conformations will be held fixed. Backbone flexibility will be modeled implicitly, through the protein dielectric constant. One can also perform several design calculations, each with a slightly different backbone conformation, or a single, multibackbone calculation. This last protocol is significantly more complicated and expensive [36], and is not described here. 2.1.3 Unfolded Protein State

If we allow mutations (active positions) in the protein, we will need to evaluate their effect on stability and eliminate highly destabilizing mutations. We model the unfolded state through an energy function that is a sum over the amino acid positions [16, 37]: X uf E uf ¼ E I ðt I Þ ð2Þ I

uf

The sum is over all amino acids I. Each term E I ðt I Þ depends on the residue type tI. There are no contributions from interresidue interactions, meaning that in the unfolded state, each residue interacts with solvent and nearby backbone groups but not the surrounding side chains. This seems reasonable for a fully extended uf polypeptide chain. In practice, the individual contributions E I ðt I Þ are computed from the folded structure as follows: we compute the energy of each side chain, including its interactions with solvent, its own backbone, and the backbone of the previous and following residues. This rather crude model is used below to filter out highly destabilizing mutations. Note that for some other applications, uf especially whole protein redesign, the E I ðt I Þ must be chosen more carefully, and normally require a complex empirical optimization [37]. 2.1.4 Energy Matrix

We precompute and store in a matrix the interaction energies between all residue pairs, taking into account all residue types (active positions) and rotamers (active or inactive). This calculation is done by the protX program, through a library of command scripts, using the energy function of Eq. (1). Given a particular set of mutations and rotamers, the energy can then be obtained by summing contributions from this matrix, with one important caveat. The solvation terms (GB, especially) are not rigorously

242

Nicolas Panel et al.

pairwise-additive: they are not a sum over residue pairs. Specifically, the GB model assigns to each atom i a “solvation radius” bi, which depends on the conformation of the whole system [9]. To overcome this, two methods are available. We can adopt a “Native Environment Approximation” (NEA), where each solvation radius is computed for the wildtype or native structure, then kept fixed [38]. Or, we can adopt a modified GB model, which computes more information and stores it in the energy matrix, then requires additional operations during the subsequent MC simulations. This variant is called the Fluctuating Dielectric Boundary method [39]. 2.1.5 Monte Carlo Simulations to Explore Sequences and Structures

The energy matrix is read by the C module protMC, which explores the space of sequences and structures. Three exploration methods are available; only Monte Carlo (MC) is considered here [40, 41]. MC can use a single “replica”, exploring a single trajectory. Or it can use multiple replicas, usually 4–8, with distinct temperatures, which occasionally exchange their temperatures. The method is known as “Replica Exchange” MC, or REMC. All the methods output multiple “snapshots”, sampled along the MC trajectory. MC moves correspond to rotamer changes at one or two positions, mutations at one or two positions, or a rotamer change at one position and a mutation at another. The REMC temperatures usually range from about 50 to 1000 K, with the very hot replica easily moving among local energy minima for effective sampling. Adaptive MC is also possible, as outlined in the next section.

2.1.6 Proteus Software Files and Documentation

Proteus 3.0 is freely available to academic and government scientists, from https://proteus.polytechnique.fr. Industrial scientists should contact the corresponding author. The distribution includes source code, binaries for Intel processors, extensive test cases, and detailed documentation.

2.2 Adaptive Landscape Flattening to Design PDZ–Peptide Binding Affinity

To design ligand binding means optimizing a free energy difference between bound and unbound states. This is not tractable by most CPD methods, such as simulated annealing or plain Monte Carlo (MC). Most studies have used heuristic methods that optimize the bound state energy, a very different property. Recently, a new method was proposed, using MC simulations and adaptive importance sampling. The energy landscape in sequence space is flattened adaptively for the unbound state, over the course of an MC simulation, thanks to a bias potential [22, 42]. The bias B is constructed such that all sequences reach comparable populations. B is then essentially the sequence free energy with its sign changed. Next, the bias is included in a simulation of the complex, where it “subtracts out” the unbound state. Thus, negative design of the unbound state is achieved. Remarkably, the result is a Boltzmann distribution where sequence populations measure their affinities.

2.2.1 General Method

Computational Design of Binding 2.2.2 Stage 1: Flattening the Unbound State

243

MC will be run for the unbound state with the protMC program, controlled by a configuration file adapt.conf. Since we plan to mutate the peptide, the unbound simulation will correspond to the peptide alone. If we wanted to mutate the protein, we would simulate the protein alone. The file adapt.conf indicates which mutations are allowed for positions that are active (four peptide positions in our application [22]). During the adaptation, it is best uf to include reasonable values for the unfolded energies E I ðt I Þ (Eq. 2), which are readily obtained with Proteus, based on the extended peptide picture (above). Thus, adapt.conf includes lines such as the following:

ALA 7.54 ARG -52.58 Etc

Adapt.conf also contains information that controls the form of the bias potential and its update schedule. Full details are in the Proteus manual (https://proteus.polytechnique.fr). At this point, we run protMC. protMC.exe < adapt.conf > adapt.log

Output files are: l

bias.dat: evolution of the bias during the MC trajectory,

l

proteus_adapt.seq: visited sequences,

l

output.ener: the energy of visited sequences.

At this point, we copy the final bias from bias.dat to a new file, bias. in. 2.2.3 Stage 2: Simulating the Bound State

The next step is to run an MC simulation of the complex, including the bias potential (which effectively subtracts out the unbound state). Thus, simulating the complex with the apo bias will now lead to peptide sequences that are populated according to their Tiam1 binding free energy (sic). The MC simulation is controlled by a file similar to adapt.conf above. No adaptation is done; rather the obtained bias is made available in the file bias.in. Once the simulation is done, affinity-based sampling is finished. Sequence populations will now lead directly to binding affinities. For two sequences s and r sampled in both states, we denote p’s, p’r the biased holo populations and ps, pr the biased apo populations (with the same bias). We can obtain the binding free energy difference as

244

Nicolas Panel et al.

ΔG s  ΔG r ¼ kT ln

p0s p þ kT ln s p0r pr

ð3Þ

Extraction of the populations from the MC simulations and calculation of the free energy is done by Proteus with a python script. For sequences of interest, such as the tightest binders, 3D structure models can be computed from the rotamer information with a bash script. 2.2.4 Application to the Tiam1–Sdc1 Complex

3

An application to the Tiam1–Sdc1 complex was reported recently [22]. Four of the last five peptide positions were allowed to mutate into all types except Gly or Pro. The position numbers were 4 to 1, following the usual “backward” convention for PDZ binders. The C-terminal position (position 0) was kept as in the wild-type peptide (Ala), because the Sdc1 backbone arrangement does not allow large side chains at this position. There were 104,976 possible sequences. Thanks to the adaptive method, a large fraction were sampled. For nine variants, relative binding free energies were available from experiment or high-level, alchemical MD free energy simulations. Excluding one large error, the mean unsigned errors from eight variants was 0.8 kcal/mol. Figure 2 shows the sequences sampled, in the form of a sequence logo, with populations given by their relative binding free energies. The logo is compared to one that represents an experimental library of Tiam1-binding peptides [43]. Positions P0 and P1 are the most important for PDZ binding specificity. Position P0 occupies three main types experimentally, C, A, and F, but was held fixed during the simulations. Of the four positions allowed to mutate, P1, P3, and P4 are highly variable in both the MC and the experimental logos. Of the top ten MC types at these positions, 7 or 8 are present in the experimental logo, and vice versa, with somewhat different occupancies. Position P2 is more conserved, both experimentally and in the simulations. Of the top four experimental types, Y, F, M, T, all but T are in the top five MC types. While T has less than 1% occupancy in the MC sequences, the chemically similar types A, C, and S are highly populated. Overall, the two logos are in reasonable agreement.

A Medium-Throughput Design Approach After high-throughput design, one can use a more costly model to characterize a few dozen of the top CPD candidates with increased accuracy [44]. The model uses MD with explicit solvent to sample conformations, then scores them with a free energy function where solvent is modeled implicitly. MD is done for the PDZ–peptide complex, typically for 80–100 ns. Several hundred snapshots are taken from the trajectory. The unbound state is modeled by using the same snapshots and simply separating the two partners. We refer to this as a single-trajectory approach. A two-trajectory variant that

245

Adaptive MC

Experimental

Computational Design of Binding

Fig. 2 Top: Sequence logo from an experimental library of peptides that bind Tiam1 [43]. Each column corresponds to a peptide position; P0 is the C-terminus; the last five positions are shown. Types have heights proportional to their abundancy. Bottom: Logo from the MC simulation of the Tiam1–peptide complex, where sequences are populated by affinity

uses separate simulations of the unbound peptide is also described. The free energy function uses implicit solvent, which includes a Poisson–Boltzmann electrostatic term plus a Surface Area (SA) and a van der Waals (vdW) term. Taking the free energy difference between the bound and unbound structures and averaging over all the snapshots we obtain a binding free energy. Obviously, some free energy components are not accounted for, such as translational entropy. But they are expected to cancel out when we compare several peptide (or protein) variants. The free energy function requires some parameterization and testing before it can be applied. Here, we outline the method, its parameterization using experimental binding free energies and its application to designed Sdc1 sequences. The method is also schematized in Fig. 3. 3.1 Explicit-Solvent MD to Characterize PDZ–Peptide Complexes

MD is done with explicit solvent and only a few restraints on the complex. Thus, the fixed backbone approximation from the CPD stage is removed. Simulations lengths of 50–100 ns are used, which allows the free energy function to converge. It also gives time for the backbone to rearrange, if necessary. If, for a predicted variant, the backbone rearranges significantly and/or the peptide starts to detach, it suggests the particular complex is unstable and may invalidate the CPD prediction.

246

Nicolas Panel et al.

Fig. 3 Medium-throughput protocol with MD sampling and PB/LIE scoring. The complex is simulated with explicit solvent; snapshots are extracted; solvent positions are discarded. The unbound state is obtained from each snapshot by removing one partner or the other. The snapshots are processed by a continuum model to obtain the electrostatic free energy component; surface areas and the van der Waals interaction are computed to obtain the nonpolar components. The binding free energy ΔG is a weighted sum of the components, timeaveraged (brackets) [44] 3.1.1 Conformational Restriction of the Peptide N-Terminus

The peptide positions most important for binding are near its C-terminus. A difficulty is that with a short, 8-residue peptide, the N-terminus has large fluctuations, detaching and reattaching on a 10 ns timescale. These fluctuations cannot be adequately sampled with 100 ns MD runs. We expect they are not too dependent on mutations near the C-terminus and so it is better to suppress them. Thus, we introduced weak, “flat-bottomed” restraints at the peptide N-terminus, which only acted if the first two residues moved more than 1 Å away from their starting position, as measured by their contact distances to the PDZ domain.

3.1.2 Force Field and MD Simulations

We used the Amber force field, consistent with the highthroughput CPD stage. For PDZ domains, it is important to use the recent ff14SB release, which gives improved helix backbone structures. We used the TIP3P water model, which has been extensively tested in combination with the Amber force field. The combination gives excellent agreement with available X-ray structures

Computational Design of Binding

247

for the overall structure during microsecond MD runs, and with experimental backbone amide order parameters, which measure conformational fluctuations. We modeled PDZ–peptide complexes based on X-ray structures involving four peptides: Sdc1, Caspr4, Neurexin, and a “consensus” peptide from a combinatorial peptide library [45]. The peptides were bound to either the wild-type Tiam1 PDZ domain (WT) or a variant containing four amino acid changes (quadruple mutant or QM). The four complexes were WT–Sdc1 (PDB 4GVD) [33], WT–consensus (PDB 3KZE) [46], QM–Caspr4 (PDB 4NXQ) [47], and QM–Neurexin (PDB 4NXR) [47]. Our truncated icosahedral simulation box included around 10,900 water molecules. Each PDZ–peptide variant was prepared by immersing it in the same solvent box, deleting overlapping waters, and running about 500 ps of equilibration at room temperature with gradually decreasing harmonic restraints, initially on all solute atoms and finally on only backbone Cα’s. Structure preparation was done with the protX module of Proteus, while equilibration was done with NAMD. Other programs can also be used, such as Charmm, Amber, or Gromacs. MD production was run for 40 ns, then continued until the free energy function converged or 100 ns were reached. This took up to 2 days on a single GPU processor. 3.2 Relative Binding Free Energies 3.2.1 The Free Energy Function

To obtain the binding free energy estimate ΔG, we used the following ansatz for the free energy: ΔG ¼ αhΔE vdW i þ βhΔG elec i þ γ hΔA i þ δ

ð4Þ

Here, α, β, and γ are adjustable constants. ΔGelec is an electrostatic free energy difference between the bound and unbound states, computed with a PB model. Brackets indicate averaging over the structural snapshots taken at regular intervals along the MD trajectory of the solvated complex. To represent the unbound state, we took each snapshot from the MD trajectory of the complex and simply moved the protein and the peptide apart. Thus, the same snapshot is used for all three solutes. ΔA is the change in the solute molecular surface upon binding (which is negative). ΔEvdW is the van der Waals interaction energy between the protein and the peptide. Solute–solvent and solvent–solvent van der Waals contributions are not explicitly included. PB calculations were done with the Charmm program, while the SA and vdW terms were computed with protX. The solute dielectric constant ϵ S was set to 8. The last term, δ, is a constant that vanishes when we consider the relative binding free energies ΔΔG of the various complexes. We refer to the free energy ansatz as a PB/LIE free energy, for PB Linear Interaction Energy, since Eq. (4) is a weighted sum of interactions.

248

Nicolas Panel et al.

3.2.2 Fitting to Experimental Binding Free Energies

Experimental binding free energies were available for 44 complexes [44]. Four very weak binders were excluded. To obtain relative values, the free energies were compared to either the WT:Sdc1 value or the QM:Caspr4 value, depending on which X-ray structure was used to model each variant. This left 38 independent ΔG values. Three involving the neurexin peptide were left aside and the remaining 35 were used to fit the adjustable constants α, β, and γ. Extensive cross-validation was done, where subsets of the data were excluded from the fit; this gave similar fitted values and errors.

3.3

For the 35 complexes with experimental ΔG values, the mean unsigned and rms errors were 0.43 and 0.55 kcal/mol. The Pearson correlation between the experimental and computed values was R ¼ 0.64. The three largest errors were 1.31, 1.13, and 1.09 kcal/ mol and included two Caspr4 complexes. Thus, there were no very large errors and the mean error is very small, and corresponds to chemical accuracy (less than the thermal energy kT).

Selected Results

3.3.1 Mean Errors

3.3.2 Scoring Sequences from CPD

From the top 30 peptide variants predicted by CPD to be the tightest binders, we chose 14 for medium-throughput study. Results (reported here) are given in Table 1. The adaptive MC results are also given. MC predicted that several variants should bind more strongly than Sdc1. PB/LIE, on the other hand, predicts that 12 of the variants bind less strongly, with ΔGs that are between 0.1 and 0.6 kcal/mol less favorable than Sdc1. Two are predicted to bind as strongly as Sdc1. None show improved binding. Since the mean PB/LIE error is around 0.5 kcal/mol, we conclude that at best, the high-throughput design may have produced a few peptides with weakly improved affinities. It may be possible to improve binding by designing other peptide positions or allowing ncAAs.

3.4 Other Variants of the Model

Claims are sometimes made that PB is a better electrostatic model than GB for binding and other properties. Here, switching to GB gave very slight increases of mue and rmse, to 0.55 and 0.66 kcal/ mol, respectively. Notice that we used GB for CPD.

3.4.1 GB Instead of PB 3.4.2 Lazaridis–Karplus Instead of SA

The Lazaridis–Karplus solvent model was parameterized elsewhere for nonpolar solvation effects [32]. It was applied to the PDZ– peptide complexes without any reparameterization, in combination with the GB electrostatic term. The mue and rmse values were 0.59 and 0.69 kcal/mol, respectively, almost the same as with GB + SA.

3.4.3 Two-Trajectory Model for Peptide Flexibility

The flexibility of the unbound peptide is in principle sequencedependent, which could influence relative affinities. Therefore, we also considered a model where the binding process is divided into two steps. In the first, we apply restraints to the peptide, forcing it to be close to its bound conformation; in the second it binds

Computational Design of Binding

249

Table 1 Relative Tiam1 binding free energies ΔΔG of designed peptides from PB/LIE ΔΔG ΔG

Sequencea

PBb

vdWb

SAb

c

c

ETMNA

3.25

48.10

1310.1

3.46

0.05

1.2

EEMNA

2.99

49.48

1307.6

3.49

0.02

1.1

EEFYA

2.84

58.67

1349.3

3.51

0.00

0.0

YECEA

3.01

55.62

1371.0

3.61

0.11

1.3

YTCDA

2.26

50.67

1313.2

3.77

0.16

1.3

ESMTA

2.24

48.65

1305.9

3.70

0.18

1.2

YSCDA

1.88

49.70

1295.0

3.71

0.20

1.3

ETMTA

2.50

49.47

1331.0

3.71

0.20

1.1

ETMEA

1.06

46.81

1247.8

3.79

0.28

1.1

EEMSA

1.64

49.27

1329.5

3.92

0.41

1.2

YNCTA

+1.20

38.10

1111.0

3.98

0.47

1.3

YTCVA

1.87

57.31

1398.6

3.98

0.47

1.3

YTCTA

0.93

50.07

1321.8

4.05

0.54

1.5

YNCYA

1.56

57.28

1401.7

4.07

0.56

1.3

YSCTA

0.46

49.27

1311.0

4.14

0.63

1.4

PB/LIE

d

Proteus

PB/LIE binding free energies ΔG (Eq. 4) for selected peptide sequences, in kcal/mol. ΔΔG values are relative to Sdc1 (in bold, “EEFYA”) a The five C-terminal amino acids are listed. The N-terminal ones are always TKQ b Specific ΔG contributions, in kcal/mol or Å2 (SA) c Component weights (Eq. 4) were α ¼ 0.02 (vdw), β ¼ 0.25 (elec), and γ ¼ 4 cal/mol/Å2 d CPD values (adaptive MC with Proteus)

Tiam1. The first free energy contribution is deduced from the fraction of bound conformations in MD simulations of the unbound peptide. The second is computed from the PB/LIE model as before. Hence, two trajectories are used (per peptide variant). In practice, the first contribution is small and noisy, and several 100 ns are needed for convergence. Its sequencedependence is weak. With this model, the rms error did not improve. 3.4.4 Three-Trajectory Model

For two complexes, WT:Sdc1 and QM:Caspr4, we performed much longer MD simulations and applied a 3-trajectory method to the PB free energy component. The unbound peptides were simulated for 400 ns, the two complexes for 500 ns, and the unbound proteins for 1000 ns each. The binding free energy difference between the two variants was in excellent agreement with the single-trajectory value.

250

Nicolas Panel et al.

3.4.5 Comparison to some PBSA or GBSA Approaches Applied to Other Systems

4

Table 2 summarizes the performance of some earlier PBSA and GBSA approaches. Several achieve high correlations with experiments. Only five outperform the Null model and only one achieves (for a much smaller dataset) the small errors reported here.

Concluding Notes We conclude with a series of practical observations, or “lessons learned” that we have derived from our work and that should help readers apply the methods presented above to related systems: 1. High-throughput CPD of the PDZ protein for binding can be handled in the same way as the peptide design illustrated in Subheading 2. A slightly simpler approach was reported earlier [37] that did not employ (yet-undiscovered) adaptive MC. 2. We recently reported redesign of the entire PDZ protein, the first successful whole-protein design with a nonempirical, physics-based energy function [60]. 3. A key to the accurate PB/LIE predictions described in Subheading 3 was the use of the Eq. (4) empirical ansatz. A pure PBSA or GBSA free energy is unlikely to succeed without empirical weights. 4. The solute dielectric constant ϵ S for the PB component (Eq. 4 in Subheading 3) is an empirical parameter. Choosing a value of 8 is physically reasonable [61]. Choosing a different value (such as 4) led to a nearly proportional change in the PB weight β (Eq. 4) and thus similar results. 5. Another key to PB/LIE success was good conformational sampling, thanks to moderately long MD runs and suppression of slow (evidently unimportant) N-terminal fluctuations of the peptide. 6. Our PB/LIE model (Subheading 3) made no attempt to model contributions from conformational entropy changes; we assume normal mode or quasi-harmonic models would not be predictive for our systems and MD trajectory lengths. 7. The PB/LIE model could not handle very weak binders, which were left out of the fit. Despite using MD, it also could not handle the conformational changes between Sdc1-like peptides and Cask-like peptides, which have distinctly different backbone arrangements. Therefore, these two data sets each had their own reference complex [23]. 8. The tradeoff for the excellent PB/LIE accuracy (Subheading 3.3) is the need to fit α, β, γ. Transferability to other PDZ domains should be good but was not yet tested.

Target

DHFR

DHFR

Alpha-thrombin 7

Avidin

Cytochrome C

Neuraminidase

P450cam

Penicillopepsin

HIV-I protease

FXa

Hsp90

[50]

[50]

[51]

[51]

[51]

[51]

[51]

[51]

[52]

[52]

[52]

RNA aptamer

RNA aptamer

[55]

MMP-2

[54]

Ligands: RNA

[53]

Ligands: peptidomimetics

HIV-I protease

[49]

6

5

8

16

20

20

7

9

8

18

7

22

22

12

Avidin

9

PB

PB

PB

PB

PB

PB

PB

PB

PB

PB

PB

PB

GB

PB

PB

PB

?

1

1

1

1

1

2

1

4

1

1

4

1

1

?

1

Sample size Elec. model ϵ S

[48]

Ligands: small molecules

Reference

Table 2 Performance of PBSA and GBSA in selected studies

NM

NM

NM

NM

NM

NM

NM

NM

NM

NM

SA

SA

PB, vdW, SA

NM

–

NM

PB, vdW, SA, TS NM

PB, vdW, SA, TS NM

PB, vdW, SA, TS NM

SA

SA

SA

SA

SA

SA

SA

SA

SA

SA

3.0 1.2 0.8 1.4

[12.8; 6.8] 2.8 [16.5; 9.9] 1.0 [12.8; 8.6] 0.6 [10.5; 5.1] 1.2

3.8 6.6

3.5 [11.5; 6.0] 6.5

[1.1; 5.6]

3.6

6.1

5.8

[7.9; 5.5]

[15.0; 7.8] 3.0

2.9

1.7

1.5

2.2

1.8

1.5

2.0

2.3

0.9

2.6

0.8

(continued)

0.75 1.2

0.66 1.2

0.84 1.9

0.63 1.6

0.54 0.8

0.72 1.8

0.41 2.1

0.68 0.8

0.68 2.2

11.9 12.4 0.30 0.7 [11.5; 3.7] 2.3

[7.1; 3.8]

5.3

[20.4; 4.5] 7.8

0.93 4.7

2.6

[12.4; 4.0] 19.8 20.7 0.80 2.1 8.1

2.7

1.3

[15.0; 5.5] 16.4 18.0 0.94 2.4

8.0

[15.0; 5.5] 7.2

0.74 1.1

4.9

2.7

1.2

[11.9; 7.8] 1.1

0.96 4.1

mue rmse

__Nulld__

0.91 2.4

3.8

mue rmse R

Performance

[5.0;20.4] 3.3

ΔG a Weighted terms bEntropy method rangec

Computational Design of Binding 251

Target

HIV-I protease

HIV-1 gp41

AIRE-PHD1

[57]

[58]

[59]

35

9

29

4

14

PB

PB

PB

PB

PB SA

SA

8

1 PB, vdW, SA

SA

var SA

1

1

Sample size Elec. model ϵ S

–

NM

QH

NM

NM

2.0

[0.6; 1.3]

[0.2; 2.1]

[2.4; 1.1]

0.4

6.7

–

0.6

7.7

–

0.6

2.3

0.64 0.4

0.74 0.5

0.71 –

0.96 1.0

0.5

0.6

–

1.2

0.9

mue rmse

__Nulld__

0.82 0.8

mue rmse R

Performance

[13.8;-10.6] 0.5

[0.8; 2.6]

ΔG a Weighted terms bEntropy method rangec

Free energies in kcal/mol a Terms in the free energy ansatz b Method used to estimate conformational entropy contributions: normal mode (NM) or quasi-harmonic (QH) approximation c Range of experimental binding free energies d Null model (all binding free energies equal to the experimental mean)

Our work [23] Tiam1

Abl-SH3

[56]

Ligands: peptides

Reference

Table 2 (continued)

252 Nicolas Panel et al.

Computational Design of Binding

253

9. A third, low-throughput approach (not described above) introduces mutations alchemically, one or two at a time, during explicit solvent MD. It does not require any parameter adjustment. It was applied to Tiam1–peptide binding and gave even higher accuracy [44]. References 1. Malisi C, Schumann M, Toussaint NC et al (2012) Binding pocket optimization by computational protein design. PLoS One 7: e52505–e52505 2. Feldmeier K, Ho¨cker B (2013) Computational protein design of ligand binding and catalysis. Curr Opin Chem Biol 17:929–933 3. Tinberg CE, Khare SD, Dou J et al (2013) Computational design of ligand-binding proteins with high affinity and selectivity. Nature 501:212–216 4. Stoddard BL (ed) (2016) Methods in molecular biology: design and creation of ligand binding proteins. Springer Verlag, New York 5. Srinivasan J, Cheatham TE, Cieplak P et al (1998) Continuum solvent studies of the stability of DNA, RNA, and phosphoramidateDNA helices. J Am Chem Soc 120 (37):9401–9409 6. Jorgensen WL (2004) The many roles of computation in drug discovery. Science 303:1813–1818 ¨ sterberg F, Almlo¨f M et al 7. Brandsdal BO, O (2003) Free energy calculations and ligand binding. Adv Protein Chem 66:123–158 8. Carlsson J, Ande´r M, Nervall M et al (2006) Continuum solvation models in the linear interaction energy method. J Phys Chem B 110:12034–12041 9. Simonson T (2007) Free energy calculations: approximate methods for biological macromolecules. Springer Ser Chem Phys 86:423–461 10. Gallicchio E, Levy RM (2011) Recent theoretical and computational advances for modeling protein-ligand binding affinities. Adv Protein Chem Struct Biol 85:27–80 11. Simonson T (2013) Protein: ligand recognition: simple models for electrostatic effects. Curr Pharm Des 19:4241–4256 12. Harris RC, Mackoy T, Fenley MO (2015) Problems of robustness in Poisson-Boltzmann binding free energies. J Chem Theory Comput 11:705–712 13. Wang C, Nguyen PH, Pham K et al (2016) Calculating protein-ligand binding affinities with MMPBSA: method and error analysis. J Comput Chem 37:2436–2446

14. Chakavorty A, Li L, Alexov E (2016) Electrostatic component of binding energy: interpreting predictions from Poisson–Boltzmann equation and modeling protocols. J Comput Chem 37:2495–2507 15. Katkova EV, Onufriev AV, Aguilar B et al (2017) Accuracy comparison of several common implicit solvent models and their implementations in the context of protein-ligand binding. J Mol Graph Model 72:70–80 16. Simonson T, Gaillard T, Mignon D et al (2013) Computational protein design: the proteus software and selected applications. J Comput Chem 34:2472–2484 17. Simonson T (2019) The Proteus software for computational protein design, Ecole Polytechnique, Paris. https://proteus.polytechnique.fr 18. Li Z, Yang Y, Zhan J et al (2013) Energy functions in de novo protein design: current challenges and future prospects. Annu Rev Biophys 42:315–335 19. Alford RF, Leaver-Fay A, Jeliazkov JR et al (2017) The Rosetta all-atom energy function for macromolecular modeling and design. J Chem Theory Comput 13:3031–3048 20. Simonson T (2015) The physical basis of ligand binding. In: Cavasotto CN (ed) In silico drug discovery and design: theory, methods, challenges, and applications. CRC Press 21. Roux B, Simonson T (1999) Implicit solvent models. Biophys Chem 78:1–20 22. Villa F, Panel N, Chen X et al (2018) Adaptive landscape flattening in amino acid sequence space for the computational design of protein: peptide binding. J Chem Phys 149:072302–072302 23. Panel N, Sun YJ, Fuentes EJ et al (2017) A simple PB/LIE free energy function accurately predicts the peptide binding specificity of the Tiam1 PDZ domain. Front Mol Biosci 4:65–65 24. Saven JG (2010) Computational protein design: advances in the design and redesign of biomolecular nanostructures. Curr Opin Colloid Interface Sci 15:13–17

254

Nicolas Panel et al.

25. Pantazes RJ, Grisewood MJ, Maranas CD (2011) Recent advances in computational protein design. Curr Opin Struct Biol 21:467–472 26. Der BS, Kuhlman B (2013) Strategies to control the binding mode of de novo designed protein interactions. Curr Opin Struct Biol 23:639–646 27. Moal IH, Moretti R, Baker D et al (2013) Scoring functions for protein-protein interactions. Curr Opin Struct Biol 23:862–867 28. Zanghellini A (2014) De novo computational enzyme design. Curr Opin Biotechnol 29:132–138 29. Khoury GA, Smadbeck J, Kieslich CA et al (2014) Protein folding and de novo protein design for biotechnological applications. Trends Biotechnol 32:99–109 30. Simonson T, Ye-Lehmann S, Palmai Z et al (2016) Redesigning the stereospecificity of tyrosyl-tRNA synthetase. Proteins 84:240–253 31. Cornell WD, Cieplak P, Bayly CI et al (1996) A second generation force field for the simulation of proteins, nucleic acids, and organic molecules J. Am. Chem. Soc. 1995, 117, 51795197. J Am Chem Soc 118:2309–2309 32. Michael E, Polydorides S, Simonson T et al (2017) Simple models for nonpolar solvation: parameterization and testing. J Comput Chem 38:2509–2519 33. Liu X, Shepherd TR, Murray AM et al (2013) The structure of the Tiam1 PDZ domain/ phospho-syndecan1 complex reveals a ligand conformation that modulates protein dynamics. Structure 21:342–354 34. Tuffery P, Etchebest C, Hazout S et al (1991) A new approach to the rapid determination of protein side chain conformations. J Biomol Struct Dyn 8:1267–1289 35. Gaillard T, Panel N, Simonson T (2016) Protein side chain conformation predictions with an MMGBSA energy function. Proteins 84:803–819 36. Druart K, Bigot J, Audit E et al (2016) A hybrid Monte Carlo scheme for multibackbone protein design. J Chem Theory Comput 12:6035–6048 37. Mignon D, Panel N, Chen X et al (2017) Computational design of the Tiam1 PDZ domain and its ligand binding. J Chem Theory Comput 13:2271–2289 38. Polydorides S, Simonson T (2013) Monte carlo simulations of proteins at constant pH with generalized born solvent, flexible sidechains, and an effective dielectric boundary. J Comput Chem 34:2742–2756

39. Villa F, Mignon D, Polydorides S et al (2017) Comparing pairwise-additive and many-body generalized born models for acid/base calculations and protein design. J Comput Chem 38:2396–2410 40. Frenkel D, Smit B (1996) Understanding molecular simulation, Computational science series, vol 1. Academic Press, New York 41. Mignon D, Simonson T (2016) Comparing three stochastic search algorithms for computational protein design: Monte Carlo, replica exchange Monte Carlo, and a multistart, steepest-descent heuristic. J Comput Chem 37:1781–1793 42. Bhattacherjee A, Wallin S (2013) Exploring protein-peptide binding specificity through computational peptide screening. PLoS Comput Biol 9:e1003277 43. Shepherd TR, Hard RL, Murray AM et al (2011) Distinct ligand specificity of the Tiam1 and Tiam2 PDZ domains. Biochemistry 50:1296–1308 44. Panel N, Villa F, Fuentes EJ et al (2018) Accurate PDZ/peptide binding specificity with additive and polarizable free energy simulations. Biophys J 114:1091–1102 45. Songyang Z, Fanning AS, Fu C et al (1997) Recognition of unique carboxyl-terminal motifs by distinct PDZ domains. Science 275:73–77 46. Shepherd TR, Klaus SM, Liu X et al (2010) The Tiam1 PDZ domain couples to syndecan1 and promotes cell-matrix adhesion. J Mol Biol 398:730–746 47. Liu X, Speckhard DC, Shepherd TR et al (2016) Distinct roles for conformational dynamics in protein-ligand interactions. Structure 24:2053–2066 48. Kuhn B, Kollman PA (2000) A ligand that is predicted to bind better to avidin than biotin: insights from computational fluorine scanning. J Am Chem Soc 122:3909–3916 49. Wang L, Magliery TJ, Liu DR et al (2000) A new functional suppressor tRNA/aminoacyltRNA synthetase pair for the in vivo incorporation of unnatural amino acids into proteins [16]. J Am Chem Soc 122:5010–5011 50. Rastelli G, Del Rio A, Degliesposti G et al (2010) Fast and accurate predictions of binding free energies using MM-PBSA and MM-GBSA. J Comput Chem 31:797–810 51. Hou T, Wang J, Li Y et al (2011) Assessing the performance of the MM/PBSA and MM/GBSA methods. 1. The accuracy of binding free energy calculations based on molecular dynamics simulations. J Chem Inf Model 51:69–82

Computational Design of Binding 52. Ben-Shalom IY, Pfeiffer-Marek S, Baringhaus KH et al (2017) Efficient approximation of ligand rotational and translational entropy changes upon binding for use in MM-PBSA calculations. J Chem Inf Model 57:170–189 53. Hou T, Guo S, Xu X (2002) Predictions of binding of a diverse set of ligands to gelatinase-A by a combination of molecular dynamics and continuum solvent models. J Phys Chem B 106:5527–5535 54. Gouda H, Kuntz ID, Case DA et al (2003) Free energy calculations for theophylline binding to an RNA aptamer: comparison of MM-PBSA and thermodynamic integration methods. Biopolymers 68:16–34 55. Hu G, Ma A, Wang J (2017) Ligand selectivity mechanism and conformational changes in guanine Riboswitch by molecular dynamics simulations and free energy calculations. J Chem Inf Model 57:918–928 56. Hou T, Chen K, McLaughlin WA et al (2006) Computational analysis and prediction of the binding motif and protein interacting partners

255

of the Abl SH3 domain. PLoS Comput Biol 2: e1–e1 57. Stoica I, Sadiq SK, Coveney PV (2008) Rapid and accurate prediction of binding free energies for saquinavir-bound HIV-1 proteases. J Am Chem Soc 130:2639–2648 58. Venken T, Krnavek D, Mu¨nch J et al (2011) An optimized MM/PBSA virtual screening approach applied to an HIV-1 gp41 fusion peptide inhibitor. Proteins 79:3221–3235 59. Spiliotopoulos D, Spitaleri A, Musco G (2012) Exploring PHD fingers and H3K4me0 interactions with molecular dynamics simulations and binding free energy calculations: AIREPHD1, a comparative study. PLoS One 7: e46902 60. Opuu V, Sun YJ, Hou T et al (2020) A physicsbased energy function allows the computational redesign of a PDZ domain. Sci Rep 10:11150 61. Simonson T (2003) Electrostatics and dynamics of proteins. Rep Prog Phys 66:737–737

Chapter 15 Mechanoregulation of PDZ Proteins, An Emerging Function Elsa Bazellie`res and Andre´ Le Bivic Abstract Mechanical forces have emerged as essential regulators of cell organization, proliferation, migration, and polarity to regulate cellular and tissue homeostasis. Changes in forces or loss of the cellular response to them can result in abnormal embryonic development and diseases. Over the past two decades, many efforts have been put in deciphering the molecular mechanisms that convert forces into biochemical signals, allowing for the identification of many mechanotransducer proteins. Here we discuss how PDZ proteins are emerging as new mechanotransducer proteins by altering their conformations or localizations upon force loads, leading to the formation of macromolecular modules tethering the cell membrane to the actin cytoskeleton. Key words PDZ proteins, Forces, Tight junctions, Adherens junctions, Actomyosin

1

A Brief Introduction on Mechanotransduction/Mechanoregulation in Biology All the cells and tissues of the body are subject to external and internal forces. These forces can affect the shape and intracellular organization of cells, their proliferation, their migration, and their intercellular interactions. Forces influence the development of embryos [1] as well as cell functions and homeostasis in the adult [2]. Moreover, many disease states are characterized by changes in these forces and/or a loss of the normal cellular response to them [3]. Over the last decade, mechanotransduction has emerged as a key process in development and diseases. Mechanotransduction can be defined as a cellular event that converts a mechanical input such as fluid shear stress (blood vessels), stretch (lung, intestine), osmotic forces (urinary tract), mechanical load (bone, muscle) [4, 5] as well as the impact of the stiffness of the extracellular matrix (ECM) that surrounds most cells leading to a biochemical response [6]. In well-studied examples of mechanotransduction, proteins can undergo force-induced changes into conformations that lead to modification in affinity for binding partners or catalytic activity. The mechanical load triggers biochemical changes that can

Jean-Paul Borg (ed.), PDZ Mediated Interactions: Methods and Protocols, Methods in Molecular Biology, vol. 2256, https://doi.org/10.1007/978-1-0716-1166-1_15, © Springer Science+Business Media, LLC, part of Springer Nature 2021

257

258

Elsa Bazellie`res and Andre´ Le Bivic

propagate via the activation of specific signaling pathways [7]. In a simple context, the mechanical load can accelerate the association or dissociation of protein–ligand bonds [8]. In the case of adhesion proteins, that exhibit catch-bond behavior, mechanical loads can produce changes in the conformation of the proteins that lead to high affinity for a binding partner [9–11]. Deciphering how fundamental physical principles are controlled by protein interactions is central to understand how forces are converted into biochemical signals. Elucidation of the specificity, selectivity, and regulatory mechanisms involved in protein–protein interactions can therefore provide important insights into many biological processes such as cell proliferation, cell migration, and cell polarity. Structural studies have revealed that mechanosensitive proteins with multiple domains and flexible interdomain interfaces can pass through multiple conformations [12, 13]. For example, a bent conformation can open up to a straighter arrangement of domains along the direction of the applied force. These changes in conformation result in modification of the affinity for the binding partners, exposure or hiding of different catalytic domains that will trigger differential intracellular signaling responses or even redistribution of the protein within the cell [14–16]. Among the proteins that mediate protein–protein interactions, there is the large PDZ (postsynaptic density 95/Disc large/Zonula occludens) family. Most of the PDZ proteins are multimodular scaffold proteins and often contain multiple PDZ domains which can interact with various binding partners and thereby assemble supramolecular signaling complexes [17]. As PDZ domains interact with motifs present in many proteins, understanding the regulatory mechanism of PDZ mediated interactions is important to gain insight into biological processes. So far, posttranslational modifications, autoinhibition, and allosteric interactions have been proposed to regulate PDZ-mediated interactions and thus intracellular signaling [18], but little is known about the impact of mechanical inputs on PDZ proteins. In this review, we will focus on the current knowledge on the mechanoregulation of PDZ proteins, focusing on few recent findings.

2

PDZ Proteins as Mechanotransducers Biological mechanotransducers can be defined as a single protein or a protein complex that produce or enable a chemical signal in response to mechanical stimuli. These mechanotransducers can participate in mechanoreception and mechanotransmission as direct mechanosensitive structure, which respond by altering their conformation upon force loads, or as second line mechanotransducers. Among the mechanotransducer proteins studied so far, the PDZ protein family is emerging as a new pool of protein sensitive to

Mechanoregulation of PDZ Proteins

259

forces, but only few examples are known so far. In epithelial cells, the adhesion machineries at the cell–cell or cell–substrate interfaces are known to play an important role in mechanotransduction [10, 19]. These machineries are composed of several proteins, and many of them are part of the PDZ family [20]. Being integrated in a complex modular array tethering the cell membrane to the cell cytoskeleton, the adhesion scaffolding PDZ proteins might constitute an important link to transduce the mechanical signals into an appropriate cell response to maintain cellular homeostasis. We will thus focus on PDZ proteins from the adhesion machinery that have been recently shown to be regulated by forces in a direct or indirect manner (Fig. 1).

3 3.1

Force Regulated PDZ Proteins ZO1

ZO1 is a tight junction scaffolding protein that belongs to the MAGUK (membrane associated guanylated kinase homolog) family. It is characterized by the occurrence of three PDZ, one SH3 and a GUK (guanylate kinase homologous domain) domains [21]. ZO1 is a cytoplasmic protein, that anchors actin filaments through its actin binding domain at its C terminal region [22, 23]. At the N-terminus, the PDZ1 and PDZ3 domains bind to the tight junction membrane associated proteins, Claudins and JAM-A, and to TAZ, a member of Hippo pathway, while the PDZ2 domain promotes heterodimerization between ZO1 and either ZO2 or ZO3 [22, 24–26]. A larger region encompassing the PDZ3, SH3, U5, and GUK domains (ZPSG-1) interacts with and recruits to junctions the transmembrane TJ protein occludin, but also DbpA/ZONAB (DNA binding protein A/ZO1 associated Nucleic Acid-binding protein) [27, 28]. The sequestration of DbpA by ZO1 and ZO2 at the junctions of confluent monolayers inhibits its nuclear activity that regulates gene expression and cell proliferation [27–30]. Recently, Spadaro et al. demonstrated that ZO1 stretches upon mechanical forces [31]. In 2011, Then et al. have already proposed an impact of membrane tension on the localization of ZO1 [32]. Under hyperosmotic stress that generates an increase in membrane tension, the actin cytoskeleton is reorganized with the appearance of a dense F-actin cortical ring. In this condition, ZO1 expression is increased, and it colocalizes with the newly formed actin ring [32]. As ZO1 is anchored to the actin filaments through its C-terminal region, this change in actin organization exerts direct pulling forces on the ZO1 protein. When present as a heterodimer with ZO2 within cells, ZO1 is in a stretched configuration that allows binding to DpbA. However, the depletion of ZO2 together with an inhibition of myosin contractility promotes a folding of the N-terminal and C-terminal end of ZO-1, and releases the

260

Elsa Bazellie`res and Andre´ Le Bivic

Fig. 1 Hypothetical role of the PDZ protein in the strengthening and stabilization of intercellular adhesions. (a) At the level of the tight junctions (TJs), under low tension the mechanotransducer ZO1 is in an inactive

Mechanoregulation of PDZ Proteins

261

interaction with DpbA and Occludin. By applying linearly increasing forces to purified full-length single ZO1 molecules with magnetic tweezers, Spadaro et al. could demonstrate that both the C-terminal region and the ZPSG-1 module of ZO1 that comprises the PDZ3, SH3, U5, and GUK domains unfold at forces ranging from 5 to 20 pN, releasing the autoinhibition interaction between ZPSG and C terminal domain of ZO1 [31]. ZPSG-1 domain is not only important for the junctional localization of ZO1 but also for its interaction with Occludin, DbpA, and thus for barrier formation and epithelial polarization [33–38]. Forces act as an allosteric effector by stretching ZO1 protein to promote the interaction of ZPSG1 with its ligands, occludin and DbpA. While the stretched conformation of ZO1 is the active conformation, the folded conformation is the inactive form in which the ZPSG domain is autoinhibited. In its inactive form, junctional ZO1 remains anchored to the membrane through binding of its N-terminal domain with interactors such as TAZ or Claudins, while the C-terminal half is intramolecularly autoinhibited. The release of ZPSG-1 domains may also regulate the interaction of other proteins such as α-Catenin, Afadin, JAM-A, Vinculin and Shroom2 [39]. 3.2

MUPP1

MUPP1 belongs to the family of multi-PDZ proteins and contains a L27 domain at its N-terminal region followed by 13 PDZ domains [40]. MUPP1 is a structural paralog of PATJ, and both share several binding partners such as PALS1, PAR-6, AMOT, Jeap, ZO3, Claudins, or Nectins [41]. Lanaspa et al. by the use of osmolarity changes demonstrated that both acute and chronic hyperosmolarity in inner medullary collecting duct 3 cells induce an increase in the expression of MUPP1, ZO1, and Afadin [42]. As MUPP1 expression increases, it localizes to the apical side of the membrane at the level of the Tight Junctions (TJs). To survive hyperosmotic stresses and to maintain the integrity of the cell sheet with efficient barrier functions, cells have to adapt through

ä Fig. 1 (continued) conformation, and the second line mechanotransducers Afadin and Mupp1 are in the cytoplasm. The actin filaments are exerting low mechanical loads. Under high tension, generated by actomyosin bundles, ZO1 unfolds and unmasks several domains that leads to the binding with Occludin and JAM together with the recruitment of DbpA and Afadin. Afadin is then able to recruit Rap2c generating a positive feedback loop on RHO allowing for actin contractions and thus force generation. This increase in forces results in the recruitment of Mupp1 at the TJs, that will bind to ZO1, ZO2, and JAM but also potentially to the CRB3A/Pals1 polarity complex. The formation of this macromolecular complex that tethers transmembrane proteins (Claudins, Occludins, JAM) to the actin cytoskeleton is key for the strengthening and the stabilization of the TJs. (b) In the lateral domain of the cells, at the level of the lateral Adherens Junctions (LAJs), the actin filaments that binds to the E-cadherin–catenin complex, exert low tension and SCRIBBLE is in the cytoplasm. Above the lateral junctions, the E-cadherin–catenin complex is link to vinculin and binds to actomyosin bundles that are under high mechanical loads. SCRIBBLE is also enriched at the AJ level, and plays a key role in the stabilization of the AJs

262

Elsa Bazellie`res and Andre´ Le Bivic

reorganization of the actin cytoskeleton as mentioned previously but also through differential expression of proteins, that move to the junctions to counterbalance the changes in membrane tension. Depletion of MUPP1, results in a disruption of the TJ barrier, with a drop of transepithelial resistance of about 25% [42]. Molecularly, depletion of MUPP1, in hyperosmotic conditions, triggers a decreased expression and loss of membrane localization of Claudin4, a TJ protein [42]. These data clearly show that MUPP1 is important in the maintenance of the epithelial homeostasis and confirm previous study that pointed out a predominant role of MUPP1 in the disruption of both TJs barrier and apicobasal polarity [43–45]. A more recent study shows that MUPP1 stability and degradation in endothelial cells depends on the regulation of PDZRN3, an E3 ubiquitin ligase PDZ domain containing ring finger 3 protein [46]. The interactions between PDZRN3, MUPP1, PAR3, and aPKC regulate the stabilization of TJs in endothelial cells. Perturbation of PDZRN3 expression induces degradation of MUPP1 that correlates with destabilization of the actin cytoskeleton and disruption of the TJs. All these studies point toward a role of MUPP1 in the stabilization and maintenance of actin cytoskeleton and TJs in cells. 20 years ago, it was suggested that MUPP1 could change its conformation through the interaction of its PDZ10 domain with the 5-hydroxytryptamine type 2C receptors [47]. This interaction induces a conformational change in MUPP1 and triggers a clustering of the 5-hydroxytryptamine type 2C receptors at the cell membrane, triggering downstream signal transduction pathways. It is of interest that different PDZ domains of MUPP1 can bind to CADM1, a transmembrane cell adhesion protein, with different affinities. The PDZ2 of MUPP1, for example, presents the highest affinity for CADM1, but when this PDZ is absent, CADM1 can still interact with other PDZ domains of MUPP1. These data could point to a change in conformation of MUPP1 that would unmask the PDZ2 domain leading to a strong binding to CADM1 [48, 49]. In the case of CADM1 and maybe of other interactors of MUPP1, the interactions with multiple domains can be an alternative when some domains are either occupied by other ligands or hidden by an autoinhibitory conformation of MUPP1. These multi–low-affinity interactions might then serve as transient interaction contacts before MUPP1 reaches its stretched configuration. However, the unfolding/folding regulation of MUPP1 is still under debate as only indirect evidence has been obtained to date. 3.3

PAR3

PAR3 is part of the Partitioning defective proteins and presents a CR1 oligomerization domain, 3 PDZ domains and aPKC binding domain. A recent study pointed out that cortical forces are responsible for the clustering of PAR3 on the cell cortex. In C. elegans, inhibition of actin contraction with blebbistatin or actin filament

Mechanoregulation of PDZ Proteins

263

polymerization causes a reduction of the PAR3 clusters on the cortex, demonstrating that higher cortical tension can drive PAR3 cluster formation at the cortex [50]. The actomyosin membrane associated network generates flows [51], which promote symmetry breaking along the anterior posterior axis through the advection of polarity components [52–54]. The polarized actomyosin contractions pull actin networks along the membrane toward the anterior region while triggering local disassembly and turnover via increased local tension, resulting in flow of material [55]. Based on the advective flow model proposed by Goerhing et al. [53] that predicts that diffusivity and turnover of cortical polarity proteins should be slow and stable enough to be passively transport by the advective cortical flow, Wang et al. [50] postulate a “clustering and stabilization” hypothesis for PAR3. By using fluorescence recovery after photobleaching, they demonstrate that the half-life of PAR3 and PKC are shorter before symmetry breaking. During the early to middle establishment phase, when the clusters are formed, the halflife increases. Molecularly, during this phase the activity of Cdc42 is reduced allowing PKC3 and PAR6 to associate with PAR3 clusters at the cortex and thus facilitates effective transport by advective cortical flows. During the maintenance phase, the half-lives of PAR3 and PKC are shorter, when the clusters disassemble, concomitant with an increased activity of Cdc42 that prevents PKC and PAR6 association with PAR3. As cortical tensions trigger a dynamic equilibrium between an unclustered and clustered form of PAR3 correlating with its association with PAR6 and PKC, it is tempting to speculate that PAR3 may undergo structural changes by mechanical stretching, which could relieve the amino-terminal CR1 domain from intramolecular inhibitions, thus promoting PAR3 oligomerization [56]. These mechanically induced conformational changes may directly activate the ability of the CR1 domain to facilitate its oligomerization, and proper localization in cells [56]. Furthermore, similar cortical flows are observed during cell division or cell migration, and then could, if proven, also explain the interaction between the different members of the PAR complex.

4

PDZ Proteins “Second Line” Mechanotransducers In this part, we will describe how some PDZ proteins interact with proteins that are activated upon forces. These interactions may trigger a change in the conformation or in the localization of the “second line” mechanotransducers, thereby regulating different signal pathways.

264

Elsa Bazellie`res and Andre´ Le Bivic

4.1

PAR6

PAR6 is part of the Partitioning defective proteins in C. elegans, and encodes a protein with PKC binding domain (PB1), a semiCdc42/Rac1 interactive binding (CRIB) and a PDZ domain [57]. Recently a PDZ binding motif (PBM) was identified at its C-terminal region [58]. PAR6 in many species is involved in cell division, cell polarization, and cell migration, processes that are characterized by changes in cellular forces [59–61]. PAR6 function is able to segregate to the apical or leading edge of epithelial cells upon activation by Cdc42-GTP [57, 62]. Accumulating evidence indicates that Cdc42 responds to and is activated upon different mechanical loads such as hyperosmolarity, shear stress, and intercellular increased tension [63]. Upon exposure to mechanical loads, Cdc42 translocates from cytosol to the membrane and is concomitantly activated [64]. Activation of Cdc42 by exchange of GDP for GTP triggers actin polymerization and generation of tension [65, 66]. Binding of activated Cdc42-GTP to the CRIB domain of PAR6 favors the interaction with CRUMBS over the interaction with PALS1 through a change in conformation of the CRIB-PDZ domains of PAR6 [67]. Upon Cdc42-GTP binding, a portion of the flexible CRIB motif folds into a stable conformation with the PDZ domain [68], triggering an increase affinity for the interaction with CRUMBS, and releasing the interaction with PALS1. The modular nature of PAR6 may allow the mutually exclusive interaction of CRUMBS and PALS1 with its PDZ domains, regulating the assembly and localization of different polarity complexes depending on the cellular context. The localization of PAR6 to the plasma membrane depends on both the PDZ and the PBM domains, as both deletions caused a strong mislocalization of PAR6 to the cytosol. The PBM domain binds with different affinities to the PDZ1 and PDZ3 of PAR3, whose clustering is mediated by an increase in cortical tension as previously mentioned [50]. The weak but multivalent interaction of one PAR3 molecule with two PAR6 molecules might allow the assembly a large cluster of PAR complexes at the cell membrane. The formation of the large-scale cluster of PAR complexes can facilitate the actomyosin dependent advective transport of several PAR6 proteins during the establishment of polarity but also the formation of large clusters made with several proteins such as aPKC, PALS1, or CRUMBS that can serve different functions.

4.2

DLG

DLG is a MAGUK protein that presents an L27 domain, 3 PDZ domains, an SH3 domain, a HOOK domain and a Guanylate like domain. DLG is a member of the basolateral polarity complex SCRIBBLE, and is involved in processes such as cell division, cell migration and cell polarity [69] The N-terminal L27 domain interacts with the two L27 domains of calcium/calmodulin-dependent serine protein kinase (CASK) [70]. CASK is a membrane-associated guanylate kinase and a scaffolding protein, that has been

Mechanoregulation of PDZ Proteins

265

demonstrated to recruit or organize other proteins at the plasma membrane to coordinate signal transduction pathways within the cytoplasm and nucleus [71]. A study done in 2016, using a cell mechanical stretch device shows that CASK expression and localization to the basal membrane is needed for the inhibition of proliferation of cells under cyclic stretch [72]. The authors show that CASK interacts with ß1-integrin, however, the tension-driven interaction between these two proteins still has to be formally proven. Depletion of CASK results in aberrant proliferation of cells under mechanical stress demonstrating that CASK localization under tension is important for the mechanoregulation of the cell division [72]. In 2019, Porter et al. demonstrated that the direct interaction of CASK with DLG is required for normal cortical recruitment of NUMA, a key component of spindle orientation machinery [73]. Disruption of this interaction affects the integrity of epithelial architecture and results in misoriented cell division that give rise to multilumen cyst in 3D. Since the formation of 3D cysts generates an increase in tension at the cell–cell interface [74], it is tempting to speculate that CASK is recruited at the cell–cell interface in a force-dependent manner and then will recruit DLG allowing the formation of well polarized 3D cysts. DLG interacts with the C terminal PDZ binding domain of CD97 and together these proteins are part of the adherens junctional signaling complex composed of E-cadherin and catenins [75]. This macromolecular complex is linked to the F actin cortex and is thus submitted to cellular forces. The localization of CD97 at the plasma membrane is actin dependent as blocking actin polymerization and elongation prevents its membrane localization. When present at the cell membrane CD97 strengthens the adherens junction (AJs) [76] while deletion of its PDZ binding domain results in the loss of cell–cell contacts [77] upon mechanical shear stress. Taken together these data show a role of CD97 in the mechanoregulation of cell–cell contacts. In a recent study, mechanical stimulation of epithelial cells applied by using shear stress or wound assay results in a rapid phosphorylation of Ser740 of CD97. This phosphorylation disrupts the binding of DLG1 to the PDZ binding domain of CD97, and correlates with a disorganization of the actin cytoskeleton. PKC contributes to the mechanical force induced cellular responses [78] and is a potential candidate to phosphorylate CD97 at S740. PKCα interacts via its PBM with PDZ3 domain of DLG1 [79]. Upon mechanical stress, PKCα might be recruited to the cell membrane via DLG1 and will then trigger phosphorylation of CD97 causing F-actin depolymerization, loss of cell–cell contact, and DLG1 detachment. This signaling pathway when activated induces depolymerization of actin and loss of cell–cell contacts, that will result in a relaxation of the tension at this particular cellular junction avoiding tissue breaking.

266

4.3

Elsa Bazellie`res and Andre´ Le Bivic

Afadin

Afadin is a filamentous actin binding protein with two Ras domains, a forkhead association domain, a dilute domain, a PDZ domain, and three proline-rich domains. Afadin is implicated in many cellular processes from cell survival, cell proliferation to cell migration and formation of the apical junctions in epithelial cells [80–83]. Afadin can interact both directly or indirectly with the actin cytoskeleton through several partners such as JAM-A, ZO1/ZO2, vinculin, and α-actinin [84–88]. Afadin is thus a strong candidate to be involved in mechanoregulated processes as many of its partners have been already linked to mechanotransduction. Interestingly, depletion of Afadin, JAM-A or double depletion of ZO1/ZO2 results in similar phenotypes with increased contractility triggered by activation of RhoA and phosphorylation of myosin-light chain [89–91]. In response to ZO1/ZO2 double depletion, Afadin is recruited to the cellular cell contacts, and inhibiting contractility perturbs the homogeneous localization of Afadin. Thus, myosin contractility is essential for maintaining Afadin uniform distribution along the zonula adherens (ZA). Removing Afadin in ZO1/ZO2 depleted cells specifically altered actomyosin architecture at the ZA of tricellular junctions where actin cables are anchored. This perturbation is accompanied by discontinuities in the E-cadherin, claudin, and occludin stainings [92]. Taken together, this data shows a synergy between ZO and Afadin depletions in disrupting tissue integrity under tension. Afadin may therefore act as a robust protein scaffold that maintains ZA architecture at tricellular junctions. JAM-A, a component of the TJs, via its PDZ binding motif, can associate with signaling molecules such as scaffold PDZ proteins, ZO1/2, and Afadin, as well as with the guanine exchange factor PDZ-GEF2 [83, 93]. Monteiro et al. demonstrate that JAM-A interacts with Afadin, ZO2, and PDZ-GEF1 to activate the small GTPase Rap2c [89]. The activation of Rap2c controls the contraction of the apical cytoskeleton regulating the epithelial permeability to prevent cell injury. In this study, the authors also mention that Afadin is able to immunoprecipitate a doublet of JAM-A that might represent phosphorylated forms of JAM-A, although they did not investigate this phosphorylation but another team did. Scott at al show that mechanical stimulation can trigger phosphorylation of JAM-A [94]. Forces were applied by using fluid shear stress or by using magnetic tweezer. After applying forces, several biochemical analyses were performed that demonstrate that tension induces rapid phosphorylation of JAM-A S284. This phosphorylation of JAM-A induces activation of RhoA by PKCζ triggering cell stiffening by modifying actin cytoskeleton. These results clearly demonstrate that JAM-A is a direct transducer of mechanical forces. Interestingly, when JAM-A is localized at the TJs which are under high levels of tension [95] it is phosphorylated [96]. It is thus tempting to speculate that mechanical tension, above a certain

Mechanoregulation of PDZ Proteins

267

threshold might induce the phosphorylation of JAM-A. This phosphorylation will control the affinity for Afadin thereby controlling the contraction of apical cytoskeleton that allows for the maintenance of ZA and TJs. 4.4

SCRIBBLE

SCRIBBLE is a classical multimodular scaffold protein that contains 16 leucine-rich repeats (LRR) and four PDZ domains. N-terminal LRR domain can associate with LGL and is required for the association with the lateral cortex and the establishment of apicobasal polarity [97, 98]. The C-terminal PDZ domains of SCRIBBLE can interact with a diverse range of proteins such as β-Pix [99], Vangl2 [100], and LGN [101], allowing for cell polarization, cell migration [102], establishment of planar cell polarity, and cell division, respectively. In epithelial cells, SCRIBBLE is required for normal intercellular adhesion by stabilizing E-cadherin–p120 catenin at the plasma membrane [103]. 10 years ago, E-cadherin complex was tagged as a mechanosensor complex. Since then many studies using different approaches such as FRET sensor, magnetic tweezer, and cell stretching have confirmed that E-cadherin can sense changes in mechanical forces within an epithelial monolayer [104–106]. The role of SCRIBBLE in mechanotransduction has, however, not been investigated so far. Some hints based on the function of SCRIBBLE in the stabilization of E-cadherin point to a role in the mechanoregulation of AJs. E-cadherin receptors form adhesive clusters that are coupled to the contractile actomyosin cortex [107, 108]. At the cellular level, E-cadherin adhesion not only binds cells together but also mechanically couples the contractile modules of neighboring cells together to generate junctional tension [109]. This junctional tension has been mainly treated as homogeneous along the cell–cell interface. However, E-cadherin forms clusters of different sizes along the junctions between cells [108–110]. Interestingly, despite overall basolateral localization, SCRIBBLE is enriched at the ZAs, where it stabilizes E-cadherin–p120 catenin adhesions [111, 112]. At the level of the ZAs, clusters of stabilized E-cadherin are linked to large actomyosin bundles [113–115]. Below the AJs, at the lateral adherens junctions (LAJs), E-cadherin is coupled to less aligned actomyosin network. Contractile tension is thereby greater at the ZA than at the LAJ. SCRIBBLE could be enriched at the ZA in a tension-dependent manner, and thereby by stabilizing E-cadherin could be implicated in a feedback positive loop in the generation of forces at the ZAs. This tension driven localization of SCRIBBLE to E-cadherin can also take place during the orientation of the mitotic spindle to align epithelial cell division. In 2017, Hart et al., by applying uniaxial stretch on epithelial cells, showed that cell division aligned with the stretch axis [116]. The orientation of this cell division axis requires trans-engagement of E-cadherin adhesions and involves tension-dependent recruitment of LGN to

268

Elsa Bazellie`res and Andre´ Le Bivic

E-cadherin. LGN can directly interact with the cytosolic tail of E-cadherin that localizes LGN at the cell–cell contacts [117]. Increase in tension does not trigger a polarized distribution of E-cadherin, while LGN and myosin IIA are polarized suggesting that an additional intermediate, LGN interacting protein might be involved in its polarized localization. This protein could be SCRIBBLE as other ones (DLG, Afadin, ERM proteins) were ruled out in this study. SCRIBBLE is indeed able to interact with LGN, and the interaction between LGN and E-cadherin is reduced in cells depleted for SCRIBBLE. The formation of this ternary complex is also important for proper cell division [101]. Taken together, all these data point to SCRIBBLE as a likely “second line” mechanotransducer to trigger signaling pathways downstream of the mechanosensitive E-cadherin units.

5

Concluding Remarks In the last two decades, mechanical stimuli have emerged as essential regulators of several biological processes. Mechanotransduction has revealed a new layer of control of the interaction between proteins, and will potentially lead to global guiding principles for the organization of complex living systems. The convergent interests from physicists and biologists to understand the complexity of integrated systems have led to the development of new biophysical tools. These new technological advances such as stretching devices, optical/magnetic tweezers, or atomic force microscopy have enabled to measure and apply controlled forces on cells. Applying different forces on cells trigger several cellular responses and activate different signaling pathways that have led to the identification of an emerging mechanotransducing function for PDZ proteins. These mechanotransducers can participate in mechanoreception and mechanotransmission as direct mechanosensitive components or as second line mechanotransducers. Pulling forces generated by the actomyosin network can stretch molecules exposing their cryptic binding sites or cryptic phosphorylation sites, triggering specific signaling pathways [5, 118]. In the case of ZO1, physical forces are responsible for the stretching of the molecule that is needed for its correct localization while allowing for the interaction with its different partners. This activated unfolded ZO1 protein is important for the maintenance of the TJs. The actomyosin network can thus directly by its contractions pull on proteins but also on the actin cytoskeleton, resulting in flow of material that leads to the formation of polarized clusters of proteins at the cell cortex that allow for adaptation of cell cytoskeleton [51, 119]. Under higher cortical tension PAR3 clusters and MUPP1 is recruited to the cell cortex. The tension-dependent clustering of PAR3 could promote its oligomerization at the cell

Mechanoregulation of PDZ Proteins

269

cortex where it could recruit PAR6 and aPKC, leading to the establishment of anteroposterior or apicobasal polarity. The recruitment of the PDZ protein MUPP1 to the TJs is key to maintaining the integrity of the TJ barrier during hyperosmotic stress. Finally, loading forces can regulate cell adhesion machineries [105, 106, 120]. The adhesive contacts can adapt to external forces, by modifying their binding to specific proteins, their structural conformations, their stability but also their links to the actin cytoskeleton. Some PDZ proteins such as SCRIBBLE, Afadin, or ZO1 can, upon loading forces, strengthen the cell adhesion machinery at the ZA and TJs localization respectively to preserve epithelial homeostasis by allowing for the proper polarization of the tissue and the tethering of the adhesions to the actin cytoskeleton. In these cases, the connections between adhesions are strengthened by assembly of new components, however in other cases the connections can be severed [121]. DLG is located at cell–cell contacts and upon forces generation will recruit aPKC at the ZA. The recruitment of aPKC by DLG at the ZA can then regulate the load of tension at the cell– cell contact by controlling the depolymerization of actin and disruption of cell–cell contact leading to a relaxation of the tissue when the forces are too elevated. The active nature of cell adhesion machineries implies that the time from the mechanosensing to the mechanoresponse is slow and occurs within seconds to minutes allowing the occurrence of many different interactions between proteins [122–124]. In the case of PDZ scaffolds proteins, this could trigger specific interactions allowing for the cells to have a multimodal adaptive “repertoire” to properly adapt their responses to the change in forces.

Acknowledgments Elsa Bazellie`res and Andre´ Le Bivic are supported by CNRS. This project was developed in the context of the LabEx INFORM (ANR-11-LABX-0054) and of the A*MIDEX project (ANR-11IDEX-0001-02) funded by the “Investissements d’Avenir” French Government program. References 1. Petridou NI, Heisenberg C (2019) Tissue rheology in embryonic organization. EMBO J 38:e102497 2. Salvi AM, DeMali KA (2018) Mechanisms linking mechanotransduction and cell metabolism. Curr Opin Cell Biol 54:114–120 3. Jaalouk DE, Lammerding J (2009) Mechanotransduction gone awry. Nat Rev Mol Cell Biol 10:63–73

4. Sidhwani P, Yelon D (2019) Fluid forces shape the embryonic heart: insights from zebrafish. Curr Top Dev Biol 132:395–416 5. Pruitt BL, Dunn AR, Weis WI, Nelson WJ (2014) Mechano-transduction: from molecules to tissues. PLoS Biol 12:e1001996 6. Dzamba BJ, DeSimone DW (2018) Extracellular matrix (ECM) and the sculpting of

270

Elsa Bazellie`res and Andre´ Le Bivic

embryonic tissues. Curr Top Dev Biol 130:245–274 7. Hoffman BD, Grashoff C, Schwartz MA (2011) Dynamic molecular processes mediate cellular mechanotransduction. Nature 475:316–323 8. Depoil D, Dustin ML (2014) Force and affinity in ligand discrimination by the TCR. Trends Immunol 35:597–603 9. Chen W, Lou J, Evans EA, Zhu C (2012) Observing force-regulated conformational changes and ligand dissociation from a single integrin on cells. J Cell Biol 199:497–512 10. Hu X, Margadant FM, Yao M, Sheetz MP (2017) Molecular stretching modulates mechanosensing pathways: molecular stretching modulates mechanosensing pathways. Protein Sci 26:1337–1351 11. Pannekoek W-J, de Rooij J, Gloerich M (2019) Force transduction by cadherin adhesions in morphogenesis. F1000Res 8:1044 12. Cailliez F, Lavery R (2005) Cadherin mechanics and complexation: the importance of calcium binding. Biophys J 89:3895–3903 13. Springer TA, Dustin ML (2012) Integrin inside-out signaling and the immunological synapse. Curr Opin Cell Biol 24:107–115 14. Rutherford SL, Zuker CS (1994) Protein folding and the regulation of signaling pathways. Cell 79:1129–1132 15. Papagrigoriou E, Gingras AR, Barsukov IL, Bate N, Fillingham IJ, Patel B, Frank R, Ziegler WH, Roberts GC, Critchley DR, Emsley J (2004) Activation of a vinculin-binding site in the talin rod involves rearrangement of a fivehelix bundle. EMBO J 23:2942–2951 16. Smith ML, Gourdon D, Little WC, Kubow KE, Eguiluz RA, Luna-Morris S, Vogel V (2007) Force-induced unfolding of fibronectin in the extracellular matrix of living cells. PLoS Biol 5:e268 17. Manjunath GP, Ramanujam PL, Galande S (2018) Structure function relations in PDZdomain-containing proteins: implications for protein networks in cellular signalling. J Biosci 43:155–171 18. Lee H-J, Zheng JJ (2010) PDZ domains and their binding partners: structure, specificity, and modification. Cell Commun Signal 8:8 19. Charras G, Yap AS (2018) Tensile forces and mechanotransduction at cell–cell junctions. Curr Biol 28:R445–R457 20. Heinemann U, Schuetz A (2019) Structural features of tight-junction proteins. Int J Mol Sci 20:6020

21. Stevenson BR, Siliciano JD, Mooseker MS, Goodenough DA (1986) Identification of ZO-1: a high molecular weight polypeptide associated with the tight junction (zonula occludens) in a variety of epithelia. J Cell Biol 103:755–766 22. Fanning AS, Jameson BJ, Jesaitis LA, Anderson JM (1998) The tight junction protein ZO-1 establishes a link between the transmembrane protein occludin and the actin cytoskeleton. J Biol Chem 273:29745–29753 23. Itoh M, Nagafuchi A, Moroi S, Tsukita S (1997) Involvement of ZO-1 in cadherinbased cell adhesion through its direct binding to α catenin and actin filaments. J Cell Biol 138:181–192 24. Itoh M, Furuse M, Morita K, Kubota K, Saitou M, Tsukita S (1999) Direct binding of three tight junction-associated Maguks, Zo-1, Zo-2, and Zo-3, with the Cooh termini of claudins. J Cell Biol 147:1351–1363 25. Wittchen ES, Haskins J, Stevenson BR (1999) Protein interactions at the tight junction. Actin has multiple binding partners, and ZO-1 forms independent complexes with ZO-2 and ZO-3. J Biol Chem 274:35179–35185 26. Remue E, Meerschaert K, Oka T, Boucherie C, Vandekerckhove J, Sudol M, Gettemans J (2010) TAZ interacts with zonula occludens-1 and -2 proteins in a PDZ-1 dependent manner. FEBS Lett 584:4175–4180 27. Balda MS, Matter K (2000) The tight junction protein ZO-1 and an interacting transcription factor regulate ErbB-2 expression. EMBO J 19:2024–2033 28. Spadaro D, Tapia R, Jond L, Sudol M, Fanning AS, Citi S (2014) ZO proteins redundantly regulate the transcription factor DbpA/ZONAB. J Biol Chem 289:22500–22511 29. Balda MS, Garrett MD, Matter K (2003) The ZO-1-associated Y-box factor ZONAB regulates epithelial cell proliferation and cell density. J Cell Biol 160:423–432 30. Sourisseau T, Georgiadis A, Tsapara A, Ali RR, Pestell R, Matter K, Balda MS (2006) Regulation of PCNA and cyclin D1 expression and epithelial morphogenesis by the ZO-1-regulated transcription factor ZONAB/DbpA. Mol Cell Biol 26:2387–2398 31. Spadaro D, Le S, Laroche T, Mean I, Jond L, Yan J, Citi S (2017) Tension-dependent stretching activates ZO-1 to control the

Mechanoregulation of PDZ Proteins junctional localization of its interactors. Curr Biol 27:3783–3795.e8 32. Then C, Bergler T, Jeblick R, Jung B, Banas B, Kr€amer BK (2011) Hypertonic stress promotes the upregulation and phosphorylation of zonula occludens 1. Nephron Physiol 119:p11–p21 33. Fanning AS, Ma TY, Anderson JM (2002) Isolation and functional characterization of the actin binding region in the tight junction protein ZO-1. FASEB J 16:1835–1837 34. Fanning AS, Little BP, Rahner C, Utepbergenov D, Walther Z, Anderson JM (2007) The unique-5 and -6 motifs of ZO-1 regulate tight junction strand localization and scaffolding properties. Mol Biol Cell 18:721–731 35. Umeda K, Ikenouchi J, Katahira-Tayama S, Furuse K, Sasaki H, Nakayama M, Matsui T, Tsukita S, Furuse M, Tsukita S (2006) ZO-1 and ZO-2 independently determine where claudins are polymerized in tight-junction strand formation. Cell 126:741–754 36. Ikenouchi J, Umeda K, Tsukita S, Furuse M, Tsukita S (2007) Requirement of ZO-1 for the formation of belt-like adherens junctions during epithelial cell polarization. J Cell Biol 176:779–786 37. Rodgers LS, Beam MT, Anderson JM, Fanning AS (2013) Epithelial barrier assembly requires coordinated activity of multiple domains of the tight junction protein ZO-1. J Cell Sci 126:1565–1575 38. Odenwald MA, Choi W, Buckley A, Shashikanth N, Joseph NE, Wang Y, Warren MH, Buschmann MM, Pavlyuk R, Hildebrand J, Margolis B, Fanning AS, Turner JR (2017) ZO-1 interactions with F-actin and occludin direct epithelial polarization and single lumen specification in 3D culture. J Cell Sci 130:243–259 39. Fanning AS, Anderson JM (2009) Zonula occludens-1 and -2 are cytosolic scaffolds that regulate the assembly of cellular junctions. Ann N Y Acad Sci 1165:113–120 40. Ullmer C, Schmuck K, Figge A, Lu¨bbert H (1998) Cloning and characterization of MUPP1, a novel PDZ domain protein. FEBS Lett 424:63–68 41. Asse´mat E, Bazellie`res E, Pallesi-Pocachard E, Le Bivic A, Massey-Harroche D (2008) Polarity complex proteins. Biochim Biophys Acta 1778:614–630 42. Lanaspa M, Andres-Hernando A, Rivard C, Dai Y, Berl T (2008) Hypertonic stress increases claudin-4 expression and tight junction integrity in association with MUPP1 in

271

IMCD3 cells. Proc Natl Acad Sci U S A 105:15797–15802 43. Latorre IJ, Roh MH, Frese KK, Weiss RS, Margolis B, Javier RT (2005) Viral oncoprotein-induced mislocalization of select PDZ proteins disrupts tight junctions and causes polarity defects in epithelial cells. J Cell Sci 118:4283–4293 44. Shin K, Straight S, Margolis B (2005) PATJ regulates tight junction formation and polarity in mammalian epithelial cells. J Cell Biol 168:705–711 45. Straight SW, Shin K, Fogg VC, Fan S, Liu C-J, Roh M, Margolis B (2004) Loss of PALS1 expression leads to tight junction and polarity defects. Mol Biol Cell 15:1981–1990 46. Sewduth RN, Kovacic H, Jaspard-Vinassa B, Jecko V, Wavasseur T, Fritsch N, Pernot M, Jeaningros S, Roux E, Dufourcq P, Couffinhal T, Dupla`a C (2017) PDZRN3 destabilizes endothelial cell-cell junctions through a PKCζ-containing polarity complex to increase vascular permeability. Sci Signal 10:464 47. Be´camel C, Figge A, Poliak S, Dumuis A, Peles E, Bockaert J, Lu¨bbert H, Ullmer C (2001) Interaction of serotonin 5-hydroxytryptamine type 2C receptors with PDZ10 of the multi-PDZ domain protein MUPP1. J Biol Chem 276:12974–12982 48. Fujita E, Tanabe Y, Imhof BA, Momoi MY, Momoi T (2012) A complex of synaptic adhesion molecule CADM1, a molecule related to autism spectrum disorder, with MUPP1 in the cerebellum. J Neurochem 123:886–894 49. Baliova M, Juhasova A, Jursky F (2014) Using a collection of MUPP1 domains to investigate the similarities of neurotransmitter transporters C-terminal PDZ motifs. Biochem Biophys Res Commun 454:25–29 50. Wang S-C, Low TYF, Nishimura Y, Gole L, Yu W, Motegi F (2017) Cortical forces and CDC-42 control clustering of PAR proteins for Caenorhabditis elegans embryonic polarization. Nat Cell Biol 19:988–995 51. Munro E, Nance J, Priess JR (2004) Cortical flows powered by asymmetrical contraction transport PAR proteins to establish and maintain anterior-posterior polarity in the early C. elegans embryo. Dev Cell 7:413–424 52. Goehring NW, Hoege C, Grill SW, Hyman AA (2011) PAR proteins diffuse freely across the anterior–posterior boundary in polarized C. elegans embryos. J Cell Biol 193:583–594 53. Goehring NW, Trong PK, Bois JS, Chowdhury D, Nicola EM, Hyman AA, Grill SW (2011) Polarization of PAR proteins

272

Elsa Bazellie`res and Andre´ Le Bivic

by advective triggering of a pattern-forming system. Science 334:1137–1141 54. Mayer M, Depken M, Bois JS, Ju¨licher F, Grill SW (2010) Anisotropies in cortical tension reveal the physical basis of polarizing cortical flows. Nature 467:617–621 55. McFadden WM, McCall PM, Gardel ML, Munro EM (2017) Filament turnover tunes both force generation and dissipation to control long-range flows in a model actomyosin cortex. PLoS Comput Biol 13:e1005811 56. Benton R, Johnston DS (2003) Drosophila PAR-1 and 14-3-3 inhibit Bazooka/PAR-3 to establish complementary cortical domains in polarized cells. Cell 115:691–704 57. Joberty G, Petersen C, Gao L, Macara IG (2000) The cell-polarity protein Par6 links Par3 and atypical protein kinase C to Cdc42. Nat Cell Biol 2:531–539 58. Renschler FA, Bruekner SR, Salomon PL, Mukherjee A, Kullmann L, Schu¨tz-Stoffregen MC, Henzler C, Pawson T, Krahn MP, Wiesner S (2018) Structural basis for the interaction between the cell polarity proteins Par3 and Par6. Sci Signal 11:517 59. Uroz M, Wistorf S, Serra-Picamal X, Conte V, Sales-Pardo M, Roca-Cusachs P, Guimera` R, Trepat X (2018) Regulation of cell cycle progression by cell-cell and cell-matrix forces. Nat Cell Biol 20:646–654 60. Ladoux B, Me`ge R-M, Trepat X (2016) Front-rear polarization by mechanical cues: from single cells to tissues. Trends Cell Biol 26:420–433 61. Trepat X, Wasserman MR, Angelini TE, Millet E, Weitz DA, Butler JP, Fredberg JJ (2009) Physical forces during collective cell migration. Nat Phys 5:426–430 62. Etienne-Manneville S, Hall A (2003) Cell polarity: Par6, aPKC and cytoskeletal crosstalk. Curr Opin Cell Biol 15:67–72 63. Lewis A, Di Ciano C, Rotstein OD, Kapus A (2002) Osmotic stress activates Rac and Cdc42 in neutrophils: role in hypertonicityinduced actin polymerization. Am J Phys Cell Phys 282:C271–C279 64. Li S, Chen BPC, Azuma N, Hu Y-L, Wu SZ, Sumpio BE, Shyy JY-J, Chien S (1999) Distinct roles for the small GTPases Cdc42 and Rho in endothelial responses to shear stress. J Clin Invest 103:1141–1150 65. Tang DD, Gunst SJ (2004) The small GTPase Cdc42 regulates actin polymerization and tension development during contractile stimulation of smooth muscle. J Biol Chem 279:51722–51728

66. Plutoni C, Bazellieres E, Le Borgne-RochetM, Comunale F, Brugues A, Se´veno M, Planchon D, Thuault S, Morin N, Bodin S, Trepat X, Gauthier-Rouvie`re C (2016) P-cadherin promotes collective cell migration via a Cdc42-mediated increase in mechanical forces. J Cell Biol 212:199–217 67. Whitney DS, Peterson FC, Kittell AW, Egner JM, Prehoda KE, Volkman BF (2016) Binding of crumbs to the Par-6 CRIB-PDZ module is regulated by Cdc42. Biochemistry 55:1455–1461 68. Garrard SM, Capaldo CT, Gao L, Rosen MK, Macara IG, Tomchick DR (2003) Structure of Cdc42 in a complex with the GTPasebinding domain of the cell polarity protein, Par6. EMBO J 22:1125–1133 69. Su W-H, Mruk DD, Wong EWP, Lui W-Y, Cheng CY (2012) Polarity protein complex Scribble/Lgl/Dlg and epithelial cell barriers. Adv Exp Med Biol 763:149–170 70. Lee S, Fan S, Makarova O, Straight S, Margolis B (2002) A novel and conserved proteinprotein interaction domain of mammalian Lin-2/CASK binds and recruits SAP97 to the lateral surface of epithelia. Mol Cell Biol 22:1778–1791 71. Zhu Z-Q, Wang D, Xiang D, Yuan Y-X, Wang Y (2014) Calcium/calmodulin-dependent serine protein kinase is involved in exendin4-induced insulin secretion in INS-1 cells. Metabolism 63:120–126 72. Jiang M, Qiu J, Zhang L, Lu¨ D, Long M, Chen L, Luo X (2016) Changes in tension regulates proliferation and migration of fibroblasts by remodeling expression of ECM proteins. Exp Ther Med 12:1542–1550 73. Porter AP, White GRM, Mack NA, Malliri A (2019) The interaction between CASK and the tumour suppressor Dlg1 regulates mitotic spindle orientation in mammalian epithelia. J Cell Sci 132:jcs230086 74. Hannezo E, Prost J, Joanny J-F (2014) Theory of epithelial sheet morphology in three dimensions. Proc Natl Acad Sci U S A 111:27–32 75. Takeichi M (2014) Dynamic contacts: rearranging adherens junctions to drive epithelial remodelling. Nat Rev Mol Cell Biol 15:397–410 76. Becker S, Wandel E, Wobus M, Schneider R, Amasheh S, Sittig D, Kerner C, Naumann R, Hamann J, Aust G (2010) Overexpression of CD97 in intestinal epithelial cells of transgenic mice attenuates colitis by strengthening adherens junctions. PLoS One 5:e8507

Mechanoregulation of PDZ Proteins 77. Hilbig D, Sittig D, Hoffmann F, Rothemund S, Warmt E, Quaas M, Stu¨rmer J, Seiler L, Liebscher I, Hoang NA, K€as JA, Banks L, Aust G (2018) Mechanodependent phosphorylation of the PDZ-binding Motif of CD97/ADGRE5 modulates cellular detachment. Cell Rep 24:1986–1995 78. Sivaramakrishnan S, Schneider JL, Sitikov A, Goldman RD, Ridge KM (2009) Shear stress induced reorganization of the keratin intermediate filament network requires phosphorylation by protein kinase C ζ. Mol Biol Cell 20:2755–2765 79. O’Neill AK, Gallegos LL, Justilien V, Garcia EL, Leitges M, Fields A, Hall RA, Newton AC (2011) PKCα interacts with its novel substrate discs large homolog (DLG) 1 to promote cellular migration. J Biol Chem jbc. M111.294603 80. Kanzaki N, Ogita H, Komura H, Ozaki M, Sakamoto Y, Majima T, Ijuin T, Takenawa T, Takai Y (2008) Involvement of the nectinafadin complex in PDGF-induced cell survival. J Cell Sci 121:2008–2017 81. Sato T, Fujita N, Yamada A, Ooshio T, Okamoto R, Irie K, Takai Y (2006) Regulation of the assembly and adhesion activity of E-cadherin by nectin and afadin for the formation of adherens junctions in Madin-Darby canine kidney cells. J Biol Chem 281:5288–5299 82. Ooshio T, Kobayashi R, Ikeda W, Miyata M, Fukumoto Y, Matsuzawa N, Ogita H, Takai Y (2010) Involvement of the interaction of afadin with ZO-1 in the formation of tight junctions in Madin-Darby canine kidney cells. J Biol Chem 285:5003–5012 83. Severson EA, Lee WY, Capaldo CT, Nusrat A, Parkos CA (2009) Junctional adhesion molecule A interacts with afadin and PDZ-GEF2 to activate Rap1A, regulate β1 integrin levels, and enhance cell migration. Mol Biol Cell 20:1916–1925 84. Asada M, Irie K, Morimoto K, Yamada A, Ikeda W, Takeuchi M, Takai Y (2003) ADIP, a novel afadin- and α-actinin-binding protein localized at cell-cell adherens junctions. J Biol Chem 278:4103–4111 85. Mandai K, Nakanishi H, Satoh A, Obaishi H, Wada M, Nishioka H, Itoh M, Mizoguchi A, Aoki T, Fujimoto T, Matsuda Y, Tsukita S, Takai Y (1997) Afadin: a novel actin filament-binding protein with one PDZ domain localized at cadherin-based cell-tocell adherens junction. J Cell Biol 139:517–528

273

86. Mandai K, Nakanishi H, Satoh A, Takahashi K, Satoh K, Nishioka H, Mizoguchi A, Takai Y (1999) Ponsin/ SH3P12: an l-afadin– and vinculin-binding protein localized at cell–cell and cell–matrix adherens junctions. J Cell Biol 144:1001–1018 87. Ooshio T, Irie K, Morimoto K, Fukuhara A, Imai T, Takai Y (2004) Involvement of LMO7 in the association of two cell-cell adhesion molecules, nectin and E-cadherin, through afadin and alpha-actinin in epithelial cells. J Biol Chem 279:31365–31373 88. Takai Y, Nakanishi H (2003) Nectin and afadin: novel organizers of intercellular junctions. J Cell Sci 116:17–27 89. Monteiro AC, Sumagin R, Rankin CR, Leoni G, Mina MJ, Reiter DM, Stehle T, Dermody TS, Schaefer SA, Hall RA, Nusrat A, Parkos CA (2013) JAM-A associates with ZO-2, afadin, and PDZ-GEF1 to activate Rap2c and regulate epithelial barrier function. Mol Biol Cell 24:2849–2860 90. Fanning AS, Van Itallie CM, Anderson JM (2012) Zonula occludens-1 and -2 regulate apical cell structure and the zonula adherens cytoskeleton in polarized epithelia. Mol Biol Cell 23:577–590 91. Miyata M, Rikitake Y, Takahashi M, Nagamatsu Y, Yamauchi Y, Ogita H, Hirata K, Takai Y (2009) Regulation by afadin of cyclical activation and inactivation of Rap1, Rac1, and RhoA small G proteins at leading edges of moving NIH3T3 cells. J Biol Chem 284:24595–24609 92. Choi W, Acharya BR, Peyret G, Fardin M-A, Me`ge R-M, Ladoux B, Yap AS, Fanning AS, Peifer M (2016) Remodeling the zonula adherens in response to tension and the role of afadin in this response. J Cell Biol 213:243–260 93. Ebnet K, Schulz CU, Zu Brickwedde M-KM, Pendl GG, Vestweber D (2000) Junctional adhesion molecule interacts with the PDZ domain-containing proteins AF-6 and ZO-1. J Biol Chem 275:27979–27988 94. Scott DW, Tolbert CE, Burridge K (2016) Tension on JAM-A activates RhoA via GEF-H1 and p115 RhoGEF. Mol Biol Cell 27:1420–1430 95. Priya R, Yap AS (2015) Active tension: the role of cadherin adhesion and signaling in generating junctional contractility. Curr Top Dev Biol 112:65–102 96. Iden S, Misselwitz S, Peddibhotla SSD, Tuncay H, Rehder D, Gerke V, Robenek H, Suzuki A, Ebnet K (2012) aPKC

274

Elsa Bazellie`res and Andre´ Le Bivic

phosphorylates JAM-A at Ser285 to promote cell contact maturation and tight junction formation. J Cell Biol 196:623–639 97. Albertson R (2004) Scribble protein domain mapping reveals a multistep localization mechanism and domains necessary for establishing cortical polarity. J Cell Sci 117:6061–6070 98. Zeitler J, Hsu CP, Dionne H, Bilder D (2004) Domains controlling cell polarity and proliferation in the Drosophila tumor suppressor Scribble. J Cell Biol 167:1137–1146 99. Audebert S, Navarro C, Nourry C, ChasserotGolaz S, Le´cine P, Bellaiche Y, Dupont J-L, Premont RT, Sempe´re´ C, Strub J-M, Van Dorsselaer A, Vitale N, Borg J-P (2004) Mammalian Scribble forms a tight complex with the betaPIX exchange factor. Curr Biol 14:987–995 100. Kallay LM, McNickle A, Brennwald PJ, Hubbard AL, Braiterman LT (2006) Scribble associates with two polarity proteins, lgl2 and vangl2, via distinct molecular domains. J Cell Biochem 99:647–664 101. Wang X, Dong B, Zhang K, Ji Z, Cheng C, Zhao H, Sheng Y, Li X, Fan L, Xue W, Gao W-Q, Zhu HH (2018) E-cadherin bridges cell polarity and spindle orientation to ensure prostate epithelial integrity and prevent carcinogenesis in vivo. PLoS Genet 14:e1007609 102. Osmani N, Vitale N, Borg J-P, Etienne-Manneville S (2006) Scrib controls Cdc42 localization and activity to promote cell polarization during astrocyte migration. Curr Biol 16:2395–2405 103. Lohia M, Qin Y, Macara IG (2012) The Scribble polarity protein stabilizes E-cadherin/p120-catenin binding and blocks retrieval of E-cadherin to the golgi. PLoS One 7:e51130 104. le Duc Q, Shi Q, Blonk I, Sonnenberg A, Wang N, Leckband D, de Rooij J (2010) Vinculin potentiates E-cadherin mechanosensing and is recruited to actin-anchored sites within adherens junctions in a myosin II-dependent manner. J Cell Biol 189:1107–1115 105. Borghi N, Sorokina M, Shcherbakova OG, Weis WI, Pruitt BL, Nelson WJ, Dunn AR (2012) E-cadherin is under constitutive actomyosin-generated tension that is increased at cell–cell contacts upon externally applied stretch. Proc Natl Acad Sci U S A 109:12568–12573 106. Bazellie`res E, Conte V, Elosegui-Artola A, Serra-Picamal X, Bintanel-Morcillo M, Roca˜ oz JJ, Sales-Pardo M, Cusachs P, Mun

Guimera` R, Trepat X (2015) Control of cellcell forces and collective cell dynamics by the intercellular adhesome. Nat Cell Biol 17:409–420 107. Wu SK, Lagendijk AK, Hogan BM, Gomez GA, Yap AS (2015) Active contractility at E-cadherin junctions and its implications for cell extrusion in cancer. Cell Cycle 14:315–322 108. Engl W, Arasi B, Yap LL, Thiery JP, Viasnoff V (2014) Actin dynamics modulate mechanosensitive immobilization of E-cadherin at adherens junctions. Nat Cell Biol 16:584–591 109. Wu SK, Gomez GA, Michael M, Verma S, Cox HL, Lefevre JG, Parton RG, Hamilton NA, Neufeld Z, Yap AS (2014) Cortical F-actin stabilization generates apical-lateral patterns of junctional contractility that integrate cells into epithelia. Nat Cell Biol 16:167–178 110. Lugo-Martı´nez V-H, Petit CS, Fouquet S, Le Beyec J, Chambaz J, Pinc¸on-Raymond M, Cardot P, Thenet S (2009) Epidermal growth factor receptor is involved in enterocyte anoikis through the dismantling of E-cadherinmediated junctions. Am J Physiol Gastrointest Liver Physiol 296:G235–G244 111. Choi J, Troyanovsky RB, Indra I, Mitchell BJ, Troyanovsky SM (2019) Scribble, Erbin, and Lano redundantly regulate epithelial polarity and apical adhesion complex. J Cell Biol 218:2277–2293 112. Yoshihara K, Ikenouchi J, Izumi Y, Akashi M, Tsukita S, Furuse M (2011) Phosphorylation state regulates the localization of Scribble at adherens junctions and its association with Ecadherin-catenin complexes. Exp Cell Res 317:413–422 113. Meng W, Mushika Y, Ichii T, Takeichi M (2008) Anchorage of microtubule minus ends to adherens junctions regulates epithelial cell-cell contacts. Cell 135:948–959 114. Priya R, Yap AS, Gomez GA (2013) E-cadherin supports steady-state Rho signaling at the epithelial zonula adherens. Differentiation 86:133–140 115. Truong Quang B-A, Mani M, Markova O, Lecuit T, Lenne P-F (2013) Principles of E-cadherin supramolecular organization in vivo. Curr Biol 23:2197–2207 116. Hart KC, Tan J, Siemers KA, Sim JY, Pruitt BL, Nelson WJ, Gloerich M (2017) E-cadherin and LGN align epithelial cell divisions with tissue tension independently of cell shape. Proc Natl Acad Sci U S A 114(29):

Mechanoregulation of PDZ Proteins E5845–E5853. https://doi.org/10.1073/ pnas.1701703114 117. Gloerich M, Bianchini JM, Siemers KA, Cohen DJ, Nelson WJ (2017) Cell division orientation is coupled to cell–cell adhesion by the E-cadherin/LGN complex. Nat Commun 8:1–11 118. del Rio A, Perez-Jimenez R, Liu R, RocaCusachs P, Fernandez JM, Sheetz MP (2009) Stretching single talin rod molecules activates vinculin binding. Science 323:638–641 119. Mitrossilis D, Fouchard J, Pereira D, Postic F, Richert A, Saint-Jean M, Asnacios A (2010) Real-time single-cell response to stiffness. PNAS 107:16518–16523 120. Choquet D, Felsenfeld DP, Sheetz MP (1997) Extracellular matrix rigidity causes strengthening of integrin–cytoskeleton linkages. Cell 88:39–48

275

121. Stehbens SJ, Paszek M, Pemble H, Ettinger A, Gierke S, Wittmann T (2014) CLASPs link focal-adhesion-associated microtubule capture to localized exocytosis and adhesion site turnover. Nat Cell Biol 16:561–573 122. Sackmann E, Smith A-S (2014) Physics of cell adhesion: some lessons from cell-mimetic systems. Soft Matter 10:1644–1659 123. Riveline D, Zamir E, Balaban NQ, Schwarz US, Ishizaki T, Narumiya S, Kam Z, Geiger B, Bershadsky AD (2001) Focal contacts as mechanosensors externally applied local mechanical force induces growth of focal contacts by an mdia1-dependent and rockindependent mechanism. J Cell Biol 153:1175–1186 124. Kaverina I, Krylyshkina O, Beningo K, Anderson K, Wang Y-L, Small JV (2002) Tensile stress stimulates microtubule outgrowth in living cells. J Cell Sci 115:2283–2291

Chapter 16 Rational Design of PDZ Domain Inhibitors: Discovery of Small Organic Compounds Targeting PDZ Domains Laurent Hoffer, Philippe Roche, and Xavier Morelli Abstract PDZ domains, which belong to protein–protein interaction networks, are critical for regulating important biological processes such as scaffolding, trafficking, and signaling cascades. Interfering with PDZ-mediated interactions could affect these numerous biological processes. Thus, PDZ domains have emerged as promising targets to decipher biological phenomena and potentially treat cancer and neurological diseases. In this minireview, we focus on the discovery and design of small molecule inhibitors to modulate PDZ domains. These compounds interfere with endogenous protein partners from the PDZ domain by binding at the protein–protein interface. While peptides or peptidomimetic ligands were described to modulate PDZ domains, the focus of this review is on small organic compounds. Key words PDZ domain, PDZ inhibitors, Rational design, Screening, Molecular modeling

1

Introduction Postsynaptic density protein 95/Drosophila disc large tumor suppressor/Zonula occludens 1 (PDZ) proteins are a family of proteins that contain at least one PDZ domain. The binding of protein partners to PDZ domains mediates various processes, such as the formation of protein networks, the immobilization of these proteins in the correct cellular compartment and the ability of promoting scaffolding, trafficking, and signaling events [1–4]. The PDZ domain is a common structural domain of approximately 90 amino acids found in signaling proteins of many organisms. The global structure is conserved and usually consists of 5–6 β-strands and 2 α-helixes. In general, PDZ domains have a shallow binding site located between one β-strand and one α-helix that recognizes the C-terminus (terminal carboxylate and last residues) of their protein partners (Fig. 1). C-terminal residues from a protein partner form an additional antiparallel β-sheet by interacting with residues of the PDZ domain through hydrogen bonds with backbone atoms. In addition, the

Jean-Paul Borg (ed.), PDZ Mediated Interactions: Methods and Protocols, Methods in Molecular Biology, vol. 2256, https://doi.org/10.1007/978-1-0716-1166-1_16, © Springer Science+Business Media, LLC, part of Springer Nature 2021

277

278

Laurent Hoffer et al.

Fig. 1 3D crystallographic structure of the second PDZ domain of Syntenin (Synt-PDZ-2). (a) General fold of Synt-PDZ-2. The protein is shown as cartoon representation and the carboxylate binding loop is highlighted. (b) Structure of “Synt-PDZ-2/peptide” complex (PDB ID: 1OBY). The PDZ domain and peptide (C-terminus from Syndecan-4: TNEFYA) are displayed as surface and sticks representation, respectively. The shallow hydrophobic subpocket, recognizing the terminal hydrophobic tail from the peptide partner, is highlighted. (c) Detailed view of “Synt-PDZ-2/peptide” complex. The PDZ domain and its peptide partner are displayed as cartoon and sticks representation, respectively. Direct hydrogen bonds between both entities are represented as black dashed lines. Syntenin residues involved in hydrogen bonds with the peptide partner are colored in pink. (d) Two-dimensional interaction diagram between PDZ-2 domain of Syntenin and TNEFYA peptide partner. 3D structures were generated using Pymol (www.pymol.org) and MOE (https://www.chemcomp.com) was used to create the 2D interaction diagram

C-terminal carboxylate group from the protein partner interacts with the backbone of a conserved loop in the PDZ domain through a canonically conserved hydrogen bond network [2]. More precisely, PDZ domains typically require a hydrophobic residue at the C-terminal position of their protein partner, and this residue fits into a small hydrophobic pocket near the carboxylate binding loop. Typical hydrophobic residues are valine, leucine and isoleucine. By convention, P0 refers to the C-terminal residue of the peptide, and P-n refers to the nth amino acid of the peptide starting from its C-terminal end. PDZ domains have been classified according to their specificity for PDZ ligands. Consensus sequences

Rational Design of PDZ Domain Inhibitors

279

were built to define these groups [5]. Briefly, “class 1” PDZ domains recognize C-terminal sequences with either a serine or threonine at the 2 position. Similarly, “class 2” PDZ domains bind to C-terminal sequences with a large hydrophobic or aromatic residue at the 2 position. Finally, the consensus sequence recognized by “class 3” PDZ domains includes a negatively charged residue at the 2 position. Other positions, such as 1 and 3, have fewer constraints regarding the nature of the residue sidechain. Due to these small constraints, many peptides can bind different PDZ domains [6], so achieving high specificity is expected to be challenging. It is also known that some PDZ domains are able to bind to lipids. For instance, syntenin1 can interact with phosphatidylinositol phosphates with an affinity in the micromolar range [7].

2

PDZ Domains as Potential Drug Targets PDZ domains, which are critical for regulating important biological processes, have emerged as promising targets to treat cancer and neurological diseases [1, 8–13]. Early inhibitors of PDZ domains consisted of using short amino acid sequences representing the key C-terminal residues from endogenous partners. Later, modified peptides, including thioketone and nonnatural residues, were reported as biological tools to study PDZ domains [14, 15]. Bivalent peptides were also shown to exhibit high affinity for PDZ domains by simultaneously interacting with multiple PDZ domains from the same protein [16]. In addition, TAT-derived bivalent peptides, which contain cell permeability tags, were also developed as efficient probes [16]. Despite their potential high affinities, peptides may suffer from protein degradation by proteases and cell permeability issues. Thus, an appealing alternative strategy is to develop small molecule inhibitors for oral administration that bypass these issues. However, targeting PDZ domains may be very challenging because this requires tackling the general problem of protein–protein interactions (PPIs). It has been shown that modulating PPIs using small organic compounds is difficult due to the nature of the interface [17]. In general, such interfaces are large and flat with several small adjacent subpockets and are not expected to be easily druggable [18, 19]. Several screening studies concluded, as expected, that PDZ domains are mainly undruggable targets to be modulated by small organic molecules and even fragments [20, 21]. Indeed, the screening of fragment-like compounds is considered a powerful tool to assess the druggability of a given target [21]. These disappointing results are not truly surprising from a structural point of view; the nature of the PDZ domain binding pocket appears to be shallow with only a few putative

280

Laurent Hoffer et al.

Fujii et al JACS (2003)

Marcotte et al ProteinSci (2017)

Grandy et al JBioChem (2009)

Shan et al ChemBioDrugDes (2012)

Fujii et al, BioMedChemLet (2007) Fujii et al, CancerRes (2007)

Lin et al SciReports (2018)

Ma et al JCAMD (2018)

Thorsen et al PNAS (2010)

Bach et al OrgBioChem (2010)

Choi et al BioorgMedChem (2016)

Saupe et al ChemMedChem (2011)

Joshi et al AngewChem (2006)

Bouzidi et al BioMedChemLet (2013) Hori et al FrontierPharm (2018)

Kim et al MolecMed (2016)

Vogrig et al ACSChemBio (2013)

Mayasundari et al BioMedChemLet (2008)

Lee et al AngewChem (2009)

Vargas et al ChemMedChem (2014)

Cartier et al PLOS (2015)

Bach et al MedChemComm (2016)

Kegelman et al PNAS (2017)

Fig. 2 2D structures and references of small organic inhibitors discussed in the minireview

interacting hotspots. Despite the fact that PDZ domains are poorly suited for the development of small organic probes, different studies were reported in which inhibitor compounds were discovered. The ultimate goals are to develop probes to decipher PDZ-related biology and potential therapeutic drugs that target PDZ domains. However, most compounds exhibited inhibition in the 10 micromolar range, despite intensive structure–activity relationship (SAR) studies where close analogs are synthesized and tested. These compounds were discovered using either a screening-based strategy, molecular modeling experiment or a combination of both approaches. Currently reported PDZ inhibitors are quickly reviewed below, and their 2D structures are depicted in Fig. 2. Fujii et al. reported the first cell-permeable irreversible inhibitor targeting a PDZ domain (MAGI3-PDZ2) in 2003 [22]. The designed compound was a potent inhibitor of the interaction between PTEN and MAGI3, presenting an original covalent mode of action. Docking studies using the DOCK tool [23] suggested that the indole-3-carbinol moiety could mimic the hydrophobic end of the partner while preorienting chemical groups toward crucial residues from the MAGI3-PDZ2 domain. It should be noted that this compound was covalently bound to the MAGI3PDZ2 domain through the histidine residue H372. Starting from this covalent binder [22], Fujii et al. reported a reversible version of their indole-containing inhibitor for the same PDZ domain [24]. As previously described, DOCK software guided the rational

Rational Design of PDZ Domain Inhibitors

281

design of new compounds to mimic the four C-terminal residues of the protein partner. This ligand was able to displace the reference peptide probe in a concentration-dependent manner. However, the affinity of the compound for MAGI3-PDZ was relatively weak, as a competition was observed with concentrations in the 100 μM range. In another study, the same compound was able to disrupt the interaction between the Frizzed-7 Wnt receptor and the PDZ domain of Dishevelled [25], leading to the downregulation of the canonical Wnt signaling and suppression of tumor cell growth. According to this study, the compound was among the first nonpeptide inhibitors to show therapeutic efficacy through the disruption of a PDZ PPI. Finally, this indole core was also used to target the PDZ domain of NHERF1 [26]. Similar to the study from Fujii et al., the DOCK tool was employed to suggest putative inhibitors designed around the indole moiety [23]. In contrast to previous work, an additional carboxylate group was incorporated at the end of the flexible aliphatic sidechain of the compound to mimic the aspartic residue from the reference partner. In 2013, Vogrig et al. published small compounds able to disrupt the PSD-95-PDZ1/5-HT2A receptor interaction leading to an antihyperalgesic activity [27]. These inhibitors were discovered using a structure-based approach that combined molecular modeling and NMR. A series of indole analogues were synthesized on the basis of docking studies using AutoDock Vina software [28]. Their ability to bind to the first PDZ domain (PDZ1) of the PSD-95 protein was then assessed using NMR experiments. The best compound exhibited a moderate IC50 value of 190 μM. PDZ inhibitors, which were able to disrupt the PSD-95PDZ2/GluN2B PPI, were discovered using a rational “click chemistry” strategy [29]. The aim was to mimic the TAV/SAV tripeptide PSD-95 ligand using triazole-containing compounds that were easily synthesized from reactive azide and alkyne moieties. The triazole heterocycle was chosen because it was previously reported as a potential amide bioisostere while being more rigid and not recognized by protein peptidases. One triazole-containing compound inhibited the PSD-95-PDZ2/GluN2B interaction with an affinity similar to that measured for the SAV tripeptide. ITC experiments concluded that the compound had a low affinity for the PSD-95-PDZ2 domain (Kd value in the 600 μM range). Finally, molecular docking simulations using the Glide tool [30] suggested that the triazole-containing compound interacts with the PDZ2 domain in a similar way as the TAV tripeptide. In 2010, Thorsen et al. reported an organic compound (FSC231) able to bind to the PDZ domain from PICK1 [9]. The inhibitor was identified using a fluorescent polarization assay by screening approximately 44,000 compounds. This inhibitor exhibited an affinity similar to that measured for peptide ligands (C-terminal end from endogenous protein partners) in the 10 μM

282

Laurent Hoffer et al.

range. The AutoDock tool [31–33] was employed to predict the binding mode of the compound within the PDZ binding site. Molecular dynamics simulations were performed with the AMBER package [34] to refine the predicted binding mode. During subsequent SAR studies around FSC231, Bach et al. managed to slightly improve the Ki value by replacing chlorine atom in the meta position with a trifluoromethyl group [35]. More potent chemical inhibitors targeting the PDZ domain from PICK1 were recently published [36]. These inhibitors were able to modulate the amyloid beta-mediated synaptic dysfunction by interfering with the PICK1-PDZ/GluA2 PPI. Such potent compounds in the submicromolar range are interesting for the biological study of memory mechanisms and may be used as potential treatments for neurodegenerative disorders. An integrated strategy involving high-throughput screening, structure-based drug design, and biochemical and cellular assays was used to discover potent small molecule PICK1-PDZ inhibitors. This structure-based strategy relied on determining the protein–ligand structure using X-ray crystallography, followed by intensive SAR studies to increase the potency of the series. The X-ray crystallography experiments were more difficult than expected because conventional methods and conditions failed to produce any cocrystal structures. A new approach called the “lock and chop” method was developed to tackle this issue [37]. Analysis of the X-ray crystal structure of the complex (PDB ID: 6AR4) revealed that the chemical compound was located in the expected binding pocket and was able to tightly interact with a phenylalanine sidechain that was not targeted by the endogenous peptide. The most potent compound from the series exhibited an approximately 200-fold better potency than that from the C-terminal of the GluA2 partner. Selectivities with respect to other reference PDZ domains were also measured for this series of compounds. This remarkable integrated study reported the highest affinity for small molecule inhibitors of the PDZ domain known to date, with an IC50 value of 70 nM. The X-ray crystal structure, deposited in the protein databank (PDB ID: 6AR4), corresponds to a compound from the series with an IC50 value of 600 nM. Unfortunately, these compounds were unable to cross the blood brain barrier, preventing any potential use as drug candidates. However, they can still be used as chemical probes in biology studies due to their high potency. NMR spectroscopy-based screening allowed the detection of weakly binding inhibitors for the PDZ domain from AF6, which is an essential component of cell junctions [38]. A dissociation constant (Kd) value of 100 μM, which is in the same range as last residues from the endogenous EphB2 partner, was obtained with one analog designed around the rhodanine core. More intensive SAR studies around the same core were published several years later [39]. The design of new compounds was guided by molecular

Rational Design of PDZ Domain Inhibitors

283

modeling using the 3D structure of the PDZ domain of interest. The new derivatives were again evaluated using an NMR-based approach. The best compound, a mixture of diastereoisomers, exhibited a 5 μM affinity for AF6-PDZ. Molecular docking using the MOE package (http://www.chemcomp.com) was performed to identify the most likely active stereoisomer from the diastereoisomer mixture. Dishevelled (Dvl) is an essential protein in the Wnt signaling pathway that relies on its PDZ domain for transduction of downstream signals [11]. Interestingly, a known sulindac drug was shown to inhibit the canonical Wnt signaling pathway by binding to the PDZ domain of Dvl [40]. NMR experiments enabled the determination of the “sulindac / PDZ domain” complex and concluded that sulindac is located within the peptide binding pocket of the PDZ domain. Finally, a Ki value of 10 μM was measured for sulindac in a competitive assay with a reference peptide. Additional chemical compounds, which bind to the Dvl-PDZ domain in the low micromolar range affinity, were discovered using a protocol involving both molecular modeling and NMR spectroscopy [11]. In silico experiments employed a structure-based pharmacophore search using the Unity module from Sybyl [41] to identify small organic compounds that could mimic the binding mode of the Dapper protein partner within the PDZ domain. Then, the FlexX docking tool [42] was used to confirm the ability of selected compounds to act as potential PDZ domain binders. NMR spectroscopy was used as an experimental method to validate (or not) the selected compounds. A benzoic acid molecule, which displayed the most significant chemical shift perturbations, exhibited a 10 μM Kd value. Finally, in vivo studies confirmed its ability to reduce the growth rate of prostate cancer cell lines. A similar strategy was used in another study to target the PSD95-PDZ domain [43]. The binding of the best identified fragment (quinoline-2,7-dicarboxylic acid) was confirmed using NMR experiments. Another study focused on the Dvl protein identified small molecules that perturb the Dvl-PDZ/CXXC5 PPI [44]. Inhibition of this interaction may have potential interest in bone anabolic osteoporosis therapy by enhancing osteoblast differentiation [44]. More than 50 analogs were synthesized to explore the chemical space around the hit while also trying to increase the microsomal stability and optimize the physicochemical properties. Binding modes of representative compounds from the series were predicted using molecular docking with DOCKER from the Discovery Studio package [45]. The best compound, which exhibited a Dvl-PDZ binding affinity of 8 μM, successfully rescued bone loss in an ovariectomized mouse model. These authors also reported additional studies that focused on the same Dvl protein and employed various computational approaches to identify other PDZ inhibitors [46, 47]. For example, in the work reported by Ma et al. [47], X-ray

284

Laurent Hoffer et al.

crystal structures of Dvl-PDZ bound to an organic compound and snapshots from molecular dynamics simulations of the Dvl-PDZ/ peptide complex guided the creation of pharmacophore models. These models combined with a virtual screening of a large chemical library allowed the identification of compounds that could mimic the binding mode of reference molecules. Fluorescence spectroscopy and NMR experiments confirmed the binding of several compounds at the Dvl-PDZ–CXXC5 interface, and the best compound had a Kd value of 22 μM. Using a combination of NMR, quantitative structure–activity relationship (QSAR) and structure-based pharmacophore filtering, Shan et al. identified and optimized inhibitors for the Dvl-PDZ domain [48]. This series of compounds essentially consists of merging a benzoic acid moiety with two protein residues. The best compound from the series exhibited a Ki value of 1.5 μM in the fluorescence polarization assay. Potential binding modes were predicted using the Glide docking tool [30] and matched those from endogenous partners. Hori et al. reported new inhibitors for the Dvl-PDZ domain using the “NMR/Docking Performance Index” (NMR-DPI) protocol, which relies on both NMR and molecular docking experiments [49]. Several reference inhibitors were investigated with GOLD as the docking engine [50, 51] to select the best scoring scheme using different scoring functions (ChemScore, GoldScore, and ChemPLP) and with and without consensus scoring. The best scoring protocol was then employed for virtual screening with a focused library (approximately 5 K compounds). In total, 13 compounds were selected to be experimentally tested, and several of them showed partial proliferation inhibition activity against a triplenegative breast cancer cell line. Saupe et al. reported in 2011 a study about the Shank3-PDZ domain [10]. The ChemBioNet library was first screened using a fluorescence polarization assay, and one natural product-like scaffold (cyclopentyl-tetrahydroquinoline-carboxylates) emerged as a PDZ inhibitor. SAR studies around this core were performed to optimize its potency. The best compound analog exhibited a Ki value in the 10 μM range, and the binding of the compound within the PDZ domain was confirmed by NMR experiments. Then, X-ray crystallography studies were used to determine the structure of Shank3-PDZ/inhibitor complex (PDB ID: 3O5N) and to confirm its ability to mimic the C-terminal end of the protein partner. Kegelman et al. disclosed one chemical inhibitor, which targets the first PDZ domain of syntenin (Synt-PDZ1), using an integrated strategy involving an NMR-based screening of fragment-like compounds, SAR studies, and molecular modeling [13]. More precisely, approximately 5 K fragments were initially evaluated using an NMR-based screening. Two nonoverlapping fragments were identified, and a structure-based linking strategy was employed to

Rational Design of PDZ Domain Inhibitors

285

merge them. Ultimately, a combination of molecular docking studies using GOLD software [50, 51] and SAR studies produced a fused compound that exhibited a dissociation constant (Kd) of 21 μM for Synt-PDZ1. From a biological point of view, glioblastoma multiform (GBM) is one of the most aggressive cancers and is associated with short survival times and poor response to radiotherapy because of its invasive properties. It has been shown that syntenin is overexpressed in this kind of cancer. Both genetic and pharmacological strategies to modulate syntenin reduced invasion gains in GBM cells following radiation. Finally, intraperitoneal administration of the developed inhibitor improved the survival of brain tumor-bearing mice. Finally, we also reported an integrated study by combining proteomic and genetic techniques with structural biochemistry and molecular modeling, providing a detailed discovery of small compounds targeting the PDZ domain from GRASP55 [52]. The impacts on germ cell Golgi remodeling and spermatogenesis after administration of the compound were also studied. First, X-ray crystal structures of GRASP55 in complex with JAM-C or JAM-B were obtained and revealed that GRASP55 underwent conformational changes with respect to its free conformation, with the latter being more open. An in silico protocol involving high-throughput docking and pharmacophore filtering was performed to identify potential GRASP55-PDZ inhibitors. A library of 200,000 compounds dedicated to PPIs was used, and these compounds were docked into the binding site of the closed GRASP55-PDZ conformation using the Surflex docking tool [53], thereby producing millions of poses. Pharmacophore filtering, using Unity package from Sybyl [41], was then used to extract compounds that could mimic the canonical binding mode of the JAM peptides. Approximately 50 molecules were purchased and tested experimentally using homogeneous time-resolved fluorescence (HTRF), leading to the identification of a chemical compound that inhibited the GRASP55-PDZ–JAM interaction with an IC50 of 8 μM. Unfortunately, despite intensive efforts, an X-ray crystal structure for this compound in complex with GRASP55-PDZ was not obtained. In the end, the biological relevance of the GRASP55-PDZ–JAM-C interaction in spermatogenesis was validated using both genetic ablation of the encoding GRASP55 gene and disruption of this PPI using a small organic compound. Treatment of mice with the inhibitor induced premature release of spermatids and germ cell loss; thus, this inhibitor has potential to be used as male contraception.

286

3

Laurent Hoffer et al.

Conclusion There is growing interest in the development of compounds able to modulate PPIs, as they control a large number of physiological events and are involved in many diseases [54, 55]. However, PPIs are considered challenging targets for the development of chemical probes or drugs. PDZ domains belong to PPI networks, and therefore are essentially considered poor druggable targets. This is confirmed by the screening of large compound libraries in which no high affinity hits were yet identified [20]. Despite this classification as being poor druggable targets, dozens of studies have reported small organic compounds able to disrupt PPIs between PDZ domains and their endogenous protein partners. Various strategies involving a combination of experimental screening, biophysical methods, molecular modeling, and organic chemistry were employed to tackle the development of PDZ inhibitors. However, most nonpeptide compounds reported thus far exhibited moderate affinity in the 10 μM range. Promisingly, several integrative studies recently reported submicromolar compounds, enabling future opportunities for the use of PDZ inhibitors as important tools to decipher biological processes and as potential therapeutics to treat some cancers and neurological diseases.

References 1. Wang NX, Lee HJ, Zheng JJ (2008) Therapeutic use of PDZ protein-protein interaction antagonism. Drug News Perspect 21 (3):137–141 2. Chi CN, Bach A, Strømgaard K, Gianni S, Jemth P (2012) Ligand binding by PDZ domains. Biofactors 38(5):338–348. https:// doi.org/10.1002/biof.1031 3. Dev KK (2004) Making protein interactions druggable: targeting PDZ domains. Nat Rev Drug Discov 3(12):1047–1056. https://doi. org/10.1038/nrd1578 4. Grillo-Bosch D, Choquet D, Sainlos M (2013) Inhibition of PDZ domain-mediated interactions. Drug Discov Today Technol 10(4): e531–e540. https://doi.org/10.1016/j. ddtec.2012.10.003 5. Songyang Z, Fanning AS, Fu C, Xu J, Marfatia SM, Chishti AH, Crompton A, Chan AC, Anderson JM, Cantley LC (1997) Recognition of unique carboxyl-terminal motifs by distinct PDZ domains. Science 275(5296):73–77. https://doi.org/10.1126/science.275. 5296.73 6. Wiedemann U, Boisguerin P, Leben R, Leitner D, Krause G, Moelling K, Volkmer-

Engert R, Oschkinat H (2004) Quantification of PDZ domain specificity, prediction of ligand affinity and rational design of super-binding peptides. J Mol Biol 343(3):703–718. https://doi.org/10.1016/j.jmb.2004.08.064 7. Wawrzyniak AM, Vermeiren E, Zimmermann P, Ivarsson Y (2012) Extensions of PSD-95/discs large/ZO-1 (PDZ) domains influence lipid binding and membrane targeting of syntenin-1. FEBS Lett 586 (10):1445–1451. https://doi.org/10.1016/j. febslet.2012.04.024 8. Aarts M, Liu Y, Liu L, Besshoh S, Arundine M, Gurd JW, Wang YT, Salter MW, Tymianski M (2002) Treatment of ischemic brain damage by perturbing NMDA receptor- PSD-95 protein interactions. Science 298(5594):846–850. https://doi.org/10.1126/science.1072873 9. Thorsen TS, Madsen KL, Rebola N, Rathje M, Anggono V, Bach A, Moreira IS, StuhrHansen N, Dyhring T, Peters D, Beuming T, Huganir R, Weinstein H, Mulle C, Strømgaard K, Rønn LC, Gether U (2010) Identification of a small-molecule inhibitor of the PICK1 PDZ domain that inhibits hippocampal LTP and LTD. Proc Natl Acad Sci U S

Rational Design of PDZ Domain Inhibitors A 107(1):413–418. https://doi.org/10. 1073/pnas.0902225107 10. Saupe J, Roske Y, Schillinger C, Kamdem N, Radetzki S, Diehl A, Oschkinat H, Krause G, Heinemann U, Rademann J (2011) Discovery, structure-activity relationship studies, and crystal structure of nonpeptide inhibitors bound to the Shank3 PDZ domain. ChemMedChem 6 (8):1411–1422. https://doi.org/10.1002/ cmdc.201100094 11. Grandy D, Shan J, Zhang X, Rao S, Akunuru S, Li H, Zhang Y, Alpatov I, Zhang XA, Lang RA, Shi DL, Zheng JJ (2009) Discovery and characterization of a small molecule inhibitor of the PDZ domain of dishevelled. J Biol Chem 284 (24):16256–16263. https://doi.org/10. 1074/jbc.M109.009647 12. Patra CR, Rupasinghe CN, Dutta SK, Bhattacharya S, Wang E, Spaller MR, Mukhopadhyay D (2012) Chemically modified peptides targeting the PDZ domain of GIPC as a therapeutic approach for cancer. ACS Chem Biol 7(4):770–779. https://doi.org/10. 1021/cb200536r 13. Kegelman TP, Wu B, Das SK, Talukdar S, Beckta JM, Hu B, Emdad L, Valerie K, Sarkar D, Furnari FB, Cavenee WK, Wei J, Purves A, De SK, Pellecchia M, Fisher PB (2017) Inhibition of radiation-induced glioblastoma invasion by genetic and pharmacological targeting of MDA-9/Syntenin. Proc Natl Acad Sci U S A 114(2):370–375. https://doi.org/10.1073/pnas.1616100114 14. Bach A, Chi CN, Olsen TB, Pedersen SW, Røder MU, Pang GF, Clausen RP, Jemth P, Strømgaard K (2008) Modified peptides as potent inhibitors of the postsynaptic density95/N-methyl-D-aspartate receptor interaction. J Med Chem 51(20):6450–6459. https://doi.org/10.1021/jm800836w 15. Bach A, Eildal JN, Stuhr-Hansen N, Deeskamp R, Gottschalk M, Pedersen SW, Kristensen AS, Strømgaard K (2011) Cellpermeable and plasma-stable peptidomimetic inhibitors of the postsynaptic density-95/Nmethyl-D-aspartate receptor interaction. J Med Chem 54(5):1333–1346. https://doi. org/10.1021/jm1013924 16. Bach A, Clausen BH, Møller M, Vestergaard B, Chi CN, Round A, Sørensen PL, Nissen KB, Kastrup JS, Gajhede M, Jemth P, Kristensen AS, Lundstro¨m P, Lambertsen KL, Strømgaard K (2012) A high-affinity, dimeric inhibitor of PSD-95 bivalently interacts with PDZ1-2 and protects against ischemic brain damage. Proc Natl Acad Sci U S A 109(9):3317–3322. https://doi.org/10.1073/pnas.1113761109

287

17. Wells JA, McClendon CL (2007) Reaching for high-hanging fruit in drug discovery at protein-protein interfaces. Nature 450 (7172):1001–1009. https://doi.org/10. 1038/nature06526 18. Morelli X, Bourgeas R, Roche P (2011) Chemical and structural lessons from recent successes in protein-protein interaction inhibition (2P2I). Curr Opin Chem Biol 15 (4):475–481. https://doi.org/10.1016/j. cbpa.2011.1005.1024 19. Bourgeas R, Basse MJ, Morelli X, Roche P (2010) Atomic analysis of protein-protein interfaces with known inhibitors: the 2P2I database. PLoS One 5(3):e9598. https://doi. org/10.1371/journal.pone.0009598 20. Chen X, Longgood JC, Michnoff C, Wei S, Frantz DE, Bezprozvanny L (2007) Highthroughput screen for small molecule inhibitors of Mint1-PDZ domains. Assay Drug Dev Technol 5(6):769–783. https://doi.org/10. 1089/adt.2007.092 21. Hajduk P, Huth J, Fesik S (2005) Druggability indices for protein targets derived from NMR-based screening data. J Med Chem 48 (7):2518–2525. https://doi.org/10.1021/ jm049131r 22. Fujii N, Haresco JJ, Novak KA, Stokoe D, Kuntz ID, Guy RK (2003) A selective irreversible inhibitor targeting a PDZ protein interaction domain. J Am Chem Soc 125 (40):12074–12075. https://doi.org/10. 1021/ja035540l 23. Ewing TJ, Makino S, Skillman GA, Kuntz ID (2001) DOCK 4.0: search strategies for automated molecular docking of flexible molecule databases. J Comput Aided Mol Des 15 (5):411–428 24. Fujii N, Haresco JJ, Novak KA, Gage RM, Pedemonte N, Stokoe D, Kuntz ID, Guy RK (2007) Rational design of a nonpeptide general chemical scaffold for reversible inhibition of PDZ domain interactions. Bioorg Med Chem Lett 17(2):549–552. https://doi.org/10. 1016/j.bmcl.2006.10.006 25. Fujii N, You L, Xu Z, Uematsu K, Shan J, He B, Mikami I, Edmondson LR, Neale G, Zheng J, Guy RK, Jablons DM (2007) An antagonist of dishevelled protein-protein interaction suppresses beta-catenin-dependent tumor cell growth. Cancer Res 67 (2):573–579. https://doi.org/10.1158/ 0008-5472.CAN-06-2726 26. Mayasundari A, Ferreira AM, He L, Mahindroo N, Bashford D, Fujii N (2008) Rational design of the first small-molecule antagonists of NHERF1/EBP50 PDZ domains. Bioorg Med Chem Lett 18

288

Laurent Hoffer et al.

(3):942–945. https://doi.org/10.1016/j. bmcl.2007.12.038 27. Vogrig A, Dorr L, Bouzidi N, Boucherle B, Wattiez AS, Cassier E, Vallon G, Ripoche I, Abrunhosa-Thomas I, Marin P, Nauton L, Thery V, Courteix C, Lian LY, Ducki S (2013) Structure-based design of PDZ ligands as inhibitors of 5-HT(2A) receptor/PSD-95 PDZ1 domain interaction possessing antihyperalgesic activity. ACS Chem Biol 8 (10):2209–2216. https://doi.org/10.1021/ cb400308u 28. Trott O, Olson AJ (2010) AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J Comput Chem 31 (2):455–461. https://doi.org/10.1002/jcc. 21334 29. Bach A, Pedersen TB, Strømgaard K (2016) Design and synthesis of triazole-based peptidomimetics of a PSD-95 PDZ domain inhibitor. MedChemComm 7(3):531–536. https://doi. org/10.1039/C5MD00445D 30. Friesner RA, Banks JL, Murphy RB, Halgren TA, Klicic JJ, Mainz DT, Repasky MP, Knoll EH, Shelley M, Perry JK, Shaw DE, Francis P, Shenkin PS (2004) Glide: a new approach for rapid, accurate docking and scoring. 1. Method and assessment of docking accuracy. J Med Chem 47(7):1739–1749. https://doi.org/10. 1021/jm0306430 31. Morris GM, Goodsell DS, Halliday RS, Huey R, Hart WE, Belew RK, Olson AJ (1998) Automated docking using a Lamarckian genetic algorithm and an empirical binding free energy function. J Comput Chem 19 (14):1639–1662 32. Goodsell DS, Morris GM, Olson AJ (1996) Automated docking of flexible ligands: applications of AutoDock. J Mol Recognit 9(1):1–5 33. Morris GM, Huey R, Lindstrom W, Sanner MF, Belew RK, Goodsell DS, Olson AJ (2009) AutoDock4 and AutoDockTools4: automated docking with selective receptor flexibility. J Comput Chem 30(16):2785–2791. https://doi.org/10.1002/jcc.21256 34. Case DA, Cheatham TE, Darden T, Gohlke H, Luo R, Merz KM, Onufriev A, Simmerling C, Wang B, Woods RJ (2005) The Amber biomolecular simulation programs. J Comput Chem 26(16):1668–1688. https://doi.org/10. 1002/jcc.20290 35. Bach A, Stuhr-Hansen N, Thorsen TS, Bork N, Moreira IS, Frydenvang K, Padrah S, Christensen SB, Madsen KL, Weinstein H, Gether U, Strømgaard K (2010) Structure-activity relationships of a small-molecule inhibitor of the PDZ domain of PICK1. Org Biomol Chem 8

(19):4281–4288. https://doi.org/10.1039/ c0ob00025f 36. Lin EYS, Silvian LF, Marcotte DJ, Banos CC, Jow F, Chan TR, Arduini RM, Qian F, Baker DP, Bergeron C, Hession CA, Huganir RL, Borenstein CF, Enyedy I, Zou J, Rohde E, Wittmann M, Kumaravel G, Rhodes KJ, Scannevin RH, Dunah AW, Guckian KM (2018) Potent PDZ-domain PICK1 inhibitors that modulate amyloid beta-mediated synaptic dysfunction. Sci Rep 8(1):13438. https://doi. org/10.1038/s41598-018-31680-3 37. Marcotte DJ, Hus JC, Banos CC, Wildes C, Arduini R, Bergeron C, Hession CA, Baker DP, Lin E, Guckian KM, Dunah AW, Silvian LF (2018) Lock and chop: a novel method for the generation of a PICK1 PDZ domain and piperidine-based inhibitor co-crystal structure. Protein Sci 27(3):672–680. https://doi.org/ 10.1002/pro.3361 38. Joshi M, Vargas C, Boisguerin P, Diehl A, Krause G, Schmieder P, Moelling K, Hagen V, Schade M, Oschkinat H (2006) Discovery of low-molecular-weight ligands for the AF6 PDZ domain. Angew Chem Int Ed Engl 45 (23):3790–3795. https://doi.org/10.1002/ anie.200503965 39. Vargas C, Radziwill G, Krause G, Diehl A, Keller S, Kamdem N, Czekelius C, Kreuchwig A, Schmieder P, Doyle D, Moelling K, Hagen V, Schade M, Oschkinat H (2014) Small-molecule inhibitors of AF6 PDZ-mediated protein-protein interactions. ChemMedChem 9(7):1458–1462. https:// doi.org/10.1002/cmdc.201300553 40. Lee HJ, Wang NX, Shi DL, Zheng JJ (2009) Sulindac inhibits canonical Wnt signaling by blocking the PDZ domain of the protein Dishevelled. Angew Chem Int Ed Engl 48 (35):6448–6452. https://doi.org/10.1002/ anie.200902981 41. Tripos Sybyl. St. Louis, MO 42. Kramer B, Rarey M, Lengauer T (1999) Evaluation of the FLEXX incremental construction algorithm for protein-ligand docking. Proteins 37(2):228–241. https://doi.org/10.1002/( SICI)1097-0134(19991101)37:23.0.CO;2-8 43. Bouzidi N, Deokar H, Vogrig A, Boucherle B, Ripoche I, Abrunhosa-Thomas I, Dorr L, Wattiez AS, Lian LY, Marin P, Courteix C, Ducki S (2013) Identification of PDZ ligands by docking-based virtual screening for the development of novel analgesic agents. Bioorg Med Chem Lett 23(9):2624–2627. https://doi. org/10.1016/j.bmcl.2013.02.100 44. Kim HY, Choi S, Yoon JH, Lim HJ, Lee H, Choi J, Ro EJ, Heo JN, Lee W, No KT, Choi

Rational Design of PDZ Domain Inhibitors KY (2016) Small molecule inhibitors of the Dishevelled-CXXC5 interaction are new drug candidates for bone anabolic osteoporosis therapy. EMBO Mol Med 8(4):375–387. https:// doi.org/10.15252/emmm.201505714 45. AccelrysSoftware Discovery Studio. San Diego, CA 46. Choi J, Ma S, Kim HY, Yun JH, Heo JN, Lee W, Choi KY, No KT (2016) Identification of small-molecule compounds targeting the dishevelled PDZ domain by virtual screening and binding studies. Bioorg Med Chem 24 (15):3259–3266. https://doi.org/10.1016/j. bmc.2016.03.026 47. Ma S, Choi J, Jin X, Kim HY, Yun JH, Lee W, Choi KY, No KT (2018) Discovery of a smallmolecule inhibitor of Dvl-CXXC5 interaction by computational approaches. J Comput Aided Mol Des 32(5):643–655. https://doi.org/10. 1007/s10822-018-0118-x 48. Shan J, Zhang X, Bao J, Cassell R, Zheng JJ (2012) Synthesis of potent dishevelled PDZ domain inhibitors guided by virtual screening and NMR studies. Chem Biol Drug Des 79 (4):376–383. https://doi.org/10.1111/j. 1747-0285.2011.01295.x 49. Hori K, Ajioka K, Goda N, Shindo A, Takagishi M, Tenno T, Hiroaki H (2018) Discovery of potent disheveled/Dvl inhibitors using virtual screening optimized with NMR-based docking performance index. Front Pharmacol 9:983. https://doi.org/10. 3389/fphar.2018.00983 50. Jones G, Willett P, Glen R, Leach A, Taylor R (1997) Development and validation of a genetic algorithm for flexible docking. J Mol

289

Biol 267(3):727–748. https://doi.org/10. 1006/jmbi.1996.0897 51. Verdonk ML, Cole JC, Hartshorn MJ, Murray CW, Taylor RD (2003) Improved proteinligand docking using GOLD. Proteins 52 (4):609–623. https://doi.org/10.1002/prot. 10465 52. Cartier-Michaud A, Bailly AL, Betzi S, Shi X, Lissitzky JC, Zarubica A, Serge´ A, Roche P, Lugari A, Hamon V, Bardin F, Derviaux C, Lembo F, Audebert S, Marchetto S, Durand B, Borg JP, Shi N, Morelli X, Aurrand-Lions M (2017) Genetic, structural, and chemical insights into the dual function of GRASP55 in germ cell Golgi remodeling and JAM-C polarized localization during spermatogenesis. PLoS Genet 13(6):e1006803. https://doi.org/10.1371/journal.pgen. 1006803 53. Jain AN (2003) Surflex: fully automatic flexible molecular docking using a molecular similaritybased search engine. J Med Chem 46 (4):499–511. https://doi.org/10.1021/jm. 020406h 54. Petta I, Lievens S, Libert C, Tavernier J, De Bosscher K (2016) Modulation of proteinprotein interactions for the development of novel therapeutics. Mol Ther 24(4):707–718. https://doi.org/10.1038/mt.2015.214 55. Ivanov AA, Revennaugh B, Rusnak L, Gonzalez-Pecchi V, Mo X, Johns MA, Du Y, Cooper LAD, Moreno CS, Khuri FR, Fu H (2018) The OncoPPi portal: an integrative resource to explore and prioritize proteinprotein interactions for cancer target discovery. Bioinformatics 34(7):1183–1191. https://doi. org/10.1093/bioinformatics/btx743

INDEX A Actomyosin................................. 261, 263, 264, 266–268 Adherens junctions (AJ) ............................................... 222 Amide-to ester mutation ....................197, 206, 210–211 AVLX-144 ................................................... 158, 159, 163

B BIAcore......................................................................75–86 Biomolecular fragment complementation assay.......... 187 Biophysical techniques .................................................... 90

C Calcium/calmodulin-dependent serine protein kinase (CASK) ........................... 34, 138–140, 143 Computational approach ....................................... 66, 237

E Electropherogram superimposition ............................... 67 Epithelial cell polarity ................................................... 224 Equilibrium ......................... 62, 78, 84, 95, 99, 149–154 Expressed protein ligation (EPL)....................... 194–197, 199, 200, 211, 212

F Fluorescence ............................................... 66, 67, 73, 99, 107, 116, 137–147, 150, 151, 153, 161, 163, 169, 170, 180, 181, 184–186 Fluorescence anisotropy ............................................... 138 Fluorescence polarization (FP) ............................ 73, 161, 163, 169, 170, 180, 181, 184–186, 188 Folding ................................................... 3, 90, 91, 97–99, 104, 105, 114–119, 127, 130, 149–155 Forces.....................................................95, 238, 246, 247

G GAL4 ............................................................................. 1, 2

H Holdup assay .................................................... 61–72, 180

I Identity ..............................................90, 92–94, 138, 227

Implicit solvent............................................ 239, 240, 245 Inhibition......................................................157, 224–226 Isothermal titration calorimetry (ITC).......................127, 129, 130, 132, 170, 175, 180, 181, 186, 187, 190

K Kinetics ....................................................... 77, 78, 81–84, 86, 115, 149, 150, 152–155, 180, 196

L Ligand binding.................... 41, 117, 127, 137–147, 242 Lipid interactions ......................................................75–86 Liposomes....................................... 76–78, 80, 82, 83, 86 L1 sensor chips......................................78, 82, 83, 85, 86

M MC simulations .................................................... 242, 243 Molecular mechanics............................................ 238–240 Molecular modeling............................280, 281, 283–286

N NanoBiT ..............................................180, 183, 187–189 Native chemical ligation (NCL)................. 194, 212, 213 Next-generation sequencing (NGS) ...................... 42, 43, 46, 51, 56, 58

P Pathogenesis ...................... 221, 223, 225–227, 230, 232 PDZ-Binding Motifs (PBM).................................. 3, 5, 7, 11, 12, 14, 19, 24, 41, 61–73, 75, 86, 89, 99, 115, 117, 126, 129, 132, 137, 139–142, 145, 179–190, 217–232 PDZ/PBM interaction ............................... 137, 138, 145 Peptide interactions ..........................................75, 81, 82, 84, 85, 125–132 Phage display ........................................... 41–58, 126, 130 Phosphorylation ..............................................42, 48, 126, 179–190, 197, 206, 224, 239 Protein complexes .............................................17, 18, 21, 22, 24, 26, 30–31, 34, 37, 76, 127, 129, 223, 258 Protein design ...................................................... 237, 250 Protein engineering ...................................................... 116 Protein modifications.................................................... 193 Protein-protein binding................................................ 218

Jean-Paul Borg (ed.), PDZ Mediated Interactions: Methods and Protocols, Methods in Molecular Biology, vol. 2256, https://doi.org/10.1007/978-1-0716-1166-1, © Springer Science+Business Media, LLC, part of Springer Nature 2021

291

PDZ MEDIATED INTERACTIONS: METHODS

292 Index

AND

PROTOCOLS

Protein-protein interaction (PPIs) ..................... 1, 61, 89, 127, 137, 157, 169, 170, 179, 187–189, 193, 217, 258, 279 Protein sample......................................................... 93, 95, 106–109, 112, 130, 131, 146, 204–206 Protein stability ............................................................. 146 Proteomics.....................17, 32, 38, 41–58, 93, 126, 285 Proteus program .................................................. 243, 247 PSD-95/Discs Large/ZO-1 (PDZ).........................1–14, 17–39, 41–58, 61–72, 75–86, 89–121, 125–132, 137–147, 149–176, 179–190, 193–215, 217–232, 237–253 domains ..................................... 3, 18, 41, 61, 75, 89, 126, 137, 143, 149, 152, 179, 194, 217, 230 inhibitors ................................................................. 159 lipids..................................................................... 75–86 proteins ...........................................19, 34, 76, 84, 85, 126, 137, 141, 143, 146, 150, 197, 219, 222, 224, 226, 231, 237, 250 Purification .....................................4, 8, 9, 17–39, 42–45, 47–49, 51, 52, 59, 62, 106, 109, 138–140, 146, 162, 170, 175, 181, 183, 198–206, 213 Purity ....................................................24, 37, 51, 52, 76, 90–92, 99, 103, 105–110, 139, 165, 166, 194, 204, 213

Q Quality control ........................................ 25, 29, 118, 120 Quantification ............................................ 20, 46, 47, 58, 61, 91, 92, 95, 107, 108, 140, 207

R Rational design..................................................... 277–286 Replication..................................221, 226, 227, 230, 232

S SA sensor chip ............................................. 77, 79, 84, 85

Screening .................................................... 2–4, 7, 10, 11, 13, 14, 159, 163, 279–282, 284, 286 Scribble ............................................... 125–132, 138–140, 145, 223–226, 231, 261, 267–269 Scribble module .......................................... 125, 127, 128 Sensor chip (SA)................................................. 77, 81, 82 SH3-containing guanine nucleotide exchange factor (SGEF) .................................. 128, 139, 145 Solid-phase peptide synthesis ............................. 159, 163, 165, 166, 193, 200, 206 Specificity ........................................... 69, 70, 75, 90, 126, 127, 137, 138, 196, 197, 219, 238, 244, 258, 278, 279 Stability ...................................... 86, 90, 97–99, 104, 105, 114–119, 146, 151, 152, 162, 171, 173, 188, 223, 237–239, 241, 262, 269, 283 Stroke...................................................158, 159, 163, 172 Structural biology ......................................................... 282 Structure ..................................................... 90, 91, 97, 99, 114, 116, 118, 126, 127, 130, 132, 149, 150, 217, 237, 241, 242, 244, 247, 248, 258, 277, 278, 280, 282–285

T Tight junction (TJs)............................217, 222, 259–261 Two-hybrid array ............................................. 11–13, 126

V Virus...........................140, 221–223, 225–227, 230, 232

X X-ray crystallography .................127, 129, 130, 282, 284

Y Yeast strains ............................................. 2, 4, 6, 9, 11, 14 Yeast two-hybrid (Y2H) ........................... 1–14, 126, 230