Recombinant Proteins in Plants: Methods and Protocols (Methods in Molecular Biology, 2480) 9781071622407, 9781071622414, 1071622404

Citation preview

Methods in Molecular Biology 2480

Stefan Schillberg Holger Spiegel Editors

Recombinant Proteins in Plants Methods and Protocols

METHODS

IN

MOLECULAR BIOLOGY

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

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

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

Recombinant Proteins in Plants Methods and Protocols

Edited by

Stefan Schillberg and Holger Spiegel Fraunhofer IME, Aachen, Nordrhein-Westfalen, Germany

Editors Stefan Schillberg Fraunhofer IME Aachen, Nordrhein-Westfalen, Germany

Holger Spiegel Fraunhofer IME Aachen, Nordrhein-Westfalen, Germany

ISSN 1064-3745 ISSN 1940-6029 (electronic) Methods in Molecular Biology ISBN 978-1-0716-2240-7 ISBN 978-1-0716-2241-4 (eBook) https://doi.org/10.1007/978-1-0716-2241-4 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022 Chapters 6, 8, and 17 are licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/). For further details see license information in the chapter. This work is subject to copyright. All rights are solely and exclusively licensed 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 Various classes of proteins such as enzymes, fluorescent proteins, antibodies, hormones, toxins, and membrane proteins are essential tools in the chemical, cosmetic, agricultural, and pharmaceutical industries enabling sustainable and defined processes that are more difficult to realize with chemical synthesis and small molecules due to their rather simple structure and lower specificity. However, advances in biotechnological research and bioeconomic applications rely on the provision of sufficient quantities of proteins of high quality and functionality. Genetic engineering has revolutionized protein production by heterologous expression of designed genes in various organisms. Since the first recombinant protein expression in Escherichia coli in 1976 also multiple eukaryotic protein production systems including yeast, insect, and mammalian cells have been developed, providing the opportunity for more complex, glycosylated protein production. Together with E. coli, eukaryotic cell systems such as the yeast Pichia pastoris and Chinese hamster ovary (CHO) cells now form the mainstay of industrial protein production. Since more than 30 years also plants have been proven as a suitable platform for the production of recombinant proteins including complex multi-domain and/or glycosylated proteins. In contrast to the conventional prokaryotic, insect and mammalian cell-based systems, plants can be grown in different phenotypes, i.e., whole plants, hairy roots as well as suspended cells or tissue. This flexibility expands possible production scenarios and the associated benefits, such as fast and easy scaling through increasing parallel cultivation of whole plants or secretion of the target protein into the culture medium in the case of suspension cells. Besides the stable integration of the target gene into the plant genome, transient plant expression systems have been developed allowing the rapid provision of recombinant proteins in the milligram to gram range within less than a week. Furthermore, recombinant proteins can be produced in plants without using animal-derived proteins or bacterial endotoxins (e.g., for media preparation), which is beneficial for certain research applications (e.g., plant-produced hormones for animal cell culture) or production of pharmaceutical proteins for some religious communities, vegans, or people with animal allergies. Also the plant matrix itself might be of advantage—plant-produced vaccines or therapeutics, or cosmetics may even be used without purification avoiding expensive purification steps. It is therefore not surprising that many researches have implemented plant-based production systems to provide sufficient amounts of high-quality proteins for their research applications, and some R&D efforts have reached market maturity as diagnostic, cosmetic, and pharmaceutical proteins. Despite the advantages described the spread of plant production platforms is often complicated by the implementation of appropriate and easy-to-use process protocols. In addition, further optimization with respect to production levels and product recovery and purification is required to compete with industrial microbial and mammalian production systems. Therefore, this book provides understandable and user-friendly descriptions for recombinant protein production in different plant systems and subsequent downstream processing as well as strategies to optimize protein expression and recovery. Besides detailed laboratory protocols, some chapters are also more descriptive to provide additional information, e.g., on technology assessment, and regulation and legal aspects. The book begins with a short overview on recombinant protein production in plants (Chapter 1) referring to and putting into context the remaining 17 chapters. Part I includes protocols for

v

vi

Preface

recombinant protein production in important plant systems: (1) whole plants such as tobacco (Nicotiana tabacum) and cereals (barley—Hordeum vulgare, wheat—Triticum aestivum), (2) cell suspension cultures (N. tabacum cv. BY-2), (3) transient expression in leaves and whole plants (N. benthamiana), as well as more emerging systems like the cellfree expression in tobacco BY-2 lysates (Chapters 2–8). The second part describes strategies to improve the subsequent process step of protein recovery and purification and also takes into account the techno-economic modeling and simulation of the plant manufacturing process (Chapters 9–11). Part III focusses on different strategies to optimize productivity exploiting cloning and fusion protein approaches, tools to increase yield in transient expression experiments, and statistical designs to maximize downstream processing as well as genome editing to manipulate glycosylation for improved protein quality (Chapters 12–16). The fourth and final part provides additional information on the regulation and freedom to operate analysis of plant-produced proteins and is of particular interest when products are driven towards the market (Chapters 17 and 18). The protocols and information provided in this book will be useful to newcomers but also to more experienced researchers interested in expanding their expertise or even plan to start a business in the field of plant-based protein production. The described procedures enable researchers to produce proteins of high quality and functionality for various research applications in molecular biology, cell biology, biochemistry, crystallization, biohybrid technologies, synthetic biology, etc. We are grateful to all authors for sharing their procedures and protocols developed and tested in their labs as well as the associated knowledge and experience, which is of major importance to distribute the technology of recombinant protein production in plants. Aachen, Germany

Stefan Schillberg Holger Spiegel

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

v ix

1 Recombinant Protein Production in Plants: A Brief Overview of Strengths and Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stefan Schillberg and Holger Spiegel

1

PART I

PLANT PRODUCTION SYSTEMS AND APPLICATIONS

2 Production of Recombinant Proteins in Transgenic Tobacco Plants . . . . . . . . . . . 17 Tim H. Szeto, Pascal M. W. Drake, Audrey Y-H. Teh, Nicole Falci Finardi, Ashleigh G. Clegg, Mathew J. Paul, Rajko Reljic, and Julian K-C. Ma 3 Molecular Farming in Seed Crops: Gene Transfer into Barley (Hordeum vulgare) and Wheat (Triticum aestivum) . . . . . . . . . . . . . . . . . . . . . . . . . 49 Eszter Kapusi and Eva Stoger 4 Cell Biology Methods to Study Recombinant Proteins in Seeds. . . . . . . . . . . . . . . 61 Elsa Arcalı´s, Emanuela Pedrazzini, Ulrike Ho¨rmann-Dietrich, Alessandro Vitale, and Eva Stoger 5 Production of Recombinant Glycoproteins in Nicotiana tabacum BY-2 Suspension Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Catherine Navarre and Franc¸ois Chaumont 6 Production of Recombinant Proteins by Agrobacterium-Mediated Transient Expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 Holger Spiegel, Stefan Schillberg, and Greta No¨lke 7 Specific Packaging of Custom RNA Molecules into Cowpea Mosaic Virus-like Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 Hadrien Peyret and George P. Lomonossoff 8 Plant-Based Cell-Free Transcription and Translation of Recombinant Proteins. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Matthias Buntru, Simon Vogel, Ricarda Finnern, and Stefan Schillberg

PART II

DOWNSTREAM PROCESSING

9 Strategies for Efficient and Sustainable Protein Extraction and Purification from Plant Tissues. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 Johannes F. Buyel 10 Improving Recombinant Protein Recovery from Plant Tissue Using Heat Precipitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 Holger Spiegel

vii

viii

11

Contents

Technoeconomic Modeling and Simulation for Plant-Based Manufacturing of Recombinant Proteins. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 Matthew J. McNulty, Somen Nandi, and Karen A. McDonald

PART III 12

13

14

15

16

Optimization of Vectors and Targeting Strategies Including GoldenBraid and Genome Editing Tools: GoldenBraid Assembly of Multiplex CRISPR/Cas12a Guide RNAs for Gene Editing in Nicotiana benthamiana . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Beatriz Gonza´lez, Marta Vazquez-Vilar, Javier Sa´nchez-Vicente, and Diego Orza´ez Advanced Fusion Strategies for the Production of Functionalized Potato Virus X Virions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Christina Dickmeis and Ulrich Commandeur Knockout of Glycosyltransferases in Nicotiana benthamiana by Genome Editing to Improve Glycosylation of Plant-Produced Proteins . . . . . Julia Jansing and Luisa Bortesi A Bioluminescent Agrobacterium tumefaciens for Imaging Bacterial Metabolic Activity in Planta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Philippe V. Jutras, Isobel Dodds, and Renier A. L. van der Hoorn Statistical Designs to Improve Downstream Processing . . . . . . . . . . . . . . . . . . . . . . Johannes F. Buyel

PART IV 17

18

OPTIMIZATION STRATEGIES

193

215

241

285 295

REGULATORY ISSUES

Regulation of Molecular Farming Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313 Penny A. C. Hundleby, Marc-Andre´ D’Aoust, Carolyn Finkle, Judith Atkins, and Richard M. Twyman Freedom to Operate Analysis of Molecular Farming Projects . . . . . . . . . . . . . . . . . 335 Harry Thangaraj

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

343

Contributors ELSA ARCALI´S • Department of Applied Genetics and Cell Biology, University of Natural Resources and Life Sciences, Vienna, Austria JUDITH ATKINS • Medicago R & D, Que´bec, QC, Canada LUISA BORTESI • Aachen-Maastricht Institute for Biobased Materials (AMIBM), Maastricht University, Geleen, The Netherlands MATTHIAS BUNTRU • Fraunhofer Institute for Molecular Biology and Applied Ecology IME, Aachen, Germany JOHANNES F. BUYEL • Fraunhofer Institute for Molecular Biology and Applied Ecology IME, Aachen, Germany; Institute for Molecular Biotechnology, Worringerweg 1, RWTH Aachen University, Aachen, Germany FRANC¸OIS CHAUMONT • Louvain Institute of Biomolecular Science and Technology, UCLouvain, Louvain-la-Neuve, Belgium ASHLEIGH G. CLEGG • Hotung Molecular Immunology Unit, St. George’s University of London, Institute for Infection and Immunity, London, UK ULRICH COMMANDEUR • Institute for Molecular Biotechnology, RWTH Aachen University, Aachen, Germany MARC-ANDRE´ D’AOUST • Medicago R & D, Que´bec, QC, Canada CHRISTINA DICKMEIS • Institute for Molecular Biotechnology, RWTH Aachen University, Aachen, Germany ISOBEL DODDS • Plant Chemetics Lab, Department of Plant Sciences, University of Oxford, Oxford, UK PASCAL M. W. DRAKE • Hotung Molecular Immunology Unit, St. George’s University of London, Institute for Infection and Immunity, London, UK NICOLE FALCI FINARDI • Hotung Molecular Immunology Unit, St. George’s University of London, Institute for Infection and Immunity, London, UK CAROLYN FINKLE • Medicago R & D, Que´bec, QC, Canada RICARDA FINNERN • LenioBio GmbH, Du¨sseldorf, Germany BEATRIZ GONZA´LEZ • IBMCP-CSIC-UPV, Valencia, Spain ULRIKE HO¨RMANN-DIETRICH • Department of Applied Genetics and Cell Biology, University of Natural Resources and Life Sciences, Vienna, Austria PENNY A. C. HUNDLEBY • John Innes Centre, Norwich Research Park, UK JULIA JANSING • Aachen-Maastricht Institute for Biobased Materials (AMIBM), Maastricht University, Geleen, The Netherlands PHILIPPE V. JUTRAS • Plant Chemetics Lab, Department of Plant Sciences, University of Oxford, Oxford, UK ESZTER KAPUSI • Department of Applied Genetics and Cell Biology, University of Natural Resources and Life Sciences, Vienna, Austria GEORGE P. LOMONOSSOFF • Department of Biochemistry and Metabolism, John Innes Centre, Norwich, UK JULIAN K-C. MA • Hotung Molecular Immunology Unit, St. George’s University of London, Institute for Infection and Immunity, London, UK

ix

x

Contributors

KAREN A. MCDONALD • Department of Chemical Engineering, University of California, Davis, CA, USA; Global HealthShare Initiative, University of California, Davis, CA, USA MATTHEW J. MCNULTY • Department of Chemical Engineering, University of California, Davis, CA, USA SOMEN NANDI • Department of Chemical Engineering, University of California, Davis, CA, USA; Global HealthShare Initiative, University of California, Davis, CA, USA CATHERINE NAVARRE • Louvain Institute of Biomolecular Science and Technology, UCLouvain, Louvain-la-Neuve, Belgium GRETA NO¨LKE • Fraunhofer Institute for Molecular Biology and Applied Ecology IME, Aachen, Germany DIEGO ORZA´EZ • IBMCP-CSIC-UPV, Valencia, Spain MATHEW J. PAUL • Hotung Molecular Immunology Unit, St. George’s University of London, Institute for Infection and Immunity, London, UK EMANUELA PEDRAZZINI • Istituto di Biologia e Biotecnologia Agraria, CNR, Milan, Italy HADRIEN PEYRET • Department of Biochemistry and Metabolism, John Innes Centre, Norwich, UK RAJKO RELJIC • Hotung Molecular Immunology Unit, St. George’s University of London, Institute for Infection and Immunity, London, UK JAVIER SA´NCHEZ-VICENTE • IBMCP-CSIC-UPV, Valencia, Spain STEFAN SCHILLBERG • Fraunhofer Institute for Molecular Biology and Applied Ecology IME, Aachen, Germany; Department of Phytopathology, Justus Liebig University Giessen, Giessen, Germany HOLGER SPIEGEL • Fraunhofer Institute for Molecular Biology and Applied Ecology IME, Aachen, Germany EVA STOGER • Department of Applied Genetics and Cell Biology, University of Natural Resources and Life Sciences, Vienna, Austria TIM H. SZETO • Hotung Molecular Immunology Unit, St. George’s University of London, Institute for Infection and Immunity, London, UK AUDREY Y-H. TEH • Hotung Molecular Immunology Unit, St. George’s University of London, Institute for Infection and Immunity, London, UK HARRY THANGARAJ • Independent Consultant, (Residential Address Withheld), London, UK RICHARD M. TWYMAN • TRM Ltd, Scarborough, UK RENIER A. L. VAN DER HOORN • Plant Chemetics Lab, Department of Plant Sciences, University of Oxford, Oxford, UK MARTA VAZQUEZ-VILAR • IBMCP-CSIC-UPV, Valencia, Spain ALESSANDRO VITALE • Istituto di Biologia e Biotecnologia Agraria, CNR, Milan, Italy SIMON VOGEL • Fraunhofer Institute for Molecular Biology and Applied Ecology IME, Aachen, Germany

Chapter 1 Recombinant Protein Production in Plants: A Brief Overview of Strengths and Challenges Stefan Schillberg

and Holger Spiegel

Abstract The first recombinant proteins were produced in microbes and animal cells cultivated in bioreactors. These systems have become the standard for industrial-scale recombinant protein manufacturing. Later, the production of recombinant proteins was demonstrated in whole plants, which differ morphologically from cell-based systems and require completely different cultivation conditions. Over time, additional plant-based production platforms were established, including hairy roots and cell suspension cultures, which are more similar to conventional cell-based systems in terms of morphology, procedures, and equipment requirements. In this brief overview of the field, we explain why plant-based systems are becoming increasingly attractive for the production of valuable proteins with scientific and commercial applications, but also highlight the challenges that these systems must overcome to achieve more widespread industrial utilization. We discuss various laboratory protocols and approaches for the production of recombinant proteins in plants, as well as strategies to optimize yields, and the regulatory and legal framework. Key words Biopharmaceuticals, Downstream processing, Molecular farming products, Optimization strategies, Plant-based expression systems, Plant molecular farming

1

Why Are Plants Used for Protein Production? The production of recombinant proteins in plants became possible following the first successful transformation of plant cells in 1983 [1] and the creation of the first transgenic plants in 1984 [2]. Interest in the genetic modification of plants grew very quickly, and initially the main application was to improve the agronomic properties of food, feed and ornamental crops, allowing the creation of new commercial plant varieties. However, the potential for recombinant protein expression in plants was soon discovered. The first such reports involved the production of functional full-length antibodies in the green alga Chlamydomonas reinhardtii (published in 1987) [3] and in tobacco plants (published in 1989) [4]. The successful production of human serum albumin in transgenic

Stefan Schillberg and Holger Spiegel (eds.), Recombinant Proteins in Plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480, https://doi.org/10.1007/978-1-0716-2241-4_1, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

1

2

Stefan Schillberg and Holger Spiegel

potato and tobacco plants, as well as tobacco cell cultures derived from the transgenic line, was reported in 1990 [5]. Since then, numerous and diverse recombinant proteins, including complex multi-domain and transmembrane proteins, have been produced in various plant species (for review see [6–8]). This approach is often described as plant-based molecular farming (see Note 1). At first, the inexpensive cultivation of transgenic plants in the open field was seen as a driver to reduce overall production costs, providing a competitive advantage over cell-based expression systems that require purpose-built facilities. However, the production of therapeutic proteins requires controlled and contained facilities such as a greenhouse or growth chamber to avoid environmental contamination and to maintain product consistency and safety (see Note 2). Furthermore, the recovery of recombinant proteins from plant tissue (and the removal of endogenous plant proteins during purification) was found to cost more than established platforms, mainly because there was no efficient and widely applicable approach to allow the secretion of recombinant proteins into the medium, which is the typical strategy used in cell suspension cultures [7]. Despite these drawbacks, plants offer some unique advantages over conventional microbial and mammalian cell-based production systems (Fig. 1), and these are summarized below.

Strengths

Challenges

Speed of production when using transient expression systems

Lower productivity compared to industrial microbial and mammalian systems

Safe plant matrix avoiding expensive purification when using food plants

Higher protein degradation due to high protease activity in some plant systems

Animal-free production avoiding the use of any animal-derived reagents

Higher downstream processing costs for proteins produced in whole plants

Consumer acceptance for ‘green’ production of lifestyle products

Rather uncertain regulatory landscape compared to industrial production platforms

Improved protein functionality when requiring altered (plant-specific) glycosylation

No accessible commercial plant production system with clear IP portfolio

Fig. 1 Strengths of plants for the production of recombinant proteins and current challenges to make this platform more competitive

A Brief Overview of Strengths and Challenges

3

First, plants offer two broad approaches for recombinant protein production, which are suitable for different target markets. Traditionally, the transgene encoding the protein of interest is stably integrated into the plant genome, resulting in a transgenic line that transmits the ability to produce the same recombinant protein to subsequent generations via the seeds or vegetative propagation. This is ideal for the continuous, long-term, large-scale production of proteins with a large demand, although it takes additional time to generate and propagate a sufficient number of transgenic plants. Alternatively, proteins can be produced within a few days by transient expression, which does not require the creation of transgenic plants (see Note 3). A widely used transient expression method is agroinfiltration, in which an Agrobacterium tumefaciens suspension carrying the plant expression vector is introduced into leaf tissue by vacuum infiltration or using a needleless syringe (see Note 4). Several approaches for transient expression by agroinfiltration have been described, and current protocols can be found in this book [9–11]. The speed of transient expression is particularly relevant in emergencies. In this scenario, cell-based production platforms tend to be occupied by routine commercial manufacturing processes and cannot be adapted quickly to emerging needs. Transient expression systems are more flexible because they are easy to scale up simply by cultivating more plants, producing Agrobacterium in larger bioreactors, and carrying out the agroinfiltration with whole plants in large vacuum tanks to generate gram quantities of the final product [12]. Therefore, proteins for diagnostic assays (such as lateral flow and ELISA tests) or for the development of vaccines can be provided within a few weeks or months after confirming the sequence of the pathogen, which is ideal for new outbreaks of epidemic and pandemic diseases including influenza and COVID-19 [13]. Second, the use of food or feed crops (which generally lack harmful substances) removes the need to purify recombinant proteins from the plant tissue matrix. For example, oral and injectable veterinary vaccines and therapeutics can be administered directly as unprocessed or minimally processed plant tissues/extracts, thus avoiding the expense of rigorous purification and satisfying the need for inexpensive veterinary products compared to human biopharmaceuticals. This is a major economic strength of plants because recombinant proteins produced bacteria and mammalian cells must be extensively purified to ensure the removal of bacterial endotoxins and potential mammalian pathogens, respectively (see Notes 5 and 6). Third, plant systems remove the need for any animal-derived reagents during the production process. Media for the cultivation of microbial cells often contain animal proteins as a nutritional source, and mammalian cells of course originate from animals. Accordingly, plants are particularly suitable for the production of

4

Stefan Schillberg and Holger Spiegel

biopharmaceuticals for people with animal allergies, vegans and some religious communities. It is also advantageous to produce diagnostic proteins without contaminating animal proteins or endotoxins to avoid interference in tests with mammalian cells, tissues, or organs (see Note 7). Fourth, the sustainability of plants can be used to promote the distribution of lifestyle products. The first such product was recombinant human epidermal growth factor produced in barley for use in a skin care serum, and several additional products of this kind are in the development pipeline. Like the veterinary biopharmaceuticals described above, proteins used as cosmetic ingredients do not necessarily require extensive purification, so this would similarly reduce manufacturing costs. However, unpurified extracts from green tissue retain the original green coloring, and extracts from cereal grains are rich in starch, which makes the resulting products sticky. Therefore, cosmetic proteins are currently purified from plant tissue to avoid these undesirable effects. Even so, the use of plants for the production of recombinant proteins as cosmetic products may improve the perception of the final product because plants are recognized as more natural and environmentally sustainable than production systems based on microbes or mammalian cells. Fifth, plants can be used to generate unique and tailored glycan structures. Many recombinant proteins contain N-linked glycans, but glycans synthesized in plants contain core β(1,2)-xylose and α(1,3)-fucose residues and thus differ slightly from the glycans present in humans. This led to concerns that glycoproteins produced in plants may induce immune responses in human patients. However, this potential drawback also brings advantages, because vaccines and certain biopharmaceuticals for cancer immunotherapy may benefit from enhanced immunogenicity when plant glycans bind to lectins on the surface of dendritic cells and thereby stimulate the activity of antigen-presenting cells. Furthermore, some proteins carrying plant-derived glycans show superior activity to their native counterparts. One example is the production of plant allergens as diagnostic reagents. These require the proper presentation of plant glycans to enable the detection of IgE antibodies that recognize cross-reactive plant carbohydrate determinants. Another example is recombinant glucocerebrosidase (see Note 8). This enzyme is produced in carrot cells and is targeted to the vacuole because the vacuole-specific mannose-tipped glycan chains improve the uptake of the protein by human macrophages [14]. These unique strengths make plants particularly suitable for the production of certain classes of proteins (Fig. 2). For example, several companies offer animal-free plant-derived diagnostic and pharmaceutical proteins, the first of which was avidin produced in maize and commercialized more than 20 years ago (see Note 9). The plant matrix may also enhance the effectiveness of vaccines and

A Brief Overview of Strengths and Challenges

5

Animal-free proteins

Animal pharmaceuticals

Cosmetic ingredients

Plant-specific glycoproteins

Emergency products

Animal-free diagnostics for cell culture and pharmaceuticals for vegans, allergy sufferers or religious communities

No purification required for animal vaccines and therapeutics providing cheaper products

‘Green’ production of cosmetic proteins, which may not been purified when using food plants

Plant-specific glycans improving binding and functionality of diagnostic and therapeutic proteins

Diagnostic and pharmaceutical proteins being available quickly, e.g. in pandemics

Fig. 2 Protein products that benefit from the advantages of plants as a production platform

therapeutics by encapsulating them in intact cells and cell compartments such as chloroplasts, protecting them from digestion and thus prolonging their activity (see Note 10). Other protein products that can benefit from production in plants are cosmetic ingredients, glycoproteins and emergency proteins, as described above.

2

Which Plants Are Used for Protein Production? Recombinant protein production has been reported in a broad range of plant species, including whole plants (e.g., alfalfa, barley, maize, pea, potato, rice, safflower, soybean, strawberry, tobacco, tomato, and wheat), hairy roots (mainly tobacco), floating or suspended plant tissue (e.g., moss and duckweed), and cell suspension cultures (e.g., alfalfa, Arabidopsis, carrot, rice, soybean, tobacco, and tomato) [6]. This wide range of production platforms offers immense flexibility with respect to cultivation formats, expression levels, downstream processing (see Note 11) and application scenarios. However, a smaller number of plant species and systems have been prioritized because they excel in terms of product yields, handling and speed. This also allows the concentration of efforts to improve productivity even further (see Subheading 3). Tobacco (Nicotiana tabacum) and its smaller relative N. benthamiana are the workhorses of molecular farming because genetic transformation and cultivation are straightforward, and both species produce large amounts of biomass per unit area and thus high yields of recombinant protein. Both species are suitable for stable transformation [15] and transient expression [9–11]. Cereals such as barley, maize, rice, and wheat have also been established as powerful production systems [16, 17]. Dry cereal grains protect

6

Stefan Schillberg and Holger Spiegel

recombinant proteins against proteolytic degradation allowing storage for months or even years before extraction without any special cooling requirements. However, the genetic transformation of cereals is more challenging compared to tobacco and the regeneration of whole plants usually takes longer, although reliable protocols based on Agrobacterium and particle bombardment can be used to generate transgenic cereals for recombinant protein production [16]. Tobacco is not only used for protein production in whole plants, it is also the most popular source of tissue explants such as hairy roots and cell suspension cultures for recombinant protein production. N. tabacum cv. Bright Yellow 2 (BY-2) is the cultivar of choice for suspension cell cultures because these cells proliferate rapidly and are easy to transform [18] (see Note 12). In contrast to intact plants, plant cells are grown in vitro under defined conditions in shake flasks or larger-scale bioreactors, and therefore resemble the cultivation of microbes and mammalian cells (and can likewise retain recombinant protein inside the cell or secrete it to the medium). BY-2 cells are derived from roots and accordingly do not need light to grow, allowing the use of inexpensive sugar-based media that promote growth at very high cell densities [19]. One of the most recent plant molecular farming platforms is cell-free in vitro transcription and translation. Cell-free systems based on wheat germ extracts have been used for a long time to screen protein expression in microtiter plates, but larger-scale production is hampered by the expensive and time-consuming lysate preparation method. Recently, a cell-free protein biosynthesis system based on tobacco BY-2 cells has been established, which is more scalable and achieves yields of up to 3 mg/mL, significantly higher than any other eukaryotic expression system [7] (see Note 13). The BY-2 cell-free lysates can be prepared easily at low cost, and reaction volumes of 10 mL are currently possible. Ongoing work aims to reach the multi-liter range, which would allow several grams of protein to be produced in each reaction. Cell-free systems also allow the biosynthesis of proteins that are cytotoxic or unstable in living cells [20]. Another major advantage is that protein biosynthesis can be decoupled from lysate preparation. Lysates can be prepared in advance and stored for a long time at
80  C, and recombinant proteins can then be produced on demand within 24–48 h by thawing the lysates. This allows the just-in-time production of proteins even in emergency situations.

3

What Challenges Remain to Be Addressed? Like other recombinant protein production platforms, the general aim of plant-based systems is to increase productivity and reduce manufacturing costs for industrial applications. In this context,

A Brief Overview of Strengths and Challenges

7

several challenges must be overcome to further improve recombinant protein production (Fig. 1) in plants and various strategies have been devised to achieve this. Compared to industrial protein production in microbial and mammalian cells, plants still fall behind in terms of productivity for most proteins. Protein stability (and consequently accumulation levels) in plants can be increased by targeting recombinant proteins to a suitable cell compartment. Recombinant proteins generally achieve the best folding and highest stability if they are directed to the same compartments used for native production—for example, antibodies are secreted to the endomembrane system of mammalian cells and similarly accumulate to the highest levels when secreted to the apoplast or retained in the endoplasmic reticulum of plant cells. However, the ideal plant cell compartment for many proteins is unknown, and different targeting strategies must be tested empirically. Here, modular cloning strategies based on Golden Gate technology (such as MoClo or GoldenBraid) facilitate the preparation of expression vectors with different targeting signals [21]. Alternatively, a recombinant protein or peptide can be fused to a carrier protein, which confers better stability and/or improves recovery during downstream processing and purification [11]. One drawback of this approach is that the fusion partner must be cleaved off precisely and removed if the recombinant protein is required in an intact form, as is the case for pharmaceutical proteins. Although larger-scale production is not yet possible in the cell-free expression system based on BY-2 cells, higher productivity may be possible because the vacuole (containing most of the proteases) is removed during lysate preparation, and a protease inhibitor cocktail is added to the reaction mix [20] (see Note 14). Improving parameters in the agroinfiltration process can increase yields during transient expression, for example by optimizing the bacterial cultivation medium and infiltration buffer, as well as vacuum conditions and subsequent incubation. Bioluminescent bacteria can be used to monitor the effect such optimization steps, allowing the easy and noninvasive detection and quantification of living bacteria by bioluminescence imaging, which correlates with recombinant protein expression [22]. The quality of plant-derived recombinant proteins must also be optimized. N-linked glycosylation is similar in plants and the mammalian cells typically used for the industrial production of recombinant glycoproteins, but (as stated above) plant glycoproteins contain core β(1,2)-xylose and α(1,3)-fucose residues that are not present in mammals and may affect the folding and stability of plant-derived glycoproteins and their interactions with receptors. This has been addressed by the modification of host plants by genome editing to remove the fucosyltransferase and xylosyltransferase genes responsible for the synthesis of plant-specific glycans [23].

8

Stefan Schillberg and Holger Spiegel

Another challenge is the recovery of recombinant proteins from plant tissues and the removal of large quantities of plantderived insoluble fibers and cell wall fragments as well as endogenous proteins. For example, larger blade-based homogenization devices can be used to extract proteins batch-wise or continuously from plant tissue, and serial filtration steps have established to remove insoluble compounds from plant extracts [24]. In addition, unwanted particles can be removed by flocculation, a process in which fine particulates clump together into a floc and can be removed by filtration. Alternatively, endogenous plant proteins can be removed prior to chromatography by heat precipitation [25]. However, this is only feasible if the target protein is thermostable, and the temperature must be selected carefully to denature and precipitate most of the host cell proteins while ensuring the recombinant protein stays soluble. Nevertheless, such procedures can significantly improve recombinant protein recovery and reduce manufacturing costs. Given that each step in the extraction and purification of recombinant proteins can affect other steps, optimization can involve changes to multiple process settings and the identification and accommodation of complex parameter interactions. Such interactions are not revealed by conventional one-factor-at-a-time approaches involving the testing of process factors individually and sequentially. Instead multiple factors must be tested simultaneously using statistical designs, reducing the number of experiments and therefore the time and costs. These “design of experiments” approaches help to identify and optimize important process parameters [26]. In the economic analysis of plant molecular farming processes, this is particularly relevant because downstream processing accounts for most of the manufacturing costs [7]. The techno-economic evaluation of molecular farming processes is therefore necessary to identify technical process and economic product challenges, and several analytical tools have been developed to determine the competiveness of recombinant protein production in plants [27]. Compared to well-established microbial and mammalian cellbased systems, the regulation of molecular farming processes and products is rather new, and involves added complexities based on the cultivation of plant systems in the open field or in containment, the product and its intended application, as well as the regulatory authorities in different regions of the world. As more products enter the market, the regulations are becoming clearer and individual examples can be described in more detail to support future processes and products [28]. Freedom-to-operate (FTO) and intellectual property (IP) issues must also be considered when using molecular farming processes in a commercial setting [29]. Suitable IP-protected processes providing FTO have been established by several molecular farming companies. However, molecular farming

A Brief Overview of Strengths and Challenges

9

platforms (expression vectors and host plants, along with corresponding protocols) that can be purchased for commercial use are not yet available, in contrast to the well-proven high-performance platforms based on microbial and mammalian cells. The lack of access to plant systems that can be used commercially by any company interested in this technology may the biggest hurdle for the wider dissemination of molecular farming.

4

Notes 1. The terms biofarming and gene farming are occasionally used as alternatives to molecular farming. The variant spelling pharming can be used for biopharmaceutical products, which are described as plant-made pharmaceuticals (PMPs). 2. Cultivation in the open field involves the risk that the plant material is contaminated with pesticides, fertilizers, bird droppings, dead insects or other animals, and other forms of environmental pollution. Ensuring that these contaminants are removed during downstream processing requires additional purification steps and analytical verification, adding to the overall costs. 3. In transient expression, proteins accumulate in the intact plants or infiltrated leaves within 3–5 days. The timing of maximum protein accumulation must be tested individually for each protein and mainly depends on the number of cells transfected by the bacteria (and the number of surrounding cells to which the vector spreads, if it is a deconstructed viral vector) and the stability of the target protein. 4. A video showing the infiltration of plant leaves can be found at the following URL: https://www.youtube.com/watch? v¼wDt7s9euS4A 5. Yeast cells can also be used for the administration of unpurified recombinant proteins because they do not produce endotoxins and are not a source of mammalian pathogens. 6. Transient expression by infiltration requires the removal of endotoxins if the product is intended for administration to humans. These endotoxins are produced by the Agrobacterium cells filling the spaces between plant cells. 7. Animal-derived proteins can also be avoided in transient expression procedures by replacing animal-derived proteins in the Agrobacterium cultivation medium with plant-derived nutrient components [30]. However, endotoxins will be still present in the infiltrated plant tissue (see Note 6). 8. The enzyme glucocerebrosidase is used to treat Gaucher’s disease, a lysosomal storage disorder. A recombinant form of this enzyme produced in carrot cells is marketed as taliglucerase

10

Stefan Schillberg and Holger Spiegel

alfa (Elelyso) and production in plant cells results in the synthesis of plant-specific glycans. Interestingly, a low incidence of induced anti-plant glycan antibodies was found in Gaucher’s disease patients after up to 30 months of enzyme replacement therapy [14]. Detailed evaluation of clinical safety and efficacy endpoints indicated that anti-plant glycan antibodies did not affect the safety or efficacy of taliglucerase alfa in patients. This study demonstrated that the risk of plant-specific glycans might be overestimated and has to be determined on a proteinspecific manner. 9. Avidin produced in maize is still sold by Sigma-Aldrich (now Merck) under the catalog no. A8706). 10. One example for such an approach is a poultry vaccine produced in tobacco cells. The vaccine is the recombinant hemagglutinin-neuraminidase (HN) glycoprotein, one of two viral surface glycoproteins and the major surface antigen that induces neutralizing antibodies and a protective immune response in domestic poultry and other avian species. The HN protein was extracted from the tobacco cells and the crude extract was injected into young chickens to protect them against the Newcastle disease virus [19]. The animal health vaccine was developed by Dow AgoSciences LLC and approved by the United States Department of Agriculture (USDA) in 2006. Even though the unpurified HN protein conferred full protection in virus challenge studies, the poultry vaccine was not commercialized by Dow AgroSciences following a strategic business decision. 11. Recombinant proteins accumulating in whole plants can be extracted from leaf tissue or secreted via the roots (rhizosecretion), whereas recombinant proteins produced in hairy roots, cell suspension cultures and tissue explants are usually secreted and recovered from the culture medium, but can be also extracted from the root/cells/tissue. The choice between secretion and accumulation within the cells/tissue may have a major impact on the final product yields. 12. Cell suspension cultures derived from the tobacco cultivar N. tabacum cv. NT-1 show similar growth characteristics and transformation efficiencies as BY-2 cells and are also often used for recombinant protein production [31]. 13. In addition to the widely used Escherichia coli lysates, several eukaryotic cell-free systems have been described, including rabbit reticulocyte lysate, yeast cell extract, Chinese hamster ovary cell lysate, HeLa cell lysate, and wheat germ extract. 14. Nucleases present in the vacuole are also removed during lysate preparation, which limits the degradation of the expression vector added to the reaction mix to initiate synthesis of the target protein.

A Brief Overview of Strengths and Challenges

11

Acknowledgments We thank Dr. Richard M Twyman for editorial assistance. The authors would like to thank the members of the Newcotiana (760331) and Pharma-Factory (774078) consortia, both funded by the EU, for stimulating discussions on the potential and challenges of plants for the production of recombinant proteins. References 1. Herrera-Estrella L, Depicker A, Van Montagu M, Schell J (1983) Expression of chimaeric genes transferred into plant cells using a Ti-plasmid-derived vector. Nature 303:209–213 2. De Block M, Herrera-Estrella L, van MCE M, Schell J, Zambryski PC (1984) Expression of foreign genes in regenerated plants and in their progeny. EMBO J 3:241681–241689 3. Stieger M (1987) Versuche zur Integration und Expression chim€arer Immunoglobuline in Pflanzen. Dissertation University of Cologne 4. Hiatt A, Cafferkey R, Bowdish K (1989) Production of antibodies in transgenic plants. Nature 342:76–78. https://doi.org/10. 1038/342076a0 5. Sijmons PC, Dekker BM, Schrammeijer B, Verwoerd TC, van den Elzen PJ, Hoekema A (1990) Production of correctly processed human serum albumin in transgenic plants. Biotechnology 8:217–221. https://doi.org/ 10.1038/nbt0390-217 6. Spiegel H, Sto¨ger E, Twyman RM, Buyel JF (2018) Current status and perspectives of the molecular farming landscape. In: Kermode AR, Jiang L (eds) Molecular farming: applications, challenges, and emerging areas. John Wiley & Sons, Inc, Hoboken, New Jersey. https://doi. org/10.1002/9781118801512 7. Schillberg S, Raven N, Spiegel S, Rasche S, Buntru M (2019) Critical analysis of the commercial potential of plants for the production of recombinant proteins. Front Plant Sci 10: 720. https://doi.org/10.3389/fpls.2019. 00720 8. Schillberg S, Finnern R (2021) Plant molecular farming for the production of valuable proteins – critical evaluation of achievements and future challenges. J Plant Physiol 258-259: 153359. https://doi.org/10.1016/j.jplph. 2020.153359 9. Spiegel H, Schillberg S, No¨lke G (2022) Production of recombinant proteins by agrobacterium-mediated transient expression. In: Schillberg S, Spiegel H (eds)

Recombinant protein production in plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480. Springer, New York. https://doi.org/10.1007/978-1-0716-22414_6 10. Peyret H, Lomonossoff GP (2022) Specific packaging of custom RNA molecules into cowpea mosaic virus-like particles. In: Schillberg S, Spiegel H (eds) Recombinant protein production in plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480. Springer, New York. https://doi.org/10.1007/978-10716-2241-4_7 11. Dickmeis C, Commandeur U (2022) Advanced fusion strategies for the production of functionalized potato virus X virions. In: Schillberg S, Spiegel H (eds) Recombinant protein production in plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480. Springer, New York. https://doi.org/ 10.1007/978-1-0716-2241-4_13 12. Spiegel H, Boes A, Voepel N, Beiss V, Edgue G, Rademacher R, Sack M, Schillberg S, Reimann A, Fischer R (2015) Application of a scalable plant transient gene expression platform for malaria vaccine development. Front Plant Sci 6:1169. https://doi. org/10.3389/fpls.2015.01169 13. Capell T, Twyman RM, Armario-Najera V, Ma KCM, Schillberg S, Christou P (2020) Potential applications of plant biotechnology against SARS-CoV-2. Trends Plant Sci 25:635–643. https://doi.org/10.1016/j.tplants.2020. 04.009 14. Rup B, Alon S, Amit-Cohen BC, Brill Almon E, Chertkoff R, Tekoah Y (2017) Immunogenicity of glycans on biotherapeutic drugs produced in plant expression systems -the taliglucerase alfa story. PLoS One 12: e0186211. https://doi.org/10.1371/journal. pone.0186211 15. Szeto TH, Drake PMW, Teh AYH, Falci Finardi N, Clegg AG, Paul MJ, Reljic R, Ma JKC (2022) Production of recombinant proteins in transgenic tobacco plants. In:

12

Stefan Schillberg and Holger Spiegel

Schillberg S, Spiegel H (eds) Recombinant protein production in plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480. Springer, New York. https://doi.org/ 10.1007/978-1-0716-2241-4_2 16. Kapusi E, Stoger E (2022) Molecular farming in seed crops: gene transfer into barley (Hordeum vulgare) and wheat (Triticum aestivum). In: Schillberg S, Spiegel H (eds) Recombinant protein production in plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480. Springer, New York. https://doi.org/10.1007/978-1-0716-22414_3 17. Arcalı´s E, Pedrazzini E, Ho¨rmann-Dietrich U, Vitale A, Stoger E (2022) Cell biology methods to study recombinant proteins in seeds. In: Schillberg S, Spiegel H (eds) Recombinant protein production in plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480. Springer, New York. https://doi.org/ 10.1007/978-1-0716-2241-4_4 18. Navarre C, Chaumont F (2022) Production of recombinant glycoproteins in Nicotiana tabacum BY-2 suspension cells. In: Schillberg S, Spiegel H (eds) Recombinant protein production in plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480. Springer, New York. https://doi.org/10.1007/978-10716-2241-4_5 19. Schillberg S, Raven N, Fischer R, Twyman RM, Schiermeyer A (2013) Molecular farming of pharmaceutical proteins using plant suspension cell and tissue cultures. Curr Pharm Des 19: 5531–5542. https://doi.org/10.2174/ 1381612811319310008 20. Buntru M, Vogel S, Finnern R, Schillberg S (2022) Plant-based cell-free transcription and translation of recombinant proteins. In: Schillberg S, Spiegel H (eds) Recombinant protein production in plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480. Springer, New York. https://doi.org/ 10.1007/978-1-0716-2241-4_8 21. Gonza´lez B, Vazquez-Vilar M, Sa´nchezVicente J, Orza´ez D (2022) Optimization of vectors and targeting strategies including GoldenBraid and genome editing tools: GoldenBraid assembly of multiplex CRISPR/ Cas12a guide RNAs for gene editing in Nicotiana benthamiana. In: Schillberg S, Spiegel H (eds) Recombinant protein production in plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480. Springer, New York. https://doi.org/10.1007/978-1-07162241-4_12

22. Jutras PV, Dodds I, van der Hoorn RAL (2022) A bioluminescent Agrobacterium tumefaciens for imaging bacterial metabolic activity in planta. In: Schillberg S, Spiegel H (eds) Recombinant protein production in plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480. Springer, New York. https://doi.org/10.1007/978-1-07162241-4_15 23. Jansing J, Bortesi L (2022) Knockout of glycosyltransferases in Nicotiana benthamiana by genome editing to improve glycosylation of plant-produced proteins. In: Schillberg S, Spiegel H (eds) Recombinant protein production in plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480. Springer, New York. https://doi.org/10.1007/978-1-07162241-4_14 24. Buyel JF (2022) Strategies for efficient and sustainable protein extraction and purification from plant tissues. In: Schillberg S, Spiegel H (eds) Recombinant protein production in plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480. Springer, New York. https://doi.org/10.1007/978-1-07162241-4_9 25. Spiegel H (2022) Improving recombinant protein recovery from plant tissue using heat precipitation. In: Schillberg S, Spiegel H (eds) Recombinant protein production in plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480. Springer, New York. https://doi.org/10.1007/978-1-0716-22414_10 26. Buyel JF (2022) Statistical designs to improve downstream processing. In: Schillberg S, Spiegel H (eds) Recombinant protein production in plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480. Springer, New York. https://doi.org/10.1007/978-1-07162241-4_16 27. McNulty MJ, Nandi S, McDonald KA (2022) Technoeconomic modeling and simulation for plant-based manufacturing of recombinant proteins. In: Schillberg S, Spiegel H (eds) Recombinant protein production in plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480. Springer, New York. https://doi.org/10.1007/978-1-0716-22414_11 28. Hundleby PAC, D’Aoust MA, Finkle C, Atkins J, Twyman RM (2022) Regulation of molecular farming products. In: Schillberg S, Spiegel H (eds) Recombinant protein production in plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480. Springer,

A Brief Overview of Strengths and Challenges New York. https://doi.org/10.1007/978-10716-2241-4_17 29. Thangaraj H (2022) Freedom to operate analysis of molecular farming projects. In: Schillberg S, Spiegel H (eds) Recombinant protein production in plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480. Springer, New York. https://doi.org/ 10.1007/978-1-0716-2241-4_18 30. Houdelet M, Galinski A, Holland T, Wenzel K, Schillberg S, Buyel J (2017) Animal component-free agrobacterium tumefaciens cultivation media for better GMP-compliance

13

increases biomass yield and pharmaceutical protein expression in Nicotiana benthamiana. Biotechnol J 12:1600721. https://doi.org/ 10.1002/biot.201600721 31. Ullisch D, Mu¨ller CA, Maibaum S, Kirchhoff J, Schiermeyer A, Schillberg S, Roberts JL, Treffenfeldt W, Bu¨chs J (2012) Comprehensive characterization of two different Nicotiana tabacum cell lines leads to doubled GFP and HA protein production by media optimization. J Biosci Bioeng 113:242–248. https://doi. org/10.1016/j.jbiosc.2011.09.022

Part I Plant Production Systems and Applications

Chapter 2 Production of Recombinant Proteins in Transgenic Tobacco Plants Tim H. Szeto, Pascal M. W. Drake, Audrey Y-H. Teh, Nicole Falci Finardi, Ashleigh G. Clegg, Mathew J. Paul, Rajko Reljic, and Julian K-C. Ma Abstract Nicotiana tabacum (the tobacco plant) has numerous advantages for molecular farming, including rapid growth, large biomass and the possibility of both cross- and self-fertilization. In addition, genetic transformation and tissue culture protocols for regeneration of transgenic plants are well-established. Here, we describe the production of transgenic tobacco using Agrobacterium tumefaciens and the analysis of recombinant proteins, either in crude plant extracts or after purification, by enzyme-linked immunosorbent assays, sodium dodecyl sulfate polyacrylamide gel electrophoresis with western blotting and surface plasmon resonance. Key words Agrobacterium tumefaciens, ELISA, Nicotiana tabacum, Recombinant protein, SDS PAGE, Surface plasmon resonance, Transgenic, Western blotting

1

Introduction Recently, the production of recombinant proteins using transient expression in Nicotiana benthamiana has become prevalent within the plant-based systems, exemplified by the production of the ZMapp™ anti-Ebola monoclonal antibody (mAb) cocktail [1] and the establishment of small and medium-sized enterprises which employ the technology, such as Medicago Inc. (Quebec, Canada), Kentucky BioProcessing (Owensboro, Kentucky, USA) and iBio (Bryan, Texas, USA). Nevertheless, expression in transgenic plants remains relevant, indeed the two plant-derived recombinant pharmaceuticals currently marketed, Elelyso™ and InterBerry α®, are produced using the transgenic method, in carrot

Tim H. Szeto, Pascal M.W. Drake, Audrey Y-H. Teh and Nicole Falci Finardi contributed equally to this work. Stefan Schillberg and Holger Spiegel (eds.), Recombinant Proteins in Plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480, https://doi.org/10.1007/978-1-0716-2241-4_2, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

17

18

Tim H. Szeto et al.

cells and strawberries respectively [2, 3]. Transgenic rice is also used by Ventria Bioscience (Fort Collins, Colorado, USA) for the generation of multiple recombinant protein products. Transgenic plants offer several potential advantages including a multiplicity of production platforms, the possibility of expression on a vast agricultural scale for proteins required in massive quantities, and in some cases an established regulatory framework [4, 5]. Here, we describe the production of recombinant proteins in transgenic tobacco (Nicotiana tabacum) and we will focus on pharmaceuticals, a wide variety of which have been expressed in this species [4, 5]. Many of the milestones in plant biotechnology have been achieved in this species; Nicotiana tabacum was the first transgenic plant produced [6] and the first examples of recombinant pharmaceutical expression in plants were also in this species [7, 8]. Tobacco has numerous advantages for molecular farming including rapid growth, large biomass, well-established genetic transformation and tissue culture protocols, and the possibility for self- and cross-fertilization [5, 9]. Numerous Nicotiana cultivars have been generated including low nicotine varieties [10]. In the European Union, GMP protocols and regulatory frameworks have been established in this species ultimately leading to a Phase I clinical trial of an anti-HIV antibody [11]. A variety of techniques are used in the production and analysis of recombinant proteins in tobacco, and we will focus on protocols used routinely in our research group. Agrobacterium tumefaciens-mediated transformation is the method of choice for producing transgenic tobacco plants. Recombinant pharmaceuticals can also be produced in tobacco transgenic “hairy roots” generated using Agrobacterium rhizogenes [12] but this will not be discussed in the present chapter. Agrobacterium tumefaciens causes crown gall disease in nature by the transmission of a DNA sequence (T-DNA) into the nucleus of the plant cells. So-called binary vector systems are employed by researchers to generate transgenic plants: here, the gene(s) of interest (GOI) is contained in the binary vector and is transferred by the action of virulence genes on a helper plasmid into the plant nucleus [13]. Typically, binary vectors are small plasmids (~5–10 kb) which are amenable to molecular cloning, which have a relaxed origin of replication for maintenance in E. coli and Agrobacterium, and which contain a bacterial selectable marker. The GOI, bracketed by suitable promoter and terminator sequences [5], is inserted by molecular cloning between left and right border Agrobacterium DNA sequences from which T-DNA transfer is initiated. In addition to the GOI, the T-DNA also contains a plant selectable marker: the neomycin phosphotransferase II gene (npt-II) conferring resistance to aminoglycoside antibiotics is frequently used in tobacco for the selection of transformed cells and transgenic plants [14]. Once molecular cloning of the GOI into the binary vector has

Recombinant Protein Production in Tobacco

19

been completed, it is isolated from E. coli and transformed into A. tumefaciens usually using either freeze-thaw or electroporation protocols [15]. Transgenic tobacco plants are generally produced using the leaf disc method [16]. Leaf discs are incubated in A. tumefaciens containing the GOI which is transferred into the plant nucleus. Transgenic plants are regenerated from the leaf disc using a tissue culture protocol consisting of culture on growth medium containing plant growth regulators to generate shoots, antibiotic to remove agrobacteria, and selective agent (usually the antibiotic kanamycin) to select transformed shoots. Shoots are removed from the leaf disc and rooted in vitro prior to transfer of the plantlet to soil. Transgenic plants generated using A. tumefaciens often contain multiple insertions of the GOI [17]. Plants with a single insertion may be desired, and these can be identified using a variety of molecular techniques [18]. Following fertilization, foreign genes are transferred to progeny by Mendelian inheritance. Putatively transgenic tobacco plants can be analyzed using a number of techniques to confirm appropriate expression of recombinant protein. We routinely employ enzyme-linked immunosorbent assays (ELISAs) and sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS PAGE) with western blotting for this undertaking, and these techniques can be used with either crude leaf extracts or purified material [19]. It is advisable to use both ELISA and SDS PAGE/western blotting when analyzing transgenic plants, as the techniques will provide different but complementary information about the recombinant protein being expressed. In addition, both can be used to measure recombinant protein concentration which is usually expressed as mg recombinant protein/kg plant tissue fresh weight. ELISA is highly sensitive, simple, relatively rapid, and exists in different formats—direct, indirect, or sandwich (also called capture) (Fig. 1). However, all ELISAs follow the same basic principle of immobilizing a protein on a solid surface that can interact with another protein in solution (e.g., an antigen and an antibody). Detection is with an enzyme substrate that can catalyze a chromogenic or chemiluminescent reaction (e.g., horseradish peroxidase). This reaction is compared to that generated with a standard of known concentration. To prevent an over-saturated signal, the reaction can be stopped by the addition of sodium hydroxide, sulfuric acid, or hydrochloric acid [20]. Reaction absorbance readings are recorded by spectrophotometry at 400–600 nm depending on the conjugate used. The different types of ELISA have advantages and disadvantages [20]. For example, indirect ELISA may give a stronger signal than direct if the secondary antibody is polyclonal, as several antibodies (conjugated to enzyme) may bind to different epitopes on the interacting antibody. Sandwich ELISAs are particularly suitable for detecting low concentrations of antigen,

20

Tim H. Szeto et al.

Fig. 1 ELISA formats. Direct ELISA involves coating one protein (e.g., an antigen) on the ELISA plate, then detecting it with another enzyme-conjugated protein (e.g., an antibody) leading to a colorimetric or chemiluminescence reaction when enzyme substrate is added. In indirect ELISA, the interacting antibody is not conjugated to the enzyme. Instead, a secondary antibody which recognizes the interacting antibody is conjugated with the enzyme. In a sandwich ELISA format, the antigen is first captured by a pre-coated antibody. Then, the interacting antibody and/or secondary antibody is introduced

as the coating antibody will concentrate any antigen present. Recombinant proteins produced by transgenic tobacco have been characterized using direct [21], indirect [22], and sandwich [23] ELISA methodologies. Here, we show the subtle differences in methodology between ELISA formats, how to interpret the results and calculate the recombinant protein yield. SDS PAGE and western blotting are frequently employed to analyze putatively transgenic tobacco plants. Tissue samples are macerated in suitable buffer, with or without a reducing agent such as β-mercaptoethanol, and supernatant containing recombinant protein is boiled in SDS, which results in the latter binding to the protein forming a negatively charged complex, with the magnitude of the charge positively correlated to protein size [24]. When loaded in a PAGE gel and subjected to electrophoresis, proteins migrate toward the anode with speed of movement inversely proportional to size. Protein standards of known molecular weight are run at the same time and comparison with these allows assessment of recombinant protein size. Proteins in a sample can be visualized by Coomassie blue or silver staining of the gel [24]. Recombinant proteins can be definitively identified by subjecting the gel to western blotting using an appropriate antibody conjugated to an enzyme for detection by color development or chemiluminescence. In addition to visualization of the protein and confirmation of size, western blotting will often give an indication of protein

Recombinant Protein Production in Tobacco

21

degradation, as multiple bands corresponding to breakdown products may be seen on the blot. It is often instructive to analyze recombinant proteins using both reducing and non-reducing western blotting; for example, when analyzing mAb production, the former will show individual bands corresponding to light and heavy chains and the latter will demonstrate the formation of assembled antibody (for an example see [25]). Transgenic tobacco plants can be self- or cross-fertilized by dusting donor pollen onto the stigma of the recipient flower to generate seed which can subsequently be germinated and analyzed for GOI expression. Typically, plants expressing the highest levels of recombinant protein in the T0 generation will be chosen for generation of progeny. If required, plants homozygous for the GOI can be identified by backcrossing with wild-type plants. In some cases, for production of complex proteins such as secretory antibody or the B-cell receptor, different components of the molecule have been produced in separate transgenic plant lines and brought together to produce the final assembled protein by sexual crossing and screening of progeny [19, 25]. Once GOI expression has been confirmed, recombinant protein can be extracted and purified from transgenic plants. This downstream processing accounts for a large proportion of the costs of production in the plant platform and much work has gone into the improvement of protein extraction methods in recent years [26]. Extraction and purification procedures vary with nature of the plant tissue being processed and the characteristics of the recombinant protein in question, but protocols may include maceration, filtration, and centrifugation to remove plant tissue, precipitation of contaminating proteins, concentration of recombinant protein, purification by affinity chromatography and further polishing steps [26–28]. Protein G and A, isolated from Streptococcus and Staphylococcus species, have been used extensively for the purification of monoclonal antibodies (mAbs) [29]. Here, the Protein A chromatography-based method which is employed for purification of human IgG and Fc-fusion proteins from transgenic tobacco will be described in detail. Recombinant proteins are also frequently purified by co-expression with affinity tags, for example use of the His-tag allows purification of recombinant protein using nickel column affinity chromatography [30]. Purified recombinant proteins can be analyzed using diverse structural and functional assays. In this chapter, we describe the technique of surface plasmon resonance (SPR). SPR is a real-time, label-free method for characterizing and quantifying the binding kinetics of protein-protein interactions. A method for surface plasmon excitation was first developed in 1968 [31] and since then, research into excitation methods and sensor chips has made SPR extremely accurate and sensitive [32].

22

Tim H. Szeto et al.

Fig. 2 Different stages of ligand-analyte interaction during SPR on a sensorgram (left) and a representation of the interaction between the ligand (Y-shaped molecule) immobilized on the dextran matrix (black line) covalently attached to the gold chip surface (yellow line) and the injected analyte (circle) at each stage of the cycle (right)

In SPR, one interacting protein (the ligand) is fixed on the surface of a gold chip, the other interacting protein (the analyte) is introduced to the protein bound to the chip surface. There are several stages of interaction when this happens in a direct assay (Fig. 2): (1) Baseline, when the ligand is bound to the chip. One of the methods for doing this is amine coupling, which involves activation of carboxymethyl groups on a carboxymethyldextrancoated gold chip by reaction with N-hydroxysuccinimide, followed by the covalent bonding of the protein ligands to the surface of the chip through amide bonds formed between the activated surface and primary amines of lysin residues within the protein amino acid sequence, and blocking the excess activated carboxyl with ethanolamine [33]. (2) Association phase (Ka): injected analyte interacts with the bound ligand. (3) Dissociation phase (Kd): the separation of ligand and analyte over time. (4) A regeneration stage may be required to regenerate the chip surface. For example, a low pH glycine buffer is used to regenerate a Protein A coupled chip by removing any antibody ligand bound to the Protein A [34]. In capture assays, a molecule that interacts with the ligand is first immobilized on the surface of the chip which will hold the ligand on the chip surface before the analyte is injected. The overall changes in response during association and dissociation are depicted in a sensorgram (Fig. 2), which need to be fitted with an appropriate kinetic binding model (typically a 1:1

Recombinant Protein Production in Tobacco

23

interaction model) before Ka, Kd and KD can be interpreted. KD is the ratio between dissociation (Kd) and association (Ka)—as such, two analytes may have the same KD value but exhibit different rates of association and dissociation. The final result (visualized as the sensorgram) is affected by different parameters: binding kinetics, flow rate, temperature, pH, salt concentrations, interaction stoichiometry, and analyte concentration. Here, we describe the use of a direct assay to determine mAb concentration and a capture assay to determine the interaction between a mAb and human FcγRI receptor [35].

2

Materials

2.1 Freeze-Thaw Method for Agrobacterium tumefaciens Transformation

The freeze–thaw method for transformation of Agrobacterium tumefaciens has been adapted from [36]. 1. Liquid nitrogen. 2. 10 mM CaCl2. 3. Luria-Bertani broth (LB): 10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl. 4. Semi solidified LB medium: LB solidified with 15 g/L agar (see Note 1). 5. Appropriate antibiotics for selection of agrobacteria.

2.2 AgrobacteriumMediated Transformation and In Vitro Regeneration of Plantlets

1. LB broth: 10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl. 2. Semisolidified LB medium: LB solidified with 15 g/L agar containing appropriate antibiotics for selection of agrobacteria. 3. Commercial bleach e.g., Domestos™. 4. Sterile distilled water. 5. Shoot Regeneration Medium (Murashige and Skoog medium (MS) [37]: 30 g/L sucrose, 1 mg/L 6-Benzlyaminopurine, 0.1 mg/L 1-Napthaleneacetic acid, made semi-solid with 8 g/ L agar. 6. Rooting Medium: MS medium containing 30 g/L sucrose, made semi-solid with 8 g/L agar. 7. Carbenicillin. 8. Kanamycin.

2.3 Transfer of In Vitro-Grown Plantlets to Soil, CrossFertilization, Seed Collection

1. All-purpose compost. 2. Soluble plant nutrient.

24

2.4

Tim H. Szeto et al.

ELISA

1. Mixer Mill MM400: Bench top homogenizer for dry, wet and cryogenic grinding of small amounts of sample. Requires Eppendorf tubes to contain the sample and buffer as well as two 3.165 mm steel ball bearings to aid in homogenization. 2. HIV-1 gp120 antigen (HIV Reagent Programme): Recombinant HIV-1 BaL gp120 spike protein, produced in mammalian HEK293T cells. 3. Phosphate buffered saline (PBS): To prepare 10x PBS stock, add 80 g of NaCl, 2 g of KCL, 14.4 g of Na2HPO4, and 2.4 g KH2PO4 to 800 mL of deionized water. Adjust pH to 7.4 and add deionized water to 1 L. Sterilized by autoclaving to store. To make 1x PBS stock, dilute 10
PBS 1 in 10 with deionized water. 4. Wash buffer: 1 mL of Tween™ 20 into 1 L PBS (0.1% v/v solution). Due to its high viscosity it is preferable to use a syringe instead of a pipette when handling Tween20. 5. Deionized water: All deionized water should be purified to the sensitivity of at least 5 MΩ-cm at 25  C. 6. Blocking buffer: Make 5% w/v non-fat dried milk powder (NFDM) or 2.5% w/v Bovine Serum Albumin (BSA) solution in Wash buffer. 7. 3,30 ,5,50 -Tetramethylbenzidine (TMB) can come as a liquid substrate system or in tablet form. Prepare TMB substrate according to the manufacturer’s instructions. 8. 2 M H2SO4: Verify the molarity of the sulfuric acid. Dissolve an appropriate amount of acid in deionized water to make 2 M sulfuric acid (see Note 2).

2.5 SDS Polyacrylamide Gel Electrophoresis

1. XCell SureLock Mini-Cell Electrophoresis System. 2. NuPAGE® Bis-Tris precast gels (Invitrogen). 3. NuPAGE® MOPS or MES SDS Running Buffer (20
) (Invitrogen). 4. NuPAGE® 4
LDS Sample Buffer (Invitrogen). 5. Precision Plus Protein™ All Blue Prestained Protein Standards (Bio-Rad). 6. Coomassie stain. 7. β-Mercaptoethanol. 8. Heating block.

2.6 Semi-Dry Enhanced Chemiluminescence (ECL) Western Blot

1. Semi-dry transfer apparatus. 2. NuPAGE® Transfer Buffer (20
) (Invitrogen). 3. Nitrocellulose/polyvinylidene fluoride (PVDF) membrane. 4. Blotting paper.

Recombinant Protein Production in Tobacco

25

5. Methanol (if using PVDF). 6. Tweezers. 7. Blocking buffer (5% w/v NFDM in PBS, store at 4  C). 8. Wash buffer (0.1% v/v Tween20 in dH2O). 9. ECL Prime Western Blotting System (GE). 10. Antibodies—primary antibody (antigen specific), secondary antibody (horseradish peroxidase (HRP) or alkaline phosphatase (AP) conjugated) as required. 2.7 Purification of Recombinant Protein from Tobacco by Affinity Chromatography

1. Retort stand and clamps. 2. Chromatography column (1–1.5 cm diameter). 3. 20% v/v ethanol. 4. Protein A resin. 5. PBS pH 7.4. 6. Benchtop blender. 7. Miracloth. 8. Centrifuge. 9. pH meter. 10. Vacuum pump/syringe and 0.22 μm filters. 11. Peristaltic pump. 12. Elution buffer (0.1 M glycine pH 2.7). 13. Neutralization buffer (1 M Tris base, pH unadjusted). 14. Spectrophotometer. 15. Dialyze tubing or cassette. 16. Magnetic stirrer. 17. Centrifugal protein concentrator.

2.8 Surface Plasmon Resonance

1. Biacore X100 (Cytiva): Machine for label-free, real-time analysis of biomolecular interactions using surface plasmon resonance. 2. CM5 chip (Cytiva): A general purpose chip for analyzing biomolecular interactions. Comprised of a carboxymethylated dextran matrix covalently attached to a gold surface. One binding partner is captured on the surface of the chip while the other is flowed over the chip. 3. Deionized water: All deionized water used in the Biacore system has to be purified to the sensitivity of at least 10 MΩ-cm at 25  C and filtered with 0.25 μm disc filters. 4. Immobilization buffer: 10 mM sodium acetate pH 5.0. 5. 1-Ethyl-3-(3-dimethylaminopropyl)-carbodiimide Part of the amine coupling kit (Cytiva).

(EDC):

26

Tim H. Szeto et al.

6. N-Hydroxysuccinimide (NHS): Used with EDC to activate the chip surface. Part of the amine coupling kit (Cytiva). 7. Ethanolamine: Used to block excess activated carboxyl after amine coupling during immobilization. Part of the amine coupling kit (Cytiva). 8. Running buffer (HBS-EP+): 10 mM HEPES, 150 mM NaCl, 3 mM EDTA and 0.005% surfactant P-20, pH 7.4. Available as 10x solution (Cytiva). 9. 50 mM NaOH: Make 1 M NaOH by mixing 4 g of NaOH with 80 mL of water. Top it up to 100 mL. Add 1 mL of 1 M NaOH to 19 mL of deionized water. Filter the solution using a 0.25 μm filter. 10. Regeneration solution: 10 mM glycine–HCl, pH 1.5 (Cytiva).

3

Methods All procedures should be undertaken following local health and safety regulations. Procedures described in Subheadings 3.1 and 3.2 should be performed under sterile conditions using aseptic media and equipment.

3.1 Freeze–Thaw Method for Agrobacterium tumefaciens Transformation

1. Inoculate 50 mL LB containing appropriate antibiotics (see Note 3) with Agrobacterium. 2. Shake at 28  C in the dark to an OD600 of 1.0. 3. Centrifuge the agrobacteria at 3000
g for 15 min. 4. Remove supernatant and resuspend the bacterial pellet in 1 mL of ice-cold 10 mM CaCl2. 5. Transfer 100 μL aliquots of the bacterial suspension into sterile 1.5 mL Eppendorf tubes and then place in liquid nitrogen. 6. Use frozen agrobacteria directly for DNA transformation or store at 80  C. 7. Pipette 10 μL of binary vector from a plasmid preparation onto the surface of 100 μL competent frozen agrobacteria and incubate for 5 min at 37  C in a water bath. 8. Add 1 mL of LB to the bacteria and shake at 28  C for 4 h. 9. Centrifuge the bacteria for 2 min at 12,000
g and resuspend the pellet in 100 μL LB. 10. Spread 50 μL of the bacterial suspension onto 25 mL of LB medium in a 9 cm Petri dish solidified by the addition of 15 g/ L agar containing appropriate antibiotics. Seal plate with Parafilm and incubate in dark at 28  C until colonies form (2–3 days).

Recombinant Protein Production in Tobacco

27

11. Re-streak individual colonies on selective LB in separate Petri dishes and incubate at 28  C for in the dark for 2 days. Confirm presence of binary vector containing GOI by PCR using suitable primers. 12. Inoculate agrobacteria in LB broth, grow to OD600 of 1.0. Take 850 μL and mix with 15 μL of sterile glycerol in a cryogenic tube. Store at 80  C. 3.2 AgrobacteriumMediated Transformation of and In Vitro Regeneration of Plantlets

1. Remove Agrobacterium from 80  C and with an inoculating loop streak an aliquot of frozen bacteria to 25 mL solidified LB medium in a 9 cm Petri dish. Seal with Parafilm and incubate at 28  C for 2 days. 2. Using an inoculating loop, inoculate 10 mL of LB containing appropriate antibiotics with Agrobacterium. Ensure bacteria are well dispersed by vigorous manual shaking. Shake at 28  C to an OD600 of 1.0. For leaf disc transformation take 4 mL and add to 16 mL of sterile distilled water in a 9 cm Petri dish. 3. Remove leaves from tobacco plants (we use Nicotiana tabacum cv. Petit Havana SR1) growing in compost and immerse them in 10% v/v commercial bleach (see Note 4). Leave for 10 min. Wash with 7 changes of sterile distilled water. 4. Using a scalpel, cut 0.5–1 cm2 discs from tobacco leaves. Immerse for 5 min in the diluted Agrobacterium suspension. 5. Briefly blot dry leaf discs on sterile filter paper and place on 25 mL of Shoot Regeneration Medium in a 9 cm Petri dish, 10–20 discs/dish. Leave for 2 days at 25  C with 16 h photoperiod. 6. Transfer leaf discs to 25 mL Shoot Regeneration Medium in a Petri dish containing 500 mg/L carbenicillin and 200 mg/L kanamycin, 10 discs/dish (see Note 5). 7. Subculture to fresh medium every 4 weeks. 8. Developing shoots should be removed using a scalpel when they reach a convenient size (approx. 0.5 cm). If individual shoot clones are required, a single shoot should be removed from a particular leaf disc and the disc should then be discarded (shoots selected from different discs must obviously have developed from different cells and therefore are not clonal). Shoots should be transferred to Rooting Medium containing 500 mg/ L carbenicillin and 200 mg/L kanamycin, 2–3 shoots/40 mL medium in a 175 mL glass jar. If roots have not formed after 14 days, 2–3 mm should be cut from the base of the shoot with a scalpel and the shoot transferred to fresh Rooting Medium containing 500 mg/L carbenicillin and 200 mg/L kanamycin.

28

Tim H. Szeto et al.

3.3 Transfer of In Vitro-Grown Plantlets to Soil, CrossFertilization, Seed Collection

1. Gently cut agar around rooted plantlet with a scalpel and with forceps remove plantlet and place in water in a suitable receptacle. With fingers gently remove agar from around the base of the roots taking care not to damage the plant. Leave plantlets in water until ready to transfer to compost. 2. Place compost in a plant pot and water. With a pencil end or finger, poke a 1–2 cm hole in compost. Place plantlet in hole and cover roots with compost. Gently press fingers down around base of shoot to firm up compost. 3. Place pot with plantlet in a propagator with vented lid. Initially keep propagator lid on with vents closed. After 7 days open vents and after another 24 h remove the lid. Maintain at 25  C with 16 h photoperiod. 4. Water plants as required with water containing soluble plant nutrient. 5. When flowers form, plants can be cross-fertilized if required. Cut petals on an immature recipient flower with a scalpel and remove anthers to prevent self-fertilization. Remove mature anther with pollen from donor plant and rub anther on stigma to transfer pollen. Cover pollinated flower with light polythene bag to prevent unwanted cross-pollination from other donor plants (see Note 6). 6. When seed pods are mature, cut the top with scissors and invert into an Eppendorf tube or another suitable container for seed collection (see Note 7). 7. Seeds can be germinated by sprinkling on the surface of wet compost in a plant pot maintained in a seed propagator. 8. Homozygotes among these progenies may be identified by growing them to the flowering stage and backcrossing with wild-type plants. However, care must be taken to analyze sufficient progeny of these crosses, as plants with multiple T-DNA insertions on different chromosomes will give rise to a high proportion of progeny expressing the GOI when backcrossed with wild-type plants.

3.4

ELISA

3.4.1 Direct ELISA for Measuring the Presence of α(1,3)-Fucose

Three ELISAs are described here—direct (Subheading 3.4.1), indirect (Subheading 3.4.2), and sandwich (Subheading 3.4.3). For direct ELISA, measuring the presence of α(1,3)-fucose in plant proteins is used as an example, indirect ELISA is illustrated by characterizing HIV antigen binding capacity of mAb VRC01 and finally we give the method for determining recombinant VRC01 concentration in plant extract by sandwich ELISA. The majority of plant glycans contain α(1,3)-fucose and β(1,2)xylose. Some applications require the removal of these residues from plant-made antibodies to improve e.g., the effector functions

Recombinant Protein Production in Tobacco

29

of the antibodies [38, 39]. The enzymes fucosyltransferase (FucT) and xylosyltransferase (XylT) can be deactivated using gene silencing or genome editing techniques to remove the fucose and xylose residues respectively from recombinant proteins. 1. Conjugate HRP to anti-fucose antibody using a commercially available kit (see Note 8). 2. To extract total soluble protein, add 100 mg of leaves from putative FucT knockout tobacco plants, with two 3.175 mm steel ball bearings and 300 μL PBS in a 1.5 mL Eppendorf tube. Insert Eppendorf tubes into a Mixer Mill MM400 and homogenize for 3 min (see Note 9). Wild-type plants are used as positive control. 3. Spin the Eppendorf tubes for 15 min at 9500
g at 4  C. Pipette the supernatant (the total soluble protein) into a new tube. 4. Make a 1 in 4 dilution of the total soluble protein in PBS. Pipette 100 μL of the dilution into a well on the first row of a transparent, flat-bottomed ELISA plate. Pipette 50 μL of PBS in subsequent rows or columns. Perform a 1 in 2 serial dilution across or down the plate (pipette 50 μL from the preceding well to the current well, mix well and pipette 50 μL from the current well to the next well). For background subtraction, coat the wells with PBS. 5. Seal the ELISA plate with a plate sealer or clingfilm to prevent evaporation. Incubate plate for 1–2 h at 37  C. 6. Remove the plant supernatant and wash the plate with 2
wash buffer and 1
deionized water. Dry the plates before the proceeding. 7. Pipette 200 μL of Blocking buffer into all wells. Seal plate and incubate plate for 1 h at 37  C. 8. Dilute 1 in 1500 HRP-conjugated anti-fucose antibody in Blocking buffer (5 mL per plate). Optimum antibody dilution can be determined beforehand. 9. Wash the plates with 2
wash buffer and 1
deionized water. Dry the plates before proceeding. 10. Pipette 50 μL of HRP-conjugated anti-fucose antibody in Blocking buffer (step 8) into all wells. Seal plate and incubate plate for 1 h at 37  C. 11. Wash the plates again with 2
wash buffer and 1
deionized water. Dry the plates before proceeding. 12. Pipette 50 μL of TMB solution into all the wells. Blue color with intensity proportional to concentration of α(1,3)-fucose will develop.

30

Tim H. Szeto et al.

13. Stop the reaction by adding 50 μL of 2 M sulfuric acid (use with caution). The wells will turn yellow. 14. Measure the absorbance at 450 nm using a 96-well plate reader. 3.4.2 Indirect ELISA for Characterizing HIV Antigen Binding Capacity

VRC01 is a broadly neutralizing antibody (bNAb) that targets the CD4-binding site of the HIV-1 envelope protein [40]. VRC01 present in crude extracts from transgenic expressing this mAb [41] will bind to the HIV-1 antigen gp120 coated on ELISA plates. In turn, it will be detected using an anti-human heavy chain antibody conjugated with HRP. 1. Prepare HIV-1 gp120 antigen in PBS at a concentration of 1 μg/mL. Coat the wells of a 96-well transparent, flatbottomed plate with 50 μL of the diluted antigen. 2. Seal the ELISA plate with a plate sealer or clingfilm to prevent evaporation. Incubate plate for 1–2 h at 37  C, or overnight at 4  C. 3. Remove the antigen and wash the plate with 2
wash buffer and 1
deionized water. Dry the plates before proceeding. 4. Pipette 200 μL of Blocking buffer into all wells. Seal plate and incubate plate for 1 h at 37  C. 5. To extract total soluble protein, 100 mg of leaves from transgenic VRC01 tobacco plants, two 3.175 mm steel ball bearings and 300 μL PBS are added into a 1.5 mL Eppendorf tube. The Eppendorf tubes are inserted into a Mixer Mill MM400 and homogenized for 3 min (see Note 9). Wild-type plants are used as a negative control. 6. Spin the Eppendorf tubes for 15 min at 9500
g at 4  C. Pipette the supernatant (the total soluble protein) into a new tube. 7. Wash the plates with 2
wash buffer and 1
deionized water. Dry the plates before proceeding. 8. Perform a serial dilution of the total soluble protein with PBS. Total volume in the wells should be 50 μL. 9. Seal plate and incubate plate for 2 h at 37  C, or 4  C overnight. 10. Wash the plates with 2
wash buffer and 1
deionized water. Dry the plates before proceeding. 11. Dilute 1 in 1000 HRP-conjugated anti-human heavy (γ) chain antibody in Blocking buffer (5 mL per plate). Pipette 50 μL of antibody in Blocking buffer into all the wells. 12. Seal plate and incubate plate for 1 h at 37  C.

Recombinant Protein Production in Tobacco

31

13. Wash the plates with 2
wash buffer and 1
deionized water. Dry the plates before proceeding. 14. Pipette 50 μL of TMB solution into all the wells (blue color will develop). 15. Stop the reaction by adding 50 μL of 2 M sulfuric acid (yellow color will develop). 16. Measure the absorbance at 450 nm using a 96-well plate reader. 3.4.3 Sandwich ELISA for Determining VRC01 Concentration

Two antibodies recognizing the light and heavy chain of the human anti-HIV antibody VRC01 can be used to determine the concentration of intact VRC01 in plant crude extracts using a sandwich ELISA. One of the antibodies, in its non-conjugated form, is used as a capture antibody, the other, which is HRP-conjugated, is used to detect the captured VRC01. 1. Dilute 1 in 200 anti-heavy (γ) chain in PBS. Coat the wells of a 96-well transparent, flat-bottomed plate with 50 μL of the diluted antibody. 2. Seal the ELISA plate with a plate sealer or clingfilm to prevent evaporation. Incubate plate for 1–2 h at 37  C, or overnight at 4  C. 3. Remove the antigen and wash the plate with 2
wash buffer and 1
deionized water. Dry the plates before proceeding. 4. Pipette 200 μL of Blocking buffer into all wells. Seal plate and incubate plate for 1 h at 37  C. 5. To extract total soluble protein, 100 mg of leaves from transgenic VRC01 tobacco plants, two 3.175 mm steel ball bearings and 300 μL PBS are added into a 1.5 mL Eppendorf tube. The Eppendorf tubes are inserted into a Mixer Mill MM400 and homogenized for 3 min (see Note 9). Wild-type plants are used as a negative control. 6. Spin the Eppendorf tubes for 15 min at 9500
g at 4  C. Pipette the supernatant (the total soluble protein) into a new tube. 7. Wash the plates with 2
wash buffer and 1
deionized water. Dry the plates before proceeding. 8. Perform a serial dilution of the total soluble protein with PBS. Total volume in the wells should be 50 μL. Known concentrations of commercially available human IgG kappa are used as a standard (see Note 10). 9. Seal plate and incubate plate for 2 h at 37  C, or 4  C overnight. 10. Wash the plates with 2
wash buffer and 1
deionized water. Dry the plates before proceeding.

32

Tim H. Szeto et al.

11. Dilute 1 in 1000 HRP-conjugated anti-human light (kappa) chain antibody in Blocking buffer (5 mL per plate). Pipette 50 μL of antibody in Blocking buffer into all the wells. 12. Seal plate and incubate plate for 1 h at 37  C. 13. Wash the plates with 2
wash buffer and 1
deionized water. Dry the plates before proceeding. 14. Pipette 50 μL of TMB solution into all the wells. The wells containing plant-made VRC01 will generate blue color. 15. Stop the reaction when the most diluted wells of the standard curve are just turning blue by adding 50 μL of 2 M sulfuric acid. The wells will turn yellow. 16. Measure the absorbance at 450 nm using a 96-well plate reader. Use the absorbance readings of the standard curve to determine the concentration of VRC01 in transgenic plant samples, multiply the concentration with the dilution factor as appropriate. Expression level can be presented as concentration of protein in mg/kg leaf fresh weight or as percentage of total soluble protein (see Note 11). 3.5 SDS Polyacrylamide Gel Electrophoresis

1. Prepare a sufficient volume of each protein sample for intended analyses by mixing with an appropriate volume of 4
concentrated sample buffer to achieve 1
(see Note 12). If analyzing reduced protein samples as well, also prepare a set of duplicate samples with the addition of 5% v/v β-mercaptoethanol in a fume hood (see Note 13). 2. Heat the samples between 70 and 90  C for 5–10 min in a heating block (see Note 14). 3. Remove the samples from the heating block and allow to cool to room temperature before pulse centrifuging to remove condensation on lids. The samples are ready to be loaded onto the gel. Alternatively, the samples can be frozen (20  C) until required. 4. Choose the precast gel with appropriate acrylamide concentration and select either MOPS or MES running buffer appropriate for the separation of the protein(s) of interest. Dilute 50 mL of the 20
running buffer into 950 mL dH2O (see Note 15). This is sufficient for running one gel tank (maximum of 2 gels). Scale as required. 5. Remove the precast gel from packaging and briefly rinse in dH2O. Do not forget to remove the white tape at the bottom of the gel cassette before assembly into tank. Position the Buffer Core in the XCell SureLock Mini-Cell Electrophoresis System and slide the gel(s) into position either side of the Buffer Core, ensuring the comb(s) of the gel(s) face inward toward the Buffer Core to create the upper buffer chamber.

Recombinant Protein Production in Tobacco

33

Clamp into the tank with the Gel Tension Wedge. Fill the upper buffer chamber with running buffer until it just overflows. Check that the buffer level doesn’t drop and there is no leaking into the lower buffer chamber. Carefully remove the gel comb by sliding it up in a smooth motion and keeping it horizontal. Removing it at an angle may damage the gel wells resulting in samples spilling or mixing during loading. Continue pouring running buffer into the upper chamber (this helps to rinse the wells) allowing it to overflow into the lower chamber until the desired level is reached (at least above the level of the white tape removed earlier). In addition to completing the circuit, the buffer in the lower chamber also provides heat dissipation from the gel during electrophoresis aiding more uniform migration. 6. Load 5 μL of the prestained protein standard and 5–15 μL of the samples (see Note 16). 7. Carefully top up the upper chamber buffer level if it has dropped significantly during sample loading. Attach the lid and connect the cables to the power supply. Run the gel at 100–150 V for 60 min (the dye front will have reached the bottom of the gel) or until the desired separation has been achieved. 8. Remove the gel cassette from the apparatus and pry it open with a spatula. The gel will adhere to one side of the cassette (discard the other). Cut off the floppy wells at the top of the gel and also remove the gel foot using the spatula. 9. For total protein visualization, submerge the gel into a container of Coomassie stain and agitate on an orbital shaker/ rocker for 20–30 min (or overnight—see Note 17). 10. Discard the Coomassie stain, rinse the gel with dH2O, and add fresh dH2O to destain the gel (see Note 18). Destain gel until protein bands are clear and distinct from background (gels can be left to destain overnight on the orbital shaker/rocker—see Note 17). 11. Record an image of the gel outcome. 3.6 Semi-Dry Enhanced Chemiluminescence (ECL) Western Blot

1. Add 50 mL 20
transfer buffer to 950 mL dH2O. 2. Cut blotting paper (4–6 pieces) and a transfer membrane (nitrocellulose or PVDF) a fraction larger than the size of the gel to ensure coverage. 3. Soak the paper and membrane in separate trays of transfer buffer. 4. Assemble the “blotting sandwich” as follows: use tweezers or gloves to handle the edge of the papers and membrane. Do not use tweezers to handle the gel. Layer in the following order

34

Tim H. Szeto et al.

(1) 2–3 wet paper on the anode surface, (2) membrane, (3) gel, and (4) 2–3 wet paper (see Note 19). 5. Close the lid (cathode) and run at 20 V for 30–45 min. 6. After transfer, peel each layer off and rinse the membrane briefly in dH2O. Successful transfer is indicated by the appearance of the prestained protein standards on the membrane and none remaining on the gel. The gel may be stained with Coomassie to determine whether the protein has fully transferred. 7. Pour 20 mL blocking solution onto the membrane to block for 30–60 min on an orbital shaker/rocker (or store at 4  C overnight). 8. Pour off the blocking solution and add primary antibody to new blocking solution (15–20 mL) at an appropriate dilution (see Note 20). Incubate for 30–60 min on an orbital shaker/ rocker (or store at 4  C overnight). 9. Pour off the blocking solution and primary antibody. Rinse the membrane three times for 5 min each time with wash buffer on the orbital shaker/rocker. If the primary antibody is conjugated to HRP or AP proceed to detection step 12. 10. Pour off the wash buffer and add secondary antibody to new blocking solution (15–20 mL) at an appropriate dilution (see Note 20). Incubate for 30–60 min on an orbital shaker/rocker (or store at 4  C overnight). 11. Pour off the blocking solution and secondary antibody. Rinse the membrane three times for 5 min each time with wash buffer on the orbital shaker/rocker. 12. Prepare the 1:1 mix of ECL Prime detection reagents (see Note 21). Cut a sheet of acetate larger than the dimensions of the membrane. 13. Holding the membrane by the top edge using tweezers, remove it from the wash container and dry it of excess wash buffer by holding it vertically and touching the bottom edge to a tissue. The excess wash buffer should drain down the membrane and be absorbed by the tissue. The membrane should be moist. 14. Place the membrane face up on the sheet of acetate (on a flat surface) and apply the ECL detection reagent to cover the entire membrane evenly and allow to incubate for 5 min. 15. Lifting one side of the acetate sheet drain off excess ECL detection reagent onto a tissue paper. 16. Image and record the result using an appropriate gel documentation imaging system (see Fig. 3 for an example of an ECL non-reducing western blot).

Recombinant Protein Production in Tobacco

35

Fig. 3 A non-reducing enhanced chemiluminescence (ECL) western blot. Putatively transgenic tobacco plants have been screened to test for expression of two proteins: B-cell receptor complex- associated alpha chain linked to cyan fluorescence protein and hemagglutin epitope tags (αCFPHA) and B-cell receptor complex- associated beta chain linked to the FLAG epitope, which when present together will assemble into a heterodimer. Plants 1, 2, 4, and 7 have both proteins as the heterodimer forms an immunoreactive band at ~85 kDa. A plant expressing αCFPHA is included as a control. Plants were probed with an antibody detecting αCFPHA 3.7 Purification of Recombinant Protein from Tobacco by Affinity Chromatography

1. Prepare capture column by rinsing the column with 20% ethanol. 2. Apply Protein A resin to column and allow to settle (see Note 22). 3. Wash resin bed with 5
column volume dH2O followed by 10
column volume PBS, and store column vertically at 4  C until required. Do not allow resin to run dry. 4. Harvest the leaves expressing protein of interest and weigh (see Note 23). 5. Add 3
volumes/weight ice-cold PBS and blend using benchtop blender on high setting for 1 min-bursts until a homogenous consistency is reached (see Note 24). 6. Filter resulting homogenate using Miracloth to remove larger pieces of plant debris. 7. Adjust the pH of the filtered homogenate to 7–8 (see Note 25). 8. Centrifuge homogenate at 21,000 to 24,000
g at 4–10  C for 30–60 min. 9. Filter clarified supernatant through a 0.22 μm filter (see Note 26) to prevent blockage of the capture resin.

36

Tim H. Szeto et al.

10. Collect an aliquot of sample for later analyses. 11. Gently apply a small volume of the filtered solution onto the capture column resin (see Note 27). 12. Deliver the sample into the column using the peristaltic pump maintaining a constant flowrate of approximately 1 drop/1–2 s (see Note 28). 13. Wash the capture resin with 5–10
column volume of ice-cold PBS (see Note 29). 14. Elute the protein of interest in 0.5–1 mL fractions with 0.1 M glycine pH 2.7 and neutralizing eluate with 1 M Tris base (pH unadjusted) (see Note 30). 15. Determine IgG-containing fractions using a suitable spectrophotometer (e.g., Nanodrop). 16. Pool and dialyze the IgG-containing fractions against PBS at 4  C overnight with constant stirring (see Note 31). 17. Recover the sample and concentrate using a centrifugal device of 50 kDa molecular weight cut off (MWCO). 18. The purified IgG can be stored in the fridge in the short term or frozen for longer term storage. 3.8 Surface Plasmon Resonance

All experiments described here are performed on a Biacore X100 (Cytiva) at 25  C with a CM5 sensor chip. Other chips are available for different assay needs. There are three steps in a direct SPR assay: (1) Ligand immobilization, (2) Analyte introduction and dissociation, (3) Regeneration of chip surface. For capture assays, the capture molecule is immobilized instead of the ligand. An additional step is introduced to capture the ligand on the chip surface before the analyte is introduced. Although the same chip is used in the assays, Protein A is defined as a ligand in direct assays and as capture molecule in capture assays. Detailed methods and further applications can be found in the Biacore Assay Handbook (Cytiva; available online at cytivalifesciences.co.jp). Evaluation is performed using the Biacore X100 Evaluation Plus software.

3.8.1 Immobilization of Protein A onto CM5 Chip for Direct and Capture Assays

1. Bring a CM5 sensor chip up to room temperature at least 30 min before experiment. The chip has two Flow cells. 2. Dock the chip in the carrier and connect Running buffer to the system. Prime with Running buffer to flush the system with the Running buffer (see Note 32). 3. Create a new wizard template in the Biacore X100 Control software by selecting “Wizards”. Under “Surface preparation”, select “Immobilization”. Select “New” to create a new template.

Recombinant Protein Production in Tobacco

37

4. Choose the type of sensor chip (CM5), the flow cell to immobilize the ligand and “Amine” for the method for capture. 5. Select “Aim for” under “immobilization method” and type 4000 Resonance Units (RUs; see Note 33). Alternatively, if injection time to achieve required RUs is known, select “Specify time and flow rate”. Injection time is usually set to 2 min, but longer times may be required. The flow rate is normally set to a slow 5 μL/min to increase ligand absorption. 6. If the “Aim for” option is selected, include 50 mM NaOH Wash solution to remove bound proteins during the immobilization scouting phase of the program. 7. Dilute Staphylococcus aureus Protein A in immobilization buffer at 35 μg/mL in volumes specified by the wizard (see Note 34). 8. Add the reagents for immobilization (EDH, NHS, and Ethanolamine) along with Protein A diluted in immobilization buffer to their assigned places in the sample rack. 9. Start the run by selecting “Next” and “Start”. At the end of the run, a table will display the immobilization level. 10. Repeat process for the other flow cell. 11. Proceed to Subheadings 3.8.2 or 3.8.3. Alternatively, the sensor chip can be kept in Running buffer at 4  C until further use. 3.8.2 Direct Assay for Measuring Concentration of Purified VRC01

This method uses a calibrated assay, interpolating unknown concentrations using the response (measured in Response Unit, or RU) of known concentrations on a standard curve. 1. If it is not already been set up, bring the CM5 chip coated with Protein A (see previous section) up to room temperature at least 30 min before experiment. Dock the chip into the machine along with Running buffer and flush the system with Running buffer. 2. Optimize the method and injection time for analyte removal to regenerate the chip surface. In the case of VRC01, two 30 s injections of 10 mM glycine, pH 1.5 at 10 μL/min can be used for surface regeneration of the Protein A-coated chip. 3. Create a new wizard template in the Biacore X100 Control software by selecting “Wizards”. Under “Assay”, select “concentration” followed by “using calibration”. Select “New” to create a new template. 4. Select one of the Flow cells (e.g., Flow cell 2) to introduce the standards and unknown samples. Don’t inject any protein into the other Flow cell so it can be used as a blank correction. Specify the number of regeneration injections to both Flow cells.

38

Tim H. Szeto et al.

5. Choose “Prime before run” if the system has not been set up and flushed with Running buffer. You can also choose to run start-up cycles. 6. Specify contact time of the sample and regeneration cycles. In this case, 180 s will be used. 7. Dilute 1000, 500, 250, 125, 62.5, 31.25, and 15.625 ng/mL of human IgG kappa standard in Running buffer, together with dilutions of purified VRC01 protein (see Note 35). 8. Input the concentrations of the standard in the wizard template, select “Next” and input the names and dilutions of the unknowns. 9. Add the required amount of blank (Running buffer), standards, unknown samples, Regeneration solution, and deionized water for injection port washing to their assigned places on the rack. 10. Start the run by selecting “Next” and “Start”. At the end of the run, a sensorgram showing the response curves for the standard and the unknowns are shown. 11. Analyze the sensorgrams by using the Biacore X100 Evaluation software. Select “Concentration analysis”. The program will perform blank subtraction, plot the standard curve and calculate the concentration of the unknown proteins. Select “finish” to save the evaluation. 12. At the end of the experiment, the sensor chip can be kept in Running buffer at 4  C until further use. 3.8.3 Capture Assay for Measuring Binding Kinetics of Glycoengineered VRC01

There are various studies that suggest apart from neutralizing the virus directly, anti-HIV bNAbs like VRC01 [40] can control persistent viral reservoirs [42–44], most probably through Fc receptor interactions [45, 46]. FcγRI receptor is one example of Fc receptors that a bNAb will interact with to recruit immune cells to remove or neutralize targets. The majority of plant glycans contain α(1,3)fucose and β(1,2)-xylose, the removal of which have been shown to improve effector functions of plant-made antibodies [38, 39]. Glyco-modified mAbs like VRC01 can be made in glycoengineered plants which have their xylosyltransferase and fucosyltransferase enzymes knocked out [47]. Here, the binding kinetics of glycol-modified mAb VRC01 to FcγRI receptor is determined by SPR and compared to VRC01 with wild-type glycosylation patterns. 1. Bring the CM5 chip coated with Protein A up to room temperature at least 30 min before experiment. Dock the chip into the machine along with Running buffer and flush the system with Running buffer.

Recombinant Protein Production in Tobacco

39

Fig. 4 Sensorgram showing high affinity interaction between capture molecule and ligand, with minimal dissociation of the ligand with the capture molecule. Furthermore, the regeneration method is able to remove most of the captured ligand and return the signal back to baseline

2. Perform optimization experiments to determine if there is very high affinity binding between the capture molecule (Protein A) and the ligand (VRC01). This is usually indicated by a nearly flat line in the dissociation phase on the sensorgram (Fig. 4). The capture molecule should not interact with the analyte and the capture step should be highly reproducible. 3. Optimize the method for the removal of the ligand and injection time to regenerate the chip surface so the signal is very close to baseline (Fig. 4). In this case, 10 mM glycine, pH 1.5 will be used to regenerate the chip surface. The “Regeneration scouting” wizard template can be used to perform this optimization. 4. Determine the level of captured ligand (RL) necessary to reach an Rmax of 50 RUs using the formula: Rmax ¼

MW analyte
RL
S MW ligand

where MW—molecular weight (Da). RL—level of immobilized/captured ligand (RU), i.e., 50 RUs. S—stoichiometric ratio. 5. Create a new wizard template in the Biacore X100 Control software by selecting “Wizards”. Under “Assay”, select “Custom assay wizard”. Select “New” to create a new template.

40

Tim H. Szeto et al.

6. Select one of the Flow cells to introduce the standards and unknown samples, while the other Flow cell will be used for blank subtraction (option “2–1”). Input chip type and select “Binding analysis” for evaluation purpose. 7. Choose “Prime before run” if the system has not been set up and flushed with Running buffer. Run start-up cycles can also be chosen. 8. Add 1 capture step and the number of regeneration steps required. Set the flow rate of capture to 20 μL/min through Flow cell 2 for the capture step, and the optimized regeneration flow rate(s) through both Flow cells. Add stabilization period if required. 9. Dilute purified VRC01 made in glycoengineered and wild-type plants 1 in 1000 in Running buffer. Input sample names and contact times into sample table. Start with 30 s and increase if necessary. 10. Add the required amount of blank (Running buffer), unknown samples, Regeneration solution and deionized water for injection port washing to their assigned places on the rack. 11. Use the resulting RUs to determine the injection time to reach RL. Injections of 30–100 s are recommended for kinetic assays. Repeat scouting experiment (steps 9 and 10) if necessary. 12. Create a new wizard template in the Biacore X100 Control software by selecting “Wizards”. Create a new “Custom assay wizard” template. 13. Input information on the chip and the optimized parameters for ligand capture and regeneration. Single-cycle or multi-cycle kinetic analysis can be chosen (see Note 36). Here, the multicycle kinetic analysis is used, with surface regenerations using regeneration buffer in between cycles. 14. In the cycle protocol, introduce 3 start-up cycles using Running buffer before the actual kinetic analyses. 15. Select the correct parameters in the program to capture glycolmodified VRC01 (the ligand) in Flow cell 2 at 20 μL/min following the optimized procedures and injection times while Flow cell 1 is used as a baseline reference. 16. After ligand capture, program the Biacore to inject 0, 3.75, 7.5, 15, 30, and 60 nM of human FcγRI (the analyte) into both Flow cells for 135 s, followed by dissociation time of 600 s at a flow rate of 40 μL/min. Begin with a lower analyte concentration and gradually increase. 17. Program a similar kinetic analysis to 15 and 16, but with VRC01 produced in non-glycoengineered plants.

Recombinant Protein Production in Tobacco

41

18. Add required amount of blank (Running buffer), ligands, different analyte concentrations and Regeneration solution to their assigned places on the rack. 19. Start the run by selecting “Next” and the “Start”. At the end of the run, the Biacore will display a sensorgram showing the response curves for the standard and the unknowns. 20. Analyze the sensorgrams by using the Biacore X100 Evaluation software. Select the “Kinetics/Affinity” button. The program will perform blank correction and align the curves to start of analyte injection. Each ligand group will be analyzed separately. Select response curves from one group (e.g., VRC01 with modified glycans). 21. Select “Next”. Remove portions of the plots not associated with the association or dissociation curve. Select “Kinetics”. 22. Select the binding model that fits the presumed interaction between ligand and analyte. 23. The fitted curves will appear as a black line overlaying the experimental data (Fig. 5). Check the fit by observing the overlap between the fitted curve and the experimental data. Confirm by examining the “Residual plot”. The plot will indicate any deviations between the fitted curves and experimental data by displacement from the zero line (noise level should be in the order of 2 RU). 24. Repeat for the other ligand group (e.g., VRC01 with wild-type glycosylation). 25. At the end of the experiment, the sensor chip can be kept in Running buffer at 4  C until further use.

Fig. 5 Example of experimental data overlaid by fitted curves on a sensorgram

42

4

Tim H. Szeto et al.

Notes 1. Agrobacteria may form clumps in LB medium. If this occurs, Yeast-Mannitol medium (0.4 g/L yeast extract, 10 g/L mannitol, 0.1 g/L NaCl, 0.2 g/L MgSO47H2O, 0.5 g/L K2HPO43H2O) may be used instead (and can also be used for the leaf disc transformation procedure). Agrobacterium suspensions generated in Yeast-Mannitol medium may be less dense than in LB and should be diluted accordingly prior to leaf disc transformation. 2. Always add acid to water and not the other way around, as this is a highly exothermic reaction. Start off with the maximum volume of water which can be used to make up the final concentration and slowly add small quantities of sulfuric acid at a time until the 2 M concentration is reached. 3. Agrobacterium chromosome and helper plasmid may contain genes conferring resistance to particular antibiotics. The medium should contain appropriate antibiotics for the bacterial strain. The binary vector will also confer resistance to an antibiotic through its bacterial selectable marker; this should be to an antibiotic to which the Agrobacterium containing helper plasmid is sensitive, to allow selection of bacteria containing the binary vector. 4. Alternatives to commercial bleach for sterilization of plant material include sodium hypochlorite, 70% v/v ethanol, hydrogen peroxide and calcium hypochlorite. We use 175 mL glass jars for sterilization and consequently pick leaves approximately 10 cm in length. In our experience, individual leaves vary considerably in ability to regenerate shoots. As the reason for this is not obvious, and as it is not possible to tell which leaves will be better than others, it is prudent to use several different leaves for the transformation process. 5. If agrobacteria are resistant to carbenicillin, we use cefotaxime for the eradication process at a concentration of 500 mg/L. This antibiotic is very photosensitive and should be kept in the dark prior to use. Other antibiotics such as timentin may also be used. 6. For self-fertilization, cover immature flower with a light polythene bag to prevent cross-fertilization. The covered flower will mature and self-pollinate without further assistance. 7. Mature tobacco seed pods are brown in color and desiccated in appearance. Shaking of mature pods leads to audible seed agitation confirming suitability for harvesting. 8. If an HRP-conjugated primary antibody is not available for a direct ELISA, an indirect ELISA using HRP-conjugated

Recombinant Protein Production in Tobacco

43

antibody is recommended for most applications. If necessary, HRP can be conjugated to unlabeled antibodies using a commercial kit (e.g., EZ-link™ activated peroxidase antibody labeling kit by Thermo Scientific Pierce). 9. Total soluble protein can also be extracted using pestle and mortar. 10. Optimize the assay so the signals of the unknown proteins fall within the standard curve. 11. To determine total soluble protein, measure 1 μL of crude extract at absorbance at 280 nm using a Nanodrop. Alternatively, protein concentration can be determined using Bradford assay [48] or BCA assay [49] using known concentrations of BSA. 12. A total volume of 50–100 μL of each sample should suffice for a PAGE and western blot with repeats. The thin walls of the 0.2 mL microfuge tubes will allow for more efficient heating (denaturing) of the sample. 13. β-mercaptoethanol has a strong offensive odor and is a respiratory irritant. It should therefore be used in a fume hood. Alternatively, use of other reducing agents (e.g., 100 mM DTT, or 50 mM TCEP) will obviate these undesirable effects. 14. Place a thermometer into one heating block tube hole. To ensure even heating of the samples, fill all the heating block tube holes with dH2O and place the tubes of samples in the heated water. Due to evaporation, the heating block tube holes may need to be topped up with dH2O (a squirt bottle is suitable for this). If heating samples containing β-mercaptoethanol, operate the heating block in a fume hood. 15. Running buffers are stable at room temperature for several months, and can be reused 2–3 times before discarding. 16. Smaller volumes can be difficult to measure accurately and, depending on well volume, larger volumes risk overflow and lane contamination. If analyzing purified proteins, aim to load 0.5–1 μg/well. This will result in a thin, horizontal Coomassie blue stained band. For crude extracts this will be more empirical. Overloading a well will result in broad bands masking other bands and confounding analyses. During migration, concentrated/overloaded samples may diffuse horizontally and encroach on neighboring lanes. Allowing a lane’s separation between samples and loading 1
sample loading buffer into the unused wells can help mitigate horizontal diffusion. 17. When staining/destaining overnight, prevent evaporation by providing a lid for the container. 18. Position the gel in the middle of the container. Roll up some tissue paper and wet it. Place it around the gel and add dH2O

44

Tim H. Szeto et al.

until the gel is submerged. Place the gel on the orbital shaker/ rocker to destain. The tissue paper adsorbs the Coomassie dye from the water, reducing the need to change the water. 19. Avoid air bubbles between layers by placing an edge of the paper/membrane/gel on the anode and allowing the paper/ membrane/gel to roll onto the previous layer. Remove any air bubbles by rolling a clean pipette over each layer after it is laid. Ensure the entire sandwich is still moist after assembly (add transfer buffer dropwise if needed). Avoid scratching the anode surface by using plastic tweezers. 20. Do not apply the antibody directly onto the membrane. Tilt the container to pool the blocking solution into one corner and pipette the antibody into it. Swirl gently to mix. 21. A total of 1 mL ECL Prime 1:1 mix is sufficient for a membrane ~8
8 cm. If the surface is uneven, parts of the membrane may have more solution. Gently tilt the acetate sheet to redistribute the solution to parts of the membrane that have less. 22. The Protein A resin (agarose or sepharose) is usually supplied as a 50% slurry, so for an X mL column 2X mL of slurry needs to be applied. The amount of capture resin to use will depend on the binding capacity and the expected/estimated expression of the protein of interest. For instance, in our laboratory a 1–2 mL Protein A bed (2–4 mL 50% slurry) is usually sufficient for capturing IgG expressed from ~30 plants. 23. It is best to work immediately with the freshly harvested leaves. However, if circumstances don’t allow, the plant material can be flash frozen in liquid nitrogen and stored at 80  C until required. 24. The cooled PBS will mitigate heating of the plant material during the blending process to preserve the integrity of the protein of interest. Alternatively, carry out the blending in a cold room. 3
volume/weight example: for 100 g of plant material add 300 mL of ice-cold PBS. 25. If the pI of the protein of interest is similar to the binding pH, then adjust the pH to at least one pH unit above/below the pI. 26. Small volumes (30 plants). Cover the base of a 3- to 4-week-old plant such that the soil is retained in the pot during the infiltration procedure. Invert the plant into a beaker containing the Agrobacterium suspension and place the beaker in the center of the vacuum desiccator unit. Ensure that all the leaves are submerged in the Agrobacterium suspension. Close and seal the desiccator and apply vacuum for 60 s at negative pressure of 25 inHg (170 mbar). Break the vacuum gently and return the infiltrated plant to the growth room. More than one plant can be infiltrated at the same time depending on the size of the desiccator unit.

Infiltrated leaf tissue is typically harvested 6–8 days post-infiltration and can be processed fresh or snap-frozen and stored at 80  C. 1. Homogenize infiltrated leaf tissue in 3 volumes of Extraction buffer using a blender or mortar and pestle. 2. Filter over Miracloth. 3. Clarify at 13,000
g, 20 min, 4  C. 4. Add 5
Precipitation buffer for a final concentration of 4% (w/v) PEG-6000 and 0.2 M NaCl (see Note 4). 5. Stir overnight at 4–8  C. 6. Pellet at 13,000
g, 20 min, 4  C. 7. Discard supernatant and resuspend pellet in an appropriate volume (typically 0.5 mL per gram of processed leaf tissue) of 0.1
sodium phosphate buffer (10 mM sodium phosphate). Allow to resuspend on a shaker located in a 4–8  C cold room for 1–2 h then resuspend by pipetting. 8. Clarify at 27,000
g, 20 min, 4  C. 9. Filter supernatant over 0.2 μm syringe filter. 10. Pellet in an ultracentrifuge at 118,000
g, 2h30min, 4  C. 11. Discard supernatant and add appropriate volume of 0.1
sodium phosphate buffer (typically about 1 mL for every 50 grams of processed leaf tissue) to pellet and allow to resuspend on a shaker located in a 4–8  C cold room overnight, then finish resuspending by pipetting up and down.

Custom RNA Packaging into VLPs

109

12. Clarify twice in microcentrifuge at 16,000
g for 20 min. 13. Depending on downstream application, the VLPs may be treated with micrococcal nuclease (37  C for 20 min) to digest any unpackaged nucleic acid that was carried over during purification. The nuclease can then be deactivated by the addition of EGTA (10 mM final concentration), then both nuclease and EGTA can be removed by buffer exchange against 0.1
sodium phosphate buffer over a centrifugal filter with a 100 kDa molecular weight cutoff (for example, Amicon Ultra-15 Centrifugal Filter Units). 14. VLPs and the RNA packaged within are stable for months at 4–8  C. Sodium azide (to a final concentration of 3 mM) may be added for long-term storage. 15. If desired, the preparation of CPMV VLPs can be analyzed by SDS-PAGE and transmission electron microscopy (see Fig. 2), and the RNA content can be analyzed by denaturing agarose gel electrophoresis, northern blotting or RT-PCR after RNA extraction, which is typically carried out by phenol:chloroform extraction followed by precipitation of RNA by lithium chloride [2]. Yields of 1 mg of purified VLPs per gram of infiltrated leaf tissue can be expected, as quantified by protein quantification (such as a BCA assay).

Fig. 2 CPMV VLPs packaging a custom RNA of choice after purification. Left: an SDS-PAGE gel showing the Large (L) and two electrophoretic forms of the Small (S) coat proteins derived from the VP60 precursor. Middle: denaturing agarose gel showing RNA extracted from VLPs by phenol:chloroform extraction and lithium chloride precipitation. The two visible bands correspond to CPMV RNA-1 and the custom RNA of choice, which in this case was ~1 kb in size. Right: transmission electron micrograph of CPMV VLPs, stained on a carbon-coated grid with 2% (w/v) uranyl acetate. Scale bar is 100 nm

110

Hadrien Peyret and George P. Lomonossoff

This method will yield concentrated and purified VLPs (Fig. 2). However, such a preparation of VLPs will be composed of a mix of three distinct populations of particles: some will be devoid of RNA (empty or eVLPs), some will contain CPMV RNA-1, and some will contain the custom RNA of choice. For most applications this is not a problem, but if necessary these distinct types of VLPs can be separated by ultracentrifugation in a caesium chloride density gradient due to their different buoyant densities caused by varying RNA content.

4

Notes 1. The VP60 coat protein precursor should be expressed from an expression cassette that directs high protein yield, but which cannot be replicated by the CPMV RNA-1–encoded replication machinery. The pHREAC overexpression vector is ideal for this (and is available on Addgene: Plasmid #134908), but alternative overexpression vectors would also work. If VP60 is expressed from an expression cassette that can be replicated by the CPMV replication machinery, the VP60 transcript will be packaged in the resulting VLPs. 2. In contrast to VP60, the custom sequence of choice must be expressed from an expression cassette that can be replicated by the CPMV replication machinery, because RNA packaging requires viral replication [2]. This requires the sequence to be placed between the wild-type 50 and 30 UTRs of CPMV RNA-2. 3. Resuspension to OD600 of 0.6 usually requires 5–10 volumes of MMA to that of the starting culture. Although each culture tends to grow at a different rate, allowing cultures to grow to stationary phase generally ensures that all cultures have similar densities. 4. Some protocols for the purification of CPMV include a chloroform:butanol step before the PEG precipitation. While this does provide greater purity of virions containing wt RNA-1 and RNA-2, it drastically reduces yield of particles that contain no RNA (eVLPs) or small (60
C. 1. Fill the 20-L working volume water bath with heat precipitation buffer. Connect one piece of silicone tubing to the pump inlet and another to the outlet. Place the loose ends of the tubing into the water bath. 2. Set the water bath temperature to 65
C. Activate the pump with a volumetric flow rate of 5 L/min. Allow the water bath to reach thermal equilibrium for 20 min and confirm that the temperature is 65  1
C. Adjust the water bath temperature set point if necessary and allow the system to reach thermal equilibrium as before. 3. Place 0.1 kg of leaf biomass in a 230  230  230 mm autoclave basket, avoiding compression of the leaf material or other damage, and close the lid of the basket. Place the basket into the water bath using appropriate personal protective equipment (see Notes 2 and 3).

132

Johannes F. Buyel

4. Incubate the leaves for 2 min in the heat precipitation buffer. Monitor the liquid temperature throughout the incubation period. Increase the incubation time by 1 min for every
C temperature drop that may occur after the basket of leaves has been placed in the water bath. 5. Remove the basket from the water bath and let residual liquid drain from the leaves for 1 min in a sink. Then manually transfer the leaves to a blender for immediate extraction (see Subheading 3.2, see Notes 4 and 5). 6. Repeat steps 3–5 with fresh leaf material until all harvested biomass has been processed. 3.2 Extraction and Clarification 3.2.1 Extraction Using a Blender

This procedure is used to release target proteins from leaves into a suitable buffer for stabilization. The method can be applied to most protein products. 1. Make sure the blender is disconnected from its power source. Do not work in the blender bucket while it is mounted on the blender motor. Place 0.1 kg (wet mass) of harvested or heattreated (see Subheading 3.1) leaf material in the blender and add 0.3 L of extraction buffer to achieve a 1:3 biomass–buffer ratio (see Notes 6 and 7). 2. Mount the blender bucket on the motor, connect the blender to a power source and homogenize the leaves for 3  30 s with 30 s breaks (see Notes 8 and 9; Fig. 2). Remove a 1-mL sample for analysis (see Note 10). 3. An optional flocculation step can be interspersed at this stage (see Subheading 3.3.1).

Fig. 2 Extraction operations. (a) Blender-based homogenization unit consisting of motor (1) and beaker (2). (b) Scale-down screw press with inlet (1), press section (2) as well as collection vessels for liquids (3) and solids (4)

Extraction and Purification

133

4. Repeat steps 1 and 2 until all leaf biomass has been processed. 5. Collect and combine all extracts in a single 5-L beaker (see Note 11). Adjust the pH to 8.0 using 0.4 M trisodium phosphate. Take a 1-mL sample for analysis and monitor the turbidity of the pooled extract by diluting 0.25 mL in 9.75 mL of extraction buffer in a turbidimeter cuvette. 6. If a flocculation unit operation was included (see step 3 and Subheading 3.3.1), skip this step. Otherwise, place the 5-L beaker (see step 5) on a magnetic stir plate and drop a magnetic stirrer into the beaker. Start stirring at ~80 rpm to avoid the formation of sediment or a floating phase of cell debris while minimizing the vortex size. Adjust the stirrer speed if necessary. 7. Connect the tri clamp to 6.4 mm hose adapters onto both ends of the Masterflex 24 tubing and mount the center region of the tubing into the peristaltic pump. Fasten the quick-release lever of the pump to hold the tubing in place. 8. Place a BP-410 bag filter above a fresh 5-L beaker, place the outlet end of the Masterflex 24 tubing in the bag filter and place the feed end of the tubing in the beaker containing the pooled extract (see Notes 12 and 13). 9. Pump the extract to the bag filter at a flow rate of 0.15 L/min. Monitor the turbidity of the filtrate at 2-min intervals by diluting 1-mL aliquots with 9 mL extraction buffer in a turbidimeter cuvette. 10. Once all extract has been filtered, mix the pooled bag filtrate in the 5-L beaker, determine the pool turbidity (see step 9) and take a 1-mL sample for analysis. 11. Place the 5-L beaker with the bag filtrate on a magnetic stir plate, add a magnetic stirrer and agitate the filtrate at ~50 rpm. 12. An optional filter additive step can be introduced at this stage if flocculation was carried out in step 3 (see Subheading 3.3.2). 13. In the peristaltic pump, replace the 24 tubing with the 16 tubing. Mount a tri clamp to 3.1 mm hose adapter onto the outlet of the 16 tubing (see Notes 12 and 14). Place the free tubing end in a beaker containing 0.2 L fresh extraction buffer (see Note 15). 14. Assemble the Velapad filter housing according to the manufacturer’s instructions. Stack the filter layers so that KS50P is next to the perforated metal plate. Make sure that the smooth surface of both filter layers is facing toward the metal plate (see Note 16; Fig. 3). 15. Connect the inlet (at the top) of the filter housing to a T-connector in an upright manner using appropriate tri clamps and gaskets. Connect the tri clamp adapter for the 16 tubing to the free T-connector outlet facing in the horizontal direction.

134

Johannes F. Buyel

Fig. 3 Filtration apparatus. (a) The bag filtration setup consists of the plant extract to be processed (here flocculated with Polymin P and therefore not agitated) (1), tubing (2) and a peristaltic pump (3), as well as the bag filter (4) and a beaker to collect the bag filtrate (5). (b) The depth filtration setup consists of a stir plate (1), the bag filtrate to be processed (2), tubing (3), and a peristaltic pump (4) to feed the filter assembly composed of a manometer (5) and the actual housing (6) from which the depth filtrate is collected in a beaker (7) that is monitored for mass and turbidity using a scale (8) and turbidimeter (9), respectively

16. Connect a tri clamp 50 ¾00 to 25 ¼00 adapter onto the free T-connector end facing in the vertical direction. Open the filter housing vent valve and start the pump at maximum speed. 17. Close the vent valve once liquid starts to leak from it. Reduce the pump speed to 30 mL/min. 18. Once liquid starts to leak from the top of the tri clamp 50 ¾00 to 25 ¼00 adapter, reduce the flow rate to 12 mL/min and mount the manometer to the 50 ¾00 end of the adapter. 19. Flush the filter with fresh extraction buffer and stop the pump before air is aspirated into the tubing (see Note 17). 20. Quickly place the free end of the tubing in the beaker with the bag filtrate, place a fresh 5-L beaker under the bottom outlet of the filter housing (see Note 18) and continue pumping at 12 mL/min.

Extraction and Purification

135

21. At 2-min intervals, monitor the filter housing pressure on the manometer and the turbidity of the depth filtrate by collecting 10 mL filtrate at the housing outlet in a turbidimeter cuvette for direct measurement. 22. Continue pumping until no liquid drains from the filter housing. If the pressure reaches 0.2 MPa (2.0 barg) before the feed is empty, stop pumping, place the filter housing assembly over the bag filtrate beaker, open the vent valve and drain the liquid inside the filter housing back into the bag filtrate beaker. Then cycle through steps 14–22 until all bag filtrate is processed. 23. Mix the pooled depth filtrate, determine its turbidity (see step 21) and take a 1-mL sample for analysis. 24. Place the 5-L beaker with the depth filtrate on a magnetic stir plate, add a magnetic stirrer and agitate the depth filtrate at ~50 rpm. 25. Mount a fresh tri clamp 25 ¼00 to 3.1 mm hose adapter onto one end of a piece of 16 tubing and connect it to the inlet of a Sartopore 2150 capsule filter. Mount the central part of the tubing in the peristaltic pump and place the filter outlet above a fresh 5-L beaker. 26. Place the loose end of the tubing in the 5-L beaker containing the depth filtrate (see Note 15). Open the vent valve of the filter and start pumping at 0.1 L/min. Close the valve once liquid starts to leak from it. 27. Continue pumping until all depth filtrate has passed through the membrane filter (see Note 19). Mix the pooled membrane (sterile) filtrate, determine the turbidity using an undiluted 10-mL aliquot and take a 1-mL sample for analysis. 28. An optional ultrafiltration/diafiltration step can be introduced at this stage if necessary (see Subheading 3.3.3). 29. The conditioned, clarified extract can now be processed by chromatography (see Subheading 3.4). 3.2.2 Extraction Using a Screw Press (Alternative)

This procedure can be used as an alternative to extraction using a blender (Subheading 3.2.1) if the products can withstand low pH (below 6.0). The benefit of this method is that the extract volume is smaller and the dispersed solids are less abundant compared to the blender-based method. 1. Assemble a screw press juicer according to the manufacturer’s instructions. Select a power setting of 150 W and 80 rpm (Fig. 2). 2. Continuously feed individual tobacco leaves into the screw press inlet with a speed of about six leaves per minute (see Note 20).

136

Johannes F. Buyel

3. Take a 1-mL sample from the pooled liquid fraction for analysis (see Note 21). 4. Continue from step 13 under Subheading 3.2.1 (see Note 22). 3.3 Enhanced Clarification and Conditioning 3.3.1 Flocculation of Dispersed Particles in Plant Extracts (Optional)

This procedure increases the effective diameter of dispersed particles in plant extracts making it easier to remove them, or even allowing removal during an earlier process step (e.g., bag instead of depth filtration), which greatly increases filter capacity and thus reduces the cost of consumables (Fig. 4a). 1. Quickly add 20 mL of flocculant solution to 0.4 L of the extract from Subheading 3.2.1 step 4 and mix for 5 s in the blender (see Note 23). 2. Carefully transfer the flocculating extract into a fresh 5-L beaker. Continue extracting as described under Subheading 3.2.1 steps 1 and 2 while flocculating as described under Subheading 3.3.3 step 1 until all leaf biomass has been processed. Whenever transferring the flocculating extract from the blender to the 5-L collection beaker, avoid splashing or otherwise agitating the liquid already in the beaker (Fig. 3). 3. After the final transfer of extract to the beaker, incubate the flocculating extract for an additional 20 min without agitation and, at the end of the incubation period, take a 1-mL sample from the liquid fraction for analysis using a long serological pipette. 4. At this stage, continue with bag filtration as described under Subheading 3.2.1 step 7.

3.3.2 Filter Additives (Conditional)

This procedure should only be carried out in combination with flocculation. It increases the capacity of depth filters further, thus reducing consumables costs. 1. If flocculation (see Subheading 3.3.1) was carried out, add 10 g of CelluFluxx F15 to 1.0 L of bag filtrate from Subheading 3.2.1 step 11 while increasing the agitation rate from 50 to 200 rpm (see Notes 24 and 25). 2. Continue with depth filtration as described under Subheading 3.2.1 step 13.

3.3.3 Ultrafiltration/ Diafiltration (Optional)

This procedure reduces the volume to be processed in subsequent steps, thereby accelerating the process and potentially reducing investment costs because smaller devices will be sufficient. It can also be used to condition the product into a buffer that is suitable for further purification. 1. Mount the Hydrosart membrane cassette to the Slice 200 holder according to the manufacturer’s instructions (Fig. 4b). Connect the flow restriction throttles to the lower permeate and the retentate ports of the holder.

Extraction and Purification

137

Fig. 4 Improved processing techniques. (a) Effect of flocculants on plant extract clarification showing raw extract (1), flocculant addition (2), bag filtrate (3), and depth filtrate (4) for a setup without (top) and with (bottom) the flocculant Polymin P. Floating (1), liquid (2), and sediment (3) phases will form after 5–15 min. (b) Ultrafiltration/diafiltration setup consisting of membrane filtrate (1, feed), tubing (2), pump control unit (3), membrane holder with cassette (4), collection beaker (5), and scale (6)

2. Connect the pressure sensors to the inlet and retentate ports. Mount the permeate pressure sensor to the lower permeate port and seal the upper port with a Luer lock dead-end adapter. 3. Connect the data wires of the pressure sensors to the corresponding sockets of the pump control unit. Connect the scale to the control unit. 4. Place a beaker containing 0.2 L of extraction buffer on the scale and position the loose ends of the tubing connected to the inlet and retentate ports in the beaker. Place the loose end of the tubing connected to the permeate port into a sink. 5. Set the pump flow rate to 0.1 L/min and start the pump. Leaving the retentate throttle fully open, regulate the permeate throttle to adjust the transmembrane pressure (TMP) to 0.05 MPa (0.5 bar) according to the observed pressure values and Eq. 1. TMP ¼

pinlet þ pretentate  ppermeate 2

ð1Þ

6. Stop the pump when the buffer is depleted and before air is aspirated into the inlet tubing. Place the loose ends of the inlet and retentate tubing into the 5-L beaker containing the membrane filtrate (see Subheading 3.2.1 step 27). Place the loose end of the permeate tubing into a clean 5-L beaker and position this beaker on the scale. 7. Restart the pump. Adjust the permeate throttle as required to maintain a TMP of 0.05 MPa during operation (see Note 26).

138

Johannes F. Buyel

8. Record the inlet, retentate and permeate pressures as well as the permeate beaker mass at 1-min intervals (see Note 27). 9. Stop the pump once the retentate volume falls to 0.2 L corresponding to a ~tenfold concentration and take a 1-mL sample for analysis (see Notes 28 and 29). 3.4

Chromatography

This procedure is used to remove process and product related impurities. ¨ KTA pure system 1. Connect the inlet tubing of pump A of the A to a 1-L glass flask containing equilibration buffer. Connect the inlet tubing of pump B to a 0.5-L flask containing elution buffer. ¨ KTA 2. Start a new run in the Unicorn control software of the A chromatography device. Set the monitoring wavelengths to 220, 260 and 280 nm (see Note 30). 3. Set the pressure alarm to 0.8 MPa and use the pump flush command to equilibrate pump B with elution buffer and then pump A with equilibration buffer. Place the active system outlet tubing in a fresh 5-L beaker. 4. Set the flow rate to 0.5 mL/min and disconnect the tubing shortcut between the column inlet valve and column outlet valve at the column outlet valve. Remove the dead-end plug from the column inlet. 5. While holding the column in one hand and the shortcut tubing in the other, use the buffer dripping from the shortcut tubing to fill the now-open column inlet with liquid. Once fully covered with liquid, connect the loose end of the shortcut tubing to the column inlet. 6. Quickly remove the dead-end plug from the column outlet tubing. Use the buffer dripping from the column outlet tubing to fill the connection hose of the column outlet valve with liquid, then connect the column outlet tubing to the column outlet valve. 7. Set the flow path in the Unicorn software to the appropriate column valve position. Increase the flow rate to 2.5 mL/min and equilibrate the column with 10 column volumes (CV) of equilibration buffer. Execute the autozero command on all UV signals. 8. Stop the flow and carefully transfer the pump A inlet tubing from the equilibration buffer flask into the beaker containing the clarified extract (see Subheading 3.2.1 step 27) or concentrate (see Subheading 3.3.3 step 5). Execute a note command in the Unicorn software to indicate the start of the loading phase (see Note 31).

Extraction and Purification

139

9. Set the flow rate to 0.5 mL/min and increase it by 0.5 mL/min every 15 s until it reaches 2.5 mL/min (see Note 32). Continue to load the extract and stop the flow before air is aspirated into the tubing of pump A. 10. Mix the flow-through fraction collected in the 5-L beaker connected to the system outlet tubing and remove a 1-mL sample for analysis. Place the outlet into a fresh 5-L beaker. 11. Briefly rinse the outside of the pump A tubing with equilibration buffer and carefully place it back into the equilibration buffer flask. Set the flow rate to 0.5 mL/min and increase it by 0.5 mL/min every 15 s until it reaches 2.5 mL/ min. 12. Wash the column with 10 CV of equilibration buffer. Then, mix the wash fraction collected in the 5-L beaker connected to the system outlet tubing and take a 1-mL sample for analysis. Place the outlet into a fresh 5-L beaker. 13. Set the percentage of buffer B from 0% to 100%. Once the UV signal at 280 nm increases to more than 30 mAU, place the system outlet tubing into a fresh 0.5-L flask and collect the elution peak. 14. Once the UV signal at 280 nm falls below 75 mAU, place the system outlet tubing into a fresh 0.5-L flask. Mix the liquid in the first 0.5-L flask and adjust the pH in that flask to 7.5 using 0.4 M trisodium phosphate (see Note 33). Take a 1-mL sample for analysis. 15. Store the first 0.5-L flask containing the antibody-rich elution peak at 4
C for immediate use or freeze at 80
C (see Note 34). 16. Continue the chromatography for an additional 5 CV, then stop the flow and replace the equilibration buffer and elution buffer flasks with flasks containing deionized water. Flush pumps A and B. 17. Resume a flow rate of 2.5 mL/min and equilibrate the column in deionized water for 10 CV (see Note 35).

4

Notes 1. Nicotiana benthamiana leaf material and stems can also be used, but if this is from a transient expression experiment it may contain viable bacteria and additional safety precautions that are not described here may be required. 2. Place a stainless steel block on top of the basket to prevent flotation. 3. Leaves can be placed into the precipitation buffer without a basket, which can increase biomass throughput. However, such

140

Johannes F. Buyel

a free flotation approach reduces control over heat exposure time and leaf damage and can thus result in higher batch-tobatch variability or product loss. 4. Plants can be stored for extended periods of time after blanching and before the start of the extraction (e.g., more than 30 min on ice or several weeks frozen at 20
C has been successful [16]). However, product stability may decline with increasing storage times and thus immediate processing is recommended. 5. Heat treatment is typically carried out by the submersion of intact leaves in a heat precipitation buffer, but heat treatment can be implemented at other stages of the extraction and purification process as discussed in recent publications [13, 16–18]. However, later heat treatment steps attract substantially higher costs in terms of equipment and consumables, for example because additional filtration steps may be necessary to remove precipitated host cell proteins. 6. The extraction buffer to biomass ratio as well as the buffer composition and pH can be adapted for the needs of a specific target protein, type of biomass and/or for the intended chromatographic purification process (see Subheading 3.4). For example, a reduced salt concentration can facilitate immediate loading onto ion exchange resins once the extract is clarified [19], whereas a pH below 5.0 can help to precipitate (additional) host cell proteins [12]. A design of experiments approach can be useful to optimize the extraction conditions including the buffer (see Chapter 18). 7. Alternative blender formats are available from other manufacturers and can allow larger quantities of leaf biomass to be processed per run. Using a custom blending device as previously described [20] can be beneficial because it mimics the extraction conditions of pilot-scale or production-scale facilities if designed accordingly and will therefore facilitate process scale-up. 8. Ensure that the leaves do not clog the blender or stick to the walls. If necessary, stop the blender, disconnect it from the power source, and then loosen any leaves blocking the blade. Use a spatula to push down leaves that stick to the blender bucket wall above the liquid level. Close the blender bucket, reconnect the blender to the power source, and then continue the homogenization. 9. The duration of the extraction step can be adjusted if the biomass is a nonleaf material, obtained from another plant species or if another blender is used (see Note 7). However, homogenization times >3 min were not found to enhance the

Extraction and Purification

141

release of an antibody retained in the endoplasmic reticulum or a fluorescent protein targeted to the chloroplasts in tobacco leaves [11]. 10. The analysis carried out on these samples is product specific. Typically, such analysis includes absolute protein quantification using Bradford reagent or stained polyacrylamide gels [21], specific quantitation by surface plasmon resonance spectroscopy, ELISA and western blotting, or purity analysis by analytical size exclusion chromatography [22]. 11. The beaker may be placed into a bucket of ice-cold water if the product is sensitive to degradation or prone to adverse reactions. This modification can be applied to all steps involving storage in beakers. Similarly, cooled extraction buffer and vessels can be used to reduce the temperature of the extract, filtrates and elution fractions. 12. The direction of flow can be adjusted on the pump. 13. The bag and depth filtration steps may be replaced by a single precoat filtration step as previously described [23]. 14. Two peristaltic pumps can be used to reduce handling effort. 15. A metal tube of appropriate diameter can be mounted to the free end of the 16 tubing to increase mass and thus avoid flotation in the extraction buffer and bag filtrate. The volume of fresh extraction buffer volume in mL should be ~10 times the filter area in cm2. 16. Other filter layers may be selected based on the particle size distribution in the bag filtrate as described [11]. If flocculants are used, the screening of other filter layers is recommended because polymer addition shifts the particle size distribution toward larger diameter. 17. The filter can be drained of extraction buffer before loading the bag filtrate to avoid sample dilution by pumping until no liquid drains from the bottom outlet of the filter. Then, the manometer must be disconnected from the assembly, the vent valve must be opened and the filter housing must be filled with bag filtrate as described for flushing the filter with fresh extraction buffer. 18. The initial filtrate will mostly be liquid from flushing the filter with extraction buffer and this clear and colorless liquid can be discarded. The plant filtrate will typically have a slightly brown tint indicating when filtrate collection should begin. 19. The Sartopore 2 filter will be sufficient for most applications processing ~0.5 kg of biomass. The safety of operation can be increased if a manometer is mounted to the membrane filter using a T-connector as described for the depth filter (see Subheading 3.2.1 steps 15–18).

142

Johannes F. Buyel

20. The feeding speed of the screw press can be adjusted if other material is used, such as N. benthamiana leaves, and should be about ~200 g biomass/min. 21. The solids can be reprocessed using a blade-based homogenizer (see Subheading 3.2.1) to assess the fraction of product that remains in the residual plant biomass. 22. Flocculation (see Subheading 3.3.1) can also be applied to the screw press extract. However, an additional device (similar to a blender) is required to ensure the brief but rapid mixing of the extract following the addition of the flocculant solution. Furthermore, the type and concentration of the flocculant may require additional optimization, especially if material other than tobacco leaves is used. 23. The concentration and type of flocculant should be adjusted if an extraction buffer other than the one described here and/or plant material other than tobacco leaves is processed. A detailed analysis of flocculants and corresponding optimization procedures can be found elsewhere [24]. 24. It is not necessary to use filter aids as well as flocculants, but filter aids can increase depth filter capacity compared to a setup using flocculants alone whereas using filter aids without flocculants often has minimal benefits in terms of increased filter capacity [25]. 25. The type and concentration of filter additives should be adjusted if feeds are prepared with biomass other than tobacco leaves, or if different flocculation conditions are used. Variations have been described in the literature [14]. 26. Running the ultrafiltration/diafiltration step for more than 20 min can result in a temperature increase in the retentate due to the input of mechanical energy. Cooling the retentate beaker can help to maintain the integrity and activity of heatsensitive products. 27. A spreadsheet macro or control system can be used to record pump flow, pressures and permeate mass automatically. 28. Once the retentate volume falls to 0.2 L, a different buffer can be added to the retentate beaker (the one that contained the membrane filtrate) to initiate buffer exchange, for example if a low salt concentration is required for protein capture by ion exchange chromatography. About 3–4 cycles of concentration and buffer addition should be carried out to achieve a close to complete buffer exchange. The operation should yield a concentrated protein solution. 29. If the product has a molecular mass below 30 kDa, an ultrafiltration/diafiltration membrane with a cutoff of 100 kDa can be used instead. In this case, the product will be found in the

Extraction and Purification

143

permeate instead of the retentate and a single diafiltration step should be applied to improve product recovery (see Note 28). Furthermore, an additional ultrafiltration/diafiltration step using a 5 kDa cutoff membrane can be applied afterward to concentrate the product-containing permeate, especially if extensive diafiltration was used to improve product recovery. 30. The 220-nm channel is generally suitable for proteins but a different channel may be used if the product has a specific absorption maximum at a different wavelength, for example if the product is a fluorescent protein. ¨ KTA pure system, a 31. Depending on the configuration of the A sample valve can be used to avoid the need to place the pump A tubing in the feed beaker and the manual setting a load mark. 32. Increasing the CV will allow faster loading at the expense of higher costs for columns and resins, especially when affinity resins are used. 33. Concentrated phosphate buffer (pH 8.0) can be preadded to the 0.5-L product peak collection flask for immediate pH neutralization. The volume and molarity of the neutralization buffer should be adjusted according to the elution pH and CV. A good starting point is a buffer volume of 1 CV and a molarity of twice the elution buffer concentration. 34. It may be necessary to add stabilizers such as glycerol to the neutralized antibody eluate before freezing to avoid activity losses during storage. 35. Additional washing and cleaning steps (e.g., using 0.5 M sodium hydroxide) can be included and should be carried out according to the manufacturer’s instructions. The column can be stored long term in 20% v/v ethanol after cleaning.

Acknowledgments I would like to thank Dr. Richard M Twyman for editorial assistance. This work was funded in part by the Fraunhofer-Gesellschaft Internal Programs under Grant No. Attract 125-600164 and the state of North-Rhine-Westphalia under the Leistungszentrum grant no. 423 “Networked, adaptive production.” The author has no conflict of interest to declare. References 1. Fischer R, Buyel JF (2020) Molecular farming – the slope of enlightenment. Biotechnol Adv 40:107519. https://doi.org/10.1016/j.bio techadv.2020.107519 2. Buyel JF (2018) Plant molecular farming – integration and exploitation of side streams to

achieve sustainable biomanufacturing. Front Plant Sci 9:1893. https://doi.org/10.3389/ fpls.2018.01893 3. Klutz S, Magnus J, Lobedann M et al (2015) Developing the biofacility of the future based on continuous processing and single-use

144

Johannes F. Buyel

technology. J Biotechnol 213:120–130. https://doi.org/10.1016/j.jbiotec.2015. 08.388 4. Xu S, Gavin J, Jiang R et al (2016) Bioreactor productivity and media cost comparison for different intensified cell culture processes. Biotechnol Prog 33(4):867–878. https://doi. org/10.1002/btpr.2415 5. Buyel JF, Twyman RM, Fischer R (2015) Extraction and downstream processing of plant-derived recombinant proteins. Biotechnol Adv 33(6 Pt 1):902–913. https://doi. org/10.1016/j.biotechadv.2015.04.010 6. Hassan S, Keshavarz-Moore E, Ma J et al (2014) Breakage of transgenic tobacco roots for monoclonal antibody release in an ultrascale down shearing device. Biotechnol Bioeng 111(1):196–201. https://doi.org/10.1002/ bit.25006 7. Nichols ME, Stanislaus T, Keshavarz-Moore E et al (2002) Disruption of leaves and initial extraction of wildtype CPMV virus as a basis for producing vaccines from plants. J Biotechnol 92(3):229–235 8. Madeira LM, Szeto TH, Ma JK et al (2016) Rhizosecretion improves the production of Cyanovirin-N in Nicotiana tabacum through simplified downstream processing. Biotechnol J 11(7):910–919. https://doi.org/10.1002/ biot.201500371 9. Komarnytsky S, Borisjuk NV, Borisjuk LG et al (2000) Production of recombinant proteins in tobacco guttation fluid. Plant Physiol 124(3): 927–933. https://doi.org/10.1104/Pp.124. 3.927 10. Kingsbury NJ, McDonald KA (2014) Quantitative evaluation of E1 endoglucanase recovery from tobacco leaves using the vacuum infiltration-centrifugation method. Biomed Res Int 2014:483596. https://doi.org/10. 1155/2014/483596 11. Buyel JF, Fischer R (2014) Scale-down models to optimize a filter train for the downstream purification of recombinant pharmaceutical proteins produced in tobacco leaves. Biotechnol J 9(3):415–425. https://doi.org/10. 1002/biot.201300369 12. Hassan S, van Dolleweerd CJ, Ioakeimidis F et al (2008) Considerations for extraction of monoclonal antibodies targeted to different subcellular compartments in transgenic tobacco plants. Plant Biotechnol J 6(7): 733–748. https://doi.org/10.1111/j. 1467-7652.2008.00354.x 13. Buyel JF, Hubbuch J, Fischer R (2016) Comparison of tobacco host cell protein removal methods by blanching intact plants or by heat

treatment of extracts. J Vis Exp (114):e54343. https://doi.org/10.3791/54343 14. Buyel JF, Fischer R (2014) Flocculation increases the efficacy of depth filtration during the downstream processing of recombinant pharmaceutical proteins produced in tobacco. Plant Biotechnol J 12(2):240–252. https:// doi.org/10.1111/pbi.12132 15. Sack M, Rademacher T, Spiegel H et al (2015) From gene to harvest: insights into upstream process development for the GMP production of a monoclonal antibody in transgenic tobacco plants. Plant Biotechnol J 13(8):1094–1105. https://doi.org/10.1111/pbi.12438 16. Buyel JF, Gruchow HM, Boes A et al (2014) Rational design of a host cell protein heat precipitation step simplifies the subsequent purification of recombinant proteins from tobacco. Biochem Eng J 88(15 July 2014):162–170. https://doi.org/10.1016/j.bej.2014.04.015 17. Menzel S, Holland T, Boes A et al (2018) Downstream processing of a plant-derived malaria transmission-blocking vaccine candidate. Protein Expres Purif 152:122–130. https://doi.org/10.1016/j.pep.2018.07.012 18. Menzel S, Holland T, Boes A et al (2016) Optimized blanching reduces the host cell protein content and substantially enhances the recovery and stability of two plant-derived malaria vaccine candidates. Front Plant Sci 7(159):1–15. https://doi.org/10.3389/fpls. 2016.00159 19. Buyel JF, Fischer R (2014) Generic chromatography-based purification strategies accelerate the development of downstream processes for biopharmaceutical proteins produced in plants. Biotechnol J 9(4):566–577. https://doi.org/10.1002/biot.201300548 20. Buyel JF, Fischer R (2015) A juice extractor can simplify the downstream processing of plant-derived biopharmaceutical proteins compared to blade-based homogenizers. Process Biochem 50(5):859–866. https://doi.org/ 10.1016/j.procbio.2015.02.017 21. Bradford MM, Simonian MH, Smith JA (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72(10):248–254 22. Wang T, Lucey JA (2003) Use of multi-angle laser light scattering and size-exclusion chromatography to characterize the molecular weight and types of aggregates present in commercial whey protein products. J Dairy Sci 86(10):3090–3101 23. Buyel JF, Gruchow HM, Fischer R (2015) Depth filters containing diatomite achieve

Extraction and Purification more efficient particle retention than filters solely containing cellulose fibers. Front Plant Sci 6(1134):1–11. https://doi.org/10.3389/ fpls.2015.01134 24. Buyel JF (2016) Procedure to evaluate the efficiency of flocculants for the removal of dispersed particles from plant extracts. J Vis Exp

145

(110):e53940. https://doi.org/10.3791/ 53940 25. Buyel JF, Opdensteinen P, Fischer R (2014) Cellulose-based filter aids increase the capacity of depth filters during the downstream processing of plant-derived biopharmaceutical proteins. Biotechnol J 10(4):584–591. https:// doi.org/10.1002/biot.201400611

Chapter 10 Improving Recombinant Protein Recovery from Plant Tissue Using Heat Precipitation Holger Spiegel Abstract Plants are increasingly viewed as suitable expression hosts for the production of recombinant proteins, especially when oxidative folding and/or posttranslational modification is essential for protein stability and functionality. In contrast to traditional platforms such as yeast and mammalian cells, where the product is secreted into the culture medium, recombinant proteins expressed in plants are usually retained within the cells so additional effort is required during extraction and purification. Various extraction processes are used to release soluble proteins from plant tissues, followed by clarification to remove fibers and particulates before the target protein is purified. Fermentation media generally contain few proteins, making it easier to recover a secreted product, whereas the green juice extracted from plants usually contains a large number of host proteins that interfere with target isolation and purification. In this chapter, we describe the use of heat precipitation to remove a large portion of the host cell proteins, thus improving the efficiency of subsequent purification steps and the quality of the purified recombinant protein. Key words Blanching, Downstream processing, Host cell proteins, Recombinant protein production

1

Introduction The production of recombinant proteins in intact plants by transient expression [1, 2] or stable transformation [3] is an attractive alternative to other heterologous expression systems, especially when a complex recombinant protein requires oxidative folding, posttranslational modification, and/or the assembly of multiple subunits [4]. Transient expression involving the delivery of viral replicons by Agrobacterium tumefaciens [5–7] has achieved significant yield improvements combined with unmatched speed and flexibility. Furthermore, the ability of plants to produce virus-like particles (VLPs) offers new opportunities in vaccine production [8, 9]. However, although plant cells can produce such proteins efficiently [10], and transient expression enables the rapid development and implementation of manufacturing processes, only a few

Stefan Schillberg and Holger Spiegel (eds.), Recombinant Proteins in Plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480, https://doi.org/10.1007/978-1-0716-2241-4_10, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

147

148

Holger Spiegel

commercial processes are currently based on protein expression in intact plants. In contrast to competing platforms based on mammalian cells [11] or microbes such as the bacterium Escherichia coli and the yeast Pichia pastoris [12], vegetative plants do not generally provide an efficient way to secrete the recombinant protein into the culture medium as a convenient starting point for purification. The production of recombinant proteins in intact plants therefore requires the harvesting of plant biomass, which is usually pressed or homogenized for extraction, and must be clarified by centrifugation and/or filtration [13, 14] before the purification of the target protein. In contrast to the culture supernatant of mammalian or microbial cells, the clarified plant extract (green juice) contains a large number of diverse host cell proteins that are released from the cells along with the target protein. Depending on the host plant, the extract may contain various oxidases, such as polyphenol oxidases (PPOs) [15], and proteases [16] that may have a negative impact on target protein quality, stability and functionality [17]. The abundance of host cell proteins complicates the establishment of efficient purification strategies, especially in the absence of product-specific affinity ligands for the rapid capture and enrichment of target proteins in the first purification step. For recombinant proteins with limited stability, it is important to keep downstream process times as short as possible to ensure rapid purification and storage under optimal conditions. Given the issues set out above, methods to deplete a significant portion of host cell proteins and to inactivate or remove harmful proteins like PPOs [18] from the green juice before further purification are highly valuable [19]. A heat precipitation step can remove a large portion of endogenous proteins from the plant extract. Many proteins denature and become insoluble at temperatures exceeding 50  C and can by removed by centrifugation, whereas heat-stable proteins remain intact and soluble. Although the applicability of this approach depends on the unique features of the target protein (e.g., general thermal stability, or number of disulfide bridges), heat precipitation and the removal of endogenous proteins is an efficient way to optimize the purification of recombinant proteins from plant extracts. In this chapter, we describe a workflow that can be used to optimize protein purification by heat precipitation. The procedure can be applied to any heat-stable recombinant protein produced in plants. If possible, it is advantageous to address the heat stability of the target protein during candidate development, especially in situations where there is some flexibility in the design of the protein (e.g., selection of suitable domains or targeted modifications when working on vaccine candidates or diagnostic reagents). In our laboratory, we implemented this workflow in a vaccine development project during the identification of candidate proteins and

Heat Precipitation for Improved Recombinant Protein Recovery

149

antigens by iterative design followed by testing the temperature stability and subsequent functionality of multidomain fusion proteins [20, 21]. The success of heat precipitation as a purification step (in terms of protein yields, quality, and process step optimization) depends on the heat stability of the target protein. An overview of the approach, as well as materials, methods, and helpful strategies, is provided in the following sections.

2

Materials The starting point for the procedures described later in the chapter should be plant biomass (probably leaf tissue) containing the target recombinant protein. The method was developed and tested using Nicotiana benthamiana plants transfected with A. tumefaciens [22] carrying a classic binary expression vector such as pTRA [23]. This transient expression method is described in Chapter 6. Based on our experience of recombinant protein expression in plants using different methods, we are convinced that the procedures in this chapter will also work on other species used for transient expression or stable transformation, especially tobacco (Nicotiana tabacum), although some modifications would be necessary to adjust the volume of extraction buffers and the filters used for fiber removal. Prepare all solutions using deionized water and use them at room temperature (22  C) unless noted otherwise. Strictly adhere to all safety regulations that apply to the chemicals described herein. For example, appropriate personal protective equipment should be worn, including gloves and safety glasses, when handling concentrated acids or bases. Also, follow local waste disposal guidelines and any regulations that apply due to the use of genetically modified plant material (and A. tumefaciens if a transient expression strategy is used). 1. Extraction buffer (PBS, pH 7.4): 8 g/L NaCl, 0.2 g/L KCl, 1.44 g/L Na2HPO4, 0.24 g/L KH2PO4. Adjust to pH 7.4 with HCl. Optional: 1.9 g/L Na2S2O5.

2.1 Buffers, Reagents and Consumables 2.2

Equipment

l

Ceramic mortar and pestle.

l

Commercial blender (e.g., Waring).

l

Commercial juicer (e.g., Angel Juicer).

l

Laboratory centrifuge.

l

Laboratory scale.

l

Heating block.

l

Magnetic stirrer.

l

Temperature-controlled water bath.

l

Gradient PCR machine.

150

3

Holger Spiegel

Methods Here we describe methods that can be used to experimentally assess the feasibility of heat precipitation and implement this procedure as initial step in the purification of a recombinant protein produced in plants. The various steps of the workflow are summarized in Fig. 1.

3.1 Extraction of Total Soluble Proteins

Typically, the first downstream processing step is the extraction of total soluble protein (TSP) from the plant biomass. Depending on the available equipment and considerations regarding future scaleup strategies, the harvested plant material can be disrupted by: (a) grinding using a mortar and a pestle, (b) homogenization using a blender, or (c) extraction of the liquid phase using a juicer. Option (a) is limited to laboratory-scale experiments because it is only suitable for small volumes, and liquid nitrogen is required to achieve efficient extraction. Option (b) is applicable to almost all scales exceeding 10–50 g of biomass given the commercial availability of blending devices that range from 100 mL to 100 L in capacity (and even larger devices are available as custom solutions). Unlike (a) and (b), option (c) compresses rather than homogenizes the tissue and therefore results in a crude extract with a much lower proportion of fibers and particulate matter, and there is no stringent need to use buffers during the extraction process.

Fig. 1 Workflow including a heat precipitation step to deplete host cell proteins. Transgenic plants (a) or plants used for transient expression by infiltration with Agrobacterium tumefaciens (b) are harvested (c) and the leaves can be immediately immersed in a preheated buffer (d) in a procedure known as blanching. Alternatively, extracts can be prepared by grinding (e), juicing (f), or blending (g), and the extracts are then heated under different temperature and buffer conditions (h). The blanched leaves from step (d) are also extracted by blending (g). The extracts are then analyzed by SDS-PAGE (i), immunoblot (j), and eventually by SPR spectroscopy (k)

Heat Precipitation for Improved Recombinant Protein Recovery

151

The process steps for the three different methods are set out below. Special care is required when working with plant material that has been infiltrated with A. tumefaciens because this may contain viable bacterial cells that must be handled and inactivated according to local GMO safety regulations. 1. Grinding. (a) Remove middle veins from leaves. (b) Transfer up to 10 g of leaf material into a ceramic mortar (see Note 1). (c) Shock-freeze the material with 50 mL liquid nitrogen (see Note 2). Wait for the liquid nitrogen to evaporate before proceeding with homogenization. (d) Use the pestle to grind the leaf material to a fine powder. (e) Add 2–3 volumes (relative to leaf weight) of extraction buffer to the powder and extract by further grinding until it forms a homogeneous green juice. (f) Proceed with heat (Subheading 3.3).

precipitation

experiments

(g) Optional: Centrifuge the green juice at >10,000  g for 10 min to remove insoluble fibers and particles. 2. Blending. (a) Transfer leaf material to a blender (see Note 3). (b) Add 2–3 volumes (relative to leaf weight) of precooled (4–8  C) extraction buffer. (c) Blend at highest power setting for 5 min. (d) Proceed with heat (Subheading 3.2).

precipitation

experiments

(e) Optional: Centrifuge the green juice at >10,000  g for 10 min to remove insoluble fibers and particles. 3. Juicing. (a) Transfer 2–3 volumes (relative to leaf weight) of precooled (4–8  C) extraction buffer into the outlet vessel of the juicer. (b) Run plant material through the juicing equipment (see Note 4). (c) Proceed with heat (Subheading 3.2).

precipitation

experiments

(d) Optional: Centrifuge the green juice at >10,000  g for 10 min to remove insoluble fibers and particles. 3.2 Heat Precipitation

Heat precipitation is an efficient strategy to remove host cell proteins from plant extracts but its applicability depends on the identification of conditions under which the target protein remains more

152

Holger Spiegel

stable than most of the host cell proteins, and therefore stays in solution while the host cell proteins denature and precipitate. There is no reliable way to predict the thermal stability of a protein, so this must be determined empirically (see Note 5). Using heating blocks or gradient thermocyclers, it is possible to perform extractions at different temperatures, optionally also varying other conditions such as the buffer pH (see Note 6), to assess the yield and functionality of the target protein and the success of host cell protein depletion. This screening procedure can be carried out using the following steps: 1. Prepare aliquots of extracts, either centrifuged or crude (see Subheading 3.1). The volume of the extract must match the incubation device (e.g., 1.5-mL tubes or 0.2-mL PCR vials). 2. Incubate for 5–10 min at different preset temperatures in the range 40–80  C using steps of 5  C or minimally 10  C. 3. Stop the heating by transferring the vials to ice and incubate them for 5 min before proceeding. 4. Centrifuge samples at >10,000  g for 10 min to remove fibers and particles as well as precipitated endogenous plant cell proteins. 5. Collect the supernatant for analysis (Subheading 3.4). The samples should be analyzed to determine optimal conditions that precipitate most of the endogenous plant proteins while keeping as much of the soluble recombinant protein in the supernatant as possible. Once identified, it is advisable to confirm the optimal temperature in further experiments and also to determine the influence of experimental deviations (temperature, incubation time, variations in the buffer components) on the outcome. 3.3 Blanching of Intact Plant Tissues

As an alternative to heating the plant extract, it is also possible to apply the heat precipitation step to intact plant tissue (typically leaves) in a process known as blanching. Harvested leaves are incubated in a prewarmed bath of buffer to facilitate the precipitation of host cell proteins within the intact issue. Although this is more technically demanding than the heat precipitation of extracts, the advantage of this method is that some proteins that negatively affect protein quality are inactivated early in the process, before any contact with the recombinant protein. Blanching is more difficult at the screening scale, so this procedure should be considered during the development of larger-scale processes, especially in cases where PPOs and proteases are known to have a negative impact on the protein of interest. Blanching results in wet leaf biomass that cannot easily be extracted by grinding or pressing (procedures (a) and (c) in Subheading 3.1), so extraction using a blender (b) is highly recommended.

Heat Precipitation for Improved Recombinant Protein Recovery

153

1. Prepare a vessel containing 1–5 L of prewarmed buffer (see Note 7) including a magnetic stir bar. 2. Transfer harvested leaves (typically 50–100 g) into the prewarmed buffer. 3. Incubate the leaves for 5–10 min (see Note 8) in the prewarmed buffer while stirring. 4. Remove the biomass from the buffer vessel, and measure the weight gain caused by buffer carryover. 5. Add 2–3 volumes (relative to initial leaf weight) of extraction buffer minus the weight gain determined in the previous step. 6. Proceed with extraction in a blender (method (b) in Subheading 3.1). The resulting extract can be used directly for sample analysis (see Note 9 and Subheading 3.4). 3.4

Sample Analysis

Several methods (see Note 10) can be used to determine how much endogenous plant cell protein has been removed from the extract by heat incubation. A rapid way to estimate the depletion of host cell proteins is the measurement of TSP using a Bradford assay [24]. Although this method does not differentiate between the host cell proteins and protein of interest, it will show how temperature affects the protein content under the chosen buffer conditions. The samples should also be fractionated by SDS-PAGE followed by gel staining with Coomassie Brilliant Blue to gauge the depletion of host cell proteins relative to the protein of interest [25], and immunoblotting [26] using antibodies specific for the target protein to determine how much of the target protein is lost. Figure 2 shows an example of temperature-dependent protein recovery. At 40  C, only 50% of the TSP remains in the sample. At 70  C, 90% of the initial TSP content has been removed, while the recombinant target protein remains soluble with minimal losses. Following the acquisition of quantitative data for TSP and the target protein, such a plot can be used to determine a temperature window that allows the removal of TSP while retaining as much of the target protein as possible. In the example shown in Fig. 2, 70  C would be the temperature of choice but higher or lower temperatures may be suitable for other target proteins with different properties. For heat-stable target proteins, > 90% of TSP can be removed without significant loss of the target protein [21]. For more labile target proteins, the temperature that maximizes the difference between TSP and target protein recovery should be considered as an appropriate starting point for further process optimization. Importantly, neither TSP measurement nor the visualization of proteins separated by SDS-PAGE provides information regarding the functionality and/or proper folding of a given protein. In most

154

Holger Spiegel

Protein recovery(%)

100

B 80

C

60

D 40

A

20 0 20

30

40

50

60

70

80

Incubation temperature (°C) Fig. 2 Temperature-dependent precipitation of endogenous plant proteins (A) and enrichment of target proteins (B–D). Possible scenarios that can be encountered when investigating heat precipitation as a potential component of protein purification. Curve A represents the temperature-dependent precipitation of N. benthamiana TSP after incubation for 10 min. Curves B, C, and D illustrate the behavior of recombinant target proteins with low (D), medium (C) and high (B) temperature stability. The optimal temperature range for the recovery of each target protein while removing most of the endogenous plant proteins is indicated by a box

cases, proper folding is absolutely essential for protein activity and this cannot be confirmed by solubility alone. Accordingly, before heat precipitation is considered as process step, protein functionality must be confirmed using a suitable method. For example, an enzyme-linked immunosorbent assay (ELISA) or surface plasmon resonance (SPR) spectroscopy with natural ligands such as target protein receptors or, if available, conformation-dependent monoclonal antibodies [27]. A large number of disulfide bonds often indicates higher thermostability, but it should be noted that reducing agents such as meta-bisulfide, which some protocols include in the extraction buffer to act as an antioxidant, can disrupt disulfide bonds during incubation at higher temperatures and may abolish the functionality of certain proteins [21]. 3.5 Scale-up Considerations

For production processes that deal with larger amounts of biomass, blanching can be performed in larger vessels or temperaturecontrolled tanks containing prewarmed buffer [20]. Heat precipitation of green juice can be carried out in tanks or customized pipe coil heat exchangers (see Note 11). The incubation times should be adjusted according to the mixing performance of the chosen system.

Heat Precipitation for Improved Recombinant Protein Recovery

4

155

Notes 1. Adjust the amount of plant material to suit the mortar. As a rule of thumb, never use more plant material than one quarter of the estimated mortar capacity, to ensure safe and efficient grinding. 2. Always use protective eyewear and take extra care when handling liquid nitrogen. 3. Commercial blenders come in a range of volumes and materials. When working with material from transient expression experiments, it is important that all parts of the blender can be properly decontaminated by cleaning and/or autoclaving to ensure the inactivation of all GMOs. 4. There are several different types of commercial juicers. When working with material from transient expression experiments, it is important that all parts of the juicer can be properly decontaminated by cleaning and/or autoclaving to ensure the inactivation of all GMOs. 5. In our experience, proteins with a high content of cysteine residues and disulfide bridges tend to be more stable at high temperatures. 6. The pH may have a strong influence on target protein solubility. If the target protein precipitates at moderate temperatures (40–60  C), test buffers with pH values 1.5–2 units above and below the theoretical pI, which may make the protein more soluble. 7. Prewarmed buffer (adjusted to the temperature determined in screening experiments) should be used in any blanchingBlanching step. If a temperature-controlled vessel is available, this may help with experimental reproducibility. 8. The incubation time should be held constant within the experiment. The minimum incubation time should be 5 min, and longer incubation times of up to 10 min should be considered for larger volumes. 9. After the heat precipitation step, a centrifugation step is necessary to remove the precipitated host cell proteins. For convenience and process efficiency, the removal of fibers and particulate matter before the heat precipitation step is unnecessary. However, the clarification of crude extracts before heat precipitation may simplify the use of pipe-based heat exchangers when the process is scaled up. 10. The analysis of the samples is product specific. Typically, such analysis includes absolute protein quantification using Bradford reagent [24] or stained polyacrylamide gels, and specific quantitation by SPR spectroscopy, ELISA and immunoblotting.

156

Holger Spiegel

11. When considering the use of pipe-based heat exchangers in a scaled-up heat precipitation step, keep in mind that proteins precipitated during the procedure may clog the pipes. This can be avoided by choosing a system with sufficient pipe diameters.

Acknowledgments We thank Dr. Richard M Twyman for editorial assistance. This work was funded in part by the Fraunhofer Zukunftsstiftung and a Fraunhofer-Gesellschaft Internal Program under Grant No. Attract 125-600164. References 1. Krenek P, Samajova O, Luptovciak I et al (2015) Transient plant transformation mediated by Agrobacterium tumefaciens: principles, methods and applications. Biotechnol Adv 33:1024–1042 2. Fischer R, Vaquero-Martin C, Sack M et al (1999) Towards molecular farming in the future: transient protein expression in plants. Biotechnol Appl Biochem 30(Pt 2):113–116 3. Daniell HH, Streatfield SJS, Wycoff KK (2001) Medical molecular farming: production of antibodies, biopharmaceuticals and edible vaccines in plants. Trends Plant Sci 6:219–226 4. Ma JK-C, Christou P, Chikwamba R et al (2013) Realising the value of plant molecular pharming to benefit the poor in developing countries and emerging economies. Plant Biotechnol J 11:1029–1033 5. Werner S, Breus O, Symonenko Y (2011) High-level recombinant protein expression in transgenic plants by using a double-inducible viral vector. Proc Natl Acad Sci U S A 108: 14061–14066. https://doi.org/10.1073/ pnas.1102928108 6. Musiychuk K, Stephenson N, Bi H et al (2007) A launch vector for the production of vaccine antigens in plants. Influenza Other Respir Viruses 1:19–25 7. Sainsbury FF, Lomonossoff GPG (2008) Extremely high-level and rapid transient protein production in plants without the use of viral replication. Plant Physiol 148:1212–1218 8. Marsian J, Lomonossoff GPG (2016) Molecular pharming-VLPs made in plants. Curr Opin Biotechnol 37:201–206 9. Scotti N, Rybicki EP (2013) Virus-like particles produced in plants as potential vaccines. Expert Rev Vaccines 12:211–224

10. Tschofen M, Knopp D, Hood E (2016) Plant molecular farming: much more than medicines. Annu Rev Anal Chem 12:271–294. https://doi.org/10.1146/annurev-anchem071015-041706 11. O’Flaherty R, Bergin A, Flampouri E et al (2020) Mammalian cell culture for production of recombinant proteins: a review of the critical steps in their biomanufacturing. Biotechnol Adv 43:107552. https://doi.org/10.1016/j. biotechadv.2020.107552 12. Mattanovich D, Branduardi P, Dato L et al (2012) Recombinant protein production in yeasts. Methods Mol Biol 824:329–358. https://doi.org/10.1007/978-1-61779-4339_17 13. Buyel JF, Twyman RM, Fischer R (2015) Extraction and downstream processing of plant-derived recombinant proteins. Biotechnol Adv 33:902–913. https://doi.org/10. 1016/j.biotechadv.2015.04.010 14. Buyel JF, Fischer R (2013) Processing heterogeneous biomass: overcoming the hurdles in model building. Bioengineered 4:21–24. https://doi.org/10.4161/bioe.21671 15. Yoruk R, Marshall MR (2003) Physicochemical properties and function of plant polyphenol oxidase: a review. J Food Biochem 27:361–422 16. Jutras PV, Dodds I, van der Hoorn RAL (2020) Proteases of Nicotiana benthamiana: an emerging battle for molecular farming. Curr Opin Biotechnol 61:60–65. https://doi. org/10.1016/j.copbio.2019.10.006 17. Mandal MK, Ahvari H, Schillberg S et al (2016) Tackling unwanted proteolysis in plant production hosts used for molecular farming. Front Plant Sci 7:267. https://doi.org/10. 3389/fpls.2016.00267

Heat Precipitation for Improved Recombinant Protein Recovery 18. Wang X, Yang L, Liu J et al (2020) Comparison of the biochemical properties and thermal inactivation of polyphenol oxidase from three lily bulb cultivars. J Food Biochem 44(10): e13431. https://doi.org/10.1111/jfbc. 13431 19. Buyel JF, Gruchow HM, Boes A et al (2014) Rational design of a host cell protein heat precipitation step simplifies the subsequent purification of recombinant proteins from tobacco. Biochem Eng J 88:162–170 20. Menzel S, Holland T, Boes A et al (2016) Optimized blanching reduces the host cell protein content and substantially enhances the recovery and stability of two plant-derived malaria vaccine candidates. Front Plant Sci 7: 159. https://doi.org/10.3389/fpls.2016. 00159 21. Beiss V, Spiegel H, Boes A et al (2015) Heatprecipitation allows the efficient purification of a functional plant-derived malaria transmission-blocking vaccine candidate fusion protein. Biotechnol Bioeng 112:1297–1305. https://doi.org/10.1002/bit.25548 22. Kapila J, DeRycke R, VanMontagu M et al (1997) An agrobacterium-mediated transient

157

gene expression system for intact leaves. Plant Sci 122:101–108 23. McLean S, Hunter CN (2009) An enzymecoupled continuous spectrophotometric assay for magnesium protoporphyrin IX methyltransferases. Anal Biochem 394:223–228 24. Bradford MM (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72: 248–254 25. Smith BJ (1984) SDS polyacrylamide gel electrophoresis of proteins. Proteins Methods Mol Biol 1:41–56. https://doi.org/10.1385/089603-062-8 26. Towbin H, Staehelin T, Gordon J (1979) Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Natl Acad Sci U S A 76:4350–4354 27. Jason-Moller ML, Murphy M, Bruno J (2006) Overview of biacore systems and their applications. Curr Protoc Protein Sci Chapter 19:Unit 1 9 . 1 3 . h t t p s : // d o i . o r g / 1 0 . 1 0 0 2 / 0471140864.ps1913s45

Chapter 11 Technoeconomic Modeling and Simulation for Plant-Based Manufacturing of Recombinant Proteins Matthew J. McNulty, Somen Nandi, and Karen A. McDonald Abstract Technoeconomic modeling and simulation is a critical step in defining a manufacturing process for evaluation of commercial viability and to focus experimental process research and development efforts. Technoeconomic analysis (TEA) is increasingly demanded alongside scientific innovation by both public and private funding agencies to maximize efficiency of resource allocation. It is particularly important for plant-based manufacturing, and other nontraditional recombinant protein production platforms, to explicitly demonstrate the manufacturing potential and to identify critical technical and economic challenges through robust technoeconomic analysis. In addition, in silico process modeling and TEA of scaled biomanufacturing facilities allows rapid evaluation of the impacts of process and economic changes on capital expenditures (CAPEX, also sometimes referred to as total capital investment), operational expenditures (OPEX, also known as total manufacturing costs or total production costs), cost of goods sold (COGS, also known as unit production costs), and profitability metrics such as net present value (NPV) and discounted cash flow rate of return (DCROR, also known as internal rate of return or IRR). These models can also be used to assess environmental, health, and safety impact of a designed biomanufacturing facility to evaluate its sustainability and environmental-friendliness. Here we describe a general method for performing technoeconomic modeling and simulation for and environmental assessment of plant-based manufacturing of recombinant proteins. Key words Plant-based recombinant proteins, Plant molecular pharming, Technoeconomic analysis (TEA)

1

Introduction Technoeconomic analysis (TEA) is the evaluation of a manufacturing process on technical viability, economic viability, and environmental impact. Technoeconomic analyses can range from simple back-of-the-envelope calculations completed in several hours to rigorous and highly detailed evaluations performed over the course of months. The cost, accuracy, and scope of the analysis is dictated by the use case of the analysis in the product’s life cycle. Table 1 details the types of design estimates and Fig. 1 details where

Stefan Schillberg and Holger Spiegel (eds.), Recombinant Proteins in Plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480, https://doi.org/10.1007/978-1-0716-2241-4_11, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

159

160

Matthew J. McNulty et al.

Table 1 Types of technoeconomic design estimates, a description of their estimate bases, and the probable accuracy range

Level Type of design estimate

Description

Accuracy range (%)

1

Order-of-magnitude estimate (ratio estimate)

Based on similar previous project data


30

2

Study estimate ( favored estimate)

Based on knowledge of major equipment items


30

3

Preliminary estimate (scope estimate)

Based on sufficient data to permit the estimate to
20 be budgeted

4

Definitive estimate (project control estimate)

Based on almost complete process data

5

Detailed estimate (contractor’s Based on complete engineering drawings, estimate) specification, and site surveys


10
5

Adapted from [1]

Fig. 1 A simple illustration of a plant-made product life cycle and an approximation of where in that life cycle a given level technoeconomic design estimate is executed. An asterisk (*) indicates the level of design estimates that we have identified as the most pressing need in the plant-made product community, and consequently the target design estimates detailed in this method. (Adapted from [2])

in the product life cycle different types of design estimates are generally executed. The most immediate need of the plant-based manufacturing community is to assess, and demonstrate, the commercial potential of lab- or pilot-scale manufacturing process data. Often these analyses are performed prior to extensive process research and development to identify “economic hot spots” and help guide the selection of materials, reagents, and unit procedures and focus experimental and pilot studies. For these reasons and more, the

Techno-Economic Modeling and Simulation

161

use of technoeconomic analyses is well established for commercialization of chemical, petrochemical, and biofuels manufacturing. Similarly, technoeconomic analyses are credited as a driving force in the commercial success of nonpharmaceutical plant-based manufacturing and are expected to play a similar role in pharmaceutical plant-based manufacturing [3]. Per Fig. 1, this is typically met by execution of a Level 2 or Level 3 technoeconomic design estimate. The focus of this chapter is to detail a technoeconomic modeling and simulation method to meet this need with particular focus on plant-made recombinant protein production. Readers can expect to be able to answer questions such as: does the lab-scale manufacturing process demonstrate potential commercial viability? What is the current process bottleneck? How can research and development efforts be effectively allocated to reach commercial feasibility? How do manufacturing costs vary with facility production level, protein expression level, and/or downstream recovery? Which unit procedures and/or cost items contribute the most to the cost of goods sold (COGS)? Which equipment items contribute the most to the total capital investment (also referred to as capital expenditure (CAPEX))? What are the main contributors to environmental, health, and safety (EHS) impact, and how does the impact compare with alternative production schemes?

2

Materials Technoeconomic analyses rely on an often-complex series of mass and energy balances layered with equipment and scheduling constraints, all of which drive economic and profitability calculations. These calculations can be performed manually via spreadsheet or, more commonly, through use of a process simulation tool (PST). A PST is a software with built-in equipment sizing, material and energy balances, operations scheduling, and technoeconomic framework and capabilities. A list of commercially available PSTs commonly used in biomanufacturing is included in Table 2; some companies have developed their own in-house tools.

Table 2 A list of process simulation tools useful for technoeconomic analyses Software

Company

Aspen Plus

Aspen Technology, Inc.

BioSolve

a

Biopharm Services Limited

CHEMCAD SuperPro Designer a

Chemstations, Inc. ®a

Intelligen, Inc.

Indicates the most common tools used in biomanufacturing, based on the authors’ experiences

162

Matthew J. McNulty et al.

The technoeconomic model and simulation method detailed in this chapter is based on the use of SuperPro Designer. A free trial version of SuperPro Designer (http://www.intelligen.com/demo. html) can be used to view and manipulate existing technoeconomic models. The full paid version of SuperPro Designer is required to execute the complete method detailed in this chapter. Lite and Intermediate editions are also available at reduced cost and may be present a viable solution for modeling a restricted number of process steps. 2.1 Technical Considerations of SuperPro Designer

2.2 Plant-Based Manufacturing Limitations with SuperPro Designer

SuperPro Designer is a powerful PST with over 140 unit procedures/operations typically used in biomanufacturing, drag-anddrop construction of process flow diagrams, built-in capabilities for mass and energy balances, equipment sizing and cost models, scheduling and more, supported by an extensive suite of charts and reports on facility performance and economics and a chemical component and mixture database for some commonly used chemicals. This greatly enhances and simplifies the technoeconomic analysis as compared to manual spreadsheet methodology. However, the software structure in turn imposes constraints, some of which are important to note for readers new to technoeconomic analysis and SuperPro Designer. Key considerations and limitations include the following. l

Limited thermophysical property database for biochemical systems making vapor–liquid and liquid–liquid equilibrium calculations more challenging.

l

Mass balances are used. Other units of accounting (e.g., activity units, cell number/density, viral genome copies) must be converted to an equivalent mass.

l

Facility scheduling capabilities are limited, with advanced functions segmented into Intelligen’s complimentary software product, SchedulePro. Without this complimentary software, SuperPro Designer scheduling limitations include that only a single-product facility can be designed, every batch must be identical, every year of operation during the facility lifetime is assumed to be the same (although production level as a percentage of the maximum capacity can be modified each year), and scheduling resets at the end of every year (e.g., partial batches extending past year end do not complete in the following year, and are not counted in costs or production level, instead the recipe scheduling begins anew).

Upstream unit procedures for plant-based manufacturing, from the various methods of plant cultivation (e.g., indoor hydroponic, greenhouse, outdoor field) to plant-specific transfection strategies (e.g., agroinfiltration) to plant harvesting, are not included in

Techno-Economic Modeling and Simulation

163

SuperPro Designer’s built-in unit procedures/operations. However, users can adapt the “Generic Box” unit procedures that SuperPro Designer provides to implement these operations. It is worth noting that SuperPro Designer does allow for integration with other software (e.g., Microsoft Excel, Matlab) through the Component Object Model (COM) library, a feature which provides the information for accessing SuperPro Designer variable-related values. This feature is recommended for advanced users and can be used to plug-in existing model behavior generated in a different software to augment the kinetic modeling options available in SuperPro Designer for a specific unit procedure (e.g., plant cultivation performance dependence on daily temperature and humidity levels). The authors have found that local agricultural extension hubs that deal with agricultural production and economics can serve as valuable resources in populating the plant-specific unit procedure and equipment information. Other useful resources include published literature on agricultural economics and direct contact with regional farms, vertical agriculture companies, and agricultural businesses. Realistic input data for plant-specific unit procedures are critically important for the analysis. A significant shortcoming in the technoeconomics of plant-based manufacturing is the lack of publicly available standard databases for plant-specific unit procedures. Development of such databases would immensely benefit the plantbased manufacturing research community to assess lab- or pilotscale results and translate them into commercial feasibility. 2.3 Technoeconomic and Plant-Based Manufacturing Resources

Plant-based manufacturing is a relatively nascent yet growing area, arguably on the cusp of realizing more mainstream commercial successes. There are currently about ten commercial plant-based production facilities worldwide. Research in plant-based manufacturing to date has spanned a diverse set of production platforms, ranging in production hosts from leafy greens to grain seeds to hairy roots, and in cultivation strategies from transient indoor hydroponics to transgenic outdoor field growth. However, despite this breadth in methodology, there is only a limited number of TEAs, which cover a few of the many plant-based manufacturing methodologies. This chapter aims to provide a more holistic perspective on developing plant-based manufacturing TEAs to address this gap. The published plant-based manufacturing TEAs, as shown in Table 3, represent an indispensable resource for readers new to TEA. Much of the information built into the models described in the publication can be accessed and leveraged for future analyses. In fact, in several cases the models themselves are available for download, viewing, and direct manipulation for readers interested in exploring them further.

164

Matthew J. McNulty et al.

Table 3 A list of recently published plant-based manufacturing technoeconomic analyses spanning multiple industries, target molecules, expression systems, degree of product purity required, and facility production level Production (kg/year)

Source

Transgenic suspension culture

25

Corbin et al. [4]

Antimicrobial protein

Transgenic whole plant

500

McNulty et al. [5]

Pharmaceutical

Griffithsin

Transient whole plant

20

Alam et al. [6]

Pharmaceutical

mAb

Transient whole plant

300

Nandi et al. [7]

Reagent

HRP

Transient whole plant

5

Walwyn et al. [8]

Pharmaceutical Biofuel

BChE Cellulase enzyme

Transient whole plant Transgenic whole plant

25 3  106

Tuse´ et al. [9]

Pharmaceutical

hLF

Transgenic plant seed

600

Nandi et al. [10]

Industry

Target molecule

Expression system

Pharmaceutical

BChE

Food Safety

BChE butyrylcholinesterase, mAb monoclonal antibody, HRP horseradish peroxidase enzyme, hLF human lactoferrin

Additional publications that are recommended to the reader include those related to reporting by commercial-scale plant-based manufacturing companies. This currently includes work published by iBio in 2015 [11], Protalix in 2015 [12], Medicago in 2010 [13], and Ventria in 2005 [10].

3

Method The following method is designed to guide the reader through the process of translating lab- and/or pilot-scale data into a biomanufacturing facility technoeconomic model via project scoping, in silico scale-up, and performance exploration with sensitivity and scenario analysis. Figure 2 displays a high-level graphical representation of the complete method workflow.

3.1

Process Creation

In the following section the reader will be defining high-level project targets, managing project expectations, and getting key stakeholder buy-in. 1. Define project scope. This includes defining the project scope, deliverables, timeline, and key stakeholders (this may involve project management, funding agency point-of-contact, a board of advisors, business strategists, and research collaborators, to name a few).

Techno-Economic Modeling and Simulation

165

Fig. 2 Graphical representation technoeconomic simulation and modeling workflow presented in the method detailed in this chapter. The hashed line rectangle represents a recommended Stage Gate, an important step in project management in which the project progress is reviewed and approved by key stakeholders

2. Define the facility design premises. This includes the regulatory framework within which the product is expected to be governed (see Note 1), the mode of operation (batch, semicontinuous, or continuous), the general manufacturing strategy and facility (greenfield—new construction on undeveloped land, contract manufacturing organization (CMO), and/or expansion of an existing facility), expression system/plant host, the anticipated annual operating factor (hours per year of operating time), the facility location (see Note 2), the production demand (see Note 3), and facility lifetime (see Note 4). 3. Define the product specifications. This includes the required product purity and quality specifications, the desired final product stream composition, and the value of by-products. 4. Develop a process block diagram with select critical process parameter and key performance indicator ranges defined (see Note 5), as shown in Fig. 3. 5. Develop a more detailed process flow sheet with specific raw materials, consumables, and specific unit procedures appropriate to the manufacturing production scale (see Note 6). 3.1.1 Stop Gate I

Here the reader will be assembling a process creation report and communicating with key stakeholders for a Stop Gate before proceeding to process synthesis. This step is critical for setting project expectations. Precise contents and format of the report and/or presentation will depend on the specific project and agreement with key stakeholders. Below are generally recommended steps.

166

Matthew J. McNulty et al.

Fig. 3 An example block flow process diagram with defined general process schematic and facility performance parameters, such as expression level, downstream recovery, final product purity, and annual production capacity. This example is loosely based on the commercial manufacturing process simulated in our recently published technoeconomic analysis of antimicrobial protein production in whole plants for food safety applications [5]. FW fresh weight, MF microfiltration, UF ultrafiltration, CEX cation exchange, DF diafiltration

1. Confirm engagement of key stakeholders and come to an agreement on preferred communication format and frequency. 2. Detail a preliminary project scope based on current knowledge, being as specific as possible to identify any mismatched expectations early on. Include questions on ill-defined and/or unknown aspects of the project scope, prepared in such a way to actively engage the key stakeholders. 3. Detail project roles and responsibilities linked to precise deliverables that are mapped to a project timeline. 3.2 Process Synthesis

In the following section the reader will be developing the “base case” technoeconomic model, iteratively building layers of mass and energy balances (Subheading 3.2.1); labor and scheduling (Subheading 3.2.2); equipment, consumables, and utilities (Subheading 3.2.3), branches and sections (Subheading 3.2.4); economics (Subheading 3.2.5); and environmental, health, and safety

Techno-Economic Modeling and Simulation

167

impact (Subheading 3.2.6), followed by developing a report and initiating review with key stakeholders at Stop Gate II (Subheading 3.2.7). 3.2.1 Mass and Energy Balances

Here the reader will be developing the first layer of the model, the mass and energy balances, and consequently the layout of unit procedures and stream connections in the visual design environment (VDE). The first step in process simulation is generally to select either batch or continuous mode of operation and annual operating time available per year. This method focuses on detailing process modeling and simulation for a facility run in batch mode operation, which is currently most commonly used in plant-made recombinant protein production processes. For reference, see the recent work of Pleitt et al. 2019 for an example of technoeconomic evaluation of continuous downstream biomanufacturing [14]. The reader will be walked through how to register pure components and stock mixtures (solutions commonly used throughout the process such as buffers and cleaning agents). This is followed by populating the process flow sheet with unit procedures, a sequence of actions performed in single equipment item, defining each of those actions, which are referred to as operations, connecting the populated unit procedures with stream lines to indicate flow directionality, and populating both the operations and input streams with registered components and stock mixtures. Process review recommended during mass and energy balance development is primarily to assess stream contents at each at step of the process flow. Two tools to assist in this review: a) the unit procedure activity overview window (VDE > Unit Procedure > Procedure Activity Overview) for a table showing the status of the contents at the start of every operation, and b) enable view of information tags on process streams (VDE > Stream > Style > Edit Style > Display Also > Info Tag checkbox > Info Tag) for a simple view of stream contents including temperature, pressure, and total mass/volumetric flow. 1. Begin process simulation and modeling on SuperPro Designer. Fig. 4 introduces the SuperPro Designer user interface and highlights key terminology referenced in the remainder of the method. 2. Check the Pure Components and Stock Mixtures databanks (Main Ribbon > Databanks > Pure Components OR Stock Mixtures) for required processing stream constituents, including both raw material inputs and reaction products (see Note 7). 3. Register the required Pure Components and Stock Mixtures (Main Ribbon > Tasks > Pure Components OR Stock Mixtures > Register/Edit, View Properties) from existing databanks

168

Matthew J. McNulty et al.

Fig. 4 The SuperPro Designer® user interface annotated with key terms used throughout the methodology detailed in this chapter. The primary user interface includes a series of toolbars and the visual design environment (VDE). A portion of our recently published technoeconomic analysis on semicontinuous bioreactor production of biopharmaceuticals in transgenic rice cell suspension cultures [4] is used for this illustration

or by adding a new item not previously stored in a databank with any processing stream constituents that are missing from the existing set by registering new items (see Notes 8 and 9). 4. Drop-in the first (high upstream) unit procedure defined in the detailed process flow block diagram (Subheading 3.1, step 3) to the VDE (Main Ribbon > Unit Procedures). Use “Connect Mode” (Main Toolbar > Connect Mode) to add stream connection lines to the unit procedure input/output ports on the VDE to define the number, and relationship of, stream inputs/outputs (see Note 10). 5. Initialize the first unit procedure by defining the input/output streams with registered ingredients (VDE > Stream > Registered Ingredients > Add Ingredient), setting total flowrates (VDE > Stream > Total Flowrates) and ingredient mass composition (VDE > Stream > Composition), and populating the unit procedure with an operation sequence (VDE > Unit Procedure > Add/Remove Operations). Only operations compatible with the given unit procedure are available to be added to the sequence list. 6. Initialize each operation in the sequence in turn (VDE > Unit Procedure > Operation Data > Select Operation > Oper. Cond’s). Operation-specific conditions required to perform the mass balances are required at minimum (e.g., for the “Transfer In” operation, the inlet stream must be defined; for the “Stoichiometric Reaction” operation, the reaction sequence must be defined).

Techno-Economic Modeling and Simulation

169

7. Use SuperPro Designer’s “Solve mass & energy balances” button (Main Toolbar > Solve ME Balances) to confirm that the unit procedure has been properly initialized. Any relevant warnings and/or errors associated with the mass and energy balances as they are defined will population in the Error Output window below the VDE. 8. Repeat steps 3–7 for each unit procedure defined in the detailed process flow block diagram (Subheading 3.1, step 3) until the entire process flow diagram has been described in the VDE (see Note 11). 9. Define the main product stream and flow basis for which all material and economic reports are referenced by (Main Ribbon > Tasks > Stream Classification > Main Product/Revenue). 10. Add relevant cleaning operations to unit procedures (see Note 12). 3.2.2 Labor and Scheduling

Here the reader will be defining the labor and scheduling of the model. The reader will be provided with information on how to define process scheduling constraints, which inform the number of times that the full process flow sheet (also referred to as a recipe) is executed per year, followed by how to define the manufacturing operator positions (termed labor types) executing the recipe. Then the operations are populated with specific labor types and labor rates. Lastly, the duration of, and coordination between, operations is defined to determine scheduling and ultimately the total duration of the recipe. Process review recommended during labor and scheduling development is to assess scheduling connections between different unit procedures and operations within each procedure, and operators required to meet labor demands. Two tools to assist in this review: a) the Operations Gantt Chart (Main Toolbar > Charts > Gantt Charts > Equipment GC) to confirm scheduling connections, and b) the Labor Demand chart (Main Toolbar > Charts > Labor) to assess the number of operators required at any given time. 1. Define the high-level process scheduling constraints (e.g., number of batches per year, annual operating time) in the Recipe Scheduling Information window (Main Ribbon > Tasks > Recipe Scheduling Information) (see Note 13). 2. Check the Labor databank (Main Ribbon > Databanks > Labor Types) and populate with any labor types that are missing from the existing set by registering any new items (see Note 14). Labor types that are currently used in the existing process can be viewed alongside their annual demand, for reference (Main Ribbon > Tasks > Other Resources > Labor).

170

Matthew J. McNulty et al.

3. Designate the labor type and labor rate associated with each operation in each unit procedure (VDE > Unit Procedure > Operation Data > Select Operation > Labor, etc. > Labor) (see Note 15). 4. Set the duration of each operation, given by the setup time and process time, in each unit procedure (VDE > Unit Procedure > Operation Data > Select Operation > Oper. Cond’s > Duration) (see Note 16). 5. Set the scheduling of each operation in each unit procedure (VDE > Unit Procedure > Operation Data > Scheduling) (see Note 17). 3.2.3 Equipment, Consumables, and Utilities

Here the reader will be defining the equipment items that house the unit procedures, as well as the specific consumables and utilities used with the equipment during execution of the unit procedure. It is recommended that the equipment items be first defined using “Design Mode,” wherein the equipment size is calculated by SuperPro Designer to be able to meet the throughput requirements of the unit procedure operations and to process the full stream volume. The reader will be walked through how to assign equipment items to unit procedures. This is followed by defining and populating the equipment with consumables, items used by equipment resources for a limited duration (e.g., use hours, recipe cycles) before they must be disposed of and replaced, and then defining and populating the operations with utilities, useful resources required for equipment operation (e.g., power, heat transfer agents). Process review recommended during equipment, consumables and utilities development includes checking equipment-related scheduling and facility power demands. Three tools to assist in this review: (a) the Equipment Occupancy chart (Main Ribbon > Charts > Equipment Occupancy), (b) the Equipment Gantt Chart (Main Ribbon > Charts > Gantt Charts > Equipment GC), and (c) the Power Demand chart (Main Ribbon > Charts > Power). 1. Designate the equipment associated with each unit procedure, and the respective equipment attributes (e.g., material of construction, sizing dimensions, number of units) (VDE > Unit Procedure > Equipment Data > Equipment). By default, SuperPro Designer assigns each unit procedure to a distinct equipment sized according to the mass and energy balance calculations using “Design Mode” (see Notes 18–20). 2. Check the Consumables databanks (Main Ribbon > Databanks > Consumables) and populate with any processing consumables that are missing from the existing set by registering any new items. Consumables that are currently used in the

Techno-Economic Modeling and Simulation

171

existing process can be viewed alongside their annual demand, for reference (Main Ribbon > Tasks > Other Resources > Consumables). 3. Set the consumable type and replacement frequency for each unit procedure (VDE > Unit Procedure > Equipment Data > Consumables) (see Note 21). 4. Check the Power Types databank (Main Ribbon > Databanks > Power Types) and populate with any power types that are missing from the existing set by register any new items (see Note 22). Power types that are currently used in the existing process can be viewed alongside their annual demand, for reference (Main Ribbon > Tasks > Other Resources > Power). 5. Check the Heat Transfer Agents databank (Main Ribbon > Databanks > Heat Transfer Agents) and populate with any heat transfer agent types that are missing from the existing set by registering any new items. Heat transfer agent types that are current used in the existing process can be viewed alongside their annual demand, for reference (Main Ribbon > Tasks > Other Resources > Heat Transfer Agents). 6. Set the power and heat transfer agent types and rate for each operation within each unit procedure (VDE > Unit Procedure > Operation Data > Select Operation > Labor, etc. > Auxiliary Utilities). 3.2.4 Branches and Sections

Here the reader will be defining sections, a grouped set of unit procedures, and branches, a grouped set of sections, to enable future information gathering specifically on these groupings. For many bioprocesses it is useful to at least have an upstream and downstream differentiation (either by branches, if additional granulation/grouping is desired, or sections) since these process areas might have different regulatory requirements that could impact things like QA/QC and labor costs. This is particularly relevant in some plant-based manufacturing processes. Process review recommended during branches and sections development is simply to verify that proper unit procedure allocations have been made to branches and sections. Two tools to assist in this verification include: (a) the Edit Branch button (Section & Branches Toolbar > Edit Branch), and (b) the Materials and Streams report (Main Ribbon > Reports > Materials & Streams). 1. Define process branches, as needed (Sections & Branches Toolbar > New Branch). 2. Define process sections, as desired, either directly within the active process branch (Sections & Branches Toolbar > New Section). The section sequence within in a specific process branch can be modified to reflect the desired process flow (Sections & Branches Toolbar > Edit Branch > Section Sequence).

172

Matthew J. McNulty et al.

Fig. 5 An example cash flow diagram for a plant-based manufacturing facility that illustrates the key stages of the project lifetime. The cash flow arrows are not drawn to a particular scale

3. Add unit procedures into the relevant process branches sections (with a Unit Procedure highlighted in the VDE. . . Sections & Branches Toolbar > Add to Section). 3.2.5 Economics

Here the reader will be defining the cost structure for the technoeconomic model developed so far. Figure 5 displays an example cash flow diagram to illustrate the key economic stages of a project lifetime that should be considered when developing the cost structure in this method. The reader will be walked through how to define project time valuation (e.g., year of analysis, construction period, project lifetime), project financing (e.g., debt incurred and loan period), operating capacity for each year (i.e., ability to set gradual facility production ramp-up), and the manufacturing cost, as defined by the costs associated with operation, namely the facility operating expenditure (OPEX) ($/year), referred to as total annual operating cost in SuperPro Designer, constituents (materials, facilitydependent costs, labor-dependent costs, laboratory/QC/QA, consumables, utilities, waste treatment/disposal, transportation, miscellaneous, and other), by the costs associated with capital assets and their fixed capital costs, namely capital expenditure (CAPEX) ($), referred to as Total Capital Investment in SuperPro Designer, including direct fixed capital (comprised of direct costs that

Techno-Economic Modeling and Simulation

173

typically includes costs for listed and unlisted purchased equipment, equipment installation, piping, instrumentation, insulation, electrical facilities, buildings, yard improvement, auxiliary facilities and indirect costs that typically includes engineering, construction, contractors fee and contingency), working capital, start-up and validation, up-front research and development, up-front royalties. In most cases of Stage 2 and Stage 3 design estimates, manufacturing costs will be sufficient economic information to assess feasibility. The OPEX normalized by production level, known as cost of goods sold (COGS), and also termed as Unit Production Cost in SuperPro Designer, is often used for decisionmaking. However, there are instances in which profitability can provide essential information on economic viable at this stage of design estimate. Profitability calculations will be detailed in which a product selling price is defined to determine total revenue, and from there to subtract OPEX (including depreciation) and income taxes, to determine the net profit after taxes. Annual cash flows are defined as the net profit after taxes plus depreciation minus any capital expenditures; annual cash flows are used to determine profitability metrics. Common profitability metrics are Return on Investment (ROI) (%), a simple metric of investment efficiency that returns annual growth rate, Discounted Flow Rate of Return (DCFROR) (also referred to as Internal Rate of Return (IRR)) (%), a more complex and time-valued metric of investment efficiency for the annual return, and Payback Time (years), the time required for the facility to reach the break-even point when positive cash flows offset the initial negative cash flows associated with capital expenditures. Readers should refer to the works of Peters et al. [1] and Turton et al. [15] for more detail in estimation of capital cost and scaling, estimation of manufacturing costs, generation of cash flow diagrams, and profitability analyses including the time value of money. Process review recommended during the economics development includes checking correctness and completeness of economic allocations through multiple lenses (e.g., asking questions such as, what is the ratio of upstream to downstream costs, and what are the top economic determinants). Three tools to assist in this review include a) the Economic Evaluation report (Main Ribbon > Reports > Economic Evaluation), b) the Cash Flow Analysis report (Main Ribbon > Reports > Cash Flow Analysis), and c) the Itemized Cost report (Main Ribbon > Reports > Itemized Cost). The report outputs can be customized (Main Ribbon > Reports > Report Options) and entirely custom reports can be defined (Main Ribbon > Reports > Custom Excel Report).

174

Matthew J. McNulty et al.

Note: numbered steps labeled with an asterisk (*) indicate those related to profitability calculations which are not necessary for manufacturing cost calculations. 1. Define the project time valuation, including year of analysis, construction period, start-up period, project lifetime, inflation rate, and net present value (NPV) interest/discount rate of return (Main Ribbon > Edit > Process Options > Economic Evaluation Parameters > Time Valuation) (see Note 23)*. 2. Define the project financing, including the amount of debt incurred, associated loan characteristics, and method for calculation of depreciation (Main Ribbon > Edit > Process Options > Economic Evaluation Parameters > Time Valuation) (see Note 24)*. 3. Define project operating capacity for each year within the project lifetime, including a product failure to account for manufacturing not meeting specifications and needing to be scrapped (Main Ribbon > Edit > Process Options > Economic Evaluation Parameters > Production Level)*. 4. Define remaining project economic evaluation parameters, including income taxes, advertising and selling expenses, and royalty expenses (Main Ribbon > Edit > Process Options > Economic Evaluation Parameters > Misc)*. 5. Define the purchasing price, selling price, and waste treatment or disposal costs for all registered pure component and stock mixtures (Main Ribbon > Tasks > Pure Components OR Stock Mixtures > Register, Edit/View Properties > View/Edit the Selected Component OR Mixture Properties > Economics) and for all output streams (Main Ribbon > Tasks > Stream Classification) (see Note 25). 6. Define equipment purchasing price and adjustments (e.g., installation cost, maintenance costs) for each equipment item (VDE > Unit Procedure > Equipment Data > Purchase Cost OR Adjustments). By default, SuperPro Designer will initialize these values with built-in cost models and adjustments. 7. Define the purchasing price and disposal cost for all registered consumables (Main Ribbon > Tasks > Other Resources > Consumables > View/Edit Properties > Properties > Cost Data). 8. Define the purchasing price for all registered power types (Main Ribbon > Tasks > Other Resources > Heat Transfer Agents > View/Edit Properties > Properties > Price). 9. Define the purchasing price for all registered heat transfer agents (Main Ribbon > Tasks > Other Resources > Power Types > View/Edit Properties > Properties > Agent Cost).

Techno-Economic Modeling and Simulation

175

10. Define the cost rate for all register labor types (Main Ribbon > Tasks > Other Resources > Labor Types > View/Edit Properties > Properties > Cost Data). The cost rate can be defined either by a Detailed Estimate (including a basic rate plus factors like benefits and supervision) or by Lumped Estimate. Additionally, define the Labor Time Estimation, the fraction of work time devoted to process-related activity (see Note 26). 11. Define the operating cost adjustments for all process sections (with a section selected from the Sections & Branches Toolbar drop-down menu. . . Sections & Branches Toolbar > Section Operating Cost Adjustments). This includes estimations of facility-dependent costs (e.g., maintenance, depreciation, insurance), lab/quality control/quality assurance costs (as a fraction of the section’s labor costs), and miscellaneous expenses (e.g., research and development, validation) (see Note 27). 12. Define the capital cost adjustments for all process sections (with a section selected from the Section & Branches Toolbar drop-down menu. . . Sections & Branches Toolbar > Section Capital Cost Adjustments). This includes estimations of direct fixed capital costs (e.g., cost contributions of direct costs like piping, instrumentation, and indirect costs like engineering and construction), cost allocation factors, and miscellaneous costs (e.g., working capital, upfront research and development, startup and validation costs) (see Note 28). 13. Define stream classifications for each of the output streams (Main Ribbon > Tasks > Stream Classification). Set the disposal cost for waste streams and the selling price of revenue streams (see Note 29). 14. Use SuperPro Designer’s “Perform economic calculations” button (Main Toolbar > Perform Economic Calculations) to update the economic calculations with the newly populated values. Any relevant warnings and/or errors associated with the economic values as they are defined will population in the Error Output window below the VDE. 15. View manufacturing cost metrics (COGS, OPEX, CAPEX) and profitability metrics (ROI, IRR after tax, Payback Time) through the Executive Summary (Main Ribbon > View > Executive Summary) or by generating Economic Evaluation and/or Cash Flow Analysis reports, as referred to in the aforementioned process review tools (see Note 30). 3.2.6 Environmental Impact

Here the reader will be using the technoeconomic model mass balance data to identify environmental, health, and safety “hotspots” that could suggest benefits in changes in raw materials, or

176

Matthew J. McNulty et al.

incorporation of processes to reduce stream flows or detoxification prior to disposal, and to quantify the amount of water, raw materials, and consumables required to produce a kilogram of product. In addition, these metrics can be used to compare different biomanufacturing facility designs from a “green engineering” perspective. The reader will be walked through two complimentary methods—an Environmental, Health, and Safety (EHS) assessment and a Process Mass Intensity (PMI) assessment. The EHS assessment detailed here is based on the semiquantitative short-cut method described by Biwer and Heinzle [16] that provides EHS metrics for process inputs and outputs that incorporates the degree of hazardousness of a particular component and the amount of that component used in the process. A diagram overview of the method is shown in Fig. 6. Examples demonstrating how this analysis is performed for a variety of bioprocesses are presented in Heinzle et al. [17].

Fig. 6 Diagram showing the steps in the Environmental, Health, and Safety assessment for biomanufacturing facilities. Analysis steps (solid line white fill) leading to EHS metric use cases (dotted line gray fill) illustrate different ways in which the EHS metrics can be used. (Adapted from [16])

Techno-Economic Modeling and Simulation

177

The PMI metric is a simpler tool, originally implemented for small molecule bioproduction by the Green Chemistry Institute Pharmaceutical Roundtable, that has recently been used in the biomanufacturing industry to quantify the mass (kg) of materials (water, raw materials, consumables) used to produce a kg of product (active pharmaceutical ingredient or API) [18]. Here we describe the steps to implement these analyses and benchmarks that can be used for comparison. It is important to note that these analyses are not explicitly supported by SuperPro Designer, and as such, the steps are intended to be completed on a spreadsheet supported by SuperPro Designer-generated data. 1. Set up a table or spreadsheet containing the mass flowrates for all registered pure components in the technoeconomic model. This information can be easily accessed through the Materials and Streams Report (Main Ribbon > Reports > Materials & Streams) (see Note 31). 2. Separate registered pure component mass flowrates into input and output flowrates. Environmental, Health, and Safety Assessment

Here the reader is walked through the EHS assessment using the Biwer and Heinzle [16] method. There are fourteen EHS impact categories, some of which are combined into six EHS impact groups as shown in Fig. 7. Four of the impact groups are important and/or relevant for process input components (Resources, Grey Input, Component Risk, and Organisms) and four impact groups are relevant for process output components (Component Risk, Organisms, Air, Water/Soil). Suppose Nin is the number of different components in input streams and Nout is the number of components in output streams. 1. Determine the Mass index for each pure component i, MIi, where
 mass of component i ðkgÞ kg i ¼ MIi kg MP mass of main product ðkg MPÞ where i ¼ 1, . . .Nin for input and i ¼ 1, Nout for outputs and enter these values in the spreadsheet for each pure component input or output (see Note 32). 2. Classify the hazardousness of each pure component in each of the fourteen EHS impact categories with ranking of an A, B, or C, where A corresponds to the most hazardous/toxic/ environmentally unfriendly/least sustainable classification and C corresponds to the least hazardous/nontoxic/environmentally benign/sustainable classification with B as an intermediate classification (see Note 33). Enter the A/B/C classification for each input component and each output component into the spreadsheet for each of the fourteen EHS impact categories.

178

Matthew J. McNulty et al.

Fig. 7 Diagram showing the Environmental, Health, and Safety impact categories for biomanufacturing facilities and how each impact category maps onto the EHS impact groups, which are defined by either input impact (boxed in solid blue line) or output impact (boxed in dotted red line). Categorization are according to the Biwer and Heinzle (2004) method [16]

3. Determine the A/B/C classification for the four EHS impact groups, using the “worst case” classification from the EHS impact categories in cases where multiple EHS impact categories are combined in an EHS impact group. For example, if component i is classified as a B in acute toxicity, A for chronic toxicity and B for endocrine disruption potential, that component would be classified as an A for the organism impact group. Enter the A/B/C classification for the four EHS impact groups for both the input and output components in the spreadsheet. 4. For each of the input and output components i, assign a numerical value to A, B, and C. Biwer and Heinzle suggest two methods for this depending on whether the arithmetic or multiplicative weighting is used [16]. For method illustration, we will use the arithmetic weighting formula in which A ¼ 1, B ¼ 0.3 and C ¼ 0 in subsequent steps. 5. For each of the input and output components i, determine the Environment Factor, EFi, by using a formula to convert A/B/

Techno-Economic Modeling and Simulation

179

C classifications defined with numerical values into an overall Environmental Factor using either the arithmetic or multiplicative weighting [16]. For method illustration, we define EFi as the average numerical value of the classifications over the four relevant EHS impact groups. For example, if an output component i had an A for component risk, a B for organisms, C for air and A for water/soil, the EFi would calculated as EFi ¼

1 þ 0:3 þ 0 þ 1 ¼ 0:575 4

With this approach EF values will range from 0 to 1, with 0 being the best case and 1 being the worst case. Enter the EF values for each input and each output component into the spreadsheet. 6. For each input component and each output component i determine the Environment Index, EIi ¼ EFi  MIi, and enter it into the spreadsheet. The higher the numerical value is, the more deleterious (from either an environmental, health or safety standpoint) that component is in the process indicating that it may be worthwhile to consider an alternative compound or try to reduce the amounts used/produced in the process. 7. The overall Environmental Index of the inputs EIin can be determined by summing EIi over all Nin and the overall Environmental Index of the outputs EIout can be determined by summing EIi over all Nout. These values can be “benchmarked” by to the best case (0) and the worst case P comparing in or N out ( N MI ) (Note 34). i i¼1 Process Mass Index (PMI)

The PMI is a measure of how much “material”P is required/conin sumed to make a kg of product, so it is similar to N i¼1 MIi but also includes the mass of consumables. This is particularly helpful in comparing water use between different types of manufacturing facilities since potable water is becoming a more precious resource. The facility PMI can be calculation from the following equation: PMIðkg=kg productÞ ¼ PMIWater ðkg=kg productÞ þ PMIRaw Materials ðkg=kg productÞ þ PMIConsumables ðkg=kg productÞ

1. Determine the PMIWater for the process from the table or spreadsheet of pure component mass flowrates. Assuming that process water (PR), purified water (PW) and water for injection (WFI) are made on-site from municipal water though reverse osmosis and distillation process with efficiencies typically found in the biopharmaceutical industry, the PMIWater can

180

Matthew J. McNulty et al.

PMIwater

be determined from the following formula where MP(kg) is the mass of product produced per year.   W municipal ðkgÞ þ 1:25W PR ðkgÞ þ 1:25W PW ðkgÞ þ 1:41W WFI ðkgÞ ¼ M P ðkgÞ 2. Determine the PMI for raw materials. In this analysis since water is considered separately, the PMIRaw Materials can be determined from the process input components as described in Subheading 3.2.6 Environmental, Health, and Safety Assessment but excluding water as a component. PMIRaw

Materials

¼

N in X

MIi ðkg=kg productÞ

i¼1

3. Record the demand for each registered consumable (e.g., bioreactor bag, media bag, chromatography resin, filter and membrane, including primary packing material) in a table or spreadsheet. This information can be easily accessed from an Economic Evaluation report or registered Consumables view window (Main Ribbon > Tasks > Other Resources > Consumables). 4. Record the mass (in kg) for each registered consumable i, Ci, in a table or spreadsheet (see Note 35). 5. In some cases, the consumable material can be used multiple times (e.g., chromatography resins, some filters). To take this into consideration (see Note 36), the consumable utilization factor Uf is defined as Uf ¼

Number of cycles the material is used per batch Number of cycles the material can be used before replacement 6. Determine PMIConsumables from the demand and mass of consumables used annually in the process using the following equation   U f C bag þ U f C resin þ U f C filter þ U f C membrane PMIConsumables ¼ MP 7. Calculate the sum of the PMI values for water, raw materials and consumables to get the PMI for the facility (see Note 37).

3.2.7 Stop Gate II

Here the reader will be assembling a status report and communicating with key stakeholders for a Stop Gate before proceeding to process analysis. Precise scope and format of the report and/or presentation will depend on the initial project scope and agreement with key stakeholders. Below are generally recommended steps to assembling a robust status report to convey the completed base case technoeconomic model and simulation.

Techno-Economic Modeling and Simulation

181

1. Detail a review of the initial, or most recent, project scope including the specific agreed upon deliverables and timeline. 2. Detail a review of the technoeconomic methodology, including the specific process simulation tool employed for analysis as well as a compilation of the various types of information sources (e.g., lab-scale data, pilot-scale data, literature results, similar existing facilities) used as the basis for technoeconomic parameters. 3. Detail the key project assumptions built into the model and simulation, including the previously agreed upon bases for process design, production specifications, and process flow sheet with critical process parameter and key performance indicator ranges, as well as the bases for OPEX and CAPEX (and profitability) economic calculations (e.g., maintenance costs, utility types, depreciation method, pricing for select cost items). 4. Detail the technical manufacturing results of the model and simulation. This may include a progression of processing metrics (e.g., product mass, concentration, purity, and recovery) and their status at each unit procedure. 5. Detail the economic results of the model and simulation. This may include a snapshot summary of key economic metrics (e.g., COGS, OPEX, CAPEX, ROI, IRR after tax) and more detailed breakdowns of those metrics by process section (e.g., harvest, concentration, capture) or cost item category (e.g., raw material, consumables, labor-dependent). 6. Review the remaining work to meet the project scope, including the specific deliverables, timelines, a proposed path, and any questions. Be prepared to iterate on the base case model and simulation based on key stakeholder feedback. This may require several iterations before there is agreement amongst all key stakeholders to proceed to the next step of the project scope. 3.3

Process Analysis

3.3.1 Price Sensitivity

In the following section, the reader will be manipulating the base case technoeconomic model developed in process synthesis (Subheading 3.2) to gain insight into the manufacturing design space and inform design and/or research and development directions. This will be performed in this method by evaluating price sensitivity (Subheading 3.3.1), scenario analysis (Subheading 3.3.2), and alternate scenarios (Subheading 3.3.3), followed by developing a report and initiating review with key stakeholders at the final Stop Gate III (Subheading 3.3.4). Here the reader will be performing univariate price sensitivity analyses. For two classes of price parameters, purchase price and product selling price, the reader will be introduced to possible

182

Matthew J. McNulty et al.

methods of identifying how to select parameters for sensitivity analysis, and then be led through a process of executing a sensitivity analysis. In this analysis, only the cost structure of the facility model is modified. 1. Assemble a list of purchase price parameters for sensitivity analysis using criteria of cost contribution, uncertainty in parameter value, and expected change in parameter value. The relative weighting of these criteria and the list size is based on individual project scope. The economic evaluation or itemized cost reports provide a list of cost item factor contributions that can be used to form cost contribution rankings. 2. Define the extent of variation to be tested for each parameter in the purchase price sensitivity analysis, usually structured plus/ minus a percentage of the base case value. 3. Update the purchase price values in SuperPro Designer, followed by running the “Perform economic calculations” button to update the economics, which can then be recorded on a spreadsheet for graphical interpretation. This is recommended for CAPEX purchase price sensitivity. Alternatively, the facility simulation economics can be manually updated with the adjustment of the OPEX purchase price parameter value, as these costs do not factor into additional downstream calculations. 4. Define a set of product selling prices. This can be set up as exploratory, based on intervals of percentage change from the base case value, or this can be derived from calculations on market analysis (e.g., testing market entry strategies). 5. Update the product selling price in SuperPro Designer, and relevant profitability metrics of choice (e.g., ROI, IRR after tax) can be pulled from the updated economic calculations and recorded on a spreadsheet for graphical interpretation. 3.3.2 Scenario Analysis

Here the reader will be performing univariate process parameter scenario analysis. It is often beneficial to elucidate the technoeconomic impact of varying high-level facility performance parameters including key performance indicators (e.g., expression level/titer), critical quality attributes (e.g., product purity), and market parameters (e.g., yearly production level). These insights can be crucial for allocating research and development resources, general decision-making, and defining business strategy. In this analysis, the technical performance, facility design modifications, and cost structure of the model are modified. 1. Assemble a list of process parameters for scenario analysis, and extent of variation, based on project scope, complimentary analyses, and/or additional key stakeholder feedback.

Techno-Economic Modeling and Simulation

183

2. Generate a clone of the base case model for each process parameter value to be tested in SuperPro Designer. The model will change significantly in this analysis and it is important to isolate these changes from the base case model. 3. Update the model with the process parameter setpoint selected. This update will result in a cascade of process performance changes. 4. Modify the model to negate off-target impact of the process parameter variation. Usually, a series of modifications are needed to accomplish this. Some of these changes can be automatically calculated by SuperPro Designer (e.g., equipment in Design Mode will resize accordingly to the new stream composition and volume/mass basis), while other aspects will require manual intervention (e.g., variation in expression level will subsequently result in changes to production level and product purity, which will have to be adjusted for). 5. Generate reports for each model and compile desired technical/economic outputs on a spreadsheet for graphical interpretation. 3.3.3 Alternate Scenarios

Here the reader will be performing comparative alternate scenario analysis. Alternate scenarios can range from evaluating a single new unit procedure all the way to evaluating an entirely new upstream section or, in some cases, complete process flow sheets. It may be valuable to chart out the preferred manufacturing among several similar options (e.g., cultivation of plants using a greenhouse or a controlled environmental facility). Additionally, it may also be valuable to generate comparisons of drastically different manufacturing options (e.g., plant cell suspension culture or whole-plant transient expression). 1. Follow the steps detailed in process creation (Subheading 3.1) as closely as is needed for the alternate scenario and the extent of process flow sheet modification. 2. Follow the steps detailed in section process synthesis (Subheading 3.2) as needed. Either revise the existing facility model(s) or start with a new model. 3. Generate reports for the new model(s) and compile desired technical/economic outputs alongside the base case model outputs on a spreadsheet for graphical interpretation.

3.3.4 Stop Gate III

Here the reader will be assembling a project summary report and communicating with key stakeholders for a final Stop Gate before concluding project work. This report and/or presentation is often viewed as the primary project deliverable, and as such, should be given commensurate attention. Precise scope and format of the report and/or presentation will depend on the initial project

184

Matthew J. McNulty et al.

scope and agreement with key stakeholders. Below are generally recommended steps to assembling a robust status report to convey the completed technoeconomic analysis. 1. Begin by updating and compiling information communicated in the previous two Stop Gates, namely the project scope and base case model. 2. Generate a single-page executive summary with key findings from the analysis, usually including a note on economic viability for scaling up to pilot or commercial scale, to serve as a reference for key stakeholders. 3. Generate an extended length report and/or presentation with detailed findings of the base case model and process analysis, as framed by the project scope. Include figures that convey key takeaways for use by key stakeholders in external communications after the project has concluded.

4

Notes 1. Regulatory framework will constrain the manufacturing process to meeting proper standards for design, monitoring, and control. For example, manufacturing of biopharmaceutical proteins intended for commercial use in the USA must follow current good manufacturing practices (cGMP). 2. The facility location is an important technoeconomic consideration and will influence factors such as raw materials and utility availability and pricing, labor supply, taxation, and environmental and legal restrictions. 3. Forecasts of production demand are most commonly based on market analyses. It is important to consider that production demand may change over time (e.g., dynamic market penetration forecasts). 4. Facility lifetime is only an important design premise if one intends to perform a profitability analysis as a part of their technoeconomic method. This is not needed if the economic insights sought are scoped at manufacturing costs (OPEX, CAPEX, COGS). 5. Critical quality attributes and key performance indicators may not yet be defined at the time of the technoeconomic analysis. Therefore, valuable parameters to include at this stage may be loosely defined using working process knowledge. For the typical plant-based manufacturing process, parameters such as expression level, downstream recovery, and product purity are generally recognized as important parameters for which key stakeholder agreement should be established early in the project.

Techno-Economic Modeling and Simulation

185

6. When building a technoeconomic model from lab-scale data, it is recommended that a notable amount of time is devoted to considering commercial-scale equivalent unit operations (e.g., while liquid nitrogen–assisted homogenization and extraction may have been used at the lab scale, one might consider employing a screw press at commercial scale simulation). 7. User-populated databanks can be transferred between projects. This can save considerable start-up time if one is building their analysis starting from another technoeconomic model. 8. One shortcut to registering the pure components and stock mixtures is to initialize the new registration using a source for default property values (e.g., registering “transgenic rice cell” can be initialized with biomass default property values). It is also important to note that the economics and pollutant categories will be visited later in the method. 9. It is important to consider the different types/qualities of water, ranging from potable water for process cooling to water-for-injection (WFI) for the last chromatography step and bulk drug substance formulation. 10. The stream inlet/outlet arrows of some unit procedure icons are designed to handle specific stream types (e.g., tank vent stream). For detailed information on the restrictions of a given unit procedure icon stream inlets/outlets, navigate to the “Help” menu in the Main Ribbon and search for that unit procedure by name. 11. The authors have found that this iterative method of process synthesis in SuperPro Designer’s visual design environment results in a streamlined and hassle-free troubleshooting process. Alternatively, SuperPro Designer does possess a “Simulation Control” toolbar with a set of functions to enable process breakpoints and partial simulations for a more classical programming approach to troubleshooting. 12. Cleaning operations for plant-based manufacturing at commercial scale generally include Steam-in-Place and/or Cleanin-Place. Cleaning heuristics that the authors have found to be relevant and useful for development of cleaning operations can be found in the works of Chisti [19], Bremer and Brent [20], and Davies et al. [21]. 13. It is important to consider the operating schedule of the facility when defining the annual operating time; is the facility operating according to 24 h/7 days per week, 24/5, 8/5, 8/7? SuperPro Designer is primarily suited for 24/7 manufacturing. As mentioned earlier, SchedulePro can be used to augment the native scheduling capabilities, which includes simulation of different facility operating schedules like 8 h per 5 days per week.

186

Matthew J. McNulty et al.

14. Default labor types do not necessarily accurately reflect the upstream operators of whole-plant–based manufacturing, whom have been historically less expensive given the lower complexity of plant growth operations. The target product and facility location may also influence labor type pay rates. 15. The labor rate can be strongly influenced by the product industry and regulations. For example, cGMP operation often requires one operator to perform the commercial manufacturing task while another is dedicating to observing, verifying, and signing off as a witness for critical operations such as preparing/adding media and/or buffers, and equipment sterilization. The labor amount needs to reflect additional expenditures of this nature. 16. Process time can either be set by the user, calculated based on a mass or volume flowrate (for some operations), or set as matched to the duration of another operation by defining a master–slave relationship. In this last case, one should consider if the labor of the Slave operation should be accounted for or negated. In some situations (e.g., Tank 1 transfer out is a slave to Tank 2 transfer in), it may be appropriate to negate the labor of the Slave operation when the labor for both operations is adequately reflected in the Master operation. 17. Scheduling the connections of operations is one place that a new practitioner of technoeconomic analysis may struggle. It is important to consider the scheduling bottleneck. Scheduling may also need to be reconsidered after equipment allocations are completed in Subheading 3.2.3. 18. A single piece of equipment can be assigned to multiple unit procedures; there does not have to be a one-to-one mapping. Equipment is assigned to the unit procedure in the “Selection” subwindow of the “Equipment” tab. 19. Equipment can be defined in either Design or Rating Mode. In Design Mode, equipment is sized according to SuperPro Designer calculations. In Rating Mode, the equipment size is user-defined and fixed, and the throughput/scheduling is determined. 20. The number of equipment units can be increased to split the processing between multiple identical equipment units. Additionally, “Stagger Mode” can be enabled, which generates identical sets of equipment units, only one set of which is for any given batch. This is an effective strategy to debottleneck the manufacturing scheduling when the equipment in question is identified as a bottleneck. 21. The default SuperPro Designer Consumables database is a very useful tool. However, the default property values for use cycles have not been developed based on plant-based manufacturing

Techno-Economic Modeling and Simulation

187

data. One may find that the use cycles are significantly higher or lower than evidenced in supporting plant-based manufacturing data. 22. Power types and their subsequent pricing structures are geographically dependent and should be defined with the facility location in mind. It is important to consider fees and/or discounts associated with certain thresholds of power demand, which are generally embedded in power consumption calculations. 23. The user is required to specify an NPV interest/discount rate of return. Typically, 7% is the value used here unless key stakeholders provide a specific value. 24. Depreciation can be applied to the total depreciable capital investments using common methods such as straight line, declining balance, and sum of the years digit methods, along with specification of the depreciation period and the salvage value. Depreciation is specified in the “Time Valuation” tab, but section- and equipment-specific depreciation calculations are controlled through the “Section Operating Cost Adjustments” window. 25. Values defined for output streams can be selected to override the base component and/or stock mixture economic property value in the technoeconomic simulation by checking the box “Is Cost/Price Set by User.” 26. The often-lower complexity of upstream whole-plant–based manufacturing, as referenced in Note 15 (labor), also generally results in a higher labor time estimation, as less time is needed for paperwork in simpler operations. 27. Generally, the operating cost adjustments for downstream processes are significantly higher than for upstream processes. This is mainly attributed to the higher labor, lab/QA/QC, ongoing validation, and maintenance costs as the stream progresses closer to the final product. 28. In general, the capital cost adjustments for upstream wholeplant–based manufacturing are lower than those of traditional bioreactor-based processes. This can be attributed to several factors, including lower capital complexity for plant cultivation and reduced startup and validation costs due to the linear scalability of whole plants. Within whole-plant systems, the heuristics for capital cost adjustments are that indoor agriculture > greenhouse > open field cultivation. 29. The selling price of the product can be difficult to determine at this stage of development, largely depending on the extent of market analysis and business model development. If the reader does not have quality information at hand to estimate a selling

188

Matthew J. McNulty et al.

price, the authors recommend that the reader focus on the cost of goods sold, termed Unit Production Cost in SuperPro Designer, rather than the profitability based on the selling price. Later in the method, price sensitivity analysis represents a valuable tool for exploring viable selling price options, as desired. 30. Positive cash flows following production startup are of course a necessary but not sufficient requirement for profitability. 31. Components involved in internal streams (those connecting one unit procedure to another unit procedure), utilities, and/or consumables are not included in this analysis. 32. The Materials and Streams Report provides pure component mass flowrates on an annual (kg/year), batch (kg/batch), and main product (kg/kg MP) basis. The main product basis can be used directly here. 33. This classification, although, somewhat arbitrary is based on information about the component often found in material data safety sheets or U.S. National Fire Protection Agency (NFPA) rating (see Table 1 in Biwer and Heinzle [16]). 34. Because there is an overall mass balance for the entire process, the summation of mass over all input components is equal to the summation of mass overall output components. 35. The mass of consumables is not a property field in SuperPro Designer. Mass data must be manually identified and compiled. 36. The consumables demand listed in SuperPro Designer accounts for use cycles by default. This manual calculation is not required for demand data obtained from SuperPro Designer. 37. For comparison Budzinski et al. [18] presents the average PMI values for fourteen commercial biomanufacturing production runs for monoclonal antibody production using mammalian cell culture (although it should be noted that cleaning solutions were not included in the analysis): PMIwater¼ 7711 kg/ kg P, PMIRaw Materials ¼ 551 kg/kg P, PMIConsumables ¼ 65 kg/ kg P, giving a total PMI of 8327 kg/kg P. References 1. Peters MS, Timmerhaus KD, West RE (2003) Plant design and economics for chemical engineers. McGraw-Hill, New York 2. Petrides D, Carmichael D, Siletti C et al (2019) Bioprocess simulation and economics. In: Essentials in fermentation technology. Springer, Cham, pp 273–305 3. Fischer R, Buyel JF (2020) Molecular farming – the slope of enlightenment. Biotechnol Adv 40:107519

4. Corbin JM, McNulty MJ, Macharoen K et al (2020) Technoeconomic analysis of semicontinuous bioreactor production of biopharmaceuticals in transgenic rice cell suspension cultures. Biotechnol Bioeng 117:bit.27475 5. McNulty MJ, Gleba Y, Tuse´ D et al (2020) Techno-economic analysis of a plant-based platform for manufacturing antimicrobial proteins for food safety. Biotechnol Prog 36: e2896

Techno-Economic Modeling and Simulation 6. Alam A, Jiang L, Kittleson GA et al (2018) Technoeconomic modeling of plant-based Griffithsin manufacturing. Front Bioeng Biotechnol 6:102 7. Nandi S, Kwong AT, Holtz BR et al (2016) Techno-economic analysis of a transient plantbased platform for monoclonal antibody production. MAbs 8:1456–1466 8. Walwyn DR, Huddy SM, Rybicki EP (2015) Techno-economic analysis of horseradish peroxidase production using a transient expression system in Nicotiana benthamiana. Appl Biochem Biotechnol 175:841–854 9. Tuse´ D, Tu T, McDonald KA (2014) Manufacturing economics of plant-made biologics: case studies in therapeutic and industrial enzymes. Biomed Res Int 2014:256135 10. Nandi S, Yalda D, Lu S et al (2005) Process development and economic evaluation of recombinant human lactoferrin expressed in rice grain. Transgenic Res 14:237–249 11. Holtz BR, Berquist BR, Bennett LD et al (2015) Commercial-scale biotherapeutics manufacturing facility for plant-made pharmaceuticals. Plant Biotechnol J 13:1180–1190 12. Tekoah Y, Shulman A, Kizhner T et al (2015) Large-scale production of pharmaceutical proteins in plant cell culture-the protalix experience. Plant Biotechnol J 13:1199–1208 13. D’Aoust M-A, Couture MM-J, Charland N et al (2010) The production of hemagglutinin-based virus-like particles in plants: a rapid, efficient and safe response to

189

pandemic influenza. Plant Biotechnol J 8: 607–619 14. Pleitt K, Somasundaram B, Johnson B et al (2019) Evaluation of process simulation as a decisional tool for biopharmaceutical contract development and manufacturing organizations. Biochem Eng J 150:107252 15. Turton R, Shaeiwitz JA, Bhattacharyya D et al (2018) Analysis, synthesis, and design of chemical processes. Pearson, London 16. Biwer A, Heinzle E (2004) Environmental assessment in early process development. J Chem Technol Biotechnol 79:597–609 17. Heinzle E, Biwer AP, Cooney CL (2007) Development of Sustainable bioprocesses: modeling and assessment. Wiley, Hoboken 18. Budzinski K, Blewis M, Dahlin P et al (2019) Introduction of a process mass intensity metric for biologics. N Biotechnol 49:37–42 19. Chisti Y (1999) Modern systems of plant cleaning. In: Encyclopedis of food microbiology. Academic Press, London, pp 1806–1815 20. Bremer PJ, Seale RB (2010) Clean-in-Place (CIP). In: Encyclopedia of industrial biotechnology. Wiley, Hoboken 21. Davies S, Sykes T, Philips M et al (2015) Hygienic design and Cleaning-In-Place (CIP) systems in breweries. In: Brewing microbiology: managing microbes, ensuring quality and valorising waste. Elsevier, Amsterdam, pp 221–239

Part III Optimization Strategies

Chapter 12 Optimization of Vectors and Targeting Strategies Including GoldenBraid and Genome Editing Tools: GoldenBraid Assembly of Multiplex CRISPR/Cas12a Guide RNAs for Gene Editing in Nicotiana benthamiana Beatriz Gonza´lez, Marta Vazquez-Vilar, Javier Sa´nchez-Vicente, and Diego Orza´ez Abstract New breeding techniques, especially CRISPR/Cas, could facilitate the expansion and diversification of molecular farming crops by speeding up the introduction of new traits that improve their value as biofactories. One of the main advantages of CRISPR/Cas is its ability to target multiple loci simultaneously, a key feature known as multiplexing. This characteristic is especially relevant for polyploid species, as it is the case of Nicotiana benthamiana and other species of the same genus widely used in molecular farming. Here, we describe in detail the making of a multiplex DNA construct for genome editing in N. benthamiana using the GoldenBraid modular cloning platform. In this case, the procedure is adapted for the requirements of LbCas12a (Lachnospiraceae bacterium Cas12a), a nuclease whose cloning strategy differs from that of the more often used SpCas9 (Streptococcus pyogenes Cas9) enzyme. LbCas12a-mediated edition has several advantages, as its high editing efficiency, described for different plant species, and its T/ A-rich PAM sequence, which expands the range of genomic loci that can be targeted by site-specific nucleases. The protocol also includes recommendations for the selection of protospacer sequences and indications for the analysis of editing results. Key words Biofactories, CRISPR/Cas12a, Editing, GoldenBraid, Multiplexing, Nicotiana benthamiana

1

Introduction Plant biofactories for molecular farming are a relatively new type of crops that have not yet been the subject of breeding programs. Paradoxically, beneficial traits associated with molecular farming are often different or even opposite to those pursued in food crops. Permissiveness to viral infections, impeded flowering or modified glycosylation profiles are examples of nontraditional trait objectives associated to plant biofactories. In this context, new breeding

Stefan Schillberg and Holger Spiegel (eds.), Recombinant Proteins in Plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480, https://doi.org/10.1007/978-1-0716-2241-4_12, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

193

194

Beatriz Gonza´lez et al.

techniques and specially genome editing tools offer an extraordinary opportunity to accelerate the breeding of plant biofactories, therefore improving their efficiency and economic viability. The popular CRISPR/Cas9 gene editing technology has been successfully applied to Nicotiana benthamiana and Nicotiana tabacum, two widely used plants in molecular farming [1]. Both are allotetraploid species, therefore they usually feature extended gene families that complicate the full inactivation of a given gene function by gene editing. Fortunately, with CRISPR/Cas technology it is possible to target several genes simultaneously by transforming plants with multigene constructs carrying tandem arrays of gRNAs, an approach usually referred to as “multiplexing.” The design and assembly of multiplexing constructs is greatly facilitated by modular cloning strategies based on the Golden Gate technology such as MoClo [2] or GoldenBraid (GB) [3]. In our laboratory we developed and maintain the so-called GB platform, a modular cloning system supported by a number of software tools that facilitate the design of multigene constructs in silico. GB also provides a collection of standard DNA elements (“parts”), many of them available at the Addgene repository. DNA parts are catalogued in hierarchical levels, where basic elements such as promoters, CDSs, and terminators, are classified as level 0 parts. Level 1 parts are combinations of level 0 parts, usually forming transcriptional units (TUs). Level 1 parts are combined binarily in iterative cloning reactions to produce higher level (>1) multigene elements called modules. Detailed protocols for the GB-assisted assembly of multigene constructs, including CRISPR/Cas9 multiplex constructs have been described elsewhere [4]. More recently, new RNA-guided site-directed nucleases other than Cas9 are being incorporated to the breeder’s toolbox, expanding the capabilities of the technology. In particular, the nuclease Cas12a offers interesting new features such as a T-rich PAM sequence that facilitates the targeting of noncoding genomic regions [5, 6]. Very recently, thermostable versions of Cas12a have been reported that significantly increase the targeting efficiency of this nuclease [7]. In this chapter we will focus on the assembly of multiplex constructs for Cas12a. Due to the distinctive features of the Cas12a gRNAs, the assembly of tandem arrays of protospacers for this nuclease involve cloning protocols that differ substantially from those employed for Cas9. As an example, we provide here a detailed description of the construction of a fourguide Cas12a multiplex construct targeting four independent loci in N. benthamiana. An efficiency test based on transient expression of the final construct is also described. Agroinfiltration-based gRNA efficiency tests are advisable prior to stable transformation in N. benthamiana.

GB-Adapted CRISPR/Cas12a-Based Gene Editing in N. benthamiana

2

195

Materials

2.1 GoldenBraid Plasmids

The set of GB plasmids required for Cas12a-mediated gene editing is listed below. These are described in detail in [3, 8], and they are available at the Addgene repository (https://www.addgene.org/ kits/orzaez-goldenbraid2/). All sequences are accessible at GB cloning website (https://gbcloning.upv.es/search/features/) using the GB database ID. 1. GB destination vectors: The pUPD2 vector is the primary GB destination vector for cloning basic (level 0) DNA parts. The α level destination vectors pDGB3α1 and α2 are modified versions of the pCambia binary plasmid, and they are used for the initial cloning of single transcriptional units (TUs). The Ω level destination vectors pDGB3Ω1 and Ω2 are used for combining several TUs (referred to as modules). Both TUs and modules assembled in α and Ω level destination vectors can be combined in binary GB reactions to generate complex multigene constructs, and also can be used directly in plant transformation. 2. Cas12a transcriptional unit: GB3720 is a GB level 1 plasmid (transcriptional unit) for the expression of the human codonoptimized Cas12a endonuclease from Lachnospiraceae sp., driven by the CaMV 35S promoter. 3. RNA Polymerase III promoter: GB1443 is a level 0 GB part that contains the first scaffold (direct repeats, DR) of the LbCas12a endonuclease guide RNA, regulated by the Arabidopsis thaliana U6-26 RNA PolIII promoter. 4. Silencing suppressor P19: GB1203 is a GB level 1 plasmid (transcriptional unit) for the expression of the silencing suppressor P19 of TBSV, which is necessary to increase the transient expression levels of a recombinant protein.

2.2

Software Tools

1. DNA sequence editor software for plasmid design and sequence analysis (e.g., https://www.benchling.com/) and GB cloning software (https://gbcloning.upv.es/). 2. A software tool for genome editing analysis (e.g., https://ice. synthego.com/#/ or http://shinyapps.datacurators.nl/tide/). 3. N. benthamiana genome database with genome annotation (i.e., Solgenomics (https://solgenomics.net/tools/blast/) or nbenth (https://apollo.nbenth.com/annotator/index).

2.3

GB Cloning

1. GB-adapted DNA synthesis fragment including the protospacer (guide) sequences designed for multiplexing CRISPR/ Cas12a-based gene editing.

196

Beatriz Gonza´lez et al.

2. Restriction enzymes BsmBI-v2 and BsaI-HFv2 (NEB) and CutSmart Buffer (NEB). 3. T4 DNA ligase (Promega) and 10X T4 DNA ligase buffer. 4. Restriction enzymes corresponding buffers.

NotI,

EcoRI,

and

NcoI

and

5. BSA. Prepare aliquots of 1 mL at 1 mg/mL (dissolved in milliQ H2O) and store at
20  C. 6. Thermocycler for the GB reactions. 7. Agarose gel and electrophoresis unit. 8. UV Transilluminator. 2.4 Bacteria Transformation and Culture

1. Laminar flow cabinet for bacteria cultures preparation. 2. Competent TOP10 Escherichia coli cells prepared with the Mix&Go E.coli Transformation Buffer Kit (ZymoResearch), following manufacturer’s instructions. We recommend preparing aliquots of 50–100 μL, stored at
80  C. Homemade electrocompetent cells can also be used. 3. Agrobacterium tumefaciens strain GV3101 C58C1 homemade electrocompetent cells (see Note 1). 4. Sterile Lysogeny Broth (LB) medium: 10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl. 5. Sterile LB agar medium: 10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl, 15 g/L agar. 6. Antibiotic stocks (1000X) of chloramphenicol (34 mg/mL, dissolved in 100% ethanol), kanamycin (50 mg/mL, dissolved in milliQ H2O), spectinomycin (50 mg/mL, dissolved in milliQ H2O) and rifampicin (50 mg/mL, dissolved in DMF or DMSO). Store at
20  C. Those antibiotics dissolved in milliQ H2O are filter-sterilized. 7. Stocks of screening reagents for blue/white selection of positive clones: X-Gal (20 mg/mL, dissolved in DMF) and IPTG (0.4 M, dissolved in milliQ and filter-sterilized). Store at
20  C. 8. Sterile SOC medium: 20 g/L tryptone, 5 g/L yeast extract, 10 mM NaCl, 2.5 mM KCl (named as SOB medium). Added after sterilization: 10 mM MgCl2, 10 mM MgSO4, 20 mM glucose (see Note 2). Prepare aliquots of SOC medium (1 mL) and store at room temperature (RT). 9. Electroporator (1 mm Gap).

and

electroporation

plastic

cuvettes

10. E.Z.N.A. Plasmid DNA Miniprep Kit (for E. coli plasmid isolation) (Omega Bio-tek). 11. QIAprep Spin Miniprep Kit (for A. tumefaciens plasmid isolation) (QIAGEN).

GB-Adapted CRISPR/Cas12a-Based Gene Editing in N. benthamiana

197

12. Spectrophotometer for the quantification of DNA concentration (e.g., NanoDrop® ND-1000 UV-Vis). 13. Primers for sequencing the DNA insert cloned into the pUPD2 vector (see Note 3). 14. Sterile glycerol solution (10% (see Note 1) and 50–70% (for bacterial glycerol stocks)). 15. Shaker and growing chamber at 28 and 37  C. 2.5 Plant Transient Transformation

1. Agroinfiltration Buffer: 10 mM MES (pH 5.6, adjusted with KOH), 10 mM MgCl2, 200 μM acetosyringone. Prepared freshly the agroinfiltration day, and stored in dark at RT (the bottle can be wrapped in aluminum foil) (see Note 4). 2. Sterile 1 mL syringe without needle. 3. Horizontal rolling mixer. 4. Spectrophotometer equipment to measure the bacterial optical density and corresponding cuvettes (Semi-micro 1.5 mL). 5. 4–5-week-old Nicotiana benthamiana plants (growing conditions: 24  C day–20  C night in a 16 h light–8 h dark cycle).

2.6 DNA Extraction and PCR Amplification

1. Liquid N2 and freezer. 2. Cork-borer (approx. 1.5 cm diameter). 3. Small pestles or steel beads. 4. Mixer Mill MM400 Retsch (optional). 5. CTAB Buffer: 2% (w/v) CTAB ((1-Hexadecyl)trimethylammonium bromide, 98%), 1.4 M NaCl, 20 mM EDTA (pH 8.0, adjusted with NaOH), 100 mM Tris–HCl (pH 8.0), 2% (v/v) β-mercaptoethanol, 0.115% (v/v) RNaseA (optional: Omega Bio-tek). Prepared freshly the same day and stored at RT. See Note 5. 6. Water bath at 65  C. 7. Chloroform:isoamyl alcohol (24:1). 8. Isopropanol. 9. 80% ethanol. 10. H2O mQ (or TE buffer). 11. Spectrophotometer for the quantification of DNA concentration (e.g., NanoDrop® ND-1000 UV-Vis). 12. MyTaq DNA Polymerase, including 5X MyTaq Reaction Buffer (Bioline Meridian Bioscience). 13. Set of specific primers for amplification of the targeted sites. 14. Thermocycler for PCR amplification. 15. Agarose gel and electrophoresis unit.

Beatriz Gonza´lez et al.

198

16. UV Transilluminator. 17. ExoSAP-IT (Applied Biosystems) or NucleoSpin Gel and PCR cleanup Kit (Macherey-Nagel), for purification of PCR amplification products.

3

Methods Here we describe a detailed protocol for Cas12a-mediated gene editing in Nicotiana benthamiana that could be used as guidance for all researchers interested in performing plant genetic engineering with CRISPR/Cas12a. In particular, we show as an example the design and functional validation of a four-gRNAs multiplexing construct for targeting TFL1 (Terminal Flower 1-like), FT (Flowering locus T), XT1 (Beta-1,2-xylosyltransferase), and CBP (Chlorophyll a-b binding protein) genes simultaneously in N. benthamiana plants.

3.1 Identification of Target Gene(s) of Interest for Editing

In the case study shown here, we chose to edit TFL1, FT, XT1, and CBP genes to get a knock-out of the proteins they encode. In a first step, a search of the gene sequences of interest is necessary to define the target sites for editing. N. benthamiana gene sequences can be searched either at Solgenomics (https://solgenomics.net/tools/ blast/) or at nbenth (https://apollo.nbenth.com/annotator/ index). We recommend using nbenth site, where a chromosome level assembly of the N. benthamiana genome can be found. In this case, we used tBLASTn tool for searching the four target gene sequences in the genome of N. benthamiana, using homologues in other plant species (e.g., Arabidopsis) as queries. The accession numbers of the genes selected for targeting are listed in Table 1.

Table 1 Accession numbers of target genes of interest Name gene TFL1-3.1

Description

Accession number Solgenomics

Accession number nbenth

Protein TERMINAL FLOWER 1-like

Niben101Scf01028g01003.1 NbD008407.1

TFL1-14.1 Protein TERMINAL FLOWER 1-like

Niben101Scf00906g00014.1 NbD007529.1

FT

Protein FLOWERING LOCUS T

Niben101Scf01519g10008.1 NbD012042.1

XT1

Beta-1,2-xylosyltransferase

Niben101Scf04205g03008.1 NbD028193.1

CBP

Chlorophyll A-B binding protein Niben101Scf02459g00024.1 NbD018463.1

GB-Adapted CRISPR/Cas12a-Based Gene Editing in N. benthamiana

3.2 Considerations for the Selection of RNA Guides for LbCas12a

199

Different tools that assist in the selection of the 20–23 nucleotides (nt) protospacer sequences targeting the genes of interest can be used (e.g., Benchling (https://www.benchling.com/), CRISPOR [9], CINDEL [10], CRISPR-DT [11]). For LbCas12a protospacers, we recommend the CRISPOR tool (http://crispor.tefor.net/ ). Besides, there are different parameters affecting gRNAs selection (listed below), for which several considerations must be taken into account: 1. Protospacer Adjacent Motif (PAM) recognition: LbCas12a recognizes AT-rich sequences with a 50 -TTTN-30 PAM, situated at 50 of the protospacer sequence [5, 12]. However, it shows robust activity only for PAM sequence TTTV (V: A/C/G), whereas the sequence TTTT shows a lower efficiency. Indeed, the highest activities of LbCas12a were reported with PAM sequences with a
1 ‘A’ nucleotide (PAM sequence: TTTA) [13]. On the other hand, the sequence TTV has been also reported as Cas12a PAM, but with some variants of Cas12a it showed lower efficiency than TTTV [5]. 2. On- and off-target scores: Guides with the highest on-target scores and the lowest off-target scores are selected, and these values can be calculated with different algorithms. For off-target scores determination, it is necessary that the target plant genome is available in the software. In reference to on-target scores, CRISPOR calculates them with the DeepCpf1 algorithm [14]. Additionally, on-target scores can be calculated at http://big.hanyang.ac.kr/gDesigner/, which uses the CINDEL algorithm [10]. However, all the algorithms developed so far for on-target score prediction are based on data obtained from experiments in silico and/or performed in mammalian cells. Therefore, these scores cannot be directly extrapolated for the prediction of in planta performance of guide RNAs. Despite the continuous improvement of these tools, there is not always a clear correlation between the ranking scores and in vivo efficiency [15]. 3. Specific characteristics of Cas12a: Cas12a cleaves the target DNA site at 18–23 bp distal to the 30 side of the PAM, producing staggered breaks (leaving 50 overhangs of 5–8 nt) [16]. Cas12a requires only a single RNA molecule (a mature crRNA), which consists of a 19–20 nt direct repeat (DR) sequence (the RNA scaffold situated at the 50 end of the structure) and 20–23 nt protospacer or guide RNA (gRNA) sequence [17]. The full size of the protospacer (23 nt) is not necessary for the activity of Cas12a, and even shortening the base pairing of spacer up to a minimum of 19 bp does not influence the activity [13]. Additionally, Cas12a has

200

Beatriz Gonza´lez et al.

Table 2 Sequence of protospacers designed for LbCas12a targeting genes of interest Name gRNA

Sequence (50 -30 )

crRNA TFL1 (3.1 and 14.1)

ACTGTGGGACTGAATGAATC

crRNA FT

CAAGATCTATTGGCCTAAGA

crRNA XT1

TATGTAGGTGTATTTGGAAT

cRNA CBP

GAGGACAAACTACATCCAGG

endoribonuclease activity for processing the mature crRNAs, so multiplex constructs can be easily created without additional cleavable structures and/or RNase activities [16, 17]. 4. Position of the target site in the gene of interest: For gene knockouts, a general rule is to target the first exon, paying attention to downstream in-frame ATGs that could function as alternative translation start sites. On the other hand, multiplexing editing, that is, arraying different guides in a single CRISPR construct, can be used as an effective option to target several regions of the gene, increasing the knockout efficiency [8]. 3.3 Design of Cas12a Multiplexed crRNA Guides with GoldenBraid

Multiplex constructs comprising several crRNA guides can be easily build following a stepwise protocol that can be searched at https:// gbcloning.upv.es/tools/cas12multiplexing/. Once the target sites are defined (Table 2) using the tools and following the criteria described above and in Vazquez-Vilar et al. [8], the protospacer sequences require a so-called domestication step, consisting in the addition of the Cas12a-scaffold sequence (50 -TAATTTCTACTAAGTGTAGAT-30 ), plus the standard GB overhangs at 50 (prefix) and 30 (suffix) ends of the whole polycistronic fragment. This allows the synthetic fragment to proceed through the next cloning steps in the multiplexing GB scheme (Fig. 1). The domestication step is software assisted at the site https:// gbcloning.upv.es/tools/cas12multiplexing_domestication/ by the Multiple Cas12a gRNA domesticator (tM12D) tool. The first decision to be taken is the number of crRNA guides sequences that will be included on the multiplexing array. In this example we will choose the LbCas12a Multiplexing 4X option, to generate a polycistronic array with four gRNAs (see Note 6). After filling in the protospacers sequences in the web form, the tM12D tool will produce as output a “domesticated” DNA sequence to be synthesized (between 100 and 500 bp in size, depending on the multiplexing number) (see Note 7). This output polycistronic array sequence, named also (protospacer-scaffold)n (n ¼ number of repetitions), also contains a Cas12a-scaffold and a Poly-T tail

GB-Adapted CRISPR/Cas12a-Based Gene Editing in N. benthamiana

201

Fig. 1 Schematic representation of multiplexing GoldenBraid DNA assembly. Three consecutive steps of restriction–ligation reactions are required to have a final construct comprising the polycistronic gRNA

202

Beatriz Gonza´lez et al.

situated at the 30 end of the last protospacer included. This is required for the correct expression and cleavage of each guide from the polycistronic array. Note that contrary to Cas9 multiplexing design [4], due the endoribonuclease activity of Cas12a, no interspaced tRNA molecules are required for the correct processing of the protospacer in Cas12a polycistronic arrays [8]. All GB parts generated and used in this example are listed in Table 3. 3.4 Cloning the (ProtospacerScaffold)n in pUPD2 to Generate a Level 0 Part

To proceed with the assembly, first it is necessary to clone the synthetic fragment (Subheading 3.3; 4X polycistronic gRNA array in this example) into the pUPD2 (the GB level 0 destination vector), using a GB restriction–ligation reaction (Fig. 1). 1. Set the single pot reaction up by mixing 60 femtomoles of the synthesized product (or 20 femtomoles if the synthetic fragment was ordered as a plasmid, see Note 7), 20 femtomoles of pUPD2, 8 U of BsmBI, 3 U of T4 ligase, 1.5 μL of BSA (1 mg/mL), and 1.5 μL of 10X T4 ligase buffer in a 15 μL reaction (see Note 8). 2. Incubate the BsmBI restriction–ligation reaction in a thermocycler with the following program: 37 C10min+35(37 C3min+16 C4min)+50 C10min + 80  C  10 min (see Note 9).

Table 3 List of GB plasmids generated and used for this protocol Name

GB ID

Level Category

pUPD2_4X cRNAs-TFL1-FT-XT1-CBP

GB2870 0

CRISPR polycistronic 4X

pUPD2_U6-26:LbCas12aDR

GB1443 0

PROM DPolIII+DRCas12

α1_4X gRNA expression cassette- TFL1-FT-XT1-CBP GB2871 1

TU

α2_P35S:LbCas12a:T35S

GB3720 1

TU

Ω1 _4XgRNA expression cassette_ LbCas12a

GB3058 2

Module

α2_P35S:P19:TNos

GB1203 1

TU

All sequences are accessible at GB cloning website (https://gbcloning.upv.es/search/features/) using the GB database ID

ä Fig. 1 (continued) expression cassette and the LbCas12a expression cassette. In a first reaction, the synthesized (protospacer-scaffold)n fragment is cloned in a level 0 destination vector (pUPD2). In a second reaction, this resulting level 0 part will be assembled with a RNAPolIII promoter (pU6-26) followed by an LbCas12a scaffold (Scf) into an alpha destination vector. In a final step, this level 1 construct generated will be assembled into an omega destination vector together with an LbCas12a expression cassette, to have the final level 2 construct that will be used for plant gene editing. GB-adapted overhangs: A: CTCG, B: TGAG, 0: GGAG, 1: AGAT, 2: CGCT, 3: GGAG, 4: GTCA, 5: CGCT. LB: Left-border and RB: Right border, for T-DNA insertion by agrobacterium-mediated transformation. P35S: CaMV 35S promoter. T35S: CaMV 35S terminator

GB-Adapted CRISPR/Cas12a-Based Gene Editing in N. benthamiana

203

3. Mix 3 μL (1–5 μL) of the reaction with 50–100 μL of Mix&Go cells previously thawed on ice. Let it stand on ice for 5 min and recover the cells by adding 300 μL of SOC. Shake the 1.5 mL tube for 1 h at 37  C and spread the whole volume in LB agar/ chloramphenicol (34 μg/mL)/IPTG (0.4 mM)/X-Gal (40 μg/mL) petri dish to select positive clones. pUPD2 vector confers resistance to chloramphenicol. Incubate the plate overnight (o/n) at 37  C (see Note 10). 4. Next day, there will be colonies transformed with the intact pUPD2 vector (distinguished by their blue color) and those transformed with the polycistronic array construct cloned in the pUPD2 (distinguished by their white color). Pick three white colonies into 3 mL of LB/chloramphenicol and grow the cultures o/n in a shaker (220 rpm approx.) at 37  C. 5. Isolate the plasmid from the cultures of each transformed clone using the E.Z.N.A. Miniprep Kit, following the manufacturer’s instructions. Perform a restriction analysis of the purified plasmids to confirm the presence of the DNA insert into the destination vector with the visualization of the expected restriction bands. Several software tools (e.g., Benchling) can be used to select the best enzyme for each specific plasmid sequence (see Note 11). 6. Sequence the DNA insert of the plasmid of one of the clones showing the expected restriction bands with a primer that anneals on the pUPD2 vector to confirm the sequence of the synthesized fragment (see Note 3). 7. Make glycerol stock of the culture of the double-confirmed positive clone, by adding glycerol to a final concentration of 20–25% and storing at
80  C, to keep a stock of this level 0 part for further experiments (see Note 12). 3.5 Level 1 Polycistronic crRNA Guide Expression Cassette Assembly

The next GB cloning step consists on a restriction–ligation reaction to assembly the final polycistronic gRNA expression cassette in a GB level 1 destination vector (Fig. 1) (see Note 13). The level 0 (protospacer-scaffold)n part from the Subheading 3.4 is assembled with another level 0 part comprising an RNAPolIII promoter and the first 50 end LbCas12a-scaffold. Using the Multiple Cas12a gRNA assembler (tM12A) tool (https://gbcloning.upv.es/tools/ cas12multiplexing_assembly/), the in silico sequence assembly of the resulting level 1 transcriptional unit (TU) part is generated [8]. 1. Set the restriction–ligation reaction up by mixing 20 femtomoles of pUPD2_U626:LbCas12aDR (GB1443), 20 femtomoles of pUPD2_LbCas12aMultiplexing4X (GB2870 in this example), 20 femtomoles of the pDBG3α1 destination vector (pDGB3_alpha1), 8 U of BsaI, 3 U of T4 ligase, 1.5 μL of BSA (1 mg/mL) and 1.5 μL of 10X T4 ligase buffer in a 15 μL reaction (see Note 8).

204

Beatriz Gonza´lez et al.

2. Incubate the BsaI restriction–ligation reaction in a thermocycler with the following program: 37 C10min+35(37 C3min+16 C4min)+50 C10min + 80  C  10 min (see Note 9). 3. Mix 3 μL (1–5 μL) of the reaction with 50–100 μL of Mix&Go cells previously thawed on ice. Let it stand on ice for 5 min and recover the cells by adding 300 μL of SOC. Shake the 1.5 mL tube for 1 h at 37  C and spread the whole volume in LB agar/ kanamycin (50 μg/mL)/IPTG (0.4 mM)/X-Gal (40 μg/mL) petri dish to select positive clones. pDGB3_alpha vectors confer resistance to kanamycin. Incubate the plate o/n at 37  C (see Note 10). 4. Next day, there will be colonies transformed with the intact pDGB3α1 vector (distinguished by their blue color) and those transformed with the gRNA expression cassette (distinguished by their white color). Pick three white colonies into 3 mL of LB/kanamycin and grow the cultures o/n in a shaker (220 rpm approx.) at 37  C. 5. Isolate the plasmid from the cultures of each transformed clone using the E.Z.N.A. Miniprep Kit, following the manufacturer’s instructions. Perform a restriction analysis of the purified plasmids to confirm the presence of the insert into the destination vector with the visualization of the expected restriction bands (see Notes 14 and 15). 6. Make glycerol stock of the culture of the confirmed positive clone, by adding glycerol to a final concentration of 20–25% and storing at
80  C, to keep a stock of this level 1 part for further experiments (see Note 12). 3.6 Final T-DNA Expression Vector Assembly (Level 2 Part)

Once the level 1 TU part comprising the polycistronic gRNA expression cassette is generated (Subheading 3.5), the last GB cloning step is set to combine this with a second TU comprising the LbCas12a expression cassette (other level 1 part) in a GB omega-level destination vector (Fig. 1). Using the GB Binary Assembler tool (https://gbcloning.upv.es/do/bipartite/), the in silico final sequence assembly (namely a level 2 module) is ready to be used in CRISPR/Cas12a-mediated gene editing [8]. It should be mentioned that this software tool allows also the binary assembly with other gRNA expression TUs (single or polycistronic) or modules, such as selection markers for stable transformation experiments (e.g., kanamycin resistance gene (GB0226) or DsRed fluorescent protein (GB0360)), to produce increasingly complex multigene structures (see Note 16). 1. Set the restriction–ligation reaction up by mixing 20 femtomoles of pDGB3α2_ P35S:LbCas12a:T35S (GB3720), 20 femtomoles of pDGB3α1_U626:LbCas12aMultiplexing4X (GB2871 in this example), 20 femtomoles of the pDBG3Ω

GB-Adapted CRISPR/Cas12a-Based Gene Editing in N. benthamiana

205

destination vector (i.e., pDGB3_omega1 in this example), 8 U of BsmBI, 3 U of T4 ligase, 1.5 μL of BSA (1 mg/mL), and 1.5 μL of 10X T4 ligase buffer in a 15 μL reaction (see Note 8). 2. Incubate the BsmBI restriction–ligation reaction in a thermocycler with the following program: 37 C10min+35(37 C3min+16 C4min)+50 C10min + 80  C  10 min (see Note 9). 3. Mix 3 μL (1–5 μL) of the reaction with 50–100 μL of Mix&Go cells previously thawed on ice. Let it stand on ice for 5 min and recover the cells by adding 300 μL of SOC. Shake the 1.5 mL tube for 1 h at 37  C and spread the whole volume in LB agar/ spectinomycin (50 μg/mL)/IPTG (0.4 mM)/X-Gal (40 μg/ mL) petri dish to select positive clones. pDGB3_omega vectors confer resistance to spectinomycin. Incubate the plate o/n at 37  C (see Note 10). 4. Next day, there will be colonies transformed with the intact pDGB3Ω vector (distinguished by their blue color) and those transformed with the module comprising the LbCas12a TU and the gRNA expression cassette (distinguished by their white color). Pick three white colonies into 3 mL of LB/spectinomycin and grow the cultures o/n in a shaker (220 rpm approx.) at 37  C. 5. Isolate the plasmid from the cultures of each transformed clone using the E.Z.N.A. Miniprep Kit, following the manufacturer’s instructions. Perform a restriction analysis of the purified plasmids to confirm the presence of the correct assembly fragment into the destination vector, with the visualization of the expected restriction bands (see Notes 15 and 17). 6. Make glycerol stock of the culture of the confirmed positive clone, by adding glycerol to a final concentration of 20–25% and storing at
80  C, to keep a stock of this level 2 part (GB3058 in this example) for further experiments (see Note 12). 3.7 Transient Expression in N. benthamiana Leaves

The efficiency of CRISPR/Cas12a-mediated gene editing, using the final LbCas12a/Multiplexing gRNA array expression vector generated following Subheadings 3.1–3.6, is assessed in Agrobacterium tumefaciens-mediated N. benthamiana transient expression experiments to verify the mutation efficiency of each guide targeted to knock-out the genes of interest. 1. Transform 200 ng of the plasmid GB3058 (cloned in Subheading 3.6) into a 50–100 μL homemade electrocompetent cells aliquot of A. tumefaciens strain GV3101 C58C1 (see Note 1) by electroporation at 1440 V. Collect the cells in 500 μL LB and incubate for 2 h at 28  C on a shaker (190 rpm). Spread 30–50 μL on an LB agar/spectinomycin (50 μg/mL)/rifampicin (50 μg/mL) petri dish. Incubate the plate at 28  C for

206

Beatriz Gonza´lez et al.

2 days. In parallel, transform also into C58C1 cells the GB1203 level 1 TU part (incubate on an LB agar/kanamycin (50 μg/ mL)/rifampicin (50 μg/mL) petri dish) containing the silencing suppressor P19 of TBSV. This plasmid is necessary to increase the transient expression levels of our construct in coinfiltration assays. 2. Pick two colonies of C58C1 transformed with GB3058 into 5 mL of LB/spectinomycin/rifampicin and two colonies of C58C1 transformed with GB1203 into 5 mL of LB/kanamycin/rifampicin. Grow the cultures in a shaker at 28  C for 2 days (190 rpm), to prepare the precultures of both constructs. 3. Isolate the plasmid from each culture using QIAprep Kit, following the manufacturer’s instructions. Perform a restriction analysis of the purified plasmids (GB3058) to confirm the presence of the correct insert. The same restriction enzyme used in the Subheading 3.6, step 5 can be used. 4. Inoculate 5 mL of LB/spectinomycin/rifampicin with 20–50 μL of the GB3058 preculture and 5 mL of LB/kanamycin/rifampicin with 20–50 μL of the GB1203 preculture. Grow the cultures o/n in a shaker at 28  C (190 rpm). 5. Pellet the bacterial cells by centrifugation at 8000  g (or 4500 rpm) for 10 min. 6. Remove the supernatant (by inversion) and resuspend the pellet in 5 mL of agroinfiltration buffer (see Note 4). Incubate for 2 h at RT in a horizontal rolling mixer in dark (tubes can be wrapped with aluminum foil). 7. Measure the optical density (OD) at 600 nm for each culture. Dilute the cell suspension with agroinfiltration buffer to a final OD600nm 0.1 in a final volume of 15 mL. 8. Mix the agrobacterium cultures transformed with each plasmid GB3058 (CRISPR/Cas12a construct) and GB1203 (P19) in a 1:1 ratio to perform the coinfiltration. 9. Infiltrate with the Agrobacterium culture the leaf intercellular spaces of three or four fully expanded leaves of a 4–5-week-old N. benthamiana plant. Use a 1 mL needle-free syringe to exert a pressure through the abaxial surface of the leaf, holding the adaxial surface with the finger. More than one puncture will be necessary to infiltrate the whole leaf surface (Fig. 2). 3.8 Functional Validation of the Generated Construct for Multiplexing Gene Editing

Following the agrobacterium-mediated transient expression, the presence of mutations (indels) in the target genes due to effect of Cas12a and the designed crRNA guides is analyzed (Fig. 2). Here, we show the functionality of the four crRNA guides (assembled as part of a 4X gRNA expression cassette) targeting four loci of the N. benthamiana genome (Fig. 3), by determining the editing efficiency as described in the next steps.

Fig. 2 Schematic representation of a functional validation assay of gene editing with transient transformation in N. benthamiana. N. benthamiana plants are agrobacterium-mediated transformed by syringe-infiltration with the T-DNA expression vector, ready to be used in CRISPR/Cas12a-based gene editing. Agroinfiltrated leaf-disks are collected to perform the genomic DNA extraction, and the PCR amplification of the regions where the target sites are defined. The editing efficiencies of each guide in the target genes are analyzed using the algorithms TIDE (http://shinyapps.datacurators.nl/tide/) and ICE (https://ice.synthego.com/#/). (Figure created with Biorender.com)

208

Beatriz Gonza´lez et al.

Fig. 3 Editing efficiency analysis of each crRNA guide used in the 4X polycistronic gRNA expression cassette for TFL1, FT, XT1 and CBP genes. Table shows the average value of the edition frequencies for each target gene, from three independent agroinfiltrated leaves, analyzed by both algorithms. Default parameters were used for the TIDE software tool. Bars represent mean  SD, n ¼ 3 independent replicates

1. Collect three leaf disks from each agroinfiltrated leaf at 5 days post infiltration (dpi), using a 1.5 cm cork-borer (75 mg of tissue approx.), in a 1.5 mL tube and freeze in liquid N2 (see Note 18). 2. Proceed with the genomic DNA extraction using the CTAB protocol. First, prepare enough volume of CTAB Buffer for all samples and keep at RT (see Note 5). 3. Grind the leaf-disks adding one steel bead per tube and using a Mixer Mill M400 (Retsch) for 45 s or 1 min at 30 Hz (cycles/s) (once or more times depending on the quantity of tissue), avoiding the material defrosting. For 75 mg (approx.) of sample (frozen powder), add 600 μL of CTAB buffer (see Note 19). For a better extraction, a small pestle can be used for further grinding of the tissue together with the CTAB Buffer. Vortex and keep on ice until all samples are processed. 4. Incubate for 45 min at 65  C. Mix by inversion every 10–15 min to keep the plant material in suspension. 5. Add 600 μL of chloroform:isoamyl alcohol (24:1). Vortex. 6. Centrifuge for 15 min at 14,000  g (or 13,000 rpm) at RT. 7. Take the upper phase and transfer to a clean 1.5 mL tube, taking care not to take any of the interphase or chloroform phases. For the samples where 600 μL CTAB Buffer were added, a volume around 400–500 μL is usually recovered. This and the next steps will be performed on ice, unless noted otherwise.

GB-Adapted CRISPR/Cas12a-Based Gene Editing in N. benthamiana

209

8. If the upper phase is not clear, repeat steps 5–7. 9. Add 1 volume of isopropanol. Mix by inversion. Precipitate on ice for 5 min. 10. Centrifuge for 10 min at 14,000  g (or 13,000 rpm) at 4  C. 11. Remove the supernatant (with pipet) and wash the pellet with 600 μL of 80% ethanol (ice-cold). Vortex. 12. Centrifuge for 5 min at 14,000  g (or 13,000 rpm) at 4  C. 13. Remove supernatant (with pipet). Spin (1 min at 14,000  g (or 13,000 rpm)) to remove any rest of ethanol and let the pellet dry at RT for 15–30 min. 14. Resuspend the pellet in 15–50 μL sterile H2OmQ (or TE buffer), according to the initial quantity of material. Let it stand at RT for about 30 min, mix well (gently) and let at RT for additional 30 min; or let it stand o/n at 4  C and then mix well. Keep your DNA extractions at
20  C for next steps. 15. Proceed with PCR amplification of the genomic DNA region of the genes of interest containing the target site of the crRNA guides selected in Subheading 3.2. Design specific primers, upstream and downstream of each gRNA, for each specific gene to amplify the regions where Cas12a-derived indels will be found (see Note 20). For the example shown here, the primers are listed in Table 4. Set the PCR reaction up by mixing 100 ng of genomic DNA, 1 U of MyTaq DNA polymerase, 5 μL of 5X MyTaq Reaction Buffer, 0.2 mM forward primer and 0.2 mM reverse primer in a 25 μL reaction; and follow the program: Table 4 List of primers used for PCR amplification Name

Sequence (50 -30 )

Tm

NbTFL1-3.1_FW

GCTTACTGTGTCCTGTATAAAACTG

54

NbTFL1-3.1_RV

GTAATCTCATATGTGGTATCGGAG

52

NbTFL1-14.1_FW

GAGTAGATATTCCAAGACAGCAAC

53

NbTFL1-14.1_RV

CACTGGAAAACCTCTGTAACAAC

54

NbFT_FW

CTAGAAAACCTATGGCTATAAGGG

52

NbFT_RV

GTTCTCGAGAGGTATAATATAGGC

52

NbXT1_FW

AACCACTTTTCCTCGTCGGAAA

56

NbXT1_RV

TAACTATTCAACTAAAGCTTCAAACAG

52

NbCBP-AB_FW

TGTCTAGACTGGTGCATTACTTC

54

NbCBP-AB_RV

GTTGCCAAAAGGATCACTCAAAT

54

210

Beatriz Gonza´lez et al.

95 C3min+35(95 C15s+Tm15s+72 C40s)+50 C10 min + 80  C  10 min (see Note 21). Verify the correct PCR amplification band using 1% agarose gel electrophoresis. 16. Purify the PCR amplification products using ExoSAP-IT, following the manufacturer’s indications. A PCR volume of 5 μL is enough (see Note 22). 17. Sequence the PCR amplified band using one of the primers used in step 15 for each pair. 18. Proceed with the analysis of edition efficiency in each region for each crRNA guide. We recommend using a software tool for Tracking of Indels by DEcomposition (TIDE) analysis (i.e., http://shinyapps.datacurators.nl/tide/) or Inference of CRISPR Edits (ICE) (i.e., https://ice.synthego.com) to analyze the differences in the sequences between the samples from leaves transformed with gene editing construct and the control samples. After upload the DNA sequence files, both algorithms accurately reconstruct the spectrum of indels from the sequence traces. The website tools report the identity and frequency of each detected indel in each analyzed region and the frequencies of edition (indel percentage) of each analyzed gRNA. The indel percentage indicates the levels of editing in each agroinfiltrated sample. The value of KO-Score is also reported with ICE algorithm, which indicates the percentage of edition of an indel that could produce a knock-out of the gene (Fig. 2). Even though these tools are designed for Cas9based indels, both algorithms can provide valid editing efficiencies for Cas12a, and both provide similar trends in indel size and frequency results, although absolute values often differ. Besides, they are very robust with an indel detection sensitivity of 2–4%. However, with ICE analysis tool, the selection of the software parameters is not required, and a higher number of samples is easier to process [18, 19]. We have noticed that crRNA guides with low editing efficiencies show lower indel frequencies with ICE algorithm in comparison to TIDE. These discrepancies are probably due to the complexity of the mutations present in the samples obtained from transient analysis. Applying both algorithms, we show in this example the calculated editing efficiency of each designed crRNA guide for the corresponding target gene (two TFL1 homologues, FT, XT1 and CBP genes) in a N. benthamiana transient expression experiment (Fig. 3). According to our previous experience in functional validation assays based on transient expression, editing efficiency levels usually range between 0 and 30%. gRNAs with indel percentage above 10% (TIDE values) are considered as valid for undertaking stable plant transformation, which usually results in heterozygous or even biallelic mutations.

GB-Adapted CRISPR/Cas12a-Based Gene Editing in N. benthamiana

4

211

Notes 1. A protocol is provided to prepare A. tumefaciens homemade electrocompetent cells. Day 1: Prepare a fresh plate of agrobacterium (from a glycerol stock) by streaking on an LB agar/ rifampicin (50 μg/mL) petri dish and incubate at 28  C for 2 days. Day 3: Prepare a preculture of agrobacterium by picking a single colony in 5 mL of LB/rifampicin and incubate o/n on a shaker (approx. 190 rpm) at 28  C. Additionally, prepare 1 L of 1 mM HEPES buffer (pH 7.4) and store at 4  C. Day 4: Inoculate 500 mL of LB/rifampicin (in a 2 L flask) with 270 μL of the preculture, taking 2 mL of the medium prior to adding the cells to use as blank next day (for OD600 measurements). Grow o/n (for about 12–13 h) in a shaker (190 rpm) at 28  C until the final OD600 is around 0.5. Day 5: Transfer the flask to ice and cool the cells for 15–30 min. Then, distribute the culture in two centrifuge bottles (250 mL to each, or adjusting the volume for each bottle until the weight of both is exactly the same) and centrifuge the cells at 4000  g (or 5000 rpm) for 15 min at 4  C. Remove the supernatant (by inversion) and resuspend the pellet of each bottle in 250 mL of ice-cold 1 mM HEPES by shaking continuously on ice (gently, do not use a pipet). A small volume of HEPES (around 10 mL) can be added initially for the resuspension of the pellet, and then continue adding until the final volume is reached (250 mL). Then, centrifuge both bottles at 4000  g for 15 min at 4  C. Remove the supernatant again and resuspend the pellet of each bottle in 125 mL of 1 mM HEPES, as before. Repeat the centrifugation step and resuspend each pellet in 5 mL of 1 mM HEPES. Transfer the volume to 13 mL tubes and repeat the centrifugation step. Remove the supernatant and resuspend the final pellets in 1–1.5 mL of ice-cold 10% glycerol. Aliquot into 1.5 mL tubes (around 50 μL each) and freeze directly in liquid N2. Store at
80  C. 2. We recommend preparing sterile stock solutions (by filtration) of 2 M Mg Stock (1 M MgCl2 and 1 M MgSO4) and 2 M glucose. 3. Recommended sequencing primers for pUPD2 inserts are: 50 -GCTTTCGCTAAGGATGATTTCTGG-30 (Forward) and/or 50 -CAGGGTGGTGACACCTTGCC-30 (Reverse). 4. We recommend preparing previously both sterile stock solutions (by autoclaving) of 100 mM MES (pH 5.6, adjusted with KOH) and 1 M MgCl2. The agroinfiltration day, prepare a fresh concentrated acetosyringone solution (200 mM, dissolved in DMSO in dark) for the volume required of agroinfiltration buffer.

212

Beatriz Gonza´lez et al.

5. A stock buffer (stored at RT) can be prepared with NaCl, EDTA and Tris–HCl. Add the rest of components just before use, adjusting quantities to the final required volume of CTAB buffer. Notice here that CTAB (added in the first place) takes long time to dissolve. The stock buffer can be warmed up to 50  C to accelerate the process. 6. GB cloning CRISPR tools support assemblies of polycistronic arrays with up to six crRNA guides for LbCas12a only. 7. Target sequences will be assembled in the final construct in the same order that they were typed in the web form tool. GB accepts Cas12a target sequences ranging from 20 to 23 nt. Synthetic DNA fragments of up to two crRNAs guides can be purchased as gBlocks from IDT (https://eu.idtdna.com/ pages/products/genes-and-gene-fragments/doublestranded-dna-fragments/gblocks-gene-fragments). DNA fragments with more crRNAs guides (3–6), therefore including more repetitions of the Cas12a scaffold, can be purchased from GenScript (https://www.genscript.com/gene_synthesis. html?src¼pullmenu). The final sequence of the DNA fragment synthesized for this example with four crRNAs guides is 50 -TGATCGTCTCGCTCGAGATACTGTGGGACTGAATGAATCTAATTTCTACTAAGTGTAGATCAAGATCTATTGGCCTAAGATAATTTCTACTAAGTGTAGATTATGTAGGTGTATTTGGAATTAATTTCTACTAAGTGTAGATGAGGACAAACTACATCCAGGTAATTTCTACTAAGTGTAGATTTTTTTTTCGCTTGAGAGAGACGATCG-30 . 8. Alternatively, GB reactions can be performed in restriction enzyme buffer. In this case the reaction needs to be supplemented with 1 mM ATP for ligase activity. The formula to calculate the volume of a DNA preparation that has to be used to add N femtomoles is: Volume (μL) ¼ (N femtomoles  10
15  (bp size DNA fragment)  617.96  109)/ DNA concentration (ng/μL). 9. GB restriction–ligation reactions can be programmed up to 50 cycles to increase effectiveness of complex cloning reactions (i.e., including more than 4 GB DNA parts). If the program has to be stopped between the cycles number 35 and 50, it is recommended doing it during the restriction step, to decrease the number of blue colonies. 10. In the transformation step, LB can also be used for recovering the cells, instead of SOC. However, if this is done, the efficiency of transformation could be lower. In the shaking incubation step, 30 min is also enough. 11. The recommended restriction enzymes to verify positive clones in pUPD2 are BsaI or NotI.

GB-Adapted CRISPR/Cas12a-Based Gene Editing in N. benthamiana

213

12. Before the bacterial plasmid isolation, drop 3 μL from the overnight cultures of the clones in LB agar/antibiotic petri dish and incubate o/n at 37  C. Store then at 4  C. Once positive clones are double-confirmed, subculture the aliquot from the petri dish in 3 mL of LB/antibiotic, to make glycerol stocks. 13. If different multiplexing gRNA expression cassettes will be assembled, the different level 0 polycistronic gRNA array parts can be cloned in the pUPD2 in parallel following the same protocol of Subheading 3.4. 14. The recommended restriction enzyme to verify positive clones in alpha 1 vector is EcoRI. 15. The confirmation of the correct assembly of the insert in alphaand omega-destination vectors by sequencing is not necessary in principle, since no PCR/DNA synthesis step is involved. 16. Due to this binary assembly, only combinations alpha1–alpha2 (to insert in an omega vector) or omega1–omega2 (to insert in an alpha vector) are allowed. 17. The recommended restriction enzyme to verify positive clones in omega 1 vector is NcoI. 18. In the sample collection step, avoid the leaf veins, which are a source of variability due to ineffective infiltration. 19. Alternatively, leaf disks can be directly grinded (fresh or frozen) in CTAB buffer using a small pestle. For 600 μL of CTAB Buffer, the material may range between 50 and 150 mg, but for lower quantity of sample, adjust the volume. 20. Primers designed for the amplification of the region where target guides are found, must be gene-specific, avoiding the amplification of possible homologous genes. 21. For PCR reaction program, the melting temperature (Tm) should be adjusted for each pair of primers, considering the Tm of each primer (Table 4). In this example, the Tm used for all PCRs was 55  C. Additionally, DNA polymerase extension time should be adjusted according to the amplicon length, following DNA polymerase manufacturer’s instructions. 22. For the purification of low number of PCR products or when agarose gel band purification is required, the NucleoSpin PCR cleanup kit can be used instead of ExoSAP-IT, using the whole PCR volume for the purification.

Acknowledgments This work was funded by H2020 EU projects 760331 Newcotiana and 774078 Pharma-Factory. M.V. is recipient of a Generalitat Valenciana and Fondo Social Europeo postdoctoral grant.

214

Beatriz Gonza´lez et al.

References 1. Molina-Hidalgo FJ, Vazquez-Vilar M, D’Andrea L et al (2021) Engineering metabolism in Nicotiana species: a promising future. Trends Biotechnol 39:901. https://doi.org/ 10.1016/j.tibtech.2020.11.012 2. Marillonnet S, Gru¨tzner R (2020) Synthetic DNA assembly using Golden Gate cloning and the hierarchical modular cloning pipeline. Curr Protoc Mol Biol 130:e115. https://doi. org/10.1002/cpmb.115 3. Vazquez-Vilar M, Quijano-Rubio A, Fernandez-Del-Carmen A et al (2017) GB3.0: a platform for plant bio-design that connects functional DNA elements with associated biological data. Nucleic Acids Res 45: 2196–2209. https://doi.org/10.1093/nar/ gkw1326 4. Vazquez-Vilar M, Bernabe´-Orts JM, Fernandez-Del-Carmen A et al (2016) A modular toolbox for gRNA-Cas9 genome engineering in plants based on the GoldenBraid standard. Plant Methods 12:10. https://doi. org/10.1186/s13007-016-0101-2 5. Bandyopadhyay A, Kancharla N, Javalkote VS et al (2020) CRISPR-Cas12a (Cpf1): a versatile tool in the plant genome editing tool box for agricultural advancement. Front Plant Sci 11:1589. https://doi.org/10.3389/fpls. 2020.584151 6. Bernabe´-Orts JM, Casas-Rodrigo I, Minguet EG et al (2019) Assessment of Cas12amediated gene editing efficiency in plants. Plant Biotechnol J 17:1971–1984. https:// doi.org/10.1111/pbi.13113 7. Schindele P, Puchta H (2020) Engineering CRISPR/LbCas12a for highly efficient, temperature-tolerant plant gene editing. Plant Biotechnol J 18:1118–1120. https://doi.org/ 10.1111/pbi.13275 8. Vazquez-Vilar M, Garcia-Carpintero V, Selma S et al (2021) The GB4.0 platform, an all-inone tool for CRISPR/Cas-based multiplex genome engineering in plants. Front Plant Sci 12:689937. https://doi.org/10.3389/ fpls.2021.689937 9. Concordet J-P, Haeussler M (2018) CRISPOR: intuitive guide selection for CRISPR/ Cas9 genome editing experiments and screens. Nucleic Acids Res 46:W242–W245. https:// doi.org/10.1093/nar/gky354

10. Kim HK, Song M, Lee J et al (2017) In vivo high-throughput profiling of CRISPR–Cpf1 activity. Nat Methods 14:153–159. https:// doi.org/10.1038/nmeth.4104 11. Zhu H, Liang C (2019) CRISPR-DT: designing gRNAs for the CRISPR-Cpf1 system with improved target efficiency and specificity. Bioinformatics 35:2783–2789. https://doi.org/ 10.1093/bioinformatics/bty1061 12. Paul B, Montoya G (2020) CRISPR-Cas12a: functional overview and applications. Biom J 43:8–17. https://doi.org/10.1016/j.bj. 2019.10.005 13. To´th E, Varga E´, Kulcsa´r PI et al (2020) Improved LbCas12a variants with altered PAM specificities further broaden the genome targeting range of Cas12a nucleases. Nucleic Acids Res 48:3722–3733. https://doi.org/ 10.1093/nar/gkaa110 14. Kim HK, Min S, Song M et al (2018) Deep learning improves prediction of CRISPR-Cpf1 guide RNA activity. Nat Biotechnol 36: 239–241. https://doi.org/10.1038/nbt. 4061 15. Naim F, Shand K, Hayashi S et al (2020) Are the current gRNA ranking prediction algorithms useful for genome editing in plants? PLoS One 15:e0227994. https://doi.org/10. 1371/journal.pone.0227994 16. Zetsche B, Gootenberg JS, Abudayyeh OO et al (2015) Cpf1 is a single RNA-guided endonuclease of a class 2 CRISPR-Cas system. Cell 163:759–771. https://doi.org/10.1016/j. cell.2015.09.038 17. Safari F, Zare K, Negahdaripour M et al (2019) CRISPR Cpf1 proteins: structure, function and implications for genome editing. Cell Biosci 9: 36. https://doi.org/10.1186/s13578-0190298-7 18. Conant D, Hsiau T, Rossi N et al (2022) Inference of CRISPR edits from Sanger trace data. CRISPR J 5(1):123–130. https://doi.org/10. 1089/crispr.2021.0113 19. Bennett EP, Petersen BL, Johansen IE et al (2020) INDEL detection, the ‘Achilles heel’ of precise genome editing: a survey of methods for accurate profiling of gene editing induced indels. Nucleic Acids Res 48:11958–11981. https://doi.org/10.1093/nar/gkaa975

Chapter 13 Advanced Fusion Strategies for the Production of Functionalized Potato Virus X Virions Christina Dickmeis and Ulrich Commandeur Abstract Plant virions are ideal for nanotechnology applications because they are structurally diverse and can selfassemble naturally, allowing for large-scale production in plants by molecular farming. Potato virus X (PVX) is particularly amenable due to the unique properties of its filamentous and flexible capsid, but efficient strategies are required to adapt the surface properties of PVX, such as the attachment of proteins and peptides. This chapter describes the selection and utilization of 2A ribosomal skip sequences, allowing the presentation of heterologous proteins and peptides as N-terminal fusions to the PVX coat protein at different densities. Another strategy for the rapid modification of PVX capsids is the plug-and-display module of the SpyTag/SpyCatcher system. The SpyTag can be presented on the PVX surface, allowing for the attachment of any protein fused to the SpyCatcher sequence. Key words Coat protein, FDMV 2A sequence, Functionalization, Nanoparticles, Plant virus, PVX, SpyTag/SpyCatcher

1

Introduction The molecular self-organization of proteins into higher-order structures has enormous potential for applications at the interface of biology, chemistry and materials science. Such structures can be used as biocompatible and biodegradable building blocks that coordinate functions based on their assembly behavior [1, 2]. Plant viruses are becoming increasingly important in this context due to their precise self-organization into monodisperse particles, providing a sustainable alternative to synthetic nanoparticles. Plant viruses comprise one or more types of coat protein (CP) subunits that assemble with the virus genome to form infectious particles. Modified versions, known as plant virus nanoparticles (VNPs), are used for a variety of biohybrid applications including the display of metal complexes to increase the surface area of batteries [3] or to improve the sensitivity of diagnostic tests

Stefan Schillberg and Holger Spiegel (eds.), Recombinant Proteins in Plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480, https://doi.org/10.1007/978-1-0716-2241-4_13, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

215

216

Christina Dickmeis and Ulrich Commandeur

[4]. They are also used as a presentation system for biological signals in biomaterials, for example, for cell proliferation or cell attachment [5], and as contrast agents carrying dyes for imaging [6]. Plant VNPs are well tolerated and noninfectious in mammals [7, 8], making them attractive as components of biomedical nanomaterials. Furthermore, plant-based expression systems offer sustainable and highly scalable VNP production platforms, an approach known as molecular farming. Potato virus X (PVX) provides unique and attractive features for VNP design because its viral capsid is filamentous and highly flexible. PVX is the type member of the genus Potexvirus in the family Alphaflexiviridae [9] and carries a plus-sense single-stranded RNA genome 6.4 kb in length. The 50 -end has a methylguanosine cap and the 30 -end is polyadenylated. The genome contains five open reading frames (ORFs) encoding the RNA-dependent RNA polymerase (RdRp), three nonstructural proteins (triple gene block, TGB) and the CP [10–12]. The polymerase is the only viral protein required for replication and is translated directly from the genomic RNA. The other ORFs are controlled by subgenomic promoter-like sequences that regulate the production of subgenomic RNAs for translation [13, 14]. The TGB proteins are named according to their sizes (25, 12, and 8 kDa) and, together with the CP (25 kDa), mediate cell-to-cell transport [15, 16]. The genomic RNA is packed into a filamentous and flexible virus particle formed by the CPs [17]. PVX has a helical structure, with the N-termini of the coat proteins presented externally, allowing the display of heterologous proteins and peptides by fusion to the N-terminus of the CP [18, 19]. A fusion is only possible if certain criteria are met [20]. The isoelectric point of the fused peptides must fall within the range 5.2–9.2 [21, 22] or must be adjusted by including a compensatory sequence such as the acidic DEADDAED peptide [23]. Added peptides generally should not exceed 60 amino acids in length [24], although some longer peptide fusions have been successful, including the iLOV marker protein domain with 113 aa [25]. Cell-to-cell movement is inhibited by the presence of too many tryptophan residues [26]. Furthermore, serine and threonine [26, 27] are essential for phosphorylation and glycosylation, thereby stabilizing the particles by creating a surrounding water shell [21, 28]. Finally, the presence of too many positively charged amino acids can make the virus unstable [22], resulting in the selection of compensatory deletion mutants over several passages through a PVX host. The main reasons for problems with the display of larger peptides is steric hindrance, which blocks the assembly of CP fusions. However, this limitation can be overcome by mixing wild-type CPs with the fusion constructs, allowing the fusion proteins to be interleaved with wild-type CPs, a strategy that even works with

Functionalized PVX Virions

217

bulky adducts such as green fluorescent protein (GFP) [19]. Although it is possible to express wild-type and fusion proteins separately, this laborious solution can be circumvented by integrating the 2A sequence linker from a picornavirus such as foot-and-mouth disease virus (FMDV). This causes a ribosomal skip to occur during translation, producing a mixture of the fusion protein, the wild-type CP and the free fusion partner, thus enabling the presentation of larger peptides and whole proteins [19, 29]. FMDV belongs to the family Picornaviridae, which encodes all its proteins in a single ORF [30]. The 2A sequence varies among different picornaviruses, but shares a highly conserved core of 18 amino acids [29]. The ribosomal skip takes place between the C-terminal glycine residue of the 2A peptide and the N-terminal proline residue of the 2B protein [31–33]. The skipping frequency, expressed as the molar ratio of fusion protein to the individual proteins, depends on the sequence and length of the 2A region [31, 34]. The analysis of modified 2A sequences in wheat germ extract and rabbit reticulocyte lysates revealed “cleavage activities” of up to ~99% [32, 35]. The tuning of 2A sequences with respect to the skipping frequency, allows the density of heterologous proteins displayed on the surface of PVX particles to be varied considerably (see Fig. 1a and Table 1). Large proteins can be displayed by ensuring high skipping frequencies and thus a low ratio of fusion protein to CP, whereas the dense display of smaller peptides can be achieved by selecting 2A variants with low skipping frequencies. Another promising strategy for the plug-and-display modification of PVX is the SpyTag/SpyCatcher system, which also allows the presentation of more complex proteins from other heterologous expression systems. The SpyTag/SpyCatcher system achieves simple and rapid bioconjugation by presenting a peptide tag (SpyTag) that forms a specific covalent bond with any protein displaying the cognate SpyCatcher ligand [36]. The system is derived from the CnaB2 domain of the Streptococcus pyogenes fibronectin-binding protein FbaB. This domain is highly stable due to the presence of a single isopeptide bond between lysine and aspartic acid, allowing it to withstand low pH and temperatures up to 100  C [37]. The SpyTag/SpyCatcher system was generated by splitting the CnaB2 domain into the SpyTag peptide (13 amino acids) and SpyCatcher protein (116 amino acids). SpyTag and SpyCatcher form a specific, covalent and irreversible isopeptide bond within minutes after mixing. This bond is extremely stable and can form under various harsh conditions including wide ranges of temperature, pH and buffer compositions [36]. The system has already been used for protein cyclization, hydrogel engineering, antibody engineering, cell labeling and macromolecular assembly [38–43]. The short SpyTag peptide can be fused to the CP of PVX via a linker to adjust the isoelectric point [23]. The PVX CP-SpyTag fusion is capable of

218

Christina Dickmeis and Ulrich Commandeur

Fig. 1 Representation of cloning constructs for different PVX fusion strategies. The viral RNA is transcribed into cDNA and introduced into a plant expression vector under the control of the cauliflower mosaic virus (CaMV) 35S promoter (P35S) and terminator (T35S). The PVX vector codes for the RNA-dependent RNA-polymerase (RdRp), the triple gene block (TGB) proteins (p25, p12, and p8), and the coat protein (CP). The N-terminus of the CP is exposed on the virion surface after particle assembly and allows a fusion of a protein of interest (POI). The viral proteins are produced from subgenomic promoter-like sequences (small arrows). (a) Larger POIs can hinder the assembly and thus need a ribosomal skip sequence like the foot-and-mouth disease virus (FDMV) 2A peptide. The sequence can be adapted to control the ratio of fusion protein to wild-type CP and the free POI. The 2A sequence remains fused to the POI. The fusion product can then be assembled with wild-type CP into a hybrid viral particle. (b) Another possibility for the display of larger proteins is the presentation of the small SpyTag sequence (ST) and a subsequent coupling with a SpyCatcher (SC) fused to the POI, which can be produced in any expression system. This is a simple plug-and-display module allowing the rapid adaption of the POI for multiple downstream applications

Functionalized PVX Virions

219

Table 1 Peptides that facilitate different fusion strategies to the PVX coat protein (CP) Peptide

Amino acid sequence

References

2A10

SRQLNFDLLKLAGDVESQPG

[31]

2A31

SRQLNFDLLKLAGDVESHPG

[31]

2A39

SRQLNFDLLKLAGDVEFNPG

[31]

2A56

SRQLNFDLLKLAGDVQSNPG

[31]

2A70

DLLKLAGDVESNPG

[33]

2A80

SRGACQLLNFDLLKLAGDVESNPG

[33]

2A90

SRQSRLNFDLLKLAGDVESNPG

[31]

2A*

SGSRNFDLLKLAGDVESNPGP

[6]

SpyTag

MAHIVMVDAYKPTKDEADDAEDPGP

[23]

The 2A sequences have different ribosomal skip efficiencies (e.g., 2A10 has a theoretical efficiency of 10% which leads to ~10% free CP and target protein and ~90% fusion product). The 2A* sequence is the reference 2A sequence which was frequently used in PVX vectors. The SpyTag sequence is followed by the DEADDAED sequence to adjust the isoelectric point of the peptide, which is required for successful particle assembly. Any SpyCatcher fusion protein can then be attached to the particle surface

correct self-assembly and systemic infection (see Fig. 1b). Any protein-SpyCatcher fusion can then be covalently linked to the resulting VNP allowing rapid adaptation for a broad range of applications. Since the coupling is performed after VNP formation, the strategy avoids possible interference of the fused protein with assembly.

2

Materials Prepare all solutions with deionized water (ddH2O). A suitable quality (resistivity ~18.2 MΩ cm at 25  C) can be achieved using a membraPure system (Aquintus).

2.1 Genetic Engineering

The following enzymes and cloning materials are used to introduce sequences into viral CPs. 1. Suitable restriction enzymes (e.g., NheI and BspEI). 2. If combinations of restriction enzymes cannot be used in the same buffer, digest plasmid DNA sequentially and purify after each reaction using the MSB® Spin PCRapace Kit (Invitek by STRATEC Biomedical AG) or similar. 3. A DNA polymerase with proofreading activity (e.g., Pfu DNA polymerase, Promega) is used for PCR experiments, and control PCRs are carried out using Taq DNA polymerase (e.g., GoTaq® DNA polymerase, Promega).

220

Christina Dickmeis and Ulrich Commandeur

4. Calf intestinal phosphatase (CIP, NEB). 5. T4 DNA ligase (Promega). 6. 1.2% (w/v) agarose in 1 TAE buffer (40 mM Tris base, 20 mM acetic acid, 1 mM EDTA). 7. 0.3 mg/L ethidium bromide. 8. GeneRuler™ 100 bp plus and 1-kb ladders (Thermo Fisher Scientific). 9. Wizard® SV Gel and PCR Clean-Up System (Promega). 10. Primer containing the SpyTag or 2A sequences including suitable restriction enzyme recognition sites (see Table 2).

Table 2 Primers used for cloning procedures Primer name

Primer 50 ! 30

2A10

ATGTATCCGGATCTAGACAACTTAATTTTGATCTTCTTAAGCTTGCAGGAGATG TTGAATCTCAACCCGGGCCCGCGAGC

2A31

ATGTATCCGGATCTAGACAACTTAATTTTGATCTTCTTAAGCTTGCAGGAGATG TTGAATCTCATCCCGGGCCCGC

2A39

ATGTATCCGGATCTAGACAACTTAATTTTGATCTTCTTAAGCTTGCAGGAGATG TTGAATTTAATCCCGGGCCCGC

2A56

ATGTATCCGGATCTAGACAACTTAATTTTGATCTTCTTAAGCTTGCAGGAGATG TTCAATCTAATCCCGGGCCCGC

2A70

ATGTATCCGGATCTAGAGATCTTCTTAAGCTTGCAGGAGATGTTGAATCTAA TCCCGGGCCCGC

2A80

ATGTATCCGGATCTAGAGGAGCTTGTCAACTTCTTAATTTTGATCTTCTTAAGC TTGCAGGAGATGTTGAATCTAATCCCGGGCCCGC

2A90

ATGTATCCGGATCTAGACAACTTAATTTTGATCTTCTTAAGCTTGCAGGAGATG TTGAATCTAATCCCGGGCCCGC

CX1

TTGAAGAAGTCGAATGCAGC

CX2

CTAGATGCAGAAACCATAAG

CX8

AGCTCTGCTGATGCCGTTGG

M13 rev

ACACAGGAAACAGCTATGAC

M13 uni

GTTGTAAAACGACGGCCAGT

NheI-STCP

AAAGCTAGCGCTCATATTGTTATGGTTGATGCTTATAAGCCTAC TAAGCCCGCGAGCACAACACAGCC

The 2A primers are used to introduce the new 2A sequences into the coat protein (CP) via PCR, amplifying the complete CP with M13 rev as second primer. Primer NheI-ST can be used to add the SpyTag sequence to the cp gene. Primers CX1, CX2, and CX8 can be used for RT-PCR and sequencing

Functionalized PVX Virions

221

11. Escherichia coli DH5α cells (genotype fhuA2 lac(del)U169 phoA glnV44 Φ800 lacZ(del)M15 gyrA96 recA1 relA1 endA1 thi-1 hsdR17). 12. LB medium: 0.5% (w/v) yeast extract, 1% (w/v) NaCl, 0.5% (w/v) tryptone/peptone. 13. LB plates: LB medium with 1.5% (w/v) agar, supplemented with suitable antibiotics depending on plasmid (e.g., 50 μg/ mL kanamycin or 100 μg/mL ampicillin). 14. Plasmids: (I) an infectious PVX cDNA clone (e.g., pCXI for PVX inoculation [44]); (II) suitable vectors for heterologous expression of SpyCatcher fusion products (e.g., pTRAkt-ERHCel12A-SC-His6 [23]); (III) vectors for transient expression in plants in other compartments; and (IV) standard E. coli expression vectors such as pET22b. 15. Pure Yield™ Plasmid Miniprep Kit (Promega) for the isolation of small amounts of DNA, and the Pure Yield™ Midiprep Kit (Promega) for larger amounts. 2.2 PVX Particle Production

1. A pCXI derivative (i.e., infectious PVX cDNA clone yielding virions exposing the fusion construct of choice). 2. 4-week-old Nicotiana benthamiana plants. 3. Celite 545 (Carl Roth).

2.3 Plant Virus Purification

1. 50–100 g of infected plant material (freshly harvested or stored at 80  C). 2. Laboratory blender. 3. Extraction buffer: 0.1 M sodium phosphate buffer (pH 8.0) containing 0.2% (v/v) 2-mercaptoethanol, 10% (v/v) ethanol (cooled). 4. Triton X-100 (Carl Roth). 5. 1 M NaCl with 20% (w/v) PEG: dissolve 58.44 g NaCl, 100 g PEG 6000 (Carl Roth) and 100 g PEG 8000 (Carl Roth), in 1 L ddH2O while stirring. 6. Sucrose gradient: prepare 10% (w/v) and 45% (w/v) sucrose in 0.01 M sodium phosphate buffer (pH 7.2) containing 0.01 M EDTA. Use 12 mL 10% (w/v) and 12 mL 45% (w/v) sucrose solution with a gradient mixer to generate the gradient. 7. 0.05 M sodium phosphate buffer (pH 8.0) containing 1% (v/v) Triton X-100. 8. 0.01 M sodium phosphate buffer (pH 7.2). 9. Miracloth (Merck Millipore). 10. Ultracentrifuge XPN-80 (Beckman Coulter) with swingbucket rotors SW 41 Ti and SW 32 Ti.

222

Christina Dickmeis and Ulrich Commandeur

11. Ultra-Clear centrifuge tubes: 14  89 mm for SW 41 Ti and 25  89 mm for SW 32 Ti (Beckman Coulter). 12. Benchtop centrifuge. 13. Spectrophotometer. 2.4 Protein Expression

1. Agrobacterium tumefaciens GV3101 pMP90RK (DSMZ) transformed with pTRAkt-POI-SpyCatcher-His6 (see Note 1).

2.4.1 Transient Expression in Plants

2. YEP medium: 0.5% (w/v) beef extract, 0.1% (w/v) yeast extract, 0.5% (w/v) peptone, 0.5% (w/v) sucrose, 2 mM MgSO4 (pH 7.4). 3. YEP plates: 1.5% (w/v) agar in YEP medium supplemented with 50 μg/mL rifampicin, 100 μg/mL carbenicillin and 50 μg/mL kanamycin. 4. 1 M 2-(N-morpholino)-ethanesulfonic acid (MES). Adjust pH to 5.6 with 1 M KOH. 5. 40% (w/v) glucose. 6. 200 mM acetosyringone dissolved in DMSO. 7. 2 infiltration medium: 10% (w/v) sucrose, 0.36% (w/v) glucose, 0.86% (w/v) Murashige and Skoog (MS) salts (pH 5.6). 8. N. benthamiana plants, 4–6 weeks old. 9. 1-mL needleless syringe.

2.4.2 Expression in E. coli

1. Suitable E. coli expression strain (see Note 2) such as BL21 (DE3) star (Thermo Fisher Scientific): FompT hsdSB (rB, mB) galdcmrne131 (DE3). 2. LB plates (see Subheading 2.1). 3. TB medium: 24 g/L yeast extract, 12 g/L tryptone, 4 mL/L glycerol. Dissolve components in 900 mL ddH2O and sterilize. After cooling to room temperature, add 100 mL (10) phosphate buffer (0.17 M KH2PO4, 0.72 M K2HPO4). 4. 1 M isopropyl β-D-1-thiogalactopyranoside (IPTG).

2.5 Protein Purification

1. Ni-NTA agarose (Qiagen) columns prepared according to the manufacturer’s instructions. 2. Phosphate-buffered saline (1 PBS): 137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KH2PO4 (pH 7.4). 3. 10 mM imidazole in PBS. 4. 150–300 mM imidazole solutions in PBS. 5. 50 mg/mL lysozyme stock solution. 6. 20% (v/v) Triton X-100. 7. Lysis buffer: 50 mM Tris–HCl (pH 8.5), 300 mM NaCl, 10% glycerol, 10 mM imidazole.

Functionalized PVX Virions

223

8. Wash buffer: 50 mM Tris–HCl (pH 8.5), 300 mM NaCl, 10% glycerol, 30 mM imidazole. 9. Elution buffer: 50 mM Tris–HCl (pH 8.5), 300 mM NaCl, 10% glycerol, 250 mM imidazole. 10. 0.01 M phosphate buffer (pH 7.2) or 20 mM Tris–HCl (pH 7.2). 11. Dialysis tubing. 12. Vivaspin® 6 columns (Sartorius). 2.6 Protein and Particle Analysis 2.6.1 Imaging

1. Handheld UV lamp (7000 μW, Novodirect). 2. Green light (515 nm): KL 2500 LCD lamp for stereomicroscopy (Schott AG) with LEE color film primary red number 106 (Thoman). 3. UV light (260 nm) Blak-Ray® B-100 YP UV Lamp (UVP). 4. D5300 camera (Nikon). 5. ImageJ (bundled with 64-bit Java 1.8.0_112).

2.6.2 RNA Analysis

1. RNeasy Plant Mini Kit (Qiagen). 2. DNase I. 3. 0.5-mL Eppendorf tubes. 4. Coating buffer: 15 mM Na2CO3, 35 mM NaHCO3 (pH 9.6). 5. PBS-T: 0.05% (v/v) Tween 20 in PBS. 6. M-MLV Reverse Transcriptase RNase H Minus, Point Mutant (M-MLV RT [H], Promega). 7. RNase H (Carl Roth). 8. RNase-free water. 9. Primers (e.g., oligo-dT). 10. dNTPs. 11. 1 TAE buffer. 12. Primers for cDNA amplification (see Table 2).

2.6.3 SDS-PAGE/Western Blotting

1. Loading dye (5 reducing buffer): 62.5 mM Tris–HCl (pH 6.8), 30% (v/v) glycerol, 4% (w/v) SDS, 10% (v/v) 2-mercaptoethanol, 0.05% (w/v) bromophenol blue. 2. 12% polyacrylamide gels + sodium dodecylsulfate (SDS): stacking gel total monomer concentration (T) ¼ 4%, cross-linking degree (C) ¼ 2.7%, pH 6.8; resolving gel T ¼ 12%, C ¼ 2.7%, pH 8.8, 0.4% SDS. 3. 30% (w/v) 37.5:1 acrylamide–bisacrylamide solution (Roth). Store in the dark at 4  C. 4. 10% ammonium persulfate. Store stock solution at 20  C.

224

Christina Dickmeis and Ulrich Commandeur

5. N,N,N0 ,N0 -Tetramethylethylenediamine (TEMED) (Roth). Store stock solution in the dark at 4  C. 6. 10 SDS-PAGE running buffer (pH 8.3): 250 mM Tris base, 2 M glycine, 1% SDS. Store at room temperature. 7. ColorPlus Prestained Protein Ladder Broad Range (P7719, New England Biolabs). 8. Coomassie Brilliant Blue staining solution: 0.25% (w/v) Coomassie Brilliant Blue G-250, 50% (v/v) methanol, 10% (v/v) acetic acid. 9. Coomassie destaining solution: 5% (v/v) methanol, 7.5% (v/v) acetic acid. 10. 1 PBS 11. 5% (w/v) skimmed milk in PBS. 12. Amersham™ Protran™ Nitrocellulose membrane 0.45 μm (GE Healthcare). 13. Trans-Blot Turbo (Bio-Rad). 14. Nitroblue tetrazolium chloride (NBT) and 5-bromo-4-chloro3-indolylphosphate (BCIP) p-toluidine salt stock solution: 3.33% (w/v) NBT and 1.65% (w/v) BCIP in dimethylformamide. 15. Semidry blotting buffer: 48 mM Tris base, 39 mM glycine, 20% (v/v) methanol (pH 9.6). 16. Alkaline phosphatase (AP) buffer: 100 mM Tris–HCl (pH 9.6), 100 mM NaCl, 5 mM MgCl2. 17. Primary antibodies: α-PVX (DSMZ), α-mCherry (α-DsRed antibody, GeneTex) and α-GFP (GeneTex). 18. Detection antibodies: monoclonal AP-conjugated goat antirabbit (GARAP) IgG secondary antibody (Dianova). 2.7 SpyTag/ SpyCatcher Coupling

1. Isolated virus and protein of interest (concentrated stocks).

2.8 Transmission Electron Microscopy (TEM)

1. Clean surface (e.g., Parafilm® or grid holder pad).

2.8.1 Adsorption Grids

2. 100 mM Tris–HCl, pH 6.0 (see Note 3).

2. Pioloform-coated 400-mesh nickel grids (Plano) or 400-mesh copper grids (Plano) covered with formvar 15/95E (SigmaAldrich) and a carbon film. 3. Filter paper (fiber free). 4. 1% (w/v) uranyl acetate (pH 4.3). 5. Zeiss EM 10 transmission electron microscope.

2.8.2 Immunosorbent TEM

1. Clean surface (e.g., Parafilm® or grid holder pad). 2. Pioloform-coated 400-mesh nickel grids (Plano). 3. Filter paper (fiber free).

Functionalized PVX Virions

225

4. 0.5% (v/v) bovine serum albumin (BSA) in PBS. 5. PBS-T. 6. 1% (w/v) uranyl acetate (pH 4.3). 7. Primary antibodies α-PVX, α-His6 and if possible an antibody against the protein of interest. 8. Secondary antibodies goat anti-rabbit or goat anti-mouse IgG labeled with 15-nm gold particles (BB Solution) for detection. 9. Zeiss EM 10 transmission electron microscope.

3

Methods

3.1 Genetic Engineering 3.1.1 Construction of PVX Clones

This section describes the generation of PVX infectious cDNA clones with genetically engineered CPs, utilizing different 2A sequences as well as the SpyTag sequence to capture proteins fused to the SpyCatcher ligand. The fluorescent proteins mCherry (28.8 kDa) and GFP (27 kDa) are generally too large for direct fusion to the PVX CP because the resulting VNPs are unstable. Fusion via the FDMV 2A sequence [45] leads to the production of the fusion product as well as free CP, thus allowing the assembly of hybrid PVX particles displaying large fusion proteins (see Fig. 1). Various 2A sequences (see Table 1) have been used to present fluorescent proteins on the particle surface at different densities (see Note 4). For all cloning procedures, follow standard guidelines [46]. 1. Clone the target sequence as an N-terminal CP fusion by replacing the gfp gene with the gene of the target protein in pCXI. The replacement can be achieved by excising the gfp sequence with restriction enzymes NheI and BspEI (see Fig. 1). 2. New 2A sequences can be added to the cp gene by amplification using a forward primer containing the new 2A sequence along with the M13 reverse primer (see Table 2). The primers also incorporate the NheI and BspEI recognition sites to facilitate cloning. 3. The SpyTag sequence can be introduced by PCR using forward primer (NheI-ST-CP), which as above is paired with the M13 universal primer (see Note 5). This amplifies the complete cp gene allowing subsequent restriction digestion with NheI and SalI. 4. Dephosphorylate the vectors with CIP after digestion and ligate with T4 ligase overnight at 16  C. 5. Introduce ligation products into E. coli DH5α cells and select colonies on LB plates supplemented with ampicillin overnight at 37  C.

226

Christina Dickmeis and Ulrich Commandeur

6. Isolate plasmids from overnight liquid cultures using the Pure Yield™ Plasmid Miniprep or Midiprep kits, and sequence the modified DNA. 3.1.2 Construction of SpyCatcher Fusion Proteins

This section describes the cloning of SpyCatcher fusion proteins for expression in plants and E. coli, although other expression hosts can be used if necessary. The SpyCatcher sequence allows N-terminal and C-terminal fusions and either strategy can be used depending on the protein of interest (see Note 6). Clone the fusion protein sequence in the desired orientation with a flexible linker such as (G4S)3 between the SpyCatcher and protein of interest. Use an expression vector that includes a His6 tag to facilitate purification. 1. Use the pTRAkt-ERH-Cel12A-SC-His6 vector as source for the 2A sequence. We advise using Gibson assembly or gene splicing by overlap extension PCR (SOE-PCR) for the construction of fusion proteins (see Note 7). 2. Introduce ligation products into E. coli DH5α cells and select colonies on LB plates supplemented with ampicillin overnight at 37  C. 3. Isolate plasmids from overnight liquid cultures using the Pure Yield™ Plasmid Miniprep or Midiprep kits, and sequence the modified DNA. 4. Introduce the target sequence into A. tumefaciens (see Subheading 3.4.1) or an E. coli expression strain (see Subheading 3.4.2).

3.2 PVX Particle Production

1. Prepare 10 μg of DNA from Subheading 3.1.1, step 6 in 100 μL of water for each leaf that will be inoculated. 2. Dust 4-week-old N. benthamiana plants with Celite 545 and rub 3–4 mature leaves per plant, equally distributed (~8 plants for 100 g infected leaf material) with the DNA solution from the previous step. Incubate the plants in a phyto-chamber with constant light (25,000–30,000 lux) at 26  C for 12 h, and in the dark at 20  C for 12 h. 3. Harvest leaf material 14–21 days post inoculation (dpi) depending on the infection status, which is determined by symptom distribution and fluorescence (see Subheading 3.6.1, Table 3 and Fig. 2). 4. Take 100 g of leaf material for virus purification using the modified protocol from CIP (International Potato Center, Lima, Peru) (see Note 8).

3.3 Plant Virus Purification

1. Homogenize 100 g of leaf material with two volumes (w/v) of ice-cold extraction buffer in a blender and filter the extract through three layers of Miracloth.

Functionalized PVX Virions

227

Table 3 Progress of infections caused by PVX particles presenting fluorescent proteins with different 2A sequences Fluorescent protein

2A sequence

First spots [dpi]

Systemic infection [dpi]

iLOV

10 31 39 56 70 80 90 Reference

7–10 7–9 7–8 5–7 4–5 4 3–4 4

12–13 10–11 10 9–10 8 7 5 6

mCherry

56 70 80 90 Reference

6–7 6–7 4 4 4

12–13 10–12 8–9 8–9 6–7

GFP

90 Reference

4 5

9 8–9

The table shows the average number of days post inoculation (dpi) before the first visible spots of a local infection and systemic infection. The reference sequence is the frequently used 2A sequence from the literature [20, 45, 47]

2. Clarify the filtrate by centrifugation at 7800  g for 20 min at 4  C. 3. Process the supernatant by adding 1% (v/v) Triton X-100, stir for 1 h at 4  C, and clarify by centrifugation at 5500  g for 20 min at 4  C. 4. Process the supernatant by adding NaCl/PEG solution to achieve a final concentration of 0.2 M NaCl and 4% (w/v) PEG (MW 6000–8000). Stir for 1 h at 4  C and then incubate for 1 h at room temperature (see Note 9). 5. Precipitate the viral particles by centrifugation at 7800  g for 10 min at 4  C. 6. Resuspend the pellet carefully in 4 mL 0.05 M phosphate buffer (pH 8.0) with 1% (v/v) Triton X-100, rinse the tubes immediately with 2 mL of the same buffer, and combine the samples. 7. Clarify the mixture by centrifugation at 7800  g for 10 min at 4  C. 8. Prepare a 24-mL sucrose gradient (10–45% (w/v)) with a gradient mixer in 25  89 mm tubes and carefully load the samples on top. 9. Centrifuge in a swing-bucket rotor at 96,500  g for 75 min at 4  C (SW 32 Ti ¼ 23,700 rpm).

228

Christina Dickmeis and Ulrich Commandeur

Fig. 2 Nicotiana benthamiana plants infected with PVX vectors containing different 2A sequences presenting ilOV (a) and other fluorescent proteins (GFP and mCherry) in comparison (b). The plants were monitored 14 days post inoculation and are visualized under the specific excitation conditions for the fluorescent proteins

Functionalized PVX Virions

229

10. Collect 1.5-mL gradient fractions from bottom to top (see Note 10). 11. Analyze all fractions by SDS-PAGE (see Note 11). 12. Select and combine fractions with the highest concentration of the desired fusion proteins in 14  89 mm tubes and dilute combined fractions at least in the same volume of 0.01 M phosphate buffer (pH 7.2). 13. Sediment the virus particles by ultracentrifugation at 102,600  g for 3–5 h at 4  C (SW 41 Ti ¼ 38,000 rpm). 14. Carefully remove the supernatant and resuspend the pellet in 0.2 mL 0.01 M phosphate buffer (pH 7.2) and stir overnight at 4  C. 15. Clarify the solution by centrifugation at 5000  g for 10 min at 4  C. 16. Determine the plant virus concentration in the supernatant by measuring the OD260nm and using the extinction coefficient εPVX ¼ 2.97 mL/mg cm at 260 nm (see Note 12). 3.4 Protein Expression

3.4.1 Transient Expression in Plants

The ideal expression host depends on the nature of the protein displayed on the particle surface. Plants are suitable for complex proteins that require posttranslational modification. These modifications can be influenced by targeting different subcellular compartments. Bacterial expression systems are better for the rapidly scalable production of structurally simple proteins (e.g., those of prokaryotic origin). 1. Prepare A. tumefaciens infiltration cultures (see Note 13) by cultivating A. tumefaciens containing an appropriate plasmid at 26  C in YEP medium overnight. Supplement cultures with 10 mM MES (pH 5.6), 10 mM glucose and 20 μM acetosyringone after 24 h, and incubate for a further day. Adjust the OD600nm to 0.5–1.0 with 2  infiltration medium, supplement with 200 μM acetosyringone, and incubate for 30 min at room temperature. 2. Infiltrate leaves of N. benthamiana plants (4–6 weeks old) with a needleless syringe and incubate in a phyto-chamber with constant light (25,000–30,000 lux) at 26  C for 12 h, and in the dark at 20  C for 12 h. 3. Harvest the infiltrated leaves 5–6 dpi and use freshly or store at 80  C until further use (see Note 14).

3.4.2 Expression in E. coli

1. Prepare E. coli expression cultures by introducing the plasmid of choice into the desired E. coli strain, and confirm transformation by standard colony PCR.

230

Christina Dickmeis and Ulrich Commandeur

2. Inoculate 200 mL TB medium including the appropriate antibiotics (e.g., 100 μg/mL ampicillin) with a positive colony and incubate the culture by shaking (180 rpm, 37  C) until the OD600nm reaches ~0.5. 3. Induce expression by adding the appropriate inducer for the promoter, such as 1 mM IPTG for the lac promoter (see Note 15). 4. Incubate the cultures for at least 3 h at 37  C, shaking at 180 rpm (see Note 16), and harvest the cells by centrifugation (5000  g for 20 min at 4  C). 5. Use the pelleted cells directly for purification or store the cells at 20  C. 3.5 Protein Purification

3.5.1 Purification of Proteins Expressed in Plants

Proteins of interest fused to the SpyCatcher sequence and including a His6 tag can be purified by immobilized metal ion affinity chromatography (IMAC) (see Note 17). 1. Grind plant material in two volumes of PBS and centrifuge at 15,000  g for 15 min at 4  C. 2. Dilute the supernatant 1:10 with PBS and apply to 500-μL Ni-NTA agarose columns (see Note 18). Collect a sample from the flow-through fraction. 3. Wash columns twice with ten volumes (v/v Ni-NTA agarose) of 10 mM imidazole in PBS. 4. Elute with one volume (v/v Ni-NTA agarose) of 150–300 mM imidazole in PBS (see Note 19). 5. Dialyze the eluate against 0.01 M phosphate buffer (pH 7.2) or 20 mM Tris–HCl (pH 7.2) to remove the imidazole and concentrate using Vivaspin 6 columns (if needed).

3.5.2 Purification of Proteins Expressed in E. coli

1. Resuspend the cell pellet (see Subheading 3.4.2) in 10 mL lysis buffer and add 2 mg/mL lysozyme. Lyse the cells by shaking gently on ice for 30 min. 2. Sonicate the cell suspension (60% amplitude, cycle 0.5) for 4  45 s with pauses of 45 s on ice. 3. Add 1% (v/v) Triton X-100 and store for 30 min on ice. 4. Remove the cell debris by centrifugation at 14,000  g for 30 min at 4  C. 5. Apply the supernatant to 500-μL Ni-NTA agarose columns (see Note 20). Collect a sample from the flow-through fraction. 6. Wash columns twice with ten volumes (v/v Ni-NTA agarose) of wash buffer. 7. Elute with one volume (v/v Ni-NTA agarose) of elution buffer.

Functionalized PVX Virions

231

8. Dialyze the eluate against 0.01 M phosphate buffer (pH 7.2) or 20 mM Tris–HCl (pH 7.2) to remove the imidazole and concentrate with Vivaspin 6 columns (if needed). 3.6 Protein and Particle Analysis 3.6.1 Visualization of Fluorescence (See Fig. 2)

1. Monitor inoculated plants (see Subheading 3.2) daily for fluorescence using a suitable excitation source. For iLOV and GFP, fluorescence can be monitored under UV light. For mCherry, use a green light source (515 nm) and a red filter. 2. Confirm the presence of fluorescent proteins by SDS-PAGE (see Subheading 3.6.3). 3. Prepare samples by mixing with 5 loading dye but do not boil the samples prior to loading because this will denature the proteins and abolish fluorescence. 4. Prepare 12% (w/v) polyacrylamide resolving gels, load the samples and run the gel to fractionate the proteins (see Subheading 3.6.3). 5. Image the gel under a UV light, a green light with red filter, or suitable excitation conditions for other fluorescent proteins.

3.6.2 RNA Analysis

It is possible to analyze the total virus RNA isolated from plants (steps 1 and 2) or the encapsulated RNA (steps 3–6) by RT-PCR and agarose gel electrophoresis (steps 7–11). 1. Characterize total virus RNA by grinding 300 mg of plant material (see Subheading 3.2) under liquid nitrogen and isolating total (plant and virus) RNA using the RNeasy Plant Mini Kit (Qiagen) according to the manufacturer’s instructions. Determine the concentration at 260 nm. 2. Prepare samples containing 3 μg RNA and incubate them with DNase I to remove DNA contamination. 3. For the characterization of RNA encapsidated in virus particles, virions can be immunocaptured in 0.5-mL Eppendorf tubes. Coat tubes for at least 4 h at 30  C with 100 μL α-PVX antibody diluted 1:100 (v/v) in coating buffer. 4. Discard the coating solution and wash the tubes three times with 500 μL PBS-T before adding 75 μL plant sap (see Subheading 3.3, step 2) and incubate overnight at 4  C. 5. Wash the tubes sequentially with 500 μL PBS-T, 500 μL PBS and 500 μL distilled water. 6. The isolated particles can be used directly for RT-PCR. 7. Reverse-transcribe isolated RNA (from steps 2 or 6) using M-MLV RT [H]. 8. Mix samples with 27 μL distilled RNase-free water and 0.2 μM primer (e.g., oligo-dT) and denature for 10 min at 80  C before cooling on ice for primer annealing.

232

Christina Dickmeis and Ulrich Commandeur

9. Supplement mixture with M-MLV RT 5 reaction buffer, 0.5 mM dNTPs and 100 U M-MLV RT [H], and top up to 50 μL with distilled water. Perform reverse transcription reaction as recommended by the manufacturer. 10. Digest RNA with RNase H following cDNA synthesis. 11. Amplify the desired region on the viral genome using primers CX1, CX2 and CX8 (see Table 2) and resolve products by agarose gel electrophoresis in 1 TAE buffer. 3.6.3 Gel Electrophoresis and Immunoblotting (Fig. 3)

1. Resolve protein samples by SDS-PAGE after mixing plant sap (see Subheading 3.3, step 2 or Subheading 3.5.1), 1–4 μg of purified particles (see Subheading 3.3) or the coupling reaction (see Subheading 3.7) with 5 loading dye and boiling for 5–10 min. 2. Stain one gel with Coomassie Brilliant Blue staining solution, gently shaking the gel for 20 min. 3. Remove the staining solution for reuse, and wash the gel with water. 4. Remove background with destaining solution. Replace the destaining solution when it darkens and repeat until the gel is sufficiently destained. 5. For the specific detection of proteins, blot a second unstained polyacrylamide gel onto a nitrocellulose membrane by semidry blotting according to standard procedures or the transfer cell manufacturer’s instructions. 6. Block membranes for 1 h with 5% (w/v) skimmed milk in PBS. 7. Incubate membranes with primary antibodies (α-PVX, α-His6, or antibodies specific for the target protein of interest) diluted 1:5000 in PBS for at least 1 h. 8. Wash membranes with PBS and incubate with AP-conjugated secondary antibody diluted 1:5000 in PBS for 2 h. 9. Wash membranes several times with PBS and incubate in AP buffer for 5 min. 10. Develop the blot with 100 μL NBT/BCIP in 10 mL AP buffer. Stop the reaction by removing the NBT/BCIP solution and washing with water. 11. The ratio of fused to nonfused or coupled and noncoupled proteins can be determined by ImageJ analysis of the gels and blots (see Fig. 3).

3.7 SpyTag/ SpyCatcher Coupling Reactions

For the coupling of SpyCatcher fusion proteins to PVX particles carrying the SpyTag, the components can be mixed in any suitable buffer.

Functionalized PVX Virions

233

Fig. 3 Different PVX-based presentation strategies using 2A sequences and SpyTag/SpyCatcher coupling. (a) Upper panel shows samples from plants in which different 2A sequences are used for the presentation of mCherry on PVX. The samples were analyzed by SDS-PAGE and the gel is illuminated for the specific excitation of mCherry. This shows that certain 2A sequences (2A70) can favor the accommodation of larger proteins. Lower panel shows the plotting of one of the gel lanes in ImageJ. N.b.—Nicotiana benthamiana noninfected plant sap, ref—reference 2A sequence. (b) Purified PVX particles presenting the SpyTag (PVX-ST) were coupled with IMAC-purified tyrosine-ammonia lyases (TALs) fused to the SpyCatcher sequence (SC). Western blots were analyzed with α-PVX and GARAP (upper panel) and α-His6 and GAMAP (lower panel). (c) Calculation

234

Christina Dickmeis and Ulrich Commandeur

1. Start with a molar excess of 3:1 SpyCatcher fusion protein to PVX-SpyTag. According to our experiences 3 μg of PVX is sufficient to analyze the reaction, revealing the optimum coupling conditions (see Note 21). 2. Mix the calculated amounts of proteins and particles for the coupling reaction in 50 μL 100 mM Tris–HCl buffer (pH 6.0) or test different pH values. 3. Incubate the coupling reaction overnight at 4  C (see Note 22). 3.8 Density of Fusion Proteins Presented on the Particle Surface

Use the results of SDS-PAGE analysis and western blotting (see Subheading 3.6.3) from plant sap, purified particles or coupling reactions, after testing different 2A sequences (see Fig. 3). 1. Load the image of the gels or blots into ImageJ software. 2. Select a rectangular lane and analyze it by defining the lane under the function “analyze gels – select first lane” and then by using the function “plot lanes.” 3. Determine the peak area of the fused and nonfused protein of interest or CP. 4. The ratio of the peak areas can be used to determine the amount of fusion protein.

3.9 Transmission Electron Microscopy 3.9.1 Adsorption Grids

3.9.2 Immunosorbent Transmission Electron Microscopy (ISEM)

1. For direct adsorption, prepare drops of ~40 μL PBS with 10 μg of purified particles on a clean surface and incubate Pioloformcoated 400-mesh nickel grids for 20 min in these drops. 2. Counterstain the loaded grids (from step 1) with 1% (w/v) uranyl acetate (pH 4.3) before analysis using, for example, a Zeiss EM 10 TEM. 1. Incubate Pioloform-coated 400-mesh nickel grids with a virusspecific antibody or antiserum diluted 1:10 in PBS for 20 min. 2. Wash off unbound antibodies with ~3 mL PBS-T by holding the grid with forceps and applying drops of PBS-T. 3. Block grids with 0.5% (v/v) BSA in PBS for 15 min. 4. Incubate grids with 10 μg of the particle preparation in 40 μL PBS for 20 min, followed by washing with PBS. 5. Captured particles can be analyzed for engineered peptides exposed on the surface by immunogold staining. Incubate the

ä Fig. 3 (continued) of the efficiency of different 2A sequences for the fusion of iLOV and mCherry to the PVX coat protein (CP). The efficiencies were calculated using ImageJ after analyzing the gel lanes (see (a), lower panel). For mCherry, only fusions using 2A sequences with skipping rates 56% are possible, and iLOV fusions tend to have higher skipping rates for the lower efficiencies than expected. Data were collected over different infection series and different time points after inoculation

Functionalized PVX Virions

235

preparations for 2 h with a virus-specific antibody from a different origin as the capture antibody, or a specific anti-peptide antibody diluted 1:100 in PBS. 6. Incubate grids overnight with secondary antibodies labeled with 15-nm gold particles diluted 1:50 in PBS. 7. Wash grids thoroughly with PBS and then with distilled water. 8. Contrast with 1% (w/v) uranyl acetate (pH 4.3) before analysis using, for example, a Zeiss EM 10 TEM.

4

Notes 1. The SpyCatcher sequence can be added to the C-terminus or N-terminus of the fusion partner, so the strategy should depend on which location is better tolerated in terms of retaining protein activity or interactions. Various linker sequences can also be inserted between the SpyCatcher and its fusion partner. We recommend starting with flexible linkers with a high serine/glycine content such as (G4S)3. 2. The best expression strain depends on the properties of the target protein. E. coli BL21 star is a good general choice because it promotes mRNA stability, but some target proteins require the coexpression of chaperons and thus need different expression strains. The SpyTag/SpyCatcher system is compatible with any expression host. 3. The reaction can be tested at different pH values for optimization, but pH 6.0 achieves robust coupling with the PVX-SpyTag system. 4. The fusion protein can also influence the activity of the 2A sequence [48]. We therefore recommend testing at least three different sequences. 5. The direct fusion of the SpyTag sequence to the CP of PVX was not successful [23]. Additional amino acids that restore the isoelectric point are therefore necessary for the successful presentation of SpyCatcher fusion proteins. 6. If both N-terminal and C-terminal SpyCatcher fusions are compatible with protein activity, we recommend selecting the strategy that achieves the best expression level. 7. The detailed cloning procedure will depend on the fusion protein and vector, and it is not possible to consider all variants here. 8. The number of plants needed to harvest at least 100 g of leaf material can vary from construct to construct and needs to be determined empirically. Some fusion proteins may influence the severity of infection symptoms, potentially reducing the

236

Christina Dickmeis and Ulrich Commandeur

biomass. PVX vectors with efficient 2A sequences expressing the iLOV domain generate small amounts of biomass due to the severe symptoms [25]. The average time needed for infection to develop when using different 2A fusion constructs is shown in Table 3. 9. Precipitation can be improved by incubating the supernatant with NaCl and PEG overnight, if the target protein is stable under these conditions. 10. Gradient fractions can be collected by puncturing the bottom of the centrifugation tubes with a needle connected to a flexible tube. Collect the drawn liquid in 1.5-mL reaction tubes. 11. To identify fractions with functional fluorescent protein/CP fusions, avoid boiling the samples for SDS-PAGE and visualize the fluorescence directly in the gel before staining with Coomassie Brilliant Blue. Boiling will denature the proteins and abolish their intrinsic fluorescence and staining will obscure the fluorescent bands. 12. Read the absorbance at 260 nm and 280 nm and calculate the A260:280 ratio to determine sample purity. A ratio of 1.2  0.1 indicates pure PVX preparations. 13. Coexpression of the tomato bushy stunt virus silencing suppressor p19 can improve product yields in N. benthamiana. 14. Plant material can be stored after harvesting by immediately freezing at 80  C. 15. Optimal inducer concentrations can vary for different plasmids and E. coli strains and should be tested based on the recommendations provided with those materials. 16. Optimal expression conditions depend on the expression host and target protein. Expression levels can be optimized by testing different temperatures, media, and expression times. 17. Analyze samples from each step by SDS-PAGE. The purification procedure can be improved by using different concentrations of imidazole for the washing steps and for the elution steps. The imidazole concentrations tolerated by the target protein on the column during the washing step should be tested empirically. We recommend increasing the imidazole concentration stepwise from 10 to 50 mM. 18. Undiluted plant sap frequently plugs the column so we recommend dilution. 19. The amount of imidazole needed to elute the target protein can differ from protein to protein. The minimal amount of imidazole needed for elution should be used, and this must be determined by testing different concentrations ranging from 150 to 300 mM.

Functionalized PVX Virions

237

20. To avoid plugged columns, the lysate should be passed through a 0.45 μm filter. 21. Use wild-type/unmodified virus particles as controls. 22. For different proteins, different incubation times and temperatures can also be advantageous. For PVX-SpyTag, robust coupling was achieved at lower temperatures. References 1. Palmer LC, Stupp SI (2008) Molecular selfassembly into one-dimensional nanostructures. Acc Chem Res 41:1674–1684. https://doi. org/10.1021/ar8000926 2. Gazit E (2007) Self-assembled peptide nanostructures: the design of molecular building blocks and their technological utilization. Chem Soc Rev 36:1263–1269. https://doi. org/10.1039/b605536m 3. Royston E, Ghosh A, Kofinas P et al (2008) Self-assembly of virus-structured high surface area nanomaterials and their application as battery electrodes. Langmuir 24:906–912. https://doi.org/10.1021/la7016424 4. Tinazzi E, Merlin M, Bason C et al (2015) Plant-derived chimeric virus particles for the diagnosis of primary sjo¨gren syndrome. Front Plant Sci 6:1080. https://doi.org/10.3389/ fpls.2015.01080 5. Dickmeis C, Kauth L, Commandeur U (2021) From infection to healing: the use of plant viruses in bioactive hydrogels. Wiley Interdiscip Rev Nanomed Nanobiotechnol 13:e1662 6. Shukla S, Dickmeis C, Nagarajan AS et al (2014) Molecular farming of fluorescent virus-based nanoparticles for optical imaging in plants, human cells and mouse models. Biomater Sci 2:784. https://doi.org/10.1039/ c3bm60277j 7. Manchester M, Singh P (2006) Virus-based nanoparticles (VNPs): platform technologies for diagnostic imaging. Adv Drug Deliv Rev 58:1505–1522. https://doi.org/10.1016/j. addr.2006.09.014 8. Wen AM, Steinmetz NF (2016) Design of virus-based nanomaterials for medicine, biotechnology, and energy. Chem Soc Rev 45: 4074–4126. https://doi.org/10.1039/ c5cs00287g 9. Adams MJ, Antoniw JF, Bar-Joseph M et al (2004) The new plant virus family Flexiviridae and assessment of molecular criteria for species demarcation. Arch Virol 149:1045–1060 10. Morozov SY, Lukasheva LI, Chernov BK et al (1987) Nucleotide-sequence of the open reading frames adjacent to the coat protein

cistron in potato virus-X genome. FEBS Lett 213:438–442. https://doi.org/10.1016/ 0014-5793(87)81538-7 11. Huisman MJ, Linthorst HJ, Bol JF, Cornelissen JC (1988) The complete nucleotide sequence of potato virus X and its homologies at the amino acid level with various plusstranded RNA viruses. J Gen Virol 69: 1789–1798. https://doi.org/10.1099/00221317-69-8-1789 12. Chapman S, Kavanagh T, Baulcombe D (1992) Potato virus X as a vector for gene expression in plants. Plant J 2:549–557. https://doi.org/ 10.1046/j.1365-313X.1992.t01-24-00999.x 13. Draghici H-K, Varrelmann M (2009) Evidence that the linker between the methyltransferase and helicase domains of potato virus X replicase is involved in homologous RNA recombination. J Virol 83:7761–7769. https://doi.org/ 10.1128/JVI.00179-08 14. Kim KH, Hemenway C (1997) Mutations that alter a conserved element upstream of the potato virus X triple block and coat protein genes affect subgenomic RNA accumulation. Virology 232:187–197. https://doi.org/10. 1006/viro.1997.8565 15. Chapman S, Hills G, Watts J, Baulcombe D (1992) Mutational analysis of the coat protein gene of potato virus X: Effects on virion morphology and viral pathogenicity. Virology 191: 223–230. https://doi.org/10.1016/00426822(92)90183-P 16. Santa Cruz S, Baulcombe D (1995) Analysis of potato virus X coat protein genes in relation to resistance conferred by the genes Nx, Nb and Rx1 of potato. J Gen Virol 76:2057–2061. https://doi.org/10.1099/0022-1317-768-2057 17. Baratova LA, Grebenshchikov NI, Dobrov EN et al (1992) The organization of potato virus X coat proteins in virus particles studied by tritium planigraphy and model building. Virology 188:175–180. https://doi.org/10.1016/ 0042-6822(92)90747-D 18. Koenig R, Torrance L (1986) Antigenic analysis of potato virus X by means of monoclonal

238

Christina Dickmeis and Ulrich Commandeur

antibodies. J Gen Virol 67:2145–2151. https://doi.org/10.1099/0022-1317-6710-2145 19. Cruz SS, Chapman S, Roberts AG et al (1996) Assembly and movement of a plant virus carrying a green fluorescent protein overcoat. Proc Natl Acad Sci 93:6286–6290. https://doi. org/10.1073/pnas.93.13.6286 20. Ro¨der J, Dickmeis C, Commandeur U (2019) Small, smaller, nano: new applications for potato virus X in nanotechnology. Front Plant Sci 10:158. https://doi.org/10.3389/fpls. 2019.00158 21. Lico C, Capuano F, Renzone G et al (2006) Peptide display on potato virus X: molecular features of the coat protein-fused peptide affecting cell-to-cell and phloem movement of chimeric virus particles. J Gen Virol 87: 3103–3112. https://doi.org/10.1099/vir.0. 82097-0 22. Uhde-Holzem K, Fischer R, Commandeur U (2007) Genetic stability of recombinant potato virus X virus vectors presenting foreign epitopes. Arch Virol 152:805–811. https://doi. org/10.1007/s00705-006-0892-y 23. Ro¨der J, Fischer R, Commandeur U (2017) Engineering potato virus X particles for a covalent protein based attachment of enzymes. Small 13. https://doi.org/10.1002/smll. 201702151 24. Uhde-Holzem K, McBurney M, Tiu BDB et al (2016) Production of immunoabsorbent nanoparticles by displaying single-domain protein A on Potato Virus X. Macromol Biosci 16: 231–241. https://doi.org/10.1002/mabi. 201500280 25. Ro¨der J, Dickmeis C, Fischer R, Commandeur U (2018) Systemic infection of Nicotiana benthamiana with Potato virus X nanoparticles presenting a fluorescent iLOV polypeptide fused directly to the coat protein. Biomed Res Int 2018:9328671. https://doi.org/10. 1155/2018/9328671 26. Lico C, Capuano F, Renzone G et al (2006) Peptide display on Potato virus X: molecular features of the coat protein-fused peptide affecting cell-to-cell and phloem movement of chimeric virus particles. J Gen Virol 87: 3103–3112. https://doi.org/10.1099/vir.0. 82097-0 27. Betti C, Lico C, Maffi D et al (2012) Potato virus X movement in Nicotiana benthamiana: new details revealed by chimeric coat protein variants. Mol Plant Pathol 13:198–203. https://doi.org/10.1111/j.1364-3703.2011. 00739.x

28. Baratova LA, Fedorova NV, Dobrov EN et al (2004) N-Terminal segment of Potato virus X coat protein subunits is glycosylated and mediates formation of a bound water shell on the virion surface. Eur J Biochem 271:3136–3145. https://doi.org/10.1111/j.1432-1033.2004. 04243.x 29. Ryan MD, King AMQ, Thomas GP (1991) Cleavage of foot-and-mouth disease virus polyprotein is mediated by residues located within a 19 amino acid sequence. J Gen Virol 72: 2727–2732. https://doi.org/10.1099/00221317-72-11-2727 30. Robertson BH, Grubman MJ, Weddell GN et al (1985) Nucleotide and amino acid sequence coding for polypeptides of foot-andmouth disease virus type A12. J Virol 54: 651–660. https://doi.org/10.1128/jvi.54.3. 651-660.1985 31. Donnelly MLL, Hughes LE, Luke G et al (2001) The “cleavage” activities of foot-andmouth disease virus 2A site-directed mutants and naturally occurring “2A-like” sequences. J Gen Virol 82:1027–1041. https://doi.org/ 10.1099/0022-1317-82-5-1027 32. Donnelly MLL, Luke G, Mehrotra A et al (2001) Analysis of the aphthovirus 2A/2B polyprotein “cleavage” mechanism indicates not a proteolytic reaction, but a novel translational effect: a putative ribosomal “skip”. J Gen Virol 82:1013–1025. https://doi.org/10. 1099/0022-1317-82-5-1013 33. Ryan MD, Drew J (1994) Foot-and-mouth disease virus 2A oligopeptide mediated cleavage of an artificial polyprotein. EMBO J 13: 9 2 8 – 9 3 3 . h t t p s : // d o i . o r g / 1 0 . 1 0 0 2 / 14651858.CD001431.pub4.Copyright 34. Minskaia E, Nicholson J, Ryan MD (2013) Optimisation of the foot-and-mouth disease virus 2A co-expression system for biomedical applications. BMC Biotechnol 13:67. https:// doi.org/10.1186/1472-6750-13-67 35. Donnelly MLL, Gani D, Flint M et al (1997) The cleavage activities of aphthovirus and cardiovirus 2A proteins. J Gen Virol 78:13–21. https://doi.org/10.1099/0022-1317-781-13 36. Zakeri B, Fierer JO, Celik E et al (2012) Peptide tag forming a rapid covalent bond to a protein, through engineering a bacterial adhesin. Proc Natl Acad Sci 109:E690–E697. https://doi.org/10.1073/pnas.1115485109 37. Hagan RM, Bjo¨rnsson R, McMahon SA et al (2010) NMR spectroscopic and theoretical analysis of a spontaneously formed lys-asp isopeptide bond. Angew Chem Int Ed 49:

Functionalized PVX Virions 8421–8425. https://doi.org/10.1002/anie. 201004340 38. Schoene C, Fierer JO, Bennett SP, Howarth M (2014) SpyTag/Spycatcher cyclization confers resilience to boiling on a mesophilic enzyme. Angew Chem Int Ed 53:6101–6104. https:// doi.org/10.1002/anie.201402519 39. Sun F, Zhang W-B, Mahdavi A et al (2014) Synthesis of bioactive protein hydrogels by genetically encoded SpyTag-SpyCatcher chemistry. Proc Natl Acad Sci 111:11269–11274. https://doi.org/10.1073/pnas.1401291111 40. Fierer JO, Veggiani G, Howarth M (2014) SpyLigase peptide–peptide ligation polymerizes affibodies to enhance magnetic cancer cell capture. Proc Natl Acad Sci 111: E1176–E1181. https://doi.org/10.1073/ pnas.1315776111 41. Chen AY, Deng Z, Billings AN et al (2014) Synthesis and patterning of tunable multiscale materials with engineered cells. Nat Mater 13: 5 1 5 – 5 2 3 . h t t p s : // d o i . o r g / 1 0 . 1 0 3 8 / nmat3912 42. Zhou C, Hong Y, Zhang X (2013) Applications of nanostructured calcium phosphate in tissue engineering. Biomater Sci 1:1012–1028. https://doi.org/10.1039/c3bm60058k

239

43. Bin ZW, Sun F, Tirrell DA, Arnold FH (2013) Controlling macromolecular topology with genetically encoded SpyTag-SpyCatcher chemistry. J Am Chem Soc 135:13988–13997. https://doi.org/10.1021/ja4076452 44. Lee KL, Uhde-Holzem K, Fischer R et al (2014) Virus hybrids as nanomaterials. Methods Protoc Methods Mol Biol 1108:3–21. https://doi.org/10.1007/978-1-62703751-8 45. Baulcombe DC, Chapman S, Santa Cruz S (1995) Jellyfish green fluorescent protein as a reporter for virus infections. Plant J 7: 1045–1053. https://doi.org/10.1046/j. 1365-313X.1995.07061045.x 46. Sambrook J, Green M (2012) Molecular cloning: a laboratory manual, 4th edn. Cold Spring Harbour Laboratory, New York 47. Lee KL, Uhde-Holzem K, Fischer R et al (2014) Genetic engineering and chemical conjugation of Potato virus X. Methods Mol Biol 1108:3–21. https://doi.org/10.1007/978-162703-751-8_1 48. Minskaia E, Ryan MD (2013) Protein coexpression using FMDV 2A: Effect of “linker” residues. Biomed Res Int 2013:291730. https://doi.org/10.1155/2013/291730

Chapter 14 Knockout of Glycosyltransferases in Nicotiana benthamiana by Genome Editing to Improve Glycosylation of Plant-Produced Proteins Julia Jansing and Luisa Bortesi Abstract Plants are excellent production hosts for the in vivo synthesis of complex glycosylated proteins such as antibodies. The plant N-glycosylation machinery is largely similar to that found in humans and other mammalian organisms, which is an advantage in comparison to microbial production systems in particular. However, there are some differences in the identity and chemical linkage of the sugars that plants and mammals use to build their N-glycans. These differences can affect important properties of glycosylated proteins produced recombinantly in plants. Here we describe the complete procedure of multiplex targeted gene knockout with CRISPR/Cas9 in Nicotiana benthamiana in order to eliminate the undesirable sugars α-1,3-fucose and β-1,2-xylose from the plant N-glycans. The workflow includes target gene identification, guide RNA design and testing, plant transformation, and the analysis of the regenerated transgenic plants by Sanger sequencing, immunoblot, and mass-spectrometric analysis of recombinant and endogenous proteins. Key words Allotetraploid genome, CRISPR/Cas9, Genome editing, Glycosyltransferase, Nicotiana benthamiana, Targeted gene knockout

1

Introduction The development of the CRISPR/Cas9 genome editing system [1] has made it easier for scientists to change the genetic makeup of an organism to make it better suited for a specific purpose. Plants, with their rather large and often complex genomes [2, 3], are no exception, but the experimental workflow to edit a gene or even several genes of interest with CRISPR/Cas9 is long, includes a wide range of methods, and may seem daunting to researchers with little genome editing experience. This chapter aims to provide a comprehensive and complete workflow describing how to knock out multiple glycosyltransferase genes in the plant species Nicotiana benthamiana. The workflow described here is based on research

Stefan Schillberg and Holger Spiegel (eds.), Recombinant Proteins in Plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480, https://doi.org/10.1007/978-1-0716-2241-4_14, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

241

242

Julia Jansing and Luisa Bortesi

previously performed by the authors [4], but the methods and experimental strategies can be adapted for other plant species and target genes. 1.1 Gene Knockout with CRISPR/Cas9

Genome editing with CRISPR/Cas9 is based on the interaction of the Cas9 nuclease, the single chimeric guide RNA (gRNA) [1] in complex with the Cas9, and the DNA target sequence. The main difference between Cas9 and previous programmable DNA nucleases is that the recognition of its target sequence is based on Watson–Crick base pairing between the gRNA and target DNA. The previous generation of programmable nucleases such as TALENs (Transcription Activator-like Effector Nucleases) [5] or ZFNs (Zinc-finger Nucleases) [6] on contrast relies on proteinDNA binding to identify its target sequence. Designing a new gRNA to target Cas9 to another DNA sequence is much easier and faster than protein-engineering TALENs or ZFNs [7]. Once Cas9 is in complex with a gRNA, it starts scanning the DNA for a short sequence called protospacer-adjacent motif, or PAM [8]. For the commonly used Cas9 from Streptococcus pyogenes, the PAM sequence is “5-NGG-3”. Other CRISPR nucleases have different PAM sequences. When Cas9 finds a PAM and the DNA and its bound gRNA are complementary, Cas9 cleaves the DNA three base pairs upstream of the PAM sequence [1, 9, 10]. This DNA break is a threat to the cell viability and genome stability [11], and is quickly repaired by non-homologous end joining (NHEJ) in higher plants [12]. NHEJ repair is imperfect and occasionally results in small deletions or insertions at the break site [13]. For the purpose of gene-knockout, this is the required repair outcome. When that insertion or deletion at the repaired break site leads to a frameshift, the downstream sequence normally contains numerous premature stop codons. The strict quality control of the mRNA molecules transcribed from the mutated gene will then lead to nonsense-mediated mRNA decay and consequently the functional gene knockout [14, 15]. The presence of one or more premature stop-codons upstream of an exon–exon junction in a processed mRNA molecule is recognized by the quality control machinery and leads to the degradation of the mRNA.

1.1.1 Multiplex Gene Knockout

One of the main advantages of the CRISPR/Cas9 system is that it is efficient enough to edit several genes at the same time. The term multiplex gene editing is used when several genomic sequences are targeted at the same time. In this chapter, two β-1,2-xylosyltransferase and four α-1,3-fucosyltransferase genes will be targeted by several gRNAs each. An efficient way to express multiple gRNAs from a single vector are polycistronic tRNA/ gRNA (PTG) genes [16]. In these polycistronic constructs, the

Multiplex Glycosyltransferase Knockout in N. benthamiana

243

Fig. 1 Schematic of the polycistronic tRNA/gRNA gene (PTG) expression cassette under control of a plant U6 promoter. The gray boxes represent the tRNA sequence, colored rhombi the different gRNA sequences, and black squares the gRNA scaffold sequence, which is identical for all gRNAs. (Figure adapted from [4])

gRNA sequences are flanked by tRNA sequences that are cleaved by the host tRNA processing machinery (see Fig. 1). This processing is very precise and releases gRNA molecules efficiently [16]. 1.2 Outline of the Workflow

The workflow is divided into three subsections. The first section describes the process of generating transgenic plants including (1) how to identify and validate the sequence of the target gene, (2) how to design and test gRNAs for it, and (3) how to transform and regenerate plants with the knockout construct. The next section explains the screening procedures for the transgenic plants you will regenerate: (4) on the genetic level, (5) by immunoblot analysis, (6) by mass-spectrometric analysis of a recombinantly produced antibody, and (7) by mass-spectrometric analysis of the leaf proteome. You will perform a subset or all of these four procedures in the analysis of the different plant generations you are going to work with. Different plant generations will be necessary because most likely you will not obtain a complete gene knockout in the first generation and will have to go through several rounds of self-pollination to identify a knockout line. Even if you obtain a full knockout early, the inheritability of the mutations has to be confirmed over several generations. For each generation, the experimental strategy for plant screening is to start with the cheaper and quicker methods to analyze all plants, and to select lines that will be analyzed with the next, more complex and expensive analysis. In principle, all plants could be screened by MALDI-TOF mass spectrometry with highly informative results, but the cost would be prohibitive for most research groups. The strategy outlined here allows to narrow down the set of samples with each consecutive analysis, so that only a few selected plants have to be analyzed with the expensive mass spectrometry methods. Figure 2 outlines the screening workflow for the T1 generation onward. The final section describes how to perform manual crossing of transgenic N. benthamiana lines in order to combine properties of

244

Julia Jansing and Luisa Bortesi

Fig. 2 Workflow for the screening of transgenic plants of the T1 and later generations

different knockout lines in one plant line, for example by crossing an α-1,3-fucosyltransferase knockout line and a β-1,2-xylosyltransferase knockout line.

2

Materials Prepare all buffer, media, and reaction mixes with ultrapure water at room temperature, unless indicated otherwise. Standard molecular biology tools and consumables are needed to conduct the experiments: a set of micropipettes covering volumes from 0.2 to

Multiplex Glycosyltransferase Knockout in N. benthamiana

245

1000 μL, pipette tips, sterile 1.5- and 2.0-mL reaction tubes, wet ice, and an insulated ice container. Storage space at room temperature (RT),
4  C,
20  C, and
80  C is required. For the following methods, external services were used by the authors and are therefore not described in detail in this chapter: Sanger sequencing, Next generation sequencing, mass-spectrometric analysis of antibody heavy chains by LC-ESI-MS, and massspectrometric analysis of endogenous leaf proteins by MALDI-TOF. 2.1 Target Gene Resequencing 2.1.1 Spin-Column Extraction of Plant Genomic DNA

1. Wild-type N. benthamiana Rdr1 insertion genotype [17]. 2. Spin-column kit for genomic DNA extraction from plant tissue, for example, Macherey & Nagel NucleoSpin Plant II. 3. Mortar and pestle or alternative tools for cell disruption. 4. Liquid nitrogen. 5. Small metal spatula or spoon. 6. 96–100% ethanol. 7. Thermal heating-block or water bath. 8. Microcentrifuge for 1.5 and 2.0 mL reaction tubes. 9. DNA quantification spectrophotometer.

2.1.2 Polymerase Chain Reaction of Genomic DNA

device,

for

example,

a

UV-Vis

1. Hot Start High-Fidelity PCR reagents (e.g., Q5 Hot Start High-Fidelity 2 Master Mix, New England Biolabs). 2. Target-gene specific primer pairs with a max. distance of ~1.0 kb. 3. 0.2 mL PCR tubes with lids or 96-well PCR plates with resealable lid. 4. PCR Thermocycler.

2.1.3 Agarose Gel Electrophoresis

1. 50 TAE buffer: 2 M Tris, 1 M acetic acid, 100 mM EDTA. 2. Loading dye. 3. DNA marker with an appropriate fragment size range. 4. Agarose solution: 0.8–2.0% (w/v) agarose in 1 TAE buffer. Mix 6.4 gr (0.8%), 9.6 gr (1.2%) or 16 gr (2.0%) agarose with 800 mL 1 TAE in a 1 L Schott flask. Cover with loosely screwed lid and heat in microwave under constant supervision until agarose is completely dissolved and the solution looks clear. Allow to cool to 60  C, then add 0.25 μg/mL ethidium bromide before use and mix carefully. The solution can be stored at 60  C for 1 week (see Note 1). 5. Horizontal gel electrophoresis system with power supply. 6. UV Imaging system (e.g., Gel Doc XR+, Bio-Rad).

246

Julia Jansing and Luisa Bortesi

2.1.4 Spin-Column Purification of Linear DNA Fragments

1. Scalpel. 2. UV Imaging system. 3. Spin column kit for purification of linear DNA from agarose gel, for example, Macherey & Nagel NucleoSpin Gel and PCR Clean-up kit. 4. 96–100% ethanol. 5. Thermal heating block or water bath. 6. Vortex mixer. 7. DNA quantification spectrophotometer.

device,

for

example,

a

UV-Vis

2.1.5 Sanger Sequencing

1. Sequencing primers.

2.1.6 Analysis of Sanger Sequencing Results

1. Template sequence for the gene(s) of interest.

2.2 gRNA Design, Cloning and Testing

1. gRNA sequence synthesized as linear DNA fragment. Reconstitute the lyophilized DNA with sterile water to a final concentration of 50 ng/μL.

2.2.1 Cloning of gRNA Test Constructs

2. Software for aligning sequences and visualizing sequencing chromatograms, for example, Clone Manager (Scientific & Educational Software, Denver, USA).

2. Suitable binary plasmid for the expression of Cas9 and gRNA in plant tissue, either a single plasmid with Cas9 and gRNA gene on the same T-DNA or as separate plasmids for Cas9 and gRNA. 3. High-fidelity PCR reagents and suitable primers for amplification of the synthesized linear DNA fragment, see Subheading 2.1.2. 4. Material for agarose gel electrophoresis, see Subheading 2.1.3. 5. Material for DNA purification from agarose gel, see Subheading 2.1.4. 6. Restriction enzymes and suitable buffer. 7. Heating block. 8. T4 DNA Ligase and Ligase reaction buffer.

2.2.2 Transformation of Chemically Competent E. coli

1. Commercial or homemade chemically competent E. coli DH5α cells or another suitable recA- cloning strain (e.g., Subcloning Efficiency DH5α Competent E. coli, Thermo Fisher Scientific). 2. Heating block. 3. SOC (super optimal broth with catabolite repression) medium: 2% (w/v) tryptone, 0.5% (w/v) yeast extract, 10 mM NaCl, 2.5 mM KCl, 10 mM MgSO4, 10 mM MgCl2, 20 mM glucose. Store sterile prepared medium in 1 mL aliquots at
20  C.

Multiplex Glycosyltransferase Knockout in N. benthamiana

247

4. Shaker with 37  C. 5. Microcentrifuge. 6. Sterile Drigalski spatula. 7. LB medium cultivation plates: 1.0% (w/v) peptone, 0.5% (w/v) yeast extract, 0.1% (w/v) NaCl, 15 g/L agar, adjust to pH 7.0. Autoclave, add the appropriate antibiotic after medium is cooled but still liquid (e.g., 100 μg/mL ampicillin or 50 μg/ mL kanamycin), and pour plates before medium hardens. Keep plates at 4  C until use, but no longer than 1 month. 2.2.3 Colony PCR

1. LB plates with appropriate antibiotic, see Subheading 2.2.2. 2. Routine PCR reagents, for example, with Taq polymerase. 3. Thermocycler. 4. Agarose gel electrophoresis solutions and equipment, see Subheading 2.1.3.

2.2.4 Liquid Cultivation of E. coli

1. LB medium: 1.0% (w/v) peptone, 0.5% (w/v) yeast extract, 0.1% (w/v) NaCl, 15 g/L agar, adjust to pH 7.0. Autoclave and store at RT until use. 2. Cultivation vessel, for example, reaction tube for small-scale cultivation of 4–5 mL, or shaking flask for larger volumes. 3. Shaker in 37  C environment.

2.2.5 Plasmid Preparation from E. coli

1. Kit for plasmid preparation from E. coli, for example, Macherey & Nagel NucleoSpin Plasmid Mini kit. 2. 96–100% ethanol. 3. Microcentrifuge. 4. Vortex mixer. 5. Heating block. 6. DNA quantification spectrophotometer.

2.2.6 Preparation of Electrocompetent Agrobacterium tumefaciens Cells for Electroporation

device,

for

example,

UV-Vis

1. A. tumefaciens GV3101::pMP90RK wild-type cells as 50 μL glycerol stock, or other Agrobacterium strain depending on the binary plasmid that is used (see Note 2). 2. YEB medium with rifampicin and kanamycin: 0.5% (w/v) beef extract, 0.1% (w/v) yeast extract, 0.5% (w/v) peptone, 0.5% (w/v) sucrose, 2 mM MgSO4, pH 7.0. Autoclave and store liquid medium without antibiotics at RT. Add 25 μg/mL rifampicin and 25 μg/mL kanamycin freshly before use. 3. Sterile 1000 mL (or 2 500 mL) cultivation flask with lid or covering that allows aeration.

248

Julia Jansing and Luisa Bortesi

4. Shaker at 28  C. 5. Sterile 50 mL centrifugation tubes suitable for 4000  g. 6. Refrigerated centrifuge for 50 mL tubes, for example, Eppendorf centrifuge 5810R. 7. Sterile ddH2O. 8. Cold, sterile 10% (v/v) glycerol. 9. Liquid nitrogen and appropriate container. 2.2.7 Transformation of Electrocompetent A. tumefaciens

1. Electrocompetent A. tumefaciens GV3101::pMP90RK, see Subheading 2.2.6. 2. Prechilled electroporation cuvette with 2 mm electrode gap. 3. Electroporation device, for example, Eppendorf Multiporator. 4. YEB medium without antibiotics. 5. Sterile Drigalski spatula. 6. Shaker at 28  C. 7. YEB selection plates with 25 μg/mL rifampicin, 25 μg/mL kanamycin and 50 μg/mL carbenicillin (see Note 2): 0.5% (w/v) beef extract, 0.1% (w/v) yeast extract, 0.5% (w/v) peptone, 0.5% (w/v) sucrose, 2 mM MgSO4, 15 g/L agarose, pH 7.0. Autoclave and add antibiotics after the medium has cooled down. Mix carefully and pour plates. Let plates cool to RT lightly covered in sterile environment. Store at 4  C in sealed plastic bag until use, but no longer than 2 weeks.

2.2.8 Colony PCR of Electroporated A. tumefaciens

1. YEB plates with appropriate antibiotic, see Subheading 2.2.6. 2. Routine PCR reagents, for example, with Taq polymerase. 3. Thermocycler. 4. Agarose gel electrophoresis solutions and equipment, see Subheading 2.1.3.

2.2.9 Liquid Cultivation of A. tumefaciens

1. YEB medium with antibiotics, see Subheading 2.2.6. 2. Cultivation vessel, for example, reaction tube for 4–5 mL culture volume or shaking flasks for larger volumes. 3. Shaker at 28  C.

2.2.10

Agroinfiltration

1. OD600 Biophotometer. 2. Semi micro cuvettes. 3. YEB medium. 4. 2 infiltration medium: 10% (w/v) sucrose, 0.4% (w/v) glucose, 0.1% (w/v) Ferty 2 Mega, pH 5.6.

Multiplex Glycosyltransferase Knockout in N. benthamiana

249

5. Acetosyringone. 6. Disposable 1 mL syringes without needles. 7. 6–9-week-old N. benthamiana plants. 8. Plant trays. 9. Plant cultivation chamber with biosecurity level 1, set to 22  C and a 16 h photoperiod. 2.2.11 Analysis of gRNA Efficiency

1. Material for genomic DNA extraction from plants, see Subheading 2.1.1. 2. Target-gene specific primers and reagents for high-fidelity PCR, see Subheading 2.1.2. The primers for Ion Torrent sequencing need barcodes and adapter sequences: the forward primer includes the P1 adapter sequence (50 -CCTCTCTATG GGCAGTCGGTGAT -30 ), the reverse primer the A adapter sequence (50 - CCATCTCATCCCTGCGTGTCTCCGACT CAG-30 ) and the sample-specific barcode. 3. Equipment and material for agarose gel electrophoresis, see Subheading 2.1.3, and purification of PCR products from agarose gels, see Subheading 2.1.4. 4. Capillary electrophoresis: Agilent DNA 1000 kit, Agilent Bioanalyzer 2100 (Agilent Technologies). 5. Ion Library Equalizer kit (Life Technologies). 6. Ion PGM Sequencing 400 kit and AMPure beads (Life Technologies). 7. Ion Personal Genome Machine with an Ion 318 chip (Life Technologies). 8. DNASTAR Lasergene software package (DNAStar, Inc.).

2.3 Stable Transformation of Plants 2.3.1 Cloning of the Transformation Construct

1. Synthetic polycistronic tRNA-gRNA gene (PTG) [16] as plasmid. Dissolve lyophilized plasmid according to the manufacturer’s instructions and use 0.2–0.5 μL to transform recAcompetent E. coli, see Subheading 2.2.2, cultivate one positive clone, see Subheading 2.2.4, and purify the plasmid, see Subheading 2.2.5. 2. Suitable binary plasmid for plant transformation including a Cas9 cassette, an nptII cassette for plant selection, and suitable restriction sites to insert the PTG. The genetic elements should preferably be separated by insulators [18] or scaffold attachment regions (SAR) [19] to isolate them from each other transcriptionally. 3. Restriction enzymes and suitable buffer. 4. Material for agarose gel electrophoresis, see Subheading 2.1.3.

250

Julia Jansing and Luisa Bortesi

5. Material for DNA purification from agarose gel, see Subheading 2.1.4. 6. Heating block. 7. T4 DNA Ligase and Ligase reaction buffer. 8. Material for the transformation of chemically competent E. coli, see Subheading 2.2.2. 9. Material for colony PCR, see Subheading 2.2.3. 10. Material for liquid cultivation of E. coli, see Subheading 2.2.4, and plasmid preparation from E. coli, see Subheading 2.2.5. 2.3.2 Agrotransformation and Agroinfiltration

1. Electrocompetent A. tumefaciens, see Subheading 2.2.6. 2. Material for transformation of A. tumefaciens, see Subheading 2.2.7.

electrocompetent

3. Material for colony PCR of electroporated A. tumefaciens, see Subheading 2.2.8. 4. Material for liquid cultivation of A. tumefaciens, see Subheading 2.2.9. 5. Material for agroinfiltration Subheading 2.2.10. 2.3.3 Plant Transformation and Regeneration

of

leaf

materials,

see

1. 70% (v/v) ethanol. 2. Sterilization solution: dilute Domestos, Chlorox, or an equivalent 5.25% solution of sodium hypochlorite 1:1 with water. Prepare freshly. 3. Sterile H2O. 4. Scalpel. 5. Standard Petri dishes. 6. MS II medium: MS medium (0.44% (w/v) MS salts with vitamins, 2% (w/v) sucrose, 0.2 mg/L Thiamine-HCl, 0.8% (w/v) agar) with 1 mg/L 6-benzylaminopourine, 0.1 mg/L naphthalenic acid, 100 mg/L kanamycin, 200 mg/L cefotaxime, pH 5.8. Autoclave the medium without antibiotics, let it cool and add kanamycin and cefotaxime under sterile conditions before pouring plates. MS salts with vitamins from Duchefa (M 0222). 7. MS III medium: MS medium (0.44% (w/v) MS salts with vitamins, 2% (w/v) sucrose, 0.2 mg/L Thiamine-HCl, 0.8% (w/v) agar) with 100 mg/L kanamycin, 200 mg/L cefotaxime, pH 5.8. Autoclave the medium without antibiotics, let it cool and add kanamycin and cefotaxime under sterile conditions before pouring the plates.

Multiplex Glycosyltransferase Knockout in N. benthamiana

251

8. Sterile Weck jars. 9. Incubation chamber or cabinet for plant tissue culture plates with controlled temperature range, light intensity and light cycle. 10. Jiffies. 11. Standard soil and pots. 12. Incubation chamber or greenhouse with controlled temperature range, light intensity and light cycle. 2.4 Genetic Analysis of Regenerated Plants

1. Regenerated plants of the T0 generation.

2.4.1 Fast and Easy Extraction of Plant Genomic DNA

3. 0.2 mL PCR tubes with lids or 96-well PCR plates with resealable lid.

2. Wild-type plant as control.

4. Sampling tool with 2–3 mm diameter for leaf sampling. Alternatively, the blunt end of a 200 μL pipette tip can be used. 5. Small forceps. 6. Buffer A: 100 mM NaOH, 2% (v/v) Tween 20. Store at RT. 7. Buffer B: 100 mM Tris–HCl, 2 mM EDTA, pH 2.0. Store at RT. 8. Thermocycler.

2.4.2 Target Gene Amplification 2.4.3 Clean-Up of PCR Products

1. Material for high-fidelity PCR amplification of the target gene region of the target gene(s), see Subheading 2.1.2. Smaller number of samples: 1. Material for spin-column purification of PCR products, see Subheading 2.1.4. Larger number of samples: 1. Macherey & Nagel NucleoFast 96 PCR ultrafiltration kit. 2. Suitable vacuum manifold or microplate centrifuge. 3. Pipetting robot or multichannel pipette.

2.4.4 Sanger Sequencing of Purified PCR Products

1. Allele-specific primers. 2. Sterile deionized H2O. 3. Template sequence files and software for the analysis of Sanger sequencing chromatograms, see Subheading 2.1.5.

2.5 Immunoblot Analysis of Regenerated Plants

1. Leaf disk sampling tool or scalpel.

2.5.1 Dot Blot Analysis

4. 70% ethanol.

2. Scales. 3. Electric micro pestle, or mortar and pestle.

252

Julia Jansing and Luisa Bortesi

5. 10x PBS (Phosphate-buffered saline): 1.37 M NaCl, 27 mM KCl, 81 mM Na2HPO4, 15 mM KH2PO4. Autoclave and store at RT. Adjust pH 7.4 in 1 PBS. 6. Cold extraction buffer: 1 PBS, 500 mM NaCl, 10 mM Na2SO4. Store at 4  C before use. 7. Microcentrifuge. 8. Nitrocellulose membrane. 9. Whatman paper. 10. Container for membrane incubation. 11. 2D or 3D rocking shaker or similar device. 12. TBS-T (Tris-buffered saline with Tween 20): 20 mM Tris, 150 mM NaCl, 0.05% (v/v) Tween 20, pH 7.5. To add Tween 20, cut off the tip of a 1000 μL pipette tip and pipet the required volume slowly. 13. Skim milk powder in TBS-T: 1 TBS-T, 2% (w/v) skim milk powder. Dissolve milk powder under stirring for at least 30 min at RT before use. 14. Primary antibodies: Polyclonal rabbit-anti-α-1,3-fucose (Agrisera Antibodies), polyclonal rabbit-anti-β-1,2-xylose (Agrisera Antibodies). 15. Secondary antibody: Goat-anti-rabbit (H + L) AP-labelled (Jackson ImmunoResearch). 16. Alkaline Phosphatase (AP) buffer: 100 mM Tris, 100 mM NaCl, 5 mM MgCl2, pH 9.6. 17. AP substrate: dilute 100 μL BCIP/NBT in 10 mL AP buffer. This amount is sufficient for one blotting membrane of ~60 cm2. 2.5.2 Protein Extraction for SDS-PAGE

1. Leaf disk sampling tool or scalpel. 2. Scales. 3. Electric micro pestle, or mortar and pestle. 4. 70% ethanol. 5. Cold extraction buffer: 1 PBS, 500 mM NaCl, 10 mM Na2SO4. Store at 4  C before use. 6. Microcentrifuge.

2.5.3 SDS-PAGE

SDS-PAGE can be performed using commercial systems and precast gels as described above. Alternatively, self-made gels and buffers can be used with established protocols. 1. Microcentrifuge. 2. 4 NuPAGE LDS Sample Buffer.

Multiplex Glycosyltransferase Knockout in N. benthamiana

253

3. 10 NuPAGE Reducing Agent. 4. Thermoblock. 5. NuPAGE Novex 4–12% Bis-Tris protein gels. 6. NuPAGE MES SDS Running Buffer. 7. NuPAGE Antioxidant. 8. PageRuler Prestained Protein Ladder (Thermo Fisher) or alternative suitable protein standards. 9. Vertical gel electrophoresis system with power supply. 2.5.4 MicrowaveAssisted Coomassie Staining of Acrylamide Gels

Gel staining is performed according to a Coomassie Blue R-250 staining protocol by Wong et al. [20] adapted from Fairbanks et al. [21]. 1. Container for staining. 2. Microwave. 3. 2D or 3D rocking shaker or similar device. 4. Tissue paper. 5. Fairbanks A: 0.05% (w/v) Coomassie R-250, 25% (v/v) isopropyl alcohol, 10% (v/v) acetic acid. Store at RT. 6. Fairbanks B: 0.005% (w/v) Coomassie R-250, 10% (v/v) isopropyl alcohol, 10% acetic acid. Store at RT. 7. Fairbanks C: 0.002% (w/v) Coomassie R-250, 10% (v/v) acetic acid. Store at RT. 8. Fairbanks D: 10% (v/v) acetic acid. Store at RT.

2.5.5 Western Blot Analysis

1. Unstained SDS-PAGE gel with samples and controls. 2. Nitrocellulose membrane. 3. Whatman paper. 4. Blotting tank for wet transfer with lid, power supply, ice pack, sponges and cassettes. 5. Shallow tray or alternative container for assembly. 6. Cold blotting buffer: 25 mM Tris, 192 mM glycine, 20% (v/v) methanol). 7. Container for membrane incubation. 8. 2D or 3D rocking shaker or similar device. 9. 5% skim milk in TBS-T, see Subheading 2.5.1. 10. Primary and secondary antibodies, see Subheading 2.5.1. 11. AP buffer and substrate, see Subheading 2.5.1.

254

Julia Jansing and Luisa Bortesi

2.6 MassSpectrometric Analysis of Antibodies Produced in Regenerated Plants

1. Binary plasmid with heavy and light chain genes of a human antibody, for example, 2G12 [22].

2.6.1 Recombinant Antibody Expression

3. Material and equipment Subheading 2.2.10.

2.6.2 Antibody Purification

1. Infiltrated frozen leaf material.

2. Material and equipment for transformation of A. tumefaciens, see Subheading 2.2.7, colony PCR, see Subheading 2.2.8, and cultivation, see Subheading 2.2.9. for

agroinfiltration,

see

2. Scales. 3. Mortar and pestle. 4. Cold protein extraction buffer: 1 PBS, 500 mM NaCl, 10 mM Na2SO4. Store at 4  C before use. 5. Miracloth (Merck). 6. 1 M Tris pH 8.0. 7. 25 or 50 mL centrifugation tubes. 8. Refrigerated centrifuge for 25 and 50 mL tubes, for example, Eppendorf centrifuge 5810R. 9. Chromatography column with filter. 10. Protein A Ceramic HyperD F (Pall). 11. 1 PBS, see Subheading 2.5.1. 12. 2G12 elution buffer: 100 mM glycine, 100 mM fructose, pH 3.6. 13. 1 M acetate buffer pH 4.75. 14. UV-Vis spectrophotometer, for example, ND-1000 (Thermo Fisher Scientific).

2.6.3 SDS-PAGE

NanoDrop

1. Antibody elution fraction from Subheading 2.6.2. 2. Material and equipment for SDS-PAGE as described in Subheading 2.5.2. 3. Material and equipment Subheading 2.5.4.

2.6.4 LC-ESI-MS

for

Coomassie

staining,

1. 3–6 SDS-PAGE gel lanes of antibody heavy chain. 2. Scalpel. 3. Deionized H2O.

see

Multiplex Glycosyltransferase Knockout in N. benthamiana

2.7 MassSpectrometric Analysis of Endogenous Leaf Proteins of Regenerated Plants

255

1. 5–6 young leaves from different parts of the plant. 2. 50 mL falcon tubes.

2.7.1 MALDI-TOF-MS

2.8 Manual Crossing of Plants

1. Plants of similar age of the two or more lines to be crossed. 2. Labels. 3. Small paper bags.

3

Methods

3.1 Generation of Transgenic Plants 3.1.1 Target Gene Resequencing Primer Design

1. Retrieve the genomic or mRNA sequence of your gene of interest from a public database such as NCBI. 2. BLAST the gene sequence against the genome sequence to search for homologous genes. For several Solanaceae species such as N. benthamiana, genome drafts are available on https://solgenomics.net [23]. Search nucleotide sequence for nucleotide sequence, and then translated nucleotide to translated nucleotide sequence. This allows the identification of less closely related gene variants with a nevertheless similar amino acid sequence. As a rule of thumb, allotetraploid plants such as N. benthamiana have at least two variants per gene, one from each parental genome. If the protein encoded by your gene of interest has a preserved motif or catalytic site, this can be used to identify more distantly related gene homologues as well. 3. Extract the genomic sequence covering everything from 1 kb upstream of the gene sequence to 1 kb downstream. 4. Identify exon and intron sequences and annotate your sequence file accordingly. 5. Prescreen the exons to identify 2–3 you will target with Cas9. It is not advisable to use the last exon as target, because this can result in an incomplete gene knockout later on. If a conserved motif or catalytic site is available, then the respective exon should be included as potential target. For genes with numerous exons, exons in the first half or 2/3 of the gene are preferable. If your gene of interest encodes a membraneanchored protein, targeting a region between membrane anchor and protein itself could result in a free form of the protein. Check the selected exons for the availability of the NGG PAM sequence. Select 2–3 exons to proceed.

256

Julia Jansing and Luisa Bortesi

6. Align all sequences found for one gene, for example, XylT 1 and XylT 2, to identify regions suitable for designing primers that are specific for each gene variant. The goal is to amplify only one gene variant with a given primer pair. To ensure that the Sanger sequencing later on covers complete exons, it is recommended to design primers for noncoding sequences such as introns or the 50 or 30 UTR. Design primer pairs suitable for the amplification of the exons selected in the previous step. Allow for a minimum distance of at least 50 nt between a primer site and the region for which you need sequencing information (see Note 3). 7. Order the primers as standard DNA oligonucleotides. Spin-Column Extraction of Plant Genomic DNA

1. Prepare all components of the genomic DNA extraction kit according to the manufacturer’s instructions. 2. Harvest 100 mg of young leaf material from a wild-type N. benthamiana plant. Store on ice or at 4  C for processing within a day, or at
20  C for longer storage. 3. Homogenize the leaf material with mortar and pestle in liquid nitrogen to a fine powder. Once the nitrogen evaporated but before the powder thaws, transfer the powder to a new tube with a chilled spatula or little spoon and proceed with the extraction according to the kit instructions (see Note 4). 4. Quantify the eluted genomic DNA with a UV-Vis spectrophotometer or alternative device. Concentrations >10 ng/μL are sufficient for PCR amplification. Store at 4  C, or aliquot and store at
20  C. Avoid repeated freeze–thaw cycles.

Polymerase Chain Reaction of Genomic DNA

1. Resolving and diluting DNA oligonucleotides: briefly centrifuge the tube containing the lyophilized oligonucleotide to collect the material at the tube bottom. Add the required amount of water to obtain a stock concentration of 100 μM (i.e., multiply the nmol amount of the primer in the tube by 10 to obtain the amount of water in μL needed to make a 100 μM stock solution). Dissolve on ice for 30 min, vortex, and store at
20  C. Dilute the stock solution 1:10 with water to prepare the 10 μM working stock (i.e., mix 10 μL of 100 μM primer solution with 90 μL water), and store at
20  C (see Note 5). 2. Thaw 10 μM primer stocks and Q5 Hot Start High-Fidelity 2 Master Mix or your high-fidelity PCR reagents of choice on ice. Vortex the 2 Master Mix briefly until any precipitate dissolves. 3. Prepare a PCR master mix for each primer pair containing everything except the genomic DNA (Table 1). Calculate an

Multiplex Glycosyltransferase Knockout in N. benthamiana

257

Table 1 PCR master mix composition 1 (20 μL)

11 (Master mix for ten samples)

Q5 2 master mix

10 μL

110 μL

Forward primer 10 μM

1 μL (0.5 μM)

11 μL

Reverse primer 10 μM

1 μL (0.5 μM)

11 μL

Genomic DNA

0.5–1.0 μL

–

ddH2O

To 20 μL

To 11  19.5 μL or to 11 19.0 μL

additional 10% of reaction mix for pipetting inaccuracies. The final reaction volume per sample is 20 μL (see Note 6). 4. Briefly vortex the PCR master mix and spin down in a microcentrifuge. Dispense the required volume (20 μL minus the volume of genomic DNA) into PCR tubes or plates and add the genomic DNA. Close the lid and tap lightly to mix. Store on ice until thermocycling. 5. PCR program for Q5 Hot Start polymerase: initial denaturation (98  C, 2 min), 30–34 amplification cycles (98  C 10 s, 60–72  C 30 s, 72  C 20–30 s/kb), final extension step (72  C, 2 min). Hold at 4–10  C. Follow the manufacturer’s instruction if working with a different PCR system (see Note 7). Agarose Gel Electrophoresis

1. Prepare the agarose gel at least 30–60 min before loading to allow for proper gelation. Use 0.8% agarose to separate large DNA fragments >1 kb, 1.2% agarose for DNA fragments ~0.4–2 kb, and 2% agarose for small DNA fragments Set Measurements. . . and select Area, Mean gray value and Min & Max gray value (see Note 14). 3. Open the ROI Manager, select Analyse > Tools > ROI Manager. . . 4. Use the appropriate selection tool (Circle or Freehand) to select a region of interest (ROI) (see Note 15). 5. Click on Add in the ROI Manager. 6. Repeat the selection process and add ROIs. 7. When all ROIs are selected, click on Measure in the ROI Manager to open the Results window. 8. Copy and paste the results into a spreadsheet for further analysis (see Note 16). 9. Subtract background values from each image based on the average value of noninfiltrated zones on the leaves.

3.3.3 Apply a Look-Up Table (LUT) to Color Images Using ImageJ

1. Import images (.tiff file) into ImageJ to apply a LUT (Fig. 1) (see Note 17). 2. Import LUT, select File > Import > LUT. . ., find the LUTs subfolder in the ImageJ program folder and select the appropriate LUT (see Note 18). 3. Scale the image brightness, select Image > Adjust > Brightness/ Contrast. . . and click on Auto (see Note 19). 4. Add a calibration bar, select Analyse > Tools > Calibration bar. . . . 5. Save the image.

3.3.4 Leaf Discs Data Acquisition

1. Add 200 μL water in wells of a 96-well plate. 2. Punch discs from agroinfiltrated leaves (6 mm diameter max) and gently add one leaf disc per well (see Note 20). 3. Insert the plate in the Infinite M200 reader (see Subheading 2.3), and keep the plate in the dark for 5 min prior to bioluminescence measurements (see Note 12).

AgroLUX Luminescence

291

4. Set the plate reader to luminescence measurement with an integration time of 1 s. 5. Save the data in a spreadsheet for further analysis.

4

Notes 1. N. benthamiana plants are widely used by plant molecular biologists to transiently express genes by agroinfiltration [11]. AgroLux can be used to monitor Agrobacterium activity in various plant species that are suitable for agroinfiltration [12]. 2. AgroLux (unique ID RSPJ07) is a bioluminescent derivative of Agrobacterium tumefaciens strain GV3101 that carries a singlecopy genomic integration of the luxABCDE operon driven by the constitutive nptII promoter [9]. The strain is available upon request. 3. AgroLux can be used as a bioluminescent reporter postagroinfiltration by mixing other Agrobacterium cultures carrying expression vectors or as a recipient of binary plasmids of interest [9]. 4. Common binary expression plasmids have a kanamycin resistance gene and can be selected by supplementing the LB medium with 50 μg/mL kanamycin (Kan50). 5. Before using freshly prepared AgroLux competent cells, it is a good practice to test an aliquot. Spread the bacterial cells on different plates containing common antibiotics, such as kanamycin (Kan50), streptomycin (Strep50) or carbenicillin (Carb100). AgroLux cells should not grow on any plate containing antibiotics, except for gentamicin (Gem50) and rifampicin (Rif25). 6. AgroLux can also be transformed through electroporation. See ref. [13] for alternative methods for A. tumefaciens transformation. 7. The youngest, fully expanded leaves of 4–6-week-old N. benthamiana are easy to infiltrate from the abaxial (lower) side with a needleless syringe and cause a good transient protein expression. When coexpression with P19 silencing inhibitor [14], mix the culture containing the binary plasmid encoding P19 with the culture containing the gene of interest (normally 1:1) prior to infiltration. 8. For transient protein expression, it is recommended to harvest plant tissue at 5 dpi. For imaging bioluminescence of AgroLux, bioluminescence can be monitored throughout 7 dpi and more.

292

Philippe V. Jutras et al.

9. If images are not taken within 30 min after the leaves have been detached, it is recommended to keep the detached leaves in a plastic box with a wet tissue to maintain humidity high. 10. The tray position of the imaging system can be adjusted to measure 2–6 leaves in one image. 11. Bioluminescence levels are stronger when leaves are imaged from the abaxial (lower) side where they have been infiltrated. 12. Keeping the leaves in the dark before bioluminescence measurements reduces the background coming from the chloroplasts. 13. ImageJ allows image pixel quantification. It is recommended to follow the Image Intensity Processing protocol from ImageJ documentation (https://imagej.nih.gov/ij). 14. When setting measurements, tick the relevant boxes. Min & Max gray value shows the background (Min) and if pixels are saturated (Max). 15. Alternatively to the selection tools, select Image > Adjust > Threshold. . . and adjust the pixel thresholds to highlight only the ROIs. Then, use the Wand (tracing) tool to select ROIs and add your selections to the ROI Manager. 16. Use the Mean values for selected ROI intensities. 17. Applying a LUT allows to display grey values as colorized pixels. The colorized pixels in the pseudocolored image reflect differences in pixel intensity. 18. To determine the appropriate LUT, select Image > Color > Display LUTs. . . to open the look-up tables. 19. To compare multiple images, select Set to adjust the brightness with specific values. 20. It is important that leaf discs are floating at the top of the water and are not submerged. Leaf discs should be placed with the abaxial side up to obtain the brightness signals and reduce variability.

Acknowledgments This work was financially supported by ERC project 616449 “GreenProteases” (RvdH); H2020 project 760331 “Newcotiana” (PVJ); the Quebec Government’s research funding body FRQNT (PVJ); and BBSRC iCASE (ID).

AgroLUX Luminescence

293

References 1. Yao Z, Zhang BS, Prescher JA (2018) Advances in bioluminescence imaging: new probes from old recipes. Curr Opin Chem Biol 45:148–156 2. Fleiss A, Sarkisyan KS (2019) A brief review of bioluminescent systems. Curr Genet 65: 877–882 3. Soldan R, Sanguankiattichai N, Bach-Pages M, Bervoets I, Huang WE, Preston GM (2021) From macro to micro: a combined bioluminescence-fluorescence approach to monitor bacterial localization. Environ Microbiol 23:2070–2085 4. Gregor C, Gwosch KC, Sahl SJ, Hell SW (2018) Strongly enhanced bacterial bioluminescence with the ilux operon for single-cell imaging. Proc Natl Acad Sci U S A 115: 962–967 5. Brodl E, Winkler A, Macheroux P (2018) Molecular mechanisms of bacterial bioluminescence. Comput Struct Biotechnol J 16: 551–564 6. Yeh HW, Ai HW (2019) Development and applications of bioluminescent and chemiluminescent reporters and biosensors. Annu Rev Anal Chem 12:129–150 7. Fan F, Crooks C, Lamb C (2008) Highthroughput quantitative luminescence assay of the growth in planta of Pseudomonas syringae chromosomally tagged with Photorhabdus luminescens luxCDABE. Plant J 53:393–399 8. Xu X, Miller SA, Baysal-Gurel F, Gartemann KH, Eichenlaub R, Rajashekara G (2010) Bioluminescence imaging of Clavibacter

michiganensis subsp. michiganensis infection of tomato seeds and plants. Appl Environ Microbiol 76:3978–3988 9. Jutras PV, Soldan R, Dodds I, Schuster M, Preston GM, Van der Hoorn RAL (2021) AgroLux: bioluminescent Agrobacteria to improve molecular pharming and study plant immunity. Plant J (resubmitted) 10. Hwang HH, Yu M, Lai EM (2017) Agrobacterium-mediated plant transformation: biology and applications. Arab B 15: e0186 11. Bally J, Jung H, Mortimer C, Naim F, Philips JG, Hellens R, Bombarely A, Goodin MM, Waterhouse PM (2018) The rise and rise of Nicotiana benthamiana: a plant for all reasons. Annu Rev Phytopathol 56:405–426 12. Zhang Y, Chen M, Siemiatkowska B, Toleco MR, Jing Y, Strotmann V, Zhang J, Stahl Y, Fernie AR (2020) A highly efficient agrobacterium-mediated method for transient gene expression and functional studies in multiple plant species. Plant Commun 1:100028 13. Wise AA, Liu Z, Binns AN (2006) Three methods for the introduction of foreign DNA into Agrobacterium. Methods Mol Biol 343:43–53 14. Van der Hoorn RAL, Rivas S, Wulff BB, Jones JDG, Joosten MHAJ (2003) Rapid migration in gel filtration of the Cf-4 and Cf-9 resistance proteins is an intrinsic property of Cf proteins and not because of their association with highmolecular-weight proteins. Plant J 35: 305–315

Chapter 16 Statistical Designs to Improve Downstream Processing Johannes F. Buyel Abstract The efficient extraction and purification of recombinant proteins from leaf and seed tissues is often a challenging task, involving multiple steps that must be optimized by identifying and accommodating complex parameter interactions. Conventional one-factor-at-a-time approaches fail to reveal these complex interactions and often result in sub-optimal processes with unnecessary costs. Here, we describe generic considerations to identify global optima for the extraction and purification of recombinant proteins from complex plant matrices. The corresponding experiments can help to streamline downstream processing by reducing the time, costs, and number of unit operations. The procedure involves the knowledge-based selection of factors for screening, the systematic design and analysis of experiments, and the iterative refinement of suitable conditions. The resulting descriptive models can be used to guide process scale-up and offer scientific justifications for process development decisions in negotiations with regulatory authorities. Key words Descriptive models, Design of experiments, Multiparameter optimization, One-factor-ata-time (OFAT), Pareto optimality, Process optimization

Abbreviations DoE FDS TSP

1

Design of experiments Fraction of design space Total soluble protein

Introduction Numerous publications have described the benefits of producing recombinant pharmaceutical proteins in plants, including the low cultivation costs compared to fermenter-based systems [1–3]. However, downstream processing has been a major cost driver in the past because recombinant proteins expressed in plants usually accumulate inside tissues such as leaves and seeds, and must be recovered by cellular disruption, which releases large quantities of host

Stefan Schillberg and Holger Spiegel (eds.), Recombinant Proteins in Plants: Methods and Protocols, Methods in Molecular Biology, vol. 2480, https://doi.org/10.1007/978-1-0716-2241-4_16, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

295

296

Johannes F. Buyel

Fig. 1 Typical stages of downstream process optimization using statistical experiments. First, factors and their interactions with a relevant effect on a response (black line) are identified from the large pool of parameters by a simple screening approach using linear models (stage 1, orange). Next, factor ranges are adjusted in the direction of the expected optimum (arrows) and the factor effects on the response(s) are quantified using models of adequate complexity in a response surface methodology (stage 2, green). Finally, the optimal region selected from the response surface is verified by independent confirmation runs (stage 3, blue)

cell proteins and other impurities [4–6]. The higher costs of downstream processing have been addressed and overcome by innovations that improve the efficiency of clarification [7]. These improvements were achieved in many cases by applying statistical experimental designs [8–10] that facilitate the knowledge-based, structured and iterative optimization of extraction, filtration, flocculation, and purification. Here, we describe how such experimental strategies can be derived for individual downstream processing steps with limited information about the corresponding unit operation (see Subheading 3.2.1) or based on an existing body of data, mechanistic understanding, and expert knowledge (see Subheading 3.2.2). The generic approach consists of screening, optimization and confirmation (Fig. 1) [11, 12], and is structured into stages of information collection (see Subheading 3.1), design setup (see Subheading 3.2), experimental work (see Subheading 3.3), and data analysis (see Subheading 3.4).

2

Materials Because of the generic nature of the approach, the list of special materials is short. Other materials appropriate for the downstream process to be optimized should be selected by the experimenter. 1. Design-Expert software (Statease Inc., USA)—the most recent version (see Note 1).

DSP Improvement

297

2. Spreadsheet software package (e.g., Microsoft Excel). 3. Relevant literature about the unit operation to be optimized.

3

Methods

3.1 Plan a Design of Experiments Strategy

This procedure is used to collect information relevant for the subsequent construction of an experimental design, also known as a design of experiments (DoE) approach (Fig. 2).

3.1.1 Define the Context, Goal, and General Procedure of the Experiment

1. Describe the context of the experiment to focus all subsequent considerations, ideally in one sentence. For example, the context can be the purification of a recombinant protein by chromatography. 2. Define a goal for the experiment that is specific, measurable, achievable, relevant and time-bound (SMART). For example, within 3 weeks, purify the target protein such that it represents 95% of the total soluble protein while achieving an overall recovery of 70%. 3. Inform all staff involved in the experiment that the experimental plan (see Subheading 3.3) is to be followed and changes in the run order cannot be made.

3.1.2 Identify Relevant Factors and Responses for Inclusion in the Design

1. Define one or several responses for measurement that characterize the downstream step to be optimized, for example product purity and step recovery. The responses should be continuous numeric parameters (see Notes 2–5). 2. For each response, define a minimum detectable difference between two experimental outcomes that is regarded as relevant, for example a difference in purity of 3% total soluble protein (TSP). Estimate the standard deviation of the system based on previous experiments (see Note 6). 3. Use the available literature, data from previous experiments and/or expert knowledge to identify parameters that might affect the responses in the experimental setting defined under Subheading 3.1.1 (see Note 7). 4. As a team, set up a parameter–effect matrix to rate the potential effects of the parameters on the responses, including the expected type of correlation (e.g., linear or quadratic) and potential interactions (see Note 8). Involve several people to ensure substantial expertise is available. 5. Categorize the parameters in the parameter–effect matrix into: (1) irrelevant parameters that are expected to have no effect on the responses; (2) random error terms that cannot be controlled in the experiment and may even be unknown (e.g., ambient pressure); (3) fix parameters that may affect the

298

Johannes F. Buyel

Fig. 2 Workflow to collect relevant information before setting up a statistical experimental plan. External data (orange) as well as experimental, laboratory and application specific considerations (blue) are incorporated at different stages to trim the conditions that are ultimately transferred to the software used to build the design of experiments

DSP Improvement

299

responses but that remain constant during the experiment due to process constraints (e.g., column volume); and (4) factors that will be varied in a defined manner in the experiment (see Note 9). 6. Assign a type and subtype to each factor. Factors can be numeric type with continuous (e.g., concentration) or discrete (e.g., number of filtration steps) subtypes, or categoric type with nominal (e.g., devices A–Z) or ordinal (e.g., blender power low or high) subtypes (see Note 10). 7. For each numeric factor, define a relevant value range based on the current mechanistic understanding (see Notes 11 and 12). Keep in mind physicochemical boundaries such as compound solubility as well as equipment constraints such as maximum operational pressure. 8. For each categoric factor, define the levels that will be tested in the design. 3.2 Prepare a Specific Experimental Design

This procedure compiles a statistical design based on the planning (see Subheading 3.1) and development status of the process step to be optimized.

3.2.1 Set Up a Factorial Screening Design

Selected this procedure if there is little or no prior experimental information about the dependencies between factors and responses, or if the number of factors is large (>7). If there is a good understanding of the relevant factors, skip this section and follow the steps under Subheading 3.2.2. 1. Start Design-Expert and select “New Design.” In the “Factorial” node, choose “Randomized” and then “Regular Two-Level.” From the overview pane, select a design that matches the total number of numeric and categoric factors selected under Subheading 3.2.1, step 6 and that has a resolution of at least V (see Note 13). 2. If the expected effect complexity is of a higher order than linear for any of the selected factors (see Subheading 3.1.2, step 4), enter “5” in the “Center points per block.” As a start, choose the number of replicates and number of blocks as “1” (see Notes 14 and 15). 3. Continue to the next, optional page and confirm the alias structure if any and proceed to the next page. 4. Enter factor names, units, types, and subtypes. For each numeric factor, extract the low and high values of the range defined before (see Subheading 3.1.2, step 7). For each categoric factor enter the level names in the corresponding fields (see Note 16).

300

Johannes F. Buyel

5. On the next page, enter the response names, units, the difference to detect, and standard deviation (see Subheading 3.1.2, step 2). 6. Open the base model menu by clicking on the “Edit model . . .” button and either manually select the factors and interactions that may have an effect on the response based on the previous assessment (see Subheading 3.1.2, step 4) or use the “Process order” dropdown menu for quick selection (see Notes 17 and 18). 7. Continue to the next page and make sure that the power is >0.80 for each factor and interaction. If the power is 0.80 and differ by less than 0.2. 8. Evaluate the diagnostics in the corresponding tab to confirm the quality of the model and detect potential extreme values in the dataset that have a strong influence on the model. Specifically, data in the “Normal plot” and “Predicted vs. Actual” should scatter close to a straight line, indicating a normal distribution of the residuals and a good predictive power of the model, respectively; the “Residuals vs. Predicted,” “Residuals vs. run,” “Residuals vs. Factor,” “DFFITS,” and “DFBETAS” should exhibit random scatter; and no points should fall outside the boundary in the “Cook’s distance” and “Leverage” plots (see Note 29). 9. If the Box–Cox plot indicates that a transformation can help to improve the model quality, go back to step 3, transform the data accordingly and iterate through steps 4–9 (see Note 30). 10. In the “Model graphs” tab, visualize the evaluated model. Pay special attention to inconsistent model predictions, for example predicted response values (especially at the edges of the design) that are by a factor or even several orders of magnitude larger than the smallest or highest values determined experimentally. If such inconsistencies occur, go back to step 5 and modify the model, for example by reducing its complexity. 11. Use the “Numerical” sub-node in the “Optimization” node to optimize a single response or multiple responses simultaneously. Select the appropriate numerical optimization goal (minimize, maximize, etc.) for each response (see Note 31). 12. Go to the “Solutions” tab which will automatically trigger the optimization calculation based on the entered criteria and list the results in a table if the “Report” illustration style is selected in the functions ribbon. The “Graphs” tab can help to localize the calculated optimum in the parameter space (see Notes 32 and 33). 13. In the “Post analysis” node, go to “Confirmation” and enter “3” into the “Response data” field to use at least three runs to confirm the optimal condition. These runs will be conducted as a new set of experiments and used to verify the model predictions. 14. Enter the response values obtained for the confirmation runs in the corresponding response data fields. Go to the “Confirmation” tab and check if the average confirmation results (mean of the confirmation runs for a specific condition) fall within the 95% prediction interval of the model. A match confirms the predictive power of the model. A mismatch indicates a low-quality model and additional runs or a refined model may be required (see Notes 35 and 36).

DSP Improvement

4

305

Notes 1. The author uses the most recent versions of Design-Expert and GNU Octave (open source) software for all designs and numeric work due to the ease of handling and open source status, respectively, but several alternative software packages are available including JMP, MatLab, MiniTab, Modde, Statistica and others. 2. Responses that use discrete, ordinal or even subjective grading should be avoided as they are incompatible with the statistical analysis implemented in the DoE software. Binary responses like yes or no, 0 or 1, can be converted into a quasinumeric form by increasing the number of replicates per factor combination and using the resulting likelihood of the binary outcome instead. 3. The response should yield values with high precision (i.e., low standard deviation) and accuracy (i.e., zero or few interfering parameters) to improve the model quality. Suitable detection devices and methods should therefore be optimized before the experiment if necessary. 4. A response can be a combination of several primary parameters. For example, the purity is often defined as the target protein concentration in a sample divided by the total protein concentration in that sample. 5. When selecting a response, consider the complexity of its measurement. For example, a screening experiment can consist of more than 50 runs, so the response and the method required to obtain it should be compatible with this sample throughput. 6. If there is no estimate for the standard deviation of the experiment, use half the initially defined effect size as a conservative estimate. 7. Parameters that cannot be controlled directly in the experiment may be considered as long as they can affect the response, for example ambient pressure. 8. In a biotechnological context, interactions between two and even among three factors are commonly observed, whereas higher-order interactions are often negligible. 9. Make sure that the levels of fixed parameters are tightly controlled and that as many random error terms as possible are monitored. 10. Avoiding categoric factors can simplify a design by reducing the number of runs. In some cases, an apparently categoric factor may be converted to a numeric counterpart. For example, instead of membranes A, B and C, the permeability of the

306

Johannes F. Buyel

membranes can be used as a discrete numeric factor. However, care must be taken if the characteristic selected to convert the categoric factor into a numeric one (e.g., permeability) does not account for all relevant properties of the categoric factor. 11. When considering numeric factors with continuous variation, avoid using values or intervals that are difficult or impossible to implement precisely. For example, ambient temperature can typically be controlled only within
2  C, so continuous variation using values such as 27.2  C, 25.9  C, and 29.3  C should be avoided. Instead, select a number of discrete levels for such factors, matching the anticipated base model (see Subheading 3.1.2, step 4 and Note 22). The number of levels is relevant for optimal designs because factorial and central composite designs have a predefined number of factor levels. 12. If the value range of a discrete numeric factor spans several orders of magnitude, consider transforming the actual values (e.g., logarithmically) before entering them into the design software to achieve a close to equidistant distribution of the levels. This greatly improves model prediction and stability. The same recommendation also applies to the extraction of the mean of a continuous numeric factor for factorial screening designs. 13. If any interaction between two factors can be excluded, a resolution IV design may also be acceptable. Otherwise, factor interactions will be aliased, which means that an analysis will not be able to distinguish if the interaction between factors A and B or the interaction between factors C and D was responsible for the observed effect. 14. If the power of a factorial design is 0.70–0.80, increase the number of center points until the power is >0.80. If the power is 10 for the values of a response. The most useful transformation can be obtained from the Box–Cox plot in the “Diagnostics” tab in the “Diagnostics” section of the “Diagnostics tool” described under Subheading 3.4.1, step 8 and Subheading 3.4.2, step 9. A log10 transformation is often appropriate. 27. If prompted by the software, choose to restore the model hierarchy because this will result in more stable models [13]. 28. Factors that are expected to have an effect may sometimes turn out to have intermediate significance (0.05 < p < 0.10). In such cases, it can be helpful to establish models with and without these factors to inspect their relevance, that is, effect on the position of the optimum. 29. In the diagnostic plots, any pattern in the residual scattering can indicate systematic problems such as nonnormal distribution, unequal variances, background trends or flawed data and should be investigated in detail. A precise analysis of what needs to be done is beyond the scope of this chapter. Briefly, a sigmoidal shape in the normal plot, a chevron shape in the residuals vs. predicted, and a corresponding suggestion in the Box–Cox plot indicate that data transformation may be necessary. A trending in the residuals vs. run can indicate background effects (such as equipment wearing out) that may be compensated by block building. Points outside the boundaries of Cook’s distance or leverage indicate an imbalanced design that may require augmentation. 30. The suggestions for data transformation may get stuck in an infinite loop of transformation A > B > A. At this point, compare the performance of the models based on the coefficients of determination and diagnostics, and select the most appropriate model and transformation. 31. Optimization goals can also be defined for factors. The relevance of the parameters to be optimized can be balanced using the importance dropdown menu (balance between different parameters) and weights definition (relevance of values and ranges for a single parameter). 32. If two or more solutions share the highest desirability, increase the stringency of the optimization criteria, for example by increasing the upper limit in case of a maximization, or by fine tuning the weights and importance of the parameters to be optimized. 33. The optimal solutions can be exported to spreadsheet software for further analysis, for example to build histograms revealing the frequency of specific factor settings associated with high or low response values.

DSP Improvement

309

34. If center points are included in the design, check for the curvature in the model. If the curvature is not significant, expand the parameter space in the direction of the optimum identified for the response(s). If center points are included and curvature is significant, analyze the direction of curvature with respect to the optimization goal. For example, if the goal is to maximize the response, the curvature can either indicate a maximum or a minimum within the current parameter space. In the former case, proceed as directed under Subheading 3.2.2 to set up an optimization design because the optimum is probably located within the parameter space investigated in the screening design. If in the example the curvature indicates a minimum, conduct a new screening design including the factors identified as being significant and consider substantially expanding the parameter space in the direction of the optimum (i.e., there may be no overlap in the parameter ranges of the initial and second factorial screening design). 35. To augment a design, go to the “Design” node and select “Design Tools” from the menu bar. Choose “Augment design” and “augment” to add further runs to the design while retaining dependence on the available data. For example, if the predicted R2 is low (