Ribosome display for selection and evolution of affibody molecules

SEBASTIAN GRIMM

Doctoral Thesis in Biotechnology Stockholm, Sweden 2011

Ribosome display for selection and evolution Ribosome display for selection and evolution of affibody molecules of affibody molecules

Sebastian Grimm Sebastian Grimm

Doctoral Thesis Doctoral Thesis

KTH Royal Institute of Technology KTH Royal Institute of Technology School of Biotechnology School of Biotechnology Stockholm 2011 Stockholm 2011

II II

© Sebastian Grimm © Sebastian Grimm Stockholm 2011 Stockholm 2011

Royal Institute of Technology Royal Institute of Technology School of Biotechnology School of Biotechnology AlbaNova University Center AlbaNova University Center SE-106 91 Stockholm SE-106 91 Stockholm Sweden Sweden

Printed by Printed by AJ E-print AB AJ E-print AB Oxtorgsgatan 9, 11157 Stockholm Oxtorgsgatan 9, 11157 Stockholm ISBN 978-91-7415-952-3 ISBN 978-91-7415-952-3 TRITA BIO Report 2011:9 TRITA BIO Report 2011:9 ISSN 1654-2312 ISSN 1654-2312 III III

Sebastian Grimm (2011): Ribosome display for selection and evolution of affibody Sebastian Grimm (2011): Ribosome display for selection and evolution of affibody molecules. School of Biotechnology, Royal Institute of Technology (KTH), Stockholm, molecules. School of Biotechnology, Royal Institute of Technology (KTH), Stockholm, Sweden. Sweden.

Abstract Abstract

Affinity proteins are invaluable tools in biotechnological and medical applications. This thesis is about Affinity proteins are invaluable tools in biotechnological and medical applications. This thesis is about combinatorial protein engineering principles for the generation of novel affinity proteins to purify combinatorial protein engineering principles for the generation of novel affinity proteins to purify mouse immunoglobulin, detect a potential cancer marker protein or inhibit a cell proliferation mouse immunoglobulin, detect a potential cancer marker protein or inhibit a cell proliferation pathway. pathway.

In a first study, ribosome display was for the first time applied to the selection of so-called affibody In a first study, ribosome display was for the first time applied to the selection of so-called affibody molecules, including the design of a ribosome display gene cassette, initial test enrichment molecules, including the design of a ribosome display gene cassette, initial test enrichment experiments and the selection of binders against murine IgG1. One of the selected binders (ZMAB25) experiments and the selection of binders against murine IgG1. One of the selected binders (ZMAB25) showed a highly selective binding profile to murine IgG1, which was exploited in the recovery of two showed a highly selective binding profile to murine IgG1, which was exploited in the recovery of two different mouse monoclonal IgG1 antibodies from a bovine immunoglobulin-containing background. different mouse monoclonal IgG1 antibodies from a bovine immunoglobulin-containing background. Ribosome display was further applied to the selection of affibody molecules binding to SATB1, a Ribosome display was further applied to the selection of affibody molecules binding to SATB1, a suggested marker protein for metastasizing adenocarcinoma. The study also included the selection of suggested marker protein for metastasizing adenocarcinoma. The study also included the selection of VHH antibody fragments from a naïve gene repertoire displayed on phage. Binders from both classes VHH antibody fragments from a naïve gene repertoire displayed on phage. Binders from both classes of protein scaffolds could be isolated that selectively recognized SATB1 but not its close homologue of protein scaffolds could be isolated that selectively recognized SATB1 but not its close homologue SATB2, and were used to detect endogenous SATB1 in Jurkat cells by immunofluorescence SATB2, and were used to detect endogenous SATB1 in Jurkat cells by immunofluorescence microscopy. The well-established phage display technology was used to select affibody molecules microscopy. The well-established phage display technology was used to select affibody molecules binding to H-Ras and Raf-1, both involved in the mitogen-activated protein kinase (MAPK) pathway binding to H-Ras and Raf-1, both involved in the mitogen-activated protein kinase (MAPK) pathway and playing a central role in the control of cell proliferation, survival and differentiation. An isolated and playing a central role in the control of cell proliferation, survival and differentiation. An isolated affibody molecule denoted ZRAF322 was found to selectively bind to Raf-1 and inhibit the interaction affibody molecule denoted ZRAF322 was found to selectively bind to Raf-1 and inhibit the interaction between H-Ras and Raf-1 in vitro. In a continued effort, ribosome display was applied to the affinity between H-Ras and Raf-1 in vitro. In a continued effort, ribosome display was applied to the affinity maturation of the ZRAF322 variant in a novel approach, based on repetitive cycles of diversification by maturation of the ZRAF322 variant in a novel approach, based on repetitive cycles of diversification by error-prone PCR of the entire affibody gene and ribosome display selection, mimicking the principles error-prone PCR of the entire affibody gene and ribosome display selection, mimicking the principles of natural evolution. The method involved a monitoring of the progress of evolution and variants of of natural evolution. The method involved a monitoring of the progress of evolution and variants of

ZRAF322 with 13- to 26-fold improved affinities were obtained, that contained different combinations of ZRAF322 with 13- to 26-fold improved affinities were obtained, that contained different combinations of single or double amino acid substitutions in either previously randomized or framework positions. single or double amino acid substitutions in either previously randomized or framework positions. Implications of the substitutions for binder stability and selectivity were also investigated, showing Implications of the substitutions for binder stability and selectivity were also investigated, showing that a higher affinity could be associated with a lower thermal melting point and that affinity-improved that a higher affinity could be associated with a lower thermal melting point and that affinity-improved variants showed uncompromised binding selectivity to the hRaf-1 target. variants showed uncompromised binding selectivity to the hRaf-1 target.

Keywords: affibody, ribosome display, phage display, combinatorial protein engineering, Keywords: affibody, ribosome display, phage display, combinatorial protein engineering, library, murine IgG1, SATB1, Ras, Raf. library, murine IgG1, SATB1, Ras, Raf. © Sebastian Grimm 2011 © Sebastian Grimm 2011 IV IV

V V

List of publications List of publications

This thesis is based on the following four papers, which are referred to in the text by their This thesis is based on the following four papers, which are referred to in the text by their Roman numerals (I-IV) and can be found in the appendix. Roman numerals (I-IV) and can be found in the appendix.

I Grimm, S., Yu, F. and Nygren, P.-Å. (2011). Ribosome display selection of a murine I Grimm, S., Yu, F. and Nygren, P.-Å. (2011). Ribosome display selection of a murine IgG1 Fab binding affibody molecule allowing species selective recovery of IgG1 Fab binding affibody molecule allowing species selective recovery of monoclonal antibodies. Molecular Biotechnology, DOI: 10.1007/s12033-010-9367-1. monoclonal antibodies. Molecular Biotechnology, DOI: 10.1007/s12033-010-9367-1.

II Grimm, S.*, Klooster, R.*, Bisschop, I.J.M., Gruselius, J., Nygren, P.-Å.* and van II Grimm, S.*, Klooster, R.*, Bisschop, I.J.M., Gruselius, J., Nygren, P.-Å.* and van der Maarel, S.M.* (2011). Single domain affinity proteins for the detection of the der Maarel, S.M.* (2011). Single domain affinity proteins for the detection of the genome organizer protein SATB1. Manuscript. genome organizer protein SATB1. Manuscript.

III Grimm, S.*, Lundberg, E.*, Yu, F., Shibasaki, S., Vernet, E., Skogs, M., Nygren, P.- III Grimm, S.*, Lundberg, E.*, Yu, F., Shibasaki, S., Vernet, E., Skogs, M., Nygren, P.- Å., Gräslund, T. (2010). Selection and characterisation of affibody molecules Å., Gräslund, T. (2010). Selection and characterisation of affibody molecules inhibiting the interaction between Ras and Raf in vitro. New Biotechnology 27:766-73. inhibiting the interaction between Ras and Raf in vitro. New Biotechnology 27:766-73.

IV Grimm, S., Salahshour, S. and Nygren, P.-Å. (2011). Affinity maturation of an IV Grimm, S., Salahshour, S. and Nygren, P.-Å. (2011). Affinity maturation of an affibody molecule binding to human Raf-1 via non-targeted in vitro evolution. affibody molecule binding to human Raf-1 via non-targeted in vitro evolution. Manuscript. Manuscript.

*These authors contributed equally to this work. *These authors contributed equally to this work. All papers were reproduced with permission from the copyright holders. All papers were reproduced with permission from the copyright holders.

VI VI

VII VII

“Perfection is achieved, not when there is nothing more to add, “Perfection is achieved, not when there is nothing more to add, but when there is nothing left to take away.” but when there is nothing left to take away.”

Antoine de Saint-Exupéry Antoine de Saint-Exupéry

VIII VIII

IX IX

Contents Contents

1 Protein engineering ...... 1 1 Protein engineering ...... 1 1.1 Brief history...... 1 1.1 Brief history...... 1 1.2 Protein traits and how to analyze them...... 3 1.2 Protein traits and how to analyze them...... 3 1.2.1 Affinity and selectivity ...... 3 1.2.1 Affinity and selectivity ...... 3 1.2.2 Methods to analyze affinity and selectivity ...... 6 1.2.2 Methods to analyze affinity and selectivity ...... 6 1.2.3 Stability ...... 9 1.2.3 Stability ...... 9 1.2.4 Methods to analyze stability ...... 10 1.2.4 Methods to analyze stability ...... 10 1.2.5 Enzyme activity ...... 11 1.2.5 Enzyme activity ...... 11 1.2.6 Methods to analyze enzyme activity...... 11 1.2.6 Methods to analyze enzyme activity...... 11 1.2.7 BOX: recombinant DNA technology...... 13 1.2.7 BOX: recombinant DNA technology...... 13 1.3 Protein engineering by rational approaches ...... 14 1.3 Protein engineering by rational approaches ...... 14 1.4 Protein engineering by directed evolution...... 17 1.4 Protein engineering by directed evolution...... 17 1.4.1 Introduction...... 17 1.4.1 Introduction...... 17 1.4.2 Sources of diversity...... 19 1.4.2 Sources of diversity...... 19 2 Protein selection systems...... 23 2 Protein selection systems...... 23 2.1 Common properties ...... 23 2.1 Common properties ...... 23 2.2 Phage display...... 25 2.2 Phage display...... 25 2.3 Cell surface display ...... 30 2.3 Cell surface display ...... 30 2.4 Ribosome display ...... 32 2.4 Ribosome display ...... 32 2.5 mRNA display ...... 35 2.5 mRNA display ...... 35 2.6 Other display systems...... 36 2.6 Other display systems...... 36 2.7 Compartment systems ...... 37 2.7 Compartment systems ...... 37 3 Protein scaffolds for molecular recognition ...... 40 3 Protein scaffolds for molecular recognition ...... 40 3.1 Introduction ...... 40 3.1 Introduction ...... 40 3.2 Antibodies...... 40 3.2 Antibodies...... 40 3.2.1 Natural polyclonal or monoclonal antibodies ...... 40 3.2.1 Natural polyclonal or monoclonal antibodies ...... 40 3.2.2 Engineering of antibody effector functions ...... 42 3.2.2 Engineering of antibody effector functions ...... 42 3.2.3 Recombinant generation of antibody fragments and monoclonal antibodies...... 44 3.2.3 Recombinant generation of antibody fragments and monoclonal antibodies...... 44 3.2.4 Single domain antibody fragments ...... 45 3.2.4 Single domain antibody fragments ...... 45 3.3 Alternative protein scaffolds ...... 46 3.3 Alternative protein scaffolds ...... 46 3.3.1 Anticalins ...... 47 3.3.1 Anticalins ...... 47 X X

3.3.2 Affibody molecules ...... 48 3.3.2 Affibody molecules ...... 48 3.3.3 Adnectins...... 51 3.3.3 Adnectins...... 51 3.3.4 DARPins...... 52 3.3.4 DARPins...... 52 3.3.5 Other alternative protein scaffolds ...... 52 3.3.5 Other alternative protein scaffolds ...... 52 4 Present investigation...... 55 4 Present investigation...... 55 4.1 Ribosome display selection of a murine IgG1 Fab binding affibody molecule allowing 4.1 Ribosome display selection of a murine IgG1 Fab binding affibody molecule allowing species selective recovery of monoclonal antibodies (I)...... 56 species selective recovery of monoclonal antibodies (I)...... 56 4.2 Single domain affinity proteins for the detection of the genome organizer protein 4.2 Single domain affinity proteins for the detection of the genome organizer protein SATB1 (II)...... 62 SATB1 (II)...... 62 4.3 Selection and characterisation of affibody molecules inhibiting the interaction between 4.3 Selection and characterisation of affibody molecules inhibiting the interaction between Ras and Raf in vitro (III) ...... 68 Ras and Raf in vitro (III) ...... 68 4.4 Affinity maturation of an affibody molecule binding to human Raf-1 via non-targeted in 4.4 Affinity maturation of an affibody molecule binding to human Raf-1 via non-targeted in vitro evolution (IV)...... 72 vitro evolution (IV)...... 72 4.5 Conclusions and future perspectives...... 79 4.5 Conclusions and future perspectives...... 79 Acknowledgements ...... 80 Acknowledgements ...... 80 References...... 82 References...... 82

XI XI

Abbreviations Abbreviations aa Amino acid aa Amino acid ABD Albumin binding domain ABD Albumin binding domain ABP Albumin binding protein ABP Albumin binding protein ARM Affibody-ribosome-mRNA ARM Affibody-ribosome-mRNA CD Circular dichroism CD Circular dichroism CDR Complementarity determining region CDR Complementarity determining region DNA Deoxyribonucleic acid DNA Deoxyribonucleic acid ELISA Enzyme-linked immunosorbent assay ELISA Enzyme-linked immunosorbent assay Fab Fragment antigen binding (antibody) Fab Fragment antigen binding (antibody) FACS Fluorescence-activated cell sorting FACS Fluorescence-activated cell sorting FBS Fetal bovine serum FBS Fetal bovine serum Fc Fragment crystallizable (antibody) Fc Fragment crystallizable (antibody) GFP Green fluorescent protein GFP Green fluorescent protein GST Glutathione S-transferase GST Glutathione S-transferase HSA Human serum albumin HSA Human serum albumin Ig Immunoglobulin Ig Immunoglobulin IgG Immunoglobulin G IgG Immunoglobulin G

Kd Equilibrium dissociation constant Kd Equilibrium dissociation constant kDa Kilodalton kDa Kilodalton mAb Monoclonal antibody mAb Monoclonal antibody MAPK Mitogen-activated protein kinase MAPK Mitogen-activated protein kinase mRNA Messenger ribonucleic acid mRNA Messenger ribonucleic acid PCR Polymerase chain reaction PCR Polymerase chain reaction pKa Acid dissociation constant in logarithmic scale pKa Acid dissociation constant in logarithmic scale PrEST Protein epitope signature tag PrEST Protein epitope signature tag RBD Ras binding domain RBD Ras binding domain SATB1 Special AT-rich sequence binding protein 1 SATB1 Special AT-rich sequence binding protein 1 scFv Single-chain variable fragment (antibody) scFv Single-chain variable fragment (antibody) SPG Streptococcal protein G SPG Streptococcal protein G SPR Surface plasmon resonance SPR Surface plasmon resonance STED Stimulated Emission Depletion STED Stimulated Emission Depletion

Tm Melting temperature Tm Melting temperature VHH Variable domain of a camelid heavy chain antibody VHH Variable domain of a camelid heavy chain antibody XII XII

Sebastian Grimm 1 Sebastian Grimm 1

1 Protein engineering 1 Protein engineering 1.1 Brief history 1.1 Brief history

Proteins are polymers of amino acids, generally with a defined three-dimensional structure. Proteins are polymers of amino acids, generally with a defined three-dimensional structure. Together with nucleic acids, lipids and polysaccharides, proteins are the fundamental building Together with nucleic acids, lipids and polysaccharides, proteins are the fundamental building blocks of life on Earth and fulfill a plethora of functions such as transport of matter (e.g. blocks of life on Earth and fulfill a plethora of functions such as transport of matter (e.g. transferrin), catalysis of reactions (enzymes), defense towards pathogens (e.g. antibodies), transferrin), catalysis of reactions (enzymes), defense towards pathogens (e.g. antibodies), support of cell structure (e.g. collagen), generation of force (e.g. myosin), regulation of cell support of cell structure (e.g. collagen), generation of force (e.g. myosin), regulation of cell metabolism (hormones), regulation of water flow (aquaporins) or transport of electrons (e.g. metabolism (hormones), regulation of water flow (aquaporins) or transport of electrons (e.g. cytochrome), all of which are vital to the survival of higher vertebrates. Natural selection over cytochrome), all of which are vital to the survival of higher vertebrates. Natural selection over three billion years has adapted proteins to fulfill their function with great efficiency, as three billion years has adapted proteins to fulfill their function with great efficiency, as demonstrated by enzymes such as superoxide-dismutase or acetylcholine esterase that demonstrated by enzymes such as superoxide-dismutase or acetylcholine esterase that calalyze reactions with rates close to the diffusion limit 1, polymerases with error rates as low calalyze reactions with rates close to the diffusion limit 1, polymerases with error rates as low as one in one million bases 2 or spider silk with tensile strengths comparable to high-grade as one in one million bases 2 or spider silk with tensile strengths comparable to high-grade steel 3. steel 3. It took us humans until the year of 1839 to get a glimpse of the atomic composition of It took us humans until the year of 1839 to get a glimpse of the atomic composition of proteins, which Mulder referred to as “albuminous matter” 4, and not until 1902 did we know proteins, which Mulder referred to as “albuminous matter” 4, and not until 1902 did we know that amino acids in proteins are linked through repeating peptide bonds 5; 6. Since the mid 20th that amino acids in proteins are linked through repeating peptide bonds 5; 6. Since the mid 20th century, our knowledge on proteins has expanded rapidly, with the amino acid sequence of century, our knowledge on proteins has expanded rapidly, with the amino acid sequence of insulin determined in 1951 by Sanger and Tuppy 7 and the protein structures of myoglobin insulin determined in 1951 by Sanger and Tuppy 7 and the protein structures of myoglobin and oxy-myoglobin determined in 1957 by means of X-ray crystallography 8; 9. The and oxy-myoglobin determined in 1957 by means of X-ray crystallography 8; 9. The opportunity emerged to pinpoint key features of protein function to certain amino acid opportunity emerged to pinpoint key features of protein function to certain amino acid residues, and, based on this information, to manipulate proteins in a systematic and residues, and, based on this information, to manipulate proteins in a systematic and knowledge-based manner, in other words to engineer proteins. Protein engineering started knowledge-based manner, in other words to engineer proteins. Protein engineering started with rather humble and often non-selective approaches in the beginning of the late 1970s by with rather humble and often non-selective approaches in the beginning of the late 1970s by directing organic chemistry specifically at macromolecular surfaces, using mostly acylating or directing organic chemistry specifically at macromolecular surfaces, using mostly acylating or alkylating chemicals to modify nucleophilic amino acid side chains. The breakthrough came alkylating chemicals to modify nucleophilic amino acid side chains. The breakthrough came when recombinant DNA techniques such as DNA cleavage, ligation 10; 11; 12; 13 and site when recombinant DNA techniques such as DNA cleavage, ligation 10; 11; 12; 13 and site directed mutagenesis 14 became available in the early 1980s (please refer to section 1.2.7). directed mutagenesis 14 became available in the early 1980s (please refer to section 1.2.7). This toolbox allowed for the editing of specific positions in a gene, and the recombination of This toolbox allowed for the editing of specific positions in a gene, and the recombination of gene fragments, followed by expression of the recombinant protein in a microbial host. gene fragments, followed by expression of the recombinant protein in a microbial host. Winter, Fersht and coworkers were the first to take advantage of the pioneering Winter, Fersht and coworkers were the first to take advantage of the pioneering oligonucleotide-based site-directed mutagenesis protocols developed by Smith 14, when they oligonucleotide-based site-directed mutagenesis protocols developed by Smith 14, when they substituted single amino acid residues of the tyrosyl tRNA synthetase from Bacillus substituted single amino acid residues of the tyrosyl tRNA synthetase from Bacillus stearothermophilus and studied the effects on the catalytic rate and substrate affinity 15. The stearothermophilus and studied the effects on the catalytic rate and substrate affinity 15. The rapidly growing understanding of protein function paved the way for protein engineers to rapidly growing understanding of protein function paved the way for protein engineers to improve or alter existing protein traits such as affinity, selectivity or stability according to the improve or alter existing protein traits such as affinity, selectivity or stability according to the requirements of the intended applications. For example, the alkaline protease of Bacillus requirements of the intended applications. For example, the alkaline protease of Bacillus 2 Ribosome display for selection and evolution of affibody molecules 2 Ribosome display for selection and evolution of affibody molecules amyloliquefaciens (subtilisin) which today is commonly used as an ingredient in laundry or amyloliquefaciens (subtilisin) which today is commonly used as an ingredient in laundry or dishwashing detergents was early engineered by Wells and coworkers for increased stability dishwashing detergents was early engineered by Wells and coworkers for increased stability at extremes of pH, elevated temperatures or oxidizing conditions found in the presence of at extremes of pH, elevated temperatures or oxidizing conditions found in the presence of bleach 16. Perry and Wetzel engineered a disulphide bond into T4 lysozyme to stabilize the bleach 16. Perry and Wetzel engineered a disulphide bond into T4 lysozyme to stabilize the protein towards thermal inactivation 17. Further, Clarke and coworkers reported two modified protein towards thermal inactivation 17. Further, Clarke and coworkers reported two modified forms of the lactate dehydrogenase: one which is specific for a new substrate and one which forms of the lactate dehydrogenase: one which is specific for a new substrate and one which lacks allosteric regulation 18. If no structural information is available, amino acid residues that lacks allosteric regulation 18. If no structural information is available, amino acid residues that affect a certain protein trait can systematically be pin-pointed using alanine-scanning affect a certain protein trait can systematically be pin-pointed using alanine-scanning mutagenesis that was originally described by Cunningham and Wells in 1989 19. In their mutagenesis that was originally described by Cunningham and Wells in 1989 19. In their work, single alanine mutations within the human growth hormone (hGH) were introduced to work, single alanine mutations within the human growth hormone (hGH) were introduced to reveal amino acid side chains that promote binding to the hGH receptor. reveal amino acid side chains that promote binding to the hGH receptor. Rational protein engineering approaches however have their limitations. It is to date Rational protein engineering approaches however have their limitations. It is to date very challenging to predict the outcome of site-directed mutations on for example the affinity very challenging to predict the outcome of site-directed mutations on for example the affinity or stability of a protein. This applies particularly to binding proteins such as antibodies, were or stability of a protein. This applies particularly to binding proteins such as antibodies, were a number of amino acids need to be precisely positioned to form a functional binding site, and a number of amino acids need to be precisely positioned to form a functional binding site, and the differing orientation of a single amino acid side chain can result in a loss of binding 20. the differing orientation of a single amino acid side chain can result in a loss of binding 20. The example of hemoglobin demonstrates that subtle changes such as the binding of the small The example of hemoglobin demonstrates that subtle changes such as the binding of the small molecule oxygen can affect an entire protein structure 21. To overcome these limitations, the molecule oxygen can affect an entire protein structure 21. To overcome these limitations, the power of natural selection can be harnessed for the engineering of proteins in so-called power of natural selection can be harnessed for the engineering of proteins in so-called library-based approaches. library-based approaches. Since Darwin’s historical work "On the origin of the species" 22 we know that Since Darwin’s historical work "On the origin of the species" 22 we know that generation of diversity and selection are the driving forces of evolution, which applies not generation of diversity and selection are the driving forces of evolution, which applies not only to living species but also to molecular systems ex vivo. To mimic natural selection only to living species but also to molecular systems ex vivo. To mimic natural selection processes in a test-tube, a technique to generate protein diversity and a selection or screening processes in a test-tube, a technique to generate protein diversity and a selection or screening system to isolate variants showing an improved function under a given experimental condition system to isolate variants showing an improved function under a given experimental condition are needed. One way of generating diversity in nucleic acids became available with the are needed. One way of generating diversity in nucleic acids became available with the invention of polymerase chain reaction by Mullis and coworkers in 1987 23, that allows for the invention of polymerase chain reaction by Mullis and coworkers in 1987 23, that allows for the numeral exponential amplification of DNA fragments starting even from a single copy of a numeral exponential amplification of DNA fragments starting even from a single copy of a template molecule. When performed under certain conditions, diversity can be introduced by template molecule. When performed under certain conditions, diversity can be introduced by an intentional random incorporation of mutations during amplification cycles (error-prone an intentional random incorporation of mutations during amplification cycles (error-prone PCR). Alternatively, diversity at intended positions can be obtained through cassette PCR). Alternatively, diversity at intended positions can be obtained through cassette mutagenesis, using mixtures of oligonucleotides designed and synthesized to contain mutagenesis, using mixtures of oligonucleotides designed and synthesized to contain degenerate codons as building blocks for cassette assembly (see chapter 1.4.2). A powerful degenerate codons as building blocks for cassette assembly (see chapter 1.4.2). A powerful system for protein selection, linking genotype and phenotype, was described in 1985 when system for protein selection, linking genotype and phenotype, was described in 1985 when Smith for the first time displayed peptides on the f1 filamentous phage 24. Later, other Smith for the first time displayed peptides on the f1 filamentous phage 24. Later, other selection systems such as bacterial display 25, ribosome display 26 and mRNA display 27 were selection systems such as bacterial display 25, ribosome display 26 and mRNA display 27 were reported (see chapter 2). The combination of diverse protein libraries with selection systems reported (see chapter 2). The combination of diverse protein libraries with selection systems allowed scientists to cast larger nets. Hanes and coworkers selected antibodies with picomolar allowed scientists to cast larger nets. Hanes and coworkers selected antibodies with picomolar Sebastian Grimm 3 Sebastian Grimm 3 affinity from a large synthetic library displayed on ribosomes 28; 29 and Seelig and Szostak affinity from a large synthetic library displayed on ribosomes 28; 29 and Seelig and Szostak reported a novel enzyme selected from a library of more than 1012 protein variants based on a reported a novel enzyme selected from a library of more than 1012 protein variants based on a non-catalytic scaffold, using mRNA display 30. non-catalytic scaffold, using mRNA display 30. The ultimate challenge for a protein engineer is to create novel proteins with designed The ultimate challenge for a protein engineer is to create novel proteins with designed functions. A rational “maquette” approach was reported by Koder and Anderson. Based on functions. A rational “maquette” approach was reported by Koder and Anderson. Based on amino acid helix propensities, a four-helix bundle protein scaffold was designed to amino acid helix propensities, a four-helix bundle protein scaffold was designed to accommodate four haem groups, capable of oxygen binding and transport 31. Using an accommodate four haem groups, capable of oxygen binding and transport 31. Using an iteration between sequence design and structure prediction in silico, Kuhlman and Damtas iteration between sequence design and structure prediction in silico, Kuhlman and Damtas from the Baker lab reported the engineering of a novel globular protein fold with atomic from the Baker lab reported the engineering of a novel globular protein fold with atomic accuracy 32. In another study, Keefe and Szostak isolated a new ATP-binding protein from a accuracy 32. In another study, Keefe and Szostak isolated a new ATP-binding protein from a large library of 6x1012 random 11-mer peptides displayed on mRNA, a masterpiece large library of 6x1012 random 11-mer peptides displayed on mRNA, a masterpiece challenging the limits of combinatorial engineering and demonstrating that novel proteins challenging the limits of combinatorial engineering and demonstrating that novel proteins with designated functions can be selected if the library is big enough 33. with designated functions can be selected if the library is big enough 33.

1.2 Protein traits and how to analyze them 1.2 Protein traits and how to analyze them

There are generally distinct traits that describe the properties of a protein and can be There are generally distinct traits that describe the properties of a protein and can be addressed by a protein engineer. Some of these traits describe intrinsic properties, such as addressed by a protein engineer. Some of these traits describe intrinsic properties, such as stability. Others describe binding characteristics, such as affinity and selectivity, mechanistic stability. Others describe binding characteristics, such as affinity and selectivity, mechanistic properties, such as enzyme activity, or a protein’s behavior in a biological system, such as properties, such as enzyme activity, or a protein’s behavior in a biological system, such as immunogenicity and half-life. In this chapter, I will focus on the protein traits affinity, immunogenicity and half-life. In this chapter, I will focus on the protein traits affinity, selectivity, stability and enzyme activity and present a selection of methods to analyze them. selectivity, stability and enzyme activity and present a selection of methods to analyze them.

1.2.1 Affinity and selectivity 1.2.1 Affinity and selectivity

Affinity describes the strength of an interaction between two molecules 34. Under Affinity describes the strength of an interaction between two molecules 34. Under standardized temperature and pressure conditions, the affinity between a protein called A and standardized temperature and pressure conditions, the affinity between a protein called A and its ligand called B is quantified by the equilibrium association constant: its ligand called B is quantified by the equilibrium association constant:

Ka = [AB]/([A][B]) (1) Ka = [AB]/([A][B]) (1)

Here, [A] and [B] are the concentrations of free protein and ligand and [AB] is the Here, [A] and [B] are the concentrations of free protein and ligand and [AB] is the concentration of the complex formed between protein and ligand at equilibrium. However, in concentration of the complex formed between protein and ligand at equilibrium. However, in protein research the equilibrium dissociation constant is often used instead: protein research the equilibrium dissociation constant is often used instead:

Kd = 1/Ka (2) Kd = 1/Ka (2)

4 Ribosome display for selection and evolution of affibody molecules 4 Ribosome display for selection and evolution of affibody molecules

There is a relation between the equilibrium and the kinetics of binding reactions and Kd can There is a relation between the equilibrium and the kinetics of binding reactions and Kd can also be expressed as: also be expressed as:

Kd = koff/kon (3) Kd = koff/kon (3)

-1 -1 -1 -1 Here, kon is the association rate constant [M s ] that determines how fast two molecules bind Here, kon is the association rate constant [M s ] that determines how fast two molecules bind -1 -1 to each other and koff is the dissociation rate constant [s ] that determines how fast the to each other and koff is the dissociation rate constant [s ] that determines how fast the complex dissociates. The binding of two molecules is driven by thermodynamics as complex dissociates. The binding of two molecules is driven by thermodynamics as quantified in the Gibbs-Helmholtz free energy of binding: quantified in the Gibbs-Helmholtz free energy of binding:

ΔG = -RTln(Ka) (4) ΔG = -RTln(Ka) (4)

Here, R is the gas constant and T is the temperature in Kelvin. The free energy of binding has Here, R is the gas constant and T is the temperature in Kelvin. The free energy of binding has both enthalpic (H) and entropic (S) contributors and can under constant temperature and both enthalpic (H) and entropic (S) contributors and can under constant temperature and pressure conditions be written as: pressure conditions be written as:

ΔG = ΔH-TΔS (5) ΔG = ΔH-TΔS (5)

The enthalpic contributors are short or long-ranging electrostatic forces between pairs of The enthalpic contributors are short or long-ranging electrostatic forces between pairs of opposite charges, hydrogen bonds between partial charges within a distance of about 2 Å, and opposite charges, hydrogen bonds between partial charges within a distance of about 2 Å, and very short-range van der Waals interactions between fluctuating electron polarizations in very short-range van der Waals interactions between fluctuating electron polarizations in nearby atoms. The main entropic contributors are ordered water molecules on the protein or nearby atoms. The main entropic contributors are ordered water molecules on the protein or ligand surface that are released to the bulk solvent upon ligand binding, and the decrease of ligand surface that are released to the bulk solvent upon ligand binding, and the decrease of peptide chain entropy upon ligand binding. The latter is less important in rigid protein peptide chain entropy upon ligand binding. The latter is less important in rigid protein scaffolds and ligands with low translational and rotational degrees of freedom 35. scaffolds and ligands with low translational and rotational degrees of freedom 35. Proteins that were generated with the purpose to bind to other molecules (affinity Proteins that were generated with the purpose to bind to other molecules (affinity proteins) of different sources and kind are used in a variety of applications, both in vitro such proteins) of different sources and kind are used in a variety of applications, both in vitro such as chromatography and detection assays or in vivo such as molecular imaging and therapy. as chromatography and detection assays or in vivo such as molecular imaging and therapy. The different applications have different affinity requirements. An equilibrium dissociation The different applications have different affinity requirements. An equilibrium dissociation constant as high as 10-3 M was shown to be sufficient to achieve a dynamic equilibrium constant as high as 10-3 M was shown to be sufficient to achieve a dynamic equilibrium between a protein immobilized on an affinity resin and a ligand in solution, allowing for between a protein immobilized on an affinity resin and a ligand in solution, allowing for separation of the ligand from non-interacting solutes 36. Other applications such as molecular separation of the ligand from non-interacting solutes 36. Other applications such as molecular tumor imaging may require very high affinities (pM range) to obtain high tumor/blood ratios tumor imaging may require very high affinities (pM range) to obtain high tumor/blood ratios as shown in a mouse model for Her2 imaging with affibody molecules of 50 nM and 22 pM as shown in a mouse model for Her2 imaging with affibody molecules of 50 nM and 22 pM affinities 37. Related to therapy, a study with full-length antibodies mediating antibody- affinities 37. Related to therapy, a study with full-length antibodies mediating antibody- dependent cellular cytotoxicity (ADCC) on cultured tumor cells has shown that an antibody dependent cellular cytotoxicity (ADCC) on cultured tumor cells has shown that an antibody with a 5x10-10 M affinity for the tumor-antigen EpCAM could mediate ADCC more with a 5x10-10 M affinity for the tumor-antigen EpCAM could mediate ADCC more efficiently than an antibody with a 2x10-8 M affinity targeting the same epitope on EpCAM 38. efficiently than an antibody with a 2x10-8 M affinity targeting the same epitope on EpCAM 38. Sebastian Grimm 5 Sebastian Grimm 5

However, in a mouse model for the treatment solid tumors with scFv antibody fragments, However, in a mouse model for the treatment solid tumors with scFv antibody fragments, moderate affinities in the 10-7 M range gave better tumor penetration than affinities of 10-9 M moderate affinities in the 10-7 M range gave better tumor penetration than affinities of 10-9 M or higher 39. In case of the T-cell engager MT103, a bispecific agent binding to CD3 or higher 39. In case of the T-cell engager MT103, a bispecific agent binding to CD3 expressed on cytotoxic T cells and CD19 on Non-Hodgkin's lymphoma, a moderate affinity to expressed on cytotoxic T cells and CD19 on Non-Hodgkin's lymphoma, a moderate affinity to CD3 allows T cells to dissociate in due time and mediate killing of multiple tumor cells CD3 allows T cells to dissociate in due time and mediate killing of multiple tumor cells (Andreas Wolf, personal communication). As illustrated by the latter examples, the impact of (Andreas Wolf, personal communication). As illustrated by the latter examples, the impact of affinity in therapy settings is still controversial and dependent on a multitude of factors, affinity in therapy settings is still controversial and dependent on a multitude of factors, among others pharmacokinetics and immune effector functions. For the identification of among others pharmacokinetics and immune effector functions. For the identification of protein interactions in vitro, the situation is less complex. From the law of mass action protein interactions in vitro, the situation is less complex. From the law of mass action follows: the higher the affinity between A and B, the higher will be the concentration of follows: the higher the affinity between A and B, the higher will be the concentration of complex AB at given concentrations of A and B. Hence, assuming that A is an immobilized complex AB at given concentrations of A and B. Hence, assuming that A is an immobilized antibody in an immunoassay for detection of a low abundant marker protein B in an analyte antibody in an immunoassay for detection of a low abundant marker protein B in an analyte sample, then a high affinity A will result in a more sensitive detection of B. Also from the law sample, then a high affinity A will result in a more sensitive detection of B. Also from the law of mass action follows that, if marker protein B is present at high concentrations, a weak of mass action follows that, if marker protein B is present at high concentrations, a weak antibody affinity may be preferable to get a broader dynamic range between different patient antibody affinity may be preferable to get a broader dynamic range between different patient samples. samples.

For high affinities, both a very fast kon and a very slow koff are required. The For high affinities, both a very fast kon and a very slow koff are required. The association rate constant is in practice limited by diffusion and rotational alignment of the association rate constant is in practice limited by diffusion and rotational alignment of the binding partners to about 2x106 M-1s-1 40 and can be engineered by manipulating long-ranging binding partners to about 2x106 M-1s-1 40 and can be engineered by manipulating long-ranging electrostatic steering forces that help molecules to “find” their dedicated binding sites 41. The electrostatic steering forces that help molecules to “find” their dedicated binding sites 41. The importance of such electrostatic interactions for the association of two proteins was importance of such electrostatic interactions for the association of two proteins was experimentally verified by Schreiber and Fersht, working with mutants of proteins denoted experimentally verified by Schreiber and Fersht, working with mutants of proteins denoted barnase and barstar 42. Furthermore, a stable and well-folded protein structure is crucial for a barnase and barstar 42. Furthermore, a stable and well-folded protein structure is crucial for a fast association, since any structural re-arrangement prior to binding will delay the association fast association, since any structural re-arrangement prior to binding will delay the association and give rise to complexes with relatively low affinity 43. To my knowledge, only few protein and give rise to complexes with relatively low affinity 43. To my knowledge, only few protein 44 44 engineering approaches were reported that selectively and intentionally address kon . The engineering approaches were reported that selectively and intentionally address kon . The dissociation rate constant koff is not diffusion-limited and can more readily be addressed in dissociation rate constant koff is not diffusion-limited and can more readily be addressed in selections, for example using a molar excess of competitor 45. Zahnd and coworkers selections, for example using a molar excess of competitor 45. Zahnd and coworkers calculated that a maximum ratio of competitor antigen to selectable antigen and a selection calculated that a maximum ratio of competitor antigen to selectable antigen and a selection time corresponding to the reciprocal of the off-rate constant that one is intending to find in the time corresponding to the reciprocal of the off-rate constant that one is intending to find in the -3 -1 46 -3 -1 46 library (e.g. 1000 s for koff = 10 s ) are suitable parameters for selecting slow off-rates . In library (e.g. 1000 s for koff = 10 s ) are suitable parameters for selecting slow off-rates . In practice, the affinity of a peptide-binding single-chain variable fragment (scFv) to its ligand practice, the affinity of a peptide-binding single-chain variable fragment (scFv) to its ligand was improved 500-fold, employing off-rate selection with a 1000-fold molar excess of was improved 500-fold, employing off-rate selection with a 1000-fold molar excess of competitor 47. competitor 47.

Selectivity is a measure of the ability of a molecule to discriminate between different Selectivity is a measure of the ability of a molecule to discriminate between different substrates (in case of enzymes) or ligands (in case of binding proteins), a crucial feature of substrates (in case of enzymes) or ligands (in case of binding proteins), a crucial feature of any biological system. According to the International Union of Pure and Applied Chemistry any biological system. According to the International Union of Pure and Applied Chemistry (IUPAC), the ultimate degree of selectivity is defined as specificity. Importantly, in the case (IUPAC), the ultimate degree of selectivity is defined as specificity. Importantly, in the case 6 Ribosome display for selection and evolution of affibody molecules 6 Ribosome display for selection and evolution of affibody molecules of binding proteins, the observed degree of selectivity is dependent on both affinity and ligand of binding proteins, the observed degree of selectivity is dependent on both affinity and ligand concentrations. Assuming a case with a binding protein showing a 1000-fold higher affinity to concentrations. Assuming a case with a binding protein showing a 1000-fold higher affinity to its "true" ligand A as compared to a "false" ligand B, an experimental condition involving a its "true" ligand A as compared to a "false" ligand B, an experimental condition involving a 1000-fold higher concentration of ligand B would compensate for its lower affinity, leading to 1000-fold higher concentration of ligand B would compensate for its lower affinity, leading to an observed low selectivity. The same applies to an enzyme with different apparent an observed low selectivity. The same applies to an enzyme with different apparent equilibrium dissociation constants KM and different catalytic rates kcat for different substrates equilibrium dissociation constants KM and different catalytic rates kcat for different substrates (see section 1.2.5). A good example of the importance of selectivity are aminoacyl t-RNA (see section 1.2.5). A good example of the importance of selectivity are aminoacyl t-RNA synthetases, responsible for loading tRNA with the correct amino acid for protein synthesis. synthetases, responsible for loading tRNA with the correct amino acid for protein synthesis. The alanyl-tRNA synthetase discriminates alanine against glycine by a factor of 250, The alanyl-tRNA synthetase discriminates alanine against glycine by a factor of 250, supporting that an enzyme’s active site can recognize the absence of a single methylene group supporting that an enzyme’s active site can recognize the absence of a single methylene group on its substrate 48. Other proteins however, such as Staphylococcus protein A or Streptococcus on its substrate 48. Other proteins however, such as Staphylococcus protein A or Streptococcus protein G, were evolutionary adapted to bind to groups of different proteins, hence having a protein G, were evolutionary adapted to bind to groups of different proteins, hence having a much broader selectivity. There are several good examples of the engineering of protein much broader selectivity. There are several good examples of the engineering of protein selectivity. Based on structural information, an enzymes substrate binding pocket could be selectivity. Based on structural information, an enzymes substrate binding pocket could be modified to accommodate a different substrate, as demonstrated as early as 1988 by Wilks modified to accommodate a different substrate, as demonstrated as early as 1988 by Wilks and coworkers 49. The affinity and selectivity of an antibody fragment for testosterone could and coworkers 49. The affinity and selectivity of an antibody fragment for testosterone could be improved, employing a negative selection step against closely related steroids during be improved, employing a negative selection step against closely related steroids during phage display selection (for phage display, please refer to chapter 2.2). A study of Boström phage display selection (for phage display, please refer to chapter 2.2). A study of Boström and coworkers illustrates the possibility to engineer a dual selectivity (Her2 and VEGF) into and coworkers illustrates the possibility to engineer a dual selectivity (Her2 and VEGF) into the binding site of an antibody and shows that, using combinatorial approaches, one binding the binding site of an antibody and shows that, using combinatorial approaches, one binding site can be engineered for high affinity to two different ligands 50. site can be engineered for high affinity to two different ligands 50.

1.2.2 Methods to analyze affinity and selectivity 1.2.2 Methods to analyze affinity and selectivity

The affinity between a protein and different ligands is an important parameter to describe the The affinity between a protein and different ligands is an important parameter to describe the protein’s natural function in basic research or to determine its suitability for different protein’s natural function in basic research or to determine its suitability for different applications in biotechnology. The observed equilibrium dissociation constants of applications in biotechnology. The observed equilibrium dissociation constants of biologically relevant protein interactions fall into an extremely wide range of about 10-4 to 10- biologically relevant protein interactions fall into an extremely wide range of about 10-4 to 10- 16 M 51. 16 M 51. Classically, the equilibrium dissociation constant between a protein A and its ligand B Classically, the equilibrium dissociation constant between a protein A and its ligand B is determined by titration of a fixed concentration of A with various concentrations of B, and is determined by titration of a fixed concentration of A with various concentrations of B, and recording a signal that correlates with the concentration of the bound complex AB at recording a signal that correlates with the concentration of the bound complex AB at equilibrium. Importantly, ambient analyte conditions should be fulfilled, where the formation equilibrium. Importantly, ambient analyte conditions should be fulfilled, where the formation of AB does not significantly deplete the initial concentration of ligand in solution 52. For of AB does not significantly deplete the initial concentration of ligand in solution 52. For titration, concentrations of B should be chosen, that range from well above to well below the titration, concentrations of B should be chosen, that range from well above to well below the equilibrium dissociation constant. The signal that correlates with the concentration of AB is equilibrium dissociation constant. The signal that correlates with the concentration of AB is then modeled against the concentration of B and the obtained binding isotherm is fitted to then modeled against the concentration of B and the obtained binding isotherm is fitted to determine Kd. This approach is frequently used in e.g. equilibrium dialysis, isothermal determine Kd. This approach is frequently used in e.g. equilibrium dialysis, isothermal Sebastian Grimm 7 Sebastian Grimm 7 titration calorimetry (ITC) 53, enzyme-linked immunosorbent assays (ELISA) 54 or cell titration calorimetry (ITC) 53, enzyme-linked immunosorbent assays (ELISA) 54 or cell surface display and fluorescence-activated cell sorting (FACS) 55. A general limitation of surface display and fluorescence-activated cell sorting (FACS) 55. A general limitation of these methods is, however, that the kinetic parameters kon and koff are not accessible and that these methods is, however, that the kinetic parameters kon and koff are not accessible and that it may, particularly for high affinity interactions, take a long time to reach equilibrium. it may, particularly for high affinity interactions, take a long time to reach equilibrium. Conducting the assay before equilibrium is reached may result in erroneous numbers Conducting the assay before equilibrium is reached may result in erroneous numbers (Andersson and colleagues, unpublished data). When using ELISA, affinity complexes need (Andersson and colleagues, unpublished data). When using ELISA, affinity complexes need to withstand the washing, and labeling of molecules may affect their binding properties 51; 56. to withstand the washing, and labeling of molecules may affect their binding properties 51; 56.

A revolution in the field took place in 1990 with the commercial launch of a system A revolution in the field took place in 1990 with the commercial launch of a system that implemented an optical interface phenomenon denoted surface plasmon resonance (SPR) that implemented an optical interface phenomenon denoted surface plasmon resonance (SPR) for the measurement of biomolecular interactions in real time, i.e. allowing the measurement for the measurement of biomolecular interactions in real time, i.e. allowing the measurement of the rate constants kon and koff to determine Kd without the need to reach equilibrium. SPR of the rate constants kon and koff to determine Kd without the need to reach equilibrium. SPR was applied to the characterization of protein interactions in all four studies in this thesis and I was applied to the characterization of protein interactions in all four studies in this thesis and I will therefore explain its physical fundamentals in detail. will therefore explain its physical fundamentals in detail.

Surface plasmons are collectively oscillating electrons at the boundary of metal, such Surface plasmons are collectively oscillating electrons at the boundary of metal, such as gold, and dielectric. The electron oscillation has a certain momentum and propagates along as gold, and dielectric. The electron oscillation has a certain momentum and propagates along the boundary. Furthermore, it is sensitive towards any changes within the boundary, such as the boundary. Furthermore, it is sensitive towards any changes within the boundary, such as binding of proteins. A surface plasmon can be excited with a light beam that is parallely binding of proteins. A surface plasmon can be excited with a light beam that is parallely polarized to the plane of incidence and hits the metal layer through a glass prism. To achieve polarized to the plane of incidence and hits the metal layer through a glass prism. To achieve resonance or energy transfer to the surface plasmon, the momentum component of the resonance or energy transfer to the surface plasmon, the momentum component of the incident light that is in direction of the boundary needs to equal the momentum of the surface incident light that is in direction of the boundary needs to equal the momentum of the surface plasmon. When light hits the boundary from different angles, one of these angles will produce plasmon. When light hits the boundary from different angles, one of these angles will produce a momentum component that equals the surface plasmons momentum, and light will then be a momentum component that equals the surface plasmons momentum, and light will then be absorbed and the angle of a shadow can be detected. The protein interaction comes into play absorbed and the angle of a shadow can be detected. The protein interaction comes into play when a protein A is immobilized on a functionalized boundary and binds to its ligand B when a protein A is immobilized on a functionalized boundary and binds to its ligand B which is in a buffer flowed over the boundary. The surface plasmon will “sense” the binding, which is in a buffer flowed over the boundary. The surface plasmon will “sense” the binding, its momentum will change and thus the angle of the absorbed incident light. The angle of the its momentum will change and thus the angle of the absorbed incident light. The angle of the absorbed light (response signal) is plotted against time to yield a “sensorgram”, and a series of absorbed light (response signal) is plotted against time to yield a “sensorgram”, and a series of such sensorgrams for different ligand concentrations can be fitted with different kinetic such sensorgrams for different ligand concentrations can be fitted with different kinetic models to obtain kon, koff and Kd (Figure 1). Besides the measurement of kinetic parameters, a models to obtain kon, koff and Kd (Figure 1). Besides the measurement of kinetic parameters, a main advantage of this technique is the label-free detection of the ligand (B). A typical SPR main advantage of this technique is the label-free detection of the ligand (B). A typical SPR binding cycle starts with an association phase, where ligand (B) is flowed over immobilized binding cycle starts with an association phase, where ligand (B) is flowed over immobilized protein (A) and the response signal increases, if desired until equilibrium is reached. This is protein (A) and the response signal increases, if desired until equilibrium is reached. This is followed by a dissociation phase where ligand-free buffer is flowed over the boundary and followed by a dissociation phase where ligand-free buffer is flowed over the boundary and ligand (B) is allowed to dissociate, the response signal decreases. Prior to the next binding ligand (B) is allowed to dissociate, the response signal decreases. Prior to the next binding cycle, a pulse of regeneration solution of e.g. low pH or high salt concentration ensures that cycle, a pulse of regeneration solution of e.g. low pH or high salt concentration ensures that the binding of any remaining ligand (B) is disrupted (Figure 1) 57. the binding of any remaining ligand (B) is disrupted (Figure 1) 57. 8 Ribosome display for selection and evolution of affibody molecules 8 Ribosome display for selection and evolution of affibody molecules

Figure 1: SPR biosensor principle and typical SPR sensorgram. (a) Buffer solution containing Figure 1: SPR biosensor principle and typical SPR sensorgram. (a) Buffer solution containing ligand (red) is flowed over a gold surface with an affinity protein (green) immobilized. Light hits the ligand (red) is flowed over a gold surface with an affinity protein (green) immobilized. Light hits the gold surface from different angles (θ) simultaneously (yellow sectors). At the gold-buffer boundary, a gold surface from different angles (θ) simultaneously (yellow sectors). At the gold-buffer boundary, a surface plasmon with the momentum kpl is excited by the incident light and propagates along the surface plasmon with the momentum kpl is excited by the incident light and propagates along the boundary. For excitation to occur, the momentum of the incident light in direction of the boundary (kx) boundary. For excitation to occur, the momentum of the incident light in direction of the boundary (kx) has to equal the plasmons momentum kpl, which is fulfilled for kx1 before ligand binding and kx2 after has to equal the plasmons momentum kpl, which is fulfilled for kx1 before ligand binding and kx2 after ligand binding. A shadow is detected at angles θ1 or θ2, respectively. (b) The variable angle θ is ligand binding. A shadow is detected at angles θ1 or θ2, respectively. (b) The variable angle θ is plotted against time in an SPR sensorgram. The resulting changes in θ of the association phase (1), plotted against time in an SPR sensorgram. The resulting changes in θ of the association phase (1), dissociation phase (2) and re-generation (3) are denoted. dissociation phase (2) and re-generation (3) are denoted.

A general concern in SPR and other techniques to determine affinity are avidity A general concern in SPR and other techniques to determine affinity are avidity effects resulting from multiple binding sites of ligand B that can simultaneously interact with effects resulting from multiple binding sites of ligand B that can simultaneously interact with protein A molecules immobilized in close proximity. A recent technical realization requiring protein A molecules immobilized in close proximity. A recent technical realization requiring only very sparse immobilization levels with long inter-molecule distances can circumvent only very sparse immobilization levels with long inter-molecule distances can circumvent such avidity effects 58. Other instrumentations not based on the SPR phenomenon, such as the such avidity effects 58. Other instrumentations not based on the SPR phenomenon, such as the quartz crystal microbalance are commercially available 59, and the application of cheap silicon quartz crystal microbalance are commercially available 59, and the application of cheap silicon nanowires for biomolecule sensing 60 from mass production can be envisioned. nanowires for biomolecule sensing 60 from mass production can be envisioned. The ability of a binding protein to discriminate between different ligands, i.e. its The ability of a binding protein to discriminate between different ligands, i.e. its selectivity, can be measured by challenging the protein with as many different ligands as selectivity, can be measured by challenging the protein with as many different ligands as possible, either together in a complex mixture or spatially separated on an array. Such possible, either together in a complex mixture or spatially separated on an array. Such complex ligand mixtures may be cell lysates, blood serum and plasma or even whole cells or complex ligand mixtures may be cell lysates, blood serum and plasma or even whole cells or tissue slices that can be probed in various affinity applications such as Western blotting, tissue slices that can be probed in various affinity applications such as Western blotting, immunofluorescence, immunoprecipitation or affinity chromatography for a selective immunofluorescence, immunoprecipitation or affinity chromatography for a selective detection or capture of the ligand. Protein arrays however offer the advantage of a parallel detection or capture of the ligand. Protein arrays however offer the advantage of a parallel identification of many ligands to assess cross-reactivity, and the possibility to compare and identification of many ligands to assess cross-reactivity, and the possibility to compare and quantify the selectivity profiles as signal/background ratios 61. An attractive alternative to quantify the selectivity profiles as signal/background ratios 61. An attractive alternative to ”classical” planar arrays are suspensions of spectrally distinguishable beads. Each set can be ”classical” planar arrays are suspensions of spectrally distinguishable beads. Each set can be coupled with a different protein and multiple sets of beads can be combined to create an array coupled with a different protein and multiple sets of beads can be combined to create an array in suspension. Such bead arrays are then incubated with the ligand of interest and analyzed on in suspension. Such bead arrays are then incubated with the ligand of interest and analyzed on Sebastian Grimm 9 Sebastian Grimm 9 a flow cytometer that can assign binding signals to the different bead identities. A possible a flow cytometer that can assign binding signals to the different bead identities. A possible advantage to planar arrays is also the flexibility of excluding or including certain beads, and advantage to planar arrays is also the flexibility of excluding or including certain beads, and that the binding reactions take place in an environment that allows for representation of native that the binding reactions take place in an environment that allows for representation of native protein structures. protein structures.

1.2.3 Stability 1.2.3 Stability

Folded proteins can be denatured to a more or less unfolded state by changing their physical Folded proteins can be denatured to a more or less unfolded state by changing their physical or chemical environments, such as temperature, solvent or pH. Denaturation is for some or chemical environments, such as temperature, solvent or pH. Denaturation is for some proteins regarded as a reversible process between two states and can be treated as a chemical proteins regarded as a reversible process between two states and can be treated as a chemical equilibrium [N] ⇔ [D], where [N] is the protein concentration of the native state and [D] is equilibrium [N] ⇔ [D], where [N] is the protein concentration of the native state and [D] is the protein concentration of the denatured state. The thermodynamic stability of a protein is the protein concentration of the denatured state. The thermodynamic stability of a protein is given by the free energy of unfolding 62: given by the free energy of unfolding 62:

ΔGD-N = ΔHD-N - TΔSD-N (6) ΔGD-N = ΔHD-N - TΔSD-N (6)

Protein stability towards temperature-mediated denaturation (thermal stability) is for such Protein stability towards temperature-mediated denaturation (thermal stability) is for such two-state proteins characterized by the melting temperature (Tm), at which half of the protein two-state proteins characterized by the melting temperature (Tm), at which half of the protein molecules are in their native state and the other half in the denatured state. Tm can molecules are in their native state and the other half in the denatured state. Tm can experimentally be determined using differential scanning calorimetry or circular dichroism experimentally be determined using differential scanning calorimetry or circular dichroism (see next section). In general, high thermal stabilities are desirable to minimize the fraction of (see next section). In general, high thermal stabilities are desirable to minimize the fraction of denatured protein at the temperature of production or application, which otherwise could lead denatured protein at the temperature of production or application, which otherwise could lead to lower specific activities or susceptibility to aggregation or proteolytic degradation 63; 64. to lower specific activities or susceptibility to aggregation or proteolytic degradation 63; 64. This particularly applies to DNA polymerases for PCR reactions or enzymes in washing This particularly applies to DNA polymerases for PCR reactions or enzymes in washing powder. Thermal stability can be attributed to protein structural characteristics and non- powder. Thermal stability can be attributed to protein structural characteristics and non- covalent interactions between amino acid residues. Comparison of structures and sequences covalent interactions between amino acid residues. Comparison of structures and sequences from hyperthermophile, thermophile and mesophile microorganisms has resulted in the from hyperthermophile, thermophile and mesophile microorganisms has resulted in the identification of features such as an increased number of salt bridges, better hydrogen bonding identification of features such as an increased number of salt bridges, better hydrogen bonding and a better inner (hydrophobic core) packing of proteins from microorganisms living at and a better inner (hydrophobic core) packing of proteins from microorganisms living at elevated temperatures 65; 66. elevated temperatures 65; 66. Solvent denaturation can be achieved with chaotropic agents such as urea or Solvent denaturation can be achieved with chaotropic agents such as urea or guanidinium chloride at high concentrations and may result in complete unfolding of the guanidinium chloride at high concentrations and may result in complete unfolding of the protein. A denaturant can either directly stabilize unfolded groups via interaction or by protein. A denaturant can either directly stabilize unfolded groups via interaction or by altering the properties of the solvent 67. An approach to use protein library technology to altering the properties of the solvent 67. An approach to use protein library technology to functionally select for proteins with improved thermodynamic stability under elevated functionally select for proteins with improved thermodynamic stability under elevated temperature or denaturant concentration was described by Jung and colleagues 68. A scFv temperature or denaturant concentration was described by Jung and colleagues 68. A scFv library was displayed on phage and repetitive selection cycles were performed at temperatures library was displayed on phage and repetitive selection cycles were performed at temperatures 10 Ribosome display for selection and evolution of affibody molecules 10 Ribosome display for selection and evolution of affibody molecules of up to 60° C or GdnHCl concentrations of up to 3.5 M to select for more stable fluorescein of up to 60° C or GdnHCl concentrations of up to 3.5 M to select for more stable fluorescein binders. binders. The primary reasons for acid- or base-pH-induced denaturation are protonatable, salt- The primary reasons for acid- or base-pH-induced denaturation are protonatable, salt- bridge forming groups buried in the inside of the protein that are lost upon protonation or de- bridge forming groups buried in the inside of the protein that are lost upon protonation or de- protonation, as well as a highly positive or highly negative overall charge of the protein at low protonation, as well as a highly positive or highly negative overall charge of the protein at low or high pH, causing electrostatic repulsions. Tollinger and colleagues developed an NMR or high pH, causing electrostatic repulsions. Tollinger and colleagues developed an NMR spectroscopic method to determine pKa values of titratable groups in the folded and unfolded spectroscopic method to determine pKa values of titratable groups in the folded and unfolded state of the Src homology 3 domain as a model protein. The authors suggest that the changes state of the Src homology 3 domain as a model protein. The authors suggest that the changes of pKa values between the two states are responsible for the pH dependence of the protein’s of pKa values between the two states are responsible for the pH dependence of the protein’s thermodynamic stability 69. thermodynamic stability 69.

1.2.4 Methods to analyze stability 1.2.4 Methods to analyze stability

One frequently used technique for assessing protein stability is differential scanning One frequently used technique for assessing protein stability is differential scanning calorimetry (DSC), which measures the heat absorbed by a protein in solution while it is calorimetry (DSC), which measures the heat absorbed by a protein in solution while it is slowly heated through its melting transition. The melting transition or denaturation of a slowly heated through its melting transition. The melting transition or denaturation of a protein is highly endothermic and there is a large uptake of heat with a maximum at its Tm. protein is highly endothermic and there is a large uptake of heat with a maximum at its Tm. Practically, a buffer solution containing a defined concentration of protein and a reference Practically, a buffer solution containing a defined concentration of protein and a reference sample of buffer solution are heated electrically 70 and a trace of heat flux [µW] against sample of buffer solution are heated electrically 70 and a trace of heat flux [µW] against temperature or time is recorded. The temperature at the maximum of the peak is Tm, where temperature or time is recorded. The temperature at the maximum of the peak is Tm, where half of the proteins are unfolded, and the area of the peak after subtraction of the reference half of the proteins are unfolded, and the area of the peak after subtraction of the reference trace corresponds to the enthalpy of unfolding ΔHD-N. The entropy and hence free energy of trace corresponds to the enthalpy of unfolding ΔHD-N. The entropy and hence free energy of unfolding can be obtained from the area under the curve of Cp/T plotted against T, where Cp is unfolding can be obtained from the area under the curve of Cp/T plotted against T, where Cp is the heat capacity. the heat capacity. Another useful technique to assess protein stability and protein secondary structure Another useful technique to assess protein stability and protein secondary structure composition and content is circular dichroism (CD). Here, left and right circularly polarized composition and content is circular dichroism (CD). Here, left and right circularly polarized light are simultaneously passed through a protein-containing buffer solution. Due to the light are simultaneously passed through a protein-containing buffer solution. Due to the different refractive indexes and absorbances of the opposing circularly polarized light, an different refractive indexes and absorbances of the opposing circularly polarized light, an elliptic polarization hits the detector and an ellipticity can be determined. The wavelength of elliptic polarization hits the detector and an ellipticity can be determined. The wavelength of the incident light is variable, and spectra ranging from 160 to 250 nm are commonly recorded the incident light is variable, and spectra ranging from 160 to 250 nm are commonly recorded for protein secondary structure determination. Proteins with a high content of α-helices show for protein secondary structure determination. Proteins with a high content of α-helices show an ellipticity maximum at 192 nm, followed by a global minimum at 207 nm and a local an ellipticity maximum at 192 nm, followed by a global minimum at 207 nm and a local minimum at 222 nm. To determine the melting temperature Tm, a single characteristic minimum at 222 nm. To determine the melting temperature Tm, a single characteristic wavelength such as 222 nm is chosen and the ellipticity is determined for temperatures wavelength such as 222 nm is chosen and the ellipticity is determined for temperatures varying through the melting transition. At the inflection point of the obtained graph, half of varying through the melting transition. At the inflection point of the obtained graph, half of the proteins are unfolded and T equals Tm. Accordingly, protein stability towards solvent the proteins are unfolded and T equals Tm. Accordingly, protein stability towards solvent denaturation can be studied by varying denaturant concentration instead of temperature. denaturation can be studied by varying denaturant concentration instead of temperature. Sebastian Grimm 11 Sebastian Grimm 11

1.2.5 Enzyme activity 1.2.5 Enzyme activity

Enzymes are mostly proteins catalyzing chemical reactions that define cellular metabolism Enzymes are mostly proteins catalyzing chemical reactions that define cellular metabolism with up to 1017-fold rate accelerations 71 and about 10% of the human proteome is predicted to with up to 1017-fold rate accelerations 71 and about 10% of the human proteome is predicted to have enzymatic activity 72. A single substrate enzyme catalyzed reaction can be described by have enzymatic activity 72. A single substrate enzyme catalyzed reaction can be described by the Michaelis-Menten equation: the Michaelis-Menten equation:

V = [E]0[S]kcat/(KM+[S]) (7) V = [E]0[S]kcat/(KM+[S]) (7)

Here, V is the initial rate of the reaction, [E]0 is the concentration of enzyme, [S] is the Here, V is the initial rate of the reaction, [E]0 is the concentration of enzyme, [S] is the concentration of substrate, kcat is the catalytic turnover rate and KM the Michaelis-Menten concentration of substrate, kcat is the catalytic turnover rate and KM the Michaelis-Menten constant. The first-order rate constant kcat describes how fast enzyme-substrate (ES) complex constant. The first-order rate constant kcat describes how fast enzyme-substrate (ES) complex is transformed into enzyme and product. KM may be treated as an apparent equilibrium is transformed into enzyme and product. KM may be treated as an apparent equilibrium dissociation constant of an enzyme and its substrate, similar to Kd for affinity proteins, if the dissociation constant of an enzyme and its substrate, similar to Kd for affinity proteins, if the conversion of substrate to product is much slower than the dissociation of substrate from the conversion of substrate to product is much slower than the dissociation of substrate from the enzyme. An enzyme’s activity or ability to convert a certain amount of substrate into product enzyme. An enzyme’s activity or ability to convert a certain amount of substrate into product within a given time depends both on KM (the formation of ES) and kcat (the conversion of ES within a given time depends both on KM (the formation of ES) and kcat (the conversion of ES into E and P). Enzymes are evolved for adequate activity in their biological context and into E and P). Enzymes are evolved for adequate activity in their biological context and towards their natural substrate, which can sometimes be further improved as shown for a C- towards their natural substrate, which can sometimes be further improved as shown for a C- terminally truncated glucanase 73. However, an industrial application frequently requires terminally truncated glucanase 73. However, an industrial application frequently requires activity for non-natural substrates or under non-natural physical or chemical conditions. activity for non-natural substrates or under non-natural physical or chemical conditions. Employing consecutive rounds of gene diversification and screening in combination with a Employing consecutive rounds of gene diversification and screening in combination with a surrogate substrate, Glieder an coworkers could convert a fatty acid monooxygenase into an surrogate substrate, Glieder an coworkers could convert a fatty acid monooxygenase into an alkane hydroxylase 74, with possible applications in hydrocarbon oxidations. You and Arnold alkane hydroxylase 74, with possible applications in hydrocarbon oxidations. You and Arnold could isolate mutants of the serine endopeptidase subtilisin with activity in 60% dimethyl could isolate mutants of the serine endopeptidase subtilisin with activity in 60% dimethyl formamide 75. formamide 75.

1.2.6 Methods to analyze enzyme activity 1.2.6 Methods to analyze enzyme activity

Prior to measuring an enzyme’s activity, the chemical changes during conversion of substrate Prior to measuring an enzyme’s activity, the chemical changes during conversion of substrate to product need to be identified. Enzymes can catalyze a variety of different reactions such as to product need to be identified. Enzymes can catalyze a variety of different reactions such as oxidation-reductions, group transfers, eliminations, isomerizations, condensations and others oxidation-reductions, group transfers, eliminations, isomerizations, condensations and others 76 76 . For the determination of KM and kcat, the initial rate of the conversion of substrate to . For the determination of KM and kcat, the initial rate of the conversion of substrate to product in relation to substrate concentration needs to be measured. One approach is to product in relation to substrate concentration needs to be measured. One approach is to continuously monitor the disappearance of substrate or appearance of product by measuring continuously monitor the disappearance of substrate or appearance of product by measuring spectroscopic measurable properties such as UV/vis absorption, fluorescence emission or spectroscopic measurable properties such as UV/vis absorption, fluorescence emission or phosphorescence. For enzymes such as kinases or proteases, artificial fluorophore-containing phosphorescence. For enzymes such as kinases or proteases, artificial fluorophore-containing substrates may be used, that undergo fluorescence emission change upon phosphorylation or substrates may be used, that undergo fluorescence emission change upon phosphorylation or cleavage 77. For substrates and products without spectroscopic measurable properties, the cleavage 77. For substrates and products without spectroscopic measurable properties, the 12 Ribosome display for selection and evolution of affibody molecules 12 Ribosome display for selection and evolution of affibody molecules reaction may be coupled to a second, enzymatic reaction that involves for example the reaction may be coupled to a second, enzymatic reaction that involves for example the chromogenic coenzyme NADH. chromogenic coenzyme NADH. If a continuous measurement is not feasible, reactions may be stopped or quenched If a continuous measurement is not feasible, reactions may be stopped or quenched after a given time and substrate and product can be fractionated for quantification using after a given time and substrate and product can be fractionated for quantification using techniques such as high performance liquid chromatography (HPLC), electrophoresis or techniques such as high performance liquid chromatography (HPLC), electrophoresis or selective precipitation 78. For both continuous and discontinuous assays, the initial rates V = - selective precipitation 78. For both continuous and discontinuous assays, the initial rates V = - dS/dt = dP/dt can be determined for different substrate concentration and, assuming dS/dt = dP/dt can be determined for different substrate concentration and, assuming

Michaelis-Menten kinetics, KM is obtained from a plot of V against substrate concentration. Michaelis-Menten kinetics, KM is obtained from a plot of V against substrate concentration.

For very high substrate concentrations, V converts against the maximum rate Vmax and kcat is For very high substrate concentrations, V converts against the maximum rate Vmax and kcat is obtained from Vmax = kcat[ET], where [ET] is the total concentration of enzyme. obtained from Vmax = kcat[ET], where [ET] is the total concentration of enzyme. Sebastian Grimm 13 Sebastian Grimm 13

1.2.7 BOX: recombinant DNA technology 1.2.7 BOX: recombinant DNA technology

Recombinant DNA technology is the protein engineer’s toolbox that opened the door to the Recombinant DNA technology is the protein engineer’s toolbox that opened the door to the study of sequence-function or genotype-phenotype relationships. Among the tools are well- study of sequence-function or genotype-phenotype relationships. Among the tools are well- characterized enzymes that selectively cleave DNA at certain sequence motifs (restriction characterized enzymes that selectively cleave DNA at certain sequence motifs (restriction endonucleases) and enzymes that can covalently join DNA fragments (ligases). DNA shuttling endonucleases) and enzymes that can covalently join DNA fragments (ligases). DNA shuttling constructs (vectors) allow for propagation and expression of recombinant DNA in a host cell. constructs (vectors) allow for propagation and expression of recombinant DNA in a host cell. Frequently used hosts are Escherichia coli, yeast, insect cells or mammalian cells. DNA Frequently used hosts are Escherichia coli, yeast, insect cells or mammalian cells. DNA polymerases catalyze the polymerization of deoxynucleotides into a DNA strand. In polymerases catalyze the polymerization of deoxynucleotides into a DNA strand. In combination with a pair of specific oligonucleotides (primers), certain fragments of the template combination with a pair of specific oligonucleotides (primers), certain fragments of the template can be mass-amplified in a polymerase chain reaction (PCR). Alterations or extensions of the can be mass-amplified in a polymerase chain reaction (PCR). Alterations or extensions of the template sequence can be introduced with the applied primers. Often, DNA-polymerase is used template sequence can be introduced with the applied primers. Often, DNA-polymerase is used in combination with its cousin reverse transcriptase, an enzyme that also catalyzes the in combination with its cousin reverse transcriptase, an enzyme that also catalyzes the polymerization of deoxynucleotides into a DNA strand, but accepts RNA instead of DNA as polymerization of deoxynucleotides into a DNA strand, but accepts RNA instead of DNA as template. The combination allows for the mass amplification and handling of mRNA-encoded template. The combination allows for the mass amplification and handling of mRNA-encoded genes. Another application of DNA polymerases is the sequencing of DNA. The still most genes. Another application of DNA polymerases is the sequencing of DNA. The still most widely used technique uses di-deoxynucleotides for chain termination and was originally widely used technique uses di-deoxynucleotides for chain termination and was originally described by Sanger in 1975 79. Today, other sequencing methods such as Pyrosequencing 80, described by Sanger in 1975 79. Today, other sequencing methods such as Pyrosequencing 80, Solexa and SOLiD sequencing 81 are used for high-resolution and deep sequencing of whole Solexa and SOLiD sequencing 81 are used for high-resolution and deep sequencing of whole genomes or larger sample cohorts. genomes or larger sample cohorts.

Figure 2: The recombinant DNA technology toolbox. (1) Primed by an oligonucleotide, Figure 2: The recombinant DNA technology toolbox. (1) Primed by an oligonucleotide, reverse transcriptase converts mRNA into cDNA. (2) cDNA is mass-amplified by DNA reverse transcriptase converts mRNA into cDNA. (2) cDNA is mass-amplified by DNA polymerase and two primers. One of these primers introduces a site-directed mutation (red). The polymerase and two primers. One of these primers introduces a site-directed mutation (red). The mutated DNA is cleaved by restriction endonucleases (3) to create protruding, sticky ends and mutated DNA is cleaved by restriction endonucleases (3) to create protruding, sticky ends and ligated with vector DNA by DNA ligase (4) for transformation to a microbial host (5). The host ligated with vector DNA by DNA ligase (4) for transformation to a microbial host (5). The host allows for conservation and amplification of the mutated DNA or for production of the encoded allows for conservation and amplification of the mutated DNA or for production of the encoded recombinant protein. recombinant protein. 14 Ribosome display for selection and evolution of affibody molecules 14 Ribosome display for selection and evolution of affibody molecules

1.3 Protein engineering by rational approaches 1.3 Protein engineering by rational approaches

The protein traits discussed in chapter 1.2 can be addressed using either rational or The protein traits discussed in chapter 1.2 can be addressed using either rational or combinatorial library-based protein engineering approaches. Rational approaches are often combinatorial library-based protein engineering approaches. Rational approaches are often guided by structural information and involve (a) chemical or enzymatic modifications, (b) guided by structural information and involve (a) chemical or enzymatic modifications, (b) genetic fusions or deletions, (c) site-directed mutagenesis or (d) de novo protein design. genetic fusions or deletions, (c) site-directed mutagenesis or (d) de novo protein design. (a) Individual amino acid side chains can be selectively chemically modified, if their (a) Individual amino acid side chains can be selectively chemically modified, if their characteristic atom composition and chemical properties differ from the rest of the residues in characteristic atom composition and chemical properties differ from the rest of the residues in the protein. For example, cysteine is the only amino acid that contains a sulfhydryl group and the protein. For example, cysteine is the only amino acid that contains a sulfhydryl group and can selectively react with maleimides in a Michael addition type of reaction. The primary can selectively react with maleimides in a Michael addition type of reaction. The primary amino group of lysine residues can react as a nucleophile with N-hydroxysuccinimide (NHS) amino group of lysine residues can react as a nucleophile with N-hydroxysuccinimide (NHS) activated ester groups to form a peptide bond. Both chemistries allow for site-directed activated ester groups to form a peptide bond. Both chemistries allow for site-directed addition of functional groups such as biotin, polyethylene glycol or fluorophores. A lysine addition of functional groups such as biotin, polyethylene glycol or fluorophores. A lysine residue with a very perturbed pKa was reported in the active site of the catalytic antibody residue with a very perturbed pKa was reported in the active site of the catalytic antibody m38C2 and selectively reacts with β-diketones to form a covalent bond, i.e. allowing the m38C2 and selectively reacts with β-diketones to form a covalent bond, i.e. allowing the coupling of small drug molecules to the antibody 82. coupling of small drug molecules to the antibody 82. (b) For protein detection, purification or to increase the solubility, stretches of oligo- (b) For protein detection, purification or to increase the solubility, stretches of oligo- amino acids such as His6, Flag or Strep tag or whole proteins such as glutathione S-transferase amino acids such as His6, Flag or Strep tag or whole proteins such as glutathione S-transferase (GST), maltose binding protein (MBP) or N utilization substance protein A (NusA) can (GST), maltose binding protein (MBP) or N utilization substance protein A (NusA) can genetically be fused to the protein of interest 83. To harness avidity effects of multiple, genetically be fused to the protein of interest 83. To harness avidity effects of multiple, closely-situated binding sites, oligomerization domains such as the coiled-coil assembly closely-situated binding sites, oligomerization domains such as the coiled-coil assembly domain of the cartilage oligomeric matrix protein 84 can be fused to ligands for affinity domain of the cartilage oligomeric matrix protein 84 can be fused to ligands for affinity applications. Immunotoxin-fusions combine the selectivity of an antibody with the cell-killing applications. Immunotoxin-fusions combine the selectivity of an antibody with the cell-killing potency of a toxin, such as domains of the Pseudomonas aeruginosa exotoxin A 85. Yet potency of a toxin, such as domains of the Pseudomonas aeruginosa exotoxin A 85. Yet another class of fusions are peptides and proteins that are recruited to increase the half-life of another class of fusions are peptides and proteins that are recruited to increase the half-life of therapeutic proteins in a patient’s circulation. Among these are unstructured peptides that therapeutic proteins in a patient’s circulation. Among these are unstructured peptides that mainly extend the hydrodynamic radius of the protein 86, albumin itself and proteins to bind mainly extend the hydrodynamic radius of the protein 86, albumin itself and proteins to bind albumin such as engineered antibody fragments (see chapter 3) or a streptococcal albumin albumin such as engineered antibody fragments (see chapter 3) or a streptococcal albumin binding domain (ABD). The latter two examples increase the hydrodynamic radius via binding domain (ABD). The latter two examples increase the hydrodynamic radius via binding to albumin and harness endosomal recycling mechanisms via albumin binding to the binding to albumin and harness endosomal recycling mechanisms via albumin binding to the neonatal Fc receptor (FcRn), i.e. hitch-hiking the albumin-mediated recycling 87. neonatal Fc receptor (FcRn), i.e. hitch-hiking the albumin-mediated recycling 87. (c) Site directed mutagenesis became available in 1978 14 and was extensively used for (c) Site directed mutagenesis became available in 1978 14 and was extensively used for the structure-guided engineering of proteins, particularly enzymes. Winter and Fersht studied the structure-guided engineering of proteins, particularly enzymes. Winter and Fersht studied the tyrosyl-transfer RNA synthetase, an enzyme responsible for the ATP-dependent ligation the tyrosyl-transfer RNA synthetase, an enzyme responsible for the ATP-dependent ligation of tyrosine to its cognate tRNA. A detailed understanding of the structure and reaction of tyrosine to its cognate tRNA. A detailed understanding of the structure and reaction mechanism of the enzyme were the prerequisite for the exchange of certain amino acids. At mechanism of the enzyme were the prerequisite for the exchange of certain amino acids. At first residue cysteine 35 was mutated to serine, because it formed a hydrogen bond to the first residue cysteine 35 was mutated to serine, because it formed a hydrogen bond to the ribose part of ATP in the crystal structure and the mutated enzyme was indeed less active than ribose part of ATP in the crystal structure and the mutated enzyme was indeed less active than Sebastian Grimm 15 Sebastian Grimm 15 the wild-type. Other, more distant residues were also mutated and some of them were shown the wild-type. Other, more distant residues were also mutated and some of them were shown to significantly contribute to the catalytic rate, probably by binding to the γ-phosphate of ATP to significantly contribute to the catalytic rate, probably by binding to the γ-phosphate of ATP 88. A non-enzyme example is the engineering of insulin, an essential hormone for the 88. A non-enzyme example is the engineering of insulin, an essential hormone for the regulation of carbohydrate metabolism, and one of the first recombinantly produced regulation of carbohydrate metabolism, and one of the first recombinantly produced biological drugs for patients with type I diabetes. Based on crystal structure information, A biological drugs for patients with type I diabetes. Based on crystal structure information, A chain asparagine 21 was exchanged with glycine and two arginine residues were added to the chain asparagine 21 was exchanged with glycine and two arginine residues were added to the carboxy terminus of the B chain to obtain an insulin variant with a prolonged duration of carboxy terminus of the B chain to obtain an insulin variant with a prolonged duration of action 89. These substitutions elevated the isoelectric point pI from 5.4 to 6.7 and stabilized a action 89. These substitutions elevated the isoelectric point pI from 5.4 to 6.7 and stabilized a hexameric crystal packing, i.e. insulin precipitates after subcutaneous injection in the patient hexameric crystal packing, i.e. insulin precipitates after subcutaneous injection in the patient and is slowly released. Structural information is however not always available for all proteins. and is slowly released. Structural information is however not always available for all proteins. In such cases, single amino acid residues can systematically be exchanged to alanine residues In such cases, single amino acid residues can systematically be exchanged to alanine residues (alanine scan), and the impact of substitutions on the trait can be investigated 90. (alanine scan), and the impact of substitutions on the trait can be investigated 90. (d) De novo protein design involves the identification of functional proteins from (d) De novo protein design involves the identification of functional proteins from novel amino acid sequences in silico. Mainly due to our incomplete understanding of protein novel amino acid sequences in silico. Mainly due to our incomplete understanding of protein folding and dynamics, this approach is among the most challenging of all protein engineering folding and dynamics, this approach is among the most challenging of all protein engineering endeavors. Typically, force field functions are used that approximate all different non- endeavors. Typically, force field functions are used that approximate all different non- covalent interactions within the unfolded or folded state of a protein. The potential energy of covalent interactions within the unfolded or folded state of a protein. The potential energy of the force field function is then iteratively minimized to predict the folded structure of the the force field function is then iteratively minimized to predict the folded structure of the protein, i.e. novel amino acid sequences that fold into desired structures can be identified. De protein, i.e. novel amino acid sequences that fold into desired structures can be identified. De novo protein design has for example been used to create proteins with an α-helical bundle or novo protein design has for example been used to create proteins with an α-helical bundle or α-helix and β-sheet structure 91. Koder and colleagues could design a four-helix bundle α-helix and β-sheet structure 91. Koder and colleagues could design a four-helix bundle protein capable of accommodating haem groups in its hydrophobic core that mediate oxygen protein capable of accommodating haem groups in its hydrophobic core that mediate oxygen binding 31. binding 31. 16 Ribosome display for selection and evolution of affibody molecules 16 Ribosome display for selection and evolution of affibody molecules

Protein Trait Approach Achievement Ref. Protein Trait Approach Achievement Ref. Site-directed mutagenesis, Incorporation of novel transition Site-directed mutagenesis, Incorporation of novel transition Thioredoxin Affinity 93 Thioredoxin Affinity 93 molecular modeling metal binding site molecular modeling metal binding site Structure-based PDZ domains with 0.1-96 µM K 94; Structure-based PDZ domains with 0.1-96 µM K 94; PDZ domain Affinity d PDZ domain Affinity d computational design for peptide ligands 95 computational design for peptide ligands 95 Humanized antibody directed Humanized antibody directed Antibody Affinity CDR loop grafting 96 Antibody Affinity CDR loop grafting 96 against human lymphocytes against human lymphocytes Pentameric peptide with 1 nM Pentameric peptide with 1 nM Peptide Avidity Fusion to coiled-coil domain 84 Peptide Avidity Fusion to coiled-coil domain 84 avidity to B-cell lymphoma Ig avidity to B-cell lymphoma Ig Lactate Conversion into malate Lactate Conversion into malate Selectivity Site-directed mutagenesis 18 Selectivity Site-directed mutagenesis 18 dehydrogenase dehydrogenase dehydrogenase dehydrogenase

T4 lysozyme Stability Insertion of a disulphide bond Increased thermal stability 17 T4 lysozyme Stability Insertion of a disulphide bond Increased thermal stability 17

Tyrosyl-tRNA Catalytic Tyrosyl-tRNA Catalytic Site-directed mutagenesis Catalytic mechanism revealed 88 Site-directed mutagenesis Catalytic mechanism revealed 88 synthetase activity synthetase activity Pharmaco- Catalytic antibody reacts with Pharmaco- Catalytic antibody reacts with Antibody Site-directed attachment of drug 82 Antibody Site-directed attachment of drug 82 kinetics its substrate kinetics its substrate Pharmaco- Engineering of crystal 8 to 26 h duration of action in Pharmaco- Engineering of crystal 8 to 26 h duration of action in Insulin 89 Insulin 89 kinetics contacts humans kinetics contacts humans Designed 4-helix Accommodation of haem Novel oxygen carrier protein Designed 4-helix Accommodation of haem Novel oxygen carrier protein Transport 31 Transport 31 bundle protein groups designed bundle protein groups designed

Table 1: Rational approaches in protein engineering. Examples of rational protein engineering Table 1: Rational approaches in protein engineering. Examples of rational protein engineering efforts are grouped according to the addressed protein trait. Columns include the engineered protein, efforts are grouped according to the addressed protein trait. Columns include the engineered protein, the addressed protein trait, a short description of the rational approach used and a short summary of the addressed protein trait, a short description of the rational approach used and a short summary of the achievement. the achievement. Sebastian Grimm 17 Sebastian Grimm 17

1.4 Protein engineering by directed evolution 1.4 Protein engineering by directed evolution 1.4.1 Introduction 1.4.1 Introduction

The emergence of recombinant DNA technology in combination with protein structure The emergence of recombinant DNA technology in combination with protein structure information envisioned to tailor any protein trait according to the designers intention. This information envisioned to tailor any protein trait according to the designers intention. This prospect was however not yet fulfilled, for mainly two reasons. First, proteins are highly prospect was however not yet fulfilled, for mainly two reasons. First, proteins are highly cooperative macromolecules, meaning that many amino acids may contribute together to a cooperative macromolecules, meaning that many amino acids may contribute together to a protein trait and the contribution of a single amino is hard to assign. Second, protein traits are protein trait and the contribution of a single amino is hard to assign. Second, protein traits are strongly influenced by their solvent surrounding, generating another level of complexity that strongly influenced by their solvent surrounding, generating another level of complexity that complicates predictions considerably. As an example, an attempt by Schiweck and Skerra to complicates predictions considerably. As an example, an attempt by Schiweck and Skerra to reshape an antibody binding site by the transplantation of designed hypervariable loops failed reshape an antibody binding site by the transplantation of designed hypervariable loops failed due to the differing orientation of a single tryptophan residue 20, supporting the notion that due to the differing orientation of a single tryptophan residue 20, supporting the notion that already subtle changes can greatly affect a protein trait. already subtle changes can greatly affect a protein trait.

To overcome these limitations, the power of natural selection can be harnessed for To overcome these limitations, the power of natural selection can be harnessed for protein engineering in a method referred to as directed evolution. In principle, a directed protein engineering in a method referred to as directed evolution. In principle, a directed evolution experiment involves repetitive rounds of diversification and selection. To get a evolution experiment involves repetitive rounds of diversification and selection. To get a picture of this procedure, one can imagine a fitness landscape with a protein space consisting picture of this procedure, one can imagine a fitness landscape with a protein space consisting of all possible amino acid combinations on the x and y axes and the corresponding protein of all possible amino acid combinations on the x and y axes and the corresponding protein trait or fitness parameter plotted on the z axis. This landscape is populated by protein variants trait or fitness parameter plotted on the z axis. This landscape is populated by protein variants after diversification, and a selection system is used to sample the fitness landscape for those after diversification, and a selection system is used to sample the fitness landscape for those variants with the best fitness parameter such as affinity or stability (Figure 3). Employing variants with the best fitness parameter such as affinity or stability (Figure 3). Employing recombinant DNA technology, diversity is introduced by the design of a DNA library where recombinant DNA technology, diversity is introduced by the design of a DNA library where some general considerations need to be done. some general considerations need to be done. The size of the library: the generally accepted dogma is “big is beautiful”. The bigger The size of the library: the generally accepted dogma is “big is beautiful”. The bigger the library, the better protein trait can be expected due to a denser population of the fitness the library, the better protein trait can be expected due to a denser population of the fitness landscape 97. In practice, the library size is limited by the selection system to be used (see landscape 97. In practice, the library size is limited by the selection system to be used (see chapter 2) and according to an in silico study, a sparse population of a larger fitness landscape chapter 2) and according to an in silico study, a sparse population of a larger fitness landscape may yield higher affinity binders than a dense population of a smaller fitness landscape 98. may yield higher affinity binders than a dense population of a smaller fitness landscape 98. The location of the varied positions: this depends on the protein and the addressed The location of the varied positions: this depends on the protein and the addressed protein trait. Affinity requires primarily amino acid positions on surface-exposed, binding- protein trait. Affinity requires primarily amino acid positions on surface-exposed, binding- competent peptide loops or secondary structure motifs of a protein scaffold to be diversified. competent peptide loops or secondary structure motifs of a protein scaffold to be diversified. However, to increase the affinity of a firstly selected binder, also core mutations may be However, to increase the affinity of a firstly selected binder, also core mutations may be applied that stabilize the protein scaffold. Examples of scaffolds are llama antibody fragments applied that stabilize the protein scaffold. Examples of scaffolds are llama antibody fragments for loops, affibody molecules for secondary structure motifs and designed ankyrin repeat for loops, affibody molecules for secondary structure motifs and designed ankyrin repeat proteins for combinations thereof. A detailed description of protein scaffolds for molecular proteins for combinations thereof. A detailed description of protein scaffolds for molecular recognition is given in chapter 3. An increased stability may require amino acid positions in recognition is given in chapter 3. An increased stability may require amino acid positions in the proteins hydrophobic core to be diversified. Here, information on the protein structure is the proteins hydrophobic core to be diversified. Here, information on the protein structure is 18 Ribosome display for selection and evolution of affibody molecules 18 Ribosome display for selection and evolution of affibody molecules very helpful to guide the choice of positions, and helps to limit the size of the library that very helpful to guide the choice of positions, and helps to limit the size of the library that exponentially increases with the total number of randomized amino acid positions. exponentially increases with the total number of randomized amino acid positions.

Figure 3: Protein fitness landscapes. (a) A selectable protein trait or fitness parameter is plotted Figure 3: Protein fitness landscapes. (a) A selectable protein trait or fitness parameter is plotted against protein sequence space to obtain a fitness landscape, which is populated by a protein library against protein sequence space to obtain a fitness landscape, which is populated by a protein library and sampled by a protein selection system. The transition from black over red and yellow to white and sampled by a protein selection system. The transition from black over red and yellow to white indicates increasing fitness. Most of the sequences presumably yield non-functional proteins (black). indicates increasing fitness. Most of the sequences presumably yield non-functional proteins (black). (b) Such fitness landscapes can range from very rugged with multiple local maxima (left) to smooth (b) Such fitness landscapes can range from very rugged with multiple local maxima (left) to smooth with one global maximum (right). The latter maximum can iteratively be reached in a directed with one global maximum (right). The latter maximum can iteratively be reached in a directed evolution experiment through repeated rounds of diversification and selection, while a rugged evolution experiment through repeated rounds of diversification and selection, while a rugged landscape contains non-functional, i.e. non-selectable proteins within local maxima and is hence less landscape contains non-functional, i.e. non-selectable proteins within local maxima and is hence less suitable. (c) Local maxima may occur that “trap” a directed evolution path and prevent it from suitable. (c) Local maxima may occur that “trap” a directed evolution path and prevent it from reaching the global maximum (red trace). However, other such paths may exist, that reach the global reaching the global maximum (red trace). However, other such paths may exist, that reach the global maximum (green trace). Figure reprinted with permission 99. maximum (green trace). Figure reprinted with permission 99. Sebastian Grimm 19 Sebastian Grimm 19

The amino acid composition of the varied positions: charged residues such as arginine, The amino acid composition of the varied positions: charged residues such as arginine, lysine and glutamate as well as the aromatic residue tyrosine are frequently found in binding lysine and glutamate as well as the aromatic residue tyrosine are frequently found in binding interfaces 100 and may be beneficial for engineering affinity. However, also libraries interfaces 100 and may be beneficial for engineering affinity. However, also libraries composed of solely serine and tyrosine were reported to yield high affinity binders 101, but composed of solely serine and tyrosine were reported to yield high affinity binders 101, but have been debated 102. Helix-breaking combinations of proline and glycine should be avoided have been debated 102. Helix-breaking combinations of proline and glycine should be avoided in alpha-helices and cysteine-mediated dimers may lead to unwanted avidity effect in the in alpha-helices and cysteine-mediated dimers may lead to unwanted avidity effect in the selection step. selection step.

1.4.2 Sources of diversity 1.4.2 Sources of diversity

A protein engineer can introduce diversity by different means: employing natural diversity, A protein engineer can introduce diversity by different means: employing natural diversity, oligonucleotide cassette mutagenesis, error-prone PCR, gene shuffling or combinations oligonucleotide cassette mutagenesis, error-prone PCR, gene shuffling or combinations thereof (Figure 4). A well-described natural repertoire of diversity is that of binding proteins thereof (Figure 4). A well-described natural repertoire of diversity is that of binding proteins generated by the immune system of vertebrates, denoted antbodies. The variable gene pool generated by the immune system of vertebrates, denoted antbodies. The variable gene pool can be isolated from B cell precursors in the bone marrow or from B cells circulating in the can be isolated from B cell precursors in the bone marrow or from B cells circulating in the blood or in peripheral lymphatic organs (naïve library). An immunization with a desired blood or in peripheral lymphatic organs (naïve library). An immunization with a desired antigen prior to isolation may enrich the gene pool for binder-encoding genes on the B cell antigen prior to isolation may enrich the gene pool for binder-encoding genes on the B cell level (immune library). Natural repertoires from various sources, such as mice 103, sharks 104 level (immune library). Natural repertoires from various sources, such as mice 103, sharks 104 or humans 105 have been employed for the isolation of binders. or humans 105 have been employed for the isolation of binders. A widely-used, synthetic approach for creating diversity is oligonucleotide cassette A widely-used, synthetic approach for creating diversity is oligonucleotide cassette mutagenesis. During oligonucleotide synthesis, mixtures of nucleotides can be incorporated at mutagenesis. During oligonucleotide synthesis, mixtures of nucleotides can be incorporated at the intended positions, resulting in mixed base triplets (codons) encoding different amino the intended positions, resulting in mixed base triplets (codons) encoding different amino acids. The complete set of amino acids is encoded using NNN codons, where N = A, T, G or acids. The complete set of amino acids is encoded using NNN codons, where N = A, T, G or C. However, also all three STOP-codons (TAA, TGA and TAG) are encoded by NNN, C. However, also all three STOP-codons (TAA, TGA and TAG) are encoded by NNN, resulting in a considerable number of sequences with premature STOP-codons. Due to the resulting in a considerable number of sequences with premature STOP-codons. Due to the degeneracy of the genetic code, the 20 amino acids will not be evenly distributed in an NNN degeneracy of the genetic code, the 20 amino acids will not be evenly distributed in an NNN library. Leucin will be encoded by six possible codons, while tryptophan will be encoded by library. Leucin will be encoded by six possible codons, while tryptophan will be encoded by only a single codon. An alternative is the use of NNK codons, where K = G or T, which still only a single codon. An alternative is the use of NNK codons, where K = G or T, which still code for all the 20 amino acids but only the amber TAG STOP-codon. Furthermore, the code for all the 20 amino acids but only the amber TAG STOP-codon. Furthermore, the amino acid distribution of the NNK codon is more even than that of the NNN codon. Other amino acid distribution of the NNK codon is more even than that of the NNN codon. Other interesting codons are NAN that encodes only polar amino acids, while NTN encodes only interesting codons are NAN that encodes only polar amino acids, while NTN encodes only apolar ones. Such codons are less diverse and allow the investigation of more positions at the apolar ones. Such codons are less diverse and allow the investigation of more positions at the same degree of library coverage than more degenerate codons. In a study of Fellouse and same degree of library coverage than more degenerate codons. In a study of Fellouse and colleagues, a TMT codon was used, where M = A or C, to construct a “binary” antibody colleagues, a TMT codon was used, where M = A or C, to construct a “binary” antibody library encoding only tyrosin and serin residues on the antigen binding site. From this library, library encoding only tyrosin and serin residues on the antigen binding site. From this library, 106 106 an antibody with a Kd of 60 nM has been selected . In another study, Rajpal and colleagues an antibody with a Kd of 60 nM has been selected . In another study, Rajpal and colleagues used combinations of oligonucleotides, each encoding single aa substitutions on the antigen used combinations of oligonucleotides, each encoding single aa substitutions on the antigen 20 Ribosome display for selection and evolution of affibody molecules 20 Ribosome display for selection and evolution of affibody molecules binding site, to improve the affinity of an antibody for TNF-α by 500- to 870-fold. This binding site, to improve the affinity of an antibody for TNF-α by 500- to 870-fold. This approach was termed look-through mutagenesis and in principle, each individual aa residue approach was termed look-through mutagenesis and in principle, each individual aa residue on the antigen binding site was substituted with any of the nine aa A, S, L, Y, D, Q, K, H or P on the antigen binding site was substituted with any of the nine aa A, S, L, Y, D, Q, K, H or P that represent the side chain chemistries of the 20 natural amino acids. Affinity-improving that represent the side chain chemistries of the 20 natural amino acids. Affinity-improving 107 107 substitutions were selected and rationally combined to reach a final Kd of 1.1 pM . substitutions were selected and rationally combined to reach a final Kd of 1.1 pM . An elegant approach is to couple readily synthesized trinucleotide building blocks to An elegant approach is to couple readily synthesized trinucleotide building blocks to library cassettes 108; 109. Therewith, STOP-codons or any other codons such as those coding library cassettes 108; 109. Therewith, STOP-codons or any other codons such as those coding for cysteines can be completely avoided, and any desired amino acid combination can be for cysteines can be completely avoided, and any desired amino acid combination can be achieved. Today, such libraries generated by yet another approach, built on double-stranded achieved. Today, such libraries generated by yet another approach, built on double-stranded DNA triplets, is offered by a commercial antibody producer. DNA triplets, is offered by a commercial antibody producer.

A non-directed method for creating diversity is the random generation of single base A non-directed method for creating diversity is the random generation of single base mutations by a “sloppy” DNA polymerase activity, denoted error-prone PCR. The original mutations by a “sloppy” DNA polymerase activity, denoted error-prone PCR. The original heat-stable Taq DNA polymerase has an error-rate of about 1 in 9000 nucleotides 110. heat-stable Taq DNA polymerase has an error-rate of about 1 in 9000 nucleotides 110. Assuming a gene of 300 bp, about 30 PCR cycles would result in an average of one mutation Assuming a gene of 300 bp, about 30 PCR cycles would result in an average of one mutation per gene. The error-rate of Taq DNA polymerase can be increased under error-prone reaction per gene. The error-rate of Taq DNA polymerase can be increased under error-prone reaction conditions such as (1) a higher concentration of MgCl2 to stabilize non-complementary base conditions such as (1) a higher concentration of MgCl2 to stabilize non-complementary base 111; 112 111; 112 pairs, (2) the addition of MnCl2, (3) varying the ratio of nucleotides or (4) including pairs, (2) the addition of MnCl2, (3) varying the ratio of nucleotides or (4) including nucleotide analogues such as 8-oxo-dGTP or dITP 113; 114 or (5) by the use of engineered DNA nucleotide analogues such as 8-oxo-dGTP or dITP 113; 114 or (5) by the use of engineered DNA polymerase mutants. A general concern in error-prone PCR reactions are biases, which can be polymerase mutants. A general concern in error-prone PCR reactions are biases, which can be examined by comparing the rate of transitions (purine/purine exchanges; Ts) with the rate of examined by comparing the rate of transitions (purine/purine exchanges; Ts) with the rate of transversions (purine/pyrimidine exchanges; Tv), or the rate of AT to GC exchanges with the transversions (purine/pyrimidine exchanges; Tv), or the rate of AT to GC exchanges with the rate of GC to AT exchanges. In general, Taq DNA polymerase generates more transitions rate of GC to AT exchanges. In general, Taq DNA polymerase generates more transitions than transversions 115 and more AT to GC than GC to AT exchanges 95, resulting in an than transversions 115 and more AT to GC than GC to AT exchanges 95, resulting in an enrichment of GC in sequence. These biases can be overcome by the random incorporation of enrichment of GC in sequence. These biases can be overcome by the random incorporation of the three different degenerated bases dITP, dPTP and dKTP, that can in a subsequent PCR the three different degenerated bases dITP, dPTP and dKTP, that can in a subsequent PCR reaction pair with any of the four regular bases with different probabilities. The desired reaction pair with any of the four regular bases with different probabilities. The desired frequency and distribution of mutations can be adjusted with the concentration and ratio of the frequency and distribution of mutations can be adjusted with the concentration and ratio of the different degenerated bases 116. Another way is to mix Taq DNA polymerase with another, different degenerated bases 116. Another way is to mix Taq DNA polymerase with another, engineered DNA polymerase with an inversed mutational spectrum 95. However, a principal engineered DNA polymerase with an inversed mutational spectrum 95. However, a principal concern remains due to the degeneracy of the genetic code: the change to any of the 20 amino concern remains due to the degeneracy of the genetic code: the change to any of the 20 amino acids requires the change of up to three bases within a single codon, an event very unlikely to acids requires the change of up to three bases within a single codon, an event very unlikely to occur under error-prone PCR conditions. Single base substitutions within a codon allow for in occur under error-prone PCR conditions. Single base substitutions within a codon allow for in average 5.7 different amino acid substitutions 117. Furthermore, most of the random occurring average 5.7 different amino acid substitutions 117. Furthermore, most of the random occurring mutations will be adverse 118 and a powerful selection system is needed to isolate the few mutations will be adverse 118 and a powerful selection system is needed to isolate the few favorable mutants. favorable mutants.

Sebastian Grimm 21 Sebastian Grimm 21

Error-prone PCR and the selection of a set of improved protein variants is often Error-prone PCR and the selection of a set of improved protein variants is often followed by a DNA shuffling procedure that allows for recombination of the improved followed by a DNA shuffling procedure that allows for recombination of the improved variants to find even better variants. DNA shuffling can also be performed directly on a set of variants to find even better variants. DNA shuffling can also be performed directly on a set of cognate genes that one would like to assemble into mosaics, e.g. 10 homologous enzyme cognate genes that one would like to assemble into mosaics, e.g. 10 homologous enzyme genes. In its original description by Stemmer, a collection of highly homologous sequences is genes. In its original description by Stemmer, a collection of highly homologous sequences is fragmented by DNaseI digestion, and the fragments are assembled into full-length mosaics by fragmented by DNaseI digestion, and the fragments are assembled into full-length mosaics by repeated cycles of annealing in the presence of DNA polymerase followed by a regular PCR repeated cycles of annealing in the presence of DNA polymerase followed by a regular PCR in the presence of primers designed to complement the ends of the strands 119. With this in the presence of primers designed to complement the ends of the strands 119. With this method, for example chimeric enzymes with improved activity and stability can be isolated. method, for example chimeric enzymes with improved activity and stability can be isolated. Crameri and colleagues shuffled four genes of the β-lactamase cephalosporinase from four Crameri and colleagues shuffled four genes of the β-lactamase cephalosporinase from four different organisms and could, after a single round of shuffling, isolate clones with 270-540 different organisms and could, after a single round of shuffling, isolate clones with 270-540 improved activity 120. Another DNA shuffling method employs a PCR-based staggered improved activity 120. Another DNA shuffling method employs a PCR-based staggered extension process. Primers anneal to a template strand and are briefly extended by a DNA extension process. Primers anneal to a template strand and are briefly extended by a DNA polymerase. Short fragments are produced that, in following cycles may switch the template polymerase. Short fragments are produced that, in following cycles may switch the template and prime the extension of other templates that are homologous in sequence, thereby and prime the extension of other templates that are homologous in sequence, thereby generating full-length chimera over several cycles 121. Yet another DNA shuffling approach generating full-length chimera over several cycles 121. Yet another DNA shuffling approach does not require any temperature cycling. A single stranded and uracil-containing “transient does not require any temperature cycling. A single stranded and uracil-containing “transient DNA scaffold” mediates the annealing of short, homologous gene fragments. Flaps are DNA scaffold” mediates the annealing of short, homologous gene fragments. Flaps are digested, gaps are filled and nicks are ligated, followed by the removal of the scaffold strand digested, gaps are filled and nicks are ligated, followed by the removal of the scaffold strand and the preparation of double-stranded DNA. According to the authors, more crossovers can and the preparation of double-stranded DNA. According to the authors, more crossovers can be achieved than with the above described methods, allowing for a higher frequency and be achieved than with the above described methods, allowing for a higher frequency and better resolution of crossovers 122. better resolution of crossovers 122. This chapter was about the historical development of protein engineering, protein traits and This chapter was about the historical development of protein engineering, protein traits and methods to measure them, the protein engineers toolbox: recombinant DNA technology, as methods to measure them, the protein engineers toolbox: recombinant DNA technology, as well as rational and combinatorial approaches for the engineering of such protein traits. A well as rational and combinatorial approaches for the engineering of such protein traits. A general overview is illustrated in Figure 4. The following chapter will tie on combinatorial general overview is illustrated in Figure 4. The following chapter will tie on combinatorial approaches with selection systems to isolate improved protein variants from diverse libraries. approaches with selection systems to isolate improved protein variants from diverse libraries.

22 Ribosome display for selection and evolution of affibody molecules 22 Ribosome display for selection and evolution of affibody molecules

Figure 4: Rational and directed evolution approaches in protein engineering. Genes are Figure 4: Rational and directed evolution approaches in protein engineering. Genes are represented as double-helices and their encoded proteins as ellipsoids. The computer symbolizes in represented as double-helices and their encoded proteins as ellipsoids. The computer symbolizes in silico calculations. Mutated nucleotide and amino acid positions are colored. Rational approaches silico calculations. Mutated nucleotide and amino acid positions are colored. Rational approaches involve the selective chemical modification of amino acid side chains (1), fusions to proteins or involve the selective chemical modification of amino acid side chains (1), fusions to proteins or peptides (2), site-directed mutagenesis (3) and de novo protein design (4). Directed evolution peptides (2), site-directed mutagenesis (3) and de novo protein design (4). Directed evolution approaches involve gene shuffling (5), error-prone PCR (6) and oligonucleotide cassette mutagenesis approaches involve gene shuffling (5), error-prone PCR (6) and oligonucleotide cassette mutagenesis (7). (7).

Sebastian Grimm 23 Sebastian Grimm 23

2 Protein selection systems 2 Protein selection systems 2.1 Common properties 2.1 Common properties

In the previous chapter, different approaches to generate diverse populations of protein In the previous chapter, different approaches to generate diverse populations of protein encoding genes for a directed protein evolution experiment were discussed. The focus of this encoding genes for a directed protein evolution experiment were discussed. The focus of this chapter will be on systems allowing for a functional selection of variants with improved chapter will be on systems allowing for a functional selection of variants with improved properties among those populations: protein selection systems. properties among those populations: protein selection systems. In nature, certain traits within a diverse population become more frequent in In nature, certain traits within a diverse population become more frequent in subsequent generations because they confer individuals a fitness advantage when certain subsequent generations because they confer individuals a fitness advantage when certain environmental stress factors are present, such as predators, drought or coldness. These traits environmental stress factors are present, such as predators, drought or coldness. These traits are part of the individuals phenotype or observable characteristics, such as their size, shape or are part of the individuals phenotype or observable characteristics, such as their size, shape or color. Each phenotype has its corresponding genotype, the genetic blueprint that is passed on color. Each phenotype has its corresponding genotype, the genetic blueprint that is passed on to next generation and ensures the heritability and propagation of the phenotype. In the to next generation and ensures the heritability and propagation of the phenotype. In the laboratory, so called protein selection experiments mimic three fundamental features of laboratory, so called protein selection experiments mimic three fundamental features of natural selection: (1) a diverse starting pool, (2) a coupling between genotype and phenotype, natural selection: (1) a diverse starting pool, (2) a coupling between genotype and phenotype, i.e. a linkage of a protein with its encoding nucleic acid is established that allows for both an i.e. a linkage of a protein with its encoding nucleic acid is established that allows for both an identification of the selected protein variants by recombinant DNA technology and a identification of the selected protein variants by recombinant DNA technology and a continued diversification and selection process, (3) a selection pressure that confers the continued diversification and selection process, (3) a selection pressure that confers the protein variants with the desired traits a selection advantage, i.e. a functional selection of the protein variants with the desired traits a selection advantage, i.e. a functional selection of the fittest. fittest. A powerful protein selection experiment has to meet some general requirements. You A powerful protein selection experiment has to meet some general requirements. You get what you select for, i.e. the conditions should ideally be such that only the desired protein get what you select for, i.e. the conditions should ideally be such that only the desired protein trait is selected for, and that biases due to for example protein expression levels, microbial trait is selected for, and that biases due to for example protein expression levels, microbial growth or avidity effects are avoided. The enrichment should be efficient. A good benchmark growth or avidity effects are avoided. The enrichment should be efficient. A good benchmark is the enrichment factor per cycle for binders against a non-binder background and values of is the enrichment factor per cycle for binders against a non-binder background and values of up to 106-108-fold were reported 123, while a value of approximately 103-fold is more common up to 106-108-fold were reported 123, while a value of approximately 103-fold is more common 24. As different selection systems are associated with different aspects of protein expression 24. As different selection systems are associated with different aspects of protein expression and presentation, it is important to make sure that the library members are functionally and presentation, it is important to make sure that the library members are functionally represented. For example, some proteins as antibody fragments may require an oxidizing represented. For example, some proteins as antibody fragments may require an oxidizing environment for their proper folding, which makes selection systems based on intracellular environment for their proper folding, which makes selection systems based on intracellular expression less suitable 124. expression less suitable 124. The size of a library will determine the outcome of a selection. Larger functional The size of a library will determine the outcome of a selection. Larger functional libraries will cover more sequence and structure space and thus potentially yield better libraries will cover more sequence and structure space and thus potentially yield better performing molecules, provided the selection conditions are sufficiently stringent 97. The performing molecules, provided the selection conditions are sufficiently stringent 97. The speed of characterization of selected proteins to some extent depends on the selection system, speed of characterization of selected proteins to some extent depends on the selection system, and post-selection work may consume more time than the actual selection procedure. and post-selection work may consume more time than the actual selection procedure. Generally, a fast selection and subsequent characterization of selected clones, in combination Generally, a fast selection and subsequent characterization of selected clones, in combination with the possibility for automation or multiplexing is desirable 125. Protein selection systems with the possibility for automation or multiplexing is desirable 125. Protein selection systems 24 Ribosome display for selection and evolution of affibody molecules 24 Ribosome display for selection and evolution of affibody molecules fall into different categories: display systems with a physical linkage between phenotype and fall into different categories: display systems with a physical linkage between phenotype and genotype, either covalent or mediated by a ribosome, phage particle or cell, and compartment genotype, either covalent or mediated by a ribosome, phage particle or cell, and compartment systems with cognate pairs of phenotype and genotype spatially separated from other library systems with cognate pairs of phenotype and genotype spatially separated from other library members by a membrane-confined volume. Both categories can either be dependent on a host members by a membrane-confined volume. Both categories can either be dependent on a host cell’s biology or cell independent (see Figure 5 and Table 2). cell’s biology or cell independent (see Figure 5 and Table 2).

Figure 5: Schematic overview of different protein selection systems. The foreign proteins and their Figure 5: Schematic overview of different protein selection systems. The foreign proteins and their cognate genes are indicated in green. Systems (a-f) are display systems while systems (g) and (h) are cognate genes are indicated in green. Systems (a-f) are display systems while systems (g) and (h) are compartment systems. (a) phage display: the foreign protein is fused to the M13 phage pIII and compartment systems. (a) phage display: the foreign protein is fused to the M13 phage pIII and incorporated in the phage coat; (b) cell surface display: exemplarily, OmpA anchors the foreign incorporated in the phage coat; (b) cell surface display: exemplarily, OmpA anchors the foreign protein to the cell surface; (c) ribosome display: a stalled ribosome attaches the foreign protein to its protein to the cell surface; (c) ribosome display: a stalled ribosome attaches the foreign protein to its cognate mRNA; (d) mRNA display: puromycin links the foreign protein covalently to its cognate cognate mRNA; (d) mRNA display: puromycin links the foreign protein covalently to its cognate mRNA; (e) CIS display: RepA is a cis-acting protein that binds to its template DNA; (f) plasmid mRNA; (e) CIS display: RepA is a cis-acting protein that binds to its template DNA; (f) plasmid display: NF-κB homodimer binds to its template plasmid DNA; (g) water-in-oil droplets: display: NF-κB homodimer binds to its template plasmid DNA; (g) water-in-oil droplets: compartments of water in oil enclose cognate pairs of foreign protein and DNA. Exemplarily, compartments of water in oil enclose cognate pairs of foreign protein and DNA. Exemplarily, enzymes may convert substrate into product (red) within these droplets; (h) protein fragment enzymes may convert substrate into product (red) within these droplets; (h) protein fragment complementation assay: a foreign scavenger protein binds to a bait protein (red) and binding complementation assay: a foreign scavenger protein binds to a bait protein (red) and binding reconstitutes the function of two halves of β-lactamase (Bla). reconstitutes the function of two halves of β-lactamase (Bla). Sebastian Grimm 25 Sebastian Grimm 25

2.2 Phage display 2.2 Phage display

In 1985, when filamentous bacteriophage vectors were already widely used for cloning and In 1985, when filamentous bacteriophage vectors were already widely used for cloning and preparation of single stranded DNA templates for sequencing of inserts, a fundamental paper preparation of single stranded DNA templates for sequencing of inserts, a fundamental paper was published by Smith 24 that paved the way for library based protein engineering projects. was published by Smith 24 that paved the way for library based protein engineering projects. Foreign DNA fragments derived from the E. coli EcoRI restriction endonuclease were Foreign DNA fragments derived from the E. coli EcoRI restriction endonuclease were genetically fused to the gene for protein III in f1 filamentous phage DNA. The fusion proteins genetically fused to the gene for protein III in f1 filamentous phage DNA. The fusion proteins could be incorporated into correctly assembled virion coats inside which the encoding vector could be incorporated into correctly assembled virion coats inside which the encoding vector resided. Smith could enrich EcoRI-peptide displaying phages more than 1000-fold from a resided. Smith could enrich EcoRI-peptide displaying phages more than 1000-fold from a non-displaying phage background using immobilized peptide-binding antibodies. The non-displaying phage background using immobilized peptide-binding antibodies. The displaying phages retained their ability to infect E. coli and could pass their genetic displaying phages retained their ability to infect E. coli and could pass their genetic information to progenitor phages. Five years later, Scott and Smith reported the first peptide information to progenitor phages. Five years later, Scott and Smith reported the first peptide library displayed on phage. From some 107 different clones, peptides resembling the epitope library displayed on phage. From some 107 different clones, peptides resembling the epitope of an antibody could be selected over several rounds 126. In the same year, the display of of an antibody could be selected over several rounds 126. In the same year, the display of human growth hormone 127 and the first antibody fragments 128 were reported. In 1991, human growth hormone 127 and the first antibody fragments 128 were reported. In 1991, antibody fragments binding to an antigen could be selected from a library displayed on phage antibody fragments binding to an antigen could be selected from a library displayed on phage 129. During the following years, various protein traits were addressed on phage, including 129. During the following years, various protein traits were addressed on phage, including catalysis 130, stability 131 and selectivity 132. Selections are typically performed in vitro, but catalysis 130, stability 131 and selectivity 132. Selections are typically performed in vitro, but sometimes even in higher organisms 133; 134. sometimes even in higher organisms 133; 134. The filamentous bacteriophages F1, fd and M13 are a group of single-stranded DNA The filamentous bacteriophages F1, fd and M13 are a group of single-stranded DNA containing viruses with the ability to infect E. coli cells containing the F conjugative pilus as a containing viruses with the ability to infect E. coli cells containing the F conjugative pilus as a receptor. After infection, phage DNA is converted to and replicated as double stranded DNA, receptor. After infection, phage DNA is converted to and replicated as double stranded DNA, from which also the genes are transcribed. New single stranded genomes are produced and from which also the genes are transcribed. New single stranded genomes are produced and assembled with produced phage capsid proteins in the bacterial envelope and hundreds of new assembled with produced phage capsid proteins in the bacterial envelope and hundreds of new phage particles are released from the cell 135. The most commonly used phage for display is phage particles are released from the cell 135. The most commonly used phage for display is M13 which obtained its prefix according to the city Munich where it was first isolated 136. Its M13 which obtained its prefix according to the city Munich where it was first isolated 136. Its single stranded, covalently closed genome confines the length of the wild-type phage particle single stranded, covalently closed genome confines the length of the wild-type phage particle and is encapsulated by around 2700 copies of the major coat protein pVIII. The ends of the and is encapsulated by around 2700 copies of the major coat protein pVIII. The ends of the phage are decorated with about 5 copies of pVII and pIX on one end and 5 copies of pIII and phage are decorated with about 5 copies of pVII and pIX on one end and 5 copies of pIII and pVI on the other (Figure 6). Using phage vectors with single copies of coat protein genes, pVI on the other (Figure 6). Using phage vectors with single copies of coat protein genes, foreign proteins are commonly displayed on pIII 97; 137, while short peptides can also be foreign proteins are commonly displayed on pIII 97; 137, while short peptides can also be displayed on pVIII 138; 139. For an efficient display, the foreign proteins need to be efficiently displayed on pVIII 138; 139. For an efficient display, the foreign proteins need to be efficiently translocated to the bacterial envelope or periplasmic space. The commonly used general translocated to the bacterial envelope or periplasmic space. The commonly used general secretory (Sec) pathway translocates readily translated proteins in their unfolded state. Some secretory (Sec) pathway translocates readily translated proteins in their unfolded state. Some fast-folding and stable proteins such as thioredoxin or DARPins (see section 3.3.4) are not fast-folding and stable proteins such as thioredoxin or DARPins (see section 3.3.4) are not efficiently translocated when employing the Sec pathway. Alternative systems exist though efficiently translocated when employing the Sec pathway. Alternative systems exist though and the signal recognition particle (SRP)-dependent pathway translocates the nascent and the signal recognition particle (SRP)-dependent pathway translocates the nascent polypeptide chain during translation and was proposed for display of very stable and fast- polypeptide chain during translation and was proposed for display of very stable and fast- 26 Ribosome display for selection and evolution of affibody molecules 26 Ribosome display for selection and evolution of affibody molecules folding proteins and scFv antibody fragments 140; 141. A third system is the twin-arginine folding proteins and scFv antibody fragments 140; 141. A third system is the twin-arginine translocation (Tat)-mediated pathway that translocates proteins in their native, folded translocation (Tat)-mediated pathway that translocates proteins in their native, folded conformation and was used for the display of proteins that fold in the E. coli cytoplasm 142. conformation and was used for the display of proteins that fold in the E. coli cytoplasm 142. These three systems can be recruited by simply choosing an appropriate signal sequence. These three systems can be recruited by simply choosing an appropriate signal sequence. Two concerns arise with display on pIII, denoted type 3 display: first, the Two concerns arise with display on pIII, denoted type 3 display: first, the multivalency of multiple displayed copies of foreign protein may lead to avidity effects that multivalency of multiple displayed copies of foreign protein may lead to avidity effects that complicate affinity discrimination and hence the selection of high affinity binders. Notably, in complicate affinity discrimination and hence the selection of high affinity binders. Notably, in other settings such avidity effects may be irrelevant or even desired. Second, pIII is important other settings such avidity effects may be irrelevant or even desired. Second, pIII is important for the infection of E. coli and larger inserts impair the phage infectivity. for the infection of E. coli and larger inserts impair the phage infectivity.

Two modifications of protein 3 display are today frequently used which reduce the Two modifications of protein 3 display are today frequently used which reduce the copy number of displayed foreign protein on infective phage particles. Type 33 display: the copy number of displayed foreign protein on infective phage particles. Type 33 display: the phage genome encodes both wild-type pIII and pIII fused to the foreign protein. Mosaic phage genome encodes both wild-type pIII and pIII fused to the foreign protein. Mosaic phages decorated with both wild-type and fusion pIII will thus be produced. Type 3+3 phages decorated with both wild-type and fusion pIII will thus be produced. Type 3+3 display: a hybrid of M13 phage and plasmid, denoted phagemid, encodes pIII fused to the display: a hybrid of M13 phage and plasmid, denoted phagemid, encodes pIII fused to the foreign protein and contains a gene conferring antibiotic resistance, origins of replication (ori) foreign protein and contains a gene conferring antibiotic resistance, origins of replication (ori) for both M13 and E. coli and a phage packaging signal. All the remaining phage proteins, for both M13 and E. coli and a phage packaging signal. All the remaining phage proteins, including wild-type pIII, are encoded by a helper-phage that also contains a gene conferring including wild-type pIII, are encoded by a helper-phage that also contains a gene conferring antibiotic resistance and a defect phage packaging signal. Noteworthy, type 3+3 display was antibiotic resistance and a defect phage packaging signal. Noteworthy, type 3+3 display was not only developed to reduce the copy number of displayed proteins, but also to use the not only developed to reduce the copy number of displayed proteins, but also to use the simple phagemid after selection for easy production of soluble proteins in E. coli. simple phagemid after selection for easy production of soluble proteins in E. coli. The first type 3+3 display was described by Bass and co-workers, who displayed the The first type 3+3 display was described by Bass and co-workers, who displayed the human growth hormone (hGH) on phage and could enrich hGH-displaying phages from a non human growth hormone (hGH) on phage and could enrich hGH-displaying phages from a non hGH-displaying background over 5000-fold during biopanning with hGH receptor-coated hGH-displaying background over 5000-fold during biopanning with hGH receptor-coated beads 127. The E. coli origin of replication and the gene conferring antibiotic resistance allow beads 127. The E. coli origin of replication and the gene conferring antibiotic resistance allow for propagation of the phagemid in E. coli. Upon infection with helper phage, the two for propagation of the phagemid in E. coli. Upon infection with helper phage, the two different resistance genes allow for selection of bacteria that contain both phagemid and different resistance genes allow for selection of bacteria that contain both phagemid and helper phage genomes. With the induction of phage protein synthesis, also here mosaic helper phage genomes. With the induction of phage protein synthesis, also here mosaic phages decorated with both wild-type and fusion pIII are produced. Because of the defective phages decorated with both wild-type and fusion pIII are produced. Because of the defective packaging signal of the helper phage, preferentially phagemid genomes are incorporated in packaging signal of the helper phage, preferentially phagemid genomes are incorporated in the phage particles. To statistically reach monovalent display, the phagemid-encoded pIII the phage particles. To statistically reach monovalent display, the phagemid-encoded pIII fusion protein contains an amber STOP-codon between the foreign protein and pIII. In an fusion protein contains an amber STOP-codon between the foreign protein and pIII. In an amber suppressor E. coli strain, only a fraction of fusion proteins will be produced. The great amber suppressor E. coli strain, only a fraction of fusion proteins will be produced. The great majority of pIII will originate from the wild-type copy encoded by the helper phage, and thus majority of pIII will originate from the wild-type copy encoded by the helper phage, and thus a majority of wild-type pIII-only bearing phagemid particles will be produced, and only about a majority of wild-type pIII-only bearing phagemid particles will be produced, and only about 1-10% will correspond to mosaic phagemid particles displaying a single copy of the foreign 1-10% will correspond to mosaic phagemid particles displaying a single copy of the foreign protein alongside with several copies of wt pIII 143. Notably, the amber STOP was mainly protein alongside with several copies of wt pIII 143. Notably, the amber STOP was mainly Sebastian Grimm 27 Sebastian Grimm 27 introduced to be able to shuttle between suppressor and non-suppressor strains for display or introduced to be able to shuttle between suppressor and non-suppressor strains for display or soluble production purposes, respectively. soluble production purposes, respectively.

Figure 6: Schematic representation of M13 phage display systems. The wild-type M13 phage is Figure 6: Schematic representation of M13 phage display systems. The wild-type M13 phage is encapsulated by about 2700 copies of mayor coat protein pVIII. The ends of the wild-type phage encapsulated by about 2700 copies of mayor coat protein pVIII. The ends of the wild-type phage contain about 5 copies of pIII and pVI and 5 copies of pVII and pIX, respectively. The single-stranded contain about 5 copies of pIII and pVI and 5 copies of pVII and pIX, respectively. The single-stranded DNA genome encodes all phage proteins and contains a phage packaging signal. Type 3 display: the DNA genome encodes all phage proteins and contains a phage packaging signal. Type 3 display: the phage contains one copy of the foreign gene fused to gene III and about five copies of the encoded phage contains one copy of the foreign gene fused to gene III and about five copies of the encoded fusion protein are displayed. Type 33 display: the phage contains both a copy of the wild-type gene III fusion protein are displayed. Type 33 display: the phage contains both a copy of the wild-type gene III and the foreign gene fused to gene III, resulting in the display of a mixture of wild-type pIII and fusion and the foreign gene fused to gene III, resulting in the display of a mixture of wild-type pIII and fusion protein. Type 3+3 display: a helper phage contains all phage genes including wild-type gene III and no protein. Type 3+3 display: a helper phage contains all phage genes including wild-type gene III and no functional phage packaging signal, while a phagemid contains the foreign gene fused to gene III and a functional phage packaging signal, while a phagemid contains the foreign gene fused to gene III and a functional phage packaging signal, resulting in the display of on average no or one copy of fusion functional phage packaging signal, resulting in the display of on average no or one copy of fusion protein on the phagemid. protein on the phagemid.

A system for multivalent display in combination with helper phage infection was described by A system for multivalent display in combination with helper phage infection was described by Rondot and co-workers 144. Here, helper phage lacking a functional copy of p3 in its genome Rondot and co-workers 144. Here, helper phage lacking a functional copy of p3 in its genome is produced in an engineered E. coli strain that provides wild-type p3 for the assembly of is produced in an engineered E. coli strain that provides wild-type p3 for the assembly of 28 Ribosome display for selection and evolution of affibody molecules 28 Ribosome display for selection and evolution of affibody molecules helper phages. Upon infection with such denoted “hyper phages”, phagemid particles are helper phages. Upon infection with such denoted “hyper phages”, phagemid particles are produced that only display the phagemid-encoded p3 fused to the foreign protein, thus produced that only display the phagemid-encoded p3 fused to the foreign protein, thus allowing for an increase of the fraction of foreign protein-displaying phagemids. Furthermore, allowing for an increase of the fraction of foreign protein-displaying phagemids. Furthermore, a trypsin cleavage site was introduced between p3 and the displayed foreign protein, allowing a trypsin cleavage site was introduced between p3 and the displayed foreign protein, allowing for elution by enzymatic cleavage and at the same time restoring the wild-type p3 phenotype for elution by enzymatic cleavage and at the same time restoring the wild-type p3 phenotype for efficient infection 144. for efficient infection 144. A typical protein selection experiment employing phage display starts with a library A typical protein selection experiment employing phage display starts with a library gene pool that is ligated with linearized phage or phagemid DNA using recombinant DNA gene pool that is ligated with linearized phage or phagemid DNA using recombinant DNA techniques and competent E. coli cells are transformed with successfully circularized vectors. techniques and competent E. coli cells are transformed with successfully circularized vectors. In case of type 3+3 display, phagemid DNA-containing cells are infected with helper phage to In case of type 3+3 display, phagemid DNA-containing cells are infected with helper phage to produce phage particles. The transformation and subsequent culturing and helper phage produce phage particles. The transformation and subsequent culturing and helper phage infection steps in practice limit the library size 135. A typical phage display selection cycle for infection steps in practice limit the library size 135. A typical phage display selection cycle for isolation of binders to a given antigen is illustrated in Figure 7. Phages are exposed to the isolation of binders to a given antigen is illustrated in Figure 7. Phages are exposed to the antigen, which can either be immobilized on a solid support such as a paramagnetic bead or antigen, which can either be immobilized on a solid support such as a paramagnetic bead or microtiter plate (solid phase), or in solution followed by post-incubation capture onto a solid microtiter plate (solid phase), or in solution followed by post-incubation capture onto a solid support (solution phase). Non-bound or weakly-bound phages are removed by repeatedly support (solution phase). Non-bound or weakly-bound phages are removed by repeatedly washing the solid support. The selection pressure can e.g. be adjusted with the concentration washing the solid support. The selection pressure can e.g. be adjusted with the concentration of antigen, the number and duration of washing steps or the concentration of a competitor of antigen, the number and duration of washing steps or the concentration of a competitor antigen added. Subsequently, bound phages are eluted by different means such as low pH, antigen added. Subsequently, bound phages are eluted by different means such as low pH, high pH, enzymatic cleavage of the displayed protein or by the addition of molecules that high pH, enzymatic cleavage of the displayed protein or by the addition of molecules that compete with the phages for binding to the target molecule 97. The eluted phages are allowed compete with the phages for binding to the target molecule 97. The eluted phages are allowed to infect E. coli cells to amplify the phage pool for the following selection cycle. This method to infect E. coli cells to amplify the phage pool for the following selection cycle. This method is also referred to as “biopanning”, in analogy to the gold-washing procedure where gold is also referred to as “biopanning”, in analogy to the gold-washing procedure where gold grains are isolated from a background of unwanted materials. grains are isolated from a background of unwanted materials. Theoretically, one selection cycle would be enough to isolate binding phages from an Theoretically, one selection cycle would be enough to isolate binding phages from an excess of non-binding phages. In practice, due to a presence of non-selectively binding excess of non-binding phages. In practice, due to a presence of non-selectively binding phages, several consecutive selection cycles are typically needed. The elution by enzymatic phages, several consecutive selection cycles are typically needed. The elution by enzymatic cleavage in combination with the introduction of a protease cleavage site within the helper cleavage in combination with the introduction of a protease cleavage site within the helper phage wild-type pIII was described to reduce the background 145 and is routinely used in the phage wild-type pIII was described to reduce the background 145 and is routinely used in the McCafferty lab (personal communication). To avoid the selection of phages with undesired McCafferty lab (personal communication). To avoid the selection of phages with undesired binding selectivity, a pre-selection is commonly included. Prior to exposure to the target binding selectivity, a pre-selection is commonly included. Prior to exposure to the target molecule, phages are exposed to the solid support lacking the target. Titers of phages eluted molecule, phages are exposed to the solid support lacking the target. Titers of phages eluted from the pre-selection can be compared with titers of phages eluted from the selection and from the pre-selection can be compared with titers of phages eluted from the selection and high selection titers as compared to low pre-selection titers is an indicator for a successful high selection titers as compared to low pre-selection titers is an indicator for a successful selection. After the panning procedure, individual clones are screened for their binding selection. After the panning procedure, individual clones are screened for their binding properties on typically antigen-coated ELISA plates. The selected clones are applied in properties on typically antigen-coated ELISA plates. The selected clones are applied in different formats, such as phages, soluble protein expressed from phagemid in periplasmic different formats, such as phages, soluble protein expressed from phagemid in periplasmic extracts or purified protein 97. Different expression levels of different clones, that may lead to extracts or purified protein 97. Different expression levels of different clones, that may lead to Sebastian Grimm 29 Sebastian Grimm 29 biased binding signals obtained from such screenings need to be considered. An alternative is biased binding signals obtained from such screenings need to be considered. An alternative is to sub-clone the selected clones into another expression system 29 or to sequence selected to sub-clone the selected clones into another expression system 29 or to sequence selected clones and search for repeatedly occurring motifs that indicate an enrichment of particular clones and search for repeatedly occurring motifs that indicate an enrichment of particular lineages 146. After such a preliminary screening, the most promising clones are produced, lineages 146. After such a preliminary screening, the most promising clones are produced, purified and further characterized for traits such as affinity, stability and selectivity. Selection purified and further characterized for traits such as affinity, stability and selectivity. Selection for catalytic activity can be achieved by displaying the substrate together with the enzyme for catalytic activity can be achieved by displaying the substrate together with the enzyme clone on the same phage and isolation of product-displaying phage particles 130 or by affinity clone on the same phage and isolation of product-displaying phage particles 130 or by affinity selection for binding to a transition-state analogue 147. selection for binding to a transition-state analogue 147.

Probably due to its robustness, the possibility to generate relatively large libraries and Probably due to its robustness, the possibility to generate relatively large libraries and no need of costly laboratory reagents or equipment, phage display is still the most widely used no need of costly laboratory reagents or equipment, phage display is still the most widely used selection system. selection system.

Figure 7: Illustration of a typical phage display selection cycle. An initial pool of foreign protein Figure 7: Illustration of a typical phage display selection cycle. An initial pool of foreign protein displaying phages is exposed to an antigen (dark blue) immobilized on a solid support. Non-binding displaying phages is exposed to an antigen (dark blue) immobilized on a solid support. Non-binding phages are washed away, binding phages are eluted and allowed to infect E. coli cells to produce phages are washed away, binding phages are eluted and allowed to infect E. coli cells to produce phages for an additional selection cycle or to analyze the selected clones. phages for an additional selection cycle or to analyze the selected clones. 30 Ribosome display for selection and evolution of affibody molecules 30 Ribosome display for selection and evolution of affibody molecules

2.3 Cell surface display 2.3 Cell surface display

Cell surface display involves the expression of multiple copies of a recombinant protein on a Cell surface display involves the expression of multiple copies of a recombinant protein on a typically microbial host surface and offers the possibility to monitor the phenotype during typically microbial host surface and offers the possibility to monitor the phenotype during selection using flow cytometry. selection using flow cytometry.

The most frequently used hosts for display of heterologuous proteins on a cell surface The most frequently used hosts for display of heterologuous proteins on a cell surface are yeast (Saccharomyces cerevisiae) and E. coli. While yeast is a fungus encapsulated by a are yeast (Saccharomyces cerevisiae) and E. coli. While yeast is a fungus encapsulated by a single membrane and a thick layer of glucan, chitin and glycoproteins, E. coli is a gram(-) single membrane and a thick layer of glucan, chitin and glycoproteins, E. coli is a gram(-) bacterium, encapsulated by two membranes and a thin murein sacculus. The functional bacterium, encapsulated by two membranes and a thin murein sacculus. The functional display of a heterologuous protein on the E. coli outer membrane was described in 1992 148, display of a heterologuous protein on the E. coli outer membrane was described in 1992 148, when Francisco and co-workers fused β-lactamase to the Lpp-OmpA hybrid protein for when Francisco and co-workers fused β-lactamase to the Lpp-OmpA hybrid protein for secretion and insertion into the outer membrane. The next year, the same group could display secretion and insertion into the outer membrane. The next year, the same group could display a scFv antibody fragment on E. coli and enrich scFv producing cells from a 105-fold excess of a scFv antibody fragment on E. coli and enrich scFv producing cells from a 105-fold excess of control cell background by means of FACS 25. Peptide libraries were inserted between control cell background by means of FACS 25. Peptide libraries were inserted between surface-exposed sites of E. coli outer membrane protein LambB and flagellum protein FliC surface-exposed sites of E. coli outer membrane protein LambB and flagellum protein FliC and, similarily to the biopanning procedure in phage display, selected by sequential binding and, similarily to the biopanning procedure in phage display, selected by sequential binding and elution 149; 150. The display of a library of scFv for affinity maturation was reported in and elution 149; 150. The display of a library of scFv for affinity maturation was reported in 1998, and a clone with an approximately three-fold better affinity could be isolated using flow 1998, and a clone with an approximately three-fold better affinity could be isolated using flow cytometry 151. Various display scaffolds involving different fusion partners have been cytometry 151. Various display scaffolds involving different fusion partners have been described, including outer membrane proteins such as OmpA, OmpX, LambB, fimbrial described, including outer membrane proteins such as OmpA, OmpX, LambB, fimbrial structure proteins such as FliC or FimH, as well as autotransporters such as AIDA-I, the IgA structure proteins such as FliC or FimH, as well as autotransporters such as AIDA-I, the IgA protease from Neisseria gonorrhoeae or the protease from enterohemorrhagic E. coli protease from Neisseria gonorrhoeae or the protease from enterohemorrhagic E. coli O157:H7 EspP. Depending on the scaffold, foreign peptides or proteins can be attached to the O157:H7 EspP. Depending on the scaffold, foreign peptides or proteins can be attached to the N-terminus, to the C-terminus or inserted in loop regions. In a commercial system denoted N-terminus, to the C-terminus or inserted in loop regions. In a commercial system denoted “FliTrx”, peptides are inserted in the active site loop of thioredoxin, which in term is inserted “FliTrx”, peptides are inserted in the active site loop of thioredoxin, which in term is inserted in a solvent exposed domain of the repeating flagellum protein FliC 152; 153. A non-covalent in a solvent exposed domain of the repeating flagellum protein FliC 152; 153. A non-covalent format for the display of full-size non-glycosylated antibodies on E. coli was described by format for the display of full-size non-glycosylated antibodies on E. coli was described by Mazor and co-workers. Here, antibodies are secreted into the periplasm and captured on the Mazor and co-workers. Here, antibodies are secreted into the periplasm and captured on the periplasmic face of the inner membrane by a membrane-associated Ig-binding protein. The periplasmic face of the inner membrane by a membrane-associated Ig-binding protein. The outer membrane is permeabilized and spheroblasts decorated with antibodies can be labeled outer membrane is permeabilized and spheroblasts decorated with antibodies can be labeled with fluorescent antigen and sorted on a flow cytometer. The spheroblasts provide only a with fluorescent antigen and sorted on a flow cytometer. The spheroblasts provide only a contemporary linkage of genotype an phenotype and are not viable after FACS sorting. The contemporary linkage of genotype an phenotype and are not viable after FACS sorting. The selected gene pool is transformed into fresh cells prior to the next sorting cycle 154. selected gene pool is transformed into fresh cells prior to the next sorting cycle 154. The possibility to employ FACS for the identification of library members with The possibility to employ FACS for the identification of library members with improved traits may be an advantage of cell surface display over other display systems. improved traits may be an advantage of cell surface display over other display systems. Between 5x104-105 copies of library members can be displayed on a single E. coli cell 25 and, Between 5x104-105 copies of library members can be displayed on a single E. coli cell 25 and, when incubating the cell with a fluorescently labeled antigen, flow cytometry allows for the when incubating the cell with a fluorescently labeled antigen, flow cytometry allows for the detection of fluorescence signals that correlate with the affinity and number of the displayed detection of fluorescence signals that correlate with the affinity and number of the displayed Sebastian Grimm 31 Sebastian Grimm 31 library members on a cell. Using multi-color FACS systems, the cells can be labeled with a library members on a cell. Using multi-color FACS systems, the cells can be labeled with a second fluorescent probe of another emission spectrum, that recognizes a constant part of the second fluorescent probe of another emission spectrum, that recognizes a constant part of the displayed library members. The dual-labeling allows for normalization of the fluorescence displayed library members. The dual-labeling allows for normalization of the fluorescence signal to the number of displayed library members, and thus fine affinity discrimination using signal to the number of displayed library members, and thus fine affinity discrimination using FACS sorting. A representative aliquot of the library can be analyzed to adjust an optimal FACS sorting. A representative aliquot of the library can be analyzed to adjust an optimal sorting gate in beforehand. Georgiou and co-workers observed that the mean fluorescence sorting gate in beforehand. Georgiou and co-workers observed that the mean fluorescence intensity resulting from different displayed scFv mutants binding to an antigen varied exactly intensity resulting from different displayed scFv mutants binding to an antigen varied exactly as expected from the differences in antigen affinity in solution 155. as expected from the differences in antigen affinity in solution 155.

Hence, cell surface display allows for the affinity determination on cell which Hence, cell surface display allows for the affinity determination on cell which facilitates the characterization of selected clones. Enrichment factors as high as 106 per cycle facilitates the characterization of selected clones. Enrichment factors as high as 106 per cycle were reported using cell surface display in combination with FACS sorting 25. The selection were reported using cell surface display in combination with FACS sorting 25. The selection stringency can be adjusted as described in the phage display section but also with the flow- stringency can be adjusted as described in the phage display section but also with the flow- cytometric sorting settings. No elution step is required to recover selected clones. There is cytometric sorting settings. No elution step is required to recover selected clones. There is however a general limitation of cell surface display and FACS, speed. A state of the art FACS however a general limitation of cell surface display and FACS, speed. A state of the art FACS can sort up to 70000 cells per second, which practically limits the sortable library size to can sort up to 70000 cells per second, which practically limits the sortable library size to about 108 when sorting for four hours and oversampling the library ten times. Hence, a about 108 when sorting for four hours and oversampling the library ten times. Hence, a biopanning step to enrich for binders and reduce the overall library size needs to be included biopanning step to enrich for binders and reduce the overall library size needs to be included when surpassing these numbers. Another concern when displaying multiple copies of library when surpassing these numbers. Another concern when displaying multiple copies of library members on the same cell in combination with an oligomeric antigen are avidity effects and members on the same cell in combination with an oligomeric antigen are avidity effects and hence undesired biases that cannot be accounted for during FACS. Outer membrane protein hence undesired biases that cannot be accounted for during FACS. Outer membrane protein scaffolds, such as the E. coli LamB, may diffuse laterally along the membrane and come in scaffolds, such as the E. coli LamB, may diffuse laterally along the membrane and come in close proximity, i.e. providing multiple binding sites to the antigen 153; 156. A particular close proximity, i.e. providing multiple binding sites to the antigen 153; 156. A particular concern with E. coli as host for display is its viability during FACS sorting 152. In comparison concern with E. coli as host for display is its viability during FACS sorting 152. In comparison to yeast or gram(+) bacteria, E. coli does not contain a thick layer of proteoglycan that makes to yeast or gram(+) bacteria, E. coli does not contain a thick layer of proteoglycan that makes it tolerable to high shear forces. it tolerable to high shear forces. Yeast surface display was pioneered by Wittrup and co-workers, who reported the Yeast surface display was pioneered by Wittrup and co-workers, who reported the selection of a scFv with decreased antigen dissociation rate from a library displayed on S. selection of a scFv with decreased antigen dissociation rate from a library displayed on S. cerevisiae in 1997 157. The library was C-terminally fused to the Aga2p subunit of the mating cerevisiae in 1997 157. The library was C-terminally fused to the Aga2p subunit of the mating adhesion receptor, that allows for anchoring to a-agglutinin. The inclusion of a 10-residue c- adhesion receptor, that allows for anchoring to a-agglutinin. The inclusion of a 10-residue c- myc tag allows for normalization of expression levels 157. Per definition, the mating adhesion myc tag allows for normalization of expression levels 157. Per definition, the mating adhesion receptor needs to be accessible on the cell surface for haploid yeast cells to find their mating receptor needs to be accessible on the cell surface for haploid yeast cells to find their mating partner, which makes it very suitable for the display of foreign proteins. Yeast possesses a partner, which makes it very suitable for the display of foreign proteins. Yeast possesses a secretory apparatus that includes chaperones, oxidative protein folding and, unlike in E. coli, secretory apparatus that includes chaperones, oxidative protein folding and, unlike in E. coli, N-linked glycosylation and thus makes it a host applicable for the display of secreted N-linked glycosylation and thus makes it a host applicable for the display of secreted eukaryotic proteins such as antibody fragments, growth factors or complex cell surface eukaryotic proteins such as antibody fragments, growth factors or complex cell surface receptors 55. In a direct comparison to phage display, using the same library towards the same receptors 55. In a direct comparison to phage display, using the same library towards the same antigen, yeast display was found to yield more high affinity clones and also required less antigen, yeast display was found to yield more high affinity clones and also required less effort 158. Since its introduction in 1997, yeast display has become popular and its potential effort 158. Since its introduction in 1997, yeast display has become popular and its potential 32 Ribosome display for selection and evolution of affibody molecules 32 Ribosome display for selection and evolution of affibody molecules has been demonstrated in a number of publications. The affinity of a scFv for fluorescein- has been demonstrated in a number of publications. The affinity of a scFv for fluorescein- biotin could be improved over 1000-fold to as low as 48 fM 159. Marks and co-workers could biotin could be improved over 1000-fold to as low as 48 fM 159. Marks and co-workers could engineer the selectivity of a botulinum neurotoxin binder to recognize both type A1 and type engineer the selectivity of a botulinum neurotoxin binder to recognize both type A1 and type A2, employing a dual-selection strategy 160. The display level of proteins fused to Aga2p was A2, employing a dual-selection strategy 160. The display level of proteins fused to Aga2p was shown to correlate with their thermal stability and soluble expression levels 161 and several shown to correlate with their thermal stability and soluble expression levels 161 and several groups could select recombinant proteins such as MHCI or the EGFR ectodomain with groups could select recombinant proteins such as MHCI or the EGFR ectodomain with engineered stability and soluble expression from randomly mutated libraries displayed on engineered stability and soluble expression from randomly mutated libraries displayed on yeast 162; 163. In an interesting study by Kranz and colleagues, the different sedimentation yeast 162; 163. In an interesting study by Kranz and colleagues, the different sedimentation characteristics of yeast and mammalian cells in a density medium were exploited for the characteristics of yeast and mammalian cells in a density medium were exploited for the isolation of TCR mutants specific for MHCI or MHCII. Upon centrifugation, yeast cells isolation of TCR mutants specific for MHCI or MHCII. Upon centrifugation, yeast cells binding to mammalian cells are selectively retained above the high density medium 164. binding to mammalian cells are selectively retained above the high density medium 164. Another proposed host for display of foreign proteins is the gram(+) Staphylococcus Another proposed host for display of foreign proteins is the gram(+) Staphylococcus carnosus, a bacterial strain classified as food-grade and widely used for meat fermentation carnosus, a bacterial strain classified as food-grade and widely used for meat fermentation processes, such as the ripening of sausages 165. It is encapsulated by a single cell membrane processes, such as the ripening of sausages 165. It is encapsulated by a single cell membrane and a thick murein sacculus, which makes it resistant to high shear forces as those in the and a thick murein sacculus, which makes it resistant to high shear forces as those in the FACS nozzle, but reduces the efficiency of plasmid uptake as compared to E. coli. A vector FACS nozzle, but reduces the efficiency of plasmid uptake as compared to E. coli. A vector for display of recombinant proteins fused to a propeptide, albumin binding protein and for display of recombinant proteins fused to a propeptide, albumin binding protein and membrane anchor motif was described by Samuelson and co-workers in 1995 166. The membrane anchor motif was described by Samuelson and co-workers in 1995 166. The significant increase in transformation efficiency from about 102 to 2.5x105 transformants per significant increase in transformation efficiency from about 102 to 2.5x105 transformants per µg of plasmid DNA described in 2007 167 allowed for the display and screening of the first µg of plasmid DNA described in 2007 167 allowed for the display and screening of the first libraries, such as an affibody molecule library pre-enriched for TNF-α-binding during one libraries, such as an affibody molecule library pre-enriched for TNF-α-binding during one preceding cycle of phage display 168. The selection of affibody molecules with sub-nanomolar preceding cycle of phage display 168. The selection of affibody molecules with sub-nanomolar affinities for Her3 from a 2nd generation library was described recently 167 and first selections affinities for Her3 from a 2nd generation library was described recently 167 and first selections from a bigger naïve affibody library are in preparation (unpublished, personal communication from a bigger naïve affibody library are in preparation (unpublished, personal communication John Löfblom). Another promising approach is the display of peptide libraries on S. carnosus John Löfblom). Another promising approach is the display of peptide libraries on S. carnosus for the mapping of antibody epitopes, that allows for the identification of continuous and, to for the mapping of antibody epitopes, that allows for the identification of continuous and, to some extent, non-continuous epitopes 169. Taken together, cell display systems have the main some extent, non-continuous epitopes 169. Taken together, cell display systems have the main advantage in allowing for a monitoring of the phenotype to select for using flow cytometry. advantage in allowing for a monitoring of the phenotype to select for using flow cytometry. Most work was done on yeast and E. coli, while S. carnosus may become a serious Most work was done on yeast and E. coli, while S. carnosus may become a serious alternative. alternative.

2.4 Ribosome display 2.4 Ribosome display

As illustrated in the previous chapters, both phage and cell display systems rely on living cells As illustrated in the previous chapters, both phage and cell display systems rely on living cells to mediate the linkage between genotype and phenotype, either directly or indirectly for the to mediate the linkage between genotype and phenotype, either directly or indirectly for the production of phages. This dependency is one possible limitation of the library size, since production of phages. This dependency is one possible limitation of the library size, since DNA needs to be transformed into cells. Furthermore, a “true” directed evolution experiment DNA needs to be transformed into cells. Furthermore, a “true” directed evolution experiment Sebastian Grimm 33 Sebastian Grimm 33 involves repetitive cycles of diversification and selection. In case of cell dependent selection involves repetitive cycles of diversification and selection. In case of cell dependent selection systems, the introduction of new diversity in between selection cycles by means of error- systems, the introduction of new diversity in between selection cycles by means of error- prone PCR or DNA shuffling (see section 1.4.2) requires a cumbersome switch between in prone PCR or DNA shuffling (see section 1.4.2) requires a cumbersome switch between in vivo selection and in vitro diversification, or the use of particular bacterial mutator strains vivo selection and in vitro diversification, or the use of particular bacterial mutator strains such as E. coli mutD5 170. A selection system that allows for selections to be entirely such as E. coli mutD5 170. A selection system that allows for selections to be entirely performed in vitro was initially described by Mattheakis in 1994 and further tested and performed in vitro was initially described by Mattheakis in 1994 and further tested and developed by Plückthun and co-workers 26; 29. In principle, a library containing DNA- developed by Plückthun and co-workers 26; 29. In principle, a library containing DNA- constructs devoid of STOP-codons is in vitro transcribed to yield library mRNA, which is in constructs devoid of STOP-codons is in vitro transcribed to yield library mRNA, which is in vitro translated in a cell extract derived from typically E. coli, wheat germ or rabbit vitro translated in a cell extract derived from typically E. coli, wheat germ or rabbit reticulocytes. Transcription and translation can either be performed in two consecutive reticulocytes. Transcription and translation can either be performed in two consecutive reactions (uncoupled) or simultaneously (coupled). Due to the absence of a STOP-codon and reactions (uncoupled) or simultaneously (coupled). Due to the absence of a STOP-codon and the addition of Mg2+ ions, mRNA (genotype), ribosome and the newly synthesized the addition of Mg2+ ions, mRNA (genotype), ribosome and the newly synthesized polypeptide (phenotype) stay physically attached in a ternary complex. These complexes enter polypeptide (phenotype) stay physically attached in a ternary complex. These complexes enter a selection cycle that is performed in principle as described in the phage display section, with a selection cycle that is performed in principle as described in the phage display section, with some distinctions. After each selection cycle, mRNA molecules of captured ternary some distinctions. After each selection cycle, mRNA molecules of captured ternary complexes are reverse transcribed and amplified in a PCR reaction to generate template for complexes are reverse transcribed and amplified in a PCR reaction to generate template for the next selection cycle (Figure 8). the next selection cycle (Figure 8).

Figure 8: Illustration of a typical ribosome display selection cycle. A library of linear DNA Figure 8: Illustration of a typical ribosome display selection cycle. A library of linear DNA fragments devoid of STOP-codons is in vitro transcribed to yield library mRNA which on its part is in fragments devoid of STOP-codons is in vitro transcribed to yield library mRNA which on its part is in vitro translated in a cell extract. The absence of a STOP-codon and addition of Mg2+-ions stalls vitro translated in a cell extract. The absence of a STOP-codon and addition of Mg2+-ions stalls ribosomes on their cognate mRNA and yields ternary complexes of mRNA, ribosome and displayed ribosomes on their cognate mRNA and yields ternary complexes of mRNA, ribosome and displayed protein. These complexes are then exposed to an immobilized antigen, non-binding complexes are protein. These complexes are then exposed to an immobilized antigen, non-binding complexes are washed away and the mRNA of binding complexes is eluted and isolated upon addition of the Mg2+- washed away and the mRNA of binding complexes is eluted and isolated upon addition of the Mg2+- chelator EDTA. Isolated mRNA is reverse transcribed and PCR-amplified for an additional selection chelator EDTA. Isolated mRNA is reverse transcribed and PCR-amplified for an additional selection cycle or post-selection analysis. cycle or post-selection analysis. 34 Ribosome display for selection and evolution of affibody molecules 34 Ribosome display for selection and evolution of affibody molecules

A ribosome display construct designed for use with E. coli extracts typically consists of a T7 A ribosome display construct designed for use with E. coli extracts typically consists of a T7 promoter that allows for mRNA synthesis, followed by a ribosome binding site (RBS) that promoter that allows for mRNA synthesis, followed by a ribosome binding site (RBS) that can base-pair with ribosomal RNA, thus recruiting the ribosome to the downstream start can base-pair with ribosomal RNA, thus recruiting the ribosome to the downstream start codon where protein synthesis is initiated. Depending on the origin of the cell free extract, a codon where protein synthesis is initiated. Depending on the origin of the cell free extract, a Shine-Dalgarno (prokaryote) or Kozak (eukaryote) sequence is used to initiate translation. Shine-Dalgarno (prokaryote) or Kozak (eukaryote) sequence is used to initiate translation.

The open reading frame frequently starts with a protein detection tag such as the RGS-His6 The open reading frame frequently starts with a protein detection tag such as the RGS-His6 tag or the FLAG tag, followed by the library of binding proteins and a spacer protein. The tag or the FLAG tag, followed by the library of binding proteins and a spacer protein. The spacer should provide the displayed library the flexibility to fold outside of the ribosome spacer should provide the displayed library the flexibility to fold outside of the ribosome tunnel. At least 23-30 amino acids of spacer are required for the displayed protein to exit tunnel. At least 23-30 amino acids of spacer are required for the displayed protein to exit completely from the tunnel 171. Mattheakis initially used a 72 aa spacer for the display of completely from the tunnel 171. Mattheakis initially used a 72 aa spacer for the display of peptides, while 116 amino acids of spacer were reported to be more efficient in displaying peptides, while 116 amino acids of spacer were reported to be more efficient in displaying proteins than shorter spacer sequences 172. Protein fragments of different origins have been proteins than shorter spacer sequences 172. Protein fragments of different origins have been applied as spacer, such as a glycine and serine rich region of M13 phage pIII, the Cκ domain applied as spacer, such as a glycine and serine rich region of M13 phage pIII, the Cκ domain of a light chain of an antibody 173, a proline-rich sequence of E. coli TonB, a helical segment of a light chain of an antibody 173, a proline-rich sequence of E. coli TonB, a helical segment of E. coli TolA 172 or λ phage protein D 174. A schematic illustration of a typical ribosome of E. coli TolA 172 or λ phage protein D 174. A schematic illustration of a typical ribosome display construct is given in the present investigations section, (I). display construct is given in the present investigations section, (I). Recently, a 17-mer peptide sequence derived from the E. coli secretion monitor Recently, a 17-mer peptide sequence derived from the E. coli secretion monitor protein SecM was found to cause elongation arrest during peptide synthesis in the ribosome protein SecM was found to cause elongation arrest during peptide synthesis in the ribosome tunnel 175 and was used to prepare homogeneous stalled ribosome nascent chain complexes in tunnel 175 and was used to prepare homogeneous stalled ribosome nascent chain complexes in vivo and in vitro 176. In three publications of Ohashi, Matsuura or Osada 174; 177; 178, the SecM vivo and in vitro 176. In three publications of Ohashi, Matsuura or Osada 174; 177; 178, the SecM elongation arrest sequence was placed downstream of the spacer sequence and thereby elongation arrest sequence was placed downstream of the spacer sequence and thereby possibly further stabilized ternary complexes. However, no comparison with classical spacer possibly further stabilized ternary complexes. However, no comparison with classical spacer constructs has been performed so far. constructs has been performed so far. The application of a cell extract for in vitro protein synthesis was reported as early as The application of a cell extract for in vitro protein synthesis was reported as early as 1961, long before the era of biotechnology, when Nirenberg and Matthaei solved the genetic 1961, long before the era of biotechnology, when Nirenberg and Matthaei solved the genetic code by translating synthetically produced RNA molecules (including poly-uracil) in an E. code by translating synthetically produced RNA molecules (including poly-uracil) in an E. coli extract 179. Mattheakis originally used a commercially available S30 extract prepared coli extract 179. Mattheakis originally used a commercially available S30 extract prepared from E. coli strain SL119 26. Plückthun and co-workers routinely use an in-house prepared from E. coli strain SL119 26. Plückthun and co-workers routinely use an in-house prepared extract based on the RNaseI deficient strain E. coli MRE600 in coupled or uncoupled extract based on the RNaseI deficient strain E. coli MRE600 in coupled or uncoupled reactions 172. Exo- and endocucleases originating from such natural extracts that rapidly reactions 172. Exo- and endocucleases originating from such natural extracts that rapidly degrade library mRNA during the in vitro translation procedure at 37° C are a general concern degrade library mRNA during the in vitro translation procedure at 37° C are a general concern during ribosome display, and ternary complexes need to be kept cold in the presence of during ribosome display, and ternary complexes need to be kept cold in the presence of RNase inhibitor to avoid degradation of the mRNA. The introduction of 5´- and 3´-stem loops RNase inhibitor to avoid degradation of the mRNA. The introduction of 5´- and 3´-stem loops in constructs to protect the library mRNA from exonucleases has been reported 29. in constructs to protect the library mRNA from exonucleases has been reported 29. Furthermore, E. coli possesses an SsrA RNA-mediated mechanism to rescue stalled Furthermore, E. coli possesses an SsrA RNA-mediated mechanism to rescue stalled ribosomes and tag the translated protein for proteolytic degradation. To neutralize this ribosomes and tag the translated protein for proteolytic degradation. To neutralize this reaction, an oligonucleotide complementary to SsrA RNA can be added to the E. coli extract. reaction, an oligonucleotide complementary to SsrA RNA can be added to the E. coli extract. In general, relatively short times are used for translation reactions, typically between 6 and 20 In general, relatively short times are used for translation reactions, typically between 6 and 20 Sebastian Grimm 35 Sebastian Grimm 35 minutes in an uncoupled reaction in E. coli extract 172. Taussig and coworkers have described minutes in an uncoupled reaction in E. coli extract 172. Taussig and coworkers have described the use of the commercially available rabbit reticulocyte lysates in a coupled reaction. While the use of the commercially available rabbit reticulocyte lysates in a coupled reaction. While in prokaryotic systems, ribosomes are efficiently dissociated and mRNA released upon in prokaryotic systems, ribosomes are efficiently dissociated and mRNA released upon addition of EDTA, rabbit reticulocyte lysate-based systems require the in situ recovery, and addition of EDTA, rabbit reticulocyte lysate-based systems require the in situ recovery, and an elegant approach involving a sensitive single-primer PCR was described 180. In a an elegant approach involving a sensitive single-primer PCR was described 180. In a comparison of the E. coli and rabbit reticulocyte ribosome display systems, Hanes and comparison of the E. coli and rabbit reticulocyte ribosome display systems, Hanes and coworkers could not detect any advantage of using the eukaryotic system when translating a coworkers could not detect any advantage of using the eukaryotic system when translating a fluorescein binding scFv. In this comparison, however, eukaryote ribosomes were eluted with fluorescein binding scFv. In this comparison, however, eukaryote ribosomes were eluted with PBS containing 20 mM EDTA 181. Yet another attractive alternative is an E. coli-based PBS containing 20 mM EDTA 181. Yet another attractive alternative is an E. coli-based system reconstituted from recombinantly produced and purified components and E. coli system reconstituted from recombinantly produced and purified components and E. coli ribosomes 182. Among the purified components are initiation, elongation and release factors, ribosomes 182. Among the purified components are initiation, elongation and release factors, aminoacyl-tRNA synthetases, T7 RNA polymerase and tRNAs. This extract is virtually aminoacyl-tRNA synthetases, T7 RNA polymerase and tRNAs. This extract is virtually devoid of RNase (with the ribosome preparation as the only source) and was shown to give devoid of RNase (with the ribosome preparation as the only source) and was shown to give higher post-selection yields of recovered mRNA as compared to E. coli S30 extract during higher post-selection yields of recovered mRNA as compared to E. coli S30 extract during ribosome display selection 183. Depending on the degree of mRNA degradation, it may be ribosome display selection 183. Depending on the degree of mRNA degradation, it may be necessary to “rescue” selected clones by RT-PCR amplification of the library gene element necessary to “rescue” selected clones by RT-PCR amplification of the library gene element only, followed by ligation into a circular construct that is used as template for a second PCR only, followed by ligation into a circular construct that is used as template for a second PCR reaction to yield functional and full library gene cassettes for the next selection cycle 184. reaction to yield functional and full library gene cassettes for the next selection cycle 184. The openness of ribosome display allows for the introduction of additional diversity in The openness of ribosome display allows for the introduction of additional diversity in between selection cycles using error-prone PCR, or by gene shuffling (see section 1.4.2). This between selection cycles using error-prone PCR, or by gene shuffling (see section 1.4.2). This feature makes ribosome display a dynamic selection system that can mimic the natural feature makes ribosome display a dynamic selection system that can mimic the natural affinity maturation process of antibodies, where somatic mutations introduce additional affinity maturation process of antibodies, where somatic mutations introduce additional diversity during the development of B-cells. In one study, Hanes and coworkers reported the diversity during the development of B-cells. In one study, Hanes and coworkers reported the selection and evolution of scFv antibody fragments with equilibrium dissociation constants as selection and evolution of scFv antibody fragments with equilibrium dissociation constants as low as 82 pM, when combining ribosome display with error-prone PCR. The introduced point low as 82 pM, when combining ribosome display with error-prone PCR. The introduced point mutations accounted for an up to 40-fold increase in affinity as compared to the progenitor mutations accounted for an up to 40-fold increase in affinity as compared to the progenitor clones 28. In another study, the combination of error-prone PCR and gene shuffling in clones 28. In another study, the combination of error-prone PCR and gene shuffling in subsequent cycles yielded an increased population of affinity-improved variants, and the subsequent cycles yielded an increased population of affinity-improved variants, and the highest affinity clone, as compared to error-prone PCR only 185. highest affinity clone, as compared to error-prone PCR only 185.

2.5 mRNA display 2.5 mRNA display

An ingenious and elegant display system named mRNA display was first described by An ingenious and elegant display system named mRNA display was first described by Roberts and Szostak in 1997 27. Similarly to ribosome display, a DNA library is in vitro Roberts and Szostak in 1997 27. Similarly to ribosome display, a DNA library is in vitro transcribed to yield library mRNA. Now, in mRNA display, the mRNA library template is transcribed to yield library mRNA. Now, in mRNA display, the mRNA library template is covalently attached to a linker of DNA or polyethylenglycol containing the translation- covalently attached to a linker of DNA or polyethylenglycol containing the translation- terminator puromycin. During in vitro translation, the ribosome will reach the 3´-end of the terminator puromycin. During in vitro translation, the ribosome will reach the 3´-end of the 36 Ribosome display for selection and evolution of affibody molecules 36 Ribosome display for selection and evolution of affibody molecules mRNA and puromycin will enter its A-site and get covalently linked to the nascent mRNA and puromycin will enter its A-site and get covalently linked to the nascent polypeptide chain. This trick allows for a direct covalent linkage of genotype and phenotype polypeptide chain. This trick allows for a direct covalent linkage of genotype and phenotype in a 1:1 stoichiometry. The RNA is typically reverse transcribed for stabilization and the in a 1:1 stoichiometry. The RNA is typically reverse transcribed for stabilization and the complexes are selected in a biopanning procedure. Library sizes as high as 1013 have been complexes are selected in a biopanning procedure. Library sizes as high as 1013 have been reported, that allowed for the selection of new binding proteins from a completely random reported, that allowed for the selection of new binding proteins from a completely random peptide library 33 or the selection of an enzyme from a non-catalytic library scaffold 30. The peptide library 33 or the selection of an enzyme from a non-catalytic library scaffold 30. The latter study involved the covalent coupling of the genotype to its dedicated substrate. The latter study involved the covalent coupling of the genotype to its dedicated substrate. The display of a proteome library on mRNA for biopanning against calmodulin and selection of display of a proteome library on mRNA for biopanning against calmodulin and selection of calmodulin binding proteins was reported in 2004 186. Together, these studies demonstrate the calmodulin binding proteins was reported in 2004 186. Together, these studies demonstrate the potential of mRNA display for sampling an exceptionally large protein sequence space and potential of mRNA display for sampling an exceptionally large protein sequence space and suggest that proteins with new functions or new folds that were not created during natural suggest that proteins with new functions or new folds that were not created during natural evolution can be isolated. evolution can be isolated.

2.6 Other display systems 2.6 Other display systems

The above described display systems are most common and widely used. There are however The above described display systems are most common and widely used. There are however also some more exotic systems that are elegant from an engineering perspective and worth also some more exotic systems that are elegant from an engineering perspective and worth looking at. One such system is CIS display. Here, gene cassettes consisting of a protein looking at. One such system is CIS display. Here, gene cassettes consisting of a protein library fused to a gene encoding the replication initiator protein RepA are in vitro transcribed library fused to a gene encoding the replication initiator protein RepA are in vitro transcribed and translated. RepA is a cis-acting DNA-binding protein that will bind to its cognate DNA and translated. RepA is a cis-acting DNA-binding protein that will bind to its cognate DNA and thus establish a genotype-phenotype pair. A prerequisite to avoid hybrid pairings is that and thus establish a genotype-phenotype pair. A prerequisite to avoid hybrid pairings is that RepA exclusively binds to the template DNA from which it was expressed. The DNA-protein RepA exclusively binds to the template DNA from which it was expressed. The DNA-protein pairs can be used for biopanning procedures and the functionality of the system was shown pairs can be used for biopanning procedures and the functionality of the system was shown when selecting peptide binders against different protein targets 187. Another such system is when selecting peptide binders against different protein targets 187. Another such system is plasmid display. Fusion proteins of a library and a DNA-binding protein such as the lac plasmid display. Fusion proteins of a library and a DNA-binding protein such as the lac repressor 188 or NF-κB homodimer 189 are expressed in cells to allow the fusion proteins to repressor 188 or NF-κB homodimer 189 are expressed in cells to allow the fusion proteins to bind to its cognate plasmid template. Cells are lysed and, as in CIS-display, plasmid-protein bind to its cognate plasmid template. Cells are lysed and, as in CIS-display, plasmid-protein pairs are panned against given protein targets. Selected plasmids need to be transformed into pairs are panned against given protein targets. Selected plasmids need to be transformed into fresh cells prior to the next selection cycle. With this approach, Schatz could identify a fresh cells prior to the next selection cycle. With this approach, Schatz could identify a peptide sequence motif that is biotinylated in vivo by the E. coli biotin holoenzyme synthetase peptide sequence motif that is biotinylated in vivo by the E. coli biotin holoenzyme synthetase BirA by panning the in vivo biotinylated peptide library against streptavidin-coated beads 188. BirA by panning the in vivo biotinylated peptide library against streptavidin-coated beads 188. A third system exploits the ability of bacteriophage P2 endonuclease P2A to form a covalent A third system exploits the ability of bacteriophage P2 endonuclease P2A to form a covalent bond with its own DNA phosphoester backbone. Hence, a coupled in vitro transcription and bond with its own DNA phosphoester backbone. Hence, a coupled in vitro transcription and translation of libraries fused to P2A yields covalently linked pairs of protein library and translation of libraries fused to P2A yields covalently linked pairs of protein library and cognate DNA that were successfully used to select anti-tetanus toxoid scFvs from a library cognate DNA that were successfully used to select anti-tetanus toxoid scFvs from a library prepared from human lymphocytes 190. prepared from human lymphocytes 190. Sebastian Grimm 37 Sebastian Grimm 37

2.7 Compartment systems 2.7 Compartment systems

All hitherto discussed selection systems rely on a physical linkage of genotype and All hitherto discussed selection systems rely on a physical linkage of genotype and phenotype. This physical linkage is however not mandatory if genotypes and phenotypes are phenotype. This physical linkage is however not mandatory if genotypes and phenotypes are spatially confined by a membrane, cell wall or barrier of other kind. Just as in nature, where spatially confined by a membrane, cell wall or barrier of other kind. Just as in nature, where the genes and their products are compartmentalized in individual cells. One approach to the genes and their products are compartmentalized in individual cells. One approach to artificially create such compartments is the dispersion of an aqueous solution in oil, yielding artificially create such compartments is the dispersion of an aqueous solution in oil, yielding droplets as small as ≥ 1 µm. A volume of 1 ml would accommodate at least 1010 droplets, to droplets as small as ≥ 1 µm. A volume of 1 ml would accommodate at least 1010 droplets, to get an idea of the possible diversity. In an elegant study of Tawfik and Griffiths, DNA get an idea of the possible diversity. In an elegant study of Tawfik and Griffiths, DNA molecules encoding either a DNA methyltransferase (M.HaeIII) or a control enzyme were molecules encoding either a DNA methyltransferase (M.HaeIII) or a control enzyme were mixed with a cell extract that allowed for transcription and translation of the encoded mixed with a cell extract that allowed for transcription and translation of the encoded enzymes. The mixture was dispersed in oil and the encoded enzymes were transcribed and enzymes. The mixture was dispersed in oil and the encoded enzymes were transcribed and translated. Only DNA in droplets containing the M.HaeIII gene was methylated and thus translated. Only DNA in droplets containing the M.HaeIII gene was methylated and thus protected against a subsequent HaeIII digestion step, allowing for the selection for target- protected against a subsequent HaeIII digestion step, allowing for the selection for target- specific DNA methylation 191. This approach was later used for library selection of active, specific DNA methylation 191. This approach was later used for library selection of active, mutated methyltransferases 192 or to alter the sequence specificity of M.HaeIII 193. The mutated methyltransferases 192 or to alter the sequence specificity of M.HaeIII 193. The applications of water-in-oil emulsions are manifold and include emulsion PCR [78], covalent applications of water-in-oil emulsions are manifold and include emulsion PCR [78], covalent DNA display, the latter for example based on fusion proteins between library members and a DNA display, the latter for example based on fusion proteins between library members and a DNA-methyltransferase that can covalently bind to a 5-fluorodeoxycytidine base at the DNA-methyltransferase that can covalently bind to a 5-fluorodeoxycytidine base at the terminus of the DNA fragment 194 or cognate pairs of biotinylated library DNA and protein- terminus of the DNA fragment 194 or cognate pairs of biotinylated library DNA and protein- streptavidin fusions generated in an emulsion (streptavidin-biotin linkage in emulsions; streptavidin fusions generated in an emulsion (streptavidin-biotin linkage in emulsions; STABLE) 195. Today, uniform and monodisperse droplets of water-in-oil can be created at a STABLE) 195. Today, uniform and monodisperse droplets of water-in-oil can be created at a rate of 10.000 per second with microfluidic devices, that also allow the fusion, splitting, rate of 10.000 per second with microfluidic devices, that also allow the fusion, splitting, mixing and sorting of droplets 196. Such devices envision to greatly accelerate the screening mixing and sorting of droplets 196. Such devices envision to greatly accelerate the screening for protein traits in a droplet format. Another format are aqueous droplets coated with a layer for protein traits in a droplet format. Another format are aqueous droplets coated with a layer of oil in a water phase, denoted water-in-oil-in-water emulsion, and was reported by Bernath of oil in a water phase, denoted water-in-oil-in-water emulsion, and was reported by Bernath and colleagues 197. Such droplets can be sorted by a flow cytometer and the potential of this and colleagues 197. Such droplets can be sorted by a flow cytometer and the potential of this technique was demonstrated in the selection from enzyme libraries with fluorogenic technique was demonstrated in the selection from enzyme libraries with fluorogenic substrates 198. substrates 198. Another approach is to employ natural compartments, cells, for selection of protein Another approach is to employ natural compartments, cells, for selection of protein traits, as realized in the protein fragment complementation assay (PCA). Essentially, a traits, as realized in the protein fragment complementation assay (PCA). Essentially, a reporter protein is split in two parts that are fused to a “bait” protein and a “scavenger” reporter protein is split in two parts that are fused to a “bait” protein and a “scavenger” protein, respectively. Both are expressed within the same E. coli cell and only upon binding of protein, respectively. Both are expressed within the same E. coli cell and only upon binding of bait to scavenger, the two halves of the reporter protein are brought into proximity and bait to scavenger, the two halves of the reporter protein are brought into proximity and reconstitute its function, such as fluorescence (split GFP), antibiotic resistance (split β- reconstitute its function, such as fluorescence (split GFP), antibiotic resistance (split β- lactamase) or synthesis of nucleic acids and amino acids (split dihydrofolate reductase). The lactamase) or synthesis of nucleic acids and amino acids (split dihydrofolate reductase). The bait may be an antigen and the scavenger a library of binding proteins. Split GFP systems bait may be an antigen and the scavenger a library of binding proteins. Split GFP systems allow for affinity selection using FACS sorting 199, while split β-lactamase and dihydrofolate allow for affinity selection using FACS sorting 199, while split β-lactamase and dihydrofolate 38 Ribosome display for selection and evolution of affibody molecules 38 Ribosome display for selection and evolution of affibody molecules reductase allow for affinity selection using Darwinian survival of bacteria in the presence of reductase allow for affinity selection using Darwinian survival of bacteria in the presence of antibiotic or absence of an essential amino acid 200; 201. In principle, the combination of PCA antibiotic or absence of an essential amino acid 200; 201. In principle, the combination of PCA and Darwinian survival selection is appealing since it does neither require the purification of and Darwinian survival selection is appealing since it does neither require the purification of the antigen, nor repeating cycles of biopanning or FACS. the antigen, nor repeating cycles of biopanning or FACS. This chapter should have brought the reader to the state-of-the-art of a blend of This chapter should have brought the reader to the state-of-the-art of a blend of selection systems that all have their unique specifications, strengths and weaknesses. Notably, selection systems that all have their unique specifications, strengths and weaknesses. Notably, there is no perfect selection system and the choice mainly depends on the scientific context there is no perfect selection system and the choice mainly depends on the scientific context and the available resources such as a flow cytometer. The next chapter will be less and the available resources such as a flow cytometer. The next chapter will be less methodological but provide molecular insights in the carriers of phenotypic diversity. methodological but provide molecular insights in the carriers of phenotypic diversity.

Selection system Genotype linkage Cell dependency Selection system Genotype linkage Cell dependency Phage display physical yes Phage display physical yes Protein Trait Approach Achievement Library Ref. Protein Trait Approach Achievement Library Ref. scFv library displayed on scFv with K = 10 nM (i) scFv library displayed on scFv with K = 10 nM (i) scFv library Affinity d 129 scFv library Affinity d 129 fd phage for hapten fd phage for hapten Panning in the presence of Cross-reactivity Panning in the presence of Cross-reactivity Fab library Selectivity (r) 105 132 Fab library Selectivity (r) 105 132 steroid competitor decreased (1 to 0.03%) steroid competitor decreased (1 to 0.03%) Herceptin light chain CDR Dual selectivity for Herceptin light chain CDR Dual selectivity for Fab library Selectivity (r) 50 Fab library Selectivity (r) 50 re-randomization Her2 and VEGF re-randomization Her2 and VEGF Panning in the presence of Up to 4 kcal/mol Panning in the presence of Up to 4 kcal/mol scFv library, Stability (r) 107 68 scFv library, Stability (r) 107 68 GdnCl or elevated T improved stability GdnCl or elevated T improved stability Proteolysis of nuclease Selection of more stable Proteolysis of nuclease Selection of more stable Nuclease Stability (r) 104 131 Nuclease Stability (r) 104 131 displayed on fd phage nucleases displayed on fd phage nucleases Catalytic Panning against transition Candidate catalytic scFv Catalytic Panning against transition Candidate catalytic scFv scFv library (i) 147 scFv library (i) 147 activity state analogue selected activity state analogue selected Catalytic Enzyme, substrate on same Enrichment of product- Catalytic Enzyme, substrate on same Enrichment of product- Nuclease (m) 130 Nuclease (m) 130 activity M13 phage displaying phages activity M13 phage displaying phages Selection system Genotype linkage Cell dependency Selection system Genotype linkage Cell dependency Cell surface display physical yes Cell surface display physical yes Protein Trait Approach Achievement Library Ref. Protein Trait Approach Achievement Library Ref. Display on E. coli OmpA, 3-fold lower K for Display on E. coli OmpA, 3-fold lower K for scFv library Affinity d (r) 151 scFv library Affinity d (r) 151 maturation hapten maturation hapten Display of full-length IgG Kd = 21-440 nM for B. Display of full-length IgG Kd = 21-440 nM for B. IgG library Affinity (i) 154 IgG library Affinity (i) 154 on E. coli cell membrane anthracis antigen on E. coli cell membrane anthracis antigen Display on yeast Aga2p, 48 fM fluorescein (r) 105-7 Display on yeast Aga2p, 48 fM fluorescein (r) 105-7 scFv library Affinity 159 scFv library Affinity 159 competition, maturation binder evolved competition, maturation binder evolved

Display on yeast, look- Kd = 1.1 pM for TNF-α Display on yeast, look- Kd = 1.1 pM for TNF-α scFv library Affinity (r) 107 107 scFv library Affinity (r) 107 107 through mutagenesis (500-870 fold better) through mutagenesis (500-870 fold better) S. carnosus display of (r) S. carnosus display of (r) Affibody library Affinity nM affinity to ErB3 202 Affibody library Affinity nM affinity to ErB3 202 affinity maturation library affinity maturation library Display on yeast, FACS, K = 115/87 pM for Display on yeast, FACS, K = 115/87 pM for scFv library Selectivity d (r) 108 160 scFv library Selectivity d (r) 108 160 two sequential antigens botulinum toxin A1/A2 two sequential antigens botulinum toxin A1/A2 Yeast surface expression, MHCI with improved 162; Yeast surface expression, MHCI with improved 162; MHCI library Stability (r) 106 MHCI library Stability (r) 106 FACS for high expression stability and solubility 163 FACS for high expression stability and solubility 163 Sebastian Grimm 39 Sebastian Grimm 39

Selection system Genotype linkage Cell dependency Selection system Genotype linkage Cell dependency Ribosome display physical no Ribosome display physical no Protein Trait Approach Achievement Library Ref. Protein Trait Approach Achievement Library Ref. Peptide library Affinity Panning against antibody Antibody binders (r) 26 Peptide library Affinity Panning against antibody Antibody binders (r) 26 Panning against Enrichment of binders Panning against Enrichment of binders scFv library Affinity (m) 29 scFv library Affinity (m) 29 hemagglutinin over several cycles hemagglutinin over several cycles K = 82 pM for bovine K = 82 pM for bovine scFv library Affinity Directed evolution d (r) 109 28 scFv library Affinity Directed evolution d (r) 109 28 insulin insulin 174; 174; Identification of Ab (r) Identification of Ab (r) Peptide library Affinity Panning against Ab 177; Peptide library Affinity Panning against Ab 177; epitopes epitopes 178 178 L-shuffling during affinity 22-fold decreased K (r) L-shuffling during affinity 22-fold decreased K (r) scFv library Affinity d 185 scFv library Affinity d 185 maturation with L-shuffling maturation with L-shuffling

Selection system Genotype linkage Cell dependency Selection system Genotype linkage Cell dependency mRNA display physical no mRNA display physical no Protein Trait Approach Achievement Library Ref. Protein Trait Approach Achievement Library Ref. 18 selection rounds against Novel ATP-binding 18 selection rounds against Novel ATP-binding Random peptide Affinity (r) 1013 33 Random peptide Affinity (r) 1013 33 ATP proteins selected ATP proteins selected Panning on BIAcore High enrichment Panning on BIAcore High enrichment scFv library Affinity (n) 1012 123 scFv library Affinity (n) 1012 123 biosensor, epPCR factors, Kd = 0.7-3.8 nM biosensor, epPCR factors, Kd = 0.7-3.8 nM Panning of proteome Identification of novel Panning of proteome Identification of novel Human cDNA Affinity 1012 186 Human cDNA Affinity 1012 186 cDNA library Calmodulin binders cDNA library Calmodulin binders Retinoid-X-rec. Catalytic Enzymes ligate substrate to Novel RNA ligase Retinoid-X-rec. Catalytic Enzymes ligate substrate to Novel RNA ligase (r) 1012 30 (r) 1012 30 scaffold activity their cognate genes selected scaffold activity their cognate genes selected

Selection system Genotype linkage Cell dependency Selection system Genotype linkage Cell dependency Compartment spatial Yes/no Compartment spatial Yes/no

Protein Trait Approach Achievement Library Ref. Protein Trait Approach Achievement Library Ref. K = 14-27 nM for K = 14-27 nM for Affibody library Affinity TEM-1 β-lactamase PCA d (r) 109 200 Affibody library Affinity TEM-1 β-lactamase PCA d (r) 109 200 TNF-α TNF-α Fatty acid Mutagenesis, screening and Conversion into alkane Fatty acid Mutagenesis, screening and Conversion into alkane Selectivity (r) 74 Selectivity (r) 74 monooxygen. shuffling hydroxylase monooxygen. shuffling hydroxylase Enzyme activity in 60% Enzyme activity in 60% Subtilisin Stability Mutagenesis and screening (r) 75 Subtilisin Stability Mutagenesis and screening (r) 75 formamide formamide Cephalo- Catalytic 270-540 fold improved Cephalo- Catalytic 270-540 fold improved Gene shuffling (r) 120 Gene shuffling (r) 120 sporinase activity activity sporinase activity activity Catalytic Evolution of Ebg into β- Catalytic Evolution of Ebg into β- Ebg Double emulsions, FACS (r) 198 Ebg Double emulsions, FACS (r) 198 activity galactosidase activity galactosidase

Table 2: Combinatorial approaches in protein engineering. In the header lines, different protein Table 2: Combinatorial approaches in protein engineering. In the header lines, different protein selection systems and some of their major characteristics are denoted. For each selection system, a few selection systems and some of their major characteristics are denoted. For each selection system, a few examples of studies are listed that address different protein traits. The columns include the selected examples of studies are listed that address different protein traits. The columns include the selected protein, the addressed protein trait, a short description of the approach used for selection and a short protein, the addressed protein trait, a short description of the approach used for selection and a short summary of the achievement of the study, the library type, size and the corresponding reference. summary of the achievement of the study, the library type, size and the corresponding reference. Library types include: (i) immune, (m) model, (n) naïve and (r) randomized. Library types include: (i) immune, (m) model, (n) naïve and (r) randomized. 40 Ribosome display for selection and evolution of affibody molecules 40 Ribosome display for selection and evolution of affibody molecules

3 Protein scaffolds for molecular recognition 3 Protein scaffolds for molecular recognition 3.1 Introduction 3.1 Introduction

In order to explain the process of immunity, Ehrlich postulated in 1901 that “side chains” on In order to explain the process of immunity, Ehrlich postulated in 1901 that “side chains” on living cells can bind toxins involved in infections 203 and was later awarded the Nobel prize living cells can bind toxins involved in infections 203 and was later awarded the Nobel prize for his contributions to understanding immunology. Ehrlichs side chain theory gave rise to a for his contributions to understanding immunology. Ehrlichs side chain theory gave rise to a “therapia magna sterilisans” and the search for a magic bullet for the treatment of diseases. “therapia magna sterilisans” and the search for a magic bullet for the treatment of diseases. He envisioned an agent with high affinity to a disease-causing pathogen and an efficacy at He envisioned an agent with high affinity to a disease-causing pathogen and an efficacy at concentrations harmless for the patient 204. Today we know that nature’s solution for such concentrations harmless for the patient 204. Today we know that nature’s solution for such agents with high affinity are antibodies, proteins consisting of several polypeptide chains, agents with high affinity are antibodies, proteins consisting of several polypeptide chains, with the ability to bind a manifold of different antigens, from small haptens to large with the ability to bind a manifold of different antigens, from small haptens to large macromolecules. With the invention of recombinant DNA technology, engineered antibody macromolecules. With the invention of recombinant DNA technology, engineered antibody fragments were reported that are simpler in composition and can be produced at high yields in fragments were reported that are simpler in composition and can be produced at high yields in microbial hosts. The concept of variable protein surfaces for molecular recognition was then microbial hosts. The concept of variable protein surfaces for molecular recognition was then further expanded to non-antibody derived or alternative protein scaffolds. In principle, a further expanded to non-antibody derived or alternative protein scaffolds. In principle, a stable natural protein is chosen and serves as framework for the engineering of a novel stable natural protein is chosen and serves as framework for the engineering of a novel binding site into its surface. Such binding sites can be composed of loops, secondary structure binding site into its surface. Such binding sites can be composed of loops, secondary structure motifs such as α-helices and β-strands or combinations thereof. One of the first reported motifs such as α-helices and β-strands or combinations thereof. One of the first reported alternative protein scaffolds for molecular recognition was the affibody molecule 205, that this alternative protein scaffolds for molecular recognition was the affibody molecule 205, that this thesis is based upon. thesis is based upon.

3.2 Antibodies 3.2 Antibodies 3.2.1 Natural polyclonal or monoclonal antibodies 3.2.1 Natural polyclonal or monoclonal antibodies

Antibodies or immunoglobulins are an essential part of the adaptive immune system of higher Antibodies or immunoglobulins are an essential part of the adaptive immune system of higher vertebrates. In response to an immunogen such as a parasite, bacterium, virus, pollen or vertebrates. In response to an immunogen such as a parasite, bacterium, virus, pollen or cancer cell, antibodies are secreted into body fluids by specialized plasma B cells. The cancer cell, antibodies are secreted into body fluids by specialized plasma B cells. The antibodies natural function is to selectively bind to antigens such as proteins, nucleic acids, antibodies natural function is to selectively bind to antigens such as proteins, nucleic acids, lipids, carbohydrates or haptens present on the immunogen. Antibodies have evolved to lipids, carbohydrates or haptens present on the immunogen. Antibodies have evolved to protect the host by either a functional neutralization, a tagging for subsequent recognition by protect the host by either a functional neutralization, a tagging for subsequent recognition by other specialized immune cells, or for cell surface located antigens, the activation of a cascade other specialized immune cells, or for cell surface located antigens, the activation of a cascade of complement proteins resulting in pore formation and cell lysis. Immunoglobulins are, with of complement proteins resulting in pore formation and cell lysis. Immunoglobulins are, with some exceptions, composed of two identical heavy and two identical light polypeptide chains, some exceptions, composed of two identical heavy and two identical light polypeptide chains, that are interconnected with disulphide bonds. The light chains consist of one variable domain that are interconnected with disulphide bonds. The light chains consist of one variable domain

VL and one constant domain CL. The heavy chains consist of one variable domain VH and VL and one constant domain CL. The heavy chains consist of one variable domain VH and three to four constant domains CH1, CH2, CH3 and CH4. All these domains are of a typical three to four constant domains CH1, CH2, CH3 and CH4. All these domains are of a typical immunoglobulin domain fold of two stacked β-sheets that are stabilized with conserved immunoglobulin domain fold of two stacked β-sheets that are stabilized with conserved disulphide bonds. Igs are classified according to their heavy chain type (α/IgA, δ/IgD, ε/IgE, disulphide bonds. Igs are classified according to their heavy chain type (α/IgA, δ/IgD, ε/IgE, Sebastian Grimm 41 Sebastian Grimm 41

γ/IgG, µ/IgM) and light chain type (λ, κ). It should be noted that even antibodies of the very γ/IgG, µ/IgM) and light chain type (λ, κ). It should be noted that even antibodies of the very same class of heavy and light chain can be dissimilar in amino acid sequence in their constant same class of heavy and light chain can be dissimilar in amino acid sequence in their constant domains due to the existence of allelic variants. domains due to the existence of allelic variants.

Figure 9: Schematic illustration of an IgG molecule and a single-chain fragment variable (scFv) Figure 9: Schematic illustration of an IgG molecule and a single-chain fragment variable (scFv) IgG derivative. The IgG consists of variable domains (V) and constant domains (C) on heavy chain IgG derivative. The IgG consists of variable domains (V) and constant domains (C) on heavy chain (H) and light chain (L). All domains contain disulphide-bonds (S-S) and heavy and light chains are (H) and light chain (L). All domains contain disulphide-bonds (S-S) and heavy and light chains are also linked with disulphide bonds. Fab: fragment antigen-binding, Fc: fragment crystallizable. The also linked with disulphide bonds. Fab: fragment antigen-binding, Fc: fragment crystallizable. The scFv consists of variable domains of heavy and light chain (VH, VL) joined with a peptide linker. scFv consists of variable domains of heavy and light chain (VH, VL) joined with a peptide linker.

The antigen binding site is located on the tips of the Y-shaped antibody molecules and The antigen binding site is located on the tips of the Y-shaped antibody molecules and both VL and VH contribute with three variable loops each, also denoted complementary both VL and VH contribute with three variable loops each, also denoted complementary determining regions (CDR1/2/3) that form a continuous and very plastic binding surface 206. determining regions (CDR1/2/3) that form a continuous and very plastic binding surface 206. The stunning variability and plasticity of the antigen binding site is probably the reason for its The stunning variability and plasticity of the antigen binding site is probably the reason for its ability to bind a manifold of different antigens with high affinities. These adaptive binding ability to bind a manifold of different antigens with high affinities. These adaptive binding characteristics can be explained by a high sequence diversity of the six variable loops, several characteristics can be explained by a high sequence diversity of the six variable loops, several possible CDR loop conformations or canonical structures, fine-tuning of shape possible CDR loop conformations or canonical structures, fine-tuning of shape complementarity by single somatic mutations in the CDR framework or Verniet region 207 and complementarity by single somatic mutations in the CDR framework or Verniet region 207 and finally the possibility of an induced fit mechanism that allows for structural adaptation to the finally the possibility of an induced fit mechanism that allows for structural adaptation to the antigen 206. antigen 206. Upon cleavage with the enzyme papain, antibodies can be split at their flexible hinge Upon cleavage with the enzyme papain, antibodies can be split at their flexible hinge region into antigen-binding (Fab) or F(ab)2 fragments and crystallizable (Fc) fragments. The region into antigen-binding (Fab) or F(ab)2 fragments and crystallizable (Fc) fragments. The Fab fragment of IgA2, IgM and IgE classes and the Fc fragment of all antibody classes Fab fragment of IgA2, IgM and IgE classes and the Fc fragment of all antibody classes contain glycosylation sites 208. The Fc fragments mediate a number of class-dependent contain glycosylation sites 208. The Fc fragments mediate a number of class-dependent antibody characteristics and effector functions such as endosomal recycling of IgG leading to antibody characteristics and effector functions such as endosomal recycling of IgG leading to an extended half-life in the circulation, activation of effector cells such as natural killer cells an extended half-life in the circulation, activation of effector cells such as natural killer cells 42 Ribosome display for selection and evolution of affibody molecules 42 Ribosome display for selection and evolution of affibody molecules or phagocytes and complement cascade initiation. Also IgM pentamerization and IgA or phagocytes and complement cascade initiation. Also IgM pentamerization and IgA dimerization are Fc-mediated. dimerization are Fc-mediated. The possibility to use antibodies for the medication of diseases was discovered in The possibility to use antibodies for the medication of diseases was discovered in 1891, when von Behring used serum of cured diphtheria patients for the treatment of diseased 1891, when von Behring used serum of cured diphtheria patients for the treatment of diseased patients. He was later awarded the Nobel prize in 1901. Polyclonal antibody preparations patients. He was later awarded the Nobel prize in 1901. Polyclonal antibody preparations derived from serum were used for the treatment of infectious diseases and snake bites. derived from serum were used for the treatment of infectious diseases and snake bites. Depending on how they are purified, such preparations may however be of very uncertain Depending on how they are purified, such preparations may however be of very uncertain composition and may cause severe side reactions in patients 209. A breakthrough came when composition and may cause severe side reactions in patients 209. A breakthrough came when antibodies of uniform structure and composition could be obtained from single, immortalized antibodies of uniform structure and composition could be obtained from single, immortalized murine B cell clones, a discovery that bestowed Milstein and Köhler the Nobel prize in 1975 murine B cell clones, a discovery that bestowed Milstein and Köhler the Nobel prize in 1975 210. In principle, a laboratory mouse is exposed to an immunogen. Antibody producing plasma 210. In principle, a laboratory mouse is exposed to an immunogen. Antibody producing plasma cells are harvested and fused with immortal cancer B cells to obtain immortal hybrid cell lines cells are harvested and fused with immortal cancer B cells to obtain immortal hybrid cell lines (hybridomas) as a renewable source for monoclonal antibody production. (hybridomas) as a renewable source for monoclonal antibody production.

Hybridoma technology however has its limitations and antibodies can only be Hybridoma technology however has its limitations and antibodies can only be generated if the antigen is not toxic or not conserved across species. Besides, the procedure is generated if the antigen is not toxic or not conserved across species. Besides, the procedure is laborious and parameters such as affinity and epitope selectivity are difficult to control. laborious and parameters such as affinity and epitope selectivity are difficult to control. Equilibrium dissociation constants below a certain threshold of about 100 pM have been Equilibrium dissociation constants below a certain threshold of about 100 pM have been reported but are not actively selected for 211. reported but are not actively selected for 211. The accessibility of the antibody genotype encoded by the hybridoma allows for the The accessibility of the antibody genotype encoded by the hybridoma allows for the cloning of murine antibody genes and grafting of murine CDRs onto human antibody cloning of murine antibody genes and grafting of murine CDRs onto human antibody frameworks to generate humanized antibodies for therapy, such as Herceptin. Alternatively, frameworks to generate humanized antibodies for therapy, such as Herceptin. Alternatively, fully human antibodies can today be obtained by immunization of transgenic mice with a fully human antibodies can today be obtained by immunization of transgenic mice with a human immunoglobulin gene repertoire 212 or recombinantly as described in section 3.2.3. human immunoglobulin gene repertoire 212 or recombinantly as described in section 3.2.3.

3.2.2 Engineering of antibody effector functions 3.2.2 Engineering of antibody effector functions

Antibodies can be further engineered for therapeutic application to improve their effector Antibodies can be further engineered for therapeutic application to improve their effector functions, half-life and immunogenicity properties. The dominant antibody class found in functions, half-life and immunogenicity properties. The dominant antibody class found in human serum is IgG 208; 213 that is N-linked glycosylated at asparagine residue 297 with a human serum is IgG 208; 213 that is N-linked glycosylated at asparagine residue 297 with a branched oligosaccharide facing to the inside of the Fc. The oligosaccharide affects the branched oligosaccharide facing to the inside of the Fc. The oligosaccharide affects the overall Fc structure and thereby its affinity to different Fcγ receptors on immune effector overall Fc structure and thereby its affinity to different Fcγ receptors on immune effector cells. Three sets of IgG glycoforms exist in human serum that may contain a core fucose cells. Three sets of IgG glycoforms exist in human serum that may contain a core fucose residue of particular interest 208. Therapeutic antibodies are typically produced in rodent residue of particular interest 208. Therapeutic antibodies are typically produced in rodent mammalian cell lines such as CHO cells and such antibody preparations contain the core mammalian cell lines such as CHO cells and such antibody preparations contain the core fucose residue 214. Core fucose removal from the oligosaccharide was shown to increase the fucose residue 214. Core fucose removal from the oligosaccharide was shown to increase the antibody affinity to the FcγRIIIa found on natural killer (NK) cells, and hence increase the antibody affinity to the FcγRIIIa found on natural killer (NK) cells, and hence increase the ADCC, a desirable effect when targeting tumor cells that can be marked by the antibody for ADCC, a desirable effect when targeting tumor cells that can be marked by the antibody for NK cell-mediated killing. Defucosylated antibodies can be produced in a CHO-knockout cell NK cell-mediated killing. Defucosylated antibodies can be produced in a CHO-knockout cell Sebastian Grimm 43 Sebastian Grimm 43 line devoid of the FUT8 gene which encodes the fucosyltransferase 213 or from the line devoid of the FUT8 gene which encodes the fucosyltransferase 213 or from the supplementation of an inhibitory fucose analogue during antibody production (Peter Senter, supplementation of an inhibitory fucose analogue during antibody production (Peter Senter, personal communication). Recently, yeast strains capable of secreting antibodies with human personal communication). Recently, yeast strains capable of secreting antibodies with human glycosylation profiles were engineered in a “tour de force” by knocking out four genes for glycosylation profiles were engineered in a “tour de force” by knocking out four genes for yeast glycosylation and introducing as many as 14 additional glycosylation genes 213. yeast glycosylation and introducing as many as 14 additional glycosylation genes 213. The IgG glycosylation does however not significantly affect the affinity to the FcRn The IgG glycosylation does however not significantly affect the affinity to the FcRn receptor 215 that mediates antibody endosomal recycling and half-life extension. Here, amino receptor 215 that mediates antibody endosomal recycling and half-life extension. Here, amino acid mutations at the CH2-CH3 domain interface can alter the FcRn binding and hence Ab acid mutations at the CH2-CH3 domain interface can alter the FcRn binding and hence Ab half-life in vivo 216. half-life in vivo 216.

Besides ADCC, complement dependent cytotoxicity (CDC) is important for antibody Besides ADCC, complement dependent cytotoxicity (CDC) is important for antibody mediated killing of target cells. The complement protein C1q is known to bind to the CH2 mediated killing of target cells. The complement protein C1q is known to bind to the CH2 domain of the antibody Fc region and engage CDC. As demonstrated for the human domain of the antibody Fc region and engage CDC. As demonstrated for the human therapeutic IgG1 Rituximab targeting CD20 on non-Hodgkins B cell lymphoma, two residues therapeutic IgG1 Rituximab targeting CD20 on non-Hodgkins B cell lymphoma, two residues in the CH2 domain can be exchanged to achieve a stronger binding to C1q and an enhanced in the CH2 domain can be exchanged to achieve a stronger binding to C1q and an enhanced CDC 217. CDC 217. Antibodies can be produced such as to bear two selectivities on the same molecule Antibodies can be produced such as to bear two selectivities on the same molecule (i.e. bi-specific). One approach is to fuse two different mouse hybridomas to a hybrid (i.e. bi-specific). One approach is to fuse two different mouse hybridomas to a hybrid hybridoma or quadroma. Such cells secrete Ig molecules composed of up to 10 different hybridoma or quadroma. Such cells secrete Ig molecules composed of up to 10 different combinations of heavy and light chains, i.e. the yield of functional bispecific antibodies is combinations of heavy and light chains, i.e. the yield of functional bispecific antibodies is very low. Due to a species-selective pairing preference of Ab heavy and light chains, the yield very low. Due to a species-selective pairing preference of Ab heavy and light chains, the yield can be improved by fusing a mouse hybridoma with a rat hybridoma 218. The first bi-specific can be improved by fusing a mouse hybridoma with a rat hybridoma 218. The first bi-specific antibody approved for therapy was Removab, targeting the tumor antigen EpCAM with one antibody approved for therapy was Removab, targeting the tumor antigen EpCAM with one arm and CD3 on cytotoxic T cells with the other arm. This mAb consists of a mouse κ-light arm and CD3 on cytotoxic T cells with the other arm. This mAb consists of a mouse κ-light chain, a mouse IgG2a-heavy chain, a rat λ-light chain and a rat IgG2b-heavy chain and is chain, a mouse IgG2a-heavy chain, a rat λ-light chain and a rat IgG2b-heavy chain and is produced in a rat-mouse hybrid hybridoma 213. A dual selectivity can more elegantly be produced in a rat-mouse hybrid hybridoma 213. A dual selectivity can more elegantly be achieved by recombinantly engineering the antibody heavy chains for heterodimerization. The achieved by recombinantly engineering the antibody heavy chains for heterodimerization. The

CH3 domain provides the most extensive interaction between the heavy chains and in a knobs- CH3 domain provides the most extensive interaction between the heavy chains and in a knobs- into-holes approach, a bulky amino acid side chain “knob” is placed in one CH3 and a small into-holes approach, a bulky amino acid side chain “knob” is placed in one CH3 and a small

“whole” in its CH3 counterpart. Sterical complementarity is hence only granted if “whole” in its CH3 counterpart. Sterical complementarity is hence only granted if heterogonous heavy chains associate 219. Other approaches include the fusion of an additional heterogonous heavy chains associate 219. Other approaches include the fusion of an additional variable domain to the N-terminus of either heavy and light chains 220, the covalent linkage of variable domain to the N-terminus of either heavy and light chains 220, the covalent linkage of antibody fragments with different selectivities or fusion partner mediated oligomerizations of antibody fragments with different selectivities or fusion partner mediated oligomerizations of antibody fragments, to name a few 221. antibody fragments, to name a few 221. Coming back to Ehrlich who envisioned a magic bullet with efficacy and high affinity Coming back to Ehrlich who envisioned a magic bullet with efficacy and high affinity to a disease-causing pathogen. First attempts to create such a magic bullet with antibodies to a disease-causing pathogen. First attempts to create such a magic bullet with antibodies conjugated to toxins were already described in 1967, when Ghose and colleagues conjugated to toxins were already described in 1967, when Ghose and colleagues noncovalently coupled a polyclonal anti-Ehrlich carcinoma antibody preparation with noncovalently coupled a polyclonal anti-Ehrlich carcinoma antibody preparation with 44 Ribosome display for selection and evolution of affibody molecules 44 Ribosome display for selection and evolution of affibody molecules radioiodine. Ehrlich carcinoma cells treated with such an antibody preparation did not radioiodine. Ehrlich carcinoma cells treated with such an antibody preparation did not develop tumors after injection in mice, in contrast to carcinoma cells treated with non-iodated develop tumors after injection in mice, in contrast to carcinoma cells treated with non-iodated antibodies 222. In 1972, Ghose reported the noncovalent coupling of the chemotherapeutic antibodies 222. In 1972, Ghose reported the noncovalent coupling of the chemotherapeutic toxin chlorambucil to antibodies for the treatment of mice pre-inoculated with tumor cells 223. toxin chlorambucil to antibodies for the treatment of mice pre-inoculated with tumor cells 223. Mainly due to the inefficient non-covalent toxin coupling and probably also due to the Mainly due to the inefficient non-covalent toxin coupling and probably also due to the heterogeneity of the polyclonal antibody preparations, these early attempts did not result in heterogeneity of the polyclonal antibody preparations, these early attempts did not result in significant improvements of conventional chemotherapy. Today however, the first significant improvements of conventional chemotherapy. Today however, the first radioactively labeled monoclonal antibodies have been FDA approved, such as Zevalin and radioactively labeled monoclonal antibodies have been FDA approved, such as Zevalin and Bexxar 213 and additional antibodies that are covalently coupled via engineered linkers to Bexxar 213 and additional antibodies that are covalently coupled via engineered linkers to toxins such as DNA alkylating agents, maytansoid or mertansine have entered clinical trials. toxins such as DNA alkylating agents, maytansoid or mertansine have entered clinical trials. Despite their success in therapy applications and the fact that antibodies are Despite their success in therapy applications and the fact that antibodies are undisputedly the most widely used protein scaffold for molecular recognition, some undisputedly the most widely used protein scaffold for molecular recognition, some disadvantages can be named. Firstly, full-sized antibodies are rather large and highly complex disadvantages can be named. Firstly, full-sized antibodies are rather large and highly complex proteins consisting of several polypeptide chains and a structure dependent on glycosylations proteins consisting of several polypeptide chains and a structure dependent on glycosylations and correct disulphide bond formations. Thus, rather laborious manufacturing in mammalian and correct disulphide bond formations. Thus, rather laborious manufacturing in mammalian cell lines or engineered yeast is required 206, although some successful attempts to produce cell lines or engineered yeast is required 206, although some successful attempts to produce full-length but non-glycosylated antibodies in E. coli have been described 154; 224. full-length but non-glycosylated antibodies in E. coli have been described 154; 224. Furthermore, some applications such as the blocking of an immune-modulator receptor or the Furthermore, some applications such as the blocking of an immune-modulator receptor or the neutralization of cytokines in serum do not require immune effector functions, and other neutralization of cytokines in serum do not require immune effector functions, and other applications such as molecular imaging benefit from a small molecular weight and no applications such as molecular imaging benefit from a small molecular weight and no endosomal recycling for a fast renal clearance. This has, besides intellectual property rights, endosomal recycling for a fast renal clearance. This has, besides intellectual property rights, motivated the development of antibody fragments and binding proteins based on non- motivated the development of antibody fragments and binding proteins based on non- immunoglobulin scaffolds for molecular recognition. immunoglobulin scaffolds for molecular recognition.

3.2.3 Recombinant generation of antibody fragments and monoclonal antibodies 3.2.3 Recombinant generation of antibody fragments and monoclonal antibodies

An alternative route to animal immunization and hybridoma technology allows for the An alternative route to animal immunization and hybridoma technology allows for the generation of antibody fragments and full-length monoclonal antibodies with desired generation of antibody fragments and full-length monoclonal antibodies with desired properties from natural or synthetic library sources. This route was established during properties from natural or synthetic library sources. This route was established during progress made in the field of antibody fragment cloning, engineering and production as well progress made in the field of antibody fragment cloning, engineering and production as well as selection system technology. as selection system technology. In 1988, the productions of three different functional antibody fragment formats in E. In 1988, the productions of three different functional antibody fragment formats in E. coli were published in Science, firstly non-covalently associated VL and VH by Skerra and coli were published in Science, firstly non-covalently associated VL and VH by Skerra and 124 124 Plückthun , secondly VL carboxyl terminus covalently joined with a designed peptide linker Plückthun , secondly VL carboxyl terminus covalently joined with a designed peptide linker to the VH amino terminus (the single-chain fragment variable or scFv, see Figure 9) by Bird to the VH amino terminus (the single-chain fragment variable or scFv, see Figure 9) by Bird and colleagues 225 and thirdly a complete Fab fragment by Better and colleagues 226. Both the and colleagues 225 and thirdly a complete Fab fragment by Better and colleagues 226. Both the scFv and the Fab proved to be versatile formats for protein engineering. They can be tailored scFv and the Fab proved to be versatile formats for protein engineering. They can be tailored Sebastian Grimm 45 Sebastian Grimm 45 using recombinant DNA technology, produced at high yields in relatively cheap microbial using recombinant DNA technology, produced at high yields in relatively cheap microbial hosts and combined to a variety of bispecific or trispecific formats. hosts and combined to a variety of bispecific or trispecific formats. In principle, the in vitro generation of antibody fragments with desired affinity and In principle, the in vitro generation of antibody fragments with desired affinity and selectivity involves the construction of a library encoding the variable VL and VH antibody selectivity involves the construction of a library encoding the variable VL and VH antibody domains in a format suitable for display on a phage, ribosome or cell. The two most domains in a format suitable for display on a phage, ribosome or cell. The two most frequently used formats are scFv and Fab. Libraries of displayed fragments are then subjected frequently used formats are scFv and Fab. Libraries of displayed fragments are then subjected to a biopanning or screening procedure as described in chapter 2. Library sources can be to a biopanning or screening procedure as described in chapter 2. Library sources can be immune or naïve natural repertoires as well as synthetic, generated by oligonucleotide immune or naïve natural repertoires as well as synthetic, generated by oligonucleotide cassette mutagenesis as described in section 1.4.2. In comparison to animal immunization and cassette mutagenesis as described in section 1.4.2. In comparison to animal immunization and hybridoma technology, in vitro selection provides the protein engineer more set screws, e.g. hybridoma technology, in vitro selection provides the protein engineer more set screws, e.g. to assure a functional representation of the antigen, direct the selection towards particular to assure a functional representation of the antigen, direct the selection towards particular epitopes or select for other traits than affinity. Recent efforts include the selection of scFv epitopes or select for other traits than affinity. Recent efforts include the selection of scFv against a panel of around 100 different antigens ranging from proteins to small haptens and against a panel of around 100 different antigens ranging from proteins to small haptens and resulting in the identification of 435 unique binders 227 or the selection of scFv against 20 resulting in the identification of 435 unique binders 227 or the selection of scFv against 20 related SH2 domains that yielded 379 selectively binding scFvs 228. In both studies, scFvs related SH2 domains that yielded 379 selectively binding scFvs 228. In both studies, scFvs were selected from naïve natural repertoires. In a third study, the library was constructed were selected from naïve natural repertoires. In a third study, the library was constructed using oligonucleotide cassette mutagenesis in combination with trinucleotide building blocks, using oligonucleotide cassette mutagenesis in combination with trinucleotide building blocks, and Fab fragments were displayed as disulphide-linked to M13 phage pIII. Here, Fab and Fab fragments were displayed as disulphide-linked to M13 phage pIII. Here, Fab fragments with up to pM affinities for 23 different antigens were obtained 229. Also antibody fragments with up to pM affinities for 23 different antigens were obtained 229. Also antibody fragments can be armed with cytotoxic or radioactive payloads and as compared to full-length fragments can be armed with cytotoxic or radioactive payloads and as compared to full-length antibodies, such smaller and monovalent fragments are an interesting alternative for cancer antibodies, such smaller and monovalent fragments are an interesting alternative for cancer imaging and radiotherapy applications. They may more efficiently penetrate the tumor and are imaging and radiotherapy applications. They may more efficiently penetrate the tumor and are faster cleared from the circulation, i.e. reducing radiotoxic side effects 221. In vitro selected faster cleared from the circulation, i.e. reducing radiotoxic side effects 221. In vitro selected antibody fragments can if needed be converted into a full-length antibody format and antibody fragments can if needed be converted into a full-length antibody format and produced in eukaryote cell culture. At present, only one therapeutic antibody approved by produced in eukaryote cell culture. At present, only one therapeutic antibody approved by FDA has been generated involving phage display selection 213. FDA has been generated involving phage display selection 213.

3.2.4 Single domain antibody fragments 3.2.4 Single domain antibody fragments

It was during an immunology practical, when students refused to purify human serum because It was during an immunology practical, when students refused to purify human serum because it may contain HIV. The human serum was replaced with an exotic camel serum sample left it may contain HIV. The human serum was replaced with an exotic camel serum sample left over from an unrelated experiment and the camel serum was purified on protein A and protein over from an unrelated experiment and the camel serum was purified on protein A and protein G columns, a serendipity that led to the discovery of naturally occurring antibodies devoid of G columns, a serendipity that led to the discovery of naturally occurring antibodies devoid of light chains 230. This kind of antibodies are to our present knowledge exclusively found in light chains 230. This kind of antibodies are to our present knowledge exclusively found in camelids such as camels, llamas and cartilaginous fishes such as nurse sharks. In both cases, camelids such as camels, llamas and cartilaginous fishes such as nurse sharks. In both cases, the variable antigen binding entity is constituted of a single variable immunoglobulin domain the variable antigen binding entity is constituted of a single variable immunoglobulin domain denoted VHH (camelids) or V-NAR (cartilaginous fishes). The VH of “classical” antibodies denoted VHH (camelids) or V-NAR (cartilaginous fishes). The VH of “classical” antibodies has a relatively hydrophobic face, usually associated with the light chain, which makes it has a relatively hydrophobic face, usually associated with the light chain, which makes it 46 Ribosome display for selection and evolution of affibody molecules 46 Ribosome display for selection and evolution of affibody molecules aggregation-prone in solution when produced alone 221. It was shown that this face is less aggregation-prone in solution when produced alone 221. It was shown that this face is less hydrophobic in the VHH domain 231. Furthermore, VHHs also carry an unusually long CDR3 hydrophobic in the VHH domain 231. Furthermore, VHHs also carry an unusually long CDR3 loop that can fold back on this face, forming a small hydrophobic core and further stabilizing loop that can fold back on this face, forming a small hydrophobic core and further stabilizing the VHH domain fold 232. Stable genetically engineered single VHH domains with a MW of the VHH domain fold 232. Stable genetically engineered single VHH domains with a MW of approximately 15 kDa can be produced in yeast or E. coli. Some VHH antibody fragments approximately 15 kDa can be produced in yeast or E. coli. Some VHH antibody fragments contain an intramolecular disulphide bond between CDR1 and CDR3, which may assist in contain an intramolecular disulphide bond between CDR1 and CDR3, which may assist in shaping the loop structure 231 and may require protein folding in an oxidizing environment. shaping the loop structure 231 and may require protein folding in an oxidizing environment. Due to their protruding CDR3 loop, VHHs are suitable for the recognition of clefts as those Due to their protruding CDR3 loop, VHHs are suitable for the recognition of clefts as those found in enzyme substrate pockets, which are usually avoided by conventional antibodies 233. found in enzyme substrate pockets, which are usually avoided by conventional antibodies 233. VHHs with desired selectivity can be obtained from synthetic, naïve or immune libraries that VHHs with desired selectivity can be obtained from synthetic, naïve or immune libraries that are displayed on e.g. phage and selected for antigen binding. Applications being investigated are displayed on e.g. phage and selected for antigen binding. Applications being investigated for VHHs include in vitro diagnostics, nanosensors as well as in vivo cancer imaging, the for VHHs include in vitro diagnostics, nanosensors as well as in vivo cancer imaging, the treatment of rheumatoid arthritis, Crohn disease or thrombosis 221. Since VHHs are not treatment of rheumatoid arthritis, Crohn disease or thrombosis 221. Since VHHs are not derived from a human protein scaffold, their immunogenicity may be a concern for therapy derived from a human protein scaffold, their immunogenicity may be a concern for therapy applications. From the lessons learned with camelid heavy chain antibodies, VH domains of applications. From the lessons learned with camelid heavy chain antibodies, VH domains of mouse or human origin have been re-visited. A set of hydrophobic residues in the light chain mouse or human origin have been re-visited. A set of hydrophobic residues in the light chain interface was substituted with more hydrophilic residues and the CDR3 loop length was interface was substituted with more hydrophilic residues and the CDR3 loop length was extended for making them more aggregation-resistant and stable, i.e. achieving similar extended for making them more aggregation-resistant and stable, i.e. achieving similar biophysical properties as seen for VHH antibody fragments 234. Such engineered fragments biophysical properties as seen for VHH antibody fragments 234. Such engineered fragments were termed domain antibodies (dAbs) and can be selected from synthetic libraries displayed were termed domain antibodies (dAbs) and can be selected from synthetic libraries displayed on e.g. phage, as demonstrated for a range or targets such as enzymes, cytokines or on e.g. phage, as demonstrated for a range or targets such as enzymes, cytokines or transcription factors 234. Also dAbs can be produced at high yields in microorganisms such as transcription factors 234. Also dAbs can be produced at high yields in microorganisms such as yeast. yeast.

3.3 Alternative protein scaffolds 3.3 Alternative protein scaffolds

With the development of phage display, affinity ligands against various targets could be With the development of phage display, affinity ligands against various targets could be selected from libraries of linear or disulphide-constrained, cyclic peptides 235. It was found selected from libraries of linear or disulphide-constrained, cyclic peptides 235. It was found that such constrained peptides yield higher affinity ligands than linear peptides, probably due that such constrained peptides yield higher affinity ligands than linear peptides, probably due to a reduced entropic penalty upon target binding 236; 237. Another way to constrain peptides or to a reduced entropic penalty upon target binding 236; 237. Another way to constrain peptides or single variegated amino acid positions is to present them on a rigid protein framework, single variegated amino acid positions is to present them on a rigid protein framework, analogous to CDRs presented on an antibody. analogous to CDRs presented on an antibody.

In principle, a naturally occurring protein is chosen as framework or scaffold for In principle, a naturally occurring protein is chosen as framework or scaffold for genetic variation of a set of surface-exposed amino acid residues to obtain a library for genetic variation of a set of surface-exposed amino acid residues to obtain a library for selection of binding proteins with novel selectivities using protein selection systems. selection of binding proteins with novel selectivities using protein selection systems. Scaffolds derived from protein families with high affinities for different natural ligands may Scaffolds derived from protein families with high affinities for different natural ligands may Sebastian Grimm 47 Sebastian Grimm 47 be particularly suitable, since such original binding activities can give valuable clues to where be particularly suitable, since such original binding activities can give valuable clues to where to introduce variegation to obtain binders to novel targets. to introduce variegation to obtain binders to novel targets. In addition to these considerations, a protein scaffold should ideally meet some In addition to these considerations, a protein scaffold should ideally meet some general requirements. Firstly, the scaffold should have intrinsic favorable biophysical general requirements. Firstly, the scaffold should have intrinsic favorable biophysical properties. As such it should be stable, soluble and well-folding as a monomer. Secondly, it properties. As such it should be stable, soluble and well-folding as a monomer. Secondly, it should ideally tolerate the variation of a sufficiently large number amino acids for creation of should ideally tolerate the variation of a sufficiently large number amino acids for creation of the library without a too high penalty in its biophysical properties, i.e. any library member the library without a too high penalty in its biophysical properties, i.e. any library member should ideally have similar, class-specific biophysical properties. Thirdly, it should be easily should ideally have similar, class-specific biophysical properties. Thirdly, it should be easily produced at high yields in cheap, microbial expression systems such as E. coli or yeast, which produced at high yields in cheap, microbial expression systems such as E. coli or yeast, which also is a requirement for the use of common selection systems. Fourthly, the scaffold should also is a requirement for the use of common selection systems. Fourthly, the scaffold should consist of a single polypeptide chain and have a structural integrity that ideally does not consist of a single polypeptide chain and have a structural integrity that ideally does not depend on any disulphide bonds. The latter condition allows for an incorporation of a unique depend on any disulphide bonds. The latter condition allows for an incorporation of a unique cysteine residue for site-directed labeling. Depending on the intended application, resistance cysteine residue for site-directed labeling. Depending on the intended application, resistance to proteolytic digestion by e.g. serum proteases or resistance to sodium hydroxide treatment to proteolytic digestion by e.g. serum proteases or resistance to sodium hydroxide treatment as used for sanitizing chromatography columns may be desired. Furthermore, an application as used for sanitizing chromatography columns may be desired. Furthermore, an application in therapy requires a low immunogenicity, and scaffolds derived from human proteins or very in therapy requires a low immunogenicity, and scaffolds derived from human proteins or very small scaffolds that provide no or very few T-cell epitopes may be a good choice. The field of small scaffolds that provide no or very few T-cell epitopes may be a good choice. The field of alternative protein scaffolds currently shows a trend of consolidation towards a few classes alternative protein scaffolds currently shows a trend of consolidation towards a few classes that are tested extensively in different applications including clinical trials 206. Some of them, that are tested extensively in different applications including clinical trials 206. Some of them, including the affibody molecule class that this thesis is based upon, will be presented here in including the affibody molecule class that this thesis is based upon, will be presented here in more detail. more detail.

3.3.1 Anticalins 3.3.1 Anticalins

Anticalins are derived from lipocalins, a family of proteins occurring in different organisms Anticalins are derived from lipocalins, a family of proteins occurring in different organisms such as bacteria, plants, insects or vertebrates, with the natural function to bind, store or such as bacteria, plants, insects or vertebrates, with the natural function to bind, store or transport organic compounds such as vitamins, lipids or steroids 238. Lipocalins are about 21 transport organic compounds such as vitamins, lipids or steroids 238. Lipocalins are about 21 kDa in size and share a β-barrel fold consisting of eight anti-parallel β-strands winding kDa in size and share a β-barrel fold consisting of eight anti-parallel β-strands winding around a central axis, and typically a C-terminal α-helix that is associated to the β-barrel. around a central axis, and typically a C-terminal α-helix that is associated to the β-barrel. Lipocalins also typically contain one or two disulphide bonds, as commonly found in proteins Lipocalins also typically contain one or two disulphide bonds, as commonly found in proteins secreted from eukaryotes 239. secreted from eukaryotes 239. Like in a cup or calyx, the ligand entry site is expanded and forms the entry to an inner Like in a cup or calyx, the ligand entry site is expanded and forms the entry to an inner pocket for ligand binding. The β-barrel fold is highly conserved, while the ligand entry site pocket for ligand binding. The β-barrel fold is highly conserved, while the ligand entry site which includes four loops connecting the β-strands is very dissimilar between different which includes four loops connecting the β-strands is very dissimilar between different lipocalins. In earlier studies, Skerra and colleagues used the bilin-binding protein (BBP) from lipocalins. In earlier studies, Skerra and colleagues used the bilin-binding protein (BBP) from Pieris brassicacea as scaffold for the randomization of 16 positions located within the four Pieris brassicacea as scaffold for the randomization of 16 positions located within the four loops or adjacent β-strands at the ligand entry side. From this library, binders with nM loops or adjacent β-strands at the ligand entry side. From this library, binders with nM affinities for the haptens fluorescein and digoxigenin could be selected using phage display affinities for the haptens fluorescein and digoxigenin could be selected using phage display 48 Ribosome display for selection and evolution of affibody molecules 48 Ribosome display for selection and evolution of affibody molecules

240. From the lipocalins natural binding profiles, one would expect that such engineered 240. From the lipocalins natural binding profiles, one would expect that such engineered lipocalins, termed anticalins, would be well-suited to accommodate novel selectivities lipocalins, termed anticalins, would be well-suited to accommodate novel selectivities towards haptens, which these studies confirm. Later libraries were based on lipocalin towards haptens, which these studies confirm. Later libraries were based on lipocalin scaffolds of human origin such as ApoD, NGAL and Tcl to possibly minimize problems with scaffolds of human origin such as ApoD, NGAL and Tcl to possibly minimize problems with immunogenicity in later therapeutic applications in humans. From these libraries, involving a immunogenicity in later therapeutic applications in humans. From these libraries, involving a different set of positions than earlier when hapten targets were addressed, binders to protein different set of positions than earlier when hapten targets were addressed, binders to protein targets such as CTLA-4 to stimulate T-cell proliferation or VEGF to block tumor targets such as CTLA-4 to stimulate T-cell proliferation or VEGF to block tumor angiogenesis could be selected using phage display 240, and a VEGF binder is currently in angiogenesis could be selected using phage display 240, and a VEGF binder is currently in clinical trials. Theses studies together suggest the versatility of anticalins for binding both clinical trials. Theses studies together suggest the versatility of anticalins for binding both small haptens and larger macromolecules such as proteins. Since most anticalins contain small haptens and larger macromolecules such as proteins. Since most anticalins contain disulphide bonds, they require an oxidizing environment for proper folding. The secretion in disulphide bonds, they require an oxidizing environment for proper folding. The secretion in the oxidizing E. coli periplasm was described and a co-secretion of proteins that support the the oxidizing E. coli periplasm was described and a co-secretion of proteins that support the formation and isomerization of disulphide bonds may improve the yield of functional protein formation and isomerization of disulphide bonds may improve the yield of functional protein 241. 241.

3.3.2 Affibody molecules 3.3.2 Affibody molecules

A small scaffold for the selection of binding proteins was described in 1995 by Nord and A small scaffold for the selection of binding proteins was described in 1995 by Nord and colleagues 205 and is derived from the B domain of the Staphylococcus aureus cell surface colleagues 205 and is derived from the B domain of the Staphylococcus aureus cell surface receptor protein A. The natural function of protein A is to bind immunoglobulins from host receptor protein A. The natural function of protein A is to bind immunoglobulins from host serum and to thereby evade and modulate the immune system 242. All of the five highly serum and to thereby evade and modulate the immune system 242. All of the five highly homologous immunoglobulin-binding domains of protein A, denoted E, D, A, B and C, share homologous immunoglobulin-binding domains of protein A, denoted E, D, A, B and C, share a three-helix bundle fold. The 58 aa B domain was engineered by exchange of glycine for a three-helix bundle fold. The 58 aa B domain was engineered by exchange of glycine for alanine at position 29 for an increased resistance towards hydroxylamine, used to cleave alanine at position 29 for an increased resistance towards hydroxylamine, used to cleave fusion proteins at Asn-Gly junctions in biotechnology applications, and by domain end fusion proteins at Asn-Gly junctions in biotechnology applications, and by domain end substitution to facilitate head-to-tail dimerization at DNA level 243 to become the Z domain. In substitution to facilitate head-to-tail dimerization at DNA level 243 to become the Z domain. In contrast to the B domain that binds to both Fab and Fc fragments of IgG in a species, subclass contrast to the B domain that binds to both Fab and Fc fragments of IgG in a species, subclass and sequence dependent manner, the Z domain has only its Fc binding retained. With 58 and sequence dependent manner, the Z domain has only its Fc binding retained. With 58 amino acids and a molecular weight of about 6 kDa, the Z domain is only half of the size of amino acids and a molecular weight of about 6 kDa, the Z domain is only half of the size of an antibody variable domain (VH or VL) but still accommodates a binding interface of about an antibody variable domain (VH or VL) but still accommodates a binding interface of about 800-900 Å2, that is typically found in protein interfaces 244. The Z domain is rapidly-folding 800-900 Å2, that is typically found in protein interfaces 244. The Z domain is rapidly-folding 245 and does not contain any cysteine residues or disulphide bonds, which allows for the 245 and does not contain any cysteine residues or disulphide bonds, which allows for the production and folding in reducing environment, including the E. coli cytoplasm. Further, the production and folding in reducing environment, including the E. coli cytoplasm. Further, the small size allows for production of affibody molecules, and site-specifically modified variants small size allows for production of affibody molecules, and site-specifically modified variants thereof, by chemical synthesis (i.e. solid phase peptide synthesis) 246. Thirteen surface- thereof, by chemical synthesis (i.e. solid phase peptide synthesis) 246. Thirteen surface- exposed residues on helices one and two of the Z domain, involving nine positions exposed residues on helices one and two of the Z domain, involving nine positions participating in the interaction with immunoglobulins and four additional ones located at the participating in the interaction with immunoglobulins and four additional ones located at the same molecular surface were originally randomized to re-shape the immunoglobulin binding same molecular surface were originally randomized to re-shape the immunoglobulin binding Sebastian Grimm 49 Sebastian Grimm 49 face and allow for the selection of affibody molecules that are based on protein Z but face and allow for the selection of affibody molecules that are based on protein Z but equipped with novel binding selectivities [31]. equipped with novel binding selectivities [31]. Early libraries were typically created using NN(G/T) or (C/A/G)NN degenerate Early libraries were typically created using NN(G/T) or (C/A/G)NN degenerate codons, while the latest libraries are built of defined sets of trinucleotide codons and are based codons, while the latest libraries are built of defined sets of trinucleotide codons and are based on a further engineered scaffold for higher thermal stability, higher hydrophilicity and more on a further engineered scaffold for higher thermal stability, higher hydrophilicity and more efficient peptide synthesis of library members 247. Selection systems used include efficient peptide synthesis of library members 247. Selection systems used include predominantly phage display, but also protein fragment complementation assay 200, phage in predominantly phage display, but also protein fragment complementation assay 200, phage in combination with Staphylococcus cell surface display 202 and, as described in this thesis, combination with Staphylococcus cell surface display 202 and, as described in this thesis, ribosome display 248 has been used. Binders have been obtained against a manifold of targets, ribosome display 248 has been used. Binders have been obtained against a manifold of targets, including Taq DNA polymerase 249, transferrin 250, TNF-α 200, Her2 251, IGF-1R 252, other including Taq DNA polymerase 249, transferrin 250, TNF-α 200, Her2 251, IGF-1R 252, other affibody molecules 253; 254, IgA 255, HIV gp120 256, CD28 257 but also the amyloid beta peptide affibody molecules 253; 254, IgA 255, HIV gp120 256, CD28 257 but also the amyloid beta peptide 258. So far, the structures of four different affibody molecules in complex with their respective 258. So far, the structures of four different affibody molecules in complex with their respective targets have been determined. Two of these complexes correspond to so called anti-idiotypic targets have been determined. Two of these complexes correspond to so called anti-idiotypic affibody molecules, i.e. affibody molecules that bind protein Z or another affibody molecule affibody molecules, i.e. affibody molecules that bind protein Z or another affibody molecule used as target for selections. These interactions were found to be characterized by relatively used as target for selections. These interactions were found to be characterized by relatively flat protein surfaces and a retained 3-helix bundle fold. The structure of the affibody molecule flat protein surfaces and a retained 3-helix bundle fold. The structure of the affibody molecule

ZHER2 in complex with the Her2 extracellular domain, with an affinity of 22 pM, also revealed ZHER2 in complex with the Her2 extracellular domain, with an affinity of 22 pM, also revealed a rather flat recognized surface and a retained 3-helix bundle fold 259. Interestingly, in this a rather flat recognized surface and a retained 3-helix bundle fold 259. Interestingly, in this thorough study, also residues other than the 13 initially randomized ones participate in the thorough study, also residues other than the 13 initially randomized ones participate in the binding interaction. These three studies may suggest a general epitope preference to rather flat binding interaction. These three studies may suggest a general epitope preference to rather flat protein surfaces and a conserved 3-helix affibody fold. The structure of the amyloid beta protein surfaces and a conserved 3-helix affibody fold. The structure of the amyloid beta peptide binding affibody molecule ZAB3 in complex with the amyloid-β peptide however peptide binding affibody molecule ZAB3 in complex with the amyloid-β peptide however disproves this conclusion. ZAB3 contains a single cysteine residue in one of the 13 randomized disproves this conclusion. ZAB3 contains a single cysteine residue in one of the 13 randomized positions and forms a S-S-bridged homodimer. Here, helix one of each affibody unit adapts a positions and forms a S-S-bridged homodimer. Here, helix one of each affibody unit adapts a β-strand structure to form a shared β-sheet with the amyloid-β peptide 258. This structure β-strand structure to form a shared β-sheet with the amyloid-β peptide 258. This structure firstly proves that affibody molecules can also bind peptides and secondly, it demonstrates firstly proves that affibody molecules can also bind peptides and secondly, it demonstrates that affibody molecules may change secondary structure when in complex with their target. that affibody molecules may change secondary structure when in complex with their target. Due to its small size, high affinity and favorable pharmacokinetics such as tumor Due to its small size, high affinity and favorable pharmacokinetics such as tumor penetration and fast renal clearance, the affibody molecule ZHER2 proved to be a suitable penetration and fast renal clearance, the affibody molecule ZHER2 proved to be a suitable affinity reagent for the radioimaging of Her2-expressing tumors in a mouse xenograft model affinity reagent for the radioimaging of Her2-expressing tumors in a mouse xenograft model 37 and has already entered clinical studies in humans for in vivo imaging of breast cancer 260. 37 and has already entered clinical studies in humans for in vivo imaging of breast cancer 260.

The possibility to chemically synthesize ZHER2 facilitates the incorporation of non-natural The possibility to chemically synthesize ZHER2 facilitates the incorporation of non-natural functional groups that can be used for site-directed labeling with e.g. radionuclides 261, a functional groups that can be used for site-directed labeling with e.g. radionuclides 261, a distinct advantage in comparison to bigger and more complex scaffolds. distinct advantage in comparison to bigger and more complex scaffolds.

Therapeutic affibody applications have been investigated. In a study by Rafal and Therapeutic affibody applications have been investigated. In a study by Rafal and colleagues, a truncated form of the Pseudomonas exotoxin A (PE38) was fused to ZHER2 for a colleagues, a truncated form of the Pseudomonas exotoxin A (PE38) was fused to ZHER2 for a selective delivery of the toxin to Her2-expressing cancer cells, and a correlation between Her2 selective delivery of the toxin to Her2-expressing cancer cells, and a correlation between Her2 50 Ribosome display for selection and evolution of affibody molecules 50 Ribosome display for selection and evolution of affibody molecules expression and PE38-mediated inhibition of protein synthesis was observed 262. In a second expression and PE38-mediated inhibition of protein synthesis was observed 262. In a second study, drug-containing polymeric nanoparticles were coated with ZHER2 for targeted drug study, drug-containing polymeric nanoparticles were coated with ZHER2 for targeted drug 263 263 delivery to Her2-expressing cells . In a third study, lipid-conjugated affibody ZEGFR was delivery to Her2-expressing cells . In a third study, lipid-conjugated affibody ZEGFR was inserted in drug-containing liposomes for targeted drug delivery to EGFR-expressing cells 264. inserted in drug-containing liposomes for targeted drug delivery to EGFR-expressing cells 264.

The pharmacokinetic parameters of dimeric ZHER2 for tumor radiotherapy in a mouse model The pharmacokinetic parameters of dimeric ZHER2 for tumor radiotherapy in a mouse model could be dramatically improved by fusion to an albumin binding domain (ABD), resulting in could be dramatically improved by fusion to an albumin binding domain (ABD), resulting in a half-life extension from 0.64 to 36 h and a 25-fold reduction in renal uptake as compared to a half-life extension from 0.64 to 36 h and a 25-fold reduction in renal uptake as compared to 265 265 non-fused ZHER2 dimer . non-fused ZHER2 dimer . Other investigated applications of affibody molecules include their use as ligands in Other investigated applications of affibody molecules include their use as ligands in affinity purification 254, detection in bioassays 266, viral retargeting 267 or even as temperature affinity purification 254, detection in bioassays 266, viral retargeting 267 or even as temperature sensitive DNA polymerase inhibitors to achieve hot-start conditions in PCR 268. sensitive DNA polymerase inhibitors to achieve hot-start conditions in PCR 268.

Figure 10: Cartoon and Connolly surface representation of the Z domain. The 13 amino acid side Figure 10: Cartoon and Connolly surface representation of the Z domain. The 13 amino acid side chains in helix 1 and 2 that are variegated in a classical affibody library are highlighted in blue. The chains in helix 1 and 2 that are variegated in a classical affibody library are highlighted in blue. The isoleucine 31 side chain is located in the middle of the binding face and highlighted in red. isoleucine 31 side chain is located in the middle of the binding face and highlighted in red. Sebastian Grimm 51 Sebastian Grimm 51

3.3.3 Adnectins 3.3.3 Adnectins

Another interesting and thoroughly-validated class of alternative protein scaffold are Another interesting and thoroughly-validated class of alternative protein scaffold are adnectins, that are engineered from the 10th human fibronectin type III domain (10Fn3). adnectins, that are engineered from the 10th human fibronectin type III domain (10Fn3). Fibronectin is an abundant protein constituent of the extracellular matrix and body fluids such Fibronectin is an abundant protein constituent of the extracellular matrix and body fluids such as plasma and contains multiple head-to-tail connected type III domains 269. These domains as plasma and contains multiple head-to-tail connected type III domains 269. These domains are structurally similar to a variable domain of an antibody and consist of a sandwich of two are structurally similar to a variable domain of an antibody and consist of a sandwich of two β-sheets with three variable loops on each of two poles that resemble an antibody’s CDRs. β-sheets with three variable loops on each of two poles that resemble an antibody’s CDRs. Despite of its high content of proline residues in the loop regions, the 10Fn3 domain has been Despite of its high content of proline residues in the loop regions, the 10Fn3 domain has been shown to refold rapidly even in the presence of 0.64 M guanidine hydrochloride 270. With a shown to refold rapidly even in the presence of 0.64 M guanidine hydrochloride 270. With a Tm of above 80° C 271 and a free energy of unfolding of -6.1 kcal/mol 272, it is relatively Tm of above 80° C 271 and a free energy of unfolding of -6.1 kcal/mol 272, it is relatively stable and unlike immunoglobulin-domains, the 10Fn3 does not contain any disulphide bonds stable and unlike immunoglobulin-domains, the 10Fn3 does not contain any disulphide bonds and thus retains its stability under reducing conditions. Taken together, these parameters make and thus retains its stability under reducing conditions. Taken together, these parameters make it a very suitable scaffold for the introduction of novel binding selectivities. it a very suitable scaffold for the introduction of novel binding selectivities.

To investigate this possibility, Koide and colleagues aligned the sequences of 17 To investigate this possibility, Koide and colleagues aligned the sequences of 17 naturally occurring 10Fn3 domains and decided to variegate the variable, adjacent BC and FG naturally occurring 10Fn3 domains and decided to variegate the variable, adjacent BC and FG loops with five randomized aa positions, respectively. A library of 108 clones was displayed loops with five randomized aa positions, respectively. A library of 108 clones was displayed on phage and binders against ubiquitin could be selected 273. To further explore other than on phage and binders against ubiquitin could be selected 273. To further explore other than these loops for the selection of binders, Batori and colleagues inserted four glycine residues in these loops for the selection of binders, Batori and colleagues inserted four glycine residues in the six loops and found that for one out of the six loops, the insertion was highly destabilizing the six loops and found that for one out of the six loops, the insertion was highly destabilizing the overall scaffold, while the insertions in the remaining five loops were tolerated 274. In the overall scaffold, while the insertions in the remaining five loops were tolerated 274. In most of the selection experiments performed to date, either two or three loops located on one most of the selection experiments performed to date, either two or three loops located on one of the poles (BC, FG or BG, DE, FG loops) were diversified and binders with µM to nM of the poles (BC, FG or BG, DE, FG loops) were diversified and binders with µM to nM affinities have been obtained. Applied selection systems include phage, yeast and mRNA affinities have been obtained. Applied selection systems include phage, yeast and mRNA display. Second generation libraries involving error-prone PCR and DNA shuffling display. Second generation libraries involving error-prone PCR and DNA shuffling procedures have yielded clones with pM affinities 275. The reported antigens are to date procedures have yielded clones with pM affinities 275. The reported antigens are to date exclusively proteins, and tyrosine residues seem to be enriched in the antigen binding exclusively proteins, and tyrosine residues seem to be enriched in the antigen binding interface 271. Comparison studies of adnectin-libraries of different composition suggest that a interface 271. Comparison studies of adnectin-libraries of different composition suggest that a complete set of 20 amino acids yields higher affinity binders than a binary set of tyrosine and complete set of 20 amino acids yields higher affinity binders than a binary set of tyrosine and serine residues 102 and that a tailored set of amino acids frequently found in the antibody serine residues 102 and that a tailored set of amino acids frequently found in the antibody heavy chain CDR3 that also takes preferences of the 10Fn3 wild type scaffold into account heavy chain CDR3 that also takes preferences of the 10Fn3 wild type scaffold into account yields significantly more binders than a fully diversified library 276. An adnectin directed yields significantly more binders than a fully diversified library 276. An adnectin directed against VEGF-R2 was PEGylated for an extension of if half-life in the circulation and has against VEGF-R2 was PEGylated for an extension of if half-life in the circulation and has already entered clinical phase II trials for the therapy of solid tumors 271. already entered clinical phase II trials for the therapy of solid tumors 271. 52 Ribosome display for selection and evolution of affibody molecules 52 Ribosome display for selection and evolution of affibody molecules

3.3.4 DARPins 3.3.4 DARPins

Repeat proteins are composed of structural units, such as the Ankyrin, WD, leucine-rich or Repeat proteins are composed of structural units, such as the Ankyrin, WD, leucine-rich or tricopeptide repeats, and mediate a multitude of protein interactions in different species. A tricopeptide repeats, and mediate a multitude of protein interactions in different species. A feature of this class of proteins is the modular size of the binding face that allows repeat feature of this class of proteins is the modular size of the binding face that allows repeat proteins to fit their binding partners 277. Notably, the immune system of jawless fishes that proteins to fit their binding partners 277. Notably, the immune system of jawless fishes that may represent an evolutionary remaining of early vertebrates employs leucine-rich repeats may represent an evolutionary remaining of early vertebrates employs leucine-rich repeats instead of immunoglobulins for the generation of antigen-specific binding proteins 278. The instead of immunoglobulins for the generation of antigen-specific binding proteins 278. The ankyrin repeat is typically 33 amino acid long and composed of a β-turn followed by two anti- ankyrin repeat is typically 33 amino acid long and composed of a β-turn followed by two anti- parallel α-helices 279. Designed ankyrin-repeat proteins (DARPins) were assembled from two parallel α-helices 279. Designed ankyrin-repeat proteins (DARPins) were assembled from two components in a method referred to as consensus design 279. The first component is a designed components in a method referred to as consensus design 279. The first component is a designed ankyrin-repeat module obtained from a consensus sequence of a set of aligned natural ankyrin-repeat module obtained from a consensus sequence of a set of aligned natural ankyrin-repeats that was refined based on structural information. The rationale is that such a ankyrin-repeats that was refined based on structural information. The rationale is that such a consensus sequence would contain all conserved amino acids and represent an extraordinarily consensus sequence would contain all conserved amino acids and represent an extraordinarily stable scaffold. Such designed ankyrin-repeats can be aligned side-by-side to share a common stable scaffold. Such designed ankyrin-repeats can be aligned side-by-side to share a common hydrophobic core. The second DARPin component corresponds to modified, naturally hydrophobic core. The second DARPin component corresponds to modified, naturally occurring capping repeats that were included at the N- or C-terminus, respectively, to seal the occurring capping repeats that were included at the N- or C-terminus, respectively, to seal the hydrophobic core. Typically two to six consensus modules are sealed with two capping hydrophobic core. Typically two to six consensus modules are sealed with two capping repeats. From this DARPin design, libraries were prepared by randomizing 7 amino acids per repeats. From this DARPin design, libraries were prepared by randomizing 7 amino acids per consensus module on β-turn and the first α-helix 279 and binders against e.g. MBP, p38, Her2, consensus module on β-turn and the first α-helix 279 and binders against e.g. MBP, p38, Her2, JNK2 or APH have been selected using either ribosome or phage display 280; 281. The latter JNK2 or APH have been selected using either ribosome or phage display 280; 281. The latter required the signal recognition particle-mediated translocation for an efficient display of the required the signal recognition particle-mediated translocation for an efficient display of the DARPins, as mentioned in section 2.2. An impressive equilibrium dissociation constant of DARPins, as mentioned in section 2.2. An impressive equilibrium dissociation constant of 100 pM was obtained for a DARPin selected against Her4 from a first-generation library 281. 100 pM was obtained for a DARPin selected against Her4 from a first-generation library 281. As all other hitherto discussed alternative scaffolds, also DARPins have entered clinical trials As all other hitherto discussed alternative scaffolds, also DARPins have entered clinical trials and an application for molecular imaging is investigated 282. and an application for molecular imaging is investigated 282.

3.3.5 Other alternative protein scaffolds 3.3.5 Other alternative protein scaffolds

The variety of described alternative protein frameworks is manifold and includes scaffolds The variety of described alternative protein frameworks is manifold and includes scaffolds recruited for the display of constrained peptides, scaffolds with intrinsic fluorescence or recruited for the display of constrained peptides, scaffolds with intrinsic fluorescence or catalytic activity that facilitate molecular detection or scaffolds engineered to yield peptide catalytic activity that facilitate molecular detection or scaffolds engineered to yield peptide binders. Others are such as DARPins based on repeat proteins and yet another class are very binders. Others are such as DARPins based on repeat proteins and yet another class are very small disulphide-stabilized domains 280. small disulphide-stabilized domains 280.

For the display of constrained peptides and as described for adnectins, a surface- For the display of constrained peptides and as described for adnectins, a surface- exposed loop of a scaffold protein is typically chosen, that tolerates the insertion of a stretch exposed loop of a scaffold protein is typically chosen, that tolerates the insertion of a stretch of foreign amino acids. One example is the cytotoxic T lymphocyte-associated antigen 4 of foreign amino acids. One example is the cytotoxic T lymphocyte-associated antigen 4 (CTLA-4). Here, an RGD peptide motif flanked by randomized positions was inserted in a (CTLA-4). Here, an RGD peptide motif flanked by randomized positions was inserted in a CDR3-like loop and displayed on phage for bio-panning against a human integrin 283. Some CDR3-like loop and displayed on phage for bio-panning against a human integrin 283. Some Sebastian Grimm 53 Sebastian Grimm 53 pharmaceutically relevant drug targets are plasma proteases and hence, stable and small pharmaceutically relevant drug targets are plasma proteases and hence, stable and small protease inhibitors have been engineered to accommodate novel enzyme selectivities in protease inhibitors have been engineered to accommodate novel enzyme selectivities in typically extended peptide loops that are positioned and constrained by a small, disulphide- typically extended peptide loops that are positioned and constrained by a small, disulphide- stabilized scaffold. Examples are Kunitz domain-based inhibitors of plasma kallikrein 284 or a stabilized scaffold. Examples are Kunitz domain-based inhibitors of plasma kallikrein 284 or a complex of human tissue factor and factor VIIa 285 of the blood coagulation cascade. More complex of human tissue factor and factor VIIa 285 of the blood coagulation cascade. More examples of scaffold-constrained peptides are thioredoxin A 286 or staphylococcal nuclease examples of scaffold-constrained peptides are thioredoxin A 286 or staphylococcal nuclease 287. 287.

The green fluorescent protein (GFP) is an intrinsically fluorescent protein with a wide The green fluorescent protein (GFP) is an intrinsically fluorescent protein with a wide range of applications in molecular biology. It is appealing to combine selective binding with range of applications in molecular biology. It is appealing to combine selective binding with fluorescence activity in one single protein domain. GFP however has no natural ligand and fluorescence activity in one single protein domain. GFP however has no natural ligand and hence no naturally evolved binding face to engineer. Furthermore, loop variegations resulted hence no naturally evolved binding face to engineer. Furthermore, loop variegations resulted in a loss of fluorescence in early studies 288. Anyhow, variants of GFP have been selected that in a loss of fluorescence in early studies 288. Anyhow, variants of GFP have been selected that tolerate aa insertions in two of the β-strand connecting loops and binders with retained tolerate aa insertions in two of the β-strand connecting loops and binders with retained fluorescence and nM affinities could be selected 289. Similarily, β-lactamase was engineered fluorescence and nM affinities could be selected 289. Similarily, β-lactamase was engineered to bind with 10 to 20 nM affinity to ferritin. Such an engineered enzyme could be used as to bind with 10 to 20 nM affinity to ferritin. Such an engineered enzyme could be used as alternative to enzyme-conjugated antibodies in immunoassays 290. alternative to enzyme-conjugated antibodies in immunoassays 290. The T-cell receptor binds to peptides presented on MHCI molecules, with equilibrium The T-cell receptor binds to peptides presented on MHCI molecules, with equilibrium dissociation constants in the typically 1-50 µM range 291. Single-chain TCR versions have dissociation constants in the typically 1-50 µM range 291. Single-chain TCR versions have been engineered for high stability and solubility 292 and served as scaffold for the selection of been engineered for high stability and solubility 292 and served as scaffold for the selection of binders to e.g. a staphylococcal enterotoxin with low nM affinity 293. Other small peptide binders to e.g. a staphylococcal enterotoxin with low nM affinity 293. Other small peptide binding protein scaffolds are derived from the Src homology domain 2 (SH2) 294, domain 3 binding protein scaffolds are derived from the Src homology domain 2 (SH2) 294, domain 3 (SH3) 295 or the PDZ domain 296 which are all naturally evolved to bind peptides. (SH3) 295 or the PDZ domain 296 which are all naturally evolved to bind peptides. The consensus design described for the engineering of DARPins was also applied to The consensus design described for the engineering of DARPins was also applied to leucine-rich repeat proteins 297 and armadillo repeat proteins 298. Armadillo repeat proteins leucine-rich repeat proteins 297 and armadillo repeat proteins 298. Armadillo repeat proteins naturally bind to extended peptides and designed Armadillo repeat modules with selectivity to naturally bind to extended peptides and designed Armadillo repeat modules with selectivity to different amino acids may in future be combined to generate binders to any given peptide different amino acids may in future be combined to generate binders to any given peptide sequence. sequence. Knottins are very small scaffolds that are entropically stabilized by multiple Knottins are very small scaffolds that are entropically stabilized by multiple disulphide bonds, remarkably stable against thermal or chemical unfolding 299 and naturally disulphide bonds, remarkably stable against thermal or chemical unfolding 299 and naturally found in toxins such as the scorpion toxin. They have been engineered to bind to e.g. metal found in toxins such as the scorpion toxin. They have been engineered to bind to e.g. metal ions 300, HIV gp120 301 or integrin 302. Due to their small size of about 4 kDa, knottins are as ions 300, HIV gp120 301 or integrin 302. Due to their small size of about 4 kDa, knottins are as affibody molecules suitable for molecular imaging 302. An overview figure on all discussed affibody molecules suitable for molecular imaging 302. An overview figure on all discussed protein scaffolds is given in Figure 11. protein scaffolds is given in Figure 11. 54 Ribosome display for selection and evolution of affibody molecules 54 Ribosome display for selection and evolution of affibody molecules

Figure 11: Protein scaffolds for molecular recognition. All scaffolds are in the same scale for Figure 11: Protein scaffolds for molecular recognition. All scaffolds are in the same scale for comparability. The scale bar indicates a size of 6 nm. PDB ID numbers are given in brackets. (1-5): comparability. The scale bar indicates a size of 6 nm. PDB ID numbers are given in brackets. (1-5): antibody and antibody fragments, (6-14): alternative protein scaffolds, (15) protein scaffold with antibody and antibody fragments, (6-14): alternative protein scaffolds, (15) protein scaffold with enzymatic activity. The figures were prepared using PyMOL and the following crystal or NMR enzymatic activity. The figures were prepared using PyMOL and the following crystal or NMR structures: (1) murine IgG2a (1IGT), (2) Fab (1GGB), (3) scFv (1P4I), (4) VHH (1OP9), (5) IgNAR structures: (1) murine IgG2a (1IGT), (2) Fab (1GGB), (3) scFv (1P4I), (4) VHH (1OP9), (5) IgNAR variable domain (2YWY), (6) lipocalin (1BBP chain A), (7) 10th fibronectin type III domain (1FNA), variable domain (2YWY), (6) lipocalin (1BBP chain A), (7) 10th fibronectin type III domain (1FNA), (8) protein Z (1Q2N), (9) CTLA-4 extracellular part (1DQT), (10) PDZ domain (1KWA chain A), (8) protein Z (1Q2N), (9) CTLA-4 extracellular part (1DQT), (10) PDZ domain (1KWA chain A), (11) SH3 domain (1BU1 chain A), (12) Kunitz inhibitor BPTI (1CBW chain D), (13) scyllatoxin (11) SH3 domain (1BU1 chain A), (12) Kunitz inhibitor BPTI (1CBW chain D), (13) scyllatoxin (1D5Q), (14) DARPin (1MJ0), (15) TEM-1 β-lactamase (1M40). (1D5Q), (14) DARPin (1MJ0), (15) TEM-1 β-lactamase (1M40).

Sebastian Grimm 55 Sebastian Grimm 55

4 Present investigation 4 Present investigation

In this chapter, the four studies that this thesis is based upon will be summarized, some In this chapter, the four studies that this thesis is based upon will be summarized, some conclusions will be drawn and future perspectives presented. conclusions will be drawn and future perspectives presented.

All four studies are centered around protein engineering and protein library All four studies are centered around protein engineering and protein library technology for the development of novel affinity reagents. Of primary interest has been the technology for the development of novel affinity reagents. Of primary interest has been the investigation of ribosome display technology for selections from affibody molecule libraries. investigation of ribosome display technology for selections from affibody molecule libraries. As discussed earlier in this thesis, ribosome display is an in vitro selection system, performed As discussed earlier in this thesis, ribosome display is an in vitro selection system, performed in a test tube, using purified RNA polymerases and a cell extract for transcription and in a test tube, using purified RNA polymerases and a cell extract for transcription and translation, respectively. The openness of the system has two important implications. Firstly, translation, respectively. The openness of the system has two important implications. Firstly, large libraries can rapidly be generated on a nucleic acid level using recombinant DNA large libraries can rapidly be generated on a nucleic acid level using recombinant DNA technology. Secondly, additional diversity can easily be introduced in-between selection technology. Secondly, additional diversity can easily be introduced in-between selection cycles, facilitating a step-wise evolution process. cycles, facilitating a step-wise evolution process.

To explore for the first time these potential advantages for work with affibody To explore for the first time these potential advantages for work with affibody molecule libraries was the primary motivation for this thesis. The individual studies have molecule libraries was the primary motivation for this thesis. The individual studies have involved phases of method development, library construction, selection and characterization involved phases of method development, library construction, selection and characterization of isolated affinity reagents. The addressed target proteins included a class of murine of isolated affinity reagents. The addressed target proteins included a class of murine monoclonal antibodies (I), a suggested cancer marker protein (II) and a protein involved in an monoclonal antibodies (I), a suggested cancer marker protein (II) and a protein involved in an important cell proliferation pathway (III and IV), reflecting the width of fields that can be important cell proliferation pathway (III and IV), reflecting the width of fields that can be addressed by affinity technology and with relevance for biotechnology, diagnostic and addressed by affinity technology and with relevance for biotechnology, diagnostic and therapeutic applications. therapeutic applications.

56 Ribosome display for selection and evolution of affibody molecules 56 Ribosome display for selection and evolution of affibody molecules

4.1 Ribosome display selection of a murine IgG1 Fab binding affibody 4.1 Ribosome display selection of a murine IgG1 Fab binding affibody molecule allowing species selective recovery of monoclonal antibodies molecule allowing species selective recovery of monoclonal antibodies (I) (I)

Ribosome display was originally described by Mattheakis 26 and further developed and Ribosome display was originally described by Mattheakis 26 and further developed and refined by Plückthun and colleagues 303. To date, ribosome display has been applied to the refined by Plückthun and colleagues 303. To date, ribosome display has been applied to the selection of different target binding entities, including peptides 174, scFvs 28, DARPins 45 and selection of different target binding entities, including peptides 174, scFvs 28, DARPins 45 and mammalian receptor domains 304. In this study, the possibility to use ribosome display for the mammalian receptor domains 304. In this study, the possibility to use ribosome display for the selection of affibody molecules was investigated. Ribosome display would allow for a rapid selection of affibody molecules was investigated. Ribosome display would allow for a rapid generation of different kinds of affibody libraries that could be used to explore novel generation of different kinds of affibody libraries that could be used to explore novel randomization strategies or identify aa positions that are important for the scaffold stability. randomization strategies or identify aa positions that are important for the scaffold stability. Large libraries may allow for the selection of high affinity binders or closely-related clones Large libraries may allow for the selection of high affinity binders or closely-related clones that facilitate subsequent re-randomizations for affinity maturation. Also, affibody molecules that facilitate subsequent re-randomizations for affinity maturation. Also, affibody molecules could be recovered by different principles during the biopanning procedure, eluting the could be recovered by different principles during the biopanning procedure, eluting the affibody-encoding mRNA and not involving the affibody-antigen interaction. affibody-encoding mRNA and not involving the affibody-antigen interaction. Initially, an expression cassette was designed for the in vitro transcription of affibody Initially, an expression cassette was designed for the in vitro transcription of affibody library-encoding mRNA and the translation of affibody molecules that are allowed to fold library-encoding mRNA and the translation of affibody molecules that are allowed to fold outside of the ribosome tunnel but stay tethered via a spacer protein to their cognate mRNA, outside of the ribosome tunnel but stay tethered via a spacer protein to their cognate mRNA, i.e. establishing a linkage of genotype and phenotype. This cassette consists of a T7 promoter i.e. establishing a linkage of genotype and phenotype. This cassette consists of a T7 promoter (pT7) for the initiation of in vitro transcription, a Shine-Dalgarno sequence (SD) for ribosome (pT7) for the initiation of in vitro transcription, a Shine-Dalgarno sequence (SD) for ribosome binding and the initiation of in vitro translation, and a "dummy" gene that can easily be binding and the initiation of in vitro translation, and a "dummy" gene that can easily be replaced with an affibody library gene fragment encoding affibody helices one and two. Helix replaced with an affibody library gene fragment encoding affibody helices one and two. Helix three is kept constant in classical affibody libraries and is encoded in the expression cassette, three is kept constant in classical affibody libraries and is encoded in the expression cassette, but can also be replaced if desired. Helix three is C-terminally fused to a spacer sequence that but can also be replaced if desired. Helix three is C-terminally fused to a spacer sequence that is derived from the M13 phage protein 3 (p3 spacer) gene and was used previously for is derived from the M13 phage protein 3 (p3 spacer) gene and was used previously for ribosome display 29. The spacer provides the affibody molecule the space and flexibility to ribosome display 29. The spacer provides the affibody molecule the space and flexibility to fold outside of the ribosome tunnel. It was also decided to include a 21-residue element fold outside of the ribosome tunnel. It was also decided to include a 21-residue element derived from the E. coli secretion monitor protein SecM. This motif is known to interact with derived from the E. coli secretion monitor protein SecM. This motif is known to interact with the ribosome tunnel, particularly via its proline 166 arrest point 175, and could possibly further the ribosome tunnel, particularly via its proline 166 arrest point 175, and could possibly further stabilize affibody-ribosome-mRNA (ARM) complexes. The SecM motif can optionally be stabilize affibody-ribosome-mRNA (ARM) complexes. The SecM motif can optionally be excluded if desired. Importantly, the gene cassette is devoid of STOP-codon to ensure that excluded if desired. Importantly, the gene cassette is devoid of STOP-codon to ensure that ARM complexes are not disrupted in the presence of E. coli release factors. For a schematic ARM complexes are not disrupted in the presence of E. coli release factors. For a schematic overview of the designed gene cassette, please refer to Figure 12. overview of the designed gene cassette, please refer to Figure 12. Ribosome display involves cell-free extracts for in vitro transcription and translation Ribosome display involves cell-free extracts for in vitro transcription and translation and it was decided to use the E. coli-based PURE extract 182 which is, except from the and it was decided to use the E. coli-based PURE extract 182 which is, except from the ribosomes, prepared from recombinantly or chemically produced and purified components. It ribosomes, prepared from recombinantly or chemically produced and purified components. It is therefore virtually devoid of RNases or proteases and was reported to yield more recovered is therefore virtually devoid of RNases or proteases and was reported to yield more recovered mRNA than classical S30 extracts 183. mRNA than classical S30 extracts 183. Sebastian Grimm 57 Sebastian Grimm 57

Figure 12: Affibody molecule ribosome display system design. Schematic overview of the Figure 12: Affibody molecule ribosome display system design. Schematic overview of the constructed ribosome display system. (a) An expression cassette was assembled in a vector denoted constructed ribosome display system. (a) An expression cassette was assembled in a vector denoted pRD, containing a T7 promoter, a Shine Dalgarno sequence and a FLAG epitope encoding sequence, pRD, containing a T7 promoter, a Shine Dalgarno sequence and a FLAG epitope encoding sequence, followed by a cloning window and a sequence encoding a bipartite spacer element containing a 88- followed by a cloning window and a sequence encoding a bipartite spacer element containing a 88- residue phage M13 protein 3 sequence (p3 spacer) followed by a 21-residue element derived from the residue phage M13 protein 3 sequence (p3 spacer) followed by a 21-residue element derived from the E. coli secretion monitor protein SecM. The Arg residue (in brackets) indicates the extension of the E. coli secretion monitor protein SecM. The Arg residue (in brackets) indicates the extension of the SecM motif by this codon during PCR amplification prior to transcription; (b, c) for initial enrichment SecM motif by this codon during PCR amplification prior to transcription; (b, c) for initial enrichment studies, gene fragments encoding either ZWT-ABD or ZTaq proteins were inserted in the expression studies, gene fragments encoding either ZWT-ABD or ZTaq proteins were inserted in the expression cassette as indicated; (d) for affibody molecule library construction, two oligonucleotides (Zlib-for and cassette as indicated; (d) for affibody molecule library construction, two oligonucleotides (Zlib-for and Zlib-rev) encoding randomized helices 1 and 2 of the three helix bundle scaffold were extended Zlib-rev) encoding randomized helices 1 and 2 of the three helix bundle scaffold were extended together utilizing mutually overlapping sequences for annealing, followed by cloning into the together utilizing mutually overlapping sequences for annealing, followed by cloning into the expression cassette where the extension products replaced a dummy fragment. A gene fragment expression cassette where the extension products replaced a dummy fragment. A gene fragment encoding helix 3 was here already present in the pRD vector. Restriction sites and annealing sites for encoding helix 3 was here already present in the pRD vector. Restriction sites and annealing sites for some of the oligonucleotide primers used during cloning, reverse transcription and PCR work are some of the oligonucleotide primers used during cloning, reverse transcription and PCR work are indicated. indicated.

Another important component are the primers used for reverse transcription of Another important component are the primers used for reverse transcription of recovered mRNA and PCR amplification of cDNA to obtain functional expression cassettes recovered mRNA and PCR amplification of cDNA to obtain functional expression cassettes for the next selection cycle. The reverse transcription primer should be very selectively for the next selection cycle. The reverse transcription primer should be very selectively annealing at temperatures of 50 to 55° C to minimize the reverse transcription of an excess of annealing at temperatures of 50 to 55° C to minimize the reverse transcription of an excess of yeast RNA added to protect recovered affibody-encoding mRNA from degradation. The yeast RNA added to protect recovered affibody-encoding mRNA from degradation. The forward primer for cDNA amplification should be long enough to in a single step re-introduce forward primer for cDNA amplification should be long enough to in a single step re-introduce the T7 promotor. The reverse PCR primer may be shorter. Firstly, nine silent mutations in the the T7 promotor. The reverse PCR primer may be shorter. Firstly, nine silent mutations in the SecM motif described previously 177 were introduced to reduce the G/C content and obtain SecM motif described previously 177 were introduced to reduce the G/C content and obtain cleaner RT-PCR products. Secondly, a short 20 bp primer with a melting temperature of 48° cleaner RT-PCR products. Secondly, a short 20 bp primer with a melting temperature of 48° C was chosen for reverse transcription, followed by longer primers for PCR amplification. C was chosen for reverse transcription, followed by longer primers for PCR amplification. This combination was found to yield clean RT-PCR products that could directly be used for This combination was found to yield clean RT-PCR products that could directly be used for mRNA preparation. It was hypothesized that such a nested RT-PCR primer combination may mRNA preparation. It was hypothesized that such a nested RT-PCR primer combination may further increase the selectivity. further increase the selectivity.

As a proof of the functionality of the system, test enrichment experiments were performed. As a proof of the functionality of the system, test enrichment experiments were performed.

ZWT-ABD and ZTaq-encoding gene cassettes were prepared, that are of different size and from ZWT-ABD and ZTaq-encoding gene cassettes were prepared, that are of different size and from 58 Ribosome display for selection and evolution of affibody molecules 58 Ribosome display for selection and evolution of affibody molecules

which mRNA was transcribed and mixed in a ratio of 1:1000 (ZWT-ABD:ZTaq). The ZWT moiety which mRNA was transcribed and mixed in a ratio of 1:1000 (ZWT-ABD:ZTaq). The ZWT moiety of the fusion ZWT-ABD binds to IgG, while ZTaq does not, and IgG-coated paramagnetic beads of the fusion ZWT-ABD binds to IgG, while ZTaq does not, and IgG-coated paramagnetic beads were used in a biopanning procedure to selectively enrich for the bigger ZWT-ABD-encoding were used in a biopanning procedure to selectively enrich for the bigger ZWT-ABD-encoding construct during two cycles of ribosome display. As documented on an agarose-gel (Figure construct during two cycles of ribosome display. As documented on an agarose-gel (Figure

13), the bigger ZWT-ABD-encoding construct clearly got enriched over two selection cycles, i.e. 13), the bigger ZWT-ABD-encoding construct clearly got enriched over two selection cycles, i.e. proving the functionality of the system. proving the functionality of the system.

Figure 13: Results from initial enrichment experiments. Agarose gel electrophoresis analysis of Figure 13: Results from initial enrichment experiments. Agarose gel electrophoresis analysis of PCR products obtained using DNA templates produced after reverse transcription of mRNA recovered PCR products obtained using DNA templates produced after reverse transcription of mRNA recovered after different numbers of rounds of IgG affinity selection of ternary complexes containing ZWT-ABD after different numbers of rounds of IgG affinity selection of ternary complexes containing ZWT-ABD proteins. Lane 1: marker DNA; lane 2: PCR product obtained after reverse transcription and proteins. Lane 1: marker DNA; lane 2: PCR product obtained after reverse transcription and amplification of an initial 1:1000 mixture of ZWT-ABD mRNA in a background of ZTaq mRNA; lane 3: amplification of an initial 1:1000 mixture of ZWT-ABD mRNA in a background of ZTaq mRNA; lane 3: PCR product obtained after a first round of enrichment; lane 4: PCR product obtained after a second PCR product obtained after a first round of enrichment; lane 4: PCR product obtained after a second round of enrichment. Arrows indicate marker DNA bands and the expected sizes for amplicons round of enrichment. Arrows indicate marker DNA bands and the expected sizes for amplicons corresponding to amplification of ZWT-ABD and ZTaq constructs, respectively. corresponding to amplification of ZWT-ABD and ZTaq constructs, respectively.

Thereupon, an affibody library was constructed from two degenerated Thereupon, an affibody library was constructed from two degenerated oligonucleotides that were extended with each other, and the extension product was inserted oligonucleotides that were extended with each other, and the extension product was inserted in the ribosome display expression cassette (Figure 12). Here, 13 positions on affibody helix in the ribosome display expression cassette (Figure 12). Here, 13 positions on affibody helix one and two described previously 306 were variegated using NN(G/T) degenerated codons. one and two described previously 306 were variegated using NN(G/T) degenerated codons. From the library ligation product, the library size was estimated as 1011. A murine monoclonal From the library ligation product, the library size was estimated as 1011. A murine monoclonal IgG1 was chosen as target protein for ribosome display selection of affibody molecules from IgG1 was chosen as target protein for ribosome display selection of affibody molecules from the newly constructed library. Murine IgG1 is the most common antibody subclass produced the newly constructed library. Murine IgG1 is the most common antibody subclass produced in hybridoma cultures and frequently used in immunoassays. Protein A and protein G are in hybridoma cultures and frequently used in immunoassays. Protein A and protein G are typically used as affinity reagents for the purification of antibodies from such culture typically used as affinity reagents for the purification of antibodies from such culture supernatants. While protein A has no or very weak affinity to this particular antibody supernatants. While protein A has no or very weak affinity to this particular antibody subclass, protein G binds well to mIgG1 but also binds to bovine immunoglobulins that are subclass, protein G binds well to mIgG1 but also binds to bovine immunoglobulins that are present in fetal bovine serum supplemented to the hybridoma culture medium. A protein A- present in fetal bovine serum supplemented to the hybridoma culture medium. A protein A- derived affinity reagent with a selective binding profile for mIgG1 could therefore be a very derived affinity reagent with a selective binding profile for mIgG1 could therefore be a very valuable biotechnological tool for the purification or detection of mIgG1. valuable biotechnological tool for the purification or detection of mIgG1. Sebastian Grimm 59 Sebastian Grimm 59

To avoid the selection against antibody variable domains, three different mIgG1 To avoid the selection against antibody variable domains, three different mIgG1 molecules were chosen as alternating antigens. The selection pressure was gradually increased molecules were chosen as alternating antigens. The selection pressure was gradually increased by decreasing the antigen concentration and increasing the number of bead washing steps. by decreasing the antigen concentration and increasing the number of bead washing steps. After four selection cycles, three affibody clones and one hybrid clone presumably created by After four selection cycles, three affibody clones and one hybrid clone presumably created by recombination during post-selection cloning work were found to be enriched and three of recombination during post-selection cloning work were found to be enriched and three of those bound to all three target mIgG1 on a BIAcore biosensor with low µM to low nM those bound to all three target mIgG1 on a BIAcore biosensor with low µM to low nM affinities. The different affinities to different antibody clones of the same subclass are an affinities. The different affinities to different antibody clones of the same subclass are an interesting finding in itself and may indicate structural differences in the target mIgG1 interesting finding in itself and may indicate structural differences in the target mIgG1 antibodies. Due to the absence of any cysteine residues and the ability to also bind to all target antibodies. Due to the absence of any cysteine residues and the ability to also bind to all target mAbs when immobilized on a biosensor surface, the variant denoted ZMAB25 was chosen for mAbs when immobilized on a biosensor surface, the variant denoted ZMAB25 was chosen for further characterization. This variant showed a very narrow binding preference and further characterization. This variant showed a very narrow binding preference and exclusively bound to mIgG1(κ) among a bigger panel of tested Igs. To map the epitope of exclusively bound to mIgG1(κ) among a bigger panel of tested Igs. To map the epitope of

ZMAB25, a competitive binding assay was conducted, where 10 nM of target mIgG1 mAb3 ZMAB25, a competitive binding assay was conducted, where 10 nM of target mIgG1 mAb3 were incubated with increasing molar excesses of protein G C2-C3 construct. This construct were incubated with increasing molar excesses of protein G C2-C3 construct. This construct has two known epitopes on Igs, one on CH1 domain of the Fab part and one between CH2 and has two known epitopes on Igs, one on CH1 domain of the Fab part and one between CH2 and

CH3 domains on the Fc part and was found to compete with ZMAB25 for binding to mAb3 CH3 domains on the Fc part and was found to compete with ZMAB25 for binding to mAb3 (Figure 14). (Figure 14).

Figure 14: Mapping of ZMAB25 binding site by competitive binding analyses. Possible binding sites Figure 14: Mapping of ZMAB25 binding site by competitive binding analyses. Possible binding sites on mouse IgG1 for the ZMAB25 variant were mapped using a competitive binding analysis employing an on mouse IgG1 for the ZMAB25 variant were mapped using a competitive binding analysis employing an immunoglobulin binding protein with known binding sites. (a) Schematic figure showing an IgG immunoglobulin binding protein with known binding sites. (a) Schematic figure showing an IgG antibody protein with its different regions and the binding sites for SPG in Fc and Fab, respectively, antibody protein with its different regions and the binding sites for SPG in Fc and Fab, respectively, indicated. (b) Biosensor sensorgrams resulting from duplicate injections of 10 nM mAb3 alone (grey indicated. (b) Biosensor sensorgrams resulting from duplicate injections of 10 nM mAb3 alone (grey diamonds); mAb3 in a 100 nM (grey squares), 1 µM (grey circles) or 10 µM (grey triangles) solution diamonds); mAb3 in a 100 nM (grey squares), 1 µM (grey circles) or 10 µM (grey triangles) solution of the recombinant two-domain protein G construct (C2-C3) over a sensorchip surface containing the of the recombinant two-domain protein G construct (C2-C3) over a sensorchip surface containing the

ZMAB25 variant. ZMAB25 variant. 60 Ribosome display for selection and evolution of affibody molecules 60 Ribosome display for selection and evolution of affibody molecules

To exclude one of the two epitopes, mAb3 was papain-digested into Fc and Fab parts To exclude one of the two epitopes, mAb3 was papain-digested into Fc and Fab parts and the product was passed over ZMAB25 immobilized on an affinity column. Western blot and the product was passed over ZMAB25 immobilized on an affinity column. Western blot analysis with detection for antibody Fc or Fab part revealed that the Fab part of mAb3 was analysis with detection for antibody Fc or Fab part revealed that the Fab part of mAb3 was selectively captured by ZMAB25 (Figure 15). Taken together, the experiments suggest that the selectively captured by ZMAB25 (Figure 15). Taken together, the experiments suggest that the epitope of ZMAB25 resides on a site functionally overlapping with protein G C2-C3 fragment on epitope of ZMAB25 resides on a site functionally overlapping with protein G C2-C3 fragment on the antibody Fab part. the antibody Fab part.

Figure 15: Western Blot analysis of mAb3 Fab fragment capture after papain cleavage. The Figure 15: Western Blot analysis of mAb3 Fab fragment capture after papain cleavage. The mouse IgG1 mAb3 was papain-digested and the sample applied on a HiTrap™ HP column with ZMAB25 mouse IgG1 mAb3 was papain-digested and the sample applied on a HiTrap™ HP column with ZMAB25 immobilized as ligand for capture of the targets for the ligand. Panel (a) was stained with a polyclonal immobilized as ligand for capture of the targets for the ligand. Panel (a) was stained with a polyclonal goat anti-mouse IgG Fc HRP conjugate. Lane 1: SeeBlue® Plus2 Pre-Stained standard; lane 2: mAb3 goat anti-mouse IgG Fc HRP conjugate. Lane 1: SeeBlue® Plus2 Pre-Stained standard; lane 2: mAb3 sample before cleavage; lane 3: mAb3 sample after cleavage; lane 4: column flow-through fraction; sample before cleavage; lane 3: mAb3 sample after cleavage; lane 4: column flow-through fraction; lanes 5, 6: eluate fractions. Panel (b) was stained with a polyclonal goat anti-mouse IgG F(ab)2 HRP lanes 5, 6: eluate fractions. Panel (b) was stained with a polyclonal goat anti-mouse IgG F(ab)2 HRP conjugate. Lane 1: SeeBlue® Plus2 Pre-Stained standard; lane 2: mAb3 sample before cleavage; lane conjugate. Lane 1: SeeBlue® Plus2 Pre-Stained standard; lane 2: mAb3 sample before cleavage; lane 3: mAb3 sample after cleavage; lane 4: column flow-through fraction; lanes 5, 6: eluate fractions. The 3: mAb3 sample after cleavage; lane 4: column flow-through fraction; lanes 5, 6: eluate fractions. The black arrows to the left indicate molecular weights in kilo Dalton (kDa). black arrows to the left indicate molecular weights in kilo Dalton (kDa).

In chromatography tests, no binding could be observed between ZMAB25 used as ligand and In chromatography tests, no binding could be observed between ZMAB25 used as ligand and bovine Ig present in an FBS containing sample, indicating a possible use for ZMAB25 in bovine Ig present in an FBS containing sample, indicating a possible use for ZMAB25 in selective capture of mIgG1 from samples containing a bovine Ig background. To test this, selective capture of mIgG1 from samples containing a bovine Ig background. To test this, samples with spiked-in mAb3 or a non-target mIgG1 were prepared that resemble hybridoma samples with spiked-in mAb3 or a non-target mIgG1 were prepared that resemble hybridoma culture supernatants and these samples were passed through columns containing either the culture supernatants and these samples were passed through columns containing either the

“golden standard” streptococcal protein G (SPG) or ZMAB25. In all cases, mIgG1 could be “golden standard” streptococcal protein G (SPG) or ZMAB25. In all cases, mIgG1 could be captured and eluted. Biosensor analysis of the eluate fractions revealed, however, that the captured and eluted. Biosensor analysis of the eluate fractions revealed, however, that the sample eluted from the SPG column contained detectable amounts of bovine sample eluted from the SPG column contained detectable amounts of bovine immunoglobulins, while samples eluted from the ZMAB25 column did not (Figure 16), i.e. immunoglobulins, while samples eluted from the ZMAB25 column did not (Figure 16), i.e. demonstrating the applicability of ZMAB25 as ligand for the selective capture of two different demonstrating the applicability of ZMAB25 as ligand for the selective capture of two different mouse IgG1 mAbs from samples resembling hybridoma supernatants and containing FBS. mouse IgG1 mAbs from samples resembling hybridoma supernatants and containing FBS. Sebastian Grimm 61 Sebastian Grimm 61

Taken together, this study showed that affibody molecules can be selected from a Taken together, this study showed that affibody molecules can be selected from a combinatorial library using ribosome display. The selection of very closely-related variants, combinatorial library using ribosome display. The selection of very closely-related variants, as would be expected from the large library size, was however not observed. This could be of as would be expected from the large library size, was however not observed. This could be of two reasons. Firstly, the library size on the level of functional ARM complexes was smaller two reasons. Firstly, the library size on the level of functional ARM complexes was smaller than estimated or secondly, the switching of antigens during selection constrained the than estimated or secondly, the switching of antigens during selection constrained the diversity of selectable clones. diversity of selectable clones.

Furthermore, the affibody molecule ZMAB25 was selected that bound to several murine Furthermore, the affibody molecule ZMAB25 was selected that bound to several murine mIgG1 in a highly species-selective manner and may become a valuable tool for mIgG1 mIgG1 in a highly species-selective manner and may become a valuable tool for mIgG1 detection or purification. detection or purification.

Figure 16: Binding studies of ZMAB25 or protein G with fetal bovine serum and species-selective Figure 16: Binding studies of ZMAB25 or protein G with fetal bovine serum and species-selective affinity recovery of two monoclonal antibodies. (a) SDS-PAGE analysis (reducing conditions) of affinity recovery of two monoclonal antibodies. (a) SDS-PAGE analysis (reducing conditions) of samples, flow through (FT) and eluate (E) fractions from affinity chromatography experiments using samples, flow through (FT) and eluate (E) fractions from affinity chromatography experiments using either a protein G (SPG) column or a ZMAB25 column, and samples 10% FBS with in-spiked mouse either a protein G (SPG) column or a ZMAB25 column, and samples 10% FBS with in-spiked mouse IgG1 mAb3 (lanes 1-6) or 10% FBS with in-spiked mouse IgG1 8F11 (lanes 7-9). M: LMW-SDS IgG1 mAb3 (lanes 1-6) or 10% FBS with in-spiked mouse IgG1 8F11 (lanes 7-9). M: LMW-SDS molecular weight marker; lane 1: 10% FBS + mAb3 sample; lane 2: SPG column, FT; lane 3: SPG molecular weight marker; lane 1: 10% FBS + mAb3 sample; lane 2: SPG column, FT; lane 3: SPG column, E; lane 4: 10% FBS + mAb3 sample; lane 5: ZMAB25 column, FT; lane 6: ZMAB25 column, E; column, E; lane 4: 10% FBS + mAb3 sample; lane 5: ZMAB25 column, FT; lane 6: ZMAB25 column, E; lane 7: 10% FBS + mAb 8F11 sample; lane 8: ZMAB25 column, FT; lane 9: ZMAB25 column, E. The black lane 7: 10% FBS + mAb 8F11 sample; lane 8: ZMAB25 column, FT; lane 9: ZMAB25 column, E. The black arrows to the right indicate nominal molecular weights of antibody heavy and light chains, arrows to the right indicate nominal molecular weights of antibody heavy and light chains, respectively. The numbers to the left indicate molecular weights in kilo Dalton (kDa). (b) Results from respectively. The numbers to the left indicate molecular weights in kilo Dalton (kDa). (b) Results from a biosensor analysis of pH neutralized eluates from the SPG and ZMAB25 columns, originating from the a biosensor analysis of pH neutralized eluates from the SPG and ZMAB25 columns, originating from the use of the 10% FBS sample containing in-spiked mAb3 (1-4) or mAb 8F11 (5, 6). On separate flow use of the 10% FBS sample containing in-spiked mAb3 (1-4) or mAb 8F11 (5, 6). On separate flow cell surfaces, anti-cow IgG and anti-mouse Ig antibodies were immobilized, and samples were cell surfaces, anti-cow IgG and anti-mouse Ig antibodies were immobilized, and samples were injected. (1), black solid trace: SPG column eluate, anti-cow IgG surface; (2), black dashed trace: SPG injected. (1), black solid trace: SPG column eluate, anti-cow IgG surface; (2), black dashed trace: SPG column eluate, anti-mouse Ig surface; (3), grey solid trace: ZMAB25 column eluate, anti-cow IgG column eluate, anti-mouse Ig surface; (3), grey solid trace: ZMAB25 column eluate, anti-cow IgG surface; (4), grey dashed trace: ZMAB25 column eluate, anti-mouse Ig surface; (5), black dotted trace: surface; (4), grey dashed trace: ZMAB25 column eluate, anti-mouse Ig surface; (5), black dotted trace:

ZMAB25 column eluate, anti-cow IgG surface; (6), grey dotted trace: ZMAB25 column eluate, anti-mouse ZMAB25 column eluate, anti-cow IgG surface; (6), grey dotted trace: ZMAB25 column eluate, anti-mouse Ig surface. Ig surface.

62 Ribosome display for selection and evolution of affibody molecules 62 Ribosome display for selection and evolution of affibody molecules

4.2 Single domain affinity proteins for the detection of the genome 4.2 Single domain affinity proteins for the detection of the genome organizer protein SATB1 (II) organizer protein SATB1 (II)

The special AT-rich sequence binding protein 1 (SATB1) is primarily expressed at high levels The special AT-rich sequence binding protein 1 (SATB1) is primarily expressed at high levels in thymocytes 307 and was reported to form cage-like network structures in the nuclei of these in thymocytes 307 and was reported to form cage-like network structures in the nuclei of these cells 308. SATB1 includes an N-terminal PDZ domain for homodimerization, followed by two cells 308. SATB1 includes an N-terminal PDZ domain for homodimerization, followed by two CUT domains and a C-terminal homeobox that together orchestrate DNA-binding to so-called CUT domains and a C-terminal homeobox that together orchestrate DNA-binding to so-called matrix attachment regions 309. In a cell biological context, Cai and colleagues found that matrix attachment regions 309. In a cell biological context, Cai and colleagues found that SATB1 is tethering chromatin into loops 308 and recruiting chromatin-remodeling factors to SATB1 is tethering chromatin into loops 308 and recruiting chromatin-remodeling factors to control gene transcription in T cells 307. In a disease context, SATB1 was by the same group control gene transcription in T cells 307. In a disease context, SATB1 was by the same group reported to promote breast cancer tumor growth and metastasis. A significant correlation reported to promote breast cancer tumor growth and metastasis. A significant correlation between SATB1 expression in human adenocarcinoma tissue specimens and poor prognosis between SATB1 expression in human adenocarcinoma tissue specimens and poor prognosis for patient survival was found 310. The prognostic relevance of SATB1 was however for patient survival was found 310. The prognostic relevance of SATB1 was however questioned in two recent and independent studies of SATB1 mRNA expression levels in questioned in two recent and independent studies of SATB1 mRNA expression levels in breast cancer cell lines or patient samples 311; 312. breast cancer cell lines or patient samples 311; 312. The present study was performed as part of an EU framework project aiming at the The present study was performed as part of an EU framework project aiming at the application of ultra-high resolution fluorescence microscopy techniques 313 for the detection application of ultra-high resolution fluorescence microscopy techniques 313 for the detection and diagnosis of breast or prostate cancer. To further elucidate the role of SATB1 in breast and diagnosis of breast or prostate cancer. To further elucidate the role of SATB1 in breast cancer and possibly harness SATB1 structural features for patient stratification using cancer and possibly harness SATB1 structural features for patient stratification using Stimulated Emission Depletion (STED) microscopy 314, very small and SATB1-selective Stimulated Emission Depletion (STED) microscopy 314, very small and SATB1-selective affinity reagents would be very suitable tools. STED microscopy approaches resolutions affinity reagents would be very suitable tools. STED microscopy approaches resolutions where the affinity reagent becomes limiting to the resolution and fluorescently labelled, small where the affinity reagent becomes limiting to the resolution and fluorescently labelled, small affinity proteins may be superior to comparably large antibodies or even larger complexes of affinity proteins may be superior to comparably large antibodies or even larger complexes of primary and labelled secondary antibodies. Furthermore, small affinity proteins may be more primary and labelled secondary antibodies. Furthermore, small affinity proteins may be more suitable for accessing fine structures in cells that antibodies. suitable for accessing fine structures in cells that antibodies. This work describes the selection and characterization of two classes of small single domain This work describes the selection and characterization of two classes of small single domain affinity proteins binding to SATB1: affibody molecules and VHH antibody fragments (see affinity proteins binding to SATB1: affibody molecules and VHH antibody fragments (see section 3.2.4). Affibody molecules were selected from the library described in (I), mixed with section 3.2.4). Affibody molecules were selected from the library described in (I), mixed with an aliquot of a novel library including also a 14th NN(G/T)-randomized Ile31 position, using an aliquot of a novel library including also a 14th NN(G/T)-randomized Ile31 position, using four consecutive rounds of ribosome display. Ile31 is located in the middle of the affibody four consecutive rounds of ribosome display. Ile31 is located in the middle of the affibody binding face and may constrain addressable epitopes on the antigen. VHH antibody fragments binding face and may constrain addressable epitopes on the antigen. VHH antibody fragments were selected from a naïve, natural repertoire using two consecutive rounds of phage display. were selected from a naïve, natural repertoire using two consecutive rounds of phage display. The targets used for selection were recombinant SATB1 fragments SATB1(aa1-360) and The targets used for selection were recombinant SATB1 fragments SATB1(aa1-360) and SATB1(aa361-763) for VHH selection and SATB1(aa346-763) for affibody selection (Figure SATB1(aa361-763) for VHH selection and SATB1(aa346-763) for affibody selection (Figure 17). 17).

Sebastian Grimm 63 Sebastian Grimm 63

Figure 17: Schematic representation of the special AT-rich sequence binding protein 1. From left Figure 17: Schematic representation of the special AT-rich sequence binding protein 1. From left to right: the PDZ-like dimerization domain (PDZ); the two DNA-binding CUT domains (CUT1, 2) to right: the PDZ-like dimerization domain (PDZ); the two DNA-binding CUT domains (CUT1, 2) and the DNA-binding homeobox domain (HB). Recombinant fragments of SATB1 used in this work and the DNA-binding homeobox domain (HB). Recombinant fragments of SATB1 used in this work are indicted with black arrows and corresponding amino acid sequence positions are given in brackets. are indicted with black arrows and corresponding amino acid sequence positions are given in brackets.

Selected clones from both classes were tested for concentration-dependent binding to Selected clones from both classes were tested for concentration-dependent binding to their respective antigens. Eleven VHH antibody fragments and two affibody molecules were their respective antigens. Eleven VHH antibody fragments and two affibody molecules were found to bind to the C-terminal half of SATB1 and were further analyzed for binding to any found to bind to the C-terminal half of SATB1 and were further analyzed for binding to any of the three DNA-binding domains contained within this fragment. Eight of the 11 tested of the three DNA-binding domains contained within this fragment. Eight of the 11 tested VHHs recognized one of the three domains, while the two affibody molecules did not bind to VHHs recognized one of the three domains, while the two affibody molecules did not bind to any of these domains, suggesting that their epitopes are located outside of the tested domains any of these domains, suggesting that their epitopes are located outside of the tested domains (Figure 18). (Figure 18).

Figure 18: ELISA mapping and competition studies of selected VHH and affibody molecules. Figure 18: ELISA mapping and competition studies of selected VHH and affibody molecules. Absorbance obtained using single concentrations of VHH or affibody clones in an ELISA assay Absorbance obtained using single concentrations of VHH or affibody clones in an ELISA assay involving different SATB1 fragments containing the first cut domain (cut1), the first and the second involving different SATB1 fragments containing the first cut domain (cut1), the first and the second cut domain (cut1 + cut2), the homeobox, SATB1(aa346-763) or a control protein as coating antigens. cut domain (cut1 + cut2), the homeobox, SATB1(aa346-763) or a control protein as coating antigens. VHH clones are grouped according to epitope preference, as denoted in diagonal text. VHH clones are grouped according to epitope preference, as denoted in diagonal text. 64 Ribosome display for selection and evolution of affibody molecules 64 Ribosome display for selection and evolution of affibody molecules

Since SATB1 is highly homologous in sequence to the close relative SATB2, the Since SATB1 is highly homologous in sequence to the close relative SATB2, the same 11 VHHs and two affibody molecules were tested for binding selectivity between same 11 VHHs and two affibody molecules were tested for binding selectivity between SATB1 and SATB2 in an ELISA assay. Most of the VHHs that recognized one of the three SATB1 and SATB2 in an ELISA assay. Most of the VHHs that recognized one of the three conserved DNA-binding domains showed cross-reactivity to SATB2. This is not surprising, conserved DNA-binding domains showed cross-reactivity to SATB2. This is not surprising, since these domains are up to 85% identical in sequence between SATB1 and SATB2. since these domains are up to 85% identical in sequence between SATB1 and SATB2.

Among the two affibody clones, ZSATB1:2 was found to selectively bind to SATB1, while Among the two affibody clones, ZSATB1:2 was found to selectively bind to SATB1, while

ZSATB1:1 showed some cross-reactivity to SATB2 (Figure 19a). ZSATB1:1 showed some cross-reactivity to SATB2 (Figure 19a).

Figure 19: Binding selectivity to SATB1 or SATB2 and biosensor affinity determination. (a) Figure 19: Binding selectivity to SATB1 or SATB2 and biosensor affinity determination. (a) ELISA absorbance from different selected VHH or affibody molecules binding to either ELISA absorbance from different selected VHH or affibody molecules binding to either SATB1(aa346-763) (grey bars), SATB2(339-729) (black bars) or a control protein (light grey bars) SATB1(aa346-763) (grey bars), SATB2(339-729) (black bars) or a control protein (light grey bars) are shown. (b) Different concentrations of VHH 2D2 (black dots), 4D10 (black squares), 4E2 (grey are shown. (b) Different concentrations of VHH 2D2 (black dots), 4D10 (black squares), 4E2 (grey triangles up), 4F6 (grey triangles down), 4H1(grey diamonds), ZSATB1:1 (grey dots) or ZSATB1:2 (grey triangles up), 4F6 (grey triangles down), 4H1(grey diamonds), ZSATB1:1 (grey dots) or ZSATB1:2 (grey squares) were flowed over a SATB1(aa1-360) (2D2) or SATB1(aa346-763) (all other binders) surface. squares) were flowed over a SATB1(aa1-360) (2D2) or SATB1(aa346-763) (all other binders) surface. The obtained equilibrium binding responses were plotted against analyte concentration and fitted (grey The obtained equilibrium binding responses were plotted against analyte concentration and fitted (grey or black traces) for Kd determination. A table of the obtained equilibrium dissociation constants is or black traces) for Kd determination. A table of the obtained equilibrium dissociation constants is shown to the right. shown to the right. Sebastian Grimm 65 Sebastian Grimm 65

The selectivity of all the 13 clones of both binder classes selected against the C- The selectivity of all the 13 clones of both binder classes selected against the C- terminal half and an additional four VHH clones selected against the N-terminal half of terminal half and an additional four VHH clones selected against the N-terminal half of SATB1 was further evaluated in IF co-localization studies with GFP-SATB1 or GFP-SATB2 SATB1 was further evaluated in IF co-localization studies with GFP-SATB1 or GFP-SATB2 expressed in MCF7 breast cancer cells, respectively. The ELISA SATB1/SATB2 binding expressed in MCF7 breast cancer cells, respectively. The ELISA SATB1/SATB2 binding selectivity profiles could in all but one case be confirmed in this assay. Among the four VHH selectivity profiles could in all but one case be confirmed in this assay. Among the four VHH antibody fragments selected against the N-terminal half of SATB1, two clones were binding antibody fragments selected against the N-terminal half of SATB1, two clones were binding selectively to SATB1 in this assay. Example of excellent IF cell staining results of VHH 2D2, selectively to SATB1 in this assay. Example of excellent IF cell staining results of VHH 2D2,

VHH 4D10 and ZSATB1:2, backed-up with flow cytometric analyses of corresponding cell VHH 4D10 and ZSATB1:2, backed-up with flow cytometric analyses of corresponding cell lysate pullouts, are shown in Figure 20. Notably, VHH 4D10 recognized both SATB1 and lysate pullouts, are shown in Figure 20. Notably, VHH 4D10 recognized both SATB1 and SATB2. SATB2. Based on these initial binding and selectivity analyses, eight VHH antibody fragments Based on these initial binding and selectivity analyses, eight VHH antibody fragments and the two affibody molecules were chosen for affinity determination on a BIAcore and the two affibody molecules were chosen for affinity determination on a BIAcore biosensor. Due to a biphasic binding behavior, VHH affinities should be considered as biosensor. Due to a biphasic binding behavior, VHH affinities should be considered as apparent and were ranging from 81 nM to 2.4 µM. Three VHH clones were non-selectively apparent and were ranging from 81 nM to 2.4 µM. Three VHH clones were non-selectively binding in this assay and affinities could not be determined. Affibody molecules ZSATB1:1 and binding in this assay and affinities could not be determined. Affibody molecules ZSATB1:1 and

ZSATB1:2 were binding with 807 and 320 nM equilibrium dissociation constants, respectively ZSATB1:2 were binding with 807 and 320 nM equilibrium dissociation constants, respectively (Figure 19b). (Figure 19b).

Also considering equilibrium dissociation constants, VHH 2D2 and ZSATB1:2 were Also considering equilibrium dissociation constants, VHH 2D2 and ZSATB1:2 were chosen as lead candidates for the detection of endogenous SATB1 in Jurkat T cells in two chosen as lead candidates for the detection of endogenous SATB1 in Jurkat T cells in two different experiments. Firstly, in a pull-out study of SATB1 from Jurkat cell lysate, followed different experiments. Firstly, in a pull-out study of SATB1 from Jurkat cell lysate, followed by detection of SATB1 on a western blot. Secondly, in a co-localization study of VHH 2D2, by detection of SATB1 on a western blot. Secondly, in a co-localization study of VHH 2D2,

ZSATB1:2 and a SATB1-selective polyclonal control antibody in Jurkat cells, using fluorescence ZSATB1:2 and a SATB1-selective polyclonal control antibody in Jurkat cells, using fluorescence microscopy. Interestingly, ZSATB1:2 specifically precipitated a single band also recognized by a microscopy. Interestingly, ZSATB1:2 specifically precipitated a single band also recognized by a SATB1-selective detection antibody, while VHH 2D2 precipitated three bands also SATB1-selective detection antibody, while VHH 2D2 precipitated three bands also recognized by the detection antibody. The presence of several bands may indicate splice recognized by the detection antibody. The presence of several bands may indicate splice variants or other, cross-reactive proteins. Using fluorescence microscopy, a clear, nuclear co- variants or other, cross-reactive proteins. Using fluorescence microscopy, a clear, nuclear co- localization between the SATB1-selective antibody and ZSATB1:2 was observed. The co- localization between the SATB1-selective antibody and ZSATB1:2 was observed. The co- localization between VHH 2D2 and ZSATB1:2 was not as clear, since VHH 2D2 also showed localization between VHH 2D2 and ZSATB1:2 was not as clear, since VHH 2D2 also showed some cytoplasmic staining (Figure 21). some cytoplasmic staining (Figure 21).

Taken together, these findings suggest that ZSATB1:2 and VHH 2D2 are two promising Taken together, these findings suggest that ZSATB1:2 and VHH 2D2 are two promising reagents for the detection of endogenous SATB1 using ultra-high resolution microscopy reagents for the detection of endogenous SATB1 using ultra-high resolution microscopy techniques. The unique cysteine residue of ZSATB1:2 allows for site-specific labelling with a techniques. The unique cysteine residue of ZSATB1:2 allows for site-specific labelling with a fluorophore and hence the use of ZSATB1:2 for detection without the need of any secondary fluorophore and hence the use of ZSATB1:2 for detection without the need of any secondary detection reagent. Due to its moderate affinity, ZSATB1:2 was applied as a head-tail trimer to detection reagent. Due to its moderate affinity, ZSATB1:2 was applied as a head-tail trimer to harness avidity effects and may need to be affinity matured for usage as a monomer. harness avidity effects and may need to be affinity matured for usage as a monomer. 66 Ribosome display for selection and evolution of affibody molecules 66 Ribosome display for selection and evolution of affibody molecules

Figure 20: Binding analysis to GFP-SATB1 and GFP-SATB2 and IF co-localization studies. (a) Figure 20: Binding analysis to GFP-SATB1 and GFP-SATB2 and IF co-localization studies. (a)

VHH 2D2, VHH 4D10, ZSATB1:2 or control affibody coated beads were incubated with GFP-SATB1 VHH 2D2, VHH 4D10, ZSATB1:2 or control affibody coated beads were incubated with GFP-SATB1 containing cell lysate (grey filled histograms), GFP-SATB2 containing cell lysate (grey histograms) or containing cell lysate (grey filled histograms), GFP-SATB2 containing cell lysate (grey histograms) or cell lysate (black histograms). Bead counts are plotted against GFP fluorescence in a logarithmic cell lysate (black histograms). Bead counts are plotted against GFP fluorescence in a logarithmic scale. (b) MCF7 cells expressing GFP-SATB1 or GFP-SATB2 (green, middle panels) were stained scale. (b) MCF7 cells expressing GFP-SATB1 or GFP-SATB2 (green, middle panels) were stained with DAPI (blue, left panels) and fluorescently labeled VHH 2D2, VHH 4D10, ZSATB1:2 or VHH with DAPI (blue, left panels) and fluorescently labeled VHH 2D2, VHH 4D10, ZSATB1:2 or VHH control (red, right panels). control (red, right panels). Sebastian Grimm 67 Sebastian Grimm 67

Figure 21: Binding analysis of endogenous SATB1 in Jurkat cell lysate and in vivo co-localization Figure 21: Binding analysis of endogenous SATB1 in Jurkat cell lysate and in vivo co-localization studies in Jurkat cells. (a) Cell lysates or immunoprecipitates (IP) and the corresponding non-bound studies in Jurkat cells. (a) Cell lysates or immunoprecipitates (IP) and the corresponding non-bound fractions (NB) originating from ZSATB1:2, affibody control, VHH 2D2 or VHH control coated beads, fractions (NB) originating from ZSATB1:2, affibody control, VHH 2D2 or VHH control coated beads, respectively, were blotted and the membrane was stained for SATB1 with rabbit anti-SATB1 antibody respectively, were blotted and the membrane was stained for SATB1 with rabbit anti-SATB1 antibody PRS4679. Molecular weights in kDa are indicated to the left. (b) Jurkat cells were co-stained with PRS4679. Molecular weights in kDa are indicated to the left. (b) Jurkat cells were co-stained with either ZSATB1:2 (red), VHH 2D2 (green) and DAPI (first row) or ZSATB1:2 (red), rabbit anti-SATB1 either ZSATB1:2 (red), VHH 2D2 (green) and DAPI (first row) or ZSATB1:2 (red), rabbit anti-SATB1 antibody PRS4679 (green) and DAPI (second row) and analyzed by fluorescence microscopy. antibody PRS4679 (green) and DAPI (second row) and analyzed by fluorescence microscopy. 68 Ribosome display for selection and evolution of affibody molecules 68 Ribosome display for selection and evolution of affibody molecules

4.3 Selection and characterization of affibody molecules inhibiting the 4.3 Selection and characterization of affibody molecules inhibiting the interaction between Ras and Raf in vitro (III) interaction between Ras and Raf in vitro (III)

Mitogen-activated protein kinase (MAPK) signaling cascades regulate cell proliferation, Mitogen-activated protein kinase (MAPK) signaling cascades regulate cell proliferation, survival and differentiation in eukaryotes and can be involved in the development of diseases survival and differentiation in eukaryotes and can be involved in the development of diseases such as cancer. Of particular interest is the extracellular signal-regulated kinase (ERK) such as cancer. Of particular interest is the extracellular signal-regulated kinase (ERK) MAPK pathway which includes Ras GTPases and their downstream mediators, Raf MAPK pathway which includes Ras GTPases and their downstream mediators, Raf serine/threonine kinases. Both protein families are frequently found mutated in e.g. pancreatic serine/threonine kinases. Both protein families are frequently found mutated in e.g. pancreatic cancer, colon cancer, non-small lung cancer (Ras) or melanoma, papillary thyroid cancer, cancer, colon cancer, non-small lung cancer (Ras) or melanoma, papillary thyroid cancer, colon cancer (Raf) 315. colon cancer (Raf) 315.

Figure 22: The (ERK) MAPK pathway. A growth factor (GF) mediates growth factor receptor Figure 22: The (ERK) MAPK pathway. A growth factor (GF) mediates growth factor receptor (GFR) dimerization and phosphorylation (P). Adaptor proteins recruit guanine nucleotide exchange (GFR) dimerization and phosphorylation (P). Adaptor proteins recruit guanine nucleotide exchange factors (GEF) to the plasma membrane. These accelerate the exchange of Ras-bound GDP with GTP, factors (GEF) to the plasma membrane. These accelerate the exchange of Ras-bound GDP with GTP, Ras changes conformation and becomes active. GTPase-activating proteins (GAP) catalyze the Ras changes conformation and becomes active. GTPase-activating proteins (GAP) catalyze the hydrolysis of GTP and return Ras in its non-active conformation. Active Ras recruits Raf kinase to the hydrolysis of GTP and return Ras in its non-active conformation. Active Ras recruits Raf kinase to the plasma membrane where it becomes phosphorylated, activated and initiates the MAPK plasma membrane where it becomes phosphorylated, activated and initiates the MAPK phosphorylation cascade. The signal is relayed into the nucleus where transcription factors (TF) phosphorylation cascade. The signal is relayed into the nucleus where transcription factors (TF) induce the expression of a number of genes that orchestrate cell differentiation, proliferation and induce the expression of a number of genes that orchestrate cell differentiation, proliferation and growth. growth. Sebastian Grimm 69 Sebastian Grimm 69

In response to extracellular stimuli such as growth factors binding to growth factor In response to extracellular stimuli such as growth factors binding to growth factor receptors, guanine nucleotide-exchange factors (GEF) get recruited to the plasma membrane, receptors, guanine nucleotide-exchange factors (GEF) get recruited to the plasma membrane, where they accelerate the release of GDP from membrane-anchored Ras. Ras proteins thereon where they accelerate the release of GDP from membrane-anchored Ras. Ras proteins thereon bind the more abundant GTP and switch to their biologically active conformation. Active Ras bind the more abundant GTP and switch to their biologically active conformation. Active Ras proteins recruit Raf proteins to the plasma membrane, where other mediators induce their proteins recruit Raf proteins to the plasma membrane, where other mediators induce their kinase activity. Raf proteins forward the signal to the MAPK pathway, leading to cell kinase activity. Raf proteins forward the signal to the MAPK pathway, leading to cell differentiation, proliferation and growth 316. Notably, this description of the (ERK) MAPK differentiation, proliferation and growth 316. Notably, this description of the (ERK) MAPK pathway is greatly simplified. pathway is greatly simplified.

Ras proteins include H-Ras, N-Ras and K-Ras and posses an internal GTPase activity Ras proteins include H-Ras, N-Ras and K-Ras and posses an internal GTPase activity which limits the lifetime of their GTP-bound, active state. Mutations can abrogate this which limits the lifetime of their GTP-bound, active state. Mutations can abrogate this GTPase activity, render Ras proteins constitutively active and cause un-controlled cell GTPase activity, render Ras proteins constitutively active and cause un-controlled cell growth, division and ultimately cancer 316. Raf proteins are kinases and include Raf-1, A-Raf growth, division and ultimately cancer 316. Raf proteins are kinases and include Raf-1, A-Raf and B-Raf which can form heterodimers 317 and share two regions involved in Ras binding: and B-Raf which can form heterodimers 317 and share two regions involved in Ras binding: the Ras-binding domain (RBD) and a cysteine-rich domain (CRD) 318. the Ras-binding domain (RBD) and a cysteine-rich domain (CRD) 318. Affinity reagents that selectively inhibit certain protein-protein interactions inside of Affinity reagents that selectively inhibit certain protein-protein interactions inside of the cell may be valuable tools for the understanding of cell biology, serve as leads for the the cell may be valuable tools for the understanding of cell biology, serve as leads for the design of small molecules or may in future be used as inhibitory molecules themselves after design of small molecules or may in future be used as inhibitory molecules themselves after gene transfer (gene therapy). Due to the dependence of antibody and antibody fragment gene transfer (gene therapy). Due to the dependence of antibody and antibody fragment structures on disulphide bonds, these protein classes are not very suitable for an intracellular structures on disulphide bonds, these protein classes are not very suitable for an intracellular application in a reducing environment that may disrupt disulphide bonds 319. Affibody application in a reducing environment that may disrupt disulphide bonds 319. Affibody molecules however do not contain any scaffold-stabilizing disulphide bonds and should retain molecules however do not contain any scaffold-stabilizing disulphide bonds and should retain their binding properties in the cellular cytoplasm. their binding properties in the cellular cytoplasm. In the present study, the possibility to select affibody molecules that selectively bind In the present study, the possibility to select affibody molecules that selectively bind to H-Ras and the RBD of Raf-1 was investigated, using a 3x109-sized affibody molecule to H-Ras and the RBD of Raf-1 was investigated, using a 3x109-sized affibody molecule library displayed on phage. A fragment of H-Ras was chosen as antigen, that includes all library displayed on phage. A fragment of H-Ras was chosen as antigen, that includes all necessary parts for its GTPase and Raf-binding activity but omits the membrane-anchored and necessary parts for its GTPase and Raf-binding activity but omits the membrane-anchored and cysteine-containing C-terminal amino acids. A fragment of Raf-1 (ABP-hRaf-15-143) was cysteine-containing C-terminal amino acids. A fragment of Raf-1 (ABP-hRaf-15-143) was chosen as antigen that contains the RBD but omits other domains, to direct the selection chosen as antigen that contains the RBD but omits other domains, to direct the selection against the RBD and hence increase the chance of selecting binders that interfere with the against the RBD and hence increase the chance of selecting binders that interfere with the

Ras-Raf interaction. Both fragments were produced and purified as His6-ABP fusion proteins, Ras-Raf interaction. Both fragments were produced and purified as His6-ABP fusion proteins, and tested for functional folding in a biosensor binding assay. Phages were panned against and tested for functional folding in a biosensor binding assay. Phages were panned against biotinylated H-Ras, either in solution (track 1) or as immobilized onto paramagnetic beads biotinylated H-Ras, either in solution (track 1) or as immobilized onto paramagnetic beads

(track 2), while phages were panned against ABP-hRaf-15-143 immobilized on beads, only (track 2), while phages were panned against ABP-hRaf-15-143 immobilized on beads, only (track 3). After four rounds of biopanning, clones were sequenced, and based on their amino (track 3). After four rounds of biopanning, clones were sequenced, and based on their amino acid sequence and the frequency of their appearance, which indicates their enrichment in the acid sequence and the frequency of their appearance, which indicates their enrichment in the clone population, one affibody clone from each selection track was produced and purified for clone population, one affibody clone from each selection track was produced and purified for further analysis. A selective binding of variants ZRAS122 and ZRAS220 (selected against H-Ras) further analysis. A selective binding of variants ZRAS122 and ZRAS220 (selected against H-Ras) 70 Ribosome display for selection and evolution of affibody molecules 70 Ribosome display for selection and evolution of affibody molecules

as well as ZRAF322 (selected against ABP-hRaf-15-143) to their respective antigen was observed as well as ZRAF322 (selected against ABP-hRaf-15-143) to their respective antigen was observed in a biosensor binding assay (Figure 23) and the selectivity was further confirmed in a dot blot in a biosensor binding assay (Figure 23) and the selectivity was further confirmed in a dot blot binding assay including 17 human control proteins. Notably, both ZRAS122 and ZRAS220 contain binding assay including 17 human control proteins. Notably, both ZRAS122 and ZRAS220 contain a cysteine residue and can form cysteine-bridged homodimers under non-reducing conditions, a cysteine residue and can form cysteine-bridged homodimers under non-reducing conditions, which may explain the biphasic dissociation seen in the biosensor traces. The apparent which may explain the biphasic dissociation seen in the biosensor traces. The apparent affinities of ZRAS122 and ZRAS220 head-tail dimers under reducing conditions were determined affinities of ZRAS122 and ZRAS220 head-tail dimers under reducing conditions were determined as 283 and 79 nM, respectively. The affinity of the ZRAF322 monomer was determined as 1.9 as 283 and 79 nM, respectively. The affinity of the ZRAF322 monomer was determined as 1.9 µM (Figure 23). µM (Figure 23).

Figure 23: Real-time interaction analysis of the selected affibody ligands. (A-C) Overlay of Figure 23: Real-time interaction analysis of the selected affibody ligands. (A-C) Overlay of sensorgrams obtained after injection of his6-ZRAS122 (A) or his6-ZRAS220 (B) at a concentration of 1 µM, sensorgrams obtained after injection of his6-ZRAS122 (A) or his6-ZRAS220 (B) at a concentration of 1 µM, over surfaces with immobilized his6-ABP-HRas (solid black line), his6-ABP (solid gray line) or human over surfaces with immobilized his6-ABP-HRas (solid black line), his6-ABP (solid gray line) or human

IgG (dotted black line). (C) Overlay of sensorgram obtained after injection of his6- ZRAF322 at a IgG (dotted black line). (C) Overlay of sensorgram obtained after injection of his6- ZRAF322 at a concentration of 1 µM over either a surface with immobilized ABP-hRaf-15-143 (solid black line), his6- concentration of 1 µM over either a surface with immobilized ABP-hRaf-15-143 (solid black line), his6- ABP (solid gray line) or human IgG (dotted black line). (D) Overlay of sensorgrams obtained after ABP (solid gray line) or human IgG (dotted black line). (D) Overlay of sensorgrams obtained after injection of ZRAS122 at different concentrations over a flow cell with immobilized his6-ABP-HRas: 125 injection of ZRAS122 at different concentrations over a flow cell with immobilized his6-ABP-HRas: 125 nM ( ), 250 nM ( ), 500 nM ( ), 1 µM ( ), 2 µM ( ) and 4 µM ( ). (E) Overlay of sensorgrams nM ( ), 250 nM ( ), 500 nM ( ), 1 µM ( ), 2 µM ( ) and 4 µM ( ). (E) Overlay of sensorgrams obtained after injection of ZRAS220 at different concentrations over a flow cell with immobilized his6- obtained after injection of ZRAS220 at different concentrations over a flow cell with immobilized his6- ABP-HRas: 31 nM ( ), 63 nM ( ), 125 nM ( ), 250 nM ( ), 500 nM ( ) and 1 µM ( ). (F) Overlay of ABP-HRas: 31 nM ( ), 63 nM ( ), 125 nM ( ), 250 nM ( ), 500 nM ( ) and 1 µM ( ). (F) Overlay of sensorgrams obtained after injection of his6- ZRAF322 at different concentrations over a flow cell with sensorgrams obtained after injection of his6- ZRAF322 at different concentrations over a flow cell with immobilized ABP-hRaf-15-143: 125 nM ( ), 250 nM ( ), 500 nM ( ), 1 µM ( ), 2 µM ( ), 4 µM ( ), 8 immobilized ABP-hRaf-15-143: 125 nM ( ), 250 nM ( ), 500 nM ( ), 1 µM ( ), 2 µM ( ), 4 µM ( ), 8 µM ( ), 16 µM ( ) and 32 µM ( ) (G) Equilibrium responses obtained in (F) was plotted against the µM ( ), 16 µM ( ) and 32 µM ( ) (G) Equilibrium responses obtained in (F) was plotted against the affibody concentration used and a sigmoidal shaped curve was fitted to the results. affibody concentration used and a sigmoidal shaped curve was fitted to the results.

Further characterizations in competition assays showed that ZRAS122 competes for Further characterizations in competition assays showed that ZRAS122 competes for binding to H-Ras with ZRAS220, which suggests that the epitopes of both binders overlay with binding to H-Ras with ZRAS220, which suggests that the epitopes of both binders overlay with Sebastian Grimm 71 Sebastian Grimm 71 each other. Finally, it was tested if any of the three binders could block the interaction each other. Finally, it was tested if any of the three binders could block the interaction between the RBD of Raf-1 and H-Ras on a BIAcore biosensor. The assay for ZRAF322 was between the RBD of Raf-1 and H-Ras on a BIAcore biosensor. The assay for ZRAF322 was conducted as follows. Firstly, H-Ras was captured on an HSA surface via its ABP affinity conducted as follows. Firstly, H-Ras was captured on an HSA surface via its ABP affinity handle. Secondly the RBD was flowed over the surface, either alone or pre-incubated with a handle. Secondly the RBD was flowed over the surface, either alone or pre-incubated with a molar excess of affibody molecules. The sensorgrams showed that pre-incubation with molar excess of affibody molecules. The sensorgrams showed that pre-incubation with

ZRAF322, but not with another variant, resulted in a complete inhibition of the H-Ras - RBD ZRAF322, but not with another variant, resulted in a complete inhibition of the H-Ras - RBD interaction. This experiment was repeated with varying concentrations of ZRAF322 and a dose- interaction. This experiment was repeated with varying concentrations of ZRAF322 and a dose- dependent inhibition could be observed (Figure 24). The latter results make ZRAF322 a very dependent inhibition could be observed (Figure 24). The latter results make ZRAF322 a very promising molecule for the interference with the (ERK) MAPK pathway inside of cells. promising molecule for the interference with the (ERK) MAPK pathway inside of cells.

Taken together, affibody molecules that selectively bind to H-Ras or Raf-1 could be Taken together, affibody molecules that selectively bind to H-Ras or Raf-1 could be selected using phage display. Of these, ZRAF322 was shown to bind to the RBD and inhibit the selected using phage display. Of these, ZRAF322 was shown to bind to the RBD and inhibit the interaction between Ras and Raf in vitro. This variant may interfere with the (ERK) MAPK interaction between Ras and Raf in vitro. This variant may interfere with the (ERK) MAPK pathway when delivered to cancer cells. However, due to its rather low affinity, an affinity pathway when delivered to cancer cells. However, due to its rather low affinity, an affinity maturation is probably required. maturation is probably required.

Figure 24: Competition study between H-Ras and his6-(ZRAF322)2 for binding to Raf-1-RBD-GST Figure 24: Competition study between H-Ras and his6-(ZRAF322)2 for binding to Raf-1-RBD-GST using a biosensor. (a) Schematic representation of the setup. his6-ABP-HRas was injected over a using a biosensor. (a) Schematic representation of the setup. his6-ABP-HRas was injected over a surface with immobilized HSA (I, II, III), followed by an injection of Raf-1-RBD-GST (abbreviated surface with immobilized HSA (I, II, III), followed by an injection of Raf-1-RBD-GST (abbreviated

RBD) (II) or RBD pre-incubated with his6-(ZRAF322)2 (III, trefoils). (b) Overlay of sensorgrams RBD) (II) or RBD pre-incubated with his6-(ZRAF322)2 (III, trefoils). (b) Overlay of sensorgrams obtained after injecting RBD (black solid line, case II in panel A), or RBD pre-incubated with his6- obtained after injecting RBD (black solid line, case II in panel A), or RBD pre-incubated with his6-

(ZRAF322)2 (black dashed line, case III in panel A), or without injection of RBD (grey solid line, case I (ZRAF322)2 (black dashed line, case III in panel A), or without injection of RBD (grey solid line, case I in Panel A). Injection of his6-ABP-HRas was started at 1 and stopped at 2. Injection of RBD (with or in Panel A). Injection of his6-ABP-HRas was started at 1 and stopped at 2. Injection of RBD (with or without his6-(ZRAF322)2) was started at 3 and stopped at 4. (c) Case III in Panel A was repeated with without his6-(ZRAF322)2) was started at 3 and stopped at 4. (c) Case III in Panel A was repeated with different concentrations of his6-(ZRAF322)2. Responses obtained 45 seconds after injection stop of RBD different concentrations of his6-(ZRAF322)2. Responses obtained 45 seconds after injection stop of RBD pre-incubated with his6-(ZRAF322)2 was plotted against the concentration of his6-(ZRAF322)2. The response pre-incubated with his6-(ZRAF322)2 was plotted against the concentration of his6-(ZRAF322)2. The response at the start of injection of RBD pre-incubated with his6-(ZRAF322)2 was set to zero. at the start of injection of RBD pre-incubated with his6-(ZRAF322)2 was set to zero.

72 Ribosome display for selection and evolution of affibody molecules 72 Ribosome display for selection and evolution of affibody molecules

4.4 Affinity maturation of an affibody molecule binding to human Raf-1 4.4 Affinity maturation of an affibody molecule binding to human Raf-1 via non-targeted in vitro evolution (IV) via non-targeted in vitro evolution (IV)

As mentioned in chapter 2 of this thesis, a “true” protein in vitro evolution experiment can be As mentioned in chapter 2 of this thesis, a “true” protein in vitro evolution experiment can be regarded as one involving repetitive rounds of diversification and selection, unlike a selection regarded as one involving repetitive rounds of diversification and selection, unlike a selection experiment that starts with a diverse pool, followed by repetitive rounds of selection without experiment that starts with a diverse pool, followed by repetitive rounds of selection without further diversification. Due to its openness and independence from transformation of cells, further diversification. Due to its openness and independence from transformation of cells, ribosome display is a suitable system for such an in vitro evolution experiment. The DNA is ribosome display is a suitable system for such an in vitro evolution experiment. The DNA is readily accessible for the introduction of diversity within selection cycles. readily accessible for the introduction of diversity within selection cycles.

The aim of this study was to further develop the ribosome display system described in The aim of this study was to further develop the ribosome display system described in (I) for the affinity maturation of affibody molecules by non-targeted in vitro evolution (I) for the affinity maturation of affibody molecules by non-targeted in vitro evolution principles. Due to its rather low affinity but interesting epitope preference, the Raf-1 binder principles. Due to its rather low affinity but interesting epitope preference, the Raf-1 binder

ZRAF322 was chosen as a suitable candidate. The ribosome display system described in (I) was ZRAF322 was chosen as a suitable candidate. The ribosome display system described in (I) was combined with error-prone PCR of the entire affibody gene before and in-between two combined with error-prone PCR of the entire affibody gene before and in-between two consecutive selection rounds. This is in contrast to traditional affinity maturations of affibody consecutive selection rounds. This is in contrast to traditional affinity maturations of affibody molecules, typically involving a re-randomization of some of the 13 initially randomized molecules, typically involving a re-randomization of some of the 13 initially randomized positions on helix one and two. The evolution process was followed by sequencing of positions on helix one and two. The evolution process was followed by sequencing of individual clones and biosensor-based binding analysis of the entire clone pool to its antigen individual clones and biosensor-based binding analysis of the entire clone pool to its antigen (Figure 25). (Figure 25). Initially, error-prone PCR conditions were adjusted to yield about 1 to 2 amino acid Initially, error-prone PCR conditions were adjusted to yield about 1 to 2 amino acid substitutions per affibody gene. The number of substitutions and an even distribution substitutions per affibody gene. The number of substitutions and an even distribution throughout the affibody gene were confirmed by DNA sequencing, and the library was throughout the affibody gene were confirmed by DNA sequencing, and the library was inserted into the ribosome display vector pRD2 for preparation of the ribosome display inserted into the ribosome display vector pRD2 for preparation of the ribosome display expression cassette. Two consecutive selection cycles were either performed at 4° C, followed expression cassette. Two consecutive selection cycles were either performed at 4° C, followed by ambient temperature (route I) or ambient temperature only (route II). In both routes, by ambient temperature (route I) or ambient temperature only (route II). In both routes, antigen concentration was reduced from 80 to 20 nM and the washing time prolonged from 6 antigen concentration was reduced from 80 to 20 nM and the washing time prolonged from 6 to 60 min to increase the selection stringency. After a first round of biopanning against ABP- to 60 min to increase the selection stringency. After a first round of biopanning against ABP- hRaf-15-143, the same antigen that was also used in (III), representative aliquots of the selected hRaf-15-143, the same antigen that was also used in (III), representative aliquots of the selected pools were inserted into pAff35, in frame with the ABD35 affinity handle for capture on an pools were inserted into pAff35, in frame with the ABD35 affinity handle for capture on an HSA-containing biosensor surface, while another aliquot served as template for a second HSA-containing biosensor surface, while another aliquot served as template for a second round of error-prone PCR, insertion into pRD2 and ribosome display selection (Figure 25). round of error-prone PCR, insertion into pRD2 and ribosome display selection (Figure 25). Sebastian Grimm 73 Sebastian Grimm 73

Figure 25: Schematic overview of the affinity maturation strategy. The gene encoding ZRAF322 is Figure 25: Schematic overview of the affinity maturation strategy. The gene encoding ZRAF322 is amplified from vector DNA (a) to yield linear DNA template for a subsequent error-prone PCR amplified from vector DNA (a) to yield linear DNA template for a subsequent error-prone PCR reaction (c). The pool of mutated ZRAF322-clones (ZRAF322*) is inserted in vector pRD2 to introduce reaction (c). The pool of mutated ZRAF322-clones (ZRAF322*) is inserted in vector pRD2 to introduce promotor (pT7) and pIII-SecM spacer protein for ribosome display (d). Gene cassettes devoid of promotor (pT7) and pIII-SecM spacer protein for ribosome display (d). Gene cassettes devoid of

STOP-codon and including promoter, ZRAF322* and spacer are amplified, transcribed and translated in STOP-codon and including promoter, ZRAF322* and spacer are amplified, transcribed and translated in vitro for ribosome display selection (f). Selected mRNA molecules are reverse transcribed (g) and vitro for ribosome display selection (f). Selected mRNA molecules are reverse transcribed (g) and PCR-amplified to obtain template for the subsequent selection cycle (h) or for ligation to a PCR-amplified to obtain template for the subsequent selection cycle (h) or for ligation to a hexahistidine-albumin binding domain (H6-ABD35) tag in an expression vector (i, j). For genotypic hexahistidine-albumin binding domain (H6-ABD35) tag in an expression vector (i, j). For genotypic analyses, E. coli cells are transformed and the affibody genes sequenced (k, m). For phenotypic pool analyses, E. coli cells are transformed and the affibody genes sequenced (k, m). For phenotypic pool analyses, transformed cells are induced for expression and capture of H6-ABD35 tagged ZRAF322* on an analyses, transformed cells are induced for expression and capture of H6-ABD35 tagged ZRAF322* on an HSA-coated biosensor surface (l). HSA-coated biosensor surface (l).

ABD35 is a protein derived from a naturally occurring 46-residue three-helix bundle ABD35 is a protein derived from a naturally occurring 46-residue three-helix bundle albumin binding domain (ABD) and affinity engineered to bind HSA with an equilibrium albumin binding domain (ABD) and affinity engineered to bind HSA with an equilibrium dissociation constant of 50-500 fM 320. This very high affinity makes it a suitable handle for a dissociation constant of 50-500 fM 320. This very high affinity makes it a suitable handle for a sustained capture on HSA. Pools of ABD-affibody fusion proteins were captured from crude sustained capture on HSA. Pools of ABD-affibody fusion proteins were captured from crude E. coli cell lysate on an HSA-containing BIAcore biosensor surface, and a recombinant target E. coli cell lysate on an HSA-containing BIAcore biosensor surface, and a recombinant target containing the Ras-binding domain of hRaf-1 (RBD) was injected at a single 200 nM containing the Ras-binding domain of hRaf-1 (RBD) was injected at a single 200 nM concentration. For comparison, the sensorgram traces were normalized to 1000 resonance concentration. For comparison, the sensorgram traces were normalized to 1000 resonance units (RU) prior to RBD injection. After a first selection cycle, no increase in RBD binding units (RU) prior to RBD injection. After a first selection cycle, no increase in RBD binding response was observed in comparison to the parental clone ZRAF322, indicating no increase in response was observed in comparison to the parental clone ZRAF322, indicating no increase in 74 Ribosome display for selection and evolution of affibody molecules 74 Ribosome display for selection and evolution of affibody molecules overall affinity of the pool to RBD. After the second selection cycle, however, the binding overall affinity of the pool to RBD. After the second selection cycle, however, the binding responses to RBD had approximately doubled for both selection routes, indicating a clear responses to RBD had approximately doubled for both selection routes, indicating a clear increase in overall affinity of the pool to RBD (Figure 26). Notably, the dissociation of increase in overall affinity of the pool to RBD (Figure 26). Notably, the dissociation of

ABD35 and HSA observed was somewhat faster than expected from a previously published ABD35 and HSA observed was somewhat faster than expected from a previously published -5 -6 320 -5 -6 320 koff between 1.5x10 and 1.5x10 /s . koff between 1.5x10 and 1.5x10 /s .

Figure 26: Biosensor based pool analysis of relative target affinity during selection process. E. Figure 26: Biosensor based pool analysis of relative target affinity during selection process. E. coli cell lysates containing pools of mutated ABD35- ZRAF322 clones obtained after selection cycle 1 coli cell lysates containing pools of mutated ABD35- ZRAF322 clones obtained after selection cycle 1 (black dotted traces) or selection cycle 2 (black solid traces) from either route I (a, b) or route II (c, d) (black dotted traces) or selection cycle 2 (black solid traces) from either route I (a, b) or route II (c, d) were injected over an HSA-surface (ii), respectively, followed by injection of the antigen RBD (iii). E. were injected over an HSA-surface (ii), respectively, followed by injection of the antigen RBD (iii). E. coli cell lysate containing ABD35- ZRAF322 only (same black dashed trace for route I and II) was coli cell lysate containing ABD35- ZRAF322 only (same black dashed trace for route I and II) was included as comparison. For reference subtraction, a duplicate of each cell lysate sample was flowed included as comparison. For reference subtraction, a duplicate of each cell lysate sample was flowed over the HSA surface, followed by injection of a buffer sample. All traces were normalized to 1000 over the HSA surface, followed by injection of a buffer sample. All traces were normalized to 1000 RU at the point of antigen injection. Traces before reference subtraction are shown in (a) and (c). RU at the point of antigen injection. Traces before reference subtraction are shown in (a) and (c). Corresponding traces obtained from antigen injections after reference subtraction are shown in (b) and Corresponding traces obtained from antigen injections after reference subtraction are shown in (b) and (d). (d). Sebastian Grimm 75 Sebastian Grimm 75

Sequencing of individual clones selected after the first selection round, from both routes, Sequencing of individual clones selected after the first selection round, from both routes, revealed that 40-50% of the clones still corresponded to the original ZRAF322 sequence, while revealed that 40-50% of the clones still corresponded to the original ZRAF322 sequence, while mutations were relatively even spread in the rest of the clones. An exception was a specific mutations were relatively even spread in the rest of the clones. An exception was a specific A27V mutation that was observed in 11% of the clones. After the second selection round, the A27V mutation that was observed in 11% of the clones. After the second selection round, the picture had changed dramatically, and about 83% of clones from both routes contained the picture had changed dramatically, and about 83% of clones from both routes contained the A27V substitution, and also some other substitutions such as N10Y, A12V and I31V became A27V substitution, and also some other substitutions such as N10Y, A12V and I31V became enriched. Notably, some of these positions do not belong to the 13 originally randomized enriched. Notably, some of these positions do not belong to the 13 originally randomized positions for the selection of ZRAF322. Four clones were chosen for further, phenotypic positions for the selection of ZRAF322. Four clones were chosen for further, phenotypic analysis. One clone (ZRAF322:01) contained the A27V substitution only, while the other three analysis. One clone (ZRAF322:01) contained the A27V substitution only, while the other three clones contained also A12V (ZRAF322:02), N10Y (ZRAF322:03) or I31V (ZRAF322:04) substitutions, clones contained also A12V (ZRAF322:02), N10Y (ZRAF322:03) or I31V (ZRAF322:04) substitutions, respectively. The presence of homodimers or oligomers was excluded using gel filtration respectively. The presence of homodimers or oligomers was excluded using gel filtration chromatography and it was found that the dominant A27V substitution alone contributed a chromatography and it was found that the dominant A27V substitution alone contributed a 13-fold increase in affinity and the A12V and I31V substitutions contributed an additional, 13-fold increase in affinity and the A12V and I31V substitutions contributed an additional, approximately two-fold increase, respectively (Figure 27c). approximately two-fold increase, respectively (Figure 27c). The question arose, whether or not these variants had a retained three-helical overall The question arose, whether or not these variants had a retained three-helical overall structure, and if their thermal stability was in some way affected by the substitutions. To structure, and if their thermal stability was in some way affected by the substitutions. To investigate this, CD spectra and melting curves were recorded, together with ZWT and ZRAF322 investigate this, CD spectra and melting curves were recorded, together with ZWT and ZRAF322 as references. Interestingly, all clones showed almost overlapping spectra, suggesting similar as references. Interestingly, all clones showed almost overlapping spectra, suggesting similar α-helical contents in all cases (Figure 28a). Thermal melting curves indicated that the A27V α-helical contents in all cases (Figure 28a). Thermal melting curves indicated that the A27V substitution alone slightly increased the thermal stability, whereas A12V and N10Y substitution alone slightly increased the thermal stability, whereas A12V and N10Y substitutions had no significant effect while the I31V substitution slightly reduced the thermal substitutions had no significant effect while the I31V substitution slightly reduced the thermal stability of the molecule (Figure 28b). stability of the molecule (Figure 28b). The I31 position is of particular importance in the affibody scaffold, since it is located The I31 position is of particular importance in the affibody scaffold, since it is located in the middle of the binding face and may constrain addressable epitopes. Still, this position in the middle of the binding face and may constrain addressable epitopes. Still, this position was conserved in all previously published affibody libraries. It was therefore interesting to was conserved in all previously published affibody libraries. It was therefore interesting to note the appearance of an isoleucine to valine substitution at this position, that is associated note the appearance of an isoleucine to valine substitution at this position, that is associated with an increased affinity but decreased thermal stability. Valine differs from isoleucine only with an increased affinity but decreased thermal stability. Valine differs from isoleucine only by a single methyl group and this substitution may reflect a beneficiary effect on affinity from by a single methyl group and this substitution may reflect a beneficiary effect on affinity from having a smaller aliphatic residue in this position. having a smaller aliphatic residue in this position. 76 Ribosome display for selection and evolution of affibody molecules 76 Ribosome display for selection and evolution of affibody molecules

Figure 27: Characterization of affinity matured affibody variants. (a) ZRAF322 (black), ZRAF322:01 Figure 27: Characterization of affinity matured affibody variants. (a) ZRAF322 (black), ZRAF322:01

(blue), ZRAF322:02 (red), ZRAF322:03 (orange), ZRAF322:04 (green) and ZWT-Cys were eluted from a gel filtration (blue), ZRAF322:02 (red), ZRAF322:03 (orange), ZRAF322:04 (green) and ZWT-Cys were eluted from a gel filtration chromatography column. The absorbance at 215 nm is plotted against the elution volume. The ZWT-Cys chromatography column. The absorbance at 215 nm is plotted against the elution volume. The ZWT-Cys reference forms dimers in solution, dimer and monomer peak are indicated with arrows. (b) Overlay reference forms dimers in solution, dimer and monomer peak are indicated with arrows. (b) Overlay plot of biosensor response traces for the parental ZRAF322 (black), ZRAF322:01 (blue), ZRAF322:02 (red), plot of biosensor response traces for the parental ZRAF322 (black), ZRAF322:01 (blue), ZRAF322:02 (red),

ZRAF322:03 (orange) and ZRAF322:04 (green) obtained after injection of 1 µM solutions of the respective ZRAF322:03 (orange) and ZRAF322:04 (green) obtained after injection of 1 µM solutions of the respective variants over a 2950 RU ABP-hRaf-15-143 sensor surface. (c) Deduced amino acid sequences and variants over a 2950 RU ABP-hRaf-15-143 sensor surface. (c) Deduced amino acid sequences and binding parameters of selected ZRAF322 variants. The wild type Z domain sequence (ZWT) is shown as binding parameters of selected ZRAF322 variants. The wild type Z domain sequence (ZWT) is shown as reference. Dots indicate positions that were randomized in the phage library for selection of ZRAF322 reference. Dots indicate positions that were randomized in the phage library for selection of ZRAF322 and boxes indicate positions corresponding to the three α-helices of the wild type Z domain. Mutated and boxes indicate positions corresponding to the three α-helices of the wild type Z domain. Mutated amino acid positions in ZRAF322 variants are written in bold. The kinetic rate constants (kon, koff) and the amino acid positions in ZRAF322 variants are written in bold. The kinetic rate constants (kon, koff) and the equilibrium dissociation constants (Kd) are given to the right. equilibrium dissociation constants (Kd) are given to the right.

Sebastian Grimm 77 Sebastian Grimm 77

Figure 28: Secondary structure content, thermal melting point and refolding analyses. Circular Figure 28: Secondary structure content, thermal melting point and refolding analyses. Circular dichroism (CD) data was recorded in PBS at pH 7.4 and a protein concentration of 62 µM. (a) Overlay dichroism (CD) data was recorded in PBS at pH 7.4 and a protein concentration of 62 µM. (a) Overlay of CD spectra recorded at a constant temperature of 25 °C for ZRAF322 (black), ZRAF322:01 (blue), of CD spectra recorded at a constant temperature of 25 °C for ZRAF322 (black), ZRAF322:01 (blue),

ZRAF322:02 (red), ZRAF322:03 (orange), ZRAF322:04 (green), and ZWT (grey) samples before thermal melting. ZRAF322:02 (red), ZRAF322:03 (orange), ZRAF322:04 (green), and ZWT (grey) samples before thermal melting.

(b) Normalized melting curves recorded at 221 nm. (c) CD spectra of ZRAF322:04, either before (blue) or (b) Normalized melting curves recorded at 221 nm. (c) CD spectra of ZRAF322:04, either before (blue) or after (red) thermal melting. after (red) thermal melting.

An additional question was, whether the selected variants retained their binding An additional question was, whether the selected variants retained their binding selectivity despite the introduced substitutions. To address this, biotinylated variants were selectivity despite the introduced substitutions. To address this, biotinylated variants were probed with their antigen in mixture with a set of 76 different human protein fragments probed with their antigen in mixture with a set of 76 different human protein fragments (PrEST fragments) used as controls. Each of the 77 proteins were immobilized on a unique (PrEST fragments) used as controls. Each of the 77 proteins were immobilized on a unique sort of color-coded beads. Binding was detected via subsequent incubation with fluorescently sort of color-coded beads. Binding was detected via subsequent incubation with fluorescently labeled streptavidin and readings in a Luminex flow cytofluometer. The original ZRAF322 labeled streptavidin and readings in a Luminex flow cytofluometer. The original ZRAF322 showed a very selective binding profile and encouragingly, the good selectivity was also more showed a very selective binding profile and encouragingly, the good selectivity was also more or less retained for the higher affinity variants ZRAF322:01-04 (Figure 29). or less retained for the higher affinity variants ZRAF322:01-04 (Figure 29). Taken together, this study represents the first time that mutations were introduced in- Taken together, this study represents the first time that mutations were introduced in- between selection cycles of affibody molecules. With this approach, a careful eye needs to be between selection cycles of affibody molecules. With this approach, a careful eye needs to be kept on the stability and selectivity of the molecules, since certain substitutions such as I31V kept on the stability and selectivity of the molecules, since certain substitutions such as I31V may affect the overall scaffold stability. A pre-selection step could in future be included to may affect the overall scaffold stability. A pre-selection step could in future be included to deplete molecules with a non-selective binding profile. Nevertheless, the study generated deplete molecules with a non-selective binding profile. Nevertheless, the study generated variants of ZRAF322 with up to 26-fold improved affinity, slightly improved thermal stability variants of ZRAF322 with up to 26-fold improved affinity, slightly improved thermal stability and retained selectivity which may be interesting candidates for an intracellular application. and retained selectivity which may be interesting candidates for an intracellular application. 78 Ribosome display for selection and evolution of affibody molecules 78 Ribosome display for selection and evolution of affibody molecules

Figure 29: Analysis of affibody binding selectivity using a collection of recombinant human Figure 29: Analysis of affibody binding selectivity using a collection of recombinant human protein fragment controls (PrESTs). (a) PrESTs are coupled to color-coded beads and contain the protein fragment controls (PrESTs). (a) PrESTs are coupled to color-coded beads and contain the ABP-tag. Median fluorescence intensities (MFI) obtained from an anti-ABP (Ab1) and labeled ABP-tag. Median fluorescence intensities (MFI) obtained from an anti-ABP (Ab1) and labeled secondary (Ab2) antibody served as an estimate for coupled protein levels. MFI obtained from secondary (Ab2) antibody served as an estimate for coupled protein levels. MFI obtained from biotinylated affibody variants (Z) and labeled Streptavidin (SA) were normalized with the biotinylated affibody variants (Z) and labeled Streptavidin (SA) were normalized with the corresponding ABP-signals. (b) ABP MFI signals are plotted against bead IDs. (c-g) Relative, corresponding ABP-signals. (b) ABP MFI signals are plotted against bead IDs. (c-g) Relative, normalized MFI obtained from affibody clones binding to the different PrESTs. The grey bar of bead normalized MFI obtained from affibody clones binding to the different PrESTs. The grey bar of bead

ID3 indicates MFI or relative MFI signals obtained from target ABP-hRaf-15-143-coated beads. (h) The ID3 indicates MFI or relative MFI signals obtained from target ABP-hRaf-15-143-coated beads. (h) The relative target MFI signal was divided by the mean relative non-target MFI signal to obtain relative target MFI signal was divided by the mean relative non-target MFI signal to obtain signal/background ratios. The standard deviations obtained from triplicate measurements are denoted. signal/background ratios. The standard deviations obtained from triplicate measurements are denoted. Sebastian Grimm 79 Sebastian Grimm 79

4.5 Conclusions and future perspectives 4.5 Conclusions and future perspectives

The possibility to create myriads of physically linked genotype-phenotype pairs has The possibility to create myriads of physically linked genotype-phenotype pairs has revolutionized the field of protein engineering. Harnessing natural selection principles, such revolutionized the field of protein engineering. Harnessing natural selection principles, such genotype-phenotype pairs are today widely used for the isolation of tailored affinity reagents genotype-phenotype pairs are today widely used for the isolation of tailored affinity reagents with high affinities and selectivities towards an array of different target structures. A recent with high affinities and selectivities towards an array of different target structures. A recent estimate of the number of human protein-encoding genes is 20500 321, and it should only be a estimate of the number of human protein-encoding genes is 20500 321, and it should only be a question of time until selective, high affinity binders to the entire proteome become available. question of time until selective, high affinity binders to the entire proteome become available. Such reagents could be characterized and validated for different applications with different Such reagents could be characterized and validated for different applications with different structural representations of the antigens. Validated sets of binders could be used by structural representations of the antigens. Validated sets of binders could be used by laboratories or hospitals for standardized binding assays, diagnostics or therapies. Due to their laboratories or hospitals for standardized binding assays, diagnostics or therapies. Due to their structural simplicity and ease of production by e.g. peptide synthesis, affibody molecules are structural simplicity and ease of production by e.g. peptide synthesis, affibody molecules are an appealing alternative to the today most widely used monoclonal or polyclonal antibodies. an appealing alternative to the today most widely used monoclonal or polyclonal antibodies. Once selected, characterized and published, their sequences are an open source for the Once selected, characterized and published, their sequences are an open source for the production of, according to Anfinsen’s dogma 322, uniform and defined binding proteins in production of, according to Anfinsen’s dogma 322, uniform and defined binding proteins in laboratories around the world. The selection and characterization of affibody molecules is still laboratories around the world. The selection and characterization of affibody molecules is still very manual and time consuming and an automation of at least parts of the process could very manual and time consuming and an automation of at least parts of the process could increase the pace and reduce experimental variation. Ribosome display is a selection system increase the pace and reduce experimental variation. Ribosome display is a selection system that is entirely performed in vitro and genotype-phenotype pairs are generated in very small that is entirely performed in vitro and genotype-phenotype pairs are generated in very small volumes of cell-free translation reactions. It can be envisioned to use a liquid handling robot volumes of cell-free translation reactions. It can be envisioned to use a liquid handling robot for the in vitro translation, biopanning, mRNA isolation and RT-PCR amplification steps. To for the in vitro translation, biopanning, mRNA isolation and RT-PCR amplification steps. To overcome limitations associated with the stability of the mRNA and ternary mRNA- overcome limitations associated with the stability of the mRNA and ternary mRNA- ribosome-protein complexes, the development of novel RNase-free in vitro translation ribosome-protein complexes, the development of novel RNase-free in vitro translation systems such as the PURE extract 182 or a further engineered SecM-based ribosome stalling systems such as the PURE extract 182 or a further engineered SecM-based ribosome stalling sequence motif may be useful. An ideal translation system should be economical, devoid of sequence motif may be useful. An ideal translation system should be economical, devoid of proteases and RNases and rich in ribosomes to produce high numbers of functional ternary proteases and RNases and rich in ribosomes to produce high numbers of functional ternary complexes. The SecM motif could be further engineered by variegation of the 17-mer peptide complexes. The SecM motif could be further engineered by variegation of the 17-mer peptide sequence and ribosome display selection at environmental stress conditions such as elevated sequence and ribosome display selection at environmental stress conditions such as elevated temperatures. temperatures. As part of this thesis, affibody molecules were described that selectively bind to mIgG1, H- As part of this thesis, affibody molecules were described that selectively bind to mIgG1, H- Ras, Raf-1 and SATB1. These binders should be further tested for their intended final Ras, Raf-1 and SATB1. These binders should be further tested for their intended final applications: the purification or detection of mIgG1, the interference with the (ERK) MAPK applications: the purification or detection of mIgG1, the interference with the (ERK) MAPK pathway or the detection of SATB1 in breast cancer cells. Furthermore, it would be very pathway or the detection of SATB1 in breast cancer cells. Furthermore, it would be very interesting to obtain co-crystal structures of the affibody molecules in complexes with their interesting to obtain co-crystal structures of the affibody molecules in complexes with their respective antigens. Such a study would finalize the epitope determination and provide respective antigens. Such a study would finalize the epitope determination and provide valuable insights in the modes of antigen recognition to e.g. explain the 13-fold affinity valuable insights in the modes of antigen recognition to e.g. explain the 13-fold affinity improvement of ZRAF322 due to a single A27V substitution. improvement of ZRAF322 due to a single A27V substitution. 80 Ribosome display for selection and evolution of affibody molecules 80 Ribosome display for selection and evolution of affibody molecules

Acknowledgements Acknowledgements

It has been a rugged but steady path on the PhD thesis landscape (Figure 3), with moments of It has been a rugged but steady path on the PhD thesis landscape (Figure 3), with moments of great joy, local minima to overcome and a global maximum in reach. I hope you will enjoy great joy, local minima to overcome and a global maximum in reach. I hope you will enjoy reading this book and would like to dedicate the following pages to all of you who inspired reading this book and would like to dedicate the following pages to all of you who inspired me, taught me, supported me, believed in me, collaborated with me, cared for me, or simply me, taught me, supported me, believed in me, collaborated with me, cared for me, or simply shared a good time. shared a good time.

Per-Åke - before you took me on as a PhD student, I met you in the corridor and excitedly Per-Åke - before you took me on as a PhD student, I met you in the corridor and excitedly presented you an idea for a new, ingenious protein selection system. You were reflecting a presented you an idea for a new, ingenious protein selection system. You were reflecting a few seconds and could name me the exact paper where just that system was already described. few seconds and could name me the exact paper where just that system was already described. Discouraged for a moment, I realized that this is someone who really knows his subject. I Discouraged for a moment, I realized that this is someone who really knows his subject. I want to thank you for having employed me as a PhD student. You gave me all the trust and want to thank you for having employed me as a PhD student. You gave me all the trust and freedom to start with a challenging project in your little, exclusive group. We built up a new freedom to start with a challenging project in your little, exclusive group. We built up a new selection system for affibodies (with some "Fingerspitzengefühl") and finally got it to work, selection system for affibodies (with some "Fingerspitzengefühl") and finally got it to work, which I am very proud of. I felt that whenever it mattered you were standing by my side and which I am very proud of. I felt that whenever it mattered you were standing by my side and supported me. Torbjörn - you picked me up when I had for the first time arrived in supported me. Torbjörn - you picked me up when I had for the first time arrived in Stockholm on a late evening, with a suitcase and an address somewhere outside of Stockholm Stockholm on a late evening, with a suitcase and an address somewhere outside of Stockholm in Tyresö. Thank you for having taken me on as a Master student, and for all the discussions in Tyresö. Thank you for having taken me on as a Master student, and for all the discussions later. In that context I also want to thank Emma L who taught me phage display and later. In that context I also want to thank Emma L who taught me phage display and supervised my Master thesis. Mathias and Stefan - you two are fantastic in motivating and supervised my Master thesis. Mathias and Stefan - you two are fantastic in motivating and engaging people for science and make floor 3 a unique place to work at, in an atmosphere of engaging people for science and make floor 3 a unique place to work at, in an atmosphere of trust and honesty. Besides, I couldn’t think of a cleaning day being such an event before I met trust and honesty. Besides, I couldn’t think of a cleaning day being such an event before I met our cleaning general Stefan. There is a German saying "Ordnung ist das halbe Leben". In that our cleaning general Stefan. There is a German saying "Ordnung ist das halbe Leben". In that half are certainly your cleaning days. Feifan - I am still amazed about the happiness and half are certainly your cleaning days. Feifan - I am still amazed about the happiness and energy you are daily spreading. It was a pleasure to teach you as a Master student. And now energy you are daily spreading. It was a pleasure to teach you as a Master student. And now you are PhD student yourself. I wish you lot’s of low nM affibodies and look forward to come you are PhD student yourself. I wish you lot’s of low nM affibodies and look forward to come to your thesis defense some day. Thank you also to Per-Åke Sr, expert in protein fragment to your thesis defense some day. Thank you also to Per-Åke Sr, expert in protein fragment complementation questions and house constructions, who not only shared the interest for complementation questions and house constructions, who not only shared the interest for protein engineering, but also for exotic plants. My two other, former Master students, Joel G protein engineering, but also for exotic plants. My two other, former Master students, Joel G and Samaneh: you passed two very challenging projects with flying colors and I wish you and Samaneh: you passed two very challenging projects with flying colors and I wish you two successful careers in academia and industry. two successful careers in academia and industry. The "Bänkpressare" Johanna, Henrik, Johan and Charlotte - thanks for friendship, all the The "Bänkpressare" Johanna, Henrik, Johan and Charlotte - thanks for friendship, all the dinners and "Blodomloppen" together. Until now my time got steadily closer to Henrik’s. dinners and "Blodomloppen" together. Until now my time got steadily closer to Henrik’s. Will there be another race? My "rumskamrater": Johan L - it was you who recommended me Will there be another race? My "rumskamrater": Johan L - it was you who recommended me switching to a Mac. I got dependent and thank you for that. Maybe I should thank Steve Jobs switching to a Mac. I got dependent and thank you for that. Maybe I should thank Steve Jobs here as well. I had an entertaining "skrivbordsgranne" with you. Just have to name ghost icons here as well. I had an entertaining "skrivbordsgranne" with you. Just have to name ghost icons and Gandalf. Johanna, for the good spirit and atmosphere in our room. Marcus, your and Gandalf. Johanna, for the good spirit and atmosphere in our room. Marcus, your Cornflakes are still there and only get better. The Lyngman trophy went to house 10, but now Cornflakes are still there and only get better. The Lyngman trophy went to house 10, but now Sebastian Grimm 81 Sebastian Grimm 81 that parts of house 10 are moving to floor 3 it may come back to where it belongs. Erik, that parts of house 10 are moving to floor 3 it may come back to where it belongs. Erik, thanks for all the scientific discussions, questions of the day, the Dynabead tube-holder and thanks for all the scientific discussions, questions of the day, the Dynabead tube-holder and strategy game evenings. Sverker for providing me your nice illustrator brushes. Thanks also strategy game evenings. Sverker for providing me your nice illustrator brushes. Thanks also to Malin, Jorge, Karin and Beata. My present colleagues: Joel L, who rescued a precious to Malin, Jorge, Karin and Beata. My present colleagues: Joel L, who rescued a precious sample when the freeze dryer malfunctioned and was always very helpful. The same applies sample when the freeze dryer malfunctioned and was always very helpful. The same applies to Ulrika, thank your for teaching me the Luminex and some R-scripting. Håkan who shared to Ulrika, thank your for teaching me the Luminex and some R-scripting. Håkan who shared late evenings with me in the office while thesis writing, thank you for discussing science and late evenings with me in the office while thesis writing, thank you for discussing science and helping me with practical thesis matters. We should soon realize our protein engineering helping me with practical thesis matters. We should soon realize our protein engineering "gedankenexperiment". Thanks also to Anna H and Johan S for the company. To all the "gedankenexperiment". Thanks also to Anna H and Johan S for the company. To all the innebandy players from floor 3 and house 10 for hard but fair, weekly matches. I really innebandy players from floor 3 and house 10 for hard but fair, weekly matches. I really enjoyed it and it made me become more Swedish. Jochen - for thoroughly reading the thesis enjoyed it and it made me become more Swedish. Jochen - for thoroughly reading the thesis manuscript and for weekend walks across Söder during thesis writing, I really needed some manuscript and for weekend walks across Söder during thesis writing, I really needed some fresh air sometimes. John - I remember us discussing anticalins while spreading some E. coli fresh air sometimes. John - I remember us discussing anticalins while spreading some E. coli cells before I started my PhD. Thank you for all the discussions from then to now, for cells before I started my PhD. Thank you for all the discussions from then to now, for perfectly organizing our conference trip to Keystone and for reading the manuscript. Johan R perfectly organizing our conference trip to Keystone and for reading the manuscript. Johan R - thank you for all the help with the flow cytometer and for sparing your SATB2 antibodies. - thank you for all the help with the flow cytometer and for sparing your SATB2 antibodies. Nina B - for your company during the proteomic workshop in Lisbon. Bahram - for the spirit Nina B - for your company during the proteomic workshop in Lisbon. Bahram - for the spirit of a Bobby McFerrin song, Mattias and Bahram for all the excellent DNA sequences and for of a Bobby McFerrin song, Mattias and Bahram for all the excellent DNA sequences and for being part of the soul of floor 3. The protein factory for helping out during protein being part of the soul of floor 3. The protein factory for helping out during protein purifications, and not to forget Emma Ö for handling all the deliveries, preparation of agar purifications, and not to forget Emma Ö for handling all the deliveries, preparation of agar plates, media and more. You deserve my special thanks. Joje - sometimes it’s nice, we kept it plates, media and more. You deserve my special thanks. Joje - sometimes it’s nice, we kept it on ice and soon it’s time to celebrate for both of us. Thank you for all the good lab-bench on ice and soon it’s time to celebrate for both of us. Thank you for all the good lab-bench company and for the VM-tips that gave a little bonus. Camilla, Anna P - I am glad that I met company and for the VM-tips that gave a little bonus. Camilla, Anna P - I am glad that I met you two, thank you for all the company during the years. Jens - for inviting me to your family you two, thank you for all the company during the years. Jens - for inviting me to your family over Midsommar, and exercising Swedish with me. Nima - for all the trips to Gålö. Eric - for over Midsommar, and exercising Swedish with me. Nima - for all the trips to Gålö. Eric - for all the IT support. Thank you also to all the committed people at plan 3 for allowing me to all the IT support. Thank you also to all the committed people at plan 3 for allowing me to participate in all the fine Swedish traditions. To all of my Swedish teachers from the KTH participate in all the fine Swedish traditions. To all of my Swedish teachers from the KTH unit for language and communication, especially Lars Cederwall. Thank you also to my unit for language and communication, especially Lars Cederwall. Thank you also to my collaboration partners Rinse Klooster from Leiden University Medical Center and Eva collaboration partners Rinse Klooster from Leiden University Medical Center and Eva Allerbring from Karolinska Institutet for fruitful collaborations. A special thank you to Arne Allerbring from Karolinska Institutet for fruitful collaborations. A special thank you to Arne Skerra for teaching me about protein engineering and recommending me for a Master thesis Skerra for teaching me about protein engineering and recommending me for a Master thesis project at KTH. My friends from Germany - Thomas, Bernd, Rasmus, Matthias and project at KTH. My friends from Germany - Thomas, Bernd, Rasmus, Matthias and Solomon for keeping contact and believing in our friendship throughout my stay in Sweden. Solomon for keeping contact and believing in our friendship throughout my stay in Sweden. And last but not least, I want to thank my family. You are always there for me and that gives And last but not least, I want to thank my family. You are always there for me and that gives me the strength to take challenging tasks. Danke! me the strength to take challenging tasks. Danke! 82 Ribosome display for selection and evolution of affibody molecules 82 Ribosome display for selection and evolution of affibody molecules

References References

1. Berg, J. M., Tymoczko, J. L. & Stryer, L. (2002). Biochemistry. 5th ed., W. H. Freeman 1. Berg, J. M., Tymoczko, J. L. & Stryer, L. (2002). Biochemistry. 5th ed., W. H. Freeman and Co., New York. and Co., New York. 2. Johnson, K. A. (2010). The kinetic and chemical mechanism of high-fidelity DNA 2. Johnson, K. A. (2010). The kinetic and chemical mechanism of high-fidelity DNA polymerases. Biochim Biophys Acta 1804, 1041-8. polymerases. Biochim Biophys Acta 1804, 1041-8. 3. Linke, W. A. (2010). Biomaterials: spider strength and stretchability. Nat Chem Biol 6, 3. Linke, W. A. (2010). Biomaterials: spider strength and stretchability. Nat Chem Biol 6, 702-3. 702-3. 4. Mulder, G. J. (1839). Ueber die Zusammensetzung einiger thierischen Substanzen. J. 4. Mulder, G. J. (1839). Ueber die Zusammensetzung einiger thierischen Substanzen. J. Prakt. Chem. Vol. 16, pp. 129-152. Prakt. Chem. Vol. 16, pp. 129-152. 5. Hofmeister, F. (1902). Über den bau des eiweissmoleküls. Ergeb. Physiol. Vol. 1, pp. 5. Hofmeister, F. (1902). Über den bau des eiweissmoleküls. Ergeb. Physiol. Vol. 1, pp. 759-802. 759-802. 6. Fischer, H. E. (1902). Ueber die hydrolyse der proteïnstoffe. Chem Ztg Vol. 26, pp. 939. 6. Fischer, H. E. (1902). Ueber die hydrolyse der proteïnstoffe. Chem Ztg Vol. 26, pp. 939. 7. Sanger, F. & Tuppy, H. (1951). The amino-acid sequence in the phenylalanyl chain of 7. Sanger, F. & Tuppy, H. (1951). The amino-acid sequence in the phenylalanyl chain of insulin. I. The identification of lower peptides from partial hydrolysates. Biochem J 49, insulin. I. The identification of lower peptides from partial hydrolysates. Biochem J 49, 463-81. 463-81. 8. Perutz, M. F., Rossmann, M. G., Cullis, A. F., Muirhead, H., Will, G. & North, A. C. 8. Perutz, M. F., Rossmann, M. G., Cullis, A. F., Muirhead, H., Will, G. & North, A. C. (1960). Structure of haemoglobin: a three-dimensional Fourier synthesis at 5.5-A. (1960). Structure of haemoglobin: a three-dimensional Fourier synthesis at 5.5-A. resolution, obtained by X-ray analysis. Nature 185, 416-22. resolution, obtained by X-ray analysis. Nature 185, 416-22. 9. Kendrew, J. C., Bodo, G., Dintzis, H. M., Parrish, R. G., Wyckoff, H. & Phillips, D. C. 9. Kendrew, J. C., Bodo, G., Dintzis, H. M., Parrish, R. G., Wyckoff, H. & Phillips, D. C. (1958). A three-dimensional model of the myoglobin molecule obtained by x-ray (1958). A three-dimensional model of the myoglobin molecule obtained by x-ray analysis. Nature 181, 662-6. analysis. Nature 181, 662-6. 10. Linn, S. & Arber, W. (1968). Host specificity of DNA produced by Escherichia coli, X. 10. Linn, S. & Arber, W. (1968). Host specificity of DNA produced by Escherichia coli, X. In vitro restriction of phage fd replicative form. Proc Natl Acad Sci USA 59, 1300-6. In vitro restriction of phage fd replicative form. Proc Natl Acad Sci USA 59, 1300-6. 11. Meselson, M. & Yuan, R. (1968). DNA restriction enzyme from E. coli. Nature 217, 11. Meselson, M. & Yuan, R. (1968). DNA restriction enzyme from E. coli. Nature 217, 1110-4. 1110-4. 12. Olivera, B. M. & Lehman, I. R. (1967). Linkage of polynucleotides through 12. Olivera, B. M. & Lehman, I. R. (1967). Linkage of polynucleotides through phosphodiester bonds by an enzyme from Escherichia coli. Proc Natl Acad Sci USA 57, phosphodiester bonds by an enzyme from Escherichia coli. Proc Natl Acad Sci USA 57, 1426-33. 1426-33. 13. Zimmerman, S. B., Little, J. W., Oshinsky, C. K. & Gellert, M. (1967). Enzymatic 13. Zimmerman, S. B., Little, J. W., Oshinsky, C. K. & Gellert, M. (1967). Enzymatic joining of DNA strands: a novel reaction of diphosphopyridine nucleotide. Proc Natl joining of DNA strands: a novel reaction of diphosphopyridine nucleotide. Proc Natl Acad Sci USA 57, 1841-8. Acad Sci USA 57, 1841-8. 14. Hutchison, C. A., Phillips, S., Edgell, M. H., Gillam, S., Jahnke, P. & Smith, M. (1978). 14. Hutchison, C. A., Phillips, S., Edgell, M. H., Gillam, S., Jahnke, P. & Smith, M. (1978). Mutagenesis at a specific position in a DNA sequence. J Biol Chem 253, 6551-60. Mutagenesis at a specific position in a DNA sequence. J Biol Chem 253, 6551-60. 15. Winter, G., Fersht, A. R., Wilkinson, A. J., Zoller, M. & Smith, M. (1982). Redesigning 15. Winter, G., Fersht, A. R., Wilkinson, A. J., Zoller, M. & Smith, M. (1982). Redesigning enzyme structure by site-directed mutagenesis: tyrosyl tRNA synthetase and ATP enzyme structure by site-directed mutagenesis: tyrosyl tRNA synthetase and ATP binding. Nature 299, 756-8. binding. Nature 299, 756-8. 16. Wells, J. A. & Estell, D. A. (1988). Subtilisin--an enzyme designed to be engineered. 16. Wells, J. A. & Estell, D. A. (1988). Subtilisin--an enzyme designed to be engineered. Trends Biochem Sci 13, 291-7. Trends Biochem Sci 13, 291-7. 17. Perry, L. J. & Wetzel, R. (1984). Disulfide bond engineered into T4 lysozyme: 17. Perry, L. J. & Wetzel, R. (1984). Disulfide bond engineered into T4 lysozyme: stabilization of the protein toward thermal inactivation. Science 226, 555-7. stabilization of the protein toward thermal inactivation. Science 226, 555-7. 18. Clarke, A. R., Atkinson, T. & Holbrook, J. J. (1989). From analysis to synthesis: new 18. Clarke, A. R., Atkinson, T. & Holbrook, J. J. (1989). From analysis to synthesis: new ligand binding sites on the lactate dehydrogenase framework. Part II. Trends Biochem ligand binding sites on the lactate dehydrogenase framework. Part II. Trends Biochem Sci 14, 145-8. Sci 14, 145-8. 19. Cunningham, B. C. & Wells, J. A. (1989). High-resolution epitope mapping of hGH- 19. Cunningham, B. C. & Wells, J. A. (1989). High-resolution epitope mapping of hGH- receptor interactions by alanine-scanning mutagenesis. Science 244, 1081-5. receptor interactions by alanine-scanning mutagenesis. Science 244, 1081-5. 20. Schiweck, W. & Skerra, A. (1997). The rational construction of an antibody against 20. Schiweck, W. & Skerra, A. (1997). The rational construction of an antibody against cystatin: lessons from the crystal structure of an artificial Fab fragment. J Mol Biol 268, cystatin: lessons from the crystal structure of an artificial Fab fragment. J Mol Biol 268, 934-51. 934-51. Sebastian Grimm 83 Sebastian Grimm 83

21. Baldwin, J. & Chothia, C. (1979). Haemoglobin: the structural changes related to ligand 21. Baldwin, J. & Chothia, C. (1979). Haemoglobin: the structural changes related to ligand binding and its allosteric mechanism. J Mol Biol 129, 175-220. binding and its allosteric mechanism. J Mol Biol 129, 175-220. 22. Darwin, C. (1859). On the Origin of Species by means of Natural Selection, or the 22. Darwin, C. (1859). On the Origin of Species by means of Natural Selection, or the preservation of favoured races in the struggle for life, John Murray, London. preservation of favoured races in the struggle for life, John Murray, London. 23. Mullis, K. B. & Faloona, F. A. (1987). Specific synthesis of DNA in vitro via a 23. Mullis, K. B. & Faloona, F. A. (1987). Specific synthesis of DNA in vitro via a polymerase-catalyzed chain reaction. Meth Enzymol 155, 335-50. polymerase-catalyzed chain reaction. Meth Enzymol 155, 335-50. 24. Smith, G. P. (1985). Filamentous fusion phage: novel expression vectors that display 24. Smith, G. P. (1985). Filamentous fusion phage: novel expression vectors that display cloned antigens on the virion surface. Science 228, 1315-7. cloned antigens on the virion surface. Science 228, 1315-7. 25. Francisco, J. A., Campbell, R., Iverson, B. L. & Georgiou, G. (1993). Production and 25. Francisco, J. A., Campbell, R., Iverson, B. L. & Georgiou, G. (1993). Production and fluorescence-activated cell sorting of Escherichia coli expressing a functional antibody fluorescence-activated cell sorting of Escherichia coli expressing a functional antibody fragment on the external surface. Proc Natl Acad Sci USA 90, 10444-8. fragment on the external surface. Proc Natl Acad Sci USA 90, 10444-8. 26. Mattheakis, L. C. (1994). An in vitro Polysome Display System for Identifying Ligands 26. Mattheakis, L. C. (1994). An in vitro Polysome Display System for Identifying Ligands from very Large Peptide Libraries. Proc Natl Acad Sci USA 91, 9022-9026. from very Large Peptide Libraries. Proc Natl Acad Sci USA 91, 9022-9026. 27. Roberts, R. W. & Szostak, J. W. (1997). RNA-peptide fusions for the in vitro selection 27. Roberts, R. W. & Szostak, J. W. (1997). RNA-peptide fusions for the in vitro selection of peptides and proteins. Proc Natl Acad Sci USA 94, 12297-302. of peptides and proteins. Proc Natl Acad Sci USA 94, 12297-302. 28. Hanes, J., Schaffitzel, C., Knappik, A. & Pluckthun, A. (2000). Picomolar affinity 28. Hanes, J., Schaffitzel, C., Knappik, A. & Pluckthun, A. (2000). Picomolar affinity antibodies from a fully synthetic naive library selected and evolved by ribosome antibodies from a fully synthetic naive library selected and evolved by ribosome display. Nat Biotechnol 18, 1287-92. display. Nat Biotechnol 18, 1287-92. 29. Hanes, J. & Plückthun, A. (1997). In vitro selection and evolution of functional proteins 29. Hanes, J. & Plückthun, A. (1997). In vitro selection and evolution of functional proteins by using ribosome display. Proc Natl Acad Sci USA 94, 4937-42. by using ribosome display. Proc Natl Acad Sci USA 94, 4937-42. 30. Seelig, B. & Szostak, J. W. (2007). Selection and evolution of enzymes from a partially 30. Seelig, B. & Szostak, J. W. (2007). Selection and evolution of enzymes from a partially randomized non-catalytic scaffold. Nature 448, 828-31. randomized non-catalytic scaffold. Nature 448, 828-31. 31. Koder, R. L., Anderson, J. L. R., Solomon, L. A., Reddy, K. S., Moser, C. C. & Dutton, 31. Koder, R. L., Anderson, J. L. R., Solomon, L. A., Reddy, K. S., Moser, C. C. & Dutton, P. L. (2009). Design and engineering of an O(2) transport protein. Nature 458, 305-9. P. L. (2009). Design and engineering of an O(2) transport protein. Nature 458, 305-9. 32. Kuhlman, B., Dantas, G., Ireton, G. C., Varani, G., Stoddard, B. L. & Baker, D. (2003). 32. Kuhlman, B., Dantas, G., Ireton, G. C., Varani, G., Stoddard, B. L. & Baker, D. (2003). Design of a novel globular protein fold with atomic-level accuracy. Science 302, 1364- Design of a novel globular protein fold with atomic-level accuracy. Science 302, 1364- 8. 8. 33. Keefe, A. D. & Szostak, J. W. (2001). Functional proteins from a random-sequence 33. Keefe, A. D. & Szostak, J. W. (2001). Functional proteins from a random-sequence library. Nature 410, 715-8. library. Nature 410, 715-8. 34. Goodrich, J. A. & Kugel, J. F. (2007). Binding and kinetics for molecular biologists, 34. Goodrich, J. A. & Kugel, J. F. (2007). Binding and kinetics for molecular biologists, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. 35. Skerra, A. (2000). Engineered protein scaffolds for molecular recognition. J Mol 35. Skerra, A. (2000). Engineered protein scaffolds for molecular recognition. J Mol Recognit 13, 167-87. Recognit 13, 167-87. 36. Ohlson, S., Lundblad, A. & Zopf, D. (1988). Novel approach to affinity 36. Ohlson, S., Lundblad, A. & Zopf, D. (1988). Novel approach to affinity chromatography using "weak" monoclonal antibodies. Anal Biochem 169, 204-8. chromatography using "weak" monoclonal antibodies. Anal Biochem 169, 204-8. 37. Orlova, A., Magnusson, M., Eriksson, T. L., Nilsson, M., Larsson, B., Hoiden- 37. Orlova, A., Magnusson, M., Eriksson, T. L., Nilsson, M., Larsson, B., Hoiden- Guthenberg, I., Widstrom, C., Carlsson, J., Tolmachev, V., Stahl, S. & Nilsson, F. Y. Guthenberg, I., Widstrom, C., Carlsson, J., Tolmachev, V., Stahl, S. & Nilsson, F. Y. (2006). Tumor imaging using a picomolar affinity HER2 binding affibody molecule. (2006). Tumor imaging using a picomolar affinity HER2 binding affibody molecule. Cancer Res 66, 4339-48. Cancer Res 66, 4339-48. 38. Velders, M. P., van Rhijn, C. M., Oskam, E., Fleuren, G. J., Warnaar, S. O. & Litvinov, 38. Velders, M. P., van Rhijn, C. M., Oskam, E., Fleuren, G. J., Warnaar, S. O. & Litvinov, S. V. (1998). The impact of antigen density and antibody affinity on antibody- S. V. (1998). The impact of antigen density and antibody affinity on antibody- dependent cellular cytotoxicity: relevance for immunotherapy of carcinomas. Br J dependent cellular cytotoxicity: relevance for immunotherapy of carcinomas. Br J Cancer 78, 478-83. Cancer 78, 478-83. 39. Adams, G. P., Schier, R., McCall, A. M., Simmons, H. H., Horak, E. M., Alpaugh, R. 39. Adams, G. P., Schier, R., McCall, A. M., Simmons, H. H., Horak, E. M., Alpaugh, R. K., Marks, J. D. & Weiner, L. M. (2001). High affinity restricts the localization and K., Marks, J. D. & Weiner, L. M. (2001). High affinity restricts the localization and tumor penetration of single-chain fv antibody molecules. Cancer Res 61, 4750-5. tumor penetration of single-chain fv antibody molecules. Cancer Res 61, 4750-5. 40. Northrup, S. H. & Erickson, H. P. (1992). Kinetics of protein-protein association 40. Northrup, S. H. & Erickson, H. P. (1992). Kinetics of protein-protein association explained by Brownian dynamics computer simulation. Proc Natl Acad Sci USA 89, explained by Brownian dynamics computer simulation. Proc Natl Acad Sci USA 89, 3338-42. 3338-42. 84 Ribosome display for selection and evolution of affibody molecules 84 Ribosome display for selection and evolution of affibody molecules

41. Marvin, J. S. & Lowman, H. B. (2003). Redesigning an antibody fragment for faster 41. Marvin, J. S. & Lowman, H. B. (2003). Redesigning an antibody fragment for faster association with its antigen. Biochemistry 42, 7077-83. association with its antigen. Biochemistry 42, 7077-83. 42. Schreiber, G. & Fersht, A. R. (1996). Rapid, electrostatically assisted association of 42. Schreiber, G. & Fersht, A. R. (1996). Rapid, electrostatically assisted association of proteins. Nat Struct Biol 3, 427-31. proteins. Nat Struct Biol 3, 427-31. 43. Dyson, H. J. & Wright, P. E. (2005). Intrinsically unstructured proteins and their 43. Dyson, H. J. & Wright, P. E. (2005). Intrinsically unstructured proteins and their functions. Nat Rev Mol Cell Biol 6, 197-208. functions. Nat Rev Mol Cell Biol 6, 197-208. 44. Lu, D., Shen, J., Vil, M. D., Zhang, H., Jimenez, X., Bohlen, P., Witte, L. & Zhu, Z. 44. Lu, D., Shen, J., Vil, M. D., Zhang, H., Jimenez, X., Bohlen, P., Witte, L. & Zhu, Z. (2003). Tailoring in vitro selection for a picomolar affinity human antibody directed (2003). Tailoring in vitro selection for a picomolar affinity human antibody directed against vascular endothelial growth factor receptor 2 for enhanced neutralizing activity. against vascular endothelial growth factor receptor 2 for enhanced neutralizing activity. J Biol Chem 278, 43496-507. J Biol Chem 278, 43496-507. 45. Zahnd, C., Pecorari, F., Straumann, N., Wyler, E. & Pluckthun, A. (2006). Selection 45. Zahnd, C., Pecorari, F., Straumann, N., Wyler, E. & Pluckthun, A. (2006). Selection and Characterization of Her2 Binding-designed Ankyrin Repeat Proteins. J Biol Chem and Characterization of Her2 Binding-designed Ankyrin Repeat Proteins. J Biol Chem 281, 35167-35175. 281, 35167-35175. 46. Zahnd, C., Sarkar, C. A. & Plückthun, A. (2010). Computational analysis of off-rate 46. Zahnd, C., Sarkar, C. A. & Plückthun, A. (2010). Computational analysis of off-rate selection experiments to optimize affinity maturation by directed evolution. Protein Eng selection experiments to optimize affinity maturation by directed evolution. Protein Eng Des Sel 23, 175-84. Des Sel 23, 175-84. 47. Zahnd, C., Spinelli, S., Luginbuhl, B., Amstutz, P., Cambillau, C. & Pluckthun, A. 47. Zahnd, C., Spinelli, S., Luginbuhl, B., Amstutz, P., Cambillau, C. & Pluckthun, A. (2004). Directed in vitro evolution and crystallographic analysis of a peptide-binding (2004). Directed in vitro evolution and crystallographic analysis of a peptide-binding single chain antibody fragment (scFv) with low picomolar affinity. J Biol Chem 279, single chain antibody fragment (scFv) with low picomolar affinity. J Biol Chem 279, 18870-7. 18870-7. 48. Tsui, W. C. & Fersht, A. R. (1981). Probing the principles of amino acid selection using 48. Tsui, W. C. & Fersht, A. R. (1981). Probing the principles of amino acid selection using the alanyl-tRNA synthetase from Escherichia coli. Nucleic Acids Res 9, 4627-37. the alanyl-tRNA synthetase from Escherichia coli. Nucleic Acids Res 9, 4627-37. 49. Wilks, H. M., Hart, K. W., Feeney, R., Dunn, C. R., MUIRHEAD, H., Chia, W. N., 49. Wilks, H. M., Hart, K. W., Feeney, R., Dunn, C. R., MUIRHEAD, H., Chia, W. N., Barstow, D. A., Atkinson, T., Clarke, A. R. & Holbrook, J. J. (1988). A specific, highly Barstow, D. A., Atkinson, T., Clarke, A. R. & Holbrook, J. J. (1988). A specific, highly active malate dehydrogenase by redesign of a lactate dehydrogenase framework. active malate dehydrogenase by redesign of a lactate dehydrogenase framework. Science 242, 1541-4. Science 242, 1541-4. 50. Bostrom, J., Yu, S.-F., Kan, D., Appleton, B. A., Lee, C. V., Billeci, K., Man, W., 50. Bostrom, J., Yu, S.-F., Kan, D., Appleton, B. A., Lee, C. V., Billeci, K., Man, W., Peale, F., Ross, S., Wiesmann, C. & Fuh, G. (2009). Variants of the antibody herceptin Peale, F., Ross, S., Wiesmann, C. & Fuh, G. (2009). Variants of the antibody herceptin that interact with HER2 and VEGF at the antigen binding site. Science 323, 1610-4. that interact with HER2 and VEGF at the antigen binding site. Science 323, 1610-4. 51. Kleanthous, C. (2000). Protein-protein recognition, Oxford University Press, Oxford. 51. Kleanthous, C. (2000). Protein-protein recognition, Oxford University Press, Oxford. 52. Ekins, R. P. (1989). Multi-analyte immunoassay. J Pharm Biomed Anal 7, 155-68. 52. Ekins, R. P. (1989). Multi-analyte immunoassay. J Pharm Biomed Anal 7, 155-68. 53. Velazquez-Campoy, A., Ohtaka, H., Nezami, A., Muzammil, S. & Freire, E. (2004). 53. Velazquez-Campoy, A., Ohtaka, H., Nezami, A., Muzammil, S. & Freire, E. (2004). Isothermal titration calorimetry. Curr Protoc Cell Biol Chapter 17, Unit 17 8. Isothermal titration calorimetry. Curr Protoc Cell Biol Chapter 17, Unit 17 8. 54. Bobrovnik, S. A. (2003). Determination of antibody affinity by ELISA. Theory. J 54. Bobrovnik, S. A. (2003). Determination of antibody affinity by ELISA. Theory. J Biochem Biophys Methods 57, 213-36. Biochem Biophys Methods 57, 213-36. 55. Gai, S. A. & Wittrup, K. D. (2007). Yeast surface display for protein engineering and 55. Gai, S. A. & Wittrup, K. D. (2007). Yeast surface display for protein engineering and characterization. Curr Opin Struct Biol 17, 467-73. characterization. Curr Opin Struct Biol 17, 467-73. 56. Van Regenmortel, M. H., Altschuh, D., Chatellier, J., Christensen, L., Rauffer-Bruyere, 56. Van Regenmortel, M. H., Altschuh, D., Chatellier, J., Christensen, L., Rauffer-Bruyere, N., Richalet-Secordel, P., Witz, J. & Zeder-Lutz, G. (1998). Measurement of antigen- N., Richalet-Secordel, P., Witz, J. & Zeder-Lutz, G. (1998). Measurement of antigen- antibody interactions with biosensors. J Mol Recognit 11, 163-7. antibody interactions with biosensors. J Mol Recognit 11, 163-7. 57. Cooper, M. A. (2002). Optical biosensors in drug discovery. Nat Rev Drug Discov 1, 57. Cooper, M. A. (2002). Optical biosensors in drug discovery. Nat Rev Drug Discov 1, 515-28. 515-28. 58. Papalia, G. A., Baer, M., Luehrsen, K., Nordin, H., Flynn, P. & Myszka, D. G. (2006). 58. Papalia, G. A., Baer, M., Luehrsen, K., Nordin, H., Flynn, P. & Myszka, D. G. (2006). High-resolution characterization of antibody fragment/antigen interactions using High-resolution characterization of antibody fragment/antigen interactions using Biacore T100. Anal Biochem 359, 112-9. Biacore T100. Anal Biochem 359, 112-9. 59. Dixon, M. C. (2008). Quartz crystal microbalance with dissipation monitoring: enabling 59. Dixon, M. C. (2008). Quartz crystal microbalance with dissipation monitoring: enabling real-time characterization of biological materials and their interactions. J Biomol Tech real-time characterization of biological materials and their interactions. J Biomol Tech 19, 151-8. 19, 151-8. 60. Lin, T. W., Hsieh, P. J., Lin, C. L., Fang, Y. Y., Yang, J. X., Tsai, C. C., Chiang, P. L., 60. Lin, T. W., Hsieh, P. J., Lin, C. L., Fang, Y. Y., Yang, J. X., Tsai, C. C., Chiang, P. L., Pan, C. Y. & Chen, Y. T. (2010). Label-free detection of protein-protein interactions Pan, C. Y. & Chen, Y. T. (2010). Label-free detection of protein-protein interactions Sebastian Grimm 85 Sebastian Grimm 85

using a calmodulin-modified nanowire transistor. Proc Natl Acad Sci USA 107, 1047- using a calmodulin-modified nanowire transistor. Proc Natl Acad Sci USA 107, 1047- 52. 52. 61. Michaud, G. A., Salcius, M., Zhou, F., Bangham, R., Bonin, J., Guo, H., Snyder, M., 61. Michaud, G. A., Salcius, M., Zhou, F., Bangham, R., Bonin, J., Guo, H., Snyder, M., Predki, P. F. & Schweitzer, B. I. (2003). Analyzing antibody specificity with whole Predki, P. F. & Schweitzer, B. I. (2003). Analyzing antibody specificity with whole proteome microarrays. Nat Biotechnol 21, 1509-12. proteome microarrays. Nat Biotechnol 21, 1509-12. 62. Fersht, A. (1999). Structure and mechanism in protein science : a guide to enzyme 62. Fersht, A. (1999). Structure and mechanism in protein science : a guide to enzyme catalysis and protein folding, W.H. Freeman, New York ; Basingstoke. catalysis and protein folding, W.H. Freeman, New York ; Basingstoke. 63. Mahler, H. C., Friess, W., Grauschopf, U. & Kiese, S. (2009). Protein aggregation: 63. Mahler, H. C., Friess, W., Grauschopf, U. & Kiese, S. (2009). Protein aggregation: pathways, induction factors and analysis. J Pharm Sci 98, 2909-34. pathways, induction factors and analysis. J Pharm Sci 98, 2909-34. 64. Kwon, W. S., Da Silva, N. A. & Kellis, J. T., Jr. (1996). Relationship between thermal 64. Kwon, W. S., Da Silva, N. A. & Kellis, J. T., Jr. (1996). Relationship between thermal stability, degradation rate and expression yield of barnase variants in the periplasm of stability, degradation rate and expression yield of barnase variants in the periplasm of Escherichia coli. Protein Eng Des Sel 9, 1197-202. Escherichia coli. Protein Eng Des Sel 9, 1197-202. 65. Parthasarathy, S. & Murthy, M. R. (2000). Protein thermal stability: insights from 65. Parthasarathy, S. & Murthy, M. R. (2000). Protein thermal stability: insights from atomic displacement parameters (B values). Protein Eng Des Sel 13, 9-13. atomic displacement parameters (B values). Protein Eng Des Sel 13, 9-13. 66. Renugopalakrishnan, V. & Lewis, R. V. (2006). Bionanotechnology: Proteins to 66. Renugopalakrishnan, V. & Lewis, R. V. (2006). Bionanotechnology: Proteins to Nanodevices, Springer, Dordrecht. Nanodevices, Springer, Dordrecht. 67. Schellman, J. A. (1994). The thermodynamics of solvent exchange. Biopolymers 34, 67. Schellman, J. A. (1994). The thermodynamics of solvent exchange. Biopolymers 34, 1015-26. 1015-26. 68. Jung, S., Honegger, A. & Plückthun, A. (1999). Selection for improved protein stability 68. Jung, S., Honegger, A. & Plückthun, A. (1999). Selection for improved protein stability by phage display. J Mol Biol 294, 163-80. by phage display. J Mol Biol 294, 163-80. 69. Tollinger, M., Crowhurst, K. A., Kay, L. E. & Forman-Kay, J. D. (2003). Site-specific 69. Tollinger, M., Crowhurst, K. A., Kay, L. E. & Forman-Kay, J. D. (2003). Site-specific contributions to the pH dependence of protein stability. Proc Natl Acad Sci USA 100, contributions to the pH dependence of protein stability. Proc Natl Acad Sci USA 100, 4545-50. 4545-50. 70. Privalov, P. L., Tiktopulo, E. I., Venyaminov, S., Griko Yu, V., Makhatadze, G. I. & 70. Privalov, P. L., Tiktopulo, E. I., Venyaminov, S., Griko Yu, V., Makhatadze, G. I. & Khechinashvili, N. N. (1989). Heat capacity and conformation of proteins in the Khechinashvili, N. N. (1989). Heat capacity and conformation of proteins in the denatured state. J Mol Biol 205, 737-50. denatured state. J Mol Biol 205, 737-50. 71. Radzicka, A. & Wolfenden, R. (1995). A proficient enzyme. Science 267, 90-3. 71. Radzicka, A. & Wolfenden, R. (1995). A proficient enzyme. Science 267, 90-3. 72. Romero, P., Wagg, J., Green, M. L., Kaiser, D., Krummenacker, M. & Karp, P. D. 72. Romero, P., Wagg, J., Green, M. L., Kaiser, D., Krummenacker, M. & Karp, P. D. (2005). Computational prediction of human metabolic pathways from the complete (2005). Computational prediction of human metabolic pathways from the complete human genome. Genome Biol 6, R2. human genome. Genome Biol 6, R2. 73. Wen, T. N., Chen, J. L., Lee, S. H., Yang, N. S. & Shyur, L. F. (2005). A truncated 73. Wen, T. N., Chen, J. L., Lee, S. H., Yang, N. S. & Shyur, L. F. (2005). A truncated Fibrobacter succinogenes 1,3-1,4-beta-d-glucanase with improved enzymatic activity Fibrobacter succinogenes 1,3-1,4-beta-d-glucanase with improved enzymatic activity and thermotolerance. Biochemistry 44, 9197-205. and thermotolerance. Biochemistry 44, 9197-205. 74. Glieder, A., Farinas, E. T. & Arnold, F. H. (2002). Laboratory evolution of a soluble, 74. Glieder, A., Farinas, E. T. & Arnold, F. H. (2002). Laboratory evolution of a soluble, self-sufficient, highly active alkane hydroxylase. Nat Biotechnol 20, 1135-9. self-sufficient, highly active alkane hydroxylase. Nat Biotechnol 20, 1135-9. 75. Chen, K. & Arnold, F. H. (1993). Tuning the activity of an enzyme for unusual 75. Chen, K. & Arnold, F. H. (1993). Tuning the activity of an enzyme for unusual environments: sequential random mutagenesis of subtilisin E for catalysis in environments: sequential random mutagenesis of subtilisin E for catalysis in dimethylformamide. Proc Natl Acad Sci USA 90, 5618-22. dimethylformamide. Proc Natl Acad Sci USA 90, 5618-22. 76. Frey, P. A. & Hegeman, A. D. (2007). Enzymatic reaction mechanisms, Oxford 76. Frey, P. A. & Hegeman, A. D. (2007). Enzymatic reaction mechanisms, Oxford University Press, Oxford ;. University Press, Oxford ;. 77. Lakowicz, J. R. (2006). Principles of fluorescence spectroscopy, Springer, New York. 77. Lakowicz, J. R. (2006). Principles of fluorescence spectroscopy, Springer, New York. 78. Harris, T. K. & Keshwani, M. M. (2009). Measurement of enzyme activity. Methods 78. Harris, T. K. & Keshwani, M. M. (2009). Measurement of enzyme activity. Methods Enzymol 463, 57-71. Enzymol 463, 57-71. 79. Sanger, F. & Coulson, A. R. (1975). A rapid method for determining sequences in DNA 79. Sanger, F. & Coulson, A. R. (1975). A rapid method for determining sequences in DNA by primed synthesis with DNA polymerase. J Mol Biol 94, 441-8. by primed synthesis with DNA polymerase. J Mol Biol 94, 441-8. 80. Ronaghi, M. (2001). Pyrosequencing sheds light on DNA sequencing. Genome Res 11, 80. Ronaghi, M. (2001). Pyrosequencing sheds light on DNA sequencing. Genome Res 11, 3-11. 3-11. 81. Schuster, S. C. (2008). Next-generation sequencing transforms today's biology. Nat 81. Schuster, S. C. (2008). Next-generation sequencing transforms today's biology. Nat Methods 5, 16-8. Methods 5, 16-8. 86 Ribosome display for selection and evolution of affibody molecules 86 Ribosome display for selection and evolution of affibody molecules

82. Doppalapudi, V. R., Tryder, N., Li, L., Aja, T., Griffith, D., Liao, F. F., Roxas, G., 82. Doppalapudi, V. R., Tryder, N., Li, L., Aja, T., Griffith, D., Liao, F. F., Roxas, G., Ramprasad, M. P., Bradshaw, C. & Barbas, C. F., 3rd. (2007). Chemically programmed Ramprasad, M. P., Bradshaw, C. & Barbas, C. F., 3rd. (2007). Chemically programmed antibodies: endothelin receptor targeting CovX-Bodies. Bioorg Med Chem Lett 17, 501- antibodies: endothelin receptor targeting CovX-Bodies. Bioorg Med Chem Lett 17, 501- 6. 6. 83. Hammarström, M., Hellgren, N., van Den Berg, S., Berglund, H. & Härd, T. (2002). 83. Hammarström, M., Hellgren, N., van Den Berg, S., Berglund, H. & Härd, T. (2002). Rapid screening for improved solubility of small human proteins produced as fusion Rapid screening for improved solubility of small human proteins produced as fusion proteins in Escherichia coli. Protein Sci 11, 313-21. proteins in Escherichia coli. Protein Sci 11, 313-21. 84. Terskikh, A. V., Le Doussal, J. M., Crameri, R., Fisch, I., Mach, J. P. & Kajava, A. V. 84. Terskikh, A. V., Le Doussal, J. M., Crameri, R., Fisch, I., Mach, J. P. & Kajava, A. V. (1997). "Peptabody": a new type of high avidity binding protein. Proc Natl Acad Sci (1997). "Peptabody": a new type of high avidity binding protein. Proc Natl Acad Sci USA 94, 1663-8. USA 94, 1663-8. 85. Pastan, I., Hassan, R., Fitzgerald, D. J. & Kreitman, R. J. (2006). Immunotoxin therapy 85. Pastan, I., Hassan, R., Fitzgerald, D. J. & Kreitman, R. J. (2006). Immunotoxin therapy of cancer. Nat Rev Cancer 6, 559-65. of cancer. Nat Rev Cancer 6, 559-65. 86. Schellenberger, V., Wang, C. W., Geething, N. C., Spink, B. J., Campbell, A., To, W., 86. Schellenberger, V., Wang, C. W., Geething, N. C., Spink, B. J., Campbell, A., To, W., Scholle, M. D., Yin, Y., Yao, Y., Bogin, O., Cleland, J. L., Silverman, J. & Stemmer, Scholle, M. D., Yin, Y., Yao, Y., Bogin, O., Cleland, J. L., Silverman, J. & Stemmer, W. P. (2009). A recombinant polypeptide extends the in vivo half-life of peptides and W. P. (2009). A recombinant polypeptide extends the in vivo half-life of peptides and proteins in a tunable manner. Nat Biotechnol 27, 1186-90. proteins in a tunable manner. Nat Biotechnol 27, 1186-90. 87. Roopenian, D. C. & Akilesh, S. (2007). FcRn: the neonatal Fc receptor comes of age. 87. Roopenian, D. C. & Akilesh, S. (2007). FcRn: the neonatal Fc receptor comes of age. Nat Rev Immunol 7, 715-25. Nat Rev Immunol 7, 715-25. 88. Leatherbarrow, R. J., Fersht, A. R. & Winter, G. (1985). Transition-state stabilization in 88. Leatherbarrow, R. J., Fersht, A. R. & Winter, G. (1985). Transition-state stabilization in the mechanism of tyrosyl-tRNA synthetase revealed by protein engineering. Proc Natl the mechanism of tyrosyl-tRNA synthetase revealed by protein engineering. Proc Natl Acad Sci USA 82, 7840-4. Acad Sci USA 82, 7840-4. 89. Sheldon, B., Russell-Jones, D. & Wright, J. (2009). Insulin analogues: an example of 89. Sheldon, B., Russell-Jones, D. & Wright, J. (2009). Insulin analogues: an example of applied medical science. Diabetes Obes Metab 11, 5-19. applied medical science. Diabetes Obes Metab 11, 5-19. 90. Morrison, K. L. & Weiss, G. A. (2001). Combinatorial alanine-scanning. Curr Opin 90. Morrison, K. L. & Weiss, G. A. (2001). Combinatorial alanine-scanning. Curr Opin Chem Biol 5, 302-7. Chem Biol 5, 302-7. 91. Mandell, D. J. & Kortemme, T. (2009). Computer-aided design of functional protein 91. Mandell, D. J. & Kortemme, T. (2009). Computer-aided design of functional protein interactions. Nat Chem Biol 5, 797-807. interactions. Nat Chem Biol 5, 797-807. 92. Hellinga, H. W. (1996). Metalloprotein design. Curr Opin Biotechnol 7, 437-41. 92. Hellinga, H. W. (1996). Metalloprotein design. Curr Opin Biotechnol 7, 437-41. 93. Hellinga, H. W., Caradonna, J. P. & Richards, F. M. (1991). Construction of new ligand 93. Hellinga, H. W., Caradonna, J. P. & Richards, F. M. (1991). Construction of new ligand binding sites in proteins of known structure. II. Grafting of a buried transition metal binding sites in proteins of known structure. II. Grafting of a buried transition metal binding site into Escherichia coli thioredoxin. J Mol Biol 222, 787-803. binding site into Escherichia coli thioredoxin. J Mol Biol 222, 787-803. 94. Reina, J., Lacroix, E., Hobson, S. D., Fernandez-Ballester, G., Rybin, V., Schwab, M. 94. Reina, J., Lacroix, E., Hobson, S. D., Fernandez-Ballester, G., Rybin, V., Schwab, M. S., Serrano, L. & Gonzalez, C. (2002). Computer-aided design of a PDZ domain to S., Serrano, L. & Gonzalez, C. (2002). Computer-aided design of a PDZ domain to recognize new target sequences. Nat Struct Biol 9, 621-7. recognize new target sequences. Nat Struct Biol 9, 621-7. 95. Vanhercke, T., Ampe, C., Tirry, L. & Denolf, P. (2005). Reducing mutational bias in 95. Vanhercke, T., Ampe, C., Tirry, L. & Denolf, P. (2005). Reducing mutational bias in random protein libraries. Anal Biochem 339, 9-14. random protein libraries. Anal Biochem 339, 9-14. 96. Riechmann, L., Clark, M., Waldmann, H. & Winter, G. (1988). Reshaping human 96. Riechmann, L., Clark, M., Waldmann, H. & Winter, G. (1988). Reshaping human antibodies for therapy. Nature 332, 323-7. antibodies for therapy. Nature 332, 323-7. 97. Bradbury, A. R. M. & Marks, J. D. (2004). Antibodies from phage antibody libraries. J 97. Bradbury, A. R. M. & Marks, J. D. (2004). Antibodies from phage antibody libraries. J Immunol Methods 290, 29-49. Immunol Methods 290, 29-49. 98. Munoz, E. & Deem, M. W. (2008). Amino acid alphabet size in protein evolution 98. Munoz, E. & Deem, M. W. (2008). Amino acid alphabet size in protein evolution experiments: better to search a small library thoroughly or a large library sparsely? experiments: better to search a small library thoroughly or a large library sparsely? Protein Eng Des Sel 21, 311-7. Protein Eng Des Sel 21, 311-7. 99. Romero, P. A. & Arnold, F. H. (2009). Exploring protein fitness landscapes by directed 99. Romero, P. A. & Arnold, F. H. (2009). Exploring protein fitness landscapes by directed evolution. Nat Rev Mol Cell Biol 10, 866-76. evolution. Nat Rev Mol Cell Biol 10, 866-76. 100. Lo Conte, L., Chothia, C. & Janin, J. (1999). The atomic structure of protein-protein 100. Lo Conte, L., Chothia, C. & Janin, J. (1999). The atomic structure of protein-protein recognition sites. J Mol Biol 285, 2177-98. recognition sites. J Mol Biol 285, 2177-98. 101. Fellouse, F. A., Esaki, K., Birtalan, S., Raptis, D., Cancasci, V. J., Koide, A., Jhurani, 101. Fellouse, F. A., Esaki, K., Birtalan, S., Raptis, D., Cancasci, V. J., Koide, A., Jhurani, P., Vasser, M., Wiesmann, C., Kossiakoff, A. A., Koide, S. & Sidhu, S. S. (2007). P., Vasser, M., Wiesmann, C., Kossiakoff, A. A., Koide, S. & Sidhu, S. S. (2007). Sebastian Grimm 87 Sebastian Grimm 87

High-throughput generation of synthetic antibodies from highly functional minimalist High-throughput generation of synthetic antibodies from highly functional minimalist phage-displayed libraries. J Mol Biol 373, 924-40. phage-displayed libraries. J Mol Biol 373, 924-40. 102. Hackel, B. J. & Wittrup, K. D. (2010). The full amino acid repertoire is superior to 102. Hackel, B. J. & Wittrup, K. D. (2010). The full amino acid repertoire is superior to serine/tyrosine for selection of high affinity immunoglobulin G binders from the serine/tyrosine for selection of high affinity immunoglobulin G binders from the fibronectin scaffold. Protein Eng Des Sel 23, 211-9. fibronectin scaffold. Protein Eng Des Sel 23, 211-9. 103. Huse, W. D., Sastry, L., Iverson, S. A., Kang, A. S., Alting-Mees, M., Burton, D. R., 103. Huse, W. D., Sastry, L., Iverson, S. A., Kang, A. S., Alting-Mees, M., Burton, D. R., Benkovic, S. J. & Lerner, R. A. (1989). Generation of a large combinatorial library of Benkovic, S. J. & Lerner, R. A. (1989). Generation of a large combinatorial library of the immunoglobulin repertoire in phage lambda. Science 246, 1275-81. the immunoglobulin repertoire in phage lambda. Science 246, 1275-81. 104. Dooley, H., Flajnik, M. F. & Porter, A. J. (2003). Selection and characterization of 104. Dooley, H., Flajnik, M. F. & Porter, A. J. (2003). Selection and characterization of naturally occurring single-domain (IgNAR) antibody fragments from immunized sharks naturally occurring single-domain (IgNAR) antibody fragments from immunized sharks by phage display. Mol Immunol 40, 25-33. by phage display. Mol Immunol 40, 25-33. 105. de Haard, H. J., van Neer, N., Reurs, A., Hufton, S. E., Roovers, R. C., Henderikx, P., 105. de Haard, H. J., van Neer, N., Reurs, A., Hufton, S. E., Roovers, R. C., Henderikx, P., de Bruïne, A. P., Arends, J. W. & Hoogenboom, H. R. (1999). A large non-immunized de Bruïne, A. P., Arends, J. W. & Hoogenboom, H. R. (1999). A large non-immunized human Fab fragment phage library that permits rapid isolation and kinetic analysis of human Fab fragment phage library that permits rapid isolation and kinetic analysis of high affinity antibodies. J Biol Chem 274, 18218-30. high affinity antibodies. J Biol Chem 274, 18218-30. 106. Fellouse, F. A., Li, B., Compaan, D. M., Peden, A. A., Hymowitz, S. G. & Sidhu, S. S. 106. Fellouse, F. A., Li, B., Compaan, D. M., Peden, A. A., Hymowitz, S. G. & Sidhu, S. S. (2005). Molecular recognition by a binary code. J Mol Biol 348, 1153-62. (2005). Molecular recognition by a binary code. J Mol Biol 348, 1153-62. 107. Rajpal, A. (2005). A general method for greatly improving the affinity of antibodies by 107. Rajpal, A. (2005). A general method for greatly improving the affinity of antibodies by using combinatorial libraries. Proc Natl Acad Sci USA 102, 8466-8471. using combinatorial libraries. Proc Natl Acad Sci USA 102, 8466-8471. 108. Glaser, S. M., Yelton, D. E. & Huse, W. D. (1992). Antibody engineering by codon- 108. Glaser, S. M., Yelton, D. E. & Huse, W. D. (1992). Antibody engineering by codon- based mutagenesis in a filamentous phage vector system. J Immunol 149, 3903-13. based mutagenesis in a filamentous phage vector system. J Immunol 149, 3903-13. 109. Virnekäs, B., Ge, L., Plückthun, A., Schneider, K. C., Wellnhofer, G. & Moroney, S. E. 109. Virnekäs, B., Ge, L., Plückthun, A., Schneider, K. C., Wellnhofer, G. & Moroney, S. E. (1994). Trinucleotide phosphoramidites: ideal reagents for the synthesis of mixed (1994). Trinucleotide phosphoramidites: ideal reagents for the synthesis of mixed oligonucleotides for random mutagenesis. Nucleic Acids Res 22, 5600-7. oligonucleotides for random mutagenesis. Nucleic Acids Res 22, 5600-7. 110. Tindall, K. R. & Kunkel, T. A. (1988). Fidelity of DNA synthesis by the Thermus 110. Tindall, K. R. & Kunkel, T. A. (1988). Fidelity of DNA synthesis by the Thermus aquaticus DNA polymerase. Biochemistry 27, 6008-13. aquaticus DNA polymerase. Biochemistry 27, 6008-13. 111. Vartanian, J. P., Henry, M. & Wain-Hobson, S. (1996). Hypermutagenic PCR involving 111. Vartanian, J. P., Henry, M. & Wain-Hobson, S. (1996). Hypermutagenic PCR involving all four transitions and a sizeable proportion of transversions. Nucleic Acids Res 24, all four transitions and a sizeable proportion of transversions. Nucleic Acids Res 24, 2627-31. 2627-31. 112. Eckert, K. A. & Kunkel, T. A. (1990). High fidelity DNA synthesis by the Thermus 112. Eckert, K. A. & Kunkel, T. A. (1990). High fidelity DNA synthesis by the Thermus aquaticus DNA polymerase. Nucleic Acids Res 18, 3739-44. aquaticus DNA polymerase. Nucleic Acids Res 18, 3739-44. 113. Zaccolo, M., Williams, D. M., Brown, D. M. & Gherardi, E. (1996). An approach to 113. Zaccolo, M., Williams, D. M., Brown, D. M. & Gherardi, E. (1996). An approach to random mutagenesis of DNA using mixtures of triphosphate derivatives of nucleoside random mutagenesis of DNA using mixtures of triphosphate derivatives of nucleoside analogues. J Mol Biol 255, 589-603. analogues. J Mol Biol 255, 589-603. 114. Spee, J. H., de Vos, W. M. & Kuipers, O. P. (1993). Efficient random mutagenesis 114. Spee, J. H., de Vos, W. M. & Kuipers, O. P. (1993). Efficient random mutagenesis method with adjustable mutation frequency by use of PCR and dITP. Nucleic Acids Res method with adjustable mutation frequency by use of PCR and dITP. Nucleic Acids Res 21, 777-8. 21, 777-8. 115. Biles, B. D. & Connolly, B. A. (2004). Low-fidelity Pyrococcus furiosus DNA 115. Biles, B. D. & Connolly, B. A. (2004). Low-fidelity Pyrococcus furiosus DNA polymerase mutants useful in error-prone PCR. Nucleic Acids Res 32, e176. polymerase mutants useful in error-prone PCR. Nucleic Acids Res 32, e176. 116. Wong, T. S., Roccatano, D., Loakes, D., Tee, K. L., Schenk, A., Hauer, B. & 116. Wong, T. S., Roccatano, D., Loakes, D., Tee, K. L., Schenk, A., Hauer, B. & Schwaneberg, U. (2008). Transversion-enriched sequence saturation mutagenesis Schwaneberg, U. (2008). Transversion-enriched sequence saturation mutagenesis (SeSaM-Tv+): a random mutagenesis method with consecutive nucleotide exchanges (SeSaM-Tv+): a random mutagenesis method with consecutive nucleotide exchanges that complements the bias of error-prone PCR. Biotechnol J 3, 74-82. that complements the bias of error-prone PCR. Biotechnol J 3, 74-82. 117. Patrick, W. M. & Firth, A. E. (2005). Strategies and computational tools for improving 117. Patrick, W. M. & Firth, A. E. (2005). Strategies and computational tools for improving randomized protein libraries. Biomol Eng 22, 105-12. randomized protein libraries. Biomol Eng 22, 105-12. 118. Bershtein, S., Segal, M., Bekerman, R., Tokuriki, N. & Tawfik, D. S. (2006). 118. Bershtein, S., Segal, M., Bekerman, R., Tokuriki, N. & Tawfik, D. S. (2006). Robustness-epistasis link shapes the fitness landscape of a randomly drifting protein. Robustness-epistasis link shapes the fitness landscape of a randomly drifting protein. Nature 444, 929-32. Nature 444, 929-32. 119. Stemmer, W. P. (1994). DNA shuffling by random fragmentation and reassembly: in 119. Stemmer, W. P. (1994). DNA shuffling by random fragmentation and reassembly: in vitro recombination for molecular evolution. Proc Natl Acad Sci USA 91, 10747-51. vitro recombination for molecular evolution. Proc Natl Acad Sci USA 91, 10747-51. 88 Ribosome display for selection and evolution of affibody molecules 88 Ribosome display for selection and evolution of affibody molecules

120. Crameri, A., Raillard, S. A., Bermudez, E. & Stemmer, W. P. (1998). DNA shuffling of 120. Crameri, A., Raillard, S. A., Bermudez, E. & Stemmer, W. P. (1998). DNA shuffling of a family of genes from diverse species accelerates directed evolution. Nature 391, 288- a family of genes from diverse species accelerates directed evolution. Nature 391, 288- 91. 91. 121. Zhao, H. & Zha, W. (2006). In vitro 'sexual' evolution through the PCR-based staggered 121. Zhao, H. & Zha, W. (2006). In vitro 'sexual' evolution through the PCR-based staggered extension process (StEP). Nat Protoc 1, 1865-71. extension process (StEP). Nat Protoc 1, 1865-71. 122. Coco, W. M., Levinson, W. E., Crist, M. J., Hektor, H. J., Darzins, A., Pienkos, P. T., 122. Coco, W. M., Levinson, W. E., Crist, M. J., Hektor, H. J., Darzins, A., Pienkos, P. T., Squires, C. H. & Monticello, D. J. (2001). DNA shuffling method for generating highly Squires, C. H. & Monticello, D. J. (2001). DNA shuffling method for generating highly recombined genes and evolved enzymes. Nat Biotechnol 19, 354-9. recombined genes and evolved enzymes. Nat Biotechnol 19, 354-9. 123. Tabata, N., Sakuma, Y., Honda, Y., Doi, N., Takashima, H., Miyamoto-Sato, E. & 123. Tabata, N., Sakuma, Y., Honda, Y., Doi, N., Takashima, H., Miyamoto-Sato, E. & Yanagawa, H. (2009). Rapid antibody selection by mRNA display on a microfluidic Yanagawa, H. (2009). Rapid antibody selection by mRNA display on a microfluidic chip. Nucleic Acids Res 37, e64. chip. Nucleic Acids Res 37, e64. 124. Skerra, A. & Plückthun, A. (1988). Assembly of a functional immunoglobulin Fv 124. Skerra, A. & Plückthun, A. (1988). Assembly of a functional immunoglobulin Fv fragment in Escherichia coli. Science 240, 1038-41. fragment in Escherichia coli. Science 240, 1038-41. 125. Mersmann, M., Meier, D., Mersmann, J., Helmsing, S., Nilsson, P., Graslund, S., 125. Mersmann, M., Meier, D., Mersmann, J., Helmsing, S., Nilsson, P., Graslund, S., Colwill, K., Hust, M. & Dubel, S. (2009). Towards proteome scale antibody selections Colwill, K., Hust, M. & Dubel, S. (2009). Towards proteome scale antibody selections using phage display. N Biotechnol 27, 118-28. using phage display. N Biotechnol 27, 118-28. 126. Scott, J. K. & Smith, G. P. (1990). Searching for peptide ligands with an epitope library. 126. Scott, J. K. & Smith, G. P. (1990). Searching for peptide ligands with an epitope library. Science 249, 386-90. Science 249, 386-90. 127. Bass, S., Greene, R. & Wells, J. A. (1990). Hormone phage: an enrichment method for 127. Bass, S., Greene, R. & Wells, J. A. (1990). Hormone phage: an enrichment method for variant proteins with altered binding properties. Proteins 8, 309-14. variant proteins with altered binding properties. Proteins 8, 309-14. 128. McCafferty, J., Griffiths, A. D., Winter, G. & Chiswell, D. J. (1990). Phage antibodies: 128. McCafferty, J., Griffiths, A. D., Winter, G. & Chiswell, D. J. (1990). Phage antibodies: filamentous phage displaying antibody variable domains. Nature 348, 552-4. filamentous phage displaying antibody variable domains. Nature 348, 552-4. 129. Clackson, T., Hoogenboom, H. R., Griffiths, A. D. & Winter, G. (1991). Making 129. Clackson, T., Hoogenboom, H. R., Griffiths, A. D. & Winter, G. (1991). Making antibody fragments using phage display libraries. Nature 352, 624-8. antibody fragments using phage display libraries. Nature 352, 624-8. 130. Pedersen, H., Holder, S., Sutherlin, D. P., Schwitter, U., King, D. S. & Schultz, P. G. 130. Pedersen, H., Holder, S., Sutherlin, D. P., Schwitter, U., King, D. S. & Schultz, P. G. (1998). A method for directed evolution and functional cloning of enzymes. Proc Natl (1998). A method for directed evolution and functional cloning of enzymes. Proc Natl Acad Sci USA 95, 10523-8. Acad Sci USA 95, 10523-8. 131. Sieber, V., Pluckthun, A. & Schmid, F. X. (1998). Selecting proteins with improved 131. Sieber, V., Pluckthun, A. & Schmid, F. X. (1998). Selecting proteins with improved stability by a phage-based method. Nat Biotechnol 16, 955-60. stability by a phage-based method. Nat Biotechnol 16, 955-60. 132. Hemminki, A., Niemi, S., Hoffren, A. M., Hakalahti, L., Soderlund, H. & Takkinen, K. 132. Hemminki, A., Niemi, S., Hoffren, A. M., Hakalahti, L., Soderlund, H. & Takkinen, K. (1998). Specificity improvement of a recombinant anti-testosterone Fab fragment by (1998). Specificity improvement of a recombinant anti-testosterone Fab fragment by CDRIII mutagenesis and phage display selection. Protein Eng Des Sel 11, 311-9. CDRIII mutagenesis and phage display selection. Protein Eng Des Sel 11, 311-9. 133. Pasqualini, R. & Ruoslahti, E. (1996). Organ targeting in vivo using phage display 133. Pasqualini, R. & Ruoslahti, E. (1996). Organ targeting in vivo using phage display peptide libraries. Nature 380, 364-6. peptide libraries. Nature 380, 364-6. 134. Joyce, J. A., Laakkonen, P., Bernasconi, M., Bergers, G., Ruoslahti, E. & Hanahan, D. 134. Joyce, J. A., Laakkonen, P., Bernasconi, M., Bergers, G., Ruoslahti, E. & Hanahan, D. (2003). Stage-specific vascular markers revealed by phage display in a mouse model of (2003). Stage-specific vascular markers revealed by phage display in a mouse model of pancreatic islet tumorigenesis. Cancer Cell 4, 393-403. pancreatic islet tumorigenesis. Cancer Cell 4, 393-403. 135. Barbas, C. F. (2000). Phage display : a laboratory manual, Cold Spring Harbor 135. Barbas, C. F. (2000). Phage display : a laboratory manual, Cold Spring Harbor Labortory Press, Cold Spring Harbor, NY. Labortory Press, Cold Spring Harbor, NY. 136. Messing, J. (1991). Cloning in M13 phage or how to use biology at its best. Gene 100, 136. Messing, J. (1991). Cloning in M13 phage or how to use biology at its best. Gene 100, 3-12. 3-12. 137. Smith, G. P. & Scott, J. K. (1993). Libraries of peptides and proteins displayed on 137. Smith, G. P. & Scott, J. K. (1993). Libraries of peptides and proteins displayed on filamentous phage. Methods Enzymol 217, 228-57. filamentous phage. Methods Enzymol 217, 228-57. 138. Iannolo, G., Minenkova, O., Petruzzelli, R. & Cesareni, G. (1995). Modifying 138. Iannolo, G., Minenkova, O., Petruzzelli, R. & Cesareni, G. (1995). Modifying filamentous phage capsid: limits in the size of the major capsid protein. J Mol Biol 248, filamentous phage capsid: limits in the size of the major capsid protein. J Mol Biol 248, 835-44. 835-44. 139. Malik, P., Terry, T. D., Gowda, L. R., Langara, A., Petukhov, S. A., Symmons, M. F., 139. Malik, P., Terry, T. D., Gowda, L. R., Langara, A., Petukhov, S. A., Symmons, M. F., Welsh, L. C., Marvin, D. A. & Perham, R. N. (1996). Role of capsid structure and Welsh, L. C., Marvin, D. A. & Perham, R. N. (1996). Role of capsid structure and membrane protein processing in determining the size and copy number of peptides membrane protein processing in determining the size and copy number of peptides displayed on the major coat protein of filamentous bacteriophage. J Mol Biol 260, 9-21. displayed on the major coat protein of filamentous bacteriophage. J Mol Biol 260, 9-21. Sebastian Grimm 89 Sebastian Grimm 89

140. Steiner, D., Forrer, P., Stumpp, M. T. & Pluckthun, A. (2006). Signal sequences 140. Steiner, D., Forrer, P., Stumpp, M. T. & Pluckthun, A. (2006). Signal sequences directing cotranslational translocation expand the range of proteins amenable to phage directing cotranslational translocation expand the range of proteins amenable to phage display. Nat Biotechnol 24, 823-31. display. Nat Biotechnol 24, 823-31. 141. Thie, H., Schirrmann, T., Paschke, M., Dubel, S. & Hust, M. (2008). SRP and Sec 141. Thie, H., Schirrmann, T., Paschke, M., Dubel, S. & Hust, M. (2008). SRP and Sec pathway leader peptides for antibody phage display and antibody fragment production pathway leader peptides for antibody phage display and antibody fragment production in E. coli. N Biotechnol 25, 49-54. in E. coli. N Biotechnol 25, 49-54. 142. Paschke, M. & Hohne, W. (2005). A twin-arginine translocation (Tat)-mediated phage 142. Paschke, M. & Hohne, W. (2005). A twin-arginine translocation (Tat)-mediated phage display system. Gene 350, 79-88. display system. Gene 350, 79-88. 143. Clackson, T. & Wells, J. A. (1994). In vitro selection from protein and peptide libraries. 143. Clackson, T. & Wells, J. A. (1994). In vitro selection from protein and peptide libraries. Trends Biotechnol 12, 173-84. Trends Biotechnol 12, 173-84. 144. Rondot, S., Koch, J., Breitling, F. & Dubel, S. (2001). A helper phage to improve 144. Rondot, S., Koch, J., Breitling, F. & Dubel, S. (2001). A helper phage to improve single-chain antibody presentation in phage display. Nat Biotechnol 19, 75-8. single-chain antibody presentation in phage display. Nat Biotechnol 19, 75-8. 145. Jestin, J. L., Volioti, G. & Winter, G. (2001). Improving the display of proteins on 145. Jestin, J. L., Volioti, G. & Winter, G. (2001). Improving the display of proteins on filamentous phage. Res Microbiol 152, 187-91. filamentous phage. Res Microbiol 152, 187-91. 146. Grimm, S., Lundberg, E., Yu, F., Shibasaki, S., Vernet, E., Skogs, M., Nygren, P. A. & 146. Grimm, S., Lundberg, E., Yu, F., Shibasaki, S., Vernet, E., Skogs, M., Nygren, P. A. & Graslund, T. Selection and characterisation of affibody molecules inhibiting the Graslund, T. Selection and characterisation of affibody molecules inhibiting the interaction between Ras and Raf in vitro. N Biotechnol 27, 766-73. interaction between Ras and Raf in vitro. N Biotechnol 27, 766-73. 147. McCafferty, J., Fitzgerald, K. J., Earnshaw, J., Chiswell, D. J., Link, J., Smith, R. & 147. McCafferty, J., Fitzgerald, K. J., Earnshaw, J., Chiswell, D. J., Link, J., Smith, R. & Kenten, J. (1994). Selection and rapid purification of murine antibody fragments that Kenten, J. (1994). Selection and rapid purification of murine antibody fragments that bind a transition-state analog by phage display. Appl Biochem Biotechnol 47, 157-71; bind a transition-state analog by phage display. Appl Biochem Biotechnol 47, 157-71; discussion 171-3. discussion 171-3. 148. Francisco, J. A., Earhart, C. F. & Georgiou, G. (1992). Transport and anchoring of beta- 148. Francisco, J. A., Earhart, C. F. & Georgiou, G. (1992). Transport and anchoring of beta- lactamase to the external surface of Escherichia coli. Proc Natl Acad Sci USA 89, 2713- lactamase to the external surface of Escherichia coli. Proc Natl Acad Sci USA 89, 2713- 7. 7. 149. Brown, S. (1992). Engineered iron oxide-adhesion mutants of the Escherichia coli 149. Brown, S. (1992). Engineered iron oxide-adhesion mutants of the Escherichia coli phage lambda receptor. Proc Natl Acad Sci USA 89, 8651-5. phage lambda receptor. Proc Natl Acad Sci USA 89, 8651-5. 150. Lu, Z., Murray, K. S., Van Cleave, V., LaVallie, E. R., Stahl, M. L. & McCoy, J. M. 150. Lu, Z., Murray, K. S., Van Cleave, V., LaVallie, E. R., Stahl, M. L. & McCoy, J. M. (1995). Expression of thioredoxin random peptide libraries on the Escherichia coli cell (1995). Expression of thioredoxin random peptide libraries on the Escherichia coli cell surface as functional fusions to flagellin: a system designed for exploring protein- surface as functional fusions to flagellin: a system designed for exploring protein- protein interactions. Biotechnology 13, 366-72. protein interactions. Biotechnology 13, 366-72. 151. Daugherty, P. S., Chen, G., Olsen, M. J., Iverson, B. L. & Georgiou, G. (1998). 151. Daugherty, P. S., Chen, G., Olsen, M. J., Iverson, B. L. & Georgiou, G. (1998). Antibody affinity maturation using bacterial surface display. Protein Eng Des Sel 11, Antibody affinity maturation using bacterial surface display. Protein Eng Des Sel 11, 825-32. 825-32. 152. Binder, U., Matschiner, G., Theobald, I. & Skerra, A. (2010). High-throughput sorting 152. Binder, U., Matschiner, G., Theobald, I. & Skerra, A. (2010). High-throughput sorting of an Anticalin library via EspP-mediated functional display on the Escherichia coli cell of an Anticalin library via EspP-mediated functional display on the Escherichia coli cell surface. J Mol Biol 400, 783-802. surface. J Mol Biol 400, 783-802. 153. Daugherty, P. S. (2007). Protein engineering with bacterial display. Curr Opin Struct 153. Daugherty, P. S. (2007). Protein engineering with bacterial display. Curr Opin Struct Biol 17, 474-80. Biol 17, 474-80. 154. Mazor, Y., Van Blarcom, T., Mabry, R., Iverson, B. L. & Georgiou, G. (2007). Isolation 154. Mazor, Y., Van Blarcom, T., Mabry, R., Iverson, B. L. & Georgiou, G. (2007). Isolation of engineered, full-length antibodies from libraries expressed in Escherichia coli. Nat of engineered, full-length antibodies from libraries expressed in Escherichia coli. Nat Biotechnol 25, 563-5. Biotechnol 25, 563-5. 155. Georgiou, G., Stathopoulos, C., Daugherty, P. S., Nayak, A. R., Iverson, B. L. & 155. Georgiou, G., Stathopoulos, C., Daugherty, P. S., Nayak, A. R., Iverson, B. L. & Curtiss, R. (1997). Display of heterologous proteins on the surface of microorganisms: Curtiss, R. (1997). Display of heterologous proteins on the surface of microorganisms: from the screening of combinatorial libraries to live recombinant vaccines. Nat from the screening of combinatorial libraries to live recombinant vaccines. Nat Biotechnol 15, 29-34. Biotechnol 15, 29-34. 156. Gibbs, K. A., Isaac, D. D., Xu, J., Hendrix, R. W., Silhavy, T. J. & Theriot, J. A. (2004). 156. Gibbs, K. A., Isaac, D. D., Xu, J., Hendrix, R. W., Silhavy, T. J. & Theriot, J. A. (2004). Complex spatial distribution and dynamics of an abundant Escherichia coli outer Complex spatial distribution and dynamics of an abundant Escherichia coli outer membrane protein, LamB. Mol Microbiol 53, 1771-83. membrane protein, LamB. Mol Microbiol 53, 1771-83. 157. Boder, E. T. & Wittrup, K. D. (1997). Yeast surface display for screening combinatorial 157. Boder, E. T. & Wittrup, K. D. (1997). Yeast surface display for screening combinatorial polypeptide libraries. Nat Biotechnol 15, 553-7. polypeptide libraries. Nat Biotechnol 15, 553-7. 90 Ribosome display for selection and evolution of affibody molecules 90 Ribosome display for selection and evolution of affibody molecules

158. Bowley, D. R., Labrijn, A. F., Zwick, M. B. & Burton, D. R. (2007). Antigen selection 158. Bowley, D. R., Labrijn, A. F., Zwick, M. B. & Burton, D. R. (2007). Antigen selection from an HIV-1 immune antibody library displayed on yeast yields many novel from an HIV-1 immune antibody library displayed on yeast yields many novel antibodies compared to selection from the same library displayed on phage. Protein Eng antibodies compared to selection from the same library displayed on phage. Protein Eng Des Sel 20, 81-90. Des Sel 20, 81-90. 159. Boder, E. T., Midelfort, K. S. & Wittrup, K. D. (2000). Directed evolution of antibody 159. Boder, E. T., Midelfort, K. S. & Wittrup, K. D. (2000). Directed evolution of antibody fragments with monovalent femtomolar antigen-binding affinity. Proc Natl Acad Sci fragments with monovalent femtomolar antigen-binding affinity. Proc Natl Acad Sci USA 97, 10701-5. USA 97, 10701-5. 160. Garcia-Rodriguez, C., Levy, R., Arndt, J. W., Forsyth, C. M., Razai, A., Lou, J., Geren, 160. Garcia-Rodriguez, C., Levy, R., Arndt, J. W., Forsyth, C. M., Razai, A., Lou, J., Geren, I., Stevens, R. C. & Marks, J. D. (2007). Molecular evolution of antibody cross- I., Stevens, R. C. & Marks, J. D. (2007). Molecular evolution of antibody cross- reactivity for two subtypes of type A botulinum neurotoxin. Nat Biotechnol 25, 107-16. reactivity for two subtypes of type A botulinum neurotoxin. Nat Biotechnol 25, 107-16. 161. Shusta, E. V., Kieke, M. C., Parke, E., Kranz, D. M. & Wittrup, K. D. (1999). Yeast 161. Shusta, E. V., Kieke, M. C., Parke, E., Kranz, D. M. & Wittrup, K. D. (1999). Yeast polypeptide fusion surface display levels predict thermal stability and soluble secretion polypeptide fusion surface display levels predict thermal stability and soluble secretion efficiency. J Mol Biol 292, 949-56. efficiency. J Mol Biol 292, 949-56. 162. Jones, L. L., Brophy, S. E., Bankovich, A. J., Colf, L. A., Hanick, N. A., Garcia, K. C. 162. Jones, L. L., Brophy, S. E., Bankovich, A. J., Colf, L. A., Hanick, N. A., Garcia, K. C. & Kranz, D. M. (2006). Engineering and characterization of a stabilized alpha1/alpha2 & Kranz, D. M. (2006). Engineering and characterization of a stabilized alpha1/alpha2 module of the class I major histocompatibility complex product Ld. J Biol Chem 281, module of the class I major histocompatibility complex product Ld. J Biol Chem 281, 25734-44. 25734-44. 163. Kim, Y. S., Bhandari, R., Cochran, J. R., Kuriyan, J. & Wittrup, K. D. (2006). Directed 163. Kim, Y. S., Bhandari, R., Cochran, J. R., Kuriyan, J. & Wittrup, K. D. (2006). Directed evolution of the epidermal growth factor receptor extracellular domain for expression in evolution of the epidermal growth factor receptor extracellular domain for expression in yeast. Proteins 62, 1026-35. yeast. Proteins 62, 1026-35. 164. Richman, S. A., Healan, S. J., Weber, K. S., Donermeyer, D. L., Dossett, M. L., 164. Richman, S. A., Healan, S. J., Weber, K. S., Donermeyer, D. L., Dossett, M. L., Greenberg, P. D., Allen, P. M. & Kranz, D. M. (2006). Development of a novel strategy Greenberg, P. D., Allen, P. M. & Kranz, D. M. (2006). Development of a novel strategy for engineering high-affinity proteins by yeast display. Protein Eng Des Sel 19, 255-64. for engineering high-affinity proteins by yeast display. Protein Eng Des Sel 19, 255-64. 165. Wernerus, H., Lehtio, J., Teeri, T., Nygren, P. A. & Stahl, S. (2001). Generation of 165. Wernerus, H., Lehtio, J., Teeri, T., Nygren, P. A. & Stahl, S. (2001). Generation of metal-binding staphylococci through surface display of combinatorially engineered metal-binding staphylococci through surface display of combinatorially engineered cellulose-binding domains. Appl Environ Microbiol 67, 4678-84. cellulose-binding domains. Appl Environ Microbiol 67, 4678-84. 166. Samuelson, P., Hansson, M., Ahlborg, N., Andréoni, C., Götz, F., Bächi, T., Nguyen, T. 166. Samuelson, P., Hansson, M., Ahlborg, N., Andréoni, C., Götz, F., Bächi, T., Nguyen, T. N., Binz, H., Uhlén, M. & Ståhl, S. (1995). Cell surface display of recombinant proteins N., Binz, H., Uhlén, M. & Ståhl, S. (1995). Cell surface display of recombinant proteins on Staphylococcus carnosus. J Bacteriol 177, 1470-6. on Staphylococcus carnosus. J Bacteriol 177, 1470-6. 167. Lofblom, J., Kronqvist, N., Uhlen, M., Stahl, S. & Wernerus, H. (2007). Optimization 167. Lofblom, J., Kronqvist, N., Uhlen, M., Stahl, S. & Wernerus, H. (2007). Optimization of electroporation-mediated transformation: Staphylococcus carnosus as model of electroporation-mediated transformation: Staphylococcus carnosus as model organism. J Appl Microbiol 102, 736-47. organism. J Appl Microbiol 102, 736-47. 168. Kronqvist, N., Lofblom, J., Jonsson, A., Wernerus, H. & Stahl, S. (2008). A novel 168. Kronqvist, N., Lofblom, J., Jonsson, A., Wernerus, H. & Stahl, S. (2008). A novel affinity protein selection system based on staphylococcal cell surface display and flow affinity protein selection system based on staphylococcal cell surface display and flow cytometry. Protein Eng Des Sel 21, 247-55. cytometry. Protein Eng Des Sel 21, 247-55. 169. Rockberg, J., Lofblom, J., Hjelm, B., Uhlen, M. & Stahl, S. (2008). Epitope mapping of 169. Rockberg, J., Lofblom, J., Hjelm, B., Uhlen, M. & Stahl, S. (2008). Epitope mapping of antibodies using bacterial surface display. Nat Methods 5, 1039-45. antibodies using bacterial surface display. Nat Methods 5, 1039-45. 170. Low, N. M., Holliger, P. H. & Winter, G. (1996). Mimicking somatic hypermutation: 170. Low, N. M., Holliger, P. H. & Winter, G. (1996). Mimicking somatic hypermutation: affinity maturation of antibodies displayed on bacteriophage using a bacterial mutator affinity maturation of antibodies displayed on bacteriophage using a bacterial mutator strain. J Mol Biol 260, 359-68. strain. J Mol Biol 260, 359-68. 171. Makeyev, E. V., Kolb, V. A. & Spirin, A. S. (1996). Enzymatic activity of the 171. Makeyev, E. V., Kolb, V. A. & Spirin, A. S. (1996). Enzymatic activity of the ribosome-bound nascent polypeptide. FEBS Lett 378, 166-70. ribosome-bound nascent polypeptide. FEBS Lett 378, 166-70. 172. Schaffitzel, C., Hanes, J., Jermutus, L. & Pluckthun, A. (1999). Ribosome display: an in 172. Schaffitzel, C., Hanes, J., Jermutus, L. & Pluckthun, A. (1999). Ribosome display: an in vitro method for selection and evolution of antibodies from libraries. J Immunol vitro method for selection and evolution of antibodies from libraries. J Immunol Methods 231, 119-35. Methods 231, 119-35. 173. He, M. & Taussig, M. J. (1997). Antibody-ribosome-mRNA (ARM) complexes as 173. He, M. & Taussig, M. J. (1997). Antibody-ribosome-mRNA (ARM) complexes as efficient selection particles for in vitro display and evolution of antibody combining efficient selection particles for in vitro display and evolution of antibody combining sites. Nucleic Acids Res 25, 5132-4. sites. Nucleic Acids Res 25, 5132-4. Sebastian Grimm 91 Sebastian Grimm 91

174. Osada, E., Shimizu, Y., Akbar, B. K., Kanamori, T. & Ueda, T. (2009). Epitope 174. Osada, E., Shimizu, Y., Akbar, B. K., Kanamori, T. & Ueda, T. (2009). Epitope mapping using ribosome display in a reconstituted cell-free protein synthesis system. J mapping using ribosome display in a reconstituted cell-free protein synthesis system. J Biochem 145, 693-700. Biochem 145, 693-700. 175. Nakatogawa, H. & Ito, K. (2002). The ribosomal exit tunnel functions as a 175. Nakatogawa, H. & Ito, K. (2002). The ribosomal exit tunnel functions as a discriminating gate. Cell 108, 629-36. discriminating gate. Cell 108, 629-36. 176. Evans, M. S., Ugrinov, K. G., Frese, M. A. & Clark, P. L. (2005). Homogeneous stalled 176. Evans, M. S., Ugrinov, K. G., Frese, M. A. & Clark, P. L. (2005). Homogeneous stalled ribosome nascent chain complexes produced in vivo or in vitro. Nat Methods 2, 757-62. ribosome nascent chain complexes produced in vivo or in vitro. Nat Methods 2, 757-62. 177. Ohashi, H., Shimizu, Y., Ying, B. W. & Ueda, T. (2007). Efficient protein selection 177. Ohashi, H., Shimizu, Y., Ying, B. W. & Ueda, T. (2007). Efficient protein selection based on ribosome display system with purified components. Biochem Biophys Res based on ribosome display system with purified components. Biochem Biophys Res Commun 352, 270-6. Commun 352, 270-6. 178. Matsuura, T., Yanagida, H., Ushioda, J., Urabe, I. & Yomo, T. (2007). Nascent chain, 178. Matsuura, T., Yanagida, H., Ushioda, J., Urabe, I. & Yomo, T. (2007). Nascent chain, mRNA, and ribosome complexes generated by a pure translation system. Biochem mRNA, and ribosome complexes generated by a pure translation system. Biochem Biophys Res Commun 352, 372-7. Biophys Res Commun 352, 372-7. 179. Nirenberg, M. W. & Matthaei, J. H. (1961). The dependence of cell-free protein 179. Nirenberg, M. W. & Matthaei, J. H. (1961). The dependence of cell-free protein synthesis in E. coli upon naturally occurring or synthetic polyribonucleotides. Proc Natl synthesis in E. coli upon naturally occurring or synthetic polyribonucleotides. Proc Natl Acad Sci USA 47, 1588-602. Acad Sci USA 47, 1588-602. 180. He, M. & Taussig, M. J. (2007). Eukaryotic ribosome display with in situ DNA 180. He, M. & Taussig, M. J. (2007). Eukaryotic ribosome display with in situ DNA recovery. Nat Methods 4, 281-8. recovery. Nat Methods 4, 281-8. 181. Hanes, J., Jermutus, L., Schaffitzel, C. & Pluckthun, A. (1999). Comparison of 181. Hanes, J., Jermutus, L., Schaffitzel, C. & Pluckthun, A. (1999). Comparison of Escherichia coli and rabbit reticulocyte ribosome display systems. FEBS Lett 450, 105- Escherichia coli and rabbit reticulocyte ribosome display systems. FEBS Lett 450, 105- 10. 10. 182. Shimizu, Y., Inoue, A., Tomari, Y., Suzuki, T., Yokogawa, T., Nishikawa, K. & Ueda, 182. Shimizu, Y., Inoue, A., Tomari, Y., Suzuki, T., Yokogawa, T., Nishikawa, K. & Ueda, T. (2001). Cell-free translation reconstituted with purified components. Nat Biotechnol T. (2001). Cell-free translation reconstituted with purified components. Nat Biotechnol 19, 751-5. 19, 751-5. 183. Villemagne, D., Jackson, R. & Douthwaite, J. A. (2006). Highly efficient ribosome 183. Villemagne, D., Jackson, R. & Douthwaite, J. A. (2006). Highly efficient ribosome display selection by use of purified components for in vitro translation. J Immunol display selection by use of purified components for in vitro translation. J Immunol Methods 313, 140-8. Methods 313, 140-8. 184. Zahnd, C., Amstutz, P. & Pluckthun, A. (2007). Ribosome display: selecting and 184. Zahnd, C., Amstutz, P. & Pluckthun, A. (2007). Ribosome display: selecting and evolving proteins in vitro that specifically bind to a target. Nat Methods 4, 269-79. evolving proteins in vitro that specifically bind to a target. Nat Methods 4, 269-79. 185. Chodorge, M., Fourage, L., Ravot, G., Jermutus, L. & Minter, R. (2008). In vitro DNA 185. Chodorge, M., Fourage, L., Ravot, G., Jermutus, L. & Minter, R. (2008). In vitro DNA recombination by L-Shuffling during ribosome display affinity maturation of an anti- recombination by L-Shuffling during ribosome display affinity maturation of an anti- Fas antibody increases the population of improved variants. Protein Eng Des Sel 21, Fas antibody increases the population of improved variants. Protein Eng Des Sel 21, 343-51. 343-51. 186. Shen, X., Valencia, C. A., Szostak, J. W., Szostak, J., Dong, B. & Liu, R. (2005). 186. Shen, X., Valencia, C. A., Szostak, J. W., Szostak, J., Dong, B. & Liu, R. (2005). Scanning the human proteome for calmodulin-binding proteins. Proc Natl Acad Sci Scanning the human proteome for calmodulin-binding proteins. Proc Natl Acad Sci USA 102, 5969-74. USA 102, 5969-74. 187. Odegrip, R., Coomber, D., Eldridge, B., Hederer, R., Kuhlman, P. A., Ullman, C., 187. Odegrip, R., Coomber, D., Eldridge, B., Hederer, R., Kuhlman, P. A., Ullman, C., FitzGerald, K. & McGregor, D. (2004). CIS display: In vitro selection of peptides from FitzGerald, K. & McGregor, D. (2004). CIS display: In vitro selection of peptides from libraries of protein-DNA complexes. Proc Natl Acad Sci USA 101, 2806-10. libraries of protein-DNA complexes. Proc Natl Acad Sci USA 101, 2806-10. 188. Schatz, P. J. (1993). Use of peptide libraries to map the substrate specificity of a 188. Schatz, P. J. (1993). Use of peptide libraries to map the substrate specificity of a peptide-modifying enzyme: a 13 residue consensus peptide specifies biotinylation in peptide-modifying enzyme: a 13 residue consensus peptide specifies biotinylation in Escherichia coli. Biotechnology 11, 1138-43. Escherichia coli. Biotechnology 11, 1138-43. 189. Speight, R. E., Hart, D. J., Sutherland, J. D. & Blackburn, J. M. (2001). A new plasmid 189. Speight, R. E., Hart, D. J., Sutherland, J. D. & Blackburn, J. M. (2001). A new plasmid display technology for the in vitro selection of functional phenotype-genotype linked display technology for the in vitro selection of functional phenotype-genotype linked proteins. Chem Biol 8, 951-65. proteins. Chem Biol 8, 951-65. 190. Reiersen, H., Lobersli, I., Loset, G. A., Hvattum, E., Simonsen, B., Stacy, J. E., 190. Reiersen, H., Lobersli, I., Loset, G. A., Hvattum, E., Simonsen, B., Stacy, J. E., McGregor, D., Fitzgerald, K., Welschof, M., Brekke, O. H. & Marvik, O. J. (2005). McGregor, D., Fitzgerald, K., Welschof, M., Brekke, O. H. & Marvik, O. J. (2005). Covalent antibody display--an in vitro antibody-DNA library selection system. Nucleic Covalent antibody display--an in vitro antibody-DNA library selection system. Nucleic Acids Res 33, e10. Acids Res 33, e10. 92 Ribosome display for selection and evolution of affibody molecules 92 Ribosome display for selection and evolution of affibody molecules

191. Tawfik, D. S. & Griffiths, A. D. (1998). Man-made cell-like compartments for 191. Tawfik, D. S. & Griffiths, A. D. (1998). Man-made cell-like compartments for molecular evolution. Nat Biotechnol 16, 652-6. molecular evolution. Nat Biotechnol 16, 652-6. 192. Lee, Y.-F., Tawfik, D. S. & Griffiths, A. D. (2002). Investigating the target recognition 192. Lee, Y.-F., Tawfik, D. S. & Griffiths, A. D. (2002). Investigating the target recognition of DNA cytosine-5 methyltransferase HhaI by library selection using in vitro of DNA cytosine-5 methyltransferase HhaI by library selection using in vitro compartmentalisation. Nucleic Acids Res 30, 4937-44. compartmentalisation. Nucleic Acids Res 30, 4937-44. 193. Cohen, H. M., Tawfik, D. S. & Griffiths, A. D. (2004). Altering the sequence specificity 193. Cohen, H. M., Tawfik, D. S. & Griffiths, A. D. (2004). Altering the sequence specificity of HaeIII methyltransferase by directed evolution using in vitro compartmentalization. of HaeIII methyltransferase by directed evolution using in vitro compartmentalization. Protein Eng Des Sel 17, 3-11. Protein Eng Des Sel 17, 3-11. 194. Bertschinger, J. & Neri, D. (2004). Covalent DNA display as a novel tool for directed 194. Bertschinger, J. & Neri, D. (2004). Covalent DNA display as a novel tool for directed evolution of proteins in vitro. Protein Eng Des Sel 17, 699-707. evolution of proteins in vitro. Protein Eng Des Sel 17, 699-707. 195. Doi, N. & Yanagawa, H. (1999). STABLE: protein-DNA fusion system for screening of 195. Doi, N. & Yanagawa, H. (1999). STABLE: protein-DNA fusion system for screening of combinatorial protein libraries in vitro. FEBS Lett 457, 227-30. combinatorial protein libraries in vitro. FEBS Lett 457, 227-30. 196. Griffiths, A. D. & Tawfik, D. S. (2006). Miniaturising the laboratory in emulsion 196. Griffiths, A. D. & Tawfik, D. S. (2006). Miniaturising the laboratory in emulsion droplets. Trends Biotechnol 24, 395-402. droplets. Trends Biotechnol 24, 395-402. 197. Bernath, K., Hai, M., Mastrobattista, E., Griffiths, A. D., Magdassi, S. & Tawfik, D. S. 197. Bernath, K., Hai, M., Mastrobattista, E., Griffiths, A. D., Magdassi, S. & Tawfik, D. S. (2004). In vitro compartmentalization by double emulsions: sorting and gene (2004). In vitro compartmentalization by double emulsions: sorting and gene enrichment by fluorescence activated cell sorting. Anal Biochem 325, 151-7. enrichment by fluorescence activated cell sorting. Anal Biochem 325, 151-7. 198. Mastrobattista, E., Taly, V., Chanudet, E., Treacy, P., Kelly, B. T. & Griffiths, A. D. 198. Mastrobattista, E., Taly, V., Chanudet, E., Treacy, P., Kelly, B. T. & Griffiths, A. D. (2005). High-throughput screening of enzyme libraries: in vitro evolution of a beta- (2005). High-throughput screening of enzyme libraries: in vitro evolution of a beta- galactosidase by fluorescence-activated sorting of double emulsions. Chem Biol 12, galactosidase by fluorescence-activated sorting of double emulsions. Chem Biol 12, 1291-300. 1291-300. 199. Jackrel, M. E., Cortajarena, A. L., Liu, T. Y. & Regan, L. (2009). Screening libraries to 199. Jackrel, M. E., Cortajarena, A. L., Liu, T. Y. & Regan, L. (2009). Screening libraries to identify proteins with desired binding activities using a split-GFP reassembly assay. identify proteins with desired binding activities using a split-GFP reassembly assay. ACS Chem Biol 5, 553-62. ACS Chem Biol 5, 553-62. 200. Lofdahl, P. A., Nord, O., Janzon, L. & Nygren, P. A. (2009). Selection of TNF-alpha 200. Lofdahl, P. A., Nord, O., Janzon, L. & Nygren, P. A. (2009). Selection of TNF-alpha binding affibody molecules using a beta-lactamase protein fragment complementation binding affibody molecules using a beta-lactamase protein fragment complementation assay. N Biotechnol 26, 251-9. assay. N Biotechnol 26, 251-9. 201. Mossner, E., Koch, H. & Pluckthun, A. (2001). Fast selection of antibodies without 201. Mossner, E., Koch, H. & Pluckthun, A. (2001). Fast selection of antibodies without antigen purification: adaptation of the protein fragment complementation assay to select antigen purification: adaptation of the protein fragment complementation assay to select antigen-antibody pairs. J Mol Biol 308, 115-22. antigen-antibody pairs. J Mol Biol 308, 115-22. 202. Kronqvist, N., Malm, M., Gostring, L., Gunneriusson, E., Nilsson, M., Hoiden 202. Kronqvist, N., Malm, M., Gostring, L., Gunneriusson, E., Nilsson, M., Hoiden Guthenberg, I., Gedda, L., Frejd, F. Y., Stahl, S. & Lofblom, J. (2010). Combining Guthenberg, I., Gedda, L., Frejd, F. Y., Stahl, S. & Lofblom, J. (2010). Combining phage and staphylococcal surface display for generation of ErbB3-specific Affibody phage and staphylococcal surface display for generation of ErbB3-specific Affibody molecules. Protein Eng Des Sel. molecules. Protein Eng Des Sel. 203. Ehrlich, P. (1901). Die Seitenkettentheorie und ihre Gegner. Muenchner medizinische 203. Ehrlich, P. (1901). Die Seitenkettentheorie und ihre Gegner. Muenchner medizinische Wochenschrift, 2123-2124. Wochenschrift, 2123-2124. 204. Winau, F., Westphal, O. & Winau, R. (2004). Paul Ehrlich--in search of the magic 204. Winau, F., Westphal, O. & Winau, R. (2004). Paul Ehrlich--in search of the magic bullet. Microbes Infect 6, 786-9. bullet. Microbes Infect 6, 786-9. 205. Nord, K., Nilsson, J., Nilsson, B., Uhlen, M. & Nygren, P. A. (1995). A combinatorial 205. Nord, K., Nilsson, J., Nilsson, B., Uhlen, M. & Nygren, P. A. (1995). A combinatorial library of an alpha-helical bacterial receptor domain. Protein Eng Des Sel 8, 601-8. library of an alpha-helical bacterial receptor domain. Protein Eng Des Sel 8, 601-8. 206. Gebauer, M. & Skerra, A. (2009). Engineered protein scaffolds as next-generation 206. Gebauer, M. & Skerra, A. (2009). Engineered protein scaffolds as next-generation antibody therapeutics. Curr Opin Chem Biol 13, 245-55. antibody therapeutics. Curr Opin Chem Biol 13, 245-55. 207. Foote, J. & Winter, G. (1992). Antibody framework residues affecting the conformation 207. Foote, J. & Winter, G. (1992). Antibody framework residues affecting the conformation of the hypervariable loops. J Mol Biol 224, 487-99. of the hypervariable loops. J Mol Biol 224, 487-99. 208. Arnold, J. N., Royle, L., Dwek, R. A., Rudd, P. M. & Sim, R. B. (2005). Human 208. Arnold, J. N., Royle, L., Dwek, R. A., Rudd, P. M. & Sim, R. B. (2005). Human immunoglobulin glycosylation and the lectin pathway of complement activation. Adv immunoglobulin glycosylation and the lectin pathway of complement activation. Adv Exp Med Biol 564, 27-43. Exp Med Biol 564, 27-43. 209. Walsh, G. (2007). Pharmaceutical Biotechnology Concepts and Applications, Wiley. 209. Walsh, G. (2007). Pharmaceutical Biotechnology Concepts and Applications, Wiley. 210. Kohler, G. & Milstein, C. (1975). Continuous cultures of fused cells secreting antibody 210. Kohler, G. & Milstein, C. (1975). Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256, 495-7. of predefined specificity. Nature 256, 495-7. Sebastian Grimm 93 Sebastian Grimm 93

211. Rathanaswami, P., Roalstad, S., Roskos, L., Su, Q. J., Lackie, S. & Babcook, J. (2005). 211. Rathanaswami, P., Roalstad, S., Roskos, L., Su, Q. J., Lackie, S. & Babcook, J. (2005). Demonstration of an in vivo generated sub-picomolar affinity fully human monoclonal Demonstration of an in vivo generated sub-picomolar affinity fully human monoclonal antibody to interleukin-8. Biochem Biophys Res Commun 334, 1004-13. antibody to interleukin-8. Biochem Biophys Res Commun 334, 1004-13. 212. Lonberg, N. (2005). Human antibodies from transgenic animals. Nat Biotechnol 23, 212. Lonberg, N. (2005). Human antibodies from transgenic animals. Nat Biotechnol 23, 1117-25. 1117-25. 213. Walsh, G. (2010). Biopharmaceutical benchmarks 2010. Nat Biotechnol 28, 917-24. 213. Walsh, G. (2010). Biopharmaceutical benchmarks 2010. Nat Biotechnol 28, 917-24. 214. Mori, K., Iida, S., Yamane-Ohnuki, N., Kanda, Y., Kuni-Kamochi, R., Nakano, R., 214. Mori, K., Iida, S., Yamane-Ohnuki, N., Kanda, Y., Kuni-Kamochi, R., Nakano, R., Imai-Nishiya, H., Okazaki, A., Shinkawa, T., Natsume, A., Niwa, R., Shitara, K. & Imai-Nishiya, H., Okazaki, A., Shinkawa, T., Natsume, A., Niwa, R., Shitara, K. & Satoh, M. (2007). Non-fucosylated therapeutic antibodies: the next generation of Satoh, M. (2007). Non-fucosylated therapeutic antibodies: the next generation of therapeutic antibodies. Cytotechnology 55, 109-14. therapeutic antibodies. Cytotechnology 55, 109-14. 215. Jung, S. T., Reddy, S. T., Kang, T. H., Borrok, M. J., Sandlie, I., Tucker, P. W. & 215. Jung, S. T., Reddy, S. T., Kang, T. H., Borrok, M. J., Sandlie, I., Tucker, P. W. & Georgiou, G. (2010). Aglycosylated IgG variants expressed in bacteria that selectively Georgiou, G. (2010). Aglycosylated IgG variants expressed in bacteria that selectively bind FcgammaRI potentiate tumor cell killing by monocyte-dendritic cells. Proc Natl bind FcgammaRI potentiate tumor cell killing by monocyte-dendritic cells. Proc Natl Acad Sci U S A 107, 604-9. Acad Sci U S A 107, 604-9. 216. Dall'Acqua, W. F., Kiener, P. A. & Wu, H. (2006). Properties of human IgG1s 216. Dall'Acqua, W. F., Kiener, P. A. & Wu, H. (2006). Properties of human IgG1s engineered for enhanced binding to the neonatal Fc receptor (FcRn). J Biol Chem 281, engineered for enhanced binding to the neonatal Fc receptor (FcRn). J Biol Chem 281, 23514-24. 23514-24. 217. Idusogie, E. E., Wong, P. Y., Presta, L. G., Gazzano-Santoro, H., Totpal, K., Ultsch, M. 217. Idusogie, E. E., Wong, P. Y., Presta, L. G., Gazzano-Santoro, H., Totpal, K., Ultsch, M. & Mulkerrin, M. G. (2001). Engineered antibodies with increased activity to recruit & Mulkerrin, M. G. (2001). Engineered antibodies with increased activity to recruit complement. J Immunol 166, 2571-5. complement. J Immunol 166, 2571-5. 218. Lindhofer, H., Mocikat, R., Steipe, B. & Thierfelder, S. (1995). Preferential species- 218. Lindhofer, H., Mocikat, R., Steipe, B. & Thierfelder, S. (1995). Preferential species- restricted heavy/light chain pairing in rat/mouse quadromas. Implications for a single- restricted heavy/light chain pairing in rat/mouse quadromas. Implications for a single- step purification of bispecific antibodies. J Immunol 155, 219-25. step purification of bispecific antibodies. J Immunol 155, 219-25. 219. Carter, P. (2001). Bispecific human IgG by design. J Immunol Methods 248, 7-15. 219. Carter, P. (2001). Bispecific human IgG by design. J Immunol Methods 248, 7-15. 220. Wu, C., Ying, H., Grinnell, C., Bryant, S., Miller, R., Clabbers, A., Bose, S., McCarthy, 220. Wu, C., Ying, H., Grinnell, C., Bryant, S., Miller, R., Clabbers, A., Bose, S., McCarthy, D., Zhu, R. R., Santora, L., Davis-Taber, R., Kunes, Y., Fung, E., Schwartz, A., D., Zhu, R. R., Santora, L., Davis-Taber, R., Kunes, Y., Fung, E., Schwartz, A., Sakorafas, P., Gu, J., Tarcsa, E., Murtaza, A. & Ghayur, T. (2007). Simultaneous Sakorafas, P., Gu, J., Tarcsa, E., Murtaza, A. & Ghayur, T. (2007). Simultaneous targeting of multiple disease mediators by a dual-variable-domain immunoglobulin. Nat targeting of multiple disease mediators by a dual-variable-domain immunoglobulin. Nat Biotechnol 25, 1290-7. Biotechnol 25, 1290-7. 221. Holliger, P. & Hudson, P. J. (2005). Engineered antibody fragments and the rise of 221. Holliger, P. & Hudson, P. J. (2005). Engineered antibody fragments and the rise of single domains. Nat Biotechnol 23, 1126-36. single domains. Nat Biotechnol 23, 1126-36. 222. Ghose, T., Cerini, M., Carter, M. & Nairn, R. C. (1967). Immunoradioactive agent 222. Ghose, T., Cerini, M., Carter, M. & Nairn, R. C. (1967). Immunoradioactive agent against cancer. Br Med J 1, 90-3. against cancer. Br Med J 1, 90-3. 223. Ghose, T. & Nigam, S. P. (1972). Antibody as carrier of chlorambucil. Cancer 29, 223. Ghose, T. & Nigam, S. P. (1972). Antibody as carrier of chlorambucil. Cancer 29, 1398-400. 1398-400. 224. Simmons, L. C., Reilly, D., Klimowski, L., Raju, T. S., Meng, G., Sims, P., Hong, K., 224. Simmons, L. C., Reilly, D., Klimowski, L., Raju, T. S., Meng, G., Sims, P., Hong, K., Shields, R. L., Damico, L. A., Rancatore, P. & Yansura, D. G. (2002). Expression of Shields, R. L., Damico, L. A., Rancatore, P. & Yansura, D. G. (2002). Expression of full-length immunoglobulins in Escherichia coli: rapid and efficient production of full-length immunoglobulins in Escherichia coli: rapid and efficient production of aglycosylated antibodies. J Immunol Methods 263, 133-47. aglycosylated antibodies. J Immunol Methods 263, 133-47. 225. Bird, R. E., Hardman, K. D., Jacobson, J. W., Johnson, S., Kaufman, B. M., Lee, S. M., 225. Bird, R. E., Hardman, K. D., Jacobson, J. W., Johnson, S., Kaufman, B. M., Lee, S. M., Lee, T., Pope, S. H., Riordan, G. S. & Whitlow, M. (1988). Single-chain antigen- Lee, T., Pope, S. H., Riordan, G. S. & Whitlow, M. (1988). Single-chain antigen- binding proteins. Science 242, 423-6. binding proteins. Science 242, 423-6. 226. Better, M., Chang, C. P., Robinson, R. R. & Horwitz, A. H. (1988). Escherichia coli 226. Better, M., Chang, C. P., Robinson, R. R. & Horwitz, A. H. (1988). Escherichia coli secretion of an active chimeric antibody fragment. Science 240, 1041-3. secretion of an active chimeric antibody fragment. Science 240, 1041-3. 227. Hust, M., Meyer, T., Voedisch, B., Rulker, T., Thie, H., El-Ghezal, A., Kirsch, M. I., 227. Hust, M., Meyer, T., Voedisch, B., Rulker, T., Thie, H., El-Ghezal, A., Kirsch, M. I., Schutte, M., Helmsing, S., Meier, D., Schirrmann, T. & Dubel, S. A human scFv Schutte, M., Helmsing, S., Meier, D., Schirrmann, T. & Dubel, S. A human scFv antibody generation pipeline for proteome research. J Biotechnol. antibody generation pipeline for proteome research. J Biotechnol. 228. Pershad, K., Pavlovic, J. D., Graslund, S., Nilsson, P., Colwill, K., Karatt-Vellatt, A., 228. Pershad, K., Pavlovic, J. D., Graslund, S., Nilsson, P., Colwill, K., Karatt-Vellatt, A., Schofield, D. J., Dyson, M. R., Pawson, T., Kay, B. K. & McCafferty, J. Generating a Schofield, D. J., Dyson, M. R., Pawson, T., Kay, B. K. & McCafferty, J. Generating a 94 Ribosome display for selection and evolution of affibody molecules 94 Ribosome display for selection and evolution of affibody molecules

panel of highly specific antibodies to 20 human SH2 domains by phage display. Protein panel of highly specific antibodies to 20 human SH2 domains by phage display. Protein Eng Des Sel 23, 279-88. Eng Des Sel 23, 279-88. 229. Rothe, C., Urlinger, S., Lohning, C., Prassler, J., Stark, Y., Jager, U., Hubner, B., 229. Rothe, C., Urlinger, S., Lohning, C., Prassler, J., Stark, Y., Jager, U., Hubner, B., Bardroff, M., Pradel, I., Boss, M., Bittlingmaier, R., Bataa, T., Frisch, C., Brocks, B., Bardroff, M., Pradel, I., Boss, M., Bittlingmaier, R., Bataa, T., Frisch, C., Brocks, B., Honegger, A. & Urban, M. (2008). The human combinatorial antibody library HuCAL Honegger, A. & Urban, M. (2008). The human combinatorial antibody library HuCAL GOLD combines diversification of all six CDRs according to the natural immune GOLD combines diversification of all six CDRs according to the natural immune system with a novel display method for efficient selection of high-affinity antibodies. J system with a novel display method for efficient selection of high-affinity antibodies. J Mol Biol 376, 1182-200. Mol Biol 376, 1182-200. 230. Muyldermans, S. (2006). Single Domain Antibodies Nanobodies, Potent Mini- 230. Muyldermans, S. (2006). Single Domain Antibodies Nanobodies, Potent Mini- antibodies for Various Applications. BIOforum Europe 9/2006. antibodies for Various Applications. BIOforum Europe 9/2006. 231. Muyldermans, S., Baral, T. N., Retamozzo, V. C., De Baetselier, P., De Genst, E., 231. Muyldermans, S., Baral, T. N., Retamozzo, V. C., De Baetselier, P., De Genst, E., Kinne, J., Leonhardt, H., Magez, S., Nguyen, V. K., Revets, H., Rothbauer, U., Kinne, J., Leonhardt, H., Magez, S., Nguyen, V. K., Revets, H., Rothbauer, U., Stijlemans, B., Tillib, S., Wernery, U., Wyns, L., Hassanzadeh-Ghassabeh, G. & Stijlemans, B., Tillib, S., Wernery, U., Wyns, L., Hassanzadeh-Ghassabeh, G. & Saerens, D. (2009). Camelid immunoglobulins and nanobody technology. Vet Immunol Saerens, D. (2009). Camelid immunoglobulins and nanobody technology. Vet Immunol Immunopathol 128, 178-83. Immunopathol 128, 178-83. 232. Bond, C. J., Marsters, J. C. & Sidhu, S. S. (2003). Contributions of CDR3 to V H H 232. Bond, C. J., Marsters, J. C. & Sidhu, S. S. (2003). Contributions of CDR3 to V H H domain stability and the design of monobody scaffolds for naive antibody libraries. J domain stability and the design of monobody scaffolds for naive antibody libraries. J Mol Biol 332, 643-55. Mol Biol 332, 643-55. 233. De Genst, E., Silence, K., Decanniere, K., Conrath, K., Loris, R., Kinne, J., 233. De Genst, E., Silence, K., Decanniere, K., Conrath, K., Loris, R., Kinne, J., Muyldermans, S. & Wyns, L. (2006). Molecular basis for the preferential cleft Muyldermans, S. & Wyns, L. (2006). Molecular basis for the preferential cleft recognition by dromedary heavy-chain antibodies. Proc Natl Acad Sci USA 103, 4586- recognition by dromedary heavy-chain antibodies. Proc Natl Acad Sci USA 103, 4586- 91. 91. 234. Holt, L. J., Herring, C., Jespers, L. S., Woolven, B. P. & Tomlinson, I. M. (2003). 234. Holt, L. J., Herring, C., Jespers, L. S., Woolven, B. P. & Tomlinson, I. M. (2003). Domain antibodies: proteins for therapy. Trends Biotechnol 21, 484-90. Domain antibodies: proteins for therapy. Trends Biotechnol 21, 484-90. 235. Smith, G. P. & Petrenko, V. A. (1997). Phage Display. Chem Rev 97, 391-410. 235. Smith, G. P. & Petrenko, V. A. (1997). Phage Display. Chem Rev 97, 391-410. 236. Giebel, L. B., Cass, R. T., Milligan, D. L., Young, D. C., Arze, R. & Johnson, C. R. 236. Giebel, L. B., Cass, R. T., Milligan, D. L., Young, D. C., Arze, R. & Johnson, C. R. (1995). Screening of cyclic peptide phage libraries identifies ligands that bind (1995). Screening of cyclic peptide phage libraries identifies ligands that bind streptavidin with high affinities. Biochemistry 34, 15430-5. streptavidin with high affinities. Biochemistry 34, 15430-5. 237. Nygren, P.-A. & Skerra, A. (2004). Binding proteins from alternative scaffolds. J 237. Nygren, P.-A. & Skerra, A. (2004). Binding proteins from alternative scaffolds. J Immunol Methods 290, 3-28. Immunol Methods 290, 3-28. 238. Flower, D. R., North, A. C. & Attwood, T. K. (1993). Structure and sequence 238. Flower, D. R., North, A. C. & Attwood, T. K. (1993). Structure and sequence relationships in the lipocalins and related proteins. Protein Sci 2, 753-61. relationships in the lipocalins and related proteins. Protein Sci 2, 753-61. 239. Frand, A. R., Cuozzo, J. W. & Kaiser, C. A. (2000). Pathways for protein disulphide 239. Frand, A. R., Cuozzo, J. W. & Kaiser, C. A. (2000). Pathways for protein disulphide bond formation. Trends Cell Biol 10, 203-10. bond formation. Trends Cell Biol 10, 203-10. 240. Skerra, A. (2008). Alternative binding proteins: anticalins - harnessing the structural 240. Skerra, A. (2008). Alternative binding proteins: anticalins - harnessing the structural plasticity of the lipocalin ligand pocket to engineer novel binding activities. FEBS J plasticity of the lipocalin ligand pocket to engineer novel binding activities. FEBS J 275, 2677-83. 275, 2677-83. 241. Schlapschy, M., Grimm, S. & Skerra, A. (2006). A system for concomitant 241. Schlapschy, M., Grimm, S. & Skerra, A. (2006). A system for concomitant overexpression of four periplasmic folding catalysts to improve secretory protein overexpression of four periplasmic folding catalysts to improve secretory protein production in Escherichia coli. Protein Eng Des Sel 19, 385-90. production in Escherichia coli. Protein Eng Des Sel 19, 385-90. 242. Foster, T. J. (2005). Immune evasion by staphylococci. Nat Rev Microbiol 3, 948-58. 242. Foster, T. J. (2005). Immune evasion by staphylococci. Nat Rev Microbiol 3, 948-58. 243. Nilsson, B., Moks, T., Jansson, B., Abrahmsen, L., Elmblad, A., Holmgren, E., 243. Nilsson, B., Moks, T., Jansson, B., Abrahmsen, L., Elmblad, A., Holmgren, E., Henrichson, C., Jones, T. A. & Uhlen, M. (1987). A synthetic IgG-binding domain Henrichson, C., Jones, T. A. & Uhlen, M. (1987). A synthetic IgG-binding domain based on staphylococcal protein A. Protein Eng Des Sel 1, 107-13. based on staphylococcal protein A. Protein Eng Des Sel 1, 107-13. 244. Nygren, P.-Å. (2008). Alternative binding proteins: Affibody binding proteins 244. Nygren, P.-Å. (2008). Alternative binding proteins: Affibody binding proteins developed from a small three-helix bundle scaffold. FEBS J 275, 2668-2676. developed from a small three-helix bundle scaffold. FEBS J 275, 2668-2676. 245. Arora, P., Oas, T. G. & Myers, J. K. (2004). Fast and faster: a designed variant of the B- 245. Arora, P., Oas, T. G. & Myers, J. K. (2004). Fast and faster: a designed variant of the B- domain of protein A folds in 3 microsec. Protein Sci 13, 847-53. domain of protein A folds in 3 microsec. Protein Sci 13, 847-53. Sebastian Grimm 95 Sebastian Grimm 95

246. Engfeldt, T., Renberg, B., Brumer, H., Nygren, P. A. & Karlstrom, A. E. (2005). 246. Engfeldt, T., Renberg, B., Brumer, H., Nygren, P. A. & Karlstrom, A. E. (2005). Chemical synthesis of triple-labelled three-helix bundle binding proteins for specific Chemical synthesis of triple-labelled three-helix bundle binding proteins for specific fluorescent detection of unlabelled protein. Chembiochem 6, 1043-50. fluorescent detection of unlabelled protein. Chembiochem 6, 1043-50. 247. Feldwisch, J., Tolmachev, V., Lendel, C., Herne, N., Sjoberg, A., Larsson, B., Rosik, 247. Feldwisch, J., Tolmachev, V., Lendel, C., Herne, N., Sjoberg, A., Larsson, B., Rosik, D., Lindqvist, E., Fant, G., Hoiden-Guthenberg, I., Galli, J., Jonasson, P. & Abrahmsen, D., Lindqvist, E., Fant, G., Hoiden-Guthenberg, I., Galli, J., Jonasson, P. & Abrahmsen, L. (2010). Design of an optimized scaffold for affibody molecules. J Mol Biol 398, 232- L. (2010). Design of an optimized scaffold for affibody molecules. J Mol Biol 398, 232- 47. 47. 248. Grimm, S., Yu, F. & Nygren, P.-Å. (2011). Ribosome Display Selection of a Murine 248. Grimm, S., Yu, F. & Nygren, P.-Å. (2011). Ribosome Display Selection of a Murine IgG(1) Fab Binding Affibody Molecule Allowing Species Selective Recovery Of IgG(1) Fab Binding Affibody Molecule Allowing Species Selective Recovery Of Monoclonal Antibodies. Mol biotechnol. Monoclonal Antibodies. Mol biotechnol. 249. Gunneriusson, E., Nord, K., Uhlen, M. & Nygren, P. (1999). Affinity maturation of a 249. Gunneriusson, E., Nord, K., Uhlen, M. & Nygren, P. (1999). Affinity maturation of a Taq DNA polymerase specific affibody by helix shuffling. Protein Eng Des Sel 12, Taq DNA polymerase specific affibody by helix shuffling. Protein Eng Des Sel 12, 873-8. 873-8. 250. Grönwall, C., Sjöberg, A., Ramström, M., Höidén-Guthenberg, I., Hober, S., Jonasson, 250. Grönwall, C., Sjöberg, A., Ramström, M., Höidén-Guthenberg, I., Hober, S., Jonasson, P. & Ståhl, S. (2007). Affibody-mediated transferrin depletion for proteomics P. & Ståhl, S. (2007). Affibody-mediated transferrin depletion for proteomics applications. Biotechnol J 2, 1389-98. applications. Biotechnol J 2, 1389-98. 251. Wikman, M., Steffen, A.-C., Gunneriusson, E., Tolmachev, V., Adams, G. P., Carlsson, 251. Wikman, M., Steffen, A.-C., Gunneriusson, E., Tolmachev, V., Adams, G. P., Carlsson, J. & Ståhl, S. (2004). Selection and characterization of HER2/neu-binding affibody J. & Ståhl, S. (2004). Selection and characterization of HER2/neu-binding affibody ligands. Protein Eng Des Sel 17, 455-62. ligands. Protein Eng Des Sel 17, 455-62. 252. Li, J., Lundberg, E., Vernet, E., Larsson, B., Hoiden-Guthenberg, I. & Graslund, T. 252. Li, J., Lundberg, E., Vernet, E., Larsson, B., Hoiden-Guthenberg, I. & Graslund, T. (2010). Selection of affibody molecules to the ligand-binding site of the insulin-like (2010). Selection of affibody molecules to the ligand-binding site of the insulin-like growth factor-1 receptor. Biotechnol Appl Biochem 55, 99-109. growth factor-1 receptor. Biotechnol Appl Biochem 55, 99-109. 253. Eklund, M., Axelsson, L., Uhlen, M. & Nygren, P. A. (2002). Anti-idiotypic protein 253. Eklund, M., Axelsson, L., Uhlen, M. & Nygren, P. A. (2002). Anti-idiotypic protein domains selected from protein A-based affibody libraries. Proteins 48, 454-62. domains selected from protein A-based affibody libraries. Proteins 48, 454-62. 254. Wållberg, H., Löfdahl, P.-K., Tschapalda, K., Uhlén, M., Tolmachev, V., Nygren, P.-K. 254. Wållberg, H., Löfdahl, P.-K., Tschapalda, K., Uhlén, M., Tolmachev, V., Nygren, P.-K. & Ståhl, S. (2010). Affinity recovery of eight HER2-binding affibody variants using an & Ståhl, S. (2010). Affinity recovery of eight HER2-binding affibody variants using an anti-idiotypic affibody molecule as capture ligand. Protein Expr Purif. anti-idiotypic affibody molecule as capture ligand. Protein Expr Purif. 255. Rönnmark, J., Grönlund, H., Uhlén, M. & Nygren, P.-A. (2002). Human 255. Rönnmark, J., Grönlund, H., Uhlén, M. & Nygren, P.-A. (2002). Human immunoglobulin A (IgA)-specific ligands from combinatorial engineering of protein A. immunoglobulin A (IgA)-specific ligands from combinatorial engineering of protein A. Eur J Biochem 269, 2647-55. Eur J Biochem 269, 2647-55. 256. Wikman, M., Rowcliffe, E., Friedman, M., Henning, P., Lindholm, L., Olofsson, S. & 256. Wikman, M., Rowcliffe, E., Friedman, M., Henning, P., Lindholm, L., Olofsson, S. & Stahl, S. (2006). Selection and characterization of an HIV-1 gp120-binding affibody Stahl, S. (2006). Selection and characterization of an HIV-1 gp120-binding affibody ligand. Biotechnol Appl Biochem 45, 93-105. ligand. Biotechnol Appl Biochem 45, 93-105. 257. Sandström, K., Xu, Z., Forsberg, G. & Nygren, P.-A. (2003). Inhibition of the CD28- 257. Sandström, K., Xu, Z., Forsberg, G. & Nygren, P.-A. (2003). Inhibition of the CD28- CD80 co-stimulation signal by a CD28-binding affibody ligand developed by CD80 co-stimulation signal by a CD28-binding affibody ligand developed by combinatorial protein engineering. Protein Eng Des Sel 16, 691-7. combinatorial protein engineering. Protein Eng Des Sel 16, 691-7. 258. Grönwall, C., Jonsson, A., Lindström, S., Gunneriusson, E., Ståhl, S. & Herne, N. 258. Grönwall, C., Jonsson, A., Lindström, S., Gunneriusson, E., Ståhl, S. & Herne, N. (2007). Selection and characterization of Affibody ligands binding to Alzheimer (2007). Selection and characterization of Affibody ligands binding to Alzheimer amyloid beta peptides. J Biotechnol 128, 162-83. amyloid beta peptides. J Biotechnol 128, 162-83. 259. Eigenbrot, C., Ultsch, M., Dubnovitsky, A., Abrahmsén, L. & Härd, T. (2010). 259. Eigenbrot, C., Ultsch, M., Dubnovitsky, A., Abrahmsén, L. & Härd, T. (2010). Structural basis for high-affinity HER2 receptor binding by an engineered protein. Proc Structural basis for high-affinity HER2 receptor binding by an engineered protein. Proc Natl Acad Sci USA 107, 15039-44. Natl Acad Sci USA 107, 15039-44. 260. Baum, R. P., Prasad, V., Muller, D., Schuchardt, C., Orlova, A., Wennborg, A., 260. Baum, R. P., Prasad, V., Muller, D., Schuchardt, C., Orlova, A., Wennborg, A., Tolmachev, V. & Feldwisch, J. (2010). Molecular imaging of HER2-expressing Tolmachev, V. & Feldwisch, J. (2010). Molecular imaging of HER2-expressing malignant tumors in breast cancer patients using synthetic 111In- or 68Ga-labeled malignant tumors in breast cancer patients using synthetic 111In- or 68Ga-labeled affibody molecules. J Nucl Med 51, 892-7. affibody molecules. J Nucl Med 51, 892-7. 261. Tran, T. A., Rosik, D., Abrahmsen, L., Sandstrom, M., Sjoberg, A., Wallberg, H., 261. Tran, T. A., Rosik, D., Abrahmsen, L., Sandstrom, M., Sjoberg, A., Wallberg, H., Ahlgren, S., Orlova, A. & Tolmachev, V. (2009). Design, synthesis and biological Ahlgren, S., Orlova, A. & Tolmachev, V. (2009). Design, synthesis and biological evaluation of a multifunctional HER2-specific Affibody molecule for molecular evaluation of a multifunctional HER2-specific Affibody molecule for molecular imaging. Eur J Nucl Med Mol Imaging 36, 1864-73. imaging. Eur J Nucl Med Mol Imaging 36, 1864-73. 96 Ribosome display for selection and evolution of affibody molecules 96 Ribosome display for selection and evolution of affibody molecules

262. Zielinski, R., Lyakhov, I., Jacobs, A., Chertov, O., Kramer-Marek, G., Francella, N., 262. Zielinski, R., Lyakhov, I., Jacobs, A., Chertov, O., Kramer-Marek, G., Francella, N., Stephen, A., Fisher, R., Blumenthal, R. & Capala, J. (2009). Affitoxin--a novel Stephen, A., Fisher, R., Blumenthal, R. & Capala, J. (2009). Affitoxin--a novel recombinant, HER2-specific, anticancer agent for targeted therapy of HER2-positive recombinant, HER2-specific, anticancer agent for targeted therapy of HER2-positive tumors. J Immunother 32, 817-25. tumors. J Immunother 32, 817-25. 263. Alexis, F., Basto, P., Levy-Nissenbaum, E., Radovic-Moreno, A. F., Zhang, L., Pridgen, 263. Alexis, F., Basto, P., Levy-Nissenbaum, E., Radovic-Moreno, A. F., Zhang, L., Pridgen, E., Wang, A. Z., Marein, S. L., Westerhof, K., Molnar, L. K. & Farokhzad, O. C. E., Wang, A. Z., Marein, S. L., Westerhof, K., Molnar, L. K. & Farokhzad, O. C. (2008). HER-2-targeted nanoparticle-affibody bioconjugates for cancer therapy. (2008). HER-2-targeted nanoparticle-affibody bioconjugates for cancer therapy. ChemMedChem 3, 1839-43. ChemMedChem 3, 1839-43. 264. Beuttler, J., Rothdiener, M., Muller, D., Frejd, F. Y. & Kontermann, R. E. (2009). 264. Beuttler, J., Rothdiener, M., Muller, D., Frejd, F. Y. & Kontermann, R. E. (2009). Targeting of epidermal growth factor receptor (EGFR)-expressing tumor cells with Targeting of epidermal growth factor receptor (EGFR)-expressing tumor cells with sterically stabilized affibody liposomes (SAL). Bioconjug Chem 20, 1201-8. sterically stabilized affibody liposomes (SAL). Bioconjug Chem 20, 1201-8. 265. Tolmachev, V., Orlova, A., Pehrson, R., Galli, J., Baastrup, B., Andersson, K., 265. Tolmachev, V., Orlova, A., Pehrson, R., Galli, J., Baastrup, B., Andersson, K., Sandstrom, M., Rosik, D., Carlsson, J., Lundqvist, H., Wennborg, A. & Nilsson, F. Y. Sandstrom, M., Rosik, D., Carlsson, J., Lundqvist, H., Wennborg, A. & Nilsson, F. Y. (2007). Radionuclide therapy of HER2-positive microxenografts using a 177Lu-labeled (2007). Radionuclide therapy of HER2-positive microxenografts using a 177Lu-labeled HER2-specific Affibody molecule. Cancer Res 67, 2773-82. HER2-specific Affibody molecule. Cancer Res 67, 2773-82. 266. Lundberg, E., Hoiden-Guthenberg, I., Larsson, B., Uhlen, M. & Graslund, T. (2007). 266. Lundberg, E., Hoiden-Guthenberg, I., Larsson, B., Uhlen, M. & Graslund, T. (2007). Site-specifically conjugated anti-HER2 Affibody molecules as one-step reagents for Site-specifically conjugated anti-HER2 Affibody molecules as one-step reagents for target expression analyses on cells and xenograft samples. J Immunol Methods 319, 53- target expression analyses on cells and xenograft samples. J Immunol Methods 319, 53- 63. 63. 267. Magnusson, M. K., Henning, P., Myhre, S., Wikman, M., Uil, T. G., Friedman, M., 267. Magnusson, M. K., Henning, P., Myhre, S., Wikman, M., Uil, T. G., Friedman, M., Andersson, K. M., Hong, S. S., Hoeben, R. C., Habib, N. A., Stahl, S., Boulanger, P. & Andersson, K. M., Hong, S. S., Hoeben, R. C., Habib, N. A., Stahl, S., Boulanger, P. & Lindholm, L. (2007). Adenovirus 5 vector genetically re-targeted by an Affibody Lindholm, L. (2007). Adenovirus 5 vector genetically re-targeted by an Affibody molecule with specificity for tumor antigen HER2/neu. Cancer Gene Ther 14, 468-79. molecule with specificity for tumor antigen HER2/neu. Cancer Gene Ther 14, 468-79. 268. Löfblom, J., Feldwisch, J., Tolmachev, V., Carlsson, J., Ståhl, S. & Frejd, F. Y. (2010). 268. Löfblom, J., Feldwisch, J., Tolmachev, V., Carlsson, J., Ståhl, S. & Frejd, F. Y. (2010). Affibody molecules: engineered proteins for therapeutic, diagnostic and Affibody molecules: engineered proteins for therapeutic, diagnostic and biotechnological applications. FEBS Lett 584, 2670-80. biotechnological applications. FEBS Lett 584, 2670-80. 269. Pankov, R. & Yamada, K. M. (2002). Fibronectin at a glance. J Cell Sci 115, 3861-3. 269. Pankov, R. & Yamada, K. M. (2002). Fibronectin at a glance. J Cell Sci 115, 3861-3. 270. Plaxco, K. W., Spitzfaden, C., Campbell, I. D. & Dobson, C. M. (1996). Rapid 270. Plaxco, K. W., Spitzfaden, C., Campbell, I. D. & Dobson, C. M. (1996). Rapid refolding of a proline-rich all-beta-sheet fibronectin type III module. Proc Natl Acad Sci refolding of a proline-rich all-beta-sheet fibronectin type III module. Proc Natl Acad Sci USA 93, 10703-6. USA 93, 10703-6. 271. Lipovsek, D. (2011). Adnectins: engineered target-binding protein therapeutics. Protein 271. Lipovsek, D. (2011). Adnectins: engineered target-binding protein therapeutics. Protein Eng Des Sel 24, 3-9. Eng Des Sel 24, 3-9. 272. Plaxco, K. W., Spitzfaden, C., Campbell, I. D. & Dobson, C. M. (1997). A comparison 272. Plaxco, K. W., Spitzfaden, C., Campbell, I. D. & Dobson, C. M. (1997). A comparison of the folding kinetics and thermodynamics of two homologous fibronectin type III of the folding kinetics and thermodynamics of two homologous fibronectin type III modules. J Mol Biol 270, 763-70. modules. J Mol Biol 270, 763-70. 273. Koide, A., Bailey, C. W., Huang, X. & Koide, S. (1998). The fibronectin type III 273. Koide, A., Bailey, C. W., Huang, X. & Koide, S. (1998). The fibronectin type III domain as a scaffold for novel binding proteins. J Mol Biol 284, 1141-51. domain as a scaffold for novel binding proteins. J Mol Biol 284, 1141-51. 274. Batori, V., Koide, A. & Koide, S. (2002). Exploring the potential of the monobody 274. Batori, V., Koide, A. & Koide, S. (2002). Exploring the potential of the monobody scaffold: effects of loop elongation on the stability of a fibronectin type III domain. scaffold: effects of loop elongation on the stability of a fibronectin type III domain. Protein Eng Des Sel 15, 1015-20. Protein Eng Des Sel 15, 1015-20. 275. Hackel, B. J., Kapila, A. & Wittrup, K. D. (2008). Picomolar affinity fibronectin 275. Hackel, B. J., Kapila, A. & Wittrup, K. D. (2008). Picomolar affinity fibronectin domains engineered utilizing loop length diversity, recursive mutagenesis, and loop domains engineered utilizing loop length diversity, recursive mutagenesis, and loop shuffling. J Mol Biol 381, 1238-52. shuffling. J Mol Biol 381, 1238-52. 276. Hackel, B. J., Ackerman, M. E., Howland, S. W. & Wittrup, K. D. (2010). Stability and 276. Hackel, B. J., Ackerman, M. E., Howland, S. W. & Wittrup, K. D. (2010). Stability and CDR composition biases enrich binder functionality landscapes. J Mol Biol 401, 84-96. CDR composition biases enrich binder functionality landscapes. J Mol Biol 401, 84-96. 277. Forrer, P., Stumpp, M. T., Binz, H. K. & Pluckthun, A. (2003). A novel strategy to 277. Forrer, P., Stumpp, M. T., Binz, H. K. & Pluckthun, A. (2003). A novel strategy to design binding molecules harnessing the modular nature of repeat proteins. FEBS Lett design binding molecules harnessing the modular nature of repeat proteins. FEBS Lett 539, 2-6. 539, 2-6. Sebastian Grimm 97 Sebastian Grimm 97

278. Herrin, B. R., Alder, M. N., Roux, K. H., Sina, C., Ehrhardt, G. R., Boydston, J. A., 278. Herrin, B. R., Alder, M. N., Roux, K. H., Sina, C., Ehrhardt, G. R., Boydston, J. A., Turnbough, C. L., Jr. & Cooper, M. D. (2008). Structure and specificity of lamprey Turnbough, C. L., Jr. & Cooper, M. D. (2008). Structure and specificity of lamprey monoclonal antibodies. Proc Natl Acad Sci USA 105, 2040-5. monoclonal antibodies. Proc Natl Acad Sci USA 105, 2040-5. 279. Binz, H. K., Stumpp, M. T., Forrer, P., Amstutz, P. & Pluckthun, A. (2003). Designing 279. Binz, H. K., Stumpp, M. T., Forrer, P., Amstutz, P. & Pluckthun, A. (2003). Designing repeat proteins: well-expressed, soluble and stable proteins from combinatorial libraries repeat proteins: well-expressed, soluble and stable proteins from combinatorial libraries of consensus ankyrin repeat proteins. J Mol Biol 332, 489-503. of consensus ankyrin repeat proteins. J Mol Biol 332, 489-503. 280. Binz, H. K., Amstutz, P. & Pluckthun, A. (2005). Engineering novel binding proteins 280. Binz, H. K., Amstutz, P. & Pluckthun, A. (2005). Engineering novel binding proteins from nonimmunoglobulin domains. Nat Biotechnol 23, 1257-68. from nonimmunoglobulin domains. Nat Biotechnol 23, 1257-68. 281. Steiner, D., Forrer, P. & Pluckthun, A. (2008). Efficient selection of DARPins with sub- 281. Steiner, D., Forrer, P. & Pluckthun, A. (2008). Efficient selection of DARPins with sub- nanomolar affinities using SRP phage display. J Mol Biol 382, 1211-27. nanomolar affinities using SRP phage display. J Mol Biol 382, 1211-27. 282. Zahnd, C., Kawe, M., Stumpp, M. T., de Pasquale, C., Tamaskovic, R., Nagy- 282. Zahnd, C., Kawe, M., Stumpp, M. T., de Pasquale, C., Tamaskovic, R., Nagy- Davidescu, G., Dreier, B., Schibli, R., Binz, H. K., Waibel, R. & Pluckthun, A. (2010). Davidescu, G., Dreier, B., Schibli, R., Binz, H. K., Waibel, R. & Pluckthun, A. (2010). Efficient tumor targeting with high-affinity designed ankyrin repeat proteins: effects of Efficient tumor targeting with high-affinity designed ankyrin repeat proteins: effects of affinity and molecular size. Cancer Res 70, 1595-605. affinity and molecular size. Cancer Res 70, 1595-605. 283. Hufton, S. E., van Neer, N., van den Beuken, T., Desmet, J., Sablon, E. & 283. Hufton, S. E., van Neer, N., van den Beuken, T., Desmet, J., Sablon, E. & Hoogenboom, H. R. (2000). Development and application of cytotoxic T lymphocyte- Hoogenboom, H. R. (2000). Development and application of cytotoxic T lymphocyte- associated antigen 4 as a protein scaffold for the generation of novel binding ligands. associated antigen 4 as a protein scaffold for the generation of novel binding ligands. FEBS Lett 475, 225-31. FEBS Lett 475, 225-31. 284. Dennis, M. S., Herzka, A. & Lazarus, R. A. (1995). Potent and selective Kunitz domain 284. Dennis, M. S., Herzka, A. & Lazarus, R. A. (1995). Potent and selective Kunitz domain inhibitors of plasma kallikrein designed by phage display. J Biol Chem 270, 25411-7. inhibitors of plasma kallikrein designed by phage display. J Biol Chem 270, 25411-7. 285. Lee, G. F. & Kelley, R. F. (1998). A novel soluble tissue factor variant with an altered 285. Lee, G. F. & Kelley, R. F. (1998). A novel soluble tissue factor variant with an altered factor VIIa binding interface. J Biol Chem 273, 4149-54. factor VIIa binding interface. J Biol Chem 273, 4149-54. 286. Colas, P., Cohen, B., Jessen, T., Grishina, I., McCoy, J. & Brent, R. (1996). Genetic 286. Colas, P., Cohen, B., Jessen, T., Grishina, I., McCoy, J. & Brent, R. (1996). Genetic selection of peptide aptamers that recognize and inhibit cyclin-dependent kinase 2. selection of peptide aptamers that recognize and inhibit cyclin-dependent kinase 2. Nature 380, 548-50. Nature 380, 548-50. 287. Norman, T. C., Smith, D. L., Sorger, P. K., Drees, B. L., O'Rourke, S. M., Hughes, T. 287. Norman, T. C., Smith, D. L., Sorger, P. K., Drees, B. L., O'Rourke, S. M., Hughes, T. R., Roberts, C. J., Friend, S. H., Fields, S. & Murray, A. W. (1999). Genetic selection of R., Roberts, C. J., Friend, S. H., Fields, S. & Murray, A. W. (1999). Genetic selection of peptide inhibitors of biological pathways. Science 285, 591-5. peptide inhibitors of biological pathways. Science 285, 591-5. 288. Abedi, M. R., Caponigro, G. & Kamb, A. (1998). Green fluorescent protein as a 288. Abedi, M. R., Caponigro, G. & Kamb, A. (1998). Green fluorescent protein as a scaffold for intracellular presentation of peptides. Nucleic Acids Res 26, 623-30. scaffold for intracellular presentation of peptides. Nucleic Acids Res 26, 623-30. 289. Pavoor, T. V., Cho, Y. K. & Shusta, E. V. (2009). Development of GFP-based 289. Pavoor, T. V., Cho, Y. K. & Shusta, E. V. (2009). Development of GFP-based biosensors possessing the binding properties of antibodies. Proc Natl Acad Sci U S A biosensors possessing the binding properties of antibodies. Proc Natl Acad Sci U S A 106, 11895-900. 106, 11895-900. 290. Legendre, D., Vucic, B., Hougardy, V., Girboux, A. L., Henrioul, C., Van Haute, J., 290. Legendre, D., Vucic, B., Hougardy, V., Girboux, A. L., Henrioul, C., Van Haute, J., Soumillion, P. & Fastrez, J. (2002). TEM-1 beta-lactamase as a scaffold for protein Soumillion, P. & Fastrez, J. (2002). TEM-1 beta-lactamase as a scaffold for protein recognition and assay. Protein Sci 11, 1506-18. recognition and assay. Protein Sci 11, 1506-18. 291. van der Merwe, P. A. & Davis, S. J. (2003). Molecular interactions mediating T cell 291. van der Merwe, P. A. & Davis, S. J. (2003). Molecular interactions mediating T cell antigen recognition. Annu Rev Immunol 21, 659-84. antigen recognition. Annu Rev Immunol 21, 659-84. 292. Shusta, E. V., Holler, P. D., Kieke, M. C., Kranz, D. M. & Wittrup, K. D. (2000). 292. Shusta, E. V., Holler, P. D., Kieke, M. C., Kranz, D. M. & Wittrup, K. D. (2000). Directed evolution of a stable scaffold for T-cell receptor engineering. Nat Biotechnol Directed evolution of a stable scaffold for T-cell receptor engineering. Nat Biotechnol 18, 754-9. 18, 754-9. 293. Kieke, M. C., Sundberg, E., Shusta, E. V., Mariuzza, R. A., Wittrup, K. D. & Kranz, D. 293. Kieke, M. C., Sundberg, E., Shusta, E. V., Mariuzza, R. A., Wittrup, K. D. & Kranz, D. M. (2001). High affinity T cell receptors from yeast display libraries block T cell M. (2001). High affinity T cell receptors from yeast display libraries block T cell activation by superantigens. J Mol Biol 307, 1305-15. activation by superantigens. J Mol Biol 307, 1305-15. 294. Malabarba, M. G., Milia, E., Faretta, M., Zamponi, R., Pelicci, P. G. & Di Fiore, P. P. 294. Malabarba, M. G., Milia, E., Faretta, M., Zamponi, R., Pelicci, P. G. & Di Fiore, P. P. (2001). A repertoire library that allows the selection of synthetic SH2s with altered (2001). A repertoire library that allows the selection of synthetic SH2s with altered binding specificities. Oncogene 20, 5186-94. binding specificities. Oncogene 20, 5186-94. 295. Hiipakka, M., Poikonen, K. & Saksela, K. (1999). SH3 domains with high affinity and 295. Hiipakka, M., Poikonen, K. & Saksela, K. (1999). SH3 domains with high affinity and engineered ligand specificities targeted to HIV-1 Nef. J Mol Biol 293, 1097-106. engineered ligand specificities targeted to HIV-1 Nef. J Mol Biol 293, 1097-106. 98 Ribosome display for selection and evolution of affibody molecules 98 Ribosome display for selection and evolution of affibody molecules

296. Schneider, S., Buchert, M., Georgiev, O., Catimel, B., Halford, M., Stacker, S. A., 296. Schneider, S., Buchert, M., Georgiev, O., Catimel, B., Halford, M., Stacker, S. A., Baechi, T., Moelling, K. & Hovens, C. M. (1999). Mutagenesis and selection of PDZ Baechi, T., Moelling, K. & Hovens, C. M. (1999). Mutagenesis and selection of PDZ domains that bind new protein targets. Nat Biotechnol 17, 170-5. domains that bind new protein targets. Nat Biotechnol 17, 170-5. 297. Stumpp, M. T., Forrer, P., Binz, H. K. & Pluckthun, A. (2003). Designing repeat 297. Stumpp, M. T., Forrer, P., Binz, H. K. & Pluckthun, A. (2003). Designing repeat proteins: modular leucine-rich repeat protein libraries based on the mammalian proteins: modular leucine-rich repeat protein libraries based on the mammalian ribonuclease inhibitor family. J Mol Biol 332, 471-87. ribonuclease inhibitor family. J Mol Biol 332, 471-87. 298. Parmeggiani, F., Pellarin, R., Larsen, A. P., Varadamsetty, G., Stumpp, M. T., Zerbe, 298. Parmeggiani, F., Pellarin, R., Larsen, A. P., Varadamsetty, G., Stumpp, M. T., Zerbe, O., Caflisch, A. & Pluckthun, A. (2008). Designed armadillo repeat proteins as general O., Caflisch, A. & Pluckthun, A. (2008). Designed armadillo repeat proteins as general peptide-binding scaffolds: consensus design and computational optimization of the peptide-binding scaffolds: consensus design and computational optimization of the hydrophobic core. J Mol Biol 376, 1282-304. hydrophobic core. J Mol Biol 376, 1282-304. 299. Martin, L., Barthe, P., Combes, O., Roumestand, C. & Vita, C. (2000). Engineering 299. Martin, L., Barthe, P., Combes, O., Roumestand, C. & Vita, C. (2000). Engineering Novel Bioactive Mini-Proteins on Natural Scaffolds. Tetrahedron 56, 9451-9460. Novel Bioactive Mini-Proteins on Natural Scaffolds. Tetrahedron 56, 9451-9460. 300. Vita, C., Roumestand, C., Toma, F. & Menez, A. (1995). Scorpion toxins as natural 300. Vita, C., Roumestand, C., Toma, F. & Menez, A. (1995). Scorpion toxins as natural scaffolds for protein engineering. Proc Natl Acad Sci USA 92, 6404-8. scaffolds for protein engineering. Proc Natl Acad Sci USA 92, 6404-8. 301. Vita, C., Drakopoulou, E., Vizzavona, J., Rochette, S., Martin, L., Menez, A., 301. Vita, C., Drakopoulou, E., Vizzavona, J., Rochette, S., Martin, L., Menez, A., Roumestand, C., Yang, Y. S., Ylisastigui, L., Benjouad, A. & Gluckman, J. C. (1999). Roumestand, C., Yang, Y. S., Ylisastigui, L., Benjouad, A. & Gluckman, J. C. (1999). Rational engineering of a miniprotein that reproduces the core of the CD4 site Rational engineering of a miniprotein that reproduces the core of the CD4 site interacting with HIV-1 envelope glycoprotein. Proc Natl Acad Sci USA 96, 13091-6. interacting with HIV-1 envelope glycoprotein. Proc Natl Acad Sci USA 96, 13091-6. 302. Miao, Z., Ren, G., Liu, H., Kimura, R. H., Jiang, L., Cochran, J. R., Gambhir, S. S. & 302. Miao, Z., Ren, G., Liu, H., Kimura, R. H., Jiang, L., Cochran, J. R., Gambhir, S. S. & Cheng, Z. (2009). An engineered knottin peptide labeled with 18F for PET imaging of Cheng, Z. (2009). An engineered knottin peptide labeled with 18F for PET imaging of integrin expression. Bioconjug Chem 20, 2342-7. integrin expression. Bioconjug Chem 20, 2342-7. 303. Hanes, J., Jermutus, L., Weber-Bornhauser, S., Bosshard, H. R. & Plueckthun, A. 303. Hanes, J., Jermutus, L., Weber-Bornhauser, S., Bosshard, H. R. & Plueckthun, A. (1998). Ribosome display efficiently selects and evolves high-affinity antibodies in (1998). Ribosome display efficiently selects and evolves high-affinity antibodies in vitro from immune libraries. Proc Natl Acad Sci USA, 6. vitro from immune libraries. Proc Natl Acad Sci USA, 6. 304. Schimmele, B., Grafe, N. & Pluckthun, A. (2005). Ribosome display of mammalian 304. Schimmele, B., Grafe, N. & Pluckthun, A. (2005). Ribosome display of mammalian receptor domains. Protein Eng Des Sel 18, 285-294. receptor domains. Protein Eng Des Sel 18, 285-294. 305. Amstutz, P., Pelletier, J. N., Guggisberg, A., Jermutus, L., Cesaro-Tadic, S., Zahnd, C. 305. Amstutz, P., Pelletier, J. N., Guggisberg, A., Jermutus, L., Cesaro-Tadic, S., Zahnd, C. & Plückthun, A. (2002). In vitro selection for catalytic activity with ribosome display. J & Plückthun, A. (2002). In vitro selection for catalytic activity with ribosome display. J Am Chem Soc 124, 9396-403. Am Chem Soc 124, 9396-403. 306. Nord, K., Gunneriusson, E., Ringdahl, J., Ståhl, S., Uhlén, M. & Nygren, P. A. (1997). 306. Nord, K., Gunneriusson, E., Ringdahl, J., Ståhl, S., Uhlén, M. & Nygren, P. A. (1997). Binding proteins selected from combinatorial libraries of an alpha-helical bacterial Binding proteins selected from combinatorial libraries of an alpha-helical bacterial receptor domain. Nat Biotechnol 15, 772-7. receptor domain. Nat Biotechnol 15, 772-7. 307. Cai, S., Lee, C. C. & Kohwi-Shigematsu, T. (2006). SATB1 packages densely looped, 307. Cai, S., Lee, C. C. & Kohwi-Shigematsu, T. (2006). SATB1 packages densely looped, transcriptionally active chromatin for coordinated expression of cytokine genes. Nat transcriptionally active chromatin for coordinated expression of cytokine genes. Nat Genet 38, 1278-88. Genet 38, 1278-88. 308. Cai, S., Han, H.-J. & Kohwi-Shigematsu, T. (2003). Tissue-specific nuclear architecture 308. Cai, S., Han, H.-J. & Kohwi-Shigematsu, T. (2003). Tissue-specific nuclear architecture and gene expression regulated by SATB1. Nat Genet 34, 42-51. and gene expression regulated by SATB1. Nat Genet 34, 42-51. 309. Dickinson, L. A., Dickinson, C. D. & Kohwi-Shigematsu, T. (1997). An atypical 309. Dickinson, L. A., Dickinson, C. D. & Kohwi-Shigematsu, T. (1997). An atypical homeodomain in SATB1 promotes specific recognition of the key structural element in homeodomain in SATB1 promotes specific recognition of the key structural element in a matrix attachment region. J Biol Chem 272, 11463-70. a matrix attachment region. J Biol Chem 272, 11463-70. 310. Han, H.-J., Russo, J., Kohwi, Y. & Kohwi-Shigematsu, T. (2008). SATB1 310. Han, H.-J., Russo, J., Kohwi, Y. & Kohwi-Shigematsu, T. (2008). SATB1 reprogrammes gene expression to promote breast tumour growth and metastasis. Nature reprogrammes gene expression to promote breast tumour growth and metastasis. Nature 452, 187-193. 452, 187-193. 311. Iorns, E., Hnatyszyn, H. J., Seo, P., Clarke, J., Ward, T. & Lippman, M. (2010). The 311. Iorns, E., Hnatyszyn, H. J., Seo, P., Clarke, J., Ward, T. & Lippman, M. (2010). The role of SATB1 in breast cancer pathogenesis. J Natl Cancer Inst 102, 1284-96. role of SATB1 in breast cancer pathogenesis. J Natl Cancer Inst 102, 1284-96. 312. Hanker, L. C., Karn, T., Mavrova-Risteska, L., Ruckhäberle, E., Gaetje, R., Holtrich, 312. Hanker, L. C., Karn, T., Mavrova-Risteska, L., Ruckhäberle, E., Gaetje, R., Holtrich, U., Kaufmann, M., Rody, A. & Wiegratz, I. (2010). SATB1 gene expression and breast U., Kaufmann, M., Rody, A. & Wiegratz, I. (2010). SATB1 gene expression and breast cancer prognosis. Breast. cancer prognosis. Breast. 313. Chi, K. (2009). Super-resolution microscopy: breaking the limits. Nat Meth 6, 15-18. 313. Chi, K. (2009). Super-resolution microscopy: breaking the limits. Nat Meth 6, 15-18. Sebastian Grimm 99 Sebastian Grimm 99

314. Donnert, G., Keller, J., Medda, R., Andrei, M. A., Rizzoli, S. O., Luhrmann, R., Jahn, 314. Donnert, G., Keller, J., Medda, R., Andrei, M. A., Rizzoli, S. O., Luhrmann, R., Jahn, R., Eggeling, C. & Hell, S. W. (2006). Macromolecular-scale resolution in biological R., Eggeling, C. & Hell, S. W. (2006). Macromolecular-scale resolution in biological fluorescence microscopy. Proc Natl Acad Sci USA 103, 11440-5. fluorescence microscopy. Proc Natl Acad Sci USA 103, 11440-5. 315. Roberts, P. J. & Der, C. J. (2007). Targeting the Raf-MEK-ERK mitogen-activated 315. Roberts, P. J. & Der, C. J. (2007). Targeting the Raf-MEK-ERK mitogen-activated protein kinase cascade for the treatment of cancer. Oncogene 26, 3291-310. protein kinase cascade for the treatment of cancer. Oncogene 26, 3291-310. 316. Malumbres, M. & Barbacid, M. (2003). RAS oncogenes: the first 30 years. Nat Rev 316. Malumbres, M. & Barbacid, M. (2003). RAS oncogenes: the first 30 years. Nat Rev Cancer 3, 459-65. Cancer 3, 459-65. 317. Hatzivassiliou, G., Song, K., Yen, I., Brandhuber, B. J., Anderson, D. J., Alvarado, R., 317. Hatzivassiliou, G., Song, K., Yen, I., Brandhuber, B. J., Anderson, D. J., Alvarado, R., Ludlam, M. J. C., Stokoe, D., Gloor, S. L., Vigers, G., Morales, T., Aliagas, I., Liu, B., Ludlam, M. J. C., Stokoe, D., Gloor, S. L., Vigers, G., Morales, T., Aliagas, I., Liu, B., Sideris, S., Hoeflich, K. P., Jaiswal, B. S., Seshagiri, S., Koeppen, H., Belvin, M., Sideris, S., Hoeflich, K. P., Jaiswal, B. S., Seshagiri, S., Koeppen, H., Belvin, M., Friedman, L. S. & Malek, S. (2010). RAF inhibitors prime wild-type RAF to activate Friedman, L. S. & Malek, S. (2010). RAF inhibitors prime wild-type RAF to activate the MAPK pathway and enhance growth. Nature 464, 431-5. the MAPK pathway and enhance growth. Nature 464, 431-5. 318. Roy, S., Lane, A., Yan, J., McPherson, R. & Hancock, J. F. (1997). Activity of plasma 318. Roy, S., Lane, A., Yan, J., McPherson, R. & Hancock, J. F. (1997). Activity of plasma membrane-recruited Raf-1 is regulated by Ras via the Raf zinc finger. J Biol Chem 272, membrane-recruited Raf-1 is regulated by Ras via the Raf zinc finger. J Biol Chem 272, 20139-45. 20139-45. 319. Cattaneo, A. & Biocca, S. (1999). The selection of intracellular antibodies. Trends 319. Cattaneo, A. & Biocca, S. (1999). The selection of intracellular antibodies. Trends Biotechnol 17, 115-21. Biotechnol 17, 115-21. 320. Jonsson, A., Dogan, J., Herne, N., Abrahmsén, L. & Nygren, P.-A. (2008). Engineering 320. Jonsson, A., Dogan, J., Herne, N., Abrahmsén, L. & Nygren, P.-A. (2008). Engineering of a femtomolar affinity binding protein to human serum albumin. Protein Eng Des Sel of a femtomolar affinity binding protein to human serum albumin. Protein Eng Des Sel 21, 515-27. 21, 515-27. 321. Clamp, M., Fry, B., Kamal, M., Xie, X., Cuff, J., Lin, M. F., Kellis, M., Lindblad-Toh, 321. Clamp, M., Fry, B., Kamal, M., Xie, X., Cuff, J., Lin, M. F., Kellis, M., Lindblad-Toh, K. & Lander, E. S. (2007). Distinguishing protein-coding and noncoding genes in the K. & Lander, E. S. (2007). Distinguishing protein-coding and noncoding genes in the human genome. Proc Natl Acad Sci U S A 104, 19428-33. human genome. Proc Natl Acad Sci U S A 104, 19428-33. 322. Anfinsen, C. B. (1973). Principles that govern the folding of protein chains. Science 322. Anfinsen, C. B. (1973). Principles that govern the folding of protein chains. Science 181, 223-30. 181, 223-30.