Shark Attack: high affinity binding proteins derived from shark vNAR domains by stepwise in vitro affinity maturation.

G Model

ARTICLE IN PRESS

BIOTEC-6679; No. of Pages 10

Journal of Biotechnology xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Journal of Biotechnology journal homepage: www.elsevier.com/locate/jbiotec

Shark Attack: High affinity binding proteins derived from shark vNAR domains by stepwise in vitro affinity maturation Stefan Zielonka a , Niklas Weber a , Stefan Becker b , Achim Doerner b , Andreas Christmann a , Christine Christmann a , Christina Uth a , Janine Fritz a , Elena Schäfer a , Björn Steinmann a , Martin Empting c,1 , Pia Ockelmann d,e , Michael Lierz f , Harald Kolmar a,∗ a

Institute for Organic Chemistry and Biochemistry, Technische Universität Darmstadt, Alarich-Weiss-Strasse 4, D-64287 Darmstadt, Germany Protein Engineering and Antibody Technologies, Merck Serono, Merck KGaA, Frankfurter Straße 250, D-64293 Darmstadt, Germany c Helmholtz-Institute for Pharmaceutical Research Saarland (HIPS), Department Drug Design and Optimization, Saarland University, Campus C2.3, D-66123 Saarbrücken, Germany d Goethe-University Frankfurt, Faculty of Biosciences, Max-von-Laue-Str. 13, D-60438 Frankfurt am Main, Germany e University Hospital Frankfurt, Department of Anesthesiology, Intensive-Care Medicine and Pain Therapy, Theodor-Stern-Kai 7, D-60590 Frankfurt am Main, Germany f Clinic for Birds, Reptiles, Amphibians and Fish, Justus-Liebig University, Gießen, Frankfurter Str. 91-93, D-35392 Giessen, Germany b

a r t i c l e

i n f o

Article history: Received 21 February 2014 Received in revised form 8 April 2014 Accepted 28 April 2014 Available online xxx Keywords: Shark vNAR antibody Semi-synthetic library Stepwise in vitro affinity maturation Yeast surface display

a b s t r a c t A novel method for stepwise in vitro affinity maturation of antigen-specific shark vNAR domains is described that exclusively relies on semi-synthetic repertoires derived from non-immunized sharks. Target-specific molecules were selected from a CDR3-randomized bamboo shark (Chiloscyllium plagiosum) vNAR library using yeast surface display as platform technology. Various antigen-binding vNAR domains were easily isolated by screening against several therapeutically relevant antigens, including the epithelial cell adhesion molecule (EpCAM), the Ephrin type-A receptor 2 (EphA2), and the human serine protease HTRA1. Affinity maturation was demonstrated for EpCAM and HTRA1 by diversifying CDR1 of target-enriched populations which allowed for the rapid selection of nanomolar binders. EpCAM-specific vNAR molecules were produced as soluble proteins and more extensively characterized via thermal shift assays and biolayer interferometry. Essentially, we demonstrate that high-affinity binders can be generated in vitro without largely compromising the desirable high thermostability of the vNAR scaffold. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Antibody-based molecules are successfully and frequently used for a plethora of biotechnological and biomedical applications. They function as versatile ligands that selectively target antigens. Applications range from affinity purification of target proteins over immunodiagnostics to disease treatment. The number of 28 currently marketed antibodies and several hundred of candidates in clinical trials is showing clear evidence that those biologics are extremely attractive for clinical development (Buss et al., 2012; Reichert, 2012, 2013). Canonical antibodies are structurally complex, large hetero-tetrameric proteins which consist of two heavy chains and two light chains. The antigen-binding site is composed

∗ Corresponding author. Tel.: +49 6151 16 4742; fax: +49 6151 16 5399. E-mail address: [email protected] (H. Kolmar). 1 Present address.

of one variable domain of the light chain and one variable domain of the heavy chain. Despite their successful application in various fields, antibody molecules encounter several limitations owing to their inherent structural complexity and size. For example, they usually demonstrate only a restricted ability to access recessed cryptic epitopes (Barelle et al., 2009; Muller et al., 2012). Furthermore, the mobility i.e. tissue penetration of those molecules is affected by reason of their large size (Mordenti et al., 1999). The need for smaller binding units led to the development of scaffolds based on conventional antibodies such as Fab-fragments or single-chain Fv fragments, which have in common that they are heterodimers of heavy and light chain domains. More recently, single domain antibodies based on single human antibody fragments of the VH or VL domain have been engineered (Dudgeon et al., 2012; Holt et al., 2003; Nelson, 2010). The cartilaginous fish and camelids possess natural antibodies composed only of heavy chains (Greenberg et al., 1995; Hamers-Casterman et al., 1993). Hence, the antigen-binding site is formed by only one single domain,

http://dx.doi.org/10.1016/j.jbiotec.2014.04.023 0168-1656/© 2014 Elsevier B.V. All rights reserved.

Please cite this article in press as: Zielonka, S., et al., Shark Attack: High affinity binding proteins derived from shark vNAR domains by stepwise in vitro affinity maturation. J. Biotechnol. (2014), http://dx.doi.org/10.1016/j.jbiotec.2014.04.023

G Model BIOTEC-6679; No. of Pages 10 2

ARTICLE IN PRESS S. Zielonka et al. / Journal of Biotechnology xxx (2014) xxx–xxx

referred to as vNAR and VHH, respectively. Variable domains of the immunoglobulin new antigen receptor (IgNAR) and camelid heavy chain antibodies are extraordinary stable proteins, with vNAR being the smallest antibody-like binding domain known in the animal kingdom to date (Barelle et al., 2009; Dooley and Flajnik, 2006; Goodchild et al., 2011; Liu et al., 2007a; Muyldermans, 2001; Roux et al., 1998). In addition, antibody fragments of camelids and sharks have been shown to target more hidden epitopes such as the active site of enzymes (Desmyter et al., 1996; Stanfield et al., 2004). IgNAR was first identified in the nurse shark Ginglymostoma cirratum in 1995 by Flajnik and co-workers (Greenberg et al., 1995). It is a homodimer of heavy chains and each chain consists of five constant domains followed by a hinge-like region and the variable domain. The vNAR domain contains four (hyper)variable loops, designated as CDR1, hypervariable loop 2 (HV2), hypervariable loop 4 (HV4) and CDR3, respectively (Fig. 1A). Due to a short C strand and a very short C strand, the CDR2 loop, referred to as HV2, wraps around the bottom of the molecule (Barelle et al., 2009; Nuttall, 2012). Another non-conventional region with enhanced mutation frequency, termed hypervariable loop 4 (HV4) is located between HV2 and CDR3 (Kovalenko et al., 2013). IgNAR diversifies by somatic hypermutation in an antigen-driven manner (Diaz et al., 1998, 1999; Dooley and Flajnik, 2005, 2006; Stanfield et al., 2007). The vNAR domain was successfully utilized to generate binding molecules against a range of targets from immunized, non-immunized, semi-synthetic and artificial vNAR libraries, using phage display and ribosome display as platform technologies (Dooley et al., 2003; Goodchild et al., 2011; Kopsidas et al., 2006; Liu et al., 2007a,b; Muller et al., 2012; Nuttall et al., 2001, 2002, 2003; Ohtani et al., 2013; Shao et al., 2007). Recently, vNAR molecules with picomolar affinities to their target have been isolated from an immunized shark by phage display (Muller et al., 2012). Furthermore, improvements of binding affinities have been achieved using error prone PCR and Q␤ replicase for in vitro affinity maturation (Kopsidas et al., 2006; Nuttall et al., 2004). Immunization still is the preferred route for the isolation of high-affinity binders from sharks. However, it has been shown that immunization of G. cirratum and Squalus acanthias is a time consuming procedure (Dooley and Flajnik, 2005; Muller et al., 2012). Moreover, there is no guarantee to obtain highly specific binders from immunized shark libraries. For example, Dooley and colleagues were unable to detect an antigen-specific response for IgNAR from immunized small-spotted catshark, Scyliorhinus canicula, and only half of the sharks immunized in a study with human serum albumin as antigen (HSA) showed a significant antibody titer (Crouch et al., 2013; Muller et al., 2012). Herein, we describe the rapid isolation of antigen-binding vNARs with artificial CDR3 loops against three different diseaserelated targets, namely EpCAM, HTRA1 and EphA2 from the bamboo shark (Chiloscyllium plagiosum) vNAR repertoire using yeast surface display (YSD) as platform technology combined with subsequent affinity maturation of target-binding vNARs. The epithelial cell adhesion molecule (EpCAM) is a membrane glycoprotein overexpressed on most carcinomas. It functions as a signaling receptor involved in the transformation of malignant cells. Besides, it was identified as a marker of cancer-initiating cells (Imrich et al., 2012; Ni et al., 2012; van der Gun et al., 2010). Human serine protease HTRA1 is an oligomeric serine protease involved in protein quality control. It is associated with a multitude of severe diseases such as arthritis, cancer or Alzheimer’s disease (Truebestein et al., 2011). The receptor tyrosine kinase EphA2 is overexpressed and functionally altered in aggressive tumors and it is believed that those aberrations promote growth and invasiveness of malignant cells (Kinch and Carles-Kinch, 2003). High levels of EphA2 have been reported in breast cancer, colon cancer, prostate cancer, non-small cell lung cancers and aggressive melanomas (Walker-Daniels et al.,

2003). To the best of our knowledge, this approach has not been reported to date. 2. Materials and methods 2.1. Media and reagents dYT medium was composed of 16 g/L tryptone, 10 g/L yeast extract, 5 g/L sodium chloride and 25 mg/L chloramphenicol. YPD medium contained 20 g/L tryptone, 20 g/L dextrose and 10 g/L yeast extract. SD-CAA medium was composed of 1.7 g/L yeast nitrogen base without amino acids and ammonium sulfate, 5 g/L ammonium sulfate, 5 g/L Bacto casamino acids, 20 g/L dextrose, 8.6 g/L NaH2 PO4 × H2 O, and 5.4 g/L Na2 HPO4 . SG-CAA medium was prepared similarly except for the substitution of 20 g/L dextrose by galactose. Additionally 10% (w/v) polyethylene glycol 8000 (PEG 8000) was incorporated. Phosphate-buffered saline (PBS) contained 8.1 g/L NaCl, 0.75 g/L KCl, 1.13 g/L Na2 HPO4 and 0.27 g/L KH2 PO4 , pH 7.4. Recombinant human His-tagged EpCAM was purchased from AcroBiosystems, (rh) EphA2 from R&D systems and recombinant mouse EphA2 from Sino Biological Inc. The catalytic domain from residue 158 to 373 of human HTRA1 was engineered for shifting oligomeric equilibrium to monomeric state by exchange of the residues Y168A, F171A and F278A mainly involved in trimerization. The protein was produced via gene expression using pET21d (Stratagene) in Escherichia coli strain BL21 at 16 ◦ C, 180 rpm for 24 h induced with 1 mM IPTG at OD600 of about 0.8. Cells were resuspended and lysed in 100 mM Tris–HCl 150 mM NaCl, pH 8.0. After cell disruption, his-tagged HTRA1 was purified using HisTrap (GE Healthcare), concentrated by ultrafiltration and further purified using a size exclusion Superdex 200 column 16/60 (GE Healthcare) equilibrated with 150 mM Tris–HCl, 150 mM NaCl, pH 8.0. After buffer exchange to 0.1 M sodium carbonate, pH 9.0, to a 2 mg/mL solution a 7-fold molar excess of FITC (Thermo Fisher) was added. Reaction was carried out over night at 4 ◦ C in the dark. Fluorescent labeled HTRA1-mono was purified using a desalting mini spin column (Thermo Fisher) equilibrated with 100 mM Tris–HCl, 150 mM NaCl, pH 8.0 according to the manufacturer’s protocol and stored at −80 ◦ C. 2.2. Animals Three specimen of the bamboo shark (C. plagiosum) in the age of approximately eight months were obtained from the Vivarium of the Staatliches Museum für Naturkunde, Karlsruhe. Blood samples were harvested from the caudal vein from anesthetized animals (MS-222) and subsequently transferred into Tri® Reagent BD (Sigma–Aldrich). Total RNA was extracted from whole blood according to the manufacturer’s protocol (Sigma–Aldrich). Procedures were conducted in accordance with the national laws § 4 Abs. 3 of the German Tierschutzgesetz (TierSchG, animal welfare act). Permission number: V 54–19 c 20 15 (1) Gl 18/19 Nr. A 35/2011, Regierungspräsidium Giessen, Germany (Regional council Giessen). 2.3. cDNA cloning and analysis of the natural vNAR repertoire of the bamboo shark The Omniscript® Reverse Transcription Kit (Qiagen) was used for cDNA synthesis as described in the manufacturer’s protocol. Five reactions of pooled total RNA from one individual were carried out in parallel, each containing about 2 ␮g of total RNA as template using an oligo(dT)18 primer. From each reaction 5 ␮L were used for the follow-up gene-specific amplification of the vNAR domains using the primers bamboo/nat up and bamboo/nat lo (Table S1).


G Model BIOTEC-6679; No. of Pages 10


3

Fig. 1. Analysis of the natural vNAR repertoire of the bamboo shark. (A) Structure of a vNAR binding domain (left: ribbon model, right: surface) indicating variable regions CDR1, HV2, HV4, and CDR3. Disulfide bond is shown in yellow. The model was generated on the basis of a published X-ray structure (PDB-ID: 1VES) using YASARA structure (Duan et al., 2003) and rendered with POVRay using radiosity (left) and subsurface light transport (right). (B) Length distribution of CDR3 of the vNAR repertoire from nonimmunized bamboo sharks. (C) Consensus-sequence of the natural vNAR repertoire from non-immunized bamboo sharks. A set of approx. 1200 sequences was analyzed. The primer-mediated parts at the respective N-terminus and C-terminus are not shown. CDR1 is shaded blue, HV2 and HV4 are shaded red, and CDR3 is shaded green. Consensus of residues at each position is indicated by a black bar. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

PCR conditions were as follows: 94 ◦ C for 2 min, 30 cycles of 30 s at 94 ◦ C, 30 s at 55 ◦ C and 40 s at 72 ◦ C, followed by 72 ◦ C for 7 min. PCR products were pooled and the vNAR repertoire was analyzed using next-generation sequencing on 454 GS Junior System platforms (Roche) according to the manufacturer’s instructions. 2.4. Library construction and establishment of randomized sublibraries for affinity maturation Total RNA from three individuals were used as template for cDNA synthesis using the Omniscript® Reverse Transcription Kit (Qiagen) and OdT-Oligonucleotides. From each individual three reactions were processed in parallel, each containing approximately 2 ␮g of RNA. From each reaction 5 ␮L were used for the follow-up gene-specific amplification as mentioned above. The PCR amplified vNAR products were used as starting material for the semi-synthetic library construction. The initial library was established in three consecutive PCR-steps (Fig. 2A). For all reactions the conditions were: 94 ◦ C for 2 min, 35 cycles of 30 s at 94 ◦ C, 30 s at 55 ◦ C and 40 s at 72 ◦ C, followed by 72 ◦ C for 7 min. Primer sequences are listed in Table S1. In the first reaction PCR amplified vNAR product from the natural repertoire was used as template with the primer combinations FR1/CDR1/Tyr up and FR3 lo. The forward primer replaced Cys by Tyr and incorporated a marginal diversity within CDR1. The PCR product was purified via Wizard® SV Gel and PCR Clean-up System (Promega). Subsequently, the second PCR was performed to fully randomize CDR3 using the primer-combination FR1 up and CDR3/rand/FR4 lo. After purification by Wizard® SV Gel and PCR Clean-up System, the PCR product was used as a template for the third reaction using primers with overlaps up- and downstream of the NheI and BamHI restriction sites of the pCT-plasmid (Boder and Wittrup, 1997), respectively (GR up and GR lo). The pCT vector was digested with NheI and BamHI and purified via Wizard® SV Gel and PCR Clean-up System. For electroporation 1–2 ␮g of the digested plasmid and 6–8 ␮g of insert were used. After 1 hour incubation at 30 ◦ C (1:1 YPD

and 1 M sorbitol) library size was calculated by dilution plating. Yeast cells (EBY 100) were transferred into SD-CAA medium. Stocks were stored at −80 ◦ C. For yeast surface display cells were grown overnight at 30 ◦ C in SD-CAA medium, transferred into SG-CAA medium and incubated for 1–2 days at 20 ◦ C. For affinity maturation of the particular enriched antigenbinding population, plasmid-DNA was isolated from yeast cells after sorting the cells via FACS and incubation at 30 ◦ C for two days. Sublibraries with totally randomized CDR1 were constructed in a 3-step PCR with the following conditions: 94 ◦ C for 2 min, 35 cycles of 30 s at 94 ◦ C, 30 s at 55 ◦ C and 40 s at 72 ◦ C, followed by 72 ◦ C for 7 min. In the first PCR reaction, the primer pair CDR1rand up/GR lo was used to randomize CDR1 followed by two consecutive PCR reactions with primer pairs FR1 up/GR lo and GR up/GR lo, respectively. Gap repair cloning and transformations were executed as described above. 2.5. Binding assays on the yeast surface and library screening for the isolation of target-specific vNAR molecules Flow cytometry was used to analyze presentation on the yeast surface and for single clone analysis. About 107 cells were labeled consecutively with anti-cMyc antibody (monoclonal, mouse, made in-house) or with anti-HA-tag antibody (polyclonal, rabbit, eBioscience, diluted 1:10), anti-mouse IgG biotin conjugate (goat, Sigma–Aldrich, diluted 1:10 in PBS) or anti-rabbit-biotin (goat, Sigma–Aldrich) and streptavidin–allophycocyanin conjugate (eBioscience, diluted 1:10) for at least 10 min on ice. For single clone analysis, vNAR-presenting cells were either incubated with biotinylated, FITC-labeled or His-tagged antigen for 15 min on ice and subsequently stained with Steptavidin-APC (diluted 1:10) or PentaHis Alexa Fluor 488 conjugate (Qiagen, diluted 1:30) for 10 min. When biotinylated antigen was used, analysis of surface presentation was conducted using anti-cMyc antibody and anti-mouse FITC conjugate (goat, Sigma–Aldrich, diluted 1:10). Plasmid DNA from positive clones was isolated and transformed into E. coli Top




4

10 competent cells for plasmid rescue and sent out for sequencing with pCT-seq lo or pCT-seq up oligonucleotide (Table S1). Affinities of isolated vNAR variants on yeast cells were determined as described (Chao et al., 2006; VanAntwerp and Wittrup, 2000). Flow cytometric analysis was performed to determine cMyc normalized antigen-binding. Hence, exclusively cells displaying the vNAR domain were included for KD calculation. At least eight different antigen concentrations were used and cells were incubated for at least 1 h with the respective antigen. For the final calculation, cMyc-normalized RFU were plotted against antigen concentration. Library screening for the isolation of antigen-specific vNAR molecules was performed on a MoFlo cell sorter (Beckman Coulter) and analyzed via Summit 4.3. For two-dimensional screening, cells were first labeled for detection of surface presentation via cMyc labeling. After washing with ice-cold PBS, cells were resuspended in PBS containing the desired concentration of antigen and incubated on ice for at least 30 min. Subsequently, antigen-binding was detected using streptavidin–allophycocyanin conjugate or PentaHis Alexa Fluor 488 conjugate (diluted 1:10 or 1:30, respectively, 10 min on ice). For the first rounds of sorting of the initial library, approximately 2 × 108 cells were analyzed and sorted. Consecutive rounds were at least performed with a 10-fold excess of cells that were collected in the previous round to ensure coverage of the enriched population. Sublibrary screening was executed with at least 10-fold the number of cells of the respective library diversity. 2.6. Soluble expression of vNAR constructs and protein purification Selected vNARs were expressed in the pMX vector (Wentzel et al., 1999) that introduces maltose binding protein (MalE) and a hexahistine tag for fusion protein expression and purification (Fig. 4A). E. coli BMH 71-18 cells containing the respective plasmid were grown to an OD600 between 0.7 and 1 in dYT medium containing 25 mg/L chloramphenicol and protein production was induced with 1 mM IPTG. Cells were grown overnight (approx. 16 h) at 29 ◦ C and harvested by centrifugation. The periplasmic fraction was isolated as described elsewhere (Minsky et al., 1986). Respective recombinant vNAR proteins were purified by metal chelate affinity chromatography (HisTrap, GE Healthcare), equilibrated in PBS, pH 7.4. Protein was eluted using a linear imidazole gradient (0–100% of 1 M imidazole over 30 min). For tabacco etch virus protease (TEV) cleavage, 10-fold TEV buffer (1.5 M NaCl, 500 mM Tris–HCl, pH 7.5 and recombinant TEV protease (ratio 1:20) were added and incubated over night at 4 ◦ C. Final purification was performed using gel filtration on either a Superdex 200 pg 16/60 or on a Superdex 75 pg 16/60 (GE Healthcare), equilibrated with PBS, pH 7.4. 2.7. Thermal shift assay Measurements were performed in quadruplicates on a BioRad 96CFX RT-PCR detection system with 0.5 ◦ C/30 s to 99 ◦ C. Tm values were obtained from melting curves using the corresponding BioRad analysis software. All reactions were performed in PBS, pH 7.4 in the presence of SYPRO Orange (diluted 1:1000, Sigma–Aldrich) and 0.1–0.5 mg/mL protein. 2.8. Binding kinetics The Octet RED96 system (FortéBio) was used to determine the equilibrium dissociation constant (KD ). All assays were performed with Streptavidin dip and read biosensors in kinetics buffer (PBS, pH 7.4, 0.1% (w/v) BSA, 0.02% Tween-20). Sensors were loaded with biotinylated vNARs at approx. 10 ␮g/mL. A target-unspecific vNAR was used as negative control and tested at the highest antigen

concentration. Kinetic data sets were fitted using 1:1 Langmuir binding via the manufacturer’s data analysis software with Savitzky–Golay filtering. 3. Results 3.1. Library construction and analysis Deep sequencing of the vNAR repertoire from blood circulating lymphocytes of a non-immunized bamboo shark revealed that vNAR diversity in this species is mainly CDR3 based, while CDR1 as well as HV2 and HV4 only display minor variations (Fig. 1C). Library construction on S. cerevisiae allows for the facile generation of over 108 clones, which supposably exceeds the natural vNAR diversity found in non-immunized animals. As a consequence, to extend natural vNAR diversity, we constructed a semi-synthetic vNAR library by polymerase chain reaction where in the framework of the natural vNAR repertoire CDR3 was fully randomized. This was achieved by the incorporation of trinucleotide mixtures encoding 19 amino acids except cysteine into the corresponding oligonucleotide (Table S1). Framework regions and the hypervariable loops were used from non-immunized repertoires of a cohort of three bamboo sharks. In analogy to the natural repertoire (Fig. 1B), we designed the CDR3 to comprise 12 randomized residues. The cysteine residue of CDR1 was replaced by tyrosine to avoid the presence of an unpaired cysteine at that position. The library was established in three subsequent PCR reactions (Fig. 2A) and a yeast surface display library with a diversity of approximately 2 × 108 unique clones was constructed by transformation of yeast in a homologous recombination-based process referred to as plasmid gap repair (Benatuil et al., 2010). Diversity of amino acid distribution in the artificial CDR3 was verified by deep sequencing (Fig. S1). Library encoded vNARs were displayed on the yeast cell surface via fusion to cell wall protein Aga2p, which is anchored to the cell wall by association with Aga1p (Boder and Wittrup, 1997). The vNAR molecules were flanked by a HA-tag and cMyc-tag at the N-terminus and C-terminus, respectively, for detection of surface exposed vNAR (Fig. 2B). Surface presentation of vNAR variants was characterized by indirect fluorescence labeling of the cMyc-epitope and HA-tag (Fig. 2C and D). Within three days post-induction, there was nearly total HA-tag expression of the library detectable and nearly 90% of cMyc-labeling, both indicating expression and high copy number display of vNAR domains on the surface of S. cerevisiae. 3.2. Selection of antigen-binding semi-synthetic vNAR domains The semi-synthetic vNAR library was screened via YSD and FACS against three different targets, EpCAM, HTRA1, and EphA2, respectively. Initial FACS sorting rounds were performed with 1 ␮M target. Target-binding was detected by indirect immunofluorescence using either biotinylated target protein followed by cell staining with streptavidin–allophycocyanin conjugate or by using an Alexa Fluor 488 labeled anti-His-tag antibody. To avoid off-target binding against detection reagents the labeling strategy was alternated from Penta-His antibody immunostaining (for his-tagged EpCAM and EphA2) to biotinylated antigens after the first rounds of the particular screening experiments. In case of HTRA1, all experiments were executed using directly FITC-labeled protein. Within three rounds of FACS sorting we were able to enrich cells displaying an antigen-specific signal for each respective target (Figs. 3 and S2A and C). Single clones were analyzed for antigen-binding and FACSpositive clones were sequenced. Ten unique vNAR sequences were




5

Fig. 2. Library construction and analysis. (A) Schematic representation of PCR-based library design. PCR amplified vNAR fragments from blood circulating lymphocytes of a cohort of three bamboo sharks were used as template. In a first PCR the framework was amplified. The cysteine residue in CDR1 was replaced by tyrosine and a marginal diversity was introduced via the forward primer. In a subsequent reaction the total vNAR molecule was constructed and an artificial CDR3 was introduced. The 3rd reaction was carried out to incorporate homologous sites for gap repair cloning. (B) Illustration of vNAR yeast surface display (Doerner et al., 2014). vNARs are presented on the surface of S. cerevisiae as Aga2p fusions with an N-terminal HA-epitope and a C-terminal cMyc-tag, respectively, for the detection of surface expression. Histogram of cMyc (C) and HA-tag (D) surface expression of the constructed vNAR library assessed by indirect immunofluorescent labeling and flow cytometry. Grey: negative control (cells after 3 days induction of gene expression) labeled with detection reagents only; black: immunofluorescence staining after 3 days induction of gene expression.

identified for EpCAM-binding, four for HTRA1 as well as four for EphA2 target binding (Fig. S3). For the isolation of vNARs that bind both human and murine EphA2, which may be useful for animal studies, the population enriched for EphA2 binders (after sort 3) was subjected to two more rounds of selection via YSD using murine EphA2 which resulted in two vNARs with the desired promiscuous specificity. Affinity titration on cells revealed equilibrium binding constants for EpCAM- and EphA2-binding clones in the single-digit micromolar to sub-micromolar range, with an average affinity of 0.7 ␮M and 0.4 ␮M, respectively (Table 1, initial screen). However, for HTRA1 only variants with very low affinities were obtained.

3.3. Affinity maturation of target-enriched semi-synthetic vNAR domains The design of the sublibraries for affinity maturation was based on the enriched target-binding vNAR populations against EpCAM and HTRA1, respectively, that were obtained after three screening rounds. To this end, DNA was isolated from the enriched population and used as a template for the randomization of five residues in CDR1. Similar to randomization of CDR3 of the initial library, diversifications were performed by incorporation of triplet codons into the corresponding oligonucleotide (Table S1). To obtain targetbinders with higher affinity, the target protein concentration was


G Model

ARTICLE IN PRESS


S. Zielonka et al. / Journal of Biotechnology xxx (2014) xxx–xxx

6

Fig. 3. Library screening against EpCAM (upper row) and screening of an ␣-EpCAM sublibrary after randomization of CDR1 (lower row). Sorting gates and target concentrations are shown. Yeast cells were labeled for simultaneous detection of surface presentation and antigen-binding (in Maturation Round 3, double labeling was omitted since vNAR cMyc surface expression was >90%). After each round a resort was performed. Cells in the sorting gate were isolated, grown and induced for the next round of selection. After the initial screen (round 3) the enriched binders were used as template for affinity maturation. Percentage of cells in sorting gate: (R1) 0.17%; (R2) 1.03% and (R3) 8.04%. Target-positive cells after affinity maturation in sorting gate: (MR1) 0.1%; (MR2) 4.8% and (MR3) 47.2%.

Table 1 Equilibrium dissociation constants (KD ) determined by yeast surface display. vNAR clone

Target

Screen

Parental CDR3

KD (nM)

H2 H3 H5 H8 H10 1013 1014 1017 5005 R23 K3 K7 K9 K10 CK1 CK2 CK4 CK5 mmE1 mmE2

EpCAM EpCAM EpCAM EpCAM EpCAM EpCAM EpCAM EpCAM EpCAM EpCAM HTRA1 HTRA1 HTRA1 HTRA1 HTRA1 HTRA1 HTRA1 HTRA1 EphA2 EphA2

initial initial initial initial initial CDR1mat CDR1mat CDR1mat CDR1mat CDR1mat initial initial initial initial CDR1mat CDR1mat CDR1mat CDR1mat initial initial

– – – – – n.d. n.d. H3 H3 H3 – – – – K7 K7 n.d. K7 – –

741 ± 267 1301 ± 616 1030 ± 524 212 ± 161 301 ± 187 7±3 5±2 6±3 14 ± 7 24 ± 9 1300 ± 158 >10,000 6200 ± 5629 8300 ± 2513 3200 ± 889 550 ± 258 450 ± 241 290 ± 179 368 ± 144 482 ± 334

KD s from initial screening experiments (initial) and after affinity maturation (CDR1mat) are shown for different targets. Parental CDR3: CDR3 of binders after affinity maturation resulting from a clone of the initial library screen. n.d.: CDR3 was not found after sequencing of binders selected from initial screen. KD s for EphA2 binders were measured with (rh) EphA2.

stepwise reduced, aimed at enhancing selection stringency (Figs. 3 and S2B). To isolate a multitude of affinity-enhanced binders and to avoid out-competition of binders, the target concentration was increased to 50 nM for ␣-EpCAM CDR1 affinity maturation in the last screening round. Six different binders against EpCAM and four different binders against HTRA1 were identified (Fig. S3). Interestingly, EpCAM-binding clones 1017, 5005 and R23 contained the same CDR3 as EpCAM-binding clone H3 from initial screenings. However, the CDR1 were unique in all of those clones. In addition clones 1013, 1014 and R31 showed novel and unique CDR3 loops, which have not been identified after sequencing of binders from the initial library. The same findings were made for the HTRA1 sublibrary. Clones CK1, CK2 and CK5 showed an identical CDR3 as clone K7S from the initial screenings and clone CK4 showed a unique CDR3. Indeed, all clones with unique CDR3 loops after affinity maturation contained a unique CDR1, resulting as a consequence of the sublibrary design and high stringency screening. Five out of six clones against EpCAM and all four unique clones against HTRA1 after CDR1 affinity maturation were selected for yeast cell surface affinity titration (Table 1, CDR1mat). All EpCAM binders exhibited high-affinity binding in the low nanomolar range. The average affinity was 11.2 nM, which corresponds to an enhancement of affinity by the factor of 65. Clones CK2 and CK5 from the HTRA1 screen, which were progenies of clone K7S, showed affinities of 550 nM and 290 nM, respectively, compared to over


G Model

ARTICLE IN PRESS


S. Zielonka et al. / Journal of Biotechnology xxx (2014) xxx–xxx

7

Table 2 Binding kinetics and Tm of soluble vNAR variants selected against EpCAM. vNAR clone H3 H5 5005

Target EpCAM EpCAM EpCAM

Screen initial initial CDR1mat

Parental CDR3

Kon (M−1 s−1 )

– – H3

6.0 × 10 4.3 × 104 3.0 × 105 4

Koff (s−1 ) −2

8.9 × 10 1.0 × 10−1 1.1 × 10−2

KD (nM)

Tm (◦ C)

1495 2342 37

75.1 ± 0.5 71.5 ± 0.6 64.9 ± 0.6

Binding kinetics were measured on the Octet® RED96 System. Tm was determined via Thermal Shift assays.

10 ␮M prior to affinity maturation. These findings clearly highlight the contribution of CDR1 to antigen-binding. 3.4. Expression and characterization of EpCAM-binding vNAR domains Previous work of Wittrup and colleagues showed that the dissociation constant calculated by affinity titration on the yeast surface correlates well with measurements of soluble protein (Gai and Wittrup, 2007; Lipovsek et al., 2007). To examine if this correlation is also consistent for vNAR domains selected by yeast surface display, we chose EpCAM binders H3 and H5, obtained from initial screenings with micromolar affinities as well as nanomolar EpCAM binder 5005, identified after CDR1 affinity maturation, for protein production in E. coli (Fig. 4). All vNAR domains were expressed as MalE-fusion proteins in the pMX vector for periplasmic production using E. coli strain BMH 71-18 as host (Wentzel et al., 1999). After overnight production, the fusion proteins were first purified using a nickel affinity column followed by TEV-protease cleavage. The

monomeric vNAR domains were further purified and isolated by gel filtration chromatography. Binding kinetics were obtained through biolayer interferometry measurements on the FortéBio Octet RED96 with immobilized vNAR domains on the sensor tips and soluble EpCAM as antigen (Table 2; Fig. S4). Binder 5005 showed an approximately 40fold enhanced affinity compared to the parental vNAR H3. Both, improvements in on-rate kinetics (5-fold) as well as in off-rate kinetics (8.1-fold) contributed to the enhanced dissociation constant. Additionally, specificity for all three produced EpCAM binders was tested through biolayer interferometry measurements against murine EGFR. For all vNARs non-specific binding was not observed (data not shown). Specificity of all three human EpCAM-binding vNARs was more deeply scrutinized via binding assays on the yeast surface (Fig. S5). For all tested vNARs, no off-target binding was detected. However, all vNAR proteins displayed affinity to murine EpCAM, a feature that might be desirable for animal studies. To further analyze the produced proteins, the thermal stability was determined in a thermal shift assay (Table 2; Fig. S6). Tm

Fig. 4. Production of recombinant vNAR proteins. (A) Scheme for vNAR production as MalE-fusion protein. Ala(3): Triple alanine linker; His(6): hexahistidine-tag for protein purification; TEV site: site for specific cleavage by tabacco etch virus protease (TEV). IMAC purified protein was first digested via TEV protease and in a final step purified using GFC. (B) Elution profile of IMAC-purified TEV-cleaved vNAR 5005 protein on a Superdex 200 pg 16/60 equilibrated in PBS, pH 7.4 and run with a flow rate of 1 mL/min. Arrows indicate approximate molecular masses (kDa) of protein standards. The peak eluting at approx. 100 mL corresponds to vNAR and the peak eluting at ∼78 mL corresponds to MalE. (C) Purified vNAR proteins after size-exclusion chromatography.




8

values varied from approximately 65 ◦ C to over 75 ◦ C, indicating a high thermal stability, with binder 5005 possessing the lowest Tm of 64.9 ◦ C. However, its parental molecule, H3 showed the highest thermal stability (Tm = 75.1 ◦ C), indicating that in this particular case enhancement of affinity is at the cost of thermal stability.

4. Discussion In this present study we describe the successful development of an in vitro affinity maturation process for the generation of high-affinity vNAR domains using yeast surface display as platform technology obviating the need of animal immunization. The isolation of affinity-matured binders was performed in a twostep process. Screening from a semi-synthetic vNAR library in which CDR3 was totally randomized resulted in the selection of binders comprising moderate affinities to their target, that were improved by diversification of CDR1 and sublibrary screening. Previous experiments to isolate vNAR antibodies against EpCAM directly from the vNAR repertoire of non-immunized bamboo sharks using yeast surface display were unsuccessful (data not shown). Deep sequencing of bamboo shark vNARs confirmed that the primary vNAR diversity of this species is mainly restricted to CDR3, while the other regions (CDR1, HV2, HV4) demonstrate only minor variations (Diaz et al., 1998). Expanding the set of variants by generating a semi-synthetic library of 2 × 108 variants with fully randomized CDR3 proved to be a valid strategy to obtain antigen-binding vNAR molecules. Interestingly, in the original vNAR repertoire one cysteine residue in CDR1 was conserved that presumably forms a disulfide bond with a Cys residue in CDR3 (Fig. 1C). In sharks, three different types of vNARs have been characterized based on the position of non-canonical Cys residues (Barelle et al., 2009). An additional type of IgNAR, termed type IV, possesses only one canonical disulfide bond. Except IgNAR type III, which has limited sequence diversity in CDR3, all types give rise to antibodies with high affinities (Dooley et al., 2003; Liu et al., 2007a; Muller et al., 2012; Nuttall et al., 2001, 2002; Ohtani et al., 2013). The semisynthetic library is cysteine-free and also has Cys in CDR1 replaced by Tyr. Therefore it may be possible that the enhanced conformational flexibility of CDR3 compared to the disulfide-bond constrained natural loop contributes to accommodation of targetprotein interaction and allows for the successful isolation of a large set of different vNARs, albeit with low affinity to their target. It will be interesting to analyze the screening outcome from a vNAR library of similar diversity that contains Cys in the randomized CDR3 loop and a potentially disulfide bond-forming Cys in CDR1 with respect to clonal diversity of isolated binders and their target affinities. Semi-synthetic Cys-biased CDR3 randomized Type II libraries were constructed by Nuttall and collegues who isolated vNAR molecules specifically targeting Gingipain K protease from Porphyromonas ginglivalis (Nuttall et al., 2001). In another approach, the same group established a dramatically expanded library by the incorporation of both synthetic and naturally occurring CDR3 sequences (Nuttall et al., 2002). From this repertoire, Gingipain protease specific vNAR molecules were obtained with the best binder having a KD in the range of 130 nM. On the other hand, Goldman and co-workers were also able to isolate high-affinity binders from a semi-synthetic CDR3 mutagenized vNAR library against different antigens (Liu et al., 2007b) that were lacking the inter-loop disulfide bond. In conclusion, the inter-loop disulfide bond between CDR1 and CDR3, if present likely contributes to binding affinity and structural stability but is not absolutely required to obtain high affinity vNARs. Stepwise affinity optimization by first randomizing CDR3 and accumulating low affinity binders followed by diversification of CDR1 in this population of enriched variants proved a useful

strategy to obtain vNARs with affinities in the low nanomolar range. This in vitro affinity maturation method resembles the natural strategy of the immune system of nurse sharks (Diaz et al., 1998) to select clones from a primary IgNAR repertoire that is nearly entirely CDR3 based, followed by affinity maturation of complementary-determining regions and hypervariable loops after antigen exposure, with a significant bias towards replacement substitutions in CDR1. The stepwise screening strategy is similar to the isolation of camelized human VH domains by phage display, where low affinity binders were selected from a CDR3 randomized VH domain. Subsequent diversification of CDR1 or CDR2 residues provoked a significant increase in affinity (Davies and Riechmann, 1995, 1996). While in these studies a single VH scaffold was chosen for CDR randomization, we used a repertoire of natural scaffold molecules from non-immunized animals. This may contribute to the successful isolation of a large set of stable binders since small variations in the scaffold sequence are known to have a large impact on folding stability and protein solubility (Ewert et al., 2003a; Ewert et al., 2003b). Although affinity matured EpCAM-binders selected with this method comprised low nanomolar affinities similar to those obtained from nurse sharks immunized with hen egg lysozyme (Dooley et al., 2003), none of the selected binders showed subnanomolar affinities, that were recently obtained by immunization of S. acanthias with human serum albumin (Muller et al., 2012). Nevertheless, subsequent third-round randomization of HV2 or HV4, respectively, that potentially contribute to the paratope may result in variants with further enhanced affinities. Since the development of yeast surface display, pioneered by Boder and Wittrup in 1997, it was extensively shown that this cellular display technology can be successfully utilized to engineer antibody-like molecules and also a wide range of alternative binding scaffolds (Boder and Wittrup, 1997; Feldhaus et al., 2003; Gai and Wittrup, 2007; Glotzbach et al., 2013; Stone et al., 2012; Walker et al., 2009; Wozniak-Knopp et al., 2010). One of the benefits of this display method is control over the selection process, as it harbors the potential of single cell real-time and on-line analysis and subsequent characterization of individual library candidates. Another beneficial attribute is the possibility to co-select for high-level expression and stability simultaneously to the binding functionality (Doerner et al., 2014; Gai and Wittrup, 2007). Moreover, proteins are believed to be most likely well folded since quality control machineries exist in S. cerevisiae for proper protein folding (Ellgaard et al., 1999). In addition, isolated binders can be instantaneously characterized in terms of affinity and specificity without the need for soluble expression (Gai and Wittrup, 2007; VanAntwerp and Wittrup, 2000). Finally, sampling of libraries supposedly is more comprehensive with yeast display than with phage display (Bowley et al., 2007). This might be particularly important for the selection of rare binders from large library e.g. naïve or artificial repertoires. This study adds vNAR antibody domains to the growing list of scaffold molecules that can be easily isolated and characterized from large libraries using this powerful display technology. Competing interest The authors declare no competing interests. Author contributions HK, SZ, AD and SB directed the project. AC, CC and JF performed deep sequencing and data analysis. NW and SZ performed FACS experiments and analyzed data. BS, ME, ML, and PO established the protocols for animal handling. AD, CU, ES, NW, and SZ developed the purification protocol, biolayer interferometry measurements




and thermal shift assays. ME performed vNAR structure modeling. Reagents, materials and analysis tools were contributed by AC, AD, BS, HK, JF, ME, ML, NW, PO, SB, and SZ. HK and SZ drafted the manuscript. Acknowledgments This paper is dedicated to Karl-Erich Jaeger on the occasion of his 60th birthday. This work was partially funded by a Ph.D. fellowship of the Merck’sche Gesellschaft für Kunst und Wissenschaft to S.Z. The project was supported in part by Federal Ministry of Education and Research (BMBF) in frame of the consortium Nanokat. We thank Siegfried Neumann for general advice and support, Johann Kirchhauser of the Staatliches Museum für Naturkunde, Karlsruhe for providing the sharks, Rolf Landvogt at Aqua Natura, Leun and Joachim Nilz for taking care of the animals. Moreover, we are grateful to Stewart D. Nuttall and Michael Ehrmann for kind advice concerning the primer design and for delivering the vector pET21d-HTRA1prot, respectively. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jbiotec. 2014.04.023. References Barelle, C., Gill, D.S., Charlton, K., 2009. Shark novel antigen receptors—the next generation of biologic therapeutics? Adv. Exp. Med. Biol. 655, 49–62. Benatuil, L., Perez, J.M., Belk, J., Hsieh, C.M., 2010. An improved yeast transformation method for the generation of very large human antibody libraries. Protein Eng. Des. Sel. 23, 155–159. Boder, E.T., Wittrup, K.D., 1997. Yeast surface display for screening combinatorial polypeptide libraries. Nat. Biotechnol. 15, 553–557. 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 antibodies compared to selection from the same library displayed on phage. Protein Eng. Des. Sel. 20, 81–90. Buss, N.A., Henderson, S.J., McFarlane, M., Shenton, J.M., de Haan, L., 2012. Monoclonal antibody therapeutics: history and future. Curr. Opin. Pharmacol. 12, 615–622. Chao, G., Lau, W.L., Hackel, B.J., Sazinsky, S.L., Lippow, S.M., Wittrup, K.D., 2006. Isolating and engineering human antibodies using yeast surface display. Nat. Protoc. 1, 755–768. Crouch, K., Smith, L.E., Williams, R., Cao, W., Lee, M., Jensen, A., Dooley, H., 2013. Humoral immune response of the small-spotted catshark, Scyliorhinus canicula. Fish Shellfish Immunol. 34, 1158–1169. Davies, J., Riechmann, L., 1995. Antibody VH domains as small recognition units. Biotechnology (N Y) 13, 475–479. Davies, J., Riechmann, L., 1996. Affinity improvement of single antibody VH domains: residues in all three hypervariable regions affect antigen binding. Immunotechnology 2, 169–179. Desmyter, A., Transue, T.R., Ghahroudi, M.A., Thi, M.H., Poortmans, F., Hamers, R., Muyldermans, S., Wyns, L., 1996. Crystal structure of a camel single-domain VH antibody fragment in complex with lysozyme. Nat. Struct. Biol. 3, 803–811. Diaz, M., Greenberg, A.S., Flajnik, M.F., 1998. Somatic hypermutation of the new antigen receptor gene (NAR) in the nurse shark does not generate the repertoire: possible role in antigen-driven reactions in the absence of germinal centers. Proc. Natl. Acad. Sci. U.S.A. 95, 14343–14348. Diaz, M., Velez, J., Singh, M., Cerny, J., Flajnik, M.F., 1999. Mutational pattern of the nurse shark antigen receptor gene (NAR) is similar to that of mammalian Ig genes and to spontaneous mutations in evolution: the translesion synthesis model of somatic hypermutation. Int. Immunol. 11, 825–833. Doerner, A., Rhiel, L., Zielonka, S., Kolmar, H., 2014. Therapeutic antibody engineering by high efficiency cell screening. FEBS Lett. 588, 278–287. Dooley, H., Flajnik, M.F., 2005. Shark immunity bites back: affinity maturation and memory response in the nurse shark, Ginglymostoma cirratum. Eur. J. Immunol. 35, 936–945. Dooley, H., Flajnik, M.F., 2006. Antibody repertoire development in cartilaginous fish. Dev. Comp. Immunol. 30, 43–56. Dooley, H., Flajnik, M.F., Porter, A.J., 2003. Selection and characterization of naturally occurring single-domain (IgNAR) antibody fragments from immunized sharks by phage display. Mol. Immunol. 40, 25–33. Duan, Y., Wu, C., Chowdhury, S., Lee, M.C., Xiong, G., Zhang, W., Yang, R., Cieplak, P., Luo, R., Lee, T., Caldwell, J., Wang, J., Kollman, P., 2003. A point-charge force field

9

for molecular mechanics simulations of proteins based on condensed-phase quantum mechanical calculations. J. Comput. Chem. 24, 1999–2012. Dudgeon, K., Rouet, R., Famm, K., Christ, D., 2012. Selection of human VH single domains with improved biophysical properties by phage display. Methods Mol. Biol. 911, 383–397. Ellgaard, L., Molinari, M., Helenius, A., 1999. Setting the standards: quality control in the secretory pathway. Science 286, 1882–1888. Ewert, S., Honegger, A., Pluckthun, A., 2003a. Structure-based improvement of the biophysical properties of immunoglobulin VH domains with a generalizable approach. Biochemistry 42, 1517–1528. Ewert, S., Huber, T., Honegger, A., Pluckthun, A., 2003b. Biophysical properties of human antibody variable domains. J. Mol. Biol. 325, 531–553. Feldhaus, M.J., Siegel, R.W., Opresko, L.K., Coleman, J.R., Feldhaus, J.M., Yeung, Y.A., Cochran, J.R., Heinzelman, P., Colby, D., Swers, J., Graff, C., Wiley, H.S., Wittrup, K.D., 2003. Flow-cytometric isolation of human antibodies from a nonimmune Saccharomyces cerevisiae surface display library. Nat. Biotechnol. 21, 163–170. Gai, S.A., Wittrup, K.D., 2007. Yeast surface display for protein engineering and characterization. Curr. Opin. Struct. Biol. 17, 467–473. Glotzbach, B., Reinwarth, M., Weber, N., Fabritz, S., Tomaszowski, M., Fittler, H., Christmann, A., Avrutina, O., Kolmar, H., 2013. Combinatorial optimization of cystine-knot peptides towards high-affinity inhibitors of human matriptase-1. PLoS ONE 8, e76956. Goodchild, S.A., Dooley, H., Schoepp, R.J., Flajnik, M., Lonsdale, S.G., 2011. Isolation and characterisation of Ebolavirus-specific recombinant antibody fragments from murine and shark immune libraries. Mol. Immunol. 48, 2027–2037. Greenberg, A.S., Avila, D., Hughes, M., Hughes, A., McKinney, E.C., Flajnik, M.F., 1995. A new antigen receptor gene family that undergoes rearrangement and extensive somatic diversification in sharks. Nature 374, 168–173. Hamers-Casterman, C., Atarhouch, T., Muyldermans, S., Robinson, G., Hamers, C., Songa, E.B., Bendahman, N., Hamers, R., 1993. Naturally occurring antibodies devoid of light chains. Nature 363, 446–448. Holt, L.J., Herring, C., Jespers, L.S., Woolven, B.P., Tomlinson, I.M., 2003. Domain antibodies: proteins for therapy. Trends Biotechnol. 21, 484–490. Imrich, S., Hachmeister, M., Gires, O., 2012. EpCAM and its potential role in tumorinitiating cells. Cell Adh. Migr. 6, 30–38. Kinch, M.S., Carles-Kinch, K., 2003. Overexpression and functional alterations of the EphA2 tyrosine kinase in cancer. Clin. Exp. Metastasis 20, 59–68. Kopsidas, G., Roberts, A.S., Coia, G., Streltsov, V.A., Nuttall, S.D., 2006. In vitro improvement of a shark IgNAR antibody by Qbeta replicase mutation and ribosome display mimics in vivo affinity maturation. Immunol. Lett. 107, 163–168. Kovalenko, O.V., Olland, A., Piche-Nicholas, N., Godbole, A., King, D., Svenson, K., Calabro, V., Muller, M.R., Barelle, C.J., Somers, W., Gill, D.S., Mosyak, L., Tchistiakova, L., 2013. Atypical antigen recognition mode of a shark immunoglobulin new antigen receptor (IgNAR) variable domain characterized by humanization and structural analysis. J. Biol. Chem. 288, 17408–17419. Lipovsek, D., Lippow, S.M., Hackel, B.J., Gregson, M.W., Cheng, P., Kapila, A., Wittrup, K.D., 2007. Evolution of an interloop disulfide bond in high-affinity antibody mimics based on fibronectin type III domain and selected by yeast surface display: molecular convergence with single-domain camelid and shark antibodies. J. Mol. Biol. 368, 1024–1041. Liu, J.L., Anderson, G.P., Delehanty, J.B., Baumann, R., Hayhurst, A., Goldman, E.R., 2007a. Selection of cholera toxin specific IgNAR single-domain antibodies from a naive shark library. Mol. Immunol. 44, 1775–1783. Liu, J.L., Anderson, G.P., Goldman, E.R., 2007. Isolation of anti-toxin single domain antibodies from a semi-synthetic spiny dogfish shark display library. BMC Biotechnol. 7, 78. Minsky, A., Summers, R.G., Knowles, J.R., 1986. Secretion of beta-lactamase into the periplasm of Escherichia coli: evidence for a distinct release step associated with a conformational change. Proc. Natl. Acad. Sci. U.S.A. 83, 4180–4184. Mordenti, J., Cuthbertson, R.A., Ferrara, N., Thomsen, K., Berleau, L., Licko, V., Allen, P.C., Valverde, C.R., Meng, Y.G., Fei, D.T., Fourre, K.M., Ryan, A.M., 1999. Comparisons of the intraocular tissue distribution, pharmacokinetics, and safety of 125 I-labeled full-length and Fab antibodies in rhesus monkeys following intravitreal administration. Toxicol. Pathol. 27, 536–544. Muller, M.R., Saunders, K., Grace, C., Jin, M., Piche-Nicholas, N., Steven, J., O’Dwyer, R., Wu, L., Khetemenee, L., Vugmeyster, Y., Hickling, T.P., Tchistiakova, L., Olland, S., Gill, D., Jensen, A., Barelle, C.J., 2012. Improving the pharmacokinetic properties of biologics by fusion to an anti-HSA shark VNAR domain. MAbs 4, 673–685. Muyldermans, S., 2001. Single domain camel antibodies: current status. J. Biotechnol. 74, 277–302. Nelson, A.L., 2010. Antibody fragments: hope and hype. MAbs 2, 77–83. Ni, J., Cozzi, P.J., Duan, W., Shigdar, S., Graham, P.H., John, K.H., Li, Y., 2012. Role of the EpCAM (CD326) in prostate cancer metastasis and progression. Cancer Metastasis Rev. 31, 779–791. Nuttall, S.D., 2012. Overview and discovery of IgNARs and generation of VNARs. Methods Mol. Biol. 911, 27–36. Nuttall, S.D., Humberstone, K.S., Krishnan, U.V., Carmichael, J.A., Doughty, L., Hattarki, M., Coley, A.M., Casey, J.L., Anders, R.F., Foley, M., Irving, R.A., Hudson, P.J., 2004. Selection and affinity maturation of IgNAR variable domains targeting Plasmodium falciparum AMA1. Proteins 55, 187–197. Nuttall, S.D., Krishnan, U.V., Doughty, L., Nathanielsz, A., Ally, N., Pike, R.N., Hudson, P.J., Kortt, A.A., Irving, R.A., 2002. A naturally occurring NAR variable domain binds the Kgp protease from Porphyromonas gingivalis. FEBS Lett. 516, 80–86. Nuttall, S.D., Krishnan, U.V., Doughty, L., Pearson, K., Ryan, M.T., Hoogenraad, N.J., Hattarki, M., Carmichael, J.A., Irving, R.A., Hudson, P.J., 2003. Isolation and


G Model BIOTEC-6679; No. of Pages 10 10


characterization of an IgNAR variable domain specific for the human mitochondrial translocase receptor Tom70. Eur. J. Biochem. 270, 3543–3554. Nuttall, S.D., Krishnan, U.V., Hattarki, M., De Gori, R., Irving, R.A., Hudson, P.J., 2001. Isolation of the new antigen receptor from wobbegong sharks, and use as a scaffold for the display of protein loop libraries. Mol. Immunol. 38, 313–326. Ohtani, M., Hikima, J., Jung, T.S., Kondo, H., Hirono, I., Aoki, T., 2013. Construction of an artificially randomized IgNAR phage display library: screening of variable regions that bind to hen egg white lysozyme. Mar. Biotechnol. (NY) 15, 56–62. Reichert, J.M., 2012. Marketed therapeutic antibodies compendium. MAbs 4, 413–415. Reichert, J.M., 2013. Which are the antibodies to watch in 2013? MAbs 5, 1–4. Roux, K.H., Greenberg, A.S., Greene, L., Strelets, L., Avila, D., McKinney, E.C., Flajnik, M.F., 1998. Structural analysis of the nurse shark (new) antigen receptor (NAR): molecular convergence of NAR and unusual mammalian immunoglobulins. Proc. Natl. Acad. Sci. U.S.A. 95, 11804–11809. Shao, C.Y., Secombes, C.J., Porter, A.J., 2007. Rapid isolation of IgNAR variable singledomain antibody fragments from a shark synthetic library. Mol. Immunol. 44, 656–665. Stanfield, R.L., Dooley, H., Flajnik, M.F., Wilson, I.A., 2004. Crystal structure of a shark single-domain antibody V region in complex with lysozyme. Science 305, 1770–1773. Stanfield, R.L., Dooley, H., Verdino, P., Flajnik, M.F., Wilson, I.A., 2007. Maturation of shark single-domain (IgNAR) antibodies: evidence for induced-fit binding. J. Mol. Biol. 367, 358–372.

Stone, J.D., Chervin, A.S., Aggen, D.H., Kranz, D.M., 2012. T cell receptor engineering. Methods Enzymol. 503, 189–222. Truebestein, L., Tennstaedt, A., Monig, T., Krojer, T., Canellas, F., Kaiser, M., Clausen, T., Ehrmann, M., 2011. Substrate-induced remodeling of the active site regulates human HTRA1 activity. Nat. Struct. Mol. Biol. 18, 386–388. van der Gun, B.T., Melchers, L.J., Ruiters, M.H., de Leij, L.F., McLaughlin, P.M., Rots, M.G., 2010. EpCAM in carcinogenesis: the good, the bad or the ugly. Carcinogenesis 31, 1913–1921. VanAntwerp, J.J., Wittrup, K.D., 2000. Fine affinity discrimination by yeast surface display and flow cytometry. Biotechnol. Prog. 16, 31–37. Walker-Daniels, J., Hess, A.R., Hendrix, M.J., Kinch, M.S., 2003. Differential regulation of EphA2 in normal and malignant cells. Am. J. Pathol. 162, 1037–1042. Walker, L.M., Bowley, D.R., Burton, D.R., 2009. Efficient recovery of high-affinity antibodies from a single-chain Fab yeast display library. J. Mol. Biol. 389, 365–375. Wentzel, A., Christmann, A., Kratzner, R., Kolmar, H., 1999. Sequence requirements of the GPNG beta-turn of the Ecballium elaterium trypsin inhibitor II explored by combinatorial library screening. J. Biol. Chem. 274, 21037–21043. Wozniak-Knopp, G., Bartl, S., Bauer, A., Mostageer, M., Woisetschlager, M., Antes, B., Ettl, K., Kainer, M., Weberhofer, G., Wiederkum, S., Himmler, G., Mudde, G.C., Ruker, F., 2010. Introducing antigen-binding sites in structural loops of immunoglobulin constant domains: Fc fragments with engineered HER2/neu-binding sites and antibody properties. Protein Eng. Des. Sel. 23, 289–297.


Selection of phage antibodies by binding affinity. Mimicking affinity maturation.

Shark attack: the emergency presentation and management.

Selecting high-affinity binding proteins by monovalent phage display.

Therapeutic Potential of Shark Anti-ICOSL VNAR Domains is Exemplified in a Murine Model of Autoimmune Non-Infectious Uveitis.

Antibody engineering--successful affinity maturation in vitro.

Shark attacks and shark diving.

Membrane attack complex of complement: generation of high-affinity phospholipid binding sites by fusion of five hydrophilic plasma proteins.

High-affinity androgen binding proteins in syringeal tissues of songbirds.

Engineering of bispecific affinity proteins with high affinity for ERBB2 and adaptable binding to albumin.

Affinity labeling of binding sites in proteins by sensitized photooxidation.

Fatal tiger shark, Galeocerdo cuvier attack in New Caledonia erroneously ascribed to great white shark, Carcharodon carcharias.

High-affinity binding of water by proteins is similar in air and in organic solvents.

Affinity probes for the avermectin binding proteins.

Affinity maturation of antibodies.

In vitro evolution and affinity-maturation with Coliphage qβ display.

High-affinity phlorizin binding in Mytilus gill.

Nuclear binding of progesterone in hen oviduct. Role of acidic chromatin proteins in high-affinity binding.

Methylmercury in dried shark fins and shark fin soup from American restaurants.

Determinants of high-affinity DNA binding by the glucocorticoid receptor: evaluation of receptor domains outside the DNA-binding domain.

Cucurbiturils: from synthesis to high-affinity binding and catalysis.

Cytosolic delivery of siRNA by ultra-high affinity dsRNA binding proteins.

High affinity binding of adrenocortical and gonadal steroids by plasma proteins of Australian marsupials.

Affinity purification and affinity characterization of carbohydrate-binding proteins in bovine kidney.

High Affinity Binding of Indium and Ruthenium Ions by Gastrins.