Single-domain antibodies against SARS-CoV-2 RBD from a two-stage phage screening of universal and focused synthetic libraries

BACKGROUND

Coronavirus disease 2019 (COVID-19) is an evolving global pandemic, and nanobodies, as well as other single-domain antibodies (sdAbs), have been recognized as a potential diagnostic and therapeutic tool for infectious diseases. High-throughput screening techniques such as phage display have been developed as an alternative to in vivo immunization for the discovery of antibody-like target-specific binders.

METHODS

We designed and constructed a highly diverse synthetic phage library sdAb-U (single-domain Antibody - Universal library ) based on a human framework. The SARS-CoV-2 receptor-binding domain (RBD) was expressed and purified. The universal library sdAb-U was panned against the RBD protein target for two rounds, followed by monoclonal phage ELISA (enzyme-linked immunosorbent assay) to identify RBD-specific binders (the first stage). High-affinity binders were sequenced and the obtained CDR1 and CDR2 sequences were combined with fully randomized CDR3 to construct a targeted (focused) phage library sdAb-RBD, for subsequent second-stage phage panning (also two rounds) and screening. Then, sequences with high single-to-background ratios in phage ELISA were selected for expression. The binding affinities of sdAbs to RBD were measured by an ELISA-based method. In addition, we conducted competition ELISA (using ACE2 ectodomain S19-D615) and SARS-CoV-2 pseudovirus neutralization assays for the high-affinity RBD-binding sdAb39.

RESULTS

Significant enrichments were observed in both the first-stage (universal library) and the second-stage (focused library) phage panning. Five RBD-specific binders were identified in the first stage with high ELISA signal-to-background ratios. In the second stage, we observed a much higher possibility of finding RBD-specific clones in phage ELISA. Among 45 selected RBD-positive sequences, we found eight sdAbs can be well expressed, and five of them show high-affinity to RBD (EC50 < 100nM). We finally found that sdAb39 (EC50 ~ 4nM) can compete with ACE2 for binding to RBD.

CONCLUSION

Overall, this two-stage strategy of synthetic phage display libraries enables rapid selection of SARS-CoV-2 RBD sdAb with potential therapeutic activity, and this two-stage strategy can potentially be used for rapid discovery of sdAbs against other targets.

Antibody Isolation from Human Synthetic Libraries of Single-Chain Antibodies and Analysis Using NGS

Antibody libraries came into existence 30 years ago when the accumulating sequence data of immunoglobulin genes and the advent of PCR technology made it possible to clone antibody gene repertoires. Phage display (most common) and additional display and screening technologies were applied to pan out desired binding specificities from antibody libraries. As other antibody discovery tools, phage display is not an off-the-shelf technology and not offered as a kit but rather requires experience and expertise for making it indeed very useful.Next-generation sequencing (NGS) coupled with bioinformatics is a powerful tool for analyzing large amount of DNA sequence output of the panning. Here, we demonstrate how NGS analysis of phage biopanning (phage-Seq) of complex antibody libraries can facilitate the antibody discovery process and provide insights regarding the biopanning process (see Fig. 1).

Construction of Synthetic VHH Libraries in Ribosome Display Format

Single-domain antibodies, or VHH, represent an attractive molecular basis to design affinity proteins with favorable properties. Beyond high affinity and specificity for their cognate target, they usually show high stability and high production yields in bacteria, yeast, or mammalian cells. In addition to these favorable properties, their ease of engineering makes them useful for many applications. Until the past few years, the generation of VHH involved the immunization of a Camelidae with the target antigen, followed by a phage display selection using phage libraries encoding the VHH repertoire of the animal blood sample. However, this approach is constrained by the accessibility to the animals, and the output relies on the animal's immune system.Recently, synthetic VHH libraries have been designed to avoid the use of animals. Here, we describe the construction of VHH combinatorial libraries and their use for the selection of binders by ribosome display, a fully in vitro selection technique.

Preparation of Immune and Synthetic VNAR Libraries as Sources of High-Affinity Binders

The shark-derived autonomous variable antibody domains known as VNARs are attractive tools for therapeutic and diagnostic applications due to their favorable properties like small size (approximately 12 kDa), high thermal and chemical stability, and good tissue penetration. Currently, different techniques have been reported to generate VNAR domains against targets of therapeutic interest. Here, we describe methods for the preparation of an immune VNAR library based on bacteriophage display, and for the preparation of a synthetic library of VNAR domains using a modified protocol based on Kunkel mutagenesis. Finally, we describe procedures for in silico maturation of a VNAR using a bioinformatic approach to obtain higher affinity binders.

A Comparison Between Cell, Protein and Peptide-Based Approaches for Selection of Nanobodies Against CD44 from a Synthetic Library

BACKGROUND

The hyaluronic acid receptor CD44, is a cancer stem cell biomarker, playing important roles in cell adhesion, tumor progression and drug-resistance. Therefore, CD44 is a potential target for cancer treatment and its blockade could result in multi-factorial therapeutic effects.

METHODS

Nanobodies against CD44 were isolated from a synthetic library with a diversity of 5×1011 CFU/ml using the phage display technique. Three approaches were used for isolation of nanobodies fragments including peptide-, protein- and cell-based panning.

RESULTS

Nanobodies from cell-based panning displayed more specificity compared to protein or peptide-based panning. Our results show that cell-based panning is the most efficient method for isolation of a specific single domain antibody fragment to CD44 from a synthetic phage displayed library.

CONCLUSION

The isolated nanobodies could successfully recognize and bind cells that express the CD44 surface antigen.

Fc Engineering: Tailored Synthetic Human IgG1-Fc Repertoire for High-Affinity Interaction with FcRn at pH 6.0

The therapeutic efficacy of an antibody drug depends on the variable domains and on the constant crystallizable fragment (Fc). IgG variable domains have been the targets of extensive molecular engineering in search of more specific binders with higher affinities for their targets. Similarly, Fc engineering approaches have led to modulating both the immune effector responses and serum half-lives of therapeutic antibodies. A high-affinity interaction between the IgG Fc and neonatal Fc receptor (FcRn) at a slightly acidic pH can protect IgG molecules from undergoing lysosomal or serum proteinase-induced degradation. Here we describe an optimized protocol for the development of a tailored, synthetic human Fc repertoire to select Fc mutants which show highly pH-restricted FcRn binding with high affinity.

Antibody Isolation From a Human Synthetic Combinatorial and Other Libraries of Single-Chain Antibodies

Antibody libraries came into existence 25 years ago when the accumulating sequence data of immunoglobulin genes and the advent of the PCR technology made it possible to clone antibody gene repertoires. Phage display (most common) and additional display and screening technologies were applied to pan out desired binding specificities from antibody libraries. "Synthetic" or "semisynthetic" libraries are from naive-non-immunized source and considered to be a source for many different targets, including self-antigens.As other antibody discovery tools, phage display is not an off-the-shelf technology and not offered as a kit but rather requires experience and expertise for making it indeed very useful. Here we present application notes that expand the usefulness of antibody phage display as a very versatile and robust antibody discovery tool.

Construction of a Synthetic Antibody Gene Library for the Selection of Intrabodies and Antibodies

Libraries of antibody fragments displayed on filamentous phages have proved their value to generate human antibodies against virtually any target. We describe here a simple protocol to make large and diverse libraries based on a single or a limited number of frameworks. The approach is flexible enough to be used with any antibody format, either single-chain (scFv, VHH) or multi-chain (Fv, Fab, (Fab')2), and to target in a single step the six complementarity-determining regions-or any other part-of the antibody molecule. Using this protocol, libraries larger than 10¹⁰ can be easily constructed in a single week.

Generation of Highly Selective MMP Antibody Inhibitors

Inhibiting individual MMPs of biomedical importance with high selectivity is critical for both fundamental research and therapy development. Here we describe the methods for discovery of inhibitory monoclonal antibodies from synthetic human antibody phage display libraries carrying convex paratopes encoded by long complementarity-determining region (CDR)-H3 segments. We demonstrate the application of this technique for isolation of highly specific and potent antibody inhibitors of human MMP-14.

Construction of Synthetic Antibody Libraries

Libraries of antibody fragments displayed on filamentous phages are now a widely used approach to isolate antibodies against virtually any target. We describe a simple protocol to make large and diverse libraries based on a single or a limited number of frameworks. The approach is flexible enough to be used with any antibody format, either single-chain (scFv, VHH) or multi-chain (Fv, Fab, (Fab')2), and to target in a single step the six complementarity-determining regions-or any other part-of the antibody molecule. Using this protocol, libraries larger than 10¹⁰ can be constructed in a single week.

Ligation-based assembly for constructing mouse synthetic scFv libraries by chain shuffling with in vivo-amplified VH and VL fragments

In vitro assembly of two or three PCR fragments using primers is a common method of constructing scFv fragments for display on the surface of phage. However, mismatch annealing often occurs during in this step, leading to cloning and display of incomplete Fab or scFv fragments. To overcome this limitation, we developed a ligation-based two-fragment assembly (LTFA) protocol that involved separately cloning VH and Vκ fragments into the high-copy-number plasmid pUC18. The VH and Vκ fragments had randomized complementarity-determining region 3 (CDR3) and were joined with a peptidyl linker composed of (G4S)3. Using this approach, complete sequences of scFv fragments were successfully constructed, and the sequencing of 83 scFv clones revealed that none of the sequences, including the linker region, contained deletions or mutations. In contrast, linker sequences generated using a conventional two-fragment PCR assembly (TFPA) protocol often contained sequence anomalies, including large truncations. Using the LTFA protocol, a final library size of 1.0×10(8)cfu was achieved. Examination of the amino acid profiles of the generated scFv fragments within the randomized regions introduced using degenerate codons did not detect any bias from that expected based on stochastic distribution. After several cycles of panning with this library, antigen-specific scFvs against two reference antigens, hen egg lysozyme and streptavidin were detected. In addition, scFvs with specificity against peptidyl antigens in the loop region of the Medaka ortholog of human C6orf89, which encodes a histone deacetylase enhancer that interacts with the bombesin receptor, were also obtained. The LTFA protocol developed here is robust and allows for the easy construction of integral scFv fragments compared with conventional TFPA. Utilizing LTFA, other CDRs can be readily combined. This approach also allows for the in vitro maturation of scFv fragments by separately introducing randomization in CDRs or using error-prone PCR for the amplification of pre-selected sequences as a template scaffold.