Applications of Yeast Surface Display Technology for Development of Therapeutic Proteins and Peptides

Abstract

The yeast surface display technology have evolved into a versatile tool for research and industrial applications. The works presented in this thesis aimed at the application of yeast surface display technology for developing therapeutic proteins and peptides and engineering their affinity, specificity and stability. I applied the yeast surface display technology to construction of fully synthetic human Fab libraries with large diversity for rapid isolation of human antibody candidates against given targets, humanization of phosphorylated epitope specific non-human antibody and development of novel neuropilin-1 binding peptides. Firstly, I report a large synthetic human Fab yeast surface-displayed library generated by stepwise optimization of yeast mating conditions. Yeast surface-displayed antibody libraries provide an efficient and quantitative screening resource for given antigens, but suffer from typically modest library sizes owing to low yeast transformation efficiency. Yeast mating is an attractive method for overcoming the limit of yeast transformation to construct a large, combinatorial antibody library, but the optimal conditions have not been reported. Yeast mating conditions were optimized in the order of cell density, media pH, and cell growth phase, yielding a mating efficiency of ~58% between the two haploid cells carrying HC and LC libraries. I constructed two combinatorial Fab libraries with VH-CDR3 of 9 or 11 residues in length with colony diversities of more than 109 by one round of yeast mating between the two haploid HC and LC libraries, with modest diversity sizes of ~107. The synthetic human Fab yeast-displayed libraries exhibited relative amino acid compositions in each position of the six CDRs that were very similar to those of the designed repertoires, suggesting that they are extremely functional in isolating human Fab antibodies with improved biophysical properties and reduced immunogenicity for given antigens. Secondly, I report humanization of a chicken pT231 antibody specific to a tau protein-derived peptide carrying the phosphorylated threonine at 231 residue (pT231-peptide) as a model for better understanding of the phosphoepitope recognition mechanism. There have been no humanization reports of phosphospecific non-human antibodies until now. By using yeast surface-displayed combinatorial library with permutations of 11 FR residues potentially affecting CDR loop conformations, I efficiently identified critical 5 FR residues, the back mutation of which with the corresponding chicken residues completely recovered the pT231-peptide binding affinity and specificity of humanized antibody, while simple grafting six CDRs of the chicken antibody into a homologous human framework (FR) template resulted in complete loss of the pT231-peptide binding. Importantly back mutation of FR 76 residue of VH (H76) (Asn to Ser) was critical in preserving the pT231-binding motif conformation via allosteric regulation of ArgH71, which closely interacts with ThrH52 and SerH52a residues on VH-CDR2 to induce the unique phosphate-binding bowl-like conformation. My humanization approach of CDR grafting plus permutations of FR residues by combinatorial library screening can be applied for non-human antibodies containing unique binding motif on CDRs specific to post-translationally modified epitopes. The successfully humanized antibodies in this study could be used for diagnostic and/or therapeutic purposes associated with the pT231 tau protein, and also served as a valuable template for generating pThr-peptide focused human antibody library to bypass animal immunization. Lastly, I report immunoglobulin Fc-fused NRP1-specific peptides deviating from CendR. VEGF and Sema3 family ligands specifically bind to the arginine-binding pocket in the b1 domain of NRP1 through a C-terminal (R/K)XX(R/K) amino acid sequence motif, where X stands for any amino acid and these interactions trigger NRP1-mediated vascular permeability. In fact, all known proteins and peptides binding to the VEGF-binding region in NRP1-b1 domain share the basic sequence motif at the free C-terminus of these ligands and peptides and this distinct motif must be exposed, with a stringent requirement for the Arg (or rarely Lys) residue at their C-terminus; this prerequisite for binding to NRP1 is known as the “C-end rule” (CendR). Moreover, previous studies have reported that lose their binding activity and biological function upon substitution of their C-terminal Arg or Lys residue with another amino acid. However, the development of non-CendR peptides is essential for the future practical applications because the C-terminal Arg and Lys residues of Fc-fused CendR peptides may be removed by carboxypeptidases in cell culture or blood circulation after systemic administration, as shown with antibody. Here, I screened a yeast surface-displayed Fc-fused non-CendR peptide library against NRP1 and isolated Fc-V12, wherein V12 peptide comprising 12 amino acids has a PPRV sequence at its C-terminal end. Although Fc-V12 lacked the CendR motif, it showed selective binding to the VEGF-binding site of NRP1 and triggered cellular internalization of NRP1, resulting in enhanced extravasation into tumor tissues and tumor tissue penetration of the Fc-fused peptide along with the co-injected chemical drug in tumor-bearing mice. Through a saturation mutagenesis study, I identified that the Val residue at the C-terminus of Fc-V12 is essential for NRP1 binding. I further improved NRP1 affinity of Fc-V12 (KD = ~761 nM) through directed evolution of the upstream sequence of PPRV to obtain Fc-V12-33 (KD = ~17.4 nM), which exhibited enhanced NRP1-mediated vascular permeability as compared with Fc-V12. My results provide a new avenue for the development of optimized NRP1-targeting peptides without the restriction of CendR peptides.