ӣ��Ӱ��

Topic

Structural studies of antibody recognition of clinically significant antigens

Department of Biochemistry and Microbiology

Date & location

• Friday, December 5, 2025
• 10:00 A.M.
• Clearihue Building, Room B021

Examining Committee

Supervisory Committee

• Dr. Stephen Evans, Department of Biochemistry and Microbiology, ӣ��Ӱ�� (Supervisor)
• Dr. Alisdair Boraston, Department of Biochemistry and Microbiology, UVic (Member)
• Dr. Caroline Cameron, Department of Biochemistry and Microbiology, UVic (Member)
• Dr. Jeremy Wulff, Department of Chemistry, UVic (Outside Member)

External Examiner

• Dr. Kenneth Ng, Department of Chemistry and Biochemistry, University of Windsor

Chair of Oral Examination

• Dr. Adam Krawitz, Department of Psychology, UVic

Abstract

Twenty structures of the Fabs (fragment antigen-binding) of 3 distinct antibodies that are either specific or have potential cross-specificity for clinically significant antigens, unliganded and, if possible, in complex are presented in 3 data chapters that explore their specificity, cross reactivity and, when appropriate, polyspecific potential.

Chapter 2 is focussed on the Fab from monoclonal antibody (mAb) 6A7, which was reported in 2007¹ to bind a Bax peptide. Bax is a pro-apoptotic member of the BCL-2 family that resides primarily in the cytoplasm of healthy cells. When activated, it will migrate to the mitochondrial outer membrane and induces cell death.^1,2 We selected 6A7 for study because a sequence alignment showed the combining site was likely to contain a germline-coded recognition pocket for the bacterial sugar 3-deoxy-D-manno-oct-2-ulosonic acid (Kdo), opening the possibility that 6A7 was able to bind both a peptide and a carbohydrate antigen, i.e. mAb 6A7 may be a truly polyspecific antibody. Interestingly, while the structures of several co-crystals of 6A7 soaked in antigen at high molar excess all showed the sugar in the expected position in the combining site (which almost completely overlaps with the paratope of the peptide), the Fab itself was shown by Surface Plasmon Resonance (SPR) studies not to display any observable affinity for Kdo. The likely reason for the observed lack of affinity could be traced to a somatic mutation in 6A7 that re-positioned an adjacent germline-residue deep inside the combining site pocket in a manner that significantly weakened but did not exclude the sugar from binding. Modelling indicates that mutation of a single residue back to germline may result in an antibody that shows true polyspecificity.

Chapter 3 picks up where Chapter 2 left off. MAb S23-24 belongs to a large family of characterized antibodies that all use the same heavy and light chain V-genes that code for the CDRs, L1, L2, H1, H2, and a few residues in L3 and H3 that make up most of the germline Kdo-recognition pocket. This family is named after its first characterized member, S25-2, which is cross-reactive toward many different Kdo-based antigens, while other members of the family bind various Kdo-based antigens with greater or lesser specificity largely depending on the identity of the D and J genes that code for most of CDR H3. While the near-germline S25-2 and the homologous antibody S25-39 are highly cross-reactive, S23-24 is significantly more affinity matured but, like S25-2 and S25-39, still appears to bind
antigen though a mechanism of conformational selection. In all conformations the CDR H3 is oriented away from the terminal Kdo binding pocket, leading to less contact with the internal saccharide residues and hence higher cross-reactivity.

Further, and in a fascinating turn, S23-24 was shown also to have measurable affinity for the Bax peptide, indicating that the germline-coded Kdo-recognition pocket can indeed provide the starting point for affinity maturation into an antibody like 6A7.

Chapter 4 is focussed on the humanized mAb h-5E5, which is based on the murine mAb 5E5 specific for the mucin Tn antigens PAPGS*T*AP and APGS*T*AP, where * stands for a GalNAc residue. Given the known structure of the Fv from murine 5E5, the determination of the structure of h-5E5 would offer insight first into the effect of humanization of an important potential ana-cancer therapeutic agent and second into antibodies that recognize glycopeptides (i.e. antigens that contain both carbohydrate and peptide moieties). Surprisingly, despite the known high avidity of murine 5E5 for the mucin Tn antigens, and the known low-resolution structure of murine 5E5 Fv in complex with the APGST*AP antigen, thousands of attempts to crystallize the humanized Fab with Tn antigen failed to produce any liganded crystals. Measurement of the affinity of h-5E5 showed it to be significantly lower than that expected for the murine structure. Examination of the unliganded crystal structure reveals that the humanization process has caused the domain associations of h-5E5 to shift significantly beyond the corresponding parameters in the murine 5E5 structure. This shift altered the combining site, which in turn may explain the observed loss of affinity.