50 years of antibody numbering schemes: a statistical and structural evaluation reveals key differences and limitations

Abstract

The complementarity determining region (CDR) of antibodies represents the most diverse region both in terms of sequence and structural characteristics, playing the most critical role in antibody recognition and binding for immune responses. Over the past decades, several numbering schemes have been introduced to define CDRs based on sequence. However, the existence of diverse numbering schemes has led to potential confusion, and a comprehensive evaluation of these schemes is lacking. We employ statistical analyses to quantify the diversity of CDRs compared to the framework regions. Comparative analyses across different numbering schemes demonstrates notable variations in CDR definitions. The Kabat and AbM numbering schemes tend to incorporate more conserved residues into their CDR definitions, whereas CDRs defined by the Chothia and IMGT numbering schemes display greater diversity, sometimes missing certain loop residues. Notably, we identify a critical residue, L29, within the kappa light chain CDR1, which appears to act as a pivotal structural point within the loop. In contrast, most numbering schemes designate the topological equivalent point in the lambda light chain as L30, suggesting the need for further refinement in the current numbering schemes. These findings shed light on regional sequence and structural conservation within antibody sequence databases while also highlighting discrepancies stemming from different numbering schemes. These insights yield valuable guidelines for the precise delineation of antibody CDRs and the strategic design of antibody repertoires, with practical implications in developing innovative antibody-based therapeutics and diagnostics. GRAPHIC ABSTRACT O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=89 SRC="FIGDIR/small/602590v1_ufig1.gif" ALT="Figure 1"> View larger version (35K): org.highwire.dtl.DTLVardef@2240a7org.highwire.dtl.DTLVardef@c2981dorg.highwire.dtl.DTLVardef@190984eorg.highwire.dtl.DTLVardef@1ff5e00_HPS_FORMAT_FIGEXP M_FIG C_FIG HIGHLIGHTSO_LIAmino acid distribution analysis conducted on 124,317 sequences archived in the abYsis database showed lower conservation among CDR residues, as defined by Kabat numbering, compared to the framework region. C_LIO_LIBoth LCDR3 and HCDR3 exhibited particularly high levels of diversity compared to the other four CDR loops, highlighting additional functional roles of the CDR3 loops relative to the others. C_LIO_LITwo distinct residues, L29 and H30, displayed highly conserved hydrophobic side chain characteristics and played unique conformational roles within the CDR1 loop, contributing to the formation of an M-shaped structure. This residue divides the C-terminus CDR1 loop, which was observed to be less conserved and have greater length diversity, and the N-terminus CRD1 loop. C_LIO_LIThe Kabat numbering scheme appears to overestimate the LCDR2 and HCDR2 loops and underestimate HCDR1, only encompassing the C-terminal loop of the M-shaped CDR1 loop. C_LIO_LIStructural analysis revealed that a residue within LCDR1 was numbered differently in the two subtypes (kappa and lambda) of the variable light domain. This discrepancy suggests the need for further refinement in the current major numbering schemes. C_LI