Structural consensus among antibodies defines the antigen binding site

Abstract

The Complementarity Determining Regions (CDRs) of antibodies are assumed to account for the antigen recognition and binding and thus to contain also the antigen binding site. CDRs are typically discerned by searching for regions that are most different, in sequence or in structure, between different antibodies. Here, we show that ~20% of the antibody residues that actually bind the antigen fall outside the CDRs. However, virtually all antigen binding residues lie in regions of structural consensus across antibodies. Furthermore, we show that these regions of structural consensus which cover the antigen binding site are identifiable from the sequence of the antibody. Analyzing the predicted contribution of antigen binding residues to the stability of the antibody-antigen complex, we show that residues that fall outside of the traditionally defined CDRs are at least as important to antigen binding as residues within the CDRs, and in some cases, they are even more important energetically. Furthermore, antigen binding residues that fall outside of the structural consensus regions but within traditionally defined CDRs show a marginal energetic contribution to antigen binding. These findings allow for systematic and comprehensive identification of antigen binding sites, which can improve the understanding of antigenic interactions and may be useful in antibody engineering and B-cell epitope identification.

Figure 1. Structure-based identification of ABRs.

(A) Using the non-redundant set of all Ab-Ag complexes in the PDB, (B) we created a multiple structure alignment of the Abs. Residues that are in contact with the Ag were identified by searching for structurally aligned positions that systematically create contacts with the Ag (black and grey solid circles) and disregarded positions that contact the Ag only sporadically (open shapes). (C) The contacting positions were mapped to the sequence representation of the multiple structure alignment (bold letters). The stretches of amino acids in which at least 10% of the Abs are in contact with the Ag were defined as ABRs (white rectangle).

Figure 2. Automated ABRs Identification

(A) Sequence based ABRs identification. A BLAST search is performed using the query Ab sequence versus the dataset of non-redundant PDB Abs. Using the best hit from the BLAST search, the query and annotated Abs FRs are aligned and hence the query sequence ABRs are inferred based on the location of the annotated sequence ABRs in the MSTA. (B) Structure based ABRs identification. A BLAST search is performed using the sequence of the query Ab versus our dataset of Abs. Using the best hit from the BLAST search, the query and annotated Abs are structurally aligned. The ABRs of the query Ab are inferred based on the location of the annotated Ab ABRs in the MSTA.

Figure 3. Total number of residues identified by each method for all Ab-Ag complexes in the test set.

L1–L3 are ABR/CDR1-3 of the light chain. H1–H3 are ABR/CDR1-3 of the heavy chain. Total light and heavy are the sum of all identified residues in the light and heavy chains respectively.

Figure 4. Recall and precision of Ag binding sites identification.

Average precision and recall were calculated for the Abs in the test set for Paratome, Kabat, Chothia and IMGT methods. Error bars represent standard error of the mean.

Figure 5. Average Ag binding sites recall and precision of light and heavy chains for all ABRs/CDRs.

(A) Average Ag binding sites precision (B) Average Ag binding sites recall. Error bars represent standard error of the mean.

Figure 6. Contribution of Paratome-unique and CDR-unique residues to the binding energy in Ab-Ag complexes.

(A) The distributions of ΔΔG values of an in-silico alanine scan analysis of Paratome-unique, CDRs-unique and CDR Ag binding residues. ΔΔG values ranging between −0.25 and 0.25 were defined as neutral. ΔΔG values<−0.25 were defined as stabilizing. ΔΔG values >0.25 were defined as destabilizing. It is clear from the distributions that typically, a Paratome-unique residue is at least as energetically important as a residue in the CDRs, while a CDR unique residue is less energetically important relative to residues within the CDRs that are identified by Paratome. (B)+(C) A detailed analysis of anti-IL-15 Ab with human IL-15 (PDB 2xqb). (B) The surface of the Ag chain is rendered according to atom charge. Due to the hydrogen bond with antigenic GLU53, TYR49 is located in high proximity to the Ag. Ab LEU46 is located in proximity to antigenic LEU52. (C) Seven residues from L2 (green, solid ribbon) interact with the Ag (orange, solid ribbon). Two of them (LEU46 and TYR49) are not identified as part of the CDR by any other CDR identification method. These Paratome-unique residues and the antigenic residues they contact (LEU52 and GLU53) are depicted in sticks. LEU46 forms a hydrophobic interaction with LEU52. TYR49 forms a hydrogen bond with antigenic GLU53 as well as a cation- п interaction with ARG53 of the Ab. Both contribute substantially to Ag binding (see text).

Figure 7. Comparison of consensus and method-specific residues.

A light chain is analysed by all methods. In the top comparison, the residues identified by Paratome (grey, top line) are compared to those identified by Kabat (grey, second line). For L1, for example, both methods agree on the fragment ESVDSYGKSFMH, however, according to Kabat, CDR L1 includes also three amino acids N-terminal to this fragment (RAS, marked with a grey box) while according to Paratome these are not included in ABR L1. ABR L2, is identified by Paratome to be longer than the CDR L2 identified by Kabat by four residues (VLIY, marked with a dashed box).