Antigen-binding specificities of antibodies are primarily determined by seven residues of VH.

Although the antigen-binding pocket of all antibodies consists of VL + VH dimers (where VL and VH represent immunoglobulin light and heavy chain regions, respectively), subgroups of their VH largely determine their antigen-binding specificities. This VH subgroup dependence automatically relegates subsidiary roles to VL as a whole and to the complementarity-determining region 3 (CDR-3) of VH encoded by independent diversity (D) and joining (J) coding segments in determining antigen-binding specificities of individual antibodies. As a sequel to our previous paper, which emphasized the role conserved residues in CDR-1 and CDR-2 of VH play in general shaping of the primordial antigen-binding cavity, here we propose that the three short clusters of amino acid sequences in CDR-1 and CDR-2 that are placed in the immediate vicinity of the tryptophan loop primarily determine subgroup-dependent antigen preference of individual VH, therefore, antibodies. The three clusters are the 31st to 35th positions of CDR-1 and the 50th to 52nd and 58th to 60th positions of CDR-2. Of those, the 32nd, 34th, 51st, and 59th positions tend to be occupied by tyrosine, methionine, isoleucine, and tyrosine, respectively. Nevertheless, free amino acid substitutions at the remaining seven sites can generate 20(7) or 1.28 X 10(9) varieties of amino acid sequence combinations. Some of these astronomically numerous sequence combinations no doubt contribute to the maintenance of the vast repertoire of antigen-combining diversity, which might be as large as 10(7), whereas others serve to vary binding affinities toward the same antigen. Ironically, but not surprisingly, a single nonconservative amino acid substitution at one of these sites often suffices to change the antigen preference of VH from one to another, whereas more substitutions affecting two or more clusters are apparently required to change the binding affinity toward the same antigen. In the case of mouse anti-p-azophenylarsonate antibodies, the principle of VH subgroup dependence is violated, their VH belonging to either subgroup 1 or 3. It appears that the mouse genome lacks anti-p-azophenylarsonate germ line VH, residues of CDR-3 derived from one particular JH coding segment coming to rescue to cope with this unnatural man-made antigen.