3-DIMENSIONAL STRUCTURE OF MURINE ANTI-PARA-AZOPHENYLARSONATE FAB-36-71 .2. STRUCTURAL BASIS OF HAPTEN BINDING AND IDIOTYPY

Comparisons between the structures and solvent-accessible surfaces of the antigen-binding fragments of two murine anti-p-azophenylarsonate monoclonal antibodies, one bearing a major cross-reactive idiotype of A/J strain mice (36-71) and one lacking the idiotype (R19.9; Lascombe et al., 1989), highlight the structural basis for the determination of hapten affinity and idiotypy. Since the sequence of R19.9 is identical with the germline-encoded sequence at 16 positions in both heavy-chain and light-chain variable regions where somatic mutations and junctional differences have occurred to produce the 36-71 sequence, the structure of R19.9 can be used to model the structure of the germline-encoded antibody (36-65) in the regions around these sites. These 16 sequence differences exclude the third heavy-chain complementarity-determining region because R19.9 utilizes a D gene segment not associated with the predominant idiotype, which is 4 residues longer than the canonical D gene segment utilized in the sequences of 36-71 and 36-65. This difference between the structures of R19.9 and 36-71 does not affect the validity of using the structure of R 19.9 to model the structure of 36-65 since the third heavy-chain complementarity-determining region is highly solvent-exposed in both 36-71 and R19.9, and does not interact with any of these 16 sites. Comparing the structures of 36-71 and R19.9 suggests that only three of the differences in the heavy-chain sequences, and three of the differences in the light-chain sequences of 36-71 and 36-65, increase the affinity for hapten. The substitutions in the light chain appear to affect the binding constant for hapten by altering the overall conformation of the third heavy-chain complementarity-determining region. Regions where the solvent-accessible surfaces of 36-71 and R19.9 differ have been located. Presumably, these differences in accessible surface area account for the lack of reactivity with anti-idiotypic antibodies for R19.9 versus 36-71. Positions in the sequence where differences are known to affect the binding of anti-p-azophenylarsonate antibodies to monoclonal anti-idiotope antibodies generally map to the region around the hapten-binding site. Not unexpectedly, substitutions at these locations in the sequence do not always appear to affect anti-idiotope binding directly but may alter idiotype/anti-idiotype affinity indirectly by altering the conformation of contact residues.