It is well known that single amino acid substitution of antigenic determinants dramatically decreases the affinity to antibodies. This phenomenon is strikingly similar to our previous observations in amino acid substitution of the three-fragment complex of horse cytochrome c. As described in another report, we hypothesize that four closed loops consisting of contacting groups mediates delocalized interaction to generate extra energy to stabilize cytochrome c and that substitution of an amino acid which is a part of this closed loop would disrupt delocalized interaction. To investigate whether a specifically recognized amino acid is a part of a hypothetical closed loop formed across or within the interface between an antigen and an antibody we have isolated 7 monoclonal antibodies to yeast holo- or apo-iso-l-cytochrome c as described in the previous years. In the present studies, we have quantitatively determined the affinities of these monoclonals with respect to yeast holo- and apo-iso-l-cytochrome c a panel of evolutionarily related cytochromes c, apocytochromes c, and homologous and hybrid fragment complexes. The results taken together with the previous data have permitted us to assign specifically recognized amino acids as follows. IgG monoclonals: 4-74-6, Leu 63 (yeast numbering) and/or Asn 67 and/or Asn 68; 4- 126-6, Glu 93; 4-145-10, Thr 74; 2-96-12, Asp 65; 2-34-19, Lys 59; and 10-28-86, trimethyl-Lys 77. IgM monoclonal 39-14, Pro 30 and His 31. With the exception of mAbs 4-14-10 and 39-14 these monoclonals are of high affinity. A calculation with mAb 4-126-6 in which replacement of Glu 93 by alanine results in a decrease in affinity by a factor of 10,000 appears to show that the sum or conventional interactions such as electrostatics (or hydrogen bond), hydrophobic interaction and van der Waals interaction does not totally account for the decrease in affinity. Thus, we suggest that some new, extra interatomic interaction sensitive to differences in configurations of atomic groups may be involved in antigen recognition as predicted by the closed interaction loop hypothesis.
