Theoretical study of relations between bay‐region and K‐region indices of carcinogenicity

It has been recognized for several years that the K‐region index of Pullman and the bay‐region carbocation delocalization energy ( BCDE ) of Jerina and co‐workers are quantitatively correlated since they both succeed about equally well in rank‐ordering the same class of molecules for degree of carcinogenicity. However, the underlying reasons for the correlation of these indices has been a puzzle since they refer to different (though related) molecules, and also to different regions on the molecules. In this study we first demonstrate that the K regions of a molecule can be distinguished as being “primitive” or “nonprimitive,” and that it is the primitive K regions that best correlate with BCDE . We then subject both BCDE and primitive K‐region indices to a perturbational molecular orbital ( PMO ) analysis and demonstrate the existence of a formal correlation between their values. From this demonstration we obtain two interesting insights. The first is an understanding of the connection between overall molecular topology and magnitude of these indices (for example, why is a bay region more “active” when attached to a long straight chain of rings than when attached to a shorter straight chain or to a bent chain?). The second is the recognition that certain methyl substitutions must have correlated effects on these two indices while other methyl substitutions have uncorrelated effects.