An examination of the expected degree of sequence similarity that might arise in proteins that have converged to similar conformational states: The impact of such expectations on the search for homology between the structurally similar domains of rhodanese

Abstract

Recognition of structural similarity in proteins invites the inference of homology even when the amino acid sequences are not highly similar. The influence of structural similarity on both the genetic tests for amino acid sequence similarity and the inference of homology was examined by statistical methods. Structure-dependent compositions of amino acid sequences representing segments of secondary and supersecondary structure were examined for the preferential occurrence of structurally similar amino acids that are also similar according to the genetic criteria of Fitch (1970), Dayhoff (1979) and McLachlan (1971). These analyses revealed that: (1) the preferential occurrence of structurally similar amino acids in analogous secondary structures should not give rise to a statistically significant sequence similarity; (2) some positional amino acid preferences in secondary and supersecondary structures score highly in sequence similarity tests; (3) the preferential occurrence of non-polar, β-branched amino acids in the parallel β-pleated sheets of α β proteins constitutes an important structural bias in genetic tests for sequence similarity. These findings may be applied to sequence comparisons whenever the conformational states are either known experimentally or may be inferred from predictive analysis of the amino acid sequence. The corrections for structure-dependent compositions were made in a search for homology between the two structurally similar, globular domains of bovine liver rhodanese. Despite these corrections and earlier failures to observe significant sequence similarity, a statistically significant sequence similarity was detected, supporting the inference that the domains are internally paralogous, i.e. intraspecies products of a partial internal duplication of an ancestral gene.