Covariation of residues in the homeodomain sequence family.

Homeodomains are 60 amino acid DNA binding domains found in numerous eukaryotic transcription factors. The homeodomain family is a useful system for studying sequence-structure relationships because several hundred sequences are known and the structures of several homeodomains have been determined. Covariation of amino acid residues in the homeodomain family has been investigated to see whether strongly covariant residue pairs can be understood in terms of the structure and function of these domains. Among 16 strongly covariant pairs examined, 2 are explained by the ability to form salt bridges, and 9 appear related to the DNA binding function of the proteins. For the remaining 5 pairs, the rationale for covariance remains unclear and the likelihood of artifactual correlations is discussed in the context of experimental and evolutionary biases in the selection of sequences. No significant correlation was found between covariance and structural proximity in the hydrophobic core.