Intra-Protein Coevolution Is Increasingly Functional with Greater Proximity to Fertilization.

Intramolecular coevolution of amino acid sites has repeatedly been studied to improve predictions on protein structure and function. Thereby, the focus was on bacterial proteins with available crystallographic data. However, intramolecular coevolution has not yet been compared between protein sets along a gradient of functional proximity to fertilization. This is especially true for the potential effect of external selective forces on intraprotein coevolution. In this study, we investigated both aspects in equally sized sets of mammalian proteins representing spermatozoa, testis, entire body, and liver. For coevolutionary analyses, we derived the proportion of covarying sites per protein from amino acid alignments of 10 mammalian orthologues each. In confirmation of the validity of our coevolution proxy, we found positive associations with the nonsynonymous or amino acid substitution rate in all protein sets. However, our coevolution proxy negatively correlated with the number of protein interactants (node degree) in male reproductive protein sets alone. In addition, a negative association of our coevolution proxy with protein hydrophobicity was significant in sperm proteins only. Accordingly, the restrictive effect of protein interactants was most pronounced in male reproductive proteins, and the tendency of sperm proteins to form internal structures decreased the more coevolutionary sites they had. Both aspects illustrate that the share of outward and thus functional coevolution increases with greater proximity to fertilization. We found this conclusion confirmed by additional comparisons within sperm proteins. Thus, sperm proteins with high hydrophobicity had the lowest proportions of covarying sites and, according to gene annotations, localized more frequently to internal cellular structures. They should therefore be less exposed to postcopulatory forms of sexual selection. Their counterparts with low hydrophobicity had larger proportions of covarying sites and more often resided at the cell membrane or were secreted. At the cellular level, they are thus closer to externally induced forces of postcopulatory selection which are known for their potential to increase substitution rates. In addition, we show that the intermediary status of the testicular protein set in correlation analyses is probably due to a special combination of reproductive and somatic involvements.