Positive selection shaped the convergent evolution of independently expanded kallikrein subfamilies expressed in mouse and rat saliva proteomes

Abstract

We performed proteomics studies of salivas from the genome mouse (C57BL/6 strain) and the genome rat (BN/SsNHsd/Mcwi strain). Our goal was to identify salivary proteins with one or more of three characteristics that may indicate that they have been involved in adaptation: 1) rapid expansion of their gene families; 2) footprints of positive selection; and/or 3) sex-limited expression. The results of our proteomics studies allow direct comparison of the proteins expressed and their levels between the sexes of the two rodent species. Twelve members of the Mus musculus species-specific kallikrein subfamily Klk1b showed sex-limited expression in the mouse saliva proteomes. By contrast, we did not find any of the Rattus norvegicus species-specific kallikrein subfamily Klk1c proteins in male or female genome rat, nor transcripts in their submandibular glands. On the other hand, we detected expression of this family as transcripts in the submandibular glands of both sexes of Sprague-Dawley rats. Using the CODEML program in the PAML package, we demonstrate that the two rodent kallikrein subfamilies have apparently evolved rapidly under the influence of positive selection that continually remodeled the amino acid sites on the same face in the members of the subfamilies. Thus, although their kallikrein subfamily expansions were independent, this evolutionary pattern has occurred in parallel in the two rodent species, suggesting a form of convergent evolution at the molecular level. On the basis of this new data, we suggest that the previous speculative function of the species-specific rodent kallikreins as important solely in wound healing in males be investigated further. In addition to or instead of that function, we propose that their sex-limited expression, coupled with their rapid evolution may be clues to an as-yet-undetermined interaction between the sexes.

Figure 1. Identification of rat subfamily kallikrein transcripts in submandibular glands of the genome (BN/SsNHsd/Mcwi strain) and Sprague-Dawley rats.

Submandibular gland cDNA libraries were produced by reverse transcription of total RNA. Kallikrein transcripts were amplified by PCR using primers produced from the designs published by and separated on a 2% agarose gel. The samples in each gel photo are identified by the primer nomenclature of with the rat gene nomenclature of listed parenthetically in this figure legend. Lanes 2–11: rKLK1 (Klk1); rKLK2 (Klk1c2); rKLK3 (Klk1c3); rKLK4 (Klk1c4); rKLK6 (Klk1c6); rKLK7 (Klk1c7); rKLK8 (Klk1c8); rKLK9 (Klk1c9); rKLK10 (Klk1c10); rKLK12 (Klk1c12). Lane 1: 100 bp DNA ladder; Lane12: GAPDH cDNA control and Lane 13: beta-actin cDNA control. It should be noted that, in the case of the two Sprague-Dawley rat gel panels, there is some curvature that causes the controls on the right end to appear to have somewhat different molecular weights than in the genome rat gel panels but this is not the case.

Figure 2. A comparison of the ω+ sites found in the mouse and rat species-specific kallikrein families, using the CODEML program (an independent analysis for each subfamily).

The alignment of the basal Klk1 sequences used in this figure contains a three-residue gap so that the numbering system will be compatible with Table 2. The numbering system begins with the first amino acid residue of the cleaved proenzyme (i.e., the active form of the enzyme). This changes the numbers of the active site triad (sites indicated with ! and in green typeface) from His57 to His41, and with the three-residue gap Asp112 and Ser205, become Asp99 and Ser192. The number of the D199 that influences substrate specificity (site indicated with #) is changed to 186. Asterisks mark the sites with posterior probabilities greater than 0.9 calculated without removing GENECONV sequences. Plus signs (+) mark the sites with posterior probabilities greater than 0.9 calculated after GENECONV sequences have been removed. Amino acid residues identical between the mouse and rat Klk1 sequences are shown in gray typface, while differences are shown in black. Sites positive both with and without GENECONV sequences are identified with blue typeface in mouse and red in rat. Half-cystine residues are shown in yellow typeface and oriented to each other with vertical dashes.

Figure 3. Positive selection at sites on molecular models of mouse and rat species-specific kallikreins.

Because both rodent Klk1 kallikreins mapped on the d1gvza model, that was used with each Klk1 sequence to produce the mouse and rat molecular models with PyMol. Residues in red are selected with a BEB posterior probability >99%; those in green with a BEB posterior probability >95%; and those in blue with a BEB posterior probability >90%. Panels A and D: Cartoon model of mouse and rat subfamily kallikreins, respectively, with α-helices shown as spiral tapes and β-sheets shown as flat arrows. The active site triad is shown in spheres colored light orange (His41), pink (Asp96) and light blue (Ser189). The two Greek key motifs, consisting of β-sheets can be seen above and below the active site residues. Solid models of mouse (Panels B and C) and rat (Panels E and F) show the fronts (B and E, where B is the same orientation as A, and E is the same orientation as D) and backs (C and F) of the two species-specific kallikreins. Essentially all the sites under selection map to the front surfaces of both structures.