Phylogenetic analyses uncover a novel clade of transferrin in nonmammalian vertebrates

Abstract

Transferrin is a protein super-family involved in iron transport, a central process in cellular homeostasis. Throughout the evolution of vertebrates, transferrin members have diversified into distinct subfamilies including serotransferrin, ovotransferrin, lactoferrin, melanotransferrin, the inhibitor of carbonic anhydrase, pacifastin, and the major yolk protein in sea urchin. Previous phylogenetic analyses have established the branching order of the diverse transferrin subfamilies but were mostly focused on the transferrin repertoire present in mammals. Here, we conduct a comprehensive phylogenetic analysis of transferrin protein sequences in sequenced vertebrates, placing a special focus on the less-studied nonmammalian vertebrates. Our analyses uncover a novel transferrin clade present across fish, sauropsid, and amphibian genomes but strikingly absent from mammals. Our reconstructed scenario implies that this novel class emerged through a duplication event at the vertebrate ancestor, and that it was subsequently lost in the lineage leading to mammals. We detect footprints of accelerated evolution following the duplication event, which suggest positive selection and early functional divergence of this novel clade. Interestingly, the loss of this novel class of transferrin in mammals coincided with the divergence by duplication of lactoferrin and serotransferrin in this lineage. Altogether, our results provide novel insights on the evolution of iron-binding proteins in the various vertebrate groups.

Fig. 1.

Phylogenetic tree of vertebrate transferrins based on Bayesian and ML analyses. Major groups of vertebrate transferrins are resolved, including the newly identified clade (clade IV). Clade III is divided into several subfamilies, some of which emerged from more recent duplications. Numbers on the branches indicate Bayesian posterior probabilities followed by bootstrap support in the ML analyses. If the partition does not appear in the ML tree, this is indicated with an asterisk. The final log likelihood of the ML tree was −40,597.96204 and the gamma shape parameter 1.336.

Fig. 2.

Topology for a subset of transferrin proteins used for the branch-specific model. Branches subtending the newly identified group and serotransferrins were assumed to have different rate from a background. The branch of novel transferrin indicates ω > 0 shows that the genes have undergone position selection. # indicates the branches that have different rate of substitution (dN/dS = ω).

Fig. 3.

Type 2 functional divergence. (a) Thirteen amino acids residues with posterior ratio more than 8 have been observed from type 2 functional divergence. The observation of amino acids location that have high posterior ratio (R > 8) shows that each cluster has a conserved radical change between each other. (b) The 3D protein for transferrin protein in Lates calcarifer was modeled to show the location of residues affected by radical changes. Five of the residues that have high posterior ratio found to form loop, five forming α-helix, whereas another three forming the β-strand. Tyr₅₂₄ is the functional residue for iron binding, which has been substituted to Asp in novel sequences. Arg₁₂₁ and Arg₄₆₀ involve in anion binding, which has been substituted to Trp and Ser in the novel sequences, respectively. The substitutions of important residues might suggest that those sites may responsible to perform a different function after the duplication event. The locations of the residues are calculated based on the Gallus gallus ovotransferrin sequence (2D3I) obtained from PDB database (last accessed November 1, 2012).