Evolution of the relaxin-like peptide family

Abstract

Background: The relaxin-like peptide family belongs in the insulin superfamily and consists of 7 peptides of high structural but low sequence similarity; relaxin-1, 2 and 3, and the insulin-like (INSL) peptides, INSL3, INSL4, INSL5 and INSL6. The functions of relaxin-3, INSL4, INSL5, INSL6 remain uncharacterised. The evolution of this family has been contentious; high sequence variability is seen between closely related species, while distantly related species show high similarity; an invertebrate relaxin sequence has been reported, while a relaxin gene has not been found in the avian and ruminant lineages.

Results: Sequence similarity searches of genomic and EST data identified homologs of relaxin-like peptides in mammals, and non-mammalian vertebrates such as fish. Phylogenetic analysis was used to resolve the evolution of the family. Searches were unable to identify an invertebrate relaxin-like peptide. The published relaxin cDNA sequence in the tunicate, Ciona intestinalis was not present in the completed C. intestinalis genome. The newly discovered relaxin-3 is likely to be the ancestral relaxin. Multiple relaxin-3-like sequences are present in fugu fish (Takifugu rubripes) and zebrafish (Danio rerio), but these appear to be specific to the fish lineage. Possible relaxin-1 and INSL5 homologs were also identified in fish and frog species, placing their emergence prior to mammalia, earlier than previously believed. Furthermore, estimates of synonymous and nonsynonymous substitution rates (dN/dS) suggest that the emergence of relaxin-1, INSL4 and INSL6 during mammalia was driven by positive Darwinian selection, hence these peptides are likely to have novel and in the case of relaxin-1, which is still under positive selection in humans and the great apes, possibly still evolving functions. In contrast, relaxin-3 is constrained by strong purifying selection, demonstrating it must have a highly conserved function, supporting its hypothesized important neuropeptide role.

Conclusions: We present a phylogeny describing the evolutionary history of the relaxin-like peptide family and show that positive selection has driven the evolution of the most recent members of the family.

Figure 1

Multiple sequence alignment of the relaxin-like peptide family. Amino acid sequences of the B and A domains only were aligned using ClustalW, then edited by hand using Seaview to remove gaps. This alignment was then used for all phylogenetic analyses. Newly identified sequences are highlighted in bold and italics. Invariant cysteine residues are indicated by asterisks (*) and the relaxin specific B-chain motif [RxxxRxxI/V] is shown. Sequences are clustered into subfamilies (A and B) based on primary sequence similarity and phylogenetic analysis. Hsa = Homo sapiens, Pt = Pan troglodytes, Mmul = Maca mulatta, Mm = Mus musculus, Rn = Rattus norvegicus, Cf = Canis familiaris, Ss= Sus scrofa, Re = Rana esculenta, Me = Macropus eugenii, Xl = Xenopus laevis, Xt = Xenopus tropicalis, Dr = Danio rerio, Tr = Takifugu rubripes, Gg = Gallus gallus, Om = Oncorhynchus mykiss.

Figure 2

Evolutionary relationships among relaxin-like peptides. Topology shown is a consensus tree based on MP (maximum parsimony), ML (maximum likelihood) and NJ (neighbour-joining) analysis of the amino acid alignment shown in figure 1. Consensus tree was produced and edited using TreeView to correlate topology with known genomic information about the family. Human insulin used as an outgroup. Where possible, confidence values are shown at branches: * >50%, ** >75%, all other branches are inferred. Hsa = Homo sapiens, Pt = Pan troglodytes, Mmul = Maca mulatta, Mm = Mus musculus, Rn = Rattus norvegicus, Cf = Canis familiaris, Ss = Sus scrofa, Re = Rana esculenta, Me = Macropus eugenii, Xl = Xenopus laevis, Xt = Xenopus tropicalis, Dr = Danio rerio, Tr = Takifugu rubripes, Gg = Gallus gallus, Om = Oncorhynchus mykiss.

Figure 3

Reconciled tree for the relaxin-like peptide family. The consensus tree of relaxin-like peptides (figure 2) from human, chimpanzee, mouse, dog, rat, pig, wallaby, chicken, fugu fish, zebrafish, rainbow trout, R. esculenta, X. laevis and X. tropicalis was reconciled using GeneTree with a species tree complied from a phylogeny of model organisms [65]. Squares indicate duplication events, red dotted lines indicate absent genes, either lost from those species (in grey), or not yet sequenced. While used to construct the gene tree as an outgroup, insulin has been removed from the reconciled tree. Hsa = Homo sapiens, Pt = Pan troglodytes, Mmul = Maca mulatta, Mm = Mus musculus, Rn = Rattus norvegicus, Cf = Canis familiaris, Ss = Sus scrofa, Re = Rana esculenta, Me = Macropus eugenii, Xl = Xenopus laevis, Xt = Xenopus tropicalis, Dr = Danio rerio, Tr = Takifugu rubripes, Gg = Gallus gallus, Om = Oncorhynchus mykiss.

Figure 4

Synonymous (d_S) and nonsynonymous (d_N) substitution rate estimates for individual B and A domains of each relaxin-like gene. Substitution rates (d_N/d_S) were estimated using the Yang and Neilsen, 2000 method as implemented in yn00 in the PAML suite. Human and chimpanzee comparisons were used for RLN1 and INSL6; human and rhesus monkey comparisons were used for INSL4 and human, mouse comparisons were used for RLN2, RLN3, INSL3 and INSL5.

Figure 5

Phylogeny of mammalian RLN2 and INSL4 genes. Tree generated using Tree-Puzzle using a gamma distribution, the Dayhoff model of substitution and 10 000 puzzling steps. Confidence values are shown as percentages on each branch. The INSL4 branch (labeled A on the tree) was tested for positive selection. Hsa = Homo sapiens, Mm = Mus musculus, Rn = Rattus norvegicus, Ss = Sus scrofa, Me = Macropus eugenii, Mmul = Maca mulatta, Cd = Camelus dromedaries, Gc = Galago crassicaudatus, Pt = Pan troglodytes, Ggo = Gorilla gorilla, Ph = Papio hamadryas, Fc = Felis catus, Cf = Canis familiaris, Oc = Oryctolagus cuniculus, Ma = Mesocricetus auratus, Ec = Equus caballus.