Relaxin gene family in teleosts: phylogeny, syntenic mapping, selective constraint, and expression analysis

Abstract

Background: In recent years, the relaxin family of signaling molecules has been shown to play diverse roles in mammalian physiology, but little is known about its diversity or physiology in teleosts, an infraclass of the bony fishes comprising approximately 50% of all extant vertebrates. In this paper, 32 relaxin family sequences were obtained by searching genomic and cDNA databases from eight teleost species; phylogenetic, molecular evolutionary, and syntenic data analyses were conducted to understand the relationship and differential patterns of evolution of relaxin family genes in teleosts compared with mammals. Additionally, real-time quantitative PCR was used to confirm and assess the tissues of expression of five relaxin family genes in Danio rerio and in situ hybridization used to assess the site-specific expression of the insulin 3-like gene in D. rerio testis.

Results: Up to six relaxin family genes were identified in each teleost species. Comparative syntenic mapping revealed that fish possess two paralogous copies of human RLN3, which we call rln3a and rln3b, an orthologue of human RLN2, rln, two paralogous copies of human INSL5, insl5a and insl5b, and an orthologue of human INSL3, insl3. Molecular evolutionary analyses indicated that: rln3a, rln3b and rln are under strong evolutionary constraint, that insl3 has been subject to moderate rates of sequence evolution with two amino acids in insl3/INSL3 showing evidence of positively selection, and that insl5b exhibits a higher rate of sequence evolution than its paralogue insl5a suggesting that it may have been neo-functionalized after the teleost whole genome duplication. Quantitative PCR analyses in D. rerio indicated that rln3a and rln3b are expressed in brain, insl3 is highly expressed in gonads, and that there was low expression of both insl5 genes in adult zebrafish. Finally, in situ hybridization of insl3 in D. rerio testes showed highly specific hybridization to interstitial Leydig cells.

Conclusions: Contrary to previous studies, we find convincing evidence that teleosts contain orthologues of four relaxin family peptides. Overall our analyses suggest that in teleosts: 1) rln3 exhibits a similar evolution and expression pattern to mammalian RLN3, 2) insl3 has been subject to positive selection like its mammalian counterpart and shows similar tissue-specific expression in Leydig cells, 3) insl5 genes are highly represented and have a relatively high rate of sequence evolution in teleost genomes, but they exhibited only low levels of expression in adult zebrafish, 4) rln is evolving under very different selective constraints from mammalian RLN. The results presented here should facilitate the development of hypothesis-driven experimental work on the specific roles of relaxin family genes in teleosts.

Figure 1

Synteny maps. Synteny maps comparing the orthologues of the relaxin family loci (RFL) and the genes flanking them in humans (H. sapiens) and five species of teleosts (G. aculeatus, D. rerio, T. nigroviridis, T. rubripes and O. latipes). 1A) RFLA locus contains the INSL5 gene in humans and its teleostean paralogues, insl5a/insl5b; 1B) RFLB locus in humans hosts four relaxin family genes, namely INSL6, INSL4, RLN2 and RLN1; in teleosts this locus is represented by rln gene found as a single copy in all of the analyzed teleost genomes except for D. rerio, in which it is absent; 2A) RFLCI locus is represented by RLN3 in humans, and its paralogues, rln3a/rln3b in teleosts; 2B) RFLCII locus hosts INSL3 in humans, while 3 of the 5 studied teleosts contain single copy orthologues, insl3. The chromosome number (in Roman numerals) and map position of each gene in humans are given on the right. On the left, the genes orthologous to the human RFL are shown by orange hexagons in the central shaded section, and RFL paralogue that arose via the whole genome duplication shown as a white hexagon. Genes flanking the RFL that are syntenic in humans and teleosts are indicated by orange rectangles; the map position of each gene in teleosts is listed in Additional File 1: Table S3. Tandem duplicate copies of genes appear as two rectangles. Genes shown as white rectangles are genes identified on the same chromosome but in more distant locations (Additional File 1: Table S1). The genes PDE4B/SLC35D1/SERBP1/RPE65 (RFLA); JAK2 (RFLB); TNPO2/RFX1/ASF1B/SLC27A1/GLT25D1 (RFLCI); and MED26/NR2F6/UNC13A/KCNN1/MAST3 (RFLCII) were all retained in duplicate in 3 or more species (Additional File 1: Table S1). One gene, NXNL1 RFLCI) was retained tandemly duplicated in 3 species. Three of the 4 RFL linkage groups contained a copy of JAK, and 2 of the 4 contained copies of PDE, SMARCA, RFX and MAST genes.

Figure 2

Synteny maps. Synteny maps comparing the orthologues of the relaxin family loci (RFL) and the genes flanking them in humans (H. sapiens) and five species of teleosts (G. aculeatus, D. rerio, T. nigroviridis, T. rubripes and O. latipes). 1A) RFLA locus contains the INSL5 gene in humans and its teleostean paralogues, insl5a/insl5b; 1B) RFLB locus in humans hosts four relaxin family genes, namely INSL6, INSL4, RLN2 and RLN1; in teleosts this locus is represented by rln gene found as a single copy in all of the analyzed teleost genomes except for D. rerio, in which it is absent; 2A) RFLCI locus is represented by RLN3 in humans, and its paralogues, rln3a/rln3b in teleosts; 2B) RFLCII locus hosts INSL3 in humans, while 3 of the 5 studied teleosts contain single copy orthologues, insl3. The chromosome number (in Roman numerals) and map position of each gene in humans are given on the right. On the left, the genes orthologous to the human RFL are shown by orange hexagons in the central shaded section, and RFL paralogue that arose via the whole genome duplication shown as a white hexagon. Genes flanking the RFL that are syntenic in humans and teleosts are indicated by orange rectangles; the map position of each gene in teleosts is listed in Additional File 1: Table S3. Tandem duplicate copies of genes appear as two rectangles. Genes shown as white rectangles are genes identified on the same chromosome but in more distant locations (Additional File 1: Table S1). The genes PDE4B/SLC35D1/SERBP1/RPE65 (RFLA); JAK2 (RFLB); TNPO2/RFX1/ASF1B/SLC27A1/GLT25D1 (RFLCI); and MED26/NR2F6/UNC13A/KCNN1/MAST3 (RFLCII) were all retained in duplicate in 3 or more species (Additional File 1: Table S1). One gene, NXNL1 RFLCI) was retained tandemly duplicated in 3 species. Three of the 4 RFL linkage groups contained a copy of JAK, and 2 of the 4 contained copies of PDE, SMARCA, RFX and MAST genes.

Figure 3

Origins of relaxin family genes in teleosts (bottom) and humans (top) determined by synteny map analyses. The ancestral Relaxin Family Loci (AncRFL) that are hypothesized to have been present in the common ancestor of teleosts and humans (tetrapods) are shown in the middle. We infer that AncRFLC duplicated giving rise to RFLCI (rln3) and RFLCII (insl3) prior to the divergence of teleosts and tetrapods. Names for the RFL proposed by Park et al. (2008) are given in brackets and underlined, those not used by Park et al., but inferred from this analyses are given in brackets with a dotted underline. The whole genome duplication (WGD) event resulted in two copies (paralogues) of each of the relaxin family genes in teleosts. AncRFLA gave rise to INSL5 and two paralogues in teleosts, insl5a and insl5b. AncRFLB was the predecessor of three human genes INSL4, INSL6 and RLN2, while the latter additionally underwent a recent duplication in primates producing RLN1. In teleosts, the RFLB gene, rln, is assumed to be orthologous to human RLN2. AncRFLC is hypothesized to have diverged into two loci: RFLCI harbouring RLN3 and the teleostean paralogues, rln3a and rln3b, and RFLCII, harbouring INSL3 and insl3. Duplicated copies of insl3 and rln in teleosts are believed to have been lost due to non-functionalization.

Figure 4

Alignment of the deduced amino acid relaxin sequences from mammals and teleost species used for the phylogenetic analysis. Conserved residues are boxed. Location of the relaxin receptor binding motif residues (RXXXRXXI/V), B-chain, A-chain, and twin dibasic junctions (B/C and C/A) are shown. Amino acids that are underlined are those identified as potential candidates of codon-specific positive selection using the branch-site model A analyses (see text for details), but only the two that are in bold and underlined were found to have a significant probability of being subject to positive selection with a BEB probability >0.95.

Figure 5

Phylogenetic reconstruction of the relationship among relaxin family DNA sequences. Phylogenetic tree reconstructed using the minimum evolution algorithm (a distance method) and including only the first two positions of each codon and employing the Tamura-3-parameter + Γ model of sequence evolution. Numbers at each node indicate the bootstrap values. Genes located at each of the four relaxin family loci, insl5 (RFLA), rln (RFLB), rln3 (RFLCI) and insl3 (RFLCII), are shown in the same colour. Paralogous copies of insl5 (insl5a and insl5b) and rln3 (rln3a and rln3b) that arose after the teleost WGD are indicated. Mammalian INSL6 is a tandemly duplicated member of the relaxin family that is linked and paralogous to mammalian RLN.

Figure 6

Relative fold increase (with standard deviation) in mRNA expression of five relaxin family genes in six tissues relative to their expression in eye and normalized by the expression of the housekeeping gene, b2m (see text for details). Relative expression of relaxin family genes rln3a, rln3b, insl3, insl5a and insl5b were assessed in the brain, gill, heart, gut, ovary and testis dissected from adult zebrafish.

Figure 7

Whole mount in situ hybridization of insl3 on zebrafish testis. A) Overview of the positive insl3 in situ hybridization signal in zebrafish testis, clearly showing positive insl3 in situ hybridization signal in the interstitial area. B) Detailed view of A, showing that only the cytoplasm of the Leydig cells in zebrafish testis shows the positive in situ hybridization signal. Blood vessels (encircled by dashes) containing erythrocytes are often visible in the Leydig cell clusters. The seminiferous tubules (ST), containing Sertoli cells and germ cells in different stages of spermatogenesis, remain completely unstained.