Figure 1.
Simplified phylogenetic tree showing the evolutionary relationships among the major deuterostome taxa (with the common names of organisms in brackets) and their ancestors discussed in this paper.
The hypothetical ancestral genome predicted by Nakatani et al. (“N”) probably belongs to an organism that existed just before 2R in early vertebrates. On the other hand, Putnam et al. (“P”) reconstructed the karyotype of the common ancestor of amphioxus and the olfactores (tunicates+vertebrates). 1R, 2R and 3R mark the three rounds of whole genome duplication that happened throughout vertebrate evolution. Tree topology adapted from Putnam et al [20].
Figure 2.
Reconstruction of the genetic events that led to the diversification of RXFP3-type receptors and RLN/INSL hormones in vertebrates based on the “fission” scenario of ancestral genome rearrangement.
The genomic origins of the hypothetical ancestral relaxin (AncRln-like) and Rxfp3/4 receptor (AncRxfp3/4) genes can be traced to a single chromosome in the vertebrate ancestor that had not yet been through 2R (Pre-2R vertebrate ancestor). The ancestral linkage group harboring AncRln-like and AncRxfp3/4-like sequentially underwent duplication, fission and another duplication yielding 5 distinct linkage groups (agnathan and gnathostome ancestor) harboring the ligand and receptor genes. Subsequently, tetrapods completely lost RXFP3-2 and often RXFP3-3 genes, but retained all of the post-2R ligand gene duplicates. Teleosts, on the other hand, retained both all of the ligand and receptor post-2R gene duplicates, suggesting that RXFP3-2 and RXFP3-3 acquired important functions in the pre-3R teleost ancestor. The duplicates of rxfp3-2 and rxfp3-3 were again retained in the post-3R teleost ancestor along with those of rln3 and insl5 (indicating their possible ligand-receptor relationships). Lastly, in vertebrates the RLN locus underwent multiple local duplications, resulting in the emergence of INSL4 in all eutherians, and INSL6 and RLN1 only in apes, whose RLN2 is orthologous to RLN of other eutherians. For simplicity, tetrapod and eutherian ancestor linkage groups are only shown to contain the fragments (e.g. A0, A2–A5) harboring the genes of interest; thus they should not be confused with actual chromosomes. Blue circles and squares represent receptor and their ligand genes respectively. Crossed circles represent pseudogenes (red, if they are verified in databases, blue if they are hypothetical). SSD: small-scale duplication. The first letter of ancestral gene names is capitalized.
Figure 3.
Reconstruction of the genetic events that led to the diversification of RXFP1/2-type receptor genes in vertebrates.
Symbols and linkage group numbering same as in Figure 2.
Table 1.
Explanation of the nomenclature used for the hypothetical ancestral genes that gave rise to the three gene families discussed in this study (receptors RXFP3/4, RXFP1/2 and their ligands RLN/INSL) via three rounds of WGD (1R, 2R and teleost-specific 3R).
Figure 4.
The evolution and genetic linkage of RLN/INSL (ligand) and RXFP3/4 (receptor) loci in the pre-3R teleost ancestor and three species of teleost fish.
Notice that among the three fish species analysed, medaka's genome and rln/insl-rxfp gene sets are the most preserved and resemble those of the teleost ancestor. Tetraodon experienced lineage-specific loss of two genes, rln3b and rxfp3-1, which may indicate their co-evolution as a ligand-receptor pair. The rxfp4 gene in zebrafish seems to have been replaced with an extra (zebrafish-specific) copy of an rxfp3-3 gene. The syntenic linkage between rln and insl3 (ligand) and rxfp3-1 and rxfp3-2(b) (receptor) genes has been conserved in all three teleosts since the post-2R ancestor (see Figure 2). Overall this scheme demonstrates that the rln/insl-rxfp system in teleosts has taken a slightly different, and seemingly more complicated, evolutionary pathway compared to other vertebrates. Chromosome numbers in extant species are shown as numbers and in the teleost ancestor as letters.
Figure 5.
Phylogenetic reconstruction of the evolutionary relationship among vertebrate RXFP3/4 protein sequences.
Reconstruction performed as outlined in methods with G = 0.91 and I = n/a. Numbers at each node indicate the bootstrap values (only values exceeding 50% shown). Teleost rxfp3-2 underwent duplication yielding two 3R-paralogs, rxfp3-2a and rxfp3-2b, while teleostean ancestral rxfp3-3 was duplicated giving rise to typically three rxfp3-3 loci in modern teleosts: 3R generated rxfp3-3a and rxfp3-3b, while a local duplication generated rxfp3-3a1 and rxfp3-3a2. Solely in zebrafish, rxfp3-3a2 duplicated again giving rise to rxfp3-3a3, an event which appears to have occurred coincidently with the exclusive loss of rxfp4 in zebrafish.
Figure 6.
Phylogenetic reconstruction of the evolutionary relationship among vertebrate RXFP1/2 protein sequences.
Phylogenetic tree reconstructed as outlined in methods with G = 0.958 and I = 0.034. Numbers at each node indicate the bootstrap values (only values exceeding 50% shown). Due to their incomplete nature, not all sequences from our created database (see Methods) were included in this tree (e.g. zebrafish rxfp2a and rxfp2b and medaka rxfp2).