Fig 1.
Gene-inactivating mutations in RXFP2 in four afrotherian species.
(A) The exon-intron structure of the coding region of the RXFP2 gene is shown as boxes (exons, drawn to scale) and lines (introns, not drawn to scale). A vertical red line/arrowhead indicates a frameshifting deletion/insertion, with the number of deleted/inserted bases given above. Stop codon mutations are shown as a black vertical line. Splice site mutations are indicated by the mutated dinucleotide. A blue vertical line indicates a frame-preserving deletion. Red boxes are exons that are either deleted or accumulated numerous mutations that destroy any sequence similarity. All inactivating mutations were validated by unassembled genome sequencing reads stored in the SRA. Elephant, rock hyrax, and aardvark have an intact RXFP2 gene and are not shown. A filled star indicates mutations that we confirmed by PCR and Sanger sequencing in the lesser hedgehog tenrec; the exon 17 frameshift was also found in the greater hedgehog tenrec (S5A and S5B Fig). An open star indicates mutations that we confirmed by PCR and sequencing in the dugong, the sister species of the manatee (S5C and S5D Fig). (B-I) Examples of inactivating mutations and their validation by unassembled SRA reads. RXFP2, relaxin/insulin-like family peptide receptor 2; SRA, Sequence Read Archive.
Fig 2.
Gene-inactivating mutations and critical amino acid mutations in INSL3.
(A) Functional domains of the INSL3 protein and the exon-intron structure of the INSL3 gene. Inactivating mutations are as in Fig 1, frame-preserving insertions/deletions are shown as blue lines/arrowheads. Exons but not introns are drawn to scale. Elephant, rock hyrax, and aardvark have an intact INSL3 and are not shown. (B) While the cape golden mole does not exhibit any gene-inactivating mutations (A), an INSL3 protein alignment of the A- and B-chain shows mutations (red background) at amino acids that are critical for structure and function of the mature hormone (gray background) [46–48]. Disulfide bonds between Cys residues are indicated by blue lines. Residues that are affected by frameshifting deletions in the underlying tenrec or cape elephant shrew nucleotide sequence are indicated by asterisks. For these two species, we ignored these frameshifts and used the ancestral reading frame. Species in red font are testicond. Note that elephant and rock hyrax have no mutations at any of the critical sites. INSL3, insulin-like 3.
Fig 3.
RXFP2 and INSL3 are only lost in several testicond afrotherian lineages.
The exon-intron structures of RXFP2 (18 coding exons) and INSL3 (2 coding exons) are shown. A yellow rectangle indicates an intact exon. Exons with inactivating mutations (stop codon or splice site mutations, frameshifts) are indicated in orange; completely deleted or highly diverged exons are indicated in red. Gray exons could not be examined because the respective genome assembly is incomplete at this locus (assembly gap). We could not examine the RXFP2 in the fragmented pangolin genome, because different parts of RXFP2 locus align to at least 5 different short scaffolds. Neither exons nor introns are drawn to scale. INSL3, insulin-like 3; RXFP2, relaxin/insulin-like family peptide receptor 2.
Fig 4.
Evolution of RXFP2 and estimated time intervals during which RXFP2 came under neutral evolution.
Red boxes represent a lower and upper bound for an estimated time interval during which RXFP2 started to evolve neutrally (Materials and methods, S2 Table), which shows that the gene was lost independently at different time points. Gray boxes represent an estimated interval for species divergence times taken from TimeTree [57] (see also S4 Table). Species in blue font are testicond. The phylogenetic position of tenrecs, golden moles, elephant shrews is well supported by morphological and molecular characters [23, 24, 26]. Uncertain phylogenetic relationships are shown as a polytomy. Note that while the loss of RXFP2 in the elephant shrew is estimated to have happened at the base of the lineage, this loss is clearly independent from the much later loss in the tenrec and in the golden mole lineage. Evidence for the loss of RXFP2 in the dugong and the greater hedgehog tenrec was obtained by PCR and sequencing experiments (S5 Fig). RXFP2, relaxin/insulin-like family peptide receptor 2.