- Open Access
Independent pseudogenizations and losses of sox15 during amniote diversification following asymmetric ohnolog evolution
BMC Ecology and Evolution volume 21, Article number: 134 (2021)
Four ohnologous genes (sox1, sox2, sox3, and sox15) were generated by two rounds of whole-genome duplication in a vertebrate ancestor. In eutherian mammals, Sox1, Sox2, and Sox3 participate in central nervous system (CNS) development. Sox15 has a function in skeletal muscle regeneration and has little functional overlap with the other three ohnologs. In contrast, the frog Xenopus laevis and zebrafish orthologs of sox15 as well as sox1-3 function in CNS development. We previously reported that Sox15 is involved in mouse placental development as neofunctionalization, but is pseudogenized in the marsupial opossum. These findings suggest that sox15 might have evolved with divergent gene fates during vertebrate evolution. However, knowledge concerning sox15 in other vertebrate lineages than therian mammals, anuran amphibians, and teleost fish is scarce. Our purpose in this study was to clarify the fate and molecular evolution of sox15 during vertebrate evolution.
We searched for sox15 orthologs in all vertebrate classes from agnathans to mammals by significant sequence similarity and synteny analyses using vertebrate genome databases. Interestingly, sox15 was independently pseudogenized at least twice during diversification of the marsupial mammals. Moreover, we observed independent gene loss of sox15 at least twice during reptile evolution in squamates and crocodile-bird diversification. Codon-based phylogenetic tree and selective analyses revealed an increased dN/dS ratio for sox15 compared to the other three ohnologs during jawed vertebrate evolution.
The findings revealed an asymmetric evolution of sox15 among the four ohnologs during vertebrate evolution, which was supported by the increased dN/dS values in cartilaginous fishes, anuran amphibians, and amniotes. The increased dN/dS value of sox15 may have been caused mainly by relaxed selection. Notably, independent pseudogenizations and losses of sox15 were observed during marsupial and reptile evolution, respectively. Both might have been caused by strong relaxed selection. The drastic gene fates of sox15, including neofunctionalization and pseudogenizations/losses during amniote diversification, might be caused by a release from evolutionary constraints.
In mammals, the Sex-determining region Y (Sry)-type high mobility group (HMG) box (Sox) family of genes includes approximately 20 members. The family is divided into eight groups (A–H) based on sequence identity in the DNA-binding domain HMG box [1, 2]. Interestingly, groups A and G contain only one member, Sry and Sox15, respectively. In contrast, most Sox groups are comprised of three closely related genes. Notably, Sry and sox15 may have diverged independently from the sox3 orthologs in common ancestors of therian mammals and vertebrates, respectively [3, 4]. Sox3 shares higher sequence identity with Sox1 or Sox2 than Sox15 in mammals, with group B1 comprising three members (Sox1, Sox2, and Sox3) [1, 5]. Vertebrate ohnologs are paralogs generated through the two rounds of whole-genome duplication (2R-WGD) in the common ancestor of vertebrates [6,7,8]. Synteny analysis data revealed ohnologous relationships for these four members. The two ancestral genes sox1/2 and sox3/15 emerged from soxB1 in the common ancestor of vertebrates in the first round of WGD, followed by the generation of these four genes in the second round of WGD . Mammalian Sox15 has orthologous relationships with teleost sox19a/b and amphibian soxD (now termed sox15), although sox19a/b and soxD/sox15 belong to groups B1 and I, respectively . The pairwise sequence identities between sox15, sox19a/b, and soxD/sox15 are not particularly high, which likely resulted in the different gene names conferred during an era of limited genome information.
Both zebrafish sox19a/b and the African clawed frog Xenopus laevis soxD/sox15 have high expression levels and function in central nervous system (CNS) development [4, 9, 10]. There are, to our knowledge, no reports of mammalian sox15 involvement in neurogenesis. In contrast, mammalian Sox15 is expressed in embryonic stem cells and satellite cells involved in muscle regeneration [11, 12]. Moreover, we described the specific expression of Sox15 in the placenta during mouse embryogenesis in placenta-derived trophoblast giant cells and placental stem cells, and demonstrated that Sox15 became pseudogenized in the marsupial opossum [5, 13, 14]. Based on the findings of Sox15 orthologs in mammals, actinopterygian (bony) fish and amphibians, we concluded that Sox15 evolved through neofunctionalization for placental formation and/or myogenesis in eutherian diversification . Therefore, there appears to be little functional overlap between sox15 and sox19a/b or soxD/sox15 orthologs.
The collective findings indicate that sox15 orthologs appear to have evolved dramatically . However, little is known about sox15 in birds, reptiles, non-eutherian mammals, non-anuran amphibians, and chondrichthyan fishes. To clarify the molecular evolution of sox15 orthologs, we retrieved orthologous sequences of sox15 from genome databases of amniote vertebrates, urodele/apoda amphibians, and chondrichthyan fishes, and performed synteny analysis. Notably, we identified independent losses and pseudogenizations of sox15 in reptiles, including birds and marsupial mammals, respectively. The fates of sox15 with independent pseudogenization/loss and neofunctionalization represent a rare example of gene fate during vertebrate evolution, although bmp16 was recently reported as an example of independent gene loss among the three ohnologs [15, 16]. Our evolutionary analyses revealed the highest dN/dS value of sox15 among the four ohnologs and its independent pseudogenizations/losses during amniote diversification.
Independent pseudogenizations of Sox15 during marsupial mammalian evolution
We previously reported that Sox15 is a pseudogene in the marsupial opossum, although Sox15 has a function in eutherian mammals [13, 14]. To examine whether pseudogenization of Sox15 occurred in the common ancestor of marsupials, we searched for Sox15 orthologs in seven other marsupial genomes using genome browser and tblastn analysis using the amino acid sequences encoded by eutherian Sox15 (Additional file 1: Table S1) and performed synteny analysis using the Fxr2 and Mpdu1 genes adjacent to Sox15. Annotated orthologs of Sox15 or Sox15-like sequences were identified between Fxr2 and Mpdu1 in the genomes of all seven species. The Sox15-like sequence of Tammar wallaby (Macropus eugenii) did not contain its HMG box-coding region, which was likely caused by sequencing gaps. In contrast, the Sox15-like sequences of both Tasmanian devil (Sarcophilus harrisii) and fairy possum (Gymnobelideus leadbeateri) contained pseudogenization signatures evident as in-frame stop codons and frame shift mutations in the HMG box-encoding regions (Figs. 1a, 2a, and Additional file 1: Fig. S1). However, the Sox15-like sequence of the other four species—common brushtail possum (Trichosurus vulpecula), thylacine (Thylacinus cynocephalus), koala (Phascolarctos cinereus), and wombat (Vombatus ursinus)—contained an intact 75 amino acid HMG box-encoding region (Figs. 1a, 2a, and Additional file 1: Fig. S1). Importantly, the following two observations indicated that the pseudogenizations of the Sox15 orthologs must have arisen independently in the three distinct ancestors of Tasmanian devil, fairy possum, and opossum. First, the sequences adjacent to the in-frame stop codon and frameshift mutation sites differed from each other. Second, each order of marsupials, Dasyuromorphia (green in Fig. 2a) containing Tasmanian devil and thylacine and Phalangeriformes (yellow in Fig. 2a) containing the fairy possum and common brushtail possum, harbored both pseudogenized and non-pseudogenized Sox15 genes (Fig. 2a). These results indicate the independent pseudogenizations of Sox15 during marsupial divergence.
Absence of pseudogenizations/losses of Sox15 in both eutherian and monotreme mammals
We previously reported the potential neofunctionalization role of mouse Sox15 in placental development [5, 13, 14]. In addition, Lee et al.  reported the involvement of Sox15 in myogenesis. Based on our finding of Sox15 pseudogenization during marsupial divergence, we searched for Sox15 orthologs in NCBI gene databases from as many different eutherian mammalian genomes as possible. Annotation for Sox15 was found in all the 148 genomes examined. Synteny of the Sox15 orthologs in nine of 148 species is shown in Fig. 1a. All 148 Sox15 sequences contained predicted open reading frames (ORFs). In addition, all the eutherian mammals, except polar bear (Ursus maritimus), harbored Sox15 between Mpdu1 and Fxr2 (Fig. 1a). We discuss later whether the Sox15 ortholog exists in U. maritime. To clarify the fate of Sox15 in monotreme mammals, we also searched for Sox15 in the platypus and echidna genomes and found Sox15 annotations between Mpdu1 and Fxr2 (Fig. 1a).
Independent gene losses of sox15 during reptilian evolution
Independent pseudogenizations of sox15 during marsupial evolution were identified (Fig. 1a). Next, we searched for sox15 orthologs in reptilian lineages using the amino acid sequence encoded by sox15 from Chinese softshell turtle (Pelodiscus sinensis) or the common wall lizard (Podarcis muralis). The search revealed the amino acid sequences of sox15 and its adjacent fxr2 and mpdu1 genes as tblastn queries from 34 squamates (lizard and snake), 22 testudine (turtle and tortoise), four crocodilians, and 503 bird species (Additional file 1: Table S1).
In the 34 squamates genomes, the tblastn hits and synteny analysis identified sox15 orthologs in 17 species from five different lineages, including Teiidae, Lacertidae, Agamidae, Viperidae, and Colubridae. No sox15-like sequences were found in the other 17 species (Fig. 2b and Additional file 1: Table S2). Blastn-based synteny analysis using Easyfig  indicated that sox15 was lost in nine of the 17 species from four different lineages: Gekkonidae, Varanidae, Pythonidae, and Elapidae (Figs. 1b, 2b, Additional file 1: S3a–o and Table S2). We categorized the other eight as “unknown” because fxr2 and/or mpdu1 orthologs were not detected (Fig. 2b and Additional file 1: Table S2).
Among the 22 testudine species, sox15 orthologs were identified in 17 species by the tblastn search and synteny analysis. There were no sox15-like sequences in the other five species (Fig. 2c and Additional file 1: Table S2). Synteny analysis with Easyfig indicated that the nine of the 11 species belonging to the Testudinoidea superfamily had sox15-like sequences between fxr2 and mpdu1 orthologs, whereas the Pinta Island giant tortoise (Chelonoidis abingdonii) had no sox15-like sequence between fxr2 and mpdu1 (Fig. 2c, Additional file 1: Fig. S3r–y, and Table S2). In the other four species, sox15 was categorized as “unknown” because fxr2 and/or mpdu1 orthologs were not detected (Fig. 2c and Additional file 1: Table S2). In the four crocodilian genomes, no sox15-like sequences were found in the tblastn search (Fig. 2c, Additional file 1: Fig. S3p, q, and Table S2).
Synteny analysis with Easyfig indicated that sox15 was lost in Alligator sinensis and Crocodylus porosus belonging to the Alligatoridae and Crocodylidae families, respectively (Fig. 1, Fig. 2c and Additional file 1: Fig. S3p, q). The other two species were categorized as “unknown”, because fxr2 and/or mpdu1 orthologs were not detected (Fig. 2c). Neither sox15 nor fxr2 orthologs were found in 503 bird genomes in the expected chromosomal regions (Fig. 2c and Additional file 1: Table S2), indicating that the sequences including the two genes must have been deleted in the ancestor of birds.
Collectively, these findings revealed independent gene losses of sox15 at least twice during species diversity of squamates and crocodile-birds on reptile evolution.
Absence of gene losses/pseudogenizations of sox15 in the examined amphibian and actinopterygian fish genomes
Orthologous relationships have been described among mammalian sox15, teleost fish sox19a/b, and frog soxD/sox15 . Thus, we examined the fate of sox15 orthologs in four amphibian species and 12 actinopterygian fish species. In amphibians, there were sox15 orthologs between fxr2 and mpdu1 in the two caecilian species, Geotrypetes seraphini and Rhinatrema bivittatum, and between ephb4 and mpdu1 in the two frog species, X. tropicalis and X. laevis (Fig. 1c). In actinopterygian fishes, nine out of ten teleost species had sox15 orthologs between mpdu1 and eif4a1, and Takifugu rubripes ortholog of sox15 was located on the same chromosome as mpdu1 and eif4a1 were (Fig. 1c). We also found sox15 orthologs between fxr2 and mpdu1 in two nonteleost fish species, Acipenser ruthenus and Erpetoichthys calabaricus.All 16 genomes harbored sox15 orthologs containing the predicted ORF (Fig. 1c).
Absence of introns in cartilaginous fish sox15
Next, we examined the fate of sox15/sox19 in cartilaginous fish (Chondrichthyes) and searched for sox15 orthologs in eight elasmobranch and two holocephalan genomes (Additional file 1: Table S1) using the amino acid sequence of ropefish Erpetoichthys calabaricus SOX19 as a query. We found sox15-like sequences in three elasmobranch species (Pristis pectinata, Chiloscyllium punctatum, and C. plagiosum), but not in the two holocephalan species. The genome information of the other five elasmobranch species contained no sox15-like or adjacent mpdu1-like sequences. Synteny analysis of the elasmobranch sox15-like sequences based on FGNENESH gene annotation and phylogenetic tree reconstruction of soxB1/G ohnologs indicated that the P. pectinata and C. plagiosum sequences should correspond to sox15 orthologs (Fig. 1c and Additional file 1: Fig. S4). In addition, the C. punctatum transcriptome database (https://transcriptome.riken.jp/squalomix/blast/) revealed that the sox15 gene could be transcribed (Additional file 1: Fig. S5).
Notably, a sequence comparison between the sox15 transcript and its genomic region in C. punctatum indicated that the sox15 gene consists of one exon, although almost all the sox15 orthologs in mammals, reptiles, amphibians, and actinopterygian fish examined in this study consisted of two exons with the same splicing sites. In contrast, no jawless fish orthologs of sox15 were identified in the three species, Petromyzon marinus, Lethenteron camtschaticum, and Eptatretus burgeri.
sox15 has the highest d N/d S value among the four ohnologs
Independent pseudogenizations and losses of sox15 were evident only in amniote divergence and not in anamniotes (Figs. 1, 2, and Additional file 1: Table S2). Additionally, Sox15 could acquire new roles in skeletal muscle and placental development through neofunctionalization in eutherian diversification . Why did drastic gene fates, including gene losses/pseudogenizations and neofunctionalization, occur in sox15 during amniote divergence? To answer this question, we constructed phylogenetic trees of four soxB1/G ohnologous proteins—SOX1, SOX2, SOX3, and SOX15—in jawed vertebrates using maximum likelihood and Bayesian methods (Additional file 1: Fig. S4). Each tree showed a relatively longer branch length in the SOX15 clade than in SOX1, SOX2, or SOX3. Interestingly, longer branch lengths of SOX15 (SOX19) in amniotes and amphibians were observed compared to the lengths in actinopterygian fish and those of SOX1, SOX2, and SOX3 in jawed vertebrates. These findings suggest that there might be different selective pressures on the molecular history of sox15 by evolutionary lineages in vertebrates, or in contrast to the other three ohnologs. To explore this idea we reconstructed a codon-based phylogenetic tree of jawed vertebrate soxB1/G ohnologs and calculated the ratio of nonsynonymous and synonymous substitution (dN/dS; ω) values for several vertebrate lineages (Fig. 3 and Additional file 1: Fig. S6). The ω values of sox1, sox2, sox3, and sox15 in jawed vertebrates were 0.0093, 0.0046, 0.0083, and 0.0204, respectively (Fig. 3). Likelihood ratio tests revealed that the ω value of sox15 significantly deviated from those of the other three soxB1/G ohnologs (p = 7.9 × 10–4, 1.7 × 10–5, and 4.9 × 10–4 for sox1, sox2, and sox3, respectively; Additional file 1: Table S4). We next divided jawed vertebrates into three classes—chondrichthyes (cartilaginous fishes), actinopterygii (bony fishes), and sarcopterygii (lobe-finned fish and tetrapod species), and calculated the ω value of sox15. Importantly, actinopterygian sox15 had a lower ω value (0.0063) than that of chondrichthyes (0.0345) and sarcopterygii (0.0392), which was statistically supported by χ2 tests (p = 9.3 × 10–15 and 3.0 × 10–15, respectively; Additional file 1: Table S4 and Fig. 3). Because the sarcopterygian class had different evolutionary distances of sox15 in the codon-based phylogenetic tree, we further divided the class into three groups—coelacanth and non-anuran amphibians (Sar1), anuran amphibians (Sar2), and amniotes (Sar3)—and examined each ω value of sox15 from a node of sarcopterygian common ancestors. These results and χ2 test results revealed that the ω value of Sar3 was statistically significantly higher (0.0533) than those of Sar1 (0.0173) and Sar2 (0.0261) (p = 6.5 × 10–4 and 0.018, respectively; Additional file 1: Table S4 and Fig. 3). These findings indicate that the higher ω value of sox15 than sox1, sox2, and sox3 could be mainly affected by those of sox15 in cartilaginous fishes and amniotes.
Relaxed selection has participated in an asymmetric evolution of sox15 among the four ohnologous soxB1/G members and is involved in pseudogenizations/losses of sox15
To elucidate whether the molecular evolution of sox15 is under relaxed or intensified selection, we performed RELAX tests  of the gene in three classes of jawed vertebrates: chondrichthyes, actinopterygii, and sarcopterygii. sox3 was used as a reference branch to calculate the relaxation or intensification parameter (k) of sox15 in each class. The RELAX test data (Table 1) revealed that sox15 might have evolved under relaxed selection in chondrichthyes and sarcopterygii (k = 0.25 [p = 3.1 × 10–4] and 0.40 [2.0 × 10–4], respectively), and even in all jawed vertebrates (k = 0.72, p = 8.0 × 10–4). Although the k value of actinopterygian sox15 was also lower than 1 (k = 0.75), the χ2 test did not dismiss the null hypothesis of ‘k = 1’, suggesting that sox15 and sox3 have similarly evolved under purifying selection in bony fishes. These results indicate that the relaxed selection might have contributed to the higher ω value of sox15 than that of sox3 during cartilaginous fish and sarcopterygian evolution, and induced an asymmetric evolution of sox15 among the four ohnologous soxB1/G members during jawed vertebrate evolution.
The three soxB1 subfamily genes (sox1, sox2, and sox3) and sox15 share ohnologous relationships in vertebrates, but only sox15 orthologs do not belong to the soxB1 subfamily . In this study, we examined the fate of the unique ohnologous member sox15 during vertebrate evolution. Figure 4 summarizes this study and represents proposed model for molecular evolution of sox15 and its ohnologous members sox1-3 s during vertebrate evolution. Although we found sox15-like sequences in the three jawless vertebrate genomes, we could not conclude that those are sox15 orthologs by synteny and phylogenetic analyses. In contrast, within the cartilaginous fish, we identified sox15 orthologs in three elasmobranchs, P. pectinata, C. punctatum, and C. plagiosum, but could not find orthologs to sox15 or any of its flanking genes in two holocephalans, Callorhinchus milii and Hydrolagus affinis. It is possible that sox15 and its adjacent genes were deleted in the ancestor of holocephalan fishes during cartilaginous fish evolution. Interestingly, we found an intron-free sox15 in C. punctatum genomes (Figs. 1c, 4, and Additional file 1: Table S1). All other orthologs of sox15 examined in this study consisted of two exons. Because vertebrate sox1, sox2, and sox3 are single exon genes, Okuda et al.  reported that sox15 should have acquired an intron during vertebrate evolution after or during the second round of WGD. Our results suggest that sox15 acquired an intron in the ancestor of bony (Osteichthyes) fish after the divergence between cartilaginous and bony fish (Fig. 4). It will be interesting to clarify whether the spread of the intron-containing sox15 within the ancestral population of bony fish might be neutrally or positively selected.
We previously reported the neofunctionalization and pseudogenization of Sox15 in eutherian mice and marsupial opossums, respectively [5, 13, 14]. Our synteny analysis revealed that all the eutherian mammals examined except for the polar bear (U. maritimus), have Sox15 between Mpdu1 and Fxr2 (Fig. 1a). We found only the exon 2 sequence of Sox15 between Mpdu1 and Fxr2 in the U. maritimus genome database; there was a 1018 bp gap upstream of exon 2. It is conceivable that this gap could hide exon 1 of Sox15. These findings suggest the presence of Sox15 for neofunctionalization in the common ancestor of eutherian mammals, resulting in no or almost no pseudogenizations/losses of Sox15 during eutherian evolution (Fig. 4). Interestingly, sox15 was independently pseudogenized during marsupial evolution and lost at least twice during reptile evolution (Figs. 1, 2, and 4). Our search revealed that the marsupial koala and wombat orthologs of Sox15 were annotated among Mpdu1 and Fxr2. These predicted amino acid sequences from the two orthologs comprising two exons and one intron shared high sequence identities with those of eutherian mice (Additional file 1: Fig. S2). In addition, their mRNA expression in both koala and wombat was confirmed by RNA-seq alignments, including 4 and 13 samples, respectively (https://www.ncbi.nlm.nih.gov/nuccore/XM_020966486.1; https://www.ncbi.nlm.nih.gov/nuccore/XM_027857251.1). Importantly, the koala and wombat sox15 mRNAs have ORFs, encoding putative 228 and 216 amino acid sequences, respectively, which could function as SOX15. In contrast, there have been no reports of losses/pseudogenizations of sox1, sox2, or sox3. Why did only sox15 orthologs have drastic gene fates during amniote divergence among the four ohnologs? The selection analysis revealed the highest dN/dS value of sox15 among the four ohnologs during jawed vertebrate evolution. In addition, the amniote dN/dS value of sox15 was the highest among the five classes (chondrichthyes, actinopterygii and Sar1-3) in jawed vertebrates (Fig. 3). These findings suggest that the relatively high dN/dS value could be connected to the divergent gene fates of sox15 during amniote divergence.
What is the difference in genetic backgrounds between independent pseudogenization and gene losses of sox15 in marsupials and reptiles? Some inversions and deletions appeared to happen on the chromosomal regions around sox15 during reptile-bird evolution (Fig. 1b, Additional file 1: Fig. S2), while there were almost no chromosomal inversions around sox15 in mammals, including marsupials (Fig. 1a). It is possible that the deletions or inversions followed by deletions on the chromosomal region around sox15 happened to result in independent losses of sox15 during reptile diversification.
We could not identify any jawless fish orthologs of sox15, but found a soxB1-like gene in the NCBI gene database of sea lamprey (P. marinus) (Gene ID: 116956344). In addition, there was no synteny between this gene and any soxB1/G member, and the predicted amino acid sequence from the gene did not belong to any clade of SOX1, SOX2, SOX3, or SOX15 in jawed vertebrates (Additional file 1: Fig. S4). Because the common ancestor of agnathans underwent genome duplication , it would be interesting to investigate why only one soxB1 gene remained in sea lamprey and how the gene evolved.
In general, ohnologs exhibit shared expression and functional redundancy prior to genome duplication, even in allotetraploidization by hybridization between two related species, as suggested by findings from the frog Xenopus laevis . An additional WGD (3R-WGD) also occurred in the teleost ancestor . Seebrafish sox19a and sox19b, which are co-orthologous to sox15, are presumptually derived from 3R-WGD . We confirmed two sox19 genes in three teleost species including zebrafish (Fig. 1c). However, we could not find one of two presumptive sox19 genes or any of its flanking genes in other five teleost fish. Then why did two copies of sox19 remain in the three teleost fish? sox1, sox2, sox3, and sox15 were predominantly expressed in the developing CNS in non-amniotes [4, 9, 21,22,23]. Zebrafish sox19a and sox19b also showed specific expression in the developing CNS [4, 10]. It is possible that expression redundancy in the CNS has been involved in the retention of the soxB1/G members including teleost sox19a and sox19b.
It is believed that cis-element evolution resulted in the differentiation of expression patterns between ohnologs . Mutations in the cis-regulatory elements for the CNS expression of sox15 might cause the gene to escape from the spatio-temporal expression regime the other three ohnologs have been subjected to. Then molecular evolution of sox15 under slightly relaxed purifying selection during amniote diversification might be involved in the divergent gene fates of sox15, neofunctionalization in the ancestor of eutherian mammals, pseudogenization during marsupial diversification, and losses during reptile diversification. It has been proposed that the fate of duplicated genes could be affected by the intrinsic properties of genomic regions harboring them as well as functional constraints on their roles . Then another possibility is that the instability of the genomic region during amniote evolution has caused relatively high mutation rates of both the cis-regulatory elements and the coding sequence of sox15, resulting in the drastic gene fates.
The evolutionary analyses revealed independent pseudogenizations and losses of sox15 during marsupial mammalian and reptile evolution, respectively, although no pseudogenizations/losses of sox15 were observed in the other classes of jawed vertebrates. sox15 showed the highest dN/dS value among the four ohnologous soxB1/G members, and higher dN/dS values in marsupials and reptiles than in eutherians. Moreover, we found that relaxed selection has been involved in asymmetric evolution of sox15 among the four ohnologs, which might have been one of the factors for the sox15’s drastic gene fates including pseudogenizations, losses, and neofuctionalization during amniote diversification. We propose that sox15 might have been released from some evolutionary constraints including expression and/or functional redundancy in the CNS, unlike sox1, sox2, and sox3 in the common ancestor of amniotes.
Identification of sox15 orthologs from genome databases
Genome sequences from 571 vertebrate species were downloaded in FASTA format from NCBI (https://www.ncbi.nlm.nih.gov/). Local databases were created using BLAST (v.2.9.0 + ; https://blast.ncbi.nlm.nih.gov/Blast.cgi?CMD=Web&PAGE_TYPE=BlastD ocs&DOC_TYPE = Download). Gene searches using tblastn (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PROGRAM=tblastn&PAGE_TYPE=BlastSearch&LINK_LOC=blasthome)  were performed using the default option (task = tblastn; evalue = 10; db_gencode = 1; max_intron_length = 0; matrix = BLOSUM62; comp_based_stats = 2). For marsupials, XP_020822145.1 (Phascolarctos cinereus SOX15), XP_020822143.1 (Phascolarctos cinereus MPDU1), and XP_020822141.1 (Phascolarctos cinereus FXR2) were used for query amino acid sequences. For squamates, XP_028559641.1 (Podarcis muralis SOX15), XP_028558914.1 (Podarcis muralis MPDU1), and XP_028557780.1 (Podarcis muralis FXR2) were used for query amino acid sequences. For Testudines, XP_014432453.1 (Pelodiscus sinensis SOX15), XP_025043921.1 (Pelodiscus sinensis MPDU1), and XP_006112154.1 (Pelodiscus sinensis FXR2) were used for query amino acid sequences. For crocodilians, XP_014432453.1 (Pelodiscus sinensis SOX15), XP_025049549.1 (Alligator sinensis MPDU1), and XP_025049532.1 (Alligator sinensis FXR2) were used for query amino acid sequences. For birds, XP_014432453.1 (Pelodiscus sinensis SOX15), XP_032940188.1 (Catharus ustulatus MPDU1), and XP_025049532.1 (Alligator sinensis FXR2) were used for query amino acid sequences. For Chondrichthyes, XP_028654266.1 (Erpetoichthys calabaricus SOX19), XP_028654265.1 (Erpetoichthys calabaricus MPDU1), and XP_028654267.1 (Erpetoichthys calabaricus FXR2) were used for query amino acid sequences. Sequence similarity of the top hit sequence to the SOX15 sequence in the database of the most similar species was analyzed using megablast (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PROGRAM=blastn&PAGE_TYPE=BlastSearch&LINK_LOC=blasthome) using BLAST, followed by manual synteny analysis. Gene annotation of sox15 orthologs was performed with Softberry-FGENESH (http://www.softberry.com/berry.phtml?topic=case_study_animal&no_menu=on)  based on a species-specific gene-finding parameter [28, 29].
Molecular phylogenetic analysis
The coding sequences of sox1, sox2, sox3, and sox15 in various vertebrate species were translated to amino acid sequences. Multiple alignment of results was performed using MAFFT version 7.427 (https://mafft.cbrc.jp/alignment/software/) . Multiple alignments of nucleotide sequences were deduced using Pal2nal version 14 (http://www.bork.embl.de/pal2nal)  with the amino acid alignments. Ambiguous sites of the alignments were removed using trimAl version 1.2 (http://trimal.cgenomics.org/)  with option-gappyout. Maximum likelihood trees were inferred by IQ-TREE (http://www.iqtree.org/) , where the most fitting amino acid and nucleotide substitution rate was estimated by ModelFinder in part of analysis by IQ-TREE . A Bayesian phylogenetic tree was inferred by Mrbayes 3.2.7a (https://nbisweden.github.io/MrBayes/download.html) . Two MCMC chains were run 300,000 times and sampled 100 times to analyze the convergence of the statistics by Tracer version 1.7.1 (http://beast.community/tracer). dN/dS was estimated by codeml in Paml4.8 (http://abacus.gene.ucl.ac.uk/software/paml.html) . A branch model was used to calculate the dN/dS ratio for each group. RELAX  in HYPHY version 2.5.8 (https://github.com/veg/hyphy) was used to detect the relaxed or intensified selection for the sox15 ortholog in each lineage.
χ2 tests were performed using Microsoft Excel for Mac version 16.42 for LRT value in codeml and RELAX analysis.
Availability of data and materials
All data generated or analyzed during this study are included in the main manuscript, figures, tables, and supplementary information file. Raw data of phylogenetic tree inference and analyses of codeml and RELAX are available in the figshare repository, https://doi.org/10.6084/m9.figshare.14400617.v1.
Central nervous system
Sex-determining region Y
High mobility group
- Sox :
Sex-determining region Y (Sry)-Type high mobility group (HMG) box
Two rounds of whole-genome duplication
- d N/d S :
Ratio of nonsynonymous substitution sites per synonymous substitution site
Likelihood ratio tests
Wegner M. From head to toes: the multiple facets of Sox proteins. Nucleic Acids Res. 1999;27:1409–20.
Bowles J, Schepers G, Koopman P. Phylogeny of the SOX family of developmental transcription factors based on sequence and structural indicators. Dev Biol. 2000;227:239–55.
Foster JW, Marshall Graves JA. An SRY-related sequence on the marsupial X chromosome: implications for the evolution of the mammalian testis-determining gene. Proc Natl Acad Sci U S A. 1994;91:1927–31.
Okuda Y, Yoda H, Uchikawa M, Furutani-Seiki M, Takeda H, Kondoh H, et al. Comparative genomic and expression analysis of group B1 sox genes in zebrafish indicates their diversification during vertebrate evolution. Dev Dyn. 2006;235:811–25.
Ito M. Function and molecular evolution of mammalian Sox15, a singleton in the SoxG group of transcription factors. Int J Biochem Cell Biol. 2010;42:449–52.
Ohno S. Evolution by gene duplication. 1970.
Holland PWH, Garcia-Fernandez J, Williams NA, Sidow A. Gene duplications and the origins of vertebrate development. Development. 1994;120(SUPPL):125–33.
Dehal P, Boore JL. Two rounds of whole genome duplication in the ancestral vertebrate. PLoS Biol. 2005;3:e314.
Mizuseki K, Kishi M, Shiota K, Nakanishi S, Sasai Y. SoxD: an essential mediator of induction of anterior neural tissues in Xenopus embryos. Neuron. 1998;21:77–85.
Okuda Y, Ogura E, Kondoh H, Kamachi Y. B1 SOX coordinate cell specification with patterning and morphogenesis in the early zebrafish embryo. PLoS Genet. 2010;6:36.
Lee H-J, Goring W, Ochs M, Muhlfeld C, Steding G, Paprotta I, et al. Sox15 is required for skeletal muscle regeneration. Mol Cell Biol. 2004;24:8428–36.
Maruyama M, Ichisaka T, Nakagawa M, Yamanaka S. Differential roles for Sox15 and Sox2 in transcriptional control in mouse embryonic stem cells. J Biol Chem. 2005;280:24371–9.
Yamada K, Kanda H, Tanaka S, Takamatsu N, Shiba T, Ito M. Sox15 enhances trophoblast giant cell differentiation induced by Hand1 in mouse placenta. Differentiation. 2006;74:212–21.
Yamada K, Kanda H, Aihara T, Takamatsu N, Shiba T, Ito M. Mammalian Sox15 gene: promoter analysis and implications for placental evolution. Zoolog Sci. 2008;25:313–20.
Feiner N, Meyer A, Kuraku S. Evolution of the vertebrate Pax4/6 class of genes with focus on its novel member, the Pax10 gene. Genome Biol Evol. 2014;6:1635–51.
Feiner N, Motone F, Meyer A, Kuraku S. Asymmetric paralog evolution between the “cryptic” gene Bmp16 and its well-studied sister genes Bmp2 and Bmp4. Sci Rep. 2019;9:1–13. https://doi.org/10.1038/s41598-019-40055-1.
Sullivan MJ, Petty NK, Beatson SA. Easyfig: a genome comparison visualizer. Bioinformatics. 2011;27:1009–10.
Delsuc F, Philippe H, Tsagkogeorga G, Simion P, Tilak MK, Turon X, et al. A phylogenomic framework and timescale for comparative studies of tunicates. BMC Biol. 2018;16:1–14.
Smith JJ, Kuraku S, Holt C, Sauka-Spengler T, Jiang N, Campbell MS, et al. Sequencing of the sea lamprey (Petromyzon marinus) genome provides insights into vertebrate evolution. Nat Genet. 2013;45:415–21.
Session AM, Uno Y, Kwon T, Chapman JA, Toyoda A, Takahashi S, et al. Genome evolution in the allotetraploid frog Xenopus laevis. Nature. 2016;538:336–43. https://doi.org/10.1038/nature19840.
Hoegg S, Brinkmann H, Taylor JS, Meyer A. Phylogenetic timing of the fish-specific genome duplication correlates with the diversification of teleost fish. J Mol Evol. 2004;59:190–203.
Prior HM, Walter MA. Sox genes: architects of development. Mol Med. 1996;2:405–12.
Uchikawa M, Kamachi Y, Kondoh H. Two distinct subgroups of Group B Sox genes for transcriptional activators and repressors: their expression during embryonic organogenesis of the chicken. Mech Dev. 1999;84:103–20.
Ochi H, Tamai T, Nagano H, Kawaguchi A, Sudou N, Ogino H. Evolution of a tissue-specific silencer underlies divergence in the expression of pax2 and pax8 paralogues. Nat Commun. 2012. https://doi.org/10.1038/ncomms1851.
Hara Y, Takeuchi M, Kageyama Y, Tatsumi K, Hibi M, Kiyonari H, et al. Madagascar ground gecko genome analysis characterizes asymmetric fates of duplicated genes. BMC Biol. 2018;16:1–19.
Gerts EM, Yu YK, Agarwala R, Schäffer AA, Altschul SF. Composition-based statistics and translated nucleotide searches: Improving the TBLASTN module of BLAST. BMC Biol. 2006;4:1–14.
Solovyev V, Kosarev P, Seledsov I, Vorobyev D. Automatic annotation of eukaryotic genes, pseudogenes and promoters. Genome Biol. 2006;7(Suppl 1):1–12.
Pyron RA, Burbrink FT, Wiens JJ. A phylogeny and revised classification of Squamata, including 4161 species of lizards and snakes. BMC Evol Biol. 2013;13:1. https://doi.org/10.1186/1471-2148-13-93.
Crawford NG, Parham JF, Sellas AB, Faircloth BC, Glenn TC, Papenfuss TJ, et al. A phylogenomic analysis of turtles. Mol Phylogenet Evol. 2015;83:250–7. https://doi.org/10.1016/j.ympev.2014.10.021.
Katoh K, Standley DM. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol Biol Evol. 2013;30:772–80.
Suyama M, Torrents D, Bork P. PAL2NAL: robust conversion of protein sequence alignments into the corresponding codon alignments. Nucleic Acids Res. 2006;34(Web serv):609–12.
Capella-Gutiérrez S, Silla-Martínez JM, Gabaldón T. trimAl: a tool for automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics. 2009;25:1972–3.
Nguyen LT, Schmidt HA, Von Haeseler A, Minh BQ. IQ-TREE: a fast and effective stochastic algorithm for estimating maximum-likelihood phylogenies. Mol Biol Evol. 2015;32:268–74.
Kalyaanamoorthy S, Minh BQ, Wong TKF, Von Haeseler A, Jermiin LS. ModelFinder: fast model selection for accurate phylogenetic estimates. Nat Methods. 2017;14:587–9.
Ronquist F, Teslenko M, Van Der Mark P, Ayres DL, Darling A, Höhna S, et al. Mrbayes 3.2: efficient bayesian phylogenetic inference and model choice across a large model space. Syst Biol. 2012;61:539–42.
Yang Z. PAML 4: Phylogenetic analysis by maximum likelihood. Mol Biol Evol. 2007;24:1586–91.
Wertheim JO, Murrell B, Smith MD, Pond SLK, Scheffler K. RELAX: detecting relaxed selection in a phylogenetic framework. Mol Biol Evol. 2015;32:820–32.
The Xenopus tropicalis genome browser (http://viewer.shigen.info/xenopus/index.php) was in part supported by Hiroshima University Amphibian Research Center through National BioResource Project (NBRP) of AMED.
This work was partially supported by Grant-in-Aid for Scientific Research, Japan Society for the Promotion of Science (18K06389) to MI.
Ethics approval and consent to participate
Consent for publication
The authors declare no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Genome IDs used in the study. Table S2. The presence ( +) and absence (-) of sox15, mpdu1, and fxr2 on reptile genomes used in Fig. 2a, b. Table S3. Genbank numbers and range of the genes used in the analyses. Table S4. Likelihood ratio test for ω values in Fig. 4. Figure S1. Multiple alignment of the HMG box-encoding nucleotide sequences of marsupial pseudogenized and non-pseudogenized. Figure S2. Predicted amino acid sequences of marsupial SOX15 from koala (Phascolarctos cinereus) and wombat (Vombatus ursinus). Figure S3. Easyfig analysis for the locus encoding fxr2 and mpdu1 genes in reptiles. Figure S4. Phylogenetic relationships of vertebrate soxB1/G ohnologous proteins. (a) Maximum likelihood and (b) Bayesian phylogenetic trees were shown. A total of 108 aa sequences containing 273 sites were used for this tree inference. The JTT + F + I + Γ4 model was selected as the best-fit model in this dataset and used for the inference. The invertebrate SOXB1 clade was rooted. Values of (a) the 1000 times ultrafast bootstrap test and (b) the Bayesian posterior probability are shown at each node. Only the 90 ≦ bootstrap values and 0.90 ≦ posterior probabilities were shown. The scale bars indicate aa substitutions per site. Figure S5. An intron-less structure of Chiloscyllium punctatum ortholog of sox15 by blastn hit.
About this article
Cite this article
Ogita, Y., Tamura, K., Mawaribuchi, S. et al. Independent pseudogenizations and losses of sox15 during amniote diversification following asymmetric ohnolog evolution. BMC Ecol Evo 21, 134 (2021). https://doi.org/10.1186/s12862-021-01864-z
- Gene loss
- d N/d S