Lessons from an unusual vertebrate sex-determining gene

Abstract

So far, very few sex-determining genes have been identified in vertebrates and most of them, the so-called ‘usual suspects’, evolved from genes which fulfil essential functions during sexual development and are thus already tightly linked to the process that they now govern. The single exception to this ‘usual suspects’ rule in vertebrates so far is the conserved salmonid sex-determining gene, sdY (sexually dimorphic on the Y chromosome), that evolved from a gene known to be involved in regulation of the immune response. It is contained in a jumping sex locus that has been transposed or translocated into different ancestral autosomes during the evolution of salmonids. This special feature of sdY, i.e. being inserted in a ‘jumping sex locus’, could explain how salmonid sex chromosomes remain young and undifferentiated to escape degeneration. Recent knowledge on the mechanism of action of sdY demonstrates that it triggers its sex-determining action by deregulating oestrogen synthesis that is a conserved and crucial pathway for ovarian differentiation in vertebrates. This result suggests that sdY has evolved to cope with a pre-existing sex differentiation regulatory network. Therefore, ‘limited options’ for the emergence of new master sex-determining genes could be more constrained by their need to tightly interact with a conserved sex differentiation regulatory network rather than by being themselves ‘usual suspects’, already inside this sex regulatory network.

This article is part of the theme issue ‘Challenging the paradigm in sex chromosome evolution: empirical and theoretical insights with a focus on vertebrates (Part I)’.

1. Genetic sex determination in vertebrates is mostly governed by the ‘usual suspects’ sex-determining genes

Sex determination, the process that initiates the development of sexually differentiated male and female reproductive organs, is a prerequisite for sexual reproduction that is widespread in the animal and plant kingdoms. In vertebrates, sex determination is generally viewed as being either genetically driven by so-called master sex-determining genes (GSD, for genetic sex determination), or environmentally driven by diverse environmental factors (ESD, for environmental sex determination) [1,2]. Among ESD systems, the major role of temperature on sex determination has been well-described in reptiles [3,4]. But more complex sex determination systems relying on a GSD system that can be influenced by environmental effects have also been described, like the thermal effects on sex determination in many fish species [5].

In vertebrates, GSD systems are governed by master sex-determining (MSD) genes and most of the currently known vertebrate MSD genes are what have been named the ‘usual suspects’, i.e. genes known to fulfil essential functions during sexual development that are thus already tightly linked to the sex differentiation pathway that they now control [6]. These ‘usual suspects’ actually belong to a small bunch of gene families that seem to have been independently and recurrently selected to produce new MSD genes (figure 1a) ([7]; see also [8], this issue, for more details). Among them, MSD genes from the SOX (SRY-related HMG-box) gene family have been recruited at least twice in vertebrates, including both the Sry (sex determining region of Y chromosome) MSD gene of most eutherian mammals [9,10] and the sox3 gene that has been characterized as the MSD gene of the Indian ricefish, Oryzias dancena [11]. The DMRT (doublesex/mab-3 related transcription factor) family has been characterized as an important gene family involved in sexual development in vertebrates, insects and nematodes [12], making it the most conserved MSD gene family in animals known so far. In vertebrates, dmrt1bY in the medaka Oryzias latipes [13,14] and Dm-W in the African clawed frog Xenopus laevis [15] have both been recruited as MSD genes through duplications of their respective autosomal paralogous Dmrt1 copies. In the half-smooth tongue sole Cynoglossus semilaevis [16], as well as in birds [17], a dosage sensitive mechanism of Dmrt1 triggers sex determination with two Z chromosome copies inducing testicular differentiation in males (ZZ) and one Z copy (haploinsufficiency) inducing ovarian differentiation in females (ZW). Many members of the transforming growth factor beta (TGF-β) pathway have also been selected as MSD genes in vertebrates (see [8], this issue), including TGF-β ligands with the anti-Mullerian hormone (amh) [18–20], the gonadal soma-derived factor (gsdf) [21,22] or the growth/differentiation factor 6a (gdf6a) [23] genes, but also TGF-β type II and type I receptors with the anti-Mullerian hormone receptor type 2 (amhr2) [24,25] and the bone morphogenetic protein receptor type IBb (bmpr1bb) [26] genes. Finally, a new family of ‘usual suspect’ MSD genes should probably be created with members of the steroidogenic pathway involving steroid enzymes like the 17β-hydroxysteroid dehydrogenase 1 (hsd17b1) that has recently been strongly suggested to be a conserved MSD gene in both amberjack and yellowtail (genus Seriola) [27,28].

Figure 1.

The vertebrate sex differentiation network is under the control of the ‘usual suspects’ master sex-determining genes. (a) In vertebrates, most of the currently known vertebrate master sex-determining genes are ‘usual suspects’ (TGF-β, Sox, Dmrt and steroid enzyme gene families) genes known to fulfil essential functions during sexual development and that are already tightly linked to the sex differentiation network that they now control. (b) The salmonid MSD gene, sdY (sexually dimorphic on the Y chromosome), is the single exception to the ‘usual suspects’ rule in vertebrates as it evolved from a gene involved in the regulation of the immune response. In salmonid females (XX), Foxl2 (Forkhead box L2) and Nr5a1 (Nuclear Receptor Subfamily 5 Group A Member 1) synergistically activate the aromatase (cyp19a1a) gene leading to the production of the aromatase enzyme and of feminizing oestrogenic steroids (E2). These E2s are then able to stimulate foxl2 expression allowing the establishment of a positive feedback loop of regulation that will increase E2 synthesis in the early differentiating gonads, inducing ovarian differentiation. In males (XY), by blocking Foxl2 action, SdY prevents the establishment of this positive feedback loop of regulation and the synthesis of E2, thereby allowing testicular differentiation to proceed.

By contrast with this relative abundance of ‘usual suspect’ MSD genes, only a single exception to the rule has been characterized to date in vertebrates, namely the salmonid MSD gene, sdY (sexually dimorphic on the Y chromosome), that evolved from a gene involved in the regulation of the immune response [29–31]. This over-representation of ‘usual suspect’ MSD genes has been used to support the ‘limited options’ idea [32,33], which states that sex determination evolution could be constrained to use a few sets of ancestral chromosomes and/or genes that would be better at becoming sex chromosomes and/or MSD genes. But the recruitment of MSD genes outside the canonical sex determination/differentiation pathway has also been demonstrated in invertebrates. For instance, in the house fly, Musca domestica, the male MSD gene Mdmd (Musca domestica male determiner) originated from a duplication of the spliceosomal factor gene CWC22 (nucampholin) [34], and, in the silkworm, Bombyx mori, a female-specific W chromosome-derived PIWI-interacting RNA acts as a MSD gene [35].

2. Salmonid sex and the discovery of a unique case of unusual vertebrate master sex-determining gene

Salmonids belong to a teleost fish clade of both economic and ecological importance. They are also important fish models; rainbow trout, for instance, has often been described from an experimental point of view as the fish equivalent of laboratory rats in mammals [36]. Sex determination has then always been a long-standing question in this group fuelled both by applied aspects with regards to sex control in aquaculture [37], but also because of the intriguing question of how sex was determined in that fish family with regards to the quite recent whole genome duplication (WGD) event that they experienced [38,39].

The sex-determination system of salmonids is generally described as being strictly genetically driven, with a male heterogametic (XX/XY) system that has been characterized in most species of the subfamily Salmoninae that contains rainbow trout and Pacific salmons (genus Oncorhynchus), Atlantic salmon and brown trout (genus Salmo) and charrs (genus Salvelinus) [40,41]. The long quest for the identification of salmonid sex chromosomes and MSD gene(s) was first based on genetic mapping strategies searching for quantitative trait loci (QTLs) linked with sex [42–49]. But despite the identification and characterization of sex chromosomes and some sex-specific sequences [45,50–53] no putative MSD gene was characterized using these genetic approaches. The discovery of the salmonid MSD gene actually came from comparative transcriptomic analysis of rainbow trout male and female embryonic gonads, that led to the identification of a gene specifically expressed in differentiating testes [30]. This gene, named sdY for ‘sexually dimorphic on the Y chromosome’, was found to be only present in male genomes thus fulfilling an important characteristic of a potential MSD gene. In addition, the sdY sequence was also found to be contained within a previously characterized Y chromosome-specific bacterial artificial chromosome (BAC) sequence named Omy-Y1 [53], adding supporting evidence that this gene could be the rainbow trout MSD gene. But the definitive proof came from functional evidence demonstrating that this gene is both necessary and sufficient to induce and to maintain rainbow trout testicular differentiation [30,31]. These functional studies make sdY one of the very few well characterized MSD genes in vertebrates.

As mentioned above, many salmonids have a male heterogametic sex determination system (XX/XY). But comparative genetics studies demonstrated that they do not share the same sex chromosomes, raising questions about the uniqueness of such MSD gene in that fish family [42,54]. By looking at the extent of conservation of the rainbow trout MSD gene, Yano et al. [29] found that most Salmonid species exhibit a complete male sex-linkage of sdY. In Esociformes, which is the sister group of salmonids, sdY was found to be absent in Northern pike [29], which has a duplicated copy of amh as MSD gene [18]. These results led to the conclusion that sdY is a conserved and salmonid-specific gene and that most salmonids do share the same MSD gene despite not having the same Y chromosomes. This indeed echoes with the initial hypothesis formulated by Woram et al. [42], that a conserved salmonid sex determining locus has been transposed or translocated to different ancestral autosomes during the evolution of salmonids, resulting in many different and species-specific Y sex chromosomes. This MSD gene transposition ability, often referred as a ‘jumping sex locus’, has not been reported in another vertebrate so far, but has been characterized both in plants [55] and insects [56]. For instance, in the phorid fly, Megaselia scalaris, the MSD gene has been observed on different sex chromosomes suggesting it is a ‘jumping’ MSD gene [34]. In rainbow trout, the analysis of 800 kb around the sdY locus suggested the presence of transposons, ribosomal DNA, repetitive elements and a few single-copy genes such as CREB-regulated transcription activator (crtc) and cAMP responsive element binding (creb) [53]. Comparative analyses of the sdY locus in rainbow trout, Chinook and Atlantic salmon then revealed a small 4.1 kb conserved region that also contains potential elements necessary for transposition, such as transposase and RNA-directed DNA polymerase [57]. These findings suggest that this conserved region is the minimal region needed to trigger masculinization in salmonids and that sdY is contained in a transposable cassette. In salmonids, it is interesting to note that this transposition mechanism still seems to be an active process as suggested in Atlantic salmon, Salmo salar, by genome-wide association analyses and fluorescence in situ hybridization (FISH) experiments, showing that different sdY loci are found on three or even four different (sex) chromosomes [58–61]. This demonstrates that the sdY locus transposition was still active even after Atlantic salmon speciation. This transposition mechanism would be regulated by a copy-and-paste mechanism as the sdY locus is predicted to contain both transposase and RNA-directed DNA polymerase [57]. Additional evidence of a recent and copy-and-paste transposition activity of the sdY locus is also confirmed by the fact that a small number of both males and females of Atlantic salmon can have additional, but non-functional, sdY autosomal copies [62]. Such non-functional copies of sdY have been postulated to explain the complete absence of sdY sex linkage in some salmonid species, for instance in several whitefish species (subfamily Coregoninae) [29]. But sdY sex-linkage discrepancies have also been reported in other salmonid species with the detection of a few sdY-positive females [63–66], or sdY-negative males [29,63,67], raising hypotheses that sdY could have been replaced at least partially by another MSD gene in some species [67] or that its mechanism of action could rely on a more complex dosage effect [66] than the simple presence/absence of a classical dominant MSD gene. In that regards, we found (S Bertho et al. 2021, unpublished results) that sdY-positive females in Chinook salmon, Oncorhynchus tshawytscha, that have been characterized by cytogenetic approaches as being apparent XY females [68,69], can be explained by a non-functional sdY copy due to a single missense mutation in their coding sequence. Other studies have also found a high proportion of female-to-male sex reversal in both rainbow trout, O. mykiss, and sockeye salmon, O. nerka, populations. These deviations may be explained either by masculinizing mutations [61,62] and/or by masculinization effects of high temperature [70–72]. Altogether this shows that a definitive conclusion on whether sdY has strictly conserved its MSD gene function in all salmonids awaits more precise studies. These studies should better take into consideration the existence of potential non-functional sdY copies and/or the effects of environmental factors on sex determination. It is also interesting to note that salmonids seem to reuse some sex chromosomes, suggesting in line with the ‘limited options’ idea [32,33] that some of these ancestral salmonid chromosomes would be better at becoming sex chromosomes [73].

Another unusual and interesting feature of sdY is that, in contrast to all previously known ‘usual suspects’ vertebrate MSD genes, sdY evolved from the duplication of a gene encoding for an interferon-regulatory factor 9 (Irf9) protein that is known to be involved in the regulation of the immune response [30]. This duplication is, however, not linked to the salmonid WGD [38,39] as salmonids already have two irf9 paralogs clearly stemming from this WGD [30]. In comparison to Irf9, which has both a protein–protein interaction domain (IAD) and a DNA transactivation domain, SdY is truncated in its N-terminal domain and only kept the C-terminal IAD domain of Irf9. This strongly suggested that SdY could mediate its sex-determining effects through protein–protein interactions. Indeed, the search for potential interacting partners uncovered that SdY is interacting with the forkhead box L2 (Foxl2) protein [74], a transcription factor well known for its crucial role in ovarian differentiation [75]. By binding to Foxl2, SdY is able to prevent the Foxl2 and Nr5a1 (Nuclear Receptor Subfamily 5 Group A Member 1) synergistic activation of the aromatase (cyp19a1a) gene, blocking the positive feedback loop of regulation required for synthesis of the feminizing oestrogenic steroids [76] in the early differentiating gonads (figure 1b). Thus, SdY is not a male-promoting factor as, despite being a male MSD gene, it acts as an anti-female factor by blocking a default female differentiation regulatory network, and thereby allowing testicular differentiation to proceed. This anti-female action may be, however, not the only function of SdY as we cannot rule out the existence of a simultaneous and yet unknown pro-testicular action of SdY.

3. Lesson 1 from sdY: how can sex chromosomes stay ‘young’ and escape from degeneration?

New sex chromosomes can evolve from autosomes through the acquisition of new sex-determining loci that can emerge through two major processes often referred to as allelic diversification and duplication [1,2,77]. The allelic diversification process involves the fixation of an allelic variant conferring a new sex-determining function to a gene previously located on an autosome. This allelic variant may become fixed in the population. As a heritable trait in the heterogametic sex this will result in the emergence of a new (proto-) sex chromosome from an ancestral autosomal pair. The duplication/insertion process was first described in the medaka, O. latipes and involves the duplication of an autosomal gene (i.e. dmrt1 in O. latipes) that will evolve a new sex-determining function after its insertion (i.e. dmrt1bY in O. latipes) at a genomic location that will become the new heterogametic sex chromosome [14]. Following the birth of a new sex-determining locus, sex chromosomes can evolve through an extension of the non-recombining region that is initially restricted to the founding allelic variant or duplicated region of the new MSD gene. This process can eventually lead to the degeneration or erosion of the heterogametic sex chromosome in genomic regions surrounding this new MSD and can even cover almost the entire chromosome except for a pseudoautosomal region. However, this evolution towards a highly differentiated sex chromosome does not seem mandatory as there are more and more examples of old but yet still undifferentiated sex chromosomes [18,21,24,78]. Differences between such undifferentiated sex chromosomes can even be restricted to a few single nucleotide polymorphisms, like for instance in some pufferfish species of Takifugu in which recombination suppression did not seem to have spread out from the single Y-defining allele bearing the sex-determining function in the amhr2 gene for at least 5 Myr [24].

Both birds and eutherian mammals have kept their respective W and Y sex chromosomes over long evolutionary periods, with eutherian mammals having kept their Y chromosome for more than 180 Myr [79]. But this stability seems to be more an exception to the rule compared to the high number of sex chromosomes and/or MSD genes transitions that have been reported in other classes of vertebrates [1,2,7]. This surprising longevity went along with a high specialization and differentiation of the mammalian Y chromosome [80], that has undergone a large degeneration process. A prediction was made reasoning that the degeneration process would proceed at the same pace as that of the human Y chromosome and in analogy those of all other mammalian species are doomed to an ineluctable death [79,81]. In consequence, it was concluded that homomorphic¹ sex chromosomes should be ‘young’ sex chromosomes. Within vertebrates a large number of species have homomorphic sex chromosomes, and two models have been proposed to explain how heterogametic sex chromosomes could escape from their potential detrimental degeneration. The first one, known as the ‘high turnover’ model (figure 2a), is based on the idea of a regular replacement of the MSD genes that also would lead to the replacement of sex chromosomes [84,85], a process that has been found often in many fish species. However, the switch of a MSD gene does not necessarily imply the loss of the previous sex chromosome as has been postulated in a recent alternative to the ‘high turnover’ theory [86]. Such an intra-chromosomal MSD gene switch, by shifting the sex determination locus far away from its initial location, would probably lead to the same rejuvenating effect as in the classical ‘high turnover’ model, at least for highly homomorphic sex chromosomes as have been observed for instance in some Takifugu species [87]. The second theory, named the ‘fountain of youth’, is based on the fact that many species exhibit sex-specific patterns of recombination with one phenotypic sex having a higher recombination rate than the other, a notion known as heterochiasmy. The ‘fountain of youth’ model then postulates that recombination of sex chromosomes will be favoured in occasional sex-reversed XY females in species having a higher female recombination rate and that this will provide opportunities to recombine the Y heterogametic chromosome, allowing a regular purge of its deleterious mutations [88]. Interestingly, sdY, by being contained in a transposable cassette, or a mobile sex locus, has the ability to continuously reshuffle the sex chromosomes and thus to prevent their decay. This strategy may provide a third explanation for homomorphic sex chromosomes, the ‘jumping sex locus’ model (figure 2c), that would keep the idea from the ‘fountain of youth’ model that new, here in the sense of rejuvenated, sex chromosomes might actually harbour old sex-determining genes [88]. But in contrast with the ‘fountain of youth’, the salmonid ‘jumping sex locus’ keeps the idea of the ‘high turnover’ model, namely a complete switch of the sex locus to a new location, on a different chromosome that becomes the sex chromosome (figure 2). This ‘jumping sex locus’ strategy, when installed, could be seen as being a good and cost-effective solution because keeping the same MSD gene would not need all the tight physiological adjustments that are probably needed to cope with a new MSD gene, including the precise regulation of expression levels and timing of expression of all the members of the sex-differentiation downstream network. However, this system probably requires a precise regulation of the number of simultaneous functional translocations to prevent bias in sex ratio and multiple sex chromosomes conflicts.

Figure 2.

How the ‘jumping sex locus’ of salmonids fits with known models explaining how sex chromosomes stay undifferentiated. (a) The ‘high turnover’ model stipulates that the regular replacement of MSD genes would lead to the replacement of sex chromosomes, preventing them from getting old and degenerated. This theory is presented here exemplified by the MSD gene turnover from the dmrt1bY gene located on linkage group (LG) 1 (LG1) in Oryzias latipes, to the gsdfY gene located on LG12 in Oryzias luzonensis. (b) The ‘fountain of youth’ model initially described based on results from the frog, Rana temporaria, postulates that occasional sex reversed individuals of the heterogametic sex (here XY females) can provide opportunities for recombination between the heterogametic chromosomes allowing deleterious mutations to be purged from the heterogametic sex chromosome. This model implies that the same MSD gene is kept on the same (LG2) rejuvenated sex chromosome. (c) In salmonids, the same MSD gene, i.e. sdY, is regularly translocated to a different sex chromosome also preventing its decay. Such a ‘jumping sex locus’ transition is presented here with the example of the brown trout, Salmo trutta, sdY locus on LG12.2 that has been translocated on LG14.2 in rainbow trout, Oncorhynchus mykiss. Salmonid LGs are defined with regards to an ancestral and non-duplicated salmonid outgroup, i.e. the Northern pike, Esox lucius [73,82]. Chromosome switches are shown from grey to blue.

4. Lesson 2 from sdY: is there constrained evolution for master sex-determining genes to cope with a conserved sex-differentiation downstream network?

With the sdY gene being the single ‘usual suspect’ MSD outlier known so far in vertebrates, its mechanism of action was then of great interest from an evolutionary point of view. Based on gene ontology inference from its irf9 ancestor, which is involved in type I interferon signalling, a first simple hypothesis was to assume that sdY would trigger its sex-determining effects by using the same regulatory network as its ancestor. However, the immune pathway has not been characterized as being important for the regulation of the vertebrate sex-differentiation regulatory network. This then raised the question of how a gene outside of the classical sex-differentiation regulatory network could trigger its sex-determining functions? Did the salmonid sex-differentiation regulatory network have to cope with a completely novel MSD function and would now rely on a previously unsuspected immune pathway? Or on the contrary, did this novel MSD gene have to cope with the pre-existing sex differentiation regulatory network by evolving a totally new function not related to its ancestral immune function? After showing that SdY determines sex in rainbow trout by interacting directly with a key member of the classical gonadal sex-differentiation regulatory network, i.e. Foxl2, we [74] brought support to a neofunctionalization hypothesis that implies that sdY evolved a new function to cope with a pre-existing classical sex-differentiation regulatory network. This led us to suggest that innovation at the top of the vertebrate sex-determination regulatory network may be highly constrained by the need for a novel master SD gene, which is able to cope with the regulatory machinery of the conserved vertebrate sex-differentiation regulatory network. This ‘limited option’ idea provided a framework hypothesis that only a small subset of genes and chromosomes, because they are better at doing the job, would be independently and repeatedly selected as new vertebrate master SD genes [32]. But now, based on the known mechanism of action of sdY, we propose that ‘limited options’ for new MSD genes are actually more constrained by their need to tightly interact with a conserved sex-differentiation regulatory network. The existence of unusual MSD genes like sdY suggests that innovation at the top of the sex-determination hierarchy is possible only if they are able to hijack the pre-existing sex-differentiation regulatory network, like sdY did by being able to deregulate oestrogen synthesis that is a conserved and crucial pathway for ovarian differentiation [76].

5. Conclusion

The salmonid sdY gene is an unusual vertebrate MSD gene for multiple reasons. It is part of a jumping sex locus that has been conserved for a long period of time, as this gene is present in all extant salmonid species, suggesting that it arose by a local gene duplication in the common ancestor of all extant salmonid species, 80–95 Ma [38,39]. It is also the single outlier MSD gene currently known in vertebrates with regards to the ‘usual suspects’ rule. This longevity of sdY as a conserved MSD gene in salmonids is in the same range as Sry in mammals, and may have been favoured by its ability to jump from one chromosome to another, combined with a mechanism of action tightly controlling a pivotal step of the gonadal sex-differentiation regulatory network, i.e. the ability to regulate oestrogen synthesis required for ovarian differentiation [76]. But whether this MSD role has conserved its sex-determining function in all salmonid species remains to be better explored as maintaining the same MSD gene for a long period seems like a difficult endeavour, with complex evolutionary trajectories including gains and losses at the top (sex chromosomes and sex determination), and downstream compromises (sex-differentiation) as have been found in some fish families [78]. Considering the complexity of the mechanisms of sex determination and their plasticity this question still needs more investigation and the discovery of other unusual MSD genes, like sdY, would probably fuel additional models for the tightly linked evolution of sex chromosomes, sex determination and sex-differentiation.

Data accessibility

This article has no additional data.

Authors' contributions

S.B., A.H., M.S. and Y.G. wrote the manuscript.

Competing interests

We declare we have no competing interests.

Funding

Part of this work was supported by Agence Nationale de la Recherche (ANR) grant no. ANR ANR-11-BSV7-0016 (SDS project) and grants to M.S. by the Deutsche Forschungsgemeinschaft (Scha408/12-1, 10-1).