Dmrt1 genes at the crossroads: a widespread and central class of sexual development factors in fish


A. Herpin, University of Wuerzburg, Physiological Chemistry, Am Hubland, D-97074 Wuerzburg, Germany
Fax: +49 931 888 4150
Tel: +49 (0)931 888 4153


A plethora of corroborative genetic studies led to the view that, across the animal kingdom, the gene-regulatory cascades triggering sexual development bear little resemblance to each other. As a result, the common emerging picture is that the genes at the top of the cascade are not conserved, whereas the downstream genes have homologues in a much broader spectrum of species. Among these downstream effectors, a gene family involved in sex differentiation in organisms as phylogenetically divergent as corals, Caenorhabditis elegans, Drosophila, frogs, fish, birds and mammals is the dmrt gene family. Despite the attention that Dmrt1 factors have received, to date it has not been elucidated how Dmrt1s mediate their activities and putative downstream targets have yet to be characterized. However, a remarkable amount of descriptive expression data has been gathered in a large variety of fish, particularly with respect to early gonadal differentiation and sex change. This minireview aims at distilling the current knowledge of fish dmrt1s, in terms of expression and regulation. It is shown how gonadal identities correlate with dimorphic dmrt1 expression in gonochoristic and hermaphroditic fish species. It is also described how sex steroid hormones affect gonadal identity and dmrt1 expression. Emphasis is also given to recent findings dealing with transcriptional, post-transcriptional, post-translational and functional regulations of the dmrt1a/dmrt1bY gene pair in medaka.


The phenomenon of two different sexes and consequently the necessity to make a developmental decision for an embryo to become male or female (the so-called sex-determination process), and the further differentiation of the whole organism into two distinct phenotypes, are common throughout the animal, plant and fungi kingdoms. Nevertheless, with respect to animals at least, decades of elegant genetic studies have led to the global picture that the gene-regulatory cascades triggering sexual differentiation from Caenorhabditis elegans and Drosophila to mammals bear little resemblance to each other. Hence, although developmental cascades are generally headed by highly conserved universal master regulators that determine the developmental fate of a cell lineage to a given tissue or organ during embryogenesis, all the evidence suggests that sex determination might disobey the conventional rules of evolutionary conservation. The common picture emerging here is that the genes at the top of the cascade are not conserved, whereas the downstream genes have homologues in a much broader spectrum of species [1,2]. For example, SRY, the male sex-determining gene of mammals, has not been detected outside the eutherians (placental mammals). Conversely, known downstream effectors involved in gonadogenesis or gonadal differentiation like, for example, Wt1, Sox-9, Bmps and Amh (see [3] for a review) are present in all vertebrates including fish [4] and for most of them even in protostomes.

Among the downstream candidate genes, a gene family involved in sex differentiation in organisms as phylogenetically divergent as C. elegans, Drosophila, frogs, fish, birds, mammals and corals is the dmrt gene family [5]. The prototype members of this group of factors are the Drosophila doublesex (dsx) and Caenorhabditis mab-3 genes. The Dmrt group of molecules is characterized by a conserved DNA-binding motif known as the Doublesex- and Mab-3-related (DM) domain. Being a noncanonical cysteine-rich DNA-binding motif, this domain has two highly intertwined finger structures that chelate one zinc ion each, and binds to the minor groove of the DNA [6]. Dmrts were originally described to play important roles during sex determination in flies and worms by regulating several aspects of somatic sexual dimorphism. They were also reported to be able to substitute for each other across species, indicating that their function is possibly interchangeable and that sex determination in invertebrates might rely on conserved molecules, at least at the bottom of the cascade [7]. Consistently, many of the subsequently characterized metazoan Dmrt homologues were predominantly expressed in the developing gonads. Thus, this widespread class of factors commonly appeared to be directly involved in sex determination. Although homology relationships of dmrt gene family members across all the metazoans have not been established, for vertebrates it has been shown that the prototype member of the gene family, designated dmrt1, is most closely related to the Drosophila dsx and C. elegans mab-3 genes in structure and by means of sex-determination/differentiation function. Gonadal dmrt1 expression is generally detected at higher levels in testes than ovaries.

The deep interest in Dmrt1 in the field of sex determination in fish came with the discovery of a dmrt1 homologue on the Y chromosome of the fish medaka (Oryzias latipes). This Y-chromosomal gene is the product of a gene duplication of the autosomal dmrt1a gene and was designated dmrt1bY [8] or Dmy [9]. It was shown to be the only functional gene in the whole Y-specific region of the sex chromosome [10]. Mutations affecting this gene result in male-to-female sex reversal [11]. In addition, dmrt1bY transgene-induced testis development in genetic females (XX) definitively pointed out that it is not only necessary, but also sufficient for triggering male development [12]. Considering that dmrt1bY has all the features of the master regulator of testicular differentiation in medaka (see [13] for review) and because of the discovery of sex-chromosome-linked dmrt1s in other vertebrates (DM-W in Xenopus [14] and dmrt1 in birds [15] for example), it was tempting to speculate, at least for teleosts, that dmrt1s might have a universal and top control function during sex determination. However, the absence of a dmrt1bY gene even in closely related Medaka species ruled this out [16]. Nevertheless, factually it did not exclude Dmrt1, in general, as an important conserved effector of testis development, including spermatogenesis.

Despite the attention that Dmrt1 factors have received, to date it has not been elucidated how Dmrt1s mediate their activities and putative downstream targets have yet to be characterized [17]. However, a remarkable amount of descriptive expression data has been gathered in a large number of different fish species, particularly in the context of early gonadal development, gonadal differentiation and sex change. This minireview aims at distilling current knowledge about the expression and regulation of dmrt1s in fish towards a more general picture. Emphasis is also given to recent findings dealing with transcriptional, post-transcriptional, post-translational and functional regulation of the dmrt1a/dmrt1bY gene pair in medaka.

Gonadal dmrt1 gene expression across different fish species

An amazing variety of sex-determining systems is found in fish. Although information is emerging about sex determination in lampreys, sharks, rays and sturgeons, most of our knowledge stems from studies on teleost fish. Hence, this minireview mainly concentrates on that group. A considerable number of teleost species are hermaphrodites, switching either from first being males (protandrous) to become female or vice versa (protogynous). Nevertheless, the majority of teleosts are gonochoristic, meaning that they exist as males and females regardless of the primary sex determination initiating process being environmental (temperature, social) or genetic (XY or ZW).

Gonadal dimorphic dmrt1 expression in gonochoristic species

Male-restricted expression of dmrt1 has been reported for North African catfish Clarias gariepinus [18], rare minnow Gobiocypris rarus [19], Nile tilapia Oreochromis niloticus [20], medaka Oryzias latipes [21] and olive flounder Paralichthys olivaceus [22]. In lake sturgeon Acipenser fulvescens [23], zebrafish Danio rerio [24], Atlantic cod Gadus morhua [25], pejerrey Odontesthes bonariensis [26], rainbow trout Oncorhynchus mykiss [27], shovelnose sturgeon Scaphirhynchus platorynchus [28] and southern catfish Silurus meridionals [29] a strong male-biased expression appears as the general rule, although some dmrt1 expression could be detected in ovaries (see Table 1). Interestingly, when detected in the ovary, dmrt1 expression is consistently seen in the germ cells (Gadus morhua [25] and Danio rerio [24]), whereas much broader and less restricted expression territories are seen within the testis. With respect to a gonadal function of Dmrt1, its early expression in the somatic part of the male gonad anlage (Oreochromis niloticus [20] and Oryzias latipes [21]) would infer a role correlated with Sertoli cell lineage specification and subsequently during testicular differentiation. The specific expression in spermatogonia and spermatocytes reported for Clarias gariepinus [18], Danio rerio [24] and Gadus morhua [25] are clearly consistent with a role at some stage of spermatogenesis in these species.

Table 1.   Gonadal expression of dmrt1 genes across the fish kingdom.
Species Gonadal expressionExpression levelsExpression localizationMethodsRef
  1. G, gonochoric; pA, protandrous; pG, protogynous; TSD, temperature-dependent sex determination; n.i., not investigated; IC, immunocytochemistry; ISH, in situ hybridization.

Acanthopagrus schlegeli pATestisHigher in mature testisn.i.PCR[30]
Acipenser fulvescens GOvary and testisHigh in testesn.i.PCR[23]
Clarias gariepinus GTestisOva-testisSpermatogonia, spermatocytesPCR, IC, western blot[18]
Danio rerio GTestis and ovaryHigh in testesSpermatogonia, spermatocytes spermatids and developing oocytesPCR, ISH[24]
Epinephelus coioides pGTestisSpermatogonia, spermatocytesPCR, IC, western blot[32]
Gadus morhua GTestis and ovaryDuring spermatogenesisGerm cells (testis and ovary)PCR, ISH[25]
Gobiocypris rarus GTestisn.i.PCR[19]
Halichoeres tenuispinis pGTestis Northern blot[33]
Monopterus albus pGTestis, ovotestis and ovary (sex-specific splice variantsHigh in testesGonadal epithelium, undifferentiated germ cells (splice variants)PCR, ISH, Northern blot[34]
Odontesthes bonariensis TSDPrimordial gonadsDuring testicular differentiationn.i.PCR[26]
Oncorhynchus mykiss GTestis and ovaryHigher in testesDifferentiating testisPCR, Northern blot[27]
Oreochromis niloticus GTestisIn sex-reversed testesSertoli and epithelial cells of the efferent ductPCR, ISH[20]
Oryzias latipesGDmrt1a: testisSpermatogonial supporting cells, pre-Sertoli,PCR, ISH, IC[21,54]
 Dmrt1bY: testisSertoli cells and testicular interstitial tubules  
Paralichthys olivaceus GTestisn.i.PCR[22]
Scaphirhynchusplatorynchus GTestis and ovaryHigher in testesn.i.PCR[28]
Sparus auratus pATestisDecreases during testicular involutionn.i.PCR[31]
Silurus meridionalis GOvary and testisHigh in testes during masculinizationn.i.PCR[29]
Takifugu rubripes GTestis and ovaryHigh in testesSertoli cellsPCR, ISH[57]
Xiphophorus maculatus GTestisSpermatogonia, Sertoli cellsPCR, ISH[58]

Another remarkable piece of information towards the understanding of Dmrt1 function(s) is coming from gonochoric fish that are annual breeders (Clarias gariepinus [18], Oncorhynchus mykiss [27] and Silurius meridionalis [29]). In these species, fish undergo a seasonal pattern of gonadal resting and recrudescence rather than being continuously mature individuals. In general, for males, abundant dmrt1 expression during preparatory and prespawning and spermatogenesis periods was seen, in contrast to a gradual decrease thereafter during spawning/spermination. This indicates that dmrt1 may have an important role during testicular recrudescence and particularly during spermatogenesis.

Hence, for all gonochoristic fish species investigated to date, the dmrt1 expression pattern was always shown to be intimately linked to male gonadogenesis and further differentiation (Table 1).

Dmrt1 expression in protogynous and protandrous hermaphroditic species

In hermaphrodite fish (protogynous or protandrous), the developmental pathways leading to either testicular or ovarian establishment have to be plastic and susceptible to the sex-inversion signals considerably beyond embryogenesis and early larval stages, whereas in gonochoristic species the developmental decision towards male or female is finally and irreversibly taken long before adulthood is reached. In this context, dmrt1 expression dynamics were consistently shown to parallel either the development (protogynous; black porgy Acanthopagrus schlegeli [30], gilthead seabream Sparus auratus [31]) or regression (protandrous; grouper Epinephlus coioides [32], wrasse Halichoeres tenuispinis [33], rice field eel Monopterus albus [34]) of the testes. This confirms the abovementioned role during testicular development and/or spermatogenesis. Of note, in pejerrey (Odontesthes bonariensis), a teleost with a temperature-dependant sex determination system, developmental expression of dmrt1 is perfectly correlated with the rearing temperature (up at male-determining temperatures and down at female-determining temperatures) [26].

Dmrt1 expression in fish and other vertebrates, what does it tell us?

In some fish species, dmrt1 expression is seen only in somatic cells, whereas other fish have clearly additional expression in the germ cell lineage (Table 1). This difference in cell types expressing dmrt1 might reflect species-specific differences in testicular structure and development. A dual dmrt1 cell lineage expression in Sertoli and germ cells is the hallmark of mammalian dmrt1s. Surprisingly, although mouse dmrt1 is detected in the bipotential gonad, knockout male mice have defects only during postnatal testis differentiation [35]. Although this observation might lead to the assumption that germline expression is dispensable, conditional dmrt1 inactivation in either the Sertoli cells or the germ cells indicated that mouse Dmrt1 is indeed required for radial migration of germ cells and survival of gonocytes. It is also required autonomously for proper Sertoli cell differentiation [36]. Hence, it is seen that mouse Dmrt1 might not play a major role during early testis differentiation, but rather appears to be required later for male gonadal differentiation. Interestingly, also expressed in the primordial gonads at the time of sex determination, the Z-linked dmrt1 gene in chicken [15, 37] and the W-linked DM-W gene in frog [14, 38] have been shown to be the major male and female determinants, respectively. Altogether, it appears that when earlier in the cascade of sex determination, the role of Dmrt1 is first to be an inducer of sex determination. Later on, when still or only expressed at later stages after the gonad is formed and being by implication at a more downstream position within the cascade, its task is restricted to a maintenance function essentially in Sertoli cells.

Other dmrt genes expressed in the fish gonads

The developmental expression of dmrt1 has been well studied in the context of gonadal induction and maintenance, illuminating its important function. But what about the other dmrt genes? Table 2 summarizes the expression pattern of these genes in fish during development and in the gonads. Although less-extensively studied, two main tendencies can already be deduced from these data. First, fish dmrt family members (Dmrt2, -3, -4, -5) exhibit conserved expression during the earliest stages of embryonic development in various organs, including the undifferentiated gonads. Second, later during development, these genes usually remain expressed in a subset of adult organs including spinal cord, brain and gonads. Noteworthy, male-specific gonadal expression could be observed for dmrt3 in medaka [39] and dmrt4 in medaka [39] and olive flounder [40] (Table 2). By contrast, in tilapia dmrt4 expression is exclusively detected in the ovary [41]. Finally, both male and female gonadal expression was reported for dmrt2 in medaka [39] and dmrt3 and -5 in zebrafish [42,43]. This expression discrepancy regarding the dmrt paralogues may indicate a possible functional switch between those in different phylogenetic lineages. Remarkably, when reported, non-dmrt1 gene expression generally occurs in developing germ cells (Table 2). In terms of inferred function(s), this incidentally indicates that paralogs of dmrt1 in fish, although obviously not involved in the first steps of gonadogenesis, might be implicated in the later processes leading to the proper development of germ cells.

Table 2.   Other dmrt genes having gonadal expression in fish.
GenesSpeciesGonadal expressionExpression in nongonadal tissuesRef
dmrt2 Oryzias latipes Testis and ovaryEmbryogenesis, somites, pharyngeal arches and brain[39]
dmrt3 Danio rerio Spermatogonia, spermatocytes
Developing oocytes
Olfactory placodes, neural tube
  Oryzias latipes TestisSpinal cord[39]
dmrt4 Oreochromis aureus OvaryEmbryogenesis, brain[41]
  Oryzias latipes TestisEmbryogenesis, nasal and otic placodes, telencephalon, branchial arches[39]
  Paralichthys olivaceus TestisDuring somotogenesis, gills and brain[40]
dmrt5 Danio rerio Testis (weak) and ovary (weaker): both in developing germ cellsEmbryogenesis, brain[42]

Effects of sex steroid hormones on gonadal identity and dmrt1 expression

Sex steroids have local, direct effects on germ cell development, but also act as endocrine hormones to influence other cell types and organs involved in sex differentiation. This multilevel control is especially complex in fish and involves a multitude of biochemical and physiological pathways to provide the necessary plasticity for gonadal development (see [4] for review). In that context, understanding the changes in dmrt1 expression following steroid treatment is of prime interest in order to link the molecular cellular events with the extracellular hormonal signalling system in gonad development.

Studies employing fish exposed to estrogens (or substances mimicking estrogen activities) are sparse but consistent in the reported effects on dmrt1 regulation (Fig. 1). In rare minnow [19], pejerrey [26] and zebrafish [44], estrogen exposure resulted in cessation of male gonad development and sex reversal. This was always correlated with a pronounced decrease in dmrt1 mRNA levels. Of note, in the same conditions, rainbow trout dmrt1 expression was only slowly inhibited [45], indicating that a reduced permissive amount of Dmrt1 expression might not be totally incompatible with active ovarian differentiation. In addition, in pejerrey, a fish with strong temperature-dependant sex determination, by combining different raising temperatures with E2 treatments, Fernandino et al. [26] could surmise that low dmrt1 and high cyp19a1a (aromatase) expression is connected to ovarian differentiation, whereas the opposite is true for testicular development. Furthermore, in females, cyp19a1a expression increased 1 and 2 weeks before the onset of dmrt1 and the first morphological signs of ovarian differentiation respectively, suggested that biologically active estrogen regulates dmrt1 expression [26].

Figure 1.

 The fish dmrt1 regulatory network or the current knowledge of dmrt1 gene regulation in fish. In many fish species, indirect dmrt1 transcriptional regulations have been described upon steroid treatment. (Upper) Steroid-induced dmrt1 regulation. Whereas feminizing substances having an estrogen-like activity (4-Nonylphenol and 17-alpha/beta estradiol) lead to dmrt1 transcriptional downregulation, masculinizing treatments (androgen, testosterone, aromatase inhibitors, estrogen antagonist or gnRHa) have been shown to conversely activate dmrt1 expression. (Lower) Proven direct regulations affecting dmrt1 transcription. In zebrafish, the transcription factor Sox5, although not itself sexually dimorphically expressed, was shown to directly downregulate dmrt1 transcription during development. In addition, in medaka and tilapia direct Dmrt1 transcriptional activity was revealed by respectively downregulating dmrt1bY and Cyp19a1a promoter activities.

Neurohormones (GnRHa) and either androgens, aromatase inhibitors or estrogen receptor antagonists have been shown to be very potent in manipulating the sexual phenotype of fish [4] (Fig. 1). These treatments, when applied to gonochoristic or hermaphroditic species, always resulted in a clear morphological masculinization process correlated with Dmrt1 upregulation (Fig. 1). Of note, some studies also pointed out the concomitant downregulation of cyp19a1a expression [46,47]. It then appears that dmrt1 could be one of the major regulators upstream of this enzyme in fish. It could be shown in trout that masculinizing treatments (1,4,6-androstatriene-3,17-dione) were inducing rapid and strong transcriptional upregulation of testicular markers like dmrt1, dax1 and pdgfra [46]. This upregulation was even interpreted as an essential step required for active masculinization. Into that direction, Dmrt1 and Dax1 have recently been shown to directly downregulate cyp19a1a promoter activity in the fish ovary [47,48]. Given the abovementioned observation that estrogens repress male differentiation it appears that, once initiated, factors of the male pathway downregulate the hormone. Hence, a feedback loop between dmrt1, cyp19a1a, and by implication the estrogen/androgen balance, becomes apparent. Dmrt1 expression modulation upon steroid treatments could then be a key effector of the induced gonadal identity change (Fig. 1). Similarly, in chicken, it could be shown that Dmrt1 also downregulates aromatase expression [37]. Overall, it is now clear that, at least in fish Dmrt1-regulating aromatase expression and by implication the estrogen/androgen balance that would also feedback (negatively or positively respectively) on dmrt1 expression, creates a complex regulatory loop combining transcriptional regulation with steroid hormonal activity (Fig. 1). The main question remaining is whether this loop aims only at activating the male pathway, or repressing the female one, or both.

In zebrafish, the transcription factor Sox5, although not itself sexually dimorphically expressed, was shown to directly downregulate dmrt1 transcription during development. This, together with a possible negative regulation of sox5 on cyp19a1a reported in the red-spotted grouper (Epinephelus akaara) [49] (Fig. 1), would constitute a perfect core for the transcriptional regulation network of dmrt1 and cyp19a1 in gonadal development.

Expression, regulation and functions of dmrt1a/dmrt1bY in medaka

In the medaka, which has XY–XX sex determination, dmrt1bY, the duplicated copy of dmrt1a on the Y chromosome was shown to be the dominant master regulator of male development [8], similar to Sry in mammals. Although many of the earliest cellular and morphological events initiated by Sry have been characterized, little is known about how the initial molecular activity of Sry is translated into cellular structure and organ morphology. Interestingly, Dmrt1, the ancestor of Dmrt1bY, is one of the downstream effectors of Sry in the male pathway.

In medaka, the duplicated copy of dmrt1 has acquired an upstream position in the sex-determining cascade. Remarkably, this evolutionary novelty, which is predicted to require a rewiring of the regulatory network, was brought about by co-option of ‘ready-to use’ pre-existing cis-regulatory elements carried by transposing elements. Further, it was shown that Dmrt1bY was able to bind to one of these elements, called Izanagi, within its own promoter, leading to significant repression of its own transcription [50] (Fig. 2). Interestingly, the autosomal Dmrt1a can bind to this site. Thus the Izanagi element enables the self- and cross-regulation of dmrt1bY expression by Dmrt1 proteins (Fig. 2). During the early stages, when the primordial gonad is determined towards testes, the exclusively expressed Dmrt1bY alone exerts sex-determining functions [9,51,52]. Noticeably, during this same period an 11-nucleotide protein-binding motif located in the 3′-UTR of dmrt1bY mediates gonad-specific mRNA stability [53] (Fig. 2). This motif is conserved in the 3′-UTRs of a wide range of dmrt1 orthologous genes from flies to mammals, indicating that different systems may employ an evolutionary conserved RNA regulatory mechanism for this gene [53].

Figure 2.

 Medaka dmrt1a/dmrt1bY regulations and functions. Grey arrows illustrate the different levels for which active dmrt1a/dmrt1bY regulation mechanisms could be shown. Transcriptional regulation: the feedback autoregulation of dmrt1bY promoter activity and transregulation by its paralogue Dmrt1a is a key mechanism of dmrt1bY transcriptional tuning. Post-transcriptional regulation: a highly conserved cis-regulatory motif directs differential gonadal synexpression of dmrt1 transcripts during gonadal development. Post-translational regulation: Dmrt1a and Dmrt1bY have a short half-life and consequently a high turnover. Functions: Dmrt1bY inhibition of germ cell proliferation might be part of its known male determining function.

Later during development of the juvenile fish and in the adult testes, where both dmrt1 genes have been shown to be expressed, it is of note that the newly generated duplicate dmrt1bY is kept back under tight transcriptional regulation of the ancestral dmrt1a gene [53]. In addition to the transcriptional regulation events, it could be shown that at any developmental stages, Dmrt1bY protein was subject to an intensive turnover due to rapid degradation [54].

With respect to its biochemical function, Dmrt1bY and the other Dmrt1s also in fish appear to act as transcription factors. This is evident from the nuclear localization of Dmrt1 fusion proteins [54,55] and studies showing direct effects of Dmrt1 on reporter gene expression as well as binding to a cognate motif in electric mobility shift assays [47,50].

Finally, linking the earliest sexual dimorphic trait to its expression dynamic, Dmrt1bY was shown to be possibly responsible for the male-specific primordial germ cell mitotic arrest [55] (Fig. 2). Indeed, functional evidence showed that expression of Dmrt1bY leads to negative regulation of male primordial germ cell proliferation prior to sex determination at the sex-determination stage [55]. This suggests that in XY medaka males, Dmrt1bY-driven primordial germ cell number regulation, as well as determination of pre-Sertoli cells, is the primary event by which the whole gonad (germline and soma) would be specified through a directional cross-talk from pre-Sertoli and Sertoli cells with the primordial germ cells. Interestingly, at this point, a parallel can be drawn with Dmrt1 function studies in mice. The lack of dmrt1 in mutant mice caused a high incidence of teratomas and resulted in a failure of germ cells to arrest mitosis [56]. Thus, Dmrt1 in mice and similarly Dmrt1bY in medaka appear to be regulators of germ cell proliferation.


To conclude, it seems that the longstanding hypothesis suggesting that the molecular sequence of sex-determination cascades might disobey the conventional rules of evolutionary developmental is now very well supported experimentally by data gathered in fish. Indeed, regarding Dmrt1, it is now obvious that because of consistent expression patterns in the gonads, and although necessarily acting at different stages of the sex-determining cascade, these effectors must individually fulfil similar and highly conserved functions. Hence, beyond the fish sphere, data recently published in Xenopus and chicken (see [14, 15] this minireview series) about dmrt genes being demonstrated to be of first importance for gonadal determination support the scheme that, whatever the sex-determination system, more comparative studies of dmrt1 are required in order to draw the first lines of a global core regulatory network for sex determination.