SRY and the Standoff in Sex Determination

Abstract

SRY was identified as the mammalian sex-determining gene more than 15 yr ago and has been extensively studied since. Although many of the pathways regulating sexual differentiation have been elucidated, direct downstream targets of SRY are still unclear, making a top down approach difficult. However, recent work has demonstrated that the fate of the gonad is actively contested by both male-promoting and female-promoting signals. Sox9 and Fgf9 push gonads towards testis differentiation. These two genes are opposed by Wnt4, and possibly RSPO1, which push gonads toward ovary differentiation. In this review, we will discuss the history of the field, current findings, and exciting new directions in vertebrate sex determination.

X, Y, AND THE GONAD

MOST ADULT MAMMALS exhibit clear sexual dimorphism that manifests externally and internally. Early studies of the sexes catalogued anatomical and endocrinological differences between males and females while the big question remained elusive: What makes them different to begin with? The first genetic clue came when Theophilus Shickel Painter (1) discovered that all males were XY and all females were XX. This led to the conclusion that genetic sex is determined when an X- or Y-bearing sperm meets an X-bearing egg. However, during early embryonic development, the two sexes are indistinguishable. The link between the heteromorphic sex chromosomes and gender was a big step, but the relationship between the sex chromosomes and the mechanism that determines male vs. female fate was still unclear.

The dimorphic development of male and female embryos originates with a fate decision that occurs in the fetal gonad. This unique bipotential organ begins as an indifferent tissue. At about 7–8 wk of development in humans, or midgestation in the mouse, the gonad adopts a testis or ovary fate. The importance of the gonad as the regulator of sexual development was first demonstrated by Alfred Jost (2). Working in France at the end of World War II, Jost performed a seminal embryological experiment that revealed the connection between the gonad and sexual fate and set the tone for the study of sex determination for decades. Jost removed undifferentiated gonads from fetal rabbits in utero and found that all of the embryos, regardless of their X/Y composition, developed female anatomy including the development of an oviduct, a uterus, a vagina, and female external genitalia. From these experiments, he concluded that male sexual development requires a gonad whereas female sexual development could be considered a default state that can be established in the absence of a gonad.

Jost demonstrated that there are at least two active factors associated with the testis that promote male sexual development. One of them is testosterone, which actively promotes male development of the epididymis, vas deferens, and male genitalia. However, Jost also showed that a second factor produced by the testis [now known to be anti-Müllerian hormone (3)] is responsible for inducing regression of the Müllerian duct, the antecedent of the oviduct, uterus, and vagina in females. Based on these experiments, Jost predicted the existence of a male sex-determining pathway the primary role of which was to regulate development of the gonad as a testis. The idea that male sexual development was active and female sexual development was passive emerged from this work and formed the conceptual framework for subsequent research.

THE Y OF SRY

How do the X- and Y-chromosomes alter the sexual fate of the gonad? After the work of Painter and Jost, it was not clear whether the presence of the Y or the lack of two X’s was responsible for male development. The answer came in 1959 when two groups examined the human conditions known as Turner’s Syndrome and Klinefelter’s Syndrome. Individuals with Turner’s Syndrome are XO (having only one sex chromosome) and are phenotypically female, whereas individuals with Klinefelter’s Syndrome are XXY and phenotypically male (4,5). This positively identified the Y-chromosome as the factor that engenders maleness, and the study of how the Y was directing male development began in earnest. Taking a genetic approach, groups began to whittle down the Y-chromosome looking for the testis-determining region (called TDYin humans, or testis-determining factor, Tdf, in mice). In 1990, the Y was pared down to a single gene SRY (sex-determining region of the Y; Sry in mice) thought to hold the secret of the switch between male and female development (6).

Here the sex chromosome mapping analysis and embryological experiments began to converge. Consistent with the implications of Jost’s experiments, Sry was shown to be expressed in the XY gonad (7,8). Conclusive proof came shortly thereafter with the aid of transgenic mice. XY mice lacking Sry develop an ovary and follow female sexual development pathways (9,10), and XX mice with a transgenic autosomal copy of Sry develop a testis and male secondary sex characteristics, although they are sterile in the absence of a Y-chromosome, which is required for spermatogenesis (11,12). These experiments firmly established the relationship between the Y-chromosome, the gonad, and sex determination.

SRY IN SUPPORTING CELLS

Mouse chimera experiments conducted before the time that Sry was isolated predicted that the male sex determinant would be required only in the Sertoli cell lineage (13,14). These chimeras, constructed between XX and XY embryos, showed a strong bias for XY cells in the Sertoli cell lineage, but not in other lineages of the testis (14). Many experiments suggest that, consistent with theoretical predictions, the cells that are precursors to the Sertoli cell lineage are the same cells that can give rise to follicle (granulosa) cells in the ovary (12,13,14,15,16,17). Therefore, this lineage is referred to as the “supporting cell lineage” based on its role in supporting development of germ cells in both sexes.

Sry is expressed in these supporting cell lineage precursors in the XY gonad. Primary sex determination, at least in mammals, appears to be focused on the cell fate decision that occurs in this precursor population when cells chose to differentiate as either Sertoli or follicle cells. Both the spatial and temporal regulation of Sry levels are critical for its ability to pattern the gonad. Sry is expressed in XY gonads, specifically in cells that initiate differentiation as Sertoli cells, at the stage when the gonad first forms as a bipotential primordium (7,8,15,19,20,21,22). The gonad represents a unique environment because ectopic expression of Sry outside this tissue does not lead to differentiation of Sertoli cells (23). The 5′-region of the Sry gene has been shown to be hypomethylated in the gonad, but hypermethylated in other tissues (24). Low or delayed expression leads to the development of a gonad containing a mixture of male and female tissues (an ovotestes) (20,25). Although expression of Sry may be transient in individual cells, based on the persistent Sry transgenic reporters, Sry-enhanced green fluorescent protein and Sry-alkaline phosphatase, it is likely that all Sertoli cells express Sry at some time during their differentiation (15,21).

SRY THE PROTEIN

SRY was the founding member of the Sox [Sry-related high-mobility group (HMG) box] family of transcription factors, although it appears to have evolved from an X-linked member of the family, Sox3 (26,27,28). The defining characteristic of this family is the presence of a HMG DNA-binding domain (29). SRY is thought to bind and bend DNA and may also be involved in chromatin remodeling (29,30,31,32). Although a binding site for the protein was predicted through PCR selection assays, the 6-bp sequence occurs so frequently in the genome that it has not been useful in defining targets of SRY in vivo (31,32,33). Consistent with the importance of the DNA binding function of SRY, most sex-reversing mutations occur within the HMG domain (33). Furthermore, this domain is the only conserved feature across mammalian SRYs. Regions outside this domain have evolved so rapidly that they are difficult to align between species (34).

As would be expected of a transcription factor, SRY is imported into the nucleus, and both importin-β and calmodulin have been shown to play a role in its translocation in vitro (35,36). From there, the water becomes murkier because SRY has been proposed to act as both an activator (37,38) and a repressor (39,40,41) of transcription. One report has even suggested that SRY may act as a direct modulator of splicing (42).

BELOW THE SWITCH

Only 10% of sex reversal cases are linked to mutations in SRY, indicating that other genes must also be critical in the sex determination pathway. The first gene known to be expressed downstream of Sry is Sox9, a closely related family member. Sox9 is expressed in all Sertoli cells, and both loss and gain of function evidence indicates that it is critical to initiate the testis pathway. Deletion of the Sox9 gene leads to male-to-female sex reversal in humans (43,44,45) and mice (46), whereas duplication or overexpression of Sox9 leads to female-to-male sex reversal (47,48,49). These experiments indicate that Sox9 can function as a substitute for Sry itself, even producing fertile males when expressed at sufficient levels in XY Sry null embryos (50). Sox9 may be the only direct target of SRY; however, direct evidence for regulation of Sox9 by SRY is still missing (16,21).

Interestingly, only an average of 90% of Sertoli cells in XX↔XY chimeras were XY cells (14). The occurrence of XX cells in this population indicated that non-cell-autonomous factors must also operate during mammalian sex determination and must be capable of recruiting XX cells to differentiate along the Sertoli pathway. Although at one time this finding was puzzling, it has proved to be conceptually enlightening. Several extracellular signaling pathways have been implicated in recruiting the cells of the gonad to the testis pathway. For example, prostaglandin D2 (PDG2) promotes the Sertoli cell fate and induces Sox9 expression in XX cells in vitro (16,51,52,53,54), whereas fibroblast growth factor 9 (Fgf9) has been shown to induce Sox9 in XX cells in vitro and is required for maintenance of Sox9 expression and testis differentiation in vivo (17,55,56). Further support for the role of extracellular signaling molecules in development of the gonad comes from the observation that in XX↔XY chimeras where the XY component is more than 30%, the gonad still differentiates as a testis (14). These findings strongly argue for the ability of extracellular signals to recruit cells to the testis developmental pathway.

Proliferation is also critical to testis development and is up-regulated in XY gonads soon after Sry expression begins. This results in the expansion of Sertoli precursor cells (56,57,58). Disruption of proliferation in vivo through disruption of the Fgf9 pathway, or through the use of proliferation inhibitors, leads to male-to-female sex reversal (17,58). Taken together, the XX↔XY chimera experiments, along with the data showing that reduced expression of Sry leads to sex reversal (20,59,60,61), have led to the interpretation that there is a critical threshold effect in the gonad when the primordium is balanced between ovarian and testis fates. Both the level of SRY produced by individual cells and the number of cells producing it may contribute to this threshold effect.

This precarious balance of the gonad between these two developmental pathways is likely what confers its dual potential. Normally, the system employs complex cell signaling loops that reinforce a single fate decision in the supporting cell lineage and recruit all gonadal cells behind the testicular or ovarian pathway. Defects in these reinforcing signaling loops may explain many disorders of incomplete sexual development that manifest as gonadal dysgenesis, ovotestis formation, ambiguous ductal or genitalia development, or a combination of these features.

THE TROUBLE WITH XX MALES

When SRY was discovered in 1990 and shown to encode a DNA-binding protein, it fit well with the prediction that the sex-determining gene would be a master regulator that activated many downsteam pathways leading to testis development. However, the persistent fly in the ointment was the occurrence of XX males that show no evidence of the SRY gene. If the male sex-determining gene is required in the bipotential gonad to activate testis development, how does an XX gonad establish testis organogenesis in the absence of SRY? A parsimonious explanation proposed by McElreavey et al. (39) and Goodfellow and Lovell-Badge (62) was that SRY actually acts as a negative regulator of a hypothetical gene Z, the role of which is to repress the testis pathway. In such a double-negative system, it is easy to explain how a single mutation rendering Z inactive could lead to initiation of testis development in the absence of SRY. Recent experiments in the field have developed this concept of inhibition further.

REMODELING THE Z HYPOTHESIS: SEX IN THE BALANCE

The Fgf- and Wnt-signaling pathways have been implicated in regulating the fate of the gonad. Mutations in Fgf9 lead to male-to-female sex reversal, whereas mutations in Wnt4 lead to partial female-to-male sex reversal (17,63,64,65). In the absence of Fgf9, Sry is expressed normally in the XY gonad and Sox9 expression is initiated. However Sox9 expression is rapidly silenced in the absence of Fgf9, and the cells of the XY gonad express genes characteristic of the female pathway, namely, Wnt4, Follistatin, and Bmp2 (17,65). These data indicate that Fgf9 is required to promote the male pathway and suppress the female pathway.

Wnt4 appears to play a reciprocal role in the female pathway. In the absence of Wnt4, Fgf9 and Sox9 expression are transiently up-regulated in the XX gonad. Activation of Sox9 occurs in this case in the absence of the Sry gene, indicating that expression of Sox9 does not require Sry but can be accomplished simply by down-regulating Wnt4. This suggests that Wnt4, like the hypothetical gene Z, normally represses the male pathway (namely, Sox9) in the XX gonad. However, unlike Z, Wnt4 is also required for male development (64). The current model is an evolution from a simple repressor of a repressor to a model of balanced opposition, both sides of which may be important to establish the initial bipotential field of cells.

Sex reversal in Wnt4 null XX gonads is incomplete, suggesting that there may be additional factors that reinforce the female pathway. The canonical Wnt signaling pathway acts by stabilizing β-catenin, which then acts as a cofactor for transcriptional activation. Recently, mutations in the R-SPONDIN1 (RSPO1) gene were identified in human XX patients with testis development (66). Importantly, this is the first human mutation that results in complete female-to-male sex reversal. RSPO1 has also been shown to activate the β-catenin signaling pathway (67), which raises the possibility that Wnt4 and Rspo1 act cooperatively to block the male pathway in XX gonads. Duplication of the human chromosome 1 region, where both WNT4 and RSPO1 map, led to disruption of testis development in humans (68). Surprisingly, overexpression of Wnt4 had only minor effects on testis development in mice (68,69,70). This could be related to an inappropriate spatial or temporal expression of these transgenes, or it could mean that ectopic expression of both Wnt4 and Rspo1 would be required.

There is a precedent for the regulation of cell fate decisions through this seemingly disparate collection of molecules. During chondrogenesis, Wnt signaling activates β-catenin, whereas and Fgf signaling activates Sox9 signaling, respectively, activate β-catenin and SOX9 in chondrocyte precursors (71,72,73,74,75). In chondrocytes, β-catenin and SOX9 have been shown to bind together and trigger the degradation of both proteins until one of the two prevails (73,74).

By analogy, this raises the possibility of an intracellular stand-off between these pathways in supporting cell precursors during the bipotential stage of gonad development (Fig. 1). Primary sex determination might occur when one of these pathways gains the upper hand, silences the other pathway, and triggers a cell fate decision resulting in the differentiation of either Sertoli or follicle cells. Although competition between SOX9 and β-catenin is an appealing concept from many points of view, Wnt4 can also signal through noncanonical pathways (76,77), and no evidence has been presented thus far to indicate that Wnt signaling acts through β-catenin in the gonad.

Figure 1.

Reinforcing Loops that Propagate Testis or Ovary Fate

The somatic precursors of the supporting cell lineage can become Sertoli cells or follicle cells dependent upon whether male- (FGF9) or female- (WNT4) promoting signals dominate the field. In a male environment, these cells become Sertoli cells, organize testis development, promote Leydig cell development, and regulate germ cells that arrest in G0/G1 [PGC (G1)]. In a female environment, the somatic precursors become follicle cell precursors. Their organization into follicle structures is dependent on meiotic germ cells [PGC (Mei)]. It is likely that reinforcing signals (dotted lines) exist between fetal follicle or Sertoli cells and theca or Leydig cells, respectively, although this has not yet been shown. Question marks denote putative interactions between Leydig or theca cells and germ cells.

These observations paint a picture that is markedly different from the classic male-active/female-passive model. It now appears that the bipotential gonad is the battleground between two active and opposing signaling pathways that converge on the regulation of the Sox9/Fgf9 loop (Fig. 2). Sox9 is a conserved element of the male pathway in all vertebrates examined, and even in one species of fly (78). In mammals, Sry tips this balanced Fgf/Wnt antagonism toward testis development by boosting expression of Sox9. This model provides an easy explanation for the occurrence of XX individuals who develop a testis in the absence of Sry. It is also easy to imagine how this plastic system could respond to diverse regulators such as temperature or behavioral cues that control sex determination in other species.

Figure 2.

Model of Opposed Signals in Mammalian Sex Determination

A, In the bipotential gonad, male-promoting (SOX9 and FGF9) and female-promoting (WNT4 and possibly RSPO1) hold each other in check. B, The presence of SRY (XY) reinforces the positive feedback between SOX9 and FGF9, which then out-competes the female signals and drives testis differentiation. In the absence of SRY (XX), the female-promoting signals shut down the male loop and drive ovarian differentiation.

SPECIES WITHOUT SRY

SRY is critical for male development in most mammals, but what about species in which there is no SRY? A few rodents appear to have lost SRY, and even a Y-chromosome, but retain fertile males and females (79). Also, to be precise, only therian mammals have SRY. Monotremes have X- and Y-chromosomes (in fact, the platypus has five Xs and Ys), but no SRY gene (80,81,82). Because SRY likely evolved after the therian/monotreme split between 310 and 130 million years ago (80,83), other vertebrates must have SRY-independent sex-determining mechanisms.

Many alternative sex-determining mechanisms do exist. Interestingly, the primary switches controlling sex determination are highly divergent across species, far more than would be expected considering the conservation among other developmental pathways. Two commonly cited examples of strong evolutionary conservation are the Pax6/eyeless genes that govern development of the eye from Drosophila to humans (84) and the Hox genes that regulate the body axis across phyla (85). However, sex determination stands apart as an emphatic exception to this rule. The reason for the radical variations in sex-determining mechanisms, even among closely related species, is unclear and often debated (86).

Traditionally, sex-determining mechanisms are broadly spit into two categories: genetic sex determination (GSD) and environmental sex determination (ESD). However, even within a single category, there is diversity. For example, mammals and birds use GSD, but mammals have male heterogamety (XX/XY), whereas birds have female heterogamety (ZW/ZZ) (87). Furthermore, the sex chromosomes of mammals and birds appear to have evolved from entirely different autosomes (88,89). On the other side, there is variation in temperature-dependent sex determination, a common form of ESD in which the temperature of egg incubation determines the sex. Temperature-dependent sex determination is found in all crocodilians and many turtles, but the ranges of temperatures, or the end result of a given temperature, are quite disparate (90,91,92). To further complicate matters, there is also no phylogenetic consistency in sex-determining mechanisms across groups of vertebrates. ESD and GSD (either XX/XY or ZW/ZZ) appear to have evolved independently multiple times within fish, lizard, and amphibian lineages (86).

It was once proposed that GSD and ESD were so different as to be mutually exclusive (93,94). This has been reconsidered in light of evidence suggesting that all sex-determining mechanisms may have underlying genetic components (95,96). Indeed, ESD and GSD have even been shown to coexist within some species (97). Now, rather than being seen as Manichean opposites, GSD and ESD are probably best viewed as distinct engines with partially overlapping mechanisms (86,98).

So what, if anything, can these radically different methods of sex determination tell us about mammalian development? Despite the variability of the switches at the top of the sex-determining cascade, the structure of the adult testis is morphologically well conserved in all vertebrates. This has led to the idea that the pathways downstream of the switch are conserved. That is to say, regardless of how evolution has altered the top of the cascade, the movers and shakers down the chain that direct gonadal differentiation are the same. Comparative analysis between mammals, the most well studied, and other vertebrates have lent support to this model. Almost every mammalian gene known to play a role in early gonad development has been found to have similar expression patterns in birds and reptiles (87,91,99,100,101). Notably, Sox9, the gene that is the likely vortex of the fate determination decision in the supporting cells of the gonad, appears to be conserved in testis development in every vertebrate examined to date.

One attractive aspect of the current opposing pathways model is that it can parsimoniously explain the existence of a wide array of sex-determining mechanisms. The evolutionary appearance or disappearance of mechanisms that affect the balance between Sox9/Fgf9 and Wnt4/Rspo1 could be very rapid. For example, temperature might affect the biochemical activity of a component in the Wnt or Fgf pathway. Alternatively, or additionally, allelic variants might evolve that segregate in the population and push the gonad in one direction or the other, eventually founding a new sex chromosome (102). Multiple mechanisms may simultaneously exist within a species, one of which emerges as the predominant modus operandi at any one time (86,97,98). However, the underlying system could remain very plastic in its ability to respond to a variety of sex-determining switches.

PROPAGATING THE DECISION

Sex determination is a strongly canalized process. Once a fate decision is made within the supporting cell lineage of the gonad, the effects of feedback loops at hormonal and physiological levels reinforce the male or female pathway. This may also be true at the intracellular and intercellular levels. Feed-forward loops that reinforce intracellular decisions have proved to be a common theme across many developing tissues, and the gonad is no exception. As previously stated, Sox9 works in a feed-forward loop with Fgf9. These two genes up-regulate each other and generate a cellular momentum moving toward Sertoli cell fate. Layered onto this intracellular loop is an extracellular mechanism that works to recruit other cells in the gonad to the developmental plan through the expression of additional secreted signals. Paracrine signals, such as prostaglandin D2 (51,52,53,54), also reinforce the male pathway. Leydig cells, the male steroidogenic cells, develop as the result of desert hedgehog (Dhh) and platelet-derived growth factor A (PdgfA) signals produced by Sertoli cells (103,104). Furthermore, morphogenetic changes, such as testis cord formation, appear to reinforce the differentiation of Sertoli cells and the establishment of the male pathway. Wnt signaling represses this male loop and drives the differentiation of follicle cells. Although it is less well understood, this promotes a female signaling network between follicle cells, the female steroidogenic precursors (theca cells), and germ cells. This establishes an ovarian environment hostile to male development (Ref.105 and Fig. 2). These local reinforcing environments in the gonad likely explain the observation that ovarian and testicular portions of an ovotestis occupy distinct domains.

GERM CELLS IN THE CONTESTED PRIMORDIUM

Although germ cells are not critical for testis morphogenesis, they are required for follicle formation in the ovary (106). Most experiments in mammals are consistent with the idea that the ovarian pathway initiates normally in the absence of germ cells but fails to result in follicle formation and a functional ovary at birth. When germ cells are absent, the production of testosterone from the testis is not lost; however, the cycling hormonal balance in the ovary is completely disrupted.

Indirect evidence indicates that germ cells initiate dimorphic development soon after they enter the gonad. Male and female germ cells show differences in sensitivity to mutations in Fgf9 and to Mvh (mouse vasa homolog) by 11.5 d post coitum (65,107). Germ cells that enter an ovary initiate meiosis soon after 13.5 d post coitum. Retinoic acid, which is likely produced by the adjacent mesonephros, has been shown to induce this process in XX gonads (18,108). In contrast, germ cells that enter a testis arrest in G0/G1 of the mitotic cycle until after birth. Sertoli cells play a critical role in blocking the entry of germ cells into meiosis in XY gonads, perhaps through the production of Cyp26b1, an enzyme that degrades retinoic acid (18). There is evidence that meiotic germ cells can disrupt testis development in vitro, inhibiting male-specific migration and cord formation (105). Thus development of germ cells in concordance with the sexual fate of the somatic cells in the gonad may be important not only for fertility, but may be critical to reinforce the fate of the gonad and further canalize development of a testis or ovary.

FINAL REMARKS

Many questions are still outstanding about how these intracellular and extracellular pathways work to establish sex determination in the gonad. It is clear that the same pathways that regulate the development of many other organs are also involved in the development of the gonad. However, the system is adapted to regulate the unique dimorphic developmental fate of this primordial organ. During the bipotential phase of gonad development, opposing pathways hold the cells of the gonad in a bipotential state until it is launched down the male or female pathway. SRY remains central as the primary trigger for the initiation of testis development in mammals. Whether there is a parallel transcription factor regulating the initiation of ovary development (i.e. a female-determining gene) is unknown, but not theoretically required. Regardless of whether or not one is found, it is clear that initiation of the ovarian pathway involves the active regulation of many genes and is not simply a passive/default developmental process.

Although it has been valuable to place SRY within this signaling network in the last few years, the mechanism through which the gene functions has not yet been revealed. Future work will address this important unresolved issue and will focus on the integration between the intracellular/intercellular signaling pathways that regulate cell fate and structural morphogenesis of the testis and ovary. Current efforts to establish networks of gene interactions will be critical to provide a framework for future investigations.