Epigenetics of sex determination and gonadogenesis

Abstract

Epigenetics is commonly defined as the study of heritable changes in gene function that cannot be explained by changes in DNA sequence. The three major epigenetic mechanisms for gene expression regulation include DNA methylation, histone modifications, and non-coding RNAs. Epigenetic mechanisms provide organisms with the ability to integrate genomic and environmental information to modify the activity of their genes for generating a particular phenotype. During development, cells differentiate, acquire, and maintain identity through changes in gene expression. This is crucial for sex determination and differentiation, which are among the most important developmental processes for the proper functioning and perpetuation of species. This review summarizes studies showing how epigenetic regulatory mechanisms contribute to sex determination and reproductive organ formation in plants, invertebrates, and vertebrates. Further progress will be made by integrating several approaches, including genomics and Next Generation Sequencing to create epigenetic maps related to different aspects of sex determination and gonadogenesis. Epigenetics will also contribute to understand the etiology of several disorders of sexual development. It also might play a significant role in the control of reproduction in animal farm production and will aid in recognizing the environmental versus genetic influences on sex determination of sensitive species in a global change scenario. Developmental Dynamics 242:360–370, 2013. © 2013 Wiley Periodicals, Inc.

INTRODUCTION

Epigenetics is a very exciting area of biology that currently is undergoing an impressive expansion. The scope of epigenetics is the subject of debate (Eccleston et al., 2007) and thus several definitions have been proposed (Bird, 2007). Here, I will use the definition suggested by Russo et al. (1996): epigenetics is “the study of mitotically and/or meiotically heritable changes in gene function that cannot be explained by changes in DNA sequence.” The term “heritable” in the definition has generated some confusion since actually it has two different implications: these changes are heritable from one cell to their daughter cells during mitotic cell division but also during the formation of gametes through meiotic cell division, and thus can be passed from parents to offspring (Gilbert and Epel, 2009). Accepted examples of epigenetic regulatory events or phenomena include mating type silencing in yeast, temperature-dependent vernalization in plants, position-effect variegation, gametic imprinting, and X-chromosome inactivation in mammals (Wakimoto, 1998; Brock and Fisher, 2005).

Epigenetic mechanisms for the regulation of gene expression typically include DNA methylation, modification of histones and histone variants, and the presence of non-coding RNAs (ncRNAs) (Brock and Fisher, 2005) and this is the order that will be followed in this review. However, it should be noted that, first, these mechanisms occur simultaneously and, second, that the order does not necessarily reflect what is actually occurring at the cellular level. Thus, in an attempt to put forward an operational description of epigenetics, Berger et al. (2009) proposed three categories of signals that sequentially operate in the establishment of a stably heritable epigenetic state. The first, called epigenators, is a signal received from the cell environment, e.g., differentiation signals or temperature variations. Everything occurring upstream of the first event on the chromosome is part of the epigenator signal; not only the environmental cue itself but also the subsequent signaling pathways (Fig. 1). The epigenator signal may be transient, enough to trigger the epigenetic phenotype but not necessary for its maintenance. The second type of signal, the epigenetic initiator, is a responding signal in the cell that establishes a local chromatin context at a precise location in response to the epigenator signal. Epigenetic initiators need some sort of sequence recognition since they must be capable of identifying the precise coordinates of a chromatin structure (Berger et al., 2009). Examples include DNA-binding proteins and ncRNAs such as XIST, which is sufficient to silence the mammalian X chromosome. The epigenetic initiators do not dissipate after their actions since in general they are signals that involve an autoregulatory positive feedback loop. The third type of signal, the epigenetic maintainer, is a sustaining signal that is not sufficient to initiate but rather perpetuates the chromatin change, carrying the epigenetic state through subsequent generations. Examples of the latter include DNA methylation, histone modifications, and histone variants (Berger et al., 2009). Nevertheless, one important feature of epigenetic modifications is their possibility of reversibility. Thus, epigenetic modifications represent a balance between maintenance and reversibility (James and Renard, 2010).

Figure 1. Integration of genetic and environmental information to produce a given phenotype through epigenetic regulatory mechanisms. Three levels are depicted: (1) Epigenators such as temperature; (2) Epigenetic initiators such as DNA-binding proteins (DNA-BP) or non-coding RNAs (ncRNA) such as XIST attach to DNA (double blue lines); and (3) Epigenetic maintainers such as DNA-modifying enzymes like DNA methyltransferases (DNMT), histone tail (purple balls) modifying and demodifying enzymes such as histone acetyl transferases (HAT) or histone deacetylases (HDAC), or histone variants (green circle vs. red circle). The epigenetic signaling pathway is based on Berger et al. (2009) and the possible outcomes on Turner (2007).

During development, cells differentiate and acquire identity in response to interactions with each other or with the environment. Changes in gene expression patterns are an essential feature of cell differentiation, and the conservation of these patterns is a key component of maintaining cell identity (Brock and Fisher, 2005; Kiefer, 2007). Thus, epigenetic mechanisms provide organisms with the ability to modify the activity of their genes in response to changes in the internal or external environment, integrating genomic and environmental information to generate a particular phenotype (Turner, 2009). One of the most important developmental processes for the proper reproductive competence and perpetuation of species is sex determination and differentiation of the gonads.

Sex determination is the genetic or environmental process by which the sex (gender, male or female) of an individual is established in a simple binary fate decision. There are two major types of sex-determining mechanisms: genotypic sex determination (GSD), where sex is determined at conception and genetic differences are expected between the sexes; and environmental sex determination (ESD), where there are no consistent genetic differences between sexes and sex is determined after fertilization in response to an environmental cue. On the other hand, sex differentiation is mainly the process by which an undifferentiated gonad is transformed into an ovary or a testis but also encompasses other aspects of differentiation, including the development of associated sex ducts and genitalia and the establishment of sex-specific brain differences (Valenzuela and Lance, 2004; Penman and Piferrer, 2008). Embryonic gonads are thus unique in that they are the only organs that can develop in two mutually exclusive phenotypes. Gonadal somatic cells are initially bipotential, subjected to the action of antagonistic signaling pathways and transcription networks (Kim and Capel, 2006). Sexual fate is determined by activating the testis or ovarian pathway and repressing the alternative pathway with many genes being expressed in a sexually dimorphic manner (Munger and Capel, 2012). In this developmental context, where many genes must be activated or suppressed in a tight spatial and temporal fashion (Barrionuevo et al., 2012), epigenetic mechanisms for the regulation of gene expression become of critical importance. Further, the heritability of epigenetic marks adds a new twist to the complexity of genetic versus environmental control of gene expression during the crucial events of gonadogenesis.

Here, I review recent studies that investigate epigenetic mechanisms related to sex determination and sex differentiation, referring mostly to vertebrates but invertebrates or plants are also mentioned when appropriate. The review is not meant to be comprehensive but, rather, it intends to emphasize emerging general features and generate interest for further research. On the other hand, chromatin transitions, specifically in germ cells, are still poorly understood when compared to those in somatic cells. The epigenetic transitions during germ cell development and entry into meiosis have been reviewed elsewhere (Ewen and Koopman, 2010; Kota and Feil, 2010). In addition, the epigenetic mechanisms of gene regulation that take place in the gonadal germ cells after sex differentiation, i.e., during spermatogenesis (Khalil and Wahlestedt, 2008; McIver et al., 2012) and oogenesis (Bromfield et al., 2008), have been discussed elsewhere, as have the epigenetic marks in mature gametes and early embryo development (Hales et al. 2011). As for non-gonadal sex differentiation, increasing evidence suggests that epigenetic mechanisms can contribute to the establishment and maintenance of some aspects of gonadal hormone-induced activational and organizational effects on neuronal substrates during sexual differentiation of the brain. Epigenetic programming of neuroendocrine and behavioral phenotypes frequently occurs sex specifically (Levenson and Sweatt, 2005), pointing to sex differences in brain epigenetics as a possible determinant (Menger et al., 2010).

DNA METHYLATION AND SEX DETERMINATION

Methylation of the DNA molecule in the nucleus is one of the best-characterized epigenetic modifications. A methyl group replaces carbon 5′ of a deoxycitidine next to a guanidine (CpG). The reaction is catalyzed by a group of enzymes called DNA methyltransferases (DNMTs). There are two important DNMTs: DNMT1 methylates the unmethylated opposing pair of an hemimethylated site and thus is called maintenance DNMT1 because it is responsible for the copy of the existing methylation profile during cell division (hence the heritability of the epigenetic mark). The other is DNMT3 and it places methylation marks on previously unmethylated CpGs and thus is responsible for the de novo DNA methylation (Hermann et al., 2004). CpGs are usually methylated across the genome. CpG islands are regions of the genome with elevated CpG content that are normally associated with promoter or regulatory regions. Changes in methylation levels in these CpG islands are associated with gene expression regulation. Further, genomic regions with varying DNA methylation, e.g., between two developmental stages, are called differentially methylated regions (DMR). Two conditions related to development in which DNA methylation is present are X-chromosome inactivation and genomic imprinting. In the latter condition, differential DNA methylation patterns at control regions in the male and female germ line result in either maternal- or paternal-specific expression at a subset of genes in the genome despite identical DNA sequence of the two parental chromosomes (Strogantsev et al., 2012).

In plants, epigenetic mechanisms involved in sex determination have been described so far in few cases, including maize, Zea mays ssp. mays, where a factor maintains repressed epigenetic stages (Parkinson et al., 2007). However, how widespread these mechanisms are in plant sex determination is at present unknown. One of these cases concerns the melon, Cucumis melo, the model plant for Cucurbitaceae. In the melon, the gene CmACS-7 encoding an ethylene biosynthesis enzyme in andromonoecious individuals (i.e., bearing male flowers in the main stem and female or hermaphrodite flowers on axillary branches) and the transcription factor CmWIP1 interact to control the development of male, female, and hermaphrodite flowers. In gynoecious lines (bearing only female flowers), the transition from male to female flowers results from epigenetic changes in the CmWIP1 promoter caused by the insertion of a transposon, Gyno-hAT. This transposon is required for initiation and maintenance of the spreading of DNA methylation to the CmWIP1 promoter (Martin et al., 2009).

Gorelick (2003) hypothesized that dioecy and sex chromosomes originated in ancestral diploid hermaphrodites as a pair of autosomes in which one chromosome had more methyl groups near to a sex-controlling region than did its homologue. Methylation would suppress transcription, including loci for gamete production hence transforming hermaphrodites into males or females. Differential methylation would also suppress recombination, increasing the speed of Muller's ratchet. The same hypothesis was also postulated by Jablonka (2004). One of the predictions of this point of view was that species with ESD require homomorphic sex chromosomes and that small environmental changes can alter the methylation patterns of sex-related loci, therefore determining the sex of individuals (Gorelick, 2003). It is well established that environmental factors can alter gene expression and thus the resulting phenotype. A good example is ESD, whereby changes in one environmental factor during critical periods of early development are able to influence the sex of the offspring. Temperature-dependent sex determination (TSD) is a form of ESD, and sex ratio shifts in response to temperature are common in fish and reptiles. In non-mammalian vertebrates, aromatase (cyp19a1) is a key steroidogenic enzyme that converts androgens into estrogens, essential for ovarian differentiation. In fish and reptiles, temperature-induced masculinization is invariably associated with cyp19a1 gene repression (Ospina-Álvarez and Piferrer, 2008; Ramsey and Crews, 2009). However, the mechanism linking temperature during early development and sex ratios has been the subject of much debate (Lance, 2009).

In the European sea bass, a fish with a polygenic system of sex determination where genetics and temperature contribute equally to sexual fate (Piferrer et al., 2005; Vandeputte et al., 2007), and where the thermosensitive period (TSP) is located during the larval stages, well before the gonads form, Navarro-Martín et al. (2011) showed that juvenile males have double the DNA methylation levels of females in the cyp19a1 promoter. Exposure to high temperature coinciding with the TSP increased the methylation of this promoter in females, and an inverse relationship between methylation levels and cyp19a1 expression was observed, indicating that induced masculinization involves DNA methylation-mediated control of cyp19a1 gene expression. Different CpGs within the cyp19a1 promoter exhibited different sensitivity to temperature. However, the increased methylation of the cyp19a1 promoter was detected in the gonads but not in the brain, indicating that it is not a generalized effect of temperature. Furthermore, the effects of temperature were also observed in sexually undifferentiated fish and were not altered by estrogen treatment. Thus, methylation of the cyp19a1 promoter causes lower expression of cyp19a1 in temperature-masculinized fish. Functional studies in vitro showed that induced methylation of the cyp19a1 promoter suppressed the ability of Sf-1 and Foxl2 to stimulate transcription. These findings constitute the first example of an epigenetic mechanism mediating temperature effects on sex ratios in a vertebrate. In the same study, we observed that a CpG loci differentially methylated by temperature and adjacent to a Sox transcription factor–binding site was conserved across some species. Based on this study, it is tempting to suggest that DNA methylation of the aromatase promoter may be an essential component of the long-sought-after mechanism connecting environmental temperature and sex ratios in vertebrate species with temperature-dependent sex determination (Navarro-Martín et al., 2011). It will be interesting to find out whether a similar mechanism operates in reptiles with TSD, in which the TSP coincides with sex differentiation, i.e., in which the time lag between temperature exposure and sex differentiation observed in the sea bass does not apply. In addition, TSD has evolved independently many times. Thus, while it is likely that many mechanisms converge in aromatase regulation, there could be other mechanisms besides DNA methylation that could mediate temperature effects.

Many sequential hermaphrodites integrate both internal and external cues to regulate the process of sex-change. The black porgy (Acanthopagrus schlegeli) is a protandrous hermaphrodite fish of the family Sparidae that has been the subject of much work concerning the genes and molecular endocrinology involved in the male-to-female transition (Wu et al., 2010). During the first two reproductive cycles, the ovarian part of the gonad remains inactive and oocytes do not pass the primary oocyte stage. Furthermore, the inactive ovarian tissue could not be activated by estradiol-17β (E₂) treatment but instead by surgical removal of the testicular tissue, indicating that somehow the presence of testicular tissue was inhibiting the development of the ovarian tissues. No differences were observed in global methylation levels among the different ovarian phases (previtellogenic, vitellogenic, and mature oocytes). However, the DNA methylation level of the cyp19a1 promoter was higher in inactive ovaries than in active ones, suggesting that cyp19a1 expression was controlled by an epigenetic mechanism in addition to classical transcriptional activators of cyp19a1 such as Sf-1 and Foxl2 (Wu et al., 2012a). In contrast, the promoter of dmrt1, a gene that is not only important for testis differentiation but also for sexual fate determination and natural sex change (Wu et al., 2012b), showed no difference in the levels of DNA methylation at any developmental stage in the testis based on MeDIP assays (Guan-Chung Wu and Ching-Fong Chang, personal communication). Thus, changes in cyp19a1 promoter methylation are likely to be found between sexes in other gonochoristic species, as a response to temperature in sensitive species, and during the process of sex-change in hermaphroditic ones. Because of the combination of genetic and environmental influences on resulting sex ratios, fish, along with reptiles, represent a fecund group of animals in which to pursue further the study of the implication of epigenetic mechanisms in the processes of sex determination and gonadogenesis.

As we have seen above, DNMTs are essential for DNA methylation and there are several studies that relate the expression of DNMTs with different aspects of development in various animal models. However, few studies relate changes in DNMT expression specifically with sex determination or gonadogenesis. The specific roles for the different DNMT family members in de novo and maintenance methylation in germ cells of the developing mouse testis and ovary were investigated by La Salle et al. (2004). They compared the temporal expression patterns of the DNMTs in the male and female germ cells before, during, and after sexual differentiation, including the timing of methylation reprogramming events that take place between E10.5 and E12.5 in the primordial germ cells colonizing the embryonic gonads (Reik et al., 2001; Sasaki and Matsui, 2008). Real-time RT-PCR was used to study simultaneously the expression profiles of the three DNMT3 to search for candidate DNMTs that could be involved in establishing methylation patterns in both germ lines. DNMT1 (DNA methylation maintenance) was used as a control. Results suggest an interaction between DNMT3a and DNMT3l based on their corresponding gene expression profiles, the latter being the predominant DNMT3 enzyme present at high levels in the postnatal female germ line at the time of acquisition of DNA methylation patterns. DNMT1 and DNMT3b expression levels peaked concomitantly, shortly after birth in the male, consistent with a role in the maintenance of methylation patterns in proliferating spermatogonia (La Salle et al., 2004).

While the characterization of epigenetic changes in germ cells is valuable, their role in somatic cell precursors is perhaps even more important, given the critical role of somatic cells in the determination of sexual fate. Genes that regulate sex determination and differentiation are mostly expressed in these cells and must be subjected to a complex mechanism of regulation. In normal ontogenetic development, Sry, the sex-determining gene in therian mammals, initiates testis differentiation by directing the development of supporting cell precursors as Sertoli rather than granulosa and its expression is restricted to a subset of gonadal somatic cells at 10.5–12.5 days postcoitum (dpc) in the mouse. Thus, both the elaborate spatial and temporal regulation of Sry are critical for its ability to pattern the gonad (DiNapoli and Capel, 2008; Hiramatsu et al., 2009). In 8.5-dpc embryos, when the Sry gene is not yet expressed, sodium bisulfite sequencing revealed that its promoter region was hypermethylated. However, at 11.5 dpc this region was specifically hypomethylated in the gonad, while it remained hypermethylated in tissues where Sry was not expressed, suggesting that Sry is controlled epigenetically by a mechanism involving DNA methylation (Nishino et al., 2004). Hypomethylation of gene promoters is typically associated with genes becoming transcriptionally active and hence the observed changes in the Sry are somewhat to be expected. Nevertheless, the study of changes in DNA methylation in the promoter region of the other two dozen or so of genes currently related to sex determination and differentiation deserves further study.

HISTONE MODIFICATIONS AND SEX DETERMINATION

Nucleosomes represent the functional unit of chromatin and consist of a histone octamer (two of each H2A, H2B, H3, and H4) (Luger and Richmond, 1998). Nucleosomes exert regulatory effects on transcription and thus on gene through posttranslational covalent modifications of amino acid residues (Jenuwein and Allis, 2001). Trimethylation of lysine 4 (H3K4me3) constitutes an important H3 modification change associated with transcriptional activity, whereas methylation of lysine 9 (H3K9me) has the opposite effect. There is evidence that H3K9me3 can also attract and activate DNMTs (Turner, 2009). Histone acetylation also constitutes one of the most important histone modifications and takes place thanks to enzymes called histone acetyl transferases (HATs). The histone deacetylases (HDAC), on the other hand, reverse the epigenetic marks imposed by the HATs. Many of the enzymes involved in chromatin remodeling are sensitive to changes in environmental and metabolic cues and therefore work as sensors through which environmental agents can alter gene expression (Turner, 2009).

Coccoids (scale insects), including mealybugs, are sexually dimorphic plant parasites that exhibit a wide variety of chromosomal systems. In many species, paternal chromosomes undergo facultative heterochromatization in male embryos and are later eliminated from the male germline during spermatogenesis such that males exclusively transmit an identical set of maternal chromosomes through sperm. This unusual cytological system is named the lecanoid system (Brown and Nelson-Rees, 1961). In such species, the sex of the offspring is determined by whether or not paternal chromosomes are inactivated in the cytoplasm of the egg right after fertilization (Haig, 1993). In mealybugs with a lecanoid chromosome system such as Planococcus citri, Buglia and Ferraro (2004) used antibodies to methylated lysine 9 of histone H3 (H3K9me) and heterochromatin protein 1 (HP1) to investigate the involvement of these epigenetic modifications in the phenomenon of imprinting. They found that the gametes originating from a given meiosis, although carrying the same genome, differed in the levels of both H3K9me and HP1, one of them being more heavily labeled by both antibodies, indicating a relationship between imprinting and sex determination.

The example above illustrates how a specific epigenetic “mark” can be associated with a particular sex. However, how these marks are “read” is still poorly understood in the context of sex determination. There are few studies that may provide a clue. One of them concerns PHD Finger Protein 7 (PHF7). PHF7 is a protein conserved from insects to mammals that exhibits male-specific expression in the germline, from germline stem cells through spermatogonia in Drosophila. Thus, Phf7 appears to be important for germline stem cell maintenance in fruitfly males (Yang et al., 2012). Expression of Phf7 promotes spermatogenesis in XX germ cells and ectopic expression in female germ cells ablates the germline. Furthermore, human PHF7 rescues Drosophila Phf7 mutants. Interestingly, both the human and Drosophila proteins bind histone H3 N-terminal tails, with a preference for dimethyl lysine 4 (H3K4me2). Based on this evidence, Yang et al. (2012) proposed that Phf7 acts as a conserved epigenetic "reader" that activates the male germline sexual program.

As we have seen above, histone H3 lysine 9 methylation (H3K9me) is a crucial epigenetic mark of heterochromatin formation and transcriptional silencing and, as with most covalent histone lysine modifications, this modification is reversible. Thus, H3K9 methylation and demethylation could also be involved in the transcriptional regulation of genes involved in germ cells sex determination and gonadogenesis. As for H3K9 methylation, G9a is a major mammalian H3K9 methyltransferase essential for mouse embryogenesis. Tachibana et al. (2007) showed that G9a-deficient germ cells exhibited perturbation of synchronous synapsis in meiotic prophase. Mono- and di-methylation of H3K9 (H3K9me1 and 2) in G9a-deficient germ cells were significantly reduced and G9a-regulated genes were overexpressed during meiosis. Finally, they showed that H3K9me1 and 2 are dynamically and sex-differentially regulated during the meiotic prophase. This genetic and biochemical evidence strongly suggests that G9a-mediated epigenetic gene silencing is crucial for proper meiotic prophase progression and that a specific set of H3K9 methyltransferase(s) and demethylase(s) coordinately regulate gametogenesis (Tachibana et al., 2007).

In 2005, the Jumonji gene (Jmj) was identified and characterized as a critical nuclear factor for mouse embryogenesis, playing important roles in cardiovascular development, neural tube fusion, hematopoiesis, and liver development (Jung et al., 2005). Subsequently, Okada et al. (2007) demonstrated that H3K9me2/1-specific demethylase JHDM2A (JmjC-domain-containing histone demethylase 2A, also known as JMJD1A) directly binds to and controls the expression of transition nuclear protein 1 (Tnp1) and protamine 1 (Prm1) genes, the products of which are required for packaging and condensation of sperm chromatin. Furthermore, by using a loss of function approach, they demonstrated that mice deficient in JHDM2A exhibited post-meiotic chromatin condensation defects. This work, therefore, demonstrated that JHDM2A was essential for spermatogenesis (Okada et al., 2007). JHDM2A catalyzes removal of H3K9 mono- and dimethylation and also regulates metabolic genes related to energy homeostasis since JHDM2a−/− mice develop adult onset obesity and other metabolic disorders (Inagaki et al., 2009).

However, what is the direct evidence linking histone modifications and sex determination through regulation of key somatic sex-determining genes? Recently, it has been shown that XY mice deficient for the H3K9 demethylating enzyme referred to above often exhibit sex reversal and sometimes complete sex-reversal, developing as fertile females. Through RNA and protein expression analysis, Tachibana et al. (2012) found that loss of H3K9 demethylation led to a strong downregulation of Sry expression during embryogenesis. Specifically, the H3K9 demethylating enzyme accumulates at the Sry locus in wild-type XY bipotential somatic cells and leads to a significant increase in H3K9 dimethylation and decrease in H3K4 trymethylation of the Sry locus in these cells. These findings provide the first evidence for a critical role of histone methylation and demethylation on mammalian sex determination (Tachibana et al., 2012).

Although the role of Polycomb group (PcG; repressors) and trithorax group (TrxG; activators) proteins is well established in the maintenance of the appropriate gene expression profiles during development (Schuettengruber et al., 2007), less is known about their involvement in sex determination and gonadogenesis. The chromobox homolog 2 (CBX2, also known as M33) is the mouse homologue of Polycomb. Katoh-Fukui et al. (1998) showed that half of mice mutant for CBX2 died before weaning but survivors showed hypoplastic gonads and male-to-female sex reversal. The formation of genital ridges was retarded in both XX and XY homozygous mutants. Gonadal growth defects appeared near the time of Sry expression, suggesting that CBX2 deficiency caused sex reversal by interfering with steps upstream of Sry (Kato-Fukui et al.,1998). However, how CBX2 regulates gonadogenesis remained unknown. Recently, transcriptomic and immunohistochemical analyses have shown that the expression of a variety of genes encoding transcription factors essential for gonadal development, including Sry, Sox9, Lhx9, Ad4BP/SF-1, Dax-1, Gata4, Arx, and Dmrt1, was affected in CBX2 KO gonads. Crossing CBX2 KO mice with mice overexpressing Sry or Sox9 rescued male-to-female sex reversal, although testes were still hypoplastic. This indicates that CBX2 is involved in testis differentiation through regulation of Sry gene expression but that sex and size on the gonads depend on different regulatory gene networks (Katoh-Fukui et al., 2012).

NON-CODING RNAs AND SEX DETERMINATION

Non-coding RNAs (ncRNAs) are functional RNA molecules that are not translated into protein and have been implicated in some of the most studied complex epigenetic phenomena (Martiensen, 1996; Costa, 2008). ncRNAS are classified according to their nucleotide (nt) length, structure, and function (Zhou et al., 2010). The best-characterized ncRNAs in terms of epigenetic regulation are long ncRNAs (lncRNAs; >200 nt), and microRNAs (miRNAs; 19–25 nt). lncRNAs such as roX and XIST have been implicated in dosage compensation in Drosophila and Mus musculus, respectively. Dosage compensation is a phenomenon present in animals with chromosomal sex determination whereby the sex-specific gene expression inequality, due to the different number of sex chromosomes, is inhibited through epigenetic chromatin modification of one of the two sex chromosomes (Angelopoulou et al., 2008). On the other hand, miRNAs' main function is the fine-tuning of translational regulation by repression or degradation of undesired mRNAs. In general, ncRNAs regulate gene transcription by the directed recruitment of epigenetic silencing complexes to homology-containing loci in the genome (Chuang and Jones, 2007). Identifying tissue-specific ncRNAs is an essential first step toward understanding the biological functions of these molecules, which include the regulation of sexual fate determination.

The participation of ncRNAs in sex determination has been demonstrated in plants. In maize (Zea mays), male and female flowers are grouped in inflorescences called tassel and ears, respectively. Sex determination occurs through abortion of female carpels in the tassel and arrest of male stamens in the ear. The indeterminate spikelet1 (ids1) is a member of the APETALA2 floral homeotic transcription factor family that is required for spikelet meristem determinacy. Chuck et al. (2007) found that tasselseed4 (ts4) encodes miRNA 172 (miR-172), which targets APETALA2, and that ts4 mutations permit carpel development in the tassel while increasing meristem branching. These results indicate that sex determination and attainment of meristem fate share a common pathway since sexual identity in maize is acquired by limiting floral growth through negative regulation of the floral homeotic pathway (Chuck et al., 2007).

In insects, the action miRNAs on gonadogenesis has been explored in dipterans such as Drosophila (Jin and Xie, 2007) and in the mosquito Aedes aegypti (Bryant et al, 2010), the Dengue fever main transmission vector, which possess meroistic ovaries, a highly modified ovarian type in which oogonia split into both oocytes and nurse cells (Irles et al., 2009). It has also been studied in species such as Blattella germanica (Dictyoptera, Blattellidae), which possess the most primitive, panoistic type of ovaries, in which all oogonia are eventually transformed into oocytes (Irles et al., 2009). In the latter, depletion of Dicer-1 (Dcr1), a key enzyme required for miRNA formation, resulted in sterile females exhibiting deep alterations in oocyte development (Tanaka and Piulachs, 2012). In Aedes, miR-275 was identified and found to be indispensable for egg production. Given the role of miRNAs in many aspects of insect reproduction (Bellés et al., 2012), these findings suggest that new miRNAs will be identified that are related to sex determination proper.

In fish, miRNA distribution in organs of the reproductive axis (brain and gonads) has recently been reported in the Atlantic halibut, Hippoglossus hippoglossus (Bizuayehu et al., 2012), where it was found that several miRNAs have sex-biased expression. Further, the levels of some of the miRNAs were altered after masculinization by androgen or an aromatase inhibitor, suggesting that some of the miRNAs respond to hormonal signaling.

Sex determination in birds has been mostly studied in the chicken, which has a ZZ (male): ZW (female) chromosomal system. Males have twice the dosage of Z-linked genes relative to females and, in contrast to mammals, there is no large-scale dosage compensation. Only a small number of Z-linked genes are compensated, in this case by upregulation in ZW females. Recent findings suggest that gene upregulation could be mediated by a long ncRNA called male-hypermethylated (MHM), which is expressed soon after fertilization from the single Z chromosome in females. In males, MHM is transcriptionally silenced by methylation (Teranishi et al., 2001). Differences between sexes are so clear that unambiguous sexing of chickens based on this sex-specific methylation pattern is possible (Caetano and Ramos, 2008). Furthermore, the MHM locus is strongly enriched for acetylation of histone H4 at lysine residue 16, a modification linked with euchromatin in female but not male chromosomes (Bisoni et al., 2005). Interestingly, this specific histone modification is also enriched along the length of the up-regulated Drosophila melanogaster male X chromosome, where it plays a vital role in the dosage compensation process (Bisoni et al., 2005). A recent study using expression analysis and retroviral-mediated mis-expression investigated the potential role of MHM in chicken embryonic development (Roeszler et al, 2012). Whole-mount in situ hybridization confirmed that MHM is only expressed in females, notably in gonads but also in other organs. Mis-expression of MHM resulted in hypertrophy of tissues where it was expressed, notably the gonads, which lost the characteristic avian asymmetrical left-right development. Interestingly, it was found that MHM mRNA accumulates on the female Z chromosome very close to the DMRT1 locus. In chicken, DMRT1 is required for male sex determination (Smith et al., 2009). In males, MHM mis-expression impaired gonadal expression of DMRT1, suggesting that in addition to dosage compensation MHM might have a role in chicken gonadal sex differentiation (Roeszler et al., 2012). If confirmed, this would be a nice example of multi-layer epigenetic regulation, where DNA methylation, histone modification, and ncRNAs act together towards a common goal.

Sexually dimorphic expression of ncRNAs in chicken sex determination and gonadogenesis is by no means restricted to lncRNAs and can include miRNas such as miR-363, although is role in sex differentiation is still not clear (Huang et al., 2010). Bannister et al. (2009) reported the identification of miRNA 202* (miR-202*) with male-biased expression and hypothesized that it could be implicated in testicular differentiation. To address this hypothesis, they induced feminization by the injection of eggs at embryonic day 4.5 (E4.5) with E₂ and analyzed changes in miR-202* expression, which was reduced to female levels and correlated with reduced expression of the testis-associated genes DMRT1 and SOX9, and with up-regulation of the ovary-associated genes FOXL2 and CYP19A. On the other hand, female gonads treated at E3.5 with an aromatase inhibitor, which blocks estrogen synthesis, were masculinized by E9.5. In this case, miR-202* expression was increased and correlated with down-regulation of FOXL2 and CYP19A and up-regulation of DMRT1 and SOX9 (Bannister et al., 2011). Thus, up-regulation of miR-202* coincides with testicular differentiation in embryonic chicken gonads.

The situation described above for the chicken provides a good example of how different epigenetic modifications contribute to ensure proper gene expression related to sexual fate determination. However, the situation is complex taking into account the number of possible sexually dimorphic ncRNAs. In the mouse gonads, several ncRNAs have also been characterized (Ro et al. 2007; Ahn et al., 2010). Mishima et al. (2008) performed small RNA library sequencing in adult mouse testis and ovary and obtained 10,852 and 11,744 small RNA clones, respectively, which included 6,630 (159 genes) and 10,192 (154 genes) known miRNAs. Of these, 55 miRNAs were detected highly, exclusively, or predominantly in adult mouse testis and ovary, and two novel miRNAs were discovered. Male-biased expression of miRNAs occurred on the X-chromosome. However, the functional role of these miRNAs is not known. Studies in vivo and ex vivo, e.g., using organ culture on agar plates, will be required to define the spatial and temporal pattern of expression of these miRNAS and to show that their mis-expression leads to altered phenotypes or changes in gene expression of important genes such as Sry, Sox9, or Dmrt1. A promising approach is the characterization of gonad-specific miRNAs. Takada et al. (2012) examined 21 mouse adult tissues and found miR-202 to be testis- and ovary-specific. Recently, Chen et al. (2012) used microarrays to identify many new ncRNAs, including lncRNAs and miRNAS, displaying sexually dimorphic expression during early gonad differentiation and verified the expression of some of them by qRT-PCR. However, the role of these ncRNAs during sex determination and gonadogenesis is at present unknown and it will have to be characterized in the near future.

Also recently, a high-throughput RNA-sequencing project was initiated to investigate the relationship between lncRNAs and sex determination in ruminants. In the goat, Foxl2 lies inside a complex locus containing several lncRNAs. The Polled Intersex Syndrome (PIS) is a mutation that induces XX sex reversal through deletion of regulatory regions upstream of Foxl2 (Pannetier et al., 2012a). Two lncRNAs, PISTR1 and PFOXic, are well characterized. Current studies are aimed at establishing a link between these lncRNAs, chromatin conformation, and regulation of FOXL2 (Pannetier et al., 2012b).

PERSPECTIVES

In this review, I have attempted to summarize our current knowledge of epigenetic regulation of gene expression related to sex determination and gonadogenesis, providing some examples of DNA methylation, posttranscriptional histone modifications, and the emerging role of ncRNAs.

Epigenetics will contribute further to the study of the events of sex determination and gonadogenesis. In the future, we need to understand how chromatin becomes altered to activate or suppress the transcription of cognate genes involved in those events, bearing in mind that most likely important genes still remain to be discovered. We also need to know how epigenetic alterations become stabilized so that chromatin modifications are passed to daughter cells during cell division and thus gene expression patterns are conserved through gonadogenesis. This is the hallmark of true epigenetic regulation. However, in this respect it should be remembered that while histone modifications are indeed part of the epigenetic mechanisms of gene regulation, not all post-translational histone modifications are epigenetic in nature, e.g., those that play a role in more dynamic processes, such as transcriptional induction and DNA repair, are not epigenetic (Berger et al., 2009). Thus, given a particular histone modification, discerning whether it represents an epigenetic modification or not is another added challenge. Importantly, the same could be said regarding ncRNAs given that their direct actions are dynamic. This sort of distinction is a major challenge in this field.

Many important questions remain to be answered. Just to name a few that come to mind: (1) How different is the overall epigenetic regulation of sex determination in GSD systems when compared to in ESD systems? (2) What are the main epigenators and epigenetic initiators that connect environment with genotype to produce a certain phenotype? (3) What are the major epigenetic maintainers in GSD? Are they conserved through evolution? For example, within vertebrates, is there a particular epigenetic regulatory mechanism that predominates in a given animal group? (4) Also within vertebrates, is there any difference in the degree of complexity in the epigenetic mechanisms between groups such as birds and mammals, in which sex determination is more canalized, as opposed to groups such as fish, amphibian, and reptiles, where sex determination is more plastic and thus more easily influenced by environmental factors? As for challenges, since it appears that cyp19a1 regulation is essential for female differentiation in all-nonmammalian vertebrates (Guiguen et al., 2009; Lance, 2009), it seems that the study of the epigenetic regulation of this important gene should be a priority. Also, since multiple lines of evidence suggest that many elements of the sex differentiation pathway converge on the stabilization or disruption of Sox9 expression (Munger and Capel 2012), the study of the epigenetic regulation of Sox9 (e.g., Furumatsu and Asahara, 2010) should also be a top priority. Other genes such as Dmrt1, Rspo1, and so on also deserve attention. During sex determination and gonadogenesis there is a very high transcriptomic activity in the gonads. Thus, the simultaneous presence of different epigenetic modifications poses a formidable challenge to the interpretation of functional studies aimed at deciphering the influence of a particular modification in a process that per se is already complex.

The use of a combination of technical approaches will be essential to gain new insights into the role of epigenetic modifications in sex determination and gonadogenesis. To study the potential role of DNA methylation and histone modifications, one approach is the use of DNMT inhibitors such as 5′aza-deoxycytidine (5-aza) or histone deacetylase inhibitors such as trichostatin A (TSA). Thus, 5-aza and TSA have begun to be used to examine the potential role of DNA methylation and histone acetylation in a turtle with TSD, Trachemys scripta, by examining changes in Dmrt1 expression when exposed to a male-producing temperature (Bieser and Wibbels, 2012).

Progress will also be made by combining different analytical techniques, including genomic and proteomic studies. Genome-wide amplification strategies, with or without reduced representation (Meissner et al., 2008; Laird, 2010) and the use of next generation sequencing (NGS) approaches (e.g., Ahn et al., 2010), bioinformatics, and evolutionary analysis using both model and non-model organisms will allow us to interrogate the complexities of the many different types of epigenetic modifications that act together. Improvements in cell purification and chromatin immunoprecipitation (ChIP) procedures will facilitate the generation of epigenetic maps. To illustrate, in order to understand how chromatin reorganizes during gonadogenesis, Maatouk et al. (2012) performed genome-wide DNaseI hypersensitivity mapping (DNaseI-seq), an assay that allows locating and characterizing regulatory regions of the genome based on the property that they are depleted of nucleosomes. Another interesting approach is the study of Singh et al. (2011). Using a ChIP-on-chip approach, they mapped different histone modifications, including H3K9ac, H3K9me3, and H3K27me3, and CTCF (a.k.a. 11-zinc finger protein)-binding sites while DNA methylation analysis of selected CpG islands was determined using bisulfite sequencing. The combination of these analyses revealed the transcriptional potential of all protein-coding genes in the Y chromosome including the sex-determining gene SRY. This study represents the first large-scale epigenetic analysis of the human Y chromosome, linking a number of cis-elements to epigenetic regulatory mechanisms, and enabling an understanding of such mechanisms in Y chromosome–linked disorders (Singh et al., 2011).

This brings us to another interesting emerging idea: it is tempting to speculate that, in the same way that the dysregulation of some epigenetic marks has been associated with several types of cancers (Esteller, 2009; Prensner and Chinnaiyan 2011), similar dysregulations are likely to be behind several types of disorders of sexual development (DSDs). More than half of the human DSDs cannot be explained by alterations to the most well-characterized genes involved in sex determination and gonadogenesis (Munger and Capel, 2012). Thus, it is possible that alterations of the epigenetic regulatory mechanisms or epigenetic states are responsible for the genesis and/or fixation of certain DSDs. To illustrate, consider the study of Mitsuhashi et al. (2010) in which the gonadal tissue and fibroblasts were examined in a 17-year-old woman suspected of having DSD by cytogenetic, histological, and molecular analyses. The patient was found to have the 46,X,inv(Y)(p11.2q11.2) karyotype and streak gonads with abnormally prolonged SRY expression. Further, the acetylated histone H3 levels in the SRY region were significantly high relative to those of the normal male. It was concluded that since SRY is epistatic in the sex-determination pathway, the prolonged SRY expression possibly induced a destabilizing effect on the expressions of the downstream sex-determining genes, suggesting that correct regulation of SRY expression is crucial for normal male sex differentiation, even if SRY is translated normally (Mitsuhashi et al., 2010).

Last, but not least, in addition to the diagnosis of human DSDs, and given the importance of sex-related differences in epigenetic marks during early development (Bermejo-Álvarez et al., 2008), epigenetics holds a promising future in the control of reproduction in farm animal production (James and Renard, 2010). Thus, epigenetics could help us to better understand how environmental factors can affect the reproduction of these animals. In addition, there is now mounting evidence that heritable variation in ecologically relevant traits can be generated through a suite of epigenetic mechanisms, even in the absence of genetic variation (Bossdorf et al., 2008). Further, environmental impact experienced by parents may induce epigenetic responses in the offspring in wild populations (Milligan et al., 2009). Therefore, epigenetics can significantly improve our understanding of the environmental versus genetic influences on sex determination of sensitive species and the responses of organisms in a global change scenario.

Given the growing interest in this area of research, undoubtedly many new insights will emerge on the role of epigenetic modifications on sex determination and gonadogenesis. The study of such modifications offers researchers an activity of great conceptual richness and with many challenges. Their research goals can be enthusiastically pursued over the coming years with the potential for making an original contribution to this exciting field.

ACKNOWLEDGMENTS

I thank Drs. Blanche Capel and Minoru Tanaka for giving me the opportunity to write this review, which originated after presentations at the 6th International Symposium on Vertebrate Sex Determination, held in Kona, Hawaii, April 23–27, 2010. Thanks also to Silvia Joly for her help with the references. Research at the author's lab was funded by Spanish government grants “Aquagenomics” (Program Consolider-Ingenio 2012, CDS2007–0002) and “Epigen-Aqua” (AGL2010–15939), and a Government of Catalonia grant (XRAq) to F.P.