Cloning and expression of candidate sexual development genes in the cane toad (Bufo marinus)



The development of the reproductive system in bufonids (true toads) is unique in several respects: sexual differentiation occurs later than in other anurans, and toads develop a Bidder's organ, a rudimentary ovary that can be manipulated in males to produce mature oocytes. To illuminate the genesis of this unusual reproductive system, we isolated from the cane toad (Bufo marinus) the orthologues of several known vertebrate sex-determining genes, determined their primary structure, and studied their expression by reverse transcriptase-polymerase chain reaction and in situ hybridization of tissue sections. We report here that cane toad Sox9, Dmrt1, and p450aromatase (Cyp19a1) are highly homologous to their counterparts in other vertebrates. They show profiles of expression that generally follow patterns observed in other taxa, but with some novel features. Our data suggest that these genes likely play key roles in sex determination and early gonad development in bufonids. Developmental Dynamics 238:2430–2441, 2009. © 2009 Wiley-Liss, Inc.


Sexual development in vertebrates involves a highly conserved group of interacting genes that vary in their roles between different taxa. The mammalian sex determining system is the best studied and understood among vertebrates and is known to hinge on the activity of a single male determining gene, Sry (Sex-determining Region Y gene; Gubbay et al.,1990; Sinclair et al.,1990). To date only three primary sex determining genes have been discovered: Sry in therian mammals, DmY (Dmrt1by) in two species of fish (Oryzias latipes and Oryzias curvinotus; Matsuda et al.,2002,2003; Kondo et al.,2003), and Dm-W in Xenopus laevis (Yoshimoto et al.,2008).

Sry in therian mammals (marsupials and placentals) induces differential development of the gonad, the initial point of sexual differentiation. Transgenesis experiments have shown that it is capable of inducing male development in chromosomally female (XX) mice (Koopman et al.,1991). The SRY protein functions through the up-regulation of Sox9 during its short expression window in mice (10.5 to 11.5 days post coitum [dpc]), thus establishing the male-determining cascade. Homologues of Sry are absent in all nontherian mammals and basal vertebrates (Kettlewell et al.,2000; Wallis et al.,2007), indicating that other sex-determining triggers must operate in these taxa.

DmY, located on the Y chromosome of the Japanese medaka fish (Oryzias latipes) is responsible for male development in that species. DmY is a paralogue of Dmrt1 (Doublesex-and mab-3-related transcription factor 1), a candidate male-determining gene in a variety of animals (Raymond et al.,1999), and is thought to have evolved through a genome duplication event that occurred in ancestral ray finned fishes (Amores et al.,1998; Meyer and Schartl,1999). Gain- and loss-of-function experiments have demonstrated that Dmy is both necessary and sufficient for male sex-determination in medaka (Matsuda et al.,2002,2007; Nanda et al.,2002).

DM-W has been shown through transgenesis experiments to be a female-determining gene located on the W sex chromosome of X. laevis (Yoshimoto et al.,2008). Like DmY, DM-W is thought to have evolved from a duplicate copy of Dmrt1 in this species, likely resulting from the tetraploid state of this species. X. tropicalis, a diploid member of the Xenopus genus, does not carry DM-W. Thus, genomic duplication events in teleost fish and in tetraploid species may have promoted the evolution of novel sex determining genes from “idle” paralogues.

Aside from these examples, the sex-determining trigger mechanisms in most vertebrate groups are presently unknown, but are likely to be diverse (Morrish and Sinclair,2002). For example, crocodilians and some turtles are known to have temperature-dependent sex determination (TSD). TSD implicates egg incubation temperature as the major influence on the initiation of a male or female pathway in a developing embryo, mediated by means of unknown molecular mechanisms. Snakes, birds, and amphibians, on the other hand, are thought to have genotypic sex determination (GSD), suggesting a genetic, hence, chromosomal difference between males and females. In some cases this involves a ZZ male/ZW female chromosome system, while in others an XX female/XY male system. In ZZ/ZW groups, it is not yet clear whether dosage of a Z-linked male-determining gene or a dominant, W-linked female-determining genes holds the key to sex determination (Ellegren,2001). Pogona vitticeps, a ZZ/ZW species, is considered to have a gene dosage dependent system, whereas X. laevis, another ZZ/ZW species, has been shown to have a W-linked female-determining gene (Quinn et al.,2007; Yoshimoto et al.,2008).

Despite the diversity in primary sex-determining switch mechanisms used in vertebrates, subsequent molecular steps involved in testis or ovary development appear to be more conserved. Sox9 (Sry-Related High Mobility Group Box 9) is an autosomal gene that encodes a pivotal transcription factor in the sex determination pathway (Kent et al.,1996; Morais da Silva et al.,1996). As for Sry, loss of function of Sox9 results in complete XY sex reversal in humans and mice (Wagner et al.,1994). Sox9 has been shown to act downstream of Sry and is able to replace Sry in inducing testis development in mice (Vidal et al.,2001). Unlike Sry, however, the role of Sox9 is not restricted to mammals. Sox9 is expressed during testis development in TSD species such as the red-eared slider turtle (Trachemys scripta; Spotila et al.,1998) and the American alligator (Alligator mississippiensis; Western et al.,1999a,b), in XX/XY GSD species such as the Japanese wrinkled frog (Rana rugosa; Takase et al.,2000), and in ZZ/ZW GSD species such as chicken (Kent et al.,1996), suggesting that Sox9 is an important testis development gene in all vertebrates.

Dmrt1 is conserved among all vertebrates and may represent the only gene associated with the regulation of sexual differentiation across the animal kingdom (Zarkower,2002). It is related to the sex-determining genes Mab-3 (male abnormal-3) in C. elegans and doublesex in Drosophila melanogaster (Raymond et al.,1998). Dmrt1 and its invertebrate orthologues encompass a DM-domain, a zinc finger-like DNA-binding motif that is diagnostic for this group of genes and has greater than 84% homology within vertebrates (Raymond et al.,1998; Osawa et al.,2005). Because Dmrt1 is a proposed ancestral sex determination gene, orthologues have been cloned and analyzed in a variety of nonmammalian vertebrate species such as the African clawed frog (Xenopus laevis - Osawa et al.,2005), Japanese wrinkled frog (Rana rugosa - Shibata et al.,2002), red-eared slider turtle (Trachemys scripta - Kettlewell et al.,2000), and chicken (Gallus sp. - Smith et al.,1999a). Furthermore, Dmrt1 has been mapped to the chicken Z chromosome, implicating it as the key gene in a proposed dosage-sensitive male sex-determination in this species (Nanda et al.,1999).

P450aromatase (Cyp19a1, Cytochrome p450 family19 subfamily a polypeptide 1) is encoded by the p450arom (CYP19a) gene and is responsible for the synthesis of estrogen from androgens in vertebrates. An increase in aromatase expression appears to be the primary signal of ovary differentiation in nonmammalian vertebrates such as birds, reptiles, and various amphibians (Smith et al.,1997; Kuntz et al.,2003; Sakata et al.,2006; Ramsey et al.,2007). Elbrecht and Smith (1992), initially reported successful female to male sex reversal in chicken by exposing them to aromatase inhibitors. Since then, a variety of expression studies have shown that aromatase is indeed a female specific factor that is primarily expressed in the ovaries of this species (Yoshida et al.,1996; Andrews et al.,1997). Amphibians are also strongly influenced by aromatase, and express this gene when exposed to female-inducing factors such as hormones and temperatures (Ohtani et al.,2003; Akatsuka et al.,2005). Furthermore, treatment with aromatase inhibitors has a masculinizing effect in urodeles, similar to the situation in chickens (Chardard and Dournon,1999).

The present study aimed to isolate and study candidate sex-determining genes in the cane toad (Bufo marinus) using a homology approach. The family Bufonidae encompasses over 500 recognized species to date (Frost,2009), with a natural geographic range spanning every continent except Australia and Antarctica. This distinctive group of amphibians has also evolved a unique system of sexual development, with a structure known as the Bidder's organ in the cephalic portion of each gonad (Falconi et al.,2007). The Bidder's organ contains immature oocytes and is a fundamental ovary that is capable of maturation and oocyte development. However, no true hermaphroditic individuals, populations, or species have been described to date among amphibians (Eggert,2004) and no known function has been ascribed to the Bidder's organ.

Although frogs and toads have become common as model organisms in developmental biology, there remains much controversy as to their sexual development. The primordial gonadal tissue seems to arise from cells in the mesonephros (Hayes,1998). In both sexes, this primordial tissue is composed of a cortex and a medulla separated by a basal lamina (Falconi et al.,2007). During testis development, the medulla develops further while the cortex either regresses or fuses with the medulla by means of the breakdown of the basal lamina separating the two layers (Hayes,1998; Falconi et al.,2007). During ovarian development, germ cells are concentrated in the cortex while regression of the medulla creates the hollow ovarian cavity (Ogielska and Kotusz,2004; Falconi et al.,2007); ovarian growth comes to depend mainly on the increasing number and size of diplotene oocytes (Ogielska and Kotusz,2004). At later stages, the metameric ovarian sacs (lobes) form, with the number of lobes depending on the species (Ogielska and Kotusz,2004). These characteristic lobes are generally used to differentiate an ovary from a testis, which has a cylindrical or cone shape in Bufo marinus (Fig. 1D,G). Genes regulating these processes have not been identified, but are likely to be amphibian homologues of genes implicated in other vertebrate species.

Figure 1.

Morphological differentiation of male and female gonads. A: Undifferentiated metamorph. B: A 20-mm female. C: A 30-mm female. D: A 60-mm female. E: A 20-mm male. F: A 30-mm male. G: A 60-mm male. Arrows indicate ovarian lobes. *, undifferentiated gonad; BO, Bidder's organ; mes, mesonephros; O, ovary; T, testis; FB, fat body. Scale bars = 0.5 mm in A; 2 mm in B–G.

We report here the entire coding sequence of cane toad homologues of Sox9, Dmrt1, and P450arom, arguably three of the more ubiquitous genes involved in the sexual differentiation mechanisms of vertebrates. In addition, we describe the spatiotemporal expression of these genes in the male and female gonads of developing toads. While we have established novel expression patterns in this species, our findings are consistent with conserved key roles for these genes in gonadal development in vertebrates.


Cloning and Characterization of the Full-Length cDNA of B. marinus Sox9

We first isolated a toad orthologue of Sox9, because of the importance and known conservation of this gene in a wide range of vertebrates. Degenerate primers were designed based on Sox9 sequences from B. gargarizans, X. laevis, R. rugosa, H. sapiens, and G. gallus (Table 1). Reverse transcriptase-polymerase chain reaction (RT-PCR) from testis and ovary cDNA yielded a 286-bp sequence, which was further expanded to the entire 1632-bp coding sequence using Rapid Amplification of cDNA Ends (RACE). The sequence (GenBank-FJ697174) was verified as the orthologue of vertebrate Sox9 using BLAST analysis ( The initiation codon was 95-bp downstream from the 5′-end and delineated an open reading frame (ORF) of 1446 bp followed by a TAG stop codon and an 88-bp 3′ untranslated region (UTR). The ORF encoded a 482 amino acid protein. The HMG DNA-binding domain started at amino acid 103 and ended at amino acid 181 (Supp. Fig. S1, which is available online).

Table 1. PCR Primers
Cloning PCR 

The cane toad SOX9 amino acid sequence was >90% homologous to that of other amphibians and to chicken, alligator, mouse, and medaka Sox9 (86%, 87%, 84%, and 71% identity, respectively). Additionally, the HMG domain of cane toad showed ≥ 97% homology compared with that of other amphibians, reptiles, mammals, and fish. It contains an amphibian-specific valine substitution in position 163 and a serine in position 149, which the two toad species share with chicken, alligator, and mouse, but not with the two frog species (Table 2; Supp. Fig. S1).

Table 2. Amino Acid Homology
SpeciesSox9 (%)Dmrt1 (%)P450 (%)
B. gargarizans98
R. rugosa926784

Spatial and Temporal Expression of Sox9

Having isolated the B. marinus orthologue of Sox9, we next set out to determine its profile of expression in gonadal vs. nongonadal tissues. RT-PCR analysis revealed that Sox9 mRNA is expressed at various levels in all 30-mm specimen tissues except muscle (Fig. 2). Testes and ovaries both expressed Sox9, as did the male Bidder's organ, a tissue type unknown in other vertebrates.

Figure 2.

Expression pattern of Sox9, Dmrt1, and p450arom in various tissues of 30-mm B. marinus determined by reverse transcriptase-polymerase chain reaction (RT-PCR). 18S was used as a positive control for the RT-PCR. All tissue samples except for ovary were taken from a male specimen.

To assess the expression of Sox9 through gonadal development in this species, we analyzed specimens at the time of metamorphosis (metamorph), designated as a presexual differentiation time-point, at 20-mm stage, when differentiation is under way, at 30-mm stage, representing early postsexual differentiation, when ovarian lobes are clearly visible, and at 60-mm stage, when differentiation has stopped yet specimens are still sexually immature (Fig. 1A–G). cDNA samples generated from the gonads of various stages were used in quantitative real-time PCR (qRT-PCR) to analyze expression levels of Sox9 mRNA during development. Because we are unable to reliably sex metamorphs, qRT-PCR was performed on triplicates of pooled, presumably mixed sex samples at this early stage. Samples from 20-mm, 30-mm, and 60-mm specimens were separated by sex. qRT-PCR results showing greater than 10% variation within a triplicate run of a sample were discarded.

Sox9 was expressed at the tadpole stage (not shown), but due to small sample sizes and high variability, we could not discern a conclusive pattern. At the metamorph stage, we observed a low level of expression in the gonads (Fig. 3A). However, at 20 mm, expression of Sox9 was greatly up-regulated in the testis, approximately 30-fold higher than in the ovary. By the 30-mm stage, expression in the developing testis had declined but was still higher than in the ovary. Of interest, in the 60-mm specimens, this sex-specific profile of Sox9 expression appeared to reverse, with the testis expressing at basal levels, while the ovary expressed Sox9 robustly (Fig. 3A).

Figure 3.

Expression of putative sexual development genes in the gonads of B. marinus at various developmental stages. A: Sox9. B: Dmrt1. C: p450arom. Each developmental time point is represented by three pools of ≥ 20 wild caught individuals and each pool was tested in triplicate. The gray line represents ovaries, and the black line represents testes. Each time point represents the mean of the normalized gene expression ratio of each sample. Error bars indicate variation (standard error) between sample pools. Metamorph samples were combined male and female gonads.

These results were confirmed in 30-mm specimens by in situ hybridization to tissue sections using a B. marinus Sox9-specific RNA probe. We observed cytoplasmic expression of Sox9 in previtellogenic oocytes, approximately equivalent to stages I and II as described by Rasar and Hammes (2006) for X. laevis (Fig. 4A–C). Sox9 expression in males was restricted to the testis cords in the developing Sertoli cells of the 30-mm specimens (Fig. 4J–L).

Figure 4.

AR: In situ hybridization analysis of expression of putative sexual development genes in the gonads of B. marinus. Localization of Sox9 (A,B,J,K), Dmrt1 (D,E,M,N), and p450arom (G,H,P,Q) transcripts in gonads from 30-mm specimens. Ovaries (top panel: A–I) and testes (bottom panel: J–R) were sectioned and hybridized with probes specific to each gene. C,F,I,L,O,R: Negative controls using sense probes generated from each gene. OC, ovarian cavity. Scale bars = 1 mm in A,D, 100 μm in B,C,E,F, 100 μm in H,I, 0.5 mm in G, 50 μm in J,L,M,O,P, R, 5 μm in K, 10 μm in N,Q.

Cloning and Characterization of the Full-Length cDNA of B. marinus Dmrt1

Using a similar strategy, a 1203-bp sequence was isolated from B. marinus containing the complete ORF of Dmrt1 cDNA (GenBank-FJ697175) The initiation codon was 98-bp downstream from the 5′-end with an ORF of 1,047 nucleotides to a TGA stop codon followed by a 56-bp 3′ UTR. The 349 amino acid cane toad DMRT1 protein was moderately homologous to both other amphibians and all other vertebrates compared (Table 2; Supp. Fig. S2). The DM domain starts with amino acid 22 and ends with amino acid 87. While the entire coding region was only moderately conserved overall, the DM domain amino acid sequence had a much higher level of homology to that of Rana rugosa (90%), Xenopus tropicalis (96%), chicken (96%), mouse (96%), medaka (80%) (Table 2; Supp. Fig. S2).

Spatial and Temporal Expression of Dmrt1

RT-PCR analysis revealed that Dmrt1 transcripts were expressed only in the gonads in both males and females, including the male Bidder's organ (Fig. 2). Additionally, when the expression level of this gene was tested in the gonads across the various stages by qRT-PCR, very low expression was observed in tadpoles (not shown) and in the pooled metamorph gonads of unknown sex (Fig. 3B). At the 20-mm stage, when differentiation has commenced, the ovary and testis both showed a sharp increase in Dmrt1 expression, with the ovary at a slightly higher level than the testis. At the 30-mm stage, when the ovary and testis had become visibly dimorphic, Dmrt1 expression in both sexes decreased, with no significant difference in expression levels between males and females (Fig. 3B). In 60-mm males, expression increased to again to a comparable level as in 20-mm specimens, while ovarian expression decreased to a basal level. However, when in situ hybridization was performed on 30-mm specimens, we were unable to detect Dmrt1 transcripts in the ovary (Fig. 4D–F). Expression in the testis was confined to the testis cords, primarily in the Sertoli cells in a similar pattern to Sox9 (Fig. 4M–O).

Cloning and Characterization of the Full-Length cDNA of B. marinus p450arom

A 2149-bp sequence was isolated from B. marinus containing the complete ORF of p450arom cDNA (GenBank-FJ697173) The initiation codon was 82 bp from the 5′-end with an ORF of 1506 nucleotides to a TAG stop codon followed by a 558-bp 3′ UTR, with no polyadenylation signal detected. The putative 502 amino acid cane toad p450arom protein sequence showed relatively high homology with those of other anurans (Rana rugosa, 84%; Xenopus tropicalis, 80%) and lower homology with respect to other vertebrates: chicken, 69%; alligator, 72%; mouse, 67%; and medaka, 53% (Table 2; Supp. Fig. S3).

The many putative DNA-binding domains of p450arom showed various degrees of homology (Supp. Fig. S4). The membrane-anchor domain showed a high homology of 90% when compared with R. rugosa but only 55% when compared with X. tropicalis with the lowest homology (25%) to chicken. In the alpha helix region, conservation was higher, sharing 88% homology with other anurans (R. rugosa and X. tropicalis) and 75%, 77%, 80%, and 63% (chicken, alligator, mouse, and medaka, respectively). The Ozol's peptide domain showed a relatively high level of conservation with mouse (95%) and moderate to high homology with other vertebrates except for medaka (54%). The aromatic region was the best conserved, with 100% identity between cane toads, R. rugosa and chicken, and 91% with respect to X. tropicalis, alligator, and mouse. Medaka showed the lowest homology at 83%. The heme-binding domain showed > 70% conservation between all taxa compared (Supp. Fig. S4).

Spatial and Temporal Expression of p450arom

RT-PCR analysis revealed that p450arom transcript was detected in several tissues excluding heart, liver, and muscle (Fig. 2). qRT-PCR was then used to examine the expression level of this gene in gonads of various stages in cane toads. A basal level of expression was observed in tadpoles (not shown) and in pooled samples of unsexed metamorph gonads (Fig. 3C). At the 20-mm stage, we observed a very low level of expression, almost equivalent to the metamorph stage. The 30-mm stage, when male and female gonads are fully differentiated, a slight increase in expression was observed, but there was no appreciable difference between the developing ovary and testis. In the 60-mm gonads, the ovary showed a significant increase in expression when compared with the 20- and 30-mm stages while in the 60-mm testis, basal level expression was maintained (Fig. 3C). By in situ hybridization, we were able to detect expression of p450arom in the peripheral follicle cells away from the ovarian cavity where oocytes are known to be in earlier stages of development compared with those in the center (Ogielska and Kotusz,2004; Fig. 4G–I). In the 30-mm testis, p450arom transcripts were detected within the testis cords, however, unlike Dmrt1 and Sox9, expression was not restricted to the Sertoli cells. (Fig. 4P–R).


Sex determination and gonadal development show an unusual evolutionary plasticity among developmental mechanisms, which are otherwise highly conserved. Moreover, the sexual differentiation pathway has recently been recognized as a plausible target for invasive species control, by means of molecular mechanism disruption (Gutierrez and Teem,2006; Cotton and Wedekind,2007). We have undertaken a study of several putative sex determination and gonad development genes in cane toads, a prolific invasive species in Australia, with the purpose of illuminating the molecular basis of these processes. We determined for the first time the entire coding sequence of cane toad Sox9, Dmrt1, and P450arom. These genes show a high level of sequence conservation in this species, and their spatiotemporal expression patterns during sexual development were to some extent conserved but showed some potentially novel features not previously reported in anurans.

Although cane toad sex determination mechanisms remain unknown, we gained important insights into the structure of several putative sexual development genes as well as their likely function in B. marinus. The Sox9 protein sequence showed the highest homology to the Rana rugosa Sox9α as would be expected from the close taxonomic relationship. Although a truncated Sox9β, lacking an HMG domain, was reported from the frog by Takase et al. (2000), we did not find any such transcript in the cane toad. A similar pattern emerges when cane toad Dmrt1 is compared with other species. The highest homology was seen with its closest taxonomic relative Rana rugosa and the lowest with its most distant relative, medaka. Interestingly, within the functional (DM) domain of DMRT1 protein, homology was, surprisingly, lowest to Rana rugosa among the species compared (X. tropicalis, chicken, and mouse). Rana rugosaDmrt1 seems to have diverged much more in expression and homology than its phylogenetic position would suggest, perhaps reflecting a functional change (Shibata et al.,2002). p450arom, the third gene sequenced, although with lower overall homology, still followed the general trend of showing higher sequence homology to Rana rugosa than to other species studied. Thus, Sox9, Dmrt1, p450arom showed a reproducible pattern of sequence homology that mirrored the phylogenetic history of vertebrates.

While their gene sequences traced an almost perfect taxonomic pattern among vertebrates, the expression patterns of these genes were mostly as predicted from studies in other species, but had deviated from expected profiles in some cases. In the cane toad, we detected expression of Sox9 at the young metamorph stage before the presumed point of sexual differentiation, suggesting a continuous expression of the gene in the developing gonads of both sexes, similar to the situation in chicken, fish, and frog (Morais da Silva et al.,1996; Spotila et al.,1998; Takase et al.,2000; Ijiri et al.,2008). Thus, Sox9 expression generally conformed to that seen in other vertebrates, and its expression in the gonads of cane toads supports its likely function as an ancient tetrapod testis-associated gene (Harley et al.,2003; Nakamoto et al.,2005; DiNapoli and Capel,2008).

In all therian mammals studied to date (with the exception of select Ellobius sp. and Tokudaia sp.), Sry is the master sex-determining gene with Sox9 implicated as a downstream target (Soullier et al.,1998; Baumstark et al.,2001; Sekido and Lovell-Badge,2008). Sox9 mutations have been implicated in sex reversal in mammals both male to female and vice versa, suggesting an ability to function in sex determination of this group without Sry (DiNapoli and Capel,2008). On the other hand, Sox9 has been excluded as the candidate sex-determining gene in the platypus (Ornithorhynchus anatinus) and in mole voles (Ellobius lutescens), a placental mammal species, neither of which use Sry as the sex-determining switch (Baumstark et al.,2001; Rens et al.,2004; Wallis et al.,2007). Furthermore, studies on other vertebrates such as various fish species (Nakamoto et al.,2005; Ijiri et al.,2008), chicken (Morais da Silva et al.,1996; Smith et al.,1999b), and alligator (Western et al.,1999b) have shown that Sox9 is up-regulated well after the initiation of the male pathway in the developing testis, relegating its role to a participant in testis development, rather than sex determination. Of interest, Shoemaker et al. (2007) have recently shown that in the turtle (T. scripta), a TSD species, Sox9 expression increases early at male producing temperatures. Furthermore, this increase occurs at a developmental stage where the sexual fate of the embryo is still reversible. Our present data indicate a relatively early up-regulation of Sox9 in cane toad testes as well, suggesting that it may be associated with sexual differentiation in this species, similar to the mammalian system.

Curiously, we observed Sox9 expression in both testicular and ovarian tissues of developing cane toads, similar to observations in the frog R. rugosa Sox9 (Takase et al.,2000). Expression is markedly increased in the testis of 20-mm specimens, very early in differentiation. Nonetheless, the initial high testicular expression is down-regulated while expression in the ovary increases as the animals develop to the 60-mm stage. In R. rugosa, Sox9α is expressed at 2 months of age in both sexes, while ovarian expression is not observed in adults of 2 years of age (Takase et al.,2000). Some duplicate isoforms of Sox9 are also expressed in ovaries in medaka (Sox9a2) and rice field eel (Sox9a; Nakamoto et al.,2005; Liu et al.,2007). However, our in situ hybridizations revealed cytoplasmic expression of Sox9 in previtellogenic oocytes and not in ovarian somatic tissue, suggesting that in B. marinus, Sox9 transcripts in the ovary are associated with the oocytes as opposed to having a function in the somatic development of the ovary.

Substantial expression of Sox9 in the ovary is likely due to the high proportion of previtellogenic oocytes in the ovaries of 60-mm female specimens. Additionally, down-regulation in the testis was not only observed in the pooled samples, but also individual animals tested (personal observation), suggesting that Sox9 is indeed down-regulated at this stage. We are confident that we have analyzed a monotypic Sox9 transcript in our experiments, because single bands were observed after electrophoresis of RT-PCR products, and these were sequenced to verify the presence of a single transcript. Furthermore, qRT-PCR primers were designed across an 1800-bp intron and showed a single peak in electrophoretograms. Although it is conceivable that a second, female-specific isoform may be present in this species, none was identified in the 5–10 males and females that were used to clone the initial sequence.

Localization of Sox9 transcripts in the ooplasm is most likely a result of expression by the previtellogenic oocytes, similar to Sox9 in medaka and sox9b in zebrafish (Yokoi et al.,2002; Rodriguez-Mari et al.,2005). In X. tropicalis, El Jamil et al. (2008) found Sox9 protein in the cytoplasm of previtellogenic oocytes, which was transported into the nucleus post vitellogenesis. Furthermore, in support of the idea that these genes are indeed expressed in the oocytes themselves and not maternally derived, most maternally derived mRNAs in germ cells are usually genes associated with cell cycle functions and other cellular processes, as opposed to developmental genes such as Sox9 which function in later development of the embryo (Vallee et al.,2008).

We did, however, find some evidence for ovarian expression of Dmrt1 in the cane toad. Dmrt1 was expressed at a basal level during the metamorph stage, similar to Xenopus (Osawa et al.,2005) and chicken (Smith et al.,2003) but not Rana rugosa (Shibata et al.,2002; Aoyama et al.,2003). In R. rugosa, Dmrt1 is expressed after sexual differentiation, during testis development, again reiterating the distinct expression pattern of Dmrt1 in that species (Shibata et al.,2002; Aoyama et al.,2003). Furthermore, using qRT-PCR we were able to detect an initial up-regulation of Dmrt1 at a similar level in both ovaries and testes at the 20-mm stage, with the ovarian expression being slightly higher than that in testis. In contrast to the RT-PCR data, we were unable to detect Dmrt1 transcripts in the ovary at this stage using in situ hybridization. This apparent discrepancy may have been due to the contribution of closely related loci and/or potential splice variants, such that we may have detected the expression of more than one transcript using qRT-PCR. However, we consider this possibility unlikely because the qRT-PCR primers used for the analysis were designed from the cloned sequence of B. marinus Dmrt1 and when compared with the R. rugosa Dmrt1, -2, -3, and -5, only matched Dmrt1. Furthermore, while cloning the Dmrt1 gene, we did not encounter any truncated or otherwise modified transcripts of this gene to indicate the presence of an alternate copy. More likely, we detected expression of Dmrt1 transcript in the ovary using qRT-PCR due to the pooling gonads of 20 individuals, whereas in situ data were obtained from a single sample. It is possible that some of the individuals had ovaries that were developmentally mismatched to those used for the in situ hybridization experiments, or were mis-sexed and were in fact testes. A more detailed analysis using qRT-PCR on individual ovaries would likely resolve this discrepancy.

In contrast to Sox9 and Dmrt1, p450arom is known to be an active gene involved in female differentiation in most vertebrates. In cane toads, a low expression level in undifferentiated animals continued into the 20-mm and 30-mm stages when sexual differentiation is already complete, as can be observed morphologically. Various degrees of p450arom expression before sexual differentiation have also been observed in the chicken and turtle, with in-depth analyses of differential expression between presumptive testes and ovaries (Villalpando et al.,2000; Ramsey et al.,2007). However, due to our inability to sex undifferentiated individuals, we cannot assess the undifferentiated gonadal expression in such detail. Of interest, in the eel (Anguilla japonica), p450arom mRNA levels were found to be lower during the initiation of vitellogenesis and to increase during the final maturation period of the ovarian follicles, due to role of estrogens in the promoting and maintaining of oocyte growth (Ijiri et al.,2003; Nunez and Applebaum,2006). However, in our in situ hybridization experiments on B. marinus, we detected p450arom mRNA in tissues in the periphery of the ovary where the oocytes are known to be in the early germ cell stages in anurans (Ogielska and Kotusz,2004). Ovarian p450arom activity in relation to mature oocytes is primarily consigned to follicle cells surrounding the germ cells (Redshaw and Nicholls,1971). It is likely that expression of p450arom in the periphery of the gonad is specific to presumptive supportive cells involved in the promotion of maturation of early germ cells into mature oocytes. These oocytes, as they mature, migrate closer to the ovarian cavity (Ogielska and Kotusz,2004). A similar expression pattern in the periphery is observed in the ovary of T. scripta at female-producing temperatures (Ramsey et al.,2007). Investigation using in situ hybridization on specific time points in the development of the ovary would likely reveal a pattern of early peripheral expression followed by expression in mature oocytes at later time points.

Furthermore, we found that p450arom was also transcribed in testes in 30-mm cane toads, similar to the expression pattern seen in some other vertebrates such as the leopard gecko (Endo and Park,2005), the frog X. laevis (Miyashita et al.,2000), and the mouse, where it is required for spermatogenesis (Robertson et al.,1999). p450arom expression is also influenced by seasonal and hormonal fluctuations (Nunez and Applebaum,2006), likely explaining the high variation we observe in our wild-caught samples.

In addition to studying the expression of these genes in the developing gonads, we also performed a nonquantitative expression study on 30-mm cane toad tissues using RT-PCR. Because the 30-mm stage represents a time point between sex determination and sexual differentiation, it is an ideal stage for the analysis of candidate sexual development genes. Dmrt1 was the only gene whose expression was restricted to the gonadal tissue in both male and female, and the Bidder's organ. This gonad specific expression of Dmrt1 is also observed in X. laevis, R. rugosa, as well as in other taxa (Shibata et al.,2002; Yoshimoto et al.,2006; Lei et al.,2007). The conserved gonad-specific expression pattern in many vertebrate lineages reflects its ancient association with sex-determination and sexual differentiation within both invertebrates and vertebrates (Raymond et al.,1998,2000). p450arom showed a wider range of tissue expression, with transcripts detected in several tissues including the gonads and the brain, the two characteristic localizations of expression in all species studied to date (Kuntz et al.,2003; Endo and Park,2005; van Nes and Andersen,2006). Many species of teleost fish express a second isoform of p450arom in the brain. However, when we sequenced the relatively small amplicon (100 bp), we did not find any significant variation indicative of a second transcript (Ijiri et al.,2003). Nonetheless, due to the small size of this amplicon, we cannot exclude the existence of an alternative transcript in B. marinus. Sox9 was the most widely expressed gene of the three, with transcripts in eight of nine tissues examined in 30-mm individuals. The only tissue lacking expression was muscle, a profile reported previously in X. laevis (Osawa et al.,2005).

In summary, we sequenced and characterized the B. marinus orthologues of Sox9, Dmrt1, and p450arom. All three were highly homologous to their counterparts in other species and even classes, suggesting an evolutionarily conserved role in cane toads as in other groups. Of interest, although sequence homology levels were lowest when compared with fish (as would be expected due to their phyletic relationship), expression profiles showed a significant functional similarity to fish, sometimes more than to higher vertebrates (Kuntz et al.,2003). This pattern may represent transcriptional modifications of sex genes correlating with the boundary between amniotes and nonamniotes, and we are currently working on a more in-depth comparative study to answer this question. Nonetheless, Sox9 shows a sexually dimorphic expression during the key period of sexual differentiation in this species, as it does in mammals. Although Dmrt1 and P450arom show no obvious dimorphic expression during sexual differentiation, they are expressed in both testes and ovaries, suggesting they may play roles in these tissues that are yet to be elucidated.


Animal Collection and Tissue Harvesting

Newly metamorphosed (metamorph), 20-mm (± 2 mm), 30-mm (± 2 mm), and 60-mm (± 5 mm) B. marinus were collected from the grounds of The University of Queensland, St. Lucia campus. All metamorphosed animals were collected and measured snout to urostyle length as described by Reading (1991). Gonadal tissues were dissected and either used immediately or stored in RNA-later (Invitrogen, Carlsbad, CA) at −80°C until later use.

RNA Extractions and cDNA Synthesis

Total RNA was isolated from four to six individual male and female gonads (excluding Bidder's organ) with the Qiagen RNeasy kit according to manufacturers' directions (Qiagen, Valencia, CA). RNA was extracted from the gonad without the Bidder's organ from metamorphosed specimens. Total RNA was used as a template for RT-PCR, which was performed using the Superscript III kit and the Oligo(dT) primer set (Invitrogen).

PCR, Cloning and Sequencing, and RACE-PCR

Partially degenerate PCR primers were designed from conserved regions of Sox9: Bufo gargarizans (DQ884961), Xenopus laevis (NM 001090807), Rana rugosa (AB035888), Homo sapiens (BC 056420.1), Gallus gallus (NM 204281); p450arom: Rana rugosa (AB178482), Xenopus laevis (BC079750), Xenopus tropicalis (BC118823), Pleurodeles waltl (AY135485), Cynops pyrrhogaster (AB164064), Hynobius retardatus (AB204518); Dmrt1: Xenopus laevis (NM 001096500), Rana rugosa (AB272609), Pleurodeles waltl (DQ265802), Homo sapiens (NM 021951), Mus musculus (NM 015826) (Table 1). PCR was performed using 0.5 μl of undiluted cDNA as template (∼500 ng). PCR conditions were: 1.5× ThermoPol buffer (New England Biolabs), 0.6 mM dNTPs, 0.4 mM primers (each) and 1.25U Taq DNA polymerase (New England Biolabs) in a 25-μl reaction. After amplification, fragments of the expected size (Sox9-286 bp; Dmrt1-252 bp; P450-634 bp) were gel-extracted using QIAquick gel extraction kit (Qiagen) and cloned using the pGemT-Easy vector system (Promega, Madison, WI) according to the manufacturers' directions. Sequencing was performed at the Australian Genome Research Facility (Brisbane, Australia). The inserts were sequenced in both directions using T7 and SP6 primers. All sequences were assembled using Sequencher 4.6 (GeneCodes, Ann Arbor, MI). Amplified sequences were verified using NCBI BLAST analysis (NCBI BLAST, and aligned to homologues in other species using ClustalW Web software (Thompson et al.,1994). After partial length sequences were confirmed, specific primers were designed for Rapid Amplification of cDNA Ends (RACE) using the 5′/3′ RACE kit 2nd generation (Roche) according to manufacturers' directions.

Quantitative RT-PCR

The qRT-PCR was performed using cDNA generated from B. marinus gonad tissue (excluding Bidder's organ) as described above (“RNA extractions and cDNA synthesis” section). Total RNA from pooled samples of 20–30 individuals was used as a template for RT-PCR, which was performed using the Superscript III and random hexamers (Invitrogen). Each stage had a minimum of three distinct pooled samples of ≥ 20 individuals. Metamorph sample pools were composed of combined, undifferentiated male and female gonads. After reverse transcription, qRT-PCR was performed with SYBR Green master mix (Applied Biosystems). Specific primer sets were designed from B. marinus orthologues of the genes and spanned putative introns larger than 3 kb when compared with mouse and frog (Xenopus tropicalis). All primers were tested using standard RT-PCR and were shown by sequencing to produce a single product. Each sample was run in triplicate and a ΔCT calculation was performed according to Bookout and Mangelsdorf (2003). The ribosomal 18S gene, which showed a consistent cycle time (CT) ranging from 5 to 10 in all tissues and developmental stages, was used as a calibrator. Additionally, total RNA was isolated using the above methods from brain, heart, lung, liver kidney, muscle, testis, ovary, and the male Bidder's organ of ∼30-mm juvenile cane toads. Total RNA was used as template for RT-PCR using the Superscript III kit with oligo(dT) primer (Invitrogen). Synthesized cDNA was used as a template for PCR amplification using primers for Sox9, Dmrt1, or p450arom. The 18S primers were used as a control to check fidelity of cDNA. PCR products were subsequently cloned and sequenced as above to check for specificity and authenticity.

In Situ Hybridization

Antisense and sense RNA probes were prepared from subclones of cane toad-specific probes. Probes for Sox9 (794-1772), P450arom (254-1504), and Dmrt1 (262-1059) were designed from cloned sequences and section in situ hybridization was performed according to Wilhelm et al. (2007). Briefly, gonads were dissected from 30-mm male and female specimens for Sox9 and Dmrt1, P450arom. After dissections, the specimens were fixed with paraformaldehyde, embedded with paraffin and sectioned at 7 μm. Sections were dewaxed, rehydrated, and treated with proteinase K at room temperature. Afterward, sections were washed with phosphate buffered saline (PBS) and refixed with 4% paraformaldehyde/PBS, acetylated and prehybridized with hybridization solution (50% formamide, 5× standard saline citrate [SSC], 5× Denhardt's, 250 μg/ml yeast RNA, 500 μg/ml herring sperm DNA). Hybridization with digoxigenin-labeled probes were performed overnight at 58°C. After hybridization, sections were washed with SSC and NT buffer (150 mM NaCl, 50 mM Tris-HCl, pH 7.5) before incubating for 2–3 hr with blocking solution (10% heat-inactivated sheep serum in NT buffer) in a humidified chamber. Anti-digoxigenin antibody (Roche Applied Sciences) at 1:2,000 dilution in blocking solution was added to the slides and incubated overnight at 4 °C. Unbound antibodies were removed by washing in NT buffer and the sections were treated in NTM buffer (100 mM NaCl, 100 mM Tris-HCl, pH 9.5, 50 mM MgCl2). Subsequently, sections were incubated in color solution (0.175 mg of 5-bromo-4-chloro-3-indolyl phosphate (Roche Applied Science), 0.35 mg of nitro blue tetrazolium (Roche Applied Science) per ml of NTM buffer) until staining was satisfactory.


We thank Zara Borg for help with cloning and RACE, Dagmar Wilhelm for help with in situ hybridization experiments, Josephine Bowles for critically reading the manuscript, and Koopman laboratory members for insightful discussions on technical aspects of the research Finally, we thank the two anonymous reviewers for valuable comments on previous drafts of this manuscript. This work was funded by the Invasive Animals Cooperative Research Center, the Queensland State Government and the Australian Research Council (ARC). Peter Koopman is a Federation Fellow of the ARC.