Animals

All fish used in this study originated from stocks kept at the Faculty of Science, Niigata University, Japan. They were reared at 27°C±2°C, under a photoperiod of 14 h light and 10 h dark. From two inbred strains of medaka, Hd-rR and HNI (Hyodo-Taguchi and Sakaizumi, 1993), we established a Y-congenic strain, Hd-rR.Y^HNI. XY fish of the Hd-rR.Y^HNI strain have a genetic background derived from the Hd-rR strain, and a Y chromosome that includes the sex-determining gene from the HNI strain (Matsuda et al., 1998). Next, we established a recombinant strain, Hd-rR.Y^HNI(R1), in which recombination occurred between the X and Y chromosomes (Matsuda et al., 2002). Genotypic sex (XY or XX) of the fish used in this study was determined by establishing the presence or absence of the DMY gene by PCR on DNA extracted from a caudal fin clip (Shinomiya et al., 1999; 2004).

Histological observation

Gonads of mature fish at about six months after hatching were dissected and fixed in Bouin's solution. Specimens were embedded in paraffin, sectioned serially at 5-μm thickness, and stained with hematoxylin and eosin for observation under a light microscope.

Treatment with methyltestosterone

Fish were reared in aged tap water containing methyltestoster-one (MT; Sigma Chemical Co., St. Louis, MO) at a concentration of 10 μg/L. The MT solution was changed every three days. At water changing, the fish were anesthetized with ethylene glycol monophenyl ether (0.025%), and the number of papillary processes on the anal fin rays was counted under a dissection microscope.

Determination of P450c17 genotypes

Genotypes of the P450c17 gene, which is responsible for both androgen and estrogen synthesis, were determined by PCR analysis. The intron and exon structure of P450c17 was estimated on the basis of the cDNA sequence from ovarian follicles of the orange-red variety of medaka (DDBJ Accession No. D87121) and the genomic sequence of the Hd-rR inbred strain (Medaka Genome Project:  http://dolphin.lab.nig.ac.jp/medaka/index.php, version 1.0, scaffold 557) (Fig. 2). P450c17 PCR primers were C17-14F (5′-ACA GAA GTA CGG ACA GAC CT-3′), C17-14R (5′-AAC AGA GCG AGC AGA TGA CG-3′), and C17-24F (5′-TGT TTG GAG AGG GTT CTG-3′). The thermal cycling conditions were 5 min at 94°C for 1 cycle; 30 sec at 94°C, 30 sec at 55°C, and 60 sec at 72°C for 30 cycles; and 5 min at 72°C for 1 cycle. After the PCR reaction, PCR products were electrophoresed on 1% agarose gel.

Fig. 2.

The 5-prime end genomic region of the P450c17 gene in the Hd-rR inbred strain. Arrows show the positions of the P450c17 primers used for amplification. Boxes indicate exons; lines, introns. The nucleotide sequences of the primers are given in the text.

Phenotype of XY fish with no secondary sex characters

We found XY fish with female-like fins and no UGP in the Hd-rR.Y^HNI(R1) strain (Fig. 1E, F). When XY fish with no secondary sex characters were paired with either normal females or normal males, spawning did not occur. We performed a histological examination of the gonads from 11 mature XY fish with no secondary sex characters at six months after hatching. All 11 XY fish had testes, and spermatogenesis appeared to have been proceeding normally (Fig. 3A, B). That is, these XY fish were neither intersex, nor neuter, nor sex-reversed female, but males without secondary sex characters.

Fig. 3.

Cross-section photomicrographs of normal and scl gonads. (A) Normal testis; (B) testis of an scl XY; (C) normal ovary; (D) spermatogenesis observed in an scl XX gonad at six months after hatching; (E) an scl XX gonad at six months after hatching, showing oocytes and spermatozoa. Asterisk, efferent duct in a testis; OC, ovarian cavity. Scale bars except E, 100 μm; E, 50 μm.

Induction of papillary processes by exogenous MT

Since male secondary sex characters of the medaka have been reported to be induced by the administration of androgens (Okada and Yamashita, 1944), the lack of male secondary sex characters in the fish in this study was thought to be caused by either deficiency of androgen or low sensitivity to androgen. To examine the sensitivity of the fish to androgen, we treated XY fish with no secondary sex characters and normal Hd-rR females with MT. In addition, because the presence of the Y chromosome might affect the development of papillary processes, we also treated sex-reversed XY females of the Hd-rR.Y^HNI(R1) strain, which were induced by the treatment of embryos with estradiol-17β(Iwamatsu, 1999). We used three individuals in each group.

In all groups, the papillary processes began to appear between three and six days after the beginning of treatment, and then increased in number (Figs. 4, 5). The time-course of change in the number of papillary processes did not differ among the three groups.

Fig. 4.

Photographs of methyltestosterone (MT)-treated fish after 21 days of treatment. (A, B) An MT-treated XY female; (C, D) an MT-treated XY fish with no secondary sex characters. p.p., papillary processes.

Fig. 5.

Time-course of changes in the number of papillary processes on the anal fin rays of normal XX females, sex-reversed XY females, and XY fish with no secondary sex characters treated with methyltestosterone (10 μg/ml). Vertical bars indicate SD around the mean (n=3).

Genetic features of fish with no secondary sex characters

Because the XY fish with no secondary sex characters had mature spermatozoa but did not reproduce naturally, we performed artificial insemination (Iwamatsu, 1973), using a testis from an XY fish to obtain offspring from these fish. The testis was dissected, and mature ova from the Hd-rR strain were artificially inseminated in a saline solution (111.2 mM NaCl, 5.4 mM KCl, 1.1 mM CaCl₂, 0.6 mM MgSO₄, pH adjusted to 7.3 with NaHCO₃; Iwamatsu, 1973). We obtained 16 adult F₁ fish, all of which showed secondary sex characters. We then crossed the F₁ females with the F₁ males to obtain 68 F₂ fish, and obtained XY fish without male secondary sex characters and XX fish without a UGP (Table 1). The ratio of fish with no sex characters to normal fish in the F₂ was just one-third (17:51), and there was no significant difference in this ratio between XX (8:27) and XY (9:24) fish (χ² test, P<0.05), indicating that the phenotype of no secondary sex characters is caused by a single recessive mutation on an autosome. Therefore, we named the mutant scl (sex character less).

Table 1.

Number and genetic and phenotypic sex in F₁ and F₂ offspring from an XY scl mutant and Hd-rR female cross.

We performed a histological examination of gonads from five XX fish with no UGP at about six months after hatching. Spermatogenesis had occurred in the gonads of all five fish; however, oocytes at the diplotene stage were also present in two of them. In the testes of all 11 scl XY fish, cysts of type B spermatogonia, spermatocytes, and spermatids were aligned in seminiferous tubules toward the central region. In contrast, in gonads of the XX fish with no UGP, these cysts were in disorderly alignment in comparison with those in the testes of normal XY and scl-XY fish, and neither an ovarian cavity nor efferent duct was identified. That is, the structure of these XX gonads appeared to be different from those of both normal ovaries and testes (Fig. 3A, C, D, E).

Identification of a candidate gene for scl

The induction of papillary processes by MT in the scl XY mutant suggests that a lack of androgens might be responsible for the scl phenotype. In addition, the scl XX mutants had no UGP. Therefore, estrogen might also be lacking in the scl mutants. The fact that both male and female secondary sex characters were absent in the scl XX and XY mutants indicates that the scl phenotype might be induced by the deficiency of an enzyme related to an early steroid hormone-producing pathway.

Of the genes that encode for enzymes related to the early steroidogenic pathway in the medaka, only the P450c17 gene had been cloned when we began this study. P450c17 encodes an enzyme with multiple catalyzing functions at a branching point of the biosynthesis pathway of cortisol and sex steroids, including androgens and estrogens. Thus, we believed that P450c17 could be a candidate scl mutant gene. We examined the P450c17 genotypes of wild-type, scl, and F₁ fish. Using a combination of C17-24F and C17-14R primers, a 1.2-kilobase pair (kb) band was detected in all genotypes. On the other hand, with a combination of C17-14F and C17-14R primers, a 1.5-kb band was observed only in the wild type and the F₁, but not in the scl mutants (Fig. 6). In other words, the C17-24F/C17-14R primer set could amplify both the wild-type and mutant alleles, but the C17-14F/C17-14R primer set could not amplify the mutant allele. Thus, the scl mutation was thought to be located between primer sites C17-14F and C17-24F.

Fig. 6.

PCR products of P450c17 in wild-type, scl, and F₁ fish obtained using (A) primer set C17-24F/C17-14R and (B) primer set C17-14F/C17-14R. Using primer set C17-24F/C17-14R, a 1.2-kb band was observed in all genotypes. However, with primer set C17-14F/C17-14R, a 1.5-kb band was present in wild-type and F₁ fish, but not in scl mutant fish.

Next, we carried out linkage analysis between the P450c17 genotype and scl phenotype. All 30 scl mutants lacked the 1.5-kb band, demonstrating that the scl phenotype was completely linked to the mutant genotype of P450c17. Analysis of medaka genome data (Medaka Genome Project:  http://dolphin.lab.nig.ac.jp/medaka/index.php, version 1.0, scaffold 557) showed that the P450c17 gene is located in autosomal linkage group 15, consistent with the results of our genetic cross, i.e., that the scl phenotype is caused by a single recessive mutation on an auto-some.

Expression mechanism of scl phenotypes

Secondary sex characters are known to be induced by sex steroid hormones secreted from gonads in vertebrates, including the medaka. Here, in the scl mutants, neither male nor female secondary sex characters were seen. Moreover, the induction of papillary processes in the scl mutant treated with MT suggests that the processes of expression of male secondary sex characters after binding of androgen to the receptor are normal. Therefore, the ability to synthesize androgen seems defective in the scl mutant. Furthermore, since a UGP, which is a female secondary sex character, does not appear in the scl mutant, the ability to synthesize estrogen also seems defective. These results strongly suggest that the scl phenotype is caused by a lack of sex steroids—both androgens and estrogens. In fact, concentrations of testosterone and 17β-estradiol did not reach detectable levels in the scl mutants (Suzuki et al., in press). Another mutant that lacks male secondary sex characters and has testes in XY individuals is the testicular feminization mutant (Tfm) in mammals. However, Tfm exhibits complete androgen insensitivity caused by a mutation of the androgen receptor (Lyon and Hawkes, 1970). Accordingly, we suggest that scl is a novel sex differentiation mutant in vertebrates.

The sex steroidogenic pathway and associated enzymes have been clarified (reviewed by Devlin and Nagahama, 2002), and P450c17, a sex steroid synthesis-related gene, has been cloned in the medaka (DDBJ Accession No. D87121). Our results suggest that the P450c17 gene of the scl mutants is distinguishable from that of normal fish. Using the C17-14F and C17-24F primer set, a 1.5-kb band was amplified in normal and F₁ fish, but not in the scl mutants. We then confirmed that the DNA sequence of the Hd-rR inbred strain was the same as that of the scl mutant at the C17-14F primer site (data not shown). These results indicate that a very large DNA fragment is inserted in the region between the C17-14F and C17-24F primer sites. Therefore, scl may be a loss-of-function P450c17 mutant caused by the insertion.

P450c17 deficiency in humans results in impaired production of cortisol and sex steroids, leading to overproduction of mineralocorticoids, as well as hypertension, pseudohermaphroditism, and a delay in sexual maturation (Kater and Biglieri, 1994; Yanase, 1995; New, 2003). In the scl mutant, the loss-of-function P450c17 mutation appears to cause a loss of sex steroids, apparently resulting in the complete absence of both male and female secondary sex characters. In addition, XY fish with the scl mutation do not show the pseudohermaphroditism evident in humans with the P450c17 mutation.

Mutation of P450c17 appears to cause impaired production of cortisol, which is vital to the existence of vertebrates, including fish. Why, then, is the scl mutant alive in the absence of P450c17 function? Recently, Zhou et al. (2007a and 2007b) identified a novel type of P450c17 (P450c17-II) in several fishes, including the medaka. P450c17-II in the medaka and tilapia possesses only 17α-hydroxylase activity, without any 17,20 lyase activity, unlike conventional P450c17 (P450c17-I), which has both of these activities. In the head kidney, P450c17-II was strongly expressed, whereas no signal for P450c17-I expression was detected. Our data suggest that scl is a loss-of-function mutant of P450c17-I. Taken together, these results indicate that because of the presence of a normal P450c17-II, which is responsible for the production of cortisol, the scl mutant is viable despite its deficiency in the P450c17-I gene.

Role of androgens in medaka spermatogenesis

The scl XY fish had well-developed testes with many spermatozoa, similarly to normal male fish. Artificial insemination using a testis from an scl XY fish showed that the spermatozoa of the scl XY fish were normal. Androgens are believed to be necessary in mammals for expression of the normal male phenotype, including the initiation and maintenance of spermatogenesis (reviewed by Collins et al., 2003; Holdcraft and Broun, 2004). In Tfm mutants and androgen receptor knockout mice, spermatogenesis is arrested in pachytene spermatocytes (Lyon and Hawkes, 1970; reviewed by Yeh et al., 2002). A deficiency in P450c17 causes infertility in male mice, and many sperm of these mice are morphologically abnormal (Liu et al., 2005). In teleost fish, androgens are also believed to be important for spermatogenesis. When 11-ketotestosterone, an androgen of fish, was added to testicular organ culture, spermatogenesis, from the proliferation of spermatogonia to spermiogenesis, was induced in the Japanese eel, Anguilla japonica (Miura et al., 1991). The action of 11-ketotestosterone in the promotion of spermatogenesis has also been recognized in goldfish (Kobayashi et al., 1991) and Hucho perryi (Amer et al., 2001). However, our results from the scl mutant suggest that spermatogenesis in medaka can occur without androgens. Saiki et al. (1997) showed that the spermatocytes of the medaka develop into fertile sperm in cell culture without contact with somatic cells. The report is consistent with normal spermatogenesis in XY scl mutants. Therefore, the scl mutant may provide novel insight into the relationship between spermatogenesis and androgens in vertebrates.

We note that a progestin, 17α,20β-dihydroxy-4-pregnen-3-one (DHP), also plays an important role not only in final maturation but also in the initiation of meiosis in Japanese eel spermatogenesis (Miura et al., 2006). However, relationship between DHP and spermatogenesis in the scl mutants is unclear.

Spermatogenesis in the scl XX mutant

Mature scl XX fish had a testis instead of ovaries, and spermatogenesis seemed to occur normally, although oocytes at the diplotene stage were observed in the gonads of two scl XX fish. Several studies in teleost fish have suggested that estrogens are involved in ovarian differentiation and development (reviewed by Fostier et al., 1983). Suzuki et al. (2004) reported that in the medaka, aromatase inhibitor (AI) treatment suppressed ovarian cavity formation, and spermatogenesis was occasionally induced in the ovaries of AI-treated adult fish. However, AI treatment did not affect early oogenesis. These results suggest that endogenous estrogen is not essential for sex determination nor early oogenesis; we thus assume that early-stage oogenesis occurs normally in scl-XX fish, but that the ovarian cavity does not develop in the absence of estrogen. Subsequently, spermatogenesis may occur in scl XX gonads after they reach maturity and ultimately only spermatogenesis may be observed, but not oogenesis. In fact, in scl XX gonads, early oogenesis is not affected, but ectopic spermatogenesis occurs in gonads at the adult stage (from 70 days after hatching) (Suzuki et al., in press). At that time, the cysts may be aligned in a disorderly fashion, unlike those in the testes of normal males.

Estrogen receptor α and β knockout (αβERKO) mice show normal ovarian development at an early stage; however, the ovaries of adult αβERKO XX individuals exhibit transdifferentiation of follicles to structures resembling seminiferous tubules of the testes after puberty (Couse et al., 1999; Dupont et al., 2000). These reports suggest that in mammals, estrogen is required for the maintenance of ovaries, but not for early ovarian development. Since the gonadal development in αβERKO XX mice is similar to that of the scl XX medaka, it appears that estrogen is required in vertebrates for maintenance of the ovary, but not for sex determination nor early oogenesis. However, the mechanism of the sex reversal in the XX scl mutant is unknown, and it is unclear whether steroids other than androgens and estrogens (for example, DHP) are involved. Thus, the mechanism of the sex reversal needs further investigation.

We found a viable P450c17 loss-of-function mutant in medaka. Although P450c17 has been isolated in many vertebrates, including several teleost fish (for example, rainbow trout [Sakai et al., 1992], Japanese eel [Kazeto et al., 2000] and zebra fish [Wang and Ge, 2004]), a P450c17 loss-of-function mutant had not yet been found in vertebrates. Thus, scl may serve as an in-vivo animal model for analyzing the role of sex steroids in sex determination and gametogenesis.