Introduction

Once a regional heterogeneity is established between a sex chromosome pair, the recombination is reduced and the Y and W chromosomes are forced to commence degeneration, even to the extent of causing eventual death of the whole chromosome and missing a master sex determining gene (Charlesworth and Charlesworth 2000; Charlesworth et al. 2005; Bergero and Charlesworth 2009; Graves 2006, 2008; Aitken and Graves 2002). The degeneration is accomplished by accumulation of deleterious mutations on the native genes, amplification of the repeated sequence and transposable element, along with structural rearrangements such as inversion, translocation and deletion (Bachtrog 2003; Steinemann and Steinemann 2005; Bergero and Charlesworth 2009). The sequential step of degeneration process from homomorphic state to highly degenerated is well seen through the phylogenetic stages in snakes and birds (Ohno 1967; Ray-Chaudhuri et al. 1971; Ray-Chaudhuri and Singh 1972; Takagi and Sasaki 1974; Nishida-Umehara et al. 2007). In eutherians, the degeneration process is independently processed since divergence from marsupials, by the addition of an autosomal part to both of the X and Y chromosomes (Waters et al. 2001). Even within eutherians, the members of degenerated genes on the Y chromosomes are different, indicating that the degeneration process of Y chromosome occurs independently in every species (Graves 2008; Waters et al. 2007). In contrast, quite a few such investigations on W chromosome have been carried out (Matsubara et al. 2006; Tsuda et al. 2007), and no investigations to compare the mechanism of degeneration between Y and W chromosomes were conducted. Generally, the latter issue could not be approached for investigation because their chromosomal origins are always different as seen in the avian W and mammalian Y chromosomes (Nanda et al. 1999).

Japanese frog Rana rugosa has two sex determining chromosome systems, existing separately, in the geographic populations (Fig. 1a) and provides us a unique opportunity to examine the difference in the process of degeneration of the Y and W chromosomes and of the W chromosomes from local populations within a species. The Northwestern Japan population, which is widely distributed along the Sea of Japan from north to south in Honshu, has a ZZ/ZW sex chromosome system, while the Central Japan population has an XX/XY sex chromosome system (Fig. 1b). These two systems share the evolutionary origin at hybridization between the two ancestral main populations (Ogata et al. 2003; Miura 2007): the Y and Z or X and W are morphologically homologous with each other, and they all share the same chromosomal origin (Fig. 1a). Between the Y and X or Z and W chromosomes, the structural differences are explained by two inversions: A chiasma at meiosis is observed at just one region of the pseudoautosomal region and almost no recombination occurs at the other part of the chromosomes (Miura et al. 1997, 2009). In this study, we produced gynogenetic diploids of WW and YY genotypes to detect degenerated, lethal genes on the sex chromosomes. In medaka fish, Oryzias latipes and common guppy, Lebistes reticulatus, the production of YY individuals is successful for detecting recessive, lethal genes on the Y chromosomes (Yamamoto 1964; Haskins et al. 1970). We found that the WW and YY embryos died during different developmental stages, and even WW embryos from two local groups died at different stages of development. We discuss the mechanisms of degeneration of the W and Y chromosomes and of sex chromosome differentiation.

Fig. 1
figure 1

The sex chromosomes of Rana rugosa and the geographic localities of specimens used. a The evolution of the XY and ZW sex chromosomes. Z and Y chromosomes, shown in blue, share the origin at chromosome 7 from the Western Japan population, while W and X chromosomes, shown in red, derived from chromosome 7 (in orange) of Eastern Japan population. Pericentromeric inversion, indicated by a circle with an arrowhead, occurred twice during the evolution of the chromosomes 7, once after emigration of an ancient population from Korea Peninsula to Western Japan, and then on the chromosome 7 from Eastern Japan population when hybridized with the Western Japan population. b The geographic localities of five ZW populations and one XY population used in this study: the cities of localities 1–6 are described in “Materials and methods”

Full size image

Materials and methods

Diploid gynogenesis, artificial sex reversal, and sexing

Specimens of ZZ males and ZW females used for experiments were collected from five populations: locality 1, Hirosaki City in Aomori Prefecture; locality 2, Niigata City of Niigata Prefecture; locality 3, Toyama City of Toyama Prefecture; locality 4, Kanazawa City of Ishikawa Prefecture; and locality 5, Fukui City of Fukui Prefecture. XY males and XX females were collected from Hamamatsu City of Aichi Prefecture (locality 6).

Gynogenetic diploids were produced according to the method of Ohtani et al. (2003). The ovulation was induced by injection of a pituitary gland extraction of the frog, Rana nigromaculata, into the abdominal cavity of females. The eggs were fertilized with UV-irradiated sperm and were subjected to a cold shock at approximately 1°C for 1 h to suppress extrusion of the second polar body. The tadpoles were fed on boiled spinach in a glass container until metamorphosis at room temperature, and frogs after metamorphosis were fed on crickets and reared at plastic container at room temperature.

To artificially produce ZW males, the genetic ZW tadpoles at 30 days after fertilization were injected with 500 μg of testosterone propionate per tadpole (Enarmon dissolved in sesame oil, Teikokuzoki Inc.) into their abdominal cavities and were reared until maturation. Genotypic sex of frogs was identified using PCR-restriction fragment length polymorphism of a sex-linked gene ADP/ATP translocase (AAT) according to the method of Ohtani et al. (2003). Genotypes of hybrids WY, WX, ZX, and ZY were identified based on RFLP profiles of both AAT and AR genes. Since W-AR is inactivated, WY could be discriminated from ZX in a crossing of ZW female and XY and from XY in a crossing of WY female and XY male (Ogata et al. 2008).

RT-PCR and Northern blot hybridization

Total RNA of gynogenetic diploids WW, ZZ, XX, and YY were extracted from the embryos (ZZ and WW of Niigata population at 10 dpf; ZZ and WW of Kanazawa population at 40 dpf; XX and YY at hatching), and those of ZZ and ZW were from adult livers of Kanazawa population using a guanidinium thiocyanate-CsCl method (Miura et al. 1998). Complementary DNA (cDNA) was synthesized using 1 μg of the total RNA as the template in 20 μl of reaction solution containing 1 μl of reverse transcriptase (SuperScript II, InVitrogen) and 25mer poly dT oligomer as the primer according to the manufacture’s instruction. In addition, 0.5 μl of the cDNA solution was amplified in 50 μl solution containing 0.2 μl of Ex Taq polymerase (TaKaRa) and 1 μl each of 12.5 μM forward and reverse primers for one cycle at 94°C for 1 min and for 35 cycles for AR and 28 cycles for elongation factor 1 α (EF1α), at 94°C for 40 s, 62°C for 40 s, and 72°C for 1 min, followed by one cycle at 72°C for 2 min. The primer sequences of AR and EF1α are the same as those described by Ohtani et al. (2003). Northern blot hybridization was performed according to the method of Miura et al. (1998), using AR fragments of 2.5 kbp as the probe.

Results

WW embryos

The Northwestern Japan population has a female heterogametic sex determining system with heteromorphic sex chromosomes (Fig. 1). We chose five populations, locations of which range from north to south, and produced gynogenetic ZZ and WW embryos. The WW embryos from three populations (localities 1–3) died of edema within 10 days after hatching in contrast to the viable ZZ embryos (Fig. 2a and Table 1). In contrast, the WW embryos from the other two populations (locality numbers 4 and 5) passed the corresponding lethal stage with no symptom, but gradually fell into a state of under-nourishment despite feeding and finally died within 50 days after fertilization (Fig. 2b and Table 1).

Fig. 2
figure 2

Gynogenetic diploids WW and YY. Gynogenetic diploids ZZ and WW produced from a ZW female of Niigata population are shown in a and those from a ZW female of Kanazawa population are shown in b. The gynogenetic diploids YY and WW produced from a WY hybrid female between ZW female of Kanazawa and XY male of Hamamatsu populations are shown in c. Bar, 5 mm

Full size image
Table 1 Survivorship of gynogenetic diploids Rana rugosa produced from ZW females of five local populations
Full size table

In order to verify whether the lethality of WW embryos is due to maternal allelic inactivation or degeneration (deleterious mutation) of the sex-linked genes responsible for development, we produced ZW sex-reversed males and crossed them with normal ZW females to obtain WW embryos, comprising both of the maternal and paternal W chromosomes. Testosterone propionate was injected into the abdomen of the tadpoles at 30 dpf, and finally 100% and 47.3% of the ZW female tadpoles from Kanazawa (K) and Niigata populations (N), respectively, were reversed to phenotypic fertile males. The sex-reversed ZW males (K and N) were crossed with normal ZW females and produced reciprocal WW embryos at a ratio of around 25% (Table 2). The WW embryos produced by crossing intra-population males and females showed the same lethality as that of the gynogenetic WW embryos. In contrast, the WW reciprocal hybrid embryos between groups N and K survived and normally matured. We conclude that the lethal genes on the W chromosomes underwent deleterious mutations, and they are not shared by the group N or K.

Table 2 Crossing of a ZW female with a sex-reversed ZW male
Full size table

Androgen receptor gene (AR) of W chromosome

AR located on the W chromosome is known to be extremely low in expression (Ohtani et al. 2003). We confirmed the AR expression in WW embryos from two different ZW populations using RT-PCR and Northern blot hybridization. The PCR product of AR was not detected in the WW embryos from Kanazawa or Niigata population, while ZZ, XX, and YY embryos showed a distinct AR product (Fig. 3a). The same results were obtained by Northern blot hybridization (Fig. 3b). Almost no signal was detected on RNA of the WW embryos from the Niigata or Kanazawa population. The hybridized signal of ZW females was weaker than that of ZZ males (Fig. 3b). Therefore, the regulatory mechanism of W-AR allele is probably inactivated both in the Niigata and Kanazawa populations of ZW type.

Fig. 3
figure 3

RT-PCR and Northern hybridization of a sex-linked androgen receptor gene (AR). Amplified product of AR is not detected in WW embryos from Niigata (N) or Kanazawa (K) population, while ZZ, YY and XX ones showed a distinct AR band (a). AR hybridization signal is detected in ZZ RNA of both populations, but not in WW (b). Signal of ZW RNA is weaker than that of ZZ

Full size image

YY embryos

In order to obtain gynogenetic YY embryos, we first produced WY hybrid by crossing a ZW female from the Kanazawa population and an XY male from Hamamatsu population. The WY embryos normally grew to female adults. Then, we produced gynogenetic YY embryos from the WY hybrid mature females. The YY embryos, about half of the clutch, died of cardiac edema at a very early developmental stage just after hatching (Fig. 2c and Table 3), while the WW embryos survived the lethal stage but finally died of under-nourishment at a much later stage, 40–50 dpf, despite feeding just like the WW embryos produced from a ZW female. Next, the WY females were crossed with normal XY males. The resultant YY embryos, about 25% of the offspring, died of the same cardiac edema at the same stage as those of gynogenetic YY embryos (Table 3). We conclude that the death of YY embryos was caused by a deleterious mutation, but not due to maternal inactivation, of one or more developmental gene(s) on the Y chromosome.

Table 3 Production of YY embryos from WY hybrid females
Full size table

Discussion

Y and W chromosome degeneration

We discovered that the Y and W chromosomes of the frog R. rugosa have already undergone deleterious mutations of the genes responsible for the development, and the lethal genes are not shared with each other. Since the W and Y chromosomes are represented by the same chromosome 7 and share their recent evolutionary origin (Miura et al. 1998; Ogata et al. 2002), it is proved for the very first time that one partner of the sex chromosomes, Y for XY and W for ZW, has undergone degeneration independently, immediately after forming a heteromorphy regardless of the heterogametic sex (Fig. 4).

Fig. 4
figure 4

Degeneration of the W and Y chromosomes. AR on the W chromosome is inactivated in both groups K and N, probably at very early stage in the primary ZW population. In contrast, lethal genes L WK and L WN, respectively, appear independently on the W chromosomes of the groups K and N. In the XY population, a different lethal gene L Y specifically appears on the Y chromosome

Full size image

In the ZW system of this frog species, five populations examined here are divided into two groups N and K based on the lethality of WW embryos, and the two types of lethal genes are not shared by the two groups. This observation indicates that the W chromosomes independently underwent degeneration after divergence of the two groups originated from a common primary population. It is also implied that the W chromosome itself involves no evolutionary prior program for degeneration.

On the contrary, the sex-linked AR gene is not the case. Its W allele seems to be degenerated in the upstream regulatory region because the expression is extremely low. This observation was first reported by Ohtani et al. (2003) and later confirmed by Yokoyama et al. (2009) and now by us in the present study. The inactivated W allele is shared by the two groups N and K. This observation suggests that the degeneration of W-AR began just after or at the origin of the ZW system and before divergence of the two groups and appearance of the lethal genes (Fig. 4). Precedence of W-AR inactivation to degeneration of the sex-linked developmental genes probably indicates its important role for ZW sex determining system, as proposed by Ohtani et al. (2003); Miura (2007) and Nakamura (2009), because the W-AR inactivation results in a sexually dimorphic expression: The AR expression level in ZW female gonads of tadpoles is almost half of ZZ ones. In fact, the treatment of ZZ male embryos with flutamide, an AR antagonist, induced upregulation of aromatase in gonads and then ovary differentiation at 30 dpf (Ohtani et al. 2003). This is a kind of “loss of function” analysis for sex determination by AR. Further critical functional analyses, such as transgenesis for overexpression of AR in ZW females and knockdown of AR in ZZ males, will be necessary to prove the sex determination role of AR.

Sex chromosome differentiation

The Z and W chromosomes of R. rugosa are similar in size and the W is still intact, with no abundant accumulation of constitutive heterochromatin at chromosomal level (Nishioka et al. 1993, 1994). Molecularly, W chromosome specific region is detected by comparative genomic hybridization around the pericentromeric region of the long arm, where a cluster of telomeric sequence of the ancestral chromosome before inversion continues to remain and transposase-like elements are also distributed (Miura et al. 2009). The molecular processing of sex chromosome differentiation and degeneration is now on going on the W chromosome, and this may have been triggered by creation of the non-recombining region with its partner Z chromosome through two inversions.

Besides, in amphibians, such a species with heteromorphic sex chromosomes is quite rare, occupying around 4% of the species examined, and most of the amphibian species still possess a homomorphic sex chromosome (Schmid et al. 1991; Eggert 2004). Basically, in highly evolved frog species, meiotic bivalents of homomorphic sex chromosomes and autosomes pair with their partners in a ring-shaped configuration, showing no chiasma except for the terminal regions of both arms (Molescalchi 1973; Okumoto 1980; Miura 1994). In fact, the recombination rate of genes is extremely lower in males than females in Rana species (Sumida and Nishioka 1994; Nishioka and Sumida 1994b). If there were almost zero recombination in males, the Y chromosomes must experience degeneration because its life is eternally isolated from X chromosomes with no exchange of chromosomal materials, but this is not the actual case. Then, Perrin (2009) proposed that the homomorphy of abundant amphibian species could be explained by the occurrence of a rare sex reversal of XY male to female followed by a recombination revival in the XY females. This theory could be a case in point. However, despite almost no actual recombination between Y-linked genes themselves in males, a recombination between the sex-linked loci (or Y chromosomal specific region) and a male-determining locus definitely occurs at a low rate, to around 10%, in Rana frog species (Wright et al. 1983; Nishioka and Sumida 1994a; Sumida and Nishioka 1994; Miura 1994). Such a recombination seems to be sufficient to release the Y-linked genes from isolation and inhibit degeneration of the Y chromosome. Therefore, we propose another candidate mechanism that inhibits evolution of amphibian sex chromosomes toward heteromorphy. It is a rare occurrence of chromosomal rearrangement, particularly an inversion, in amphibian chromosomes. In 980 anuran species whose karyotypes are examined, 83% have a diploid chromosome number of 22–26 (Kuramoto 1990). The 2n = 26 chromosomes, consisting of five large and eight small pairs, of nine Rana species examined are almost completely homologous with each other with no chromosomal mutations, and particularly, the large chromosome pairs are highly conserved over genera (Miura 1995). These observations demonstrate a very high conservation of karyotypes in amphibians. We proved in the present study that if inversion occurs in one of the sex chromosomes, the degeneration process soon sets in even in a frog regardless of the heterogametic sex. Therefore, no occurrence of inversion on the sex chromosomes (on the large sex chromosomes), particularly at the evolutionary branching point of a large taxon, seems to be a reason for the evolutionary continuing homomorphy of abundant amphibian sex chromosomes. Together, a male determining gene on the large chromosome (1–4) is frequently replaced with another one in every species and even in local populations of one species, as seen in the Rana species (reviewed by Miura 2007). This frequent turnover of sex determining gene may not allow the sex chromosomes enough time to differentiate and leave them in a homomorphic state for a long period of time.

In summary, the Y and W chromosomes of R. rugosa have already undergone deleterious mutations on the genes for embryonic development. The lethal genes are not shared by the Y or W chromosomes and even by the W chromosomes from local populations. Thus, the degeneration of the Y and W chromosomes is processed independently and involves no evolutionary, prior common scenario. For the onset of degeneration of sex chromosomes, evolution of non-recombining region by inversions may be critically important. In amphibian species, the karyotypic evolution is highly conserved, and thus, the rare occurrence of inversion on the sex determining chromosomes could be one reason why they are kept in a homomorphic state. Surprisingly, S. Ohno had already predicted the critically important role of inversion for sex chromosome differentiation, as early as 1967 (Ohno 1967).