Introduction

In teleosts, a number of species change their sex from female to male (protogynous), male to female (protandrous), or both directions (bi-directional sex change; Atz 1964; Reinboth 1970; Kuwamura and Nakashima 1998). These sex changes are dramatic events that convert both gametogenesis and behavior between female and male within a single individual. Additionally, in several species, morphological changes such as in body shape and coloration accompany the process of sex change (Shapiro 1979). Many attempts have been made to clarify the physiological mechanism of sex change, including the physiological roles of several substances such as monoamines (Larson et al. 2003), arginine vasotocin (Grober and Sunobe 1996), gonadotropin-releasing hormone (Kramer et al. 1993), gonadotropin (Koulish and Kramer 1989), and sex steroid hormones (reviewed by Devlin and Nagahama 2002). Of these substances, most attention has been focused on the role of sex steroids.

In some protogynous and protandrous species, high levels of estradiol-17β (E2), the major estrogen in vertebrates, and low levels of 11-ketotestosterone (11KT), the most active androgen in teleosts, have been detected in the circulation of fish in the female phase and vice versa for fish in the male phase. The androgen 11KT and 11β-hydroxylase, the key enzyme to synthesize 11KT, correlate with sex change in protogynous and protandrous species (Devlin and Nagahama 2002). Additionally, in many protogynous species, treatment with 11KT can result in female to male sex change (Devlin and Nagahama 2002). An exception is the rice-field eel Monopterus albus, in which 11KT administration failed to induce sex change (Tang et al. 1974). Treatment with 11KT also induces a body color change in the parrotfish Sparisoma viride (Cardwell and Liley 1991) and the wrasse Thalassoma bifasciatum (Grober et al. 1991) associated with the protogynous sex change. Therefore, gonadal androgen is generally accepted as the key hormone in the development of male characteristics, including gonad and body morphology, in many hermaphrodite species.

Aromatase is the enzyme that converts androgen to estrogen. Its enzymatic activity (Lee et al. 2001), immunolocalization (Sunobe et al. 2005), and the transcript level of the gonadal form of aromatase (cyp19a1a; Liu et al. 2004; Wong et al. 2006) have been linked to sexual dimorphism and sex change in hermaprodites. In addition, administration of E2 has resulted in male to female changes in the protandrous black porgy Acanthopagrus schlegeli (Lee et al. 2001), while blocking estrogen synthesis using an aromatase inhibitor (AI) suppresses this protandrous sex change (Lee et al. 2002). AI induces the female to male sex change in protogynous species such as the goby Coryphopterus nicholsii (Kroon and Liley 2000), the wrasse Halichoeres trimaculatus (Higa et al. 2003), and the grouper Epinephelus merra (Bhandari et al. 2004). Moreover, experimental manipulation of E2 levels by AI and E2 administration leads to the induction of sex change in each direction in the coral goby, a species in which bi-directional sex changes occur (Kroon et al. 2005). These results indicate that estrogen plays a role in female development and maintenance, and that suppression of aromatase (i.e., estrogen biosynthesis) may lead to testicular differentiation of the male phenotype. Although it remains unclear whether the steroidogenic change initially comes from androgen or estrogen biosynthesis, sex steroids appear to be a significant factor controlling gonadal sex in hermaphrodite species.

The wrasse Pseudolabrus sieboldi exhibits diandric protogyny with small initial-phase (IP) males (primary males), IP females, and large terminal-phase (TP) males (Nakazono 1979). TP males arise from one of two processes: sex change in an IP female (secondary male) or a role and color change in an IP male. In autumn, P. sieboldi spawns almost daily, and there is a diurnal pattern of oogenesis in females and spermatogenesis in secondary males (Matsuyama et al. 1997, 1998). Recently, we clarified the synthetic pathway of sex steroids in the ovarian follicles and secondary testes of this species (Ohta et al. 2001; Sundaray et al. 2003), and demonstrated that the pathways for E2 and 11KT synthesis branch from androstendione (AD). From AD, E2 is synthesized by aromatase and 17β-hydroxysteroid dehydrogenase (17β-HSD) activities in the ovarian follicle, while 11KT is synthesized in the secondary testes with 11β-hydroxylase, 17β-HSD, and 11β-HSD activities. These results clearly indicate that the steroidogenic pathway is converted from an ovarian to a testicular mode in association with gonadal sex changes and suggest that each sex steroid plays a role in the maintenance and/or change of sex.

The purpose of this study was to understand the relationship between body coloration and serum steroid levels during sex change in P. sieboldi. Nakazono (1979) previously obtained a few individuals with hermaphroditic gonads in March and June and suggested that sex change occurs during the non-spawning season. However, sampling was not conducted throughout the year, and consequently no information on sex change and body color was obtained from winter to spring just after the spawning season. Therefore, we examined adult IP and TP individuals of P. sieboldi over a range of body lengths throughout the year. First, the gonads were examined histologically, and we recorded the monthly occurrence of each gonad type, i.e., ovary, testis, or transitional (intersexual). Second, individual body color was analyzed quantitatively using a newly developed method. Finally, circulating levels of 11KT and E2 were measured, and the relationship of sex steroids with body color and gonadal histology was investigated.

Materials and methods

Fish and sampling

Every month throughout the year, adult 9–19 IP and 8–10 TP fish were sampled. From April to November 2000, the fish were caught by hook and line around the rocky shore near the Fisheries Laboratory of Kyushu University (33.8′N; 130.4′E), Fukuoka Prefecture, Japan. IP and TP fish were transferred to the Fisheries Laboratory and kept in a 1,000-l tank with filtered seawater under conditions of natural day length and water temperature. Because spawning occurs within 1 or 2 days after transfer to a tank during the spawning season (Matsuyama et al. 1998), we held fish for 5 days to acclimate before performing the experiments.

From December 2000 to March 2001, rough weather prevented us from collecting fish in the field; therefore, we collected specimens from stocks maintained at the Fisheries Laboratory during this period. Eighteen IP and ten TP fish were kept in a 1,000-l tank under the conditions described above. Five tanks were prepared, and all fish in a tank were removed each month, from December to April, for experimentation. Throughout the year, 329 adult individuals were obtained and used for histological observation, and 308 fish whose fins were not injured were used for the body color analysis. We found that the sex change in P. sieboldi took place and proceeded during the non-spawning season. For serum steroid measurements, we used 43 fish including female, secondary male derived from female, and inter-sexual phase individuals collected between February and July.

Experimental procedure

After anesthetization with 50 ppm of ethyl-4-aminobenzoate (benzocaine), each fish was placed in a transparent glass container filled with seawater. A 200-megapixel digital camera (C-920, Olympus, Japan) and four photograph lights were fixed to a camera stand (CS-5, LPL SHOJI K.K., Japan) in a dark room. The left lateral view of an individual fish were digitally photographed, together with a color separation guide (Kodak, NY, USA) that was also attached to the stand 600 mm under the camera. After taking the picture, blood samples were collected from the caudal vessel using syringes with 25-gauge needles and then centrifuged (2,000 rpm, 4°C, 20 min). The separated serum was stored at −30°C until assayed for sex steroid levels. Gonads were fixed in Bouin’s solution overnight, dehydrated, and embedded in Technovit resin (Kulzer, Wehrheim, Germany). For light microscopy, 4 μm thick sections were cut and stained with 1% toluidine blue. To clarify the difference of male germ cell occurrence among the gonadal parts, cross sections of the anterior, middle, and posterior of left and right gonads (six parts in total) were examined for the occurrence of male germ cells in the transitional gonads. Although no significant differences were observed between each gonadal part with respect to the area of male germ cells/gonad, we cut a section of the right posterior part for all individuals (data not shown). The types of ovarian and testicular germ cells were classified as in previous studies (Matsuyama et al. 1997, 1998; Morita et al. 1998).

Body color analysis

Images of each fish were stored as JPEG files and analyzed using Adobe Photoshop 5.5 (Adobe Systems Inc, CA, USA). To equalize the standard colors among the images, RGB values of the black and white squares of the color separation guide in each picture were adjusted to R = 5, G = 5, B = 5 and R = 245, G = 245, B = 245 by adjusting the input levels of red, green, and blue, respectively. In this equalization process, the hue values of the red and yellow squares of the color separation guide were always 0° and 55°, respectively. Subsequently, the hue value of each body part of the fish was determined using the “eyedropper” tool in Adobe Photoshop. For the fins, the areas between the third spine and the first ray of the anal fin, between the second and third spine (anterior) and the eighth and ninth ray (posterior) of the dorsal fin, and between the rays on an extension of the lateral line of the tail fin were measured, avoiding line patterns and spots.

In P. sieboldi, the body color of IP females is mainly red and yellow, therefore, the value of the hue was sequentially expressed by establishing a range, from 0° to ±180°, with red set as 0°.

Serum steroid levels

Serum E2 levels were measured using the Estradiol EIA Kit (Cayman Chemical, MI, USA), while 11KT levels were determined using an enzyme-linked immunosorbent assay (ELISA) following the same procedure as for testosterone (Morita et al. 1997; Ohta et al. 2001). The intra- and interassay coefficients of variation were 9.6 and 2.8% for E2, and 10.7 and 7.2% for 11KT, and competition curves for serum collected from P. sieboldi were almost parallel to the standard curves (ANCOVA P > 0.05).

Statistics

Statistical analyses were carried out with the StatView software program (SAS Institute Japan, Japan). Mann–Whitney U test was performed to analyze differences in body color among body length within IP and TP individuals, between IP male and females, and between female and other sexual stages. Serum steroid levels for the different gonadal stages were compared using a one-way analysis of variance (ANOVA) followed by the Tukey–Kramer test. Pearson’s correlation coefficient followed by Fisher’s z transformation was used to analyze the relationship between hue vales and steroid levels.

Results

Monthly occurrence of different gonadal stages

Spawning was observed in the tanks in October and November. The gonadosomatic index (GSI; gonad weight/whole body weight × 100) of females increased dramatically during the spawning season and decreased rapidly thereafter (Fig. 1).

Fig. 1
figure 1

Monthly variation in the gonadosomatic index of Pseudolabrus sieboldi. Closed circles females, closed squares TP males derived from IP males and females, open triangles Stage 1 and 2, open diamonds IP males. Sample size corresponds to Table 1 

Full size image

Immediately after the spawning season, vitellogenic oocytes in the ovary regressed and disappeared, whereas pre-vitellogenic oocytes remained and were observed throughout the year. Similarly, pre-vitellogenic oocytes were present in transitional gonads at an early stage of sex change. There were no morphological differences in the pre-vitellogenic oocytes of the ovary (Fig. 2a) and the transitional gonads; therefore, the gonadal stage of sex change was evaluated based on the developmental stage of testicular tissue. Gonads in which the cysts of spermatocytes appeared sporadically but ubiquitously among pre-vitellogenic oocytes are referred to as stage 1 (Fig. 2b), whereas gonads in which spermatids and/or spermatozoa had developed are referred to as stage 2 (Fig. 2c). The secondary testes of P. sieboldi (Fig. 2d), which are found in sex-changed females, can be generally distinguished from the primary testes derived from IP males by the presence of an ovarian cavity remnant, ovarian wall, and peripheral sperm duct (Nakazono 1979; Matsuyama et al. 1997). However, a number of testes obtained during the non-spawning season could not be identified as primary or secondary testes because of the drastic regression of gonadal structure after the spawning season. Consequently, the primary and secondary testes of TP individuals (TP primary males derived from IP primary males and TP secondary males derived from females) were pooled and are presented in Tables 1 and 2 as testes of TP males.

Fig. 2
figure 2

Histology of the gonads of Pseudolabrus sieboldi at each stage during sex change. Gonads were collected during the non-spawning season. a Ovary, b stage 1, c stage 2, d secondary testis of TP secondary male derived from female. O oocyte (pre-vitellogenic), SC spermatocytes, ST spermatids, SZ spermatozoa, OC ovarian cavity or ovarian cavity remnant

Full size image
Table 1 Monthly occurrence of fish with different stages of gonads in P. siebold i
Full size table
Table 2 Total length of fish with different stage of gonad in P. sieboldi
Full size table

IP fish with ovaries and TP fish with testes were found in every month throughout the year (Table 1). Fish with stage 1 transitional gonads were first collected in December, just after the spawning season, and were collected continually until September, just before the next spawning season. Fish with stage 2 gonads were first collected in February and throughout the non-spawning season. No fish with transitional gonads were observed during the spawning season. Gonads in intersexual stages (stage 1 and 2), i.e., those undergoing a sex change, were seen in 34 (10.3%) of the 329 fish examined. The total length of IP fish with ovaries ranged from 98 to 162 mm, and TP fish with testes ranged from 120 to 230 mm (Table 2). The total length of fish with intersexual gonads ranged from 114 to 189 mm. Ten IP fish with primary testes were also collected and ranged in length from 106 to 134 mm.

Body color analysis

The anal fin, dorsal fin, and tail fin were selected for body color analysis because there were clear differences in the color of these structures between fish with ovaries and those with secondary testes. The anal fin and the anterior and posterior parts of the dorsal fin showed a change in hue from yellow to red. In contrast, the hue of the tail fin changed from red to yellow. Of the body parts examined, the hue of the anal fin changed most, with the hue value gradually decreasing with progressing gonadal stage during the sex change process (Figs. 3, 4). Differences in fin color did not correlate with body size within IP and TP individuals, and between IP males and females (P > 0.05).

Fig. 3
figure 3

Hue values of the anal, tail, and dorsal fins at each stage during sex change in Pseudolabrus sieboldi. F female, S1 stage 1, S2 stage 2, TPM TP male derived from IP male and female. Numbers indicate sample size. Each value is the mean ± SEM. Asterisks represent significant difference from female (*P < 0.05, ** P < 0.01, ***P < 0.001)

Full size image
Fig. 4
figure 4

Anal fin colouration of Pseudolabrus sieboldi during the sex change process. Gonadal stages of individuals: a ovary, b stage 1, c stage 2, d secondary testis of TP secondary male derived from female

Full size image

Serum sex steroid levels

During the sex change process, serum E2 levels fluctuated between 0.54 and 1.06 ng/ml; however, these changes were not significant (Fig. 5). In contrast, serum 11KT levels remained between 0.90 and 0.98 ng/ml in fish with ovaries and stage 1 gonads, but increased significantly at stage 2 (1.51 ng/ml), and peaked in fish with secondary testes (2.29 ng/ml).

Fig. 5
figure 5

Serum steroid levels at each gonadal stage during sex change in Pseudolabrus sieboldi. O ovary, S1 stage 1, S2 stage 2, T secondary testis of TP secondary male derived from female. Numbers indicate sample size. Each value is the mean ± SEM. Significant differences between groups are indicated by asterisks (*P < 0.05, **P < 0.01)

Full size image

The relationship between the hue value of the anal fin and the serum steroid level was plotted for all fish (Fig. 6). The hue value of the anal fin decreased significantly with increasing serum 11KT levels (all four stages: r = −0.67, P < 0.0001, n = 43; stage 1 and 2: r = −0.53, P = 0.0087, n = 23). On the other hand, serum E2 levels did not show a significant change in association with the hue value of the anal fin (all four stages: r = −0.07, P < 0.6756; stage1 and 2: r = −0.06, P = 0.7730). Pearson correlation coefficients between serum steroid levels and hue value of four fins are shown in Table 3. The hue values of the tail fin and the anterior of dorsal fin also showed significant correlations with serum 11KT levels, whereas the posterior of dorsal fin significantly correlated with E2 levels during stage 1 and 2.

Fig. 6
figure 6

Relationship between anal fin hue value and serum steroid level. Open diamonds females, closed circles stage 1, open triangles stage 2, closed squares secondary males

Full size image
Table 3 Correlation between serum steroids and hue of four different fins
Full size table

Discussion

We were able to quantify body color change during sex change in P. sieboldi, and showed that the anal fin coloration of this species correlates with both the stage of sex change and serum 11KT levels. To our knowledge, this is the first direct demonstration of the correlation between the body coloration and serum androgen level in sex-changing species, and will provide a model system to analyze the physiological mechanism of sex change especially with respect to body coloration.

The spawning season of P. sieboldi spanned October and November, which is consistent with previous studies (Matsuyama et al. 1997, 1998; Morita et al. 1998). As suggested by Nakazono (1979), individuals undergoing sex change were obtained outside the 2-month spawning season. This also supports the hypothesis that sex change in many protogynous species takes place outside the reproductive season (Shapiro 1987). In this study, we must take into consideration that fish obtained in the winter season were examined after being kept in tanks for 1–4 months. Our recent experiment demonstrated that P. sieboldi changes sex according to social status (Ohta et al. 2003). Therefore, it is possible that sex change was induced by the conditions in the tank. Body colorations at each sexual stage were not distinct from fish collected during other seasons (data not shown), and the GSI values at each gonadal stage were not markedly different between the fish collected in April 2000 and those sampled in April 2001 that had been reared in tanks for 5 months beginning in November 2000. Therefore, P. sieboldi is capable of changing sex in captivity as well as under natural conditions.

This is the first study to quantitatively analyze the changes in body color that accompany sex change in a fish. The results indicate an excellent correlation between the hue value of the anal fin and the stage of gonadal sex change. Generally, body color change in fish has been divided into two categories; one is a physiological color change, which is attributed to rapid motile responses of chromatophores, and the other is a morphological color change, which results from changes in morphology and density of chromatophores (Fujii 2000; Sugimoto 2002). The coloration of anal, dorsal and tail fin in P. sieboldi did not change rapidly, even when fish were pale or in a dark environment (data not shown). Therefore, the color changes that we observed in three fins were probably a morphological change. Further study of the change in chromatophore densities will provide direct evidence for the mechanism of the alteration in three fins coloration of P. sieboldi. Meanwhile, monitoring changes in the color of the anal fin is a useful approach for identifying the sex of P. sieboldi individuals. Accordingly, we could easily distinguish the sex-change status of a particular individual based on the color of its anal fin.

Field research has previously demonstrated that TP males show their ventral side to females at the beginning of their mating behaviour, when females enter the TP male’s territory (Nakazono 1979). This behaviour and the fin colour change lead us to speculate that the colour of the anal fin is an important trait communicating the sexuality of a male to a female. It has been suggested that male-specific coloration is important in female mate choice in other protogynous wrasses (Warner and Schultz 1992; Kuwamura et al. 2000). However, these colorations may be unnecessary for starting the mating behaviours because sex changing individuals spawn with females even before the completion of their colour changes (Godwin et al. 1996; Sakai et al. 2002). Future experiments should be conducted to explore the role of fin colouration in the mating success of male P. sieboldi.

Here, we found that serum levels of 11KT increased in association with the development of testicular tissue during sex change in P. sieboldi. In our previous studies of P. sieboldi, we showed that the pathways to E2 and 11KT synthesis in ovarian follicles and testes branch from AD (Ohta et al. 2001, Sundaray et al. 2003). In ovarian follicles, E2 is synthesized from AD via estrone (E1), but not via testosterone (T). In the testes of P. sieboldi, 11KT is produced from AD via T. Therefore, T is not produced in the ovarian follicles, but is synthesized in testicular tissue, and 11KT is synthesized only after T production. Moreover, we have previously shown that 11KT induces spermatogenesis when 11KT is implanted intraperitoneally into female fish (Ohta et al. 2003). Although extra-follicular productions of T and 11KT are possible because of the presence of circulating 11KT in females (Ohta et al. 2001; this study), increased levels of 11KT probably play roles in promoting gonadal change from females to males in P. sieboldi. These effects of 11KT may be general in gonadal change of protogynous species, as seen in other species (Devlin and Nagahama 2002).

We found a significant correlation between serum 11KT levels and the hue values of the anal fin, tail fin, and anterior of dorsal fin. This result agrees with a previous study that the treatment of IP females with 11KT induces the change to TP males (Ohta et al. 2003). Therefore, 11KT probably plays a role in the transformation of IP females to TP males in P. sieboldi as in other protogynous species (Cardwell and Liley 1991; Grober et al. 1991). The body color change that occurs in the transition from IP female to TP male in protogynous hermaphroditic fish is traditionally thought to be similar to the expression of secondary sexual characteristics in other vertebrates (Brantley et al. 1993). Secondary characteristics of males are generally induced by androgen, not only in fish, but also in amphibians, reptiles, birds, and mammals (Borg 1994; Hews and Moore 1995; Emerson et al. 1997; Eens et al. 2000; Sinisi et al. 2003). In addition, the development of some male characteristics is mediated by the androgen receptor (AR), including the development of the gonopodium of mosquitofish (Ogino et al. 2004) and the male reproductive organs of mice (Yeh et al. 2002). Further studies are needed to determine whether the AR is involved in the body color change in fish. The results of such studies will also provide insight into the general mechanism behind the morphological development of secondary characteristics in male vertebrates, including body shape and coloration.

On the other hand, the hue value of the posterior of dorsal fin showed a significant correlation with E2 levels during stage 1 and 2. The role of estrogen for body color change is unknown. Further study must be performed to clarify the mechanism of color change in the posterior of dorsal fin.

During the process of gonadal sex change in P. sieboldi, serum E2 levels did not change drastically. This may be due to the fact that the serum levels of sex steroids in P. sieboldi were analyzed in the non-spawning season when their sex changes occur. In some protogynous and bi-directional species, serum E2 levels decrease dramatically at the onset of the gonadal sex change from female to male, in which fish from which circulating levels of female hormones are obtained have actively vitellogenic and maturational oocytes (wrasse Thalassoma duperrey; Nakamura et al. 1989, parrotfish Sparisoma viride; Cardwell & Liley 1991, goby Gobiodon histrio; Kroon et al. 2003). Indeed, serum E2 levels in females were 6–24 times higher in the spawning season (Ohta et al. 2001) than that in the non-spawning season that we examined. Such seasonal changes are also observed in many hermaphrodites (e.g., Yeung and Chan 1987a, b; Guiguen et al. 1993; Bhandari et al. 2004) that undergo sex changes during the non-breeding season. Moreover, recent studies have revealed that E2 is also involved in male spermatogenesis (Miura et al. 1999; Amar et al. 2001). Therefore, the dramatic decrease in E2 levels does not appear necessary for the protogynous sex change in the gonads and body coloration.

In contrast, recent studies have revealed that the sex changes in some protogynous and bi-directional sex-changing species were induced by implanting females with the non-steroidal AI fadrozole (gobies; Kroon and Liley 2000; Kroon et al. 2005, wrasse; Higa et al. 2003, grouper; Bhandari et al. 2004). These sex changes may be induced by the change in the circulation profiles of E2. Our study revealed that the profiles of serum E2 levels are completely different between females and males in the spawning season (Ohta et al. 2001; Sundaray et al. 2003). Female E2 levels fluctuated in the range of 5.03–18.52 ng/ml throughout 1 day in the spawning season, whereas a surge of E2 was observed around 1500 hours in males, reaching 36.8 from 0.2 ng/ml. Thus, the constant levels of E2 may be necessary for female maintenance. The physiological role of estrogen in males, as well as the changes in the diurnal profiles of E2 levels in individuals undergoing sex change, should be analyzed in P. sieboldi. Additionally, it is noteworthy that serum 11KT levels in female P. sieboldi were relatively low throughout the year, including both spawning and non-spawning season (Morita et al. 1997; this study). This likely plays a role in preventing testicular development and TP body coloration during the female phase in P. sieboldi.

In conclusion, we quantitatively evaluated the change in body color accompanying gonadal sex change from female to male in P. sieboldi. Based on the analysis of serum steroids combined with the results of our earlier experiments, we propose that 11KT production in P. sieboldi acts as a promoter of sex change. These features of P. sieboldi have allowed the characterization of all steroid hormones produced in the gonads of this species, as well as their synthetic pathways (Ohta et al. 2001, Ohta and Matsuyama 2002; Sundaray et al. 2003). P. sieboldi is useful as a model species in which to investigate the endocrinological and molecular mechanisms of sex change in fish.