[Lyric] Differential Responses of Brain Gonad and Muscle Steroid Levels to Changes in Social Status and Sex in a Sequential and Bidirectional Hermaphroditic Fish (Varenka Lorenzi*, Ryan L. Earley, Matthew S. Grober )

Abstract

Sex steroids can both modulate and be modulated by behavior, and their actions are mediated by complex interactions among multiple hormone sources and targets. While gonadal steroids delivered via circulation can affect behavior, changes in local brain steroid synthesis also can modulate behavior. The relative steroid load across different tissues and the association of these levels with rates of behavior have not been well studied. The bluebanded goby (Lythrypnus dalli) is a sex changing fish in which social status determines sexual phenotype. We examined changes in steroid levels in brain, gonad and body muscle at either 24 hours or 6 days after social induction of protogynous sex change, and from individuals in stable social groups not undergoing sex change. For each tissue, we measured levels of estradiol (E2), testosterone (T) and 11-ketotestosterone (KT). Females had more T than males in the gonads, and more E2 in all tissues but there was no sex difference in KT. For both sexes, E2 was higher in the gonad than in other tissues while androgens were higher in the brain. During sex change, brain T levels dropped while brain KT increased, and brain E2 levels did not change. We found a positive relationship between androgens and aggression in the most dominant females but only when the male was removed from the social group. The results demonstrate that steroid levels are responsive to changes in the social environment, and that their concentrations vary in different tissues. Also, we suggest that rapid changes in brain androgen levels might be important in inducing behavioral and/or morphological changes associated with protogynous sex change.

Introduction

Sex steroids have an important role in modulating social behavior [1], and social interactions can, in turn, influence steroid levels [2], [3], [4], [5]. For instance, territorial challenges simulated by acoustic playbacks increase 11-ketotestosterone (KT), the most potent fish androgen, in Gulf toadfish, Opsanus beta [6] and non-invasive administration of KT, in turn, modulates male calling behavior [7]. Although the vast majority of studies on steroids and aggression focus on males, female responses do not always align with those of males. Female song sparrows Melospiza melodia show decreased testosterone (T) in response to territorial intrusion [8] while female California mice Peromyscus californicus show no change in T [9]. In the cooperative breeding cichlid Neolamprologus pulcher, female and male residents and intruders have higher T and KT than controls that did not receive a territorial challenge, but only male aggression correlates with T [10]. Estrogen levels in these females were not affected by social challenge [9], [10], [8]. All of these studies examined changes in circulating steroids, but plasma levels can be different from local tissue levels [11]. While the gonads are an important source of steroids and probably a major contributor to circulating levels, there is evidence that a variety of tissues, including the brain, also are capable of steroidogenesis. Homogenates of brain tissues from goldfish Carassius auratus, toadfish, and Atlantic salmon Salmo salar parr produced large amounts of estrogen, and ovaries and testes synthesized much less estrogen than brain tissue [12], [13]. In fish, male blood KT levels are typically high, but there is evidence that this could be due to extragonadal production: after castration, implanting male sticklebacks Gasterosteus aculeatus with the KT precursor 11-ketoandrostenedione (OA) produced high plasma KT levels [14]. In male rainbow trout Oncorhynchus mykiss, highest OA to KT conversion activity was found in liver, testis and kidney, followed quite closely by intestine, brain and spleen, while OA activity in muscle was negligible [15]. Because muscle is not considered an important site of steroidogenesis, its steroid levels should be reflecting circulating levels, and so we included it in our analysis as a type of negative control.

We measured steroids in the brain because local synthesis of steroids could affect behavior through fast, non-genomic mechanisms that can act in a time frame more consistent with rapid behavioral change [16], [17], [18], [19], [20], [21]. In vivo microdialysis of male zebra finch Taeniopygia guttata brain showed local changes in steroid levels in response to social stimuli; playback of male songs caused an increase in estradiol (E2) and a decrease in T while presenting females to males caused an increase in E2 but not T [22]. These changes in brain steroid concentrations were not reflected in plasma steroid levels [22]. Some studies have found a correlation between plasma and tissue steroid levels (see below), but few have examined hormone concentrations across different tissues within an individual. For example, there is a strong correlation between blood and mucus KT, but not between blood and muscle KT in koi Cyprinus carpio [23]. In the ratfish Hydrolagus colliei, there is a significant positive correlation between plasma and muscle T and KT, but not E2 [24]. A study on the protogynous Thalassoma duperrey showed some discrepancies between gonad and plasma steroids [25]. In vitro gonad culture showed that testes of territorial dominant males produced much more KT than females and newly sex changed males, and gonads of sex changed males produced more KT than female gonads. When comparing plasma KT, however, neither types of male differed from females. Testosterone production by the gonad was higher in all males relative to females, but there was no significant sex difference in plasma T levels [25].

Given the mounting evidence that circulating steroid levels are not always concordant with tissue specific levels, we tested whether steroid levels in brain, gonad and muscle show sex or status differences in steroid concentrations and, if so, whether they are differentially responsive to changes in the social environment in the bluebanded goby Lythrypnus dalli, a bidirectional sequential hermaphrodite. Sex change in this species is controlled by social status [26], [27], and their small size allows for the expression of natural behavior and sex reversal in a controlled laboratory setting. In L. dalli, females can be as aggressive as males and form linear dominance hierarchies [27], so it is also an ideal species for examining differences in steroids among females within different status classes. Natural changes in or manipulation of sex steroids are associated with sex change (reviewed by Frisch [28]). In protandrous species, E2 levels increase with transition from male to female, and administering E2 to males induces sex change. In protogynous species, KT levels increase during sex change, and KT administration can induce gonadal sex change in females. Testosterone in many species does not vary during sex change. The vast majority of these studies on sex change are based on ‘circulating’ levels of hormone measured from whole body homogenates [29], [30], plasma [31], [32], [33], [34], [35], [36], [37] or fish holding water (Earley et al. unpublished data).

In the present experiment, we tested the hypothesis that tissue hormone concentrations are different between males and females, and that E2 drops and KT increases during sex change (Fig.1). Our null prediction is that the hormonal profiles of all tissues would show the same directional change in response to changes in the social environment. Deviations from this null prediction would suggest that different mechanisms control hormone concentrations in different tissues.Materials and Methods

Ethics Statement

The research conducted herein was approved by the Georgia State University IACUC protocol No. A06030.

Experimental Design

Lythrypnus dalli is a small benthic marine goby (20–50 mm adult standard length) in which large males typically dominate a nesting territory containing multiple females [38]. Under laboratory conditions, these fish form linear dominance hierarchies when kept in social groups similar to those found in nature [27]. The fish used in this study were collected off the coast of Catalina Island, CA, in June 2007 by SCUBA diving (permit #803036-03 to VL), and were initially housed in communal holding tanks at the Wrigley Institute of Environmental Sciences on Catalina Island. From these animals, thirty experimental social groups were established in indoor seawater tables with continuous seawater supply and exposure to the natural summer light cycle. Each social group consisted of a male (the largest animal in the tank), a large (alpha) female, a medium-sized (beta) female that was at least 3 mm smaller than alpha, and a small (gamma) female; these size differences ensure stable hierarchy positions. Behavioral observations consisted of 10 min sessions, and we recorded the following behaviors: approaches, displacements received, displacements given, attacks, and threat displays. Approach behavior is defined as a fish moving within one body length of another fish, and can be considered a form of challenge to the other fish; the fish being approached can react or not, and this often depends on its social status. If the approach ends with the subordinate fish moving away, then it is scored as a displacement (see below) and is a good measure of relative dominance. The attack is a very quick approach aimed directly towards the other fish and causing the fish that was attacked to rapidly move away. Attacks happen very fast and if the subordinate fish is too slow to escape, the attack can end with physical contact. Conversely, displacements are much less aggressive in nature and describe situations where the dominant fish just swims in the proximity of the subordinate fish, which moves away. Submission is measured as the number of times a fish moves away from another member of the social group in response to approaches, attacks, or displacements. Finally, threat displays occur when two fish approach one another raising the dorsal fin and opening the opercula while swimming sideways or in circle.
In ten groups, fish were kept in the tank for 7 days without any social manipulation, and these are referred to as “stable” groups. Behavior was observed on day 6 in the morning and in the afternoon, and on day 7 in the morning. Tissue samples were collected from all group members (male and all females) on day 7. In the other twenty social groups, the male was removed from each tank after 6 days to induce sex change in the alpha female. Behavior was observed on day 5 in the morning and in the afternoon, on day 6 in the morning plus one hour after male removal in the afternoon, and on day 7 (the day after male removal) in the morning. In ten of these groups, tissues were collected from the alpha and beta females 24 hours after male removal, and these will be referred to as “sex change 24 h” groups. In the remaining ten groups, tissues were collected from alpha and beta females 6 days after male removal. In these groups, additional behavioral observations were performed 5 days after male removal in the morning and in the afternoon, and 6 days after male removal in the morning. These will be referred to as “sex change 6d” groups. For each fish, we recorded standard length, and captured a digital image of the external genitalia (genital papilla) at the start and at the end of the experiment. The shape of the genital papilla is a reliable indicator of the sex of the fish and is expressed as length/width ratio. A ratio of 1.2 or less is typical of a round female papilla, while a ratio >1.6 is typical of a thin pointed male papilla; transitional fish have intermediate values [27]. We used AxioVision (Zeiss Inc.) imaging software to take papilla measurements from digital images. Fish were then killed via exposure to a lethal dose of tricaine methanesulfonate (MS222). Brain, gonads and caudal muscle were collected, and a digital image of the gonads was taken to confirm sex and reproductive status. Each tissue was put in a microcentrifuge vial and immediately frozen in liquid nitrogen. All the fish were killed between 1400 and 1600 h. Previous work showed that L. dalli water-borne androgen levels peak in the morning and E2 levels peak in the evening [39], so we assayed tissues when hormones do not fluctuate dramatically to avoid differences due to diurnal rhythms. The tissue samples were shipped frozen to Georgia State University and stored at −80°C until processing for steroid content.

Steroid Analysis

Steroid hormones were extracted from tissue following a solid-phase extraction protocol (modified from [40]. Briefly, each frozen tissue sample was weighed, transferred to borosilicate vials, and homogenized in 350 µl of buffer in an ice-cold water bath (0.1 M phosphate buffer for brain and gonad samples [pH 7.5], 0.1 M borate buffer [pH 7.5] for muscle samples). After homogenization, 1.5 ml methanol was added to each sample, and they were quickly vortexed and returned to the ice bath. When all samples were homogenized, they were vortexed for 60 min and then stored at 4°C overnight. We repeated the same protocol for 2 control blank samples without any tissues: one containing 0.1 M phosphate buffer and one with 0.1 M borate buffer. The next day, samples were vortexed again for 20 min at room temperature, and centrifuged at 1027×g for 10 min at 4°C. The supernatant was decanted, and 16 ml of water was added to dilute the methanol. The homogenized samples were connected through high purity tubing (Saint-Gobain Tygon formulation 2275) to Sep-Pak C18 columns (Waters) mounted onto a vacuum manifold. Before starting, the columns were primed with 2×2 ml washes of methanol followed by 2×2 ml washes of distilled water. After drawing the samples through the columns, the columns were rinsed with 2×2 ml of distilled water to purge salts, and then total hormones were eluted with 2×2 ml of methanol. The eluted methanol was collected in borosilicate vials, dried under a stream of nitrogen, and the resulting hormone residue was resuspended in 350.5 µl of a solution made of 95% EIA buffer and 5% ethanol. After vortexing for 15 min, samples were stored at 4°C overnight. Samples were vortexed for 45 min prior to assay. EIA kits (Cayman Chemicals, Inc.) for E2, KT and T were used to quantify hormone in the samples; kit protocols were strictly followed. Assay plates were read on a spectrophotometer (Opsys MR, Dynex Technologies) at a wavelength of 405 nm after 90 min of incubation. See supporting information text for details on the validation of the hormone assays.

Statistical Analysis

For stable groups, we compared hormone levels among males and alpha, beta and gamma females. We could not perform a multivariate analysis because levels of E2, T, and KT in each fish are not independent, so we performed three separate ANOVAs, one for each hormone. We tested for normality and homogeneity of variance. To achieve normality, T and E2 levels were log transformed and KT was square root transformed. When the main effects in our statistical models were significant, we performed Tukey’s HSD post hoc tests to evaluate differences among the levels. For interaction terms, the number of possible comparisons were numerous, but only a subset of them were relevant to our hypotheses. We thus generated a priori linear contrasts to evaluate differences among the levels of our interaction terms. For sex changing groups, we performed two different analyses. In the first, we performed ANOVAs comparing hormone levels of alpha (dominant) females initiating sex change after male removal at 24 hours and 6 days after male removal to alpha females of stable groups with the male present, and beta-ranking females ascending to dominant female position in sex-changing groups to beta females in stable groups with the male present. This analysis tests whether steroid profiles in sex-changing alphas deviated significantly from stable alphas and whether betas whose status was elevated to alpha differ from individuals who remain beta. In the second analysis, we performed ANOVAs to compare alphas in the sex-changing groups to males, and betas in the sex changing groups to alphas in stable groups. This analysis tests whether steroid profiles in sex-changing alphas approached the profile of stable males, and whether betas whose status was elevated exhibit steroid profiles comparable to those of established alphas. Paired t-tests were used to compare values of papilla ratio for each animal at the start versus the end of the experiment. We performed correlations between behavior and hormone levels and between hormones in different tissues. The ANOVAs and linear contrasts were performed with SAS version 8.1 (SAS Institute Inc.), while the correlation analyses were performed with JMP (SAS Institute Inc.). All data are reported as mean ± SEM.

Results

Genital Papilla

None of the fish in stable groups changed sex, based upon the shape of their external genitalia. Initial and final genital papilla length/width ratios were not significantly different for males (paired t test: t9 1.33, p 0.215; initial 3.28±0.20; final 3.50±0.15), alphas (paired t test: t9 1.29, p 0.229; initial 1.09±0.04; final 1.12±0.04), betas (paired t test: t9 0.64, p 0.540; initial 1.09±0.04; final 1.12±0.04) or gammas (paired t test: t9 0.90, p 0.394; initial 1.13±0.07; final 1.16±0.03) from stable groups.

Alpha females in sex change treatments did not show any significant changes in the papilla ratio 24 h after male removal (paired t test: t8 1.39, p 0.202; initial 1.12±0.02; final 1.16±0.03), but the papilla shifted towards the male typical shape 6d after male removal (paired t test: t9 2.51, p 0.033; initial 1.17±0.04; final 1.38±0.07). Beta females in sex change treatments did not show any significant changes in the papilla ratio 24 h (paired t test: t8 0.67, p 0.523; initial 1.08±0.02; final 1.11±0.02) or 6d after male removal (paired t test: t9 0.54, p 0.601; initial 1.16±0.03; final 1.18±0.02).

Hormones in Stable Groups

Because of the large number of statistical comparisons conducted in the following sections, we generated a full list of significant linear contrasts for stable groups in Table S1 of the supporting information. All results in this section refer to Table S1 unless otherwise noted.

Estradiol.

There were significant status, tissue, and status×tissue effects on E2 concentrations (Table 1). Overall, females had significantly higher E2 than males. Linear contrasts revealed that for all tissues, alpha, beta, and gamma females had significantly higher E2 than males. Beta and gamma females had significantly higher gonad E2 than alphas. In females, gonads had significantly higher E2 than the other tissues, and brain had significantly more than muscle (Fig. 2). In males, gonad and brain E2 did not differ but both had significantly higher E2 than muscle (Fig. 2).Figure 2. Steroid concentrations in stable groups.
Concentrations of estradiol (E2), testosterone (T) and 11-ketotestosterone (KT) in stable groups. Lines between bars represent significant differences.
Doi:10.1371/journal.pone.0051158.g002Table 1. Summary of ANOVAs comparing sex steroids across status classes and tissue types.
Doi:10.1371/journal.pone.0051158.t001
Androgens.

For both T and KT, there were significant main effects of status and tissue (Table 1). Beta and gamma females had more T than males, and gammas had more T and KT than alpha females (Tukey’s HSD, p<0.05). Brain had significantly higher T and KT than muscle and gonad (Tukey’s HSD, p<0.05; Fig. 2). Ovaries of females pooled across status classes had significantly higher T than testes (Table S1; female gonad vs. male gonad). This effect was driven by gammas and betas whose gonads had significantly more T than male gonads (Table S1 and Fig. 2).

Hormones in Sex Changing Groups: Comparing Alpha and Beta Females

We compared hormone levels of alpha females in stable groups with those of sex-changing alphas and hormone levels of beta females in stable groups with those of betas rising to the alpha female position at 24 h and 6d after male removal. The full list of significant linear contrasts and their statistics are reported in Table S2, and a summary of the three-way interaction for E2 in Table S3 of the supporting information. All results in this section refer to Table S2 unless otherwise noted.

Estradiol.

Estradiol varied significantly with time, status, and tissue in the sex changing groups (Table 1). The significant three-way interaction (time×status×tissue) was driven largely by differences in E2 among tissue types. Linear contrasts revealed significantly higher E2 in the gonad than in brain and muscle (Fig. 3), and in brain than muscle for alphas (stable and sex-changing) and betas (stable and rising to alpha position) at all time points (Table S3). Betas rising to the alpha position had significantly higher E2 (pooled tissues) at 6d than at 24 h and compared to beta females of stable groups; this significant difference was likely driven by gonad E2, which was the only tissue to vary with time (Fig. 3 and Table S3). Six days after male removal, betas rising to the alpha position had significantly higher E2 than sex-changing females, which also appeared to be driven by differences in gonad E2 (Fig. 3 and Table S3).Figure 3. Steroid concentrations in sex changing groups.

Concentrations of estradiol (E2), testosterone (T) and 11-ketotestosterone (KT) in stable and sex changing groups. The legends “24 hours” and “6 days” refer to time after male removal.
Doi:10.1371/journal.pone.0051158.g003

Androgens.

Testosterone was higher in betas (stable and rising to alpha position) than alphas (stable and sex-changing), an effect driven mainly by gonad T (Fig. 3). Sex-changing females had higher KT at 24 h and 6d after male removal than in alpha females of stable groups. For both alphas (stable and sex-changing) and betas (stable and rising to alpha position), KT was significantly higher in brain than in gonad and muscle while brain T was significantly higher than gonad in stable and 6d groups and higher than muscle at all time points (Fig. 3). Alpha and beta females of stable groups had higher brain and muscle T than alphas and betas from sex changing groups, respectively, at the 24 h and 6d time points (Fig. 3). Brain KT was significantly higher at 24 h and 6d after male removal than in stable groups, but only at 24 h in the gonad (Fig. 3). In muscle, KT was higher in stable groups than after male removal.

Hormones in Sex Changing Groups: Comparing with Males

We compared hormone levels of sex-changing females (alphas) with those of males, and hormones of betas rising to alpha position at 24 h and 6d after male removal with those of dominant alpha females in stable groups (Table 1). Males had significantly lower E2 than sex-changing females (Fig. 3). Beta females rising to alpha position had significantly higher gonad E2, and lower brain and muscle T than dominant females in stable groups (Fig. 3). Sex-changing females had significantly higher brain KT than males, but lower muscle KT and lower brain and muscle T (Fig. 3). Due to space constraints, we present the details of this result section in the text of the supporting information, the full list of significant linear contrasts in Table S4, and a summary of the three-way interaction for E2 in Table S5 of the supporting information.

Hormone-by-Hormone Correlations

We examined the relationship between tissue steroid concentrations and due to the large number of comparisons we report details in Table 2. All results in this section refer to Table 2 with some depicted as figures (Fig. 4 and Fig. S1 and Fig. S2 in supporting information).Figure 4. Correlations between different steroids.
Correlations between concentration of KT and T in the muscle (A) and in the gonad (B) across all fish, and in the muscle of alpha females in stable groups (C) and in sex changing groups (D). Correlations between concentration of KT and T (E) and between E2 and KT in the gonad of alpha females. A solid line means that the associated p value is significant (p
Doi:10.1371/journal.pone.0051158.g004Table 2. Relationships among hormones both within and across tissue types.
Doi:10.1371/journal.pone.0051158.t002
Among-tissue relationships for a given hormone.

First, we compared levels of the same hormone in different tissues to test whether changes occur across all tissues in a uniform way. When all fish were included in the analysis, there was a significant positive relationship between gonad KT and brain KT (Fig. S1A) but it was not significant when considering each class of fish separately except for gammas in stable groups. Testosterone in the brain and gonad were also significantly positively related only for gamma females in stable groups. Muscle KT and gonad KT were significantly positively related for betas rising to the alpha position at 6d and sex-changing alphas at 24 h following male removal (Fig. S1D).

Muscle and brain E2 concentrations were significantly positively related at the population level but this trend did not hold for any of the individual status classes. There was a significant positive relationship between gonad and brain E2 in betas rising to the alpha position at 24 h (Fig. S1B) and sex-changing alphas at 6d following male removal (Fig. S1C).
Within-tissue relationships among hormones.

Second, we analyzed whether there was any relationship between different hormones in the same tissue. There were highly significant positive relationships between muscle KT and muscle T (Fig. 4A) as well as gonad KT and gonad T (Fig. 4B); these relationships were very consistent across the sexes, status classes, and time points. Only sex-changing alphas at 24 h and 6d following male removal failed to show relationships between the two androgens in gonad and muscle (Fig. 4C–E). Gonad E2 and KT concentrations were significantly negatively related only for sex-changing alphas at 6d following male removal (Fig. 4F). Brain KT and T concentrations were significantly positively related only in betas rising to the alpha position at 24 h and 6d following male removal (Fig. S2A–B).
Hormone by Behavior Correlations

We examined the relationships between steroid concentrations in each tissue type and the behavior (approaches, attacks, displacements given, courtship, threat displays and displacement received) of animals in both stable groups and after male removal. All results described in this section refer to Table 3 with some depicted as figures (Fig. 5 and Fig. S3). With few exceptions, all hormone-behavior relationships occurred in sex-changing females and betas rising to the alpha position.Figure 5. Correlations between hormones and behavior.

Correlation between the total rate of approaches performed after male removal and concentration of KT in the gonad (A), and in the muscle (B); and between the rate of attacks performed the day after male removal and concentration of E2 in the brain (C). The values refer to alpha females in sex changing groups collected 24 hours and 6 days after male removal. A solid line means that the associated p value is significant (p<0.05) while dotted lines refer to non-significant correlations.
Doi:10.1371/journal.pone.0051158.g005Table 3. Summary of hormone-behavior relationships.
Doi:10.1371/journal.pone.0051158.t003

Stable groups.

Males of stable groups that were more aggressive (e.g., approaches, displacements given) had significantly less gonad E2. Aside from this, there were no significant relationships between tissue hormones and the behavior of animals in stable groups.

Sex-changing females.

There was a significant positive relationship between both muscle and gonad (but not brain) KT of sex-changing females 24 h after male removal and the total number of approaches delivered to subordinate group members in the first day (Fig. 5A–B) as well as the number of approaches delivered in the first hour after male removal. Brain E2 of sex-changing females 6d after male removal also was significantly positively related to attacks delivered on the first day following male removal (Fig. 5C).

Betas rising.

Brain T and KT of betas rising to the alpha position 24 h after male removal were significantly positively related to the number of times they were displaced by the sex-changing female following male removal (Fig. S3A–B). Brain T and KT of betas 6d after male removal instead showed a significant negative relationship with the rate of displacements received 5 days after male removal (Fig. S3C–D).

Discussion

Our results demonstrate that levels of steroid hormones are altered during sex change but that these alterations do not follow the same pattern in all tissues. In addition, the strong positive relationship between androgens within different tissues is lost during sex change.

We also found a sex difference for T and E2 but not for KT. The higher E2 levels we found in female tissue are consistent with previous work on L. dalli that measured levels of steroids from water samples [39], and with the higher aromatase activity in females relative to males [17]. While water samples did not identify any sex difference in T nor KT [39], the present study shows that females have higher T than males in the gonads. Since brain levels of T are much higher than gonad levels and are not sexually dimorphic, they might mask the gonadal sex difference when using methods that average over the whole body (e.g., plasma or water-borne samples). Also, the very low levels of KT detected in L. dalli suggest that the lack of sex difference in KT is due to low levels in males rather than unusually high levels in females. These results are different from many fish species where males have much higher KT levels than females [41], but similar to what was found in another bidirectional hermaphrodite, Gobiodon histrio, where whole body concentrations of KT were below detection limits in both males and females [30]. 11-ketotestosterone is often viewed as the key hormone for regulating male-typical morphology in teleost fishes [41] and the lack of sex differences in KT in sequential bi-directional sex changers (like L. dalli and G. histrio) is consistent with their reduced overall sexual dimorphism and with the hypothesis that these fish reduce KT production across their life history so as to maintain the capacity for adult sex change in both directions.

Different Tissues Present Different Concentrations of Steroids

During sex change, different tissues responded independently to the same social manipulation, demonstrating that tissues can be autonomous when it comes to steroid synthesis, binding, or uptake in response to changes in the social environment. We were surprised to find very few significant correlations when comparing the same hormone across different tissues within individuals. One might expect that if steroids are completely free to reach all organs through circulation, then there should be a positive correlation in steroid levels across tissues within individuals. For example, fish whose gonad is producing high levels of E2 would also have more E2 in brain and/or muscle. Instead, our results indicate that each tissue is remarkably independent with regard to steroid concentrations. This result challenges the central dogma that steroids produced in the gonad represent ‘the’ whole body regulator of steroid actions, and could be viewed as a broad application of the law of mass action, which suggests that the concentration and affinity of receptors, binding globulins, and synthetic/degradative enzymes determines the functional capacity of a signaling biochemical in a given tissue. Future studies must consider all these factors together in a tissue to fully understand the mechanism of action of steroid hormones.

The results for androgens were unexpected; both T and KT were much higher in the brain than in gonad and muscle. The most likely explanation is that the brain is an important site for androgen synthesis. There is evidence that fish brain homogenates can produce T, E2 and KT [12], [13], [15], [42]. An alternative explanation is that androgens are sequestered from circulation and accumulated in the brain by androgen receptors and/or steroid binding globulins (SBG) at a concentration higher than the gonadal source. Steroid binding globulins can transport steroid hormones from circulation into cells [43], and can be internalized by neurons in rats [44]. Also, different rates of steroid degradation associated with each specific tissue could explain the different levels of T and KT that we observed among tissue types.

The only strong hormone correlation that held true for all females in stable groups was between KT and T in both muscle and gonad. Similarly, in male Tilapia zillii, plasma KT was positively correlated with plasma T [45]. An obvious explanation is that T is a precursor of KT but for the same reason, we would expect to see the same correlation also in the brain. 11-ketotestosterone might not be directly synthesized in the brain or its rate of synthesis could be independently and locally regulated such that product and substrate are not tightly coupled. Also, KT could be binding SBG in the brain with a different affinity than T [46], [47], [48] affecting its metabolism and precluding direct relationships with its precursor.

We found no significant difference in T between gonad and muscle. Testosterone present in the muscle could be coming through circulation or could be produced by muscle tissue but there is no evidence for any 17β-HSD activity in muscle of male rainbow trout [15].

Because we found correlations between KT and its precursor T, we also expected correlations between T and another of its metabolites, E2. In the cichlid N. pulcher T correlated with both E2 and KT in plasma [10]. On the other hand, in the protandrous Amphiprion melanopus T correlated with E2 in females but not in males, and none of the correlations were present in fish undergoing sex change [49]. In L. dalli, we did not find a correlation between T and E2 in any of the tissues, which suggests that E2 could be synthesized from estrone after aromatization of androstenedione as in the protogynous wrasse Pseudolabrus sieboldi [36]. Further work is necessary to determine whether L. dalli employs this alternative pathway.

Steroid Responses During Sex Change

After removal of the male from the social groups, we observed changes in steroid levels in tissues of sex-changing females supporting a critical role of KT in the regulation of sex change. Comparing the steroid response of alpha females with beta females rising to the alpha position allows one to partition hormonal responses due to changes in social status from those due to sex change. Specifically, KT in alphas increases at 24 h and 6d relative to alphas in stable groups, while it does not change significantly in any tissue of betas rising to alphas. These results demonstrate that changes in KT are closely associated with sex change, rather than a change in social status following male removal. In L. dalli, KT can induce masculinization of the genital papilla [50], so the increase is consistent with the male typical papilla ratio we found in sex changing alpha females at 6 days. The rise in KT during sex change might be needed to induce spermatogenesis or the onset of male typical behavior, but is not necessary for the maintenance of male function, as indicated by the lack of sex differences in KT in stable groups. In stoplight parrotfish, serum KT increased in females during sex change, and then again when changing color into the terminal male phase [33] supporting the hypothesis that KT is required during transitions from one reproductive state to another. The mechanism of action of KT is not clear yet, but administration of KT induced the start of sex change and depressed levels of E2 in stoplight parrotfish [51], and in the honeycomb grouper Epinephelus merra [32]. Increased KT during sex change might function primarily to inhibit E2 production allowing oocytes to regress and gonadal sex change to proceed.

We found increased KT in both brain and gonads of sex changing alpha females, so extra-gonadal synthesis of KT could be essential to initiate sex change. Interestingly, in stable groups alpha females had the lowest levels of KT in all tissues and this might make the rapid increase in brain and gonadal KT that happens during sex change even more dramatic and effective.

In L. dalli, aromatase activity decreases during the first stages of sex change in the brain but not in the gonad [17] so we expected a drop in brain E2 levels during sex change, however this was not the case. Changes in E2 levels do not seem to play a major role in the initiation of sex change. In fact, sex changing alphas 6 days after male removal still have E2 levels similar to alpha females in stable groups and much higher than males. Our results show that the ovary in L. dalli has the highest levels of E2 while in other teleosts aromatase activity is typically reported to be higher in brain than gonad [52], [13] L. dalli brain has very high level of T, which is the substrate for aromatase, so we would expect much higher levels of E2 in the brain than in the gonad if aromatase activity was equivalent.

Since in L. dalli the gonad produces much more E2 than the brain, gonadal E2 coming into the brain from circulation could swamp our ability to detect differences in brain E2 levels during the early phases of sex change, or brain aromatase and E2 could just be modulated at a very fine local scale, consistent with a synaptocrine regulation of E2 in the brain [53].

In contrast with what happens to KT, concentrations of T in L. dalli decrease in brain and muscle of both alpha and beta females after male removal, suggesting an association of T with the change in social circumstance rather than the induction of sex change per se. Other animals have been shown to change steroid levels after being presented with an individual of the opposite sex [54], [55], [56]. The presence of males might increase T levels in female L. dalli in stable groups while male removal might precipitate a decrease in T synthesis. It is worth noting that the strong relationship between KT and T in the other females was lacking in alphas undergoing sex change (Fig. 4). Both alphas and betas do show the drop in T brain levels in response to male removal suggesting that the increase in KT which is only present in alphas, is responsible of the lack of correlation between T and KT. Further work will be needed to understand the mechanism peculiar to the sex changing fish that can induce this rapid increase in KT.
Androgens and Aggression

We found a positive relationship between androgens and aggression in alpha females but only after male removal when the social situation became unstable. For sex changing females at 24 h, gonad and muscle KT positively correlated with the total rate of approaches to other fish, and muscle KT 6d after male removal with the rate of displacement of subordinate females. We found that alpha female E2 levels correlated with aggression 6d after male removal. The time shortly after male removal is the most critical phase in the onset of sex change because that is when the dominant female takes the place of the male in the hierarchy. Thus, aggression levels displayed by an alpha female at that stage might affect hormone levels on the following days. On the day of male removal, L. dalli sex changing females often show increased aggression (e.g., [26], [17]) that can last for several days, but then returns to levels similar to stable groups. Most of the significant correlations we found involve the behavior after male removal, so sex steroids might affect behavior only in situations of social instability [4] as suggested by the “challenge hypothesis” [5]. The vast majority of studies finding a relationship between androgens and aggression use the resident/intruder paradigm, while in our experiment the fish already established a stable social structure so the hormone-behavior responses might be less dramatic.

Conclusions

Our results show that different tissues within the same animal often have very different steroid levels, and these tissues can respond differently to the same social stimulus. Thus, measures of circulating levels of steroids may not provide adequate information to describe the local steroidal milieu and thus the primary effects of steroid action, which take place at the local level. Our data suggest that the gonad in L. dalli is the main source of estrogen and that the brain is an important source of androgens. Future studies should seek to determine whether these tissue differences are due to differential rates of synthesis/catabolism or active compartmentalization by steroid binding globulins or steroid receptors. Androgens seem to be an important candidate for steroids that translate changes in the social environment into changes in sexual phenotype. In fact, during the first 6 days of sex change, E2 levels stay stable in all tissues, T decreases dramatically in brain and muscle, and KT increases in the brain. The drop in T could act as a chemical message to signal the lack of mating opportunity, but it is not sufficient to induce sex change without the increase in KT, which only occurs in the sex changing alpha female that is now the new dominant animal. We propose that an increase in brain KT and an associated drop in brain T may serve as a steroidal switch that initiates sex change. Future studies also should consider regional differences in brain steroid levels, and the potential for pulsatile steroid secretion and neurosteroids other than those measured in this study (e.g. DHEA-S) to mediate the behavioral and physiological responses to changes in status and sex. Socially regulated sex change provides an excellent model in which to study the evolution of novel mechanisms for allowing social interactions to regulate steroid signaling and reproductive phenotype.

Supporting InformationFigure S1.
Steroid correlations across tissues. Correlations between concentration of KT in the gonad versus KT in the brain (A) across all fish from stable and sex changing groups, between concentrations of E2 in the gonad versus the brain of beta (B) and alpha (C) females, and between concentration of KT in the muscle versus the gonad of alpha females (D). A solid line means that the associated p value is significant (p<0.05) while dotted lines are associated with non-significant correlations.Figure S2.
Correlations across hormones. Correlation between concentration of T versus KT in the brain of beta females in stable groups (A), and in sex changing groups (B). Correlation between concentration of E2 in the brain versus KT in the gonad (C) of alpha females. A solid line means that the associated p value is significant (pFigure S3.
Correlation between hormones and behavior. Correlation between the rate of displacements received 1 day after male removal and brain concentration of T (A), and KT (B); between the rate of displacements received in the afternoon 5 days after male removal and brain concentration of T (C) and KT (D); between the total rate of displacements received after male removal and concentration of KT in the gonad (E); and between the rate of displacements given before male removal and concentration of E2 in the gonad (F). The values are from beta females in sex changing groups collected 24 hours and 6 days after male removal. A solid line means that the associated p value is significant (p<0.05) while dotted lines refer to non-significant correlations.Table S1.
List of all linear contrasts that are significant or show a trend towards significant values for fish in stable groups.Table S2.
List of all linear contrasts that are significant or show a trend towards significant values for sex changing groups. These linear contrasts compare alpha and beta females in stable groups and in groups undergoing sex change (24 hours and 6 days).Table S3.
Three-way linear contrasts comparing females.Table S4.
Linear contrasts between stable males and females in sex changing groups.Table S5.
Three-way linear contrasts comparing males and females.Methods S1.

Acknowledgments

We would like to thank David Sinkiewicz and Vincente Colunga for help in hormone extraction, Bridget Wynn, Scott Wilson, Jeff Glenn and Yong Ah for helping with behavioral observations, the staff at USC Wrigley Institute for Environmental Studies for logistical assistance, and two anonymous reviewers for helpful comments on the manuscript.
Author Contributions

Conceived and designed the experiments: VL RLE MSG. Performed the experiments: VL RLE. Analyzed the data: VL RLE. Contributed reagents/materials/analysis tools: RLE MSG. Wrote the paper: VL RLE MSG.

References

1. Nelson RJ (2005) An introduction to behavioral endocrinology. Sinauer Associate, Inc. Sunderland, MA.
2. Liley NR, Breton B, Fostier A, Tan ESP (1986) Endocrine changes associated with spawning behavior and social stimuli in a wild population of rainbow trout (Salmo gairdneri): I. Males. Gen Comp Endocrinol 62: 145–156. Find this article online
3. Oliveira RF (2009) Social behavior in context: hormonal modulation of behavioral plasticity and social competence. Integrat Comp Biol 49: 423–440. Find this article online
4. Oliveira RF, Hirschenhauser K, Carneiro LA, Canário AVM (2002) Social modulation of androgen levels in male teleost fish. Comparative Biochemistry and Physiology Part B: Biochem. Mol Biol 132: 203–215. Find this article online
5. Wingfield JC (1984) Environmental and endocrine control of reproduction in the song sparrow, Melospiza melodia : II. Agonistic interactions as environmental information stimulating secretion of testosterone. Gen Comp Endocrinol 56: 417–424. Find this article online
6. Remage-Healey L, Bass AH (2005) Rapid elevations in both steroid hormones and vocal signaling during playback challenge: a field experiment in Gulf toadfish. Horm Behav 47: 297–305. Find this article online
7. Remage-Healey L, Bass AH (2006) From social behavior to neural circuitry: steroid hormones rapidly modulate advertisement calling via a vocal pattern generator. Horm Behav 50: 432–441. Find this article online
8. Elekonich MM, Wingfield JC (2000) Seasonality and hormonal control of territorial aggression in female song sparrows (Passeriformes: Emberizidae: Melospiza melodia). Ethology 106: 493–510. Find this article online
9. Davis ES, Marler CA (2003) The progesterone challenge: steroid hormone changes following a simulated territorial intrusion in female Peromyscus californicus. Horm Behav 44: 185–198. Find this article online
10. Desjardins JK, Hazelden MR, Van der Kraak GJ, Balshine S (2006) Male and female cooperatively breeding fish provide support for the “Challenge Hypothesis”. Behav Ecol 17: 149–154. Find this article online
11. Taves MD, Ma C, Heimovics SA, Saldanha CJ, Soma KK (2011) Measurement of steroid concentrations in brain tissue: methodological considerations. Front Endocrinol 2: 1–13. Find this article online
12. Andersson E, Borg B, Lambert JGD (1988) Aromatase activity in brain and pituitary of immature and mature Atlantic salmon (Salmo salar L.) parr. Gen Comp Endocrinol 72: 394–401. Find this article online
13. Pasmanik M, Callard GV (1985) Aromatase and 5[alpha]-reductase in the teleost brain, spinal cord, and pituitary gland. Gen Comp Endocrinol 60: 244–251. Find this article online
14. Mayer I, Borg B, Schulz RD (1990) Seasonal changes in and effect of castration/androgen replacement on the plasma levels of five androgens in the male three-spined stickleback, Gasterosteus aculeatus L. Gen Comp Endocrinol. 79: 23–30. Find this article online
15. Schulz R, Blum V (1991) Extragonadal 17[beta]-hydroxysteroid dehydrogenase activity in rainbow trout. Gen Comp Endocrinol 82: 197–205. Find this article online
16. Balthazart J, Ball GF (2006) Is brain estradiol a hormone or a neurotransmitter? Trends Neurosci 29: 241–249. Find this article online
17. Black MP, Balthazart J, Baillien M, Grober MS (2005) Socially induced and rapid increases in aggression are inversely related to brain aromatase activity in a sex-changing fish, Lythrypnus dalli. Proc R Soc B-Biol Sci 272: 2435–2440. Find this article online
18. Cornil CA, Ball GF, Balthazart J (2006) Functional significance of the rapid regulation of brain estrogen action: Where do the estrogens come from? Brain Res 1126: 2–26. Find this article online
19. Remage-Healey L, Bass AH (2006b) A rapid neuromodulatory role for steroid hormones in the control of reproductive behavior. Brain Res 1126: 27–35. Find this article online
20. Schmidt KL, Pradhan DS, Shah AH, Charlier TD, Chin EH, et al. (2008) Neurosteroids, immunosteroids, and the balkanization of endocrinology. Gen Comp Endocrinol 157: 266–274. Find this article online
21. Woolley CS (2007) Acute effects of estrogen on neuronal physiology. Annual Rev Pharmacol Toxicol 47: 657–680. Find this article online
22. Remage-Healey L, Maidment NT, Schlinger BA (2008) Forebrain steroid levels fluctuate rapidly during social interactions. Nature Neurosci 11: 1327–1334. Find this article online
23. Schultz DR, Perez N, Tan CK, Mendez AJ, Capo TR, et al. (2005) Concurrent levels of 11-ketotestosterone in fish surface mucus, muscle tissue and blood. J App Ichthyol 21: 394–398. Find this article online
24. Barnett L, Earley R, Ebert D, Cailliet G (2009) Maturity, fecundity, and reproductive cycle of the spotted ratfish, Hydrolagus colliei. Mar Biol 156: 301–316. Find this article online
25. Nakamura M, Hourigan TF, Yamauchi K, Nagahama Y, Grau EG (1989) Histological and ultrastructural evidence for the role of gonadal steriod hormones in sex change in the protogynous wrasse Thalassoma duperrey. Environ Biol Fishes 24: 117–136. Find this article online
26. Reavis RH, Grober MS (1999) An integrative model of sex change: social, behavioural and neurochemical changes in Lythrypnus dalli (Pisces) Acta Ethol. 2: 51–60. Find this article online
27. Rodgers EW, Earley RL, Grober MS (2007) Social status determines sexual phenotype in the bi-directional sex changing bluebanded goby Lythrypnus dalli. J Fish Biol 70: 1660–1668. Find this article online
28. Frisch A (2004) Sex-change and gonadal steroids in sequentially-hermaphroditic teleost fish. Rev Fish Biol Fisheries 14: 481–499. Find this article online
29. Kroon FJ, Liley NR (2000) The role of steroid hormones in protogynous sex change in the blackeye goby, Coryphopterus nicholsii (Teleostei: Gobiidae) Gen Comp Endocrinol. 118: 273–283. Find this article online
30. Kroon FJ, Munday PL, Pankhurst NW (2003) Steroid hormone levels and bi-directional sex change in Gobiodon histrio.. J Fish Biol 62: 153–167. Find this article online
31. Bhandari RK, Alam MA, Higa M, Soyano K, Nakamura M (2005) Evidence that estrogen regulates the sex change of honeycomb grouper (Epinephelus merra), a protogynous hermaphrodite fish. J. Exp. Zool. Part A-Comp Exp Biol 303A: 497–503. Find this article online
32. Bhandari RK, Alam MA, Soyano K, Nakamura M (2006) Induction of female-to-male sex change in the honeycomb grouper (Epinephelus merra) by 11-ketotestosterone treatments. Zool Sci 23: 65–69. Find this article online
33. Cardwell JR, Liley NR (1991) Hormonal control of sex and color change in the stoplight parrotfish, Sparisoma viride. Gen Comp Endocrinol 81: 7–20. Find this article online
34. Kobayashi M, Aida K, Stacey NE (1991) Induction of testis development by implantation of 11-ketotestosterone in female goldfish. Zool Sci 8: 389–393. Find this article online
35. Lone KP, Al-Ablani S, Al-Yaqout A (2001) Steroid hormone profiles and correlative gonadal histological changes during natural sex reversal of sobaity kept in tanks and sea-cages. J Fish Biol 58: 305–324. Find this article online
36. Ohta K, Mine T, Yamaguchi A, Matsuyama M (2001) Steroidogenic Pathway to Estradiol-17β synthesis in the ovarian follicles of the protogynous wrasse, Pseudolabrus sieboldi.. Zool Sci 18: 937–945. Find this article online
37. Yeung WSB, Chan STH (1987) A radioimmunoassay study of the plasma levels of sex steriods in the protandrous, sex-reversing fish Rhabdosargus sarba (Sparidae) Gen Comp Endocrinol. 66: 353–363. Find this article online
38. Behrents KC (1983) The comparative ecology and interactions between two sympatric gobies (Lythrypnus dalli and Lythrypnus zebra) PhD Dissertation, University of Southern California.
39. Lorenzi V, Earley RL, Rodgers EW, Pepper DR, Grober MS (2008) Diurnal patterns and sex differences in cortisol, 11-ketotestosterone, testosterone, and 17 beta-estradiol in the bluebanded goby (Lythrypnus dalli). Gen Comp Endocrinol 155: 438–446. Find this article online
40. Newman AEM, Chin EH, Schmidt KL, Bond L, Wynne-Edwards KE, et al. (2008) Analysis of steroids in songbird plasma and brain by coupling solid phase extraction to radioimmunoassay. Gen Comp Endocrinol 155: 503–510. Find this article online
41. Borg B (1994) Androgens in teleost fishes. Comp. Biochem. Physiol. 109C, 219–245.
42. Diotel N, Do Rego JL, Anglade I, Vaillant C, Pellegrini E, et al. (2011) The Brain of Teleost Fish, a Source, and a Target of Sexual Steroids. Frontiers in Neuroscience 5: 137. Find this article online
43. Avvakumov GV, Zhuk NI, Strel’chyonok OA (1986) Subcellular distribution and selectivity of the protein-binding component of the recognition system for sex-hormone-binding protein-estradiol complex in human decidual endometrium. Biochim. Biophys Acta Gen Sub 881: 489–498. Find this article online
44. Caldwell JD, Shapiro RA, Jirikowski GF, Suleman F (2007) Internalization of sex hormone-binding globulin into neurons and brain cells in vitro and in vivo. Neuroendocrinol 86: 84–93. Find this article online
45. Neat FC, Mayer I (1999) Plasma concentrations of sex steroids and fighting in male Tilapia zillii. J Fish Biol 54: 695–697. Find this article online
46. Miguel-Queralt S, Avvakumov GV, Blázquez M, Piferrer F, Hammond GL (2005) Sea bass (Dicentrarchus labrax) sex hormone binding globulin: molecular and biochemical properties and phylogenetic comparison of its orthologues in multiple fish species. Mol Cell Endocrinol 229: 21–29. Find this article online
47. Miguel-Queralt S, Knowlton M, Avvakumov GV, Al-Nouno R, Kelly GM, et al. (2004) Molecular and functional characterization of sex hormone binding globulin in zebrafish. Endocrinol 145: 5221–5230. Find this article online
48. Pasmanik M, Callard GV (1986) Characteristics of a testosterone-estradiol binding globulin (TEBG) in goldfish serum. Biol Reprod 35: 838–845. Find this article online
49. Godwin JR, Thomas P (1993) Sex-change and steroid profiles in the protandrous anemonefish Amphiprion melanopus (Pomacentridae, Teleostei). Gen Comp Endocrinol 91: 144–157. Find this article online
50. Carlisle SL, Marxer-Miller SK, Canario AVM, Oliveira RF, Carneiro L, et al. (2000) Effects of 11-ketotestosterone on genital papilla morphology in the sex changing fish Lythrypnus dalli. J Fish Biol 57: 445–456. Find this article online
51. Cardwell JR, Liley NR (1991) Androgen control of social status in males of a wild population of stoplight parrotfish, Sparisoma viride (Scaridae). Horm Behav 25: 1–18. Find this article online
52. Li GL, Liu XC, Lin HR (2007) Seasonal changes of serum sex steroids concentration and aromatase activity of gonad and brain in red-spotted grouper (Epinephelus akaara). Anim Reprod Sci 99: 156–166. Find this article online
53. Saldanha CJ, Remage-Healey L, Schlinger BA (2011) Synaptocrine Signaling: Steroid Synthesis and Action at the Synapse. Endocr Rev doi: 10.1210/er.2011-0004.
54. Kamel F, Mock EJ, Wright WW, Frankel AI (1975) Alterations in plasma concentrations of testosterone, LH, and prolactin associated with mating in the male rat. Horm Behav 6: 277–288. Find this article online
55. Macrides F, Bartke A, Dalterio S (1975) Strange females increase plasma testosterone levels in male mice. Science 189: 1104–1106. Find this article online
56. Pinxten R, de Riddera E, Eensa M (2003) Female presence affects male behavior and testosterone levels in the European starling (Sturnus vulgaris). Horm Behav 44: 103–109. Find this article online