Abstract
Increasing evidence suggests the crucial role of estrogen in ovarian differentiation of nonmammalian vertebrates including fish. The present study has investigated the plausible role of Foxl2 in ovarian differentiation through transcriptional regulation of aromatase gene, using monosex fry of tilapia. Foxl2 expression is sexually dimorphic, like Cyp19a1, colocalizing with Cyp19a1 and Ad4BP/SF-1 in the stromal cells and interstitial cells in gonads of normal XX and sex-reversed XY fish, before the occurrence of morphological sex differentiation. Under in vitro conditions, Foxl2 binds to the sequence ACAAATA in the promoter region of the Cyp19a1 gene directly through its forkhead domain and activates the transcription of Cyp19a1 with its C terminus. Foxl2 can also interact through the forkhead domain with the ligand-binding domain of Ad4BP/SF-1 to form a heterodimer and enhance the Ad4BP/SF-1 mediated Cyp19a1 transcription. Disruption of endogenous Foxl2 in XX tilapia by overexpression of its dominant negative mutant (M3) induces varying degrees of testicular development with occasional sex reversal from ovary to testis. Such fish display reduced expression of Cyp19a1 as well as a drop in the serum levels of 17β-estradiol and 11-ketotestosterone. Although the XY fish with wild-type tilapia Foxl2 (tFoxl2) overexpression never exhibited a complete sex reversal, there were significant structural changes, such as tissue degeneration, somatic cell proliferation, and induction of aromatase, with increased serum levels of 17β-estradiol and 11-ketotestosterone. Altogether, these results suggest that Foxl2 plays a decisive role in the ovarian differentiation of the Nile tilapia by regulating aromatase expression and possibly the entire steroidogenic pathway.
GONADAL SEX DIFFERENTIATION is an indispensable and fundamental process that ensures the sexual integrity of a vertebrate through the proper development of the bipotential gonad into a testis or an ovary. Although vast information has been available on the sex-determining gene (SRY/Sry) that induces mammalian testicular differentiation and its downstream genes, very little is known about the female pathway. Unlike mammals, estrogen plays a crucial role in the ovarian sex differentiation of nonmammalian vertebrates, including fish (1, 2). Even though the tilapia gonad remains morphologically undifferentiated until 25 d after hatching (dah), the female-specific expression of Cyp19a1, encoding aromatase that catalyzes the conversion of androgen to estrogen, starts from around 5 dah, before the onset of morphological differentiation of the ovary. Treatment of tilapia XX fry with fadrozole (aromatase inhibitor) or tamoxifen (estrogen receptor antagonist) caused complete sex change to functional males. This sex reversal could be rescued by simultaneous treatment with 17β-estradiol (E2), providing solid proof to the notion that endogenous estrogen is the natural inducer of ovarian differentiation (1, 3, 4).
There are two aromatase genes described in fish, Cyp19a1 and Cyp19a2 (5–8). Cyp19a1 was found to encode the fish ovarian type aromatase, involved in the production of estrogens in the fish gonad. The fish Cyp19a1 promoter region contains Ad4BP/SF-1 binding sites, suggesting Ad4BP/SF-1 as a likely candidate engaged in the regulation of Cyp19a1 at the transcriptional level, and investigations using medaka and tilapia have provided more insight on this aspect (9, 10). However, the role of Ad4BP/SF-1 in sex-specific transcriptional regulation of Cyp19a1 is unclear, as the fish Ad4BP/SF-1 failed to show any difference in the pattern of expression between the male and female sex during early sex differentiation (Ijiri, S., H. Kaneko, D. S. Wang, and Y. Nagahama, unpublished data and present study). Meanwhile, there have been several studies implicating the forkhead (FH) transcription factor, Foxl2, in ovarian development, granulosa cell differentiation, and thus the proper maintenance of ovarian function (11–16). It is the earliest known sex dimorphic marker, expressed in the somatic cells during early development and later in granulosa cells surrounding the oocytes (12, 17–20).
A number of studies have provided information on varying ranges of phenotypes, where Foxl2 was either mutated or knocked out, demonstrating its role in ovarian differentiation. The pathologies varied from blepharophimosis/ptosis/epicanthus inversus syndrome and premature ovarian failure in humans, where Foxl2 was mutated (11, 21), to a total absence of secondary follicles and oocyte atresia in the Foxl2 knockout mouse (14, 15). Foxl2 was suggested even as a repressor of the male pathway during female gonad development, because the mouse XX gonads without Foxl2, proceeded with the genetic program for somatic testis determination, even though these gonads had oocytes in the meiotic prophase initially (22). In order to delineate the entire molecular mechanisms through which Foxl2 carries out its role in the ovary, the identification of its target genes is a prerequisite.
The steroidogenic acute regulatory (StAR) gene was shown as a candidate, containing multiple putative FH consensus sites to some of which Foxl2 bound directly to repress its transcription (13). The expression of StAR was derepressed when Foxl2 had dominant negative mutations within it, causing premature ovarian failure. However, a recent review has suggested that derepression of StAR transcription in mice is not the cause for the early follicular block at the primordial follicular stage (23), leaving room for the involvement of some other Foxl2-mediated mechanisms. However, there is hardly any information about the other target genes of Foxl2 to bolster its precise function in ovarian determination/differentiation. Goats with polled intersex syndrome (PIS), where Foxl2 function was disrupted, displayed a reduction in the expression of aromatase also (24, 25), suggesting the latter to be controlled by Foxl2. Further evidence to support this notion was obtained from the studies on chicken, rainbow trout, and medaka fish, where Foxl2 expression was found to be correlated with the spatial and temporal expression of aromatase (17, 19, 20, 26) during sex differentiation and later follicular development. Furthermore, the preliminary data on the quantitative expression of Foxl2 also corroborated its positive correlation with the expression patterns of aromatase from as early as 5 dah (Ijiri, S., H. Kaneko, D. S. Wang, and Y. Nagahama, unpublished data) in the Nile tilapia.
Unfortunately, none of the aforementioned studies could furnish any direct evidence, either in vitro or in vivo, to prove the role of Foxl2 in the regulation of Cyp19/Cyp19a1, except for a recent study (27) published during the preparation of this manuscript, which provided an insight into the regulation of aromatase by Foxl2. The mechanistic relationship between Foxl2 and estrogen production and its physiological relevance in the process of ovarian sex differentiation continue to be uncovered. In the present study we have made an effort to uncover the possible mechanisms by which Foxl2 influences the undifferentiated gonad, using an array of in vitro and in vivo techniques, including transgenic approaches using wild-type and dominant-negative mutant forms of Foxl2. Our data imply that Foxl2 plays an important role in promoting gonadal differentiation toward the female pathway, either by binding directly to the promoter of Cyp19a1 or interacting with Ad4BP/SF-1 to enhance the Cyp19a1 transcription and thereby estrogen production specifically in the female gonad.
RESULTS
Expression of Foxl2 and Ad4BP/SF-1 Is Positively Correlated with Cyp19a1 Spatially and Temporally
To determine whether the expression patterns of Ad4BP/SF-1 and Foxl2 are positively correlated with the expression of Cyp19a1 spatially and temporally, in situ hybridization and immunocytochemistry were performed continuously from 5 dah. However, data for 5, 10, and 20 dah normal, as well as sex-reversed, gonads only are shown in Fig. 1a. In mammals, the sexual dimorphic expression of genes and gonadal morphogenesis start early in the embryogenesis. In contrast, in tilapia sexual dimorphic expression of genes starts from 5 dah, and morphological sex differentiation, such as formation of ovarian cavity (a special structure found in the teleost ovary, which does not have any analogous structure in the mammalian gonad), starts around 25 dah. Before morphological sex differentiation, the gonadal structure of tilapia is very simple, consisting mainly of interstitial cells and a few germ cells (28).
Colocalization of Ad4BP/SF-1, Foxl2, and Cyp19a1 in Tilapia Gonad Ad4BP/SF-1 expression was detected by in situ hybridization, whereas Cyp19a1 and Foxl2 were detected by immunohistochemistry. a, Expression of Ad4BP/SF-1, Cyp19a1, and Foxl2 during gonadal sex differentiation. A–F, Gonads from 5 dah; G–L, gonads from 10 dah; M–X, gonads from 20 dah; S–U, XX gonad after the induction of sex reversal with MT; V–X, XY gonad after the induction of sex reversal with E2 (see Materials and Methods). Arrow, Blood vessel (the blood vessel is not observable at 5 dah); arrowhead, germ cell. Scale bar, 10 μm. Tilapia Ad4BP/SF-1 is seen in the outermost epithelium in the gonads of both sexes at 5 dah (A and D). Thereafter, it is expressed in the stromal cells near the blood vessels and interstitial cells but disappears from outermost epithelium (G, J, M, P, S, and V). No sexual dimorphism is seen up to 20 dah. Expression of Cyp19a1 is found in the stromal cells near the blood vessel specifically in XX gonads from 10 dah (H). No expression of Cyp19a1 is observed in XY gonads, except for estrogen-induced XY sex reversal (W). Cyp19a1 positive cells (H) coincide with Ad4BP/SF-1-positive cells (G) in the XX gonads, but not in XY gonads (K and J). Foxl2 is seen in epithelial cells in the outermost epithelium of the XX gonads at 5 dah (C), but not in XY gonads (F). Thereafter, Foxl2 is expressed in the nucleus of the interstitial cells and stromal cells in XX gonads dominantly, but not in XY gonads. Foxl2 is also colocalized with Cyp19a1 in the stromal cells near the blood vessels and the interstitial cells during gonadal sex differentiation (I and O). In gonads after the induction of sex reversal, Foxl2 and Cyp19a1 expressions are up-regulated in the XY gonad (W and X) and down-regulated in the XX gonad (T and U). b, Expression of Ad4BP/SF-1, Cyp19a1, and Foxl2 in previtellogenic and vitellogenic ovaries. Ad4BP/SF-1, Cyp19a1, and Foxl2 are colocalized in interstitial cells (i) in previtellogenic ovary (A–C) and in granulosa cells (arrowhead) in vitellogenic ovary (D–F). Ad4BP/SF-1 mRNAs and Cyp19a1 were also found to be colocalized in theca cells (arrow); however, Foxl2 is never observed in the theca cells. Scale bar, A–C, 30 μm; inset of C, D–F, 50 μm. Rev., Sex reversed gonad.
At 5 dah, Ad4BP/SF-1 and Foxl2 were seen in the most outer epithelium of the gonads. Although Ad4BP/SF-1 was seen in the gonads of both sexes, Foxl2 was observed only in XX gonads. Cyp19a1 was not detectable in the gonads of either sex at 5 dah. Thereafter, Ad4BP/SF-1 was expressed in the stromal cells near the blood vessels and interstitial cells but disappeared from the outermost epithelium. No sexual dimorphism in Ad4BP/SF-1 expression was seen up to 20 dah. Cyp19a1 and Foxl2 were found to be expressed in the gonads of 10 and 20 dah normal XX and sex-reversed XY (E2 treated) fish. In these fish, the expression of these proteins was observed in the stromal cells near the blood vessel. Expression of these two genes was also found in the interstitial cells. On the contrary, there was hardly any trace of Cyp19a1 and Foxl2 expression in the 10 and 20 dah normal XY and sex-reversed XX [17α-methyltestostorone (MT)-treated] fish. These results have clearly demonstrated that during gonadal sex differentiation, Ad4BP/SF-1, Cyp19a1, and Foxl2 were colocalized in the stromal cells near the blood vessel and interstitial cells of female gonads, but not male gonads. In the later stages of ovarian development, Ad4BP/SF-1, Foxl2, and Cyp19a1 were colocalized in the interstitial cells of the previtellogenic ovary and granulosa cells of the vitellogenic follicles (Fig. 1b). These data indicated that the expression of Foxl2 and Cyp19a1 is related to phenotypic sex of the tilapia.
Foxl2 Enhances Ad4BP/SF-1-Activated Cyp19a1 Gene Expression in Human Embryonic Kidney (HEK) 293 Cells
Ad4BP/SF-1 alone could activate Cyp19a1 gene transcription in luciferase assay using HEK293 cells (Fig. 2A and supplemental Fig. 1 published as supplemental data on The Endocrine Society’s Journals Online web site at http://mend.endojornals.org). Foxl2 enhanced the Ad4BP/SF-1-activated Cyp19a1 expression (Fig. 2A), when it was cotransfected with Ad4BP/SF-1. Further, the impact of Foxl2 alone on tilapia Cyp19a1 gene transcription also was examined using HEK293 cells. Only a small change was observable in the luciferase activity, which calls into question the ability of Foxl2 to alter single-handedly the Cyp19a1 gene transcription.
Effect of Foxl2 and Ad4BP/SF-1 on Tilapia Cyp19a1 Promoter Activity in HEK293 Cells (A) and TM3 Cells (B) Foxl2 alone (100 ng) or together with Ad4BP/SF-1 (100 ng) expression vectors were cotransfected into HEK293 cells (A) and TM3 cells (B) with tilapia Cyp19a1 −2346-bp promoter construct (500 ng/well). The total amount of transfected plasmids, including the pRL-TK control vector (100 ng/well), was adjusted to 1.0 μg with empty vectors. Firefly and Renilla luciferase activities were measured 48 h after transfection. Relative luciferase activity was calculated by dividing the firefly luciferase activity with the Renilla luciferase activity. Results are the mean ± sd of triplicate transfections.
Foxl2 Alone Can Activate Tilapia Cyp19a1 Gene Transcription in TM3 Cells
Similar to HEK293 cells, when the mouse testicular cell line TM3 (established from Leydig cells) were transfected transiently with Ad4BP/SF-1 expression vector and tilapia Cyp19a1-promoter construct, luciferase activity was found to be enhanced. In contrast to the situation in HEK293 cells, there was a 5-fold increase in the luciferase activity when Foxl2 alone was used in the transfection. Furthermore, cotransfection of Foxl2 and Ad4BP/SF-1 increased the luciferase activity in an additive and dose-dependent manner (Fig. 2B).
Identification of Foxl2 Binding Site in the Cyp19a1 Promoter Region by Sequential Deletion of the Promoter
To determine the precise location to which Foxl2 binds for transcriptional activation of the Cyp19a1 gene, a series of Cyp19a1 promoter deletion constructs were generated and transfected into TM3 cells. Relative luciferase activity was measured to estimate the Foxl2 response element(s). Deletion of the upstream region from −2346 to −696 bp had no effect on activation mediated by Foxl2 alone. In contrast, activation by Foxl2 was considerably decreased by deletion of the −696 to −222 bp. However, deletion of the −438/−119 bp from the 2346-bp promoter had no effect on transcription mediated by Foxl2, indicating that the Foxl2 binding site should be lying between −696 to −438 bp. The search of this promoter region resulted in the finding of a core element ACAAATA, from −545 to −538 bp, similar to the 7-bp core recognition motif, 5′-(G/A)(T/C)(C/A)AA(C/T) A-3′ to which various FH factors bind (13, 29, 30). As expected, mutation of this element to ctcccgc considerably decreased the transcription activation efficiency of Foxl2 (Fig. 3). This result suggests that ACAAATA could be the possible binding site for Foxl2 in the Cyp19a1 promoter region.
5′-Deletion Mapping of the Foxl2-Binding Region on Tilapia Cyp19a1 Promoter (−2346 bp) Expression vectors (100 ng/well) for Foxl2 (+/−) and the pRL-TK control vector (100 ng/well) were cotransfected with tilapia Cyp19a1 promoter deletion constructs (500 ng/well) into TM3 cells. The total amount of transfected plasmids was adjusted to 1.0 μg with empty vectors. Firefly and Renilla luciferase activities were measured 48 h after transfection. Relative luciferase activity was calculated by dividing the firefly luciferase activity with the Renilla luciferase activity. Results are the mean ± sd of triplicate transfections.
Confirmation of Foxl2 Binding Site by EMSAs
The significance of the Foxl2 binding site was further confirmed by EMSAs (Fig. 4) using probe and mutated probe designed according to the sequences between −557 to −524 bp of Cyp19a1. In vitro-translated Foxl2 proteins with c-Myc tag, bound to the radiolabeled oligonucleotides, resulting in the formation of a specific band of protein/DNA complex. This did not occur when the mutated probe was used. Cold competitor (10–100×) could displace the above band in a dose-dependent manner. Furthermore, the band was supershifted by the addition of c-Myc monoclonal antibody. The same was observed when mouse Foxl2 also was used, proving its ability to bind to the same consensus sequence as well.
EMSAs Showing Binding of Foxl2 Protein to the Tilapia Cyp19a1 Promoter Radiolabeled Cyp19a1 promoter probe was incubated with in vitro-translated recombinant tilapia Foxl2 (tFoxl2) and mouse Foxl2 (mFoxl2) protein with the c-Myc-tag. Excess unlabeled probe and unlabeled mutated probe (M-probe) were used to demonstrate specific binding between Foxl2 and the Cyp19a1 promoter. The positions of the Foxl2-DNA probe complexes are shown with arrows. Mab, Monoclonal antibody.
Regions of the Foxl2 Protein Responsible for Its DNA Binding and Transactivation Function
To determine which regions of the Foxl2 protein are most important for its function, Foxl2 mutants were constructed by deleting the amino and carboxyl termini of the wild-type protein, sequentially (Fig. 5A). In tilapia Foxl2, the putative DNA-binding domain (FH) is located at the amino terminus similar to mammalian Foxl2s. However, the polyalanine tracts and the polyproline repeats found in mammalian Foxl2 were absent in fish. EMSAs with Foxl2 mutants M2, M3, and M4 showed that M3 and M4 can bind to the normal Cyp19a1 oligonucleotide probe but not the mutated probe, whereas M2 (without FH domain) could not bind to either of the probes (Fig. 5B). Foxl2 mutation constructs were cotransfected into TM3 cells with tilapia Cyp19a1 promoter-reporter construct. Measurement of the luciferase activity revealed that the conserved FH domain and the carboxyl terminus of Foxl2 were important for its activation ability because deletions of the carboxyl terminus (M3, M4, and M5) and the FH domain (M2) resulted in a complete loss of activity, whereas deletion of the amino terminus (M1) had very little effect (Fig. 5C). The mutant M3 showed a dominant-negative effect over the tFoxl2 (supplemental Fig. 2 published as supplemental Fig. 2 on The Endocrine Society’s Journals Online web site at http://mend.endojornals.org). Therefore, it was used later for the loss-of-function study in vivo.
Regions of the Foxl2 Protein Responsible for Its Activation Function A, Domain arrangement of Foxl2 and Foxl2 mutants. Foxl2 constructs with sequential deletions of the amino terminus, FH domain, and the carboxyl terminus were generated. B, EMSAs showing binding of Foxl2 protein to the tilapia Cyp19a1 promoter through FH. Radiolabeled normal Cyp19a1 promoter probe (N-probe) and mutated probe (M-probe) were incubated with in vitro-translated mutated tilapia Foxl2 protein (M2, M3, and M4). Arrows indicate the protein-DNA complexes. C, Cyp19a1 −2346- bp promoter analysis showing carboxyl terminus of Foxl2 as important for its transactivation function in TM3 cells. WT, Wild type.
Demonstration of the Interaction between Foxl2 and Ad4BP/SF-1 by Mammalian Two-Hybrid Assay, Pull-Down Assay, and Double Mutation of Foxl2 and Ad4BP/SF-1
To prove the interaction between Foxl2 and Ad4BP/SF-1, mammalian two-hybrid assays were performed. Tilapia Ad4BP/SF-1, Foxl2, and its deletion mutants (M1-M5) were subcloned into the pBIND vector and the pACT vector to produce the GAL4- and VP16-fusion proteins, respectively. The pBIND-Ad4BP/SF-1 and pACT-Foxl2/pACT-Foxl2-mutants were transfected along with pG5 lucVector into HEK293 cells. Even though expression of the fusion protein GAL4-Ad4BP/SF-1 plus VP-16 showed relatively high luciferase activity as compared with the other controls, when the pGAL4 and pVP16 fusion constructs were transfected along with pG5 lucVector into HEK293 cells, Foxl2 and deletion mutants, M1 and M3–M5, promoted higher luciferase activity as compared with the controls (Fig. 6). By contrast, cotransfection of M2 had no effect as compared with Gal4-Ad4BP/SF-1 alone. Same results were obtained when GAL4-Ad4BP/SF1-ligand-binding domain (LBD) fusion construct was used in the experiment (data not shown).
Foxl2 and Ad4BP/SF-1 Interaction as Revealed by the Mammalian Two-Hybrid Assays Tilapia Ad4BP/SF-1, Foxl2, and its deletion mutants (M1–M5) were subcloned into the pBIND vector and the pACT vector to produce the GAL4- and VP16-fusion proteins, respectively. The pBIND vector expressed the R. reniformis luciferase was also used to normalize the transfection efficiency. The pBIND-Ad4BP/SF-1 and pACT-Foxl2 or pACT-M1 or pACT-M2 or pACT-M3 or pACT-M4 or pACT-M5 plasmids were transfected along with pG5 lucVector, which contains five Gal4 binding sites into HEK293 cells. The cells were lysed 2 d after transfection, and the amount of firefly luciferase and Renilla luciferase were quantitated using the Dual-Luciferase Reporter Assay System. Relative luciferase activity was calculated by dividing the firefly luciferase activity with the Renilla luciferase activity. Results are presented as the mean ± sd of data from triplicate transfections.
The interaction between Foxl2 and Ad4BP/SF-1 was further investigated by pull-down assays. Images of the coimmunoprecipitation of fluorescent hemagglutinin (HA)-tagged Ad4BP/SF-1, Ad4BP/SF-1-LBD, and c-Myc-tagged Foxl2, M2-M5 are shown in Fig. 7. Tilapia Ad4BP/SF-1 showed intense interaction with Foxl2 of both mouse and tilapia and the same occurred with mouse Ad4BP/SF-1. In the case of M2–M5, interaction was observed between Ad4BP/SF-1 and M3-M5, but not M2. When Ad4BP/SF-1-LBD was used to pull-down Foxl2 and M2, a strong band was observed only in the lane for Foxl2 but not in the lane for M2. In addition, strong interaction was also observed between tilapia Lrh-1 and Foxl2.
Foxl2 and Ad4BP/SF-1 Interaction as Revealed by in Vitro Pull-Down Assays Ad4BP/SF-1, Lrh-1, and Foxl2 ORFs and Foxl2 deletion mutants (M2–M5) were subcloned into the pGADT7 and pGBKT7. For each protein indicated, both c-Myc- and HA-fusion proteins were obtained by in vitro translation and labeled by FluroTect GreenLYS in vitro Translation Labeling System. Coimmunoprecipitations were performed using the MATCHMAKER Co-IP Kit according to the manufacturer’s instructions. Reactions were carried out with the equivalent amount of each protein. Each protein was coimmunoprecipitated with both c-Myc and HA antibodies along with the protein A beads. Data from only one direction (HA) are shown in the figure. Coprecipitated proteins were analyzed by SDS-PAGE. Fluorescent image of the gel was scanned and visualized by laser fluoroimager: lanes 1–6, interactions between tilapia tAd4BP/SF-1 and tFoxl2, tFoxl2 mutations (M2–M5) and mouse Foxl2 (mFoxl2); lanes 7 and 8, interactions between mouse mAd4BP/SF-1 and tFoxl2 and mFoxl2; lane 9, interaction between tilapia Lrh-1 and tFoxl2; lanes 10 and 11, interactions between the tAd4BP/SF-1-LBD and tFoxl2 and M2.
The interaction between Foxl2 and Ad4BP/SF-1 was further confirmed by using single and double mutations of Foxl2 and Ad4BP/SF-1 in Cyp19a1 promoter assay as shown in Fig. 8. M1, M4, and M5 exhibited the ability to enhance the Ad4BP/SF-1-activated Cyp19a1 promoter activity even though more weakly than that of tFoxl2, whereas M2 showed no effect. Interestingly, M3 showed a strong suppression effect on Ad4BP/SF-1-activated Cyp19a1 promoter activity. When Ad4BP/SF-1-LBD was used in the assay, the LBD alone or together with Foxl2 or its mutants M1–M3 could not show any effect.
Single or Double Mutations of Foxl2 and Ad4BP/SF-1 on Tilapia Cyp19a1 Promoter Activity in HEK293 Cells The expression vectors carrying Foxl2 or its mutants (M1–M5) (100 ng) alone or in combination with Ad4BP/SF-1 or its mutant Ad4BP/SF-1-LBD (100 ng) were cotransfected into HEK293 cells along with tilapia Cyp19a1 −696-bp promoter construct (500 ng/well). The total amount of transfected plasmids was adjusted to 1.0 μg with pcDNA3.1 empty vectors (Mock). Luciferase activity was measured 48 h after transfection. Results are represented as the mean ± sd of triplicate transfections.
Overexpression of Dominant-Negative Mutant Form (M3) of Foxl2 in XX Fish
Because M3 was found to have a dominant-negative effect over wild-type Foxl2 in luciferase assay (supplemental Fig. 2), this was further analyzed by overexpression in transgenic fish, so as to ascertain the function of Foxl2 in vivo. M3 cDNA was subcloned into the multiple cloning sites downstream to the cytomegalovirus (CMV) promoter of the pIRES-hrGFP-1a vector. After injection into the fertilized egg, the transgene and the humanized recombinant green fluorescent protein (hrGFP) sequence were transcribed as a single mRNA but were translated as two separate proteins. Mosaic expression of GFP signal was observed in the degenerating ovaries of the 2- to 6-month-old XX fish with M3 expression (Fig. 9, A and B). Histological analysis revealed that these ovaries had a varying degree of degeneration starting from the posterior to anterior direction, with maximum degeneration in that area where the GFP signal was very intense, denoting high expression levels of the transgene in that particular area (Fig. 9A). Transgenic XX tilapia displayed reduced expression of Cyp19a1, with disfigured oocytes and follicle cells leading to follicular atresia (Fig. 9H) and, in some cases, there was a complete sex change from ovary to testis (Fig. 9G). In the case of complete sex reversal, it occurred through sequential changes that started with oocyte atresia and was followed by partial development of the testicular tissue with no germ cells as revealed by vasa staining (Fig. 9, D–F), consequently leading to the ultimate change from ovary to functional testis marked by the appearance of new germ cells. The fish with partial sex change were characterized by ovaries with either a delay in the formation of the ovarian cavity, or an opened ovarian cavity even at 4 months after hatching (Fig. 9C), whereas another group showed the formation of a distinct efferent duct-like structure in the ovary (data not shown).
Transgenic Overexpression of Foxl2 in XY Gonad (J and K) and Its Dominant-Negative Mutant (M3) in XX Gonad (A–H) A, Ovary from a 5-month-old XX fish after overexpression of M3 to show the degeneration of the posterior region (a, anterior; p, posterior; r, right gonad; and l, left gonad); inset, GFP signal in the posterior degenerated region. B, The ovary from an 8-month- old XX fish showing the mosaic expression of GFP (M3) signal in dark field; inset, bright field. C, Ovary section from 5-month-old XX fish showing the open ovarian cavity (arrow). D–F, Antivasa staining to show the initiation of degeneration (D), partial degeneration (E), and complete degeneration (F) of the ovary and the appearance of testicular structure (indicated by arrows in E and F). G, Hematoxylin-eosin staining of the gonad section of a 4-month-old XX fish to show the complete sex reversal from ovary to functional testis. H, Antiaromatase staining to show the decrease in aromatase (asterisk) expression level, deformed oocytes, and misshapen follicle cells in comparison with the control (I). J, Testis section from a 3-month-old XY fish, stained with aromatase antibody to show the loss of germ cells, proliferation of somatic cells, and appearance of aromatase-positive cells/cell clusters after overexpression of Foxl2 compared with the control testis (inset). K, Sections from a 4-month-old XY fish stained with SCC antibody to show the proliferation of the interstitial cells (s) and degeneration of testicular tissue, especially spermatogonia in the periphery region (dp) after overexpression of Foxl2, compared with the control (L).
Overexpression of tFoxl2 in XY Fish
Unlike the fish with M3 overexpression, the fish with tFoxl2 overexpression never exhibited a complete sex reversal, despite significant structural changes in the gonads of these fish. The somatic cells were found to be proliferating around the large blood vessel in the dorsal region, with induced aromatase gene expression (Fig. 9J). In addition, increased aromatase expression was observed in the proliferating somatic cells near the efferent duct and in interstitial cells neighboring the spermatogonia. When the peripheral region of the testis with overexpressed tFoxl2 manifested degeneration, the interstitial cells adjacent to the spermatogenic cysts toward the inner area showed tremendous growth (Fig. 9K), which probably made these gonads look like ovaries outwardly. These gonads were three times larger than the control testis in diameter (Fig. 9L). These interstitial cells were cholesterol side chain cleavage enzyme (SCC) positive (Fig. 9K), indicating them to be steroidogenically active and, in the control fish, SCC-positive cells were far less in number (Fig. 9L). However, these SCC-positive cells were aromatase negative when stained with aromatase antibody (data not shown). Above all, the spermatogenic cysts in the transgenic testis also were enlarged in size compared with those of the control testis (Fig. 8, K and L).
Serum E2 and 11-Ketotestosterone (11-KT) Levels of Transgenic and Control Fish
E2 and 11-KT are the major estrogen and androgen found in tilapia serum, respectively. To determine whether or not Foxl2 really influences the estrogen and androgen production, we collected blood samples from the 6-month-old transgenic and control fish and measured the E2 and 11-KT levels (Fig. 10). The serum E2 level of the control females (14.7 ng/ml) was 15 times higher than that of the males (0.9 ng/ml), whereas the serum 11-KT level of the male (0.9 ng/ml) was two times higher than that of the female fish (0.4 ng/ml). The level of E2 was much higher than the level of 11-KT in the control female fish, as expected. No such difference in the levels of these two hormones was observed in the control male fish of the same age. Overexpression of tFoxl2 in the XY male resulted in increased levels of both E2 and 11-KT, E2 being at the same level as that of XX control fish, but with a 10-fold higher 11-KT level than that of the XY control fish. In contrast, overexpression of M3 in the XX female resulted in low levels of both E2 and 11-KT. These transgenic fish had an E2 level (2.1 ng/ml) that was seven times lower than that of the XX control fish but two times higher than that of the XY control fish and an 11-KT level (0.1 ng/ml) much lower than that of both the XX (0.4 ng/ml) and XY (0.9 ng/ml) control fish.
Impact of Transgenic Overexpression of Foxl2 and Its Dominant-Negative Mutant (M3) on Tilapia Serum E2 and 11-KT Levels Results are presented as the mean ± sd. Sample numbers are shown on the figure. Ctrl, Control.
DISCUSSION
There have been numerous studies since the discovery of Foxl2, to elucidate its putative role in female sex differentiation and maintenance. Despite all these attempts, the function of Foxl2 and the mechanism through which it acts continue to be an enigma because no study has defined the targets of Foxl2, other than StAR. Hence, it was quite tempting to search the possibility of aromatase, encoded by Cyp19/Cyp19a1 as the downstream target of Foxl2, because of the convincing evidence from the studies on goats with PIS (24, 25). Cyp19a1 is the key enzyme involved in the production of endogenous estrogen, which has been shown to be the natural inducer of ovarian differentiation in tilapia (1, 3, 4). This made the need for fathoming the possible role of Foxl2 in the female-specific enhancement of estrogen via Cyp19a1 increasingly important.
Foxl2 Regulates Cyp19a1 Gene Expression by Direct Binding to the Promoter
In transfection assays using TM3 cells, we observed that Foxl2 alone can activate aromatase gene transcription. Further, by use of sequential deletion mutants of Cyp19a1 promoter region in transfection studies along with EMSAs, the response element to which Foxl2 binds was identified. Furthermore, EMSAs and luciferase assays were carried out using Foxl2 deletion mutants M1–M5 to prove the binding between Foxl2 and Cyp19a1. The deletion mutants established that the FH domain was indispensable for Foxl2 to bind to its recognition motif on Cyp19a1. However, although FH domain was sufficient for Foxl2 to bind to the DNA, luciferase assays revealed that the FH domain alone was not sufficient for Foxl2 to carry out its transactivation function. The transactivation domain was found to be situated at the C terminus because the deletion mutants without this region (M3, M4, and M5) could not activate the Cyp19a1 transcription. The finding was consistent with the features of FOXJ2, which activated transcription mainly through its proline-rich domain (H/P domain) and acidic blob (AB domain) located at the C terminus of the protein (29, 30). However, unlike FOXJ2, there is no transactivation domain found to be located at the N terminus of Foxl2.
Foxl2 Regulates Cyp19a1 Gene Expression by Interacting with Ad4BP/SF-1
In contrast to the situation found in TM3 cells, in which Foxl2 alone activated the Cyp19a1 gene transcription, the luciferase assays using HEK293 cells showed that Foxl2 alone had no effect on the Cyp19a1 promoter activity; however, in the presence of Ad4BP/SF-1, it enhanced Ad4BP/SF-1-induced activation of the Cyp19a1 in a synergistic manner. A similar finding has been reported in goats, i.e. Foxl2 had an enhancing effect on goat Cyp19 transcription in sheep granulosa cells, whereas such an effect was not observed in COS7 cells (27). We also tested and found that Foxl2 could enhance the tilapia Cyp19a1 transcription in COS7 cells in the presence of Ad4BP/SF-1 (data not shown). One of the obvious differences between these cell lines is that TM3 cells and granulosa cells are steroid-producing cells that contain endogenous Ad4BP/SF-1 whereas HEK293 cells and COS7 cells are not. These results strongly suggest that Ad4BP/SF-1 has an essential role in activating Cyp19a1 gene transcription. It is well documented in mammals that the behavior of the transcription factors is largely dependent on cell type/line (27, 31, 32), and the difference in the nature of the cell line used here also could be another reason for the contradictory results.
Mammalian two-hybrid and pull-down assays were performed to procure more substantial proof of the interaction between Foxl2 and Ad4BP/SF-1. In the mammalian two-hybrid assays, tFoxl2 and its mutants M1 and M3–M5 showed interaction with Ad4BP/SF-1, whereas M2 mutant without the FH domain failed to show any interaction. This was further verified by the pull-down assays. Unlike tFoxl2 and M3–M5, M2 could not pull down Ad4BP/SF-1 or Ad4BP/SF-1-LBD, indicating that the FH domain of Foxl2 and LBD of Ad4BP/SF-1 were instrumental in effecting the interaction between them. In the luciferase assay with single mutation of Foxl2 and Ad4BP/SF-1, M2 again was unable to enhance the Ad4BP/SF-1-mediated promoter activity; Ad4BP/SF-1-LBD, the mutant of Ad4BP/SF-1 without the DNA-binding domain, lost the ability to transactivate the promoter, indicating that the binding of Ad4BP/SF-1 to Cyp19a1 is essential. In the case of the Ad4BP/SF-1 and Foxl2 double mutations (Ad4BP/SF-1-LBD plus Foxl2 M1–M3), the luciferase activity was found to be more or less the same as the control. Collectively, these findings prompt us to postulate that the regulation of Cyp19a1 by Ad4BP/SF-1 and Foxl2 must be accomplished through a physical interaction between them in vivo. The spatial and temporal colocalization of these three factors in the tilapia ovary has clarified this idea further.
The three genes discussed in the present study are quite conserved throughout the vertebrate genomes. Hence, the present investigation checked whether mammalian Foxl2 also acts in a similar fashion to that of the fish Foxl2 to regulate the transcription of Cyp19. It was found that comparatively lower doses of mouse Foxl2 could enhance the Ad4BP/SF-1-activated Cyp19 transcription (supplemental Fig. 3 published as supplemental data on The Endocrine Society’s Journals Online web site at http://mend.endojornals.org). As already mentioned, mouse Foxl2 did show interaction with mouse Ad4BP/SF-1 in the pull-down assays also. In addition, mouse Foxl2 could elevate the levels of luciferase activity when TM3 cells were cotransfected with mouse Foxl2 and tilapia Cyp19a1 (data not shown). It is relevant to state here that the consensus sequence ATAAAACA, to which FH factors usually bind, was seen in the promoter regions of mouse and human Cyp19 PII, at −147/−139 bp and −150/−142 bp, respectively, and moreover, there are only three amino acid differences between mouse and tilapia Foxl2 protein in FH domain sequences (18). This could very well explain why mouse Foxl2 could bind to the tilapia probe.
Foxl2 Is the Key Factor in Controlling Female-Specific Aromatase Gene Expression and Estrogen Production in Vivo
As already mentioned, the orphan nuclear receptor, Ad4BP/SF-1, has been shown to be an important regulator of steroidogenic P450s (33–39). Binding of the Ad4BP/SF-1 protein to specific target sequences in the ovary-specific Cyp19 gene promoter has been implicated in the regulation of cAMP-induced P450arom expression in the human and rat ovary (32, 33, 36). A number of studies have suggested a potential role for Ad4BP/SF-1 in the transcriptional regulation of Cyp19a1 in fish also, including medaka and tilapia (7–10). More recent studies (17, 19, 20, 26, 27) have implicated the involvement of Foxl2 in the regulation of Cyp19/Cyp19a1 promoters in vertebrates. Because estrogen is suggested as the natural inducer of early ovarian differentiation in tilapia, it was envisaged that Foxl2 might regulate Cyp19a1 in a sex-specific manner, in association with Ad4BP/SF-1, as suggested by our in vitro data.
Results from the present investigation have corroborated the earlier findings that Foxl2 and Cyp19a1 are colocalized spatially and temporally in the female (19, 20, 26, 27), making them the earliest known markers for ovarian sex differentiation (12). In contrast, Ad4BP/SF-1 did not show a distinguishable sexual dimorphism in its expression pattern until 20 dah, when morphological sex differentiation is initiated in tilapia. Moreover, Foxl2 and Cyp19a1 were found to be down-regulated in the sex-reversed XX fish and up-regulated in the sex-reversed XY fish, whereas Ad4BP/SF-1 did not show any changes in its expression pattern in either situation. This prompted us to examine the action of Foxl2 in vivo with the help of transgenic experiments because it had both sexual dimorphic expression pattern and the ability to enhance the transcription of Cyp19a1 either independently or synergistically with Ad4BP/SF-1.
As expected from the in vitro studies, the CMV promoter-directed overexpression of Foxl2 resulted in the up-regulation of Cyp19a1 in the XY fish. The morphological appearance of this testis was like an ovary, and the sections of this gonad revealed degeneration of the testicular tissue in the periphery. The inner areas of the testis were more or less maintained, but the spermatogenic cysts, as well as the interstitial cells between them, showed enlargement as if the cells had proliferated more. This raises the question whether Foxl2 has any direct or indirect role in cell proliferation, a new area that warrants further research. This could also be due to the sharp increase in the levels of E2, which were equivalent to that of the normal female. The situation was quite tantalizing, but overexpression of tFoxl2 never caused complete sex reversal. It was found that the expression of Foxl2 was mostly mosaic in these transgenic gonads, and this might explain why there was no sex reversal in these fish. Moreover, these fish had surprisingly high levels of 11-KT also, which might have aided the inner area of the testis to maintain its phenotypic properties. It is noteworthy that degeneration had occurred invariably in the dorsal periphery region adjacent to the blood vessel. Only a few spermatogonia remained in the degenerated area.
On the other hand, expression of the M3 in the XX gonad, which showed varying degrees of gonad morphologies, could succeed in reversing the sex completely. Some of the gonads failed to form an ovarian cavity, which is usually formed around 2 months after hatching, and yet some other individuals had gonads with open ovarian cavity even at 5 months after hatching. The expression of aromatase was diminished, rendering them incapable of producing estrogen, which is required for normal gonadal development (4, 40). Well in accord with these data, estrogen was found to be reduced to significantly lower levels in these fish compared with the control females, proving the role of Foxl2 in the production of estrogen through the regulation of aromatase. This result seems to indicate that M3 had a dominant-negative effect on tFoxl2 in vivo, which was also proved in vitro (supplemental Fig. 2). Earlier, it was reported that the short deletion mutants of Foxl2 have a dominant-negative effect (13, 24). However, the promoter assay demonstrated that M3 could suppress even the Ad4BP/SF-1-mediated promoter activity (Fig. 8). Even though the mechanism of this suppression is not clear as of now, this prompts us to assume that the effect of transgenic overexpression of M3 in vivo might not only have resulted from its dominant-negative effect, but also from a suppression of Ad4BP/SF-1-mediated aromatase transcription. Hence, it is proved beyond a doubt that Ad4BP/SF-1-mediated activation of Cyp19a1 is not sufficient for the differentiation of the bipotential gonad into ovary and that Foxl2 has a role in the regulation of Cyp19a1. In other words, in the absence of the presumptive male sex-determining gene, Foxl2 enhances the Ad4BP/SF-1 mediated Cyp19a1 transcription activation, so as to provide the appropriate hormonal milieu required for ovarian development in tilapia. Meanwhile, another male-specific factor might suppress the Ad4BP/SF-1-mediated Cyp19a1 transcription in XY gonad so as to produce little or no estrogen, which favors the testicular differentiation.
Foxl2 Is a Possible General Regulator of Steroidogenesis
The modulated levels of 11-KT and E2 in the transgenic fish raise a pertinent question whether or not Foxl2 has a role in the overall steroidogenesis, because both these hormones are the end products of the entire steroidogenic pathway. It was quite surprising to note that 11-KT levels were increased with E2 by overexpression of tFoxl2 in XY fish, whereas M3 overexpression in the XX fish resulted in reduced E2 and 11-KT to levels that were even lower than those of the control male. This scenario suggests the absence of even the substrates for the synthesis of the above hormones and points to the involvement of Foxl2 in the modulation of other steroidogenic enzymes also. The preliminary data from the luciferase assays using the promoter regions of the genes encoding the major steroidogenic enzymes of both medaka and human also supported the idea that Foxl2 might regulate steroidogenesis in a more general manner (data not shown). Moreover, the tissue distribution pattern also has supported this notion, because Foxl2 expression was found in the brain-pituitary-gonad axis (18). Our gene expression studies indicated that Foxl2 and Ad4BP/SF-1 were colocalized in more cell types in the gonad than the aromatase-positive cells, a finding also reported in medaka (20). This indicates that Foxl2 and Ad4BP/SF-1 may have functions other than regulation of aromatase gene expression in the early gonad, including the possible regulation of other steroidogenic enzymes that provide substrates for aromatase. Low levels of expression of Foxl2 were detected in the adult testis of tilapia, suggesting the possible involvement of Foxl2 in the steroidogenesis of even the male gonad during adulthood.
The other transcriptional factors implicated in aromatase gene regulation, such as Wt1, Ad4BP/SF-1, Dax1, and LRH-1, were also found to be involved in the regulation of other steroidogenic enzymes (41–44). It was seen from this study that Foxl2 could also interact with LRH-1, which has a very high structural resemblance to Ad4BP/SF-1 (45), enhancing the Lrh-1-activated transcription of Cyp19a1 in the same way as Ad4BP/SF-1 (supplemental Fig. 4 published as supplemental data on The Endocrine Society’s Journals Online web site). Furthermore, some recent studies have shown Foxl2 as binding to the response elements in the promoters of GnRH receptor and StAR to regulate these genes in the human (13, 46). Therefore, Foxl2 might be regulating the ovarian differentiation not only through the transcriptional regulation of Cyp19a1, but also through the regulation of the entire steroidogenic pathway.
Foxl2 Has a Decisive Role in Early Ovarian Differentiation
A preponderance of evidence from a number of studies, including our own, requires us to revise the age-old concept that the female pathway of sex differentiation occurs by default. Although the male sex-determining gene of tilapia remains unknown, Foxl2 can be considered as the proovary, but antitestis gene because the disruption of Foxl2 could stimulate the XX tilapia to reverse its sex from female to male partially or completely, whereas the overexpression of Foxl2 in the XY fish resulted in degeneration of the testicular structure and a rise in estrogen levels. However, the absence of sex reversal in the latter group might preclude categorization of Foxl2 as an antitestis gene. Presumably, the inner pool of the intact spermatogenic cysts and other cells might disappear later because the spermatogonia had already undergone atresia, depriving the gonad of its capacity to replenish. On the other hand, studies showing the female to male sex reversal in the PIS goats (24, 25), blepharophimosis/ptosis/epicanthus inversus syndrome with or without premature ovarian failure in humans with Foxl2 mutations, and the induction of the genetic program for somatic testis determination in mice lacking Foxl2 strongly suggest a role of Foxl2 in ovarian sex differentiation (11, 21, 22). The present study has gone into greater depth to uncover the mechanism by which Foxl2 carries out its role during the early ovarian sex differentiation.
In conclusion, Foxl2 apparently has a decisive role in early ovarian sex differentiation, by activating the Cyp19a1 transcription either directly or in conjunction with Ad4BP/SF-1, to enhance estrogen production. The association between Foxl2 and Ad4BP/SF-1 thus provides the undifferentiated female gonad with its proper hormonal environment to differentiate as an ovary. The in vitro data, together with the in vivo data from the present study, provide concrete proof to support the above notion. Additionally, Foxl2 possesses the potential to act as an antitestis gene, by interfering with normal testicular differentiation, as seen from the XY fish with overexpression of tFoxl2.
MATERIALS AND METHODS
Animals
Nile tilapia, Oreochromis niloticus, were kept in recirculating fresh water tanks at 26 C before use. Monosex genetic female (XX), male (XY), sex-reversed female (XY, treated with E2) and male (XX, treated with MT) fry were obtained and maintained as described previously (3). All animal experiments conformed to the Guide for Care and Use of Laboratory Animals and were approved by the Institutional Committee of Laboratory Animal Experimentation (Institute for Basic Biology).
In Situ Hybridization and Immunohistochemistry
Tilapia gonads were dissected and fixed in 4% paraformaldehyde in 0.1 m phosphate buffer (pH 7.4) at 4 C overnight. After fixation, the tissues were embedded in paraffin, and cross-sections were cut at 5 μm. Digoxigenin-labeled sense and antisense probes were transcribed in vitro from linearized pGEM-T easy-Foxl2 (AY554172), -Ad4BP/SF-1 (AB060814), and -Cyp19a1 (U72071) cDNAs using a RNA labeling kit (Roche Diagnostics, GmbH, Indianapolis, IN). In situ hybridization was performed as described previously (47). For immunohistochemistry, the dissected gonads were fixed in Bouin’s solution at room temperature overnight. Immunostaining of aromatase, Foxl2, vasa, SCC (cholesterol side chain cleavage enzyme) etc., was performed as described previously (3). Some of the immunostained sections were counterstained with hematoxylin.
Plasmid Constructs and Site-Directed Mutagenesis
Tilapia Cyp19a1 gene promoter fragments (−2346 bp, blunt ligation into pGL3 SmaI site) were generated by PCR and subcloned into the pGL3-basic vector (Promega Corp., Madison, WI).
The tilapia Cyp19a1 5′-flanking region was cut with restriction enzymes and inserted into pGL3-basic vector to generate the deletion mutants. In case of no restriction site, the cutting site was introduced with the help of gene-specific primers. Mutant without the −438 /−119-bp promoter fragments was constructed by cutting with EcoT22I and PstI. Luciferase plasmid bearing a mutation in the −557/−524-bp Foxl2 binding motif was constructed by PCR-mediated mutagenesis using primers containing the mutations (see EMSAs).
The transcription factors Ad4BP/SF-1 and Foxl2 were subcloned for expression into pcDNA3.1 (Invitrogen, Carlsbad, CA) from the original clones, using gene-specific open reading frame (ORF) primers. The constructs with the deletion mutations on the 5′- and 3′-regions of Foxl2 were prepared by designing gene-specific primers at the desired positions. The mutants without the FH domain were generated by introducing BamH1 cutting site before and after FH domain. Ad4BP/SF-1-LBD (134–486 amino acids) was amplified using gene-specific primers. All mutants were cloned into the pcDNA 3.1 expression vector.
Plasmids used in transfection experiments were purified using either QIAfilter Plasmid Midi Kit or QIAprep Spin Miniprep Kit (QIAGEN Sciences, Boston, MA). The constructs and mutations and the orientation of the inserts were confirmed by direct sequencing.
Cell Culture, Transient Transfections, and Luciferase Assays
HEK293 cells and the mouse testicular cell line TM3, established from Leydig cells, were grown in DMEM (Sigma Chemical Co., St. Louis, MO) supplemented with 10% fetal bovine serum (JRH Biosciences, Lenexa, KS) and 1× penicillin-streptomycin-glutamine (Invitrogen) with 5% CO2 at 37 C. HEK293 cells and TM3 cells were transfected using Lipofectamine (Invitrogen) with the following plasmids: 1) 0.5 μg of normal or deletion constructs of Cyp19a1 promoter cloned into pGL3-basic luciferase reporter plasmids; 2) 0.05 μg-0.5 μg of pcDNA3.1 expression plasmid (Invitrogen), containing the cDNAs encoding Ad4BP/SF-1 (or LBD) and Foxl2 (or Foxl2 deletion mutants M1-M5); and 3) pRL-TK (Promega Corp.), 100 ng/well, Renilla luciferase employed as an internal control for transfection efficiency. The day before transfection, cells were seeded into 24-well plates. At the time of transfection, HEK293 cells and TM3 cells were 95% and 65% confluent, respectively. The transfection solution was made of 100 μl of Opti-MEM I reduced serum medium containing precomplexed DNA, and 2 μl of Lipofectamine reagent. Cells were washed in PBS 48 h after transfection and lysed in 100 μl luciferase lysis buffer. Firefly luciferase and Renilla luciferase readings were obtained using the Dual-Luciferase Reporter Assay System (Promega) and LUMAT LB9507 luminometer (Berthold Technologies GmbH & Co. KG, Bad Wildbad, Germany).
EMSAs
EMSAs were basically performed as described previously (48). The TNT T7 Coupled Reticulocyte Lysate System (Promega) was used to generate in vitro-translated tilapia and mouse recombinant Foxl2 proteins and mutants M2, M3, and M4. Tilapia Cyp19a1 oligonucleotide probes were designed based on the sequence between −557/-524 (sense 5′-TGGCTCTAATTAAAACAAATAGCTTTGATTTTGAA-3′ (GenBank accession no. AB089924), and for mutated probes the putative Foxl2 binding consensus sequence was changed (sense 5′-TGGCTCTAATTAAActcccgcGCTTTGATTTTGAA-3′; the lowercase nucleotides indicate the mutations). The annealed oligonucleotides were labeled with α-32P dCTP by end fill in from both ends using Klenow fragment (Takara, Otsu, Shiga, Japan). The unincorporated α-32P dCTP label was removed using a Sephadex-25 column. Protein-DNA binding reactions were performed using 3 μl protein in a 20 μl volume of binding buffer, the radiolabeled probe (10,000 cpm), and 1 μg of polydeoxyinosinic deoxycytidylic acid. After 15 min incubation on ice, 2 μl of loading dye was added, and the samples were fractionated on a 5% polyacrylamide gel at 130 V for 1.5 h. The gels were dried and exposed to a BAS-III imaging plate, and the hybridization signals were analyzed by a BAS 2000 Bio-Image Analyzer (Fuji Film Co. Ltd., Tokyo, Japan). Competition experiments were performed in the presence of 10- to 100-fold molar excess of unlabeled probes that were added 15 min before the labeled probe.
Mammalian Two-Hybrid Assays
Mammalian two-hybrid assays was performed using the CheckMate Mammalian Two-Hybridization System (Promega) according to the manufacturer’s instruction. Briefly, tilapia Ad4BP/SF-1, FOXL2, and its deletion mutants M1–M5 were cloned into the pBIND and the pACT vectors to generate fusion proteins with the DNA-binding domain of GAL4 and the activation domain of VP16, respectively. The pBIND vector expressing the Renilla reniformis luciferase was used to normalize the transfection efficiency. The pGAL4 and VP16 fusion constructs were transfected along with pG5 luc vector into HEK293 cells. The cells were lysed 2 d after transfection, and the amount of Renilla luciferase and firefly luciferase were quantitated using the Dual-Luciferase Reporter Assay System (Promega). Results are presented as the mean ± sd of data from triplicate replicates. Interaction between the two test proteins, as GAL4 and VP16 fusion constructs, results in an increase in firefly luciferase expression over the negative controls.
In Vitro Pull-Down Assays
Ad4BP/SF-1, Ad4BP/SF-1-LBD, Lrh-1, Foxl2 ORFs, and Foxl2 deletion mutants M2–M5 were cloned into the pGADT7 and pGBKT7 vectors. c-Myc- and HA-fusion proteins were obtained by in vitro translation with TNT-T7 Coupled Reticulocyte Lysate Systems (Promega), and labeled by FluroTect GreenLys in vitro Translation Labeling System (Promega). Here, the proteins were labeled at lysine residues with a green-fluorescent fluorophore. Coimmunoprecipitations were performed using the MATCHMAKER Co-IP Kit (BD Biosciences CLONTECH, Palo Alto, CA) according to the manufacturer’s instructions. The coimmunoprecipitated complexes were analyzed by SDS-PAGE, followed by scanning and visualization using laser fluoroimager (Typhoon 9400 Variable Mode Imager; Amersham Biosciences, Piscataway, NJ). Western blot analyses of the immunoprecipitated materials were performed with anti-HA or anti-c-Myc antibody to confirm the scanned image.
Overexpression of tFoxl2 and Its Dominant-Negative Mutant by Transgenesis
cDNAs for Foxl2 and its dominant-negative mutant were subcloned into the multiple cloning sites downstream of the CMV promoter of the pIRES-hrGFP-1a vector (Stratagene, La Jolla, CA). In vivo transgenic overexpression of Foxl2 in XY fish and its dominant-negative mutant in XX fish were carried out by injection of these GFP constructs into the fertilized eggs. The gonads of the injected fish were examined by monitoring the GFP signal, and later these gonads were subjected to both histological and immunohistochemical analyses after 2–6 months of injection, using antibodies against hrGFP, aromatase, SCC, and vasa. Blood samples were collected from the caudal veins of the 6-month-old transgenic as well as control fish. Serum E2 and 11-KT (the key androgenic steroid found in tilapia) levels were measured using the Estradiol EIA Kit and 11-ketotestosterone EIA Kit (Cayman Chemical Co., Ann Arbor, MI). Sample purification and assays were performed according to the manufacturer’s instructions.
Acknowledgments
This work was supported in part by Grant-in-Aid for Scientific Research from the Solution-Oriented Research for Science and Technology Research Project of Japan Science and Technology Corporation, the Ministry of Education, Science, Sports and Culture of Japan, and Environmental Endocrine Disruptor Studies from the Ministry of the Environment.
Disclosure Statement: The authors have nothing to disclose.
Abbreviations
- Ad4BP/SF-1
Ad4 binding protein/steroidogenic factor 1;
- CMV
cytomegalovirus;
- dah
days after hatching;
- E2
17β-estradiol;
- FH
forkhead;
- HA
hemagglutinin;
- HEK
human embryonic kidney;
- hrGFP
humanized recombinant green fluorescent protein;
- 11-KT
11-ketotestosterone;
- LBD
ligand-binding domain;
- MT
17α-methyltestostorone;
- ORF
open reading frame;
- PIS
polled intersex syndrome;
- SCC
cholesterol side chain cleavage enzyme;
- StAR
steroidogenic acute regulatory protein.
