Roles of anti-Müllerian hormone and its duplicates in sex determination and germ cell proliferation of Nile tilapia

Abstract

Duplicates of amh are crucial for fish sex determination and differentiation. In Nile tilapia, unlike in other teleosts, amh is located on X chromosome. The Y chromosome amh (amhΔ-y) is mutated with 5 bp insertion and 233 bp deletion in the coding sequence, and tandem duplicate of amh on Y chromosome (amhy) has been identified as the sex determiner. However, the expression of amh, amhΔ-y, and amhy, their roles in germ cell proliferation and the molecular mechanism of how amhy determines sex is still unclear. In this study, expression and functions of each duplicate were analyzed. Sex reversal occurred only when amhy was mutated as revealed by single, double, and triple mutation of the 3 duplicates in XY fish. Homozygous mutation of amhy in YY fish also resulted in sex reversal. Earlier and higher expression of amhy/Amhy was observed in XY gonads compared with amh/Amh during sex determination. Amhy could inhibit the transcription of cyp19a1a through Amhr2/Smads signaling. Loss of cyp19a1a rescued the sex reversal phenotype in XY fish with amhy mutation. Interestingly, mutation of both amh and amhy in XY fish or homozygous mutation of amhy in YY fish resulted in infertile females with significantly increased germ cell proliferation. Taken together, these results indicated that up-regulation of amhy during the critical period of sex determination makes it the sex-determining gene, and it functions through repressing cyp19a1a expression via Amhr2/Smads signaling pathway. Amh retained its function in controlling germ cell proliferation as reported in other teleosts, while amhΔ-y was nonfunctionalized.

Introduction

Sex determination and differentiation are inherently fascinating topics to reproductive scientists. To date, more than 20 different master sex-determining (SD) genes have been identified on different sex chromosomes of vertebrates. Interestingly, 13 of these genes (amh, amhr2, bmpr1b, gsdf, and gdf6) belong to the transforming growth factors (TGF-β) signaling pathway (Pan et al. 2021). Fish is the largest group of vertebrates which exhibit a striking variety of SD gene (Schultheis et al. 2009; Gammerdinger and Kocher 2018; Li and Gui 2018; Kottler et al. 2020). Amh has been reported to be the SD gene in several fish species, including the Patagonian pejerrey (Odontesthes hatcheri) (Hattori et al. 2012), Nile tilapia (Oreochromis niloticus) (Li et al. 2015), lingcod (Ophiodon elongatus) (Rondeau et al. 2016), Northern pike (Esox lucius) (Pan et al. 2019), and threespine stickleback (Gasterosteus aculeatus) (Peichel et al. 2020). Besides the ligand, amhr2 has also been reported to be the SD gene of the tiger puffer fish (Takifugu rubripes) (Kamiya et al. 2012) and the yellow perch (Perca flavescens) (Feron et al. 2020). It is worth noting that all of the above mentioned TGF-β pathway molecules with the exception of amhr2 in the puffer fish arose via gene duplication. These results indicated that Amh signaling is crucial for fish sex determination. However, the following issue remains unknown: How are the SD functions of Amh-signaling molecules accomplished during fish gonadal induction? Or how does that gonadal Amh regulatory network interact with the canonical gonadal gene regulatory network? It is also interesting to known whether the SD duplicated from amh still retains the original function of amh.

In Nile tilapia, a series of studies revealed the association of LG23 with sex seems to be highly conserved in many laboratory strains, wild and farmed populations (Baroiller and D'Cotta 2016; Cáceres et al. 2019; Sissao et al. 2019; Curzon et al. 2020; Triay et al. 2020; Taslima et al. 2021). A SD region which hosts the amh (it should be amhΔ-y or amhy, however, at that time, it was unclear whether it was amhΔ-y or amhy) gene, was identified through quantitative trait loci (QTL) mapping (Shirak et al. 2006; Eshel et al. 2011, 2012). Later, a Y-linked amh on LG23 with 233 bp deletion in exon 7, finally named as amhΔ-y by our group, was identified (Eshel et al. 2014; Li et al. 2015). In addition, the amhΔ-y with a 5 bp insertion in exon 6 and 233 bp deletion in exon 7 results in a truncated protein lacking the TGF-β binding domain as revealed by Western blot, and several insertions and deletions in the promoter region of amhΔ-y were detected (Li et al. 2015). A duplicate of amh with a missense single nucleotide polymorphism (SNP) in exon 2 and a 5,608 bp promoter loss, which is located immediately downstream of amhΔ-y and designated as amhy, was identified as the SD gene in a Japan strain of Nile tilapia (Li et al. 2015). Y chromosome (LG23) harboring amhΔ-y and amhy with 5,608 bp deletion in the promoter region but without the missense SNP in exon 2 (same allele of amhy in Japan strain) was also observed in the GIFT strain (Taslima et al. 2021). Recently, the genome assembly from individuals of 2 wild Nile tilapia populations revealed that the tandem duplication which containing amhΔ-y and amhy was a duplication of 21.5 kb located on the Y chromosome (LG23). The breakpoints on the X chromosome that led to the Y chromosome duplication are between 2 genes (oaz1 and dot1l) that flank amh. The oaz1 is found upstream of amh and dot1l is found downstream. The duplication occurs after the third exon of the oaz1 gene and ends before the last exon of dot1l. The breakpoints take place within introns, and consequently, the duplication does not include the whole coding sequence for oaz1 or dot1l. The insertion of this duplication on the Y chromosome could have taken place downstream within the dot1l gene (Triay et al. 2020). Evidence obtained from gene knockout and overexpression study from our group indicated that amhy, rather than amhΔ-y, was responsible for male sex determination in Nile tilapia (Li et al. 2015). Despite this, fine mapping using whole-genome sequencing demonstrated amh as a major gene for sex determination in farmed Nile tilapia (Cáceres et al. 2019), and more recently, amhΔ-y was suggested to be associated with genetic sex determination as the insertions, deletions and SNPs in amhΔ-y were conserved among different Nile tilapia strains (Curzon et al. 2020). Even though these results all point to the amhΔ-y-amhy region as the SD locus of Nile tilapia, it is controversial which of the 2 Y-linked genes, amhy or amhΔ-y, is the SD gene. In addition, whether the variations in the promoter or the SNP (C/T) in exon 2 of amhy make it a SD gene remains to be elucidated.

Previous studies showed that amh was strongly expressed in differentiating gonads at 19 days post fertilization (dpf, ∼15 dah, days after hatching) in Nile tilapia (Ijiri et al. 2008; Poonlaphdecha et al. 2011). Later, dimorphic expression of amh gene was detected in embryos from 5 dpf, and increased sharply in male gonads until 35 dpf, indicating that this gene might be involved in sexual differentiation (Eshel et al. 2014). After the identification of the 3 amh copies, all of them were found to be expressed in XY gonads from 9 dpf (Li et al. 2015). These expression studies indicated that all the 3 amh copies expressed in the gonads during sex differentiation while the exact expression profile and cell type of each copy during early sex determination remains to be investigated.

Amh induces the regression of Müllerian duct in male mammals (Durlinger et al. 1999, 2002; Josso et al. 2001; Alvaro Mercadal et al. 2015). Although teleosts lack Müllerian ducts, they have amh gene which is proved to be involved in germ cell proliferation, gametogenesis and fertility (Pfennig et al. 2015; Lin et al. 2017; Liu et al. 2020; Zhang et al. 2020a, 2020b). Amh is produced in male Sertoli cells and female granulosa cells and inhibits germ cell proliferation and differentiation in both sexes (Pfennig et al. 2015). As mentioned above, 3 amh duplicates, amh, amhΔ-y and amhy, exist in Nile tilapia, it is interesting to known the function of each copy in germ cell proliferation, especially in YY fish which harbor 2 copies of amhy and amhΔ-y but lack amh.

In mammals, Amh binds to its type II receptor Amhr2, and then induces the phosphorylation of Smad1/5/8 (R-Smad), the phosphorylated R-Smad associates with co-Smad (Smad4), the activated R-Smad/co-Smad complex translocate to the nucleus and regulate the transcription of target genes (Gouédard et al. 2000; Orvis et al. 2008). Amh has a negative and inhibitory role in the expression of aromatase Cyp19a1 (Sacchi et al. 2016). Cyp19a1a is the key gene essential for estrogen synthesis and ovarian differentiation in teleosts (Guiguen et al. 2010; Wu et al. 2010; Nakamoto et al. 2018; Li et al. 2019). Knockout of male pathway genes or overexpression of female pathway genes up-regulate cyp19a1a expression and increase estrogen level so as to promote ovarian differentiation in fish (Wang et al. 2007, 2010; Chakraborty et al. 2016; Wu et al. 2020). Dysfunctional Amh signaling is responsible for the up-regulated cyp19a1a expression and sex reversal in the medaka (Oryzias latipes) hotei (amhr2) mutants (Nakamura et al. 2012). In Patagonian pejerrey, knockdown amhy in XY embryos resulted in the up-regulation of foxl2 and cyp19a1a mRNAs and the development of ovaries (Hattori et al. 2012). In salmonids, the SD gene sdy prevents the activation of cyp19a1a promoter in cooperation with foxl2 (Bertho et al. 2018). Therefore, it is interesting to know whether Amh/AmhΔ-y/Amhy determines male sex by repressing the expression of cyp19a1a.

In this study, we analyzed the promoter differences and expression profiles of amh, amhΔ-y and amhy during sex determination and differentiation and established their single, double and triple mutants in XY and YY fish, and amhy and cyp19a1a double mutants in XY fish of Nile tilapia. We provided multiple evidence to confirm that amhy is the SD gene. Amhy inhibited the transcription of cyp19a1a through Amhr2/Smads signaling pathway to determine the male sex. The X-linked amh was essential for normal germ cell proliferation. In the absence of amh, amhy but not amhΔ-y could compensate for the function of amh.

Materials and methods

Tilapia maintenance

The founder strain of the Nile tilapia (O. niloticus) was obtained from Japan (introduced from Egypt in Africa, 1970s), and kept in recirculating freshwater tanks at 26°C under natural photoperiod. All XY progenies were obtained by mating the YY males with the XX wild type (WT) females, and all XX progenies were obtained by mating the XX phenotypic males with the XX-WT (Sun et al. 2014b). Animal experiments were conducted in accordance with the regulations of the Guide for Care and Use of Laboratory Animals and were approved by the Committee of Laboratory Animal Experimentation at Southwest University (No. IACUC-20181015-12, October 15, 2018).

Establishment of amh/amhΔ-y/amhy mutant lines

XY mutants with amhΔ-y, amhy single, amhΔ-y, and amhy double mutation were obtained in our previous study (Li et al. 2015). In this study, new F0 mutants with different mutation types from target sites 1 and 2 were selected to establish mutant lines. The target sites were located in exon 1 and 2 (target sites 1 and 2) of amh/amhΔ-y/amhy. Due to the high homology of amh, amhΔ-y, and amhy, CRISPR/Cas9 disrupted the 3 genes at both target sites. XY fish with amh single mutation (amh⁻) and XX fish with amh heterozygous mutation (amh^+/−) were obtained by mating F0 XX female founders with XY-WT males. XY fish with amhΔ-y (amhΔ-y⁻) and amhy single (amhy⁻), and amhΔ-y and amhy double mutation (amhΔ-y⁻; amhy⁻) were obtained by mating F0 XY male founders with XX-WT females. Restriction enzyme digestion and Sanger sequencing were performed to confirm the insertions or deletions leading to frame-shift mutations. These mutations resulted in truncated Amh, AmhΔ-y, and Amhy proteins without the TGF-β domain. XY fish with amh and amhy double mutation (amh⁻; amhy⁻) were obtained by mating XY-amhy⁻ phenotypic females with XY-amh⁻ males. XY fish with amh and amhΔ-y double mutation (amh⁻; amhΔ-y⁻) were obtained by mating XY-amhΔ-y⁻ males with XX-amh^+/− females. XY fish with amh, amhΔ-y and amhy triple mutation were obtained by mating XY-amhΔ-y⁻; amhy⁻ phenotypic females with XY-amh⁻ males. YY fish with amhy heterozygous mutation (YY-amhy^+/−) were obtained by mating XY-amhy⁻ phenotypic females with XY-WT males. YY fish with amhy homozygous mutation (YY-amhy^−/−) were obtained by mating XY-amhy⁻ phenotypic females with YY-amhy^+/− males. XY-amhy⁻ phenotypic females were mated with XX-cyp19a1a^−/− phenotypic males (Zhang et al. 2017) to obtain XY-amhy⁻; cyp19a1a^+/− females. XY-amhy⁻; cyp19a1a^−/− fish were obtained by cross XY-amhy⁻; cyp19a1a^+/− females with XX-cyp19a1a^−/− phenotypic males.

Masculinization and feminization of the mutants

The XY fish with amhy single or amh and amhy double mutation were masculinized by administration of letrozole, which is the third-generation aromatase inhibitor, and is widely used for suppressing estrogen production (Liao et al. 2014). Letrozole treatment was performed to induce female to male sex reversal in this study. The XY fish with amh mutation and YY fish with amhy heterozygous mutation were feminized by estradiol-17β (E2) treatment. Letrozole or E2 (Sigma-Aldrich, MO, USA) was dissolved in the 95% ethanol for 4 mg/ml concentration, then the fish diet (Shengsuo Feed Technology, Shandong, China) was sprayed with the ethanol containing the letrozole or E2 to a concentration of 200 μg/g diet. Letrozole and E2 administration were performed from 10 dpf, and lasted for 30 d. And then, fish were fed with normal commercial diet. The gonadal phenotypes were determined by histological examination. All treatments were performed in the thermostatic aquaria, and the water temperature was controlled at 26°C.

Genotyping, sampling and histological examination

XY Fish with amh/amhΔ-y/amhy single, double or triple mutation were identified by PCR amplification the regions containing the mutations and followed by restriction enzyme digestion, and verified by Sanger sequencing. The genetic sex of the mutant fish was determined by sex-linked marker (Marker 5) as described previously (Sun et al. 2014b). The body weight and gonad weight were measured before fish were sampled for histological analyses. The gonads were fixed in Bouin’s solution for 24 h at room temperature, and then dehydrated and embedded in paraffin. Tissue blocks were sectioned at 5 μm for H&E staining.

Follicle counting

Follicles from the median cross section of ovaries (n = 6) of each genotype (XX-WT, XY-amhy⁻, XY-amhΔ-y⁻; amhy⁻, XY-amh⁻; amhy⁻, XY-amh⁻; amhΔ-y⁻; amhy⁻; and YY-amhy^−/− fish) were counted for statistical analyses. The histological classification of the oocytes was performed as described previously (Coward and Bromage 1998; Liu et al. 2020).

Measurement of serum E2

Blood was collected from the caudal vein from 6 fish of each genotype, and was stored at 4°C overnight. Blood samples were centrifuged at 8,000 g for 5 min at 4°C. The serum was collected and stored at –80°C until use. Concentration (ng/ml) of E2, the main bioactive estrogen in fish, was measured using the Enzyme Immunoassay (EIA) Kit (Cayman, Ann Arbor, USA) according to the manufacturer’s instructions.

Fecundity and fertilization rate assessment

Fecundity and fertilization rate assessment were performed as described previously (Liu et al. 2020). The fecundity of XY and YY phenotypic females was examined as follows: matured phenotypic females (XY-amhy⁻, XY-amh⁻; amhy⁻, XY-amh⁻; amhΔ-y⁻; amhy⁻; and YY-amhy^−/− mutants, 300 dpf) were paired with XY-WT adult males. XX-WT matured females (300 dpf) were used as control. The spawned eggs were collected and counted, and then incubated in re-circulating systems at 26°C for the analysis of fertilization rate at 1 dpf. The fertilization rates of XY males with amhΔ-y or amh mutation were assessed via mating the mutants with XX-WT females in a spawning pool. XY-WT males were used as control. Developing embryos were maintained in the re-circulating water system at 26°C. The fertilization rate was calculated at 1 dpf using a stereomicroscope. Five independent crosses were analyzed for fecundity and fertilization rate assessment of each genotype.

Immunohistochemical analysis

Amh, Vasa (germ cell marker), Cyp11c1 (11β-hydroxylase, the key enzyme for 11-KT synthesis), Cyp19a1a and 42Sp50 (Oocytes marker) antibodies were prepared by our laboratory as reported previously (Sun et al. 2014a; Li et al. 2015; Chen et al. 2017; Zhang et al. 2017; Zheng et al. 2020). Primary antibodies against Amh, Vasa, PCNA (marker for cell proliferation, Cell Signaling), Cyp11c1, Cyp19a1a, and 42Sp50 were diluted at ratio of 1:500, 1:1,000, 1:500, 1:500, 1:2,000, and 1:500 for use. HRP-conjugated goat anti-rabbit IgG (1:1,000, Proteintech, Guangzhou, China) or Alexa Fluor 488-, 594-conjugated secondary antibodies (1:1,000; Thermo Fisher Scientific, Waltham, USA) were used to detect the primary antibody. Diaminobenzidine tetrachloride (DAB) was applied for the color reaction of PCNA. The nuclei were counterstained with DAPI (Invitrogen, Carlsbad, USA). Images were captured under a Laser confocal microscope (Olympus FV3000, Tokyo, Japan).

RT-PCR and Real-time PCR

The head, tail and viscera of the XY larvae were removed at 4, 6, 8, 10, and 14 dpf, and only the trunk containing gonads was collected for RNA extraction. Gonads of the XY fish at 30, 90, and 180 dpf were dissected for RNA extraction. Total RNA (1.0 μg) was extracted and reverse transcribed using PrimeScript RT Master Mix Perfect real-time PCR Kit (Takara, Dalian, China) according to the manufacturer’s instructions. RT-PCR Products amplified by F2/R1 (Fig. 2a) were sub-cloned into the pMD-19T vector, and positive clones were sequenced to calculate the proportion of amh and amhy. The primers used for Real-time PCR were validated. Real-time PCR was performed on an ABI7500 machine (Applied Biosystems, Waltham, USA), according to the protocol of SYBR1 Premix Ex TaqTM II (Takara, Dalian, China). The relative abundance of amh and amhy mixture in XY fish was evaluated using the formula R = 2^−ΔΔCt (Livak and Schmittgen, 2001). The reference gene β-actin was used to normalize the expression values. Primer sequences used for real-time PCR were listed in Supplementary Table 2.

Fig. 2.

Expression profile of amh/amhΔ-y/amhy in XY and YY fish. a) Schematic representation of the ORF of amh, amhΔ-y, and amhy. ORF of amh and amhy were the same except the SNP (C/T) in exon 2. Five bp (ATGTC) insertion in the exon 6 and 233 bp deletion in exon 7 were detected in ORF of amhΔ-y, and resulted in early termination of genetic coding. b) Relative mRNA expression level of amh and amhy in XY fish. Real-time PCR primers (F1/R1) could amplify amh and amhy but not amhΔ-y. Products amplified by F2/R1 primer were sequenced to distinguish the expression of amh and amhy by the difference of SNP (C/T) in exon 2. Relative mRNA expression of amh and amhy was calculated by dividing the real-time PCR results based on the proportion of C and T. c) Expression profile of Amh, AmhΔ-y, and Amhy in XX, XY, and YY fish during sex determination and differentiation by immunofluorescence. d) Expression of Amh, AmhΔ-y, and Amhy in the XY-amhΔ-y⁻ and XY-amhy⁻ and XY-amh⁻ mutants at 8 dpf by immunofluoresence. dpf, day post fertilization; Ins, Insertion; Del, Deletion. Differences of the relative mRNA level between amh and amhy were tested by Student's t-test, *P < 0.05; **P < 0.01.

Luciferase assay

Nile tilapia cyp19a1a promoter (–2,346 bp) in pGL3-basic vector and constructs of sf1 and foxl2 in pcDNA3.1 were prepared previously (Wang et al. 2007). Open reading frames of Nile tilapia amh, amhy, amhΔ-y, amhr2, smad1/5/8, and smad4a/b/c/d were sub-cloned into the pcDNA3.1 vector (Invitrogen, Carlsbad, USA). The constructs and the orientation of the inserts were confirmed by Sanger sequencing. Plasmids were purified using QIAfilter Plasmid Midi Kit (Qiagen, Hilden, Germany). All primer sequences used for constructing recombinant vectors were listed in Supplementary Table 2. Potential binding sites for transcription factors within the cyp19a1a promoters were predicted using the MatInspector program (http://www.genomatix.de). Cell culture, transient transfections and luciferase assays were performed as reported previously (Wang et al. 2007). Briefly, HEK293 cells (ATCC number: CRL-1573) were maintained in DMEM supplemented with 10% FBS, penicillin (100 IU/ml) and streptomycin (100 μg/ml) at 37°C in a humidified atmosphere containing 5% CO₂. HEK293 cells were transfected using Lipofectamine 2000 (Invitrogen, Carlsbad, USA) with the following plasmids: (1) 400 ng/well plasmids of cyp19a1a promoter cloned into pGL3-basic luciferase reporter; (2) 50–250 ng/well of pcDNA3.1 expression plasmids, containing the ORFs encoding Foxl2, Sf1, Smad1/5/8, Smad4a/b/c/d, Amhr2, Amh, AmhΔ-y, or Amhy; Empty pcDNA3.1 expression plasmid was used to supplement the total amount of plasmid in each well. (3) pRL-TK, 100 ng/well. Renilla luciferase was 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 were 50%–70% confluent. The transfection solution was made of 50 μl/well of DMEM containing precomplexed DNA, and 2 μl of Lipofectamine reagent. Cells were washed in PBS after 48 h 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, Madison, USA) and Luminoskan Ascent Luminometer (Thermo Fisher Scientific, Waltham, USA). Normalized luciferase activity was obtained by dividing the Firefly luciferase activity by the Renilla luciferase activity.

Electrophoretic mobility shift assay

Electrophoretic mobility shift assay (EMSA) experiments were performed as described previously (Dai et al. 2021). Briefly, biotin-labeled and unlabeled double-stranded DNA probes containing intact potential binding sites for Smad4b were prepared according to the manufacturer’s protocol of the Chemiluminescent EMSA Kit (Beyotime, Shanghai, China). The related primers were listed in Supplementary Table 2. Nuclear proteins were extracted from HEK293 cells overexpressing the Nile tilapia smad4b gene with NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Fisher Scientific, Waltham, USA). Then, 3–5 μg nuclear protein was incubated with 500 nM biotin-labeled probes in 1× binding buffer for 10 min at room temperature. For the competition analysis, a 10-, 50-, 100-fold molar excess of the unlabeled probe (cold competitor) was added to the nuclear extracts prior to the addition of the labeled probe. Protein-DNA complexes were separated by electrophoresis in 4% polyacrylamide gels with 0.5× TBE at 100 V for 90 min, and then transferred to nylon membranes. The biotin-labeled DNA on the membrane was detected using the Chemiluminescent EMSA Kit (Beyotime, Shanghai, China).

Western blot

Western blot was performed following the protocol reported previously (Zhang et al. 2017). Total protein was extracted from HEK293 cells transfected with plasmids. The protein lysates were resolved by SDS/PAGE on 12% Tris·glycine gels followed by transfer to nitrocellulose membranes. Unspecific binding was blocked with 5% BSA in Tris-buffered saline with Tween-20 (TBST) [10 mM Tris (pH = 7.9); 150 mM NaCl; 0.1% Tween] for 1 h at room temperature. Incubation with P-Smad1/5/8 and Smad1 primary antibodies (Cell Signaling Technology, Beverly, USA) at a dilution of 1:500 was performed overnight at 4°C. After washing with TBST for 3 times, the membranes were incubated with HRP conjugated secondary antibody (1:1,000, Thermo Fisher Scientific, Waltham, USA) in blocking solution for 1 h. The abundance of α-Tubulin was examined as a loading control using rabbit anti-α-Tubulin antibody (Cell Signaling Technology, Beverly, USA) at a dilution of 1:1,000. Specific signal was detected with Pierce ECL Western Blotting Substrate (Thermo Fisher scientific, Waltham, USA) and was visualized on a Fusion FX7 (Vilber Lourmat, East Sussex, France).

Data analyses

All data were presented as the mean ± SD from at least 3 independent experiments. Statistical comparisons were made using Student’s t-test when comparing 2 groups. One-way ANOVA was performed for comparisons with more than 2 groups followed by Turkey test. Statistics analyses were performed using GraphPad Prism 8 (GraphPad Software, La Jolla, USA). In all analyses, a value of P < 0.05 was considered to be statistically significant.

Results

amhy is responsible for male sex determination

To confirm the phenotype of our previously published mutations, mutant lines in this study were newly generated. In this study, multiple new mutant lines, including amh (+2 bp or −5 bp), amhΔ-y (+2 bp or −13 bp) and amhy (−4 bp or −19 bp) single mutation, amh and amhΔ-y (amh +2 bp and amhΔ-y −13 bp) and amh and amhy (amh +2 bp and amhy −4 bp) double mutation, amhΔ-y and amhy (amhΔ-y −4 bp and amhy −19 bp) double mutation, amh, amhΔ-y and amhy (amh +2 bp, amhΔ-y −4 bp and amhy −19 bp) triple mutation in XY fish, and YY males with amhy heterozygous (YY-amhy^+/−, −4 bp) and homozygous (YY-amhy^−/−) mutation, were established in Nile tilapia (Supplementary Figure 1; Supplementary Table 1). Single mutation of amh (n = 31 males, 0 female) or amhΔ-y (n = 30 males, 0 female) and double mutation of amh and amhΔ-y (n = 15 males, 0 female) did not cause sex reversal in XY fish (Fig. 1, a–f). However, sex reversal occurred in the mutants with amhy single (n = 44 females, 0 male), amh and amhy double (n = 24 females, 0 male) and amh, amhΔ-y and amhy triple (n = 15 females, 0 male) mutation, indicating that sex reversal occurred only when amhy was mutated in XY fish (Fig. 1, g–j). Homozygous mutation of amhy (n = 26 females, 0 male) in YY males also resulted in male to female sex reversal, while heterozygous mutation of amhy (n = 28 males, 0 female) could not (Fig. 1, k–n). These genetic evidence strongly demonstrate that amhy was responsible for male sex determination in Nile tilapia.

Fig. 1.

Sexual phenotypes of XY and YY fish with amh/amhΔ-y/amhy mutation at 90 dpf. a,b) Schematic representation of amh/amhΔ-y/amhy genes position on the X- and Y-chromosome. The distance (Kb) between ORF (open reading frame) of two genes was indicated. c-n) Phenotypes of gonads from XY fish with amh, amhΔ-y, or amhy single mutation (amh⁻, amhΔ-y⁻, and amhy⁻), amh and amhΔ-y (amh⁻; amhΔ-y⁻), amh and amhy (amh⁻; amhy⁻), amhΔ-y and amhy (amhΔ-y⁻; amhy⁻) double mutation, amh, amhΔ-y and amhy (amh⁻; amhΔ-y⁻; amhy⁻) triple mutation, and YY fish with amhy heterozygous (YY-amhy^+/−) or homozygous (YY-amhy^−/−) mutation. Sg, spermatogonia; Sc, spermatocytes; Og, oogonia; Oc, oocytes.

The SNP (C/T) in exon 2 of amhy is polymorphism, while variations in the promoter are conserved in amhy from different populations

Three other populations derived from genetically improved farmed tilapia (GIFT, introduced from WorldFish Center in 2006) were collected from Guangdong (Gd), Yunnan (Yn), and Hubei (Hb) Province of China. Variations reported in the promoter and the coding sequence of amhy were also analyzed. The 5,608 bp deletion in the promoter of amhy was detected in the XY individuals from these 4 Nile tilapia populations (Supplementary Figure 2a). The SNP (C to T change) in exon 2 of amhy was variable among different tilapia populations (Supplementary Figure 2b). Sequence alignments analysis revealed that the 5,608 bp deletion and several other variations in the promoter were conserved (Supplementary Figure 3). The promoters of amh were variable among different Nile tilapia populations compared with that of amhΔ-y (Supplementary Figure 4). These data indicated that the SNP (C/T) in exon 2 is polymorphism, while variations in the amhy promoter are conserved from different Nile tilapia populations.

Expression of amhy is higher than amh in XY gonad during sex determination and differentiation

We analyzed the detailed mRNA expression profile of amh and amhy in XY fish by real-time PCR followed by cloning and sequencing the amplified bands to identify gene of origin. Real-time PCR primer (R1) were designed located in the 233 bp deletion region to exclude amhΔ-y, so the real-time PCR results were the combination of amh and amhy using primer F1/R1 (Fig. 2a). Reverse transcription-PCR (RT-PCR) products that containing the SNP (C/T) on exon 2 were amplified by primers F2/R1, and then sequenced to calculate the proportion of C/T (amh/amhy). The relative mRNA expression of separate amh or amhy was calculated by dividing the real-time PCR results based on the SNP (C/T) proportion (Supplementary Figure 5).

Similar expression level of amh and amhy were observed at 4 dpf. Increased expression of amhy was detected at 6, 8, and 10 dpf (the critical period of sex determination) compared with that at 4 dpf, and expression of amhy was significantly higher than that of amh at these time points. The expression of amhy decreased gradually and expression of amh increased gradually at 14, 30, 90, and 180 dpf (Fig. 2b). Immunofluorescence analysis was performed to examine the expression of Amh, AmhΔ-y, and Amhy in the XX, XY, and YY fish. The Amh antibody could recognize Amh, AmhΔ-y, and Amhy (Li et al. 2015). Signal of Amh, AmhΔ-y, and Amhy was detected in the Sertoli cells of the XY and YY fish, but was absent in the XX fish at 6 and 8 dpf, indicating that the signal was mainly from AmhΔ-y and Amhy (Fig. 2, c-i–vi). In the XX fish, weak signal of Amh was detected at 10 dpf, and the signal increased at 14, 18, and 30 dpf. In both the XY and YY fish, strong signal of Amh, AmhΔ-y, and Amhy was detected at 8 and 10 dpf. At 14, 18, and 30 dpf, strong signal was observed in the XY fish while only weak signal was detected in the YY male fish, indicating the increase of Amh expression and decrease of AmhΔ-y and Amhy expression at these time points (Fig. 2, c-vii–xviii). In order to distinguish the signal between Amh, AmhΔ-y, and Amhy, Immunofluorescence analysis was performed in XY mutants with amh, amhΔ-y or amhy mutation at 8 dpf. Extensive signal of Amh, AmhΔ-y, and Amhy was detected in the Sertoli cells of the XY-amhΔ-y⁻ and XY-amh⁻ fish, but weak signal was detected in the XY-amhy⁻, indicating that the strong signal in XY-WT testis was mainly from Amhy but not AmhΔ-y or Amh at 8 dpf (Fig. 2d). Expression profile analysis revealed that higher expression of Amhy was detected in the critical period of sex determination compared with that of Amh and AmhΔ-y in XY fish.

Transcription of cyp19a1a is repressed by Amhy signaling

In order to investigate the possible mechanism of amhy determining the sex of Nile tilapia, the target genes regulated by amhy were analyzed. In this study, we performed dual-luciferase reporter assays in human HEK293 cells to evaluate the effects of Amhy on the activity of cyp19a1a promoter. Effects of Smad1/5/8 and 4 Smad4 (Smad4a, Smad4b, Smad4c, and Smad4d) alone were tested to evaluate which of them had effects on the transcription activity of cyp19a1a. In the presence of both Sf1 and Foxl2, Smad1/5/8 and Smad4a/b/c/d alone all could significantly increase the transcriptional activity of cyp19a1a. Smad1 and Smad4b displayed the highest effects among the Smads tested and were selected for the subsequent experiments (Supplementary Figure 6a).

Smad1 and Smad4b enhanced the transcription activity of cyp19a1a in a dose dependent manner from 10 ng to 250 ng in the presence of Sf1 and Foxl2 (Fig. 3a). We next examined whether the cyp19a1a promoter responses to Amhy, Amh, and AmhΔ-y overexpression. In the presence of Smad1, Smad4b, Amhr2, Sf1, and Foxl2, the transcription of cyp19a1a was significantly decreased by Amhy and Amh but not AmhΔ-y in a dose dependent manner (Fig. 3b and Supplementary Figure 6, b and c). Left out one of the Amhr2/Smad1/Smad4b/Foxl2/Sf1, the cyp19a1a promoter did not respond to Amhy treatment. The significance of the Smad4b binding site in the cyp19a1a promoter was further confirmed by EMSA using probe and mutated probe designed according to the sequences between −325 and −355 bp of cyp19a1a promoter. Nucleoproteins, extracted from HEK293 cells overexpressing Smad4b, bound to the biotin-labeled oligonucleotides and resulted in the formation of a specific band of protein/DNA complex. Cold competitor (unlabeled oligonucleotides, 1–100 μmol), but not mutated cold competitor, could displace the above band in a dose dependent manner (Fig. 3c). We further examined whether repressed transcription of cyp19a1a by Amhy was mediated by Smad1/5/8 phosphorylation. Constructs containing smad1, amhr2, and amhy were cotransfected into HEK293 cells. Western blot analysis showed that the phosphorylation of Smad1/5/8 was increased in a dose dependent manner by overexpression of Amhy (Fig. 3d). Specific signal of P-Smad1/5/8 was detected in the XY-WT fish, but was absent in the XY fish with amhy mutation at 8 dpf (Fig. 3e). Inhibitor (LDN193189) of the Amh type I receptor Alk2/3/6 (Yu et al. 2008) was used to further confirm the phosphorylation of Smad1/5/8 was regulated by Amhy. Dual-luciferase reporter assay showed that administration of LDN193189 recovered the transcriptional activity of cyp19a1a promoter repressed by Amhy (Supplementary Figure 6d). Consistently, administration of LDN193189 into HEK293 cells decreased the Amhy induced phosphorylation of Smad1/5/8 in a dose dependent manner (Supplementary Figure 6e). These results indicated that Amhy repressed the transcription of cyp19a1a via Smad1/5/8 and Smad4b in XY fish.

Fig. 3.

Amhy represses the transcription of cyp19a1a. a) Effects of Smad1, Smad4b, Foxl2, and Sf1 cotransfection on the transcriptional activity of tilapia cyp19a1a promoter in HEK293 cells. b) Effects of Amhy on the transcriptional activity of tilapia cyp19a1a promoter in the presence of Smad1, Smad4b, Amhr2, Foxl2, and Sf1. Results were presented as the mean ± SD of triplicate transfections. “+” represents 100 ng/well constructs. Different letters above the error bar indicate statistical differences (P < 0.05), as determined by one-way ANOVA followed by Turkey test. c) EMSAs showing binding of Smad4b protein to the tilapia cyp19a1a promoter. Smad4b bound to the biotinylated probe covering the potential binding site in the cyp19a1a promoter in a dose-dependent manner; the binding was competitively inhibited by unlabeled cold probe. Unlabeled mutated probe failed to compete for binding of Smad4b to the biotinylated probes. Arrow indicates the protein/DNA complex. d) Phosphorylation of Smad1/5/8 in the HEK293 cells cotransfected with Smad1, Amhr2, and Amhy expression vectors. e) Phosphorylation of Smad1/5/8 in the XX-WT, XY-WT, and XY-amhy⁻ fish at 8 dpf (day post fertilization).

Sex reversal in the XY-amhy⁻ fish is rescued by mutation of cyp19a1a

Expression of Cyp19a1a in the XX fish was detected at 10 dpf, and increased at later stages (12 dpf and 15 dpf), but it was not detected in the XY fish. Mutation of amhy but not amh or amhΔ-y in XY fish resulted in the expression of Cyp19a1a as observed in XX females (Fig. 4a). It is unclear how the sexual phenotype will be determined if both amhy and cyp19a1a are lost in XY fish. To answer this question, we generated amhy and cyp19a1a double mutants (XY-amhy⁻; cyp19a1a^−/−) by crossing XY-amhy⁻; cyp19a1a^+/− phenotypic females with XX-cyp19a1a^−/− phenotypic males. At 90 dpf, the gonads of XY-amhy⁻ developed as normal ovaries (Fig. 4b), while the gonads of XX-cyp19a1a^−/− fish (n = 6, fish examined by histological staining), XY-amhy⁻; cyp19a1a^−/− fish (n = 6) and letrozole treated XY-amhy⁻ fish (n = 6) developed as testes (Fig. 4, c–e). The genetic sex (XX and XY) of these mutants was genotyped by sex specific marker (Sun et al. 2014b). Consistently, by immunofluorescence, specific signal of male leydig cell marker Cyp11c1 (Zheng et al. 2020) was detected in the gonads of these phenotypic males (Fig. 4, f–i), and specific signal of female marker 42Sp50 was only detected in the gonads of XY-amhy⁻ fish (Fig. 4, j–m). These results indicated that loss of cyp19a1a or Letrozole treatment could rescue the sexual phenotype of XY-amhy⁻ fish.

Fig. 4.

Mutation of cyp19a1a rescued the sexual phenotype in XY-amhy⁻ fish. a) Expression of Cyp19a1a in the gonads of XX-WT, XY-WT, and XY-amhy⁻ fish during sex determination. b–e) Histological examination of gonads from XY-amhy⁻, XX-cyp19a1a^−/−, XY-amhy⁻; cyp19a1a^−/− and XY-amhy⁻ + AI (aromatase inhibitor, Letrozole) at 90 dpf using H&E staining. Homozygous mutation of cyp19a1a resulted in female to male sex reversal in Nile tilapia. Here, we use XX-cyp19a1a^−/− as a control. f–m) Expression of male (Cyp11c1) and female (42Sp50) marker analyzed by immunofluorescence. Nuclei were counterstained with DAPI. AI, aromatase inhibitor (letrozole); Oc, oocytes; Sg, spermatogonia; Sc, spermatocytes. Scale bars: 15 μm.

amh is involved in germ cell proliferation and fecundity

Morphological and histological analysis showed that the ovaries of adult (300 dpf) XY phenotypic females with amhy single or with amhΔ-y and amhy double mutation developed normally compared with that of the XX-WT (Fig. 5, a-i–vi and c). However, hypertrophic ovaries and abnormal folliculogenesis due to uncontrolled proliferation of germ cells were observed in the XY phenotypic females with amh and amhy double or with amh, amhΔ-y, and amhy triple mutation and YY-amhy^−/− phenotypic females (Fig. 5, a-vii–xii and c). The gonadosomatic index (GSI) showed no differences in the XY phenotypic females with amhy single or with amhΔ-y and amhy double mutation, while significantly increased (P < 0.05) in the XY phenotypic females with amh and amhy double or with amh, amhΔ-y, and amhy triple mutation and YY-amhy^−/− mutants compared with that of XX-WT (Fig. 5b). There were no differences in the expression of Cyp19a1a, serum E2 levels, spawned eggs, hatching rate and spawning frequency among XY phenotypic females with amhy single or with amhΔ-y and amhy double mutation and XX-WT (Supplementary Figure 7). In contrast, the XY phenotypic females with amh, and amhy double or with amh, amhΔ-y, and amhy triple mutation and YY-amhy^−/− phenotypic females were infertile with no eggs spawned (data not shown). Normal folliculogenesis was observed in the E2 induced XY phenotypic females with amh single mutation and YY phenotypic females with amhy heterozygous mutation (Supplementary Figure 8, a–e), fecundity assay indicated that these E2 treated mutants were fertile (Supplementary Figure 8, f–g). All these results indicated that the amh was essential for controlling female germ cell proliferation and fertility, and in the absence of amh, amhy could compensate for the function of amh.

Fig. 5.

Effects of amh/amhΔ-y/amhy mutation on germ cell proliferation and fecundity. a) Morphological and histological analyses of gonads from mutants at 300 dpf. Phenotypes of XY-amhy⁻, XY-amhΔ-y⁻; amhy⁻, XY-amh⁻; amhy⁻, XY-amh⁻; amhΔ-y⁻; amhy⁻, and YY-amhy^−/− mutants (n = 6). b) Gonadosomatic index (GSI, n = 6). c) Relative follicle counting (n = 6). Follicles from the median cross sections of ovaries were counted for statistical analyses. d) Detection of germ cell proliferation by immunohistochemical analysis of PCNA expression (i–v) at 90 dpf. e) Number of PCNA positive cells (n = 6). PCNA positive cells from the median sections of the testes were counted for statistical analyses. f) GSI (n = 6). g) Fertilization rate (n = 6). Results were presented as the mean ± SD in b, c, and e–g. Different letters above the error bar indicate statistical differences (P < 0.05), as determined by one-way ANOVA followed by the Turkey test. Og, oogonia, I–IV, phase I to phase IV follicle. Spg, spermatogonia; Spc, spermatocytes; Spt, spermatids; AI, aromatase inhibitor Letrozole. XY-amhy⁻ and XY-amh⁻; amhy⁻ fish were rescued by letrozole (200 μg/g diet) treatment.

Increased proliferation of spermatogonia was revealed by PCNA (cell proliferation marker) staining in the XY fish with amh but not amhΔ-y single mutation compared with that of the XY-WT fish (Fig. 5, d-i–iii and e). In order to investigate whether amhy could compensate for the function of amh in controlling the proliferation of spermatogonia, XY fish with amhy single and amh and amhy double mutation were induced as phenotypic males by Letrozole. Significantly increased proliferation of spermatogonia was observed in the XY male fish with amh/amhy double mutation compared with that of the XY male fish with amh, amhΔ-y, or amhy single mutation (Fig. 5, d-iv–v and e). The GSI of the XY male fish with amh, amhΔ-y, or amhy single mutation showed no difference while significantly increased in the XY male fish with amh and amhy double mutation compared with that of the XY-WT (Fig. 5f). The fertilization rates of the XY male fish with amh or amhΔ-y single mutation and amh and amhy double mutation were similar to that of the XY-WT (Fig. 5g). All these results indicated that X-linked amh was required for normal spermatogonia proliferation, in the absence of amh, amhy could compensate for the function of amh.

Discussion

In this study, we analyzed the expression profiles of amh, amhΔ-y, and amhy during sex determination and differentiation, and analyzed their functions in sex determination and germ cell proliferation, and revealed the possible mechanism of how amhy determines sex in Nile tilapia.

Multiple lines of evidence supports amhy as the SD gene

A previous study revealed that single mutation of amhy or double mutation of amhΔ-y and amhy resulted in male to female sex reversal in XY tilapia, while single mutation of amhΔ-y could not (Li et al. 2015). Consistently, in this study, we established several new mutant lines and found that male to female sex reversal occurred in XY fish with amhy single, amhy and amhΔ-y and amh, amhΔ-y and amhy triple mutation. These results demonstrated that sex reversal occurred only when amhy was mutated in XY fish. In contrast, amh/amhΔ-y single and double mutation resulted in no sex reversal. In addition, homozygous mutation of amhy in YY males also resulted in male to female sex reversal. All these evidence from mutants in XY and YY fish strongly support amhy, rather than amh or amhΔ-y, as the SD gene of Nile tilapia.

Differences of the coding region between amhΔ-y and amh/amhy was the absence and/or presence of 233 bp deletion in exon 7 and 5 bp insertion in exon 6. The 233 bp deletion and 5 bp insertion in amhΔ-y were confirmed to be 100% linked with sex in Nile tilapia from 6 strains (Curzon et al. 2020). Consistently, both 233 bp deletion and 5 bp insertion in amhΔ-y were detected in amhΔ-y from 4 Nile tilapia strains in the present study. A study proposed that amhΔ-y was the sex determiner of Nile tilapia via regulation of amhy expression (Curzon et al. 2020). However, amhΔ-y encodes protein rather than noncoding sequence. In order to confirm that amhΔ-y can regulate the expression of amhy, targeted deletion of whole amhΔ-y gene is necessary. But it is difficult to delete amhΔ-y specifically due to the high similarity of these 3 copies. In addition, the 5 bp insertion and the 233 bp deletion in amhΔ-y resulted in a truncated protein without the TGF-β domain which do not transcriptionally regulate the expression of cyp19a1a suggesting that it do not function in Amh signal transduction. The 233 bp deletion in exon 7 of amhΔ-y was considered to be an ideal marker for distinguishing XX and XY Nile tilapia from different strains (Eshel et al. 2014; Curzon et al. 2020; Triay et al. 2020; Taslima et al. 2021). However, the 233 bp sequence was still present in the exon 7 of amhΔ-y, with the presence of amhy, from some male individuals of the Lake Kou (Burkina Faso) and Lake Koka (Ethiopia-East Africa) populations (Sissao et al. 2019; Triay et al. 2020). These results indicated that the 233 bp deletion was polymorphism existing at different frequency in different populations/strains. As for amh, various insertions or deletions were also detected in the promoter region among different Nile tilapia strains in a recent study (Jiang et al. 2021). These polymorphisms, together with the phenotypes of amh/amhΔ-y single and double mutation, excluded amh and amhΔ-y as the SD gene in Nile tilapia.

Up-regulation of amhy expression during the critical period of sex determination makes it the SD gene

The expression patterns of SD genes are often sex specific and spatiotemporal specific (Nagahama et al. 2020). The male SD gene dmy of medaka is specifically expressed in the Sertoli cells of XY gonads during gonadal differentiation and development (Kobayashi et al. 2004). The female SD gene DM-W of Xenopus laevis is expressed transiently only in the ZW gonads during the sex determination period (Yoshimoto et al. 2010). In Patagonian pejerrey, amhy is expressed much earlier than that of the autosomal amh in the presumptive Sertoli cells of the XY males during testicular differentiation (Hattori et al. 2012). Similarly, the SD gene amhy of Northern pike starts its expression before any molecular and morphological sex differentiation (Pan et al. 2019). Consistently, our results showed that amhy was expressed earlier and higher in the Sertoli cells of the XY male gonads during sex determination compared with that of amh in XY Nile tilapia. Probably, the specific expression pattern of amhy in XY gonads makes it the SD gene in tilapia.

Previously, 2 changes, one was the deletion of the 5,608 bp in the promoter and the other was the SNP (C/T) in exon 2, were suggested to be responsible for amhy to become the SD gene (Li et al. 2015). However, the SNP change (from C to T) in exon 2 of amhy was not detected in some Nile tilapia strains with amhy as the sex determiner (Curzon et al. 2020; Triay et al. 2020). In this study, the SNP (C/T) in exon 2 of amhy was also variable among different Nile tilapia populations from China. These results indicated that the SNP was a polymorphism existing at different frequency in different populations/strains, and completely ruled out the C to T change in amhy responsible for male sex determination. On the other hand, the 5,608 bp deletion of amhy promoter, which was conserved in different Nile tilapia strains (Sissao et al. 2019; Jiang et al. 2021), may be responsible for the higher expression of amhy in XY fish. In addition, some other subtle variations including deletions and insertions exist in amhy promoter compared with that of amh, these variations may also be responsible for the earlier and higher expression of Amhy during sex determination. Further investigations are needed to confirm whether these variations or regulatory factors are responsible for the altered expression of amhy from amh during sex determination.

Amhy determines male sex by repressing cyp19a1a expression

Different members of the TGF-β family have been identified as the SD genes in different fish species. For example, amhy takes over male sex determination in the Patagonian pejerrey, Nile tilapia and Northern pike (Hattori et al. 2012; Li et al. 2015; Pan et al. 2019). In fugu, a trans-species missense SNP in amhr2 is associated with male sex determination (Kamiya et al. 2012). In yellow perch, a duplicate of the amhr2 on the Y chromosome (amhr2y) becomes the candidate SD gene (Feron et al. 2020). In Atlantic herring (Clupea harengus), a truncated form of a Bmp1b receptor bmpr1bby, originated from gene duplication and translocation, becomes the candidate male SD gene (Rafati et al. 2020). In addition, gdf6y, encoding a TGF-β family growth factor, is the master SD gene in killifish (Nothobranchius furzeri) (Reichwald et al. 2015). Therefore, TGF-β signaling pathway plays an important role in fish sex determination. However, the molecular mechanism of how TGF-β members determine fish sex is largely unknown.

The salmonid master SD gene sdy represses the expression of cyp19a1a by hijacking the classical vertebrate sex differentiation cascade by interacting with Foxl2 and Sf1. Hence, by blocking a positive loop of regulation needed for the synthesis of estrogens in the early differentiating gonad, consequently allowing testicular differentiation to proceed (Bertho et al. 2018). Reciprocal expression of amh and cyp19a1a was reported in several fish species including tilapia (Wang and Orban 2007; Ijiri et al. 2008; Poonlaphdecha et al. 2013; Sarida et al. 2019; Huang et al. 2020). Our data showed that Amhy inhibited the expression of cyp19a1a through the Amhr2/Smads signaling pathway. In Nile tilapia, Foxl2 could interact with Sf1 to enhance the Sf1 mediated cyp19a1a transcription (Wang et al. 2007). A recent study has also obtained similar results, activation of the medaka cyp19a1a promoter requires the presence of both Foxl2 and Sf1 (Bertho et al. 2018). Synergistic interaction between Smads and Foxl2 in control of fshβ transcription has been well documented (Tran et al. 2011; Roybal et al. 2014; Ongaro et al. 2020; Weis-Banke et al. 2020). Consistently, Smad1 and Smad4 could enhance the transcription of cyp19a1a in a dose dependent manner in the presence of Foxl2 and Sf1 in HEK293 cells, indicating possible interaction of Smad1/4 with Foxl2 and Sf1. TGF-β Family members regulate the expression of target genes by phosphorylation of R-smad (Smad1/5/8) and then recruiting Co-Smad (Smad4) to form a complex into the nucleus. We showed that in HEK293 cells, Amhy inhibited the expression of cyp19a1a via Amhr2 and Alk2/3/6, phosphorylation of Smad1/5/8, the activation of Smad4. The R-smad/Co-Smad complex bound to the promoter and repressed the cyp19a1a transcription probably by interaction with Foxl2 and Sf1. Mutation of cyp19a1a or chemical inhibition of the synthesis of estrogen in XY-amhy⁻ mutants rescued the sex reversal phenotype, which further showed that amhy determines the sex of Nile tilapia by inhibiting the expression of cyp19a1a.

Dmrt1 is an important transcription factor implicated in male sex determination and differentiation in teleosts (Chen et al. 2014; Lin et al. 2017; Dai et al. 2021). In zebrafish, early cyp19a1a-mediated suppression of dmrt1 is required to establish a bipotential ovary and initiate female fate acquisition (Webster et al. 2017), and introduction of the dmrt1 mutation into the cyp19a1a mutants could rescue the all-male phenotype of the latter (Wu et al. 2020; Romano et al. 2020). In Nile tilapia, the expression of foxl2 and cyp19a1a in XX fish is up-regulated from 9 dpf (5 dah) compared with XY fish, while the expression of dmrt1 in XY fish is up-regulated from 10 dpf (Ijiri et al. 2008), indicating that the up-regulation of dmrt1 in XY fish depend on the suppression of cyp19a1a. Consistently, mutation of cyp19a1a resulted in elevated expression of dmrt1 and female to male sex reversal in Nile tilapia (Zhang et al. 2017). Introduction of the dmrt1 mutation into the cyp19a1a mutants rescued the female to male sex reversal phenotype in Nile tilapia (data not shown) as reported in zebrafish (Romano et al. 2020; Wu et al. 2020). All these results indicated that Amhy determines the male sex of Nile tilapia by repressing cyp19a1a expression and E2 production through the Amhr2/Smads signaling pathway, which in turn, results in elevated dmrt1 expression and initiates the male sex differentiation (Fig. 6).

Fig. 6.

A proposed model for Amh/AmhΔ-y/Amhy on tilapia sex determination and germ cell proliferation. a) Diagram showing neofunctionalization of amhy and nonfunctionalization of amhΔ-y after tandem duplication on the Y chromosome. The X-linked Amh retained its function, control of germ cell proliferation, in tilapia as reported in other teleosts. The Y-linked tandem duplicate amhy was neofunctionalized as a male SD gene, while the amhΔ-y was nonfunctionalized. In the absence of Amh, Amhy could compensate for the function of Amh in control of germ cell proliferation. b) Diagram of Amhy in Nile tilapia sex determination. Amhy functions through binding its type II receptor Amhr2, resulted in recruitment of the type I receptor (Alk2/3/6) to form a heteromeric signaling complex, in which the type I receptor is trans-phosphorylated by the type II receptor. The activated type I receptor then phosphorylates R-Smad (Smad1, 5, and 8), the phosphorylated R-Smad associate with co-Smad (Smad4). The activated R-Smad/co-Smad complexes translocate to the nucleus to interact with transcription factors to repress the expression of cyp19a1a driven by Sf1/Foxl2, and results in low E2 level and elevated expression of Dmrt1, and initiates the male differentiation.

amhy could compensate for the function of amh in control of germ cell proliferation

It is well-documented that amh is important for controlling germ cell proliferation in teleosts. Homozygous mutation of amh in female fish resulted in excessive proliferation of germ cells and hypertrophic ovaries in zebrafish (Lin et al. 2017; Yan et al. 2019; Zhang et al. 2020b) and Nile tilapia reported by our group (Liu et al. 2020). Consistently, mutation of amh in male zebrafish resulted in excessive proliferation of spermatogonia (Lin et al. 2017; Yan et al. 2019; Zhang et al. 2020b). In this study, loss of both amh and amhy in XY tilapia resulted in significantly increased proliferation of spermatogonia. All these data suggested that the function of Amh was conservative in controlling germ cell proliferation during teleosts gametogenesis. Homozygous mutation of amhy in YY fish, or amh and amhy double and amh, amhΔ-y, and amhy triple mutation in XY fish resulted in phenotypic females displaying same phenotypes in XX Nile tilapia with amh homozygous mutation reported by our group (Liu et al. 2020). These results strongly indicated that amhy could compensate for the function of amh in controlling germ cell proliferation in Nile tilapia (Fig. 6).

Conclusion

In this study, we analyzed the functions of the 3 copies of amh in sex determination and germ cell proliferation of Nile tilapia through different mutation combinations, and revealed the possible mechanism of how amhy determines the male sex. The elevated expression of amhy during sex determination may be the key reason for amhy to become a SD gene. Amhy represses the expression of cyp19a1a via the Amhr2/Smads signaling pathway, resulting in the elevated expression of dmrt1, thus initiating male differentiation. In addition, our data revealed that amhΔ-y was dispensable for sex determination and germ cell proliferation, while amh was essential for normal germ cell proliferation. In the absence of amh, amhy could compensate for the function of amh in controlling germ cell proliferation. This study provided a good case for understanding the neofunctionalization and nonfunctionalization of the duplicated genes involved in sex determination of teleosts.

Data availability

Strains and plasmids are available upon request. The authors affirm that all data necessary for confirming the conclusions of the article are present within the article, figures, and tables.

Supplemental material is available at GENETICS online.