Introduction

Sex determination and sex differentiation are important developmental events in the life cycle of all sexually reproducing animals [1]. In fish, sex determination is complicated, and various types, including gonochorism and hermaphrodites, have been reported [2]. Monopterus albus is a freshwater fish with a snake-like body shape that is widely distributed in Asia from India to Asiatic Russia [3]. Owing to its high nutritional and medical value, this species is widely cultured in China, with up to 0.39 million tons being produced in 2019 [4]. M. albus is a hermaphroditic fish whose sex changes from female through intersex to male during its life cycle [5]. Recently, many studies from various perspectives were conducted to analyse the mechanism of sex reversal and abundant data were generated that support their insight [6,7,8,9]. However, the underlying mechanism of sex reversal remains unclear.

Epigenetic regulation plays a crucial role in sex determination and sex differentiation [10, 11]. DNA methylation, an important epigenetic modification, is generally associated with gene silencing [12] and play important roles in developmental process, especially sex differentiation [13]. In mice, DNA methylation levels of the promoter region of sry exhibited an inverse relationship with gene expression. DNA methylation levels were low in the E11.5 gonad while the expression level reached the peak [14]. In the red-eared slider turtle (Trachemys scripta elegans), KDM6B exhibits temperature-dependent sexually dimorphic expression, and knockdown of Kdm6b at 26 °C promoted male-to-female sex reversal [15]. In some fish, such as sea bass, temperature influences sex determination. In temperature-masculinized fish, high temperatures increased the cyp19a promoter methylation levels in females and methylation of cyp19a promoter decreased cyp19a expression [16]. Assay for Transposase-Accessible Chromatin using sequencing (ATAC-seq) is a novel technique used for the identification of open chromatin [17]. Chromatin is in an uncondensed state when open, and is accessible to gene transcriptional machinery and regulatory element binding. However, when condensed, chromatin is inactive and in a transcriptionally inaccessible state [18].

Here, we produced and analysed comprehensive DNA methylome, transcriptome, and chromatin accessibility maps of M. albus gonads over the course of gonad development. We found that chromatin accessibility was weakly correlated with the gene expression. On the contrary, a strongly negative correlation was observed between DNA methylation status and gene expression. Based on the expression profile, differentially expressed genes (DEGs) were divided into several groups, and we identified that DNA methylation of regulatory elements, especially the transcription start site (TSS), was strongly correlated with gene expression, which was up-regulated from ovary to ovotestis. Chromatin accessibility maps showed that these gene promoter regions were in low accessibility states which suggested that DNA methylation could play crucial roles in gonadal development. Plastin-2 was selected as the key candidate gene as per integration of low accessibility states, hypomethylation, and up-regulated expression. We determined that promoter methylation of plastin-2 was negatively correlated with gene expression in developing gonads and degenerated gonads treated with 17α-methyltestosterone (MT). We used zebularine to inhibit methyltransferase and found that the methylation status of three CpG sites could regulate plastin-2 gene expression. Finally, we found that transcription factor dmrt1 and foxl2 bind to the promoter of plastin-2 and were responsible for transcriptional activation of the plastin-2 gene. These data suggested that DNA methylation could impact the binding of transcription factor dmrt1 and foxl2 through methylation and demethylation to regulate plastin-2 expression during gonadal development.

Methods

Animals and samples

Rice field eel (M. albus) were obtained from the Breeding Center of Yangtze River Fisheries Research Institute, Chinese Academy of Fishery Sciences, China. Nine eels of three different growth stages (female, intersex, and male) were killed by MS222 administration according to Yangtze River Fisheries Research Institute Care Committee (No. 2013001) [6]. The gonads were collected and divided into three parts. One was preserved in 4% paraformaldehyde (pH 7.5) to prepare tissue sections; the second was frozen in liquid nitrogen and then stored at − 80 °C for RNA extraction; and the third was stored in ethanol for DNA extraction. The DNA and RNA was extracted as per the method described in a previous study [19]. Concentration of DNA and RNA were detected using an Agilent 2100 Bioanalyzer (Agilent Technologies) and the integrity was identified through agarose gel electrophoresis.

Nanopore sequencing

Genomic DNA of the gonads from male, female, and intersex eels were extracted using the HiPure Tissue & Blood DNA Kit (Magen). The concentration and quality of DNA were detected using an Agilent 2100 Bioanalyzer (Agilent Technologies) and the integrity was detected using 1% agarose gel electrophoresis. Two microgrammes of gDNA was repaired using a NEB Next FFPE DNA Repair Mix kit (New England Biolabs). The ONT template prep kit (Oxford Nanopore Technologies) was used to prepare the template according the manufacturer’s instructions. The library was sequenced using the ONT sequencing reagent kit (Oxford Nanopore Technologies) according to the manufacturer’s instructions on the ONT PromethION48 platform with PromethION Flow Cells [20]. The fast5 format data were converted to fastq format for QC analysis using Basecall with ONT’s Guppy software [21]. The fastq data were further filtered to remove the adapters, short reads (length < 500 bp), and low-quality reads (MeanQual < 6).

Identification and analysis of differentially methylated regions

Minimap2 [22] was used to analyse the sequence depth and alignment efficiency and to align the clean-reads to the reference genome. Nanopolish software with the Hidden Markov Model was used to detected the DNA methylation sites [23]. SMART2 software [24] was used to detect the differentially methylated region (DMR) and differentially methylated loci (DML). When the P value was < 0.05 and the methylation specificity was > 0.2, the CpG site or region was regarded as a DML or DMR. ChIPseeker software was used to locate the DML or DMR in the corresponding region in the genome [25].

RNA-seq library generation and sequencing

Total RNA was extracted applying the TRIzol method according to the manufacturer’s instructions. NanoDrop 2000 (Thermo Fisher Scientific) and Bioanalyzer 2100 (Agilent Technologies) were used to determine the concentration and quality. RNA integrity was determined using 1% agarose gel electrophoresis. RNA-seq libraries were constructed in different gonad developmental stages (female, intersex, and male) in M. albus. The Ribo-Zero™ MagneticKit (Epicentre) was used to remove the rRNA, followed by reverse transcription of these RNAs to construct the cDNA library using NEBNext Ultra Directional RNALibrary Prep Kit (New England Biolabs). We then prepared paired-end sequencing on an Illumina Nova Seq6000 (Illumina San Diego). Clean data were obtained by removing reads containing adapters, reads containing poly-N, and low-quality reads from raw data. Q20, Q30, GC-content, and sequence duplication level of the clean data were calculated.

Identification and analysis of differentially expressed genes

Differential expression analysis of two groups was performed using the DESeq R package (1.10.1) [26]. The threshold value was determined in multiple tests using the Benjamini–Yekutieli method to control the false discovery rate (FDR) [27, 28]. Criteria for DEGs were a false discovery rate < 10−3 and an expression level differing at least twofold between two groups.

ATAC-seq library construction, sequencing, and mapping

For ATAC-seq library construction, gonads from three females, three intersex, and three males were collected. Approximately 50,000 cells were used for ATAC-seq library construction according to a previous description [29, 30]. The cells were washed with cold PBS and then suspended with cold lysis buffer (10 mM Tris, pH 7.4, 10 mM NaCl, 3 mM MgCl2, and 0.1% IGEPAL CA-630). The DNA was purified after incubation at 37 °C for 30 min with the Tn5 transposase. Transposed DNA fragments were used as the template for PCR amplification. The final DNA libraries were run on the Illumina NextSeq 500. Low-quality reads and adapters were removed and the Bowtie2 software was used to map the high-quality reads to the genome [31]. DeepTools v3.2 was used to map the density distribution of reads in the 3 kb intervals up- and down-stream of the TSS of each gene [32].

Identification of ATAC peaks and DARs

Reads mapped to mitochondrial DNA and unplaced scaffolds were removed. MACS2 V2.1.1 [33] was used to identify ATAC peaks according to the following steps: removing redundant reads; then adjusting the read position; calculating peak enrichment; and estimating the empirical FDR. When the FDR was < 0.05, it was selected as the identified peak. To identify the differentially accessible regions (DARs), DiffBind version 2.6.6 [34] was used to detect the difference peak. The affinity score was calculated based on the read count which was then used as the input of the DESeq2 software [26] to screen for DARs in each group. When the fold change was > 1.5 and the P value was < 0.05, it was identified as a DAR.

17α-Methyltestosterone treatment

Two hundred and forty M. albus larvae at 60 days post-fertilization 60 per group, were used to produce sex reversal M. albus by sex steroid exposure as described in a previous study [35]. The control group underwent no treatment. Dry powders of MT were dissolved in 95% ethanol and then diluted in water. Larvae were immersed daily for 10–12 h in water containing MT at a concentration of 100 µg/L (MT1), 200 µg/L (MT2), and 300 µg/L (MT3). Forty mg of dry powders of MT were dissolved in 95% ethanol and then mixed with 200 g daphnia to a concentration of 200 µg/g MT. The mixture was kept at room temperature for 24 h and then stored at 4 °C. The trial continued for 2 months. The gonads were then collected and divided into three parts. One was preserved in 4% paraformaldehyde (pH 7.5) to prepare tissue sections, the second was frozen in liquid nitrogen and then stored at − 80 °C for RNA extraction, and the third was stored in ethanol for DNA extraction.

Treatment of primordial gonadal cells with zebularine

Gonad tissues were collected from healthy M. albus specimens, and the tissues were washed with cold foetal bovine serum (FBS), the gonads were then cut into small pieces and placed on a 70-μm nylon mesh and gently pushed through the mesh with constant dripping of cold Dulbecco's modified Eagle medium (DMEM) (Thermo Fisher Scientific) containing 10% FBS (BioInd), 1% heparin and penicillin–streptomycin (P/S) (Thermo Fisher Scientific). The cell suspension was carefully collected and centrifuged at 500×g for 10 min at 4 °C. The supernatant was discarded and the cells were resuspended in 3 mL DMEM containing 10% FBS and 1% P/S. Finally, the cells were seeded in a 6-well plate (Corning) at a density of 1 × 106 cells/well. The cells were incubated in a 5% CO2 incubator at 28 °C. After overnight culture, non-adherent cells were carefully washed off and the adherent cells were exposed to zebularine at concentrations of 50, 100, and 200 mg for 48 h. The cells were harvested for RNA preparation and DNA extraction to detect the change in expression profile and DNA methylation level.

Plasmid constructs

Expression plasmids, pcDNA3.1-foxl2 and pcDNA3.1-dmrt1, were constructed from the pcDNA 3.1(+) plasmid using the double restriction endonucleases NheI and EcoRI (NEB). The promoter sequence of plastin-2 (XP_020477222) was obtained from the genomic database. Three deletion fragments (1575, 1074, and 436 bp) of the promoter were amplified from genomic DNA according to the designed primers (Additional file 1: Table S5), and cloned into the pGL3-basic vector (Promega) using KpnI and XhoI. Site-directed mutagenesis for the dmrt1 and foxl2 binding sites were performed using a Fast Site-Directed Mutagenesis Kit (TIANGEN) and the primers described in Additional file 1: Table S5.

Cell culture, transfection, and dual-luciferase reporter assays

HEK293T cells were obtained from the Center of Animal Science and Animal Medicine, Shandong Agricultural University and cultured at 37 °C in DMEM (Thermo Fisher Scientific) containing 10% FBS (BioInd) and 1% P/S (Thermo Fisher Scientific) for 24 h before transfection. The cells were seeded onto the 24-well plates at a concentration of 1 × 105 per well. The DMEM was removed and opti-MEM medium was added to incubate the cells. Five hundred nanogramme recombinant constructs and 50 ng pRL-TK were co-transfected into the cells in 400 µl opti-MEM medium using Lipofectamine™ 3000 (Invitrogen) according the manufacturer’s instructions and incubated at 37 °C for 6 h. Subsequently, the opti-MEM medium was removed and the cells was incubated for 48 h in DMEM. Cells were collected, and luciferase activity assays were performed using a Dual-Luciferase kit (Promega).

In situ hybridization

To assess plastin-2 expression in gonadal cells, two pairs of primers were designed (Additional file 1: Table S5). A 433 bp cDNA fragment was amplified from the plastin-2 Open Reading Frame (ORF) domain. Primers were used to amplify the T7 promoter sequence synthetic probe, and PCR products were purified using a QIAquick Gel Extraction Kit (QIAGEN). Transcription was performed using the MEGAshortscript T7 High Yield Transcription Kit (Thermo Fisher Scientific) to obtain probes. In situ hybridization was performed as described in a previous study [36] using anti-digoxigenin-AP Fab fragments (Roche) as the antibody, while BCIP/NBT (Beyotime) was used to detect the positive signal.

Quantitative real-time PCR

The expression of various tissues, developing gonads, and gonadal cells in M. albus were analysed by quantitative real-time PCR (qRT-PCR) according to a previous description [37]. ef1a was used as the internal control gene, cDNA was used as a template, tissues from three individuals were used, and each reaction was repeated three times. qRT-PCR was performed as follows: 95 °C for 30 s; 40 cycles of 95 °C for 5 s, 60 °C for 30 s, and 72 °C for 30 s; and then 72 °C for 5 min.

Bisulphite PCR methylation analysis

Bisulphite PCR methylation analysis was conducted according to a previous description [38]. Genomic DNA from at least 15 individuals in one group was extracted using the TIANamp Genomic DNA Kit (Tiangen). Concentration of DNA was detected using an Agilent 2100 Bioanalyzer (Agilent Technologies) and the integrity was identified via agarose gel electrophoresis. Mixed DNA were treated using the DNA methylation kit (Zymo) following the manufacturer's protocol. Primers were designed using the online MethPrimer design software (http://www.urogene.org/methprimer/). PCR amplification was conducted using treated DNA as the template, and the purified DNA was cloned into the PMD-18T vector. Each group sequence had 15–20 positive clones. The methylation level was analysed using the DNA methylation analysis platform (http://services.ibc.uni-stuttgart.de/BDPC/BISMA/).

Statistics

All statistical tests were performed using SPSS 22.0 (IBM). The difference in expression was analysed using one-way analysis of variance followed by Duncan multiple comparison tests. Significance was set at P < 0.05, and different letters indicated significantly difference. Differences in mean expression level and methylation level among groups were determined using an independent sample t-test. Differences in the ratio of methylated to unmethylated CpG at each site were assessed using a Chi-square test followed by Fisher's exact test. A P value of less than 0.05 was considered significant.

Results

Genome-wide DNA methylation level profiling

To understand the DNA methylation changes during M. albus gonadal development, we collected the samples from the three important developmental stages female (ovary), intersex (ovotestis), and male (testis) (Fig. 1A). Whole-genome bisulphite sequencing was conducted and 36.1 GB, 31.7 GB, and 32.9 GB clean data were obtained from the ovary, ovotestis, and testis groups, respectively (Additional file 1: Table S1). The raw sequence data were deposited in the National Genomics Data Center [39] under the accession number CRA007182. The string average depths were 45.1 (ovary), 39.5 (ovotestis), and 41.1 (testis). All clean-reads were assembled and mapped to the M. albus reference genome [40]. Means of 99.65% (ovary), 99.57% (ovotestis), and 99.57% (testis) reads were mapped to the genome (Additional file 1: Table S1). To detect the DNA methylation dynamics, global CpG methylation levels during gonadal development were analysed and a scale of methylation level between 0.75 and 1 in ovaries was higher than that in ovotestes and testes. However, a scale of methylation level between 0.25 and 1 showed no dramatic differences among the ovary, ovotestis, and testis (Fig. 1C). The distribution of global CpG methylation levels in ovary, ovotestis, and testis showed a similar bimodal distribution with a peak outside gene region (Fig. 1D). To further detect the DNA methylation dynamics in gonadal development, 23,265 hyper- and 7416 hypo-DMRs in the OVT/OV groups, and 55,841 hyper- and 10,342 hypo-DMRs in the TE/OV groups (Fig. 1E). Taken together, 15,496 hypo-DMRs were obtained. To study whether hypo-DMRs were associated with regulation activity, an ATAC-seq library was constructed from the same samples containing three ovaries, three ovotestis, and three testes. DMR exhibited much lower ATAC-seq signals than their neighbouring region in the OV group, while ATAC-seq signals were absent in the OVT and TE groups (Fig. 1F). Dynamic change of the methylation profile from ovary to testis showed that methylation level decreased gradually (Fig. 1F). These results suggested that DNA methylation may play a key role in the development of ovary to testis.

Fig. 1
figure 1

DNA methylation dynamic alteration during the gonadal development. A Scheme of HE section from developing gonad. B Pearson correlation coefficient of genome-wide CpG methylation among the developing gonad. C Global CpG methylation levels and total CpG with low (0,0.25), medium (0.25,0.75) and high (0.75,1) methylation level during gonadal development. D Distribution of genome-wide CpG methylation level of each stage of gonadal cell. E Number of DMRs identified between the two groups. F ATAC-seq signals and DNA methylation levels over 10-kb regions centred on a total of 15496 hypo-DMRs

Full size image

Genes activation associated with the TSS methylation status

To study the relationship between DNA methylation status and transcriptomic dynamics during gonadal development, RNA-seq was conducted using the same samples containing three ovaries, three ovotestis and three testes. A total of 4527 DEGs containing 469 down-regulated and 4058 up-regulated genes displayed > twofold differential transcript abundance in the OVT/OV group (Fig. 2A). In the TE/OV group, a total of 11,853 DEGs containing 5523 down-regulated and 6330 up-regulated genes (Fig. 2A). The up-regulated genes in the OVT/OV group were enriched in biological functions including translation, cell adhesion, spermatid differentiation, and immune response (Fig. 2B). All DEGs were divided into seven groups according to the expression profiles (Fig. 2C). Nine hundred and three genes in cluster 4 exhibited up-regulated expression from ovary to ovotestis which was then down-regulated from ovotestis to ovary. The expression profile of cluster 4 suggested that these genes could play key roles in sex reversal. To examine DNA methylation dynamics of regulatory elements associated with the expression of these genes, DNA methylation changes were analysed for these gene promoters and the centre region. DNA methylation levels showed no remarkable difference in promoter and centre regions among the groups. However, in the promoter region, the methylation status was hypo while in the centre region, the methylation status was hyper (Fig. 2D). These results suggested that the promoter region methylation status might regulate the expression of these genes.

Fig. 2
figure 2

DNA methylation changes in TSS region activated gene expression. A MA plots for differentially expressed genes during gonadal development. Black dots represent genes with no significantly change. Red and green dots represent significantly up-regulated and down-regulated genes during gonadal development, respectively (P < 0.05). B Gene ontology terms associated with significantly differentially expressed genes. C Different expression profile during the gonadal development. D DNA methylation level over 10-kb region centre and promoter of the differentially expressed genes from cluster 4

Full size image

Gene activation is not associated with decreasing chromatin accessibility

To test the hypothesis that accessible chromatin is key in regulating gene transcription, ATAC-seq libraries from the same samples were generated. The raw sequence data were deposited in the National Genomics Data Center under the accession number CRA007367. About 100 MB of total clean data were obtained from each sample and > 95% of reads were mapped to the genome (Additional file 1: Table S2). The peak number in the three groups ranged from 153,317 to 306,358. We found that about 25% of ATAC peaks were in the promoter region (3 kb region around the TSS), 35% of ATAC peaks were distally intergenic to TSS (Additional file 1: Table S3) while no peak was found in the 3ʹ- and 5ʹ-UTRs. To study the dynamics of chromatin accessibility during gonadal development, the differential accessibility regions (DARs) were identified and a total of 53,540 and 41,010 DARs exhibited > twofold differential accessibility in the OVO/OV and TE/OV groups (Additional file 1: Table S4). To determine whether these dynamic chromatin accessibility changes were associated with gene expression in gonadal development, ATAC signal and DNA methylation level of 903 genes from cluster 4 were analysed and it was found that ATAC-seq signals were decreased from ovary to ovotestis and the expression profile was increased while the DNA methylation level was decreased (Fig. 3A). Sixty-seven overlapped genes were obtained from hypo-DMRs, ATAC-decreased, and the genes from cluster 4 (Fig. 3B). These gene expressions were up-regulated from ovary to ovotestis and down-regulated from ovotestis to testis (Fig. 3C). Eleven out of 67 candidate genes were verified and only seven genes (slc22a17, pmp-22, tnfrs6, plastin-2, tdrp, pqlc1, and unc13d) displayed the expression profile of cluster 4 (Additional file 1: Fig. S1). Integrative Genomics Viewer showed that the RNA transcription level was negative with the ATAC-seq peak and DNA methylation level for plastin-2 gene, unc13d, tnfrs6, and tdrp (Fig. 3D). These results support that chromatin accessibility was not responsible for activating gene expression while DNA methylation might play a key role in gonadal development.

Fig. 3
figure 3

Gene activation with lose of the chromatin accessibility. A DNA methylation and ATAC signal changes over 10-kb promoter region of the differentially expressed genes from cluster 4. B Overlap genes of DMR hypo, ATAC decrease and RNA up-regulated from cluster 4. C Heatmap of the 67 overlap genes expression profile. D Identified candidates of sex-reversal related genes. Red boxes indicate DARs changes during sex reversal

Full size image

DNA methylation was associated with plastin-2 expression in developing gonads

To detect the sex reversal mechanism of plastin-2 during gonadal development, we analysed the association between RNA transcription and DNA methylation. We firstly analysed the sequence of the promoter region (GenBank accession no: ON692924) using MethPrimer software and found a region around the TSS with high CpG content; the methylation primers were designed according this sequence (Additional file 1: Table S5). We found that plastin-2 expression was significantly higher in the ovotestis than in the ovary and testis (P < 0.05, Fig. 4A). To detect the distribution of plastin-2 in gonads, in situ hybridization was conducted to identify the expression of plastin-2 in ovary, testis, and ovotestis cells while the sense probe had no signal (Additional file 1: Fig S2). We found a strongly positive signal in granulosa cells and cytoplasm in ovaries; in ovotestis, plastin-2 was strongly expressed in granulosa cells and spermatogonia whereas, in testis, plastin-2 was strongly detected in spermatocytes and spermatids (Fig. 4B). The methylation status among the same samples during gonadal development were detected (Additional file 1: Table S6) and we found that methylation levels were significantly higher in the testis and ovary (94.36% and 93.18%, respectively), than in ovotestis (87.66%; P < 0.05, Fig. 4C). The results suggested that methylation status was negatively associated with the expression level.

Fig. 4
figure 4

Expression and methylation level of plastin-2 in ovary, ovariotestis and testis of Monopterus albus. A Expression profile of plastin-2 in ovary, ovariotestis and testis detected by qRT-PCR; B RNA expression of plastin-2 in ovary, ovariotestis and testis detected by in situ hybridization; C DNA methylation level of plastin-2 in ovary, ovariotestis and testis of Monopterus albus. Oo: oocytes; Sp: spermatocytes; Yv: yolk vesicle; Spg: spermatogonia. 4,5,6 is the large magnification of frame area in 1,2,3, respectively

Full size image

Expression and DNA methylation status of plastin-2 after MT treatment

To further examine the sex reversal mechanism of plastin-2, MT was used to treat the larvae for 2 months. Histological examination showed that the ovary was degenerated after MT treatment (Fig. 5A). The number of oocytes and cellular volume decreased. Additionally, connective tissues were increased (Fig. 5A). Expression of dnmt1 and dnmt2 were significantly decreased while plastin-2 expression was significantly increased (P < 0.05, Fig. 5B). The results suggest that DNA methylation status undergoes changes. Bisulphite sequencing PCR (BSP) examined the DNA methylation status after MT treatment (Additional file 1: Table S7). We found that the DNA methylation level was significantly decreased (P < 0.05, Fig. 5C).

Fig. 5
figure 5

Expression and DNA methylation level of plastin-2 in ovary and degenerated ovary treated by MT. A Histology sections of ovary and degenerated ovary treated by MT; B expression profile of dnmt1, dnmt2 and plastin-2 between the ovary and degenerated ovary treated by MT; C. methylation level of plastin-2 promoter between the ovary and degenerated ovary treated by MT

Full size image

DNA methylation could regulate plastin-2 expression

To detect the role of promoter methylation in plastin-2 expression, zebularine was used to inhibit the methylation level in the primordial gonadal cells. We found that dnmt1 and dnmt2 expression were significantly decreased at a concentration of 50 mg compared with that of the other genes (P < 0.05, Fig. 6A, B), while expression of plastin-2 was significantly increased (P < 0.05, Fig. 6C). DNA methylation level was lower than the control group after zebularine treatment but was not significantly different (P > 0.05, Fig. 6D), while methylation level of CpG sites 5,6, and 17 were significantly decreased in the zebularine group (P < 0.05, Additional file 1: Table S8, Fig. 6E). These results suggested that zebularine inhibited the methylation level of CpG sites to promote gene expression.

Fig. 6
figure 6

Expression and methylation level of plastin-2 in primordial gonadal cells treated by zebularine. A Expression profile of dnmt1 in various concentrations of zebularine; B expression profile of dnmt2 in various concentrations of zebularine; C expression profile of plastin-2 at concentrations of 50 mg zebularine; D methylation level of plastin-2 promoter after treated by zebularine; E. comparison of methylation level in each CpG sites between zebularine and control group

Full size image

Dmrt1 and foxl2 activate plastin-2 promoter

To explore the regulatory elements, potential binding sites for transcription factors were predicted using the JASPAR online software. Several sex-related transcription factor binding sites were found in the promoter region (Fig. 7A). To determine the exact binding sites, luciferase reporter assays with a series of deletions were conducted and the luciferase activities were significantly higher than the basic group in the three deletion constructs (Fig. 7B). Luciferase activities exhibited that key regulatory elements ranged from − 1 to − 424 which includes one dmrt1 and two foxl2 binding sites (Fig. 7B). To determine the potential role of the transcriptional factors dmrt1 and foxl2 in the plastin-2 promoter, site mutants were constructed using the PGL3-Mn3 plasmid as the template. Three mutants including Mut-dmrt1, Mut-foxl2a, and Mut-foxl2b were obtained (Fig. 7C). Luciferase activities were significantly decreased after the dmrt1 binding site mutation in group of pGL3-Mn3 + pcDNA3.0-dmrt1 (P < 0.05, Fig. 7D). Additionally, we found that the foxl2a binding site mutation caused significant down-regulation of the luciferase activity in the pGL3-Mn3 + pcDNA3.0-foxl2 group (P < 0.05), while the foxl2b binding site mutation caused no significant difference (P > 0.05, Fig. 7D). The findings suggested that dmrt1 and foxl2 could bind to activate the plastin-2 promoter, while mutation of any of them caused a decrease in luciferase activity, except for the foxl2b binding site.

Fig. 7
figure 7

Luciferase assay detected the binding sites and activation of transcriptional factor. A Schematic showing the dmrt1 and foxl2 binding sites and CpG sites. B Luciferase assay showing the activity of deletions constructs. C Schematic showing the mutation of dmrt1 and foxl2, and the wild type. D Luciferase assay showing site mutation of promoter in 293T cells. The mean ± SEM was from three independent experiments, *P < 0.05 shows significantly difference

Full size image

Discussion

Epigenetic modifications, including DNA methylation and chromatin accessibility, have been proposed as the molecular mechanisms underlying sex differentiation [15, 41, 42]. In the present study, we used M. albus, a good model for studying sex reversal [43], to investigate the potential role of DNA methylation and chromatin accessibility in gonadal development. Gonads from three key stages of the developmental process were selected to investigate with multiple epigenomics assays. We found a weak correlation between gene expression and chromatin accessibility during sex reversal. However, we found that DNA methylation status had a strong negative correlation with gene expression in the dynamic changes. The observation suggested that DNA methylation might play a crucial role during sex reversal. Another study also exhibited that DNA methylation and chromatin accessibility work together to regulate gene expression in zebrafish regeneration [44].

To detect how epigenetic modification affects regulatory elements to alter gene expression, we detected hundreds of genes activated in sex reversal using RNA-seq. Epigenetic modification of regulatory elements was identified and we found that chromatin accessibility was also weakly responsible for gene expression while DNA methylation status was strongly negatively correlated with gene expression, especially for the TSSs. These regulatory elements were in a hypomethylated but low accessibility state in sex reversal. Whether epigenetic modifications other than DNA methylation in these regulatory elements contribute to sex reversal should be further studied. Global epigenome profiling not only has advantages for detecting genomic regulatory elements but it also identified important regulatory genes. To further explore the association of DNA methylation with sex reversal, up-regulated genes with hypomethylated profiles of promoter and low accessibility states were selected as the putative candidate genes for further study.

Plastin-2, called L-plastin, an actin-binding protein [45] known for regulating T-cell adhesion and immune synapse formation [46]. Recently, another study showed that plastin-2 is expressed in the ovary and plays a crucial role in reproduction [47, 48]. Additionally, we found that the binding sequence of oestrogen receptor was in the promoter region of plastin-2 and its expression was up-regulated after dihydrotestosterone and oestradiol induction [49, 50]. However, the function of plastin-2 involved in sex differentiation was unclear. RNA-seq showed that plastin-2 expression was significantly up-regulated from ovary to ovotestis the gonadal development while the promoter was hypomethylated and in a low accessibility state. The results suggested that DNA methylation was strongly negatively associated with gene expression. To further detect whether DNA methylation was closely related to gene expression, we detected the DNA methylation level of plastin-2 promoter and plastin-2 gene expression during the key stages of gonadal development. We found that plastin-2 gene expression was negatively correlated with promoter methylation status, and that plastin-2 was expressed in granulosa cells and cytoplasm in ovaries whereas in spermatocytes and spermatids in testes. After larvae were treated with MT, we found that ovaries were degenerated and that plastin-2 expression was significantly up-regulated with a hypomethylated promoter profile. In another study in humans, plastin-2 expression was up-regulated after oestradiol and dihydrotestosterone treatment [49]. Additionally, dnmt1, dnmt2, and DNA methyltransferase1 and 2 expression was down-regulated, which suggested that methyltransferase could be involved in sex reversal. To detect whether methylase activity inhibits up-regulation of plastin-2 expression during sex reversal, zebularine, which inhibits DNA methyltransferase during replication through covalent binding [51], was used to treat primordial gonadal cells to inhibit DNA methylation. The methylation level was low in the treated group with no significant difference. We determined three CpG sites whose methylation levels were significantly down-regulated and were negatively associated with plastin-2 expression. To determine the exact binding sites that activate plastin-2 expression in the promoter, binding sites of transcription factors were detected and luciferase activity was analysed We found dmrt1 and foxl2 can bind to the plastin-2 promoter and were responsible for transcriptional activation of the plastin-2 gene. Dmrt1 and foxl2 are known as the marked genes in sex differentiation that have been reported in many species [52,53,54]. In addition, many studies have determined that promoter methylation can inhibit the binding of transcription factors in the gene promoter region, resulting in the decline of gene transcription level or even termination of transcription, to regulate gene expression [55, 56]. Taken together, these data suggest that DNA methylation could impact the binding of transcription factor dmrt1 and foxl2 through methylation and demethylation to regulate plastin-2 expression during gonadal development.

Perspectives and significance

Monopterus albus is a hermaphroditic fish that undergoes sex reversal from ovary to testis via ovotestis during gonadal development which is consider an ideal model for genetic research. The present study on the gonadal development of M. albus provides novel insights from epigenetic modification to help elucidate the potential molecular mechanism. It will facilitate the sex control and biomarker exploration in M. albus. In the future, epigenetic markers will be explored to identify the un-reversal female to improve the fecundity which exhibited low level in the normal female due to the small body size. Function studies on epigenetic modification are needed to elucidate the molecular mechanism of M. albus gonadal development and to reveal the sex differentiation mechanism of the hermaphroditic animal.