Abstract
N6-methyladenosine (m6A), catalyzed by Mettl3 methyltransferase, is a highly conserved epigenetic modification in eukaryotic messenger RNA (mRNA). Previous studies have implicated m6A modification in multiple biological processes, but the in vivo function of m6A has been difficult to study, because mettl3 mutants are embryonic lethal in both mammals and plants. In this study, we have used transcription activator-like effector nucleases and generated viable zygotic mettl3 mutant, Zmettl3m/m, in zebrafish. We find that the oocytes in Zmettl3m/m adult females are stalled in early development and the ratio of full-grown stage (FG) follicles is significantly lower than that of wild type. Human chorionic gonadotropin-induced ovarian germinal vesicle breakdown in vitro and the numbers of eggs ovulated in vivo are both decreased as well, while the defects of oocyte maturation can be rescued by sex hormone in vitro and in vivo. In Zmettl3m/m adult males, we find defects in sperm maturation and sperm motility is significantly reduced. Further study shows that 11-ketotestosterone (11-KT) and 17β-estradiol (E2) levels are significantly decreased in Zmettl3m/m, and defective gamete maturation is accompanied by decreased overall m6A modification levels and disrupted expression of genes critical for sex hormone synthesis and gonadotropin signaling in Zmettl3m/m. Thus, our study provides the first in vivo evidence that loss of Mettl3 leads to failed gamete maturation and significantly reduced fertility in zebrafish. Mettl3 and m6A modifications are essential for optimal reproduction in vertebrates.
TO date, >100 chemical modifications have been identified in eukaryotic RNA. Among these, N6-methyladenosine (m6A) is the most abundant RNA modification and the first one shown to be reversible in eukaryotic messenger RNA (mRNA) (Desrosiers et al. 1974; Fu et al. 2014).
A multicomponent methyltransferase complex, including methyltransferase-like 3 (Mettl3) (Bokar et al. 1997), methyltransferase-like 14 (Mettl14) (J. Liu et al. 2014), Wilms tumor 1-associated protein (Wtap) (Ping et al. 2014), and KIAA1429 (Schwartz et al. 2014), is responsible for adenosine methylation, while fat mass and obesity-associated protein (Fto) and alkB homolog 5 (Alkbh5) have been identified as the demethylase for m6A (Jia et al. 2011; Zheng et al. 2013). Additionally, m6A can be recognized by multiple RNA-binding proteins, such as the YTH domain family proteins (Ythdf1–3, Ythdc1) and the heterogeneous nuclear ribonucleoprotein (HNRNP) protein families (HNRNPA2B1 and HNRNPC) (X. Wang et al. 2014, 2015; Alarcon et al. 2015a; Liu et al. 2015; Xiao et al. 2016). Together, these m6A methyltransferases, demethylases, and binding proteins function as “writers,” “erasers,” and “readers” of m6A and have implicated m6A in a variety of biological processes, such as mRNA metabolism (X. Wang et al. 2014, 2015), microRNAs maturation (Alarcon et al. 2015b), spermatogenesis (Hsu et al. 2017; Xu et al. 2017), stemness (Batista et al. 2014; Geula et al. 2015), circadian rhythm (Fustin et al. 2013), and disease(Cui et al. 2017).
First identified as a component of the methyltransferase complex, Mettl3 is highly conserved in eukaryotes from yeast to humans (Bokar et al. 1997; Yue et al. 2015). In Saccharomyces cerevisiae, Ime4 (inducer of meiosis 4; homolog of METTL3) has an important role in the initiation of meiosis and sporulation. Deletion of ime4 leads to the loss of m6A defects in sporulation (Clancy et al. 2002; Schwartz et al. 2013). In Arabidopsis thaliana, MT-A (mRNA adenosine methylase; homolog of METTL3) is mainly expressed in dividing tissues, particularly the reproductive organs. Inactivation of MT-A in A. thaliana results in an embryonic lethal phenotype with developmental arrest at the globular stage (Zhong et al. 2008; Bodi et al. 2012). The Drosophila melanogaster homolog of mettl3, Ime4, is mainly expressed in ovaries and testes, and is required for Notch signaling during oogenesis. Loss of Ime4 in fly leads to a lethal phenotype (Hongay and Orr-Weaver 2011), but this finding is challenged by recent studies finding that Drosophila Ime4-null mutants are viable and fertile, though flightless, and m6A is required for female-specific alternative splicing of sxl (Haussmann et al. 2016; Lence et al. 2016). In mammals, Mettl3 affects cell division, differentiation, reprogramming, and spermatogenesis. Knockout of mettl3 in mouse embryonic stem cells impairs differentiation and mettl3−/− mice are embryonic lethal (Y. Wang et al. 2014; Chen et al. 2015; Geula et al. 2015). Recently, Yang’s laboratory generated germ cell conditional mettl3 knockout mice and found that Mettl3 is essential for spermatogenesis by regulating spermatogonial differentiation and meiosis (Xu et al. 2017). Thus, Mettl3 is clearly indispensable in different organisms, implicating an essential role for m6A in embryonic development.
Furthermore, He’s laboratory reports that one-third of zebrafish maternal mRNAs are highly methylated at m6A, and clearance of m6A, mediated by Ythdf2, is critical for the maternal-to-zygotic transition (Zhao et al. 2017). Morpholino knockdown of wtap and mettl3 in zebrafish embryos leads to tissue differentiation defects and the expression of somite marker myod (Ping et al. 2014). Recently, Liu’s group demonstrated the critical function of m6A modification in the fate determination of hematopoietic stem/progenitor cells during zebrafish embryogenesis (Zhang et al. 2017). However, the in vivo functions of Mettl3 and m6A function in zebrafish adults remains poorly understood. To elucidate the role of Mettl3 and m6A in zebrafish reproduction, we employed transcription activator-like effector nucleases (TALENs)-mediated genomic editing technique and generated a null allele of mettl3 (Zmettl3m/m). We found that Zmettl3m/m zebrafish are viable, in contrast to the lethal phenotype in mammals and plants. Loss of Mettl3 leads to failed gamete maturation and impaired fertility that are accompanied by altered sex hormone synthesis and gonadotropin signaling. Our study provides the first line of evidence supporting that Mettl3-mediated modifications in vivo are essential for zebrafish reproduction.
Materials and Methods
Zebrafish maintenance
AB strain zebrafish used in this study were maintained and raised in recirculation systems at 28.5° under an alternating 14 hr:10 hr light/dark cycle. All animal experiments were conducted in accordance with the Guiding Principles for the Care and Use of Laboratory Animals and were approved by the Institute of Hydrobiology, Chinese Academy of Sciences.
Establishment of mettl3 mutant zebrafish lines
The TALEN target was the first exon of mettl3. The half-site length was 17 bp while the position “0” was T, and the spacer length was 15 bp. The pCS2-TALEN-ELD/KKR plasmids for TALENs were constructed as described (Y. Liu et al. 2014). The final TALEN plasmids were linearized by NotI and transcribed using the mMESSAGE mMACHINE Sp6 Kit (Ambion, Austin, TX). TALEN mRNAs (300–500 pg) were microinjected into one-cell stage wild-type (WT) zebrafish embryos. Mutations were confirmed by competitive PCR and sequencing (Luo et al. 2015); the primers mett3-F, mett3-R1, and mett3-R2 for PCR are listed in Supplemental Material, Table S1. The injected embryos were raised to adulthood and outcrossed with WT to obtain F1 offspring. The adults of F1 heterozygotes (mettl3m/+) were then outcrossed with WT again. F2 heterozygotes with the same mutation were raised to adulthood and self-crossed to obtain F3 zebrafish, which would give rise to zygotic deficiency mutant lines (Zmettl3m/m), then the Zmettl3m/m self-crossed to obtain maternal and zygotic deficiency mutant lines (MZmettl3m/m). The flowchart for mutant line development is shown in Figure S1 in File S1.
Generation of Mettl3 overexpressing transgenic zebrafish lines
The mettl3 sequence was cloned into laboratory stocks of the pSK-GFP (Tol2-CMV-GFP-pA-CMV-MCS-pA-Tol2) vectors to get Mettl3 overexpression construction (Tol2-CMV-GFP-pA-CMV-mettl3-pA-Tol2; Figure S2A in File S1). mRNA encoding Tol2 transposase (100 ng/ml) and Tol2-based Mettl3 overexpression construction (50 ng/ml) were co-injected into one-cell stage WT zebrafish embryos to generate the Mettl3 overexpression line (OE-mettl3). A flowchart of OE-mettl3 lines is shown in Figure S2B in File S1. The positive embryos expressed fluorescence (Figure S2C in File S1). We crossed the Mettl3 overexpression transgenic line (OE-mettl3) with the mettl3 knockout line (Zmettl3m/m) to obtain the mettl3-specific knockout and ectopic expression lines (OE-KO) to rescue the homozygotes.
RNA isolation and real-time quantitative PCR (qPCR)
Total RNA samples were isolated from the embryos, tissues, FG stage follicles, and pituitaries of WT and Zmettl3m/m zebrafish using TRIzol Reagent (Invitrogen, Carlsbad, CA). The amount and purity of the RNA were determined by spectrophotometry and agarose gel electrophoresis. After a treatment with RNase-free DNase (Promega, Madison, WI), the total RNA was reverse-transcribed to complementary DNA (cDNA) using Rever Tra Ace M-MLV (TOYOBO, Osaka, Japan) with random primers. qPCR was carried out on an ABI 7900HT RT-PCR System using 2×SYBR Green mix (TOYOBO, Japan). The expression level of mettl3 was normalized to that of β-actin, and the expression levels of target genes in ovaries and testes were normalized to that of the internal control ef1a. The gene names and primers used in this study are listed in Table S1.
Quantitative analysis of the m6A level using LC-MS/MS
Experiments followed the published procedures in Jia et al. (2011). The mRNAs were isolated from WT and Zmettl3m/m total RNA using the PolyATtract Isolation Systems (Promega). Two hundred nanograms of mRNA was digested by nuclease P1 (2 U) at 50° for 1 hr, followed by the addition of NH4HCO3 (100 mM) and alkaline phosphatase (0.5 U). After incubation at 37° for 1 hr, the solution was diluted to 100 μl, and 10 μl of the solution was injected into LC-MS/MS. Nucleosides were separated by high-performance liquid chromatography coupled with triple-quadrupole tandem mass spectrometry and quantified using the nucleoside to base ion mass transitions of 282.2–150 (m6A), and 268.2–136 (A). Quantification was performed in comparison with the standard curve obtained from pure nucleoside standards running on the same batch of samples. The ratio of m6A to A was calculated based on the calculated concentrations.
In situ hybridization
To label the primordial germ cells (PGCs), we used vasa as the marker gene (Yoon et al. 1997). In situ hybridization was performed as described (Thisse and Thisse 2008). To verify the localization of mettl3 in zebrafish gonads, fluorescent in situ hybridization on sections was performed with our published methods (Song et al. 2015). Fluorescence photomicrographs were collected using a laser scanning confocal microscope (Zeiss LSM710). Sense RNA probe was used as a negative control. Primers used in in situ hybridization can be found in Table S1.
Western blot
For Western blot analysis, tissues from WT and Zmettl3m/m fish were lysed by the Total Protein Extraction Kit (Sangon Biotech, Shanghi, China). The lysates were first separated on a 12.5% SDS-PAGE gel and transferred onto PVDF membranes. The separated proteins were immunoblotted with rabbit anti-Mettl3 (15073-1-AP; Proteintech) and rabbit anti-β-actin (bs-0061R; Bioss). After incubation overnight at 4°, the membrane was washed in phosphate buffered saline containing Tween-20 for 3 × 10 min and incubated with an HRP-conjugated anti-rabbit secondary antibody (#7074; Cell Signaling Technology). The signals were detected with Immobilon Western Chemiluminescent HRP Substrate (Millipore) and visualized using ImageQuant LAS 4000 mini system (GE Healthcare).
Histological analysis and fertility assessment
After MS-222 anesthesia, we dissected the intact gonadal tissues from WT and Zmettl3m/m adult zebrafish (4 months postfertilization) and calculated the gonadosomatic index (GSI; gonad weight/body weight × 100%). The gonad was fixed in 4% PFA (Sigma, St. Louis, MO) overnight at 4° and embedded in paraffin. Then the sections (7 μm) were stained with hematoxylin and eosin (H&E) and examined under a microscope (Olympus BX53). The staging systems on oogenesis and spermatogenesis were identified as described previously (Selman et al. 1993; Shang et al. 2006; Leal et al. 2009; Q. Wang et al. 2015). As described in a previous report (Chu et al. 2014; Zhang et al. 2015), adult zebrafish (WT and Zmettl3m/m) were transferred to a breeding aquaria at the ratio of one male to one female. Each pair of the following were used to assess the fertilization rate: WT male × WT female; WT male × Zmettl3m/m female; Zmettl3m/m male × WT female; Zmettl3m/m male × Zmettl3m/m female; OE-mettl3 female × WT male; OE-mettl3 male × WT female. The numbers of eggs ovulated and ovulation rate (ovulation rate = number of spawned females/total number of females × 100%) were assessed by natural mating of WT, Zmettl3m/m, and OE-mettl3 females with WT males, respectively. If the females did not spawn after mating with males, the experiment was repeated 7 days later. Females were considered sterile after three such failed attempts.
Isolation and incubation of ovarian follicles
After being dissected from WT and Zmettl3m/m females, the follicles of different stages were manually separated and divided into five groups based on their size and vitellogenic state: primary growth stage (PG, ∼0.15 mm); previtellogenic stage (PV, ∼0.25 mm); early vitellogenic stage (EV, ∼0.35 mm); midvitellogenic stage (MV, ∼0.45 mm); full grown stage (FG, ∼0.65 mm). The incubation of FG stage follicles was carried out according to a previous report (Li et al. 2015). Briefly, the cells were incubated with human chorionic gonadotropin (hCG; 100 IU/ml) in medium 199 (Gibco) for 6 hr at 28°, and oocytes were examined microscopically for ooplasmic clearing, which occurs due to proteolytic cleavage of vitellogenin and is indicative of oocyte maturation (Selman et al. 1994; Li et al. 2015). No hormone was added in the control groups. Groups of four mutants and four controls were assayed in three replicate wells for a total of 24.
Intraperitoneal injection in adult zebrafish
The procedure was performed according to a previous report (Kinkel et al. 2010). After anesthetization, the Zmettl3m/m females that did not spawn naturally and WT females were weighted, and then injected intraperitoneally with 0.65% isotonic saline (control) or the maturation-inducing hormone 17α,20β-dihydroxyprogesterone (17α-20β-DHP, 5 μg/g; Cayman Chemical, Ann Arbor, MI). After 4 hr, the ovaries were carefully dissected for quantification of oocyte maturation.
Sperm motility assessments
The sperm motility assessment was performed according to our previous report (Chen et al. 2016). Fresh semen samples were obtained by manual extrusion. Thereafter, 1 μl of the diluted semen was dropped onto a 2X-CEL Chamber slide and analyzed with a CEROS sperm tracker (Hamilton Torne, Beverly, MA). The analysis parameter included: average path velocity (VAP), curvilinear velocity (VCL), straight-line velocity (VSL), amplitude of lateral head displacement (ALH), and beat-cross frequency (BCF).
Immunofluorescence of synaptonemal complexes
According to the method described previously (Chen et al. 2016), the synaptonemal complex was detected in spermatocyte bivalents from WT and Zmettl3m/m adult males. The immunofluorescence was stained by anti-SYCP3 and anti-SYCP1 antibodies, which were prepared as described previously (Chen et al. 2016).
Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) analysis
Gonadal tissues were dissected from 4 months postfertilization WT and Zmettl3m/m adult fish and embedded in Optimal Cutting Temperature compound (O.C.T. SAKURA). The samples were sectioned at 7 μm thickness. The TUNEL cell death assay was performed using the In Situ Cell Death Detection Kit (Roche) according to the manufacturer’s instructions. Images were obtained using a laser scanning confocal microscope (Zeiss LSM710).
Sex steroid measurements
Blood samples were collected from the caudal vein of the WT and Zmettl3m/m adults as described by Pedroso et al. (2012). Serum samples were extracted by centrifugation at 3000 × g for 15 min at 4°. 11-keto-testosterone (11-KT) and estradiol-17β (E2) were measured by competitive enzyme-linked immunosorbent assay kits (Cayman Chemical) following the manufacturer’s instructions.
Prediction of m6A modification sites
Most m6A sites are found within the consensus sequence RRACH ([G/A/U][G > A]m6AC[U > A > C]) (Dominissini et al. 2012; Meyer et al. 2012), so we wrote an in-house script to match RRACH motifs with the mRNA sequences of target genes from the NCBI database (https://www.ncbi.nlm.nih.gov/).
Statistical analyses
All data were expressed as mean ± SEM and analyzed by a Student’s t-test or a one-way ANOVA using SPSS 18.0 (SPSS, Chicago, IL). Chi-square analysis was used to identify differences in sex ratio between WT, Zmettl3m/m, and OE-KO groups. P < 0.05 was considered statistically significant. The figures were drawn using Prism 5 GraphPad Software. Results were confirmed in three independent experiments.
Results
Mettl3 expression and m6A levels are high in embryos and gonads
To study the functional role of mettl3 and m6A in zebrafish, we first examined mettl3 expression and m6A level in embryos and tissues of WT zebrafish. The results show that mettl3 is a maternally expressed gene. It is highly abundant in the early stages of embryonic development, but dramatically decreased at the 256-cell stage, and is further decreased to ∼28.4% in the dome stage (Figure 1A). Mettl3 mRNA was detected in adult tissues (brain, hypothalamus, eye, heart, liver, spleen, kidney, gill, and gonad), with especially high levels in ovaries and testes (Figure 1B).
The expression pattern of mettl3 and m6A level in zebrafish embryos and adult tissues. (A and B) qPCR analysis for the temporal and spatial expression profile of mettl3 mRNA in WT embryos (A) and adult tissues (4 months postfertilization) (B). (C and D) Quantification of the m6A/A ratio of the total mRNA purified from WT embryos (C) and adult tissues (D) by LC-MS/MS. hpf, hours postfertilization.
Similarly, liquid chromatography-tandem mass spectrometry (LC-MS/MS) measurement showed that high levels of m6A persist during the early stages of development with a temporary decrease in 12 and 24 hpf (Figure 1C). In addition, the m6A/A ratio was quantified at relatively low levels in liver and muscle, with the higher levels of expression being found in brain, ovaries, and testis (Figure 1D). In situ hybridization revealed that mettl3 was expressed in both germ cells and somatic cells during gonad development (Figure 2). The high expression levels of mettl3 and m6A in embryos and gonads led us to determine their roles in development and reproduction.
Localization of mettl3 expression in the ovary and testis by fluorescent in situ hybridization. Cell nuclei from somatic cells are indicated by arrowheads, Leydig cells are indicated by asterisks, and Sertoli cells are indicated by arrows. NC, negative control using sense probe.
Generation of mettl3 mutant using TALENs
To define the function of Mettl3 and m6A in vivo, we targeted the first exon of mettl3 using TALENs (Figure 3A and Figure S1 in File S1). Upon PCR and sequencing, two independent mutant lines were obtained with a 4-bp deletion and 6-bp insertion (−4; +6) or a 10-bp deletion (−10) in the first exon of mettl3 (Figure 3A and Figure S3A in File S1). The two mutations resulted in an open reading frame-shift and caused a truncated protein. Both truncated proteins lost the catalytic DPPW motifs and S-adenosylmethionine (SAM) binding domain (Figure S3B in File S1). Western blot analysis showed that Mettl3 protein was completely abolished in the mutants (Figure 3B). Using LC-MS/MS, we compared the m6A level of mRNA from WT zebrafish and mettl3-deficient zebrafish. We observed that the m6A/A ratio decreased significantly in brain (P < 0.001), liver (P < 0.001), muscle (P < 0.01), testis (P < 0.001), and ovaries (P < 0.001) from mettl3-deficient zebrafish (Figure 3C). Strikingly, MZmettl3m/m mutant embryos developed poorly at the dome stage (3–4 hpf) and all of them died by 24 hpf (Figure S4 in File S1 and Table S2). We therefore used the Zmettl3m/m homozygote line for subsequent experiments (marked in green, Figure S1 in File S1).
Establishment of mettl3 knockout mutant lines. (A) Schematic representation of the genomic structures of zebrafish mettl3 and the target sites of TALENs. Recognition sequences are boxed, and the spacer sequences are between the two black boxes. The start codon of translation ATG is underlined. Deletions and insertions are indicated by dotted line and red letters, respectively. (B) Western blot analysis of whole zebrafish and dissected testes from WT and mettl3 mutant adults. (C) Quantification of m6A/A ratio of the total mRNA purified from WT and mettl3 mutant adult tissues by LC-MS/MS. (Student’s t-test, *** P < 0.001; data are presented as mean ± SEM.)
Oocyte maturation is disrupted in Zmettl3m/m females
Through morphological and histological analyses, we found the GSI was significantly reduced in the Zmettl3m/m (9.2%) compared to WT (16.2%) females (Figure 4, A–C, P < 0.001). Most oocytes were arrested in the early development stages of PG, PV, EV, and MV and showed a relatively loose arrangement in ovaries of Zmettl3m/m females (Figure 4, D and E). The ratios at the PG, PV, EV, and MV stage had no significant difference with the corresponding stage of WT, but the ratio at FG follicles was significantly lower in Zmettl3m/m (13.3%) than that of WT (25.6%) (Figure 4F, P < 0.05).
Oocyte maturation defects in Zmettl3m/m females. (A) Appearance of ovaries dissected from WT and Zmettl3m/m females. (B) Gross anatomical appearance of ovaries from WT and Zmettl3m/m females. Bar, 5 mm. (C) The GSI scatterplot of WT and Zmettl3m/m females (n = 12). GSI, gonadosomatic index. (D and E) H&E staining of the ovaries from WT and Zmettl3m/m females. PG follicles are indicated by arrowheads, and PV follicles are indicated by arrows. PG, primary growth stage; PV, previtellogenic stage; EV, early vitellogenic stage; MV, midvitellogenic stage; FG, full-grown stage. Bar, 200 μm. (F) The relative distribution of different stage follicles in the WT and Zmettl3m/m females (n = 12) (* P < 0.05; data are presented as mean ± SEM).
To determine if reproductive defects were related to the disruption of oocyte maturation in mutant females, FG stage follicles were isolated and treated with hCG in vitro. As expected, in the control group without hCG treatment, the percentage of germinal vesicle breakdown is 5.8% in WT follicles, but GVBD did not happen in Zmettl3m/m follicles. After incubation with hCG, the percentage of GVBD was significantly increased to 14.5% (P < 0.05) in WT and 3.9% (P < 0.05) in mutants, respectively, though the % GVBD in mutant follicles is still lower than in WT follicles (Figure 5, A and B).The data indicate that the competency to respond to endogenous hormones was adversely affected in the follicles of Zmettl3m/m. To further test the in vivo action of steroids on oocyte maturation of Zmettl3m/m, we injected the 17α-20β-DHP into adult zebrafish. In the control group, GVBD did not happen (Figure 5C). However, GVBD was evident in the ovaries of zebrafish injected with 17α-20β-DHP (Figure 5C). Quantitation of oocyte maturation in vivo indicated significant induction of oocyte maturation in the injected Zmettl3m/m (47.0%, P < 0.05) and WT (81.5%, P < 0.001) (Figure 5D). Taken together, these data suggest that oocyte maturation is impaired in mettl3 mutant fish and the defects of oocyte maturation in mutants can be rescued by hormones in vitro and in vivo.
Defects of oocyte maturation in mutants can be rescued by sex hormone. (A) Morphology of FG follicles dissected from WT and Zmettl3m/m ovaries with incubation of hCG (100 IU/ml) after 6 hr. Follicles undergoing GVBD are marked by red arrows. Bar, 500 μm. (B) Comparison of the %GVBD in WT and Zmettl3m/m in control or hCG treatment. (C) Gross morphology of ovaries dissected from adult zebrafish 4 hr after injection of isotonic saline (control) or 17α-20β-DHP. Representative follicles undergoing GVBD are marked by red arrows. Bar, 500 μm. (D) Quantitative assessment of oocyte maturation induction by injection of 17α-20β-DHP. (n = 12, one-way ANOVA; data are presented as mean ± SEM.)
Sperm maturation is blocked in Zmettl3m/m males
Similar to mutant females, the male GSI was also significantly decreased in Zmettl3m/m (0.62%) compared to WT (1.02%) (Figure 6, A–C, P < 0.001). Histological examination revealed that the lobular cavities were filled with mature sperm, aligned in a tight and orderly manner in the testes of WT males. In contrast, the lobular cavities were smaller and contained very little or no mature sperm in the testes of the Zmettl3m/m males (Figure 6, D and E). Early stages of spermatogenesis, particularly spermatogonia (SG) and spermatocytes (SC), were proportionally 24.4% (P < 0.05) and 56.1% (P < 0.01) in Zmettl3m/m males, respectively. In contrast, the proportions of SG and SC were, respectively, only 7.5 and 26.7% in WT males, while the proportion of spermatozoa (SZ) was 50.1% in WT males and significantly reduced (P < 0.001) to 10.4% in Zmettl3m/m (Figure 6F). Sperm motility parameters, such as average path velocity (VAP), curvilinear velocity (VCL), and straight-line velocity (VSL), were also significantly reduced (P < 0.01) in Zmettl3m/m males, compared with WT sperm, but no significant difference was found in the amplitude of lateral head displacement (ALH) and BCF between Zmettl3m/m and WT males (Table 1).
Sperm maturation is affected in Zmettl3m/m male. (A) Appearance of ovaries dissected from WT and Zmettl3m/m males. (B) Gross morphological appearance of ovaries from WT and Zmettl3m/m males. Bar, 5 mm. (C) The GSI scatterplot of WT and Zmettl3m/m males. (D and E) H&E staining of testes from WT and Zmettl3m/m. SG, spermatogonia; SC, spermatocyte; ST, spermatid; SZ, spermatozoa. SG are marked by arrows. Bar, 100 μm. (F) The ratio of different stage sperms in the WT and Zmettl3m/m males. (n = 15, Student’s t-test. * P < 0.05, ** P < 0.01, *** P < 0.001; data are presented as mean ± SEM.)
In addition, we examined the synaptonemal complex in Zmettl3m/m and WT spermatocytes using immunofluorescence. Synaptonemal complex protein Sycp1 is a component of the transverse filaments of synaptonemal complexes, which marks the synapsed chromosome regions (Handel and Schimenti 2010), and Sycp3 is a component of the axial/lateral synaptonemal complex elements. We found that both Sycp1 and Sycp3 protein assembled to the synaptonemal complex, similar to WT spermatocytes (Figure 7). In addition, expression levels of several genes involved in meiosis during gametogenesis were not different between Zmettl3m/m and WT (Figure S5, A and B in File S1). These data suggest that synapsis is normal in Zmettl3m/m spermatocytes, and mettl3 mutation primarily impairs sperm maturation.
Synapsis proceeds normally in the Zmettl3m/m spermatocytes. Bar, 10 μm (n = 12).
Fertility and sex differentiation are affected in Zmettl3m/m adults
Since the maturation of gametes was severely altered, we next assessed natural fertilization rates. We found that fertilization rate in WT fish was ∼91.4%. In contrast, Zmettl3m/m females or males outcrossed with the WT adults had the fertilization rate of 42.6% (Figure 8A, P < 0.001) and 48.4% (Figure 8A, P < 0.001), respectively, and the fertilization rate in self-cross homozygotes was only 33.3% (Figure 8A, P < 0.001). The fertilization rate of OE-KO females and males outcrossed with the WT adults was increased to 87.2 and 90.7%, respectively (Figure 8A).
Fertility and sex ratios are altered in Zmettl3m/m adults. (A) The fertilization rate (fertilization rate = fertilized eggs/total eggs × 100%) of WT, Zmettl3m/m, and OE-KO adults. ♀, female, ♂, male (n = 30, one-way ANOVA). (B) The ovulation rate (ovulation rate = number of spawned females/total number of females × 100%) of WT, Zmettl3m/m, and OE-KO females crossed with WT males (n = 30, Student’s t-test). (C) The numbers of eggs ovulated by WT, Zmettl3m/m, and OE-KO females crossed with WT males (n = 30, Student’s t-test). (D) The ovulation rate of WT females crossed with WT, Zmettl3m/m, and OE-KO males (n = 30, Student’s t-test). (E) Sex ratio of WT, Zmettl3m/m, and OE-KO adults (n = 120, one-way ANOVA; data are presented as mean ± SEM).
Compared with the 96.7% ovulation rate of WT females, only 51.0% of the Zmettl3m/m females were able to spawn naturally (Figure 8B, P < 0.001), but 83.3% of OE-KO females ovulated (Figure 8B). Those Zmettl3m/m females that spawned naturally only ovulated 158 eggs, whereas WT females ovulated 382 eggs (Figure 8C, P < 0.001). Ectopic expression of Mettl3 in Zmettl3m/m increased the number of eggs ovulated to 240 in OE-KO females, indicating a degree of rescue (Figure 8C). To test the fertility of mutant males, we crossed the WT females with the WT, Zmettl3m/m, and OE-KO males. We found 94.4% of WT male × WT female crosses spawned naturally, but only 8.1% of Zmettl3m/m male × WT female pairs spawned successfully (Figure 8D, P < 0.001). Ovulation rate of WT females increased to 87.5% when crossed with OE-KO males (Figure 8D). The TUNEL assays did not reveal any major apoptotic signals in Zmettl3m/m ovaries (Figure S6 in File S1) and testes (Figure S7 in File S1).These data indicate that the fertility of Zmettl3m/m females and males is significantly lower than that in WT adults.
Compared with the 1:1 sex ratio of WT adults, the sex ratio was significantly altered in Zmettl3m/m adults with ∼85.2% being males (Figure 8E, P < 0.001). The sex ratio of OE-KO adults approached normal expectations with 44.6% of the OE-KO being males (Figure 8E). The number of vasa-positive PGCs had no significant difference between Zmettl3m/m and WT embryos (Figure S8A in File S1). The PGC numbers, quantified at the bud stage by whole-mount in situ hybridization for vasa expression, were 24.8 for WT and 23.7 for mettl3 mutants (Figure S8B in File S1), indicating that deficits in PGCs do not account for the male bias. These data suggest that both mettl3 null females and males show significantly reduced reproduction and loss of Mettl3 affects sex ratio in the mutants.
Mettl3 deletion alters expression pattern of sex hormone synthesis and gonadotropin-related genes
To further analyze the molecular mechanism underlining the fertility defects observed in mettl3 mutants, the expression levels of a panel of genes (Table S1) involved in gamete maturation and steroidogenesis were assessed in pituitaries, FG stage follicles, and testis from the Zmettl3m/m and WT lines by qPCR. Compared with WT adults, lhβ mRNA was significantly lower in the pituitaries of Zmettl3m/m females (P < 0.05), but there was no difference for fshβ, gh, prl expression level (Figure 9A). For key genes involved in steroidogenesis, oocyte maturation, and ovulation, the expression of npr, igf3, star, 3βhsd, and cyp19a1a was significantly reduced while fsta expression was significantly increased in mutant FG stage follicles. In contrast, acvr2aa, acvr2b, cyp17a2, lhr, cyp11a1, 17βhsd, hsd17b3, amh, cyp26B1, esr2a, esr2b, ptgs2a, ptgs2b, piwil, wt1a, and fshr mRNA levels were normal in Zmettl3m/m FG follicles (Figure 9C). Unlike the females, there was a significant decrease in fshβ mRNA level in the pituitaries of Zmettl3m/m males (Figure 9B, P < 0.05); Lhβ expression was also decreased, but this was not statistically significant. There was no difference in the levels of expression of gh and prl (Figure 9B). Additionally, the expression of fshr, star, cyp11a1, cyp17a1, cyp17a2, 3βhsd, and hsd17β3 was significantly reduced, while cyp26B1, yp19b, cyp19a1a, 17βhsd, esr1, esr2a, esr2b, and lhr mRNA levels were normal in the Zmettl3m/m testes (Figure 9D).
Gene expression profiles in WT and Zmettl3m/m adults. Expression of genes in pituitaries of females (A) and males (C), FG stage follicles (B), and testes (D) by qPCR. (Student’s t-test, * P < 0.05, ** P < 0.01, *** P < 0.001; data are presented as mean ± SEM.)
Since the expression of genes involved in sex hormone synthesis was changed, we assessed the serum concentrations of sex hormones in adult fish. The results indicate that 11-KT and E2 levels are significantly decreased in Zmettl3m/m females (Figure 10A, P < 0.05). Similarly, 11-KT and E2 levels were significantly decreased in Zmettl3m/m males (Figure 10B, P < 0.001).
Serum 11-KT and E2 level are significantly decreased in Zmettl3m/m adults. (A) Serum concentration of 11-KT and E2 in WT and Zmettl3m/m females. (B) Serum concentration of 11-KT and E2 in WT and Zmettl3m/m males. (n = 12, Student’s t-test, * P < 0.05, *** P < 0.001; data are presented as mean ± SEM.)
To provide supportive evidence for the relationship between m6A modification and these differentially expressed genes including 3βhsd, cyp11a1, cyp17a1, cyp17a2, cyp19a1a, fshβ, fshr, fsta, hsd17β3, igf3, lhβ, npr, and star, we predicted the probable m6A modification site within the consensus sequence RRACH. We found the mRNA sequences of these genes had many of the consensus RRACH ([G/A/U] [G > A] m6AC [U > A>C]) (Table S3) and the motifs were mainly located in CDS and 3′UTR Region (Figure S9 in File S1), so these genes are potentially targets for the m6A modification.
Discussion
Unlike the lethal effect of mettl3 knockout in mammals (Geula et al. 2015) and plants (Zhong et al. 2008; Bodi et al. 2012), we obtained viable Zmettl3m/m-deficient homozygote zebrafish. In the mettl3 mutant, we reported that m6A methylation levels were significantly lower than in the WT line. The mettl3 deletion was detrimental to gamete maturation and fertility in zebrafish. This study provides the first in vivo evidence that Mettl3, the “writer” of m6A modification, plays an important role in regulating reproduction in zebrafish.
The process of oogenesis is divided into three different phases: growth, maturation, and ovulation (Kagawa 2013). Germinal vesicle breakdown is usually regarded as a hallmark of the progress of oocyte maturation. In response to hormonal stimulation, the germinal vesicle (GV) migrates to the animal pole. As maturation proceeds the GV becomes visible under a dissecting microscope and then disappears (Suwa and Yamashita 2007; Nagahama and Yamashita 2008). Compared with WT adults, there was no significant difference in the proportion of the early developmental stages, including PG, PV, EV, and MV stages. However, the number of FG stage follicles was significantly lower in the Zmettl3m/m line than in WT female fish, and the % GVBD and numbers of eggs ovulated were also significantly decreased. The disruption of gamete maturation was successfully rescued by overexpression of mettl3 in Zmettl3m/m, indicating that the deletion of mettl3 is the cause of the disruption of gamete maturation. These results indicate that the mettl3 mutants have defects in oocyte maturation.
In female zebrafish, Lh signaling is mainly responsible for stimulating oocyte maturation and ovulation, and Fsh signaling is mainly responsible for promoting follicular growth (Zhang et al. 2014; Chu et al. 2015). Compared to WT, the pituitaries of Zmettl3m/m females expressed lower levels of lhβ but there was no difference for the fshβ. Furthermore, the expression of fsta was significantly increased, while the npr and igf3 expression were significantly reduced in mutant FG stage follicles. Fsta inhibits hCG-induced oocyte maturation (Wu et al. 2000; Ge 2005), while the npr and igf3 activate oocyte maturation and ovulation (Li et al. 2015; Tang et al. 2016). Hence, these results show oocyte maturation and ovulation are blocked in Zmettl3m/m adults at multiple steps.
In most teleost fish, three major phases compose spermatogenesis: mitotic proliferation of spermatogonia, meiosis of spermatocytes, and spermiogenesis, which is characterized by the restructuring of spermatids into flagellated spermatozoa (Schulz and Miura 2002; Nobrega et al. 2009; Schulz et al. 2010). In S. cerevisiae, the initiation of meiosis and sporulation are dependent upon m6A methylation mediated by Ime4 (Clancy et al. 2002). In mice, the ablation of Mettl3 in germ cells severely inhibited spermatogonial differentiation and blocked the initiation of meiosis (Xu et al. 2017), while in Zmettl3m/m spermatocytes, the synaptonemal complex can be assembled and synapsis was normal. However, the proportion of SG and SC stages was significantly higher, yet the proportion of SZ stage decreased. In addition, sperm motility was decreased in Zmettl3m/m males. Therefore, in the Zmettl3m/m males, the decreased levels of m6A mediated by Mettl3 do not affect spermatogonial proliferation and the first meiotic division of spermatocytes, but hinder sperm maturation. This is entirely different from the situation in mice (Hsu et al. 2017; Xu et al. 2017), suggesting Mettl3 and m6A regulate gametogenesis differently in fishes and mammals. In the pituitaries of Zmettl3m/m males, there was a significant decrease in the fshβ mRNA level. The expression of fshr was also significantly reduced in mutant testes. In male zebrafish, fshβ plays an important role in sperm maturation, while lhβ and Lhr seem to be less important (Zhang et al. 2014; Chu et al. 2015). Therefore, the defect in sperm maturation in Zmettl3m/m males is correlated to changes in Fsh signaling.
In zebrafish, the sex steroids 11-KT and E2 have significant regulatory effects on germ cells to promote gametogenesis and gamete maturation (Lubzens et al. 2010; Schulz et al. 2010). Levels of mRNA for the cholesterol transporter star and key enzymes in the steroidogenesis pathway such as 3βhsd and Cyp19a1a were significantly reduced in Zmettl3m/m mutant females. Moreover, star, cyp11a1, cyp17a1, cyp17a2, 3βhsd, and hsd17β3 were all significantly reduced in Zmettl3m/m mutant males. Serum concentrations of 11-KT and E2 were also significantly decreased in Zmettl3m/m females and males. The disruption of gamete maturation can be successfully rescued by injecting the maturation-inducing hormone 17α-20β-DHP in Zmettl3m/m. These data show decreased sex hormone levels also contribute to defects in gamete maturation. We also report that the sex ratio is significantly altered with 85.2% of adult Zmettl3m/m being male. This was not associated with any changes in PGC numbers in embryos. It is reported that m6A can also modulate sex determination by potentiating sex lethal (Sxl) alternative pre-mRNA splicing in Drosophila (Haussmann et al. 2016; Lence et al. 2016; Kan et al. 2017). In zebrafish, Mettl3 and m6A may have an important role in sexual development by regulating the expression of genes important for sex steroid synthesis.
It is reported that reversible m6A methylation can dynamically tune the stability and translation of the target mRNAs by the “erasers” of m6A. In the cytosol, Ythdf1 and Ythdf3 act in concert to affect the translation of their targets by facilitating ribosome loading in HeLa cells (X. Wang et al. 2015; Li et al. 2017), whereas Ythdf2 decreases mRNA stability by recruiting the CCR4-NOT deadenylase complex (X. Wang et al. 2014; Du et al. 2016). In the nucleus, Ythdc1 influences mRNA splicing (Xiao et al. 2016). The differentially expressed genes identified here had many m6A consensus motifs that were mainly located in CDS and 3′UTR regions, which is consistent with m6A-seq results in human and mouse (Dominissini et al. 2012; Meyer et al. 2012). We infer that the decreased m6A level may lead to differential expression of these reproduction-related genes in mettl3 mutant adults. Moreover, the m6A-binding protein Ythdf2 targets genes with the m6A that are involved in phosphorus metabolism, cell cycle processes, and reproduction (Zhao et al. 2017), supporting the relationship between m6A modifications and control of reproduction in zebrafish.
In summary, our results demonstrate that loss of Mettl3, the “writer” for m6A modifications, leads to failed gamete maturation and impaired fertility, likely as the result of decreased m6A levels and disrupted expression of genes important for sex hormone synthesis and gonadotropin signaling. We provided further evidence that these genes are likely the targets of m6A modifications. In contrast to previous studies in mammals and plants, we provide functional in vivo support for the theory that m6A is essential for zebrafish reproduction, highlighting a key role for m6A modification in gamete maturation.
Acknowledgments
We are grateful to Jie Mei (Huazhong Agricultural University) for providing help to test sperm quality parameters and Zhaolan Zhou (University of Pennsylvania School of Medicine) for his valuable discussion and critical reading of the manuscript. This work was supported financially by the National Natural Science Foundation (grant nos. 31325026, 31721005, and 31461163006) and the Chinese Academy of Sciences (grant nos. XDA08010106 and 2016FBZ03). Funding from the University of Ottawa International Research Acceleration Program (to V.L.T. and W.H.) is acknowledged with appreciation.
Author contributions: WH, HX, and CZ designed the research; HX, CZ, XW, JC, and BT developed methods and performed research; XX and MS contributed bioinformatics analysis; WH and ZZ provided reagents; WH, HX, CZ, ZZ, and VLT analyzed data; WH, HX, and VLT wrote the manuscript.
Footnotes
Supplemental material is available online at www.genetics.org/lookup/suppl/doi:10.1534/genetics.117.300574/-/DC1.
Communicating editor: M. Halpern
- Received June 17, 2017.
- Accepted November 29, 2017.
- Copyright © 2018 by the Genetics Society of America
