YTHDF2 is essential for spermatogenesis and fertility by mediating a wave of transcriptional transition in spermatogenic cells

Abstract

The dynamic and reversible regulation roles of m⁶A modification and the characterization of m⁶A readers have provided new insights into spermatogenesis at the post-transcriptional level. YTHDF2, as an m⁶A reader, has been reported to mediate the m⁶A-containing transcript decay during the mouse oocyte maturation, embryonic stem cell differentiation, neural development, and zebrafish maternal-to-zygotic transition. However, the roles of YTHDF2 in mammalian spermatogenesis are uncertain. Here, we generated germ cell-specific Ythdf2 mutants (Ythdf2-vKO) at a C57BL/6J background and demonstrated that YTHDF2 is essential for mouse spermatogenesis and fertility. Ythdf2-vKO provides oligoasthenoteratozoospermia phenotype with increased apoptosis in germ cells. High-throughput RNA-seq analysis showed that a group of mRNAs is upregulated in Ythdf2-vKO mouse testis; further analysis and MeRIP-qPCR data showed that most of the upregulated genes in Ythdf2-vKO mouse testis are modified with m⁶A and are YTHDF2 candidate binding genes. Interestingly, RNA-seq analysis combined with our previous single-cell transcriptomics data of mouse spermatogenesis pointed out the failure of a wave of transcript transition during the spermatogenesis of Ythdf2-vKO mice, which was confirmed by gene expression analysis using qPCR of diplotene spermatocytes and round spermatids obtained through fluorescence-activated cell sorting. Our study demonstrates the fundamental role of YTHDF2 during mouse spermatogenesis and provides a potential candidate for the diagnosis of male infertility with the oligoasthenoteratozoospermia syndrome.

infertility, m⁶A, oligospermia, spermatogenesis, YTHDF2 protein

Introduction

Male infertility is becoming a worldwide health problem [1,2]. Male gamete is generated through spermatogenesis, which is a highly orchestrated development process typically including the following three phases: mitosis, meiosis, and spermiogenesis [3]. During mitosis, spermatogonial stem cells undergo self-renewal or generate differentiating spermatogonia. In the subsequent meiosis, spermatocytes produce haploid round spermatids through one-time replication and two-time division. At the third stage, round spermatids undergo a series of morphological and biochemical changes to generate elongated mature spermatozoa [4,5]. These highly complex processes are regulated by epigenetics that control certain genes to express at different stages. Our previous work and other groups have reported the widely dynamic changes of the transcriptome during mammalian spermatogenesis [6,7]. Moreover, transcription and translation are uncoupled during certain stages of spermiogenesis. The later stage of spermiogenesis is transcriptionally inert [8]. Therefore, the post-transcriptional mechanisms are crucial to understand the spermatogenesis [5,9].

Among over 170 types of different RNA modifications known to date, RNA N⁶-methyladenosine (m⁶A) is the most prominently internal messenger RNA (mRNA) modification [10]. m⁶A modification has been characterized to tune gene expression by post-transcriptionally regulating the metabolism of m⁶A-containing mRNA [11], including mRNA splicing [12], export [12,13], translation [14], and decay [15,16].

The RNA m⁶A modification has been reported to be involved in extensive biological processes, such as DNA damage response [17], mouse oocyte maturation [18], zebrafish maternal-to-zygotic transition [19], embryonic stem cell self-renewal [20], and immune response [21]. In particular, emerging evidence has shown the dynamic and reversible regulatory roles of the m⁶A modification in mammalian spermatogenesis [5,9]. The m⁶A modification is catalyzed by ‘writers’, removed by ‘erasers’, and recognized by its ‘readers’ [11]. We and another group respectively revealed that methyltransferase-like 3 (METTL3) and METTL14 control the fate of spermatogonial stem cells [14,22,23]. The eraser, AlkB homologue 5 (ALKBH5), was reported to modulate the meiotic metaphase-stage spermatocytes and impact the fertility [13]. These reports demonstrated that the dynamic reversible m⁶A modification has fundamental function during mouse spermatogenesis.

The m⁶A readers are mainly referred to the YT521-B homology (YTH) family including five members, i.e. YTH domain family 1 (YTHDF1), YTHDF2, YTHDF3, YTH domain containing 1 (YTHDC1), and YTHDC2 [24,25]. YTHDC1 is required for spermatogonial development [26], and YTHDC2 is indispensable for meiotic prophase I through interacting with MEIOC [27]. YTHDF2, a reader mediating the target mRNA degradation and turnover [16,28], has been reported to be critical for mouse embryonic stem cell differentiation [29], oocyte maturation and early zygotic development [18], and zebrafish maternal-to-zygotic transition [19]. In zebrafish, an average of 70% ythdf2^−/− progeny arrested at the 1-cell stage is likely caused by a defective sperm [19]. YTHDF2 is expressed at high levels in mouse testis [18] and promotes cell adhesion of spermatogonial cells in an in vitro cell model [30]. However, Ythdf2^−/− male mouse showed no defects in spermatogenesis and fertility on a mixed genetic background [18]. So far, the biological roles of YTHDF2 in mammalian spermatogenesis in vivo are uncertain.

In the present study, we generated the germline-specific Ythdf2-knockout mouse (Ythdf2-vKO) at a C57BL/6J background. Our study showed that YTHDF2 is essential for spermatogenesis and fertility. Germline-specific deletion of Ythdf2, although having a complete spermatogenesis process, led to oligoasthenoteratozoospermia (OAT) with increased apoptosis in germ cells. RNA-seq of the testis and qPCR analysis of diplotene spermatocytes and round spermatids showed that a wave of transcripts dynamic transition was unsuccessful, which might be caused by a failure of the degradation of YTHDF2 target mRNA timely in Ythdf2-vKO mouse testis.

Materials and Methods

Mice

Ythdf2-knockout first mice were obtained from the Model Animal Research Center of Nanjing University (Nanjing, China). To knockout the first allele, a promoter-driven cassette (including LacZ and neo genes) flanked by FRT sites was inserted between exon 3 and exon 4, whereas exon 4 of Ythdf2 was flanked by LoxP sites. The knockout first allele was converted into a conditional allele by crossing Ythdf2 knockout first mice with Flp deleter mice. The resulting floxed Ythdf2 mice were crossed with Vasa-GFPCre knock-in mice (generated by Shanghai Biomodel Organism Co., Ltd., Shanghai, China) to allow specific knockout of Ythdf2 in germ cells. All mice were kept in the C57BL/6J background. The primers used for genotyping are listed in Table 1. Ythdf2^flox/∆Vasa-GFPCre (Ythdf2-vKO) was used as the experimental mouse for this study, and Ythdf2^flox/+Vasa-GFPCre (Ythdf2-CTL) was used as a control. All animal experiments were conducted in accordance with the guidelines of the Animal Care and Use Committee at Shanghai Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences (Shanghai, China).

Table 1.

Sequences of primers used in this study

Name	Sequence (5ʹ→3ʹ)	Product size (bp)	Analysis
Ythdf2-KO-R	TCACTGACAGAGGGCTTGCTTAG	403	Genotyping
Ythdf2-GT-tR1	GCTTAACCTTCTGGGCTGCTAG	530	Genotyping
Ythdf2-GT-tF1	GCATCAAATCCCTTGAGACTGGG	–	Genotyping
Vasa-GFPCre-F	GTAGAAAATGTGAGTCTTTGCTGTC	481	Genotyping
Vasa-GFPCre-R	GGTCGGGGTAGCGGCTGA		Genotyping
Tnfrsf12a-F	CAGATCCTCGTGTTGGGATT	162	qPCR
Tnfrsf12a-R	GCAGAAGTCGCTGTGTGGT		qPCR
Syk-F	AAACCGTGGCTGTGAAAATC	154	qPCR
Syk-R	ATCACCAGCATCCAGGACTC		qPCR
Plch2-F	GAGTCGCTGGACTTGGTCTC	163	qPCR
Plch2-R	CGTTCTTGTCAGCCTCATCA		qPCR
Mast1-F	AGAAACCTCCCGGTGAAAGT	195	qPCR
Mast1-R	TCTCTGCAAACGTGAGGATG		qPCR
Tubb3-F	TGAGGCCTCCTCTCACAAGT	237	qPCR
Tubb3-R	TGCAGGCAGTCACAATTCTC		qPCR
Stard8-F	CCGACCTAGCCTCAACTCAG	201	qPCR
Stard8-R	AATACATGGTGTCCCCGGTA		qPCR
Inf2-F	CTGTGCATCTATTCGCCTGA	236	qPCR
Inf2-R	TCCAGCAGCTGTAGCCCTAT		qPCR
Tspan4-F	ACCTCATGTTCGCCTTCAAC	247	qPCR
Tspan4-R	ACTAGCAGCAGCAGCACAAA		qPCR
Sh2d7-F	TCAAAACAAGCCCAATAGCC	163	qPCR
Sh2d7-R	GTGGCTCAGCTTCAGAAACC		qPCR
Kl-F	TTGCAGACTACGCAAACCTG	197	qPCR
Kl-R	CTGCAAGGCGATGGATATTT		qPCR
Prss40-F	AACAGATTCCACACCCAAGG	248	qPCR
Prss40-R	GGTGGTGGGAAAAATCCTTT		qPCR
β-actin-F	GTGACGTTGACATCCGTAAAGA	154	qPCR
β-actin-R	GCCGGACTCATCGTACTCC		qPCR
Tnfrsf12a-F	GACCTCGACAAGTGCATGGA	60	MeRIP-qPCR
Tnfrsf12a-R	GCAGAAGTCGCTGTGTGGTC
Syk-F	TCCTGATGTGGGAAGCGTTC	125	MeRIP-qPCR
Syk-R	CATCTCTCTCGGGCATCCTG
Plch2-F	CCGGGAATACGGCGAGAT	52	MeRIP-qPCR
Plch2-R	AGCCGGTGAGCCAGCGTA
Mast1-F	CTGGGAGCAGACAAGAAGGG	133	MeRIP-qPCR
Mast1-R	CAATAGGAGGCGTCTCACCATC
Inf2-F	GTGAGGGACCAGAGGATGAAGC	160	MeRIP-qPCR
Inf2-R	CCTGAGGAATCTGTCACTGCGT
Tspan4-F	AGGAAGATGTCCAAACCCG	128	MeRIP-qPCR
Tspan4-R	TCTGCTGCCACCCAAGCA
Tubb3-F	CCTGCACTGGTACACGGGC	51	MeRIP-qPCR
Tubb3-R	GCCTCGGTGAACTCCATCTCA
Stard8-F	CAATGCCTTGCCACCTCG	53	MeRIP-qPCR
Stard8-R	CAAGTGCGGAAAGATGAAGC
Sh2d7-F	CTCACTTGAAGGTGTGACCCA	134	MeRIP-qPCR
Sh2d7-R	TGGACAGCATTGAAGAGACATC
Kl-F	TGAGCTCTTACACCATGCTACT	114	MeRIP-qPCR
Kl-R	CTCAGGTGTAGCACAGAGACTT
Prss40-F	AAATCGGGCCTGGGTGGA	83	MeRIP-qPCR
Prss40-R	GCGGTCTGGGCATGCTGT

F: forward; R: reverse.

Isolation of spermatogenic cells

The testes from adult Ythdf2-vKO and control mice were processed through a two-step enzymatic digestion and Hoechst 33342 staining as described previously [14]. Different types of spermatogenic cells (leptotene/zygotene, diplotene spermatocytes, and round spermatids) were isolated and collected by Fluorescence-activated cell sorting (FACS, Becton Dickinson, Franklin Lakes, USA).

Total RNA isolation and quantitative RT-PCR

Total RNA was isolated from testis tissue or spermatogenic cells using Trizol reagent (15596018; Life Technologies, Shanghai, China) according to the manufacturer’s instructions. RNA quality was assessed with NanoDrop 2000 (Thermo Fisher Scientific, Waltham, USA), and complementary DNA was prepared by using ReverTra Ace qPCR RT Master Mix with gDNA Remover (FSQ-301; Toyobo, Osaka, Japan). RT-qPCR was performed on an Eppendorf Mastercycler PCR System using SYBR Green Mix (QPK-201; Toyobo). The primers listed in Table 1 were designed using Primer3 (Version. 0.4.0) and synthesized by Generay Biotech (Shanghai, China). The relative expression level was normalized to β-actin.

Western blot analysis

For western blot analysis, testis tissues were homogenized and lysed in 1× sodium dodecyl sulfate (SDS) loading buffer by boiling for 10 min. SDS-polyacrylamide gel electrophoresis (10%) was used to resolve the proteins, which were then transferred to a nitrocellulose membrane. After being blocked with 5% nonfat milk, the membranes were incubated with anti-YTHDF2 (1:5000; 24744-1-AP; Proteintech, Wuhan, China) overnight at 4°C. Horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG (1:500; 31460; Thermo Fisher Scientific) was used as a secondary antibody. Anti-β-actin (1:3000; AB2001; Abways, Shanghai, China) was used as a control. SuperSignal™ West Pico PLUS Chemiluminescent Substrate (34577; Thermo Fisher Scientific) was used to visualize the protein bands.

Histological analysis and immunofluorescence staining

For histological analysis, testes from Ythdf2-vKO and control mice were fixed in Bouin’s solution, embedded in paraffin, and sectioned. Hematoxylin and eosin (H&E) staining was carried out after the sections were deparaffinized and rehydrated. For immunofluorescence analysis, the testes were fixed in 4% paraformaldehyde (PFA), embedded in paraffin, and sectioned. After deparaffinization and rehydration, the sections were boiled in 10 mM sodium citrate buffer (pH 6.0) for 15 min and washed in phosphate buffered saline (PBS) with 0.1% Triton X-100. The sections were then blocked with blocking buffer (10% donkey serum and 0.1% Triton X-100 in PBS) for 60 min and incubated with primary antibodies overnight at 4°C. The following primary antibodies were used: anti-YTHDF2 (1:100; 24744-1-AP; Proteintech), mouse anti-γH2AX (1:500; 05-636; Millipore), anti-Ki-67 (1:50; 550609; BD Pharmingen, Franklin Lakes, USA), and anti-MVH (1:200; ab13840; Abcam, Cambridge, UK). Alexa Fluor 488- and 594-conjugated secondary antibodies (1:500; 711-545-152/711-585-152; Jackson ImmunoResearch Laboratories, West Grove, USA) were used.

TUNEL analysis

Testes from Ythdf2-vKO and control mice were fixed in 4% PFA, embedded in paraffin, and sectioned. After deparaffinization and rehydration, the sections were boiled in 10 mM sodium citrate buffer (pH 6.0) for 1 min, ddH₂O was added, and the sections were cooled to room temperature. The sections were blocked with blocking buffer (10% donkey serum and 0.1% Triton X-100 in PBS) for 30 min. TUNEL staining was performed using the TUNEL BrightRed Apoptosis Detection kit (A112-02; Vazyme, Nanjing, China) according to the manufacturer’s instructions. The sections were then washed with PBST three times and mounted on slides with mounting medium with DAPI-Aqueous, Fluoroshield (ab104139; Abcam). Images were observed under a fluorescence microscope (Zeiss Axio Scope A1; Carl Zeiss, Jena, Germany).

m⁶A Dot blot assays

The mRNA was purified using GenElute^™ mRNA Miniprep Kit (MRN10, Sigma Aldrich, St Louis, USA) and then loaded on nitrocellulose membranes (Hybond-N⁺, RPN303B; GE Healthcare, Wisconsin, USA). The membranes were UV cross-linked (1200×100 µJ/cm²). After being blocked with 5% nonfat milk, the membranes were incubated with anti-m⁶A (1:2000; ab151230; Abcam) overnight at 4°C. HRP-conjugated goat anti-rabbit IgG (1:500; 31460) was used as the secondary antibody. The staining with 0.05% methylene blue in 0.3 M sodium acetate (pH 5.2) was used to display the loading consistency.

Computer-assisted semen analyzer

Cauda epididymides were separated from the control and Ythdf2-vKO mice and then minced in pre-warmed (37°C) Tyrode’s buffer (Sigma-Aldrich). After being treated for 15 min at 37°C, the tissue was removed, and the sperm suspension assessment including the count and the percentages of motile and progressively motile spermatozoa were calculated by computer-assisted semen analyzer (CASA) (Hamilton Thorne, Selangor D.E., Malaysia).

RNA-Seq

The testes from the control and Ythdf2-vKO mice were homogenized and prepared for RNA extraction. RNA quality was assessed using a 2100 Bioanalyzer (Agilent Technologies, Santa Clara, USA). RNA-Seq libraries were prepared according to the protocols of the NEBNext rRNA Depletion kit (Human/Mouse/Rat) (NEB #E6310; Massachusetts, USA) and NEBNext Ultra II Directional RNA Library Prep kit for Illumina (NEB #E7760). Then, the samples were pooled for deep sequencing on the Illumina HiSeq Xten (2×150) platform at the CAS-MPG Partner Institute for Computational Biology Omics Core (Shanghai, China). The quality of the raw reads was evaluated with FastQC (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/, v0.11.5). Adaptor sequences and read sequences at both ends with Phred quality scores below 30 were trimmed. The trimmed reads were then mapped with the Hisat2 algorithm (Hisat2 v2.1.0) to target sequences. Gene expression levels were quantified by the software package HTSeq (v0.6.1p1). Differentially expressed genes were generated by using the criteria of a fold change >1.5 with FDR <0.1. The significantly differentially expressed genes were subsequently analyzed for the enrichment of biological themes with the DAVID 6.8 bioinformatics platform (https://david.ncifcrf.gov/). RNA-seq data from this study has been deposited in the GEO (GSE147574) and NODE (OEP000805, https://www.biosino.org/node/project/detail/OEP000805) database.

MeRIP-qPCR

Purified mRNAs for the MeRIP assay were prepared using the Arraystar Seq-Star TM poly (A) mRNA Isolation kit (AS-MB-006-01/02; Arraystar, Shanghai, China) and then randomly fragmented into ∼100 nucleotides. After setting aside a fraction of mRNAs as an input, the remainder was immunoprecipitated using an affinity purified anti-m⁶A rabbit polyclonal antibody (202003; Synaptic Systems, Shanghai, China) and DynabeadsTM M-280 Sheep anti-rabbit IgG (11203D; Invitrogen, Shanghai, China). The primers for qPCR analysis are listed in Table 1. The relative m⁶A enrichment of mRNAs was normalized to the input as follows:

Fold enrichment = 2^{[IP (Ct Input-Ct MeRIP)–IgG (Ct Input-Ct MeRIP)]}

Statistical analysis

Student’s t test was used for statistical analysis using GraphPad Prism software (Version 5.01). Data are presented as the mean±SEM. P<0.05 was considered as statistically significant.

Results

Generation of the germ cell-specific deletion of Ythdf2

First, we investigated the expression level of YTHDF2 at different developmental stages. Testis lysates from 3, 7, 14, 21, 28, and 56 days postpartum (dpp) were prepared for western blot analysis. The results showed that the YTHDF2 protein was at a low or undetectable level at 3 and 7 dpp stages but dramatically increased from 14 dpp and maintained a high level afterward (Fig. 1A). Given that mitosis-to-meiosis transition starts between 7 and 14 dpp, we speculated that YTHDF2 might play role(s) in the process of meiosis and/or spermiogenesis, rather than in mitosis.

Figure 1.

Generation of the germ cell-specific deletion of Ythdf2   (A) Representative western blot results of the YTHDF2 protein in lysates from wild-type testes at different developmental stages. β-Actin was used as a loading control. dpp: days postpartum. (B) A representative western blot results of the YTHDF2 protein in L/Z spermatogenic cells isolated from 2-month-old control and age-matched Ythdf2-vKO mice. β-Actin was used as a loading control. L/Z: leptotene/zygotene. (C) The efficiency of germ cell-specific deletion of Ythdf2 is shown by immunofluorescence staining of YTHDF2 (red), the leptotene and zygotene spermatocyte marker γH2AX (green) and DAPI (blue) in sections of 8-week-old control and Ythdf2-vKO mice testes. Scale bar=40 μm.

To explore the biological roles of YTHDF2 in mouse spermatogenesis, we obtained a Ythdf2-floxed line in which exon 4 of the Ythdf2 allele is flanked by loxP sites. By crossing with Vasa-GFPCre mice, the fourth exon of Ythdf2 was specifically deleted in male germ cells as early as embryonic day 15 (E15) (Supplementary Fig. S1A). Four genotypes were generated, including Ythdf2^flox/∆Vasa-GFPCre (hereafter referred to as Ythdf2-vKO), Ythdf2^flox/∆, Ythdf2^flox/+Vasa-GFPCre, and Ythdf2^flox/+. The littermate Ythdf2^flox/+Vasa-GFPCre was used as a control (referred to as Ythdf2-CTL). Both Ythdf2-vKO and Ythdf2-CTL mice were healthy and grew normally into adulthood. Consistent with Ivanova’s report [18], our immunostaining results also showed that YTHDF2 was expressed at each stage of spermatogenesis in control mice, while the YTHDF2 signal was absent in germ cells of Ythdf2-vKO mice (Fig. 1B,C), not in the total testes (Supplementary Fig. S1B), confirming the germ-cell specific mutation of Ythdf2. These data showed that Ythdf2 is dynamically expressed during mouse spermatogenesis, and germ cell-specific Ythdf2 mutants functioned well.

YTHDF2 is required for male fertility and spermatogenesis

To determine the fertility of Ythdf2-vKO mice, mating tests of Ythdf2-vKO male mice with wild-type C57BL/6J (B6) background female mice were carried out. Our data showed that Ythdf2-vKO mice were completely sterile (Fig. 2A). However, the testis size, testicular weight, and body weight of Ythdf2-vKO mice were not significantly different from those of Ythdf2-CTL mice (Fig. 2B–D).

Figure 2.

YTHDF2 is required for male fertility and spermatogenesis   (A) The infertility of germ cell-specific Ythdf2 mutants was shown through a mating assay. At 2 months of age, control (n=2) and Ythdf2-vKO mutant (n=4) mice cohabited with two female C57BL/6 mice. Vaginal plugs were checked, and pups were counted. ***P<0.001. (B) Gross morphology of representative testes from 2-month-old control and age-matched Ythdf2-vKO mice. (C,D) Comparison of testis weight (C) and body weight (D) from 2-month-old controls and age-matched Ythdf2-vKO mice. Data are presented as the mean±SEM (n=3 for each genotype). (E) H&E staining of the testes and epididymis from 2-month-old control and age-matched Ythdf2-vKO mice. Spg: spermatogonia; Spc: spermatocytes; Rs: round spermatids; Es: elongating spermatids; S: sperm. Scale bar=40 μm. (F) The sperm counts and the percentages of motile and progressively motile spermatozoa in the cauda epididymides of 2-month-old control and age-matched Ythdf2-vKO mice. Data are presented as the mean±SEM (n=5 for each genotype). ***P<0.001. (G) The morphology of representative sperm from 2-month-old control and age-matched Ythdf2-vKO mice. The red arrow indicates an abnormal head sperm. Scale bar=40 μm. (H) Statistical analysis of normal sperm from 2-month-old control and age-matched Ythdf2-vKO mice. Data are presented as the mean±SEM (n=3 for each genotype, at least 100 sperm were counted for each genotype). ***P<0.001.

To characterize the spermatogenesis of Ythdf2-vKO mice, histological analysis was performed, and the results showed that spermatogenic cells arranged normally in seminiferous tubules of Ythdf2-vKO and Ythdf2-CTL adult mice, indicating that Ythdf2-vKO mice are capable of completing the spermatogenesis process (Fig. 2E). However, the epididymides of Ythdf2-vKO mice contained much fewer mature spermatozoa than those of Ythdf2-CTL mice and were filled with a high number of degenerated germ cells (Fig. 2E). Consistent with this, CASA results showed that the sperm counts and the percentages of motile and progressively motile spermatozoa in Ythdf2-vKO mouse cauda epididymides were significantly lower than those in Ythdf2-CTL cauda epididymides (Fig. 2F). In addition, >90% of the spermatozoa in Ythdf2-vKO mouse cauda epididymides were malformed with abnormal heads (Fig. 2G,H). Collectively, these data indicated that Ythdf2-vKO mutant mice displayed an OAT phenotype, suggesting that YTHDF2 is required for mouse spermatogenesis and fertility.

Germ cell-specific Ythdf2 deletion causes wide apoptosis during spermatogenesis

Based on the above results, Ythdf2-vKO mice contained all types of spermatogenic cells in seminiferous tubules and degenerated germ cells in epididymides, so we hypothesized that apoptosis occurred in seminiferous tubules of Ythdf2-vKO mice. The TUNEL assay demonstrated that widespread apoptosis occurred in the seminiferous tubules of Ythdf2-vKO mice (Fig. 3). To analyze the proliferation level of Ythdf2-vKO mouse testis, immunofluorescence staining of Ki-67 was carried out, and no obvious change of proliferation was found in Ythdf2-vKO mouse testis compared with the control (Supplementary Fig. S2). Thus, the increase of apoptosis in Ythdf2-vKO mouse testis might contribute to the degenerated germ cells in the epididymides.

Figure 3.

Germ cell-specific Ythdf2 deletion causes severe apoptosis during spermatogenesis   (A) Representative TUNEL staining of testes from 2-month-old controls and age-matched Ythdf2-vKO mice. Staining of apoptotic cells with TUNEL (green) and the nucleus with DAPI (blue). Scale bar=40 μm. (B) Quantification of TUNEL-positive cells in the seminiferous tubules from 2-month-old controls and age-matched Ythdf2-vKO mice. Data are presented as the mean±SEM (n=3 for each genotype; tubules examined: 185 for control mice and 205 for Ythdf2-vKO mice). ***P<0.001.

YTHDF2 regulates the dynamic mRNA level during spermatogenesis

Given the reported molecular role of YTHDF2 in mRNA stability, we conducted RNA-seq on testes from Ythdf2-vKO and Ythdf2-CTL adult mice to investigate the biological mechanism underlying the defects in Ythdf2-vKO mouse spermatogenesis. Compared with Ythdf2-CTL testes, 266 genes were identified to be dysregulated in Ythdf2-vKO mouse testis (196 upregulated and 70 downregulated) (FDR < 0.1, fold change >1.5) (Fig. 4A and Supplementary Table S1A). The RNA-seq data were validated by qPCR analysis of stage-specific expressed genes (Fig. 4B). Several biological processes were affected and enriched for the upregulated genes in Ythdf2-vKO mice (Fig. 4C and Supplementary Table S1B). According to our previous report of m⁶A mRNA methylomes of mouse spermatogenic cells [14], 66% of the upregulated genes (129 genes) and 51% of the downregulated genes (36 genes) in Ythdf2-vKO mouse testis were modified with m⁶A (Fig. 4D). The distributions of m⁶A in the 129 upregulated and 36 downregulated genes at the 3ʹUTR, stop codon, and CDS in the wild-type transcriptome were 74% and 58%, respectively (Fig. 4D). Furthermore, MeRIP-qPCR analysis validated the enrichment of m⁶A in the selected genes (Fig. 4E).

Figure 4.

Germ cell-specific Ythdf2 deletion alters the expressions of specific genes modified with m⁶A during spermatogenesis   (A) Scatter plot showing the fold change of transcripts of testis tissue between 2-month-old controls and age-matched Ythdf2-vKO mice. Significantly downregulated and upregulated genes with a fold change greater than 1.5 (FDR<0.1) are shown in blue and red, respectively. Analysis was performed on two biological replicates. (B) qPCR validation of the dysregulated genes in Ythdf2-vKO mouse spermatogenesis. Genes used for analysis are expressed at the following specific stages: Un.S, undifferentiated spermatogonia; A1, type A1 spermatogonia; Prel, preleptotene spermatocytes; DS, diplotene spermatocytes; RS, round spermatids. Data are presented as the mean± SEM (n=3 for each genotype). ***P<0.001. (C) Gene ontology enrichment analysis of the upregulated genes in Ythdf2-vKO mice. The top 10 most enriched biological processes are shown. (D) The ratio of transcripts modified with m⁶A among the dysregulated genes in Ythdf2-vKO mice and the distribution of m⁶A sites were analyzed according to our previously reported m⁶A mRNA methylomes of mouse spermaogenic cells [14]. (E) MeRIP-qPCR analysis using anti-IgG and anti-m⁶A antibodies showed m⁶A enrichment in Tnfrsf12a, Syk, Plch2, Mast1, Inf2, Tspan4, Tubb3, Stard8, Sh2d7, Kl, and Prss40 in adult mouse testis. The relative m⁶A enrichment was normalized to the input. ***P<0.001.

Notably, the 129 upregulated genes with m⁶A modification in Ythdf2-vKO mouse testis were found to be mainly distributed at early stages of spermatogenesis in our previously report [14] (Fig. 5A). To confirm this trend, we further analyzed the precise distribution of the 129 upregulated genes at the single-cell level using our published single-cell transcriptomics data of mouse spermatogenesis [7] (Fig. 5B). These two analyses consistently indicate that the 129 upregulated genes with the m⁶A modification in Ythdf2-vKO mouse testis are mainly expressed at early stages of wild-type mouse spermatogenesis. The 70 downregulated genes in Ythdf2-vKO mouse testis are mainly present at later stages of wild-type mouse spermatogenesis and begin to be expressed in diplotene spermatocytes of wild-type mice (Fig. 5B). Furthermore, dot blot analysis showed that the m⁶A level increased in testis after ythdf2 depletion (Fig. 5C), which proved the failure of YTHDF2 target genes decay in Ythdf2-vKO mouse testis. To verify the expression changes of the differentially expressed genes in Ythdf2-vKO mice, we performed qPCR analysis in two types of germ cells, i.e. diplotene spermatocytes and round spermatids, collected by fluorescence-activated cell sorting from Ythdf2-vKO and Ythdf2-CTL mice testes. The qPCR analysis results confirmed the failure of a wave of transcriptome dynamic transition during mouse spermatogenesis in Ythdf2-vKO mice (Fig. 5D).

Figure 5.

Germ cell-specific Ythdf2 deletion changes the dynamic expressions of a wave of stage specific genes during spermatogenesis   (A) Mapping the upregulated genes modified with m⁶A to our previous data of mouse spermatogenesis [14]. (B) Heatmap analysis of the wild-type expression level of the upregulated genes modified with m⁶A and downregulated genes in Ythdf2-vKO mice based on our previously published paper [7]. A1, type A1 spermatogonia; In, intermediate spermatogonia; TypeBS, S phase type B spermatogonia; TypeBG2M, G2/M phase type B spermatogonia; G1, G1 phase preleptotene; ePL, early S phase preleptotene; mPL, middle S phase preleptotene; lPL, late S phase preleptotene; L, leptotene; Z, zygotene; eP, early pachytene; mP, middle pachytene; lP, late pachytene; D, diplotene; MI, metaphase I; MII, metaphase II; RS1o2, steps 1–2 spermatids; RS3o4, steps 3–4 spermatids; RS5o6, steps 5–6 spermatids; RS7o8, steps 7–8 spermatids. (C) Dot blot assay was used to analyze the total m⁶A level of Ythdf2-vKO mice testes. A total of 50 ng mRNA was loaded for each point. MB: methylene blue. (D) The changes in the dynamic expression pattern of genes during spermatogenesis of germ cell-specific Ythdf2 mutants were confirmed via qPCR analysis of two stages of spermatogenic cells (diplotene spermatocytes and round spermatids) collected by FACS. DS, diplotene spermatocytes; RS, round spermatids. Data are presented as the mean±SEM (n=3 for each genotype). ***P<0.001.

Collectively, these results showed that YTHDF2 is required for the degradation of a group of target transcripts modified with m⁶A and a wave of transcriptome dynamic transition during mouse spermatogenesis.

Discussion

Emerging evidence has shown the essential roles of m⁶A modification in mammalian spermatogenesis, particularly the characterization of m⁶A readers, which has provided new insights into post-transcriptional mechanisms in spermatogenesis [5,26,27]. However, the biological roles of YTHDF2 in mammalian spermatogenesis are uncertain. In this study, we generated a germ-cell specific Ythdf2-knockout mouse at a C57BL/6J background (Fig. 1). Our results demonstrate that YTHDF2 is critical for spermatogenesis and fertility (Fig. 2). Deletion of Ythdf2 led to OAT and increased apoptosis in germ cells (Fig. 3). High-throughput RNA-seq of the testis tissue and qPCR analysis of diplotene spermatocytes and round spermatids confirmed that the degradation of a wave of YTHDF2 target mRNA failed, and the expressions of the downstream subset genes were affected in Ythdf2-vKO mouse testis (Figs. 4 and 5).

Based on our results of the dynamic expression of YTHDF2 at different developmental stages (Fig. 1A), the reported special highest expression level of YTHDF2 in mouse testis tissue [18], and the reported clue of sperm defects in ythdf2^-/- zebrafish [19], we proposed that YTHDF2 might play critical roles in mammalian spermatogenesis. As expected, our study using the germ-cell specific knockout mouse Ythdf2-vKO showed that YTHDF2 is required for mouse spermatogenesis and fertility at the C57BL/6J background (Figs. 2 and 3). Notably, a recent study reported a similar phenotype that Ythdf2-knockout mice established by CRISPR–Cas9 showed hypofertility with mild degenerative signals in the seminiferous tubules and severe loss of sperm in the cauda epididymis [23]. Although impaired sperm maturation was confirmed by measuring gene expression in Ythdf2-KO mouse round spermatids, the related mechanism was not clarified [23]. Moreover, the apoptosis characteristic of spermatocytes and sperm after depletion of Ythdf2 needs to be explored in further study.

The high complex of spermatogenesis manifests the dynamic transcriptome and m⁶A RNA methylomes of different stage spermatogenic cells to timely tune mRNA translation and decay at the post-transcriptional level [5,14]. The installation of m⁶A on transcripts could mark them for recognition by readers. In accordance with the regulation role of YTHDF2 in mRNA stability, the dysregulated genes were mainly upregulated in Ythdf2-vKO mouse testis (Fig. 4A), most of which were modified with m⁶A (Fig. 4D,E), a natural consequence of the deletion of Ythdf2. Although cell apoptosis was not indicated as being involved according to GO analysis of the upregulated genes in Ythdf2-vKO mice, apoptosis-related genes were present, such as Tnfrsf12a and Jun (Fig. 4B and Supplementary Table S1). Among the tumor necrosis factor receptor superfamily (TNFRSF), Tnfrsf12a (also known as Fn14 and TWEAK) was reported as a weak inducer of apoptosis [31] and is involved in cell proliferation and the cell death process depending on different cells [32,33]. Tnfrsf12a is expressed at a lower level in mouse testis compared with that in other tissues [34]. Our RNA-seq data showed that germ-cell-specific depletion of ythdf2 led to the upregulation of Tnfrsf12a (Fig. 4B and Supplementary Table S1). In Ythdf2-vKO mouse testis, cell proliferation was not affected (Supplementary Fig. S1), while the apoptosis signals significantly increased compared to those in the control (Fig. 3). Therefore, how the expression of Tnfrsf12a is regulated and whether the upregulation of Tnfrsf12a induces apoptosis in Ythdf2-vKO mouse testis are worthy of further detailed study.

Like YTHDC2 [27], only a small part of transcripts among a large number of genes marked with m⁶A was targeted by different readers to accelerate decay. This phenomenon might be caused by the compensation mechanism of the other YTHDF paralogs [23,35]. In addition, there might exist more readers or more coordinating mechanisms of m⁶A-containing transcripts at the post-transcriptional level, which are worthy for investigation in the future.

Interestingly, by mapping the dysregulated transcripts Ythdf2-vKO in our previously reported single-cell transcriptomics of mouse spermatogenesis [7], we found that there was a wave of transcripts transition failure (Fig. 5). Our data and other studies in embryonic stem cell [29], oocyte mature [18], neural development [15], and zebrafish maternal-to-zygotic transition [19] together supported the hypothesis that YTHDF2 plays a crucial role in different development transiting stages through recognition and clearance of the no longer necessary targets, giving road to express genes required in the subsequent development stages. Because the amount of RNAs obtained from germ cells was not enough to perform RNA-seq, we confirmed the transcript transition failure in Ythdf2-vKO by qPCR analysis of diplotene spermatocytes and round spermatids collected by FACS (Fig. 5D). Meanwhile, m⁶A modification was reported to tune chromatin state and transcription [36]. Therefore, whether the increase of m⁶A modification in Ythdf2-vKO mouse testis (Fig. 5C) affects the chromatin state and transcription level should be further explored.

In summary, our study showed that YTHDF2 is required for mouse spermatogenesis and fertility at the C57BL/6J background. YTHDF2 mediates a wave of transcript transition in seminiferous tubule of Ythdf2-vKO mice. Given the existence of different types of germ cells in seminiferous tubule, the explanation for the dysregulation of the subset of gene expression and the increase of apoptosis in Ythdf2-vKO mouse testis requires more detailed study in the future.

Supplementary Data

Supplementary Data is available at Acta Biochimica et Biophysica Sinica online.

Funding

This work was supported by the grants from the National Key R&D Program of China (No. 2018YFC1003000) and the National Natural Science Foundation of China (Nos. 81871163 and 81801527).

Conflict of Interest

The authors declare that they have no conflict of interest.

Author notes

†

Xinxi Zhao, Zhen Lin, Yong Fan and Wenzhi Li contributed equally to this work.