O-GlcNAc transferase governs spermatogenic mitotic-to-meiotic transition and progression by coordinating transcription and alternative splicing programs

Abstract

Background

O-GlcNAcylation is a post-translational modification (PTM) uniquely catalyzed by O-GlcNAc transferase (OGT), which has been linked to tumorigenesis and neurodegeneration. However, its roles in mammalian spermatogenesis remain unexplored. This study aims to elucidate the functional mechanisms of OGT in spermatogenesis and male fertility.

Methods

We employed immunoprecipitation-mass spectrometry (IP-MS) to identify candidate O-GlcNAcylated substrates of OGT in juvenile mouse testes. To explore the physiological roles of OGT and O-GlcNAcylation, we constructed a mouse model with postnatal germ cell-specific deletion of Ogt via Stra8-Cre. In addition, we performed integrated bulk and single-cell RNA sequencing analyses to investigate the potential mechanisms by which OGT and O-GlcNAcylation deficiency impairs spermatogenesis.

Results

The results showed stage-specific OGT enrichment and O-GlcNAcylation in mouse testicular spermatogonia and early spermatocytes. Furthermore, OGT was found to interact with and O-GlcNAcylate transcription factors (e.g., HCFC1) as well as splicing regulators (e.g., SRSF1 and SF3B3) in mouse testes. Postnatal germ cell-specific Ogt deletion impaired spermatogonial differentiation, disrupted meiotic initiation and progression, and induced apoptosis, ultimately leading to male infertility. Mechanistically, Bulk RNA sequencing (RNA-seq) analysis revealed that OGT deficiency dysregulated transcriptional and alternative splicing programs, affecting genes critical for the mitotic–meiotic transition (e.g., Ythdc2 and Rbm46) and meiotic progression (e.g., Stra8, Stag3, and Syce2) in the testes. Single-cell RNA sequencing further uncovered aberrant retention of mitotic transcripts (e.g., Ccna2 and Ccnb1) in spermatocytes and impaired mRNA metabolism during spermatogonial differentiation. In addition, OGT deficiency caused cytoplasmic mislocalization and reduced expression of core transcription factors and splicing regulators in spermatocytes.

Conclusions

These findings establish that OGT and its mediated O-GlcNAcylation coordinate essential gene expression and mRNA metabolism during mitotic-to-meiotic transition and meiotic progression. Moreover, our study provides mechanistic insights into the pathogenesis of male infertility associated with O-GlcNAcylation dysregulation.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s11658-026-00972-2.

Background

Spermatogenesis is a complex and tightly regulated biological process that encompasses spermatogonial proliferation/differentiation, meiotic spermatocyte division, and spermiogenesis. This continuum is demarcated by stage-specific molecular markers: PLZF​ maintains the undifferentiated spermatogonial pool [1]; c-KIT signals commitment to differentiation [2]; STRA8, induced by retinoic acid, is essential for meiotic entry [3]; SYCP3 is a structural component of the synaptonemal complex during meiotic prophase I [4]; and DDX4/VASA is broadly expressed across germ cell stages [5]. Meiosis serves as the core event for haploid gamete formation through DNA replication, followed by two successive divisions [6]. The mitotic-to-meiotic transition is a germ-cell-specialized event that is driven by retinoic acid (RA) signaling via activation of STRA8 and MEIOSIN [7, 8]. Although the transcription factor STRA8 and its interactors are essential for meiotic entry, STRA8-mediated transcriptional activation of downstream target meiotic genes is not sufficient to activate this process, underscoring multi-level regulatory control over meiotic initiation [9]. Among these, post-transcriptional mechanisms, such as RNA modification and translational control, play crucial roles. For instance, the prevalent m6A modification dynamically regulates the stability and translation of meiosis-associated transcripts. The m6A reader YTHDC2 together with MEIOC/RBM46 post-transcriptionally repress mitotic genes, thereby enforcing meiotic entry [9–12]. Likewise, chromatin remodeling and histone modifications contribute to stage-specific gene expression programs during this switch [13, 14]. Despite advances in understanding these regulatory networks, the specific contributions of post-translational modifications (PTMs) to meiotic initiation remain poorly defined, underscoring an unresolved gap in reproductive biology with potential therapeutic relevance.

O-linked β-N-acetylglucosamine modification (O-GlcNAcylation) is a pivotal PTM that orchestrates essential cellular processes across eukaryotes. This modification is synthesized through the hexosamine biosynthesis pathway (HBP), a metabolic network that combines glucose, amino acids, fatty acids, and nucleotides to yield uridine diphosphate N-acetylglucosamine (UDP-GlcNAc), an indispensable substrate for O-GlcNAcylation [15]. O-GlcNAcylation is distinct from classical N- and O-glycosylation in that it involves monosaccharide modification without further glycan chain formation. This process occurs through a tightly controlled enzymatic cycle involving ​​O-GlcNAc transferase (OGT)​​ for addition and ​​O-GlcNAcase (OGA)​​ for removal ​[16, 17]. Moreover, unlike secretory or membrane-bound glycans, O-GlcNAc modification is predominantly localized to nuclear, cytoplasmic, and mitochondrial proteins. This binary enzymatic regulation enables precise control over critical cellular processes, such as transcriptional regulation, cell cycle control, and stress response modulation [18]. As a ubiquitous PTM, O-GlcNAcylation is distributed across virtually all human tissues and dynamically regulates cellular functions through > 5000 identified O-GlcNAc-modified proteins [19]. Perturbations in homeostasis are mechanistically linked to major human pathologies, including cancer progression, diabetes-associated metabolic dysfunction, and neurodegeneration [20–23]. Nevertheless, the functional contributions and mechanistic basis of OGT and O-GlcNAc signaling in male gametogenesis remain unresolved. A recent study by Qian et al. demonstrated a specific enrichment of O-GlcNAc in murine spermatogonia and early spermatocytes. Pharmacological O-GlcNAcase (OGA) inhibition via thiamet-G administration increases testicular O-GlcNAc levels, mechanistically linking its hyperaccumulation to meiotic defects in germ cells [24]. Furthermore, the first comprehensive mapping of the murine testicular O-GalNAcylated glycoproteome was achieved, establishing a critical reference framework for investigating O-glycosylation-mediated regulatory networks in spermiogenesis and fertilization competence [25]. However, the O-GlcNAc-modified substrates in mammalian testes and their in vivo regulatory functions during spermatogenesis remain largely elusive.

The X-linked OGT encodes the sole O-GlcNAc glycosyltransferase (OGT), an evolutionarily conserved enzyme that is ubiquitously expressed across species from Caenorhabditis elegans to humans and mediates essential protein O-GlcNAcylation modifications [26]. OGT generates three isoforms through alternative splicing, all of which share a C-terminal catalytic domain essential for glycosyltransferase activity and N-terminal tetratricopeptide repeat (TPR) domains responsible for substrate targeting and protein–protein interaction [27, 28]. Despite this conservation, the role of OGT in germline development remains poorly defined, largely because of the embryonic lethality observed in OGT-null models of Drosophila, zebrafish, and mice [26, 29, 30]. Notably, OGT depletion in C. elegans causes defects in male subfertility and spermatogenesis without embryonic lethality [31]. In mammals, OGT is indispensable for early development, as evidenced by its requirement for embryonic stem cell viability and ontogeny in mice [26]. Furthermore, conditional Ogt knockout in cerebellar granule neuron precursors (GNPs) triggered structural abnormalities (reduced cerebellar size and aberrant synaptic connectivity) and behavioral deficits in adult male mice, highlighting the pleiotropic role of OGT and O-GlcNAc in neural development [32]. Genetic screening of clinical patients identified hemizygous OGT mutations within the TPR or catalytic domains, variants directly associated with X-linked intellectual disability (XLID) and developmental delay [33, 34]. Notably, affected males carrying the OGT (c.759G > T; p.Leu254Phe) missense mutation ​​concomitantly present​​ with microorchidism and hypospadias [33], suggesting ​​that OGT deficiency contributes to reproductive development defects.

In this study, germ cell-specific Ogt ablation (Stra8-cre) revealed that O-GlcNAcylation deficiency disrupts spermatogonial differentiation and arrests meiotic transition, culminating in complete male infertility. Bulk RNA sequencing (RNA-seq) revealed widespread transcriptomic perturbations in Ogt^flox/Y; Stra8-Cre conditional knockout (referred to as Ogt cKO hereafter) testes, characterized by dysregulated genes involved in DNA replication, meiotic progression, and mitotic-to-meiotic transition, along with defective alternative splicing. Complementary single cell (sc)RNA-seq further identified OGT’s essential role in coordinating mRNA metabolism during spermatogonial differentiation and terminating mitotic programs to ensure faithful meiotic entry. Strikingly, OGT-mediated O-GlcNAcylation directly modifies core transcriptional regulators and splicing factors. Deficiency in this modification causes decreased expression and/or aberrant subcellular localization of these factors in spermatocytes, thereby disrupting their functions. Collectively, our study establishes testicular OGT and O-GlcNAcylation as a key regulator of meiotic initiation, elucidates the molecular mechanism of programmed gene expression during mitotic-meiotic transition, and provides a theoretical basis for precise diagnosis and targeted therapy of male infertility caused by O-GlcNAcylation dysregulation.

Materials and methods

Ogt mutant mice generation and ethical approval

The Ogt^flox/+ and Stra8-Cre mice were generated by Cyagen Biosciences. The Ogt gene (NCBI reference sequence: NM_139144.4) is located on the mouse chromosome X. For the targeted allele, loxP sites were inserted into introns 1 and 3 of Ogt gene to generate Ogt^flox/+ mice. Ogt^flox/+ female mice were crossed with Ogt^flox/Y male mice to generate Ogt^flox/flox female mice. Ogt^flox/flox females were crossed with Stra8-Cre males to obtain Ogt^flox/Y Stra8-Cre conditional knockout (Ogt^cKO) mice. The exon 2–3 of Ogt contains a 425 bp coding sequence, and deletion of this region results in loss of function of the mouse Ogt gene. All offspring were genotyped by PCR analysis of tail genomic DNA. Ogt wild-type (186 bp) and floxed (253 bp) were assayed by PCR with primers 5′CATCTTGTTTATCAGCTGCTTGCT3′ and 5′TCTCACATTTCAAGGTGGAGTGTA3′. The KO allele (424 bp) were confirmed by PCR using primers 5′CATCTTGTTTATCAGCTGCTTGCT3′ and 5′TACTGAGGGGAGGTATTCAGAACG3′. The Stra8-Cre allele (350 bp) was assessed by PCR with primers 5′GCTGGAGTTTCAATACCGGAGATC3′ and 5′CCTTGTATTCGACAAGCCCCAAA3′. All mice were maintained and used in accordance with the guidelines of the Institutional Animal Care and Use Committee of Anhui Medical University (approval no. LLSC20211392).

Histological analysis

Testicular and epididymal tissues from mice of various age groups were fixed overnight in Bouin’s solution (Sigma-Aldrich, SLBJ3855V), followed by sequential processing using a graded ethanol series (70%, 80%, 90%, and 100%) for dehydration. After xylene clearing and paraffin embedding, 5-μm thick sections were prepared using a rotary microtome. The sections were subjected to standard deparaffinization and rehydration procedures. Tissue morphology was subsequently examined by hematoxylin and eosin (H&E) staining, following established protocols.

Immunofluorescence, TUNEL, and surface nuclear spread analyses

For immunofluorescence analysis, the testes were fixed in 4% paraformaldehyde at 4 ˚C overnight, dehydrated in sucrose, embedded in OCT, and cryosectioned into 5-μm thick sections. Cryosections were subjected to antigen retrieval, followed by three washes with phosphate-buffered saline (PBS). After blocking, the sections were incubated sequentially with primary antibodies and corresponding secondary antibodies and mounted with 4′,6-diamidino-2-phenylindole (DAPI) (Sigma-Aldrich, F6057). In terms of TUNEL analysis, apoptotic cells were detected using the TUNEL Enzyme and Label Kit (Roche, 11,684,795,910) according to the manufacturer’s protocol. Nuclear spreads of spermatocytes were prepared using established methods [35]. The primary antibodies used for immunofluorescence and chromosome spread analyses were as follows: anti-OGT (Abcam, ab96718, 1:100, UK), anti-O-GlcNAc (Invitrogen, MA1-072, 1:100, USA), anti-OGA (Proteintech, 14,711–1-AP, 1:100, China), anti-SYCP1 (Abcam, ab15090, 1:100, UK), anti-SYCP3 (Abcam, ab97672, 1:200, UK), anti-γH2AX (Millipore, 05–636, 1:500, USA), anti-PLZF (R&D Systems, AF2944, 1:200, USA), anti-c-KIT (R&D Systems, AF1356, 1:200, USA), anti-SOX9 (Sigma-Aldrich, AB5535, 1:200, USA), anti-DDX4 (Abcam, ab27591, 1:200, UK), anti-STRA8 (Abcam, ab49602, 1:100, UK), anti-YTHDC2 (Abcam, ab176846, 1:100, UK), anti-RBM46 (ABclonal, A16605, 1:100, China), anti-CCNA2 (Abcam, ab181591, 1:200, UK), anti-pHH3 (Cell Signaling Technology, 9701 s, 1:200, USA), anti-SF3B3 (Proteintech, 14,577–1-AP, 1:100, China), anti-SRSF1 (Proteintech, 12,929–2-AP, 1:100, China), anti-ZC3H14 (Proteintech, 28,190–1-AP, 1:100, China), and anti-HCFC1 (Cell Signaling Technology, 69,690, 1:100, USA). Confocal images were acquired using a Zeiss LSM 800 microscope (Carl Zeiss AG, Jena, Germany).

Western blotting

Testicular tissues were washed with PBS and homogenized in ice-cold RIPA lysis buffer (Beyotime Institute of Biotechnology) supplemented with a protease inhibitor cocktail. Following 1 h rotational incubation at 4 °C, lysates were centrifuged (11,000 × g, 15 min, 4 °C) to obtain clarified supernatants. Protein quantification was performed using a bicinchoninic acid (BCA) assay, and 20–30 μg of proteins were separated using sodium dodecylsulphate–polyacrylamide gel electrophoresis (SDS–PAGE) and transferred onto polyvinylidene fluoride (PVDF) membranes and immunoblotted with indicated antibodies as follows: anti-OGT (Abcam, ab96718, 1:2000, UK), anti-O-GlcNAc (Invitrogen, MA1-072, 1:2000, USA), anti-PLZF (R&D Systems, AF2944, 1:2000, USA), anti-STRA8 (Abcam, ab49602, 1:2000, UK), anti-YTHDC2 (Abcam, ab176846, 1:2000, UK), anti-RBM46 (ABclonal, A16605, 1:2000, China), anti-SF3B3 (Proteintech, 14,577–1-AP, 1:2000, China), anti-SRSF1 (Proteintech, 12,929–2-AP, 1:2000, China), anti-ZC3H14 (Proteintech, 28,190–1-AP, 1:2000, China), anti-HCFC1 (Cell Signaling Technology, 69,690, 1:2000, USA), anti-THOC1 (Proteintech, 10,920–1-AP, 1:2000, China), anti-THOC5 (Proteintech, 14,862–1-AP, 1:2000, China), anti-GAPDH (Novus Biologicals, NBP2-37,828, 1:5000, USA), and anti-ACTIN (Sigma-Aldrich, A5441, 1:5000, USA).

Co-immunoprecipitation and mass spectrometry (MS)

Testicular tissues (100 mg) were lysed in 1 mL ice-cold RIPA buffer (10 mM Tris, pH 8.0, 140 mM NaCl, 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS, and 1 mM EDTA) containing 1 mM PMSF protease inhibitor. After 1 h rocking incubation at 4 °C, lysates were centrifuged at 16,100 × g for 30 min at 4 °C. The supernatant was subjected to sequential immunoprecipitation as follows: overnight incubation with primary antibodies, including anti-OGT (Abcam, ab96718, 4ug, UK), anti-SF3B3 (Proteintech, 14,577–1-AP, 4ug, China), anti-SRSF1 (Proteintech, 12,929–2-AP, 4ug, China), anti-ZC3H14 (Proteintech, 28,190–1-AP, 4ug, China), anti-HCFC1 (Cell Signaling Technology, 69,690, 4ug, USA), anti-THOC1 (Proteintech, 10,920–1-AP, 4ug, China), and anti-THOC5 (Proteintech, 14,862–1-AP, 4ug, China), followed by 3 h of binding with Protein A/G Dynabeads (Thermo Fisher). Bead-bound complexes were washed four times with RIPA buffer and denatured in 50 μL 2× SDS–PAGE loading buffer (95 °C, 12 min). For in-gel digestion coupled with MS analysis of OGT interactors, immunoprecipitated proteins were resolved on 4–12% gradient SDS–PAGE gels and silver-stained, and target bands were excised and analyzed by liquid chromatography–tandem mass spectrometry (LC–MS/MS). Potential OGT-interacting proteins identified from P21 mouse testes are listed in Supplementary Table S1.

Quantitative and semi-quantitative reverse transcription (RT)-PCR assay

Total RNA was extracted with TRIzol Reagent (Thermo Fisher Scientific) and subsequently treated with RNase-free DNase I (Invitrogen). Reverse transcription was performed using 1 μg RNA with PrimeScriptRT Master Mix (Takara, RR036A) to generate cDNA. For quantitative RT-PCR, the 1 µL diluted cDNA was amplified in a standard 20 µL reaction containing LightCycler 480 SYBR Green I Master Mix (Roche, 04887352001). All data were normalized to Gapdh expression levels. Semi-quantitative RT-PCR utilized 2 μL cDNA templates amplified with gene-specific primers, followed by agarose gel electrophoresis (concentration optimized per target). β-actin served as the loading control. Primer sequences are listed in Supplementary Table S2.

RNA-seq and data analysis

Total RNA was isolated from P10 wild-type and Ogt cKO mouse testes using TRIzol reagent (Thermo Fisher Scientific) following the manufacturer’s guidelines. At least three biological replicates per genotype were used for RNA-seq analysis. The cDNA libraries were prepared and sequenced on an Illumina NovaSeq 6000 platform (OE Biotech, Inc., Shanghai, China) generating 150 bp paired-end reads. Raw sequencing data were further subjected to quality control using FASTP [36] for adapter trimming and low-quality read removal. High-quality clean reads were aligned to the murine reference genome (GRCm39) using HISAT2 [37]. DESeq2 [38] identified differentially expressed genes (DEGs) with |log₂(fold change)|> 1 and an adjusted P-value of < 0.05. Gene Ontology (GO) analysis of DEGs was performed using the clusterProfiler package software (v4.2.1). Alternative splicing events were quantified using rMATS (v4.1.2) [39] with thresholds of |ΔPercent Spliced In (PSI)|> 0.05 and False Discovery Rate (FDR) < 0.001.

scRNA-seq library preparation and data analysis

scRNA-seq libraries

Single-cell RNA sequencing (scRNA-seq) was conducted on testicular tissues from P12 wild-type and Ogt cKO mice, with three biological replicates per genotype. Single-cell suspensions were prepared using the MobiCube High-throughput Single Cell 3 Transcriptome Kit V2.1 (MobiTech Biotechnology, PN-S050200301) and processed using the MobiNova-100 microfluidic system. The cell suspensions were calibrated to 700–1200 cells/μL and immediately loaded onto microfluidic chips for droplet generation. Subsequent reverse transcription, cDNA amplification, and library preparation were performed according to the manufacturer’s instructions. Sequencing was performed on an Illumina NovaSeq 6000 (150 bp paired-end), and subsequent bioinformatic analyses were conducted by OE Biotech Co., Ltd. (Shanghai, China) using their established pipeline.

Read filtering and alignment

Single-cell sequencing data processing commenced with the alignment of FASTQ files to the GRCm39 reference genome using MobiVision v3.2, incorporating unique molecular identifier (UMI) deduplication for transcript quantification. The resulting UMI count matrix was subjected to quality control and computational analysis using Seurat v4.0.0 [40], implementing a multiparameter filtration strategy. A set of criteria was used: cells were filtered by (1) < 200 detected genes/cell, (2) < 1000 UMIs/cell, (3) gene-to-transcript complexity ratio (log₁₀[Genes/UMI]) < 0.7, (4) > 10% mitochondrial content, and (5) > 5% hemoglobin gene expression. Potential multiplets were computationally identified using DoubletFinder v2.0.3 [41]. Normalization was performed using the LogNormalize method, scaling cellular transcripts to 10,000 UMIs/cell, followed by natural logarithm transformation.

Dimensionality reduction and clustering analysis

The top 2000 highly variable genes (HVGs) were identified using Seurat FindVariable Features (mean function = ExpMean, dispersion function = LogVMR) and integrated across batches using Harmony (v1.0). Cell clusters were resolved using graph-based clustering (FindClusters) in the integrated gene expression space, followed by dimensionality reduction using UMAP (RunUMAP) for 2D visualization.

Marker gene identification and cell type annotation

Putative marker genes for each cluster were identified using FindAllMarkers (test.use = “presto”) in Seurat, detecting cluster-specific upregulated genes. Expression patterns were validated using VlnPlot and FeaturePlot. Automated cell-type annotation was performed using SingleR (v1.4.1) [42], where the Spearman correlation of each cell with the reference transcriptome was assigned the highest-correlating predefined type, minimizing subjective bias.

Differential gene expression and functional enrichment analysis​

Differentially expressed genes (DEGs) were identified using FindMarkers (test.use = presto) with a significance threshold of |log₂FC|> 0.26 (≥ 1.2-fold change) or |log₂FC|> 0.58 (≥ 1.5-fold change), and P value < 0.05. Subsequently, GO and KEGG pathway enrichment analyses were performed using the clusterProfiler package in R software (v4.0.3) on the basis of the hypergeometric distribution.

Slingshot pseudotime analysis

Slingshot pseudotime analysis (R package v1.8.0) was used to infer lineage trajectories and cellular pseudotemporal ordering [43]. The dimensionally reduced Seurat object was converted to the SingleCellExperiment format, with spermatogonia subtype-1 designated as the trajectory root. Gene expression–pseudotime relationships were modeled via NB-GAM using tradeSeq (v1.4.0). Genes exhibiting pseudotime-dependent expression (identified using an association test) were visualized on a heat map.

Statistical analysis

Data are presented as mean ± standard error of the mean (SEM). Statistical significance was determined by a two-tailed Student’s t-test using GraphPad Prism 6.0, with significance levels defined as *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Results

OGT and O-GlcNAcylation are highly enriched in mouse testes during meiotic onset

Phylogenetic and amino acid sequence conservation analyses demonstrated that OGT is highly conserved across mammalian species (Fig. S1A, B), with human and mouse OGT proteins sharing 99% homology (Fig. S1C), suggesting that OGT is functionally conserved. To understand the biological functions of OGT and O-GlcNAc, we first determined their expression and found that OGT was ubiquitously expressed across mouse tissues, with the highest abundance observed in the lung, ovary, spleen, and testes (Fig. 1A). Western blot analysis showed that OGT, OGA, and O-GlcNAc levels peaked on postnatal days 7 (P7) and 10 (P10), followed by a significant decrease after P14 (Fig. 1B). Notably, high OGT expression at P10 coincided with meiosis initiation in mice, indicating a potentially critical role for OGT during the mitosis-to-meiosis transition.

Fig. 1.

The O-GlcNAc transferase (OGT) is highly conserved across mammals and is enriched during meiotic initiation. A Expression of OGT protein in adult mouse different tissues. GAPDH serves as a loading control. Three independent experiments were performed with biologically distinct mice (n = 3). B Expression of OGT, OGA, and O-GlcNAc protein from mouse testes collected at different time points during postnatal development. ACTIN serves as a loading control. Western blotting was performed in triplicates. C, D Immunostaining of OGT (C) and O-GlcNAc (D) in testis sections from 8-week-old control mice. Scale bar, 25 μm. E, F Distribution of OGT at different stages of spermatogenesis from 8-week-old mice. Scale bars: 25 μm (E) and 10 μm (F). G Immunofluorescence analysis of O-GlcNAc in different male germ cells from 8-week-old mice. PNA is an acrosome marker. Spg, spermatogonia; Pre-Lep, preleptotene spermatocytes; Lep, leptotene spermatocytes; Zyg, zygotene spermatocytes; Pa, pachytene spermatocytes; Dip, diplotene spermatocytes; RS, round spermatids; ES, elongating spermatids. Scale bars, 25 μm. All images are representative of data from n = 3 mice per genotype

Immunofluorescence analysis of adult testes revealed cell type- and stage-specific OGT localization, with robust nuclear expression in Sertoli cells and cytoplasmic–nuclear distribution in spermatogonia, preleptotene, and leptotene spermatocytes (Fig. 1C, E, F). Notably, OGT signals diminished during meiotic progression and were detectable at low levels in zygotene spermatocytes but were entirely absent in pachytene spermatocytes and post-meiotic spermatids (Fig. 1E, F). Consistently, O-GlcNAc was highly enriched in spermatogonia and early spermatocytes, including preleptotene, leptotene, and zygotene spermatocytes (Fig. 1D, G). In contrast, no significant O-GlcNAc signal was detected in Sertoli cells, whereas OGT expression was markedly elevated (Fig. 1C–G). In addition, co-immunostaining analysis revealed OGT (Fig. S2A, B) and O-GlcNAc (Fig. S2C) enrichment in both c-KIT⁺ differentiated spermatogonia and PLZF⁺ undifferentiated spermatogonia in the prepubertal mouse testes. These findings demonstrate that OGT-mediated O-GlcNAcylation may act as a critical modulator of spermatogonial differentiation, meiotic initiation, and progression.

OGT-mediated O-GlcNAcylation targets core transcriptional regulators and splicing factors in mouse testes

To identify OGT-binding proteins and their potential O-GlcNAcylated substrates in the testes, and further clarify the impact of OGT deficiency on spermatogenesis, we characterized the OGT protein complex using co-immunoprecipitation coupled with mass spectrometry (co-IP/MS) in mouse P21 testes (Fig. 2A, B). Hundreds of candidate proteins were identified in the OGT pull-down complexes (Table S1). As expected, OGT and the known OGT-interacting protein HCFC1 emerged as the top two hits on the basis of unique peptide counts (Fig. 2C). Notably, comparison with the recently updated O-GlcNAcAtlas 4.0 [44] revealed that 46 out of 366 identified proteins carry O-GlcNAcylation modification in other mouse tissues/cells, which provides valuable clues for further exploring the potential O-GlcNAc modification of these proteins in germ cells (Table S1). GO enrichment analysis of all identified OGT-interacting proteins revealed significant enrichment in transcriptional regulation and RNA metabolic processes. Key enriched functional categories included mRNA splicing, nuclear export, and stabilization pathways, highlighting the potential roles of OGT in coordinating transcriptional control and mRNA processing (Fig. 2D). To biochemically validate these interactions, we performed immunoprecipitation followed by western blotting, confirming that OGT binds to the splicing regulator ZC3H14, splicing factors SRSF1 and SF3B3, nuclear export factors THOC1 and THOC5, and the well-known transcription co-factor HCFC1 (Fig. 2E). Furthermore, reciprocal co-immunoprecipitation assays demonstrated that HCFC1, SRSF1, and SF3B3 not only bind to OGT but also undergo O-GlcNAc modification, thus identifying them as direct OGT substrates (Fig. 2F–H). In contrast, parallel experiments suggest that OGT is not a major interactor for ZC3H14, THOC1, and THOC5, and these proteins are also not O-GlcNAcylated substrates in mouse testes (Fig. 2I–K). Taken together, our results confirm that HCFC1, SRSF1, and SF3B3 are O-GlcNAcylated by OGT in mouse testes, implying a critical role of OGT in orchestrating testicular gene expression programs.

Fig. 2.

OGT directly O-GlcNAcylates transcriptional and splicing regulators in juvenile mouse testes. A, B OGT immunoprecipitation (IP) with anti-OGT or control IgG antibodies​​ identified protein interactomes in P21 mouse testes (A). Immunoprecipitated complexes were resolved by SDS–PAGE, silver-stained (B), and analyzed by mass spectrometry (MS). C A list of OGT-interacting protein candidates identified by IP-MS analysis from P21 mouse testes. D Functional enrichment analysis (GO) of OGT-interacting proteins​​ identified via IP-MS. E Immunoprecipitation with western blotting validated HCFC1, ZC3H14, SRSF1, SF3B3, THOC1, and THOC5 as OGT interacting partners in P14 murine testes. F–H Co-immunoprecipitation confirmed OGT-mediated O-GlcNAcylation of HCFC1 (F), SRSF1 (G), and SF3B3 (H) in P14 mouse testes. I–K Co-immunoprecipitation analyses of ZC3H14 (I), THOC1 (J), and THOC5 (K) with OGT and O-GlcNAcylation in P14 mouse testes

Loss of germline OGT triggers O-GlcNAcylation deficiency, spermatogenic arrest, and complete male sterility

To further delineate the physiological function of OGT-catalyzed O-GlcNAcylation during mammalian spermatogenesis, we established a germ cell-specific Ogt knockout model through Cre/loxP-mediated conditional gene targeting. We engineered a floxed Ogt allele (Ogt^fl) with loxP sites inserted into the intronic regions flanking exons 2–3 (Fig. S3A). Subsequently, Ogt^fl/fl mice were crossed with Stra8-Cre transgenic lines, where Cre recombinase activates in type A1 spermatogonial cells, starting at postnatal day 3 [45] (Fig. S3A, B). In addition, the successful flox insertion into Ogt and efficient Ogt mRNA deletion in the germline of cKO mice were validated by genotyping PCR and scRNA-seq, respectively (Fig. S3C, D). Despite normal viability and somatic development, Ogt cKO adults developed pronounced testicular atrophy (Fig. 3A). The testicular mass of 8-week-old Ogt cKO males was reduced to approximately 20% of that observed in age-matched control littermates (Ogt^fl/Y; without Stra8-Cre) (Fig. 3B), and Ogt cKO males were completely infertile when mated with wild-type (WT) females (Fig. 3C). Strikingly, the testis-to-body weight ratio in Ogt cKO males exhibited a progressive decline starting from postnatal day 10 (P10) onward (Fig. S3E, F), indicating that the underlying developmental defect originated prior to the meiotic prophase, potentially during meiotic initiation. Western blotting revealed residual OGT expression in P10 Ogt cKO testes, which was attributed to its persistent expression in the testicular somatic cells (Fig. 3D). Intriguingly, total O-GlcNAc levels remained undistinguishable between P10 control and Ogt cKO testes (Fig. 3D). Immunofluorescence analysis further revealed a complete loss of both OGT (Fig. 3E; Fig. S2A, B) and O-GlcNAc (Fig. 3F) signals specifically in the germ cells of Ogt cKO testes. In contrast, Sertoli cells exhibited normal expression patterns, confirming the germ cell-specific depletion of OGT and O-GlcNAcylation in the mutant testes (Fig. 3E, F).

Fig. 3.

Germ cell-specific OGT ablation disrupts O-GlcNAcylation in murine spermatogenesis. A Images from adult (8-week-old) control and Ogt cKO testes. B Testis/body weight ratio of control and Ogt cKO mice at 8 weeks old. Data are presented as mean ± SEM (n = 3). ***P < 0.001. Student’s t-test. C Litter size of adult control and Ogt cKO males. n = 4 for each genotype. Mean number of litters ± SEM. ****P < 0.0001. D Immunoblotting analysis of OGT, OGA, and O-GlcNAc protein in P10 control and Ogt cKO testes. Band intensity was quantified and normalized to ACTIN. Data are representative of three biological replicates. ***P < 0.001 (Student’s t-test). E Immunofluorescence staining of OGT (green) and the acrosome marker PNA (red) in testis sections of P56 control and Ogt cKO mice. SPG, spermatogonia; Pre-Lep, preleptotene spermatocytes; Lep-like, leptotene-like spermatocytes; Pa, pachytene spermatocytes; RS, round spermatids. Scale bar, 25 μm. F Immunostaining of O-GlcNAc (red) and PLZF (green) in P12 control and Ogt cKO testes. Scale bar, 20 μm

Consistent with the observed testicular hypotrophy, histological examination of P56 Ogt cKO testes revealed complete depletion of post-meiotic germ cells and a dramatic reduction in spermatocytes during meiotic prophase I (Fig. 4A). Seminiferous tubule staging revealed two distinct pathological patterns in Ogt cKO mice. Most tubules displayed spermatogonia/Sertoli cell-only histology or were arrested at preleptotene-like stages, whereas a subset exhibited partial progression to zygotene-like spermatocytes (Fig. 4B). In line with the histological findings, immunostaining for SYCP3, a core structural protein of the synaptonemal complex, demonstrated significantly reduced SYCP3-positive spermatocyte populations in Ogt cKO testes compared with those in controls (Fig. 4C).

Fig. 4.

Postnatal loss of Ogt causes spermatogenesis arrest and male infertility. A H&E staining of testes from 8-week-old control and Ogt cKO mice. Pac, pachytene spermatocytes; RS, round spermatids; ES, elongating spermatids; Spg, spermatogonia; Zyg, zygotene spermatocytes. Scale bar, 50 μm. B Representative images of adult Ogt cKO testis. Scale bar, 50 μm. C Immunofluorescence of SYCP3 on P18 testis from control and Ogt cKO mice. Scale bar, 20 μm. D Histological analysis of testes from control and Ogt cKO mice at P8, P10, P12, P18, and P35. Scale bar, 50 μm. E Immunofluorescence staining of DDX4 (a germ cell marker) on testicular sections from P8 and P10 control and Ogt cKO mice. Scale bars, 20 μm. The right histogram indicates the number of DDX4-positive cells per tubule. n, the number of tubules counted. Data are presented as mean ± SEM, with n = 3 mice for both the P8 and P10 groups. ns, nonsignificant. ***P < 0.001. Student’s t-test

To investigate stage-specific defects during the first wave of spermatogenesis in Ogt cKO testes, we performed hematoxylin and eosin (H&E) staining of testes from P8, P10, P12, P18, and P35 mice. In the control testes, germ cells progressed through sequential meiotic stages: leptotene spermatocytes emerged by P10, advanced to zygotene by P12, and reached pachytene by P18, with mature elongated spermatids fully formed at P35 (Fig. 4D). In contrast, Ogt cKO testes exhibited progressive germ cell depletion, beginning with a subtle depletion at postnatal day 10 (P10), becoming pronounced by P12, and culminating in the complete absence of pachytene spermatocytes and post-meiotic spermatids at P18 and P35. The remaining germ cells in Ogt cKO testes exhibited aberrant arrest at zygotene-like stages. Notably, P8 Ogt cKO testes showed no detectable germ cell loss compared with controls (Fig. 4D). Immunofluorescence analysis using the germ cell-specific marker DDX4 demonstrated a comparable number of DDX4-positive germ cells in P8 Ogt cKO and control testes, but revealed an obvious reduction in germ cells in Ogt cKO testes by P10 relative to age-matched controls (Fig. 4E). Collectively, these results demonstrate that OGT-catalyzed O-GlcNAcylation is essential for male germ cell development and maintenance of fertility.

OGT deficiency impairs spermatogonial differentiation and meiotic initiation

Given the pronounced DDX4-positive germ cell depletion in the P10 Ogt cKO testes, we investigated whether OGT deficiency disrupts spermatogonial differentiation and meiotic initiation. Co-immunostaining of the undifferentiated spermatogonia marker, PLZF, and the germ cell marker DDX4 revealed comparable PLZF-positive cell numbers between control and Ogt cKO mice at both P8 (Fig. 5A, C) and P10 (Fig. 5B, C). Remarkably, this unchanged, undifferentiated spermatogonial pool persisted in adult Ogt cKO testes despite the progressive loss of DDX4-positive germ cells (Fig. S4A, B). However, further immunostaining for the differentiated spermatogonia marker c-KIT indicated a significant loss of c-KIT-positive cells in P10 Ogt cKO versus control testes, indicating impaired differentiation (Fig. 5D, E). Critically, although PLZF⁺STRA8⁺ cells persisted in both genotypes, PLZF⁻STRA8⁺ preleptotene spermatocytes were virtually undetectable in Ogt cKO testes (Fig. 5F, G), demonstrating the essential role of OGT in initiating meiosis. Collectively, OGT deficiency blocks spermatogonial differentiation and inhibits meiotic entry during spermatogenesis.

Fig. 5.

OGT is required for spermatogonia differentiation and meiotic initiation and progression. A Immunofluorescence staining of testicular sections from P8 control and Ogt cKO mice with anti-PLZF and anti-DDX4 antibodies. Scale bar, 50 μm. B Immunofluorescence of PLZF (green) and DDX4 (red) on P10 testicular sections from control and Ogt cKO mice. Scale bar, 50 μm. C Quantification of PLZF-positive germ cells per tubule in P8 (A) and P10 (B) control and Ogt cKO mice. n, the counted number of PLZF-positive tubules. Three different mice per genotype were examined. ns, non-significant. Student’s t-test. D Immunofluorescence analysis of P10 testes from control and Ogt cKO mice with antibodies to c-KIT (green) and DDX4 (red). Scale bar, 50 μm. E The ratio of c-KIT-positive cells to DDX4-positive cells per tubule (D) was quantified. Briefly, the numbers of c-KIT-positive and DDX4-positive cells within the same tubule were counted, and their ratio was calculated. Data are derived from n tubules containing c-KIT-positive cells, obtained from three different mice. Statistical significance was determined by an unpaired t-test. F Immunofluorescence staining of PLZF (red) and STRA8 (green) on P10 testicular sections from control and Ogt cKO mice. Scale bar, 20 μm. G The histogram shows the quantification of the number of PLZF⁻STRA8⁺ cells per tubule in (F). n, the number of PLZF⁻STRA8⁺ tubules counted. The data are from three independent biological replicates. Data are presented as mean ± SEM, ***P < 0.001 by Student’s t-test. H Nuclear spreads from P20 control and Ogt cKO testes were immunostained with antibodies against synaptonemal complex protein SYCP3 (green) and DSB marker γH2AX (red). Scale bars, 20 μm. I Quantitative distribution of spermatocytes across meiotic prophase I substages (leptotene, zygotene, pachytene, and diplotene). J TUNEL assay detecting apoptosis in testis sections from control and Ogt cKO mice at P10, P12, and P18. Scale bar, 50 μm. K Quantification of TUNEL-positive cells per tubule. n, the number of tubules counted. Three mice per genotype were counted. Data are presented as mean ± SEM, *P < 0.05, **P < 0.01

OGT safeguards meiotic progression and spermatocyte viability

To characterize the meiotic defects in homologous chromosome synapsis and DNA double-strand breaks (DSBs) repair within Ogt cKO spermatocytes, we firstly performed co-immunostaining of SYCP3 (axial elements) and γH2AX (DSB marker) in P20 testes, a stage when control mice complete meiosis I (Fig. 5H). Control testes contained predominantly pachytene and diplotene spermatocytes. In contrast, Ogt cKO testes showed accumulation of preleptotene, leptotene, and zygotene spermatocytes, with rare pachytene/diplotene cells (Fig. 5I). Notably, despite pronounced meiotic arrest caused by OGT deficiency, both the formation and repair of DSBs remained unaffected (Fig. 5H). Next, we analyzed chromosomal synapses by immunostaining the synaptonemal complex components SYCP3 (axial elements) and SYCP1 (central elements) in P14 control and Ogt cKO testes. SYCP1 signals marked synapsed regions, with synapsis initiating at the zygotene stage and completing fully at the pachytene stage in both genotypes, except for the XY bodies (Fig. S5A, B). This finding confirmed that intact chromosomal synapses persisted following OGT loss. In addition, the meiotic-arrest phenotype indicates susceptibility to apoptosis. Subsequent TUNEL quantification confirmed a significant increase in apoptotic spermatocytes in Ogt cKO testes at P10, P12, and P18 compared with that in controls (Fig. 5J, K), demonstrating the essential role of OGT in meiotic progression and spermatocyte survival.

OGT ablation disrupts global gene expression and induces aberrant splicing in testes

Given that OGT-interacting partners are functionally enriched in mRNA transcription and alternative splicing pathways (Fig. 2D), we first performed RNA-sequencing (RNA-seq) on testes from control and Ogt cKO mice at the meiotic entry stage (P10). Principal component analysis (PCA) demonstrated genotype-specific clustering (Fig. 6A), and hierarchical analysis confirmed differential expression patterns (Fig. 6B). We identified 203 upregulated and 684 downregulated genes, including the essential meiotic entry regulators Stra8 and Meiosin (Fig. 6C). To broaden our investigation, GO analysis revealed that downregulated genes participate in “meiotic progression,” “spermatogenesis,” “DNA metabolic processes,” “DNA replication,” and “the mitotic-to-meiotic transition” (Fig. 6D). Further RT–qPCR assays confirmed significant downregulation of meiotic transcriptional networks in P10 Ogt cKO mice relative to controls, including regulators of chromosome axis (Stag3 and Rec8), double-strand break (DSB) formation (Spo11, Msh4, and Msh5), synapsis components (Sycp1, Sycp2, and Sycp3), recombination machinery (Dmc1, Brca2, Rad51, and Tex11), and cell cycle progression (Rbm46 and Ythdc2) (Fig. 6E). In addition, western blot analysis further revealed significantly reduced STRA8, RBM46, and YTHDC2 expression in P10 Ogt cKO testes compared with that in controls. In contrast, PLZF levels remained comparable after OGT depletion (Fig. 6F). Given the established role of the YTHDC2/RBM46 complex in regulating the mitotic-to-meiotic transition during spermatogenesis, we assessed its localization in Ogt cKO testes via immunofluorescence. In controls, YTHDC2 exhibited predominant cytoplasmic localization in spermatocytes. In contrast, cytoplasmic YTHDC2 expression was significantly reduced in Ogt cKO spermatocyte-like cells (Fig. S6A). Coincidentally, RBM46 expression was markedly diminished in these cells relative to that in controls (Fig. S6B), indicating complex dysregulation upon OGT loss.

Fig. 6.

Inactivation of Ogt disrupts the transcriptome profiles and alternative splicing in testes. A Principal component analysis (PCA) was performed on RNA-seq data obtained from P10 control and Ogt cKO mice. At least three biological replicates per genotype. B Heat map of upregulated and downregulated genes between control and Ogt cKO mice. C The volcano plot exhibits differentially expressed genes (DEGs) in P10 control and Ogt cKO testes. |Fold change|> 2 and an adjusted P-value < 0.05 were used to identify DEGs. D GO term enrichment analysis of downregulated genes in testes of Ogt cKO mice. E qPCR analysis validated selected downregulated genes (from RNA-seq data) in P10 Ogt cKO testes, with Gapdh normalization. Data are presented as mean ± SEM, n = 3. ****P < 0.0001. F Western blot detection of PLZF, STRA8, RBM46, and YTHDC2 in P10 testes from control and Ogt cKO mice. Band intensities were quantified and normalized to ACTIN from three independent experiments. **P < 0.01, ***P < 0.001, ****P < 0.0001. G Quantification of differential alternative splicing (AS) events in in P10 control versus Ogt cKO testes, with category-specific event counts. H Venn diagrams illustrating DEGs/AS event overlap between Ogt cKO and control testes. I GO term enrichment analysis of overlap genes in H. J RT–PCR analysis of Syce2 splicing in P10 control versus Ogt cKO testes. Middle panel: quantification of abnormal splicing frequency. Right panel: We excised the corresponding gel bands from the left panel, and purified the full-length and short isoforms of Syce2 RT–PCR products for Sanger sequencing. The results revealed that aberrant splicing of Syce2 gives rise to exon 3 skipping, with product sizes of 410 bp (full-length) versus 256 bp (skipped isoform). Data are presented as mean ± SEM. **P < 0.01

Using replicate multivariate analysis of transcript splicing (rMATS), we identified 495 differential splicing events in P10 Ogt cKO testes compared with the controls. These comprised 342 skipped exons (SE), 71 alternative 5′ splice sites (A5SS), 49 alternative 3′ splice sites (A3SS), 28 retained introns (RI), and 5 mutually exclusive exons (MXE) (Fig. 6G). Notably, 27 genes overlapped between these differential splicing events and differentially expressed genes (DEGs), including the core meiotic regulators Cdkl2, Stag3, Stra8, and Syce2 (Fig. 6H). GO analysis revealed enrichment of these overlapping genes in “homologous chromosome pairing,” “synaptonemal complex assembly,” and “meiotic cell cycle regulation” (Fig. 6I). We subsequently selected Syce2 and validated its mRNA splicing in P10 control and Ogt cKO testes by RT–PCR. Quantitative analysis and Sanger sequencing revealed increased exon 3 skipping in Ogt cKO testes, which causes a frameshift and further leads to abnormal translation products. These results demonstrate that OGT mediates the regulation of alternative splicing of meiosis-related genes such as Syce2 (Fig. 6J).

Single-cell transcriptome uncovers dysregulated mitotic-meiotic switching in Ogt cKO Spermatocytes

Because bulk RNA-seq measures the average expression across heterogeneous cell populations, the DEGs observed between the control and Ogt cKO testes may result partially from compositional differences. To resolve OGT-dependent transcription at a cellular resolution, we performed scRNA-seq across the spermatogenic lineage in P12 testicular cells, retaining 29,309 control and 15,669 Ogt cKO cells that passed quality control for downstream analysis (Fig. S7A). Through dimension reduction, clustering, and analysis of known testicular cell markers, we identified nine distinct cell populations: spermatogonia (SPG), spermatocytes (SPC), and seven somatic lineages: Leydig cells, peritubular myoid cells, macrophages, endothelial cells, innate lymphoid cells, stromal cells, and Sertoli cells (Fig. 7A; Fig. S7B). Quantitative analysis of the cellular composition revealed that Ogt cKO testes exhibited significant spermatocyte depletion and concomitant Sertoli cell expansion compared with controls, with no notable changes in other major lineages (Fig. S7C), indicating that OGT deficiency impaired meiotic initiation and progression.

Fig. 7.

Single-cell transcriptomic profiling of spermatocytes in control and Ogt cKO mice. A UMAP projection with clustering (control: 29,309 cells, Ogt cKO: 15,669 cells). Each dot represents a single testicular cell. SPG, spermatogonia; SPC, spermatocytes; PTM, peritubular myeloid cells. B Volcano plot of differentially expressed genes (DEGs) in spermatocyte (SPC) populations from control and Ogt cKO mice. Significance thresholds: |log₂FC|> 0.585, and P < 0.05. C, D Gene Ontology (GO) enrichment of the upregulated genes (C) and downregulated genes (D) in Ogt cKO spermatocyte (SPC) populations. E, F FPKM expression profiles in stage-specific spermatogenic cells of the upregulated genes (E) and the downregulated genes (F) in Ogt cKO spermatocytes (SPC). ​​Comparative reanalysis​​ of upregulated/downregulated genes in Ogt cKO spermatocytes against published datasets from three spermatogenic lineages: spermatogonial stem cells (SSCs), spermatocytes, and round spermatids. SPG, spermatogonial stem cells; SPC, meiotic spermatocytes; RS, round spermatids. Statistical significance was determined using the Wilcoxon rank‑sum test. ns, nonsignificant. *P < 0.05. ***P < 0.001. ****P < 0.0001. G, H Heat map analysis of gene expression patterns for upregulated (G) and downregulated (H) genes in Ogt cKO spermatocytes across three spermatogenic cell types. I Immunolabeling of P12 testis sections from control and Ogt cKO mice​​ with phospho-histone H3 Ser¹⁰(pHH3) and synaptonemal complex protein SYCP3. White lines denote spermatogonia, whereas white arrows mark aberrant spermatocyte-like cells exhibiting pHH3 signals. J Quantification of pHH3-positive spermatocytes at P12. n, the number of tubules counted. Three mice per genotype were examined. ****P < 0.0001 by Student’s t-test. K Immunofluorescence staining of SYCP3 and CCNA2 in P12 control and Ogt cKO testis sections. White lines indicate spermatogonia, and white arrows indicate SYCP3-positive spermatocyte-like cells that express CCNA2. Scale bar, 20 μm. L Quantification shows the number of CCNA2‑positive spermatocytes in P12 testes. Data are from n tubules across three mice per genotype. ****P < 0.0001 by Student’s t-test

Next, we analyzed the spermatocyte (SPC) populations in Ogt cKO testes and identified 255 upregulated genes (e.g., Ccna2, Ccnb1, Ccnd3, Ccne2, and Cdk12) and 321 downregulated genes (e.g., Sycp2, Hspa2, Hfm1, and Msh4) (Fig. 7B). GO analysis demonstrated that the upregulated genes were enriched for cell cycle regulatory functions, including “cell division,” “mitotic cell cycle phase transition,” “chromosome condensation,” “DNA replication,” and “protein phosphorylation” (Fig. 7C). Nevertheless, the downregulated genes functionally clustered in some meiotic processes, particularly “male meiosis I,” “male meiotic nuclear division,” and “meiotic cell cycle” pathways (Fig. 7D). To assess whether OGT is essential for mitotic–meiotic transcriptional regulation, we mapped DEGs from Ogt cKO spermatocytes to published spermatogenic cell transcriptomes (spermatogonial stem cells, spermatocytes, and round spermatids) [46, 47]. Notably, upregulated genes in Ogt cKO spermatocytes predominantly exhibited high expression in spermatogonial stem cells, with a progressive decline through spermatocytes to round spermatid stages (Fig. 7E, G). In contrast, the downregulated genes maintained their general expression in both SPG and SPC populations (Fig. 7F, H). We next performed a comparative analysis of DEGs between bulk RNA-seq and spermatocyte populations from scRNA-seq. The two sequencing strategies were conducted using mice at distinct postnatal developmental stages: P10 for bulk RNA-seq, and P12 for scRNA-seq. We observed a limited overlap of DEGs, with only 1 shared upregulated gene and 27 overlapping downregulated genes, such as the key meiotic regulators Sycp2, Hfm1, Brca2, Msh4, and Rad51b (Fig. S8A, B). These results indicate that scRNA-seq provides superior resolution for dissecting cell type-specific gene regulation and uncovering the aberrant overexpression of spermatogonial-enriched transcripts in spermatocytes. These findings suggest that OGT is a key regulator of transcriptional reprogramming, which is essential for accurate mitotic-to-meiotic progression.

The aforementioned experiments demonstrated that male germ cells lacking Ogt exhibit defective meiotic prophase progression and premature cell death. Crucially, Ogt cKO spermatocytes showed an aberrant upregulation of transcripts enriched in chromosomal condensation. Therefore, we tested whether Ogt cKO spermatocytes display pathologically condensed chromosomes. Phosphorylated H3 at serine 10 (pHH3) serves as a metaphase marker localized to condensed chromatin in both mitotic spermatogonia and meiotic metaphase I spermatocytes [35]. In P12 control testes, pHH3 immunostaining was strictly localized to the germ cells within the basal compartment of the seminiferous tubules, specifically marking mitotic metaphase spermatogonia. Consistent with normal developmental progression, SYCP3-positive spermatocytes exhibited no pHH3 signal because germ cells at this stage had not entered meiotic metaphase I. Notably, in P12 Ogt cKO testes, a subset of SYCP3-positive spermatocyte-like cells also displayed pHH3 signals concurrent with chromosome condensation (Fig. 7I, J). To further investigate whether the cell cycle was dysregulated in Ogt cKO testes, we analyzed the expression dynamics of mitotic cyclin A2 (CCNA2), a direct cell cycle progression marker. Control spermatogonia exhibited robust CCNA2 expression that was extinguished upon meiotic commitment and became undetectable in leptotene spermatocytes. However, a proportion of Ogt cKO spermatocytes retained partially aberrant CCNA2 expression (Fig. 7K, L), demonstrating the essential role of OGT in terminating mitotic cycling during meiotic initiation.

OGT governs mRNA metabolism across spermatogonia differentiation stages

To delineate the dynamic gene expression programs during spermatogonial differentiation, we isolated and reclustered spermatogonia (SPG) populations, and identified four distinct subtypes (subtype-1 to subtype-4) (Fig. 8A). On the basis of the established spermatogonial marker expression patterns, subtype-1 cells were classified as spermatogonial progenitors, subtype-2 as early differentiated spermatogonia, subtype-3 as late-differentiated spermatogonia, and subtype-4 as an undefined population exhibiting pan-marker expression (Fig. 8B). Furthermore, slingshot pseudotime analysis delineated the developmental trajectory from subtype-1 to subtype-3 spermatogonia (Fig. 8C). Quantification analysis revealed shifted spermatogonial subtype distributions between control and Ogt cKO testes. In controls, subtype frequencies were 37.06% (subtype-1), 34.14% (subtype-2), 18.78% (subtype-3), and 10.02% (subtype-4). In Ogt cKO testes, this distribution changed significantly to 66.38%, 12.40%, 11.29%, and 9.93% respectively. The 1.8-fold increase in progenitor spermatogonia (subtype-1) coupled with a reduction in differentiated subtypes (subtype-2 and 3) indicated impaired differentiation upon OGT loss (Fig. 8D).

Fig. 8.

scRNA-seq identifies altered spermatogonial subpopulations in Ogt cKO mice. A High-resolution clustering of spermatogonia (SPG)​​ delineates four cellular states (subtype-1 to subtype-4) during spermatogonia differentiation. B Heat map of known marker gene expression across identified cell types. C Slingshot pseudotime trajectory analysis​​ of four spermatogonial (SPG) subtypes. D Proportions of spermatogonia per cell cluster in control versus Ogt cKO testes. E Pseudotime-resolved expression heat map​​ linking spermatogonial differentiation stages to cluster-specific gene modules. Representative gene markers per cluster (right panel) and associated biological processes (GO terms, left panel) mapped to corresponding modules. F, G Venn diagrams of downregulated genes (F) and upregulated genes (G) between SPG subtype-2 and subtype-3 cells. H, I GO enrichment of downregulated genes in SPG subtype-2 (H) and SPG subtype-3 (I)

To uncover the molecular drivers of disrupted spermatogonial differentiation, we applied pseudotime ordering to cluster co-expressed gene modules and visualized their dynamic expression patterns through spermatogonia (SPG) developmental trajectories (Fig. 8E). During early pseudotime development, genes enriched in “ATP synthesis-coupled electron transport,” “protein autophosphorylation,” and “regulation of mRNA poly(A) tail shorting” showed peak expression. Conversely, genes regulating transcription, mitotic G1/S transition, meiotic cell cycle, synapsis, and DNA replication dominated the mid-to-late pseudotime stages (Fig. 8E). Further DEG analysis revealed 1170 and 1185 downregulated genes (P < 0.05, log₂ fold-change > 0.26) in Ogt cKO subtype-2 and subtype-3 spermatogonia, respectively, with 537 genes shared between the subtypes (Fig. 8F). In addition, subtype-specific upregulation (P < 0.05, log₂ fold-change > 0.26) revealed 336 (subtype-2) and 488 (subtype-3) genes sharing 95 common genes (Fig. 8G). Given the substantial downregulation of both subtypes, we performed GO analysis. The results aligned with the pseudotime expression dynamics (Fig. 8E), revealing that the downregulated genes disrupted critical mid-to-late pseudotime processes, including cell cycle regulation, cell division, DNA replication, and transcription (Fig. 8H, I). In comparison, the upregulated genes demonstrated shared enrichment in “ATP synthesis-coupled electron transport” and “protein phosphorylation” pathways across both subtype-2 and subtype-3 spermatogonia (Fig. S8C, D). Furthermore, comparative analysis of DEGs derived from bulk RNA-seq and those from subtype-2 and subtype-3 spermatogonia in scRNA-seq revealed limited overlap among the upregulated genes. Notably, the overlapping downregulated genes were mainly associated with meiotic initiation and progression, including Sycp3, RBM46, Stra8, and Stag3 (Fig. S8E–H). Collectively, dysregulated mRNA metabolism that affects cell cycle control, DNA replication, and transcriptional regulation underlies the defects observed in spermatogonial differentiation.

Deficiency of OGT reduces expression and induces nuclear to cytoplasmic translocation of HCFC1, SF3B3, and ZC3H14 in spermatocytes

To further elucidate how loss of OGT and O-GlcNAcylation affects the expression and subcellular localization of OGT-interacting proteins, we analyzed these targets in testes from P10 and P12 control and Ogt cKO mice. Western blotting revealed comparable levels of SRSF1 and ZC3H14 between the two groups, but showed significant reductions in SF3B3 and the long isoform of HCFC1 in Ogt cKO testes (Fig. 9A). Consistent with the unaltered protein abundance of SRSF1, its subcellular localization remained unchanged in Ogt cKO testes (Fig. 9B, C). Strikingly, HCFC1 (Fig. 9D, E), SF3B3 (Fig. 9F, G), and ZC3H14 (Fig. 9H, I) exhibited nuclear localization in controls, but displayed cytoplasmic mislocalization or expression loss in Ogt-deficient spermatocytes. Collectively, our findings reveal that OGT exerts its regulatory function by interacting with and O-GlcNAcylating key transcription factors (e.g., HCFC1) and splicing regulators (e.g., SRSF1, SF3B3) in spermatocytes, thereby modulating their expression and subcellular localization. This OGT-mediated O-GlcNAcylation-dependent regulatory mechanism is essential for directing the gene expression programs underlying meiotic initiation and progression (Fig. 9J).

Fig. 9.

OGT deficiency alters expression and localization of its interacting substrates in testes. A Immunoblotting of P10 testicular lysates detected OGT-interacting proteins (ZC3H14, SRSF1, SF3B3, and HCFC1) expression in control and Ogt cKO testes. Bands were quantified and normalized to ACTIN. The experiments were performed three times. ***P < 0.001 by Student’s t-test. B Immunofluorescence staining of SRSF1 (green) and SYCP3 (red) in P12 control and Ogt cKO testicular sections. Scale bar, 20 μm. C Quantification of relative SRSF1 immunofluorescence intensity in P12 control and Ogt cKO spermatocytes. n, the number of tubules counted. Data are presented from three biological replicates. ns, nonsignificant. D, F, H​ Representative immunofluorescence images of HCFC1 (D), SF3B3 (F), and ZC3H14 (H) in P12 control and Ogt cKO testes. Spermatocytes are marked by SYCP3 (red). White arrows highlight expression within spermatocytes. Magnified insets highlight the reduced expression and prominent cytoplasmic mislocalization of these factors in cKO spermatocytes. Scale bars, 20 μm. E, G, I​ Corresponding quantifications of the relative fluorescence intensity for each protein in spermatocytes. Data points represent individual tubules (n) from three biological replicates. ****P < 0.0001 (Student’s t-test). J Proposed working model for OGT function in spermatogenesis. OGT and its mediated O-GlcNAcylation orchestrates spermatogenesis by modifying transcription factors (e.g., HCFC1) and splicing regulators (e.g., SRSF1 and SF3B3), thereby programming the expression of key genes (e.g., Stra8, Ythdc2, Rbm46, and Ccna2) essential for the mitotic-to-meiotic transition

Discussion

Spermatogenesis is orchestrated by complex layers of post-translational modifications (PTMs), with transcriptional suppression during meiosis arising from X-chromosome inactivation and histone replacement by protamines during spermiogenesis [48]. While phosphorylation, methylation, acetylation, and ubiquitination regulate critical processes in germ cell development [49], and aberrant SUMOylation and histone crotonylation result in spermiogenic defects and infertility [50], the functional significance of O-GlcNAcylation, a dynamic nutrient-sensitive modification that affects approximately 25% of the proteome [51], remained largely unexplored in mammalian testes. In this study, we focused on O-GlcNAc modification and its sole catalytic enzyme OGT to delineate the physiological functions of OGT in murine spermatogenesis. Our findings demonstrate that OGT is essential for orchestrating transcriptional programming during the mitotic-to-meiotic transition in male germ cells. This regulatory role is exerted through the direct binding to and O-GlcNAcylation of core transcription factors (e.g., HCFC1) and splicing regulators (e.g., SRSF1 and SF3B3).

Protein glycosylation is critically involved in spermatogenesis. While the essential role of endoplasmic reticulum-resident N-linked glycosylation is well established, with deficiencies in its enzymatic machinery (e.g., MAN2A2, RPN1) known to cause male infertility [52, 53], the target substrates and functional relevance of O-linked glycosylation, particularly O-GlcNAcylation, has remained largely unexplored in the testis [54]. In contrast to other O-glycosylation forms, Liu et al. recently provided a systematic profile of O-GalNAcylation in developing mouse testes. Their work mapped 349 O-GalNAc glycoproteins with 799 precise glycosites at P24 and 12 weeks, revealing potential roles for these modified proteins in spermatogenesis and fertilization [25]. Interestingly, previous studies have shown that O-GlcNAc transferase (OGT) and O-GlcNAcylation are present in the testis, epididymis, and immature caput sperm. During epididymal maturation, however, O-GlcNAcylation levels in sperm decrease significantly, suggesting a dynamic regulation of this PTM during functional maturation [55].

OGT complexes direct O-GlcNAcylation to metabolically regulate protein function, with their dysregulation driving metabolic disease [51]. Its functional significance has been demonstrated in diverse cellular processes, including the maturation of the transcriptional cofactor HCFC1 [56], modulation of stress responses [57], and the regulation of stem cell fate [58]. However, emerging evidence on the global impact of O-GlcNAcylation (e.g., in oocyte maturation) is primarily derived from pharmacological OGT/OGA inhibition studies [51], leaving the functional consequences of its complete loss in vivo uncharacterized. To address this, we employed a Stra8-Cre-mediated conditional knockout of Ogt specifically in male germ cells. This model revealed that OGT deficiency triggers a dual defect in spermatogonial differentiation and meiotic entry, culminating in meiotic arrest, spermatocyte apoptosis, and male infertility. While PLZF-positive spermatogonial stem cells were maintained, c-KIT-positive differentiated spermatogonia were significantly depleted, and PLZF⁻STRA8⁺ preleptotene spermatocytes were virtually absent, underscoring the essential role of OGT in germ cell development.

A central finding is the phenotypic convergence between germ cell-specific Ogt ablation and mutations in RNA-binding proteins such as YTHDC2 and RBM46, both resulting in a blockade at the meiotic entry checkpoint [9, 12]. Correlating with transcriptional downregulation, protein levels of the YTHDC2–RBM46 complex were reduced in Ogt cKO testes, which was further corroborated by immunostaining showing diminished expression of YTHDC2 and RBM46 in spermatocytes, indicating that impaired complex function contributes to the spermatogenic defects. This suggests that the OGT regulation pathway converges with​ RNA regulatory networks to form a coordinated regulatory module​ essential for driving the developmental transition. Mechanistically, we delineate that OGT orchestrates this transition probably through two complementary axes operating on distinct substrate classes. First, it directly regulates transcriptional programming by modifying the transcriptional coactivator HCFC1. HCFC1, which is essential for cell cycle progression and is recruited to meiotic gene promoters by factors such as histone demethylase KDM2A [59, 60], emerges as a high-occupancy substrate in our glycoproteome. Its O-GlcNAcylation, required for proteolytic maturation, likely modulates the activity of complexes driving the expression of key meiotic genes such as Stra8, Meiosin, Spo11, and Sycp1[14]. Second, OGT exerts direct control over pre-mRNA splicing fidelity by glycosylating core spliceosome components, including the serine/arginine-rich protein SRSF1 and SF3B3. The functional consequence is evidenced by the aberrant splicing of meiosis-critical genes such as Syce2 upon OGT loss, a phenotype mirroring a Srsf1 conditional knockout model, where disrupted splicing leads to meiotic arrest in mice [61]. Notably, while HCFC1 and SF3B3 exhibited reduced expression and a nuclear-to-cytoplasmic shift in Ogt cKO spermatocytes, SRSF1 levels and localization remained unchanged. We therefore speculate that OGT-mediated O-GlcNAcylation likely modulates SRSF1 activity, such as its ability to bind RNA or protein partners, rather than its abundance or localization. The splicing factor SF3B3 participates in regulating alternative splicing of numerous mRNAs during meiotic and haploid spermatogenesis by forming a complex with Ran‑binding protein 9 (RANBP9) and the splicing factor HNRNPM [62]. This dual-level regulation positions of OGT and O-GlcNAcylation as a key integrator, ensuring that metabolic and developmental cues are precisely coupled to execute the mitotic-to-meiotic transition.

Single-cell clustering analysis revealed a significantly reduced proportion of spermatocytes in Ogt cKO mice, indicating that OGT deficiency impairs meiosis. Despite the marked downregulation of meiosis-related genes, residual spermatocytes in Ogt cKO mice exhibited normal synaptonemal complex assembly and γH2AX accumulation. This suggests that early meiotic events initiate correctly upon OGT loss. However, subsequent expression of key meiotic genes is severely dysregulated, consequently causing meiotic arrest and likely checkpoint-driven apoptosis. Interestingly, a set of core mitotic genes was aberrantly upregulated in these mutant spermatocytes, suggesting a failure to properly silence the mitotic program. This, along with the persistent expression of the mitotic markers pHH3 and CCNA2 in Ogt cKO spermatocytes, a phenotype that mirrors Ythdc2 and Rbm46 mutants [9, 12], confirms that OGT facilitates the mitotic-to-meiotic transition by regulating stage-specific transcriptional programming. Although YTHDC2 was detected in OGT pull-down complexes, co-immunoprecipitation did not support an interaction between OGT and the YTHDC2/RBM46 complex (data not shown), suggesting that the latter is not a glycosylation substrate. Thus, the observed defects in this complex are likely an indirect consequence of OGT loss. Furthermore, certain candidate interactors such as THOC1 and THOC5, which may not bind to OGT, could potentially serve as part of a molecular scaffold that localizes OGT in proximity to its target proteins. Notably, mounting evidence indicates that OGT exerts critical cellular functions beyond its canonical O-GlcNAc transferase activity. Recent studies have uncovered nonenzymatic roles for OGT, including acting as a molecular scaffold via its N-terminal TPR domains to mediate protein–protein interactions [63], regulating stress granule dynamics independent of catalysis, and contributing to cellular homeostasis through structural or adaptor functions that do not require glycosyltransferase activity [64, 65]. Therefore, we cannot exclude the possibility that the meiotic defects in Ogt cKO mice arise, at least in part, from the loss of these unidentified noncatalytic functions of OGT, which may operate in parallel with or independently of O-GlcNAcylation to govern spermatogenesis. These findings collectively suggest that OGT orchestrates meiotic entry through a multi-faceted regulatory network.

The broader implications of our work extend to understanding how nutrient availability and cellular metabolism influence germ cell fate. O-GlcNAcylation acts as a nutrient-sensitive sensor, directly linking the cellular levels of UDP-GlcNAc, the end product of the hexosamine biosynthetic pathway (HBP) [15], to the precise regulation of spermatogenesis. This positions the HBP–OGT axis as a functional bridge​ between metabolic flux and germline development. Consequently, systemic or local metabolic disturbances, such as those occurring in diabetes and obesity, could compromise male fertility [66, 67] by altering the O-GlcNAc modification landscape of key meiotic regulators, thereby disrupting the mitosis-to-meiosis transition. Our findings provide a plausible mechanistic explanation for the well-documented but poorly understood association between metabolic disorders and subfertility in men. Furthermore, the evolutionary conservation of OGT and its modified human orthologs suggests that this pathway is likely relevant in human biology. Therefore, the OGT-interacting proteome we identified in mice could guide the discovery of diagnostic O‑GlcNAc markers in human seminal plasma or testicular biopsies [68]. Moreover, our Ogt cKO model, which recapitulates meiotic arrest, provides a direct preclinical platform to test whether pharmacological or nutritional modulation of O‑GlcNAc cycling can rescue spermatogenic failure. Future studies should aim to characterize the O-GlcNAc modification profile in testicular samples from both mice and men with idiopathic infertility, and to evaluate whether fine‑tuning OGT activity constitutes a viable therapeutic strategy for male infertility subtypes associated with meiotic impairment.

We acknowledge several limitations that also chart future directions. The cellular heterogeneity of testicular tissue likely diluted germ cell-specific signals in our whole-tissue glycoproteomic analysis, potentially masking germline-specific glycosylation alterations in Ogt cKO samples (data not shown). To address this limitation, future studies should employ advanced germ cell sorting techniques coupled with ultrasensitive glycoproteomic approaches, such as electron transfer dissociation (ETD) or complementary methods including higher-energy C-trap dissociation (HCD)-pd-ETD, to define the bona fide germline O-GlcNAcylome and map distinct O-GlcNAc modification sites. Furthermore, although we identified key OGT substrates, a critical next step will be to validate the functional necessity of O-GlcNAcylation at specific residues via in vivo site-directed mutagenesis. Finally, the phenotypic parallel with the YTHDC2/RBM46 complex, despite the absence of a direct physical interaction, suggests an indirect regulatory relationship, warranting deeper investigation into the potential crosstalk between O-GlcNAcylation and m6A-dependent pathways during meiosis.

Conclusions

Our study comprehensively elucidates the essential function of OGT and its mediated O-GlcNAcylation in murine spermatogenesis. We establish that OGT plays a pivotal role in O-GlcNAcylating key transcriptional and splicing regulators, thereby orchestrating programmed gene expression networks that ensure faithful germ cell transition from mitosis to meiosis. These findings elucidate the regulatory control of OGT in meiotic entry and progression, providing a mechanistic foundation for understanding O-GlcNAc metabolic dysregulation in male infertility.

Abbreviations

OGT

O-GlcNAc transferase

PTM

Post-translational modification

HBP

Hexosamine biosynthesis pathway

OGA

O-GlcNAcase

IP-MS

Immunoprecipitation and mass spectrometry

H&E

Hematoxylin and eosin

DSBs

DNA double-strand breaks

PCA

Principal component analysis

rMATS

Replicate multivariate analysis of transcript splicing

DEGs

Differentially expressed genes

SPG

Spermatogonia

SPC

Spermatocytes

pHH3

Phosphorylated H3 at serine 10

ETD

Electron transfer dissociation

HCD

Higher-energy C-trap dissociation

RANBP9

Ran‑binding protein 9

Author contributions

Z.D., C.W., M.L., and K.H. contributed equally to this work. R.G., Z.D., Y.C., Q.S., and H.X. conceived and designed the study. C.W., M.L., K.H., X.L., Z.C., K.L., and H.C. performed the bench experiments and data analyses. R.G. and Z.D performed the bioinformatic analysis. Z.D. and R.G. wrote the manuscript. Q.S., H.X., R.G., and Y.C. revised the manuscript. All authors read and approved the manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (grant numbers: 82301862 and 82201834), the Natural Science Foundation of Anhui Province (grant number: 2208085QH233), the Basic and Clinical Cooperative Research Promotion Program of Anhui Medical University (grant number: 2021xkjT030), the Key Discipline Construction Fund of the Leading Hospital Project at the First Affiliated Hospital of Anhui Medical University (grant number: 060201001), and the University Natural Foundation of Anhui Educational Committee (2022AH010072).

Data availability

Data is available upon reasonable request to the corresponding author. Single-cell and bulk RNA sequencing data have been deposited in the ArrayExpress database under accession numbers E-MTAB-15514 and E-MTAB-15517, respectively.

Declarations

Ethics approval and consent to participate

All animal experiments were approved by the Ethics Committee of the Association of Laboratory Animal Sciences at Anhui Medical University, China (approval no. LLSC20211392; approval date: 18 October 2021). This Ethics Committee operates in full compliance with the guidelines of the International Council for Laboratory Animal Science (ICLAS).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.