mRNA-Seq of testis and liver tissues reveals a testis-specific gene and alternative splicing associated with hybrid male sterility in dzo

Abstract

Alternative splicing (AS) plays an important role in the co-transcription and post-transcriptional regulation of gene expression during mammalian spermatogenesis. The dzo is the male F1 offspring of an interspecific hybrid between a domestic bull (Bos taurus ♂) and a yak (Bos grunniens ♀) which exhibits male sterility. This study aimed to identify the testis-specific genes and AS associated with hybrid male sterility in dzo. The iDEP90 program and rMATS software were used to identify the differentially expressed genes (DEG) and differential alternative splicing genes (DSG) based on RNA-seq data from the liver (n = 9) and testis (n = 6) tissues of domestic cattle, yak, and dzo. Splicing factors (SF) were obtained from the AmiGO2 and the NCBI databases, and Pearson correlation analysis was performed on the differentially expressed SFs and DSGs. We focused on the testis-specific DEGs and DSGs between dzo and cattle and yak. Among the top 3,000 genes with the most significant variations between these 15 samples, a large number of genes showed testis-specific expression involved with spermatogenesis. Cluster analysis showed that the expression levels of these testis-specific genes were dysregulated during mitosis with a burst downregulation during the pachynema spermatocyte stage. The occurrence of AS events in the testis was about 2.5 fold greater than in the liver, with exon skipping being the major AS event (81.89% to 82.73%). A total of 74 DSGs were specifically expressed in the testis and were significantly enriched during meiosis I, synapsis, and in the piRNA biosynthesis pathways. Notably, STAG3 and DDX4 were of the exon skipping type, and DMC1 was a mutually exclusive exon. A total of 36 SFs were significantly different in dzo testis, compared with cattle and yak. DDX4, SUGP1, and EFTUD2 were potential SFs leading to abnormal AS of testis-specific genes in dzo. These results show that AS of testis-specific genes can affect synapsis and the piRNA biosynthetic processes in dzo, which may be important factors associated with hybrid male sterility in dzo.

The dzo is the interspecific hybrid male offspring of a domestic cattle (Bos taurus) and a yak (Bos grunniens), and displays arrested spermatogenesis at meiosis that causes male sterility. This study discovered differential alternative splicing of testis-specific genes and revealed the potential splicing factors associated with hybrid male sterility in dzo, highlighting the important role of alternative splicing in their meiotic arrest.

Introduction

Crossbreeding of a yak (Bos grunniens) with a domestic cattle (Bos taurus) is widely practiced on the Qinghai-Tibet Plateau as a means of increasing milk or meat production via heterosis. Male F1 to F3 offspring of interspecies hybrid are infertile due to spermatogenetic arrest. This male infertility makes it difficult to maintain the dominant favorable traits in subsequent generations, preventing the long-term utilization of the benefits of heterosis. Dzo is the Tibetan name for the male F1 offspring of this interspecific hybridization. Understanding the mechanisms of male infertility in dzo is an important topic in biological and animal genetic science (Zhang et al., 2023).

Spermatogenesis involves the mitotic proliferation of spermatogonia, meiosis, and spermiogenesis, and the successive and simultaneous execution of three different gene expression processes (da Cruz et al., 2016; Griswold, 2016). Each stage requires the interaction of multiple factors whose expression is precisely controlled in time and space (White-Cooper and Davidson, 2011). Dysfunction of gene expression or gene regulation at any stage will impair male fertility. The seminiferous tubules of dzo contain a few primary spermatocytes but lack spermatid cells, and spermatogenesis occurs at meiotic arrest (Lou et al., 2014; Zhang et al., 2019a). Several studies have attempted to unravel the potential mechanisms underlying hybrid male sterility in dzo using omics technology. Transcriptomic and proteomic analyses have revealed significant down-regulation of the genes and proteins associated with cell cycles and meiosis in dzo testis, and spermatogenesis genes were barely expressed at all (Cai et al., 2017; Wu et al., 2020, 2023a; Zhang et al., 2023). Using single cell RNA sequencing analysis, the obstruction of dzo spermatogenesis has been shown to perhaps originate in the differentiation stage of undifferentiated spermatogonia, and to be aggravated during mitosis and meiosis (Mipam et al., 2023). A number of epigenetic studies have been conducted to determine the reasons for the robust burst of misregulated gene expression in dzo testis. Whole-genome bisulfite-seq and small RNA-seq have revealed that the promoter hypermethylation-associated silencing of the PIWI/piRNA pathway genes disrupt the production of piRNA in dzo (Zhang et al., 2020; Phakdeedindan et al., 2022). Abnormal expression of miRNA in dzo spermatogenic cells may block the transition of male germ cells from mitosis to meiosis (Xu et al., 2020). Assays for transposase-accessible chromatin using high-throughput sequencing have revealed different chromatin accessibility between dzo and cattle or yak testis (Cao et al., 2022; Zhang et al., 2023). The transcription factor MYBL1 may be a candidate master regulator for dysregulation of the pachytene spermatocyte genes in dzo (Zhang et al., 2023). Meanwhile, the levels and distributions of the histone methylation (H3K4me3, H3K9me1, H3K9me3, and H4K20me3) were strikingly different in the meiotic chromosomes of dzo spermatocytes (Li et al., 2020). Taken together, these results suggest that abnormal gene expression and regulation may adversely affect the formation and differentiation of dzo spermatogenic cells. Therefore, discovering the testis-specific genes and their regulatory factors causing this male sterility may be a key step in understanding the molecular mechanism of sterility in dzo.

Alternative splicing (AS) increases transcriptome and proteome complexity and plays an essential role in tissue development, including spermatogenesis (Wu et al., 2022). AS often acts in an organ-, tissue-, or cell-type-specific manner, and can globally reprogram the splicing landscape in the mitotic-to-meiotic transition during male meiosis (Kalsotra and Cooper, 2011; Schmid et al., 2013). The abundance of AS events reached 95% to 100% in human genes and 63% in mouse genes (Barbosa-Morais et al., 2012; Merkin et al., 2012). Exon skipping events accumulate in the brain and testis, suggesting a tissue-specific nature of AS (Barbosa-Morais et al., 2012; Schmid et al., 2013). Although the roles of most spliced forms of specific genes during spermatogenesis remain unclear, several genes important for spermatogenesis have specific splice variants in different developmental stages. For example, meiotic recombination protein SPO11 is required for the formation of DNA double-strand breaks (DSB). Spo11 generate the two isoforms, Spo11α and Spo11β, by exon skipping. Spo11α is expressed in spermatocytes during the pre-pachytene stage and may be necessary for effective XY chromosome pairing in later stages. Spo11β is highly expressed in the early leptotene and zygotene stages to produce sufficient DSBs (Kauppi et al., 2011; Cesari et al., 2020). The enhancer of zeste homolog 2 (EZH2) is highly expressed in mitotic germ cell populations and spermatogonial stem cells and can cause three forms of H3K27 methylation. In contrast, exon 14 skipped EZH2 (ex14D-EZH2) expression increases only after the onset of meiosis and does not maintain H3K27me3 (Mu et al., 2018).

In addition, numerous transacting regulators of pre-mRNA splicing are important for spermatogenesis. Sequence-specific RNA-binding proteins (RBP) bind to pre-mRNA to control AS (Fu and Ares, 2014). Splicing factors (SF) are a class of RBPs that regulate variable splicing. Serine/arginine-rich splicing factors (SRSFs) are key components in executing pre-RNA processing events. SRSF1 and SRSF2 affect the expression and AS of STRA8, a key factor in spermatogenesis (Lei et al., 2023a, 2023b). SRSF2 in Sertoli cells is essential for testicular development and spermatogenesis in mice (Xie et al., 2023). Polypyrimidine tract-binding protein 2 (PTBP2) is a strong regulator of AS and many mis-spliced isoforms of genes critical for germinal-Sertoli cells appear in PTBP2-deficient mouse testis (Hannigan et al., 2017). hnRNPH1 recruits PTBP2 and SRSF3 to modulate alternative splicing in the testis (Feng et al., 2022). BCAS2 affected AS of DAZL (lacking exon 8) and is involved in spermatogonia and the transition to meiosis (Liu et al., 2017).

In bovine, an overwhelming proportion of tissue-specific transcripts (61%) were attributed to the testis, and more than 80% of the transcription start sites used by testis-specific isoforms were only active in the testis (Halstead et al., 2021). Lactate dehydrogenase C (LDHC), bovine vasa homology (Bvh), and b-Boule are essential proteins for spermatogenesis, and their splicing variants are downregulated in dzo testis compared with yak or cattle (Huang et al., 2012; Luo et al., 2013; Li et al., 2014). To our knowledge, there are currently no studies of genome-wide analysis of AS in dzo. Previous omics studies have only compared dzo with domestic cattle or yak, and only with regard to the testis tissues. Adult testis exhibits high cell heterogeneity and contains different types of germ and somatic cells. Our previous study revealed that most upregulated genes in dzo testis were related to metabolic processes (Zhang et al., 2023). The liver is an essential metabolic organ, and its metabolism is controlled by insulin and other metabolic hormones. Therefore, liver cells were used as a proxy for the somatic cells to identify the differentially expressed genes (DEG) and differential alternative splicing genes (DSG) specifically expressed in dzo testis. The SFs that potentially regulate testicular-specific differential AS events were analyzed. The results revealed that the testis-specific genes and AS were associated with hybrid male sterility in dzo.

Materials and Methods

Animals

Liver tissues were collected from cattle (n = 3; 2.5 years), yaks (n = 2; 4.5 years), and dzo (n = 4; 4.5 years) for mRNA-seq. Testicular tissue samples had already been obtained from the center of the testes of sexually mature cattle (n = 2; 2.5 years), yaks (n = 2; 4.5 years), and dzo (n = 2; 4.5 years) for mRNA-seq and histological analysis in our previous study (Zhang et al., 2019a, 2023). All experimental procedures were approved by the Institutional Animal Care and Use Committee of Southwest University (Chongqing, China).

RNA-Seq library preparation, sequencing

The nine liver tissue samples from cattle, yak, and dzo were sent to Novogene Biotechnology Ltd. (Beijing, China) for RNA extraction, library construction, and sequencing. Total RNA was extracted using Trizol reagent (Invitrogen, Carlsbad, CA, USA). Total RNA was reverse transcribed into cDNA using a TruSeq RNA Sample Preparation Kit (Illumina, San Diego, CA, USA). The double-stranded cDNA was purified using an MPure XP system (Beckman Coulter, Beverly, USA). The purified double-stranded cDNA was subjected to end repair, poly (A) addition, and splicing of the sequencing adaptor. A paired-end RNA-seq library was compiled using an Illumina HiSeq 4000 (2 × 150 bp read length). In addition, we integrated the six testis mRNA-Seq datasets from our previous study (Zhang et al., 2023).

Gene expression pattern analysis

Clean data (clean reads) were obtained from the raw data (raw reads) in a fastq format by removing reads containing adapters, and reads containing ploy-N and low quality reads. At the same time, the Q20, Q30, and GC contents of the clean data were calculated. The sequencing quality was assessed using FastQC v.0.11.8 (Babraham Institute, Cambridge, UK). Low-quality reads and remaining adapters were trimmed using Trimmomatic v.0.36. The clean reads were then aligned to the Bos_taurus reference genome ARS-UCD1.2 and the yak reference genome BosGru_v.2.0 using Hisat2 v.2.0.5 (Kim et al., 2015) and only the uniquely mapped reads were retained. Read counts were calculated using HTSeq-count. DEGs were identified using the DESeq2 R package (v.1.16.0) (Love et al., 2014), with the criteria |log₂ fold change (FC)| >1 and P < 0.05.

The mitochondrial genes were removed from the datasets. A total of 21,043 genes or novel transcripts with more than 20 read counts in at least two samples were used for the subsequent analyses. Principal component analysis (PCA) was performed using the NovoMagic Cloud Platform (https://magic.novogene.com). The read count data of these 21,043 genes were sent to iDEP90, an integrated web application for differential expression and pathway analysis of RNA-Seq data (Ge et al., 2018). We chose the top 3,000 most variable genes for K-means cluster analysis and Gene Ontology (GO) biological process enrichment evaluation. The fold change of dzo, compared with cattle or yak, was converted using a log₂-transformation, thereby enabling directions and fold changes to be viewed directly. The graphs of gene expression were produced using GraphPad Prism 8.0 (Inc., La Jolla, CA, USA).

Identification of AS events and functional enrichment of DSGs

The rMATS software can identify five types of AS events, including alternative 3ʹ splice sites (A3SS), alternative 5ʹ splice sites (A5SS), mutually exclusive exons (MXE), retained introns (RI), and exon skipping (ES), as well as calculating the corresponding percent spliced in (PSI) values between the groups (Figure 2A). The comparison groups of cattle, yak, and dzo for the liver (CY_L vs C_L, CY_L vs Y_L, Y_L vs C_L) and testis (CY_T vs C_T, CY_T vs Y_T and Y_T vs C_T) samples were analyzed, using the Bos_taurus reference genome ARS-UCD1.2. Statistical graphs of AS events were generated using GraphPad Prism 8.0. The AS events with P < 0.05 were defined as DSGs and were screened for DSGs between the same tissues in cattle, yak, and dzo. Venn diagrams were plotted using the website tool (http://bioinformatics.psb.ugent.be/webtools/Venn/). GO and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses of the DSGs were performed by reference to the GO and KEGG databases, respectively. Significant GO and KEGG terms were identified using Fisher’s exact test.

Figure 2.

Analysis of alternative splicing events and differential alternative splicing genes. (A) Five types of alternative splicing (AS) events. (B) Five types of AS differentially splicing genes (DSG) in the liver tissue samples. (C) Five types of AS DSGs in testis tissue samples. (D) Venn plots of DSGs showing gene clusters C, D, E, and F in dzo testis tissues. The numbers represent the number of genes in each category. (E) Enrichment analysis of the GO biological process clusters of C and D overlapping differential genes (ODG). (F) Enrichment analysis of the GO biological processes of ODGs in clusters E and F. (G) The expression of ODGs (STAG3, DMC1, KASH5, GPAT2, DDX4, and TDRD9) involved in synapsis and piRNA biosynthetic process in testicular tissue of cattle, yak and dzo. (H) The KEGG pathways of DSGs occurring in the testis of dzo, compared with cattle and yak. (I) The mRNA expression of DSGs in the liver and testis tissues of dzo, compared with cattle and yak. The x-axis represents dzo and cattle; the y-axis represents dzo and yak.

Analyses of SFs and the regulatory network of SFs and DSGs

SFs were obtained using AmiGO2 (https://amigo.geneontology.org/) and the NCBI database for the functional annotation of the cattle genome. The differential expression of SFs in tissues between cattle and dzo was analyzed using DESeq2 (v.1.16.0). Volcano plots displaying gene markers were generated using ggplot2 in the R package (v.3.4.4), applying the differential expression screening criteria |log₂FC|>1 and P < 0.05. We conducted a Pearson’s correlation analysis to determine the relationships between the AS event PSI values of the overlapping differential genes (ODG) of clusters E-F and the expression of differential SFs. Significant correlations were assigned at a critical value of |r| > 0.95 and P < 0.01. Based on the significance results produced by the Pearson’s correlation analyses, Cytoscape (v.3.10.0) was used to generate a potential SF-AS regulatory network.

Validation of AS events using RT-PCR and RT-qPCR

The NCBI (https://www.ncbi.nlm.nih.gov) database was used to localize the region of the gene where AS occurred. The full-length transcript and splice transcript sequences of the gene were compared using the UCSC database (https://genome.ucsc.edu) to determine the type of AS event. AS events in DDX4 and MYBL1, which are associated with meiosis and piRNA biosynthesis, were validated using RT-PCR with β-actin as the internal reference gene. Specific primers were designed for the regions where AS occurred in the target genes using the NCBI primer design software (Supplementary Table S1). Primers were synthesized by GENEWIZ Biotechnology (Suzhou, China). The RT-PCR reactions were performed on an S1000 system (Bio-Rad, Hercules, CA, USA) using Premix Taq (Taq Version 2.0 plus dye, TaKaRa, Dalian, China). The reaction system consisted of Premix Taq 5 µL, 0.4 µL each of forward and reverse primers, cDNA 1 µL, and ddH₂O 3.2 µL. Reactions were conducted under the following conditions: pre-denaturation at 95 °C for 5 min, 35 cycles of denaturation at 95 °C for 30 s, annealing at 60 °C for 30 s, and finally extension at 72 °C for 10 min.

Reverse transcription-quantitative polymerase chain reaction (RT-qPCR) was used to verify the expression levels of full-length transcripts and splicing variants of DSGs. The RT-qPCR reactions were performed on a CFX96 system (Bio-Rad) using TB Green Premix Ex Taq II (TaKaRa, Dalian, China). The reaction system was prepared according to the manufacturer’s instructions. Reactions were conducted under the following conditions: pre-denaturation at 95 °C for 3 min, 40 cycles of denaturation at 95 °C for 10 s, and annealing at 60 °C for 30 s. Melting analysis was then performed at temperatures ranging from 65 °C to 90 °C, at increments of 0.2 °C. β-Actin and GAPDH were used as reference genes. The mRNA relative fold change was analyzed using the CFX-96 system’s relative quantification software, based on the 2^−△△Ct method.

Statistical analysis

RT-qPCR data were expressed as mean ± SD. For all statistical analyses, P < 0.05 was considered significant. One-way ANOVA was conducted for all statistical significance analyses using the GraphPad Prism 8.0 software.

Results

We generated mRNA-seq data from the testis and liver samples. The testis samples were taken from two cattle, two yak, and two dzo; the liver samples were taken from three cattle, two yak, and four dzo and had an average of 43 million raw reads of each tissue per individual (Supplementary Table S2). An average of 38.4 million (88.8%) clean reads of each tissue per individual were uniquely mapped to the Bos_taurus reference genome ARS-UCD1.2 (Supplementary Table S2). An average of 37.3 million (86.3%) clean reads of each tissue per individual were uniquely mapped to the yak reference genome BosGru_v2.0 (Supplementary Table S2). The mitochondrial genes of dzo and yak both showed the same maternal expression pattern (Supplementary Figure S1), showing that all of the individual dzo in this study were the result of crosses between a female yak and a domestic bull, which was consistent with our previous study based on the phylogenetic analysis of mitochondrial DNA (mtDNA) sequences (Zhang et al., 2019a). All of the mitochondrial genes were removed, but non-coding genes or specific gene families were not excluded. The mapping data from the Bos_taurus ARS-UCD1.2 reference genome were used for subsequent analyses.

Identification of testis-specific genes involved in spermatogenesis

A total of 21,043 genes or novel transcripts with more than 20 read counts in at least two samples were used for the subsequent analyses. PCA analysis showed that the 15 samples clearly clustered on the PC1 axis and accounted for 77.23% of the variance (Figure 1A). This indicated significant differences between the testis and liver tissues of the cattle, yaks, and dzo. The PC2 axis accounted for 8.64% of the variance, which was due to differences between fertile and infertile testis (Figure 1A). Exploratory data analysis showed that thousands of genes were differently expressed between the testis and liver tissues. Based on the distribution of gene expression variances, we chose the top 3,000 genes for cluster analysis (Figure 1B). Based on the within-group sum of squares plot (Figure 1C) and spermatogenesis stages, we chose K = 6 for gene clustering and analysis of the enriched GO terms (Figure 1D and E). Of these 3,000 genes, those showing higher expression in the liver were involved in the metabolic process categories. These genes were concentrated in clusters A (n = 693) and B (n = 254) (Figure 1D and E). The scatter plot showed no significant difference in the expression of clusters A and B genes in dzo liver tissues compared with cattle or yak (Figure 1F, −1 ≤ log₂FC ≤ 1). However, the expression of these genes was disordered in dzo testis tissues (Figure 1F, −3 ≤ log₂FC ≤ 3). The genes more highly expressed in the testis corresponded to four different stages of spermatogenesis: cluster F (n = 574) for the mitotic, cluster E (n = 291) for the meiotic, cluster D (n = 877) for the spermatid, and cluster C (n = 311) for the spermiogenesis stages (Figure 1D and E). The genes in cluster F were both up- and downregulated in dzo, compared with cattle and yak (Figure 1F, −2 ≤ log₂FC ≤ 3). The genes that were significantly downregulated in dzo showed fold changes that changed in a cascading manner from clusters E to C (Figure 1F, log₂FC ≤ −2). The genes in cluster F were most enriched in the mitosis stage of spermatogenesis; e.g., in GO categories, sister chromatid separation, mitotic cell cycle, sister chromatid separation, and mitotic nuclear division (Figure 1E). The genes in cluster E were most enriched in the GO categories of male gamete generation, meiotic cell cycle, and meiotic nuclear division (Figure 1E). These results were consistent with our previous study showing that mitotic and meiotic stage might be the key point of meiosis arrest in dzo (Zhang et al., 2023).

Figure 1.

Disorder of testis-specific gene expression in dzo. (A) Principal component analysis of gene expression differences between the liver and testis tissues of domestic cattle, yak, and dzo. (B) Standard deviations of 21,043 genes. (C) The elbow rule determines the optimal K value of the top 3,000 genes using K-means with the largest difference. (D) Heat map of the aggregation of differentially expressed genes between the tissues of cattle, yak, and dzo. (E) Functional enrichment of A–F cluster genes. (F) The scatterplot compares the differences in gene expression levels between dzo and its parental species (cattle, x-axis; yak, and y-axis). The green box contains the scatter plot of testicular tissue gene expression levels.

Identification of AS events in testis and liver tissues

Five types of AS events were identified in all of the tissues tested, including A5SS, A3SS, MXE, RI, and ES (Figure 2A). When AS occurs in a gene, it results in sequence changes that may affect transcriptome and proteome complexity. Regarding the liver tissue comparisons: dzo comparing with cattle (CY_L vs C_L) revealed 11,378 AS events in 5429 transcripts; (CY_L vs Y_L) revealed 11,126 AS events in 5452 transcripts; (Y_L vs C_L) revealed 10,837 genes in 5285 transcripts. A total of 19,493 genes or novel transcripts were expressed in the liver. AS events therefore occurred in 27.11% to 27.97% of transcripts in each liver tissue sample. Regarding the testis tissue comparisons: 29,711 AS events were detected in 9531 transcripts between dzo and cattle (CY_T vs C_T); 28697 AS events were detected in 9365 transcripts (CY_T vs Y_T); and 26545 AS events were detected in 8811 transcripts (Y_T vs C_T), respectively. The probability of AS events in the testis tissues was 42.08% to 45.52% of the total 20,937 expressed transcripts. The number of AS events in the testis tissues were about 2.5-fold higher than in the liver tissue; the major AS events were ES (81.89~82.73%), and A5SS, only with 1.10% to 1.63% (Supplementary Table S3).

Identification of the different AS genes in dzo testis tissues

Using rMATS, we found 92, 75, and 254 genes with significantly different AS events in liver tissue between dzo and cattle (CY_L vs C_L), dzo and yak (CY_L vs Y_L), and yak and cattle (Y_L vs C_L), respectively (P < 0.05, Figure 2B). These genes were defined as DSGs. These DSGs accounted for 0.81%, 0.67%, and 2.34% of the AS events in these three comparison groups, respectively. In the testis tissues, 1001 (3.37%), 1022 (3.56%), and 839 (3.16%) DSGs were identified between CY_T vs C_T, CY_T vs Y_T, and Y_T vs C_T, respectively (Figure 2C). ES was the major splicing type in these DSGs. The AS type with the lowest occurrence in liver tissue was A5SS, while it was A3SS in testis tissue. Both the number of DSGs and the proportion of AS events were higher in the testis than in the liver.

Enrichment analysis of the different spliced genes in dzo testis tissue

To better understand the importance of AS in gene regulation during dzo spermatogenesis, we further explored the significantly differential AS events in testis-specific genes. A total of 2053 testis-specific genes in clusters C–F were subjected to Venn analysis comparing the DSGs of CY_T vs C_T and CY_T vs Y_T. The results showed that 79 genes in clusters C–D, and 74 genes in clusters E–F overlapped with DSGs (Figure 2D, blue and red circles). For clarity, we designated the genes with differences in mRNA expression levels and AS levels as ODG. The GO and KEGG functional enrichment analyses showed that the 79 ODGs in clusters C–D were enriched in biological processes such as intracellular signal transduction, flagellated sperm motility, spermatogenesis, and regulation of mitotic cytokinesis (Figure 2E). The 74 ODGs in clusters E–F were associated with male meiosis I, synapsis, piRNA biosynthetic process, and reciprocal meiotic recombination (Figure 2F). The expressions of the six ODGs STAG3, DMC1, KASH5, GPAT2, DDX4, and TDRD9 involved in synapsis and piRNA biosynthesis in the three bovine testis are shown in Figure 2G. It can be seen that, except for GPAT2, the expression of the other five genes was significantly downregulated in dzo. These results suggest that genes involved in synapsis and piRNA biosynthesis occurring in AS events may be associated with spermatogenesis arrest in dzo.

There were 360 genes with significant AS events that occurred in dzo testis, compared with both cattle and yak testis (Figure 2D, purple circle). These 360 DSGs were enriched in the hedgehog signaling pathway, cell adhesion, glycerophospholipid metabolism, and biosynthesis of amino acids (Figure 2H). Compared with cattle and yak, these 360 DSGs were not significantly differently expressed in dzo liver samples, but their expression was disordered in dzo testis, with a fold change of between −2 to 2 (Figure 2I).

Identification of the SFs associated with AS events in mitotic and meiotic genes in dzo testis tissues

In order to explore the upstream regulators governing differential AS events, we screened 557 SFs from the AmiGOs and NCBI databases and finally annotated 168 SFs in Ensemble for further analysis. Thirty-six SFs were differentially expressed between dzo and cattle testis, of which 20 were downregulated and 16 were upregulated (Figure 3A, |log₂ FC|>1 and P < 0.05). Sixteen SFs belonged to DSGs in the CY_T vs C_T and CY_T vs Y_T comparisons. Of these, seven SFs genes simultaneously underwent differential AS and were differentially expressed in the testis (Figure 3B). DDX4 was not expressed in the liver, but six other genes (CELF2, CELF6, PTBP1, RBM4B, PRPF39, SUGP1) were expressed in the liver, but were not significantly differently expressed in comparisons between dzo liver and cattle or yak liver (data not shown). Six SFs were of the ES splicing type, and only CELF2 was of type A5SS (Figure 3B). We were unable to obtain sufficiently purified spermatogenic cell populations to investigate the stage-specific expression of these SFs in bovine. Data from mice (da Cruz et al., 2016) showed that the CELF2, PTBP1, PRPF39, and SUGP1 genes were turned on in the majority of somatic cells (2C) and in leptotene or zygotene spermatocytes (LZ), and that the expression peaks of DDX4 was observed in pachytene spermatocytes (PS) (Figure 3C). These overlapping differential SFs may be able to regulate both the splicing of other genes and the splicing of themselves.

Figure 3.

Splicing factor difference analysis and correlation analysis. (A) Volcano plot of the differences in splicing factors (SF). Red indicates upregulation and blue indicates down-regulation in dzo testis. Green circles show overlapping differential genes (ODG) for SFs. (B) Basic characteristics of SFs: DSG_SF refers to AS level differences in SFs; DEG_SF refers to gene level differences in SFs (left); the expression of 7 SFs with different mRNA levels and AS levels (right). (C) Expression patterns of genes in purified mouse spermatogenic cell populations: 2C, most contained somatic cells; LZ, leptotene and zygotene spermatocytes; PS, pachytene spermatocytes; RS, round spermatids. Data from da Cruz et al. (2016). (D) ODGs correlation network diagram for SFs and clusters E–F in dzo testis tissue. The colors represent log₂fold change: red indicates upregulation of gene expression; blue indicates downregulation. The inner circle shows ODGs and the outer circle shows differential SFs. The green area in the outer circle shows the ODGs of SFs. (E) Sequence alignment of DDX4 and MYBL1 full-length transcripts and splicing variants (left); agarose gel electrophoresis of genes and the β-actin internal reference (right). (F) mRNA expression levels of DDX4 and MYBL1 full-length transcripts and splicing variants based on reverse transcription-quantitative PCR. The full-length transcripts and splice variants of DDX4 and MYBL1 are represented by DDX4-Full, DDX4-ES, MYBL1-Full, and MYBL1-ES, respectively.

A Pearson’s correlation analysis was performed between the expression levels of 36 differentially expressed SFs and the PSI of 74 ODGs in clusters E and F. Correlations were deemed significant when |r| > 0.95 and P < 0.01. The correlation network diagram of the 36 SFs and 48 ODGs is shown in Figure 3D. Of the 48 ODGs, four were upregulated and 20 were downregulated. Among them, SFs EFTUD2 and SUGP1 were significantly correlated with 15 and 14 of these 48 ODGs, respectively. These genes were downregulated in dzo testis, e.g., DMC1 and TDRD9, and were involved in synapsis and the piRNA pathway. The expression peak of DMC1 appeared in LZ, and the expression peak of TDRD9 appeared in PS (Figure 3C).

Most interestingly, the Pearson’s correlation analysis showed that testis-specific splicing factor DDX4 was associated with testis-specific transcription factor MYBL1 AS events. The AS events involving the DDX4 and MYBL1 genes were validated using RT-PCR. UCSC results showed the occurrence of the AS of these two genes (Figure 3E, left). The AS events of DDX4 and MYBL1 were of the ES type, with exons 3 and 15 skipped, separately. We further analyzed the PCR products of AS events using agarose gel electrophoresis (Figure 3E, right) and observed the expected amplification products of the different transcripts of these two genes. The expression of full-length transcripts was higher than that of the splicing variants, and the expression of the splicing variants in the testis of dzo was lower than those in cattle or yak. RT-PCR showed that exon 3 skipped DDX4 transcripts were almost never expressed in dzo testis but were expressed at high levels in cattle and yak (Figure 3E, right). However, two types of MYBL1 transcript were expressed at low levels in dzo testis, although the full-length transcript of MYBL1 was significantly downregulated in dzo, compared with cattle and yak (Figure 3E). RT-qPCR showed that the full-length transcripts, and the splice variants of DDX4 and MYBL1 were significantly downregulated in the dzo testis (Figure 3F). Our previous study showed that MYBL1 is the master regulator of dysregulated PS genes and that pachytene piRNA expression is lower in interspecific hybrid dzo (Zhang et al., 2020, 2023). These results indicate that SFs may affect AS events in synapsis and the piRNA biosynthesis processes in dzo testis.

Discussion

Interspecific hybridization plays a vital role in enriching animal diversity and offspring often displays hybrid male sterility (HMS), usually having obstructed spermatogenesis, which leads to reproductive isolation. Compared with fertile hybrids, male-biased genes and genes involved in spermatogenesis are disordered in sterile hybrids, causing HMS (Civetta, 2016; Mack and Nachman, 2017). Dzo can be used as an ideal livestock model of HMS. Most previous studies on them have only used testicular tissue to attempt to reveal the molecular mechanism of their HMS (Xu et al., 2020; Mipam et al., 2023; Zhang et al., 2023).

In this study, liver tissue was used as proxy for the somatic cells and was compared with testis tissue. This approach excluded genes that may be associated with somatic cells in the testis, and allowed us to focus on testis-specific genes related to male infertility in dzo (Figure 1D). Clusters A and B genes were highly expressed in liver tissue and not significantly different between dzo and cattle or yak. However, they exhibited disordered expression in dzo testis. The metabolism-related genes were testis-specific disordered, suggesting that testis somatic cell function was also problematic in dzo. Previous studies have shown that glial cell-derived neurotrophic factor (GDNF) and the genes involved in retinoic acid biogenesis were altered in the Sertoli cells of dzo, highlighting the role of Sertoli cells and their derived factors in HMS (Wu et al., 2023b).

Compared with the liver, as a proxy for the somatic tissues, the testis-specific DEGs were significantly enriched in spermatogenesis. It was impossible to determine whether these gene expression disorders were the cause or the result of male sterility. Based on the K-means cluster analysis, we found that gene expression was disordered in the biological processes of mitosis and meiosis, and in the burst downregulation in the pachynema spermatocytes. Genes in the clusters C to F were enriched in pathways related to spermatogenesis, consistent with previous research on dzo male sterility (Mipam et al., 2023; Zhang et al., 2023). The disordered expression of testis-specific genes (clusters E-F) was enriched in mitosis and meiosis, which was also consistent with previous studies (Wu et al., 2020; Mipam et al., 2023; Zhang et al., 2023). Genes in clusters C and D were involved in the post-meiosis and spermiogenesis stages, as the result of there being no specific spermatogenesis cells in dzo (Zhang et al., 2019a). Therefore, the reasons for gene expression disorder during mitosis and meiosis were key to exploring the obstruction of spermatogenesis in dzo. In total, the identification of testis-specific genes and their stage-specific expression patterns is helpful in further exploration of the reasons for the failure of spermatogenesis in dzo.

AS events occurred much more often in testis tissue than in liver tissue. This is because spermatogenesis is an extremely complex and coordinated process. Various genes are dynamically expressed in every type of spermatogenic cell and the development of the testis is regulated by higher AS levels than other tissues (Soumillon et al., 2013; Song et al., 2020; Niayale et al., 2021). There were a large number of DSGs in the CY_T vs C_T and CY_T vs Y_T comparisons (Figure 2C). Using Venn analysis, it was found that 153 of the testis-specific genes were differentially spliced in dzo and the ODGs were enriched in pathways related to synapsis and the piRNA biosynthetic process. During meiosis, synapsis and homologous recombination are interdependent processes and are prerequisites for the formation of crossovers (CO). The abnormalities or impairments caused by COs are likely to lead to infertility (Ma et al., 2022; Fan et al., 2023). DNA meiotic recombinase 1 (DMC1) (Hinch et al., 2020), stromal antigen 3 (STAG3) (Llano et al., 2014), and KASH domain containing 5 (KASH5) (Horn et al., 2013) are key proteins for DSB repair and homologous chromosome pairing. Studies have shown that the splicing variant of DMC1 may reduce its binding affinity to the regulatory factor BRCA2 (Ahlawat et al., 2016). Sequence variation of STAG3 and KASH5 can lead to a meiotic arrest in male mice (van der Bijl et al., 2019; Yang et al., 2022). DMC1, STAG3, and KASH5 had AS events and were downregulated in dzo testis. Differential abundance proteins in yak and dzo were related to DSB repair during meiosis (Wu et al., 2023a). As a type of small RNA that is highly expressed in testis tissue, piRNA promotes spermatogenesis in a stage-specific manner and is comprised of pre-pachytene piRNA and pachytene piRNA (Li et al., 2013). Pachytene piRNAs are produced from pachytene spermatocytes to round spermatids during spermatogenesis, which protects the genomic integrity of germ cells by silencing active transposable elements (de Mateo and Sassone-Corsi, 2014). Among the DSGs, DDX4, TDRD9, and GPAT2 are involved in piRNA biosynthesis (Kuramochi-Miyagawa et al., 2010; Wenda et al., 2017; Shiromoto et al., 2019). Of them, DDX4 and TDRD9 differ at the mRNA and AS levels, but GPAT2 is only different at the AS level. Our previous studies have shown that downregulation of piRNA-related genes leads to a decrease in the number of piRNAs in the pachytene stage in dzo testis (Zhang et al., 2020). Therefore, differential splicing in dzo testis may affect the successful completion of meiosis and lead to spermatogenesis failure.

SFs play a central role in AS regulation by interacting with sequence elements of pre-RNA (Chen and Manley, 2009). This study explored the correlation between differentially expressed SFs and the PSI of ODGs and found that PTBP1, SUGP1, and EFTUD2 were highly correlated with the AS of testis-specific genes. Deletion of PTBP1 leads to the severe arrest of spermatogonial cell proliferation accompanied by increased apoptotic cell death (Senoo et al., 2019). A previous study showed that a large number of splicing errors occurred when SUGP1 were knockdown (Zhang et al., 2019b). EFTUD2 can inhibit N6-methyladenosine (m⁶A) and affect mRNA accessibility, thereby affecting mRNA stability (He et al., 2023). In the testis of dzo, SUGP1 expression was downregulated and EFTUD2 expression was upregulated (Figure 3A). A previous study showed that differentially methylated genes (e.g., m⁶A) were enriched in pathways related to spermatogenesis in dzo, such as homologous recombination and meiotic spindle organization (Wang et al., 2022). The expression of the m⁶A-associated gene ALKBH5 was specifically downregulated in dzo testis (Chen et al., 2022). ALKBH5-mediated m⁶A modification can influence the stability and splicing of mRNAs with long 3ʹ-UTRs in spermatocytes and round spermatids (Tang et al., 2018). These results suggest that the disorder of gene expression and AS events in dzo are co-regulated by multiple factors.

Interestingly, the testis-specific gene DDX4 was also found to be significantly correlated as a splicing factor with the AS events of the transcription factor MYB proto-oncogene like 1 (MYBL1). The expression peaks of DDX4 and MYBL1 appeared during the early stages of meiotic prophase I (Figure 3C). DDX4 can form a complex with RNA binding motif 5 (RBM5) as an RBP to regulate various RNA processing events (Wu et al., 2022). In previous studies, it was found that MYBL1 might be the master transcriptional regulator for significant mis-regulation of the expression of PS genes in dzo (Zhang et al., 2023). Previous studies have shown that the transcription factor MYBL1 initiates the burst of piRNA production during early meiosis in mouse testes (Li et al., 2013; Yu et al., 2023). In this study, significantly different expressions of AS genes were significantly enriched during meiosis I and synapsis, and in the piRNA biosynthesis pathways. In summary, we speculate that SF expression disorders may regulate the abnormal AS of genes in the testis of dzo.

Conclusions

Testis-specific gene expression was dysregulated in mitosis and meiosis in dzo and showed a burst downregulation from pachynema spermatocytes. The proportion of AS in testis tissue was higher than that in liver tissue, and ES was the main AS type. Genes involved with meiosis, synapsis, and the piRNA biosynthetic pathway showed differential AS and were downregulated in dzo testis tissue. DDX4, SUGP1, and EFTUD2 were identified as the potential SFs leading to abnormal testis-specific AS genes in dzo. Therefore, AS during spermatogenesis may be an important factor affecting meiotic arrest in dzo. Further research is needed to fully explain the role of differentially expressed SFs in dzo testis, to enable a better understanding of the mechanism of hybrid male infertility.

Glossary

Abbreviations

AS

alternative splicing

RBP

RNA binding protein

DSB

DNA double-strand breaks

SF

splicing factor

DEG

differentially expressed gene

mRNA

messenger RNA

miRNA

microRNA

piRNA

piwi-interacting RNA

lncRNA

long non-coding RNA

DSG

differential alternative splicing gene

PCA

principal component analysis

GO

Gene Ontology

A3SS

alternative 3ʹ splice sites

A5SS

alternative 5ʹ splice sites

MXE

mutually exclusive exons

RI

retained introns

PSI

percent spliced in

KEGG

Kyoto Encyclopedia of Genes and Genomes

ES

exon skipping

FC

fold change

mtDNA

mitochondrial DNA

CY_L vs C_L

dzo liver compared to cattle liver

CY_L vs Y_L

dzo liver compared to yak liver

Y_L vs C_L

yak liver compared to cattle liver

CY_T vs C_T

dzo testis compared to cattle testis

CY_T vs Y_T

dzo testis compared to yak testis

Y_T vs C_T

yak testis compared to cattle testis

DSG

differential splicing gene

ODG

overlapping differential gene

2C

the majority of somatic cells

LZ

leptotene or zygotene spermatocytes

PS

pachytene spermatocytes

HMS

hybrid male sterility

CO

crossovers

Conflict of Interest Statement

Authors have no conflicts of interest.