AtSNU13 modulates pre-mRNA splicing of RBOHD and ALD1 to regulate plant immunity.

Pre-mRNA splicing is a significant step for post-transcriptional modifications and functions in a wide range of physiological processes in plants. Human NHP2L binds to U4 snRNA during spliceosome assembly; it is involved in RNA splicing and mediates the development of human tumors. However, no ortholog has yet been identified in plants. Therefore, we report At4g12600 encoding the ortholog NHP2L protein, and AtSNU13 associates with the component of the spliceosome complex; the atsnu13 mutant showed compromised resistance in disease resistance, indicating that AtSNU13 is a positive regulator of plant immunity. Compared to wild-type plants, the atsnu13 mutation resulted in altered splicing patterns for defense-related genes and decreased expression of defense-related genes, such as RBOHD and ALD1. Further investigation shows that AtSNU13 promotes the interaction between U4/U6.U5 tri-snRNP-specific 27 K and the motif in target mRNAs to regulate the RNA splicing. Our study highlights the role of AtSNU13 in regulating plant immunity by affecting the pre-mRNA splicing of defense-related genes.

M6A-modified lncRNA FAM83H-AS1 promotes colorectal cancer progression through PTBP1.

LncRNA plays a crucial role in cancer progression and targeting, but it has been difficult to identify the critical lncRNAs involved in colorectal cancer (CRC) progression. We identified FAM83H-AS1 as a tumor-promoting associated lncRNA using 21 pairs of stage IV CRC tissues and adjacent normal tissues. In vitro and in vivo experiments revealed that knockdown of FAM83H-AS1 in CRC cells inhibited tumor proliferation and metastasis, and vice versa. m6A modification is critical for FAM83H-AS1 RNA stability through the writer METTL3 and the readers IGF2BP2/IGFBP3. PTBP1-an RNA binding protein-is responsible for the FAM83H-AS1 function in CRC. T4 (1770-2440nt) and T5 (2440-2743nt) on exon 4 of FAM83H-AS1 provide a platform for PTBP1 RRM2 interactions. Our results demonstrated that m6A modification dysregulated the FAM83H-AS1 oncogenic role by phosphorylated PTBP1 on its RNA splicing effect. In patient-derived xenograft models, ASO-FAM83H-AS1 significantly suppressed the growth of gastrointestinal (GI) tumors, not only CRC but also GC and ESCC. The combination of ASO-FAM83H-AS1 and oxaliplatin/cisplatin significantly suppressed tumor growth compared with treatment with either agent alone. Notably, there was pathological complete response in all these three GI cancers. Our findings suggest that FAM83H-AS1 targeted therapy would benefit patients primarily receiving platinum-based therapy in GI cancers.

SART3 reads methylarginine-marked glycine- and arginine-rich motifs.

Glycine- and arginine-rich (GAR) motifs, commonly found in RNA-binding and -processing proteins, can be symmetrically (SDMA) or asymmetrically (ADMA) dimethylated at the arginine residue by protein arginine methyltransferases. Arginine-methylated protein motifs are usually read by Tudor domain-containing proteins. Here, using a GFP-Trap, we identify a non-Tudor domain protein, squamous cell carcinoma antigen recognized by T cells 3 (SART3), as a reader for SDMA-marked GAR motifs. Structural analysis and mutagenesis of SART3 show that aromatic residues lining a groove between two adjacent aromatic-rich half-a-tetratricopeptide (HAT) repeat domains are essential for SART3 to recognize and bind to SDMA-marked GAR motif peptides, as well as for the interaction between SART3 and the GAR-motif-containing proteins fibrillarin and coilin. Further, we show that the loss of this reader ability affects RNA splicing. Overall, our findings broaden the range of potential SDMA readers to include HAT domains.

MRS2 missense variation at Asp216 abrogates inhibitory Mg2+ binding, potentiating cell migration and apoptosis resistance.

Mitochondrial magnesium (Mg2+) is a crucial modulator of protein stability, enzymatic activity, ATP synthesis, and cell death. Mitochondrial RNA splicing protein 2 (MRS2) is the main Mg2+ channel in the inner mitochondrial membrane that mediates influx into the matrix. Recent cryo-electron microscopy (cryo-EM) human MRS2 structures exhibit minimal conformational changes at high and low Mg2+, yet the regulation of human MRS2 and orthologues by Mg2+ binding to analogous matrix domains has been well established. Further, a missense variation at D216 has been identified associated with malignant melanoma and MRS2 expression and activity is implicated in gastric cancer. Thus, to gain more mechanistic and functional insight into Mg2+ sensing by the human MRS2 matrix domain and the association with proliferative disease, we assessed the structural, biophysical, and functional effects of a D216Q mutant. We show that the D216Q mutation is sufficient to abrogate Mg2+-binding and associated conformational changes including increased α-helicity, stability, and monomerization. Further, we reveal that the MRS2 matrix domains interact with ~µM affinity, which is weakened by up to two orders of magnitude in the presence of Mg2+ for wild-type but unaffected for D216Q. Finally, we demonstrate the importance of Mg2+ sensing by MRS2 to prevent matrix Mg2+ overload as HeLa cells overexpressing MRS2 show enhanced Mg2+ uptake, cell migration, and resistance to apoptosis while MRS2 D216Q robustly potentiates these cancer phenotypes. Collectively, our findings further define the MRS2 matrix domain as a critical Mg2+ sensor that undergoes conformational and assembly changes upon Mg2+ interactions dependent on D216 to temper matrix Mg2+ overload.

Interleukin-17 directly stimulates tumor infiltrating Tregs to prevent cancer development.

Background: Interleukin-17 (IL-17) family cytokines promote protective inflammation for pathogen resistance, but also facilitate autoimmunity and tumor development. A direct signal of IL-17 to regulatory T cells (Tregs) has not been reported and may help explain these dichotomous responses. Methods: We generated a conditional knockout of Il17ra in Tregs by crossing Foxp3-YFP-Cre mice to Il17ra-flox mice (Il17ra ΔTreg mice). Subsequently, we adoptively transferred bone marrow cells from Il17ra ΔTreg mice to a mouse model of sporadic colorectal cancer (Cdx2-Cre +/Apc F/+), to selectively ablate IL-17 direct signaling on Tregs in colorectal cancer. Single cell RNA sequencing and bulk RNA sequencing were performed on purified Tregs from mouse colorectal tumors, and compared to those of human tumor infiltrating Treg cells. Results: IL-17 Receptor A (IL-17RA) is expressed in Tregs that reside in mouse mesenteric lymph nodes and colon tumors. Ablation of IL-17RA, specifically in Tregs, resulted in increased Th17 cells, and exacerbated tumor development. Mechanistically, tumor-infiltrating Tregs exhibit a unique gene signature that is linked to their activation, maturation, and suppression function, and this signature is in part supported by the direct signaling of IL-17 to Tregs. To study pathways of Treg programming, we found that loss of IL-17RA in tumor Tregs resulted in reduced RNA splicing, and downregulation of several RNA binding proteins that are known to regulate alternative splicing and promote Treg function. Conclusion: IL-17 directly signals to Tregs and promotes their maturation and function. This signaling pathway constitutes a negative feedback loop that controls cancer-promoting inflammation in CRC.

Reverse engineering neuron type-specific and type-orthogonal splicing-regulatory networks using single-cell transcriptomes.

Cell type-specific alternative splicing (AS) enables differential gene isoform expression between diverse neuron types with distinct identities and functions. Current studies linking individual RNA-binding proteins (RBPs) to AS in a few neuron types underscore the need for holistic modeling. Here, we use network reverse engineering to derive a map of the neuron type-specific AS regulatory landscape from 133 mouse neocortical cell types defined by single-cell transcriptomes. This approach reliably inferred the regulons of 350 RBPs and their cell type-specific activities. Our analysis revealed driving factors delineating neuronal identities, among which we validated Elavl2 as a key RBP for MGE-specific splicing in GABAergic interneurons using an in vitro ESC differentiation system. We also identified a module of exons and candidate regulators specific for long- and short-projection neurons across multiple neuronal classes. This study provides a resource for elucidating splicing regulatory programs that drive neuronal molecular diversity, including those that do not align with gene expression-based classifications.

Expression and Purification of Human U1-70K (snRNP70) and its BAD Domains Using an E. coli Expression System.

U1-70K (snRNP70) serves as an indispensable protein component within the U1 complex, assuming a pivotal role in both constitutive and alternative RNA splicing processes. Notably, U1-70K engages in interactions with SR proteins, instigating the assembly of the spliceosome. This protein undergoes regulation through phosphorylation at multiple sites. Of significant interest, U1-70K has been implicated in Alzheimer's disease, in which it tends to form detergent-insoluble aggregates. Even though it was identified more than three decades ago, our understanding of U1-70K remains notably constrained, primarily due to challenges such as low levels of recombinant expression, susceptibility to protein degradation, and insolubility. In endeavoring to address these limitations, we devised a multifaceted approach encompassing codon optimization, strategic purification, and a solubilization protocol. This methodology has enabled us to achieve a high yield of full-length, soluble U1-70K, paving the way for its comprehensive biophysical and biochemical characterization. Furthermore, we provide a detailed protocol for the preparation of phosphorylated U1-70K. This set of protocols promises to be a valuable resource for scientists exploring the intricate web of U1-70K-related mechanisms in the context of RNA splicing and its implications in neurodegenerative disorders and other disorders and biological processes. © 2024 The Authors. Current Protocols published by Wiley Periodicals LLC. Basic Protocol 1: Expression and purification of full-length U1-70K from E. coli Support Protocol 1: Making chemically competent BL21 Star pRARE/pBB535 cells Basic Protocol 2: Phosphorylation of full-length U1-70K using SRPK1 Support Protocol 2: Purification of SRPK1 Basic Protocol 3: Expression and purification of U1-70K BAD1 from E. coli Basic Protocol 4: Phosphorylation of U1-70K BAD1 using SRPK1 Basic Protocol 5: Expression and purification of U1-70K BAD2 from E. coli.

Interferon-Regulated Expression of Cellular Splicing Factors Modulates Multiple Levels of HIV-1 Gene Expression and Replication.

Type I interferons (IFN-Is) are pivotal in innate immunity against human immunodeficiency virus I (HIV-1) by eliciting the expression of IFN-stimulated genes (ISGs), which encompass potent host restriction factors. While ISGs restrict the viral replication within the host cell by targeting various stages of the viral life cycle, the lesser-known IFN-repressed genes (IRepGs), including RNA-binding proteins (RBPs), affect the viral replication by altering the expression of the host dependency factors that are essential for efficient HIV-1 gene expression. Both the host restriction and dependency factors determine the viral replication efficiency; however, the understanding of the IRepGs implicated in HIV-1 infection remains greatly limited at present. This review provides a comprehensive overview of the current understanding regarding the impact of the RNA-binding protein families, specifically the two families of splicing-associated proteins SRSF and hnRNP, on HIV-1 gene expression and viral replication. Since the recent findings show specifically that SRSF1 and hnRNP A0 are regulated by IFN-I in various cell lines and primary cells, including intestinal lamina propria mononuclear cells (LPMCs) and peripheral blood mononuclear cells (PBMCs), we particularly discuss their role in the context of the innate immunity affecting HIV-1 replication.

Rice floury endosperm26 encoding a mitochondrial single-stranded DNA-binding protein is essential for RNA-splicing of mitochondrial genes and endosperm development.

Endosperm, the major storage organ in cereal grains, determines the grain yield and quality. Mitochondria provide the energy for dry matter accumulation, in the endosperm development. Although mitochondrial single-stranded DNA-binding proteins (mtSSBs) play a canonical role in the maintenance of single-stranded mitochondrial DNA, their molecular functions in RNA processing and endosperm development remain obscure. Here, we report a defective rice endosperm mutant, floury endosperm26 (flo26), which develops abnormal starch grains in the endosperm. Map-based cloning and complementation experiments showed that FLO26 allele encodes a mitochondrial single-stranded DNA-binding protein, named as mtSSB1.1. Loss of function of mtSSB1.1 affects the transcriptional level of many mitochondrially-encoded genes and RNA splicing of nad1, a core component of respiratory chain complex I in mitochondria. As a result, dysfunctional mature nad1 led to dramatically decreased complex I activity, thereby reducing ATP production. Our results reveal that mtSSB1.1 plays an important role in the maintenance of mitochondrial function and endosperm development by stabilizing the splicing of mitochondrial RNA in rice.

Aberrant spliceosome activity via elevated intron retention and upregulation and phosphorylation of SF3B1 in chronic lymphocytic leukemia.

Splicing factor 3b subunit 1 (SF3B1) is the largest subunit and core component of the spliceosome. Inhibition of SF3B1 was associated with an increase in broad intron retention (IR) on most transcripts, suggesting that IR can be used as a marker of spliceosome inhibition in chronic lymphocytic leukemia (CLL) cells. Furthermore, we separately analyzed exonic and intronic mapped reads on annotated RNA-sequencing transcripts obtained from B cells (n = 98 CLL patients) and healthy volunteers (n = 9). We measured intron/exon ratio to use that as a surrogate for alternative RNA splicing (ARS) and found that 66% of CLL-B cell transcripts had significant IR elevation compared with normal B cells (NBCs) and that correlated with mRNA downregulation and low expression levels. Transcripts with the highest IR levels belonged to biological pathways associated with gene expression and RNA splicing. A >2-fold increase of active pSF3B1 was observed in CLL-B cells compared with NBCs. Additionally, when the CLL-B cells were treated with macrolides (pladienolide-B), a significant decrease in pSF3B1, but not total SF3B1 protein, was observed. These findings suggest that IR/ARS is increased in CLL, which is associated with SF3B1 phosphorylation and susceptibility to SF3B1 inhibitors. These data provide additional support to the relevance of ARS in carcinogenesis and evidence of pSF3B1 participation in this process.

Comprehensive Analyses of the Effects of the Small-Molecule Inhibitor of the UHM Domain in the Splicing Factor U2AF1 in Leukemia Cells.

Mutations in RNA splicing factor genes including SF3B1, U2AF1, SRSF2, and ZRSR2 have been reported to contribute to development of myeloid neoplasms including myelodysplastic syndrome (MDS) and secondary acute myeloid leukemia (sAML). Chemical tools targeting cells carrying these mutant genes remain limited and underdeveloped. Among the four proteins, mutant U2AF1 (U2AF1mut) acquires an altered 3' splice site selection preference and co-operates with the wild-type U2AF1 (U2AF1wt) to change various gene isoform patterns to support MDS cells survival and proliferation. U2AF1 mutations in MDS cells are always heterozygous and the cell viability is reduced when exposed to additional insult affecting U2AF1wt function. To investigate if the pharmacological inhibition of U2AF1wt function can provoke drug-induced vulnerability of cells harboring U2AF1 mut , we conducted a fragment-based library screening campaign to discover compounds targeting the U2AF homology domain (UHM) in U2AF1 that is required for the formation of the U2AF1/U2AF2 complex to define the 3' splice site. The most promising hit (SF1-8) selectively inhibited growth of leukemia cell lines overexpressingU2AF1 mut and human primary MDS cells carrying U2AF1 mut . RNA-seq analysis of K562-U2AF1mut following treatment with SF1-8 further revealed alteration of isoform patterns for a set of proteins that impair or rescue pathways associated with endocytosis, intracellular vesicle transport, and secretion. Our data suggested that further optimization of SF1-8 is warranted to obtain chemical probes that can be used to evaluate the therapeutic concept of inducing lethality to U2AF1 mut cells by inhibiting the U2AF1wt protein.

Identification and characterization of a novel intronic splicing mutation in CSF1R-related leukoencephalopathy.

AIMS: Colony stimulating factor 1 receptor (CSF1R)-related leukoencephalopathy is a rapidly progressing neurodegenerative disease caused by CSF1R gene mutations. This study aimed to identify and investigate the effect of a novel intronic mutation (c.1754-3C>G) of CSF1R on splicing. METHODS: A novel intronic mutation was identified using whole-exome sequencing. To investigate the impact of this mutation, we employed various bioinformatics tools to analyze the transcription of the CSF1R gene and the three-dimensional structure of its encoded protein. Furthermore, reverse transcription polymerase chain reaction (RT-PCR) was performed to validate the findings. RESULTS: A novel mutation (c.1754-3C>G) in CSF1R was identified, which results in exon 13 skipping due to the disruption of the 3' splice site consensus sequence NYAG/G. This exon skipping event was further validated in the peripheral blood of the mutation carrier through RT-PCR and Sanger sequencing. Protein structure prediction indicated a disruption in the tyrosine kinase domain, with the truncated protein showing significant structural alterations. CONCLUSIONS: Our findings underscore the importance of intronic mis-splicing mutations in the diagnosis and management of CSF1R-related leukoencephalopathy.

SpliceAPP: an interactive web server to predict splicing errors arising from human mutations.

BACKGROUND: Splicing variants are a major class of pathogenic mutations, with their severity equivalent to nonsense mutations. However, redundant and degenerate splicing signals hinder functional assessments of sequence variations within introns, particularly at branch sites. We have established a massively parallel splicing assay to assess the impact on splicing of 11,191 disease-relevant variants. Based on the experimental results, we then applied regression-based methods to identify factors determining splicing decisions and their respective weights. RESULTS: Our statistical modeling is highly sensitive, accurately annotating the splicing defects of near-exon intronic variants, outperforming state-of-the-art predictive tools. We have incorporated the algorithm and branchpoint information into a web-based tool, SpliceAPP, to provide an interactive application. This user-friendly website allows users to upload any genetic variants with genome coordinates (e.g., chr15 74,687,208 A G), and the tool will output predictions for splicing error scores and evaluate the impact on nearby splice sites. Additionally, users can query branch site information within the region of interest. CONCLUSIONS: In summary, SpliceAPP represents a pioneering approach to screening pathogenic intronic variants, contributing to the development of precision medicine. It also facilitates the annotation of splicing motifs. SpliceAPP is freely accessible using the link https://bc.imb.sinica.edu.tw/SpliceAPP . Source code can be downloaded at https://github.com/hsinnan75/SpliceAPP .

Splicing-specific transcriptome-wide association uncovers genetic mechanisms for schizophrenia.

Recent studies have highlighted the essential role of RNA splicing, a key mechanism of alternative RNA processing, in establishing connections between genetic variations and disease. Genetic loci influencing RNA splicing variations show considerable influence on complex traits, possibly surpassing those affecting total gene expression. Dysregulated RNA splicing has emerged as a major potential contributor to neurological and psychiatric disorders, likely due to the exceptionally high prevalence of alternatively spliced genes in the human brain. Nevertheless, establishing direct associations between genetically altered splicing and complex traits has remained an enduring challenge. We introduce Spliced-Transcriptome-Wide Associations (SpliTWAS) to integrate alternative splicing information with genome-wide association studies to pinpoint genes linked to traits through exon splicing events. We applied SpliTWAS to two schizophrenia (SCZ) RNA-sequencing datasets, BrainGVEX and CommonMind, revealing 137 and 88 trait-associated exons (in 84 and 67 genes), respectively. Enriched biological functions in the associated gene sets converged on neuronal function and development, immune cell activation, and cellular transport, which are highly relevant to SCZ. SpliTWAS variants impacted RNA-binding protein binding sites, revealing potential disruption of RNA-protein interactions affecting splicing. We extended the probabilistic fine-mapping method FOCUS to the exon level, identifying 36 genes and 48 exons as putatively causal for SCZ. We highlight VPS45 and APOPT1, where splicing of specific exons was associated with disease risk, eluding detection by conventional gene expression analysis. Collectively, this study supports the substantial role of alternative splicing in shaping the genetic basis of SCZ, providing a valuable approach for future investigations in this area.

A rare ACAN non-canonical splicing-site intron variant results in familial short stature.

OBJECTIVE: ACAN gene variants, prevalent monogenic defects linked to short stature, are characterized by impaired cartilage generation in growth plates. We aimed to unravel the genetic basis of short stature in a specific pedigree by investigating the role of a novel non-canonical splicing-site variant, c.630-13G > A, within the ACAN gene. METHOD: Sanger sequencing was used for pedigree verification, and the effects of this variant on mRNA splicing were analyzed through minigene assay. RESULTS: The study revealed that this variant led to the creation of a previously unreported splice site in the fourth intron, resulting in the incorporation of an 11 bp sequence from the intron into the final transcript. This alteration led to a frameshift and formation of a premature termination codon, impacting the structure of the aggrecan protein. CONCLUSIONS: We document the pathogenicity of an ACAN non-canonical splicing-site variant, emphasizing the significance of considering intronic variants during genetic testing.

A 3'-pre-tRNA-derived small RNA tRF-1-Ser regulated by 25(OH)D promotes proliferation and stemness by inhibiting the function of MBNL1 in breast cancer.

BACKGROUND: We explored the potential novel anticancer mechanisms of 25-hydroxyvitamin D (25(OH)D), a vitamin D metabolite with antitumour effects in breast cancer. It is stable in serum and is used to assess vitamin D levels in clinical practice. Transfer RNA-derived small RNAs are small noncoding RNAs that generate various distinct biological functions, but more research is needed on their role in breast cancer. METHODS: Small RNA microarrays were used to explore the novel regulatory mechanism of 25(OH)D. High-throughput RNA-sequencing technology was used to detect transcriptome changes after 25(OH)D treatment and tRF-1-Ser knockdown. RNA pull-down and high-performance liquid chromatography-mass spectrometry/mass spectrometry were used to explore the proteins bound to tRF-1-Ser. In vitro and in vivo functional experiments were conducted to assess the influence of 25(OH)D and tRF-1-Ser on breast cancer. Semi-quantitative PCR was performed to detect alternative splicing events. Western blot assay and qPCR were used to assess protein and mRNA expression. RESULTS: The expression of tRF-1-Ser is negatively regulated by 25(OH)D. In our breast cancer (BRCA) clinical samples, we found that the expression of tRF-1-Ser was higher in cancer tissues than in paired normal tissues, and was significantly associated with tumour invasion. Moreover, tRF-1-Ser inhibits the function of MBNL1 by hindering its nuclear translocation. Functional experiments and transcriptome data revealed that the downregulation of tRF-1-Ser plays a vital role in the anticancer effect of 25(OH)D. CONCLUSIONS: In brief, our research revealed a novel anticancer mechanism of 25(OH)D, unveiled the vital function of tRF-1-Ser in BRCA progression, and suggested that tRF-1-Ser could emerge as a new therapeutic target for BRCA.

The Functional Relationship Between RNA Splicing and the Chromatin Landscape.

Chromatin is a highly regulated and dynamic structure that has been shown to play an essential role in transcriptional and co-transcriptional regulation. In the context of RNA splicing, early evidence suggested a loose connection between the chromatin landscape and splicing. More recently, it has been shown that splicing occurs in a co-transcriptional manner, meaning that the splicing process occurs in the context of chromatin. Experimental and computational evidence have also shown that chromatin dynamics can influence the splicing process and vice versa. However, much of this evidence provides mainly correlative relationships between chromatin and splicing with just a few concrete examples providing defined molecular mechanisms by which these two processes are functionally related. Nevertheless, it is clear that chromatin and RNA splicing are tightly interconnected to one another. In this review, we highlight the current state of knowledge of the relationship between chromatin and splicing.

Genetic variation of circHIBADH enhances prostate cancer risk through regulating HNRNPA1-related RNA splicing.

The current study aimed to investigate associations of circRNAs and related genetic variants with risk of prostate cancer (PCa) as well as to elucidate biological mechanisms underlying the associations. By using the MiOncoCirc database, we first compared expression levels of circRNAs between 25 paired PCa and adjacent normal tissues to identify risk-associated circRNAs. We then used logistic regression models to evaluate associations between genetic variants in candidate circRNAs and PCa risk among 4662 prostate cancer patients and 3114 healthy controls, and identified circHIBADH rs11973492 as a significant risk-associated variant (odds ratio = 1.20, 95% confidence interval: 1.08-1.34, P = 7.06 × 10 -4) in a dominant genetic model, which altered the secondary structure of the corresponding RNA chain. In the in silico analysis, we found circHIBADH to sponge and silence 21 RNA-binding proteins (RPBs) enriched in the RNA splicing pathway, among which HNRNPA1 was identified and validated as a hub RBP using an external RNA-sequencing data as well as the in-house (four tissue samples) and publicly available single-cell transcriptomes. Additionally, we demonstrated that HNRNPA1 might influence hallmarks including MYC, DNA repair, and E2F target signaling pathways, thereby promoting carcinogenesis. In conclusion, genetic variants in circHIBADH may act as a sponge and inhibitor of RNA splicing-associated RBPs including HNRNPA1, playing an oncogenic role in PCa.

Self-supervised learning on millions of primary RNA sequences from 72 vertebrates improves sequence-based RNA splicing prediction.

Language models pretrained by self-supervised learning (SSL) have been widely utilized to study protein sequences, while few models were developed for genomic sequences and were limited to single species. Due to the lack of genomes from different species, these models cannot effectively leverage evolutionary information. In this study, we have developed SpliceBERT, a language model pretrained on primary ribonucleic acids (RNA) sequences from 72 vertebrates by masked language modeling, and applied it to sequence-based modeling of RNA splicing. Pretraining SpliceBERT on diverse species enables effective identification of evolutionarily conserved elements. Meanwhile, the learned hidden states and attention weights can characterize the biological properties of splice sites. As a result, SpliceBERT was shown effective on several downstream tasks: zero-shot prediction of variant effects on splicing, prediction of branchpoints in humans, and cross-species prediction of splice sites. Our study highlighted the importance of pretraining genomic language models on a diverse range of species and suggested that SSL is a promising approach to enhance our understanding of the regulatory logic underlying genomic sequences.