Genome-wide association between branch point properties and alternative splicing

Most human introns are recognized via multiple and tissue-specific branchpoints

Pineda and Bradley demonstrate that almost all human introns contain multiple branchpoints. Approximately three-quarters of constitutive introns exhibit tissue-specific branchpoint usage.

Although branchpoint recognition is an essential component of intron excision during the RNA splicing process, the branchpoint itself is frequently assumed to be a basal, rather than regulatory, sequence feature. However, this assumption has not been systematically tested due to the technical difficulty of identifying branchpoints and quantifying their usage. Here, we analyzed ∼1.31 trillion reads from 17,164 RNA sequencing data sets to demonstrate that almost all human introns contain multiple branchpoints. This complexity holds even for constitutive introns, 95% of which contain multiple branchpoints, with an estimated five to six branchpoints per intron. Introns upstream of the highly regulated ultraconserved poison exons of SR genes contain twice as many branchpoints as the genomic average. Approximately three-quarters of constitutive introns exhibit tissue-specific branchpoint usage. In an extreme example, we observed a complete switch in branchpoint usage in the well-studied first intron of HBB (β-globin) in normal bone marrow versus metastatic prostate cancer samples. Our results indicate that the recognition of most introns is unexpectedly complex and tissue-specific and suggest that alternative splicing catalysis typifies the majority of introns even in the absence of differences in the mature mRNA.

Modeling positional effects of regulatory sequences with spline transformations increases prediction accuracy of deep neural networks

Motivation

Regulatory sequences are not solely defined by their nucleic acid sequence but also by their relative distances to genomic landmarks such as transcription start site, exon boundaries or polyadenylation site. Deep learning has become the approach of choice for modeling regulatory sequences because of its strength to learn complex sequence features. However, modeling relative distances to genomic landmarks in deep neural networks has not been addressed.

Results

Here we developed spline transformation, a neural network module based on splines to flexibly and robustly model distances. Modeling distances to various genomic landmarks with spline transformations significantly increased state-of-the-art prediction accuracy of in vivo RNA-binding protein binding sites for 120 out of 123 proteins. We also developed a deep neural network for human splice branchpoint based on spline transformations that outperformed the current best, already distance-based, machine learning model. Compared to piecewise linear transformation, as obtained by composition of rectified linear units, spline transformation yields higher prediction accuracy as well as faster and more robust training. As spline transformation can be applied to further quantities beyond distances, such as methylation or conservation, we foresee it as a versatile component in the genomics deep learning toolbox.

Availability and implementation

Spline transformation is implemented as a Keras layer in the CONCISE python package: https://github.com/gagneurlab/concise. Analysis code is available at https://github.com/gagneurlab/Manuscript_Avsec_Bioinformatics_2017.

Contact

avsec@in.tum.de or gagneur@in.tum.de.

Supplementary information

Supplementary data are available at Bioinformatics online.

Landscape of the spliced leader trans-splicing mechanism in Schistosoma mansoni

Spliced leader dependent trans-splicing (SLTS) has been described as an important RNA regulatory process that occurs in different organisms, including the trematode Schistosoma mansoni. We identified more than seven thousand putative SLTS sites in the parasite, comprising genes with a wide spectrum of functional classes, which underlines the SLTS as a ubiquitous mechanism in the parasite. Also, SLTS gene expression levels span several orders of magnitude, showing that SLTS frequency is not determined by the expression level of the target gene, but by the presence of particular gene features facilitating or hindering the trans-splicing mechanism. Our in-depth investigation of SLTS events demonstrates widespread alternative trans-splicing (ATS) acceptor sites occurring in different regions along the entire gene body, highlighting another important role of SLTS generating alternative RNA isoforms in the parasite, besides the polycistron resolution. Particularly for introns where SLTS directly competes for the same acceptor substrate with cis-splicing, we identified for the first time additional and important features that might determine the type of splicing. Our study substantially extends the current knowledge of RNA processing by SLTS in S. mansoni, and provide basis for future studies on the trans-splicing mechanism in other eukaryotes.

Invention and Early History of Exon Skipping and Splice Modulation

Since its discovery in 1977, much has been known about RNA splicing and how it plays a central role in human development, function, and, notably, disease. Defects in RNA splicing account for at least 10% of all genetic disorders, with the number expected to increase as more information is uncovered on the contribution of noncoding genomic regions to disease. Splice modulation through the use of antisense oligonucleotides (AOs) has emerged as a promising avenue for the treatment of these disorders. In fact, two splice-switching AOs have recently obtained approval from the US Food and Drug Administration: eteplirsen (Exondys 51) for Duchenne muscular dystrophy, and nusinersen (Spinraza) for spinal muscular atrophy. These work by exon skipping and exon inclusion, respectively. In this chapter, we discuss the early development of AO-based splice modulation therapy-its invention, first applications, and its evolution into the approach we are now familiar with. We give a more extensive history of exon skipping in particular, as it is the splice modulation approach given the most focus in this book.

CHARGE syndrome: a recurrent hotspot of mutations in CHD7 IVS25 analyzed by bioinformatic tools and minigene assays

CHARGE syndrome is a rare genetic disorder mainly due to de novo and private truncating mutations of CHD7 gene. Here we report an intriguing hot spot of intronic mutations (c.5405-7G > A, c.5405-13G > A, c.5405-17G > A and c.5405-18C > A) located in CHD7 IVS25. Combining computational in silico analysis, experimental branch-point determination and in vitro minigene assays, our study explains this mutation hot spot by a particular genomic context, including the weakness of the IVS25 natural acceptor-site and an unconventional lariat sequence localized outside the common 40 bp upstream the acceptor splice site. For each of the mutations reported here, bioinformatic tools indicated a newly created 3' splice site, of which the existence was confirmed using pSpliceExpress, an easy-to-use and reliable splicing reporter tool. Our study emphasizes the idea that combining these two complementary approaches could increase the efficiency of routine molecular diagnosis.

The landscape of human mutually exclusive splicing

Mutually exclusive splicing of exons is a mechanism of functional gene and protein diversification with pivotal roles in organismal development and diseases such as Timothy syndrome, cardiomyopathy and cancer in humans. In order to obtain a first genomewide estimate of the extent and biological role of mutually exclusive splicing in humans, we predicted and subsequently validated mutually exclusive exons (MXEs) using 515 publically available RNA‐Seq datasets. Here, we provide evidence for the expression of over 855 MXEs, 42% of which represent novel exons, increasing the annotated human mutually exclusive exome more than fivefold. The data provide strong evidence for the existence of large and multi‐cluster MXEs in higher vertebrates and offer new insights into MXE evolution. More than 82% of the MXE clusters are conserved in mammals, and five clusters have homologous clusters in Drosophila. Finally, MXEs are significantly enriched in pathogenic mutations and their spatio‐temporal expression might predict human disease pathology.

Molecular basis of differential 3' splice site sensitivity to anti-tumor drugs targeting U2 snRNP

Several splicing-modulating compounds, including Sudemycins and Spliceostatin A, display anti-tumor properties. Combining transcriptome, bioinformatic and mutagenesis analyses, we delineate sequence determinants of the differential sensitivity of 3′ splice sites to these drugs. Sequences 5′ from the branch point (BP) region strongly influence drug sensitivity, with additional functional BPs reducing, and BP-like sequences allowing, drug responses. Drug-induced retained introns are typically shorter, displaying higher GC content and weaker polypyrimidine-tracts and BPs. Drug-induced exon skipping preferentially affects shorter alternatively spliced regions with weaker BPs. Remarkably, structurally similar drugs display both common and differential effects on splicing regulation, SSA generally displaying stronger effects on intron retention, and Sudemycins more acute effects on exon skipping. Collectively, our results illustrate how splicing modulation is exquisitely sensitive to the sequence context of 3′ splice sites and to small structural differences between drugs.

Several families of natural compounds target core components of the pre-mRNA splicing machinery and display anti-tumor activity. Here the authors show that particular sequence features can be linked to drug response, and that drugs with very similar chemical structures display substantially different effects on splicing regulation.

Predicting human splicing branchpoints by combining sequence-derived features and multi-label learning methods

Background

Alternative splicing is the critical process in a single gene coding, which removes introns and joins exons, and splicing branchpoints are indicators for the alternative splicing. Wet experiments have identified a great number of human splicing branchpoints, but many branchpoints are still unknown. In order to guide wet experiments, we develop computational methods to predict human splicing branchpoints.

Results

Considering the fact that an intron may have multiple branchpoints, we transform the branchpoint prediction as the multi-label learning problem, and attempt to predict branchpoint sites from intron sequences. First, we investigate a variety of intron sequence-derived features, such as sparse profile, dinucleotide profile, position weight matrix profile, Markov motif profile and polypyrimidine tract profile. Second, we consider several multi-label learning methods: partial least squares regression, canonical correlation analysis and regularized canonical correlation analysis, and use them as the basic classification engines. Third, we propose two ensemble learning schemes which integrate different features and different classifiers to build ensemble learning systems for the branchpoint prediction. One is the genetic algorithm-based weighted average ensemble method; the other is the logistic regression-based ensemble method.

Conclusions

In the computational experiments, two ensemble learning methods outperform benchmark branchpoint prediction methods, and can produce high-accuracy results on the benchmark dataset.

A heuristic model for computational prediction of human branch point sequence

BACKGROUND

Pre-mRNA splicing is the removal of introns from precursor mRNAs (pre-mRNAs) and the concurrent ligation of the flanking exons to generate mature mRNA. This process is catalyzed by the spliceosome, where the splicing factor 1 (SF1) specifically recognizes the seven-nucleotide branch point sequence (BPS) and the U2 snRNP later displaces the SF1 and binds to the BPS. In mammals, the degeneracy of BPS motifs together with the lack of a large set of experimentally verified BPSs complicates the task of BPS prediction in silico.

RESULTS

In this paper, we develop a simple and yet efficient heuristic model for human BPS prediction based on a novel scoring scheme, which quantifies the splicing strength of putative BPSs. The candidate BPS is restricted exclusively within a defined BPS search region to avoid the influences of other elements in the intron and therefore the prediction accuracy is improved. Moreover, using two types of relative frequencies for human BPS prediction, we demonstrate our model outperformed other current implementations on experimentally verified human introns.

CONCLUSION

We propose that the binding energy contributes to the molecular recognition involved in human pre-mRNA splicing. In addition, a genome-wide human BPS prediction is carried out. The characteristics of predicted BPSs are in accordance with experimentally verified human BPSs, and branch site positions relative to the 3'ss and the 5'end of the shortened AGEZ are consistent with the results of published papers. Meanwhile, a webserver for BPS predictor is freely available at http://biocomputer.bio.cuhk.edu.hk/BPS .

Multilayered Control of Alternative Splicing Regulatory Networks by Transcription Factors

Networks of coordinated alternative splicing (AS) events play critical roles in development and disease. However, a comprehensive knowledge of the factors that regulate these networks is lacking. We describe a high-throughput system for systematically linking trans-acting factors to endogenous RNA regulatory events. Using this system, we identify hundreds of factors associated with diverse regulatory layers that positively or negatively control AS events linked to cell fate. Remarkably, more than one-third of the regulators are transcription factors. Further analyses of the zinc finger protein Zfp871 and BTB/POZ domain transcription factor Nacc1, which regulate neural and stem cell AS programs, respectively, reveal roles in controlling the expression of specific splicing regulators. Surprisingly, these proteins also appear to regulate target AS programs via binding RNA. Our results thus uncover a large "missing cache" of splicing regulators among annotated transcription factors, some of which dually regulate AS through direct and indirect mechanisms.

Whole gene sequencing identifies deep-intronic variants with potential functional impact in patients with hypertrophic cardiomyopathy

BACKGROUND

High throughput sequencing technologies have revolutionized the identification of mutations responsible for genetic diseases such as hypertrophic cardiomyopathy (HCM). However, approximately 50% of individuals with a clinical diagnosis of HCM have no causal mutation identified. This may be due to the presence of pathogenic mutations located deep within the introns, which are not detected by conventional sequencing analysis restricted to exons and exon-intron boundaries.

OBJECTIVE

The aim of this study was to develop a whole-gene sequencing strategy to prioritize deep intronic variants that may play a role in HCM pathogenesis.

METHODS AND RESULTS

The full genomic DNA sequence of 26 genes previously associated with HCM was analysed in 16 unrelated patients. We identified likely pathogenic deep intronic variants in VCL, PRKAG2 and TTN genes. These variants, which are predicted to act through disruption of either splicing or transcription factor binding sites, are 3-fold more frequent in our cohort of probands than in normal European populations. Moreover, we found a patient that is compound heterozygous for a splice site mutation in MYBPC3 and the deep intronic VCL variant. Analysis of family members revealed that carriers of the MYBPC3 mutation alone do not manifest the disease, while family members that are compound heterozygous are clinically affected.

CONCLUSION

This study provides a framework for scrutinizing variation along the complete intronic sequence of HCM-associated genes and prioritizing candidates for mechanistic and functional analysis. Our data suggest that deep intronic variation contributes to HCM phenotype.

DDX54 regulates transcriptome dynamics during DNA damage response

The cellular response to genotoxic stress is mediated by a well-characterized network of DNA surveillance pathways. The contribution of post-transcriptional gene regulatory networks to the DNA damage response (DDR) has not been extensively studied. Here, we systematically identified RNA-binding proteins differentially interacting with polyadenylated transcripts upon exposure of human breast carcinoma cells to ionizing radiation (IR). Interestingly, more than 260 proteins, including many nucleolar proteins, showed increased binding to poly(A)+ RNA in IR-exposed cells. The functional analysis of DDX54, a candidate genotoxic stress responsive RNA helicase, revealed that this protein is an immediate-to-early DDR regulator required for the splicing efficacy of its target IR-induced pre-mRNAs. Upon IR exposure, DDX54 acts by increased interaction with a well-defined class of pre-mRNAs that harbor introns with weak acceptor splice sites, as well as by protein-protein contacts within components of U2 snRNP and spliceosomal B complex, resulting in lower intron retention and higher processing rates of its target transcripts. Because DDX54 promotes survival after exposure to IR, its expression and/or mutation rate may impact DDR-related pathologies. Our work indicates the relevance of many uncharacterized RBPs potentially involved in the DDR.

The determinants of alternative RNA splicing in human cells

Alternative splicing represents an important level of the regulation of gene function in eukaryotic organisms. It plays a critical role in virtually every biological process within an organism, including regulation of cell division and cell death, differentiation of tissues in the embryo and the adult organism, as well as in cellular response to diverse environmental factors. In turn, studies of the last decade have shown that alternative splicing itself is controlled by different mechanisms. Unfortunately, there is no clear understanding of how these diverse mechanisms, or determinants, regulate and constrain the set of alternative RNA species produced from any particular gene in every cell of the human body. Here, we provide a consolidated overview of alternative splicing determinants including RNA-protein interactions, epigenetic regulation via chromatin remodeling, coupling of transcription-to-alternative splicing, effect of secondary structures in pre-RNA, and function of the RNA quality control systems. We also extensively and critically discuss some mechanistic insights on coordinated inclusion/exclusion of exons during the formation of mature RNA molecules. We conclude that the final structure of RNA is pre-determined by a complex interplay between cis- and trans-acting factors. Altogether, currently available empirical data significantly expand our understanding of the functioning of the alternative splicing machinery of cells in normal and pathological conditions. On the other hand, there are still many blind spots that require further deep investigations.

Protein 4.1R Exon 16 3' Splice Site Activation Requires Coordination among TIA1, Pcbp1, and RBM39 during Terminal Erythropoiesis

Exon 16 of protein 4.1R encodes a spectrin/actin-binding peptide critical for erythrocyte membrane stability. Its expression during erythroid differentiation is regulated by alternative pre-mRNA splicing. A UUUUCCCCCC motif situated between the branch point and the 3' splice site is crucial for inclusion. We show that the UUUU region and the last three C residues in this motif are necessary for the binding of splicing factors TIA1 and Pcbp1 and that these proteins appear to act in a collaborative manner to enhance exon 16 inclusion. This element also activates an internal exon when placed in a corresponding intronic position in a heterologous reporter. The impact of these two factors is further enhanced by high levels of RBM39, whose expression rises during erythroid differentiation as exon 16 inclusion increases. TIA1 and Pcbp1 associate in a complex containing RBM39, which interacts with U2AF65 and SF3b155 and promotes U2 snRNP recruitment to the branch point. Our results provide a mechanism for exon 16 3' splice site activation in which a coordinated effort among TIA1, Pcbp1, and RBM39 stabilizes or increases U2 snRNP recruitment, enhances spliceosome A complex formation, and facilitates exon definition through RBM39-mediated splicing regulation.

A birth of bipartite exon by intragenic deletion

BACKGROUND

Disease-causing mutations that activate transposon-derived exons without creating a new splice-site consensus have been reported rarely, but they provided unique insights into our understanding of structural motifs required for inclusion of intronic sequences in mature transcripts.

METHODS

We employ a combination of experimental and computational techniques to characterize the first de novo bipartite exon activation in genetic disease.

RESULTS

The exon originated from two separate introns as a result of an in-frame COL4A5 deletion associated with a typical Alport syndrome. The deletion encompassed exons 38 through 41 and activated a cryptic 3' and 5' splice site that were derived from intron 37 and intron 41, respectively. The deletion breakpoint was in the middle of the new exon, with considerable complementarity between the two exonic parts, potentially bringing the cryptic 3' and 5' splice site into proximity. The 3' splice site, polypyrimidine tract and the branch site of the new exon were derived from an inactive, 5' truncated LINE-1 retrotransposon. This ancient LINE-1 copy sustained a series of mutations that created the highly conserved AG dinucleotide at the 3' splice site early in primate development. The exon was fully included in mature transcripts and introduced a stop codon in the shortened COL4A5 mRNA, illustrating pitfalls of inferring disease severity from DNA mutation alone.

CONCLUSION

These results expand the repertoire of mutational mechanisms that alter RNA processing in genetic disease and illustrate the extraordinary versatility of transposed elements in shaping the new exon-intron structure and the phenotypic variability.

Role and convergent evolution of competing RNA secondary structures in mutually exclusive splicing

Exon or cassette duplication is an important means of expanding protein and functional diversity through mutually exclusive splicing. However, the mechanistic basis of this process in non-arthropod species remains poorly understood. Here, we demonstrate that MRP1 genes underwent tandem exon duplication in Nematoda, Platyhelminthes, Annelida, Mollusca, Arthropoda, Echinodermata, and early-diverging Chordata but not in late-diverging vertebrates. Interestingly, these events were of independent origin in different phyla, suggesting convergent evolution of alternative splicing. Furthermore, we showed that multiple sets of clade-conserved RNA pairings evolved to guide species-specific mutually exclusive splicing in Arthropoda. Importantly, we also identified a similar structural code in MRP exon clusters of the annelid, Capitella teleta, and chordate, Branchiostoma belcheri, suggesting an evolutionarily conserved competing pairing-guided mechanism in bilaterians. Taken together, these data reveal the molecular determinants and RNA pairing-guided evolution of species-specific mutually exclusive splicing spanning more than 600 million years of bilaterian evolution. These findings have a significant impact on our understanding of the evolution of and mechanism underpinning isoform diversity and complex gene structure.

Large-scale analysis of branchpoint usage across species and cell lines

The coding sequence of each human pre-mRNA is interrupted, on average, by 11 introns that must be spliced out for proper gene expression. Each intron contains three obligate signals: a 5′ splice site, a branch site, and a 3′ splice site. Splice site usage has been mapped exhaustively across different species, cell types, and cellular states. In contrast, only a small fraction of branch sites have been identified even once. The few reported annotations of branch site are imprecise as reverse transcriptase skips several nucleotides while traversing a 2–5 linkage. Here, we report large-scale mapping of the branchpoints from deep sequencing data in three different species and in the SF3B1 K700E oncogenic mutant background. We have developed a novel method whereby raw lariat reads are refined by U2snRNP/pre-mRNA base-pairing models to return the largest current data set of branchpoint sequences with quality metrics. This analysis discovers novel modes of U2snRNA:pre-mRNA base-pairing conserved in yeast and provides insight into the biogenesis of intron circles. Finally, matching branch site usage with isoform selection across the extensive panel of ENCODE RNA-seq data sets offers insight into the mechanisms by which branchpoint usage drives alternative splicing.

Global analysis of pre-mRNA subcellular localization following splicing inhibition by spliceostatin A

Spliceostatin A (SSA) is a methyl ketal derivative of FR901464, a potent antitumor compound isolated from a culture broth of Pseudomonas sp no. 2663. These compounds selectively bind to the essential spliceosome component SF3b, a subcomplex of the U2 snRNP, to inhibit pre-mRNA splicing. However, the mechanism of SSA's antitumor activity is unknown. It is noteworthy that SSA causes accumulation of a truncated form of the CDK inhibitor protein p27 translated from CDKN1B pre-mRNA, which is involved in SSA-induced cell-cycle arrest. However, it is still unclear whether pre-mRNAs are uniformly exported from the nucleus following SSA treatment. We performed RNA-seq analysis on nuclear and cytoplasmic fractions of SSA-treated cells. Our statistical analyses showed that intron retention is the major consequence of SSA treatment, and a small number of intron-containing pre-mRNAs leak into the cytoplasm. Using a series of reporter plasmids to investigate the roles of intronic sequences in the pre-mRNA leakage, we showed that the strength of the 5' splice site affects pre-mRNA leakage. Additionally, we found that the level of pre-mRNA leakage is related to transcript length. These results suggest that the strength of the 5' splice site and the length of the transcripts are determinants of the pre-mRNA leakage induced by SF3b inhibitors.

The role of alternative splicing in cancer

The functional capacity of cells is defined by the transcriptome. Many recent studies have identified variations in the transcriptome of tumors due to alternative splicing changes, as well as mutations in splicing factors and regulatory signals in most tumor types. Some of these alterations have been linked to tumor progression, metastasis, therapy resistance, and other oncogenic processes. Here, we describe the different mechanisms that drive splicing changes in tumors and their impact in cancer. Motivated by the current evidence, we propose a model whereby a subset of the splicing patterns contributes to the definition of specific tumor phenotypes, and may hold potential for the development of novel clinical biomarkers and therapeutic approaches.

Cryptic splicing events in the iron transporter ABCB7 and other key target genes in SF3B1-mutant myelodysplastic syndromes

The splicing factor SF3B1 is the most frequently mutated gene in myelodysplastic syndromes (MDS), and is strongly associated with the presence of ring sideroblasts (RS). We have performed a systematic analysis of cryptic splicing abnormalities from RNA sequencing data on hematopoietic stem cells (HSCs) of SF3B1-mutant MDS cases with RS. Aberrant splicing events in many downstream target genes were identified and cryptic 3' splice site usage was a frequent event in SF3B1-mutant MDS. The iron transporter ABCB7 is a well-recognized candidate gene showing marked downregulation in MDS with RS. Our analysis unveiled aberrant ABCB7 splicing, due to usage of an alternative 3' splice site in MDS patient samples, giving rise to a premature termination codon in the ABCB7 mRNA. Treatment of cultured SF3B1-mutant MDS erythroblasts and a CRISPR/Cas9-generated SF3B1-mutant cell line with the nonsense-mediated decay (NMD) inhibitor cycloheximide showed that the aberrantly spliced ABCB7 transcript is targeted by NMD. We describe cryptic splicing events in the HSCs of SF3B1-mutant MDS, and our data support a model in which NMD-induced downregulation of the iron exporter ABCB7 mRNA transcript resulting from aberrant splicing caused by mutant SF3B1 underlies the increased mitochondrial iron accumulation found in MDS patients with RS.

Large-scale remodeling of a repressed exon ribonucleoprotein to an exon definition complex active for splicing

Polypyrimidine-tract binding protein PTBP1 can repress splicing during the exon definition phase of spliceosome assembly, but the assembly steps leading to an exon definition complex (EDC) and how PTBP1 might modulate them are not clear. We found that PTBP1 binding in the flanking introns allowed normal U2AF and U1 snRNP binding to the target exon splice sites but blocked U2 snRNP assembly in HeLa nuclear extract. Characterizing a purified PTBP1-repressed complex, as well as an active early complex and the final EDC by SILAC-MS, we identified extensive PTBP1-modulated changes in exon RNP composition. The active early complex formed in the absence of PTBP1 proceeded to assemble an EDC with the eviction of hnRNP proteins, the late recruitment of SR proteins, and binding of the U2 snRNP. These results demonstrate that during early stages of splicing, exon RNP complexes are highly dynamic with many proteins failing to bind during PTBP1 arrest.

DOI:http://dx.doi.org/10.7554/eLife.19743.001

Integrative classification of human coding and noncoding genes through RNA metabolism profiles

Pervasive transcription of the human genome results in a heterogeneous mix of coding RNAs and long noncoding RNAs (lncRNAs). Only a small fraction of lncRNAs have demonstrated regulatory functions, thus making functional lncRNAs difficult to distinguish from nonfunctional transcriptional byproducts. This difficulty has resulted in numerous competing human lncRNA classifications that are complicated by a steady increase in the number of annotated lncRNAs. To address these challenges, we quantitatively examined transcription, splicing, degradation, localization and translation for coding and noncoding human genes. We observed that annotated lncRNAs had lower synthesis and higher degradation rates than mRNAs and discovered mechanistic differences explaining slower lncRNA splicing. We grouped genes into classes with similar RNA metabolism profiles, containing both mRNAs and lncRNAs to varying extents. These classes exhibited distinct RNA metabolism, different evolutionary patterns and differential sensitivity to cellular RNA-regulatory pathways. Our classification provides an alternative to genomic context-driven annotations of lncRNAs.

The RNA-binding landscape of RBM10 and its role in alternative splicing regulation in models of mouse early development

Mutations in the RNA-binding protein, RBM10, result in a human syndromic form of cleft palate, termed TARP syndrome. A role for RBM10 in alternative splicing regulation has been previously demonstrated in human cell lines. To uncover the cellular functions of RBM10 in a cell line that is relevant to the phenotype observed in TARP syndrome, we used iCLIP to identify its endogenous RNA targets in a mouse embryonic mandibular cell line. We observed that RBM10 binds to pre-mRNAs with significant enrichment in intronic regions, in agreement with a role for this protein in pre-mRNA splicing. In addition to protein-coding transcripts, RBM10 also binds to a variety of cellular RNAs, including non-coding RNAs, such as spliceosomal small nuclear RNAs, U2 and U12. RNA-seq was used to investigate changes in gene expression and alternative splicing in RBM10 KO mouse mandibular cells and also in mouse ES cells. We uncovered a role for RBM10 in the regulation of alternative splicing of common transcripts in both cell lines but also identified cell-type specific events. Importantly, those pre-mRNAs that display changes in alternative splicing also contain RBM10 iCLIP tags, suggesting a direct role of RBM10 in these events. Finally, we show that depletion of RBM10 in mouse ES cells leads to proliferation defects and to gross alterations in their differentiation potential. These results demonstrate a role for RBM10 in the regulation of alternative splicing in two cell models of mouse early development and suggests that mutations in RBM10 could lead to splicing changes that affect normal palate development and cause human disease.

Alternative splicing of U2AF1 reveals a shared repression mechanism for duplicated exons

The auxiliary factor of U2 small nuclear ribonucleoprotein (U2AF) facilitates branch point (BP) recognition and formation of lariat introns. The gene for the 35-kD subunit of U2AF gives rise to two protein isoforms (termed U2AF35a and U2AF35b) that are encoded by alternatively spliced exons 3 and Ab, respectively. The splicing recognition sequences of exon 3 are less favorable than exon Ab, yet U2AF35a expression is higher than U2AF35b across tissues. We show that U2AF35b repression is facilitated by weak, closely spaced BPs next to a long polypyrimidine tract of exon Ab. Each BP lacked canonical uridines at position -2 relative to the BP adenines, with efficient U2 base-pairing interactions predicted only for shifted registers reminiscent of programmed ribosomal frameshifting. The BP cluster was compensated by interactions involving unpaired cytosines in an upstream, EvoFold-predicted stem loop (termed ESL) that binds FUBP1/2. Exon Ab inclusion correlated with predicted free energies of mutant ESLs, suggesting that the ESL operates as a conserved rheostat between long inverted repeats upstream of each exon. The isoform-specific U2AF35 expression was U2AF65-dependent, required interactions between the U2AF-homology motif (UHM) and the α6 helix of U2AF35, and was fine-tuned by exon Ab/3 variants. Finally, we identify tandem homologous exons regulated by U2AF and show that their preferential responses to U2AF65-related proteins and SRSF3 are associated with unpaired pre-mRNA segments upstream of U2AF-repressed 3′ss. These results provide new insights into tissue-specific subfunctionalization of duplicated exons in vertebrate evolution and expand the repertoire of exon repression mechanisms that control alternative splicing.

Cancer-associated SF3B1 mutants recognize otherwise inaccessible cryptic 3' splice sites within RNA secondary structures

Recurrent mutations in core splicing factors have been reported in several clonal disorders, including cancers. Mutations in SF3B1, a component of the U2 splicing complex, are the most common. SF3B1 mutations are associated with aberrant pre-mRNA splicing using cryptic 3’ splice sites (3’SS) but the mechanism of their selection is not clear. To understand how cryptic 3’SS are selected, we performed comprehensive analysis of transcriptome-wide changes to splicing and gene expression associated with SF3B1 mutations in patient samples as well as an experimental model of inducible expression. Hundreds of cryptic 3’SS were detectable across the genome in cells expressing mutant SF3B1. These 3’SS are typically sequestered within RNA secondary structures and poorly accessible compared to their corresponding canonical 3’SS. We hypothesized that these cryptic 3’SS are inaccessible during normal splicing catalysis and that this constraint is overcome in spliceosomes containing mutant SF3B1. This model of secondary structure-dependent selection of cryptic 3’SS was found across multiple clonal processes associated with SF3B1 mutations (myelodysplastic syndrome and chronic lymphocytic leukemia). We validated our model predictions in mini-gene splicing assays. Additionally, we found deregulated expression of proteins with relevant functions in splicing factor-related diseases both in association with aberrant splicing and without corresponding splicing changes. Our results show that SF3B1 mutations are associated with a distinct splicing program shared across multiple clonal processes and define a biochemical mechanism for altered 3’SS choice.

High Resolution Genomic Scans Reveal Genetic Architecture Controlling Alcohol Preference in Bidirectionally Selected Rat Model

Investigations on the influence of nature vs. nurture on Alcoholism (Alcohol Use Disorder) in human have yet to provide a clear view on potential genomic etiologies. To address this issue, we sequenced a replicated animal model system bidirectionally-selected for alcohol preference (AP). This model is uniquely suited to map genetic effects with high reproducibility, and resolution. The origin of the rat lines (an 8-way cross) resulted in small haplotype blocks (HB) with a corresponding high level of resolution. We sequenced DNAs from 40 samples (10 per line of each replicate) to determine allele frequencies and HB. We achieved ~46X coverage per line and replicate. Excessive differentiation in the genomic architecture between lines, across replicates, termed signatures of selection (SS), were classified according to gene and region. We identified SS in 930 genes associated with AP. The majority (50%) of the SS were confined to single gene regions, the greatest numbers of which were in promoters (284) and intronic regions (169) with the least in exon's (4), suggesting that differences in AP were primarily due to alterations in regulatory regions. We confirmed previously identified genes and found many new genes associated with AP. Of those newly identified genes, several demonstrated neuronal function involved in synaptic memory and reward behavior, e.g. ion channels (Kcnf1, Kcnn3, Scn5a), excitatory receptors (Grin2a, Gria3, Grip1), neurotransmitters (Pomc), and synapses (Snap29). This study not only reveals the polygenic architecture of AP, but also emphasizes the importance of regulatory elements, consistent with other complex traits.

Alcohol Used Disorder (AUD) or Alcoholism extracts a great societal cost in terms of human suffering. Understanding the genetic basis is critical to comprehend, treat and prevent this disease, but difficult in humans, as choice is influenced by nature and nurture. To discover its genetic basis, we used an animal model system that controlled for genetic and non-genetic factors through randomization, study replication, long-term divergent selection, and a controlled environment. We conducted whole genome sequencing in breeds that were either compulsive excessive drinkers or completely abstinent. We discovered consistent alterations in several genes and neurological pathways previously unassociated with alcoholism. These results strengthened our understanding of the genetic basis of alcoholism and revealed potential genetic- and neurological-based treatments.

The alternative splicing program of differentiated smooth muscle cells involves concerted non-productive splicing of post-transcriptional regulators

Alternative splicing (AS) is a key component of gene expression programs that drive cellular differentiation. Smooth muscle cells (SMCs) are important in the function of a number of physiological systems; however, investigation of SMC AS has been restricted to a handful of events. We profiled transcriptome changes in mouse de-differentiating SMCs and observed changes in hundreds of AS events. Exons included in differentiated cells were characterized by particularly weak splice sites and by upstream binding sites for Polypyrimidine Tract Binding protein (PTBP1). Consistent with this, knockdown experiments showed that that PTBP1 represses many smooth muscle specific exons. We also observed coordinated splicing changes predicted to downregulate the expression of core components of U1 and U2 snRNPs, splicing regulators and other post-transcriptional factors in differentiated cells. The levels of cognate proteins were lower or similar in differentiated compared to undifferentiated cells. However, levels of snRNAs did not follow the expression of splicing proteins, and in the case of U1 snRNP we saw reciprocal changes in the levels of U1 snRNA and U1 snRNP proteins. Our results suggest that the AS program in differentiated SMCs is orchestrated by the combined influence of auxiliary RNA binding proteins, such as PTBP1, along with altered activity and stoichiometry of the core splicing machinery.

Rapid evolutionary turnover underlies conserved lncRNA-genome interactions

Many long noncoding RNAs (lncRNAs) can regulate chromatin states, but the evolutionary origin and dynamics driving lncRNA-genome interactions are unclear. We adapted an integrative strategy that identifies lncRNA orthologs in different species despite limited sequence similarity, which is applicable to mammalian and insect lncRNAs. Analysis of the roX lncRNAs, which are essential for dosage compensation of the single X chromosome in Drosophila males, revealed 47 new roX orthologs in diverse Drosophilid species across ∼40 million years of evolution. Genetic rescue by roX orthologs and engineered synthetic lncRNAs showed that altering the number of focal, repetitive RNA structures determines roX ortholog function. Genomic occupancy maps of roX RNAs in four species revealed conserved targeting of X chromosome neighborhoods but rapid turnover of individual binding sites. Many new roX-binding sites evolved from DNA encoding a pre-existing RNA splicing signal, effectively linking dosage compensation to transcribed genes. Thus, dynamic change in lncRNAs and their genomic targets underlies conserved and essential lncRNA-genome interactions.

Polymorphic tandem repeats within gene promoters act as modifiers of gene expression and DNA methylation in humans

Despite representing an important source of genetic variation, tandem repeats (TRs) remain poorly studied due to technical difficulties. We hypothesized that TRs can operate as expression (eQTLs) and methylation (mQTLs) quantitative trait loci. To test this we analyzed the effect of variation at 4849 promoter-associated TRs, genotyped in 120 individuals, on neighboring gene expression and DNA methylation. Polymorphic promoter TRs were associated with increased variance in local gene expression and DNA methylation, suggesting functional consequences related to TR variation. We identified >100 TRs associated with expression/methylation levels of adjacent genes. These potential eQTL/mQTL TRs were enriched for overlaps with transcription factor binding and DNaseI hypersensitivity sites, providing a rationale for their effects. Moreover, we showed that most TR variants are poorly tagged by nearby single nucleotide polymorphisms (SNPs) markers, indicating that many functional TR variants are not effectively assayed by SNP-based approaches. Our study assigns biological significance to TR variations in the human genome, and suggests that a significant fraction of TR variations exert functional effects via alterations of local gene expression or epigenetics. We conclude that targeted studies that focus on genotyping TR variants are required to fully ascertain functional variation in the genome.

IntSplice: prediction of the splicing consequences of intronic single-nucleotide variations in the human genome

Precise spatiotemporal regulation of splicing is mediated by splicing cis-elements on pre-mRNA. Single-nucleotide variations (SNVs) affecting intronic cis-elements possibly compromise splicing, but no efficient tool has been available to identify them. Following an effect-size analysis of each intronic nucleotide on annotated alternative splicing, we extracted 105 parameters that could affect the strength of the splicing signals. However, we could not generate reliable support vector regression models to predict the percent-splice-in (PSI) scores for normal human tissues. Next, we generated support vector machine (SVM) models using 110 parameters to directly differentiate pathogenic SNVs in the Human Gene Mutation Database and normal SNVs in the dbSNP database, and we obtained models with a sensitivity of 0.800±0.041 (mean and s.d.) and a specificity of 0.849±0.021. Our IntSplice models were more discriminating than SVM models that we generated with Shapiro-Senapathy score and MaxEntScan::score3ss. We applied IntSplice to a naturally occurring and nine artificial intronic mutations in RAPSN causing congenital myasthenic syndrome. IntSplice correctly predicted the splicing consequences for nine of the ten mutants. We created a web service program, IntSplice (http://www.med.nagoya-u.ac.jp/neurogenetics/IntSplice) to predict splicing-affecting SNVs at intronic positions from -50 to -3.

Axial Spondylometaphyseal Dysplasia Is Caused by C21orf2 Mutations

Axial spondylometaphyseal dysplasia (axial SMD) is an autosomal recessive disease characterized by dysplasia of axial skeleton and retinal dystrophy. We conducted whole exome sequencing and identified C21orf2 (chromosome 21 open reading frame 2) as a disease gene for axial SMD. C21orf2 mutations have been recently found to cause isolated retinal degeneration and Jeune syndrome. We found a total of five biallelic C21orf2 mutations in six families out of nine: three missense and two splicing mutations in patients with various ethnic backgrounds. The pathogenic effects of the splicing (splice-site and branch-point) mutations were confirmed on RNA level, which showed complex patterns of abnormal splicing. C21orf2 mutations presented with a wide range of skeletal phenotypes, including cupped and flared anterior ends of ribs, lacy ilia and metaphyseal dysplasia of proximal femora. Analysis of patients without C21orf2 mutation indicated genetic heterogeneity of axial SMD. Functional data in chondrocyte suggest C21orf2 is implicated in cartilage differentiation. C21orf2 protein was localized to the connecting cilium of the cone and rod photoreceptors, confirming its significance in retinal function. Our study indicates that axial SMD is a member of a unique group of ciliopathy affecting skeleton and retina.

The role of gene polymorphism in HLA class I splicing

Among the large number of human leucocyte antigen (HLA) alleles, only a few have been identified with a nucleotide polymorphism impairing correct splicing. Those alleles show aberrant expression levels, due to either a direct effect of the polymorphism on the normal splice site or to the creation of an alternative splice site. Furthermore, in several studies, the presence of alternatively spliced HLA transcripts co-expressed with the mature spliced transcripts was reported. We evaluated the splice site sequences of all known HLA class I alleles and found that, beside the consensus GT and AG sequences at the intron borders, there were some other highly conserved nucleotides for the different class I genes. In this review, we summarize the splicing mechanism and evaluate what is known today about alternative splicing of HLA class I genes.

Cancer-associated SF3B1 mutations affect alternative splicing by promoting alternative branchpoint usage

Hotspot mutations in the spliceosome gene SF3B1 are reported in ∼20% of uveal melanomas. SF3B1 is involved in 3'-splice site (3'ss) recognition during RNA splicing; however, the molecular mechanisms of its mutation have remained unclear. Here we show, using RNA-Seq analyses of uveal melanoma, that the SF3B1(R625/K666) mutation results in deregulated splicing at a subset of junctions, mostly by the use of alternative 3'ss. Modelling the differential junctions in SF3B1(WT) and SF3B1(R625/K666) cell lines demonstrates that the deregulated splice pattern strictly depends on SF3B1 status and on the 3'ss-sequence context. SF3B1(WT) knockdown or overexpression do not reproduce the SF3B1(R625/K666) splice pattern, qualifying SF3B1(R625/K666) as change-of-function mutants. Mutagenesis of predicted branchpoints reveals that the SF3B1(R625/K666)-promoted splice pattern is a direct result of alternative branchpoint usage. Altogether, this study provides a better understanding of the mechanisms underlying splicing alterations induced by mutant SF3B1 in cancer, and reveals a role for alternative branchpoints in disease.

Vemurafenib-resistant BRAF selects alternative branch points different from its wild-type BRAF in intron 8 for RNA splicing

One mechanism of resistance of the melanoma-associated BRAF kinase to its small molecule inhibitor vemurafenib is by point mutations in its intron 8 resulting in exons 4–8 skipping. In this report, we carried out in vitro BRAF RNA splicing assays and lariat RT-PCR to map the intron 8 branch points in wild-type and BRAF mutants. We identify multiple branch points (BP) in intron 8 of both wild-type (wt) and vemurafenib-resistant BRAF RNA. In wt BRAF, BPs are located at -29A, -28A and -26A, whereas in a vemurafenib-resistant BRAF splicing mutant, BPs map to -22A, -18A and -15A, proximal to the intron 8 3′ splice site. This finding of a distal-to-proximal shift of the branch point sequence in BRAF splicing in response to point-mutations in intron 8 provides insight into the regulation of BRAF alternative splicing upon vemurafenib resistance.

Deciphering targeting rules of splicing modulator compounds: case of TG003

Background

Recent advances in the development of small chemical compounds that can modulate RNA splicing brought excitement to the field of splicing-targeting therapy. Splicing-targeting therapy tries to ameliorate the disease by altering the exon combination of transcripts to reduce the undesired effect of genetic mutations. However, the knowledge and tools to understand factors contributing to splicing modulator compound sensitivity have been lacking. Our goal was to establish a method to characterize sequence features found in compound sensitive exons.

Results

Here we developed a comparative transcriptomic approach to explore features that make an exon sensitive to a chemical compound. In this study, we chose TG003, a potential drug for Duchenne muscular dystrophy, and performed RNA-sequencing on samples from human and mouse skeletal muscle cells, with and without TG003 treatments. We compared TG003 responsiveness between homologous exon pairs and identified 21 pairs in which human exons were skip-enhanced but not mouse exons. We compared the sequence features; splice site scores, number of splicing factor binding sites, and properties of branch sequence and polypyrimidine tracts, and found that polypyrimidine tracts were stronger (longer stretches and richer content of consecutive polypyrimidine) in the mouse TG003 insensitive exons. We also compared the features between TG003 skip-enhanced and insensitive exons within the species, and discovered that human TG003 skip-enhanced exons were shorter and had less splicing factor binding sites than the group of human TG003 insensitive exons. Mouse insensitive exons homologous to human TG003 skip-enhanced exons shared these properties. Our results suggested that these features are prerequisites for TG003 skip-enhanced exons and weak polypyrimidine tracts are defining features, which were supported by a decision tree analysis on all cassette exons in human.

Conclusions

In this study we established a comparative transcriptomic approach, which shed lights on how small chemical compounds modulate RNA splicing. The results described here was the first attempt to decipher the targeting rules of a splicing modulator compound. We expect that this approach would contribute to the precise understanding of the mechanism of TG003-induced splicing modulation, expand target diseases of splicing modulators in general, as well as the development of new splicing modulators.

Electronic supplementary material

The online version of this article (doi:10.1186/s12867-015-0044-6) contains supplementary material, which is available to authorized users.

The CYP2C19 Intron 2 Branch Point SNP is the Ancestral Polymorphism Contributing to the Poor Metabolizer Phenotype in Livers with CYP2C19*35 and CYP2C19*2 Alleles

CYP2C19 rs12769205 alters an intron 2 branch point adenine leading to an alternative mRNA in human liver with complete inclusion of intron 2 (exon 2B). rs12769205 changes the mRNA reading frame, introduces 87 amino acids, and leads to a premature stop codon. The 1000 Genomes project (http://browser.1000genomes.org/index.html) indicated rs12769205 is in linkage disequilibrium with rs4244285 on CYP2C19*2, but found alone on CYP2C19*35 in Blacks. Minigenes containing rs12769205 transfected into HepG2 cells demonstrated this single nucleotide polymorphism (SNP) alone leads to exon 2B and decreases CYP2C19 canonical mRNA. A residual amount of CYP2C19 protein was detectable by quantitative proteomics with tandem mass spectrometry in CYP2C19*2/*2 and *1/*35 liver microsomes with an exon 2 probe. However, an exon 4 probe, downstream from rs12769205, but upstream of rs4244285, failed to detect CYP2C19 protein in livers homozygous for rs12769205, demonstrating rs12769205 alone can lead to complete loss of CYP2C19 protein. CYP2C19 genotypes and mephenytoin phenotype were compared in 104 Ethiopians. Poor metabolism of mephenytoin was seen in persons homozygous for both rs12769205 and rs4244285 (CYP2C19*2/*2), but with little effect on mephenytoin disposition of CYP2C19*1/*2, CYP2C19*1/*3, or CYP2C19*1/*35 heterozygous alleles. Extended haplotype homozygosity tests of the HapMap Yorubans (YRI) showed both haplotypes carrying rs12769205 (CYP2C19*35 and CYP2C19*2) are under significant natural selection, with CYP2C19*35 having a higher relative extended haplotype homozygosity score. The phylogenetic tree of the YRI CYP2C19 haplotypes revealed rs12769205 arose first on CYP2C19*35 and that rs4244285 was added later, creating CYP2C19*2. In conclusion, rs12769205 is the ancestral polymorphism leading to aberrant splicing of CYP2C19*35 and CYP2C19*2 alleles in liver.

Human genomics. Effect of predicted protein-truncating genetic variants on the human transcriptome

Accurate prediction of the functional effect of genetic variation is critical for clinical genome interpretation. We systematically characterized the transcriptome effects of protein-truncating variants, a class of variants expected to have profound effects on gene function, using data from the Genotype-Tissue Expression (GTEx) and Geuvadis projects. We quantitated tissue-specific and positional effects on nonsense-mediated transcript decay and present an improved predictive model for this decay. We directly measured the effect of variants both proximal and distal to splice junctions. Furthermore, we found that robustness to heterozygous gene inactivation is not due to dosage compensation. Our results illustrate the value of transcriptome data in the functional interpretation of genetic variants.

Identification of U2AF(35)-dependent exons by RNA-Seq reveals a link between 3' splice-site organization and activity of U2AF-related proteins

The auxiliary factor of U2 small nuclear RNA (U2AF) is a heterodimer consisting of 65- and 35-kD proteins that bind the polypyrimidine tract (PPT) and AG dinucleotides at the 3′ splice site (3′ss). The gene encoding U2AF35 (U2AF1) is alternatively spliced, giving rise to two isoforms U2AF35a and U2AF35b. Here, we knocked down U2AF35 and each isoform and characterized transcriptomes of HEK293 cells with varying U2AF35/U2AF65 and U2AF35a/b ratios. Depletion of both isoforms preferentially modified alternative RNA processing events without widespread failure to recognize 3′ss or constitutive exons. Over a third of differentially used exons were terminal, resulting largely from the use of known alternative polyadenylation (APA) sites. Intronic APA sites activated in depleted cultures were mostly proximal whereas tandem 3′UTR APA was biased toward distal sites. Exons upregulated in depleted cells were preceded by longer AG exclusion zones and PPTs than downregulated or control exons and were largely activated by PUF60 and repressed by CAPERα. The U2AF(35) repression and activation was associated with a significant interchange in the average probabilities to form single-stranded RNA in the optimal PPT and branch site locations and sequences further upstream. Although most differentially used exons were responsive to both U2AF subunits and their inclusion correlated with U2AF levels, a small number of transcripts exhibited distinct responses to U2AF35a and U2AF35b, supporting the existence of isoform-specific interactions. These results provide new insights into function of U2AF and U2AF35 in alternative RNA processing.

Transcriptome sequencing reveals potential mechanism of cryptic 3' splice site selection in SF3B1-mutated cancers

Mutations in the splicing factor SF3B1 are found in several cancer types and have been associated with various splicing defects. Using transcriptome sequencing data from chronic lymphocytic leukemia, breast cancer and uveal melanoma tumor samples, we show that hundreds of cryptic 3’ splice sites (3’SSs) are used in cancers with SF3B1 mutations. We define the necessary sequence context for the observed cryptic 3’ SSs and propose that cryptic 3’SS selection is a result of SF3B1 mutations causing a shift in the sterically protected region downstream of the branch point. While most cryptic 3’SSs are present at low frequency (<10%) relative to nearby canonical 3’SSs, we identified ten genes that preferred out-of-frame cryptic 3’SSs. We show that cancers with mutations in the SF3B1 HEAT 5-9 repeats use cryptic 3’SSs downstream of the branch point and provide both a mechanistic model consistent with published experimental data and affected targets that will guide further research into the oncogenic effects of SF3B1 mutation.

A key goal of cancer genomics studies is to identify genes that are recurrently mutated at a rate above background and likely contribute to cancer development. Many such recurrently mutated genes have been identified over the last few years, but we often do not know the underlying mechanisms by which they contribute to cancer growth. Unexpectedly, several genes in the spliceosome, the collection of RNAs and proteins that remove introns from transcribed RNAs, are recurrently mutated in different cancers. Here, we have examined mutations in the splicing factor SF3B1, a key component of the spliceosome, and identified a global splicing defect present in different cancers with SF3B1 mutations by comparing the expression of splice junctions using generalized linear models. While prior studies have reported a limited number of aberrant splicing events in SF3B1-mutated cancers, we have established that SF3B1 mutations are associated with usage of hundreds of atypical splice sites at the 3’ end of the intron. We have identified nucleotide sequence requirements for these cryptic splice sites that are consistent with a proposed mechanistic model. These findings greatly expand our understanding of the effect of SF3B1 mutations on splicing and provide new targets for determining the oncogenic effect of SF3B1 mutations.

Identification of common genetic variants controlling transcript isoform variation in human whole blood

An understanding of the genetic variation underlying transcript splicing is essential to dissect the molecular mechanisms of common disease. The available evidence from splicing quantitative trait locus (sQTL) studies has been limited to small samples. We performed genome-wide screening to identify SNPs that might control mRNA splicing in whole blood collected from 5,257 Framingham Heart Study participants. We identified 572,333 cis sQTLs involving 2,650 unique genes. Many sQTL-associated genes (40%) undergo alternative splicing. Using the National Human Genome Research Institute (NHGRI) genome-wide association study (GWAS) catalog, we determined that 528 unique sQTLs were significantly enriched for 8,845 SNPs associated with traits in previous GWAS. In particular, we found 395 (4.5%) GWAS SNPs with evidence of cis sQTLs but not gene-level cis expression quantitative trait loci (eQTLs), suggesting that sQTL analysis could provide additional insights into the functional mechanism underlying GWAS results. Our findings provide an informative sQTL resource for further characterizing the potential functional roles of SNPs that control transcript isoforms relevant to common diseases.

Genome-wide discovery of human splicing branchpoints

During the splicing reaction, the 5′ intron end is joined to the branchpoint nucleotide, selecting the next exon to incorporate into the mature RNA and forming an intron lariat, which is excised. Despite a critical role in gene splicing, the locations and features of human splicing branchpoints are largely unknown. We use exoribonuclease digestion and targeted RNA-sequencing to enrich for sequences that traverse the lariat junction and, by split and inverted alignment, reveal the branchpoint. We identify 59,359 high-confidence human branchpoints in >10,000 genes, providing a first map of splicing branchpoints in the human genome. Branchpoints are predominantly adenosine, highly conserved, and closely distributed to the 3′ splice site. Analysis of human branchpoints reveals numerous novel features, including distinct features of branchpoints for alternatively spliced exons and a family of conserved sequence motifs overlapping branchpoints we term B-boxes, which exhibit maximal nucleotide diversity while maintaining interactions with the keto-rich U2 snRNA. Different B-box motifs exhibit divergent usage in vertebrate lineages and associate with other splicing elements and distinct intron–exon architectures, suggesting integration within a broader regulatory splicing code. Lastly, although branchpoints are refractory to common mutational processes and genetic variation, mutations occurring at branchpoint nucleotides are enriched for disease associations.

Exon circularization requires canonical splice signals

Circular RNAs (circRNAs), an abundant class of noncoding RNAs in higher eukaryotes, are generated from pre-mRNAs by circularization of adjacent exons. Using a set of 15 circRNAs, we demonstrated their cell-type-specific expression and circular versus linear processing in mammalian cells. Northern blot analysis combined with RNase H cleavage conclusively proved a circular configuration for two examples, LPAR1 and HIPK3. To address the circularization mechanism, we analyzed the sequence requirements using minigenes derived from natural circRNAs. Both canonical splice sites are required for circularization, although they vary in flexibility and potential use of cryptic sites. Surprisingly, we found that no specific circRNA exon sequence is necessary and that potential flanking intron structures can modulate circularization efficiency. In combination with splice inhibitor assays, our results argue that the canonical spliceosomal machinery functions in circRNA biogenesis, constituting an alternative splicing mode.

Single-nucleotide resolution mapping of the Gossypium raimondii transcriptome reveals a new mechanism for alternative splicing of introns

Alternative splicing (AS) is a vital genetic mechanism that enhances the diversity of eukaryotic transcriptomes. Here, we generated 8.3 Gb high-quality RNA-sequencing data from cotton (Gossypium raimondii) and performed a systematic, comparative analysis of AS events. We mapped 85% of the RNA-sequencing data onto the reference genome and identified 154368 splice junctions with 16437 as events in 10197 genes. Intron retention constituted the majority (40%) of all AS events in G. raimondii. Comparison across 11 eukaryote species showed that intron retention is the most common AS type in higher plants. Although transposable elements (TEs) were found in only 2.9% of all G. raimondii introns, they are present in 43% of the retained introns, suggesting that TE-insertion may be an important mechanism for intron retention during AS. The majority of the TE insertions are concentrated 0-40 nt upstream of the 3'-splice site, substantially altering the distribution of branch points from preferred positions and reducing the efficiency of intron splicing by decreasing RNA secondary structure flexibility. Our data suggest that TE-insertion-induced changes in branch point-site distribution are important for intron retention-type AS. Our findings may help explain the vast differences in intron-retention frequencies between vertebrates and higher plants.

Splicing code modeling

How do cis and trans elements involved in pre-mRNA splicing come together to form a splicing "code"? This question has been a driver of much of the research involving RNA biogenesis. The variability of splicing outcome across developmental stages and between tissues coupled with association of splicing defects with numerous diseases highlights the importance of such a code. However, the sheer number of elements involved in splicing regulation and the context-specific manner of their operation have made the derivation of such a code challenging. Recently, machine learning-based methods have been developed to infer computational models for a splicing code. These methods use high-throughput experiments measuring mRNA expression at exonic resolution and binding locations of RNA-binding proteins (RBPs) to infer what the regulatory elements that control the inclusion of a given pre-mRNA segment are. The inferred regulatory models can then be applied to genomic sequences or experimental conditions that have not been measured to predict splicing outcome. Moreover, the models themselves can be interrogated to identify new regulatory mechanisms, which can be subsequently tested experimentally. In this chapter, we survey the current state of this technology, and illustrate how it can be applied by non-computational or RNA splicing experts to study regulation of specific exons by using the AVISPA web tool.

In silico to in vivo splicing analysis using splicing code models

With the growing appreciation of RNA splicing's role in gene regulation, development, and disease, researchers from diverse fields find themselves investigating exons of interest. Commonly, researchers are interested in knowing if an exon is alternatively spliced, if it is differentially included in specific tissues or in developmental stages, and what regulatory elements control its inclusion. An important step towards the ability to perform such analysis in silico was made with the development of computational splicing code models. Aimed as a practical how-to guide, we demonstrate how researchers can now use these code models to analyze a gene of interest, focusing on Bin1 as a case study. Bridging integrator 1 (BIN1) is a nucleocytoplasmic adaptor protein known to be functionally regulated through alternative splicing in a tissue-specific manner. Specific Bin1 isoforms have been associated with muscular diseases and cancers, making the study of its splicing regulation of wide interest. Using AVISPA, a recently released web tool based on splicing code models, we show that many Bin1 tissue-dependent isoforms are correctly predicted, along with many of its known regulators. We review the best practices and constraints of using the tool, demonstrate how AVISPA is used to generate high confidence novel regulatory hypotheses, and experimentally validate predicted regulators of Bin1 alternative splicing.

Intronic sequence elements impede exon ligation and trigger a discard pathway that yields functional telomerase RNA in fission yeast

The fission yeast telomerase RNA (TER1) precursor harbors an intron immediately downstream from its mature 3' end. Unlike most introns, which are removed from precursor RNAs by the spliceosome in two sequential but tightly coupled transesterification reactions, TER1 only undergoes the first cleavage reaction during telomerase RNA maturation. The mechanism underlying spliceosome-mediated 3' end processing has remained unclear. We now demonstrate that a strong branch site (BS), a long distance to the 3' splice site (3' SS), and a weak polypyrimidine (Py) tract act synergistically to attenuate the transition from the first to the second step of splicing. The observation that a strong BS antagonizes the second step of splicing in the context of TER1 suggests that the BS-U2 snRNA interaction is disrupted after the first step and thus much earlier than previously thought. The slow transition from first to second step triggers the Prp22 DExD/H-box helicase-dependent rejection of the cleaved products and Prp43-dependent "discard" of the splicing intermediates. Our findings explain how the spliceosome can function in 3' end processing and provide new insights into the mechanism of splicing.

Impact of base pair identity 5' to the spliceosomal branch site adenosine on branch site conformation

The branch site helix from Saccharomyces cerevisiae with pseudouridine (ψ) incorporated in a phylogenetically conserved position of U2 snRNA features an extrahelical branch site adenosine (A) that forms a base triple interaction with the minor groove edge of a widely conserved purine(U2 strand)-pyrimidine(intron strand) (R(U2)-Y(intron)) base pair two positions upstream. In these studies, NMR spectra of a duplex in which 2-aminopurine (2ap), a fluorescent analog of adenine lacking the proposed hydrogen bond donor, was substituted for the branch site A, indicated that the substitution does not alter the extrahelical position of the branch site residue; thus, it appears that a hydrogen bond between the adenine amino group and the R-Y pair is not obligatory for stabilization of the extrahelical conformation. In contrast, reversal of the orientation of A(U2)-U(intron) to U(U2)-A(intron) resulted in an intrahelical position for the branch site A or 2ap. Fluorescence intensity of 2ap substituted for the branch site A with the original R(U2)-Y(intron) orientation (AU or GC) was high, consistent with an extrahelical position, whereas fluorescence in helices with the reversed R-Y orientation, or with a mismatched pair (A-U → G•A or U•C), was markedly quenched, implying that the residue was stacked in the helix. The A 5' to the branch site residue was not extrahelical in any of the duplexes. These findings suggest that the R(U2)-Y(intron) base pair orientation in the ψ-dependent branch site helix plays an important role in positioning the branch site A for recognition and/or function.

The RNA-binding landscapes of two SR proteins reveal unique functions and binding to diverse RNA classes

Background

The SR proteins comprise a family of essential, structurally related RNA binding proteins. The complexity of their RNA targets and specificity of RNA recognition in vivo is not well understood. Here we use iCLIP to globally analyze and compare the RNA binding properties of two SR proteins, SRSF3 and SRSF4, in murine cells.

Results

SRSF3 and SRSF4 binding sites mapped to largely non-overlapping target genes, and in vivo consensus binding motifs were distinct. Interactions with intronless and intron-containing mRNAs as well as non-coding RNAs were detected. Surprisingly, both SR proteins bound to the 3' ends of the majority of intronless histone transcripts, implicating SRSF3 and SRSF4 in histone mRNA metabolism. In contrast, SRSF3 but not SRSF4 specifically bound transcripts encoding numerous RNA binding proteins. Remarkably, SRSF3 was shown to modulate alternative splicing of its own as well as three other transcripts encoding SR proteins. These SRSF3-mediated splicing events led to downregulation of heterologous SR proteins via nonsense-mediated decay.

Conclusions

SRSF3 and SRSF4 display unique RNA binding properties underlying diverse cellular regulatory mechanisms, with shared as well as unique coding and non-coding targets. Importantly, CLIP analysis led to the discovery that SRSF3 cross-regulates the expression of other SR protein family members.