Exploring the functional impact of alternative splicing on human protein isoforms using available annotation sources

Alternative isoform expression of key thermogenic genes in human beige adipocytes

Background

The beneficial effect of thermogenic adipocytes in maintaining body weight and protecting against metabolic disorders has raised interest in understanding the regulatory mechanisms defining white and beige adipocyte identity. Although alternative splicing has been shown to propagate adipose browning signals in mice, this has yet to be thoroughly investigated in human adipocytes.

Methods

We performed parallel white and beige adipogenic differentiation using primary adipose stem cells from 6 unrelated healthy subjects and assessed differential gene and isoform expression in mature adipocytes by RNA sequencing.

Results

We find 777 exon junctions with robust differential usage between white and beige adipocytes in all 6 subjects, mapping to 562 genes. Importantly, only 10% of these differentially spliced genes are also differentially expressed, indicating that alternative splicing constitutes an additional layer of gene expression regulation during beige adipocyte differentiation. Functional classification of alternative isoforms points to a gain of function for key thermogenic transcription factors such as PPARG and CITED1, and enzymes such as PEMT, or LPIN1. We find that a large majority of the splice variants arise from differential TSS usage, with beige-specific TSSs being enriched for PPARγ and MED1 binding compared to white-specific TSSs. Finally, we validate beige specific isoform expression at the protein level for two thermogenic regulators, PPARγ and PEMT.

Discussion

These results suggest that differential isoform expression through alternative TSS usage is an important regulatory mechanism for human adipocyte thermogenic specification.

Mutant U2AF1-Induced Mis-Splicing of mRNA Translation Genes Confers Resistance to Chemotherapy in Acute Myeloid Leukemia

Patients with primary refractory acute myeloid leukemia (AML) have a dismal long-term prognosis. Elucidating the resistance mechanisms to induction chemotherapy could help identify strategies to improve AML patient outcomes. Herein, we retrospectively analyzed the multiomics data of more than 1,500 AML cases and found that patients with spliceosome mutations had a higher risk of developing refractory disease. RNA splicing analysis revealed that the mis-spliced genes in refractory patients converged on translation-associated pathways, promoted mainly by U2AF1 mutations. Integrative analyses of binding and splicing in AML cell lines substantiated that the splicing perturbations of mRNA translation genes originated from both the loss and gain of mutant U2AF1 binding. In particular, the U2AF1S34F and U2AF1Q157R mutants orchestrated the inclusion of exon 11 (encoding a premature termination codon) in the eukaryotic translation initiation factor 4A2 (EIF4A2). This aberrant inclusion led to reduced eIF4A2 protein expression via nonsense-mediated mRNA decay. Consequently, U2AF1 mutations caused a net decrease in global mRNA translation that induced the integrated stress response (ISR) in AML cells, which was confirmed by single-cell RNA sequencing. The induction of ISR enhanced the ability of AML cells to respond and adapt to stress, contributing to chemoresistance. A pharmacologic inhibitor of ISR, ISRIB, sensitized U2AF1 mutant cells to chemotherapy. These findings highlight a resistance mechanism by which U2AF1 mutations drive chemoresistance and provide a therapeutic approach for AML through targeting the ISR pathway.

SIGNIFICANCE

U2AF1 mutations induce the integrated stress response by disrupting splicing of mRNA translation genes that improves AML cell fitness to enable resistance to chemotherapy, which can be targeted to improve AML treatment.

Alternative splicing in prostate cancer progression and therapeutic resistance

Prostate cancer (CaP) remains the second leading cause of cancer deaths in western men. CaP mortality results from diverse molecular mechanisms that mediate resistance to the standard of care treatments for metastatic disease. Recently, alternative splicing has been recognized as a hallmark of CaP aggressiveness. Alternative splicing events cause treatment resistance and aggressive CaP behavior and are determinants of the emergence of the two major types of late-stage treatment-resistant CaP, namely castration-resistant CaP (CRPC) and neuroendocrine CaP (NEPC). Here, we review recent multi-omics data that are uncovering the complicated landscape of alternative splicing events during CaP progression and the impact that different gene transcript isoforms can have on CaP cell biology and behavior. We discuss renewed insights in the molecular machinery by which alternative splicing occurs and contributes to the failure of systemic CaP therapies. The potential for alternative splicing events to serve as diagnostic markers and/or therapeutic targets is explored. We conclude by considering current challenges and promises associated with splicing-modulating therapies, and their potential for clinical translation into CaP patient care.

SpliceProt 2.0: A Sequence Repository of Human, Mouse, and Rat Proteoforms

SpliceProt 2.0 is a public proteogenomics database that aims to list the sequence of known proteins and potential new proteoforms in human, mouse, and rat proteomes. This updated repository provides an even broader range of computationally translated proteins and serves, for example, to aid with proteomic validation of splice variants absent from the reference UniProtKB/SwissProt database. We demonstrate the value of SpliceProt 2.0 to predict orthologous proteins between humans and murines based on transcript reconstruction, sequence annotation and detection at the transcriptome and proteome levels. In this release, the annotation data used in the reconstruction of transcripts based on the methodology of ternary matrices were acquired from new databases such as Ensembl, UniProt, and APPRIS. Another innovation implemented in the pipeline is the exclusion of transcripts predicted to be susceptible to degradation through the NMD pathway. Taken together, our repository and its applications represent a valuable resource for the proteogenomics community.

A cis-regulatory element regulates ERAP2 expression through autoimmune disease risk SNPs

Single-nucleotide polymorphisms (SNPs) near the ERAP2 gene are associated with various autoimmune conditions, as well as protection against lethal infections. Due to high linkage disequilibrium, numerous trait-associated SNPs are correlated with ERAP2 expression; however, their functional mechanisms remain unidentified. We show by reciprocal allelic replacement that ERAP2 expression is directly controlled by the splice region variant rs2248374. However, disease-associated variants in the downstream LNPEP gene promoter are independently associated with ERAP2 expression. Allele-specific conformation capture assays revealed long-range chromatin contacts between the gene promoters of LNPEP and ERAP2 and showed that interactions were stronger in patients carrying the alleles that increase susceptibility to autoimmune diseases. Replacing the SNPs in the LNPEP promoter by reference sequences lowered ERAP2 expression. These findings show that multiple SNPs act in concert to regulate ERAP2 expression and that disease-associated variants can convert a gene promoter region into a potent enhancer of a distal gene.

Efficient Detection of the Alternative Spliced Human Proteome Using Translatome Sequencing

Alternative splicing (AS) isoforms create numerous proteoforms, expanding the complexity of the genome. Highly similar sequences, incomplete reference databases and the insufficient sequence coverage of mass spectrometry limit the identification of AS proteoforms. Here, we demonstrated full-length translating mRNAs (ribosome nascent-chain complex-bound mRNAs, RNC-mRNAs) sequencing (RNC-seq) strategy to sequence the entire translating mRNA using next-generation sequencing, including short-read and long-read technologies, to construct a protein database containing all translating AS isoforms. Taking the advantage of read length, short-read RNC-seq identified up to 15,289 genes and 15,906 AS isoforms in a single human cell line, much more than the Ribo-seq. The single-molecule long-read RNC-seq supplemented 4,429 annotated AS isoforms that were not identified by short-read datasets, and 4,525 novel AS isoforms that were not included in the public databases. Using such RNC-seq-guided database, we identified 6,766 annotated protein isoforms and 50 novel protein isoforms in mass spectrometry datasets. These results demonstrated the potential of full-length RNC-seq in investigating the proteome of AS isoforms.

Identifying genes with conserved splicing structure and orthologous isoforms in human, mouse and dog

BACKGROUND

In eukaryote transcriptomes, a significant amount of transcript diversity comes from genes' capacity to generate different transcripts through alternative splicing. Identifying orthologous alternative transcripts across multiple species is of particular interest for genome annotators. However, there is no formal definition of transcript orthology based on the splicing structure conservation. Likewise there is no public dataset benchmark providing groups of orthologous transcripts sharing a conserved splicing structure.

RESULTS

We introduced a formal definition of splicing structure orthology and we predicted transcript orthologs in human, mouse and dog. Applying a selective strategy, we analyzed 2,167 genes and their 18,109 known transcripts and identified a set of 253 gene orthologs that shared a conserved splicing structure in all three species. We predicted 6,861 transcript CDSs (coding sequence), mainly for dog, an emergent model species. Each predicted transcript was an ortholog of a known transcript: both share the same CDS splicing structure. Evidence for the existence of the predicted CDSs was found in external data.

CONCLUSIONS

We generated a dataset of 253 gene triplets, structurally conserved and sharing all their CDSs in human, mouse and dog, which correspond to 879 triplets of spliced CDS orthologs. We have released the dataset both as an SQL database and as tabulated files. The data consists of the 879 CDS orthology groups with their detailed splicing structures, and the predicted CDSs, associated with their experimental evidence. The 6,861 predicted CDSs are provided in GTF files. Our data may contribute to compare highly conserved genes across three species, for comparative transcriptomics at the isoform level, or for benchmarking splice aligners and methods focusing on the identification of splicing orthologs. The data is available at https://data-access.cesgo.org/index.php/s/V97GXxOS66NqTkZ .

Uncovering the impacts of alternative splicing on the proteome with current omics techniques

The high‐throughput sequencing of cellular RNAs has underscored a broad effect of isoform diversification through alternative splicing on the transcriptome. Moreover, the differential production of transcript isoforms from gene loci has been recognized as a critical mechanism in cell differentiation, organismal development, and disease. Yet, the extent of the impact of alternative splicing on protein production and cellular function remains a matter of debate. Multiple experimental and computational approaches have been developed in recent years to address this question. These studies have unveiled how molecular changes at different steps in the RNA processing pathway can lead to differences in protein production and have functional effects. New and emerging experimental technologies open exciting new opportunities to develop new methods to fully establish the connection between messenger RNA expression and protein production and to further investigate how RNA variation impacts the proteome and cell function.

This article is categorized under:

RNA Processing > Splicing Regulation/Alternative Splicing

Translation > Regulation

RNA Evolution and Genomics > Computational Analyses of RNA

Integrated Proteotranscriptomics of Human Myometrium in Labor Landscape Reveals the Increased Molecular Associated With Inflammation Under Hypoxia Stress

During labor, a variety of coordinated physiological and biochemical events cause the myometrium to transition from a quiescent to contractile state; the molecular mechanisms responsible for this transition, however, remain unclear. To better understand this transition at a molecular level, the global transcriptome and proteome of human myometrial samples in labor and those not in labor were investigated through RNA sequencing (RNA-seq) and quantitative liquid chromatography–tandem mass spectrometry (LC-MS/MS) via data-independent acquisition (DIA) and parallel reaction monitoring (PRM) methods. Furthermore, an integrated proteotranscriptomic analysis was performed to explore biological processes and pathway alterations during labor; this analysis identified 1,626 differentially expressed mRNAs (1,101 upregulated, 525 downregulated) and 135 differentially expressed proteins (97 upregulated, 38 downregulated) in myometrium between nonlabor and in labor groups. The comprehensive results of these analyses showed that the upregulated mRNAs and proteins increased inflammation under hypoxia stress in the myometrium under labor, and related proteins and cytokines were validated by PRM and Luminex assays. Our study confirmed the biological process of inflammation and hypoxia in laboring myometrium at the transcriptome and proteome levels and provided recourse to discover new molecular and biological changes during labor.

Structural basis of the selective activation of enzyme isoforms: Allosteric response to activators of β1- and β2-containing AMPK complexes

AMP-activated protein kinase (AMPK) is a key energy sensor regulating the cell metabolism in response to energy supply and demand. The evolutionary adaptation of AMPK to different tissues is accomplished through the expression of distinct isoforms that can form up to 12 complexes, which exhibit notable differences in the sensitivity to allosteric activators. To shed light into the molecular determinants of the allosteric regulation of this energy sensor, we have examined the structural and dynamical properties of β1- and β2-containing AMPK complexes formed with small molecule activators A-769662 and SC4, and dissected the mechanical response leading to active-like enzyme conformations through the analysis of interaction networks between structural domains. The results reveal the mechanical sensitivity of the α2β1 complex, in contrast with a larger resilience of the α2β2 species, especially regarding modulation by A-769662. Furthermore, binding of activators to α2β1 consistently promotes the pre-organization of the ATP-binding site, favoring the adoption of activated states of the enzyme. These findings are discussed in light of the changes in the residue content of β-subunit isoforms, particularly regarding the β1Asn111 → β2Asp111 substitution as a key factor in modulating the mechanical sensitivity of β1- and β2-containing AMPK complexes. Our studies pave the way for the design of activators tailored for improving the therapeutic treatment of tissue-specific metabolic disorders.

Identification and characterization of novel splice variants of human farnesoid X receptor

Farnesoid X receptor (FXR, NR1H4) is a ligand-activated nuclear receptor, which regulates bile acid, lipid and glucose metabolism. Due to these functions, FXR has been investigated as a potential drug target for the treatment of liver diseases, such as primary biliary cholangitis and non-alcoholic steatohepatitis. Based on the previously described four splice variants, it has been suggested that alternative promoter usage and splicing may have an impact on total FXR activity as a result of encoding functionally diverse variants. Here we aimed for a systematic analysis of human hepatic FXR splice variants. In addition to the previously described FXRα1-4, we identified four novel splice variants (FXRα5-8) in human hepatocytes, which resulted from previously undetected exon skipping events. These newly identified isoforms displayed diminished DNA binding and impaired transactivation activities. Isoform FXRα5, which suppressed the transactivation activity of the functional isoform FXRα2, was further characterized as deficient in heterodimerization, coactivator recruitment and ligand binding. These findings were further supported by molecular dynamics simulations, which offered an explanation for the behavior of this isoform on the molecular level. FXRα5 exhibited low uniform expression levels in nearly all human tissues. Our systematic analysis of FXR splice variants in human hepatocytes resulted in the identification of four novel FXR isoforms, which all proved to be functionally deficient, but one novel variant, FXRα5, also displayed dominant negative activity. The possible associations with and roles of these novel isoforms in human liver diseases require further investigation.

TNMplot.com: A Web Tool for the Comparison of Gene Expression in Normal, Tumor and Metastatic Tissues

Genes showing higher expression in either tumor or metastatic tissues can help in better understanding tumor formation and can serve as biomarkers of progression or as potential therapy targets. Our goal was to establish an integrated database using available transcriptome-level datasets and to create a web platform which enables the mining of this database by comparing normal, tumor and metastatic data across all genes in real time. We utilized data generated by either gene arrays from the Gene Expression Omnibus of the National Center for Biotechnology Information (NCBI-GEO) or RNA-seq from The Cancer Genome Atlas (TCGA), Therapeutically Applicable Research to Generate Effective Treatments (TARGET), and The Genotype-Tissue Expression (GTEx) repositories. The altered expression within different platforms was analyzed separately. Statistical significance was computed using Mann–Whitney or Kruskal–Wallis tests. False Discovery Rate (FDR) was computed using the Benjamini–Hochberg method. The entire database contains 56,938 samples, including 33,520 samples from 3180 gene chip-based studies (453 metastatic, 29,376 tumorous and 3691 normal samples), 11,010 samples from TCGA (394 metastatic, 9886 tumorous and 730 normal), 1193 samples from TARGET (1 metastatic, 1180 tumorous and 12 normal) and 11,215 normal samples from GTEx. The most consistently upregulated genes across multiple tumor types were TOP2A (FC = 7.8), SPP1 (FC = 7.0) and CENPA (FC = 6.03), and the most consistently downregulated gene was ADH1B (FC = 0.15). Validation of differential expression using equally sized training and test sets confirmed the reliability of the database in breast, colon, and lung cancer at an FDR below 10%. The online analysis platform enables unrestricted mining of the database and is accessible at TNMplot.com.

Investigating the complex interplay between genotype and high-fat-diet feeding in the lactating mammary gland using the Tph1 and Ldlr knockout models

Obesity is a prevailing problem across the globe. Women who are obese have difficulty initiating and sustaining lactation. However, the impact of genetics and diet on breastfeeding outcomes is understudied. Here we explore the effect of diet and genotype on lactation. We utilized the low-density lipoprotein receptor (Ldlr-KO) transgenic mouse model as an obesity and hypercholesterolemia model. Additionally, we used the tryptophan hydroxylase 1 (Tph1-KO) mouse, recently identified as a potential anti-obesogenic model, to investigate if addition of Tph1-KO could ameliorate negative effects of obesity in Ldlr-KO mice. We created a novel transgenic mouse line by combining the Ldlr and Tph1 [double knockout (DKO)] mice to study the interaction between the two genotypes. Female mice were fed a low-fat diet (LFD; 10% fat) or high-fat diet (HFD; 60% fat) from 3 wk of age through early [lactation day 3 (L3)] or peak lactation [lactation day 11 (L11)]. After 4 wk of consuming either LFD or HFD, female mice were bred. On L2 and L10, dams were milked to investigate the effect of diet and genotype on milk composition. Dams were euthanized on L3 or L11. There was no impact of diet or genotype on milk protein or triglycerides (TGs) on L2; however, by L10, Ldlr-KO and DKO dams had increased TG levels in milk. RNA-sequencing of L11 mammary glands demonstrated Ldlr-KO dams fed HFD displayed enrichment of genes involved in immune system pathways. Interestingly, the DKO may alter vesicle budding and biogenesis during lactation. We also quantified macrophages by immunostaining for F4/80+ cells at L3 and L11. Diet played a significant role on L3 (P = 0.013), but genotype played a role at L11 (P < 0.0001) on numbers of F4/80+ cells. Thus the impact of diet and genotype on lactation differs depending on stage of lactation, illustrating complexities of understanding the intersection of these parameters.NEW & NOTEWORTHY We have created a novel mouse model that is focused on understanding the intersection of diet and genotype on mammary gland function during lactation.

DIGGER: exploring the functional role of alternative splicing in protein interactions

Alternative splicing plays a major role in regulating the functional repertoire of the proteome. However, isoform-specific effects to protein-protein interactions (PPIs) are usually overlooked, making it impossible to judge the functional role of individual exons on a systems biology level. We overcome this barrier by integrating protein-protein interactions, domain-domain interactions and residue-level interactions information to lift exon expression analysis to a network level. Our user-friendly database DIGGER is available at https://exbio.wzw.tum.de/digger and allows users to seamlessly switch between isoform and exon-centric views of the interactome and to extract sub-networks of relevant isoforms, making it an essential resource for studying mechanistic consequences of alternative splicing.

tappAS: a comprehensive computational framework for the analysis of the functional impact of differential splicing

Recent advances in long-read sequencing solve inaccuracies in alternative transcript identification of full-length transcripts in short-read RNA-Seq data, which encourages the development of methods for isoform-centered functional analysis. Here, we present tappAS, the first framework to enable a comprehensive Functional Iso-Transcriptomics (FIT) analysis, which is effective at revealing the functional impact of context-specific post-transcriptional regulation. tappAS uses isoform-resolved annotation of coding and non-coding functional domains, motifs, and sites, in combination with novel analysis methods to interrogate different aspects of the functional readout of transcript variants and isoform regulation. tappAS software and documentation are available at https://app.tappas.org.

DIFFUSE: predicting isoform functions from sequences and expression profiles via deep learning

MOTIVATION

Alternative splicing generates multiple isoforms from a single gene, greatly increasing the functional diversity of a genome. Although gene functions have been well studied, little is known about the specific functions of isoforms, making accurate prediction of isoform functions highly desirable. However, the existing approaches to predicting isoform functions are far from satisfactory due to at least two reasons: (i) unlike genes, isoform-level functional annotations are scarce. (ii) The information of isoform functions is concealed in various types of data including isoform sequences, co-expression relationship among isoforms, etc.

RESULTS

In this study, we present a novel approach, DIFFUSE (Deep learning-based prediction of IsoForm FUnctions from Sequences and Expression), to predict isoform functions. To integrate various types of data, our approach adopts a hybrid framework by first using a deep neural network (DNN) to predict the functions of isoforms from their genomic sequences and then refining the prediction using a conditional random field (CRF) based on co-expression relationship. To overcome the lack of isoform-level ground truth labels, we further propose an iterative semi-supervised learning algorithm to train both the DNN and CRF together. Our extensive computational experiments demonstrate that DIFFUSE could effectively predict the functions of isoforms and genes. It achieves an average area under the receiver operating characteristics curve of 0.840 and area under the precision-recall curve of 0.581 over 4184 GO functional categories, which are significantly higher than the state-of-the-art methods. We further validate the prediction results by analyzing the correlation between functional similarity, sequence similarity, expression similarity and structural similarity, as well as the consistency between the predicted functions and some well-studied functional features of isoform sequences.

AVAILABILITY AND IMPLEMENTATION

https://github.com/haochenucr/DIFFUSE.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.