A compendium of RNA-binding motifs for decoding gene regulation

Landscape of chimeric RNAs in COVID-19 patient blood

Despite the availability of efficacious vaccines, COVID-19 persists and our knowledge of how SARS-CoV-2 infection affects host transcriptomics remains incomplete. Transcriptome analysis, which has progressed our understanding of the patient response to SARS-CoV-2 infection, can be enhanced by considering chimeric transcript expression. Here we assess and characterize chimeric RNAs found in the whole blood of 178 COVID-19 patients. STAR-Fusion, SOAPfuse, and EricScript were used to detect chimeric RNAs resulting in over 30,000 predictions with approximately 500 high-confidence predictions that were found by more than one software and filtered based on exon annotations around the chimeric splice junction. GO term enrichment performed on the 5' and 3' parental genes of chimeric RNAs found in severe and critical patients resulted in pathways known to be affected in these patients, such as erythroid differentiation. Motif enrichment analysis of sequences proximal to chimeric splice junctions found in COVID-19 patients versus those found in GTEx whole blood revealed two RNA binding proteins previously implicated with coronavirus infection, PTBP1 and SFPQ. We discovered a chimeric RNA that correlated with COVID-19 disease status and appeared to be dependent upon a loss of PTBP1's function as a splicing repressor. Overall, we found over 350 novel COVID-19-specific chimeric RNAs not detectable in GTEx whole blood that may also serve as biomarkers for viral infection.

Structural basis for the RNA binding properties of mouse IGF2BP3

IGF2BP family proteins (IGF2BPs) contain six tandem RNA-binding domains (RBDs), resulting in highly complex RNA binding properties. Dissecting how IGF2BPs recognize their RNA targets is essential for understanding their regulatory roles in gene expression. Here, we have determined the crystal structures of mouse IGF2BP3 constructs complexed with different RNA substrates. Our structures reveal that the IGF2BP3-RRM12 domains can recognize CA-rich elements up to 5-nt in length, mainly through RRM1. We also captured the antiparallel RNA-binding mode of the IGF2BP3-KH12 domains, with five nucleotides bound by KH1 and two nucleotides bound by KH2. Furthermore, our structural and biochemical studies suggest that the IGF2BP3-KH12 domains could recognize the "zipcode" RNA element within the β-actin mRNA. Finally, we analyzed the similarities and differences of the RNA-binding properties between the KH12 and KH34. Our studies provide structural insights into RNA target recognition by mouse IGF2BP3.

Pumilio differentially binds to mRNA 3' UTR isoforms to regulate localization of synaptic proteins

In neuronal cells, the regulation of RNA is crucial for the spatiotemporal control of gene expression, but how the correct localization, levels, and function of synaptic proteins are achieved is not well understood. In this study, we globally investigate the role of alternative 3' UTRs in regulating RNA localization in the synaptic regions of the Drosophila brain. We identify direct mRNA targets of the translational repressor Pumilio, finding that mRNAs bound by Pumilio encode proteins enriched in synaptosomes. Pumilio differentially binds to RNA isoforms of the same gene, favoring long, neuronal 3' UTRs. These longer 3' UTRs tend to remain in the neuronal soma, whereas shorter UTR isoforms localize to the synapse. In cultured pumilio mutant neurons, axon outgrowth defects are accompanied by mRNA isoform mislocalization, and proteins encoded by these Pumilio target mRNAs display excessive abundance at synaptic boutons. Our study identifies an important mechanism for the spatiotemporal regulation of protein function in neurons.

Transcriptome-wide alternative mRNA splicing analysis reveals post-transcriptional regulation of neuronal differentiation

Alternative splicing (AS) plays an important role in neuronal development, function, and disease. Efforts to analyze the transcriptome of AS in neurons on a wide scale are currently limited. We characterized the transcriptome-wide AS changes in SH-SY5Y neuronal differentiation model, which is widely used to study neuronal function and disorders. Our analysis revealed global changes in five AS programs that drive neuronal differentiation. Motif analysis revealed the contribution of RNA-binding proteins (RBPs) to the regulation of AS during neuronal development. We concentrated on the primary alternative splicing program that occurs during differentiation, specifically on events involving exon skipping (SE). Motif analysis revealed motifs for polypyrimidine tract-binding protein 1 (PTB) and ELAV-like RNA binding protein 1 (HuR/ELAVL1) to be the top enriched in SE events, and their protein levels were downregulated after differentiation. shRNA knockdown of either PTB and HuR was associated with enhanced neuronal differentiation and transcriptome-wide exon skipping events that drive the process of differentiation. At the level of gene expression, we observed only modest changes, indicating predominant post-transcriptional effects of PTB and HuR. We also observed that both RBPs altered cellular responses to oxidative stress, in line with the differentiated phenotype observed after either gene knockdown. Our work characterizes the AS changes in a widely used and important model of neuronal development and neuroscience research and reveals intricate post-transcriptional regulation of neuronal differentiation.

Noncanonical circRNA biogenesis driven by alpha and gamma herpesviruses

Herpesviruses require the host transcriptional machinery, inducing significant changes in gene expression to prioritize viral transcripts. We examined alpha- and gamma-herpesvirus alterations to a type of alternative splicing, namely circular RNA (circRNA) synthesis. We developed "Circrnas in Host And viRuses anaLysis pIpEline" (CHARLIE) to facilitate viral profiling. This method identified thousands of back-splicing variants, including circRNA common to lytic and latent phases of infection. Ours is the first report of Herpes Simplex Virus-1 circRNAs, including species derived from ICP0 and the latency-associated transcript. We characterized back-splicing cis- and trans-elements, and found viral circRNAs resistant to spliceosome perturbation and lacking canonical splice donor-acceptors. Subsequent loss-of-function studies of host RNA ligases (RTCB, RLIG1) revealed instances of decreased viral back splicing. Using eCLIP and 4sU-Sequencing, we determined that the KSHV RNA-binding protein, ORF57, enhanced synthesis for a subset of viral and host circRNAs. Our work explores unique splicing mechanisms driven by lytic infection, and identifies a class of transcripts with the potential to function in replication, persistence, or tumorigenesis.

Identification of RNA processing factor LSM5 as a new adverse biomarker in nasopharyngeal carcinoma

Dysregulated RNA processing is crucial in nasopharyngeal carcinoma (NPC) progression. Our research aimed to evaluate the prognostic values of RNA-processing genes (RPGs) in NPC through bioinformatic analysis of the GSE12452 and GSE102349 datasets, identifying differentially expressed RNA-processing genes (DE-RPGs). A prognostic model was developed using univariate and multivariate Cox analysis, with effectiveness assessed by ROC analysis. The correlation between risk scores, immune characteristics, and chemotherapy sensitivity was also analyzed. Model gene expression was validated by RT-qPCR, Western Blot, and immunohistochemistry, alongside functional assays. Bioinformatics indicated that RNA binding motif protein 20 (RBM20) and LSM5 are prognostic RPGs, with the ROC curve confirming their predictive ability for survival. Significant differences in drug sensitivity were noted between high- and low-risk groups. Experimental validation showed LSM5 is overexpressed in NPC tissues, correlating with poorer prognosis, and its down-regulation inhibits cell proliferation and migration. Thus, LSM5 is identified as a new adverse biomarker in NPC, with implications for targeted therapy and prognosis improvement.

Dissecting RNA selectivity mediated by tandem RNA-binding domains

RNA-protein interactions are pivotal to proper gene regulation. Many RNA-binding proteins possess multiple RNA-binding domains; however, how these domains interplay to select and regulate RNA targets remains poorly understood. Here, we investigate three multi-domain proteins, Musashi-1, Musashi-2, and unkempt, which share a high degree of RNA specificity, a common feature across RNA-binding proteins. We used massively parallel in vitro assays with unprecedented depth with random or naturally-derived RNA sequences and find that individual domains within a protein can have differing affinities, specificities, and motif spacing preferences. We conducted large scale competition assays between these proteins and determined how individual protein specificities and affinities influence competitive binding. Integration of binding and regulation in cells with in vitro specificities showed that target selection involves a combination of the protein intrinsic specificities described here, but cellular context is critical to drive these proteins to motifs in specific transcript regions. Finally, evolutionarily conserved RNA regions displayed evidence of binding multiple RBPs in cultured cells, and these RNA regions represent the highest affinity targets. This work emphasizes the importance of in vitro and in cultured cells studies to fully profile RNA-binding proteins and highlights the complex modes of RNA-protein interactions and the contributing factors in target selection.

Immiscible proteins compete for RNA binding to order condensate layers

Biomolecular condensates mediate diverse and essential cellular functions by compartmentalizing biochemical pathways. Many condensates have internal subdomains with distinct compositional identities. A major challenge lies in dissecting the multicomponent logic that relates biomolecular features to emergent condensate organization. Nuclear paraspeckles are paradigmatic examples of multi-domain condensates, comprising core and shell layers with distinct compositions that are scaffolded by the lncRNA NEAT1, which spans both layers. A prevailing model of paraspeckle assembly proposes that core proteins bind directly and specifically to core-associated NEAT1 domains. Combining informatics and biochemistry, we unexpectedly find that the essential core proteins FUS and NONO bind and condense preferentially with shell-associated NEAT1 domains. The shell protein TDP-43 exhibits similar NEAT1 domain preferences on its own but forms surfactant-like shell layers around core protein-driven condensates when both are present. Together, experiments and physics-based simulations suggest that competitive RNA binding and immiscibility between core and shell proteins orders paraspeckle layers. More generally, we propose that sub-condensate organization can spontaneously arise from a balance of collaborative and competitive protein binding to the same domains of a lncRNA.

Significance

The cellular milieu is spatially organized into compartments called biomolecular condensates that exhibit rich internal organization which shapes their functions. Often comprising multiple proteins and RNAs, a major question concerns how molecular-scale features relate to emergent condensate forms. Here we study nuclear paraspeckles, archetypal multi-domain condensates comprising distinct core and shell layers assembled around a layer-spanning RNA scaffold. We find that different proteins associated with each layer all bind preferentially to the same shell-associated domains of the RNA scaffold. Core and shell proteins are inherently immiscible, establishing a competition that redirects core proteins to a suboptimal, core-associated domain of the scaffold. Our work reveals how a combination of competitive RNA binding and protein immiscibility can spatially organize multicomponent condensate subdomains.

RBPsuite 2.0: an updated RNA-protein binding site prediction suite with high coverage on species and proteins based on deep learning

BACKGROUND

RNA-binding proteins (RBPs) play crucial roles in many biological processes, and computationally identifying RNA-RBP interactions provides insights into the biological mechanism of diseases associated with RBPs.

RESULTS

To make the RBP-specific deep learning-based RBP binding sites prediction methods easily accessible, we developed an updated easy-to-use webserver, RBPsuite 2.0, with an updated web interface for predicting RBP binding sites from linear and circular RNA sequences. RBPsuite 2.0 has a higher coverage on the number of supported RBPs and species compared to the original RBPsuite, supporting an increased number of RBPs from 154 to 353 and expanding the supported species from one to seven. Additionally, RBPsuite 2.0 replaces the CRIP built into RBPsuite 1.0 with iDeepC, a more accurate RBP binding site predictor for circular RNAs. Furthermore, RBPsuite 2.0 estimates the contribution score of individual nucleotides on the input sequences as potential binding motifs and links to the UCSC browser track for better visualization of the prediction results.

CONCLUSIONS

RBPsuite 2.0 is an updated, more comprehensive webserver for predicting RBP binding sites in both linear and circular RNA sequences. It supports more RBPs and species and provides more accurate predictions for circular RNAs. The tool is freely available at http://www.csbio.sjtu.edu.cn/bioinf/RBPsuite/ .

Species-specific circular RNA circDS-1 enhances adaptive evolution in Talaromyces marneffei through regulation of dimorphic transition

Thermal adaptability is a crucial characteristic for mammalian pathogenic fungi that originally inhabit natural ecosystems. Thermally dimorphic fungi have evolved a unique ability to respond to host body temperature by shifting from mycelia to yeast. The high similarity of protein-coding genes between these fungi and their relatives suggests the indispensable but often overlooked roles of non-coding elements in fungal thermal adaptation. Here, we systematically delineated the landscape of full-length circRNAs in both mycelial and yeast conditions of Talaromyces marneffei, a typical thermally dimorphic fungus causing fatal Talaromycosis, by optimizing an integrative pipeline for circRNA detection utilizing next- and third-generation sequencing. We found T. marneffei circRNA demonstrated features such as shorter length, lower abundance, and circularization-biased splicing. We then identified and validated that circDS-1, independent of its parental gene, promotes the hyphae-to-yeast transition, maintains yeast morphology, and is involved in virulence regulation. Further analysis and experiments among Talaromyces confirmed that the generation of circDS-1 is driven by a T. marneffei-specific region in the flanking intron of circDS-1. Together, our findings not only provide fresh insights into the role of circRNA in fungal thermal adaptation but also reveal a novel molecular mechanism for the adaptive evolution of functional circRNAs derived from intronic mutations.

RNA gain-of-function mechanisms in short tandem repeat diseases

As adaptors, catalysts, guides, messengers, scaffolds, and structural components, RNAs perform an impressive array of cellular regulatory functions often by recruiting RNA-binding proteins (RBPs) to form ribonucleoprotein complexes (RNPs). While this RNA-RBP interaction network allows precise RNP assembly and the subsequent structural dynamics required for normal functions, RNA motif mutations may trigger the formation of aberrant RNP structures that lead to cell dysfunction and disease. Here, we provide our perspective on one type of RNA motif mutation, RNA gain-of-function mutations associated with the abnormal expansion of short tandem repeats (STRs) that underlie multiple developmental and degenerative diseases. We first discuss our current understanding of normal polymorphic STR functions in RNA processing and localization followed by an assessment of the pathogenic roles of STR expansions in the neuromuscular disease myotonic dystrophy. We also highlight ongoing questions and controversies focused on STR-based insights into the regulation of nuclear RNA processing and export as well as the relevance of the RNA gain-of-function pathomechanism for other STR expansion disorders in both coding and noncoding genes.

Antisense oligonucleotide-mediated TRA2β poison exon inclusion induces the expression of a lncRNA with anti-tumor effects

Upregulated expression of the oncogenic splicing factor TRA2β occurs in human tumors partly through decreased inclusion of its autoregulatory non-coding poison exon (PE). Here, we reveal that low TRA2β-PE inclusion negatively impacts patient survival across several tumor types. We demonstrate the ability of splice-switching antisense oligonucleotides (ASOs) to promote TRA2β-PE inclusion and lower TRA2β protein levels in pre-clinical cancer models. TRA2β-PE-targeting ASOs induce anti-cancer phenotypes and widespread transcriptomic alterations with functional impact on RNA processing, mTOR, and p53 signaling pathways. Surprisingly, the effect of TRA2β-PE-targeting ASOs on cell viability are not phenocopied by TRA2β knockdown. Mechanistically, we find that the ASO functions by both decreasing TRA2β protein and inducing the expression of TRA2β-PE-containing transcripts that act as long non-coding RNAs to sequester nuclear proteins. Finally, TRA2β-PE-targeting ASOs are toxic to preclinical 3D organoid and in vivo patient-derived xenograft models. Together, we demonstrate that TRA2β-PE acts both as a regulator of protein expression and a long-noncoding RNA to control cancer cell growth. Drugging oncogenic splicing factors using PE-targeting ASOs is a promising therapeutic strategy.

DHX9 helicase impacts on splicing decisions by modulating U2 snRNP recruitment in Ewing sarcoma cells

Ewing sarcomas (ESs) are biologically aggressive tumours of bone and soft tissues caused by chromosomal translocations yielding in-frame fusion proteins driving the neoplastic transformation. The DNA/RNA helicase DHX9 is an important regulator of cellular processes often deregulated in cancer. Using transcriptome profiling, our study reveals cancer-relevant genes whose splicing is modulated by DHX9. Immunodepletion experiments demonstrate that DHX9 impacts on the recruitment of U2 small nuclear RNP (snRNP) onto the pre-mRNA. Analysis of structure and sequence features of DHX9 target exons reveal that DHX9-sensitive exons display shorter flanking introns and contain HNRNPC and TIA1 consensus motifs. A prominent target of DHX9 is exon 11 in the Cortactin (CTTN) gene, which is alternatively spliced to generate isoforms with different activities in cell migration and tumour invasion. Alternative inclusion of the exon 11 in CTTN gene is one of the most recurrent isoform switches in multiple cancer types, thus highlighting the pivotal role of DHX9 in defining the tumour phenotype. Biochemical analyses reveal that DHX9 binding promotes the recruitment of U2snRNP, SF3B1, and SF3A2 to the splice sites flanking exon 11. These findings uncover a new role of DHX9 in the control of co-transcriptional splicing in ES, which may represent a new druggable target to counteract ES malignancy.

Endo-bind-n-seq: identifying RNA motifs of RNA binding proteins isolated from endogenous sources

RNA binding proteins (RBPs) are crucial regulators of gene expression and critically depend on the specific recognition of their target RNAs. Accordingly, a selection of methods to analyze RBP specificities has been developed, including protein-RNA crosslinking and sequencing (CLIP) and in vitro selection methods such as SELEX, RNA compete or RNA bind-n-seq. However, limitations like the availability for purified recombinant proteins and custom microarray platforms (RNAcompete) or extensive sequencing depth and sophisticated bioinformatic data processing (CLIP) may limit a broader implementation of these methods. Here, we present an RNA bind-n-seq method that uses short random RNA pools and enables multiple rounds of selection. This results in strong motif enrichment with low positional variance thus reducing sequencing depth requirements. Furthermore, we have coupled our protocol to immunoprecipitation of tagged or endogenous RBPs from cultured cells or tissue samples, eliminating the need for recombinant proteins. Our method also allows for the identification of indirect RNA motifs of proteins that are integral parts of multiprotein RNPs and result in physically more relevant RNA motifs.

Iron-mediated post-transcriptional regulation in Toxoplasma gondii

Iron is required to support almost all life; however, levels must be carefully regulated to maintain homeostasis. Although the obligate parasite Toxoplasma gondii requires iron, how it responds upon iron limitation has not been investigated. Here, we show that iron depletion triggers significant transcriptional changes in the parasite, including in iron-dependent pathways. We find that a subset of T. gondii transcripts contain stem-loop structures, which have been associated with post-transcriptional iron-mediated regulation in other cellular systems. We validate one of these (found in the 3' UTR of TGME49_261720) using a reporter cell line. We show that the presence of the stem-loop-containing UTR is sufficient to confer accumulation at the transcript and protein levels under low iron. This response is dose and time-dependent and is specific for iron. The accumulation of transcript is likely driven by an increased reporter mRNA stability under low iron. Interestingly, we find iron-mediated changes in mRNA stability in around 400 genes. To examine the potential mechanism of this stability, we tested aconitase interaction with mRNA in low iron and found 43 enriched transcripts, but no specific interaction with our reporter UTR. However, the endogenous UTR led to maintenance of protein levels and increased survival of the parasite under low iron. Our data demonstrate the existence of iron-mediated post-transcriptional regulation in Toxoplasma for the first time; and suggests iron-mediated regulation may be important to the parasite in low iron environments.

A novel NLP-based method and algorithm to discover RNA-binding protein (RBP) motifs, contexts, binding preferences, and interactions

RNA-binding proteins (RBPs) are essential modulators in the regulation of mRNA processing. The binding patterns, interactions, and functions of most RBPs are not well-characterized. Previous studies have shown that motif context is an important contributor to RBP binding specificity, but its precise role remains unclear. Despite recent computational advances to predict RBP binding, existing methods are challenging to interpret and largely lack a categorical focus on RBP motif contexts and RBP-RBP interactions. There remains a need for interpretable predictive models to disambiguate the contextual determinants of RBP binding specificity in vivo . Here, we present a novel and comprehensive pipeline to address these knowledge gaps. We devise a Natural Language Processing-based decomposition method to deconstruct sequences into entities consisting of a central target k -mer and its flanking regions, then use this representation to formulate the RBP binding prediction task as a weakly supervised Multiple Instance Learning problem. To interpret our predictions, we introduce a deterministic motif discovery algorithm designed to handle our data structure, recapitulating the established motifs of numerous RBPs as validation. Importantly, we characterize the binding motifs and binding contexts for 71 RBPs, with many of them being novel. Finally, through feature integration, transitive inference, and a new cross-prediction approach, we propose novel cooperative and competitive RBP-RBP interaction partners and hypothesize their potential regulatory functions. In summary, we present a complete computational strategy for investigating the contextual determinants of specific RBP binding, and we demonstrate the significance of our findings in delineating RBP binding patterns, interactions, and functions.

Heterogeneous nuclear ribonucleoprotein C promotes non-small cell lung cancer progression by enhancing XB130 mRNA stability and translation

BACKGROUND

XB130, a classical adaptor protein, exerts a critical role in diverse cellular processes. Aberrant expression of XB130 is closely associated with tumorigenesis and aggressiveness. However, the mechanisms governing its expression regulation remain poorly understood. Heterogeneous nuclear ribonucleoprotein C (hnRNPC), as an RNA-binding protein, is known to modulate multiple aspects of RNA metabolism and has been implicated in the pathogenesis of various cancers. We have previously discovered that hnRNPC is one of the candidate proteins that interact with the 3' untranslated region (3'UTR) of XB130 in non-small cell lung cancer (NSCLC). Therefore, this study aims to comprehensively elucidate how hnRNPC regulates the expression of XB130 in NSCLC.

MATERIALS AND METHODS

We evaluated the expression of hnRNPC in cancer and assessed the correlation between hnRNPC expression and prognosis in cancer patients using public databases. Subsequently, several stable cell lines were constructed. The proliferation, migration, invasion, and epithelial-mesenchymal transition (EMT) of these cells were detected through Real-time cellular analysis, adherent colony formation, wound healing assay, invasion assay, and Western blotting. The specific regulatory manner between hnRNPC and XB130 was investigated by Real-time quantitative PCR, Western blotting, RNA pull‑down assay, dual‑luciferase reporter assay, RNA immunoprecipitation, and Co-Immunoprecipitation.

RESULTS

We identified that hnRNPC expression is significantly elevated in NSCLC and correlates with poor prognosis in patients with lung adenocarcinoma. HnRNPC overexpression in NSCLC cells increased the expression of XB130, subsequently activating the PI3K/Akt signaling pathway and ultimately promoting cell proliferation and EMT. Additionally, overexpressing XB130 in hnRNPC-silenced cells partially restored cell proliferation and EMT. Mechanistically, hnRNPC specifically bound to the 3'UTR segments of XB130 mRNA, enhancing mRNA stability by inhibiting the recruitment of nucleases 5'-3' exoribonuclease 1 (XRN1) and DIS3-like 3'-5' exoribonuclease 2 (DIS3L2). Furthermore, hnRNPC simultaneously interacted with the eukaryotic initiation factor 4E (eIF4E), a component of the eIF4F complex, facilitating the circularization of XB130 mRNA and thereby increasing its translation efficiency.

CONCLUSIONS

HnRNPC overexpression promotes NSCLC progression by enhancing XB130 mRNA stability and translation, suggesting that hnRNPC might be a potential therapeutic and prognostic target for NSCLC.

The regulatory landscape of 5' UTRs in translational control during zebrafish embryogenesis

The 5' UTRs of mRNAs are critical for translation regulation during development, but their in vivo regulatory features are poorly characterized. Here, we report the regulatory landscape of 5' UTRs during early zebrafish embryogenesis using a massively parallel reporter assay of 18,154 sequences coupled to polysome profiling. We found that the 5' UTR suffices to confer temporal dynamics to translation initiation and identified 86 motifs enriched in 5' UTRs with distinct ribosome recruitment capabilities. A quantitative deep learning model, Danio Optimus 5-Prime (DaniO5P), identified a combined role for 5' UTR length, translation initiation site context, upstream AUGs, and sequence motifs on ribosome recruitment. DaniO5P predicts the activities of maternal and zygotic 5' UTR isoforms and indicates that modulating 5' UTR length and motif grammar contributes to translation initiation dynamics. This study provides a first quantitative model of 5' UTR-based translation regulation in development and lays the foundation for identifying the underlying molecular effectors.

A deep learning model for characterizing protein-RNA interactions from sequences at single-base resolution

Protein-RNA interactions play pivotal roles in regulating transcription, translation, and RNA metabolism. Characterizing these interactions offers key insights into RNA dysregulation mechanisms. Here, we introduce Reformer, a deep learning model that predicts protein-RNA binding affinity from sequence data. Trained on 225 enhanced cross-linking and immunoprecipitation sequencing (eCLIP-seq) datasets encompassing 155 RNA-binding proteins across three cell lines, Reformer achieves high accuracy in predicting binding affinity at single-base resolution. The model uncovers binding motifs that are often undetectable through traditional eCLIP-seq methods. Notably, the motifs learned by Reformer are shown to correlate with RNA processing functions. Validation via electrophoretic mobility shift assays confirms the model's precision in quantifying the impact of mutations on RNA regulation. In summary, Reformer improves the resolution of RNA-protein interaction predictions and aids in prioritizing mutations that influence RNA regulation.

RASP v2.0: an updated atlas for RNA structure probing data

RNA molecules function in numerous biological processes by folding into intricate structures. Here we present RASP v2.0, an updated database for RNA structure probing data featuring a substantially expanded collection of datasets along with enhanced online structural analysis functionalities. Compared to the previous version, RASP v2.0 includes the following improvements: (i) the number of RNA structure datasets has increased from 156 to 438, comprising 216 transcriptome-wide RNA structure datasets, 141 target-specific RNA structure datasets, and 81 RNA-RNA interaction datasets, thereby broadening species coverage from 18 to 24, (ii) a deep learning-based model has been implemented to impute missing structural signals for 59 transcriptome-wide RNA structure datasets with low structure score coverage, significantly enhancing data quality, particularly for low-abundance RNAs, (iii) three new online analysis modules have been deployed to assist RNA structure studies, including missing structure score imputation, RNA secondary and tertiary structure prediction, and RNA binding protein (RBP) binding prediction. By providing a resource of much more comprehensive RNA structure data, RASP v2.0 is poised to facilitate the exploration of RNA structure-function relationships across diverse biological processes. RASP v2.0 is freely accessible at http://rasp2.zhanglab.net/.

OncoSplicing 3.0: an updated database for identifying RBPs regulating alternative splicing events in cancers

Alternative splicing (AS) is a crucial mechanism to regulate gene expression and protein complexity. RNA-binding proteins (RBPs) play an important role in regulating abnormal alternative splicing in cancers. However, few resources are available to identify specific RBPs responsible for regulating individual AS event. We have developed the OncoSplicing database for integrative analysis of clinically relevant alternative splicing events in TCGA cancers. Here, we further updated the OncoSplicing database by performing correlation analysis between the splicing and mRNA expression data from the TCGA cancers or GTEx tissues, mapping known RNA-binding motifs and eCLIP-seq peaks to all AS events, conducting splicing analysis for RNA-seq data from RBP perturbation experiments in the ENCODE project, and integrating exon and intron sequences for each AS event. With this updated database, users can easily identify potential RBPs responsible for the queried AS event and obtain sequences to design AS-specific primers and minigene constructs for experiment validation. Overall, compared to the previous version, the substantially updated OncoSplicing database (www.oncosplicing.com) offers a more valuable resource for users to identify RBPs responsible for regulating alternative splicing events in cancers.

A generative framework for enhanced cell-type specificity in rationally designed mRNAs

mRNA delivery offers new opportunities for disease treatment by directing cells to produce therapeutic proteins. However, designing highly stable mRNAs with programmable cell type-specificity remains a challenge. To address this, we measured the regulatory activity of 60,000 5' and 3' untranslated regions (UTRs) across six cell types and developed PARADE (Prediction And RAtional DEsign of mRNA UTRs), a generative AI framework to engineer untranslated RNA regions with tailored cell type-specific activity. We validated PARADE by testing 15,800 de novo-designed sequences across these cell lines and identified many sequences that demonstrated superior specificity and activity compared to existing RNA therapeutics. mRNAs with PARADE-engineered UTRs also exhibited robust tissue-specific activity in animal models, achieving selective expression in the liver and spleen. We also leveraged PARADE to enhance mRNA stability, significantly increasing protein output and therapeutic durability in vivo. These advancements translated to notable increases in therapeutic efficacy, as PARADE-designed UTRs in oncosuppressor mRNAs, namely PTEN and P16, effectively reduced tumor growth in patient-derived neuroglioma xenograft models and orthotopic mouse models. Collectively, these findings establish PARADE as a versatile platform for designing safer, more precise, and highly stable mRNA therapies.

NaP-TRAP reveals the regulatory grammar in 5'UTR-mediated translation regulation during zebrafish development

The cis-regulatory elements encoded in an mRNA determine its stability and translational output. While there has been a considerable effort to understand the factors driving mRNA stability, the regulatory frameworks governing translational control remain more elusive. We have developed a novel massively parallel reporter assay (MPRA) to measure mRNA translation, named Nascent Peptide Translating Ribosome Affinity Purification (NaP-TRAP). NaP-TRAP measures translation in a frame-specific manner through the immunocapture of epitope tagged nascent peptides of reporter mRNAs. We benchmark NaP-TRAP to polysome profiling and use it to quantify Kozak strength and the regulatory landscapes of 5' UTRs in the developing zebrafish embryo and in human cells. Through this approach we identified general and developmentally dynamic cis-regulatory elements, as well as potential trans-acting proteins. We find that U-rich motifs are general enhancers, and upstream ORFs and GC-rich motifs are global repressors of translation. We also observe a translational switch during the maternal-to-zygotic transition, where C-rich motifs shift from repressors to prominent activators of translation. Conversely, we show that microRNA sites in the 5' UTR repress translation following the zygotic expression of miR-430. Together these results demonstrate that NaP-TRAP is a versatile, accessible, and powerful method to decode the regulatory functions of UTRs across different systems.

EDCLoc: a prediction model for mRNA subcellular localization using improved focal loss to address multi-label class imbalance

BACKGROUND

The subcellular localization of mRNA plays a crucial role in gene expression regulation and various cellular processes. However, existing wet lab techniques like RNA-FISH are usually time-consuming, labor-intensive, and limited to specific tissue types. Researchers have developed several computational methods to predict mRNA subcellular localization to address this. These methods face the problem of class imbalance in multi-label classification, causing models to favor majority classes and overlook minority classes during training. Additionally, traditional feature extraction methods have high computational costs, incomplete features, and may lead to the loss of critical information. On the other hand, deep learning methods face challenges related to hardware performance and training time when handling complex sequences. They may suffer from the curse of dimensionality and overfitting problems. Therefore, there is an urgent need for more efficient and accurate prediction models.

RESULTS

To address these issues, we propose a multi-label classifier, EDCLoc, for predicting mRNA subcellular localization. EDCLoc reduces training pressure through a stepwise pooling strategy and applies grouped convolution blocks of varying sizes at different levels, combined with residual connections, to achieve efficient feature extraction and gradient propagation. The model employs global max pooling at the end to further reduce feature dimensions and highlight key features. To tackle class imbalance, we improved the focal loss function to enhance the model's focus on minority classes. Evaluation results show that EDCLoc outperforms existing methods in most subcellular regions. Additionally, the position weight matrix extracted by multi-scale CNN filters can match known RNA-binding protein motifs, demonstrating EDCLoc's effectiveness in capturing key sequence features.

CONCLUSIONS

EDCLoc outperforms existing prediction tools in most subcellular regions and effectively mitigates class imbalance issues in multi-label classification. These advantages make EDCLoc a reliable choice for multi-label mRNA subcellular localization. The dataset and source code used in this study are available at https://github.com/DellCode233/EDCLoc .

Single-nucleus transcriptome atlas of orbitofrontal cortex in amyotrophic lateral sclerosis with a deep learning-based decoding of alternative polyadenylation mechanisms

Amyotrophic lateral sclerosis (ALS) and frontotemporal lobar degeneration (FTLD) are two age-related and fatal neurodegenerative disorders that lie on a shared disease spectrum. While both disorders involve complex interactions between neuronal and glial cells, the specific cell-type alterations and their contributions to disease pathophysiology remain incompletely understood. Here, we applied single-nucleus RNA sequencing of the orbitofrontal cortex, a region affected in ALS-FTLD, to map cell-type specific transcriptional signatures in C9orf72-related ALS (with and without FTLD) and sporadic ALS cases. Our findings reveal disease- and cell-type-specific transcriptional changes, with neurons exhibiting the most pronounced alterations, primarily affecting mitochondrial function, protein homeostasis, and chromatin remodeling. A comparison with independent datasets from different cortical regions of C9orf72 and sporadic ALS cases showed concordance in several pathways, with neuronal STMN2 and NEFL showing consistent up-regulation between brain regions and disease subtypes. We also interrogated alternative polyadenylation (APA) as an additional layer of transcriptional regulation, demonstrating that APA events are not correlated with identified gene expression changes. To interpret these events, we developed APA-Net, a deep learning model that integrates transcript sequences with RNA-binding protein expression profiles, revealing cell type-specific patterns of APA regulation. Our atlas illuminates cell type-specific pathomechanisms of ALS/FTLD, providing a valuable resource for further investigation.

At-RS31 orchestrates hierarchical cross-regulation of splicing factors and integrates alternative splicing with TOR-ABA pathways

Alternative splicing is essential for plants, enabling a single gene to produce multiple transcript variants to boost functional diversity and fine-tune responses to environmental and developmental cues. At-RS31, a plant-specific splicing factor in the Serine/Arginine (SR)-rich protein family, responds to light and the Target of Rapamycin (TOR) signaling pathway, yet its downstream targets and regulatory impact remain unknown.To identify At-RS31 targets, we applied individual-nucleotide resolution crosslinking and immunoprecipitation (iCLIP) and RNAcompete assays. Transcriptomic analyses of At-RS31 mutant and overexpressing plants further revealed its effects on alternative splicing.iCLIP identified 4,034 At-RS31 binding sites across 1,421 genes, enriched in CU-rich and CAGA RNA motifs. Comparative iCLIP and RNAcompete data indicate that the RS domain of At-RS31 may influence its binding specificity in planta, underscoring the value of combining in vivo and in vitro approaches. Transcriptomic analysis showed that At-RS31 modulates diverse splicing events, particularly intron retention and exitron splicing, and influences other splicing modulators, acting as a hierarchical regulator.By regulating stress-response genes and genes in both TOR and abscisic acid (ABA) signaling pathways, At-RS31 may help integrate these signals, balancing plant growth with environmental adaptability through alternative splicing.

DeepFusion: A deep bimodal information fusion network for unraveling protein-RNA interactions using in vivo RNA structures

RNA-binding proteins (RBPs) are key post-transcriptional regulators, and the malfunctions of RBP-RNA binding lead to diverse human diseases. However, prediction of RBP binding sites is largely based on RNA sequence features, whereas in vivo RNA structural features based on high-throughput sequencing are rarely incorporated. Here, we designed a deep bimodal information fusion network called DeepFusion for unraveling protein-RNA interactions by incorporating structural features derived from DMS-seq data. DeepFusion integrates two sub-models to extract local motif-like information and long-term context information. We show that DeepFusion performs best compared with other cutting-edge methods with only sequence inputs on two datasets. DeepFusion's performance is further improved with bimodal input after adding in vivo DMS-seq structural features. Furthermore, DeepFusion can be used for analyzing RNA degradation, demonstrating significantly different RBP-binding scores in genes with slow degradation rates versus those with rapid degradation rates. DeepFusion thus provides enhanced abilities for further analysis of functional RNAs. DeepFusion's code and data are available at http://bioinfo.org/deepfusion/.

From genes to traits: Trends in RNA-binding proteins and their role in plant trait development: A review

RNA-binding proteins (RBPs) are essential for cellular functions by attaching to RNAs, creating dynamic ribonucleoprotein complexes (RNPs) essential for managing RNA throughout its life cycle. These proteins are critical to all post-transcriptional processes, impacting vital cellular functions during development and adaptation to environmental changes. Notably, in plants, RBPs are critical for adjusting to inconsistent environmental conditions, with recent studies revealing that plants possess, more prominent, and both novel and conserved RBP families compared to other eukaryotes. This comprehensive review delves into the varied RBPs covering their structural attributes, domain base function, and their interactions with RNA in metabolism, spotlighting their role in regulating post-transcription and splicing and their reaction to internal and external stimuli. It highlights the complex regulatory roles of RBPs, focusing on plant trait regulation and the unique functions they facilitate, establishing a foundation for appreciating RBPs' significance in plant growth and environmental response strategies.

Canada's contributions to RNA research: past, present, and future perspectives

The field of RNA research has provided profound insights into the basic mechanisms modulating the function and adaption of biological systems. RNA has also been at the center stage in the development of transformative biotechnological and medical applications, perhaps most notably was the advent of mRNA vaccines that were critical in helping humanity through the Covid-19 pandemic. Unbeknownst to many, Canada boasts a diverse community of RNA scientists, spanning multiple disciplines and locations, whose cutting-edge research has established a rich track record of contributions across various aspects of RNA science over many decades. Through this position paper, we seek to highlight key contributions made by Canadian investigators to the RNA field, via both thematic and historical viewpoints. We also discuss initiatives underway to organize and enhance the impact of the Canadian RNA research community, particularly focusing on the creation of the not-for-profit organization RNA Canada ARN. Considering the strategic importance of RNA research in biology and medicine, and its considerable potential to help address major challenges facing humanity, sustained support of this sector will be critical to help Canadian scientists play key roles in the ongoing RNA revolution and the many benefits this could bring about to Canada.

Systematic analysis of the target recognition and repression by the Pumilio proteins

RNA binding proteins orchestrate the post-transcriptional fate of RNA molecules, but the principles of their action remain poorly understood. Pumilio (PUM) proteins bind 3' UTRs of mRNAs and lead to mRNA decay. To comprehensively map the determinants of recognition of sequences by PUM proteins in cells and to study the binding outcomes, we developed a massively parallel RNA assay that profiled thousands of PUM-binding sites in cells undergoing various perturbations or RNA immunoprecipitation. By studying fragments from the NORAD long non-coding RNA, we find two features that antagonize repression by PUM proteins - G/C rich sequences, particularly those upstream of the PUM recognition element, and binding of FAM120A, which limits the repression elicited by PUM-binding sites. We also find that arrays of PUM sites separated by 8-12 bases offer particularly strong repression and use them to develop a particularly sensitive reporter for PUM repression. In contrast, PUM sites separated by shorter linkers, such as some of those found in NORAD, exhibit strong activity interdependence, likely mediated by competition between PUM binding and formation of strong secondary structures. Overall, our findings expand our understanding of the determinants of PUM protein activity in human cells.

ALS-associated FUS mutation reshapes the RNA and protein composition of stress granules

Stress granules (SG) are part of a cellular protection mechanism where untranslated messenger RNAs and RNA-binding proteins are stored upon conditions of cellular stress. Compositional variations due to qualitative or quantitative protein changes can disrupt their functionality and alter their structure. This is the case of different forms of amyotrophic lateral sclerosis (ALS) where a causative link has been proposed between the cytoplasmic de-localization of mutant proteins, such as FUS (Fused in Sarcoma), and the formation of cytotoxic inclusions. Here, we describe the SG transcriptome in neuroblastoma cells and define several features for RNA recruitment in these condensates. We demonstrate that SG dynamics and RNA content are strongly modified by the incorporation of mutant FUS, switching to a more unstructured, AU-rich SG transcriptome. Moreover, we show that mutant FUS, together with its protein interactors and their target RNAs, are responsible for the reshaping of the mutant SG transcriptome with alterations that can be linked to neurodegeneration. Our data describe the molecular differences between physiological and pathological SG in ALS-FUS conditions, showing how FUS mutations impact the RNA and protein composition of these condensates.

Long-read subcellular fractionation and sequencing reveals the translational fate of full-length mRNA isoforms during neuronal differentiation

Alternative splicing (AS) alters the cis-regulatory landscape of mRNA isoforms, leading to transcripts with distinct localization, stability, and translational efficiency. To rigorously investigate mRNA isoform-specific ribosome association, we generated subcellular fractionation and sequencing (Frac-seq) libraries using both conventional short reads and long reads from human embryonic stem cells (ESCs) and neural progenitor cells (NPCs) derived from the same ESCs. We performed de novo transcriptome assembly from high-confidence long reads from cytosolic, monosomal, light, and heavy polyribosomal fractions and quantified their abundance using short reads from their respective subcellular fractions. Thousands of transcripts in each cell type exhibited association with particular subcellular fractions relative to the cytosol. Of the multi-isoform genes, 27% and 19% exhibited significant differential isoform sedimentation in ESCs and NPCs, respectively. Alternative promoter usage and internal exon skipping accounted for the majority of differences between isoforms from the same gene. Random forest classifiers implicated coding sequence (CDS) and untranslated region (UTR) lengths as important determinants of isoform-specific sedimentation profiles, and motif analyses reveal potential cell type-specific and subcellular fraction-associated RNA-binding protein signatures. Taken together, our data demonstrate that alternative mRNA processing within the CDS and UTRs impacts the translational control of mRNA isoforms during stem cell differentiation, and highlight the utility of using a novel long-read sequencing-based method to study translational control.

A bioinformatic approach for the prediction and functional classification of Toxoplasma gondii long non-coding RNAs

Long non-coding RNAs (lncRNAs) have emerged as significant players in diverse cellular processes, including cell differentiation. Advancements in computational methodologies have facilitated the prediction of lncRNA functions, enabling insights even in non-model organisms like pathogenic parasites, in roles such as parasite development, antigenic variation, and epigenetics. In this work, we focus on the apicomplexan Toxoplasma gondii differentiation process, where the infective stage, tachyzoite, can develop into the cysted stage, bradyzoite, under stress conditions. Using a publicly available transcriptome dataset, we predicted putative lncRNA sequences associated with this differentiation process. Notably, a substantial proportion of these putative lncRNAs exhibited stage-specific expression, particularly at the bradyzoite stage. Furthermore, co-expression patterns between coding transcripts and putative TglncRNAs suggest their involvement in shared processes, such as bradyzoite development. Putative TglncRNA loci analysis revealed their potential influence on the expression of nearby coding genes, including subtelomeric genes unique to the T. gondii genome. Finally we propose a k-mer analysis approach to predict putative functional relationships between characterized lncRNAs from model organisms like Homo sapiens and the putative T. gondii lncRNAs. Our perspective led to predict putative T. gondii lncRNA that potentially could act mediating DNA damage repair pathways, opening a new study field to validate this kind of adaptive mechanisms of T. gondii in response to stress conditions.

PTBP1 mediates Sertoli cell actin cytoskeleton organization by regulating alternative splicing of actin regulators

Spermatogenesis is a biological process within the testis that produces haploid spermatozoa for the continuity of species. Sertoli cells are somatic cells in the seminiferous epithelium that orchestrate spermatogenesis. Cyclic reorganization of the Sertoli cell actin cytoskeleton is vital for spermatogenesis, but the underlying mechanism remains largely unclear. Here, we report that the RNA-binding protein PTBP1 controls Sertoli cell actin cytoskeleton reorganization by programming alternative splicing of actin cytoskeleton regulators. This splicing control enables ectoplasmic specializations, the actin-based adhesion junctions, to maintain the blood-testis barrier and support spermatid transport and transformation. Particularly, we show that PTBP1 promotes actin bundle formation by repressing the inclusion of exon 14 of Tnik, a kinase present at the ectoplasmic specialization. Our results thus reveal a novel mechanism wherein Sertoli cell actin cytoskeleton dynamics are controlled post-transcriptionally by utilizing functionally distinct isoforms of actin regulatory proteins, and PTBP1 is a critical regulatory factor in generating such isoforms.

Phylogenomic instructed target analysis reveals ELAV complex binding to multiple optimally spaced U-rich motifs

ELAV/Hu RNA-binding proteins are gene-specific regulators of alternative pre-mRNA processing. ELAV/Hu family proteins bind to short AU-rich motifs which are abundant in pre-mRNA, making it unclear how they achieve gene specificity. ELAV/Hu proteins multimerize, but how multimerization contributes to decode degenerate sequence environments remains uncertain. Here, we show that ELAV forms a saturable complex on extended RNA. Through phylogenomic instructed target analysis we identify the core binding motif U5N2U3, which is repeated in an extended binding site. Optimally spaced short U5N2U3 binding motifs are key for high-affinity binding in this minimal binding element. Binding strength correlates with ELAV-regulated alternative poly(A) site choice, which is physiologically relevant through regulation of the major ELAV target ewg in determining synapse numbers. We further identify a stem-loop secondary structure in the ewg binding site unwound upon ELAV binding at three distal U motifs. Base-pairing of U motifs prevents ELAV binding, but N6-methyladenosine (m6A) has little effect. Further, stem-loops are enriched in ELAV-regulated poly(A) sites. Additionally, ELAV can nucleate preferentially from 3' to 5'. Hence, we identify a decisive mechanism for ELAV complex formation, addressing a fundamental gap in understanding how ELAV/Hu family proteins decode degenerate sequence spaces for gene-specific mRNA processing.

Mature microRNA-binding protein QKI suppresses extracellular microRNA let-7b release

Argonaute (AGO), a component of RNA-induced silencing complexes (RISCs), is a representative RNA-binding protein (RBP) known to bind with mature microRNAs (miRNAs) and is directly involved in post-transcriptional gene silencing. However, despite the biological significance of miRNAs, the roles of other miRNA-binding proteins (miRBPs) remain unclear in the regulation of miRNA loading, dissociation from RISCs and extracellular release. In this study, we performed protein arrays to profile miRBPs and identify 118 RBPs that directly bind to miRNAs. Among those proteins, the RBP quaking (QKI) inhibits extracellular release of the mature microRNA let-7b by controlling the loading of let-7b into extracellular vesicles via additional miRBPs such as AUF1 (also known as hnRNPD) and hnRNPK. The enhanced extracellular release of let-7b after QKI depletion activates Toll-like receptor 7 (TLR7) and promotes the production of proinflammatory cytokines in recipient cells, leading to brain inflammation in the mouse cortex. Thus, this study reveals the contribution of QKI to the inhibition of brain inflammation via regulation of extracellular let-7b release.

Circular RNA landscape in extracellular vesicles from human biofluids

BACKGROUND

Circular RNAs (circRNAs) have emerged as a prominent class of covalently closed single-stranded RNA molecules that exhibit tissue-specific expression and potential as biomarkers in extracellular vesicles (EVs) derived from liquid biopsies. Still, their characteristics and applications in EVs remain to be unveiled.

METHODS

We performed a comprehensive analysis of EV-derived circRNAs (EV-circRNAs) using transcriptomics data obtained from 1082 human body fluids, including plasma, urine, cerebrospinal fluid (CSF), and bile. Our validation strategy utilized RT-qPCR and RNA immunoprecipitation assays, complemented by computational techniques for analyzing EV-circRNA features and RNA-binding protein interactions.

RESULTS

We identified 136,327 EV-circRNAs from various human body fluids. Significantly, a considerable amount of circRNAs with a high back-splicing ratio are highly enriched in EVs compared to linear RNAs. Additionally, we discovered brain-specific circRNAs enriched in plasma EVs and cancer-associated EV-circRNAs linked to clinical outcomes. Moreover, we demonstrated that EV-circRNAs have the potential to serve as biomarkers for evaluating immunotherapy efficacy in non-small cell lung cancer (NSCLC). Importantly, we identified the involvement of RBPs, particularly YBX1, in the sorting mechanism of circRNAs into EVs.

CONCLUSIONS

This study unveils the extensive repertoire of EV-circRNAs across human biofluids, offering insights into their potential as disease biomarkers and their mechanistic roles within EVs. The identification of specific circRNAs and the elucidation of RBP-mediated sorting mechanisms open new avenues for the clinical application of EV-circRNAs in disease diagnostics and therapeutics.

Chromatin-associated lncRNA-splicing factor condensates regulate hypoxia responsive RNA processing of genes pre-positioned near nuclear speckles

Hypoxia-induced alternative splicing (AS) regulates tumor progression and metastasis. Little is known about how such AS is controlled and whether higher-order genome and nuclear domain (ND) organizations dictate these processes. We observe that hypoxia-responsive alternatively spliced genes position near nuclear speckle (NS), the ND that enhances splicing efficiency. NS-resident MALAT1 long noncoding RNA, induced in response to hypoxia, regulates hypoxia-responsive AS. MALAT1 achieves this by organizing the SR-family of splicing factor, SRSF1, near NS and regulating the binding of SRSF1 to pre-mRNAs. Mechanistically, MALAT1 enhances the recruitment of SRSF1 to elongating RNA polymerase II (pol II) by promoting the formation of phase-separated condensates of SRSF1, which are preferentially recognized by pol II. During hypoxia, MALAT1 regulates spatially organized AS by establishing a threshold SRSF1 concentration near NSs, potentially by forming condensates, critical for pol II-mediated recruitment of SRSF1 to pre-mRNAs.

ADAR3 modulates neuronal differentiation and regulates mRNA stability and translation

ADAR3 is a catalytically inactive member of the family of adenosine deaminases acting on RNA (ADARs). Here we have investigated its function in the context of the developing mouse brain. The expression of ADAR3 gradually increases throughout embryogenesis and drops after birth. Using primary cortical neurons, we show that ADAR3 is only expressed in a subpopulation of in vitro differentiated neurons, which suggests specific functions rather than being a general regulator of ADAR editing in the brain. The analysis of the ADAR3 interactome suggested a role in mRNA stability and translation, and we show that expression of ADAR3 in a neuronal cell line that is otherwise ADAR3-negative changes the expression and stability of a large number of mRNAs. Notably, we show that ADAR3 associates with polysomes and inhibits translation. We propose that ADAR3 binds to target mRNAs and stabilizes them in non-productive polysome complexes. Interestingly, the expression of ADAR3 downregulates genes related to neuronal differentiation and inhibits neurofilament outgrowth in vitro. In summary, we propose that ADAR3 negatively regulates neuronal differentiation, and that it does so by regulating mRNA stability and translation in an editing-independent manner.

Reconstructing the sequence specificities of RNA-binding proteins across eukaryotes

RNA-binding proteins (RBPs) are key regulators of gene expression. Here, we introduce EuPRI (Eukaryotic Protein-RNA Interactions) - a freely available resource of RNA motifs for 34,736 RBPs from 690 eukaryotes. EuPRI includes in vitro binding data for 504 RBPs, including newly collected RNAcompete data for 174 RBPs, along with thousands of reconstructed motifs. We reconstruct these motifs with a new computational platform - Joint Protein-Ligand Embedding (JPLE) - which can detect distant homology relationships and map specificity-determining peptides. EuPRI quadruples the number of known RBP motifs, expanding the motif repertoire across all major eukaryotic clades, and assigning motifs to the majority of human RBPs. EuPRI drastically improves knowledge of RBP motifs in flowering plants. For example, it increases the number of Arabidopsis thaliana RBP motifs 7-fold, from 14 to 105. EuPRI also has broad utility for inferring post-transcriptional function and evolutionary relationships. We demonstrate this by predicting a role for 12 Arabidopsis thaliana RBPs in RNA stability and identifying rapid and recent evolution of post-transcriptional regulatory networks in worms and plants. In contrast, the vertebrate RNA motif set has remained relatively stable after its drastic expansion between the metazoan and vertebrate ancestors. EuPRI represents a powerful resource for the study of gene regulation across eukaryotes.

The MTR4/hnRNPK complex surveils aberrant polyadenylated RNAs with multiple exons

RNA surveillance systems degrade aberrant RNAs that result from defective transcriptional termination, splicing, and polyadenylation. Defective RNAs in the nucleus are recognized by RNA-binding proteins and MTR4, and are degraded by the RNA exosome complex. Here, we detect aberrant RNAs in MTR4-depleted cells using long-read direct RNA sequencing and 3' sequencing. MTR4 destabilizes intronic polyadenylated transcripts generated by transcriptional read-through over one or more exons, termed 3' eXtended Transcripts (3XTs). MTR4 also associates with hnRNPK, which recognizes 3XTs with multiple exons. Moreover, the aberrant protein translated from KCTD13 3XT is a target of the hnRNPK-MTR4-RNA exosome pathway and forms aberrant condensates, which we name KCTD13 3eXtended Transcript-derived protein (KeXT) bodies. Our results suggest that RNA surveillance in human cells inhibits the formation of condensates of a defective polyadenylated transcript-derived protein.

An ancient competition for the conserved branchpoint sequence influences physiological and evolutionary outcomes in splicing

Recognition of the intron branchpoint during spliceosome assembly is a multistep process that defines both mRNA structure and amount. A branchpoint sequence motif UACUAAC is variably conserved in eukaryotic genomes, but in some organisms more than one protein can recognize it. Here we show that SF1 and Quaking (QKI) compete for a subset of intron branchpoints with the sequence ACUAA. SF1 activates exon inclusion through this sequence, but QKI represses the inclusion of alternatively spliced exons with this intron branchpoint sequence. Using mutant reporters derived from a natural intron with two branchpoint-like sequences, we find that when either branchpoint sequence is mutated, the other is used as a branchpoint, but when both are present, neither is used due to high affinity binding and strong splicing repression by QKI. QKI occupancy at the dual branchpoint site directly prevents SF1 binding and subsequent recruitment of spliceosome-associated factors. Finally, the ectopic expression of QKI in budding yeast (which lacks QKI) is lethal, due at least in part to widespread splicing repression. In conclusion, QKI can function as a splicing repressor by directly competing with SF1/BBP for a subset of branchpoint sequences that closely mirror its high affinity binding site. This suggests that QKI and degenerate branchpoint sequences may have co-evolved as a means through which specific gene expression patterns could be maintained in QKI-expressing or non-expressing cells in metazoans, plants, and animals.

Deciphering 3'UTR Mediated Gene Regulation Using Interpretable Deep Representation Learning

The 3' untranslated regions (3'UTRs) of messenger RNAs contain many important cis-regulatory elements that are under functional and evolutionary constraints. It is hypothesized that these constraints are similar to grammars and syntaxes in human languages and can be modeled by advanced natural language techniques such as Transformers, which has been very effective in modeling complex protein sequence and structures. Here 3UTRBERT is described, which implements an attention-based language model, i.e., Bidirectional Encoder Representations from Transformers (BERT). 3UTRBERT is pre-trained on aggregated 3'UTR sequences of human mRNAs in a task-agnostic manner; the pre-trained model is then fine-tuned for specific downstream tasks such as identifying RBP binding sites, m6A RNA modification sites, and predicting RNA sub-cellular localizations. Benchmark results show that 3UTRBERT generally outperformed other contemporary methods in each of these tasks. More importantly, the self-attention mechanism within 3UTRBERT allows direct visualization of the semantic relationship between sequence elements and effectively identifies regions with important regulatory potential. It is expected that 3UTRBERT model can serve as the foundational tool to analyze various sequence labeling tasks within the 3'UTR fields, thus enhancing the decipherability of post-transcriptional regulatory mechanisms.

GenerRNA: A generative pre-trained language model for de novo RNA design

The design of RNA plays a crucial role in developing RNA vaccines, nucleic acid therapeutics, and innovative biotechnological tools. However, existing techniques frequently lack versatility across various tasks and are dependent on pre-defined secondary structure or other prior knowledge. To address these limitations, we introduce GenerRNA, a Transformer-based model inspired by the success of large language models (LLMs) in protein and molecule generation. GenerRNA is pre-trained on large-scale RNA sequences and capable of generating novel RNA sequences with stable secondary structures, while ensuring distinctiveness from existing sequences, thereby expanding our exploration of the RNA space. Moreover, GenerRNA can be fine-tuned on smaller, specialized datasets for specific subtasks, enabling the generation of RNAs with desired functionalities or properties without requiring any prior knowledge input. As a demonstration, we fine-tuned GenerRNA and successfully generated novel RNA sequences exhibiting high affinity for target proteins. Our work is the first application of a generative language model to RNA generation, presenting an innovative approach to RNA design.

RNAelem: an algorithm for discovering sequence-structure motifs in RNA bound by RNA-binding proteins

Motivation

RNA-binding proteins (RBPs) play a crucial role in the post-transcriptional regulation of RNA. Given their importance, analyzing the specific RNA patterns recognized by RBPs has become a significant research focus in bioinformatics. Deep Neural Networks have enhanced the accuracy of prediction for RBP-binding sites, yet understanding the structural basis of RBP-binding specificity from these models is challenging due to their limited interpretability. To address this, we developed RNAelem, which combines profile context-free grammar and the Turner energy model for RNA secondary structure to predict sequence-structure motifs in RBP-binding regions.

Results

RNAelem exhibited superior detection accuracy compared to existing tools for RNA sequences with structural motifs. Upon applying RNAelem to the eCLIP database, we were not only able to reproduce many known primary sequence motifs in the absence of secondary structures, but also discovered many secondary structural motifs that contained sequence-nonspecific insertion regions. Furthermore, the high interpretability of RNAelem yielded insightful findings such as long-range base-pairing interactions in the binding region of the U2AF protein.

Availability and implementation

The code is available at https://github.com/iyak/RNAelem.

A spatiotemporally resolved atlas of mRNA decay in the C. elegans embryo reveals differential regulation of mRNA stability across stages and cell types

During embryonic development, cells undergo dynamic changes in gene expression that are required for appropriate cell fate specification. Although both transcription and mRNA degradation contribute to gene expression dynamics, patterns of mRNA decay are less well understood. Here, we directly measure spatiotemporally resolved mRNA decay rates transcriptome-wide throughout C. elegans embryogenesis by transcription inhibition followed by bulk and single-cell RNA sequencing. This allows us to calculate mRNA half-lives within specific cell types and developmental stages, and identify differentially regulated mRNA decay throughout embryonic development. We identify transcript features that are correlated with mRNA stability and find that mRNA decay rates are associated with distinct peaks in gene expression over time. Moreover, we provide evidence that, on average, mRNA is more stable in the germline than in the soma and in later embryonic stages than in earlier stages. This work suggests that differential mRNA decay across cell states and time helps to shape developmental gene expression, and it provides a valuable resource for studies of mRNA turnover regulatory mechanisms.

CircHTT(2,3,4,5,6) - co-evolving with the HTT CAG-repeat tract - modulates Huntington's disease phenotypes

Circular RNA (circRNA) molecules have critical functions during brain development and in brain-related disorders. Here, we identified and validated a circRNA, circHTT(2,3,4,5,6), stemming from the Huntington's disease (HD) gene locus that is most abundant in the central nervous system (CNS). We uncovered its evolutionary conservation in diverse mammalian species, and a correlation between circHTT(2,3,4,5,6) levels and the length of the CAG-repeat tract in exon-1 of HTT in human and mouse HD model systems. The mouse orthologue, circHtt(2,3,4,5,6), is expressed during embryogenesis, increases during nervous system development, and is aberrantly upregulated in the presence of the expanded CAG tract. While an IRES-like motif was predicted in circH TT (2,3,4,5,6), the circRNA does not appear to be translated in adult mouse brain tissue. Nonetheless, a modest, but consistent fraction of circHtt(2,3,4,5,6) associates with the 40S ribosomal subunit, suggesting a possible role in the regulation of protein translation. Finally, circHtt(2,3,4,5,6) overexpression experiments in HD-relevant STHdh striatal cells revealed its ability to modulate CAG expansion-driven cellular defects in cell-to-substrate adhesion, thus uncovering an unconventional modifier of HD pathology.

Post-transcriptional mechanisms controlling neurogenesis and direct neuronal reprogramming

Neurogenesis is a tightly regulated process in time and space both in the developing embryo and in adult neurogenic niches. A drastic change in the transcriptome and proteome of radial glial cells or neural stem cells towards the neuronal state is achieved due to sophisticated mechanisms of epigenetic, transcriptional, and post-transcriptional regulation. Understanding these neurogenic mechanisms is of major importance, not only for shedding light on very complex and crucial developmental processes, but also for the identification of putative reprogramming factors, that harbor hierarchically central regulatory roles in the course of neurogenesis and bare thus the capacity to drive direct reprogramming towards the neuronal fate. The major transcriptional programs that orchestrate the neurogenic process have been the focus of research for many years and key neurogenic transcription factors, as well as repressor complexes, have been identified and employed in direct reprogramming protocols to convert non-neuronal cells, into functional neurons. The post-transcriptional regulation of gene expression during nervous system development has emerged as another important and intricate regulatory layer, strongly contributing to the complexity of the mechanisms controlling neurogenesis and neuronal function. In particular, recent advances are highlighting the importance of specific RNA binding proteins that control major steps of mRNA life cycle during neurogenesis, such as alternative splicing, polyadenylation, stability, and translation. Apart from the RNA binding proteins, microRNAs, a class of small non-coding RNAs that block the translation of their target mRNAs, have also been shown to play crucial roles in all the stages of the neurogenic process, from neural stem/progenitor cell proliferation, neuronal differentiation and migration, to functional maturation. Here, we provide an overview of the most prominent post-transcriptional mechanisms mediated by RNA binding proteins and microRNAs during the neurogenic process, giving particular emphasis on the interplay of specific RNA binding proteins with neurogenic microRNAs. Taking under consideration that the molecular mechanisms of neurogenesis exert high similarity to the ones driving direct neuronal reprogramming, we also discuss the current advances in in vitro and in vivo direct neuronal reprogramming approaches that have employed microRNAs or RNA binding proteins as reprogramming factors, highlighting the so far known mechanisms of their reprogramming action.

Structural basis for RNA recognition by the C-terminal RRM domain of human RBM45

RBM45 is an RNA-binding protein with roles in neural development by regulating RNA splicing. Its dysfunction and aggregation are associated with neurodegenerative diseases such as amyotrophic lateral sclerosis (ALS) and frontotemporal lobar dementia (FTLD). RBM45 harbors three RRM domains that potentially bind RNA. While the recognitions of RNA by its N-terminal tandem RRM domains (RRM1 and RRM2) have been well understood, the RNA-binding property of its C-terminal RRM (RRM3) remains unclear. In this work, we identified that the RRM3 of the RBM45 sequence specifically binds RNA with a GACG sequence, similar but not identical to those recognized by the RRM1 and RRM2. Further, we determined the crystal structure of RBM45RRM3 in complex with a GACG sequence-containing single-stranded DNA. Our structural results, together with the RNA-binding assays of mutants at key amino acid residues, revealed the molecular mechanism by which RBM45RRM3 recognizes an RNA sequence. Our finding on the RNA-binding property of the individual RRM module of RBM45 provides the foundation for unraveling the RNA-binding characteristics of full-length RBM45 and for understanding the biological functions of RBM45.

Rbm24 modulates neuronal RNA splicing to restrict cognitive dysfunction

Synaptic dysfunction is associated with early neurodegenerative changes and cognitive deficits. Neuronal cell-specific alternative splicing (AS) programs exclusively encode unique neuron- and synapse-specific proteins. However, it remains unclear whether splicing disturbances in neurons influence the pathogenesis of cognitive impairment. Here, we observed that RNA-binding motif protein 24 (RBM24) expression was decreased in Alzheimer's disease (AD) patients. Using conditional RBM24 knockout mice, we demonstrated that deletion of RBM24 in the brain resulted in learning and memory impairment. Electrophysiological recordings from hippocampal slices from mice lacking RBM24 revealed multiple defects in excitatory synaptic function and plasticity. Furthermore, RNA sequencing and splicing analysis showed that RBM24 regulates a network of genes related to cognitive function. Deletion of RBM24 disrupted the AS of synapse-associated genes, including GluR2 and Prrt1, the major disease genes involved in cognitive impairment and memory loss, leading to cognitive dysfunction. Together, our results suggest that the regulation of mRNA splicing by RBM24 is a key process involved in maintaining normal synaptic function and provide novel mechanistic insights into the pathogenesis of AD.