Biology subjects

Ule, J.

Publications and source records attributed to Ule, J..

At least 19 recordsLinked to original sources

bioRxiv2026

A combinatorial pre-mRNP retention factor network couples nuclear speckle architecture to splicing quality control

Cells must prevent incompletely spliced transcripts from entering the cytoplasm, yet how splicing status is recognized and converted into selective nuclear retention remains unclear. Here, we identify a network of RNA-binding proteins that retain pre-mRNPs within nuclear speckles. Proteomic analysis of splicing intermediates accumulated in nuclear speckles identified a set of retention factors including BCLAF1, THRAP3 and SRSF7 that anchor pre-mRNPs to the SON scaffold and are required for pre-mRNP retention in speckles. Distinct RNA-binding patterns and transcript-specific retention factor requirements suggest that multiple mRNP features tune retention combinatorially. Super-resolution imaging revealed that the SON and SRRM2 speckle scaffolds form interdigitated networks with RNA enriched along networks and within interstitial speckle lumens. This architecture remodels in response to pre-mRNP load and is disrupted upon depletion of retention factors. Together, these findings define a pre-mRNP quality-control pathway that links splicing state to nuclear retention and the dynamic organization of nuclear speckles.

cell biology↗

bioRxiv2026

PARNET: A CLIP-SEQ-BASED FOUNDATION MODEL FOR RNA SEQUENCE REPRESENTATION LEARNING

RNA-binding proteins (RBPs) orchestrate a complex combinatorial regulatory "code" that governs RNA splicing, stability, localization, and translation. Learning the relationship between RNA sequences and these processes is a central challenge in genomics. Foundation models, notably RNA language models, have emerged as the dominant approach, learning general-purpose representations from unlabeled sequence at scale. While RNA language models have demonstrated impressive performance across a broad range of downstream tasks, they generally learn from sequence reconstruction objectives alone, lacking direct connections to the regulatory principles that govern RNA function. Here we introduce Parnet, an RNA foundation model trained directly and exclusively on experimental CLIP-seq data. Parnet is a multi-task foundation model trained end-to-end on 223 eCLIP-seq experiments spanning 150 RBPs to predict base-resolution RBP binding profiles directly from RNA sequence. This CLIP-seq pretraining strategy departs fundamentally from the masked-language-modeling paradigm, anchoring learned RNA representations directly in measured protein-RNA interactions rather than sequence statistics. Parnet substantially outperforms its single-task predecessor RBPNet in binding profile and motif recovery, generalizes to unseen cell types and iCLIP data, and recapitulates position-dependent splicing regulation. Frozen Parnet embeddings, without task-specific fine-tuning, match or exceed the performance of both task-specific tools, as well as larger self-supervised RNA and genomic language models across diverse downstream tasks, including RNA biotype classification, lncRNA chromatin localization, translational efficiency, splice-site recognition, intron retention, and non-coding variant effect prediction. Importantly, Parnet remains mechanistically interpretable, tracing predictions back to the specific RBPs and motifs that drive them. These results establish the RBP interactome as a compact, functionally sufficient, and interpretable basis for foundation model pretraining in RNA biology.

bioinformatics↗

bioRxiv2026

CDKL5 phosphorylates neuronal ELAVL proteins to promote mRNA binding, protein synthesis and visual cortex development

Loss-of-function mutations in the X-linked CDKL5 gene lead to a severe neurodevelopmental disorder characterized by early-onset epilepsy, known as CDKL5 Deficiency Disorder (CDD). Despite its clinical significance, the physiological substrates of the serine/threonine kinase CDKL5 and its roles in neuronal development remain poorly understood. To address this, we performed quantitative phosphoproteomics analysis in Cdkl5 knockout (KO) mouse brains, identifying 22 CDKL5 substrates involved in diverse cellular functions. Among these, we focused on the neuronal RNA-binding proteins (nELAVLs) ELAVL2, ELAVL3, and ELAVL4, as these represented the only evolutionarily conserved phosphorylation and are known regulators of neuronal differentiation. Through kinase assays and individual-nucleotide resolution crosslinking and immunoprecipitation (iCLIP), we found that CDKL5 phosphorylates S119/131 in ELAVL2/3/4, promoting their cytoplasmic localization and enhancing their binding to target mRNAs at 3UTRs. Loss of CDKL5 activity in neurons caused reduced new protein synthesis, as measured by puromycin incorporation; this phenotype was rescued by knockdown of the nELAVL inhibitor long non-coding RNA, RNY3, revealing an essential function of CDKL5 in enhancing protein synthesis via nELAVL phosphorylation. To investigate the in vivo functions of nELAVL phosphorylations, we generated Elavl2/3/4 phosphomutant mice and found that collectively nELAVL phosphorylations are required for viability. Proteomic and transcriptomic analyses of Elavl2/3 homozygous phosphomutants, which exhibited sub-viability, revealed compensatory upregulation of ELAVL4 and synaptic proteins. Functionally, in Elavl2/3/4 triple heterozygous mice Neuropixels recordings in the primary visual cortex showed deficits in receptive field properties and orientation tuning, revealing the role of nELAVL phosphorylation for accurate cortical circuit formation. Our study uncovers a crucial role for CDKL5 in regulating nELAVL-mediated protein synthesis and the development of cortical circuits.

neuroscience↗

bioRxiv2026

Targeted 3'-end RNA sequencing uncovers cryptic polyadenylation in Huntington's disease linked to somatic instability and CAG repeat purity

Huntingtons disease (HD) is a progressive neurodegenerative disorder caused by expanded CAG repeats in the first exon of the HTT gene, which encodes for huntingtin (HTT) protein. Full-penetrance is established at 40 repeats, but beyond, somatic repeat instability in the brain and CAG repeat purity modulate disease onset and severity. Previous studies have described that expanded repeats induce the incomplete splicing of HTT intron 1 to express the most pathogenic HTT isoform, known as HTT1a. Yet, the lack of a robust and sensitive method to evaluate HTT RNA-misprocessing has limited our understanding of HTT1a expression in HD pathophysiology. Here we describe a targeted RNA sequencing approach, known as 3-end targeted RNA sequencing or 3TRS, to simultaneously quantify multiple HTT transcripts generated by canonical and cryptic polyadenylation in several HD models. We show that activation of HTT cryptic polyadenylation is highly selective and requires long and uninterrupted CAG repeat expansions. In HD knock-in mice and human postmortem brain, cryptic HTT expression strongly correlates with brain-specific somatic repeat instability, supporting a model where ultralong and unstable CAG repeats drive toxicity by activating HTT RNA-misprocessing. Overall, 3TRS provides a robust framework to investigate HTT1a biogenesis and expression and to evaluate HTT-lowering therapeutic strategies. GRAPHICAL ABSTRACT O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=140 SRC="FIGDIR/small/697463v2_ufig1.gif" ALT="Figure 1"> View larger version (44K): org.highwire.dtl.DTLVardef@168d8e1org.highwire.dtl.DTLVardef@76ba3forg.highwire.dtl.DTLVardef@bdaf93org.highwire.dtl.DTLVardef@151e581_HPS_FORMAT_FIGEXP M_FIG C_FIG

molecular biology↗

bioRxiv2025

Increased HCN1 activity in human excitatory neurons drives excessive network bursting in Dravet syndrome

Dravet syndrome (DS) is a severe childhood epilepsy caused by mutations of the sodium channel NaV1.1. These mutations are thought to compromise the ability of inhibitory interneurons to regulate network activity, leading to seizure events. However, standard treatments to restore inhibition have limited efficacy, suggesting the existence of additional pathological mechanisms. Here, we use hiPSC-derived neuronal networks containing both excitatory and inhibitory neurons to show that excessive bursting activity in DS cultures is driven by excitatory neurons. This rise in bursting frequency is caused by the increased expression of HCN1 "pacemaker" channels in excitatory neurons, and bursting activity can be normalised using a channel blocker. With this work, we propose a new pathophysiological mechanism in DS and identify HCN1 as a novel therapeutic target.

neuroscience↗

bioRxiv2025

Characterisation of LAMP1- and LAMP2A-positive organelles in neurons

LAMP1 and LAMP2A are abundant proteins of late endosomal/lysosomal compartments, which are often used interchangeably to label what is thought to be the same pool of organelles, potentially obscuring their unique physiological roles. Here, we characterised the transport dynamics of LAMP1- and LAMP2A-positive compartments in human iPSC-derived cortical neurons. We found that axonal LAMP1-positive organelles move more slowly in the retrograde direction, pause more frequently, and show a broader velocity distribution in the anterograde direction than LAMP2A-positive vesicles, suggesting they are distinct compartments with differential trafficking behaviour. To explore the molecular mechanism underlying these differences, we characterised with high spatiotemporal precision, the protein interactomes of LAMP1 and LAMP2A-positive compartments through proximity labelling, using full-length LAMP1 or LAMP2A fused to the light-activated biotin ligase LOV-Turbo. We identified and validated the endosomal protein, ZFYVE16, as a novel member of LAMP1 and LAMP2A interactomes. We suggest that LAMP2A-positive organelles represent a subset of LAMP1-positive compartments, which are surprisingly enriched in synaptic vesicle proteins. Summary statementLAMP1- and LAMP2A-positive organelles have different axonal transport dynamics and form distinct organelle pools characterised by specific protein compositions.

neuroscience↗

bioRxiv2025

Defining the ordered pathway for ZAP-mediated RNA decay

Zinc-finger Antiviral Protein (ZAP)-mediated RNA decay (ZMD) restricts replication of viruses containing CpG dinucleotide clusters. However, why ZAP isoforms differ in antiviral activity and how they recruit cofactors to mediate RNA decay is unclear. Therefore, we determined the ordered events of the ZMD pathway. The long ZAP isoform preferentially binds viral RNA and has distinct binding motifs compared to the short isoform. The endoribonuclease KHNYN then cleaves viral RNA at positions of ZAP binding. The 5 cleavage fragment undergoes TUT4/TUT7-mediated 3 uridylation and degradation by DIS3L2. The 3 cleavage fragment is degraded by XRN1. ZAP and TRIM25 interact with KHNYN, TUT7, DIS3L2 and XRN1 in a RNase-resistant manner. Viral infection promotes the interaction between TRIM25 with these enzymes, leading to viral RNA decay while also decreasing the abundance of cellular transcripts. Overall, the long isoform of ZAP recruits key enzymes to assemble an RNA decay complex on viral RNA.

immunology↗

bioRxiv2025

PRMT activity promotes global 3' UTR shortening in proliferating cells

Protein methyltransferase (PRMT)-catalysed arginine methylation is a widespread post-translational modification that regulates numerous RNA-binding proteins and frequently becomes dysregulated in cancer. While PRMT inhibitors have shown promise as an anti-cancer strategy, greater understanding of the downstream pathways linking arginine methylation to tumour-promoting phenotypes is needed to improve patient stratification and develop more effective therapeutic approaches. Here, we reveal arginine methylation as a critical regulator of alternative polyadenylation (APA) patterns that are fundamental to tumour progression. 3' RNA-sequencing assays uncover a rapid and global shift toward longer 3' UTR isoforms upon dual (symmetric and asymmetric) methylation (DMAi), impacting a broad range of cellular proliferation and signalling genes. Arginine methylation is required for sustaining proximal poly(A) site usage under high proliferative demand, as DMAi treatment blocks use of such sites in activated T cells, various cancer cell lines and patient-derived lung organoids. DMAi also counteracts the 3' UTR shortening caused by reduced CFIM25 expression, which normally promotes oncogenic isoforms. DMAi treatment affects APA in many of the same mRNAs as impaired cleavage and polyadenylation activity, and these mRNAs contain characteristic signatures such as high GC-content and long 3 UTRs. This systematic impact of PRMT activity on APA regulation broadens the potential utility of PRMT inhibitors as therapeutic agents for both cancer and immune-related diseases.

molecular biology↗

bioRxiv2025

CAR-SPLASH identifies nascent pre-mRNA structures implicated in kinetic coupling and alternative splicing

Pre-mRNA splicing is kinetically coupled to transcription as shown by the widespread effects of transcription speed on alternative splicing outcomes. The molecular basis for such kinetic coupling is incompletely understood, but one potential mechanism is through elongation rate dependent alternative folding pathways of the nascent pre-mRNA. To search for RNA structures in nascent pre-mRNA, we modified SPLASH (1) (Sequencing of Psoralen Crosslinked, Ligated And Selected Hybrids) for use with Chromatin Associated RNA. We applied this new method called CAR-SPLASH to cells expressing WT and slow mutant RNA polymerase II and identified > 3000 intramolecular RNA structures of which > 400 are proximal to splice sites. Antisense oligonucleotide (ASO) disruption of several such structures that sequester splice sites has a major impact on alternative splicing outcomes, even though the ASOs do not directly disrupt splice sites. ASO disruption of novel regulatory elements that we designate "RNA kinetic switches", modified alternative splicing of NISCH Exon 18, GAK Exon 7 and MEGF8 Exon 14 in a way that depends on the rate of transcription elongation. We propose that these switches mediate kinetic coupling via effects of transcription speed on folding of nascent RNA structures that modulate alternative splicing, and that many nascent RNA structures can thereby serve as new targets for splice modifying ASOs. Significance StatementAlternative splicing (AS) of mRNAs is a major regulator of gene expression that is frequently disrupted is disease. AS is affected by the speed of RNA polymerase II that synthesizes mRNA precursors, but how such kinetic coupling works is not well understood. We developed a method, CAR-SPLASH, to capture RNA structures in nascent RNA chains by cross-linking with psoralen. CAR-SPLASH identified novel structural regulatory elements we call "RNA kinetic switches" that control AS in a way that depends on the speed of transcription. RNA kinetic switches are new targets for splice modifying antisense oligonucleotides that could have therapeutic value.

molecular biology↗

bioRxiv2024

Short RNA chaperones promote aggregation-resistant TDP-43 conformers to mitigate neurodegeneration

Aberrant aggregation of the prion-like, RNA-binding protein TDP-43 underlies several debilitating neurodegenerative proteinopathies, including amyotrophic lateral sclerosis (ALS). Here, we define how short, specific RNAs antagonize TDP-43 aggregation. Short, specific RNAs engage and stabilize the TDP-43 RNA-recognition motifs, which allosterically destabilizes a conserved helical region in the prion-like domain, thereby promoting aggregation-resistant conformers. By mining sequence space, we uncover short RNAs with enhanced activity against TDP-43 and diverse disease-linked variants. The solubilizing activity of enhanced short RNA chaperones corrects aberrant TDP-43 phenotypes in optogenetic models and ALS patient-derived neurons. Remarkably, an enhanced short RNA chaperone mitigates TDP-43 proteinopathy and neurodegeneration in mice. Our studies reveal mechanisms of short RNA chaperones and pave the way for the development of short RNA therapeutics for fatal TDP-43 proteinopathies.

biochemistry↗

bioRxiv2024

Integrative analysis of higher-order transcriptome organisation and RNA condensation principles

Complex RNA-protein networks play a pivotal role in the formation of many types of biomolecular condensates. How intrinsic RNA features contribute to condensate formation however remains unclear. Here, we integrate tailored transcriptomics assays to identify a distinct class of developmental condensation-prone RNAs termed smOOPs (semi-extractable, orthogonal organic phase separation-enriched RNAs). These transcripts are localised to larger intracellular foci, form denser RNA-RNA interaction subnetworks than expected and are heavily bound by RNA binding proteins (RBPs). Using an explainable deep learning framework, we reveal that smOOPs harbor characteristic sequence composition with lower sequence complexity, increased intramolecular folding and specific RBP binding patterns. Intriguingly, these RNAs encode proteins bearing extensive intrinsically disordered regions and are markedly predicted to be involved in biomolecular condensates, indicating an interplay between RNA- and protein-based features in phase separation. This work advances our understanding of condensation-prone RNAs and provides a versatile resource to further investigate RNA-driven condensation principles.

molecular biology↗

bioRxiv2023

Creation of de novo cryptic splicing for ALS/FTD precision medicine

A system enabling the expression of therapeutic proteins specifically in diseased cells would be transformative, providing greatly increased safety and the possibility of pre-emptive treatment. Here we describe "TDP-REG", a precision medicine approach primarily for amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD), which exploits the cryptic splicing events that occur in cells with TDP-43 loss-of-function (TDP-LOF) in order to drive expression specifically in diseased cells. In addition to modifying existing cryptic exons for this purpose, we develop a deep-learning-powered algorithm for generating customisable cryptic splicing events, which can be embedded within virtually any coding sequence. By placing part of a coding sequence within a novel cryptic exon, we tightly couple protein expression to TDP-LOF. Protein expression is activated by TDP-LOF in vitro and in vivo, including TDP-LOF induced by cytoplasmic TDP-43 aggregation. In addition to generating a variety of fluorescent and luminescent reporters, we use this system to perform TDP-LOF-dependent genomic prime editing to ablate the UNC13A cryptic donor splice site. Furthermore, we design a panel of tightly gated, autoregulating vectors encoding a TDP-43/Raver1 fusion protein, which rescue key pathological cryptic splicing events. In summary, we combine deep-learning and rational design to create sophisticated splicing sensors, resulting in a platform that provides far safer therapeutics for neurodegeneration, potentially even enabling preemptive treatment of at-risk individuals. One-Sentence SummaryWe engineer TDP-43-regulated cryptic exons, enabling exceptionally precise activation of gene therapies in diseased neurons.

molecular biology↗

bioRxiv2023

The FXR1 network acts as signaling scaffold for actomyosin remodeling

It is currently not known whether mRNAs fulfill structural roles in the cytoplasm. Here, we report the FXR1 network, an mRNA-protein (mRNP) network present throughout the cytoplasm, formed by FXR1-mediated packaging of exceptionally long mRNAs. These mRNAs serve as underlying condensate scaffold and concentrate FXR1 molecules. The FXR1 network contains multiple protein binding sites and functions as a signaling scaffold for interacting proteins. We show that it is necessary for RhoA signaling-induced actomyosin reorganization to provide spatial proximity between kinases and their substrates. Point mutations in FXR1, found in its homolog FMR1, where they cause Fragile X syndrome, disrupt the network. FXR1 network disruption prevents actomyosin remodeling--an essential and ubiquitous process for the regulation of cell shape, migration, and synaptic function. These findings uncover a structural role for cytoplasmic mRNA and show how the FXR1 RNA-binding protein as part of the FXR1 network acts as organizer of signaling reactions.

cell biology↗

bioRxiv2023

Flow: a web platform and open database to analyse, store, curate and share bioinformatics data at scale

Ever-increasing volumes of sequencing data offer potential for large-scale meta-analyses to address significant biological questions. However, challenges such as insufficient data processing information, data quality concerns, and issues related to accessibility and curation often present obstacles. Additionally, most experimental biologists lack the time and expertise needed to independently analyse, manage and share their own data. To overcome these hurdles, we present Flow, a web-based platform that links bioinformatic analysis and database solutions with a user-friendly interface and web API. Flow currently accommodates a range of genomics methods and further DSL2-compliant Nextflow pipelines can be added via a simple JSON schema file. Deployable on local systems or cloud services, an instance is freely accessible to academic researchers at https://flow.bio.

bioinformatics↗

bioRxiv2023

Mutual homeostasis of charged proteins

Protein dosage is regulated to maintain cellular homeostasis and health. The dosage of proteins containing disordered low complexity domains (LCDs) must be particularly well-controlled to prevent aberrant disease, yet no mechanism to maintain homeostasis has been identified1, 2. Here we report a mutual homeostatic mechanism that controls the concentration of such proteins, termed interstasis, in which proteins with similar LCDs co-regulate their combined dosage through collective negative feedback. We focused on the mechanism that exploits the fundamental multivalency of GA-rich RNA regions that encode charged LCDs, including those with arginine-enriched mixed charge domains (R-MCDs). Modest variations in the abundance of an R-MCD protein change the properties of nuclear speckles, a protein-RNA condensate, selectively trapping multivalent GA-rich mRNAs to promote their nuclear retention. This interstasis depends on conserved codon biases, shared by amniotes, which enhance the multivalency of GA-rich regions encoding charged LCDs. The threshold of interstasis is modulated by CLK kinases, which affect the nuclear speckle localisation of proteins such as TRA2B, key binder of GA-rich RNAs. Notably, many classes of LCDs are encoded by RNA regions containing multivalency-enhancing codon biases, each preferentially bound by specific proteins, suggesting that interstasis might co-regulate many classes of functionally related LCD-containing proteins through dose-sensitivity of various types of protein-RNA condensates.

molecular biology↗

bioRxiv2023

Interrogation of RNA-protein interaction dynamics in bacterial growth

Characterising RNA-protein interaction dynamics is fundamental to understand how bacteria respond to their environment. In this study, we have analysed the dynamics of 91% of the Escherichia coli expressed proteome and the RNA-interaction properties of 271 RNA-binding proteins (RBPs) at different growth phases. We find that 68% of RBPs differentially bind RNA across growth phases and characterise 17 previously unannotated proteins as bacterial RBPs including YfiF, a ncRNA-binding protein. While these new RBPs are mostly present in Proteobacteria, two of them have human orthologs in the form of mitochondrial proteins associated with rare metabolic disorders. Moreover, we reveal novel RBP functions for proteins such as the chaperone HtpG, a new stationary phase tRNA-binding protein. Altogether, we provide the first dynamic RBPome of a bacterium, showcasing how this approach can reveal the function of uncharacterised proteins, and identify critical RNA-protein interactions for cell growth which could inform new antimicrobial therapies.

microbiology↗

bioRxiv2023

Abundant capped RNAs are derived from mRNA cleavage at 3'UTR G-Quadruplexes

The 3 untranslated region (3UTR) plays a crucial role in determining mRNA stability, localisation, translation and degradation. Cap analysis gene expression (CAGE), a method for the detection of capped 5 ends of mRNAs, additionally reveals a large number of apparently 5 capped RNAs derived from 3UTRs. Here we provide the first direct evidence that these 3UTR-derived RNAs are indeed capped and often more abundant than the corresponding full-length mRNAs. By using a combination of AGO2 enhanced individual nucleotide resolution UV crosslinking and immunoprecipitation (eiCLIP) and CAGE following siRNA knockdowns, we find that these 3UTR-derived RNAs likely originate from AGO2-mediated cleavage, and most often occur at locations with potential to form RNA-G-quadruplexes and are enriched by RNA-binding protein UPF1. High-resolution imaging and long-read sequencing analysis validates several 3UTR-derived RNAs, demonstrates their abundance and shows that they tend not to co-localise with the parental mRNAs. We also find that production of 3UTR-derived RNA could explain the previously reported role of a 3UTR G-quadruplex in regulating the production of APP protein. Taken together, we provide new insights into the origin and abundance of 3UTR-derived RNAs, show the utility of CAGE-seq for their quantitative detection, and provide a rich dataset for exploring new biology of a poorly understood new class of RNAs. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=73 SRC="FIGDIR/small/538568v3_ufig1.gif" ALT="Figure 1"> View larger version (15K): org.highwire.dtl.DTLVardef@3f2ca7org.highwire.dtl.DTLVardef@18ca1b6org.highwire.dtl.DTLVardef@1ccdd8dorg.highwire.dtl.DTLVardef@e58938_HPS_FORMAT_FIGEXP M_FIG C_FIG

molecular biology↗

bioRxiv2023

ORC1 binds to cis-transcribed RNAs for efficient activation of replication origins

Cells must coordinate the activation of thousands of replication origins dispersed throughout their genome. Active transcription is known to favor the formation of mammalian origins, although the role that RNA plays in this process remains unclear. We show that the ORC1 subunit of the human Origin Recognition Complex interacts with RNAs transcribed from genes with origins in their transcription start sites (TSSs), displaying a positive correlation between RNA binding and origin activity. RNA depletion, or the use of ORC1 RNA-binding mutant, result in inefficient activation of proximal origins, linked to impaired ORC1 chromatin release. ORC1 RNA binding activity resides in its intrinsically disordered region, involved in intra- and inter-molecular interactions, regulation by phosphorylation, and phase-separation. We show that RNA binding favors ORC1 chromatin release, by regulating its phosphorylation and subsequent degradation. We propose that fluctuating concentrations of RNA during the cell cycle may play a sequential role in controlling origins through interaction with this flexible region of ORC1. Our results unveil a novel non-coding function of RNA as a dynamic component of the chromatin, orchestrating the activation of replication origins. One sentence summaryThe human origin recognition complex subunit 1 ORC1, binds to RNAs transcribed from genes with origins of replication at the TSS, which is required for optimal origin activation.

molecular biology↗