HydRA: Deep-learning models for predicting RNA-binding capacity from protein interaction association context and protein sequence.

RNA-binding proteins (RBPs) control RNA metabolism to orchestrate gene expression and, when dysfunctional, underlie human diseases. Proteome-wide discovery efforts predict thousands of RBP candidates, many of which lack canonical RNA-binding domains (RBDs). Here, we present a hybrid ensemble RBP classifier (HydRA), which leverages information from both intermolecular protein interactions and internal protein sequence patterns to predict RNA-binding capacity with unparalleled specificity and sensitivity using support vector machines (SVMs), convolutional neural networks (CNNs), and Transformer-based protein language models. Occlusion mapping by HydRA robustly detects known RBDs and predicts hundreds of uncharacterized RNA-binding associated domains. Enhanced CLIP (eCLIP) for HydRA-predicted RBP candidates reveals transcriptome-wide RNA targets and confirms RNA-binding activity for HydRA-predicted RNA-binding associated domains. HydRA accelerates construction of a comprehensive RBP catalog and expands the diversity of RNA-binding associated domains.

Extraordinary clinical response to ibrutinib in low-grade ovarian cancer guided by organoid drug testing.

Low-grade serous ovarian cancer (LGSOC) typically responds poorly to standard platinum-based chemotherapy and new therapeutic approaches are needed. We describe a remarkable response to targeted therapy in a patient with platinum-resistant, advanced LGSOC who had failed standard-of-care chemotherapy and two surgeries. The patient was in rapid decline and entering hospice care on home intravenous (i.v.) opioid analgesics and a malignant bowel obstruction requiring a G-tube. Genomic analysis of the patient's tumor did not indicate obvious therapeutic options. In contrast, a CLIA-certified drug sensitivity assay of an organoid culture derived from the patient's tumor identified several therapeutic choices, including Bruton's tyrosine kinase (BTK) inhibitor ibrutinib, as well as the EGFR inhibitors afatinib and erlotinib. Following off-label administration of daily ibrutinib as monotherapy, the patient had an exceptional clinical turnaround over the following 65 weeks with normalization of CA-125 levels, resolution of the malignant bowel obstruction, halting of pain medications, and improvement of performance status from ECOG 3 to ECOG 1. After 65 weeks of stable disease, the patient's CA-125 levels began to rise, at which point the patient discontinued ibrutinib and began taking afatinib as monotherapy. The patient's CA-125 levels remained stable for an additional 38 weeks but due to anemia and rising CA-125 levels, the patient switched to erlotinib and is currently being monitored. This case highlights the clinical utility of ex vivo drug testing of patient-derived tumor organoids as a new functional precision medicine approach to identify effective personalized therapies for patients who have failed standard-of-care treatments.

Global analysis of RNA-binding proteins identifies a positive feedback loop between LARP1 and MYC that promotes tumorigenesis.

In addition to genomic alterations, aberrant changes in post-transcriptional regulation can modify gene function and drive cancer development. RNA-binding proteins (RBPs) are a large class of post-transcriptional regulators that have been increasingly implicated in carcinogenesis. By integrating multi-omics data, we identify LARP1 as one of the most upregulated RBPs in colorectal cancer (CRC) and demonstrate its oncogenic properties. We perform LARP1:RNA interactome profiling and unveil a previously unexplored role for LARP1 in targeting the 3'UTR of oncogenes in CRC. Notably, we identify the proto-oncogenic transcription factor MYC as a key LARP1-regulated target. Our data show that LARP1 positively modulates MYC expression by associating with its 3'UTR. In addition, antisense oligonucleotide-mediated blocking of the interaction between LARP1 and the MYC 3'UTR reduces MYC expression and in vitro CRC growth. Furthermore, a systematic analysis of LARP1:protein interactions reveals IGF2BP3 and YBX1 as LARP1-interacting proteins that also regulate MYC expression and CRC development. Finally, we demonstrate that MYC reciprocally modulates LARP1 expression by targeting its enhancer. In summary, our data reveal a critical, previously uncharacterized role of LARP1 in promoting CRC tumorigenesis, validate its direct regulation of the proto-oncogene MYC and delineate a model of the positive feedback loop between MYC and LARP1 that promotes CRC growth and development.

Overriding FUS autoregulation in mice triggers gain-of-toxic dysfunctions in RNA metabolism and autophagy-lysosome axis.

Mutations in coding and non-coding regions of FUS cause amyotrophic lateral sclerosis (ALS). The latter mutations may exert toxicity by increasing FUS accumulation. We show here that broad expression within the nervous system of wild-type or either of two ALS-linked mutants of human FUS in mice produces progressive motor phenotypes accompanied by characteristic ALS-like pathology. FUS levels are autoregulated by a mechanism in which human FUS downregulates endogenous FUS at mRNA and protein levels. Increasing wild-type human FUS expression achieved by saturating this autoregulatory mechanism produces a rapidly progressive phenotype and dose-dependent lethality. Transcriptome analysis reveals mis-regulation of genes that are largely not observed upon FUS reduction. Likely mechanisms for FUS neurotoxicity include autophagy inhibition and defective RNA metabolism. Thus, our results reveal that overriding FUS autoregulation will trigger gain-of-function toxicity via altered autophagy-lysosome pathway and RNA metabolism function, highlighting a role for protein and RNA dyshomeostasis in FUS-mediated toxicity.

Tissue-selective restriction of RNA editing of CaV1.3 by splicing factor SRSF9.

Adenosine DeAminases acting on RNA (ADAR) catalyzes adenosine-to-inosine (A-to-I) conversion within RNA duplex structures. While A-to-I editing is often dynamically regulated in a spatial-temporal manner, the mechanisms underlying its tissue-selective restriction remain elusive. We have previously reported that transcripts of voltage-gated calcium channel CaV1.3 are subject to brain-selective A-to-I RNA editing by ADAR2. Here, we show that editing of CaV1.3 mRNA is dependent on a 40 bp RNA duplex formed between exon 41 and an evolutionarily conserved editing site complementary sequence (ECS) located within the preceding intron. Heterologous expression of a mouse minigene that contained the ECS, intermediate intronic sequence and exon 41 with ADAR2 yielded robust editing. Interestingly, editing of CaV1.3 was potently inhibited by serine/arginine-rich splicing factor 9 (SRSF9). Mechanistically, the inhibitory effect of SRSF9 required direct RNA interaction. Selective down-regulation of SRSF9 in neurons provides a basis for the neuron-specific editing of CaV1.3 transcripts.

Genetic mutations in RNA-binding proteins and their roles in ALS.

Mutations in genes that encode RNA-binding proteins (RBPs) have emerged as critical determinants of neurological diseases, especially motor neuron disorders such as amyotrophic lateral sclerosis (ALS). RBPs are involved in all aspects of RNA processing, controlling the life cycle of RNAs from synthesis to degradation. Hallmark features of RBPs in neuron dysfunction include misregulation of RNA processing, mislocalization of RBPs to the cytoplasm, and abnormal aggregation of RBPs. Much progress has been made in understanding how ALS-associated mutations in RBPs drive pathogenesis. Here, we focus on several key RBPs involved in ALS-TDP-43, HNRNP A2/B1, HNRNP A1, FUS, EWSR1, and TAF15-and review our current understanding of how mutations in these proteins cause disease.

Protein-RNA Networks Regulated by Normal and ALS-Associated Mutant HNRNPA2B1 in the Nervous System.

HnRNPA2B1 encodes an RNA binding protein associated with neurodegeneration. However, its function in the nervous system is unclear. Transcriptome-wide crosslinking and immunoprecipitation in mouse spinal cord discover UAGG motifs enriched within ∼2,500 hnRNP A2/B1 binding sites and an unexpected role for hnRNP A2/B1 in alternative polyadenylation. HnRNP A2/B1 loss results in alternative splicing (AS), including skipping of an exon in amyotrophic lateral sclerosis (ALS)-associated D-amino acid oxidase (DAO) that reduces D-serine metabolism. ALS-associated hnRNP A2/B1 D290V mutant patient fibroblasts and motor neurons differentiated from induced pluripotent stem cells (iPSC-MNs) demonstrate abnormal splicing changes, likely due to increased nuclear-insoluble hnRNP A2/B1. Mutant iPSC-MNs display decreased survival in long-term culture and exhibit hnRNP A2/B1 localization to cytoplasmic granules as well as exacerbated changes in gene expression and splicing upon cellular stress. Our findings provide a cellular resource and reveal RNA networks relevant to neurodegeneration, regulated by normal and mutant hnRNP A2/B1. VIDEO ABSTRACT.

Distinct and shared functions of ALS-associated proteins TDP-43, FUS and TAF15 revealed by multisystem analyses.

The RNA-binding protein (RBP) TAF15 is implicated in amyotrophic lateral sclerosis (ALS). To compare TAF15 function to that of two ALS-associated RBPs, FUS and TDP-43, we integrate CLIP-seq and RNA Bind-N-Seq technologies, and show that TAF15 binds to ∼4,900 RNAs enriched for GGUA motifs in adult mouse brains. TAF15 and FUS exhibit similar binding patterns in introns, are enriched in 3' untranslated regions and alter genes distinct from TDP-43. However, unlike FUS and TDP-43, TAF15 has a minimal role in alternative splicing. In human neural progenitors, TAF15 and FUS affect turnover of their RNA targets. In human stem cell-derived motor neurons, the RNA profile associated with concomitant loss of both TAF15 and FUS resembles that observed in the presence of the ALS-associated mutation FUS R521G, but contrasts with late-stage sporadic ALS patients. Taken together, our findings reveal convergent and divergent roles for FUS, TAF15 and TDP-43 in RNA metabolism.

Genome-wide approaches to dissect the roles of RNA binding proteins in translational control: implications for neurological diseases.

Translational control of messenger RNAs (mRNAs) is a key aspect of neurobiology, defects of which can lead to neurological diseases. In response to stimuli, local translation of mRNAs is activated at synapses to facilitate long-lasting forms of synaptic plasticity, the cellular basis for learning, and memory formation. Translation, as well as all other aspects of RNA metabolism, is controlled in part by RNA binding proteins (RBPs) that directly interact with mRNAs to form mRNA-protein complexes. Disruption of RBP function is becoming widely recognized as a major cause of neurological diseases. Thus understanding the mechanisms that govern the interplay between translation control and RBP regulation in both normal and diseased neurons will provide new opportunities for novel diagnostics and therapeutic intervention. As a means of studying translational control, genome-wide methods are emerging as powerful tools that have already begun to unveil mechanisms that are missed by single-gene studies. Here, we describe the roles of RBPs in translational control, review genome-wide approaches to examine translational control, and discuss how the application of these approaches may provide mechanistic insight into the pathogenic underpinnings of RBPs in neurological diseases.

LIN28 binds messenger RNAs at GGAGA motifs and regulates splicing factor abundance.

LIN28 is a conserved RNA-binding protein implicated in pluripotency, reprogramming, and oncogenesis. It was previously shown to act primarily by blocking let-7 microRNA (miRNA) biogenesis, but here we elucidate distinct roles of LIN28 regulation via its direct messenger RNA (mRNA) targets. Through crosslinking and immunoprecipitation coupled with high-throughput sequencing (CLIP-seq) in human embryonic stem cells and somatic cells expressing exogenous LIN28, we have defined discrete LIN28-binding sites in a quarter of human transcripts. These sites revealed that LIN28 binds to GGAGA sequences enriched within loop structures in mRNAs, reminiscent of its interaction with let-7 miRNA precursors. Among LIN28 mRNA targets, we found evidence for LIN28 autoregulation and also direct but differing effects on the protein abundance of splicing regulators in somatic and pluripotent stem cells. Splicing-sensitive microarrays demonstrated that exogenous LIN28 expression causes widespread downstream alternative splicing changes. These findings identify important regulatory functions of LIN28 via direct mRNA interactions.

Differential regulation of N-Myc and c-Myc synthesis, degradation, and transcriptional activity by the Ras/mitogen-activated protein kinase pathway.

Myc transcription factors are important regulators of proliferation and can promote oncogenesis when deregulated. Deregulated Myc expression in cancers can result from MYC gene amplification and translocation but also from alterations in mitogenic signaling pathways that affect Myc levels through both transcriptional and post-transcription mechanisms. For example, mutations in Ras family GTPase proteins that cause their constitutive activation can increase cellular levels of c-Myc by interfering with its rapid proteasomal degradation. Although enhanced protein stability is generally thought to be applicable to other Myc family members, here we show that c-Myc and its paralog N-Myc respond to oncogenic H-Ras (H-Ras(G12V)) in very different ways. H-Ras(G12V) promotes accumulation of both c-Myc and N-Myc, but although c-Myc accumulation is achieved by enhanced protein stability, N-Myc accumulation is associated with an accelerated rate of translation that overcomes a surprising H-Ras(G12V)-mediated destabilization of N-Myc. We show that H-Ras(G12V)-mediated degradation of N-Myc functions independently of key phosphorylation sites in the highly conserved Myc homology box I region that controls c-Myc protein stability by oncogenic Ras. Finally, we found that N-Myc and c-Myc transcriptional activity is associated with their proteasomal degradation but that N-Myc may be uniquely dependent on Ras-stimulated proteolysis for target gene expression. Taken together, these studies provide mechanistic insight into how oncogenic Ras augments N-Myc levels in cells and suggest that enhanced N-Myc translation and degradation-coupled transactivation may contribute to oncogenesis.

CRELD2: gene mapping, alternate splicing, and comparative genomic identification of the promoter region.

CRELD2 is the second member of the CRELD family of proteins. The only other CRELD family member, encoded by CRELD1, is also known as the AVSD2 gene as mutations in CRELD1 are associated with cardiac atrioventricular septal defects (AVSD). Like CRELD1, CRELD2 is ubiquitously expressed during development and by mature tissues. Recently, a specific CRELD2 isoform (CRELD2beta) was implicated as a regulator of alpha4beta2 nicotinic acetylcholine receptor expression, suggesting that the CRELD family has widely diverse biological roles in both developmental events and subsequent cell function. Here we report additional characterization of CRELD2, which was undertaken to further our understanding of this important family. Mapping of CRELD2 by FISH shows that it maps to 22q13 rather than the GenBank reported locus of 22p13. Comparative genomic analysis of upstream sequences shows a discrete region that is highly conserved among diverse species with hallmark features indicative of a promoter region. Functional analysis demonstrates that this region has promoter activity. Consistent with widespread expression of CRELD2, this region is GC-rich and lacks a TATA box. Overall, the highest levels of CRELD2 expression occur in adult endocrine tissues. However, alternative splicing of CRELD2 is extensive with positive identification of several splice variants expressed by most normal fetal and adult tissues. Confirmed splice variants encode 5 different CRELD2 isoforms that differ significantly in composition indicating that CRELD2 function is varied and as yet poorly understood.