Biology subjects

Bae, B.

Publications and source records attributed to Bae, B..

7 recordsLinked to original sources

bioRxiv2025

Allelic chromatin structure is a pervasive feature of imprinted domains and functions cooperatively with cis-acting long non-coding RNAs

Genomic imprinting is an epigenetic phenomenon in which genes exhibit restricted or biased expression from one allele according to parental origin. Imprinted gene expression plays crucial roles in embryonic growth and brain development. Higher-order chromatin structure has long been associated with gene regulation, particularly in the context of spatial enhancer-promoter interactions. Because imprinted genes exhibit allele-biased expression, it is compelling to ask whether differences in the three-dimensional organization of maternal and paternal genomes underlie this regulation. Using a capture Hi-C approach, we identified parental allele-specific higher-order chromatin structures across multiple imprinted domains in the mouse brain. These allele-specific structural features largely stem from annotated imprinting control regions (ICRs), concomitant with allele-specific binding of CTCF. The transcription start sites of active and inactive alleles of imprinted genes engage in distinct distal chromatin interactions that differ in number and in the epigenetic states of their contact regions. CRISPR interference (CRISPRi) screening identified a distal cis-regulatory element that modulates imprinted expression at the Mest-Copg2 locus in neurons, with its regulatory activity closely linked to allele-specific chromatin interactions. Further investigation revealed that both a cis-acting long non-coding RNA and allele-specific enhancer-promoter architecture modulates Mest-Copg2 imprinted expression in neurons. Together, this study highlights the interplay between chromatin structure and regulatory landscapes that modulate allele-specific expression of imprinted genes.

genetics↗

bioRxiv2024

Physics-inspired accuracy estimator for model-docked ligand complexes

Model docking, which refers to ligand docking into the protein model structures, is becoming a promising avenue in drug discovery with the advances in artificial intelligence (AI)-based protein structure prediction. However, a significant challenge remains; even when sampling was successful in model docking, typical docking score functions fail to identify correct solutions for two-thirds of them. This discrepancy between scoring and sampling majorly arises because these scoring functions poorly tolerate minor structural inaccuracies. In this work, we propose a deep neural network named DENOISer to address the scoring challenge in model-docking scenario. In the network, ligand poses are ranked by the consensus score of two independent sub-networks: the Native-likeness prediction and the Binding energy prediction networks. Both networks incorporate physical knowledge as inductive bias in order to enhance pose discrimination power while ensuring tolerance to small interfacial structural noises. Combined with Rosetta GALigandDock sampling, DENOISer outperformed existing docking tools on the PoseBusters model-docking benchmark set as well as on a broad cross-docking benchmark set. Further analyses reveal the physics-based components and the consensus ranking approach are the two most crucial factors contributing to its ranking success. We expect DENOISer may assist future drug discovery endeavors by providing more accurate structural models for protein-ligand complexes. The network is freely available at https://github.com/HParklab/DENOISer.

bioinformatics↗

bioRxiv2024

Mapping snoRNA-target RNA interactions in an RNA binding protein-dependent manner with chimeric eCLIP

Small nucleolar RNAs (snoRNAs) are non-coding RNAs that function in ribosome and spliceosome biogenesis, primarily by guiding modifying enzymes to specific sites on ribosomal RNA (rRNA) and spliceosomal RNA (snRNA). However, many orphan snoRNAs remain uncharacterized, with unidentified or unvalidated targets, and studies on additional snoRNA-associated proteins are limited. We adapted an enhanced chimeric eCLIP approach to comprehensively profile snoRNA-target RNA interactions using both core and accessory snoRNA binding proteins as baits. Using core snoRNA binding proteins, we confirmed most annotated snoRNA-rRNA and snoRNA-snRNA interactions in mouse and human cell lines and called novel, high-confidence interactions for orphan snoRNAs. While some of these interactions result in chemical modification, others may have modification-independent functions. We then showed that snoRNA ribonucleoprotein complexes containing certain accessory proteins, like WDR43 and NOLC1, enriched for specific subsets of snoRNA-target RNA interactions with distinct roles in ribosome and spliceosome biogenesis. Notably, we discovered that SNORD89 guides 2-O-methylation at two neighboring sites in U2 snRNA that are important for activating splicing, but also appear to ensure imperfect splicing for a subset of near-constitutive exons. Thus, chimeric eCLIP of snoRNA-associating proteins enables a comprehensive framework for studying snoRNA-target interactions in an RNA binding protein-dependent manner, revealing novel interactions and regulatory roles in RNA biogenesis.

genomics↗

bioRxiv2023

Allelic chromatin structure primes imprinted expression of Kcnk9 during neurogenesis

Differences in chromatin state inherited from the parental gametes influence the regulation of maternal and paternal alleles in offspring. This phenomenon, known as genomic imprinting, results in genes preferentially transcribed from one parental allele. While local epigenetic factors such as DNA methylation are known to be important for the establishment of imprinted gene expression, less is known about the mechanisms by which differentially methylated regions (DMRs) lead to differences in allelic expression across broad stretches of chromatin. Allele-specific higher-order chromatin structure has been observed at multiple imprinted loci, consistent with the observation of allelic binding of the chromatin-organizing factor CTCF at multiple DMRs. However, whether allelic chromatin structure impacts allelic gene expression is not known for most imprinted loci. Here we characterize the mechanisms underlying brain-specific imprinted expression of the Peg13-Kcnk9 locus, an imprinted region associated with intellectual disability. We performed region capture Hi-C on mouse brain from reciprocal hybrid crosses and found imprinted higher-order chromatin structure caused by the allelic binding of CTCF to the Peg13 DMR. Using an in vitro neuron differentiation system, we show that on the maternal allele enhancer-promoter contacts formed early in development prime the brain-specific potassium leak channel Kcnk9 for maternal expression prior to neurogenesis. In contrast, these enhancer-promoter contacts are blocked by CTCF on the paternal allele, preventing paternal Kcnk9 activation. This work provides a high-resolution map of imprinted chromatin structure and demonstrates that chromatin state established in early development can promote imprinted expression upon differentiation.

molecular biology↗

bioRxiv2023

Coordination of Alternative Splicing and Alternative Polyadenylation revealed by Targeted Long Read Sequencing

Nervous system development is associated with extensive regulation of alternative splicing (AS) and alternative polyadenylation (APA). AS and APA have been extensively studied in isolation, but little is known about how these processes are coordinated. Here, the coordination of cassette exon (CE) splicing and APA in Drosophila was investigated using a targeted long-read sequencing approach we call Pull-a-Long-Seq (PL-Seq). This cost-effective method uses cDNA pulldown and Nanopore sequencing combined with an analysis pipeline to resolve the connectivity of alternative exons to alternative 3 ends. Using PL-Seq, we identified genes that exhibit significant differences in CE splicing depending on connectivity to short versus long 3UTRs. Genomic long 3UTR deletion was found to alter upstream CE splicing in short 3UTR isoforms and ELAV loss differentially affected CE splicing depending on connectivity to alternative 3UTRs. This work highlights the importance of considering connectivity to alternative 3UTRs when monitoring AS events.

molecular biology↗

bioRxiv2022

Tail-tape-fused virion and non-virion RNA polymerases of a thermophilic virus with an extremely long tail

Thermus thermophilus bacteriophage P23-45 encodes a giant 5,002-residue tail tape measure protein (TMP)1 that defines the length of its extraordinarily long 800 nm tail2,3. We found that the N-terminal portion of P23-45 TMP is an unusual RNA polymerase (RNAP) homologous to cellular and viral two-barrel RNAPs. The TMP-fused virion RNAP transcribes pre-early phage genes, including a gene that encodes another, non-virion RNAP, that transcribes early and some middle phage genes. We determined the crystal structures of both P23-45 RNAPs. The non-virion RNAP has a crab claw-like architecture similar to previously reported two-barrel RNAPs. The virion RNAP adopts a unique flat structure without a clamp, which likely reflects the requirement for its extrusion through the narrow channel in the phage tail for delivery into the cell. Structure and sequence comparisons of the P23-45 RNAPs with other phage and cellular RNAPs suggest that, despite the extensive functional differences, the two P23-45 RNAPs originate from an ancient gene duplication in an ancestral phage. Our findings demonstrate remarkable adaptability of two-barrel RNAPs that can be attained within a single virus species.

molecular biology↗

bioRxiv2019

Precise removal of Calm1 long 3′ UTR isoform by CRISPR-Cas9 genome editing impairs dorsal root ganglion development in mice.

Most mammalian genes are subject to Alternative cleavage and PolyAdenylation (APA), often resulting in alternative length 3' UTR isoforms. Thousands of extended or long 3' UTR variants are preferentially expressed in neuron-enriched tissues of metazoans. However, the in vivo functions of these long 3' UTR isoforms are largely unknown. Calmodulin 1 (Calm1) is a key integrator of calcium signaling that is required for correct neural development. Calm1 generates short (Calm1-S) and long 3' UTR (Calm1-L) mRNA isoforms via APA. We found Calm1-S to be broadly expressed across mouse tissues, whereas Calm1-L expression was largely restricted to neural tissues, including the dorsal root ganglion (DRG). Using CRISPR-Cas9 genome editing, a series of mouse deletion lines were generated that successfully eliminated expression of Calm1-L while maintaining expression of Calm1-S. One of these lines, Calm1{Delta}3' UTR, carried a 163 bp deletion surrounding the distal polyA site. Examination of Calm1{Delta}3' UTR embryos revealed disrupted development of the DRG. In Calm1{Delta}3' UTR DRG explant cultures undergoing axon outgrowth, we observed a dramatic increase in axon fasciculation. These results demonstrate a physiological role for Calm1-L in DRG development, and more generally, establish a genome-editing strategy to study in vivo functions of long 3' UTR isoforms. Author SummaryMore than half of all human genes generate alternative mRNA isoforms which differ in the length of their 3 Untranslated regions (3 UTRs). Through a process called Alternative Cleavage and Polyadenylation thousands of broadly expressed genes preferentially express long 3 UTR variants in brain tissues whereas their short 3 UTR counterparts are more broadly expressed. A challenge to study the functions of these transcripts has been to generate loss of function mutant animals that lack a long 3 UTR isoform but maintain expression of the corresponding short 3 UTR isoform. Here, we used the precise, rapid, and efficient approach of CRISPR genome-editing to generate long 3 UTR mutant mice. These mice, which do not express the long 3 UTR of the Calmodulin 1 (Calm1) gene, exhibit impairment in the development of sensory neurons, including increased fasciculation of axons and aberrant cell body migration. This finding is important because it provides conclusive genetic evidence for a neural function of a long 3 UTR isoform in an animal. The CRISPR genome-editing approach used here can be applied to the study of neuron-enriched long 3 UTR isoforms, which number in the thousands and have largely unexplored functions.

genetics↗