Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Review
. 2024 Sep;25(9):683-700.
doi: 10.1038/s41580-024-00739-7. Epub 2024 May 21.

Splicing regulation through biomolecular condensates and membraneless organelles

Affiliations
Review

Splicing regulation through biomolecular condensates and membraneless organelles

Jimena Giudice et al. Nat Rev Mol Cell Biol. 2024 Sep.

Abstract

Biomolecular condensates, sometimes also known as membraneless organelles (MLOs), can form through weak multivalent intermolecular interactions of proteins and nucleic acids, a process often associated with liquid-liquid phase separation. Biomolecular condensates are emerging as sites and regulatory platforms of vital cellular functions, including transcription and RNA processing. In the first part of this Review, we comprehensively discuss how alternative splicing regulates the formation and properties of condensates, and conversely the roles of biomolecular condensates in splicing regulation. In the second part, we focus on the spatial connection between splicing regulation and nuclear MLOs such as transcriptional condensates, splicing condensates and nuclear speckles. We then discuss key studies showing how splicing regulation through biomolecular condensates is implicated in human pathologies such as neurodegenerative diseases, different types of cancer, developmental disorders and cardiomyopathies, and conclude with a discussion of outstanding questions pertaining to the roles of condensates and MLOs in splicing regulation and how to experimentally study them.

PubMed Disclaimer

Conflict of interest statement

Competing interests

The authors declare no competing interests.

Figures

Figure 1.
Figure 1.. Fundamentals of splicing regulation.
a. The spliceosome (not shown) recognizes and binds to splicing consensus sequences at the 5’ and 3’ ends of the introns, known respectively as 5’ splice site (5’ss) and 3’ss, thereby allowing the removal of the intron and the joining of the adjacent exons. When the sequences of the splice sites are very similar to a consensus sequences, the splice sites are usually considered “strong” and constitutive splicing occurs (left). When the sequences of the splice sites differ from the consensus, the splice sites are often considered “weak” and can be sub-optimally used, resulting in alternative splicing and the formation of different mRNA transcripts (right). b. In some conditions, “strong” splice sites undergo alternative splicing and “weak” splice sites are constitutively used. This is possible because, in addition to the strength of the splice sites, trans-acting factors (RNA-binding proteins (RBPs) and splicing factors) that bind to cis-regulatory sequences within the pre-mRNA influence splicing outcomes: binding to exonic splicing enhancer (ESE) or intronic splicing enhancer (ISE) sequences favors exon inclusion, whereas binding to exonic splicing silencer (ESS) and intronic splicing silencer (ISS) sequences favors exon skipping.
Figure 2.
Figure 2.. Splicing regulation through phase separation and its physiological relevance.
a. Molecular properties of RNA-binding proteins (RBPs) that influence their condensation propensity include post-translational modifications (PTMs), which can alter their charge, size and hydrophobicity; RNA-binding domains, which mediate their interaction with RNA; and intrinsically disordered regions (IDRs), which mediate protein–protein associations. Note that IDRs are enriched in PTMs owing to their high accessibility and that these PTMs can in turn regulate condensation. b. RNA-binding fox-1 homolog 1 (RBFOX1) interacts with the large assembly of splicing regulators (LASR) complex through its C-terminus and together they assemble into organelle-like, high-order protein complexes. A low complexity (LC) sequence contains a Tyr (Y)-rich domain that mediates RBFOX–LASR high-order assembly and is required for splicing activation by RBFOX1 (Ref.). Mutations (red asterisk) in these Tyr residues block the capacity of RBFOX1 for high-order assembly, but not its interaction with LASR; however, these mutations do prevent proper RBFOX1 functions in splicing regulation. c. A-kinase anchoring protein 95 kDa (AKAP95) interacts with the MLL2 complex, RNA polymerase II (Pol II) and different splicing factors; its two zinc-finger domains (ZF) are involved in binding to pre-mRNA introns and splicing regulation. An IDR allows AKAP95 to form dynamic and liquid-like droplets that function in splicing regulation. Both a Tyr-to-Ser (Y-to-S) mutation that disrupts AKAP95 condensation and a Tyr-to- Phe (Y-to-F) mutation that enhances AKAP95 phase separation but reduces the dynamicity of the condensates significantly impairs AKAP95 activity in regulating splicing and promoting cancer. DDX5, probable ATP-dependent RNA helicase; HNRNPs, heterogeneous nuclear ribonucleoproteins.
Figure 3.
Figure 3.. Regulation of condensate formation and properties through alternative splicing.
a. Most heterogeneous nuclear ribonucleoprotein A (HNRNPA) and HNRNPD proteins contain Gly-Tyr (GY)-rich intrinsically disordered regions (IDRs) that are encoded by exons that are constitutively spliced in non-mammals, but alternatively spliced in mammals. In mammals, those exons can be included or skipped and, in this way, fine-tune the propensity of these RNA-binding proteins to form biomolecular condensates and expand their capacity to regulate gene expression. b. HNRNPD-like (HNRNPDL) has three splice isoforms, which differ in the presence or absence of an Arg (R)-rich IDR at the N-terminus and a Tyr (Y)-rich IDR at the C-terminus. HNRNPDL1 includes both IDRs and can undergo phase separation; HNRNPDL2 only includes the Y-rich IDR and forms amyloid aggregates; and HNRNPDL3 lacks both IDRs and is soluble–. c. TAR DNA-binding protein 43 (TDP43) is encoded by the TARDBP gene. The full-length TDP43 mRNA isoform includes the entire exon 6 which contains a stop codon. The full length protein (TDP43-fl) contains a Gly (G)-rich low complexity (LC) domain at the C-terminus, which facilitates nuclear aggregation and mediates its functions in splicing regulation. Usage of the same 3’ splice site (3’ss) than the TDP43-fl but two alternative 5’ splice sites (5’ss) within exon 6 before the stop codon gives rise to two short TDP43 isoforms (TDP43-s1 and TDP34-s2) that include only a small portion of exon 6 and the entire exon 7 within their coding sequence. Therefore, the C-termini of TDP43-s1 and TDP34-s2 lack the Gly-rich low complexity domain present in TDP43-fl, but includes a putative nuclear export signal (NES) not found in TDP43-fl. Thus, the short TDP43 variants aggregate in the cytoplasm, where they sequester TDP43-fl, and lack splicing regulation capacity. NLS, nuclear localization signal.
Figure 3.
Figure 3.. Regulation of condensate formation and properties through alternative splicing.
a. Most heterogeneous nuclear ribonucleoprotein A (HNRNPA) and HNRNPD proteins contain Gly-Tyr (GY)-rich intrinsically disordered regions (IDRs) that are encoded by exons that are constitutively spliced in non-mammals, but alternatively spliced in mammals. In mammals, those exons can be included or skipped and, in this way, fine-tune the propensity of these RNA-binding proteins to form biomolecular condensates and expand their capacity to regulate gene expression. b. HNRNPD-like (HNRNPDL) has three splice isoforms, which differ in the presence or absence of an Arg (R)-rich IDR at the N-terminus and a Tyr (Y)-rich IDR at the C-terminus. HNRNPDL1 includes both IDRs and can undergo phase separation; HNRNPDL2 only includes the Y-rich IDR and forms amyloid aggregates; and HNRNPDL3 lacks both IDRs and is soluble–. c. TAR DNA-binding protein 43 (TDP43) is encoded by the TARDBP gene. The full-length TDP43 mRNA isoform includes the entire exon 6 which contains a stop codon. The full length protein (TDP43-fl) contains a Gly (G)-rich low complexity (LC) domain at the C-terminus, which facilitates nuclear aggregation and mediates its functions in splicing regulation. Usage of the same 3’ splice site (3’ss) than the TDP43-fl but two alternative 5’ splice sites (5’ss) within exon 6 before the stop codon gives rise to two short TDP43 isoforms (TDP43-s1 and TDP34-s2) that include only a small portion of exon 6 and the entire exon 7 within their coding sequence. Therefore, the C-termini of TDP43-s1 and TDP34-s2 lack the Gly-rich low complexity domain present in TDP43-fl, but includes a putative nuclear export signal (NES) not found in TDP43-fl. Thus, the short TDP43 variants aggregate in the cytoplasm, where they sequester TDP43-fl, and lack splicing regulation capacity. NLS, nuclear localization signal.
Figure 4.
Figure 4.. Spatial regulation of splicing through nuclear membraneless organelles.
Genomic regions with transcribed genes concentrated in 3D space are in close proximity to nuclear speckles, which support concentration of splicing factors, and the pre-mRNAs of these genes are spliced efficiently; genes that are farther away from speckles (right-hand side of the figure) are spliced less efficiently owing to the reduced local concentration of splicing factors,,,,. The “interface splicing” model was recently proposed for serine-arginine (SR) rich splicing factor 1 (SRSF1) and heterogeneous nuclear ribonucleoprotein A1 (HNRNPA1). The outer layer of the speckle is enriched in HNRNPA1, which binds to introns of pre-mRNAs, whereas the interior layer of the speckle is enriched in SRSF1, which binds to exons of the pre-mRNAs,. This distribution facilitates the localization of the exon-intron boundary at the interface of the two layers, where active spliceosomes are located and execute the splicing reaction,. The model proposes a certain order for the regulation of splicing: first, the RNA is positioned at the nuclear speckle interface based on the recognition of binding motifs by SRSF1 and HNRNPA1; then, the spliceosome assembles; and finally, the splicing reaction takes (inset). Some retained-intron RNAs (RI-RNAs) in nuclear speckles can be efficiently spliced post-transcriptionally, whereas RI-RNAs far from speckles (top right corner of the figure) are spliced less efficiently and may undergo decay. The transition from transcription initiation condensates, which include transcription factors and the Mediator complex, to elongation and co-transcriptional splicing condensates is regulated by phosphorylation (P) of the C-terminal domain (CTD) of RNA polymerase II (Pol II) (left). CDK7, cyclin-dependent kinase 7.
Figure 5.
Figure 5.. Condensate-mediated splicing regulation in health and disease.
a. TAR DNA-binding protein 43 (TDP43) mutants (A315T and A315E) present in familial amyotrophic lateral sclerosis (ALS) might favor the transition from reversible to irreversible, pathogenic TDP43 aggregation. Lys84 (K84) acetylation reduces the nuclear import of TDP43. Lys136 acetylation impairs the RNA-binding and splicing capabilities of TDP43, and failure to interact with RNA enhances its capacity to undergo phase separation through its low complexity (LC) domain, thereby favoring the formation of pathological, insoluble TDP43 aggregates . Sirtuin-1 deacetylates Lys136, thereby reducing the aggregation propensity of TDP43. b. Methyl-CpG-binding protein 2 (MeCP2) normally forms co-condensates and assemblies of the MeCP2–RNA-binding fox-1 homolog (RBFOX1)–large assembly of splicing regulators (LASR) complex that regulate co-transcriptional splicing. MeCP2 mutations (red asterisk) in Rett syndrome reduce co-condensation and impair the binding of RBFOX1–LASR to key neuronal pre-mRNAs, thereby affecting their alternative splicing,. c. A-kinase anchoring protein 95 kDa (AKAP95) and the deubiquitinase ubiquitin specific peptidase 42 (USP42) form liquid-like condensates that regulate splicing programs that support cancer growth,. Through its positively charged C-terminal intrinsically disordered region (IDR), USP42 condensation is important for its colocalization with the pleiotropic regulator 1 (PLRG1), a spliceosome component, and nuclear speckles. Depletion of AKAP95, USP42, or PLRG1 markedly reduced cancer cell growth and caused alternative splicing changes in numerous genes including some involved in cell growth,. V354M, a colon cancer mutation in an RNA-recognition motif (RRM) of RBM10, enhances RBM10 condensation, which may explain how this it and related RBM10 mutations affect cancer-associated splicing,,,. d. RBM20 forms nuclear foci that regulate alternative splicing of transcripts from multiple chromosomes. Mutations in the serine-arginine-rich domain of RMB20 are associated with cardiomyopathies; they alter its location to the cytoplasm, where it forms condensates that sequester mRNAs and proteins,,. NLS, nuclear localization signal; Pol II, RNA polymerase II; Y-LC, Tyr-rich low complexity domain.

Similar articles

Cited by

References

    1. Berget SM, Moore C & Sharp PA Spliced segments at the 5’ terminus of adenovirus 2 late mRNA. Proceedings of the National Academy of Sciences of the United States of America 74, 3171–3175, doi:10.1073/pnas.74.8.3171 (1977). - DOI - PMC - PubMed
    1. Chow LT, Gelinas RE, Broker TR & Roberts RJ An amazing sequence arrangement at the 5’ ends of adenovirus 2 messenger RNA. Cell 12, 1–8, doi:10.1016/0092-8674(77)90180-5 (1977). - DOI - PubMed
    1. Pan Q, Shai O, Lee LJ, Frey BJ & Blencowe BJ Deep surveying of alternative splicing complexity in the human transcriptome by high-throughput sequencing. Nature genetics 40, 1413–1415, doi:10.1038/ng.259 (2008). - DOI - PubMed
    1. Wang ET et al. Alternative isoform regulation in human tissue transcriptomes. Nature 456, 470–476, doi:10.1038/nature07509 (2008). - DOI - PMC - PubMed
    1. Tilgner H et al. Deep sequencing of subcellular RNA fractions shows splicing to be predominantly co-transcriptional in the human genome but inefficient for lncRNAs. Genome research 22, 1616–1625, doi:10.1101/gr.134445.111 (2012). - DOI - PMC - PubMed

LinkOut - more resources