Epithelial splicing regulatory protein 1 (esrp1) of the elephant shark is a 684-residue RNA-binding protein of the ESRP branch of the hnRNP-related PTHR13976 family. It carries an N-terminal RNase H-like fold followed by three RNA recognition motifs (RRMs), and is the ortholog of mammalian ESRP1: it groups with tetrapod ESRP1 sequences (about 78% identity to human ESRP1 over its full length) rather than with ESRP2 or hnRNP H/F. Mammalian ESRP1 is a nuclear, epithelial cell-type-specific regulator of alternative pre-mRNA splicing that binds GU-rich (UGG/GGU) cis-elements in introns and controls a large program of epithelial splice isoforms, best known for promoting the FGFR2 IIIb isoform. Alternative 5' splice-site use at the end of exon 12 produces nuclear and cytoplasmic ESRP1 isoforms, and the cytoplasmic form binds UTRs and acts post-transcriptionally. The shark protein is expected to share the conserved mRNA-binding and splicing-regulatory activities; it has not been studied directly.
Summary: InterPro2GO mapping from the RBD-domain superfamily (IPR035979) and RNase H-like superfamily (IPR036397). The protein is a nucleic acid-binding protein, but the characterised activity of its ESRP1 orthologs is sequence-specific binding of pre-mRNA and mRNA via the RRMs; there is no evidence of DNA binding, and the RNase H-like fold is not known to bind or cleave nucleic acid in ESRPs.
Reason: Correct but too general. The target groups with tetrapod ESRP1 inside the ESRP clade (file:CALMI/A0A4W3GVU1/A0A4W3GVU1-bioinformatics/RESULTS-rereview.md), which descends from PAINT node PTN001018050 where mRNA binding (GO:0003729) was placed by IBD from fusilli, ESRP1 and ESRP2. Human ESRP1 has mRNA binding by IDA (PMID:19285943). Replace with GO:0003729, the term used in core_functions.
Summary: Combined automated (InterPro RRM / UniRule) assignment of RNA binding. True: the protein has three RRMs and its ESRP1 orthologs bind GU-rich elements in target pre-mRNAs. PAINT also places RNA binding at the deep node PTN000360053. The more specific mRNA binding is, however, supported for the ESRP clade.
Reason: The term is correct but less specific than the evidence supports. Human ESRP1 binds FGFR2 pre-mRNA intronic elements and mRNA UTRs (PMID:19285943, PMID:36238894) and the ESRP-clade PAINT node carries GO:0003729 mRNA binding; refine to that term (also used in core_functions), consistent with the nucleic acid binding row.
Summary: UniProt SubCell mapping (ARBA-derived) of nucleus. Human ESRP1 is nuclear by IDA (PMID:19285943) and its splicing-regulatory function is nuclear. ESRP1 also has cytoplasmic isoforms produced by alternative 5' splice-site choice at the end of exon 12, so nucleus is not exclusive, but it is the compartment of the core activity. PAINT places nucleoplasm at PTN000360053; nucleus is as precise as the direct ortholog evidence for this gene model allows.
Reason: Nuclear localisation is experimentally established for the ortholog (human ESRP1, same ESRP1 clade by phylogeny) and is the site of splicing regulation. The ARBA source itself is not used as support.
Summary: Proposed process annotation. ESRP1 itself does the regulatory work: it binds intronic GU-rich cis-elements in target pre-mRNAs and enhances or silences adjacent exons (FGFR2 IIIb/IIIc being the defining case, plus hundreds of epithelial splicing events). This is participation, not a downstream consequence. Human ESRP1 carries GO:0043484 by IDA (PMID:19285943) and PAINT placed GO:0043484 by IBD at PTN000360053, an ancestor of the ESRP clade; the shark protein groups with ESRP1 and would inherit the IBA if it were in the PANTHER tree. The previous version of this review proposed the child term GO:0048024 (regulation of mRNA splicing, via spliceosome); both the human curators and PAINT chose GO:0043484, so this review follows them.
Reason: Gene product directly performs the regulation via sequence-specific pre-mRNA binding; relationship to the experimentally characterised ortholog is established by phylogenetic placement in the ESRP1 clade, not by best hit alone.
In the NJ tree the target groups with tetrapod ESRP1 (Xenopus, mouse, human)
Core Functions
Sequence-specific binding of GU-rich (UGG/GGU) elements in target pre-mRNAs via three RRMs, through which the protein enhances or represses inclusion of adjacent alternative exons in the nucleus, directing an epithelial splicing program (inferred from the ESRP1 ortholog).
Q: Does the elephant shark esrp1 gene produce nuclear and cytoplasmic isoforms through an alternative 5' splice site at the end of the exon equivalent to mammalian exon 12, and is its expression epithelium-restricted as in mammals?
Suggested Experiments
Experiment: Express tagged shark esrp1 in an ESRP-depleted mesenchymal cell line with an FGFR2 IIIb/IIIc minigene to test whether it rescues epithelial exon IIIb inclusion, and use CLIP to confirm binding to GU-rich intronic elements.
External Prediction Reviews
These computational predictions are reviewed separately from the GOA annotation set used for this review. The assessments below are from this project and do not constitute official GO annotations or endorsement by GO/UniProt. They are not included in the existing annotation review above.
Elephant shark esrp1, an ESRP1 ortholog that groups with tetrapod ESRP1 and retains the characterized ESRP1 nuclear-localization peptide. The single ProtNLM2 prediction (nucleus) is correct and is already carried by the gene in GOA.
Prediction method: ProtNLM2 · Version: UniProt 2024_06 pilot
Review rationale: The target is the elephant shark ESRP1 ortholog: it carries the ESRP family and ESRP1 RRM1 InterPro signatures and groups with tetrapod ESRP1 in a neighbour-joining tree. Its sequence contains the peptide GLSPPPCKLPCLSPP (UniProt sequence), which experiments on mouse Esrp1 showed is sufficient for nuclear localization, while isoforms lacking CKLP are mainly cytoplasmic (PMID:28634384). Nucleus is therefore a justified ortholog transfer, though not exclusive of a cytoplasmic isoform pool. The gene already carries nucleus in GOA (accepted in the gene review), so the prediction is correct but not novel.
PMID:28634384: "we identified a fifteen-amino-acid stretch “GLSPPPCKLPCLSPP” that is sufficient for the nuclear localization while the same peptide excluding CKLP showed a cytoplasmic localization"
PMID:28634384: "isoforms containing the CKLP amino acids (derived from the distal 5′ss) were predominantly nuclear while isoforms lacking the CKLP (derived from the proximal 5′ss) were predominantly localized to the cytoplasm"