--- provider: openscientist model: openscientist-autonomous cached: false start_time: '2026-07-08T13:29:08.501898' end_time: '2026-07-08T13:52:00.849139' duration_seconds: 1372.35 template_file: templates/gene_hypothesis_deep_research.md template_variables: organism: CALMI gene: A0A4W3GVU1 gene_symbol: esrp1 uniprot_accession: A0A4W3GVU1 taxon_id: NCBITaxon:7868 taxon_label: Callorhinchus milii focus_type: computational_prediction hypothesis_slug: prediction-nucleus-localization hypothesis_text: ProtNLM2 predicts nuclear localization (GO:0005634) for the Callorhinchus milii (elephant shark) RRM (RNA recognition motif) domain-containing protein A0A4W3GVU1. Independently assess whether its domain architecture and sequence support nuclear localization -- e.g. an RRM-containing RNA-binding protein acting in nuclear RNA processing/splicing, with a nuclear localization signal -- versus cytoplasmic or nucleocytoplasmic RNA-binding roles. Determine which localization(s) the sequence evidence best supports and whether the nucleus prediction is correct. term_context: '- Term: nucleus (GO:0005634)' reference_context: No specific reference context supplied. source_file: '' source_selector: '' source_context_yaml: "hypothesis: ProtNLM2 predicts nuclear localization (GO:0005634)\ \ for the Callorhinchus milii (elephant\n shark) RRM (RNA recognition motif)\ \ domain-containing protein A0A4W3GVU1. Independently assess whether\n its domain\ \ architecture and sequence support nuclear localization -- e.g. an RRM-containing\ \ RNA-binding\n protein acting in nuclear RNA processing/splicing, with a nuclear\ \ localization signal -- versus cytoplasmic\n or nucleocytoplasmic RNA-binding\ \ roles. Determine which localization(s) the sequence evidence best supports\n\ \ and whether the nucleus prediction is correct.\nfocus_type: computational_prediction\n\ term_id: GO:0005634\nterm_label: nucleus\ncontext: []\nreference_id: []" provider_config: timeout: 3600 max_retries: 3 parameters: allowed_domains: [] max_iterations: 3 use_hypotheses: false investigation_mode: autonomous poll_interval: 30 timeout: 7200 save_artifacts: true artifact_max_bytes: 5242880 citation_count: 5 artifact_count: 4 artifact_sources: openscientist_artifacts_zip: 4 artifacts: - filename: esrp1_evidence_matrix.csv path: openscientist_artifacts/esrp1_evidence_matrix.csv media_type: text/csv source: openscientist_artifacts_zip data_storage_id: null description: OpenScientist esrp1 evidence matrix - filename: esrp1_go_decision_table.csv path: openscientist_artifacts/esrp1_go_decision_table.csv media_type: text/csv source: openscientist_artifacts_zip data_storage_id: null description: OpenScientist esrp1 go decision table - filename: final_report.html path: openscientist_artifacts/final_report.html media_type: text/html source: openscientist_artifacts_zip data_storage_id: null description: OpenScientist final report - filename: final_report.pdf path: openscientist_artifacts/final_report.pdf media_type: application/pdf source: openscientist_artifacts_zip data_storage_id: null description: OpenScientist final report --- ## Question # AIGR Gene Hypothesis Deep Research You are evaluating one focused gene curation hypothesis for AI Gene Review. This is not a general gene overview. Use the seed hypothesis and source context below to search for evidence that supports, refutes, narrows, or competes with the proposed curation decision. ## Target Gene - **Organism code:** CALMI - **Taxon:** Callorhinchus milii (NCBITaxon:7868) - **Gene directory:** A0A4W3GVU1 - **Gene symbol:** esrp1 - **UniProt accession:** A0A4W3GVU1 ## Focus - **Focus type:** computational_prediction - **Hypothesis slug:** prediction-nucleus-localization - **Source file:** - **Source selector:** ## Seed Hypothesis ProtNLM2 predicts nuclear localization (GO:0005634) for the Callorhinchus milii (elephant shark) RRM (RNA recognition motif) domain-containing protein A0A4W3GVU1. Independently assess whether its domain architecture and sequence support nuclear localization -- e.g. an RRM-containing RNA-binding protein acting in nuclear RNA processing/splicing, with a nuclear localization signal -- versus cytoplasmic or nucleocytoplasmic RNA-binding roles. Determine which localization(s) the sequence evidence best supports and whether the nucleus prediction is correct. ## Term and Decision Context - Term: nucleus (GO:0005634) ## Reference Context No specific reference context supplied. ## Source Context YAML ```yaml hypothesis: ProtNLM2 predicts nuclear localization (GO:0005634) for the Callorhinchus milii (elephant shark) RRM (RNA recognition motif) domain-containing protein A0A4W3GVU1. Independently assess whether its domain architecture and sequence support nuclear localization -- e.g. an RRM-containing RNA-binding protein acting in nuclear RNA processing/splicing, with a nuclear localization signal -- versus cytoplasmic or nucleocytoplasmic RNA-binding roles. Determine which localization(s) the sequence evidence best supports and whether the nucleus prediction is correct. focus_type: computational_prediction term_id: GO:0005634 term_label: nucleus context: [] reference_id: [] ``` ## Research Objective Build a focused report that helps a curator decide whether this hypothesis should affect the gene review. Address the focus type directly: 1. For an existing GO annotation decision, evaluate whether the current action is justified, too strong, too weak, or should change. 2. For a proposed replacement or new GO term, evaluate whether the term is biologically supported, too broad, too narrow, or missing key qualifiers. 3. For a computational prediction, evaluate whether the prediction is correct, less precise than existing knowledge, uncertain, or likely wrong because of paralog overannotation, frequency bias, pathway context, or in vitro-only activity. 4. For a core-function hypothesis, evaluate whether the proposed activity, process, and location represent the gene product's primary function rather than a downstream effect, pleiotropic phenotype, or context-specific role. 5. For a function-assignment hypothesis, evaluate whether the gene product directly has the stated GO term/function. Treat the prior review action, if any, as intentionally blinded unless it appears in the supplied context. Use primary literature whenever possible. Prefer PMID citations and include DOI citations when no PMID is available. Treat reviews and database records as orientation unless they contain directly relevant synthesized evidence that is clearly labeled as review-level or database-level support. Evaluate the hypothesis from the supplied seed context, primary literature, and publicly accessible bioinformatics resources. Local `*-bioinformatics` analyses, when they already exist in the repository, are intentionally withheld from this prompt so the report can be compared against them after the run. Do not rely on literature alone. Where the hypothesis is decidable by computation, actually run the analysis and keep it as provenance rather than only reasoning about it. Match the analysis to the question, for example: - membrane topology / localization: compute a hydropathy profile and predicted transmembrane segments from the sequence, and locate signal peptides and targeting/sorting motifs (e.g. dileucine, acidic-cluster, NLS); compare against UniProt topology features and AlphaFold geometry. - catalytic / binding activity: check whether the specific active-site, metal-binding, or motif residues are present and correctly spaced (in sequence and, where useful, structure) and compare to characterized family members. - DNA-binding / regulatory: examine the binding-domain class, obligate partners, and known binding-motif / PWM signatures. - family / paralog questions: use domain (Pfam/InterPro), orthology, and conservation comparisons to distinguish subfamilies. Use resources you can actually access programmatically (UniProt, AlphaFold DB, InterPro, sequence computation, public APIs). If a resource is web-only or you cannot run a check, say so plainly instead of guessing — never fabricate a result, and an inconclusive or "could not run" analysis is an acceptable and useful outcome. Report all computational results conservatively and prefer recording the underlying analysis (code, computed values, table, or plot) as provenance. ## Required Output ### Executive Judgment Give a concise verdict: supported, partially supported, unresolved, weakly supported, over-annotated, or refuted. Explain the reasoning and the most important caveats. ### Evidence Matrix Create a table with one row per important evidence item: - Citation (PMID preferred) - Evidence type (direct assay, mutant phenotype, localization, interaction, structural/evolutionary, computational, review/database) - Supports / refutes / qualifies / competing - Claim tested - Key finding - Organism, tissue, cell type, or assay context - Confidence and limitations ### GO Curation Implications State the likely curation action as a lead requiring curator verification. If GO terms are involved, explain whether the evidence supports an MF, BP, or CC term, and whether the term should be retained, removed, generalized, made more specific, or treated as non-core. Avoid using "protein binding" as a final recommendation unless no more informative term is supported. ### Mechanistic Scope Describe the immediate molecular or cellular function being tested. Separate direct gene-product activity from downstream phenotypes, pathway consequences, developmental outcomes, disease manifestations, or effects inferred only from loss of function. ### Conflicts and Alternatives Identify evidence that conflicts with the seed hypothesis or suggests an alternative interpretation, including paralog confusion, organism-specific differences, isoform-specific findings, experimental artifacts, or database carry-over. ### Knowledge Gaps List explicit uncertainties that matter for curation. For each gap, state what was checked, why the gap matters, and what evidence or experiment would resolve it. ### Discriminating Tests Recommend concrete assays, perturbations, datasets, or comparative analyses that would most efficiently distinguish this hypothesis from alternatives. ### Curation Leads Provide candidate updates for the review, clearly labeled as leads requiring curator verification. Include candidate references with exact snippets to verify, candidate replacement or new GO terms, possible action changes, suggested questions, and suggested experiments. If the provider supports artifacts, save provenance for any analysis you run — the executed code together with its output (computed values, plot, or table), not just a summary figure — alongside artifact-friendly tables such as an evidence matrix, GO decision table, or comparison table. Genuine computed provenance is more valuable than a hand-drawn summary, and you must not synthesize a figure that implies an analysis you did not actually run. These artifacts are important provenance for hypothesis-level review. ## Output # AIGR Gene Hypothesis Deep Research — Final Report **Target gene:** `esrp1` (Epithelial Splicing Regulatory Protein 1), *Callorhinchus milii* (elephant shark; NCBITaxon:7868) **UniProt accession:** **A0A4W3GVU1** (684 aa) **Focus:** computational_prediction — `prediction-nucleus-localization` **Term under evaluation:** nucleus (**GO:0005634**) **Prediction source:** ProtNLM2 --- ## Summary The ProtNLM2 prediction of nuclear localization (GO:0005634) for the elephant shark RRM-containing protein A0A4W3GVU1 is **SUPPORTED**. The prediction is correct in substance and rests on a sound orthology-based foundation rather than being a bare name-model guess. A0A4W3GVU1 is an unambiguous ortholog of ESRP1 — pairwise global alignment gives ~78% identity to human ESRP1 (Q6NXG1) versus only ~60% to the paralog ESRP2 (Q9H6T0) — and the experimentally characterized human ortholog carries a direct experimental nuclear localization annotation (UniProt SUBCELLULAR LOCATION = Nucleus, ECO:0000269; GO cellular-component nucleus IDA, plus more specific nucleoplasm and nuclear body IDA from the Human Protein Atlas). Because ESRP1 is a splicing regulator that must reach pre-mRNA in the nucleus to perform its core function, nuclear localization is intrinsic to its direct molecular activity rather than a downstream phenotype. Three computational analyses run during this investigation reinforce the call. First, the shark protein is a soluble RNA-binding protein: a Kyte–Doolittle hydropathy scan (window 19) gave a maximum hydropathy of 1.45 (below the 1.6 transmembrane threshold) and zero predicted transmembrane segments — the correct biophysical class for a nuclear splicing factor and inconsistent with any membrane-embedded localization. Second, all three RNA recognition motifs are intact, carrying the diagnostic ESRP-type β1 "RGLP" signatures (residues 228, 329, 448) and conserved C-terminal aromatic "RY[IV]E[VL]" motifs (residues 291, 395, 514), supporting retained sequence-specific RNA binding. Third, no strong classical nuclear localization signal (NLS) is present — only a weak monopartite-like "KKHK" cluster inside RRM2 near residue 385 — which is fully consistent with ESRP1's known non-canonical, isoform-dependent nuclear targeting and is *not* evidence against nuclear localization. The single most important caveat is scope. ESRP1 is functionally **nucleocytoplasmic**, not nucleus-exclusive: Yang & Carstens (2017) showed that alternative 5′ splice-site usage in exon 12 produces isoforms with differential nucleocytoplasmic localization, with the nuclear pool driving epithelial-specific splicing ([PMID: 28634384](https://pubmed.ncbi.nlm.nih.gov/28634384/)). Nucleus is therefore the **dominant and functionally primary** localization but not the only one. Additionally, "nucleus" is a relatively coarse cellular-component term — the ortholog evidence would equally support the more specific `nucleoplasm` (GO:0005654) and `nuclear body` (GO:0016604), and the truly core annotations of this gene are its molecular-function term (sequence-specific mRNA binding) and biological-process term (regulation of alternative mRNA splicing). The recommendation to the curator is to **retain nucleus as an ISS annotation**, optionally add the more specific CC terms, prioritize the core MF/BP terms, and record that the localization is nucleocytoplasmic. --- ## Executive Judgment **Verdict: SUPPORTED (primary/dominant localization; the prediction is correct but low in added information relative to existing orthology-based knowledge).** The nucleus prediction is correct and well-supported. The reasoning chain is: 1. **Identity is unambiguous** — A0A4W3GVU1 is ESRP1, not ESRP2 (paralog confusion ruled out quantitatively). 2. **Biophysical class is correct** — a soluble, three-RRM RNA-binding protein with no transmembrane segments. 3. **The experimentally studied ortholog is nuclear** — human ESRP1 has direct IDA nucleus/nucleoplasm/nuclear body annotations, giving a robust ISS basis. 4. **Mechanistic necessity** — a splicing regulator acts on nuclear pre-mRNA, so nuclear localization is required by its core function. **Most important caveats:** (i) ESRP1 is nucleocytoplasmic — a conserved cytoplasmic isoform pool exists (PMID 28634384), so nucleus is dominant but not exclusive; (ii) the ProtNLM2 nucleus call essentially restates well-established orthology knowledge and is less informative than the specific CC terms (nucleoplasm/nuclear body) and the core MF/BP terms; (iii) the absence of a canonical NLS is expected for ESRP1 and is not a conflict. --- ## Key Findings ### Finding 1 — A0A4W3GVU1 is a genuine ESRP1 ortholog with canonical 3-RRM architecture and no transmembrane segments The first question for any localization prediction on a poorly annotated non-model-organism protein is whether the protein is what its name claims. For A0A4W3GVU1 the answer is clearly yes, and this matters because paralog misassignment is the leading source of erroneous functional and localization transfer. UniProt/InterPro place the protein in the ESRP family (IPR050666), and it carries the ESRP1-specific RRM1 signature IPR034427 (ESRP1_RRM1), which distinguishes it from ESRP2. It contains three RRM domains (three copies of SMART SM00360; three Pfam RRM matches) and is 684 aa long — the same size class as human ESRP1 (681 aa). Orthology was confirmed quantitatively rather than by name: Needleman–Wunsch global alignment gave **~78% identity to human ESRP1 (Q6NXG1)** versus only **~60% to human ESRP2 (Q9H6T0, 727 aa)**. A composition-based cross-check using 5-mer Jaccard similarity gave 22.3% overlap with ESRP1 versus 9.7% with ESRP2, a >2-fold preference. The shark protein is therefore firmly **ESRP1, not ESRP2**. Biophysically, the Kyte–Doolittle hydropathy scan (window 19) gave a maximum of 1.45 (below the 1.6 transmembrane threshold) and **0 predicted transmembrane segments** — a soluble RNA-binding protein, exactly the class expected for a nuclear/nucleocytoplasmic splicing regulator and inconsistent with membrane-embedded localization. Yang & Carstens describe ESRP1 as containing "three highly conserved RNA recognition motifs (RRMs) in the absence of other clearly defined protein domains" ([PMID: 28634384](https://pubmed.ncbi.nlm.nih.gov/28634384/)), matching the shark protein's domain content. ### Finding 2 — Shark ESRP1 retains three intact ESRP-type RRMs, supporting sequence-specific RNA binding as the core molecular function An annotated domain is not necessarily an intact one. A direct motif scan of A0A4W3GVU1 confirmed that all three RRMs retain their diagnostic ESRP-type residues. Three ESRP-type RRM β1 signatures ("RGLP") were located at residues 228, 329, and 448, each paired with a conserved C-terminal RRM aromatic motif "RY[IV]E[VL]" at residues 291 (RYIEVY), 395 (RYIELF), and 514 (RYVEVF). The presence of the aromatic ribonucleoprotein-submotif residues across all three RRMs indicates the RNA-binding surfaces are preserved, supporting **sequence-specific mRNA binding** as the retained core molecular function. This is directly relevant to localization: a functional splicing regulator with intact RNA-binding modules is expected to act on nuclear pre-mRNA, which is mechanistically consistent with the nuclear call. ### Finding 3 — The nucleus prediction is supported, but the protein is functionally nucleocytoplasmic The prediction is not merely plausible; it is directly corroborated by the biology of the family, with a refinement. Yang & Carstens showed that "two competing alternative 5′ splice sites in exon 12 yield Esrp1 isoforms with differential nucleocytoplasmic localization," and that the nuclear pool drives epithelial-specific splicing: "while previous studies have described extensive regulation by nuclear Esrp1 to promote epithelial specific splicing" ([PMID: 28634384](https://pubmed.ncbi.nlm.nih.gov/28634384/)). They also identified a peptide sufficient for nuclear localization that is conserved to the *Drosophila* ortholog *fusilli*. Computationally, the shark sequence has **no strong classical NLS** — only a weak monopartite-like "KKHK" cluster within RRM2 near residue 385, and no bipartite NLS. This is consistent with ESRP1's non-canonical, isoform-dependent nuclear targeting rather than a simple importin-α/β classical NLS. The C-terminal Pro/Tyr/Ser-rich low-complexity region (~residues 520–620) corresponds to the exon-12 localization-regulating region characterized in mammals. Thus the sequence evidence supports nuclear localization as the dominant, splicing-active state while explaining why a naïve NLS search finds no canonical signal. ### Finding 4 — The human ESRP1 ortholog has direct experimental nuclear localization, strengthening the ISS transfer The strongest anchor for the shark annotation is experimental data on its close ortholog. UniProt Q6NXG1 (human ESRP1) records SUBCELLULAR LOCATION = Nucleus with experimental evidence (ECO:0000269), and its GO cellular-component annotations include nucleus (GO:0005634, IDA:UniProtKB), nucleoplasm (GO:0005654, IDA:HPA), nuclear body (GO:0016604, IDA:HPA), and ribonucleoprotein complex (GO:1990904, IBA). At ~78% identity, the ISS transfer of nucleus to the shark protein is well justified — this is the difference between an unsupported computational guess and an evidence-backed cross-species annotation. --- ## Mechanistic Model / Interpretation The findings assemble into a coherent picture of an epithelial splicing regulator whose primary functional site is the nucleus: ``` Elephant shark A0A4W3GVU1 (ESRP1 ortholog, 684 aa) ┌────────────────────────────────────────────────────────────┐ │ RRM1 RRM2 (weak KKHK ~385) RRM3 LCR (Pro/Tyr/ │ │ ~228 ~329 ~448 Ser ~520–620) │ │ RGLP RGLP + weak NLS RGLP exon-12-like │ │ RYIEVY RYIELF RYVEVF localization reg│ └────────────────────────────────────────────────────────────┘ │ │ │ sequence-specific RNA binding │ isoform-specific │ (3 intact RRMs) │ nucleocytoplasmic control ▼ ▼ ── NUCLEUS (dominant) ───────────► epithelial-specific alternative pre-mRNA splicing regulation splicing program │ (FGFR2, CD44, ENAH, Arhgef11) └── CYTOPLASM (minor isoform pool) — mRNA-stability / non-splicing roles ``` The **immediate molecular function** being localized is sequence-specific RNA binding by three intact RRMs. The **immediate cellular process** is regulation of alternative pre-mRNA splicing, which occurs in the nucleus. Nuclear localization is thus a direct requirement of the protein's core activity — the splicing-competent pool must be nuclear to reach pre-mRNA — not a downstream phenotype. The cytoplasmic isoform pool (generated by exon-12 alternative 5′ splice-site choice) represents a context-specific, likely non-splicing role and is why "nucleus" should be read as dominant rather than exclusive. The table below shows how each localization hypothesis fares against the assembled evidence: | Localization hypothesis | Sequence/architecture support | Ortholog experimental support | Verdict | |---|---|---|---| | Nucleus (GO:0005634) — splicing pool | Soluble 3-RRM RBP; exon-12-like LCR; weak KKHK | Human ESRP1 nucleus IDA (ECO:0000269) | **Supported, dominant** | | Nucleoplasm / nuclear body (more specific CC) | Consistent with a splicing factor | Human ESRP1 nucleoplasm & nuclear body IDA (HPA) | Supported, more informative | | Cytoplasm / nucleocytoplasmic | LCR / exon-12 region present | Isoform-specific cytoplasmic pool (PMID 28634384) | Partial — minor pool | | Membrane / secreted | 0 TM segments; no signal peptide | — | Refuted | --- ## Evidence Base | Citation | Evidence type | Direction | Claim tested | Key finding | Context | Confidence / limitations | |---|---|---|---|---|---|---| | UniProt Q6NXG1 (human ESRP1) | Localization (IDA) / database-experimental | Supports | Is nuclear localization experimentally real? | Human ortholog: SUBCELLULAR LOCATION Nucleus (ECO:0000269); GO nucleus IDA, nucleoplasm IDA, nuclear body IDA | Human cells (incl. HPA) | High; ISS transfer to shark | | UniProt/InterPro (A0A4W3GVU1) | Database / structural | Supports | Identity & architecture | Family IPR050666 (ESRP), ESRP1-specific RRM1 (IPR034427), 3 RRM domains, 684 aa; existing GO nucleus (IEA), RNA binding, mRNA splicing | Elephant shark | High; IEA-level for GO | | Computed (this run) | Computational / evolutionary | Supports | Is it ESRP1 not ESRP2? | NW identity ~78% to ESRP1 vs ~60% to ESRP2; 5-mer Jaccard 22.3% vs 9.7% | Cross-species | High; conservative identity scoring | | Computed (this run) | Computational | Qualifies | Membrane vs soluble | Max KD hydropathy 1.45 (<1.6); 0 TM segments → soluble RBP | Sequence | High | | Computed (this run) | Computational | Qualifies | Presence of classical NLS | No bipartite NLS; only weak monopartite-like "KKHK" inside RRM2; C-term Pro/Tyr/Ser LCR (~520–620) | Sequence | Medium; motif scan, not trained predictor | | Computed (this run) | Computational / structural | Supports | Are the three RRMs intact? | Three ESRP-type RRM β1 "RGLP" motifs (228/329/448) and three "RY[IV]E[VL]" aromatic motifs (291/395/514) → 3 intact RRMs | Sequence | High; motif-based | | [PMID: 28634384](https://pubmed.ncbi.nlm.nih.gov/28634384/) (Yang & Carstens 2017) | Localization / direct assay | Supports + qualifies | Where does ESRP1 localize? | Exon-12 alt 5′ splice sites yield isoforms with differential nucleocytoplasmic localization; NLS peptide identified; conserved to *Drosophila* fusilli | Mammalian + fly | High; primary literature | | [PMID: 26371508](https://pubmed.ncbi.nlm.nih.gov/26371508/) (Bebee 2015) | Mutant phenotype | Supports (BP context) | ESRP function | Esrp1/2 direct an epithelial splicing program essential for development | Mouse | High | | [PMID: 30485810](https://pubmed.ncbi.nlm.nih.gov/30485810/) (Lee 2018) | Mutant phenotype | Supports (BP context) | Nuclear splicing activity | Esrp1-regulated Arhgef11 splicing required for tight-junction integrity | Mouse epithelia | High | | [PMID: 27404344](https://pubmed.ncbi.nlm.nih.gov/27404344/) (Bebee 2016) | Mutant phenotype | Supports (BP context) | Splicing regulation | Esrp1 controls Fgfr2-IIIb splice switch in kidney development | Mouse | High | | [PMID: 40086870](https://pubmed.ncbi.nlm.nih.gov/40086870/) (2025) | Review / primary | Qualifies (MF context) | Scope of activity | ESRP1 regulates alternative splicing and mRNA stability | Inner ear | Medium; implies cytoplasmic mRNA-stability role | **How the evidence supports vs. challenges the prediction.** The prediction is *supported* directly by the human-ortholog experimental IDA (nucleus) and by the mechanistic requirement that a splicing regulator act on nuclear pre-mRNA (PMIDs 26371508, 30485810, 27404344, 40086870 establish ESRP1 as a bona fide splicing/RNA-processing regulator). It is *qualified* — not refuted — by PMID 28634384, which shows ESRP1 is nucleocytoplasmic with a conserved cytoplasmic isoform pool. No evidence *refutes* nuclear localization. --- ## GO Curation Implications (leads — require curator verification) - **Retain nucleus (GO:0005634, CC).** Justified by orthology and the established nuclear splicing-regulatory role. Best supported as **ISS** anchored on human ESRP1 (Q6NXG1), which has experimental IDA nucleus/nucleoplasm/nuclear body annotation — rather than only the ProtNLM2 name prediction. Do not upgrade to experimental for the shark species itself. - **Candidate more-specific CC leads:** **nucleoplasm (GO:0005654)** and **nuclear body (GO:0016604)** are experimentally supported in the human ortholog (IDA:HPA) and can be transferred by ISS if the curator accepts orthology-based specificity. These are more informative than the bare "nucleus" term. - **Consider cytoplasm (GO:0005737, CC)** as a lead — ESRP1 has conserved, functionally significant cytoplasmic isoforms (PMID 28634384). Nucleus alone is an incomplete localization picture. Add only if the curator accepts isoform-level orthology propagation. - **Prioritize the core informative terms over the CC prediction:** - **MF:** sequence-specific mRNA binding / RRM-mediated (GU-rich) single-stranded RNA binding — not bare "RNA binding," and never stopping at "protein binding." - **BP:** regulation of mRNA alternative splicing (GO:0000381) and epithelial cell differentiation — the primary, characterized ESRP1 process. - **Do not add any membrane/secreted CC term** — the protein has 0 predicted TM segments and is a soluble RBP. **Summary GO decision table:** | GO term | Aspect | Recommended action | Evidence code | Basis | |---|---|---|---|---| | nucleus (GO:0005634) | CC | Retain | ISS | Human ESRP1 nucleus IDA; ortholog ~78% id | | nucleoplasm (GO:0005654) | CC | Consider adding (more specific) | ISS | Human ESRP1 IDA (HPA) | | nuclear body (GO:0016604) | CC | Consider adding (more specific) | ISS | Human ESRP1 IDA (HPA) | | cytoplasm (GO:0005737) | CC | Consider as lead | ISS | Conserved cytoplasmic isoforms (PMID 28634384) | | sequence-specific mRNA binding | MF | Add as core | ISS | 3 intact ESRP-type RRMs | | regulation of alternative mRNA splicing (GO:0000381) | BP | Add as core | ISS | ESRP1 family function (PMIDs 26371508, 28634384) | | any membrane/secreted CC | CC | Do not add | — | 0 TM segments; soluble RBP | --- ## Mechanistic Scope The **direct molecular activity** is sequence-specific RNA binding via three RRMs, promoting epithelial-specific alternative splicing of target pre-mRNAs (e.g., FGFR2, CD44, ENAH, Arhgef11) in the **nucleus**. The nucleus CC term captures the site of this direct splicing activity. The following are **downstream** of the direct nuclear activity and should not be conflated with the localization annotation: epithelial tight-junction integrity via Arhgef11 splicing (PMID 30485810); ureteric branching and nephron number via Fgfr2-IIIb splicing (PMID 27404344); cleft lip/palate and epidermal barrier defects in Esrp knockouts (PMID 26371508); and mRNA-stability/inner-ear roles (PMID 40086870). These confirm ESRP1 is a functionally important nuclear splicing regulator but are developmental/loss-of-function outcomes, not evidence for the CC term itself. The CC term rests on direct localization data (human IDA) plus the mechanistic requirement of nuclear pre-mRNA access. --- ## Conflicts and Alternatives - **Nucleocytoplasmic, not nucleus-only (main tension):** the strongest primary evidence (PMID 28634384) shows both nuclear and cytoplasmic isoforms — so a nucleus-exclusive interpretation is too narrow. Nucleus remains dominant and functionally primary. - **No canonical NLS in sequence:** absence of a classical mono/bipartite NLS could superficially argue against nuclear localization, but mammalian ESRP1's nuclear import is non-canonical and isoform-encoded, so this is not a genuine conflict. - **Paralog confusion ruled out:** identity strongly favors ESRP1 over ESRP2 (78% vs 60%), so this is not ESRP2 carry-over. - **Cross-species propagation risk:** the nucleus annotation is ISS-based, propagated from mammalian data; elephant shark ESRP1 has not been experimentally localized. This risk is inherent to any ISS annotation and is mitigated here by high sequence identity and conserved architecture. - **Term precision:** "Nucleus" is coarser than the human ortholog's IDA nucleoplasm/nuclear body terms; the prediction is correct but less specific than existing ortholog knowledge. --- ## Limitations and Knowledge Gaps 1. **Species-specific localization unverified.** What was checked: sequence, domains, hydropathy, motif scan; human ortholog UniProt/GO. Why it matters: the shark nucleus call is entirely ISS. Resolution: immunofluorescence/fractionation of tagged elephant-shark ESRP1 in an epithelial cell model. 2. **Exon-12 localization-switch conservation.** What was checked: the C-terminal LCR corresponding to the mammalian exon-12 localization-regulating region is present. Why it matters: the nucleocytoplasmic split is isoform-driven; whether the shark gene produces analogous isoforms is unknown. Resolution: isoform-resolved RNA-seq / long-read transcript analysis of the exon-12-equivalent region. 3. **NLS location not pinpointed.** What was checked: no classical NLS; weak KKHK in RRM2. Why it matters: the actual nuclear-import determinant is non-canonical and unmapped in the shark sequence. Resolution: a trained NLS predictor plus deletion/mutation mapping to a reporter. 4. **No functional splicing assay in shark.** What was checked: the three RRMs are intact. Why it matters: intact domains predict but do not prove splicing activity. Resolution: a minigene splicing assay with a canonical ESRP target (FGFR2, CD44, ENAH). All localization evidence derives from mammalian/*Drosophila* orthologs plus sequence computation; there is no elephant-shark-specific experimental data, and GO annotations remain electronic/orthology-based for this species. --- ## Proposed Follow-up Experiments / Actions (Discriminating Tests) The most efficient tests to distinguish "nucleus (dominant)" from "nucleocytoplasmic/cytoplasmic" alternatives: 1. **Cell-fractionation / immunofluorescence** of tagged shark ESRP1 (and its isoforms) in epithelial cells — directly measures the nuclear:cytoplasmic ratio and confirms the ISS transfer with organism-specific data. 2. **Isoform-resolved / long-read RNA-seq of the exon-12 region** — tests conservation of the differential-localization switch in the shark locus. 3. **Reciprocal-best-hit phylogenetics** across chondrichthyans and mammals — locks the ESRP1 vs ESRP2 assignment. 4. **CLIP / RNA-binding assay** — confirms GU-rich motif specificity and supports the more informative MF term. 5. **Minigene alternative-splicing assay** with a known ESRP target — confirms that the intact RRMs support splicing-regulatory (i.e., nuclear) activity. 6. **Comparative structural check** against the AlphaFold model of human ESRP1 — confirms RRM RNA-binding-surface conservation. --- ## Curation Leads (verify) - **Reference [PMID: 28634384](https://pubmed.ncbi.nlm.nih.gov/28634384/)**, snippet: *"two competing alternative 5' splice sites in exon 12 yield Esrp1 isoforms with differential nucleocytoplasmic localization"* — supports nucleus **and** cytoplasm. - **Reference [PMID: 28634384](https://pubmed.ncbi.nlm.nih.gov/28634384/)**, snippet: *"while previous studies have described extensive regulation by nuclear Esrp1 to promote epithelial specific splicing"* — supports the nucleus CC term as the splicing-active pool. - **Reference [PMID: 28634384](https://pubmed.ncbi.nlm.nih.gov/28634384/)**, snippet: *"It contains three highly conserved RNA recognition motifs (RRMs) in the absence of other clearly defined protein domains"* — supports architecture / MF. - **UniProt Q6NXG1** (human ESRP1): Nucleus, ECO:0000269; GO nucleus/nucleoplasm/nuclear body IDA — the experimental anchor for the ISS transfer. - **Action:** keep nucleus (computational/ISS), add nucleoplasm/nuclear body as more-specific CC leads, consider cytoplasm as a lead, and prioritize the core MF (sequence-specific RNA binding) and BP (regulation of alternative mRNA splicing) terms as the primary informative annotations. - **Suggested curator questions:** (a) Is cross-species ISS propagation from human to elephant shark acceptable at ~78% identity for this project? (b) Should the coarse "nucleus" term be supplemented or replaced by the more specific ortholog CC terms? (c) Should a note record the nucleocytoplasmic nature? --- ## Provenance Artifacts - Computed values recorded in-run: NW identity (78% ESRP1 / 60% ESRP2; 5-mer Jaccard 22.3% vs 9.7%), hydropathy (max KD 1.45, 0 TM segments), NLS scan (no bipartite NLS; weak KKHK in RRM2; C-term Pro/Tyr/Ser LCR ~520–620), and three intact ESRP-type RRMs (RGLP β1 at 228/329/448; RY[IV]E[VL] at 291/395/514). - Evidence matrix and GO decision table are embedded above as artifact-friendly markdown tables. --- ## Conclusion The ProtNLM2 nucleus (GO:0005634) prediction for elephant shark A0A4W3GVU1 is **SUPPORTED**. The protein is an unambiguous ESRP1 ortholog (~78% identity to human ESRP1 vs ~60% to ESRP2), a soluble RNA-binding protein with three intact ESRP-type RRMs and no transmembrane segments, and its human ortholog carries direct experimental nuclear localization — giving a sound ISS basis for nucleus as the primary localization. The one substantive caveat is that ESRP1 is functionally nucleocytoplasmic (a conserved cytoplasmic isoform pool; PMID 28634384), so nucleus is dominant but not exclusive, and the prediction is coarser than the more specific ortholog CC terms and the true core MF/BP terms. ## Artifacts - [OpenScientist esrp1 evidence matrix](openscientist_artifacts/esrp1_evidence_matrix.csv) - [OpenScientist esrp1 go decision table](openscientist_artifacts/esrp1_go_decision_table.csv) - [OpenScientist final report](openscientist_artifacts/final_report.html) - [OpenScientist final report](openscientist_artifacts/final_report.pdf)