AIGR Gene Hypothesis Deep Research — B4MAQ2 (DROVI)

Hypothesis slug: prediction-cytoplasm-localization Target: B4MAQ2, Drosophila virilis (NCBITaxon:7244), gene Dvir\GJ15622 Focus type: computational_prediction Term under evaluation: cytoplasm (GO:0005737) Seed prediction source: ProtNLM2

Summary

The ProtNLM2 prediction that B4MAQ2 is cytoplasmic (GO:0005737) is correct but under-informative. Domain architecture, orthology, and predicted structure converge on an unambiguous identity: B4MAQ2 is Exportin-5 (XPO5), an importin-β-family (karyopherin-β), Ran-GTP-dependent nucleocytoplasmic transport receptor, not a generic cytoplasmic protein. Because karyopherin-β receptors shuttle through the nuclear pore, they genuinely occupy the cytoplasm (where Ran-GTP hydrolysis releases cargo) as well as the nucleus (where cargo is loaded) and the nuclear pore/envelope (translocation). The prediction therefore names one true compartment of a shuttling receptor while omitting the nucleus and — more importantly — the receptor's defining molecular function and biological process.

The identification is watertight. The UniProt record (1,238 aa, soluble; GRAVY = 0.007; no transmembrane span; no signal peptide) carries a diagnostic exportin-5 domain set: an Importin-β N-terminal Ran-binding domain (Pfam PF03810 / IBN_N), an Exportin-1/Importin-β-like domain (InterPro IPR013598), an Exportin-5 C-terminal domain (Pfam PF19273 / InterPro IPR045478), the Xpo1 domain (PF08389), and an all-α ARM/HEAT solenoid fold. PANTHER assigns the subfamily PTHR11223:SF3 = EXPORTIN-5. A full-length Needleman–Wunsch alignment gives 85.5% identity to D. melanogaster Ranbp21/Exportin-5 (Q9VWE7/CG12234) — a clean 1:1 ortholog — and k-mer paralog discrimination excludes confusion with CRM1/XPO1, IPO5, CSE1L/XPO2, TNPO1, and KPNB1. The AlphaFold model (mean pLDDT 82.9) confirms the all-α HEAT/ARM superhelical solenoid characteristic of the karyopherin fold.

The correct curation posture is therefore to retain cytoplasm (GO:0005737) as a secondary cellular-component annotation, add nucleus (GO:0005634), and lead with the transport-receptor molecular function (GO:0005049) and Ran binding (GO:0031267) plus the nuclear-export biological processes (GO:0006405 / GO:0035281 / GO:0006611). A curator who accepts "cytoplasm" alone records a true fact while leaving the annotation set silent about what the gene actually does. The chief caveat is that all evidence is computational or orthology-based: no direct experimental localization or function has been published for the D. virilis protein itself, and exportin-5 orthologues are known to be functionally divergent across species, so cargo-level details should be transferred "by similarity" with an ISS/ISO evidence code.

Executive Judgment

Verdict: Partially supported (correct but under-informative / incomplete).

The seed prediction is not refuted — Exportin-5 has an obligatory cytoplasmic phase in its transport cycle, so GO:0005737 is factually valid. The problem is one of completeness and informativeness, not truth. Collapsing this shuttling receptor to a single "cytoplasm" CC term discards its molecular identity: a receptor that binds Ran-GTP and cargo (pre-miRNAs, tRNAs, and certain proteins) in the nucleus and releases them in the cytoplasm. The most important caveats are: (1) no direct experimental data exist for B4MAQ2 itself — the functional attribution rests on 85.5% identity to the experimentally characterized D. melanogaster ortholog; and (2) exportin-5 orthologues diverge functionally across species (16963774), so cargo preferences should be annotated conservatively.

Key Findings

Finding 1 — B4MAQ2 is Exportin-5, an importin-β-family transport receptor; "cytoplasm" is correct but incomplete

The sequence and annotation evidence is diagnostic of exportin-5/XPO5. The UniProt record (B4MAQ2, D. virilis, gene Dvir\GJ15622) describes a 1,238-residue soluble protein — computed GRAVY = 0.007, no continuous transmembrane span, no signal peptide — ruling out a membrane or secreted assignment and establishing that whatever compartment it occupies, it does so as a soluble factor.

The domain architecture is decisive. B4MAQ2 carries an Importin-β N-terminal Ran-binding domain (Pfam PF03810 / IBN_N, residues 34–100; InterPro IPR001494) — the hallmark of the karyopherin-β superfamily and the surface that engages Ran-GTP; an Exportin-1/Importin-β-like domain (InterPro IPR013598, residues 114–274); an Exportin-5 C-terminal domain (Pfam PF19273; InterPro IPR045478, residues 319–1193) that is essentially pathognomonic for exportin-5; an Xpo1 domain (PF08389); and an all-α ARM/HEAT solenoid (SUPFAM SSF48371; Gene3D 1.25.10.10). Family classifiers agree: PANTHER PTHR11223:SF3 = EXPORTIN-5 at the subfamily level, and the UniProt SIMILARITY line states the protein "Belongs to the exportin family."

Importantly, the existing IEA GO annotations already reflect shuttling-receptor biology rather than cytoplasm alone: they include C:cytoplasm (GO:0005737) and C:nucleus (GO:0005634), plus F:nuclear export signal receptor activity (GO:0005049), F:small GTPase binding (GO:0031267, Ran), P:protein export from nucleus (GO:0006611), and P:RNA export from nucleus (GO:0006405). The ProtNLM2 prediction of cytoplasm alone is a strict subset of — and less informative than — what the automated pipeline already captures.

This finding is anchored to primary literature on exportin-5 as a class. PMID:15134074 classifies "exportin-5 (Exp5), a Ran-dependent importin-beta-related transport receptor, [that] mediates nuclear export of miRNA precursors (pre-miRNAs)," implying a nucleus + cytoplasm shuttle rather than cytoplasm alone. PMID:15254228 establishes the cargo set: "Exportin-5 is a nuclear export receptor for certain classes of double-stranded RNA (dsRNA), including pre-micro-RNAs, viral hairpin RNAs, and some tRNAs." PMID:20951941 confirms that "Exportin 5 (XPO5) mediates pre-miRNA nuclear export," a nucleus-to-cytoplasm process.

Finding 2 — B4MAQ2 is a 1:1 ortholog of Drosophila Exportin-5 (Ranbp21/dmExp5), which experimentally exports pre-miRNAs and tRNAs

A full-length Needleman–Wunsch global alignment of B4MAQ2 (1,238 aa) against D. melanogaster Ranbp21/Exportin-5 (Q9VWE7, 1,241 aa) yields 1,063/1,244 = 85.5% identity — establishing a clean 1:1 orthology and licensing functional transfer with high confidence. Among human karyopherin-β paralogs, B4MAQ2's k-mer (5-mer) containment is highest to human XPO5 (Q9HAV4, 1.6%) and negligible to CRM1/XPO1 (0.2%), IPO5 (0.4%), CSE1L/XPO2 (0.4%), TNPO1 (0.0%), and KPNB1 (0.1%). This paralog discrimination matters: the karyopherin-β superfamily shares the same HEAT-repeat fold across many importins and exportins, and the analysis specifically pinpoints exportin-5 and rules out CRM1/XPO1 or the importins.

The D. melanogaster ortholog (CG12234; aliases dmExp5/Exp5/RanBP21) carries the same Exportin-5 C-terminal Pfam domain (PF19273) and has direct experimental support. PMID:16963774 reports: "we found that Drosophila exportin-5 binds pre-miRNAs and that amongst the exportin-5 orthologues tested, it shows the highest affinity for tRNAs. The knockdown of Drosophila exportin-5 in cultured cells decreased the amounts of tRNA as well as miRNA." Because Drosophila lacks a dedicated exportin-t, exportin-5 carries the tRNA-export load — a lineage-specific functional emphasis. The same paper confirms "Exportin-5, an evolutionarily conserved nuclear export factor belonging to the importin-beta family of proteins," matching B4MAQ2's IBN_N/exportin domain architecture. Its title — Exportin-5 orthologues are functionally divergent among species — is also the key caveat: family placement is robust, but quantitative cargo preferences vary and should be annotated "by similarity."

Finding 3 — The AlphaFold model confirms an all-α HEAT/ARM superhelical solenoid (karyopherin fold)

The AlphaFold DB model AF-B4MAQ2-F1 (v6, 1,238 residues) has a mean pLDDT of 82.9 (85% of residues > 70; 46% > 90), i.e., a confident model. Backbone φ/ψ dihedral analysis shows the structure is ~74% α-helical with negligible genuine β-sheet, and the molecule has a radius of gyration of 36.8 Å with an elongated principal-axis ratio (~1.9) — the signature of a curved, elongated superhelical solenoid built from stacked α-helical (HEAT/ARM) repeats. This is precisely the importin-β/karyopherin fold and is incompatible with a membrane protein (consistent with GRAVY 0.007 and no signal peptide) or a compact globular enzyme.

Structurally, this reinforces the localization logic: a karyopherin carries no classical NLS or signal peptide of its own. Its subcellular distribution is not encoded by an intrinsic targeting motif but is dictated by the Ran-GTP gradient — it binds cargo + Ran-GTP in the nucleus, translocates through the nuclear pore, and releases cargo upon Ran-GTP hydrolysis in the cytoplasm. Localization is an emergent, dynamic property of the transport cycle rather than a fixed compartment, which is exactly why a single "cytoplasm" CC term underserves the record.

Mechanistic Model / Interpretation

Exportin-5 operates as a directional cargo shuttle powered by the nucleocytoplasmic Ran-GTP gradient. The cycle, and how it maps onto cellular compartments, is:

        NUCLEUS (high Ran-GTP)                     CYTOPLASM (low Ran-GTP; RanGAP)
   ┌───────────────────────────────┐          ┌───────────────────────────────────┐
   │  Exportin-5 + Ran-GTP          │          │  Exportin-5 (apo) recycles back    │
   │        + cargo                 │          │       into nucleus                 │
   │  (pre-miRNA / tRNA / JAZ etc.) │  ── NPC ─▶  Ran-GTP hydrolysis (RanGAP)      │
   │        │                       │          │       │                            │
   │        ▼                       │          │       ▼                            │
   │  ternary export complex  ──────┼──────────┼─▶ complex disassembles;            │
   │                                │          │   cargo released into cytoplasm    │
   └───────────────────────────────┘          └───────────────────────────────────┘
            ▲                                            │
            └──────────── apo-exportin-5 returns ────────┘

Because the protein spends functionally essential time in both the nucleoplasm and the cytoplasm, and transits the nuclear pore complex, any single-compartment CC annotation is a partial description. The defining molecular action is export receptor activity (binding Ran-GTP via the IBN_N domain, recognizing cargo via the HEAT-repeat solenoid and the exportin-5 C-terminal domain), and the process it drives is RNA/protein export from the nucleus.

The following table maps the sequence/structure evidence onto the appropriate GO annotations:

Evidence Base

Computational provenance generated in this investigation (findings F001–F003): UniProt feature/domain parse (PF03810, PF19273, IPR045478, PANTHER PTHR11223:SF3); Kyte–Doolittle hydropathy GRAVY = 0.007 with no TM span; Needleman–Wunsch global alignment B4MAQ2 vs Q9VWE7 = 85.5% identity (1063/1244); k-mer paralog discrimination (highest to human XPO5); AlphaFold AF-B4MAQ2-F1 secondary-structure and geometry analysis (~74% α-helix, Rg 36.8 Å, axis ratio ~1.9, mean pLDDT 82.9).

GO Curation Implications

Lead recommendation (requires curator verification):

1. Retain GO:0005737 (cytoplasm) as a secondary CC annotation — factually correct (obligatory cytoplasmic phase of the transport cycle) but it must not stand alone or be the headline.
2. Add GO:0005634 (nucleus) as a CC annotation (already in the existing IEA set); the ProtNLM2 prediction must not narrow the record to cytoplasm only. Optionally consider GO:0005643 (nuclear pore) or GO:0005635 (nuclear envelope) for the translocation step.
3. Lead with molecular function: GO:0005049 (nuclear export signal receptor activity) and GO:0031267 (small GTPase/Ran binding) — these define the gene product.
4. Add biological process: GO:0006405 (RNA export from nucleus) and the more specific GO:0035281 (pre-miRNA export from nucleus) under ISS transfer; retain GO:0006611 (protein export from nucleus) "by similarity"; consider GO:0006409 (tRNA export) as an organism-specific lead.
5. Evidence codes: because no assay exists on the D. virilis protein, functional and process terms transferred from D. melanogaster Exp5 should carry ISS/ISO with a with/from reference to the ortholog, not IDA.

Do not finalize the record as "cytoplasm" only, and do not fall back to the uninformative "protein binding" — the evidence supports the specific export-receptor MF and nuclear-export BP terms above.

Mechanistic Scope

The immediate molecular activity tested is receptor-mediated nucleocytoplasmic transport: B4MAQ2/Exportin-5 directly binds Ran-GTP (via the IBN_N domain) and cargo (pre-miRNAs, tRNAs, and specific proteins such as JAZ), forming a ternary export complex that translocates through the nuclear pore and releases cargo upon Ran-GTP hydrolysis in the cytoplasm. This is the direct gene-product activity a curator should annotate.

Downstream and context-specific phenomena that must not be conflated with the core function include: the effect of XPO5 loss on miRNA biogenesis/tumor suppression (20951941, 21346411 — a downstream human-cancer consequence, not the D. virilis molecular function); the ARF6/GRP1 cytoplasmic hand-off and microvesicle loading of pre-miRNA cargo (31235936); and viral manipulation of the Ran cofactor (22593162). These illuminate the pathway but are not the gene product's own activity and should not be transferred as direct annotations to B4MAQ2.

Conflicts and Alternatives

• Is the prediction simply wrong? No. Cytoplasm is not contradicted — Exportin-5 genuinely occupies the cytoplasm. The issue is completeness/informativeness, not truth.
• Paralog confusion. The main family risk is misassignment within the karyopherin-β superfamily (CRM1/XPO1, IPO5, CSE1L, transportins), because IBN_N (PF03810) is shared across the whole superfamily. However, the exportin-5-specific C-terminal domain (PF19273), PANTHER SF3, the computed 85.5% full-length identity to Dmel Exportin-5, and negligible k-mer similarity to other paralogs make CRM1/importin confusion unlikely. Length (1,238 aa) also matches XPO5 (~1,204–1,241 aa).
• Species divergence. 16963774 explicitly reports that exportin-5 orthologues are functionally divergent among species, with Drosophila exportin-5 showing unusually high tRNA affinity (compensating for the absence of exportin-t). Cargo-level details from D. melanogaster should be transferred to D. virilis only "by similarity."
• Cross-species transfer is within-genus (D. virilis ↔ D. melanogaster), making functional/localization transfer low-risk. No isoform-specific or experimental-artifact conflicts were identified for B4MAQ2.

Limitations and Knowledge Gaps

1. No direct experimental data on B4MAQ2 itself. All conclusions are computational/sequence-based or transferred from orthologs. Why it matters: it distinguishes "cytoplasm-only" from "shuttling receptor." Resolution: GFP-fusion imaging or subcellular fractionation in D. virilis cells; in vitro Ran-GTP/pre-miRNA binding assay with recombinant B4MAQ2.
2. Steady-state distribution unknown for the fly protein. Mammalian XPO5 is often reported predominantly cytoplasmic at steady state; the nuclear/cytoplasmic ratio in flies is unverified. Resolution: quantitative imaging under Ran-gradient perturbation.
3. Quantitative cargo preferences unknown for D. virilis. Given documented cross-species divergence, the tRNA-vs-pre-miRNA balance may differ from D. melanogaster. Resolution: comparative binding/export assays.
4. AlphaFold provides fold, not localization. The structure confirms the karyopherin solenoid but cannot establish compartment occupancy. Resolution: localization must come from imaging/fractionation.

Discriminating Tests

1. Subcellular fractionation / live-cell imaging in D. virilis (or heterologous) cells with tagged B4MAQ2, under Ran-GTP gradient perturbation (RanGAP/RanGEF manipulation), to demonstrate active shuttling rather than static cytoplasmic residence — this directly distinguishes "cytoplasm-only" from "shuttling receptor."
2. In vitro reconstitution: recombinant B4MAQ2 + Ran-GTP + candidate cargo (pre-miR-30, tRNA) to confirm ternary-complex formation and export competence (mirroring 15356295).
3. RNAi knockdown in D. virilis-derived cells measuring mature miRNA and tRNA levels, replicating the D. melanogaster result (16963774) to confirm functional conservation.
4. Domain-swap / C-terminal truncation (removing the PF19273 region) to test the requirement of the exportin-5 C-terminal domain for cargo/Ran-GTP ternary complex formation, as implicated by cancer-associated truncations (21346411).
5. Heterokaryon shuttling assay to demonstrate nucleocytoplasmic cycling directly.

Proposed Follow-up Actions (Curation Leads)

All items are leads requiring curator verification.

• Action on the seed term: Retain GO:0005737 (cytoplasm) as secondary CC; do not accept it as the sole/primary localization or the headline annotation.
• Add CC: GO:0005634 (nucleus); evaluate GO:0005643 (nuclear pore) / GO:0005635 (nuclear envelope).
• Add/lead MF: GO:0005049 (nuclear export signal receptor activity); GO:0031267 (Ran/small GTPase binding).
• Add/lead BP (ISS, by similarity to Dmel Exp5, Q9VWE7): GO:0006405 (RNA export from nucleus), GO:0035281 (pre-miRNA export from nucleus), GO:0006611 (protein export from nucleus); consider GO:0006409 (tRNA export) as an organism-specific lead.
• Evidence code: ISS/ISO with with/from = D. melanogaster Exportin-5 (Q9VWE7 / CG12234) for functional terms; keep IEA where automated.
• Candidate references to verify (with exact snippets):
• 16963774 — "Drosophila exportin-5 binds pre-miRNAs … shows the highest affinity for tRNAs … knockdown … decreased the amounts of tRNA as well as miRNA" (Drosophila-specific functional support).
• 15254228 — "Exportin-5 is a nuclear export receptor for certain classes of double-stranded RNA … pre-micro-RNAs, viral hairpin RNAs, and some tRNAs" (core MF/BP).
• 15134074 — "exportin-5 (Exp5), a Ran-dependent importin-beta-related transport receptor, mediates nuclear export of miRNA precursors" (family/nucleocytoplasmic).
• Suggested question for the curator: Should the ProtNLM2 cytoplasm prediction be flagged as "correct but under-informative," with the record upgraded to the export-receptor function set (nucleus CC + export MF/BP) rather than accepting the CC prediction at face value?

Provenance / Analyses Run

• UniProt fetch of B4MAQ2: name, sequence (1,238 aa), features, Pfam/InterPro/PANTHER/SUPFAM cross-refs, GO (recorded above).
• Kyte–Doolittle hydropathy (window = 19): GRAVY 0.007, no continuous TM span → soluble protein.
• AlphaFold model AF-B4MAQ2-F1 (v6): mean pLDDT 82.9 (85% > 70, 46% > 90); backbone-dihedral analysis ≈ 74% α-helix with negligible genuine β-sheet, radius of gyration 36.8 Å, elongated principal-axis ratio 1.9 → all-α curved HEAT/ARM superhelical solenoid (canonical importin-β/karyopherin fold); no transmembrane helices.
• Full-length Needleman–Wunsch alignment B4MAQ2 vs Dmel Ranbp21/Exportin-5 (Q9VWE7): 1063/1244 = 85.5% identity; 5-mer containment 55.2%.
• k-mer (5-mer) family discrimination vs human karyopherin-β paralogs (XPO5 top; CRM1/IPO5/CSE1L/TNPO1/KPNB1 negligible); UniProt confirming Dmel PF19273 ortholog = Ranbp21/CG12234.
• PubMed evidence retrieval: PMIDs 16963774, 15134074, 15254228, 15356295, 20951941, 21346411, 31235936, 22593162.

Limitations: bioinformatic/homology-based; no wet-lab localization or function for the D. virilis protein itself.