AIGR Gene Hypothesis Deep Research — GEI-17 (Q94361): SAP/SIM-Mediated Ligase-Independent Transcription-Inhibitor Activity

Target: Caenorhabditis elegans GEI-17 (UniProt Q94361) · Term under evaluation: transcription regulator inhibitor activity (GO:0140416) · Focus: function_assignment · Hypothesis slug: gei17-sap-sim-coregulator

Executive Judgment

Verdict: REFUTED as a molecular-function mechanism / OVER-ANNOTATED as a GO MF term — despite a partially correct structural premise.

The seed hypothesis makes a three-part claim: (1) GEI-17 retains the PIAS N-terminal SAP domain; (2) it retains SUMO-interacting motifs (SIMs); and (3) these modules together confer a ligase-independent transcriptional co-repressor / transcription-regulator-inhibitor activity (GO:0140416) analogous to human PIAS1, operating independently of the SP-RING SUMO ligase. Our investigation finds that the structural premise is partly salvageable but the functional conclusion does not hold, for three convergent reasons:

1. The SAP premise is sequence-undetectable but structurally plausible. No sequence-profile method (Pfam PF02037, PROSITE PS50800, CATH-Gene3D, SUPERFAMILY) calls a SAP domain in GEI-17, whereas all three reference orthologs (human PIAS1, PIAS4, Drosophila Su(var)2-10) are annotated by multiple independent methods. However, AlphaFold shows GEI-17 retains a confidently folded, helix-hairpin-helix N-terminal module that superposes onto the PIAS1 SAP core at 1.55 Å Cα-RMSD. So a cryptic, highly diverged SAP-like fold is present — but the DNA-binding helix-2 and the LXXLL motif invoked by the hypothesis are degenerate (GEI-17 "LQSII" vs PIAS1 "LQVLL").

2. The SAP domain is the wrong module even in human PIAS1. Decisively, the ligase-independent transcription-inhibitor activity of human PIAS1 does not map to the SAP domain. A SAP-domain mutant and a catalytically dead (ligase-dead) PIAS1 both retain IRF3 inhibition, which instead maps to a C-terminal acidic region (PMID:24036127). The original human PIAS1 GO:0140416 basis (PMID:9724754) is STAT1 sequestration — a protein-protein mechanism, not SAP–DNA co-repression. So even if GEI-17's SAP-like fold were intact, SAP conservation would not establish GO:0140416.

3. GEI-17's only documented worm transcription-repressive role is ligase-DEPENDENT. GEI-17 does inhibit transcription in C. elegans — specifically piRNA transcription (PMID:40316696) — but this operates through its SUMO E3 ligase (SP-RING) activity and SUMOylation-driven condensate regulation, directly contradicting the hypothesis's "independent of its SP-RING ligase" claim.

Bottom line for the curator: GEI-17's GO:0140416 annotation is IBA-only (phylogenetic propagation from the PIAS/PANTHER node), with no experimental C. elegans support. The gene product's primary, experimentally-grounded molecular function is SUMO E3 ligase activity (GO:0061665). Any transcriptional repression GEI-17 mediates is best captured as a downstream biological process (e.g., negative regulation of transcription, GO:0045892) executed via its catalytic ligase activity — not as a direct, ligase-independent transcription-regulator-inhibitor molecular function. The most important caveat is that the structural SAP-like fold is genuinely present (so a curator should not claim "SAP is absent"); the point is that its presence does not license the MF term.

Key Findings

Finding 1 — GEI-17 lacks a sequence-detectable SAP domain, unlike all three reference orthologs

InterPro integrated annotation calls the SAP domain (IPR003034) in human PIAS1 (O75925, residues 11–45), PIAS4 (Q8N2W9, 12–46) and Drosophila Su(var)2-10 (Q7KNF5, 2–36) by multiple independent methods — Pfam PF02037, PROSITE PS50800, CATH-Gene3D G3DSA:1.10.720.30, and SUPERFAMILY SSF68906. For GEI-17 (Q94361), none of these methods call a SAP domain; only the PINIT domain (residues 203–367) and the SP-RING/MIZ zinc-finger (400–485) are detected. GEI-17 carries a long, largely disordered N-terminal extension — PINIT begins at residue ~203 versus ~119–124 in the orthologs.

Direct Smith–Waterman alignment (BLOSUM62) of the PIAS1 SAP core (residues 6–46) against each N-terminus quantifies the divergence: PIAS1-vs-PIAS4 scores 133 (63% identity); PIAS1-vs-Su(var)2-10 scores 61 (41% identity); PIAS1-vs-GEI-17 scores only 36 over a 27-aa window (37% identity). A shuffle null model (n=300 shuffles) gives GEI-17 an observed z = 3.93, p ≈ 0.003 — a statistically detectable but weak remnant, confined to SAP helix-1 (GEI-17 "VMKLRVHDLQSII" vs PIAS1 "VMSLRVSELQVLL"). Critically, the DNA-binding second helix ("RNKHGRKHELLTKALHLLK" in PIAS1) and the LXXLL motif are not conserved (GEI-17 "LQSII" vs PIAS1 "LQVLL"). This is the first and most direct refutation of the hypothesis's sequence premise: the specific DNA-contacting residues that the seed invokes are gone.

Finding 2 — GO:0140416 is IBA-only in GEI-17; the human basis is STAT1 sequestration, not SAP co-repression

QuickGO shows GEI-17 (Q94361) carries GO:0140416 (transcription regulator inhibitor activity) solely with evidence code IBA (GO_REF:0000033; phylogenetic propagation from the PANTHER PIAS-family node PTN000845825, family PTHR10782). There is no experimental (IDA/IMP/IPI) C. elegans annotation for this term. The human PIAS1 GO:0140416 annotation rests on an IDA from Liu et al. 1998 (PMID:9724754(https://pubmed.ncbi.nlm.nih.gov/9724754/)): "PIAS1, but not other PIAS proteins, blocked the DNA binding activity of Stat1 and inhibited Stat1-mediated gene activation in response to interferon." This is a protein-sequestration mechanism dependent on the PIAS1–STAT1 interaction — not on SAP–DNA binding or the LXXLL motif invoked by the seed. Moreover, C. elegans lacks a canonical interferon/JAK–STAT1 axis, so the specific documented mechanism underlying the human annotation does not transfer to the worm. The IBA propagation therefore carries a function whose experimental anchor is both mechanistically and organismally mismatched to GEI-17.

Finding 3 — GEI-17's SIMs are experimentally characterized for SUMO-scaffolding, not transcriptional repression

Pelisch et al. 2017 (PMID:27939944(https://pubmed.ncbi.nlm.nih.gov/27939944/)) show experimentally that "GEI-17 and another RC component, the kinase BUB-1, contain functional SUMO interaction motifs (SIMs), allowing them to recruit SUMO modified proteins, including KLP-19, into the RC [meiotic ring complex]." Our motif scan of GEI-17 recovers acidic-flanked hydrophobic SIM-like cores (e.g., residues 472–475 and regions near 545–548 and 597–600 within C-terminal acidic/disordered stretches), consistent with SIM presence. So the second structural premise of the hypothesis — that GEI-17 has SIMs — is correct. However, these SIMs function in a SUMO–SIM scaffolding network for chromosome congression during oocyte meiosis, operating downstream of GEI-17's SUMO E3 ligase activity — not in transcriptional co-repression. All experimentally characterized GEI-17 roles (meiotic ring-complex assembly, PMID:27939944; mitotic chromosome assembly / cell-cycle progression, PMID:25475837(https://pubmed.ncbi.nlm.nih.gov/25475837/)) are SUMO-ligase-dependent. The SIMs are real but repurposed for a catalytic/scaffolding pathway, not for the ligase-independent repression the seed proposes.

Finding 4 — AlphaFold rescues the SAP-fold premise (1.55 Å RMSD) even though sequence profiles miss it

AlphaFold models show the PIAS SAP region is confidently folded in all orthologs (PIAS1 res 11–45 mean pLDDT 93.7; PIAS4 12–46, 93.5; Su(var)2-10 2–36, 89.0). GEI-17 (Q94361) likewise has a confidently folded N-terminal module at res ~21–80 (mean pLDDT ~86–89; 100% of res 21–60 above pLDDT 70), flanked by a long disordered linker (res 81–200, mean pLDDT ~36) before PINIT (203+). CA(i,i+4) helix mapping shows GEI-17 res 20–85 forms helical runs [7,11,8,10] and PIAS1 res 6–50 forms [4,11,12] — both consistent with a SAP helix-hairpin-helix core. Kabsch superposition of the sequence-aligned Cα atoms (PIAS1 SAP residues 11–36 ↔ GEI-17 27–47, 26 residue pairs) gives Cα-RMSD = 1.55 Å.

This is the key nuance for the curator: GEI-17 retains a structurally SAP-like folded module even though Pfam/PROSITE/CATH/SUPERFAMILY do not classify it as SAP, because the sequence has diverged beyond profile-detection limits. The structural premise of the hypothesis is therefore partly vindicated — but with the same retained caveat: the DNA-binding helix-2 and the LXXLL motif are not conserved. A fold without its functional residues does not recover the function.

Finding 5 — Human PIAS ligase-independent transcription inhibition is SAP-INDEPENDENT, undercutting the seed's core mechanism

This is the decisive functional finding. Li et al. 2013 (PMID:24036127(https://pubmed.ncbi.nlm.nih.gov/24036127/)) dissected exactly the ligase-independent transcription-inhibitor activity the seed invokes and found it is neither ligase-dependent nor SAP-dependent:

• Ligase-independent: "SUMO E3 ligase activity dead mutant PIAS1/C350S still had the comparable inhibitory function with WT PIAS1."
• SAP-independent: "PIAS1 with a mutation in the SAP domain retained the inhibitory function in virus-induced IFN transcription."
• The responsible module is elsewhere: "The C-terminal region of PIAS1 around a cluster of acidic amino acids is critical for the interaction with IRF3 and the inhibitory functions of PIAS1."

Consistent with this, Tolkunova et al. 2007 (PMID:17991485(https://pubmed.ncbi.nlm.nih.gov/17991485/)) showed PIASy inhibits Oct4-mediated transcription by sequestering Oct4 to the nuclear periphery, and "These modes of PIASy action are uncoupled from its sumoylation activity." In other words, even where PIAS proteins do exhibit genuine ligase-independent transcriptional inhibition, the mechanism is transcription-factor protein-protein sequestration via C-terminal/acidic regions, not SAP-domain DNA binding or the LXXLL motif. The seed hypothesis's specific causal claim — that the SAP domain confers the activity — is directly contradicted in the very paralog it uses as its model.

Finding 6 — GEI-17 does inhibit transcription in C. elegans, but via ligase-DEPENDENT condensate regulation

Zhu et al. 2025 (PMID:40316696(https://pubmed.ncbi.nlm.nih.gov/40316696/)) provide the direct worm evidence that GEI-17 represses transcription — and show it is ligase-dependent. A screen for regulators of piRNA transcription "isolated the SUMO E3 ligase GEI-17 as inhibiting and the SUMO protease TOFU-3 as promoting piRNA transcription foci formation, thereby regulating piRNA production." The effect operates through SUMOylation/deSUMOylation of the USTC transcription complex and phase-separation of transcriptional condensates — i.e., GEI-17's SUMO E3 ligase (SP-RING) activity produces the transcription-inhibitory outcome (deSUMOylation by TOFU-3 promotes foci; SUMOylation by GEI-17 opposes them). This is the opposite of the hypothesis's "independent of its SP-RING ligase" claim. Reinforcing a ligase-dependent, transcription-factor-SUMOylation mode, Fergin et al. 2022 (PMID:35666766(https://pubmed.ncbi.nlm.nih.gov/35666766/)) show GEI-17-dependent "sumoylation of the ETS transcription factor LIN-1 at K169 is necessary for the proper contraction of the ventral vulA toroids" during vulval development — GEI-17 acting on a transcription factor by catalytic SUMOylation, not by ligase-independent co-repression.

Mechanistic Model / Interpretation

The investigation can be summarized as a decision tree in which the hypothesis is tested at three independent gates, failing the functional gates even where it passes a structural one:

SEED HYPOTHESIS: GEI-17 SAP + SIMs → ligase-INDEPENDENT
         transcription-regulator-inhibitor activity (GO:0140416)
  │
   ┌──────┴───────────────────────────────────────────────┐
   │ GATE 1: Does GEI-17 retain the SAP domain?            │
   ├───────────────────────────────────────────────────────┤
   │ Sequence profiles (Pfam/PROSITE/CATH/SUPERFAMILY): NO │  F001
   │ AlphaFold fold superposition (1.55 Å RMSD):   PARTIAL │  F004
   │ DNA-binding helix-2 + LXXLL residues:              NO │  F001/F004
   │ → Fold cryptically present; functional residues GONE  │
   └──────┬────────────────────────────────────────────────┘
  │ (even granting the fold…)
   ┌──────┴────────────────────────────────────────────────┐
   │ GATE 2: Does the SAP domain confer the activity        │
   │         in the model paralog (human PIAS1)?           │
   ├────────────────────────────────────────────────────────┤
   │ PIAS1 ΔSAP mutant RETAINS IRF3 inhibition:          NO │  F005
   │ PIAS1 ligase-dead (C350S) RETAINS inhibition:       NO │  F005
   │ Activity maps to C-TERMINAL ACIDIC region:    (not SAP)│  F005
   │ Original GO basis = STAT1 SEQUESTRATION:      (not SAP)│  F002
   │ → SAP is the WRONG module; mechanism is TF seq.       │
   └──────┬─────────────────────────────────────────────────┘
  │ (so is GEI-17's worm repression ligase-independent?)
   ┌──────┴─────────────────────────────────────────────────┐
   │ GATE 3: In C. elegans, is GEI-17 repression            │
   │         independent of its SP-RING ligase?           │
   ├─────────────────────────────────────────────────────────┤
   │ piRNA transcription repression = SUMO-ligase DEPENDENT │  F006
   │ LIN-1 regulation = catalytic SUMOylation (K169):  DEP. │  F006
   │ SIMs = SUMO scaffolding of meiotic RC (ligase pathway) │  F003
   │ → Repression is a DOWNSTREAM BP of catalytic activity  │
   └─────────────────────────────────────────────────────────┘

CONCLUSION: GO:0140416 (MF) NOT supported for GEI-17.
Primary MF = SUMO E3 ligase (GO:0061665).
Transcription repression = downstream BP (e.g., GO:0045892),
                   executed VIA ligase activity.

The coherent narrative: GEI-17 is, first and foremost, the C. elegans PIAS-family SUMO E3 ligase. Its experimentally documented biology — meiotic ring-complex assembly, mitotic chromosome assembly, telomere anchoring, vulval LIN-1 regulation, and piRNA condensate control — is uniformly ligase-dependent. The seed hypothesis attempts to graft onto GEI-17 a ligase-independent co-repressor function borrowed from human PIAS1, routed through the SAP domain and LXXLL motif. But that route is a misreading of the human data: even in PIAS1 the ligase-independent transcription-inhibitor activity is SAP- and ligase-independent, mapping to a C-terminal acidic TF-sequestration module. GEI-17 retains a cryptic SAP-like fold but not its DNA-binding residues, and its worm repression runs entirely through catalysis. The GO:0140416 annotation is thus an IBA artifact of phylogenetic propagation from a family node whose experimental anchor (human STAT1/IRF3 inhibition) neither depends on the conserved module nor exists in the worm's signaling repertoire.

Evidence Matrix

GO Curation Implications

Lead for curator verification — likely action: do NOT retain GO:0140416 as a direct molecular-function annotation for GEI-17; treat as non-core / downstream.

• GO:0140416 (transcription regulator inhibitor activity, MF): The evidence does not support this as a direct molecular function of GEI-17. It is IBA-only, propagated from the PIAS/PANTHER family node, whose experimental anchor (human PIAS1 STAT1/IRF3 inhibition) (i) is mechanistically a TF-sequestration activity mapping to a C-terminal acidic region rather than the conserved SAP domain, and (ii) does not depend on the module the seed invokes. Recommended action: remove or demote this MF annotation for GEI-17, or at minimum flag it as an uninformative IBA carry-over not supported by the worm's experimental record. If retained for phylogenetic-consistency reasons, it should carry a NOT/under-review qualifier pending experimental validation.

• Primary molecular function — GO:0061665 (SUMO ligase activity) / SUMO transferase: This is the experimentally supported MF and should be the anchor annotation (PMID:25475837; PMID:27939944; PMID:35666766; PMID:40316696 all rely on GEI-17 catalytic activity).

• Biological process — GO:0045892 (negative regulation of DNA-templated transcription) or a piRNA-transcription-specific child: GEI-17's documented transcription-repressive effect on piRNA genes (PMID:40316696) is real and supportable as a BP, but it is executed via SUMO ligase activity and condensate regulation — a downstream process, not a direct MF. A curator could add a BP term with IMP evidence from PMID:40316696, explicitly noting ligase dependence.

• Avoid "protein binding" as a terminal recommendation: The more informative supported terms are SUMO ligase activity (MF) and negative regulation of transcription (BP, ligase-dependent); these should be used rather than the uninformative generic binding term.

Mechanistic Scope

Immediate molecular activity being tested: a ligase-independent transcription-regulator-inhibitor (co-repressor) activity attributed to the SAP domain + SIMs. This direct activity is not supported. What is directly supported is SUMO E3 ligase (transferase) activity (SP-RING/MIZ domain, res 400–485), aided by PINIT (203–367) substrate positioning and C-terminal SIMs for SUMO-chain/scaffold engagement.

Separation of direct activity from downstream effects:
- Direct gene-product activity: catalytic SUMO conjugation onto substrate lysines (e.g., LIN-1 K169); SIM-mediated recruitment of SUMO-modified partners (KLP-19) into higher-order assemblies.
- Pathway/cellular consequence: SUMOylation of the USTC transcription complex modulates phase-separated transcriptional condensates, thereby repressing piRNA transcription — a downstream BP, not a direct MF of binding/inhibiting the transcription machinery.
- Developmental/phenotypic outcome: meiotic chromosome congression, mitotic chromosome assembly, telomere anchoring, vulval toroid contraction — pleiotropic phenotypes of the catalytic ligase, not evidence for a distinct co-repressor MF.
- Inferred-from-loss-of-function: the piRNA screen phenotype (PMID:40316696) demonstrates repression but attributes it to ligase activity; no loss-of-function data isolate a ligase-independent SAP-driven repression.

Conflicts and Alternatives

1. Paralog over-generalization / frequency bias. The hypothesis extrapolates from "human PIAS1 has GO:0140416" to "GEI-17 has it," but the human activity (a) does not require the SAP domain (PMID:24036127) and (b) originally rests on STAT1 sequestration (PMID:9724754), an axis absent in C. elegans. This is a classic IBA over-annotation pattern.

2. Fold vs. function mismatch. AlphaFold shows a conserved SAP-like fold (1.55 Å RMSD), which could tempt a curator toward the MF. But the DNA-binding helix-2 and LXXLL residues are degenerate, and the fold is not the functional determinant even in PIAS1. Structural conservation here is a red herring for GO:0140416.

3. SIMs real but repurposed. The hypothesis correctly predicts SIMs (PMID:27939944), but their documented role is SUMO-scaffolding of the meiotic ring complex — a ligase-pathway function — not transcriptional co-repression. Correct prediction, wrong functional attribution.

4. Organism-specific mechanism. C. elegans GEI-17 transcription repression (piRNA; PMID:40316696) is ligase-dependent, directly contradicting the seed's "independent of its SP-RING ligase." The worm mechanism is catalytic SUMOylation + condensate control, not DNA-binding co-repression.

5. Alternative interpretation favored by the data: GO:0140416's phenotypic signal in GEI-17 is fully explained as a downstream consequence of SUMO ligase activity, requiring no separate co-repressor MF.

Limitations and Knowledge Gaps

Discriminating Tests

1. Ligase-dead rescue in the native assay (highest value). Express catalytically dead GEI-17 (SP-RING Cys mutant) in a gei-17 loss-of-function background and test rescue of piRNA transcription repression (PMID:40316696 assay) and LIN-1-dependent vulval phenotypes (PMID:35666766). If repression requires catalysis, the ligase-independent hypothesis is refuted in vivo; if a dead mutant still represses, the hypothesis gains support.

2. SAP-module deletion/point mutants. Delete the AlphaFold-defined N-terminal SAP-like module (res ~21–80) or mutate its residual helix-1 residues; test for any loss of transcription repression separable from catalysis. Parallels the decisive PIAS1 ΔSAP experiment (PMID:24036127).

3. DNA/chromatin-binding assay of the isolated N-terminal module. EMSA or chromatin pulldown to test whether the cryptic SAP-like fold binds DNA at all, given helix-2/LXXLL degeneracy.

4. C-terminal acidic-region mapping. By analogy to PIAS1's IRF3-binding region, test whether a GEI-17 C-terminal fragment sequesters any germline transcription factor independent of SUMOylation.

5. Comparative orthology refinement. Formal structure-guided alignment (e.g., Foldseek/DALI) of GEI-17 vs. PIAS1/PIAS4/Su(var)2-10 SAP folds to quantify which functional SAP residues are retained vs. lost — strengthens the "fold-present, function-absent" curation note.

Curation Leads (require curator verification)

Candidate action change:
- Remove / demote GO:0140416 (transcription regulator inhibitor activity, MF) as a direct annotation for GEI-17; it is IBA-only with no worm experimental support and rests on a human mechanism (SAP- and ligase-independent TF sequestration) that neither transfers to the worm nor depends on the conserved module.
- Retain/strengthen GO:0061665 (SUMO ligase activity, MF) as the primary molecular function.
- Consider adding GO:0045892 (negative regulation of DNA-templated transcription, BP) with IMP evidence from PMID:40316696, annotated as ligase-dependent (downstream of catalytic activity).

Candidate references with exact snippets to verify:
- PMID:24036127(https://pubmed.ncbi.nlm.nih.gov/24036127/): "PIAS1 with a mutation in the SAP domain retained the inhibitory function in virus-induced IFN transcription" and "SUMO E3 ligase activity dead mutant PIAS1/C350S still had the comparable inhibitory function with WT PIAS1." → SAP- and ligase-independence of the human activity.
- PMID:9724754(https://pubmed.ncbi.nlm.nih.gov/9724754/): "PIAS1, but not other PIAS proteins, blocked the DNA binding activity of Stat1 and inhibited Stat1-mediated gene activation in response to interferon." → Human GO:0140416 basis is STAT1 sequestration.
- PMID:40316696(https://pubmed.ncbi.nlm.nih.gov/40316696/): "isolated the SUMO E3 ligase GEI-17 as inhibiting ... piRNA transcription foci formation." → Worm repression is ligase-dependent.
- PMID:27939944(https://pubmed.ncbi.nlm.nih.gov/27939944/): "GEI-17 ... contain functional SUMO interaction motifs (SIMs), allowing them to recruit SUMO modified proteins ... into the RC." → SIMs real, function is SUMO scaffolding.
- PMID:35666766(https://pubmed.ncbi.nlm.nih.gov/35666766/): "sumoylation of the ETS transcription factor LIN-1 at K169 is necessary for the proper contraction of the ventral vulA toroids." → Catalytic TF regulation.

Suggested curator questions:
- Is the GEI-17 GO:0140416 IBA propagation appropriate given the family node's experimental anchor is SAP- and ligase-independent and organism-mismatched?
- Should the worm piRNA repression phenotype be annotated as a BP with explicit ligase dependence rather than inherited as an MF?

Suggested experiments: ligase-dead rescue (test #1 above) and SAP-module mutant (test #2) are the two most decisive and should be flagged as the experiments that would settle the annotation.

Evidence Base Summary

The strongest primary-literature pillars, in order of decisiveness for this hypothesis:

1. PMID:24036127(https://pubmed.ncbi.nlm.nih.gov/24036127/) (Li et al. 2013) — decisive refutation of the SAP mechanism: human PIAS1 ΔSAP and ligase-dead mutants both retain transcription inhibition; activity maps to the C-terminal acidic region.
2. PMID:40316696(https://pubmed.ncbi.nlm.nih.gov/40316696/) (Zhu et al. 2025) — decisive worm evidence against ligase-independence: GEI-17 represses piRNA transcription via SUMO ligase activity and condensate regulation.
3. PMID:9724754(https://pubmed.ncbi.nlm.nih.gov/9724754/) (Liu et al. 1998) — the human GO:0140416 IDA basis is STAT1 sequestration, not SAP co-repression.
4. PMID:17991485(https://pubmed.ncbi.nlm.nih.gov/17991485/) (Tolkunova et al. 2007) — independent paralog confirms ligase-independent PIAS repression works by TF sequestration.
5. PMID:27939944(https://pubmed.ncbi.nlm.nih.gov/27939944/) (Pelisch et al. 2017) — GEI-17 SIMs are real but function in SUMO scaffolding of the meiotic ring complex.
6. PMID:35666766(https://pubmed.ncbi.nlm.nih.gov/35666766/) (Fergin et al. 2022) & PMID:25475837(https://pubmed.ncbi.nlm.nih.gov/25475837/) — reinforce that GEI-17's TF/chromosome roles are catalytic SUMO-ligase functions.

Database/computational evidence (InterPro profile methods; AlphaFold superposition) establishes the "fold present, functional residues absent" structural picture: GEI-17 escapes SAP profile detection yet retains a 1.55 Å-RMSD SAP-like fold with a degenerate DNA-binding helix-2/LXXLL motif.

Conclusion

The seed hypothesis is refuted as a molecular-function mechanism and over-annotated as GO:0140416, with a partially correct structural premise that must be stated carefully. GEI-17 does retain a cryptic, structurally SAP-like fold and genuine SIMs — but its DNA-binding residues are degenerate, the SAP domain is not the module that confers transcription-inhibitor activity even in human PIAS1, and GEI-17's only documented worm transcription-repressive role is SUMO-ligase-dependent. The curator should treat GO:0140416 as an unsupported IBA carry-over, anchor GEI-17's annotation on SUMO E3 ligase activity (GO:0061665), and, if desired, capture its transcription repression as a downstream, ligase-dependent biological process.

Artifacts

• OpenScientist final report
• OpenScientist final report