gei-17 (C. elegans, UniProt Q94361) - curation notes

Sources used

• gei-17-uniprot.txt (Swiss-Prot entry GEI17_CAEEL, 780 aa, six isoforms)
• gei-17-deep-research-openscientist.md (not edited; claims checked below)
• Cached primary papers (full text unless noted): PMID:15654100, PMID:16549501,
  PMID:19111656, PMID:22761594, PMID:25475837, PMID:25873636, PMID:27939944,
  PMID:24297748; abstract only: PMID:16701625, PMID:40316696.
• Fetched with just fetch-pmid during this review: PMID:34003111 (PIE-1 SUMOylation),
  PMID:34003109 (HDAC1 SUMOylation), PMID:31243051 (BUB-1/CLS-2 in meiosis),
  PMID:35666766 (tissue-specific SUMO inhibition, vulva), PMID:17908915 (SMK-1/PPH-4.1),
  PMID:40316696 (piRNA condensates; abstract only).

Identity and domains

• PIAS family; sole PIAS/Siz-type E3 in the worm, co-ortholog of human PIAS1/PIAS2
  (PANTHER PTHR10782:SF94) and fly Su(var)2-10.
• PINIT domain aa 203-367; SP-RING zinc finger aa 400-485 with Zn(2+) ligands at
  431/433/454/457 [file:worm/gei-17/gei-17-uniprot.txt "L->A: Greatly reduces E3 ligase activity and"]
• C-terminal SIMs (aa 423-602 of isoform f) bind SUMO-modified proteins
  PMID:27939944

Molecular function: SUMO E3 ligase (strong, multiple IDA)

• MUS-101/TopBP1 SUMOylation in vitro PMID:15654100
• POLH-1 SUMOylation in vitro and in vivo PMID:19111656
  PMID:19111656
• AIR-2 SUMOylation, SUMO chain formation, SP-RING L417A mutant
  PMID:25475837
• KLP-19 SUMOylation, auto-SUMOylation PMID:27939944
• BUB-1 SUMOylation dose-dependent PMID:31243051

Biological roles

DNA damage tolerance / TLS (core)

• GEI-17 SUMOylates POLH-1 at damaged chromatin and protects it from CRL4-Cdt2 degradation
  PMID:19111656
• Suppression by cdt-2 co-depletion; GEI-17 is not needed for POLH-1 catalysis per se
  PMID:19111656
• Epistatic with polh-1;polk-1 PMID:22761594
• Note: PMID:22761594 twice miscalls gei-17 a "SUMO protease"; this is a slip in that paper.

Checkpoint silencing in embryos (indirect consequence)

• PMID:16549501
• Mechanism is suppression of fork stalling, i.e. removal of the activating signal
  rather than action on ATL-1/CHK-1: PMID:16549501
  -> keep GO:1904290 as non-core.

Mitotic and meiotic chromosome segregation

• Mitosis: PMID:25475837
  GEI-17 at metaphase plate PMID:25475837
• Oocyte meiosis: GEI-17 assembles the midbivalent ring complex by SUMO conjugation + SIM
  binding PMID:27939944
  PMID:27939944
• Anaphase I: BUB-1 and CLS-2 dynamics PMID:31243051
• Possible NEW for mitotic metaphase chromosome alignment (GO:0007080) not proposed:
  the mitotic alignment phenotype is described mainly for UBC-9/SUMO; kept as a question.

Transcriptional repression (non-core)

• tbx-2 autorepression requires SUMO pathway incl. GEI-17 PMID:25873636
• TBX-2 Y2H interaction PMID:25873636
• GEI-17-dependent SUMOylation of TBX-2 itself in worms was not shown; deep research
  overstates "GEI-17 SUMOylates ... TBX-2".
• Vulva: LIN-1 K169 SUMOylation is required for vulA contraction (PMID:35666766); AID of
  GEI-17 phenocopies SUMO loss in VPCs.

Germline piRNA/nuclear silencing (redundant with PIE-1)

• gei-17 nulls alone do not desilence the piRNA sensor; with pie-1[K68R] they cause
  complete desilencing and sterility PMID:34003111
• HDA-1 SUMOylation/NuRD formation depend redundantly on PIE-1 and GEI-17
  PMID:34003109
• GEI-17 inhibits piRNA transcription foci (PMID:40316696 abstract only).
• No GOA annotation currently; not proposed as NEW: genetic redundancy evidence and
  no direct demonstration that GEI-17 (rather than PIE-1) SUMOylates HDA-1 in vivo.
  Raised as a suggested question.

Nuclear organization

• Telomere anchoring at the nuclear periphery in embryos requires GEI-17
  PMID:24297748
  Substrate unknown; no NEW proposed.

Deep research (openscientist) check

Verified against primary papers: E3 activity (PMID:15654100, 19111656, 25475837, 27939944),
POLH-1 protection (19111656), checkpoint silencing (16549501), meiotic ring complex
(27939944, 31243051), telomeres (24297748), LIN-1 K169 (35666766). Overstatements: TBX-2
SUMOylation by GEI-17 (only Y2H + RNAi). Omission: PIE-1/HDA-1 germline silencing
(PMID:34003111, PMID:34003109).

IBA notes

• JAK-STAT (GO:0046426) and transcription regulator inhibitor activity (GO:0140416) IBAs
  come from a PIAS1/Su(var)2-10 node (PTN000845825). No gei-17 evidence for STAT
  inhibition. GO:0046426 is REMOVED (C. elegans has no JAK, PMID:28874466; recorded as a member
  exception in the PTHR10782 family review); GO:0140416 is marked over-annotated.

Literature review integration (2026-10-04)

• Liongue & Ward 2013 (PMID:24058787) added to the GO:0046426 IBA REMOVE. The canonical JAK-STAT pathway was
  assembled in the bilaterian ancestor PMID:24058787,
  so the eumetazoan PIAS node (PTN000845825) is a sound IBD placement, and nematode JAK absence is a
  lineage-specific loss that needs an IRD.
• Palvimo 2007 (PMID:18031232, abstract only) added for PIAS-family background PMID:18031232.
  It also says PMID:18031232.
  This qualifies the GO:0140416 MARK_AS_OVER_ANNOTATED reasoning, which rests on GEI-17 transcriptional
  effects being SUMOylation-dependent. The action is unchanged, but this is worth revisiting.
• No actions changed.

Cross-gene briefing for GO editors and PAINT curators: https://claude.ai/artifact/4MLfbbbWWnLchDvspwdPuM

OpenScientist: ligase-independent repression via SAP/SIMs? (2026-10-04)

Report: gei-17-hypotheses/gei17-sap-sim-coregulator/openscientist.md (neutral framing; the
review's MARK_AS_OVER_ANNOTATED on GO:0140416 was withheld from the run).

• Verdict: refuted as a mechanism for GO:0140416; the over-annotation call stands.
• Checked against repo data before wiring: GEI-17's UniProt entry lists PINIT and the SP-RING
  zinc finger but no SAP domain; human PIAS1's lists IPR003034 SAP_dom and PS50800.
• Human PIAS1 ligase-independent inhibition is SAP-independent PMID:24036127 and ligase-independent PMID:24036127; for PIASy and Oct4
  PMID:17991485.
• Worm repression is ligase-dependent PMID:40316696.
• From the run, not re-derived: a cryptic SAP-like fold may remain in the N-terminus (AlphaFold),
  but the DNA-binding helix is not conserved. This resolves the Palvimo 2007 caveat recorded
  earlier: RING-independent co-regulation by PIAS proteins does not run through the SAP domain.
• Lead, not acted on: GEI-17's repression of piRNA transcription could be captured as a process
  term downstream of its ligase activity (PMID:40316696, abstract only). Decisive experiment: a
  ligase-dead GEI-17 rescue in the piRNA transcription assay.

Correction: GO:0140416 is UNDECIDED, not over-annotated (2026-10-05)

The previous section, and the MARK_AS_OVER_ANNOTATED action it supported, misread the term.
GO:0140416 transcription regulator inhibitor activity is defined as inhibiting a transcription
regulator "via direct binding and/or post-translational modification", so inhibition through
GEI-17's SUMO ligase activity is within the term. The OpenScientist run tested a narrower claim,
ligase-independent repression through the SAP domain and SIMs, and its refutation applies only
to that route. It does not refute the term.

• Binding route through the SAP domain: not supported (no SAP domain in GEI-17; PIAS1's
  ligase-independent inhibition does not need it, PMID:24036127).
• Modification route: plausible. GEI-17 inhibits piRNA transcription condensates built by the
  USTC complex (PRDE-1, SNPC-4, TOFU-4, TOFU-5) in a SUMOylation-dependent way PMID:40316696, but the cached abstract does not show direct SUMOylation
  of a USTC component, and the full text is unavailable.
• Action changed to UNDECIDED; propagation_review root cause back to UNRESOLVED. The PTHR10782
  family review's node text was updated to match. Prompted by the ai4c-reviewer finding on
  ai4curation/ai-gene-review#4199, whose suggested fix (a member exception or REMOVE) would
  have hardened the misreading.
• Decisive test: whether GEI-17 SUMOylates PRDE-1 or SNPC-4 directly, and whether that SUMOylation
  is what blocks condensate formation.

Addendum: the binding route is not excluded either (2026-10-05)

The re-review on ai4curation/ai-gene-review#4199 pointed out that the donor activity in
PMID:24036127 is ligase- and SAP-independent and maps to "the C-terminal region of PIAS1 around
a cluster of acidic amino acids", so the sharper test is whether GEI-17 keeps that module.
GO:0140416's definition was confirmed against QuickGO: "A molecular function regulator that
inhibits the activity of a transcription regulator via direct binding and/or post-translational
modification."

• A positional scan (gei-17-bioinformatics/acidic_clusters.py, RESULTS.md) finds PIAS1 has
  one acidic cluster (467-476, SSDEEEEEPS, next to its SIM VIDL), 62 aa past the SP-RING. GEI-17
  has a SIM-like core followed by an acidic run (IITLSDDDDEEL, 545-556) 63 aa past its SP-RING.
  The module is positionally conserved; an earlier eyeball reading that it sat elsewhere compared
  it with GEI-17's extreme C-terminus, which PIAS1's cluster is not at either.
• So neither route is excluded. Decisive tests: (1) whether GEI-17 SUMOylates PRDE-1 or SNPC-4
  directly and that blocks condensate formation; (2) whether a ligase-dead GEI-17 (e.g. the
  L417A SP-RING mutant) still binds and inhibits a transcription regulator, and whether that
  needs the 545-556 SIM-acidic module.
• The stale ROLE_CONFLATION failure mode, left over from the retracted argument, was removed
  from the GO:0140416 propagation review.