UniProt: Q09436 (MEX6_CAEEL) · WormBase: WBGene00003231 · ORF: AH6.5 · 467 aa
Gene: mex-6 ("Muscle EXcess 6"). Paralog: mex-5 (Q9XUB2, WBGene00003230).
These are a working research journal for the AI GO-annotation review. Provenance is
recorded inline as [PMID:xxxxx "verbatim quote"]. All cached publications for this
gene are abstract-only (full_text_available: false in each publications/PMID_*.md),
so experimental IDA/IPI/IGI annotations are curated by WormBase from full text I cannot
see — I defer to the curator on those per project guidelines and use UNDECIDED only where
neither the abstract nor UniProt can corroborate.
mex-5 and mex-6 are "two nearly identical genes" PMID:10882103. Almost every mechanistic / biochemical study foregrounds mex-5:
- The RNA-recognition biochemistry (poly-U tract, low specificity, discriminator residue)
was done on MEX-5 PMID:17264081. There is no equivalent mex-6-specific biochemistry.
- The P-granule/phase-separation mRNA-competition mechanism (PMID:27594427) was demonstrated
for MEX-5, not mex-6.
Therefore, for mex-6 I treat shared-family/paralog properties as inference and flag them,
and I rely on the sources that name mex-6 explicitly for the strongest calls.
Redundant polarity/soma-germline function with mex-5. mex-5 and mex-6 together link
PAR cortical asymmetry to downstream cytoplasmic protein asymmetries; loss of both (not
either alone) produces the strong phenotype PMID:10882103.
UniProt summarizes: mex-6 "Functions with mex-5 to affect embryonic viability, establish
soma germline asymmetry in embryos and establish plk-1, pie-1, mex-1, and pos-1 asymmetry
in embryos ... Also affects formation of intestinal cells" (Q09436 CC FUNCTION, cites
PubMed:10882103, PubMed:18199581).
MEX-6 is one of two cytoplasmic CCCH-finger proteins acting in the establishment phase
of zygote polarity. PMID:12588843. MEX-6 localization dynamics were imaged by GFP fusion
alongside MEX-5 PMID:12588843.
MEX-6 binds polo kinases PLK-1 and PLK-2 via their polo-box domains (direct, mex-6
named). PMID:18199581. This underpins the IPI
annotations GO:0019901 (protein kinase binding) and GO:0019904 (protein domain specific
binding), curated by WormBase with with = plk-1 (WBGene00004042) and plk-2
(WBGene00004043). UniProt: mex-6 "Interacts (probably when phosphorylated on Thr-190)
with plk-1 (via POLO box domain) and plk-2 (via POLO box domain)" (Q09436 CC SUBUNIT).
The polo docking is primed by MBK-2/DYRK2 phosphorylation of a Thr near the fingers.
In MEX-5 the primed residue is T186; the paralogous residue in MEX-6 is Thr-190
(UniProt MOD_RES 190 "Phosphothreonine", MUTAGEN T190A "Severe reduction in binding to
plk-1 and plk-2"; MUTAGEN T190E phosphomimetic "severely abolishes interaction with plk-1
and plk-2"), both citing PubMed:18199581. The paper's abstract describes the mechanism on
MEX-5's T186 PMID:18199581, with
mex-6 assayed in the full text (UniProt MUTAGEN evidence is ECO:0000269|PubMed:18199581).
MEX-6 is required (redundantly with mex-5) for PLK-1 asymmetry. UniProt DISRUPTION
PHENOTYPE: "RNAi-mediated knockdown in mex-5 zu199 mutant background causes a loss in
plk-1 asymmetric distribution during the first embryonic cell divisions"
(ECO:0000269|PubMed:18199581). This is the sensitized (mex-5 null) background that exposes
the mex-6 contribution — the clearest statement of a mex-6 role beyond mex-5.
Supports GO:0032880 (regulation of protein localization), curated IGI with with = mex-5.
Subcellular location: cytoplasm. UniProt SUBCELLULAR LOCATION "Cytoplasm"
(ECO:0000250|UniProtKB:Q9XUB2, i.e. by similarity to mex-5). Also curated experimentally:
GO:0005737 cytoplasm IDA from PMID:12588843.
PTM (by similarity): "Phosphorylation on Ser-457 by par-1 promotes localization of the
protein to the anterior cytoplasm of the zygote" (Q09436 MOD_RES 457, ECO:0000250|
UniProtKB:Q9XUB2) — inferred from mex-5, consistent with an anterior gradient like mex-5.
MF (molecular function):
- GO:0035925 mRNA 3'-UTR AU-rich region binding (IBA) — MODIFY. Ancestral TTP/ZFP36
(AU-rich) specificity; nematode MEX proteins diverged (discriminator residue) and MEX-5
binds poly-U with low specificity, not AREs PMID:17264081.
Replace with the more accurate, already-curated GO:0003730 (mRNA 3'-UTR binding).
- GO:0000289 nuclear-transcribed mRNA poly(A) tail shortening (IBA) — MARK_AS_OVER_ANNOTATED.
Ancestral TTP deadenylation function (Reactome R-CEL-450513); no direct mex-6 deadenylation
evidence; MEX-5/6 literature is about polarity/translational control, not decay.
- GO:0000900 mRNA regulatory element binding translation repressor activity (IBA) — UNDECIDED.
Mechanism of MEX-5/6 translational control is not a clean sequence-specific repressor;
the best-supported translational output (zif-1 3'UTR) is positive (relieving POS-1
repression). No direct mex-6 evidence. (Mirrors mex-5 review.)
- GO:0005829 cytosol (IBA) — ACCEPT (cytoplasmic; consistent with IDA cytoplasm).
- GO:0160134 protein-RNA sequence-specific adaptor activity (IBA) — MODIFY to GO:0003729
mRNA binding. Binding is not sequence-specific (poly-U, low specificity in the paralog).
- GO:0003729 mRNA binding (IEA InterPro) — ACCEPT (core MF; CCCH TZF).
- GO:0046872 metal ion binding (IEA InterPro) — ACCEPT (zinc in CCCH fold).
- GO:0019901 protein kinase binding (IPI PMID:18199581) — ACCEPT (binds PLK-1/PLK-2).
- GO:0019904 protein domain specific binding (IPI PMID:18199581) — ACCEPT (polo-box).
- GO:0003730 mRNA 3'-UTR binding (IDA PMID:17264081) — ACCEPT, defer to curator. The
cited paper's abstract characterizes MEX-5, but WormBase curated this as a mex-6 IDA from
full text; mex-6 is a genuine mRNA/3'UTR-binding TZF protein. Core MF.
CC (cellular component):
- GO:0005737 cytoplasm (IEA SubCell) — ACCEPT.
- GO:0005737 cytoplasm (ISS from mex-5 Q9XUB2) — ACCEPT (ortholog transfer; also IDA-backed).
- GO:0005737 cytoplasm (IDA PMID:12588843) — ACCEPT (experimental, core location).
- GO:0043186 P granule (IDA PMID:12588843) — KEEP_AS_NON_CORE. Experimental WormBase IDA;
MEX-6 is predominantly anterior cytoplasmic (where P granules dissolve), so P-granule
association is at best transient/minor and not the core localization. I do NOT REMOVE an
experimental IDA whose full text I cannot read (unlike mex-5, mex-6 has no dedicated
phase-separation paper establishing it dissolves rather than resides in P granules).
- GO:0043186 P granule (IEA ARBA, GO_REF:0000117) — MARK_AS_OVER_ANNOTATED. Pure
machine-learning (ARBA) prediction of P-granule residency for an anterior-enriched TZF
protein; redundant with the IDA and exactly the kind of electronic over-propagation to
down-weight. Localization (if real) is already captured non-core by the IDA.
BP (biological process):
- GO:0017148 negative regulation of translation (IEA, logical inference from GO:0000900) —
UNDECIDED (inherits the uncertainty of GO:0000900; directionality doesn't match the
positive zif-1 output).
- GO:0032880 regulation of protein localization (IGI PMID:18199581, with mex-5) — ACCEPT.
mex-6 (redundantly with mex-5) establishes plk-1/pie-1/mex-1/pos-1 asymmetry
(UniProt CC FUNCTION; DISRUPTION PHENOTYPE for plk-1 asymmetry).
First attempt: just deep-research-falcon worm mex-6 --fallback perplexity-lite — the falcon
(Edison) provider timed out after 600s, and the perplexity-lite fallback returned HTTP 401
(insufficient_quota), so no deep-research file was produced on the first pass. A second
falcon-only retry was launched; the wrapper again reported a 600s timeout, but the underlying
Edison client did complete just after the timeout and wrote a genuine report
(mex-6-deep-research-falcon.md, Edison Scientific Literature, 33 citations, 636s duration,
+ _artifacts/). That file is retained as a committed research artifact. Its citations use
falcon's internal authorYYYY pages format (not PMIDs), and the validator does not check
non-PMID quotes, so — per project guidance — I did NOT cite the falcon file in the review;
every supporting_text in the review is instead a verbatim substring of a cached PMID abstract,
grep-verified. The falcon report is consistent with (and corroborates) the review: it independently
describes mex-6 as the partially redundant paralog of mex-5, the CCCH-TZF poly-U RNA binding, the
PAR-1/PP2A diffusion gradient, PLK-1/2 polo-box association, and the zif-1/ZIF-1 anti-germ-plasm
axis. No content in the review was taken uncritically from it, and nothing was fabricated.
Annotations that could not be corroborated from cached primary sources remain UNDECIDED
(GO:0000900, GO:0017148).
All existing annotations, authored NEW rows, description, core functions and references
were assessed. Source fields were preserved. Exact source-row identities and individual
outcomes are recorded in
projects/IBA_REVIEW/rereview-2026-09-20/germ-granule-rna-regulation.yaml.
The repository provider reports and the global OpenScientist cache were checked by
exact gene/accession. No target-specific OpenScientist adjudication was found; the
one incidental PGL/MEX name match is a WAGO-1 report, not an adjudication of these genes.
No duplicate provider request was launched by this reviewer.
The actual PTHR12547 topology places the exact target below PTN002648882 for all
five source IBAs. mex-6-paint-lineage.json preserves original assertions, lineage
and raw-response hash. Direct primary evidence establishes ARE binding by BOTH
paralogs (PMID:17264081) and tagged MEX-5/MEX-6 on germline granules
(PMID:12588843). Low selectivity relative to TTP does not mean absent ARE recognition.
The maternal Wnt paper (PMID:24131629) demonstrates mom-2 repression; positive zif-1
regulation is a different target context. MEX-5 biochemical binding is distinguished
from combined MEX-5/6 in vivo depletion. The primary-evidence note in mex-5 records
short indexed full-text passages and the blocked-download/abstract-cache limitations.
The Falcon reports contribute PLK recruitment, polarity and maternal RNA-turnover
context. The MEX-5 report's missing sequence-motif evidence is a retrieval limitation;
the MEX-6 report's MEX-5 structural results are not relabeled direct MEX-6 assays.
Tail-shortening and protein-RNA adaptor IBAs are retained as inherited process/activity
assertions. Absence of a purified target deadenylase assay is not a loss-of-function
argument: an RNA adaptor can contribute to a step catalysed by another protein.
The existing authored axis-specification proposal is retained because the proteins
help execute spatial RNA/protein patterning and feedback on PAR-domain growth;
this is distinct from their molecular binding annotations. No new proposals were added.