actC (Q93P57), Myxococcus xanthus — curation notes

Provenance of this file

No deep-research provider API keys exist in this container, so just deep-research
was not run and no -deep-research-<provider>.md file was created. Everything
below comes from the cached publications in publications/, the UniProt flat file,
Europe PMC / PubMed metadata, the UniProt and QuickGO REST APIs, and InterPro/PANTHER
records. Quotes are verbatim from the cited source.

GONUTS (gowiki.tamu.edu), the CACAO wiki that holds the student evidence pages, was
re-checked on 2026-11-01 as the CACAO project page asks. It is no longer merely
offline: https://gowiki.tamu.edu/wiki/index.php/Main_Page returns HTTP 410 Gone
with the page "GONUTS, the GO wiki, has been retired" ("The GO wiki that was at
gowiki.tamu.edu is no longer online."). The Wayback availability API has no snapshot
for MYXXA:ACTC, MYXXA:Q93P57, or the Gronewold/Kaiser reference page. The
student-evidence check is therefore permanently unavailable, not temporarily so, and
the CACAO rows below are judged from the papers alone under the curator-deference rule.

What the protein is

Q93P57 is a 437-residue unreviewed (TrEMBL) entry, PE 4: Predicted;, with no
functional UniProt comments. Its architecture is two-part:

residues ~1–270: an NAD(P)-binding Rossmann-like domain (Gene3D 3.40.50.720,
SUPFAM SSF51735, InterPro IPR036291 / IPR051783 "NAD(P)-dependent oxidoreductase-like").
residues 273–437: a GNAT (GCN5-related N-acetyltransferase) domain —
DOMAIN 273..437 with /note="N-acetyltransferase", plus
PROSITE; PS51186; GNAT; 1., Pfam; PF00583; Acetyltransf_1; 1. and
CDD; cd04301; NAT_SF; 1.

No catalytic activity has ever been assayed for ActC. There is no experimental
molecular-function datum in the literature; the whole experimental record is genetic.

The act operon and what the cited paper shows for actC

The one paper behind every CACAO row is PMID:11359579 (Gronewold & Kaiser 2001,
Mol Microbiol). It is abstract-only in our cache (full_text_available: false)
and is not open access (Europe PMC: isOpenAccess: N, inEPMC: N, no PMCID).

The abstract is nonetheless specific about actC:

Note what the abstract attributes to which gene. The sporulation phenotype named in
the abstract is actA's and actB's, not actC's:
PMID:11359579.
actC's stated result is a 15 h advance in the timing of CsgA production at an
unchanged maximum level.

Two later papers from the same authors pin down what the 2001 paper showed for actC
and are the basis for judging the two developmental rows:

PMID:11807078 (2002, abstract-only in cache): mutants in each act gene were
examined against four reporters.
PMID:11807078
and
PMID:11807078.
So actC shifts the timing of aggregation and sporulation gene expression; the
levels ultimately reached are wild type. Early, C-signal-independent development is
untouched:
PMID:11807078.
PMID:17189369 (2007, full text available) restates the 2001 paper's actC
morphological phenotype explicitly, citing it as reference 17:
PMID:17189369.
This is the sentence that grounds the fruiting-body row in the cited paper.
It also states the direction of ActC's effect:
PMID:17189369
and offers a mechanism hypothesis:
PMID:17189369

Why the sporulation claim is still defensible, if peripheral:
PMID:17189369 and
PMID:17189369.
A gene that sets when C-signal rises therefore does modulate when the sporulation
threshold is crossed — but no actC sporulation-efficiency defect is reported anywhere
we can read, and the 2002 reporters reach wild-type levels.

The operon has five genes, not four: PMID:17189369.
actE was described after the 2001 paper, whose abstract says "four genes".

Which entity performs the step

ActC's own action is on csgA expression: it restrains the rise of C-signal
production. Fruiting-body morphogenesis and sporulation are executed by the
C-signalling cascade downstream, not by ActC. The CACAO rows are nonetheless
regulation of terms, which is the correct GO device for a regulator that acts
through a signal, so the terms are not a "phenotype read as process" failure. What is
missing from GOA is the direct claim — negative regulation of csgA gene expression —
which is proposed as a NEW row.

Molecular-function rows: domain prediction, and one wrong propagation

GO:0016747, GO:0016407 and GO:0016740 all trace to the real GNAT match
(InterPro IPR000182, UniRule UR000414522 via PROSITE-ProRule PRU00532). The domain is
specifically an N-acetyltransferase signature, which GO:0008080 states and the
three annotated terms do not.

GO:0004029 aldehyde dehydrogenase (NAD+) activity (TreeGrafter, GO_REF:0000118,
PANTHER node PTN005180820) does not fit. NAD+-dependent aldehyde dehydrogenases
(EC 1.2.1.3) are the ALDH superfamily — Pfam PF00171 (Aldedh), SUPFAM SSF53720 — and
Q93P57 matches none of those; its NAD(P)-binding domain is the Rossmann-like
3.40.50.720 / SSF51735 fold used by short-chain dehydrogenases/reductases and
NAD-dependent epimerases/dehydratases. PANTHER itself places the protein in
PTHR48079 whose name is "PROTEIN YEEZ" (local interpro/panther/panther.obo) and
whose InterPro integration IPR051783 is described as a family of ketoreductases,
epimerases and dehydratases, with no aldehyde dehydrogenase among the members listed
in interpro/panther/PTHR48079/PTHR48079-entries.csv.

The propagation is not specific to this gene. A QuickGO query on
withFrom=PANTHER:PTN005180820 (2026-11-01) returns 2,648 annotations, every one
of them either GO:0004029 or GO:0005737, i.e. ~1,324 proteins receive this
aldehyde-dehydrogenase assertion from that single node. Recorded as a source_defects
entry against GO Central / PANTHER.

just paint-losses MYXXA actC reports "Q93P57 is not a leaf of this tree; losses not
checked. No IRD or IKR loss on this gene's lineage." (The row is a TreeGrafter graft,
not a PAINT IBA.)

The predicted failure shape: cross-isolate transfer — ABSENT

The screen flagged PMID:11359579 because its CACAO rows span two NCBI taxon ids,
NCBITaxon:34 (Myxococcus xanthus) and NCBITaxon:246197 (M. xanthus strain
DK1622), and the Gag-shaped hypothesis was that one experiment had been annotated onto
entries of a different isolate without re-grounding. It had not. Checked concretely:

Which strain the paper worked in. DK1622, the standard laboratory strain and the
reference genome. The 2001 paper's own GenBank submission carries
RC STRAIN=DK1622 on Q93P57, and the follow-up paper states that the submitted
sequence is the genome sequence: PMID:17189369
(CP000113 is the DK1622 genome). DK1622 is also the wild-type comparator throughout
PMID:17189369.
Which strain Q93P57 is from. DK1622, as above. The species-level taxon id on the
entry reflects how the authors deposited AF350253, not a different organism.
Whether the paper assays actC itself. Yes — the abstract reports the actC null
mutant's 15 h advance in CsgA production, and PMID:17189369 attributes the actC
fruiting-body phenotype to this same reference.

So the two taxon ids denote one organism and one locus, exactly the alternative the
screen design anticipated. No cross-isolate transfer; the shape is absent.

What is there instead: redundant UniProt entries fragmenting one operon's annotations

The taxon span is real but its cause is entry redundancy, and it has a consequence
worth recording. Each act gene exists twice in UniProtKB — once from the 2001
paper's own AF350253 submission at taxon 34, and once from the DK1622 genome
(CP000113) at taxon 246197:

gene	taxon 34 entry (AF350253)	taxon 246197 entry (DK1622 proteome)	CACAO rows landed on
actA	Q93P59 (342 aa, `GN actA`)	Q1D7F9 (342 aa, `GN actA MXAN_3213`)	Q93P59
actB	Q93P58 (553 aa, `GN actB`)	Q1D7F8 (553 aa, `GN actB MXAN_3214`)	Q1D7F8
actC	Q93P57 (437 aa, `GN actC`)	Q1D7F7 (437 aa, no gene name, `MXAN_3215`)	Q93P57
actD	Q93P56	MXAN_3216	—
actE	Q93P55	MXAN_3217	—

Verified by fetching both sequences: Q93P57 and Q1D7F7 are byte-identical (437 aa,
same md5 of the sequence) and share the same RefSeq protein WP_011553262.1. They are
the same protein of the same strain under two accessions and two taxon ids.

Three consequences:

The three CACAO rows for one paper are split across both entry sets — actA and actC
on the taxon-34 entries, actB on the DK1622 entry. Nothing was transferred between
organisms; the curators simply reached whichever entry carried the gene symbol
(actB is a gene name only on the DK1622 entry; actC is a gene name only on the
taxon-34 entry, since Q1D7F7 has no GN Name).
The experimental annotations sit on the entry that is not in the reference
proteome. Q1D7F7 carries DR Proteomes; UP000002402; Q93P57 carries none. So
Q1D7F7 — the ActC that appears in the DK1622 proteome, in STRING, KEGG and
EnsemblBacteria — has only the three IEA rows (GO:0005737, GO:0016747,
GO:0004029, checked 2026-11-01) and none of the experimental developmental
annotations.
The defect is UniProt's, not CACAO's: an obsolete-style redundant pair that should be
merged. Recorded as a source_defects entry against UniProt on the
GO:0031155 row.

Summary of judgements

term	code	action	one-line reason
GO:0031155 regulation of reproductive fruiting body development	IMP	ACCEPT	ΔactC gives premature formation of many small fruiting bodies, attributed to this paper by PMID:17189369
GO:0009966 regulation of signal transduction	IMP	MODIFY → GO:0009968	sign is known: ActC restrains the rise of C-signal production
GO:0043937 regulation of sporulation	IMP	KEEP_AS_NON_CORE	real but downstream and timing-only; the abstract's sporulation phenotypes are actA's and actB's
GO:0004029 aldehyde dehydrogenase (NAD+) activity	IEA	REMOVE	wrong fold; over-propagated from PTN005180820 to ~1,300 proteins
GO:0005737 cytoplasm	IEA	ACCEPT	soluble protein, no TM or signal peptide; right granularity for a bacterium
GO:0016407 acetyltransferase activity	IEA	MODIFY → GO:0008080	GNAT domain is an N-acetyltransferase signature
GO:0016747 acyltransferase activity, transferring...	IEA	MODIFY → GO:0008080	same domain, same refinement
GO:0016740 transferase activity	IEA	TOO_GENERAL	near-root term entailed by GO:0008080
GO:0010629 negative regulation of gene expression	IMP	NEW	ActC's own action: it inhibits the rise of csgA expression