pilM (PA5044, G3XD28) notes

No deep-research provider was available; this is my own literature review
(PubMed MCP plus cached publications).

Identity

UniProt G3XD28 (PILM_PSEAE), 354 aa, Swiss-Prot reviewed; locus PA5044,
first gene of the pilMNOPQ operon. Matches the worklist accession.
Actin-like / FtsA-MreB fold ATP-binding protein (IPR005883 PilM,
TIGR01175, Pfam PF11104 PilM_2; PANTHER PTHR32432:SF3, a subfamily that
PANTHER names after EutJ and that also contains the ethanolamine
utilization protein EutJ).

Function

Identified in a twitching-motility transposon screen; pilM-P mutants lack
functional fimbriae (twitching, fimbria-specific phage, EM), and PilM shares
motifs with FtsA/MreB/actin ATPase domains PMID:7565110.
PilM/N/O/P form an inner-membrane alignment subcomplex; PilM, although
nominally cytoplasmic, co-fractionates with the inner membrane; loss of any
component destabilises the others and the PilQ secretin PMID:19857645.
The cytoplasmic N-terminal tail of PilN binds PilM PMID:23457250.
Crystal structures of P. aeruginosa PilM and PilM-PilN(1-12) chimera: PilM
binds ATP; PilN binding monomerises PilM and closes the domain around the
ATP site; PilM binds PilB and PilT directly (pull-down) and PilC by
two-hybrid PMID:27022027.
Thermus PilM binds ATP but no ATP hydrolysis was detected PMID:21596754;
Myxococcus PilM-N(1-16) binds but does not hydrolyse ATP PMID:26851283.
So PilM is an ATP-binding, not an ATPase, protein.
pilM deletion blocks PilNO disulfide heterodimer formation in vivo
PMID:27474743.

Summary: PilM is the cytoplasmic, ATP-binding, actin-like component of the
PilMNOP alignment subcomplex of the type IVa pilus machine. It docks on the
inner membrane through the PilN N-terminus and links the alignment subcomplex
to the motor (PilB, PilT, PilC). It is not a pilus fibre subunit.

GO term check

GO:0044096 type IV pilus: "A short filamentous structure on the surface of a
bacterial cell distinguished from other pili by post-translational
N-methylation of the pilin monomers." This is the fibre. PilM is
cytoplasmic/inner-membrane-associated and is not part of the fibre. There is
no GO CC for the type IV pilus assembly machine.
Comparator check (QuickGO, exact GO:0044096, experimental codes): IDA rows are
on pilins/minor pilins (pilA, fimU, pilV, pilW, pilX) and PilY1; the IMP rows
are PseudoCAP pilM, pilN, pilO, pilP (all PMID:7565110) and fimV. The
pilMNOP IMP rows all come from the same mutant screen, i.e. "mutant has no
pili" was annotated as a location.

PANTHER node PTN001286216

QuickGO IBA rows for both GO:0044096 and GO:0043107 from PTN001286216 (23
leaves each, WITH UniProtKB:G3XD28 only). Leaves include genuine PilM
orthologs (Neisseria NMB1808, Xanthomonas XCC3232, Shewanella SO_0281,
Geobacter GSU2032, Synechocystis, Deinococcus DR_0770, Vibrio VC_2634), but
also ethanolamine utilization protein EutJ of E. coli K-12 (P77277, b2454) and
Salmonella LT2 (P0A206, STM2462), an EutJ-named Fusobacterium protein
(Q8RI50), Listeria lmo1161 and other uncharacterised FtsA-like proteins (UniProt
REST, PANTHER PTHR32432:SF3 for all). EutJ belongs to the eut microcompartment
operon and is not a pilus protein, so the node is broader than the PilM clade:
both IBDs over-propagate to EutJ leaves.

PseudoCAP IBA-seed assessment

PTN node	node term	n_seeds	leaf IBAs	seed annotation (term, evidence, PMID)	verdict	rationale
PTN001286216	GO:0044096 type IV pilus	1 (sole)	23 (20 taxa)	GO:0044096 type IV pilus, located_in, IMP, PMID:7565110 (PseudoCAP)	WRONG	A mutant phenotype (no fimbriae) annotated as a location. PilM is a cytoplasmic, inner-membrane-docked component of the assembly machine, not part of the surface fibre the term defines. The IBD has no other support; it also reaches EutJ proteins.
PTN001286216	GO:0043107 type IV pilus-dependent motility	1 (sole)	23 (20 taxa)	GO:0043107, IMP, PMID:7565110 and PMID:19857645 (PseudoCAP)	SOUND (seed); node too broad	The seed is right for PilM: a core T4P machine component whose loss abolishes twitching. But the node includes E. coli and Salmonella EutJ and other non-pilus FtsA-like proteins, so the clade-wide claim over-propagates; GO_CENTRAL asked to move the IBD to a PilM-only node.

Decisions summary

ACCEPT: cytoplasm (EXP, IEA); type IV pilus-dependent motility (IBA, IMP x2);
type IV pilus assembly (IMP x2).
REMOVE: type IV pilus (IMP PseudoCAP; IBA); cell division (IEA, SHS2 FtsA
domain mapping).
NEW: ATP binding (IDA, PMID:27022027); cytoplasmic side of plasma membrane
(IC from PMID:19857645 fractionation + PilN-tail binding PMID:23457250).

Leaf evidence check (2026-11-06)

Audit of the leaf claims in the GO_CENTRAL source defects on PTN001286216.
Leaf membership and subfamily confirmed with
uv run python projects/PSEUDOCAP/panther_tree.py node PTN001286216 --leaves
(23 leaves, all PTHR32432:SF3); UniProt flat files fetched for all 23 accessions.

Discriminator found: the seed and 14 other leaves match InterPro IPR005883 PilM
with Pfam PF11104 PilM_2 (10 of those also NCBIfam TIGR01175 pilM); four leaves
match InterPro IPR013366 EutJ with Pfam PF14450 FtsA and none of the PilM
signatures; four match neither family. No leaf other than G3XD28 carries any
experimental GO annotation.

Leaf	Name and its evidence	Experimental annotations	Verdict
P77277 (E. coli b2454)	RecName "Ethanolamine utilization protein EutJ", curated (Reviewed); PE 3	none; FUNCTION by ECO:0000250 from P0A206. Lies in the E. coli eut operon (CC CAUTION)	Claim survives on domain + operon context, not on the name: IPR013366 EutJ, PF14450 FtsA, no PilM signature
P0A206 (S. typhimurium STM2462)	RecName "Ethanolamine utilization protein EutJ", curated; GN from ECO:0000303 PubMed:7868611; PE 2	no experimental GO rows, but curated ECO:0000269 FUNCTION/DISRUPTION PHENOTYPE from eutJ deletion and eutJ/eutS overexpression (PubMed:10464203, 16585748, 27063436), placing it in the 17-gene eut operon and its microcompartment	Claim survives, now on experimental curated statements
Q8RI50 (F. nucleatum FN1783)	SubName "Ethanolamine utilization protein eutJ" {ECO:0000313\|EMBL}; PE 3	none	Name is not evidence; the EutJ claim is downgraded to the signature match (IPR013366, no PilM signature; Pfam HSP70 rather than FtsA)
Q8Y7V8 (L. monocytogenes lmo1161)	SubName "Lmo1161 protein" {ECO:0000313\|EMBL}; PE 4 Predicted	none	Not a pilus protein on two checkable grounds: IPR013366 EutJ / PF14450 FtsA with no PilM signature, and gene neighbourhood (lmo1160 phosphate propanoyltransferase, lmo1164 corrinoid adenosyltransferase) places it in the B12-dependent catabolic microcompartment locus
A0A0H2W2G1 (Y. pestis YPO0146)	SubName "Uncharacterized protein"; PE 4	none	Carries IPR005883 PilM / PF11104, so the Enterobacteriaceae subclade PTN002089606 is mixed, not an "EutJ branch"

Correction to the earlier entry: the tree offers no PilM-only node for the
seed. The smallest clade holding G3XD28 with other PilM-signature leaves is
PTN001596234, which already contains P77277 and P0A206, and the
Enterobacteriaceae clade below it is mixed (above). The requested change is
therefore NOT/IKR on the four EutJ-signature leaves, not a move.

Confidence: GO:0043107 defect PROBABLE -> CONFIRMED; GO:0044096 defect
PROBABLE -> CONFIRMED (the decisive fact for the latter is G3XD28's own
GO:0005737 cytoplasm EXP:UniProtKB annotation plus PMID:19857645, independent
of any leaf name).