lpxA (PA3644, Q9X6P4) notes

Identity

Acyl-[acyl-carrier-protein]--UDP-N-acetylglucosamine O-acyltransferase
(UDP-GlcNAc acyltransferase), EC 2.3.1.129, 258 aa, homotrimeric
left-handed beta-helix of hexapeptide repeats (TIGR01852 lipid_A_lpxA,
IPR010137, HAMAP MF_00387, PANTHER PTHR43480). Swiss-Prot annotation is
entirely rule-based (ECO:0000255|HAMAP-Rule:MF_00387).
Catalyses the first, committed step of lipid A synthesis: transfer of a
(3R)-hydroxyacyl chain from ACP to the 3-OH of UDP-GlcNAc
[file:PSEAE/lpxA/lpxA-uniprot.txt "Glycolipid biosynthesis; lipid IV(A) biosynthesis; lipid IV(A)"]

The P. aeruginosa PAO1 lpxA gene was cloned by expression in E. coli and the
enzyme characterised; it is acyl-chain-length-specific for C10 rather than
the C14 of E. coli LpxA
PMID:9440522
PMID:9440522
PMID:9440522
and the lipid A made in that strain carries 3-hydroxydecanoate at positions 3
and 3'
PMID:9440522
This paper is cited in the UniProt entry (reference [1]) but is not used in
GOA.
The product specificity matches the native P. aeruginosa lipid A, whose
3 and 3' positions carry 10:0(3-OH)
PMID:1904818

Kulshin et al. 1991, Eur J Biochem: a chemical/NMR/MS structural
characterisation of P. aeruginosa lipid A from two wild-type strains and the
rough mutant PAC 605
PMID:1904818
The organism is right, but the abstract mentions no lpxA, enzyme assay or
lpxA mutant: it describes an analysis of the purified glycolipid, published
before the gene was cloned (1991 vs 1998) and before the PAO1 genome. Only
the abstract was read (no PMC copy, 2026-10-29), so it is probably, not
certainly, not a direct assay of PA3644; the WRONG_REFERENCE defect was
downgraded from CONFIRMED to PROBABLE in the 2026-10-29 audit.
The claim is nevertheless correct: LpxA catalyses the first step of lipid A
synthesis, lipid A is the LPS anchor, and the Pseudomonas enzyme's C10
specificity was demonstrated directly in PMID:9440522. So the right fix is
to repoint the reference (and the paper that is cited reinforces the
product: 10:0(3-OH) at 3 and 3').

PTHR43480 (268 members, all in subfamily SF1), entirely Gram-negative
bacteria: Enterobacteriaceae, Burkholderia, Pseudomonas, Rickettsia,
Brucella, Chlamydia, Campylobacter, Helicobacter, Vibrio, Xanthomonas,
Neisseria, Rhizobium and so on.
QuickGO (2026-10-28): the node propagates two terms, GO:0009103 and
GO:0005829, each to 24 leaves; the GO:0009103 leaves are lpxA genes of
E. coli, Helicobacter, Chlamydia, Vibrio, Bacteroides, Thermotoga,
Bradyrhizobium, Shewanella and others, i.e. exactly the lipid-A-making
clade.
The MF IBA on this gene comes from a different node, PTN000602373, whose
seeds are E. coli LpxA and the Arabidopsis locus AT4G29540, i.e. the
acyltransferase activity is placed deeper than the bacterial LpxA clade.
E. coli LpxA (P0A722) carries GO:0009245 lipid A biosynthetic process and
GO:0008780 by IDA (PMID:3549716) but has no experimental GO:0009103, which
is why the LPS IBD rests on the PseudoCAP row alone.
Note that GO:0009245 lipid A biosynthetic process is not a descendant of
GO:0009103 (it sits under lipooligosaccharide/glycolipid biosynthesis), so
the two are separate claims rather than a general/specific pair; both are
true of LpxA, and the lipid A term is the more mechanism-specific of the two.

PTN node	node term	n_seeds	leaf IBAs	seed annotation (term, evidence, PMID)	verdict	rationale
PTN000602377	GO:0009103 lipopolysaccharide biosynthetic process	1 (sole seed)	24 (22 taxa)	GO:0009103, IDA, PMID:1904818 (PseudoCAP)	SOUND (term and node); reference WRONG	The term is correct for LpxA and for the whole node: PTHR43480 is the LpxA family of Gram-negative bacteria, every member catalyses the first step of lipid A synthesis, and lipid A is the LPS anchor, so an LPS-biosynthesis IBD suits the clade. The cited paper, however, is a 1991 structural study of P. aeruginosa lipid A whose abstract mentions no enzyme assay and no lpxA; it predates the cloning of the gene, so the IDA is probably not direct evidence for PA3644 (abstract only read). The direct evidence exists in PMID:9440522, where the cloned PAO1 lpxA is assayed as a 10-carbon-specific UDP-GlcNAc acyltransferase and complements an E. coli lpxA mutant. Recorded as a PROBABLE WRONG_REFERENCE PseudoCAP defect; the IBD itself stands, and would be sharper as GO:0009245 lipid A biosynthetic process, for which E. coli LpxA already has an IDA.

ACCEPT: acyl-ACP--UDP-GlcNAc O-acyltransferase activity (IBA, IEA), lipid A
biosynthetic process (IEA), LPS biosynthetic process (IBA, IDA with a
repointed reference), cytosol (IBA).
MODIFY: cytoplasm -> cytosol; lipid biosynthetic process -> lipid A
biosynthetic process; transferase activity -> acyl-ACP--UDP-GlcNAc
O-acyltransferase activity.
No REMOVE. One PSEUDOCAP WRONG_REFERENCE defect.