phzC1 (PA4212 / PA1901, Q7DC82) notes

2026-10-28 review (PseudoCAP IBA-seed worklist, rank 103)

No deep-research provider keys were available; literature gathered by PubMed
search and from the cached publications. The completed phzM and phzS reviews in
this repository were read for pathway context.

Identity

UniProt Q7DC82 (unreviewed, PE 3), 405 aa, "Phospho-2-dehydro-3-deoxyheptonate
aldolase" (DAHP synthase, EC 2.5.1.54), class-II DAHP synthase family, Pfam
PF01474 DAHP_synth_2, PANTHER PTHR21337:SF0
[file:PSEAE/phzC1/phzC1-uniprot.txt].
The entry covers both copies of the duplicated phenazine operon: gene names
phzC1 and phzC2, ordered locus names PA1901 and PA4212
[file:PSEAE/phzC1/phzC1-uniprot.txt]. The two proteins are identical, so GOA
rows and the PANTHER leaf (phzC1 = PA1901) refer to the same sequence.
PAO1 also has non-phenazine DAHP synthases, including PA2843 (Q9I000), which
matters for the propagation analysis below.
The enzyme catalyses the first step of the shikimate pathway: D-erythrose
4-phosphate + phosphoenolpyruvate -> 3-deoxy-D-arabino-heptulosonate
7-phosphate [file:PSEAE/phzC1/phzC1-uniprot.txt]. In the phz operon it
increases flux towards chorismate, the substrate of PhzE.

What PMID:11591691 (the seed reference) shows

Mavrodi et al. cloned both operons PMID:11591691 and showed
that each is sufficient in a heterologous host
PMID:11591691.
The evidence is therefore at the level of the operon: PseudoCAP applied the
same IDA row to phzA1, phzB1, phzC1, phzE1, phzF1, phzG1 and the phzA2/B2/G2
copies (QuickGO, reference PMID:11591691). Only the abstract is cached (PMC
has no usable full text for this article), so the individual-gene
contributions cannot be read off the paper. The conclusion that PhzC belongs
to the pathway does not rest on this paper alone, however.

Independent support for PhzC's role

The PAO1 phz-cluster DAH7PS has been purified and crystallised. It is kinetically
competent (KM 17 and 16 microM for PEP and E4P) and, unlike the housekeeping
enzyme, is not feedback-inhibited PMID:30242059. The same
work contrasts it explicitly with the housekeeping paralogue PA2843
PMID:30242059.
In the close relative P. chlororaphis, deleting the orthologous phzC collapses
phenazine output and complementation restores it
[PMID:35053084 "As shown in Figure 4, no morphological difference was found,
whereas phzC mutant strain HT66DphzC displayed a reduction in PCN
production."; PMID:35053084 "After the complementation of gene phzC in the
corresponding DAHP synthase-deficient strain, the production of PCN was
significantly restored."], while the housekeeping aroF/aroG/aroH deletions
barely affect it.

So, unlike several other seeds in this worklist, the seed claim itself is right:
PhzC performs a step - the committed entry into the shikimate pathway - whose
product is consumed by phenazine biosynthesis, and the enzyme is dedicated to
that role rather than to primary metabolism.

Where the propagation goes wrong

PTN000481936 is a Proteobacteria node of PTHR21337 with five leaves:
Helicobacter pylori HP_0134, Bradyrhizobium diazoefficiens blr3778,
Xanthomonas campestris dhs1, P. aeruginosa phzC1 (PA1901) and
P. aeruginosa PA2843. The GO:0002047 IBD at this node is sole-seeded by
Q7DC82, and all five leaves now carry phenazine biosynthetic process by IBA.
Four of them are housekeeping DAHP synthases of organisms, or of a paralogous
locus, with no phenazine pathway: H. pylori, Xanthomonas and
Bradyrhizobium are not phenazine producers, and PA2843 is the P. aeruginosa
primary-metabolism enzyme that the structural work above distinguishes from the
phz-cluster one. The DAHP synthase activity itself is asserted, correctly, at
the family root PTN000481914 (two seeds), so nothing is lost by removing the
phenazine IBD from the Proteobacteria node.

Annotation decisions

GO:0002047 (IDA, IBA): ACCEPT for this gene (core), because PhzC supplies the
pathway's own shikimate flux and the gene is part of the phenazine operon;
but a GO_CENTRAL defect is filed against the node, whose other four leaves
should not inherit the term. The IDA code is operon-level evidence, noted in
the review but not treated as a defect.
GO:0003849 (IBA from the family root, IEA from EC): ACCEPT, core activity.
GO:0009073 aromatic amino acid family biosynthetic process (IEA from
InterPro): MARK_AS_OVER_ANNOTATED - this paralogue is the secondary-metabolism
enzyme and is not feedback-regulated by aromatic amino acids; the primary
metabolism role belongs to PA2843 and the aro genes.

PseudoCAP IBA-seed assessment

PTN node	node term	n_seeds	leaf IBAs	seed annotation (term, evidence, PMID)	verdict	rationale
PTN000481936	GO:0002047 phenazine biosynthetic process	1 (sole)	5 GO_Central (5 GOA rows; no TreeGrafter)	GO:0002047, IDA, PMID:11591691	SOUND seed, WRONG node	The seed is right: PhzC is the phz-operon DAH7PS, it is kinetically characterised and allosterically unregulated, which marks it as a secondary-metabolism enzyme, and deleting the orthologue in P. chlororaphis collapses phenazine production. The caveat on the evidence code is that Mavrodi et al. tested whole operons, so the IDA is operon-level. The node, however, is a generic Proteobacteria DAHP synthase clade: besides phzC1 it contains Helicobacter pylori HP_0134, Xanthomonas dhs1, Bradyrhizobium blr3778 and - revealingly - P. aeruginosa's own housekeeping DAHP synthase PA2843, which structural work explicitly contrasts with the phz enzyme. All four now carry phenazine biosynthetic process by IBA. The term should sit on the phz-cluster enzymes only, not on this node.

Leaf evidence check (2026-11-06)

Audit of the GO_CENTRAL defect on PTN000481936. The load-bearing claim was
that leaf PA2843 (Q9I000) is P. aeruginosa's housekeeping type II DAH7PS,
distinguished from the phz-cluster enzyme by structural work. That claim is
backed by a real publication, not by a name, and the citation has been made
explicit. Leaf membership confirmed with panther_tree.py node PTN000481936 --leaves: 5 leaves, all PTHR21337:SF0.

Leaf	Name and its evidence	Experimental annotations found	Verdict
Q9I000 (P. aeruginosa PA2843, 448 aa)	`RecName` + EC 2.5.1.54 from UniRule RuleBase RU363071 (ECO:0000256) — i.e. a rule-derived name, not the basis of our claim; PE 1 (evidence at protein level)	No experimental GO rows (`GO:0003849` IBA, `GO:0002047` IBA, `GO:0009073` IEA). The structural/functional evidence is in the literature, not the GAF: PMID:30242059 names PA2843 by locus tag as `PaeDAH7PSPA2843`, states it "is not found within the phzA–G PCA biosynthetic cluster", reports its solved structure (PDB 5UXM, its ref 33) with the inserted alpha2a/alpha2b helices and Tyr/Trp inhibition, and concludes the type II enzymes "diverged to deliver type II enzymes that function within either primary or secondary metabolism"	SURVIVES, strengthened. Independent sequence check: Q7DC82 is 405 aa and Q9I000 448 aa, exactly the 43-residue difference the paper reports between the two, and Q7DC82 is the merged entry for both PA1901 and PA4212, so the characterised `PaeDAH7PSPA1901` is the seed protein
O24947 (Helicobacter pylori HP_0134, 449 aa)	`RecName` + EC from RuleBase RU363071; PE 3	None (IBA/IEA only)	Name not used as evidence. Retained facts: full-length type II by sequence length; no `phz` gene in H. pylori 26695 (UniProt gene query, 2026-11-06)
Q8PC31 (Xanthomonas campestris dhs1, 478 aa)	`RecName` + EC from RuleBase RU363071; PE 3	None (IBA/IEA only)	As above (ATCC 33913)
Q89NQ6 (Bradyrhizobium diazoefficiens blr3778, 462 aa)	`RecName` + EC from RuleBase RU363071; PE 3	None (IBA/IEA only)	As above (USDA 110)
Q7DC82 (phzC1 = PA4212/PA1901, 405 aa, the seed)	`RecName` + EC from RuleBase RU363071; gene names ECO:0000313	EMBL	PseudoCAP IDA (PMID:11591691); kinetics and structure in PMID:30242059

phz gene query (rest.uniprot.org, 2026-11-06, gene names
phzB/D/E/F/G/M/S): 0 hits in H. pylori 26695 (85962), 0 in X. campestris pv.
campestris ATCC 33913 (190485), 0 in B. diazoefficiens USDA 110 (224911),
against 6 in P. aeruginosa PAO1 (208964: phzD1, phzD2, phzF1, phzF2,
phzM, phzS). This is a gene-name query, so it is suggestive of cluster
absence rather than proof, and the entry says so.

Note that the RuleBase RecNames here are legitimate family evidence under the
repository's UniProt-quoting rules (UniRule is a curated system), which is why
membership in the type II DAH7PS family is accepted for all five leaves; what
is not taken from them is any statement about which pathway an individual
enzyme serves.

Confidence raised PROBABLE -> CONFIRMED: PMID:30242059 is cached full text, it
states the decisive facts outright, and the 43-residue length difference was
checked here against the two sequences.