PslC is a predicted glycosyltransferase that adds one sugar of the Psl repeat
unit. Its role in biofilm formation is through Psl synthesis (biofilm formation
by definition includes production of the matrix polymers, so the claim is true
but derived). The process it directly performs is polysaccharide (Psl)
biosynthesis; GO has no Psl-specific term, so GO:0000271 is proposed as NEW (IMP).
| PTN node | node term | n_seeds | leaf IBAs | seed annotation (term, evidence, PMID) | verdict | rationale |
|---|---|---|---|---|---|---|
| PTN002463206 | GO:0044010 single-species biofilm formation | 1 (sole seed) | 7 (5 taxa) | GO:0044010 single-species biofilm formation, IDA, PMID:19659934 (PseudoCAP) | INDIRECT-PHENOTYPE | The paper shows a pslC deletion abolishes Psl and surface attachment: a mutant phenotype (should be IMP), and biofilm formation is the downstream outcome of Psl synthesis, not something conserved in the GT clade. The node's other leaves are glycosyltransferases for unrelated glycans: Salmonella RfbN O-antigen rhamnosyltransferase (P26403), Bradyrhizobium ExoU (Q89D69, succinoglycan-type), P. aeruginosa FgtA flagellin glycosyltransferase (Q9I4N8), PA1390 (Q9I3V4), Geobacter Q74CL7 and Xanthomonas Q8P4N7. Biofilm formation does not suit a clade of GT-2 enzymes with diverse acceptors. The IBD should be removed; the safe node-level claim is glycosyltransferase activity. |
Node leaves were taken from the PAINT leaf GAF (.cache/paint/leaf.gaf.gz).
Audit of the leaf claims in the GO_CENTRAL source defect on PTN002463206.
Membership confirmed with
uv run python projects/PSEUDOCAP/panther_tree.py node PTN002463206 --leaves
(7 leaves; PslC and RfbN in PTHR43685:SF13, the other five in SF2); UniProt
flat files fetched for all seven.
| Leaf | Name and its evidence | Experimental annotations | Verdict |
|---|---|---|---|
| P26403 (S. typhimurium RfbN) | RecName, curated (Reviewed, ECO:0000305); GN from ECO:0000303 PubMed:1710759; PE 1 | FUNCTION and PATHWAY by ECO:0000269 PubMed:7684736: rhamnosyltransferase adding rhamnose to the galactosyl-undecaprenyl-diphosphate LPS O-antigen intermediate | Claim survives and strengthens: experimentally characterised, product is LPS O-antigen, and it is PslC's closest relative (same subfamily SF13) |
| Q89D69 (B. diazoefficiens "exoU") | SubName "UDP-hexose transferase" {ECO:0000313|EMBL}; PE 4 Predicted | none | "Succinoglycan ExoU" identity does not survive: predicted name only. What is checkable: PF00535 GT-2, 313 aa |
| Q9I4N8 (P. aeruginosa "FgtA", PA1091) | SubName "Flagellar glycosyl transferase, FgtA" {ECO:0000313|EMBL}; PE 4 | none | Flagellin-glycosyltransferase identity does not survive. Stronger replacement fact: 1694 aa with Pfam PF04464 / InterPro IPR007554 TagF-like and no GT-2 match at all, so it does not even share PslC's catalytic domain |
| Q9I3V4 (P. aeruginosa PA1390) | SubName "Probable glycosyl transferase"; PE 4 | none | Uncharacterised; PF00535, 344 aa |
| Q74CL7 (G. sulfurreducens GSU1510.1) | SubName "Glycosyltransferase"; PE 4 | none | Uncharacterised; PF00535, 264 aa |
| Q8P4N7 (X. campestris XCC3670) | RecName derived from ECO:0000259|Pfam:PF00535; PE 4 | none | Uncharacterised; PF00535, 231 aa |
Net effect: the "unrelated glycans" framing was overclaimed, because five of the
six leaves are uncharacterised and their glycan assignments were predicted
names. The argument is rebuilt on (a) RfbN's experimentally curated O-antigen
role, (b) the absence of any experimental annotation on the other six, (c)
domain content including FgtA's non-GT-2 architecture, and (d) the seed being a
deletion phenotype. Confidence held at PROBABLE: the checks strengthened the
one verifiable leaf but the "not a conserved function" step still rests partly
on absence of evidence.