traB (F plasmid, Escherichia coli K-12) — curation notes

UniProt: P41067 (TRAB1_ECOLI), 475 aa. Encoded on plasmid F and on plasmid
IncFII ColB2, not on the chromosome. UniProt assigns it to "the TraB family"
and gives one functional statement: "Involved in F pilus assembly."

Identity and role

TraB is a structural subunit of the outer-membrane core complex of the F-type
type IV secretion system, the VirB10 counterpart of that machine. The
cryo-EM structure of the core complex of an F-family system resolves the
architecture directly: the complex is built from an outer ring of TraK and
TraV together with a central cone made entirely of TraB
PMID:35046412.
The same work establishes that these three proteins are functionally
load-bearing rather than incidental
PMID:35046412.

Note on strain scope: that structure was solved for pED208, an F-family
plasmid, so it establishes the role for the F-type system rather than for
P41067's own polypeptide. P41067 is the F plasmid's TraB.

The F transfer region is the reference conjugation system in which this gene
sits PMID:7915817,
and its genes cover pilus synthesis, mating-pair stabilization, surface
exclusion, DNA nicking and transfer, and regulation
PMID:7915817.
TraB belongs to the first of those groups.

The ATP hydrolysis annotation is attributed to the wrong protein

P41067's only GOA row is GO:0016887 ATP hydrolysis activity, IDA, from
PMID:20172994, assigned by CACAO. That paper is not about this protein. Its
abstract states the subject explicitly
PMID:20172994,
and its title names the plasmid as well.

Four independent facts place the paper's protein apart from P41067:

  1. Plasmid. The paper studies pKM101. P41067 is recorded by UniProt as
    encoded on plasmid F and plasmid IncFII ColB2.
  2. Family. P41067 belongs to the TraB family. The pKM101 protein is a
    VirB4 homologue; the corresponding UniProt entry Q46698 is annotated
    "Belongs to the TrbE/VirB4 family".
  3. Length. P41067 is 475 aa. Q46698 is 866 aa, consistent with VirB4-family
    ATPases being the largest component of the machine.
  4. Where F keeps that role. F's own VirB4-family ATPase is a different
    gene product, TraC (P18004). If F TraB were the VirB4 homologue, F would
    have two.

The collision is purely nomenclatural: Tra symbols are not stable between
conjugative plasmids, and on pKM101 the symbol TraB happens to name the
VirB4 homologue that on F is called TraC.

Worth recording even for the correct protein: the paper's ATP hydrolysis
result is conditional on how the protein was prepared. The membrane-purified
form did not hydrolyse ATP, and only the soluble hexameric form did
PMID:20172994.
So the annotation would need care even on Q46698.

Conclusion: REMOVE. This is not a case of second-guessing an assay whose full
text is unavailable. The cached abstract states the plasmid and the family
outright, which is the condition CLAUDE.md sets for making that call, and the
paper's protein has its own accession.

Open questions