Notes: Shigella flexneri sctN / Spa47 / SpaL (UniProt P0A1C1)

Curation journal for the AI gene review. All assertions carry provenance.

Identity

Core biology

Spa47 is a bona fide, oligomerization-activated ATPase. The Dickenson lab
produced the first purified active enzyme:
PMID:26947936
and showed activity tracks oligomeric state
PMID:26947936.
A Walker-A lysine mutant (K165A) is catalytically dead
PMID:26947936
and fails to complement a spa47 null for invasion.

Kinetics (UniProt BIOPHYSICOCHEMICAL PROPERTIES, from PMID:29595954):
KM(ATP) = 181 µM monomer / 114 µM trimer; kcat 0.22 s⁻¹ monomer, 0.84 s⁻¹ trimer.
These are hydrolysis rates — there is no reported synthesis activity for this
enzyme anywhere in the literature.

Structure. Crystal structures of Spa47(80-430) WT and active-site mutants
(PDB 5SWJ/5SWL/5SYP/5SYR) support an activated hexamer model with shared
inter-protomer active sites
PMID:27770024,
PMID:27770024.
Nucleotide-bound structures (PDB 5YBH/5YBI/5ZT1) define the ATP site
PMID:30013545.
Interfacial residues supporting oligomerization were mapped in
PMID:31162724,
PMID:31162724.

Homo-oligomerization is the activation mechanism, so the two IntAct
identical protein binding rows (GO:0042802, IPI, PMID:30460850 and
PMID:31162724) are informative rather than boilerplate self-interaction. The
inhibitor paper also purified the oligomer
PMID:30460850
and showed the inhibitors act without dissolving it
PMID:30460850.

Location and complex. Native Spa47 is cytoplasmic and in large complexes
PMID:18657109,
PMID:18657109.
Cryo-ET of intact S. flexneri injectisomes places the hexamer at the hub of
the cytoplasmic sorting platform
PMID:25583506,
aligned below the export gate
PMID:25583506.
Interactions with MxiK/SctK and MxiN/SctL were established genetically and
biochemically
PMID:12864857;
MxiN/SctL regulates the enzyme differentially by oligomeric state
PMID:29595954.
Injectisome targeting is a property of the N-terminal domain, not of catalysis
PMID:31978132.

Energetics: why the proton terms are wrong for Spa47

Spa47 is a soluble cytoplasmic protein with no transmembrane segment. The
proton motive force is used by the Shigella T3SS, but it is used by the
inner-membrane export gate MxiA/SctV, not by the ATPase:

So: depending on the PMF is not the same as transporting protons, and
hydrolysing ATP is not the same as synthesising it. GO:1902600 and
GO:0046933/GO:0015986 all belong to other proteins (SctV) or to other
families (F1-beta).

The ATP-synthase propagation

GOA row: GO:0046933 proton-transporting ATP synthase activity, rotational mechanism, IEA, GO_REF:0000118 (TreeGrafter), WITH/FROM PANTHER:PTN001807734.

From projects/TREEGRAFTER/rotary_atpase/node_placement.tsv:

So the failure is not "the target sits outside the clade the seeds represent" —
it is that the IBD was placed one node too deep, at the duplication that
separates F1-beta from the T3SS/flagellar export ATPases. All the experimental
seeds for GO:0046933 (E. coli AtpD, human ATP5F1B, yeast ATP2, S. pombe atp2)
lie in the PTN008558588 child. Moving the IBD from PTN008558586 to
PTN008558588 would keep every seeded annotation and stop the function leaking
into the export-ATPase sister clade. (This is a slightly different — and more
precise — framing than the CAUVC/fliI sibling review, which described the target
as outside the clade; the tables show the clade relationship explicitly, and the
fix is a node move rather than an IRD.)

The InterPro2GO rows GO:0046034 and GO:1902600 come from IPR004100
(F1/V1/A1 alpha/beta N-terminal domain), a domain genuinely shared with F1
ATPases — the domain is present, but the mapped terms describe rotary
ATP-synthase biology that this protein does not do. GO:0015986 is a
GOC logical inference (GO_REF:0000108) over the wrong GO:0046933 row and
falls with it.

Note the contrast within the same GOA file: the InterPro2GO rows sourced from
IPR005714 (Type 3 secretion system ATPase SctN) are all correct
(GO:0016887, GO:0030254, GO:0030257, GO:0005737). It is specifically the
IPR004100 (shared F1 domain) and PANTHER-family rows that misfire.

Decisions taken

Term Evidence / source Action
GO:0005524 ATP binding IEA IPR000194, IPR020003 ACCEPT
GO:0005737 cytoplasm EXP PMID:18657109 ACCEPT
GO:0005737 cytoplasm EXP PMID:25583506 ACCEPT
GO:0005737 cytoplasm IEA GO_REF:0000120 ACCEPT
GO:0008564 protein-exporting ATPase activity IEA EC:7.4.2.8 ACCEPT (core MF)
GO:0015986 PMF-driven ATP synthesis IEA GO_REF:0000108 from GO:0046933 REMOVE
GO:0016887 ATP hydrolysis activity IEA IPR005714 ACCEPT
GO:0030254 protein secretion by T3SS IEA IPR005714 ACCEPT
GO:0030257 T3SS complex IEA IPR005714 ACCEPT
GO:0042802 identical protein binding IPI PMID:30460850 ACCEPT
GO:0042802 identical protein binding IPI PMID:31162724 ACCEPT
GO:0046034 ATP metabolic process IEA IPR004100 MARK_AS_OVER_ANNOTATED
GO:0046933 ATP synthase activity, rotational IEA GO_REF:0000118 PTN001807734 REMOVE
GO:1902600 proton transmembrane transport IEA IPR004100 REMOVE

No NEW terms proposed. Unlike the CAUVC/fliI sibling (which was missing
GO:0008564 and the flagellum-assembly child term), this GOA already carries
the full correct set: the export-ATPase MF via the correct EC, ATP hydrolysis,
the T3SS process and complex terms, and the cytoplasm location. Invasion /
virulence phenotypes of spa47 mutants are downstream consequences of failed
effector secretion (the effectors do the invading), so no host-interaction BP
term is proposed — that would be the indirect-effect over-annotation CLAUDE.md
warns against. There is no GO term for the T3SS cytoplasmic sorting platform;
GO:0030257 is the most precise CC available (checked against QuickGO — the
only T3SS CC terms are GO:0030257 and the other secretion-system complexes).