Curation notes: Pseudomonas putida KT2440 fadA (Q88L01, PP_2137)

These notes record the reasoning behind the two non-routine curation actions in
fadA__Q88L01-ai-review.yaml: the REMOVE of GO:0010124 and the NEW
GO:0036125. No OpenScientist deep-research report was generated for this gene, so
everything below is anchored to the cached UniProt records and GOA tables already
in the repository.

Identity and family assignment

The reviewed UniProt entry is a HAMAP-rule-typed FadA: the rule MF_01620 names
the protein 3-ketoacyl-CoA thiolase, assigns EC 2.3.1.16 and Rhea RHEA:21564,
places it in the fatty acid beta-oxidation pathway, and calls it the
"Fatty acid oxidation complex subunit beta"
[file:PSEPK/fadA__Q88L01/fadA__Q88L01-uniprot.txt "HAMAP; MF_01620; FadA; 1."].

Critically, the PANTHER assignment is to the FadA-specific subfamily, not to
the family root:

So both UniProt's rule-based typing and its subfamily assignment describe a
beta-oxidation thiolase, and neither mentions aromatic-compound catabolism.

Why GO:0010124 (phenylacetate catabolic process) is removed

The annotation is a TreeGrafter electronic inference — GOA records it as
IEA/GO_REF:0000118 with WITH/FROM = PANTHER:PTN002466592
[file:PSEPK/fadA__Q88L01/fadA__Q88L01-goa.tsv]. There is no experimental,
author, or curator-reviewed evidence behind it, so per the repository's guidance
this is an over-propagated electronic inference that may be argued against on
biological grounds rather than a curator judgment to defer to.

The decisive biological argument is that KT2440 runs phenylacetate catabolism
through a dedicated paa operon that has its own thiolase
, curated separately
in this repository:

Gene Accession Locus Product GO:0010124 source
paaJ Q88HS3 PP_3280 3-oxoadipyl-CoA/3-oxo-5,6-dehydrosuberyl-CoA thiolase (EC 2.3.1.174) TreeGrafter, PANTHER:PTN001291485
paaH Q88HS1 PP_3282 3-hydroxyadipyl-CoA dehydrogenase (EC 1.1.1.35) InterPro2GO, InterPro:IPR011967
paaF Q88HR9 PP_3284 Enoyl-CoA hydratase-isomerase (EC 4.2.1.17) — (carries GO:0006635 only)

The load-bearing point is that the thiolase step of the phenylacetate route is
already carried out by a different protein
, and fadA's own typing excludes it
from that route: paaJ/PP_3280 is a distinct gene product with its own EC number
(2.3.1.174) acting on the ring-cleavage intermediates, while Q88L01 is assigned to
the FadA-specific subfamily PTHR43853:SF11 and typed by HAMAP rule MF_01620,
both of which confine it to fatty acid beta-oxidation. Under the repository's
guidance that is an over-propagated electronic inference that can be argued
against on biological grounds.

A secondary, non-decisive observation is that paaJ receives GO:0010124 from a
different TreeGrafter node (PTN001291485) than the one that put the term on fadA
(PTN002466592). This is consistent with the term reaching fadA by thiolase-fold
propagation, but it does not settle the matter on its own: two nodes could in
principle both legitimately carry GO:0010124 if two thiolase clades each act in
phenylacetate catabolism. It is recorded as corroboration, not as the argument.

Note for coherence with the neighbouring ACCEPT: PTN002466592 is also among
the sources of the GO:0006635 fatty acid beta-oxidation annotation this review
accepts, where it is joined by UniRule:UR000080052 and UniPathway:UPA00659
[file:PSEPK/fadA__Q88L01/fadA__Q88L01-goa.tsv]. The node itself is a
beta-oxidation node; it is the phenylacetate term sitting on it that is the
anomaly, so removing GO:0010124 does not undercut GO:0006635.

Taken together, GO:0010124 on fadA is family-wide electronic propagation across
the thiolase fold and should be removed rather than merely flagged.

The paralog context matters for the same reason and is recorded here for
auditability: thiolase-fold propagation of pathway-specific terms is expected to
hit multiple KT2440 thiolases, so the presence of GO:0010124 on this entry is not
evidence of a fadA-specific phenylacetate role.

Why GO:0036125 (fatty acid beta-oxidation multienzyme complex) is added

UniProt states the quaternary structure explicitly:
"Heterotetramer of two alpha chains (FadB) and two beta chains (FadA)"
[file:PSEPK/fadA__Q88L01/fadA__Q88L01-uniprot.txt]. The corresponding alpha
subunit is curated in this repository as genes/PSEPK/fadB. GOA carries no
cellular-component term for this complex on Q88L01 — only GO:0005737 cytoplasm —
so the part_of complex annotation is a genuine gap rather than a
re-statement of an existing term, and it is asserted at the level UniProt
actually supports (complex membership, not a new molecular function).

Open point

The chain-length range over which the FadBA complex operates, relative to the
other KT2440 thiolase paralogs, is not established by any record consulted here;
it is carried in suggested_questions rather than asserted.