AADACL3 catalytic-residue and membrane-topology audit

Generated by analyze.py (uv run python analyze.py). Query: ADCL3_HUMAN / Q5VUY0, 407 aa, UniProt protein existence 2: Evidence at transcript level.

AADACL3's entire GO record is family- or fold-derived electronic inference. This audit asks whether the two premises behind it hold: that the catalytic machinery of the 'GDXG' lipolytic enzyme family is actually conserved in AADACL3, and that AADACL3 is membrane-integral rather than secreted.

1. Catalytic residues are present in the AADACL3 sequence

Oxyanion-hole motif 119-121 reads HGG; the nucleophile elbow 191-196 reads GDSFGG.

2. Triad conservation against characterised relatives

Global pairwise alignment (BLOSUM62, gap -11/-1). Two different counts are reported because only the stricter one supports a conservation claim. "on annotated site" counts how many of AADACL3's three active-site residues align to a position UniProt annotates as an active site in the partner; "conserved" additionally requires the residue to be identical. HIDH shows why the distinction matters: AADACL3's Ser193 lands on HIDH's annotated site 164, but that residue is a threonine, so it counts for the first column and not the second.

3. PROSITE PS01174 (GDXG-family nucleophile serine site)

Pattern (fetched from PROSITE): [LIVM]-x-[LIVMF]-[SA]-G-D-S-[CAS]-G-[GA]-x-[LI]-[CAVT]

4. Membrane topology: predicted, and what UniProt records

5. Pfam PF07859 (Abhydrolase_3) match segmentation

6. Subfamily-level versus fold-level signature assignment

Which signature carries an activity inference matters more than whether one matches at all: a fold signature constrains architecture, a subfamily signature constrains the reaction. Spans are residue ranges on each protein; score is the member-database e-value where the API reports one (InterPro entries themselves carry no score). An e-value of 0.0 is the API reporting underflow below double precision, not a literal zero.

• IPR017157 - Arylacetamide deacetylase (InterPro, subfamily)
• PIRSF037251 - Arylacetamide deacetylase (PIRSF member signature)
• IPR013094 - Alpha/beta hydrolase fold-3 (InterPro, fold)
• IPR050300 - GDXG lipolytic enzyme (InterPro, family)

7. Is the nucleophile a serine in every source cited for the hydrolase IBA?

GO:0017171 serine hydrolase activity is defined by mechanism - a serine nucleophile activated by a proton relay through an acidic and a basic residue - so asserting it at a phylogenetic node propagates a serine nucleophile to every descendant, and a single non-serine member makes a plain transfer unsafe. (It does not make the term unreachable: PAINT can annotate the ancestral node and mark the divergent descendant with a NOT.) This reads the residue at each source's first annotated active site. WITH/FROM tokens are taken from the gene's own AADACL3-goa.tsv so they cannot drift from the record under review; model-organism identifiers are resolved through the Alliance API and Arabidopsis loci through UniProt, and anything that cannot be resolved to a single reviewed entry is reported as unresolved rather than dropped.

Audited: GO:0016787 (IBA, GO_REF:0000033). 17 WITH/FROM tokens, 14 resolved to a protein, 14 with a readable nucleophile, of which 13 are serine. Non-serine: HIDH_SOYBN T164.

Interpretation

All numbers above are produced by the script; the reading of them is the reviewer's. Limitations: topology calls are Phobius/TMHMM/SignalP predictions retrieved from InterProScan, not experiments, and they cannot say which membrane. No structure of any AADACL3 protein exists, so triad geometry is inferred from alignment, not from coordinates. Nothing here demonstrates catalytic activity; it only establishes whether the residues that would be required for activity are present.