AIGR Deep Research — AprA (Q72DT2) plasma membrane localization (GO:0005886) OpenScientist openscientist-autonomous 2 artifacts 2026-09-20T18:10:24.021369

OpenScientist final report (text/html)

OpenScientist final report (application/pdf)

AIGR Deep Research — AprA (Q72DT2) plasma membrane localization (GO:0005886)

Gene: aprA — adenylyl-sulfate reductase, alpha subunit
Organism: Nitratidesulfovibrio (Desulfovibrio) vulgaris Hildenborough (NCBITaxon:882)
UniProt: Q72DT2 (664 aa; flavoprotein; SdhA/FrdA/AprA superfamily)
Focus: function_assignment — does AprA directly localize to the plasma membrane (GO:0005886)?

Executive Judgment

Verdict: Partially supported / weakly supported as a bare CC term — refine, do not accept as-is.

AprA is fundamentally a soluble cytoplasmic flavoenzyme: the exact-strain enzyme was purified to homogeneity as a soluble αβ (α2β2) protein (PMID 8174563), and Q72DT2 has no transmembrane helix. The plasma-membrane annotation currently carried by UniProt (GO:0005886) rests only on IEA:TreeGrafter, a phylogenetic graft from the SdhA/FrdA membrane-anchored flavoprotein family, and is therefore weak and imprecise.

However, the hypothesis is not simply wrong. There is genuine experimental evidence for a peripheral, partially stable membrane-associated pool of AprAB, recruited to the cytoplasmic face of the integral membrane QmoABC complex (co-purification + blue-native PAGE in the exact strain, PMID 23842468; direct electron transfer, PMID 26768116). Crucially, this is a peripheral/extrinsic association driven by protein–protein interaction, not intrinsic membrane residency. Membrane interaction via Qmo alone does not establish residency.

Bottom line for the curator: The primary location is cytoplasm (GO:0005737). A membrane pool is real but peripheral; if captured, it is best modeled as extrinsic component of the cytoplasmic side of the plasma membrane (GO:0031234) supported by IPI evidence, not as bare "plasma membrane." Separately, the co-annotated GO:0000104 succinate dehydrogenase activity is a paralog overannotation and should be removed.

Evidence Matrix

Citation	Evidence type	Supports/Refutes/Qualifies	Claim tested	Key finding	Context	Confidence & limitations
PMID 8174563 (Verhagen 1994)	Direct biochemistry / purification	Refutes intrinsic membrane residency; supports soluble	Is AprAB a soluble enzyme?	Purified to homogeneity as soluble αβ (67.8/25.6 kDa), α2β2, one Fe-S cluster + FAD; no membrane anchor	D. vulgaris Hildenborough (exact strain)	High for solubility; predates Qmo-complex concept, did not test membrane fraction
PMID 23842468 (Krumholz 2013)	Interaction / co-purification / BN-PAGE	Qualifies (supports peripheral membrane pool)	Does AprAB associate with the membrane?	Membrane band contains 3 Qmo subunits + AprA + AprB; AprA/QmoA/QmoB tagging reciprocally co-purifies AprA/B; "Qmo-Apr complex at least partially stable in protein extracts"	D. vulgaris Hildenborough + D. alaskensis G20	High for interaction; shows association via Qmo, i.e., peripheral, not integral; does not quantify soluble vs membrane partition
PMID 26768116 (Duarte 2016)	Direct assay (electrochemistry)	Qualifies (interaction, not residency)	Is the AprAB–Qmo link functional?	Direct QmoABC→AprAB electron transfer; Qmo essential for efficient electron delivery to sustain catalysis	D. desulfuricans ATCC 27774	High for electron transfer; orthologous strain; demonstrates functional interaction, not localization
PMID 20581180 (Zane 2010)	Mutant phenotype	Supports Qmo being the membrane partner	Is Qmo the transmembrane electron conduit to Apr?	ΔqmoABC cannot grow on sulfate; Qmo is a transmembrane complex delivering electrons to Apr	D. vulgaris Hildenborough	High; establishes Qmo as membrane-integral partner of soluble Apr
UniProt Q72DT2 (database)	Computational / database	Competing (weak basis for GO:0005886; flags overannotation)	What is the current annotation basis?	GO:0005886 = IEA:TreeGrafter; GO:0000104 SDH activity = IEA:TreeGrafter; family = SdhA/FrdA/AprA (PF00890); 0 TM	Sequence/family	Orientation only; IEA graft from membrane-anchored SDH/FRD relatives

GO Curation Implications (leads — require curator verification)

GO term	Aspect	Current basis	Recommended action	Rationale
GO:0005886 plasma membrane	CC	IEA:TreeGrafter	Do not retain as bare term / generalize-then-specify. Replace with GO:0031234 extrinsic component of cytoplasmic side of plasma membrane (or GO:0019898 extrinsic component of membrane) with IPI evidence from PMID 23842468 / 26768116.	AprAB is peripherally recruited via Qmo; no TM helix; bare "plasma membrane" overstates residency.
GO:0005737 cytoplasm	CC	(absent)	Add as primary location (IDA/IEA)	Soluble αβ enzyme purified from non-membrane fraction (PMID 8174563); no membrane anchor.
GO:0000104 succinate dehydrogenase activity	MF	IEA:TreeGrafter	Remove	Paralog overannotation from SdhA/FrdA/AprA family; AprA catalyzes APS reduction, not succinate oxidation.
GO:0009973 adenylyl-sulfate reductase activity	MF	IEA:UniProtKB-EC	Retain (core function)	Correct primary molecular function.
GO:0009055 electron transfer activity	MF	IEA:TreeGrafter	Retain (defensible)	Consistent with FAD/Fe-S redox chemistry and Qmo electron transfer.

Avoids "protein binding" as the recommendation; the interaction is captured more informatively via the extrinsic-membrane CC term plus retention of the APS reductase MF term.

Computed Provenance (Iteration 2 sequence analysis)

Kyte–Doolittle hydropathy analysis of the Q72DT2 sequence (664 aa) fetched live from UniProt:

Metric	Value	Interpretation
Max KD hydropathy, 19-res window	1.232 (res 25–43)	Below the ~1.6 transmembrane threshold → no TM-length hydrophobic segment
N-terminal max window (res 1–30)	1.232	No signal-peptide-like hydrophobic n-region
GRAVY (grand avg hydropathy)	−0.371	Hydrophilic → soluble character
UniProt annotated TM features	0	Consistent with the above

Provenance CSV saved to /tmp/aprA_artifacts/seq_membrane_analysis.csv. This independent sequence check corroborates that AprA has no intrinsic membrane anchor or secretion signal; any plasma-membrane association is necessarily peripheral/extrinsic (Qmo-mediated), not integral.

QuickGO annotation provenance (live query, EBI): All 7 GO annotations for Q72DT2 are electronic (IEA); none are experimental. Critically, GO:0005886 plasma membrane carries the strong located_in (residency) qualifier but is assigned solely by TreeGrafter/PANTHER (GO_REF:0000118) — the same source as the erroneous GO:0000104 succinate dehydrogenase, GO:0050660 FAD binding, GO:0009055 electron transfer, and GO:0009061 anaerobic respiration. Only GO:0009973 (adenylyl-sulfate reductase, UniProt EC, GO_REF:0000003) and GO:0016491 (oxidoreductase, InterPro) come from non-PANTHER sources.

GO id	Aspect	Qualifier	Evidence	AssignedBy / Ref
GO:0005886 plasma membrane	CC	located_in	IEA	TreeGrafter / GO_REF:0000118
GO:0000104 succinate dehydrogenase activity	MF	enables	IEA	TreeGrafter / GO_REF:0000118
GO:0009055 electron transfer activity	MF	enables	IEA	TreeGrafter / GO_REF:0000118
GO:0050660 FAD binding	MF	enables	IEA	TreeGrafter / GO_REF:0000118
GO:0009061 anaerobic respiration	BP	involved_in	IEA	TreeGrafter / GO_REF:0000118
GO:0009973 adenylyl-sulfate reductase activity	MF	enables	IEA	UniProt / GO_REF:0000003
GO:0016491 oxidoreductase activity	MF	enables	IEA	InterPro / GO_REF:0000002

The key curation point: despite direct experimental interaction/complex data existing (PMID 23842468, 26768116), the CC annotation is a pure phylogenetic graft asserting residency (located_in). It is both under-evidenced and over-specified — the experimental record supports a peripheral, Qmo-recruited pool (an IPI-supportable extrinsic component CC), not intrinsic plasma-membrane residency.

Mechanistic Scope

Immediate molecular function (direct): two-electron reduction of adenosine-5′-phosphosulfate (APS) → sulfite + AMP, using FAD and a [4Fe-4S] cluster (AprA = flavoprotein subunit; AprB = Fe-S subunit).
Immediate cellular context (direct): soluble αβ enzyme in the cytoplasm that transiently/peripherally docks onto the cytoplasmic face of the membrane-integral QmoABC complex to receive electrons from the menaquinone pool.
Downstream/inferred (not localization evidence): dissimilatory sulfate reduction growth phenotype; ΔqmoABC lethality on sulfate. These reflect the pathway, not AprA membrane residency.

Conflicts and Alternatives

Peripheral pool vs. residency: Membrane co-fractionation (Krumholz) genuinely conflicts with a purely cytoplasmic model, but is fully explained by peripheral recruitment to Qmo — not intrinsic membrane insertion. Both can be true (dual pool).
Paralog confusion (SdhA/FrdA): The FAD_binding_2 family drives TreeGrafter to graft membrane-anchoring and succinate-dehydrogenase terms from SDH/FRD, where the flavoprotein subunit is tethered to membrane cytochromes. This is the likely origin of GO:0005886 and the certain origin of the erroneous GO:0000104.
Partition unknown: No study quantifies the soluble:membrane ratio of AprAB in vivo, so "resident membrane protein" cannot be asserted.

Knowledge Gaps

Gap	What was checked	Why it matters	Resolving experiment
Quantitative soluble vs. membrane partition of AprAB	Literature (PMID 8174563 soluble; 23842468 membrane pool) — no partition data	Determines whether "membrane" is a minor or major pool; affects whether a CC membrane term is warranted at all	Quantitative cell fractionation / Western of AprA in soluble vs. membrane fractions ± Qmo
Dependence of membrane pool on Qmo	Co-purification only	Confirms peripheral (Qmo-dependent) vs. independent membrane binding	AprA localization in ΔqmoABC background
Structural interface / topology of Apr–Qmo	Electrochemistry (26768116), no cryo-EM cited here	Confirms extrinsic cytoplasmic-face geometry	Cryo-EM of the Qmo–Apr supercomplex

Discriminating Tests

Fractionation ± high salt / carbonate wash: peripheral proteins strip off at high pH/salt — would confirm extrinsic (not integral) membrane association.
AprA localization in ΔqmoABC: loss of the membrane pool would prove Qmo-dependent peripheral recruitment.
Cryo-EM / crosslinking of the Apr–Qmo supercomplex: defines the cytoplasmic-face docking geometry.
Family-aware GO re-annotation: verify that GO:0005886/GO:0000104 trace to SDH/FRD ancestral grafts and are not experimentally supported for AprA.

Curation Leads (require curator verification)

Candidate references to attach: PMID 8174563 (soluble αβ enzyme, exact strain), PMID 23842468 (Qmo-Apr membrane co-complex, exact strain), PMID 26768116 (direct Qmo→Apr electron transfer), PMID 20581180 (Qmo membrane partner, phenotype).
Verify snippet — 23842468: "This provides clear evidence for the presence of a Qmo-Apr complex that is at least partially stable in protein extracts of D. vulgaris and D. alaskensis."
Verify snippet — 8174563: "Adenosine phosphosulfate reductase from Desulfovibrio vulgaris Hildenborough has been purified to homogeneity and was found to consist of two subunits."
Verify snippet — 26768116: "the QmoABC complex is essential for efficient electron delivery to AprAB, in order to sustain catalysis."
Candidate GO changes: add GO:0005737 (cytoplasm, primary); replace bare GO:0005886 with GO:0031234 (extrinsic component of cytoplasmic side of plasma membrane, IPI); remove GO:0000104 (succinate dehydrogenase activity, paralog overannotation); retain GO:0009973.
Suggested curator question: Is there any quantitative fractionation establishing AprA as a majority membrane pool? If not, treat plasma-membrane residency as non-core / peripheral only.
Suggested experiments: carbonate-wash fractionation; AprA localization in ΔqmoABC.

Limitations

Literature search was constrained to PubMed abstracts; UniProt was queried live (GO evidence codes, family, TM count). No local bioinformatics files were used. Snippets are quoted verbatim from retrieved abstracts and should be re-verified against source.