AASDHPPT review notes

2026-08-08 setup and provenance

just fetch-gene human AASDHPPT retrieved the reviewed UniProt Q9NRN7 record and 32 live QuickGO rows, which the repository seeder normalized to 27 review tuples.
Falcon deep research failed with Edison HTTP 402. The configured perplexity-lite fallback also failed with HTTP 401 because the provider quota was exhausted. No provider-branded research file was created or substituted with manual content.
just fetch-gene-pmids human AASDHPPT completed for all 11 source PMIDs. Manual synthesis below therefore uses cached primary literature, database records, and explicitly identified external-source checks.

GOA normalization audit

The 32 live rows reduce to 27 source tuples only because the normalization key collapses repeated WITH/FROM rows. PMID:25416956 has three otherwise identical GO:0005515/IPI/enables rows (TRAF2, SIAH1, and USP22), and PMID:32296183 has four (CRX, TRAF2, EFHC2, and USP22 isoform 2). These two groups account for all five collapsed rows; their seven partner identifiers must remain as unioned supporting_entities, not be silently discarded.
Four live annotation extensions also require explicit preservation because the local TSV export has no extension column: (1) GO:0008897/IDA/PMID:19933275 has input UniProtKB:O75891 (ALDH1L1) and is causally upstream of GO:0016155 (formyltetrahydrofolate dehydrogenase activity); (2) GO:0008897/IDA/PMID:21238436 has input UniProtKB:Q3SY69 (ALDH1L2); (3) GO:0051604/IDA/PMID:19933275 has input UniProtKB:O75891; and (4) GO:0070062/HDA/PMID:19056867 is part of UBERON:0001088 (urine). The unusual first causal extension was rechecked against the live QuickGO API on 2026-08-08 and is present exactly as recorded. The fourth extension describes the sampled biofluid, not a stable AASDHPPT compartment.
No source row is negated or isoform-targeted. The isoform-2 identifier appearing in a WITH/FROM field belongs to the interaction partner USP22, not to AASDHPPT.

Direct evidence for the human gene product

The gene was first connected to the yeast LYS5 pathway by cloning a human cDNA and testing it in a yeast knockout. The cache states that the authors "have identified a full-length human cDNA homologous to the yeast LYS5 gene" and that the "homolog encodes alpha-aminoadipate dehydrogenase phosphopantetheinyl transferase" PMID:11286508. This is functional evidence for the human coding sequence in a heterologous complementation assay; it is not direct evidence for a human stable complex with AASDH.
Purified human AASDHPPT transfers the CoA-derived 4'-phosphopantetheine group to carrier-protein serines. The founding study says the "enzyme is capable of transferring the 4'-phosphopantetheine moiety of coenzyme A" and identifies the "human cytosolic multifunctional fatty acid synthase and the acyl carrier protein" "associated independently with human mitochondria" as acceptors PMID:12815048. It concludes that humans use a broad-specificity enzyme for post-translational phosphopantetheinylation PMID:12815048.
Structural work directly resolved human AASDHPPT with donor and acceptor substrates: "We have now determined the structure of human PPT in complex with CoA and the ACP domain of human FAS (PDB ID: 2CG5)." The FASN ACP interaction is a catalytic enzyme-substrate interface—"The interaction between the two proteins takes place on a large interface and is mostly of hydrophobic character."—rather than evidence for a persistent complex PMID:18022563. Magnesium is coordinated by the CoA pyrophosphate and by Asp129/Glu181 PMID:18022563. Mutagenesis supports Glu181 and Lys185 as acid/base catalytic residues PMID:18022563.
ALDH1L1 is a directly demonstrated human substrate. Purified recombinant human AASDHPPT plus CoA and Mg2+ restored the 10-formyltetrahydrofolate dehydrogenase activity of apo-ALDH1L1/FDH PMID:19933275. Ser354 is the acceptor site PMID:19933275. Human A549-cell knockdown independently connects the endogenous transferase to this modification PMID:19933275. AASDHPPT therefore acts upstream of ALDH1L1-mediated folate catabolism; it does not itself catalyze 10-formyltetrahydrofolate breakdown.
ALDH1L2 is another directly demonstrated human substrate. The study expressed and purified human mtFDH PMID:21238436, found that "PPT modifies mtFDH exclusively at Ser375", and showed that the AASDHPPT/CoA/Mg2+ reaction made recombinant ALDH1L2 dehydrogenase-competent PMID:21238436. This supports the has_input ALDH1L2 extension and the causal, non-core folate-process annotation.
Human HepG2 chemical proteomics broadened the substrate inventory by identifying known phosphopantetheinylated proteins and their modified sites [PMID:32537878 "proteins in HepG2 cells. In combination with a quantitative chemical proteomic"; PMID:32537878 "4'-phosphopantetheinylated proteins with high confidence, and unambiguously"]. It also reported "a potential 4'-phosphopantetheinylation site" in "the protein of mitochondrial dehydrogenase/reductase SDR family member 2" PMID:32537878. Because the cached record is abstract-only and describes DHRS2 as potential, DHRS2 should remain a candidate substrate rather than a confidently assigned direct AASDHPPT target.

Localization: human evidence versus mouse orthology evidence

Direct human fractionation in the founding study supports a predominant cytosolic pool PMID:12815048. The current reviewed record likewise says Cytoplasm, cytosol [UniProt:Q9NRN7 "Cytoplasm, cytosol"]. Cytosol is therefore an experimentally grounded human location, but neither source establishes exclusivity.
PMID:41061852 substantially revises the older model by demonstrating a mitochondrial-matrix pool and an N-terminal targeting determinant. Its principal mechanistic system was explicitly orthologous mouse material: "To test whether PPT is required for mitochondrial oxidative metabolism, we used CRISPR/Cas-9 to mutate Aasdhppt in C2C12 mouse skeletal myoblasts." In that system endogenous PPT was detected in both mitochondrial and post-mitochondrial fractions PMID:41061852, matrix localization was supported by protease protection and split-Venus complementation PMID:41061852, and the first 36 residues were sufficient to target reporters to the matrix PMID:41061852.
The same C2C12 experiments connect the matrix pool to NDUFAB1 phosphopantetheinylation and mtFAS: "From these data, we conclude that PPT is required for 4’PP modification of NDUFAB1 and downstream mtFAS activity [15]." Loss reduced holo- and octanoyl-NDUFAB1, lipoylation, respiratory-complex abundance, and respiration, with full-length rescue. This is strong mammalian orthology evidence and is biologically consistent with the earlier human in-vitro NDUFAB1/mitochondrial-ACP result, but the localization and downstream-process experiments should not be represented as human IDA without replication in human cells or tissue.
PMID:41061852 also carries bounded human-allele evidence: substitutions corresponding to human AASDHPPT ClinVar VUS were modeled in the C2C12 rescue system, and "all variants, except for PPT-GFP P12S, exhibited weaker rescue than wild type PPT" PMID:41061852. The affected substitutions were R47H, A72T, W79G, R83C, and K96E. This strengthens the paper's human-disease relevance, but it remains a mouse-cell functional model and does not directly establish endogenous human matrix localization or clinical pathogenicity.
The resulting synthesis is dual localization: a major cytosolic human pool and a smaller mitochondrial-matrix pool strongly demonstrated in mouse C2C12 cells. The earlier hypothesis that all mitochondrial substrates are modified before import is no longer sufficient, but direct evidence for the abundance, regulation, and physiological importance of the matrix pool in humans remains incomplete.

Substrate and complex model

High-confidence human substrates are the ACP domain of FASN, mitochondrial ACP/NDUFAB1, ALDH1L1 at Ser354, and ALDH1L2 at Ser375. AASDH is supported by human-cDNA complementation and the conserved lysine-pathway model but lacks the same site-resolved human biochemical evidence in the cached set. DHRS2 is a chemical-proteomic candidate. Prokaryotic ACP/PCP modification demonstrates biochemical breadth but is not a normal human biological role.
Reactome models the canonical cytosolic reaction precisely: "Cytosolic AASDHPPT (alpha-aminoadipic semialdehyde dehydrogenase-phosphopantetheinyl transferase) catalyzes the transfer of a phosphopantetheine moiety from coenzyme A to serine 2156 within the ACP domain of FAS (fatty acyl synthase)." [Reactome:R-HSA-199202]. This is an appropriate core human reaction.
The protein is monomeric [PMID:12815048 "The human enzyme, which is expressed mainly in the cytosolic compartment in a wide range of tissues, is a 329-residue, monomeric protein."; UniProt:Q9NRN7 "Monomer."]. The crystallographic AASDHPPT-FASN ACP complex is a transient substrate complex. The GOA interactions with TRAF2, SIAH1, USP22/USP22-2, CRX, and EFHC2 come from large interaction screens and do not establish a stable AASDHPPT complex, regulator, or additional catalytic substrate. No stable complex should be placed in core_functions.

Isoform and sequence caveats

The reviewed Q9NRN7 record is 309 amino acids and lists two splice isoforms. Isoform 2 changes residue 138 [UniProt:Q9NRN7 "R -> T (in isoform 2)"] and lacks residues 139-309 [UniProt:Q9NRN7 "Missing (in isoform 2)"]. It therefore omits Glu181 and Lys185, which the structural study identifies as essential catalytic residues. No cached experiment establishes isoform-2 protein abundance, localization, or transferase activity, so all core catalytic claims should be scoped to the canonical displayed isoform unless specifically tested.
PMID:11286508 predicted a 309-aa product, whereas the PMID:12815048 abstract calls the characterized enzyme a 329-residue protein. The structure spans the current 309-aa numbering. Treat the 329-aa statement as a historical sequence-description discrepancy, not as evidence for a third isoform or an alternative functional protein.

Curation rationale

Accept GO:0008897 holo-[acyl-carrier-protein] synthase activity and GO:0000287 magnesium ion binding as core molecular functions: both are directly supported by human biochemistry, structures, and mutagenesis. Accept cytosol as a directly demonstrated human location; consider mitochondrial matrix only with the explicit mouse-orthology evidence boundary.
Modify generic GO:0051604 protein maturation to GO:0018070 peptidyl-serine phosphopantetheinylation. The papers define the covalent serine modification and its acceptor sites precisely.
Keep acts_upstream_of_positive_effect GO:0009258 annotations as valid but non-core. AASDHPPT activates ALDH1L1/ALDH1L2; the substrate enzymes perform the folate-catabolic chemistry.
Modify the Reactome-derived GO:0015939 pantothenate metabolic process annotation to the specific phosphopantetheinylation process. Reactome correctly notes that pantothenate is upstream of CoA [Reactome:R-HSA-199220 "Vitamin B5 ((R)-pantothenate, PanK), is an essential precursor for the synthesis of the metabolic cofactor Coenzyme A (CoA-SH) (Robishaw and Neely 1985) and is the prosthetic group of acyl carrier protein (ACP) (Joshi et al. 2003)."], but AASDHPPT transfers a moiety from CoA and does not metabolize pantothenate itself.
Mark generic GO:0005515 protein-binding annotations as over-annotated. The FASN contact is already part of substrate recognition during catalysis, while the other pairs are screening observations without a demonstrated mechanism. Keep the exact partner provenance. Keep urinary-exosome localization as non-core HDA: the paper profiled a biofluid vesicle preparation PMID:19056867, not selective export or an exosomal enzymatic role.

Knowledge gaps and useful experiments

Confirm endogenous human AASDHPPT in purified mitochondrial matrix fractions and by microscopy/proximity labeling, then test whether disrupting the N-terminal targeting region selectively reduces human NDUFAB1 4'-phosphopantetheinylation, mtFAS, lipoylation, and respiration.
Quantify cytosolic-versus-matrix partitioning across human tissues and metabolic states, and determine whether mitochondrial import involves cleavage, regulated routing, or competition with cytosolic substrate capture.
Test isoform 2 directly for protein stability, localization, CoA/Mg2+ binding, and activity on FASN ACP and NDUFAB1; transcript detection alone cannot establish a functional isoform.
Map endogenous site occupancy and AASDHPPT dependence for AASDH and DHRS2 in human cells. This would distinguish direct physiological substrates from complementation or chemical-proteomic candidates.
Determine whether any recurrent high-throughput interaction partner regulates AASDHPPT or is a substrate. Until targeted reciprocal/co-localization and functional assays exist, none supports a stable complex or core binding function.