DHTKD1

Summary: Refine the correct mitochondrial localization to mitochondrial matrix.

Reason: The UniProt location mapping correctly places DHTKD1 in mitochondria. Human Reactome explicitly places Q96HY7 in the matrix through the DHTKD1–ThDP–Mg2+ entity and OADH complex, and the mouse ortholog has a curated matrix IDA. These combined data justify the more precise matrix term, already represented by rows 7 and 14–16; the human fluorescence assay by itself does not resolve the matrix.

Proposed replacements: mitochondrial matrix

Summary: Refine a valid E1 oxidoreductase class to 2-oxoadipate dehydrogenase activity.

Reason: GO:0016624 is a direct ancestor of GO:0160166 and applies to the E1 reaction: the substrate-derived intermediate reduces and acylates the oxidized DLST lipoyl group. The InterPro/ARBA mapping is therefore chemically sound, contrary to the earlier complex-only rationale. Human substrate-selective biochemistry supports GO:0160166, already present in rows 3 and 12. DHTKD1 performs the E1 step; DLST and DLD perform the subsequent CoA transfer and reoxidation.

Proposed replacements: 2-oxoadipate dehydrogenase activity

Summary: Thiamine diphosphate binding is an integral feature of the catalytic E1 activity.

Reason: The IPR011603 mapping includes thiamine pyrophosphate binding, and human DHTKD1 structures and activity assays independently establish the ThDP cofactor. This accepted binding function is an integral part of the GO:0160166 E1 catalytic mechanism represented by the single core unit. The engineered structural and activity constructs are not alternative natural products.

Summary: The Rhea mapping identifies the specific DHTKD1 E1 reaction.

Reason: RHEA:69576 and the current GO definition describe decarboxylation of 2-oxoadipate coupled to reductive glutarylation of a protein-bound lipoyl group. Recombinant human biochemistry supports this activity. Glutaryl-CoA and NADH are products of the complete multienzyme sequence, not products independently generated by DHTKD1 alone.

Summary: Retain the positive BioPlex association with PGLYRP3 as non-core.

Reason: The human Q96LB9–Q96HY7 record IM-25778-3723 is a positive anti-tag co-immunoprecipitation in HEK293T, with C-terminally tagged PGLYRP3 as bait and DHTKD1 as prey. AP-MS supports co-complex association, not purified binary binding. PGLYRP3 has peptidoglycan-recognition and bactericidal functions, but these data do not establish a DHTKD1 signaling-receptor or enzyme-binding function. Its secretory biology and DHTKD1 mitochondrial biology also leave the native association unresolved. Preserve the measured association without assigning a catalytic or immune role. The scoped project instruction retains such supported generic interactions as non-core when a justified finer binding class is unavailable.

Summary: Refine the validated BMAL1 interaction to DNA-binding transcription factor binding.

Reason: The positive human Q96HY7–O00327-8 pair is supported by validated Y2H record IM-25472-124467 and companion pooling/array records. GO:0140297 specifies binding to a DNA-binding transcription factor; BMAL1 has established sequence-specific E-box recognition and transcriptional activator function. The tested 625-residue isoform 8 lacks canonical residue 274 but retains the bHLH DNA-binding domain. These data support the partner-class refinement as qualitative binding capacity in tagged yeast assays. The term does not assert transcriptional or circadian regulation by DHTKD1. Native colocalization, endogenous complex formation and affinity remain unestablished, and the study-wide orthogonal validation is not claimed for this individual pair. Unlike the PGLYRP3 associations, this pair has an independently established applicable finer partner class; the distinction is not based on ranking yeast assays above human-cell AP-MS.

Proposed replacements: DNA-binding transcription factor binding

Summary: Retain the second BioPlex PGLYRP3 association as non-core.

Reason: The positive human Q96LB9–Q96HY7 record IM-29278-17115 identifies C-terminally tagged PGLYRP3 bait and DHTKD1 prey in HEK293T anti-tag co-immunoprecipitation. This is co-complex evidence, not direct binary-binding proof. The partner’s bacterial-recognition/killing function does not establish the signal-initiation requirement of signaling receptor binding, nor an enzymatic partner function suitable for a specific refinement. Preserve the association while its native compartment and physiological role remain unresolved. The scoped project instruction retains such supported generic interactions as non-core when a justified finer binding class is unavailable.

Summary: Retain the orthology-supported mitochondrial matrix annotation.

Reason: The source ortholog is mouse Dhtkd1/A2ATU0, which has a curated matrix IDA from PMID:28545977, rather than solely a transfer back from human. The original mouse paper was available only as an abstract, so its precise localization assay is not independently reconstructed. Human Reactome explicitly places the Q96HY7-containing DHTKD1/OADH assemblies in the matrix, and human imaging independently supports mitochondria. The combined evidence supports the biological location without claiming submitochondrial resolution from fluorescence.

Summary: Refine amino acid catabolism to the supported lysine and tryptophan pathways.

Reason: The mouse donor A2ATU0 has experimental amino acid and lysine-catabolism IMP annotations. Independently, human patient fibroblasts accumulate isotope-labeled 2-oxoadipate from labeled lysine, and wild-type DHTKD1 complementation reduces that accumulation. The E1 reaction acts on the common 2-oxoadipate intermediate of lysine and tryptophan breakdown. Thus the two specific pathway terms describe catalytic participation, with tryptophan supported by pathway convergence and reaction assignment rather than a tryptophan tracer experiment. Lysine is already represented by row 9; this refinement supersedes the separate historical NEW tryptophan proposal.

Proposed replacements: L-lysine catabolic process L-tryptophan catabolic process

Summary: Retain the experimentally grounded orthology annotation to lysine catabolism.

Reason: Mouse Dhtkd1/A2ATU0 has a curated lysine-catabolism IMP from PMID:28545977. Human patient-fibroblast complementation and deuterated-lysine tracing independently connect DHTKD1 to disposal of lysine-derived 2-oxoadipate. Human E1 biochemistry identifies the catalytic step, so this is direct pathway participation rather than an inference from disease necessity alone.

Summary: Refine HPA mitochondrial localization using the combined matrix evidence.

Reason: HPA reports supported mitochondrial staining with antibody HPA037949 in A-431, MCF7 and U2OS cells. This confirms organelle localization but does not resolve the matrix. The more precise matrix replacement follows the independently curated human Reactome Q96HY7 entity and mouse ortholog IDA, while retaining the original HPA evidence identity. The matrix assertion is already represented by rows 7 and 14–16.

Proposed replacements: mitochondrial matrix

Summary: Refine the mitochondrial proteome annotation to matrix using independent compartment evidence.

Reason: The original MitoCoP Table S1 identifies Q96HY7 as the majority and simplified protein ID: sheet A row 305 is experimentally identified, belongs to MitoCoP and scores Class 1 in subtractive proteomics, Cluster 1 in spatial proteomics and Class 1 in importomics; sheet B row 278 confirms the gene-level entry. This directly corroborates mitochondria. The matrix replacement uses the separate human Reactome Q96HY7 matrix entity and mouse matrix IDA, not the workbook’s compiled GO-matrix label. Its membrane-association category is Ambiguous, so no membrane-specific location is inferred.

Proposed replacements: mitochondrial matrix

Summary: Retain the direct human biochemical annotation for the specific E1 activity.

Reason: The verified PMID:29191460 abstract reports recombinant human E1a with approximately 49-fold catalytic-efficiency preference for 2-oxoadipate over 2-oxoglutarate. Full text for that paper was not recovered; later accessible human biochemical/structural studies corroborate the reaction and its component-specific chemistry. Lower oxoglutarate activity is not absent activity, and the 2020 docked or tentative substrate models are not established substrate-bound crystal structures.

Summary: Retain DHTKD1 membership in the oxoadipate dehydrogenase complex.

Reason: The original reconstitution report assigns human E1a to a functional complex with DLST and DLD. Later human biochemical assembly and endogenous co-immunoprecipitation/PLA provide independent support. DHTKD1 is the E1 component; reported hybrid association with OGDH does not negate OADH membership or require a second core function. The original source identity is preserved despite independent corroboration.

Summary: Reactome explicitly locates the DHTKD1 E1 reaction and protein in the matrix.

Reason: Event R-HSA-9858321 has catalyst 9853486, active unit DHTKD1 dimer R-HSA-9858579 and physical complex R-HSA-9858580. The dimer contains R-HSA-9858584, which contains the Q96HY7 protein entity R-HSA-6792669 in mitochondrial matrix. This is a target-specific component path, not localization inferred merely from another enzyme’s reaction.

Summary: The DLD reaction carries a valid matrix annotation through the DHTKD1-containing complex.

Reason: Event R-HSA-9858589 uses a DLD active unit, but its catalyst’s physical entity is OADH complex R-HSA-9858580, whose explicit components include the Q96HY7-containing DHTKD1 dimer in matrix. The cellular-component annotation is therefore supported by membership. It does not assign DLD-mediated lipoamide reoxidation or NADH production to DHTKD1.

Summary: The DLST reaction carries a valid matrix annotation through the DHTKD1-containing complex.

Reason: Event R-HSA-9858590 has DLST as the active unit and OADH complex R-HSA-9858580 as the physical entity. The complex explicitly contains the DHTKD1 dimer and Q96HY7 matrix entity. This supports the location while preserving the division of labor: transfer of glutaryl to CoA is the DLST step, not an autonomous DHTKD1 activity.

Summary: Refine the directly observed mitochondrial localization using independent matrix evidence.

Reason: PMID:23141294 reports endogenous DHTKD1 colocalization with MitoTracker in human HepG2 cells. That experiment establishes mitochondria, not matrix. The matrix replacement uses the separately curated human Reactome Q96HY7 entity and experimentally annotated mouse ortholog, and is already represented by rows 7 and 14–16. The original IDA/source fields are retained and the evidence for the finer compartment is explicitly separate.

Proposed replacements: mitochondrial matrix

Summary: Retain the energy-metabolism perturbation phenotype as non-core.

Reason: Three siRNAs reduced DHTKD1 expression; their equimolar mixture decreased ATP and NAD(H) pools in human HEK293T cells (PMID:23141294). Those pool measurements do not identify a direct reaction. Retention also has independent performed-step support: DHTKD1 catalyzes E1 decarboxylation and reductive glutarylation of DLST-bound lipoamide in the OADH pathway. The PMID:32633484 supplement distinguishes isolated E1 DCPIP reduction from NADH production by reconstituted DHTKD1, DLST and DLD. DLST transfers glutaryl to CoA and DLD subsequently reduces NAD+ while reoxidizing the lipoyl group, as represented in the separate Reactome reactions. DHTKD1 therefore performs part of this precursor- and redox-generating sequence; it is not merely required for an unexplained phenotype. The broad process is retained as contextual to the specific amino-acid-catabolic core. No autonomous DHTKD1 NADH production, ATP synthesis or DLD catalytic activity is asserted.

Summary: Retire the separate historical NEW proposal because row 8 already retains the tryptophan-pathway refinement.

Reason: This was a prior reviewer proposal, not an imported GOA row. The row 8 replacement already includes GO:0006569 alongside the lysine term, and the single catalytic core retains the convergent tryptophan-pathway role. The separate NEW object is therefore retired as duplicate coverage, not as a withdrawal of the biological claim. The former lysine-only quotation did not itself establish tryptophan catabolism; current support is the E1 reaction and the pathway convergence described in PMID:32633484. Historical term, evidence and reference identifiers remain unchanged.