D2HGDH oxidizes D-2-hydroxyglutarate to 2-oxoglutarate. The 2004 work combines purified rat enzyme with human full-open-reading-frame expression in HEK293 cells; the 2021 study uses human residues 51–521, with an N-terminal MBP tag removed and a C-terminal His6 tag retained. Both support the reaction, without identifying a unique physiological electron acceptor. FAD binds in the catalytic pocket, and zinc-containing substrate complexes plus mutagenesis explain the metal-coordinating active site. [PMID:15070399; PMID:15609246; PMID:33431826]
D-malate is not a weak incidental substrate in the human assay: its turnover is comparable to, and at the stated concentration slightly above, D-2-hydroxyglutarate. D-lactate turnover is weaker but positive. Meso-tartrate oxidation is reported for purified rat enzyme, including results not shown. These performed chemical reactions are retained as non-core processes; their physiological fluxes are not established. Neither L-malate/TCA-cycle chemistry nor an NAD-dependent decarboxylating activity follows from these experiments. [PMID:15070399; PMID:33431826]
Zinc, cobalt and manganese stimulate enzyme activity under the measured conditions. Magnesium and calcium did not stimulate the purified rat enzyme. These are biochemical observations; the GO response terms describe cellular or organismal responses. Three positive process transfers exceed the assays. The two original NOT assertions remain literal and unresolved at that broader process scope; the negative enzyme observations are preserved rather than reversed. PMID:15070399
The three official products are Q8N465-1 (521 residues), Q8N465-2 (307) and Q8N465-3 (319). Each sequence was reconstructed from the actual annotated splice edits and matched to its official FASTA. The short products omit much of the canonical C-terminal substrate/metal-binding region. This supports caution about equal activity, not a claim that their inactivity was measured. The predicted 1–13 transit peptide, Reactome's modeled 14–521 chain and the experimental 51–521 crystallography construct are separate facts. No original GOA isoform field was added or changed.
MitoCoP Table S1 explicitly includes D2HGDH: protein group 8117, sheet A row 290 and sheet B row 263. The target has class 1 subtractive proteomics, cluster 1 spatial profiling, class 1 importomics and 73.55% subtractive-proteomics mean sequence coverage. Its membrane-association classification is peripheral. These are actual target measurements; the matrix annotation column is separately imported prior knowledge. HPA reports supported mitochondrial ICC/IF in Hep-G2 and RT-4 cells using HPA060548. Neither source establishes individual splice-product localization. Reactome's actual Q8N465 catalyst supplies the accepted matrix model and 2-oxoglutarate pathway context. [PMID:34800366; GO_REF:0000052; Reactome:R-HSA-880007; Reactome:R-HSA-880009]
The mitochondrial IBA is not rejected for its donor count or human self evidence. The actual PTHR43716 tree and MSA place the 521-residue target beneath PTN000202300; the literal 2017-02-28 IBD has the same rat, yeast and human descendants as the source row. Rat P84850 and yeast P46681 were resolved through actual official cross-references and tree leaves after direct MOD-ID API queries failed. Complete current pages contain the relevant experimental mitochondrial assertions. The genuine October 12 IBD capture was reused with its original timestamp; it was not presented as a fresh October 16 retrieval.
The DPP8/9-inhibitor study provides much more than a proteomic association. Full-length D2HGDH is depleted; proteasome/cullin-pathway inhibition blocks the effect; removal of residues 1–51 increases stability and prevents depletion. Transferring those residues to CARD8-ZUC permits inhibitor-induced GSDMD cleavage and LDH release in engineered CARD8-null THP-1 cells. Nevertheless, native D2HGDH is the degradation target, and the fusion remains an engineered substrate. These results do not establish a native D2HGDH process of destabilizing other proteins. The existing process is marked over-annotated while all construct evidence is retained. [PMID:33053349, Results and Figure 4 caption; Methods cloning and limited proteolysis]
The two clinical activity sources with only complete abstracts available (PMID:16037974 and PMID:20020533) explicitly report impaired mutant enzyme activity. The catalytic function is independently secure, so their experimental annotations are retained with curator deference and without invented protocol details. PMID:15609246 differs: its cache is abstract-only, but the original PMC1196381 HTML Table 1 and radiochemical notes were independently recovered and inspected. Sequence-resource and clinical-context caches were also assessed for relevance; they do not add enzyme functions simply because they were downloaded.
The evidence guide provides public original/DOI locators, exact MitoCoP cells and explicit assay/product limits. It is authored navigation, not independent supporting evidence. Original publication, UniProt and GOA cache bytes were preserved. Three existing publication caches and the pending shared PMID:15815621 cache were guarded; the genuine new 15815621 fetch was byte-identical to the pending source. The latter is sequence-resource context and is not used to establish an additional function.
The normal Falcon request failed with a genuine authentication error. Its configured perplexity-lite fallback also failed because of provider quota. Both responses and terminal receipts are preserved; the synthesis here is manual research and an independently attributed chemistry consultation, not a fabricated provider report. No NEW annotation is proposed. The primary unknowns are the physiological electron acceptor, short-product activity and intact-cell D-malate flux.
Protein destabilization can involve a protein acting in cis; it does not require an effect on another protein. The earlier “destabilizing other proteins” wording is superseded by the distinction between performing a native protein-destabilization step and being degradation cargo. The positive N-terminal deletion and CARD8 fusion evidence remains preserved. It establishes an engineered transferable susceptibility element but does not establish the native D2HGDH process assignment. PMID:33053349
For PMID:20020533, the abstract reports mutations in 24 of 50 patients, rather than 24 distinct mutations or 24 individually expressed mutant constructs. All mutation-positive patients had impaired enzyme activity. Patients with normal enzyme activity had no detected mutations; the converse, that every mutation-negative patient had normal activity, is not asserted. This patient-enzyme evidence supports the retained catalytic annotation with curator deference. It narrows the earlier combined clinical-source wording: PMID:16037974 reports mutant overexpression, whereas PMID:20020533 reports patient measurements. Detailed original protocols for those two sources remain unavailable.
The rat study PMID:15070399 provides a direct experimental observation relevant to the matrix inference: D2HGDH activity is readily released from mitochondria by freezing and thawing. Its Discussion interprets this as probable matrix solubility. That ortholog evidence now explicitly supports the retained human matrix annotation. Human MitoCoP identifies D2HGDH and categorizes it as peripherally associated using extraction measurements; the category is not a mutually exclusive compartment and does not contradict a releasable matrix pool. The matrix column in the workbook remains prior annotation context. Neither study establishes exclusive free solubility in living human cells. Exact cache excerpts occur once on the matrix annotation, with species and inference limits in its reason.
The two NOT ion-response rows deliberately remain unresolved. The negative rat enzyme observations are accepted, but they neither establish nor contradict the broader human cellular-process exclusion. The positive rows instead promote measured stimulation of isolated enzyme into an unshown cellular process and remain over-annotated. This asymmetry concerns the claim being adjudicated, not publication access; accessible full text can leave the cross-species negative-process inference unresolved. Neither action reverses an original NOT qualifier or denies a measured metal effect.
The primary source note makes the separately recovered PMID:15609246 Table 1 and assay notes auditable through the original public URL and DOI. The annotation carries an exact abstract excerpt, while the cache remains abstract-only. Actual external full-text recovery means the reference's full_text_unavailable flag remains false. The source note is authored navigation, not an independent experiment and not a substitute quotation fabricated from an unavailable cache.
D-lactate remains a weak secondary reaction supported by kinetics, binding and independent rat substrate experiments, without a physiological clearance claim. Both existing FAD-binding mappings remain within the single catalytic core because the evidence concerns the redox cofactor and its oxidized state; this is not a claim that ancestor breadth creates extra coverage. The electron-acceptor question now tests the specific ETF/ETF-dehydrogenase hypothesis already proposed in the 2004 Discussion. The genuine provider failures and manually researched source scope remain as previously documented. All 27 original source objects, both NOT flags, three products, annotation actions and the existing core are unchanged.