This authored note records where the original experiments can be checked. It does not replace the primary papers or claim that an abstract-only publication cache contains the full article. No publication cache was modified.
Achouri et al. (2004), PMID:15070399, original PMC1133759, DOI 10.1042/BJ20031933. Relevant locators are Experimental, Subcellular fractionation; Results, mitochondrial fractionation; and Discussion, Physiological and pathophysiological implications. The fractionation used rat liver. Activity concentrated in mitochondrial fractions; the Discussion reports release after freezing and thawing and uses this observation to infer probable matrix solubility. This is biochemical support for a rat matrix pool, not direct imaging of the human matrix or a measured human targeting-peptide cleavage site.
Morgenstern et al. (2021), PMID:34800366, original PMC8664129, DOI 10.1016/j.cmet.2021.11.001. Table S1, sheet A, row 290, protein group 8117, identifies Q8N465/D2HGDH. Column 76 is the Peripheral category; columns 77 and 78 respectively contain 13.830724243501532 and 65.80154063856689 percent. The Column explanations sheet, rows 77–79, defines the category through clustering and those two quantities as the means of carbonate and sonication pellet-fraction measurements. These are protocol-specific extraction measurements, not percentages of D2HGDH proved to occupy a particular mitochondrial compartment in vivo. The separate matrix annotation column represents prior compartment knowledge. The existing evidence index retains the literal target cells and their headers.
Peripheral association need not exclude matrix residence or a pool released by freeze–thawing. The rat and human observations use different preparations and methods; they do not establish equal partitioning, permanent integral-membrane insertion, exclusive free solubility or human membrane sidedness. The human Reactome matrix assignment remains a curated inference corroborated by this bounded biochemical context.
Struys et al. (2005), PMID:15609246, original PMC1196381, DOI 10.1086/427890. The original HTML was independently recovered and its Table 1, caption, note and footnote were personally read. The repository publication cache remains abstract-only. The annotation therefore quotes the cache abstract and points here for the external-original assay scope.
Table 1 compares wild-type human enzyme with Val444Ala and Ile147Ser after transient expression in HEK-293 cells. Both mutants have less activity than wild type. The note specifies sequence-confirmed mutant plasmids, a radiochemical assay with 2 micromolar D-2-hydroxyglutarate and no added cobalt. The footnote states that nontransfected-cell activity was subtracted and gives means with SEM from three or four transfections. The study also discusses the difficulty of resolving the D- and L-enzyme contributions in low-activity patient samples. No absolute rate-unit reinterpretation, activity assignment to the two short splice products or physiological electron-acceptor identity is inferred here. Full-text access and cache completeness are separate facts.
Yang et al. (2021), PMID:33431826, original PMC7801405, DOI 10.1038/s41421-020-00227-0. Results, Substrate specificity and stereo selectivity, Figure 3 and Methods, Enzymatic activity assay, report weak human D-lactate turnover together with kinetics and measurable binding. This evidence, plus the independent rat turnover and competition in Achouri et al., supports a secondary performed reaction. It is stronger than merely comparing two small signals, but does not establish substantial D-lactate flux in vivo. The long human construct, residues 51–521, was assayed; the short products were not.
The oxidized FAD pocket and the redox-cofactor cycle are represented by the two existing cofactor-binding terms within one catalytic core. Their retention is not an assertion that a broader term adds independent logical coverage. Achouri et al., Discussion, Physiological and pathophysiological implications, proposes ETF/ETF-dehydrogenase transfer from the clinical accumulation pattern in glutaric aciduria type II. That is a specific candidate to test, not a direct D2HGDH-to-ETF transfer assay.