Hmgcs2 catalyzes the condensation of acetyl-CoA with acetoacetyl-CoA to form
HMG-CoA in the mitochondrial matrix. This is the committed and rate-limiting
step of ketogenesis. The product HMG-CoA is cleaved by HMG-CoA lyase (Hmgcl) to
yield acetoacetate (the first ketone body) and acetyl-CoA.
Key references:
- PMID:1971108
- PMID:8097464
Glucagon activates HMG-CoA synthase by decreasing succinylation. In fed rats,
the enzyme is ~40% succinylated and inactive; glucagon lowers succinyl-CoA and
decreases succinylation to <10%, activating the enzyme.
- PMID:1967579
Insulin inhibits HMGCS2 gene expression through the forkhead transcription
factor FKHRL1 (FOXO3a). An FKHRL1-responsive element (AAAAATA) at -211 bp
mediates transcriptional repression.
- PMID:12027802
Protein rapidly increases in response to cAMP, dexamethasone, starvation, fat
feeding, and diabetes; decreased by insulin and refeeding.
- PMID:7902069
Dexamethasone effects: decreases mit. HMG-CoA synthase activity in suckling
rat liver and intestine.
- PMID:9546617
- PMID:9798904
UniProt notes that SIRT5 desuccinylates HMGCS2; succinylation at Lys-83 and
Lys-310 inhibits activity.
Expressed in liver, intestine, and kidney of suckling rats. Postnatal increase
correlates with glucagon levels. Expression disappears from intestine and kidney
upon weaning to high-carbohydrate diet but is re-induced by high-fat diet.
- PMID:8099282
- PMID:8620869
Atypical expression in subcutaneous adipose tissue of male rats, dependent on
age (from 9 weeks) and sex (higher in males). Testosterone-dependent.
- PMID:10357839
Expression modulated by bacterial species, particularly butyrate-producing
bacteria (Clostridium paraputrificum).
- PMID:14686922
Most IEP annotations represent expression changes (mRNA or protein levels)
in response to various stimuli. These are valid IEP annotations as they show
differential expression, but they do NOT indicate direct involvement of
Hmgcs2 in those processes -- they indicate the gene is regulated by those
conditions as part of metabolic adaptation.
The BioReason SFT document accurately identifies:
1. The thiolase-like fold architecture
2. The core HMG-CoA synthase catalytic activity
3. Mitochondrial matrix localization
4. Connection to ketogenesis
However, it makes several errors:
- The UniProt summary incorrectly states HMG-CoA is "converted to acetoacetate by
HMG-CoA reductase (HMGCR)". Actually HMG-CoA lyase (HMGCL) converts HMG-CoA to
acetoacetate. HMGCR converts HMG-CoA to mevalonate in the cytosol. This is a
significant factual error in the BioReason document.
- The "GO Term Predictions" sections are empty -- no actual predictions were made
- BioReason does NOT cite any specific PMIDs to support its claims
The GO:0010142 (farnesyl diphosphate biosynthetic process, mevalonate pathway) and
GO:0008299 (isoprenoid biosynthetic process) annotations are from phylogenetic/
InterPro inference. While the cytosolic HMGCS1 feeds the mevalonate pathway for
sterol/isoprenoid synthesis, the mitochondrial HMGCS2 predominantly feeds
ketogenesis. The IBA annotation lumps both paralogs. These annotations are
misleading for the mitochondrial isoform.
This is an IEA annotation from UniProtKB-KW. The mitochondrial isoform is not
directly involved in cholesterol biosynthesis; that is the cytosolic HMGCS1.
This is a clear case of paralog confusion.
ISS/ISO from human P54868. Not informative about specific function.
Reviewed all 55 rows against the 29 existing references, available primary abstracts, UniProt and PTHR43323. Newly retrieved PMID:7961793 shows mitochondrial HMG-CoA synthase restoring growth and acetate-to-cholesterol labeling in engineered mevalonate-auxotrophic CHO cells. This disproves an absolute compartment barrier but does not establish physiological rat isoprenoid flux; the mevalonate/FPP and isoprenoid rows are UNDECIDED pending adjudication. The GOA source PTN000222418 remains in the current PAINT slice for synthase/acetyl-CoA functions, but its FPP assertion is absent; this version discrepancy cannot prove why PAINT changed. Acetyl-CoA metabolism, acyltransferase activity and mitochondrial parent terms are valid core properties, and homodimer-based identical-protein binding is retained as non-core. Lung-development, adipose-development and amino-acid-response judgments are UNDECIDED because abstract-only coverage cannot overrule the curated specific experimental claims. The temperature-response challenge is retained in the conditions explicitly reported as showing no effect. New source: PMID:7961793(https://pubmed.ncbi.nlm.nih.gov/7961793/).
Read the entire mevalonate-and-isoprenoid-synthesis/openscientist.md report and both delivered CSV artifacts. The report supports the established ketogenic core and emphasizes the HMGCR/mitochondrial compartment distinction. It also acknowledges absence of native compartment-resolved isotope tracing. Those findings do not establish universal exclusion from the pathway. Its recommended removal/NOT rests partly on a failed literature search: PMID:7961793, independently checked against PubMed, explicitly reports cholesterol labeling after mitochondrial HMG-CoA synthase complementation and mitochondrial immunolocalization. The cached paper is abstract-only; no unobserved transport mechanism, rat construct identity or physiological in-vivo flux is inferred. The engineered system is evidence of capacity, not a new core rat pathway assertion.
Live QuickGO returned 55 P22791 annotations, including GO:0008299 IEA (InterPro IPR000590/IPR010122), contrary to the report and its CSV. The earlier GO:0010142 rows are absent from that current release, matching the previously observed PAINT-slice drift, but the original source rows remain intact. Snapshot: projects/IBA_REVIEW/rereview-2026-09-20/hmgcs2-live-annotations.json. The report's evidence-code hierarchy is not a substitute for an ancestral-node argument. All three reviewed pathway rows remain UNDECIDED; no NOT assertion is added. Native flux and curator clarification remain follow-up questions.