HMG-CoA Synthase: How Isoforms Regulate Ketogenesis

HMG-CoA synthase is the enzyme that builds 3-hydroxy-3-methylglutaryl-coenzyme A, a molecule sitting at a critical fork in metabolism. Depending on which version of the enzyme is at work and where it operates in the cell, HMG-CoA either feeds into cholesterol production or gets converted into ketone bodies, the backup fuel your body burns when glucose runs low. Humans carry two distinct forms of this enzyme, each with its own job, its own location, and its own set of consequences when something goes wrong. The biology around HMG-CoA synthase touches cholesterol management, fasting physiology, liver disease, cancer drug resistance, and even plant defense systems.

Two Isoforms, Two Very Different Jobs

The reason HMG-CoA synthase comes up in so many different medical contexts is that mammals have two versions of it, encoded by separate genes. HMGCS1, the cytosolic isoform, operates in the cell’s main compartment and feeds the mevalonate pathway. Mevalonate is a precursor for cholesterol and a family of molecules called isoprenoids that are involved in cell signaling, membrane structure, and protein modification.¹ HMGCS1 represents the first committed step of that pathway, making it a gatekeeper for cholesterol synthesis.²

HMGCS2, the mitochondrial isoform, lives inside the cell’s energy-producing compartments and has an entirely different purpose: making ketone bodies. It is the rate-limiting enzyme in ketogenesis, meaning it controls the pace at which the liver converts fatty acids into ketones like beta-hydroxybutyrate.³ The liver is the dominant ketone factory; although the kidney also turns on HMGCS2 during fasting, studies using kidney-specific knockout mice found only a trivially small drop in circulating ketones, while liver-specific knockouts wiped out fasting ketosis entirely.⁴ HMGCS2 also appears in the small intestine, particularly in suckling animals, where ketone production supports the energy demands of early development.⁵

How the Enzyme Works

Both isoforms belong to the thiolase-fold family, a group of enzymes that share a common structural blueprint built around a long internal tunnel where reactions take place. At the bottom of that tunnel sits a cysteine residue, which is the key player in the chemistry. In the bacterial version from Staphylococcus aureus, this is Cys-111, flanked by a histidine and an asparagine that together form a catalytic triad. The histidine activates the cysteine so it can launch a chemical attack on incoming acetyl-CoA, stripping off the acetyl group and attaching it to itself.⁶

The reaction proceeds in three stages. First, the enzyme grabs an acetyl group from acetyl-CoA and holds it on that active-site cysteine. Second, acetoacetyl-CoA arrives and condenses with the tethered acetyl group. Third, the finished product, HMG-CoA, is released through a water-assisted cleavage. Structural studies that captured the enzyme mid-reaction showed that a single glutamic acid residue acts as a general base in both the condensation and the release steps, performing double duty across the catalytic cycle.⁷ Crystal structures from Enterococcus faecalis confirmed that the coenzyme A portion and the acetoacetate fragment end up covalently linked to the active-site cysteine through a thioester bond during the process.⁸

How the Body Tunes HMGCS2 During Fasting

Because HMGCS2 controls how fast the liver makes ketones, the body needs to crank it up when food is scarce and dial it back when glucose is plentiful. One of the main ways this happens is through a chemical tag called acetylation. In mice lacking the protein SIRT3, a mitochondrial enzyme that removes acetyl groups, HMGCS2 becomes heavily acetylated at three specific lysine positions and its activity drops. When mice fast and SIRT3 is present, it strips those acetyl marks off HMGCS2, boosting ketone production. Mice without SIRT3 show lower beta-hydroxybutyrate levels during fasting, confirming that this deacetylation matters for normal ketone output.⁹

Acetylation is not the only chemical switch. Another modification called succinylation also suppresses HMGCS2 activity, and the protein SIRT5 removes those marks. In one study, the short-chain fatty acid sodium butyrate stimulated HMGCS2 activity by promoting SIRT5-mediated desuccinylation at two specific sites. The effect was significantly reduced in mice lacking SIRT5, confirming that this is a distinct regulatory pathway from the SIRT3 system.¹⁰

For the cholesterol-pathway isoform HMGCS1, regulation happens largely at the level of gene expression. A transcription factor called SREBP2 acts as a master switch for cholesterol biosynthesis genes, including HMGCS1. When cholesterol levels in the cell drop, SREBP2 ramps up the expression of HMGCS1, HMGCR (the statin target), and other enzymes in the pathway. Conversely, research has shown that overexpression of the protein TDP-43 can suppress SREBP2 and, downstream, reduce HMGCS1 expression.¹¹

When HMGCS2 Is Missing From Birth

Inherited deficiency of mitochondrial HMG-CoA synthase is extremely rare, but the consequences are serious. Because the enzyme is the bottleneck for ketone production, children who lack functional HMGCS2 cannot mount a ketone response when they need it most, typically during illness or prolonged fasting. The classic presentation includes dangerously low blood sugar without the expected rise in ketones (called hypoketotic hypoglycemia), along with metabolic acidosis, an enlarged fatty liver, and brain dysfunction that can progress to coma.¹² Episodes are typically triggered by catabolic stress: a stomach bug, a fever, or simply going too long without eating.

One complicating factor for diagnosis is that low blood sugar is not always present. A case report of a 20-month-old who presented with brain dysfunction and metabolic acidosis after a brief illness noted that the child was actually normoglycemic before receiving intravenous glucose. The blood sample was extremely fatty, with high triglycerides. This suggests clinicians should not rule out HMGCS2 deficiency solely because blood sugar looks normal at presentation.¹³

Diagnosis relies heavily on urine organic acid testing. During episodes of metabolic crisis, affected children excrete a distinctive pattern of metabolites. A compound called 4-hydroxy-6-methyl-2-pyrone (4-HMP) stands out as one of the most specific markers. When researchers applied cutoffs for 4-HMP and another metabolite in urine samples, they flagged all decompensation episodes from confirmed HMGCS2-deficient patients with a positive predictive value of 80%.¹⁴ Plasma findings can also help: elevated levels of acetylcarnitine and a specific hydroxylated carnitine species may be detectable during acute episodes, and urine organic acids can sometimes reveal the diagnosis even when the critical blood sample was collected late.¹⁵

Fatty Liver Disease and the Ketogenesis Connection

The relationship between HMGCS2 and liver health extends well beyond rare genetic disorders. When HMGCS2 is disrupted in mouse models, the liver loses its ability to efficiently burn fat, and the result looks a lot like metabolic dysfunction-associated steatotic liver disease (MASLD, formerly known as non-alcoholic fatty liver disease). Loss-of-function models showed that knocking out HMGCS2 impaired overall hepatic fat oxidation and produced a liver phenotype resembling steatohepatitis.¹⁶

More recent work has pushed the consequences beyond the liver itself. Mice lacking HMGCS2 that were fed a high-fat diet showed accelerated weight gain, worsened glucose intolerance, elevated serum triglycerides and cholesterol, and greater hepatic fat accumulation compared to normal mice on the same diet. The study also found that these mice developed worse cognitive impairment, linking the liver’s beta-hydroxybutyrate production to brain health via reduced hippocampal insulin resistance and neuroinflammation.¹⁷ The implication is that ketogenesis is not just a backup energy system. It appears to play an active role in keeping the liver healthy and protecting distant organs.

HMGCS1 and Cancer Drug Resistance

Cancer cells often rewire their metabolism, and both HMG-CoA synthase isoforms have turned up in surprising roles in tumor biology. For HMGCS1, the cholesterol-pathway enzyme, research has connected it to drug resistance in several cancer types. In acute myeloid leukemia cells, HMGCS1 was shown to protect against the cytotoxic effects of standard chemotherapy drugs. Cells that became resistant to cytarabine or adriamycin were actually more sensitive to HMGCS1 inhibition than their non-resistant counterparts, suggesting that these resistant cells had become more dependent on the enzyme. The mechanism involved HMGCS1 protecting the cell’s internal stress-response machinery, specifically the endoplasmic reticulum and mitochondria, from damage under chemotherapy-induced stress.¹⁸

An even more unexpected finding came from cervical cancer research, where HMGCS1 was found to physically relocate into mitochondria in cisplatin-resistant cells. Directing HMGCS1 specifically to mitochondria, rather than to the nucleus or the cell’s main compartment, was enough on its own to confer cisplatin resistance. This mitochondrial HMGCS1 regulated mitochondrial gene transcription and respiratory function, a role no one had previously attributed to this enzyme. Depleting HMGCS1 or blocking mitochondrial transcription re-sensitized the resistant cells to cisplatin, and combining cisplatin with inhibitors targeting either approach produced effects greater than either treatment alone.¹⁹

On the HMGCS2 side, the story is more about silencing than overexpression. In clear cell renal carcinoma, researchers found that the HMGCS2 gene’s promoter region was heavily methylated, a chemical modification that effectively shuts gene expression down. Higher methylation correlated with worse outcomes for patients. Treating the cells with a demethylating agent restored HMGCS2 transcription, suggesting the enzyme may normally act as a brake on tumor cell movement and that cancer cells gain an advantage by turning it off epigenetically.

Inhibitors and Drug Development

The most well-known inhibitor of HMG-CoA synthase is hymeglusin, a naturally occurring beta-lactone compound originally isolated from fungi. Hymeglusin works by permanently attaching itself to the active-site cysteine, the same residue the enzyme uses to grab its acetyl-CoA substrate. Studies with purified rat liver and hamster enzymes showed that hymeglusin covalently binds to Cys-129, and the enzyme’s activity was completely eliminated when roughly two molecules of hymeglusin attached per molecule of enzyme.²⁰

Hymeglusin also works on bacterial HMG-CoA synthases, which matters because certain pathogens rely on the mevalonate pathway for essential metabolites. The crystal structure of hymeglusin bound to the Enterococcus faecalis enzyme revealed that the inhibitor fits into a narrow tunnel almost completely sealed off from the surrounding water, making the bond particularly stable compared to the more exposed binding pocket in eukaryotic versions of the enzyme.²¹ This structural difference could be exploited to design inhibitors selective for bacterial enzymes, potentially leading to new antibiotics. High-throughput screening has already identified small molecules that inhibit bacterial HMG-CoA synthase with sub-micromolar potency.²²

On the therapeutic side for human disease, the same Cys-129 that hymeglusin targets has attracted interest in lipid-lowering drug development. Ligustilide, a compound found in several traditional medicinal plants, was shown to irreversibly bind Cys-129 of HMGCS1 through a reactive metabolic intermediate. This binding reduced cholesterol synthesis and improved lipid metabolism in experimental models of dyslipidemia. The approach is conceptually different from statins, which block HMGCR (the enzyme one step downstream). Targeting HMGCS1 instead could offer an alternative or complementary strategy for managing cholesterol.

Brown Fat and Thermogenesis

Both HMG-CoA synthase isoforms play a role in the function of brown adipose tissue, the fat that generates heat rather than storing energy. Knocking down either mitochondrial HMGCS2 or cytosolic HMGCS1 in brown fat cells led to marked reductions in UCP1, the protein that drives heat production, along with significant drops in uncoupled mitochondrial respiration, the metabolic process underlying thermogenesis.²³ The fact that both isoforms matter is striking. It suggests the mevalonate pathway contributes to brown fat function through more than just cholesterol production, likely by supplying isoprenoid intermediates needed for signaling and membrane assembly in these metabolically demanding cells.

HMG-CoA Synthase in Plants

Plants have their own version of HMG-CoA synthase, and it occupies a similarly strategic position in their metabolism. In plants, the mevalonate pathway supplies precursors for sterols and terpenoids, compounds that play structural roles in membranes and serve as chemical defenses. Overexpressing HMG-CoA synthase from Brassica juncea (Indian mustard) in the model plant Arabidopsis ramped up sterol biosynthesis genes and increased the production of sterols like beta-sitosterol and stigmasterol. These overexpressing plants germinated earlier, showed less cell death in response to oxidative stress, and were more resistant to the fungal pathogen Botrytis cinerea.²⁴

Interestingly, the enzyme’s effects on plant defenses involve a trade-off. Overexpression of HMG-CoA synthase promotes root growth but suppresses the production of glucosinolates, sulfur-containing compounds that many plants use as pest deterrents. Conversely, inhibiting the enzyme reduced sterol levels but boosted glucosinolate biosynthesis.²⁵ This metabolic tug-of-war between growth-promoting sterols and defensive glucosinolates illustrates a broader principle in plant biology: resources channeled into one branch of metabolism often come at the expense of another.

An Ancient Gene Duplication That Changed Vertebrate Biology

The existence of two HMG-CoA synthase isoforms in mammals is not an accident of recent evolution. Comparative analysis of all known mitochondrial and cytosolic HMG-CoA synthase protein sequences shows a high degree of conservation near one end of the protein that decreases progressively toward the other, a pattern consistent with the two forms having diverged from a single ancestor gene somewhere between 400 and 900 million years ago. The human mitochondrial isoform’s gene maps to chromosome 1p12-p13.²⁶

The researchers who mapped this divergence proposed something intriguing: the gene duplication that produced the mitochondrial isoform may have been one of the enabling events in early vertebrate evolution. Before the duplication, fatty acid breakdown and leucine breakdown both produced acetoacetyl-CoA, but there was no efficient way to convert that into ketone bodies inside mitochondria. The appearance of HMGCS2 inside mitochondria created the ketogenesis pathway, linking pre-existing metabolic routes and generating a lipid-derived fuel source that the energy-hungry vertebrate brain could use during fasting. If that hypothesis holds, the ketone body system that powers your brain between meals exists because of an ancient copy-paste event in a genome hundreds of millions of years ago.