Repository navigation
Cholesterol synthesis, hydroxylation and bile acid synthesis reactions: enzyme and compartment corrections #1108
Description
Several cholesterol hydroxylation and bile acid synthesis reactions had the wrong enzyme, the wrong compartment or chemistry that cannot happen. #1106 removes impossible or duplicated reactions writes the CYP27A1 steps as mitochondrial monooxygenations, and removes cholesterol-synthesis shortcuts that bypassed HSD17B7 and MSMO1. Bile acid synthesis tasks still pass.
1. CH25H 26- and 20-hydroxylations removed (#1106)
| Change | Model entities | Evidence |
|---|---|---|
| ER cholesterol 26-hydroxylation with CH25H: deprecated | MAR01785 | CH25H only makes 25-hydroxycholesterol (UniProt O95992, PMID 9852097). The only human sterol 26-hydroxylase is mitochondrial CYP27A1, already modelled as MAR01762. 26-hydroxycholesterol still reaches the ER through MAR01824 |
| ER cholesterol -> 20-hydroxycholesterol with CH25H: deprecated | MAR01781 | Same as above. Cholesterol 20-hydroxylation is done by CYP11A1 in mitochondria. No ER 20-hydroxylase is known |
| 20-hydroxycholesterol[r] -> 7alpha-hydroxycholest-4-en-3-one (HSD3B1/2): deprecated. ER 20-hydroxycholesterol (MAM00592r) deprecated | MAR01783, MAM00592r | This was the only consumer of the MAR01781 product. As written, the reaction also moves the hydroxyl from C20 to C7alpha, which a 3beta-HSD cannot do |
2. Sterols converted to 7alpha-hydroxy-C4 without a 7alpha-hydroxylation (#1106)
Four reactions turned a sterol without a 7alpha-hydroxyl into 7alpha-hydroxycholest-4-en-3-one in one dehydrogenase step. They balance only because the formula changes cancel out. They also bypass CYP7A1/CYP7B1. HSD3B7 is active only on 7alpha-hydroxylated sterols (PMID 11067870, UniProt Q9H2F3).
| Reaction (before) | Change | Notes |
|---|---|---|
| MAR01787: 26-hydroxycholesterol[r] + NAD+ -> 7alpha-hydroxycholest-4-en-3-one, no gene | deprecated | The real route is CYP7B1 (MAR01574/MAR01765) followed by HSD3B7 (MAR01766), and both are in the model |
| MAR03259: the same reaction in [m], named as an aminoacylase (EC 3.5.1.14) | deprecated | No enzyme, and the EC does not match |
| MAR02184: 25-hydroxycholesterol[r] + NADP+ <=> 7alpha-hydroxycholest-4-en-3-one, HSD3B7 | deprecated (no merge target) | The correct HSD3B7 step of the 25-hydroxylation branch is MAR01778, which is fed by CYP7B1 (MAR01777). The cited PMID 12679481 does not describe 25-hydroxycholesterol |
| MAR02881: 20-hydroxycholesterol[m] + NAD+ <=> 7alpha-hydroxycholest-4-en-3-one, HSD3B1/2 | deprecated | HSD3B1/2 act on C19/C21 steroids (P14060, P26439) |
After these removals, two reactions each became the only reaction of their metabolite and would have been dead ends. Both were deprecated together with those metabolites:
- MAR01578: cholesterol -> 20-hydroxycholesterol in [m], no gene. MAM00592m deprecated.
- MAR02895: 5beta-cholestane-3alpha,7alpha-diol -> 7alpha-hydroxycholest-4-en-3-one in [m], no gene. MAM01182m deprecated.
3. CYP27A1 as a mitochondrial monooxygenase (#1106)
CYP27A1 is a mitochondrial inner-membrane P450 that uses O2 and adrenodoxin, which the model writes with NADPH (UniProt Q02318; Rhea 14373, Rhea 40231, Rhea 34627; PMID 11412116, PMID 1708392). The separate NAD(P)+ dehydrogenase ECs 1.1.1.161 and 1.2.1.40 were deleted by IUBMB because their activity is part of this monooxygenase.
| Change | Model entities | Notes |
|---|---|---|
| Tetrol -> 26-al written as a monooxygenation: tetrol + NADPH + H+ + O2 -> 26-al + NADP+ + 2 H2O. Was an NADP+ dehydrogenation | MAR01610 | Now carries KEGG R03507 and RHEA:40231, moved from MAR01608 |
| NAD+ dehydrogenase copy in [m] merged | MAR01609 -> MAR01610 | |
| NADH-driven copy of the aldehyde oxidation merged | MAR01613 -> MAR01614 | CYP27A1 takes its electrons from NADPH via adrenodoxin. The NADH copy formed an NADH -> NADPH loop with MAR01614 |
| Reversible NADP+ aldehyde dehydrogenation with CYP27A1 in [m] merged | MAR03214 -> MAR01614 | |
| Cytosolic cholesterol 26-hydroxylase with CYP27A1 merged | MAR01786 -> MAR01762 | Cholesterol and 26-hydroxycholesterol cross c <-> m (MAR01915, MAR01764) |
| Cytosolic triol 26-hydroxylation with CYP27A1 merged | MAR01604 -> MAR01605 | The triol and tetrol cross c <-> m (MAR01602, MAR01772) |
| Cytosolic reversible NADP+ aldehyde oxidation, no gene, merged | MAR03210 -> MAR01611 | |
| Cytosolic tetrol + NAD+ -> 26-al: GPR CYP27A1 -> ADH1B, EC 1.1.1.1 | MAR01608 | Purified human liver ADH beta2beta2 oxidises this tetrol (PMID 6338006). There are also rat liver cytosol data (PMID 5723340). The evidence is in vitro only |
| Cytosolic 26-al + NAD+ -> 26-oate: gene and EC removed | MAR01611 | A mitochondrial monooxygenase cannot catalyse a cytosolic NAD+ reaction. The reaction is kept because rat liver cytosol performs it (PMID 5914340). No human enzyme has been identified |
4. Cholesterol synthesis: 3-ketosteroid reduction and C4 demethylation (#1106)
In the second C4-demethylation round of cholesterol synthesis, HSD17B7 reduces 3-keto-4-methylzymosterol, MSMO1 oxidises the remaining 4alpha-methyl group to a carboxylate, NSDHL decarboxylates it to zymosterone, and HSD17B7 reduces zymosterone to zymosterol. HSD17B7 is the only 3-ketosteroid reductase of this pathway (PMID 12829805, UniProt P56937), and all three enzymes are in the ER membrane (UniProt). In the model, the ER copy of this pathway (from Recon3D) bypassed HSD17B7 and MSMO1 through reactions that no enzyme carries out, so HSD17B7 was essential in some ftINIT context models and not in others (gene-essentiality run on #1095).
| Change | Model entities | Evidence |
|---|---|---|
| Zymosterone -> zymosterol [r]: GPR HSD17B4 -> HSD17B7, EC 4.2.1.17 -> 1.1.1.270 | MAR00115 | RHEA:33459 is HSD17B7 in UniProt, citing PMID 12829805 (the reaction's own reference) and PMID 11165030. HSD17B4 is the peroxisomal D-bifunctional protein, and EC 4.2.1.17 is enoyl-CoA hydratase. The cytosolic twin MAR01509 already had HSD17B7 |
| 3-keto-4-methylzymosterol reduction written with the correct cofactor. Before: 3-keto-4-methylzymosterol + NADP+ + 3 H+ <=> 4alpha-methylzymosterol + NADPH (a reduction that made NADPH, charge -4). After: 3-keto-4-methylzymosterol + NADPH + H+ <=> 4alpha-methylzymosterol + NADP+ | MAR01500 [c], MAR02301 [r] | RHEA:36379, HSD17B7 in UniProt (PMID 20659585) |
| Deprecated: 3-keto-4-methylzymosterol + NAD(P)+ + O2 -> zymosterone + CO2 + NAD(P)H (Recon3D C4STMO2Pr/C4STMO2r, GPR NSDHL or MSMO1) | MAR00117, MAR00118 | Lumps the second demethylation into one step that skips the HSD17B7 reduction and the MSMO1 oxidation (3 O2 and 3 NAD(P)H per methyl group) and even produces NAD(P)H. MetaNetX MNXR96419 carries the same lumped equation. The correct steps exist in the cytosolic copy (MAR01500, MAR01502-MAR01505, MAR01509) |
| Deprecated: 4alpha-methylzymosterol + H2O -> zymosterol + CO2 (no gene) | MAR02302 | Removes a methyl group as CO2 without an oxidant |
| MSMO1 EC 4.2.1.17 -> 1.14.18.9 | MAR00116 | UniProt Q15800 |
With the shortcuts removed, the ER copy stops at 4alpha-methylzymosterol[r] and MAR00115 has no substrate; cholesterol is made by the complete cytosolic copy. The two copies are reconciled in #1117.
Checks
All touched reactions stay mass and charge balanced. The essential and verification tasks pass, as do the bile acid synthesis tasks (glyco- and tauro-cholate and chenodeoxycholate). Each metabolite in these blocks keeps at least one producer and one consumer.
Open points (not changed)
- MAR03817 (mitochondrial 26-oate + NADH -> 26-al, ALDH2/ALDH1B1/ALDH7A1) is written as the reduction of the acid back to the aldehyde. The direction and the genes need a separate decision.
- The pentaol node MAM01091 (MAR01804, MAR01805, MAR01807, MAR03774) looks like the 26-aldehyde written as its hydrate.
- MAR01778 and MAR01585 (HSD3B7) have direction and cofactor questions: all documented HSD3B7 activities are NAD+-dependent oxidations.
- MAR01765 (CYP7B1) is cytosolic, but CYP7B1 is an ER enzyme.
Implemented in #1106 (sections 1-4). The GPRs that #1090 joined on some of these reactions are corrected in #1105 (see the comment on #626).
Activity
[-]Cholesterol hydroxylation and bile acid synthesis reactions: enzyme and compartment corrections[/-][+]Cholesterol synthesis, hydroxylation and bile acid synthesis reactions: enzyme and compartment corrections[/+]edkerk commentedon Oct 5, 2026
Fixed in Human 2.1.0.