ERG19 / MVD1 (Saccharomyces cerevisiae) — Review Notes

UniProt: P32377 (MVD1_YEAST) · SGD: S000005326 (MVD1) · ORF: YNR043W · Chr XIV
Standard SGD name ERG19; aliases MVD1, MPD. PDB: 1FI4.

Identity and core function

ERG19 encodes diphosphomevalonate decarboxylase (mevalonate-5-diphosphate
decarboxylase; MDD/MVD), EC 4.1.1.33. It catalyzes the third and final step
of the mevalonate module of isoprenoid/ergosterol biosynthesis:

(R)-5-diphosphomevalonate + ATP → isopentenyl diphosphate (IPP) + ADP + CO2 + phosphate
(Rhea:RHEA:23732) [UniProt; PMID:8626466 "recombinant ... enzyme is a homodimer of 43-kDa subunits with a specific activity of 2.4 units/mg"]

This is an ATP-dependent decarboxylation: ATP phosphorylates the 3-hydroxyl of
diphosphomevalonate, and the resulting phosphate is eliminated together with the
carboxyl group (decarboxylation), yielding IPP. So although classed as a lyase
(carboxy-lyase, EC 4.1.1.x), the chemistry involves an ATP-dependent
phosphotransfer step — the enzyme is part of the GHMP kinase superfamily
(galactokinase, homoserine kinase, mevalonate kinase, phosphomevalonate kinase)
PMID:15169949.

IPP produced here is the universal C5 building block of all isoprenoids; in yeast
the downstream pathway makes farnesyl diphosphate (FPP), then squalene →
lanosterol → ergosterol, plus dolichol, ubiquinone, heme A, and substrates
for protein prenylation [UniProt FUNCTION; PMID:28904410].

Essentiality and genetics

• Essential gene for viability. ERG19 disruption is lethal; ergosterol
  auxotrophy in the pathway cannot bypass the block at this step PMID:9244250.
• A temperature-sensitive erg19 allele carries a single Leu79→Pro change in a
  conserved sequence, conferring thermosensitivity [PMID:9244250; UniProt MUTAGEN L79].
• Original erg19 mutant strains were isolated as ergosterol auxotrophs blocked in
  MDD PMID:1779710.
• High-level overexpression of wild-type ERG19 lowers steady-state sterol
  accumulation, suggesting a regulatory role in the mevalonate pathway PMID:9244250.

Structure

• Homodimer [PMID:11698677 SUBUNIT; PMID:8626466].
• Crystal structure PDB 1FI4, 2.27 Å, residues 1–396, solved as part of a
  structural-genomics effort on sterol/isoprenoid enzymes PMID:11698677.
• Two-domain GHMP-kinase-like fold (N-terminal + GHMP kinase C-terminal domain;
  Gene3D 3.30.230.10 + 3.30.70.890; SUPFAM ribosomal-S5-D2-like + GHMP kinase C).
• Substrate-binding residues (modeled by similarity): 19–22, 74, 153–158, 209
  (bind (R)-5-diphosphomevalonate) [UniProt FT BINDING].

Active-site / catalytic residues

• Asp302 is a crucial catalytic residue: D302N and D302A reduce k_cat ~10^3 and
  ~10^5-fold respectively, while still binding nucleotide — assigns a catalytic
  (general base) role to Asp302 PMID:15169949.
• Lys18 influences the active site (K18M: 30-fold activity decrease, 16-fold
  higher K_m for ATP) but is not essential for chemistry PMID:15169949.
  (Note: this study was on the GHMP/MDD active site generally; numbering per the
  enzyme studied — the conserved Asp/Lys are present in the yeast enzyme.)

Localization

• Bulk proteomic/GFP data place the protein in the cytoplasm/cytosol
  [PMID:14562095 HDA cytoplasm; IBA cytosol GO_Central].
• A vacuole annotation (NAS) derives from the ergosterol-biosynthesis review,
  which describes the FPP/"second module" enzymes as acting in the vacuole
  PMID:28904410. This is a review-level, non-experimental assertion and conflicts
  with the high-throughput cytosolic localization; treat as weaker evidence. The
  UniProt comment likewise states the second module is "carried out in the vacuole"
  but cites the review, not direct localization of MDD.
• Abundance: ~13,700 molecules/cell in log-phase SD medium [UniProt MISCELLANEOUS,
  PMID:14562106].

Pathway placement (mevalonate module)

acetyl-CoA → (HMG-CoA, HMG-CoA reductase ERG13/HMG1/HMG2) → mevalonate →
(ERG12, mevalonate kinase) → mevalonate-5-P → (ERG8, phosphomevalonate kinase) →
mevalonate-5-PP → (ERG19/MVD1) → IPP → (IDI1 isomerase) → DMAPP →
(ERG20, FPP synthase) → FPP → sterol/ergosterol + dolichol + ubiquinone + prenylation.

UniPathway: UPA00057 / UER00100 (IPP from (R)-mevalonate, step 3/3).

Annotation review orientation

• MF (diphosphomevalonate decarboxylase activity, GO:0004163) — well supported
  by direct biochemistry (IDA PMID:8626466, PMID:15169949) and IMP (PMID:9244250,
  PMID:1779710). Clear core function. The multiple redundant IEA/RCA copies are
  acceptable but the IDA/IMP-backed ones are the primary evidence.
• carboxy-lyase activity (GO:0016831) — correct parent of GO:0004163, but a
  generalization; MODIFY/redundant relative to the specific term.
• BP terms: isopentenyl diphosphate biosynthetic process, mevalonate pathway
  (GO:0019287) is the most precise direct-process term → core. Ergosterol
  biosynthetic process (GO:0006696), sterol biosynthetic process (GO:0016126),
  isoprenoid biosynthetic process (GO:0008299), FPP biosynthetic process,
  mevalonate pathway (GO:0010142) are broader / downstream pathway memberships —
  keep but assess core vs non-core (the enzyme's direct product is IPP, not FPP
  or ergosterol; downstream-process terms are pathway-context, candidates for
  KEEP_AS_NON_CORE rather than as the molecular role).
• CC: cytosol/cytoplasm well supported; vacuole is weak (NAS from a review).
• ATP binding (GO:0005524, UniProtKB-KW) is correct biochemically (ATP is a
  co-substrate) though not in the GOA stub list — present in UniProt DR lines.

Provenance note

Automated deep research succeeded on a third attempt:
ERG19-deep-research-falcon.md (falcon / "Edison Scientific Literature", 17 citations,
~19 min runtime). The first two attempts failed only because the wrapper's default
600 s cap killed the provider mid-run; re-running with --timeout 3600 let it finish.
(The WARNING - agentapi not found in PATH line in the logs is benign and unrelated to
the timeout.) The review is grounded in the UniProt record (P32377), the six cached
primary publications (PMID:8626466, 15169949, 9244250, 1779710, 28904410, 14562095),
the structural-genomics paper for PDB 1FI4 (PMID:11698677), the PANTHER family data
(PTHR10977), and this deep-research report.

Deep research (falcon) — corroboration & additions

The falcon report independently confirms the core review and contradicts nothing:

• Reaction / cofactors: confirms the ATP-dependent decarboxylation of
  mevalonate-5-diphosphate to IPP, and additionally states the reaction requires
  Mg2+ and is coupled to ATP hydrolysis (products IPP + CO2 + ADP)
  [Cordier 1999, doi:10.1023/a:1006181720100; Garay 2026 preprint,
  doi:10.20944/preprints202605.0182.v1]. This reinforces the added ATP-binding
  molecular function; Mg2+ dependence is family/review-level here (flagged as inferred
  by the report) and is not annotated as metal binding in UniProt P32377, so no
  metal-ion GO term was added.
• Mechanism: GHMP-kinase-superfamily mechanism via ATP-dependent phosphorylation of
  the substrate C3-hydroxyl (transient 3-phospho-MVAPP) then decarboxylation/elimination
  — consistent with the Asp302/Lys18 active-site biochemistry [Garay 2026].
• Localization: independently described as a cytosolic homodimer / cytoplasmic
  precursor-module enzyme [Johnston 2020 Yeast, doi:10.1002/yea.3452; Garay 2026]. This
  further supports flagging the NAS vacuole annotation (GO:0005773) as over-annotated.
• Essentiality: an Arabidopsis MVD cDNA complements a yeast thermosensitive
  MVD-deficient strain and rescues the lethal ERG19 deletion [Cordier 1999] —
  additional support for essentiality alongside PMID:9244250.
• Regulation (context, non-core): ERG genes including ERG19 are controlled by
  Upc2/Ecm22 (sterol-responsive) and Hap1/Rox1/Mot3 (heme/oxygen); early-module genes
  (ERG8, ERG13, ERG19) are down-regulated as oxygen decreases [Jorda & Puig 2020 Genes,
  doi:10.3390/genes11070795 = PMID:32679672].
• Applied context (non-core): ERG19 is included in MVA-pathway overexpression
  cassettes for terpenoid production (e.g. 528 mg/L taxadiene, Karaca 2024) and is
  discussed as part of the essential fungal sterol axis for antifungal targeting
  (Gutierrez-Perez & Cramer 2025; fungal vs human homolog <50% identity).

None of these change the annotation actions; they corroborate the core MF/BP/CC calls
and the vacuole over-annotation flag.