GLUD1 (Glutamate dehydrogenase 1, mitochondrial) — review notes

UniProt: P00367 (DHE3_HUMAN). HGNC:4335. Gene: GLUD1 (Synonym GLUD). 558 aa precursor,
with an N-terminal mitochondrial transit peptide (residues 1..53); mature chain 54..558.
EC 1.4.1.3.

Deep research: GLUD1-deep-research-falcon.md is committed and was checked against this review on
2026-10-11. It agrees on the core annotation (a mitochondrial-matrix NAD(P)-dependent L-glutamate
dehydrogenase, EC 1.4.1.3, a homohexamer of two trimers with an antenna and pivot helix, GTP
inhibiting and ADP activating, leucine activating) and adds four points worth recording:

• It is explicit that leucine stimulates the enzyme but is not the amino acid converted, and that
  glutamine supplies glutamate only after an upstream glutaminase reaction. GLUD1 is not a
  glutamine-cleaving enzyme, a leucine dehydrogenase or a glutamate transporter.
• Reversibility does not fix a direction per tissue. The recombinant hGDH1 ammonia Km rises from
  12.8 mM at pH 8.0 to 35.0 mM at pH 7.5 and 57.5 mM at pH 7.0, so acidification reduces capacity
  for reductive amination. These are assay values, not tissue ammonia measurements.
• The Fang 2002 mutagenesis separates the regulatory mechanisms cleanly: half-maximal GTP
  inhibition at 42 nM for wild type against 210 microM for H454Y and 3.1 microM for S448P, both
  retaining ADP and leucine responses, while engineered R463A abolishes ADP activation with little
  effect on GTP or leucine. This supports annotating GTP inhibition and ADP activation separately.
• It keeps GLUD2 findings out of GLUD1's annotations, which matters for the renal and brain
  claims: both paralogs are in human proximal tubule, so the pathway-level estimate that the GDH
  route accounts for at least 25% of renal ammonia cannot be assigned to GLUD1 alone, and hGDH1 is
  predominantly glial while hGDH2 has substantial neuronal expression. It also warns against
  reading the 2023 SIRT4/ornithine transcarbamylase result as evidence about GLUD1: GLUD1 is not a
  urea-cycle enzyme.

Core biology (from UniProt P00367 and cited literature)

• Function. "Mitochondrial glutamate dehydrogenase that catalyzes the conversion of
  L-glutamate into alpha-ketoglutarate. Plays a key role in glutamine anaplerosis by
  producing alpha-ketoglutarate, an important intermediate in the tricarboxylic acid
  cycle" [file:human/GLUD1/GLUD1-uniprot.txt "Mitochondrial glutamate dehydrogenase that catalyzes the"].
  Also "Plays a role in insulin homeostasis" and "May be involved in learning and memory
  reactions by increasing the turnover of the excitatory neurotransmitter glutamate (By
  similarity)".
• Catalytic activity (dual cofactor). UniProt lists two reactions:
• L-glutamate + NAD(+) + H2O = 2-oxoglutarate + NH4(+) + NADH + H(+) (RHEA:15133), EC 1.4.1.3
• L-glutamate + NADP(+) + H2O = 2-oxoglutarate + NH4(+) + NADPH + H(+) (RHEA:11612), EC 1.4.1.3
  So the enzyme uses both NAD+ and NADP+ — the umbrella MF is GO:0004353
  "L-glutamate dehydrogenase [NAD(P)+] activity" (parents GO:0004352 NAD+ and GO:0004354 NADP+).
  PMID:11254391 (bovine GDH structures): "only animal GDH utilizes both NAD(H) or NADP(H)
  with comparable efficacy".
• Reaction direction / reversibility. Fang et al.: "Glutamate dehydrogenase (GDH)
  catalyses the reversible oxidative deamination of l-glutamate to 2-oxoglutarate in the
  mitochondrial matrix" PMID:11903050.
  Reactome models both directions (R-HSA-70600 forward, R-HSA-70589 reverse).
• Allosteric regulation. "Subject to allosteric regulation. Activated by ADP
  (PubMed:11903050). Inhibited by GTP and ATP ... ADP can occupy the NADH binding site and
  activate the enzyme (PubMed:16023112). Inhibited by SIRT4 (PubMed:16959573)"
  [file:human/GLUD1/GLUD1-uniprot.txt]. Leucine is an allosteric activator; ADP sensitizes
  the enzyme to leucine PMID:11032875.
  Nonactivated GLUD1 is potently inhibited by GTP (IC50 = 0.20 microM)
  PMID:11032875.
• Quaternary structure. Homohexamer [file:human/GLUD1/GLUD1-uniprot.txt "Homohexamer (By similarity)"];
  ADP-ribosylation occurs on one subunit per catalytically active homohexamer
  PMID:16023112.
• PTM / SIRT4. "SIRT4 is a mitochondrial enzyme that uses NAD to ADP-ribosylate and
  downregulate glutamate dehydrogenase (GDH) activity" PMID:16959573.
  ADP-ribosylation site is Cys172 in the mature-numbering UniProt feature (paper identifies
  the reactive Cys as "Cys119" in their synthetic-gene numbering) PMID:16023112.

Localization

• Mitochondrion + endoplasmic reticulum. UniProt SUBCELLULAR LOCATION: "Mostly translocates
  into the mitochondria, only a small amount of the protein localizes to the endoplasmic
  reticulum" [file:human/GLUD1/GLUD1-uniprot.txt "only a small amount of the protein"].
  Experimental basis PMID:19448744: "while most of the hGDHs translocate into the
  mitochondria (a process associated with cleavage of the signal sequence), part of the
  protein localizes to the endoplasmic reticulum" PMID:19448744.
• Reactome places the active enzyme in the mitochondrial matrix (GO:0005759).
• Cytoplasm (GO:0005737, IDA PMID:18688271): Rosso et al. showed "GLUD1 localizes to the
  mitochondria as well as the cytoplasm" whereas GLUD2 is mitochondria-specific
  PMID:18688271. This is a
  minor pool; the mitochondrial matrix is the primary site of function.

Disease

• Hyperinsulinemic hypoglycemia, familial, 6 (HHF6 / HI/HA syndrome; MIM:606762). Autosomal
  dominant, "hypoglycemia due to congenital hyperinsulinism combined with persistent
  hyperammonemia" [file:human/GLUD1/GLUD1-uniprot.txt "combined with"]. Caused by gain-of-function
  regulatory mutations that desensitize GDH to GTP inhibition: "Mutations of glutamate
  dehydrogenase cause the hyperinsulinism/hyperammonemia syndrome by desensitizing glutamate
  dehydrogenase to allosteric inhibition by GTP" PMID:11502802. Loss of GTP inhibition leaves leucine
  activation unopposed, producing exaggerated leucine-stimulated insulin release
  PMID:11502802.

Interactions

• KLHL22 (Q53GT1): cryo-EM structure of CUL3(KLHL22)-RBX1 bound to a GDH1 hexamer;
  "CULLIN3KLHL22-RBX1 ligase mediated the polyubiquitination of GDH1 in vitro"
  PMID:37788672.
  So KLHL22 is a CUL3 substrate-adapter that ubiquitinates GDH1. UniProt records this
  IntAct interaction (NbExp=10).
• GLUD2 (P49448): paralog; recorded IntAct interaction.
• High-throughput interactome IPIs (BioPlex: PMID:28514442, PMID:33961781; cell-map:
  PMID:40205054) contribute generic "protein binding" annotations.

Annotation review reasoning (summary)

• Core molecular function = GO:0004353 L-glutamate dehydrogenase [NAD(P)+] activity
  (dual cofactor). The cofactor-specific children GO:0004352 (NAD+) and GO:0004354 (NADP+)
  are experimentally supported (NAD+ EXP/IDA in multiple papers; NADP+ EXP PMID:11032875)
  and kept as ACCEPT / non-core specializations.
• Regulatory-ligand binding (ADP GO:0043531, GTP GO:0005525, L-leucine GO:0070728, NAD+
  GO:0070403) are genuine allosteric/cofactor-binding functions; ADP binding is core to the
  ADP-activated physiological state, others kept as accepted non-core.
• Localization: mitochondrion/mitochondrial matrix are core; ER and cytoplasm are minor real
  pools kept as non-core.
• Insulin-secretion and substantia-nigra/neuro terms are real physiology but downstream/
  tissue-specific → KEEP_AS_NON_CORE (or over-annotation for HEP proteomics).
• Bare "protein binding" (GO:0005515) IPIs → MARK_AS_OVER_ANNOTATED (uninformative; the
  meaningful partner KLHL22 is captured structurally but the GO term itself is not
  informative).
• No experimental annotation is REMOVEd. Wrong-branch / redundant IEA (oxidoreductase
  activity, generic amino acid metabolic process) → MODIFY or KEEP_AS_NON_CORE per policy.
  
  

2026-11-05 ORGANELLES ER batch 2 (endoplasmic reticulum rows)

Rows: GO:0005783 IEA (GO_REF:0000044, SL-0095) and IDA (PMID:19448744), both located_in; no ER sub-compartment rows; no is_active_in.

• UniProt: "Endoplasmic reticulum {ECO:0000269|PubMed:19448744}" with note "Mostly translocates into the mitochondria, only a small amount of the protein localizes to the endoplasmic reticulum." So the IEA and IDA rows carry one claim.
• PMID:19448744 (abstract only, no PMC): hGDH1-EGFP overexpressed in COS 7, HeLa, CHO, HEK293, SH-SY5Y; PMID:19448744; the ~95 kDa full-length (unprocessed) band was in the ER-containing cytosol fraction.
• Topology: FT TRANSIT 1..53 (mitochondrion); no ER signal peptide, TM or retention motif, so any ER pool is cytosolic-face; stays at root.
• Outcome: both rows stay KEEP_AS_NON_CORE (precedent AHCY, ER batch 1: overexpressed fusion against an ER marker); reasons added (rows previously had none) and UniProt + abstract quotes added. Close call with MARK_AS_OVER_ANNOTATED (tagged, overexpressed, unprocessed precursor), stated in the reasons.