GLOD4 (Q9HC38, HGNC:14111; formerly C17orf25) review notes

Why this gene was selected

GLOD4 is a glyoxalase-domain orphan: UniProt places it in the glyoxalase I family by sequence
similarity (ECO:0000305) but assigns no EC number and no catalytic residues, and until 2026 its
entire GOA molecular-function record was GO:0045296 cadherin binding from a single
proximity-labelling experiment. A 2026 PNAS paper assigns it a first real enzymatic activity,
and an unusual one.

Reference verification

PMID:41628334 verified against PubMed: Wright S, Dang VC, Hussain S, et al. (22 authors),
"Selective peroxynitrite-mediated protein nitration catalyzed by glyoxalase domain containing
protein 4." Proc Natl Acad Sci USA 2026 Feb 2;123(6):e2515002123. doi:10.1073/pnas.2515002123.
PMC12890929. Full text cached.

PMID:25468996 verified: Guo Z, Neilson LJ, Zhong H, et al., "E-cadherin interactome complexity
and robustness resolved by quantitative proteomics." Sci Signal 2014;7(354):rs7.
PMID:34800366 verified: Morgenstern M et al., Cell Metab 2021;33(12):2464-2483 (MitoCoP).
PMID:23533145 verified: In-depth proteomic analyses of exosomes from expressed prostatic
secretions in urine, Proteomics 2013.

The claim, and how strong it is

Headline:
PMID:41628334
PMID:41628334

What is actually shown, layer by layer:

Biochemistry. Recombinant and RBC-purified GLOD4 nitrates an α-synuclein peptide and
full-length α-syn in the presence of peroxynitrite; MS localises the modification
PMID:41628334. Kinetics follow Michaelis-Menten
PMID:41628334 and stopped-flow shows GLOD4
consumes the cosubstrate PMID:41628334.

Cosubstrate specificity (a real strength). The paper rules out the obvious alternatives:
PMID:41628334 and PMID:41628334

Active site. PMID:41628334 and
PMID:41628334
This is the single most important control: a structure-guided catalytic-residue mutant that
loses activity is what separates an enzyme from a protein that merely accelerates a chemical
reaction non-specifically.

Substrate selectivity. A 23,000-protein HuProt array gave a very short hit list
PMID:41628334, and
γ-synuclein, which lacks the C-terminal tyrosines, was not nitrated. Selectivity is the core
argument of the paper, since spontaneous peroxynitrite chemistry is not selective.

Cells and animals. PMID:41628334 and in mice
PMID:41628334, with the gene-dose dependence being the
persuasive part.

Localization. PMID:41628334 - which qualifies UniProt's mitochondrion-only
subcellular location, itself inferred (ECO:0000305) from a 2003 interaction paper.

Limitations to state honestly

1. One paper, one laboratory, commercial interest. Almost all authors are at Nitrase
   Therapeutics, and the paper itself declares
   PMID:41628334 That is not a
   reason to disbelieve the data, but it is a reason to record the claim as single-sourced and
   awaiting independent replication.
2. The category is unusual. "Enzymatically catalysed protein tyrosine nitration" is a new
   enzyme class; the authors are explicit that it is a proposal
   PMID:41628334 and hedge again in the discussion
   PMID:41628334
3. Mechanism is not established. The paper states that work to dissect the catalytic
   mechanism is ongoing and that it is not yet known whether zinc is the catalytic metal.
4. No organismal phenotype. PMID:41628334 So the physiological importance of the activity is undetermined, even though the
   biochemical contribution to α-syn nitration is large.
5. Two of the four substrates are in vitro only (PPM1B, PQBP1), as the authors note.
6. The glyoxalase assignment is untouched. No glyoxalase/lactoylglutathione lyase activity
   has ever been demonstrated for GLOD4, and the paper reframes the family relationship
   structurally rather than functionally
   PMID:41628334. The "glyoxalase domain" in the gene name remains a fold description, not an
   activity, and no EC number is assigned.

The GO term problem

There is no GO term for protein nitration at all - not as a molecular function and not as a
biological process. Searching QuickGO for "nitration" returns nothing; the nearest existing
terms are wrong chemistry or wrong direction:

• GO:0017014 protein nitrosylation - addition of a nitric oxide group (S-nitrosylation of
  cysteine). Different atom count, different residue, different chemistry. The paper
  specifically excludes NO and NO-derived oxidants as GLOD4 cosubstrates.
• GO:0072541 peroxynitrite reductase activity and GO:0062213 peroxynitrite isomerase activity - these consume peroxynitrite but yield nitrite or nitrate, i.e. they are
  detoxification activities. GLOD4 also accelerates peroxynitrite decomposition, so there is a
  superficial resemblance, but the product is a nitrated protein, not a free anion. Using either
  would assert a detoxifying role, the opposite of what is described.
• GO:0018212 peptidyl-tyrosine modification is obsolete.

So the honest options were a wrong term or no term. Two proposed_new_terms are filed instead
(one MF, one BP), and the NEW annotations placed in the file are deliberately non-committal:
GO:0140096 catalytic activity, acting on a protein as an explicitly flagged placeholder that
is true but uninformative, plus GO:0008270 zinc ion binding and GO:0005829 cytosol, which
are specific and independently supported.

Curation position taken

• GO:0005739 mitochondrion (IEA and HTP) -> ACCEPT for both. GLOD4 is in the MitoCoP
  high-confidence mitochondrial proteome and the 2026 paper's own fractionation finds it in
  mitochondria, though with more in cytosol.
• GO:0045296 cadherin binding (HDA) -> MARK_AS_OVER_ANNOTATED. The source is BioID
  proximity labelling of the E-cadherin cytoplasmic tail, which reported
  PMID:25468996 Proximity within a labelling
  radius is not binding; a 561-protein list from an abundant-cytosolic-protein-rich compartment
  is the classic source of spurious cadherin binding annotations, and nothing else connects
  GLOD4 to adherens junctions.
• GO:0070062 extracellular exosome (HDA) -> KEEP_AS_NON_CORE. GLOD4 was genuinely detected
  in prostatic-secretion exosomes, but abundant cytosolic proteins are routine in exosome
  proteomes and this is not where the enzyme acts.
• NEW GO:0140096 catalytic activity, acting on a protein (IDA) - placeholder for the
  nitrase activity, flagged as such.
• NEW GO:0008270 zinc ion binding (IDA) - ICP-MS on purified protein plus loss of activity
  in the H8A/E70A metal-site mutants.
• NEW GO:0005829 cytosol (IDA) - subcellular fractionation of H4 cells.

Open questions

• Is the activity reproducible outside Nitrase Therapeutics? This is the first thing an
  independent laboratory should check, ideally with the recombinant protein and the published
  active-site mutants.
• What is the catalytic mechanism, and is Zn2+ the catalytic metal or a structural one? The
  authors leave this open and list Mn, Ni and Fe as alternatives.
• Does GLOD4 have any glyoxalase-family activity? The fold says VOC superfamily; no metabolic
  substrate has ever been tested systematically.
• Why is a nitrase in mitochondria as well as cytosol, and are there mitochondrial substrates?
  The α-syn work is all cytosolic.
• GLOD4-KO mice are healthy to 18 months. What, then, is the physiological function of selective
  α-syn nitration, and does removing it change synuclein pathology rather than baseline biology?