AMDHD1 (amidohydrolase domain-containing protein 1) is the human imidazolonepropionase (EC 3.5.2.7), a cytosolic metal-dependent amidohydrolase that catalyses the third step of the universal L-histidine degradation pathway. It hydrolyses the carbon-nitrogen bond of 4-imidazolone-5-propanoate to yield N-formimidoyl-L-glutamate (N-formimino-L-glutamate), positioned between urocanate hydratase (UROC1) and glutamate formimidoyltransferase (FTCD) in the His-to-glutamate route. The protein belongs to the metallo-dependent hydrolase superfamily (HutI family; Amidohydro_1 fold) and binds one catalytic zinc or iron ion per subunit coordinated by conserved active-site residues. The human enzyme is expressed most highly in liver and skeletal muscle. Beyond its metabolic role, one report describes AMDHD1 acting as a tumor suppressor in cholangiocarcinoma by stabilizing SMAD4 and promoting TGF-beta/SMAD signaling, a secondary activity that is not yet independently established.
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0006548 L-histidine catabolic process | IBA GO_REF:0000033 | ACCEPT | Summary: Core biological process. Phylogenetically inferred (IBA) membership in the imidazolonepropionase clade correctly places AMDHD1 in L-histidine catabolism, the third step of which it catalyses. Consistent with UniProt PATHWAY (L-histidine degradation into L-glutamate, step 3/3) and Reactome. Reason: ACCEPT because AMDHD1 itself performs a step of the process rather than merely being required for it: it catalyses the conversion of 4-imidazolone-5-propanoate to N-formimidoyl-L-glutamate, the third of the four reactions that take histidine to glutamate. The IBD node behind this IBA sits at the base of the HutI/imidazolonepropionase clade, which is defined by exactly this activity, and the seed includes the characterized Pseudomonas aeruginosa imidazolonepropionase HutI (UniProtKB:Q9HU91 in the WITH/FROM, with PANTHER:PTN000505272); human AMDHD1 retains the full Amidohydro_1 fold and both metal-ligating positions, so there is no evidence of divergence or activity loss that would exclude it from the clade. Supporting Evidence: file:human/AMDHD1/AMDHD1-uniprot.txt glutamate; N-formimidoyl-L-glutamate from L-histidine: step 3/3. |
| GO:0050480 imidazolonepropionase activity | IBA GO_REF:0000033 | ACCEPT | Summary: Core molecular function. IBA to imidazolonepropionase activity matches the UniProt catalytic-activity/EC assignment (EC 3.5.2.7, 4-imidazolone-5-propanoate + H2O = N-formimidoyl-L-glutamate) and the HutI-family metallo-hydrolase fold. This is the defining catalytic activity of the gene product. Reason: ACCEPT: this is the gene's defining molecular function and the enables qualifier is correct, since AMDHD1 is a single-subunit amidohydrolase that performs the hydrolysis itself. The phylogenetic inference is the strongest evidence available for the human protein - UniProt still labels it "Probable imidazolonepropionase" and derives the EC assignment by similarity to UniProtKB:P42084, and Reactome states that the existence of the human enzyme is inferred from high-throughput screening rather than a direct assay. A short WITH/FROM list is not weak support here: the node is seeded by a well-characterized bacterial HutI, and the human protein retains the substrate-binding (159, 192, 263), product-binding (159, 336) and metal-ligating (260, 334) residues the mechanism requires. Direct assay of the human enzyme remains the main open item and is listed under suggested_experiments. Supporting Evidence: file:human/AMDHD1/AMDHD1-uniprot.txt Reaction=4-imidazolone-5-propanoate + H2O = N-formimidoyl-L-glutamate; |
| GO:0005737 cytoplasm | IEA GO_REF:0000120 | KEEP AS NON CORE | Summary: Correct but non-specific location. UniProt records experimental cytoplasmic localization (Cytoplasm, ECO:0000269|PubMed:39143229), and the soluble histidine-catabolic enzymes act in the cytosol. Retained as a non-core location; the more precise GO:0005829 cytosol (TAS Reactome) is the preferred term. Reason: KEEP_AS_NON_CORE rather than MODIFY, because unlike the generic catalytic-activity rows this one is not merely a pipeline artefact: it is the term UniProt's own SUBCELLULAR LOCATION line asserts, and that line is experimental (ECO:0000269|PubMed:39143229), so the row records a real observation at the granularity at which the observation was made. The cytosol child (GO:0005829, TAS Reactome) is more precise and is the one carried into core_functions, which is why this parent is non-core rather than core. Supporting Evidence: file:human/AMDHD1/AMDHD1-uniprot.txt SUBCELLULAR LOCATION: Cytoplasm {ECO:0000269|PubMed:39143229}. |
| GO:0006548 L-histidine catabolic process | IEA GO_REF:0000002 | ACCEPT | Summary: Same core BP as the IBA annotation, here from InterPro2GO (HutI signature IPR005920). Correct and consistent with the enzyme's role. Reason: ACCEPT: the InterPro2GO rule fires on IPR005920, the family-specific imidazolonepropionase signature rather than a superfamily-level match, so it arrives at precisely the right process term. Agreement with the IBA and the Reactome TAS from three independent routes is a reason to keep it, not to prune it as duplicative. Supporting Evidence: file:human/AMDHD1/AMDHD1-uniprot.txt glutamate; N-formimidoyl-L-glutamate from L-histidine: step 3/3. |
| GO:0016787 hydrolase activity | IEA GO_REF:0000002 | MODIFY | Summary: Correct but over-general. This is a high-level ancestor of the specific catalytic term GO:0050480 imidazolonepropionase activity, which is available and evidenced. Replace with the specific molecular function. Reason: MODIFY on granularity, not correctness. The rule fires on IPR006680 (Amidohydro_1), a Pfam-level domain shared across the whole metallo-dependent hydrolase superfamily - deaminases, dihydroorotases, ureases - which is why it can only emit the branch-root hydrolase term. GO:0016787 is an ancestor of the GO:0050480 annotation AMDHD1 already carries by IBA and IEA, so it adds nothing; the specific term is the right replacement. Proposed replacements: imidazolonepropionase activity Supporting Evidence: file:human/AMDHD1/AMDHD1-uniprot.txt Reaction=4-imidazolone-5-propanoate + H2O = N-formimidoyl-L-glutamate; |
| GO:0016810 hydrolase activity, acting on carbon-nitrogen (but not peptide) bonds | IEA GO_REF:0000002 | MODIFY | Summary: Correct but over-general ancestor of GO:0050480. AMDHD1 does hydrolyse a C-N bond, but the precise imidazolonepropionase activity term should be used. Reason: MODIFY on granularity. The chemistry asserted is right - UniProt's FUNCTION line describes hydrolytic cleavage of a carbon-nitrogen bond - but the InterPro signature behind the row (IPR011059, metal-dependent hydrolase composite domain) identifies the fold rather than the substrate, so the term stops two levels above GO:0050480. Replace with the specific activity, which is independently supported. Proposed replacements: imidazolonepropionase activity Supporting Evidence: file:human/AMDHD1/AMDHD1-uniprot.txt FUNCTION: Catalyzes the hydrolytic cleavage of the carbon-nitrogen bond |
| GO:0016812 hydrolase activity, acting on carbon-nitrogen (but not peptide) bonds, in cyclic amides | IEA GO_REF:0000120 | MODIFY | Summary: Accurate description of the reaction class (hydrolysis of the cyclic amide of 4-imidazolone-5-propanoate) but still a parent of the specific catalytic term. Replace with GO:0050480 imidazolonepropionase activity. Reason: MODIFY on granularity. This is the closest of the three generic hydrolase-activity rows - the imidazolone ring really is a cyclic amide - but it is still a direct ancestor of GO:0050480, which the same UniProt pipeline also emits from the RHEA/EC mapping. Keeping both the parent and the child from one pipeline adds no information, so the parent is replaced. Proposed replacements: imidazolonepropionase activity Supporting Evidence: file:human/AMDHD1/AMDHD1-uniprot.txt Reaction=4-imidazolone-5-propanoate + H2O = N-formimidoyl-L-glutamate; |
| GO:0050480 imidazolonepropionase activity | IEA GO_REF:0000120 | ACCEPT | Summary: Core molecular function, independently supported by the RHEA/EC mapping (RHEA:23660, EC 3.5.2.7). Redundant with the IBA annotation to the same term but correct and mutually reinforcing. Reason: ACCEPT: the RHEA:23660 / EC 3.5.2.7 mapping reaches the specific catalytic term directly from the reaction UniProt records for this entry, which is a different route to the same conclusion as the phylogenetic IBA. Both ultimately trace back to the characterized bacterial enzyme rather than to a human assay; that is a reason to flag the open experimental question, not to reject the annotation. Supporting Evidence: file:human/AMDHD1/AMDHD1-uniprot.txt ChEBI:CHEBI:77893; EC=3.5.2.7; |
| GO:0005515 protein binding | IPI PMID:28514442 Architecture of the human interactome defines protein commun... | REMOVE | Summary: Bare "protein binding" from a proteome-scale AP-MS interactome screen (BioPlex 2.0), supporting the AMDHD1-KLHL23 (UniProtKB:Q8NBE8) interaction. The paper is a systematic HEK293T HA-FLAG AP-MS study that does not characterise AMDHD1 individually or assign it a biological function. The term is uninformative and does not represent a core function. Reason: REMOVE per the generic-protein-binding policy: the defect is absence of functional information, not a claim exceeding the evidence, so MARK_AS_OVER_ANNOTATED is explicitly excluded for GO:0005515. Removal does not assert the interaction is false - AMDHD1-KLHL23 is reproduced in BioPlex 3.0 and UniProt records it with NbExp=3. MODIFY is unavailable because no evidence-backed, more informative molecular function exists: KLHL23 is a Kelch/BTB adaptor, but neither screen tests whether AMDHD1 is a CUL3 substrate, so proposing a substrate-recognition or ubiquitin-pathway term would exceed the evidence. Note that the one study assigning AMDHD1 a protein-stabilising role names SMAD2/3/4, not KLHL23 (PubMed:39143229, not cached here), so it does not license a replacement term for this row either. Supporting Evidence: PMID:28514442 Here we present BioPlex 2.0 (Biophysical Interactions of ORFeome-derived complexes), which uses robust affinity purification-mass spectrometry methodology to elucidate protein interaction networks file:human/AMDHD1/AMDHD1-uniprot.txt Q96NU7; Q8NBE8: KLHL23; NbExp=3; IntAct=EBI-21895891, EBI-2512246; |
| GO:0005515 protein binding | IPI PMID:33961781 Dual proteome-scale networks reveal cell-specific remodeling... | REMOVE | Summary: Same uninformative "protein binding", from BioPlex 3.0 (Huttlin 2021), again capturing the AMDHD1-KLHL23 (UniProtKB:Q8NBE8) AP-MS interaction in 293T/HCT116 cells. High-throughput, with no defined biological meaning for AMDHD1's metabolic role. Reason: REMOVE for the same reason as the BioPlex 2.0 row: a bare GO:0005515 annotation carries no functional information, which the policy treats as grounds for removal rather than for MARK_AS_OVER_ANNOTATED. This screen is the independent replication that makes the interaction itself credible, which is precisely why removal is a statement about the term and not about the measurement. No replacement term is proposed, because the study assigns AMDHD1 no function. Supporting Evidence: PMID:33961781 Through affinity-purification mass spectrometry, we have created two proteome-scale, cell-line-specific interaction networks. file:human/AMDHD1/AMDHD1-uniprot.txt Q96NU7; Q8NBE8: KLHL23; NbExp=3; IntAct=EBI-21895891, EBI-2512246; |
| GO:0006548 L-histidine catabolic process | TAS Reactome:R-HSA-70921 | ACCEPT | Summary: Core biological process, author-asserted via Reactome's "Histidine catabolism" pathway, in which AMDHD1 catalyses step 3 (4-imidazolone-5-propanoate to N-formimino-L-glutamate). Correct. Reason: ACCEPT: Reactome's pathway lays out the four steps from histidine to glutamate and places AMDHD1 at the third, so the gene product performs part of the work of the process and involved_in is literally true. The flanking steps in the same pathway (HAL, UROC1, FTCD) are annotated to the same process, and the upstream end of the route has independent human genetic support from histidinemia. Supporting Evidence: Reactome:R-HSA-70921 The major pathway of histidine catabolism, annotated here, proceeds in four steps to yield glutamate |
| GO:0016812 hydrolase activity, acting on carbon-nitrogen (but not peptide) bonds, in cyclic amides | TAS Reactome:R-HSA-70906 | MODIFY | Summary: Correct reaction class from Reactome (the 4-imidazolone-5-propanoate + H2O reaction) but over-general relative to the specific catalytic term. Replace with GO:0050480 imidazolonepropionase activity. Reason: MODIFY on granularity: Reactome's own reaction description names the substrate and product explicitly, so the specific term GO:0050480 is what the cited statement actually supports, and the cyclic-amide parent understates it. The same Reactome entry is candid that the human enzyme's existence is inferred from high-throughput screening rather than direct assay - that affects how strongly any activity term is held, not which term fits best. Proposed replacements: imidazolonepropionase activity Supporting Evidence: Reactome:R-HSA-70906 Cytosolic imidazolonepropionase catalyzes the hydrolysis of 4-imidazolone-5-propanoate to form N-Formimino-L-glutamate. |
| GO:0005829 cytosol | TAS Reactome:R-HSA-70906 | ACCEPT | Summary: Correct and appropriately precise location. Reactome places the imidazolonepropionase reaction in the cytosol, consistent with the soluble histidine-catabolic enzymes and with UniProt's experimental cytoplasmic localization. This is the preferred location term (over the parent GO:0005737 cytoplasm). Reason: ACCEPT: the cytosolic assignment agrees with the protein's own features - Q96NU7 has no signal peptide, transmembrane segment or organellar targeting sequence - and with the compartment of the flanking pathway enzymes, all of which are soluble cytosolic proteins. It is also consistent at a coarser grain with UniProt's experimental Cytoplasm annotation. This is the most specific supported location and is the one used in core_functions. Supporting Evidence: Reactome:R-HSA-70906 Cytosolic imidazolonepropionase catalyzes the hydrolysis of 4-imidazolone-5-propanoate to form N-Formimino-L-glutamate. |
| GO:0003674 molecular_function | ND GO_REF:0000015 | ACCEPT | Summary: Root molecular function ND annotation: when it was made, no molecular function had been curated for this gene product from literature. Reason: ACCEPT under the ND acceptance rule (projects/NO_DATA.md). The evidence this review cites for the molecular function is inferred (from orthologs, family membership or sequence features) or comes from other species; none of it is literature on this gene product in this species. This was judged from the review's own content during the NO_DATA re-action pass, not from a new literature search. Earlier action: REMOVE. Possibly stale: the 2026-11-06 ND citation pass found literature on this gene in this species that may identify a molecular function; see AMDHD1-notes.md, section "ND literature-search citations (2026-11-06)". A curator should check it before this ACCEPT stands. |
| GO:0046872 metal ion binding | ISS file:human/AMDHD1/AMDHD1-uniprot.txt | NEW | Summary: Proposed new annotation. AMDHD1 is a metallo-dependent hydrolase that binds one catalytic divalent/trivalent metal ion per subunit - zinc or ferric iron - ligated by residues 260 and 334 in the HutI active site. GOA currently carries no metal-binding term for AMDHD1, although the orthologous bacterial imidazolonepropionase does. Reason: NEW because the binding is done by AMDHD1 itself: the metal is held by the protein's own side chains (BINDING 260 and BINDING 334, each annotated for both Zn(2+) and Fe(3+)) and is the Lewis acid that activates the water molecule for C-N bond hydrolysis, so this is the cofactor case rather than a necessity-only relationship. The comparator check supports rather than refutes the proposal: the Pseudomonas putida HutI ortholog carries both GO:0008270 zinc ion binding and GO:0005506 iron ion binding (IEA, GO_REF:0000104), and human ACMSD, in the same metallo-dependent hydrolase superfamily, carries GO:0008270 by IDA from two papers - so metal binding is the convention for this family, and its absence on AMDHD1 is a coverage gap in the UniProt cofactor pipeline for this entry rather than a deliberate curatorial omission. The parent term GO:0046872 is proposed rather than GO:0008270 zinc ion binding or GO:0005506 iron ion binding, because UniProt asserts the two metals as an explicit either/or at equal evidential strength: Zn(2+) and Fe(3+) are both COFACTOR entries carrying ECO:0000250|UniProtKB:P42084, the Note reads "Binds 1 zinc or iron ion per subunit", and residues 260 and 334 each have a BINDING line for either metal (Zn from P42084, Fe from A0KF84). COFACTOR order is not a ranking, so neither specific term can be asserted without narrowing a claim the source deliberately leaves open. Evidence is ISS, because the cofactor and both ligating residues are assigned in UniProt by similarity to characterized bacterial enzymes; the human protein has not been purified and its metal content not measured, which is why the metal identity is raised under suggested_questions and why the specific term should be substituted once the human enzyme is assayed. Supporting Evidence: file:human/AMDHD1/AMDHD1-uniprot.txt Name=Zn(2+); Xref=ChEBI:CHEBI:29105; file:human/AMDHD1/AMDHD1-uniprot.txt Name=Fe(3+); Xref=ChEBI:CHEBI:29034; file:human/AMDHD1/AMDHD1-uniprot.txt Note=Binds 1 zinc or iron ion per subunit. file:human/AMDHD1/AMDHD1-uniprot.txt Belongs to the metallo-dependent hydrolases superfamily. |
Loading supporting content…
Download this section (compressed HTML)Q: UniProt lists both Zn(2+) and Fe(3+) as cofactors, ligated by the same two residues (260 and 334), with the assignment transferred from bacterial orthologs. Which metal does the human enzyme actually use in liver and skeletal muscle, and does the identity of the metal change kcat or substrate specificity?
Q: The human enzyme has never been purified or assayed; its activity and EC assignment are inherited from UniProtKB:P42084, and Reactome states that the existence of the human enzyme is inferred only from high-throughput screening. What are the steady-state kinetic parameters of recombinant human AMDHD1, and do they match the bacterial enzyme closely enough to justify the "probable" qualifier being dropped?
Q: One report describes AMDHD1 as a cholangiocarcinoma tumor suppressor that stabilizes SMAD4 and enhances TGF-beta signaling. Is this activity separable from catalysis - does a catalytically dead active-site mutant still stabilize SMAD4 - and is the effect reproducible outside biliary tract tumors?
Q: The only reproducible interactor from proteome-scale screens is the Kelch/BTB adaptor KLHL23 (NbExp=3). Is AMDHD1 a CUL3-KLHL23 ubiquitylation substrate, and if so does that pathway set histidine-catabolic flux by controlling AMDHD1 abundance?
Experiment: Express and purify recombinant human AMDHD1, then measure 4-imidazolone-5-propanoate hydrolysis (coupled or direct assay for N-formimidoyl-L-glutamate formation) to obtain Km and kcat. In parallel, determine the bound metal and its stoichiometry by ICP-MS on as-purified protein, and repeat the kinetics after metal removal and reconstitution with Zn(2+) or Fe(3+) separately. Assay H260A and H334A variants alongside, checking fold integrity by circular dichroism and thermal shift so that loss of catalysis can be distinguished from loss of structure. This would convert the activity from a by-similarity inference into a direct human measurement and settle the Zn-versus-Fe question.
Hypothesis: Recombinant human AMDHD1 is a genuine imidazolonepropionase whose turnover depends on an active-site divalent metal held by His-260 and His-334, and substituting either residue abolishes activity without preventing folding.
Type: recombinant enzyme kinetics with ICP-MS metal quantification
Experiment: Knock out AMDHD1 in a hepatocyte line (or primary human hepatocytes) and follow the fate of uniformly labelled [13C6]-L-histidine by LC-MS, quantifying urocanate, 4-imidazolone-5-propanoate, N-formimino-L-glutamate and glutamate. Accumulation of the AMDHD1 substrate with loss of downstream label would establish the step in human cells; rescue with wild-type but not active-site mutant AMDHD1 would tie the flux to catalysis rather than to the protein's presence.
Hypothesis: AMDHD1 is required for histidine-catabolic flux in human hepatocytes, and the pathway stalls at 4-imidazolone-5-propanoate when it is lost.
Type: stable-isotope metabolic flux analysis in a knockout line
Experiment: Compare wild-type AMDHD1 and a catalytically inactive active-site variant for their ability to co-immunoprecipitate SMAD2/3/4, to slow SMAD4 ubiquitination and proteasomal turnover in cycloheximide-chase and ubiquitination assays, and to restore TGF-beta-dependent reporter activity and cell-cycle arrest in AMDHD1-depleted cells. A catalysis-independent effect would support annotating a second, moonlighting molecular function; a catalysis-dependent one would suggest the phenotype is metabolic and should not be annotated as protein stabilisation.
Hypothesis: The SMAD4-stabilising, TGF-beta-promoting activity reported in cholangiocarcinoma is independent of imidazolonepropionase catalysis.
Type: structure-function separation of moonlighting activity
Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)