Nudt2 (Apah1) encodes the mouse asymmetrical diadenosine tetraphosphate (Ap4A) hydrolase, a small (147 aa) soluble Nudix hydrolase of the PTHR21340:SF0 subfamily that is the orthologue of human NUDT2 (89% identical) and C. elegans ndx-4. By homology to the biochemically characterised pig, human and nematode enzymes it cleaves Ap4A (and related dinucleoside polyphosphates such as Gp4G) asymmetrically to a nucleoside triphosphate plus a nucleoside monophosphate (AMP + ATP for Ap4A), using the Nudix-box glutamates and divalent metal ions. Its main physiological role is to set intracellular Ap4A levels: disruption of the human gene raises Ap4A about 175-fold, and in immunologically activated mast cells the mammalian enzyme degrades the Ap4A that releases MITF and USF2 from HINT1 inhibition, terminating Ap4A second-messenger signalling. The same active site also removes 5' caps from RNA: recombinant mouse Nudt2 decaps m7G- and unmethylated G-capped RNA (releasing m7GMP and m7GDP), dpCoA- and FAD-capped RNA in vitro, and human NUDT2 cleaves Ap4A-capped RNA. In human patient cells loss of NUDT2 decapping alters mRNA half-life, and biallelic NUDT2 loss causes a recessive neurodevelopmental disorder. The enzyme also removes the 5'-triphosphate of uncapped RNAs one phosphate at a time, leaving a 5'-monophosphate end that the exonuclease XRN1 can degrade; Nudt2-knockout mice are viable and breed normally, but their bone-marrow-derived cells support 10- to 30-fold greater growth of viruses that generate 5'-triphosphate RNA. Mouse-specific data on localization are limited to proteomic detection, including in mitochondrial fractions; human NUDT2 is nuclear and cytoplasmic and the Drosophila orthologue is predominantly nuclear.
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0004081 bis(5'-nucleosyl)-tetraphosphatase (asymmetrical) activity | IBA GO_REF:0000033 | ACCEPT | Summary: Phylogenetic inference of asymmetrical Ap4A hydrolase activity from PANTHER node PTN000482012 (PTHR21340:SF0), grounded in experimental data for pig NUDT2 (P50584), Plasmodium falciparum C0H4F3 and C. elegans ndx-4. Reason: Mouse Nudt2 sits within the SF0 Ap4A hydrolase subfamily, retains the Nudix box (residues 43-64) and the Tetra_PHTase signature (IPR003565), and there is no evidence of divergence or loss of catalytic residues. The node placement is sound and the activity is the defining function of this subfamily. Supporting Evidence: PMID:11738085 It hydrolyses Ap4A with a K(m) of 7 microM and k(cat) of 27 s(-1) producing AMP and ATP as products. PMID:18644867 we provided here evidence that the "Nudix" type 2 gene product, Ap(4)A hydrolase, is responsible for Ap(4)A degradation following the immunological activation of mast cells. |
| GO:0004081 bis(5'-nucleosyl)-tetraphosphatase (asymmetrical) activity | IEA GO_REF:0000120 | ACCEPT | Summary: Automated mapping from the UniProt catalytic activity (EC 3.6.1.17, RHEA:22484) to the asymmetrical bis(5'-nucleosyl)-tetraphosphatase activity term. Reason: The EC assignment is correct for this subfamily and the GO term is the precise core activity; redundant with the IBA/ISS rows but correct. Supporting Evidence: PMID:11738085 It hydrolyses Ap4A with a K(m) of 7 microM and k(cat) of 27 s(-1) producing AMP and ATP as products. |
| GO:0004081 bis(5'-nucleosyl)-tetraphosphatase (asymmetrical) activity | ISS GO_REF:0000024 | ACCEPT | Summary: Curator transfer of Ap4A hydrolase activity from the pig orthologue (P50584). Reason: Pig and mouse Ap4A hydrolases are close mammalian orthologues; transfer of the core catalytic activity is well justified. Mouse Nudt2 itself has not been assayed on Ap4A, but the catalytically inactive EE-QQ Nudix-box mutant of mouse Nudt2 lost cap-hydrolysing activity, showing the Nudix active site is functional. Supporting Evidence: PMID:32432673 Highly conserved glutamic acid residues essential for catalysis within the Nudix motif were substituted with lysine in Nudt16 (EE-KK), glutamines in Nudt2 and Nudt3 (EE-QQ). |
| GO:0004081 bis(5'-nucleosyl)-tetraphosphatase (asymmetrical) activity | ISS PMID:11738085 Cloning, characterisation and crystallisation of a diadenosi... | ACCEPT | Summary: Curator transfer of Ap4A hydrolase activity from the C. elegans orthologue ndx-4 (Q9U2M7), whose recombinant protein was kinetically characterised. Reason: Same subfamily (PTHR21340:SF0) and same core activity; ndx-4 behaves as a typical animal Ap4A hydrolase, so the transfer to mouse is sound, though more distant than the pig transfer. Supporting Evidence: PMID:11738085 The purified recombinant enzyme behaves as a typical animal Ap4A hydrolase. |
| GO:0005739 mitochondrion | HDA PMID:18614015 A mitochondrial protein compendium elucidates complex I dise... | KEEP AS NON CORE | Summary: High-throughput mass-spectrometry detection of Nudt2 in mitochondrial preparations from mouse tissues (MitoCarta compendium study). Reason: Experimental proteomic evidence that MGI curators annotated from the full dataset; Nudt2 is not discussed in the paper text, so the strength of its mitochondrial assignment cannot be checked here. Ap4A hydrolase activity is classically cytosolic/nuclear in animal cells, and direct localization data for human NUDT2 show nucleus and cytoplasm. The human mitochondrial matrix annotation is a Reactome pathway-placement artefact (removed in the human review), not independent support. A minor mitochondrial pool is possible but its functional relevance is unknown; the MGI HDA alone justifies retaining this as non-core rather than removing it. Supporting Evidence: PMID:18614015 we performed mass spectrometry, GFP tagging, and machine learning to create a mitochondrial compendium of 1098 genes and their protein expression across 14 mouse tissues. file:mouse/Nudt2/Nudt2-deep-research-openscientist.md a definitive mouse localization (nuclear vs. nucleocytoplasmic, chromatin association) has not been directly measured. |
| GO:0006167 AMP biosynthetic process | IBA GO_REF:0000033 | MARK AS OVER ANNOTATED | Summary: Process term propagated from PTN000482012 because AMP is a product of Ap4A hydrolysis; the node IBD is seeded only by C. elegans ndx-4 IDA rows. Reason: AMP is released as a by-product of degrading Ap4A, not as the purpose of a biosynthetic pathway; cellular AMP pools come from de novo and salvage purine synthesis and ATP turnover. The enzyme's process is diadenosine tetraphosphate catabolism (GO:0015967). Consistent with the ndx-4 review, where the seeding IDA rows were themselves marked over-annotated. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: ROLE CONFLATION Sources checked: PANTHER:PTN000482012 Node-level process term grounded in the reaction product of Ap4A hydrolysis; the catabolic process term is the appropriate one. WB:WBGene00003581 · ndx-4 Sole seed; its IDA rows are product-based process annotations, marked over-annotated in genes/worm/ndx-4. Supporting Evidence: PMID:11738085 It hydrolyses Ap4A with a K(m) of 7 microM and k(cat) of 27 s(-1) producing AMP and ATP as products. |
| GO:0006754 ATP biosynthetic process | IBA GO_REF:0000033 | MARK AS OVER ANNOTATED | Summary: Process term propagated from PTN000482012 because ATP is a product of Ap4A hydrolysis; the node IBD is seeded only by C. elegans ndx-4 IDA rows. Reason: ATP generated from micromolar Ap4A is negligible relative to oxidative and substrate-level phosphorylation; the enzyme removes Ap4A rather than making ATP. GO:0015967 diadenosine tetraphosphate catabolic process is the appropriate process term. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: ROLE CONFLATION Sources checked: PANTHER:PTN000482012 Node-level process term grounded in the reaction product of Ap4A hydrolysis; the catabolic process term is the appropriate one. WB:WBGene00003581 · ndx-4 Sole seed; its IDA rows are product-based process annotations, marked over-annotated in genes/worm/ndx-4. Supporting Evidence: PMID:11738085 always generating ATP as one of the products |
| GO:0006915 apoptotic process | ISS PMID:11937063 The crystal structure of diadenosine tetraphosphate hydrolas... | REMOVE | Summary: Sequence-similarity transfer from C. elegans ndx-4, whose own apoptosis annotation is a TAS based on a speculative sentence in a crystal-structure paper. Reason: The source statement is hedged ("potentially") background in a structural study, and the worm TAS source has been removed in genes/worm/ndx-4. No experiment links Nudt2 or any Ap4A hydrolase directly to execution of apoptosis; the hypothesis rests on Ap3A/Ap4A ratio models. A transfer of an unsupported statement adds no evidence, so this ISS should be removed. Supporting Evidence: PMID:11937063 Ap(4)A hydrolase has a key role in regulating the intracellular Ap(4)A levels and hence potentially the cellular response to metabolic stress and/or differentiation and apoptosis via the Ap(3)A/Ap(4)A ratio. |
| GO:0008796 bis(5'-nucleosyl)-tetraphosphatase activity | IEA GO_REF:0000002 | ACCEPT | Summary: InterPro Tetra_PHTase (IPR003565) mapping to the parent bis(5'-nucleosyl)-tetraphosphatase activity. Reason: Correct, if less specific than the asymmetrical activity (GO:0004081); an acceptable broader IEA. Supporting Evidence: PMID:11738085 As expected, sequence analysis shows the enzyme to be a member of the Nudix hydrolase family. |
| GO:0016787 hydrolase activity | IEA GO_REF:0000120 | MODIFY | Summary: Generic hydrolase activity from combined automated methods (ARBA rule, Nudix box signature). Reason: Correct but uninformative; the specific activity is well supported by orthology. Proposed replacements: bis(5'-nucleosyl)-tetraphosphatase (asymmetrical) activity Supporting Evidence: PMID:11738085 It hydrolyses Ap4A with a K(m) of 7 microM and k(cat) of 27 s(-1) producing AMP and ATP as products. |
| GO:0015967 diadenosine tetraphosphate catabolic process | ISS PMID:18644867 Diadenosine tetraphosphate hydrolase is part of the transcri... | NEW | Summary: Proposed process for the enzyme: degradation of Ap4A. The mammalian Nudt2 gene product is responsible for Ap4A degradation in activated mast cells, and its knockdown alters Ap4A accumulation; the worm orthologue carries this term by IDA. Reason: Nudt2 itself performs the hydrolytic step that destroys Ap4A, so it does the work of this process (not a substrate or requirement relationship). The comparator Ap4A hydrolase ndx-4 carries GO:0015967 by IDA. This replaces the product-based AMP/ATP biosynthetic process terms as the correct process annotation. Supporting Evidence: PMID:27144453 KBM-7 cells in which the NUDT2 Ap4A hydrolase gene had been disrupted (NuKO cells), causing a 175-fold increase in intracellular Ap4A PMID:18644867 The knockdown of Ap(4)A hydrolase modulated Ap(4)A accumulation, resulting in changes in the expression of MITF and USF2 target genes. PMID:11738085 The purified recombinant enzyme behaves as a typical animal Ap4A hydrolase. |
| GO:0110154 RNA decapping | IDA PMID:32432673 Mammalian Nudix proteins cleave nucleotide metabolite caps o... | NEW | Summary: Recombinant mouse Nudt2 removes m7G, unmethylated G, dpCoA and FAD caps from RNA in vitro, and the EE-QQ Nudix-box mutant loses the activity. Proposed to align the mouse review with the human NUDT2 review, which carries the same process term. Reason: Nudt2 itself catalyses the cap cleavage, so it performs the step rather than being a substrate or a requirement. The mouse protein has been assayed directly in two studies (PMID:23353937, PMID:32432673). The general term is used because the cap type that matters in cells is not known: the m7G products are mixed (m7GMP and m7GDP), so neither m7G-cap decapping child term fits, and Nudt2 loss did not change FAD-capped RNA levels in cells. Comparator check: the human orthologue NUDT2 carries GO:0110154 in its review on genetic (patient add-back) evidence, and the decapping Nudix enzymes assayed in the same mouse studies (e.g. Nudt16) are the usual carriers of decapping process terms, so this is a role the enzyme performs, not an inference from necessity. Non-core relative to Ap4A hydrolysis in mouse because the in vivo substrate is unresolved. Supporting Evidence: PMID:32432673 Additionally, we identify seven Nudix proteins (Nudt2, Nudt7, Nudt8, Nudt15, Nudt16 and Nudt19) that possess dpCoA cap decapping (deCoAping) activity in vitro, and two Nudix proteins (Nudt2 and Nudt16) possessing FAD cap decapping (deFADding) activity. PMID:23353937 Here we demonstrate that six additional mouse Nudix proteins--Nudt2, Nudt3, Nudt12, Nudt15, Nudt17, and Nudt19--have varying degrees of decapping activity in vitro on both monomethylated and unmethylated capped RNAs. |
| GO:0140818 mRNA 5'-triphosphate monophosphatase activity | ISS PMID:34824277 NUDT2 initiates viral RNA degradation by removal of 5'-phosp... | NEW | Summary: Recombinant NUDT2 releases single phosphates sequentially from 5'-triphosphorylated RNA, and the E58A Nudix-box mutant is inactive; in the same study Nudt2-knockout mouse cells support increased growth of viruses that generate 5'-PPP RNA. Proposed for consistency with the human NUDT2 review, which carries this term by IDA. Reason: The enzymology in PMID:34824277 was done with recombinant NUDT2, not stated in the text to be the mouse protein, so the activity is transferred by similarity (mouse and human NUDT2 are 89% identical, with the Nudix box conserved). What makes the transfer more than sequence-based is that the same paper made Nudt2-deficient mice and found the predicted loss-of-function phenotype in their cells, specific to viruses that produce 5'-triphosphate RNA. GO:0140818 describes the first step (PPP-RNA to PP-RNA plus phosphate). The term is restricted to mRNA, while the substrates shown are viral and in vitro transcribed RNAs; the human review raises the same scope question and it is repeated here in suggested_questions. Supporting Evidence: PMID:34824277 Notably, NUDT2 sequentially released single phosphates resulting in a di-phosphorylated dinucleotide intermediate that is further converted into a mono-phosphorylated product PMID:34824277 We concluded from these studies that mouse NUDT2 impairs the growth of PPP-RNA generating viruses in vitro, but it does not affect the accumulation of other viruses. |
Loading supporting content…
Download this section (compressed HTML)Q: Nudt2-knockout bone-marrow cells support increased growth of 5'-PPP-RNA viruses (PMID:34824277), and Nudt2 does the chemical step that starts degradation of the viral RNA. Should Nudt2 carry a virus-defence process term (e.g. GO:0051607 defense response to virus), and does the knockout show the phenotype in vivo?
Q: GO:0140818 is restricted to mRNA, but the 5'-PPP substrates shown for NUDT2 are viral and in vitro transcribed RNAs. Should the term be broadened, or a sibling created?
Q: Does Nudt2 loss in mice raise tissue Ap4A levels and alter MITF/USF2-dependent transcription in mast cells or cardiomyocytes in vivo?
Q: Is a fraction of Nudt2 genuinely resident in mitochondria (as proteomic datasets suggest), and if so does it control a mitochondrial Ap4A pool?
Q: Which RNA cap substrates (m7G, Ap4A, dpCoA, FAD) does Nudt2 decap in vivo, and should its RNA decapping be captured by a GO term? Nudt2 releases both m7GMP and m7GDP from m7G-capped RNA in vitro, so neither GO:0140932 nor GO:0140933 fits cleanly, and GO has no Ap4A-cap, FAD-cap or dpCoA-cap decapping terms.
Q: Do the in vitro dpCoA-cap and FAD-cap decapping activities of Nudt2 have any physiological substrate in mammalian cells, given that Nudt2 loss did not change FAD-capped RNA levels?
Experiment: Characterise the existing Nudt2 knockout (PMID:34824277) or a patient-variant knock-in mouse for brain Ap4A levels, mRNA half-lives of host-response/interferon genes and neurodevelopmental phenotypes, using decapping-defective versus Ap4A-hydrolysis-defective rescue alleles to separate the two activities.
Experiment: Measure Ap4A/Ap3A by LC-MS in Nudt2 knockout mouse tissues and bone-marrow-derived mast cells at baseline and after FcepsilonRI activation, together with RNA-seq of MITF/USF2 targets.
Experiment: Determine Nudt2 subcellular distribution (cytosol, nucleus, mitochondrial matrix) in mouse tissues by fractionation with protease protection and by an endogenous epitope-tagged allele.
Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)