ADPRH (ARH1) is a monomeric, magnesium-dependent 39-kDa hydrolase that erases arginine mono-ADP-ribosylation. It cleaves the N-glycosidic bond of N(omega)-(ADP-D-ribosyl)-L-arginyl residues in proteins, releasing free ADP-D-ribose and regenerating unmodified arginine (EC 3.2.2.19). It is therefore the reversing half of the arginine ADP-ribosylation cycle whose forward reaction is catalysed by the cholera-toxin-like transferase ARTC1 and, in disease, by cholera toxin itself. Catalysis requires two active-site Mg2+ ions coordinated by a vicinal acidic pair and flanking serines; substituting any of Ser54, Asp55, Asp56, Asp302 or Ser305 abolishes activity. Documented arginine-ADP-ribosylated substrates include the heterotrimeric G protein alpha subunit Gs-alpha, whose cholera-toxin-driven modification ADPRH reverses, and the muscle membrane-repair protein TRIM72/MG53. The enzyme also hydrolyses the alpha anomer of NAD+, but not beta-NAD+, linking it to the turnover of a minor NAD+ stereoisomer. Its specificity is narrow within its own family: the paralog ARH3/ADPRS acts on serine-linked ADP-ribose and on poly(ADP-ribose), and ADPRHL1/ARH2 has lost the metal-binding residues and is catalytically inactive. In mice, loss of Adprh causes cardiomyopathy, impaired myocyte membrane repair, greater susceptibility to cholera toxin, and spontaneous tumours.
Definition: Catalysis of the reaction alpha-NAD(+) + H2O = ADP-D-ribose + nicotinamide + H(+). This is the alpha-anomer-specific reaction; enzymes with this activity do not act on the physiological beta anomer.
Justification: UniProt records a second experimentally evidenced catalytic activity for P54922 - "alpha-NAD(+) + H2O = ADP-D-ribose + nicotinamide + H(+)", Rhea:RHEA:68792, substrate ChEBI:CHEBI:77017, ECO:0000269|PubMed:31599159 - for which no GO term exists and which consequently has no GO annotation on any of the six proteins shown to catalyse it. The obvious candidate is wrong for a reason that is invisible from its label - GO:0003953 "NAD+ nucleosidase activity" is defined as "Catalysis of the reaction: NAD+ + H2O = ADP-D-ribose + nicotinamide + H+" with the single xref EC 3.2.2.5, i.e. the beta-anomer reaction, and the finding being annotated is precisely that these enzymes hydrolyse alpha-NAD+ but NOT beta-NAD+. GO:0061809 is the cyclic-ADP-ribose-generating variant and GO:0050135 is the NADP+ term, so neither fits either. The term would apply immediately to ADPRH (ARH1), ADPRS (ARH3), MACROD1, MACROD2, OARD1 (TARG1) and archaeal Af1521.
Parent term: hydrolase activity, hydrolyzing N-glycosyl compounds
Mappings:
Supporting Evidence:
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0003875 ADP-ribosylarginine-[protein] hydrolase activity | IBA GO_REF:0000033 | ACCEPT | Summary: Accepted as the gene's core molecular function. This is a self-referential IBA whose WITH/FROM also names two orthologs that carry their own experimental evidence. Reason: Every WITH/FROM token was resolved and its own evidence queried. MGI:MGI:1098234 returns three mouse entries for Adprh - the Swiss-Prot P54923 plus two TrEMBL clones, so the lookup is ambiguous and is reported as such; P54923 is the reviewed one and holds GO:0003875 by IMP from PMID:8349667. RGD:2052 resolves unambiguously to the reviewed rat entry Q02589, which holds GO:0003875 by both IDA and IMP from PMID:1375222. UniProtKB:P54922 is the gene itself, which per campaign convention makes this a PAINT curator judging the function core rather than a circular inference. PANTHER:PTN009030515 is a tree node, not a protein; its reach was queried directly and is 20 gene products, all named adprh/Adprh/ADPRH, containing neither the inactive paralog ADPRHL1 nor the serine-specific ADPRS. The term's definition names exactly the reaction UniProt records for this protein with EC 3.2.2.19 and Rhea:RHEA:14885, so the granularity is right as well as the identity. Propagation Review Root cause: NO FAILURE CORE Sources checked: MGI:MGI:1098234 SUPPORTS TRANSFER Mouse Adprh. UniProt xref:mgi-1098234 returns 3 entries (P54923 Swiss-Prot; Q9CTF5 and Q3U5N4 TrEMBL); the reviewed entry P54923 carries GO:0003875 by IMP from PMID:8349667. RGD:2052 SUPPORTS TRANSFER Rat Adprh Q02589, reviewed, 362 aa. Carries GO:0003875 by IDA and IMP from PMID:1375222, i.e. its own experimental evidence for this exact term. PANTHER:PTN009030515 UNRESOLVED A PANTHER tree node, not a protein - unresolvable as a gene product by construction rather than through a lookup failure. Queried by withFrom in QuickGO - it reaches 20 gene products, all ADPRH orthologs, and excludes ADPRHL1 and ADPRS. UniProtKB:P54922 SUPPORTS TRANSFER The subject itself. A self-referential WITH/FROM records the PAINT curator's judgement that this is a core function of the node, and is not circular support. Supporting Evidence: PMID:30472116 Whereas ARH1 is specific for the N-glycolytic linkage of mono(ADP-ribosylated) arginines, ARH3 specifically cleaves the O-glycolytic linkage of mono(ADP-ribosylated) serine and, at least in vitro, other linkages as well, such as those in poly(ADP-ribose) chains. PMID:1375222 A hydrolase fusion protein, synthesized in Escherichia coli, reacted with anti-39-kDa polyclonal antibodies and exhibited Mg(2+)- and dithiothreitol-dependent hydrolase activity. |
| GO:0051725 protein de-ADP-ribosylation | IBA GO_REF:0000033 | ACCEPT | Summary: Accepted as the biological-process counterpart of the hydrolase activity, and confirmed to run in the erasing direction. Reason: The WITH/FROM here is narrower than on the GO:0003875 row - only the node and the gene itself - which matches the cached PAINT table, where the GO:0051725 IBD at PTN009030515 lists UniProtKB:P54922 as its sole seed while the GO:0003875 IBD lists three. So the process term is seeded by the human annotation alone; that is a self-referential IBA recording a curator judgement, not a defect. The term's definition, "The process of removing one or more ADP-ribose residues from a protein", is unambiguously the direction ADPRH runs in, and the gene independently holds the same term by IDA from PMID:30472116. Propagation Review Root cause: NO FAILURE CORE Sources checked: PANTHER:PTN009030515 UNRESOLVED Tree node, not a protein. Its reach for this term is the same 20 ADPRH orthologs as for GO:0003875. UniProtKB:P54922 SUPPORTS TRANSFER The subject itself, and the sole IBD seed for this term at this node per the cached PAINT table. Supporting Evidence: PMID:30472116 Arginine de-modification is catalyzed by ARH1, PARylation is removed by PARG and to a lesser extend ARH3, MARylation on glutamate/aspartate residues is hydrolyzed by macrodomain proteins, whereas the terminal modification on serine residues is removed by ARH3. |
| GO:0000287 magnesium ion binding | IEA GO_REF:0000002 | ACCEPT | Summary: Accepted. The signature is family-specific rather than a bare fold, and the term is independently confirmed on this protein by crystallography and mutagenesis. Reason: IPR012108 is "ADP-ribosylarginine hydrolase", a PIRSF-backed family entry, and its interpro2go mapping was fetched directly - it supplies exactly GO:0000287, GO:0003875 and GO:0051725. The three broader signatures ADPRH also matches (IPR005502 and IPR036705 ADP-ribosylation/Crystallin J1, IPR050792 ADP-ribosylglycohydrolase-like) map to nothing at all, which is the internal control showing InterPro2GO is being deliberately restrained here. For ADPRH itself the term is not merely inferred - UniProt records COFACTOR Mg(2+) with the note "Binds 2 magnesium ions per subunit", six Mg2+ BINDING features at residues 54, 55, 56, 302, 304 and 305, and PDB 6G28 and 3HFW both contain bound MG. Noted in suggested_questions that this same signature route also gives GO:0000287 to ADPRHL1, whose UniProt CAUTION states it lacks the metal-binding residues. |
| GO:0003875 ADP-ribosylarginine-[protein] hydrolase activity | IEA GO_REF:0000120 | ACCEPT | Summary: Accepted. Uniquely among this gene's IEA rows the WITH/FROM cites the reaction itself (RHEA:14885, EC:3.2.2.19), and both match UniProt's experimentally evidenced catalytic activity. Reason: The combined-IEA route reaches this term by four independent paths that all agree - the ARBA rule, the mouse and rat orthologs (both reviewed Swiss-Prot entries carrying their own experimental annotations to this term), the IPR012108 family signature, and the reaction identifiers. RHEA:14885 is the exact reaction UniProt records for P54922 with ECO:0000269|PubMed:30472116 and PhysiologicalDirection=left-to-right, and EC 3.2.2.19 is the enzyme's assigned EC number. This is the EC-derived row the review was asked to test against the measured reaction, and it passes exactly. Supporting Evidence: PMID:30472116 Whereas ARH1 is specific for the N-glycolytic linkage of mono(ADP-ribosylated) arginines, ARH3 specifically cleaves the O-glycolytic linkage of mono(ADP-ribosylated) serine and, at least in vitro, other linkages as well, such as those in poly(ADP-ribose) chains. |
| GO:0051725 protein de-ADP-ribosylation | IEA GO_REF:0000002 | ACCEPT | Summary: Accepted. Same family-specific signature as the magnesium row, and the process term is independently held by IDA and IBA. Reason: Third and last of the three terms interpro2go maps from IPR012108, and the only one of the three that is a biological process. It is redundant with the IDA row from PMID:30472116 rather than wrong, and redundancy across evidence routes is normal in GOA and is not grounds for removal. |
| GO:0005515 protein binding | IPI PMID:25416956 A proteome-scale map of the human interactome network. | MARK AS OVER ANNOTATED | Summary: Over-annotation. A single unreplicated yeast two-hybrid hit with a transcription-factor hub, contributing no functional information about this enzyme. Reason: The partner accession resolves to Q12800 TFCP2_HUMAN, a reviewed 502-aa canonical Swiss-Prot entry, so there is no TrEMBL or ORFeome substitution here. What fails is replication and informativeness. Expanding IntAct's records for P54922 shows the ADPRH-TFCP2 pair is logged three times under three sub-methods of one screen - "two hybrid array", "two hybrid prey pooling approach" and "validated two hybrid" - at MI-score 0.56 in Saccharomyces cerevisiae, which is what UniProt's NbExp=3 is counting. There is no orthogonal assay for the pair in any of the 17 IntAct records for ADPRH. TFCP2 is a hub with 350 distinct IntAct partners against ADPRH's 17, it is a nuclear transcription factor with no reported connection to ADP-ribosylation, and no follow-up study exists. Per CLAUDE.md a bare protein binding term is uninformative even when real, and here there is no candidate replacement term because there is no characterised functional relationship to name. |
| GO:0005515 protein binding | IPI PMID:32296183 A reference map of the human binary protein interactome. | MARK AS OVER ANNOTATED | Summary: Over-annotation, on the same grounds as the TFCP2 row but judged separately - a single unreplicated HuRI yeast two-hybrid hit with no follow-up. Reason: Q96BD6 resolves to SPSB1_HUMAN, a reviewed 273-aa canonical entry - again no partial-clone substitution. SPSB1 is a SPRY-domain SOCS-box substrate receptor for a cullin-RING ligase, so a degradation link would be biologically conceivable and this partner is the more interesting of the two; that is why it was judged on its own rather than folded into the TFCP2 verdict. But the evidence is the same shape - IntAct logs the pair under exactly the three HuRI sub-methods "two hybrid array", "two hybrid prey pooling approach" and "validated two hybrid", MI-score 0.56, one screen, no orthogonal assay, and SPSB1 carries 78 distinct IntAct partners. No publication examines ADPRH stability or ubiquitination. A specific replacement term such as a ubiquitin ligase binding term would require evidence that does not exist, so the row is marked over-annotated rather than modified. |
| GO:0005576 extracellular region | IEA GO_REF:0000107 | MARK AS OVER ANNOTATED | Summary: Over-annotation of the human protein. The whole claim traces to one western blot of rat cerebrospinal fluid, projected across species by Ensembl Compara. Reason: Tracing the chain - human P54922 holds this term only by IEA GO_REF:0000107 with WITH/FROM rat Adprh; querying that donor's own evidence, rat Q02589 holds GO:0005576 by a single IDA from PMID:9037477 assigned by RGD. The reference-projection test on PMID:9037477 returns 1 annotation on 1 entity when fully paginated, so this is one curator's call on one paper and not a bulk import. Reading that paper, its subject is intracellular immunostaining of neurons throughout rodent CNS, and the extracellular claim rests on a single 38-kDa immunoreactive band in rat CSF that the authors themselves present as a first observation. The rat IDA is not second-guessed here - only its transfer to human, which is weak on three independent grounds. ADPRH has no signal peptide and no transmembrane segment in its UniProt feature table and no SUBCELLULAR LOCATION comment at all, so no secretion route is annotated. Both of its verified substrates are intracellular - Gs-alpha on the cytoplasmic leaflet and TRIM72 in the myocyte membrane-repair machinery. And an immunoreactive band in a body fluid is equally consistent with release from damaged or turning-over cells. REMOVE is not used because the CSF observation is unexplained rather than refuted, and because GO:0005576 asserts only "the space external to the outermost structure of a cell" with no mechanism implied. Propagation Review Root cause: PROPAGATION BAD Failure modes: COMPARTMENT OR COMPLEX MISMATCH SOURCE EVIDENCE WEAK Sources checked: UniProtKB:Q02589 SUPPORTS SOURCE BUT NOT TARGET Rat Adprh, reviewed, 362 aa - a genuine ortholog, so the transfer is not a paralog error. Its GO:0005576 IDA comes from PMID:9037477, a rat CSF western blot in an immunohistochemistry paper whose main result is intracellular neuronal staining. ensembl:ENSRNOP00000034815 CIRCULAR OR REDUNDANT The Ensembl protein identifier for the same rat Adprh entry; not an independent source. Supporting Evidence: PMID:9037477 On Western blot analysis of rat cerebrospinal fluid (CSF), the anti-AAH antibodies recognized a protein with a molecular mass of 38 kDa. |
| GO:0000287 magnesium ion binding | IDA PMID:30472116 (ADP-ribosyl)hydrolases: Structural Basis for Differential S... | ACCEPT | Summary: Accepted. Two Mg2+ ions are resolved in the 1.23 A structure and every residue coordinating them is a mutagenesis-validated catalytic residue. Reason: This is the strongest of the three magnesium rows. UniProt derives from this paper a COFACTOR line naming Mg(2+) with the note "Binds 2 magnesium ions per subunit", six BINDING features (residues 54, 55, 56 for site 1; 302, 304, 305 for site 2), and five MUTAGEN entries - S54A, D55N, D56N, D302N and S305A - each annotated "Complete loss of activity". PDB 6G28 (1.23 A) and 6G2A both contain bound MG. Magnesium binding here is not incidental cofactor presence but the catalytic requirement itself, which is why it is carried into core_functions alongside the hydrolase activity rather than being kept as a non-core property. Supporting Evidence: PMID:30472116 The active sites of hARH1 and LchARH3 are structurally very similar and contain residues for the coordination of two Mg2+ ions. file:human/ADPRH/ADPRH-bioinformatics/RESULTS.md Magnesium is present in **4 of 4** structures (3HFW, 6G28, 6G2A, 6IUX). |
| GO:0003875 ADP-ribosylarginine-[protein] hydrolase activity | IDA PMID:30472116 (ADP-ribosyl)hydrolases: Structural Basis for Differential S... | ACCEPT | Summary: Accepted as the core molecular function, at the correct granularity - this paper is the head-to-head measurement that separates ARH1's arginine specificity from ARH3's serine specificity. Reason: The row was tested for the specific hazard this family presents, namely a term that belongs to a relative. It survives in both directions. Forwards, GO:0003875's definition - catalysis of H2O + N(omega)-(ADP-D-ribosyl)-L-arginyl-[protein] = ADP-D-ribose + L-arginyl-[protein] - is the exact reaction UniProt records for P54922 with Rhea:RHEA:14885 and EC 3.2.2.19, so there is no granularity slack to exploit and no more specific child exists. Backwards, none of the ARH3-specific terms has leaked onto ADPRH - the reference-projection test on this very PMID shows it annotated two entities, giving P54922 the arginine terms and the coelacanth ARH3 ortholog H3BCW1 GO:0004649, GO:0140290 and GO:0140292, in the same curation pass. The direction is also correct - this is a hydrolase term, and the GOA contains no term asserting that ADPRH performs ADP-ribosylation. Supporting Evidence: PMID:30472116 Whereas ARH1 is specific for the N-glycolytic linkage of mono(ADP-ribosylated) arginines, ARH3 specifically cleaves the O-glycolytic linkage of mono(ADP-ribosylated) serine and, at least in vitro, other linkages as well, such as those in poly(ADP-ribose) chains. |
| GO:0051725 protein de-ADP-ribosylation | IDA PMID:30472116 (ADP-ribosyl)hydrolases: Structural Basis for Differential S... | ACCEPT | Summary: Accepted as the core biological process. The removal direction is what was measured. Reason: GO:0051725 is defined as "The process of removing one or more ADP-ribose residues from a protein", which is what the assays in this paper measure. It is the correct process partner for the GO:0003875 activity, and it is also the term to which the redundant GO:0036211 IMP row is being redirected. Supporting Evidence: PMID:30472116 Arginine de-modification is catalyzed by ARH1, PARylation is removed by PARG and to a lesser extend ARH3, MARylation on glutamate/aspartate residues is hydrolyzed by macrodomain proteins, whereas the terminal modification on serine residues is removed by ARH3. |
| GO:0000287 magnesium ion binding | IDA PMID:19407395 Cloning, expression, purification and crystallization as wel... | ACCEPT | Summary: Accepted. A second, independent structural determination of the same magnesium site, from the earlier 1.9 A structure. Reason: PDB 3HFW, the structure reported in this paper, lists MG among its bound non-polymer components, and UniProt cites ECO:0000269|PubMed:19407395 alongside PubMed:30472116 for the site-1 magnesium BINDING features at residues 54, 55, 56 and 304. Two structures from independent groups agreeing on the metal is stronger than either alone, so this row is kept rather than folded into the PMID:30472116 row. Supporting Evidence: file:human/ADPRH/ADPRH-bioinformatics/RESULTS.md Magnesium is present in **4 of 4** structures (3HFW, 6G28, 6G2A, 6IUX). PMID:19407395 hARH1 has been cloned, expressed heterologously in Escherichia coli, purified and crystallized in complex with K(+) and ADP. |
| GO:0030955 potassium ion binding | IDA PMID:19407395 Cloning, expression, purification and crystallization as wel... | MARK AS OVER ANNOTATED | Summary: Over-annotation of a crystallisation additive. The potassium ion is genuinely resolved, and the term asserts nothing more than that - but it entered from the cocktail, does not recur in the two higher-resolution structures, and UniProt read the same paper and declined to record a potassium site. Reason: Reading the term's definition rather than its label is what keeps this at MARK_AS_OVER_ANNOTATED rather than REMOVE. GO:0030955 is exactly "Binding to a potassium ion (K+)", with no claim of functional requirement, and PDB 3HFW's bound non-polymer components are literally K and MG - so a resolved potassium ion does satisfy the term as written, and deleting the row would be the mirror of the error this review is guarding against elsewhere. What the row does not deserve is to stand unqualified. Three computed facts put it in the over-annotation class. First, provenance - the paper's own account is that K+ was discovered by X-ray fluorescence as a trace contaminant of poorly diffracting apo crystals and was then added deliberately as K-ADP because it improved diffraction, so the ion entered from the cocktail rather than from biology. Second, non-recurrence - across all four ADPRH structures, magnesium is present in 4 of 4 while potassium is present in 1 of 4, and that one is the lowest-resolution member of the set; neither 6G28 at 1.23 A nor 6IUX at 1.20 A contains any potassium. Third, and the point that decides it, a curator reading the same paper reached the opposite conclusion - UniProt created six Mg2+ BINDING features citing ECO:0000269|PubMed:19407395 and created no potassium feature at all, and names only Mg(2+) in COFACTOR. So the discrepancy is not this review against a curator; it is two curators reading one paper differently, which is exactly what MARK_AS_OVER_ANNOTATED is for. Whether K+ affects activity is recorded as an experiment worth doing. Supporting Evidence: PMID:19407395 A prerequisite for obtaining well diffracting crystals was the performance of X-ray fluorescence analysis on poorly diffracting apo hARH1 crystals, which revealed the presence of trace amounts of K(+) in the crystal. file:human/ADPRH/ADPRH-bioinformatics/RESULTS.md Every structure containing potassium is at the worst resolution in the set (1.92 A); all 3 better-resolved structures (6G28, 6G2A, 6IUX) contain none. |
| GO:0003875 ADP-ribosylarginine-[protein] hydrolase activity | IMP PMID:8349667 Cloning and site-directed mutagenesis of human ADP-ribosylar... | ACCEPT | Summary: Accepted. The original human characterisation, in which recombinant wild-type and point-mutant human enzyme were expressed and assayed. Reason: This 1993 paper cloned the human hydrolase, expressed it in E. coli and compared wild-type with the S103C point mutant, establishing both the activity and its dithiothreitol independence relative to the rodent enzymes. IMP is the right code for a mutant-based demonstration. The reference-projection test returns 4 annotations over 2 entities, human P54922 and mouse P54923, which is ordinary two-species curation of one paper rather than a projection across a set. Supporting Evidence: PMID:8349667 Human hydrolase and rat C108S were DTT-independent; human S103C was, however, DTT-dependent. |
| GO:0036211 protein modification process | IMP PMID:8349667 Cloning and site-directed mutagenesis of human ADP-ribosylar... | MODIFY | Summary: Redundant, not misdirected. GO:0036211 is a strict ancestor of the protein de-ADP-ribosylation term this gene already holds, so the fix is to replace it with the specific child. Reason: This row looks at first like the classic direction inversion - a term naming the forward modification applied to an enzyme that removes it - and the check was run rather than assumed. It is not an inversion. QuickGO's ancestors endpoint with relations=is_a,part_of returns GO:0036211 in the ancestor set of GO:0051725, via GO:0043412 macromolecule modification, so GO deliberately classifies de-modification under protein modification process and the term is simply the most general possible statement of what the 1993 experiments showed. Since the gene already holds GO:0051725 by IDA, IBA and IEA, the row adds nothing at this level. Replacing it with GO:0051725 preserves the 1993 reference's contribution to the process annotation at the granularity the same paper's activity row already supports. Proposed replacements: protein de-ADP-ribosylation Supporting Evidence: PMID:8349667 Mono-ADP-ribosylation of arginine is a reversible modification of proteins with NAD:arginine ADP-ribosyltransferases and ADP-ribosylarginine hydrolases catalyzing the opposing reactions in the cycle. |
Loading supporting contentβ¦
Download this section (compressed HTML)Q: Should the interpro2go mapping from IPR012108 and the ARBA rule behind GO_REF:0000120 be prevented from reaching ADPRHL1? Both routes give the catalytically inactive paralog ADPRHL1 the catalytic terms that are correct for ADPRH - GO:0003875 by IEA GO_REF:0000120 and GO:0000287 by IEA GO_REF:0000002 - even though Swiss-Prot's own entry for Q8NDY3 is titled "Inactive ADP-ribosyltransferase ARH2" and carries the CAUTION that it "lacks the metal-binding and substrate-binding residues, suggesting that it has no hydrolase activity". A census over all 31 reviewed Swiss-Prot members of PTHR16222 shows all 7 ADPRHL1 orthologs holding GO:0003875 while all 7 have lost both Asp56 and Ser305, at 42.6-47.7% identity to ADPRH where the alignment is reliable. The contrast with PAINT is the point rather than a complaint about IEA in general - PAINT put the arginine terms on PTN009030515, which reaches 20 ADPRH orthologs and no ADPRHL1, and gave the ADPRHL1 clade nothing. So the family's catalytic loss is already modelled correctly by one route while two others miss it.
Suggested experts: InterPro/GO Central (interpro2go curators), UniProt automatic annotation (ARBA)
Q: Does GO want a term for alpha-NAD(+) hydrolase activity? Six proteins - ARH1, ARH3, MacroD1, MacroD2, TARG1 and Af1521 - have a measured activity that no GO term can express, and the term that looks right, GO:0003953 NAD+ nucleosidase activity, is defined on the beta anomer via EC 3.2.2.5, which is exactly the anomer these enzymes do not act on. UniProt already encodes the reaction as Rhea:RHEA:68792 on P54922 with ECO:0000269|PubMed:31599159.
Suggested experts: GO Central molecular function curators
Q: Should the human GO:0005576 extracellular region annotation on ADPRH be withdrawn, and should the rat annotation it derives from be reconsidered? Human, mouse and rat all carry this term only through Ensembl Compara or ISO transfer from one rat IDA (PMID:9037477), whose evidence is a single 38-kDa band on a western blot of cerebrospinal fluid in a paper otherwise devoted to intracellular neuronal immunostaining. ADPRH has no signal peptide, no transmembrane segment and no SUBCELLULAR LOCATION comment in UniProt.
Suggested experts: RGD, Ensembl Compara GO projection maintainers
Q: Is potassium functionally relevant to ARH1, or is the K+ in PDB 3HFW purely a crystallisation artefact? The GO:0030955 IDA is literally satisfied by the structure, but K+ is absent from the two later, higher-resolution structures 6G28 and 6IUX, and UniProt records no potassium BINDING feature and no potassium in COFACTOR while recording six magnesium sites from the same papers.
Suggested experts: UniProt structural curators, Authors of PMID:19407395 and PMID:30472116
Experiment: Express recombinant human ADPRHL1 alongside ADPRH as a positive control and assay both against ADP-ribosyl-arginine and against ADP-ribosylated Gs-alpha, with and without Mg2+. Pair this with inductively coupled plasma mass spectrometry or X-ray fluorescence on purified ADPRHL1 to test directly whether it retains any metal at the degenerate site. A clean negative would let UniProt's CAUTION be upgraded to a measured statement and would justify an IRD at the ADPRHL1 node in PAINT.
Hypothesis: ADPRHL1 has no ADP-ribosylarginine hydrolase activity, so the GO:0003875 and GO:0000287 annotations it holds by IEA are false.
Type: enzyme assay
Experiment: Perform subcellular fractionation and immunofluorescence on human cells expressing endogenously tagged ADPRH, and in parallel test conditioned medium for the tagged protein under basal conditions and after deliberate cell lysis. The lysis control is the point - it distinguishes regulated release from leakage, which is the ambiguity the rat CSF blot cannot resolve.
Hypothesis: Human ADPRH is a cytosolic enzyme and is not present in the extracellular space; the 38-kDa immunoreactive species reported in rat cerebrospinal fluid is not secreted ADPRH.
Type: cell biology
Experiment: Measure ADP-ribosyl-arginine hydrolysis rates across a K+ titration at fixed ionic strength, with Na+ and choline as counter-ion controls, and attempt to reproduce a K+-bound structure from a potassium-free crystallisation condition. If activity is K+-independent and the ion does not reappear, the GO:0030955 annotation should be re-examined by the assigning curator.
Hypothesis: Potassium occupies a real site on ARH1 and modulates its catalytic rate.
Type: enzyme assay
Experiment: Run side-by-side kinetics for recombinant human ARH1 and ARH3 on ADP-ribosyl-arginine, 1''-O-acetyl-ADP-ribose and poly(ADP-ribose) under one set of conditions, reporting kcat/KM for each pair. A head-to-head comparison is what is missing - the existing claims for ARH1 come from a paper whose subject is ARH3, and the comparison that exists (PMID:30472116) assayed the arginine and serine linkages rather than the acetyl and polymer substrates.
Hypothesis: Human ARH1 does not hydrolyse O-acetyl-ADP-ribose or poly(ADP-ribose) at rates comparable to its arginine activity.
Type: enzyme kinetics
What is not known β curated, literature-grounded statements of the open unknowns (the inverse of core functions).
Gap: Whether human ARH1 hydrolyses O-acetyl-ADP-ribose and poly(ADP-ribose) in addition to ADP-ribosyl-arginine is unresolved in the literature this review can read, and the two sources disagree in emphasis.
OPEN BIOLOGYCURATION
What is known: Established - the arginine activity, by structure, kinetics and mutagenesis. Unestablished - whether the O-acetyl-ADP-ribose and poly(ADP-ribose) activities reported for ARH1 in a paper titled for ARH3 hold for the human enzyme at physiologically meaningful rates.
Significance: If they do, ADPRH would additionally warrant GO:0061463 and GO:0004649, both of which currently sit only on ADPRS in the human GOA. Proposing them now on the strength of a single restated sentence would be exactly the paralog leak this review was asked to guard against, so they are deliberately not proposed.
Provenance (the field's own admissions):
Gap: The subcellular localisation of human ADPRH has never been determined experimentally. UniProt assigns no SUBCELLULAR LOCATION at all, and the only location annotation in GOA is an Ensembl Compara projection of a rat cerebrospinal-fluid western blot.
OPEN BIOLOGYCURATION
What is known: Known - the protein is detected at protein level (PE 1) with peptides in PeptideAtlas and jPOST, has no signal peptide or transmembrane segment, and acts on substrates that are intracellular. Unknown - whether it is cytosolic, whether any pool is genuinely extracellular, and what the 38-kDa CSF species is.
Significance: A cytosol annotation is the expectation everyone works from and none exists; conversely the one location annotation that does exist points outside the cell on evidence from a different species. A single imaging or fractionation experiment on the human protein would resolve both.
Provenance (the field's own admissions):
Loading supporting contentβ¦
Download this section (compressed HTML)Loading supporting contentβ¦
Download this section (compressed HTML)Loading supporting contentβ¦
Download this section (compressed HTML)Loading supporting contentβ¦
Download this section (compressed HTML)