id: Q8NDY3
gene_symbol: ADPRHL1
product_type: PROTEIN
status: COMPLETE
taxon:
  id: NCBITaxon:9606
  label: Homo sapiens
description: >-
  ADPRHL1 (ARH2) is a catalytically inactive member of the ADP-ribosylglycohydrolase
  family, expressed almost exclusively in striated muscle and most strongly in the
  heart. It retains the all-alpha ADP-ribosylhydrolase fold of its active paralogues
  ADPRH (ARH1, an ADP-ribosylarginine hydrolase) and ADPRS (ARH3, a serine/poly(ADP-ribose)
  hydrolase), but has lost the aspartate pair that coordinates the catalytic magnesium
  ions and the serines that grip the adenosine-ribose of the substrate, and recombinant
  ARH2 shows no hydrolase activity in the assays that detect ARH1 and ARH3 activity.
  Its function instead depends on the vestigial substrate cleft: the two adjacent serines
  that bind adenosine-ribose in ADPRH are replaced in ADPRHL1 by a di-arginine motif
  (Arg271-Arg272), and deleting one to four residues from that loop abolishes ventricular
  myofibril assembly in Xenopus without touching any catalytic residue. Loss of ADPRHL1
  blocks striated myofibril assembly and outgrowth of the cardiac ventricle in frog
  embryos, and knockout in human embryonic stem cell-derived cardiomyocytes disorganises
  sarcomeres, reduces beta-1 integrin, paxillin and phosphorylated FAK, and impairs
  focal adhesion formation, calcium transients and electrical conduction through
  excessive activation of the ROCK-myosin II pathway. Epitope-tagged protein localises
  to stripes flanking the Z-disc, consistent with the sarcomeric and myofibrillar
  localisation recorded for the protein. No ligand, substrate or binding partner has
  been identified, so the molecular activity underlying these roles remains undefined.
alternative_products:
- name: '1'
  id: Q8NDY3-1
- name: '2'
  id: Q8NDY3-2
  sequence_note: VSP_023036
existing_annotations:
- term:
    id: GO:0000287
    label: magnesium ion binding
  evidence_type: IEA
  original_reference_id: GO_REF:0000002
  qualifier: enables
  supporting_entities:
  - InterPro:IPR012108
  review:
    summary: >-
      This is the ancestral catalytic magnesium site transferred onto a protein in which
      that site has been dismantled. The row comes from interpro2go, which maps
      IPR012108 "ADP-ribosylarginine hydrolase" to GO:0000287, GO:0003875 and GO:0051725;
      ADPRHL1 matches IPR012108 because the entry's single member signature, PIRSF016939,
      does not separate the active ARH1 proteins from the inactive ARH2 ones. Of the
      1282 proteins in IPR012108, 11 are reviewed Swiss-Prot entries, and 7 of those 11
      are the entries Swiss-Prot itself names "Inactive ADP-ribosyltransferase ARH2" and
      annotates with a CAUTION that they lack the metal-binding residues. So a majority
      of the entry's own curated members are curated as catalytically dead, and the entry
      still supplies the metal-binding term. UniProt states the loss directly on this
      accession, and a residue census confirms it against a matched control: of ADPRH's
      six UniProt-annotated Mg(2+) ligands, ADPRHL1 keeps the coordinating oxygen at only
      three and loses it at three. Written as ADPRH position to ADPRHL1 residue, the
      survivors are Ser54 to Ser56, Asp55 to Asp57 and Asp302 to Glu304, and the losses
      are Asp56 to Asn58, Asp304 to Ala306 and Ser305 to Ala307. The decisive one is Asp56 to
      Asn58, because aspartate-to-asparagine at a vicinal catalytic aspartate is the very
      substitution used to abolish this family's activity experimentally - PMID:17075046
      kills ARH3's reaction by replacing Asp77 and Asp78 with asparagine. The comparison
      is controlled rather than asserted: Dictyostelium ADPRH, a genuine ARH1 at 48.4%
      identity to human ADPRH - the same distance as ADPRHL1's 42.6-47.7% - retains 13 of
      20 ligand sites and all six Mg(2+) donors, while every ADPRHL1 orthologue retains
      6-7 sites and 2-3 donors. Retention is therefore not tracking sequence distance.
      One counter-consideration is recorded rather than suppressed: PMID:27217161's model
      notes that Asp100 of ADPRHL1 could provide an alternative coordination site for
      metals. That is a hedged modelling suggestion, presented in the source as an
      obstacle to substrate docking rather than as a functional site, and never tested -
      it does not support the annotation, whose provenance is specifically the ancestral
      catalytic site.
    action: REMOVE
    reason: >-
      Removed on the term's own provenance rather than by importing the hydrolase result.
      Magnesium binding as such has never been assayed on ARH2, and three of ADPRH's six
      Mg(2+) donor positions do survive, so the honest statement is not that ADPRHL1 binds
      no metal. All positions in this paragraph are given in ADPRH numbering, with the
      ADPRHL1 residue after the arrow: retained are Ser54 to Ser56, Asp55 to Asp57 and
      Asp302 to Glu304; lost are Asp56 to Asn58, Asp304 to Ala306 and Ser305 to Ala307.
      That reconciles with PMID:27217161, which reports three of the four critical
      aspartates lost with only Asp55 conserved - its verbatim sentence is quoted in this
      row's supported_by below, and nothing in this paragraph is a quotation. It uses a
      different denominator and a stricter criterion: its four are Asp55, Asp56, Asp302 and
      Asp304, and it scores Asp302 to Glu304 as lost because the residue is no longer an
      aspartate, whereas the donor test used here scores it retained because a glutamate
      still contributes a carboxylate oxygen. Both statements describe the same alignment. What is refuted is the only
      thing that put this term here: the row exists solely as interpro2go's mapping of
      IPR012108, that mapping describes the catalytic Mg(2+) centre of the active ARH1
      members, and that centre is demonstrably dismantled - UniProt's own CAUTION records
      the metal-binding residues as absent, and the losses include the vicinal Asp56, the
      position whose aspartate-to-asparagine substitution abolishes the paralogue's
      activity. There is no independent support for the term anywhere in the clade: all
      seven reviewed ARH2 orthologues hold it by the same single IEA route. With its sole
      basis refuted and no measurement on either side, REMOVE rather than
      MARK_AS_OVER_ANNOTATED - though the latter would also have been defensible, and the
      caveat in the summary about ADPRHL1 Asp100, the one position in this row's reasoning
      cited in ADPRHL1 rather than ADPRH numbering, is why.
    supported_by:
    - reference_id: file:human/ADPRHL1/ADPRHL1-uniprot.txt
      supporting_text: lacks the metal-binding and substrate-binding residues
    - reference_id: PMID:27217161
      supporting_text: >-
        Three of the four critical aspartates of ADPRH have been lost in mammalian
        ADPRHL1 (hADPRH D56, D302, D304 lost, only D55 is conserved), suggesting that
        ADPRHL1 cation binding may be compromised.
    - reference_id: PMID:36497109
      supporting_text: >-
        ARH2 (so far) does not exhibit any enzymatic activities [70], consistent with
        differences in primary sequences from critical residues found in ARH1 (e.g.,
        54-SDDT-57, 302-DSDS-305)
    - reference_id: file:human/ADPRHL1/ADPRHL1-bioinformatics/RESULTS.md
      supporting_text: Retention therefore is not tracking sequence distance here.
    propagation_review:
      root_cause: PROPAGATION_BAD
      failure_modes:
      - PSEUDO_OR_SUBACTIVITY_LOSS
      - WRONG_ORTHOLOG_OR_PARALOG
      source_entities:
      - source_id: InterPro:IPR012108
        source_label: ADP-ribosylarginine hydrolase (InterPro family, member signature PIRSF016939)
        source_status: SUPPORTS_SOURCE_BUT_NOT_TARGET
        comment: >-
          The signature is sound for the ARH1 members that justify it and unsound for the
          ARH2 members that outnumber them: 7 of the entry's 11 reviewed members are
          "Inactive ADP-ribosyltransferase ARH2" entries carrying a
          lack-of-catalytic-residues CAUTION. The three other InterPro signatures ADPRHL1
          matches - IPR050792, IPR005502 and IPR036705 - map to nothing in interpro2go,
          which is the internal control showing this is one over-reaching entry rather
          than a careless pipeline.
- term:
    id: GO:0003875
    label: ADP-ribosylarginine-[protein] hydrolase activity
  evidence_type: IEA
  original_reference_id: GO_REF:0000120
  qualifier: enables
  supporting_entities:
  - ARBA:ARBA00088955
  - InterPro:IPR012108
  review:
    summary: >-
      The activity this term names has been measured on ARH2 and was not found. Ono et al.
      assayed recombinant ARH2 alongside ARH1 and ARH3 and report it inactive; the Moss
      laboratory's family review states the negative across the whole assay panel and,
      specifically for the reaction this term describes, that ARH2 does not appear to be
      responsible for hydrolysis of the ADP-ribose-acceptor linkage. UniProt records the
      same on this entry. Two independent automatic routes deliver the term anyway and
      neither examines a residue. The first is interpro2go via IPR012108. The second is
      the ARBA rule cited in this row's own WITH/FROM: ARBA00088955 has a single condition
      set, FunFam id 1.10.4080.10:FF:000002 AND taxon Mammalia, and this entry's
      cross-reference reads "FunFam; 1.10.4080.10:FF:000002; ADP-ribosylarginine hydrolase
      isoform X1". That FunFam contains both the active ARH1 proteins and the inactive ARH2
      ones (ARH3 sits in FF:000001), so the rule is keyed on a structural family whose name
      asserts an activity that half its members have lost. The rule's taxon clause makes a
      testable prediction and GOA confirms it: the five mammalian ARH2 orthologues receive
      GO:0003875 via GO_REF:0000120 while Xenopus and zebrafish receive it via GO_REF:0000002
      instead. Across all seven reviewed ARH2 orthologues, GO:0003875, GO:0000287 and
      GO:0051725 are held 21 times and every single one is IEA - no experimental, ISS or
      IBA support exists anywhere in the clade. PAINT, by contrast, gets this right: the
      GO:0003875 IBA reaches 20 gene products from node PTN009030515 and all 20 are ADPRH
      orthologues, with no ARH2 among them. The specificity question was also asked in the
      other direction and answered no: no ARH2 orthologue carries any ARH3-type term
      (GO:0140290, GO:0004649, GO:0140292 all return zero), and the residue census scores
      ADPRHL1 at 6-7 of ARH3's 17 sites against 14-17 for the ARH3 clade itself. ADPRHL1 is
      not a misfiled serine hydrolase; it is not a hydrolase.
    action: REMOVE
    reason: >-
      A catalytic activity that was directly assayed on this protein and not detected,
      arriving from two fold-based pipelines neither of which inspects the catalytic
      residues. The field's own caveat - that the negative was obtained with model
      substrates - is why no replacement hydrolase term is proposed rather than why the
      term is kept: GO cannot express "might hydrolyse something unidentified", and
      inventing a general hydrolase term to hold that hedge would be an over-annotation
      of the opposite sign.
    supported_by:
    - reference_id: PMID:17075046
      supporting_text: >-
        The rate of O-acetyl-ADP-ribose hydrolysis by recombinant ARH3 was 250-fold that
        observed with ARH1; ARH2 and poly(ADP-ribose) glycohydrolase were inactive.
    - reference_id: PMID:36497109
      supporting_text: >-
        ARH2 does not appear to be responsible for hydrolysis of ADP-ribose-acceptor
        linkage.
    - reference_id: PMID:32726316
      supporting_text: >-
        The sequence changes suggest binding of ADP-ribosylated protein substrates and
        cation-mediated catalysis are both abolished in Adprhl1 and biochemical assays
        have confirmed the lack of any comparable enzymatic activity
    - reference_id: file:human/ADPRHL1/ADPRHL1-uniprot.txt
      supporting_text: showing no activity against O-acetyl-ADP-ribose
    propagation_review:
      root_cause: PROPAGATION_BAD
      failure_modes:
      - PSEUDO_OR_SUBACTIVITY_LOSS
      - WRONG_ORTHOLOG_OR_PARALOG
      source_entities:
      - source_id: ARBA:ARBA00088955
        source_label: ARBA rule, FunFam 1.10.4080.10:FF:000002 AND taxon Mammalia gives GO:0003875
        source_status: SOURCE_BAD
        comment: >-
          The rule's whole condition set is a CATH FunFam membership plus a taxon. The
          FunFam is named "ADP-ribosylarginine hydrolase isoform X1" and contains ADPRH and
          ADPRHL1 alike, so the rule grants a catalytic molecular function on the strength
          of a family name. Its Mammalia clause is visible in GOA as the split between
          GO_REF:0000120 for the five mammalian ARH2 orthologues and GO_REF:0000002 for the
          frog and zebrafish ones.
      - source_id: InterPro:IPR012108
        source_label: ADP-ribosylarginine hydrolase (InterPro family, member signature PIRSF016939)
        source_status: SUPPORTS_SOURCE_BUT_NOT_TARGET
        comment: >-
          Same entry as on the GO:0000287 and GO:0051725 rows; 7 of its 11 reviewed members
          are Swiss-Prot "Inactive ADP-ribosyltransferase ARH2" entries.
- term:
    id: GO:0030017
    label: sarcomere
  evidence_type: IEA
  original_reference_id: GO_REF:0000044
  qualifier: located_in
  supporting_entities:
  - UniProtKB-SubCell:SL-0313
  review:
    summary: >-
      The automatic mapping of UniProt SubCell SL-0313 (Sarcomere) to GO:0030017. Its input
      is this entry's SUBCELLULAR LOCATION line, "Cytoplasm, myofibril, sarcomere", which
      carries ECO:0000250|UniProtKB:Q6AZR2 - so this row and the ISS row below are the same
      underlying observation arriving by two routes rather than two independent findings,
      and they are given the same verdict. The observation itself is the Xenopus work: an
      N-terminally epitope-tagged Adprhl1 localises to stripes flanking the Z-disc. Two
      caveats are recorded rather than glossed. It is over-expressed tagged protein, and
      the same study reports that its peptide antibody was not sensitive enough to detect
      endogenous Adprhl1 in situ, so the endogenous distribution has not been imaged in any
      species. And no human localisation experiment exists at all. Kept because the
      inference is nonetheless well founded - a 74.6%-identical 1:1 ortholog, a heart- and
      skeletal-muscle-restricted expression pattern, and human knockout cardiomyocytes in
      which sarcomere arrangement is disturbed - and because a location term that is very
      likely true should not be deleted on the strength of an assay caveat.
    action: ACCEPT
    reason: >-
      A true but indirectly evidenced compartment call. Same datum as the ISS row, reached
      through UniProt's SubCell vocabulary rather than through the curator's ortholog
      transfer.
    supported_by:
    - reference_id: file:human/ADPRHL1/ADPRHL1-uniprot.txt
      supporting_text: 'SUBCELLULAR LOCATION: Cytoplasm, myofibril, sarcomere'
    - reference_id: PMID:27217161
      supporting_text: Recombinant Adprhl1 can localize to stripes adjacent to the Z-disc
- term:
    id: GO:0051725
    label: protein de-ADP-ribosylation
  evidence_type: IEA
  original_reference_id: GO_REF:0000002
  qualifier: involved_in
  supporting_entities:
  - InterPro:IPR012108
  review:
    summary: >-
      The biological process that the absent hydrolase activity would carry out, transferred
      from the same IPR012108 signature and falling with it. The direction of this term was
      checked explicitly rather than assumed, because this family is exactly where reaction
      direction goes wrong: GO:0051725 is the removal of ADP-ribose, which is the correct
      direction for an ARH-family protein, so this is not the error of annotating a protein
      as performing the modification it removes. It fails for a different reason - ADPRHL1
      has never been shown to remove ADP-ribose from anything, and was assayed for exactly
      that and found inactive. Nothing else in the literature supports the process either.
      The closest is PMID:35816343, in which a germline p.Asp78Val allele expressed in
      prostate cells activates PARP1 and raises poly(ADP-ribose) levels while wild-type
      ADPRHL1 suppresses proliferation; that is consistent with wild-type ADPRHL1 restraining
      cellular ADP-ribosylation somehow, but it is ectopic expression of a single variant
      in cancer cell lines with no demonstration that ADPRHL1 acts on ADP-ribose itself,
      and it cannot support a de-ADP-ribosylation annotation. Note that Asp78 is not one of
      the ligand positions in the residue census, so the variant does not bear on the
      catalytic question either.
    action: REMOVE
    reason: >-
      A process term that exists solely as the downstream consequence of a catalytic
      activity this protein does not have. Removed on the same evidence as the GO:0003875
      row, and from the same signature.
    supported_by:
    - reference_id: PMID:36497109
      supporting_text: >-
        ARH2 does not appear to be responsible for hydrolysis of ADP-ribose-acceptor
        linkage.
    - reference_id: PMID:17075046
      supporting_text: >-
        The rate of O-acetyl-ADP-ribose hydrolysis by recombinant ARH3 was 250-fold that
        observed with ARH1; ARH2 and poly(ADP-ribose) glycohydrolase were inactive.
    - reference_id: file:human/ADPRHL1/ADPRHL1-uniprot.txt
      supporting_text: lacks the metal-binding and substrate-binding residues
    propagation_review:
      root_cause: PROPAGATION_BAD
      failure_modes:
      - PSEUDO_OR_SUBACTIVITY_LOSS
      - WRONG_ORTHOLOG_OR_PARALOG
      source_entities:
      - source_id: InterPro:IPR012108
        source_label: ADP-ribosylarginine hydrolase (InterPro family, member signature PIRSF016939)
        source_status: SUPPORTS_SOURCE_BUT_NOT_TARGET
        comment: >-
          The third of the three terms interpro2go attaches to this entry. It reaches all
          seven reviewed ARH2 orthologues, in every case as the only evidence any of them
          has for the process.
- term:
    id: GO:0003242
    label: cardiac chamber ballooning
  evidence_type: ISS
  original_reference_id: GO_REF:0000024
  qualifier: involved_in
  supporting_entities:
  - UniProtKB:Q6AZR2
  review:
    summary: >-
      A sound ortholog transfer. The WITH/FROM entity is correct for an ISS row - the
      sequence-similar protein, not a partner - and it resolves to ARHL1_XENLA, a reviewed
      Swiss-Prot entry for Xenopus laevis adprhl1, 354 aa, a 1:1 ortholog at 74.6% aligned
      identity (published as 75%). Its own annotation is IMP from PMID:27217161, where
      morpholino knockdown of all Adprhl1 variants prevents outgrowth of the ventricle while
      leaving cardiomyocyte number, electrical conduction and myofibrillar gene expression
      intact. A fully paginated reference-projection test on that PMID returns 3 annotations
      over 1 entity, so this is one curator reading one paper about one gene, not a bulk
      import and not a complex-level phenotype spread across a set. The mammalian evidence
      is thinner but not contradictory: mice lacking Adprhl1 exons 3-4 are normal, but that
      allele still produces the smaller ADPRHL1 species and so is not a null. ADPRHL1 is
      heart- and skeletal-muscle-enriched in human tissue and expressed in left ventricle
      myocardium, and human variants in the locus associate with ventricular conduction
      phenotypes, so nothing about the human gene argues against the transfer.
    action: ACCEPT
    reason: >-
      Legitimate ISS from a reviewed 1:1 ortholog at 75% identity whose underlying evidence
      is a specific, single-gene loss-of-function experiment in the same organ.
    supported_by:
    - reference_id: PMID:27217161
      supporting_text: >-
        Morpholino oligonucleotide-mediated knockdown of all Adprhl1 variants inhibits
        striated myofibril assembly and prevents outgrowth of the ventricle.
    - reference_id: PMID:36497109
      supporting_text: >-
        ARH2 is a cytoplasmic protein expressed primarily in heart and skeletal muscle and
        may be involved in cardiac myofibril assembly and cardiac chamber outgrowth
- term:
    id: GO:0030017
    label: sarcomere
  evidence_type: ISS
  original_reference_id: GO_REF:0000024
  qualifier: located_in
  supporting_entities:
  - UniProtKB:Q6AZR2
  review:
    summary: >-
      The curator's ortholog transfer of the Xenopus localisation, and the origin of the
      SubCell IEA row above. Q6AZR2 holds GO:0030017 by IMP from PMID:27217161 - IMP rather
      than IDA because the localisation was obtained from transgenically over-expressed,
      N-terminally tagged protein. The same caveats therefore apply here as to the IEA row:
      it is tagged over-expressed protein, the endogenous protein has never been imaged in
      situ in any species, and there is no human localisation experiment. It is accepted
      because the ortholog relationship is close and unambiguous, because the compartment is
      consistent with everything else known about the protein - muscle-restricted
      expression, a knockout that disorganises sarcomeres in human cardiomyocytes, and a
      reported association of mouse ARH2 with a synemin, desmin and triadin-containing
      assembly - and because the two GO:0030017 rows describe one observation and must take
      one action.
    action: ACCEPT
    reason: >-
      Correctly executed ISS from a reviewed 1:1 ortholog. The assay caveat is recorded as a
      knowledge gap rather than used to delete a compartment call that is very likely true.
    supported_by:
    - reference_id: PMID:27217161
      supporting_text: Recombinant Adprhl1 can localize to stripes adjacent to the Z-disc
    - reference_id: PMID:27217161
      supporting_text: >-
        immunocytochemistry of hearts using the same antibody was not sufficiently sensitive
        to detect endogenous Adprhl1 protein in situ
- term:
    id: GO:0055003
    label: cardiac myofibril assembly
  evidence_type: ISS
  original_reference_id: GO_REF:0000024
  qualifier: involved_in
  supporting_entities:
  - UniProtKB:Q6AZR2
  review:
    summary: >-
      The best-supported annotation on this gene, and the only one with independent human
      evidence behind it. The ISS is faithful to its donor: Q6AZR2 holds GO:0055003 by IMP
      from PMID:27217161, where knockdown blocks striated myofibril assembly and
      over-expression produces disarrayed, branching myofibrils with sarcomere division at
      the actin-Z-disc boundary. It is independently corroborated in human cells that GOA has
      not yet seen: CRISPR knockout of ADPRHL1 in H9 embryonic stem cell-derived
      cardiomyocytes leaves cardiac differentiation and myofibrillar gene expression intact
      but disturbs sarcomere arrangement. Accepted as it stands, with two additive
      recommendations recorded separately rather than by modifying this row - a more specific
      ventricular term that exists in GO but has been curated only onto TrEMBL copies of the
      Xenopus gene, and the human IMP evidence that should sit alongside this ISS. Why the
      human sarcomere result is left as a curation recommendation rather than proposed as its
      own row, when the same paper's focal adhesion result is proposed: the focal adhesion
      finding is the paper's central, quantified result, measured on four independent markers
      and closed by a pharmacological rescue, and it has no annotation in GOA at all. The
      sarcomere finding is a qualitative supplementary-figure observation with no
      quantification, and the process it would support is already annotated here and refined
      by the GO:0055005 row below - so a third overlapping myofibril term drawn from the
      weaker of the paper's two observations would add granularity the observation does not
      carry. GO:0045214 sarcomere organization was considered on that basis and declined.
    action: ACCEPT
    reason: >-
      A faithful ortholog transfer of a specific loss-of-function phenotype, now also
      supported by a human knockout. Left unmodified because the row correctly reflects the
      term its cited donor holds; the refinement is proposed additively below.
    supported_by:
    - reference_id: PMID:27217161
      supporting_text: >-
        Morpholino oligonucleotide-mediated knockdown of all Adprhl1 variants inhibits
        striated myofibril assembly and prevents outgrowth of the ventricle.
    - reference_id: PMID:37880701
      supporting_text: the arrangement of sarcomeres was less regular
- term:
    id: GO:0055005
    label: ventricular cardiac myofibril assembly
  evidence_type: ISS
  original_reference_id: GO_REF:0000024
  qualifier: involved_in
  supporting_entities:
  - UniProtKB:A0A8J0UG54
  - UniProtKB:A0A8J0UIM8
  review:
    summary: >-
      Proposed addition, and it exists only because a reference-projection test on the 2020
      CRISPR paper turned up a split donor. All three of ADPRHL1's ISS rows transfer from
      UniProtKB:Q6AZR2, the Swiss-Prot entry for Xenopus laevis adprhl1, which carries the
      2016 morpholino paper's three IMP annotations. But Xenbase curated the 2020 CRISPR
      paper onto two TrEMBL accessions for the same frog gene, A0A8J0UG54 and A0A8J0UIM8
      (adprhl1.L), which the UniProt ISS pipeline never consulted. Those entries hold
      GO:0055005 ventricular cardiac myofibril assembly, GO:0030041 actin filament
      polymerization and GO:0036342 post-anal tail morphogenesis by IMP, and none of the
      three has ever propagated to any mammal. GO:0055005 is an is_a descendant of the
      GO:0055003 that ADPRHL1 already holds, and the evidence for it is specific: CRISPR
      deletion of one to four residues from the Arg271-Arg272 loop at the centre of the
      vestigial ADP-ribosylhydrolase cleft causes loss of ventricular myofibril assembly,
      which is also the cleanest demonstration anywhere that ADPRHL1's function is
      cleft-dependent and catalysis-independent. Human relevance is direct - ADPRHL1 is
      expressed in left ventricle myocardium and its knockout disorganises sarcomeres in
      human cardiomyocytes. GO:0030041 was considered and declined for human: the Xenopus
      IMP is a morphological inference from myofibril phenotypes rather than a
      polymerization assay, and the human mechanism places any actin effect downstream of
      ROCK-myosin II. GO:0036342 is Xenopus-specific and not transferable.
    action: NEW
    reason: >-
      A more specific term than the existing GO:0055003 row, curated for the orthologous
      gene from a later paper, which failed to propagate because the two Xenopus papers were
      curated onto different accessions of the same gene. Added rather than substituted for
      GO:0055003, because that row correctly reflects the donor it cites.
    supported_by:
    - reference_id: PMID:32726316
      supporting_text: >-
        Mutant alleles encode discrete loss of 1, 3 or 4 amino acids from a di-arginine
        (Arg271-Arg272) containing peptide loop at the centre of the ancestral
        ADP-ribosylhydrolase site.
    - reference_id: PMID:32726316
      supporting_text: >-
        Thus despite lacking catalytic activity, it is the modified (adenosine-ribose)
        substrate binding cleft of Adprhl1 that fulfils an essential role during heart
        formation.
    - reference_id: PMID:37880701
      supporting_text: the arrangement of sarcomeres was less regular
- term:
    id: GO:0051894
    label: positive regulation of focal adhesion assembly
  evidence_type: IMP
  original_reference_id: PMID:37880701
  qualifier: involved_in
  review:
    summary: >-
      Proposed addition from the only functional experiment ever performed on the human
      protein, which contributes nothing to GOA at present - a fully paginated
      reference-projection test on PMID:37880701 returns zero annotations over zero
      entities. CRISPR knockout of ADPRHL1 in human H9 embryonic stem cell-derived
      cardiomyocytes reduces beta-1 integrin, phosphorylated FAK, paxillin and talin, and
      impairs focal adhesion formation; the cells detach from the culture surface, and
      calcium transients and microelectrode-array conduction are perturbed. Inhibiting ROCK
      or myosin II restores focal adhesions and rescues the electrical and calcium
      phenotypes, placing the defect downstream of excessive ROCK-myosin II activity. The
      direction matters and is stated by the experiment: knockout decreases focal adhesion
      formation, so wild-type ADPRHL1 promotes it. The regulation term rather than the bare
      GO:0048041 is chosen deliberately - ADPRHL1 is not a structural component of focal
      adhesions and the paper's mechanism is regulatory and indirect, running through the
      ROCK-myosin II pathway. Evidence code is IMP: this is a genetic knockout scored by
      phenotype, not a direct assay of ADPRHL1 acting on a focal adhesion. No term is
      proposed for the calcium and conduction phenotypes, which the same experiments show
      are downstream of the adhesion defect and are rescued by relieving it. One caveat
      recorded rather than argued away: the paper's own model makes the focal adhesion defect
      downstream of de-repressed ROCK-myosin II, so a curator might reasonably prefer a
      Rho-kinase-pathway term over the focal adhesion one. This row stays on the phenotype
      that was actually measured, on four markers, and rescued - the pathway step is inferred
      from RNA-seq and inhibitor sensitivity, and no assay places ADPRHL1 on ROCK itself.
    action: NEW
    reason: >-
      Direct human loss-of-function evidence for a regulatory role in focal adhesion
      assembly, in the correct direction, currently absent from GOA entirely.
    supported_by:
    - reference_id: PMID:37880701
      supporting_text: suggesting that ADPRHL1 deficiency disrupts the FAs in these cells
    - reference_id: PMID:37880701
      supporting_text: >-
        the expression of Paxillin, a structural protein in FAs, in KO-CM was markedly
        reduced
    - reference_id: PMID:37880701
      supporting_text: >-
        Our results demonstrated that ADPRHL1-deficient cardiomyocytes exhibited abnormal
        adhesion, calcium transients and electrophysiological activity.
core_functions:
- description: >-
    Cleft-dependent, catalysis-independent scaffolding activity required for striated
    myofibril assembly in the growing cardiac ventricle. ADPRHL1 is a pseudoenzyme whose
    ancestral ADP-ribosylhydrolase site has been converted into a di-arginine-containing
    surface (Arg271-Arg272 in place of ADPRH's adenosine-ribose-binding Ser269 and Ser270);
    small in-frame deletions in that loop abolish ventricular myofibril assembly while
    leaving cardiomyocyte specification, myofibrillar gene expression and electrical
    conduction intact. Epitope-tagged protein sits in stripes flanking the Z-disc. The
    ligand or partner engaged by the modified cleft is unknown, so no molecular function
    term is asserted.
  directly_involved_in:
  - id: GO:0055005
    label: ventricular cardiac myofibril assembly
  - id: GO:0003242
    label: cardiac chamber ballooning
  locations:
  - id: GO:0030017
    label: sarcomere
  supported_by:
  - reference_id: PMID:32726316
    supporting_text: >-
      Thus despite lacking catalytic activity, it is the modified (adenosine-ribose)
      substrate binding cleft of Adprhl1 that fulfils an essential role during heart
      formation.
  - reference_id: PMID:27217161
    supporting_text: >-
      Morpholino oligonucleotide-mediated knockdown of all Adprhl1 variants inhibits
      striated myofibril assembly and prevents outgrowth of the ventricle.
- description: >-
    Restraint of the ROCK-myosin II pathway in cardiomyocytes, permitting focal adhesion
    assembly. Loss of ADPRHL1 in human embryonic stem cell-derived cardiomyocytes
    over-activates ROCK-myosin II, reduces beta-1 integrin, phosphorylated FAK, paxillin
    and talin, and impairs focal adhesion formation, cell attachment, calcium handling and
    electrical conduction; pharmacological inhibition of ROCK or of myosin II restores
    them. Whether ADPRHL1 acts on the pathway directly is not established. No location is
    given for this function deliberately: focal adhesions and costameres are not the
    sarcomere, and the only compartment ever reported for the protein is the sarcomere, so
    carrying it over here would assert a place this activity has not been shown to occupy.
  directly_involved_in:
  - id: GO:0051894
    label: positive regulation of focal adhesion assembly
  supported_by:
  - reference_id: PMID:37880701
    supporting_text: >-
      Our results demonstrated that ADPRHL1-deficient cardiomyocytes exhibited abnormal
      adhesion, calcium transients and electrophysiological activity.
  - reference_id: PMID:37880701
    supporting_text: >-
      the expression of Paxillin, a structural protein in FAs, in KO-CM was markedly
      reduced
knowledge_gaps:
- gap_statement: >-
    ADPRHL1 has no known molecular function. No ligand, no substrate, no verified protein
    partner and no measured biochemical activity of any kind has been reported. The
    genetics say the vestigial substrate cleft is what matters, but what that cleft binds
    is undetermined - the 2020 study lists it as an open question, asking whether Adprhl1
    retains binding for a partially degraded ADP-ribose derivative such as phospho-ribose,
    or whether the cleft is itself a target for ADP-ribosylation.
  boundary: >-
    Established: the protein exists and is detected by mass spectrometry (UniProt PE 1); it
    folds on the all-alpha ADP-ribosylglycohydrolase scaffold; it is inactive against
    O-acetyl-ADP-ribose and against the ADP-ribose-acceptor linkages that ARH1 and ARH3
    cleave; three of six magnesium-coordinating positions and both adenosine-ribose serines
    are lost; and one to four residue deletions in the Arg271-Arg272 loop abolish
    ventricular myofibril assembly. Not established: any binding activity, any partner, and
    whether the loop acts as a binding surface, a protein-protein interface or something
    else. IntAct returns zero interactions for Q8NDY3, against 17 for ADPRH and 35 for
    ADPRS by the same query, so this is a real absence and not a failed lookup.
  gap_kind:
  - BIOLOGY
  dark_aspect: MF_DARK
  status: OPEN
  significance: >-
    Every molecular function currently in GOA for this gene is removed by this review, so
    after it the gene has no MF annotation at all. That is the honest state: ADPRHL1 is an
    essential cardiac protein with a completely undefined activity, and the family review
    itself cautions that the in vitro negatives may reflect the use of model substrates
    rather than genuine inertness.
  resolution: >-
    Identify what the Arg271-Arg272 cleft binds - for example by photo-crosslinking or
    pull-down from cardiomyocyte extracts using wild-type versus loop-deleted ADPRHL1 as
    bait - and re-test hydrolase activity against physiological ADP-ribosylated cardiac
    substrates rather than the small-molecule model substrates used so far.
  provenance:
  - reference_id: PMID:36497109
    supporting_text: >-
      However, the lack of ARH2 catalytic activity may be the result of the use of model
      substrates.
  - reference_id: PMID:36497109
    supporting_text: Function of ARH2 would be an interesting area of future study.
- gap_statement: >-
    The subcellular localisation of endogenous ADPRHL1 has never been imaged, in any
    species. Every localisation statement in the record - UniProt's SUBCELLULAR LOCATION
    line, the SubCell IEA row and the ISS row - descends from a single Xenopus experiment
    using transgenically over-expressed, N-terminally epitope-tagged protein.
  boundary: >-
    Established: over-expressed tagged Xenopus Adprhl1 localises to stripes flanking the
    Z-disc and a diffuse stripe at the H-zone; human ADPRHL1 is heart-, skeletal-muscle- and
    tongue-enriched and expressed in left ventricle myocardium; and human ADPRHL1 knockout
    disorganises sarcomeres. Not established: where the endogenous protein is, at native
    abundance, in any species - the antibody raised against the protein was not sensitive
    enough to detect it in situ - and nothing at all about the human protein's location.
  gap_kind:
  - BIOLOGY
  - CURATION
  dark_aspect: CC_DARK
  status: OPEN
  significance: >-
    Two of the seven GOA rows are this one observation, and it is the only positive evidence
    for the cellular component. If the tagged construct mislocalises, both rows are wrong
    together and there is no independent check anywhere in the record.
  resolution: >-
    Endogenous immunofluorescence or a knock-in fluorescent tag at the native locus in human
    iPSC-derived cardiomyocytes, with the ADPRHL1 knockout line as the specificity control.
  provenance:
  - reference_id: PMID:27217161
    supporting_text: >-
      immunocytochemistry of hearts using the same antibody was not sufficiently sensitive
      to detect endogenous Adprhl1 protein in situ
- gap_statement: >-
    Whether a mammal actually requires ADPRHL1 for heart development is untested. The only
    published mammalian allele deletes exons 3-4, and those animals are normal - but the
    allele still produces the smaller ADPRHL1 protein species, so it is not a null and
    cannot answer the question.
  boundary: >-
    Established: loss of all Adprhl1 variants in Xenopus prevents ventricle outgrowth; loss
    in human embryonic stem cell-derived cardiomyocytes impairs adhesion, calcium handling
    and conduction; and mice lacking exons 3-4 are viable and normal while retaining the
    small protein species, which localises the essential activity to the C-terminal portion.
    Not established: the phenotype of a true mammalian null. As of the 2022 family review no
    Arh2-deficient mouse had been reported; a knockout has since been described only in a
    preprint with no PubMed identifier, which this review does not rely on.
  gap_kind:
  - BIOLOGY
  dark_aspect: BP_DARK
  status: NARROWING
  significance: >-
    Three of the seven GOA rows are ISS transfers of Xenopus developmental phenotypes into
    human. They are sound as ortholog inferences, but no mammalian loss-of-function result
    yet tests them, and the one mammalian allele that exists is hypomorphic by construction.
  resolution: >-
    A conditional or full mouse Adprhl1 null in which loss of both the 40 kDa and the
    smaller protein species is confirmed by western blot, phenotyped for chamber
    morphogenesis and myofibril organisation.
  provenance:
  - reference_id: PMID:32726316
    supporting_text: >-
      Mice lacking Adprhl1 exons 3-4 are normal but production of the smaller ADPRHL1
      species is unaffected
  - reference_id: PMID:32726316
    supporting_text: What is the precise composition of the smaller 23 kDa Adprhl1 protein?
suggested_questions:
- question: >-
    To InterPro and GO Central: IPR012108 "ADP-ribosylarginine hydrolase" maps to GO:0000287,
    GO:0003875 and GO:0051725, but a majority of its own reviewed members are pseudoenzymes.
    Of the entry's 1282 proteins, 11 are Swiss-Prot; 4 are named "ADP-ribosylhydrolase ARH1"
    and 7 are named "Inactive ADP-ribosyltransferase ARH2", each carrying a CAUTION that it
    lacks the metal-binding and substrate-binding residues. The result is 21 IEA annotations
    across seven ARH2 orthologues with no experimental, ISS or IBA support anywhere behind
    them. Either the entry should be split so that PIRSF016939's ARH1 and ARH2 branches are
    separate signatures, or the three GO mappings should be withdrawn from it. Note the
    contrast that shows this is one entry and not a systemic problem: the other three
    signatures ADPRHL1 matches - IPR050792, IPR005502 and IPR036705 - map to nothing at all,
    and PAINT independently reaches the correct answer, giving GO:0003875 to 20 ADPRH
    orthologues from node PTN009030515 and to no ARH2 protein.
  experts:
  - InterPro curators
  - GO Central
- question: >-
    To UniProt (ARBA): rule ARBA00088955 assigns GO:0003875 on a single condition set, FunFam
    id 1.10.4080.10:FF:000002 AND taxon Mammalia. That FunFam is named "ADP-ribosylarginine
    hydrolase isoform X1" and contains ADPRH and ADPRHL1 alike - P54922, Q8NDY3, Q8BGK2,
    Q5XIB3, Q3ZBM1, Q5RCJ0, Q6AZR2 and Q5XJB9 all match it, while ARH3 sits in FF:000001. The
    rule therefore grants a catalytic molecular function to five mammalian pseudoenzymes on
    the strength of a structural family's name, in entries whose own CAUTION line says the
    catalytic residues are absent. A negative taxon or a negative condition excluding the ARH2
    orthologues would fix it.
  experts:
  - UniProt ARBA developers
- question: >-
    To UniProt (nomenclature): Q8NDY3's RecName is "Inactive ADP-ribosyltransferase ARH2", and
    the same name is used for all seven ARH2 orthologues. Transferase is the wrong reaction
    direction. The entry's own SIMILARITY line assigns it to the ADP-ribosylglycohydrolase
    family, its AltName is "[Protein ADP-ribosylarginine] hydrolase-like protein 1", and no
    ARH-family protein has been proposed to add ADP-ribose. Suggest "Inactive
    ADP-ribosylhydrolase ARH2", consistent with ADPRH_HUMAN "ADP-ribosylhydrolase ARH1" and
    ADPRS_HUMAN "ADP-ribosylhydrolase ARH3". This name is what GOA prints in the GENE NAME
    column of every ADPRHL1 annotation.
  experts:
  - UniProt nomenclature curators
- question: >-
    To GOA and Xenbase: the ISS pipeline and Xenbase are curating the same Xenopus laevis gene
    onto different accessions, and one whole paper is invisible as a result. All three of
    human ADPRHL1's ISS rows transfer from UniProtKB:Q6AZR2 (ARHL1_XENLA, Swiss-Prot), which
    carries the 2016 morpholino paper. Xenbase curated the 2020 CRISPR paper (PMID:32726316)
    onto A0A8J0UG54 and A0A8J0UIM8, TrEMBL entries for adprhl1.L, giving GO:0055005,
    GO:0030041 and GO:0036342 by IMP - none of which has reached any mammalian orthologue.
    Merging or cross-linking those accessions would let the later evidence propagate.
  experts:
  - GOA curators
  - Xenbase curators
- question: >-
    To GOA: PMID:37880701 is the only functional experiment ever performed on human ADPRHL1 -
    a CRISPR knockout in H9 embryonic stem cell-derived cardiomyocytes - and it contributes
    zero annotations. Beyond the focal adhesion term proposed here, it supports human IMP
    evidence for GO:0055003, which the gene currently holds only by ISS from frog.
  experts:
  - GOA curators
- question: >-
    To the ADPRH reviewers (paint/ADPRH, PR 2332): the two reviews were run independently and
    agree on every residue and mapped position across the five shared positions and five
    shared accessions. They differ on the class label at seven of those calls, because that
    review scores substitutions with hand-defined conservative groups while this one used
    BLOSUM62. The difference is entirely at aspartate-to-asparagine and serine-to-alanine,
    which BLOSUM62 scores positively. This review resolves the clash in favour of the ADPRH
    review by adding a third, mechanism-anchored test - whether the coordinating carboxylate
    or hydroxyl oxygen survives - which calls both substitutions lost. The anchor is that
    aspartate-to-asparagine at a vicinal catalytic aspartate is exactly the substitution
    PMID:17075046 uses to abolish ARH3 activity. Recorded here rather than silently absorbed.
  experts:
  - ai-gene-review PAINT campaign reviewers
suggested_experiments:
- description: >-
    Identify what the Arg271-Arg272 cleft binds. Express wild-type ADPRHL1 and the
    loop-deletion alleles characterised in PMID:32726316 as tagged baits in human iPSC-derived
    cardiomyocytes and compare interactomes; a partner lost specifically by the loop deletions
    is a candidate for the molecular function that this review could not annotate.
- description: >-
    Re-test hydrolase activity against physiological substrates. The published negatives used
    O-acetyl-ADP-ribose and other model substrates; the family review flags that as a possible
    explanation for the null result. Assay recombinant ADPRHL1 against ADP-ribosylated proteins
    purified from cardiac muscle, and against phospho-ribosylated derivatives, which the 2020
    paper raises as a candidate remaining ligand.
- description: >-
    Image endogenous ADPRHL1. Knock a fluorescent tag into the native ADPRHL1 locus in human
    iPSCs, differentiate to cardiomyocytes, and image against the knockout line as the
    specificity control. This would give the first localisation of the protein at native
    abundance in any species and the first in human, and would test whether the Z-disc-flanking
    distribution seen with over-expressed tagged Xenopus protein is genuine.
- description: >-
    Make a true mammalian null. The existing mouse allele deletes exons 3-4 and still produces
    the smaller ADPRHL1 species. A full knockout, verified by loss of both protein species on
    western blot, would test whether the Xenopus chamber-outgrowth requirement holds in a
    mammal and would validate or refute the three ISS rows in this review.
- description: >-
    Determine which ADPRHL1 protein species matters. The 40 kDa and 23 kDa forms differ in
    abundance across development and the mouse exon 3-4 allele separates them genetically.
    Define the 23 kDa species by mass spectrometry and test the two forms separately for rescue
    of the knockout phenotype in human cardiomyocytes.
references:
- id: GO_REF:0000002
  title: Gene Ontology annotation through association of InterPro records with GO
    terms
  findings:
  - statement: >-
      The route behind the GO:0000287 and GO:0051725 rows, and behind GO:0003875 in the
      non-mammalian ARH2 orthologues. Its input is IPR012108, whose interpro2go entry maps
      to exactly those three terms and whose reviewed membership is 4 ARH1 proteins and 7
      Swiss-Prot-declared inactive ARH2 proteins.
  reference_review:
    relevance: HIGH
    correctness: VERIFIED
    review_notes: >-
      The pipeline is correctly applied - the interpro2go file does map IPR012108 to these
      three terms, confirmed by downloading it and looking up all four signatures ADPRHL1
      matches. The defect is in the input signature, not the mapping machinery, and the
      other three signatures mapping to nothing is the control that shows so.
- id: GO_REF:0000024
  title: Manual transfer of experimentally-verified manual GO annotation data to orthologs
    by curator judgment of sequence similarity
  findings:
  - statement: >-
      The route behind the three ISS rows, all citing UniProtKB:Q6AZR2 (ARHL1_XENLA), a
      reviewed 1:1 ortholog at 74.6% aligned identity. Correct WITH/FROM entity type for
      ISS. The limitation is not in the transfer but in the donor set: the same Xenopus gene
      also has Xenbase-curated annotations on TrEMBL accessions that this route never sees.
  reference_review:
    relevance: HIGH
    correctness: VERIFIED
    review_notes: >-
      Q6AZR2 resolved via the UniProt REST API - reviewed Swiss-Prot, Xenopus laevis
      adprhl1, 354 aa, holding GO:0003242, GO:0055003 and GO:0030017 by IMP from
      PMID:27217161. Not a paralog and not a partial clone.
- id: GO_REF:0000044
  title: Gene Ontology annotation based on UniProtKB/Swiss-Prot Subcellular Location
    vocabulary mapping, accompanied by conservative changes to GO terms applied by
    UniProt
  findings:
  - statement: >-
      The route behind the second GO:0030017 row. Its input is SubCell SL-0313, which comes
      from this entry's SUBCELLULAR LOCATION line, itself ECO:0000250|UniProtKB:Q6AZR2 - so
      this row and the ISS GO:0030017 row are one observation reaching GOA twice.
  reference_review:
    relevance: MEDIUM
    correctness: VERIFIED
    review_notes: >-
      Correctly applied. Worth noting only that it is not independent of the ISS row, which
      is why the two rows are given the same action.
- id: GO_REF:0000120
  title: Combined Automated Annotation using Multiple IEA Methods
  findings:
  - statement: >-
      The route behind the GO:0003875 row in the five mammalian ARH2 orthologues. Its
      WITH/FROM names ARBA:ARBA00088955, whose full condition set - fetched from the UniProt
      ARBA endpoint - is FunFam id 1.10.4080.10:FF:000002 AND taxon Mammalia. The Mammalia
      clause is directly visible in GOA as the GO_REF:0000120 versus GO_REF:0000002 split
      between the mammalian and non-mammalian orthologues.
  reference_review:
    relevance: HIGH
    correctness: VERIFIED
    review_notes: >-
      Correctly applied by the pipeline; the rule itself is the problem. Verified by fetching
      the rule and checking that its FunFam condition matches ADPRHL1's own DR FunFam line,
      and by confirming that the predicted mammalian and non-mammalian reference split is
      exactly what GOA shows across the seven reviewed ARH2 orthologues.
- id: PMID:27217161
  title: The cardiac-restricted protein ADP-ribosylhydrolase-like 1 is essential for
    heart chamber outgrowth and acts on muscle actin filament assembly.
  findings:
  - statement: >-
      The experimental root of all three ISS rows. Morpholino knockdown of all Adprhl1
      variants in Xenopus laevis blocks striated myofibril assembly and prevents ventricle
      outgrowth, without affecting cardiomyocyte number, calcium signal propagation or
      myofibrillar gene expression.
    supporting_text: >-
      Morpholino oligonucleotide-mediated knockdown of all Adprhl1 variants inhibits
      striated myofibril assembly and prevents outgrowth of the ventricle.
  - statement: >-
      Structural modelling on the ADPRH crystal structure 3HFW identifies the specific
      losses in the ADPRHL1 active site, listing them in ADPRH numbering.
    supporting_text: >-
      Three of the four critical aspartates of ADPRH have been lost in mammalian ADPRHL1
      (hADPRH D56, D302, D304 lost, only D55 is conserved), suggesting that ADPRHL1 cation
      binding may be compromised.
  - statement: >-
      The same modelling records a hedged possibility that a different aspartate could
      coordinate metal, which is why the GO:0000287 verdict is argued from the annotation's
      derivation rather than from a flat claim that ADPRHL1 binds no metal.
    supporting_text: >-
      Of particular note is the aspartate, D100, that could provide an alternative
      coordination site for metals but whose presence interferes with the phosphate groups
      of forcibly docked ADP.
  - statement: >-
      The source of the sarcomere localisation, obtained with over-expressed N-terminally
      tagged protein.
    supporting_text: Recombinant Adprhl1 can localize to stripes adjacent to the Z-disc
  - statement: >-
      The caveat on that localisation, from the same study - the endogenous protein could
      not be detected in situ, so no endogenous localisation exists in any species.
    supporting_text: >-
      immunocytochemistry of hearts using the same antibody was not sufficiently sensitive
      to detect endogenous Adprhl1 protein in situ
  reference_review:
    relevance: HIGH
    correctness: VERIFIED
    review_notes: >-
      Verified against the cached full text. Fully paginated reference-projection test
      returns 3 annotations over 1 entity (Q6AZR2), so this is one curator reading one paper
      about one gene rather than a bulk import. No retraction, erratum or Crossref update-to.
- id: PMID:32726316
  title: Defective heart chamber growth and myofibrillogenesis after knockout of adprhl1
    gene function by targeted disruption of the ancestral catalytic active site.
  findings:
  - statement: >-
      The genetic demonstration that ADPRHL1's function is cleft-dependent and
      catalysis-independent: CRISPR-induced in-frame deletions of one to four residues from
      the di-arginine loop at the centre of the vestigial ADP-ribosylhydrolase site abolish
      ventricular myofibril assembly.
    supporting_text: >-
      Mutant alleles encode discrete loss of 1, 3 or 4 amino acids from a di-arginine
      (Arg271-Arg272) containing peptide loop at the centre of the ancestral
      ADP-ribosylhydrolase site.
  - statement: The authors' own summary of what that means for a pseudoenzyme.
    supporting_text: >-
      Thus despite lacking catalytic activity, it is the modified (adenosine-ribose)
      substrate binding cleft of Adprhl1 that fulfils an essential role during heart
      formation.
  - statement: >-
      The structural claim that this review's alignment independently reproduces from
      UniProt features alone - ADPRH's substrate-binding Ser269 and Ser270 map onto
      ADPRHL1's Arg271 and Arg272.
    supporting_text: >-
      The critical Adprhl1 deletion covers the exact structural position where in the active
      enzyme Adprh, two adjacent serines that support adenosine-ribose substrate binding are
      located.
  - statement: >-
      The secondary attribution for the biochemical negative, which this review anchors
      independently to PMID:17075046 and PMID:36497109.
    supporting_text: >-
      The sequence changes suggest binding of ADP-ribosylated protein substrates and
      cation-mediated catalysis are both abolished in Adprhl1 and biochemical assays have
      confirmed the lack of any comparable enzymatic activity
  - statement: >-
      The only published mammalian allele, and the reason it does not test the requirement -
      it is not a null.
    supporting_text: >-
      Mice lacking Adprhl1 exons 3-4 are normal but production of the smaller ADPRHL1
      species is unaffected
  - statement: >-
      The published identity figures used as external checks that this review's alignment
      method is sane; measured 46.6% and 74.6% against 46% and 75%.
    supporting_text: the 357 amino acid ADPRH and 354 aa ADPRHL1 share 46% sequence identity
  reference_review:
    relevance: HIGH
    correctness: VERIFIED
    review_notes: >-
      Verified against the cached full text. Its reference-projection test is what exposed
      the split-donor problem: 10 annotations over 2 entities, and both entities are TrEMBL
      accessions for Xenopus adprhl1.L curated by Xenbase, not the Swiss-Prot entry the ISS
      rows transfer from. Crossref reports a has-preprint relation and no update-to; no
      retraction or erratum.
- id: PMID:37880701
  title: The pseudoenzyme ADPRHL1 affects cardiac function by regulating the ROCK pathway.
  findings:
  - statement: >-
      The only functional experiment ever performed on the human protein. CRISPR knockout in
      H9 embryonic stem cell-derived cardiomyocytes; the cells still differentiate and still
      express cardiac genes, but adhere abnormally and show perturbed calcium handling and
      electrophysiology.
    supporting_text: >-
      Our results demonstrated that ADPRHL1-deficient cardiomyocytes exhibited abnormal
      adhesion, calcium transients and electrophysiological activity.
  - statement: >-
      The mechanism and its direction - knockout reduces focal adhesion formation via
      excessive ROCK-myosin II activity, so wild-type ADPRHL1 promotes focal adhesion
      assembly.
    supporting_text: >-
      ADPRHL1 deficiency disrupted the formation of focal adhesions in cardiomyocytes by
      excessively upregulating the ROCK
  - statement: The focal adhesion component measurements underlying that conclusion.
    supporting_text: suggesting that ADPRHL1 deficiency disrupts the FAs in these cells
  - statement: A second, independent focal adhesion marker measured in the same cells.
    supporting_text: >-
      the expression of Paxillin, a structural protein in FAs, in KO-CM was markedly reduced
  - statement: >-
      Independent human support for the cardiac myofibril assembly annotation that GOA
      currently holds only by ISS from frog.
    supporting_text: the arrangement of sarcomeres was less regular
  reference_review:
    relevance: HIGH
    correctness: VERIFIED
    review_notes: >-
      Verified against the cached full text. A fully paginated reference-projection test
      returns zero annotations over zero entities - this paper is entirely absent from GOA,
      which is the largest coverage gap on this gene. No retraction, erratum or Crossref
      update-to.
- id: PMID:17075046
  title: The 39-kDa poly(ADP-ribose) glycohydrolase ARH3 hydrolyzes O-acetyl-ADP-ribose,
    a product of the Sir2 family of acetyl-histone deacetylases.
  findings:
  - statement: >-
      The direct in vitro measurement that recombinant ARH2 has no activity, made alongside
      ARH1 and ARH3 in the same assay. This is the single most decision-relevant fact for
      this gene's molecular function rows, and it is in a paper titled for a paralogue.
    supporting_text: >-
      The rate of O-acetyl-ADP-ribose hydrolysis by recombinant ARH3 was 250-fold that
      observed with ARH1; ARH2 and poly(ADP-ribose) glycohydrolase were inactive.
  - statement: >-
      The mutagenesis that anchors this review's claim that aspartate-to-asparagine at a
      vicinal catalytic aspartate is inactivating rather than conservative - the same
      substitution ADPRHL1 carries at the position corresponding to ADPRH Asp56.
    supporting_text: >-
      hydrolysis of O-acetyl-ADP-ribose was abolished by replacement of the vicinal
      aspartates at positions 77 and 78 of ARH3 with asparagine
  full_text_unavailable: true
  reference_review:
    relevance: HIGH
    correctness: VERIFIED
    review_notes: >-
      Cached abstract-only, so nothing is asserted here beyond what the abstract states, and
      both quoted claims are in the abstract. Found by searching the family rather than the
      gene symbol - no ADPRHL1-keyed query returns a paper titled for ARH3 - and it is absent
      from the affinage record. UniProt cites this PMID as the experimental evidence for the
      mouse ARH2 entry's inactivity. No retraction, erratum or Crossref update-to.
- id: PMID:36497109
  title: ARH Family of ADP-Ribose-Acceptor Hydrolases.
  findings:
  - statement: >-
      The family review's statement of ARH2's enzymatic status, with the specific sequence
      motifs whose loss explains it. The ARH1 motifs quoted, 54-SDDT-57 and 302-DSDS-305,
      correspond exactly to the UniProt-annotated magnesium ligands used as the reference set
      in this review's residue census.
    supporting_text: >-
      ARH2 (so far) does not exhibit any enzymatic activities [70], consistent with
      differences in primary sequences from critical residues found in ARH1 (e.g.,
      54-SDDT-57, 302-DSDS-305)
  - statement: >-
      The statement bearing specifically on the reaction that GO:0003875 and GO:0051725
      name.
    supporting_text: >-
      ARH2 does not appear to be responsible for hydrolysis of ADP-ribose-acceptor linkage.
  - statement: >-
      The authors' own caveat, recorded because it is the reason no replacement hydrolase
      term is proposed and the reason this review does not claim ADPRHL1 is biochemically
      inert.
    supporting_text: >-
      However, the lack of ARH2 catalytic activity may be the result of the use of model
      substrates.
  - statement: >-
      The expression pattern and the biological role attributed to ARH2 by the family's own
      reviewers.
    supporting_text: >-
      ARH2 is a cytoplasmic protein expressed primarily in heart and skeletal muscle and may
      be involved in cardiac myofibril assembly and cardiac chamber outgrowth
  publication_type: REVIEW
  reference_review:
    relevance: HIGH
    correctness: VERIFIED
    review_notes: >-
      A review, used for the family-level statement of enzymatic status and for its explicit
      caveat, not as a primary source for any measurement; the underlying measurement is
      cited separately as PMID:17075046. From the laboratory that characterised all three ARH
      proteins. Absent from the affinage record. No retraction, erratum or Crossref update-to.
- id: PMID:35816343
  title: A Recurrent ADPRHL1 Germline Mutation Activates PARP1 and Confers Prostate
    Cancer Risk in African American Families.
  findings:
  - statement: >-
      A recurrent germline p.Asp78Val allele segregating in African-American prostate cancer
      families; the mutant activates PARP1 and increases the DNA damage response, an effect
      reversed by olaparib.
    supporting_text: the ADPRHL1 mutant activates PARP1
  - statement: >-
      Ectopic wild-type ADPRHL1 restrains proliferation in prostate cancer cells. Suggestive
      of a link to cellular ADP-ribosylation but not evidence that ADPRHL1 itself acts on
      ADP-ribose, so no annotation is proposed from it.
    supporting_text: >-
      expression of the wild-type ADPRHL1 in prostate cancer cells suppressed cell
      proliferation and oncogenesis
  reference_review:
    relevance: MEDIUM
    correctness: VERIFIED
    review_notes: >-
      Correctly cited and internally sound, but deliberately not used for any annotation: it
      is ectopic expression in cancer cell lines, one variant in one ancestry group, and it
      demonstrates no biochemical activity of ADPRHL1. Asp78 is not among the ligand
      positions in the residue census, so it does not bear on the catalytic question either.
      Crossref shows is-supplemented-by and has-preprint relations and no update-to.
- id: PMID:40801020
  title: Inhibition of ARH2 by pH/ROS-responsive nanosystem for improved lung adenocarcinoma
    immunochemotherapy.
  findings:
  - statement: >-
      siRNA knockdown of ARH2 delivered by nanoparticle in lung adenocarcinoma models; ARH2
      is reported to promote M2 macrophage polarisation via FPR2, PI3K and AKT. A different
      tissue and the opposite direction from the prostate work, and rated low confidence by
      the provider that surfaced it.
    supporting_text: >-
      ARH2 promotes M2 macrophage polarization and suppresses immune responses by regulating
      the FPR2/PI3K/AKT signaling pathway
  reference_review:
    relevance: LOW
    correctness: VERIFIED
    review_notes: >-
      Correctly cited and genuinely about ARH2, but not used for any annotation. It is a
      drug-delivery study whose gene-function claims rest on siRNA in tumour models, and its
      direction contradicts the tumour-suppressor reading of PMID:35816343 in a different
      tissue. Recorded so the next reviewer knows it was read and set aside.
- id: PMID:34492228
  title: Histone deacetylase 4 deletion broadly affects cardiac epigenetic repression
    and regulates transcriptional susceptibility via H3K9 methylation.
  findings:
  - statement: >-
      Identifies the Adprhl1 promoter as a target of HDAC4, MEF2 and SUV39H1 repression in
      cardiomyocytes. This concerns regulation of the gene, not a function of the protein,
      and yields no annotation.
  full_text_unavailable: true
  reference_review:
    relevance: LOW
    correctness: VERIFIED
    review_notes: >-
      Cached abstract-only, so nothing is asserted about its full text and no quote is taken
      from it. Included for completeness because affinage cites it; a gene whose promoter is
      a repression target is not thereby annotated with a function.
- id: PMID:16278211
  title: Identification and characterization of a mammalian 39-kDa poly(ADP-ribose)
    glycohydrolase.
  findings:
  - statement: >-
      The reference both Xenopus papers cite for ARH2's lack of comparable enzymatic
      activity. Cached abstract-only, and the abstract does not itself mention ARH2, so no
      claim in this review rests on it - the biochemical negative is anchored to
      PMID:17075046 and PMID:36497109 instead.
  full_text_unavailable: true
  reference_review:
    relevance: MEDIUM
    correctness: UNVERIFIED
    review_notes: >-
      Listed because it is the citation the primary literature uses for ARH2's inactivity,
      but marked unverified rather than verified: only the abstract is cached, the abstract
      is about ARH3, and this review declines to assert what a paper's unavailable full text
      contains. Nothing here depends on it.
- id: PMID:12070318
  title: The family of toxin-related ecto-ADP-ribosyltransferases in humans and the
    mouse.
  findings:
  - statement: >-
      UniProt's reference 1 for the ADPRHL1 isoform-1 mRNA. It surveys the
      ecto-ADP-ribosyltransferase (ARTC) family, which is a different family from the ARH
      hydrolases, and is the most likely origin of the word transferase in Q8NDY3's
      recommended name.
  full_text_unavailable: true
  reference_review:
    relevance: LOW
    correctness: VERIFIED
    review_notes: >-
      Correctly cited by UniProt for a sequence submission only. Noted here solely in support
      of the nomenclature correction request; no functional claim about ARH2 is drawn from
      it, and its full text is not available.
- id: file:human/ADPRHL1/ADPRHL1-deep-research-affinage.md
  title: Affinage mechanistic annotation for ADPRHL1 (human)
  findings:
  - statement: >-
      The provider record used as a starting lead, recorded for provider evaluation rather
      than cited as evidence. Its narrative is broadly accurate and all six of its citations
      were individually verified to concern ADPRHL1 or ARH2 - which is not a given, since on
      the sibling gene ADPRH the same provider returned citations resolving to papers about
      entirely different proteins.
  - statement: >-
      Its recall, however, missed every source that decides this review's molecular function
      verdicts: PMID:17075046 (the only direct measurement that ARH2 is inactive, titled for
      ARH3), PMID:36497109 (the family review that states the negative across the assay
      panel and supplies the model-substrate caveat) and PMID:16278211 (the reference the
      primary literature cites for ARH2's inactivity). All three were found by searching the
      family rather than the gene symbol.
  reference_review:
    relevance: MEDIUM
    correctness: VERIFIED
    review_notes: >-
      gates_passed True, six citations, all numeric PMIDs with no bioRxiv identifiers in a
      PMID-shaped field, and each verified against PubMed to be about ADPRHL1/ARH2. No
      retraction or erratum among them. Used only as a lead: no sentence from it is quoted
      as supporting_text anywhere in this review, per the campaign rule that provider prose
      must not carry a mechanistic claim. The repo validator's warning that no annotation
      references this file is therefore expected and is left standing rather than silenced
      by inventing a citation.
- id: file:human/ADPRHL1/ADPRHL1-uniprot.txt
  title: UniProtKB entry Q8NDY3 (ARHL1_HUMAN), Inactive ADP-ribosyltransferase ARH2
  findings:
  - statement: >-
      The curated record states the catalytic loss twice - once in FUNCTION and once in
      CAUTION - and its feature table contains zero BINDING and zero ACT_SITE features.
      ADPRH (P54922) has 14 such features spanning 20 distinct residue positions and
      ADPRS (Q9NX46) 14 spanning 17; ADPRHL1 has none at all.
    supporting_text: lacks the metal-binding and substrate-binding residues
  - statement: >-
      The FUNCTION line's statement of the measured negative, attributed by similarity to
      the rat and mouse entries.
    supporting_text: showing no activity against O-acetyl-ADP-ribose
  - statement: >-
      The input to the SubCell IEA row, itself ECO:0000250|UniProtKB:Q6AZR2 and therefore
      not independent of the ISS row.
    supporting_text: 'SUBCELLULAR LOCATION: Cytoplasm, myofibril, sarcomere'
  - statement: >-
      Corroborates the QuickGO finding that this gene has no IBA annotation, which is a
      deliberate PAINT judgement rather than a coverage gap.
    supporting_text: 0 GO annotations based on evolutionary models.
  reference_review:
    relevance: HIGH
    correctness: VERIFIED
    review_notes: >-
      All four quotes hand-verified with grep -F against the file in this worktree, each on a
      single physical line so that no UniProt CC continuation is crossed. The repo's
      reference validator does not check file: quotes, so this was done by hand.
- id: file:human/ADPRHL1/ADPRHL1-bioinformatics/RESULTS.md
  title: 'ADPRHL1 (ARH2, Q8NDY3): catalytic-site census against its active paralogues'
  findings:
  - statement: >-
      The identity-matched control that turns the residue observation into an argument.
      Dictyostelium ADPRH, a genuine ARH1 at 48.4% identity - the same distance as ADPRHL1's
      42.6-47.7% - keeps 13 of 20 ligand sites and all six magnesium donors, while every
      ADPRHL1 orthologue keeps 6-7 sites and 2-3 donors.
    supporting_text: Retention therefore is not tracking sequence distance here.
  - statement: >-
      The low-identity active controls pointing the same way from the other side, and the
      reason the result cannot be explained by sequence divergence alone.
    supporting_text: 'Two low-identity active enzymes point the same way: DraG at 27.5% keeps 9, and the'
  - statement: >-
      The independent reproduction of the sibling ADPRH review's panel, split into the
      objective channel (residue and mapped position) and the metric channel (class label).
    supporting_text: identical on all 25 calls
  reference_review:
    relevance: HIGH
    correctness: VERIFIED
    review_notes: >-
      Generated by the committed script catalytic_site_census.py, which fetches every
      sequence and feature live from UniProt, caches them, self-tests, break-tests its own
      reproduction gate, and regenerates this file byte-identically. Its two external checks
      reproduce published identity figures (46.6% versus 46%, 74.6% versus 75%). All quotes
      taken from it were hand-verified with grep -F because the repo validator skips file
      references.