ARG1 (Arginase-1, human, UniProtKB:P05089) — review notes

Historical research notes follow. The dated source-scope correction and normal-source closure sections supersede earlier nuclear-carryover, generic-binding, NK-cell, IFN-signaling and manganese-contribution interpretations. The current review has one catalytic core and four unresolved source-specific assertions.

Summary of function

ARG1 is the cytosolic, liver-type (type I) arginase, a binuclear manganese
metalloenzyme that catalyzes the terminal (fifth) step of the urea cycle:
L-arginine + H2O -> L-ornithine + urea (EC 3.5.3.1). It regenerates ornithine to
close the cycle and produces the urea that is excreted. It is a homotrimer, each
subunit binding two Mn(2+) ions. Beyond hepatic ureagenesis, ARG1 is
constitutively expressed in neutrophil granules and released during inflammation,
where local arginine depletion suppresses T-cell (and NK-cell) proliferation /
cytokine production — a myeloid immunoregulatory role. Loss-of-function variants
cause argininemia / arginase deficiency, a urea cycle disorder distinguished by
progressive spastic diplegia/paraparesis rather than the neonatal hyperammonemic
crises typical of proximal UCDs.

Key verbatim citations

Disease (dismech Arginase_Deficiency.yaml)

"progressive spastic diplegia or paraparesis, seizures, intellectual disability,
and growth retardation... relatively infrequent hyperammonemia compared to other
urea cycle disorders." Neurotoxicity from arginine + guanidino compounds.

GO term-definition checks (OLS)

Deep research

falcon deep-research launched (just deep-research-falcon human P05089 --alias ARG1);
FAILED after 600s ("All providers failed" — falcon endpoint timeout). No
-deep-research-falcon.md file produced. Review grounded in the UniProt record, all 9
cached publications, and dismech Arginase_Deficiency.yaml. No DR file fabricated.

2026-09-28 source-scope correction and followup

This section supersedes the earlier nuclear carryover, generic-binding removal,
NK-cell specificity, IFN-signaling interpretation and manganese contributes_to
comments. All 38 original annotation source objects and three product records
are retained. The six cytoplasm/arginine-metabolism umbrellas remain core
annotations because they describe the same cytosolic catabolic activity at
broader resolution. One arginase activity core now connects that reaction to
the urea cycle and L-arginine catabolism; its binuclear manganese requirement is
part of the catalytic description, not a partial contribution to another
molecular function.

The nine canonical PMID abstracts were read in full. Local full-text availability
is true only for 21728378 and 28813417; these records were read in targeted
sections, not treated as proof of complete experimental inspection. Raw UniProt,
GOA and all existing publication/Reactome cache bytes are unchanged. The prior
failed Falcon attempt remains the actual provider history; no provider file was
authored. Three normal Reactome caches remain pending, so this draft is not a
completed source-closure claim.

Human IFN-gamma production rather than response signaling

PMID:16709924 original Blood article,
Results and Figure 6C–D, directly tests human T-cell cytokine output following
arginine depletion by granulocyte material or recombinant human ARG1.
The Figure 6C legend states: “Human PMN arginase suppresses T-cell IFN-γ synthesis by a posttranscriptional mechanism.”
The targeted original Methods/Results read distinguishes secreted protein
measured after 48 hours from transcript measurements. The existing IMP row is
refined to GO:0032689, negative regulation of type II interferon production;
GO:0060336 concerns regulation of the response pathway. These are distinct
endpoints. The original source/evidence/relationship are preserved. The short
quotation is from the publisher body, not the abstract-only normal cache.

Target-specific uncertainties

The antiprotozoal and Th2 Ensembl transfers remain unresolved at the exact donor
experiment/transfer level. General human fungicidal or T-cell evidence does not
close those specific assertions. The independently read mouse macrophage study,
PMID:19360123 provides contextual Th2
corroboration, not a reconstructed Ensembl chain or a direct human assay. Its
official abstract was consulted; no normal cache was requested for that context.

For PMID:28813417(https://pmc.ncbi.nlm.nih.gov/articles/PMC5706633/), the inspected
CMTM6 co-IP/MS Methods and Extended Data 6c point to a specific interaction
inventory. The ARG1 row of that inventory and Supplementary Table 1 have not
been adjudicated. The original ARG1–CMTM6 interaction stays UNDECIDED; neither
generic wording nor a missing body-text match makes it wrong. The prior MISCITED
judgment is withdrawn. For PMID:21630459(https://pubmed.ncbi.nlm.nih.gov/21630459/),
the abstract reports microscopy-based nuclear purity above 99.9%. The ARG1
protein-table/peptide entry remains uninspected, so nuclear localization stays
UNDECIDED without speculation about cytoplasmic carryover or a requirement for
a known nuclear function.

Catalysis and compartment boundaries

PMID:21728378(https://pmc.ncbi.nlm.nih.gov/articles/PMC3150614/) contains human
arginase I inhibitor-binding and structural work, while its inspected enzyme
inhibition Methods use Plasmodium falciparum arginase. The human activity row is
retained using its established biochemical context; the parasite kinetic assay
is not relabeled as human. The title is reconciled to the normal cached Greek
alpha spelling. PMID:17562323(https://pubmed.ncbi.nlm.nih.gov/17562323/) explicitly
reports assayed recombinant human enzyme, and
PMID:3540966(https://pubmed.ncbi.nlm.nih.gov/3540966/) reports activity conferred
by the cloned human cDNA.

The neutrophil fungicidal and T-cell regulatory roles remain secondary to the
core hepatic catabolic role. Phagolysosomal arginine depletion is the proposed
antimicrobial mechanism in PMID:15546957. Its azurophil-granule localization does
not independently establish the separate Reactome specific-granule assertion.
That existing curated compartment is retained with the limited event-summary
scope explicit, rather than supported by an off-target granule quotation.

Official indexed records distinguish normal ARG1 catalysis,
failed ARG1 variant catalysis,
and ARG1 gene expression.
For the last event, cytosol describes the protein output, not a transcriptional
activity of the ARG1 protein. Exact normal-cache recovery and its source check
remain pending. No NEW annotation is introduced by this followup.

2026-09-28 normal-source closure

The three pending normal Reactome records are now present after Source39 recovery.
Their canonical bytes were checked against the exact independently validated archive;
all three complete event summaries and human identity fields were read. This closes
the recovery hold recorded above. R-HSA-70569 explicitly names the cytosolic ARG1
trimer and its arginine-to-ornithine/urea reaction. R-HSA-9956512 describes deficient
variants and distinguishes candidate versus characterized members; R-HSA-9959871
describes TP53 repression of ARG1 expression. Those latter two short summaries do
not reproduce all participant/compartment fields of the earlier inspected official
graphs. Their location reasons retain that distinction and independent human
cytosolic evidence, without assigning normal activity to deficient variants or
transcriptional activity to the expression product.

All 38 decisions are now made, including four explicit unresolved claims whose
source-specific evidence remains uncertain. The three newly read records do not
change any action, product, core or original source assertion. The review status
is COMPLETE; this records completion of the audit, not resolution of every
biological uncertainty. Raw gene records and every prior cache remain unchanged.

2026-09-28 PR3361 evidence-scope clarification

All 38 annotation decisions, three product records and the single arginase core
remain unchanged. The original source fields and published history are preserved.
The user-supplied session AGENTS.md requires UNDECIDED when relevant experimental
evidence cannot be accessed and defines REMOVE as an assertion unlikely to be
correct. This explicit instruction takes priority over the local review skill's
generic-binding removal default. It is not a rule attributed to CLAUDE.md.
Consequently the uninspected ARG1–CMTM6 table entry stays unresolved. No specific
binding activity is invented to replace an unadjudicated co-IP/MS observation.

The six cytoplasm/arginine-metabolism reasons now distinguish PAINT, UniProt,
InterPro, crystallographic EXP and cDNA-era TAS provenance. Independent human
cytosolic or catalytic evidence corroborates these broad assertions without
relabeling it as the original source experiment. The MF refinements identify
known catalytic or manganese specificity; retained broad CC/BP annotations are
umbrella coverage of that same function, not additional activities. Extracellular
myeloid arginase is biologically real and remains a secondary context of the same
chemistry in this compact single-core synthesis.

The two granule-exocytosis Reactome identities and summaries are checked, but
their ARG1-specific participant inventories remain uninspected. Their reference
reviews now explicitly say UNVERIFIED for that remaining scope. The deficient-
variant and gene-expression records retain the separately documented prior
official graph read by the coordinating reviewer; their short recovered summaries
do not themselves display every participant or compartment. In the expression
event ARG1 is the cytosolic protein output, not the transcriptional machinery.

The IFN-gamma-production refinement still depends on the publicly linked original
Blood Results/Figure 6C–D and targeted Methods recorded above. General cytokine
suppression in the immutable abstract cache alone does not establish the specific
endpoint. The exact short figure-legend quotation and body locator remain visible;
no publication cache was edited. The antiprotozoal transfer remains unresolved,
with the donor infection context and possible parasite-versus-host-defense effects
raised as a question rather than asserted from an uninspected experiment.