ENPP1 original-source access and evidence guide

The locators below identify the original evidence used to assess ENPP1 annotations. Access to an abstract is distinguished from reading operative Methods, Results and figure captions. Independently established functions can support retaining an annotation whose original citation remains inaccessible; they do not establish what the unread paper measured. No publication quotations are added to the review, and normal publication-cache availability flags are preserved.

Natural products and experimental preparations

The imported human UniProt P22413 record describes a 925-amino-acid type-II membrane protein and a 103–925 secreted chain inferred by similarity to mouse. That processed chain is not a separately documented splice isoform. alternative_products is reserved for UniProt splice-isoform entries and was not invented here. Soluble human plasma PC-1 is independently detected in PMID:11159191, but that experiment does not determine the human cleavage site. Mouse donor P06802 separately annotates membrane homodimer and secreted monomer; the mouse distinction is not assumed to be directly measured for every human preparation.

Recombinant extracellular fragments, domain deletions, catalytic mutants and ENPP1/ENPP2 fusions are experimental constructs. Their kinetic constants and residue coordinates must not automatically be assigned to full-length natural human ENPP1.

Original enzyme and structural studies

• PMID:23027977 — PMC original, publisher DOI. Main Methods, Results and captions were read; the supplement was not recovered. Substrate Specificity and Fig. 1 report purified mouse extracellular ENPP1, residues 92–905, ATP products and ATP/UTP/GTP/CTP kinetics. ATP has the lowest reported Km and highest calculated catalytic efficiency among these substrates; it does not have the largest kcat. Catalytic Domain/Fig. 3 resolves two zinc ions; Nuclease-Like Domain identifies a separate calcium site. Product-bound structures contain AMP, TMP, GMP or CMP. The nuclease-like fold is not an independently demonstrated nuclease assay. Molecular Determinants of Substrate Specificity/Fig. 4 distinguishes wild-type nucleotide specificity from engineered insertion-loop deletion and acquired lipid hydrolysis. The insulin discussion cites earlier work and supplies no new insulin-receptor-binding assay. Mouse coordinates are not remapped onto human ENPP1.
• PMID:25344812 — PMC original, publisher original. Main article, Methods and supplemental assay descriptions/captions were read. Figs. 1–3 combine commercial-reagent cGAMP hydrolysis, native human MDA-MB-231 knockdown, and mouse knockout tissue/plasma assays; Fig. 2 includes ATP hydrolysis. Methods identifies R&D 6136-EN-010. The current official datasheet, revised 2019, describes an ENPP1 catalytic region within ENPP2 flanking domains. It is not the historical 2014 lot certificate. Preparation-specific kinetics must not be described as measurements of the natural full-length human protein; the independent native-cell and knockout results remain positive evidence.
• PMID:25689338 and PMID:26284675 — first publisher correction, second publisher correction. The notices correct Fig. 4a chemical drawings in PMID:25344812. They do not retract its enzyme findings. The original identifier remains the experimental citation; the notices do not establish that every separately hosted manuscript image incorporates the corrections.
• PMID:30356045 — PMC original, publisher original. Main Methods/Results, Figs. 1–3, supplementary Figs. 1–3 and Table 1 were read; supplement page 2 was visually inspected. Structures use engineered mouse extracellular protein, including signal/domain modifications and T238A; crystals contain linear pApG or 3′3′-cGAMP, while intact 2′3′-cGAMP placement is modeled. The THP-1 reporter Methods separately identifies human ENPP1 plasmids. Mouse residue numbering is not assigned to current human P22413. The work supports catalytic mechanism and immune-reporter effects, without resolving a cGAMP transport mechanism.
• PMID:8001561 — publisher-indexed original. Original pages 434 and 436–439 were recovered as text; a complete PDF and its figure pixels were not obtained. Human full-length PC-1 was expressed in COS-7 and mouse Ltk− cells. Methods p. 436 and Table 1 p. 437 use a synthetic phosphonate substrate, which does not by itself establish the oligonucleotide-defined GO:0004528 reaction. Figs. 2–3/pp. 437–438 show ATP labeling and monomer/unreduced dimer bands. Fig. 4/p. 439 establishes human PC-1 surface expression by flow cytometry, with vector control.
• PMID:22285541 — publisher DOI, PubMed record. Verified metadata and abstract only; operative ENPP1 assays, constructs and controls were unavailable. The abstract explicitly reports ENPP1 oligoadenylate nuclease, so this is not dismissed as a PDE12-only or wrong-gene reference. PDE12-specific kinetics do not establish ENPP1-specific cleavage directionality or physiological access.
• PMID:28011303 — publisher DOI, ScienceDirect article, publisher abstract route, PubMed record. The normal cache is bibliography-only and is preserved unchanged. On 2026-10-24 the publisher-indexed Abstract, Highlights and limited section previews were recovered separately; direct article requests returned 403. The accessible original describes biochemical substrate experiments together with human homology modeling and docking, not a modeling-only study. Complete operative Methods, protein preparation and Table 1 were not recovered. The cAMP and human-source UTP judgments remain UNDECIDED; ATP, broad NTP and cGAMP acceptance rests on the independently inspected originals above. No exact kinetics or natural-product assignment is inferred from the preview.
• PMID:27467858 — publisher record. Abstract only despite full-text/supplement attempts. Variant activity and membrane localization are reported, but exact assay substrates and imaging controls were not inspected. Independent human surface evidence supports plasma membrane; the exact oligonucleotide phosphodiesterase-I claim remains unresolved.

Mineralization and pyrophosphate

• PMID:11159191 — PMC original. Main Methods/Results, captions and Discussion were read. Methods and Figs. 2–3 measure PPi, not orthophosphate Pi; Fig. 2 also detects soluble plasma PC-1. Fig. 9 tests human PC-1 expressed in human aortic smooth muscle cells and reports increased NTPPPH and extracellular PPi. Discussion distinguishes PC-1-dependent production from the ANK transport phenotype; ANK was not tested experimentally. These data establish extracellular PPi generation, not directed PPi transport by ENPP1.
• PMID:25644539 — publisher original, accessible author-uploaded original. Main Methods/Results and captions were read; the supplemental appendix was unavailable. Human valve-interstitial-cell NPP1/WGA confocal localization and NPP1 knockdown support membrane location and ATP response/metabolism. ENPP1 supplies ATP-to-AMP chemistry; CD73 produces adenosine, and A2a receptor signaling mediates a cellular response. Mouse A2a-receptor and HEK293T CREB experiments are distinct from the human valve-cell ENPP1 assays.
• PMID:28592560 — PMC original. Main original and captions were read; the downloaded supplement was an HTML challenge and was not treated as a PDF. Fig. 2 measures ENPP1 expression/activity in human mutant fibroblasts; it does not establish ENPP1 executing gene expression. Fig. 7 separates ENPP1 ATP-to-AMP/PPi chemistry from CD73 AMP cleavage and TNAP PPi breakdown. The hypothetical ABCC6 transport discussion does not establish ENPP1 transport. The unseen supplementary enzyme protocol is not used to infer an assay substrate.
• PMID:28652107 — PMC original. Main Methods/Results, Table 2, Fig. 3 and Table 3 were read. Human ENPP1 was expressed in mutant mice under an accelerating diet. Table 2 shows PPi correction without significant Pi differences; Fig. 3/Table 3 show incomplete, sex-dependent skeletal rescue. The study supports regulation of bone mineralization and extracellular PPi, not direct mineral deposition or Pi-specific homeostasis by this experiment. It does not exclude other ENPP1 effects on phosphate physiology.
• PMID:10513816 — publisher DOI, PubMed record. Abstract only; the operative original remained inaccessible after bounded publisher attempts. Its positive NTPPPH/PPi claims are consistent with independent human evidence. The exact transport annotation remains UNDECIDED rather than being rejected from an abstract.
• PMID:12746903 — publisher DOI, PubMed record. Abstract only. The CILP-centered article may include an ENPP1 comparator; its title does not establish miscitation. Both the source-specific dinucleotide phosphatase and precursor-metabolite/energy process remain unresolved. The cGAMP studies do not independently resolve the acid-anhydride reaction class of GO:0004551.
• PMID:28701330 — PMC original, publisher DOI. Operative Results, Fig. 1, Fig. 3 and Discussion were read from the normal full-text cache. Fig. 1 measures uptake of supplied oral PPi in mice and humans; Fig. 3 studies rescue of Enpp1-deficient calcification by supplied PPi. This is not a demonstration that ENPP1 executes the PPi transport step.
• PMID:28704395 — PLOS original, PMC original. Operative Results, Figs. 1–4 and Discussion were read from the normal full-text cache. Npt2a/Enpp1 mouse genetics, urinary PPi and supplied-PPi rescue establish altered PPi availability and calcification. The paper distinguishes ENPP1 generation from ANK transport; renal/extrarenal contributions remain unresolved. Neither the uptake nor genetic experiment identifies ENPP1 as the directed-movement agent.

Interactions, localization and cellular phenotypes

• PMID:11289049 — DOI, accessible author-uploaded original. Operative Methods and Results were read. Human K121/Q121 constructs were tested in human cells and mouse fibroblasts bearing human receptor. Figs. 2–3 show receptor capture/co-immunoprecipitation; Fig. 6 measures insulin-dependent incorporation of labeled glucose into glycogen. These demonstrate association and regulatory effects, not purified binary affinity. Radiolabeled insulin binding concerns insulin binding to its receptor, not insulin binding by ENPP1.
• PMID:17849011 — PMC original, PLOS original. Main Methods/Results and captions were read. Human constructs named K121K/K121Q were expressed in mouse 3T3-L1 cells; a separate arm used mouse-knockout embryonic fibroblasts. Fig. 2C counts cells and viability assays do not find increased death, supporting reduced population proliferation rather than reduced individual-cell size. Figs. 2–3 assess adipocyte maturation, Fig. 4 insulin-receptor/Akt phosphorylation and glucose uptake, and Fig. 5 mouse knockout differentiation. Context-dependent regulation is retained without inventing direct ENPP1–transporter binding or an adipogenesis enzyme reaction.
• PMID:7830796 — publisher original landing page. Abstract/landing page only; the PDF route did not yield operative full text. The abstract reports insulin-receptor inhibition. Later interaction, dimer and ATP studies independently support several assertions, but PAPS metabolism/binding and the exact original assays remain unresolved. No absence or miscitation inference is drawn from the abstract's focus.
• PMID:10615944 — DOI, bounded original preview. Abstract and an original preview ending in Methods were available. Results/captions were not recovered; this is contextual background rather than independent proof of a complete receptor interface.
• PMID:7737162 — PubMed DOI, alternate publisher DOI record. Abstract only. The interpretation of ATP labeling as catalytic intermediates rather than conventional protein kinase activity is contextual; species, construct and residue mapping were not inferred from an unread original.
• PMID:26065921 — PMC original, Nature original. Operative original Results/Methods were read from the normal full-text cache. Results on ABCG2 surface localization, Fig. 6f,g, Supplementary Fig. 9 and interaction Methods combine proximity ligation, co-immunoprecipitation/deletion experiments and binding with affinity-purified proteins. Human ENPP1/ABCG2 interaction is supported; the experiments do not identify ENPP1 as a transporter or general adapter. UniProt Q9UNQ0 was verified as ABCG2.
• PMID:11598187 — PMC original, publisher DOI. Main original Methods, Results and captions were read. Mouse NPP1 and rat NPP3 constructs were tested in canine MDCK cells. Figs. 2–3 show basolateral NPP1 localization; Figs. 6/9 test a sorting motif conserved in human and mouse. Fig. 8 distinguishes wild-type behavior from an engineered truncated sequence that internalizes rapidly. This supports conserved localization with an explicit ortholog boundary, not direct human imaging or proof that ENPP1 performs vesicle transport.
• PMID:17897319 — publisher original. Main human-placental-proteomics Methods/Results were recovered, but the ENPP1-specific Supplementary Table S3 entry was not. The method distinguishes significantly enriched proteins from other detections; the unresolved protein-level result prevents a confident lysosomal-membrane assessment. Lack of recovered supplementation is not evidence against the HDA annotation.
• PMID:28964717 — publisher original, accessible author-uploaded original. Author-original Methods/Results and captions were read after publisher access failed. Fig. 4f–h, pp. 296–297 compares human wild-type/C120R/C164S ENPP1 under reducing/nonreducing conditions in HEK293T cells, supporting homodimerization. Figs. 3/5, pp. 294–299 examine skin pigmentation and already identified patient melanocyte cultures; RNA-seq pathway enrichment and selected qPCR validation support altered pigmentation/state. Whether these establish acquisition of differentiated melanocyte identity remains unresolved. A precursor time-course is not imposed as a universal requirement; the actual data simply do not distinguish that transition from mature-cell dysfunction.
• PMID:15072822, PMID:1315502, PMID:9553761 — biliary-localization publisher, cloning publisher, review publisher. Abstracts only; original images, topology experiments and review source chains were not recovered. Independent human membrane and mouse basolateral evidence are attributed separately. The historical cloning sequence length does not replace the current imported human product.

Database and ontology evidence

Current QuickGO term records were retrieved for all 52 seeded GO terms. GO:0004528 specifies sequential terminal oligonucleotide hydrolysis; a generic synthetic phosphodiesterase substrate does not establish that exact reaction. GO:0004551 has pyrophosphatase/acid-anhydride ancestry, so cGAMP phosphodiester cleavage does not establish this reaction merely by producing two mononucleotides. GO:0030505 requires directed PPi movement. GO:0005044 requires polyanionic-ligand binding and endocytic delivery. GO:0030318 concerns acquisition of specialized melanocyte features. The phosphate-ion homeostasis terms concern phosphate ions, distinct from the directly measured PPi. These definitions constrain the assessed claims without globally excluding untested ENPP1 biology. Current QuickGO also places the substrate-specific GO:0106177 term under acid-anhydride hydrolases; its exact cGAMP definition and Rhea reaction are retained without using that ancestry to infer GO:0004551 activity. The core extracellular location refers to the soluble product, rather than merely the external domain of membrane-bound ENPP1; no claim is made that every natural product was individually assayed against both core substrates.

The pinned PTHR10151 PAINT table records the actual PTN002562515 IBD assertions. Human ENPP1 appearing among descendant sources is legitimate, and donor count is not a weakness. The cache has basolateral location at this node but lacks an exact plasma-membrane IBD row corresponding to the imported plasma-membrane IBA; current full-tree retrieval timed out. No tree-placement error is inferred. Independent human surface experiments establish the location. The PPi-transport concern concerns the experimental process meaning, not fabricated phylogenetic loss or divergence.

Official mouse UniProt P06802 and both QuickGO result pages were retrieved to identify actual donor records. Older mouse phosphodiesterase-I originals cited by the donor were not independently read; their curated assertion is not dismissed. The two PPi donor originals above were examined specifically to distinguish transport from generation.

Official InterPro records for IPR001604, IPR020821 and IPR020436 and the human P22413 match list were retrieved. Domain matches are real; generic domain-level scavenger-receptor/polysaccharide/immune mappings are not independent target assays. Immune-response retention instead uses the cGAMP originals. Nuclease-like architecture does not refute the separately reported ENPP1 oligoadenylate activity.

Official ARBA00027801, ARBA00029247, ARBA00088592 and ARBA00091019 records were retrieved. Their actual matching FunFam conditions were compared with human UniProt cross-references. The ARBA00098322 and ARBA00098486 endpoints returned 404; their conditions were not invented. The corresponding UTP and ATP Rhea equations were independently available.

The official Rhea SPARQL endpoint supplied equations for RHEA:14245 (ATP), 23996 (general NTP), 25277 (cAMP), 27762 (CTP), 29391 (GTP), 29395 (UTP) and 58808 (cGAMP). Correct chemical mappings do not by themselves verify the ENPP1 substrate experiments. Reactome R-HSA-196955 places its ENPP1 catalyst at the plasma membrane; the model's small-molecule compartment labels do not establish cytosolic ENPP1.

Focused follow-up assessment

The PMID:11289049 original, PC-1–IR interaction Methods (p. 832) and Figs. 2–3 (p. 833), identifies the receptor partner sufficiently to refine generic protein binding to insulin receptor binding. Receptor capture and co-immunoprecipitation are the measured association; purified binary affinity is not asserted.

The PMID:28592560 original, patient-fibroblast Results and Fig. 2C–F, measures ENPP1 activity and its own transcript. NT5E transcript is unchanged in ENPP1-mutant cells despite altered CD73 activity. Increased ALPL expression is reported in ABCC6-mutant cells, and Discussion/Fig. 7 explicitly presents a still-untested adenosine/receptor regulatory model. The distinction between an expression readout/regulatory consequence and participation in producing a gene product warrants a curator source query. Because the detailed supplementary methods remain unread, the IDA assertion is UNDECIDED; incomplete experimental coverage does not justify removal. This does not rule out downstream expression regulation.

For IPR020436, the polysaccharide-binding assessment concerns ligand specificity: a broad somatomedin-B mapping does not establish binding of the polysaccharide class. PMID:28964717 Fig. 4f–h and PMID:23027977 SMB-Like Domains establish oligomerization and domain architecture. Scavenger-receptor activity independently requires ligand delivery by endocytosis; that extra requirement is not imposed on polysaccharide binding. Both mappings overshoot the available ENPP1 evidence for different term-specific reasons.

The current QuickGO GO:0009143 ancestor service, inspected 2026-10-24, does not include GO:0006796 in the is_a/part_of ancestors. The two nucleotide-catabolism replacements remain biological refinements without asserting the proposed ancestor chain. The ATP/PPi papers remain supports of the ATP core and the general location discussion; the cGAMP core directly lists its substrate-specific originals and this guide.