EGLN1

Summary: EGLN1 regulates the cellular response to low oxygen.

Reason: PMID 16956324 measures HIF1alpha/HRE responses after PHD2 depletion in human cells and HIF-peptide hydroxylation in cell lysates. These are cellular responses to hypoxia, rather than an organism-level response inferred only from expression. Refine to GO:0071456; the activity remains the oxygen-dependent hydroxylation step that controls HIF stability.

Proposed replacements: cellular response to hypoxia

Summary: Rat PHD2 suppresses PDE4D abundance and activity, but the molecular inhibitor claim remains unresolved.

Reason: The literal donor P59722 carries an IMP enzyme-inhibitor assertion from PMID 22975349. Its accessible abstract reports lower PDE4D protein and activity after PHD2 expression, co-immunoprecipitation and hydroxylated PDE4D peptides. It does not resolve whether catalytic inhibition was distinguished from reduced enzyme abundance. GO:0004857 requires reduction of catalytic activity by a molecular function regulator. Preserve the transfer without converting a perturbation endpoint into a direct inhibitor activity; the full operative assay was not recovered.

Summary: The relevant catalytic metal is ferrous iron.

Reason: The IPR006620 domain mapping genuinely supplies broad iron-ion binding. EGLN1 has the conserved H313/D315/H374 metal-binding residues, and target-authenticated PDB 5V18 contains Fe2+. Refine to ferrous iron binding; this is a cofactor feature of the dioxygenase mechanism, not a separate catalytic function.

Proposed replacements: ferrous iron binding

Summary: EGLN1 binds the bHLH transcription factor HIF1A.

Reason: PMID 11598268 assays HPH2 association with the HIF1alpha substrate peptide and its hydroxylation. The historical HPH2 splice construct omits residues 76–177 and is not assigned to a current natural isoform. Official partner records independently establish the bHLH transcription-factor class. GO:0043425 describes binding to that class; it does not require contact with the DNA-binding domain or imply that EGLN1 itself acts as a transcription factor. The original partner and reference remain unchanged.

Proposed replacements: bHLH transcription factor binding

Summary: The OS9 association is retained as a contextual protein interaction.

Reason: PMID 15721254 explicitly reports OS9 association with PHD proteins/HIF1alpha. OS9 is an ER quality-control lectin, but that does not establish a protein-folding chaperone activity or a more informative verified binding-class term here. Retain the positive association under the project-specific generic-binding policy; do not assign EGLN1 a chaperone or ER core function from the interaction.

Summary: EGLN1 binds the bHLH transcription factor HIF1A.

Reason: PMID 16511565 reports PHD2–HIF1alpha co-immunoprecipitation in the MTA1/HDAC study, including the MG132-stabilized interaction and its reduction by MTA1. This target association is not inferred merely from the paper title. Official partner records independently establish the bHLH transcription-factor class. GO:0043425 describes binding to that class; it does not require contact with the DNA-binding domain or imply that EGLN1 itself acts as a transcription factor. The original partner and reference remain unchanged.

Proposed replacements: bHLH transcription factor binding

Summary: EGLN1 binds the peptidyl-prolyl isomerase FKBP8.

Reason: PMID 17353276 supports the FKBP8/FKBP38–PHD2 pair by two-hybrid, pull-down and co-immunoprecipitation experiments. The interaction uses FKBP8 residues 3–97 and PHD2 residues 1–169, outside the FKBP8 catalytic domain. Official FKBP8 annotation independently establishes peptidyl-prolyl isomerase activity. GO:0019899 requires binding to an enzyme, not binding its catalytic site, so enzyme binding is justified without assigning that enzyme activity to EGLN1.

Proposed replacements: enzyme binding

Summary: EGLN1 binds the bHLH transcription factor HIF1A.

Reason: PMID 17353276 maps PHD2/HIF interactions alongside the FKBP8/FKBP38 N-terminal interaction. The substrate association is preserved independently of the FKBP8 mechanism. Official partner records independently establish the bHLH transcription-factor class. GO:0043425 describes binding to that class; it does not require contact with the DNA-binding domain or imply that EGLN1 itself acts as a transcription factor. The original partner and reference remain unchanged.

Proposed replacements: bHLH transcription factor binding

Summary: EGLN1 binds the bHLH transcription factor EPAS1.

Reason: PMID 17353276 tests PHD2 catalytic-region constructs with the HIF2alpha oxygen-dependent degradation region, including the404–569 fragment. Official partner records independently establish the bHLH transcription-factor class. GO:0043425 describes binding to that class; it does not require contact with the DNA-binding domain or imply that EGLN1 itself acts as a transcription factor. The original partner and reference remain unchanged.

Proposed replacements: bHLH transcription factor binding

Summary: EGLN1 binds the bHLH transcription factor EPAS1.

Reason: PMID 19208626 uses recombinant PHD2 and HIF2alpha oxygen-dependent degradation fragments/peptides. Wild-type substrate binding and hydroxylation are compared with erythrocytosis-associated substitutions; the mutant effects do not negate the wild-type interaction. Official partner records independently establish the bHLH transcription-factor class. GO:0043425 describes binding to that class; it does not require contact with the DNA-binding domain or imply that EGLN1 itself acts as a transcription factor. The original partner and reference remain unchanged.

Proposed replacements: bHLH transcription factor binding

Summary: EGLN1 binds the bHLH transcription factor HIF1A.

Reason: PMID 20840591 studies recombinant human PHD2 residues 181–426 with HIF1alpha CODD556–574 using substrate binding and hydroxylation assays. Official partner records independently establish the bHLH transcription-factor class. GO:0043425 describes binding to that class; it does not require contact with the DNA-binding domain or imply that EGLN1 itself acts as a transcription factor. The original partner and reference remain unchanged.

Proposed replacements: bHLH transcription factor binding

Summary: EGLN1 binds the signaling scaffold LIMD1.

Reason: PMID 22286099 explicitly identifies LIMD1 as a scaffold that simultaneously binds PHD proteins and VHL. The literal PHD2–LIMD1 interaction and independently verified scaffold role support GO:0097110. This class-specific binding refinement does not make EGLN1 the scaffold or assert that every natural product enters the same complex.

Proposed replacements: scaffold protein binding

Summary: The processed HCV NS5A association is retained without transferring polyprotein functions.

Reason: PMID 24169621 identifies EGLN1 in an NS5A affinity-purification screen and tests reciprocal association in HEK293T/Huh7.5 contexts. The source participant Q9WMX2-PRO_0000037551 is processed NS5A, not the entire viral polyprotein. No informative verified binding-class replacement is established for this particular nonenzymatic participant; retain the association as non-core and do not transfer protease/polymerase activities from other polyprotein segments.

Summary: EGLN1 binds the bHLH transcription factor HIF1A.

Reason: The PMID 24681946 abstract explicitly reports increased HIF1alpha–PHD2 association following SIRT2-mediated HIF1alpha deacetylation. The literal IntAct target pair corroborates the named association; operative full-text assay details were not recovered. Official partner records independently establish the bHLH transcription-factor class. GO:0043425 describes binding to that class; it does not require contact with the DNA-binding domain or imply that EGLN1 itself acts as a transcription factor. The original partner and reference remain unchanged.

Proposed replacements: bHLH transcription factor binding

Summary: EGLN1 associates with the transcription corepressor CBFA2T3/MTG16.

Reason: PMID 25974097 detects PHD2 with MTG16 in human Raji-cell immunoprecipitates. CBFA2T3 is independently established as a transcription corepressor, making GO:0001222 an informative partner-class refinement. The assay establishes complex association rather than a purified direct interface; that limit does not require discarding the reported binding class or assigning corepressor activity to EGLN1.

Proposed replacements: transcription corepressor binding

Summary: EGLN1 binds the bHLH transcription factor HIF1A.

Reason: PMID 25974097 reports HIF1alpha/MTG16/PHD2 co-association in human Raji cells under controlled oxygen conditions. This is an association within an immunoprecipitated complex, not proof of an isolated binary interface. Official partner records independently establish the bHLH transcription-factor class. GO:0043425 describes binding to that class; it does not require contact with the DNA-binding domain or imply that EGLN1 itself acts as a transcription factor. The original partner and reference remain unchanged.

Proposed replacements: bHLH transcription factor binding

Summary: EGLN1 binds the bHLH transcription factor EPAS1.

Reason: PMID 30487161 compares wild-type human EPAS1 and variants in a mammalian two-hybrid assay with PHD2; the empty-vector control and positive wild-type signal support the association. The AC16 host and DFO treatment are kept distinct from construct identity and physiological hypoxia. Official partner records independently establish the bHLH transcription-factor class. GO:0043425 describes binding to that class; it does not require contact with the DNA-binding domain or imply that EGLN1 itself acts as a transcription factor. The original partner and reference remain unchanged.

Proposed replacements: bHLH transcription factor binding

Summary: The GRID2IP interaction records correspond to below-threshold PDZ measurements.

Reason: The literal PMID 36115835 IntAct records name GRID2IP, and the original table resolves the corresponding tested domains (GRID2IP_2, GRID2IP_1). All recovered EGLN1 measurements for this partner are below the stated binding-intensity threshold, with pKd encoded as 0 rather than a quantified affinity. The paper explicitly warns that sub-threshold results do not prove noninteraction. IntAct negative=false therefore cannot by itself establish a positive result for this source-specific pair. Preserve the imported assertion as unresolved, without declaring the partner nonbinding or transferring a positive from a different PDZ protein.

Summary: The MPDZ interaction records correspond to below-threshold PDZ measurements.

Reason: The literal PMID 36115835 IntAct records name MPDZ, and the original table resolves the corresponding tested domains (MPDZ_12, MPDZ_5, MPDZ_10). All recovered EGLN1 measurements for this partner are below the stated binding-intensity threshold, with pKd encoded as 0 rather than a quantified affinity. The paper explicitly warns that sub-threshold results do not prove noninteraction. IntAct negative=false therefore cannot by itself establish a positive result for this source-specific pair. Preserve the imported assertion as unresolved, without declaring the partner nonbinding or transferring a positive from a different PDZ protein.

Summary: The DLG4 interaction records correspond to below-threshold PDZ measurements.

Reason: The literal PMID 36115835 IntAct records name DLG4, and the original table resolves the corresponding tested domains (DLG4_1, DLG4_3, DLG4_2). All recovered EGLN1 measurements for this partner are below the stated binding-intensity threshold, with pKd encoded as 0 rather than a quantified affinity. The paper explicitly warns that sub-threshold results do not prove noninteraction. IntAct negative=false therefore cannot by itself establish a positive result for this source-specific pair. Preserve the imported assertion as unresolved, without declaring the partner nonbinding or transferring a positive from a different PDZ protein.

Summary: The APBA1 interaction records correspond to below-threshold PDZ measurements.

Reason: The literal PMID 36115835 IntAct records name APBA1, and the original table resolves the corresponding tested domains (APBA1_1, APBA1_2). All recovered EGLN1 measurements for this partner are below the stated binding-intensity threshold, with pKd encoded as 0 rather than a quantified affinity. The paper explicitly warns that sub-threshold results do not prove noninteraction. IntAct negative=false therefore cannot by itself establish a positive result for this source-specific pair. Preserve the imported assertion as unresolved, without declaring the partner nonbinding or transferring a positive from a different PDZ protein.

Summary: The TJP1 interaction records correspond to below-threshold PDZ measurements.

Reason: The literal PMID 36115835 IntAct records name TJP1, and the original table resolves the corresponding tested domains (TJP1_3, TJP1_2). All recovered EGLN1 measurements for this partner are below the stated binding-intensity threshold, with pKd encoded as 0 rather than a quantified affinity. The paper explicitly warns that sub-threshold results do not prove noninteraction. IntAct negative=false therefore cannot by itself establish a positive result for this source-specific pair. Preserve the imported assertion as unresolved, without declaring the partner nonbinding or transferring a positive from a different PDZ protein.

Summary: The DLG1 interaction records correspond to below-threshold PDZ measurements.

Reason: The literal PMID 36115835 IntAct records name DLG1, and the original table resolves the corresponding tested domains (DLG1_1, DLG1_3, DLG1_2). All recovered EGLN1 measurements for this partner are below the stated binding-intensity threshold, with pKd encoded as 0 rather than a quantified affinity. The paper explicitly warns that sub-threshold results do not prove noninteraction. IntAct negative=false therefore cannot by itself establish a positive result for this source-specific pair. Preserve the imported assertion as unresolved, without declaring the partner nonbinding or transferring a positive from a different PDZ protein.

Summary: The EGLN1 C-terminal peptide binds a PDZ domain of IL16.

Reason: PMID 36115835 Supplementary Data 1, PDZ-PBM interactome row 15085, reports IL16_4: normalized binding intensity 0.07007436 and estimated pKd 3.62865447. The normalized binding intensity exceeds the study threshold of 0.05; pKd is an estimated affinity value, not the quantity compared with that threshold. The tested EGLN1 peptide corresponds exactly to canonical residues 417–426; the assay uses isolated human PDZ domains, not a native full-length complex. The current IntAct records cover IL16 PDZ1/PDZ3, which are censored in the source table; the positive support instead comes from the original IL16_4 measurement and its human IL16 domain mapping. GO:0030165 captures the measured domain-binding activity. No experimental insert-sequence or natural-isoform equivalence is asserted.

Proposed replacements: PDZ domain binding

Summary: The SCRIB interaction records correspond to below-threshold PDZ measurements.

Reason: The literal PMID 36115835 IntAct records name SCRIB, and the original table resolves the corresponding tested domains (SCRIB_2, SCRIB_1, SCRIB_3). All recovered EGLN1 measurements for this partner are below the stated binding-intensity threshold, with pKd encoded as 0 rather than a quantified affinity. The paper explicitly warns that sub-threshold results do not prove noninteraction. IntAct negative=false therefore cannot by itself establish a positive result for this source-specific pair. Preserve the imported assertion as unresolved, without declaring the partner nonbinding or transferring a positive from a different PDZ protein.

Summary: The PDZK1 interaction records correspond to below-threshold PDZ measurements.

Reason: The literal PMID 36115835 IntAct records name PDZK1, and the original table resolves the corresponding tested domains (NHERF3_4, NHERF3_3, NHERF3_1). All recovered EGLN1 measurements for this partner are below the stated binding-intensity threshold, with pKd encoded as 0 rather than a quantified affinity. The paper explicitly warns that sub-threshold results do not prove noninteraction. IntAct negative=false therefore cannot by itself establish a positive result for this source-specific pair. Preserve the imported assertion as unresolved, without declaring the partner nonbinding or transferring a positive from a different PDZ protein.

Summary: The FRMPD2 interaction records correspond to below-threshold PDZ measurements.

Reason: The literal PMID 36115835 IntAct records name FRMPD2, and the original table resolves the corresponding tested domains (FRMPD2_1, FRMPD2_2, FRMPD2_3). All recovered EGLN1 measurements for this partner are below the stated binding-intensity threshold, with pKd encoded as 0 rather than a quantified affinity. The paper explicitly warns that sub-threshold results do not prove noninteraction. IntAct negative=false therefore cannot by itself establish a positive result for this source-specific pair. Preserve the imported assertion as unresolved, without declaring the partner nonbinding or transferring a positive from a different PDZ protein.

Summary: The MAGI2 interaction records correspond to below-threshold PDZ measurements.

Reason: The literal PMID 36115835 IntAct records name MAGI2, and the original table resolves the corresponding tested domains (MAGI2_4, MAGI2_5, MAGI2_6, MAGI2_2, MAGI2_3). All recovered EGLN1 measurements for this partner are below the stated binding-intensity threshold, with pKd encoded as 0 rather than a quantified affinity. The paper explicitly warns that sub-threshold results do not prove noninteraction. IntAct negative=false therefore cannot by itself establish a positive result for this source-specific pair. Preserve the imported assertion as unresolved, without declaring the partner nonbinding or transferring a positive from a different PDZ protein.

Summary: The NHERF4 interaction records correspond to below-threshold PDZ measurements.

Reason: The literal PMID 36115835 IntAct records name NHERF4, and the original table resolves the corresponding tested domains (NHERF4_2, NHERF4_4, NHERF4_1). All recovered EGLN1 measurements for this partner are below the stated binding-intensity threshold, with pKd encoded as 0 rather than a quantified affinity. The paper explicitly warns that sub-threshold results do not prove noninteraction. IntAct negative=false therefore cannot by itself establish a positive result for this source-specific pair. Preserve the imported assertion as unresolved, without declaring the partner nonbinding or transferring a positive from a different PDZ protein.

Summary: The LNX2 interaction records correspond to below-threshold PDZ measurements.

Reason: The literal PMID 36115835 IntAct records name LNX2, and the original table resolves the corresponding tested domains (LNX2_3, LNX2_2). All recovered EGLN1 measurements for this partner are below the stated binding-intensity threshold, with pKd encoded as 0 rather than a quantified affinity. The paper explicitly warns that sub-threshold results do not prove noninteraction. IntAct negative=false therefore cannot by itself establish a positive result for this source-specific pair. Preserve the imported assertion as unresolved, without declaring the partner nonbinding or transferring a positive from a different PDZ protein.

Summary: The PATJ interaction records correspond to below-threshold PDZ measurements.

Reason: The literal PMID 36115835 IntAct records name PATJ, and the original table resolves the corresponding tested domains (Inadl_6, Inadl_4, Inadl_10, Inadl_1, Inadl_9, Inadl_2, Inadl_5, Inadl_8, Inadl_3). All recovered EGLN1 measurements for this partner are below the stated binding-intensity threshold, with pKd encoded as 0 rather than a quantified affinity. The paper explicitly warns that sub-threshold results do not prove noninteraction. IntAct negative=false therefore cannot by itself establish a positive result for this source-specific pair. Preserve the imported assertion as unresolved, without declaring the partner nonbinding or transferring a positive from a different PDZ protein.

Summary: The LNX1 interaction records correspond to below-threshold PDZ measurements.

Reason: The literal PMID 36115835 IntAct records name LNX1, and the original table resolves the corresponding tested domains (LNX1_2, LNX1_3, LNX1_1). All recovered EGLN1 measurements for this partner are below the stated binding-intensity threshold, with pKd encoded as 0 rather than a quantified affinity. The paper explicitly warns that sub-threshold results do not prove noninteraction. IntAct negative=false therefore cannot by itself establish a positive result for this source-specific pair. Preserve the imported assertion as unresolved, without declaring the partner nonbinding or transferring a positive from a different PDZ protein.

Summary: The EGLN1 C-terminal peptide binds a PDZ domain of DLG3.

Reason: PMID 36115835 Supplementary Data 1, PDZ-PBM interactome row 8521, reports DLG3_2: normalized binding intensity 0.17058929 and estimated pKd 4.07469335. The normalized binding intensity exceeds the study threshold of 0.05; pKd is an estimated affinity value, not the quantity compared with that threshold. The tested EGLN1 peptide corresponds exactly to canonical residues 417–426; the assay uses isolated human PDZ domains, not a native full-length complex. The positive DLG3_2 domain also matches the literal IntAct fragment encompassing the official PDZ2 feature. GO:0030165 captures the measured domain-binding activity. No experimental insert-sequence or natural-isoform equivalence is asserted.

Proposed replacements: PDZ domain binding

Summary: The MAGI1 interaction records correspond to below-threshold PDZ measurements.

Reason: The literal PMID 36115835 IntAct records name MAGI1, and the original table resolves the corresponding tested domains (MAGI1_1, MAGI1_2). All recovered EGLN1 measurements for this partner are below the stated binding-intensity threshold, with pKd encoded as 0 rather than a quantified affinity. The paper explicitly warns that sub-threshold results do not prove noninteraction. IntAct negative=false therefore cannot by itself establish a positive result for this source-specific pair. Preserve the imported assertion as unresolved, without declaring the partner nonbinding or transferring a positive from a different PDZ protein.

Summary: The APBA2 interaction records correspond to below-threshold PDZ measurements.

Reason: The literal PMID 36115835 IntAct records name APBA2, and the original table resolves the corresponding tested domains (APBA2_1, APBA2_2). All recovered EGLN1 measurements for this partner are below the stated binding-intensity threshold, with pKd encoded as 0 rather than a quantified affinity. The paper explicitly warns that sub-threshold results do not prove noninteraction. IntAct negative=false therefore cannot by itself establish a positive result for this source-specific pair. Preserve the imported assertion as unresolved, without declaring the partner nonbinding or transferring a positive from a different PDZ protein.

Summary: The GRIP2 interaction records correspond to below-threshold PDZ measurements.

Reason: The literal PMID 36115835 IntAct records name GRIP2, and the original table resolves the corresponding tested domains (GRIP2_4, GRIP2_5, GRIP2_7, GRIP2_3, GRIP2_6). All recovered EGLN1 measurements for this partner are below the stated binding-intensity threshold, with pKd encoded as 0 rather than a quantified affinity. The paper explicitly warns that sub-threshold results do not prove noninteraction. IntAct negative=false therefore cannot by itself establish a positive result for this source-specific pair. Preserve the imported assertion as unresolved, without declaring the partner nonbinding or transferring a positive from a different PDZ protein.

Summary: The EGLN1 C-terminal peptide binds a PDZ domain of WHRN.

Reason: PMID 36115835 Supplementary Data 1, PDZ-PBM interactome row 17873, reports DFNB31_1: normalized binding intensity 0.05223779 and estimated pKd 3.49108388. The normalized binding intensity exceeds the study threshold of 0.05; pKd is an estimated affinity value, not the quantity compared with that threshold. The tested EGLN1 peptide corresponds exactly to canonical residues 417–426; the assay uses isolated human PDZ domains, not a native full-length complex. The current IntAct records cover WHRN PDZ2/PDZ3, which are censored in the source table; positive support instead comes from the original DFNB31_1 measurement. DFNB31 is the verified synonym of human WHRN. GO:0030165 captures the measured domain-binding activity. No experimental insert-sequence or natural-isoform equivalence is asserted.

Proposed replacements: PDZ domain binding

Summary: The GRIP1 interaction records correspond to below-threshold PDZ measurements.

Reason: The literal PMID 36115835 IntAct records name GRIP1, and the original table resolves the corresponding tested domains (GRIP1_7, GRIP1_3, GRIP1_5, GRIP1_4, GRIP1_2). All recovered EGLN1 measurements for this partner are below the stated binding-intensity threshold, with pKd encoded as 0 rather than a quantified affinity. The paper explicitly warns that sub-threshold results do not prove noninteraction. IntAct negative=false therefore cannot by itself establish a positive result for this source-specific pair. Preserve the imported assertion as unresolved, without declaring the partner nonbinding or transferring a positive from a different PDZ protein.

Summary: EGLN1 can occupy the nucleus as well as the cytoplasm.

Reason: The inspected PMID 12615973 original uses human PHD2 GFP/PK-tagged constructs in U2OS cells. PHD2 is predominantly cytoplasmic, with a nuclear fraction; this supports nonexclusive nuclear localization. Other PHD paralogs and PHD3 aggregation observations are not substituted for the target. The study does not resolve all current natural products.

Summary: PHD2 undergoes nucleocytoplasmic shuttling.

Reason: The complete PMID 19339211 abstract explicitly reports leptomycin-B-sensitive PHD2 shuttling and an N-terminal export region. It supports the nuclear pool stated in the original experimental annotation. Full source-specific images and construct sequence were not recovered, so no finer nuclear subcompartment or isoform claim is added.

Summary: The nuclear PHD2 pool is supported by transport experiments.

Reason: The complete PMID 19631610 abstract explicitly describes PHD2 cycling between nucleus and cytoplasm, with CRM1-dependent export and an N-terminal 100-residue requirement. Retain the experimental nuclear annotation without borrowing PHD1-specific classical import mechanisms or claiming inspection of unavailable image panels.

Summary: The PAINT nuclear-localization inference is compatible with direct target evidence.

Reason: The target PTN002490115 lies beneath PTN004393219 in the genuine PTHR12907 tree; the literal IBD places nucleus at that ancestor. Direct target studies12615973/19339211/19631610 support a nuclear pool. The target appearing among descendant evidence is legitimate, not circular, and no target-specific loss of this localization was identified.

Summary: The curated UniProt nuclear location agrees with target observations.

Reason: The SL-0191-to-nucleus mapping is the appropriate term for the official target nuclear location. It is supported by PHD2 localization/shuttling studies and coexists with the cytosolic pool; it does not imply exclusive nuclear residence.

Summary: The broad cytoplasm annotation is retained alongside finer cytosolic evidence.

Reason: PMID 19339211 explicitly reports PHD2 shuttling between cytoplasm and nucleus, but abstract-only access does not establish soluble cytosol in every source-specific construct. The independent HPA cytosol result and modeled Reactome target localization provide finer positive evidence elsewhere. Retaining this broad source-specific assertion avoids rewriting its measurement resolution; cytoplasm is not a separate core compartment in addition to cytosol and nucleus.

Summary: The broad cytoplasm annotation is retained alongside finer cytosolic evidence.

Reason: PMID 19631610 explicitly reports cytoplasm/nucleus cycling and CRM1-dependent nuclear export; the available abstract does not establish an organelle-excluding biochemical fraction. The independent HPA cytosol result and modeled Reactome target localization provide finer positive evidence elsewhere. Retaining this broad source-specific assertion avoids rewriting its measurement resolution; cytoplasm is not a separate core compartment in addition to cytosol and nucleus.

Summary: The broad cytoplasm annotation is retained alongside finer cytosolic evidence.

Reason: The PTHR12907 target descends from PTN004393219, whose literal IBD contains cytoplasm, including target and rat Egln3 descendant evidence. The independent HPA cytosol result and modeled Reactome target localization provide finer positive evidence elsewhere. Retaining this broad source-specific assertion avoids rewriting its measurement resolution; cytoplasm is not a separate core compartment in addition to cytosol and nucleus.

Summary: The broad cytoplasm annotation is retained alongside finer cytosolic evidence.

Reason: The inspected PMID 12615973 original shows predominantly cytoplasmic human PHD2 with a smaller nuclear pool in U2OS cells; it does not establish a soluble fraction by biochemical fractionation. The independent HPA cytosol result and modeled Reactome target localization provide finer positive evidence elsewhere. Retaining this broad source-specific assertion avoids rewriting its measurement resolution; cytoplasm is not a separate core compartment in addition to cytosol and nucleus.

Summary: The broad cytoplasm annotation is retained alongside finer cytosolic evidence.

Reason: The official target cytoplasm statement maps through UniProt subcellular location SL-0086. The independent HPA cytosol result and modeled Reactome target localization provide finer positive evidence elsewhere. Retaining this broad source-specific assertion avoids rewriting its measurement resolution; cytoplasm is not a separate core compartment in addition to cytosol and nucleus.

Summary: HPA supports EGLN1 localization in the cytosol.

Reason: The exact ENSG00000135766/Q9GZT9 HPA target page labels cytosol Supported with antibody HPA022129 in A-431, U-251MG and U-2OS contexts. The separate vesicle observation is Uncertain and is not promoted to a core location. Location-specific reliability, rather than the aggregate gene score, was inspected; no image-pixel assessment is claimed.

Summary: Cytosol is an established EGLN1 location; retain the historical TAS annotation with its source-access caveat.

Reason: Independent human HPA observations support cytosolic EGLN1, consistent with the five Reactome cytosol annotations. Accept the biologically established location and defer to the curator for the historical TAS attribution. The PMID 8606489 abstract reports rat and human smooth-muscle-cell protein detection and cytoplasmic filament localization, with substantial rat tissue analysis. The complete original was not recovered, so its precise human soluble-cytosol evidence remains unverified. Independent modern observations support the gene-level claim without reconstructing that historical assay or alleging a wrong species or gene.

Summary: The Reactome model places the EGLN1-containing entity in cytosol.

Reason: The actual R-HSA-1234159 object has a literal component/set path to EGLN1 entity 1234108, with reference accessionQ9GZT9 and cytosol GO:0005829. The target leaf spans residues 2–426. This is model-supported localization, not a new microscopy experiment; membership in a VHL/proteasome-associated event does not give EGLN1 ubiquitin-ligase or proteolytic activity.

Summary: The Reactome model places the EGLN1-containing entity in cytosol.

Reason: The actual R-HSA-1234163 object has a literal component/set path to EGLN1 entity 1234108, with reference accessionQ9GZT9 and cytosol GO:0005829. The target leaf spans residues 2–426. This is model-supported localization, not a new microscopy experiment; membership in a VHL/proteasome-associated event does not give EGLN1 ubiquitin-ligase or proteolytic activity.

Summary: The Reactome model places the EGLN1-containing entity in cytosol.

Reason: The actual R-HSA-1234173 object has a literal component/set path to EGLN1 entity 1234108, with reference accessionQ9GZT9 and cytosol GO:0005829. The target leaf spans residues 2–426. This is model-supported localization, not a new microscopy experiment; membership in a VHL/proteasome-associated event does not give EGLN1 ubiquitin-ligase or proteolytic activity.

Summary: The Reactome model places the EGLN1-containing entity in cytosol.

Reason: The actual R-HSA-1234177 object has a literal component/set path to EGLN1 entity 1234108, with reference accessionQ9GZT9 and cytosol GO:0005829. The target leaf spans residues 2–426. This is model-supported localization, not a new microscopy experiment; membership in a VHL/proteasome-associated event does not give EGLN1 ubiquitin-ligase or proteolytic activity.

Summary: The Reactome model places the EGLN1-containing entity in cytosol.

Reason: The actual R-HSA-1234183 object has a literal component/set path to EGLN1 entity 1234108, with reference accessionQ9GZT9 and cytosol GO:0005829. The target leaf spans residues 2–426. This is model-supported localization, not a new microscopy experiment; membership in a VHL/proteasome-associated event does not give EGLN1 ubiquitin-ligase or proteolytic activity.

Summary: PHD2 signaling influences intracellular iron balance in a neuronal stress context.

Reason: The verified mouse Egln1 donor record cites PMID 26818499. Its original measures cytosolic iron in dopaminergic synaptosomes after PHD2 knockdown and shows ATP13A2-dependent iron handling and protection in differentiated human neuronal cells after pharmacological PHD2 inhibition. Retain the conserved contextual process without claiming that EGLN1 itself transports iron or that catalytic Fe2+ binding alone proves iron homeostasis.

Summary: Ferrous iron binding is conserved in the EGLN catalytic lineage.

Reason: The genuine PTN004393218 IBD and target descent support Fe2+ binding. The saved MSA retains the target H313/D315/H374 metal-ligand residues across the inspected family representatives; target-authenticated PDB 5V18 provides direct corroboration. Self-donor evidence is expected and is not treated as circular.

Summary: The deposited human PHD2 structure contains ferrous iron.

Reason: Although PMID 28594552 foregrounds PHD1 in its title, its deposited PDB 5V18 includes Q9GZT9: entity 1 is GGS plus canonical EGLN1 residues 181–416, and nonpolymer entity 4 is FE2. This directly authenticates a target construct and metal identity. The paper is not rejected from its title; no independent electron-density or full-original-body inspection is claimed.

Summary: Postsynaptic-density localization is supported in the mouse neuronal donor context.

Reason: PMID 26972007 detects PHD2 in mouse brain synaptosomal/PSD fractions and in neuronal protrusions using a specificity-checked antibody. The Q91YE3 donor join is authentic. Retain the human ortholog inference as a contextual neuronal pool, without presenting mouse fractions as a direct endogenous human-neuron assay or moving it into the oxygen-sensing core.

Summary: The target-specific dioxygenase reaction is HIF-proline hydroxylation.

Reason: IPR006620 genuinely maps to the broad paired-donor oxidoreductase term. Human target biochemical evidence identifies the more informative GO:0160082 HIF-proline dioxygenase reaction. This refinement does not assert that HIF is the only possible substrate or import collagen specificity from the family name.

Proposed replacements: hypoxia-inducible factor-proline dioxygenase activity

Summary: EGLN1 is a 2-oxoglutarate-dependent dioxygenase.

Reason: Human EGLN1 biochemical experiments establish the Fe(II)/2-oxoglutarate-dependent dioxygenase reaction class as well as its HIF substrate specificity. Peptide hydroxylation and 2-oxoglutarate turnover are separately measured in the cited biochemical studies; uncoupled cosubstrate turnover alone is not evidence of peptide hydroxylation. Retain this experimentally attributed reaction-class annotation, while the more specific GO:0160082 term represents the single core function; this is not an additional core activity or a newly proposed ancestor annotation. The PMID 22955912 abstract discusses P4H-TM and also HIF-P4H-2 gene-modified mice; its title is not evidence of misattribution. Its operative source assay was not recovered, so the independent biochemical evidence supports the class without a claim to have inspected a human assay in that paper.

Summary: The rat PHD2–PDE4D association supports contextual enzyme binding.

Reason: The literal P59722 donor IPI records cite PMID 22975349 and PDE4D/P14270. The complete abstract explicitly reports PHD2–PDE4D co-immunoprecipitation, with PDE4D independently identified as a phosphodiesterase. This supports enzyme binding while the separate inhibitor mechanism remains unresolved. No narrower established phosphodiesterase-binding term was recovered in the bounded ontology search; no direct contact or all-product equivalence is claimed.

Summary: Ascorbate association is retained as a biochemical cofactor feature.

Reason: The current IPR006620 entry genuinely supplies L-ascorbic-acid binding. The human enzyme assays use ascorbate in the Fe2+/2-oxoglutarate reaction system, and the exact Reactome EGLN1 complex explicitly includes ascorbate. Retain the source claim as non-core; cofactor dependence and model membership are not described as a newly measured binding affinity, and no separate ascorbate-binding core is created.

Summary: The cited PHD2 mechanism is HIF-proline hydroxylation.

Reason: PMID 16956324 measures HIF-substrate hydroxylation in human-cell lysates and HIF signaling after PHD2 depletion. Refine the broad peptidyl-proline dioxygenase description to GO:0160082, corroborated by direct human enzyme work. This does not assign the IOP1 protein or other components the PHD2 chemistry.

Proposed replacements: hypoxia-inducible factor-proline dioxygenase activity

Summary: The conserved prolyl 4-hydroxylase activity has a specific HIF substrate term.

Reason: The literal PTN004393219 IBD, target descent and conserved catalytic residues support peptidyl-proline 4-dioxygenase activity. GO:0160082 is an authentic is_a child and describes directly tested human HIF-alpha substrates. Refine to that informative reaction without interpreting it as exclusive substrate specificity; the separate neuronal FLNA evidence remains explicit contextual biology.

Proposed replacements: hypoxia-inducible factor-proline dioxygenase activity

Summary: The cited assays establish oxygen-responsive HIF regulation rather than intracellular oxygen homeostasis.

Reason: GO:0032364 requires maintenance of a steady-state oxygen level within a cell. The inspected PMID 16956324 experiments measure HIF signaling, substrate hydroxylation and oxygen-dependent responses, not stabilization of intracellular oxygen concentration. Oxygen sensing and downstream adaptation are real, but this particular steady-state homeostasis claim exceeds the measured mechanism; the supported HIF/hypoxia annotations are retained.

Summary: Neuronal apoptosis regulation is a context-dependent family inference.

Reason: The target descends from the literal PTN004393219 IBD with experimentally grounded EGLN-family descendants. PMID 19587290 specifically found PHD1, not PHD2, knockdown protective against HCA oxidative death in rat cortical neurons; that is a conditional negative, not universal absence of EGLN1 neuronal regulation. PMID 26818499 independently shows PHD2-dependent effects in mitochondrial-stress dopaminergic models. Retain the broad inferred regulation as non-core with cell/stimulus dependence, rather than rejecting the ancestral node by donor count or treating all paralogs as interchangeable.

Summary: Angiogenesis regulation is retained as a downstream physiological context.

Reason: The verified mouse Egln1 donor IMP cites PMID 18500250. Its complete abstract reports altered angiogenic activity of conditioned medium from PHD2-silenced NIH3T3 fibroblasts, HUVEC responses and mouse Matrigel angiogenesis. This supports regulatory influence through the oxygen/HIF pathway, not an autonomous angiogenic growth-factor activity or a direct human endothelial-cell mechanism. Operative full-text details were not recovered.

Summary: Angiogenesis regulation is retained as a downstream physiological context.

Reason: The verified mouse Egln1 donor IMP cites PMID 18500250. Its complete abstract reports altered angiogenic activity of conditioned medium from PHD2-silenced NIH3T3 fibroblasts, HUVEC responses and mouse Matrigel angiogenesis. This supports regulatory influence through the oxygen/HIF pathway, not an autonomous angiogenic growth-factor activity or a direct human endothelial-cell mechanism. Operative full-text details were not recovered.

Summary: The traced positive-transcription transfer conflicts with the direction of the donor experiment.

Reason: The current Q91YE3 donor GO:0045944 IMP explicitly cites PMID 26818499 with ATP13A2 as input. The recovered original Methods quantify mRNA by qPCR; Results and Figure 3 report that PHD2 inhibition or knockdown increases ATP13A2 mRNA/protein, while PHD1/PHD3 knockdown does not. Thus this source does not establish positive regulation of transcription by PHD2 and gives the opposite expression-direction relationship. Remove this electronic assertion on that verified source-specific basis. Do not infer an opposite transcription annotation from steady-state mRNA alone, and retain the separate supported iron-homeostasis and HIF-signaling claims.

Summary: The cellular hypoxia response is conserved in the PHD lineage.

Reason: The literal PTN004393218 IBD uses the fly Hph descendant and the human target lies within its inherited lineage. Conserved catalytic residues and direct human PHD2/HIF responses support the same cellular oxygen-sensing mechanism. The short donor list is not evidence of weakness, and no target-specific loss is indicated.

Summary: Nitric oxide can modify the PHD2 catalytic system in vitro.

Reason: The complete PMID 21601578 abstract explicitly describes biochemical/biophysical reactions of PHD2 with nitric oxide at Fe2+ and cysteine residues, while leaving biological significance unresolved. Retain the original response annotation as a non-core biochemical context; do not turn it into an established physiological NO-signaling mechanism or an additional core function.

Summary: A glutamatergic postsynaptic pool is supported in mouse neurons.

Reason: The source detects PHD2 in PSD fractions and neuronal protrusions and examines vGlut1/PSD95-positive synaptic contacts. The literal mouse Q91YE3/SynGO donor cites PMID 26972007. Preserve this mechanistic context and the rodent neuronal versus transfected-cell distinction; do not report a direct endogenous human-neuron experiment or assign every current human product the same effect.

Summary: PHD2 contributes to remodeling of postsynaptic structures in the mouse neuronal context.

Reason: Mouse neuronal PHD2 depletion/deletion and active-versus-inactive rescue alter spine morphology and synapse density; FLNA perturbations connect these effects to hydroxylation-dependent protein turnover. The literal mouse Q91YE3/SynGO donor cites PMID 26972007. Preserve this mechanistic context and the rodent neuronal versus transfected-cell distinction; do not report a direct endogenous human-neuron experiment or assign every current human product the same effect.

Summary: PHD2 regulates postsynaptic FLNA turnover through the hydroxylation/VHL pathway.

Reason: The original combines neuronal FLNA accumulation after PHD2 loss, proteasome-inhibitor effects, cell-based proline hydroxylation/MS, FLNA mutants and VHL-dependent turnover. This is a regulatory hydroxylation contribution, not direct protease activity. The literal mouse Q91YE3/SynGO donor cites PMID 26972007. Preserve this mechanistic context and the rodent neuronal versus transfected-cell distinction; do not report a direct endogenous human-neuron experiment or assign every current human product the same effect.

Summary: Human EGLN1 catalyzes HIF-proline hydroxylation.

Reason: PMID 25129147 uses human EGLN1 constructs expressed in insect cells, HIF1/HIF2 substrates and biochemical kinetic assays, with cell-based depletion/rescue comparisons. The measured differences among variants do not change the wild-type catalytic identity. The source does not authenticate equivalent activity for all natural isoforms.

Summary: The original human HPH2 assay directly establishes HIF-proline hydroxylation.

Reason: PMID 11598268 uses recombinant human HPH2 and HIF1alpha peptide substrate, with VHL capture and mass-spectrometric/proline controls. The HPH2 AF229245 splice construct omits residues 76–177 and is not equated with current Q9GZT9-3 or canonical full length. The HPH1 cellular reporter and fly RNAi portions are separate evidence, not human EGLN1 assays.

Summary: The integrated enzyme mapping matches experimentally established HIF-proline dioxygenase activity.

Reason: The imported ARBA/Rhea/EC sources are preserved. Current official UniProt catalytic annotation names RHEA:48400 and EC1.14.11.29, and direct human assays independently establish the reaction. The ARBA rule endpoint did not return an inspectable rule, so its predicate-level audit remains unresolved; this access failure does not overturn the directly supported enzyme activity. The apparent HIF1B wording in one official FUNCTION sentence is not adopted as a substrate claim.

Summary: EGLN1 suppresses HIF1alpha signaling through oxygen-dependent substrate hydroxylation.

Reason: The inspected PMID 16956324 human-cell experiments show increased HIF1alpha/HRE activity after PHD2 depletion and reduced HIF-substrate hydroxylation. Together with direct human enzyme chemistry, this supports negative regulation of the HIF1alpha pathway. IOP1 perturbations and broad iron/sulfur functions are not substituted for the PHD2 mechanism.