EIF2B5

Summary: Ovarian disease association does not establish a follicle-development step.

Reason: PMID:15507143 screened 93 women with isolated premature ovarian failure for selected EIF2B2/EIF2B4/EIF2B5 variants and found none. Its discussion proposes possible follicular mechanisms but does not assay follicle formation or maturation. The independently inspected PMID:12707859 does contain positive EIF2B5 ovarioleukodystrophy cases, including R113H homozygotes and an R113H/R195C case. Those clinical observations remain valid; they do not identify a follicle-development step performed by epsilon. This is a process-scope overstatement, not a denial of ovarian disease involvement. A separate source query asks which primary reference and experimental basis were intended for the imported IMP assertion; no replacement reference is asserted.

Summary: Epsilon participates in cytoplasmic translation initiation.

Reason: The independently inspected shared PMID:27023709 Methods 12 and Results 23/Figure 2h include a human eIF2/eIF2B comparator in reconstituted HCV-IRES translation, despite the primarily S. pombe purification and crystallography. Human five-subunit reconstitution in PMID:29599213 independently establishes the complex exchange mechanism that regenerates initiation-competent eIF2. The epsilon subunit participates in that mechanism; detailed yeast structural assays are not represented as human mutant experiments.

Summary: Epsilon contributes to eIF2B initiation-factor activity.

Reason: The inspected PMID:16289705 abstract reports increased translation after eIF2/eIF2B supplementation of HeLa extracts. It does not establish the species or preparation of the supplemented factors. The broad contributes_to assertion is independently supported by the human-complex comparator and reconstitution in PMID:27023709 and PMID:29599213. The host-cell name is not used to infer the origin of purified factors.

Summary: The initiation-factor assignment is biologically supported.

Reason: The reserved-source advisory identifies PMID:8688466 as rabbit epsilon cloning plus a human U937 cDNA fragment, with only the authentic abstract accessible. The incomplete historical clone is not a natural human isoform and is not treated as a full-length functional assay. Independent human eIF2B reconstitution and translation evidence support epsilon initiation-factor activity. The NAS source identity and enables qualifier are retained without claiming this paper alone demonstrated isolated human catalysis.

Summary: The epsilon-family exchange activity is conserved.

Reason: The saved PTHR45887 tree and MSA place the exact 721-residue human sequence below the literal PTN000500978 exchange-activity IBD. The inherited assertion is supported by epsilon-family biochemistry, including recombinant rat epsilon in PMID:9139680 and human eIF2B assays. A short donor list is not weak support. Epsilon is the catalytic subunit; the regulatory-subunit relation concerns raised for alpha or beta are not automatically transferred to epsilon.

Summary: Native human eIF2B exchanges GDP on eIF2.

Reason: The inspected PMID:11323413 Methods purify a five-subunit complex from HeLa cytoplasmic extracts using ammonium-sulfate fractionation and Q/S chromatography. The complex was depleted of eIF2 and assayed for GDP dissociation and nucleotide exchange. This directly supports the source contributes_to qualifier for epsilon in the purified native complex; it is not an isolated-epsilon assay. The source Methods are recovered externally while the normal cache remains abstract-only.

Summary: Epsilon supports the exchange activity of eIF2B.

Reason: The reserved PMID:25858979 advisory documents whole-complex GDP-release assays, with CHO delta-mutant and mouse-cell experiments distinguished from epsilon biochemistry. That paper is not an isolated-epsilon activity test. Nevertheless the imported exchange function is supported by epsilon catalytic-subunit evidence and direct human-complex studies. Unlike regulatory alpha/beta, epsilon has independent family evidence for a catalytic exchange domain, so no automatic enables-to-contributes_to defect is asserted.

Summary: Human eIF2B supports the conserved exchange mechanism.

Reason: PMID:27023709 includes a reconstituted human eIF2/eIF2B translation comparator, while its detailed recombinant and crystal assays are predominantly S. pombe. The human comparator and independent human-complex biochemistry support the broad exchange activity of epsilon within eIF2B. This does not turn the yeast structural or alpha-mutant experiments into direct human-epsilon assays.

Summary: Independent rule and ortholog evidence support epsilon exchange activity.

Reason: Current ARBA00028188 condition set 20 matches all three saved target FunFam identifiers; the current IPR044123 entry maps to exchange activity. The literal rat donor Q64350 has direct experimental exchange annotations, including recombinant epsilon in PMID:9139680 and complex assays in PMID:11500362. The exact rule predicates, donor records and target identifiers are replayed; descriptions alone are not treated as rule conditions.

Summary: Patient EIF2B5 variants reduce complex exchange activity.

Reason: PMID:15054402 measures GDP dissociation in EBV-transformed patient lymphoblast extracts using rat-liver eIF2 substrate. Table 1 identifies multiple EIF2B5 genotypes with reduced exchange activity; heterozygous relatives have normal activity. This is direct human genotype-to-complex-function evidence, not purified isolated epsilon or human-substrate proof. It supports the catalytic epsilon exchange assignment without assigning every mutant the same molecular defect.

Summary: Rabbit epsilon evidence supports conserved exchange function.

Reason: The literal P47823 donor has a current experimental exchange annotation to PMID:8168527. Its authentic abstract reports that antibodies to rabbit epsilon inhibit eIF2B exchange and translation initiation. The inspected scope does not demonstrate an isolated rabbit catalytic preparation, but it supports epsilon-dependent complex function; recombinant rat epsilon and human-complex work independently establish the conserved catalytic role.

Summary: A functional GSK3 docking interaction supports protein kinase binding.

Reason: The exact PMID:10481074 IntAct records concern human Q13144-derived engineered 534–547 peptides with substitutions and phosphoserine 544 assayed against GSK3B/P49841; they do not test intact native human epsilon. Independent PMID:11500362 Results and Figure 8 test full-length and C-terminally truncated rat epsilon: HA-tagged epsilon coimmunoprecipitates with GSK3 in HEK293 cells, and removal of the C-terminal region reduces association and efficient GSK3 phosphorylation. The immunoprecipitation detects GSK3 alpha and beta together, so it is not a beta-selective docking assay. This functional interaction, alongside the literal human peptide evidence and conservation of the mammalian region, supports protein kinase binding. Rat residue numbers and engineered constructs are not relabeled as unmodified human protein; no kinase activity is assigned to epsilon. The curated UniProt PTM statement places the DYRK2 priming site at human Ser-544 by similarity (ECO:0000250); it supplies conserved-site context without converting these engineered or rat assays into direct experiments on unmodified human epsilon.

Proposed replacements: protein kinase binding

Summary: The human gamma subunit is an identified initiation-factor partner.

Reason: Selected original PMID:15060152 cloning and Figure 1 results use full-length human subunits in HEK293 cells with tagged epsilon. Gamma/Q9NR50 associates with epsilon when the two are expressed together; larger complexes are assessed separately. Gamma is an independently established eIF2B initiation-factor subunit. Translation initiation factor binding is a supported class refinement; missing current IntAct rows for this historical source are not negative evidence.

Proposed replacements: translation initiation factor binding

Summary: The human delta subunit associates within the eIF2B complex.

Reason: The selected PMID:15060152 original reports human-subunit coexpression and complex recovery. Delta/Q9UI10 is recovered with epsilon in the assembled beta/gamma/delta/epsilon complex, whereas the binary gamma interaction is distinguishable. The delta association supports binding to an initiation-factor partner without claiming a purified binary interface or isolated delta catalysis.

Proposed replacements: translation initiation factor binding

Summary: Retain the positive STX11 interaction without assigning a specific functional role.

Reason: PMID:32296183 reports human Q13144–O75558/STX11 positives in pooled, array and validated yeast-two-hybrid records, with Gal4 fusion constructs and sequence-confirmed pair verification in the study workflow. The same protein pair is supported independently by the other study: binary screening in PMID:32296183 and cellular affinity capture in PMID:33961781. Retain the experimentally recorded generic interaction as non-core. These records do not establish an epsilon-specific functional role for the interaction, so selecting syntaxin binding solely from the partner name supplies no supported functional refinement. The interaction is not dismissed as false, nor is physiological occupancy a GO term requirement. Dataset-wide orthogonal validation is not assigned to this pair, the coimmunoprecipitation record remains association evidence, and no vesicle-transport role is inferred.

Summary: Retain the positive dynein heavy chain DYNC1H1 interaction without a partner-class functional refinement.

Reason: The source-specific PMID:32814053 IntAct records report positive human Q13144–Q14204 pairs in pooled, array and validated yeast-two-hybrid configurations. These curated literal records verify the generic experimental pair, which is retained as non-core. The selected paper is abstract-only and complete operative Methods and assayed construct details were not recovered, so no exact full-length insert or native cellular interface is claimed. Partner identity alone does not establish an epsilon-specific functional role warranting dynein heavy chain binding; remove that proposed replacement. The unresolved details limit the functional refinement, not the existence of the documented positive pair. Neither a physiologically false interaction nor a wrong-gene attribution is inferred; the cytosolic GEF role does not by itself exclude additional binding.

Summary: Retain the positive kallikrein-6/KLK6 interaction without a partner-class functional refinement.

Reason: The source-specific PMID:32814053 IntAct records report positive human Q13144–Q92876 pairs in pooled, array and validated yeast-two-hybrid configurations. These curated literal records verify the generic experimental pair, which is retained as non-core. The selected paper is abstract-only and complete operative Methods and assayed construct details were not recovered, so no exact full-length insert or native cellular interface is claimed. Partner identity alone does not establish an epsilon-specific functional role warranting protease binding; remove that proposed replacement. The unresolved details limit the functional refinement, not the existence of the documented positive pair. Neither a physiologically false interaction nor a wrong-gene attribution is inferred; the cytosolic GEF role does not by itself exclude additional binding.

Summary: Retain the positive STX11 interaction without assigning a specific functional role.

Reason: PMID:33961781 has an exact human Q13144–O75558/STX11 anti-tag coimmunoprecipitation record from HCT116 cells, with a C-terminal tandem tag on the STX11 bait; this assay establishes association, not necessarily binary contact. The same protein pair is supported independently by the other study: binary screening in PMID:32296183 and cellular affinity capture in PMID:33961781. Retain the experimentally recorded generic interaction as non-core. These records do not establish an epsilon-specific functional role for the interaction, so selecting syntaxin binding solely from the partner name supplies no supported functional refinement. The interaction is not dismissed as false, nor is physiological occupancy a GO term requirement. Dataset-wide orthogonal validation is not assigned to this pair, the coimmunoprecipitation record remains association evidence, and no vesicle-transport role is inferred.

Summary: Nuclear exclusion is retained within the tested expression context.

Reason: The externally recovered PMID:12426392 original expresses human epsilon cDNA as eGFP fusions in BHK cells. Basal epsilon is excluded from nuclei; engineered NLS and export-inhibition experiments test distinct constructs and conditions. Retain the imported NOT nucleus observation as contextual, not a universal impossibility of nuclear localization. A separate endogenous mouse immunostaining experiment reports nuclear signal and is not conflated with this human-construct assay.

Summary: The mouse donor supports contextual nuclear localization.

Reason: The literal mouse Q8CHW4 donor has an IDA nuclear annotation to PMID:12388085. The original Results and Figures 1–2 captions describe endogenous epsilon staining in NIH3T3 nuclei and cytoplasm. Raising paraformaldehyde from 1% to 3% increases cytoplasmic staining relative to nuclear staining; the text does not report abolition of the nuclear signal. This limits quantitative partitioning but does not establish that all nuclear staining is artifactual. Retain the contextual non-core ISS localization, with the existing species/construct consistency exception. The separate human epsilon–eGFP experiment in BHK cells supports basal exclusion in that expression context, not universal exclusion of endogenous epsilon from all human nuclei.

Summary: Independent target evidence refines the cytoplasmic localization to cytosol.

Reason: PMID:11323413 purifies native human five-subunit eIF2B from HeLa cytoplasmic extracts; that original preparation supports the imported broad compartment. Independent target evidence supports a more precise review conclusion: HPA classifies EIF2B5 cytosol localization as enhanced, and literal Reactome paths from reactions 72670/72722 reach the cytosolic epsilon component 72523, Q13144 residues 2–721. Refine to cytosol while preserving the original IDA source and qualifier. The refinement does not imply that the old purification itself resolved every cytoplasmic compartment or that epsilon is exclusively cytosolic.

Proposed replacements: cytosol

Summary: Verified human target localization supports cytosol beyond the original construct assay.

Reason: PMID:12426392 shows human epsilon-eGFP in the cytoplasm of BHK cells with basal nuclear exclusion; the tagged human construct and hamster host remain distinct. The original images alone do not exclude every organellar or peripheral-membrane association. Independently, the human target HPA record gives enhanced cytosol localization, and Reactome directly maps Q13144 residues 2–721 to the cytosolic epsilon component of the exchange complex. Refine the broad annotation to cytosol on this combined evidence, without claiming an exclusive location or recasting the original assay as cytosolic fractionation.

Proposed replacements: cytosol

Summary: Direct target evidence permits a cytosol refinement of the ortholog-based compartment.

Reason: The literal mouse donor Q8CHW4 and ENSMUSP00000003320 chain carries direct NIH3T3 cytoplasmic immunofluorescence evidence from PMID:12388085. That donor experiment retains its mouse, endogenous-staining and stress-context scope. Independently verified human EIF2B5 HPA localization is enhanced for cytosol, and the Reactome Q13144 component is explicitly cytosolic. These target data justify refining the broad transferred compartment to cytosol. They do not change the imported IEA evidence, original reference, donor identities or qualifier, and they do not assert that epsilon lacks other contextual locations.

Proposed replacements: cytosol

Summary: HPA provides location-specific enhanced cytosol evidence.

Reason: The saved HPA EIF2B5 page explicitly classifies cytosol as enhanced. It lists HPA064370 in MCF7/PC3/U2OS and HPA069303 in HaCaT/RT4/U2OS. This is the location-specific score, not merely aggregate gene reliability; image pixels were not independently examined. The positive cytosol annotation is consistent with the soluble initiation-factor complex.

Summary: The UniProt location mapping resolves to cytosol.

Reason: The literal source SL-0091 is currently Cytosol and maps to GO:0005829; its parent is cytoplasm. The target location record and enhanced HPA cytosol result agree. This is a checked location mapping, not a default prediction from absence of signal peptides.

Summary: Reactome places epsilon-containing exchange complexes in cytosol.

Reason: The saved reaction 72670 reaches complex 72526 and its epsilon component 72523, whose reference is human Q13144, residues 2–721, in cytosol. Removal of the initiator methionine is not a separate natural isoform. The path is a curated reaction/complex localization model, not an independently performed microscopy assay.

Summary: The exchange reaction contains cytosolic epsilon.

Reason: Reactome reaction 72722 reaches epsilon component 72523 through its catalyst and complex participants. The literal leaf maps to Q13144 residues 2–721 with cytosol compartment. These exact paths support the imported curated location without implying that every modeled edge is a distinct experiment.

Summary: The epsilon sequence inherits eIF2B-complex membership.

Reason: The literal PTN000500978 complex-membership IBD lies on the saved target path, and the MSA contains the exact 721-residue human epsilon sequence. Current experimental human complex membership and cross-species donor records support the conserved assignment. Q13144 in the source list is legitimate descendant experimental grounding, not circularity.

Summary: Epsilon is present in purified native human eIF2B.

Reason: The complete operative PMID:11323413 purification and subunit characterization recover native HeLa five-subunit eIF2B. This supports epsilon complex membership directly; the exchange assay is whole-complex, and separately expressed alpha/beta preparations are not attributed to epsilon.

Summary: Human epsilon assembles with the other eIF2B subunits.

Reason: Selected original PMID:15060152 Methods and Figure 1 results use sequence-verified full-length human clones in HEK293 cells. Tagged epsilon recovers a beta/gamma/delta/epsilon complex and the five-subunit assembly, with binary gamma association distinguished. The positive complex assignment does not require every pair to form a stable isolated dimer.

Summary: The human comparator contains the eIF2B complex.

Reason: PMID:27023709 explicitly includes human eIF2B in a reconstituted translation comparator; most detailed structural and recombinant experiments are S. pombe. Human complex membership is independently corroborated by native purification and human-subunit reconstitution. The yeast title is not used to dismiss the curator assertion.

Summary: Rat complex membership supports the epsilon ortholog transfer.

Reason: The exact Q64350/ENSRNOP00000002321 chain carries experimental rat eIF2B-complex annotations, including PMID:9139680. That study expresses rat subunits and tests assembled complexes. Direct human native and recombinant complex evidence independently supports the conserved membership; donor species is preserved.

Summary: Human recombinant epsilon is part of the reconstituted assembly.

Reason: The independently inspected PMID:29599213 original uses five codon-optimized human genes expressed in E. coli. Alpha homodimer and beta/gamma/delta/epsilon tetramer reconstitution supports the human eIF2B assembly. Exact synthesized sequences reside in uninspected supplementary material, so no unsupported natural-product equivalence is claimed from the host or gene name alone.

Summary: The initiation process is cytoplasmic in this source context.

Reason: PMID:16289705 reports HeLa-extract translation supplemented with eIF2/eIF2B, including capped, uncapped and IRES reporters. The broad initiation assertion is supported, and cytoplasmic translational initiation is the correct compartment-specific process. The factor-preparation species remains unresolved in the inspected abstract; independent human comparator/reconstitution evidence supports the mechanism.

Proposed replacements: cytoplasmic translational initiation

Summary: Epsilon positively regulates the initiation step.

Reason: ARBA00035061 condition set 0 matches the target IPR003307/IPR016024 identifiers and primate lineage. The broad regulation term is true, but epsilon-mediated eIF2 nucleotide exchange promotes initiation-competent eIF2 and human reconstitution directly increases initiation. Positive regulation of translational initiation captures that defined sign and step, without assigning epsilon kinase or GTP-hydrolysis activity.

Proposed replacements: positive regulation of translational initiation

Summary: The rat donor supports a contextual heat response.

Reason: Current Q64350 carries direct PMID:9341116 heat-response evidence. Its authentic abstract describes rat H35 cells in which mild heat changes eIF2B exchange activity and protein synthesis; this supports a regulated complex response, not epsilon acting as a temperature sensor. Retain the conserved response as non-core, with the complete experimental preparation unresolved in the abstract-only original.

Summary: Independent evidence supports a contextual heat response, with the original human assay unresolved.

Reason: The accessible PMID:15723074 abstract and publisher preview provide positive human EIF2B5 astrocyte findings but do not expose the operative heat experiment; full Methods and supplementary experiments remain unavailable. That citation-specific access limit is retained. Independently, the exact rat epsilon donor Q64350 has an IDA heat-response annotation to PMID:9341116, whose authentic abstract reports changes in eIF2B exchange activity correlated with protein synthesis during and after mild heat in H35 cells. Together with the conserved catalytic role of epsilon in mammalian eIF2B, this supports the broad heat-response claim as non-core. It does not establish epsilon thermosensing or reconstruct the uninspected human IMP experiment.

Summary: Rat heat-regulated exchange is a defensible non-core transfer.

Reason: The literal Q64350 donor has current direct PMID:9341116 evidence for changes in eIF2B activity under mild heat in rat H35 cells. That supports a contextual physiological response inherited by the mammalian complex. It does not establish epsilon thermosensing or independently resolve the different human IMP/TAS source details.

Summary: The broad heat-response claim is retained as non-core using independent functional evidence.

Reason: The PMID:12499492 abstract reports severe homozygous EIF2B5 disease and metabolic findings; its specific heat-response statement remains unverified because the full original is unavailable. Independent evidence nevertheless supports the biological claim: current rat Q64350 has direct PMID:9341116 heat-response annotation, and the authentic abstract describes regulated eIF2B exchange activity associated with translation changes during and after mild heat in H35 cells. Conserved epsilon participation in this complex supports a contextual mammalian response. Retain the term as non-core while keeping the original TAS citation, species boundaries and access caveat explicit; neither febrile deterioration alone nor a sibling review is used as the functional argument.

Summary: The rat-islet donor supports a contextual glucose response.

Reason: Current Q64350 carries PMID:8567668 experimental evidence: the authentic abstract reports glucose-dependent increases in eIF2B exchange activity in rat pancreatic islets, whereas mannoheptulose did not activate eIF2B. This is positive complex-level response evidence. Retain it as non-core without claiming epsilon is a glucose receptor or that every mechanistic detail has been demonstrated in human cells.

Summary: Glucose-regulated exchange supports the ortholog transfer.

Reason: The exact rat Q64350 donor is experimentally annotated from PMID:8567668, whose abstract reports rapid glucose-dependent eIF2B activation in isolated islets. Conserved epsilon participation in this regulated complex makes the broad response defensible. The source is a rat complex assay, not a direct purified-human glucose-binding experiment.

Summary: Human epsilon is required for astrocyte development in the reported culture context.

Reason: The PMID:15723074 abstract reports few GFAP-positive astrocytes from affected EIF2B5 patient brain cultures and impaired GFAP induction after EIF2B5 RNA interference in normal human glial progenitors. This positive perturbation evidence supports a contextual developmental requirement, retained as non-core. It does not establish an astrocyte-specific catalytic activity or a direct fate-specifying mechanism. Full operative Methods and supplementary data remain unavailable; the positive accessible findings are not erased.

Summary: The source measures oligodendrocyte pathology rather than developmental progression.

Reason: Selected PMID:15217090 autopsy Methods, genotypes and Results include four EIF2B5 patients and one EIF2B2 patient. Oligodendrocyte density, regional staining, apoptotic markers and myelin abnormalities are positive observations. The authors leave increased density compatible with alternative proliferation or survival explanations. Those measurements do not establish the progression from fate commitment to mature oligodendrocytes required by GO:0014003. Preserve the human disease evidence while marking the developmental mechanism over-specific.

Summary: Developmental expression does not establish hippocampal developmental function.

Reason: The literal rat Q64350 donor carries GO:0021766 with IEP evidence from PMID:12850562. Its authentic abstract measures developmental changes in hippocampal eIF2B-subunit expression, including high late-embryonic levels followed by decline. Expression during development is not evidence that epsilon performs hippocampal morphogenesis. The source identity and positive expression observation remain intact; the process transfer overstates that evidence.

Summary: Epsilon binds translation-initiation machinery through a conserved function.

Reason: The literal PTN000500978 initiation-factor-binding IBD lies on the target path. The rat experimental donor has direct eIF2 association evidence, including PMID:11500362, while human eIF2B complex/substrate assays support the same mechanistic role. Epsilon is the catalytic subunit of eIF2B. This binding assertion describes its initiation-factor contacts in the complex exchange mechanism, not a separate claim about catalytic sufficiency. The single listed donor does not weaken the inherited evidence.

Summary: Domain and ortholog evidence support initiation-factor binding.

Reason: The current IPR044123 mapping includes GO:0031369, and the literal rat Q64350 donor carries experimental binding annotations. PMID:11500362 compares recombinant rat epsilon variants for eIF2 association and exchange in assembled complexes. The target family identity and human-complex mechanism agree. Rat residue numbers are not silently mapped onto the current human sequence.

Summary: The rat binding experiment supports conserved epsilon partner recognition.

Reason: PMID:11500362 directly compares rat epsilon variants for association with eIF2 and for exchange activity in eIF2B complexes. Its native human phosphorylation observation is a separate experiment; the detailed recombinant constructs are rat. The conserved source transfer and human-complex work support initiation-factor binding without pretending the rat mutants are human assays.

Summary: Conserved epsilon activity participates in the cellular response to ER stress.

Reason: PMID:16041584 reports heightened ATF4 induction in patient fibroblasts with eIF2B mutations despite similar broad translational inhibition and upstream kinase activation; the exact EIF2B5 genotype and operative patient experiments remain unresolved because the retrieved publisher payload was an HTML challenge. The broad response claim has independent support: PMID:12388085 Results and Figure 1 describe endogenous mouse epsilon redistribution into stress granules after thapsigargin treatment, and PMID:11500362 describes the catalytic epsilon-containing eIF2B complex whose exchange activity is inhibited by phosphorylated eIF2 during stress. Epsilon contributes the exchange step controlling translation, rather than merely being required for a downstream phenotype. Together these observations support a conserved, contextual ER-stress response as non-core. They do not reconstruct the unavailable human IMP experiment, identify epsilon as the ER sensor, or assign transcription-factor activity to it.

Summary: Clinical myelin pathology does not settle epsilon participation in myelination.

Reason: The inspected PMID:14566705 Table 1 identifies EIF2B5 patients 7 and 8 with V430A/S447L and homozygous Y495C genotypes; clinical and MRI results document severe myelin abnormalities. These observations establish disease involvement but do not demonstrate a step performed by epsilon in myelin-sheath formation or maintenance. They also do not disprove such a step. The unresolved issue for this source is whether epsilon contributes to myelination itself beyond its established translation function in affected cells. Reading the clinical study does not resolve that biological attribution. Retain UNDECIDED on this evidence-specific uncertainty, preserving the positive myelin pathology without treating disease necessity as demonstrated process participation.

Summary: The myelination assertion remains unresolved beyond the accessible glial findings.

Reason: PMID:15723074 links human EIF2B5 disease to white-matter pathology and altered astrocyte generation. Its accessible abstract reports readily generated normal-appearing oligodendrocytes but few abnormal astrocytes in patient cultures. These positive observations suggest glial pathophysiology but do not resolve the specific sheath-formation or maintenance experiment intended by GO:0042552. Full Methods and supplementary experiments remain unavailable. Preserve the curator assertion as unresolved, without assuming that uninspected experiments were absent or that myelin biology is unaffected.

Summary: The positive-apoptosis term overstates the demonstrated phosphorylation-dependent role.

Reason: The exact rat Q64350 donor annotates positive regulation of apoptosis from PMID:11756553. Its Methods use sequence-confirmed rat epsilon S535A/S539A constructs in Rat1/PC12 cells. The nonphosphorylatable mutants preserve protein synthesis and protect against GSK3 overexpression, PI3K inhibition or serum-withdrawal apoptosis; the mechanism is loss of exchange activity during proapoptotic signaling, not active epsilon promoting apoptosis. Results 23–45 and Discussion 49–54 connect GSK3-dependent inhibition of epsilon-supported translation to cell death; tau phosphorylation and cycloheximide non-rescue are mechanistic controls. This is real pathway involvement, but it does not identify an active epsilon proapoptotic step. The term overstates that role; the judgment does not rest on lack of an effect from WT overexpression and does not assert universal absence of positive regulation. No opposite NEW annotation is manufactured.

Summary: Regulation of epsilon-dependent exchange supports a contextual peptide-hormone response.

Reason: The literal rat donor Q64350 traces this annotation to PMID:8430778, whose inspected abstract compares diabetic-tissue eIF2B activity and NADPH/CKII effects without exposing a peptide-hormone challenge; that original donor experiment remains unresolved. The broader mammalian response claim is supported independently. PMID:8397507 reports insulin-dependent GSK3 inactivation in human-insulin-receptor-transfected CHO cells and identifies the largest eIF2B subunit as a kinase substrate; its abstract does not itself measure epsilon dephosphorylation after insulin. PMID:11500362 Introduction explicitly relates insulin-dependent GSK3 inactivation, epsilon dephosphorylation and eIF2B activation as prior experimental knowledge, while its own Results/Figure 8 establish a functional GSK3 docking region in rat epsilon. PMID:11756553 tests rat epsilon phosphosite mutants in Rat1/PC12 cells under GSK3, PI3K and serum-withdrawal perturbations, supporting the regulated translation mechanism but not supplying an insulin-challenge experiment. This combined evidence supports the broad response term as non-core because epsilon catalyzes the regulated exchange step. It does not establish a direct hormone receptor function, reproduce the unavailable donor assay, or convert rat phosphosite numbering into human numbering.

Summary: Regulation of epsilon-dependent exchange supports a contextual peptide-hormone response.

Reason: The literal rat donor Q64350 traces this annotation to PMID:8430778, whose inspected abstract compares diabetic-tissue eIF2B activity and NADPH/CKII effects without exposing a peptide-hormone challenge; that original donor experiment remains unresolved. The broader mammalian response claim is supported independently. PMID:8397507 reports insulin-dependent GSK3 inactivation in human-insulin-receptor-transfected CHO cells and identifies the largest eIF2B subunit as a kinase substrate; its abstract does not itself measure epsilon dephosphorylation after insulin. PMID:11500362 Introduction explicitly relates insulin-dependent GSK3 inactivation, epsilon dephosphorylation and eIF2B activation as prior experimental knowledge, while its own Results/Figure 8 establish a functional GSK3 docking region in rat epsilon. PMID:11756553 tests rat epsilon phosphosite mutants in Rat1/PC12 cells under GSK3, PI3K and serum-withdrawal perturbations, supporting the regulated translation mechanism but not supplying an insulin-challenge experiment. This combined evidence supports the broad response term as non-core because epsilon catalyzes the regulated exchange step. It does not establish a direct hormone receptor function, reproduce the unavailable donor assay, or convert rat phosphosite numbering into human numbering.

Summary: The positive effect is specifically on translational initiation.

Reason: ARBA00084181 condition set 0 matches the three target FunFam identifiers, and rat Q64350 has direct positive-translation evidence from PMID:20484009. That study overexpresses rat epsilon in muscle: it increases exchange activity and restores sepsis-impaired protein synthesis, while healthy-muscle protein synthesis need not increase. Human-complex reconstitution identifies the underlying initiation step. Refine the broad positive-translation term to positive regulation of translational initiation without claiming a universal dose response.

Proposed replacements: positive regulation of translational initiation

Summary: Human eIF2B exchange promotes translation initiation.

Reason: The independently inspected PMID:29599213 original reconstitutes human eIF2B from E. coli-expressed subunits and measures GDP exchange on eIF2. It directly supports the positive initiation mechanism, while alpha-free tetramer activity and ISRIB effects distinguish assembly regulation. Epsilon participates as the catalytic subunit; the experiment is not described as an isolated-human-epsilon assay.

Summary: Rabbit epsilon-dependent exchange supports positive initiation.

Reason: The current P47823 donor has IMP positive-initiation evidence from PMID:8168527. Its authentic abstract reports inhibition of exchange and initiation by epsilon-directed antibodies, with other subunit antibodies distinguished. Conserved epsilon function and direct human reconstitution support the transfer, without assuming antibody recognition alone proves a particular human residue or isoform.

Summary: Human astrocyte differentiation has positive contextual epsilon evidence.

Reason: PMID:15723074 reports impaired GFAP induction after EIF2B5 knockdown in normal human glial progenitors, together with fewer abnormal astrocytes from affected patient cultures. That supports a non-core requirement in the tested differentiation context. It does not establish a unique astrocyte substrate or an independent fate-regulatory enzyme. The accessible preview and abstract support this bounded conclusion, while detailed Methods and supplementary experiments remain uninspected.

Summary: The epsilon-specific T-cell signaling role is unresolved.

Reason: PMID:8626696 reports increased whole-eIF2B exchange activity and reduced GSK3 activity during human T-cell activation. The recovered selected original Results/Discussion also discuss beta-subunit abundance. Complete operative Methods and epsilon-specific perturbations are unavailable, so the direct epsilon participation claimed by the pathway annotation cannot be confidently separated from regulated translation downstream of activation. Positive whole-complex activity is preserved; beta emphasis alone is not a wrong-gene argument.