EIF2B2 is the beta regulatory subunit of cytosolic eIF2B. Native human complex purification, human subunit assembly and variant-bearing patient extracts support its contribution to nucleotide exchange and translation initiation. The exchange reaction belongs to the assembled machinery; these observations do not establish beta-only GEF activity. In the patient assay, the extracts are human and the GDP-bound eIF2 substrate is from rat liver. [PMID:11323413; PMID:15054402; PMID:15060152]
Native beta ATP/GTP cross-linking is a positive observation. Recombinant beta also binds GTP, but the inspected historical cloning passage does not establish its exact current-human insert sequence. The physiological role of nucleotide binding remains unresolved, and no beta GTPase activity or second catalytic mechanism follows from it. PMID:11323413
Six EIF2B4 associations support translation initiation factor binding. The older human-subunit paper recovered beta in larger assemblies while finding no stable isolated beta–individual-subunit association in its pairwise assay. Later Y2H and AP-MS records retain their own scope. C9orf72 association is retained as non-core generic binding because GEF/GAP labels do not by themselves establish a narrower catalytic or scaffold partner class. This follows the explicit ClinGen project instruction for supported generic associations; no pair-specific activity is invented. [PMID:15060152; PMID:25910212; PMID:31515488; PMID:32296183; PMID:33961781]
Human EIF2B2 variants are associated with vanishing-white-matter disease and severe myelination abnormalities. The original infant/autopsy evidence includes beta-associated oligodendrocyte pathology, but altered survival and apoptosis do not resolve the developmental progression defined by the imported oligodendrocyte-development term. The ovarian screen tested known variants in isolated ovarian failure and does not establish a follicle-building role. The review preserves clinical relevance while separating these observations from specific developmental mechanisms. [PMID:14566705; PMID:15217090; PMID:15507143; PMID:15776425]
Rat-ortholog heat, glucose and peptide-hormone responses concern regulated whole-factor activity; they are retained as contextual non-core functions without assigning beta a sensor or receptor role. The clinical heat-response citation and human T-cell signaling assertion retain source-specific uncertainty. Early whole-factor activity and later beta abundance after T-cell activation are distinct from a demonstrated antigen-receptor signaling step. [PMID:9341116; PMID:8567668; PMID:8430778; PMID:12499492; PMID:8626696]
The PAINT target aligns exactly to the official 351-residue sequence and inherits the literal PTN000027996 assertions. The target's own experimental evidence is valid descendant grounding. ARBA family matches are authentic, but broad developmental conclusions require more than family membership and clinical pathology. The two curated Reactome events reach the exact beta entity in cytosol. These record-level proofs are reproducible in the evidence report.
There is one official displayed product in the imported record; no artificial isoform suffix is added. The deposited human beta sequence in PDB 6K72 matches the 351-residue reference, but deposition identity is not a substitute for unavailable original Methods or an inferred contact/catalytic assay. Independently inspected shared originals distinguish chiefly yeast structural experiments from a human translation comparator, and human subunit reconstitution from isolated-subunit activity. [PMID:27023709; PMID:29599213; PMID:31048492]
The normal publication caches remain unchanged. Their full-text flags describe those files; selected genuine external originals and shared independent source assessments have separate, explicit access scopes in the source guide. No figure-pixel or uninspected supplementary-sequence claim is made. Automated research attempts failed with provider-access errors; no generated report was invented. All support is reference-based, with no new publication quotations.
All 42 imported assertion objects and qualifiers are preserved. The decisions are 19 ACCEPT, 8 MODIFY, 8 KEEP_AS_NON_CORE, 4 UNDECIDED and 3 MARK_AS_OVER_ANNOTATED. No new process annotation is proposed. Outstanding questions concern the physiological role of beta nucleotide binding and the source-specific evidence needed to resolve developmental, heat and TCR claims.
The two UniProt IDA GO:0005085 assertions from PMID:25858979 and PMID:27023709 retain their imported enables relation and ACCEPT action. A probable WRONG_QUALIFIER query asks UniProt to check contributes_to; its cause is unknown and it has not been reported upstream. Official GO relation guidance and the source-specific complex assays motivate this distinction. The supported beta function remains contribution to eIF2B exchange, with the same integrated core. Retaining the imported relation preserves provenance without endorsing it. This is not a demand that every enzyme function in isolation, a confirmed database error, or a rejection of beta participation. The source guide records the conditional query; all original assay-access limits and selected cache bytes remain unchanged.
The human disease evidence includes EIF2B2-associated vanishing-white-matter disease and ovarian failure. The original VWM report identifies beta-subunit variants, and the ovarian series identifies affected patients 944 and 992 with EIF2B2 variants. The ovarian paper leaves decreased follicle formation, increased loss, altered recruitment and interrupted maturation as alternatives. The later negative screen of selected known variants in isolated ovarian failure does not refute the earlier positive cases. The follicle-development annotation remains over-specific to the demonstrated source evidence. [PMID:11704758; PMID:12707859; PMID:15507143]
The preserved UniProt P49770 record explicitly places alpha/beta/delta in the regulatory subcomplex and gamma/epsilon in the catalytic subcomplex, with experimental attributions to PMID:25858979, PMID:27023709 and PMID:31048492. This positive architecture supports beta's contribution to assembled exchange activity. It does not alone establish a curator error in a particular annotation relation: the two source queries retain CHECK, PROBABLE confidence, unknown cause and unreported status, together with their original assay-access limits.
The oligodendrocyte and T-cell decisions remain scientifically unresolved. The autopsy study contains positive beta-associated glial pathology, but its evidence does not separate developmental progression from survival effects. The T-cell study contains positive early exchange activation and later beta accumulation, but it does not resolve whether beta performs part of receptor-initiated signaling or is regulated downstream. These judgments do not rest on an assertion that the original Results were unavailable. Selected original text was read, and the source guide states the precise access boundaries. No isolated beta catalytic or developmental assay is demanded. [PMID:15217090; PMID:8626696]
The distinct same-term source decisions are retained: uncertain source-specific evidence is not a negative assertion or a contradiction with another positive context. The older isolated beta–delta pair negative remains distinct from its positive assembled-complex recovery and later pair assays. The broad cytoplasm annotation is explicitly refined to cytosol using the independent human EIF2B2 Reactome entity and the separately attributed UniProt location. The original HeLa fractionation itself resolves only cytoplasm. [Reactome:R-HSA-72670; Reactome:R-HSA-72722; PMID:11323413] The integrated core remains contribution to the exchange machinery in cytoplasmic initiation; no additional process or core unit is added merely to repeat accepted annotations. The physiological role of beta nucleotide binding remains an open question; the ATP rationale now cites its own ATP experiment, leaving recombinant GTP evidence in the GTP annotation. All 42 imported assertions and qualifiers, core and product fields remain unchanged. The sole action change is row18 ACCEPT to MODIFY with GO:0005829; the resulting counts are 18 ACCEPT, 9 MODIFY, 8 KEEP_AS_NON_CORE, 4 UNDECIDED and 3 MARK_AS_OVER_ANNOTATED. No publication quotation is added.