GCN2 couples amino-acid and translation stress to eIF2alpha phosphorylation and altered initiation. The evidence supports one integrated eIF2alpha kinase mechanism. Reduced bulk initiation and increased translation of selected stress transcripts are compatible outcomes of that mechanism. The review retains 52 source contexts: 23 ACCEPT, 15 MODIFY, 12 KEEP_AS_NON_CORE and two UNDECIDED, with no NEW assertions or publication excerpts.
The early mammalian biochemistry uses mouse GCN2 AJ243533, expressed in human 293/293T cells or obtained from mouse liver. Human eIF2alpha is the substrate, including a Ser51 mutant control. The human GCN2 sequence reported in that study is only a 548-residue EST-derived fragment; it is not the enzyme used in the assays. Serum-withdrawal activation and the lack of an amino-acid response in that particular transfected preparation should not be generalized to every physiological stress or to every human protein product. PMID:10504407(https://pubmed.ncbi.nlm.nih.gov/10504407/)
Human SQ20B GCN2 depletion reduces eIF2alpha phosphorylation during leucine limitation. The p21 study separates those human experiments from mouse reporter, polysome and translation experiments. Its selective increase in a p21 transcript's translation is consistent with repression of bulk synthesis. The later human knockout/rescue work supports GCN2-dependent initiation control in rRNA quality control, while RNF10 performs the ubiquitination step. Neither downstream RNA decay nor ubiquitin-ligase chemistry is assigned to GCN2. PMID:26102367(https://pubmed.ncbi.nlm.nih.gov/26102367/), PMID:39947182(https://pubmed.ncbi.nlm.nih.gov/39947182/)
The collision-response experiments use human MCF10A cells. GCN2 inhibition and shRNA reduce eIF2alpha phosphorylation; kinase-dead ZAKalpha rescues this branch, while simple glutamine-starvation signaling persists without ZAK. GCN1 was captured on collided ribosomes, but GCN2 and GCN20 were not captured in that preparation. Positive yeast and mouse ribosome-association evidence is retained separately: association does not require GCN2 to be an obligatory structural ribosome subunit, and a preparation-specific failure to capture it does not establish universal absence. PMID:32610081(https://pubmed.ncbi.nlm.nih.gov/32610081/), PMID:2038314(https://pubmed.ncbi.nlm.nih.gov/2038314/), PMID:23447528(https://pubmed.ncbi.nlm.nih.gov/23447528/)
The core uses eIF2alpha kinase activity, GCN2-mediated signaling and stress-dependent negative regulation of initiation in the cytosol. The two process terms have no ancestor/descendant relation in the inspected typed ontology closure. The phosphorylation-dependent initiation term describes the same mechanism already explicit in the catalytic function and narrative. Additional broad response labels, cellular phenotypes and organismal effects do not require separate catalytic cores.
All 29 ISS contexts have exact current term matches in the captured mouse Q9QZ05 or yeast P15442 donor annotations. Those joins identify the curator's evidence chain; they are not substitutions for the experimental scope. Human expression hosts are kept distinct from mouse constructs. In the antiviral study, the Northwestern probe is viral GAR RNA, while the tRNA experiment concerns activation. The viral-translation assertion is refined to negative regulation of viral translation, preserving the measured sign. PMID:16601681(https://pubmed.ncbi.nlm.nih.gov/16601681/)
Mouse GCN2 loss promotes neurite outgrowth in primary neurons and N2a cells, while the reported NeuN result limits a general neuronal-differentiation claim. Both neuronal source contexts therefore use negative regulation of neuron projection development. The memory study's direction depends on stimulation and training strength: weak protocols improve some outcomes after GCN2 loss, whereas stronger protocols expose impaired late plasticity and memory. These are retained as contextual mouse phenotypes, without claiming human behavioral assays. PMID:23447528(https://pubmed.ncbi.nlm.nih.gov/23447528/), PMID:16121183(https://pubmed.ncbi.nlm.nih.gov/16121183/)
The immune sources distinguish dendritic-cell signaling and antigen presentation from a T-cell-intrinsic function. Feeding behavior remains a curator-supported mouse context, with the incomplete recovered Methods disclosed. PMID:24310610(https://pubmed.ncbi.nlm.nih.gov/24310610/), PMID:16054071(https://pubmed.ncbi.nlm.nih.gov/16054071/)
The IPR016255 family match is real, and the conserved kinase function is supported. Its checkpoint premise is fungal, however, and the sources do not establish the same DNA-damage-checkpoint role for human EIF2AK4. The fungal response is stimulus-dependent. The mouse UV study demonstrates translational attenuation and discusses a possible cell-cycle consequence; it does not resolve the human checkpoint transfer. This annotation remains UNDECIDED without asserting absence of any mammalian checkpoint contribution. PMID:17369398(https://pubmed.ncbi.nlm.nih.gov/17369398/), PMID:19033384(https://pubmed.ncbi.nlm.nih.gov/19033384/), PMID:12176355(https://pubmed.ncbi.nlm.nih.gov/12176355/)
The distinct cold-response annotation also remains UNDECIDED. The PMID:25329545 abstract and a selected caption were accessible, but the operative cold assay was not recovered. Independent evidence supports the source's central kinase and amino-acid stress claims; general stress biology cannot verify the cold stimulus. The positive imported experimental assertion is preserved. PMID:25329545(https://pubmed.ncbi.nlm.nih.gov/25329545/)
The three official human products are 1,649, 1,621 and 616 residues. Q9P2K8-2 lacks canonical residues 774–801 within the active kinase domain. Q9P2K8-3 changes residues 608–616 and lacks residues 617–1649. The catalytic Asp848 position refers to canonical Q9P2K8-1. These sequence observations do not demonstrate equal activity, complete inactivity, or the identity of an unspecified experimental insert. All three original product objects are retained. UniProt Q9P2K8
Six IBA annotations trace to six positive IBD records on three ancestors of the exact human target sequence in a 1,076-node tree and 1,076-sequence alignment. The complete selected target lineage has no same-term loss record. Neither a short donor list nor target self-evidence is a defect. The nuclear placement is retained as conditional localization rather than discarded for lack of direct target microscopy. ARBA00028884 has three matching InterPro/FunFam condition sets among 144; its broad output is refined using the independent GCN2 mechanism. The portable evidence package reproduces these joins without computing biological decisions.
The four inspected GO-CAM target activities describe kinase activity in starvation or GCN2 signaling and distinguish GCN2 from downstream machinery. Their citation reuse does not reconstruct inaccessible assays. No additional process is proposed merely to complete a pathway diagram. The campaign's generic-binding policy remains applicable, although this source set contains no generic protein-binding annotations.
The source guide distinguishes normal-cache contents, separately recovered operative original text and unrecovered sections. No failed download is treated as an article. No figure-image or separately downloaded supplement interpretation is claimed. The configured research providers returned authentication/quota failures, so the review uses the cited primary, curated and computational evidence and these manual notes; no generated research report was fabricated. Publication-cache bytes and flags remain unchanged. No new publication quotation was introduced.
The evidence for mouse GCN2 in polysomal fractions is independent of the paper's emphasis on IMPACT. In PMID:23447528(https://pmc.ncbi.nlm.nih.gov/articles/PMC3624466/), the Polysome Profiles Methods describe N2a-cell sucrose gradients, and Figure 5C–F explicitly detects and quantifies GCN2, GCN1 and IMPACT separately. Panel E normalizes to input; panel F normalizes to polysomal rpS6. The selected Results and Discussion retain GCN2 association after the latter normalization. This supports the existing cytoplasmic and ribosome-association interpretations without asserting direct human localization, direct molecular contact or obligatory ribosomal membership. The accessible indexed Results contain a discontinuity; the complete caption provides an independent locator for the GCN2 measurements. No figure pixels or separately downloaded supplement were assessed.
The fungal checkpoint evidence is positive. PMID:17369398(https://genesdev.cshlp.org/content/21/6/649.full.html), Figure 3B–D, links Gcn2 to UV-induced eIF2alpha phosphorylation, translation attenuation and delayed pre-RC formation. PMID:19033384(https://doi.org/10.1242/jcs.035428), Figure 2A–E, reports a Gcn2-dependent G1 delay after hydrogen peroxide; Mcm6 loading distinguishes this from a later delay that remains in the mutant. Lack of a response to every damaging stimulus does not refute DNA-damage-checkpoint participation. GO:0000077 and its parents describe checkpoint signaling rather than a universal response to DNA damage. The human transfer remains UNDECIDED because conservation of this specific checkpoint function is unresolved, not because fungal biology is presumed irrelevant or because human inactivity has been established.
The optional availability field for PMID:24310610 was omitted to avoid collapsing partial access into a binary claim. Its normal cache remains marked full text, while its inspected contents omit Methods and separately published supplements. Conversely, PMID:23447528 remains an unchanged abstract-only cache even though selected original sections were separately accessed. These access distinctions and precise primary-source locators are recorded in the source guide.
The existing cytoplasm assertions remain compatible with the more specific cytosol core location. Ribosome association stays non-core; no new direct-binding activity is inferred from fractionation. The dendritic-cell experiments in PMID:24310610 include measured downstream antigen-specific T-cell responses, so the existing contextual immune annotation is retained without inferring a T-cell-intrinsic mechanism or adding a new process annotation. All 52 actions, the three products and the single kinase core are unchanged. Reference-only links improve auditability under the standing zero-new-quotation constraint; no supporting-text values or publication-cache edits were introduced.
Uncharged-tRNA sensing and eIF2alpha phosphorylation are distinct molecular activities of the same GCN2 sensor-kinase. The core synthesis now represents both activities, rather than requiring the tRNA-binding input to be implicit in the catalytic output. The conserved tRNA-binding assignment combines the mouse donor annotation, the tRNA-dependent activation and HisRS-region controls in PMID:16601681, and the curated UniProt FUNCTION/DOMAIN statements. The latter are explicitly by similarity (ECO:0000250), and the original direct RNA-probe assay used viral RNA. No purified human tRNA-binding assay or equivalence of the three natural splice products is asserted. PMID:16601681(https://pmc.ncbi.nlm.nih.gov/articles/PMC1440839/); curated record
Amino-acid starvation is now explicit in the sensing core. Human SQ20B GCN2 depletion under leucine limitation reduces eIF2alpha phosphorylation and p21 induction (PMID:26102367, Figure 1E), separately from that paper's mouse reporter and polysome experiments. The kinase core also records activation-related autophosphorylation. Its conserved evidence is distinguished from a new human Thr899 measurement: UniProt's PTM statement is by similarity. These changes retain all existing actions. PMID:26102367(https://doi.org/10.1371/journal.pgen.1005212); PMID:16601681(https://pmc.ncbi.nlm.nih.gov/articles/PMC1440839/)
The two MF terms are not ancestors or descendants of each other. Official QuickGO is_a/part_of closures checked on 2026-10-27 also show no hierarchy between the four selected core processes (GCN2 signaling, stress-dependent inhibition of initiation, amino-acid-starvation response and autophosphorylation). GO:0004694 already reaches GO:0010998 through part_of, so the phosphorylation-dependent initiation term is not repeated as another core process. This is a refinement of core coverage, not a new imported annotation.
The nuclear IBA remains positive phylogenetic evidence: the literal IBD at PTN000113601 lies on the verified human target lineage. Its depth and sibling-clade donors are not evidence that GCN2 lost the localization, and cytoplasmic localization is not nuclear exclusion. Nuclear residence remains a non-core inference rather than a claim of observed human microscopy. Ribosome association likewise retains positive yeast and mouse evidence, particularly the separately quantified GCN2 signal in PMID:23447528 Figure 5C–F; preparation-specific non-detection in other studies does not erase it. No obligatory ribosome membership or new binding function is asserted. PAINT evidence; PMID:23447528(https://pmc.ncbi.nlm.nih.gov/articles/PMC3624466/)
UniProt's DNAJC3 interaction under hypothermic stress is weighed as a positive by-similarity statement. The unavailable cold-response experiment remains unresolved, as does transfer of the fungal checkpoint role to human EIF2AK4. The previous journal entries remain intact. This follow-up preserves 53 raw GOA rows, their 52 distinct imported contexts, all existing reference objects, three natural splice records and zero publication quotations; it adds only an explicit reference to the unchanged curated UniProt file.