CIITA supplies transcriptional coactivation rather than sequence-specific promoter recognition. Its interactions with RFX and the basal factor TAF9 link promoter-bound regulatory machinery to MHC transcription. Human complementation, promoter assays and interaction experiments support MHC class II biosynthesis as the dominant role, with positive MHC class I regulation also directly tested in the NLRC5 study. A single organizing coactivator core retains both contexts (PMID:8402893(https://pubmed.ncbi.nlm.nih.gov/8402893/), PMID:9171108(https://pubmed.ncbi.nlm.nih.gov/9171108/), PMID:10938133(https://pubmed.ncbi.nlm.nih.gov/10938133/), PMID:20639463(https://pubmed.ncbi.nlm.nih.gov/20639463/)).
CIITA also has directly measured biochemical capacities. The 2013 study tests phosphorylation of H2B Ser36, TAF7 and RAP74, with matched extract controls, in-gel denaturation/renaturation, a human-cell preparation, contaminant assessment and a TAF1-negative immunoblot. Its Methods, Results and supplementary legends also describe acetyl-CoA-dependent H3/H4 acetylation. These observations support the existing kinase annotations and one new histone acetyltransferase annotation. They do not demonstrate the complete proposed sequence of TAF7 release and transcription initiation in living cells. The original 2001 acetyltransferase article remains abstract-only in available access; the later actual assay supplies independent support (PMID:24036077(https://pubmed.ncbi.nlm.nih.gov/24036077/), PMID:11172716(https://pubmed.ncbi.nlm.nih.gov/11172716/)).
The original GTP-binding study was recovered only as journal page 1402 after three pages of an unrelated article. The page has the matching title, authors and complete Figure 1 legend. In-vitro-translated and immunoprecipitated CIITA is tested for GTP binding with RasQ61L, NF-YA, motif-mutant and protein-expression controls; COS7 nucleotide labeling is a separate assay. Independent review inspected the actual figure pixels. This supports one new GTP-binding MF, without claiming that the uninspected Methods footnotes or later nuclear-import Results were read. Nucleotide labeling alone does not establish GTP-versus-GDP selectivity, and no GTPase or new import-process annotation is proposed (PMID:10464099(https://pubmed.ncbi.nlm.nih.gov/10464099/)).
Both new molecular functions were compared with the original terms using typed ontology ancestry. Neither adds an ancestor or descendant of existing coverage, and the two proposed MFs are not ancestors of one another. No new process is proposed. The exact selected GO-CAM index has no CIITA activity hit; absence is not used as a reason to manufacture process coverage.
All 47 original source objects, including the NOT DNA-binding assertion, qualifiers, evidence codes and supporting entities, are preserved. The four fetched products retain their exact identities and sequence notes. Full-length 1130-residue constructs and tested fragments are described at the source’s scope; they are not silently mapped to every current product. The 2013 source inconsistently calls Sf9 cells Drosophila in its Results, while its Methods identify Spodoptera. The latter is the correct host identity, and no source typo is propagated.
Supported generic binding is retained as non-core under the campaign policy. More precise classes are used when the demonstrated partner and actual term definition support them: RFX5 binding is regulatory Pol II transcription-factor binding, SIRT1 binding is histone deacetylase binding, and the TAFII32/TAF9 assay is basal Pol II initiation-factor binding. The TAF9 refinement does not assert that TAF9 can never contact DNA. ZXDA–ZXDC experiments include complex-dependent association rather than uniformly purified binary affinity, and their zinc fingers or promoter occupancy do not by themselves establish sequence-specific DNA recognition (PMID:16600381(https://pubmed.ncbi.nlm.nih.gov/16600381/), PMID:17493635(https://pubmed.ncbi.nlm.nih.gov/17493635/), PMID:21890893(https://pubmed.ncbi.nlm.nih.gov/21890893/)).
The collagen/HDAC2 study is available as a complete authentic abstract and a publisher culture-system preview, but full Results and panels were not recovered. The abstract positively reports CIITA–HDAC2 co-immunoprecipitation, deacetylation, altered promoter recruitment, reporter activity, quantitative PCR and RFX5 association. Those observations support the retained context-dependent repression and binding annotations, with exact product-to-cell assignments left unspecified. HDAC2 antagonizes CIITA; it is not demonstrated to be a recruited enzyme executing CIITA-mediated collagen repression (PMID:19041327(https://pubmed.ncbi.nlm.nih.gov/19041327/)).
PML-body localization is retained as a contextual pool alongside diffuse nuclear CIITA. The PML partner isoform II remains explicit. Selected HPA HEL and U251MG images independently support nuclear enrichment, with an additional HEL cytosolic pool. Reactome explicitly places its human CIITA output entity in nucleoplasm; the event does not make CIITA a catalyst of its own production (PMID:23007646(https://pubmed.ncbi.nlm.nih.gov/23007646/), Reactome:R-HSA-1031716).
The antiviral study distinguishes CIITA-dependent transcription from downstream CD74 p41 protease-related effects. The newer HCMV study independently corroborates CIITA-dependent HLA-II expression and shows viral pUS28 antagonism, including reciprocal association after lysate mixing; neither source makes CIITA a direct antiviral protease inhibitor (PMID:32855215(https://pubmed.ncbi.nlm.nih.gov/32855215/), PMID:40608405(https://pubmed.ncbi.nlm.nih.gov/40608405/)).
The original PMID:107465 is preserved exactly, but official MGI reference J:107465 resolves to PMID:16280321. The exact MGI Ciita GO:0046677 ISO record uses that MGI reference and human P33076. This is an authenticated namespace correction, not a guess from a paper title. The original PMID itself concerns cerebellar glia.
Selected actual Methods, Results and the Figure 2 legend of the corrected source show that Hsp90 inhibition suppresses IFN-induced CIITA transcripts and promoter activity through upstream signaling. The GO response-to-antibiotic definition explicitly includes changes in gene expression. The curation concern is therefore participation by the regulated CIITA protein, not exclusion of transcriptional responses from the term. Expression of CIITA is the response readout in this experiment; it does not alone establish that CIITA performs the antibiotic-response step. The source-specific assertion is marked over-annotated without claiming a universal negative biological capability (PMID:16280321(https://pubmed.ncbi.nlm.nih.gov/16280321/)).
The genuine normal intake and publication retrievals are preserved. Falcon timed out after 90 seconds, and the configured Perplexity fallback returned HTTP 401 for insufficient quota. No provider output was fabricated. Manual primary research and two bounded independent consultations supplied the source assessment.
Current donor records and matching ARBA conditions are authenticated. The curated PAINT node is supported by actual target biology, while its historical tree, MSA and IBD placement remain uninspected. Neither donor count nor target self-inclusion is treated as a propagation failure. The compact source-evidence file holds record identifiers, URLs, hashes, definitions and scoped provenance; complete primary responses and literal donor records remain in the temporary audit packet. File citations contain no self-quoted authored claims. Primary quote strings are allocated only in the review YAML and are not repeated in this notes file or artifact metadata.
Normal validation, calculated status and rendering are run in the isolated genuine tooling workspace before the packet is sealed. Publication and history creation remain separate from this temporary review packet; campaign completion requires the qualifying review and merge.