DPF2 (Q92785; HGNC:9964), also called BAF45d or REQ, is a histone-recognition and transcription-regulatory subunit associated with SWI/SNF complexes. The review preserves all 52 imported assertions, including the NOT assertion, and distinguishes positive observations from uncertainties in their source-specific interpretation. There are no proposed NEW annotations. The two core units summarize histone recognition/recruitment and transcription-factor-associated regulation, not two unrelated proteins or an autonomous DPF2 ATPase.
Human DPF2 residues 270–391 bind H3K14-crotonylated peptides with a reported affinity of 85 nM and approximately eightfold stronger affinity than the acetylated comparison. Designer-nucleosome pulldown enriches endogenous DPF2 from HeLa S3 nuclear extract. The DPF2 structure in this study is apo; ligand-bound MOZ structures and MOZ cellular ChIP are separate experiments and are not assigned to DPF2. PMID:27775714(https://pubmed.ncbi.nlm.nih.gov/27775714/)
The 2017 human PHD-tandem ITC measurements do not show an intrinsic acetyl preference under the reported conditions: H3 and H3K14ac have approximately 1.0 and 1.1 micromolar dissociation constants, while H4 and H4K16ac have approximately 30.2 and 50.4 micromolar values. This does not erase binding to the acetylated substrates. Cellular histone association responds to acetylation conditions, and a folding-controlled pocket mutant loses target-chromatin recruitment and the ability to inhibit myeloid differentiation. The reader annotations mean recognition; they are not claims of exclusive binding or a universal affinity ranking. The abstract-only machine cache is preserved; the complete Caltech original and supplement were independently examined, including the ITC tables. PMID:28533407(https://pubmed.ncbi.nlm.nih.gov/28533407/)
DPF2 also associates with RUNX1 in the myeloid context. Full-length protein supports that association, whereas the isolated PHD tandem does not. RUNX1 methylarginine peptides did not bind in the examined assays, so a methylarginine-reader activity is not inferred. The RUNX1 partner class supports DNA-binding transcription factor binding without pretending that co-immunoprecipitation supplies a purified affinity or a resolved contact interface. PMID:28533407(https://pubmed.ncbi.nlm.nih.gov/28533407/)
A separate human-cell study maps important p52 and SWI/SNF association determinants to the N-terminal 33–84 region. GST proteins were combined with in-vitro-translated partners in reticulocyte or wheat-germ lysates, rather than a completely purified binary system. Integrated reporters, lymphotoxin-responsive BLC expression in HT29 cells, promoter recruitment and nuclear co-immunoprecipitation support cooperation with RelB/p52 and Brm-type SWI/SNF. The absence of an effect on the transient reporter and on p100 processing helps delimit the mechanism. The 2010 PHD-to-nucleosome bridging model was a proposal; later direct histone assays supply the separate reader evidence. This source prevents an account of DPF2 as uniformly repressive. PMID:20460684(https://pmc.ncbi.nlm.nih.gov/articles/PMC2903353/)
The imported NOT H3K9me2/me3-reader row remains UNDECIDED. In the actual 2018 Results and Figure 3A, wild-type GST-DPF2 binds H3K9me2 and H3K9me3 peptides; pathogenic PHD variants weaken or abolish several interactions. A mutant defect is not wild-type absence. Conversely, the positive peptide bands do not alone establish selective methylmark reading in physiological chromatin. The accepted manuscript, relevant Methods and actual Figure 3 image were inspected; neither a new positive selective-reader row nor a confident categorical negative is manufactured. PMID:29429572(https://pubmed.ncbi.nlm.nih.gov/29429572/)
Wild-type GFP-DPF2 is diffusely nuclear in the tested human HEK293 cells, whereas some PHD variants form nuclear aggregates that recruit wild-type DPF2 and BRG1. FLAG-tagged and COS7 experiments are separate controls. These engineered disease variants are not natural isoform 2. PMID:29429572(https://pubmed.ncbi.nlm.nih.gov/29429572/)
Official UniProt sequence features and both isoform FASTAs agree on Q92785-1, 391 residues, and Q92785-2, 207 residues. The shorter product deletes canonical residues 156–339, removing the C2H2 finger, all of PHD1 and the beginning of PHD2. This architecture difference does not by itself prove total inactivity, nor does retention of the N terminus prove a particular interaction. Canonical PHD-fragment experiments cannot be transferred wholesale to the short natural product. UniProt Q92785
The HPA target identity is ENSG00000133884/Q92785. Supported ICC/IF records show nucleoplasm with both inspected antibodies. HPA020880 additionally reports cytosol and centrosome in U-251MG cells and vesicles in U2OS cells; HPA056786 reports cytosol in KOLF2.1J cells. These are curated records and captions, not a new image-pixel analysis or an isoform-resolved experiment. The exported vesicle label carries the broad source identifier GO:0043231; the review selects the verified vesicle term without guessing an endosomal or lysosomal subtype. HPA subcellular record
The nucleus refinements retain both nucleoplasm and chromatin; they do not invent a chromatin-to-nucleus ontology edge. The cytoplasm refinements retain cytosol, centrosome and vesicles as nonexclusive finer pools. These are target-level refinements using independent evidence, not claims that every older source observed every pool. The IEA mouse location is itself by similarity to the experimentally supported human record and is not counted as independent localization evidence.
All six nucleoplasm Reactome annotations have inspected positive paths to DPF2 entity R-HSA-9727509, Q92785 residues 2–391. The paths traverse actual component or candidate sets in the modeled MITF-M/cBAF/npBAF events. This establishes the selected member/location paths, not exhaustive pathway traversal, identical roles for every candidate, or DPF2 ATPase activity. Reactome DPF2 entity
Exact target, partner and PMID joins were checked for all interaction rows. The 2014 binary-screen procedure required reproducible pairwise retesting and identity confirmation; HuRI used quadruplicate retesting and sequence confirmation. Dataset-level orthogonal validation is not asserted for each DPF2 pair individually. All applicable IntAct reference-sequence ROGIDs were reconstructed from sequence and taxon. The four 2017 complex-associated pairs are supported by the original co-immunoprecipitation experiments even though they were not recovered in the current IntAct query. PMID:25416956(https://pubmed.ncbi.nlm.nih.gov/25416956/), PMID:32296183(https://pubmed.ncbi.nlm.nih.gov/32296183/), PMID:28533407(https://pubmed.ncbi.nlm.nih.gov/28533407/)
The HuRI MEOX2 accession is Q6FHY5, not silently replaced by P50222. It is 303 residues and differs from the 304-residue reviewed comparator by one upstream glutamine deletion; the homeobox sequences are identical. GOLGA6L9 has a recorded sufficient-binding fragment at residues 159–432. Reference-sequence authentication does not independently resolve every experimental insert endpoint. LNX1 binding is refined to ubiquitin protein ligase binding, and TXNDC12 to enzyme binding; neither refinement transfers catalysis to DPF2. Positive generic associations are retained as non-core under the scoped project instruction.
The current PTHR45888 tree contains the DPF2 leaf PTN002472020, and the ungapped target MSA sequence exactly matches all 391 canonical residues. PTN008700032 is an actual ancestor for the broad coregulator/transcription assertions, with experimental descendant grounding. PTN002578083 is an actual ancestor for the myeloid-regulation assertion, and the human target itself supplies valid experimental grounding. A target occurring in its own WITH/FROM is not circular evidence. An initial accession-string MSA lookup was corrected to the resource's persistent-node key; no missing-sequence or phylogenetic-loss conclusion was drawn from that parser limitation.
The foundational apoptosis experiment uses mouse Requiem antisense suppression in IL-3-dependent FDCP-1 cells. Stable-transfection and conditioned-medium controls support the survival result, but they do not identify an autonomous human DPF2 execution reaction. The existing apoptosis observations are retained as non-core context; no new process is inferred from necessity alone. PMID:7961935(https://pubmed.ncbi.nlm.nih.gov/7961935/)
The older G0/G1, G1/S, stem-cell-maintenance, metaphase/anaphase, double-strand-break repair and nucleotide-excision-repair assertions remain UNDECIDED where the accessible evidence does not resolve DPF2's particular contribution. The 2014 main original establishes PBAF/BRG1/BAF180 repair biology, but the requested supplement remained inaccessible. The 2009 esBAF cache has a full-text flag yet contains abstract and discussion without operative Results/Methods. These limits are not evidence of a wrong gene, absence of a role, or a curator error. The reviewed chromatin-remodeling participation is independently supported by DPF2 recognition/recruitment and complex association. PMID:25066234(https://pmc.ncbi.nlm.nih.gov/articles/PMC4157577/), PMID:19279220(https://pubmed.ncbi.nlm.nih.gov/19279220/)
The ordinary gene/publication bootstrap completed in an isolated workspace. The genuine Falcon research attempt failed authentication and the fallback failed with an insufficient-quota response; no provider report was fabricated. Manual source work used authenticated originals, cached publications, official ontology/data records and explicit access limits. HTTP 200 challenge pages, metadata-only responses and BioC error messages were not treated as full text. Machine-generated source files and all preexisting source bytes are preserved.
Five short excerpts occur once each in the review YAML: 15 words from PMID:27775714, 7 from PMID:20460684, 15 from PMID:28533407, 6 from PMID:29429572 and 13 from PMID:7961935. These notes do not repeat them. The accompanying source-evidence file supplies public locators and exact original-read windows. Final ordinary repository validation, status normalization and publication are separate from the saved-input scientific checks.
The acetyl-reader annotations are retained. Their current definitions specify recognition of the modified histone, without an affinity-preference threshold. Human DPF2 peptide assays and histone-pocket-dependent cellular recruitment support that recognition; the 2017 measurements do not show preferential binding to the acetylated peptides. SAHA-sensitive association provides cellular context but does not by itself resolve H3K14 or H4K16 as the causal mark. The H3K9me2/3 NOT assertion remains uncertain because the cited wild-type protein binds those peptides. No positive methyl-reader annotation is introduced.
The HPA antibody records explicitly report centrosome and vesicle pools. Inspection of the saved HPA020880 U-2 OS image places the annotated puncta within the cytoplasm, supporting cytoplasmic vesicle in the complementary location set. This does not identify a vesicle subtype, trafficking role or natural product. The data are available on the HPA DPF2 subcellular page and the inspected image. Absence from one GOA export is not treated as a curator decision against independently documented localization.
The broad RNA polymerase II regulation row is refined to both activating and repressive regulation. In human cells, DPF2 links the RelB/p52 and Brm systems, is recruited to a responsive promoter and increases integrated-reporter output; this is evidence of a performed coregulatory role beyond loss-of-function necessity. Separate human myeloid experiments support repression and histone-pocket-dependent chromatin recruitment. These signed children replace the broad process in the existing row and core; no NEW process row is added. Source-specific assay and cell-context limits remain in the annotation reasons.
A proposed H3K14cr reader term records the direct crotonylated-peptide and nucleosome recognition evidence. The broad histone-reader molecular function remains an appropriate synthesis of multiple recognition activities. The schema does not allow both molecular_function and proposed_molecular_function in one core, so the ontology proposal is recorded at top level without fabricating a GO identifier or replacing a valid broad core. No new quoted words are introduced; all existing quotations remain in their original locations.
