CDKN2A

Summary: ARF promotes COMMD1 polyubiquitination in the tested human-cell context.

Reason: Selected full primary Results show ARF-dependent COMMD1 polyubiquitination and an ARF binding-deficient deletion that loses this effect. Its C terminus contributes beyond simple binding. Retain non-core participation in this complex-dependent modification; the experiment leaves recruitment versus intrinsic E3 mechanism open and does not assign ARF a ubiquitin-transferase MF.

Summary: CARF associates with human ARF.

Reason: The source abstract explicitly reports human as well as mouse ARF association with CARF and a soluble nucleoplasmic complex. Retain the supported Q9NXV6 association under the project binding instruction, without importing CARF transcriptional or catalytic activity or claiming a newly inspected purified interface. Supported generic association is retained under the explicit ClinGen project instruction.

Summary: MDM2 binding can be represented by a ligase-specific binding term.

Reason: The exact source-linked IntAct record is fluorescence microscopy, which alone does not prove a binary interface. Independent human ARF–MDM2 binding and functional experiments in PMID:9724636 support the known association and this partner-class refinement. Preserve the original source assertion and distinguish that independent evidence from the original imaging assay.

Proposed replacements: ubiquitin protein ligase binding

Summary: ARF associates with the Epstein–Barr virus EBNA-LP protein.

Reason: Exact positive IntAct records tied to this paper include Y2H and pull-down experiments for Q8N726 and Q8AZK7, with tagged constructs and N-terminal ARF mapping. Keep the viral association as non-core; imaging is complementary context, and the record does not establish an endogenous constitutive host complex. Supported generic association is retained under the explicit ClinGen project instruction.

Summary: ARF–NPM1 association is supported without a new NPM-derived activity.

Reason: The original curated P06748 pair is retained in light of independent target-specific NPM/ARF co-immunoprecipitation in PMID:16957780 Figure 4C. The selected PMID:15989956 abstract does not independently enumerate this NPM1 pair, so its exact purification evidence is not claimed newly verified. Neither NPM1 enzymatic activity nor its RNA-binding role is transferred to ARF. Supported generic association is retained under the explicit ClinGen project instruction.

Summary: ARF binds the ubiquitin ligase HUWE1/ARF-BP1.

Reason: The original abstract explicitly identifies ARF-BP1 as a HECT ligase bound and inhibited by ARF; indexed primary Figure 2 describes GST-ARF fragment mapping against translated ARF-BP1. Refine the source binding assertion to the verified partner class. Binding scope is separate from any claim that ARF itself catalyzes ubiquitin transfer.

Proposed replacements: ubiquitin protein ligase binding

Summary: Human ARF associates with NPM1 in the source antiviral experiments.

Reason: Selected original Results and Figure 4C describe NPM immunoprecipitation followed by ARF detection in inducible human NARF-E6 cells. Other experiments use mouse p19ARF, which is kept distinct. Retain this source-supported association without assigning PKR kinase activity or a new antiviral process to ARF. Supported generic association is retained under the explicit ClinGen project instruction.

Summary: NIAM/TBRG1 is an ARF-associated nuclear protein.

Reason: The source abstract identifies NIAM as an ARF- and MDM2-binding protein in the nucleus. Keep the curated Q3YBR2 association under the project binding instruction; the abstract does not establish a purified binary interface, and NIAM chromosomal-stability phenotypes are not assigned to ARF. Supported generic association is retained under the explicit ClinGen project instruction.

Summary: ARF associates with the DNA-binding transcription factor MYCN.

Reason: The source abstract explicitly reports human ARF interaction with N-Myc, mapped to N-Myc residues 140–300, and nucleolar redistribution. MYCN is the assayed sequence-specific transcription factor; the replacement describes ARF binding that partner, not ARF acquiring MYCN DNA-binding or transcriptional activity.

Proposed replacements: RNA polymerase II-specific DNA-binding transcription factor binding

Summary: ARF binds C1QBP/p32 with construct-specific evidence.

Reason: PMID:17486078 Figure 1b uses full human GST-p14ARF and translated p32; Figure 1f tests tagged human short ARF lacking residues 1–47. Endogenous short-form co-IP was unresolved. Mouse fractionation experiments and possible post-lysis full-length mouse association are kept separate; the human binding evidence supports non-core retention. Supported generic association is retained under the explicit ClinGen project instruction.

Summary: The source supports association with MDM2.

Reason: The exact positive human Q8N726–Q00987 co-immunoprecipitation record is linked to PMID:17909018. The paper title foregrounds a p16 mutation, but the curated ARF assay is not discarded on that basis. Refine the verified ligase partner class without importing p16 CDK inhibition.

Proposed replacements: ubiquitin protein ligase binding

Summary: ARF directly contributes an interaction with COMMD1.

Reason: Selected full primary Figure 1 and Results identify human ARF in Y2H, reciprocal co-IP and deletion mapping; endogenous association is detected after actinomycin D treatment. Retain the contextual COMMD1 association without asserting that co-IP establishes an invariant stoichiometric complex. Supported generic association is retained under the explicit ClinGen project instruction.

Summary: ARF associates with MDM2 in the DAXX study.

Reason: The original abstract reports HDM2 interaction with ARF and DAXX and its effects on their localization. The specific ligase-binding term captures the ARF partner class; the evidence does not imply ARF has DAXX chromatin or HDM2 catalytic activity.

Proposed replacements: ubiquitin protein ligase binding

Summary: ARF associates with DAXX.

Reason: Coexpression, co-immunoprecipitation and endogenous nuclear/nucleolar association are explicitly described in the source abstract. Retain this interaction as non-core, without converting DAXX partner identity into an ARF chromatin-modifying activity. Supported generic association is retained under the explicit ClinGen project instruction.

Summary: ARF binds the Pol II DNA-binding transcription factor EGR1.

Reason: The source connects ARF to EGR1 SUMO regulation; the existing specific EGR1-binding assertion from this same source also carries GO:0061629. Use that correct partner-class refinement while retaining the original P18146 source and distinguishing binding from EGR1 transcriptional activity.

Proposed replacements: RNA polymerase II-specific DNA-binding transcription factor binding

Summary: ARF binds MDM2 in a phosphorylation-dependent regulatory context.

Reason: The source abstract reports that Pim-mediated MDM2 phosphorylation enhances its ARF interaction. Refine to ligase binding; Pim performs the kinase reaction, and this observation does not assign kinase activity to ARF.

Proposed replacements: ubiquitin protein ligase binding

Summary: ARF–TRIP12/ULF association is retained with an explicit editorial concern.

Reason: The 2026 Expression of Concern names blot issues in Figures 1e, 3g and 4c of PMID:20208519; unavailable original data limits confidence. Its Figure 2 binding assays were not specifically named. Independent PMID:27323397 Figure 6B also detects ARF in ULF immunoprecipitates. Retain only the low-resolution association; do not treat the disputed regulatory claims as settled or call the original article retracted. Supported generic association is retained under the explicit ClinGen project instruction.

Summary: ARF–MDM2 binding is a supported ligase-partner function.

Reason: The original Q00987 assertion is retained with independent human ARF–MDM2 evidence in PMID:9724636. The selected PMID:20713054 abstract principally addresses APAK and does not by itself resolve its exact MDM2 control assay. This refinement uses the established target function, not an invented original assay or an APAK-derived activity.

Proposed replacements: ubiquitin protein ligase binding

Summary: ARF binds APAK/ZNF420 in the oncogenic-stress study.

Reason: The source abstract reports ARF-dependent binding and sequestration of APAK that competes with its association with p53. Keep the Q8TAQ5 interaction as non-core; binding alone does not establish a specific ARF catalytic or DNA-binding activity. Supported generic association is retained under the explicit ClinGen project instruction.

Summary: ARF binds the sequence-specific transcription factor SP1.

Reason: The original abstract describes ARF interaction with SP1 in regulation of MTA1 expression. Refine the binding partner class without transferring SP1 DNA recognition or transcription-factor activity to ARF.

Proposed replacements: RNA polymerase II-specific DNA-binding transcription factor binding

Summary: ARF associates with MTA1.

Reason: The source abstract reports the ARF–MTA1 interaction in addition to effects on SP1 and MTA1 abundance. Retain non-core association, without assigning ARF the molecular functions of an MTA1-containing chromatin complex. Supported generic association is retained under the explicit ClinGen project instruction.

Summary: The quantitative interaction study recovers ARF with HUWE1.

Reason: The exact positive human Q8N726–Q7Z6Z7 IntAct record is an anti-tag co-immunoprecipitation experiment linked to PMID:26496610. The tagged screen supports association with an E3 ligase, not a new direct binary interface or an inhibition assay. Refine to the partner class and preserve the original accession and screen context.

Proposed replacements: ubiquitin protein ligase binding

Summary: ARF binds NOP53/GLTSCR2.

Reason: The full source Figure 1 includes endogenous association, reciprocal tagged co-IP and a recombinant pull-down using bacterially expressed proteins. Figure 2 maps the ARF C-terminal region. Keep this specific association as non-core without transferring NOP53 functions to ARF. Supported generic association is retained under the explicit ClinGen project instruction.

Summary: Nuclear localization is part of the ARF inhibitory mechanism.

Reason: The human oligodendroglioma immunohistochemistry source reports nuclear p14ARF staining. This is consistent with the independently supported nuclear MDM2-regulatory core; it does not establish that all six externally listed products share this localization.

Summary: The HPA nucleoplasmic assignment is consistent with the core location.

Reason: Preserve the source immunofluorescence assertion as a core-compatible location, supported independently by human ARF nucleoplasmic association and trafficking experiments. This located_in statement does not claim the antibody distinguishes every ARF product.

Summary: Nucleoplasmic ARF is supported in the NIAM study.

Reason: The source describes an ARF-associated nuclear NIAM pool. Retain the finer curator-assigned nucleoplasm location, coherent with the nuclear inhibitory core and independent human trafficking evidence; do not derive NIAM chromosomal stability as an ARF core process.

Summary: The COMMD1 experiments identify a nucleoplasmic ARF pool.

Reason: Selected full Results and Figure 5 describe actinomycin D-induced ARF redistribution and endogenous COMMD1 association. The regulated pool is consistent with the nucleoplasmic core location, without implying that untreated ARF is uniformly nucleoplasmic.

Summary: The UniProt nucleoplasm mapping is biologically coherent.

Reason: The vocabulary mapping preserves a located_in assertion. Human primary ARF trafficking and binding evidence supports this compartment in the core mechanism; no new cell type or isoform specificity is inferred.

Summary: The curated event supports core-compatible nucleoplasmic residence.

Reason: The authentic Reactome event concerns MDM2/ARF-dependent TP53 SUMO modification. Retain its source located_in assertion as consistent with independently supported nucleoplasmic ARF. This location decision neither endorses every catalytic interpretation of the event nor assigns the partner complex activities to ARF.

Summary: The curated event supports core-compatible nucleoplasmic residence.

Reason: The authentic Reactome event concerns ARF-associated WRN SUMO modification. Retain its source located_in assertion as consistent with independently supported nucleoplasmic ARF. This location decision neither endorses every catalytic interpretation of the event nor assigns the partner complex activities to ARF.

Summary: The curated event supports core-compatible nucleoplasmic residence.

Reason: The authentic Reactome event concerns MDM2 sequestration by p14ARF. Retain its source located_in assertion as consistent with independently supported nucleoplasmic ARF. This location decision neither endorses every catalytic interpretation of the event nor assigns the partner complex activities to ARF.

Summary: The curated event supports core-compatible nucleoplasmic residence.

Reason: The authentic Reactome event concerns the ARF–MDM2–TP53 association. Retain its source located_in assertion as consistent with independently supported nucleoplasmic ARF. This location decision neither endorses every catalytic interpretation of the event nor assigns the partner complex activities to ARF.

Summary: The curated event supports core-compatible nucleoplasmic residence.

Reason: The authentic Reactome event concerns p14ARF nuclear translocation. Retain its source located_in assertion as consistent with independently supported nucleoplasmic ARF. This location decision neither endorses every catalytic interpretation of the event nor assigns the partner complex activities to ARF.

Summary: The curated event supports core-compatible nucleoplasmic residence.

Reason: The authentic Reactome event concerns the p14ARF interaction with the NFE2L2 coactivator complex. Retain its source located_in assertion as consistent with independently supported nucleoplasmic ARF. This location decision neither endorses every catalytic interpretation of the event nor assigns the partner complex activities to ARF.

Summary: Nucleolar ARF is a core location in the NIAM context.

Reason: Retain the experimental source localization, consistent with ARF sequestration and trafficking in its nuclear regulatory mechanism. The NIAM interaction does not require ARF to be exclusively nucleolar.

Summary: The primary COMMD1 experiments support nucleolar ARF before redistribution.

Reason: The selected full source localization experiments distinguish a nucleolar ARF pool from its nucleoplasmic redistribution after actinomycin D. Retain the normal nuclear subcompartment without erasing that regulated trafficking.

Summary: The PANO study supports nucleolar ARF.

Reason: The source abstract explicitly describes nucleolar p14ARF in human HeLa cells. This reinforces the nuclear regulatory core without assigning PANO stabilization activity to ARF or claiming a new inspection of all source imaging panels.

Summary: Endogenous human ARF is detected in nucleoli.

Reason: PMID:27323397 Figure 1A stains endogenous ARF and GLTSCR2 in HeLa nucleoli; subsequent experiments follow their redistribution. Retain the source location as core-compatible, with trafficking rather than exclusive residence.

Summary: The UniProt nucleolus mapping matches direct human evidence.

Reason: Retain the located_in mapping in agreement with endogenous human ARF immunostaining and nucleolar/nucleoplasmic trafficking. This does not extend nuclear localization to every shorter or externally listed product.

Summary: Mitochondrial localization is supported for tested ARF products.

Reason: Selected original PMID:16713577 Results distinguish human and mouse internally initiated products; PMID:17486078 independently tests human short ARF. Later human full-length ARF mitochondrial evidence also exists. Retain the broad compartment while preserving construct dependence and reports of additional nucleolar localization.

Summary: The broad mitochondrial vocabulary mapping is supported.

Reason: Human short and full-length ARF have mitochondrial evidence in particular experimental contexts. Product- and cell-dependent nuclear localization prevents treating mitochondria as the exclusive or universal location of Q8N726 products.

Summary: Mitochondrial matrix precision remains unresolved.

Reason: Reactome genuinely places the ARF–C1QBP complex in the matrix. The inspected human primary fractionation, microscopy and protease-protection experiments support mitochondrial or intramitochondrial ARF but do not resolve matrix from intermembrane space. Preserve the curated source while withholding this finer compartment; this is not a wrong-product claim.

Summary: The source provides a bounded cytosolic context.

Reason: The translated 132-residue ARF output is placed in cytosol; this does not make ARF the regulator of its own translation. Retain the original located_in assertion without elevating this event-specific context to a separate core cytosolic activity.

Summary: The source provides a bounded cytosolic context.

Reason: Cytosolic ARF is the input to nuclear translocation, so this is a transit location. Retain the original located_in assertion without elevating this event-specific context to a separate core cytosolic activity.

Summary: The source provides a bounded cytosolic context.

Reason: The ARF–C1QBP complex is in cytosol before its modeled mitochondrial translocation. Retain the original located_in assertion without elevating this event-specific context to a separate core cytosolic activity.

Summary: The source provides a bounded cytosolic context.

Reason: The event models cytosolic ARF–C1QBP complex formation, rather than a separate localization assay. Retain the original located_in assertion without elevating this event-specific context to a separate core cytosolic activity.

Summary: The source provides a bounded cytosolic context.

Reason: The event concerns C1QBP-binding-deficient ARF mutants; their defect is not generalized to wild-type ARF. Retain the original located_in assertion without elevating this event-specific context to a separate core cytosolic activity.

Summary: The source provides a bounded cytosolic context.

Reason: The V22Pfs*46 mutant fails nuclear translocation; this does not imply that normal ARF remains exclusively cytosolic. Retain the original located_in assertion without elevating this event-specific context to a separate core cytosolic activity.

Summary: ARF can promote apoptosis in defined contexts.

Reason: The ARF-family electronic mapping is consistent with human p14ARF gain-of-function and perturbation evidence. Apoptosis varies with cell state and can occur through p53-dependent and independent mechanisms, so keep the broad process non-core rather than make every ARF product constitutively apoptotic.

Summary: Human p14ARF restrains proliferation in the source positive-control experiments.

Reason: Selected original PMID:17569660 Figure 9 and BrdU Methods test p14ARF in HCT116 cells alongside MEG3, including p53-deficient comparisons. The paper title is not a reason to discard its ARF control experiments. This supports the core antiproliferative outcome without assigning MEG3 RNA activity to ARF.

Summary: The ARF family mapping captures a central antiproliferative role.

Reason: Retain the InterPro-derived assertion in agreement with direct human p14ARF proliferation and arrest experiments. This refers to the Q8N726 ARF product family and does not import the structurally unrelated p16 CDK-inhibitory mechanism.

Summary: ARF can contribute to SUMO modification through ligase regulation.

Reason: The broad Reactome SUMO reaction is consistent with direct ARF binding and modulation of TRIM28 in PMID:26055329. Retain non-core process participation through a cofactor/regulatory contribution; do not use the broad event to infer autonomous SUMO-transferase activity.

Summary: ARF enhances SUMO modification, with independent direct evidence for regulation of the SUMO ligase TRIM28.

Reason: PMID:15355988 reports ARF-dependent WRN SUMO conjugation and UBCH9 synergy but leaves its mechanism unresolved. Independent PMID:26055329 directly demonstrates ARF–TRIM28 binding and enhancement of TRIM28-dependent NPM1 SUMOylation in a recombinant assay; ARF alone does not support the NPM1 reaction under those conditions. GO:0180016 captures the established SUMO ligase regulatory activity. This does not exclude all possible substrate-specific catalytic contributions or require conventional autonomous chemistry of every E3. Preserve the original WRN source; the NPM1/TRIM28 assay is independent mechanistic support.

Proposed replacements: SUMO ligase regulator activity

Summary: The mouse ARF donor supports a bounded lymphoid antiproliferative role.

Reason: Actual Q64364 donor GOA links this assertion to PMID:15964995. Full primary genetic comparisons separate Arf loss from Ink4a loss in Bmi1-deficient mice and measure B-cell counts/composition. Retain the curated ortholog transfer as non-core, without describing it as a direct human B-cell experiment.

Summary: ARF regulates protein stability through its ligase-inhibitory mechanism.

Reason: The mouse donor chain includes PMID:12154087. Independent human p14ARF/MDM2 experiments directly support regulation of p53 stability, coherent with the molecular core. Preserve the ISS source; this is ARF regulating another protein, not an inference from ARF merely being degraded.

Summary: The source reports context-dependent MDM2 destabilization by ARF.

Reason: The original abstract describes ARF binding MDM2, MDM2 degradation and p53 stabilization. Retain this source-specific effect; other human experiments can stabilize MDM2 under different expression conditions, so no universal rule that ARF destabilizes every substrate or always lowers MDM2 is inferred.

Summary: The mouse donor supports a thymocyte antiproliferative context.

Reason: The actual Q64364 source chain points to PMID:15964995. Its full Results distinguish Arf and Ink4a deletions, thymocyte counts and immature CD4/CD8 populations in Bmi1-deficient mice. Retain the curated transfer with that developmental/genetic context, without asserting a newly demonstrated human thymus assay.

Summary: ARF positively regulates EGR1 SUMO modification.

Reason: The source abstract reports ARF-dependent EGR1 SUMOylation and loss-of-function effects, providing a regulatory role rather than mere substrate necessity. Keep this context non-core and distinguish it from intrinsic SUMO-transferase catalysis.

Summary: ARF contributes to apoptosis in the PANO perturbation context.

Reason: In the source abstract, PANO-induced apoptosis in low-serum HeLa cells is inhibited by p14ARF knockdown. Retain the experimental annotation as a context-dependent ARF contribution; PANO stabilizes ARF, and the effect is not assigned to all products or all cell conditions.

Summary: ARF increases p53-dependent transcriptional output.

Reason: Selected original Results and Figures 5–8 test p14ARF controls for p53 targets, including p21/GDF15 output and p53 promoter occupancy. Retain the transcriptional consequence as non-core; ARF regulates the MDM2–p53 axis rather than being inferred to bind promoter DNA directly.

Summary: The source supports increased Pol II target-gene output downstream of ARF.

Reason: The p14ARF comparator experiments in the MEG3 paper increase p53-dependent target expression. This supports the existing regulatory BP, without assigning sequence-specific transcription-factor activity to ARF or claiming a direct polymerase catalytic assay.

Summary: The donor phenotype supports inhibition of stem-cell self-renewing division.

Reason: Current Q64364 donor GOA links this ISS assertion to PMID:15964995. Full Figure 4 and Methods separate Arf-specific knockout from Ink4a/Arf loss and measure serial multipotent neurosphere formation. Arf loss enhances self-renewal and partially rescues Bmi1 deficiency. Refine participation in division to negative regulation of that process, preserving the mouse-to-human transfer and assay context.

Proposed replacements: negative regulation of somatic stem cell division

Summary: Cell-cycle regulation is coherent with the ARF inhibitory core.

Reason: The ARF-specific InterPro mapping is supported by the MDM2–p53 arrest mechanism and human arrest/senescence experiments. Retain the broad process without substituting p16 direct CDK inhibition or extending the full-length mechanism to every alternative product.

Summary: The transferred inhibitory activity can be specified as ubiquitin ligase inhibition.

Reason: The actual mouse Q64364 donor has IDA support from PMID:9878046 for inhibition of Mdm2-mediated p53 ubiquitination. Independent human p14ARF assays in PMID:18560357 establish the same mechanistic class. Refine the broader transferase-inhibitor term to the existing ligase-inhibitor term, preserving the original ISS assertion.

Proposed replacements: ubiquitin ligase inhibitor activity

Summary: EGR1 binding is a supported specific non-core ARF activity.

Reason: The source assigns the actual EGR1 partner, a sequence-specific Pol II transcription factor, and links the interaction to its SUMO regulation. Retain the specific binding term without transferring EGR1 DNA recognition to ARF.

Summary: ARF promotes K63-linked modification of COMMD1.

Reason: Selected original Figure 4 and Results distinguish K63-linked COMMD1 ubiquitination from the usual degradative K48 interpretation. Retain this binding-dependent non-core process contribution, without claiming ARF itself performs the ubiquitin-transfer chemistry.

Summary: Human ARF can implement a stable senescence response.

Reason: The full source tests ectopic p14ARF in hTERT-immortalized human mammary epithelial cells, with arrest, senescent morphology and p53-pathway controls. Retain this central antiproliferative response while preserving the reported resistance in HPV E6-related contexts; ARF is not universally sufficient in every cell.

Summary: Human p14ARF can signal apoptosis in the tested tumor-cell systems.

Reason: The source abstract describes adenoviral p14ARF experiments across defined p53/BAX backgrounds, caspase activation and mitochondrial-potential effects. Retain this context-dependent signaling contribution without making apoptosis the outcome of every ARF interaction.

Summary: ARF-derived peptides bind an initially disordered Hdm2 segment in a coupled folding/assembly assay.

Reason: The original PMID:12630860 abstract reports p14ARF/Hdm2 peptide coassembly. The final journal body was not recovered, but the same-inventor primary disclosure US7704703B2 experimentally describes human ARF residues 1–14 and Hdm2 residues 240–254, their initial disorder and saturable coassembly. This supports the specific disordered-domain interaction at peptide resolution and is consistent with the original Q00987 IPI curation. The disclosure is not relabeled as journal full text or independent replication, and these assays do not establish equivalent activity of every full-length ARF product.

Summary: Short ARF/HDM2 peptides form beta-rich amyloid-like assemblies, but the gene-level biological-process assertion exceeds the demonstrated scope.

Reason: PMID:12630860 positively reports assembly of excised human ARF residues 1–14 with HDM2 residues 240–254. The same-inventor primary disclosure documents insoluble beta-rich protofibrillar material at pH 7/23°C and long bundled fibrils after 100 hours at 70°C/pH 3.0–3.5. FTIR/CD and EM support secondary structure and morphology; negative Congo-red/ThT binding is not used as conclusive proof of nonamyloid identity. The stronger limitation is extrapolation from this peptide system to an unqualified ARF biological process: native cellular ARF fibril formation is not established by the inspected evidence. Retain the observed assembly capacity in the notes, preserve the original IMP/source fields, and mark the process as over-annotated rather than denying the experiment. The final journal body remains unrecovered; the patent is supporting primary disclosure, not a substitute claimed to be that body.

Summary: Human p14ARF inhibits MDM2 ubiquitin ligase activity.

Reason: The recovered full primary Figure 3 and Methods use human p14ARF in U2OS cells and measure MDM2-dependent p53 ubiquitination with ligase controls. This establishes the central inhibitory MF despite the paper title foregrounding ribosomal proteins; ARF is an explicit experimental comparator.

Summary: ARF limits ubiquitin-dependent p53 degradation.

Reason: The same human p14ARF experiments measure reduced MDM2-dependent p53 ubiquitination and increased p53 abundance/response. Retain this mechanistic process in the inhibitory core, rather than infer participation from ARF merely being a substrate.