# ANKZF1 PN review notes

Deep research provider status: the Falcon deep-research job was attempted first and timed out after 600 seconds. The configured perplexity-lite fallback was attempted next and failed with a Perplexity API 401 quota error. No provider-authored deep-research file was produced, so this review uses the cached UniProt, GOA, publication, Reactome, PANTHER, and PN projection sources directly.

ANKZF1 is best curated as a cytosolic ribosome-associated quality-control factor. UniProt summarizes the core activity as an endonuclease that cleaves polypeptidyl-tRNAs downstream of RQC to release incompletely synthesized polypeptides for degradation and states that ANKZF1-cleaved tRNAs are repaired and recycled by ELAC1 and TRNT1 [file:human/ANKZF1/ANKZF1-uniprot.txt, "Endonuclease that cleaves polypeptidyl-tRNAs downstream of"].

The direct literature support is strong for RQC and tRNA cleavage. Verma et al. showed human ANKZF1 activity in a conserved stalled-ribosome release assay: "Purified Ankzf1 ... catalyzed deacylation" and framed the pathway as preventing aggregation/proteotoxic stress when nascent chains are not degraded [PMID:29632312, "Failure to degrade the NCs leads to protein aggregation and proteotoxic stress"; PMID:29632312, "Purified Ankzf1"]. Kuroha et al. refined the mechanism in a mammalian RQC system: ANKZF1 does not act as a peptidyl-tRNA hydrolase, but "induces specific cleavage in the acceptor arm of 60S-bound P site tRNA" and releases proteasome-degradable nascent chains linked to a short tRNA fragment [PMID:30244831, "specific cleavage in the acceptor arm of 60S-bound P site tRNA"; PMID:30244831, "In conclusion, ANKZF1 induces specific tRNA cleavage in ubiquitinated 60S RNCs"]. Su et al. show ANKZF1/Vms1p precisely cleaves the terminal 3'CCA nucleotides of polypeptidyl-tRNAs on RQC complexes [PMID:31011209, "polypeptidyl-tRNAs on RQC complexes by precisely cleaving off the terminal 3'CCA"]. Yip et al. support this in cells by showing ELAC1 repair of ANKZF1-cleaved tRNAs during ribosome stalling [PMID:32075755, "Deleting ELAC1 leads to the ANKZF1-dependent accumulation of unrepaired tRNA intermediates"].

Existing GOA protein-quality-control and stalled-cytosolic-ribosome annotations are therefore core and should be accepted. The existing MF annotations to RNA endonuclease activity and catalytic activity acting on a tRNA are biologically correct but less specific than GO:0004549 tRNA-specific ribonuclease activity, so they were marked MODIFY to that term.

The IBA ERAD pathway annotation appears to be an overprojection from a broad Vms1/Cdc48 quality-control relationship. The human evidence supports cytosolic RQC and stress-dependent mitochondrial context, not a direct ERAD pathway role.

PN projection check: the Proteostasis projection already has ANKZF1 exactly covered for GO:0006515 under Translation|Cytosolic translation|Ribosome-associated QC [file:projects/PROTEOSTASIS/reports/pn_projection/pn_projected_gene_go_summary.tsv, "ANKZF1		GO:0006515	protein quality control for misfolded or incompletely synthesized proteins"]. The new PN candidate is GO:0035694 mitochondrial protein catabolic process from the mitochondrial proteostasis Vms pathway row [file:projects/PROTEOSTASIS/reports/pn_projection/pn_projected_candidate_additions.tsv, "ANKZF1		GO:0035694	mitochondrial protein catabolic process"]. The mapping source is a broad parent class for organelle-specific protein degradation, while the Vms pathway child itself is no-mapping, and the mapping audit requires manual gene-level review before changes [file:projects/PROTEOSTASIS/mappings/mitochondrial_proteostasis.yaml, "No additional direct GO mapping is appropriate from this node"; file:projects/PROTEOSTASIS/reports/pn_mapping_audit/current_mapping_scrutiny.tsv, "manual_gene_level_review_required_before_gene_review_change"]. I do not see direct human evidence that ANKZF1 catalyzes mitochondrial protein catabolism. The mitochondrial evidence is stress localization/mitochondrial integrity: ANKZF1 "translocates to the mitochondria upon cellular stress" and ANKZF1 depletion reduces mitochondrial integrity and respiration under stress [PMID:28302725, "translocates to the mitochondria upon cellular stress"; PMID:28302725, "integrity and mitochondrial respiration under conditions of cellular stress"]. Therefore GO:0035694 should not be added for ANKZF1 in this pass.

The HIV AP-MS protein-binding row is a generic interaction annotation from a large-scale host-virus screen [PMID:22190034, "physical interactions of all 18 HIV-1 proteins and polyproteins with host"]. The NK-cell membrane annotation is from a high-throughput membrane proteome that explicitly includes many proteins transiently associated with membranes [PMID:19946888, "transiently associated with membranes"]. Both are non-core/over-annotated relative to direct ANKZF1 biology.

## Falcon deep research findings (2026-06-07)

A Falcon (Edison Scientific) deep-research report was generated and reviewed. It confirms the existing RQC-centric curation and adds several literature contexts not previously captured. PMIDs below resolved via DOI/PubMed lookup.

- CONFIRMS (molecular function / process / localization): The report reiterates the existing core picture - ANKZF1 cleaves the tRNA acceptor arm/3'-CCA of P-site peptidyl-tRNA on 60S RQC complexes (not a canonical peptidyl-tRNA hydrolase), releasing nascent chains for p97/VCP extraction and proteasomal degradation, with ELAC1/TRNT1 recycling cleaved tRNAs; basal cytosolic localization with stress-induced mitochondrial translocation. No change warranted to existing annotations. The recurring "peptidyl-tRNA hydrolase" vs "tRNA endonuclease / acceptor-arm cleavage" terminology split is noted but both converge on nascent-chain release [PMID:30244831, "specific cleavage in the acceptor arm of 60S-bound P site tRNA"].

- NEW (interaction / mechanism, well-supported): The VCP/p97 interaction is mediated by a defined VCP-interacting motif (VIM, consensus RX5AAX2R) that binds the p97 N-domain; ANKZF1/ZNF744 was one of the founding members used to define this VIM cofactor family [PMID:21896481 Stapf 2011 "The general definition of the p97/valosin-containing protein (VCP)-interacting motif (VIM)..."]. This gives a concrete, informative replacement context for the generic "protein binding" annotation (VCP binding via VIM), though I am not adding a VCP-binding GO annotation in this conservative pass since there is no existing GOA row to modify.

- NEW (biological process, disease-adjacent, well-supported): ANKZF1 (with NEMF and LTN1) limits repeat-associated non-AUG (RAN) translation of toxic dipeptide-repeat products from G4C2 (C9ORF72) and CGG (FMR1) repeats; ANKZF1 depletion boosts RAN products and overexpression reduces them in reporters and C9ALS/FTD patient iPSC-derived neurons [PMID:38412259 Tseng 2024]. This is a disease-relevant consequence of the same core RQC activity, not a new molecular function.

- NEW (process context, well-supported): Mitochondrial stress upregulates a CNOT4-ZNF598 RQC pathway in which downstream steps include ANKZF1-mediated release of stalled nascent chains; ZNF598 overexpression clears faulty translation products of mitochondrial outer-membrane mRNAs and improves mitochondrial/tissue health in Drosophila disease models [PMID:38388640 Geng 2024]. Connects RQC to mitochondrial proteostasis but does not, by itself, establish ANKZF1 catalyzing mitochondrial protein catabolism.

- NEW (regulation / physiology, moderate confidence): ANKZF1 is reported as a HIF-1alpha effector/target gene; CRISPR-edited ANKZF1-inactivated human MSCs show impaired pro-angiogenic processes under hypoxia [PMID:37158785 Lei 2024]. Plausible but tangential to the core RQC function; treated as non-core context only.

- PROVISIONAL / low weight (oncology biomarker, correlative): High ANKZF1 expression associates with poorer overall and recurrence-free survival in colon cancer in a TCGA bioinformatic analysis (OS HR 2.094, 95% CI 1.188-3.689, p=0.011) [PMID:31257922 Zhou 2019]. Correlative expression-survival association; does not inform GO annotation.

- Reviews / supporting framework (not used to change annotations): Inada 2020 NAR RQC review (DOI:10.1093/nar/gkz1201) and McGirr 2025 FEBS J RQC-disease review (DOI:10.1111/febs.17217) place ANKZF1 within the canonical mammalian RQC step order; Miscicka 2024 NAR (DOI:10.1093/nar/gkae087) covers upstream ZNF598/ASCC ribosome splitting. Open Targets disease-association scores cited in the report are platform/screen-derived and not used here.

Net effect on the review: the existing annotation actions remain correct and unchanged. I added the genuinely-new primary references (Stapf 2011 VIM; Tseng 2024 RAN; Geng 2024 mito-stress RQC; Lei 2024 HIF-1alpha; Zhou 2019 colon cancer) to the references list as statement-only entries, and added a couple of suggested questions reflecting the VIM/p97 mechanism and the RAN-translation link.