YFH7 (P43591 / YFR007W / AIM12) — research notes

Species: Saccharomyces cerevisiae (S288c), NCBITaxon:559292
UniProt: P43591 (YFH7_YEAST); ORF YFR007W; SGD S000001903; aliases AIM12.
Length: 353 aa; MW 39,885. Structures: PDB 2GA8 (1.77 Å), 2GAA (1.95 Å).

Deep research provenance

just deep-research-falcon yeast YFH7 --fallback perplexity-lite was attempted twice
(2026-07-05) and timed out both times (Edison/falcon exit 143 after ~5 min; no output
file produced). Per project guidance I did NOT fabricate a -deep-research-*.md file.
These notes are compiled directly from: (1) the cached UniProt record
(YFH7-uniprot.txt), (2) the three cached primary publications
(publications/PMID_18004758.md, PMID_12868057.md, PMID_19300474.md), and (3) the
completed, reproducible bioinformatics analysis in YFH7-bioinformatics/RESULTS.md.

What is KNOWN

Enzyme / fold identity (well established, experimental + structural)

• YFH7 is an ATP-dependent kinase of the P-loop NTPase superfamily. The 2008 crystal
  structure paper solved the fold and did enzymology:
  PMID:18004758
• Fold assignment: weak similarity to the PRK/URK/PANK (phosphoribulokinase / uridine
  kinase / bacterial pantothenate kinase) subfamily of P-loop kinases, with the
  characteristic core, lid, and NMP(bind) subdomains:
  PMID:18004758
• A ~100-residue insertion distinguishes YFH7 from other family members and forms an
  "additional alpha/beta domain of novel topology":
  PMID:18004758
• Measured biophysical parameter: KM = 1178 µM for ATP (UniProt BIOPHYSICOCHEMICAL
  PROPERTIES, ECO:0000269|PubMed:18004758). This anchors the IDA ATP-turnover annotation.

ATP-binding site / motifs (structure + sequence)

• UniProt BINDING 31..39 = ATP (ChEBI:30616), ECO:0000250.
• Bioinformatics (YFH7-bioinformatics/RESULTS.md): canonical Walker A / P-loop
  GSPGSGKS at residues 31–38, matching [AG]-x(4)-G-K-[ST]; and the URK-family
  diagnostic second block VILEG at 263. Both diagnostic blocks are retained →
  catalytic scaffold intact, not an obviously degenerate pseudokinase.
• Domain xrefs: IPR027417 (P-loop NTPase), Gene3D 3.40.50.300, SUPFAM SSF52540 (fold in
  three discontinuous segments 22–54 / 156–204 / 233–349, gap ~55–155 = the insertion),
  PANTHER PTHR10285 "URIDINE KINASE", CDD cd00009 (AAA). UniProt: "Belongs to the
  YFH7 family."

Conservation

• The fold/family is ancient and pan-eukaryotic (eggNOG KOG2702 at Eukaryota;
  OrthoDB group at Eukaryota; PANTHER PTHR10285 URK family reaches into bacteria).
• The specific YFH7 protein is lineage-restricted: bioinformatics UniRef50 cluster
  common taxon = Saccharomycetaceae (budding yeasts), UniRef90 narrows to
  Saccharomyces sensu stricto (YFH7-bioinformatics/RESULTS.md). Consistent with the
  paper's "yeast-specific protein" language.

Cellular / phenotypic associations (indirect, high-throughput)

• Abundance: ~161 molecules/cell in log-phase SD medium (UniProt MISCELLANEOUS,
  ECO:0000269|PubMed:14562106).
• Localization: cytoplasm (GO IBA GO:0005737, GO_Central). No experimental yeast IDA
  localization in the cached record.
• ER biogenesis screen: yfh7Δ was recovered in a parallel deletion-mutant screen
  for genes required for ER (karmellae) biogenesis; UniProt encodes this as the FUNCTION
  hedge "could be involved in endoplasmic reticulum membrane assembly"
  (ECO:0000269|PubMed:12868057). Note: the cached PMID:12868057 abstract is a genome-wide
  screen; it does not itself assign a molecular mechanism to YFH7.
• Mitochondrial-genome maintenance: yfh7Δ increases the frequency of mitochondrial
  genome loss (petite frequency) in a computationally-driven quantitative screen; UniProt
  DISRUPTION PHENOTYPE "Increases frequency of mitochondrial genome loss"
  (ECO:0000269|PubMed:19300474). This is the source of the AIM ("Altered Inheritance of
  Mitochondria") alias. PMID:19300474 (full text cached) is a screen that assigns yeast
  ORFs a mitochondrial-biogenesis phenotype by petite-frequency assay; it does not
  determine YFH7's biochemical substrate.

What is NOT known (the primary deliverable)

• The physiological substrate / phosphoacceptor of YFH7 is unknown. The crystallographers
  could only conclude "new substrate specificity" — i.e. it is a small-molecule kinase
  but they could not identify what it phosphorylates:
  PMID:18004758
  No substrate has been identified since from sequence, structure, or the bioinformatics
  analysis here.
• The pathway / biological process in which YFH7 acts is undefined. The ER-biogenesis
  and mitochondrial-genome-loss associations are pleiotropic deletion-screen phenotypes,
  not a demonstrated molecular pathway; neither is corroborated by a mechanistic study.
• Whether the in-vitro ATPase/ATP-turnover activity reflects productive substrate
  phosphorylation (kinase turnover coupled to a real acceptor) vs. basal P-loop
  ATP hydrolysis is not resolved without the acceptor.

Curation reasoning summary (for the ai-review)

• GO:0016301 kinase activity (ISS, PMID:18004758) — ACCEPT (core). Supported by the intact
  Walker A P-loop + retained URK second block + the paper's ATP-dependent-kinase conclusion.
  Kept general ("kinase activity") deliberately: no substrate is demonstrated, so a specific
  child (e.g. a named small-molecule kinase) would over-specify.
• GO:0016887 ATP hydrolysis activity (IDA, PMID:18004758) — ACCEPT (core MF). Directly
  measured ATP turnover (KM for ATP). This is the experimentally observed activity.
• GO:0016310 phosphorylation (ISS, PMID:18004758) — ACCEPT (BP, non-core-ish but the
  process directly entailed by kinase activity). Kept as-is.
• GO:0005737 cytoplasm (IBA) — ACCEPT/KEEP_AS_NON_CORE. Phylogenetic location call, plausible
  for a soluble P-loop kinase; not core to function.
• GO:0005575 cellular_component (ND) — this is the root "no data" placeholder. KEEP_AS_NON_CORE
  (it is the standard ND stub; nothing to remove, adds no information).
• ATP binding (GO:0005524) was present in the 2026-06 UniProt DR xrefs as an IEA
  UniProtKB-KW annotation but absent from the GOA TSV. UniProt entry version 155 no longer
  exposes this GO xref, so the old proposed NEW ATP-binding row should not be carried
  forward: the Walker A P-loop is mechanistically important evidence for GO:0016301 and
  GO:0016887, but current GOA/UniProt are not asserting a standalone ATP-binding
  molecular function.

2026-09-29 IBA re-review

Re-checked the single YFH7 IBA against projects/IBA_REVIEW.md and the current
interpro/panther/PTHR10285/PTHR10285-paint.tsv cache. The GO:0005737
cytoplasm row traces to PANTHER:PTN000032154, a broad PTHR10285 ancestral node
seeded by plant, bacterial and mammalian cytoplasmic members of the
uridine-kinase/PRK-URK-PANK family. That is the correct propagation_review
source entity for the IBA; the row should not enumerate the extant
WITH/FROM proteins as if this were a pairwise transfer.

The transfer itself is still sound. YFH7 is a solved soluble P-loop kinase with
no transmembrane segment or organelle-targeting sequence, so the generic
cytoplasm term is a defensible inferred site of action while the physiological
small-molecule substrate remains unknown. I therefore added a NO_FAILURE_CORE
propagation review for PANTHER:PTN000032154 and switched the row from
KEEP_AS_NON_CORE to ACCEPT so it lines up with the cytoplasm location already
used in core_functions.

Searches for recent YFH7, YFR007W, AIM12 and P43591 papers did not find
a newer functional publication. The 2008 crystal-structure/enzymology paper
(PMID:18004758) remains the only YFH7-specific mechanistic paper found in
PubMed, and the ER-biogenesis and petite-frequency papers remain screen-level
phenotype context rather than substrate evidence.

2026-10-01 current GOA refresh

• Forced a fresh GOA/UniProt pull for YFH7. QuickGO still returns five live rows: the
  PTHR10285 cytoplasm IBA, one cellular-component ND root, kinase activity and
  phosphorylation ISS rows from the Gueguen-Chaignon structure/enzymology paper, and the
  IDA ATP hydrolysis row from the same paper.
• Copied the current GOA WITH/FROM values into supporting_entities for the IBA and
  two ISS rows. The IBA row still resolves to PANTHER:PTN000032154; the PAINT cache
  still places only a broad cytoplasm assertion there, so the existing NO_FAILURE_CORE
  propagation review remains appropriate.
• Updated the UniProt cache to entry version 155. UniProt no longer lists its previous
  GO:0005737 IBA or GO:0005524 keyword-derived ATP-binding xrefs, matching QuickGO's
  absence of a live ATP-binding row; the prior ATP-binding NEW proposal was therefore
  removed as stale and redundant with kinase/ATP hydrolysis activity.
• Searched again for recent direct YFH7/YFR007W/AIM12/P43591 literature. No newer paper
  assigning the physiological phosphoacceptor or pathway was found.