YFH7 (P43591 / YFR007W) — bioinformatics analysis

Target: Saccharomyces cerevisiae YFH7 (UniProt P43591; ORF YFR007W; alias AIM12),
a 353-aa protein annotated as an "ATP-dependent kinase" (EC 2.7.1.-), with a solved
crystal structure (PDB 2GA8 at 1.77 Å, 2GAA at 1.95 Å; Gueguen-Chaignon et al.
2008, PMID:18004758).

Question addressed: Is YFH7 a structurally intact, plausibly catalytically-competent
small-molecule (P-loop) kinase, what is its domain architecture, and how broadly is it
conserved?

Methods (reproducible)

All steps are driven by the justfile; scripts read sequence/accession from
arguments (nothing gene-specific is hardcoded). Dependencies are managed with uv
(pyproject.toml).

Run with just all (motifs offline; domains/conservation need network access to the
public EBI/UniProt REST APIs). Control tests: just test-controls.

Findings

1. Walker A / P-loop motif — present and canonical

The classic P-loop (Walker A) glycine-rich loop is detected at residues 31–38,
GSPGSGKS(T), matching the [AG]-x(4)-G-K-[ST] consensus and exactly the position
reported for the ATP-binding site in UniProt (BINDING 31..39) and in the crystal
structure. This is the ATP β/γ-phosphate–binding loop of P-loop NTPases/kinases.

2. Walker B / catalytic acidic residue — candidate present

The Walker-B heuristic (hydrophobic strand ending in D/E) yields three candidates; the
one consistent with a P-loop-kinase catalytic Mg²⁺-coordinating aspartate lies within
the downstream VILEG block (see point 4). The heuristic is weak on its own, but its
best candidate co-locates with the family's diagnostic second block, consistent with an
intact catalytic scaffold. (Walker B position is not pinned to a single residue by
sequence alone — noted as a limitation.)

3. Domain architecture — P-loop NTPase fold with a ~100-residue insertion

From UniProt + InterPro cross-references (results_domains.json):

• IPR027417 P-loop containing nucleoside triphosphate hydrolase (homologous
  superfamily), spanning 1–353.
• Gene3D 3.40.50.300 / SUPFAM SSF52540 — P-loop NTPase fold. The SUPFAM fold is
  reported in three discontinuous segments (22–54, 156–204, 233–349), leaving a gap
  at ~55–155. This ~101-residue stretch is the "additional α/β domain of novel
  topology" / ~100-residue insertion described in the crystal paper — the feature that
  distinguishes YFH7 from canonical family members.
• PANTHER PTHR10285 "URIDINE KINASE" / nucleoside-kinase domain, spanning 12–322.
• CDD cd00009 (AAA) mapped to 9–105 (the N-terminal core lobe).
• UniProt similarity comment: "Belongs to the YFH7 family."

This is fully consistent with the crystal-structure conclusion that YFH7 adopts the
PRK/URK/PANK (phosphoribulokinase / uridine kinase / pantothenate kinase) fold with
core, lid, and NMPbind subdomains plus a family-atypical insertion.

4. PRK/URK/PANK family signature — two diagnostic blocks retained

Comparing YFH7 against a bona fide family member, E. coli uridine kinase (Udk,
P0A8F4), both proteins carry the two diagnostic sequence blocks
(results_urk_compare.json):

Retention of both blocks indicates the small-molecule-kinase catalytic scaffold is
intact — i.e. YFH7 is not an obviously degenerate pseudokinase at the level of these
family fingerprints. (This is evidence for structural competence, not proof of in-vivo
catalysis or of any particular substrate.)

5. Orthology / conservation — ancient fold, budding-yeast-restricted protein

Two layers must be distinguished:

• The fold/domain is ancient and pan-eukaryotic. YFH7 belongs to eggNOG
  KOG2702 (a eukaryotic orthologous group), OrthoDB group at the Eukaryota
  level (...at2759), has an InParanoid record, and maps to PANTHER PTHR10285
  (uridine-kinase family) — all indicating the P-loop kinase family it sits in is
  deeply conserved across eukaryotes (and, via the URK family, into bacteria).
• The specific YFH7 protein is lineage-restricted. Its UniRef50 cluster (≥50 %
  identity) has a common taxon of Saccharomycetaceae (23 members: Saccharomyces,
  Vanderwaltozyma polyspora, Saccharomyces eubayanus/pastorianus, etc.), and the
  UniRef90 cluster narrows to Saccharomyces sensu stricto (results_conservation.json).
  For comparison, the control protein actin (ACT1, P60010) has a UniRef50 common taxon
  of Eukaryota (434 members) — the expected signature of a broadly conserved protein.

Interpretation: YFH7 is a budding-yeast-specific member of an ancient P-loop
small-molecule-kinase family. This matches the crystal paper's description of a
"yeast-specific protein showing weak similarity" to the PRK/URK/PANK subfamily, with a
lineage-specific insertion. The specific in-vivo substrate is not determinable from
sequence/structure alone.

Overall assessment

YFH7 is a structurally intact P-loop (PRK/URK/PANK-fold) ATP-dependent small-molecule
kinase with a canonical Walker A P-loop, retained family diagnostic blocks, and a
lineage-specific ~100-residue insertion. The evidence supports classifying it as a
catalytically-plausible small-molecule kinase (kinase activity; more specifically
consistent with a phosphotransferase acting on a small-molecule hydroxyl acceptor), but
the physiological substrate is unknown — the crystallographers could only conclude
"new substrate specificity," and no substrate has since been identified from sequence,
structure, or the analyses here. This is an MF/BP substrate gap, not a fold or
family-membership uncertainty.

Checklist

• [x] Scripts do not use hardcoded inputs or outputs (sequence/accession are CLI args;
  verified by running on control sequences).
• [x] Scripts tested on at least one other input:
  • analyze_motifs.py / compare_urk_motifs.py: positive control E. coli Udk
    (P-loop + URK block detected) and negative control yeast histone H4 (neither
    detected).
  • analyze_conservation.py: control actin ACT1 (P60010) → Eukaryota-wide cluster,
    correctly contrasting with YFH7's yeast-restricted cluster.
• [x] Analyses completed as expected (all just steps run cleanly).
• [x] Direct results of the scripts are in the folder (results_*.json).
• [x] Summary includes provenance and justification; uncertainty (substrate identity,
  exact Walker B residue) is stated explicitly.
• [?] Walker B catalytic residue is a heuristic candidate only, not pinned to a single
  residue by sequence alone (would need structure-based residue mapping from PDB 2GA8).

Data provenance / tools

• UniProt REST (rest.uniprot.org) — entry P43591, UniRef50/90 clusters.
• InterPro REST (www.ebi.ac.uk/interpro) — domain-to-protein mapping.
• Reference family member: E. coli uridine kinase Udk (P0A8F4).
• Control proteins: yeast actin ACT1 (P60010), yeast histone H4 (P02309).
• Structural/functional context: Gueguen-Chaignon et al. 2008, Proteins 71:804–812
  (PMID:18004758; PDB 2GA8, 2GAA).
• biopython 1.87, requests 2.34 (see pyproject.toml / uv.lock).