ERR1 (YOR393W) curation notes

Gene: ERR1 — Enolase-related protein 1
UniProt: P0CX10 (ERR1_YEAST); SGD: S000005920; systematic name YOR393W.
Organism: Saccharomyces cerevisiae S288C (NCBITaxon:559292).
Protein existence level: PE 3 — Inferred from homology (no direct experimental
characterization of the protein itself).

This is an understudied ("dark") gene. The primary curation goal is an honest
knowledge_gaps section plus carefully-reasoned core_functions/description
grounded in domain architecture, orthology, and the (sparse) literature — never
invented function.

1. Paralog landscape — critical for evidence attribution

S. cerevisiae has three "Enolase-Related Region" (ERR) genes plus the two true
glycolytic enolases:

Sequence identities I computed inline (ungapped, all sequences are exactly 437 aa
with no indels — enolase is a highly conserved rigid TIM-barrel fold):

• ERR1 vs ERR2 = 100.0% identical (the two proteins are the same sequence; this
  is why UniProt merged YOR393W and YPL281C into a single accession P0CX10 carrying
  two OrderedLocusNames and two EnsemblFungi/KEGG mRNA cross-references).
• ERR1 vs ERR3 = 99.5% identical (only 2 residue differences over 437).
• ERR1 vs ENO1 = 68.0% identical.
• ERR1 vs ENO2 = 67.3% identical.

Consequence: the three ERR proteins are functionally indistinguishable at the
protein level. No protein-level, sequence-based, or most biochemical assays could
attribute a phenotype specifically to ERR1 rather than ERR2/ERR3. Only
locus-specific genetics (deletion of the specific ORF, allele-specific expression)
could do so, and I have found no such ERR1-specific functional study. Any functional
claim from the literature (or from IBA/ISA propagation) is really a claim about the
ERR family, not ERR1 uniquely.

Foundational reference: Pryde, Huckle & Louis 1995 [PMID:7785338, "The first of
these shows 61%\nand 60% DNA sequence identity to Enolases 1 and 2 respectively. The
Enolase-like\nsequence appears to be species specific, with three copies being found
in all\nstrains of S. cerevisiae studied. The location of the three copies is the same\nfor
all strains."]. Note: this reference is abstract-only in our cache
(full_text_available: false) and reports DNA (not protein) identity and genomic
location — it does NOT assay enzymatic activity.

2. Domain architecture and catalytic-residue analysis (inline, from UniProt)

UniProt P0CX10 annotates a complete two-domain enolase architecture:
- Enolase N-terminal domain (Pfam PF03952 Enolase_N; IPR029017)
- Enolase C-terminal TIM-barrel domain (Pfam PF00113 Enolase_C; IPR020810)
- PANTHER PTHR11902 (ENOLASE), CDD cd03313 enolase, PROSITE PS00164 ENOLASE,
HAMAP MF_00318 Enolase, SFLD enolase.
- KW: Glycolysis, Lyase, Magnesium, Metal-binding.

UniProt catalytic/binding features (all ECO:0000250, i.e. by-similarity, not
experimental):
- ACT_SITE 212 "Proton donor" (Glu)
- ACT_SITE 346 "Proton acceptor" (Lys)
- BINDING 160, 169 substrate
- BINDING 247 Mg(2+)
- BINDING 296 Mg(2+)/substrate
- BINDING 321 Mg(2+)/substrate
- BINDING 373–376 substrate; 397 substrate

Catalytic-residue conservation check I ran inline (ERR1 vs ENO2, position-matched
because both are 437 aa, no indels):

All 8 annotated catalytic / Mg-binding / substrate-binding residues are identical
between ERR1 and ENO2. The canonical enolase catalytic machinery (Glu proton donor,
Lys proton acceptor, the D/E/D Mg2+ triad) is fully intact. ERR1 is therefore NOT a
degenerate pseudoenzyme — it retains every residue required for phosphopyruvate
hydratase (enolase) catalysis. This makes the molecular-function annotations
(phosphopyruvate hydratase activity GO:0004634; magnesium ion binding GO:0000287)
domain-defensible as a prediction.

But conserved catalytic residues establish capacity, not in-vivo activity or
role. The gap is physiological, not structural (see §4).

3. What is KNOWN vs NOT KNOWN

KNOWN (well supported)

• ERR1 is a member of the enolase family with a complete, catalytically-intact enolase
  fold (domain analysis + residue conservation above).
• There are three near-identical ERR paralogs at conserved sub-telomeric loci, present
  in all S. cerevisiae strains examined; the family is S. cerevisiae-specific
  (Pryde et al. 1995, PMID:7785338).
• By homology to ENO1/ENO2, the predicted activity is 2-phospho-D-glycerate
  hydro-lyase (EC 4.2.1.11): (2R)-2-phosphoglycerate = phosphoenolpyruvate + H2O,
  Mg2+-dependent.

NOT KNOWN (genuine gaps)

• Whether ERR1 is catalytically active as an enolase in vivo. No purified-protein
  enzyme assay attributable specifically to ERR1 exists. PE level is 3 (homology).
• The in-vivo biological role. Deleting ENO1+ENO2 is what impairs glycolysis;
  the ERR genes are not part of the characterized core glycolytic machinery. Whether
  ERR1 contributes measurably to glycolytic flux under any standard condition is
  undetermined.
• The expression condition. If ERR1 is expressed at all, under what
  condition/stage? (The ERR family has been discussed in the context of sporulation
  in the broader literature, but I have not verified an ERR1-specific expression
  dataset in our cache; do not assert this without provenance.)
• Redundancy / non-redundancy vs ERR2, ERR3, and vs ENO1/ENO2. Because ERR1 =
  ERR2 (100%) and ~ERR3 (99.5%), any function is almost certainly shared across the
  ERR trio; whether the trio does anything the true enolases do not is unknown.
• Subcellular localization is undetermined (GOA carries an ND cellular_component
  annotation). Cytosol is the default expectation for a glycolytic-type enzyme (no
  signal peptide/TM in the sequence), but this is not experimentally established for
  ERR1.

4. Annotation-by-annotation reasoning (GOA has 10 lines)

All ten GOA lines derive from homology/electronic/phylogenetic inference or are
ND placeholders — there is no experimental (IDA/IMP/IPI/IGI) annotation for ERR1.

1. GO:0004634 phosphopyruvate hydratase activity / IBA (GO_REF:0000033) — enables.
   Phylogenetic propagation from the enolase PANTHER family (PTN000224401), with
   ENO1/ENO2 among the source genes. Domain-defensible (residues intact). ACCEPT
   as the core predicted MF; but note it is inference, not measured for ERR1.
2. GO:0006096 glycolytic process / IBA (GO_REF:0000033) — involved_in. Propagated
   BP. The MF is defensible; the process (that ERR1 actually participates in
   glycolysis in vivo) is exactly the unproven physiological claim. KEEP_AS_NON_CORE
   — retain the family-level expectation but flag it is not established that ERR1
   contributes to glycolytic flux (it is not one of the two characterized glycolytic
   enolases). This is the honest treatment for a dark paralog.
3. GO:0000015 phosphopyruvate hydratase complex / IBA (GO_REF:0000033) — part_of.
   Enolase functions as a homodimer/oligomer = the phosphopyruvate hydratase complex.
   For a catalytically-competent enolase this CC is reasonable but whether ERR1
   forms/joins such a complex in vivo is unverified. KEEP_AS_NON_CORE.
4. GO:0000015 phosphopyruvate hydratase complex / IEA (GO_REF:0000002, InterPro) —
   part_of. Same term via InterPro IPR000941. Redundant electronic support for #3.
   KEEP_AS_NON_CORE.
5. GO:0000287 magnesium ion binding / IEA (GO_REF:0000002, InterPro) — enables.
   Mg2+ is the enolase cofactor; the three Mg-binding residues are conserved in ERR1
   (D247/E296/D321). Domain-defensible. ACCEPT.
6. GO:0004634 phosphopyruvate hydratase activity / IEA (GO_REF:0000120) — enables.
   Combined automated (ARBA/EC/RHEA/InterPro). Redundant with #1/#8. ACCEPT (same
   MF, electronic).
7. GO:0006096 glycolytic process / IEA (GO_REF:0000120) — involved_in. Redundant
   with #2, electronic. KEEP_AS_NON_CORE (same reasoning as #2).
8. GO:0004634 phosphopyruvate hydratase activity / ISA (PMID:7785338;
   with SGD:S000001217 ENO1, SGD:S000003486 ENO2) — enables. Sequence-similarity
   annotation from SGD based on the Pryde 1995 identity to ENO1/ENO2. The reference
   supports sequence similarity, and the MF is domain-defensible. ACCEPT, but
   note the supporting paper reports DNA identity + genomic location only (abstract
   states "61% and 60% DNA sequence identity to Enolases 1 and 2"), not enzyme assay.
   Per curation rules: do NOT REMOVE an ISA annotation on paralog grounds; the MF
   inference is sound.
9. GO:0005575 cellular_component / ND (GO_REF:0000015) — is_active_in. Root-term ND
   placeholder = "no data". Standard practice is to keep as-is. ACCEPT (it
   correctly records the absence of CC data).
10. GO:0008150 biological_process / ND (GO_REF:0000015) — involved_in. Root-term ND
    placeholder. ACCEPT (records absence of BP data).

No protein binding terms present (good). No experimental annotations to protect from
over-ruling. No negated/isoform annotations.

5. Provenance discipline

• Sequence identities and catalytic-residue table in §1–2 are computed inline in this
  session from UniProt sequences (P0CX10 ERR1, P00925 ENO2, P00924 ENO1, P0CX11 ERR2,
  P42222 ERR3). These are recorded as a file: bioinformatics-style provenance
  anchored to the UniProt record; the core numeric claims (100% ERR1=ERR2, all
  catalytic residues conserved) are reproducible from the sequences in the UniProt
  files.
• The only literature reference in GOA is PMID:7785338 (abstract-only). Quotes used in
  the review are verbatim substrings of the cached abstract.
• Falcon deep research was requested; if/when it returns it is kept as
  ERR1-deep-research-falcon.md. It must NOT be used to fabricate ERR1-specific
  experimental claims — the paralog-indistinguishability point above stands regardless.

6. Deep research status (falcon unavailable)

Falcon deep research (just deep-research-falcon yeast ERR1 --fallback perplexity-lite)
was launched at the start of the session and ran for ~24 minutes with zero output
before the wrapper gave up; a single bounded retry (8-min hard timeout) was then
attempted per protocol and also returned nothing (SIGTERM/exit 143). No
ERR1-deep-research-*.md file was produced. The Edison/falcon endpoint was effectively
hanging today. Per project rules I did NOT fabricate a -deep-research-{provider}.md
file. This review is therefore grounded entirely in:
- the UniProt record P0CX10 (domain architecture, catalytic residues, PE level),
- the QuickGO GOA export (the 10 existing annotations),
- the one cached primary reference PMID:7785338 (abstract-only), and
- inline sequence/catalytic-residue analysis run in this session (ERR1 vs
ENO1/ENO2/ERR2/ERR3).
The paralog-indistinguishability and dark-gene conclusions do not depend on deep
research; the honest knowledge_gaps section captures exactly what remains unknown.