Journal of research for the AI GO-annotation review of Saccharomyces cerevisiae ESL1.
All assertions carry inline provenance. Where a claim is about the paralog ESL2 or the
ESL1/ESL2 pair jointly, that is stated explicitly.
The UniProt record and the primary literature both establish ESL1 as a PIN-domain +
14-3-3-like (EOH) protein of the Est1/SMG family — there is no GRAM domain, no VASt
domain, and no evidence for lipid binding or lipid transfer. I proceeded on the basis of
the actual UniProt/primary-literature evidence, not the brief's mischaracterization.
All functional data come from a single paper studying esl1Δ, esl2Δ single mutants and
esl1Δ esl2Δ double mutants (and nuclease-dead alleles). Most phenotypes are reported for
the pair and are stronger in the double mutant, indicating functional redundancy
between ESL1 and ESL2. ESL1-specific single-mutant results are called out below.
NOT NMD; NOT telomere maintenance. Despite structural orthology to metazoan
hEST1A/B (SMG5/6), esl1Δ, esl2Δ and esl1Δ esl2Δ mutants have wild-type telomere
length, normal senescence/ALT kinetics (with est2Δ), normal subtelomeric silencing
(5-FOA), no TERRA accumulation, and no accumulation of NMD substrates (ade2-1;
unspliced pre-CYH2). [PMID:23893744 "Esl1 and Esl2 were not involved in
nonsense-mediated mRNA decay or telomere maintenance pathways"; "ESL1 and ESL2 do not
seem to have NMD-related functions"; "Esl1 and Esl2 are not required for
telomerase-dependent or alternative telomere maintenance mechanisms"]
→ This directly REFUTES the IBA-propagated NMD (GO:0000184), telomerase holoenzyme
complex (GO:0005697), telomeric DNA binding (GO:0042162), and telomerase RNA binding
(GO:0070034) annotations for this yeast gene.
Environment-sensing adaptive gene expression. esl1Δ esl2Δ deregulate ~50–53
transcripts (≥2-fold) — hexose transporters (HXT3/6/7), hexokinase HXK1, MAL genes,
JEN1, glycogen genes (GPH1, PGM2), and one-carbon/glycine-regulon genes (GCV1/2/3,
SHM2, ADE17) — in the direction opposite to the environmentally appropriate
response. [PMID:23893744 "absence of Esl1 and Esl2 led to more than two-fold
deregulation of ∼50 transcripts, most of which were expressed inversely to the
appropriate metabolic response to environmental nutrient supply"; "esl1Δ esl2Δ double
mutants may have a defect in adapting the expression of hexose and one-carbon
metabolism genes to environmentally appropriate requirements"]
Genetic interaction with the Rim101 pH-sensing pathway. esl1Δ esl2Δ is synthetic
sick with rim8Δ and dfg16Δ (the arrestin-like adaptor and GPCR of the Rim101 pathway),
and ESL2 interacts with RIM9/RIM13/RIM20 in an independent screen — supporting a role
parallel to Rim101 in environmental adaptation. PMID:23893744
Genome-stability / stress phenotypes (pair; some nuclease-dependent). esl1Δ esl2Δ
show elevated mitochondrial-genome instability (petite formation) and altered
sensitivity to genotoxins (better on bleomycin/HU, worse on adriamycin). Bleomycin
resistance is phenocopied by nuclease-dead alleles, implicating the PIN catalytic
residues in this phenotype. [PMID:23893744 "ESL1 and ESL2 contribute to maintenance
of mitochondrial genome stability"; "the nuclease-dead mutants phenocopied the
bleomycin sensitivity of the esl1Δ or esl2Δ mutants"]
Synthetic interaction with trf4Δ (TRAMP-complex poly(A) polymerase); esl1Δ esl2Δ
deregulate neighboring CUTs/SUTs — but this CUT/SUT change is nuclease-independent
and not directly coupled to the coding-gene changes. [PMID:23893744 "loss of Esl1 and
of Esl2 lead to impaired genetic fitness with trf4Δ, rim8Δ, and dfg16Δ"; "these
changes in transcript levels were again independent of Esl1 and Esl2 nuclease domain
integrity"]
| GO term | aspect | evidence | verdict |
|---|---|---|---|
| GO:0000184 NMD | BP | IBA | REMOVE — directly refuted for this yeast gene (PMID:23893744) |
| GO:0005697 telomerase holoenzyme complex | CC | IBA | REMOVE — refuted (no telomere role) |
| GO:0042162 telomeric repeat DNA binding | MF | IBA | REMOVE — refuted (no telomere role) |
| GO:0070034 telomerase RNA binding | MF | IBA | REMOVE — refuted (no telomere role) |
| GO:0003674 molecular_function (root) | MF | ND | KEEP_AS_NON_CORE (ND placeholder) |
| GO:0005575 cellular_component (root) | CC | ND | KEEP_AS_NON_CORE (ND placeholder) |
| GO:0008150 biological_process (root) | BP | ND | KEEP_AS_NON_CORE (ND placeholder) |
The four IBA annotations are phylogenetic propagations from the metazoan SMG5/6 / EST1A
clade. Lai et al. 2013 experimentally tested exactly these two functions (NMD and
telomere maintenance) in the yeast proteins and found them absent. This is a textbook
case where experimental data in the target species overrides an IBA propagation. Note the
NMD IBA even lists SGD:S000002614 (EBS1) among the with/from set — EBS1, not ESL1, is the
yeast SMG7 ortholog with NMD links; ESL1's inclusion in the tree does not reflect yeast
NMD function.
ESL1-deep-research-falcon.md was generated (Edison, 21 citations, ~29 min). It is
literature-limited: falcon could NOT access the key functional paper (Lai et al. 2013)
and therefore relies on domain/family inference (PIN + TPR-like/EOH, Est1/Ebs1-like
family) and on the paralogs Est1/Ebs1. It correctly flags ESL1 as poorly characterized and
notes the mammalian ESL1 = E-selectin ligand 1 (GLG1) name clash (irrelevant to yeast
YIL151C). It does not contradict this review; because it lacked the primary paper, its
inferential telomere/NMD framing is superseded by the direct experimental exclusion of
those functions in Lai et al. 2013, which I read in full. No falcon-only claims were
imported into the review.
All nine input rows were assessed; seven GOA rows and all source fields are preserved. The two authored NEW process assertions (regulation of gene expression; cellular response to nutrient levels) were withdrawn. Their former rationale was alignment with core terms. The primary double-mutant transcriptome and genetic data establish physiological effects, but do not identify the step performed by Esl1, and the user requires participation evidence for NEW. These findings remain in the biological description, references and suggested questions. The speculative custom PIN-domain-dependent regulatory MF proposal was also withdrawn; most tested transcript effects were nuclease-independent. No gene-process assertion was invented to eliminate an empty MF slot or a validation warning. No matching GO-CAM target entry was found.
The full text of Lai et al. PMID:23893744(https://doi.org/10.1534/g3.113.006924) was read, including methods and Figures 2–6. Figure 2 measures ade2-1 and pre-CYH2 RNA abundance against upf1-null positive controls in single and double esl deletions. These direct NMD-substrate negatives support the retained REMOVE for canonical NMD under the tested conditions; they do not prove every substrate or condition negative. Telomere assays instead measure length, senescence/alternative lengthening, subtelomeric silencing and transcript abundance. They do not measure Esl1–DNA or Esl1–TLC1 binding or physical holoenzyme membership. The three corresponding REMOVE calls become UNDECIDED: relevant physiological negatives create an indirect conflict with the inherited function, but do not establish molecular or complex loss. Root has requested a neutral focused adjudication; no duplicate was launched.
The actual PANTHER tree was retrieved by POST for PTHR15696. Accession P40456 is explicitly leaf PTN007651903 below ancestral PTN000403280. Related IRD node PTN008560932 is outside this lineage. Its NOT rows mention the same root as their evidence and must not be mistaken for target-lineage losses. Exact path and target/ancestral rows are retained in ESL1-paint-lineage.json and interpreted in the lineage note.
The complete Falcon report was read and added to references with its substantive limitations: it explicitly lacked the dedicated Lai paper, incorrectly treated Esl1 as the only yeast PIN-bearing EST/SMG family member, and framed NMD/telomere roles largely by analogy. Primary target data supersede that omission; the report remains useful for comparative domains and for distinguishing conserved architecture from assayed nuclease activity. Exact cache searches found no OpenScientist report for ESL1/YIL151C/P40456. Gene-specific abundance and phosphorylation findings formerly quoted as primary-paper text actually originate in UniProt; their attribution was corrected, and the two primary caches are explicitly marked abstract-only.
Description, core narrative and knowledge gaps now avoid categorical loss of all NMD/telomere-related molecular functions. All source fields and quoted supporting text were checked.
Read the full telomerase-association report, HTML rendition and all three CSV artifacts. Its recommendations exceed its evidence: IBA-only is not an annotation failure, zero STRING experimental score is not a negative assay, and identical database scores do not trace the database source. The sequence matrix lacks executable code, sequence versions and aligned sequences; neither its reported global identity nor an uninspected subfamily name establishes interface loss. The positive IBD PTN000403280 is an actual ancestor of P40456/PTN007651903, and the related loss node is outside that lineage. NMD-substrate negatives cannot refute a separate binding/complex function by analogy.
The report explicitly did not read primary full texts. Newly fetched PMID:22544908 Fig.4 and assay Results establish purified KlEst1–Ter1 UV crosslinking despite failed EMSA, not an Esl1 negative. The previously reviewed full target paper PMID:23893744 tests telomere outcomes, not physical Esl1 binding or complex incorporation. Retain UNDECIDED for all three telomerase assertions and record an expert/interface or direct-assay follow-up. No repeat OpenScientist call, no new process/nuclease term, and no change to the scoped canonical-NMD judgment.