Focus: function_assignment · slug histone-site-specificity-telomeres-and-dna-maintenance
Source: genes/yeast/HST1/HST1-ai-review.yaml (free-text hypothesis)
Iteration: 1 of 3
Verdict: PARTIALLY SUPPORTED — the compound seed hypothesis is a mix of one well-supported claim and several over-annotated / unresolved claims.
The seed bundles five independent claims. Adjudicated separately:
| Sub-claim | Verdict |
|---|---|
| (1) NAD-dependent deacetylase activity | Supported (direct) |
| (1b) Specific H3K14 or H3K9 site as the physiological substrate | Weakly supported / unresolved — direct evidence is for general NAD-dependent HDAC activity and for H4K5ac as the in vivo mark; H3K9/H3K14 terms are phylogenetic inference (IBA) only |
| (2) Participates in native subtelomeric heterochromatin | Refuted for native Hst1 — Hst1 is a gene-specific Sum1–Rfm1 corepressor; regional/telomeric silencing is the Sir2 function |
| (3) Localizes to telomeric regions | Refuted/unsupported for native Hst1 (annotation is IEA only); telomere/HM localization occurs only in the engineered SUM1‑1 gain-of-function context (at HM loci, not telomeres) |
| (4) Responds to DNA damage | Weakly supported / mostly attributable to other sirtuins — the DNA-damage-response term is IBA only; the genotoxin-sensitivity phenotype is dominated by Hst3/Hst4 (H3K56ac) |
| (5) Negatively regulates recombination | Weakly supported / likely refuted for native rDNA recombination — the founding hst1Δ study (PMID:8810037) reports no rDNA-recombination phenotype; an IMP annotation (GO:0045950) exists for a specific context, but direct control vs. secondary transcriptional effect is unresolved |
Core reasoning. Native Hst1's primary, directly-demonstrated function is NAD⁺-dependent, Sum1–Rfm1-tethered, promoter-local histone deacetylation that gene-specifically represses middle-sporulation genes, de novo NAD⁺-biosynthesis (BNA) genes, and stress genes. The seed's histone-site, telomere, subtelomeric-heterochromatin, and DNA-damage claims map almost one-to-one onto IBA/IEA (inferred, non-experimental) GO annotations that read as paralog carry-over from Sir2. This is the central curation signal.
Most important caveats. (i) "H4K5 preference is not exclusion of other sites" is fair — but the burden of proof for target-specific H3K9/H3K14 native activity is unmet by primary data. (ii) PMID:18990212 (H4K16 context effects) and PMID:27185881 (conditional HMR effects) were pre-excluded as already-settled and were not re-litigated here.
| Citation (PMID) | Evidence type | Supports/Refutes/Qualifies | Claim tested | Key finding | Context | Confidence / limitations |
|---|---|---|---|---|---|---|
| 12972620 (Bedalov 2003) | Direct assay + expression | Supports (1); qualifies (1b) | NAD-dependent activity; physiological target | Hst1 is an NAD⁺ sensor; its NAD-dependent deacetylase activity directly represses de novo NAD⁺ biosynthesis genes | S. cerevisiae, in vivo + transcriptomics | High for NAD-dependence & gene-specific repression; does not localize activity to H3K9/H3K14 |
| 17242192 (Mead 2007) | Domain-swap chimeras / mutant | Refutes (2),(3) for native Hst1 | Gene-specific vs regional silencing | Hst1→Rfm1/Sum1 = gene-specific promoter repression; Sir2→Sir3/Sir4 = telomere/HM regional silencing; specificity set by cofactors, not enzymatic difference | S. cerevisiae | High; the cleanest native Hst1-vs-Sir2 separation |
| 12612074 (McCord 2003) | Interaction + mutant | Qualifies mechanism | Targeting architecture | Rfm1 tethers Hst1 to Sum1 at specific promoters | S. cerevisiae | High; explains promoter-local (not domain-wide) action |
| 29066473 (Jaiswal 2017) | Mutant phenotype + ChIP | Qualifies (1b) | In vivo histone substrate | Loss of Sum1–Hst1 → increased H4K5ac at target loci | S. cerevisiae, vegetative | Moderate–high; identifies H4K5 (not H3K9/H3K14) as the documented mark |
| 11316790 (Rusché & Rine 2001) | Mutant / ChIP | Qualifies (2),(3) — context-gates them | Hst1 in silencing | SUM1‑1 relocalizes Sum1 to HM loci via ORC; Hst1 gives Sir-independent silencing w/ nucleosome hypoacetylation | Engineered SUM1‑1, HM loci | High; explicitly a gain-of-function, non-native context |
| 11313477 (Sutton 2001) | Mutant / genetic | Qualifies (2),(3) | Hst1 requirement in SUM1‑1 silencing | SUM1‑1 silencing at HMR requires Hst1 + its NAD⁺ activity + ORC | Engineered SUM1‑1, HMR | High; confirms context dependence |
| 26748095 (Simoneau 2016) | Genome-wide fitness + genetics | Qualifies (4) | Sirtuins & DNA-damage | Genotoxin phenotypes mainly from Hst3/Hst4 (H3K56ac), not Hst1 | S. cerevisiae | High that DNA-damage role is largely non-Hst1 |
| 24710277 (Corbi 2014) | ChIP + genetics | Qualifies (5) | NDT80/meiotic control | Sum1 (not Hst1) determines NDT80 DNA occupancy; Hst1 is the enzymatic partner | S. cerevisiae meiosis | Moderate; recombination link is indirect/transcriptional |
| 26122277 (De Las Peñas 2015, review) | Review/database | Qualifies (2) | Local vs regional silencing | In C. glabrata: Hst1+Sum1–Rfm1 = local silencing; Sir2 = regional/telomeric | Candida glabrata | Orientation only; consistent cross-species division of labor |
| 8810037 (Derbyshire 1996) | Mutant phenotype (founding characterization) | Refutes (2),(5) for native Hst1 | rDNA recombination & regional silencing | hst1Δ shows no phenotype in rDNA recombination or regional HMLα silencing; HST1 is 71% identical to SIR2 | S. cerevisiae, native loss-of-function | High; original clean deletion test — argues against native recombination/regional roles |
| UniProt/GO P53685 (computed) | Database (evidence codes) | Frames all claims | Annotation provenance | H3K14/H3K9/H4K16 MF, subtelomeric-heterochromatin BP, DNA-damage BP = IBA; telomeric-region CC = IEA; only general NAD-HDAC = IDA | Annotation audit | Direct computed provenance (see CSV artifact) |
| HST1 vs SIR2 alignment (computed) | Structural/evolutionary | Qualifies (1b) | Enzymatic-core conservation | Catalytic core 82% identical to Sir2; catalytic His310 conserved — near-identical enzyme, so site "choice" is set by targeting/cofactors, not intrinsic chemistry | Needleman-Wunsch, UniProt seqs | Direct computed provenance (see identity artifact) |
Retain (well-supported):
- MF GO:0017136 histone deacetylase activity, NAD-dependent — IDA-backed. Keep as the primary MF.
- CC GO:0000118 / Sum1–Rfm1–Hst1 (histone deacetylase complex), GO:0005634 nucleus — supported by interaction data.
- BP: gene-specific negative regulation of transcription; regulation of NAD⁺ biosynthesis (well-documented, PMID:12972620); negative regulation of meiotic/sporulation genes (PMID:17242192, 12612074, 24710277).
Scrutinize / likely too specific (paralog over-annotation):
- MF GO:0032041 (H3K14) and GO:0046969 (H3K9) — IBA only. Recommend do not assert as native experimental MF; the direct evidence supports only the general NAD-HDAC term, and the one documented in vivo mark is H4K5ac. Lead: generalize to GO:0017136, or, if a residue-specific term is desired, the best-evidenced candidate is an H4K5-directed activity (still indirect). GO:0046970 (H4K16) likewise IBA — not native-supported.
Make context-specific / non-core:
- BP GO:0031509 subtelomeric heterochromatin formation (IBA) and CC GO:0000781 telomeric region (IEA) — recommend remove or down-weight for native Hst1; these are Sir2 functions. If retained, gate to the SUM1‑1 engineered context.
- BP GO:0030466 silent mating-type heterochromatin (IMP) — retain only with a SUM1‑1 / gain-of-function qualifier; it is not native targeting.
Treat as unresolved / needs primary check:
- BP GO:0006974 DNA damage response (IBA) — recommend non-core; phenotype dominated by Hst3/Hst4.
- BP GO:0045950 negative regulation of mitotic recombination (IMP) — flag for curator: the founding hst1Δ study (PMID:8810037) found no rDNA-recombination phenotype, so verify the exact locus/context behind the IMP annotation and whether it reflects direct control vs. secondary transcriptional silencing; recommend treating as non-core pending that check.
Avoiding "protein binding": the recommended informative MF is NAD-dependent histone/protein-lysine deacetylase activity (GO:0017136), not a generic binding term.
Direct gene-product activity: NAD⁺-dependent deacetylation of promoter-proximal nucleosomal histones (in vivo mark: H4K5ac) at loci bound by the DNA-binding protein Sum1, to which Hst1 is tethered by Rfm1. This is a local, sequence-directed activity — not domain-wide heterochromatin nucleation/spreading.
Downstream / inferred (not the immediate function): transcriptional repression outputs (middle-sporulation genes; BNA NAD-biosynthesis genes; oxidative-stress genes), and any recombination or meiotic-progression phenotypes that follow from that repression. Telomeric/heterochromatic and DNA-damage "functions" are either paralog-inferred (Sir2/Hst3/Hst4) or gain-of-function (SUM1‑1) manifestations.
Artifacts: HST1_GO_evidence_matrix.csv (computed GO/evidence-code audit for P53685); HST1_vs_SIR2_identity.txt (computed Needleman-Wunsch identity: 82% catalytic-core identity to Sir2, His310 conserved).
Iteration 2 update: Added founding-paper loss-of-function evidence (PMID:8810037: hst1Δ has no rDNA-recombination or regional HMLα-silencing phenotype), which strengthens the refutation of the native recombination and heterochromatin sub-claims, and added a computed HST1-vs-SIR2 catalytic-core alignment supporting the "targeting-not-enzyme" basis of Hst1/Sir2 divergence.
Iteration 3 update: Targeted search for a primary in-vitro Hst1 histone-site assay returned none, reinforcing that H3K9/H3K14 site-specificity for native Hst1 rests on homology inference rather than direct biochemistry. Consolidated a per-term GO decision table (HST1_GO_decision_table.csv) with recommended actions.