ECM30 (YLR436C / Q06673) curation notes

Journal of research for the AI GO-annotation review. ECM30 is an understudied ("dark")
gene: SGD classifies it as a "protein of unknown function" with molecular_function and
biological_process annotated ND. The primary deliverable of this review is therefore an
honest, sharply-delimited knowledge_gaps section plus carefully-reasoned, evidence-limited
description/core_functions.

Identity and family

• UniProt entry HID1_YEAST / Q06673, RecName "Protein ECM30", AltName "Extracellular
  mutant protein 30". Systematic name YLR436C, chromosome XII. 1274 aa, 145 kDa.
  [file:yeast/ECM30/ECM30-uniprot.txt]
• Gene name origin: "ExtraCellular Mutant" — the ECM series comes from the Lussier et al.
  1997 calcofluor-white / cell-surface transposon-mutagenesis screen
  PMID:9335584. SGD attributes the ECM30 gene name to this paper. This paper is the source of
  UniProt's FUNCTION line ("Required to form the correct cell wall composition"
  {ECO:0000269|PubMed:9335584}). NOTE: the cached PMID_9335584 is abstract-only
  (full_text_available: false); the abstract reports 82 genes but does not name ECM30
  individually — the gene-specific data are in the full text, which I have not read. Per
  curation policy I do NOT remove/contradict this experimental cell-wall attribution.
• Family = HID1 family (UniProt "Belongs to the hid-1 family"; Pfam PF12722 Hid1;
  InterPro IPR026705 HID1/Ecm30; PANTHER PTHR21575 "PROTEIN HID1").
  [file:yeast/ECM30/ECM30-uniprot.txt]
• MNEMONIC-COLLISION CAVEAT: UniProt's mnemonic "HID1_YEAST" reflects family membership, not
  a human-disease connection. ECM30 is the S. cerevisiae ortholog of metazoan HID1.
• No paralog in S. cerevisiae. SGD lists an empty paralogs array for ECM30
  [SGD backend locus S000004428, paralogs: []]. (The task brief's suggestion of a paralog
  "YKL215C" is NOT supported by SGD; YKL215C = OXP1, 5-oxoprolinase, unrelated.) ECM30 is the
  single-copy yeast member of PTHR21575. By contrast S. pombe has three HID1-family
  paralogs — hid1 (SPAP27G11.12 / O13776), hid2 (Q9HDV3), hid3 (Q9P7M6)
  [file:interpro/panther/PTHR21575/PTHR21575-entries.csv].

PANTHER family membership (PTHR21575) — reviewed members

From interpro/panther/PTHR21575/PTHR21575-entries.csv (8 reviewed proteins, 4027 total,
7556 taxa):

Metazoan ortholog (human HID1, Q8IV36) — for orthology reasoning ONLY

• Human HID1 FUNCTION (UniProt Q8IV36): "Plays a role in the biogenesis of large dense core
  vesicles (LDCVs) at the trans-Golgi network; necessary for the proper sorting of cargo into
  LDCVs and is required for proper regulated exocytosis (also known as regulated secretory
  pathway)." {ECO:0000269|PubMed:37751424}
  [https://rest.uniprot.org/uniprotkb/Q8IV36.txt]
• Human HID1 localization: Cytoplasm; Golgi apparatus, trans-Golgi network membrane
  (lipid-anchor); "Shuttles between the cytosol and the Golgi apparatus."
  [https://rest.uniprot.org/uniprotkb/Q8IV36.txt]
• Human HID1 experimental GO (QuickGO, Q8IV36): IDA to Golgi trans cisterna (GO:0000138),
  Golgi medial cisterna (GO:0005797), Golgi apparatus (GO:0005794), cytosol (GO:0005829);
  IMP/ISS to dense core granule biogenesis (GO:0061110); IBA to secretory granule maturation
  (GO:0061792) and Golgi organization (GO:0007030). [QuickGO annotation search Q8IV36]
• KEY CAVEAT for propagation: LDCVs / regulated dense-core secretory granules are a
  metazoan/neuroendocrine feature. S. cerevisiae has no regulated dense-core secretory
  pathway, so the metazoan-specific processes/compartments of HID1 do NOT transfer wholesale.
  The conserved element is more plausibly a general Golgi-associated membrane-trafficking role.

What is KNOWN about ECM30 (evidence-anchored)

1. Subcellular localization: cytoplasm (cytosolic), from GFP-fusion high-throughput
   analysis [PMID:14562095 (Huh et al. global GFP localization); GOA HDA GO:0005737]. UniProt
   SUBCELLULAR LOCATION: Cytoplasm {ECO:0000269|PubMed:14562095}. Consistent with the human
   ortholog's cytosol↔Golgi shuttling (peripheral, not integral membrane).
2. Cell-wall phenotype (source of the gene name): ecm30 mutants show altered calcofluor-
   white sensitivity and abnormal relative levels of mannose and glucose [PMID:9335584; SGD
   description "may play a role in cell wall biosynthesis, mutants have abormal relative
   levels of mannose and glucose"].
3. Gap1p (general amino acid permease) sorting and transport defects in ecm30 mutants
   [SGD locus description, S000004428: "have Gap1p sorting and transport defects"]. This is a
   membrane-cargo intracellular-sorting phenotype (Gap1 traffics TGN → endosome → vacuole vs
   plasma membrane depending on nitrogen source). I could not resolve a single primary PMID
   naming ECM30 for this Gap1 defect from public search; recording it as an SGD-curated
   free-text phenotype, not asserting a specific citation.
4. Genetic interaction network centered on membrane trafficking (SGD interaction_details,
   BioGRID; n=87 genetic partners). Notable, functionally coherent hits:
5. GET pathway: GET1 (Phenotypic Suppression), GET2 (Negative Genetic + Phenotypic
   Suppression), GET3 (Phenotypic Suppression) — tail-anchored membrane-protein insertion
   at the ER.
6. Golgi/secretory: ARL1 (Phenotypic Enhancement; Golgi Arf-like GTPase), SEC14
   (Synthetic Growth Defect; TGN PI/PC transfer protein), SEC26 & SEC28 (Negative Genetic;
   COPI coatomer subunits).
7. Vacuolar/endosomal sorting: VPS1 (dynamin), VPS9, VPS17 (retromer).
8. Physical interactions (SGD/BioGRID, n=30 physical partners; mostly single-method
   high-throughput AC-MS). Trafficking-relevant: LAA1 (Affinity Capture-MS; TGN AP-1
   clathrin-adaptor-associated / AP-3 accessory), VAM6/VPS39 (Affinity Capture-MS; HOPS
   tethering complex, vacuole fusion), UBP15 (AC-MS + AC-Western; ubiquitin protease).
   Many other AC-MS hits (histones HHT1/HHT2/HTB1/HTB2, RNA-binding CAF20/PUF2/PUF3/DHH1,
   Pol II RPO21) are typical abundant-protein / RNA-tethered background for a large 145 kDa
   cytosolic protein and are not treated as functional.
9. Post-translational modification: phosphoserine at Ser635 and Ser1065 (large-scale
   phosphoproteomics) [UniProt MOD_RES; PMID:17330950, PMID:18407956, PMID:19779198]. Ser1065
   is a mitotic/Cdk1-context site (Holt et al.). Regulatory significance unknown.
10. Sequence architecture: 1274 aa, HID1 α-solenoid-type fold (Pfam Hid1 ×2 across the
    protein), multiple internal disordered/low-complexity stretches (MobiDB-lite regions at
    ~23-42, 405-439, 494-517, 803-842, 1100-1149). No transmembrane domain, no signal peptide,
    no catalytic-domain signature → consistent with a peripheral/cytosolic scaffold rather than
    an enzyme or integral membrane protein. [file:yeast/ECM30/ECM30-uniprot.txt]

What is NOT known (the real gaps)

• Molecular function (MF-dark). No biochemical activity assigned; SGD/UniProt MF = ND.
  The HID1 fold has no catalytic assignment in any organism. Whether ECM30 acts as a
  membrane-trafficking scaffold/adaptor (like human HID1 at the TGN) is inferred, not shown.
• Direct molecular partners. No validated, specific binding partner defines its activity;
  the AC-MS hits are unfiltered high-throughput. The mechanistic link to LAA1/GET/ARL1 is
  genetic/HT-physical, not a defined complex.
• Biological role / mechanism of the cell-wall phenotype. Why an ecm30 null perturbs
  cell-wall mannose/glucose and calcofluor sensitivity is unexplained — likely indirect via
  mis-sorting of cell-wall-biosynthetic cargo (cf. Gap1 mis-sorting), but this is a
  hypothesis, not established.
• Whether the metazoan HID1 Golgi/secretory role is conserved in yeast. ECM30 localizes
  to cytoplasm in the one HT study; it has no experimental Golgi localization in yeast. The
  IBA Golgi-cisterna/Golgi-organization annotations derive from metazoan orthologs performing
  a process (regulated dense-core secretion) that S. cerevisiae lacks.
• Redundancy / genetic buffering. ECM30 is non-essential and single-copy in S. cerevisiae
  (no paralog), yet the null is only modestly sick — implies functional buffering by parallel
  trafficking pathways rather than a dedicated paralog.

GO annotation review plan (existing_annotations, 9 total)

Rationale for the Golgi IBA calls: I am NOT removing an experimental annotation. These are
IBA electronic inferences over-propagated from metazoan HID1 (dense-core-vesicle
biogenesis at the TGN), a process absent in S. cerevisiae. Per CLAUDE.md, over-propagated
IEA/IBA inferences that can be argued against on biological grounds are legitimate
MARK_AS_OVER_ANNOTATED / MODIFY candidates. I keep cytosol/cytoplasm (supported) and record
the honest dark-gene status via the ND roots and the knowledge_gaps section.

Deep research status

Automated deep research did not yield a provider file. Falcon (Edison) timed out at 600s
on the initial run and again on a single retry (the shared Edison endpoint was saturated by
many concurrent sibling jobs at the time); the perplexity-lite fallback returned HTTP 401
(quota exhausted). No -deep-research-{provider}.md file was produced. This review is
therefore grounded directly in primary/curated sources: the UniProt record (Q06673), the
GOA TSV, cached publications (PMID:9335584, PMID:14562095, and the phosphoproteomics PMIDs),
the SGD locus record (S000004428: description, phenotypes, interactions, empty paralog list),
the PANTHER PTHR21575 family membership file, QuickGO annotations for the human ortholog HID1
(Q8IV36), and the human HID1 UniProt FUNCTION/localization. All assertions above carry inline
provenance. Per project policy, no self-authored content is named as a provider deep-research
file.

2026-09-21: full-gene re-review and actual ancestral placement

All nine original rows reviewed. Actual treeinfo lineage places Q06673 leaf PTN001853777 below the eukaryotic PTN000491103 Golgi/membrane/cytosol assertions. The old lack-of-TM and cytoplasm-only arguments do not exclude peripheral association; GO:0016020 explicitly includes attached proteins. The absence of dense-core secretory vesicles does not remove Golgi organization or cisternae. Restore all five inherited annotations to ACCEPT, with direct target measurements distinguished from evolutionary inference.

PMID:40899782 full text provides a new, substantive divergence lead: the budding-yeast Ecm30 sequence is greatly modified, and the organism has unstacked Golgi. Its actual functional experiments assay S. pombe SPAC17A5.16, proposed Hid1/Ftp105; they do not assay Ecm30. Correct the earlier notes' outdated assignment of Hid1 to SPAP27G11.12 (now proposed Hid3). Golgi organization is broader than stacking, so this comparative correlation does not itself establish target functional loss. A neutral focused OpenScientist question is queued after exact repository/global-cache searches found no prior report.

The existing Falcon narrative and artifact were read and their useful primary leads followed. Contrary to the earlier statement that Ecm30 has no validated specific partner, Benschop et al., PMID:20620961 includes the Results section “The Ubp15-Ecm30 Ubiquitin Protease Complex Is Involved in Methionine Metabolism” and validates the complex. Costanzo et al., PMID:20093466 directly links deletion to Gap1 sorting/transport defects and describes the complex. The local caches for those two papers are abstract-only; primary publisher/author text was checked separately. No protease activity is attributed to Ecm30, no generic protein-binding annotation is added, and perturbation phenotypes are not used to manufacture additional process annotations.

Actual tree lineages, IBD/IRD rows and GO definitions are retained in projects/IBA_REVIEW/rereview-2026-09-20/ecm30-fun19-paint-and-terms.json. The full-gene assessment retains specific experimental questions without treating them as disproof of positive ancestry.