SET6 (YPL165C / UniProt Q12529) — research notes

Research journal for the GO-annotation review of Saccharomyces cerevisiae SET6.
Provenance recorded inline as [PMID:xxxxx "verbatim quote"] or [file:... ] for the
local bioinformatics analysis.

One-line summary

SET6 is a dark budding-yeast SET-domain protein: a putative, likely catalytically
competent protein-lysine methyltransferase of the SMYD subfamily whose substrate,
in-vivo activity, and biological role are all unknown. It is the paralog of the
characterized SMYD enzyme Set5.

What is KNOWN

Domain architecture / family

UniProt (Q12529) names it "Potential protein lysine methyltransferase SET6", EC 2.1.1.-,
373 aa, with a SET domain (residues ~12–338) and assigns it to the "class V-like
SAM-binding methyltransferase superfamily". PE level 3 ("Inferred from homology"). The
only experimental FUNCTION note is chemogenomic (see below); there is no methyltransferase
assay.
InterPro/PANTHER place SET6 in PANTHER family PTHR12197 ("HISTONE-LYSINE
N-METHYLTRANSFERASE SMYD"), subfamily SF294 ("POTENTIAL PROTEIN LYSINE METHYLTRANSFERASE
SET6"); CDD cd20071 SET_SMYD (from the UniProt DR lines). So the family-level signal is
clearly SMYD-type SET domain.
SET6 and its paralog Set5 are the two yeast SMYD proteins:
[PMID:31685550 "Saccharomyces cerevisiae contains two SMYD proteins, Set5 and" ... "which
share structural elements with the mammalian SMYD enzymes."] and
PMID:31642774.

Catalytic competence (bioinformatics, this review)

Local reproducible analysis (file:yeast/SET6/SET6-bioinformatics/RESULTS.md) shows the
SET domain is present and near-complete (hmmsearch vs Pfam PF00856, domain E=6.3e-14,
residues 23–337, split-SET/SMYD architecture) and, critically, that the invariant
catalytic/SAM-binding residues are intact: the "NHSC" pocket (Asn303-His304-Ser305-
Cys306) and the C-terminal catalytic Tyr337 are conserved, exactly as in the active
enzymes Set1/Set2/Set5/SMYD2/SMYD3. In contrast, the two yeast SET proteins that are
reported to be catalytically dead — Set3 and Set4 — have the Asn-His replaced by
Arg-Arg. On primary sequence SET6 therefore groups with the active methyltransferases,
not with the pseudoenzymes. This independently reproduces the literature assertion below.
SET-domain-restricted identity ranks Set5 first (closest yeast relative), consistent with
the SMYD ortholog-pair literature.
SET6 is Cys-rich (17 Cys, 4.6%) in three clusters (an N-terminal CxxC, an internal-insert
cluster, and a C-terminal post-SET C-x-C...C), the expected SMYD-family zinc-knot/post-SET
metal-binding architecture — consistent with the RCA zinc-ion-binding annotation.
A separate canonical MYND zinc finger could not be confirmed (SET6 is 373 aa, shorter than
Set5 at 526 aa).

Literature: likely-active but substrate unknown

The most authoritative statement is from the near-complete yeast methylation-network
paper: of the yeast SET/7βS/SPOUT methyltransferases, only five are predicted to target
proteins PMID:37252976, and of the three plausibly-active ones Set6 is
singled out as the most likely genuine protein methyltransferase: PMID:37252976. The Set3/Set4 contrast is
explicit: PMID:37252976.
Same paper: SET domain enzymes are protein (not RNA/DNA/metabolite) methyltransferases:
PMID:37252976.
Whatever Set6 methylates is not yet identified: PMID:37252976. In Table 2 of that paper Set6 is listed as a putative protein/SET methyltransferase
with no assigned substrate and no disqualifying "Notes" (unlike Set3/Set4, which carry
active-site-substitution notes).

The one experimental phenotype

SET6 (YPL165C) appears in a chemogenomic (heterozygous-deletion) screen: it is "Involved
in resistance to compounds that target ergosterol biosynthesis, including fenpropimorph,
dyclonine, and alverine citrate" (UniProt FUNCTION, ECO:0000269|PubMed:14718668). UniProt
itself hedges that, because deletion has no growth effect absent the compounds, SET6 "is
more likely to be involved in compound availability" rather than in ergosterol biosynthesis
per se. The primary paper PMID:14718668 is a genome-wide small-molecule/deletion screen
("alverine citrate"); it is a fitness-profiling readout, not a molecular-function assay, so
it does not establish a methyltransferase activity, substrate, or a defined biological
process for SET6.

What is NOT known (the knowledge gaps)

Substrate — no protein (histone or non-histone) methylation site attributable to SET6
has ever been identified. The generic "histone methyltransferase" call is an
IBA/family-transfer inference from a mixed SMYD donor set; Set5 acts on H4K5, H4K8,
and H4K12 PMID:31642774 and mammalian SMYD proteins
have both histone and non-histone targets, so a histone-specific MF is not defensible
for SET6 itself.
In-vivo (or in-vitro) activity — SET6 has never been shown to transfer a methyl group
to anything; competence is inferred only from sequence-conserved active-site residues.
"Motif conservation is necessary but not sufficient for activity"
(file:yeast/SET6/SET6-bioinformatics/RESULTS.md).
Biological role — deletion has no standalone growth phenotype; the only phenotype is
drug-conditional fitness of uncertain mechanism. Subcellular localization is unmeasured:
PMID:31642774. The nucleus annotation is IBA family-transfer, not measured for SET6.

Annotation-by-annotation reasoning (GOA, 7 annotations)

#	Term	Evid	Ref	Decision	Rationale
1	GO:0016279 protein-lysine N-methyltransferase activity	IBA	GO_REF:0000033	ACCEPT	Best-supported MF: intact SET/SMYD active site + literature calling Set6 the most-likely genuine protein KMT. Putative, but this is the safest core-function call.
2	GO:0042054 histone methyltransferase activity	IBA	GO_REF:0000033	MARK_AS_OVER_ANNOTATED	Over-specific substrate class. Set5 supports the source-side histone call, but no histone substrate has been shown for Set6. The lysine-KMT MF (#1) is the defensible level.
3	GO:0005634 nucleus	IBA	GO_REF:0000033	KEEP_AS_NON_CORE	Plausible by SMYD family transfer, but Set6 localization is explicitly unmeasured; keep as non-core, low confidence.
4	GO:0006338 chromatin remodeling	IEA (from GO:0042054)	GO_REF:0000108	REMOVE	Inter-ontology inference chained off the over-annotated histone-MTase term (#2); no evidence Set6 remodels chromatin. Falls with its parent term.
5	GO:0008270 zinc ion binding	RCA	PMID:30358795	KEEP_AS_NON_CORE	Computational (zinc-proteome) prediction, corroborated by the Cys-rich SMYD zinc-knot/post-SET architecture in the bioinformatics analysis. Structural, non-core.
6	GO:0005575 cellular_component (root)	ND	GO_REF:0000015	ACCEPT	Root/ND placeholder; accept as-is per GO convention.
7	GO:0008150 biological_process (root)	ND	GO_REF:0000015	ACCEPT	Root/ND placeholder; accept as-is per GO convention.

Reference triage

PMID:37252976 (Hamey & Wilkins, PNAS 2023, near-complete yeast methylation network) —
HIGH; full text; directly adjudicates Set6 as the most-likely genuine yeast protein KMT
with unknown substrate. VERIFIED.
PMID:31642774 (Jethmalani & Green, review) — HIGH; SMYD homology, histone/non-histone
substrate range, unmeasured localization. VERIFIED (full text).
PMID:31685550 (Jaiswal et al., Set5 SMYD paper) — MEDIUM; establishes Set5/Set6 as the two
yeast SMYDs (context/paralog). VERIFIED (abstract).
PMID:30358795 (Wang et al., yeast zinc proteome) — MEDIUM; source of the RCA zinc-binding
annotation. VERIFIED (abstract; this is the RCA support).
PMID:14718668 (Giaever et al., chemogenomic profiling) — LOW; source of the only phenotype
(drug-conditional fitness); does not define molecular function. VERIFIED (abstract).
file:yeast/SET6/SET6-bioinformatics/RESULTS.md — HIGH; reproducible SET-domain integrity /
SMYD-family / zinc-architecture analysis.
GO_REF:0000015 / 0000033 / 0000108 — pipeline references (ND / IBA / inter-ontology IEA).

Deep research

just deep-research-falcon yeast SET6 --fallback perplexity-lite was attempted but timed
out (no report produced); per project policy no fabricated deep-research file was written.
These notes are the manual research journal instead.

2026-09-29 IBA re-review

Re-read the GOA IBA rows against interpro/panther/PTHR12197/PTHR12197-paint.tsv.
All three PAINT rows descend from PANTHER:PTN008534456.
Added explicit propagation_review metadata to the accepted
GO:0016279 protein-lysine N-methyltransferase activity IBA and the non-core
GO:0005634 nucleus IBA. The already-reviewed GO:0042054 histone methyltransferase activity over-propagation was left as a substrate-scoping problem on the same PTN node.
Promoted the broad lysine-methyltransferase IBA from KEEP_AS_NON_CORE to ACCEPT
because it is the only plausible SET6 enzymatic activity and the broad, family-consistent
term avoids the unsupported histone-specific substrate. It remains explicitly putative in
the summary, reason, and knowledge gaps.
Removed structural GO:0008270 zinc ion binding from core_functions. It is retained as
a non-core RCA annotation, but its role is to support the SMYD fold rather than to define a
standalone evolved activity.
Searched PubMed for 2023-2026 SET6/YPL165C/Set6 yeast papers and found no newer
direct SET6 publication beyond the already-cached Hamey and Wilkins 2023 methylation-network
synthesis.

ND literature-search citations (2026-11-06)

existing_annotations[6], GO:0008150 biological_process (ND, GO_REF:0000015), ACCEPT: the ND may be stale. These notes cite PMID:14718668, a chemogenomic heterozygous-deletion screen that UniProt curates (ECO:0000269) as SET6 being "Involved in resistance to compounds that target ergosterol biosynthesis, including fenpropimorph, dyclonine, and alverine citrate". That is a loss-of-function phenotype for this gene product in S. cerevisiae tied to a drug response, so a curator should decide whether it supports a BP annotation. UniProt itself suggests the effect may reflect compound availability. The review action was left unchanged.