SULT1A1 (human) — curation notes

UniProt: P50225 (ST1A1_HUMAN), 295 aa. EC 2.8.2.1. PANTHER family
PANTHER:PTHR11783 ("SULFOTRANSFERASE SULT").

Provenance note. No deep-research provider is reachable in this environment
(ai-gene-review exposes no deep-research subcommand here), so there is no
SULT1A1-deep-research-<provider>.md. Per the project instructions, the literature
synthesis is recorded here instead. Every assertion below is anchored to a cached
publication in publications/, to the UniProt record, or to a named web source.

1. What the protein is

SULT1A1 is a cytosolic (soluble) sulfotransferase — the "thermostable"
phenol-sulfating phenol sulfotransferase (TS-PST / P-PST-1), historically also
called HAST1, ST1A3 and aryl sulfotransferase 1. It transfers the sulfonate group
of PAPS (3'-phosphoadenosine 5'-phosphosulfate) onto the hydroxyl (or amine)
of small acceptor molecules, releasing PAP.

[file:human/SULT1A1/SULT1A1-uniprot.txt, "Sulfotransferase that utilizes 3'-phospho-5'-adenylyl sulfate
(PAPS) as sulfonate donor to catalyze the sulfate conjugation of a wide variety of acceptor molecules
bearing a hydroxyl or an amine group."]

It is an obligate homodimer (UniProt SUBUNIT: Homodimer, ECO:0000269 from
PMID:12471039, PMID:16221673, PMID:20417180), dimerising through the conserved
C-terminal KTVE motif shared across the cytosolic SULTs.

Catalytic machinery (UniProt FT):
- ACT_SITE 108 — proton acceptor (His108)
- BINDING 48..53, 130, 138, 193, 227..232, 255..259 — PAPS (CHEBI:58339)
- BINDING 106..108 — substrate (ECO:0000269|PubMed:12471039)

The structural work confirms this directly: "The refined 3D structure of
SULT1A1-PAP-2NAP revealed a single 2NAP molecule within pocket-1, positioned in a
catalytically competent manner in which the hydroxyl group is positioned 2.34 Å from
the catalytic amine group of His108 and 3.33 Å from Lys106"
[PMID:22069470 full text].

Localisation: cytoplasm/cytosol. UniProt SUBCELLULAR LOCATION: Cytoplasm
(ECO:0000269|PubMed:8093002); Reactome places every SULT1A1 reaction in the cytosol.
No membrane, signal peptide or targeting features in the record.

Tissue distribution: UniProt TISSUE SPECIFICITY: Liver, lung, adrenal, brain, platelets and skin. SULT1A1 is the dominant hepatic SULT (>50% of total SULT protein
in liver), with substantial gut expression; platelet SULT1A1 activity is the classic
surrogate phenotype used in pharmacogenetic studies
(ScienceDirect SULT1A1 topic page).

2. Substrate profile — what the enzyme is actually for

The defining feature is broad specificity with a strong preference for small
phenols, achieved by a plastic, largely hydrophobic acceptor pocket:

"For example, SULT1A1 [8] displays substrate preference for small phenolic
compounds, while SULT1E1 shows a preference for estrogen acceptors [9]."
PMID:22069470

"Cytosolic sulfotransferases (SULTs) are mammalian enzymes that detoxify a wide
variety of chemicals through the addition of a sulfate group." PMID:22069470

2.1 Small phenols / model xenobiotics (highest-confidence core)

• 4-nitrophenol, Km 0.6 µM — the canonical P-PST substrate.
  "COS-expressed HAST1 was shown to be enzymatically active in sulphating
  p-nitrophenol with high affinity (Km 0.6 microM)" PMID:8093002.
  (HAST1 = SULT1A1; HAST3 = SULT1A3.)
• 2-naphthol, 3-cyano-7-hydroxycoumarin — co-crystallised acceptors PMID:22069470.
• Drugs: acetaminophen/paracetamol, minoxidil (UniProt, by similarity;
  Reactome R-HSA-158468 and R-HSA-9753277 both place APAP sulfation on SULT1A1).

2.2 Iodothyronines — the highest-affinity documented substrates

This is the most striking kinetic fact about the enzyme and is under-represented in GO.

"In all cases, the substrate preference was 3,3'-T2 >> rT3 > T3 > T4."
PMID:10199779

"The apparent Km values of 3,3'-T2 and T3 [at 50 micromol/L
3'-phosphoadenosine-5'-phosphosulfate (PAPS)] were 1.02 and 54.9 micromol/L for
liver cytosol, 0.64 and 27.8 micromol/L for kidney cytosol, 0.14 and 29.1
micromol/L for SULT1A1, and 33 and 112 micromol/L for SULT1A3, respectively."
PMID:10199779

Km for 3,3'-T2 is 0.14 µM — ~240× tighter than SULT1A3's 33 µM, and the tightest
Km anywhere in this gene's record. UniProt also records KM=0.12 uM for 3,3'-diiodothyronine. Sulfation channels iodothyronines toward degradation:
"Sulfation is an important pathway of thyroid hormone metabolism that facilitates
the degradation of the hormone by the type I iodothyronine deiodinase" PMID:10199779.

Known unknown (the authors' own words):

"Of the latter, SULT1A1 clearly shows the highest affinity for both iodothyronines
and PAPS, but it remains to be established whether it is the prominent isoenzyme for
sulfation of thyroid hormone in human liver and kidney." PMID:10199779

There is no GO term for iodothyronine/thyroid-hormone sulfotransferase activity
(checked: GO:0008146 descendants contain nothing thyroid-related). Proposed below.

2.3 Procarcinogen bioactivation (the "dark side" of the same chemistry)

SULT1A1 O-sulfonates N-hydroxy-arylamines and N-hydroxy-heterocyclic amines,
producing unstable sulfate esters that decompose to nitrenium ions and form DNA adducts.

"In the 12 human hepatic cytosols studied, the extent of
3'-phosphoadenosine-5'-phosphosulfate-dependent DNA binding of the N-hydroxy
derivatives were all significantly correlated with levels of thermostable phenol
ST (TS-PST) activity but not with thermolabile phenol ST or dehydroepiandrosterone
ST activities." PMID:7834621

"A major human sulfotransferase, SULT1A1, metabolizes and/or bioactivates many
endogenous compounds and is implicated in a range of cancers because of its ability
to modify diverse promutagen and procarcinogen xenobiotics." PMID:12471039

Reactome R-HSA-158860 ("SULT1A1 dimer sulfonates NHABP") captures the
N-hydroxy-4-aminobiphenyl case. Mouse SULT1A knockout/humanisation studies confirm
the toxification role in vivo for 4-aminobiphenyl and methyleugenol
(Essays Biochem 2024, PMC11625864).

2.4 Dietary flavonoids

"SULT1A2 and SULT1A1 catalyze preferably and most efficiently the formation of
hesperetin 3'-O-sulfate, and SULT1C4 catalyzes preferably and most efficiently the
formation of hesperetin 7-O-sulfate." PMID:20056724

Important for annotation: hesperetin is a flavanone, and the SULT1A1 product is
the 3'-O-sulfate (B-ring). The GO term currently used, GO:0047894 flavonol 3-sulfotransferase activity, is defined as "3'-phospho-5'-adenylyl sulfate +
quercetin = adenosine 3',5'-diphosphate + H+ + quercetin 3-sulfate" — a different
substrate class (flavonol) and a different position (3, C-ring). Since hesperetin has
no C-ring 3-hydroxyl at all, no assay on hesperetin can demonstrate flavonol
3-sulfotransferase activity, which is why the argument holds without needing to know
what else the paper tested. GO:1990135 flavonoid sulfotransferase activity is the
correct term. (Note these two are siblings under GO:0008146, not parent/child.)

Full-text caveat. PMID:20056724 is abstract-only in the cache — the article is
not open access and Europe PMC reports no PMC record (isOpenAccess: N, inPMC: N),
and its text-mined annotations cover only the title and abstract. So it could not be
checked whether quercetin appears anywhere in the full methods. The MODIFY above
rests on hesperetin's structure and on the product the abstract itself names, not on
a claim about what was or was not assayed; a curator with the full text should
confirm before the term is changed.

2.5 Gut-microbial 4-ethylphenol → 4-EPS

"We observed sulfation of 4EP to 4EPS by the sulfotransferase SULT1A1 and others
during in vitro biochemical reactions" PMID:35165440

4-EPS crosses the blood–brain barrier, impairs oligodendrocyte maturation and
myelination, and produces anxiety-like behaviour in mice. UniProt records this with
ECO:0000269|PubMed:35165440 and RHEA:70607. The paper itself flags the open question:
"though the site(s) of 4EP sulfation remains unknown". Not currently in GOA —
a genuine annotation gap.

2.6 Marginal / low-affinity activities (real in vitro, weak in vivo)

These are the ones most at risk of over-annotation:

"HAST1 could also sulphate dopamine, as could HAST3 sulphate p-nitrophenol, but the
Km for these reactions were at least two orders of magnitude greater than for the
preferred substrates." PMID:8093002

"The crystal structure of SULT1A1 that we present here has PAP and one molecule of
E2 bound in a nonproductive mode in the active site." PMID:16221673

"Ethyl sulfate, a minor and direct ethanol metabolite in adult human body, has been
implicated as a biomarker for alcohol consumption and in utero exposure to ethanol."
PMID:23207770

Note also that ethanol is an aliphatic alcohol, not a phenol — so the
PMID:23207770-based GO:0004062 aryl sulfotransferase activity IDA is mis-typed for
its own evidence; GO:0004027 alcohol sulfotransferase activity is what that
experiment shows.

3. Substrate inhibition and active-site plasticity

A recurring theme across three structural papers: SULT1A1 binds a second, catalytically
unproductive acceptor molecule, which explains its partial substrate inhibition.

"An unexpected finding is that the enzyme accommodates not one but two molecules of
the xenobiotic model substrate p-nitrophenol in the active site." PMID:12471039

"In agreement with previous reports, the enzyme shows partial substrate inhibition at
high concentrations of E2." PMID:16221673

"We found that active site plasticity enables binding of different acceptors and
identified dramatic structural changes in the SULT1A1 active site leading to the
binding of a second acceptor molecule in a conserved yet non-productive manner."
PMID:22069470

This plasticity is the mechanistic explanation for the broad specificity, and is why
substrate-specific MF terms should be applied to SULT1A1 sparingly.

4. Protein–protein interactions

• SULT2B1 (O00204) — UniProt INTERACTION: P50225; O00204: SULT2B1; NbExp=8.
  Four separate large-scale interactome papers (PMID:25416956, PMID:25910212,
  PMID:31515488, PMID:33961781) each generate a bare GO:0005515 protein binding IPI
  with WITH/FROM UniProtKB:O00204. None of these assigns a molecular function.
• SULT4A1 (Q9BR01) — mechanistically informative, unlike the above:

  "Using immunoprecipitation, SULT4A1 was shown to interact with both SULT1A1 and
  SULT1A3 when expressed in human cells." PMID:32152050
  "Mutation of the conserved dimerization motif located in the C terminus of the
  sulfotransferases prevented this interaction." PMID:32152050

SULT4A1 lowers SULT1A1/1A3 protein levels in neuronal cells and may act as a
chaperone. Because the interaction is shown to require the conserved dimerisation
motif, GO:0046982 protein heterodimerization activity is a defensible, more
informative replacement for the bare protein binding term.
- Homodimerisation (GO:0042803) is experimentally established (UniProt SUBUNIT,
3 ECO:0000269 references) but is not annotated in GOA — an annotation gap.

5. Reference provenance problem: PMID:8661000 is about the paralog

Two legacy TAS annotations assigned by PINC in 2003 (GO:0008146 sulfotransferase activity and GO:0009308 amine metabolic process) cite:

Her C, Raftogianis R, Weinshilboum RM (1996) "Human phenol sulfotransferase STP2
gene: molecular cloning, structural characterization, and chromosomal localization."
Genomics 33:409–420. PMID:8661000

STP2 is SULT1A2, not SULT1A1. Checked directly against UniProt:
- P50225 (SULT1A1): GN Name=SULT1A1; Synonyms=STP, STP1;
- P50226 (SULT1A2): GN Name=SULT1A2; Synonyms=STP2; — and PMID:8661000 is
reference 4 of the P50226 entry.

The abstract is explicit that the object of study is the STP2 gene:

"We have determined the structure and chromosomal localization of the gene for one
of these two cDNAs, STP2, as a step toward understanding molecular genetic mechanisms
involved in the regulation of this enzyme activity in humans." PMID:8661000

What the full text contains (read by the project curator; the cached record here is
abstract-only, as the article is not open access and has no PMC record):

• STP2/SULT1A2 is the only gene newly characterized experimentally. The other
  sulfotransferase genes appear purely as comparisons: STP1 → SULT1A1 (its protein
  reported as 96% identical to STP2's), STM → SULT1A3 and STE → SULT1E1
  (exon–intron structure comparisons), plus rat phenol- and guinea-pig
  estrogen-sulfotransferase genes at conserved splice junctions. None of these is
  characterized in the paper.
• Exons IA and IB are alternative noncoding first exons — alternative 5′ UTRs of
  STP2, not additional genes.
• There is no new enzymology of any kind. No recombinant enzyme assays, no
  substrate-specificity experiments, no K_m / V_max or catalytic-rate measurements, no
  thermostability assays, and no functional comparison of the STP1 and STP2 proteins.
  The enzymology present is background: that thermostable PST sulfonates relatively
  simple planar phenols, and that its activity varies heritably between individuals.
  The new results are entirely molecular genetics — genomic cloning and sequencing, the
  nine-exon structure, the alternative 5′ ends, promoter-region analysis and
  chromosome-16 localization — explicitly framed as enabling later investigation of
  how enzyme activity is regulated.

So both TAS annotations are unsupported by this reference. GO:0008146 happens to be
true of SULT1A1 anyway (abundant independent experimental support), so that row
survives on its merits but should be re-sourced. GO:0009308 amine metabolic process
has neither support from this reference nor a good fit to the gene — SULT1A1 is the
phenol-preferring form, while monoamine sulfation belongs to SULT1A3.

Where the amine annotation probably came from. The only amine-adjacent statement in
the paper is its opening background sentence: "Sulfonation is an important pathway in
the biotransformation of many drugs, xenobiotics, neurotransmitters, and steroid
hormones." PMID:8661000 That is a claim about the SULT family in general, not about
this gene and not a result of this paper — a textbook bad TAS.

Action taken: MARK_AS_OVER_ANNOTATED, though REMOVE is now equally defensible.
A TAS with no traceable statement supporting it for this gene is arguably not a valid
TAS at all. It is kept at over-annotated only because the term is not biologically
false of SULT1A1, which does act on amine-bearing molecules (dopamine,
N-hydroxy-arylamines) — although note it sulfonates those on oxygen, not on
nitrogen, so even the chemistry the term implies is not this enzyme's.

Note on full_text_unavailable: true. That flag is set on this reference and on
PMID:20056724. It describes the cached record — which is what the quote validator
reasons about — and not the reviewer's access. It does not mean the full text was
never consulted.

Nomenclature confirmed via
Her et al. 1996 / STP1-STP2 genomic organisation, PMID:8912648
and OMIM 171150.

6. Polymorphism / pharmacogenetics (context, not GO-annotatable)

UniProt POLYMORPHISM: There are several alleles. The sequence shown is that of allele SULT1A1*3. The best-known variant is Arg213His (SULT1A1*2, rs1042028) —
UniProt: "R -> H (in allele SULT1A12; has a lower sulfotransferase activity)"*. It
reduces both activity and thermostability and is the basis of a very large
epidemiological literature on cancer risk and drug response. Copy-number variation at
the locus also modulates activity. These are allele-level facts and do not by
themselves justify GO annotations.

7. Loss-of-function

Sult1a1 knockout mice are viable and outwardly normal — "Sult1a1 or Sult1d1 knockouts
were healthy and showed no obvious deficiencies", in contrast to Sult1e1 (reduced
fertility) and Sult4a1 (early postnatal death)
(Essays Biochem 2024, PMC11625864).
The knockouts are protected against DNA-adduct formation by 4-aminobiphenyl and
methyleugenol, confirming the bioactivation role in vivo. So the gene's indispensable
role is chemical-defence/bioactivation rather than a developmental program.

8. Annotation decisions taken (summary)

The GO:0008146 IDA from PMID:10199779 was left as ACCEPT rather than modified to
GO:0004062: the same paper already carries a GO:0004062 EXP row, so the modification
would only have created a duplicate. What is actually missing there is a thyroid-hormone
sulfotransferase term, which is raised under proposed terms instead.

One deliberate validation warning

ai-gene-review validate reports exactly one warning:

Inconsistent review actions for term GO:0004062: ACCEPT (IBA, IEA, EXP×5, IDA×2);
MODIFY (IDA)

This is intentional and is explained in the review's reason field. The consistency
check is a good heuristic — a term is normally either right for a gene or not — but here
the divergence is evidence-level, not gene-level: SULT1A1 unambiguously has aryl
sulfotransferase activity (nine rows on phenolic acceptors), while the single
PMID:23207770 row assayed only ethanol, an aliphatic alcohol, which cannot
demonstrate a reaction GO defines as requiring a phenol. The file is therefore DRAFT
rather than COMPLETE.

Proposed new terms

1. iodothyronine sulfotransferase activity (with thyroid hormone sulfotransferase
   activity as an exact synonym) — no GO term exists for the iodothyronine sulfation
   reaction, despite it being SULT1A1's highest-affinity chemistry and having four RHEA
   reactions (RHEA:67876, 67888, 67892, 83575) in UniProt. The substrate-class label is
   preferred because the highest-affinity acceptors (3,3′-T2, rT3) are inactive
   metabolites, not the active hormones.

9. OpenScientist hypothesis runs

Three runs against genes/human/SULT1A1/SULT1A1-hypotheses/, each scoped to a single
computable question (3 iterations, 7200s job timeout). All three produced real
computation with provenance artifacts, not just literature summaries.

9.1 The endogenous-acceptor gap was mis-framed

The strongest result. Three independent analyses — cross-species orthologs, cross-paralog
comparison, and human population genetics — all trend against the hypothesis that the
acceptor pocket is tuned to a dedicated endogenous acceptor:

• Pocket less conserved than scaffold across 10 mammalian orthologs (0.836 vs 0.868).
• Pocket is the most divergent region across 8 human SULT1 paralogs (0.511 vs 0.629)
  — a specificity-determining-position signature.
• Phe247 ortholog identity 0.20 and carries a common human missense variant;
  Ile89 identity 0.50. gnomAD shows no pocket-specific missense depletion.
• The classic low-activity allele Arg213His lies outside the pocket.

Reading: the pocket is a promiscuity-optimised, specificity-diversifying module.
Dedicated specificity in this subfamily lives in the paralog — SULT1A3, via Glu146.

Honest caveat. The individual pocket-vs-scaffold tests are non-significant
(n = 11 pocket residues; MWU p = 0.499, 0.10, 0.444; permutation p = 0.102). This is
absence of evidence for focused constraint, with three lines trending the same way,
not positive proof of promiscuity. The hard anchors are the specific poorly-conserved
residues and the common Phe247 variant, not the p-values. The gap is recorded as
status: NARROWING, not resolved.

9.2 Two references the runs surfaced

Both were absent from my review; both are now cached, and both quotes were checked
verbatim against the freshly fetched abstracts rather than trusted from the reports:

• PMID:9855620 "The change of a single amino acid, E146A, was sufficient to
  transform the catalytic properties and substrate preference of SULT1A3, such that they
  closely resembled those of SULT1A1." The residue that makes an enzyme a dopamine
  sulfotransferase is exactly the one SULT1A1 lacks (Ala146). This is far stronger
  support for the dopamine over-annotation call than the Km comparison alone.
• PMID:20417180 Lu et al. 2010 — the Phe247 selectivity switch, and the third
  UniProt homodimer reference, which I had cited only in prose.

9.3 An error the runs caught in my own proposal

My first draft justified the new term partly by saying it "would also apply to SULT1A3,
SULT1B1 and SULT1E1, all of which sulfate iodothyronines." The SULT1E1 claim was
asserted without checking and is wrong. Verified directly against UniProt P49888: its
five curated reactions are estrone, 24S-hydroxycholesterol, 17β-estradiol, DHEA and
4-ethylphenol — no iodothyronine at all. SULT1A3 (four curated iodothyronine
reactions) and SULT1B1 (three) do hold. Corrected in the review.

9.4 What the iodothyronine mechanism run does and does not license

The measured claim (~240-fold Km advantage over SULT1A3) is confirmed. The pocket
mechanism — SULT1A1/1A2 pockets hydrophobic and acid-free, SULT1A3 carrying Glu89/Glu146
for a net −1.9 charge — is a coherent correlate only: there is no iodothyronine-bound
structure and no iodothyronine-specific mutagenesis, and the signature is shared with
SULT1A2 (1 of 22 pocket positions differ). Recorded in the review as a hypothesis-level
lead with a discriminating test, not as a curated fact.

Operational note: the first launch of run 1 exited 0 after ~2 min with an upstream
HTTP 502 and an empty output directory — a third failure signature beyond the two the
skill catalogues. Cleaned up and retried once; the retry succeeded.

Gaps in GOA worth filling

• GO:0042803 protein homodimerization activity — the enzyme is an obligate homodimer
  with three experimental UniProt references, and Reactome models every reaction on the
  "SULT1A1 dimer"/"SULT1A1 homodimer".
• 4-ethylphenol sulfation (RHEA:70607, PMID:35165440) is in UniProt but absent from GOA.

10. Thyroid-hormone literature added on PR review

Four papers were fetched during the OpenScientist work but initially cited only inside
the hypothesis reports. PR #2835 review flagged this; they are now in references:.
Three of them bear directly on the iodothyronine story:

• PMID:11739018 (Li et al. 2001, J Endocrinol) — independent replication of the
  substrate ordering in a second laboratory, across both common allozymes, extending the
  series with 3,5-T2 as by far the poorest acceptor:
  PMID:11739018. It also shows the
  Arg213His polymorphism changes catalytic activity and thermostability, not
  substrate specificity.
• PMID:15531517 (Ebmeier & Anderson 2004, JCEM) — SULT1A1 activity measured in
  native human thyroid cytosol across 86 surgically obtained glands. This is what moves
  iodothyronine sulfation past recombinant-enzyme-only evidence, and is why thyroid
  was added to the tissue list in the description.
• PMID:9848125 (Schuur et al. 1998) — hydroxylated PCBs inhibit 3,3'-T2 sulfation
  by hSULT1A1 but not hSULT1A3, discriminating the two closest paralogs on this
  chemistry. Note the human enzymes are recombinant (V79 cells) and the liver cytosols
  assayed are rat, so it speaks to isozyme discrimination, not to which enzyme
  carries the activity in human liver.

PMID:28109953 is a review, added for pathway orientation only and marked
relevance: LOW; it carries no annotation.

The thyroid knowledge gap stays open. PMID:15531517 measures tissue activity and
its abstract says "only 3,3'-T2 and daidzein served as substrates for the normal thyroid
SULT activities" — plural, unresolved between 1A1 and 1A3. It narrows the
recombinant-only concession without attributing the activity to the gene product by
knockdown or immunodepletion, so it does not close the gap.

10.1 A chemistry error, and its correction

A first attempt at wiring PMID:11739018 in described 3,5-T2 as "the one iodothyronine of
the series without an outer-ring hydroxyl", and cited that as licensing the 4′-
regiospecificity clause of the proposed term. That is wrong, and it was recorded in
two places.

Thyronine is 4-(4-hydroxyphenoxy)-L-phenylalanine, so the 4′-phenolic hydroxyl is part of
the scaffold and is present in every iodothyronine, 3,5-T2 included. Verified against
ChEBI:89575, whose SMILES NC(Cc1cc(I)c(Oc2ccc(O)cc2)c(I)c1)C(=O)O shows an intact
4-hydroxyphenoxy outer ring with both iodines on the inner tyrosyl ring. 3,5-T2 is
itself sulfated at the 4′-OH — 3,5-T2 sulfate is a documented metabolite. It is simply a
poor SULT1A1 acceptor.

What 3,5-T2 lacks is outer-ring iodination, and re-reading the series that way is a
sharper result than the one originally written:

So SULT1A1 selects for iodine substitution adjacent to the 4′-hydroxyl — plausibly via
the lowered phenol pKa — rather than for the hydroxyl itself. The 4′-regiochemistry in the
proposed definition is instead sourced to the identity of the products (the iodothyronine
sulfates are 4′-O-sulfates) and to the UniProt/RHEA reactions already cited.

Worth recording that the OpenScientist report had this right — it calls 3,5-T2 "an
inner-ring diiodothyronine". The error entered when the finding was transcribed into YAML.

10.2 GO label confirmations

The review asked for live lookups on two ids used in proposed_replacement_terms, which
is range: Term and therefore not label-validated by the toolchain. Confirmed against
EBI QuickGO: