AIGR Hypothesis Deep Research — Final Report

Target: Nicotiana attenuata A0A2H4GSI3 ("NaUGT1" candidate, UGT85A2_0, NIATv7_g26396)

Focus type: function_assignment
Hypothesis slug: naugt1-candidate-identity-and-nicotinate-glucosylation
Source: genes/NICAT/NaUGT1_candidate_UGT85A2_0/NaUGT1_candidate_UGT85A2_0-ai-review.yaml
Length: 485 aa · ORF: A4A49_26396 (= NIATv7_g26396) · Genome label: UGT85A2_0

Summary

The seed hypothesis proposes that N. attenuata A0A2H4GSI3 — the candidate called "NaUGT1" in the nicotine-biosynthesis project — catalyzes nicotinic acid N-glucosylation as a step of nicotine biosynthesis, and asks whether a characterized tobacco "UGT1" locus (Nitab4.5_0006222g0020) maps by orthology onto this exact sequence. After three iterations combining computed sequence analysis, curated cluster membership, and primary pathway biochemistry, the assignment is refuted / over-annotated. The protein's true substrate cannot be positively fixed from accessible data and should be reported as unresolved.

Three faulty premises each fail independently. First, the enzyme clade is wrong: by global sequence identity and by UniRef90 cluster membership, A0A2H4GSI3 sits in the iridoid/terpenoid O-glucosyltransferase clade (nearest characterized relative CrUGT8/UGT709C2, a reviewed 7-deoxyloganetic acid O-glucosyltransferase, at 59.8%), whereas the only retrievable plant N-glucosyltransferases (UGT76C4/C5, UGT72B1) belong to distinct families only ~37–38% identical. Second, the pathway framing is wrong: in tobacco, nicotine biosynthesis consumes free nicotinic acid by condensation with N-methylpyrrolinium, whereas nicotinic acid N-glucoside is a separate conjugation/storage branch (one of three fates of nicotinic acid), not a nicotine-biosynthetic step. Third, the orthology claim is unverifiable: the genuine >90%-identity N. tabacum orthologs of the target (A0A1S3YWH6 at 88.5%, A0A1S3Z7D3 at 87.0%) are themselves annotated 7-deoxyloganetic acid O-glucosyltransferases, and the cited locus Nitab4.5_0006222g0020 could not be confirmed as this protein's ortholog.

All UniProt functional terms for this accession are IEA-only (EC-based and TreeGrafter), and no direct enzyme assay of A0A2H4GSI3 exists in accessible literature. The earlier quercetin-3-O-glucosyltransferase hypothesis reflects the same electronic-annotation overreach and is likewise not positively established. The safest curation conclusion is a negative: this protein is not demonstrably a nicotinate N-glucosyltransferase, and its true acceptor substrate remains open, with the best-supported (still homology-only) lead being terpenoid/iridoid-type O-glucosylation.

Key Findings

F001 — Sequence places A0A2H4GSI3 among iridoid/terpenoid O-glucosyltransferases, not nicotinate N-glucosyltransferases

A computed Needleman–Wunsch global identity analysis (own provenance, ugt_tree.png) anchored the 485-aa target against a panel of characterized plant UGTs. Its nearest characterized relative is CrUGT8/UGT709C2, a reviewed 7-deoxyloganetic acid O-glucosyltransferase (EC 2.4.1.323), at 59.8% identity. Membership in the nominal UGT85A subfamily is weak — only ~40–43% identity (UGT85A23/CrUGT6 42.8%; Arabidopsis UGT85A1 40.3%; sorghum UGT85B1 37.5%) — and, revealingly, the target is only 43.0% identical even to its own genome-labeled N. attenuata paralog UGT85A2_3 (A0A2I2MND5), which is itself a bona fide UGT85A (70.2% to UGT85A23). Identity to the N-glucosylating Arabidopsis UGT72B1 is just 37.7%.

The protein carries a canonical PSPG box (…WAPQEEVLAHPSVGGFWTHCGWNS) with the UDP-glucose donor signature, confirming it is a genuine family-1 UDP-glycosyltransferase, but the box and identity landscape place it firmly in the O-glucosyltransferase neighborhood. All UniProt functional annotations are IEA: a 7-DLGT term assigned by EC and quercetin 3-/7-O-GT terms by TreeGrafter. No nicotinate N-GT term exists for this accession. This is significant because it means the seed hypothesis is not merely unverified — it proposes a reaction chemistry (N-glucosylation) and enzyme clade that the sequence evidence actively contradicts.

{{figure:ugt_tree.png|caption=Computed Needleman–Wunsch global %identity matrix, neighbor-joining tree, and PSPG-box extraction for A0A2H4GSI3 against characterized plant UGTs. The target clusters with iridoid/terpenoid O-glucosyltransferases (nearest characterized relative CrUGT8/UGT709C2 at 59.8%), is only ~40–43% identical to UGT85A members, and only 37.7% to the N-glucosylating UGT72B1.}}

F002 — Nicotinate N-glucosylation is a diversion, not a step, of nicotine biosynthesis; plant N-GTs belong to distinct families

Classical pyridine-nucleotide-cycle biochemistry establishes that in tobacco roots the free nicotinic acid supplied by the pyridine-nucleotide cycle is condensed with N-methylpyrrolinium to form nicotine (PMID: 24232153; PMID: 24241855; PMID: 28256030). Nicotine synthesis therefore consumes nicotinic acid; it does not glucosylate it. Metabolic surveys show nicotinic acid N-glucoside is instead one of three competing metabolic fates of nicotinic acid — NAD salvage, N-glucoside conjugation, or trigonelline formation (PMID: 18251891; PMID: 22983143) — i.e., a conjugation/storage/catabolic branch, not a step toward nicotine. Glucosylating nicotinate would drain the pool that nicotine biosynthesis needs, so the phrase "nicotinic acid N-glucosylation in nicotine biosynthesis" is mechanistically mis-stated.

Independently, a UniProt search for "(nicotinate OR nicotinic) AND glucosyltransferase" returns no annotated plant nicotinate N-GT. The retrievable plant N-glucosyltransferases — cytokinin N-GTs UGT76C4 (Q9FI98) and UGT76C5 (Q9FI97), and the xenobiotic N-GT UGT72B1 (Q9M156) — belong to the UGT76C and UGT72 families, which are only ~37–38% identical to the target. The enzymes that actually perform plant N-glucosylation are therefore not in the target's clade.

F003 — The true N. tabacum ortholog is a 7-DLGT-like O-glucosyltransferase, not a nicotinate N-GT "UGT1"

The UniRef90 cluster (>90% identity threshold) containing the target, UniRef90_A0A1S3YWH6, has exactly four members: N. tabacum A0A1S3YWH6 (representative) and A0A1S3Z7D3, N. sylvestris A0A1U7YM42, and the target itself — all named "7-deoxyloganetic acid glucosyltransferase." Computed NW identities confirm true orthology: target vs A0A1S3YWH6 = 88.5%, vs A0A1S3Z7D3 = 87.0%, vs N. sylvestris A0A1U7YM42 = 88.5%. This is the decisive test of the hypothesis's orthology claim: the genuine tobacco orthologs of A0A2H4GSI3 are O-glucosyltransferases, contradicting the seed claim that the characterized "tobacco UGT1" nicotinate N-GT (Nitab4.5_0006222g0020) is the ortholog of this exact sequence. That mapping could not be reproduced from any accessible resource.

Notably the orthologs are 524 aa versus the target's 485 aa — the target is ~39 aa shorter, consistent with a partial or truncated N. attenuata gene model, an additional caveat that weakens confident fine-specificity inference.

F004 — Convergent final verdict: refuted / over-annotated; true substrate unresolved

Three independent lines converge on the same conclusion. (1) Annotation provenance — UniProt annotations are IEA-only and name the protein a 7-DLGT / quercetin O-GT, with no nicotinate N-GT term. (2) Sequence identity — the target sits in the iridoid/terpenoid O-glucosyltransferase clade (nearest characterized CrUGT8/UGT709C2 59.8%; ~40–43% to UGT85A; N-glucosylating UGT72B1 only 37.7%). (3) Genuine orthology — its authentic >90%-identity N. tabacum orthologs (A0A1S3YWH6 88.5%; A0A1S3Z7D3 87.0%) are annotated 7-DLGT-like O-GTs, not a nicotinate N-GT "UGT1." Pathway biochemistry (PMID: 24232153; PMID: 18251891) shows nicotinic acid N-glucoside is a conjugation/storage branch, not a nicotine-biosynthetic step. No direct assay of A0A2H4GSI3 exists in accessible literature, and the cited locus Nitab4.5_0006222g0020 could not be confirmed as its ortholog. The target's true substrate should be reported as unresolved.

Mechanistic Model / Interpretation

The hypothesis conflates three separable things: a pathway role (nicotine biosynthesis), a reaction chemistry (N-glucosylation of nicotinic acid), and an ortholog identity (equating the target with a characterized tobacco "UGT1"). Each fails on its own terms.

   PLANT NICOTINIC ACID METABOLISM (what actually happens)
   -------------------------------------------------------
de novo (aspartate/quinolinate) + pyridine-nucleotide cycle
             |
     nicotinic acid  (free NA pool)
             |
+--------------------+---------------------+
|                    |                     |
   NAD salvage         N-glucoside            trigonelline
  (NaMN -> NAD)     (storage/conjugation)   (N-methyl-NA)
|                    |                     |
|            [distinct N-GT enzymes:       |
|             UGT76C / UGT72 clade]        |
v
   NICOTINE BIOSYNTHESIS (root):
      nicotinic acid + N-methylpyrrolinium ---> nicotine
      (NA is CONSUMED, not glucosylated)


   WHERE A0A2H4GSI3 ACTUALLY SITS (by sequence + orthology)
   -------------------------------------------------------
     Family-1 UGT with canonical PSPG box (UDP-glucose donor)
          |
iridoid / terpenoid O-glucosyltransferase clade
nearest characterized: CrUGT8/UGT709C2 (7-DLGT)   59.8%
true N. tabacum orthologs: A0A1S3YWH6 (88.5%), A0A1S3Z7D3 (87.0%)
          |
annotated (IEA only): "7-deoxyloganetic acid glucosyltransferase"
substrate for THIS protein: NOT directly assayed -> UNRESOLVED

The key interpretive point: plant nicotinic acid N-glucosylation forms an N-glucosidic bond to the pyridine nitrogen and is performed by enzymes chemically and phylogenetically distinct from the O-glucosyltransferases that add glucose to hydroxyl groups of iridoid/terpenoid or flavonoid acceptors. A0A2H4GSI3's sequence, PSPG box, cluster membership, and orthologs all point to O-glucosylation chemistry. Assigning it a nicotinate N-GT role would require it to (1) switch reaction chemistry from O to N, (2) jump from the UGT709/UGT85 O-GT clade to a UGT76C/UGT72-type N-GT clade, and (3) join a branch of metabolism that is not part of nicotine biosynthesis — none of which is evidenced.

Evidence Base / Evidence Matrix

GO Curation Implications (leads — require curator verification)

• Do NOT add any MF term equivalent to nicotinate/nicotinic-acid N-glucosyltransferase activity, nor a nicotine biosynthetic process BP term tied to nicotinate glucosylation. The hypothesis is refuted; such terms would be over-annotations with no supporting evidence, an incorrect pathway framing, and no orthology anchor.
• Existing IEA MF terms (GO:0102970 7-DLGT activity; GO:0080043 quercetin 3-O-GT; GO:0080044 quercetin 7-O-GT) are electronic inferences. Retain them as low-confidence IEA leads only; do not upgrade to experimental evidence codes, and flag that the closest characterized homolog (UGT709C2) is only ~60% identical and from another species.
• Most defensible positive MF lead: a generic family-level term such as UDP-glucosyltransferase / hexosyltransferase activity (GO:0035251 / GO:0016758-child), supported by the canonical PSPG box, with a low-confidence (IEA/ISS) evidence code. A specific substrate MF term is not supported by direct data.
• Name caveat: the genome label "UGT85A2" is questionable; sequence evidence favors a UGT709-like placement. Flag for re-classification under UGT nomenclature criteria.
• Avoid "protein binding" — the MF is genuinely a glycosyltransferase; the open question is the acceptor substrate, not whether it is an enzyme.

GO Decision Table (leads — require curator verification)

Mechanistic Scope

The immediate molecular function under test is a glycosyl-transfer reaction: transfer of glucose from UDP-glucose to an acceptor. The seed specifies the acceptor as nicotinic acid and the linkage as N-glucosidic (to the pyridine nitrogen). The evidence indicates the protein's chemistry is instead O-glucosylation of a terpenoid/iridoid-type (or flavonoid) hydroxyl acceptor, based on clade placement and ortholog annotation.

Separated from direct activity: (i) downstream pathway consequence — nicotine accumulation belongs to upstream enzymes (QPT, pyridine-nucleotide cycle; PMID: 28256030), not to a glucosyltransferase; (ii) N-glucoside as a metabolite — a real plant metabolite, but produced by a distinct N-GT clade and representing a storage/conjugation branch, not nicotine biosynthesis; (iii) loss-of-function inference — none available; no mutant/silencing data exist for this exact gene.

Conflicts and Alternatives

• Paralog confusion (primary risk). N. attenuata carries multiple divergent UGT85/UGT709-clade genes; the target and its own "UGT85A2_3" paralog share only 43% identity. A nicotinate activity reported for one paralog cannot be transferred to A0A2H4GSI3 without direct evidence.
• Family misassignment / IEA carry-over. Genome label "UGT85A2" conflicts with sequence evidence for UGT709 relatedness; the name is likely database carry-over. All functional terms are electronic; the earlier quercetin-3-O-GT hypothesis is the same class of inference.
• Historical naming clash. "UGT1/GT1" in tobacco historically denotes broad-specificity coumarin/flavonoid O-GTs (PMID: 11454003), not a nicotinate N-GT — a plausible source of the confusion behind the seed.
• Truncated gene model. The 485-aa target versus 524-aa orthologs suggests a partial model; missing C-terminal residues could affect the acceptor-binding region, undermining any fine-specificity claim (nicotinate or quercetin).
• In-vitro-only / organism specificity. The iridoid O-GT annotation derives from Catharanthus/Gardenia/Camptotheca assays; N. attenuata is not a canonical iridoid producer, so the physiological substrate may differ from both nicotinate and 7-deoxyloganetic acid.

Limitations and Knowledge Gaps

Analysis limitations: identity was computed with a simple match/mismatch/gap global (Needleman–Wunsch) alignment, coarser than profile-HMM/BLAST; orthology of the cited tobacco locus could not be verified programmatically; A0A2H4GSI3 has never been experimentally assayed. All computational results are inferences, explicitly separated from direct assay evidence above. Several PubMed hits returned by keyword search (neuronal nicotinic-receptor, dental-caries papers) were off-target and excluded.

Proposed Follow-up Experiments / Discriminating Tests

1. In vitro acceptor panel (definitive). Express recombinant, full-length re-annotated A0A2H4GSI3 with UDP-glucose and assay against nicotinic acid, 7-deoxyloganetic acid, 7-deoxyloganetin, quercetin, cytokinins, and coumarins; use LC-MS/NMR to determine whether any product is an N-glucoside vs O-glucoside. This single experiment resolves the entire hypothesis.
2. Reciprocal-best-hit orthology + bootstrapped gene tree. Formally test whether Nitab4.5_0006222g0020 is the true ortholog of A0A2H4GSI3 versus the 7-DLGT-annotated A0A1S3YWH6/A0A1S3Z7D3, including all Nicotiana UGT85/UGT709 paralogs and UGT76C/UGT72 N-GTs.
3. In planta silencing/tracer. VIGS/CRISPR of the N. attenuata locus with [^14C]nicotinamide tracer and targeted metabolomics for nicotinate N-glucoside vs nicotine — tests whether the gene affects the nicotinate pool at all.

Curation Leads (require curator verification)

• Action lead: Reject/withhold the nicotinate N-glucosyltransferase and nicotine-biosynthesis assignments for A0A2H4GSI3; treat existing 7-DLGT/quercetin-O-GT terms as low-confidence IEA leads only.
• Candidate references with exact snippets to verify:
• PMID: 24104568: "UGT8 possessed a high catalytic efficiency toward its exclusive iridoid substrate, 7-deoxyloganetic acid" — nearest characterized function is an iridoid O-GT.
• PMID: 24232153: "nicotinic acid is the relevant metabolite which is provided by the pyridine-nucleotide cycle and consumed for nicotine synthesis" — nicotine synthesis consumes, not glucosylates, nicotinate.
• PMID: 18251891: "Radioactivity from [carbonyl-14C]nicotinamide was incorporated into trigonelline… and/or into nicotinic acid 1N-glucoside (Na-Glc)" — N-glucoside is a conjugate/storage fate.
• PMID: 11454003: tobacco NtGT1a/b have "glucosylation activity against both flavonoids and coumarins" — historical tobacco "UGT1" is an O-GT, not a nicotinate N-GT.
• Candidate GO edits: keep/soften GO:0102970, GO:0080043, GO:0080044 as IEA leads; do not add N-glucosyltransferase MF or nicotine-biosynthesis BP terms.
• Suggested questions to submitter: What is the primary reference and enzyme family for the "tobacco UGT1" nicotinate N-GT? Is orthology one-to-one vs A0A2H4GSI3 or vs its paralog UGT85A2_3? Is the 485-aa model complete?

Artifacts / Provenance

• ugt_tree.png — computed NW %identity matrix, neighbor-joining tree (1 − %identity), and PSPG-box extraction of the target vs characterized/reference UGTs (executed code + output preserved in the iteration tool-call log).
• Executed code (UniProt fetches, Needleman–Wunsch identity matrix, UniRef90 cluster retrieval, PSPG-box extraction) is preserved in the iteration logs.

Investigation: 3 iterations, 4 confirmed findings, 22 papers reviewed. All computational results are inferences, explicitly distinguished from direct assay evidence.

Artifacts

• OpenScientist final report
• OpenScientist final report
• OpenScientist ugt tree