Horrible asian OGG1 mutant issues

1

compound 38 is not an OGG1 inhibitor. It is a newer OGG1 activator, specifically a serotonin-derived ORCA that accele
rates the AP-lyase step of OGG1-mediated repair. In the 2025 paper, it was optimized to very high potency (AC50 0.040
μM) and the authors explicitly say it “accelerated the AP-lyase activity of the disease-associated mutant Ser326Cys.”
They also model it in the DNA-bound context, which matches the way you asked to run Boltz. Sources: J Med Chem 2025 /
PMC12621200 (Giving an Enzyme Scissors: Serotonin Derivatives as Potent Organocatalytic Switches for DNA Repair Enzyme OGG1 - PMC) and the older OGG1 activator review summary RSC Med
Chem 2024 / PMC11487406 (Therapeutic upregulation of DNA repair pathways: strategies and small molecule activators - PMC).

What that means for your result:

  • Our Boltz panel puts compound 38 first in the S326C + 8-oxoG:C context: C:/Users/User.codex/
    alex_chen_nucleus_dna_download_vcf_NU-PGIK-3398%20(1).vcf/gene_sequences/boltz2_ogg1/results/
    boltz2_ogg1_dna_panel_metrics.tsv
  • But Boltz does not prove “restores mutant to WT.”
  • The literature-backed claim for compound 38 is narrower: it improves Ser326Cys AP-lyase activity.
  • If you want the stronger “back toward WT” precedent, that is actually compound F: in the 2018 Ser326Cys study, the
    authors say compound F restored variant 8-oxodG incision activity to wild-type extract levels in their mitochondrial
    system. Source: Free Radic Biol Med 2018 / PMC6098717 (Enhanced Mitochondrial DNA Repair of the Common Disease- Associated Variant, Ser326Cys, of hOGG1 through Small Molecule Intervention - PMC).

So the honest ranking is:

  • compound 38: best current first-pass candidate from your Boltz run, and real literature support for Ser326Cys
    activation
  • compound F: older scaffold, but the cleanest published “restored to WT-like activity” evidence in a Ser326Cys model
  • TH10785: strong activator scaffold overall, but more “gain a new function” than “specifically rescue Ser326Cys to
    WT”

On upregulating OGG1: yes, in principle that could partly compensate, but not in a simple linear way.

So my blunt take:

  • compound 38 looks like a plausible rescue-style activator.
  • No, we cannot say from Boltz that it restores your mutant to true WT function.
  • Yes, boosting OGG1 activity/expression is a plausible compensation strategy, but it should be treated as partial
    rescue, not a full correction.

If you want, I can do the next two useful steps:

  1. run matched WT + 8-oxoG:C + compound 38 / compound F controls so we can compare mutant vs WT directly;
  2. build a short ranked “OGG1 compensation” note covering compound 38, compound F, TH10785, CoQ10/ubiquinol, and
    SIRT3/AMPK/HDAC1 routes.
2

• Yes, but mostly as an indirect support strategy, not a clean substitute for compound 38.

My read is:

The upstream biology that matters most is:

So the practical phytochemical hierarchy is:

  1. Resveratrol
  2. Sulforaphane
  3. maybe Curcumin as a low-confidence general oxidative-stress support compound, not an OGG1-specific tool

Important caveat:

For your Ser326Cys context, the cleanest summary is:

  • phytochemicals can plausibly support the axis
  • they are much more likely to be partial compensation than true normalization to wild type
  • resveratrol and sulforaphane are the two I would take most seriously

If you want, I can do a Boltz follow-up next with a small exploratory panel:

  • OGG1 S326C + 8-oxoG:C + resveratrol
  • OGG1 S326C + 8-oxoG:C + sulforaphane
  • OGG1 S326C + 8-oxoG:C + curcumin
  • plus WT controls for the top 1-2 so we can compare directionally.
3

If you exclude resveratrol, the best remaining options are not equally good.

My ranking for your OGG1 Ser326Cys context is:

So the practical takeaway is:

  • best plant-derived OGG1 support candidate: sulforaphane
  • best accessible non-prescription OGG1-support candidate overall: ubiquinol
  • curcumin and EGCG are much weaker, more generic antioxidant plays

One caution that matters: more OGG1 is not automatically a pure win. Recent reviews emphasize that OGG1 has both repa
ir and signaling/inflammatory roles, so indiscriminate upregulation is not universally beneficial: PMID 39119702
(OGG1: An emerging multifunctional therapeutic target for the treatment of diseases caused by oxidative DNA damage - PubMed), PMC11487406 (Therapeutic upregulation of DNA repair pathways: strategies and small molecule activators - PMC)

If you want, I can do one focused next pass on:

  • sulforaphane
  • ubiquinol
  • NRF2
  • SIRT3
  • AMPK
4

You probably can’t get all of them as normal-person catalog items, because of course the fun compounds are always half-hidden in papers like tiny aristocrats.

Here’s the practical split:

Easy-ish to buy from research vendors

  • TH10785 is sold by vendors like Tocris/R&D Systems, Cayman, and MedChemExpress as an OGG1 activator. (R&D Systems)
  • TH5487 is sold by vendors like Selleck, Tocris/Bio-Techne, Axon, and Fisher’s MedChemExpress listing as an OGG1 inhibitor. (Selleck Chemicals)
  • 8-bromoguanine is available from Sigma-Aldrich. (MilliporeSigma)
  • 8-aminoguanine is also sold by research vendors such as Sigma-Aldrich, MedChemExpress, and AbMole. (MilliporeSigma)
  • Compound 30 is the one nice surprise: MedChemExpress appears to sell it as “OGG1 activator-1 (Compound 30)”. (MedChemExpress)

Probably not straightforward to buy off the shelf

  • Compound 38 looks like a 2025 paper compound, and I did not find a normal vendor listing for it. The paper presents it as a serotonin-derived ORCA and discusses its activity, but that is not the same thing as commercial availability. (PMC)
  • Compound F and compound D from the 2018 rescue paper look like tool compounds from a screening effort rather than standard catalog reagents. The paper describes them as internally identified activators and does not read like a commercial sourcing guide for suffering mortals. (PMC)

So if you want the realistic shopping list, I’d reorder it like this:

  1. TH10785
  2. TH5487
  3. Compound 30
  4. 8-bromoguanine
  5. 8-aminoguanine

And I would treat compound 38 / F / D as paper-only unless proven otherwise.

Two important annoyances:

1) Buy the right chemical form.
For OGG1 work, 8-bromoguanine is not the same as 8-bromoguanosine. Search results often push the nucleoside instead of the base because the internet enjoys sabotage. Your mechanistic probe was the base. (MilliporeSigma)

2) These are usually sold as research reagents, not supplements.
TH10785, TH5487, and compound 30 are vendor reagents meant for lab use. The listings are in the standard small-molecule research-supply ecosystem, which usually means institution/company purchasing, not “add to cart next to toothpaste.” (R&D Systems)

My blunt recommendation for what you can actually do next:

  • Boltz next with: TH10785, TH5487, compound 30, 8-bromoguanine, 8-aminoguanine
  • Keep compound 38 as the ideal literature benchmark
  • Drop compound F and D for now unless you can get a structure or a custom synthesis route

If you want the closest thing to a sane plan, it is:

  • use compound 30 as your purchasable “modern activator” surrogate,
  • TH10785 as the classic activator comparator,
  • TH5487 as inhibitor control,
  • 8-bromoguanine / 8-aminoguanine as mechanistic probes.

That gets you most of the information without needing a medicinal chemistry lab and a trust fund.

5

What changed:

  • The exact GTEx rows for the queried canonical NEIL2 UTR sites are not brain-side. The
    exact hits I found are all Heart_Left_Ventricle, and they are positive for 11770112,
    11771150, 11786406, and 11787242 with NES about +0.20 to +0.22.
  • The full-locus phased Enformer run confirms NEIL2 is still a strong regulatory locus
    in your data, but hap1 and hap2 collapse to the same effective sequence. So this is
    not a phase-complex story here. Region-level perturbation stayed 3’ UTR > 5’ UTR >
    gene body, with mean abs deltas 0.122356, 0.110448, and 0.075784.

So the clean read now is: NEIL2 is still worth keeping as a possible OGG1-backup-side
modifier, but I do not have evidence yet that your exact canonical NEIL2 UTR cluster is
giving a brain-specific compensatory expression boost. The exact observed human signal
at those sites is cardiac-positive, while the sequence-model signal remains broader and
strong.

If you want the next useful step, I’d do one of these:

  • look for brain-specific NEIL2 QTL/chromatin evidence beyond exact GTEx single-tissue
    rows
  • extract NEIL2 track-level Enformer deltas the same way we did for FOXO3 and the
    focused NEIL2 constructs