Cooking (even boiling) food damages its DNA which gets absorbed into the body's cells, which damages their DNA

#1

From FB Messenger:

An additional source of aging from cooking food has been discovered.

It has long already been known that adding heat to foods causes sugars to randomly attach to proteins in an uncontrolled version of glycosylation called glycation via heat increasing random reaction rates. Because our bodies posttranslationally modify proteins with different chemical groups including sugars to accomplish their functions, sugar mods can prevent proper breakdown of amino acids and lead to problematic incorporation of glycated endproducts which our kidneys help filter with RAGES (receptor for advanced glycation endproducts).

A new paper demonstrates that a similar process appears to occur with heat-damaged DNA in food, whose alternative bases also become incorporated into bodies after consumption. This was a suspected potential damage pathway before, but for me this study confirms it. Animals ingesting cooked food compared to uncooked food had elevated levels of double-stranded break regions and genetic rearrangements, and incorporation of alternative bases.

Cooking has long been a tradeoff between killing microbes with improving flavor versus nutrition and carcinogens and production of toxicity.

This information won’t change what I’m doing by too much I’m already on uncooked meal replacements like soylent and pills, but might be interesting to others.
h/t Denis Odinokov
https://pubs.acs.org/doi/10.1021/acscentsci.2c01247


1 Like
#2
2 Likes
#3

Is this the old glycation issue and burned food issue?

1 Like
#4

particular, within many phylogenetic groups, organisms with
larger cells and lower metabolic rates generally have larger genomes (e.g.,
Gregory 2001, 2002a,b; Vinogradov and Anatskaya 2006)

#5

Steam or boil food to cook it is best.

3 Likes
#6

And add some Carnosine Supplement to your daily routine to fend off the glycation issues.

#7

Why do you mostly consume soylent?

#8

^that was a quote in a Messenger thread, not my quote

Short version: Yes. Plants and tofu contain DNA, high heat damages it, and the same “salvage → incorporation” pathway shown in the 2023 Stanford/NIST paper can, in principle, operate. But vegetables/tofu generally contain far less DNA per gram and are cooked more gently, so the dose of damaged nucleosides is far smaller than from a heavily seared meat serving.

How much DNA is in these foods?

  • Red meat (muscle): ~1.4 mg DNA per gram of meat (wet weight). Seal meat shows ~0.6–1.1 mg/g; primate muscle ~0.96 mg/g—same order of magnitude. (Wiley Online Library, ScienceDirect)
  • Plants (per dry matter) – examples from a compiled table: broccoli 5.1 g/kg DM, cauliflower 2.8, spinach (frozen) 2.6, potato 1.0, onion 0.7. Converting to wet weight using typical moisture contents puts vegetables at ~0.1–0.6 mg DNA per gram, i.e., several‑fold less than meat. (PMC)
  • Tofu: ~84–85% water. Legumes in that table (wheat 0.6, lentil 0.8 g/kg DM) suggest tofu will sit around ~0.09–0.12 mg DNA per gram wet before processing; tofu manufacture substantially degrades DNA, which is why GMO traceability drops along the process. (PMC, PMC, PMC)

How much damaged DNA after cooking?

The Stanford/NIST team measured ~300 deoxyuridine (dU) lesions per million bases in beef after a mild roast (15 min, 220 °C)—about a 0.03% mole fraction. They argue this yields milligram‑scale dU in a large cooked meat serving because meat contains so much DNA. (PMC)

Using their 0.03% figure just for ballpark comparison:

  • 200 g seared steak
    DNA ≈ 1.4 mg/g × 200 g = 280 mg DNA → dU ≈ 0.03% ≈ 0.084 mg (84 ”g). (Wiley Online Library, PMC)

  • 150 g fajita veggies (onion+pepper, sautĂ©ed 7–10 min, not charred)
    Take onion as a measured example: DNA ≈ 0.105 mg/g (0.7 g/kg DM × ~15% DM). For 150 g mixed vegetables, order‑of‑magnitude DNA ~15 mg, so dU ≈ 0.03% ≈ ~4–5 ”g—~10–20× lower than the steak example. (Likely even lower because fajitas are cooked shorter and cooler than 220 °C roasting.) (PMC, Chemical Engineering Transactions, ResearchGate, Culinary Hill)

  • 150 g tofu entrĂ©e
    DNA roughly ≀10–15 mg pre‑processing; actual salvageable DNA is lower due to fragmentation and whey removal during tofu making → dU plausibly a few micrograms or less. (PMC, PMC)

Mechanism & risk framing

  • The 2023 study demonstrated that damaged nucleosides (notably dU, 5‑OH‑dU, 5‑OH‑dC) can be salvaged and incorporated into intestinal DNA in mice, elevating γ‑H2AX foci; the effect scaled with dose and was strongest for pyrimidines. (PMC)
  • Vegetables/tofu do form damaged bases under high heat, but because they usually contain less DNA per gram and are often cooked at lower temperatures/shorter times, the exogenous lesion dose is much smaller than from well‑done meat. This is in addition to separate genotoxins from high‑heat cooking (PAHs, HCAs in meats; acrylamide in carb‑rich plant foods), which are managed by the same advice: avoid charring and excessive frying. (PMC, PMC, ScienceDirect)

Practical tips (apply to veggies, tofu, and meat)

  • Keep temps moderate; avoid char. SautĂ©/roast lightly; don’t blacken. Chipotle‑style fajita veggies are typically sautĂ©ed ~7–10 min to “tender‑crisp,” not heavily charred. (Culinary Hill, Reddit, FoodPrint)
  • Short cook times, flip often. Limits surface overheating and oxidation. (PMC)
  • Use acidic, polyphenol‑rich marinades (lemon, herbs) to suppress oxidation products. (ACS Publications)
  • Prioritize moist‑heat methods (steam, poach, sous‑vide) when convenient. (PMC)

If you’d like, I can put together a small table estimating dU micrograms per typical serving for: a 200 g steak (grill), 150 g salmon (baked 180 °C), 150 g fajita veg (sautĂ©), and 150 g tofu (stir‑fry), using the Stanford/NIST lesion rate and moisture assumptions above.

3 Likes
#9

Pressure cooker uses lower temperature and higher pressure. What are your thoughts on stewing in a pressure cooker?

#10

That would be interesting. However cooking at higher pressure means that also the temperatur is higher. In a good pressure cooker you can go up to 120°C. It is good for sterilizing things.

#11

I was mistaken I thought pressure allowed boiling at lower temperatures.

1 Like
#12

This is doubly relevant if you have compromised OGG1


Yes, roasted coffee beans contain damaged DNA bases

Cooking results in greatly elevated levels of oxidative and deaminated DNA damage in nearly all foods, as indicated by 8-oxo-dG and dU nucleotides, with up to 250-fold increases versus raw food. ResearchGate Coffee roasting at 200°C+ is about as extreme a cooking process as exists, so the beans’ endogenous DNA gets wrecked — oxidized guanosine (8-oxo-dG), deaminated cytosine, depurinated bases, probably strand breaks everywhere. You then extract these into hot water and drink it.

The salvage pathway concern — the actually scary mechanism

The reason this matters beyond just “damaged stuff gets degraded in the gut” is the nucleotide salvage pathway. The hypothesis is that high-temperature cooking may cause significant damage to the DNA in food, and this damage might find its way into cellular DNA by metabolic salvage; exposing cultured cells to damaged 2’-deoxynucleosides resulted in elevated DNA damage and repair responses, and feeding a deaminated 2’-deoxyucleoside to mice resulted in substantial uptake into intestinal genomic DNA and promoted double-strand chromosomal breaks. ACS Publications

So the mechanism isn’t passive absorption — it’s that salvage kinases are promiscuous and will phosphorylate oxidized nucleosides, loading 8-oxo-dGTP into the dNTP pool, where it can get incorporated into replicating DNA and cause G→T transversions. The intestinal epithelium specifically is at risk because it’s absorbing this stuff before nucleases fully degrade it.

But then the paradox — coffee actually reduces DNA damage biomarkers

Here’s the Janusian flip: regular consumption of a dark roast coffee blend has a beneficial protective effect on human DNA integrity in both men and women, in a 4-week randomized controlled study. PubMed

And separately, coffee intake is associated with lower levels of oxidative DNA damage and decreasing body iron storage in healthy women, as measured by urinary 8-hydroxydeoxyguanosine. PubMed

So coffee delivers damaged DNA bases and reduces systemic DNA damage markers. The antioxidant load (chlorogenic acids, trigonelline, whatever else) apparently overwhelms the nucleoside salvage concern — or the gut nucleases degrade most dietary DNA fragments before meaningful salvage can occur.

The methodological wrinkle — isotope studies

There’s actually a direct test of whether dietary 8-oxo-dG gets absorbed. Different amounts of 15N-labeled oxidatively modified DNA were consumed orally by volunteers, and over 2 weeks no 15N-labeled 8-oxoGua or 8-oxodG were detected in either urine or DNA from circulating blood cells. PubMed Central

This pushes against the salvage concern being clinically relevant — at least for systemic exposure. The intestinal mucosa question remains somewhat open, since you could have localized salvage in gut epithelium without it showing up in blood or urine.

Multi-hypothesis framing on the net risk:

H1: Net neutral, antioxidants dominate (very likely, ~0.65) — the RCT and cohort data on urinary 8-oxo-dG consistently go in coffee’s favor. The damaged nucleoside load is real but overwhelmed.

H2: Localized intestinal mucosal risk that doesn’t appear systemically (plausible, ~0.25) — the isotope study clears blood and urine but didn’t look at intestinal epithelial DNA directly. Colorectal cancer risk from coffee is actually neutral or slightly negative (protective), which argues against this.

H3: The salvage concern is real but the biome does the cleanup (can’t rule out, ~0.10) — gut bacteria express nucleases, and the massive microbial load in the colon might degrade damaged nucleosides before they reach the epithelium in significant quantity. Untested.

What would shift this: a study directly measuring 8-oxo-dG incorporation in intestinal mucosal biopsies of heavy coffee drinkers vs. non-drinkers. That apparently doesn’t exist yet.

Your intuition that the net health picture remains positive is almost certainly correct — but the reason coffee is net-DNA-protective despite containing damaged DNA bases is genuinely unresolved at the mechanistic level. The polyphenol antioxidant story is the leading hypothesis but it’s not watertight.

===

this sharpens the question significantly and the answer is actually mechanistically interesting.

The Ser326Cys deficit — what it actually does

Even without oxidizing conditions, the S326C variant shows a 40.8% decrease in repair activity across all enzyme concentrations compared to wild type. That’s the baseline. But the really nasty part is what happens under oxidative stress: both variants increase OGG1 activity in response to oxidative stress, but peak activity of Ser326-OGG1 occurs 12 hours prior to that of Cys326-hOGG1 — with further evidence for impaired Cys326-hOGG1 repair ability specifically under conditions of oxidative stress.

The mechanism for why stress makes it worse: under conditions of cellular oxidative stress, Cys326-OGG1 protein complexes accumulate — the introduced cysteine enables disulfide-bonded dimerization under oxidizing conditions, and this complex formation was inhibited by thiol reducing agents and antioxidants, indicating oxidative stress is causally driving the inactivation.

So it’s not just a static 40% reduction — it’s a repair enzyme that specifically fails under the exact conditions (high ROS) when you’re generating the most 8-oxo-dG. The protein clumps up when you need it most.

Mitochondria are particularly exposed

Mitochondrial extracts from cells expressing hOGG1S326C were deficient in mitochondrial 8-oxodG incision activity, and cells expressing the variant showed increased cellular and mitochondrial reactive oxygen species compared to wild-type counterparts. Mitochondrial DNA has no histones, sits near the ROS production site, and relies heavily on OGG1 for 8-oxo-dG repair. The Ser326Cys variant hits mtDNA repair harder than nuclear DNA repair proportionally.

What this means for dietary 8-oxo-dG specifically

The salvage pathway concern from the previous discussion becomes more consequential here. The body has three layers of defense against 8-oxo-dG:

  1. MTH1 — sanitizes the dNTP pool by hydrolyzing 8-oxo-dGTP before it gets incorporated
  2. OGG1 — removes 8-oxo-dG from DNA after it’s been incorporated (or from endogenous oxidation)
  3. MUTYH — removes adenine mispaired with 8-oxo-dG, preventing the G→T transversion

Compromised OGG1 means layer 2 is leaky. If dietary 8-oxo-dG gets through MTH1 and incorporates, it persists longer before excision. The G→T transversion rate goes up, and the MUTYH backup is working against a higher substrate load.

Ser326Cys polymorphism of OGG1 increases the risk of colorectal cancer and is associated with significantly increased levels of 8-oxoguanine in colorectal cancer patients. The intestinal mucosa is exactly where dietary 8-oxo-dG from coffee would hit first, which is a uncomfortable specificity.

The Janusian problem for OGG1-Cys326 carriers and coffee

Here’s where it gets genuinely contradictory. Coffee’s antioxidant load (chlorogenic acids, polyphenols) reduces endogenous ROS → reduces endogenous 8-oxo-dG production → takes pressure off the already-compromised OGG1. That’s a net positive, arguably more valuable for someone with slow repair than for wild-type, because the marginal benefit of reducing substrate load is higher when clearance is impaired.

But simultaneously, the dietary 8-oxo-dG from the coffee itself is harder to clear. And critically — the oxidative environment in intestinal cells after drinking coffee (melanoidins, acrylamide metabolites, local redox shifts) could be exactly the conditions under which Cys326-OGG1 dimerizes and becomes further inactivated, right when it’s needed to clear any incorporated dietary 8-oxo-dG.

This interaction hasn’t been studied directly for coffee. Speculatively: the net effect probably still favors coffee for Ser326Cys carriers, because the antioxidant benefit dominates in epidemiology even in populations with high Cys326 allele frequency (East Asian cohorts where this variant is common haven’t shown elevated coffee-cancer associations). But the margin of safety is plausibly smaller, and the mitochondrial DNA angle is almost completely unstudied.

What would actually matter here

If someone carries Cys326/Cys326 homozygous and has elevated endogenous ROS load on top of that — say, from mitochondrial dysfunction — the compounding of impaired repair + high substrate generation means the dietary 8-oxo-dG question stops being academically interesting and becomes more practically relevant. The ceiling for how much additional oxidized base load the system can tolerate gets lower. Whether that translates to “drink less roasted coffee” is a step the data doesn’t actually support yet, but it’s a more defensible individual concern than for wild-type.

#13

The oil matters a lot and gets underappreciated

The dominant oils in East Asian cooking historically and currently are vegetable oils with high polyunsaturated fatty acid (PUFA) content — soybean oil, corn oil, rapeseed/canola, sunflower, and historically peanut oil. These are fundamentally different from lard (mostly saturated), olive oil (mostly monounsaturated oleic acid), or coconut oil (saturated) when you heat them hard.

PUFAs oxidize at high heat and produce:

  • Aldehydes — 4-hydroxynonenal (4-HNE), malondialdehyde (MDA), acrolein. These are directly genotoxic, form DNA adducts, and 4-HNE specifically forms exocyclic DNA adducts that OGG1 is NOT the repair enzyme for — that’s more NER territory
  • Lipid peroxidation products that generate secondary ROS, which then produce 8-oxo-dG endogenously
  • Trans fatty acids from repeated high-heat use of the same oil

Olive oil at high heat produces far fewer of these because oleic acid (C18:1) has one double bond; linoleic acid (C18:2) in soybean oil has two; linolenic acid (C18:3) has three. Each additional double bond dramatically increases oxidative reactivity. Lard is mostly saturated — extremely stable at heat, basically no PUFA oxidation products.

So the comparison isn’t just “frying vs not frying” — it’s specifically high-PUFA-oil frying at wok temperatures (often 250-300°C+) which is a substantially worse genotoxic environment than frying in lard or olive oil.

The wok hei problem

Wok hei requires extremely high heat and rapid cooking. The temperatures involved routinely exceed the smoke point of the oils being used, which is where lipid oxidation goes nonlinear. The indoor air in a wok kitchen contains measurable 4-HNE, acrolein, benzene, PAHs. The cook inhales this; the food absorbs some of it.

This is actually where the never-smoking Chinese women lung cancer story partially lands — repeated high-heat wok frying exposure, indoor without adequate ventilation, sustained over decades.

So does this change the OGG1 Ser326Cys picture?

Somewhat, but in a different direction than the 8-oxo-dG story. The genotoxic products of PUFA oxidation at high heat — 4-HNE adducts, propano-dG adducts from acrolein, MDA-dG adducts — are repaired by nucleotide excision repair (NER) and sometimes mismatch repair, not primarily BER/OGG1. So OGG1 Ser326Cys specifically isn’t the rate-limiting vulnerability for lipid oxidation product genotoxicity.

But — and this is important — lipid peroxidation products generate secondary ROS, which generates 8-oxo-dG, which NOW hits the compromised OGG1. So the pathway is:

high-heat PUFA oil → lipid radicals → secondary ROS → endogenous 8-oxo-dG → slower repair in Cys326 carriers

It’s indirect but real. The OGG1 vulnerability is activated by the oxidative environment created by the frying, not directly by the DNA damage products of the lipids themselves.

Why you still might not see a clean signal

The effect is probably real but embedded in a complex exposure matrix where:

  • Garlic, ginger, green onion — standard aromatics in most Asian frying — have meaningful antioxidant capacity that quenches some lipid radicals
  • High vegetable intake provides polyphenols
  • Green tea consumption is substantial in Japan/China/Korea
  • Portion sizes and total caloric intake historically lower

These are genuine confounders, not just noise. The antioxidant co-exposures from the same culinary tradition that produces the frying exposure may be substantially buffering the genotoxic load. Which is frustrating from an attribution standpoint but probably real.

The honest bottom line

High-heat PUFA oil frying is probably more genotoxic than lard or olive oil frying, the mechanism is lipid oxidation products + secondary ROS rather than direct DNA base damage, and Cys326 OGG1 carriers are plausibly more vulnerable to the secondary ROS arm of this. But the population-level signal is masked by dietary antioxidant co-exposures, cooking time brevity (stir-fry is fast), and the difficulty of detecting gene-environment interactions with this specificity in observational data.

#14

Plants do have large genomes but that doesn’t straightforwardly mean more DNA per gram of food

The genome size (C-value) and the DNA content per gram of tissue are related but not the same thing, because:

  • Ploidy — many crop plants are polyploid. Wheat is hexaploid (6 copies). Strawberry is octoploid (8 copies). More genome copies per cell, yes, but this is already baked into the C-value measurements
  • Cell size — bigger genomes often come with bigger cells, so cells per gram goes down. The DNA concentration per gram of tissue doesn’t scale linearly with genome size
  • Water content — vegetables are mostly water. A gram of broccoli is like 90% water, so DNA per dry weight is higher than DNA per fresh weight
  • Cell type composition — actively dividing meristematic tissue has more DNA synthesis happening; storage tissue (potato tuber, carrot root) has large vacuolated cells with relatively low nucleus-to-cytoplasm ratio

The actual numbers are surprisingly low

Human dietary DNA intake estimates are roughly 0.1-1g of DNA per day from food. Plants contribute substantially to this. But the ACS Central Science paper on heat-damaged food DNA found large variation in extractable DNA content across 21 food ingredients, implying widely variable levels of consumption. ResearchGate “Extractable” is doing a lot of work there — cooking, cell wall disruption, and digestion all affect how much DNA actually reaches your intestinal cells in intact enough form to matter for salvage.

The strawberry extraction trick is revealing about why

The reason strawberries work so well for classroom DNA extraction isn’t that they have unusually high DNA content per se — it’s the octoploid genome giving lots of copies, combined with soft cell walls that break easily, combined with high water content that helps with lysis. You’re extracting efficiently, not extracting a lot. The white stringy precipitate looks like a lot but is mostly water and you’re concentrating from a large volume.

Conifers have genuinely enormous genomes — Norway spruce is around 20 Gb, loblolly pine ~22 Gb, compared to human ~3 Gb. But you’re not eating pine needles in quantity. The relevant food plants are:

  • Wheat: ~17 Gb (hexaploid)
  • Maize: ~2.4 Gb
  • Rice: ~0.4 Gb (actually small)
  • Soybean: ~1.1 Gb (but tetraploid)
  • Tomato: ~0.9 Gb
  • Potato: ~0.84 Gb (tetraploid, so ~3.4 Gb effective)

So wheat is actually one of the higher-DNA-content foods by genome size, which is interesting given how much bread crust (heavily Maillard-reacted) people eat.

What cooking does to plant DNA specifically

Plant cells have cell walls that partially protect DNA from immediate degradation during cooking, but also trap it inside the tissue. High heat denatures it, causes strand breaks, oxidizes bases. The question for the dietary 8-oxo-dG story is whether cooking liberates damaged nucleosides into a form that’s bioavailable for salvage, or whether it just produces a mass of denatured polymer that gut nucleases chew through without meaningful nucleoside absorption.

The honest answer: plant DNA in cooked food is probably mostly degraded to free bases and nucleosides by intestinal nucleases before meaningful salvage can occur. The concern from the ACS paper was more acute for foods where DNA is dense and the cooking is rapid — like a quickly seared piece of meat where the interior might still have partially intact but oxidized DNA fragments reaching the small intestine before nucleases fully process them.

Heavily fried vegetables — which are cooked thoroughly, have high surface-area-to-volume ratio from slicing, and spend time at temperature — probably produce more complete DNA degradation than, say, rare meat. Which counterintuitively might make them less of a dietary 8-oxo-dG concern via the salvage pathway, even though they have more Maillard products overall.

The more relevant genotoxic exposure from fried vegetables is probably the lipid oxidation products and acrylamide rather than the plant DNA damage products specifically. The DNA story is more acute for animal-derived foods where the DNA is denser, the cooking is sometimes less complete, and the salvage pathway concern is more plausible.

#15

Time to build a low pressure cooker!