#1

@adssx kindly referred a paper to me that seems to explain the reasons why some old experiments with DNA had quite significant lifespan benefits for rodents.

1.4_Part_1.445.pdf (731.0 KB)
j.1532-5415.1973.tb01207.x.pdf (202.6 KB)
Robertson 1928 nucleic acids upon the growth and longevity of the white mouse.pdf (5.1 MB)

3 Likes
#2

Thanks for writing this.

The paper (U Gothenburg + Max Plank Institute): Ribonucleotide incorporation into mitochondrial DNA drives inflammation 2025

To test whether the increased rNMP content of mtDNA affects mtDNA release and SASP, we treated senescent cells with deoxyribonucleosides. Deoxyribonucleoside supplementation did not affect the senescent state per se as assessed by monitoring SA-β-gal activity (Extended Data Fig. 9h). However, it significantly reduced cytosolic mtDNA (Fig. 5c and Extended Data Fig. 9i) and the mtDNA sensitivity for alkali and RNase H2 treatment, indicating a lower rNMP content (Fig. 5c,d).

News report: Mitochondria Dump Their Rubbish DNA, And It Could Be Costing Us Our Health 2025

1 Like
#3

I do reference the paper in my blog post. I think the sciencealert summary is wrong. They say:

That rejection is potentially one of the key drivers behind the inflammation that comes with old age, and the negative health consequences associated with it – from certain types of cancer to neurodegenerative diseases like Alzheimer’s.

“This response provides protection against pathogens but can also promote autoimmune and inflammatory diseases and contribute to senescence and aging,” write the researchers in their published paper.

It’s yet to be determined how much this type of inflammation occurs during the normal aging process, or whether it takes place under specific conditions.

Humans are living longer than ever before, and that means all of the biological machinery in our bodies is working overtime. As the years go by, stress, damage, and inflammation stack up, leading to poor health.

If we can understand more about how to cut out some of that damage – how to stop mtDNA making these particular mistakes in replication, for example – then we’re possibly on the way to developing ways of keeping our cells in better shape in old age.

This implies they see the rejection of the flawed DNA as being the issue rather than the creation of the flawed DNA.

IMO it is very obvious that this is a mechanism that enables improved replication of mtDNA. Exactly how to do this is not yet that clear. However, I will spend some time reading up on this as it is a new dimension of interventions.

3 Likes
#4

IIRC, when I looked into the nucleotide studies on rodents I think pretty much all of them found benefits on the background of a diet deficient in nucleotides. That doesn’t mean taking nucleotide supplements will be beneficial in a person eating a normal healthy diet becauase such diets will already contain modest amount of nucleotides from the protein sources in the diet.

1 Like
#5

There are lots of complex questions here about the balance between different nucleotides. However, underlying this issue is the fact that there is no enzyme that repairs mtDNA when RNA is mistakeny included in the DNA. Only having more nNTPs around can either prevent this or ensure that when the mtDNA is replicated the replicated copy does not have this error.

#6

I don’t see how the above results have anything to do with the balance between different nucleotides. It’s not like the rodent studies gave only one or two types of nucleotides. What the studies are showing is that oral nucleotides can be beneficial if dietary intake is low.

I think you’re projecting through your “mtDNA importance in aging theory” lens here. I don’t think that mtDNA repair has much relevance to the rodent studies on the benefits of oral nucleotides. Nucleotides are used for a lot of things besides mtDNA repair and AFAIK none of the studies highlighted influence on mtDNA repair as an important part of the effects of oral nucleotides.

#7

I haven’t really responded in sufficient detail. The issue is quite complex and I probably need to structure all the information I have in another blog post. For example there is a question as to the balance between purines and pyrimidines and the levels of urate.

Also when it comes to replication the balance between nucleotides affects fidelity.

The papers above do give an indication that there is an aging pathway geroscience hypothesis available from some supplementation in this area.

This paper
https://www.nature.com/articles/s41586-025-09541-7

and various other papers show that normally about 0.25% of mtDNA is RNA and that additional dNTPs and possibly particularly pyrimidines can reduce the error rate. Which is known to be a good thing.

Nucleotides are used for a lot of things besides mtDNA repair

I don’t think they are actually used for this sort of repair. This sort of repair is not done in mitochondria and this AIUI in the nucleus does not actually need a nucleotide as it simply edits the one in situ.

1 Like
#8

sorry guys but you literally confuse things ! Consume dietary nucleotides will NOT help to balance rNTP / dNTP ratio ! Yeast nucleotides (5′-AMP, GMP, CMP, UMP) are Ribonucleotides… So you will actually do NOTHING or worst on this ratio…

A total confusion here :smile:

#9

The reason I disagree with this is that the balance of rNTPs is primarily driven by energy levels hence the ratio in the cytosol of rNTPs to dNTPs is something in the range of like 100 → 1000 to 1.

Digested ribonucleotides can be converted to deoxyribonucleotides in the cytosol and then migrated to the mitochondria. There is a key question as to the balance of different bases and dietary ribonucelotides with a skew towards the pyrimidines has the ability to feed this step. There are people who argue that this is a good approach.

However, your basic point is a good point and hence increasing dietary consumption of DNA is probably strategically the best approach. That is hard because people are not allowed to sell it as a food supplement. There are pyrimidine dNTP supplements provided as medication.

#10

John, The key variable for mtDNA fidelity is the local rNTP/dNTP ratio in the mitochondrial matrix, not the cytosolic rNTP abundance… Dietary nucleotides are dephosphorylated to nucleosides… before absorption > converting them into mitochondrial dNTPs requires RNR (NDP→dNDP), which is tightly regulated and limited in non-dividing tissus. Mitochondria rely mainly on salvage (TK2) and pyrimidine nucleotide carriers (SLC25A33/A36). Adding ribonucleotides tends to actually raise rNTP**, potentially worsening the mtDNA error pressure. Adding deoxy-nucleosides (dT, dC etc) directly raises mitochondrial dTTP/dCTP, which is the axis supported mechanistically and clinically (check “MT1621”). Sorry ^^

#11

My reading on RNR is that it is expressed even without p53 and post division.

Are there papers linking dietary nucleotides and increased mitochondrial rNTP?

When I am on the right device I will upload some of the material I have been sent.

When I read the dosing of MT1621 on the other device it was about 25g per day of each of the two deoxy-nucleosides.

#12

From a human trial in older adults:

“The results of primary outcomes showed that [dietary nucleotides] had significantly greater reduction in median DNA methylation age compared to placebo over 19 weeks.”

“Insulin sensitivity improved in the [nucleotides] group, with a significant reduction in HOMA-IR over 19 weeks.”

“our findings establish that [nucleotides] may delay biological aging and improve insulin sensitivity with a well-tolerated safety profile.”

1 Like
#13

ART-ANH-VER-IPM-240608-Verkerk-Small Molecules with Big Effects- Peptides and Nucleotides for Immune Resilience and Longevity.pdf (8.9 MB)

ART-FSN-2016-20200806-Article- Magazine-Energy, Nucleotides-Koeppel.pdf (3.0 MB)

DOC-STA-PKN-20220308-Statement-Gout-Nucleotides-Dr Peter Koeppel-March 2022.pdf (124.9 KB)

ChatGPT: (rNTP then dNTP)

The intracellular concentrations (and thus ratios) of the four ribonucleoside triphosphates (ATP, GTP, UTP, CTP) in the cytosol vary somewhat by cell type, but several studies give good estimates. Here’s a summary of what is known, and what approximate ratios you might infer.

:mag: What the studies say

  • One review on mammalian nucleotide metabolism notes: “In cells, it is typically observed that the nucleotide concentrations are in the order ATP > UTP > GTP > CTP.” (PMC)
  • In a short communication, the concentrations in human myeloid-leukaemia cells were estimated roughly as: ATP ≈ 3.5 mM, CTP ≈ 0.4 mM, GTP ≈ 0.9 mM, UTP ≈ 0.9 mM. (MDPI)
  • Another source gives a “typical” pattern: ATP > UTP > GTP > CTP; and notes CTP is around ~1 µmol/g in some cells. (PMC)
  • A methodological article used a standard: 1.0 mM ATP, 1.0 mM GTP, 0.67 mM CTP, 0.67 mM UTP as “1×” NTP mix. (PMC)

:bar_chart: Approximate ratio

Putting together the above numbers, a plausible approximate ratio (ATP : GTP : UTP : CTP) in some human-cell cytosol might be:

~ 3.5 mM : 0.9 mM : 0.9 mM : 0.4 mM
Which simplifies roughly to:
ATP : GTP : UTP : CTP ≈ 9 : 2.5 : 2.5 : 1

If you convert to simpler whole-numbers:
≈ 9 : 3 : 3 : 1

Or, if rounding further:
≈ 9 : 3 : 3 : 1 ≈ ≈ 9 : 3 : 3 : 1

:brain: Caveats & context

  • These numbers are estimates and differ by cell type, metabolic state, organism, and measurement method.
  • “Cytosol” here means the bulk soluble part of the cell; compartmentalization (mitochondria, nucleus) may have different values.
  • The ratios can shift under stress, proliferation, or disease (e.g., tumour cells might have elevated other NTPs). (MDPI)
  • The measured values are sometimes total cellular rather than strictly “cytosolic free” concentrations.
  • The functional relevance: higher ATP level makes sense because ATP powers many processes; the lower levels of CTP reflect its more limited direct usage outside nucleotide synthesis.

If you like, I can check for recent high-precision quantitative studies that give exact concentrations in specific human cell types (e.g., fibroblasts, hepatocytes) and provide a table of those values. Would that be helpful?

Here are recent, quantitative measurements of cytosolic NTP levels from human cells, using LC-MS(/MS) or validated HPLC methods. I’ve included absolute amounts and an at-a-glance ratio for each dataset.

Human cell measurements (absolute → quick ratio)

Cell / sample (method) ATP GTP UTP CTP Implied ratio (ATP:GTP:UTP:CTP)
Molm-13 (human AML) cells — molecules per cell (LC-MS/MS with isotopic standards) 2.29×10⁹ 5.05×10⁸ 5.20×10⁸ 2.07×10⁸ ~11 : 2.2 : 2.3 : 0.9. (BioMed Central)
Human leukemia bone-marrow cells (patient sample) — pmol per 10⁶ cells (LC-MS/MS) 218.3 pmol 218.3 296.0 105.7 ~2.1 : 2.1 : 2.9 : 1.0 (per 10⁶ cells). (PMC)
Mammalian cells (meta-analysis across many types) — mean conc. (µM) 3152 µM 468 567 278 ~11 : 1.5 : 1.8 : 0.9 (typical ordering ATP > UTP ≳ GTP > CTP). (考研)

Notes on the table

  • The Molm-13 dataset provides per-cell absolute counts using isotope-labeled standards (high confidence for ratios). (BioMed Central)
  • The patient bone-marrow sample shows per-million-cells amounts; ratios are comparable but not identical (biological context and extraction differences matter). (PMC)
  • The Traut review aggregates ~600 measurements across mammalian systems (gives robust typical cytosolic concentrations). (考研)

Practical takeaways

  • Across human cells, a stable pattern emerges: ATP is highest, UTP and GTP are mid-range, CTP is lowest. (考研)
  • Absolute values vary by cell type, growth state, and method, but ratios are surprisingly consistent (often near ~10–12 : 2–3 : 2–3 : 1). (BioMed Central)
  • Modern LC-MS(/MS) with stable-isotope internal standards gives the most reliable absolute numbers (see methods papers). (BioMed Central)

If you’d like, I can pull a longer panel (e.g., HeLa, HEK293T, primary fibroblasts, T cells) and normalize everything to either molecules per cell or mM (estimating cytosolic volume) so you can compare like-for-like.

Here’s what the literature says about the ratios of deoxyribonucleoside triphosphates (dNTPs) in human (or mammalian) cytosol, along with caveats. The data are more limited and variable compared to ribonucleoside triphosphates (rNTPs).

:books: What the studies report

  • A review by Gandhi & Agarwal (2011) summarised multiple measurements: they emphasise that the four canonical dNTPs (dATP, dGTP, dCTP, dTTP) are present in very low absolute concentrations (often 10-100 pmol per million cells during S-phase) and that even modest imbalances can lead to mutagenesis or replication stress. (PMC)
  • Song et al. (2005) measured mammalian tissue dNTP pools and reported that dGTP usually comprises only ~5–10% of the total dNTP pool in cultured cells. (PNAS)
  • Ferraro et al. (2010) carried out careful quantitation of cellular dNTPs in eukaryotic cells and emphasised the need for high sensitivity methods. They do provide absolute values but less direct “ratio tables” for each nucleotide. (OUP Academic)
  • The more recent review (Yagüe-Capilla et al., 2024) highlights that dNTP pool perturbations (including ratios) vary across cell cycle, cell type, disease states—but clearer “standard” ratios are rarely given. (PMC)

:bar_chart: Approximate ratio derived from data

From these data points, one can infer a rough (very approximate) ratio for human cytosolic dNTPs (ordered as dATP : dTTP : dCTP : dGTP). Note this is a generalised estimate and will vary significantly by cell type, phase of cell cycle, etc.

Estimated ratio ≈ 10 : 10 : 5 : 1

  • i.e., dATP and dTTP are the highest, dCTP somewhat lower, and dGTP is the smallest fraction (consistent with “5-10% of pool” for dGTP).
  • For example, Song et al. reported dGTP ~5-10% of total, implying the other three sum to ~90%. If dATP and dTTP are comparable and larger, you might get a ratio such as 10:10:5:1 for ease of reference.

:warning: Important caveats

  • The absolute concentrations of dNTPs are very low compared to rNTPs and vary with cell cycle (especially high in S-phase) and cell type (proliferating vs resting). (PMC)
  • Many measurements do not separate cytosolic vs mitochondrial pools cleanly; some data lumps whole‐cell or includes mitochondria.
  • Extraction and quantification of dNTPs is technically challenging (due to instability, binding to proteins, compartmental sequestration) → reported values and ratios can differ between labs. (OUP Academic)
  • The ratio may shift dramatically in disease (e.g., cancer), in mitochondrial DNA depletion syndromes, or under metabolic stress. (Cell)

:white_check_mark: Summary

  • In human cytosol: dATP and dTTP are the more abundant dNTPs, dCTP is somewhat lower, and dGTP is the lowest (often ~5-10% of total).
  • A rough working ratio: ~10 : 10 : 5 : 1 (dATP : dTTP : dCTP : dGTP).
  • Use this as a ballpark estimate — if you need precise numbers for a specific cell line or condition, you’ll want to check that exact study.

If you like, I can look up specific recent values from human primary cells (e.g., fibroblasts, T-cells) with contemporary LC-MS quantitation and provide a small table of absolute concentrations plus ratios.

#14

This is then the key question asked of chatGPT: “what research is there as to which bases are replaced by rNTPs in mtDNA when replication erroneously places an rNTP in the DNA rather than dNTP”

Great question. Short answer: all four rNMPs can get embedded in mtDNA during replication by POLγ, but they’re not incorporated equally. Which base shows up depends on POLγ’s base-specific discrimination and on the local rNTP:dNTP pool ratios.

Here’s what mapping/biochemistry studies show:

  • Human cells (in vivo mapping): Using HydEn-seq on HeLa and fibroblasts, Berglund et al. mapped rNMPs across mtDNA and found rCMP was most frequent, then rGMP; rAMP and rUMP were less common. They also saw strand-specific differences (e.g., rGMP most frequent on the H-strand), and showed that changing dNTP pools shifts both the frequency and the identity of embedded rNMPs. (PLOS)

  • Mechanism (in vitro with human POLγ): The same paper quantified base-specific misincorporation and confirmed that POLγ discriminates least well against rCTP and rGTP, explaining the rCMP/rGMP excess in vivo; rATP and rUTP are more strongly discriminated. (PLOS)

  • Polymerase discrimination (earlier kinetics): Biochemical work showed ~100-fold differences in POLγ’s ability to exclude different rNTPs, again indicating poorer exclusion for C and G ribonucleotides than for A and U/T. (PMC)

  • Pool-ratio effects: When dNTP supply is limited or rNTP:dNTP becomes skewed, total rNMP content rises and the base identity shifts accordingly (e.g., high rATP/dATP can drive more rAMP). Mouse/clinical contexts that perturb dNTP metabolism (TK2, DGUOK, MPV17 defects) alter the spectrum in predictable ways. (PLOS)

  • Genome/strand context: Recent mtDNA maps report zones and strand biases of embedded rNMPs associated with replication/transcription architecture (light-strand enrichment zones), reinforcing that incorporation is not random and is influenced by sequence/structural context. (OUP Academic)

Put simply

  • Which base gets “replaced” (i.e., the identity of the embedded rNMP) = function of POLγ’s base-specific discrimination (C≈G > A≈U) + the instantaneous rNTP:dNTP ratios in mitochondria + local sequence/strand context.
  • In typical human-cell conditions studied so far: rCMP and rGMP predominate; rAMP and rUMP are less frequent. (PLOS)

If you want, I can pull the specific bar-plots from the HydEn-seq paper (per-strand rAMP/rCMP/rGMP/rUMP counts) and tabulate them for quick reference.

#15

The obvious conclusion of this is that if supplementing with rNMPs it would not be sensible to supplement with large amounts of adenosine, but if supplementing with CMP there is a question as to whether the conversion of CMP this to a deoxyNucleotide happens at a better ratio by the residual expression of RNR than the rNTP dNTP ratio of Cytosine. (strictly in the mitochondria, but it is probably not massively different to the cytosol although that’s worth checking in the long term).

I remain of the view that additional DNA is a better idea than RNA, however.

This makes the assumption that it is the ratio for the specific base pair that drives the errors in replication.

#16

Not adding any technical content, but contributing with a lovely image from a Nature Portfolio article

IMG_1129
IMG_11291000×1000 153 KB

#17

Ribo nucleotides are sure beneficial, but it is 0% related to the dNTP tht you want to rebalance the ratio / pool. Take DNA as a supplement is possible, but still you will also get other unwanted purines… not just dT/dC (the one used in the studies, and beneficial potentially to avoid the mistake rNTP)

And yes forget about RNR, it will do the job only in people that need it the less, and only in some tissue… and only on a fraction (the NDP only!)… So lets consider ribo nucleotides from yeast as beneficial, but playing on others pathway not related to the topic here.

#18

Thank you for your comments. It is encouraging me to read up on other aspects. The basic question I need to look at is the extent to which RNR is expressed in non dividing unstressed cells.

#19

These are a good sequence of papers looking at RNR. Because of my obsession with acetlatyion I find the fact that high acetylation drives lower RNR activity particularly interesting.

However, it does appear that a low level of RNR is available for mtDNA activity. Hence I would argue the case for rCytosine supplementation is valid to top up the salvage pathway. However it clearlu would not replace salvage.

https://academic.oup.com/nar/article/38/22/8208/1047039

https://www.nature.com/articles/s41467-019-11214-9

https://www.sciencedirect.com/science/article/pii/S2352304222003245

#20

here is a chatgpt diagram :

Oral “yeast nucleotides” (rNMP: rC/rCMP, etc.)
│ (IAP/CD39/CD73 → dephosphorylation)
▼
Ribonucleosides (cytidine, uridine, …)
│ (CDA: cytidine → uridine; RNR works on NDP→dNDP, low flux in non-proliferative tissues)
▼
Poor effective conversion to mitochondrial dCTP/dTTP
│
└─> Mito matrix rNTP/dNTP ratio worsens → POLγ embeds more rNMP → mtDNA instability

Oral deoxynucleosides (dC + dT)
│ (ENT/CNT transport into cells)
▼
Mitochondria: TK2 → dCMP/dTMP → dCDP/dTDP → dCTP/dTTP
│ (+ SLC25A33/36 pyrimidine carriers)
▼
Favorable mito dNTP/rNTP ratio → POLγ fidelity → stable mtDNA

and his explanation : Why ribo-nucleotides aren’t pertinent: orally, “nucleotides” are first dephosphorylated to nucleosides, then rely on RNR (tightly regulated, low in non-dividing tissues) and suffer CDA deamination, so they don’t efficiently raise mitochondrial dNTPs and can worsen the rNTP/dNTP balance.
The practical, targeted solution: deliver deoxynucleosides dT + dC to feed TK2 directly and elevate dTTP/dCTP in the mitochondrial matrix, lowering rNMP embedding and stabilizing mtDNA

I follow “him” on that. RNR is actually low as we age + not active in main tissue without replication such as muscle and brain… So, no choice to find a way to consume dT + dC as it was done in the study. Im currently looking to create it for me ^^