Nucleotide imbalance decouples cell growth from cell proliferation

Nucleotide precursors can inhibit cell proliferation

Diverse metabolic perturbations can inhibit cell proliferation, including disruptions to pro-growth signalling pathways, amino acid availability, mitochondrial respiration and nucleotide synthesis (Fig. 1a). Obtaining nucleotides can be particularly limiting for cell proliferation19,20,21. Indeed, pharmacological inhibition of purine production with lometrexol (LTX) or pyrimidine production with brequinar (BRQ) depletes total purine or pyrimidine levels and blocks proliferation, consistent with previous studies11,22,23 (Fig. 1a and Extended Data Fig. 1a). Intriguingly, thymidine treatment has long been used to arrest and synchronize cells; however, the proximal mechanisms of this arrest, and whether this has broader implications for the regulation of nucleotide homeostasis, are less clear. Moreover, thymidine exists uniquely in the dNTP pool, and it is unclear whether perturbations to ribonucleotide (NTP) pools are equally detrimental. To investigate this, we supplemented cells with individual nucleobases and nucleosides, which can be salvaged to produce nucleotides. Nucleotide salvage preserves metabolic substrates that would otherwise be needed for de novo nucleotide synthesis. However, we found that single nucleobase or nucleoside supplementation impaired proliferation in a dose-titratable manner (Fig. 1b–f).

Fig. 1: Excess individual nucleotides can impair proliferation.figure 1

a, Proliferation rates of A549 cells cultured in standard conditions (untreated) or with 1 µM Torin1, without serum (−FBS), without arginine (−Arg), without leucine (−Leu), with 100 nM rotenone, with 5 nM oligomycin, with 1 µM lometrexol (LTX) or with 1 µM brequinar (BRQ). bf, Proliferation rates of A549 cells treated with the indicated concentration of guanine (G), thymidine (T), deoxyadenosine (dA), cytidine (C) or adenine (A). Each of these nucleobases/nucleosides can be salvaged to produce intracellular nucleotides as shown. Data are presented as mean ± standard deviation (s.d.) of three biological replicates. Source numerical data are available in source data.

Source data

Expression of nucleotide salvage and synthesis enzymes, as well as transporters, varies across cells and could affect sensitivity to individual nucleobase or nucleoside addition. Consistent with this, different cells had differential sensitivity to each species, although nucleotide precursor addition could inhibit proliferation of all cells tested, including non-transformed cells (Extended Data Fig. 1b,c). Interestingly, deoxycytidine (at concentrations up to 14 mM) was the only precursor tested that did not inhibit cell proliferation (Extended Data Fig. 1d). As most cells tested exhibited greatest sensitivity to guanylate nucleotide precursors (Extended Data Fig. 1e), we focused further mechanistic studies on understanding the effects of guanine supplementation. Importantly, a functional salvage pathway was needed for the corresponding nucleotide precursor to inhibit proliferation: cells deficient for APRT and HPRT, the enzymes that salvage adenine and guanine, were not sensitive to these precursors, and thymidine kinase-deficient 143B cells were unaffected by thymidine addition (Extended Data Fig. 1b,f).

Nucleotide imbalance impairs cell proliferation

We reasoned that salvage of single nucleobases/nucleosides might perturb relative levels of intracellular nucleotide species and measured nucleotide levels in cells with or without guanine addition. Guanine supplementation increased intracellular levels of guanylate (G) nucleotides (GTP/GDP/GMP) and unexpectedly decreased intracellular levels of adenylate nucleotides (A) (ATP/ADP/AMP) (Fig. 2a and Extended Data Fig. 2a). These data suggest that guanine salvage disrupts relative levels of purines by increasing the ratio of G to A nucleotides. Notably, providing adenine together with guanine restored the balance of G and A nucleotides (Fig. 2a). To understand how providing guanine depletes intracellular A nucleotides, we measured the contributions of salvage and de novo synthesis to intracellular purines. To assess de novo synthesis, we measured incorporation of amide-15N-glutamine into purines, as the amide nitrogen of glutamine is incorporated during AMP and GMP synthesis. To assess salvage, we measured incorporation of 13C-guanine and 13C-adenine into purines (Fig. 2b). As expected, a subset of purines in untreated cells were 15N-labelled, reflecting their production via de novo synthesis (Fig. 2c and Extended Data Fig. 2b). Providing 13C-adenine increased levels of A nucleotides, the majority of which were 13C-labelled and therefore derived from adenine salvage. Similarly, salvage of 13C-guanine accounted for increased G nucleotide levels upon guanine supplementation. Moreover, providing either 13C-adenine or 13C-guanine eliminated the contribution of de novo synthesis to intracellular purines. This likely reflects known allosteric feedback regulation of purine synthesis enzymes: A and G nucleotides can inhibit ribose-5-phosphate pyrophosphokinase and glutamine phosphoribosyl pyrophosphate amidotransferase, which catalyse the initial steps of de novo purine synthesis3,24,25. Therefore, aberrantly high G nucleotides derived from guanine salvage can inhibit de novo synthesis of both G and A nucleotides (Extended Data Fig. 2c), resulting in A nucleotide depletion.

Fig. 2: Nucleotide salvage leading to imbalanced nucleotide pools inhibits cell proliferation.figure 2

a, GTP and ATP levels in A549 cells cultured in standard conditions (none) or treated for 24 h with 200 µM guanine (G) with or without 200 µM adenine (A) as indicated. b, Schematic showing how stable isotope tracing was used to determine the source of intracellular purines. Salvage of 13C-guanine or 13C-adenine produces 13C-labelled GTP and ATP. The 15N label from amide-15N-glutamine is incorporated in de novo purine synthesis, producing 15N-labelled ATP and GTP. c, Total levels and labelling of GTP and ATP in A549 cells cultured for 24 h in medium containing amide-15N-glutamine with or without 200 µM 13C-guanine and/or 13C-adenine as indicated. d, Fold change in the specified intracellular NTP levels in A549 cells cultured with the indicated concentrations of nucleotide precursors compared with those found in cells cultured in standard conditions. e, Fold change in the specified intracellular dNTP levels in A549 cells cultured with the indicated concentrations of nucleotide precursors compared with those found in cells cultured in standard conditions. f, Fold change in the specified intracellular NTP levels in A549 or U2OS cells cultured with the indicated concentrations of nucleotide precursors compared with those found in cells cultured in standard conditions. g, Fold change in the specified intracellular dNTP levels in A549 or U2OS cells cultured with the indicated concentrations of nucleotide precursors compared with those found in cells cultured in standard conditions. h, Proliferation rates of A549 cells cultured in standard conditions (none) or treated with 200 µM G with or without 200 µM A. i, Proliferation rates of A549 cells cultured in standard conditions (none) or treated with 1 mM T with or without 1 mM C. All nucleotide levels were measured using LCMS. Fold changes in nucleotide levels were calculated from absolute concentrations presented in Extended Data Fig. 2d–g. Data are presented as mean + /- SD of 3 biological replicates. Source numerical data are available in source data.

Source data

We hypothesized that analogous imbalances in nucleotide levels account for impaired proliferation upon salvage of other nucleotide precursors (Fig. 1b–f). To test this, we used liquid chromatography–mass spectrometry (LCMS) to quantify absolute intracellular nucleotide levels upon addition of A, G, T and C precursors. (Extended Data Fig. 2d,e). At concentrations that inhibit proliferation, each precursor increased intracellular concentrations of at least one nucleotide species and decreased intracellular concentrations of at least one other nucleotide species. Consistent with T being exclusive to the dNTP pool, T treatment caused altered dNTP levels but not NTP levels. Determining the fold change in levels of each intracellular nucleotide caused by A, G, T or C addition revealed that salvage of different precursors altered relative nucleotide levels in different ways (Fig. 2d,e). Thus, a change in any specific nucleotide species does not explain decreased proliferation across these conditions. Rather, these data argue that cells are vulnerable to multiple different perturbations to the balance among nucleotide species. This led us to define nucleotide imbalance as a detrimental increase in one or more nucleotide species above normal levels along with a decrease in one or more other nucleotide species below normal levels. Importantly, this is distinct from depletion of purines, pyrimidines or all NTP or dNTP species.

To understand what degree of imbalance is needed to impair proliferation, we took advantage of differential sensitivity to nucleotide precursors across cell types. A549 cells are more sensitive than U2OS cells to G, but are less sensitive to A (Fig. 1b,f and Extended Data Fig. 1b). Comparing A549 and U2OS cells treated with concentrations of G or A that differentially impair proliferation revealed that similar magnitudes of change to nucleotide levels were detrimental to each cell type (Fig. 2f,g and Extended Data Fig. 2f,g). This suggests that differential sensitivity to each nucleotide may be attributable to varied transport or salvage activity across cell types and underscores that nucleotide imbalances are detrimental to proliferation. Indeed, providing adenine to re-establish purine balance restored proliferation of guanine-treated cells (Fig. 2d,e,h and Extended Data Fig. 2d,e,h). Providing cytidine also restored nucleotide balance and proliferation of thymidine-treated cells (Fig. 2d,e,i and Extended Data Fig. 2d,e).

Salvage of individual nucleotides altered intracellular levels of both NTPs and dNTPs. dGTP has the same molecular weight as ATP and similar chromatographic properties, and because ATP is much more abundant in cells, dGTP was not confidently distinguished by LCMS. Nevertheless, addition of each nucleotide precursor at concentrations that inhibit proliferation caused imbalances among dNTP species (Fig. 2e,g and Extended Data Fig. 2e,g), raising the possibility that imbalanced dNTPs play a role in impairing proliferation upon nucleotide precursor addition.

Nucleotide imbalance slows S phase progression

To test whether nucleotide imbalance impairs proliferation by impeding DNA replication during S phase, we monitored cell cycle progression following G treatment. We used flow cytometry to measure DNA content and incorporation of 5-ethynyl-2′-deoxyuridine (EdU) into DNA, which reflects active DNA replication (Fig. 3a). Untreated cells contain populations in all cell cycle phases; as reported in classic studies, serum starvation causes G1 arrest, while Taxol treatment causes G2/M arrest (Extended Data Fig. 3a,b)26. Consistent with its effect on proliferation (Fig. 1b), guanine treatment had a dose-dependent effect on cell cycle progression: increasing concentrations of guanine caused cells to accumulate in S phase, and at the highest concentration cells failed to incorporate EdU (Fig. 3b and Extended Data Fig. 3c). Similarly, guanine treatment for 96 h initially increased the population of cells in S phase and later prevented EdU incorporation (Extended Data Fig. 3d). Providing adenine together with guanine to restore nucleotide balance restored normal cell cycle distribution (Fig. 3c and Extended Data Fig. 3c). Imbalances caused by other nucleotide precursors also impaired S phase progression across cell types (Extended Data Fig. 3e,f). Treatment with LTX or BRQ to deplete all purines or pyrimidines, respectively, prevented EdU incorporation by 96 h but did not cause the same extent of S phase accumulation as guanine treatment (Extended Data Fig. 3d). Thus, while nucleotide imbalance impairs proliferation by slowing S phase progression, purine or pyrimidine depletion may inhibit proliferation in part through a different mechanism.

Fig. 3: Nucleotide imbalance impairs S phase progression.figure 3

a, Approach using flow cytometry to assess cell cycle phase by DNA content (as determined by propidium iodide staining) and EdU incorporation. b, Cell cycle distribution of A549 cells cultured with the indicated concentration of guanine (G) for 24 h. c, Cell cycle distribution of A549 cells treated with or without 200 µM G with or without 200 µM adenine (A) for 24 or 48 h. d, Approach to assess S phase progression. After pulsing cells with EdU, cell cycle progression of EdU-positive and EdU-negative populations was monitored. e, Cell cycle distribution of A549 cells pulsed with EdU, then cultured with or without 200 µM G for the indicated time. Percentage of total cells that are EdU-positive and in G1, S or G2/M phase is shown. f, mVenus-Gem1 fluorescent reporter to assess cell cycle dynamics in live cells. g, Representative images from live-cell imaging of A549 cells expressing mVenus-Gem1 cultured with or without 200 µM G (see also Supplementary Videos 13). h, Fraction of cells cultured with or without 200 µM G that began the experiment in G1 phase and entered S phase (assessed by live-cell imaging of A549 cells expressing mVenus-Gem1; 76 cells were analysed). i, Duration of S/G2 phase in cells cultured with or without 200 µM G (assessed by live-cell imaging of A549 cells expressing mVenus-Gem1; 115 cells were analysed). j, Cell cycle distribution of A549 cells synchronized in G2 phase with 4.5 µM RO-3306 for 18 h, then released from arrest and treated with standard culture media (untreated), 25 µM G (low G) or 200 µM G (high G) as indicated. k, dNTP levels in A549 cells 21 h after release from RO-3306 and subsequent treatment with or without low G or high G as indicated. dNTP levels in unsynchronized cells cultured with or without low G or high G for 24 h are also shown. dNTPs were measured using LCMS. Data are presented as mean ± s.d. of three biological replicates. Source numerical data are available in source data.

Source data

To more directly test whether nucleotide imbalance slows DNA replication, we pulsed cells with EdU to mark the population in S phase at t = 0, then monitored S phase progression. In untreated cells, the EdU-positive population progressed to 4 N DNA content and then back to 2 N DNA content, reflecting completion of S phase and return to G1 after cell division (Fig. 3d,e). This does not occur when cells are arrested in G1 or G2/M (Extended Data Fig. 3g). While untreated EdU-positive cells completed S phase and divided within 24 h, guanine-treated EdU-positive cells failed to divide by 24 h (Fig. 3e), suggesting that DNA replication is slowed. Further, a population of EdU-negative cells with intermediate DNA content accumulated during guanine treatment. Initially, EdU-negative cells with 2 N DNA content are in G1 phase. Thus, accumulation of EdU-negative cells at intermediate DNA content argues that cells enter S phase with nucleotide imbalance, but progression through S phase is impaired.

To define kinetics of S phase entry and duration, we performed live-cell imaging using a previously described fluorescent reporter27 where mVenus is conjugated to a truncated form of geminin, whose degradation is cell cycle dependent. Cells expressing mVenus-Gem1 have fluorescent nuclei between the G1/S transition and mitosis, allowing for specific monitoring of S phase entry and quantification of S/G2 and G1 durations (Fig. 3f,g and Supplementary Videos 13). Almost all guanine-treated cells entered S phase, but subsequently had much longer S/G2 duration than untreated cells (Fig. 3h,i). Guanine treatment also increased G1 duration in cells born after induction of nucleotide imbalance, though not to the same extent as S/G2 duration (Extended Data Fig. 3h).

We next tested whether dNTP imbalance persisted through S phase by synchronizing cells in G2 phase using the CDK1 inhibitor RO-3306 (ref. 28) and then releasing cells into the following cell cycle. Importantly, this strategy does not directly perturb cell metabolism. Untreated cells entered S phase around 15 h after release and progressed to late S phase around 21 h after release, while high concentrations of guanine resulted in slower S phase progression (Fig. 3j). Guanine treatment caused dNTP imbalance 21 h after release (Fig. 3k), demonstrating that nucleotide imbalance perturbs dNTP availability during S phase. Together, these data suggest that cells lack a mechanism to prevent S phase entry with imbalanced nucleotides, leading to impaired DNA replication and S phase progression.

Growth control pathways do not sense nucleotide imbalance

As numerous growth signalling pathways regulate nucleotide metabolism, we asked whether these pathways decrease growth in coordination with decreased proliferation under nucleotide imbalance. Although mTORC1 responds to purine depletion11,12, we found that mTORC1 signalling remains active despite nucleotide imbalance (Fig. 4a and Extended Data Fig. 4a,d). Activity of other major growth regulatory pathways, Akt and AMPK, also did not correlate with proliferation arrest under nucleotide imbalance (Extended Data Fig. 4b,c). Interestingly, decreased mTORC1 activity upon nucleotide depletion did not prevent continued growth in H1299 cells (Extended Data Fig. 4d,k). Additionally, guanine-treated cells grew aberrantly large, and incorporation of puromycin into nascent peptides showed that protein synthesis rates were unchanged (Fig. 4b,c). Indeed, guanine-treated cells synthesized protein in coordination with increasing cell volume, as measured by a YFP protein synthesis reporter29,30 (Fig. 4d and Extended Data Fig. 4e). Thus, protein concentration and overall cell density are maintained despite a larger cell size (Fig. 4e,f). These data suggest that canonical growth signalling does not directly respond to nucleotide imbalance.

Fig. 4: Nucleotide imbalance causes continued cell growth without division.figure 4

a, Phosphorylation of ribosomal protein S6 and S6 kinase (S6K) in A549 cells cultured with or without 1 µM Torin 1, or 200 µM guanine (G), 1 µM lometrexol (LTX) or 1 µM brequinar (BRQ) for the indicated time. Levels of vinculin, total S6K and total S6 are also shown. b, Proliferation rate (left) and mean volume (right) of A549 cells cultured with or without 200 µM G. c, Global protein synthesis measured by puromycin incorporation into nascent peptides in A549 cells cultured with or without 200 µM G for 96 h. Cycloheximide treatment was used as a negative control. d, Cell number (left), mean volume (centre) and protein accumulation (right) in A549 cells treated with 200 µM G. Protein accumulation was determined using a YFP reporter (Extended Data Fig. 4e). e, Protein concentration in A549 cells cultured with or without 200 µM G, calculated by dividing total protein by cell number and volume. f, Density of A549 cells cultured with or without 200 µM G for 72 h, calculated by dividing cell mass by cell volume. g, Mean volume of A549 cells treated for 96 h with the indicated concentrations of G, thymidine (T), deoxyadenosine (dA), cytidine (C) or adenine (A). h, Mean volume of A549 cells cultured with or without 200 µM G with or without 200 µM A for 96 h. i, Mean volume of A549 cells cultured with or without 1 mM T with or without 1 mM C for 96 h. j, Proliferation rate and size of A549 cells cultured in conditions that perturb cell metabolism. Data are compiled from experiments shown in Figs. 1a–f, 2h,i and 4g–i, and Extended Data Fig. 4j. Conditions are grouped into signalling disruption (Torin treatment or serum withdrawal), amino acid limitation (leucine or arginine starvation), electron transport chain (ETC) inhibition (oligomycin or rotenone treatment), purine or pyrimidine depletion (using LTX or BRQ), or nucleotide imbalance. Data are presented as mean ± s.d. of three biological replicates. Source numerical data and unprocessed blots are available in source data.

Source data

Other nucleotide imbalances also caused cells to grow aberrantly large in a dose-titratable manner that matched their anti-proliferative effects and was observed across cell types. (Fig. 4g and Extended Data Fig. 4f,g). Adding nucleobases/nucleosides at concentrations that do not affect proliferation did not change cell size (Extended Data Fig. 4h). Re-establishing nucleotide balance restored normal size in cells treated with G- or T-nucleotide precursors (Fig. 4h,i). Most other metabolic perturbations did not robustly increase cell size, with the exception of pyrimidine synthesis inhibition (Extended Data Fig. 4i). Pyrimidine depletion also did not inhibit mTORC1 signalling, as previously reported11,12. Thus, while metabolic state, growth and proliferation are normally tightly linked, cell growth is decoupled from proliferation following nucleotide imbalance (Fig. 4j).

The purine synthesis inhibitor LTX depletes both A and G nucleotides and inhibits proliferation (Extended Data Figs. 1a and 4j,k). In cells where LTX inhibits growth and mTORC1 activity, we asked whether supplementing purine-depleted cells with either adenine or guanine to cause purine imbalance decouples growth from proliferation. Adenine and guanine can reactivate mTORC1 in purine-depleted cells, but the time required for A versus G nucleotides to induce mTORC1 activity may be variable11,12. We found that both guanine and adenine could activate mTORC1 acutely following purine depletion and sustain signalling over longer time periods (Extended Data Fig. 4l,m). However, activation of growth signalling is not sufficient for proliferation: providing excess adenine or guanine did not restore proliferation (Extended Data Fig. 4n). Of note, low concentrations of adenine that do not induce nucleotide imbalance could rescue proliferation of LTX-treated cells. This may be explained by the ability of AMP deaminase to convert AMP to IMP, which can then be converted to GMP to potentially replenish both A and G nucleotides. Together, these data suggest that, while sufficient levels of either purine can restore growth, balanced levels of purines are required for proliferation. Further, while purine-depleted cells (with inactive mTORC1 signalling) accumulate in G1 phase, providing guanine to these cells caused S phase entry and subsequent S phase stalling (Extended Data Fig. 4o). We therefore hypothesized that mTORC1 activity is needed for S phase entry under nucleotide imbalance, and consistent with this, pharmacological inhibition of mTORC1 prevented guanine-treated cells from entering S phase (Extended Data Fig. 4o).

Nucleotide imbalance activates replication stress signalling

Impaired S phase progression suggests stalled DNA replication, and we therefore tested whether nucleotide imbalance causes DNA replication stress. The ATR and ATM kinases sense single-stranded DNA and DNA double-strand breaks, respectively, and their respective downstream targets, Chk1 and Chk2, are major DNA damage response (DDR) effectors (Fig. 5a)31,32,33. Guanine treatment caused robust phosphorylation of both Chk1 and Chk2, with higher concentrations of guanine that inhibit proliferation to a greater extent inducing a stronger signalling response (Fig. 5b). Interestingly, Chk1 was phosphorylated first within 24 h, followed by phosphorylation of Chk2 between 48 and 72 h. This may indicate that following purine imbalance, replication fork stalling first activates ATR, with later activation of ATM. Addition of adenine together with guanine prevented replication stress response induction (Extended Data Fig. 5a). Nucleotide imbalances induced by other precursors also activated ATR and ATM signalling, while as expected, using leucine deprivation to inhibit proliferation did not (Fig. 5c and Extended Data Fig. 5a–c). Inhibiting total purine or pyrimidine synthesis induced phosphorylation of Chk1 and Chk2 to a lesser extent than guanine treatment, consistent with fewer cells stalling in S phase in these conditions (Fig. 5d and Extended Data Fig. 3d).

Fig. 5: Replication stress signalling promotes cell survival and recovery from nucleotide imbalance.figure 5

a, ATR and ATM kinases respond to replication stress and DNA damage. ATR and ATM phosphorylate Chk1 and Chk2, respectively. b, Phosphorylation of Chk1 and Chk2 in A549 cells treated for the indicated time with the indicated concentration of guanine (G). c, Phosphorylation of Chk1 and Chk2 in A549 cells treated for the indicated amount of time with 1 mM thymidine (T), 2.5 mM adenine (A) or 1.5 mM deoxyadenosine (dA). d, Phosphorylation of Chk1 and Chk2 in A549 cells treated for the indicated time with 200 µM G, 1 µM LTX or 1 µM BRQ. Levels of vinculin are also shown in all western blots as a loading control. e, Proliferation rates of A549 cells treated with the indicated concentration of guanine with or without 50 nM of the ATR kinase inhibitor AZ20 (ATRi) as indicated. f, Proliferation rates of A549 cells cultured with or without 2 mM A, 1.5 mM dA, 200 µM G or 1 mM T, with or without 50 nM ATRi as indicated. g, Cell fate of A549 cells expressing the mVenus-Gem1 reporter that were in G1 phase at the time of addition of 200 µM G with or without 50 nM ATRi, as assessed using live-cell imaging. The fate of cells in that were in G1 at the beginning of the experiment and were not exposed to excess G is also shown (untreated). In total, 124 cells were analysed. h, Approach to assess how cells recover from treatment with excess G. Cells were cultured in medium containing 200 µM G with or without 50 nM ATRi for 4 days. Medium was then changed to untreated medium or medium containing 50 nM ATRi, and cell number was determined every 24 h for 14 days thereafter. i, A549 cell number over time after release from treatment with G with or without ATRi treatment as outlined in h. Data are presented as mean ± s.d. of three biological replicates. Source numerical data and unprocessed blots are available in source data.

Source data

Recent work demonstrated that pharmacological G nucleotide synthesis inhibition with the IMPDH inhibitor mycophenolic acid (MPA) can have dose-dependent effects: low-dose MPA increased p53 and p21 protein levels and caused cells to accumulate in G1 phase after 24 h, whereas high-dose MPA caused p21 degradation and increased the number of cells in S phase34. We tested whether guanine addition has similar dose-dependent effects on p53 and p21 levels. Consistent with its effects on ATR and ATM signalling, increasing concentrations of guanine increased p53 levels; however, higher guanine concentrations did not increase p21 degradation (Extended Data Fig. 5d). This suggests that nucleotide synthesis inhibition and excess nucleotide salvage can have differing effects on cells.

The role of ATR and ATM in the cellular response to DNA damage has been extensively studied31,32; however, only a small fraction of guanine-treated cells exhibited minor increases in DNA damage at 24 h as measured by a comet tail assay (Extended Data Fig. 5e). At that time, the signalling response is already robust, indicating that replication stress-sensing pathways are activated under nucleotide imbalance without large amounts of DNA damage. Further, the failure of metabolic regulatory mechanisms to prevent S phase entry with imbalanced nucleotides suggests that replication stress sensing constitutes the major signalling response to nucleotide imbalance.

ATR activity is required to survive nucleotide imbalance

ATR and ATM activate downstream effectors that block cell cycle progression (Extended Data Fig. 6a)31,33,35. ATR-mediated cell cycle arrest might therefore explain why nucleotide imbalance prevents proliferation. If so, inhibiting ATR would allow cells to continue proliferating despite nucleotide imbalance. Using the ATR inhibitor AZ20 (ref. 36,37), we found that, instead of restoring proliferation, ATR inhibition increased cell death following guanine treatment. (Fig. 5e and Extended Data Fig. 6b–d). ATR inhibition increased sensitivity to all nucleotide imbalances but did not increase sensitivity to purine or pyrimidine depletion (Fig. 5f and Extended Data Fig. 6e–g), consistent with a less robust induction of ATR signalling in these conditions. Together, these data suggest that replication stress signalling can be a protective mechanism to enable cell survival with nucleotide imbalance.

Live-cell imaging with the mVenus-Gem1 cell cycle reporter (Fig. 3f) showed that cells can successfully divide if they have partially completed S phase before guanine addition, though ATR inhibition caused some of these cells to arrest or die (Extended Data Fig. 6h). This is consistent with the expected kinetics of nucleotide imbalance: because these cells turn over purine nucleotide pools in approximately 24 h (ref. 18), inhibition of A nucleotide synthesis following excess guanine salvage (Fig. 2c and Extended Data Fig. 2c) would not immediately deplete A nucleotides. Thus, cells that have already partially replicated their DNA can likely complete replication before purine balance is drastically changed. In contrast, cells in G1 at the time of guanine supplementation that enter S phase with imbalanced nucleotides are unable to divide, and instead either arrest or die after entering S phase (Fig. 5g). ATR inhibition caused most cells to die upon entering S phase with imbalanced purines, suggesting that ATR is critical for preventing S phase catastrophe under nucleotide imbalance. Interestingly, daughter cells born later after induction of nucleotide imbalance and ATR inhibition were more likely to become stalled in G1 (Extended Data Fig. 6i). Replication stress in mother cells can affect G1 length and lead

留言 (0)

沒有登入
gif