Mitochondrial deoxynucleotide pool sizes in mouse liver and evidence for a transport mechanism for thymidine monophosphate
-
Contributed by Peter Reichard, October 13, 2006 (received for review September 29, 2006)
Abstract
Dividing cultured cells contain much larger pools of the four dNTPs than resting cells. In both cases the sizes of the individual pools are only moderately different. The same applies to mitochondrial (mt) pools of cultured cells. Song et al. [Song S, Pursell ZF, Copeland WC, Longley MJ, Kunkel TA, Mathews CK (2005) Proc Natl Acad Sci USA 102:4990–4995] reported that mt pools of rat tissues instead are highly asymmetric, with the dGTP pool in some cases being several-hundred-fold larger than the dTTP pool, and suggested that the asymmetry contributes to increased mutagenesis during mt DNA replication. We have now investigated this discrepancy and determined the size of each dNTP pool in mouse liver mitochondria. We found large variations in pool sizes that closely followed variations in the ATP pool and depended on the length of time spent in the preparation of mitochondria. The proportion between dNTPs was in all cases without major asymmetries and similar to those found earlier in cultured resting cells. We also investigated the import and export of thymidine phosphates in mouse liver mitochondria and provide evidence for a rapid, highly selective, and saturable import of dTMP, not depending on a functional respiratory chain. At nM external dTMP the nucleotide is concentrated 100-fold inside the mt matrix. Export of thymidine phosphates was much slower and possibly occurred at the level of dTDP.
Mitochondria contain pools of the four canonical dNTPs for the synthesis of mitochondrial (mt) DNA, separated by the mt double membrane from the corresponding cytosolic dNTP pools for nuclear DNA synthesis (1). mt DNA synthesis occurs throughout the whole life of a cell, independent of nuclear DNA synthesis in S-phase, and mt dNTPs must therefore be available also in cells outside S-phase, when dNTP pools are not required for nuclear DNA synthesis. In cultured cells we found much larger dNTP pools in the cytosol than in mitochondria, but we found roughly equal proportions among the four dNTPs in these two compartments. This applied to both cycling (2) and resting (3) cells even though pool sizes were much smaller in the latter. Pulse–chase experiments with radioactively labeled thymidine demonstrated separate metabolic compartments for mt and cytosolic dTTP pools, in rapid communication with each other. mt dTTP was formed by two separate pathways: (i) de novo synthesis from ribonucleotides in the cytosol, followed by import into mitochondria; or (ii) phosphorylation of imported thymidine by a mt thymidine kinase. Cycling cells largely depended on the first pathway whereas resting cells used the second. These results may explain why genetic defects (4–6) involving enzymes of the second pathway mostly affect tissues with terminally differentiated cells.
A recent article in PNAS by Song et al. (7) reported a large asymmetry among the four mt dNTP pools in various rat organs, with dGTP in some cases being several-hundred-fold more abundant than dTTP. In model experiments the asymmetries greatly affected the fidelity of DNA synthesis in vitro. They were suggested to lead to an increased mutability of the mt genome in vivo. The pool data appear to be at variance with the scenario for mt dNTP synthesis in cultured cells suggested by us.
Earlier analyses of the size and composition of adenosine phosphate pools of animal organs showed the crucial importance of the methodology used for the isolation of the pools before analysis (8, 9). Mainly because of the immediate onset of anaerobiosis at death, extreme precaution must be exercised not to degrade ATP to AMP and further to nucleosides before extraction from tissues. It is not unlikely that similar considerations apply also to other ribonucleoside triphosphates (rNTPs) and dNTPs. Song et al. do not report the level of ATP in their experiments (7). We decided to reinvestigate the size of mt dNTP pools and to relate our determinations to the ATP pool of the isolated mitochondria. We used livers from inbred mice rather than rat livers because they can be handled more rapidly. We then found that the measured size of dNTP pools was highly variable and depended on the ATP pool and the efficacy of our preparation procedure. The dTTP/dTDP/dTMP pool ratio was closely related to the ATP/ADP/AMP ratio, and also a sizeable fraction of the other three deoxynucleotide pools consisted of monophosphates and diphosphates. However, also when considering these complications in the determinations of total deoxyribonucleotide pools we did not find the asymmetries reported by Song et al. (7).
An important aspect of our scenario for the metabolism of mt dTTP involves a hypothetical transport mechanism between the cytosol and mitochondria for thymidine phosphates. We now present strong evidence for a specific transport of dTMP from the cytosol to mitochondria. At outside concentrations of dTMP in the nM range the nucleotide was enriched at least 100-fold in the mt matrix by a rapid, saturable, and specific uniport mechanism. dTDP and dTTP were not imported. We also found some evidence for a slow export of thymidine nucleotides with a preference for dTDP.
Results
Covariation of dNTP and rNTP Pools with ATP in Preparations of Mouse Liver Mitochondria.
We found earlier (10) that in mitochondria from cultured cells ATP amounted to one-third or slightly more of the total adenine nucleotide pool, the remainder being made up of roughly equal amounts of ADP and AMP. The pool in the cytosol instead consisted of 80–90% of ATP, and AMP was barely detectable. We found a similar difference between mitochondria and cytosol for thymidine phosphates, with the dTTP/dTDP/dTMP ratio closely following the corresponding ATP/ADP/AMP ratio (10). On that occasion, we did not investigate the behavior of other nucleotides.
We now determined in various consecutive preparations of mouse liver mitochondria the concentrations of ATP, ADP, and AMP as well as the concentrations of the triphosphates of the other common ribonucleoside and deoxyribonucleosides. We encountered large variations in the total content of ATP and the ATP/ADP/AMP ratio, probably depending on the degree of anaerobiosis during the preparation of mitochondria (8, 9). In early experiments we combined livers from several mice before homogenization and then obtained mitochondria in which 1–10% of the total adenine nucleotides was ATP, the remainder being mostly AMP. In later experiments we prepared mitochondria from only a single liver, with rapid cooling and homogenization. The yield of total adenine nucleotides then increased, with a 30–50% content of ATP.
Altogether we analyzed nine different preparations of mitochondria. In each we measured the concentrations of ATP and the four dNTPs (Fig. 1 A), in four of them also rNTPs (Fig. 1 B). Note the different scales for ATP, dNTPs, and rNTPs in Fig. 1. dNTPs and rNTPs both covaried with ATP. In Exps. 1, 2, and 3 the ATP concentrations were, respectively, 70, 290, and 970 pmol/mg, and the values for the other nucleotides were correspondingly low. In later experiments the ATP concentrations increased up to 3,200 (Exp. 7), with corresponding increases in other nucleotides. In some cases oligomycin (Exp. 5), rotenone (Exp. 6), or atractyloside (Exp. 8) was present during the preparation of mitochondria, but they did not affect the relations between dNTPs and ATP. Table 1 summarizes average ratios between triphosphates and ATP in mitochondria and also shows values for the cytosol from a single experiment (Exp. 9). In mitochondria all dNTPs amounted to ≈0.1% of ATP. The relative values for dNTPs are similar to those found earlier in resting cultured cells (3), with dCTP and dATP providing the largest pools, followed by dTTP and finally dGTP. Similar relations were found in the cytosol. In no case did we find the asymmetries reported by Song et al. (7), who reported the dGTP pool to be 10 times larger than the dTTP pool in liver mitochondria.
The measured size of the mt dNTP and rNTP pools depends on the measured size of the ATP pool. (A) Size of dNTP pools and ATP pool in consecutive preparations of liver mitochondria. The ordinate is in pmol/mg mitochondria for dNTPs and in nmol/mg mitochondria for ATP. Exp. 1 is the earliest experiment, and Exp. 9 is the most recent one. Buffer A contained 25 μg/ml oligomycin in Exp. 5, 25 μM rotenone in Exp. 6, and 50 μM atractyloside in Exp. 8. (B) Size of rNTP pools. rNTPs are given as pmol, and ATP is given as nmol. dNTP and rNTP pools were determined by the polymerase assay and by HPLC, respectively.
Ratios between the concentrations of nucleoside triphosphates and ATP in mouse liver mitochondria
For ATP the large differences in pool size were, at least in part, related to dephosphorylation of ATP to ADP and AMP during preparation of mitochondria. If other nucleoside triphosphates are dephosphorylated similarly and their maintenance depends on the ATP level, this would explain the observed relation between them and ATP shown in Fig. 1. Because we could not directly determine concentrations of monophosphates and diphosphates other than ADP and AMP, we resorted to an indirect isotopic method. In two experiments we injected a mouse with a tracer dose of [3H]thymidine, prepared liver mitochondria, and compared the amounts of radioactive dTMP, dTDP, and dTTP with those of ATP, ADP, and AMP (Table 2). We also determined the concentrations of thymidine and adenosine nucleotides after a 10-min incubation of the labeled mitochondria under the conditions detailed in Materials and Methods for transport experiments that favor ATP regeneration. In both experiments we found similar phosphorylation levels for thymidine and adenosine nucleotides in mitochondria before incubation, with ATP and dTTP representing slightly more than one-third of each total nucleotide pool. Incubation increased the relative amount of both dTTP and ATP. We found that dTTP had the same specific radioactivity before and after incubation (data not shown), demonstrating that mt dTTP, dTDP, and dTMP had reached isotope equilibrium during the in vivo labeling period and that we could use the value for the specific radioactivity of dTTP to calculate pmol from radioactivity also for dTDP and dTMP. The results then demonstrated that the phosphorylation level of thymidine nucleotides closely followed that of adenosine nucleotides (Table 2).
Relations between phosphorylation levels of adenosine and thymidine nucleotides in mouse liver mitochondria
To investigate whether a similar situation occurred with dATP, dCTP, and dGTP, we analyzed in two additional experiments all four deoxynucleoside triphosphates directly after preparation of mitochondria and after incubation of respiring mitochondria (Table 3). In the first experiment we analyzed only the mitochondria; in the second experiment we also analyzed the supernatant solution after centrifugation of the incubated mitochondria and, in addition, investigated whether the presence of ATP during incubation changed the amount of dNTPs. In both experiments all four dNTPs increased inside the mitochondria during incubation, with a minor part being released into the incubation medium (Table 3). When ADP had been added during incubation, the increase was larger and amounted to at least a doubling of the total amount of each dNTP. When added at 0.1 mM concentration, >90% of ADP was recovered as ATP in the medium in <1 min. The increase in the concentration of dNTPs must have occurred by phosphorylation of monophosphates and diphosphates, demonstrating that the isolated mitochondria in addition to triphosphate pools contained sizeable pools of monophosphates and diphosphates of all deoxynucleosides.
Increase in dNTP pools during incubation of mouse liver mitochondria
Mitochondria Import Thymidine Monophosphate.
In cycling cells the majority of the intramitochondrial deoxynucleotides is synthesized in the cytosol and imported into mitochondria (10). To study the import of thymidine nucleotides, we incubated respiring liver mitochondria with isotopically labeled dTTP or dTMP in the presence and absence of a large excess of ADP, separated the mitochondria from the incubation medium, and investigated by HPLC the distribution of nucleotides in the medium and in mitochondria.
Incubation of mitochondria with 5 nM dTMP in the medium resulted in a rapid import of the nucleotide (Fig. 2 A) with a concomitant decrease of its concentration outside (Fig. 2 B). Already after 2 min almost 40% of the total dTMP was inside, and prolonged incubation gave no further increase. However, both in the medium and in mitochondria, a fraction of dTMP was degraded to thymidine, probably by pyrimidine 5′-deoxyribonucleotidases (11), and a smaller part was phosphorylated to dTDP+dTTP. A high concentration of ATP in the medium (from ADP added at the start of the incubation) did not affect the import of dTMP. ATP did, however, decrease the degradation of dTMP both in the medium and in mitochondria and increased the phosphorylation of dTMP, but only in the medium. The finding that ATP did not increase intramitochondrial dTDP+dTTP suggests that phosphorylation of dTMP was an intramitochondrial process independent of a high external ATP concentration.
Import of dTMP into liver mitochondria. We incubated isolated mitochondria with 5 nM [3H]dTMP (45,000 cpm/pmol) with and without added 0.1 mM ADP (transformed to ATP by the mitochondria) as detailed for transport experiments in Materials and Methods. After either 2 or 10 min, mitochondria and external medium were separated by centrifugation and analyzed for the presence of labeled thymidine, dTMP, and dTDP+dTTP. (A) Import into mitochondria. (B) Remainder in external medium. Note the break in the ordinate in both panels. ■, dTDP+dTTP; ▴, dTMP; ●, thymidine. Solid lines are data from incubations without added ADP, and broken lines are data from incubations with ADP.
dTTP added at 5 nM to the incubation medium was not by itself imported into mitochondria (Fig. 3 A). The triphosphate was, however, degraded rapidly to dTDP, dTMP, and thymidine in the medium (Fig. 3 B), and both dTMP and thymidine were imported into mitochondria (Fig. 3 A). ATP decreased the degradation of dTTP, resulting in less dTMP and thymidine in the mitochondria.
Incubation of mitochondria with labeled dTTP. We incubated mitochondria for 2 and 10 min with [3H]dTTP (35,000 cpm/pmol) with and without added ADP and determined labeled thymidine compounds as described for Fig. 2. (A) Data from mitochondria. (B) Data from the external medium. No dTTP was found inside mitochondria, but dTMP derived from the breakdown of dTTP in the external medium had been imported. Note that B contains a break in the ordinate. ■,dTTP+dTTP; ▴, dTMP; ●, thymidine. Broken lines show results from the experiment with ADP.
In the experiments described so far a very low concentration of thymidine phosphates was added to the medium. Higher concentrations of external dTMP resulted in a large increase of the import with saturation >5 μM (Fig. 4). The mitochondria then also contained increasing amounts of thymidine formed by dephosphorylation of dTMP and also smaller amounts of dTDP+dTTP (Fig. 4). At 5 μM external dTMP the internal concentration of dTMP was ≈60 pmol/mg mitochondria, which translates into an intramitochondrial concentration of 60 μM, assuming a volume of 1.0 μl/mg mitochondria (12). This is a >10-fold increase in the concentration of the nucleotide. At lower, nonsaturating concentrations of external dTMP the increase is still larger. In the experiment depicted in Fig. 2 (5 nM external dTMP), the nucleotide became 100- to 200-fold concentrated.
Dependence of dTMP import on concentration of dTMP. We incubated mitochondria with increasing concentrations of [3H]dTMP (430 cpm/pmol) for 2 min and determined the amount of intramitochondrial dTMP (▴), thymidine (●), and dTTP+dTDP (■) from their radioactivity.
We found no inhibition of the import of 5 nM dTMP by a 100-fold excess of thymidine or of the monophosphates or triphosphates of the other canonical deoxynucleosides (not detailed here). A 100-fold excess of dUMP inhibited 30%. Thus, the transport system was highly specific for dTMP. In other experiments we tested the effects of 20 μM atractyloside (inhibitor of the ATP/ADP exchange), of 1 μg/ml oligomycin (inhibitor of the F1Fo ATP synthase), of 2 μM rotenone or 0.1 μg/ml antimycin A (inhibitors of the respiratory chain) or the SH-reagent N-ethylmaleimide (0.1 mM), but none had an inhibitory effect on dTMP import (data not shown). Also, omission of glutamate/malate had no effect. The inhibitors decreased or abolished the level of ATP inside mitochondria and the phosphorylation of the imported dTMP to dTTP. Thus, import of dTMP did not depend on a functional respiratory chain or on ATP hydrolysis. We found, however, a complete abolition of the import by 20 mM bathophenantroline and a decrease to 30% at 2 mM. Thiamine pyrophosphate (20 mM) inhibited by 50%.
Export of Thymidine Phosphates from Mitochondria.
As described above, mt thymidine phosphates can be labeled in vivo from [3H]thymidine and the isolated mitochondria then contain radioactive dTTP, dTDP, dTMP, and thymidine. After determination of the specific radioactivity of dTTP, we could calculate the amount of each component in pmol/mg mitochondria and found that each thymidine phosphate represented approximately one-third of the total thymidine phosphate pool. To determine the efflux of thymidine nucleotides, we incubated the labeled mitochondria and measured during a 20-min time period the concentration of each radioactive nucleotide inside mitochondria (Fig. 5 A) and in the surrounding medium (Fig. 5 B). Inside mitochondria dTTP increased rapidly and represented close to 50% of the total pool after 1 min and 60% after 10 min. This increase occurred at the expense of dTMP and thymidine, whereas dTDP maintained its original concentration (Fig. 5 A). These data confirm the activity of highly active mt dTMP (Zee-Fen Chang, personal communication) and dTDP (13) kinases. The rate of efflux of thymidine nucleotides was only moderate (Fig. 5 B) and not comparable to the rapid import of dTMP shown in Fig. 2. The largest component was thymidine, followed by dTDP, dTMP, and dTTP. A distinction among the three nucleotides cannot be made in this experiment. Interpretation of the data is complicated by degradation and/or phosphorylation of the nucleotides after they had left the mitochondria.
Export of radioactive thymidine and thymidine phosphates from mitochondria labeled in vivo from [3H]thymidine (20,000 cpm/pmol). We incubated the labeled mitochondria under standard conditions and determined the amount of radioactive dTTP (□), dTDP (■), dTMP (▴), and thymidine (●) remaining in mitochondria (A) or exported into the medium (B) between 1 and 20 min. The specific radioactivity of the thymidine phosphates was 2,200 cpm/pmol.
It seemed possible that the efflux of thymidine phosphates required an exchange with a nucleotide outside the mt membrane. To test this, we incubated prelabeled mitochondria with either nonlabeled dTMP or dTTP at concentrations ranging from 2 nM to 5 μM in the medium, with or without added 0.1 mM ADP, and determined the efflux of dTTP, dTDP, dTMP, and thymidine. dTTP added to the medium was partially degraded, and therefore also dTDP was present in the medium if required for exchange. In Fig. 6 we show averaged data from the experiments with 0.5 and 5 μM nonlabeled external nucleotides because the results at the two concentrations were very similar. Each row shows the efflux of an individual thymidine nucleotide after 2 min (open bars) or 10 min (filled bars), determined from its radioactivity. The four columns represent different incubation conditions. A comparison of the first three columns shows that an excess of neither dTMP nor dTTP in the medium affected the amount and distribution of radioactivity across the membrane. The main conclusion is therefore that efflux of thymidine phosphates did not depend on an exchange. Furthermore, dTDP was the most abundant radioactive species, in particular at the early time point. At the later time point both dTMP and dTTP increased whereas dTDP remained unchanged, indicating that dTDP was the primary translocated product but was further transformed to dTTP and dTMP in the medium. In the presence of ATP (fourth column) dTTP increased and dTMP decreased, supporting the idea that dTTP and dTMP are secondary products formed by transphosphorylation from dTDP in the medium.
Export of thymidine phosphates does not involve an exchange process. We incubated labeled mitochondria under standard conditions for 2 min (open staples) or 10 min (filled staples). First column, controls without further addition; second column, averaged data from two experiments with 0.5 and 5 μM external dTMP; third column, 0.5 and 5 μM external dTTP; fourth column, 0.5 and 5 μM dTTP plus 0.1 mM ADP. The data show the export in pmol of dTTP (first row), dTDP (second row), dTMP (third row), and thymidine (fourth row).
Discussion
We address two related questions concerning mt dNTP pools: (i) Is there a high asymmetry among the sizes of the four pools in animal organs, as reported by Song et al. (7)? If so, this would be different from the situation in cultured cells where the sizes of the four pools are quite similar. Also, (ii) is there a specific carrier mechanism for the import of one of the four dNTPs into mitochondria?
To answer the first question, we concentrated our experiments to a single organ, mouse liver, but we believe that the general conclusions drawn from our results are relevant also for other animal organs. The main question in our minds was whether the discrepancies in pool size between cultured cells and animal organs depend on an inherent difference or whether they depend on different conditions for the preparation of mitochondria. Cultured cells are cooled immediately before extraction of mitochondria and are not exposed to a period of anaerobiosis that is known to lead to degradation of ATP (8, 9) and may also affect dNTPs. Song et al. (7) used animals that were first anesthetized with diethyl ether and then killed by decapitation before preparing mitochondria after an unspecified time. The results shown in Fig. 1 and Tables 1–3 point out the crucial importance of the conditions for the preparation of liver mitochondria. The content of both ATP and each of the four rNTPs and dNTPs differed widely between various preparations. When in early experiments we combined livers from several mice, which considerably prolonged the time before homogenization, we recovered only small amounts of ATP and dNTPs but the content of AMP was high. In later experiments when we used only one liver and homogenized it within <1 min in a large volume of ice-cold buffer the yield of both ATP and dNTPs improved dramatically (Fig. 1). The increase occurred, at least in part, at the expense of monophosphates and diphosphates in the mt preparations. Our data underline the problems involved in the determination of dNTP pools in animal mitochondria and suggest that changes in pool sizes during the preparation of mitochondria may influence the final results. As evident from Fig. 1 also, our most recent analyses show some variations, and we cannot exclude some degradation of dNTPs also in our best preparation. We consider our data therefore to be only an approximation of the actual pool sizes in vivo. However, irrespective of the yield of ATP and dNTPs, we never did find the pool asymmetries reported by Song et al. (7). Instead, the proportions between the four dNTP pools from mouse liver were in good general agreement with earlier results from resting cultured cells.
Also, the reliability of the method for the determination of dNTP pools is of course important. The DNA polymerase assay used by both Song et al. (7) and us is fraught with many pitfalls. We confirmed by HPLC the correctness of our determinations of dCTP and dGTP by the polymerase assay (Exp. 1 in Table 3 and unpublished data). We could not use HPLC for dTTP and dATP because of interfering material in the mt extracts.
The second question concerned the transport mechanism involved in the communication between the cytosol and mitochondria. In earlier experiments we found a rapid exchange of thymidine nucleotides between the cytosol and mitochondria of cultured cells, suggesting the presence of a carrier system in the mt membrane. We could now directly demonstrate the existence of a mechanism highly specific for dTMP. During incubation of isolated liver mitochondria, we found that, at nM concentration, dTMP but not thymidine, dTDP, or dTTP was rapidly concentrated >100-fold from the surrounding medium into the mt matrix by a uniport mechanism. Other canonical nucleoside monophosphates or triphosphates did not compete with dTMP for import, nor did various respiratory inhibitors or atractyloside, an inhibitor of the ATP/ADP transporter, affect the import. Import of dTMP became saturated at external concentrations of dTMP of ≈1 μM. The specificity and high affinity of the carrier for dTMP suggest that we are dealing with a physiological import mechanism for thymidine nucleotides. Because accumulation of dTMP was not affected by energization it is unlikely that dTMP is transported in its anionic form. One possibility is that the nucleotide is transported in symport with H+ (or in the thermodynamically equivalent exchange with OH−), as is the case for the phosphate carrier.
Note that dTMP is the first thymidine nucleotide formed during de novo synthesis, whereas the other three deoxynucleotides are formed as diphosphates. The concentration of dTMP in the cytosol is extremely low compared with those of dTTP and dTDP (10), and an import of the monophosphate into mitochondria therefore requires a highly efficient carrier.
The movement of labeled thymidine nucleotides from mitochondria into the surrounding medium was a much slower process than the import of dTMP. The two processes appear to occur by different mechanisms, and the export may take place at the diphosphate level. However, the relevance of this result is uncertain, and further work is required on this point.
The molecular identity of the carrier(s) is not defined by the present work. Several candidates for an mt carrier of deoxynucleotides have been suggested by transport assays with bacterially overexpressed putative carrier proteins incorporated into proteoliposomes (14–17). None of them appears to have the high specificity and high sensitivity for dTMP found for the carrier described here, suggesting that the transport observed by us requires the discovery of an as yet unknown protein. On the other hand, both sensitivity and specificity may be different in artificial proteoliposomes and in the natural environment of the mt membrane. Hopefully further work will clarify also this point.
Materials and Methods
Materials.
We used 8-week-old male C57Bl/6J mice (Charles River Laboratories, Calco, Italy) over a period of 3 months. Radioactive compounds were from PerkinElmer (Monza, Italy) or Moravek (Brea, CA). Antimycin A, atractyloside, oligomycin, and rotenone were available in the laboratory. N-ethylmaleimide, bathophenantroline sulfonate, and thiamine pyrophosphate were from Sigma (St. Louis, MO).
Preparation of Mitochondria.
Our finally adapted method was a modification of that described by Costantini et al. (18) for the preparation of rat liver mitochondria. We killed one mouse that had been starved for 24 h by cervical dislocation, transferred the liver (1 g) immediately into 30 ml of ice-cold buffer A (220 mM mannitol/70 mM sucrose/5 mM Mops, pH 7.4/2 mM EGTA/0.2 mg/ml BSA), and cut the liver into small pieces. We did all work rapidly and at close to 2–4°C. We transferred the liver to an all-glass Potter homogenizer (loose pestle) in 7 ml of buffer A and homogenized it with two firm strokes. We increased the volume of the suspension to 40 ml and centrifuged it at 700 × g for 10 min to remove nuclei. Mitochondria were sedimented from the supernatant solution at 9,000 × g for 10 min, followed by washing with buffer A and a second centrifugation at 9,000 × g. We suspended the pellet in 0.2–0.4 ml of buffer A for further experiments. Each preparation of mitochondria amounted to between 25 and 35 mg of protein. The respiratory control of the mitochondria ranged from 3.4 to 8.0 with glutamate plus malate as substrates determined with a Clark oxygen electrode.
Pool Determinations.
We extracted the mt suspension with 2 ml of 60% methanol at −20°C for at least 1 h and after centrifugation heated the supernatant for 3 min in a boiling water bath. After additional centrifugation we evaporated the solution in a flash evaporator to dryness and dissolved the residue in 0.2 ml of water. Portions of this solution were processed in different ways. We measured the four dNTP pools and the specific radioactivity of dTTP by the DNA polymerase assay (19) as modified (3). We determined the ATP/ADP/AMP ratio and the dTTP/dTDP/dTMP ratio by HPLC on a C-18 column (Phenomenex, Torrance, CA) by isocratic elution (1–18 min at 0.5 ml/min, then 1 ml/min) with 0.2 M ammonium phosphate buffer (pH 3.5) for 35 min, followed by a 5-min linear gradient from ammonium phosphate to 30% methanol in ammonium phosphate and finally 15 min of 30% methanol/ammonium phosphate. Retention times were as follows: ATP, 14.0 min; ADP, 15.6 min; AMP, 24.8 min; dTTP, 17.8 min; dTDP, 20.6 min; dTMP, 30.6 min; thymidine, 40 min. Separation of all four rNTPs and dNTPs was made isocratically on a WAX column (PolyLC; Lab Service Analytica, Anzola Emilia, Italy) with 0.32 M potassium phosphate (pH 5) (20). Retention times were as follows: dTTP, 10.5 min; UTP, 12.9 min; dCTP, 17.6 min; CTP, 20.9 min; dATP, 25.2 min; ATP, 29.2 min; dGTP 53.1 min; GTP, 62.9 min. Peaks were identified from their position and from the ratio of the absorbances at 260 and 280 nm. In mt extracts, dATP was not well separated from the large ATP peak and an overlapping peak made determination of dTTP difficult. Nucleotides were quantified from the area on the chromatogram or, in transport experiments, by radioactivity. All values were normalized to 1 mg of mitochondria.
Transport Experiments.
We started the experiments by addition of the mt suspension to solutions on a shaker in a 25°C room containing buffer B (220 mM mannitol/70 mM sucrose/5 mM Mops, pH 7.4/0.1 mM EGTA/1 mM phosphate·Tris, pH 7.4/1 mM MgCl2/5 mM glutamate/2.5 mM malate/0.2 mg/ml BSA) and other reagents specified for each experiment in a final volume of 0.3 ml. After the indicated time we cooled the incubation mixtures in glass vessels in an ice bath for 3 min and centrifuged them at +4°C for 10 min at 19,000 × g. We carefully removed the supernatant for further analyses and washed the sediment twice with 1 ml of buffer A without suspension. In model experiments we found that this washing did not remove any labeled compounds from prelabeled mitochondria. We extracted the final sediment with 60% methanol and analyzed the amount of nucleotides as described above.
In Vivo Labeling of Thymidine Phosphates.
We injected i.p. one mouse with 0.5 mCi of [3H]thymidine (20,000 cpm/pmol) 10 min before removal of the liver. The mitochondria of the mouse contained thymidine phosphates with a specific radioactivity of 2,000–2,500 cpm/pmol. We used the mitochondria to measure efflux in transport experiments.
Acknowledgments
This work was supported by grants from the Italian Association for Cancer Research (to V.B.), Italian Telethon (GGP05001 to V.B.), and the Italian Ministry of Education and Research [to V.B. (Prin 2005) and P.B. (Prin 2004)].
Footnotes
- ‡To whom correspondence should be addressed. E-mail: reichard{at}bio.unipd.it
-
Author contributions: P.F., P.B., P.R., and V.B. designed research; P.F., L.N., and P.R. performed research; and P.B., P.R., and V.B. wrote the paper.
-
The authors declare no conflict of interest.
- Abbreviations:
- mt,
- mitochondrial;
- rNTP,
- ribonucleoside triphosphate.
- © 2006 by The National Academy of Sciences of the USA
