Logo of mbcLink to Publisher's site
Mol Biol Cell. 2003 Sep; 14(9): 3628–3635.
PMCID: PMC196555

New Insights into the Bioenergetics of Mitochondrial Disorders Using Intracellular ATP Reporters

Thomas Fox, Monitoring Editor


Mutations in mitochondrial DNA (mtDNA) cause impairment of ATP synthesis. It was hypothesized that high-energy compounds, such as ATP, are compartmentalized within cells and that different cell functions are sustained by different pools of ATP, some deriving from mitochondrial oxidative phosphorylation (OXPHOS) and others from glycolysis. Therefore, an OXPHOS dysfunction may affect different cell compartments to different extents. To address this issue, we have used recombinant forms of the ATP reporter luciferase localized in different cell compartments— the cytosol, the subplasma membrane region, the mitochondrial matrix, and the nucleus— of cells containing either wild-type or mutant mtDNA. We found that with glycolytic substrates, both wild-type and mutant cells were able to maintain adequate ATP supplies in all compartments. Conversely, with the OXPHOS substrate pyruvate ATP levels collapsed in all cell compartments of mutant cells. In wild-type cells normal levels of ATP were maintained with pyruvate in the cytosol and in the subplasma membrane region, but, surprisingly, they were reduced in the mitochondria and, to a greater extent, in the nucleus. The severe decrease in nuclear ATP content under “OXPHOS-only” conditions implies that depletion of nuclear ATP plays an important, and hitherto unappreciated, role in patients with mitochondrial dysfunction.


The concept that high-energy compounds are compartmentalized in cells was proposed more than 20 years ago (Erickson-Viitanen et al., 1982a blue right-pointing triangle, 1982b blue right-pointing triangle; Saks et al., 1994 blue right-pointing triangle). For example, it was suggested that in normal smooth muscle cells the contractile functions are supported by mitochondrial ATP derived from the respiratory chain and oxidative phosphorylation (OXPHOS), whereas the plasma membrane proton pumps are supported by ATP from anaerobic glycolysis (Ishida et al., 1994 blue right-pointing triangle). In addition, recent studies showed that the import of histones into the nuclei of neonatal cardiomyocytes is strictly dependent on a concerted interaction between mitochondrial ATP synthesis and the trafficking of high-energy phosphoryls (Dzeja et al., 2002 blue right-pointing triangle).

In a technical advance, targeted luciferase has been used as an ATP sensor to investigate the kinetics of the variation of ATP concentration beneath the plasma membrane, in the mitochondria, and in the cytosol of pancreatic β-cells in response to glucose stimulation (Kennedy et al., 1999 blue right-pointing triangle). These experiments demonstrated that in response to the administration of glucose and potassium, ATP levels increased in the plasma membrane of β-cells in concert with that in mitochondria, whereas cytosolic ATP showed only a transient increase. On the other hand, studies using the ATP-dependent potassium channel as an ATP sensor showed that in Xenopus oocytes and in cultured mammalian cells there was no gradient between bulk cytosolic ATP and subplasma membrane ATP, suggesting that ATP diffuses freely between these two cell compartments (Gribble et al., 2000 blue right-pointing triangle).

Despite a growing body of evidence that high-energy molecules such as ATP and phosphocreatine may be compartmentalized in cells, we still do not fully appreciate what modifications occur in intracellular ATP pools of cells affected by defects in energy metabolism, such as those caused by mutations in mitochondrial DNA (mtDNA). These mutations, which are associated with a heterogeneous group of sporadic or maternally inherited metabolic disorders (DiMauro et al., 1998 blue right-pointing triangle), cause impairment of OXPHOS, resulting in reduced mitochondrial ATP synthesis.

The question of which cell compartments are more prone to become depleted of ATP, and whether specific ATP-dependent functions are affected differentially in mitochondrial diseases, remains essentially unresolved. Clearly, a better understanding of the mechanisms by which cells cope with impaired mitochondrial ATP synthesis, and which specific ATP-dependent cellular functions are preserved, down-regulated, or abolished in these conditions, could help us better understand the pathogenesis of mitochondrial diseases.

To begin to address those issues, we have investigated the fate of intracellular ATP pools in intact living cells harboring pathogenic mtDNA mutations. We used targeted luciferase constructs to assess free ATP levels in the cytosolic, mitochondrial, intranuclear, and subplasma membrane compartments, an approach that has already been used successfully by other investigators to detect ATP in living normal cells (Maechler et al., 1998 blue right-pointing triangle; Jouaville et al., 1999 blue right-pointing triangle; Kennedy et al., 1999 blue right-pointing triangle; Porcelli et al., 2001 blue right-pointing triangle).

We chose to analyze two known pathogenic mutations. The first one is a T→G transversion at position 8993 in the ATPase 6 gene (Anderson et al., 1981 blue right-pointing triangle), encoding a subunit of the F0 portion of ATP synthase. This mutation is associated with two related syndromes, NARP (neuropathy, ataxia, and retinitis pigmentosa; Holt et al., 1990 blue right-pointing triangle) and MILS (maternally inherited Leigh syndrome; Santorelli et al., 1993 blue right-pointing triangle). Importantly, in cytoplasmic hybrid (cybrid) cells containing homoplasmic levels of the mutation, mitochondrial ATP synthesis is reduced by ∼50–80% (Manfredi et al., 2002b blue right-pointing triangle), but the remainder of the respiratory chain is relatively unaffected. The second mutation is an A→G transition at position 3243 in the gene encoding tRNALeu(UUR). This mutation is associated with MELAS (mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes; Goto et al., 1990 blue right-pointing triangle). The mutation results in globally impaired mitochondrial protein synthesis and affects all respiratory chain complexes containing mtDNA-encoded subunits (i.e., complexes I, III, IV, and V; King et al., 1992 blue right-pointing triangle). We used cells harboring wild-type mtDNA (i.e., without known pathogenic mutations) as a positive control, and ρ0 cells, which are devoid of mtDNA and therefore have no residual OXPHOS function (King and Attardi, 1989 blue right-pointing triangle), as a negative control.


Luciferase Constructs

Wild-type firefly luciferase from Photinus pyralis (accession number AB062786) is a 550-amino acid protein containing the canonical peroxisomal targeting sequence Ser-Lys-Leu at its C terminus (Gould et al., 1987 blue right-pointing triangle). To generate cytosolic luciferase (Lucc), we replaced the leucine at position 550 with a valine by in vitro mutagenesis, thereby destroying the natural peroxisomal targeting sequence (Gould et al., 1989 blue right-pointing triangle) and allowing this modified luciferase to reside free in the cytoplasm. For the mitochondrial luciferase (Lucm), Lucc was fused downstream of the sequence encoding the mitochondrial targeting signal (MTS) of subunit VIII of cytochrome c oxidase (accession number NM004074). The COX VIII targeting sequence has been used successfully in other mitochondrial targeting experiments (Rizzuto et al., 1992 blue right-pointing triangle; Rizzuto et al., 1995 blue right-pointing triangle), and a similar COX VIII-luciferase construct was shown to be targeted appropriately to the mitochondrial matrix (Jouaville et al., 1999 blue right-pointing triangle; Kennedy et al., 1999 blue right-pointing triangle; Porcelli et al., 2001 blue right-pointing triangle). For the subplasma membrane luciferase (Lucpm), Lucc was fused downstream of the second external arm of the Ht1a receptor (accession number X57829), a G-linked protein; the Ht1a sequence was shown to target a reporter gene to the plasma membrane of eukaryotic cells (Singer 1990 blue right-pointing triangle). For the nuclear luciferase (Lucn), we fused Lucc downstream of the SV-40 large T antigen (accession number NC001669) nuclear localization signal (NLS); this NLS was shown to target recombinant luciferase to the nucleus (Michels et al., 1995 blue right-pointing triangle). All the chimeric Luc constructs were inserted into the mammalian expression vector pCDNA3.0 (Invitrogen Inc., Carlsbad, CA), as described (Manfredi et al., 2002b blue right-pointing triangle).

Cell Culture and Transfection

We used four human cell lines: 1) 143B osteosarcoma cells containing wild-type mtDNA (i.e., ρ+ cells); 2) 143B206 cells that were completely devoid of mtDNA, due to long-term treatment of the 143B cells with ethidium bromide (i.e., ρ0 cells; King and Attardi, 1989 blue right-pointing triangle); 3) transmitochondrial cell hybrids (cybrids) consisting of 143B206 cells repopulated with mitochondria containing 100% 8993-G mtDNA (i.e., homoplasmic mutation in ATPase 6) from a patient with NARP (Manfredi et al., 1999 blue right-pointing triangle), and 4) cybrids of 143B206 cells repopulated with mitochondria containing 100% 3243-G mtDNA (i.e., homoplasmic mutation in tRNALeu(UUR)) from a patient with MELAS (King et al., 1992 blue right-pointing triangle). Cells were transfected with pCDNA3.0 containing the engineered Luc genes, using the transfection reagent FuGENE6 (Roche Applied Sciences, Indianapolis, IN) as described by the manufacturer. Cells were grown in 100-mm culture dishes in DMEM containing high glucose (4.5 mg/ml), 2 mM l-glutamine, 110 mg/l sodium pyruvate supplemented with 10% fetal bovine serum (FBS), and 50 μg/ml uridine. For stable transfections, osteosarcoma-derived cells were selected in 500 μg/ml of the neomycin analog Geneticin (Invitrogen Inc.).

Intracellular localization of the chimeric luciferases was assayed by immunocytochemistry in transiently transfected human 293T-HEK cells as described (Kennedy et al., 1999 blue right-pointing triangle), using rabbit polyclonal antiluciferase antibodies (Promega, Madison, WI) followed by antirabbit Cy-2–conjugated secondary antibodies (Molecular Probes, Eugene, OR). The intracellular localization of the luciferase constructs was tested by staining transfected cells with cell compartment-specific fluorescent dyes (Molecular Probes): Mitotracker Red CMX-ROS was used for mitochondria, FM 4–64 for the plasma membrane, and Hoechst 33342 for the nuclei.

Calibration of the Luciferase System

To assess the range of luciferase expression in transfected cells, the luminescence derived from 2 × 105 cells was measured with a luciferase assay kit (Promega) and was compared with that obtained from a set of purified firefly luciferase standards (Roche Applied Sciences), ranging from 1 pg to 1 ng.

To determine the dependence of light production on luciferin concentration, we constructed a luciferin dose-response curve using a fixed amount of luciferase. Typically, 100 pg, a value comparable to that found in lysates of 2 × 105 stably transfected cells, was dissolved in a 25 mM tricine, 150 mM NaCl buffer, pH 7.4, in the presence of increasing ATP concentrations, ranging from 31.2 to 500 μM.

To ascertain the kinetic properties of the targeted luciferases, we constructed ATP-luminescence curves both with purified firefly luciferase and with lysates from cells expressing the different targeted luciferases. Cells (2 × 105) were solubilized in 20 μl of a lysis buffer containing 20 mM HEPES and 0.1% Triton X-100, pH 7.2. Purified firefly luciferase (100 pg) was also dissolved in 20 μl of the same lysis buffer. Luminescence was measured on the soluble cell fractions diluted in 80 μl of a buffer that mimicked physiological intracellular conditions (physiological buffer), containing 20 mM HEPES, 140 mM KCl, 10.2 mM EGTA, 6.7 mM CaCl2, 2 mM luciferin, 20 μg/ml digitonin, 0.5 mM MgCl2, and 0.01 mM CoA, as described (Jouaville et al., 1999 blue right-pointing triangle; Kennedy et al., 1999 blue right-pointing triangle) and increasing concentrations of ATP ranging from 0 to 1000 μM. To correct for variable luciferase expression levels, luminescence was normalized to the “total potential luminescence” measured in lysates from an identically sized aliquot of transfected cells, diluted in 50 μl of the luciferase assay buffer (Promega) in the presence of excess (>1 mM) ATP.

Measurements of Cellular ATP

For luciferase assays, aliquots of 2 × 105 transfected cells were incubated in DMEM containing the following combinations of substrates and inhibitors (20 replicate samples for each condition): 1) glucose (4.5 mg/ml) plus 110 mg/l pyruvate; 2) glucose plus 1 μg/ml the ATP synthase inhibitor oligomycin; 3) pyruvate alone (110 mg/l); 4) pyruvate plus 5 ng/ml oligomycin; and 5) pyruvate plus 10 ng/ml oligomycin. The ρ0 cells were also assayed with glucose plus 1 μg/ml oligomycin in the presence of 110 mg/l pyruvate. Luminescence was measured as described (Manfredi et al., 2002b blue right-pointing triangle). Briefly, cells were placed in noncoated 24-well plastic plates and incubated with gentle rocking (to prevent cell attachment) for 1 h at 37°C in 5% CO2. Cells were collected by aspiration, pelleted immediately by centrifugation, and resuspended in 90 μl buffer containing 25 mM tricine and 150 mM NaCl, pH 7.4. Beetle luciferin (Promega) was added to the cell suspension (final concentration, 2 mM), and light emission was measured in a luminometer (MGM Instruments, Camden, CT) at 5-s intervals until the maximum value of luminescence was reached. To correct for the variability of luciferase expression in each sample, the relative luminescence values in each cell compartment were normalized to the “total potential luminescence” as described above.

To assess the total ATP content in cells incubated in the same conditions as above, we used an HPLC-based method (Manfredi et al., 2002b blue right-pointing triangle), using appropriate ATP standards. ATP content measured in total cell lysates was expressed as nmoles/mg cellular proteins. Proteins were measured with a DC protein assay kit (Bio-Rad) as recommended by the manufacturer.

Statistical analyses of differences between ATP concentration values were performed by unpaired two-tailed Student's t test.


Expression of Targeted Luciferase Constructs

We obtained light production in lysates of cells transfected with all four luciferase constructs, whereas the mock transfection of pCDNA3.0 without insert gave a negative assay.

To verify the correct subcellular localization of the chimeric luciferases, the transfected 293T-HEK cells were treated with specific fluorescent dyes, which stained the intracellular compartment of interest, and then were immunostained with antiluciferase antibodies. The overlay of images of the same field of cells photographed using appropriate light filters demonstrated a good degree of colocalization of the fluorochromes in cells transfected with Lucm, Lucpm, and Lucn (Figure 1, B–D). Cells transfected with Lucc showed a diffused staining compatible with cytosolic localization (Figure 1A).

Figure 1.
Subcellular localization of recombinant chimeric luciferases in 293T cells. Maps of the engineered luciferase genes are shown at the top. (A) Cells transfected with Lucc and immunostained with an antiluciferase antibody (green), showing a cytosolic, diffuse ...

Kinetics of Chimeric Luciferases and Calibration of the System

A standard dilution curve of firefly luciferase plotted vs. luminescence showed that light production was linearly proportional to the amount of luciferase added (Figure 2A).On the bais of this luciferase standard curve, we estimated that aliquots of 2 × 105 stably or transiently transfected cells typically expressed 50–150 pg of luciferase. Therefore, the kinetic properties of the chimeric luciferases were compared with those of 100 pg of firefly luciferase.

Figure 2.
Calibration of the luciferase ATP reporter system. (A) Luciferase titration curve. Light emission was measured with increasing amounts of luciferase. Light emission was directly proportional to the amount of luciferase added, within the limits of detection ...

A luciferin standard curve (Figure 2B) was obtained using a fixed amount of firefly luciferase in the presence of increasing concentrations of luciferin (between 100 μM and 3 mM) and ATP (between 31.2 and 500 μM). Maximal light emission was reached for all luciferin concentrations with 500 μM ATP; the absolute luminescence values were highest with 2 mM luciferin. This suggested that luciferin concentrations between 1 and 2 mM did not inhibit luciferase, whereas lower concentrations (i.e., <0.5 mM) were unable to achieve maximum luminescence in the range of luciferase present in transfected cells. Furthermore, in prokaryotes, free diffusion of luciferin inside cells at physiological pH is limited by its negative charge (Di Tomaso et al., 2001 blue right-pointing triangle). Therefore, it is possible that also in mammalian cells the intracellular concentration of luciferin may be lower than that in the buffer. For these reasons, we used 2 mM luciferin for all our assays.

To measure luciferase activity in an environment that closely reflects the intracellular one, the apparent Km values for ATP of the chimeric luciferases were determined in a buffer that mimicked physiological intracellular conditions and compared with that of purified firefly luciferase. In this buffer, the apparent Km of native firefly luciferase (Figure 2C) was ∼70 μM ATP, which was slightly above the values previously reported for Photinus pyralis luciferase (∼60 μM ATP; DeLuca and McElroy 1974 blue right-pointing triangle; Lemasters and Hackenbrock 1977 blue right-pointing triangle). The Km values of the four chimeric luciferases in soluble cell fractions of transfected cells were all higher than that of firefly luciferase. Whereas Lucm, Lucc, and Lucpm were very similar to each other, Lucn had a substantially higher Km than the other three (Figure 2C). These kinetic differences were presumably due to conformational changes resulting from the addition of the targeting peptides. The polynomial equations (unpublished data) describing the curves in Figure 2C were used to extrapolate the absolute ATP levels detected by targeted luciferases in intact cells.

Total Cellular ATP Content in Wild-type and Mutant Cells

We measured total cellular ATP content by high-pressure liquid chromatography (HPLC) of cell lysates from wild-type and mutant (i.e., NARP, MELAS, and ρ0) cells after incubation in medium containing different combinations of substrates and inhibitors (Figure 3). We first measured total ATP levels in medium containing both glucose and pyruvate (i.e., substrates for glycolysis and OXPHOS, respectively). In this “complete” medium, all cell types had similar ATP contents, ranging from 17 to 26 nmol/mg cell protein.

Figure 3.
Quantification of total cellular ATP. Wild-type, NARP, MELAS, and ρ0 cells were incubated for 2 h in medium containing the indicated combinations of substrate and inhibitors. Cells were harvested and lysed immediately in perchloric acid. Total ...

In “glycolysis-only” medium (containing glucose but lacking pyruvate, and containing 1 μg/ml oligomycin, a concentration that completely inhibits mitochondrial ATP synthase), not only the wild-type cells, but also the NARP and MELAS cells maintained almost unchanged ATP levels, suggesting that glycolysis was sufficient to provide the required cellular ATP. However, ρ0 cells, which rely exclusively on glycolysis for ATP synthesis (as they do not possess a functioning respiratory chain) showed a statistically significant reduction in total ATP content in glucose alone compared with complete medium (41% reduction; p < 2 × 106). Because these cells also lack the oligomycin-binding subunit 6 of ATPase, which is mtDNA encoded, this reduction cannot be attributed to the effect of oligomycin. One possible explanation is that in cells lacking a functioning complex I (NADH dehydrogenase-CoQ reductase), NADH derived from glycolysis can only be reoxidized by lactate dehydrogenase, using pyruvate as substrate. Therefore, in the absence of supplemented pyruvate, lactate dehydrogenase may function more slowly and NADH may accumulate and partially inhibit glycolysis. In support of this hypothesis, ATP levels were restored to 85% of the value in complete medium when pyruvate was added back to the glycolysis-only medium (our unpublished results).

In “OXPHOS-only” medium containing pyruvate as the sole energy substrate (i.e., cells were forced to utilize OXPHOS to generate ATP), there was a decline in total ATP levels in all four cell lines (Figure 3). However, there was a 45% decline in the ATP content of wild-type cells compared with the value in complete medium, whereas NARP cells, which have a partial defect in mitochondrial ATP synthesis, showed a more marked (66%) reduction in total ATP. This difference between wild-type and NARP cells in OXPHOS-only medium was statistically significant (p < 0.009). Both MELAS cells, which have a more severe respiratory chain defect (due to compromised translation of mtDNA-encoded polypeptides; King et al., 1992 blue right-pointing triangle), and ρ0 cells showed an almost complete loss of measurable ATP (≤1 nmol/mg protein).

To elicit a more severe ATP loss in NARP cells, which have a milder mitochondrial ATP synthesis defect than do MELAS cells, we used pyruvate medium containing the ATPase inhibitor oligomycin. We had previously demonstrated that low doses of oligomycin substantially decrease mitochondrial ATP synthesis in NARP, but not in wild-type, cells (Manfredi et al., 2002a blue right-pointing triangle). As expected, the addition of oligomycin did not significantly affect ATP levels in wild-type cells (Figure 3), whereas NARP cells showed further statistically significant decreases (53% [p < 5 × 106] and 76% [p < 2 × 106], in 5 and 10 ng/ml oligomycin, respectively) compared with wild-type cells treated with oligomycin. This showed that NARP mutant cells have reduced capability to maintain ATP levels when the OXPHOS machinery is put under pressure by limiting doses of inhibitors or, presumably, by increased energy demand.

Subcellular ATP Concentration in Wild-type and Mutant Cells

We measured subcellular ATP concentrations in cells incubated in different substrates, as described above. We first looked at cytosolic ATP in cells transfected with Lucc (Figure 4A). ATP concentration in the cytosol of all cell types reflected closely the trend that we observed in total cell lysates. In particular, maximal ATP concentrations (between 91 and 110 μM) were obtained in complete medium, and these values remained stable in glycolysis-only medium. In OXPHOS-only medium, wild-type cells lost 29% of their maximal ATP levels, whereas NARP cells showed a significantly greater loss of ATP compared with wild-type cells (62%; p < 0.007). MELAS and ρ0 cells showed an almost complete loss of cytosolic ATP (i.e., <1 μM).

Figure 4.
ATP concentration in intracellular compartments. Wild-type, NARP, MELAS, and ρ0 cells expressing luciferases targeted to the indicated cell compartment (A–D) were incubated in medium containing the indicated combinations of substrate and ...

We then looked at cells transfected with Lucpm (Figure 4B). ATP concentrations in the subplasma membrane region were similar to those of cytosolic ATP, suggesting that there was no real ATP “gradient” between these two cell compartments. These results are in agreement with those obtained using pancreatic β cells (Jouaville et al., 1999 blue right-pointing triangle) and cardiomyocytes (Dzeja et al., 2002 blue right-pointing triangle), where no difference in ATP concentration was found between the cytosol and the plasma membrane.

Next, we measured ATP concentrations in the mitochondrial matrix using the Lucm construct (Figure 4C). In complete medium, wild-type cells showed a high matrix ATP content (223 μM) that was more than double the level observed in the cytosol (Figure 4A). On the other hand, MELAS, NARP, and ρ0 cells all had ATP levels (101–112 μM) that were essentially identical to the respective cytosolic concentrations. This “equilibration” of ATP concentrations between the mitochondrial and cytosolic compartments in these cells was presumably the consequence of the fact that in OXPHOS-defective cells, ATP is imported from the cytosolic compartment into the mitochondrial matrix (Buchet and Godinot 1998 blue right-pointing triangle; Loiseau et al., 2002 blue right-pointing triangle). This interpretation is consistent with the observation that in wild-type cells in glycolysis-only medium, the ATP concentration in the mitochondrial matrix (86 μM) was reduced to a value comparable to that of cytosolic ATP (91 μM), because in this condition all mitochondrial ATP must be imported from the cytosol.

In ρ0 cells grown in glycolysis-only medium, mitochondrial ATP was significantly decreased (34 μM) compared with that in wild-type cells (86 μM; p < 0.002). This was unexpected, because ρ0 cells produce most of their ATP through glycolysis. However, as discussed above for total cellular ATP, this reduction was probably due to diminished glycolytic activity resulting from the lack of exogenous pyruvate. In fact, when pyruvate was added back to the glycolysis-only medium, ρ0 cells were able to restore their mitochondrial matrix ATP concentration to 87% of the value in complete medium (our unpublished results).

In OXPHOS-only medium, MELAS and ρ0 cells had virtually no detectable mitochondrial ATP, confirming that in these cells all the mitochondrial ATP is of glycolytic origin. In NARP cells, which have reduced mitochondrial ATP synthesis (Manfredi et al., 2002a blue right-pointing triangle), ATP in the matrix (30 μM) was only partially decreased compared with wild-type cells (55 μM; p < 0.0003).

Finally, we analyzed nuclear ATP content with Lucn (Figure 4D). In complete medium, nuclear ATP levels were similar in all cell lines and were only moderately decreased in MELAS and ρ0 cells in glycolysis-only medium, presumably because of the lack of pyruvate. In glycolysis-only medium, ρ0 cells showed a significant decrease of ATP concentration compared with complete medium (p < 0.004), albeit less marked than in the mitochondrial compartment. In OXPHOS-only medium there was a striking decrease in nuclear ATP. Surprisingly, wild-type cells lost ∼79% of nuclear ATP in OXPHOS-only medium compared with the level in complete medium (p < 2 × 105). ATP was undetectable in MELAS and ρ0 cells, and in NARP cells there was a marked ATP decrease to 11% of the value measured in complete medium (p < 1 × 106). This value was significantly lower than in wild-type cells (p < 3 × 106). As expected, the addition of low doses of oligomycin to the OXPHOS-only medium almost completely eliminated nuclear ATP in all cell lines.


We have shown that targeted luciferase can be used to assess the degree of ATP compartmentalization inside cells. The apparent Km for ATP in vitro has been reported to be 63 μM (DeLuca and McElroy, 1974 blue right-pointing triangle; Lemasters and Hackenbrock, 1977 blue right-pointing triangle), but it was suggested previously that, when measured in an intracellular milieu, the kinetic properties of luciferase are different from those found in vitro (Gandelman et al., 1994 blue right-pointing triangle; Maechler et al., 1998 blue right-pointing triangle; Jouaville et al., 1999 blue right-pointing triangle; Kennedy et al., 1999 blue right-pointing triangle). In our system, the kinetic properties of the targeted luciferases were dissimilar from that of native firefly luciferase as well as from each other, suggesting that the addition of leader peptides somehow affect their biochemical properties. Therefore, ATP standard curves were generated for each construct in order to determine ATP concentrations in the cytosol, subplasma membrane region, mitochondria, and nucleus with some confidence. In principle, our results suggest that a targeted luciferase approach could also be used to measure ATP concentrations in other cell compartments. These could be compartments found ubiquitously in all cells, such as the endoplasmic reticulum and the nuclear envelope, or could be compartments specific to certain cell types, such as secretory vesicles, synaptic buttons, and the contractile apparatus of myocytes.

We have investigated here the fate of intracellular free ATP pools (or more precisely, ATP available to react in a luciferin-luciferase assay) in intact wild-type cells and in cells harboring pathogenic mtDNA mutations grown under conditions in which they used preferentially either glycolytic or oxidative metabolism for ATP synthesis. We obtained evidence that ATP distribution is not uniform throughout the cell compartments that we analyzed, suggesting that the partition of high energy phosphorylated molecules in cells does not occur simply by diffusion of ATP along concentration gradients. Clearly, each ATP pool in the cell contributes to different extents to the overall cellular ATP pool. Here, we have looked at four cell compartments and found that the changes in total cellular ATP observed in different metabolic conditions (Figure 3) were reflected in some but not in all cell compartments. For example, the cytosolic ATP pool (Figure 4A), which presumably is the largest one in the cell and contains the majority of hydrolysable ATP, was the compartment that showed a better correlation, albeit not a perfect one, with the total cellular ATP pool. On the other hand, the nuclear pool (Figure 4D), ostensibly much smaller than the cytosolic one, showed a rather different behavior. These differences underscore the importance of developing tools such as targeted luciferases to investigate ATP concentrations within specific cellular compartments.

In healthy cells grown in “rich” conditions, when both glycolysis in the cytosol and OXPHOS in the mitochondria were active, the ATP concentration in the mitochondrial matrix was approximately double that in other cell compartments (Figure 4). In the same conditions, cells with impaired OXPHOS had apparently equilibrated their ATP concentrations among all the four cell compartments studied (including, notably, their mitochondria), suggesting that in these cells the bulk of ATP was generated by glycolysis and imported into mitochondria.

Under oxidative conditions, ATP levels were markedly decreased in all cell compartments of NARP cells, and even more so in MELAS cells. Conversely, wild-type cells were able to maintain ATP levels comparable to those measured in complete medium in the cytosol and in the subplasma membrane region, but showed reduced ATP levels in the mitochondria and, to an even greater extent, in the nucleus. These findings suggest that, even in normal cells, when mitochondrial OXPHOS is the sole source of high-energy phosphoryls, ATP is exported preferentially from the mitochondria to supply energy to other energy-intensive compartments, such as cytosol and subplasma membrane.

In oxidative conditions, particularly when ATP supplies in the cell become limited (e.g., after the addition of low doses of oligomycin), the nucleus appears to be the compartment most susceptible to ATP depletion, not only in mutant but also in wild-type cells. These findings are particularly relevant in light of those of Dzeja and colleagues (Dzeja et al., 2002 blue right-pointing triangle), who reported that the import of histone proteins into cardiomyocyte nuclei was dependent mainly on mitochondrial ATP in concert with phosphotransfer enzymes, such as adenylate kinase and creatine kinase.

The decrease in nuclear ATP content under OXPHOS-only conditions implies that depletion of nuclear ATP plays an important, and hitherto unappreciated, role in patients with oxidative energy dysfunction. In these patients, the output of ATP from mitochondria may become critically low, particularly in cells with higher metabolic demand and higher oxidative rates, such as myocytes and neurons, which are frequently the most affected by mitochondrial disorders. Loss of ATP in the nucleus of these cells may have potentially severe consequences, resulting in the impairment of crucial cellular functions such as nuclear import, DNA transcription, and replication.


The authors thank Dr. Michael Lin for his invaluable assistance with the manuscript. The COX8 gene was a generous gift of Dr. Rosario Rizzuto, University of Ferrara, Italy. This work was supported by grants from the U.S. National Institutes of Health (NS02179[G.M.], NS28828, NS39854, and HD32062 [E.A.S.]) and the Muscular Dystrophy Association (E.A.S., G.M).


Article published online ahead of print. Mol. Biol. Cell 10.1091/mbc.E02–12–0796. Article and publication date are available at www.molbiolcell.org/cgi/doi/10.1091/mbc.E02-12-0796.


  • Anderson, S. et al. (1981). Sequence and organization of the human mitochondrial genome. Nature 290, 457–465. [PubMed]
  • Buchet, K., and Godinot, C. (1998). Functional F1-ATPase essential in maintaining growth and membrane potential of human mitochondrial DNA-depleted rho degrees cells. J. Biol. Chem. 273, 22983–22989. [PubMed]
  • DeLuca, M., and McElroy, W.D. (1974). Kinetics of the firefly luciferase catalyzed reactions. Biochemistry 13, 921–925. [PubMed]
  • Di Tomaso, G., Borghese, R., and Zannoni, D. (2001). Assay of ATP in intact cells of the facultative phototroph Rhodobacter capsulatus expressing recombinant firefly luciferase. Arch. Microbiol. 177, 11–19. [PubMed]
  • DiMauro, S., Bonilla, E., Davidson, M., Hirano, M., and Schon, E.A. (1998). Mitochondria in neuromuscular disorders. Biochim. Biophys. Acta 1366, 199–210. [PubMed]
  • Dzeja, P.P., Bortolon, R., Perez-Terzic, C., Holmuhamedov, E.L., and Terzic, A. (2002). Energetic communication between mitochondria and nucleus directed by catalyzed phosphotransfer. Proc. Natl. Acad. Sci. USA 15, 15. [PMC free article] [PubMed]
  • Erickson-Viitanen, S., Geiger, P.J., Viitanen, P., and Bessman, S.P. (1982a). Compartmentation of mitochondrial creatine phosphokinase. II. The importance of the outer mitochondrial membrane for mitochondrial compartmentation. J. Biol. Chem. 257, 14405–14411. [PubMed]
  • Erickson-Viitanen, S., Viitanen, P., Geiger, P.J., Yang, W.C., and Bessman, S.P. (1982b). Compartmentation of mitochondrial creatine phosphokinase. I. Direct demonstration of compartmentation with the use of labeled precursors. J. Biol. Chem. 257, 14395–14404. [PubMed]
  • Gandelman, O., Allue, I., Bowers, K., and Cobbold, P. (1994). Cytoplasmic factors that affect the intensity and stability of bioluminescence from firefly luciferase in living mammalian cells. J. Biolumin. Chemilumin. 9, 363–371. [PubMed]
  • Goto Y.-i., Nonaka, I., and Horai, S. (1990). A mutation in the tRNALeu(UUR) gene associated with the MELAS subgroup of mitochondrial encephalomyopathies. Nature 348, 651–653. [PubMed]
  • Gould, S.G., Keller, G.A., and Subramani, S. (1987). Identification of a peroxisomal targeting signal at the carboxy terminus of firefly luciferase. J. Cell Biol. 105, 2923–2931. [PMC free article] [PubMed]
  • Gould, S.J., Keller, G.A., Hosken, N., Wilkinson, J., and Subramani, S. (1989). A conserved tripeptide sorts proteins to peroxisomes. J. Cell Biol. 108, 1657–1664. [PMC free article] [PubMed]
  • Gribble, F.M., Loussouarn, G., Tucker, S.J., Zhao, C., Nichols, C.G., and Ashcroft, F.M. (2000). A novel method for measurement of submembrane ATP concentration. J. Biol. Chem. 275, 30046–30049. [PubMed]
  • Holt, I.J., Harding, A.E., Petty, R.K., and Morgan-Hughes, J.A. (1990). A new mitochondrial disease associated with mitochondrial DNA heteroplasmy. Am. J. Hum. Genet. 46, 428–433. [PMC free article] [PubMed]
  • Ishida, Y., Riesinger, I., Wallimann, T., and Paul, R.J. (1994). Compartmentation of ATP synthesis and utilization in smooth muscle: roles of aerobic glycolysis and creatine kinase. Mol. Cell Biochem. 133–134, 39–50. [PubMed]
  • Jouaville, L.S., Pinton, P., Bastianutto, C., Rutter, G.A., and Rizzuto, R. (1999). Regulation of mitochondrial ATP synthesis by calcium: evidence for a long-term metabolic priming. Proc. Natl. Acad. Sci. USA 96, 13807–13812. [PMC free article] [PubMed]
  • Kennedy, H.J., Pouli, A.E., Ainscow, E.K., Jouaville, L.S., Rizzuto, R., and Rutter, G.A. (1999). Glucose generates sub-plasma membrane ATP microdomains in single islet beta-cells. Potential role for strategically located mitochondria. J. Biol. Chem. 274, 13281–13291. [PubMed]
  • King, M.P., and Attardi, G. (1989). Human cells lacking mtDNA: repopulation with exogenous mitochondria by complementation. Science 246, 500–503. [PubMed]
  • King, M.P., Koga, Y., Davidson, M., and Schon, E.A. (1992). Defects in mitochondrial protein synthesis and respiratory chain activity segregate with the tRNA(Leu(UUR)) mutation associated with mitochondrial myopathy, encephalopathy, lactic acidosis, and stroke-like episodes. Mol. Cell. Biol. 12, 480–490. [PMC free article] [PubMed]
  • Lemasters, J.J., and Hackenbrock, C.R. (1977). Kinetics of product inhibition during firefly luciferase luminescence. Biochemistry 16, 445–447. [PubMed]
  • Loiseau, D., Chevrollier, A., Douay, O., Vavasseur, F., Renier, G., Reynier, P., Malthiery, Y., and Stepien, G. (2002). Oxygen consumption and expression of the adenine nucleotide translocator in cells lacking mitochondrial DNA. Exp. Cell Res. 278, 12–18. [PubMed]
  • Maechler, P., Wang, H., and Wollheim, C.B. (1998). Continuous monitoring of ATP levels in living insulin secreting cells expressing cytosolic firefly luciferase. FEBS Lett. 422, 328–332. [PubMed]
  • Manfredi, G., Fu, J., Ojaimi, J., Sadlock, J.E., Kwong, J.Q., Guy, J., and Schon, E.A. (2002a). Rescue of a deficiency in ATP synthesis by transfer of MTATP6, a mitochondrial DNA-encoded gene, to the nucleus. Nat. Genet. 25, 25. [PubMed]
  • Manfredi, G., Gupta, N., Vazquez-Memije, M.E., Sadlock, J.E., Spinazzola, A., De Vivo, D.C., and Schon, E.A. (1999). Oligomycin induces a decrease in the cellular content of a pathogenic mutation in the human mitochondrial ATPase 6 gene. J. Biol. Chem. 274, 9386–9391. [PubMed]
  • Manfredi, G., Yang, L., Gajewski, C.D., and Mattiazzi, M. (2002b). Measurements of ATP in mammalian cells. Methods 26, 317–326. [PubMed]
  • Michels, A.A., Nguyen, V.T., Konings, A.W., Kampinga, H.H., and Bensaude, O. (1995). Thermostability of a nuclear-targeted luciferase expressed in mammalian cells. Destabilizing influence of the intranuclear microenvironment. Eur. J. Biochem. 234, 382–389. [PubMed]
  • Porcelli, A.M., Pinton, P., Ainscow, E.K., Chiesa, A., Rugolo, M., Rutter, G.A., and Rizzuto, R. (2001). Targeting of reporter molecules to mitochondria to measure calcium, ATP, and pH. Methods Cell Biol. 65, 353–380. [PubMed]
  • Rizzuto, R., Brini, M., Pizzo, P., Murgia, M., and Pozzan, T. (1995). Chimeric green fluorescent protein as a tool for visualizing subcellular organelles in living cells. Curr. Biol. 5, 635–642. [PubMed]
  • Rizzuto, R., Simpson, A.W., Brini, M., and Pozzan, T. (1992). Rapid changes of mitochondrial Ca2+ revealed by specifically targeted recombinant aequorin [published erratum appears in Nature 1992 Dec 24–31;360(6406):768]. Nature 358, 325–327. [PubMed]
  • Saks, V.A., Khuchua, Z.A., Vasilyeva, E.V., Belikova, O., and Kuznetsov, A.V. (1994). Metabolic compartmentation and substrate channelling in muscle cells. Role of coupled creatine kinases in in vivo regulation of cellular respiration—a synthesis. Mol. Cell. Biochem. 133–134, 155–192. [PubMed]
  • Santorelli, F.M., Shanske, S., Macaya, A., DeVivo, D.C., and Di-Mauro, S. (1993). The mutation at nt 8993 of mitochondrial DNA is a common cause of Leigh's syndrome. Ann. Neurol. 34, 827–834. [PubMed]
  • Singer, S.J. (1990). The structure and insertion of integral proteins in membranes. Annu. Rev. Cell Biol. 6, 247–296. [PubMed]

Articles from Molecular Biology of the Cell are provided here courtesy of American Society for Cell Biology
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...