Characterization of a dCTP Transport Activity Reconstituted from Human Mitochondria*

A protein fraction of mitochondria from human acute lymphocytic leukemia cells, which could be reconstituted into proteoliposomes to have dCTP transport activity, has been partially purified by hydroxyapatite and blue Sepharose chromatography. The dCTP transport activity in proteoliposomes was time-dependent and could be activated by Ca2+ and to a lesser extent by Mg2+. None of the other divalent cations tested could activate the transport activity. The K m value of dCTP in the presence of Ca2+ was shown to be 3 μm. dCDP but not dCMP or dCyd could inhibit the transport activity. Other deoxynucleoside triphosphates could also inhibit the uptake of dCTP with the potency dGTP = dATP > TTP. Although ATP could competitively inhibit dCTP uptake with aK i value of 8 μm, the reconstituted dCTP uptake activity was not sensitive to the ATP/ADP carrier inhibitor atractyloside or the sulfhydryl reagent N-ethylmaleimide. This suggests that the dCTP transport system studied is not the same as the ATP/ADP carrier. In conclusion, these studies describe the first functionally reconstituted mitochondrial carrier that displays an efficient transport activity for dCTP.

The inner membrane of mitochondria normally possesses very low electrophoretic permeability to most molecules. However, through the action of specific transport systems, the inner mitochondrial membrane is selectively permeable to a number of co-factors, metabolites, and nucleotides (1,2). The AAC 1 is the only mitochondrial nucleotide carrier successfully isolated and reconstituted in an active state (3). The AAC is highly selective for ADP and ATP, mediating the import of ADP from the cytoplasm and export of matrix ATP. The ATP-Mg/P i carrier is another nucleotide transport activity that has been described in isolated mitochondria (4) but has not been successfully isolated. Another member of the family of anion transport proteins that reside within the mitochondrial inner membrane is the phosphate carrier that catalyzes transport of inorganic phosphate into the mitochondrial matrix, where the phosphate is utilized for phosphorylating ADP to ATP (5).
The existence of a mechanism for mitochondrial dNTP uptake has been suggested by DNA synthesis experiments using isolated mitochondria (6 -8). These studies indicated that ex-ogenous dNTPs could be utilized by isolated mitochondria to synthesize mitochondrial DNA. In addition, based on previous studies from our laboratory, antiviral nucleoside analog triphosphates in mitochondria appear to originate from the cytoplasm (9). In these studies, the absence of 2Ј,3Ј-dideoxycytidine-induced mitochondrial toxicity in cytoplasmic dCyd kinase-deficient cells suggested that the cytoplasmic-formed metabolite 2Ј,3Ј-dideoxycytidine triphosphate is the nucleotide source in the inhibition of mitochondrial DNA synthesis. Thus, the 2Ј,3Ј-dideoxycytidine triphosphate metabolites inside mitochondria appear to originate from the cytoplasm. However, the experimental evidence supporting the existence of a dNTP carrier in intact mitochondria is only circumstantial and not definitive as for other mitochondrial metabolite carriers.
The final proof for the existence of a carrier protein is its isolation and functional reconstitution. To that end, we demonstrate that upon incorporation into lipid vesicles, partially purified mitochondrial protein catalyzes dCTP uptake. This transport activity displays unique substrate specificity and inhibitor sensitivity from that previously described for mitochondrial anion transporters. To our knowledge, this is the first report of a substantially purified preparation of this class of mitochondrial carrier activity in functional form.
Preparation of [␥-32 P]dCTP-Nucleoside diphosphate kinase (from Baker's yeast; Sigma) was used to prepare [␥-32 P]dCTP from dCDP as follows. A 1-ml mixture containing 10 units of nucleoside diphosphate kinase, 2 mM [␥-32 P]GTP (100 Ci), and 0.5 mM dCDP in 150 mM Tris acetate, pH 7.5, and 10 mM MgCl 2 was incubated at 37°C for 10 min. The reaction was stopped by the addition of 3 volumes of ice-cold methanol and incubated on ice for 15 min. After centrifugation at 14,000 ϫ g for 10 min, the supernatant containing [␥ 32 P]dCTP was evaporated to dryness and resuspended in 500 l of water. The mixture containing dCDP, GDP, [␥-32 P]GTP, and [␥-32 P]dCTP was chromatographed by ion exchange HPLC using a Whatman Partisil-SAX column (4.6 mm ϫ 25 cm) at a flow rate of 1 ml/min. The nucleotides were resolved with a gradient consisting of water to 30 mM buffer (potassium phosphate, pH 6.7) from 0 to 10 min, 30 -150 mM buffer from 10 to 15 min, 150 mM buffer from 15 to 50 min, and 300 mM buffer from 50 to 80 min. Peaks were identified by authentic standards, and the peak corresponding to [␥-32 P]dCTP was collected. The purified [␥-32 P]dCTP was diluted 5-fold with water, placed over a DEAE-Sephadex A-25 column (2-ml bed volume), and washed with 20 ml of water to remove the phosphate buffer, and the [␥-32 P]dCTP was eluted with 10 ml of ammonium formate (500 mM). The eluted sample was freeze-dried to remove the ammonium formate and resuspended to a volume of 2 ml in water. The purity of [␥ 32 P]dCTP was verified by ion exchange HPLC as described above, and the radiospecificity was determined to be 90 Ci/mmol.
Preparation of Mitochondria and Partial Purification of Proteins with dCTP Transport Activity-Mitochondria were isolated from acute lymphocytic leukemia cells after leukophoresis of a patient in blast crisis and further purified on a discontinuous sucrose gradient as described previously (9). Frozen mitochondria (100 -150 mg) were solubilized for 10 min on ice in 10 ml of a buffer containing 20 mM Hepes (pH 7.0), 1% Triton X-114, 150 mM Na 2 SO 4 , 1 mM phenylmethylsulfonyl fluoride, and 1 mM benzamidine. The detergent was separated from the aqueous phase by centrifugation at 280,000 ϫ g for 1 h at 4°C. The upper aqueous phase was decanted, and the detergent phase was resuspended to 10 ml in 20 mM Hepes (pH 7.0), 1 mM phenylmethylsulfonyl fluoride, and 1 mM benzamidine (buffer B). The detergent extract was applied to an HPT column (1 g of dry HPT/20 mg of protein) connected in tandem with a blue Sepharose column (1 ml of Blue Sepharose/2 mg of protein) equilibrated with buffer B. The unbound protein-containing fraction was eluted from the tandem columns with buffer B.
Preparation of Proteoliposomes-Asolectin was further purified as described (10). For use in transport studies, cholesterol (120 mg/ml final concentration) was added to the asolectin (200 mg/ml final concentration). Cholesterol has been reported to prevent protein-mediated leakage in reconstituted proteoliposomes (11). The asolectin was stored in chloroform under nitrogen in a light proof container at Ϫ20°C. Immediately before use, an aliquot of asolectin was dried under a stream of nitrogen. Following the addition of 2 ml of 20 mM Hepes (pH 7.0), the lipid was dispersed by sonication (Branson Sonifier 250; microtip output control ϭ 4; 80% duty cycle; on ice) until the mix appeared transparent.
Aliquots of freshly isolated protein fractions (typically 0.3 mg) were mixed with lipid vesicles (55 mg) to a final volume of 1 ml. This mixture was vortexed and then rapidly frozen in liquid nitrogen. For ATP/ADP exchange reactions and dCTP efflux experiments, proteoliposomes were prepared in the presence of 100 M ADP and 20 M [␣-32 P]dCTP, respectively. These proteoliposomes could be stored at Ϫ80°C up to 5 days without appreciable loss of dCTP transport activity. Immediately prior to assay, the samples were freeze-thawed twice (total of three cycles) using liquid nitrogen and thawed a final time in an ice-water bath followed by sonication (Branson Sonifier 250; microtip output control ϭ 2; 70% duty cycle; 30 burst; total sonication time ϭ 25 s; on ice). For ATP/ADP exchange reactions and dCTP efflux experiments, external ADP and [␣-32 P]dCTP, respectively, were removed by passing the proteoliposomes over an ion exchange column (Dowex; 1 ϫ 3.5 cm) equlibrated with 20 mM Hepes (pH 7.0) at 4°C. The first 1 ml of the turbid eluate from the Dowex column was collected and then used for transport experiments. A heterogeneous population of large, primarily unilamellar proteoliposomes suitable for transport studies are generated by this general freeze-thaw-sonicate procedure (12)(13)(14)(15).
Assay Conditions for dCTP Uptake-All measurements of nucleotide uptake into reconstituted proteoliposomes were carried out at 37°C. Transport was initiated by the addition of 4 l of radiolabeled substrate (typically [␣-32 P]dCTP, 10 M final concentration, 30 Ci/mmol) to 190 l of proteoliposomes in a final volume of 200 l. For competition assays, competitor was added to the desired concentration to the proteoliposome mix before the radiolabeled substrate. After the desired incubation time, the reaction mixture was placed over a Dowex ion exchange column (1 ϫ 3.5 cm) equilibrated with 20 mM Hepes (pH 7.0) at 4°C in order to remove the external radioactivity not transported into proteoliposomes. The liposomes were eluted with 4 ml of ice-cold equilibration buffer, collected in 16 ml of scintillation fluid, vortexed, and counted. Subtracting values obtained when a 100-fold molar excess of nonradiolabeled substrate was added 2 min prior to the radiolabeled substrate made corrections for nonspecific uptake of radioactivity. For efflux experiments, external radioactivity was removed by passing the samples through a Dowex column and collecting and counting the liposomes as described above. The transport activity was calculated by subtracting the experimental values from the control values (efflux at 4°C).

RESULTS AND DISCUSSION
Mitochondria possess a variety of specific carrier systems for the transport of metabolites across the inner mitochondria membrane. Because of the importance of mitochondrial toxicity in nucleoside analog chemotherapy (for a review, see Ref. 16), we have sought to extend our knowledge about dCTP uptake into mitochondria. Our procedure for isolation and reconstitution of active dCTP transport from human mitochondria consisted of three basic steps. First, sucrose gradient isolated mitochondria were extracted with Triton X-114. Second, hydrophobic membrane proteins in the detergent were separated from the aqueous phase. Finally, the detergent phase contain-ing solubilized membrane proteins was chromatographed on HPT and blue Sepharose. When applying these conditions to mitochondria from acute lymphocytic leukemia cells, the activity of dCTP transport was significantly enriched, but it remains a heterogeneous mixture of proteins.
In Table I the activity of the protein obtained by blue Sepharose chromatography relative to that of the HPT chromatography is reported. For comparison, we report also the data on the total mitochondrial membrane extract. However, reconstitution of a total membrane extract in liposomes and its comparison with the data obtained from a smaller number of proteins after chromatography should be done with caution and may not be fully reliable because of differences in protein-protein interactions, liposome size, and passive permeabilities. Uptake of [␣-32 P]dCTP could be detected in proteoliposomes reconstituted using crude mitochondrial protein extracts and detergent phase membrane proteins with a specific activity of 2.8 and 18 pmol/min/mg protein, respectively. However, only minimal dCTP transport activity could be detected in proteoliposomes using aqueous phase proteins after detergent phase separation. Approximately 180-and 90-fold increases in the specific activity of dCTP uptake in proteoliposomes compared with crude extracts were obtained after HPT and blue Sepharose chromatography, respectively. The decrease in specific activity after blue Sepharose chromatography compared with HPT elutes may indicate the presence of multiple dCTP transport activities resolved by blue Sepharose. Alternatively, this result may suggest the removal of a protein(s) component of a multisubunit complex responsible for dCTP transport. Furthermore, there is a large increase in the total activity recovered in the HPT eluate (i.e. 965%) relative to the starting material. The recovery of activity greater than the starting material may reflect the removal of an inhibitor or interfering activity. Further studies will be required to define the role of these factors in the above observations.
Hydroxyapatite chromatography is important because it offers a large single purification step common to mitochondrial anion transporters (17)(18)(19)(20)(21)(22)(23). Under the conditions described under "Experimental Procedures," the reconstituted protein eluate after HPT chromatography contained atractyloside-sensitive AAC activity in addition to dCTP transport activity (data not shown). To improve the purification of the dCTP carrier activity, the protein eluate after HPT chromatography was applied to a blue Sepharose column to remove the AAC activity (24). Table II shows that after HPT and blue Sepharose chromatography the internalized ADP as a counterion did not stimulate ATP uptake in reconstituted proteoliposomes. This is in contrast to HPT protein eluates reconstituted into proteoliposomes, where internalized ADP dramatically increased ATP uptake (data not shown). Furthermore, ATP uptake in proteoliposomes was observed after HPT and blue Sepharose chro- matography but was not sensitive to the AAC inhibitor atractyloside. These results demonstrate that the reconstituted protein eluate after HPT and blue Sepharose chromatography is capable of ATP transport, but this activity is not associated with AAC activity. Further studies will be required to determine whether one or more proteins are present that play a role in dCTP and ATP transport. The unbound protein fraction from HPT and blue Sepharose chromatography was functionally characterized for dCTP transport activity after incorporation into liposomes. To obtain a base line of transport activity in the absence of divalent cations, the effect of EDTA on transport using 10 M [ 32 P]dCTP in reconstituted proteoliposomes was measured (Fig. 1A). A significant dose-dependent decrease in dCTP uptake was observed with a maximum effect at 0.2 mM EDTA. We next measured the effect of 1 mM divalent cations on dCTP uptake in the presence of 0.2 mM EDTA (Fig. 1B). Mn 2ϩ and Zn 2ϩ had a negligible effect, while Cu 2ϩ and Co 2ϩ inhibited dCTP uptake. However, Ca 2ϩ dramatically increased dCTP uptake, whereas Mg 2ϩ and Ni 2ϩ moderately increased uptake. In addition, NaCl and KCl were observed to have no effect on dCTP uptake (data not shown). The effects of Ca 2ϩ and Mg 2ϩ were further investigated in Fig. 1C. A dose-dependent increase in dCTP uptake was noted with Ca 2ϩ that was saturable at about 2 mM (an 8-fold increase), whereas the effect of Mg 2ϩ was less dramatic (a 2.5-fold increase).
What effect is Ca 2ϩ having to stimulate dCTP transport activity? The transport assay described in this study consisted of several components that, in principle, could all be targets for the cations: negatively charged phospholipids, the negatively charged substrate dCTP, and the inserted carrier protein. Interaction of Ca 2ϩ with the anionic substrate dCTP is questionable because of the large discrepancy in the actual concentrations used. Interestingly, the affinity of Mg 2ϩ for ATP has been reported to be about 10 times higher than that of Ca 2ϩ (25). If Mg 2ϩ interacts with the triphosphate moiety of dCTP similar to that of ATP to stimulate dCTP transport in reconstituted proteoliposomes, the activation effect of Mg 2ϩ should be seen at lower concentrations. Second, the results presented in this paper cannot simply be due to surface charge screening of the phospholipid vesicles, since there are significant differences with respect to the stimulation power between the various cations tested. Finally, although the protein may be the most reasonable choice as the target of the activation by Ca 2ϩ , further studies will be required to address this issue.
The effect of protein concentration on the reconstitution of the dCTP transport activity was investigated by progressively increasing the protein concentration in the reconstitution mix while keeping the lipid concentration constant (55 mg/ml). In-creasing the protein concentration from 0.05 to 0.4 mg/ml resulted in a parallel increase in the transporter-mediated uptake of [ 32 P]dCTP ( Fig. 2A). There also was a slight increase in the background rate of uptake with increasing protein concentration (Ͻ2 pmol/min at 0.4 mg/ml protein), but this increase was much less pronounced than the increase in the rate of protein-mediated uptake. From these experiments, we selected 0.3 mg/ml protein as an optimum condition to characterize dCTP uptake.
In Fig. 2B the dCTP uptake catalyzed by mitochondrial protein reconstituted into liposomes has been plotted as a function of time. There is an apparent initial rapid component occurring within 1 min and a slower uptake component that reached a  steady state within 5-10 min. A temperature dependence in dCTP uptake is also observed with negligible [ 32 P]dCTP uptake when the proteoliposomes are incubated at 4 o C. Subsequent experiments were conducted at 1 min to assess the initial uptake of dCTP in reconstituted proteoliposomes. Because the insertion of protein in the lipid vesicle is in a random orientation, unidirectional transport in both directions may explain the time-dependent change in the transport rate. To measure the efflux of dCTP, proteoliposomes were preloaded with [␣-32 P]dCTP. Fig. 2C illustrates a time-dependent efflux of [␣-32 P]dCTP from proteoliposomes similar to that observed for influx. To directly address if dCTP is transported without prior breakdown, dCTP was radiolabeled with 32 P in the ␥-position only and used for uptake studies. Fig. 2D illustrates the timeand temperature-dependent uptake of [␥-32 P]dCTP. Under these conditions, [␥-32 P]dCTP uptake decreases in the presence of excess nonradiolabeled dCTP. 2 However, a 10-fold molar excess of inorganic phosphate has no effect on [␥-32 P]dCTP uptake, suggesting that protein-mediated uptake of 32 P-labeled inorganic phosphate (e.g. the phosphate transporter) plays no role in the observed transport activity. 2 Thus, it is unlikely than dCTP is broken down before uptake into proteoliposomes. In addition, the exchange of a counterion (including dCDP or ADP) for dCTP in reconstituted proteoliposomes, similar to that for ATP and ADP in the AAC nucleotide transport system, has not been observed. 2 We have thus far demonstrated the transport activity of reconstituted mitochondrial protein by measuring the uptake of radioactive substrate into proteoliposomes. With this assay method, the amount of substrate taken up can be measured, but it is not possible to obtain precise initial kinetic parameters because the radioactive substrate remaining outside the proteoliposomes has to be removed before the radioactive sub-strate transported inside the vesicles can be measured. Further studies were performed to elucidate a dose-response curve for dCTP uptake in proteoliposomes. The value obtained from a Lineweaver-Burk plot (Fig. 3) indicated an apparent K m of 3 M. The inhibition of dCTP uptake in proteoliposomes by ATP was also studied. Fig. 3 demonstrates that ATP competitively inhibits dCTP uptake. The K i value was estimated to be 8 M by a replot of the slopes versus the ATP concentration. Numerous studies in cell culture, particularly that of the T-lymphoblastic cell line CEM, have demonstrated dCTP intracellular pool sizes ranging from 8 to 38 M (26 -30). Whereas CEM cells are rapidly dividing cells in culture, another study using peripheral blood lymphocytes reported a dCTP pool size of 1.5 M in quiescent cells and 18 M in phytohemagglutinin-activated lymphocytes (31). Thus, the mitochondrial dCTP transport activity reported herein may be significant within the range of these reported intracellular dCTP concentrations.
The substrate specificity of this novel mitochondrial carrier activity was examined by investigating the ability of dCTP and its derivatives to inhibit uptake of the unidirectional, inward flux of 10 M [ 32 P]dCTP (Table III). If [ 32 P]dCTP were broken down to [ 32 P]dCDP or [ 32 P]dCMP for transport, we would expect excess dCDP or dCMP to decrease uptake by dilution of the radiospecificity of [ 32 P]dCDP or [ 32 P]dCMP derived from the action of phosphatase present in the preparation. Excess unlabeled dCTP is included as a positive control. Uptake of [ 32 P]dCTP was most affected by the triphosphate of deoxycytidine (dCTP), while the diphosphate form (dCDP) exhibited a moderate effect. In contrast, the nucleoside deoxycytidine and its monophosphate derivative (dCMP) had little effect on [ 32 P]dCTP uptake (Ͻ10% inhibition). Thus, the reconstituted transporter discriminates between different phosphate forms of deoxycytidine. The uptake of ␣and ␥-labeled [ 32 P]dCTP into proteoliposomes and the effect of excess dCDP and dCMP on uptake suggests that dCTP is the primary substrate used for transport. The effects of other nucleoside triphosphates on [ 32 P]dCTP uptake are shown in Table IV. At equimolar concentrations, the purines GTP, ATP, dATP, and dGTP moderately reduced [ 32 P]dCTP uptake in proteoliposomes with more potent inhibi-tion with a 5-fold excess of competitor. Interestingly, these purines were more effective than TTP or UTP in competing for [ 32 P]dCTP uptake in proteoliposomes. These results raise the possibility that the observed transport activity may not be specific for dCTP only but may reflect the activity of a general dNTP carrier. Further studies will be required to assess if the competition represents uptake of other dNTPs.
The sensitivity of the reconstituted carrier to externally added inhibitors was investigated using the sulfhydryl reagent N-ethylmaleimide (a potent inhibitor of the phosphate carrier) (32), atractyloside (a specific inhibitor of the AAC) (33), and phosphate. N-Ethylmaleimide, atractyloside, and phosphate had no significant inhibitory effect on dCTP transport (Table  III). However, a 50% increase in dCTP uptake occurs in the presence of atractyloside. The mechanism by which atractyloside increases dCTP uptake is unclear. These data indicate that the phosphate carrier and the AAC do not play a role in the dCTP uptake activity observed in this study. We have not assessed the uptake of other dNTPs, although it is possible that there are other carrier activities for these metabolites. To our knowledge, this is the first report of a preparation of a mitochondrial dCTP carrier activity in functional form from any source.
This transport activity suggests an important role of bioactive endogenous dNTP uptake in mitochondria. Thus, cytoplasmic pools of dNTPs may be available to mitochondria, which

TABLE III
Effect of competitors and inhibitors on the uptake of [ 32 P]dCTP by proteoliposomes Competitors were added immediately before the radiolabeled substrate. Atractyloside and NEM were added at a concentration of 1 mM 3 min before the labeled substrate. Proteoliposomes were prepared after HPT and blue Sepharose chromatography as described under "Experimental Procedures." Transport was started by the addition of 10 M [␣- 32  are continually being damaged, and for replication of their DNA. The transport activity reported in this study may also be important in mediating the mitochondrial accumulation of cytotoxic nucleoside analogs used in chemotherapy. Clinical and laboratory findings have focused attention on damage to mitochondrial function as a mechanism of toxicity of various drugs. Indeed, available evidence supports the hypothesis that the myopathy induced by anti-human immunodeficiency virus nucleoside analogs is due to inhibition of mitochondrial DNA synthesis (34,35). The existence of the mitochondrial dCTP carrier activity described in this study may provide insight into the molecular mechanism underlying nucleoside analoginduced mitochondrial toxicity and protection.