Characterization of the Saccharomyces cerevisiaeCytosine Transporter Using Energizable Plasma Membrane Vesicles*

The purine-cytosine permease is a carrier localized in the plasma membrane of the yeast Saccharomyces cerevisiae. The energetics of cytosine transport catalyzed by this permease has been studied in an artificial system obtained by fusion between proteoliposomes containing beef heart cytochromec oxidase and plasma membrane-enriched fractions of aS. cerevisiae strain overexpressing the permease. Upon addition of an energy donor, a proton-motive force (inside alkaline and negative) is created in this system and promotes cytosine accumulation. By using different phospholipids, it is shown that cytosine uptake is dependent on the phospholipids surrounding the carrier. It was demonstrated that the purine-cytosine permease is able to catalyze a secondary active transport of cytosine. By using nigericin and valinomycin, the ΔpH component of the proton-motive force is shown to be the only force driving nucleobase accumulation. Moreover, transport measurements done at two pH values have shown that alkalinization of intravesicular pH leads to a significant increase in cytosine uptake rate. Finally, no specific role of K+ ions on cytosine transport could be demonstrated in this system.

Living cells have to take up various ions and metabolites from extracellular medium. Recent reviews have listed many hydrophobic proteins of the yeast Saccharomyces cerevisiae plasma membrane involved in specific transport of a large array of molecules, such as hexoses and amino acids, as well as potassium and sulfate ions (1,2). Many of these proteins have been reported to work as secondary active carriers. Most of them are symports coupling the utilization of the proton gradient, built up by the H ϩ -ATPase of the plasma membrane, to the uptake of different kinds of molecules (1).
Purine-cytosine permease from S. cerevisiae is one of these plasma membrane carriers (3,4). In vivo, this permease seems to mediate the co-transport of proton and purine bases (adenine, hypoxanthine, and guanine) or a pyrimidine base (cytosine) (5,6), the energy source of this active transport being the proton electrochemical gradient built up by the H ϩ -ATPase (7,8). By measuring simultaneously hypoxanthine uptake and H ϩ and K ϩ fluxes, it has been proposed that purine translocation through the S. cerevisiae plasma membrane is an electroneutral base/H ϩ symport with a K ϩ antiport (9). In contrast to this, it has been proposed in other experiments carried out on a S. cerevisiae strain (lacking cytosine deaminase and overexpressing the purine-cytosine permease) in ATP-depletion conditions, that the pump works as an electrogenic proton symport (10) with a H ϩ /base stoichiometry close to 1 (11).
In addition, studies have been carried on S. cerevisiae purine-cytosine permease proficient strains carrying plasmid-encoded multiple copies of either wild type or mutated FCY2 gene, encoding the purine-cytosine permease. We have analyzed the effects of pH on in vivo uptake and in vitro equilibrium binding of nucleobases, and have shown a key role played by a protonable group of the permease for the solute binding step of the translocation process (12).
Mechanistic studies of solute uptake in whole cells cannot be analyzed in close detail for the following reasons: (a) inside the cell, the solute can be rapidly metabolized obscuring the characteristics of the uptake; (b) since the ionic content of the cytoplasm cannot be controlled, no specific study of the influence of ionic fluxes on solute transport can be performed. However, it is possible to analyze the transport in a well defined medium with an in vitro system obtained by membrane fusion between plasma membrane fractions and proteoliposomes containing a so-called "energy-producing system." Such a system that generates a proton-motive force has been widely used for the study of various transport systems in bacteria (13) and of some carriers of yeast plasma membrane (for a review, see Ref. 1).
In this study, we have used this technique to analyze the energetics of the transport catalyzed by the purine-cytosine permease of the plasma membrane of S. cerevisiae. Cytosine uptake has been analyzed in artificial vesicles obtained by fusion between proteoliposomes containing beef heart cytochrome c oxidase and plasma membrane-enriched fractions of a S. cerevisiae strain overexpressing the permease. Several properties of the transport system, which were difficult to assess in whole cells, have been characterized with these hybrid vesicles. With this system, it is demonstrated that the purine-cytosine permease is a secondary active transport driven only by the ⌬pH component of the proton-motive force. In addition, it is shown that the rate of uptake increases when the internal pH is more alkaline, and that the carrier is able to fully exchange cytosine at the steady state of accumulation. The influence of the phospholipid environment on the uptake is also investigated.
Strains and Growth Conditions-The strains used in this work have already been described (3). NC233-10B[pJDB] is a permease null strain, with a BglII-KpnI deletion in the FCY2 gene. NC233-10B[pAB4] is a permease-proficient strain, carrying plasmid-encoded multiple copies of the FCY2 gene strain in which the amount of functional permease is increased (14). Cells were grown at 28°C under agitation in a liquid medium containing yeast nitrogen base without amino acid and ammonium sulfate (1.75 g⅐liter Ϫ1 ), ammonium sulfate (5 g⅐liter Ϫ1 ), D-glucose (20 g⅐liter Ϫ1 ), and 25 mM sodium phtalate at pH 5.
Plasma Membrane Isolation-Plasma membrane-enriched fractions were isolated as already described (14). Cells harvested during the exponential phase (3 ϫ 10 7 cells⅐ml Ϫ1 ) were washed once with ice-cold water and then disrupted with glass beads (0.45 mm). The homogenate was submitted to differential centrifugations. In the purification procedure, mitochondrial material was eliminated by acid precipitation at pH 4.7. The purified membrane preparations were resuspended in 50 mM KH 2 PO 4 /KOH (K ϩ buffer) containing 10% (w/v) glycerol either at pH 5.7 or at pH 6.2 and stored at Ϫ20°C.
Extraction and Purification of Phospholipids-Lipids of plasma membrane fractions of S. cerevisiae NC233-10B[pAB4] strain and E. coli Xl1-blue strain were extracted according to Folch-Pi et al. (15). Commercially obtained phospholipids and phospholipids extracted from yeast or bacteria plasma membrane were purified with acetone/ethyl ether as described by Viitanen et al. (16). Purified phospholipids dissolved in chloroform were stored at Ϫ20°C under nitrogen.
Purification and Reconstitution of Cytochrome c Oxidase into Proteoliposomes-Cytochrome c oxidase was isolated from bovine heart according to Yu et al. (17). Its heme content was determined by spectrophotometric measurement (18). Proteoliposomes containing purified cytochrome c oxidase were prepared in K ϩ buffer at pH 5.7 or at pH 6.2 from different purified phospholipids by a dialysis procedure according to Van Leeuwen et al. (19), and stored in liquid nitrogen. Respiratory activity of these proteoliposomes was measured using a Clark electrode at various pH after addition of potassium ascorbate (20 mM), TMPD (350 M), and cytochrome c (35 M).
Fusion of Proteoliposomes with Plasma Membrane Vesicles-Cytochrome c oxidase-containing proteoliposomes were mixed with isolated plasma membrane fraction in K ϩ buffer containing MgCl 2 (1 mM) at pH 5.7 or 6.2 in a protein/phospholipid ratio of 1:20 (w/w). The mixture was frozen in liquid nitrogen, slowly thawed at room temperature, and passed 11 times through a filter (pore size: 400 nm) of an extruder (Avestin Inc.) (20). The resulting cytochrome c oxidase liposomesplasma membranes vesicles (termed as CL-PMV) were always used within a few hours. Vesicle fusion efficiency was determined by measurement of the fluorescence of the octadecyl rhodamine dye added to the medium before fusion, as described by Hoekstra et al. (21). Internal volume of CL-PMV was estimated by a filtration method and with [2-14 C]cytosine added to the medium before fusion (19).
Measurements of Transmembrane Electric Potential and pH Gradient-Generation of a proton-motive force was done by addition of an energization mixture composed of potassium ascorbate (30 mM), TMPD (150 M), and cytochrome c (15 M) to CL-PMV suspension in K ϩ buffer containing MgCl 2 (1 mM). Transmembrane electric potential (⌬, inside negative) was estimated from the TPP ϩ lipophilic cation distribution (19) using a TPP ϩ selective electrode (22). ⌬ was calculated according to Lolkema et al. (23) and De Vrij et al. (24). ⌬pH was measured according to Clément and Gould (25) with the pyranine dye and in the same energization conditions as those used for ⌬ determinations. Pyranine (300 M) was added to the fusion buffer and, after fusion, the dye remaining in the external medium was eliminated by elution through a Sephadex G-25 column.
Activity Measurements on CL-PMV-Ligand equilibrium binding measurements were performed at 4°C on CL-PMV by a centrifugation technique as already described (26) in K ϩ buffer, pH 5.7, containing NaCl (100 mM) and [2-14 C]cytosine (620 MBq⅐mmol Ϫ1 ) at concentrations ranging from 0.1 to 30 M in the presence (nonspecific binding) or absence (total binding) of 4 mM adenine. B max (maximal amount of specifically bound cytosine) and K d (app) (apparent half-saturation constant of cytosine binding) were calculated by nonlinear regression analysis of the saturation curve.
Cytosine uptake measurements were done with CL-PMV (0.8 mg of plasma membrane protein⅐ml Ϫ1 ) in K ϩ buffer containing MgCl 2 (1 mM) at pH 5.7 or 6.2 in a magnetically stirred vessel thermostated at 30°C. Energization and additions of nucleobase and ionophores were done as indicated in the figures. Ionophores and uncoupler stock solutions were made in ethanol and, when added to the incubation medium, were diluted 100-fold. At given intervals, aliquots (10 l) were withdrawn and added to 2 ml of ice-cold LiCl (0.1 M), filtered on cellulose nitrate filters (0.45 m, Schleicher & Schuell), and washed once with 2 ml of ice-cold LiCl (0.1 M).
For K T (Michaelis constant of transport) and V max (maximal rate of uptake) determinations of cytosine uptake, plasma membrane protein concentration was 0.2 mg⅐ml Ϫ1 . Aliquots (30 l) were withdrawn at 10-s intervals during the first minute of the uptake.
Miscellaneous-Protein concentration was determined by the Lowry procedure (27) for plasma membrane and by the biuret method (28) for cytochrome oxidase with bovine serum albumin as standard.

RESULTS
Characterization of Hybrid Vesicles-Plasma membrane preparations containing the purine-cytosine permease were fused with cytochrome c oxidase proteoliposomes using of a freeze-thaw-extrusion procedure. Fusion efficiency was estimated to be more than 95% as measured by the octadecyl rhodamine fluorescence method. The internal volume of the fused vesicles was close to 0.9 l⅐mg Ϫ1 of phospholipids for CL-PMV obtained with a protein/phospholipid ratio of 1:20 (w/w). Upon addition of the energization mixture, the rate of oxygen consumption measured with these vesicles was 5.4 Ϯ 0.4 mol of oxygen atom⅐min Ϫ1 ⅐mg of cytochrome c oxidase Ϫ1 . It could be stimulated by FCCP, 10 M, up to 8.9 Ϯ 1.3 mol of oxygen atom⅐min Ϫ1 ⅐mg of cytochrome c oxidase Ϫ1 and was totally inhibited by KCN (160 M).
Addition of the energization mixture to CL-PMV induced the generation (in about 5 min) of a steady state transmembrane electric potential difference (⌬) of about 95 mV, which was nearly stable for at least 30 min. It was increased by nigericin and collapsed by valinomycin addition (Fig. 1A). In these conditions, a ⌬pH of 1.1-1.2 units was simultaneously created. It was also stable for at least 30 min, slightly increased by valinomycin, and abolished by nigericin (Fig. 1B) or KCN (not shown) addition. Thus, energized CL-PMV showed a protonmotive force of about 160 mV stable for at least 30 min.
To establish whether the membrane fusion process affected the ligand binding properties of the purine-cytosine permease, cytosine equilibrium binding experiments were done on CL-PMV. Binding parameters obtained at pH 5.7 (K d (app) of 8.4 Ϯ 0.8 M and B max of 1700 Ϯ 70 pmol⅐mg of plasma membrane protein Ϫ1 ) were very similar to those obtained for the plasma membrane preparations used to prepare the CL-PMV (K d (app) of 13.6 Ϯ 0.4 M and B max of 1750 Ϯ 95 pmol⅐mg of plasma membrane protein Ϫ1 ).
the proton-motive force since: (a) it did not occur in the absence of energy supply; (b) it was completely abolished by addition of FCCP at the steady state of accumulation. Therefore, cytosine translocation is a secondary active transport system. The low amount of radioactivity taken up by CL-PMV of permease null strain in the presence of energy or by CL-PMV made with the purine-cytosine permease proficient strain in the absence of energy was mainly due to diffusion.
Adenine and hypoxanthine were also actively taken up by energized CL-PMV, but the levels of accumulation were lower than that observed for cytosine. The accumulation ratios obtained at the plateaus were 6.8 for cytosine (Fig. 2) and 3.4 for purine bases (not shown).
The curve showing initial velocities of the carrier-mediated transport versus external cytosine concentrations (Fig. 3)   Initial rates of uptake were determined on 30-l aliquots withdrawn from the incubation medium every 10 s during the first minute of the kinetic after cytosine addition. At each concentration tested, carrier-mediated initial rate of uptake was obtained from the difference between the initial rate measured in the absence (active transport and diffusion) and in the presence (diffusion) of uncoupler.

Effect of Phospholipid Composition of CL-PMV on Cytosine
Uptake-Fused vesicles of different phospholipid compositions were prepared. Their abilities to create and maintain ⌬ and ⌬pH were tested as described and were not significantly different from those described in Fig. 1 (not shown). Among the various phospholipids used for the preparation of CL-PMV, phosphatidylethanolamine from E. coli was found to be the most efficient phospholipid for active transport (Fig. 4). For this reason, all subsequent experiments described in this work were done with CL-PMV prepared with this phospholipid.
Cytosine Exchange-An important point was to know whether the solute accumulation plateau corresponded to cessation of the influx or to the same influx and efflux rates. When the accumulation plateau was reached, addition of an excess of non-labeled cytosine led to a rapid and complete efflux of the internal radioactive solute (Fig. 5A). On the other hand, when radioactive cytosine was added to energized CL-PMV pre-loaded with non-labeled solute, accumulation of the radioactive cytosine inside the vesicles was observed (Fig. 5B). These data showed that the accumulation plateau was the result of equal influx and efflux rates. Moreover, influx and efflux apparent rates of radioactive cytosine observed at the steady state of accumulation (exchange rates) were very similar and significantly higher than the corresponding apparent uptake rate.
Roles of the ⌬pH and ⌬ Components of ⌬p on Cytosine Uptake-Uptake experiments were done at two external pH values of 5.7 and 6.2; in each case, the ⌬pH and ⌬ component of the ⌬p were collapsed by use of the appropriate ionophores added to the CL-PMV suspension 1 min before energization (Fig. 6).
Addition of nigericin (⌬p ϭ ⌬) totally prevented active cytosine transport. Addition of valinomycin (⌬p ϭ Ϫ0.06 ⌬pH) and no ionophore addition (⌬p ϭ ⌬ Ϫ0.06 ⌬pH) led to active transport of cytosine for which the observed accumulation ratios were related to the existing ⌬pH component of the ⌬p (Table I). This suggests that the ⌬pH component was the only driving force of the active solute transport process. In these experiments, showing that the ⌬pH was totally collapsed and transformed at least partly in ⌬, nigericin was added at a ratio of 6.6 pmol⅐mg of phospholipids Ϫ1 , a value which is much lower than the value of 115 pmol⅐mg of phospholipids Ϫ1 used for ⌬ and ⌬pH measurements displayed in Fig. 1. Therefore, at the ratio used for cytosine uptake, nigericin had negligible or no uncoupling effect.
Moreover, at each of the external pH values tested, the apparent rate of cytosine uptake was higher in the presence of valinomycin than in its absence (Fig. 6, A and B). Valinomycin addition to energized CL-PMV transforms, at least partially (see Fig. 1), the ⌬ into ⌬pH leading to alkalinization of the internal pH. In each of the four experiments shown in Fig. 6, the cytosine accumulation ratio was related to the ⌬pH ( Table  I). The stimulation of the apparent rate of cytosine uptake observed in the presence of valinomycin could have to deal with modification of ⌬pH value or alkalinization of vesicle interior or both.
In this respect, one has to compare the cytosine uptake experiments shown in Fig. 6A in the presence of valinomycin and in Fig. 6B in the absence of valinomycin. These two experiments, done at different external pH (5.7 and 6.2, respectively), showed almost the same internal pH value and thus, different ⌬pH values. One can see that the two curves display very similar apparent rates of uptake despite the fact that ⌬pH were different in both experiments. Thus, if the CL-PMV ⌬pH controls the cytosine accumulation ratio, the apparent rate of cytosine uptake is related to the internal pH value.
Cytosine uptake experiments in buffers containing Na ϩ instead of K ϩ ions were also carried out and the effects of an Na ϩ ionophore (nonactin) and of an Na ϩ /H ϩ exchanger (monensin) were tested. The results shown in Fig. 6C were analogous to those described for K ϩ ions (Fig. 6A), confirming that the ⌬pH was the only driving force for cytosine uptake, and that the active transport did not display any strict specificity for K ϩ ions. DISCUSSION In this report, cytosine transport was analyzed in artificial energizable vesicles obtained by fusion of plasma membraneenriched fractions of S. cerevisiae containing the purine-cytosine permease and proteoliposomes containing cytochrome c oxidase. CL-PMV displayed an internal volume value very similar to that already obtained by other methods (19,29,30). The cytochrome c oxidase embedded in CL-PMV displayed specific activity, total inhibition by cyanide, and activation upon FCCP addition similar to that observed with isolated beef heart mitochondria (17). FCCP stimulation of oxygen consumption indicated that CL-PMV were not very leaky for protons. The equilibrium binding parameters values (K d (app) and B max ) measured for cytosine were almost the same for CL-PMV as those measured with plasma membrane-enriched fractions. However, because of a dramatic decrease in cytochrome c oxidase activity at pH values below 5.5, the experiments presented here were done at pH 5.7 or 6.2. These pH conditions, which were not the optimum for purine-cytosine permease activity, were acceptable since nucleobase uptake on intact cells and equilibrium binding on plasma membrane fractions were still accurately measured at these pH (12).
Upon energy supply, CL-PMV containing purine-cytosine permease were able to actively transport nucleobases in a solute/H ϩ symport process ( Fig. 2 and Table I). The apparent Michaelis constant of cytosine uptake measured for CL-PMV (Fig. 3) was not dramatically different to that of intact cells: they were 13.0 Ϯ 3.3 and 1.9 Ϯ 0.1 M, respectively. The uptake catalytic constant (calculated from B max and V max values given FIG. 4. Influence of CL-PMV phospholipid composition on the cytosine uptake. CL-PMV were made by fusion between plasma membrane fractions isolated from NC233-10B[pAB4] and proteoliposomes containing cytochrome c oxidase made with different kinds of phospholipids. q, phosphatidylethanolamine from E. coli; E, purified phospholipids from E. coli plasma membrane; ϫ, purified phospholipids from S. cerevisiae plasma membrane; Ⅺ, soybean phosphatidylcholine; and f, egg yolk phosphatidylcholine. The plasma membrane protein/phospholipid ratio was 1:20 (w/w). For all CL-PMV, cytosine uptake experiments were performed as described.
in this text) was around 2 min Ϫ1 , a value 30 times lower than that already published for intact cells which is close to 60 min Ϫ1 (14). Such reduced efficiency of a carrier is rather common in reconstituted systems and mainly attributable to effects of physical and chemical constraints imposed during plasma membrane preparation and vesicle fusion, and to the nature of the phospholipid environment of the carrier in the artificial system. Data in Fig. 2 demonstrate that cytosine uptake is a secondary active transport process depending on the presence of a proton-motive force or, at least, one of its components. The observed cytosine accumulation ratio was only 6.8 when the ⌬p was 160 mV. Various attempts have been done to improve this low accumulation ratio by modifying the preparation of CL-PMV. Instead of the extrusion procedure, sonication methods described for other carriers were tried (19,29). In addition, in the fusion procedure, the protein/phospholipid ratio, the pore size of the extruder filters, and the number of passes through the filter were also varied. The vesicles obtained from these numerous attempts always displayed the same or lower accumulation ratio (not shown). Modifications of the phospholipid environment of the carrier have also been tried by preparing CL-PMV with purified phospholipids from various sources (Fig.  4). Phosphatidylethanolamine from E. coli was more efficient than all the other phospholipids, and particularly more so than the natural purified phospholipids extracted from plasma membrane fractions of S. cerevisiae. Such behavior has already been observed for some secondary active carriers (19,31). Moreover, phosphatidylethanolamine from E. coli was also shown to confer the best carrier environment for the well investigated branched amino acid permease of Streptococcus cremoris, located in vivo in a membrane totally lacking this phospholipid (31). Clearly, more work is needed to understand why natural yeast phospholipids did not allow good uptake and accumulation of cytosine in CL-PMV.
Use of ionophores collapsing either the ⌬pH or the ⌬⌿ component of the ⌬p (Fig. 6) showed that the cytosine active transport was solely driven by the ⌬pH component and that the solute accumulation ratios observed were related with the ⌬pH (Table I). In all cases, the accumulation plateaus were stable for about 10 -15 min except for the experiment shown in Fig.  6B in the presence of valinomycin, where the curve showed a maximum instead of a plateau followed by a decrease which was related to the energy source exhaustion (as observed by the appearance of a blue color of the incubation medium due to the stable free radical form of TMPD).
Thus, if we assume a H ϩ /solute stoichiometry of 1 (11) and if the system was to reach its thermodynamic equilibrium, one FIG. 6. Effects of ionophores and pH on cytosine uptake. CL-PMV were prepared from the NC233-10B[pAB4] strain in K ϩ buffer at pH 5.7 (A) or 6.2 (B) or in a Na ϩ buffer (NaH 2 PO 4 /NaOH), pH 5.7 (C). Vesicle suspension, cytosine addition, energization, and measurements of cytosine uptake were performed as described. The appropriate ionophores were added 1 min before energization. f, no ionophore; E, valinomycin, 100 nM (panels A and B) and nonactin 100 M (panel C); Ⅺ, nigericin 100 nM (panels A and B) and monensin 100 nM (panel C). Each curve shows the results obtained from one typical experiment which was representative of several. Energy Coupling of the Purine-cytosine Permease ⌬pH unit should lead to a solute accumulation ratio of 10. In the various uptake experiments described, we observed values ranging from 32 to 58% of that value (Table I). Two main reasons could account for these results. First, in the fusion procedure, there was a large excess of cytochrome c oxidase proteoliposomes as compared with the plasma membrane fraction. Thus, in the CL-PMV population not all the energizable vesicles might contain purine-cytosine permease. Unfortunately, it was not possible to determine the amount of functional vesicles as already done for reconstituted vesicles containing other permeases, where the reconstituted material contained from 10 to 50% of functional vesicles (19,30,32). Second, because of the low affinity of the carrier in CL-PMV for the nucleobases, and because of the low specific radioactivity of the commercially available [2-14 C]cytosine, a rather high external solute concentration had to be used to measure the uptake accurately. In such conditions, as solute accumulation proceeds, it causes the internal solute concentration to reach values for which the efflux via passive diffusion becomes large for such a hydrophobic compound and, consequently, limits its accumulation ratio. This is shown by uptake analyses done at external cytosine concentrations of 25 M (Fig. 5A) and 120 M (Fig. 5B) which led to accumulation ratios of 6.8 and 3.4, respectively. This has been observed and discussed for other carriers (33). Therefore, this made it very difficult to determine accurately the H ϩ /cytosine stoichiometry. However, our data (Table I) fit with the postulated stoichiometry of 1 (11). Moreover, replacement of K ϩ by Na ϩ ions in the uptake buffer did not change the behavior of the transport (Fig. 6). Clearly, this is consistent with the results of Hopkins et al. (10). Therefore, if the nucleobase transport is electroneutral, K ϩ or Na ϩ counterflow should not occur via the permease itself, but rather through some unidentified monocation/H ϩ exchange system located in the plasma membrane. But, an alternative to this would be that the cytosine transport is actually electroneutral by itself. This means that the carrier would take up, instead of the cytosine neutral form, its anionic specie (very scarce at the experimental pH since the relevant pK a of cytosine is 12.2) with a tremendous affinity: this mechanism would be in good agreement with the fact that the cytosine accumulation ratio depends only on the ⌬pH component of the proton-motive force.
On the other hand, another important finding was the role played by the internal pH value on the uptake kinetics, alkaline values leading to stimulation of the apparent rate of cytosine uptake (Fig. 6). In this respect, it is to be noted that there was a clear dependence of the apparent rate of cytosine uptake on the internal CL-PMV pH. Weak acid deprotonation is known to be very rapid and, this event, by itself, cannot be ratelimiting in the active H ϩ /solute translocation process. However, alkalinization of the medium leads to an increase in the ratio of deprotonated/protonated forms of the permease. As a consequence, this might increase the relative amount of the carrier form involved in the rate-limiting step of the catalytic cycle, as described for the models of cytosine transport (11) and of leucine transport (34). It would be highly speculative to propose a mechanistic model based only on our present results, but these data are nevertheless very consistent with the above mentioned models.
In intact cells lacking cytosine deaminase, the cytosine accumulation ratio was close to 500 and the catalytic constant was around 60 min Ϫ1 (14). Under these conditions, there was a transmembrane ⌬pH close to 2.7 pH units, when the external medium was pH 4.5 ([ 31 P]P i NMR cytosolic pH measurements). 2 Such pH values are the optimal in vivo conditions for carrier-mediated cytosine active transport, leading to thermodynamic equilibrium between ⌬pH and the cytosine accumulation ratio (assuming a H ϩ /cytosine stoichiometry of 1). Unfortunately, in vivo ⌬pH conditions could not be mimicked in CL-PMV.
Finally, exchange experiments at the steady state level of accumulation have shown that no transinhibition by accumulated cytosine occurred, and that the observed steady state plateau of accumulation was the result of equal influx and efflux rates (Fig. 5). Such behavior has already been observed in other transport system (35). In contrast, it has been shown that yeast arginine and maltose permeases are totally irreversible (29,35). An interesting case is the histidine permease of Salmonella typhimurium. By using energized right-side-out plasma membrane vesicles, it was first shown that the transport was fully reversible (36). Recently, by using a totally reconstituted system, the same group showed it to be fully irreversible and regulated by transinhibition (37). As, in our present work, CL-PMV were prepared from plasma membrane fractions, caution is needed about the reversible aspect of cytosine accumulation and the answer to that question would be obtained by total reconstitution.
In conclusion, the CL-PMV system used in this work is a very useful tool in the study of the purine-cytosine permease. It has allowed us to make the following statements: (a) the carrier catalyzes a nucleobase-active transport process; (b) the ⌬pH component of the ⌬p is the sole driving force for solute accumulation; (c) the vesicle internal pH plays a crucial role in the uptake kinetics; and (d) the purine-cytosine permease does not show any strict specificity for K ϩ ions.
However, our observations indicate the limitations of the CL-PMV system for complete characterization of the translocation process and show that further elucidation of the transport mechanism will require a totally reconstituted system.