pH gradient-dependent phosphate transport catalyzed by the purified mitochondrial phosphate transport protein.

The mitochondrial phosphate transport protein (Wohlrab, H. (1980) J. Biol. Chem. 255, 8170-8173), which co-purifies with the adenine nucleotide translocase, has been isolated with a modified procedure resulting in an increased yield and in a preparation that is stable to storage in liquid nitrogen. The phosphate transport protein was incorporated into liposomes (phosphatidylethanolamine/phosphatidylcholine/phosphatidic acid, 2.75:1:1), and phosphate/phosphate exchange rates were determined. With pHe (extraliposomal) = pHi (intraliposomal) = 7.2, we found a Km of 2.5 mM independent of [(Pi)i] and a Vmax of 12 mumol/min . mg of protein. Parallel phosphate transport experiments were carried out with liposomes containing phosphate transport protein isolated from mitochondria inhibited by N-ethylmaleimide. Phosphate transported unidirectionally [pHe = pHi = 6.8; = 1.0 mM, (Pi)i = 0 mM] reaches a maximum at 30 s and is 0.13 and 0.05 mumol/mg for active and inhibited protein, respectively. At pHe = 6.8 and pHi = 8.0, the respective amounts are 0.45 and 0.05. At pHe = 8.0, the uptake becomes the same with active and inhibited protein (pHi = 6.8, 0.15 and 0.14; pHi = 8.0, 0.25 and 0.20). At all the above pH values, about the same uptake is observed with liposomes prepared without protein as those prepared with inhibited protein. The initial rate of protein catalyzed unidirectional flux [(Pi)e = 1.0 mM, pHe = 6.8; (Pi)i = 0 mM, pHi = 8.0) is 2 mumol/min . mg of protein or 8 mumol/min . mg of phosphate transport protein estimated without the adenine nucleotide translocase.


pmol/min=mg of phosphate transport protein estimated without the adenine nucleotide translocase.
Inorganic phosphate must be transported into the mitochondrial matrix for steady state oxidative phosphorylation of ADP to occur. We recently reported the purification of a protein from beef heart mitochondria which upon incorporation into liposomes catalyzed phosphate/phosphate exchange (1, 2). The exchange is sensitive to the known mitochondrial phosphate transport inhibitors but not to those of dicarboxylate or adenine nucleotide transport (2).
We show in this communication that the purified phosphate Grant AGO0100 and Grant PCM 8003916 from the National Science * This investigation was supported by National Institutes of Health Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. transport protein co-purified with the adenine nucleotide translocase catalyzes, after incorporation into liposomes, phosphate transport with characteristics similar to those of the protein in intact mitochondria (3-5). We specifically demonstrate that it can catalyze unidirectional phosphate transport, that the amount of phosphate accumulated depends on the intravesicular pH, and that the K , and V,,, are similar to those of mitochondria.

EXPERIMENTAL PROCEDURES
Purification of the Phosphate Transport Protein-Mitochondria were prepared from beef hearts according to published procedures (6) and stored at a concentration of 13 nmol of cytochrome b/ml (-40 "C). The cytochrome b concentration was determined as described (7). All operations were carried out at 4 "C unless noted otherwise. Thawed mitochondria (6 ml) were Centrifuged (17,000 X g, 10 rnin), and the pellet was suspended in 80 ml of medium S (10 mM sodium phosphate, 0.33 m~ EDTA, pH 7.2), kept on ice for 10 min, and centrifuged (17,000 X g, 10 min). The pellet was suspended in 80 ml of medium A (10 mM sodium phosphate, 0.1 mM EDTA, 130 m~ NaCl, 5 mM dithiothreitol, pH 7.2) and centrifuged (17,000 X g, 10 min). The pellet was resuspended in medium A and centrifuged once more. The supernatant was completely removed and the pellet suspended in medium A to yield a final volume of 5 ml. 1 ml of a 16% Triton X-100 (Sigma) solution in medium A was added while vortexing the suspension. After centrifugation (166,000 X g, 10 min, Beckman 50-Ti rotor), 1.6 g of washed SM-2 Bio-Beads (Bio-Rad) (8) were added to the supernatant. The suspension was stirred for 30 min. The SM-2 Bio-Beads decrease the Triton X-100 concentration to 1.4%. The supernatant, after the Bio-Beads had settled out, was put on a hydroxylapatite column (1 X 8.5 cm with 15 cm of tubing attached to the lower end to increase the flow rate) and fractions (about 1.2 m l ) were collected at intervals of 1.5 min. The sample was eluted with medium A which contained 0.5% Triton X-100. The hydroxylapatite (hydroxylapatite, fast flow, dry powder, Calbiochem-Behring) had been suspended in and the column equilibrated with medium B (10 m~ sodium phosphate, 0.1 mM EDTA, 130 m~ sodium sulfate, pH 7.2). The fractions with high absorbance (275 nm) were pooled. 500 pl of a freshly prepared lipid dispersion (46 mg/ml) (see below) were added immediately, the mixture was vortexed, and the air was flushed out of the test tube with high purity argon. Washed SM-2 Bio-Beads (1 g/2-ml sample) were added and the mixture was stirred for 60 min. 200-pl fractions were pipetted into argon-flushed 1.5-ml cryotubes (Vangard International, Neptune, NJ) and stored in a liquid nitrogen freezer.
The phosphate transport protein inhibited by N-ethylmaleimide was prepared from beef heart mitochondria with a modified purification procedure. The mitochondrial pellet from the first centrifugation was suspended in 30 ml of medium E (100 mM 4-morpholinepropanesulfonic acid, 0.2 mM ethylene glycol bis(P-aminoethyl ether)-N,N,N',N'-tetraacetic acid, 2 mM dithiothreitol, adjusted to pH 7.2 with NaOH) and centrifuged (17,000 x g, 10 rnin). The pellet was suspended to 10 ml with medium E, and 500 pl of 100 mM Nethylmaleimide in water was added. After 3 min on ice, the suspension was centrifuged (17,000 X g, 10 min). The pellet was taken up in 80 ml of medium S and the procedure was finished as described above.
The amount of phospholipid in each ampule was kept small to permit complete removal of solvents by blowing high purity argon into each ampule. To facilitate further solvent removal, 100 pl of diethyl ether were added to dissolve the lipids and the solvent was again removed by blowing argon into the ampules. This last step was repeated with 100 pl and then with 50 pl of diethyl ether to reduce the size of the lipid film to a minimum. The ampules were placed overnight into a high vacuum chamber (<1 p ) to remove remaining traces of solvent.

29
Commercially lyophilized phosphatidic acid (5 mg) was transferred into an empty ampule and 200 p1 of medium C, (10 m~ Tris base, 2 m~ MgC12, 10 m~ 1,4-piperazinediethanesulfonic acid, various amounts of phosphate for the intraliposomal concentration, pH was adjusted with 10 or 3 N KOH) was added. The phosphatidic acid was dispersed in a bath-type sonicator (2). 80 p l of this solution plus 170 pl of medium Ci were added to an ampule with 9.5 mg of solvent-free lipids (phosphatidylethanolamine and phosphatidylcholine).
The mixture was sonicated for 2 min at 20 "C and the pH was adjusted with sodium hydroxide. The mixture was briefly sonicated once more. This lipid dispersion was prepared fresh daily and used for the treatment of the anion exchange column (see below) and in the preparation of the liposomes that catalyze phosphate transport.
Medium A (250 pl) was added to the ampules with the solvent-free phosphatidylethanolamine/phosphatidylcholine/cardio~pin mixture, the lipids were dispersed, and their pH was adjusted. This lipid mixture was used in the phosphate transport protein purification procedure.
Incorporation of the Phosphate Transport Protein into Liposomes-The thawed solution of phosphate transport protein was passed through a small P2 column (Bio-Rad) (2) equilibrated with medium Ci plus 5 m~ dithiothreitol. Incorporation of the transport protein into liposomes was carried out by adding 80 pl of the phosphate transport protein (10 to 15 pg of protein) to a mixture of 100 pl of medium Ci plus 80 pl of liposomes in medium Ci (see above). The mixture was frozen, thawed (2, 9), and sonicated in small plastic test tubes under argon for 10 s at 13 "C in the bath-type sonicator (2).
Phosphate Transport Assays-25 pl of liposomes into which phosphate transport protein had been incorporated (0 "C) were added to 500 pl of reaction medium (medium C, 22 "C) to which carrier-free 32P, (1.7 X lo4 Bq) and various amounts of phosphate had been added.
The pH of medium C was at the pH, (extraliposomal) before phosphate was added and was readjusted after the phosphate addition. The phosphate transport was stopped with mersalyl (10 pl, 79 mM), which was prepared by adding mersalylic acid to water followed with only enough NaOH to dissolve the mersalyl. Zero time points were obtained by adding the liposomes to the reaction medium plus mersalyl. 30 s later, the mixture was placed on an anion exchange column (see below) (4-10 "C) and eluted with 3 ml of medium D (5% glycerol, 0.1 m M sodium azide). The eluate (3.5 ml) was collected in one scintillation vial, 8 ml of scintillation mixture (Liquiscint, National Diagnostics) were added and the vial was vortexed and counted. For the other time points, mersalyl was added t s after the liposomes; (30 t ) s later, the sample was placed on the anion exchange column.
The anion exchange column (2) (AG 1-X8, formate, 50-100 mesh, BioRad) was treated with 500 p1 of bovine serum albumin (30 mg/ ml), 100 p1 of liposomes (46 mg/ml of medium Ci), 520 p1 liposomes (1.77 mg/ml of medium C,) and 500 pl of liposomes (1.77 mg/ml of medium C,) before use. The affinity of the column for Pi was monitored by passing reaction mixtures without liposomes through the column after every second or third reaction mixture with liposomes. From a reaction mixture without liposomes and an acceptable commercial 3zP1 sample (1.7 X lo4 Bq) put through the column, less than 1.7 Bq eluted. With some unacceptable commercial "P, samples, 34 Bq eluted.
Other Methods-Protein was determined in the presence of sodium dodecyl sulfate and N-ethylmaleimide using the Lowry method (IO).

RESULTS AND DISCUSSION
Purification of the Phosphate Transport Protein-The mitochondrial phosphate transport protein was originally purified by centrifuging a mixture of hydroxylapatite and solubilized mitochondria. The pure protein together with the adenine nucleotide translocase was recovered from the supernatant (2). Since much of the purified protein was lost in the liquid trapped in the hydroxylapatite pellet, we increased the yield by passixg the solubilized mitochondrial membrane through a hydroxylapatite column. This preparation also consists only of the phosphate transport protein and the adenine nucleotide translocase (11). This column procedure takes more time and we therefore developed conditions that protect the active preparation during storage in a liquid nitrogen freezer. Freezing the phosphate transport protein with or without added phospholipids results in an 80% loss of reconstituted transport activity. A 30to 60-min exposure to SM-2 Bio-Beads after the addition of phospholipids and before freezing completely protected the transport activity from inhibition by freezing. Phosphate/Phosphate Exchange Rates-We extended our phosphate/phosphate exchange studies (1, 2) in order (a) to increase the reconstituted activity and ( b ) to obtain kinetic properties similar to those reported for mitochondria (5). We modified our original assay system (2) by stopping the phosphate/phosphate exchange at various times with mersalyl.
Mersalyl appears to be less competitive with phosphate than some other transport inhibitors and is thus more suitable for initial transport rate studies. We discovered that high mersalyl

30
Reconstituted Mitochondrial Phosphate Transport Activity locase. We expect therefore both proteins to bind similar amounts of Coomassie blue in sodium dodecyl sulfate-polyacrylamide gels. The stained slab gel was scanned with an LKB scanner and from the resulting integrated absorbance curve we estimated that the preparation reported in this communication consists of 25 to 30% phosphate transport protein and of 70 to 75% adenine nucleotide translocase. The VmaX for this amount of phosphate transport protein becomes 48 to 60 pmol of Pi/min.mg with a turnover number of 1630 to 2040 min". This turnover number is not too different from the 3500 min" (0 "C) determined by Coty and Pedersen (5) with rat liver mitochondria on the basis of some very general assumptions.
Unidirectional Phosphate Uptake-The liposomes were prepared with medium Ci either at pH 8.0 or 6.8 with no phosphate. The reaction medium (500 pl) contained 1 mM sodium phosphate at pH 6.8 or 8.0. Since phosphate uptake is essentially completed within 30 s, the transport was stopped only at 0 and 30 s. Our controls consisted of liposomes incorporated with the same amount of protein isolated from mitochondria inhibited by N-ethylmaleimide. Fig. 2, a and c, clearly shows that at an initial external pH of 6.8 and an internal pH of 6.8 much less Pi is taken up in 30 s than at an initial internal pH of 8.0. Uptake in the presence of transport protein inhibited by N-ethylmaleimide is almost completely absent in both cases. When, however, the external pH is raised to 8.0 (Fig. 2, b and d), the phosphate uptake in the presence of protein inhibited by N-ethylmaleimide or active protein is the same. This suggests that either the liposomes have become more leaky and phosphate is trapped by mersalyl (as discussed earlier) or that a phosphate transport pathway insensitive to N-ethylmalehide but sensitive to mersalyl is activated at the higher external pH in the protein.
Since, however, liposomes prepared without any protein show almost the same uptake pattern at the different external and internal pH values as the liposomes with the protein inhibited by N-ethylmaleimide, we conclude that the liposomes are more permeable to phosphate at the higher external pH.  The experiment was carried out as in Fig. 2c except that the transport was stopped also at times shorter than 30 s. The phosphate trapped by liposomes incorporated with phosphate transport protein inhibited by N-ethylmaleimide was subtracted. prepared with protein inhibited by N-ethylmaleimide. The initial uptake rate is about 2 pmol of Pi/min mg of protein or 8 pmol/min.mg of phosphate transport protein estimated without the adenine nucleotide translocase.
While many more studies have to be carried out with this protein in its natural or synthetic membrane environment, we can conclude from the present set of results that the phosphate transport protein catalyzes unidirectional phosphate transport that is stimulated by a membrane pH gradient.