Solubilization and Reconstitution of the Lactose Transport System from Escherichia coli*

The lactose transport system from Escherichia coli was solubilized with octylglucoside and reconstituted into liposomes by an octylglucoside dilution procedure. The reconstituted proteoliposomes exhibited lactose counterflow and membrane potential-driven lactose transport.

The lactose transport system from Escherichia coli was solubilized with octylglucoside and reconstituted into liposomes by an octylglucoside dilution procedure. The reconstituted proteoliposomes exhibited lactose counterflow and membrane potential-driven lactose transport.
The lactose transport system of Escherichia coli is responsible for the active transport of P-galactosides into the cell (1). In 1963, Mitchell (2) postulated that this system functions as a proton-substrate co-transport system which is coupled to the metabolism of the cell via the transmembrane electrochemical proton gradient. Studies with intact cells and cytoplasmic membrane vesicles have provided strong support for this concept (see Ref. 3 for review). In addition, the kinetics (4-6) and substrate specificity (7) of the transport system have been extensively studied. The lac y gene, which codes for the lactose transport protein, has been cloned on a bacterial plasmid and the nucleotide sequence of the gene has been determined (8). The transport protein has been purified in an inactive form (9).
Despite the fact that the lactose carrier represents one of the most extensively characterized active transport systems, little is known about its subunit structure, the molecular mechanism of active transport, or the mechanism by which the lactose carrier is regulated by the phosphotransferase system (10). Reconstitution of the carrier would provide an assay for the purification of the protein@) responsible for /3galactoside transport. Furthermore, reconstitution of a purified transport system would greatly facilitate the determination of the molecular mechanisms of carrier function and regulation. Solubilization and reconstitution of lactose transport into transport-negative membrane vesicles have been reported (11). In addition, Padan et al. (12) have reported that lactose transport activity is lost after extraction of membrane vesicles with sodium cholate, and that transport activity in these vesicles is restored upon addition of exogenous phospholipid followed by detergent removal. However, previous attempts to solubilize and reconstitute the lactose transport * This work was supported by Public Health Service Grant AM05736 from the National Institute of Arthritis, Metabolism and Digestive Diseases and Grant PCM 78-00859 from the National Science 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.
$ Supported by a National Science Foundation Predoctoral Fellowship. To whom correspondence should be addressed. system into liposomes have been unsuccessful.
Several reports on the reconstitution of bacterial cationsubstrate co-transport systems have appeared (13-16), one of which deals with the reconstitution of proline transport from E . coli (13).
This communication describes the solubilization of the lactose transport system and its reconstitution into liposomes prepared from E. coli phospholipid. The reconstituted proteoliposomes exhibited lactose counterflow and membrane potential-driven lactose transport.
Preparation of Acetone/Ether-washed E. coli Lipid-Chloroform/methanol-extracted E. coli lipid was acetone/ether-washed by a modification of the method of Kagawa and Racker (19). Crude lipid extract, 50 ml of chloroform/methanol (91), containing 1 g of lipid, was evaporated to 5 ml under a stream of Nz gas. The material was suspended in 100 ml of Nz-bubbled anhydrous acetone containing 2 mM 2-mercaptoethanol. The suspension was placed in a light-protected 250-ml flask under NZ gas and stirred at low speed (on top of a Styrofoam block) on a magnetic stirrer for 12 h at room temperature. The extract was filtered through Whatman No. 1 paper on a Buchner funnel with suction, and the insoluble material was scraped off the filter paper and immediately resuspended (with stirring) in 100 ml of anhydrous ether containing 2 mM 2-mercaptoethanol. This suspension was centrifuged in a glass bottle at 2500 X g (4000 rpm) for 15 min.
The supernatant was carefully decanted and placed in a small flask and then evaporated to 10 ml under a stream of Nz gas. The solution was transferred to a preweighed test tube, and the remaining ether was evaporated as above. A small amount of chloroform was added to the tube, and the lipid was dispersed as a film on the bottom and sides of the tube by rotating the tube under a stream of Nz gas. The lipid was lyophilized for 3 h to remove remaining solvent. The tube containing the lipid was weighed, and the lipid was suspended in 2 mM 2-mercaptoethanol at 50 mg/ml. The lipid was Vortex-dispersed while under NZ gas, and the suspension was stored in 1-ml aliquots under N2 gas at -80°C. One-half of the starting material was recovered after acetone/ether washing.

10583
Reconstitution of Lactose Transport System X-100, and the level in the bath was adjusted to give maximal agitation of the solution (20).
Preparation of High Pressure French Press Vesicles-Cells (1.5 to 3 liters) were grown to late logarithmic phase with continuous shaking in modified medium 63 (21), which consisted of 0.1 M potassium phosphate, pH 7.0,15m~ ammonium sulfate, and0.8 mM MgSO,. 7 H,O. The medium was supplemented with 0.4% glycerol. The cells or membrane vesicles were kept at 4°C during all subsequent steps. The cells were harvested and washed once with modified medium 63. The cells were resuspended (0.2 g wet weight/ml) in 50 mM potassium phosphate, pH 7.5, 1 mM dithiothreitol, 20 mM lactose, 5 mM MgS0, and then hand homogenized before the addition of 20 pg/ml of DNase and 0.5 mM phenylmethylsulfonyl fluoride. In some initial experiments, 0.5 mM phenanthroline (phenylmethylsulfony1 fluoride and phenanthroline are both protease inhibitors) was also added, but the presence or absence of this inhibitor did not affect the reconstitution.
Cells were disrupted by passage through an Aminco French pressure cell (model 4-3398) at 19,000 p s i . total pressure and collected in a tube in an ice bath. Unbroken cells were removed by centrifugation at 11,700 X g (10,000 rpm) for 10 min. The supernatant was carefully removed and the centrifugation was repeat.ed. The resulting supernatant was removed and the vesicles were sedimented by ultracentrifugation at 140,000 X g (45,000 rpm) for 2 h. The pellet was resuspended in 50 mM potassium phosphate, pH 7.5, 1 mM dithiothreitol, 20 mM lactose, 0.5 mM phenylmethylsulfonyl fluoride, and the centrifugation was repeated. The vesicles were resuspended in the same buffer at a protein concentration of 30 to 70 mg/ml. Vesicles were divided into 50-pl aliquots, frozen in liquid N2, and stored at Reconstitution of Lactose Transport-Lactose transport was reconstituted by the octylglucoside dilution procedure of Racker et al. (22). Two hundred ninety microliters of 50 mM potassium phosphate, pH 7.5, 5 p1 of 100 mM dithiothreitol, 17 pl of 580 mM lactose, 7.3 pl of high pressure vesicles (0.5 mg of protein) and 40 pl of acetone/etherwashed E. coli lipid (from 50 mg/ml of stock) were added to a small test tube in an ice bath, and the tube was blended on a Vortex mixer. Octylglucoside (33 p1 of a 15% solution in 50 mM potassium phosphate, pH 7.5) was added (final concentration, 1.25%) and the tube was blended on a Vortex mixer. The suspension was incubated at 4'C for 10 min, blended on a Vortex mixer again, and then centrifuged at 175,000 X g (45,000 rpm) (4OC) for 1 h. The supernatant (containing 0.2 to 0.3 mg of solubilized membrane protein) was carefully removed with a Pasteur pipette and mixed with 100 pl of bath sonicated liposomes (5 mg of lipid) plus 8.3 p1 of 15% octylglucoside (final concentration, approximately 1.25%). The mixture was blended on a Vortex mixer and then incubated at 4°C for 10 min. The suspension was then pipetted directly into 15 ml of 50 mM potassium phosphate, pH 7.5, 20 mM lactose, 1 mM dithiothreitol at room temperature, and the tube was blended on a Vortex mixer gently. The resultant proteoliposomes were sedimented by ultracentrifugation in a TY 42.1 rotor at 85,000 X g (35,000 rpm) for 1.5 h. The supernatant was decanted and the tube was wiped with a cotton-tipped applicator. The pellet was then resuspended, for counterflow, in 200 p1 (final volume) of 50 mM potassium phosphate, pH 7.5, 20 mM lactose, 1 mM dithiothreitol. Resuspension was carried out by stirring the pellet with a glass rod and squirting the suspension up and down three times with a 5O-pl Hamilton syringe. For membrane potential-driven lactose transport, lactose was eliminated from all reconstitution steps and the proteoliposomes were resuspended in 50 1.1 of 50 mM sodium phosphate, pH 7.5, 1 mM dithiothreitol. Reconstituted proteoliposomes were stored at 4OC. Membrane potential-driven uptake was assayed immediately. Counterflow was assayed within 12 h.
for counterflow by diluting 9 pl (2.3 pg of protein) of lactose-loaded Transport Assays-Reconstituted proteoliposomes were assayed proteoliposomes into 0.45 ml of 50 mM potassium phosphate, 2 mM MgS04 containing 0.9 pCi of ['4C]lactose (counterflow assay buffer). The final lactose concentration was 0.43 mM. Transport was carried out at room temperature. The tube was blended on a Vortex mixer and, at various times, 0.1-ml samples were removed and filtered bopwise onto the center of a 0.22-pM Millipore filter (type GSTF) using 25 inches of mercury vacuum suction. The filter was washed with 5 ml of ice-cold 50 mM potassium phosphate and counted in 4 ml of Bray's scintillation fluid at a I4C efficiency of 86%. A blank value, obtained by filtering 0.1 ml of assay buffer without proteoliposomes, was subtracted from all points.
Membrane potential-driven lactose transport in reconstituted proteoliposomes was assayed by diluting 7 pl of potassium phosphateloaded proteoliposomes (5.4 pg of protein) into 0.7 ml of 50 mM -80°C.
sodium phosphate, pH 7.5, containing 1.4 pCi of ['4C]lactose. The final lactose concentration was 0.23 mM. Transport was carried out at room temperature. The tube was blended on a Vortex mixer, and 0.1ml samples were removed at 15-and 45-s time points, fdtered as above, washed with 5 ml of ice-cold 50 mM sodium phosphate, and counted as above. At 60 s, 7 p1 of 1 mM valinomycin was added (to give a final concentration of 14 p~) , and the tube was blended on a Vortex mixer. At various times, 0.1-ml samples were removed, filtered, washed with 5 ml of ice-cold 50 mM sodium phosphate, and counted as above.
Reconstituted proteoliposomes (lactose-preloaded), approximately 1.5 mg of lipid phosphate, 20 pg of protein in 20 pl, were suspended in 0.5 ml of low density solution and layered on one gradient (42% sucrose cushion). High pressure French press vesicles, approximately 0.6 mg of lipid phosphate, 1 mg of protein in 14.5 pl, were suspended in 0.5 ml of low density solution and layered on top of a second, similar gradient (60% sucrose cushion). The gradients were centrifuged for 2.5 h at 58,000 rpm in an SW 65 rotor (4"C).
Protein Assays-Protein was determined by the modified Lowry procedure of Peterson (23). The microassay was used, without precipitation.
TotaZPhosphate-Total phosphate was determined by the method of Ames (24). The samples were ashed in a box type furnace at 600°C.

RESULTS AND DISCUSSION
When E. coli membrane vesicles were treated with 1.25% octylglucoside in the presence of exogenous acetone/etherwashed E. coli lipid, as described under "Experimental Procedures," 40 to 60% of the membrane protein was solubilized. This step was followed by a high speed centrifugation to remove all unextracted membrane material. The supernatant was then exposed to liposomes and diluted 30-fold into detergent-free buffer. In the present experiments, 15 to 20% of the solubilized protein was recovered in the reconstituted proteoliposomes centrjfuged following the dilution step. This represents 7.5 to 10% of the original protein in the French press vesicle material.

when preloaded with nonradioactive lactose. The counterflow phenomenon is observed when radioactive substrate is transported into a cell or vesicle and accumulates because its exit is competitively inhibited by a high internal concentration of unlabeled substrate. The subsequent rate of loss of the accumulated radioactive substrate has been shown to be dependent on the number of carriers per cell (25). The time course for counterflow in the reconstituted proteoliposomes (one-half of the accumulated lactose was retained after
70 min) and the lipid to protein ratio used for reconstitution (approximately 201) are consistent with the hypothesis that each active proteoliposome contains only one or a few active lactose carriers. Proteoliposomes which were not preloaded with lactose exhibited very low transport activity (Fig. 1, inset). Lactosepreloaded proteoliposomes assayed for counterflow in the presence of thiodigalactoside (a competitive inhibitor of lactose transport) also exhibited very low transport activity. N-Ethylmaleimide-treated lactose-preloaded proteoliposomes exhibited virtually no transport activity. The same result was obtained with proteoliposomes prepared with solubilized membrane protein from uninduced E. coli ML3 (lac y-) French press vesicles.  FIG . 2 (center). Membrane potential-driven lactose transport in reconstituted proteoliposomes. Reconstitution was carried out and membrane potential-driven transport was assayed as described under "Experimental Procedures." M , proteoliposomes were diluted into 50 mM sodium phosphate, pH 7.5, and valinomycin (final concentration, 14 p~) was added at 60 s, the time indicated by the arrow (the 60-s time point was obtained in a separate experiment, using the same batch of proteoliposomes); M , proteoliposomes were diluted into sodium phosphate, pH 7.5, containing 20 p~ CCCP, and valinomycin (14 p~) was added at 60 s; [1"(3, proteoliposomes were diluted into 50 mM potassium phosphate, pH TABLE I

Effect of lipid on reconstitution of lactose transport
Lactose counterflow in the presence and absence of 10 mM thiodigalactoside was measured at I-, 5-and 15-min time points as described under "Experimental Procedures." For Experiment 1, reconstitution was carried out as described under "Experimental Procedures." For Experiment 2, no E. coli lipid was added at the time of solubilization.
A 40-pl sample of 50 mM potassium phosphate, pH 7.5, was added to maintain equivalent volumes. For Experiment 3, asolectin was substituted for E. coli lipid at both the solubilization and detergent dilution (liposome) steps of the reconstitution. The values presented in the table represent the 15-min time point, which corresponded to the peak level of transport in each experiment. at the membrane solubilization step was required for lactose transport reconstitution. In addition, when asolectin was substituted for acetone/ether-washed E . coli lipid in the reconstitution, only 15% of the E . coli lipid reconstitution activity was recovered. Energy-depleted whole cells and French press vesicles accumulate lactose against a concentration gradient when a membrane potential, negative inside, is generated across the cytoplasmic or vesicle membrane (26). This can be accom-0 1 1 4-I 7.5, and valinomycin (14 p~) was added at 60 s; M , proteoliposomes were diluted into 50 mM sodium phosphate, pH 7.5, and gramicidin (14 p~) , instead of valinomycin, was added at 60 s;

A-A,
proteoliposomes were treated with 4 mM N-ethylmaleimide for 10 min at room temperature prior to dilution into 50 r n~ sodium phosphate, pH 7.5, and valinomycin (14 p~) was added at 60 s. FIG. 3 (right). Isopycnic density gradient centrifugation of reconstituted proteoliposomes and French press vesicles. Gradients were prepared and assays were carried out as described under "Experimental Procedures." Thirteen 0.37-ml fractions were collected from each gradient. A 25-pl aliquot from each of the French press vesicle gradient fractions was assayed for phosphate. In this gradient, the lipid phosphate peak was found in the Fist fraction (indicated by the arrow). Aliquots of 10 p1 from each of the proteoliposome gradient fractions were assayed for phosphate (M). The phosphate peak observed at the top of the gradient corresponds to potassium phosphate buffer placed on the gradient with the proteoliposomes. Each fraction was diluted into 15 ml of 50 mM potassium phosphate, pH 7.5, 20 mM lactose, I mM dithiothreitol. Proteoliposomes were then sedimented by ultracentrifugation of the suspension at 85,000 X g (35,000 rpm) for 1.5 h. Fraction 7 was the only fraction which yielded a pellet upon centrifugation. This pellet was resuspended in 15 p1 of the above buffer, and a 6-pl aliquot was assayed for counterflow at a 2-min time point (M). No lactose uptake was observed when a second 6-pl aliquot was assayed in the presence of 10 mM thiodigalactoside.
plished by diluting potassium-containing cells or vesicles into low potassium medium and adding valinomycin, thus giving rise to a potassium diffusion potential. Membrane potentialdriven lactose accumulation was demonstrated in the reconstituted proteoliposomes (Fig. 2). When potassium phosphateloaded proteoliposomes were resuspended and diluted in sodium phosphate plus radioactive lactose, uptake of [14C]1actose was observed (approximately 10 nmol of lactose/mg of protein at 60 s). The addition of valinomycin (arrow, Fig. 2) resulted in a transient 5-fold accumulation of radioactive lactose. When a similar experiment was carried out in the presence of the proton conductor CCCP, the addition of valinomycin did not give rise to transient lactose accumulation. The CCCP would be expected to block membrane potential-driven lactose accumulation, since this ionophore collapses the protonmotive force generated by the membrane potential. Dilution of potassium phosphate-loaded proteoliposomes into sodium phosphate and addition of gramicidin instead of valinomycin also resulted in no lactose accumulation. When potassium phosphate-loaded proteoliposomes were diluted into potassium phosphate instead of sodium phosphate, the addition of valinomycin resulted in no lactose accumulation, since no potassium diffusion potential was produced. In the preceding three experiments, lactose presumably equilibrated between the inside and outside of the proteoliposomes containing active transport protein. This view is confirmed by the experiment (Fig. 2) in which proteoliposomes were pretreated with N-ethylmaleimide, which inactivated the carrier and almost completely prevented entry of sugar into the proteoliposomes.

Reconstitution of Lactose Transport System
The possibility was considered that the lactose transport activity was not solubilized, but that intact French press vesicles were retained during the reconstitution procedure and were responsible for the final transport activity. However, no activity was found when lactose-loaded high pressure French press vesicles were assayed for counterflow under the same conditions as the proteoliposome counterflow assay (data not shown). Lancaster and Hinkle (26) have shown that French press vesicles pass through 0.22-pM Millipore filters, unless the vesicles are aggregated with (po1y)lysine prior to fitration. The absence of activity (by the counterflow assay) in the high pressure French press vesicle preparation used for reconstitution argues against the possibility that the reconstitution transport activity is due to contaminating French press vesicles.
A second possibility was that hybrid vesicles might be formed due to the introduction of exogenous lipid into detergent-disrupted French press vesicles or that hybrid vesicles were formed from large fragments of native membrane and exogenous lipid. This possibility was examined by density gradient centrifugation of the reconstituted vesicles and native French press vesicles in a manner similar to that utilized by Papazian et al. (27). Native membrane material, containing approximately equal amounts of protein and lipid, should sediment to a much denser region of the gradient than liposomes or proteoliposomes containing only a few protein molecules per vesicle. Two similar gradients were prepared. Onto the surface of one was placed reconstituted vesicles; the other gradient received high pressure French press vesicles. After centrifugation, 13 fractions were collected from each gradient. As shown in Fig. 3 (arrow), the peak of lipid phosphate in the French press vesicle gradient was found in the first fraction, which included the 60% sucrose cushion. This position in the gradient corresponds to a density of between 1.09 and 1.29 g/ ml. Each fraction from the proteoliposome gradient was assayed for phosphate (closed circles) and then diluted and centrifuged at high speed to pellet any intact proteoliposomes. Only Fraction 7 contained a pellet following this procedure. This pellet was found to contain lactose transport activity (open circle). The position of lactose transport activity on the gradient closely parallels the position of the only lipid phosphate peak (artificial vesicles) and corresponds to a density of 1.046 g/ml, which is close to the theoretical value for pure phospholipid vesicles (1.03 g/ml). This experiment provides strong evidence that the lactose transport system has been solubilized and transferred into an artificial vesicle or proteoliposome.
A preliminary kinetic analysis of initial rates of lactose counterflow in the reconstituted proteoliposomes (data not shown) has yielded maximum velocities of 0.5 to 1.0 pmol of lactose/min/mg of protein. These values are 5 to 10 times higher than the maximum velocity (100 nmol/min/mg of protein) published for lactose counterflow in membrane vesicles (28). Considering that only 7.5 to 10% of the original French press vesicle protein was recovered in the reconstituted proteoliposomes, these results suggest that a partial purification of the lactose transport system may have been achieved.
One of the interesting features of these experiments is the finding that the addition of acetone/ether-washed E. coli lipid at the solubilization step and the use of this material in the detergent dilution procedure significantly improves the activity of the reconstituted proteoliposomes. A requirement for exogenous phospholipid at the time of solubilization has also been demonstrated by Maron et ai. (29) for the reconstitution of a catecholamine transporter from bovine chromaffin granules. The present reconstitution method should be useful as an assay for the purification of an active lactose transport system and make possible the study of several interesting aspects of the molecular mechanism of lactose carrier function.