Purification and Reconstitution of Functional Lactose Carrier from Escherichia coli *

The lactose carrier protein of Escherichia coli was purified by a simple procedure employing differential solubilization and ion-exchange chromatography and reconstituted into liposomes by octylglucoside dilution. The proteoliposomes exhibited both membrane poten-tial-driven lactose transport and lactose counterflow. Furthermore, the purified protein was identified as the product of the lac y gene. These and other results demonstrate that the lactose carrier is the only polypeptide species essential for energy-coupled lactose transport and counterflow.

The lactose carrier protein of Escherichia coli was purified by a simple procedure employing differential solubilization and ion-exchange chromatography and reconstituted into liposomes by octylglucoside dilution. The proteoliposomes exhibited both membrane potential-driven lactose transport and lactose counterflow. Furthermore, the purified protein was identified as the product of the lac y gene. These and other results demonstrate that the lactose carrier is the only polypeptide species essential for energy-coupled lactose transport and counterflow.
The transport of /3-galactosides across the cytoplasmic membrane of Escherichia coli is mediated by the lactose transport system (1). During the last decade, a large body of evidence has accumulated to support Mitchell's hypothesis (2) that sugar-proton symport by this system is driven by a transmembrane electrochemical gradient of protons (see Ref. 3 for review). The kinetics, substrate specificity, and genetics of this transport system have been extensively studied, and in addition, Kennedy and his collaborators (4, 5) have identified a membrane-bound protein as the product of the lac y gene. Most recently, the lac y gene has been cloned in a bacterial plasmid and its product amplified (6). The sequence of the carrier protein was deduced from the DNA sequence, and the purification of the carrier in an inactive form was reported (7). Although it has been demonstrated that the lactose carrier plays an essential role in lactose transport, evidence has been presented suggesting that an additional component(s) may be required for energy-coupled transport (8).
The elucidation of the subunit composition and molecular mechanism of the lactose transport system will require the purification and reconstitution of the protein(s) responsible for lactose transport. A step in this direction was the recent solubilization and reconstitution of the lactose transport system by Newman and Wilson (9). Furthermore, Kaczorowski et al. (10) were able to label the lactose carrier specifically with the photoaffinity reagent 4-nitro[2-'H]phenyl-a-~-gdactopyranoside. Using these techniques in concert, we have now purified a single protein that was identified as the product of the lac y gene on the basis of its activity, inducibility, electrophoretic mobility, and amino acid composition.
* This work was supported in part 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 to T. Hastings Wilson. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
3 Supported by a National Science Foundation Predoctoral Fellowship and a Fellowship from the Albert J. Ryan Foundation.

EXPERIMENTAL. PROCEDURES
Bacteria-E. coli T206 (12), which carries the lac y gene in a recombinant DNA plasmid, was provided by Dr. Peter Overath, Max-Planck-Institut fur Biologie, Tdbingen, West Germany.
Materials-Sodium cholate and octylglucoside' were obtained from Calhiochem. DEAE-Sepharose CL-GB was from Pharmacia. Urea (ultra-pure) was from Bethesda Research Laboratories. Crude chloroform/methanol-extracted E. coli lipids were either purchased from Avanti Biochemicals or prepared from E. coli by the procedure of Ames (11). Filter chimneys (9-mm internal diameter) were custom made from plexiglass. Electrophoresis reagents and molecular weight standards were from Bio-Rad. ['HINPG was synthesized by Yu-Ying Lui (Isotope Synthesis Group, Hoffmann-La Roche) under the direction of Arnold Liebman and had a specific activity of 30 Ci/mmol (1 Ci = 3.7 X 10"' becquerels). All other materials were obtained as described (9).
Preparation ofMembrane Vesicles-Cells were grown and induced for the lactose transport system as described by Teather et al. (12). Vesicles were prepared either by osmotic lysis (rightside-out; Refs. 13 and 14) or by passage through a French pressure cell (9).
Purification of the Lactose Carrier-All steps were performed at 0-4 "C with 10-35 mg of membrane protein as starting material. T206 membrane vesicles were mixed in a 98:2 (protein/protein) ratio with [3H]NPG-labeled membranes. The specific activity of the mixture was 50-100 nCi/mg of membrane protein. In a typical experiment, 12.5 mg of protein of the above mixture was adjusted to 10 mg of protein/ml in 50 mM potassium phosphate, pH 7.5, 0.5 mM dithiothreitol, 10 mM lactose. An equal volume of 10 M urea (room temperature) was added dropwise while blending on a Vortex mixer. This suspension was incubated for 10 min on ice and centrifuged at 175,000 x g for 1 h. The pellet was resuspended (using a glass rod and a 1-ml syringe) in a final volume of 1.75 ml of 50 mM potassium phosphate, pH 7.5. Sodium cholate (20% w/v, pH 7.8) was added to a final concentration of 6% (w/v) while blending on a Vortex mixer. This suspension was incubated on ice for 20 min and centrifuged at 26,000 X g for 15 min. The pellet was resuspended in 5 ml of 10 mM potassium phosphate, pH 5.8, and centrifuged again at 26,000 X g for 15 min. The pellet (occasionally stored under 1 ml of the above buffer overnight on ice) was resuspended in 1.45 ml of 10 mM potassium phosphate, pH 5.8, and this suspension was mixed with 17.5 pl of 100 mM dithiothreitol, 13 mg of lactose, 131 pl of washed E . coli phospholipid (50 mg/ml) and blended on a Vortex mixer. Octylglucoside (146 p1 of a 15% (w/v) solution in 10 mM potassium phosphate, pH 5. Vortex mixer, incubated for 10 min on ice, blended again, and then centrifuged at 175,000 X g for 1 h. The supernatant solution was removed and adjusted to pH 5.8 with 10 mM H J P O~ containing 1.25% octylglucoside.
DEAE-Sepharose was prewashed once with 10 volumes of 1 M potassium phosphate, pH 5.8, and 8 times with 10 volumes of 10 mM potassium phosphate, pH 5.8. A 6-ml column (0.9 X 9 cm) was prepared and equilibrated with 10 mM potassium phosphate, pH 5.8. Equilibration was ensured by measuring the pH and conductivity of the effluent. The column was then washed with 18 ml of column buffer: IO mM potassium phosphate, pH 5.8, 1 mM dithiothreitol, 20 mM lactose, 0.25 mg of washed E . coli lipid/ml, 1.25% octylglucoside. One ml of the octylglucoside extract (approximately 300 pg of protein) was loaded on the column,2 and the lactose carrier was eluted with column buffer at a flow rate of 15 ml/h. Fifteen 1-ml fractions were collected. Fractions were stored at 0 "C and were active for reconstitution for at least 1 week.
Reconstitution of Lactose Transport-A rapid filter assay was developed for the determination of carrier activity in column fractions. The procedure utilizes 1/6 the amount of lipid as the previous method (9) and eliminates the proteoliposome centrifugation step. All steps were carried out at room temperature unless otherwise noted. Bathsonicated liposomes (16.5 pl), prepared as described (9), were mixed with 1.4 p1 of 15% octylglucoside and 55 pl of an octylglucoside membrane extract or column fraction, blended on a Vortex mixer, and then incubated on ice for 10 min. The suspension was drawn into an automatic pipette and squirted (diluted) into 2.5 ml of 50 mM potassium phosphate, pH 7.5, 1 mM dithiothreitol, 20 mM lactose which was at room temperature and blended gently on a Vortex mixer. Lactose counterflow in the reconstituted proteoliposomes was measured by filtering 0.5 ml of proteoliposomes on a 25 mm (0.22 p~; Millipore) GSTF filter at 25 inches of mercury vacuum suction, using a 9-mm internal diameter chimney. The use of a small diameter chimney to direct the proteoliposomes and ['4C]lactose onto a spot in the center of the fiter eliminates background radioactivity due to seepage of ['4C]lactose under the larger diameter chimney during the wash. As soon as the proteoliposomes were filtered, the vacuum pump was turned off and the vacuum was allowed to dissipate. The filter valve was closed and 200 p1 of 50 mM potassium phosphate, pH 7.5, containing 0.2 pCi of ['4C]lactose (the final lactose concentration was 17 p~) was placed on top of the filter (in the 9-mm chimney). It is important to use a fiter device which does not draw the lactose solution through the filter due to residual vacuum. After a 2-min incubation, the pump was turned on and the lactose solution was filtered. The chimney was immediately removed, replaced by a standard (16-mm internal diameter) chimney, and the filter was washed with 5 ml of ice-cold 50 mM potassium phosphate, pH 7.5, and counted as described (9). A background value, determined by assaying proteoliposomes in the presence of 10 mM thiodigalactosidase, was routinely subtracted. The rapid filter technique allows the entire transport assay to be carried out on the filter.
In order to determine the time course of counterflow or membrane potential-driven transport (Fig. 3), reconstituted proteoliposomes were prepared, centrifuged, and assayed essentially as previously described (9) except as noted in the figure legends.
SDS-Polyacrylamide Gel Electrophoresis-The buffer system of Laemmli (15) was used to run 12% polyacrylamide (10 cm) gels with 5% polyacrylamide (1 cm) stacking gels. Protein samples were mixed with an equal volume of 2-times concentrated sample buffer which contained 6% SDS, and 200 mM dithiothreitol instead of 2-mercaptoethanol. Solubilization was carried out at room temperature for 45 min. Samples (50 p l ) were stacked at 60 V and then electrophoresed at 120 V (constant voltage). Gels were silver stained by the method of Oakley et al. (16).
Protein Determination-Protein was assayed by the method of Schaffner and Weissmann (17). The final concentration of SDS in the assay was increased to 1%, and protein precipitated with trichloroacetic acid was fdtered on a 47-mm diameter, (0.45 p~; Amicon) microporous filter. For the determination of protein in proteoliposomes, a background WBS routinely subtracted, which was determined by processing liposomes that were "reconstituted" without added protein.  Determination of protein in E . coli membrane preparations by the Lowry procedure (18) routinely yielded values 1.5-fold higher than those obtained by the procedure of Schaffner and Weissmann. A similar overestimation has been observed in another system (19). The accuracy of the Schaffner and Weissmann procedure was confumed by the amino acid analysis of the purified lactose carrier (data not shown).  (Table I).  (Table I) '.

Purification
Recent work has demonstrated that the lactose carrier can be solubilized in a functional form with octylglucoside and reconstituted into liposomes (9). When' the urea/cholatetreated membrane was extracted with octylglucoside in the presence of washed E. coli lipid according to this procedure,

60-80%3 of the I3H]NPG label was solubilized. Since this procedure resulted in the solubilization of only 15% of the protein, this step yielded a 4-fold enrichment of the [3H]NPG
label, resulting in a 12-fold purification relative to the original membrane ( Table I).
The octylglucoside extract was loaded onto a DEAE-Sepharose column which was then developed with column buffer as described under "Experimental Procedures." The effluent was assayed for protein, C3H]NPG, and counterflow activity (Fig. 1). Although a significant amount of counterflow activity and 60% of the ['HINPG-labeled material was eluted, most of the protein (74%) was adsorbed by the column. As shown, the elution profile contained two protein peaks. The first, which coincides with the void volume of the column, contained approximately 20% of the protein and 20% of the E3H]NPG Octylglucoside extraction of rightside-out or French press vesicles which had not been treated with urea or cholate typically resulted in the solubilization of 95% of the C3H]NPG-labeled protein. We do not know why the urea/cholate-treatment leads to a slightly lower efflciency of octylglucoside solubilization. There is no indication that treatment with urea and cholate results in substantial loss of lactose transport activity following solubilization and reconstitution.  applied to the column, but very little counterflow activity. The second peak was slightly retarded by the column and contained the remainder of the protein eluted. Furthermore, this peak typically contained 50% of the ['HINPG label and at least 50% of the total counterflow activity applied to the column, as determined by the rapid filter assay (data not ~hown).~ The specific activity for counterflow (2-min uptake in the rapid filter assay) of proteoliposomes reconstituted with ' The determination of a single time point by the rapid filter assay can only yield a semiquantitative estimation of the total counterflow activity, as the time course and maximum accumulation for counterflow is affected by the concentration of carrier used in the reconstitution (22).

Amino acid composition of the purified lactose carrier
Protein (1.8 pg) was hydrolyzed in an evacuated sealed tube with 5.7 N HCI for 24 and 72 h at 110 "C. The hydrolysates were analyzed on a microbore amino acid analyzer using a fluorescamine detection system (27). The sensitivity of detection was 20 pmol. Cysteine was determined as described (28). Values given were corrected for degradation during hydrolysis. pooled material from the second peak (fractions 6, 7, and 8) was 405 nmol of lactose/mg of protein. This value was 2.8 times higher than the specific activity of proteoliposomes reconstituted with the octylglucose extract (144 nmol of lactose/mg of protein). Although this is only a rough estimation of specific activity, the difference corresponds well with the purification factor reported for the DEAE column step ( Table  I) and indicates that the carrier did not lose activity during DEAE-Sepharose chromatography. There was excellent correspondence between protein concentration, ["HINPG concentration, and counterflow activity in the second peak.
As described in Table I, the fractions of the second peak containing the highest counterflow activity (fractions 6,7, and 8 in Fig. 1) contained 14% of the ["HINPG label and 0.4% of the protein originally present in the membranes. This represents a 35-fold purification of the ["HINPG label relative to the starting material. Photolabeling studies with ['HINPG have indicated that 3% of the protein in E. coli T206 membranes is lactose carrier (data not shown). A 35-fold enrichment of the ["HINPG label suggests that a high degree of purification was achieved. This was confirmed by SDS-polyacrylamide gel electrophoresis of the purified material. Fig. 2 shows the results from SDS-polyacrylamide gel electrophoresis of urea/cholate-extracted membranes, the octylglucoside extract, and pooled DEAE fractions. The pooled DEAE fractions yielded a single broad band when stained with a highly sensitive silver procedure. The purified protein had an apparent M, = 33,000, which is in close agreement with published values for the molecular weight of the carrier as determined by SDS-polyacrylamide gel electrophoresis (5,6). When membranes were prepared from cells which had not been induced with isopropyl-thiogalactoside, the band corresponding to the purified carrier was only a minor constituent of the octylglucoside extract of urea/cholate-treated membranes (data not shown). This indicates that the purified protein is induced by isopropyl-thiogalactoside, a property expected of the product of the lacy gene in the recombinant plasmid (12). Finally, the Counterflow and membrane potential-driven lactose transport in proteoliposomes reconstituted with purified lactose carrier. A, counterflow; 100 pl of a DEAE-Sepharose fraction containing purified carrier (1.5 pg of protein) was mixed with 230 p1 of 50 mM potassium phosphate, pH 7.5, 20 mM lactose, 1.25% octylglucoside, 1 mM dithiothreitol. The solution was mixed with 100 pl of bath-sonicated liposomes and 8.3 p1 of 15% octylglucoside and reconstituted for counterflow as described (9). The proteoliposomes were resuspended with 100 p1 of 50 m M potassium phosphate, pH 7.5, 20 m M lactose, 1 mM dithiothreitol and assayed for counterflow as described (9), except that 9 pl of proteoliposomes were diluted into 0.9 ml of 50 m M potassium phosphate, pH 7.5 (no MgSO,), containing 1.8 pCi of [14C]lactose. The final lactose concentration was 0.23 mM. In addition to subtracting a blank obtained by filtering 0.1 ml of the ['4C]lactose suspension without proteoliposomes, a second blank (to correct for the presence of [3H]NPG-labeled carrier in the proteoliposomes), obtained by fdtering proteoliposomes without [14C]lactose, was subtracted from all points. 0, proteoliposomes were assayed as described above. 0, proteoliposomes were assayed in the presence of 10 m M thiodigalactoside. B, membrane potential-driven transport; 100 pl of a DEAE-Sepharose fraction containing purified carrier (1.0 pg of protein) was mixed with 230 p1 of 50 m M potassium phosphate, pH 7.5, 1.25% octylglucoside, 1 m M dithiothreitol. The solution was then mixed with 100 p1 of bath-sonicated liposomes and 8.3 pl of 15% octylglucoside and reconstituted for membrane potential-driven transport as described (9). The proteoliposomes were resuspended with 50 p1 of 50 mM potassium phosphate, pH 7 . 5 , l mM dithiothreitol and assayed for transport by diluting 6 pl of proteoliposomes into 0.6 ml of 50 m M sodium phosphate, pH 7.5, 10 p~ valinomycin, 0.23 m M lactose, containing 0.6 pCi of [14C]lactose. Aliquots (0.1 rnl) were taken at various times and filtered and counted as described (9). Blank values were subtracted as described under A. 0, proteoliposomes were assayed as described above. 0, proteoliposomes were diluted into 50 RIM sodium phosphate, pH 7.5, containing 20 p~ carbonyl cyanide-m-chlorophenyl hydrazone. The same result was obtained when proteoliposomes were diluted into 50 m M potassium phosphate instead of sodium phosphate.
amino acid composition of the purified protein (Table 11) closely matches the composition predicted from the DNA sequence of the lac y gene (7). This result indicates that the functional lactose carrier has a molecular weight similar to the value predicted (46,504) from the DNA sequence (7). It is not known why the carrier yields a lower molecular weight in SDS-polyacrylamide gels, although the highly hydrophobic amino acid composition of the carrier (7) suggests that this phenomenon may be due to unusually high binding of SDS. Furthermore, it is not known why the carrier migrates as a broad band.
Villarejo and Ping (23, 24) have reported that the lactose carrier can be resolved into two distinct bands on SDS-polyacrylamide gels when solubilized at 100 "C in the presence of high concentrations of dithiothreitol. These bands have apparent M, = 30,000 and 15,000 and it was proposed that both polypeptides are products of the lac y gene. When purified lactose carrier was prepared for electrophoresis according to this procedure, a single band (MI = 33,000) was obtained after electrophoresis (data not shown). The absence of the M, = 15,000 species in our preparation indicates that a polypeptide of this molecular weight is not essential for lactose transport activity.

A Single Polypeptide Is Required for Lactose Counterflow
and Energy-coupled Transport-When the purified carrier protein was reconstituted into liposomes as previously described (g), the resultant proteoliposomes exhibited counterflow activity, and the activity was completely inhibited by thiodigalactoside (Fig. 3A). Quantitation of the protein and [3H]NPG label associated with the reconstituted proteoliposomes indicated that 90-100% of the purified carrier (which was incubated with liposomes in the reconstitution procedure) was recovered in the proteoliposome fraction after centrifugation.
As shown in Fig. 3B, liposomes reconstituted with purified carrier also exhibited lactose accumulation when a membrane potential (interior negative) was generated with a potassium diffusion gradient (Kfin -+ K' ,,t) in the presence of valinomycin (9). Lactose accumulation was not observed when potassium-loaded proteoliposomes were diluted into medium containing the proton ionophore carbonyl cyanide-m-chlorophenyl hydrazone or medium containing potassium phosphate instead of sodium phosphate.
Additionally, liposomes reconstituted with purified carrier exhibited lactose efflux-induced dansylgalactoside fluorescence and lactose-dependent alkalinization of the medium (data not shown).

DISCUSSION
These experiments describe a relatively simple procedure for the purification of the lactose carrier protein in functional form. Moreover, the results demonstrate that only one polypeptide species, the product of the lacy gene, is necessary for lactose counterflow and energy-coupled lactose transport. In light of this finding, it is difficult to explain how certain mutations mapping outside the lactose operon (8) result in the pleiotropic loss of energy-coupled accumulation of solutes by several transport systems, including the lactose carrier, despite the ability of these cells to maintain a transmembrane electrochemical proton gradient. One possible explanation is that the transport systems in these mutants have been altered by a mutation in some component of a common regulatory system.
This and previous work (20,21,25) have demonstrated that several E. coli membrane proteins are not irreversibly denatured by treatment of the membrane with urea or cholate, although significant amounts of protein are solubilized. Furthermore, the lactose and melibiose (26) transport systems have been solubilized with octylglucoside and reconstituted. The results suggest that these procedures may have general applicability to other bacterial transport systems.