INTRODUCTION

Periplasmic permeases are complex transport systems composed of a soluble substrate-binding protein, the receptor, and a membrane-bound complex containing four subunits (2, 3). Transport through these systems is energized by ATP hydrolysis mediated by one of the membrane-bound components that is evolutionarily conserved in a large superfamily of transporters, the traffic ATPases (4) or ABC transporters (5). This superfamily includes several eukaryotic proteins, such as the multidrug resistance protein (P-glycoprotein or MDR), the cystic fibrosis transmembrane conductance regulator (CFTR), and the STE6 gene product from yeast (6, 7).

The histidine permease of Salmonella typhimurium and Escherichia coli comprises the receptor, the periplasmic histidine-binding protein HisJ (8), and a membrane-bound complex (HisQ/M/P)¹ containing two hydrophobic subunits, HisQ and HisM, and two identical copies of the ATP-binding (conserved) component, HisP (9). A model for transport through these permeases was proposed in which liganded HisJ interacts with HisQ/M/P, thus initiating ATP hydrolysis and the consequent substrate translocation (10). Several in vitro reconstituted assay systems have been developed for the biochemical analysis and purification of the histidine permease (1, 11, 12). Of these, the most responsive to in vitro analysis is reconstitution into PLS. A detergent dilution procedure has been shown to be effective for several systems (1, 13, 14, 15). However, this procedure is cumbersome, tends to give unreproducible PLS preparations, and, because it can only produce small batches of PLS that are unstable, it has not been useful for extended biochemical analyses.

Here we extensively characterize the histidine permease using a simple and reliable dialysis method for reconstitution into PLS. Various parameters have been examined including the affinity of HisQ/M/P for HisJ and ATP; the effect of ADP, internal substrate, pH, salt, and temperature on transport; the nature of the plateau; and the reversibility of the system. Evidence is provided for the existence of a substrate-binding site in HisQ/M/P.

EXPERIMENTAL PROCEDURES

The following E. coli K12 strains were used: GA298, which carries a deletion (unc702) that eliminates the F₀F₁ ATPase and the  repressor cI857 temperature-sensitive mutation; TA1889, which is isogenic with GA298, harbors plasmid pFA17 containing the S. typhimurium hisQ, hisM, and hisP genes under the temperature-sensitive control of the phage  P_L promoter, and the -lactamase gene (16); GA300, which is isogenic with GA298, but instead of pFA17 harbors plasmid pFA108 that carries the wild type S. typhimurium hisQ and hisM genes, and the hisP9085 gene (which codes for a mutant protein that has binding protein-independent ATPase activity) (17).

1-100 liters of bacterial culture were grown aerobically in fresh LB medium in the presence of ampicillin (50 µg/ml) at 30 °C until it reached an A_650 nm of about 0.5. The temperature was then raised to 42 °C, and the growth continued for 1 h. The heat-induced cells were harvested by centrifugation, washed with buffer A (50 mM MOPS/K⁺, 5 mM MgSO₄, 1 mM dithiothreitol, 0.1 mM phenylmethanesulfonyl fluoride, pH 7.0), and resuspended in the same buffer at a cell density 100-fold higher than the original culture. The cells were disrupted by three passes through a French Press cell at 10,000 psi (68.9 MPa). The unbroken cells were removed by two successive centrifugations at 1400 × g for 5 min, and the membranes were harvested by a discontinuous sucrose gradient (1 ml of 2 M sucrose at the bottom and 2 ml of 0.5 M sucrose on top, both prepared in 50 mM MOPS/K⁺, pH 7.0). Ten ml of cell extract were applied on top of each sucrose gradient and centrifuged in a SW41 rotor (Beckman) at 40,000 rpm for 1 h. The membrane fraction migrated to form a thin layer between 2 M and 0.5 M sucrose. The portion above the membrane layer (including the 0.5 M sucrose) was removed using a Pasteur pipette without disturbing the membrane layer. A Pasteur pipette was carefully inserted through the center of the membrane layer to reach the bottom of the tube, and the 2 M sucrose solution was thoroughly removed. The membrane layer was resuspended in buffer A at a protein concentration of 20-30 mg/ml and stored in liquid nitrogen in 1 ml aliquots. The typical yield was 50 mg of membrane protein/liter of culture.

Aliquots (1 ml) of membranes vesicles were diluted into solubilization buffer (buffer A containing 1.2% octylglucoside (Calbiochem), 20% (v/v) glycerol, 3.7% E. coli phospholipids (Avanti Polar Lipids Inc, Birmingham, AL), 2 mM EGTA, and 15 mM ATP) to a final membrane protein concentration of 1 mg/ml. The mixture was incubated on ice for 30 min, with occasional shaking by hand, and then centrifuged at 150,000 × g for one h. To the supernatant fraction containing the solubilized membrane protein (50% of the total membrane protein, at a concentration of 0.4-0.5 mg/ml), additional E. coli phospholipids were added to a final concentration of 10 mg/ml. The mixture was then dialyzed at 4 °C for about 24 h against a 100-fold volume of dialysis buffer (50 mM MOPS/K⁺, pH 7.0, 1 mM dithiothreitol) with one change of the dialysis buffer after the first 4 h. The resulting PLS were placed into cryotubes (1 ml aliquots) and stored in liquid nitrogen. The ATPase activity of the PLS was routinely assayed immediately after reconstitution.

ATP trapping was as described in the legend to Fig. 2. Non-encapsulated ATP was routinely removed by passing PLS (2 ml) through a pre-packed disposable PD-10 desalting gel filtration column (Sephadex G-25 M; Pharmacia Biotech Inc.) equilibrated in buffer B (50 mM MOPS/K⁺, 100 mM NaCl, pH 7.0). Other compounds (HisJ, ADP, histidine, and other amino acids) were similarly trapped.

ATP-containing PLS (400 µl) resuspended in buffer B were mixed with purified unliganded HisJ (20 µM final concentration) in a final volume of 0.5 ml and kept on ice. Transport assays were carried out at 37 °C. After a 1-min incubation at 37 °C, transport was initiated by the addition of L-[³H]histidine (20 µM final concentration) (Amersham Corp.). Aliquots (50 µl) were taken at the indicated times, diluted into 0.9 ml of ice-cold buffer B, filtered immediately through 0.22 µm GSTF Millipore filters, and washed with 1 ml of ice-cold buffer B. Filters were dried, placed in Scint A scintillation fluid (Packard), and counted using a Searle Delta 300 scintillation counter. For a time zero value, the complete reaction mixture was kept on ice, and three 50 µl-aliquots were taken and the radioactivity values averaged. Filtration of HisJ liganded with L-[³H]histidine through a GSTF filter under the transport assay conditions showed that about 10% of the liganded HisJ was retained by the filter and that the zero time value was mostly due to liganded HisJ being retained by the filter; this value was subtracted from all subsequent values. Liposomes or PLS containing HisQ/M/P (or not) had no effect on the retention of HisJ by the filter. The retention efficiency of PLS by the 0.22 µM GSTF filter had been determined to be about 80% by measuring the radioactivity retained upon filtering PLS loaded with ³⁵S-adenosine 5-O-(3-thio)triphosphate (ATP--³⁵S) under the transport assay conditions.

5(6)-Carboxyfluorescein (CF) (40 mM; Molecular Probes) was encapsulated into PLS by freeze-thawing as described for ATP. When the concentration of CF reached 40 mM, its fluorescence was almost completely self-quenched; the addition of Triton X-100 (0.2% v/v) to the PLS yielded 100% fluorescence because release of the CF into the medium resulted in its dilution and, thus, in the relief of self-quenching. The internal volume of PLS was calculated by the level of released fluorescence using a standard curve of CF fluorescence (18). Fluorescence measurements were performed with a Perkin-Elmer spectrofluorimeter (LS50) equipped with a thermostat and a continuous stirrer using excitation and emission wavelengths of 490 and 515 nm, respectively. Excitation and emission slit widths were 10 and 5.0 nm, respectively. Experiments were performed at a constant temperature of 25 °C.

PLS were diluted and allowed to settle on a Formvar-coated 200-mesh copper grid (made with a 0.5% Formvar solution) for 2 min. Excess solution was removed using filter paper. After negative staining with uranyl acetate, the grid was examined immediately at 80 kV using a transmission electron microscope (JEOL100CX). The level of magnification was calibrated with Pelco grating replicas (Ted Pelco, Inc.).

The ATPase reaction was initiated by adding ATP together with MgSO₄ (2 and 10 mM final concentration, respectively) to a suspension of PLS in 50 mM MOPS/K⁺, pH 7.5, containing 10 µM HisJ and 100 µM histidine. The final concentration of PLS in the assay was 3 mg of phospholipid/ml. Unless specified differently, the assay was carried out at 37 °C. Aliquots (100 µl) were taken at various times, mixed with an equal volume of 12% SDS, and the released inorganic phosphate was determined (19). The ATPase activity of membrane vesicles was assayed in the same way, except that PLS in the reaction mixture were replaced by about 1 mg of membrane protein/ml, and 1 mg/ml E. coli phospholipids was added. Alternatively, ATP levels in PLS were determined by the luciferin/luciferase assay (20).

Proteins were analyzed by SDS-12.5% PAGE (21) with the pH of the resolving gel adjusted to 8.65 (16). Measurement of protein concentration (22), the purification of HisJ (12) and the separation of the unliganded and liganded forms (23), and immunoprecipitation (9) have been described. ATP--³⁵S (DuPont NEN) was dissolved in water and stored at 20 °C. Radioactively labeled octylglucoside (American Radiolabeled Chemicals Inc.) was dissolved in ethanol and stored at 4 °C. CF (Molecular Probes) was dissolved in water and stored as a 0.3 M stock solution at 20 °C. Ether/acetone-precipitated E. coli total phospholipids were suspended at a concentration of 50 mg/ml in argon-saturated 2 mM 2-mercaptoethanol, briefly sonicated with a tip sonicator to obtain a viscous homogeneous suspension, and then stored in liquid nitrogen in aliquots.

RESULTS

Solubilization Conditions

Membrane vesicles from TA1889 containing HisQ/M/P were solubilized as described under "Experimental Procedures," using octylglucoside concentrations ranging from 0.3 to 1.5%. The effect of ATP, Mg²⁺ ions, glycerol, and phospholipids concentration on solubilization was determined. The extent of protein solubilization of HisQ/M/P was examined by SDS-PAGE, immunoblots using anti-HisP and -HisQ antibodies (9), and ATPase activity (after reconstitution, for detergent-treated samples). Solubilization of HisQ/M/P needs octylglucoside concentrations higher than 0.9%. ATP increases the solubilization of all membrane proteins, including porins, that appear to interfere with histidine uptake in PLS. The presence of ATP also results in partial degradation of HisP; however, phospholipids protect HisP from degradation and suppress the solubilization of porins. Glycerol, which is known to stabilize membrane proteins (14, 24), also increases the level of solubilization of HisQ/M/P. The conditions adopted for optimal solubilization of HisQ/M/P (over 90%) are: 1.2% octylglucoside (40 mM), 15 mM ATP, 20% glycerol, 5 mM Mg²⁺, 3 mg/ml of E. coli phospholipids, 1 mg/ml of protein, and incubation for 30 min on ice.

Although the HisJ-stimulated ATPase activity is freely assayable in membrane preparations, upon addition of octylglucoside, no activity can be detected. 99% of the activity is recovered if the reconstitution is performed immediately after solubilization. However, prolonged exposure of HisQ/M/P to octylglucoside irreversibly damages the ATPase activity, as shown by the loss of about 70 and 30% of activity in solubilized samples stored for 24 h at 4 and at 20 °C, respectively, in the presence of 1.2% octylglucoside before reconstitution. Therefore, all samples were reconstituted within 30 min of solubilization.

Characterization of PLS

To reconstitute HisQ/M/P into PLS, detergent was removed by dialysis, which was monitored using ¹⁴[C]-labeled octylglucoside. Fig. 1 shows that within 4 h, starting with an initial concentration of octylglucoside of 40.5 mM (1.2%), 90% is lost and 30% of the HisJ-stimulated ATPase activity is recovered. The final concentration of octylglucoside in PLS after 24 h of dialysis with one change of buffer is 0.16 mM (0.0046%). The final molar ratio of phospholipids to detergent is estimated to be about 70:1, using a phospholipid molecular weight of 700 Da. Dialysis for longer than 24 h neither removes more detergent nor increases the ATPase activity of PLS. Including hydrophobic beads (Bio-BEADS SM-2, Bio-Rad) in the dialysis buffer speeds up the dialysis process, and the final concentration of octylglucoside can be lowered to 0.027 mM (0.0008%) (Fig. 1, inset), even though it is not necessary to reach this lower level of detergent for transport assays. The PLS maintain 100% activity after one month of storage either in liquid nitrogen or at 20 °C, as followed by both ATPase and transport activities.

ATP is introduced into PLS subsequently to their formation using a freeze-thawing procedure. The presence of ATP in PLS was confirmed by separating PLS from free ATP with a gel filtration column (Fig. 2A, solid circles). Two peaks of ATP are detected, with the earlier one corresponding to the void volume and the later one corresponding to free ATP. PLS are eluted in the void volume, as identified by optical density (liposomes with 1 mg/ml E. coli phospholipids have an A_530 nm of 0.671 due to light scattering), translucent appearance, and SDS-PAGE (indicating the presence of HisP) (Fig. 2, B and C). That the ATP associated with PLS is not unspecifically bound was shown by several experiments. 80% of the ATP co-eluted with PLS, which corresponds to the PLS retention efficiency, is retained by filtration (0.22 µM filters). If PLS are exposed to ATP after freeze-thawing, the earlier peak (Fig. 2A, open circles) contains only 10% of the ATP in PLS exposed to ATP during freeze-thawing (solid circles), indicating that in the latter case, the majority of the ATP is trapped inside. Trapped ATP is readily lost if PLS are permeabilized.² After the removal of external ATP by gel filtration chromatography, histidine uptake proceeds for 20 min, indicating that the ATP associated with PLS is freely available internally.

The trapping efficiency depends on the number of freeze-thawing cycles; one and two cycles allow trapping of 50 and 85%, respectively, of the amount trapped with three cycles. Five cycles of freeze-thawing was routinely performed. Repeated freeze-thawing has no damaging effect on the ATPase activity of HisQ/M/P. Mg²⁺ is needed for both transport and ATPase activity (data not shown); therefore, ATP trapping is performed in the presence of equimolar MgSO₄. Varying concentrations of ATP, up to 30 mM, were trapped and shown to be effective for transport; higher concentrations were not used because of the correspondingly high osmolarity. The concentration finally chosen, 15 mM, is saturating for the initial rate of transport (see below). After freeze-thawing, PLS are passed several times through filters with defined pore sizes, using a commercially available extrusion apparatus; this procedure has been shown to produce unilamellar and homogeneous PLS (25). PLS with encapsulated ATP could be stored at 4 °C for at least 4 days without any loss of transport activity. Longer storage at 4 °C was avoided.

The internal volume of PLS was determined using CF (18) to be 0.75 µl/mg of phospholipids (see "Experimental Procedures"), which is similar to the volume estimated for PLS reconstituted by a dilution method (26). Alternatively, the volume can be estimated from the amount of ATP trapped, as measured both using (ATP--³⁵S) as a tracer and by the luciferin/luciferase assay, after passage of PLS through a gel filtration column. The internal volume was calculated to be 0.65 µl/mg of phospholipids, which is in good agreement with the former measurement.

The size was estimated by electron microscopy. The PLS appeared as small, mostly round vesicles, with diameters distributed according to a Gaussian curve with a mean of 48 nm (as determined from 194 vesicles). This value is consistent with other measurements on PLS prepared by dialysis (27). Assuming that PLS are perfectly spherical, the surface area and volume of a single reconstituted proteoliposome vesicle are 7.2 × 10⁵ Ų and 5.8 × 10⁷ ų, respectively. The minimal packing requirement for the phospholipids headgroup is about 50 Å² (28, 29), therefore each PLS contains 1.44 × 10⁴ phospholipids molecules. Since the phospholipid concentration is 10 mg/ml in PLS and the average molecular weight of phospholipids is 700, each ml of PLS preparation contains 6.15 × 10¹⁴ vesicles, assuming that all PLS are monolamellar (25). The total membrane protein concentration in PLS is about 0.5 mg/ml and HisQ/M/P comprises 20% of the total membrane protein. Therefore, there is an average of 1 HisQ/M/P complex/PLS vesicle.

It is important to determine whether the HisQ/M/P complexes maintains the same orientation in PLS as in vivo. The orientation in intact cells (native orientation) is such that the ATP-binding domain resides on the inner face of the membrane. HisQ/M/P embedded in PLS in the native orientation would hydrolyze ATP only when HisJ is outside and ATP is inside, while HisQ/M/P in the reverse orientation would hydrolyze ATP only with HisJ inside and ATP outside. Therefore, a comparison of ATPase activities in these respective PLS preparations would allow an estimation of the ratio between the two orientations. PLS containing HisJ inside and ATP outside (Fig. 3, left) hydrolyze 110 nmol of ATP/min/mg of protein (gray column), which is one-third of the activity with HisJ and ATP on both sides (350 nmol of ATP/min/mg of protein; dotted column). The activity due to the presence of HisJ outside and ATP inside is calculated to be 240 nmol/min/mg of protein (hatched column). Therefore, two-thirds of HisQ/M/P appears to be in the native orientation. To exclude the possibility that inefficient trapping of HisJ is responsible for the low activity of PLS with HisJ inside (which would result in poor stimulation of ATPase activity), PLS were prepared from mutant strain GA300, which has a HisJ-independent, i.e. constitutive, ATPase activity (17) (Fig. 3, right). These PLS had the same ratio of orientations as TA1889.

The orientation of HisQ/M/P was also analyzed by immunoprecipitation experiments (Fig. 3, inset). Since on average there is one HisQ/M/P/PLS vesicle, there are two populations of PLS based on the orientation of the single HisQ/M/P complex. Only PLS with HisQ/M/P embedded in the reverse orientation can be immunoprecipitated by anti-HisP antibody, which recognizes the cytoplasmic portion of HisP (30). The HisP content of PLS precipitated with anti-HisP antibody in the absence of Triton X-100 is 45% of the amount precipitated in the presence of Triton X-100, indicating that 55% of PLS have HisQ/M/P embedded in the native orientation, which is in good agreement with the value obtained from the ATPase assay above. As a negative control, it was shown that HisQ/M/P in right-side out membrane vesicles cannot be precipitated in the absence of Triton X-100 (30).

Characterization of Histidine Transport in PLS

Various parameters of transport have been analyzed in detail. Fig. 4A shows that L-histidine transport is linear for about 1 min and then gradually slows down over the next 5-10 min, reaching a plateau that is maintained for at least 20 min. The initial rate of transport, as calculated from data taken within the first minute, is 7.5 ± 0.8 nmol/min/mg of protein (an average of five experiments). Transport is completely dependent on the presence of HisJ, ATP, and HisQ/M/P (as expected) (1, 11, 12). PLS prepared from strain GA298, which does not contain any HisQ/M/P, fails to transport. The apparent K_m of HisJ for HisQ/M/P is 8 µM (Fig. 4B), which is similar to the value obtained from other studies (10).

External ATP can be removed without affecting transport.³ Both the initial rate of transport and the plateau level are dependent on the internal ATP concentration (Fig. 4C), with an apparent K_m for ATP, as determined from the initial rates, of 8 mM (Fig. 4D). External free histidine (up to 2 mM) does not affect either the rate of transport or the plateau level (in the presence of either 20 µM or 2 mM L-histidine, 2.7 nmol of histidine/min/mg of protein were transported, and the plateau was 20 nmol of histidine/mg of protein), indicating that free histidine is not the immediate substrate for transport.

The effect of pH, temperature, and salt on transport was determined (Fig. 5). The optimum external pH is about 7.5, with the internal pH kept at 7.0. It should be noted that pH values lower than 5 result in aggregation of PLS, as indicated by the appearance of turbidity. The optimum temperature is 46 °C, at which the rate is 60% higher than at 37 °C. No transport was observed below 15 °C. The optimum external salt concentration was determined using NaCl and found to be 50 mM, which has the same ionic strength as that of the internal ATP/Mg²⁺.

It has been reported previously that histidine transport in whole cells and in membrane vesicles reaches a plateau after a prolonged period of uptake (12, 31). Using membrane vesicles, it was concluded that the plateau is the result of an equilibrium between entry and exit (12). Because of the complex nature of these in vitro systems, we examined again whether the plateau is due either to an equilibrium between the efflux and uptake or to a decrease in the uptake rate, using the PLS system. The possibility that the plateau is caused by efflux countering uptake was investigated by adding a large excess of unlabeled histidine to PLS, which had transported labeled histidine and had almost reached the plateau (Fig. 6). Accumulation of labeled histidine stopped immediately, and there was no leakage of histidine over the next 10 min, indicating that there is no efflux. Therefore, the plateau must result from a decrease in the uptake rate. To confirm this conclusion, the uptake rates at different times were measured by adding labeled histidine to a sample that already contained (and transported) unlabeled histidine. Fig. 6, inset, shows that the rates of uptake decrease over time, and after 2 min the rate of uptake has become one-half the initial rate.

There are several possibilities for the decrease in uptake rate: i) inhibition by the accumulated histidine, ii) inhibition by a positive membrane potential built up by the accumulated histidine, iii) exhaustion of ATP, and iv) inhibition by ADP generated during ATP hydrolysis (17). The change in pH caused by ATP hydrolysis during transport was negligible under the assay conditions, which include 50 mM MOPS/K⁺ buffer.

The effect of increasing the internal histidine concentrations is shown in Fig. 7. Various amounts of histidine (with labeled histidine added as a tracer) were trapped into PLS by freeze-thawing, and the excess external histidine was removed by gel filtration, as described for the trapping of ATP. The concentration of the final external unlabeled histidine was estimated to be less than 1 µM, which would not affect the specific radioactivity in the transport assay (which uses 20 µM L-[³H]histidine). Both the initial rate and the plateau level are decreased when millimolar levels of histidine are present inside PLS; 15 mM histidine gives 85% inhibition of the initial rate. In order to exclude the possibility that the effect of histidine is nonspecific, the effect of two other substrates of the histidine permease, L-arginine and L-lysine, and of a non-substrate, L-glutamate (each at 15 mM), was also analyzed under the same conditions. L-arginine and L-lysine inhibit the initial rate of histidine uptake by 65% and 30%, respectively, while glutamate gives a 15% inhibition⁴ (data not shown). The K_i values for L-histidine, L-arginine, and L-lysine are estimated to be 3.2, 7.5, and 31.2 mM, respectively.⁵ These results indicate that there is a specific substrate-binding site on HisQ/M/P. Because the histidine concentration at the plateau is 0.23 mM (as calculated from the known internal volume of PLS), it cannot account for most of the plateau effect as obtained during a standard assay.

Since histidine is partially charged under the experimental conditions, the accumulation of histidine may create an electrochemical gradient across the membrane bilayer, which may be inhibitory. However, upon addition of valinomycin to eliminate any possible membrane potential, both the initial transport rate and the plateau level were unchanged.

At the plateau, about one-third of the total trapped ATP is hydrolyzed (5.4 nmol of ATP out of a total of 14.9 nmol in 100 µl of PLS, as measured by the luciferin-luciferase ATP assay). Because, as shown in Fig. 4C, both the initial rate of transport and the plateau levels are dependent on the ATP concentration, complete exhaustion of ATP would be a sufficient explanation for the plateau. Starting with an initial ATP concentration of 15 mM and assuming that all of the trapped ATP is accessible, it can be calculated that ATP is about 9 mM at the plateau. This lowered ATP concentration accounts for a 40% drop in the transport rate. The inhibitory effect of ADP was tested by trapping increasing amounts of ADP with a constant amount of ATP in PLS. Fig. 8 shows that ADP inhibits transport with a K_i of 2.1 mM. The calculated concentration of internal ADP at the plateau generated during transport (about 6 mM) would result in 60% loss of transport rate. Thus, the combined effect of ATP exhaustion and inhibition by ADP generated during ATP hydrolysis is sufficient to explain the plateau. Consistent with this notion is the fact that trapping an ATP-regenerating system (creatine kinase with 20 mM creatine phosphate, Boehringer Mannheim) increases the initial rate by 4-fold and the plateau by 2-fold.

DISCUSSION

A previous examination of the periplasmic histidine permease reconstituted into PLS by a dilution method had provided some basic insights into its properties (1). Here we have been able to analyze this permease in much more extensive detail using a newly developed PLS reconstitution method. The reconstituted permease has an efficiency that is very similar to that of the permease as studied in vivo and much higher than that obtained previously⁶ (1). Its turnover rate is 10 mol/min/mol of HisQ/M/P, which is equivalent to about 40-70% of the V_max (15-27 mol/min/mol of HisQ/M/P) estimated for the corresponding intact cells (12). This value takes into account the following considerations. i) HisQ/M/P accounts for about 20% of PLS proteins, as estimated by SDS-PAGE and Coomassie Blue staining. ii) About half of HisQ/M/P is embedded in the native orientation. iii) PLS are mostly unilamellar (25). iv) The V_max is calculated to be 10.5 nmol/min/mg of protein (under the assay conditions of 20 µM liganded HisJ, using a K_m value for HisJ of 8 µM). This turnover rate is likely to be an underestimate because some of the HisQ/M/P complexes may be inactive and some of the PLS may be multilamellar. Thus, it appears that this in vitro transport system closely resembles the conditions in vivo.

The presence of a substrate-binding site on the membrane complex of periplasmic transport systems has been postulated in the past, but never proven. Preliminary evidence was provided by the isolation of membrane proteins mutants that transport in the absence of the binding protein (32, 33, 34, 35). In addition, mutations have been characterized that are located in the membrane-bound components of the histidine permease and alter its substrate specificity (36). Our evidence that internal histidine inhibits uptake (Fig. 7) provides strong support that a substrate-binding site indeed is present in these systems. It appears unlikely that its function is regulatory, in order to slow down histidine transport, since such high internal concentrations would be rarely achieved in vivo (31). We speculate that the site is located in the "translocation pore" in HisQ/M/P (37); high internal concentrations of histidine would block such a site, thus interfering with translocation. The site might be responsible also for the substrate selectivity that the membrane-bound complex must possess, as indicated by both the binding protein-independent mutants (17, 32) and the altered specificity mutants (36).

Transport is unidirectional (Fig. 6). The absence of ATP synthesis from ADP and P_i in PLS loaded with a high histidine concentration also supports this notion.⁷ At the plateau, the internal histidine concentration is about 230 µM, with a free histidine concentration in the assay medium of 0.12 µM (38).⁸ Therefore, under the conditions of this assay, a 2000-fold concentration gradient is established, which is one order of magnitude higher than reported previously for this and similar permeases (1, 15, 39, 40). This value is closer to the 1000-fold concentration gradient obtained in vivo (2). It should be noted that transport can occur against a much higher gradient, as shown in Fig. 7, where transport proceeds even when the internal histidine concentration is much higher than 230 µM.

We show that the plateau is due to the combined effect of ATP limitation and inhibition by the ADP produced rather than to histidine efflux. These data are in contrast with results obtained in right-side out membrane vesicles where the plateau level is due to an equilibrium between entry and exit of histidine (12). It is likely that other transport systems or contaminating porins operating in the relatively crude right-side out membrane vesicles system are responsible for the exit process. A considerable amount of ATP (60-70%) is still present at the plateau. Inhibition of ATP hydrolysis by ADP contributes to the incompleteness of ATP utilization. The possibility that some of the PLS are multilamellar and the absence of transport complex in some of the PLS may also contribute to this effect, but they are not major factors, as previously suggested for the branched-chain amino acid transport system of Pseudomonas aeruginosa (39).

Successful reconstitution by dialysis has been achieved in the past for many membrane processes (41). Reconstitution of PLS by this dialysis method, coupled with the freeze-thawing procedure and the passage through micropore filters provides an easy and reproducible system for studying transport in vitro. It offers several advantages over the detergent dilution method. i) Large amounts of PLS can be prepared and stored for long periods of time without loss of activity. ii) It provides better control of the nature and amounts of reagents encapsulated because encapsulation is a separate step from reconstitution and is performed in a small volume and, thus, minimal amounts of reagents are consumed. iii) PLS are of a homogeneous size and mostly unilamellar. iv) Transport assays are very reproducible. v) High efficiency of transport activity. This method should be directly applicable to other transport systems.