Lactose Transport System of Streptococcus thermophilus

The kinetic mechanism of the lactose transport system of Streptococcus thermophilus was studied in membrane vesicles fused with cytochrome c oxidase containing liposomes and in proteoliposomes in which cytochrome c oxidase was coreconstituted with the lactose transport protein. Selective manipulation of the components of the proton (and sodium) motive force indicated that both a membrane potential and a pH gradient could drive transport. The galactoside/proton stoichiometry was close to unity. Experiments which discriminate between the effects of internal pH and delta pH as driving force on galactoside/proton symport showed that the carrier is highly activated at alkaline internal pH values, which biases the transport system kinetically toward the pH component of the proton motive force. Galactoside efflux increased with increasing pH with a pKa of about 8, whereas galactoside exchange (and counterflow) exhibited a pH optimum around 7 with pKa values of 6 and 8, respectively. Imposition of delta pH (interior alkaline) retarded the rate of efflux at any pH value tested, whereas the rate of exchange was stimulated by an imposed delta pH at pH 5.8, not affected at pH 7.0, and inhibited at pH 8.0 and 9.0. The results have been evaluated in terms of random and ordered association/dissociation of galactoside and proton on the inner surface of the membrane. Imposition of delta psi (interior negative) decreased the rate of efflux but had no effect on the rate of exchange, indicating that the unloaded transport protein carries a net negative charge and that during exchange and counterflow the carrier recycles in the protonated form.


Lactose Transport System of Streptococcus thermophilus
FUNCTIONAL RECONSTITUTION OF THE PROTEIN AND CHARACTERIZATION OF THE KINETIC MECHANISM OF TRANSPORT* (Received for publication, March 4, 1992) Catherine FoucaudS and Bert PoolmanQ From the Department of Microbiology, University of Groningen, Kerklaan 30, 9751 NN Haren, The Netherlands The kinetic mechanism of the lactose transport system of Streptococcus thermophilus was studied in membrane vesicles fused with cytochrome c oxidase containing liposomes and in proteoliposomes in which cytochrome c oxidase was coreconstituted with the lactose transport protein. Selective manipulation of the components of the proton (and sodium) motive force indicated that both a membrane potential and a pH gradient could drive transport. The galactosidelproton stoichiometry was close to unity. Experiments which discriminate between the effects of internal pH and ApH as driving force on galactosidelproton symport showed that the carrier is highly activated at alkaline internal pH values, which biases the transport system kinetically toward the pH component of the proton motive force. Galactoside efflux increased with increasing pH with a pK, of about 8 , whereas galactoside exchange (and counterflow) exhibited a pH optimum around 7 with pK. values of 6 and 8, respectively. Imposition of ApH (interior alkaline) retarded the rate of efflux at any pH value tested, whereas the rate of exchange was stimulated by an imposed ApH at pH 5.8, not affected at pH 7.0, and inhibited at pH 8.0 and 9.0. The results have been evaluated in terms of random and ordered associationldissociation of galactoside and proton on the inner surface of the membrane. Imposition of A* (interior negative) decreased the rate of efflux but had no effect on the rate of exchange, indicating that the unloaded transport protein carries a net negative charge and that during exchange and counterflow the carrier recycles in the protonated form.
The lactose transport protein (Lacs) of Streptococcus thermophilus is a polytopic membrane protein that traverses the cytoplasmic membrane most likely 12 times and contains a carboxyl-terminal hydrophilic extension of approximately 180 amino acids. The hydrophobic carrier domain of Lacs is homologous to the melibiose carrier protein (MelB) of Escherichia coli, but, with the exception of a region between putative a-helices X and XI, Lacs shares no similarity with the lactose transport protein (Lacy) of E. coli (Poolman et al., , 1992. The carboxyl terminus of Lacs is denoted * 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.
$ Permanent address: National Institute of Agronomic Research, Dairy Research Group, 78350 Jouy en Josas, France.
§ Recipient of a fellowship from the Royal Netherlands Academy of Arts and Sciences. To whom correspondence should be addressed. Tel.: 31-50-632170;Fax: 31-50-632154. IIA or enzyme I11 domain due to its similarity with IIA (enzyme 111) protein domains of various phosphonenolpyruvate:sugar phosphotransferase systems (Poolman et al., , 1992. The lactose transport system of S. thermophilus has been characterized at DNA level , and the role of the conserved (and other) histidine residues in the carrier and IIA domain has been assessed by biochemical characterization of site-directed mutants (Poolman et al., 1992). The latter studies indicate that at least part of the galactoside recognition site of Lacs, tentatively located between a-helices X and XI, may be similar to that of Lacy.
All these studies were performed with the cloned gene expressing Lacs in E. coli. The level of expression of lacs in E. coli from its own promoter, however, is low in comparison with the expression level in S. thermophilus. Therefore, in the present study the mechanism of transport was analyzed using membrane vesicles isolated from S. thermophilus as starting material. In view of the similarity of Lacs with MelB, which transports galactosides in symport with either protons, sodium, or lithium ions Leblanc et al., 1990), experiments were set up to determine the cation selectivity of Lacs and to compare the transport of a-galactoside (melibiose) and @-galactoside (TMG,' lactose). Furthermore, in a recent report, dealing with a study of Lacs in intact cells, it is concluded that Lacs acts as a lactose/ galactose antiporter (Hutkins and Ponne, 1991). The conclusion carries back to earlier observations that S. thermophilus only metabolizes the glucose moiety of lactose and that galactose is excreted into the medium stoichiometrically (Thomas and Crow, 1984). It has therefore been suggested that in vivo Lacs may facilitate lactose/galactose exchange rather than lactose/cation symport . In the present investigation, the different modes of facilitated diffusion mediated by the lactose transport protein of S. thermophilus have been analyzed in vitro. On basis of the effects of pH and membrane potential on the facilitated diffusion processes, a kinetic scheme of the translocation cycle of galactoside/proton symport and galactoside/galactoside exchange is proposed.
Isolation of Membrane Vesicles and Membrane Fusion-Membrane vesicles of S. thermophilus were prepared as described previously (Otto et al., 1982) with the following modifications: the culture was harvested during the exponential phase of growth (AGrn of 0.6-OB), the quantities of lysozyme, ribonuclease, and deoxyribonuclease were doubled, and the incubation steps were performed a t 37 "C. Cytochrome c oxidase containing liposomes (22 nmol of heme a/100 mg of phopholipid) were prepared as described  using L-a-phosphatidylethanolamine type IX obtained from E. coli (Sigma), which were further purified by aceton/ether washing (Kagawa et al., 1973), and L-a-phosphatidylcholine type XI-E from fresh egg yolk (Sigma) in a ratio of 3:l (weight/weight). Fusion between membranes vesicles and cytochrome c oxidase containing liposomes (10 mg of phospholipid/mg of vesicular protein) was performed by freeze/thaw sonication . To form unilamellar vesicles, the thawed suspension was sonicated a t 4 "C for 4 s with a macrotip at an output of 4 pm (peak to peak). By the same procedure, membrane vesicles were fused with liposomes devoid of cytochrome c oxidase. The fused membranes were collected by centrifugation (1 h a t 185,000 X g, 4 "C) and suspended to a final protein concentration of about 15 mg/ml in 50 mM potassium phosphate (KP,), pH 7.0, unless indicated otherwise.
Solubilization and Reconstitution of the Lactose Transport Protein-Membrane vesicles were extracted consecutively with 5 M urea and 6% (weight/volume) sodium-cholate in 50 mM KP,, pH 7.0 (Newman et al., 1981). Membrane proteins (1 mg of protein) were solubilized with 1.25% (weight/volume) n-octyl-P-D-ghcopyranoside (octyl glucoside) in the presence of 5 mg of a mixture of phosphatidylethanolamine/phophatidylcholine (3:1, weight/weight), 20% (volume/volume) glycerol and 50 mM KPi, pH 7.0, in a final volume of 1.0 ml, unless indicated otherwise. The suspension was incubated for 30 min at 4 "C and then centrifuged at 185,000 X g (4 "C) for 1 h. The supernatant (1 ml) was removed and mixed with 0.25 ml of 20 mg/ml of PE/PC (3:1, weight/weight) in 4% (weight/volume) octyl glucoside and 0.25 ml of 50 mM KP,, pH 7.0. The mixed micelle suspension was agitated by manual inversion of the tube and incubated for 10 min a t 4 "C. Proteoliposomes were formed by detergent dilution (35-fold) into 50 mM KPi, pH 7.0, or by detergent dialysis (against 500-fold volumes of 50 mM KPi, pH 7.0) as described (In't Veld et al., 1992).
Transport Assays-(i) Ap-driven Uptake. Fused membranes and proteoliposomes were diluted to a final protein concentration of about 0.25 mg/ml in KPM or in 25 mM potassium-Hepes, 25 mM potassium-Mes, 25 mM potassium-Pipes, 25 mM potassium-chloride containing 2 mM MgSO,, (HMP buffer), pH 5.0-8.0, supplemented with 200 p~ TMPD and 10 PM cytochrome c. The electron donor ascorbate (potassium-salt, 10 mM, final concentration) was added 2 min prior t o the initiation of transport, and the incubation mixture was kept under continuous aeration. Transport was initiated upon addition of small aliquots of radiolabeled substrates. At given time intervals samples were withdrawn, diluted with 2 ml of ice-cold 0.1 M of lithium chloride, filtered immediately on 0.45-pm cellulose nitrate filters (Schleicher & Schiiell GmbH, Dassel, Germany), and washed once with 2 ml of 0.1 M lithium chloride. Initial rates of uptake were determined from the amount of labeled substrate accumulated during the first 5 s. Transport assays were performed at 30 "C. Radioactivity was measured by liquid scintillation spectrometry.
(ii) Counterflow. Concentrated cell, membrane vesicle, fused membrane, or proteoliposome suspensions in KPM or HMP buffer were equilibrated in the presence of 2 mM TMG for 1 h at 42 "C. Aliquots (2 pl) were diluted 100-fold to final protein concentrations of 0.30 mg/ml (cells and membrane vesicles), 0.15 mg/ml (fused membranes), or 0.06 mg/ml (proteoliposomes) into KPM or HMP buffer of the indicated pH containing ["CITMG. Initial rates of counterflow were determined after 3 s of incubation. Experiments were performed a t 25 "C and uptake was assayed by filtration as described above.
(iii) Efflux and exchange. Membrane preparations were equilibrated in the presence of 2 mM ["CITMG or 5 mM [3H]melibiose a t were diluted into buffer without (efflux) or with 2 mM of TMG or 5 42 "C for 1 h as described under Counterflow, except that membranes mM of melibiose (equilibrium exchange). For efflux and exchange in the presence of artificially imposed diffusion potentials, the membranes were resuspended in the buffers specified below prior to loading with 2 mM [I4C]TMG or 5 mM [:'H]melibiose. To generate a A", valinomycin was added to the membrane suspension to a final concentration of 2 nmol/mg of protein. The membranes were washed and resuspended in 120 mM KP, or 120 mM sodium-phosphate (Nap,), pH 6.5, containing 2 mM MgS04. Potassium-loaded membranes were diluted into the same buffer (no gradient) or in Napi buffer (A", interior negative). Sodium-loaded membranes were diluted into the same buffer (no gradient) or in potassium-containing buffers (A", interior positive). For efflux and exchange in the presence of a SpH, interior alkaline, membranes were resuspended in 100 mM potassiumacetate supplemented with 2 mM MgSOl and 20 mM KP,, pH 6.5. Subsequently, the membranes were diluted into the same buffer (no gradient) or into 120 mM KP,, pH 6.5, containing 2 mM MgS04 (ApH, interior alkaline). For efflux and exchange in the presence of a SpH, interior acid, membranes were equilibrated in the presence of 120 mM KPi, pH 6.5, supplemented with 2 mM MgS04, and, subsequently, diluted 100-fold into 100 mM potassium-acetate, 20 mM KP,, 2 mM MgS04, p H 6.5. For the generation of a Ap (interior alkaline and negative or interior acid and positive), the properly oriented potassium-and acetate-diffusion gradients were combined. To impose artificial diffusion potentials in the pH range of 5 to 9, the phosphate in the buffers specified above was replaced by similar concentrations of Pipes plus Ches. Efflux and exchange were assayed at 25-37 and 7-20 "C, respectively, as specified in the legends to figures.
(iv) Imposed Ap-driven uptake. For uptake driven by artificially imposed diffusion gradients (A*, interior negative; ApH, interior alkaline and Ap, interior negative and alkaline), membranes were treated as described under Efflux and Exchange except that loading with radiolabeled substrates was omitted. The membrane preparations were diluted 100-fold into buffers supplemented with the appropriate concentration of radiolabeled substrate. Uptake was assayed a t 30 "C as described above.
Determination of the Electrical Potential Difference-The A" across the membrane (A", inside negative) was determined from the distribution of the lipophilic cation TPP+ using a TPP+-selective electrode (Shinbo et al., 1978). The A 9 was calculated from the steady state level of TPP+ accumulation and was corrected for concentration-dependent binding of the probe to the membrane (Lolkema et al., 1982).
Determination of the pH Gradient-The ApH across the membrane was determined from the fluorescence of pyranine (100 PM, final concentration) entrapped within fused membranes or proteoliposomes (Clement and Gould, 1981). External pyranine was removed by washing with the buffer indicated, and membranes were collected by centrifugation (185,000 X g, 45 min, 4 "C).
Determination of the Internal Volume-Trapped volume measurements were performed with the fluorophore calcein as described (Oko et al., 1982). A value of 8 pl/mg of protein was determined for fused membranes. Alternatively, the specific internal volume was determined from the equilibration of [I4C]TMG in the absence of Ap. This method yielded values of 4.1, 8.7, and 70.3 pl/mg of protein for the membrane vesicles, fused membranes containing cytochrome c oxidase, and proteoliposomes, respectively.
Miscellaneous-Protein was determined by the method of Lowry et al. (1953) in the presence of 0.5% SDS (Dulley and Grieve, 1975) with bovine serum albumine as a standard. P-Galactosidase activity was determined from the hydrolysis of ortho-nitrophenyl-P-D-galactopyranoside (Citti et al., 1965). Bovine heart cytochrome c oxidase was isolated according to described procedures (Yu et al., 1975).

Membrane Vesicles, Fused Membranes, and
Proteoliposones-Membrane vesicles of lactose-grown 5' . thermophilus A147 were isolated by a protocol developed for Lactococcus lactis (Otto et al., 1982). Although these membrane vesicles exhibited TMG counterflow activity, neither TMG nor alanine uptake driven by artificially imposed ion gradients could be demonstrated. Moreover, a large fraction of cytosolic pro-Lactose Transport in S. thermophilus 22089 teins (e.g. @-galactosidase) remained associated with the membranes which prevented the use of lactose as substrate in the transport assays. Fusion of the membrane vesicles with liposomes decreased the "leakiness" of the membranes (artificially imposed ion gradients were sustained for more than 2 min) and reduced the contamination with cytosolic enzymes. Fusion of the membrane vesicles with cytochrome c oxidase containing liposomes yielded membrane preparations in which, in the presence of the electron donor system ascorbate-TMPD-cytochrome c, proton motive force-driven uptake of TMG could be assayed (Fig. 1).
Alternatively, membrane vesicles were solubilized with noctyl-@-glucopyranoside in the presence of phospholipids and glycerol. Reconstitution of membrane proteins was performed by detergent dilution or detergent dialysis. Both methods were equally efficient in reconstituting TMG (lactose, melibiose) and alanine transport activities (data not shown). Coreconstitution of the streptococcal membrane proteins with bovine heart cytochrome c oxidase enabled us to demonstrate Apdriven lactose uptake in the proteoliposomes (not shown). Finally, efflux and exchange of galactosides were monoexponential and continued with pseudo-first order rate kinetics until nearly all radiolabel had disappeared from the membranes both in fused membranes and proteoliposomes. Treatment of the membranes with p-chloro-mercuribenzoic acid (100 p~, final concentration) resulted in inactivation of the lactose carrier, and these membranes displayed exit of TMG with a much lower rate constant (passive diffusion). The extent of release of sugars from the fused membranes and proteoliposomes with the carrier-mediated kinetics indicated that more than 95% of the membranes contained a lactose carrier molecule (data not shown).
Mechanism of Energy Coupling-The effect of ionophores and protonophores on the initial rate of TMG uptake, the steady state level of galactoside accumulation, and the magnitude of the components of the proton motive force were analyzed in membrane vesicles fused with cytochrome c oxi-   (Fig. 1). Nigericin, which dissipates the ApH, decreased the initial rate of uptake and lowered the TMG accumulation level (Fig. lA). Valinomycin, which dissipates the A*, had virtually no effect on the rate of uptake but lowered the steady state level of accumulation (Fig. 1B). The combination of valinomycin plus nigericin (Fig. IC) and the uncoupler carbonyl cyanide m-chlorophenylhydrazone (not shown), which abolish the total Ap, reduced transport to equilibration levels. Monensin, which converts the ApH into a ApNa+, lowered TMG transport in the presence of 5 mM sodium ions (not shown). The same TMG accumulation levels were reached irrespective of whether the ionophores (protonophores) were added prior to the initiation of transport or during the course of an experiment (indicated by arrows). Altogether, the results indicate that Lacs-mediated TMG transport is coupled to the proton rather than to a sodium motive force. Similar conclusions were reached from the uptake of lactose (@-galactoside) and melibiose (a-galactoside) in proteoliposomes and fused membranes, respectively (data not shown).
GalactosidelProton Stoichiometry-The galactosidelproton stoichiometry of Lacs-mediated transport was determined from the steady state accumulation levels of melibiose and the magnitude of the components of the Ap. Melibiose was used as substrate in these experiments because the noncarriermediated flux of the disaccharide was at least 10-fold lower than that of TMG (not shown). In general, accumulation ratios predicted by the thermodymic equilibrium levels are difficult to reach for substrates with hydrophobic properties (Maloney and Wilson, 1973;Driessen et al, 1987). At pH 6.5 a Ap of -154 mV (A* of -130 mV and ZApH of -24 mV) was generated by ascorbate-TMPD-cytochrome c oxidation in the proteoliposomes, and this Ap did not depolarize upon addition of galactosides. Steady state melibiose accumulation levels (melibiosei,/melibioseout) were reached after 30-40 min and depended on the external melibiose concentration. At thermodynamic equilibrium ApMEL/F equals n(-Ap), in which Lactose Transport in S. thermophilus A/.LMEL/F represents the melibiose concentration gradient (in mV) and n the number of protons translocated in symport with melibiose. By extrapolating A/.LMEL/F(-AP) to an external melibiose concentration of zero, the melibiose/H' stoichiometry was estimated to be one (Fig. 2). pH Dependence of Ap-driven Uptake-The effects of valinomycin and nigericin on the initial rate of TMG uptake and the components of the proton motive force were investigated further by titrating with the individual ionophores. As shown in Fig. 3A, nigericin lowered the ApH without having an effect on A\k and decreased the initial rate of TMG uptake. Valinomycin, on the other hand, dissipated the A\k and increased the ApH but had little or no effect on the initial rate of TMG uptake despite a significant drop in the total proton motive force (Fig. 3 B ) . These data are most easily explained by assuming that Lacs is activated at alkaline internal pH values (Fig. 3C). It was not possible to determine the pK, of the internal pH dependence precisely since the internal pH could only be manipulated in a narrow range without causing major changes in the driving force (Ap) of the transport process.
The effects of pH on the initial rate of TMG uptake were further analyzed at external pH values of 5.0, 6.0, 7.0, and 8.0 while the internal pH was manipulated by the ionophores valinomycin or nigericin. The transport rates increased with increasing external pH when the internal pH was raised in parallel, whereas minor effects were observed at varying external pH while the internal pH was kept constant (Fig. 4). The total Ap varied somewhat nonsystematically with the external (and internal) pH, i.e. maximal and minimal values were reached a t external pH values of 6.0 and 8.0, respectively (data not shown). Altogether, the complicated pH profiles as shown in Fig. 4 (and Fig. 3C) indicate that in the fused membranes protonation site(s) on the inner surface of the membrane control Ap-driven TMG uptake.
p H Profiles of Efflux, Exchange, and Counterflow-Efflux, equilibrium exchange, and counterflow were investigated at different pH values under conditions that Ap was zero. Concentrated membrane preparations (membrane vesicles or fused membranes) were equilibrated with 2 mM [14C]TMG (efflux and exchange) or [l*C]TMG (counterflow) and then diluted rapidly 100-fold into the same buffer devoid of TMG (efflux) or with 2 mM unlabeled (exchange) or labeled (counterflow) TMG. To measure the exchange reaction accurately Melibiose uptake by proteoliposomes (see "Experimental Procedures'') was assayed in 50 mM KPi, pH 6.5, containing 2 mM MgSO,, and supplemented with the electron donor system ascorbate-TMPDcytochrome c. Melibiose uptake was assayed for 60 min, and the steady state accumulation levels were estimated. In parallel experiments the magnitude of A* and ApH was determined. The apparent melibiose/proton stoichiometry was calculated from the melibiose concentration gradient (ApMEJF) and the Ap (=A* -ZApH). the temperature of the assay medium was set at 7 "C; efflux was assayed at 25 "C. As shown in Fig. 5A, the rate of TMG efflux increased with increasing pH with a pK, of about 8.2. Under conditions that an equimolar concentration of TMG was present externally, exit of [14C]TMG (equilibrium exchange) displayed an optimum at pH 7.2 (apparent pK values of 6.2 and 8.2). TMG counterflow exhibited a pH dependence similar to that of exchange (data not shown). Although exchange is one to two orders of magnitude faster than efflux around neutral pH, both activities become very similar at pH 10. Efflux and exchange of melibiose displayed pH profiles similar to that of TMG both in membrane vesicles (Fig. 5 B ) and in fused membranes (not shown), except that differences in rates of efflux and exchange were less pronounced than with TMG.
Effect of Membrane Potential and InternalpH on Efflux and Exchange-Galactoside/proton symport involves the net translocation of a charge (proton) across the membrane. Consequently, the reorientation of either the ternary carriergalactoside-proton complex and/or the unloaded carrier should involve net movement of charge. Since exchange is more rapid than efflux in the pH range of 5 to 9, the ratedetermining step for TMG and melibiose efflux down a concentration gradient could involve a reaction associated with the return of the unloaded carrier to the inner surface of the membrane. If net movement of charge is involved in this step, the membrane potential should affect efflux. Membrane potential (inside negative relative to outside) was imposed by means of a valinomycin-mediated potassium diffusion gradient (see "Experimental Procedures"). Since passive fluxes of melibiose are less manifest than those of TMG, the adigalactoside was used as substrate in the following experiments. The results presented in Fig. 6 show that efflux of melibiose from fused membranes was retarded by a membrane potential (interior negative) at pH 5.8 (panel A ) , pH 7.0 (panel C), pH 8.0 (panel E ) and pH 9.0 (not shown). Equilibrium exchange of melibiose, on the other hand, was not affected by the membrane potential at any pH tested (Fig. 6 B , D, and F, not shown). These results are consistent with a translocation cycle for efflux in which a negative charge moves to the inside during the reorientation of unloaded binding sites. Since exchange is unaffected by A\k it is unlikely that any of the translocation intermediates of this reaction carries a net charge.
T o discern external and internal pH effects in the pH dependences of the facilitated diffusion reactions (Fig. 5 ) , efflux and exchange of melibiose were assayed at pH 5. 8, 7.0, 8.0, and 9.0 in the presence of a pH gradient (inside alkaline relative to outside). To raise the internal pH an outwardly directed acetate diffusion gradient was imposed (see "Experimental Procedures"). As shown in Fig. 6, in the presence of a ApH the rate of efflux was reduced at all pH values tested. The relative effect of the ApH on the rate of melibiose efflux increased with increasing pH. Equilibrium exchange of melibiose was enhanced by ApH (inside alkaline) at pH 5.8 (Fig.  6B), not significantly affected at pH 7.0 ( Fig. 6 D ) , and retarded at pH 8.0 (Fig. 6 F ) and 9.0 (not shown). These results clearly indicate that the rates of efflux and exchange are differently affected by pH and that the observed pH dependence of exchange (Fig. 5 ) could be due to changes in the internal pH.

DISCUSSION
The kinetic mechanism of the lactose transport protein (Lacs) of S. thermophilus has been analyzed in fused membranes and in proteoliposomes. The effects of ionophores and FIG. 3. Effect of increasing concentrations of ionophores on TMG uptake and the magnitude of the components of the Apt and the internal pH dependence of the initial rate of TMG uptake. Fused membranes were resuspended in HMP buffer, pH 6.5, and energized by ascorbate-TMPDcytochrome c oxidation. Increasing concentrations nigericin (panel A ) and valinomycin (panel B ) were added. Conditions were similar to those described in the legend to Fig. 1 except that the final protein concentration was 0.24 mg/ml. The initial rate of TMG uptake was estimated from the amount of label accumulated over 5 s, and the magnitudes of the components of the Ap were estimated as described. Panel C, The dependence of the initial rate of TMG uptake on the internal pH (data were taken from panels A and B ) . For comparison the magnitude of the Ap is also shown. uncouplers on Ap-driven galactoside uptake indicate that transport proceeds in symport with a proton. Despite the similarities in the primary structure of Lacs and MelB , the Lacs carrier protein shows no substrate-dependent cation selectivity. By contrast, the melibiose carrier protein cotransports a-galactosides (and galactose) with H' , Na' , and to a lesser extent Li', whereas @galactosides are transported equally well with Na' and Li' but not with H' Leblanc et al., 1990). In fact, the functional characteristics of Lacs resemble more those of the lactose carrier protein (Lacy) of E. coli than those of MelB (see below).
Activation/inhibition of transport activity by pH can be envisaged in terms of two types of proton-binding sites (Poolman et al., 1987). The first type involves binding/release of the symported proton and can be interpreted as catalytic site. The second type of proton-binding site is not directly involved in the catalytic mechanism but affects transport allosterically. When Lacs facilitates Ap-driven uptake the galactoside and proton are released on the inner surface of the membrane. Since Ap-driven TMG uptake is stimulated at alkaline inter- Internal p H nal pH values one could argue that under these conditions the translocation cycle is rate-limited by the release of the catalytic proton on the inner surface of the membrane. In fact, the effect of a large drop in the A\k on the rate of Apdriven uptake can be compensated by a relatively minor increase in the ApH (internal pH) (Figs. 1 and 3). Notice that a lowering of A\k does result in a lower level of TMG uptake (Fig. lB), indicating that the steady state level of galactoside accumulation is coupled to Ap. Thus, the ApH component of the Ap not only acts as a driving force for transport but also affects transport by influencing the equilibrium between the protonated and deprotonated forms of the carrier protein or the rate of proton transfer from the protein to the solvent on the inner surface of the membrane (Fig. 7). An increase in ApH (or internal pH) would shift the equilibrium to the unprotonated form of the carrier protein, and as a result the influx of TMG is accelerated.
For efflux down a concentration gradient, a proton and galactoside molecule have to be bound by the carrier protein on the inside, and both have to be released on the outside. The inhibition of efflux by a ApH (interior alkaline) is in Lactose Transport i n S. thermophilus accordance with a lower rate of protonation on the inner surface of the membrane or a decrease in the relative abundance of the protonated carrier intermediate. Under conditions that the internal and external pH are equal (Fig. 5), an increase in pH enhances the release of the proton on the outside but at the same time decreases the protonation of the carrier on the inside. The observed increase in the rate of TMG (and melibiose) efflux with increasing pH should therefore be the resultant of two opposite pH effects, one being exerted on the inner and the other on the outer surface of the membrane. Consequently, a decrease in the rate of efflux is ultimately expected with increasing pH, i.e. at higher pH values than used in these experiments.
When saturating amounts of galactosides are present both on the inside and the outside of the membrane, i.e. in the exchange and counterflow experiments, the rates of radiolabel equilibration are faster than the rates of efflux and exhibit a pH optimum. Above pH 7 the rate of exchange decreases sigmoidally with a pK, of about 8, and below pH 7 the exchange activity decreases with a pK, of about 6. To discriminate between external and internal pH effects, the rates of exchange have been estimated at fixed external pH values and in the presence and absence of an imposed ApH (Fig. 6).
An increase in internal pH ( ApH) at external pH > 7 inhibits the rate of melibiose exchange, whereas a similar pH stimulates exchange at external pH < 7 . These observations indicate that (de)protonation steps on the inner surface of the membrane affect the rate of galactoside exchange. To interpret the effects of pH on Lacs-mediated galactoside efflux and exchange, the reactions involved in the translocation are schematically represented (Fig. 7 ) . In this scheme the ternary carrier-galactoside-proton (CLH) complex is formed on the inner surface of the membrane through the binding of either galactoside first and proton last (bold type letters) or proton first and galactoside last, the ternary complex reorients binding sites, and the galactoside and proton are released on the outside in unspecified order. From this point on efflux and exchange differ. During efflux the unloaded carrier protein reorients its binding sites, whereas during exchange the carrier cycle proceeds in the opposite direction ("backward reaction"), i.e. another galactoside molecule and proton bind, Experiments were performed in 30 mM potassium-citrate, 30 mM KPi, 30 mM potassium-Ches, 2 mM MgS04 of the indicated pH, supplemented with valinomycin (50 nM) plus nigericin (5 nM), and at a final protein concentration of 0.3 mg/ml as described under "Experimental Procedures." The release of [14C]TMG and [3H]melibiose from the membranes in the assays of efflux and exchange was monoexponential and first order rate constants ( k = ln2/tBh) could be estimated. The rates of efflux and exchange were obtained by multiplying k with the internal galactoside concentration (2 mM or 8.2 nmol/mg of protein for TMG; 5 mM or 20.5 nmol/mg of protein for melibiose). Efflux and equilibrium exchange of TMG were assayed at 25 and 7 "C, respectively; efflux and equilibrium exchange of melibiose were performed at 35 and 15 "C, respectively. For exchange of melibiose the results of two independent experiments ( 0 , O ) are shown. the ternary complex reorients its binding sites, and the proton and galactoside are released in the internal medium. In this scheme protonation and deprotonation steps (step 2 ) take place during exchange on the inner surface of the membrane if the ternary complex is formed through the binary CLcomplex (Fig. 7 , steps 1 and 2 ) . If the exchange pathway proceeds via the binary CH complex the galactoside may associate/dissociate (step 2 ' ) without association/dissociation of the proton (step 1') (Fig. 7 ) . Since an imposed ApH has converse effects on the rate of exchange at pH < 7 and pH > 7 , one could argue that exchange is rate-limited at "high" internal pH by the protonation of the carrier molecule ("forward reaction") and at low internal pH by the deprotonation (backward reaction). The resultant of the two opposing (internal) pH effects may lead to an optimum in the pH depend- (interior negative, 0) or ApH (interior alkaline, 0) (see "Experimental Procedures''), and devoid of melibiose (efflux, panels A , C, and E ) or supplemented with 5 mM melibiose (equilibrium exchange, panels B , D, and F ) . The medium pH was 5.8 (panels A and B ) , 7.0 (panels C and D ) or 8.0 (panels E and F). Efflux and exchange were assayed at 37 and 20 "C, respectively. ence of exchange. Assuming ordered binding with proton first and galactoside last (steps 1' and 2 ' ) , exchange may be inhibited by high internal pH, due to a lowering of the concentration of the binary CH intermediate, but should not be affected by a low internal pH. In that case one has to invoke additional regulatory (allosteric) (de)protonation steps t o explain the pH dependence of the rate of exchange. In conclusion, by interpreting the pH profiles in terms of catalytic (de)protonation steps the results are consistent with a mechanism in which the binding of the galactoside occurs first and proton last. The observed pH dependences of the rates of efflux and Ap-driven uptake do not allow a distinction t o be made between an ordered or random binding mechanism.
Since the rate of efflux is affected by imposed A'k, whereas the rate of exchange is not, the unloaded carrier protein most likely carries a negative charge (Cin Fig. 7). Finally, the effects of ApH and A'k on efflux and exchange indicate that the components of Ap affect different steps in the translocation process, and, as a consequence, a unique relationship between the rate of uptake and Ap upon selective manipulation of either A 9 or ApH is not to be expected.
The translocation scheme put forward for Lacs-mediated galactoside transport differs from that proposed for the Lacy protein of E. coli. These differences have for the greater part their origin in the pH dependence of the exchange reaction catalyzed by Lacs which is claimed to be independent of pH in case of Lacy Kaczorowski et al., 1979;Garcia et al., 1983;Viitanen et al., 1983;Kaback, 1990), although other studies indicate that pH effects can also be observed during Lacy-mediated exchange (Wright, 1986;Page, 1987). The effect of A\k on the rate of galactoside efflux, and the lack of a A\k effect on rate of exchange (Fig. 6) are similar for Lacs and Lacy.
It has been suggested that lactose uptake by S. thermophilus is facilitated by a galactose/lactose antiporter (Hutkins and Ponne, 1991). These studies were performed in whole cells in which the lactose taken up is immediately hydrolyzed and the galactose, used to preload the cells, is slowly metabolized,2 which complicates the analysis of the transport mechanism. The evidence presented in this paper indicates that the lactose transport protein catalyzes lactose (galactoside)/proton symport as well as homologous and heterologous exchange. The requirement for an alkaline internal pH for maximal lactose/ proton symport activity and the observation that the maximal rate of exchange is at least 10-fold higher than the maximal rate of Ap-driven uptake supports the suggestion that the lactose transport system may predominantly catalyze lactose/ galactose exchange under physiological conditions . Notice that S. therrnophilus usually grows between pH 7 and 5 , and, depending on how well the organism can regulate the intracellular pH Ap-driven lactose uptake may be far too low to meet the observed lactose utilization rates .