Energy Coupling of the @Methylgalactoside Transport System of Escherichia coZi*

coupling for the concentrative uptake of galactose via the P-methylgalactoside transport system occurs via an intermediate produced during oxidative phosphorylation but before ATP formation. The membrane-bound ATPase, necessary for ATP formation during oxidative phosphorylation, is not obligatory for energy coupling of transport but can, under anaerobic conditions, accomplish energy coupling by ATP hydrolysis. These conclusions are supported by the following findings:

These conclusions are supported by the following findings: (a) galactose uptake in intact cells can be stimulated by substrates of the respiratory chain and inhibited by inhibitors of the respiratory chain and by uncouplers of oxidative phosphorylation; (b) cells made anaerobic for the transport assay are fully transport active; (c) a mutant defective in the membrane-bound ATPase has lost transport activity unless supplied with exogenous D-lactate, and the additional presence of arsenate completely blocks the stimulation by D-lactate; (d) a mutant defective in the respiratory chain due to a defect in ubiquinone biosynthesis still exhibits transport activity, which cannot be stimulated by D-lactate.
The accumulation of substauces within a bacterial cell against the concentration gradient requires an expenditure of energy. Despite many recent studies the molecular mechanism by which energy coupling to bacterial active transport systems is accomplished remains an intriguing problem. Several different proposals have been made for the energy source of the lactose transport system. Scarborough et al. (1) reported that ATP, under conditions where the cells were made permeable to this compound, can increase the rate of downhill transport of ONPGal,' indicating that energy could be derived * This work was supported by grants from the National Institutes of Health (GM-18498) and the National Science Foundation (GB 38785X).
1 In partial fulfillment of the requirements for the Degree of Bachelor of Arts in Biochemistry.
1 The abbreviations used are: ONPGal, o-nitrophenyl-p-ngalactopyranoside; TMG, methyl-l-thio-p-n-galactopyranoside. All other sugars mentioned in the text have n-configuration. This observation led them to suggest that the energy source for the lactose system might be linked to a proton or ion gradient across the cell membrane in accordance with the chemiosmotic hypothesis of energy coupling originally proposed by hIitchel1 (6). The chemiosmotic coupling mechanism has received additional support from West (7, S), who reported that poisoned cells take up protons concomitant with the downhill transport of lactose.
Recent studies from Kaback's laboratory (9-16) have provided an entirely new approach to the problem of energy coupling iu bacteria. Using membrane vesicles devoid of cytoplasm and incapable of oxidative phosphorylation, Kaback and his colleagues (10, 11) could dcmou strate that concentrative uptake of lactose as well as other sugars and amino acids is coupled primarily to a membrane-bound, 2 Studies of active transport in ATPase-deficient mutants have led to somewhat confusing-results.
The uptake of TMG, proline, serine, and lysine by whole cells and membrane vesicles of AN120 has been reported to be normal in a recent study from Kaback's laboratory (16). Schairer and Haddock (2) also showed normal uptake of TMG in a different Mgzi-, Ca2+-stimulated ATPase mutant of E. co&. Simoni and Shallenberger (3) have reported a decreased ability to transport proline andalanine in an&her E. CO& strain deficient in this enzyme. Kaback (15) has suggested that the ATPase mutant studied by Simoni and Shallenberger may have more than one defect.
Discrepancies in the basal rates of transport, i.e. in the absence of inhibitors or energy sources, for substrates of different transport systems may be a reflection of the relative degree of dependency of these systems upon each of the two pathways by which energy can be derived for transport. 4429 by guest on March 17, 2020 http://www.jbc.org/ Downloaded from flavin-linked n-lactate dehydrogenase, but not to NADH dehydrogenase.
Therefore, active transport of lactose in membrane vesicles seems to depend on electron transport but not upon the participation of a high energy intermediate of oxidative phosphorylation.
From these data a model was proposed for the active transport of lactose and other substrates; the carrier molecules were thought to be obligatory intermediates of the respiratory chain located between the n-lactate dehydrogenase and cytochrome br (11, 12). However, the isolation of electron transfer coupling mutants, which oxidize n-lactate and other electron donors at a normal rate but cannot catalyze respiration-linked transport, necessitated the modification of this model in order to account for electron transfer in the absence of active transport (15). It was therefore suggested that the transport carrier molecules may be components of shunts off the main part rat,her than obligatory intermediates of the respiratory chain (15). While studies with membrane vesicles of E. coli have contributed to our understanding of energy coupling of the lactose transport system, as well as other sugar and amino acid transport systems, they seem not to be as useful in the st'udy of transport systems involving periplasmic binding proteins.
Vesicles of E. coli ML3 and ML35 have lost their galactose-binding protein together with a high affinity transport system for galactose, the P-methylgalactoside transport system (14). Any attempt to restore transport activity of the &methylgalactoside transport system by addition of purified galactose-binding protein to membrane vesicles has failed so far.
Neither has it been possible to demonstrate any binding of radioactively labeled galactose-binding protein to membrane vesicles.3 In the present paper we attempted to study the energy coupling of the P-methylgalactoside transport system mediated by the periplasmic galactosebinding protein in whole cells.

Bacterial Strains
Unless otherwise specified all tests were performed with strain W3092cy-(ATCC25939) (F-, g&K, ZCICY).~ This strain is endogenously induced for the /3-methylgalactoside transport system and the synthesis of the galactose-binding protein and has been described (18). Four strains were obtained from Dr. Cox: AN180 (F-, argES, thi-1, strR) (19), AN120 (F-, argE3, thi-I, strR, uncA 401) (19),AB2154 (Hfr, metE, thr-l, Zeu-6) (20), AN59 (Ilfr, thr-I, Zeu-6, u&B-) (20). Before using these strains, single cell colonies were isolated and tested for their amino acid markers by growth on minimal plates containing 0.2y0 glucose in the presence and absence of the respective amino acid. The uncA@l and the ubiB markers were tested by the ability of the strains to grow aerobically on n-lactate or succinate as sole carbon source. It was noted that strain AN59 exhibits a high reversion rate to ubiB+.
Therefore, after measuring /!-methylgalactoside transport activity, the ubiB marker had to be rechecked routinely to avoid misinterpretation caused by the outgrowth of ubiB+ reversions.

Growth oj Bacterial Strains and Preparation of Bacterial Suspension
Used for Transport Assay All strains were grown and prepared for the transport assay as previously described (21) (17).

Transport Assay
Initial Rate of Entry of Galactose-The assay for the initial rate of galactose entry was performed as described elsewhere (21) with the modifications indicated below. When energy poisons were used during the assay, they were added in maximally 100 ~1 of concentrated solutions either together with the radioactive galactose or 10 min prior to the addition of labeled galactose. Sub strates to stimulate transport were added routinely 5 min prior to the addition of labeled galactose.
Some components with limited solubility in water were dissolved in maximally 100 ~1 of dimethylsulfoxide, which by itself did not change transport activity. When uptake under anaerobic conditions was measured, the cell suspension prepared for the transport test was incubated at 37" for up to 3 hours in the presence of chloramphenicol (100 pg per ml) while oxygen-free helium was bubbled through the suspension continuously.
After temperature equilibrium to the assay tenperature 5 ml were quickly removed and transferred into a vessel filled with CO, and containing 50 ~1 of [I-%]galactose, resulting in a final concentration of 0.5 ELM. Aliquots (0.5 ml) were removed as quickly as possible and filtered as usual.
Znitial Rate of Exit of Galactose-The assay for the initial rate of exit of galactose has been described (21). To measure the effect of n-lactate on exit of preaccumulated [I-i4C]galactose, n-lactate was added to the lo-ml test suspension of preloaded cells in 20 ~1 of concentrated solution 5 min prior to the addition of excess unlabeled sugar or energy poisons. When energy poisons were utilized in the exit assay, they were added to the test suspension in 50 ~1 of solution at time 0.

Combined Treatment of Temperature and Osmotic Shock to Increase
Permeabiliiy of A TP The following procedure was done according to a suggestion of Dr. M. Cashel, National Institutes of Health.5 After growth in nutrient broth the bacterial cells were washed once with Medium A at 25-30" and centrifuged at the same temperature. They were resuspended in 0.1 M Tris-HCl, pH 7.3, at room temperature and centrifuged at 2530". The supernatant was carefully removed and the pellet was resuspended in ice-cold 1 mM Tris-HCI, pH 7.3, containing 1 mM MgC12. The cells were centrifuged at O-4" and resuspended in ice-cold Medium A to an optical density of 0.3 at 650 nm.
For the transport test this suspension was then treated as described above.

EJect of DiJerent
Energy Poisons on Initial Rate of Galactose Gptalce- Table  I shows the ability of different substances to interfere with the initial rate of entry of galactose at 23". Uncouplers of oxidative phosphorylation as well as inhibitors of glycolysis show varying degrees of inhibition.
Most of the inhibitors are temperature-dependent in their inhibitory effect, being two to three times more effective at 23" than at 15". The effect of 30 m&r potassium cyanide is rather complex. Almost no inhibition of uptake is observed when it is added together with the [1-i4C]galactose.
The inhibitory effect increases with time and no transport activity can be observed after 10 min of incubation.
This time dependence of potassium cyanide inhibition is similar to that previously described for p-HMB inhibition of galactose uptake (21). Yet in contrast to the effect of p-HMl%, glucose-stimulated exit of galactose from preloaded cells is not inhibited even after 3 min incubation with 30 mM potassium cyanide.
Effect of n-lactate on Rate oj Entry, Exit, and Steady State of Bccumulation of Galactose- Fig.  1 shows that n-lactate stimulates the initial rate of uptake of galactose about 2-fold. The rate of exit as measured by addition of glucose, sodium azide, or CCCP to cells preloaded to an internal [I-14C]galactose concentration of 1 rnx shows no significant difference whether or not n-lactate was present prior to the initiation of exit (Fig. 2). Therefore, n-lactate must act primarily on the entry process. This finding supports results obtained with energy poisons (21) and indicates energy coupling to the entry process.
As a result of the increased rate of uptake in the presence of n-lactate, the ratio of internal to external galactose is increased in cells which have reached their steady state of accumulation.
The addition of n-lactate results therefore in a 25 to 30% decrease of the external galactose concentration (Fig. 2D).
The stimulation of initial rate of entry by n-lactate also shows a temperature dependence similar to that ohserved with energy poisons.
Its maximum appears to be at 25". Eflect OJ Di$erent Metabolites on Initial Rate of Galactose r7~,take- Table  11 shows the ability of a variety of metabolites to stimulate the initial rate of uptake of galactose after incubatidn with cells for 5 min at 23". From the  Cultures grown to stationary or logarithmic growth phase were washed with ice-cold Tris-HCl buffer and then subjected to a rapid 3-fold dilution according to the procedure described by Scarborough et al. (1). Cells treated in this manner did not show any effect on the initial rate of uptake of galactose (not shown).
The presence of ATP during the transport assay or, in addition, during the shock procedure, again caused the same slight inhibition observed in untreated cells.
The same results were obtained with the Tris-EDTA procedure described by Leive (22,23), a procedure known to remove large portions of the lipopolysaccharide content of the cell. In contrast, the rate of galactose uptake is decreased considerably by subjecting the cells to a simultaneous temperature and osmotic shock suggested by M. Cashel and described under "Materials and Methods." This treatment results in a loss of small molecules from the cytoplasm, including nucleotides but not proteins5 Immunodouble diffusion test with toluenized cells showed that this treatment does not result in loss of the periplasmic galactosebinding protein (not shown). Also, when cell suspensions of shocked and nonshocked cells prepared for the transport assay were plated for viable counts, the shocked cells showed no reduction in viability as compared with nonshocked cells. Fig. 3 shows that cells made permeable to small molecules by the combined temperature and osmotic shock treatment do not show any stimulation of transport activity by the addition of ATP. In contrast, ATP still inhibits slightly, as in nontreated cells. It is clear that the treated cells have not lost part of their transport capabilities since the addition of n-lactate still stimulates the initial rate of uptake of galactose (Fig. 3) Cells prepared for the transport assay were measured at 23" but otherwise as described in the legend to Fig. 1. A, wild type strain AB2154; B, strain AN59 defective in ubiquinone biosynthesis.
Open symbols, control; closed symbols, incubation with 20 mM n-lactate 5 min prior to the transport assay; haZf,liZZed symbols, incubation with 30 mM arsenate 10 min prior to the transport assay.
cedure of Neu and Heppel (24), known to remove periplasmic proteins from the cell envelope of E. co& the transport activity of the shocked cells is dramatically reduced and no stimulation by n-lactate can be observed (Fig. 3).

Galactose Uptake and Galactose-binding Protein Synthesis in Strains Dejective in Membrane-bound ATPase and in Ubiquinone
Biosynthesis-The finding that inhibitors of both oxidative phosphorylation and glycolysis are able to inhibit galactose transport suggested that energy derived from oxidative phosphorylation is a possible but not obligatory energy pathway. This idea was supported by the finding that cells kept anaerobic under helium at 37" for times up to 3 hours in the absence of any exogenous energy source and in the presence of chloramphenicol showed the same transport activity as a parallel cell suspension which was kept aerobic. It was therefore of interest to see whether or not enzymatic defects in the respiratory chain or oxidative phosphorylation would show any interference with transport activity of galactose. Two mutants were available, one defective in ubiquinone biosynthesis (20) and one defective in the membranebound Mg*+-, Ca*+-stimulated ATPase (19). Neither strain can grow on D-lactate or succinate, and both show reduced growth rates in nutrient broth. Fig. 4 shows the transport activity of the ubiB strain AN59 in comparison to its ubiBf parent AB2154. As can be seen, the defect in ubiquinone biosynthesis reduces but does not abolish transport activity, and the addition of D-lactate to the mutant stimulates only to an insignificant extent. Arsenate inhibits the galactose uptake in both the mutant and its parent. The behavior of the ATPase defective mutant AN 120 in comparison with its parent AN180 is shown in Fig. 5. The mutant exhibits only residual transport activity. However, the addition of D-lactate almost completely restores the initial rate of uptake to the value observed in tJe parent strain, while the additional presence of arsenate completely counteracts the stimulating effect of D-lactate. This clearly shows that in the absence of a functional ATPase, the energy necessary for the uptake of galactose against the concentration gradient can be derived from an intermediate generated during oxidative phosphorylation, but before ATP formation. It should be mentioned that both parents of the respective mutants show, in comparison with our wild Cells prepared for the transport assay were measured at 23" but otherwise as described in the legend to Fig. 1. A, wild type strain AN180; B, mutant strain AN120 defective in the membrane-bound ATPase. Open symbols, control; closed symbols, incubation with 20 mM Dlactate prior to the transport assay; ha&filled symbols, incubation with 30 mu arsenate 10 min prior to the transport assay. In case of the mutant AN120 (B) D-lactate, 20 mM, was present in addition to arsenate. 4 FIG. 6. Galactose-binding protein synthesis in mutants of ubiquinone biosynthesis and the membrane-bound Mgz+-, Ca2+dependent ATPase as measured by cross-reactivity against antigalactose-binding protein antibodies. Ouchterlony immunodiffusion and preparation of the bacterial extracts were performed as previously described (25). 1, purified galactose-binding protein (0.1 mg per ml); &, extract of strain AN59; 9, extract of strain AB2154; 4, extract of strain AN120; 6, extract of strain ANlSO. All strains were grown in the presence of 1 mM fucose. type W3092cy'-, a rather low transport activity, which cannot be further induced by the presence of D-fUCOSe during growth. However, neither mutation affects the amount of galactose-binding protein synthesized, in comparison with its respective parent, as judged by the ability of toluenized cells to cross-react with antigalactose-binding protein antibodies (Fig. 6). Again, this amount is smaller than that observed in our wild type W3092cy-, and might account for the relatively low transport activity of these strains. The results presented here indicate that energy for the concentrative uptake of galactose mediated by the P-methylgalactoside transport system is derived from a high energy intermediat.e which can be generated either during oxidative phosphorylation before the formation of ATP or directly from ATP hydrolysis, via the Mg2+-, Ca2+-stimulated, membrane-bound ATPase. It was found that substrates of the respiratory chain or metabolites which can be converted in viva into respiratory substrates stimulate transport up to a-fold. Although this effect does not reveal the site at which energy coupling occurs, it does indicate that glycolysis itself is not required.
A variety of energy poisons were shown to be potent inhibitors of the initial rate of galactose uptake.
The most effective were the respiratory inhibitor cyanide and the uncouplers of oxidative phosphorylation: dinitrophenol, CCCP, azide, and arsenate. The inhibition by uncoupling agents indicates that respiration is not sufficient for the accumulation of galactose. It was also observed that aerobically grown cells could be incubated anaerobically for up to 3 hours (in the absence of protein synthesis) mithout reduction of their ability to actively transport galactose. Taken toget,her these inhibitor studies suggest that energy can be derived from an intermediate of oxidative phosphorylation under aerobic conditions or from glycolysis (via ATP hydrolysis) under anaerobic conditions. However, it should be noted that the strong inhibition of galactose uptake by cyanide does not appear consistent with the failure of anoxia to reduce transport activity. In addition, we were not able to directly demonstrate any stimulation of transport by exogenous ATE' even under conditions which are known to allow the penetration of nucleoside triphosphates (23,26).
ATP showed a slight but reproducible inhibition of transport as observed by Knappe et al. (27) for the uptake of arginine by E. coli membranes.
The failure of ATP to stimulate transport might possibly be explained by a strong feedback regulation of transport activity by the energy charge of the cell, i.e. the ratio: ((ATP) + Js(ADP))/((ATP) + (ADP) + (AMP)) (28). One would therefore expect transport activity to be high at a low energy charge in order to accumulate more metabolizable substrates, and low at a high energy charge, i.e. an inhibitory effect by a high ATP concentration.
The strongest evidence for the conclusions concerning energy coupling are derived from measurements of galactose uptake in mutants defective in oxidative phosphorylation. One mutant, AN59, which is defective in ubiquinone biosynthesis, is unable to grow on n-lactate or succinate because of its respiratory chain deficiency.
This strain also exhibits a reduced galactose transport activity which cannot be significantly stimulated by n-lactate.
Since the mutant cannot derive energy from respiration or oxidative phosphorylation, it is dependent upon the production of ATP via glycolysis for all energy-requiring processes. Thus, its reduced transport activity reflects the loss of one of the two methods of generating energy which can be coupled to galactose accumulation.
Furthermore, it was shown that the residual transport activity of this mutant was inhibited by the phosphate analogue arsenate.
This inhibition is expected based on the demonstrated ability of arsenate to reduce intracellular ATP and phosphoenolpyruvate levels in intact cells of E. coli (4). The other mutant studied, AN120, can oxidize metabolites via the electron transport chain but is deficient in the Mg2+-, Caa+stimulated, membrane-bound ATPase. Thus, the failure of this strain to grow on n-lactate or succinate results from its inability to couple respiratory chain oxidation to ATP formation. As a result, no energy for galactose uptake can be generated from *4TP.
This mutant was shown to be almost completely devoid of transport activity in the absence of added energy source. However, n-lactate stimulates galactose uptake to the level observed in the parent strain, indicating that the effect of this metabolite is not mediated by the generation of ATP. It was also shown that arsenate completely blocks the stimulatory effect of n-lactate, demonstrating that respiration alone is not enough for n-lactate to influence transport.
Arsenate acts as a "secondary" uncoupler of oxidative phosphorylation; although its precise mechanism of action is unknown, it is thought to replace phosphate by forming an unstable arsenyl intermediate which is thrn spontaneously hydrolyzed (29). Arsenat,e could therefore tlischarge an energy-rich intermediate of oxidative phosphorylation even in the absence of the enzyme responsible for the final step(s) in the generation of ATP. Such a mode of action would explain its observed effects upon galactose transport.
One might expect arsenate to cause a stronger inhibitioil 01 transport in wild type cells in view of its actions in mutants, each of which larks one of the two pathways postulated for energy coupling.
However, in cells exhibiting normal glycolysis and oxidative phosphorylation, the presence of more than one site for arsenate action could explain a reduction in the over-all effectivcness of this inhibitor, especially since the standard assay medium is high in phosphate.
Although not shown here, in a phosphatcfree medium 30 mM arsenate does inhibit the initial rate of galactose uptake by 85% and prevents the stimulation by lactate in wild type cells (W3092cy-).
As discussed earlier (see introduction) other studies of activrl transport in E. coli have also led to the conclusion that rnrrgy can be derived either from respiration or from glycolysis (2)(3)(4). The present findings further show that, at least in the cast of galactose, the link between electron flow and sugar transport is not an immediate one; respiration is not sufficient for galactosc, accumulation, but ,4TP need not be generated. Hong and Kaback (15) arrived at a similar conclusion in their studies of amino acid uptake in electron transfer coupling mutants.
The results do not give any information about the nature of the high energy intermediate which serves as the direct source' of transport energy.
In terms of current hypotheses for oxidativc phosphorylation the observations are consistent with and cannot distinguish between the generation of a proton or electrochemical gradient, a high energy compound, or an energized state of the bacterial membrane.
The scheme for energy coupling presented here bears close> analogy to the one postulated for the energy-dependent reduct'ion of NADP by NADH in E. coli membranes. The direct energy source for this transhydrogenase act,ivity is most likely a high energy intermediate which can be dissipated by uncouplers and which can be generated by either of two pathways, respirator) chain oxidation of metabolites or ATP hydrolysis in the prcsencc of a functional Mg7+-, Ca2+-stimulated ATPase (30)(31)(32)(33). In a previous paper (21) we presented evidence suggesting that energy coupling for the /3-methylgalactoside transport system occurs during the entry rather than the exit of galactose: (a) the inhibition of galactose uptake by energy poisons was too great to be accounted for by an increased rate of sugar efflux; (b) no counterflow could be demonstrated; and (c) sodium azide did not change the K, for the exit of galactoi;e induced by glucose. Thr effects of n-lactate reported here also support this conclusion. n-Lactate was used as a representative respiratory substrate anal was found to stimulate the initial rate of galactose uptake 2-t,o a-fold. In contrast, it did not induce a significant change in the rate of exit of galactose from preloaded cells either in the presence or absence of energy poison.
If energy coupling were to occur during the exit process, one would expect n-lactate to act by decreasing the exit of galactose rather than by increasing the entry.
Consequently the present observations also point to the entry process as the site of energy coupling.
The present state of knowledge precludes characterization of the molecular mechanism by which energy coupling to transport is accomplished.
The problems involved in studying such a multicomponent, membrane-localized system are precisely the ones that have caused elucidation of the mechanism of oxidative phosphorylation to be one of the most persistent and elusive problems in biochemistry.
~lcknowledgments-We wish to thank Dr. Herman Kalckar for his generous hospitality and his encouragement during this work.