Eur. J . Biochem. 102, 231 -236 (1979)

A Novel ATP-Driven Glucose Transport System in Escherichia coli Lrwin F. WAGNER, Jill D. FABRICANT, and Manfred SCHWEIGER lnstitut fur Biochemie, Naturwissenschaftliche Fakultit der Universitit Innsbruck (Received July 13, 1979)

In Escherichia coli wild-type cells and in ATPase-deficient cells (unc mutants), glucose was found to be transported mainly by an ATP-driven system. The evidence is based on experiments involving interference at different sites of energy metabolism with the use of uncouplers, arsenate, and starved cells. Furthermore, addition of succinate to starved cells increased glucose uptake only in the wild-type cells, where ATP could be regenerated. Glucose transport was also ATP-dependent in cells deficient in methyl-p-galactoside transport (a system that carries glucose specificity). It was found to be shock-sensitive in all strains tested. The novel ATP-driven glucose transport is a highaffinity (K, 3- 10 pM) and high-capacity ( V 240- 330 nmol . min-' . mg cell protein-') uptake system.

It is generally accepted that Eschrrichia coli transports glucose mainly by the phosphoenolpyruvatedependent hexose phosphotransferase system (reviewed in [1,2]). The biochemistry of this system is well established : the energy-rich phosphate is transferred to the sugar from phosphoenolpyruvate via intermediates, involving several enzymes, divalent cations and lipids. The phosphate transfer to the sugar is coupled to its translocation across the membrane. Besides the phosphotransferase system, prokaryotes possess two other major types of active transport systems for sugars, the proton-motive-forcedriven and the ATP-dependent systems (for a review, see [3]). The proton-motive force is generated by the proton-extruding oxidative reactions of electron transport or by the hydrolysis of ATP by the energytransducing ATPase [4- 61. On the other hand, ATPdriven transports are not necessarily dependent on the generation of a proton-motive force, but are directly dependent on ATP or some other high-energy phosphate compounds [7]. The interconversion of a proton-motive force and ATP does not occur in ATPase-deficient cells (unc mutants [7,8]). Thus, unc mutants in E. coli provide a unique method for distinguishing between transport systems dependent on proton-motive force and on ATP. Unexpected properties of glucose transport were observed in our laboratory in experiments dealing with the regulation of sugar uptake by the phosphotransferase system. As observed previously [9 - 131, .~

Ahhwviarion. MeGal, methyl p-D-galactoside. Enzjmc. ATPase (EC 3.6.1.8).

methyl a-D-glucoside and other phosphotransferase substrates were accumulated at an increased rate when the energy content of the membrane was released, while glucose transport was reduced. These observations prompted us to analyse the properties of the glucose uptake in greater detail. We observed glucose to be transported in E. coli wild type and in the unc mutants mainly via an ATPdependent system. The techniques used were interference at different sites of energy metabolism with uncouplers or arsenate. In addition, experiments dealing with the starvation of cells, and the subsequent addition of substrates which restore ATP, confirmed the ATP-dependent glucose uptake. This glucose transport system was independent of the methyl-8-galactoside (MeGal) system, a known ATPdriven system that also carries specificity for glucose. However, it was found to be shock-sensitive, a property characteristic of ATP-driven and binding-proteindependent systems.

METHODS Bacteria, Media and Growth of Cells The bacterial strains used in these experiments are Escherichia coli B,-1 [14], B 834 [15], K 802 [15], BH 273 (uncAB [16]) and LA 3600 (galE,arg,his, rng1,lacY [17]). The cells were grown at 30 'C to midlog phase (4-5 x lo8 cells/ml) in minimal phosphatebuffered media containing Mg2+ with a carbon source of either 22 mM glucose or 54 mM glycerol. Cells

Glucose Transport in Escherichia coli

232 1

I

I

3

I

I

5

10

ZS 0 0

.v)

-" 0 3

Cn

b

-

5.0

A2 7,

c

0

E 2.5

0

0

I

I

5

10

i

0 Time (rnin)

n

Fig. 1. ,?fleet of'uncouplers on uptake ofmethyl a-glucoside andglucose in E. coli B,-I.Cells (4 x lo8 cells/nil) were grown with 22 mM glucose (79 Ci/mol, 100 pM) (A) or ~ - [ U - ' ~ C ] g l u c o s(248 e Cijmol, 100 pM) (B) were allowed as a carbon source. Methyl a-~-[U-'~C]glucopyranoside to accumulate for 4.5 min before the uncoupling agent (0) carbonyl cyanide m-chlorophenylhydrazone (final concentration 20 pM) or (A)dinitrophenol (tinal concentration 1 mM) was added; (0) control. Transport was determined at 30 C as described

prepared for the starvation experiment were grown in glucose minimal media. Cells were pelleted, washed twice with medium without a carbon source and suspended in the same phosphate-buffered medium. Cell viability during starvation was checked by plating. Cell colonies remained constant for 7 days.

Chemicals ~-[~H]Glutamine,~-['~C]proline,~ - [ U - ' ~ C ] g l u cose, and methyl a-~-[U-'~C]glucopyranoside, were purchased from The Radiochemical Centre, Amersham. Carbonyl cyanide m-chlorophenylhydrazone was the generous gift of H.U. Schairer.

Transpprt Assays For transport measurements, exponentially growing cultures (5 x 10' cells/ml) were used. For the experiments involving arsenate, the cells were washed three times and suspended in Tris/maleate (80 mM in maleate) pH 7.0/0.4 mM MgS04. Cells subjected to osmotic shock treatment and unshocked controls were taken from the same culture with identical cell counts. Uptake was initiated by the addition of the substrates at the following concentrations : 10 pM ~-[~H]glutamine, 20 pM ~-['~C]proline, 10 nM - 1mM [U-'4C]glucose, 1 - 100 pM methyl a-[U-'4C]glucopyranoside. Aliquots of the cell suspension (0.5 ml) were removed and filtered immediately over Millipore filters (0.45 pm pore size). The filters were then washed twice with 5.0 ml 20 mM phosphate buffer, pH 7.0, and immediately removed and dried. The radioactivity was counted in toluene scintillation liquid. Uptake rates are expressed in nmol/mg cells, the latter determined from turbidity measurements and a calibration curve. Other Methods Methods for osmotic shock of exponentially growing cells and measurements of cell ATP levels have been described previously [4,18,19].

RESULTS Eflect of Uncouplers on Glucose Transport In Escherichia coli, the energized state of the membrane controls the phosphotransferase system [20]. De-energization of the membrane, for instance, leads to a stimulation of uptake of methyl a-glucoside, whereas energization represses its uptake [9-131. Such a stimulatory effect on the uptake of methyl a-glucoside in E. coli B,-' cells following the addition of the uncoupling agents dinitrophenol or carbonyl cyanide m-chlorophenylhydrazone is shown in Fig. 1A. This response also occurs in three other E. coli wild-type cells (not shown) and is similar to that observed following colicin l a treatment or T1 infection [21,22]. In parallel experiments using glucose as a potential phosphotransferase sugar, no stimulation was observed, but an inhibition of uptake following the addition of the uncouplers was seen in E. coli B,-' cells (Fig. 1 B) and also in E. coli LA 3600 (not shown). Rapid metabolism of the labelled glucose might be thought to account for this unexpected results, but the use of [U-14C]glucose and very brief pulse experiments rule out this possibility (not shown). Glucose accumulation was measured over a wide

E. F. Wagner, J. D. Fabricant, and M . Schweiger

0

10

5

0 Time (rnin)

5

233

10

Fig. 2. EfTect of' curbonyl cyunide m-chlorophenylhydruzoiie on vurious trunsport sysrems ir: an E. coli ATPase-less (unc) strain. Cells were grown and transport was measured as described in Fig. 1 and Methods. ~ - [ U - ' ~ C ] G ~ u c o (248 s e Ci/mol, 100 pM) (A), methyl a-~-[U-'~C]glucosyranoside (79 Ci/mol, 100 pM) (B), L-[~H]glutamine (21 Ci/mmol, 10 pM) (C) or ~-[U-'~C]proline (10 pCi/ 0.1 15 mg, 20 pM) (D) were allowed to accumulate for 5 min before the uncoupling agent ( 0 )carbonyl cyanide m-chlorophenylhydrazone (final concentration 20 pM) was added; (0) control

; -; 15

B

. z m E

E

10

a,

Ejfect of Arsenate on Glucose Transport

v)

2

-

concentration range (from 10nM to 1 mM) which covers the concentrations where the methyl a-glucoside uptake was determined. Glucose uptake was inhibited at all concentrations. An active transport system for glucose, other than the phosphotransferase system, could be responsible for the observed inhibition. If so, then an energy donor other than phosphoenolpyruvate, such as ATP or a proton-motive force, would be needed. Uptake dependent on the phosphotransferase system cannot easily be determined directly in vivo (e. g. by measuring the dependence on phosphoenolpyruvate). However, transports driven by ATP or a proton-motive force can be measured [23]. The unc mutation provides a method for distinguishing between these two latter transports 171. This mutation abolishes the normal functions of the proton-translocating ATPase. Since the interconversion of protonmotive force and ATP does not occur in the unc strain, a proton-motive force can be generated without the production of ATP. Thus, the uptake of glutamine (as an ATP-driven transport) and proline (as a proton-motive-force-driven transport) as well as methyl a-glucoside (as a phosphotransferase substrate) were tested using the unc strain. These substrates were compared with glucose uptake following the addition of the uncoupling agent carbonyl cyanide rn-chlorophenylhydrazone (Fig. 2). As can be seen, the proton-motive-force-driven transport is inhibited immediately after the addition of the uncoupler while the phosphotransferase-dependent uptake of methyl a-glucoside is stimulated (Fig. 1A, 2). In contrast, both glutamine and glucose show only a slight inhibition of uptake. These data suggest that glucose may, in fact, be transported by an ATP-driven system.

5

(3

0

.

0 0

5

Fig. 3. EJject of ursenate on various transport systems in E. coli ATPuse-less (unc) cells. The experiment was done as described in Fig. 2 except that arsenate (final concentration 5 mM) was added 4.5 min after substrate accumulation. (A) glucose, (B) methyl a-glucoside, (C) glutamine and (D) proline uptake. (0)Control; ( 0 )arsenate-treated cells

A more direct approach for studying the ATPdependent nature of glucose uptake was the use of arsenate as a selective inhibitory agent for the ATPdriven system [23,24]. The results from these experiments are shown in Fig.3. As expected, neither a proton-motive-force-driven system nor the phosphotransferase system for methyl a-glucoside were affected. However, the ATP-driven glutamine uptake was inhibited with arsenate. Similarly, an inhibitory effect was observed for glucose uptake. Therefore, this indicates that glucose transport is neither driven by a proton-motive force nor mediated by the phosphotransferase system. ATP (or a Related Energy-Rich Phosphate) Is Required for Glucose Transport If glucose is transported via an ATP-driven system, then cellular ATP concentration should be related to

Glucose Transport in E.scho.ichicr d i

234 B

A

30

-c ._ E

. 0)

c c

-$ 20

I

V

s

F:

0 1c

C

I I? .c,a-_-- -I - - - -.t

______--: / - -I I 5

0

10

5

1

10

Time (min)

Fig.4. Eifk't #fsuccinutr #ti thr uptakr ofg~ucoseund methyl r-gluc~~.sit/e in .strirvct/ E. cob B,- ltirid ATPusc~-les.s(unc] crlls. Starvation of Cells and uptake measurements were performed as described. Uptake was initiated by the addition of (o,.) ~i-[U-'~C]glucose (248 Cilmol, 0.5 pM) or ( A , A ) methyl r - ~ - [ U - ~ ~ C ] g ~ u c o p y r d n o(79 s i dCijmol, e 0.5 pM).The radioactivities are expressed per 0.5 mi culture (2 x lo8 cells). Trdnsport in the presence of succinate (A,., final concentration 20 m M ) was determined 8 min after succinate addition. (A) Experiments using B,-I cells; (B) uptake measured in uncAB cells

Table 1. A T P concentration.s in E. coli B, unci in an ATPa.sc.-less (unc) strnin during starvation Cells were prepared for starvation as described in Methods. Aliquots of 0.25 ml were withdrawn after various periods during starvation. Samples were treated with 0.25 mi ice-cold 0.5 M glycine, pH 7.4; 10 p1 chloroform was added and the samples were boiled for 10 min. The ATP was determined in 20.~1aliquots as described [4] Starvation time

ATP in cells of -~~

B< I

days

n m o1/ m g ~

0 2 4 6 7 7 (8 min after the addition of succinate, 20 mM)

~. .. .

imc,AB

~~~

2.1 0.7 0.2 0.2 0.1

5.6 1.1 1.o 1.4 1.8

0.4

1.9

glucose uptake. Thus, experiments were performed using starved cells in which ATP is limited. Under these conditions, the addition of an exogenous substrate leads to an increase in ATP concentration which should cause an increase in glucose uptake. Table 1 shows the intracellular ATP concentrations of a starved wild type and an unc strain. The wild type exhibits a decreased ATP level following starvation (e.g. 0.2 nmol ATP/mg cells 4 days after starvation compared to 2.1 nmol/mg for unstarved cells). The unc mutant contained a higher concentration of ATP (5.6 nmol ATP/mg cells). However, the percentage

decrease of ATP levels following starvation was similar to that in the wild type (e.g. 1.0 nmol ATP/mg cells 4 days after starvation). The addition of a substrate, such as succinate, to starved cells provides a method to measure the ATP dependence of transports directly [23]. Succinate oxidation by the respiratory chain leads to the formation of ATP exclusively via the intact ATPase. In starved wild-type cells, an increased production of ATP was observed following the addition of succinate (Table 3 ) . As expected, the unc mutants showed no increase in ATP concentrations after succinate treatment. Therefore, starved cells were used to study transport of glucose in the presence and absence of succinate. Fig.4 shows glucose and methyl r-glucoside uptake in wild type and unc mutants under increased and constant ATP concentrations. As can be seen, there is no difference in uptake of methyl x-glucoside with or without succinate. Glucose uptake was independent of added succinate in the unc cells and was in the same range as that seen in the wild-type cells without the added succinate. Glucose uptake in the wild type, however, was stimulated twofold by addition of succinate. The lack of increased glucose uptake in the unc mutants when compared to the wild-type cells indicates that glucose transport is directly supported by the level of ATP. Glucose Uptake Is Sensitive to Osmotic Shock

The uptake of glucose following osmotic shock in wild-type and MeGal-deficient cells was measured.

E. F. Wagner, J. D. Fabricant, and M . Schweiger

Glucose uptake was inhibited by 50-60%, in those cells receiving the shock (not shown). As expected, the control system involving glutamine uptake (which requires a periplasmic binding protein) was inhibited too. In a further control experiment the phosphotransferase-mediated uptake of methyl a-glucoside (substrate concentration 1 pM) was not reduced in shocked cells, but a stimulation of uptake was observed probably due to a de-energization of the membrane during the osmotic shock procedure (data not shown). Glucose transport, however, was shock-sensitive. In this system, a small fraction of the transport activity appears to be resistent to the osmotic shock. This remaining activity may reflect a phosphotransferasemediated uptake. It may also be possible that a glucose binding periplasmic protein component was not efficiently removed during osmotic shock.

Properties qf' the A TP-Driven Glucose Uptake

The ATP-dependent glucose uptake system seems to be constitutively synthesized. Cells grown on glucose accumulate glucose with a slightly increased rate when compared to glycerol-grown cells (not shown). The substrate saturation kinetics of glucose-grown wildtype cells (B\- showed an apparent K, of 10 pM and a V of 330 nmol . min-' . (mg cell protein)-'. With mutant cells, deficient in the MeGal transport system, a K, of 3 pM and a V of 240 nmol . min-' . (mg cell protein)-' was obtained. These parameters show that glucose is transported in addition to the MeGal system that carries glucose specificity [25] by a novel ATP-driven and probably binding-protein-dependent system.

DISCUSSION This study provides evidence that the majority of glucose is taken up in Escherichiu coli by an ATPdependent transport and not by the phosphotransferase system. The experiments reported here involved the use of uncouplers or arsenate to block energy metabolism in order to demonstrate the ATP-dependent transport for glucose. In these experiments, as well as in others dealing with the starvation of cells, ATP or a closely related high-energy phosphate, was found to support the uptake of glucose directly. The main transport of glucose in E. coli has been regarded previously as being mediated by the welldefined phosphotransferase system [I, 21. This assumption was derived particularly from studies in vitro and from transport experiments with the nonmetabolizable substrate methyl a-glucoside. In the studies in vitro the purified phosphotransferase enzyme system was used to analyse the mechanism of carbo-

235

hydrate phosphorylation. Glucose was reported to be a substrate for these enzymes (for review see [26]). Therefore, it was concluded that glucose is mainly transported by the phosphotransferase system. Other possible active transport systems for glucose were less extensively considered. For the transport studies, the glucose analogue methyl a-glucoside was mainly used as a glucose substitute; it was found to be taken up by theptsG system [27]. Hence, glucose was thought to enter the cell mainly via the same system [1,2]. A transport system for glucose via the phosphotransferase system has been reported in membrane vesicles [28]. There, the otherwise overshadowing ATP-driven glucose system was inactivated because the possible periplasmic binding components are removed under these conditions. The minor phosphotransferasemediated glucose transport could either be due to the ptsG or ptsM function or both [ l l ] . The ATP-driven glucose transport reported here is unlike any of the transport systems described previously [29]. First, it behaves differently from the phosphotransferase-mediated uptake of methyl 3-glucoside (Fig. 1 - 4). Second, the galactose transport system, gulp, which is capable of transporting glucose, is dependent on a proton-motive force and not on ATP [30,31]. Furthermore, the gulp system is repressed under our conditions (growth on glucose and in the absence of an inductor of the gal operon). Third, the transport system for methyl-/3-galactoside (MeGal) [32] can also be ruled out since MeGaldeficient cells (mgf)are not altered in the main glucose uptake. In addition, glucose transport in those cells is also shock-sensitive. Fourth, the kinetic parameters of the measured glucose transport are inconsistent with the MeGal, gulp, and ptsM systems. Our data on glucose transport is confirmed in various E. coli strains (in wild-type E. coli B and K12; in uric' and mgl cells) which show that in all these strains a highaffinity system ( K , 3 - 10 pM) with a high capacity ( V 240-330 nmol . min-' . mg cell protein-') is responsible for the observed ATP-dependent glucose uptake. It should be recognized that any other energyrich phosphate, which is related to ATP, could be the direct energy souce of the 'ATP-driven' transports. Very recently, acetyl phosphate was reported to serve as an energy-donor for shock-sensitive transports [33]. The experiments described here do not try to define the ultimate energy source but classify 'ATP-driven' transports, which depend on energy-rich phosphate (but not phosphoenolpyruvate), related to ATP. We cordially thank W. Boos for extensive and stimulating discussions and J . Lengeler for critical comments on the manuscript. P. J. F. Henderson and M. C. Jones-Mortimer are thanked for their valuable suggestions. We are also very grateful to H. Ponta for helpful comments. This work was supported by the Fond.y zur Fri'rderung der wissenschaftlichen Forschung, Austria, project 3 194.

236

E. F. Wagner, J. D. Fabricant, and M. Schweiger: Glucose Transport in E.scherichia coli

REFERENCES 1. Roseman, S. (1977) in Biochemistry of Membrune Transport (Semenza, G. Sr Carafoli, E., eds) pp. 582-597, SpringerVerlag, Heidelberg. 2. Saier, M. H., Jr (1977) Bacteriol. Rev. 41, 856-871. 3. Harold, F. M. (1977) Curr. Top. Bioenerg. 6, 83- 149. 4. Klein, W. L. Sr Boyer, P. D. (1972) J . Biol. Chem. 247, 72577265. 5. Scholes, P. Sr Mitchell, P. (1970) J . Bioenerg. 1, 309-323. 6. West, J. C. Sr Mitchell, P. (1974) FEBS Lett. 40, 1-4. 7. Berger, E. A. (1973) Proc. Natl Acad. Sci. U.S.A. 70, 15141518. 8. Schairer, H. U. Sr Haddock, B. A. (1972) Biochem. Biophys. Res. Commun. 48, 544 - 55 1. 9. del Campo, F. F., Hernandez-Asensio, M. Sr Ramirez, J. M. (1975) Biochem. Biophys. Res. Commun. 63, 2366-2370. 10. Hoffee, P., Englesberg, E. Sr Lamy, F. (1964) Biochim. Biophys. Acta, 79, 337- 350. 1 1 . Saier, M. H., J r Sr Moczylowski, E. G. (1978) in Bacterial Transport (Rosen, B. P., ed.) pp. 103-125, Marcel Dekker, New York and Basel. 12. Singh, A. P. Sr Bragg, P. D. (1976) FEBS Lett. 64, 169-172. 13. Wagner, E. F. (1978) Ph. D. Thesis, Technical University, Graz. 14. Hill, R. F. (1958) Biochem. Biophys. Acta, 30, 636-637. 15. Wood, W. B. (1966) J . Mol. Biol. 16, 118-133. 16 Schairer, H. U., Friedl, P., Schmid, B. J. Sr Vogel, G. (1976) Eur. J . Biochem. 66, 257 - 268. 17 Silhavy, T. J. Sr Boos, W. (1974) J . Bacteriol. 120, 424-432. 18 Boos, W. (1974) Curr. Tap. Membrunes Trans. 5, 51 - 136.

19. Heppel, L. A. (1971) in Structure and Function of Biologicul Membranes (Rothfield, L. J., ed.) pp. 224-247, Academic Press, New York. 20. Reider, E., Wagner, E. F. Sr Schweiger, M. (1979) Proc. Natl. Acud. Sci. U.S.A. (in the press). 21. Gilchrist, M. J. R . Sr Konisky, J. (1975) J . Biol. Chem. 250, 2457 - 2462. 22. Schweiger, M., Wagner, E. F., Hirsch-Kauffmann, M., Ponta, H. Sr Herrlich, P. (1978) FEBS Symp. 43, 171 - 186. 23. Rhoads, D. B. Sr Epstein, W. (1977) J . B i d . Chem. 252, 13941401. 24. Ferenci,T., Boos, W., Schwartz, M . Sr Szmelcman, S. (1977)Eur. J . Biochem. 75, 187-193. 25. Lengeler, J., Hermann, K. O., Unsold, H . J. Sr Boos, W. (1971) Eur. J . Biochem. 19, 457-470. 26. Hays, J. B. (1978) in Bacterial Transport (Rosen, B. P., ed.) pp. 43- 102, Marcel Dekker, New York and Basel. 27. Kornberg, H. Sr Smith, J. (1972) FEBSLett. 20, 270-272. 28. Kabdck, H. R. (1968) J . B i d . Chem. 243, 371 1-3724. 29. Silhavy, T. J., Ferenci, T. Sr Boos, W. (1978) in Bacteriul Trunsport (Rosen, B. P., ed.) pp. 127-169, Marcel Dekker, New York and Basel. 30. Henderson, P. J. F., Giddens, R. A. Sr Jones-Mortimer, M. C. (1977) Biochem. J . 162, 309-320. 31. Rotman, B., Ganesan, A. K. Sr Guzman, R. (1968) J . Mol. Biol. 36, 247-260. 32. Rotman, B. Sr Radojkovic, J. (1964)J. Biol. Cllem. 239, 31533156. 33. Hong, J. S., Hunt, A. G., Masters, P. S. Sr Lieberman, M . A. (1979) Proc. Natl Acad. Sci. U.S.A. 76, 1213-1217.

E. F. Wagner and M . Schweiger*, Institut fur Biochemie, Naturwissenschaftliche Fakultat der Leopold-Franzens-Universit5t Innsbruck, Innrain 52a, A-6020 Innsbruck, Austria J. D. Fabricant, Life Science Division, Techn. Inc., National Aeronautm and Space Administration, Houston, Texas, U.S.A. 77058 ~-

* T o whom correspondence should be addressed.