Sugar Transport by the Bacterial ... - Semantic Scholar

Vol. 262, No. 33, Issue of November 25, pp. 16261-16266,1987 Printed in U.S. A.

OF BIOLOGICAL CHEMISTRY THEJOURNAL C 1987 by The American Soeiety for Biochemistry and Molecular Biology, Inc.

Sugar Transportby the Bacterial PhosphotransferaseSystem RECONSTITUTION OF INDUCER EXCLUSION

IN SALMONELLA TYPHIMURIUM MEMBRANE VESICLES* (Received for publication, May 28, 1987)

Thomas P. MiskoS, Wilfrid J. Mitchell$, Norman D. Meadow, and SaulRoseman From the Department of Biology and the McCollum-Pratt Institute, TheJohns Hopkim Universiw,Baltimore, Maryland 21218

The accompanying articles (Saffen,D. W., Presper, meases. Uptake of proline, on the other hand, was K. A., Doering, T. L., and Roseman, S. (1987)J. Biol. unaffected. The resultsare therefore consistent with an hypoth16241-16253; Mitchell, W. J., Saffen, D. Chem. 262, W., and Roseman, S. (1987)J. Biol. Chem. 262, esis that dephosphorylated III?,'& is an inhibitor of 16254-16260)show that"inducer exclusion" in intact certain non-PTSpermeases. cells of Escherichia coliis regulated byIIIG'",a protein encoded bythe crr gene of the phosphoenolpyruvate:glycose phosphotransferasesystem(PTS).The The accompanying articles (1, 2) outline earlier (3-22) and present studies attempt to show a direct effect of IIIG'" on non-PTS transportsystems. present evidence for the role of the phosphoenolpyruInner membrane vesicles prepared from a wild type vate:glycose phosphotransferase system (PTS)'.' in the phestrain of Salmonella typhimurium(pts'), carrying the nomenon designated "diauxie" or the "glucoseeffect." In E. coli lactose operon on an episome, showed respira- Escherichia coli, the PTS controlscertain non-PTS sugar tion-dependent accumulation of methyl-fl-D-thiogalac- catabolic operons via two known mechanisms by regulating topyranoside (TMG) via the lactose permease. In the the appropriate specific transporter (such as the lactose perpresence of methyl-a-D-glucopyranosideor other PTS mease) that takesup the inducing molecule and by regulating sugars, TMG uptake was reducedby an amount which adenylate cyclase. Both processes are governed by the product was dependent on the relative concentrations of IIIG1" of the crr gene, and theaccompanying papers (1,2) show that and lactose permease in the vesicles. The endogenous III"', a PTS protein (23), is encoded by this gene and that 1IIG1'concentration in these vesicles was in the range 5-10 PM, similar to that found in whole cells. Methyl- cells that overproduce IIIG'"behave as predicted by an hypoa-glucoside had no effect on lactose permease activity thetical model (15). The hypothesis proposes that IIIGIcbinds in vesicles prepared from a deletionmutant strain to and inhibits the non-PTS permeases, whereas phospholacking the soluble PTS proteins Enzyme I, HPr, and IIIG1"does not or is a positive effector of the permeases. While in vitro evidence has been presented (19-21) that IIIG1".One or more of the pure proteins could be inserted into the mutant vesicles; when one of the two supports this model, several key questions have not yet been electrophoretically distinguishable forms of the phos- addressed. For example, homogeneous IIIG'"can be isolated in phocarrier protein, IIIz,'&, was inserted, both the ini- two electrophoretically distinguishable forms, III$:w and tial rate and steady state level of TMG accumulation I1Igkt (23); the fast form of the protein lacks the N-terminal were reducedby upto 40%. The second electrophoretic heptapeptide found in the slow form. The relative activities form, IIIFkt,had much less effect.A direct relationship of the two proteins in regulating non-PTS transportsystems, was observed between the intravesicular concentrasuch as the lactose permease, is not known. Furthermore, i n of inhibition of the lactose vitro experiments have not demonstrated whether IIIG1'acts tion of III'&'&,and the extent permease. No inhibition was observed when II@& was on transport systems other than the lactose permease. We added tothe outside of the vesicles, indicating that the have therefore investigated thesequestionsin the present site of interaction with the lactose permease is acces- studies using membrane vesicles isolated from Salmonella sible only from the inner face of the membrane. typhimurium along with homogeneous PTS proteins which In addition to the lactose permease,IIIg,'& was found are inserted into theintravesicular space, or are added to the to inhibit both the galactose and the melibiose per- outside. A preliminary reporthas been presented (24). * This work was supported by Grant CA21901 from the National Institutes of Health. This is contribution 1378 from the McCollumPratt Institute and Paper XXXII in the series, Sugar Transport by the Bacterial Phosphotransferase System. The preceding paper in this series is by Mitchell, W. J., Saffen, D., Roseman, S. (1987) J. Biol. Chem. 262, 16254-16260. 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. 2 Supported by Training Grant GM-57 from the National Institutes of Health. Present address: Dept. of Neurobiology, Stanford Medical School, Stanford University, Stanford, CA 94305. This work was done under the tenure of a postdoctoral fellowship from the Cystic Fibrosis Foundation. Receipt of a research travel grant from the Wellcome Trust is also acknowledged. Present address: Dept. of Brewing and Biological Sciences, Heriot-Watt University, Chambers Street, Edinburgh EH1 1HX United Kingdom.

The abbreviations used are: PTS, phosphoeno1pyruvate:glycose phosphotransferase system; PEP, phosphoenolpyruvate; HPr, histidine-containing phosphocarrier protein of the phosphotransferase system; 1IIG1',the phosphocarrier protein of the phosphotransferase system, specific for glucose and methyl-a-glucoside, which functions as part of the 11-Glc complex; IPTG, isopropyl P-D-thiogalactopyranoside; TMG, methyl-P-D-thiogalactopyranoside. * The phosphoeno1pyruvate:glycose phosphotransferase system consists of a number of protein components which function sequentially in the transfer of a phosphoryl group from PEP tosugar. The phosphoryl group is transferred via the general soluble proteins Enzyme I and HPr toone of the sugar-specific proteins, for example to II-AMa"(an integral membrane protein) or IIIG1'(a soluble and/or peripheral membrane protein). The integral membrane proteins IIBM"" and II-BG1'catalyze the transfer of the phosphoryl group from II-AM""and 1IIG1',respectively, to the sugar concomitant with the translocation of the sugar across the membrane.

16261

16262

Reconstitution of Inducer Exclusion

in S. typhimurium Membrane

EXPERIMENTAL PROCEDURES

Materials-[14C]Methyl-~-~-thiogalactopyranoside (54.7 Ci/ mmol), 2-deoxy-~-[l-'~C]glucose (54.6 Ci/mmol), and L-[U-"C]proline (230 mCi/mmol) were purchased from DuPont-New England Nuclear. Ascorbic acid, bovine serum albumin, CAMP,deoxyribonuclease I (crude, from bovine pancreas), dithioerythritol, phenazinemethosulfate, phosphoenol-pyruvate (potassium salt), D-(-)-3-phosphoglyceric acid (grade 11, disodium salt), L-prolone, sodium succinate, L-tryptophan, and all sugars were from Sigma. All other commercial chemicals were of the highest available purity. Bacterial Strains and Growth of Cells-S. typhimurium strains SB2950 (trpB223trzAptsHIcrrA49), SB3675 (trpB223pts+F'lacOc15), and SB3891 (trpB223trzAptsHIcrrA49F'tacOcl5)have been described (2, 18, 25). Cells were routinely grown in minimal salts Medium 63M supplemented with 1%(w/v) Difco casamino acids and L-tryptophan (20 pglml). For induction of appropriate transport activities, mediawere further supplemented with either 0.2% Dgalactose, 0.5% melibiose, and 5 mM CAMP,or 0.5% sodium succinate, plus 0.1% lactose and 0.5 mM IPTG. Preparation of Inner Membrane Vesicles-Vesicles were prepared from strain SB2950 as described (26). For strains SB3675 and SB3891, the procedure was modified as follows. The starter culture was prepared in 500 ml of Medium 63M containing 0.1% lactose, and approximately 150 ml of this culture (absorbance = 0.1) was used to inoculate each 2-liter batch of growth medium. Also,during treatment with penicillin G the temperature was reduced from 37 to 21 "C; this effectively increased lactose permease activity in the resultant vesicles. In all vesicle preparations used in this report, both 5-P buffer (5 mM potassium phosphate, pH 6.6, containing 10 mM magnesium sulfate) in which spheroplasts were lysed, and 50-P buffer (50 mM potassium phosphate, pH 6.6, containing 10 mM magnesium sulfate) in which vesicles werewashed and stored, were supplemented with 1 mM dithioerythritol, which stabilized lactose permease activity for several days. Proteins were inserted into the vesicles during spheroplast lysis in 5-P buffer (26). Transport Assays-Assays were conducted in glass beakers (10-ml capacity) which were first treated by filling with a 2% solution of bovine serum albumin, incubated for 30 min at 37 "C, rinsed once with distilled water, and allowed to dry in air; coating the beaker surface with protein in this way helped maintain vesicle integrity during experiments. An assay volume of 1 mlwas routinely used. Vesicles (approximately 1 mg of protein) in 50-P buffer, were incubated for 3 min at 25 "C in a gyratory water bath with shaking at 250 rpm. An energy source was then added in the form of ascorbic acid (titrated to pH 6.6 with KOH) and phenazine methosulfate to give final concentrations of 20 and 0.1 mM, respectively. This was followed immediately by addition of radioactively labeled substrate. Samples (0.1 ml) were removed at the indicated times, diluted into 10 mlof ice-cold 0.1 M LiCl, and rapidly filtered through glass-fiber discs (Whatman CF/F,2.5 cm). In assays of the melibiose transport system, the LiCl solution was replaced as diluent by 0.1 M potassium phosphate buffer, pH 6.6. Filters were dried under a heat lamp, placed in vials with 3.5 ml of premixed scintillation fluor (3a70B, RPI Corp.) and radioactivity determined by liquid scintillation counting. Purification and Assay of PTS Proteins-Enzyme I, HPr, III'&, and III$kt were purified to homogeneity as described previously (23, 27, 28); the phosphorylated forms of both species of IIIc'c have been isolated, and their relationship is given below (see "Results"). Intravesicular IIIc" concentrations were measured by rocket immunoelectrophoresis (23,26) after disruptionof the vesicles by treatment with 30% (v/v) ethanol.The proportion of IIIg/&and II& in preparations of the protein was determined by crossed rocket immunoelectrophoresis as described (23). Protein concentration in vesicle suspensions was estimated by a microbiuret method (29), with bovine serum albumin as standard. RESULTS

Inducer Exclusion in Vesicles Isolated from Wild Type and Mutant Strains-In a previous report (la),we showed that a wild type (pts') strain of S. typhimurium SB3675, carrying a constitutive lac episome, exhibited reduced levels of TMG accumulation via the lactose permease when exposed to millimolar concentrationsof the nonmetabolizable glucose analog methyl a-glucoside. Thus, thelactose permease, anE. coli transport protein, not only functioned normally in S. typhi-

Vesicles

murium but was susceptible to inhibitionby the S. typhimurium PTS proteins in the hostorganism. Inner membrane vesicles prepared from this strain behave similarly to the whole cells and show respiration-dependent uptake of TMG via the lactose permease. The vesicles also contain residual amounts of endogenous, soluble PTS proteins (26); therefore,whenthe cellshave been inducedfor the phosphoglycerate transportsystem, P E P drives uptake of glucose and methyl-a-glucoside in the resultingvesicle preparation. The vesicles exhibit inducer exclusion as illustrated inFig. 1A. Over a wide range of methyl-a-glucoside concentrations, from 10 ~ L to M 5 mM, essentially the same degree of inhibition of TMG accumulation was observed. This sensitivity to low concentrations of methyl-a-glucoside is similar to that found cells in cells of a ptsl mutant (12). On the other hand, intact of the wild type (pts') strain from which these vesicles were prepared were insensitive t o low concentrations of the PTS sugar: 50 p~ methyl-a-glucoside had no effect on TMG uptake (18).Increased sensitivity of the vesicles t o methyl-a-glucoside, compared to the whole cells, thus reflects the loss of PTS proteins, particularly of Enzyme I, which is present at low levels in the vesicle preparations (26). In common with whole cells, addition of methyl-a-glucoside to the vesicle suspension after accumulation of TMG had reached a steady state level, resulted in a net efflux of TMG until a new, reduced level of accumulation was reached. This reduced accumulation was identical to that obtained when methyl-a-glucoside was present throughout the experiment (Fig. 1B).Mannose, glucose, and N-acetylglucosamine, all PTS sugars, also promoted inducer exclusion to about the same extentin wild type (SB3675) vesicles, whereas L-glucose and L-fucose did not (data not shown). Vesicles prepared from S. typhimurium SB3891, a strain

0

0

2

4

6

0

2 4 6 TIME ( m i d

FIG. 1. Effect of methyl-a-glucoside on TMG uptake via the lactose permease in membrane vesicles. Vesicles were prepared from cells grown in the standard medium (see "Experimental Procedures") supplemented with 0.5% sodium succinate, 0.1% lactose, and 0.5 mM IPTG. TMG uptake was assayed as described under "Experimental Procedures" at 25 "C. Incubation mixtures contained in final volumes of1.0ml: 50 mM potassium phosphate, pH 6.6, 10 mM magnesium sulfate, 1 mM dithioerythritol, 0.7-0.9 mg of vesicle protein, 20 mM ascorbate, 0.1 mM phenazine methosulfate, and 1 mM [I4C]TMG (0.77 mCi/mmol). Vesicles prepared from wild type cells, SB3675, are shown in A and B; vesicles from SI33891 (ptsHIcrr) in C. Other conditions are: A, 0, 0 duplicate incubation mixtures with no additions; the area within the hatched lines, 0.01, 0.1, or 5 mM methyl-a-glucoside; A, incubation mixture without ascorbate and phenazine methosulfate. B, 0,. duplicate incubations, no additions; 0, 0.1 mM methyl-a-glucoside; A, 0.1 mM methyl-a-glucoside added at 3.5 min. C , 0, no additions; 0, 5 mM methyl-a-glucoside; A, 5 mM lactose; A, no ascorbate and phenazine methosulfate. The range of variation between different vesicle preparations is shown by the control (top) curves in panels A and B.

Reconstitution of I n d u c e r Exclusion in S. typhimurium Membrane

Vesicles

16263

lacking HPr, Enzyme I, and IIIG1"because of a deletion mu- mobility and activity in in vitro PTS-sugar phosphorylation tation extending through ptsH, the p t s f , and crr genes, showed assays. IIIz&. is considerably more active and is the predoma reduced level of TMG accumulation which was insensitive inant form in the cell. II1$kt is derivedfrom IIIgi& by the vesicles proteolytic cleavage of the amino-terminal heptapeptide, a to 5 mM methyl-a-glucoside (Fig. 1C).Thus,the isolated from the mutant did not show the phenomenon of processcatalyzed by a specific membrane-boundprotease inducer exclusion and in this respect behaved similarly to the (30). III$zt accepts the high energy phosphatefrom phosphoHPr butdoes not act asa phosphoryl donor to the sugar (23). intact cells from which they were derived. When IIIg& was incorporated into SB3891 vesicles both In whole cells, we have shown that the efficiency of excluthe initial rate and the final level of TMG accumulationwere sion of TMG in the pts+ strain SB3675 is a function of the reduced (Fig. 3 A ) . III$kt was much less effective in inhibition relative cellular concentration of PTS proteins and lactose of TMG uptake (Fig. 3B). The concentration of I1I$kt incorpermease (18),and that strainsof E. coli containing elevated porated into the vesicles in the experiment shown in Fig. 3B levels of IIIG1' were foundtobe more sensitivetoPTSwas sufficient to give nearly maximal inhibition (35-40%) if mediated effects (2). In the present studies, these parameters III$zt was as active as IIIg;&, but the observed inhibition was were examined in membrane vesicles. By growing cells on only about 8%. In other experiments, similar results were different concentrationsof lactose in the presence or absenceobtained, I1I$kt was 20-30% as effective aninhibitoras of IPTG (2, 18), vesicles were obtained which contained IIIg/:w. The presence of up to 15% III$kt in a preparation of different amounts of lactose permease. In the vesicle prepa- IIIg,'& did not reduce the inhibitory effect of I I I g w(Fig. 4). rations, the concentration of IIIG1' was found to be in the Proteins such as soybean trypsin inhibitor, similar in mo, to that in whole cells. As shown in lecular weight to IIIG1",and bovine serum albumin, had no range 5-10 p ~ similar Fig. 2, the extent of inhibition of TMG accumulation in the effect on TMG transport whether present in the intravesicular presence of methyl-a-glucoside was clearly dependent on the or extravesicular space (data not shown). ratio of IIIG1'to lactose permease in whole cells as reported Both IIIgi& and III$kt were completelyineffectivewhen before, (Fig. 2 A ) and in vesicles from pts+ cells (Fig. 2B). In added externally to thevesicles. The site of action of IIIG1'is neither case was completeinhibition of TMG uptake ob- thus located at the inner face of the vesicular membrane, an served. important observation with respect to inthevitroexperiments Reconstitution of Inducer Exclusionin MembraneVesicles(see "Discussion"). We have demonstrated,by restoring PTS function to a deleAs shownin Fig. 4, the degree of inhibition of TMG tion mutant strain, that functionalsoluble PTS proteins can accumulation in the reconstituted vesicle system was dependof IIIg,'& and thelactose be introduced into the intravesicular space during preparation ent on both the intravesicular content of membrane vesicles from S. typhimurium (26). This tech- permease,activity. This result is thus inexcellent agreement nique therefore serves as the basis of an in vitro system for with that found in whole cells and vesicles from wild type identifying the protein(s) involved in inducer exclusion, uti- cells (Fig.2, A and B). In the reconstituted preparations there ~ ~III$&. w lizing vesicles prepared from strain SB3891, which carry a was no detectable conversion of I I I ~ to The data therefore demonstrate that, in the complete abfunctional lactose permeasebutare completelydevoid of 111''. Inparticular,itis possible t o sence of the soluble PTS proteins Enzyme I and HPr, the EnzymeI,HPr,and examine the role of IIIC1"in the absence of the other soluble lactose permease is inhibited by III?i:w, and to a lesser extent by III$kt. PTS proteins. Attempts toReverseInducer Exclusion-In the original When IIIG1'is purified from whole cells two forms termed IIIg;& and III$zt are found (23)which differ inelectrophoretic model for inducer exclusion, the inhibitor of the non-PTS

;i.

f

Y

1

1

1

1

0 2 4 6 8 1 0 2 4 6 0 1 0 ~ Q ~ ~ ~ / ~ - G A L A C T O S I O A S EIIIQ'c/LACTOSEPERMEASE (pglunit activity) x IO (gg/unit activity). 10'

FIG. 2. Inducer exclusion in vesicles from pts+ cells: effect of the relative levels of IIIG"and lactose permease. SI33675 cells were induced to different extents for the lactose catabolic and PTS proteins by growth on different concentrations of lactose in the presence or absence of0.5 mM IPTG (18). A , data obtained with whole cells, taken from Fig. 6 of Ref. (18),is presented for comparison. B , vesicleswere isolated from these cells, and the level of lactose permease was determined by uptake of [''CC]TMG as shown in Fig. 1, and IIIG1'(pg/mg of vesicle protein) by rocket immunoelectrophoresis. One unit of lactose permease was arbitrarily assigned the value 10 nmol of TMG transported/min/mg of vesicle protein. Inducer exclusion is expressed as the percent decrease in the steady state level of TMG accumulated by the vesicles in the presence of 5 mM methyl aglucoside.

TIME ( m i d

FIG. 3. Reconstitution of inducer exclusion by III"% comparison of I@$!& and III$kb.Vesicles were prepared from cells of SB3891 @tsHlcrr) grown in the presence of 0.5% sodium succinate, 0.1% lactose, and 0.5 mM IPTG. Uptake of TMG was assayed as described in the legend to Fig. 1.A , effect of IIIg/& on TMG uptake. Spheroplasts were lysed in the presence (0)or absence (0)of 6.7 p~ III&. The intravesicular concentration of the protein was 0.41 pg/ mgof vesicle protein (approximately 4 p M ) . Extravesicular 6.7 p~ III~/&(X) had no detectable effect on TMG uptake. Incubation mixtures contained 0.89 mg of vesicle protein/ml. B , effect of II@& onTMGuptake.Spheroplasts werelysed in the presence (m) or absence (0)of IIIgkt. The intravesicular concentration of was 0.74 pg/mg of vesicle protein (approximately 7.2 p ~ ) Incubation . mixtures contained 1.0 mg of vesicle protein/ml. It should be noted that thevesicles in panel B contained 7.2 p~ III$zt while the vesicles in panel A contained 4p M IIIg/&.

16264

Reconstitution of Inducer Exclusion in S. typhimurium Membrane Vesicles

PERMEASE (pg/unit activity) X 10

mGlC/ LACTOSE

FIG. 4. Reconstitution of inducer exclusion is dependent upon the ratio IIC%'&/lactose permease. Several preparations of membrane vesicles were isolated from SB3891 (ptsHZcrr) by lysis of spheroplasts in the presence of IIIG1'.The vesicle content of IIIG'" was varied by using different concentrationsof pure IIIgk (0)and (0)in the medium; in three experiments (A),the protein preparation consisted of 85% III%w, and15% III;zt, and the results are calculated in terms of the former. A single batchof cells was usedto prepare sixof the membranevesicle preparations (0).The III'& content of the vesicles was measured by rocket immunoelectrophoresis and lactose permease as described in Fig.1. One unit of lactose permease activity was arbitrarily assignedthe value 10 nmol of TMG transported/min/ mg of vesicle protein. Inducer exclusion is definedas the decrease in the steady state level of TMGwhen the vesicles contain 111% (expressed as percent). permeases is IIIG'",while phospho-IIIG1' is either inactive ora positive effector of the permeases. The data presentedabove clearly showthat I1IGicis an inhibitorof the lactose permease, and attempts were therefore made to reverse this effect, by phosphorylating the proteininside the vesicles. A number of experiments were conductedwith vesicles obtained from the wild type cells and from the deletion mutant.Inboth cases, the cellswere first grown inthe presence of phosphoglycerate in order toinduce the transport system for PEP. The resultsof these experimentswere inconclusive. Vesicles obtained from the wild type cells contained sufficient endogenous Enzyme I and HPt so that they could be useddirectly withvariouscombinations of methyl aglucoside and/or PEP. Onlya partial reversal of inducer exclusion was obtained by even 5 mM PEP. Even here, the results were not easily interpretable since PEP alone shows aninhibitory effect on TMG uptake(competition for energy?). The results obtained withvesicles from the deletion mutation intowhich were inserted thehomogeneous proteins IIIG1", Enzyme I, and HPr in various combinations, were even less conclusive. These results are further considered in the "Discussion."

Effects of I I P on the Galactose and Melibiose PerrneasesIn addition to lactose, metabolism of glycerol, maltose, and melibiose by E. coli and S. typhirnuriurn has been shown to be subject to PTS-mediated repression (14). Also, inducer exclusion has beensuggested to control expression of the own galactose operon of E. coli (31, 32),althoughinour studies (2, 14) we were unable to demonstrate any effect of the PTS on galactose fermentation by intact cells. This apparent discrepancyvery likely results from the fact that inducer exclusion is mediated by a stoichiometric interaction between the PTS and the permeases, and therefore, a t fully induced levels of the permease, little or noeffect of the PTS

may be observed in whole cells. Indeed, it is well known that glucose does not inhibit uptake of solutes via the lactose permease in cells fully induced for expression of the lactose operon (7, 18). Vesicles prepared from appropriately induced cells of the pts deletion strain S. typhirnuriurn SB2950 actively accumulate substrates of the galactose and melibiose permeases, and we therefore examined the effect of intravesicular IIIGIcon the activity of these transport systems. The absence of phosphotransferase activity and lactose permease in this strain permits use of 8-deoxyglucose and TMG as substratesof the galactose and melibiose systems, respectively. As shown in Fig. 5 , uptake of both analogs was decreased in the presence of intravesicular IIIg;:w. By contrast, extravesicular IIIG1"had no effect, as was found in the case of the lactose permease. Finally, accumulation of proline was unaffected by the presence of III&:w on either sideof the membrane(Fig. 5C). Thus, pure intravesicularIIIg/& specifically inhibits transport systems which are targets for inducer exclusion. DISCUSSION

As indicated in the Introduction to this paper, we suggested a model (4, 15) for the mechanism of inducer exclusion in enteric bacteria in which the phosphocarrier protein of the glucose-specific phosphotransferase system, designatedHIG1', acts as an inhibitor of certain non-PTS transport systems. a The model was based inpartontheobservationthat functional crr gene is requiredfor inducer exclusion. This gene has now been shown to be the structural gene for IIIG1" (1, 16,17). In addition, E. coli cells containing elevatedlevels of IIIG'' show increased sensitivity to inducer exclusion and more effective control of the lactose transport system when 2 5 L " -

I

I

8

TIME (minf

FIG. 5. Effect of III$& on the galactose and melibiose permeases in membrane vesicles: reconstitution of inducer excluthe deletion mutant SB2950 sion. Vesicleswerepreparedfrom (ptsHZcrr) grown in the presence of 0.2% galactose (A and C) or 0.5% melibiose plus 5 mM cyclic AMP ( B ) . Uptake of labeled solutes was measured as described in the text and in the legend to Fig. 1. A, uptake of 2-deoxyglucose viathe galactose permease. Wherepresent, the concentration of intravesicular III:& (0)was 0.8 rg/mg of vesicle protein (approximately 7.8 p ~ ) .Incubation mixtures contained 1.0 (0.73 mCi/ mgof vesicle protein and 0.1 mM 2-deoxy-~-[~~C]glucose mmol). One control contained24 p~ extravesicular III'&,(X), while the other (0)consisted of untreated vesicles. The vesicles accumulated approximately0.7 nmol of 2-deoxyglucose/mg of protein in the absence of ascorbate and phenazine methosulfate. B, uptake of TMG via the melibiose permease. Vesicles were prepared from whole cells as in A, ie. in the presence or absence of III%,. Where present inthe intravesicular space, the concentration of the protein was about 8.4 p~ (0).The same symbols are used as in A. Incubation mixtures contained 0.85mg of vesicle protein, 20 mM Licl, and 1 mM ["c] TMG (0.95 mCi/mmol). The LiCl was added immediately beforethe TMG.In the absence of ascorbateandphenazinemethosulfate, vesicles accumulated approximately2-3 nmol of TMG/mg of protein. C, uptake of proline. The vesicle preparations were aliquots of those (3.2 mCi/mmol)was used at a concentraused in A. ~-[U-'~C]proline tion of 10 PM. In the absence of ascorbate and phenazine methosulfate, the vesicles accumulated approximately 0.3 nmol proline/mgof protein.

Reconstitution of Inducer Exclusion in S. typhimurium Membrane the latter isinduced to high levels (2). Preliminary reports have appearedfrom this (24) and other laboratories (19-21) which have attempted to prove or disprove the hypothetical model by in vitro experiments; the details of our work are the subject of this paper. Osumi and Saier (19) reported that 111"'" preparations, undefined with respect to the electrophoretically slow and fast forms of the protein, bind to E. coli membranes containing higher than normal levels of the lactose permease. Nonspecific binding occurred, but binding significantly increasedwhen substrates of the lactose permease were present. In another setof experiments (20), binding of 111"'" t o proteoliposomes containing the purified lactose carrier was also measured, but there was little bindingabove the high background level of IIIG1"binding to the lipid itself. However, the presence of IIIG'' did affect the binding of one of the substrates of the lactose carrier to (201, the the reconstituted proteoliposomes. In this report remaining data are concerned with the carrier function of the lactose permease. IIIG'' showed a slight inhibitory effect on the carrier in reconstituted proteoliposomes and in counterflow experiments with membranevesicles isolated from whole cells. In these experiments, the 111"'" was apparently only in the extravesicular space. Finally ( Z l ) , partially purified III"" was shocked into plasma membrane vesicles, and significantly inhibited the uptakeof lactose.3 No effect was observed when the 111"'' was located in the extravesicularspace. In the present studies,we have used homogeneous proteins and separatelyassayed IIIgi:w and 111% for their activity. The vesicle preparations have been characterized as highly purified right-side out inner membrane vesicles (26), isolated from two S. typhimurium strains carrying the E. coli lactose operon on anepisome. One strain waspts+, and vesicles the contained endogenous IIIG'c, Enzyme I, and HPr (26), while the other strain wasa deletion mutation lacking the pts genes and therefore contained no Enzyme I, HPr, orIIIG1".The vesicles behaved as predicted from the behavior of the whole cells. That is, the vesicles isolated from the pts+strain were capable of taking up TMG, and this was inhibited by methyl-aglucoside and other PTS sugars. On the other hand, no such inhibition was observed in vesicles prepared from the deletion mutant lacking thethree PTS proteins.When IIIG'" was inserted into the mutant vesicles, the uptake and accumulation of TMG was reduced. As in the case of the whole cells, the efficiency of inhibition of TMG uptake depended on the ratio of intravesicular IIIGi' concentration to the lactose permease in the membranes, suggesting a direct interaction between the two proteins. Finally, IIIgi& was three to five times more active than III$2t in inhibiting the lactose permease. The proposed model for inducer exclusion predicts that the inhibitory effect of IIIG" should be overcome by phosphorylation of the protein. Indeed, the reportsfrom other laboratories cited above suggest that this may be the case. Binding of 111"'' to E. coli membranes and proteoliposomes containing the lactose permease was substantially decreased by the presence of Enzyme I, HPr, and PEP (19, 20), while in E. coli There were, surprisingly, "minimal effects" (21) on methyl aglucoside uptake by shocking PEP and the PTS proteins into the vesicles. The differential effects on lactose and methyl-a-glucoside uptake led the authors to suggest that soluble IIIG1' is important in regulation of the lactose permease but not in PTS-mediated methyla-glucoside transport. In our studies with vesicles from the S. typhimurium deletion mutant, SB2950 (26), uptake of the glucoside was restored to close to wild type levels when the soluble PTS proteins were inserted intravesicularly. The differential effect (lactose versus methyl-a-glucoside uptake) reported with the E . coli vesicles (21) therefore differs from the results reportedhere and remains to be explained.

Vesicles

16265

vesicles, inhibition of lactose uptake by IIIG1"was reversed by a high concentration of P E P (21). Our own studies have not led to a clear resolution of this issue, and in our view, more work is required before we understand how the other PTS proteins affect the interaction between IIIG1",phospho-III"'", and the non-PTSpermeases. The interactions between 1IIG" and adenylate cyclase (22, 33) are also affected in a complex manner in the presenceof Enzyme I and HPr. Most studiesof inducer exclusion havebeen concerned with regulation of the lactose permease. The reconstitution experiments shown in Fig. 5 provide the first direct demonstration of the ability of IIIG1"to interact with and inhibit other nonPTS transport systems, i.e. the melibiose and galactose permeases. It seems likely that a similar interaction occurs between I I P " and the transport system for maltose. Inhibition of glycerol uptake, on the other hand, is apparently the result of interaction between 111"" and the enzyme glycerol kinase (34,35) rather than with the glycerol transporter, although a direct effect on theglycerol transporter has not been excluded. All of the available data are therefore in accord with at least one part of the original model, namely that IIIG1' (i.e. IIIg&,) bindstoandinhibitsatleast some non-PTS permeases. At the same time, this protein must bind to a t least phospho-HPr, one of the soluble proteins of the PTS, to glycerol kinase, and phospho-IIIG1' must bind to the membrane protein 11-B"'" and possibly to adenylatecyclase as well (22). The binding to PTS the protein 11-B"" appears torequire at least the N-terminal heptapeptide of IIIg;:w (23), and itwill be of great interest to determine whether this site is also requiredfor the binding to the non-PTS permeases since IIIg2t is less inhibitory than IIIgi,&. In addition, the crr gene has beencompletely sequenced (1,36),and a series of mutants and partial deletions of 111"'' have been isolated which show no activity in intact cells in inducer exclusion. These molecular formsof 111"'" will be used in studies of the type indicated above. Unless the binding sitesof I1Pc to the different membrane proteins are identical, one might expect to find mutants defective in both inducer exclusion and methyl-a-glucoside phosphorylation and two additional classes of mutants, each one defective in only one of these two functions. Finally, it will also be of great interest to determine whether the various PTS andnon-PTS, which bind membraneproteins,both either IIIG'' and/or phospho-1IIG1', have homologous regions or sequences which are responsible for these binding phenomena. REFERENCES 1. Saffen, D. W., Presper, K. A,, Doering, T. L., and Roseman, S. (1987) J. Biol. Chem. 262, 16241-16253 2. Mitchell, W. J., Saffen, D. W., and Roseman, S. (1987) J. Biol. Chem. 262, 16254-16260 3. Meadow, N. D., Kukuruzinska, M. A., and Roseman, S. (1984) in Enzymes of Biological Membranes (Martonosi, A., ed) Vol. 3, 2nd ed, pp. 523-559, Plenum Press, New York 4. Postma, P. W., and Roseman, S. (1976) Biochim. Biophys. Acta 457,213-257 5. Rosen, B. P., and Kashket, E. R. (1978) in Bacterial Transport (Rosen, B. P., ed) pp. 559-620, Marcel Dekker, Inc., New York 6. Dills, S. S., Apperson, A., Schmidt, M. R., and Saier, M. H., Jr. (1980) Microbiol. Rev. 4 4 , 385-418 7. Magasanik, B. (1970) in The Lactose Operon (Beckwith, J. R., and Zipser, D., eds) pp. 189-219, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 8. Makman, R. S., and Sutherland, E. W. (1965) J. Biol. Chem. 240, 1309-1314 9. Peterkofsky, A., and Gazdar, C. (1974) Proc. Natl. Acad. Sci. U. S.A. 71, 2324-2328 10. Boniface, J., and Koch, A. L. (1967) Biochim. Biophys. Acta 1 3 5 , 756-770

16266

Reconstitution of Inducer Exclusion

i n S. typhimurium Membrane

11. Winkler, H. H., and Wilson, T. H. (1967) Biochim. Biophys. Acta 135,1030-1051 12. Saier, M. H., Jr., Simoni, R. D., and Roseman, S. (1976) J. Biol. Chem. 251,6584-6597 13. Saier, M.H., Jr., and Roseman, S. (1976) J. Biol. Chem. 251, 6598-6605 14. Saier, M.H., Jr., and Roseman, S. (1976) J. Biol. Chem. 251, 6606-6615 15. Roseman, S. (1977) Fed. Eur. Biochem. SOC.Symp. (Berlin) 4 2 , 582-597 16. Meadow, N. D., Saffen, D. W., Dottin, R. P., and Roseman, S. (1982) Proc. Natl. Acad. Sci. U. S. A. 79, 2528-2532 17. Meadow, N. D., Rosenberg, J. M., Pinkert, H. M., and Roseman, S. (1982) J. Biol. Chem. 257, 14538-14542 18. Mitchell, W. J., Misko, T. P., and Roseman, S. (1982) J. Biol. Chem. 2 5 7 , 14553-14564 19. Osumi, T., and Saier, M. H., Jr. (1982) Proc. Natl. Acad. Sci. U. S. A. 7 9 , 1457-1461 20. Nelson, S. O., Wright, J. K., and Postma, P. W. (1983) EMBO J. 2 , 715-720 21. Dills, S. S., Schmidt, M. R., and Saier, M. H., Jr. (1982) J . Cell. Biochem. 18, 239-244 22. Liberman, E., Saffen, D., Roseman, S., and Peterkofsky, A. (1986) Biochem. Biophys. Res. Commun. 141,1138-1144

Vesicles

23. Meadow, N. D., and Roseman, S. (1982) J. Biol. Chem. 257, 14526-14537 24. Misko, T. P., Mitchell, W. J., Meadow, N. D., and Roseman, S. (1982) Fed. Proc. 41, 1416 25. Cordaro, J. C., and Roseman, S. (1972) J. Bacteriot. 112, 17-29 26. Beneski, D. A., Misko, T. P., and Roseman, S. (1982) J. Biol. Chem. 2 5 7 , 14565-14575 27. Weigel, N., Waygood, E. B., Kukuruzinska, M. A., Nakazawa, A., and Roseman, S. (1982) J. Biol. Chem. 257, 14461-14469 28. Beneski, D.A., Nakazawa, A,, Weigel, N., Hartman, P. E., and Roseman, S. (1982) J. Biol. Chem. 257, 14492-14498 29. Zamenhof, S. (1957) Methods Enzymol. 3,696-704 30. Meadow, N. D., Coyle, P., Komoryia, A., Anfinsen, C. B., and Roseman, S. (1986) J. Biol. Chem. 261, 13504-13509 31. Adhya, S., and Echols, H. (1966) J. Bacteriol. 92,601-608 32. Joseph,J.,Danchin, A., and Ullmann, A. (1981) J . Bucteriol. 146,149-154 33. Reddy, P., Meadow, N., Roseman, S., and Peterkofsky, A. (1985) Proc. Natl. Acad. Sci. U. S. A . 82, 8300-8304 34. Postma, P. W., Epstein, W., Schuitema, A. R. J., and Nelson, S. 0. (1984) J. Bacteriol. 1 5 8 , 351-353 35. Novotny, M. J., Frederickson, W. L., Waygood, E. B., and Saier, M. H.,Jr. (1985) J. Bacteriol. 162,810-816 36. Nelson, 0.S., Schuitema, A. R. J., Benne, R., van der Ploeg, L. H. T., Plijter, J . S., Aan, F., and Postma, P. W. (1984) EMBO J. 3, 1587-1593