Sugar Transport by the Bacterial Phosphotransferase System

THEJOURNALOF BIOLOGICAL CHEMISTRY

Vol. 262, No. 33, Issue of November 25, pp. 16254-16260,1987 Printed in U.S. A.

(0 1987 by The American Society for Biochemistry and Molecular Biology, Inc.

Sugar Transport by the Bacterial Phosphotransferase System IN VIVO REGULATION OF LACTOSE TRANSPORT INESCHERICHIA COLI BY IIIG", A PROTEIN OF THE PHOSPHOEN0LPYRUVATE:GLYCOSE PHOSPHOTRANSFERASE SYSTEM* (Received for publication, May 28, 1987)

Wilfrid J. Mitchell$, David W. Saffene, and SaulRoseman From the McCollum-Pratt Institute and The Department of Biology, The Johns Hopkins University, Baltimore, Maryland 21218

Escherichia coli and Salmonella typhimuriumpref- sistent with the model where inducer exclusion is aferentially utilize sugar substrates of the phosphoenol- fected by a direct interactionbetween IIIG'"and a nonpyruvate:glycose phosphotransferase system (PTS) PTS transportsystem. when the growth medium also contains other sugars. This phenomenon, diauxic growth, is regulatedby the crr gene, which encodes the PTS proteinIIIG'"(Saffen, D. W., Presper, K. A., Doering, T. L., and Roseman, S. The phosphoeno1pyruvate:glycose phosphotransferase sys(1987) J. Biol. Chem. 16241-16253). We have protem (PTS)' is widely distributed in obligate and facultative posed that non-PTS permeases are regulated by their interaction with IIIG1", and in vitro studies from other anaerobic bacteria (2, 3) and is used by these cells to acculaboratories haveprovided support for thismodel, but mulate sugars. The systemconsists of a series of protein the in vivoeffects of excess IIIG1" are not known. In the components that together actas a phosphoryl transfer chain present studies, transformed cells that overproduced between the donor, phosphoenolpyruvate, and the sugar substrate which is phosphorylated as it crosses the cell memIIIG'"2- and 10-fold, respectively, were constructed from a pts+ strain of E. coli and plasmids containing brane. A schematic summaryof the PTS is shown in Fig. 1 of (1).Purification and characterization the crr gene. In the2-fold overproducer, fermentation the accompanying paper of, and growthon the non-PTS carbohydrates glycerol, of Enzyme I, HPr, and IIIGICfrom Salmonella typhimurium lactose, maltose, and melibiose wasgenerally more have been reported (4-6), as well as the specificities of the sensitive to the glucose analogue methyl-a-D-glucopyr- glucose Enzyme I1 complexes (7). anoside than in a control strain containing normal Extensive genetic and physiological studies have revealed levels of IIIG1".In addition, inhibition of lactose per- that, in addition to its primary function in sugar transport, mease activity by methyl-a-glucoside (inducer exclu- the PTS has an important role in theregulation of metabolism sion) was more effective in the 2-fold overproducer of non-PTS sugars. That is, in the presence of a PTS sugar, than in the control strain, particularlywhen the per- induction of synthesis of enzymes required for utilization of mease activity was high. The 10-fold IIIG'Coverpro- certain non-PTS sugars is prevented. This phenomenon has ducing strain had a requirement for the amino acids been known for many years as the glucose effect or diauxie methionine, isoleucine, leucine, and valinethat may or (8). There are two known facets to the control mechanism. may not be related to the increased concentration of First, the PTS has been shown to regulate adenylate cyclase IIIG1".Fermentation of non-PTS carbohydrates was also poor in the latter strain. Finally, lactose permease activity and therefore intracellular CAMP concentration(9activity was50%of that in control cells containing the 11); second, the entry of non-PTS sugars, the inducers, into same levels of @-galactosidase,and the lactose per- the cell is inhibited (12-14). This latter control mechanism is mease activity in the IIIC1" overproducer was reduced known as inducer exclusion and has been the subject of our to an extremely low level in thepresence of methyl a- own studies. Our earlierfindings may be summarized asfollows (14-16): glucoside. Thus there is an inverse relationship between the cellular concentration of IIIG'"and the ability ( a ) mutants lacking either Enzyme I or HPr (or both) are to metabolize non-PTS substrates. The results are con- hypersensitive to PTS-mediated regulation. ( b ) A second-site mutation called crr (carbohydrate repression resistant) is associated witha deficiency in IIIG1"and renders these strains * This work was supported by Grant CA 21901 from the National regulation. (c) A PTS Institutes of Health. This is Contribution 1376 from the McCollum- completely insensitive to PTS-mediated Pratt Institute, and is paper XXXI in the series Sugar Transport by sugar is effective in inducer exclusion only in the presenceof the Bacterial Phosphotransferase System. The preceding paper in this its EnzymeI1 complex. For example, mutants defective in the series is Ref. 1. The costs of publication of this article were defrayed specific glucose membrane receptor, II-BC'', are insensitive to in part by the payment of page charges. This article must therefore methyl-a-glucoside. ( d ) Inducer exclusion occurs in cells that be hereby marked "aduertisement" in accordance with 18 U.S.C. are fully induced for the susceptible non-PTS system. Section 1734 solely to indicate this fact. A model was presented in whichthe dephosphorylated form $This workwas performed under the tenure of a Postdoctoral Fellowship from the Cystic Fibrosis Foundation. Receipt of a Re- of the glucose-specific phosphocarrier protein IIIG" acted as search Travel Grant from the Wellcome Trust is also acknowledged. Present address: Dept. of Brewing and Biological Sciences, HeriotWatt University, Chambers Street, Edinburgh EH11HX United Kingdom. 3 Supported by Training Grant GM-57 from the National Institutes of Health. Present address, The Dept. of Molecular Biology and Genetics, The Johns Hopkins University Medical School, Baltimore, MD 21205.

The abbreviations used are: PTS, phosphoeno1pyruvate:glycose phosphotransferase system; HPr, histidine-containing phosphocarrier protein of the phosphotransferase system; 1IIG", the phosphocarrier protein of the phosphotransferase system, specific for glucose and methyl a-glucoside, which functions as part of the 11"" complex; IPTG, isopropyl P-D-thiogalactopyranoside;ONPG, o-nitrophenyl PD-galactopyranoside;TMG, methyl P-D-thiogalactopytanoside.

16254

Regulation of Lactose Transport by IIP' an inhibitor of certain non-PTS transport systems (2, 17,181. (The adenylate cyclase results a r e less clear (19, 20), but t h e phosphorylated form of IIIcl' may be an activator of adenylate cyclase.) The presence of a PTS sugar would lead t o a net dephosphorylation of I1IG", resulting in inhibition of inducer uptake. Enzyme I and HPr mutants, which have a reduced capacity for rephosphorylatingIIIGIC,would be extremely senhas beenconsiderably sitivetotheseeffects.Themodel strengthened by the demonstration that the crr gene is, in fact, the structural gene for IIIG1"(1, 21, 22). In vitro experiments have been reported which support the model. These results are summarized below (see "Discussion"). It has recently been shown (23, 24) that the efficiency of exclusion of inducersinwholecells is dependent on the cellular content of the non-PTS transport protein, or perthe sensitivity of the E. coli mease. In our own work (23), lactose permease to inducer exclusion was clearly reducedin cells with higher levels of permease activity.This is consistent with the suggestion that the inducer exclusion mechanism involves some type of stoichiometric interaction between the permease and a PTS component present in the cell in limiting amounts. The effect of independently increasingthe concentration of the PTS proteins could not be studied, however, sincegrowthconditionsleadingto such an increasealso resulted in lowerinduction of the lac operon.Now, with success in cloningthe crr gene (1,21), it has become possible to construct strains of E. coli that overproduce 1IIG". This report describes the properties of two such strains. A direct relationship was found between PTS-mediated inhibition of lactose permease activity and cellular I I P " concentration.

16255 TABLE I

Cellular content of IIIGk Cells were grownin the indicated media, harvested, washed, resuspended in Medium 63M, and thequantity of 111"' in each suspension was assayed by rocket immunoelectrophoresis as described under "Experimental Procedures." Results aregiven f standard deviations; the number of measurements are shown in parentheses. W"content Growth condition DS340

DS344

DS336" p g / m g cell protein

LB medium 1.85 A 0.42 (4) 3.41 f 0.55 (5) 19.7 f 2.9 (3) L-Arabinose in 1.62 f 0.06 (5) 3.42 f 0.34 (8) N.D.6 63M L-Arabinose 1.31 f 0.20 (5) 2.75 f 0.33 (4) 18.9 f 0.9 (4) casamino acids' in 63M The following strains harbor the indicated plasmids (see Fig. 1): DS344, pBR322; DS340, pDS45; DS336, pDS35. N.D., not determined. e The same cell extracts were also assayed for the PTS proteins Enzyme I and HPrby rocket immunoelectrophoresis. Antibodies were raised against the pure proteins isolated from S. typhimurium, and these proteins were used as standards in the measurements with the E. coli extracts. While both IIIG1'and HPrfrom the two species crossreact completely with a given antibody preparation, this is not true of Enzyme I, and we therefore present the relative concentrations of the latter in the three strains. The results were as follows: HPr (pg/ mg cell protein): DS344,1.73 f 0.12 (5); DS340,1.78 f 0.04 (4); DS336,1.56 f 0.21 (4). Enzyme I (relative quantities/mgcellprotein): DS344, 100 f 5.6 (5); DS340, 108 f 8.4 (4); DS336, 88 f 5.1 (4).

+

volume of the same medium and grown to mid-exponential phase in a New Brunswick gyratory shaker at 180 rpm. The temperature throughout the growth period was 37 "C. Cells were harvested and EXPERIMENTALPROCEDURES washed with Medium 63M (23) and resuspended in an appropriate Materials-Lactose and glycerol were purchased from J. T. Baker volume of Medium 63M for each experiment. In experiments where Chemical Co. Other sugars used for growth of bacteria and in fermen- the extent of lac operon induction was varied, chloramphenicol was and all galactosides were added to cultures, before harvesting, to a concentration of 50 pg/ml, tation tests, methyl-a-D-glucopyranoside, was synthesized from and was present the wash and assay buffers a t the same concentration. from Sigma. Methyl-cY-D-[14C]glucopyranoside D-[U-14C]glucose(Du Pont-New England Nuclear) by Dr. Catherine Cells were stored on ice for a maximum of 4-5 h before use. IIPk Assay-The cellular concentration of IIIGICwas assayed by was Shea in this laboratory. [14C]Methyl-~-~-thiogalactopyranoside from Du Pont-New England Nuclear. All other chemicals were of the the technique of rocket immunoelectrophoresis as described previously (6, 21, 23). Aliquots (50 p l ) of cell suspensions were treated in highest purity commercially available. Plasmids, Bacterial Strains, and Cell Growth-Plasmids were pre- a bath sonicator, for 3 X 15 s in the presence of 1.25% Triton N-101, pared, cell strains constructed, and grown as described in the accom- and 5 p1 of the resulting suspension was applied to theimmunoplate. panying manuscript (1). Plasmid pDS45, a derivative of pBR322, Where necessary, suspensions were diluted with electrophoresis carries a 1.5-kilobase fragment of the E. coli genome, while pDS35 buffer containing 0.1 mg/ml bovine serum albumin before application carries a 7.05-kilobase fragment (Fig. 1).Transformants harboring to theplate. Electrophoresis was conducted for 16-18 h at 90 V. 111'" pDS45 are tetracycline resistant, ampicillin sensitive, while pDS35 was estimated by comparison with known concentrations of the pure confers the opposite drug phenotypic behavior. Both plasmids carry IIIG1'protein from S. typhimurium (6, 21, 22). Pure E. coli 111'" gives the crr but not the other genes of the pts operon (encoding Enzyme I the same rocket heights as the protein from S. typhimurium using and HPr).A P-1 transducing lysate was prepared on E. coli JC10240 antibodies raised against the latter? Assay of Lactose Permease and 8-Galactosidase-TMG counterflow, (HfrP045srlC300::TnlOrecA56) (1,21,26) andwas used to transduce E. coli K-12 strain 1100. Transductants (resistant to tetracycline) and the rate of ONPG hydrolysis in while cells (lactose permease) were recA-. TnlO was removed by the method of Bochner et al. (27) and toluene-treated cells (&galactosidase), were assayed as described as modified by Maloy and Nunn (28). after which the presence of the (23). Spectrophotometric measurements were obtained with a PerkinrecA- mutation was again tested. The resulting strain was trans- Elmer model 557 Spectrophotometer. Uptake of Methyl-a-glucoside-Accumulation of methyl-a-glucoformed as described (1,21) with one of the plasmids pDS35, pDS45, or pBR322, to give rise to strainsDS336, DS340, and DS344, respec- side was followed by means of standard procedures for measuring tively. Transformants were isolated on MacConkey agar indicator solute uptake. Details of the assay and treatment of samples have plates, without lactose (29),containing 2% (w/v) maltose, 10 mM been described (7, 23). Cell Protein-Protein concentration of cell suspensions was measmethyl a-glucoside, and either 30 pg/ml ampicillin or 10 pg/ml tetracycline as appropriate. Theywere stored a t -20 "C in LB medium ured by a modification of the Lowry procedure (31). (30) containing 20% (v/v) glycerol. LB medium contains the following (g/liter): Bactotryptone, 10; yeast extract, 5; NaCl, 5; distilled water. RESULTS Cultures were grown in LB medium or in defined media consisting of Medium 63M supplemented with a carbon source a t a concentration IIPk Concentration-Theresults of assays of the I I P of 0.5% (w/v) and 5 pg/ml thiamine hydrochloride (23); where pres- content of strains DS336, DS340, and DS344 are presented ent, Difco casamino acids were added to a concentration of 0.1%. All i n Table I. The two strains transformed with plasmids carmedia contained either 15 pg/ml ampicillin or 10 pg/ml tetracycline as appropriate. In general, cells were inoculated intoa test tube rying the crr gene producedmore IIIG1'than the control, but containing 5ml of LB medium and grown to early exponential phase, DS336 contained considerably more of the protein than did is clear. T h e pDS35 the culture being aerated by rotation. A small aliquot was then diluted DS340. The reason for this difference not into anothertube containing 10 ml of the appropriate growth medium. After overnight growth, this culture was used to inoculate a larger * Dr. N. Meadow, unpublished results.

Regulation of Lactose Transport by I I P

16256

plasmid carries a significantly larger piece of the bacterial chromosome thanpDS45 (Fig. 1) and may thuscontain additional sites of regulation of transcription of the gene (1). Alternatively, the degree of overproduction of IIIGlCmay be dependent on the site of insertion of the bacterial segment into theplasmid, i.e. expression of the crr gene in pDS35 may be under the control of a plasmid-derived promoter. When the same plasmids were used to transform mutant strains which were ptsl- or ptsl-crr-, the same pattern of overproduction of IIIG1'was observed (data not shown).Also, neither of the two plasmids carries the genes for Enzyme I and HPr (Table I), and the transformed and control strains were found to contain similar levels of these proteins by immunoassay. Therefore, the IIIG1'overproducing strains constitute a convenient experimental system inwhich to specifically examine the role of the IIIG" protein in inducerexclusion. Fermentation of Sugars-We examinedfermentation of Hindlll

3.24

1

pDS35 11.4 Kb

\\ 2.9

8omHI HpoI

0.65

€COR1 EcoRI

""I

TABLE I1 Fermentation of PTS and non-PTS sugars

pDS45 5.1 Kb

-

both PTS sugars and non-PTS sugars by the plasmid-bearing strains, and the results are presented in Table11. Fermentation of PTS sugars was independent of the cellular content of IIIG1".In the case of non-PTS sugars, however, significant differences were found between the three strains. The control strain DS344 showedstrong fermentationof all sugars tested, and this fermentationwas largely insensitive to the presence of methyl-a-glucoside. The strain DS340 was also a strong fermenter, although in two cases (glycerol and melibiose), it was considerably weaker when methyl-a-glucoside was present. DS336, on the other hand,was considerably impaired in fermentation even in the absenceof the PTS sugars. It thus appears that thereis, at least qualitatively, aninverse correlation between the ability to ferment these non-PTS sugars and the amount of IIIG1' in the cell. Interestingly, Table I1 also illustrates the apparent difference in sensitivityof each non-PTS catabolic system to the PTS, since 10-fold overproduction of IIIcl' substantially reduced glycerol and melibiose fermentation while having littleeffect on utilizationof lactose. Fermentation of galactose was completely unaffected under the conditionsof our experiment; we have previouslyreported (14) that galactose utilization isrelatively insensitive to PTSmediated repression, although the galactose permease can be partially inhibited by IIIG1'in membrane vesicle preparations (35,43). Growth in Minimal Medium-Strain DS336 did not grow well in minimal media; it showed a doubling time on glucose, for example, of approximately 4 h. Growth on glucose was accelerated by inclusion in themedium of a low concentration of casamino acids, and the lattercould be replaced by methionine, isoleucine, leucine, and valine, each at a concentration of 25 pg/ml. By growing DS336 on glucose in the presenceof these amino acids, an inoculum for growth studies could be prepared, and the rate of growth on several carbon sources was examined and compared DS340 to and DS344 (Table 111). The lattertwo strains behaved similarlyon the carbon sources tested, while DS336 grew more slowly, except onglucose. The slower growth rate of DS336was particularly marked on

E. coli D N A

ADNA

- pBR322

Cultures were prepared in LB medium, and used to inoculate MacConkey agar indicator plates (29) containing the test sugar at a concentration of 2% (w/v), and either 30 pg/ml ampicillin (DS336) or 10 pg/ml tetracycline (DS340 and DS344). Plates were incubated for 20 h at 37 'C. The symbols used for degree of fermentation are: ++, very strong fermentation, dark red colonies; +, strong fermentation, red colonies with region of weaker fermentation at perimeter; +, weak fermentation, pale red colonies; -, no fermentation, white colonies. Methyl-

FIG. 1. Restriction maps of plasmids pDS36 and pDS45. A, the plasmid pDS35 carries the IIIGLstructural gene (err) on a 7.05kb fragment derived from the transducing page XpcysAcrrS2 (1, 21). This fragment is inserted into the BamHI site of pBR322, disrupting the tep gene. The direction of transcription of the crr gene is indicated. It is not known whether initiation of mRNA synthesis occurs at thepromoter for the teP gene or at a promoter for crr, or both. B, the plasmid pDS45 contains a 1.5-kb fragment carrying the crr gene and replaces the segment of pBR322 between the PstI and EcoRI restriction sites which carries part of the gene coding for p-lactamase. The direction of transcription of the crr gene is indicated, although the point of initiation of mRNA synthesis is unknown. This plasmid lacks the primary promoter for the 0-lactamase gene, but initiation of transcription could occur at a secondary promoter for 0-lactamase, P2, which is located near the Hind111 restriction site of pBR322 (25), or a t a promoter for the crr gene (1).The sources of DNA fragments in the plasmids, pBR322, E. coli, and X-DNA, as well as thedirection of transcription of the crr gene, are given in the figure. The numerical values give lengths of DNA in terms of the number of base pairs, expressed as kilobases.

a-glucoside sugar Test

in medium

Degree of fermentation DS344

DS340

DS336

++

++ ++ ++ ++

mM

PTS sugars Glucose Fructose Mannitol Mannose Non-PTS sugars Glycerol Lactose Maltose Melibiose Galactose

0 0 0 0

0 10 ++0 10 ++ 0 10 0

lo

0 10

++ ++ ++ ++ ++ + ++ ++ ++ ++ ++ + ++ ++++

++ ++ ++ + k

++ ++ ++ f ++ ++

f

-

++ + +-

+-

-

++

Regulation ofTransport Lactose TABLE I11

by IIP'

16257

protein on theefficiency of control of transport of a non-PTS sugar by the PTS. Thecurrent working model for PTSmediated regulation assigns to 111"" the role of an inhibitor of the transport systems for lactose, maltose, melibiose, and glycerol (2, 3, 17, 18,23, 24); IIIG1'has been shown to inhibit glycerol kinase (40-42), and it hasbeen suggested that itdoes not affect the glycerol transporter.3 Within the framework of this model, it may be expected that any alteration in the amounts of the interactingspecies will influence the efficiency of inhibition. We have shown thatwhen the lactose system is induced to high levels transport activitybecomes less sensitive to methyl-a-glucoside (23), and similar findings havebeen reported for the maltose transport system (24). However, thus far it has notbeen possible to manipulate thecellular content of PTS proteins independently of the lactose system, since growth of cells for example on glucose or lactose stimulates synthesis of the PTS system and also repressesthe lac operon. This problem is overcome in strainsDS336 and DS340, which Doubling time (h) of inocula grown in appear to beinduced normally in thepresence of IPTG. Cells Carbon source Glucose Lactate in medium of DS336 in Medium 63M containing arabinose and casamino DS344 DS340 SD336 DS344 DS340 acids grew well, and expression of the lac operon could be + + varied by exposure to IPTG for different lengths of time. PTS sugars Under the same conditions DS340 and DS344 grew more 1.06 1.06 1.14 1.16 1.18 1.10 1.14 Glucose rapidly. For eachof the three strains, the relationship between 1.58 1.48 1.80 1.70 1.70 1.70 1.64 Fructose @-galactosidase activity and lactose permease activity, measN.D.6 N.D. 1.62 1.24 1.26 2.76 1.56 Mannitol ured as the rate of hydrolysis of ONPG in toluene-treated and 1.80 1.86 N.D. 1.52 N.D. 1.48 N.D. Mannose intact cells, respectively, (32) is shown in Fig. 2 A . With each Non-PTS sugars strain the two activities were directly proportional to one 1.98 0' 2.02 1.64 3.18 2.06 0' Glycerol another, in agreement with earlier reports (23, 32). However, 1.28 1.34 2.10 1.28 1.56 1.38 1.72d Lactose while the resultsfor DS340 and DS344 are virtually identical, 1.62 1.78 1.70 2.76 1.62 1.56 2.52 Maltose cultures of DS336 show about half the amount of permease 1.78 1.54 3.44 1.64 1.94 1.62 1.92d Melibiose activity for thesame @-galactosidase content (seebelow), a -, methyl-a-glucoside absent; +, methyl-a-glucoside present. indicatingthatthepermeaseactivityin DS336 does not N.D., not determined. function as efficiently as in the other strains. 0, no growth within 20 h. Increased lag period before onset of growth. When methyl-a-glucoside is present in the permease assay, the rate of ONPG hydrolysis is markedly decreased, and it is glycerol and melibiose,which were alsofermented poorly apparent from Fig. 2, A and B that DS340 is more sensitive (Table 11), and on the PTS-sugar mannitol, which surprisingly to the PTS sugar than is the control, DS344. This inhibitory supported a growth rate less than half that of the other strains effect is considerably exacerbated in DS336 where the lactose (see "Discussion"). When theinoculum was grown in medium permeaseactivityis reduced toextremely low levels by containing glucose, even the control strainDS344, was found methyl-a-glucoside. to be very sensitive t o inclusion of methyl-a-glucoside in the DS340 and DS344 were further compared in the experiment medium. Growth of the cells in the presence of the glucoside illustrated in Fig. 3. By growing these strains on arabinose in resulted in considerable lag periods on all substrates tested the absence of casamino acids, the rate of growth was slowed (except glucose itself). Therefore, comparisonof the effects of and a greater extent of induction of the lac operon by IPTG methyl-a-glucoside on growth of the three strains was not was attained. For the parental strain,DS344, as the @-galacattempted with glucose grown inocula. tosidase activity (and consequently lactose permease activity) When inocula were prepared on lactate instead of glucose, increased,the effect of methyl-a-glucosideontherate of the control strainDS344 was found to be virtually insensitive ONPG hydrolysis progressively decreased until there was only to methyl-a-glucoside, exceptfor one carbon source, glycerol. While DS336 did not grow well on lactate, and so could not approximately 30% inhibition at maximum lactose permease be manipulated in the same way, this growth regime allowed induction. Thisdecrease in efficiency of control of the lactose a comparison of the effects of methyl-a-glucoside on DS340 permease when permease levels are increased was also found S. typhimurium (23). The and DS344. As shown in Table 111, there was little difference inourearlierexperimentswith results obtained with DS340 are, however, quite different. In in the rates of growth on the two strains, except in the case this case, 1 mM methyl-a-glucoside gave around 80% inhibiof medium containingmaltoseandmethyl-a-glucoside,in which DS340 grew considerably more slowly. Not shown in tion even at the highest permease activity. Therefore, PTSthe table is the fact that, in the presenceof methyl-a-gluco- mediated regulation of lactose transport activity is considerside, there was a lag period of several hours before growth of ably more efficient in a cell that contains twice the normal DS340 commenced on either lactose or melibiose. It therefore level of 111"". Similar results were obtained when DS340 and appears that DS340, containing twice the level ofIII"'", is The inhibition of glycerol kinase may or may not be sufficient by more sensitive to methyl-a-glucoside than is the control strain itself to account for the marked sensitivity of the glycerol uptake DS344. system to inhibition by the PTS, a point that is discussed pro and Sensitivity of the Lactose Permease to Methyl a-Glucosidecon in the published papers (40-42). For present purposes, therefore, The main objective in constructing strains that harbor inglycerol transport is meant as the combined action of the glycerol creased amounts of IIIG'" was to examine the effect of this transporter and kinase.

Growth rates of strains of E. coli Inocula weregrown at 37 "C in Medium 63M containing either 0.5% glucose or lactate. These cultures were harvested in the exponential growth phase, washed, and concentrated 10-15-fold in Medium 63M, and aliquots transferred to 50 mlof growth medium in 125-ml Erlenmeyer flasks. Growth rates are for cells in Medium 63M with the indicated sugars a t 0.5% concentration. Where present, methyl-a-glucoside was added to a concentration of 5 mM. Media, both for inocula and in growth studies, were supplemented with 5 pg/ ml thiamine-HC1 and either 15 pg/ml ampicillin (DS336 and DS344) or 10 pg/ml tetracycline (DS340); in addition, media used for DS336 contained methionine, isoleucine, leucine, and valine, each at 25 pg/ ml. Flasks were rotated at 180 rpm in a New Brunswick gyratory shaker at 37 "C. Samples were removed at regular intervals and the respective absorbances determined at 500 nm. Growth rates, expressed as doubling time in hours, were derived by plotting the logarithm of the absorbance as a function of time. At the conclusion of each growth experiment, the ability of the cells to ferment glycerol or melibiose was tested in MacConkey agar (see Table 11).

Regulation of Lactose Transport by I I p L c

16258

1000 2000 0 0.2 0.4 0.6 0.8 1.0 8-GALACTOSIDASE ACTIVITY METHYL (I-GLUCOSIDE (mM)

DS344 were grown o n lactate instead of arabinose (data not shown). The activity of the lactose transport system was also measured by means of a counterflow assay in which exchange of labeled and unlabeled substrate was followed (33). Since this assay is carried out in cells treated with azide to prevent accumulation of substrates of the lactose permease, it gives a direct measure of the ability of the permease to translocate its substratesindependent of the energy-coupling mechanism. The results of TMG counterflow assays are shown in Fig. 4. For the particular experiments illustrated, the DS340 and DS344 cultures have identical @-galactosidaseactivities and

(nmol ONPG hydrolyred/min/mg)

1

FIG. 2. Inhibition of lactose permease activity by methyl-aglucoside. Cultures (100 ml) were prepared, as described under “Experimental Procedures,” in defined medium containing 0.5% (w/ v) glucose and 0.1% (w/v) casamino acids. Cells were concentrated 20-fold by harvesting, washing, and resuspending in Medium 63M, and diluted 100-fold into fresh medium containing 0.5% (w/v) Larabinose and 0.1% casamino acids. The h c operon was induced to different extents by addition of 0.5 mM IPTG at different times during growth (23).Cultures were harvested in mid-exponential phase, washed, and resuspended in Medium 63M containing chloramphenicol as described in the test. A, permease activity as a function of @-galactosidaseactivity. To assay for total @-galactosidaseactivity, cuvettes contained 0.01-0.1 mg of cell protein (whole cells permeabilized with toluene) and 1 mM ONPG in 1 ml Medium 63M. For the whole cell (permease) assay, cuvettes contained 0.25-0.30 mg of cell protein and 1 mM ONPG in 1 mlof Medium 63M; this assay was carried out in the presence (closed symbols) and absence (open symbols) of 1 mM methyl-a-glucoside. 0, 0, DS344; 0, U, DS340; A, A, DS336. All hydrolysis rates in whole cells were corrected for the rate in the presence of 5 mM thiodigalactoside, which represents nonspecific entry of ONPG into the cell. B, inhibition of lactose permease activity as a function of methyl-a-glucoside concentration. For the cultures in A which had @-galactosidaseactivities of around 1500 nmol of ONPG hydrolyzed/min/mg cell protein, the rate of ONPG hydrolysis by whole cells (permease activity) was measured in the presence of the indicated concentrations of methyl-a-glucoside. 0, DS344; U, DS340; A, DS336.

320t -200

kz

-

’

’

’

’

DS344

”-

‘

’

’

’

z 0 L

o

k

-

t

I

”

DS 340

2

‘ 0 METHYL a-GLUCOSIDE (mM)

2

4

6

8 1 0

TIME (min)

FIG. 4. Inhibition of TMG counterflow by methyl-a-glucoFIG. 3. Inducer exclusion as a function of lac operon expression. Overnight cultures (10ml) were prepared, as described under side. Cells were prepared as described in the legend to Fig. 2. Cell “Experimental Procedures,” in defined medium containing 0.5% (w/ suspensions (approximately 20 mg of protein/ml) were loaded with v) L-arabinose. Thesecultures were diluted into 500 ml of fresh 20 mM TMG in the presence of 30 mM sodium azide at 30 “C; 50 pl medium, and after one generation of growth this was divided into of the suspension was then diluted into 3.95mlof Medium 63M 100-ml portions. Induction of the lac operon to different extents, containing 30 mM sodium azide and 0.25 mM [“CITMG (2 mCi/ collection of cells and assay of the rateof ONPG hydrolysis by whole mmol). Samples (0.4 ml) were removedat theindicated times, filtered, cells was performed as described in the legend to Fig. 2. @-Galactosid- washed, and intracellular radioactivity determined as described prease activity of the cultures, expressed as nmol of ONPG hydrolyzed/ viously (23). 0, control; 0, 1 mM methyl-a-glucoside presentin min/mg cell protein by toluene-treated cells, were as follows: DS340 diluent. P-Galactosidase activity of the cultures, expressed as nmol of U, 451; 0, 1040; 0,1751; 0,2981.DS344: U, 399; 0, 770; 0, 1392; 0, ONPG hydrolyzed/min/mg cell protein, were as follows: DS344,1324; DS340,1345; DS336,1681. 2821.

Regulation of Lactose Transport by IIP'

16259

111"'" inhibits the non-PTS permeases, are thusdirectly comparable with respect to permease activity. pothesis (2,17,18) that The control experiments show similar profiles of accumula- presumably by combining with them, while phospho-IIIG" is either inactive or stimulates these transport proteins. Recent tion of [14C]TMG, and in both cases this accumulation is sensitive.to methyl-a-glucoside. However, in agreement with in vitro experiments have attempted to test this hypothesis theresults of assays of ONPG hydrolysis(Fig. 2 A ) , the directly. For example, we report in the accompanying paper (43) thatIIIG1'inserted into membranevesicles prepared from parental strain DS344 appears to be less sensitive to the glucoside than is DS340. The DS336 culture contained 25% S. typhimurium partially inhibited thelactose, galactose, and more P-galactosidase activity, but nevertheless did not show melibiose permeases, and similar results have been reported greater permease activity than the other two strains, again for in the lactose permease in E. coli vesicles (36). In addition, agreementwiththefindings for ONPG hydrolysis. Most IIIG1"has been shown to have a small inhibitory effect on striking about the DS336 experiment, however, was the fact lactose permease activity both in natural membranes prepared that counterflow was completely eliminated in the presence froma permease overproducing strain of E. coli and ina of methyl-a-glucoside. Apparently, the small amount of per- reconstituted proteoliposome system (37). Significant binding mease activity remaining in the presenceof methyl-a-gluco- of IIIG1"to membranes and proteoliposomes has also been of ONPG demonstrated (37, 38). The bindingis quantitatively effected side, asshown by a smallbutsignificantrate hydrolysis, is insufficient toallow accumulation of [14C]TMG by substrates of the permease andlikewise IIIG'c, whenpresunder these conditions. Once again, therefore, the activity of ent, affects binding of the galactosides. The in vitro results proposed model, but there is the lactose permease and its sensitivity to methyl-a-glucosideare therefore consistent with the little information on the effects of 111"'' in vivo. This question is shown to be dependent on the cellular concentration of is particularly important since IIIG1' is thought to interact 111"". stoichiometrically, rather than catalytically, with other sysEffect of ZZpk Levels on Methyl-a-glucoside Uptake-The accumulation of methyl-a-glucoside by each of the three E. tems. Furthermore, it is believed to bind to many proteins, including thesoluble protein HPr, a membrane P T S protein, coli strains is shown in Fig. 5. It is quite clear that, despite (e.g. lactose,melibiose, containing elevated levels of III"", DS340 and DS336 do not 11-B"'", non-PTSpermeasesystem maltose), glycerol kinase, adenylate cyclase, etc. have agreater capacityfor uptake of substrates of the glucoseIn one in vivo study (39), PTS-mediated inhibition of specific PTS than does DS344. The increased sensitivity of glycerol uptake by E. coli was partially reversed by substrates the lactose transport system to methyl-a-glucoside is therefore not due to an increased rate of transport of the glucoside of the lactose and melibiose permeases. The authorssuggested into the cell. Consistent with this conclusion, preliminary that the results were explicable by cooperative binding of a experiments indicated that when methyl-a-glucosidewas re- common inhibitor, possibly III"", to thelactose and melibiose placed by 2-deoxyglucose, transport of which occurs via dif- permeases in the presenceof the respective substrates4 This interpretation implies that thein vivo level of 111"'" is limiting ferent sugar-specific components of the PTS (7, 34) and is thus independent of 1IIG", ONPG hydrolysis was again more undertheseconditions,anhypothesisthatcanbetested provided that the level of 111'' can be altered independently sensitivein DS340 and DS336 thanin DS344. Thedata presented above therefore strongly implicate the 111"'" protein in the cell. In earlierwork (23,24),PTS proteins were found to regulate as having a critical and direct role in the phenomenon of non-PTS permeasesprovided that the latter were not induced inducer exclusion. to levels high enough to overcome the inhibition by the PTS. In those studies, however, it was not possible to increase the DISCUSSION level of 111"'' without simultaneously increasing the levels of A singlemutation ina genedesignated crr(14-16) abolishes Enzyme I and HPr and depressing the levels of the non-PTS PTS-mediated repression of the utilizationof three non-PTS permeases. The availability of plasmids containing the crr sugars in S. typhimurium (glycerol, maltose, melibiose) and gene has now made itpossible to construct strains containing also of lactose in E. coli. The crr gene codes for III"'', one of increased levels of IIIG1"as an independentvariable. the glucose-specific proteins of the PTS (1, 21, 22) rather Strain DS336, the 10-foldoverproducer ofIII"", showed than regulating its synthesis. Earlier results led to the hy- markedly different properties than the controlcells (DS344) or the 2-fold overproducer, DS340. That is, DS336 required four amino acids, methionine, isoleucine, leucine, and valine for optimal growth in minimal medium, and, with the exception of glucose, grew much more slowly on all sugars tested, even including the PTS sugars fructose and mannitol. Unlike the plasmid used to constructDS340, which contains only 1.5 kilobases of the E. coli genome, and which produces only a single bacterial protein, III"", in maxi-cell experiments (21), the plasmid pDS35 used to construct the 10-fold overproducer DS336, contains a 7.05-kilobase fragment of E. coli DNA at a

t

-

0

I 8

0 2

4 6 TIME (min)

FIG. 5 . Uptake of methyl-cr-glucoside. Cells were prepared as described in the legend to Fig. 2. Assay mixtures contained 0.2-0.4 mg of cell protein and0.1 mM [14C]methyl-a-glucoside ( 5 mCi/mmol) in total volumes of 2.0 ml of Medium 63M. After incubating at 30 "C for the indicated times, 0.1-ml samples were removed, diluted 100fold with Medium 63M, and immediately filtered, and intracellular radioactivity determined as described previously (7). 0, DS336; DS340; A, DS344.

* The phenomenon, partial reversibility of methyl-a-glucoside inhibition of glycerol uptake by substrates of the lactose and melibiose permeases appears to be restricted to these systems. For instance, glycerol did notrelieve inhibition of maltose or melibiitol uptake, and maltose did not affect inhibition of melibiitol or glycerol uptake. In one sense, the results are unexpected, based on thesimple assumption that the various permeases and PTS proteins compete for a limiting quantity of 111'''. As shown here andelsewhere, the non-PTS system that responds mostsensitively to PTS-mediated inhibition is the glycerol uptake system, followed by melibiose and maltose; the least sensitive is thelactose permease.

.,

16260

Regulation of Lactose Transport by IIPk

Acknowledgments-We are grateful to Richard Bardales and Tadifferent restriction site in pBR322. Thus, we do not know whether the aminoacid requirement of DS336 and itsgrowth mara Doering for skilled technical help, and to Dorothy Regula for behavior is related solely to the high level of 111"" or to the expert editorial assistance. products of the unidentified genes carried in by the plasmid REFERENCES or both. The answer to this important question lies in linking the crr gene alone t o a potent promoter of the type found in 1. Saffen, D.W., Presper, K. A., Doering, T. L., and Roseman, S. (1987) J . Biol. Chem. 262,16241-16253 pDS35, and attempts arenow in progress to construct sucha 2. Postma, P. W., and Roseman, S. (1976) Biochim. Biophys. Acta 4 5 7 , 213257 plasmid. As will be reported elsewhere, strains carrying the N. D., Kukuruzinska, M. A., and Roseman, S. (1984) in Enzymes overproducer plasmid pDS35 arehighly unstable with respect 3. Meadow, of Bzological Membranes (Martonosi, A., ed) Vol. 3, 2nd ed. pp. 523-559, Plenum Press, New York to the synthesis of IIIG1c,and this suggests in fact that cells 4. Weigel, N., Waygood, E. B., Kukuruzinska, M. A,, Nakazawa, A., and cannot normally tolerate high levels of this protein. Roseman, S. (1982) J. Bml. Chem. 257, 14461-14469 5 . Beneski, D. A., Nakazawa, A., Weigel, N., Hartman, P. E., and Roseman, The problems encountered in growing DS336 may well be S. (1982) J. Biol. Chem. 257, 14492-14498 relatedtoPTS-mediatedcontrol over adenylate cyclase, 6 . Meadow, N. D., and Roseman, S. (1982) J. Biol. Chem. 257,14526-14537 7 . Stock, J. B., Waygood, E. B., Meadow, N. D., Postma, P., and Roseman, S. which is also thought tobe effected via IIIG1"and/or phospho(1982) J. Bid. Chem. 257,14543-14552 111"" (19, 20). However, under conditions described in this 8, Magasanik, B. (1970) in The Lactose Operon (Beckwith, J. R., and Zipser, Cold Spring D., eds)pp. 189-219, ColdSpringHarborLaboratory, report, it is possible to induce the lac operon in DS336, as Harbor. NY judged by the levels of P-galactosidase. Since lactose permease 9. Makman,'R. S . , and Sutherland, E. W. (1965) J. Biol. Chem. 2 4 0 , 13091314 is synthesized coordinately, and assuming that the permease 10. I., and Perlman, R. L. (1970) Science 169,339-344 is normally inserted into the cell membrane, we ascribe the 11. Pastan, Peterkofsky, A., and Gazdar, C. (1974) Proc. Natl. Acad. Sei. U. S. A. 71, 2324-2328 properties of lactose permease in DS336 to the high levels of 12. Koch, A. L. (1964) Biochim. Biophys. Acta 79, 177-200 111"" in the cell. 13. Winkler, H. H., and Wilson, T. H. (1967) Biochim. Biophys. Acta 1 3 5 , 1030-1051 The behavior of the three strains on fermentation plates, 14. Saier, M. H., Jr., and Roseman,S. (1976) J. Biol. Chem. 251,6606-6615 which contain a rich growth medium in addition to the car15. Saier, M. H., Jr., Simoni, R. D., and Roseman, S. (1976) J. Bid. Chem. 252,6584-6597 bohydrate being tested, generally confirmed the quantitative 16. Saier, M. H., Jr., and Roseman, S. (1976) J. Biol. Chem. 251, 6598-6605 growth behavior in minimalmedium. That is, all three strains 17. Roseman, S. (1977) Fed. Eur. Biochem. SOC. Symp. (Berlin)42, 582-597 fermented PTS sugars. The control (DS344), and the 2-fold 18. Saier, M. H., Jr. (1977) Bacteriol. Reu. 41, 856-871 19. Reddy, P., Meadow, N.,Roseman, S., and Peterkofsky, A. (1985) Proc. overproducer (DS3401, fermented the four non-PTS sugars Natl. Acad. Sci. U. S. A. 82,8300-8304 20. Liberman, E., Saffen, D., Roseman, S., and Peterkofsky, A. (1986) Biochem. approximately equally well, while DS336 showed good ferBiophys. Res. Commun. 1 4 1 , 1138-1144 mentation of lactose but poor fermentation of the other three 21. Meadow, N. D., Saffen, D. W., Dottin, R. P., and Roseman, S. (1982) Proc. Natl. Acad. Sci. U. S. A. 7 9 , 2528-2532 non-PTS sugars. Wehave noted (14-16,23) that PTS-medi22. Meadow, N. D., Rosenberg, J. M., Pinkert, H. M., and Roseman, S. (1982) ated repression is not effected equally on all non-PTSsugars. J. Biol. Chem. 2 5 7 , 14538-14542 When the fermentation plates contained methyl a-glucoside, 23. Mitchell, W. J., Misko, T. P., and Roseman, S. (1982) J . Biol. Chem. 257, 14553-14564 the fermentation behavior of bS340 was affected on glycerol 24. Nelson, S. O., Scholte, B. J., and Postma, P. W. (1982) J. Bacteriol. 150, 604-615 and melibiose, and thatof DS336 on allfour non-PTS sugars. 25. Brosius, J., Cate, R. L., and Perlmutter, A. P. (1982) J. Biol. Chern. 257, Thefermentationplateresultstakentogetherwiththe 9205-9210 growth behavior in minimal media lead to the same general 26. Csonka, L. N.,and Clark, A. J. (1980) J Bacteriol. 143, 529-530 27. Bochner, B. R., Huang, H., Schieven, G. L., and Ames, B. N. (1980) J . conclusion, namely that cells containing increased levels of Bacteriol. 143, 926-933 111"" are more sensitive to PTS-mediated repression of growth 28. Maloy, S. R., and Nunn, W. D. (1981) J. Bacteriol. 1 4 5 , 1110-1112 29. Difco Laboratories (1953) Difco Manual, 9th ed., pp. 131-132, Difco Laboand of fermentation of at least some non-PTS sugars. ratories, Detroit, MI 30. Davis, R. W., Botstein, D., and Roth,J. R. (eds) (1980) Advanced Bacterial Finally,thetransportexperiments gave resultsentirely Genetics. p. 201, Cold Spring Harbor Laboratory, Cold Spring Harbor, consistent with the growth and fermentation behavior on the New York non-PTS sugars. Thedatapresentedheretherefore agree 31. Bensadoun, A,, and Weinstein, D. (1976) Anal. Biochem. 70, 241-250 P. C., and Wilson, T. H. (1973) Biochim. Biophys. Acta 330, with the previous conclusion that inducerexclusion is a func- 32. Maloney, 196-205 33. Wong, P. T. S., and Wilson, T. H. (1970) Biochim. Biophys. Acta 196, tion of the cellular content of both IIIG1' andthetarget 336-350 transport protein (23, 24). 34. Beneski, D. A., Misko, T. P., and Roseman, S. (1982) J . Biol. Chem. 257, 14565-14575 Methods are now being developed to determine thein vivo T. P., Mitchell, W. J., Meadow, N. D., and Roseman, S. (1982) Fed. levels of the two forms of IIIG1'which differ by the N-terminal 35. Misko, Proc. 41, 1416 36. Dills, S. S., Schmidt, M. R., and Saier, M. H., Jr. (1982) J. Cell. Biochem. heptapeptide (6), as well as the corresponding phosphopro18, 239-244 teins. With the availability of such methods, and the strains 37. Nelson, S. O., Wright, J. K., and Postma, P. W. (1983) EMBO J. 2, 715720 constructed and described in this report, it may be possible T., and Saier, M. H., Jr. (1982) Proc. Natl. Acad. Sci. U. S. A. 79, to determine whether the ratio of IIIG1'to phospho-111"" is 38. Osumi, 1- A R 7 - -1 ." Akl 39. Saier, M. H., Jr., Novotny, M. J., Comeau-Fuhrman, D., Osumi, T., and indeed important in regulating the non-PTS permeases and Desai, J. D.(1983) J. Bacteriol. 1 5 5 , 1351-1357 adenylate cyclase as suggested in the original hypothesis, or 40. Postma, P. W., Epstein, W., Schuitema, A. R. J., and Nelson, S. 0. (1984) J. Bacteriol. 158. 351-353 whether Enzyme I and HPr arealso required as indicated by M. J., Frederickson, W. L., Waygood, E. B., and Saier, M. H. the most recent results with adenylate cyclase (19, 20). Fur- 41. Novotny, (1985) J . Bacterid. 1 6 2 , 810-816 thermore, these strains shouldprove important in determin- 42. DeBoer, M., Broekhuizen, C. P., and Postma, P. W. (1986) J. Bacterid. 167,393-395 ing the physiological significance of further in vitro studies 43. Misko, T. P., Mitchell, W. J., Meadow, N. D., and Roseman, S. (1987) J . Biol. Chem. 2 6 2 , 16261-16266 on the regulation of the non-PTS permeasesby the PTS. ~~~

~

~

~~