Reconstitution of the Bacterial Chemotaxis Signal Transduction ...

THE JOURNAL OF BIOLOGICAL CHEMISTRY

Vol. 266, No. 15, Issue of May 25, pp. 9764-9770, 1991 Printed in U.S.A.

0 1991 by The American Society for Biochemistry and Molecular Biology, Inc

Reconstitution ofthe Bacterial Chemotaxis Signal Transduction System from Purified Components* (Received for publication, October 9, 1990)

Elizabeth G . NinfaSP, Ann StockSll,Sherry Mowbrayll, and Jeff Stock$** $From the Departments of Molecular Biology and Chemistry, Princeton University, Princeton, New Jersey 08544, the $Department of Biochemistry, Wayne State University School of Medicine, Detroit, Michigan 48201, and the IlDepartment of Molecular Biology, Swedish University of Agricultural Sciences, UppsalaS-751 24,Sweden

In bacterial chemotaxis, transmembrane receptor proteins detect attractants and repellents in the medium and send intracellular signals that control motility. The cytoplasmic proteins that transduce information from the receptors to the flagellar motor have previously been purified and many of their enzymatic activitieshave been identified. Here we report the reconstitution of the complete signal transduction systemfrom purified components. The protein kinase, CheA, plays a central role in both the initial excitation response to stimuli as well as subsequent events associated with adaptation. Thiskinase provides phosphoryl groups to two acceptor proteins, CheY, which interacts with the flagellar motor, and CheB, which demethylates the receptors. The purified aspartate receptor, Tar, reconstituted into phospholipid vesicles, acts in conjunction with an auxiliary protein, Chew, to stimulate the rate of kinase autophosphorylation greater than 10-fold. This stimulation is inhibited by aspartate. The activity of the kinase is increased by increased levels of receptor methylation. This effect provides a mechanism that explains how changes in receptor methylation mediate adaptive responses to attractant and repellant stimuli.

to continue moving toward increasing attractant concentrations. Activities have been ascribed to many of the Che proteins. CheA is a kinase that catalyzes the transfer of a phosphoryl group from ATP to an imidazole nitrogen on one of its own histidine side chains (Stock et al., 1988; Hess et al., 1988a). This phosphoryl group is transferred to an asparticacid side chain in the CheY protein (Stock et al., 1988; Hess et al., 1988b; Sanders et al., 1989a). Phospho-CheY is thought to interact directly with the flagellar motor to cause tumbly swimming behavior (Yamaguchiet al.,1986; Ravid et al.,1986; Wolfe et al.,1987; Smith et al., 1988; Ninfa et al.,1988; Oosawa et al., 1988). Chew acts in conjunctionwith the receptors to regulate the CheA-dependent phosphorylation of CheY (Borkovich et al., 1989), and CheZ facilitates the dephosphorylation of phospho-CheY (Hess et al., 1988b). In addition to components that act directly to regulate motility by controlling CheY phosphorylation and dephosphorylation, another set of enzymes, CheR and CheB, functions indirectly by controlling levels of receptor methylation (for reviews see Stock and Stock,1987; Stock, 1990). CheR is S-adenosylmethionine-dea transferasethatcatalyzesthe pendent methylesterification of glutamate side chains in the membrane receptor proteins, and CheB is an esterase that catalyzes the hydrolysis of thesemethylesters.Attractant stimulicauseincreasedmethyltransferaseand decreased In motile bacteria such as Escherichia coli and Salmonella methylesterase activity,leading to increasedlevels of receptor methylation. High levels of receptor methylation have been typhimurium, a system of interactingproteinstransduces environmental signals into a regulatory output that controls correlated with tumblybehavior. Each of the Che proteins aswell as the aspartate receptor the flagellar motor (for reviews see Macnab, 1985; Stewart and Dahlquist,1987; Stock et al., 1991). Chemical stimuli bind have been purified to homogeneity and characterized (Simms to extracytoplasmic domains of transmembrane receptor pro- et al., 1985 and 1987; Stock et al., 1985, 1987, and 1988; Stock teins whose cytoplasmic domains communicate with signal and Stock, 1987a; Foster et al., 1985). Here we show that a defined transduction components within the cytosol. Signal transduc- these isolated components can be reconstituted into system where CheY phosphorylation is controlled by stimution involves the products of six chemotaxis or che genes, latory ligands. Under these conditions, kinase activity modis designated CheA, CheB,CheR,Chew, CheY, and CheZ. ulated by the level of receptor methylation. Increased levels These proteins process information from the receptors and of methylation cause increases in kinase activity thatlead to use it to direct swimming toward favorable environmental increased rates of CheY phosphorylation. These results proconditions. Cells control their direction of movement by mod- vide amolecular basis for understandingchemotacticreulating the probability that they tumble and changecourse. sponses in living cells: conditions that increase CheY phosso that cells tend phorylation in vitro cause increased tumbly behavior in vivo. Attractant stimuli suppress tumbly behavior *This work was supported in part by Grant AI20980 from the National Institutes of Health, Grant MV486 from the American Cancer Society and by a grant from the Lucille P. Markey Charitable Trust. 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. li Lucille P. Markey Scholar. ** To whom correspondence should be addressed Dept. of Molecular Biology, Princeton University, Princeton, NJ 08544.

EXPERIMENTALPROCEDURES

Proteins-All proteins were derived from S. typhimurium genes expressed at high levels in E. coli. The aspartate receptor with a Gln to Glu substitution at amino acid 309 (methylation site3; Terwilliger and Koshland, 1984) encoded in the plasmid pDK121 was expressed in E. coli RP3808 A(cheA-cheZ)2209 tsr leu his thi (J. S. Parkinson, University of Utah). The receptor was purified by the procedure of Foster et al. (1985); the final preparations contained 50 mM TrisHC1, pH 7.4, 10% (w/v) glycerol, 1%P-octylglucoside, and 50 mM

9764

Signal Transductionin Bacterial Chemotaxis

9765

NaCl. All other proteinswere purified as described previously (Simms 32 pl of 50 mM potassium phosphate, pH 7.5, 5 mM MgCl,, 25 pg/ml bovine serum albumin, 1 mM ATP, and 50 p M [8-I4C]ADP(ICN). All et al., 1985, 1987; Stock et al., 1985, 1987, 1988). Receptor Reconstitution-Purified aspartate receptors in octyl glu- components except ADP were mixed on ice and then incubated 15 coside were stored frozen at -20 "C. Just prior to reconstitution into min at 20 "C. The exchange reaction was initiated by addition of ["C] phospholipid vesicles these preparations were thawed by incubation ADP, and 5-pl aliquots were removed and quenched with 5 pl of 1 N at 0 "C, and then mixed with 1 volume of 2 X reconstitution buffer acetic acid at the indicated times. Radiolabeled nucleotides were for 1 h a t 23 "C.The resulting mixtures containedreceptor in 25 mM analyzed by thin layer chromatography on polyethy1eneimine)-celluTris-HC1, pH 7.4, 25 mM NaCI, 40% (w/v) glycerol,0.5% octyl lose plates using 0.8 M LiCl, 0.8 N formic acid as the mobile phase. glucoside, 0.5 mM 1,lO-phenanthroline, 0.1 mM phenylmethylsulfonyl Regions of the chromatogram containing ATP andADP were excised fluoride (to inhibit possible protease contaminants), and 1.8 mg/ml and counted by liquid scintillation spectrometry. E. coli phospholipids (Avanti PolarLipids, Inc.). A control that lacked receptor protein was prepared in parallel. Phospholipids at 10 mg/ml RESULTS were sonicated in a Branson 1200 ultrasonic cleaner for 1 min under N, immediately prior to use. Finally, octyl glucoside was removedby Reconstituted Aspartate Receptor Stimulates dialyzing samples for 18-24 h at 4 "C in a buffer consisting of 25 mM Kinase-dependent Phosphorylationof CheY Tris-HCI, pH 7.4,25 mM NaCI, 40% (w/v) glycerol,0.5 mM 1,lOphenanthroline, and 0.1 mM phenylmethylsulfonyl fluoride. ReconCheA is a kinase that catalyzes a slow transfer of phosstituted receptors and receptor-deficient liposome controls were phoryl groups from ATP to CheY in the presence of Mg2+ stored at -20 "C. Freeze-fracture Electron Microscopy-The procedure was essen- (Wylie e t al., 1988; Hess et al., 1988b). No other component tially that of Ellens et al. (1989). Aliquots of reconstituted receptor is required forthis activity. Borkoviche t al. (1989) have shown (0.1-0.3 pl of a 1 p~ solution) were sandwiched between a pair of that CheY phosphorylation is dramatically stimulated when Balzers (Nashua, NH) copper support plates at 20 "C and rapidly kinase reaction mixtures are supplemented with Chew and plunged into liquid propane. These were fractured and replicated on chemoreceptor prombar cell membrane preparations that contain a Balzers BAF 400 freeze-fracture unit at a vacuum of 4 X or better, a t -115 "C. Replicas were floated off in 3.0 N HN03 and teins. In order to characterize receptor/CheW-mediated stimwashed in agraded series of Clorox solutions (0-100%). Replicas were ulation of the kinase reaction in a defined system, we incorviewed in a Philips 300 electron microscope. porated the purified S. typhimurium aspartate receptor, Tar, Protein Phosphorylation-Unless otherwise indicated, reaction into phospholipid vesicles (Fig. 1).In the presence of C h e w mixtures contained various additions of protein and phospholipid this reconstituted receptor preparationcaused a greater than vesicles in 15-30 p1 of 50 mM potassium phosphate, pH 7.0, 5.0 mM MgCI,, 0.10 mM dithiothreitol, and [-y-32P]ATP(ICN). Receptor- 10-fold increase in the CheA-dependent phosphorylation of containing or control liposomes, in reconstitution buffer, comprised CheY (Fig. 2). Neither thereceptor nor Chew, eitherindividhalf of the reaction volume. Assaysfor CheB or CheY phosphorylation ually or together, had any detectable ability to phosphorylate were conducted in 0.40 mM ATP while CheA phosphorylation was CheY in the absence of the CheA. In the presence of CheA, measured in 50 PM ATP. Aliquots were removed at various times and neither the receptor nor Chew had effect any when they were added to SDSI-EDTA sample buffer (2.6% SDS, 30 mM EDTA, 12.5 added alone. Only the complete mixture containing all three mM Tris-HC1, pH 6.8, 12.5% (w/v) glycerol, and 1%P-mercaptoethanol), heated for 30 s a t 65 "C,and subjected to SDS-polyacrylamide components hadelevated kinase activity. The concentration of receptor, kinase, and Chew were gel electrophoresis according to the method of Laemmli (1970). Gels were either autoradiographed directly, or stained for 10 min with systematically varied in the reaction mixtures. ConcentraCoomassie Brilliant Blue R-250, and destained 1 h in 10% acetic acid. tions of receptor, kinase, and C h e w within S. typhimurium Gel slices containing the relevant protein were excised and analyzed cells are in themicromolar range (aspartate receptor,-5 pM, for radioactivity by liquid scintillation spectrometry. When this pro- kinase and Chew, -1 PM).' Uptothese levels, increased cedure was used, control reactions lacking CheA were run to provide concentrations of receptor and/or Chew caused proportional values for background radioactivity. increases in CheY phosphorylation. Although increased kiReceptor Methylation-Reconstituted receptor preparationsor control liposomes (comprising 80% of the reaction volume) were nase concentrations caused increases in CheY phosphorylaincubated with the CheR methyltransferase in 1.1 mlof 100 mM tion, at lower concentrations of kinase there was a greater potassium phosphate, pH 7.0, and 40 PM S-aden~syl[~H-rnethyl] degree of stimulation by receptor and Chew. This result is methionine (Du Pont-New England Nuclear). Aliquots (7.5 p l ) were consistent with the presence of two kinase populations in the removed at various times, and applied to 1-cm2Whatman 3MM filter papers. These squares were immediately added to 10% trichloroacetic reconstituted system: a t low concentrations of kinase a high acid, washed with methanol, air-dried, and assayed for radioactivity activity form in a complex with receptor and C h e w predominates while a t high concentrations of kinase a low activity by liquid scintillation spectrometry. ATPase Assay-ATPase activity was assayed essentially as de- form that is free in solution predominates. Since we wished scribed by Norby (1988). Proteins weremixed with reconstituted t o investigatephosphorylationreactions mediated by the receptor or control liposomes (which comprised half of the reaction receptor-Chew-kinase complex rather than by free kinase, volume) in 100 p1of 100 mM potassium phosphate, pH 7.0, 5 mM MgCl,, 0.1 mM EDTA, 0.1 mM dithiothreitol,3 mM ATP, 1 mM submicromolar concentrationsof kinase were used with relaphosphoenolpyruvate, 0.2 mM NADH, 2 units of pyruvate kinase tively high concentrations of receptor and Chew. At concentrations of C h e w significantly higher than those (Boehringer Mannheim), and 6.6 units of lactate dehydrogenase (Sigma). Reactions were conducted a t ambient temperature in a100- normally found in thecell (>5 p ~ )Chew , had an inhibitory pl microcuvette, and NADH oxidation was monitored in a Beckman effect on CheY phosphorylation (Fig. 3). This result is conDU-65 spectrophotometer which recorded absorbance at 340 nm every sistent with the effects of varying levels of c h e w expression 30 s. The reactions were initiated by the addition of kinase CheA, and the course of the reaction was monitored until a stable linear on swimming behavior i n vivo where both abnormally low or decay of absorbance was observed. The rateof decay over 15-60 min high levels of C h e w have been shown to cause dramatic decreases in tumbling behavior (Stewart e t al., 1988; Sanders was used to calculate ATP hydrolysis rates, using a value of 6220 M-' cm" for the extinction coefficient of NADH. In control experiments, et al., 1989b; Liu and Parkinson, 1989). in the absence of Che proteins, expected decreases were rapidly Receptor and Che W Increase the Steady-state Level of Phosgenerated by addition of micromolar concentrations of ADP, indicat- pho-CheY-CheY phosphorylation reachesa steady-state ing that the coupling reactions were not rate-limiting. ATPIADP Exchange Reaction-Reaction mixtures contained var- where the rate of phosphorylation by the kinase equals the of purified receptor ious combinations of proteins, with reconstituted receptor prepara- rate of dephosphorylation. In the presence tions or controlliposomes comprising half of the reaction volume, in and Chew, a higher steady-state level of phospho-CheY is The abbreviation used is: SDS, sodium dodecyl sulfate.

J. Stock, unpublished results.

Signal Transductionin Bacterial Chemotaxis

9766

FIG. 1. Aspartate receptors in phospholipid vesicles. I'anels A and R, show freeze-fracture electron micrographs of liposomes containing aspartate receptor; panel C showscontrol liposomes lacking receptor. Magnification is 75.000-fold.

Receptor Chew CheA

- + - + - + - + - " -+ + "+ + ++ ++ "

-0

-

P-CheY

+Receptor

Y

FIG.2. Signal transduction components that stimulate the phosphorylation of CheY. Proteins were mixed on ice in the indicated comhinations. The final concentrations were 3.2 p~ aspartate receptor (reconstituted into liposomes as descrihed under "Experimental Procedures"), 4p M Chew, 0.2 p M CheA, and 10 p M CheY. In mixtures lacking receptor, liposomes prepared in the absence of receptor were added. Samples were incubated at 23 "C for 2 min, and the phosphorylation reaction was initiated hy addition of [y-:'"P]ATP (finalconcentration, 0.40 m M ; specific activity, 2,270 cpm/pmol). After 10 s, reactions were quenched with SDS-EDTA buffer, and samples were analyzed for phospho-CheY hv polyacrylamide gel electrophoresis and autoradiography (see "Experimental Procedures").

PI 6

lomt -Receptor

" 0

30

0

60

90

Time [eec]

FIG. 4. Timecourse of CheY phosphorylation. Aspartate receptors (or an equalvolume of liposomes in the minusrereptor control) were mixed with CheW, kinase, antl CheS on ice. The linal concentrations of protein were 0.6 p M receptor, 4 pM Chetf'. 0.2 p M CheA, and 10 p~ CheY. After a 2-min incuhation at 23 "C. ATP WAS added to the reaction mixture (final Concentration, 0.40 p ~ specific ; activity, 1,400 cpm/pmol). At the indicated times after ATP addition, aliquots were removed and added to SDS-EIITA huffer to stop the r e a d o n . Samples were analvzed for phospho-CheYas descrihed under "Experimental Procedures." TAMEI

0

5

10

15

Chrmotaxis .signal transducing ATPasr Hydrolysis of ATP by the CheA/CheY ATPasewas coupled to the pyruvatekinaseandlactatedehydrogenasereactions as descrihed under "Experimental Procedures." Chemotaxis proteins were present at the following concentrations: CheA. 0.2 p ~ CheS, ; 10 p ~ rhrlV, ; 4 pM; receptor, 0.5 pM. Aspartate was present, where indicated. at 50

20

Chew [ru]

FIG.3. Dependence of CheY phosphorylationonChew. C h e w was addedtoreceptor,kinase,and CheY on ice, antlthe mixtures were incubated for 2 min a t 23 "C. The final concentrations of protein were 0.6 p M receptor, 0.2 p M CheA, 10 p M CheY, and Chew as indicated. ATP was added to initiate the phosphorylation reaction (final concentration, 0.40 mM; specific activity, 1,330 cpm/ pmol). After 10 s, the reactions were terminated and phospho-CheY was assayed as descrihed under "Experimental Procedures."

p.

Assay romponrntn ".

-~ -

ATP hvdrnlyri5-~ rote "

p.w/rnin

Receptor+CheW+CheA+CheY Receptor+CheW+CheA+CheY+Asp CheW+CheA+CheY Receptor+CheA+CheY Receptor+CheW+CheY Receptor+CheW+CheA

3.1

0.2 0.2 0.2