Purification and Characterization of the Membrane-associated ...

30 downloads 53 Views 4MB Size Report
Hofnung, 1985; Froshauer and Beckwith, 1984). The MalK protein is more hydrophilic and appears to be peripherally associated with the cytoplasmic surface of ...
THEJOURNAL OF BIOLOGICAL CHEMISTRY Q 1991 by The American Society for Biochemistry and Molecular Biology, h e .

Vol. 266. No. 14, Issue of May 15, pp. 8946-8951,1991 Printed in U.S. A .

Purification and Characterization of the Membrane-associated Components ofthe Maltose TransportSystem from Escherichia coli* (Received for publication, November 5,1990)

Amy L. DavidsonS and Hiroshi Nikaido From the Department of Molecular and Cell Biology, University of California, Berkeley, California 94720

Maltose is transported across the cytoplasmic membrane of Escherichia coli by a binding protein-dependent transport system. The three membrane-associated components of the transport system, the MalK, MalF, and MalG proteins, have been solubilized from the membrane andmaltose transport activity has been reconstituted in proteoliposome vesicles (Davidson, A. L., and Nikaido, H. (1990)J. Biol. Chern. 265, 42544260). A modification of the reconstitution technique is presented which permits reconstitution from the detergent dodecyl maltoside. Utilizing reconstitution of maltose transport as an assay, we have purified these proteins in the presence of n-dodecyl-j3-D-maltoside. The purified proteins catalyze both maltose transport activity and ATP hydrolysis. In all experiments, the MalF, MalG, and MalK proteins behaved as a multiprotein complex; all three proteins were immunoprecipitated using antibody prepared against MalF, and they copurified, eluting from a gel filtration column between markers of M , 160,000 and 200,000. Each complex contains two MalK, one MalF, and one MalG proteins, providing two putative sites for ATP hydrolysis. Chemical cross-linking detected specific interactions between MalF and MalG and between MalF and MalK.

of periplasmic binding protein-dependent transport systems in bacteria (Dassa and Hofnung, 1985; Duplay et al., 1984; Froshauer andBeckwith, 1984). Early studiesof these “shocksensitive” permeases suggested that transport was energized by ATP orsome other formof phosphate bondenergy (Berger and Heppel, 1974). An important feature of these systems is that the MalK protein and homologues its contain a putative nucleotide binding domain(Higgins et al., 1985). Two of these proteins, OppD, a subunit of the oligopeptide permease (Higgins etal., 1985), and HisP,a subunit of the histidine permease (Hobson et al., 1984), have been labeled with 8-azido-ATP. The presence of an ATP binding site was consistent with a model in which the hydrolysis of ATP provided the energy for maltose transport. In agreementwith this model, we have succeeded in reconstituting maltose transport activity in proteoliposome vesicles fromdetergent-solubilized membrane proteins using ATP as thesole energy source (Davidson and Nikaido, 1990): ATP hydrolysis, tightly coupled to transport, has been detected in this system, as well as in membrane vesicles (Deanet al., 1989), and in anotherreconstituted transportsystem,thehistidinepermease from Salmonella typhimurium (Bishop etal., 1989). Over the past several years, proteins not involved in periplasmic binding-dependent transport, butshowing homology to MalK, have been discovered in severalprocaryotic and eucaryotic organisms, and it is believed that these proteins also function as ATP-dependent transport systems. One exThree of the proteins required for active maltose uptake ample is the mammalian P-glycoprotein, an ATPase which into Escherichia coli, the MalF, MalG, and MalK proteins, may couple ATP hydrolysis to efflux of anti-cancer drugs have been localized to the cytoplasmic membrane (Bavoil et from multidrug-resistant cells (Hamada and Tsuruo, 1988). al., 1980; Dassa, 1990; Shuman and Silhavy,1981; Shuman et Another is the STE6gene product which mediates export of al., 1980). The MalF andMalG proteins arevery hydrophobic the yeast mating pheromone, cy-factor (Kuchler et al., 1989). and likely spanthemembrane several times(Dassaand The gene responsible for cystic fibrosis, CFTR, whose exact Hofnung, 1985; Froshauer and Beckwith, 1984). The MalK function has not yetbeen elucidated, is also a member of this protein is more hydrophilic and appears to be peripherally family (Riordan et al., 1989). associated with the cytoplasmicsurface of the membrane, In this paper,we report on thepurification of a membraneperhaps via an interaction withMalG(Bavoil et al., 1980; associated complex composed of the MalF, MalG, and MalK Shuman and Silhavy,1981). The last protein required for proteins which catalyzes the uptake of maltose inthe presence transport, the product of the malE gene, is aperiplasmic of the maltose binding protein and the concomitant hydrolysis maltose binding protein (MBP).’ This system isofone a class of ATP. * This work was supported in partby Public Health Service Grant AI-09644. 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. $- Supported by PublicHealth Service Postdoctoral Fellowship GM-11717. T o whom correspondence should be addressed Dept. of Molecular and Cell Biology, Stanley/Donner ASU, University of California,Berkeley, CA 94720. Tel.: 415-642-2546; Fax: 415-6439290. ’ The abbreviations used are: MBP, maltose binding protein; dodecyl maltoside, n-dodecyl-0-D-maltoside; DSP, dithiobis-(succinimidyl propionate); octylglucoside, octyl-P-o-glucopyranoside;SDS, sodium dodecyl sulfate.

EXPERIMENTALPROCEDURES

Bacteria,Plasmids, Growth Conditions, and Radiolabeling-The MalF, MalG, and MalK proteins were overproduced simultaneously from two plasmids; pFG23 carries the malF and malG genes under control of the trc promoter and pMRll carries themalK gene under control of the trc promoter (Davidson andNikaido, 1990; Reyes and Shurnan, 1988). These plasmids were transformed into the strain HN597 ( m a l y AuncB-C i1ur:TnlO araD lac rpsLI/F’ l a c a lacZ::Tn5, proA+ proB+) which carries a deletion of the FoFt-ATPase (unc) (Davidson and Nikaido, 1990). Freshly transformed cells were grown with aeration at 37 “C in twice-strength Luria broth(20 g of tryptone, 20 g of yeast of extract, 5 g of NaCI) with 100 rg/ml ampicillin, 34 rg/rnl chloramphenicol, and 50 pg/ml kanamycin. Whencells reached

8946

Transport Maltose Complex the Purification of a density equivalent to 0.2 AGIHI units, expressionof theplasmid encoded mal genes was induced by adding isopropyl-0-D-thiogalactoside (Sigma) toa final concentration of 0.1 mM. Cells were harvested after 3 h of induction. The strain HN93 ( F argHl lac rpsLI malT metA28)(Bavoil et al., 1980),a methionineauxotrophexpressingthe maltosegenes constitutively, was used to label cells continuously with ["S]methionine. This strainwas grown with aerationat 37 "C to stationary phase inminimal medium M63(Miller, 1972) supplementedwith 0.2% maltose, 5 pg/mlthiamine, 100 p~ arginine, and5 p~ ["S]methionine (Amersham Corp., specific activity 16 mCi/pmol). Isolation of Total Membrane-Cells were chilled to 0 "C, harvested by centrifugation at 3000 X g for 10 min, washed once with 50 mM Tris-SO,, pH 8.0, 1 mM EDTA, and resuspended in the samebuffer. They were then lysed in the presenceof DNase by two passes through a French pressurecell (American InstrumentCo.) operating at 10,000 p. s. i. Unbroken cells were removed by centrifugation a t 3000 X g for 10 min, the supernatant was transferred to a clean tube, and the centrifugation step repeated. To pellet the membrane fractions, the supernatant was subjected to centrifugation a t 100,000 x g for 1 h. The membrane pellet was resuspended in 50 mM Tris-SO,, pH 8.0, and subjected to a second centrifugation at 100,000 X g for 1 h. The final pellet was suspended in 50 mM Tris-S04, pH 8.0, a t a protein concentration of 10 mg/ml and frozen in aliquots a t -70 "C. Membranes from radiolabeled cells were prepared by the same method except that cells were lysed with 3 X 15-s burstsof a probe sonicator (Soniprep 150, Gallenkamp). Purification of MalF, MalG, and MulK-AI1 manipulations were performed a t 0-4 "C, except for the chromatographic procedures, which were performed a t room temperature. Columnswere connected to a Perkin-Elmer Series 3B high performance liquid chromatography pump and absorbance of the eluate was monitored continuously a t 280 nm. Membrane vesicles, prepared as described above, (10 mg/ml, typically 30 mg total) were incubated with 1.5 volumes of 10 M urea (ultrapure, Bio-Rad)for 10 min on ice and collected by centrifugation for 30 min at 100,000 X g. The pellet was washed once in 20 mM TrisSO,, pH 8.0, and resuspended a t a concentration of approximately 2.5 mg protein/ml in buffer containing 20 mM Tris-SO,, p H 8.0, 5 mM MgCl,, 1 mM dithiothreitol,and 20% glycerol. Proteins were solubilized from the urea-treated membrane by the addition of 1% dodecyl maltoside (n-dodecyl (3-D-maltoside, Boehringer Mannbeim). After a 20-min incubation, the particulate fraction was removed by centrifugation for 30 min at 100,000 X g. The supernatant was loaded onto a DEAE-5 PW column (Bio-Rad) equilibrated with bufferA (containing 20 mM Tris-SO,, pH 8.0, 1 mM dithiothreitol, 20% glycerol, and 0.01% dodecyl maltoside). The column was washed with 10 ml of buffer A and then eluted with a 30-ml gradient (from 0 to 250 mM Na,SO,) using buffer A plus 0.5 M Na,SO, a t a flow rate of 0.5 ml/min. The desired fractions were pooled and concentrated by centrifugationin a Centricon-30 ( M , 30,000 cutoff, Amicon). A 100-pl aliquot, usually containing 300 pg of protein, was applied to a Superose 12 column (HR 10/30, Pharmacia LKBBiotechnology Inc.) equilibrated with buffer B (50 mM KP,, pH 7.0,50mM Na2S0,, 1 mM dithiothreitol, 20% glycerol, and 0.01% dodecyl maltoside) and eluted a t a flow rate of 0.2 ml/min. Reconstitution Assay-Maltose transport activitywas assayed during the purification by reconstitution of the proteins into proteoliposome vesicles. We have shown that when ATP is trapped inside, proteoliposomes will accumulate maltose in the presence of MBP (Davidson and Nikaido, 1990). Proteoliposomes were prepared using E . coli phospholipids by an octyl glucoside (octyl-0-D-glucopyranoside) dilution procedure, as described previously (Davidson and Nikaido, 1990). From 50 to 78 pl of column fractions were mixed with 16 p1 of sonicated E. coli phospholipids (50 mg/ml stock), and 5 mM ATP and octyl glucoside were added to a final concentration of 1%. Proteoliposomes were formed upon a 25-fold dilution of the mixture into 20 mM KP,, pH 6.2, 5 mM ATP, and 1 mM dithiothreitol. To prepare proteoliposomes from proteins solubilized with dodecyl maltoside, the concentration of dodecyl maltoside was first reduced to 0.01% during chromatography on the DEAE column, and then octyl glucoside was added (see above) prior to the detergent dilution step. The initial rate of ["C]maltose uptake in the presence of purified MBP was measured as described previously (Davidson and Nikaido, 1990) except that 0.22-pm GSWP filters (MilliporeCorp.) were used to collect vesicles. All values were corrected for background counts/ min on filters in the presenceof MBP and maltose only. Assay of ATP Hydrolysis-For measurement of ATP hydrolysis, proteoliposome vesicles were prepared by incubatingthe purified

8947 protein fractions (concentrated in a Centricon-30 microconcentrator), phospholipid and octyl glucoside, with 5 mM [y-"'P]ATP (160 pCi/ pmol) in a final volume of 0.1 ml. This mixture was then diluted into medium containing no additional ATP. Proteoliposomes, which had been collected by centrifugation (Davidson and Nikaido,19901, were washed once in 20 mM KP,, pH 6.2, to remove external ATP prior to assay. Assays were carried out asdescribed for transport except that nonradioactive maltose was used, and 25-p1 aliquots of the assay medium were removed atthe indicatedtimesto 175 pl of 1 M perchloric acid and 1 mM KP,. The release of radioactive orthophosphate was assayed essentially as described by Lill et al. (Lill et al., 1989). Cross-linking of Transport Proteins-A stock solution of the crosslinkingreagent DSP (dithiobis-(succinimidylpropionate), Pierce Chemical Co.) was prepared a t a concentration of 4 mg/ml in dimethyl sulfoxide. Purified proteins were dialyzed against 50 mM KP,, pH7.0, and 20% glycerol to remove ditbiothreitol, and then 40 pg of protein was incubated with 1 pl of DSP stock solution in a final volume of 0.2 ml for 2 min a t 23 "C. The reaction was stopped by the addition of an equal volume of 0.5 M Tris-HC1, pH 6.8, mixed with gel loading buffer (0.1 M Tris-HC1, pH 6.8, 4% SDS (sodium dodecyl sulfate), 20% glycerol, and bromphenolblue) in the absence of reducing agent and subjected to two-dimensionalelectrophoresis as described below. Electrophoresis-SDS-polyacrylamide gel electrophoresis was performed according to the method of Lugtenberg (Lugtenberg et al., 1975) except that the concentration of acrylamide was adjusted to 8.5% in the separating gels. In some cases, the pH of the separating gel was changed from pH 8.8 to pH 8.6. Proteins were visualized by silver staining according to the method of Heukeshoven and Dernick (Heukeshoven and Dernick, 1988). For autoradiography, gels were soaked in H,O, dried undera vacuum, andexposed to Kodak X-Omat AR film. Radioactive bands were excised from dried gels and rebydrated. Radioactivitywas determined by treating theslices with tissue solubilizer (NCS, Amersham Corp.) as per manufacturer's instructions and counting the samplein liquid scintillant (OCS, Amersham Corp.). To separate proteins in two dimensions, lanes were cut from a first-dimension (SDS-polyacrylamide) gel immediately following the completion of electrophoresis. The gel slices were soaked in gel loading buffer for 20 min either in thepresence or absence of 10% 2mercaptoetbanol and then placed horizontally across the top of a second SDS-polyacrylamide gel. Immunoblot Analysis-Following electrophoresis, proteins were transferred to nitrocellulose as described by Towbin (Towbin et al., 1979). Immunodetection of proteins was accomplished by using polyclonal anti-MalF, anti-MalG, or anti-MalK rabbit antiserum and alkaline phosphatase conjugate anti-rabbit antibody (Sigma) as described (Blake et al., 1984). Antiserum to MalF and MalKwas raised against purified proteins (Davidson and Nikaido, 1990). Antiserum to MalG was raised against a synthetic peptide (aminoacids 175-194 of the MalG protein) kindly provided by DNAX Corp. Immunoprecipitation-The immunoprecipitation procedure described by Anderson and Blobel (Anderson and Blobel,1983) was modified in order to maintain the transport proteins in their native oligomericconfiguration. Membranepreparations from strains labeled with ["'S]methionine were treated with 1%dodecyl maltoside as described above to solubilize proteins. Following centrifugation, the dodecyl maltoside supernatantwas diluted in10 volumesof media containing 50 mM Tris-HC1 buffer, pH 7.5, 150 mM NaC1, 20% glycerol, 0.01% dodecyl maltoside, 0.1% bovine serum albumin, and 1 mM phenylmethylsulfonyl fluoride (Buffer C) and incubated 4at"C overnight with 10p1 of immune serum. Following a short centrifugation to remove insoluble complexes, 30 pl of a slurry of protein A conjugated to Sepharose4B (Pharmacia, 100 mg beads/ml) was added and the mixture shaken end-to-end for 2 h at 4 "C. The immunoglobulin-protein A complexes were sedimentedin amicrocentrifuge (Fisher, model 235B3, and the pellet was washed three times with 1 ml of buffer C, followed by three washes with 1 ml of buffer C plus 0.5 M NaCl and thentwo washes with buffer C, withoutbovine serum albumin. Between washes the complexes were sedimented in a microcentrifuge for 1min. After the finalwash, the pellets were resuspended in 40 p1 of gel loading buffer either with or without reducing agent. Samples were boiled for 5 min,thebeads were sedimented,and aliquots of the supernatant were loaded onto SDS-polyacrylamide gels. Protein Assays-Protein concentrations were determined using the method of Schaffner and Weissmann (Schaffner and Weissmann,

Purification of the Transport Maltose Complex

8948

1973) modified as described (Newman etal., serum albumin as standard.

1981), using bovine

RESULTS

Purification of MalF, MalG, and MalK Proteins-Purification of the membrane associated components of the maltose transport system began with the 10-fold overproduction of theseproteinsinthemembrane (Davidson and Nikaido, 1990). We have shown previously that proteins can be solubilized from the membrane and reconstituted into proteoliposome vesicles without any apparent loss in activity (Davidson and Nikaido, 1990). Prior tosolubilization of the transport proteins with a detergent, the membrane fraction was 0 0 ’ ” . I , I , I , treatedwith 6 M ureaas described under“Experimental Procedures.” As was true for the lactose permease (Newman 0 10 20 30 40 50 60 et al., 1981), treatment of the membrane with urea removed I FRACTION 50% of the protein without solubilizing significant amounts of the mal proteins or affecting maltose transport activity. B . . Following solubilization of the MalF, MalG, and MalK proteins with 1% dodecyl maltoside, the proteins were purified by ion exchangeand gel filtration chromatography. Transport 96 activity was monitored throughout the purificationfollowing 68 reconstitution of proteins into proteoliposome vesicles and MalK measurement of maltoseaccumulation by the vesicles as *F 43 MalF described under“Experimental Procedures.” Theproteins 3 1 were eluted inbuffer containing 0.01% dodecyl maltoside, and .. - _ + MalG we found that itwas essential to add octyl glucoside to samples 22 prior to reconstitution in order to recover full activity. The 1316 19 2 2 25 28 3 13 4 3 7 4 0 4 3 4 64 9 elution profile of DEAE andgel filtration columns are shown FRACTION in Figs. 1 and 2, respectively. Inboth cases, three major protein bands coeluted with the peak of maltose transport FIG.1. Elution profile of DEAE-5 P W column. The column activity. These havebeen identified by immunoblotting as the was run as described under “Experimental Procedures,” and 0.5-ml MalF, MalG, and MalK proteins (notshown). Because these fractions were collected. A , absorbance and transportactivity. Maltthree proteins copurified, i t seemed likely that they formed a ose transport activity was determined by reconstituting proteins into proteoliposome vesicles and assayingas described under “Experimenfunctional complex in the membrane which was not disso- tal Procedures.” Transport activities (O), reported as nanomoles of ciated upon solubilization with the detergent dodecyl malto- maltose accumulated/min/ml of fraction, are initial rates estimated side. Consistent with their presence aincomplex,the proteins from time points taken at 15,30,60, and 120 s. Protein in the eluant eluted from the Superose12 column between protein markers was monitored by UV absorbance a t 280 nm (solid line). Dashed line of MI 160,000and 200,000 (Fig. 2 A ) , earlier than the majority indicates Na2S04 gradient. B, samples (3 pl) of fractions were subof other membrane proteins of similar M , in a crude mem- jected to SDS-polyacrylamide gel electrophoresis on 8.5% gels. Proteins were visualized by silver staining. The positionsof migration of brane supernatant. The M , of the complex as judged by gel the molecular weight markers MalF, MalG and MalK proteins of and filtration is only an approximation, since the overall shape of (X lo-”) are indicated. the complex is unknown and the complex is likely to be associatedwitha detergent micelle. Dodecyl maltoside by 1990). Both ATP hydrolysis and maltose transport are deitself forms micelles of MI 50,000 (Roseveard et al., 1980). A pendent on the presence of MBP; purified MBP by itself does summary of the purification is presented in Table I and ingel not hydrolyze ATP. For reasons which are not yet clear, the form in Fig. 3. Because the intensity of the silver stain used ratio of ATP hydrolysis to maltose transport varied from 1:l in Fig. 3 is not always proportional to protein concentration, to 30:l in our previous experiments (Davidson and Nikaido, the degree of purity was judged bydensitometric scanning of 1990). When proteoliposome vesicles were prepared using the a SDS-polyacrylamide gel stained with Coomassie Blue (not purified transport proteins, we found that ATP was hydroshown). The final preparationwas approximately 90% pure. The modest increase in specific activity (Table I) reflects the lyzed a t a rate of 20 nmol/min/mg protein upon addition of protein. Thisresult demonfact that, following overexpression of the mal genes, the maltoseandmaltosebinding strates that the maltose transport complex itself functions as predominant proteins in proteoliposomes prepared from crude a n ATPase. The rate of hydrolysis presented here cannot not membranes are already those of maltose transport (Davidson and Nikaido, 1990). Densitometric scanning of Coomassie- be compared directly with the rate of transport in Table I, stained SDS-polyacrylamide gels containing these crude pro- because proteoliposomes used in transport assays were preteoliposome preparations show that as much as 30% of the pared under slightly different conditions than thoseused for protein is the transport complex (see Fig. 3 of Davidson and measurement of hydrolysis (see “Experimental Procedures”). Cross-linking of mal Proteins-Since the mal proteins apNikaido, 1990). As shown in Table I, 3% of the total activity present in the detergent supernatant was recovered following pear to exist in a complex, we searched for subunit interacthe two chromatographic steps. This low yield most likely tions using chemical cross-linking techniques. Treatment of purified proteins in dodecyl maltoside with the cross-linking reflects a loss of protein rather thana loss of activity. Hydrolysis of ATP-We have shown previouslythat, follow- reagent DSP, asdescribed under “Experimental Procedures,” of several new highmolecular ing reconstitutionof the maltose transport proteins into pro-resulted in the appearance teoliposomes, ATP trapped inside the vesicles is hydrolyzed weight bands on SDS-polyacrylamide gels. Because DSP can concomitantwithmaltoseuptake (Davidson and Nikaido, be cleaved with sulfhydryl reducing agents, the composition

2

“WT”7“-



Transport Maltose Complex thePurification of A

VO

43 kDa

200 1 6 06 8

$0.4

1

8949



1

0.1



L s

2

40

30

50

60 80 70

.

-

43

-

31

-

24

-

c MalG

FIG. 3. Summary of purification of MalF, MalG, and MalK. Protein samples (0.5 pg) from each step of the purification were boiled for 5 min in gel loading buffer and applied to a 8.5% SDS-polyacrylamide gel. Proteins were visualized by silver staining. Lane 1, total membrane; lane 2, membrane following urea treatment; lane 3, supernatant fraction following treatment with dodecyl maltoside; lane 4, pooled fractions following DEAE chromatography, lane 5, pooled fractions following gel filtration chromatographyon Superose 12. The positions of migration of the MalF, MalG, and MalK proteins and of molecular weight markers (X lo-”) are indicated.

90

FRACTION

-B

96 68

.

96 68 43

-

31

-

22

MalK MalF

ci-

Mc, 3 5 25 45 65 86 06 26 46 66 87 07 27 4

FRACTION FIG. 2. Elution profile of Superose 12 column. Column fractions (0.2 ml) were collected following gel filtration on Superose 12 as described under “Experimental Procedures.” A , absorbance (solid line) and transport activity (0)were obtained as described in the legend to Fig. 1. The position of the void volume (V,) and of proteins of known M, run under the same conditions are indicatedabove the graph. Markerswere: ovalbumin (43 kDa),bovine serum albumin (68 kDa), immunoglobulinG(160 kDa),and myosin (denatured, 200 kDa). R,samples (3 p l ) of fractions were subjected to SDS-polyacrylamide gel electrophoresis on 8.5% gels. Proteins were visualized by silver staining. The positions of migration of the MalF, MalG, and are indiMalKproteinsand of molecularweight markers (X cated.

TABLE I Summary of purification of MalF, MalG, and MalK Purification was as described under “Experimental Procedures.” Fractions 26-31 from the DEAE column were pooled and concentrated. One-third of the DEAE pool was applied to the Superose 12 column. Fractions 59-64 from the Superose 12 column were pooled to determine the finalactivity. Sample

Total protein

Total Specific activity” activity nrnol/rnin/rng nrnol,rnin ?6 of total protein activity

rn‘

-h Total membrane 31.2 -b Urea-treated membrane 14.4 Detergent supernatant 11.0 5.9‘ 64.9 100 1.1 10.8 11.2 17 DEAE pool 1.8 2.8 0.16d 11.0 Superose 12 pool I’ Initial rate of [’‘CC]maltose uptake into proteoliposome vesicles following reconstitution of transportproteins (see “Experimental Procedures”). * Not determined. ‘Estimate based on activity of preparation solubilized in octyl glucoside. Value shown has been multiplied by three as only one-third of DEAE pool was applied to Superose column.

=.\MalK r MalF

- MalG -

tt

t

(?

$2 FIG.4. Two-dimensional SDS-polyacrylamide gel electrophoresis of transport complex cross-linked by dithiobis-(succinimidyl propionate). Protein samples (0.5 pg) which had been treated with DSP asdescribed under “ExperimentalProcedures” were separated by SDS-polyacrylamide gel electrophoresis (8.5% gel) in the the first dimension in the absence of reducing agent. Gel slices were soaked in gel loading buffer in the absence ( A ) or presence ( R ) of reducing agent prior to separation in the second dimension. The direction of migration is from left to right in the firstdimension and from top to bottom in the second dimension. The subunits arising from cleavage of cross-linked products are visible along the uertical above the arrow. Proteins were visualized by silver-staining. The positions of migration of the MalF ( F ) , MalG ( G ) , and MalK ( K ) proteins are indicated.

of cross-linked products was analyzed using a two-dimensional electrophoresis technique(Fig. 4). In this system, s7.aples were subjectedto electrophoresis under nonreducingconditions in the first dimension (to maintaincross-linked complexes) and under reducing conditions in the second dimension(to cleave the cross-linkingreagent). Proteins which had not been cross-linked migrate on a diagonal in the final gel, whereas proteins which had been cross-linked migrate below the diagonal. The high molecular weight crosslinked complexes are visible on the diagonal in the twodimensional gel in Fig. 4A where the reducing agent was omitted in both dimensions. Treatment with reducing agent prior to electrophoresis in the second dimension (Fig. 4B) revealed the subunitcomposition of the high molecularweight complexes. The major product, migratingat a M , of 58,000 in the first dimension,a iscomplexof MalF andMalG. A product migrating at M , 95,000 is composed of the MalF and MalK proteins. In addition, a high molecular weight complex ( M ,

Purification of the Maltose Transport Complex

8950

TABLEI1 116,000) of all three proteins is also seen. Cross-linking beRatio of Ma1K:MalF:MalG in immunoprecipitated complex tween MalF and MalG with DSP was also monitored using Radioactive bands were excised from gels ofpH 8.8 (for MalG) and immunoblot techniques in crude membrane fractions solubilized with a variety of detergents. Cross-linkingwas detected pH 8.6 (for MalF, and MalK), (see Fig. 5, B and C). Equal counts/ when membrane proteins were solubilized in either dodecyl min were loaded onto the two gels. Results are means f S.D. for n = 3 determinations. maltoside or octyl glucoside, but not when proteins were result Experimental Ratio solubilized in Zwittergent 3-14 (Calbiochem), sodium N-lauMalK:MalF 2.2 f 0.4 ryl sarcosine (Sigma), or SDS (data not shown). Itwas also 1.8 f 0.1 Ma1K:MalG possible to detect, by immunoblotting, the linkage between Ma1F:MalG 0.85 k 0.16 MalF andMalG when whole cells or isolated membranes were treated with DSP (data not shown). These results, taken together, argue that (a) the cross-linking interactions in do- to MalF, three proteins were precipitated. The identity of decyl maltoside result from close juxtaposition of subunits these proteins as the MalF, MalG, and MalK proteins was rather than nonspecific interactions occurring in detergent confirmed by immunoblotting (Fig. 5, D and E ) . solution and (b) the interactions in solution are representative Stoichiometry of Subunits-The ability to immunoprecipiof those seen inuiuo in the nativemembrane. a complex permitted the determination tate these proteins as Immunoprecipitation of Transport Complex-Additional of the stoichiometry of the polypeptides in the transport evidence that the transport proteins exist in a multimeric complex. The complex was immunoprecipitated from cells complex was obtained through an immunoprecipitation tech-grown inminimal medium with [:”S]methionine, and the nique. We reasoned that it shouldbe possible to immunopre- proteins were separated by SDS-polyacrylamide gel electrocipitate the entirecomplex using antibody to just oneof the phoresis. Since strains carrying the overexpressing plasmids proteins in the complex. As shown in Fig. 5A,when membrane pFG23 and pMR11 did not grow in minimal media, strain proteins isolated from the overproducing strain were solubi- HN93 expressing fully induced, wild-type levels of the mal lized with 1% dodecyl maltoside and incubated with antibodygenes was usedin these experiments. The autoradiographs in were specifically Fig. 5, B and C, show that the three proteins precipitated by the antibody against MalF. As shown in Fig. A II PP I P 5 , C and E, it was possible to obtain a greater separation between MalK (upper band) and MalF(lower band) by alter_. ing the pHof the separatinggel from pH8.8 to pH8.6; under these conditions, MalG was lost from the gel. Bands were - 96 excised from the gels, and radioactivity associated with each -96 - 66 polypeptide was determined. Taking into account the number -66 of methionines in each protein based on the known DNA 96 ” -- 4 3 Mal- v ” sequences (Dassa and Hofnung, 1985; Froshauer and BeckMalF MalK* -43 - 66 with, 1984; Gilson et al., 1982), the ratios of proteins in the - 31 MalF immunoprecipitated complex are reported in Table11. Based MalG + on thesize of the complex as judged by gel filtration and the stoichiometry (Table11),the complex must containtwo MalK, one MalF, and one MalG protein; hence the molecular weight of the complex is 170,000. ”

-

I

D

rin

F

I

P

n n K F K F

DISCUSSION

We have succeeded in purifying the threemembrane-associated componentsof the maltose transport system, the MalF, 50 MalG,andMalKproteins,andinreconstituting maltose 50transport activity proteoliposome in vesicles using thepurified components. This task was simplified considerablyby the fact that the three proteins copurified under severaldifferent 33 conditions. The use of the detergentdodecyl maltoside proved 28 to be crucial to thesuccess of the purification. In preliminary FIG. 5. Immunoprecipitation of the maltose transport comstudies, little or no activitycould be recovered from columns plex. Membrane proteins solubilized in dodecyl maltoside were imoctyl glucoside. We suspect that the loss munoprecipitated as described under “Experimental Procedures” run in the detergent using either preimmune serum ( P ) or immune serum ( I ) raised of activity was a result of the dissociation of the complex against MalF. A, immunoprecipitation was from 100 pg of membrane during separation in octyl glucoside. If, in fact, dodecyl malprotein isolated from cells which overproduced the mal proteins. toside is superior in its abilitystabilize to subunit interactions Samples (5 pl) in loading buffer in the absence of reducing agent were within complexes, it may also prove usefulin thepurification separated on 8.5% gels, pH 8.8, and visualized by silver staining. B, of other hydrophobic membrane complexes. proteins were immunoprecipitatedfrom wild-type membranes labeled A basic requirement for reconstitution by dilution is that with [:%]methionine (2 x IOGcpm). Samples were boiled in gel loading the detergent have a high critical micellar concentration. In buffer in the presence of reducing agent prior to separation on an theory, dilution below the critical micellar concentration re8.5% gel, pH 8.8. C, equal volume of sample in B separated on an 8.5% gel, pH 8.6. D,samples (2 pl) from A were separated on an 8.5% sults in the instantaneous conversion of micelle to monomer, gel, pH 8.8, transferred to nitrocellulose membrane, and probed with transiently exposing hydrophobicsurfaces of lipid and protein antibody to MalF ( F ) , MalG ( G ) , or MalK ( K ) as described under to the aqueous environment and thereby promoting thefor“Experimental Procedures.”The portion of the nitrocellulose harboring Immunoglobulin G (visible at top of gel in A ) was removed prior mation of proteoliposomes. Despite the low critical micellar to the blocking step. E, as in D except that separation was on an concentration of dodecyl maltoside, we found that we were 8.5% gel, pH 8.6. able to use essentially the same detergent dilution procedure K F G

K F G

Purification of the Maltose Transport Complex

895 1

for reconstitution of transport from dodecyl maltoside as we raises the possibility that the individual proteins may not absence of a full had described previously for octyl glucoside (Davidson and takeonthesameconformationinthe Nikaido, 1990). Proteins which had been solubilized in 1% complement of domains. Second, the presence of two hydrododecyl maltoside were depleted of excess detergent by puri- philic domainsin the complex may expedite the crystallization fication in buffers containing 0.01% dodecyl maltoside and process. A recent strategy used in the crystallizationof memthen simply mixed with 1% octyl glucoside prior to dilution. brane proteins has been to introduce large hydrophilic comThe addition of high concentrations of octyl glucoside to low ponents by gene fusion to hydrophobic proteins to increase concentrations of other detergentsmay allow a simple recon- the likelihood of crystallization. We hope that by working to stitution by dilution from a range of detergents which previ- increase the yield from our purification scheme, crystallization experiments will become feasible in the near future. ously were not thought to be suitable for this purpose. We haveprovided extensive biochemicalevidence that Acknowledgments-We would like to thank Grace Hsn and John these proteins exist as a multisubunit complex. In addition to Reid of the DNAX Research Institute of Molecular and Cellular the copurification, cross-linking experiments demonstrated Biology for the gift of the syntheticpeptide used in the production of that these proteins lie in close proximity both in the memantibody to MalG. brane and in detergent solution. The position of elution from REFERENCES a gel filtration column also suggests that the proteins are migratingas a multimer. Finally, allthreeproteins were Anderson, D. J., and Blobel, G. (1983) Methods Enzymol. 96, 111immunoprecipitated by antibody prepared against asingle 120 protein, the MalF protein. Theidea that these proteinsform Bavoil, P., Hofnung, M., and Nikaido, H. (1980) J . Biol. Chem. 255, 8366-8369 a complex in the membrane was based primarily on genetic Berger, E. A., and Heppel, L. A. (1974) J. Biol. Chem. 249, 7747data suggesting that MalK was anchored to the membrane 7755 via an interaction with MalG (Shuman and Silhavy, 1981). Bishop, L., Agbayani, R., Jr., Ambudkar, S. V., Maloney, P. C., and This hypothesis was strengthened by the similarities found Ames, G. (1989) Proc. Natl. Acad. Sci. U. S. A. 86,6953-6957 between the maltose system and other homologous systems. Blake, M. S., Johnston, K. H., Russell-Jones, G. J., and Gotschlich, E. C. (1984) Anal. Biochem. 136, 175-179 As described by Hyde et al. (Hyde et al., 1990), each of these systems containsfour domains, two hydrophobic regions with Dassa, E. (1990) Mol. & Gen. Genet. 222, 33-36 approximately six transmembrane segments and two hydro- Dassa, E., and Hofnung, M. (1985) EMBO J. 4, 2287-2293 Davidson, A. L., and Nikaido, H. (1990) J. Biol. Chem. 265, 4254philic regions each witha nucleotide-binding site. In different 4260 systems, the four domains can be encoded by a single gene Dean, D.A., Davidson, A. L., and Nikaido, H. (1989) Proc. Natl. (P-glycoprotein and cystic fibrosis), two genes (white-brown Acad. Sci. U. S. A. 86, 9134-9138 in Drosophila), three genes (ribose and maltose transport in Duplay, P., Bedouelle, H., Fowler, A., Zabin, I., Saurin, W., and Hofnung, M. (1984) J. Biol. Chem. 259, 10606-10613 E. coli), or four genes (oligopeptide transport in S. typhimuFroshauer, S., and Beckwith, J. (1984) J. Biol. Chem. 259, 10896rium) (Hyde et al., 1990). 10903 Based on these homologies, one would predict that a func- Gilson, E., Nikaido, H., and Hofnung, M. (1982) Nucleic Acids Retional maltose transport complex would contain one copy search 10,7449-7458 each of the MalF and MalG proteins and two copies of the Hamada, H., and Tsuruo, T. (1988) J . Biol. Chem. 263, 1454-1458 MalK protein. Using an immunoprecipitation technique,we Heukeshoven, J., and Dernick, R. (1988) Electrophoresis 9, 28-32 have shown that this prediction is true. The presenceof two Higgins, C. F., Hiles, I. D., Whalley, K., and Jamieson, D. J. (1985) EMBO J. 4, 1033-1040 putative sites of ATP hydrolysis throughout the evolution of A. C., Weathenvax, R., and Ames, G . F.-L. (1984) Proc. Natl. these systemssuggests that both are essential for the function Hobson, Acad. Sci. U. S. A. 81, 7333-7337 of these permeases. The role of two sites deserves further Hyde, S. C., Emsley, P., Hartshorn, M. J., Mimmack, M. M., Gileadi, investigation as may it be central to the understanding of the U., Pearch, S. R., Gallagher, M. P., Gill, D. R., Hubbard, R. E., and Higgins, C. F. (1990) Nature 346,362-365 molecular mechanism of this type of permease. Kuchler, K., Sterne,R. E., and Thorner,J. (1989) E M B O J . 8,3973The complex,which was purified togreaterthan90% homogeneity, appears to retainfull maltose transport activity 3984 R., Cunningham, K., Brundage, L. A., Ito, K., Oliver, D., and and exhibits an ATPase activity which is tightly coupled to Lill, Wickner, W. (1989) E M B O J. 8,961-966 maltose transport. Itis therefore quite clear that the maltose Lugtenberg, B., Meijers, J., Peters, R., van der Hoek, P., and van transport complex hydrolyzes ATP to provide energy for Alphen, L. (1975) F E B S Lett. 58, 254-258 active transport. The site of ATP hydrolysis is most likely on Miller, J. H. (1972) Experiments in Molecular Genetics, Cold Spring Harbor Laboratory, p.431, Cold Spring Harbor, NY the MalK polypeptide as its sequence encodes a nucleotide binding site and the homologous domains in other transport Newman, M. J., Foster, D. L., Wilson, T. H., and Kaback, H. R. (1981) J. Biol. Chem. 256, 11804-11808 systems have been labeled with 8-azido-ATP (Higgins et al., Reyes, M., and Shuman, H. A. (1988) J. Bacteriol. 170, 4598-4602 1985; Hobson et al., 1984). Riordan, J. R., Rommens, J. M., Kerem, B.-S., Alon, N., Rozmahel, The mechanismby which the energyreleased by hydrolysis R., Grzelczak, Z., Zielenski, J., Lok, S., Plavsic, N., Chou, J.-L., Drumm,M. L., Iannuzzi, M. C., Collins, F. S., andTsui, L. C. results in maltose accumulation remains a major unanswered (1989) Science 245, 1066-1073 question. The determination of the crystal structure of the Roseveard, P., VanAken, T.,Baxter,J.,and Ferguson-Miller, S. entire complex would be of immense value in answering this (1980) Biochemistry 19, 4108-4115 fundamental question. Thediscovery that these proteinswill Schaffner, W., and Weissmann, C. (1973) Anal. Biochem 56, 502copurify may have several important consequences. First, 514 efforts tocrystallize any of these proteins shouldfocus on the Shuman, H. A., and Silhavy, T. J. (1981) J . Bid. Chem. 256, 560562 crystallization of the entire complex rather on an individual Shuman, H. A., Silhavy, T. J., and Beckwith, J. R. (1980) J . Biol. component. Our early observation that thesolubility properChem. 255, 168-174 ties of the proteins varied depending on whether or not all Towbin, H., Staehelin, T., and Gordon, J. (1979) Proc. Natl. Acad. threeproteins were present(DavidsonandNikaido, 1990) Sci. U. S. A . 76,4350-4354