Reconstitution of an Efficient Protein Translocation ... - Semantic Scholar

Vol. 269, No. 38, Issue of September 23, pp. 23625-23631, 1994 Printed in U.S.A.

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

Reconstitution of an Efficient Protein Translocation Machinery Comprising SecA and theThree MembraneProteins, SecY, SecE, and SecG (plZ)* (Received for publication, May 19, 1994) Mitsuharu Hanada, Ken-ichi Nishiyama, Shoji MizushimaS, and Hajime Tokudal From the Institute of Molecular and Cellular Biosciences, University of Tokyo, 1-1-1Yayoi, Bunkyo-ku, Tokyo113, Japan

A cytoplasmic membraneprotein, p12, of Escherichia coli was discovered as a new factor that stimulates the protein translocation activity reconstituted with SecA, SecY, and SecE(Nishiyama, K., Mizushima, S., and Tokuda, H. (1993)EMBO J. 12,340943415).Direct involvement of p12 in protein translocation was subsequently demonstrated in vivo by genetic studies, and the name SecG has been proposedfor p12 (Nishiyama,K., Hanada, M., and Tokuda, H.(1994) EMBO J. 13, 32724277). To elucidate the role of SecG in protein translocation and to characterize the translocation apparatus comprising these four Sec proteins, the activity of reconstituted proteoliposomes was examined in detail as afunction of the amount of each component. SecG markedly stimulated the translocation activity over wide ranges of amounts of the other threeSec proteins, indicating that none of the other three Sec proteins substitutes for the SecG function. Detailed kinetic analyses indicated that the activity of proteoliposomeswas dependent on the amount of the SecY-SecE complex when SecG was absent and the amount of the SecY-SecE-SecG complex when the proteoliposomes contained SecG. Thetranslocation activity of the latter complex was significantly higher than that of the former one. Binding of SecA to liposomes appreciably increased when they contained both SecY and SecE, whereas the further presence of SecG did not enhance the binding. On the other hand, the ATPase activity of SecA, which was dependent on proOmpA and SecY-SecE-containing proteoliposomes, was significantly enhanced when the proteoliposomes contained SecG. Taken together, these results indicate that SecG enhances the translocation activity of the apparatus after thestep of SecA targeting to SecY.SecE.

The protein translocation machinery of Escherichia coli has been reconstituted from independently purified SecY, SecE, and SecA (1)or the purified SecY.SecE complex and SecA (2). The direct involvement of these three Sec proteins in protein translocation has thus been established. The SecY.SecE complex, whether isolated chromatographically (2) orimmunoprecipitated with an anti-SecY antibody (31, contained another protein, termed band 1,however. The activity of the translocationmachinery reconstituted from SecY, SecE, and SecA is significantly lower than that of everted membrane vesicles,

* This work was supported by grants from the Ministry of Education,

Science, and Culture of Japan. 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. i Present address: Tokyo College of Pharmacy, 1432-1 Horinouchi, Hachiohji-city 192-03, Japan. 8 To whom correspondence should be addressed. Tel.:81-3-5684-0086; Fax: 81-3-3818-9435.

whereas the translocation machinery reconstituted from the SecY.SecE.band 1 complex and SecA has been reported to be comparable to that of membrane vesicles (4). A novel membrane protein, p12, has been found to remarkably stimulate the translocation activity of proteoliposomes when it was reconstituted into the proteoliposomes together with SecY and SecE (5). The gene encoding p12 has been mapped at 69 min on the E. coli chromosome (5). Furthermore, the disruption of this gene has been found to cause the accumulation of precursors of maltose-bindingprotein, p-lactamase, and OmpA (6). These observations indicate thatp12 is a general component of the translocation machinery. Based on these observations, we proposed that p12 be named SecG (6). SecG was recently found to be identical to band 1.’ These results, taken together, indicate that SecA and the three membrane components, SecY, SecE, and SecG, constitute the fundamentalunit of the proteintranslocationmachinery in E. coli. It has been reported that the protein translocation machinery, the SecGlp complex, in the membrane of the mammalian endoplasmic reticulum also comprises three protein subunits (a,p, and y subunits) (7). The amino acid sequences of the a and y subunits exhibit some similarity to those of SecY and SecE, respectively (8, 9). Although the similaritybetween the p subunit and SecG has not been clarified, these observations indicate theevolutionary conservationof components of the protein translocation machinery from prokaryotes to mammals. Our previous reconstitution studies involving SecA, SecY, and SecE revealed that theactivity of proteoliposomes did not exhibit saturation even with an excess amount of SecE (l), suggesting that SecE is present in excess over SecY in functional stoichiometry. These observations appeared to support the proposal by Bieker and Silhavy (lo), who assumed SecE functioned as a shuttle between SecA and SecY in the protein translocation pathway. On the other hand, theisolation of the SecY.SecE.band 1(SecG) complex suggests that theseproteins functiontogether. Our previousreconstitution studies were carried out withoutSecG, which plays an important role in an efficient protein translocation reaction.More detailed and careful reconstitution studies are, therefore, required. In this study, the activity of the translocation machinery reconstituted from the four Sec proteins and phospholipids was examined as a function of the amount of each component. EXPERIMENTAL PROCEDURES coli W3110 M25 (ompT) was transformed with pMAN809 and pMAN510 for the overproduction of SecE and SecY (11). E. coli KO02 (Lpp-) was transformed with pTG1, which carries the tac-secG gene, for the overproduction of SecG. The Bacterial Strains and Plasmids-E.

Douville, K., Leonard, M.,Brundage, L., Nishiyama,H., Tokuda, H., Mizushima, S., andWickner, W. (1994) J. Biol. Chem. 269, 1870518707.

23625

23626

Reconstitution of an Efficient Protein Danslocation Machinery

construction of this plasmid will be reported elsewhere. The overproduction was induced by isopropyl-P-D-thiogalactopyranoside. Materials-Octyl glucoside' was purchased from Dojindo Laboratories. ATP and creatine kinase were from Boehringer Mannheim. Proteinase K was fromMerck. Creatine phosphate wasfrom Sigma. T~-an~~S-label mixture (a of 70% [35Slmethionineand 20% [35Slcysteine, 1000 Cilmmol) was obtained from ICN. NBD-PE was from Molecular Probes, Inc. Preparation of SecA, SecG, and proOmpA-SecA (12),SecG (51,and proOmpA (13)were purified from cells overproducing the respective proteins as described. Purification of SecY and SecE-The method ofAkimaru et al. (1)was modified. Cytoplasmic membrane fractions prepared from SecY.SecEoverproducing cells were sequentially washed with urea and cholate prior t o solubilization.Washing with 5 M urea was carried out on ice for 10 min a t 5 mg of proteinfml in 50 m~ potassium phosphate (pH 7.5). The urea-washed membranes were precipitated by centrifugation at 230,000 x g for 1 h and then treated with 6% (w/v) sodium cholatecontaining 50 mM potassium phosphate (pH 7.5) onicefor 20 min. Approximately 50%of the proteins were removed fromthe membranes by these washings, whereas neither SecY nor SecE was extracted. For the purification ofSecY, the uredcholate-washed membranes were solubilizedat 5 mg of proteinfml on ice for20 min with 2.5% (w/v) octyl glucoside-containing 50 mM potassium phosphate (pH 6.951, 150 mM NaC1, 10 ~ MgSO,, l l ~ 10% glycerol, and 10 mgt'ml E. coli phospholipids. External addition of phospholipids was essential for SecY solubilization. The supernatant (10 ml) obtained on centrifugation at 230,000x g for 1 h was applied to a cation exchange column (Monos, 1 x 10 cm; Pharmacia Biotech Inc.) that had been equilibrated with 2.5% octyl glucoside-containing 50 mM potassium phosphate (pH 6.95). 150 nm NaC1, 10% glycerol. The column was washed with 60 ml of the same buffer at a flow rate of 4 mltmin. A lineargradient of NaCl(O.15-0.4 M, 20 ml) in the same buffer was then applied to elute SecY. SecY-containing fractions were combined,concentrated by membrane filtration, and then further purified by size exclusion chromatography on a Superose 12 column (1x 30 cm; Pharmacia) that had been equilibrated with 2.5% octyl glucoside-containing 50 mM potassium phosphate (pH 7.5), m 150 M NaCI, 10% glycerol. The column was developedwith the same buffer at a flow rate of 0.4 mumin. The fractions containing SecY at a purity of >80% were combined and kept frozen at -80 "C until use. SecE was extracted from the uredcholate-washed membranes at 5 mg/ml with 10 mM Tris-C1 (pH 7.5) containing 10 mM MgSO,, 10% glycerol, and 2.5% octyl glucoside on ice for 20 min. SecE solubilization did not require the external addition of phospholipids. The supernatant (10 ml) obtained upon centrifugation was applied to an anion exchange column (Mono&,1x 10 cm; Pharmacia) that had been equilibrated with 2.5% octyl glucoside-containing 10 mM Tris-C1 (pH 7.5) and 10% glycerol. SecE was recovered in the pass-through fractions, whereas many other proteins were adsorbed to the column. The pass-through fractions were combined,concentrated, and thenapplied to a desalting column to change the buffer for the next column. The fractions containing SecE were then applied t o a Monos column equilibrated with 10 mM potassium phosphate (pH 6.0) containing 10% glycerol and 2.5% octyl glucoside. The column was developed with a linear gradient of NaCl(0-0.4 M, 40 ml) in the same buffer. The fractions containing SecE were combined, concentrated, and then further purified by size exclusion chromatography as described for the purification of SecY. SecE (purity of >go%) was kept frozen at -80 "C until use. Protein Tkanslocation into Reconstituted Proteoliposomes-The indicated amounts of purified SecY, SecE, and SecG were mixed with 1.25 mgof E. coli phospholipids in a total volume of 85 pl, followedby reconstitution into proteoliposomes as previously described 11). The reconstituted proteoliposomes were suspended in 105 pl of SO mM potassium phosphate (pH 7.5), 150 mM NaCl and then subjected to the translocation assay. Protein translocation activity was measured at 37 "C in thepresence of 60 pg/ml SecA, 2 mM ATP, 2 mM MgSO,, and an ATP-regenerating system consisting of 10 m creatine phosphate and 0.25mg/ml creatine kinase, as described (1).The assay mixture (150 pl) contained 90 pl of reconstituted proteoliposomes. The 35S-labeled proOmpA D26, which is a derivative of proOmpA lacking about 250 amino acidresidues at itsC terminus, was used as a substrate (14).At the indicated times, aliquots (25pl) of the reaction mixture were withdrawn and treatedwith proteinase K on ice for30 min as described (1).

The translocated protein was detected as a proteinase K-resistant band after SDS-PAGE by fluorography. The amount of translocated protein was determined by densitometric scanning of the fluorograms with a Shimadzu CS-930chromatoscanner. The initial rates of translocation were determined from at least four different time points. SecA Binding to Proteoliposomes-Proteoliposomes were reconstituted as described above except that 125 pg of a phospholipid mixture was used instead of 1.25 mg of E. coli phospholipids. The phospholipid mixture contained E. coli phospholipids and NBD-PE, a fluorescent analogue of phosphatidylethanolamine, at a weight ratio of 49:l. The amount of phospholipids in proteoliposomes was determined from the fluorescenceintensity of NBD-PE (excitation at 480 nmand emission at 525 nm) with a Shimadzu RF-5000 spectrofluorophotometer. The assay mixture comprised in 75 pl, the indicated amount of proteoliposomesor liposomes, 50 mM potassium phosphate (pH 7.5), 150 mM NaC1, 2 mM ATP, 2 mM MgSO,, and 60 pgt'ml(O.6 w) SecA. After a 10-min incubation a t 30 "C, the assay mixture was centrifuged at 170,000x g for 1 h. The precipitates were analyzed by SDS-PAGE, followed by Coomassie Brilliant Blue staining. The assay was alsocarried out in the absence of phospholipids for correction of the nonspecifically precipitated SecA. The amounts of precipitated SecA and SecY were determined by densitometric scanning of the Coomassie Brilliant Blue-stained gels using purified SecA and SecY as standards, respectively. ATPase Activity of Sed-SecY (0.45 nmol) and SecE (1.44 nmol), with or without SecG (0.9 nmol), were mixed with 250 pg of E. coli phospholipids, followed by reconstitution into proteoliposomes. The reconstituted proteoliposomes were resuspended in 210 p1of 50 mM potassium phosphate (pH 7.5)containing 150 nm NaC1. ATPase activity was determined by a coupled spectrophotometricassay involving pyruvate kinase and lactate dehydrogenase as described (15).The cuvette contained, in 2 ml, 50 mM potassium phosphate (pH 7.5), 150 mM NaC1, 2 mM MgSO,, 3 nm phosphoenolpyruvate,0.25mM NADH, 1 mM ATP, 10 units of pyruvate kinase, 15 units of lactate dehydrogenase, 20 pdml SecB, and 200 pl of a proteoliposome suspension at 37 "C. The assay was started by the addition of SecA(2O pg). ProOmpA wassubsequently added at a final concentration of 0.45 PM. Oxidation ofNADH was continuously monitored at 340 nm with a Shimadzu UV-3000spectrophotometer. The amount of ATP hydrolyzed was calculated using a value of 6.22for the millimolar absorption coefficient of NADH. Other Methods-E.coli phospholipids were prepared as described (16).Phospholipids wereanalyzed by thin layer chromatography (Silica GEL 60, Merck). SDS-PAGE was carried out using a gel composed of 12.5% acrylamide and 0.27% N,N"methylenebisacrylamide, as described by Hussain et al. (17),for the analysis of proOmpA D26, or 12.5% acrylamide and 0.33% N,N"methylenebisacrylamide, as described by Laemmli (181,forSecA,SecY, SecE, and SecG. The labeled proOmpA D26 was synthesized in vitro in the presence of Tran'%-label (0.46mCVml) as described (19)and was partially purified as reported (20).Protein was determined by the method of Lowry et al. (21)using bovine serum albumin as a standard.

'%-

RESULTS

SecE and SecY Function Together-We previously showed that theactivity of proteoliposomes reconstituted with various amounts of SecY and a fured amount of SecE became maximum at a certain amount of SecY, whereas the activity of ones reconstituted with various amounts of SecE and a fixed amount of SecYdid notlevel off even with anexcess amount of SecE (1). The SecE preparation used in these experiments however, was, found t o still contain a considerable amount of phospholipids, which had been added together with octyl glucoside to solubilize the membranes(14).It was also found that the membrane filtration method employed t o concentrate SecY resulted in the simultaneous concentration of octyl glucoside. These factors might have interfered with the accurate determinationof the activity in relation to the amounts of these components. As described under "Experimental Procedures," the purification methods for SecY and SecE were improved to overcome these problems. SecY and SecE purified by the improved methods did not contain detectable amounts of phospholipids, as judged by thin layer chromatography. Moreover, the concentration of OCThe abbreviations used are: octyl glucoside, n-octyl-p-D-glucopyran- tyl glucoside in the final preparation was readjusted by means oside; NBD-PE, N-~7-nitrobenz-2-oxa-1,3-diazol-4-yl~-l,2-dihexadeof gel filtration chromatography. The following reconstitution canoyl-sn-glycero-3-phosphoethanolamine; PAGE, polyacrylamidegel studies were carried out using these preparations. The initial electrophoresis.

Reconstitution of a n Efficient Protein

E51

2 Y

nanslocation Machinery

23627

P

A C

0 4-

0

0

100

200

300 500 400

600

SecE (pmol)

100

FIG.2. Protein translocation activity reconstituted with fixed amounts of SecY and various amounts of SecE. Proteoliposomes were reconstituted with SecE and SecY as described in the legend to Fig. 1. The amountof SecY was fixed at 5 (O), 10 (O),or 20 (A)pmol, and the amount of SecE was varied as indicated. The initial ratesof translocation of proOmpA D26 into the proteoliposomes were determined as described in the legend to Fig. 1 and then plotted as a functionof the amount of SecE.

tB

0

50

100

Fixed amount of SecE (pmol)

FIG.1. Protein translocation activity reconstituted with fixed amounts of SecE and various amounts of SecY. Proteoliposomes were reconstituted with SecE and SecY as described under "Experimental Procedures." The amount of SecE wasfixed at 15 ( 0 ) , 5 0 (01,75 (A), or 120 (Ajpmol, and the amountof SecY was varied as indicated. The translocation of 35S-labeledproOmpA D26 was assayed in the presence of 60 pg/ml SecA and 2 mM ATP. The relative amountof proOmpA D26, which was translocated intoproteoliposomes contained in25-pl aliquots of the reaction mixture, was determineda t various timesby taking the amount of the input proOmpA D26 as 100%. A, the initial ratesof the translocation thus determined were plotteda as function of the amount of SecY for each fixed amount of SecE. B , the amountof SecY required for the half-maximal stimulation of the activity (KvJ was determinedfor each amountof SecE from an Eadie-Hofstee plotof the datashown in A and then plotted as a function of the fixed amount of SecE.

rate of translocation vanes to some extent depending on the precursor preparation. Therefore, only the activities determined with the same precursor preparation on the same day were directly compared. The translocation activity of proteoliposomes reconstituted with a fixed amount of SecE and various amounts ofSecY became maximum when a n almost equimolar amount of SecY and SecE was added, irrespective of the fixed amount of SecE (Fig. lA). The Kv,value for SecY, which represents the amount of SecY required for half-maximal stimulation of the activity, was roughly proportionalto the fixed amount of SecE (Fig. 1B1, indicating that the amount of SecY required for the maximum activity is dependent on the fixed amount of SecE. When reconstitution was carried out with a fmed amount of SecY and various amounts of SecE, the amountof SecE required to maximize the activity exceeded that of SecY (Fig. 2). However, the amount of SecE required was also dependent on that of SecY. When the amountof SecY was fured at 5 or 10 pmol, the amount of SecE required to maximize the activity was about 20-fold that of SecY. The activity reconstituted with 20 pmol of SecY did not level off with 300 pmol of SecE. Taken together, these results most likely indicate that SecY and SecE function to-

gether as a SecY6ecE complex and that thereconstituted activity represents the amountof the SecY.SecE complex, which is in equilibrium with dissociated SecY and SecE, in proteoliposomes. When the activity becomes maximum, most of the functionally reconstituted molecules of one component, the amount of which is fixed, are assumed to become part of the complex upon the addition of the othercomponent. The amount of functional SecY molecules in proteoliposomes reconstituted with equal amounts of SecY and SecE seems to be, therefore, sufficiently high for most SecE molecules to become part of the complex (Fig. 1).The amount of functional SecE molecules in proteoliposomes is assumedto reach a level that isrequired for most SecY molecules to become part of the complex when an amount of SecE more than 20-fold that of SecYis added (Fig.2). The significant difference in theSecE:SecY molar ratio for maximum activity observed in Figs. 1 and 2 seems tobe due to the fact either that thefraction of functionally reconstituted molecules is smaller with SecE than with SecY or that the SecE: SecY stoichiometry of the complex is more than 1or both. It is noteworthy that the functional amount ofSecY or SecE in proteoliposomes is proportional to the amountof it added since the reconstituted activity increased proportionally with an increase in the amount of the component added until the activity became maximum. The recoveries ofSecY (80%) and SecE (90%) in proteoliposomes were constant, irrespective of the amount added. SecG Stimulates the Activity Even in the Presence of Large Amounts of SecY and SecE-SecG (p12) was identified as a factor that stimulates the translocation activity of proteoliposomes reconstituted from fixed amounts of SecY and SecE ( 5 ) . The translocation activityuponreconstitution from various amounts of SecY and SecE was determined in thepresence and absence of SecG. Analysis of the reconstituted proteoliposomes by SDS-PAGE revealed that SecG did not affect the recoveries of SecY and SecE under the conditions employed in this study (data not shown). The activity of proteoliposomes reconstituted from 50 pmol of SecE and various amounts of SecY, with or without 150pmol of SecG, was examined (Fig. 3). SecY was essential irrespective of the presence or absence of SecG, as observed previously ( 5 ) . SecG significantly enhanced the reconstituted activity over the entire rangeof amounts of SecY examined. Eadie-Hofstee plots of the data revealed that SecG caused an approximately 5-fold

Reconstitution of an Efficient Protein DanslocationMachinery

A

L

0

-

0

-

-%cG

0

20

40

60

80

SecY (pmol) FIG.3. SecG-dependent stimulationof the translocation activity in the presenceof various amounts of SecY. Proteoliposomes were reconstituted from 50 pmol of SecE and various amounts of SecY with (0)or without (0) 150 pmol of SecG. The initial rates of translocation of proOmpA D26 into the proteoliposomes were determined as described in the legend t o Fig. 1 and then plotted as a function of the amount of SecY.

0

100

200

300

400

500

600

SecE (pmol)

LB

SecG

K,,

Vm

increase in theV,, value, whereas theK Y value ~ for SecY was not significantly affected. When proteoliposomes were reconstituted from 10 pmol of SecY, more thana stoichiometric amount of SecE as compared +SecG with SecY was required for the maximum activity even in the presence of 150 pmol of SecG (Fig. 4A). The K Y value ~ for SecE decreased, however, about %fold upon reconstitution withSecG suggesting that SecG increases the affinity of SecE (Fig. a), for complex formation. Furthermore, SecG significantly stimulatedthetranslocationactivity over theentirerange of amounts of SecE examined (Fig. 4A). SecG increased the V, value by a factor of about 4 (Fig. 4B). 0 10 20 30 40 50 60 70 80 90 The SecG-dependent stimulation of the translocationactivity Activity/SecE (O/omin/nmol) over a wide range of amounts of SecY (Fig. 3) or SecE (Fig. 4 ) indicates that SecG has a function different fromthose of SecY FIG.4. SecG decreases the amountof SecE required to maxiProteoliposomes were reconstituted machin- mize the translocation activity. and SecE and that it renders the protein translocation from 10 pmol of SecY and various amountsof SecE with (0)or without ery moreefficient. (0) 150 pmol of SecG. A, the initial rates of translocation of proOmpA Both SecY and SecE, but Not SecG, Are Required for SecA D26 into theproteoliposomes were determined as described in thelegBinding-Proteoliposomes were reconstituted from SecY and end to Fig. 1.B , the data shown in A were plotted accordingto EadieSecE, with or without SecG, and then assayed for translocation Hofstee. TheK& (pmol) value for SecE and theV, value (V,) are also in the presence of various concentrations of SecA (Fig. 5A). indicated. Although proteoliposomes containing SecG exhibited higher translocation activity than those without SecG over the entire dition of 0.6 p~ SecA to liposomes caused about 0.1 p~ to be SecA concentration range examined, both types of proteolipo- bound, leaving about 0.5 p~ unbound. These values were es~ for SecA (Fig. sentially the sameas those reportedby Hendrick andWickner somes exhibited essentially the same K Yvalue 5B 1, suggesting thatSecG does not affect the affinity of SecA for (23) for the SecA binding tophospholipids. Whenreconstitution was carried out with the normal amount of phospholipids, the protein translocation. I t has been shown that the binding of SecA to membranes amount of SecA bound to proteoliposomes containing SecY and was inhibited by an anti-SecY antibody (221, suggesting that SecE was only marginally higherthan thatbound to liposomes, SecY constitutes a part of the receptor for SecA. However, the irrespective of the presence or absenceof SecG (Fig. 6A).On the 10-fold interaction of SecA with SeeY andlor SecE has not been directly other hand, proteoliposomes reconstitutedwiththe demonstrated. Since SecA binds with high capacity to lipo- lower amount of phospholipids bound an appreciably higher somes of E. coli phospholipids (231, the effect of phospholipids amount of SecA than liposomes (Fig. 6B). However, SecG did on the detection of SecY-dependent SecA binding was exam- not affect the amount of SecA bound. Although ATP has been reported to decrease the binding of SecA to phospholipids (24), ined. Proteoliposomeswere reconstitutedwiththenormal of SecA was not appreciably amount and one-tenthof the normal amountof phospholipids the SecY4ecE-dependent binding affected by the omission of ATP (data not shown). and then assayed for SecA binding in the presence ofATP, which has been shown to decrease the SecA binding to phosWe further examined whether SecA binding requires both of SecY and SecE or only one of them. Proteoliposomes were repholipids (24). Proteoliposomes containing the same amount for SecA phospholipids were used for each assay (Fig. 6 , A and B 1. SecA constituted withSecY, SecE, or both and then assayed bound to liposomes or proteoliposomes was recovered by ultra- binding (Fig. 6C). The amount of SecA bound to SecYSecEthat bound to lipocentrifugation and then analyzed by SDS-PAGE, followed by containing proteoliposomes was larger than Coomassie Brilliant Blue staining. The amount of SecA was somes, whereas the amountof SecA bound to proteoliposomes then determinedby densitometric scanningof the gel. The ad- containing SecY or SecE alone was similar to that bound to

2

Reconstitution of an Efficient Protein Danslocation Machinery

23629

sence of SecG. These results indicate that the SecY-SecE-dependent binding of SecA detected with the proteoliposomes reconstituted with the reduced amount of phospholipids is functional. From the amount of SecYin SecY-SecE-containingproteoliposomes reconstituted with the reduced amount of phospholipids andthat ofSecA bound to thep~teoliposomes,the SecASecY stoichiometry of the SecA binding was estimated. Although the estimated stoichiometry was rather variable, depending on the experiment, a value of 2 * l was obtained as an average of more than 10 determinations. The SecY content of the SecY.SecE-containingproteoliposomes reconstituted with the reduced amount of phospholipids was 1.1%(wlw), which was considerably higher than the value (0.05-0.1%)reported for membrane vesicles (25). Hart1 et al. (22) reported that the number of high affinity binding sites for S e d , which is as10 20 30 40 50 60 70 sumed to be dependent on SecY and/or SecE, is about 0.1 nmol SecA (pglml) per mg ofmembrane protein or per mg ofmembrane lipid, since the amounts of protein and phospholipid in membranes are similar. The SecYeSecE proteoliposomes bound 1.4nmol of SecAimg of phospholipids (Fig. 6C). This value seems reasonSecG KTE Vm able since the amount of SecY in these proteoliposomes was - 3.1 0.50 10-20-fold that in membrane vesicles. A smaller amount (0.3 + 4.8 1.95 nmol/mg of phospholipids) of SecA bound was estimated from the results shown in Fig. 6B,however. SecG Enhances the ProOmpA-dependent ATPase Activity of SecA-The ATPase activity of SecAincreases upon the addition of proOmpA in the presence of membrane vesicles (26). This proOmpA-dependent ATPase activity, called translocation ATPase activity (261, is significantly higher with membrane vesicles containing overproduced SecY-SecEthan with normal membrane vesicles (15).The effect of SecY4ecE with or without SecG on the proOmpA-dependent ATPaseactivity of SecA was examined in proteoliposomes reconstituted with the re0.0 0.1 0.2 0.3 0.4 0.5 duced amount of phospholipids. ATPaseactivity stimulated by ActivityISecA (%/minlpglml) the addition of proOmpA was slightly higher with proteolipoFIG.5. SecA dependence of the translocation activity of pro- somes containing SecY and SecE than with liposomes (Fig. 8). teoliposomes withor without SecG. Proteoliposomes werereconsti- Reconstitution of SecG with SecY and SecE significantly entuted from 25 pmol of SecY and 80 pmol of SecE with (Ofor without (0) hanced the proOmpA-dependent ATPase activity. These results 300 pmol of SecG and then assayed for the translocation of proOmpA are consistent with SecG-dependent stimulation of the trans026 into the proteoliposomes in the presence of various concentrations of S e d . A, the initial rates of translocation were determined under each location activity. Taken together, these results indicate that condition and thenplotted as a function of the concentration of SecA.B , SecG stimulates the protein translocation reaction after SecA the Kv,value (pg/ml)for SecA and the V, value CV,) were determined binds to the machinery. I

.

I

.

0

1

.

: i

by an Eadie-Hofstee plot o f the data shown in A.

DISCUSSION

liposomes. The binding of SecA, therefore, required both SecY and SecE, most likely suggesting that SecY and SecE form a complex and then serve as a SecA receptor. The proteoliposomes used in the experiments shown in Fig. 6 (A and B ) were analyzed by SDS-PAGE, followed byCoomassie Brilliant Blue staining. The amounts of SecY, SecE, and SecG recovered in the proteoliposomes were then densitometrically determined using the respective purified proteins as standards. Reconstitution with the 10-fold loweramount of phospholipids was found to decrease the amount of SecY recovered in proteoliposomes by a factor of 2-3, whereas the recoveries of SecE and SecG were only marginally affected. Proteoliposomes reconstituted with the reduced amount of phospholipids were, however, found to contain an approximately 4-fold higher amount of SecY per unit of phospholipids than ones reconstituted withthe normal amount of phospholipids. The translocation activity of these proteoliposomes was examined with proOmpA D26 as a substrate, and theinitial rates of the translocation per unit of SecY in proteoliposomes were estimated (Fig. 7). The activity reconstituted with the reduced amount of phospholipids did not differ significantly from that reconstituted with the normal amount of phospholipids, irrespective of the presence or ab-

It has been proposed onthe basis of genetic studies that SecE functions as a shuttle between SecA and SecY (10). Our previous reconstitution studies indicating the requirement of an excess amount of SecE as compared with SecY for the maximum activity appeared to support this proposal (1).On the other hand, more critical and detailed reconstitution studies described in this paper revealed that the amount of one component, which was fixed, determines the amount of the other component required to maximize the activity (Figs. 1 and 2), indicating that SecY and SecE function together, most likely as a SecY.SecE complex. The requirement of both SecY and SecE for the binding of SecA also supports this conclusion. Furthermore, in the presence of a fixed amount of one component, the reconstituted activity increased linearly with an increase in the amount of the othercomponent, i.e. either SecY (Fig. I)or SecE (Fig. 21, until the activity became maximum. This simple relationship indicates that theactivity reconstituted with SecY and SecE is determined by the amount of a single factor, the SecY.SecE complex. These results also suggest thatthe SecY.SecE complex has a definite subunit stoichiometry. However, the subunit stoichiometry of the complex has not been clarified yet.

Reconstitution of an Efficient Protein Panslocation Machinery

23630

15

B

IC 10 -

5 -

SecYISecE

SecG

-

-

0

+

+

-

+

-

+

+

SecE

-

+

SecY

-

-

+ +

+

-

+

FIG.6. Both SecY and SecE arerequired for SecA binding. Proteoliposomes were reconstituted with either 1.25 mg( A )or 125 pg ( B and C ) of phospholipids containing NBD-PE and the indicated combinations of SecY (100pmol), SecE (320 pmol), and SecG (200 pmol). The amount of phospholipids in theproteoliposomes was determinedfrom the fluorescence intensity of NBD-PE. Binding of SecA was examined as described under “Experimental Procedures” using proteoliposomes containing 5.63 ( A and B ) or 3.75 (C)pg of phospholipids. Liposomes containing the same amount of phospholipids were also assayedas a control.

SecG

Phospholipids

-u+ normal

+

1MQ

SecEISecY SecG

-

+ -

+ +

FIG.7. Protein translocation activity of proteoliposomes having different SecY/phospholipidratios. The proteoliposomes used in Fig. 6( A and B ) were assayedfor the translocationof proOmpA D26. The initial rates of translocation were determined and corrected for the amount of SecY in therespective proteoliposomes. The amountof SecY in the proteoliposomes was densitometrically determined as described under “Experimental Procedures.”

FIG.8. SecG enhances the translocationATPase activity. SecY (0.45 nmol) and SecE (1.44nmol) withor without SecG (0.9nmol) were

SecG stimulated the activity even in the presence of large amounts of SecY (Fig. 31, SecE (Fig. 41, and SecA (Fig. 51, indicating that none of these Sec proteins can substitute for the SecG function. Stimulation of the activity by SecG is most likely caused by the formation of a SecY.SecE.SecG complex, which possesses significantly higher activity than does the SecY.SecE complex. SecG not only stimulated the activity but also increased the affinity of SecE for complex formation (Fig. 4). The SecY-SecE.SecG (band 1) complex has beenisolated from membraneextractsandreconstitutedinto proteoliposomes (2). These proteoliposomes exhibit appreciable translocation ATPase activity (2). We found that SecG plays a role in the stimulation of this ATPase activity (Fig. 8). It has been shown that thesolubilized SecY.SecE.SecG complex easily dissociates into subunits, causing a significant decrease in the reconstituted activity (3). These observations and the results shown inthis paper, taken together, indicate thatthe SecY.SecE.SecG complex represents the fundamental unit of the translocation machinery in membranes.

The addition of a large amountof SecG as well as of SecE was required to maximize the activity reconstituted with a fixed amount of SecY or the SecY.SecE.SecG complex. On the other hand, E. coli cells contain SecY, SecE (25), and SecG3in similar molecular amounts. It seems possible, therefore, that some other factors are involved in theformation and stabilizationof the complex in uiuo. These factors, if present, may enhance the reconstituted translocation activity by facilitating complex formation. It hasbeen reported that neither SecD nor SecF has a n appreciable effect on the activity obtained upon reconstitution with SecY and SecE (27). It would be interesting toexamine the effects of SecD and SecF on the reconstituted activity in the presence of SecG. SecA binding was found to require both SecE and SecY, but not SecG (Fig. 6).This most likely indicates that SecA is targetedtothemachinerythrough its recognition of the SecY.SecE portion of the SecY.SecE.SecG complex. Since a con-

mixed with 250 pg of E. coli phospholipids, followed by reconstitution into proteoliposomes. Thereconstituted proteoliposomes were suspended in 210 p1 of 50 mM potassium phosphate (pH 7.5) containing 150 mM NaCl and then assayedfor ATPase activity. Aliquots (200pl) of the proteoliposome suspensions were usedfor each assay as described under “Experimental Procedures.”

K. Nishiyama and H. Tokuda, unpublished observation.

Reconstitution of an Efficient Protein

TFanslocation Machinery

23631

siderable amountof SecA bound to liposomes, as reported (231, ratory for critical reading of this paper, Miyuki Yamanakafor technical reconstitution with the 10-fold lower amount of phospholipids support, and Akiko Ishii for secretarial support. was required to detect theSecY.SecE-dependent SecA binding REFERENCES (Fig. 6). S e d has been reported to interact with acidic phosAkimaru, J., Matsuyama, S., Tokuda, H., and Mizushima,S. (1991)Proc. Natl. 1. pholipids (28) and to penetrate the lipid bilayer (29). These Acad. Sci. U. S. A . 88, 6545-6549 steps are assumed to initiate the insertion of a precursor pro2. Brundage, L., Hendrick, J. P., Schiebel, E., Driessen, A. J. M., and Wickner, W. tein together withSecA into the membrane in ATP-dependan (1990) Cell 62,649-657 3. Brundage, L., Fimmel, C. J., Mizushima, S., and Wickner, W. (1992) J. Biol. ent manner (29, 30). Interaction of SecA with SecY.SecE has Chem. 267,41664170 been speculated to cause the translocation of the precursor 4. Bassilana, M., and Wickner, W. (1993) Biochemistry 32,262G2630 protein acrossthe membrane(30). Both ATP hydrolysis and the 5. Nishiyama, K., Mizushima, S., and Tokuda, H. (1993)EMBO J. 12,3409-3415 6. Nishiyama, K., Hanada, M., and Tokuda, H. (1994) EMBO J. 13,32723277 proton motive force drive the translocation (20, 30). The 7. Giirlich, D., and Rapoport, T.A. (1993) Cell 75, 615430 SecY.SecE-dependent binding of SecA demonstrated here is 8. Giirlich, D., Prehn, S., Hartmann, E., Kalies, K. U., and Rapoport, T. A. (1992) Cell 71, 489503 likely to represent this step of the protein translocation path9. Hartmann, E., Sommer, T., Prehn, S., Gorlich, D., Jentsch, S., and Rapoport, way. It is not clear, however, whether or not any functional or T. A. (1994)Nature 367, 654-657 structural difference exists between lipid-bound and 10. Bieker, K. L., and Silhavy, T. J. (1990) Cell 61,833442 SecY.SecE-bound SecA. Attempts to distinguish them by ki- 11. Matsuyama, S., Akimaru, J., and Mizushima, S . (1990) FEBS Lett. 269, 9 G 100 netic examinationwere unsuccessful because the amountof the 12. Akita, M., Sasaki, S., Matsuyama, S., and Mizushima,S . (1990) J . Biol. Chem. latter was not sufficiently high for detailed analysis. In any 265,8164-8169 case, it is important to clarify the difference in function or 13. Tani, R , Tokuda, H., and Mizushima, S.(1990) J . Biol. Chem. 265, 1734117347 structure between these two types of SecA. Kim and Oliver (31) 14. 'Ibkuda, H., Akimaru, J., Matsuyama, S., Nishiyama, K., and Mizushima, S . have reported that more than 50-fold overproduction of SecY (1991)FEES Lett. 279,233-236 and SecE caused only a 50% increase in the number of high 15. Kawasaki, S., Mizushima, S., and Tokuda, H. (1993) J. Biol. Chem. 268, 8193-8198 affinity SecA-binding sites in membrane vesicles, suggesting 16. Tokuda, H., Shiozuka, K., and Mizushima, S.(1990) Eur J. Biochem. 192, that SecY and SecE are insufficient to constitute the SecA 583-589 receptor. Although the results shown here indicate the impor- 17. Hussain, M., Ichihara, S., and Mizushima, S.(1980)J . Biol. Chem. 255,37073712 tance of both SecY and SecE for the SecA targeting, the par- 18. Laemmli, U. IC (1970) Nature 227, 680-685 ticipation of some unknown components in this step cannotbe 19. Yamane, K., Ichihara, S., and Mizushima, S . (1987)J. Biol. Chem.262, 23582362 completely excluded. 20. Tani, K., Shiozuka, K., Tokuda, H., and Mizushima, S. (1989) J. Biol. Chem. SecG had little effect on the SecY.SecE/SecA interaction, 264, 18582-18588 whereas it significantly stimulated the translocation ATPase 21. Lowry, 0. H., Rosebrough, N. J., Fan; A. L., and Randall, R. J. (1951) J. Biol. Chem. 193,265-275 activity as well as the proOmpA D26 translocation. Taken to22. Hartl, F.-U., Lecker, S., Schiebel, E., Hendrick, J. P., and Wickner, W. (1990) gether, these results indicate that SecG plays a role in protein Cell 63,269-279 translocation after the SecA.precursor complex has been tar- 23. Hendrick, J. P., and Wickner, W. (1991) J. Biol. Chem. 266, 2459G24600 24. Breukink, E., Demel, R. A,, De Korte-Kool, G . , and De Kruijff, B. (1992) geted to the SecY.SecE.SecG complex. Biochemistry 31, 1119-1124 The conditions for the reconstitution of an efficient translo- 25. Mizushima, S.,Tokuda, H., and Matsuyama,S. (1992) in Membrane Biogenesis and Protein Targeting (Neupert, W., and Lill, R., eds) pp. 21-32, Elsevier cation machinery withSecA and the three membrane proteins, Science Publishers B. V.. Amsterdam SecY, SecE, and SecG, were thus established. On the other 26. Lill, R., Cunningham, K., Brundage, L. A,, Ito, K., Oliver, D., and Wickner, W. (1989)EMBO J . 8,961-966 hand, the exact function of each membrane subunit and the S., Fujita, Y.,Sagara, K., and Mizushima, S. (1992) Biochim. subunit stoichiometry of the complex remain to be clarified. 27. Matsuyama, Biophys. Acta 1122, 77-84 The role of the proton motive force in relation to the SecG 28. Lill, R., Dowhan, W., and Wickner, W. (1990) Cell 60,271-280 function is currently underinvestigation in everted membrane 29. Ulbrandt, N. D., London, E., and Oliver, D.B. (1992) J . Biol. Chem. 267, 15184-15192 vesicles prepared from the AsecG cells. 30. Schiebel, E., Driessen, A. J. M., Hartl, F.-U., and Wickner, W. (1991) Cell 64, Acknowledgments-We

thank Dr. Shin-ichi Matsuyama of this labo-

927-939 31. Kim, Y. J., and Oliver, D. B. (1994) FEES Lett. 339, 175-180