- Email: [email protected]
Effects of Lipids on the Interaction of SecA with Model Membranes
Archives of Biochemistry and Biophysics Vol. 395, No. 1, November 1, pp. 14 –20, 2001 doi:10.1006/abbi.2001.2565, available online at http://www.idealibrary.com on
Effects of Lipids on the Interaction of SecA with Model Membranes 1 Taeho Ahn,* Joon-Sik Kim,* Byoung-Chul Lee,* and Chul-Ho Yun† ,2 *Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Taejon 305-701, Republic of Korea; and †Department of Genetic Engineering, Pai-Chai University, Taejon 302-735, Republic of Korea
Received June 14, 2001, and in revised form August 9, 2001
The effects of nonlamellar-prone lipids, diacylglycerol and phosphatidylethanolamine (PE), on the kinetic association of SecA with model membranes were examined by measuring changes in the intrinsic emission fluorescence with a stopped-flow apparatus. Upon interaction with standard liposomes composed of 50 mol% dioleolyphosphatidylcholine (DOPC) and 50 mol% of dioleoylphosphatidylglycerol (DOPG), the intrinsic fluorescence intensity of SecA was decreased after a lapse of time with a rate constant of 0.0049 s ⴚ1. When the DOPC of the standard vesicles was gradually replaced with either dioeloyl PE (DOPE) or Escherichia coli (E. coli) PE, the rate constant increased appreciably as a function of PE concentration, in the order DOPE > E. coli PE. In addition, when the PE of E. coli PE/DOPG (50/50) vesicles was replaced with more than 5 mol% dioleoylglycerol (DOG), the rate constant further increased by 40%. The incorporation of nonlamellar-prone lipids in the vesicles also enhanced the binding of SecA to model membranes in the order DOPE > E. coli PE/DOG > E. coli PE > DOPC. These results provide the first kinetic evidence for the importance of nonlamellar-prone phospholipids for the association rate of SecA with membranes. © 2001 Academic Press
Key Words: nonlamellar-prone lipids; diacylglycerol; phosphatidylethanolamine; SecA; model membranes; stopped-flow apparatus.
tory proteins in Escherichia coli (1–3). This protein exists as a homodimer of a 102-kDa subunit and is distributed in vivo about equally between the inner membrane and the cytosol (4). The protein translocation is brought about by close cooperation between SecA and other proteinous components such as SecYEG complex, SecB, and Ffh (reviewed in 5–7). Phospholipids, especially acidic phospholipids, also have been shown to play an important role in protein translocation (8, 9). SecA is an unusual water-soluble protein, which can readily penetrate into membranes and even traverse the lipid bilayer of liposomes (10) and the inner membrane of E. coli to be exposed to the periplasm (11, 12). SecA binds to various phospholipid vesicles including those of phosphatidylcholine (PC), 3 but strictly requires negatively charged phospholipids such as phosphatidylglycerol (PG) and cardiolipin for its proper function of hydrolyzing ATP and using the energy to translocate precursor proteins across the cytoplasmic membrane. This effect of the acidic phospholipids appears to be related to the extent of binding and penetration of SecA to the vesicles. Recently, in close relation to the importance of lipid composition in SecA-membrane interaction, we suggested that the PE, a nonlamellar-prone lipid, in the 3
SecA, an ATPase, is one of the central protein components of the translocation machinery for the secre1
This work was supported by Korea Research Foundation Grant (KRF-2000-015-FS0002). 2 To whom correspondence should be addressed at Department of Genetic Engineering, Pai-Chai University, Taejon 302-735, Republic of Korea. Fax: 82-42-520-5385. E-mail: [email protected]. 14
Abbreviations used: PC, phosphatidylcholine; DG, diacylglycerol; DOG, 1,2-dioleoly-sn-glycerol; DOPC, 1,2-dioleoyl-sn-glycero-3phosphocholine; DOPE, 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine; DOPG, 1,2-dioleoyl-sn-glycero-3-[phospho-rac-(1-glycerol)]; DOPE-NMe, 1,2-dioleoyl-sn-glycero-3-phospho-N-methylethanolamine; DOPE-N(Me) 2 , 1,2-dioleoyl-sn-glycero-3-phospho-N,N-dimethylethanolamine; DOGA, 1,2-dioleoyl-D-glyceramide; 2-MOG, 2-monooleoylglycerol; DOFP, 1,2-dioleoyl-1-fluoro-2,3-propanediol; LUV, large unilamellar vesicle; PE, phosphatidylethanolamine; pyrenePE, 1-palmitoyl-2-(1-pyrenedecanoyl)-sn-glycero-3-phosphoethanolamine; FRET, fluorescence resonance energy transfer; CD, circular dichroism; L-H II, lamellar to hexagonal II transition; T H, L-H II transition temperature; DTT, dithiothreitol. 0003-9861/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.
INTERACTIONS OF SecA WITH NONLAMELLAR-PRONE LIPIDS
vesicles promotes the binding and penetration of SecA into the bilayer, which enhances the ATPase activity (13). We also demonstrated that diacylglycerol (DG), another strong nonlamellar-forming lipid, enhances SecA ATPase activity, which might be ascribed to the increased binding/insertion of SecA into vesicles. DG is a neutral lipid that can lower the lamellar to hexagonal II transition (L-H II) temperature (T H) of PE significantly (14, 15). In E. coli, DG accounts for about 1% of the total membrane lipids, which depends on the growth conditions (16), although the precise role of this lipid is not known. However, in some mutant strains, about 8% of DG is present in the E. coli inner membrane (17). SecA undergoes membrane insertion/deinsertion cycling, which is dependent on ATP binding and its hydrolysis (18). This cycling concomitantly results in the stepwise translocation of preproteins across the membrane with other translocase components, SecYEG. The membrane-binding status of SecA also can be modulated by the interaction with secretory proteins and lipids (13, 19, 20). The dynamic features of SecA, therefore, upon interacting with membranes seem to be important for the efficient function of SecA in the preprotein translocation. However, the factors that specifically determine the rate for the association of SecA with membranes remain unknown. Addressing this issue is fundamental to any translocation model that employs SecA to facilitate targeting of the preprotein to membrane translocation sites. It seems worthwhile, therefore, to examine the kinetics of SecA association with membranes to understand the initial step for the translocation of secretory proteins in E. coli. In the present investigation, the possible effects of nonlamellar-prone lipids on the initial membranebinding step of SecA are examined by measuring kinetic changes in the intrinsic fluorescence intensity of SecA upon interaction with vesicles of various lipid compositions. Our results indicate that binding and the subsequent insertion rate of SecA into lipid bilayers increase with increasing nonlamellar-prone lipid contents and that those abilities of SecA are strictly dependent on the presence of PG, a major acidic phospholipid in the E. coli inner membrane. MATERIALS AND METHODS Materials. All phospholipids and DOG were purchased from Avanti Polar Lipids (Alabaster, AL) and checked for purity by thinlayer chromatography. 1-Palmitoyl-2-(1-pyrenedecanoyl)-sn-glycero3-phosphoethanolamine (pyrene-PE) was obtained from Molecular Probes (Eugene, OR). Chloroform solutions of lipids were stored in sealed ampules under argon at ⫺20°C. All other chemicals were of the highest grade commercially available. SecA preparation. SecA protein was purified from a SecA-overproducing strain (RR1/pMAN400) (21) as described (20), and stored at ⫺70°C until used. Protein concentrations were determined using bicinchoninic acid according to the manufacturer’s instruction (Sigma).
15
Liposome preparation. Vesicles of mixed lipid composition of DOPC/DOPG (50/50, by molar ratio) were used as a standard liposome throughout these investigations. The mole percentage of DOPG was fixed at 50%, but the initial 50% of DOPC in the standard vesicles was varied by replacing it upward of 50 mol% with DOPE, N or N,N-methylated derivatives of DOPE, or E. coli PE. To examine the effect of DG on the SecA-membrane interaction, E. coli PE in E. coli PE/DOPG (50/50) vesicle was further replaced with more than 5 mol% of DOG or its analogs. To prepare vesicles containing pyrenePE, 1.5 mol% of this membrane probe was incorporated into liposomes instead of DOPC, E. coli PE, or DOPE. All liposomes were prepared using the extrusion method with 100 nm pore size polycarbonate membrane, as described (22). All large unilamellar vesicles used for this work were apparently stable for at least 2 days. The change in the light-scattering intensity measured with a spectrofluorometer at 450 nm during this time was less than 10%. The concentrations of phospholipids were determined by a phosphorus assay (23) and that of pyrene-PE was determined spectrophotometrically at 342 nm using 42,000 cm ⫺1 as the molar extinction coefficient (24). Equilibrium fluorescence quenching of SecA by model membranes. The membrane-induced fluorescence quenching of SecA was monitored by measuring Trp emission intensity. A 0.4 M SecA was incubated with lipid vesicles for 30 min at 30°C before measurement. Fluorescence was recorded on a Shimadzu RF-5301 PC spectrofluorometer equipped with a constant-temperature cell holder. The excitation wavelength was 295 nm (1.5 nm slit) and the emission was from 320 to 450 nm (5 nm slit). The membrane binding/insertion of SecA into lipid bilayers was also measured by the fluorescence resonance energy transfer (FRET) between SecA and pyrene-PE incorporated into vesicles, as described (13). Stopped-flow fluorescence measurements. All kinetic experiments were carried out in 25 mM Tris-HCl, pH 7.4 containing 100 mM NaCl, 0.5 mM Na-EDTA, and 1 mM DTT at 30°C. Fluorescence changes of SecA upon interaction with liposomes were monitored using a Bio-Logic SFM-3 stopped-flow apparatus (Bio-Logic, Claix, France) fitted with an emission monochromator such that emitted light could be collected at a fixed bandpass over the entire time range. The excitation wavelength was 295 nm with a bandwidth of 10 nm. Emission intensity was monitored at 340 nm with a bandwidth of 10 nm. Fluorescence curves were the average of at least five independent traces. The data were averaged using the software provided with the stopped-flow apparatus. Fluorescence traces were analyzed using the single-exponential function. Binding amount of SecA to model membranes. The amount of SecA bound to membranes was determined as described previously (20): Briefly, the iodinated SecA (about 50,000 cpm) was incubated with the indicated amount of vesicles for 30 min at 30°C in a buffer containing 50 mM potassium phosphate, 100 mM KCl, and 1 mM DTT (pH 7.5). Protein-bound vesicles were then centrifuged in a Beckman TLA 100.2 rotor at 70,000 rpm for 60 min, and the radioactivities of pellets and supernatant were determined using a Beckman Model 5500 ␥-counter. Circular dichroism (CD) measurements. Time-dependent CD spectra of SecA were recorded on a Jasco J720 spectropolarimeter in a thermostated cuvette with a stopped-flow apparatus at 222 and 208 nm in a 1-cm path length cell. The calibration of the spectropolarimeter was performed using D-10-camphorsulfonic acid, which shows a molar ellipticity of 7800° cm 2/dmol at 290.5 nm in an aqueous solution. The ellipticity was obtained by rapid mixing of 1 M SecA with various phospholipid compositions of vesicles. Calculations of the fractional percentage of secondary structures were carried out using the algorithm of Chang et al. (25).
16
AHN ET AL.
RESULTS
Steady-state fluorescence quenching of SecA by model membranes. To examine the effect of phospholipids, especially PE, a nonlamellar-prone lipid, on the binding feature of SecA to the model membrane at steady state, we utilized the intrinsic fluorescence intensity change of SecA as a function of lipid concentration. SecA has seven Trp residues and these have been used as chromophores for spectroscopic studies. Figure 1A shows that the intrinsic fluorescence intensity of SecA was decreased by liposomes containing DOPG, an acidic phospholipid, regardless of the vesicle composition as assayed by measuring the emission intensity at 340 nm. But in the presence of E. coli PE or DOPE instead of DOPC, the lipid concentration for the halfmaximal fluorescence quenching decreased from 48 –56 to 136 M. However, we could not observe any changes in the maximal emission wavelength of SecA (results not shown). In the absence of DOPG, 100% DOPC vesicles had no effect on the fluorescence quenching, implying the strict necessity of acidic phospholipid to the interaction of SecA with membranes (results not shown). This phenomenon could result from the conformational changes such as unfolding in the region of Trp residues of SecA bound to membranes, as suggested (26), and/or the fluorescence quenching by acidic phospholipid molecules (27). Irrespective of the reasons, this result is apparently ascribed to the binding/insertion of SecA into lipid bilayers and suggests that PE molecules stimulate the association of SecA with the model membranes. These results also agree well with the previous observations (13) and provide the possibility that vesicles without extrinsic fluorescence probes could be used to investigate SecA-membrane interaction. A comparable experiment was also carried out with pyrene-PE incorporated into the membranes utilizing the FRET between Trp fluorescence of SecA and the pyrene group, which is used to examine the penetration of SecA into the lipid bilayers (26). Very similar binding curves to those of Fig. 1A were obtained indicating the stimulating effect of PE on the binding/ insertion of SecA into membranes (Fig. 1B). The results show that the extent of Trp fluorescence quenching of SecA resulting from the interaction with the membrane itself can be used to trace the binding/insertion of SecA into lipid bilayers. Effects of nonlamellar-prone lipids on the binding amount of SecA to membranes. To obtain more insight on the effect of nonlamellar-prone lipids on SecAmembrane association and the Trp fluorescence quenching of SecA upon interaction with lipid bilayers, we measured the amount of SecA bound to the vesicles as a function of lipid concentration. The incorporation
FIG. 1. Effect of nonlamellar-prone lipids on the intrinsic fluorescence of SecA in the absence (A) or presence (B) of pyrene-PE. Maximal emission fluorescence of Trp was monitored by increasing concentration of liposomes with different compositions as described under Materials and Methods. L/P represents the concentration ratio of lipid to protein. In (B), F o and F represent the emission intensities of SecA bound to membranes in the absence (F o) or presence (F) of pyrene-PE, respectively. Data points represent mean ⫾ SE of three independent experiments. A 1.5 mol% of pyrene-PE was incorporated into membranes instead of DOPC, DOPE, or E. coli PE. R.F.I. represents the relative fluorescence intensity. Each symbol represents the lipid compositions as followings: In (A) F, DOPE/DOPG (50/50, mol%); ■, DOPC/DOPG (50/50, mol%); Œ, E. coli PE/DOPG (50/50, mol%). In (B) F, DOPE/DOPG/pyrene-PE (50/48.5/1.5, mol%); ■, DOPC/DOPG/pyrene-PE (50/48.5/1.5, mol%); Œ, E. coli PE/DOPG/ pyrene-PE (50/48.5/1.5, mol%).
of PE and DOG in the vesicles enhanced the binding of SecA to model membranes compared to the case of DOPC in the order DOPE ⭌ E. coli PE/DOG ⬎ E. coli PE (Fig. 2). This result supports the quenching effect on SecA Trp fluorescence by nonlamellar-prone lipids as shown in Fig. 1. Therefore, we could deduce that the
INTERACTIONS OF SecA WITH NONLAMELLAR-PRONE LIPIDS
FIG. 2. Phospholipid-dependent binding of SecA to model membranes. The amount of SecA bound to the membrane was determined as described under Materials and Methods. 125I-labeled SecA was incubated with phospholipid vesicles for 30 min at 30°C. Vesicles with bound SecA were pelleted by ultracentrifugation, and the radioactivities of pellets and supernatant were determined by ␥-ray counting. Data points represent average values of two independent experiments: E, DOPC/DOPG (50/50, mol%); ‚, DOPE/DOPG (50/50, mol%); Œ, E. coli PE/DOPG (50/50, mol%); F, E. coli PE/DOG/DOPG (45/5/50, mol%).
amount of SecA bound to membranes is related to the degree of fluorescence decrease of SecA upon interaction with vesicles. Kinetic study of the interactions of SecA with model membranes by fluorescence stopped-flow measurements. Previous work had qualitatively demonstrated the ability of nonlamellar-prone lipids to enhance the interaction of SecA with membranes. The result was confirmed and extended here. To investigate the initial binding kinetics of SecA to the vesicles, the membrane-induced fluorescence quenching properties of SecA were examined by determining the kinetic constant of binding/insertion of the protein to membranes with a stopped-flow method. Figure 3 shows the fluorescence decay of SecA when the protein was mixed with liposomes as a function of time. As the decay curves were fitted to the single exponential equation, it is reasonable to assume first-order kinetics, at least for the initial period of the interaction. Therefore, the k value (0.0049 s ⫺1) in the presence of DOPC/DOPG (50/ 50) could be assumed as a rate constant for the binding/ insertion of SecA into vesicles. When DOPC of standard vesicles was gradually replaced by DOPE, monomethyl-DOPE (DOPE-NMe), dimethyl-DOPE (DOPE-N(Me) 2), or E. coli PE, the rate constant dramatically increased 3- to 5-fold, with a more pronounced effect of DOPE (0.0203 s ⫺1 in the presence of 50 mol% DOPE) than that of E. coli PE (0.0157 s ⫺1 in the presence of 50 mol% E. coli PE) (Fig.
17
FIG. 3. Kinetics of the SecA binding/penetration to liposomes measured by stopped-flow Trp fluorescence of SecA. Traces show the fluorescence changes of SecA that follow the mixing of equal volumes of solutions from the drive syringe: Line a, 100% DOPC; line b, DOPC/DOPG (50/50, mol%); line c, E. coli PE/DOPG (50/50, mol%); line d, DOPE/DOPG (50/50, mol%); line e, DOPG (100%). The excitation wavelength was 295 nm with a bandwidth of 10 nm. Emission intensity was monitored at 340 nm with a bandwidth of 10 nm. Fluorescence traces show the average of at least five repeats. The data were averaged using the software provided with the stoppedflow apparatus.
4). However, DOPE-NMe and DOPE-N(Me) 2 did not show rate constants comparable to DOPE, but showed a slightly increased rate over that of DOPC. These results suggest again the stimulating function of PE to SecA-membrane association. We also repeated the same kinetic experiment using the Trp fluorescence quenching of SecA by pyrene-PE
FIG. 4. Effects of PE and its N-methylated derivatives on the rate of SecA-membrane association. DOPC and DOPG (50/50, by molar ratio) were used to prepare the standard liposomes. The rate constant for SecA-membrane association was measured after each stepwise replacement of DOPC upward of 50 mol% with DOPE, DOPENMe, DOPE-N(Me) 2, or E. coli PE.
18
AHN ET AL. TABLE I
Kinetics of Trp Fluorescence Quenching of SecA Induced by Pyrene-PE-Containing Membranes a Membrane composition
Rate constant, k (s ⫺1)
100% DOPC DOPC/DOPG (50/50) DOPE/DOPG (50/50) E. coli PE/DOPG (50/50) 100% DOPG
ND b 0.0043 0.0221 0.0149 0.0282
a All rate constants for association of SecA with pyrene-PE-containing vesicles present in the additions were measured as described under Materials and Methods. b Not determined.
incorporated into membranes, which is known to represent the penetration of the protein into lipid bilayers (26). The effect of pyrene-PE incorporated into membranes on the Trp fluorescence quenching of SecA was similar to that of vesicles without any extrinsic fluorescent probes (Table I). To ascertain the enhanced rate of fluorescence quenching of SecA by nonlamellar-prone lipids, PE of E. coli PE/DOPG (50/50) vesicles was replaced with more than 5 mol% of DOG, a strong nonbilayer-forming lipid, and the rate constants were determined. The rate constants further increased by 40% in the presence of 5 mol% DOG (0.0210 s ⫺1) compared to the case without DOG (Fig. 5). We also measured the rate constants of the Trp fluorescence quenching with the DOG analogs such as 2-monooleoylglycerol (2-MOG), dioleoylglycer-
FIG. 5. Effect of DOG or its analogs on the rate of SecA-membrane association. E. coli PE in E. coli PE/DOPG vesicles (50/50, by molar ratio) was replaced by upward of 5 mol% DOG or its analogs and the rate constants for binding/penetration of SecA into membranes were determined as described under Materials and Methods.
FIG. 6. PG concentration-dependent rate constant for the fluorescence quenching of SecA. The membrane-induced fluorescence quenching of SecA was measured after each stepwise replacement of PC in 100% DOPC vesicles with DOPG.
amide (DOGA), and dioleoylfluoropropanediol (DOFP). Interestingly, DOG analogs had no apparent effect on the change of rate constants but rather the constants decreased slightly with increasing concentrations of the analogs. However, in the absence of PG, 100% DOPC vesicles and any replacement of PC with PE or DOG showed no apparent effects on the intrinsic fluorescence change of SecA (results not shown). To examine the effect of DOPG mole fraction in membranes on the rate constants, we replaced 100% DOPC with DOPG gradually and measured the kinetics of fluorescence quenching of SecA upon binding to membranes. The rates increased exponentially with increasing concentrations of DOPG in vesicles and reached 0.0270 s ⫺1 when DOPC was completely replaced by 100% DOPG (Fig. 6). As shown in Fig. 3, the extent of the SecA Trp quenching by 100% DOPG was also greater than those by any other phospholipid compositions. Quenching of Trp fluorescence may result from the interaction with acidic phospholipid, as shown previously (27). In order to examine the kinetic change of secondary structures of SecA upon interaction with membranes, we measured the time-dependent CD spectra of the protein. When the ellipticity of SecA was monitored at 208 and 222 nm by mixing with model membranes as a function of time, no apparent differences were found among DOPC/DOPG (50/50), E. coli PE/DOPG (50/50), DOPE/DOPG (50/50), and 100% DOPG vesicles. This result indicates that nonlamellar-prone lipids themselves had no effects on the secondary structures of SecA. Analysis of the CD spectrum for the SecA in the presence of standard vesicles yielded about 23% ␣-helix
INTERACTIONS OF SecA WITH NONLAMELLAR-PRONE LIPIDS
and about 10% -sheet structure. These results also agree well with previous observations (26). DISCUSSION
In the present investigation, we observed that the intrinsic Trp fluorescence of SecA was quenched by PG-containing liposomes and that the extent of quenching and its kinetics were regulated by the presence of nonlamellar-prone lipids such as PE and DG in the vesicles. The increase of PG concentration, a major acidic phospholipid in E. coli membrane, also enhanced the rate constants of vesicle-induced fluorescence quenching. However, in the absence of PG, we could not detect any apparent change of the SecA Trp fluorescence upon interaction with vesicles regardless of the presence of nonlamellar-prone lipids. These results suggest that the acidic phospholipid is strictly required for the efficient kinetic interaction of SecA with membranes, whereas nonlamellar-prone lipids show stimulatory effects in the presence of acidic phospholipid. Our results presented here provide the first kinetic evidence for distinct characteristics of nonlamellarprone phospholipids for the initial binding of SecA to membranes. These results suggest a possibility that the nonbilayer structure and/or the tendency to form the nonbilayer structure is involved in the enhanced rate of SecA interaction with PG containing-membranes in the presence of PE or DG over PC. This suggestion can be further supported by comparing the results with DOPE and E. coli PE (Fig. 4). DOPE has a lower T H (12°C) than that of E. coli (about 55°C) and is assumed to have the H II conformation at room temperature (28). However, other hypotheses should be conceived when considering the results that DOPE-NMe and DOPE-N(Me) 2 did not show intermediate effects between DOPE and DOPC on the increase of the rate constants, whereas DOPE-NMe is able to form nonbilayer structures at 73°C (29) and DOPC is not. Additional explanations such as hydrogen bonding between PE molecules and/or PE and SecA might be involved. The ethanolamine part of the headgroup of PE contains a primary amine which can form hydrogen bonds (reviewed in 30), whereas “PC-like” PE analogs do not have this property due to the change of packing density and the orientation between the headgroups. The formation of hydrogen bonds between PE molecules might result in the phase separation of membranes into PEand PG-enriched domains. Consequently, the interaction and its rate of SecA with membranes enhance as suggested previously (13). The nonbilayer lipids have been shown to play important roles in protein translocation. Rietveld et al. (31) suggested that nonbilayer lipids are required for the efficient translocation of secretory proteins across the plasma membrane. Recently, it was found that
19
nonbilayer lipids stimulate the activity of the reconstituted SecYEG, a protein translocase, of E. coli and Bacillus subtilis (32). Interestingly, the enhanced interactions of proteins and drugs with acidic phospholipids, which are caused by PE instead of PC, have also been observed in other studies. For instance, such observations were made for the signal sequence of prePhoE (33), the leader peptidase of E. coli (34), and doxorubicin, an anticancer drug (35). In addition, it was demonstrated that PE acts as a molecular chaperone in the correct assembly of lactose permease of E. coli in membranes, whereas PC does not have this function (36). We previously shown that SecA binding and penetration into the membrane are enhanced by the phase separation, indicating that SecA penetration prefers the PG clusters and the increased ATPase activity is the consequence of increased SecA population penetrated into the membrane (13). The presence of DOG also induced the exposure of phospholipid acyl chains to the surface. The exposure can promote the SecA binding to the vesicles. Taken together, the following process concerning the binding and insertion of SecA to membranes could be suggested as a probable mechanism. The first step is the binding of SecA to the negatively charged PG domains induced by PE, nonlamellar-prone phospholipids, on the membrane surface. After binding, the insertion of SecA into the lipid bilayer takes place, followed by extensive hydrophobic interaction between the protein and the lipids which is promoted by PE. This hydrophobic interaction could cause deeper penetration of the SecA into bilayers and eventually translocation across membranes. At the moment, however, we do not have a satisfactory explanation for the fluorescence quenching of SecA upon interacting with lipid bilayers. It might be ascribed to the membrane-induced conformational change of SecA such as partial unfolding, as suggested (26), and/or only fluorescence quenching of SecA Trp residues by a close contact with acidic phospholipid molecules (27). Our results suggest that the fluorescence quenching represents the binding and insertion of SecA to membranes and it can be used for the investigation of the kinetic association between SecA and lipid bilayers. This suggestion can be supported from the fact that rate constants and binding profiles are very similar between native membrane-induced fluorescence quenching of SecA and FRET for Trp-pyrene fluorophore incorporated into bilayers, which was demonstrated for the measurement of SecA penetration into vesicles (26). REFERENCES 1. Cabelli, R., Chen, J. L., Tai, P. C., and Oliver, D. B. (1988) Cell 55, 683– 692.
20
AHN ET AL.
2. Cunningham, K., Lill, R., Crooke, E., Rice, M., Moore, K., Wickner, W., and Oliver D. (1989) EMBO J. 8, 955–959. 3. Watanabe, M., and Blobel, G. (1993) Proc. Natl. Acad. Sci. USA 90, 9011–9015. 4. Cabelli, R. J., Dolan, K. M., Qian, L., and Oliver, D. B. (1991) J. Biol. Chem. 266, 24420 –24427. 5. Wickner, W., and Leonard, M. R. (1996) J. Biol. Chem. 271, 29514 –29516. 6. Economou, A. (1996) Trends Microbiol. 7, 315–320. 7. Manting, E. H., and Driessen, A. J. M. (2000) Mol. Microbiol. 37, 226 –238. 8. Van Klompenburg, W., and de Kruijff, B. (1998) J. Membr. Biol. 162, 1–7. 9. Van Voorst, F., and de Kruijff, B. (2000) Biochem. J. 347, 601– 612. 10. Ahn, T., and Kim, H. (1994) Biochem. Biophys. Res. Commun. 203, 326 –330. 11. Kim, Y. J., Rajapandi, T., and Oliver, D. B. (1994) Cell 78, 845– 853. 12. Economou, A., and Wickner, W. (1994) Cell 78, 835– 843. 13. Ahn, T., and Kim, H. (1998) J. Biol. Chem. 273, 21692–21698. 14. Epand, R. M. (1985) Biochemistry 24, 7092–7095. 15. Das, S., and Rand, R. P. (1986) Biochemistry 25, 2882–2889. 16. Raetz, C. R. H., and Newman, K. F. (1978) J. Biol. Chem. 253, 3882–3887. 17. Raetz, C. R. H., and Newman, K. F. (1979) J. Bacteriol. 137, 860 – 868. 18. Economou, A., Pogliano, J. A., Beckwith, J., Oliver, D. B., and Wickner, W. (1995) Cell 83, 1171–1181. 19. Breukink, E., Demel, R. A., de Korte-Kool, G., and de Kruijff, B. (1992) Biochemistry 31, 1119 –1124. 20. Ahn, T., and Kim, H. (1996) J. Biol. Chem. 271, 12372–12379.
21. Kawasaki, H., Matsuyama, S., Sasaki, S., Akita, M., and Mizushima, S. (1989) FEBS Lett. 242, 431– 434. 22. Ahn, T., Guengerich, P., and Yun, C-H. (1998) Biochemistry 37, 21860 –12866. 23. Vaskovsky, V. E., Kostetsky, E. Y., and Vasendin, I. M. (1975) J. Chromatogr. 114, 129 –141. 24. Galla, H. J., and Hartmann, W. (1980) Chem. Phys. Lipids 27, 199 –219. 25. Chang, C. T., Wu, C. C., and Yang, J. T. (1978) Anal. Biochem. 91, 13–31. 26. Ulbrandt, N. D., London, E., and Oliver, D. B. (1992) J. Biol. Chem. 267, 15184 –15192. 27. Nemat-Gorgani, M., and Dodd, G. (1977) Eur. J. Biochem. 74, 139 –147. 28. Ellens, H., Bentz, J., and Szoka, F. C. (1986) Biochemistry 25, 285–294. 29. Siegel, D. P., Burns, J. L., Chestnut, M. H., and Talmon, Y. (1989) Biophys. J. 56, 161–170. 30. Boggs, J. M. (1987) Biochim. Biophys. Acta 906, 353– 404. 31. Rietveld, A. G., Koorengevel, M. C., and de Kruijff, B. (1995) EMBO J. 14, 5506 –5513. 32. Van der Does, C., Swaving, J., van Klompenburg, W., and Driessen, A. J. M. (2000) J. Biol. Chem. 275, 2472–2478. 33. Van Raalte, A. L. J., Demel, R. A., Verbeekmoes, G., Breukink, E., Keller, R. A. C., and de Kruijff, B. (1996) Eur. J. Biochem. 235, 207–214. 34. Van Klompenburg, W., Paetzel, M., de Jong, J. M., Dalbey, R. E., Demel, R. A., von Heijne, G., and de Kruijff, B. (1998) FEBS Lett. 431, 75–79. 35. Speelmans, G., Staffhorst, R. W. H. M., and de Kruijff, B. (1997) Biochemistry 36, 8657– 8662. 36. Bogdanov, M., Sun, J., Kaback, H R., and Dowhan, W. (1996) J. Biol. Chem. 271, 11615–11618.
Effects of Lipids on the Interaction of SecA with Model Membranes 1 Taeho Ahn,* Joon-Sik Kim,* Byoung-Chul Lee,* and Chul-Ho Yun† ,2 *Department of Biological Sciences, Korea Advanced Institute of Science and Technology, Taejon 305-701, Republic of Korea; and †Department of Genetic Engineering, Pai-Chai University, Taejon 302-735, Republic of Korea
Received June 14, 2001, and in revised form August 9, 2001
The effects of nonlamellar-prone lipids, diacylglycerol and phosphatidylethanolamine (PE), on the kinetic association of SecA with model membranes were examined by measuring changes in the intrinsic emission fluorescence with a stopped-flow apparatus. Upon interaction with standard liposomes composed of 50 mol% dioleolyphosphatidylcholine (DOPC) and 50 mol% of dioleoylphosphatidylglycerol (DOPG), the intrinsic fluorescence intensity of SecA was decreased after a lapse of time with a rate constant of 0.0049 s ⴚ1. When the DOPC of the standard vesicles was gradually replaced with either dioeloyl PE (DOPE) or Escherichia coli (E. coli) PE, the rate constant increased appreciably as a function of PE concentration, in the order DOPE > E. coli PE. In addition, when the PE of E. coli PE/DOPG (50/50) vesicles was replaced with more than 5 mol% dioleoylglycerol (DOG), the rate constant further increased by 40%. The incorporation of nonlamellar-prone lipids in the vesicles also enhanced the binding of SecA to model membranes in the order DOPE > E. coli PE/DOG > E. coli PE > DOPC. These results provide the first kinetic evidence for the importance of nonlamellar-prone phospholipids for the association rate of SecA with membranes. © 2001 Academic Press
Key Words: nonlamellar-prone lipids; diacylglycerol; phosphatidylethanolamine; SecA; model membranes; stopped-flow apparatus.
tory proteins in Escherichia coli (1–3). This protein exists as a homodimer of a 102-kDa subunit and is distributed in vivo about equally between the inner membrane and the cytosol (4). The protein translocation is brought about by close cooperation between SecA and other proteinous components such as SecYEG complex, SecB, and Ffh (reviewed in 5–7). Phospholipids, especially acidic phospholipids, also have been shown to play an important role in protein translocation (8, 9). SecA is an unusual water-soluble protein, which can readily penetrate into membranes and even traverse the lipid bilayer of liposomes (10) and the inner membrane of E. coli to be exposed to the periplasm (11, 12). SecA binds to various phospholipid vesicles including those of phosphatidylcholine (PC), 3 but strictly requires negatively charged phospholipids such as phosphatidylglycerol (PG) and cardiolipin for its proper function of hydrolyzing ATP and using the energy to translocate precursor proteins across the cytoplasmic membrane. This effect of the acidic phospholipids appears to be related to the extent of binding and penetration of SecA to the vesicles. Recently, in close relation to the importance of lipid composition in SecA-membrane interaction, we suggested that the PE, a nonlamellar-prone lipid, in the 3
SecA, an ATPase, is one of the central protein components of the translocation machinery for the secre1
This work was supported by Korea Research Foundation Grant (KRF-2000-015-FS0002). 2 To whom correspondence should be addressed at Department of Genetic Engineering, Pai-Chai University, Taejon 302-735, Republic of Korea. Fax: 82-42-520-5385. E-mail: [email protected]. 14
Abbreviations used: PC, phosphatidylcholine; DG, diacylglycerol; DOG, 1,2-dioleoly-sn-glycerol; DOPC, 1,2-dioleoyl-sn-glycero-3phosphocholine; DOPE, 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine; DOPG, 1,2-dioleoyl-sn-glycero-3-[phospho-rac-(1-glycerol)]; DOPE-NMe, 1,2-dioleoyl-sn-glycero-3-phospho-N-methylethanolamine; DOPE-N(Me) 2 , 1,2-dioleoyl-sn-glycero-3-phospho-N,N-dimethylethanolamine; DOGA, 1,2-dioleoyl-D-glyceramide; 2-MOG, 2-monooleoylglycerol; DOFP, 1,2-dioleoyl-1-fluoro-2,3-propanediol; LUV, large unilamellar vesicle; PE, phosphatidylethanolamine; pyrenePE, 1-palmitoyl-2-(1-pyrenedecanoyl)-sn-glycero-3-phosphoethanolamine; FRET, fluorescence resonance energy transfer; CD, circular dichroism; L-H II, lamellar to hexagonal II transition; T H, L-H II transition temperature; DTT, dithiothreitol. 0003-9861/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.
INTERACTIONS OF SecA WITH NONLAMELLAR-PRONE LIPIDS
vesicles promotes the binding and penetration of SecA into the bilayer, which enhances the ATPase activity (13). We also demonstrated that diacylglycerol (DG), another strong nonlamellar-forming lipid, enhances SecA ATPase activity, which might be ascribed to the increased binding/insertion of SecA into vesicles. DG is a neutral lipid that can lower the lamellar to hexagonal II transition (L-H II) temperature (T H) of PE significantly (14, 15). In E. coli, DG accounts for about 1% of the total membrane lipids, which depends on the growth conditions (16), although the precise role of this lipid is not known. However, in some mutant strains, about 8% of DG is present in the E. coli inner membrane (17). SecA undergoes membrane insertion/deinsertion cycling, which is dependent on ATP binding and its hydrolysis (18). This cycling concomitantly results in the stepwise translocation of preproteins across the membrane with other translocase components, SecYEG. The membrane-binding status of SecA also can be modulated by the interaction with secretory proteins and lipids (13, 19, 20). The dynamic features of SecA, therefore, upon interacting with membranes seem to be important for the efficient function of SecA in the preprotein translocation. However, the factors that specifically determine the rate for the association of SecA with membranes remain unknown. Addressing this issue is fundamental to any translocation model that employs SecA to facilitate targeting of the preprotein to membrane translocation sites. It seems worthwhile, therefore, to examine the kinetics of SecA association with membranes to understand the initial step for the translocation of secretory proteins in E. coli. In the present investigation, the possible effects of nonlamellar-prone lipids on the initial membranebinding step of SecA are examined by measuring kinetic changes in the intrinsic fluorescence intensity of SecA upon interaction with vesicles of various lipid compositions. Our results indicate that binding and the subsequent insertion rate of SecA into lipid bilayers increase with increasing nonlamellar-prone lipid contents and that those abilities of SecA are strictly dependent on the presence of PG, a major acidic phospholipid in the E. coli inner membrane. MATERIALS AND METHODS Materials. All phospholipids and DOG were purchased from Avanti Polar Lipids (Alabaster, AL) and checked for purity by thinlayer chromatography. 1-Palmitoyl-2-(1-pyrenedecanoyl)-sn-glycero3-phosphoethanolamine (pyrene-PE) was obtained from Molecular Probes (Eugene, OR). Chloroform solutions of lipids were stored in sealed ampules under argon at ⫺20°C. All other chemicals were of the highest grade commercially available. SecA preparation. SecA protein was purified from a SecA-overproducing strain (RR1/pMAN400) (21) as described (20), and stored at ⫺70°C until used. Protein concentrations were determined using bicinchoninic acid according to the manufacturer’s instruction (Sigma).
15
Liposome preparation. Vesicles of mixed lipid composition of DOPC/DOPG (50/50, by molar ratio) were used as a standard liposome throughout these investigations. The mole percentage of DOPG was fixed at 50%, but the initial 50% of DOPC in the standard vesicles was varied by replacing it upward of 50 mol% with DOPE, N or N,N-methylated derivatives of DOPE, or E. coli PE. To examine the effect of DG on the SecA-membrane interaction, E. coli PE in E. coli PE/DOPG (50/50) vesicle was further replaced with more than 5 mol% of DOG or its analogs. To prepare vesicles containing pyrenePE, 1.5 mol% of this membrane probe was incorporated into liposomes instead of DOPC, E. coli PE, or DOPE. All liposomes were prepared using the extrusion method with 100 nm pore size polycarbonate membrane, as described (22). All large unilamellar vesicles used for this work were apparently stable for at least 2 days. The change in the light-scattering intensity measured with a spectrofluorometer at 450 nm during this time was less than 10%. The concentrations of phospholipids were determined by a phosphorus assay (23) and that of pyrene-PE was determined spectrophotometrically at 342 nm using 42,000 cm ⫺1 as the molar extinction coefficient (24). Equilibrium fluorescence quenching of SecA by model membranes. The membrane-induced fluorescence quenching of SecA was monitored by measuring Trp emission intensity. A 0.4 M SecA was incubated with lipid vesicles for 30 min at 30°C before measurement. Fluorescence was recorded on a Shimadzu RF-5301 PC spectrofluorometer equipped with a constant-temperature cell holder. The excitation wavelength was 295 nm (1.5 nm slit) and the emission was from 320 to 450 nm (5 nm slit). The membrane binding/insertion of SecA into lipid bilayers was also measured by the fluorescence resonance energy transfer (FRET) between SecA and pyrene-PE incorporated into vesicles, as described (13). Stopped-flow fluorescence measurements. All kinetic experiments were carried out in 25 mM Tris-HCl, pH 7.4 containing 100 mM NaCl, 0.5 mM Na-EDTA, and 1 mM DTT at 30°C. Fluorescence changes of SecA upon interaction with liposomes were monitored using a Bio-Logic SFM-3 stopped-flow apparatus (Bio-Logic, Claix, France) fitted with an emission monochromator such that emitted light could be collected at a fixed bandpass over the entire time range. The excitation wavelength was 295 nm with a bandwidth of 10 nm. Emission intensity was monitored at 340 nm with a bandwidth of 10 nm. Fluorescence curves were the average of at least five independent traces. The data were averaged using the software provided with the stopped-flow apparatus. Fluorescence traces were analyzed using the single-exponential function. Binding amount of SecA to model membranes. The amount of SecA bound to membranes was determined as described previously (20): Briefly, the iodinated SecA (about 50,000 cpm) was incubated with the indicated amount of vesicles for 30 min at 30°C in a buffer containing 50 mM potassium phosphate, 100 mM KCl, and 1 mM DTT (pH 7.5). Protein-bound vesicles were then centrifuged in a Beckman TLA 100.2 rotor at 70,000 rpm for 60 min, and the radioactivities of pellets and supernatant were determined using a Beckman Model 5500 ␥-counter. Circular dichroism (CD) measurements. Time-dependent CD spectra of SecA were recorded on a Jasco J720 spectropolarimeter in a thermostated cuvette with a stopped-flow apparatus at 222 and 208 nm in a 1-cm path length cell. The calibration of the spectropolarimeter was performed using D-10-camphorsulfonic acid, which shows a molar ellipticity of 7800° cm 2/dmol at 290.5 nm in an aqueous solution. The ellipticity was obtained by rapid mixing of 1 M SecA with various phospholipid compositions of vesicles. Calculations of the fractional percentage of secondary structures were carried out using the algorithm of Chang et al. (25).
16
AHN ET AL.
RESULTS
Steady-state fluorescence quenching of SecA by model membranes. To examine the effect of phospholipids, especially PE, a nonlamellar-prone lipid, on the binding feature of SecA to the model membrane at steady state, we utilized the intrinsic fluorescence intensity change of SecA as a function of lipid concentration. SecA has seven Trp residues and these have been used as chromophores for spectroscopic studies. Figure 1A shows that the intrinsic fluorescence intensity of SecA was decreased by liposomes containing DOPG, an acidic phospholipid, regardless of the vesicle composition as assayed by measuring the emission intensity at 340 nm. But in the presence of E. coli PE or DOPE instead of DOPC, the lipid concentration for the halfmaximal fluorescence quenching decreased from 48 –56 to 136 M. However, we could not observe any changes in the maximal emission wavelength of SecA (results not shown). In the absence of DOPG, 100% DOPC vesicles had no effect on the fluorescence quenching, implying the strict necessity of acidic phospholipid to the interaction of SecA with membranes (results not shown). This phenomenon could result from the conformational changes such as unfolding in the region of Trp residues of SecA bound to membranes, as suggested (26), and/or the fluorescence quenching by acidic phospholipid molecules (27). Irrespective of the reasons, this result is apparently ascribed to the binding/insertion of SecA into lipid bilayers and suggests that PE molecules stimulate the association of SecA with the model membranes. These results also agree well with the previous observations (13) and provide the possibility that vesicles without extrinsic fluorescence probes could be used to investigate SecA-membrane interaction. A comparable experiment was also carried out with pyrene-PE incorporated into the membranes utilizing the FRET between Trp fluorescence of SecA and the pyrene group, which is used to examine the penetration of SecA into the lipid bilayers (26). Very similar binding curves to those of Fig. 1A were obtained indicating the stimulating effect of PE on the binding/ insertion of SecA into membranes (Fig. 1B). The results show that the extent of Trp fluorescence quenching of SecA resulting from the interaction with the membrane itself can be used to trace the binding/insertion of SecA into lipid bilayers. Effects of nonlamellar-prone lipids on the binding amount of SecA to membranes. To obtain more insight on the effect of nonlamellar-prone lipids on SecAmembrane association and the Trp fluorescence quenching of SecA upon interaction with lipid bilayers, we measured the amount of SecA bound to the vesicles as a function of lipid concentration. The incorporation
FIG. 1. Effect of nonlamellar-prone lipids on the intrinsic fluorescence of SecA in the absence (A) or presence (B) of pyrene-PE. Maximal emission fluorescence of Trp was monitored by increasing concentration of liposomes with different compositions as described under Materials and Methods. L/P represents the concentration ratio of lipid to protein. In (B), F o and F represent the emission intensities of SecA bound to membranes in the absence (F o) or presence (F) of pyrene-PE, respectively. Data points represent mean ⫾ SE of three independent experiments. A 1.5 mol% of pyrene-PE was incorporated into membranes instead of DOPC, DOPE, or E. coli PE. R.F.I. represents the relative fluorescence intensity. Each symbol represents the lipid compositions as followings: In (A) F, DOPE/DOPG (50/50, mol%); ■, DOPC/DOPG (50/50, mol%); Œ, E. coli PE/DOPG (50/50, mol%). In (B) F, DOPE/DOPG/pyrene-PE (50/48.5/1.5, mol%); ■, DOPC/DOPG/pyrene-PE (50/48.5/1.5, mol%); Œ, E. coli PE/DOPG/ pyrene-PE (50/48.5/1.5, mol%).
of PE and DOG in the vesicles enhanced the binding of SecA to model membranes compared to the case of DOPC in the order DOPE ⭌ E. coli PE/DOG ⬎ E. coli PE (Fig. 2). This result supports the quenching effect on SecA Trp fluorescence by nonlamellar-prone lipids as shown in Fig. 1. Therefore, we could deduce that the
INTERACTIONS OF SecA WITH NONLAMELLAR-PRONE LIPIDS
FIG. 2. Phospholipid-dependent binding of SecA to model membranes. The amount of SecA bound to the membrane was determined as described under Materials and Methods. 125I-labeled SecA was incubated with phospholipid vesicles for 30 min at 30°C. Vesicles with bound SecA were pelleted by ultracentrifugation, and the radioactivities of pellets and supernatant were determined by ␥-ray counting. Data points represent average values of two independent experiments: E, DOPC/DOPG (50/50, mol%); ‚, DOPE/DOPG (50/50, mol%); Œ, E. coli PE/DOPG (50/50, mol%); F, E. coli PE/DOG/DOPG (45/5/50, mol%).
amount of SecA bound to membranes is related to the degree of fluorescence decrease of SecA upon interaction with vesicles. Kinetic study of the interactions of SecA with model membranes by fluorescence stopped-flow measurements. Previous work had qualitatively demonstrated the ability of nonlamellar-prone lipids to enhance the interaction of SecA with membranes. The result was confirmed and extended here. To investigate the initial binding kinetics of SecA to the vesicles, the membrane-induced fluorescence quenching properties of SecA were examined by determining the kinetic constant of binding/insertion of the protein to membranes with a stopped-flow method. Figure 3 shows the fluorescence decay of SecA when the protein was mixed with liposomes as a function of time. As the decay curves were fitted to the single exponential equation, it is reasonable to assume first-order kinetics, at least for the initial period of the interaction. Therefore, the k value (0.0049 s ⫺1) in the presence of DOPC/DOPG (50/ 50) could be assumed as a rate constant for the binding/ insertion of SecA into vesicles. When DOPC of standard vesicles was gradually replaced by DOPE, monomethyl-DOPE (DOPE-NMe), dimethyl-DOPE (DOPE-N(Me) 2), or E. coli PE, the rate constant dramatically increased 3- to 5-fold, with a more pronounced effect of DOPE (0.0203 s ⫺1 in the presence of 50 mol% DOPE) than that of E. coli PE (0.0157 s ⫺1 in the presence of 50 mol% E. coli PE) (Fig.
17
FIG. 3. Kinetics of the SecA binding/penetration to liposomes measured by stopped-flow Trp fluorescence of SecA. Traces show the fluorescence changes of SecA that follow the mixing of equal volumes of solutions from the drive syringe: Line a, 100% DOPC; line b, DOPC/DOPG (50/50, mol%); line c, E. coli PE/DOPG (50/50, mol%); line d, DOPE/DOPG (50/50, mol%); line e, DOPG (100%). The excitation wavelength was 295 nm with a bandwidth of 10 nm. Emission intensity was monitored at 340 nm with a bandwidth of 10 nm. Fluorescence traces show the average of at least five repeats. The data were averaged using the software provided with the stoppedflow apparatus.
4). However, DOPE-NMe and DOPE-N(Me) 2 did not show rate constants comparable to DOPE, but showed a slightly increased rate over that of DOPC. These results suggest again the stimulating function of PE to SecA-membrane association. We also repeated the same kinetic experiment using the Trp fluorescence quenching of SecA by pyrene-PE
FIG. 4. Effects of PE and its N-methylated derivatives on the rate of SecA-membrane association. DOPC and DOPG (50/50, by molar ratio) were used to prepare the standard liposomes. The rate constant for SecA-membrane association was measured after each stepwise replacement of DOPC upward of 50 mol% with DOPE, DOPENMe, DOPE-N(Me) 2, or E. coli PE.
18
AHN ET AL. TABLE I
Kinetics of Trp Fluorescence Quenching of SecA Induced by Pyrene-PE-Containing Membranes a Membrane composition
Rate constant, k (s ⫺1)
100% DOPC DOPC/DOPG (50/50) DOPE/DOPG (50/50) E. coli PE/DOPG (50/50) 100% DOPG
ND b 0.0043 0.0221 0.0149 0.0282
a All rate constants for association of SecA with pyrene-PE-containing vesicles present in the additions were measured as described under Materials and Methods. b Not determined.
incorporated into membranes, which is known to represent the penetration of the protein into lipid bilayers (26). The effect of pyrene-PE incorporated into membranes on the Trp fluorescence quenching of SecA was similar to that of vesicles without any extrinsic fluorescent probes (Table I). To ascertain the enhanced rate of fluorescence quenching of SecA by nonlamellar-prone lipids, PE of E. coli PE/DOPG (50/50) vesicles was replaced with more than 5 mol% of DOG, a strong nonbilayer-forming lipid, and the rate constants were determined. The rate constants further increased by 40% in the presence of 5 mol% DOG (0.0210 s ⫺1) compared to the case without DOG (Fig. 5). We also measured the rate constants of the Trp fluorescence quenching with the DOG analogs such as 2-monooleoylglycerol (2-MOG), dioleoylglycer-
FIG. 5. Effect of DOG or its analogs on the rate of SecA-membrane association. E. coli PE in E. coli PE/DOPG vesicles (50/50, by molar ratio) was replaced by upward of 5 mol% DOG or its analogs and the rate constants for binding/penetration of SecA into membranes were determined as described under Materials and Methods.
FIG. 6. PG concentration-dependent rate constant for the fluorescence quenching of SecA. The membrane-induced fluorescence quenching of SecA was measured after each stepwise replacement of PC in 100% DOPC vesicles with DOPG.
amide (DOGA), and dioleoylfluoropropanediol (DOFP). Interestingly, DOG analogs had no apparent effect on the change of rate constants but rather the constants decreased slightly with increasing concentrations of the analogs. However, in the absence of PG, 100% DOPC vesicles and any replacement of PC with PE or DOG showed no apparent effects on the intrinsic fluorescence change of SecA (results not shown). To examine the effect of DOPG mole fraction in membranes on the rate constants, we replaced 100% DOPC with DOPG gradually and measured the kinetics of fluorescence quenching of SecA upon binding to membranes. The rates increased exponentially with increasing concentrations of DOPG in vesicles and reached 0.0270 s ⫺1 when DOPC was completely replaced by 100% DOPG (Fig. 6). As shown in Fig. 3, the extent of the SecA Trp quenching by 100% DOPG was also greater than those by any other phospholipid compositions. Quenching of Trp fluorescence may result from the interaction with acidic phospholipid, as shown previously (27). In order to examine the kinetic change of secondary structures of SecA upon interaction with membranes, we measured the time-dependent CD spectra of the protein. When the ellipticity of SecA was monitored at 208 and 222 nm by mixing with model membranes as a function of time, no apparent differences were found among DOPC/DOPG (50/50), E. coli PE/DOPG (50/50), DOPE/DOPG (50/50), and 100% DOPG vesicles. This result indicates that nonlamellar-prone lipids themselves had no effects on the secondary structures of SecA. Analysis of the CD spectrum for the SecA in the presence of standard vesicles yielded about 23% ␣-helix
INTERACTIONS OF SecA WITH NONLAMELLAR-PRONE LIPIDS
and about 10% -sheet structure. These results also agree well with previous observations (26). DISCUSSION
In the present investigation, we observed that the intrinsic Trp fluorescence of SecA was quenched by PG-containing liposomes and that the extent of quenching and its kinetics were regulated by the presence of nonlamellar-prone lipids such as PE and DG in the vesicles. The increase of PG concentration, a major acidic phospholipid in E. coli membrane, also enhanced the rate constants of vesicle-induced fluorescence quenching. However, in the absence of PG, we could not detect any apparent change of the SecA Trp fluorescence upon interaction with vesicles regardless of the presence of nonlamellar-prone lipids. These results suggest that the acidic phospholipid is strictly required for the efficient kinetic interaction of SecA with membranes, whereas nonlamellar-prone lipids show stimulatory effects in the presence of acidic phospholipid. Our results presented here provide the first kinetic evidence for distinct characteristics of nonlamellarprone phospholipids for the initial binding of SecA to membranes. These results suggest a possibility that the nonbilayer structure and/or the tendency to form the nonbilayer structure is involved in the enhanced rate of SecA interaction with PG containing-membranes in the presence of PE or DG over PC. This suggestion can be further supported by comparing the results with DOPE and E. coli PE (Fig. 4). DOPE has a lower T H (12°C) than that of E. coli (about 55°C) and is assumed to have the H II conformation at room temperature (28). However, other hypotheses should be conceived when considering the results that DOPE-NMe and DOPE-N(Me) 2 did not show intermediate effects between DOPE and DOPC on the increase of the rate constants, whereas DOPE-NMe is able to form nonbilayer structures at 73°C (29) and DOPC is not. Additional explanations such as hydrogen bonding between PE molecules and/or PE and SecA might be involved. The ethanolamine part of the headgroup of PE contains a primary amine which can form hydrogen bonds (reviewed in 30), whereas “PC-like” PE analogs do not have this property due to the change of packing density and the orientation between the headgroups. The formation of hydrogen bonds between PE molecules might result in the phase separation of membranes into PEand PG-enriched domains. Consequently, the interaction and its rate of SecA with membranes enhance as suggested previously (13). The nonbilayer lipids have been shown to play important roles in protein translocation. Rietveld et al. (31) suggested that nonbilayer lipids are required for the efficient translocation of secretory proteins across the plasma membrane. Recently, it was found that
19
nonbilayer lipids stimulate the activity of the reconstituted SecYEG, a protein translocase, of E. coli and Bacillus subtilis (32). Interestingly, the enhanced interactions of proteins and drugs with acidic phospholipids, which are caused by PE instead of PC, have also been observed in other studies. For instance, such observations were made for the signal sequence of prePhoE (33), the leader peptidase of E. coli (34), and doxorubicin, an anticancer drug (35). In addition, it was demonstrated that PE acts as a molecular chaperone in the correct assembly of lactose permease of E. coli in membranes, whereas PC does not have this function (36). We previously shown that SecA binding and penetration into the membrane are enhanced by the phase separation, indicating that SecA penetration prefers the PG clusters and the increased ATPase activity is the consequence of increased SecA population penetrated into the membrane (13). The presence of DOG also induced the exposure of phospholipid acyl chains to the surface. The exposure can promote the SecA binding to the vesicles. Taken together, the following process concerning the binding and insertion of SecA to membranes could be suggested as a probable mechanism. The first step is the binding of SecA to the negatively charged PG domains induced by PE, nonlamellar-prone phospholipids, on the membrane surface. After binding, the insertion of SecA into the lipid bilayer takes place, followed by extensive hydrophobic interaction between the protein and the lipids which is promoted by PE. This hydrophobic interaction could cause deeper penetration of the SecA into bilayers and eventually translocation across membranes. At the moment, however, we do not have a satisfactory explanation for the fluorescence quenching of SecA upon interacting with lipid bilayers. It might be ascribed to the membrane-induced conformational change of SecA such as partial unfolding, as suggested (26), and/or only fluorescence quenching of SecA Trp residues by a close contact with acidic phospholipid molecules (27). Our results suggest that the fluorescence quenching represents the binding and insertion of SecA to membranes and it can be used for the investigation of the kinetic association between SecA and lipid bilayers. This suggestion can be supported from the fact that rate constants and binding profiles are very similar between native membrane-induced fluorescence quenching of SecA and FRET for Trp-pyrene fluorophore incorporated into bilayers, which was demonstrated for the measurement of SecA penetration into vesicles (26). REFERENCES 1. Cabelli, R., Chen, J. L., Tai, P. C., and Oliver, D. B. (1988) Cell 55, 683– 692.
20
AHN ET AL.
2. Cunningham, K., Lill, R., Crooke, E., Rice, M., Moore, K., Wickner, W., and Oliver D. (1989) EMBO J. 8, 955–959. 3. Watanabe, M., and Blobel, G. (1993) Proc. Natl. Acad. Sci. USA 90, 9011–9015. 4. Cabelli, R. J., Dolan, K. M., Qian, L., and Oliver, D. B. (1991) J. Biol. Chem. 266, 24420 –24427. 5. Wickner, W., and Leonard, M. R. (1996) J. Biol. Chem. 271, 29514 –29516. 6. Economou, A. (1996) Trends Microbiol. 7, 315–320. 7. Manting, E. H., and Driessen, A. J. M. (2000) Mol. Microbiol. 37, 226 –238. 8. Van Klompenburg, W., and de Kruijff, B. (1998) J. Membr. Biol. 162, 1–7. 9. Van Voorst, F., and de Kruijff, B. (2000) Biochem. J. 347, 601– 612. 10. Ahn, T., and Kim, H. (1994) Biochem. Biophys. Res. Commun. 203, 326 –330. 11. Kim, Y. J., Rajapandi, T., and Oliver, D. B. (1994) Cell 78, 845– 853. 12. Economou, A., and Wickner, W. (1994) Cell 78, 835– 843. 13. Ahn, T., and Kim, H. (1998) J. Biol. Chem. 273, 21692–21698. 14. Epand, R. M. (1985) Biochemistry 24, 7092–7095. 15. Das, S., and Rand, R. P. (1986) Biochemistry 25, 2882–2889. 16. Raetz, C. R. H., and Newman, K. F. (1978) J. Biol. Chem. 253, 3882–3887. 17. Raetz, C. R. H., and Newman, K. F. (1979) J. Bacteriol. 137, 860 – 868. 18. Economou, A., Pogliano, J. A., Beckwith, J., Oliver, D. B., and Wickner, W. (1995) Cell 83, 1171–1181. 19. Breukink, E., Demel, R. A., de Korte-Kool, G., and de Kruijff, B. (1992) Biochemistry 31, 1119 –1124. 20. Ahn, T., and Kim, H. (1996) J. Biol. Chem. 271, 12372–12379.
21. Kawasaki, H., Matsuyama, S., Sasaki, S., Akita, M., and Mizushima, S. (1989) FEBS Lett. 242, 431– 434. 22. Ahn, T., Guengerich, P., and Yun, C-H. (1998) Biochemistry 37, 21860 –12866. 23. Vaskovsky, V. E., Kostetsky, E. Y., and Vasendin, I. M. (1975) J. Chromatogr. 114, 129 –141. 24. Galla, H. J., and Hartmann, W. (1980) Chem. Phys. Lipids 27, 199 –219. 25. Chang, C. T., Wu, C. C., and Yang, J. T. (1978) Anal. Biochem. 91, 13–31. 26. Ulbrandt, N. D., London, E., and Oliver, D. B. (1992) J. Biol. Chem. 267, 15184 –15192. 27. Nemat-Gorgani, M., and Dodd, G. (1977) Eur. J. Biochem. 74, 139 –147. 28. Ellens, H., Bentz, J., and Szoka, F. C. (1986) Biochemistry 25, 285–294. 29. Siegel, D. P., Burns, J. L., Chestnut, M. H., and Talmon, Y. (1989) Biophys. J. 56, 161–170. 30. Boggs, J. M. (1987) Biochim. Biophys. Acta 906, 353– 404. 31. Rietveld, A. G., Koorengevel, M. C., and de Kruijff, B. (1995) EMBO J. 14, 5506 –5513. 32. Van der Does, C., Swaving, J., van Klompenburg, W., and Driessen, A. J. M. (2000) J. Biol. Chem. 275, 2472–2478. 33. Van Raalte, A. L. J., Demel, R. A., Verbeekmoes, G., Breukink, E., Keller, R. A. C., and de Kruijff, B. (1996) Eur. J. Biochem. 235, 207–214. 34. Van Klompenburg, W., Paetzel, M., de Jong, J. M., Dalbey, R. E., Demel, R. A., von Heijne, G., and de Kruijff, B. (1998) FEBS Lett. 431, 75–79. 35. Speelmans, G., Staffhorst, R. W. H. M., and de Kruijff, B. (1997) Biochemistry 36, 8657– 8662. 36. Bogdanov, M., Sun, J., Kaback, H R., and Dowhan, W. (1996) J. Biol. Chem. 271, 11615–11618.