854

Effect of Phosphatidylcholine on the Fusion of Vesicles

-Lactalbumin-lnduced

Byoung Sun Parka, Jeongha Kimb, Uh Hee Kirna and Hyoungman Kimb,* aDepartment of Science Education, Dankook University,Seoul, Korea, and bDepartment of Biological Science and Engineering, Korea Advanced institute of Science, P.O. Box 150, Chongyang, Seou1130-650 Korea

Previous studies on a-lactalbumin induced fusion of phosphatidylserine/phosphatidylethanolamine vesicles are extended to vesicles composed of various combinations of phosphatidylserine, phosphatidylethanolamine, phosphatidylcholine and cardiolipin. It was found that inclusion of phosphatidylcholine in the vesicles results in a depression of fusion. This depression of fusion appears to be caused by a reduction in the amount of irreversibly bound a-lactalbumin to vesicles containing phosphatidylcholine. It is suggested that in this system fusion is dependent upon the extent by which a particular protein segment penetrates the bilayer. Lipids 24, 854-858 (1989). Recently, there have been extensive investigations on protein-induced membrane fusion. Although many proteins were found to fuse model membranes under various conditions, the basic mechanism of the fusion process remains unknown. The general consensus that emerged from these investigations, however, appears to be that hydrophobic interactions between protein and lipid vesicles are a prerequisite for fusion {1-4}. Since the hydrophobic interaction between lipid vesicles and a protein generally involves the interaction of the protein with phospholipid acyl chains, we investigated the penetration of the phosphatidylserine/ phosphatidylethanolamine {PS/PE, 1:1} bilayer by several proteins including a-lactalbumin {a-LA} under conditions of fusion (5}. It was found that at low pH, a small segment of a-LA in the presence of the vesicles is protected from proteolytic digestion, and that the extent of fusion is dependent on this interaction. Hydrophobic labeling of this s e g m e n t with 3(trifluoromethyl}-3-(m-[12~I]iodophenyl}diazirine {[125I]TID) and subsequent identification of labeled amino acid residues suggested that an a-helical form of the segment protrudes partially into the bilayer. One surface of the a-helix appears to be in contact with the hydrophobic acyl chain of the lipid while the opposite surface of the helix is exposed to the aqueous environment i6). The question of why this type of interaction should induce fusion is still to be answered. The choice of PS/PE vesicles for these earlier investigations was made because such vesicles have extensively been used in studies of cation induced fusion and because PE is known to enhance the fusion process. Studies showed that Ca2+-induced fusion of vesicles containing phosphatidylcholine (PC) was generally less pronounced than fusion of vesicles containing PE {7,8}. Here we show a similar tendency for a-LA*To whom correspondenceshould be addressed. Abbreviations: a-LA, a-lactalbumin; TID, 3-(trifluoromethyl}3-(m-[125I]iodophenyl) diazirine; PE, phosphatidylethanolamine; PC, phosphatidylcholine;Tb, terbium; DPA, dipicolinic acid; PS, phosphatidylserine. LtPID$,Vol. 24, No. 10 (1989)

induced fusion. It was found that the extent of fusion is proportional to the extent of protein binding, supporting the notion that binding and subsequent penetration of a segment of the proteins are responsible for fusion.

MATERIALSAND METHODS Materials. PS {from bovine brain), PE {from bovine brain}, PC (from egg yolk}, a-LA (from bovine milk}, trypsin {from bovine pancrease), and dipicolinic acid (DPA, pyridine-2,6-dicarboxylic acid} were purchased from Sigma Chemical Co. (St. Louis, MO). TbC13 • 6H20 (99.99%} was obtained from Alfa (Danvers, MA). All chemicals were purchased in the highest purity available. The a-LA was purified by using Sephadex G-100 column chromatography. All phospholipids migrated as single spots in thin-layer chromatography on silica gel. The amount of lipid applied was 100 ~g, and the solvent used was chloroform/methanol/water 165:25:4, v/v/v). Phospholipid concentration was determined according to the method of Vaskowsky et al. (9). Proteins were determined with a Gilford 260 UVvisible spectrophotometer. Vesicle formation. Phospholipid vesicles composed of various phospholipids were prepared by the ether injection technique {10) followed by 15-s sonication. The multilamellar liposomes were removed by centrifugation at 6,500 rpm for 30 min, and the supernatant was collected. Relatively homogeneous unilameUar vesicles (-50 nm diameter} were obtained as determined after negative staining with uranyl acetate using a Japan Electron JE M 100 CX-II electron microscope. Fusion. The fusion process was followed by the terbium-dipicolinic acid (Tb-DPA} method of Wilschut et al. (11). This method monitors the extent of mixing of vesicle contents upon fusion and has been widely used for model membrane fusion studies. One population of vesicles was prepared in a solution containing 2.5 mM TbC13 and 50 mM sodium citrate, and the other population of vesicles was prepared in a solution containing 50 mM DPA {sodium salt) and 20 mM NaC1. The vesicles were purified from nonencapsulated material by gel filtration on a Sephadex G-75 column, using a buffer solution containing 1.0 mM EDTA for elution. After equal concentrations of the two types of vesicles and protein at various concentrations were mixed, the fusion process was followed by using an AmincoBowman spectrofluorometer. The leakage of the vesicles after the addition of the fusogenic agent was monitored by using PS/PE vesicles containing 1.25 mM TbC13, 25 mM sodium citrate, and 25 mM DPA {12}. Again, the nonencapsulated material was removed by gel filtration (see above}. The decrease in the fluorescence intensity after initiation of fusion resulted from the dissociation of the Tb-DPA complex when it leaked into the bulk phase containing

855 E F F E C T OF PC ON THE a-LA-INDUCED VESICLE FUSION

EDTA and from the leak-in of outside solution into the fused vesicles. Protein binding to vesicles. The centrifugation method was used for the binding studies. After incubation of a-LA with vesicles at pH 4 (82 mM acetate, 18 mM Na-acetate, 41 mM NaC1), the vesicles were sedimented and the concentration of unbound a-LA in the supernatant was determined. Bound a-LA was estimated by a comparison of the protein concentration in the original solution with the remaining concentration in the supernatant. For these experiments it was essential that there was no sedimentation equilibrium redistribution of a-LA, and that all the vesicles were sedimented under centrifugation. The vesicles prepared at pH 4 were first centrifuged at 6,800 rpm to remove heavy liposomes. The vesicles in the supernatant were then sedimented eight times with progressively increasing angular velocity from 20,000 to 34,000 rpm, each operation lasting from 60-90 min, using a Beckman SW 41 swinging-bucket rotor. In this way, the size range of the vesicles was narrowed. After the final sedimentation, the supernatant was found to be optically clear as observed by the OD at 280 nm. Phosphate analysis confirmed that the phospholipid concentration of the supernatant was negligible. The last vesicle sediment was resuspended in a buffer solution, and an aliquot of a-LA solution was added to bring the concentration of both phospholipid and protein to predetermined values. After incubation at 18~ for 30 min, the vesicles were centerfuged at 35,000 rpm, and the supernatant was analyzed for the protein concentration by measuring OD values at 280 nm. Preliminary tests showed that at least 15 min of incubation were required prior to ultracentrifugation in order to reach binding equilibria. To test the reversibility of the binding, the vesicles were incubated with a-LA at pH 4 at 18~ for 30 min, and centrifuged for 90 min at 35,000 rpm. The sediments were then resuspended in a pH 7 (2 mM Tes, 2 mM L-histidine, 100 mM NaCl) buffer solution and incubated for 48 hr at 18~ The pH of this suspension remained constant at seven throughout this period. The suspension was centrifuged for 90 min at 35,000 rpm, and the protein concentration in the supernatant was determined (OD at 280 nm}. Proteolytic digestion of vesicle-bound proteins. The possibility of a segment of these proteins being inserted into the phospholipid vesicle membrane was checked by the treatment of the vesicle-protein complex with trypsin. In these experiments, 2 mg of protein were incubated with 10 ml of vesicles suspension (1 mM Pi} for 60 min at 18~ at pH 4. The vesicles with bound protein were sedimented by centrifuging for 90 min at 30,000 rpm using a Beckman SW 41 rotor. The pellets were resuspended in 5 ml of medium containing 150 mM KC1, 25 mM imidazole, and 100 ~g of proteolytic enzyme, pH 7.5, and then incubated for one hr at 37~ The digestion was stopped by the addition of PMSF as a freshly dissolved ethanolic solution to 3 mM final concentration followed by incubation of samples at 37~ for three min. In order to cleave -S-S linkages, f3-mercaptoethanol was added to the sample solution at a 1% final concentration and incubated for 30 min at 18~ The vesicles were then

pelleted by centrifugation for 90 min at 30,000 rpm in a Beckman SW 41 rotor. The pellets were resuspended in 2 ml of buffer {25 mM imidazole, 1 mM EDTA; pH 7.5). For molecular weight determination of the protected protein portion, the lipid was extracted in three volumes of chloroform-methanol (2:1) and the protein was precipitated in 10% trichloroacetic acid at 0~ followed by centrifugation for 15 min in an Eppendorf microfuge. After two acetone washings, part of the resulting pellet of the protein fragment was solubilized in a buffer of 40% glycerol, 2.5% SDS, 0.01 M H3POJ Tris, 8 M urea, and 5% ~-mercaptoethanol (pH 6.7}. The dissolved protein was electrophoresed on 15% acrylamide gels in SDS. The remaining portion of the pellet was lyophilized. Amino acid sequencing of the extracted sample was done by the semi-micro manual Edman degradation method. Because of difficulties with the manual method, only 10 N-terminal residues of the segment were determined. RESULTS AND DISCUSSION

The time-course of fusion for all the vesicle compositions studied here was similar to that described earlier (1, Fig. 1). The initial increase in the percent MAX fluorescence intensity due to fusion was followed by the decrease in the percent MAX fluorescence intensity caused by the leakage of vesicle content. The percent MAX fluorescence is defined as the fluorescence intensity measured during fusion divided by the fluorescence intensity after the vesicle-entrapped components were mixed following detergent treatment. The time-course of the percent MAX fluorescence, when corrected for leakage according to the method of Bentz et al. (12), approached a plateau. Figure 1 shows the

UJ

uJ 4

3

)

2

3

4

5

A

A

6

7

pH FIG. I. pH dependence of initial rate of fusion of three types of vesicles. PSIPE (1:1) (~~), PS/PC (1:1) ( l - - - - - l ) , and PC/PE (1:1) (&---A). Fusion was initiated by adding a-LA solution to make up a final concentration of 100 ~g/ml (18~ Buffer solutions used are: pH 2 (5 mM KCI, 18 mM HCI, 43.4 mM NaCI); pH 3 (50 mM glycine, 11 mM HCI, 48 mM NaCI); pH 4 (82 mM acetate, 18 mM Na-acetate, 41 mM NaCI); pH 5 (14 mM acetate, 86 mM Na-acetate, 7 mM NaCI); pH 6 (2 mM Mes, 80 mM NaCI); pH 7 (2 mM Tes, 2 mM L-histidine, 100 mM NaCI). UPIDS,Vol. 24, No. 10 (1989)

856 B.S. PARK E T AL. initial rate of percent m a x i m u m fluorescence increase during the fusion of three kinds of vesicles as a function of pH. The fusion-pH profiles of these vesicles are similar, and the initial rates of fusion increase with decreasing pH. The initial rate of fusion of P S / P E vesicles is about twice that of PS/PC vesicles, and the initial rate of fusion of neutral PE/PC vesicles is much lower. In order to follow in detail the change in fusion behavior progressing from PS/PE to PS/PC vesicles, P E in the vesicles was gradually replaced by PC while m a i n t a i n i n g PS constant. The time-course of a-LAinduced fusion of these vesicles after correcting for leakage (12) is shown in Figure 2. It can be seen that the fusion is complete within about two min. Figure 3

gives the initial rate of fusion plotted against PC content. I t is of interest that the curve is sigmoidal. Figure 4 gives the concentration dependent a-LA binding to the PS/PE (1:1), PS/PC (1:1) and PC/PE (1:1) vesicles. Essentially 100% of a-LA binds to both PS/ PE (1:1) and PS/PC (1:1) vesicles regardless of the total a m o u n t of a-LA added. However, the amount of a-LA bound to PC/PE (1:1) vesicles is about 50% less than that bound to the other vesicles. These results suggest t h a t simple binding per se has little bearing on fusion. The data points contain two components: reversible binding and irreversible binding. The irreversible portion of binding can be estimated by bringing the incubation pH to 7, where there is no initial binding, and by then determining the amount of a-LA still bound to the vesicles. Figure 5 gives the amount of irrevers-

80

.~, ._,60!

~

4

W

.

-c

40

- d

"=

-e

20

gf 0

2 TIME

3

(min)

0 FIG. 2. Time course of fusion of vesicles composed of combinations of PS, PE and PC. Corrections for leakage were made according to the method of Bentz et al. (12). The vesicles are (a) PS/PE (I:I), (b) PSIPE/PC (5:4:1), (e) PS/PE/PC (I0:7:3), (d) PS/PE/ PC (5:3:2), (e) PS/PEIPC (5:2:3), (f) PS/PE/PC (5:1:4), and (g) PS/PC (I:I). At time zero, 0.I ml of 1 mg/ml a-LA solution was added into the spectrofluorometer cuvette which contains 0.9 ml of vesicle suspension {0.05mM Pi; 18~ pH 4.0).

10

20

I

I

I

30

40

50

PC CONTENT(% ) FIG. 3. Effect of PC content on o-LA-inducedvesicle fusion. The experimental conditions are the same as in Figure 2. Data were obtained by replotting Figure 2.

1000

,<

750 ~.

6(

<

_1

z

i

40

0 z

250

~ 2o I

250

I

i

l

500 750 1000 ADDED o-LA(/Jg)

FIG. 4. Binding isotherms (18~ of a-LA binding to three types of vesicles at various total protein concentrations. The vesicles are PS/PE (1:1) (~------e), PS/PC (1:1) (D---~), and PC/PE (1:1) ('_ _').

LIPIDS,VoL24, No, 10 (1989)

0

I

10

I

I

I

20 30 40 PC CONTENT( %)

i

50

FIG. 5. The amount of irreversibly bound a-LA as a function of PC content. The compositions of the vesicles are the same as in Figure 2. ( ~ ), total binding. (~-------e), the amount of irreversibly bound a-LA.

857

EFFECT OF PC ON THE a-LA-INDUCED VESICLE FUSION ibly bound a-LA plotted against PC content of the vesicles as the P E is slowly replaced b y PC, while keeping the PS content constant. The curve is strikingly similar to the fusion profile shown in Figure 3. This strongly s u g g e s t s t h a t the inhibitory effect of PC on fusion is caused b y a decrease in irreversible binding. Figure 6 shows the S D S - P A G E b a n d s of the a-LA s e g m e n t which had been protected from tryptic digestion. a-LA was first incubated with vesicles of different P E / P C ratios at p H 4 and then treated with t r y p s i n at p H 7.5 at 37~ The lipids were e x t r a c t e d with chloroform-methanol, and the protein s e g m e n t was prec i p i t a t e d w i t h 10% trichloroacetic acid prior to the S D S - P A G E . A protein band with an e s t i m a t e d molecular weight of approximately 4,000 was observed for each vesicle system. B a n d intensities became stronger with increasing PE content and corresponded to the a m o u n t of a-LA s e g m e n t protected from proteolytic digestion. I t is likely t h a t a portion of this Mr 4,000 unit is buried within the hydrophobic core of the vesicle bilayer r e g a r d l e s s of p h o s p h o l i p i d r a t i o s in the PS/PE/PC vesicles. The amino acid sequence of the first 10 residues of this unit for all the phospholipid compositions was found to be the 80--89 stretch of a-LA. This, together with the observation t h a t the Mr of all the s e g m e n t s were similar to the s e g m e n t obtained previously (1), strongly s u g g e s t s t h a t the 80-108 s e g m e n t of a-LA was protected from the tryptic digestion in all vesicles. The result s u g g e s t s t h a t the population of s e g m e n t p e n e t r a t i n g the bilayer is a k e y factor in bringing a b o u t fusion. Table 1 shows the effect of phospholipids on a-LAinduced fusion of vesicles containing b o t h acidic lipids and PC, and shows the a m o u n t of total binding and irreversible binding of a-LA to t h e s e vesicles. H e r e again, the extent of fusion and binding gradually decreases as the PC content increases. I t was reported t h a t PC has an inhibitory effect on Cae+-induced fusion while P E h a s an e n h a n c i n g

TABLE 1 Effect of Phospholipids on Fusion, Binding, and Irreversible Binding

Percent Irreversible Fusion {%)a Binding {%) Binding (%) -20 • 5 3- 1 -37 - 4 4_ 1 1.5 +- 0.5 75 • 5 11 • 2 15 • 94--4 19+- 2 2 + 1 40--3 4• 1 PC/CLI5:I) 16 • 100• 20• PC/CL{3:I) 42 • 100• 35• aThe time-course of the percent MAX fluorescence, when corrected for leakage according to the method of Bentz et aL 112), approached plateaus and these limiting values were taken as percent fusion.

Vesicles PC PC/PS (9:1) PC/PS (4:1~ PC/PS(I:I} PC/CL{9:I)

effect (8). These c o n t r a s t i n g effects were a t t r i b u t e d to differences in p h o s p h o l i p i d d e h y d r a t i o n b y Ca 2+. I t was s u g g e s t e d t h a t Ca 2+ d e h y d r a t e s the PE head group more readily t h a n the PC head group, and thus would facilitate the approach of P E vesicles and vesicle fusion (8,13,14}. Although the differential effect on dehydration m a y also be i m p o r t a n t for a-LA-induced fusion, it is clear t h a t differences in a-LA binding behavior b y these vesicles are a major reason for the discrepancies observed. The a-LA s e g m e n t s which penetrate the bilayer are identical regardless of whether or not the vesicles contain PC or PE. Thus, it appears t h a t observed differences in fusion properties depend primarily on the extent b y which a particular a-LA s e g m e n t p e n e t r a t e s the bilayer. The reason why more a-LA binds to P S / P E vesicles t h a n to PS/PC vesicles is not altogether clear. PC and PE have similar charges within the p H range studied and the orientation of the head groups of these neutral phospholipids is parallel to the bilayer surface. The main differences between these head groups are their size and degree of hydration. PC has a larger head group with a higher affinity for w a t e r (15). I t is possible t h a t the PE head groups create less of a barrier for the a-LA s e g m e n t s to p a s s into the bilayer interior. Also, PE has a propensity for forming hexagonal H H structures. W h e t h e r or not this is an i m p o r t a n t factor in the present context is not clear as yet.

ACKNOWLEDGMENT The financial support by the Korea Science and Engineering Foundation is gratefully acknowledged.

FIG. 6. SDS-PAGE patterns of the a-LA segment which had been protected in the a-LA/vesicle suspension from tryps'm digestion. See text for experimental details. Lane A: PS/PEIPC {5:4:1). Lane B: PS/PE/PC (5:3:2). Lane C: (PS/PEIPC (5:1:4}.

REFERENCES 1. Homma, M., and Ohuchi, M. (1973)J. Virol. 12, 1457-1465. 2. Perez, H., Toister, Z., Laster, Y., and Loyter, A. (1974) J. Cell BioL 63, 1-11. 3. Lucy, J.A. {1984} FEBS Lett. 166, 223-231. 4. Montecucco, C., Schiavo, G., Brunner, J., Dutlot, E., Boquet, P., and Roa, M. (1986) Biochemistry 25, 919-924. 5. Kim, J., and Kim, H. {1986}Biochemistry 25, 7867-7874. 6. Kim, J., and Kim, H. {1989) Biochirn. Biophys. Acta 983, 1-8. 7. Papahadjopoulos, D., Poste, G., Schaeffer, B.E., and Vail, W.J. {1974}Biochim. Biophys. Acta 352, 10-28. LIPIDS,Vol. 24, No. 10 (1989)

858 B.S. PARK E T AL. 8. Papahadjopoulos, D., Vail, W.J., Newton, C., Nir, S., Jacobson, K., Poste, G., and Lazo, R. (1977) Biochim. Biophys. Acta 465, 579-598. 9. Vaskowsky, R.E., Kostetsky, E.Y., and Vasendin, I.M. (1975) J. Chromatogr. 114, 129-141. 10. Deamer, D., and Bangham, A.D. (1976) Biochim. Biophys. Acta 443, 629-634. 11. Wilschut, J., Duzgunes, N., Fraley, R., and Papahadjopoulos, D. (1980} Biochemistry 19, 6011-6021. 12. Bentz, J., Duzgunes, N., and Nir, S. {1983) Biochemistry 22, 3320-3330.

LIPIDS,Vol. 24, No. 10 (1989)

13. Portis, A., Newton, C., Pangborn, W., and Papahadjopoulos, D. (1979) Biochemistry 18, 780-790. 14. Papahadjopoulos, D., Portis, A., and Pangborn, W. (1978) Ann. N. Y. Acad Sci. 308, 50-66. 15. Cevc, G., and Marsh, D. (1987) Phospholipid Bilayers, Chapter 3, Wiley-Interscience, New York.

[Received February 13, 1989; Revision accepted August 2, 1989]