Chemistry and Physics of Lipids, 57 (1991) 103---107 Elsevier ScientificPublishers Ireland Ltd.

103

Short Communication

Effect of membrane phase transition on long-time calcium-induced fusion of phosphatidylserine vesicles S.O. Leung and J.T. Ho Department of Physics and Astronomy, State University of New York at Buffalo, Buffalo, NY 14260 (U.S.A.)

(Received June 6th, 1990; revision receivedAugust 30th, 1990; accepted October 30th, 1990) Dynamic light scattering has been used to study the temperature dependenceof the extent of long-timecalcium-inducedfusion of sonicated vesiclescomposedof various natural and syntheticphosphatidylserinewith differentacyl chains. The vesiclesof each composition are found to exhibit a peak temperature in the vicinityof which the extent of fusion shows a distinct maximum. The fusion peak temperatureincreasesas the bilayergel-to-liquid-crystalphase transition temperatureincreases.The results suggest~ role played by membrane fluidity in determining fusion efficiency. Keywords: dynamic light scattering; fusion; calcium; phosphatidylserine;phase transition; vesicles.

Introduction Phospholipid vesicles have been widely used as model systems for studying membrane fusion [1]. Of particular interest is the role played by divalent cations, especially Ca 2+, in promoting the fusion of acidic phospholipid vesicles, such as phosphatidylserine, because of the possible relation to such cellular functions as exocytosis. Recent research emphasis has been placed on the examination of the initial rate o f fusion as measured by the formation of a fluorescent complex resulting from the mixing of the aqueous contents o f the fusing vesicles [2]. These studies have revealed that, as far as the initial fusion rate is concerned, a cationinduced phase transition of the bilayer lipids, which is generally expected to promote membrane instability and hence fusion, appears not to be a prerequisite for the fusion process [3,4]. It seems Correspondence to: S.O. Leung, Department of Physics and Astronomy, State Universityof New York at Buffalo,Buffalo, NY 14260, U.S.A.

important, however, that the vesicles are in an overall fluid state when the cations bind to their outside surface prior to fusion [3]. In a series of earlier experiments using dynamic light scattering to measure the resultant vesicle size after prolonged incubaton of sonicated unilameUar vesicles ( s u e ) in Ca 2+ followed by chelation with ethylenediaminetetraacetic acid (EDTA), it was found that the extent of long-term fusion thus measured showed an intriguing temperature dependence for vesicles with either bovine brain phosphatidylserine (PS) or PS mixed with zwitterionic phosphatidylcholine: there exists a temperature at which the fusion efficiency appears to show a clear maximum [5,6]. It was speculated that this temperature dependence of the long-term fusion efficiency might be related to the gel-toliquid-crystal phase transition o f the lipid bilayer in the presence of Ca 2+ prior to vesicle aggregation or fusion, but the evidence was somewhat preliminary. Irrespective o f the validity o f this explanation, further exploration of this phenomenon is important to our understanding of the effect of

0009-3084/91/$03.50 © 1991ElsevierScientificPublishers lreland Ltd. Published and Printed in Ireland

104 Ca 2+ on acidic membranes and its physiological significance. This unusual long-time aggregation and fusion process is also an interesting problem from the point of view of colloidal science. In an attempt to see whether the previous observations in PS are unique and to obtain additional insight into this puzzling problem, we have examined the temperature dependence of the long-time Ca2+-induced fusion of vesicles made up of natural and synthetic phosphatidylserine of various acylchain compositions and thus with different phase transition temperatures. Our results have revealed an interesting correlation between the fusion peak temperature and the bilayer phase transition temperature which points to a possible role played by membrane fluidity in promoting fusion. Experimental procedure Bovine brain PS, hydrogenated bovine brain phosphatidylserine (HPS) and 1,2-dimyristoyl-snglycero-3-phosphatidylserine (DMPS) (> 99% pure) were obtained from Avanti Polar Lipids. One of these lipids or their mixtures were suspended by mechanically shaking in a buffer containing 100 mM NaCI, 2 mM L-histidine, 2 mM N-Tris-(hydroxymethyl)methyl-2-aminoethanesulfonic acid and 0.1 mM EDTA adjusted to pH 7.4. The suspension was then sonicated in a Bransonic 185 Tip Sonifier for 30 rain. The sample was kept under nitrogen at all times. The temperature of the sample during mixing and sonication was regulated at several degrees above the lipid bilayer phase transition temperature. The sonicated sample was centrifuged at 40,000 × g for 20 min. The supernatant after centrifugation was diluted with buffer to a concentration of about 1 mM of lipid and used in the fusion studies. The aggregation and fusion of an SUV sample, initiated by mixing 0.5 ml of sample with 0.5 ml of buffer containing 1.8 mM of CaCI 2, was conducted with the sample tube in a water bath set at a chosen temperature. The typical incubation time to ensure the completion of the fusion process was 2.5 h, as determined by monitoring the turbidity kinetics following calcium addition. In a separate study, the turbidity increase was found to correlate well with the irreversible change in particle size as measured by dynamic light scattering. At the end of the in-

cubation period, 0.1 ral of 0.1 M EDTA in buffer was injected into the sample. The chelation process was generally completed after 15 rain, as indicated by the stabilization of the final vesicle size. The average vesicle size before and after incubation in calcium and EDTA was measured using dynamic light scattering. The initial average vesicle size was typically 50 nm, while the size after calcium addition and that after EDTA addition were typically several microns, depending on the lipid composition and temperature. Dynamic light scattering was performed using a Spectra-Physics 144 He-Ne laser as the light source and an ITT FW-130 photomultipfier tube biased by an Ortec 456 high-voltage power supply as the detector for light scattered at 60 °. The photoelectronic pulses from the photomultiplier tube, after being amplified and noise discriminated, were analyzed by a 64-channel Langley-Ford digital correlator. The time required to obtain a set of correlation function data, typically 5 to 15 rain depending on the scattering intensity, is not an important factor since all samples studied are in equilibrium. A second-degree cumulant analysis of the time-correlation function of the scattered intensity was performed on a CDC Cyber 173 computer to obtain the weighted average hydrodynamic diameter of the vesicles [5]. Turbidity measurements were made by detecting the transmitted light with a Spectra-Physics 385 photodiode, the output of which was recorded on a Houston 5233-50mniscribe strip-chart recorder. Results and Discussion We used the average vesicle size after incubation of SUV in Ca 2+ and subsequent chelation with EDTA, which measures the irreversible morphological change induced by Ca 2+ as an indicator of the extent of long-time vesicle fusion. An examination of the effect of Ca 2+ concentration was first made and it was found that the fusion process saturates at about 0.85 mM Ca 2+. An effective Ca 2+ concentration of 0.9 mM was therefore chosen for the studies on the effect of the incubation temperature. The first studies involved SUV made up of three separate types of lipids, namely, PS, DMPS and HPS. The results on the effect of incubation

105

temperature on the extent of fusion using PS S U V (data not shown) confirm the observation reported earlier [5]: as a function of increasing incubation temperature, the final vesicle size first increases and then decreases, with a clear maximum occurring at around 11 °C. Examples of new results with DMPS and HPS SUV are shown in Figs. 1 and 2. The fluctuations in the temperature dependence of the final average vesicle size are larger than the experimental accuracy and therefore represent the effect of sample variability. However, the existence of a peak is unmistakable and its location can be determined to within 2°C in repeated experiments. Thus the long-time fusion of both types of SUV also shows temperature dependences similar to that of PS, except that the maximum extent of fusion occurs at different temperatures, i.e., at 24°C for DMPS and at 38°C for HPS. The fact that DMPS and HPS each exhibit a clear fusion peak temperature is a new result indicating for the first time that the earlier observations in PS are not unique to that system. The observation that

I

HPS 8 A

E rr uJ I-hi

6

o

o oO

J L~ (/1 IJ.I

o°

°

uJ 4

> /

z o

2

o o o °

oO

o

oo

0

le I0

30

I

I 50

I

I 70

INCUBATION T E M P E R A T U R E (°C)

Fig. 2. Final average vesicle size as a function of incubation temperature for sonicated HIS vesicles in 0.9 mM calcium. I

I

I

I

I

I

_

I

DMPS

E 2 ::L

o o

n." UJ

I-

oo

UJ o

F,

° o

UJ

._1

ca UJ >

o

o ° °

°

I

J

z o

o

o •

I 10

o

°

I

I 20

INCUBATION

I

I 30

I

TEMPERATURE

I 40 I°C)

Fig. I. Final average vesicle size as a function of incubation temperature for sonicated DMPS vesicles in 0.9 mM calcium.

the fusion peak temperature is different and progressively higher for PS, DMPS and HPS is intriguing. Because of the earlier speculation about a connection of the fusion process with the bilayer phase transition, it is interesting to observe that the gel-to-liquid-crystal transition temperature of PS, DMPS and HPS vesicles is alsoprogressively higher. For PS and DMPS, the published transition temperatures are 7°C [7] and 39°C [8], respectively. For HPS, differential scanning calorimetry has been performed by Dr. Philip S. Low of Purdue University, who obtained a distinct peak at 63°C (private communication). Thus an interesting correlation between the fusion peak temperature and the phase transition temperature emerges when one is plotted against the other, as shown in Fig. 3. There appears to be an almost linear relationship between these two quantities for the three types of lipids. To test the role of the bilayer phase transition further in affecting the fusion efficiency, we have also

106

50

I1-

hi (3-

I

I

I

I

I

I

40

30

Ld

~-

2O Ops

u.I Q_

~"

P

I0

0 ~q h

I

I

I

I

I

I

I0

20

30

40

50

60

70

PHASE TRANSITION TEMPERATURE (%)

Fig. 3. Fusion peak temperature against bilayer phase transition temperature for the three acidic membranes studied.

studied the temperature dependence of the final vesicle size after long-time Ca 2+ incubation and EDTA chelation of SUV made up of five different binary mixtures of PS and HPS. We find that each binary lipid system also shows a maximum in the temperature dependence of the extent of fusion similar to that of the individual components. However, the fusion peak temperature increases as the proportion of HPS in the mixture increases. This

4O

I

I

I

I

I

I

I

I

I

I

I

uJ

3c Q_

m 2(? Z 0

t

i.u

I0 I

0

I

I

20

I

I

40

I

% of H P S i n

I

60

I

I

80

I

[ I00

PS

Fig. 4. Fusion peak temperature of vesicles o f PS-HPS mixtures as a function o f the percentage of HPS in PS.

is seen most dearly in Fig. 4, which shows an almost linear dependence of the fusion peak temperature on the percentage of HPS in the mixture. Since the phase transition temperature of a PS-HPS mixture is expected to increase monotically with the HPS concentration, this result again points to a direct relationship between the fusion peak temperature and the bilayer phase transition temperature. Several conclusions can be drawn from this work. First, the unexpected peak in the temperature dependence of fusion found earlier in PS is not unique but, in fact, appears to be quite universal, being present in DMPS, in HPS and in PS-HPS mixtures. Second, the fusion peak temperature is not equal to the bilayer phase transition temperature, with or without Ca 2+. However, an intriguing correlation seems to exist between these two quantities. Third, the fusion peak temperature in PS-HPS mixtures is very sensitive to the HPS content. Since changing the concentration of HPS in PS mainly changes the phase transition temperature of the membrane, the fusion peak is clearly affected by changes in membrane fluidity. From our previous results on PS alone, it was not clear whether the existence of the fusion peak was accidental and the possible relation of the peak temperature to any intrinsic membrane property was only speculated and not proven. In this new study, we have demonstrated for the first time that the calcium-induced fusion peak temperature is dependent on the acidic lipid composition and shows an indirect correlation with the bilayer phase transition temperature of the membrane. It should be emphasized that our experimental approach allows us to study only the long-time fusion of SUV. Thus our results are not readily comparable with those on the initial fusion rate using either SUV or large unilamellar vesicles [3,41 and are not necessarily in contradiction with those experiments. In fact, our results might also be governed in part by the effect of temperature on the production of vesicles from Ca2+-induced dehydrated cochleates upon the addition of EDTA [9]. In any case, this interesting correlation we have found does provide additional suggestive evidence about the important role played by membrane fluidity in governing the efficiency of membrane fusion, but its exact interpretation would require additional systematic studies.

107

Acknowledgements We would like to thank Dr. Philip S. Low for sharing with us his differential scanning calorimetry data. This work was supported by the National Institutes of Health (Grant No. GM-24590) and the National Science Foundation (Grant No. DMR-8717614). References

3 4 5 6 7 8

1 J. Wilschut and D. Hcekstra (1986) Chem. Phys. Lipids 40, 145--166. 2 N. Duzgunes, J. Wilschut and D. Papahadjopouios (1985) in: F. Conti et al. (Eds.) Physical Methods on

9

Biological Membranes and Their Model Systems, Plenum Press, l~¢w York, pp. 195--286. J. Wilschut, N. Duzgunes, D. Hoektra and D. Papahadjopouios (1985) Biochemistry 24, 8--14. J. Bentz, N. Duzgunes and S. Nit (1985) Biochemistry 24, 1064---1072. S.T. Sun, E.P. Day and J.T. Ho (1978) Proc. Natl. Acad. Sci. U.S.AI 75, 4325--4328. S.T. Sun, C.C. Hsang, E.P. Day and J.T. Ho (1979) Biochim. Biophys. Acta 557, 45--~52. K. Jacobson and D. Papahadjopoulos (1975) Biochemistry 14, 152--161. H. Hanser, F. Paltauf and G.G. Shipley (1982) Biochemistry 22, 1061--1067. D. Papahadjopoulos, W.J. Vail, K. Jacobson and G. Poste (1975) Biochim. Biophys. Acta 394, 483--491.