- Email: [email protected]
Interaction of dextran sulfate with phospholipid surfaces and liposome aggregation and fusion
Chemistry and Physics of Lipids, 55 (1990) 301--307
301
Elsevier ScientificPublishers Ireland Ltd.
Interaction of dextran sulfate with phospholipid surfaces and liposome aggregation and fusion K. Arnold', S. O h k i b a n d M . K r u m b i e g e P "Institute of Biophysics, Medical Center, Karl Marx University, Leipzig (Germanyj and ~Department of Biophysical Sciences, Medical School, SUN)', Buffalo (U.S.A.) (Received February 9th, 1990; revision received June 5th, 1990; accepted June 6th, 1990)
The binding of dextran sulfate to phospholipid liposomes was investigated by microelectrophoresis experiments. The polyanion binds to neutral phospholipid liposomes (DMPC and PE) only in the presence of Ca 2". If positively charged stearylamine is incorporated in the vesicles dextran sulfate is bound without Ca 2". Negatively charged phospholipids as PS do not bind dextran sulfate, even in the presence of millimolar concentrations of Ca ~'. The adsorption of dextran sulfate results in an aggregation of vesicles due to a bridging mechanism. In all cases the aggregation is followed by a disaggregation toward higher dextran sulfate concentrations. The disaggregation process starts at polymer concentrations smaller than the concentration of the onset of saturation of the adsorption. By use of the probe dilution method a fusion of small DMPC and DMPC/PE vesicles in the presence of Ca 2" and dextran sulfate was found.
Keywords: phospholipid vesicles; dextran sulfate; aggregation; fusion; electrophoresis; turbidity; fluorescence.
Introduction The formation of soluble and insoluble associations between plasma lipoproteins and glycosaminoglycans (GAG) has been studied for more than 25 years and it was found that they play a possible role in the depositions of lipoproteins in the arterial intima [1]. The usual explanation of the interaction assumes an ionic interaction between positively charged amino acids of the apolipoprotein B of LDL and the negatively charged GAG molecules [2,3]. Because phospholipid vesicles are also precipitated by heparincalcium mixtures, it was concluded that the interaction with lipoproteins can be also realized by calcium bridges between the phospholipid phosphate groups and the sulfate groups of GAG [4]. An involvement of phospholipids in the interaction of LDL with GAG was demon-
Correspondence to: K. Arnold.
strated very recently [5] and the importance of positively charged amino groups on the LDL surface was shown [6]. Dextran sulfate has a structure very similar to glycosaminoglycans and this polymer was used instead of the glycosaminoglycans in many investigations. The Interest in this polymer increased because it was used as an anti-atherosclerotic drug [7] and potent agent against HIV infection [8] and the molecular mechanisms of these actions were not explained. The purpose of the present investigation was to study the adsorption of dextran sulfate on phospholipid surfaces by use of measurements of the electrophoretic mobility and the effect on vesicle fusion and aggregation. The influence of lipid composition, NaCI and C a 2÷ o n the interaction of dextran sulfate with phospholipids is studied. Especially the study of the influence of Ca z" on the interaction represents a very relevant problem because of the relatively high extracellular Ca 2" concentrations.
0009-3084/90/$03.50 © 1990 Elsevier Scientific Publishers Ireland Ltd. Published and Printed in Ireland
302 H
Materials and Methods
3
Dimyristoylphosphatidylcholine (DMPC) from Serva, F.R.G. and egg-phosphatidylserine (PS) and egg-phosphatidylethanolamine (PS) from Avanti Polar Lipids, U.S.A. were used. Multilamellar vesicles (MLV) were prepared using the method of Bangham et al. [9]. For the preparation of small unilamellar vesicles a bathtype sonicator was applied at a frequency of 20 kHz. Optically clear suspensions were obtained. The dextran sulfate of average molecular weight of 500,000 (DS 500) was purchased from Serva, F.R.G. The fusion of vesicles was monitored by the probe dilution method which makes use of the resonance energy transfer between nitrobenzooxa-diazole (NBD) and rhodamine (Rh) covalently attached to the head group of phosphatidylethanolamine [10]. The corresponding fluorescent phospholipids N-(7-nitro-2,1,3benzo-oxa-diazole-4-yl)-PE (NBD-PE) and N(lissamine Rhodarnin B sulfonyl)dioleoyl-PE (Rh-PE) were purchased from Molecular Probes, U.S.A.. Fluorescence spectra were measured by use of the spectrofluorimeter Perkin-Elmer LS5 and the fluorescence was excited at 475 nm and the spectra were recorded between 490 and 600 am. The size of the vesicles was determined from the dynamic light scattering by use of the submicron particle analyzer Coulter Model N4S (Coulter Electronics, U.S.A.). A Parmoquant-2 (Carl Zeiss Jena, G.D.R.) operating in the dark field mode was used for the measurement of the electrophoretic mobility of MLV. In each case the device automatically calculates mean values and standard deviations of the electrophoretic mobility of a set of 100 particles. The zeta potential was calculated according to the Smolukhovski equation. The viscosity of the solution was measured with a Hoeppler viscosimeter. Results
Electrophoresis measurements
The electrophoretic mobility of pure phospha-
•
....
~.JL--'-,=|
"o_o
-2 >~,m o
"1
u~
:" W -J hi
o2 "H
o
amd DEXTRAN SULPHATE m g / m l
Fig. 1. Electrophoretic mobility of muhilamellar D M P C vesicles in the presence o f 3 m M Ca 2" (O) and D M P C vesicles containing different mole fractions o f stearylamine (A, 5 tool %; n , I0 tool %; o , 20 tool %) as a function o f dextran sulfate concentration. The vesicles were prepared in a buffer solution of I0 m M Tris, 145 m M NaCI, 0.3 m M EDTA, pH 7.4. The total lipid concentration was 0.I rag/ ml. The measurements were performed at a temperature below T (20°C).
tidylcholine (DMPC) multilamellar vesicles which are uncharged and do not move in the electric field is not changed on addition of dextran sulfate. A completely different behaviour is measured for DMPC vesicles which contain 5, 10 and 20 mol~0 stearylamine (Fig. l). These vesicles have a positive electrophoretic mobility in the absence of DS 500 and the magnitude of the mobility is dependent on the mole fraction of the positively charged stearylamine. Addition of DS 500 results in a reversal of the direction of the electrophoretic motion of the vesicles. The electrophoretic mobility is increased for higher concentrations of DS 500 and a plateau is reached for DS 500 concentrations higher than 0.01 mg/ml. The plateau values depend on the stearylamine concentration in the vesicles to a very small extent. These experiments demonstrate the adsorption of the polyanionic DS 500 on the vesicle surface. Negative charge is accumulated due to the binding and overcomes the positive charge of the vesicle surface which
303
determined the electrophoretic mobility prior to
the ad.sorption of the polymer. The eleetrophoretic mobility of DMPC MLV as a function o f the concentration o f DS 500 in the presence o f 3 mM Ca 2° is also given in Fig. 1. The electrophoretic mobility of the particles is positive in the absence o f DS 500 due to a small adsorption of Ca ~°. On addition o f DS 500 a similar behaviour as described above was observed.
,
Turbidity measurements A very strong aggregation occurs for the positively charged liposomes consisting o f mixtures o f DMPC and stearylamine. Upon addition of DS 500 such large particles are formed that a fast precipitation appears and turbidity measurements cannot be usefully applied. A systematic study of the influence of dextran sulfate, Ca 2° and NaCI on the aggregation of unilarnellar DMPC and D M P C / P E liposomes was performed. The aggregation process was monitored by measurements o f the turbidity and the sizes of vesicles were determined by dynamic light scattering. Dextran sulfate was subsequently added in a concentration range from 0.005 to 5 mg/ml (10 nM--10/aM).
0
'~
o oo,
o
"~, o',
,o
"
01 10 OEXTRAN SULPHATE, n~/rr~
tOO
Fig. 3. Particle size of sonicated DMPC/eBB-PE vesicles (10 tool% egg-PE) as a function of dextran sulfate concentration for 0 ( + ) , 50 ( I ) , 100 (&), 150 ( e ) and 200 (O) mM NaCI in the presence of 2.0 mM CaCI 2 and 10raM HEPES at pH 7.4 and above T (30°C). Dextran sulfate was added at time intervals of 3 rain. A phospholipid concentration of 0.3 rag/ ml was used.
10
$
O0l
\ ,0o
OEXTRAN SULPHATE500. mg/ml Fig. 2. Turbidity of sonicated DMPC/egg-PE vesicles (10 tool% egg-PE) as a function of dextran sulfate concentration for 0 ( + ) , 50 ( 8 ) , 100 (&), 150 (O) and 200 (O) mM NaC! in the presence of 2.0 mM CaCI 2 and 10 mM HEPES at pH 7.4 and above 7", (30°C). Dextran sulfate was added at time intervals of 3 min. A phospholipid concentration of 0.3 rag/ ml was used.
As representative examples the change o f the turbidity (Fig. 2) and the size (Fig. 3) of the D M P C / P E liposomes is given as a function of DS 500 concentration for 2 mM Ca 2" and different NaCI concentrations. In all experiments a maximum of the turbidity is observed. The maximal turbidity is a function o f the NaCI concentration and it decreases with higher NaCI concentrations. For a Ca2"/Na ÷ ratio of about 1 : 100 aggregation vanishes. The particle sizes of the phospholipid vesicles determined by dynamic light scattering are given in Fig. 3 for different NaCI concentrations as a function o f dextran sulfate. The same samples whose turbidity was given in Fig. 2 were used. The particle size increases and has a maximum at about the same concentration of dextran sulfate where the turbidity maximum was observed. In the absence of NaCI the increase of the particle size from about 55 nm to about 500 nm occurs at the lowest concentration of dextran sulfate and is unchanged over the whole range of concentrations. A decrease of the particle sizes appears for all other NaCI concentrations after passing the size maximum. However, the
304
1,, 2 . 3
2(:
/
/
>-
_o m (Z I--
/-,
~-- 05 <
/
.-/
w
e
~e
06
10
1/.
~8
"-
-
Fluorescence measurements
/
to
w
because of the strong repulsion of the negatively charged surfaces of the liposomes and the dextran sulfate molecules. At the concentrations of Ca 2. for aggregation of PS liposomes the surface potential is still negative (about 30 mV [11]).
2"2
CO 2" C O N C E N T R A T I O N , mM Fig. 4. Turbidity o f sonicated phosphatidylserin¢ vesicles as a function o f the Ca 2" concentration in the presence o f 0 mg/ ml (1), 0.075 m $ / m l (2), 0.6 m g / m l (3) and 3.1 m g / m l (4) dextran sulfate at 27°C. The samples were prepared in 0.1 M NaCI buffer solution (4 m M TES, 0.02 m M E D T A , p H 7.0).
particle size does not return to the initial size of the vesicles measured without dextran sulfate. This demonstrates an irreversible transformation of the vesicles into larger panicles. We did not find any significant difference between the aggregation behaviour of liposomes prepared from DMPC and liposomes prepared from a mixture of DMPC and PE. The mixed system contains positively charged amino groups of the PE besides the positively charged trimethyi ammonium groups of the DMPC. In Fig. 4 the influence of dextran sulfate on the Ca 2"- induced aggregation of phosphatidylserine iiposomes is given. The threshold concentration of aggregation is about 1 mM Ca 2" in the absence of dextran sulfate. This concentration is increased by addition of dextran sulfate. After addition of 0.6 mg/ml or 3.1 mg/ml dextran sulfate the aggregation starts at a Ca 2" concentration of about 0.25 mM and 1.0 mM higher than the initial Ca 2" concentration, respectively. This increase results from the binding of Ca 2" by the dextran sulfate molecules. From the data it can be estimated that 1 mg/ml dextran sulfate (2 /zM) bind about 0.3 mM Ca ~'. Despite the binding of Ca 2" to the liposome surface and to the dextran sulfate molecule a binding o f dextran sulfate to the vesicle surface does not occur
The following results of our experiments show that dextran sulfate is able to induce fusion of pure DMPC and D M P C / P E vesicles in the presence of Ca 2". The probe dilution method was used for the measurement o f the fusion of small DMPC and D M P C / P E vesicles in the presence of dextran sulfate and Ca 2". The vesicles were labeled with 1 tool% NBD-PE and 1 mol% RhPE. These vesicles were mixed with unlabeled vesicles. A fluorescence spectrum is measured which consists of the NBD-PE fluorescence (peak maximum at 525 nm) and Rh-PE fluorescence (peak maximum at 590 nm). Fusion events are detected by the dilution of the fluorescence probes due to the transfer from labeled into unlabeled vesicles. They are monitored as the increase in NBD-PE fluorescence resulting from the decreased resonance energy transfer from the donor (NBD-PE) to the acceptor (Rh-PE). The normalized NBD-PE fluorescence in dependence on subsequent additions of dextran ,.-t .J
0 mM NoCl
0s Z
z_ w
/// 10'~ OEXTRAN
I~' SULPHATE
130 rngtrnt
Fig. 5. Normalized fluorescence intensity of the N-NBD-PE probe in a mixture of 0.2 mM DMPC vesicles labelled with I mol% N-NBD-PE and I mol~0 N-Rh-PE and 0.4 mM unlabelled DMPC vesicles as a function of dextran sulfate concentration in the presence of of 2 mM CaCI 2 and 0 ( + ), 50 ( I l L 100 (A) and 200 ( e ) mM NaCI at a temperature above T (35°C). Dextran sulfate was added at time intervals of 3 rain.
305 sulfate is shown in Fig. 5 for different NaCI concentrations and a temperature o f 35 °C. The fluorescence increase o f NBD-PE has a maximum between 0.01 and 0.1 mg/ml DS 500 and decreases slightly for higher DS 500 concentrations. The increase o f the NBD-PE fluorescence was interpreted to result from fusion processes [121. The effect decreases on increase o f the NaCI concentration because o f the lowered ability o f the system for aggregation. Similar results were found for vesicles consisting of a mixture of DMPC and PE at a molar ratio o f 9:1. The slight decrease o f the NBD fluorescence observed in the concentration range o f dextran sulfate where a disaggregation of vesicles occurs requires a separate discussion. Once the vesicles are fused further increase of the concentration of dextran sulfate should not decrease the fluorescence intensity. This indicates that the observed increase in the NBD fluorescence signal probably contains a small, reversible contribution from the aggregated state beside the fusion state. This conclusion is supported by measurements o f samples which were separately prepared for each dextran sulfate concentration. As expected a much smaller increase of the fluorescence intensity was observed for high dextran sulfate concentrations indicating that a low degree of fusion occurs. The probe mixing method has also given support for the appearence o f fusion processes (data not shown). In this version o f the fluorescence assay each fluorescence probe is placed in a separate population o f vesicles and the quenching of the NED-PE fluorescence is monitored [12]. In the mixture o f both vesicle populations a drastic reduction o f the NBD-PE fluorescence is observed, even at DS 500 concentrations smaller than the concentration o f maximal aggregation. This process has a time constant o f about 2 min and a reversibility was not observed after addition of high concentrations o f dextran sulfate, NaCI and EDTA which results in disaggregation processes. We failed to apply the fluorescence fusion assay o f the internal content mixing because the vesicles become leaky at relatively low concentrations of dextran sulfate.
Discussion Kim and Nishida [4] have found from precipitation measurements that dextran sulfate interacts with phosphatidyicholine vesicles in the presence of Ca 2" and that insoluble complexes are formed. Our turbidity measurements confirm these findings and allow conclusions for other phospholipids, especially positively and negatively charged iipids. Additionally, the electrophoresis data give information about the adsorption process and the fluorescence measurements show the fusogenic action of dextran sulfate on unilamellar vesicles. An interaction o f dextran sulfate with phosphatidylcholine and phosphatidylethanolamine was only observed in the presence of Ca 2". As discussed in Ref. 4, the divalent cation can form bridges between the negative phosphate group of the lipids and the negative sulfate group o f dextran sulfate. Furthermore a participation of the positive charges of the zwitterionic head-group in the binding of the polyanion is expected due to a neutralizing o f the neighbouring phosphate groups by Ca 2". If the positively charged stearylamine is incorporated the surface potential of the liposomes becomes positive and a binding of dextran sulfate occurs also in the absence of Ca 2÷. Vesicles prepared with the negatively charged phosphatidylserine do not bind dextran sulfate, even in the presence of millimolar concentrations of Ca 2". It can be concluded that a positive surface potential (resulting from Ca 2" or incorporation of positively charged molecules in vesicles) is a prerequisite for the binding o f the anionic dextran sulfate. Increase of the NaCI concentration in the medium results in a decrease o f the interaction of dextran sulfate with neutral liposomes in the presence of Ca 2" as shown by aggregation measurements. For a Ca2"/Na" ratio of about 1:100 the interaction vanishes. The ratio of the intrinsic binding constants o f both cations to phospholipids was determined to be about the same value of 1:100 [131. This favors the idea that the competition between Na" and Ca 2" ions for the binding to the liposome surface modulates the interaction with dextran sulfate. Furthermore the
306
ionic strength in the medium influences the interaction indicating an ionic nature of the interaction. As revealed by electrophoresis the binding is saturated at relatively low concentrations of dextran sulfate. Strong electrostatic repulsive forces between free and bound macromolecules may be responsible for this behaviour supported by the fact that the electrophoretic mobility in the saturation range is rather similar for different compositions of liposomes. The structure of the adsorbed layer should depend on the repulsion between charge-carrying segments. The relatively high density of negative charges in the adsorption layer is demonstrated by the high electrophoretic mobility in the plateau region. A close packing of the polymers should not occur but the adsorption in trains is likely [14]. Assuming such a flat layer the surface coverage by dextran sulfate was calculated by application of the Gouy-Chapman theory as described in Ref. 15. With an area of about 0.65 nm 2 of a glycosyl subunit and 2.3 sulfate groups per subunit of dextran sulfate it was calculated that about 10% of the liposome surface are covered by the polymer. The aggregation induced by dextran sulfate is assumed to be mediated by a bridging mechanism of the polymer between adjacent liposomes [4]. At high concentrations of the polymer a disaggregation occurs. The comparison of the dextran sulfate concentration for maximal aggregation and for maximal electrophoretic mobility reveals insights into the processes responsible for the disaggregation. The polymer concentrations necessary for maximal aggregation are slightly lower than the dextran sulfate concentrations for the plateau value of the electrophoretic mobility. Similar results were found for colloidal systems [16]. It can be concluded that the bridging mechanism can only occur at low polymer concentrations where the surface of the particles is covered to a lesser extent. At high polymer concentrations strong electrostatic and steric repulsive forces result in a disaggregation. Our present studies have shown that dextran sulfate is able to induce fusion of small unila-
mellar DMPC and D M P C / P E vesicles in the presence of Ca 2". There are some other effects such as the shift of the phase transition temperature which indicate a change of structural properties of the bilayer by the polymer [17]. It can be imagined that the polymer establishes very narrow contacts between the bilayers by formation of bridges from bilayer to bilayer along the polymer chain. From the concerted participation of all charged groups such as the head group, the polymer and Ca z" in the interaction a merging of the outer layers of the bilayers can occur and the repulsive hydration force can be effectively reduced so enabling a strong approaching of the bilayers. The fusion of the outer layers of the contacting bilayers and the increase of the phase transition temperature can occur in such a state. So far fusion of vesicles was mainly described for cationic polyelectrolytes such as polylysine and polyamine [18]. First reports about fusion in the presence of anionic polyelectrolytes were given in Refs. 19 and 20. Unlike preceding studies the fusion processes described in this paper occur without addition of positively charged lipids and require only the presence of Ca L'. Taking into consideration the enhanced binding of Ca 2" to the phospholipid surface mediated by dextran sulfate [4] an increase of the hydrophobicity of the bilayer surface similar to the action of Ca 2. on the fusion of negatively charged phospholipids is possible [21,22]. Similar results were found by Budker et al. (private communication).
Acknowledgement The work was supported partly by a grant from the U.S. Institute of Health (GM 24840) to K.A. and S.O.
References G. Camejo (1982) Adv. Lipid Res. 19, 1--53. R. Radhakrishnamunhy, S.R. Srinivasan, P. Vijayagopal, P. Dalferes and G.S. Berenson (1982) in: R.S. Varma and R. Varma (Eds), Glycosmminogiycams and Proteogiycans in Physiological and Pathological Processes of Body Systems, S. Karger, Basel, pp. 231--251.
307 3 4 5
6
7
8
9 10 11
A. Mitterer, W.D. Eigner, J. Schurz, G. Juergens and A. Holasek (1982) Int. J. Biol. Macrornol. 4, 227--232. Y.C. Kim and T. Nishida (1977) J. Biol. Chem. 252, 1243--1249. M. Krumbiegel, O. Zschoernig, K. Arnold, O. Panasenko, T. Volnova, O.A. Azizova, A.I. Deer and K. Herrmann (1988) Chem. Phys. Lipids 48, 83--89. K. Arnold, J. Arnhold, O. Zschoernig, D. Wiegel and M. Krumbiegel (1989) Biomed. Biochim. Acta 48, 735 --742. B. Radhakrishnamurthy, H.A. Ruitz, S.R. Srinivasan, W. Preau, E.R. Dalferes and G.S. Berenson (1978) Atherosclerosis 31,217--224. H. Mitsuya, D.J. Looney, S. Kuno, R. Ueno, F. Wong-Staal and S. Broder (1988) Science 240, 649. A.D. Bangham, M.W. Hill and N.G.A. Miller (1974) Methods Membr. Biol. 1, 1---68. D.K. Struck, D. Hoekstra and R.E. Pagano (1981) Biochemistry 20, 4093---4099. S. Ohki and J. Duax (1986) Biochim. Biophys. Acta 861, 177--186.
12
13 14 15 16 17 18 19 20 21 22
N. Duezguenes, T.M. Alien, J. Fedor and D. Papahadjopoulos (1988) in: S. Ohki, D. Doyle, Th.D. Flanagan, S.W. Hui and E. Mayhew (Eds.), Molecular Mechanisms of Membrane Fusion, Plenum Press, New York, pp. 543--555. S. Ohki and R. Kurland (1981) Biochim. Biophys. Acta 645, 170---176. V. Hlady (1984) J. Colloid Interface SCi. 98, 373--384. M. Krumbiegel and K. Arnold (1990) Chem. Phys. Lipids 54, 1--7. M.J. Rosen (1978) in: Surfactants and lnterfacial Phenomena, John Wiley and Sons, New York, p. 265. M. Krumbiegel (1989) Thesis, Leipzig. N, Oku, S. Shibamoto, F. lto, H. Gondo and M. Nango (1987) Biochemistry 26, 8145--8150. M. Beigel, M. Keren-Zur, Y. Laster and A. Loyter (1988) Biochemistry 27, 660---666. M. Keren-Zur, M. Beigel and A. Loyter (1989) Biochim. Biophys. Acta 983, 253--258. S. Ohki (1988) stud. biophys. 127, 89--97. S. Ohki and K. Arnold (1990) J. Membr. Biol. 114, 195 --203.
301
Elsevier ScientificPublishers Ireland Ltd.
Interaction of dextran sulfate with phospholipid surfaces and liposome aggregation and fusion K. Arnold', S. O h k i b a n d M . K r u m b i e g e P "Institute of Biophysics, Medical Center, Karl Marx University, Leipzig (Germanyj and ~Department of Biophysical Sciences, Medical School, SUN)', Buffalo (U.S.A.) (Received February 9th, 1990; revision received June 5th, 1990; accepted June 6th, 1990)
The binding of dextran sulfate to phospholipid liposomes was investigated by microelectrophoresis experiments. The polyanion binds to neutral phospholipid liposomes (DMPC and PE) only in the presence of Ca 2". If positively charged stearylamine is incorporated in the vesicles dextran sulfate is bound without Ca 2". Negatively charged phospholipids as PS do not bind dextran sulfate, even in the presence of millimolar concentrations of Ca ~'. The adsorption of dextran sulfate results in an aggregation of vesicles due to a bridging mechanism. In all cases the aggregation is followed by a disaggregation toward higher dextran sulfate concentrations. The disaggregation process starts at polymer concentrations smaller than the concentration of the onset of saturation of the adsorption. By use of the probe dilution method a fusion of small DMPC and DMPC/PE vesicles in the presence of Ca 2" and dextran sulfate was found.
Keywords: phospholipid vesicles; dextran sulfate; aggregation; fusion; electrophoresis; turbidity; fluorescence.
Introduction The formation of soluble and insoluble associations between plasma lipoproteins and glycosaminoglycans (GAG) has been studied for more than 25 years and it was found that they play a possible role in the depositions of lipoproteins in the arterial intima [1]. The usual explanation of the interaction assumes an ionic interaction between positively charged amino acids of the apolipoprotein B of LDL and the negatively charged GAG molecules [2,3]. Because phospholipid vesicles are also precipitated by heparincalcium mixtures, it was concluded that the interaction with lipoproteins can be also realized by calcium bridges between the phospholipid phosphate groups and the sulfate groups of GAG [4]. An involvement of phospholipids in the interaction of LDL with GAG was demon-
Correspondence to: K. Arnold.
strated very recently [5] and the importance of positively charged amino groups on the LDL surface was shown [6]. Dextran sulfate has a structure very similar to glycosaminoglycans and this polymer was used instead of the glycosaminoglycans in many investigations. The Interest in this polymer increased because it was used as an anti-atherosclerotic drug [7] and potent agent against HIV infection [8] and the molecular mechanisms of these actions were not explained. The purpose of the present investigation was to study the adsorption of dextran sulfate on phospholipid surfaces by use of measurements of the electrophoretic mobility and the effect on vesicle fusion and aggregation. The influence of lipid composition, NaCI and C a 2÷ o n the interaction of dextran sulfate with phospholipids is studied. Especially the study of the influence of Ca z" on the interaction represents a very relevant problem because of the relatively high extracellular Ca 2" concentrations.
0009-3084/90/$03.50 © 1990 Elsevier Scientific Publishers Ireland Ltd. Published and Printed in Ireland
302 H
Materials and Methods
3
Dimyristoylphosphatidylcholine (DMPC) from Serva, F.R.G. and egg-phosphatidylserine (PS) and egg-phosphatidylethanolamine (PS) from Avanti Polar Lipids, U.S.A. were used. Multilamellar vesicles (MLV) were prepared using the method of Bangham et al. [9]. For the preparation of small unilamellar vesicles a bathtype sonicator was applied at a frequency of 20 kHz. Optically clear suspensions were obtained. The dextran sulfate of average molecular weight of 500,000 (DS 500) was purchased from Serva, F.R.G. The fusion of vesicles was monitored by the probe dilution method which makes use of the resonance energy transfer between nitrobenzooxa-diazole (NBD) and rhodamine (Rh) covalently attached to the head group of phosphatidylethanolamine [10]. The corresponding fluorescent phospholipids N-(7-nitro-2,1,3benzo-oxa-diazole-4-yl)-PE (NBD-PE) and N(lissamine Rhodarnin B sulfonyl)dioleoyl-PE (Rh-PE) were purchased from Molecular Probes, U.S.A.. Fluorescence spectra were measured by use of the spectrofluorimeter Perkin-Elmer LS5 and the fluorescence was excited at 475 nm and the spectra were recorded between 490 and 600 am. The size of the vesicles was determined from the dynamic light scattering by use of the submicron particle analyzer Coulter Model N4S (Coulter Electronics, U.S.A.). A Parmoquant-2 (Carl Zeiss Jena, G.D.R.) operating in the dark field mode was used for the measurement of the electrophoretic mobility of MLV. In each case the device automatically calculates mean values and standard deviations of the electrophoretic mobility of a set of 100 particles. The zeta potential was calculated according to the Smolukhovski equation. The viscosity of the solution was measured with a Hoeppler viscosimeter. Results
Electrophoresis measurements
The electrophoretic mobility of pure phospha-
•
....
~.JL--'-,=|
"o_o
-2 >~,m o
"1
u~
:" W -J hi
o2 "H
o
amd DEXTRAN SULPHATE m g / m l
Fig. 1. Electrophoretic mobility of muhilamellar D M P C vesicles in the presence o f 3 m M Ca 2" (O) and D M P C vesicles containing different mole fractions o f stearylamine (A, 5 tool %; n , I0 tool %; o , 20 tool %) as a function o f dextran sulfate concentration. The vesicles were prepared in a buffer solution of I0 m M Tris, 145 m M NaCI, 0.3 m M EDTA, pH 7.4. The total lipid concentration was 0.I rag/ ml. The measurements were performed at a temperature below T (20°C).
tidylcholine (DMPC) multilamellar vesicles which are uncharged and do not move in the electric field is not changed on addition of dextran sulfate. A completely different behaviour is measured for DMPC vesicles which contain 5, 10 and 20 mol~0 stearylamine (Fig. l). These vesicles have a positive electrophoretic mobility in the absence of DS 500 and the magnitude of the mobility is dependent on the mole fraction of the positively charged stearylamine. Addition of DS 500 results in a reversal of the direction of the electrophoretic motion of the vesicles. The electrophoretic mobility is increased for higher concentrations of DS 500 and a plateau is reached for DS 500 concentrations higher than 0.01 mg/ml. The plateau values depend on the stearylamine concentration in the vesicles to a very small extent. These experiments demonstrate the adsorption of the polyanionic DS 500 on the vesicle surface. Negative charge is accumulated due to the binding and overcomes the positive charge of the vesicle surface which
303
determined the electrophoretic mobility prior to
the ad.sorption of the polymer. The eleetrophoretic mobility of DMPC MLV as a function o f the concentration o f DS 500 in the presence o f 3 mM Ca 2° is also given in Fig. 1. The electrophoretic mobility of the particles is positive in the absence o f DS 500 due to a small adsorption of Ca ~°. On addition o f DS 500 a similar behaviour as described above was observed.
,
Turbidity measurements A very strong aggregation occurs for the positively charged liposomes consisting o f mixtures o f DMPC and stearylamine. Upon addition of DS 500 such large particles are formed that a fast precipitation appears and turbidity measurements cannot be usefully applied. A systematic study of the influence of dextran sulfate, Ca 2° and NaCI on the aggregation of unilarnellar DMPC and D M P C / P E liposomes was performed. The aggregation process was monitored by measurements o f the turbidity and the sizes of vesicles were determined by dynamic light scattering. Dextran sulfate was subsequently added in a concentration range from 0.005 to 5 mg/ml (10 nM--10/aM).
0
'~
o oo,
o
"~, o',
,o
"
01 10 OEXTRAN SULPHATE, n~/rr~
tOO
Fig. 3. Particle size of sonicated DMPC/eBB-PE vesicles (10 tool% egg-PE) as a function of dextran sulfate concentration for 0 ( + ) , 50 ( I ) , 100 (&), 150 ( e ) and 200 (O) mM NaCI in the presence of 2.0 mM CaCI 2 and 10raM HEPES at pH 7.4 and above T (30°C). Dextran sulfate was added at time intervals of 3 rain. A phospholipid concentration of 0.3 rag/ ml was used.
10
$
O0l
\ ,0o
OEXTRAN SULPHATE500. mg/ml Fig. 2. Turbidity of sonicated DMPC/egg-PE vesicles (10 tool% egg-PE) as a function of dextran sulfate concentration for 0 ( + ) , 50 ( 8 ) , 100 (&), 150 (O) and 200 (O) mM NaC! in the presence of 2.0 mM CaCI 2 and 10 mM HEPES at pH 7.4 and above 7", (30°C). Dextran sulfate was added at time intervals of 3 min. A phospholipid concentration of 0.3 rag/ ml was used.
As representative examples the change o f the turbidity (Fig. 2) and the size (Fig. 3) of the D M P C / P E liposomes is given as a function of DS 500 concentration for 2 mM Ca 2" and different NaCI concentrations. In all experiments a maximum of the turbidity is observed. The maximal turbidity is a function o f the NaCI concentration and it decreases with higher NaCI concentrations. For a Ca2"/Na ÷ ratio of about 1 : 100 aggregation vanishes. The particle sizes of the phospholipid vesicles determined by dynamic light scattering are given in Fig. 3 for different NaCI concentrations as a function o f dextran sulfate. The same samples whose turbidity was given in Fig. 2 were used. The particle size increases and has a maximum at about the same concentration of dextran sulfate where the turbidity maximum was observed. In the absence of NaCI the increase of the particle size from about 55 nm to about 500 nm occurs at the lowest concentration of dextran sulfate and is unchanged over the whole range of concentrations. A decrease of the particle sizes appears for all other NaCI concentrations after passing the size maximum. However, the
304
1,, 2 . 3
2(:
/
/
>-
_o m (Z I--
/-,
~-- 05 <
/
.-/
w
e
~e
06
10
1/.
~8
"-
-
Fluorescence measurements
/
to
w
because of the strong repulsion of the negatively charged surfaces of the liposomes and the dextran sulfate molecules. At the concentrations of Ca 2. for aggregation of PS liposomes the surface potential is still negative (about 30 mV [11]).
2"2
CO 2" C O N C E N T R A T I O N , mM Fig. 4. Turbidity o f sonicated phosphatidylserin¢ vesicles as a function o f the Ca 2" concentration in the presence o f 0 mg/ ml (1), 0.075 m $ / m l (2), 0.6 m g / m l (3) and 3.1 m g / m l (4) dextran sulfate at 27°C. The samples were prepared in 0.1 M NaCI buffer solution (4 m M TES, 0.02 m M E D T A , p H 7.0).
particle size does not return to the initial size of the vesicles measured without dextran sulfate. This demonstrates an irreversible transformation of the vesicles into larger panicles. We did not find any significant difference between the aggregation behaviour of liposomes prepared from DMPC and liposomes prepared from a mixture of DMPC and PE. The mixed system contains positively charged amino groups of the PE besides the positively charged trimethyi ammonium groups of the DMPC. In Fig. 4 the influence of dextran sulfate on the Ca 2"- induced aggregation of phosphatidylserine iiposomes is given. The threshold concentration of aggregation is about 1 mM Ca 2" in the absence of dextran sulfate. This concentration is increased by addition of dextran sulfate. After addition of 0.6 mg/ml or 3.1 mg/ml dextran sulfate the aggregation starts at a Ca 2" concentration of about 0.25 mM and 1.0 mM higher than the initial Ca 2" concentration, respectively. This increase results from the binding of Ca 2" by the dextran sulfate molecules. From the data it can be estimated that 1 mg/ml dextran sulfate (2 /zM) bind about 0.3 mM Ca ~'. Despite the binding of Ca 2" to the liposome surface and to the dextran sulfate molecule a binding o f dextran sulfate to the vesicle surface does not occur
The following results of our experiments show that dextran sulfate is able to induce fusion of pure DMPC and D M P C / P E vesicles in the presence of Ca 2". The probe dilution method was used for the measurement o f the fusion of small DMPC and D M P C / P E vesicles in the presence of dextran sulfate and Ca 2". The vesicles were labeled with 1 tool% NBD-PE and 1 mol% RhPE. These vesicles were mixed with unlabeled vesicles. A fluorescence spectrum is measured which consists of the NBD-PE fluorescence (peak maximum at 525 nm) and Rh-PE fluorescence (peak maximum at 590 nm). Fusion events are detected by the dilution of the fluorescence probes due to the transfer from labeled into unlabeled vesicles. They are monitored as the increase in NBD-PE fluorescence resulting from the decreased resonance energy transfer from the donor (NBD-PE) to the acceptor (Rh-PE). The normalized NBD-PE fluorescence in dependence on subsequent additions of dextran ,.-t .J
0 mM NoCl
0s Z
z_ w
/// 10'~ OEXTRAN
I~' SULPHATE
130 rngtrnt
Fig. 5. Normalized fluorescence intensity of the N-NBD-PE probe in a mixture of 0.2 mM DMPC vesicles labelled with I mol% N-NBD-PE and I mol~0 N-Rh-PE and 0.4 mM unlabelled DMPC vesicles as a function of dextran sulfate concentration in the presence of of 2 mM CaCI 2 and 0 ( + ), 50 ( I l L 100 (A) and 200 ( e ) mM NaCI at a temperature above T (35°C). Dextran sulfate was added at time intervals of 3 rain.
305 sulfate is shown in Fig. 5 for different NaCI concentrations and a temperature o f 35 °C. The fluorescence increase o f NBD-PE has a maximum between 0.01 and 0.1 mg/ml DS 500 and decreases slightly for higher DS 500 concentrations. The increase o f the NBD-PE fluorescence was interpreted to result from fusion processes [121. The effect decreases on increase o f the NaCI concentration because o f the lowered ability o f the system for aggregation. Similar results were found for vesicles consisting of a mixture of DMPC and PE at a molar ratio o f 9:1. The slight decrease o f the NBD fluorescence observed in the concentration range o f dextran sulfate where a disaggregation of vesicles occurs requires a separate discussion. Once the vesicles are fused further increase of the concentration of dextran sulfate should not decrease the fluorescence intensity. This indicates that the observed increase in the NBD fluorescence signal probably contains a small, reversible contribution from the aggregated state beside the fusion state. This conclusion is supported by measurements o f samples which were separately prepared for each dextran sulfate concentration. As expected a much smaller increase of the fluorescence intensity was observed for high dextran sulfate concentrations indicating that a low degree of fusion occurs. The probe mixing method has also given support for the appearence o f fusion processes (data not shown). In this version o f the fluorescence assay each fluorescence probe is placed in a separate population o f vesicles and the quenching of the NED-PE fluorescence is monitored [12]. In the mixture o f both vesicle populations a drastic reduction o f the NBD-PE fluorescence is observed, even at DS 500 concentrations smaller than the concentration o f maximal aggregation. This process has a time constant o f about 2 min and a reversibility was not observed after addition of high concentrations o f dextran sulfate, NaCI and EDTA which results in disaggregation processes. We failed to apply the fluorescence fusion assay o f the internal content mixing because the vesicles become leaky at relatively low concentrations of dextran sulfate.
Discussion Kim and Nishida [4] have found from precipitation measurements that dextran sulfate interacts with phosphatidyicholine vesicles in the presence of Ca 2" and that insoluble complexes are formed. Our turbidity measurements confirm these findings and allow conclusions for other phospholipids, especially positively and negatively charged iipids. Additionally, the electrophoresis data give information about the adsorption process and the fluorescence measurements show the fusogenic action of dextran sulfate on unilamellar vesicles. An interaction o f dextran sulfate with phosphatidylcholine and phosphatidylethanolamine was only observed in the presence of Ca 2". As discussed in Ref. 4, the divalent cation can form bridges between the negative phosphate group of the lipids and the negative sulfate group o f dextran sulfate. Furthermore a participation of the positive charges of the zwitterionic head-group in the binding of the polyanion is expected due to a neutralizing o f the neighbouring phosphate groups by Ca 2". If the positively charged stearylamine is incorporated the surface potential of the liposomes becomes positive and a binding of dextran sulfate occurs also in the absence of Ca 2÷. Vesicles prepared with the negatively charged phosphatidylserine do not bind dextran sulfate, even in the presence of millimolar concentrations of Ca 2". It can be concluded that a positive surface potential (resulting from Ca 2" or incorporation of positively charged molecules in vesicles) is a prerequisite for the binding o f the anionic dextran sulfate. Increase of the NaCI concentration in the medium results in a decrease o f the interaction of dextran sulfate with neutral liposomes in the presence of Ca 2" as shown by aggregation measurements. For a Ca2"/Na" ratio of about 1:100 the interaction vanishes. The ratio of the intrinsic binding constants o f both cations to phospholipids was determined to be about the same value of 1:100 [131. This favors the idea that the competition between Na" and Ca 2" ions for the binding to the liposome surface modulates the interaction with dextran sulfate. Furthermore the
306
ionic strength in the medium influences the interaction indicating an ionic nature of the interaction. As revealed by electrophoresis the binding is saturated at relatively low concentrations of dextran sulfate. Strong electrostatic repulsive forces between free and bound macromolecules may be responsible for this behaviour supported by the fact that the electrophoretic mobility in the saturation range is rather similar for different compositions of liposomes. The structure of the adsorbed layer should depend on the repulsion between charge-carrying segments. The relatively high density of negative charges in the adsorption layer is demonstrated by the high electrophoretic mobility in the plateau region. A close packing of the polymers should not occur but the adsorption in trains is likely [14]. Assuming such a flat layer the surface coverage by dextran sulfate was calculated by application of the Gouy-Chapman theory as described in Ref. 15. With an area of about 0.65 nm 2 of a glycosyl subunit and 2.3 sulfate groups per subunit of dextran sulfate it was calculated that about 10% of the liposome surface are covered by the polymer. The aggregation induced by dextran sulfate is assumed to be mediated by a bridging mechanism of the polymer between adjacent liposomes [4]. At high concentrations of the polymer a disaggregation occurs. The comparison of the dextran sulfate concentration for maximal aggregation and for maximal electrophoretic mobility reveals insights into the processes responsible for the disaggregation. The polymer concentrations necessary for maximal aggregation are slightly lower than the dextran sulfate concentrations for the plateau value of the electrophoretic mobility. Similar results were found for colloidal systems [16]. It can be concluded that the bridging mechanism can only occur at low polymer concentrations where the surface of the particles is covered to a lesser extent. At high polymer concentrations strong electrostatic and steric repulsive forces result in a disaggregation. Our present studies have shown that dextran sulfate is able to induce fusion of small unila-
mellar DMPC and D M P C / P E vesicles in the presence of Ca 2". There are some other effects such as the shift of the phase transition temperature which indicate a change of structural properties of the bilayer by the polymer [17]. It can be imagined that the polymer establishes very narrow contacts between the bilayers by formation of bridges from bilayer to bilayer along the polymer chain. From the concerted participation of all charged groups such as the head group, the polymer and Ca z" in the interaction a merging of the outer layers of the bilayers can occur and the repulsive hydration force can be effectively reduced so enabling a strong approaching of the bilayers. The fusion of the outer layers of the contacting bilayers and the increase of the phase transition temperature can occur in such a state. So far fusion of vesicles was mainly described for cationic polyelectrolytes such as polylysine and polyamine [18]. First reports about fusion in the presence of anionic polyelectrolytes were given in Refs. 19 and 20. Unlike preceding studies the fusion processes described in this paper occur without addition of positively charged lipids and require only the presence of Ca L'. Taking into consideration the enhanced binding of Ca 2" to the phospholipid surface mediated by dextran sulfate [4] an increase of the hydrophobicity of the bilayer surface similar to the action of Ca 2. on the fusion of negatively charged phospholipids is possible [21,22]. Similar results were found by Budker et al. (private communication).
Acknowledgement The work was supported partly by a grant from the U.S. Institute of Health (GM 24840) to K.A. and S.O.
References G. Camejo (1982) Adv. Lipid Res. 19, 1--53. R. Radhakrishnamunhy, S.R. Srinivasan, P. Vijayagopal, P. Dalferes and G.S. Berenson (1982) in: R.S. Varma and R. Varma (Eds), Glycosmminogiycams and Proteogiycans in Physiological and Pathological Processes of Body Systems, S. Karger, Basel, pp. 231--251.
307 3 4 5
6
7
8
9 10 11
A. Mitterer, W.D. Eigner, J. Schurz, G. Juergens and A. Holasek (1982) Int. J. Biol. Macrornol. 4, 227--232. Y.C. Kim and T. Nishida (1977) J. Biol. Chem. 252, 1243--1249. M. Krumbiegel, O. Zschoernig, K. Arnold, O. Panasenko, T. Volnova, O.A. Azizova, A.I. Deer and K. Herrmann (1988) Chem. Phys. Lipids 48, 83--89. K. Arnold, J. Arnhold, O. Zschoernig, D. Wiegel and M. Krumbiegel (1989) Biomed. Biochim. Acta 48, 735 --742. B. Radhakrishnamurthy, H.A. Ruitz, S.R. Srinivasan, W. Preau, E.R. Dalferes and G.S. Berenson (1978) Atherosclerosis 31,217--224. H. Mitsuya, D.J. Looney, S. Kuno, R. Ueno, F. Wong-Staal and S. Broder (1988) Science 240, 649. A.D. Bangham, M.W. Hill and N.G.A. Miller (1974) Methods Membr. Biol. 1, 1---68. D.K. Struck, D. Hoekstra and R.E. Pagano (1981) Biochemistry 20, 4093---4099. S. Ohki and J. Duax (1986) Biochim. Biophys. Acta 861, 177--186.
12
13 14 15 16 17 18 19 20 21 22
N. Duezguenes, T.M. Alien, J. Fedor and D. Papahadjopoulos (1988) in: S. Ohki, D. Doyle, Th.D. Flanagan, S.W. Hui and E. Mayhew (Eds.), Molecular Mechanisms of Membrane Fusion, Plenum Press, New York, pp. 543--555. S. Ohki and R. Kurland (1981) Biochim. Biophys. Acta 645, 170---176. V. Hlady (1984) J. Colloid Interface SCi. 98, 373--384. M. Krumbiegel and K. Arnold (1990) Chem. Phys. Lipids 54, 1--7. M.J. Rosen (1978) in: Surfactants and lnterfacial Phenomena, John Wiley and Sons, New York, p. 265. M. Krumbiegel (1989) Thesis, Leipzig. N, Oku, S. Shibamoto, F. lto, H. Gondo and M. Nango (1987) Biochemistry 26, 8145--8150. M. Beigel, M. Keren-Zur, Y. Laster and A. Loyter (1988) Biochemistry 27, 660---666. M. Keren-Zur, M. Beigel and A. Loyter (1989) Biochim. Biophys. Acta 983, 253--258. S. Ohki (1988) stud. biophys. 127, 89--97. S. Ohki and K. Arnold (1990) J. Membr. Biol. 114, 195 --203.