transport properties of model membranes ��

The Langmuir monolayer technique and voltammetric analysis were used to investigate the properties of model lipid membranes prepared from dioleoylphosphatidylcholine (DOPC), hexadecaprenol (C80), and their mixtures. Surface pressure-molecular area isotherms, current-voltage characteristics, and membrane conductance-temperature were measured. Molecular area isobars, specific molecular areas, excess free energy of mixing, collapse pressure and collapse area were determined for lipid monolayers. Membrane conductance, activation energy of ion migration across the membrane, and membrane permeability coefficient for chloride ions were determined for lipid bilayers. Hexadecaprenol decreases the activation energy and increases membrane conductance and membrane permeability coefficient. The results of monolayer and bilayer investigations show that some electrical, transport and packing properties of lipid membranes change under the influence of hexadecaprenol. The results indicate that hexadecaprenol modulates the molecular organisation of the membrane and that the specific molecular area of polyprenol molecules depends on the relative concentration of polyprenols in membranes. We suggest that hexadecaprenol modifies lipid membranes by the formation of fluid microdomains. The results also indicate that electrical transmembrane potential can accelerate the formation of pores in lipid bilayers modified by long chain polyprenols.

Polyisoprenols (polyprenols and dolichols) are found in membrane fractions of prokaryotic and eukaryotic cells (Chojnacki et al., 1987).Polyisoprenols (poly-cis-prenols) are natural products, derivatives of the C5 isoprene unit.The occurrence of a-unsaturated polyisoprenols (polyprenols) (OEwie¿ewska et al., 1994;Wanke et al., 1998) and a-saturated polyisoprenols (dolichols) (Jankowski et al., 1994;Hemming 1983; Chojnacki et al., 1987) in membrane fractions has been reported.Phosphopolyisoprenols function as hydrophobic carriers of glycosyl residues across membranes in glycosylation reactions (Bugg & Brandish, 1994).The behaviour of polyisoprenols in model lipid membranes has been studied using several techniques, including ESR (McCloskey & Troy, 1980;Lai & Schutzbach, 1984), NMR (de Ropp & Troy, 1985; Valterson et al., 1985;Knudsen & Troy 1989), fluorescence spectroscopy (Sunamoto et al., 1983;Boscoboinik et al., 1985;Vigo et al., 1984), X-ray scattering (Gruner, 1985), differential scanning calorimetry (Vigo et al., 1984), voltammetric analysis (Janas et al., 1986;1994;1998;2000) and electron microscopy (Janas et al., 2000).However, the mechanism of the interaction between polyprenol and phospholipid molecules is still unclear.Studies of monomolecular films and bilayer lipid membranes containing a mixture of membrane lipids are of considerable importance because of their relevance to numerous natural systems.This paper presents the investigations of monolayer and bilayer lipid membranes modified by a long chain polyprenol -hexadecaprenol (C 80 ).The molecule of hexadecaprenol is composed of 16 isoprene units with the structure: wT 2 C 12 aOH, where w is an isoprene residue, T is a trans-isoprene residue, C is a cis-isoprene residue, a is the a-saturated OH terminal isoprene residue and OH is the hydroxyl group.Monolayer techniques have been applied to the study of the interactions between dioleoylphosphatidylcholine and other lipids (Gaines, 1966;Costin & Barnes, 1975;Gruszecki et al., 1999a) and this technique can be used to investigate the origin and magnitude of the molecular interactions in mixed monolayers.The aim of the present work was to study the influence of polyprenol molecules on the organisation and packing of phospholipid molecules in lipid monolayers and bilayers.
Monolayer formation and isotherm recording.The monolayers were deposited by spreading a proper volume of C 80 /DOPC mixture in chloroform.Surface pressure was measured by the Wilhelmy method (Gruszecki et al., 1999a;1999b).Monomolecular layers at the air-water interface were formed in a 10 40 cm Teflon trough.The experiments were run at 21°C.Prior to isotherm recording, monolayers were equilibrated at zero pressure for 5 min to allow evaporation of chloroform.Lipid monolayers were then compressed at a speed of 0.5 mm/s.Surface pressure was measured by tensiometer PS 3 from Nima Technology and entered into computer memory.Measurement error was less than 0.1 mN/m.Deionised water was used as the subphase.The initial value of the area per molecule was 5 nm 2 .The obtained data with measurement error less than 0.1 mN/m were further elaborated by the use of Excel 5.0 worksheets (office software package, Microsoft ) and mathematical calculations were performed using Mathematica 3.0 (Wolfram Research).
In the present work we analysed mixed monolayer isotherms in terms of excess free energy of mixing, DG mix E was calculated for each mixture using the following equation (Pack et al., 1997;Markowitz et al., 1995): where A 12 is the area per molecule in the mixed film, A 1 and A 2 are molar areas in pure films, x 1 and x 2 are molar fractions of component 1 and 2, and P is the surface pressure.
The integrals correspond to the areas under the P-A isotherms, and P o is defined as the surface pressure where the monolayer components are ideally miscible.It is generally assumed that P o is close to zero.In practice, P o is commonly set to the lowest measurable surface pressure, and in this work P o was set to 0.4 mN/m.This method involves calculating the differences between the areas under the surface pressure-area isotherms of the mixtures and pure components at a specific surface pressure.
Experimentally it is found that many monolayers can be compressed to pressures considerably higher than their equilibrium spreading pressures.Eventually, however, it is impossible to increase the surface pressure further and the area of the film decreases if constant pressure is maintained or pressure falls in the film held at a constant area.This condition is referred to as the collapse point (collapse pressure (p c ); collapse area (A c )) of the monolayer under the experimental conditions.When collapse occurs, molecules are forced out of the monolayer to form agglomerates of an adjacent bulk phase (Gaines, 1966).
Bilayer formation and electrical measurements.Bilayer lipid membranes in the form of hemispheres were formed according to the technique described previously (Janas et al., 1986) on a Teflon capillary tube in unbuffered (pH 6) aqueous solution of 0.1 M and 0.2 M NaCl (at the inner and outer side of the membrane, respectively).DOPC or C 80 /DOPC mixtures used for membrane formation were dissolved in n-decane/butanol (3:1, v/v) at 10 mg/ml.The area of the macrovesicular bilayer lipid membrane was about 50 mm 2 .Saturated silver chloride electrodes were used to apply external voltage and detect the electric potentials.Electrometers were used to measure voltage distribution between the membrane and an external resistance.The area of the membrane, S, was determined by optical measurement of membrane dimensions.Temperature, T, was controlled by water circulating from an external bath.Electrical conductance of the membrane, G, was calculated from current-voltage characteristics.
Membrane permeability coefficients for Cl - ions, P Cl -, were calculated from the following equation (Tien, 1974): where c 1 and c 2 are the concentrations of NaCl inside and outside the spherical bilayer, respectively; F is the Faraday constant; E Clis the equilibrium potential for Cl -ions; G is specific membrane conductance.The ratio of ionic transference numbers (t Na +/t Cl -for sodium and chloride ions, respectively) was determined from measurements of steady-state diffusion potentials (Janas & Janas, 1995).Experimental values were fitted by the Goldman-Hodgkin-Katz equation (Gamble et al., 1982;Janas et al., 2000): which correlates the potential difference, DV m , developed between the two sides of the membrane to the activities, a Na + and a Cl -, of the sodium and chloride ions, respectively, at the inner side (i) and the outer side (o) of the membrane; V i and V o are the electric potentials at the inner and outer side of the membrane, respectively; F is the Faraday constant.
The activation energy of ion migration across membrane, E A , was determined from Arrhenius plots of normalized conductance of bilayer lipid membranes (Smith et al., 1984): where: ln[(G/C)/(G 0 /C 0 )] is the normalized conductance of the membrane; G 0 and C 0 are membrane conductance and membrane capacitance, respectively, at temperature T 0 , R is the gas constant.The normalisation of membrane conductance (with respect to the membrane capacitance measured simultaneously) corrects for any variations in the bilayer conductance, which are due to variations in bilayer area or bilayer thickness.

Monolayer experiments
The Langmuir monolayer technique was used to investigate the interaction of hexadecaprenol with lipid films.Figure 1 shows typical surface pressure-area curves (isotherms) of monolayers prepared from pure hexadecaprenol, pure DOPC, and their mixtures in various molar fractions (0.01; 0.1; 0.5; 0.9; 0.99) spread on water and compressed at a rate of 0.5 mm/s at 21°C.By changing the molar fraction, a gradual change in the shape of the isotherms is observed.One can notice from Fig. 1 that at pressures below 20 mN/m, monolayers prepared from the molar fraction 0.01 and 0.1 are more expanded than monolayers formed from the molar fractions 0.9 and 0.99.It means that lipid packing is optimal for membranes of molar fraction equal to 0.9 and 0.99. Figure 2      The value of A M was linearly extrapolated from the linear parts of the isotherms to zero surface pressure.

Bilayer experiments
The behaviour of hexadecaprenol/DOPC membranes as a function of applied potential was studied by performing current-voltage experiments.As presented in Fig. 7A, the curves are symmetric and linear for values in the potential range -20 to +20 mV.The value of the slope increases with the increased percentage of hexadecaprenol in the membrane.The dependence of membrane specific conductance, G S , on the percentage is shown on a semilogarithmic scale in Fig. 7B.The values of G S increase with the increasing percentage of hexadecaprenol (C 80 ) in the membrane.The maximal rise, up to G S max = (6.8± 1.1)´10 -7 S cm -2 , was observed for the C 80 /DOPC mole ratio equal to 0.2.The value of membrane conductance obtained for DOPC bilayers equals (4.8 ± 0.9)´10 -8 S cm -2 and is in accordance with the value (4 ± 1)´10 -8 S cm -2 reported by Gamble et al. (1982).The normalized conductance of bilayer lipid membranes was measured as a function of temperature in the range of 25-42°C.Typical trends are reported in Fig. 8A.An increase of normalized conductance is observed with increasing temperature.The Arrhenius plots are linear, the slope of the curves depending on the percentage of C 80 in the bilayer.The relationship between the value of activation energy of ion transport across the membrane, E A , and the percentage of hexadecaprenol in macrovesicular bilayers is shown in Fig. 8B.The values of activation 666 T. Janas and others 2000   energies were derived from the Arrhenius plots by least squares fitting according to eqn. 4. For lower concentrations of hexadecaprenol, a pronounced decrease of E A is observed.For higher concentrations of C 80 in the membrane, the value of E A increases slightly.The E A value decreases from 48 ± 3 kJ/mole for DOPC bilayers to the value of minimal activation energy, E A min , equal to 23 ± 2 kJ/mole for bilayers prepared from mixtures at C 80 /DOPC mole ratio 0.1 and increases slightly for higher concentrations of hexadecaprenol.Smith et al. (1984) reported activation energy equal to 35 ± 2 kJ/mole for a lecithin bilayer.Ionic transference numbers, calculated according to eqn. 3, were nearly independent of the lipid composition of the membrane and the ratio (t Cl -/t Na +) of about 1.5 was obtained at 25°C.The dependence of the membrane permeability coefficient for chloride ions, P Cl -, on the percentage of Experiments were performed at 25 ± 0.1°C.

B. Specific membrane conductance, G, versus C 80 /DOPC mole ratio.
Values of membrane conductance were derived from the linear parts of I/V curves by least-squares fitting.Each point represents the mean value ± S.D. obtained for 6-8 different macrovesicular bilayer lipid membranes.

B. Activation energy, E A , of ion migration across the membrane vs C 80 /DOPC mole ratio.
Values of activation energies were derived from Arrhenius plots by least-squares fitting according to eqn. 4 (see Materials and Methods).Each point represents the mean value ± S.D. obtained for 6-8 different macrovesicular bilayer lipid membranes.
hexadecaprenol in the bilayer is shown on a semilogarithmic scale in Fig. 9.The value of P Cl -, calculated according to eqn. 2, increases with the increasing percentage of C 80 in the bilayer.The value of P Cl -equal to (4.2 ± 0.7)´10 -11 cm/s is obtained for a DOPC bilayer.The value of this coefficient for lipid bilayers modified by long chain polyprenols is higher than for DOPC bilayers and the maximal rise, about 30-fold, is observed for C 80 /DOPC mole ratio equal to 0.2.For lower concentrations of the polyprenol in the membrane, the rate of increase of P Cl -is considerable.For higher concentrations of C 80 in the membrane, a slight increase in the value of P Cl -is observed in comparison with DOPC bilayers.

DISCUSSION
Lipid membranes modified by hexadecaprenol (C 80 ) exhibit different electrical, transport and mixing properties from DOPC monolayers and bilayers.The behaviour of hexadecaprenol-lecithin in monolayers was studied by performing Langmuir monolayer technique experiments.Surface pressurearea isotherms were measured for pure C 80 , DOPC films and their mixtures.The changes of specific molecular area, A M , free energy of mixing, DG mix E , and area per molecule at constant surface pressure, A p , show that hexadecaprenol can modify the structure of lipid aggregates.Similar conclusions were presented for xanthophyll-lecithin monolayers (Tomoaia-Cotiºel et al., 1987;N'soukpoé-Kossi et al., 1988).Negelmann et al. (1997) noted that the limiting molecular area value represents the beginning of observable intermodular forces between adjacent molecules in the monolayer.The behaviour of mixed monolayers can be considered in terms of excess free energy of mixing DG mix E .This interesting phenomenon, represented in Fig. 5, indicates two opposite processes: van der Waals interactions between polyprenol chains, and hydrogen interactions between polyprenol hydroxyl group and the oxygen atom of a water molecule.In the case of a two-component monolayer, the hydrogen bond can be additionally formed between the hydroxyl group of hexadecaprenol and the phosphate group of a DOPC molecule.Our research indicates that for high molar fractions of polyprenols van der Waals forces predominate.This means that the DOPC molecules are monodispersed and surrounded by several C 80 molecules each.One can notice that a negative deviation from the rule of additivity is usually interpreted as pointing at a mutual interaction in a two-component system, decreasing the area occupied by molecules in the mixture and lowering the excess of free energy in a monolayer.Positive deviation from the straight line in this case seems to reflect a phenomenon of molecules packed in a monolayer indicating that even small additions of C 80 disturb the layer of DOPC.The increase of the area per molecule seems to be the cause of the existence (in the mixture) of the border of phases not occupied by molecules.As can be noticed in the isothermic isobars of Fig. 2, two ex-

668
T. Janas and others 2000  With respect to the electrical properties, the measurements showed that hexadecaprenol increases membrane specific conductance which is consistent with the data for dotriacontaprenol (C 160 ) (Janas et al., 2000).Macrovesicular lipid bilayers prepared from hexadecaprenol/DOPC mixtures exhibited much higher permeability for ions in comparison with DOPC bilayers and the values of activation energy of ion transport were much lower than those for DOPC bilayers.The perfect linearity of current-voltage characteristics for hexadecaprenol-containing DOPC bilayers indicates that the increase in membrane permeability and decrease in activation energy are incompatible with the formation of carriers or selective channels by hexadecaprenol molecules.Aggregation of spin-labelled polyisoprenols in phospholipid membranes was observed even at relative concentrations not exceeding 0.005 (McCloskey & Troy, 1980).These aggregates can modulate the permeability and stability of polyisoprenolphospholipid membranes.Our investigations show that hexadecaprenol substantially decreases the energy barrier for ion migration through membranes, giving rise to an increase of ionic conductance.Lai & Schutzbach (1984) showed that dolichol promoted membrane leakage in the absence of transmembrane potential in liposomes composed of phosphatidylethanolamine and phosphatidylcholine but not in liposomes composed of phosphatidylcholine only.A strong destabilisation of phosphatidylethanolamine bilayers, but not the phosphatidylcholine ones, in the presence of a-saturated polyprenol (dolichol) was detected by Valterson et al. (1985) on studying the phase transition of the bilayers.These experiments were also performed in the absence of transmembrane potential.In contrast to their observation, we report an increase of ion permeability in the presence of transmembrane potential, observed for C 80 /DOPC bilayers containing no phosphatidylethanolamine.
The activation energy of ion transport was found to be essentially independent of temperature.This indicates that the influence of temperature on the aggregation behaviour of C 80 and DOPC molecules in the membrane is negligible, which is consistent with the observation that the temperature dependence of clustering of polyisoprenol molecules in model membranes is minimal (McCloskey & Troy, 1980).a-Saturated polyprenols were previously found to increase the motional freedom of bilayer lipid membranes (de Ropp & Troy, 1984;1985;Vigo et al., 1984;Knudsen & Troy, 1989) and plasma membranes (Wood et al., 1986;1989;Schroeder et al., 1987).McCloskey & Troy (1980) have demonstrated the existence of polyisoprenol clustering in phospholipid bilayers.We suggest the existence in C 80 /DOPC bilayers of hexadecaprenol-rich microdomains, which can form transmembrane pores.The microdomain may be stabilized by hydrogen bonds between hydroxyl group of C 80 and the ester oxygen of DOPC.These microdomains can modulate the permeability and stability of hexadecaprenol-phospholipid membranes.As analysed in the paper of Smith et al. (1984) the decrease in the activation energy of ion migration across lipid bilayers is related to the increase in the radius of the transmembrane pore.For the value of activation energy of 18 kJ/mole, the authors estimated the minimum pore radius to be about 1 nm.For C 80 /DOPC bilayers, with the activation energy for ion transport about 2-fold lower, the minimal pore radius can be estimated to be around 2 nm.The action of C 80 seems therefore to affect the formation of these pores and to increase their size.The function of polyprenyl microdomains in biological membranes may also depend on their ability to induce local changes in membrane thickness and membrane fluidity corresponding to the hydrophobic thickness and environment requirements of an integral membrane protein located in such domain, in accordance with the mattress model (Mouri-tsen & Bloom, 1984) of lipid-protein interactions in membranes.
In conclusion, the results of monolayer and bilayer investigations show that some electrical, transport and packing properties of lipid membranes change under the influence of hexadecaprenol.The results indicate that: hexadecaprenol modulates the molecular organisation of the membrane; the specific molecular area of polyprenol molecules depends on the relative concentration of polyprenols in membranes; hexadecaprenol can modify lipid membranes by the formation of fluid micro-domains.The results also indicate that electrical transmembrane potential can accelerate the formation of pores in lipid bilayers modified by long chain polyprenols.

Figure 2 .
Figure 2. The dependence of area per molecule, A p , of the C 80 /DOPC monolayer mixture on the molar fraction of hexadecaprenol at constant surface pressures of 2, 4, 8 and 16 mN/m.

Fig 5 .
The excess free energy of mixing has the minimal value equal to -4.5 ± 1 kJ/mole for C 80 /DOPC molar fraction 0.9 and maximal value equal to 8.2 ± 1 kJ/mole for C 80 /DOPC molar fraction 0.01.By definition, the excess free energy of mixing for pure lipid monolayers equals 0 kJ/mole.Figure6Arepresents the values of collapse pressure, p C , for pure C 80 , pure DOPC and mixed C 80 /DOPC monolayers.The collapse pressure of the lipid monolayers of various C 80 /DOPC molar fractions vary from 49.9 ± 1 mN/m for pure DOPC monolayer to 26.5 ± 1 mN/m for pure C 80 monolayer.The maximal value equal to 49.1 ± 1 mN/m for p C was obtained for C 80 /DOPC molar fraction of 0.01 and the minimal value equal to 15.3 ± 1 mN/m for C 80 /DOPC molar fraction of 0.99.These results indicate a greater stability of DOPC in comparison with C 80 monolayers.Figure6Bshows the collapse area A C as a function of the C 80 /DOPC molar fraction.The collapse area varies from 41.3 ± 5 Å 2 for a pure DOPC monolayer to 22.6 ± 5 Å 2 for a pure C 80 monolayer.The maximal value equal to 63 ± 5 Å 2 of p C was obtained for C 80 /DOPC molar fraction of 0.1 and the minimal one equal to 10.2 ± 5 Å 2 for C 80 /DOPC molar fraction of 0.5-0.6.

Figure 3A .
Figure 3A.Isotherms of compression of monomolecular layers of hexadecaprenol and DOPC at the air water interface.Specific molecular areas A M are found by extrapolation of the linear parts of the isotherms to zero surface pressure.The A M values shown are averages of three experiments.B. Specific molecular area, A M , as a function of molar fractions C 80 /DOPC [0-1].

Figure 4 .
Figure 4. Limiting molecular area, A ¥ , as a function of molar fraction C 80 /DOPC [0-1].The values of A ¥ were taken for surface pressure of 0.4 mN/m.

Figure 5 .
Figure 5. Excess free energy of mixing, DG mix E , as a function of molar fraction C 80 /DOPC [0-1].DG mix E was calculated for surface pressure of 75 mN/m.The values of DG mix E were calculated according to eqn. 1 (see Materials and Methods).

Figure 9 .
Figure 9. Membrane permeability coefficient for chloride ions, P Cl -, vs C 80 /DOPC mole ratio.Each point represents the mean value ± S.D. obtained from six to eight different macrovesicular bilayer lipid membranes.Experiments were performed at 25 ± 1°C.Values of P Cl -were calculated according to the eqn.2, the values of t Cl -were calculated according to eqn. 3 (see Materials and Methods).

Figure 10 .
Figure 10.Schematic drawings of a model of the molecular organisation of polyprenol and DOPC in the monolayers investigated.
(Rolland et al., 1996) 80 /DOPC molar fraction(Rolland et al., 1996).The A ¥ values correspond to the areas occupied by the molecules at the surface pressure equal to 0.4 mN/m.The maximal A ¥ value equal to 276 Å 2 was obtained for the C 80 /DOPC molar fraction 0.01, and the minimal A ¥ value equal to 83 Å 2 for molar fraction 0.99.From eqn. 1 we have calculated the excess free energy of mixing, DG mix E , for the surface pressure range 0 to 75 mN/m.The results are shown in