Effect of Cholesterol on the Formation and Hydration Behavior of Solid

Sep 17, 2009 - Daniela Pozzi,*,† Ruggero Caminiti,† Carlotta Marianecci,‡ Maria Carafa,‡ Eleonora Santucci,‡. Sofia Candeloro De Sanctis,†...
0 downloads 0 Views 937KB Size
pubs.acs.org/Langmuir © 2009 American Chemical Society

Effect of Cholesterol on the Formation and Hydration Behavior of SolidSupported Niosomal Membranes Daniela Pozzi,*,† Ruggero Caminiti,† Carlotta Marianecci,‡ Maria Carafa,‡ Eleonora Santucci,‡ Sofia Candeloro De Sanctis,† and Giulio Caracciolo*,† †

Chemistry Department, Sapienza University of Rome, P. le A. Moro 5, 00185 Rome, Italy, and ‡Dipartimento di Chimica e Tecnologie del Farmaco, Faculty of Pharmacy, Sapienza University of Rome, P. le A. Moro 5, 00185 Rome, Italy Received July 22, 2009. Revised Manuscript Received September 1, 2009

The effect of cholesterol on the formation and hydration behavior of solid-supported polysorbate 20 (Tween 20)/ cholesterol self-assemblies was investigated by means of in situ energy-dispersive X-ray diffraction in a wide range of relative humidity (0.4 0.986, both scattering distributions showed a simultaneous loss in intensity (Figure 5, panels a and b) and broadening of the Bragg peaks. These findings are due to “second-order disorder”22 promoted by free water molecules that penetrate the multibilayer systems and produce dynamical fluctuations of the bilayers around well-defined mean layer positions (affecting the intensity of Bragg peaks) as well as small variations in the bilayer separations (affecting the width of Bragg peaks).22,28 As a result of such motions, the multilayered structure fluctuates from instant to instant and such fluctuations are likely to contribute to stacking disorder. Some further considerations can be stated about the origins of the driving force for cholesterol solubilization. When the cholesterol content is high enough to exceed the maximum solubility in the membrane (a mole fraction that we refer to as (Xchol)*), excess cholesterol molecules leave the bilayer as a precipitate of cholesterol crystals. According to Huang and Feigenson,29 cholesterol in a crystal, as a pure substance, is supposed to have a constant chemical potential, (μchol)crystal. In contrast, the chemical potential of cholesterol in a bilayer, (μchol)bilayer, is a function of the bilayer composition and is expected to increases with increasing (Xchol). In a bilayer, the polar headgroups act like umbrellas, (25) Cheng, J. X.; Pautot, S.; Weitz, D. A.; Xie, X. S. Proc. Natl. Acad. Sci. U.S. A. 2003, 100, 9826. (26) Binder, H.; Kohlstrunk, B.; Pohle, W. J. Phys. Chem. B 2000, 104, 12049. (27) Binder, H. Eur. Biophys. J. 2007, 36, 265. (28) Pabst, G.; Rappolt, M.; Amenitsch, H.; Laggner, P. Phys. Rev. E 2000, 62, 4000. (29) Huang, J.; Feigenson, G. W. Biophys. J. 1999, 76, 2142.

Langmuir 2010, 26(4), 2268–2273

shielding the nonpolar part of cholesterol molecules from water. For (Xchol) < (Xchol)*, (μchol)bilayer must be less than (μchol)crystal for the bilayer to be the only stable phase.29 As the concentration of cholesterol increases, headgroups are stretched to provide coverage for the increasing number of cholesterols. When (Xchol) g (Xchol)*, the headgroups are stretched to their limit and can no longer provide shielding for all cholesterol molecules. The bilayer phase now contains the maximum number of cholesterols that the headgroups can cover.29 Some sterol molecules are therefore exposed to water, causing a sharp jump in the chemical potential of cholesterol. To lower the free energy, excess cholesterol molecules precipitate and form a second phase consisting of cholesterol crystals. Starting from this point, the bilayer phase and the cholesterol crystal phase coexist. In this two-phase region, the chemical potential of cholesterol in the bilayer must be equal to the chemical potential of cholesterol in the crystal. Our results on Tween 20/cholesterol mixed membranes can be therefore rationalized as follows. At low hydration (RH < 0.985), there are no niosomal membranes and almost no cholesterol binding sites. Accordingly, only cholesterol crystals are found to exist. When niosomal membranes form, (Xchol) is much lower than (Xchol)*, which means that (μchol)bilayer is much lower than (μchol)crystal. As a result, crystal solubilization is thermodynamically favored and cholesterol monomers can bind to Tween 20 molecules. Upon hydration, the number of cholesterol molecules associated with Tween 20 increases progressively, with the result that both (Xchol) and (μchol)bilayer increase. Lastly, when (Xchol) approaches (Xchol)*, (μchol)bilayer equals (μchol)crystal and crystallites are no longer solubilized. At full hydration (RH = 1, data not reported), the lamellar d spacing of solid-supported niosomal membranes (d ≈ 90 A˚) was found to be the same as their counterpart in aqueous solution.9 Furthermore, niosomal membranes were still found to coexist with unsolubilized cholesterol, thus showing that the limiting fraction of cholesterol in the Tween 20/cholesterol mixture is Xchol < 0.5. As evident, the hydration behavior of niosomal membranes appeared to be a rather interesting process. With the aim of providing accurate insight into the effect of hydration on the structure of niosomal bilayers on the nanoscale, EDPs along the normal to niosomal membranes were calculated. Figure 6 shows the EDP along the normal to niosomal membranes at RH=0.985. DOI: 10.1021/la9026877

2271

Article

Pozzi et al. Table 1. Interfacial Area, A, per Surfactant Molecule as a Function of RH, Calculated from Equation 2 RH

A (A˚2)

0.986 0.987 0.988 0.989 0.990 0.991 0.992 0.993 0.994

20.6 19.9 20.1 20.8 22.9 26.7 27.9 31.6 34.4

Figure 7. Structural parameters of niosomal membranes as a function of RH.

The electron densities of the headgroups, FH, and hydrocarbon tails, FC, are defined relative to the methylene electron density FCH2 (0.317 e/A˚3).28 The EDP of Figure 6 exhibits two strong Gaussians corresponding to the headgroups of Tween 20, whereas the electron-sparse region at the methyl terminus of the hydrocarbon chains is represented by another Gaussian with negative amplitude. We observe that the EPD along the normal to niosomal membranes is quite different from that of singlecomponent membranes.16,30-33 In fact, the electron density of polar Tween 20 headgroups at the edges of the profile is lower than that at the hydrophilic-hydrophobic interface (inflection point in the EDP). According to the recent results by Pandit et al.,34 who reported on the EDPs of phospholipid/cholesterol mixtures, we suggest that such an increase in electron density in the hydrophobic core (HC) is due to the presence of cholesterol (minor Gaussians in the EDP). Indeed, cholesterol is mainly positioned in the HC of niosomal membranes, with the hydrated -OH group of cholesterol being roughly located at the hydrophilic-hydrophobic interface.34,35 According to recent definitions,28 all of the relevant structural parameters of bilayer membranes can be derived from the EDP. The headgroup size, dH, can be estimated from the fwhm of the Gaussian representing the headgroup (Figure 6). As a result, the bilayer thickness of niosomal membranes, dB, is the sum of the distance between the maxima of the EDP (headgroup-headgroup distance, dHH) plus the thickness of the headgroup size including water molecules associated with the hydration shell of Tween 20 headgroups, dH (dB = dHH þ dH).28 According to geometric considerations, the thickness of the interbilayer water region can be calculated as dW = d - dB.28 Because of the possible low resolution of the EDP, the definition of the boundary between the membrane and the (30) Tristram-Nagle, S.; Petrache, H.; Nagle, J. F. Biophys. J. 1998, 75, 917. (31) Tristram-Nagle, S.; Nagle, J. F. Chem. Phys. Lipids 2004, 127, 3. (32) Liu, Y.; Nagle, J. F. Phys. Rev. E 2004, 69, 040901(R). (33) Kucerka, N.; Liu, Y.; Chu, N.; Petrache, H. I.; Tristram-Nagle, S.; Nagle, J. F. Biophys. J. 2005, 88, 2626. (34) Pandit, S. A.; Bostick, D.; Berkowitz, M. L. Biophys. J. 2004, 86, 1345. (35) Mills, T.; Huang, J.; Feigenson, G. W.; Nagle, J. F. Gen. Physiol. Biophys. 2009, 28, 126.

2272 DOI: 10.1021/la9026877

Figure 8. Total number of water molecules, nW, per surfactant headgroup as a function of RH. Solid lines are the best linear fits to the data. The dashed line separates two hydration regimes. The break, which occurs at 10 water/surfactant, is most likely due to the point of completion of the Tween 20 hydration shell.

interbilayer water region is often not trivial; consequently, the calculation of dH can be difficult. However, this is not the case for our EDXD measurements. Indeed, the high number of Bragg reflections observed (Figure 3) allowed us to obtain high-resolution EDPs (Figure 6). Figure 7 shows the changes in the structural parameters of niosomal membranes as a function of increasing RH. The lamellar d spacing (Figure 7, panel a) increases monotonously with RH without reaching a finite swelling limit. Such indefinite swelling is due to the electrostatic repulsion between opposing membranes, as reported several times in the literature.36 This finding is in good agreement with previous zeta-potential measurements on niosomes showing that vesicles have a net negative charge.10 In Figure 7, we can observe two different hydration regimes (separated by a dashed line). In the first regime, the water layer thickness, dW, exhibited minor changes (Figure 7, panel b) and the membrane thickness, dB, increased with RH until it reached its maximum expansion for RH ≈ 0.988 (dB ≈ 70 A˚). In the second regime, the membrane structure was quite insensitive to hydration, as shown by the fact that dB remains roughly constant. However, the interlamellar water layer increased remarkably. Hydration is known to be a complex process that is regulated predominantly by short-range headgroup-headgroup, headgroup-solvent, van der Waals, and long-range interbilayer electrostatic interactions. Such interactions can affect the interfacial area of surfactant molecules remarkably. To gain deeper insight into the hydration process, we calculated the area per surfactant molecule as a function of RH. Recently, Pabst and co-workers calculated the area per lipid, A, on the basis of the (36) Pozo-Navas, B.; Raghunathan, V. A.; Katsaras, J.; Rappolt, M.; Lohner, K.; Pabst, G. Phys. Rev. Lett. 2003, 91, 028101.

Langmuir 2010, 26(4), 2268–2273

Pozzi et al.

Article

Figure 9. Schematic sketch of the hydration behavior of solid-supported niosomal membranes. (a) In the early stages of formation (RH ≈ 0.985), niosomal membranes coexist with large cholesterol crystallites. (b) Upon hydration, both the membrane thickness and the water layer increase. Upon hydration, water adsorption promotes the solubilization of a large part of the crystals.

integration of the EDP.28 The developed full-q refinement model has the advantage of deriving A without further information on the specific volume of the lipid. By applying this method, the thermal behavior of phosphatidylcholine and phosphatidylethanolamine membranes could be investigated in great detail.37 To determine the area per surfactant molecule, the formalism of Lemmich et al.38 was applied, which yields   1 FnC nH A¼ FCH2 ðF -1Þ dC dH

ð2Þ

where FCH2 is the methylene electron density (0.317 e/A˚3), F = (FH - FCH2)/(FC - FCH2) is the ratio between the electron density of the headgroup (FH) and that of the hydrocarbon tails (FC), both relative to the methylene electron density, nC is the number of hydrocarbon electrons, nH is the number of headgroup electrons, dC is the hydrocarbon chain length, and dH is the headgroup size. In our calculation, we set nC=158 and nH=136, and F, dC, and dH were retrieved from the EDPs. This method, as discussed by Rappolt et al.,37 could lead to an underestimation of the area per lipid, but the relative changes should be the same. The calculated values are reported in Table 1. According to previous findings, the interfacial area was found to increase with RH.39 Upon hydration, nonionic Tween 20 molecules exhibit a preferential adsorption of hydrochloride ions40 with the result that the headgroups acquire a net negative charge. To minimize unfavorable surfactant-surfactant electrostatic repulsions, water molecules align their dipole moments toward the water bulk region, thus creating a positive potential.41 Such water molecules closely interacting with the bilayer surface are strongly ordered,41 and the surfactant headgroups are tightly packed. After fulfilment of the first hydration shell, a large interface network of hydrogen bonds between water and surfactants forms,41 resulting in a drastic expansion of the surfactant interfacial area.39,42 Another parameter of interest in hydration studies is the total number of water molecules including the molecules intercalated into the bilayer, nW. This can be estimated from the EDP by geometrical considerations28 nW ¼

  A d dHH VW 2 2

ð3Þ

where VW is the volume of one water molecule (VW ≈ 30 A˚3).28 Figure 8 shows the number of waters per surfactant molecule, nW, as a function of RH. We observe a distinct break in the curve at about 10 waters/surfactant molecule. According to the literature,42,43 the break is supposed to result from a structural change (37) Rappolt, M.; Laggner, P.; Pabst, G. Recent Res. Dev. Biophys. 2004, 3, 363. (38) Lemmich, J.; Mortensen, K.; Ipsen, J. H.; Hoenger, T.; Bauer, R.; Mouritsen, O. G. Phys. Rev. E 1996, 53, 5169. (39) Binder, H.; Gutberlet, T.; Anikin, A.; Klose, G. Biophys. J. 1998, 74, 1908. (40) Uchegbu, I. F.; Florence, A. T. Adv. Colloid Interface Sci. 1995, 58, 1. (41) Cheng, J. X.; Pautot, S.; Weitz, D. A.; Sunney, X. Proc. Natl. Acad. Sci. U.S.A. 2003, 19, 17. (42) Hristova, K.; White, S. H. Biophys. J. 1998, 74, 2419. (43) Caracciolo, G.; Pozzi, D.; Caminiti, R. Appl. Phys. Lett. 2007, 90, 183901.

Langmuir 2010, 26(4), 2268–2273

accompanying the completion of the filling of the Tween 20 hydration shell. This interpretation is strongly enforced by the structural changes in niosomal membranes occurring at the same RH (Figure 7). The observation of two different hydration regimes is in very good agreement with recent findings on the water sorption of lipid membranes.26 In these studies, the existence of distinct hydration regimes has been related to a change in water properties with increasing distance from the membrane plane. The latter phenomenon reflects the fact that the water initially incorporates into a highly ordered water structure of the primary hydration shell. Such a shell of adsorbed water differs from the subsequent layers of water because the polar headgroups of surfactant molecules strongly interact with this first layer via relatively strong water-substrate (solvent-solute) interactions.27 In such a scenario, a well-structured, low-density interfacial ‘‘network’’ of structuring waters connects the HC of the bilayer and the unperturbed bulk water. The first adsorbed water molecules can also modulate the arrangement of lipid and surfactant molecules with the headgroup locked in a highly ordered, stable structure.41 The observed process (schematic sketch given in Figure 9) completely reversed upon sample dehydration (data not reported).

4. Conclusions In this work, we investigated the effect of hydration on the structure of equimolar Tween 20/cholesterol self-assemblies over a wide range of relative humidity (0.4 < RH < 1). At low hydration, Tween 20 and cholesterol demixed with the latter molecules forming crystallites with a pseudobilayer structure (d ≈ 34 A˚). Water sorption promoted the solubilization of cholesterol crystals. When in the presence of enough cholesterol, the formation of niosomal bilayer membranes occurred (RH ≈ 0.985). Upon further hydration, two hydration regimes were identified. First, the membrane thickness increased with minor changes in the interlamellar water layer. In the second regime (RH < 0.988), the interfacial area and the total number of waters per surfactant molecule were found to increase, whereas the bilayer thickness was practically insensitive to hydration. From RH=0.988 up to full hydration, a strong increase in the lamellar d spacing was due to the enlargement of the interlamellar water layer after the completion of the Tween 20 hydration shell. Under the biologically relevant full hydration condition, niosomal membranes were found to coexist with unsolubilized cholesterol crystals and exhibited the same repeat distance as niosomes vesicles in aqueous solution.9 Aside from the conclusions of the present study, the latter observation indicates that in situ EDXD is a powerful tool for investigating the supramolecular structure of a number of surfactant/cholesterol mixtures on the nanoscale. EDXD experiments are therefore expected to provide insights into the aggregational state and hydration behavior of cholesterol molecules in surfactant-cholesterol and phospholipid-cholesterol model membranes. DOI: 10.1021/la9026877

2273