Configuration of Carbonyl Groups at the Lipid Interphases of Different

May 13, 2009 - UBA Junın 956 2°P (1113) Buenos Aires, Argentina. Received February 16, 2009. Revised Manuscript Received April 20, 2009. The purpose...
0 downloads 0 Views 911KB Size
Download PDF
pubs.acs.org/Langmuir © 2009 American Chemical Society

Configuration of Carbonyl Groups at the Lipid Interphases of Different Topological Arrangements of Lipid Dispersions Marı´ a de los Angeles Frı´ as and E. Anı´ bal Disalvo* Laboratorio de Fisicoquı´mica de Membranas Lipı´dicas, Facultad de Farmacia y Bioquı´mica, UBA Junı´n 956 2°P (1113) Buenos Aires, Argentina Received February 16, 2009. Revised Manuscript Received April 20, 2009 The purpose of this work is to analyze the conformation of the carbonyl groups of acyl phospholipids at the hydrocarbon-water interphase in different topological ensembles and phase states, such as micelles and bilayers. The separation of the band components in lipids dispersed in D2O is compared with that of PCs in a low hydrated state. When hydrated, the differences in the frequencies of the band components corresponding to the carbonyl groups identified as low hydrated and hydrated populations increase when dimyristoylphosphatidylcholine (DMPC) bilayers go from the lamellar gel to the ripple corrugated phase at the pretransition temperature. Below the pretransition, at which the membrane in the gel state is planar, the two components overlap making the deconvolution unreliable. A further analysis shows that the frequency of the highly hydrated population increases more noticeable than that corresponding to the low hydrated one following the sequence: micelles, fluid phase, ripple gel phase, and lamellar gel phase. This is confirmed by the increase in the separation of the band components when the liposomes are subjected to an osmotic dehydration suggesting that the hydrated population loses water and the dehydrated one partially hydrates. It is concluded that this behavior is a feature conferred by hydration of the different topological arrangements. The relevance of these results on the interphase properties of lipid membranes is discussed.

Introduction Most of the biological functions take place in water-membrane or water-protein interphase. In particular, the water-membrane interface presents a multifacetic structure and a complex dynamical behavior. The surface of lipid membranes, even of those constituted of a pure lipid, are highly heterogeneous in their chemical composition, lateral organization, and dynamics.1 One of the reasons for this behavior is that water confers to chemical surface subgroups of the lipids a number of conformational states of different free energy. Polar lipids swell in the presence of water to take up a number of water molecules per lipid depending on the state of the acyl chains, liquid or solid crystalline, and on the nature of the polar headgroup. In this process, water and polar headgroup arrangements derived by the lateral interaction would determine the free energy of the interphase necessary for the adsorption of additives present in the aqueous environment. The water uptake is determined by the difference between the chemical potential of water in the lipids and in the adjacent solution.2 At equilibrium, the degree of hydration is then accomplished by the presence of groups and residues on the membrane surface that are able to bind water with different types of interactions. These sites have been identified as the phosphate group, the carbonyl groups, the ether oxygens, and the hydrophobic residues of the acyl chains partially exposed to the water phase.3-6 Due to the orientation, their ability to form hydrogen bonds and hence the number and the binding energy of the water molecules may differ to great extent between them. Therefore, the (1) Soderlund, T.; Alakoskela, J.-M.; Pakkanen, A. L.; Kinnunen, P. K. J. Biophys. J. 2003, 85, 2333–2341. (2) Wennerstrom, H.; Sparr, E. Pure Appl. Chem. 2003, 75, 907–912. (3) Jendrasiak, G. L. J. Nutr. Biochem. 1996, 7(11), 599–609. (4) Arrondo, J. L.; Go~ni, F. M.; Macarulla, J. M. Biochim. Biophys. Acta 1984, 794(1), 165–168. (5) Binder, H. Eur. Biophys. J. 2007, 36(4-5), 265–279. (6) Selle, C.; Pohle, W. Biospectroscopy 1998, 4, 281–294.

Langmuir 2009, 25(14), 8187–8191

variation of the partial free energies with water content would strongly depend on the degrees of freedom of the polar head groups.2 The organization of lipid bilayers in water has been studied using a wide variety of methodologies. The problem of elucidating the orientation ordering of the different molecular moieties of the phospholipids, namely, carbonyl group, phosphate group, and hydrocarbon chains, has given relevance to FTIR analysis, since this is, together with NMR, a nonperturbing technique not requiring the addition of probe molecules.7 In this regard, the study of the conformational changes at the glycerol backbone, in particular, in relation to the exposure of the carbonyl group toward the aqueous phase, is relevant to highlight the contribution of the interfacial region to the membrane organization.8 The molecular basis of the FTIR spectroscopy, being a technique that operates at very short time scale, allows insight into the degree of hydration and nature of the hydrogen bonding of water molecules local around those groups. The elucidation of the structural and dynamical properties of this polar-nonpolar region would provide new insights into the interphase properties of lipid bilayers and its relevance to membrane function. The interphase region is defined between an internal plane (the water-hydrocarbon interface) and an external plane (the hydrodynamic shear plane between the membrane and the bulk aqueous phase).9 The inner plane divides the low dielectric constant region of the hydrocarbon phase from the polar region where carbonyl groups are localized. In this context, it becomes particularly important to consider the molecular nature of the (7) Casal, H.; Mc Elhaney, R. N. Biochemistry 1990, 29, 5423–5427. (8) Lewis, R. N. A. H.; Pohle, W.; McElhaney, R. N. Biophys. J. 1996, 70, 2736–2746. (9) Disalvo, E. A.; Lairion, F.; Martini, F.; Tymczyszyn, E.; Frı´ as, M.; Almaleck, H.; Gordillo, G. J. Biochim. Biophys. Acta 2008, 1778, 2655–2670.

Published on Web 05/13/2009

DOI: 10.1021/la900554h

8187

Article

headgroup, which is supposed to interact directly with the water. There are indications that, within this region, lipid molecules have conformational freedom in the liquid and the condensed state, although they are confined to a lamellar arrangement by basically the hydrophobic interaction and, in some extension, by water bridges and H-bonds between the headgroups. Water may penetrate into the lipid membrane in a non-homogeneous way, affecting the first C atoms in the glycerol backbone.10,11 These water defects or fingers help in understanding the stabilization of polar molecules in the membrane phase, suggesting that the limit of the water-hydrocarbon interface is not a clear-cut plane.12 One possibility opposing the sharpness of the inner plane of the interphase region is the occurrence of fluctuations in the water penetration beyond that plane and, as a consequence, of the exposure of hydrophobic regions and polar residues to water. The nature of these fluctuations has been ascribed to undulations involving local changes in area and water thickness.13 In addition to phosphate groups, carbonyls in the ester union are centers for water immobilization.14-16 The carbonyl groups are important constituents of the membrane interphase, since they contribute to the dipole potential by its own dipole and by the water that they polarize by H bond interactions.17-19 From the structural point of view, carbonyl groups as a center of hydration show complex behavior. Hubner and Blume14-16 and references therein suggested that the resolvable subcomponents of the νCdO bands of diacyl phospholipids may reflect subpopulations of hydrogenbonded and non-hydrogen-bonded ester carbonyl groups. This suggestion has been adopted in the interpretation of the νCdO bands in several studies of diacylglicerolipid bilayers.20-22 For fully hydrated phosphatidylcholines, the high- and low-frequency components of the νCdO bands cannot be assigned to the individual ester carbonyl group of acyl chains esterified at the sn1 and sn2 positions of the glycerol backbone. These studies indicate that the carbonyl groups at the sn1 and sn2 positions in phospholipids split into two populations: one hydrated (bound) and another nonhydrated (nonbound).15,16 This separation can be related to local fluctuations making the surface more heterogeneous at the membrane plane according to the water available for the CO groups establishing equilibrium between different topological arrays.23 From the structural view, an organization of a few water molecules in a pocket could constitute a transient defect restricted to carbonyl orientation at the water-membrane interface. Lipids can stabilize in different topological arrangements and phase states for which it is likely a different water organization. In this context, a change in the water activity may produce local changes of membrane packing leading to the formation of defects in the membrane interface. The local changes of water hydrogen (10) Simon, S. A.; McIntosh, T. J. Methods Enzymol. 1986, 127, 511–521. (11) Simon, S. A.; McIntosh, T .J. Proc. Natl. Acad. Sci. U.S.A. 1989, 86, 9263–9267. (12) MacCallum, J. L.; Bennett, W. F.; Tieleman, D. P. Biophys. J. 2008, 94(9), 3393–3404. (13) Nagle, J.; Tristam-Nagle, S. Biochim. Biophys. Acta 2000, 1469, 159–195. (14) Blume, A.; Hubner, W; Messner, G. Biochemistry 1988, 27, 8239–8249. (15) Hubner, W.; Blume, A. Chem. Phys. Lipids 1998, 96(1-2), 99–123. (16) Lewis, R. N. A. H.; McElhaney, R. N.; Pohle, W.; Mantsch, H. H. Biophys. J. 1994, 67, 2367–2375. (17) Gawrisch, K.; Ruston, D.; Zimmerberg, J.; Parsegian, V. A.; Rand, R. P.; Fuller, N. Biophys. J. 1992, 61, 1213–1223. (18) Diaz, S.; Amalfa, F.; Biondi, A. C.; Disalvo, E. A. Langmuir 1999, 15, 5179–5182. (19) Lairion, F.; Disalvo, E. A. J. Phys. Chem. B 2009, 113(6), 1607–1614. (20) Pohle, W.; Selle, C.; Fritzsche, H.; Bohl, M. J. Mol. Struct. 1997, 408/409, 273–277. (21) Pohle, W.; Gauger, D. R.; Fritzsche, H.; Rattay, B.; Selle, C.; Binder, H.; Bohlig, H. J. Mol. Struct. 2001, 563/564, 463–467. (22) Binder, H.; Peinel, G. J. Mol. Struct. 1985, 123, 155–163. (23) Heerklotz, H.; Epand, R. M. Biophys. J. 2001, 80(1), 271–279.

8188 DOI: 10.1021/la900554h

Frı´as and Disalvo

bonding may be reflected in the excess surface free energy of the water layer deformation contributing to the adsorption processes of compounds in the adjacent water phase, such as sugars, amphiphiles, and amino acids. In this regard, OH bonding compound molecules, inhibiting the detergent action of lysophosphatidylcholines on lipid membrane, are able to interact with carbonyl groups.24-26 For these reasons, the purpose of this work is to analyze the exposure of the carbonyl group populations at the hydrocarbon-water interphase in different topological and phase changes within the time scale of infrared spectroscopy.

Materials and Methods Lipids and Chemicals. Dimyristoylphosphatidylcholine (DMPC) and monomyristoylphosphatidylcholine (LysoPC) were obtained from Avanti Polar Lipids, Inc. (Alabaster, AL). Arbutin (4-hidroxyphenyl-β-D-glucopiranoside) and D2O were from Sigma-Aldrich (Saint Louis, MO). The purity of the lipid and the arbutin was checked by running the FTIR spectra of lyophilized samples and by differential scanning calorimetry (DSC). All other chemicals were of analytical grade and triple-distilled in water. FTIR Measurements. An FTIR Nicolet 380 spectrophotometer, provided with a DTGS detector, was used. Bands were obtained in KBr disks at a relative humidity (RH) of 20%. The water content of the lipid films was estimated by means of the spectral parameter defined as the ratio of the integral absorbance of the ν1,3OH band of water centered near 3400 cm-1 and of the integral absorbance of the C-H stretching region (3000-2750 cm-1) after baseline correction. The value obtained in this condition was coincident with that reported by Pohle et al.27 and was taken as the value corresponding to lipid in the lower hydration state. For fully hydrated samples, 3-5 mg of the dried sample was mixed with 30-50 μL of D2O by vigorously vortexing at temperatures above the Tm of lipid. The dispersion was squeezed between two AgCl windows and was mounted in a cell. This cell was placed in the holder. A total of 64 scans with a resolution of 1 cm-1 were done in each condition, and the spectra were analyzed using the mathematical software provided by the instrument. A number of different samples (no fewer than 3) were processed in order to obtain a standard deviation below the resolution of the equipment. In this condition, the values of the peaks obtained were assigned to lipid in the higher hydration state. In these cases, spectra consisted of broad overlapping bands, so Fourier deconvolution was used to estimate the frequencies of the component bands, followed by curve-fitting to obtain bandwidth and intensity (band narrowing factors: 1.6-2.2). For the osmotic stress condition, the lipids were dispersed in D2O and the liposomes formed were then suspended in a solution of D2O containing 10 mM arbutin. In this condition, an osmotic gradient due to the presence of arbutin in the outer solution was imposed across the lipid membrane. The measurements were carried out at 18 °C.

Results In Figure 1, the bands corresponding to lipids equilibrated with 20% relative humidity (A) and fully hydrated lipids (B) are shown. Lipids at low humidity give two clear bands: one at 1741 and another at 1728 cm-1. The fully hydrated lipids show bands at similar frequencies after the Fourier deconvolution and curve-fitting described in Materials and Methods: one centered at 1741.2 cm-1 and another at 1725.7 cm-1. (24) Disalvo, E. A.; Viera, L. I.; Bakas, L. S.; Senisterra, G. A. J. Colloid Interface Sci. 1996, 178(2), 417–425. (25) Diaz, S. B.; Biondi, A. C.; Disalvo, E. A. Chem. Phys. Lipids 2003, 122, 153–158. (26) Frı´ as, M. A.; Nicastro, A.; Casado, N.; Gennaro, A.; Dı´ az, S. B.; Disalvo, E. A. Chem. Phys. Lipids 2007, 147, 22–29. (27) Pohle, W.; Selle, C.; Fritzsche, H.; Binder, H. Biospectroscopy 1998, 4(4), 267–280.

Langmuir 2009, 25(14), 8187–8191

Frı´as and Disalvo

Article

Figure 2. Differences between the frequencies of bound and unbound carbonyl populations for different topological and phase states of DMPC lipid organizations.

Figure 3. Changes in the carbonyl frequency band of bound (9) and unbound ([) populations for the different topological and lamellar phase states of lysoPC and DMPC. Figure 1. FTIR spectral bands of carbonyl groups of DMPC in 20% RH and in D2O suspensions. A comparison of the CdO stretching regions of the IR spectra of the 20% RH (A) and ripple (B) phases of DMPC (14:0) samples. The solid line in part B represents the contours of the spectra acquired, and the dashed lines represent our estimates of the position and relative intensities of the component bands after deconvolution and fitting.

In Figure 2, the separation of the component bands in fully hydrated lipids is compared for different topological and phase states. The main observation is that the higher separation of the component bands correspond to the lamellar fluid phase, followed by the gel (ripple) state. However, the band separation in micelles is near the gel (planar) phase. For comparison, the band splitting in solid is included. This value is below the error of deconvolution (8 cm-1 band separation) and does not allow identification of two populations between the carbonyl groups. In Figure 3, the values for the bound and nonbound populations are compared for the different phase and topological states. The frequency of the nonbound populations (i.e., the less hydrated) maintains its value in the lamellar conformation independent of the phase state. In contrast, the frequency corresponding to the hydrated population increases following the sequence fluid < ripple< planar< solid, converging to the value of the less hydrated population. The frequency values in lysoPC micelles converge to that corresponding to the hydrated population. Langmuir 2009, 25(14), 8187–8191

Figure 4. Frequencies for the unbound and bound carbonyl populations for DMPC liposomes dispersed in arbutin above the pretransition temperature and subjected to an osmotic stress.

In Figure 4, we compare the band frequencies and the band separation of the pure lipids of DMPC in the gel (first two columns) and the ripple (third two columns) states shown in Figure 2. It is observed that when the lipids go through the pretransition the band separation increases from 7 cm-1 to 18 cm-1. DOI: 10.1021/la900554h

8189

Article

Frı´as and Disalvo

In the presence of a compound that is able to avoid the pretransition, that is, to eliminate the corrugated phase at 18 °C, such as arbutin (a hydroxyphenyl glucopyranoside),25 the difference in the values corresponding to the low hydrated and highly hydrated populations is 8 cm-1, very similar to that for the pure DMPC gel (planar) phase (10 °C). In addition, also in Figure 4 it is observed that when liposomes in the gel phase at 10 °C are subjected to an osmotic stress, due to an asymmetric distribution of arbutin between the two sides of the membrane (see Materials and Methods) at the same temperature, the difference in the frequencies corresponding to the two populations approach that for the ripple phase of pure lipids (four two columns).

Discussion As reported elsewhere, the Fourier deconvolution of the bands for lipids in the anhydrous state does not allow discrimination between two populations, because the separation between them is 4 cm-1, i.e., much less than 8 cm-1, the limit accepted for a realistic deconvolution.27 However, the contact of lipids with 20% RH of water vapor presents a clear band separation (Figure 1A). Therefore, the band separation observed in Figure 1 is a process linked to lipid swelling, i.e., lipid hydration allows stabilization of carbonyl groups in two populations: one at low frequencies and an another at the higher frequency. When dispersed in D2O in order to avoid the overlapping of H2O bands, and after Fourier deconvolution and curve fitting, similar components are apparent (Figure 1B). It is likely that the carbonyl groups at the lower frequency are forming H bonds with the water molecules. An indication of the water-carbonyl interaction is given by the measures of dipole potential. When water is displaced by hydrogen bonding compounds, the dipole potential decreases. This has been ascribed to the contribution of the dipole of water molecules attached to carbonyl groups.17 This is sustained by the fact that dehydration of carbonyl groups by the insertion of OH bonding compounds decreases the dipole potential, which is due to polarized water.28-30 A correlation between dipole potential and the number of water molecules displaced by OH bonding compounds has been published recently.9 In previous work, the hydration of the carbonyl groups was studied by following the intensity changes at the main phase transition of dipalmitoylphosphatidylcholine suspensions. The importance of the significant change in the relative intensities resides in the level of carbonyl groups in each population.15,16 However, the shift in the positions of the band components gives additional insight into the changes affecting the interphase by the phase transition. It suggests that the binding strength of the carbonyl groups with the water molecules may vary according to the phase state, which may imply a different orientation of the carbonyl dipole toward the water phase. Moreover, for a given phase state, the more pronounced change is observed in the unbound (less hydrated) population when topology of the ensemble changes from micelles to lamellar phase. Thus, the topological arrangements with large curvature radius such as micelles do not have a different degree of hydration level between carbonyl groups. However, they are well below that corresponding to the solid lipid, indicating that they are all bounded to water. This may be a consequence that in the lysoPC (28) Luzardo, M. C.; Amalfa, F.; Nun~ez, A.; Dı´ az, S.; Biondi, A. C.; Disalvo, E. Biophys. J. 2000, 78, 2452–2458. (29) Lairion, F.; Martini, F.; Diaz, S.; Disalvo, E. A. Langmuir 2002, 18(17), 6716–6717. (30) Lairion, F.; Disalvo, E. A. Langmuir 2004, 20, 9151–9155.

8190 DOI: 10.1021/la900554h

only one stereo chemical chain is present and thus no lateral effect of intramolecular chain-chain interaction takes place. As these micelles are formed by a single chain lipid, it is likely that the compromise between acyl chain packing and headgroup repulsion would result in direct contact of all carbonyl groups to the water phase. With regard to the values of the carbonyl bands in lamellar phases, it is clear from Figure 3 that the band separations observed in Figure 2 along the phase states are due to a shift of the hydrated population to higher values, which would mean lower exposure to water or localization in a less polar ambient when the membrane goes from the fluid to the gel phase. The unbound population has comparable values for all the phase states, although a slight shift to higher frequencies is insinuated for the ripple phase in comparison to gel and fluid phases (Figure 3). When comparing the results of ripple phase with those of gel phase, it appears that the same carbonyl arrangement can be induced by a temperature decrease to below the pretransition or by the insertion of an OH compound at 18 °C. A shift from 10 to 18 °C corresponds to the appearance of ripples with a concomitant increase in the band separation (compare columns 1 and 3). Above the pretransition temperature, the bilayer is stabilized in the presence of a compound that allows the band separartion to remain unchanged (columns 1 and 2). That is, the same carbonyl arrangement is obtained. This clearly suggests that the transition from the gel to the ripple phase is related to hydration as previously stated.31,32 However, these new data provide evidence that the hydration change is taking place at the carbonyl region. Similar responses have been obtained with compounds that affect the lipid interphase by substituting water.28-30 Hence, water distribution seems to be linked to the topological arrangements of the carbonyl groups at the water-hydrocarbon interphase, in this case, the presence of ripples. Additional evidence relating a topological change with hydration is given by the results in which the bilayer is subjected to osmosis. In this case, the band separation of column 3 is comparable to that of column 4 in Figure 4. Column 3 corresponds to the ripples and column 4 to a membrane which, in the absence of ripples, is subjected to an osmotic stress. Therefore, osmosis can induce similar carbonyl group exposures in the gel phase than that occurring in the spontaneous curvature of the ripple phases. However, the curvature induced by osmosis produces a shift to lower frequencies of the two populations, denoting increasing hydration. In contrast, when in isosmotic conditions the membrane is shifted from the gel (planar) (first two columns) to the ripple phase (spontaneous curvature) (third two columns); the decrease is only noted in the less hydrated population. The appearance of the ripple or corrugated phases (Pβ0 phase) has been correlated with the level of hydration31 in which osmotic stress can play a role.32 In addition, although there is extensive discussion in the literature on the ordering of headgroups related to the different hydrations of the bilayer,33,34 the present results denote differential carbonyl group exposure as a function of curvature and hydration. In micelles, both populations are largely exposed to water, probably by the absence of chain-chain interaction in the same molecule. In ripple phases, only one population can be considered exposed to water, and when the bilayer is subject to osmotic stress, both carbonyl population (31) (32) (33) (34)

Banerjee, S. Physica A 2002, 308, 89–100. Heimburg, T. Biophys. J. 2000, 78, 1154–1165. Mingtao, G.; Jack Freed, H. Biophys. J. 2003, 85(6), 4023–4040. Ho, C.; Slater, S. J.; Stubbs, C. Biochemistry 1995, 3, 6188–6195.

Langmuir 2009, 25(14), 8187–8191

Frı´as and Disalvo

appears to localize in a more polar media than in the absence of osmosis. The effects of water activity modification by osmosis and its consequences on the interface properties have been reported by other methodologies.35 The present analysis shows that there are two extreme cases in which the band separation is at its lowest value. One corresponds to micelles in which the values converge to the lowest frequency, suggesting that all the carbonyls are hydrated. In the other extreme, the values converge to the higher frequency corresponding to solid anhydrous lipids. Little change (35) Lehtonen, J. Y.; Kinnunen, P. K. Biophys. J. 1995, 68, 525–535.

Langmuir 2009, 25(14), 8187–8191

Article

is observed in the band separation between the fluid and the ripple phases. This is congruent to that reported for DPPC.16 However, different arrangements can be found between lipids in the gel and the spontaneous (ripple) or forced curvature (osmotic stress) when the changes in the components are analyzed. In consequence, the difference in frequencies between the two populations depends on the topological phase state of the surface induced by temperature shift and/or osmotic stress. This highlights the critical role of interfacial groups as a determinant of the organization of water at lipid assemblies at different mechanical and hydration stresses.

DOI: 10.1021/la900554h

8191