Changes in Dipalmitoyl Lecithin Multilayers (gel-liquid crystral transition/noncooperative/transition temperature)

Transcription

1 Proc. Nat. Acad. Sci. USA Vol. 68, No. 7, pp , July 1971 Laser Raman Investigation of the Effect of Cholesterol on Conformational Changes in Dipalmitoyl Lecithin Multilayers (gel-liquid crystral transition/noncooperative/transition temperature) JOSEP L. LIPPERT AND WARNER L. PETICOLAS Department of Chemistry, University of Oregon, Eugene, Oregon Communicated by Terrell L. ill, April 26, 1971 ABSTRACT Large and abrupt changes are observed at 380C in the 1100 cm region of the Raman spectrum of aqueous dipalmitoyl lecithin multilayers. They correspond to conformational changes due to the melting of the paraffin side chains. The addition of cholesterol to the multilayers broadens but does not abolish these changes. It is suggested that the addition of cholesterol decreases the interactions between adjacent paraffin side chains of lecithin, causing a change from a cooperative to a noncooperative gel-liquid crystal transition. Removal of water from dipalmitoyl lecithin also results in a noncooperative transition strikingly similar to that caused by addition of cholesterol. Raman spectroscopy thus provides a new and sensitive probe for analyzing the structures of membranes and their constituents. Cholesterol has long been recognized as an important component of biological membranes, yet the role of cholesterol in membranes remains unclear. Indeed, recent intensive investigations (1-7) of model systems have suggested that cholesterol acts in different ways, depending on the fatty acid composition of the model. On the one hand, experiments with multilayers of egg yolk lecithin, which contains many unsaturated paraffin chains, suggest that cholesterol significantly rigidizes that lipid. NMR signals are strongly reduced by the presence of cholesterol (1), which suggests that chain motion has been hindered. Likewise, recent electron spin resonance studies of chain motion in egg lecithin using both fatty acid (2) and cholestane (3) spin labels indicate a decrease in the mobility of the labels when cholesterol is added. In addition, it has been shown that the water conductivity of egg lecithin films falls with increased cholesterol content (4), which again suggests increased rigidity in the film. On the other hand, model systems containing saturated paraffin chains appear to be fluidized by cholesterol at room temperature. Shah and Schulman (5) observed a change from solid to liquid type films when cholesterol was added to dipalmitoyl lecithin monolayers. Joos (6) has demonstrated a reduction in the surface viscosity of distearoyl lecithin multilayers with added cholesterol. In addition, the x-ray long spacings of dipalmitoyl lecithin in water at 25 C decrease as the cholesterol concentration increases (7), which suggests that the paraffin side chains have become more fluid. Ladbrooke, Williams, and Chapman (7) have shown that addition of cholesterol to dipalmitoyl lecithin in water lowers the transition temperature, T,, between the gel and liquid crystal phase, and decreases the heat absorbed in the transition. They suggest that cholesterol controls the fluidity of the hydrocarbon chains of the phospholipid by disruption of 1572 the crystalline chain lattice of the gel phase (fluidization), and by inhibiting the flexing of chains in the liquid crystalline phase (rigidization). This suggestion provides an explanation for the apparently differing results obtained when cholesterol is added to various model systems, since experiments showing solidification of egg lecithin by cholesterol were done above the T, of that system while experiments indicating fluidization of dipalmitoyl lecithin were done below the T, of dipalmitoyl lecithin in water. Additional evidence for a change in the effect of cholesterol above and below the liquid crystal transition temperature is provided by DeGier, Mandersloot, and van Deenen (8), who show that below a certain temperature (about 6VC below the T, of the system), cholesterol added to liposomes enhances the rate of penetration of glycol and glycerol while above that temperature the reverse effect is noted. In this paper we present evidence obtained from Raman spectroscopy indicating that the effect of cholesterol on dipalmitoyl lecithin multilayers is to change the sharp, cooperative gel-liquid crystal transition to a diffuse, noncooperative event. As a result, the apparent effect of cholesterol on the multilayers is different above and below the transition temperature. METODS 1,2-Dipalmitoyl-Di-lecithin and z grade cholesterol were obtained from the Sigma Chemical Co. No major contaminants were detected by thin-layer chromatography; in addition, fluorescence was minimal and the lipids were used without further purification. Mixtures were prepared by first dissolving the two lipid components in chloroform, drying under nitrogen and then in vacuo. Water suspensions were obtained by sonicating (9) weighed amounts of lipid mixture and water for 1 hr, above the gel-liquid crystal transition temperature. Anhydrous samples were dried at 780C in vacuo for 24 hr over P205 and sealed in capillary tubes. The Raman spectrometer used in these experiments consists of a Coherent Radiation Model 53 argon ion laser, which is typically operated to give 1 W of 5145 i light incident on the sample. Raman light is collected at 90'C and analyzed in a Spex dual grating monochromator. Details of the spectrometer and counting system have been published elsewhere (10). The sample is sealed in a melting-point tube, which is placed perpendicular to the incident beam in a thermostated cell. Local heating by the incident light is 3 C at 58 C, determined by comparison of the melting temperature of a

2 ' Raman Spectra of Cholesterol-Lecithin Multilayers " : S " K32 = E C A _ V _ 1 FIG. 1. Laser Raman spectra of polycrystalline (a) hexadecane, (b) DL-dipalmitOyl lecithin monohydrate, (c) cholesterol, excited at Samples were thermostated at 80C. Resolution is about 4 cm-'. E identifies an argon emission line. myristic acid sample both in and out of the beam. This temperature rise should be fairly constant with temperature. All results reported here are corrected to give the temperature in the light beam. RESULTS Confornational effects on the Raman spectra of hydrocarbons The Raman spectra of long-chain hydrocarbons (for example hexadecane, Fig. la) contain two regions that are very susceptible to conformational changes in the paraffins. One region (40400 cm-') contains the low-frequency acoustical modes. The motion involved in this type of mode is the symmetrical stretch of the entire molecule. This region has been studied extensively by Schaufele (lla,b) for both crystalline and liquid hydrocarbons, and by Peticolas et al. (12) for crystalline polyethylene. Of interest in this work is the region cm-', in which the peaks have been assigned (13, 14) to the skeletal optical mode of the hydrocarbon chain with a motion such that alternate carbon atoms move in opposite directions along the chain axis. The Raman spectrum of C3(C2)n-2C3 solids with n > 8, shows two intense bands at 1064 and 1130 cm-' whose frequencies are independent of n within ±2 cm-i, and one very weak band at cm-' which is highly dependent on n. In hexadecane (Fig. la) this band ap-

3 1574 Chemistry: Lippert and PeticolasP I XA InI 3.0 I') c\j In a b S S c d 0....,....,I.... 1,,,) =o ~~0 8 v, cm-, FIG. 2. Raman spectra of the 1100 cm-' region of 20% (by weight) DL-dipalmitoyl lecithin sonicates in water at (a) 200C, (b) 30 C, (c) 40 C, (d) 50 C. pears at 1081 cm-', but it is shifted to 1100 cm-' inlhexadecanoic acid. These bands have all been assigned to vibrations of the all-trans configuration of the chain (13, 14).6 The Raman spectrum of liquid n-hydrocarbons in this region is characterized by a broad, intense band at 1090 cm-' in addition to much weaker bands at and cm-'. The decrease in intensity of the 1066 and 1130 cm-' bands can be attributed to a decrease in the amount of all-trans crystal structure while the broad band at 1089 cm-' is assigned to the appearance of structures containing several gauche rotations in the melted paraffin (15). These changes in the Raman spectrum around 1100 cm-' represent a sensitive probe to changes in the structure of paraffin chains. We have used this probe to study the effect of cholesterol on dipalmitoyl lecithin multilayers. Raman spectra of dipalmitoyl lecithin multilayers The Raman spectrum of DL-dipalmitoyl lecithin monohydrate is shown in Fig. lb. The most prominent features are those that are associated with the palmitate side chains and have nearly the same frequencies as the peaks in hexadecane (Fig. (Fig. la). The 1100 cm-' band corresponds exactly (unpublished observations) to an all-trans vibration in palmitic T, OC FIG. 3. Changes with temperature in the ratio of Raman peak heights of DL-dipalmitoyl lecithin at 1089 and 1128 cm-' for (a) 20% (by weight) sonicate in water of dipalmitoyl lecithin A, (b) 20% (by weight) sonicate in water of 1: -cholesteroldipalmitoyl lecithin 0, and (c) 10% (by weight) solution of dipalmitoyl lecithin in chloroform O. Similar curves are obtained by comparison of the peak heights at 1089 and 1066 cm-1. acid, which is slightly shifted from 1081 cm-' in hexadecane. For this reason we conclude that the palmitate side chains are in a trans configuration. Unfortunately, no bands appear in the sensitive low-frequency region corresponding to the 149 cm'l acoustical vibration in hexadecane. Fig. 2 shows the 1100 cm-' region of the Raman spectrum of a DL-dipalmitoyl lecithin sonicate (20% by weight in water) at several temperatures. An abrupt decrease in the intensity of the Raman bands at cm-' and cm'-- that are due to a vibration of the extended alltrans struct;ure is apparent as the temperature is changed from 30 to 40 C. At the same time a sharp increase in the intensity of the 1089 cm-' band assigned to random liquidlike configurations in observed. In Fig. 3 the curve marked with open triangles shows the change with temperature of the 1066 cm-' band, relative to the height of the 1089 cm-' band. This curve displays an abrupt change at 38-39oC. This change corresponds to the gel-liquid crystal phase transition that has previously been observed in the L1-isomer at C by microscope and differential scanning calorimetry (16). From our measurements we can conclude that this transition involves a change in the palmitate chains from an all-trans to a fluid configura-

4 PcRaman Spectra of Cholesterol-Lecithin Multilayers 1575 kc) r-_- I r- IKA 1.8 a T. OC FIG. 5. Change with temperature in the ratio of Raman peak heights of DL-dipalmitoyl lecithin at 1089 and 1128 cm-' with temperature for dipalmitoyl lecithin monohydrate *, anhydrous dipalmitoyl lecithin O, and anhydrous 1:1 cholesterol-dipalmitoyl lecithin A. I. I,I IC D C Vcm-, FIG. 4. Difference Raman spectra of the 1100 cm-' region of DL-dipalmitoyl lecithin in 20% (by weight) sonicates of 1:1 cholesterol-dipalmitoyl lecithin mixtures in water at (a) 20'C, (b) 300C, (c) 40'C, (d) 50'C. The Raman spectrum of cholesterol matched at 1673 cm-' is shown in (e) and is independent of temperature. tion. Since these curves are sharply sigmoidal, we can deduce that this transition is highly cooperative. The curve in Fig. 3 marked by open cricles shows the change in the same relative peak heights of dipalmitoyl lecithin in a 1:1 mol/mol cholesterol-lecithin sonicate, 20% by weight in water. ere, an extremely broad transition is observed. The rather weak cholesterol spectrum (Fig. lc) was subtracted from the cholesterol-lecithin spectrum using the cholesterol band at 1673 cm-' as a measure of its intensity. The resultant dipalmitoyl lecithin spectra are shown in Fig. 4 at various temperatures in the 1100 cm'- region. Fig. 3 also contains a curve showing the temperature dependence of the relative peak heights of 10% (by weight) dipalmitoyl lecithin in chloroform (open squares). In this good solvent, the paraffin side chains are no doubt in a very fluid condition and show no transition of any kind. Fig. 5 shows the temperature changes in relative peak heights for (a) dipalmitoyl lecithin monohydrate, (b) anhydrous dipalmitoyl lecithin, and (c) anhydrous, 50 mol% cholesterol-dipalmitoyl lecithin. Pure lecithin monohydrate exhibits a sharp crystal-liquid crystal transition at C (ref. 16 gives 680C for L-diplamitoyl lecithin). DISCUSSION Fig. 3 shows dramatically that the effect of cholesterol on dipalmitoyl lecithin multilayers is to greatly broaden the gel-liquid crystal transition. Thus, multilayers containing cholesterol will show properties more fluid than the pure lipid when measured below the transition temperature, while measurements above it will show more rigidity. This broadening has been reported at lower cholesterol concentrations by differential scanning colorimetry (7). owever, at 50 mol% cholesterol, the transition is so broad that it can no longer be observed by that method. This leaves open the possibility that dipalmitoyl lecithin multilayers containing equimolar amounts of dipalmitoyl lecithin and cholesterol are in a fluid condition even at room temperature (7). Fig. 3 shows, however, that the transition from largely trans configuration to a principally random configuration of palmitate side chains does occur but is spread over a 70-80QC range in temperature in the presence of cholesterol. This suggests that the mode of action of cholesterol is to change the gel-liquid crystal transition from a cooperative to a noncooperative event by decreasing the interactions between adjacent paraffin side chains in the multilayer. It is somewhat hazardous to compare these model experiments with biological processes, but the role of cholesterol in biological membranes may be to plasticize the membranes so that they are not so susceptible to temperature variations. In addition, the role of chlesterol in rigidizing membranes at temperatures above the T, of their paraffin side chains cannot be discounted. We would now ask the question, "What characteristics of the cholesterol-lecithin-water system are required to obtain a cooperative gel-liquid crystal transition?" The answer seems to be that a unique ordering of lecithin in the presence of water is required and that the absence of water or the addition

5 1576 Chemistry: Lippert and Peticolas of cholesterol disrupts this order. This is suggested by the experiments illustrated in Fig. 5. Dipalmitoyl lecithin monohydrate exhibits a sharp, cooperative crystal-liquid crystal transition at 680C. This transition is a melting of the palmitate side chains and is equivalent to the gel-liquid crystal transition in sonicated multilayers. Removal of the last traces of water, however, causes the loss of the cooperative transition. As might be expected, anhydrous 1:1 dipalmitoyl lecithin-cholesterol also shows a noncooperative transition, although shifted 30'C lower in temperature. In fact, the transition curves of anhydrous dipalmitoyl lecithin, 1:1 lecithincholesterol, and sonicated 1:1 lecithin-cholesterol are superimposable, but shifted relative to one another along the temperature axis (compare Figs. 3 and 5). X-ray studies (16) of anhydrous and monohydrated dipalmitoyl lecithin strongly suggest that removal of all water causes disordering of the lecithin crystals. The similarity between the effects of addition of cholesterol and removal of water on the shape of the transition curves (Figs. 3-5) indicates that cholesterol may somehow disorder lecithin multilayers, or at least may strongly reduce the interactions between adjacent hydrocarbon chains. These experiments show the value of Raman spectroscopy in the study of lipid configurations in aqueous solutions. They are possible only because the Raman spectrum of water is broad and quite weak in the region of interest. Infrared studies in water solutions in this frequency range are impossible because of the intense infrared absorption of water. In addition, Raman studies of membranes provide a considerable advantage over such techniques as electron microscopy and electron spin resonance, since Raman studies do not require the substitution of metal or free-radical probes, which may greatly perturb the environment that is studied. The principal drawback of Raman spectroscopy applied to biological systems is the problem of intense luminescence that is often encountered. Nevertheless, we hope to apply this technique to the study of lipid motion in selected biological systems. NOTE ADDED IN PROOF Some of the experiments illustrated in Fig. 3 have been reproduced with L-dipalmitoyl lecithin. We therefore believe there is little difference in the nature of the cooperative transition observed in either the DL or L isomers. We wish to thank 0.. Griffith and P. Jost for most helpful discussions. The authors gratefully acknowledge support from the National Science Foundation Grant No. GB W. L. L. is a National Institutes of ealth Postdoctoral Fellow. 1. Chapman, D., and S. A. Penkett, Nature, 211, 1304 (1966). 2. Waggoner, A. S., T. J. Kingzett, S. Rottschaeffer, 0.. Griffith, and A. D. Keith, Chem. Phys. Lipids, 3, 245 (1969). 3. sia, J. C.,. Schneider, and I. C. P. Smith, Chem. Phys. Lipids, 4, 238 (1970). 4. Finkelstein, A., and A. Cass, Nature, 216, 717 (1967). 5. Shah, D. O., and J.. Schulman, J. Lipid Res., 8, 215 (1967). 6. Joos, P., Chem. Phys. Lipids, 4, 162 (1970). 7. Ladbrooke, B. D., R. M. Williams, and D. Chapman, Biochim. Biophys. Acta, 150, 333 (1968). 8. De Gier, J., J. G. Mandersloot, and L. L. M. van Deenen, Biochim. Biophys. Acta, 173, 143 (1969). 9. Chapman, D., and D. J. Fluck, J. Cell Biol., 30, 1 (1966). 10. Fanconi, B., B. Tomlinson, L. A. Nafie, E. W. Small and W. L. Peticolas, J. Chem. Phys., 51, 3993 (1969). 11. (a) Schaufele, R. G., and T. Shimanouchi, J. Chem. Phys., 47, 3605 (1967); (b) Schaufele, R. G., J. Chem. Phys., 49, 4168 (1968). 12. Peticolas, W. L., G. W. ibler, J. L. Lippert, A. Peterlin, and. Olf, Appl. Phys. Lett., 18, 87 (1971). 13. Tasumi, M., and T. Shimanouchi, J. Mol. Spectrosc., 9, 261 (1962). 14. Snyder, R. G., and J.. Schachtscheider, Spectrochim. Acta, 19, 85 (1963). 15. Snyder, R. G., J. Chem. Phys., 47, 1316 (1967). 16. Chapman, D., R. M. Williams, and B. D. Ladbrooke, Chem. Phys. Lipids, 1, 445 (1967).