Langmuir 2004, 20, 1043-1046
1043
Lanosterol and Cholesterol-Induced Variations in Bilayer Elasticity Probed by 2H NMR Relaxation Gary V. Martinez,† Emily M. Dykstra,† Silvia Lope-Piedrafita,‡ and Michael F. Brown*,†,‡ Departments of Chemistry and Physics, University of Arizona, Tucson, Arizona 85721 Received October 31, 2003 The influences of lanosterol on lipid bilayers have been compared to those of cholesterol by combining deuterium (2H) NMR spin relaxation studies with segmental order parameter measurements. For bilayers of 1,2-diperdeuteriomyristoyl-sn-glycero-3-phosphocholine (DMPC-d54), the results are consistent with a square-law dependence of the 2H Zeeman relaxation rates (R1Z) on the corresponding order parameters (SCD). This behavior is indicative of relatively slow order fluctuations, for example, due to quasi-elastic bilayer disturbances. Significant differences are found in the influences of lanosterol versus cholesterol on the microscopic NMR observables; although lanosterol produces smaller order parameters than cholesterol, it leads to larger relaxation rates. By correlating the NMR relaxation behavior with the order parameters, the results are explained by a progressive reduction of the bilayer elasticity, which parallels the biosynthetic pathway from lanosterol to cholesterol.
The influences of sterols such as cholesterol on the biophysical properties of cellular membranes have drawn substantial recent interest in chemical biology1-3 as a result of their proposed roles in structures known as caveolae and rafts.4 Cholesterol-rich, liquid-ordered phases5 or condensed complexes6 including specific membrane proteins4 have been implicated in cell signaling, membrane trafficking, and disease transmission including HIV infection.7 The properties of sterols are of fundamental significance1-3 in conjunction with protein-mediated functions of biomembranes.8 Because 20-30% of the proteins of the human genome are membrane-associated,9 bilayers containing sterols are of major importance with regard to receptors and drug design. Sterol interactions with lipid bilayers have been studied using various biophysical techniques, including 2H NMR,3,10 quasi-elastic neutron scattering,2 micropipet deformation,11 and fluctuation analysis of giant vesicles.12 We report herein that the * To whom correspondence should be addressed. E-mail:
[email protected]. Fax: 520-621-8407. † Department of Chemistry. ‡ Department of Physics. (1) Bloom, M.; Evans, E.; Mouritsen, O. G. Q. Rev. Biophys. 1991, 24, 293-397. (2) Endress, E.; Heller, H.; Casalta, H.; Brown, M. F.; Bayerl, T. M. Biochemistry 2002, 41, 13078-13086. (3) Miao, L.; Nielsen, M.; Thewalt, J.; Ipsen, J. H.; Bloom, M.; Zuckermann, M. J.; Mouritsen, O. G. Biophys. J. 2002, 82, 1429-1444. (4) (a) Brown, D. A.; London, E. J. Biol. Chem. 2000, 275, 1722117224. (b) Simons, K.; Ikonen, E. Science 2000, 290, 1721-1726. (c) Anderson, R. G. W.; Jacobson, K. Science 2002, 296, 1821-1825. (5) Ipsen, J. H.; Karlstro¨m, G.; Mouritsen, O. G.; Wennerstro¨m, H. W.; Zuckermann, M. J. Biochim. Biophys. Acta 1987, 905, 162-172. (6) Radhakrishnan, A.; Anderson, T. G.; McConnell, H. M. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 12422-12427. (7) Ono, A.; Freed, E. O. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 1392513930. (8) (a) Mitchell, D. C.; Straume, M.; Miller, J. L.; Litman, B. J. Biochemistry 1990, 29, 9143-9149. (b) Brown, M. F. Chem. Phys. Lipids 1994, 73, 159-180. (9) Wallin, E.; von Heijne, G. Protein Sci. 1998, 7, 1029-1038. (10) (a) Seelig, J. Q. Rev. Biophys. 1977, 10, 353-418. (b) Oldfield, E.; Meadows, M.; Rice, D.; Jacobs, R. Biochemistry 1978, 17, 27272740. (c) Vist, M. R.; Davis, J. H. Biochemistry 1990, 29, 451-464. (d) Stohrer, J.; Gro¨bner, G.; Reimer, D.; Weisz, K.; Mayer, C.; Kothe, G. J. Chem. Phys. 1991, 95, 672-678. (e) Urbina, J. A.; Pekerar, S.; Le, H. B.; Patterson, J.; Montez, B.; Oldfield, E. Biochim. Biophys. Acta 1995, 1238, 163-176. (11) Rawicz, W.; Olbrich, K. C.; McIntosh, T.; Needham, D.; Evans, E. Biophys. J. 2000, 79, 328-339.
Chart 1
influences of sterols on solid-state 2H NMR order parameters (SCD) and relaxation rates (R1Z) of bilayers involve quasi-elastic thermal deformations of the nanostructure, corresponding to the bulk membrane elasticity.12 The NMR relaxation data suggest a progressive increase in bilayer rigidity on going from lanosterol to cholesterol, which parallels the demethylation occurring in the metabolic pathway of sterol biogenesis (Chart 1).1,13 Aqueous binary mixtures of lanosterol (1) or cholesterol (2), together with a representative phospholipid, 1,2-diperdeuteriomyristoyl-sn-glycero-3-phosphocholine (DMPCd54; 3), were investigated. Comparison of 1 with 2 shows that the major difference involves three additional methyl groups, leading to a roughening of the R face of the sterol.13 Solid-state 2H NMR spectra of random multilamellar dispersions of 3 together with either 1 or 2 were acquired and numerically inverted14 as shown in Figure 1 (see (12) Me´le´ard, P.; Gerbeaud, C.; Pott, T.; Fernandez-Puente, L.; Bivas, I.; Mitov, M. D.; Dufourcq, J.; Bothorel, P. Biophys. J. 1997, 72, 26162629. (13) (a) Yeagle, P. L.; Martin, R. B.; Lala, A. K.; Lin, H.-K.; Bloch, K. Proc. Natl. Acad. Sci. U.S.A. 1977, 74, 4924-4926. (b) Bloch, K. Crit. Rev. Biochem. 1983, 14, 47-92.
10.1021/la036063n CCC: $27.50 © 2004 American Chemical Society Published on Web 01/17/2004
1044
Langmuir, Vol. 20, No. 4, 2004
Figure 1. Representative solid-state 2H NMR spectra for (a) DMPC-d54 in the LR phase, (b) lanosterol/DMPC-d54 (1:1), and (c) cholesterol/DMPC-d54 (1:1) in the liquid-ordered phase. The samples contained 50 wt % 1H2O (20 mM Tris, pH 7.3) at 44 °C. Powder-type 2H NMR spectra (light) at 46.1 MHz (7.06 T) were numerically inverted (de-Paked) to obtain subspectra (dark) corresponding to the θ ) 0° bilayer orientation.
Supporting Information). Note that a large increase in the quadrupolar splittings occurs upon incorporation of both 1 and 2 into the DMPC-d54 (3) bilayer.10 The 2 orientational order parameters S(i) CD of the various C- H labeled segments are related to both the prominent peaks (edges or “horns”) of the experimental powder-type 2H NMR spectra (θ ) 90°) and to the de-Paked 2H NMR spectra (θ ) 0°), using the relation (i) 3 |∆ν(i) Q | ) /2χQ|SCD||P2(cos θ)|
(1)
Here, ∆ν(i) Q is the ith residual quadrupolar splitting, χQ is the static quadrupolar coupling constant (170 kHz), and P2(cos θ) ) 1/2(3 cos2 θ - 1). The segmental order parameters characterize the dynamic structure of the bilayer1,15 and are defined as 1 2 (i) S(i) CD ) /2〈3 cos β - 1〉
(2)
in which β(i) is the angle between the ith C-2H bond and the normal to the bilayer surface (director) and the angular brackets pertain to a time or ensemble average. The frequencies of each of the lines were obtained by nonlinear regression fitting of the de-Paked 2H NMR spectra to a set of Gaussian basis functions, using the curve fitting routine in the program Igor Pro (Wavemetrics; Lake Oswego, OR). Typically, Gaussians were found to give a better fit to the data than Lorentzians. Representative results of the curve fitting procedure are shown in Figure 2 for bilayers of DMPC-d54 in the liquid-crystalline (LR) state, part a, for a dispersion of lanosterol/DMPC-d54 (14) (a) McCabe, M. A.; Wassall, S. R. Solid State Nucl. Magn. Reson. 1997, 10, 53-61. (b) Sternin, E.; Scha¨fer, H.; Polozov, I. V.; Gawrisch, K. J. Magn. Reson. 2001, 149, 110-113. (15) Petrache, H. I.; Dodd, S. W.; Brown, M. F. Biophys. J. 2000, 79, 3172-3192.
Figure 2. Fitting of 2H NMR spectra (de-Paked) to Gaussian basis functions: (a) DMPC-d54 in the liquid-crystalline (LR) phase at 44 °C, (b) lanosterol/DMPC-d54 (1:1), and (c) cholesterol/ DMPC-d54 (1:1) in the liquid-ordered phase at 44 °C. Data were acquired at 76.8 MHz (11.8 T). The area under each of the peaks obtained from the Gaussian fitting procedure corresponds to the number of deuterons. Note the large increase in the residual quadrupolar couplings as a result of the presence of lanosterol or cholesterol in the liquid-ordered phase.
(1:1), part b, and for a dispersion of cholesterol/DMPC-d54 (1:1) in the liquid-ordered phase, part c. The location, intensity, and width of the Gaussian curve for each peak were the adjustable parameters optimized in the curve fitting. By fitting the 2H NMR spectra as shown in Figure 2, it was possible to estimate the number of deuterium atoms that contribute to each of the peaks. Assignment of the quadrupolar splittings to individual acyl chain segments was based on previous data for specifically deuterated phospholipids10 and for specifically deuterated DMPC containing 30 mol % cholesterol.10 Where necessary, the data were interpolated or extrapolated to obtain a highly consistent set of assignments. The 2H NMR quadrupolar frequencies for the θ ) 0° orientation were used to calculate the corresponding S(i) CD order parameters (see previous text). In addition, the 2H nuclear spin relaxation rates indicate the types of lipid motions that average the coupling tensor, which for 2H NMR is due to the electric field gradient of the C-2H bond.16 Such motions can be due to local segmental isomerizations of the flexible lipids, effective molecular rotations within the potential of mean force, or collective excitations of the bilayer.16 The spin-lattice relaxation rate is given by
R1Z ) 3/4π2χQ2[J1(ω0) + 4J2(2ω0)]
(3)
where χQ is the coupling constant and ω0 is the nuclear Larmor frequency. The symbols Jm(ω) ) ∫Gm(t) exp(-iωt) dt denote the (irreducible) spectral densities of motion, in which Gm(t) are the temporal autocorrelation functions (rank 2) of the perturbing Hamiltonian (m ) 1, 2). Figure 3, part a, shows an experimental set of partially relaxed 2H NMR spectra for a random dispersion of DMPCd54 in the liquid-crystalline (LR) phase. The sharp spectral edges represent the weak singularities due to the θ ) 90° (16) (a) Brown, M. F. J. Chem. Phys. 1982, 77, 1576-1599. (b) Brown, M. F.; Thurmond, R. L.; Dodd, S. W.; Otten, D.; Beyer, K. J. Am. Chem. Soc. 2002, 124, 8471-8484.
Langmuir, Vol. 20, No. 4, 2004 1045
Figure 3. Representative partially relaxed 2H NMR spectra of randomly oriented DMPC-d54 multilamellar dispersion in the liquid-crystalline (LR) phase at T ) 44 °C: (a) experimental 2 H NMR spectra and (b) deconvolved de-Paked spectra (θ ) 0°). The sample contained 20 mM Tris buffer at pH 7.3 (50 wt % H2O). Data were acquired at 76.8 MHz (11.8 T) using a phasecycled, inversion recovery pulse sequence, with the variable delay ranging from 5 ms to 3 s.
orientation of the bilayer normal (director axis) relative to the main magnetic field, and the weak shoulders are due to θ ) 0°. The 2H NMR spectra indicate a single bilayer environment on the time scale of ≈10-6 s; no evidence for long-lived cholesterol-containing complexes or domains is found. In Figure 3, part b, the partially relaxed 2H NMR spectra are deconvolved to yield de-Paked spectra corresponding to the θ ) 0° orientation. The increased resolution allows a more accurate determination of the R1Z rates. Following the inverting π pulse, the z magnetization appears negative and recovers to the equilibrium positive signal. The rate of z magnetization recovery is different for the individual resolved components and increases with the quadrupolar splitting. This behavior is consistent with relaxation due to order fluctuations that involve a preaveraged or residual coupling tensor rather than local segmental motions. The distribution of residual quadrupolar couplings manifests the orientational order parameters of the various C2H2 groups,10 which are plotted against the acyl chain position (i) in Figure 4, parts a and b. As can be seen, a progressive reduction occurs along the chains, where the smallest values are due to the acyl terminal C2H3 groups. The profiles of the R(i) 1Z rates as a function of acyl position (i) are markedly reminiscent of the order profiles as shown in Figure 4, parts c and d. Note the opposite effects of sterols on the order and relaxation profiles in Figure 4: the order parameters increase whereas the relaxation rates decrease. This paradox is explained by order fluctuations16 in which the residual ordering leftover from local segmental isomerizations is modulated by slower, quasi-elastic bilayer disturbances. Given a composite membrane deformation model,16 the relaxation involves a broad spectrum of three-dimensional
Figure 4. Profiles of experimental observables from 2H NMR for mixtures of lanosterol/DMPC-d54 and cholesterol/DMPCd54 at T ) 44 °C. Order parameters |S(i) CD| for the (a) sn-1 chain (filled symbols) and (b) sn-2 chain (open symbols) and spinlattice relaxation rates R(i) 1Z for the (c) sn-1 chain (filled symbols) and (d) sn-2 chain (open symbols) are shown: (9, 0) DMPC-d54 in the LR phase, ([, ]), lanosterol/DMPC-d54 (1:1), and (b, O), cholesterol/DMPC-d54 (1:1) in the liquid-ordered phase. Data were acquired at 46.1 MHz (7.06 T).
Figure 5. Dependence of relaxation rates R(i) 1Z on squared 2 order parameters |S(i) | for resolved H NMR splittings of CD DMPC-d54, showing influences of lanosterol and cholesterol at 55 °C. Data at 76.8 MHz (11.8 T) are shown: DMPC-d54 alone (9), lanosterol/DMPC-d54 (1:1; [), and cholesterol/DMPC-d54 (1:1; b). The presence of sterols leads to a large decrease in the square-law slopes, consistent with a progressive reduction in bilayer elasticity on going from lanosterol to cholesterol.
collective excitations, with effective rotations of the lipids about the bilayer normal. The picture just described gives a useful framework to explore the properties of membranes at mesoscale lengths ≈ the bilayer thickness, corresponding to relatively high frequencies (mega-Hertz range) and how they are influenced by sterols. As shown in Figure 5, for the pure DMPCd54 bilayer the functional dependence of the R(i) 1Z and order obeys a simple square law to a fairly good profiles S(i) CD degree of approximation. Assuming a composite membrane deformation model, the square law manifests the bending rigidity of the membrane on the mesoscopic length scale.16 The slopes are inversely related to the bilayer bending modulus so that an increase in the bilayer stiffness gives a smaller value and vice versa. Addition of 1 (1:1 molar ratio) yields a marked reduction of the square-law slope, and 2 gives the greatest effect of all. The reduction is interpreted as an increase in the bending energy,
1046
Langmuir, Vol. 20, No. 4, 2004
corresponding to measurements of the bulk membrane elasticity.11,12,17 Thus, the bilayer stiffness is less for lanosterol versus its metabolic product cholesterol. The more molecularly smooth van der Waals surface of the R face of cholesterol13 enables a large increase in bilayer rigidity and stabilizes the liquid-ordered phase to an even greater degree than lanosterol. Interestingly, for the systems studied the ordinate intercepts are comparable, suggesting that the local segmental motions of the flexible lipids are similar, in contrast to the longer-wavelength quasi-elastic modes. In summary, this work provides the first experimental NMR evidence that bilayers containing lanosterol are not as rigid as those containing cholesterol. The variations in bilayer stiffness are consistent with the presence of (17) Endress, E.; Bayerl, S.; Prechtel, K.; Maier, C.; Merkel, R.; Bayerl, T. M. Langmuir 2002, 18, 3293-3299.
stronger cohesive interactions between cholesterol and lipids than those for lanosterol.3 Last, the progressive increase in the bilayer rigidity on going from lanosterol to cholesterol parallels the metabolic pathway of sterol biogenesis13 and may be related to the optimization of their biophysical properties.1 Acknowledgment. We thank T. Bayerl, M. Bloom, E. Endress, and M. Zuckermann for stimulating discussions and the NIH for financial support. G.V.M. is the recipient of an NIH postdoctoral fellowship. Supporting Information Available: Experimental 2H NMR methods and data analysis and reduction. This material is available free of charge via the Internet at http://pubs.acs.org. LA036063N
