High-field, high-resolution proton "magic-angle" sample-spinning nuclear magnetic resonance spectroscopic studies of gel and liquid crystalline lipid bilayers and the effects of cholesterol

Abstract

We have obtained proton ('H) magic-angle sample-spinning (MASS) nuclear magnetic resonance (NMR) spectra of a variety of smectic liquid crystalline phases, including sodium decanoate (30.1 wt %)-decanol(38.9 wt %)-water, potassium oleate (72 wt %)-water, and 1,2-dimyristoyl-sn-glycero-3-phosphocholine (lecithin)(50 wt %)-water, in addition to investigating the effects of temperature and cholesterol (CHOL) addition on the lecithin spectrum. Our results indicate that even relatively slow (-3 kHz) MASS causes averaging of the dipolar interactions causing line broadening in the static NMR spectra, at least for the non-CHOL species. All of the major proton-containing groups are well resolved, the resolution being at least as good as obtained in previous studies of oriented samples or sonicated systems. The methylene chain protons in each liquid crystalline lipid bilayer system give rise to intense, sharp, spinning sidebands (SSBs) due to the special form of the dipolar Hamiltonian. The methyl groups of the lipids, and the trimethylammonium group in lecithin, do not yield intense SSB patterns. Addition of CHOL causes attenuation of the center-band methylene peak of the lecithin, and a corresponding increase in SSB intensity. All or nearly all of the non-CHOL protons present in the samples appear to contribute to the high-resolution spectra, within our experimental error of - 10-20%. Use of a chain-deuterated lecithin allows peaks arising from the side chain of CHOL to be observed. In the gel phase of lecithin, only the trimethylammonium peak is apparent. The high-resolution afforded by MASS of the liquid crystalline phases permits rapid determination of the spin-lattice relaxation times (TI) of all resolved resonances. In addition, the observation of numerous chemically shifted peaks permits the use of two-dimensional (2-D) NMR techniques, which can give information on the spatial proximity of the various groups in the bilayer. Taken together, our results indicate a very promising future for high-field 'H MASS NMR studies of other lipid and membrane systems because of the extremely high sensitivity of the IH nucleus and the unique ability to obtain chemical shift, TI, and 2-D information from a single sample, without recourse to isotopic labeling, macroscopic sample orientation, or ultrasonic irradiation. Nuclear magnetic resonance (NMR) spectroscopy has been used to investigate the structure of model and biological mem- branes for over 20 years, and studies of molecular motion in simple lipids and hydrocarbons can be traced back even further.'S2 The earliest studies concentrated on the 'H nucleus, which because of its high sensitivity and abundance was expected to be a useful probe of membrane structure. Early studies by Cerb~jn~,~ identified mobile lipid components in Nocardia asteroides; then Chapman et al. began an extensive series of studies of model- and biological membrane systems.+l2 The early 'H NMR experiments utilized wide-line methods, and assumed that 'H-IH dipolar interactions dominated the observed line widths, since very broad lines were, in general, obtained. These line widths could be reduced by sonicating the (liquid crystalline) lipid bilayers, or membranes, reducing their particle size and permitting faster vesicle tumbling. Shortly after the first papers by Chapman et al., numerous other groups published similar studies on related sy~tems.l~-*~ Three main questions arose from the early studies. These centered around the nature of the line-broadening mechanisms in the liquid crystalline phases, the effects of sonication (does it cause line narrowing due to a change in lipid bilayer structure, or because of increased rates of particle rotation?), and whether spin diffusion occurs in the liquid crystalline phases. These questions have (at least in part) been answered over the past 15 years. Thus, Chan et a1.,26 Tidd~,~' and Oldfield et aLZ8 showed that line broadening in various liquid crystalline phases (at least up to -90 MHz) was purely dipolar in origin, since the line widths (or effective T2 values, T2*) were field independent. This view was supported by the results of 'H 'magic-angle sample-spinning (MASS) NMR,29 by multiple-pulse line narrowing,30 and by magic-angle alignment of oriented sample^.^'-^^ The effects of sonication have been widely studied, and the early view, that sonication causes line narrowing only due to increased particle tumbling, has received considerable s~pport,~~-~~ although it is