Desmosterol may replace cholesterol in lipid membranes

Abstract

Recently, knockout mice entirely lacking cholesterol have been described as showing only a mild phenotype. For these animals, synthesis of cholesterol was interrupted at the level of its immediate precursor, desmosterol. Since cholesterol is a major and essential constituent of mammalian cellular membranes, we asked whether cholesterol with its specific impact on membrane properties might be replaced by desmosterol. By employing various approaches of NMR, fluorescence, and EPR spectroscopy, we found that the properties of phospholipid membranes like lipid packing in the presence of cholesterol or desmosterol are very similar. However, for lanosterol, a more distant precursor of cholesterol synthesis, we found significant differences in comparison with cholesterol and desmosterol. Our results show that, from the point of view of membrane biophysics, cholesterol and desmosterol behave identically and, therefore, replacement of cholesterol by desmosterol may not impact organism homeostasis.

FIGURE 1

Chemical structure of cholesterol (A), desmosterol (B), and lanosterol (C).

FIGURE 2

Smoothed ²H-NMR order parameter profiles of POPC-d₃₁ membranes in the presence of varying sterol concentrations at a temperature of 30°C. Order parameters are shown for desmosterol (solid symbols in A) and for lanosterol (solid symbols in B). In comparison, the values of pure POPC-d₃₁ (×) and cholesterol (open symbols) are shown. The steroid concentrations were 5 mol % (squares), 10 mol % (circles), 15 mol % (triangles), and 20 mol % (diamonds). In C, the average order parameter of the palmitoyl acyl chain of POPC-d₃₁ is given at varying concentration of (□) cholesterol, (♦) desmosterol, and (•) lanosterol at 30°C. The typical error of the order parameters is on the order of ±0.005 and thus covered by the symbol size in the plot.

FIGURE 3

EPR spectra of spin-labeled phosphatidylcholine analogs in large unilamellar vesicles. Spectra of C5-SL-PC (A) and C16-SL-PC (C) were recorded in LUV consisting of POPC (a), POPC/cholesterol (70:30) (b), POPC/desmosterol (70:30) (c), and POPC/lanosterol (70:30) (d) at 30°C as described in Materials and Methods. The arrows show the outer hyperfine splitting (A), which is increased in the presence of steroids. Note that for C16-SL-PC, an additional component (see high field peak of the spectra) may indicate that a small part of this lipid is organized in a different ordered domain within the membrane. From the spectra of C5-SL-PC and C16-SL-PC, an order parameter (S, panel B) and a correlation time of rotation (τ, panel D), respectively, was calculated. Data are the average ± SE of estimate of 4–5 independent measurements.

FIGURE 4

Fluorescence spectra of Laurdan in large unilamellar vesicles consisting of POPC (a), POPC/cholesterol (70:30) (b), POPC/desmosterol (70:30) (c), and POPC/lanosterol (70:30) (d) were recorded at 30°C as described in Materials and Methods (A). Fluorescence intensities were normalized to intensities measured after disruption of vesicles by Triton X-100 (0.5%). From these spectra, the generalized polarization of emission intensities GP = (I_B − I_R)/(I_B + I_R) (I_B and I_R: intensities at 437 nm and 482 nm) was calculated (average ± SE of estimate of at least six independent measurements) (B). Fluorescence spectra of Laurdan were recorded at 40°C and the GP values calculated for vesicles consisting of SPM (e), SPM/cholesterol (70:30) (f), SPM/desmosterol (70:30) (g), and SPM/lanosterol (70:30) (h) (C). (C) Data are the average ± SE of estimate of at least five independent measurements.