Ultrastructure and lipid composition of detergent-resistant membranes derived from mammalian sperm and two types of epithelial cells

Abstract

Lipid rafts are micro-domains of ordered lipids (Lo phase) in biological membranes. The Lo phase of cellular membranes can be isolated from disordered lipids (Ld phase) after treatment with 1 % Triton X-100 at 4 °C in which the Lo phase forms the detergent-resistant membrane (DRM) fraction. The lipid composition of DRM derived from Madin-Darby canine kidney (MDCK) cells, McArdle cells and porcine sperm is compared with that of the whole cell. Remarkably, the unsaturation and chain length degree of aliphatic chains attached to phospholipids is virtually the same between DRM and whole cells. Cholesterol and sphingomyelin were enriched in DRMs but to a cell-specific molar ratio. Sulfatides (sphingolipids from MDCK cells) were enriched in the DRM while a seminolipid (an alkylacylglycerolipid from sperm) was depleted from the DRM. Treatment with <5 mM methyl-ß-cyclodextrin (MBCD) caused cholesterol removal from the DRM without affecting the composition and amount of the phospholipid while higher levels disrupted the DRM. The substantial amount of (poly)unsaturated phospholipids in DRMs as well as a low stoichiometric amount of cholesterol suggest that lipid rafts in biological membranes are more fluid and dynamic than previously anticipated. Using negative staining, ultrastructural features of DRM were monitored and in all three cell types the DRMs appeared as multi-lamellar vesicular structures with a similar morphology. The detergent resistance is a result of protein-cholesterol and sphingolipid interactions allowing a relatively passive attraction of phospholipids to maintain the Lo phase. For this special issue, the relevance of our findings is discussed in a sperm physiological context.

Keywords: Cholesterol; Detergent-resistant membranes; Epithelial cell; Glycolipids; Lipid rafts; Phospholipids; Sperm; Ultrastructure.

Fig. 1

The effect of 1 % Triton X-100 at 4 °C on porcine sperm and a scheme for the separation method used to obtain a soluble and insoluble membrane fraction from sperm, McArdle and MDCK cells. a Regional differences of solubilization of sperm membranes after subjection to 1 % Triton X-100 at 4 °C. An ultrathin section of a porcine sperm head visualized with transmission electron microscopy (samples were processed for transmission electron microscopy according to the method of Tsai et al. 2010). Scale bar 2 μm. The red area (at the equatorial surface area) is shown magnified below and shows that membrane structures are lost due to Triton  X-100 solubilization. The blue area (of the apical ridge surface area) is shown magnified on the right (rotated to the right by 90°) and insoluble membrane micro-domains are indicated as DRM. b A schematic representation for separating the DRM from the soluble membrane fraction and the cellular remnants

Fig. 2

Partitioning of glycolipids and caveolin-1 in the sucrose gradient of 1 % Triton X-100 at 4 °C treated MDCK cells and sperm. The sucrose gradient of MDCK cells and sperm (cf. Fig. 1) was divided into 13 fractions of 1 ml. Proteins of fractions 1–13 were solubilized and transferred to a PVDF membrane (dot blot). Specific antibody binding was detected with enhanced chemifluorescence. For presentation purposes, dots of fractions 9–13 were aligned aside the spots of 1–8; the dots were originally spotted in multiple rows of 8 dots on one PVDF membrane and developed in the same fashion. Lipids from the 1–13 fractions were extracted, from which the glycolipids were purified and spotted on HPTLC plates, which was after development and charred with orcinol to allow purple staining of glycolipids (for method, see Gadella et al. 1993). a Dotblot and HPTLC for MDCK cells and b for boar sperm cells. The amount of sulfatides (SGalCer for structure: c for MDCK and seminolipid; SGalAAG for structure: d of fraction 13 versus fraction 5–9) was quantified according to the coloric method of Kean (1968) as modified by Radin (1984). Mean values ± SD are provided (n = 5)

Fig. 3

Ultrastructure of DRM derived from MDCK cells, McArdle cells and boar sperm. After subjecting the cells to 1 % Triton  X-100 at 4 °C, the DRM (fraction 5) was isolated as indicated in Fig. 1. A small drop of DRM was spotted on a grid that was processed for negative staining (Vennema et al. 1996) and inspected with transmission electron microscopy

Fig. 4

Cell-type specific compositions of lipid molecular species. HPLC chromatograms showing the phospholipid composition as monitored by light scattering detection: a–c the separation of sterols and molecular species of PC and SM of MDCK cells (a), McArdle cells (b) and sperm cells (c). d–f The molecular PE species of MDCK cells (d), McArdle cells (e) and sperm cells (f) after preceding isolation of PE by normal phase chromatography. Note that the cell types have a completely different molecular species composition of both PC, SM and PE and that the amount of cholesterol differs as well. Peak numbers refer to the identification of the species in Tables 1 and 2. Arrows indicate cholesterol

Fig. 5

MBCD decreases the amount of lipid in the DRM fraction. HPLC chromatograms showing the separation of molecular PC and SM species and cholesterol of DRM fractions derived from MDCK cells incubated with 0 mM (top) and 10 mM MBCD (bottom). Note that the PC and SM species composition does not change after MBCD treatment and that cholesterol is specifically extracted from the DRM fraction. Both traces are plotted on the same scale and represent DRMs from an identical amount of cells. The left inset shows the decrease in lipids after MBCD treatment and the right inset shows the MBCD-mediated, dose-dependent depletion of cholesterol from the DRM fraction. Similar results with porcine sperm have been published previously (van Gestel et al. 2005a)

Fig. 6

Proposed model for differences between a lipid ordered versus disordered bilayer. The cholesterol-induced tighter ordering of phospholipid head groups is in our model supposed to cause the detergence resistance rather than the hydrophobic fluidness of hydrogen carbon chains attached to the phospholipids. The wider distance of phospholipids in the L_d phase in our model allows intercalation of the detergent and this is the solubilization of L_d ordered lipids. As is measured in this study, under a threshold concentration of MBCD, lowered cholesterol levels from DRM do not disrupt the DRM and we expect that despite cholesterol removal the phospholipid head groups remain tightly packed. Above a threshold concentration of MBCD, the whole DRM becomes disrupted as phospholipids become disorganized in phospholipid head group packing. For MDCK, this level was reached at >10 mM MBCD while sperm DRMs became disrupted at 2–5 mM (van Gestel et al. 2005b). This difference may be explained by the lower abundance of cholesterol in the DRM of sperm when compared to MDCK, presumably making them more sensitive for cholesterol depletion. In this simplified model, we have not included information on the DRM accumulation of sphingolipids - including the ceramide based sulfatides of MDCK cells and the exclusion of the alkylacyl based seminolipid of sperm; two glycolipids with the same head group (see Fig. 2). This phenomenon shows that the cholesterol induced tighter packing of phospholipid head groups does have an impact on attraction and repulsion of glycolipids. Likewise, the specific attraction of specific membrane proteins and gangliosides is not included in this model