Globoside with spin-labelled fatty acid: bilayer lateral distribution and immune recognition.

We have critically addressed the question of lateral distribution of glycolipids in bilayer membranes, and the effect of glycolipid fatty acid chain length upon such distribution. For this purpose we synthesised the complex neutral glycosphingolipid, globoside, with spin-labelled fatty acid. Base hydrolysis to remove the natural fatty acid was found to deacetylate the GalNAc residue concomitantly, necessitating application of the synthetic route described for gangliosides by Neuenhofer et al. (Biochemistry 24, 525-532 (1985)). Globosides were produced with 18-carbon and 24-carbon fatty acids bearing a spin label at the C-16 position. Spin-labelled globosides were incorporated at 2 and 10 mol% into rigid, highly cooperative bilayer matrices of 1,2-dipalmitoylglycerophosphocholine (DPPC) and also into semi-fluid, non-cooperative membranes of DPPC/cholesterol. Recorded electron paramagnetic resonance (EPR) spectra were analysed by comparison with a library of standards representing samples of known composition. Spectra were manipulated using a computer program which permitted linear combination of standards to stimulate coexistence of laterally separated domains of different composition. The most important conclusions were as follows: (1) at least 80% of the globoside was definitely not confined to domains highly enriched in glycolipid, although there was evidence of binary-phase separation in the rigid DPPC/globoside matrix; (2) the presence of 33 mol% cholesterol reduced the evidence of globoside phase separation; (3) there was remarkably little difference in results whether the globoside fatty acid chain length was similar to that of the phospholipid host matrix or eight carbons longer. Temperature profiles derived over the phase-transition region of DPPC using spin-labelled globoside or an unattached amphiphilic spin label were consistent with these findings. The same systems lent themselves to consideration of the role of glycolipid fatty acid chan length and cholesterol in determining glycolipid crypticity in membranes: (1) polyclonal anti-globoside IgG bound to globoside in DPPC liposomes without inducing agglutination. (2) The same antibodies did agglutinate DPPC/cholesterol liposomes bearing globoside. (3) The effect of cholesterol probably was upon glycolipid dynamics or attitude in the membrane, rather than upon distribution. (4) These observations were basically unaffected by the choice of 18-carbon vs. 24-carbon glycolipid fatty acids.(ABSTRACT TRUNCATED AT 400 WORDS)

Evidence that trans-bilayer interdigitation of glycosphingolipid long chain fatty acids may be a general phenomenon.

'Interdigitation' is a term coined to describe the phenomenon whereby pure phosphatidylcholines with intramolecular fatty acid chain length heterogeneity when hydrated to form bilayers may insert the methyl ends of long fatty acids from one side across more than half of the membrane thickness to protrude amongst the acyl chains of the opposite side of the bilayer (Keough, K.M.W. and Davis, P.J. (1979) Biochemistry 18, 1453-1459; Huang, C. and Mason, J.T. (1986) Biochim. Biophys. Acta 864, 423-470). In this article we address the fate of long fatty acid chains of glycosphingolipids present as minor components in membranes of non-interdigitating phosphatidylcholines. In this pursuit, derivatives of galactosyl ceramide, lactosyl ceramide, globoside and GM1 were synthesized having either 18-carbon or 24-carbon fatty acid with a spin label covalently attached at C-16. Labelled glycolipids were incorporated at 1-2 mol% into bilayers of synthetic phosphatidylcholines, their mixtures with cholesterol, or natural egg phosphatidylcholine. In each case the C-16 carbon of the glycolipid long chain fatty acid showed considerably greater 'order' and immobility than did C-16 of the fatty acid which was similar in length to the host matrix phospholipids. We interpret this as strong evidence that the long chain fatty acid interdigitates across the mid point of the bilayer in the systems studied. Clearly this phenomenon did not require that the phospholipid host matrix have mixed chain lengths. Furthermore it was totally independent of glycolipid family: for a given host matrix and (glycolipid) fatty acid chain length the order parameter values found were the same amongst all four glycolipid families tested.

Double label freeze-etch study of the relative topography of concanavalin A and wheat germ agglutinin receptors at the myoblast surface.

This report records for the first time double-label freeze-etch electron microscopy of cells in culture. On L6 rat myoblasts, receptors for wheat germ agglutinin and concanavalin A were found distributed together in irregular granular microclusters of cell surface material up to 60 nm in diameter. Simultaneous localization of two different receptor families to such small regions using colloidal gold and ferritin to differentiate between lectin markers proved difficult in our hands. We were able to achieve the desired result using native concanavalin A and ferritin-conjugated wheat germ agglutinin, whose shadowed diameters are measurably different.

A long chain spin label for glycosphingolipid studies: transbilayer fatty acid interdigitation of lactosyl ceramide.

16-Carbon and 18-carbon fatty acids with covalently attached nitroxide free radicals have seen wide usage in membrane studies of phospholipid dynamics, orientation, and associations. However, they are inadequate for dealing with some very important questions that relate to glycosphingolipids. We report here the synthesis of a long chain (24-carbon) spin-labelled fatty acid designed for such problems. We have used both the new 24-carbon and the more conventional 18-carbon spin-labelled fatty acids to replace the natural fatty acid of lactosyl ceramide so that we may begin to compare short and long chain derivatives to analyse the molecular basis of their functional differences. Spectra seen are consistent with the view that in a bilayer host matrix the methyl end of the long fatty acid crosses the hydrophobic membrane center and interdigitates with fatty acids of phospholipids of the opposing monolayer.

Visualization of domains in rigid ganglioside/phosphatidylcholine bilayers: Ca2+ effects.

We have considered the extent to which details of lectin binding directly visualized by freeze-etch electron microscopy are consistent with current concepts of ganglioside arrangement in phosphatidylcholine bilayer membranes. Native lectins in general seem appropriate labels for this type of study. Wheat germ agglutinin, Ricinus communis agglutinin, and peanut agglutinin are adequately resolved on membrane surfaces as spherical particles of diameters 6 nm, 10 nm, and 13 nm, respectively (uncorrected for platinum shadow thickness). The finite areas covered by these markers correspond to some 56, 157, and 265 lipid molecules, respectively, on the surfaces of the shadowed rigid phosphatidylcholine matrices employed here; and this constitutes a basic limitation to the precision with which one can localize a given glycolipid receptor. Ricinus communis agglutinin provides a marker whose size permits adequate quantitation of bound material while minimally obscuring detail. Using it we estimated the size limits of GM1-enriched domains, since this is the ganglioside which has shown the greatest evidence of discontinuous distribution in our hands (Peters, M.W., Mehlhorn, I.E., Barber, K.R. and Grant, C.W.M. (1984) Biochim. Biophys. Acta 778, 419-428). Results of such analyses indicate the probable existence of phase separated domains selectively enriched in GM1 up to 60 nm in extent (5600 lipid molecules) for rigid dipalmitoylphosphatidylcholine membranes bearing up to 14 mol% GM1. Similar observations were true of rigid bilayers of dimyristoylphosphatidylcholine; however, if domains enriched in GM1 exist in fluid dimyristoylphosphatidylcholine, they are on the order of 6 nm or less in diameter (or are dispersed by lectin binding). Employing the small lectin, wheat germ agglutinin, which binds to all gangliosides, we then examined the effect of exposure to Ca2+ ions (while in the fluid state) on the ganglioside 'domain structure' referred to above in rigid dipalmitoylphosphatidylcholine host matrices. GM1, GD1a and GT1b were studied at 0, 2 and 10 mM Ca2+ concentrations. It was demonstrated by spin label measurements that the dipalmitoylphosphatidylcholine matrix retained its basic melting characteristics in the presence of added Ca2+ and ganglioside under these conditions. Within the technique's functional resolution limit of some 6 nm we were unable to identify any effect of Ca2+ in physiological concentration on ganglioside topography as reflected by bound lectin distribution.(ABSTRACT TRUNCATED AT 400 WORDS)

Freeze-etch visualization of the outer surface of myoblasts in culture.

The Pfenninger device is one of several types of specimen holders designed to permit freeze-fracture electron microscopy of cultured cells growing attached to solid substrates. It achieves this by directing a fracture plane horizontally through a monolayer of cells frozen with overlying medium (without need for prior disruption of cell attachments or relationships). The end result is a platinum-shadowed replica of the cell membrane hydrophobic interior. Here we describe the features seen when this traditional fracture step is followed by a lengthy etching step, making possible views of the cell membrane outer surface with a resolution 100 X better than that of fluorescence microscopy. Because of the technical difficulties involved, such views have in past been restricted to samples which may be handled in suspension, particularly blood cells and model membranes. Thus we have been able to examine extensive regions of the myoblast outer surface (the so-called etch face) at a magnification that permits visualization of details on the order of individual macromolecules. Prominent clumps of glycocalyx material occupying some 50% of the surface can be readily resolved as a closely spaced network of uniformly distributed 20- to 60-nm irregular granular patches. Receptors for wheat germ agglutinin were found to be associated almost exclusively with these surface prominences, so that bound lectin tended to exist in a uniform distribution of small clusters corresponding to the patches described above. When cells were not fixed until 15 min after lectin addition there was visibly more binding, but in a similar distribution. The details of cell surface architecture recorded here at a resolution of 2-4 nm are well below the limit of resolution of light microscopy and complement existing studies by fluorescence techniques. The presence of surface receptors in small patches reinforces the possibility that some literature observations of receptor interaction may be explained on the basis of direct receptor-receptor contact.

Evidence of a distribution difference between two gangliosides in bilayer membranes.

Freeze-etch electron microscopy, a platinum shadowing technique, has been used to compare the lateral distribution of several gangliosides in bilayer model membranes by directly visualizing bound lectin molecules. In particular, GM1 and GD1a, major components of brain ganglioside, were studied in phase-separated mixtures of dipalmitoyl- and dielaidoylphosphatidylcholines exposed to Ricinus communis agglutinin and wheat germ agglutinin. The distribution of glycolipid showed evidence of microheterogeneity in that bound lectin tended to occur in clusters of several or more molecules. With GD1a as receptor such clusters were small and very uniformly distributed over the membrane surface. Somewhat larger, irregularly spaced clusters of up to a dozen lectin particles were more typical of membranes bearing GM1 and, in addition, there were occasional extensive patches of bound lectin coexisting with areas apparently devoid of glycolipid receptor in phase-separated mixtures of dipalmitoyl- and dielaidoylphosphatidylcholine. Gangliosides in the latter mixtures were not obviously influenced in their lateral distribution by the presence of coexisting fluid and rigid domains. These basic observations seem to extend to bilayer membranes containing mixtures of two gangliosides. The patterns of lectin binding were not grossly affected by incubation time or history of warming and cooling. This study was extended to bilayers of pure dipalmitoylphosphatidylcholine in expectation that the distinctive features characteristic of the P beta' phase of this lipid might accentuate any behavioural differences between GM1 and GD1a. GM1 was found to exist preferentially in the 'trough' regions between P beta' ripples, while GD1a showed no apparent preferential arrangement. Given that bound lectins adequately reflect glycolipid distribution in membranes, it would appear that structurally different glycolipids from the same host membrane can assume different distributions on the basis of interactions with defined lipid host matrices.