Linda J. Johnston

The size of lipid rafts: an atomic force microscopy study of ganglioside GM1 domains in sphingomyelin/DOPC/cholesterol membranes.

Atomic force microscopy has been used to study the distribution of ganglioside GM1 in model membranes composed of ternary lipid mixtures that mimic the composition of lipid rafts. The results demonstrate that addition of 1% GM1 to 1:1:1 sphingomyelin/dioleoylphosphatidylcholine/cholesterol monolayers leads to the formation of small ganglioside-rich microdomains (40-100 nm in size) that are localized preferentially in the more ordered sphingomyelin/cholesterol-rich phase. With 5% GM1 some GM1 microdomains are also detected in the dioleoylphosphatidylcholine-rich phase. A similar preferential localization of GM1 in the ordered phase is observed for bilayers with the same ternary lipid mixture in the upper leaflet. The small GM1-rich domains observed in these experiments are similar to the sizes for lipid rafts in natural membranes but considerably smaller than the ordered bilayer domains that have been shown to be enriched in GM1 in recent fluorescence microscopy studies of lipid bilayers. The combined data from a number of studies of model membranes indicate that lateral organization occurs on a variety of length scales and mimics many of the properties of natural membranes.

Atomic force microscopy studies of ganglioside GM1 domains in phosphatidylcholine and phosphatidylcholine/cholesterol bilayers.

The distribution of ganglioside in supported lipid bilayers has been studied by atomic force microscopy. Hybrid dipalmitoylphosphatidylcholine (DPPC)/dipalmitoylphosphatidylethanolamine (DPPE) and (2:1 DPPC/cholesterol)/DPPE bilayers were prepared using the Langmuir Blodgett technique. Egg PC and DPPC bilayers were prepared by vesicle fusion. Addition of ganglioside GM1 to each of the lipid bilayers resulted in the formation of heterogeneous surfaces that had numerous small raised domains (30--200 nm in diameter). Incubation of these bilayers with cholera toxin B subunit resulted in the detection of small protein aggregates, indicating specific binding of the protein to the GM1-rich microdomains. Similar results were obtained for DPPC, DPPC/cholesterol, and egg PC, demonstrating that the overall bilayer morphology was not dependent on the method of bilayer preparation or the fluidity of the lipid mixture. However, bilayers produced by vesicle fusion provided evidence for asymmetrically distributed GM1 domains that probably reflect the presence of ganglioside in both inner and outer monolayers of the initial vesicle. The results are discussed in relation to recent inconsistencies in the estimation of sizes of lipid rafts in model and natural membranes. It is hypothesized that small ganglioside-rich microdomains may exist within larger ordered domains in both natural and model membranes.

Distribution of Ganglioside GM1 in l-α-Dipalmitoylphosphatidylcholine/Cholesterol Monolayers: A Model for Lipid Rafts1

The distribution of low concentrations of ganglioside GM1 in L-alpha-dipalmitoylphosphatidylcholine (DPPC) and DPPC/cholesterol monolayers supported on mica has been studied using atomic force microscopy (AFM). The monolayers studied correspond to a pure gel phase and a mixture of liquid-expanded (LE) and liquid-condensed (LC) phases for DPPC and to a single homogeneous liquid-ordered phase for 2:1 DPPC/cholesterol. The addition of 2.5-5% GM1 to phase-separated DPPC monolayers resulted in small round ganglioside-rich microdomains in the center and at the edges of the LC domains. Higher amounts of GM1 (10%) give numerous filaments in the center of the LC domains and larger patches at the edges. A gel phase DPPC monolayer containing GM1 showed large domains containing a network of GM1-rich filaments. The addition of GM1 to a liquid-ordered 2:1 DPPC/cholesterol monolayer gives small, round domains that vary in size from 50 to 150 nm for a range of surface pressures. Larger amounts of GM1 lead to coalescence of the small, round domains to give longer filaments that cover 30-40% of the monolayer surface for 10 mol % GM1. The results indicate that biologically relevant GM1 concentrations lead to submicron-sized domains in a cholesterol-rich liquid-ordered phase that is analogous to that found in detergent-insoluble membrane fractions, and are thought to be important in membrane microdomains or rafts. This demonstrates that AFM studies of model monolayers and bilayers provide a powerful method for the direct detection of microdomains that are too small for study with most other techniques.

Nanoscale Imaging of Epidermal Growth Factor Receptor Clustering: EFFECTS OF INHIBITORS*

The development of some solid tumors is associated with overexpression of the epidermal growth factor receptor (EGFR) and often correlates with poor prognosis. Near field scanning optical microscopy, a technique with subdiffraction-limited optical resolution, was used to examine the influence of two inhibitors (the chimeric 225 antibody and tyrosine phosphorylation inhibitor AG1478) on the nanoscale clustering of EGFR in HeLa cells. The EGFR is organized in small clusters, average diameter of 150 nm, on the plasma membrane for both control and EGF-treated cells. The numbers of receptors in individual clusters vary from as few as one or two proteins to greater than 100. Both inhibitors yield an increased cluster density and an increase in the fraction of clusters with smaller diameters and fewer receptors. Exposure to AG1478 also decreases the fraction of EGFR that colocalizes with both rafts and caveolae. EGF stimulation results in a significant loss of the full-length EGFR from the plasma membrane with the concomitant appearance of low molecular mass proteolytic products. By contrast, AG1478 reduces the level of EGFR degradation. Changes in receptor clustering provide one mechanism for regulating EGFR signaling and are relevant to the design of strategies for therapeutic interventions based on modulating EGFR signaling.

Fluorescence correlation spectroscopy with sub-diffraction-limited resolution using near-field optical probes

We report fluorescence correlation spectroscopy (FCS) measurements using near-field scanning optical microscopy (NSOM) probes to produce a sub-diffraction-limited observation area. An order of magnitude reduction in the area compared to confocal FCS has been achieved. We also demonstrate a simple means to model the autocorrelation decay due to diffusion within the excitation profile at the NSOM probe aperture. The use of probes with smaller apertures is expected to provide an additional order of magnitude reduction in the observation area, thus enabling the study of cellular membranes with higher concentrations of fluorophores than is currently possible with diffraction-limited techniques.

Near-field scanning optical microscopy probes: a comparison of pulled and double-etched bent NSOM probes for fluorescence imaging of biological samples

Bent near‐field optical probes for biological applications have been fabricated using a combination of a two‐step chemical etching method and focused ion beam milling to create a well‐defined aperture. The transmission efficiencies have been evaluated as a function of laser wavelength (λ) and aperture size (D) for both large and small core fibres. The probe transmission behaviour follows a (D/λ)3 relationship. The double‐etched probes are compared to pulled probes fabricated from highly GeO2‐doped dispersion compensating fibre and a standard single‐mode optical fibre. The transmission efficiencies of both types of pulled probes are approximately two orders of magnitude lower than double‐etched probes with similar aperture sizes. To demonstrate the utility of the various probes, their imaging performance has been evaluated for samples of polymer beads and phase‐separated phospholipid monolayers of dipalmitoylphosphatidylcholine or cholesterol/phosphatidylcholine/sphingomyelin mixtures. Both pulled and double‐etched probes are suitable for fluorescence imaging of polymer spheres. However, pulled probes are rapidly damaged at the higher input laser intensities required for fluorescence imaging of monolayer samples doped with < 1% of a fluorescent dye‐labelled lipid. The images obtained with the double‐etched probes show excellent spatial resolution and signal/noise, illustrating the potential of such probes for imaging of biological samples.

Enzymatic generation of ceramide induces membrane restructuring: Correlated AFM and fluorescence imaging of supported bilayers.

The effect of enzymatic generation of ceramide on phase separated bilayers with a mixture of co-existing fluid and liquid-ordered phases has been examined using a combination of atomic force microscopy (AFM) and fluorescence imaging. Supported lipid bilayers prepared from a DOPC/sphingomyelin/cholesterol mixture were imaged prior to, during and after incubation with sphingomyelinase by total internal reflection fluorescence (TIRF) microscopy. Enzyme treatment resulted in the growth of large dye-excluded regions. The growth kinetics for these patches are consistent with activity of a variable number of enzyme molecules in different regions of the bilayer. Correlated AFM and fluorescence imaging shows that some of the large dye-excluded patches form around the original liquid-ordered domains, which become heterogeneous in height with many raised ceramide-rich regions around their periphery. However, some of the dye-excluded patches correspond to areas of the bilayer where the initial domains have largely or partially disappeared. The dye-excluded patches observed by fluorescence are shown to be areas of increased adhesion in lateral deflection AFM images and are postulated to form by incorporation of both cholesterol and ceramide in the original fluid phase and to vary in composition throughout the bilayer. This is evident from the observation that the dye-excluded areas are all detected as areas of increased friction, but do not always show a distinct height difference in topographic images. These results highlight the utility of a multi-modal imaging approach for understanding the complex membrane restructuring that occurs upon enzymatic generation of ceramide.

NBD-cholesterol probes to track cholesterol distribution in model membranes.

A series of cholesterol (Chol) probes with NBD and Dansyl fluorophores attached to the 3-hydroxyl position via carbamate linkers has been designed and synthesized and their ability to mimic the behavior of natural cholesterol in bilayer membranes has been examined. Fluorescence spectroscopy data indicate that the NBD-labeled lipids are located in the polar headgroup region of the bilayer with their position varying with the method of fluorophore attachment and the linker length. The partitioning of the Chol probes between liquid-ordered (L(o)) and liquid-disordered (L(o)) phases in supported bilayers prepared from ternary lipid mixtures of DOPC, Chol and either egg sphingomyelin or DPPC was examined by fluorescence microscopy. The carbamate-linked NBD-Chols show a stronger preference for partitioning into L(o) domains than does a structurally similar probe with an ester linkage, indicating the importance of careful optimization of probe and linker to provide the best Chol mimic. Comparison of the partitioning of NBD probes to literature data for native Chol indicates that the probes reproduce well the modest enrichment of Chol in L(o) domains as well as the ceramide-induced displacement of Chol. One NBD probe was used to follow the dynamic redistribution of Chol in phase separated membranes in response to in situ ceramide generation. This provides the first direct optical visualization of Chol redistribution during enzymatic ceramide generation and allows the assignment of new bilayer regions that exclude dye and have high lateral adhesion to ceramide-rich regions.

Correlating Cellulose Nanocrystal Particle Size and Surface Area

Cellulose nanocrystals (CNCs) are negatively charged nanorods that present challenges for characterization of particle size distribution and surface area-two of the common parameters for characterizing nanomaterials. CNC size distributions have been measured by two microscopy methods: atomic force microscopy (AFM) and transmission electron microscopy (TEM). The agreement between the two methods is good for length measurements, after taking into consideration tip-convolution effects for AFM. However, TEM widths are almost twice as large as AFM heights-an effect that we hypothesize is due to counting of a larger fraction of laterally associated CNCs in the TEM images. Overall, the difficulty of selecting individual particles for analysis and possible bias due to selection of a specific particle size during sample deposition are the main limitations associated with the microscopy measurements. The microscopy results were compared to Z-average data from dynamic light scattering, which is a useful method for routine analysis and for examining trends in size as a function of sample treatment. Measurements as a function of sonication energy were used to provide information on the presence of aggregates in the sample. Magic-angle-spinning solid-state NMR was employed to estimate the surface area of CNCs based on the ratio of integrated spectral intensities of resonances stemming from C4 sites at the crystallite surfaces and from all C4 sites. Our approach was adapted from the application of solid-state NMR to characterize larger cellulose microfibers and appears to provide a useful estimate that overcomes the limitations of using the BET method for measuring surface areas of highly aggregated nanomaterials. The solid-state NMR results show that the lateral dimension of the CNCs is consistent with that of elementary cellulose crystallites.

Cholesterol Tuning of BK Ethanol Response Is Enantioselective, and Is a Function of Accompanying Lipids

In the search to uncover ethanols molecular mechanisms, the calcium and voltage activated, large conductance potassium channel (BK) has emerged as an important molecule. We examine how cholesterol content in bilayers of 1,2-dioleoyl-3-phosphatidylethanolamine (DOPE)/sphingomyelin (SPM) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylethanolamine (POPE)/1-palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylserine (POPS) affect the function and ethanol sensitivity of BK. In addition, we examine how manipulation of cholesterol in biological membranes modulates ethanols actions on BK. We report that cholesterol levels regulate the change in BK channel open probability elicited by 50 mM ethanol. Low levels of cholesterol (<20%, molar ratio) supports ethanol activation, while high levels of cholesterol leads to ethanol inhibition of BK. To determine if cholesterol affects BK and its sensitivity to ethanol through a direct cholesterol-protein interaction or via an indirect action on the lipid bilayer, we used the synthetic enantiomer of cholesterol (ent-CHS). We found that 20% and 40% ent-CHS had little effect on the ethanol sensitivity of BK, when compared with the same concentration of nat-CHS. We accessed the effects of ent-CHS and nat-CHS on the molecular organization of DOPE/SPM monolayers at the air/water interface. The isotherm data showed that ent-CHS condensed DOPE/SPM monolayer equivalently to nat-CHS at a 20% concentration, but slightly less at a 40% concentration. Atomic force microscopy (AFM) images of DOPE/SPM membranes in the presence of ent-CHS or nat-CHS prepared with LB technique or vesicle deposition showed no significant difference in topographies, supporting the interpretation that the differences in actions of nat-CHS and ent-CHS on BK channel are not likely from a generalized action on bilayers. We conclude that membrane cholesterol influences ethanols modulation of BK in a complex manner, including an interaction with the channel protein. Finally, our results suggest that an understanding of membrane protein function and modulation is impossible unless protein and surrounding lipid are considered as a functional unit.