Main

The simplest way to describe a membrane is as a lipid bilayer, but it is actually a complex network of lipid, protein and sugar components, the organization of which is thought to have functional significance for the cell. Although it is possible to 'see' small subsets of these components—such as a complex of membrane receptors—by X-ray crystallography, methods that allow greater late-ral resolution are necessary to study entire membrane domains. Fluorescence microscopy is often used to image differentially labeled components at such a resolution, but addition of fluorescent labels can change the physical properties of the relatively small components of biological membranes.

Thus, to measure model membrane composition, Steven Boxer of Stanford University and his colleagues turned to secondary ion mass spectrometry (SIMS), a method originally developed over 50 years ago. They used a high-resolution version of this technology, NanoSIMS, to analyze supported bilayers of two lipids labeled with different stable isotopes (Kraft et al., 2006). In this technique, the sample is bombarded with a beam of ions, and the secondary ions emitted from the sample are extracted and measured by mass spectrometry (Fig. 1). The resulting secondary ion data for each of the isotopes can be used to create an image of the distribution of the two lipids in the bilayer.

Figure 1: Schematic of the NanoSIMS analysis.
figure 1

As the raster beam passes across the supported lipid bilayer containing domains of two labeled lipids, 13C18-DSPC and 15N-DLPC, secondary ions are generated. Ions characteristic of each lipid (in red and green, respectively) can be used to identify these lipids in the NanoSIMS image to create a chemical composition map of the surface. Reprinted with permission from AAAS.

Full size image

In analyzing the NanoSIMS data, Boxer's team found that the lipid distribution varied significantly in some of the domains. The NanoSIMS data also helped explain some of the irregularities observed in the atomic force microscopy (AFM) analysis of the same samples. AFM, which provides information about the 'bumps and lumps' in samples, indicated that there were some structures of unknown origin within the observed domains. Comparison to NanoSIMS data indicated that the observed objects were not subdomains because they did not have the chemical signatures of either of the labeled lipids, and the authors were able to conclude that they were debris—just dirt.

Taken together, data produced by these two methods provide a clearer picture of the system. ”The combination is more powerful than either technique by itself,“ says Boxer. AFM gives a very high degree of lateral information, ”much higher than any version of imaging mass spectrometry can now or probably will ever give, ... and it's fantastic for looking at topography, but it does not tell you at all what the molecules are. And mass spectrometry in its many variations, including [NanoSIMS], is the ideal method to analyze what's there because it gives you a chemical map of what's there.“

Just as the combination of imaging techniques provides the most information about the membrane, different methods of preparing the membrane for imaging will aid this effort as well. Some groups are now developing methods to image membrane components in a nonsupported format (see, for example, Gonçalves et al., 2006). Boxer and colleagues continue to develop the supported bilayer system with the hope of building up to a more complex system. Other groups are developing ways of removing a part of a membrane from a cell and putting it on a surface. This is one of the avenues Boxer's team is pursuing, and he adds that ”if one could do this quickly enough, so that the molecules have not moved very far on the time scale of this ripping-apart process, then the hope is that one can interrogate the lateral composition of membranes derived from real cells. That is the dream experiment, in a sense. It just requires many technical developments.“