While there has been a large amount of research into the phase behavior of lipid membranes in the past decade driven by the biological relevance of ordered micro-domains (often termed “lipid-rafts”), images of three coexisting phases are very rare. Researchers have difficulties preparing samples with a high proportion of ordered phase, and finding reliable methods to sufficiently discriminate the phases as fluorescent dyes often prefer one phase and are excluded from the remaining two. The cover image of the January 24th issue of the Biophysical Journal shows a collection of atomic force micrographs (AFM) exhibiting three-phase coexistence in a model cell membrane under water.
Three phases are found in lipid compositions. They lie in a narrow triangle below the more commonly studied two-phase micro-domain region, which contains the liquid disordered (Ld – yellow/red) and liquid ordered (Lo – magenta) phases, together with a gel phase (Lb – blue/green). Each phase has a different degree of chain packing order leading to varying depths. A surprising and counterintuitive finding of this study is that the gel phase, while definitely solid, is more disordered and slightly lower in height than Lo. It is also structurally very weak. This is explained by the small but significant quantity of cholesterol that disrupts the ordered solid phase, while being insufficient to form the Lo phase. Relative proportions of the three phases are governed by their position in the three-phase triangle. Domain morphology is controlled by the mechanism of phase separation. We observed examples of both spinodal and nucleated domains of each phase, and in some cases both mechanisms in the same image An example of this can be seen in the cover image where a nucleated gel phase (green) is surrounded by a percolated Lo/Ld structure (yellow/magenta). Another interesting finding was the signature of a radially varying composition across the nucleated gel domains, reflecting the kinetically trapped solid state in the process of separating from a compositionally varying melt. This effect has been commonly observed in metallurgy, but not in lipids.
Each image was produced within the standard AFM analysis software, Bruker Nanoscope v1.5. Manual coloring in Photoshop has not been used, rather a standard color look-up table (No 9) is mapped directly to the topography data. The contrast was adjusted so that the lowest phase is on the yellow/red transition, and the highest phase is magenta, resulting in a blue/green middle phase. A problem with this approach is the tiny 0.6 nm difference in height between the highest and lowest phases, which calls for accurate levelling to remove image bowing artefacts common in AFM, provide defined peaks in the depth histogram, and a uniform color across each phase. This task was made even more painstaking by the presence of three phases interfering with the thresholding and masking approach normally used. The final composite was created simply in Powerpoint.
Our work is part of an EPSRC (Engineering and Physical Sciences Research Council) Program entitled CAPiTALS, which is based around the fundamental understanding of the physics governing lipid membrane curvature, asymmetry, and patterning, and the technological uses arising from this new knowledge.
– Simon Connell, Anders Aufderhorst-Roberts, Udayan Chandra.