Fluctuations caught in the act (original) (raw)
Large numbers of interacting particles show dramatic phase transitions and critical phenomena. At a critical point, mole- cular-scale interactions build up to macroscopic scales, resulting in universal behaviour1. The discovery of critical opalescence at the critical point of carbon dioxide2, caused by density fluctuations scattering light in the fluid, stimulated attempts to glimpse these fluctuations at all scales.
A monolayer of phospholipids at the interface of air and water can be brought close to a critical point by varying the temperature and surface pressure of the monolayer in a Langmuir trough3. We immobilized the monolayer by transferring it, by horizontal dipping, to a solid, hydrophilic substrate of mica using Langmuir–Blodgett techniques4. We then imaged critical fluctuations varying in size from nanometres to micrometres using contact-mode atomic-force microscopy.
Figure 1a and b shows two different lipid monolayers — dimyristoyl phosphatidylcholine (DMPC) and dipalmitoyl phosphatidylcholine (DPPC) — close to their critical points. When the monolayer passes through the critical point, the acyl chains of the lipids become disordered and the monolayer becomes thinner. As atomic force microscopy maps the height contours of the monolayer, these images show a monolayer structure with domains of one phase within the other.
Figure 1: Critical fluctuations in phospholipid monolayers imaged by atomic-force microscopy as a height-difference map.
a,b, Images of dimyristoyl phosphatidylcholine (25 × 25 μm2) and dipalmitoyl phosphatidylcholine (20 × 20 μm2) monolayers at their respective critical points. The monolayers have been transferred from an air–water interface to solid mica supports. The patterns correspond to lipid domains of one phase immersed into the other. The height difference between the light and dark areas is about 5 Å. c, Quantitative analysis of many images like those in a and b in terms of the structure factor, S(q), in arbitrary units, shown in a double-logaritmic plot as a function of the wavevector, q. The wavevector is related to distance, r, as q = 2π/r. The data for both phospholipids scale as S(q ) ≈ _q_−2 x, with x ≈ 1 for q ≤ 20 μm−1. At low q there is a deviation from this power law because of the finite size of the images analysed.
The height difference between the two types of domain corresponds to the chain-length difference between an ordered and a disordered lipid acyl chain. Near the critical point, these domains are dynamic and fluctuate strongly when the monolayer is at the air–water interface; we capture them immobilized during the transfer process. The morphology of the domains is very ramified, owing to the vanishing line tension near criticality. Domains of all sizes appear, indicating that there is no characteristic length scale — the hallmark of critical fluctuations.
Figure 1c shows a quantitative analysis of the domain patterns in terms of the structure factor, S(q), where q is the length of a two-dimensional wavevector. The structure-factor data for both DMPC and DPPC scale as a power law, S( q) ∼ q_−2_x, with x ≈ 1. Within the accuracy of the data, this is consistent with the critical-point behaviour associated with the universality class of the two-dimensional Ising model1.
Lipid domains in monolayers made of phospholipid mixtures can be immobilized and imaged using similar techniques. Mixtures of phospholipids with acyl chains of different lengths show compositional demixing on the nanometre scale5.
Monolayers are simple model systems of the lipid bilayer component of cell membranes6. Functions supported by lipid bilayers, such as phospholipase activity7 and protein binding8, are correlated with lipid-domain formation in the nanometre range that is controlled by the underlying phase transition in the bilayer. Biological membranes contain many different lipid species that are probably organized heterogeneously in the form of domains9,10,11 or ‘rafts’12 in the plane of the membrane, though the principles underlying this organization remain largely unknown. Our results suggest that fluctuations in lipid bilayer properties, for example, density or composition, may be responsible for lipid-domain formation in biomembranes.