Why are lipid rafts not observed in vivo?

Abstract

The existence of lipid rafts in live cells remains a topic of lively debate. Although large, micrometer-sized rafts are readily observed in artificial membranes, attempts to observe analogous domains in live cells place an upper limit of approximately 5 nm on their size. We suggest that integral membrane proteins attached to the cytoskeleton act as obstacles that limit the size of lipid domains. Computer simulations of a binary lipid mixture show that the presence of protein obstacles at only 5-10% by area dramatically reduces the tendency of the lipids to phase separate. These calculations emphasize the importance of spatial heterogeneity in cell membranes, which limits the transferability of conclusions drawn from artificial membranes to live cells.

FIGURE 1

Picture of Ising model where diamonds represent membrane protein obstacles and blue and red squares the two lipid components A and B.

FIGURE 2

Lipid composition probability distribution functions P(x_A) for L = 50 at the temperatures shown for a pure lipid mixture (no obstacles). Above the critical temperature, P(x_A) has one peak, and below the critical temperature it has two peaks, which are located at the concentrations of the coexisting phases. At the critical temperature for this system size (dashed curve), P(x_A) is very flat. The apparent critical temperature for this system size is T_c = 51°C. From finite size scaling, we determine that the true (infinite system) critical temperature is T_c = 40°C.

FIGURE 3

Determination of the critical temperature of the infinite systems from simulations of finite systems. The Binder ratio is plotted as a function of temperature for various system sizes. The true critical temperature is the temperature at which the Binder ratio is independent of system size, i.e., the three curves cross.

FIGURE 4

Lipid composition probability distribution functions P(x_A) for L = 50 at the temperatures shown for a lipid mixture in the presence of obstacles at 10% by area. P(x_A) at the apparent critical temperature (for this system size) is given by the dashed curve, which is for T_c = 26°C. The true critical temperature, obtained by finite size scaling, is T_c = 5°C.

FIGURE 5

Variation of the true critical temperature with the area fraction of protein obstacles. The dashed line is the prediction of the mean-field theory described in the text.

FIGURE 6

Coexistence curve of the lipid mixture for L = 50. Lines are meant to guide the eye.

FIGURE 7

Snapshot of a simulation with L = 50 and φ = 0.1 for T = 37°C, which is below T_c.