Why we should study protein binding to bicontinuous inverted cubic phases, to better understand membrane trafficking.

Abstract

Many proteins bind to membranes in a curvature-sensitive manner. This includes some adaptors/effectors in the vesicle budding and fusion stages of membrane trafficking. Recently, others showed that some proteins bind preferentially to membranes with Gaussian curvature: they bind more strongly to spherical vesicles (with positive Gaussian curvature) than to tubules with the same mean curvature but zero local Gaussian curvature. Fusion and fission pores (which are intermediates in vesicle fusion & budding) have special curvature properties. Here I show that the surfaces of such pores have curvature that is qualitatively and quantitatively different than in the membrane systems conventionally used to assess curvature-sensitive protein binding (membrane vesicles and tubules). Hence, existing methods may miss proteins that have evolved to bind to pore-relevant curvatures. However, the curvature properties of bicontinuous inverted cubic (QII) phases are quite similar to those of fusion/fission pores. Specifically, both QII phases and fusion/fission pores have negative Gaussian curvature, which is true of neither vesicles nor tubules; have more negative mean curvature than many tubules; and gradients in mean curvature that are either absent or smaller in vesicles and tubules. QII phases can be equilibrated with peptide through the aqueous phase: it has been shown that proteins with molecular weights of several tens of kDa can enter the water channel networks and penetrate into aggregates of QII phase. Thus, we might find new adaptors/effectors of trafficking by studying peptide or protein binding to preparations of QII phase. Some features of QII phases make such experiments awkward. Only certain ranges of lipid compositions form QII phases, and, in phospholipids, temperature-cycling techniques or particular electrolyte compositions may be needed. Appropriate protocols will be proposed.