Human proteolipid protein (PLP) mediates winding and adhesion of phospholipid membranes but prevents their fusion.

Proteolipid protein (PLP or lipophilin) is a highly conserved, strongly hydrophobic, integral membrane protein, and is the major protein component of central nervous system myelin. Although PLP has been implicated in many functions, its in vivo role is still uncertain. Here, we report the investigation of PLP's putative adhesive function using purified PLP and reconstituted phospholipid vesicles made of either 100% phosphatidylcholine (PC), or a mixture of 92% PC and 8% phosphatidylserine (PS), by weight. PLP-induced changes in the phospholipid bilayer surfaces were directly examined by transmission electron microscopy. We found that upon the introduction of PLP, larger lipid vesicles became smaller and unilamellar. At the PLP:lipid molar ratio of 1:20, vesicle membranes rolled onto themselves forming 'croissant'-like structures that subsequently adhered to each other. The phenomena of PLP-induced bilayer rolling and adhesion were dependent on the concentration of PLP and the period of incubation, but were independent of the presence of calcium and types of phospholipids (PC or PC:PS). Furthermore, the presence of PLP in the lipid bilayers prevented the fusion of membranes. These findings show that PLP can induce membrane 'winding' while preventing the fusion of adjacent lipid bilayers. Hence, our data provide direct evidence for PLP's suspected function of membrane adhesion, and also suggest that PLP could potentially play a role in the formation of the myelin sheath.

Topology of recombinant rat upstream binding factor.

Transmission electron microscopy and single particle electron crystallography were employed to reconstruct high-quality projection images of a recombinant, acidic tail deficient form of rat upstream binding factor. The upstream binding factor was found to be dimeric and approximately 10 nm in diameter with a central region of low density. Distinct nodes were observable, of size and spacing consistent with being HMG boxes 3 and 4. The dimerisation domain seemed most probably to be located in the internal region of the structure.