EVIDENCE FOR TRANSBILAYER, TAIL-TO-TAIL CHOLESTEROL DIMERS IN DIPALMITOYLGLYCEROPHOSPHOCHOLINE LIPOSOMES

The behavior of multilamellar liposomes of 2,3-dipalmitoyl-sn-glycero- 1-phosphocholine (DPPC) was studied by differential scanning calorimet ry (DSC) in the presence of less than or equal to 5 mol % of the amphi philic solutes methyl oleate, cholesterol, pregnenolone, and dehydroan drosterone. The DSC thermograms indicate that the solutes are miscible only with the liquid-disordered (1(d)) phase, and not with the solid- ordered (S-o) phase. The slopes of the T-m vs solute concentration cur ves confirm this conclusion: It appears that the s(o)-1(d) phase trans ition of DPPC, which corresponds to the melting of the phospholipid ch ains, can be treated as a simple melting process and, thus, could be u sed as a cryoscopic system. In that case, its melting point depression constant, K-f, can be calculated a priori from the experimentally mea sured heat of fusion per gram of DPPC, l(f), and the temperature of th e phase transition of pure DPPC, T-o, by the equation K-f = RT(o)(2)/( 1000l(f)) = 12.3 +/- 0.9 K g M(-1) cm(3). With methyl oleate as the so lute, the T-m vs methyl oleate concentration plot is linear, and from the slope we calculate K-f 12.9 +/- 0.8 K g M(-1) cm(3). Thus, methyl oleate appears to form an ideal cryoscopic system with dipalmitoylleci thin liposomes: It is fully miscible with the id phase but is apparent ly insoluble in the s(o) phase. Pregnenolone and dehydroandrosterone a lso form ideal cryoscopic systems with dipalmitoyllecithin liposomes: The T-m vs solute concentration plots are linear and yield the correct MWs for these solutes. In contrast, the T-m vs concentration plot of cholesterol is curved; the slope below 2 mol % solute corresponds to t he monomeric MW, but above 3 mol % it approaches that of a dimer. The full curve is consistent with a simple equilibrium dimerization of the solute in the 1(d) phase. Thus, cholesterol appears to be insoluble i n the s(o) phase, and it readily forms a dimer in the 1(d) phase. Spac e-filling considerations indicate that the dimer is a transbilayer one . Such a dimer of cholesterol would be consistent with a variety of ob servations reviewed by Sankaram and Thompson [Sankaram, M. B., and Tho mpson, T. E. (1990) Biochemistry 29, 10676-10684]. Since cholesterol i s insoluble in the s(o) phase, it must form below T-m a separate phase or at least separate domains. The domains, however, should not be too small since the relatively large boundary region of the microdomains would affect the thermodynamic stability of the solid phase.