Current evidence strongly suggests that a substantial fraction of the phospholipids in many membranes are arranged into fluid bilayers (1). Although the word "bilayer" is self-explanatory, the term "fluid" is not. This paper summarizes recent data on the fatty acid configurations, dynamics, and conformations that result in the fluidity of sonicated lecithin bilayers.
In n-alkanes the configurations about each C-C bond are termed trans and gauche isomers. In the gaseous and liquid states of most alkanes the trans isomer is 500-800 cal/mole more stable than either of the two gauche states (2). Therefore, a Boltzmann distribution predicts 10-12 gauche configurations for each pair of 16 carbon n-alkane chains. Several lines of evidence suggest that for sonicated lecithin above its transition temperature this value is 8-10 (3,6). From measurements of the volume and entropy changes during the endothermic transition of dipalmitoyl-lecithin, and from estimates of the bilayer thickness and surface area per molecule, TrIuble and Haynes (3) concluded that there are 3-8 gauche configurations per molecule. Nagle (4) has also treated the calorimetric data from DPL and suggests a value of 7-10 as most reasonable. The laser Raman studies of Lippert and Peticolas (5,6) provide a more direct estimate. The Raman spectra, which are sensitive to the presence of gauche configurations, of sonicated dipalmitoyl and dioleolyllecithin are similar to those of liquid n-alkanes. For dioleolyllecithin this is true on both sides of the double bond (6). Thus there are at least two gauche configurations above and at least two to three below the double bond, or a total of 8-10 per molecule.
The dynamics of fatty acid segmental motions can be inferred from the nuclear relaxation times of the methylene protons and carbons. The values of T1 for these resonances are less than an order of magnitude shorter than those for the same molecules in organic solvents (7-9) but are similar t o those for neat n-decanol(l0). The value of the correlation time estimated from the T1 data is about an order of magnitude longer than that calculated from an Arrhenius equation with an activation energy of 2.4-3.0 kcal/mole and a preexponential factor of 3.3 X 1OI2 sec -' (9, 11). These values can be construed t o yield an effective Stokes-Einstein viscosity which is within an order of magnitude of that of a liquid.
The activation energy for the T1 process is -3.0 kcal/mole (12), a value which is