Electro-optics of membrane electroporation in diphenylhexatriene-doped lipid bilayer vesicles.

The electric (linear) dichroisms observed in the membrane electroporation of salt-filled lipid bilayer vesicles (diameter O = 2 alpha = 0.32 micron; inside [NaCl] = 0.2 M) in isotonic aqueous 0.284 M sucrose-0.2 mM NaCl solution indicate orientation changes of the anisotropic light scattering centers (lipid head groups) and of the optical transition moments of the membrane-inserted probe 1,6-diphenyl-1,3,5-hexatriene (DPH). Both the turbidity dichroism and DPH absorbance dichroism show peculiar features: (1) at external electric fields E > or = Esat the time course of the dichroism shows a maximum value (reversal): Esat = 4.0 (+/- 0.2) MV m-1, T = 293 K (20 degrees C), (2) this reversal value is independent of the field strength for E > or = Esat, (3) the dichroism amplitudes exhibit a maximum value Emax = 3.0 (+/- 0.5) MV m-1, (4) for the pulse duration of 10 microseconds there is one dominant visible normal mode, the relaxation rate increases up to tau-1 approximately 0.6 x 10(6) s-1 at Esat and then decreases for E > Esat. The data can be described in terms of local lipid phase transitions involving clusters Ln of n lipids in the pore edges according to the three-state scheme C<-->HO<-->HI, C being the closed bilayer state, HO the hydrophobic pore state and HI the hydrophilic or inverted pore state with rotated lipid and DPH molecules. At E > or = Esat, further transitions HO<-->HO* and HI<-->HI* are rapidly coupled to the C<-->HO transition, which is rate-limiting. The vesicle geometry conditions a cos theta dependence of the local membrane field effects relative to the E direction and the data reflect cos theta averages. The stationary induced transmembrane voltage delta phi (theta, lambda m) = -1.5 aEf(lambda m) magnitude of cos theta does not exceed the limiting value delta phi sat = -0.53 V, corresponding to the field strength Em,sat = -delta phi sat/d = 100 MV m-1 (10(3) kV cm-1), due to increasing membrane conductivity lambda m. At E = Esat, f(lambda m) = 0.55, lambda m = 0.11 mS m-1. The lipid cluster phase transition model yields an average pore radius of rp = 0.35 (+/- 0.05) nm of the assumed cylindrical pore of thickness d = 5 nm, suggesting an average cluster size of <n> = 12 (+/- 2) lipids per pore edge. For E > Esat, the total number of DPH molecules in pore states approaches a saturation value; the fraction of DPH molecules in HI pores is 12 (+/- 2)% and that in HO pores is 48 (+/- 2)%. The percentage of membrane area P approximately (lambda m/lambda i) x 100% of conductive openings filled with the intravesicular medium of conductance lambda i = 2.2 S m-1 linearly increases from P approximately 0% (E = 1.8 MV m-1) to P = 0.017% (E = 8.5 MV m-1). Analogous estimations made by Kinosita et al. (1993) on the basis of fluorescence imaging data for sea urchin eggs give the same order of magnitude for P (0.02-0.2%). The increase in P with the field strength is collinear with the increase in concentration of HI and HI* states with the field strength, whereas the HO and HO* states exhibit a sigmoid field dependence. Therefore our data suggest that it is only the HI and HI* pore states which are conductive. It is noted that the various peculiar features of the dichroism data cannot be described by simple whole particle deformation.