We determined the change in orientation of the electronic transition dipole moment with respect to the membrane normal in the K, L, and M intermediates of bacteriorhodopsin using transient linear dichroism. Purple membranes were oriented in a 14 T magnetic field and immobilized in a gel. The oriented purple membranes were excited isotropically and the transient absorbance changes were detected with the sample between two parallel polarizers. The absorbance changes were measured as a function of wavelength, time, and angle between the orientation axis and polarizer direction. In this way, the transient changes in isotropic absorbance, linear dichroism, and linear birefringence were determined with high accuracy. The Kramers−Kronig transform of the transient linear dichroism was in excellent agreement with the transient linear birefringence and this served as a useful control on the reliability of the linear dichroism data. We developed a novel formalism to extract the anisotropies, spectra and time courses of the photocycle intermediates in a model−independent way from the combined analysis of the transient absorbance and linear dichroism data. Whereas an analysis based on transient absorbance data alone is underdetermined, we show that in combination with transient linear dichroism data a unique solution may be obtained for the early intermediates K, L, and M. The analysis makes use of the constraints that (1) the sum of the populations of the K, L, and M intermediates is constant in time and (2) the absorption for the M intermediate vanishes for λ ≥ 520 nm. For wild-type bR at pH 7 (10 °C) we obtained in this way the following wavelength-independent anisotropies for the main absorption band:  r bR = −0.145, r K = −0.140, r L = −0.132, and r M = −0.139. Similar experiments were carried out for the mutant D96A which allows more accurate experiments for the L and M intermediates under various conditions of temperature and pH (pH 7, 20 °C; pH 4.7, 20 °C; pH 4.7, 10 °C). In all cases there are very clear differences in the anisotropies and the sequence is always r bR < r K < r M < r L. The data analysis is validated by the fact that the spectra and time courses of the intermediates are in excellent agreement with previous work. Making the reasonable assumption that the order parameter characterizing the orientational distribution is the same for each intermediate, the anisotropy changes translate into small orientational changes for the transition dipole moment:  ΔθK = −0.8 ± 0.2°, ΔθL = −1.7 ± 0.2°, ΔθM = −1.1 ± 0.3°. The largest change occurs in the L intermediate. The angle with respect to the membrane normal is smaller in every intermediate than in the ground state. The simplest interpretation of the results is that after the isomerization of the C13−C14 double bond the C5−N direction remains approximately the same with the C5−C13 part of the polyene chain tilting out of the plane of the membrane.

Transient linear dichroism and linear birefringence changes in the photocycle of bacteriorhodopsin at alkaline pH were measured with magnetically oriented purple membrane samples using the method of isotropic excitation (Borucki, B.; Otto, H.; Heyn, M. P. J. Phys. Chem. B 1999, 103, 6371−6383). At pH 10.4, the accumulation of the O intermediate is negligible, and N is the only intermediate that is relevant in the M decay and the recovery of the bR ground state. We introduced a new analytical approach that is model-independent and makes use of the transient linear dichroism in combination with isotropic absorbance changes. SVD analysis reveals that only one species (spectral component), namely, the N intermediate, is involved in the M decay at pH 10.4. The spectrum of this intermediate and its average anisotropy are determined from an eigenvalue equation. The transient linear dichroism data suggest a transition between two substates of the N intermediate with different anisotropies (i.e., with different orientations of the electronic transition dipole moment). Assuming that the polar angle of the transition dipole moment with respect to the membrane normal is 70° in the bR ground state, we get 66.3° (upper limit) for the first and 67.9° (lower limit) for the second N substate. The reorientation of the chromophore in the N1 → N2 transition is probably associated with the movement of the protonated Schiff base away from Asp96, suggesting that it serves as a reprotonation switch for Asp96 by stabilizing the protonation state of the Schiff base. In contrast to measurements at neutral pH, the linear birefringence changes at alkaline pH contain a wavelength-independent component that is not due to the transient absorption changes and whose time dependence is correlated with the kinetics of the N intermediate. This result is attributed to an electro-optical effect caused by the large electric field generated by the transient negative charge on Asp96 in the N intermediate. Alternatively this change in Δn may be due to the major conformational change in the cytoplasmic region of the membrane during N.