Electron-phonon coupling and vibronic fine structure of light-harvesting complex II of green plants: Temperature dependent absorption and high-resolution fluorescence spectroscopy

Polarized, site-selected fluorescence was measured for light-harvesting complex II (LHCII), the major Chi a/b/xanthophyll binding light-harvesting complex of green plants. Upon selective excitation in the range of 679-682 nm at 4 K, separate zero-phonon lines and phonon wings could be observed, as well as sharp lines in the vibronic region of the emission: vibronic zero-phonon lines. The maximum of the phonon wing was located 22 cm(-1) to the red of the zero-phonon line. Forty-eight vibrational modes could be identified, and their Franck-Condon factors were estimated. From the vibrational frequencies it is concluded that the Chi a responsible for the emission at 4 K is monoligated and accepts a hydrogen bond on the 13(1)-keto group. Also measured was the temperature dependence of the absorption spectrum of LHCII. Using the phonon wing obtained from the fluorescence measurements and an algorithm based on linear, harmonic Franck-Condon electron-phonon coupling and temperature independent inhomogeneous broadening, the temperature dependence of the low-energy part of the Q(r) absorption spectrum could be simulated very well up to 220 K. Above this temperature, the simulated and experimental results start to deviate. From the simulations it is concluded that inhomogeneous broadening of the long-wavelength band(s) (676 nm and above) is 120 +/- 15 cm(-1) below 220 K, whereas the Huang-Rhys factor of the protein phonons is 0.6 +/- 0.1 (at 4 K). We have modeled the results from absorption, fluorescence, hole-burning, triplet-minus-singlet absorption, and fluorescence anisotropy measurements by one Gaussian inhomogeneous distribution function (peaking near 676 nm) with the spectroscopic properties of the lowest energy state(s) at 4 K. There was a significant discrepancy between the results from the simulations and the experiments. A much better agreement could be obtained by assuming either two Gaussian distributions (centered around 676 and 680 nm with an intensity ratio of 11:1) or a non-Gaussian distribution around 676 nm. Although we cannot discriminate between these two descriptions, both simulations have in common that at least nine separate electronic states per trimer are present in the 676 (and 680 nm) band.