ИСТИНА

Non-photochemical quenching (NPQ) of chlorophyll fluorescence is a fast response mechanism developed by photosynthetic organisms to protect them from excessive light illumination which can lead to irreversible damage of photosynthetic apparatus. Excitation energy transfer from a pool of chlorophylls a to lutein molecule is supposed to be the main pathway of NPQ in LHCII complex. More precisely, energy transfer rate from the S1 state of the closest chlorophyll to the S1 state of lutein governs rate of NPQ, and is proportional to the squared exciton coupling between them according to Fermi's golden rule. Optically dark S1 state of lutein (2Ag state if one assumes ideal polyene symmetry for lutein) has strong multireference character, and requires large active spaces in MCSCF to account for static correlation. We studied an impact of active space size in MCSCF calculations on exciton coupling in the aforementioned pigment pair. Ground state geometries for both pigments were optimized in vacuo at B3LYP/6-31G* level, dipole moment and transition density for S0->S1 transition were obtained using MCSCF (in CASSCF and RASSCF formulations). pi-orbitals in active space were prepared by projecting orbitals from minimal ANO-MB basis set to ANO-VDZP which was used for final calculation. For CASSCF, active spaces from (4,4) to (14,14) were examined; for RASSCF, two sets were examined: from (20,1,1;8,4,8) to (20,1,1;5,10,5) and (20,2,2;8,4,8) to (20,2,2;6,8,6). Averaging with equal weights was performed for 4 lowest states. Additionally, DMRG(20,20)[256] calculation was performed to estimate "full CI" limit of active space containing all pi orbitals. S1 state of chlorophyll a was calculated using RASSCF(20,2,2;8,4,8) to be consistent with results for lutein. Transition densities calculated as described were used to derive effective charges for transition electrostatic potential (TrESP-charges). In order to evaluate exciton coupling, charges were placed on MM-optimized structure of LHCII complex and Coulomb interactions were calculated between them. We have found significant dependence of transition dipole of lutein and, consequently, exciton coupling on active space: dipole varies from 0.40 D for CAS(6,6) (smallest active space predicting correct relative position of 2Ag state) to 0.060 D for RASSCF(20,1,1;5,10,5). Therefore, due to quadratic energy transfer rate dependence on exciton coupling, calculations in large active spaces are crucial for obtaining correct NPQ rates.