Search

Exciton model for description of experimentally determined excitation energy transfer from carotenoids to chlorophylls in the LHC-II trimer of spinach is presented. Such an approach allows connecting the excitonic states to the spatial structure of the complex and hence descriptions of advancements of the initially created excitations in space and time. Carotenoids were excited at 490 nm and at 500 nm and induced absorbance changes probed in the Chl Q(y) region to provide kinetic data that were interpreted by using the results from exciton calculations. Calculations included the 42 chlorophylls and the 12 carotenoids of the complex, Soret, Q(x) and Q(y) states of the chlorophylls, and the main absorbing S(2) state of the carotenoids. According to the calculations excitation at 500 nm populates mostly a mixed Lut S(2) Chl a Soret state, from where excitation is transferred to the Q(x) and Q(y) states of the Chl a's on the stromal side. Internal conversion of the mixed state to a mixed Lut S(1) and Chl a Q(y) state provides a channel for Lut S(1) to Chl a Q(y) energy transfer. The results from the calculations support a picture where excitation at 490 nm populates primarily a mixed neoxanthin S(2) Chl b Soret state. From this state excitation from neoxanthin is transferred to iso-energetic Chl b Soret states or via internal conversion to S(1) Chl b Q(y) states. From the Soret states excitation proceeds via internal conversion to Q(y) states of Chl b's mostly on the lumenal side. A rapid Chl b to Chl a transfer and subsequent transfer to the stromal side Chl a's and to the final state completes the process after 490 nm excitation. The interpretation is further supported by the fact that excitation energy transfer kinetics after excitation of neoxanthin at 490 nm and the Chl b Q(y) band at 647 nm (Linnanto et al., Photosynth Res 87:267-279, 2006) are very similar.

The present review describes the use of quantum chemical methods in estimation of structures and electronic transition energies of photosynthetic pigments in vacuum, in solution and imbedded in proteins. Monomeric Mg-porphyrins, chlorophylls and bacteriochlorophylls and their solvent 1:1 and 1:2 complexes were studied. Calculations were performed for Mg-porphyrin, Mg-chlorin, Mg-bacteriochlorin, mesochlorophyll a, chlorophylls a, b, c(1), c(2), c(3), d and bacteriochlorophylls a, b, c, d, e, f, g, h, plus several homologues. Geometries were optimised with PM3, PM3/CISD, PM5, ab initio HF (6-31G*/6-311G**) and density functional B3LYP (6-31G*/6-311G**) methods. Spectroscopic transition energies were calculated with ZINDO/S CIS, PM3 CIS, PM3 CISD, ab initio CIS, time-dependent HF and time-dependent B3LYP methods. Estimates for experimental transition energies were obtained from linear correlations of the calculated transition energies of 1:1 solvent complexes against experimentally recorded solution energies (scaling). According to the calculations in five-coordinated solvent complexes the magnesium atom lies out of the porphyrin plane, while in six-coordinated complexes the porphyrin is nearly planar. Charge densities on magnesium and nitrogen atoms were strongly dependent on the computational method deployed. Several dark states of low oscillator strength below the main Soret band were predicted for solvent complexes and chlorophylls and bacteriochlorophylls in protein environment. Such states, though not yet identified experimentally, might serve as intermediate states for excitation energy transfer in photosynthetic complexes. Q(y), Q(x) and Soret transition energies were found to depend on the orientation of the acetyl group and external pressure. A method to estimate site energies and dimeric interaction energies and to simulate absorption and CD spectra of photosynthetic complexes is described. Simulations for the light harvesting complexes Rhodospirillum molischianum, chlorosomes of Chlorobium tepidum and Chloroflexus aurantiacus, and LHC-II of Spinacia oleracea are presented as examples.

The semiempirical PM5 method has been used to calculate fully optimized structures of magnesium-bacteriochlorin, magnesium-chlorin, magnesium-porphin, mesochlorophyll a, chlorophylls a, b, c(1), c(2), c(3), and d, and bacteriochlorophylls a, b, c, d, e, f, g, and h with all homologous structures. Hartree-Fock/6-31G* ab initio and density functional B3LYP/6-31G* methods were used to optimize structures of methyl chlorophyllide a, chlorophyll c(1), and methyl bacteriochlorophyllides a and c for comparison. Spectroscopic transition energies of the chromophores and their 1:1 or 1:2 solvent complexes were calculated with the Zindo/S CIS method. The self-consistent reaction field model was used to estimate solvent shifts. The PM5 calculations predict planar structure of the porphyrin ring and central position of the four coordinated magnesium atoms in all pigments studied, in accord with the experimental, ab initio, and density functional results, a significant improvement as compared to the older semiempirical PM3 approach. Only small differences in PM5 and B3LYP/6-31G* or Hartree-Fock/6-31G* minimum energy geometries of the reference molecules were observed. Calculations show that in 1:1 solvent complexes, where the magnesium atom is five coordinated, the magnesium atom is shifted out of the plane of the porphyrin ring towards the solvent molecule, while the hexa coordinated 1:2 complexes are again planar. The PM5 method gives atomic charges that are comparable with those obtained from the Hartree-Fock/6-31G* and B3LYP/6-31G* calculations. The single point ZINDO/S CIS calculations with PM5 minimum energy structure gave excellent correlations between calculated and experimental transition energies of the chlorophylls and bacteriochlorophylls studied. Such correlations may be used for prediction of transition energies of the chromophores in protein binding sites. Calculations also predict existence of dark electronic states below the main Soret absorption band in all chromophores studied. The results suggest that the semiempirical PM5 method is a fairly reliable and computationally efficient method in predicting molecular parameters of porphyrin-like molecules., (Copyright 2003 Wiley Periodicals, Inc. J Comput Chem 1: 123-137, 2004)