| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Biological Sciences, Brock University, St. Catharines, Ontario L2S 3A1, Canada
Correspondence: Address reprint requests to Sergei Vasil'ev, Dept. of Biological Sciences, Brock University, St. Catharines, ON, Canada. Tel.: 905-688-5550 ext. 4163; Fax: 905-688-1855; E-mail: svassili{at}brocku.ca.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
–4). The reaction-center chromophores are organized into two symmetry-related branches that spread from a shared special pair of chlorins that form the primary electron donor. This common structural motif suggests a unique evolutionary origin.
Oxygenic photosynthesis uses two different reaction centers, photosystems II and I (PSII and PSI), to oxidize water and reduce NADP+ to NADPH. It is the most evolutionarily successful variation of photosynthesis and the major contributor to planetary photosynthetic production. PSII is the site of water oxidation and plastoquinone reduction. As the source of planetary oxygen, PSII has had a profound impact on the evolution of life (see Barber (5) and Diner and Rappaport (6) for recent reviews on PSII).
The efficiency of oxygenic photosynthesis is enhanced by light-harvesting antenna pigments that are divided into the peripheral antenna and core antenna. The peripheral antenna systems are diverse and often variable in number, whereas the core systems are constant and more closely associated with the reaction centers. In PSI, 
100 core-antenna chlorophylls (Chls) are bound by two structurally homologous polypeptides, PsaA and PsaB, which also hold the photochemically active reaction-center chromophores (7–10). In PSII, 30 core-antenna Chls are held by two smaller polypeptides, CP47 and CP43, which are bound to the D1 and D2 reaction-center polypeptides (11–13). Interestingly, CP47 and CP43 share some sequence homology and a folding motif with the antenna binding regions of PsaA and PsaB (3,4,7,8,10). We previously showed that 26 of the core-antenna Chls are structurally conserved between PSII and PSI (14). Of further evolutionary significance is the fact that the core-antenna binding regions of PsaA and PsaB also share a number of conserved chlorophyll liganding His residues with the ancient green sulfur bacteria and Heliobacteria (15,16). It is interesting that all reaction centers possessing core-antenna pigments not only share a common protein structural motif for the reaction-center chromophore-binding regions but also for the core-antenna binding regions. The conservation of this large and relatively complex protein structural motif underlines its importance in both reaction-center and core-antenna function.
X-ray structures of PSI and PSII have had a massive impact on our understanding of primary processes of photosynthesis and the structural similarities between functionally diverse reaction centers. Chromophore positions, from the x-ray structures, have provided the distance and relative orientation information required to determine pathways and efficiencies of excitation energy transfer in both PSII (17,18) and PSI (19,20). The x-ray structures of photosystems have further allowed detailed calculations of excited-state energy levels of individual antenna and reaction-center chromophores (21) and redox potentials of electron-transport cofactors (22,23).
As x-ray structures are generated from crystallized proteins at cryogenic temperatures, they represent an average static structure. At physiological temperatures, proteins exhibit significant thermally induced fluctuations in conformation. All of the factors (distance, relative orientation, and energy levels) that affect energy transfer between photosynthetic pigments are influenced by the thermal dynamics of the proteins holding the pigments. Dynamic changes in reaction-center chromophore conformation, separation, and interaction with the protein will also influence electron-transport efficiency. How do the real-time protein dynamics of a reaction center affect the efficiency of photosynthesis?
Molecular dynamics (MD) simulations provide information concerning the thermal movements of proteins and chromophores. They offer molecular trajectories of motion that have the potential to far surpass the amount of information contained in static x-ray structures. We have applied MD simulations to the PSII core complex in its native membrane environment and generated molecular trajectories with a duration of nanoseconds. We have used the time-dependent structures as the basis for kinetic models of excited-state dynamics to see how protein dynamics affects the efficiency of energy transfer within PSII. In addition, we have done quantum mechanical (QM) calculations of the excited states of individual chromophores, based on the molecular trajectories, to determine the site energy or "color" of each chromophore in PSII, as well as how that color changes in time. We have also used the MD simulations of PSII to see how changes in reaction-center chromophore separation control the rate of electron transport. How accurate are the structures and dynamics that result from our molecular dynamics simulations? We found that the MD/QM simulations accurately predicted the experimental absorbance of the PSII core complex, whereas QM calculations based on the x-ray structure did not. We also found that the MD/QM calculations correctly predicted the energy levels of the chromophores with known energy levels in the PSII core. Our results have allowed us to color the PSII core chromophores and study how protein dynamics influences both structural and spectral factors that affect energy transfer and electron-transport efficiency.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
) in conjunction with the VMD visualization program (25), and our own code have been used to set up the simulation system. For MD and energy minimization procedures we used the NAMD package (24). The simulations were performed with periodic boundary conditions in an NPT ensemble (Langevin piston coupling) at 300 K and 1 atm pressure. The long-range electrostatic interactions were calculated using the particle mesh Ewald algorithm. Both the electrostatic and the Lennard-Jones interactions had a twin-range cutoff of 8/12 Å. The multiple time step integration using an impulse-based Verlet-I method with time steps of 4 fs for long-range electrostatic forces, 2 fs for short-range nonbonded forces, and 1 fs for bonded forces was used. MD simulations were performed on the Brock Beowulf computer cluster (http://beowulf.ac.brocku.ca/beowulf/); 10 days of central processing unit time was required for a 1-ns-long simulation using 20 computing nodes.
Force fields
The all-atom AMBER-1994 force field (26) was used for protein and the GLYCAM-2000a (27) force field for headgroups of galactolipids. For chlorophyll, pheophytin, plastoquinone, and ß-carotene, we used the ab initio force field developed for the cofactors of bacterial photosynthesis (28), which was parameterized to reproduce density functional theory vibrational modes. We have modified this force field for cofactors of PSII as described below.
First we modeled the electrostatic potential field of PSII cofactors. This was accomplished by assigning partial charges to each atom. To derive atomic charges, molecular structures were first energy-minimized at a density functional level of theory with the gradient-corrected approximation of Perdew and Becke (29,30). This approximation is known to provide an accurate description of minimized geometries. To maintain consistency with the AMBER-1994 force field, we used the 6-31G* basis set. Ab initio electrostatic potential was then obtained on a set of four shells of surfaces at 1.4, 1.6, 1.8, and 2.0 times the van der Waals radii around the molecule according to a charge-fitting technique developed by the AMBER group (31). Partial atomic charges were fit to reproduce electrostatic potential on this set of grid points by a restrained electrostatic potential fit, as implemented in the RESP program from AMBER8 (UCSF, San Francisco, CA) suite. The GAUSSIAN-98 (Gaussian, Wallingford, CT) package was used for all quantum chemical calculations. Atomic charges of the oxygen evolving manganese cluster were assigned according to the redox state of its atoms in the dark-adapted (S1) state as follows: Mn1–Mn3, +3; Mn4, +2; O1–O4, –2; Ca, +2. The atomic charge of nonheme iron was set to +2.
We then ensured that equilibrium geometries of the cofactors were accurately reproduced by the force field. Pheophytin a and chlorophyll a are only slightly different from their bacterial analogs: the imidazole ring II is unsaturated and the carbonyl group attached to ring I is replaced with a -CH group in chlorophyll a and pheophytin a This modification was accomplished by selecting proper atom types from the existing force fields. Minor modification of the plastoquinone headgroup was done using existing atom types as well. We were unable to reproduce equilibrium geometry of the conjugated double-bond system of ß-carotene in a satisfactory manner using only existing atom types. To describe the equilibrium geometry of ß-carotene accurately we introduced several new carbon atom types and refined force field parameters (bond lengths, angles, and dihedrals) of the bonds involving these new atom types.
Calculation of chromophore excited-state energies
Transition energies of the lowest excited state QY of each chromophore were calculated ab initio using the single excitation configuration interaction (CIS) method. To account for the chromophore environment, all charges within 12 Å of any of the four Chl nitrogen atoms were included in the excited-state calculations (32,33). We used the STO-3G basis set for this calculation. This basis set is too small for accurate calculation of the absolute excited-state energies and produces excitation energies 0.5 eV higher than experimentally determined. Nevertheless, this small basis set allowed us to complete calculations in a reasonable computation time and it has previously been shown to correctly reproduce both relative site energies and their fluctuations (32). After the CIS calculation all excited-state energies were shifted by –0.565 eV to match the experimental room temperature peak absorbance wavelength of the PSII core complex (674 nm).
Simulation of the excited-state dynamics
A model for the light-harvesting system can be briefly described in terms of an effective Hamiltonian with diagonal elements denoting the site energies of individual chromophores and off-diagonal elements denoting the couplings between them:
 |
The couplings were computed in the point dipolar approximation according to the equation:
 |
Here 
denotes the unit vector along the QY transition of pigment i, 
is a vector connecting origins of transition dipoles i and j, 
= 32 D2 is the in situ transition dipole strength of chlorophyll a in media with a refractive index of 1.5 (34,35).
According to Förster theory, the transfer rate Tij between pigments i and j can be obtained from the elements of the Hamiltonian:
 |
is a spectral overlap between donor emission spectrum and acceptor absorption spectrum. A Stokes shift Si = 118 cm–1 was calculated from the absorption spectrum of chlorophyll (36) using the Kennard-Stepanov relation to ensure detailed balance (37).
Excitation transfer rates in the system of coupled chromophores are described by the rate matrix K which is related to the transfer matrix T
 |
To account for reversibility of the charge separation in PSII, two states representing the primary radical pair P680+Pheo– (RP) and QA were added to the rate matrix K. The transfer rates between RP, QA, and the rest of the pigments were calculated as follows:
If i = RP, 
; if j = RP, 
; if i = QA, 
and if j = RP, 
Solution of this system of differential equations requires knowledge of the rates of several photophysical processes in PSII: charge separation between primary electron donor and Pheo, KPC, its reversal 
electron transfer from Pheo– to QA, KST, and the rate of intrinsic Chl decay by fluorescence, Kdiss. Values of KST, Kdiss, and KPC are known from the literature. According to the experimental estimates, we used the following values of these rate constants: Kdiss = 0.5 ns–1 (38), KST = 2.5 ns–1 (39,40), and KPC = 1000 ns–1 (41,42), The rate of charge recombination 
= 5 ns–1 was obtained by fitting excited-state decay kinetics predicted by the model to the experimental fluorescence-decay kinetics of PSII core complex as described in Vasil'ev et al. (17,43).
The evolution of the excitation probability in time is governed by a master equation:
 |
Here 
is the vector of occupation probabilities for the excited pigments in the system and K is the rate matrix. This has the solution
 |
 |
Here, V is the matrix whose columns are the eigenvectors of K, and 
m is the eigenvalue of K associated with the column m of V.
Calculation of electron-transfer rates
Dutton and colleagues (44) have produced an empirical equation known as "Dutton's Ruler", relating electron-transfer rate to distance between cofactors:
 |
The initial constant 13 is the rate at van der Waals contact distance (R = 3.6 Å). The second term describes an approximately exponential fall-off in electron tunneling rate with distance through the insulating barrier. R is the edge-to-edge distance. The third term is the quantized Frank-Condon factor at room temperature. 
G is free energy and is the reorganization energy. To obtain an optimal value of the electron-transfer rate according to Marcus theory (45) we assumed that G = –.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
The x-ray structure (13) did not identify all molecules in the PSII core complex. There was a large space with unassigned electron density between the reaction center (RC) and CP43 and a smaller one between the RC and CP47. The inner surface of these cavities consists predominantly of hydrophobic atoms. Two lipid molecules are present in analogous locations between RC and antenna domains in photosystem I (9), which is structurally related to photosystem II. Most likely these cavities are occupied with aliphatic fatty acid chains of lipid molecules in PSII as well, which is why we chose to fill it with lipids. This was done interactively using NAMD (24) and VMD (25). Eight lipid molecules were required to fill the space between CP43 and the RC and two lipid molecules were fit between CP47 and the RC. In total we used six galactolipid molecules and four phosphatidylglycerol molecules. Use of improper molecules could affect equilibrium structure and dynamics of the simulated system. Placement of the different charged groups could also affect excited states of neighboring chromophores. However, the new, higher-resolution PSII structure (2AXT entry in PDB) published after the original submission of our manuscript supports our assumption. This structure identified five lipid molecules between CP43 and the RC and two lipids between CP47 and the RC, in the same areas where we placed eight and two lipid molecules, correspondingly. There is still some unassigned electron density left in this area in the new PSII structure, consistent with the expected high disorder of the flexible fatty acid chains in the large cavity.
The whole PSII core complex was inserted into the pre-equilibrated lipid bilayer composed of monogalactosyldiacylglycerol, digalactosyldiacylglycerol, and phosphatidylglycerol with 18:3 and 16:0 fatty acids. The lipid bilayer components were chosen to represent the most abundant lipids and fatty acids found in photosystem II preparations in an approximately correct proportion (46).
The complete system, containing 573 lipids (253 monogalactosyldiacylglycerol, 160 phosphatidylglycerol, and 160 digalactosyldiacylglycerol), was neutralized by addition of 178 Na+ ions and hydrated with 41,580 water molecules. The total number of atoms was 236,161 and the size of the system was 18 x 16 x 11 nm. To start molecular dynamics simulations we first minimized the energy of the complete system. Then the temperature was gradually raised to 300 K and the system was equilibrated for 200 ps with protein and cofactors constrained to the initial positions. All atoms were then released and a second equilibration 1 ns long was carried out with constant pressure. At this stage, anisotropic pressure coupling was applied (all periodic box dimensions were allowed to fluctuate independently). After 0.5 ns, the periodic box reached its equilibrium dimensions. All energy components remained at constant levels during the last 0.5 ns of the equilibration run, indicating that equilibrium had been reached. A snapshot of the equilibrated simulation system is shown in Fig. 1.
|
|
Some structural reorganization of this kind is expected to occur when direct protein-protein interactions between crystallographic unit cells tightly packed in a solid phase at low temperature are replaced by protein-lipid and protein-solvent interactions at physiological temperature in a liquid phase. Thus, this reorganization may be explained by an equilibration process related to changing the protein environment from a crystal (starting point of the MD run) to a water-lipid biphasic mixture. However, we cannot exclude that it may be also related to the substitution of lipids for missing/unassigned electron densities.
Calculation of individual chromophore excitation energies and their fluctuations
Experimental studies have determined the excited-state energies of only a few PSII chromophores: Chl, coordinated by His-114 in CP47 (47), PSII RC Chls, and pheophytins (48–51). Site energy levels of the remaining antenna chromophores are difficult, if not impossible, to ascertain experimentally due to the spectral congestion arising from a large number of pigments with such similar energy levels. Theoretical assignments, based on the x-ray structure (9), have been done for PSI by semiempirical calculations of Chl excited-state energies (21). Similar work on PSII has not been feasible due to the lower resolution of the available x-ray structures (12,13).
Following the method used in Mercer et al. (32) and Damjanovic et al. (33), we calculated excited-state energies for individual chromophores in PSII using the coordinates from the x-ray structure 1S5L. This required adding protons to the structure and optimizing their positions with the molecular mechanics force field. Site energies were then calculated using the CIS excited-state calculation as described in Methods. Most of the off-diagonal elements of the system Hamiltonian were <100 cm–1; only three were within 100–160 cm–1. This is less than the experimentally determined homogeneous chlorophyll linewidth (160–180 cm–1) (36,52). Thus, the absorption spectrum of PSII can be well approximated by consideration of only diagonal elements of the Hamiltonian. Based on these findings we modeled the inhomogeneously broadened experimental absorption spectrum as a sum of Gaussians 170 cm–1 wide. The resulting spectrum is shown in Fig. 3. As seen from the figure, the spectrum calculated from the x-ray coordinates failed to reproduce the experimental spectrum and had a much broader bandwidth.
|
20 fs (data not shown). Similar high-frequency modes were observed previously for both Chl in solution (32) and BChl in light-harvesting protein (33). Most probably, this mode arises from intramolecular vibration.
Autocorrelation of energy-gap fluctuations is the key quantity allowing calculation of the absorption lineshape without artificial parameters. However, for convergence of absorption spectra, about twice as many quantum chemical computations of the excited state for each chromophore would be necessary (30). This number of computations was beyond our current computing resources. In our study, we simplified calculation of the absorption lineshape and used the same procedure described above for calculation of the absorption spectrum of the x-ray structure: we dressed QY states of individual chromophores by Gaussians of the proper width and summed them up. Due to fast fluctuations, single-point calculations of the energy level from one frame of the molecular trajectory do not provide meaningful information about site energy level. To obtain mean site energy levels we calculated an average of 240 single-point excited-state calculations done with a time increment of 5 fs. We found this number and time interval of calculations to be sufficient for convergence of the mean site energy levels (data not shown).
Fluctuations occurring on longer timescales cause variation of the peak positions of individual chromophores. To characterize the amount of inhomogeneous line broadening caused by these slower motions of the pigment-lipid-protein complex we repeated our mean site energy level calculations for four different times from the molecular dynamics trajectory, each separated by 1 ns. The resulting calculated absorption spectra of the PSII core complex are shown in Fig. 3. Despite the relatively large excited-state energy fluctuations of the contributing individual chlorophylls, there was little variation of shape and peak positions of the overall spectra calculated from the four different times sampled from the molecular dynamics trajectory. These four spectra, and their average, reproduced well the experimental absorption spectra of the PSII core complex. In these calculations we used only one parameter: the width of the Gaussian distribution function 
representing the lineshape of a single chromophore. From the fit of the simulated spectra to the experimental PSII spectrum we found = 168 cm–1. This value is close to the = 159 cm–1 of Chl a in diethyl ether (36) and = 170 cm–1 of Chl a in pyridine (52). Recent theoretical modeling of experimental spectroscopic data from PSII samples estimated of Chl to be 180 cm–1 (51).
Site energies for individual chromophores determined from the QM/MD calculations are shown in Fig. 4. Values are averages of those calculated at the four different time points in the MD simulation. Some chromophores exhibited larger variations between the four time slots than others. In addition, there were differences between the average site energies of the chromophores, some with higher and some with lower energies. Interestingly, the MD/QM calculation of excited states correctly predicted the average site energy levels of a number of PSII chromophores whose energies have been determined experimentally (see Discussion for details).
|
The antenna chlorophylls of CP43 and CP47 are separated from the reaction center core chromophores by 20 Å. At this distance, the rate of energy transfer from the antenna to the reaction center to some extent limits the overall rate of excited-state decay (18,43). The distance fluctuations between the antenna and the reaction center pigments are therefore expected to affect the experimentally observable fluorescence-decay kinetics. We analyzed fluctuations of distances between the geometrical centers of the reaction center pheophytins and their nearest antenna pigment neighbors. We chose these pigments because they represent one of the major pathways of excitation energy transfer from light-harvesting antenna to the reaction center (18,43). Fluctuations of energy transfer rates in these pairs of pigments will have the largest impact on the efficiency and excited state lifetime of the PSII core complex. The time dependence of these fluctuations is shown in Fig. 5. As seen from this figure, the magnitudes of distance fluctuations were 2.5 Å and 2.1 Å for the CP43 and CP47 sides of the PSII core complex, respectively. The distance between the D1 pheophytin and the bridging antenna Chl in CP43 is slightly larger and more variable than the distance between the D2 pheophytin and the bridging Chl on the CP47 side. Due to the 1/R6 dependence of excitation-energy transfer rate on distance, the distance fluctuations we observed resulted in a factor of 2 change of excitation-energy transfer rate. It is interesting to estimate how variation of the excitation transfer rates will affect the lifetime of the system along the molecular dynamics trajectory. To achieve this, we recomputed the off-diagonal entries of the Hamiltonian for each frame of the trajectory. Diagonal entries of the Hamiltonian were kept constant. This approach is expected to account for most of the fluctuations in excitation transfer, as the latter is dominated by the dipole-dipole couplings. However, excitation transfer is not completely insensitive to the fluctuations of the diagonal elements. Future refinement of the excited dynamics involving recomputation of energy levels along the molecular dynamics trajectory will characterize contribution of the diagonal elements to the overall lifetime variation. Lifetime calculations revealed significant variations in the calculated PSII fluorescence-decay lifetimes, Fig. 5 B. Along with the fast distance fluctuations, caused by the vibrational modes of the chromophores (picosecond time domain), slow distance changes (nanoseconds) arising from long-range motions of the whole system are clearly seen in this figure.
|
|
,53,54), indicating that the rate of this electron transport step is close to optimal. The molecular dynamics, however, revealed strong fluctuations of the rate of this electron-transfer step. At certain times, the two electron-transfer cofactors come so close to each other that the rate of electron transfer increased by a factor of 10 from its average value Fig. 7 B. The lowest and highest limits for KET observed during the 8-ns simulation run were 0.9 ns–1 and 53 ns–1. The thermal dynamics of the PSII core complex modulates the rate of electron transfer such that periods with a high average electron-transfer rate (8–10 ns–1) are alternating with periods with a low rate (1.5–2 ns–1). The system spends approximately equal time in conformations with slow and fast electron-transfer rate. The resulting distribution of KET (Fig. 7 B) does not resemble a bell-shaped curve; it has several different peaks, which is why the time averaged Pheo– oxidation kinetics are expected to be multi-phasic.
|
oxidation (55). Similar differences between optimal and experimental rates of the QA-to-QB electron-transfer step have been observed in purple bacteria (44). This difference is most likely explained by reorganization energy associated with this process. The protein dynamics modulates the rate of this electron-transfer step as well as the rate of Pheo– oxidation. However, the modulation pattern is quite different in the case of QA-to-QB transfer. Most of the time, the system is found in a conformation with QB in a distal position, characterized by a slow electron-transfer rate.
Implications of protein dynamics calculations on quantum efficiency
To assess how dynamic changes in chromophore geometry influence energy-transfer efficiency and thus affect the quantum yield of photochemistry in PSII, we used kinetic models for excitation energy transfer based on each of the structural configurations resulting from the MD simulation. Quantum yield was determined as the yield of reduced quinone electron acceptors in the kinetic models. Interestingly, the distribution of quantum yields resulting from changes in antenna configuration was very narrow (only 0.25%, data not shown). We found that site-energy fluctuations affected the quantum yield much more than the orientation fluctuations. The quantum yield calculated for systems with the spectral assignments determined for the four different times from the MD simulation as described above ranged from 89% to 93%. An even larger difference of 8% was found when the locations of the spectral forms were randomly shuffled in the PSII core complex (data not shown).
The electron-transfer rate from Pheo– to QA (KST) is critical for a high yield of photochemistry. With the parameters used in our kinetic model, fluctuations of the KST lead to a variation of quantum yield within 2%. This is a very conservative estimate, based on the optimal theoretical rate of KST. With the parameters used for the kinetic model in this study, the quantum yield reached its maximal value when KST was 2 ns–1; any further increase of KST did not improve the quantum yield. Our calculations showed that if KST was scaled down by a factor of 0.5 to bring the average KST in the kinetic model to the average of experimentally determined values, the width of the quantum yield distribution increased by a factor of 2.5 and became 5%. This is the full spread of the distribution and it does not mean that the average yield of PSII becomes significantly lower. The average yield only goes down from 96% (for KST determined by Dutton's ruler) to 95% (for KST scaled by 0.5 to match the average experimental rate).
Effect of the dynamic changes in energy and electron transport on experimental observables
Processes of energy transfer and electron transport contribute to the decay kinetics of the excited state, which can be observed by time-resolved fluorescence and absorption experiments. Experimental fluorescence-decay kinetics of the PSII core complex are multi-exponential (17,56,57). To date, the nature of excited-state decay complexity has not been fully understood. The radical-pair equilibrium model (39,58) predicts two exponential kinetics where the second decay component arises from recombination of the primary radical pair. It has been suggested that a complex charge stabilization process is responsible for multi-exponential decay kinetics (17,59). However, excited-state decay will be affected by the molecular dynamics of the system as chromophore positions, transition dipole orientations, site energy levels, and electron-transfer rates are dynamically changing. To what extent does thermal disorder modulate excitation decay kinetics? To address this question, we calculated the fluctuation of the mean fluorescence lifetime during the time course of the MD simulation.
At first we studied how the excited-state decay depends upon position and orientation of the antenna chlorophylls. To accomplish this, we extracted positions and dipole directions of all chromophores from every snapshot of the atomic trajectory saved every picosecond of the MD. Using this data we calculated fluctuations in the excitation lifetime. Only the distance and orientation factors in the Förster equation were changed in this case. The time dependence of the excitation lifetime is shown in Fig. 5 B. As seen from this figure, increases in distances between the linkers and the reaction center were correlated with longer excitation lifetimes. We observed changes of 15% in the lifetime during the time course of our simulation. However, the average fluorescence-decay kinetics, obtained by averaging the individual kinetics calculated for each snapshot of the atomic trajectory, was well described by two exponential decay phases and was indistinguishable from fluorescence-decay kinetics calculated from the original x-ray configuration of chromophores. Thus, we conclude that variations of the decay kinetics caused by positional and orientational disorder did not result in the appearance of new fluorescence-decay components.
We then studied how excited-state dynamics are modulated by fluctuations of electron-transfer rate. The rate of electron transfer from Pheo to QA defines the lifetime of the slow component of the excited-state decay. We therefore expect a large degree of fluorescence-decay inhomogeneity (appearance of several slow fluorescence-decay components with different lifetimes) to arise due to thermal fluctuation of the distance between Pheo and QA. Indeed, when we calculated the average fluorescence-decay kinetics using time-dependent KST resulting from the molecular dynamics simulation we found that four decay components (77% 64 ps, 17.5% 140 ps, 4% 300 ps, and 1.5% 614 ps) were required to describe fluorescence decay. Two additional slow-decay components appeared in the excited-state decay kinetics due to the fluctuation of the rate of electron transfer. These components are similar to the experimentally observable slow fluorescence-decay components (17,60). For this calculation we used the optimal KST. Even more complex fluorescence-decay kinetics were found when we scaled KST down to bring the average KST to the average of experimentally determined values as described above.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Fluctuations in chromophore geometry arising from the dynamical motions of their polypeptide ligands were found to contribute relatively little to variation in the efficiency of energy transfer and calculated quantum yield as determined with kinetic models. The PSII antenna system is robust and delivers virtually the same high efficiency despite relatively large fluctuations of energy transfer rates from linker Chls to the core. More variation in energy-trapping efficiency was introduced by variations in the excited-state energy levels of the antenna resulting from changing chromophore conformations and changing chromophore protein and/or solvent interactions. Individual chromophores showed a range of excited-state energy variations. Due to high computational cost, we were able to perform excited-state calculations only for the system at four different times during the molecular trajectory. This was not sufficient to observe the full extent of variation in excited-state site energies. Thus, our data represent a conservative estimate of variation in energy-trapping efficiency.
The absorption spectrum of Chl is clearly broadened by its interaction with protein ligands. This is most likely driven by a strong selection pressure to increase the distribution of antenna chromophore energy levels to increase the effective light-harvesting capability of the reaction center. Protein-induced antenna spectral broadening has often been presumed to arise predominantly from variations in individual chromophore site energies which arise from site-dependent differences in protein-chromophore interactions. Our MD simulation shows that variation in the overall distribution of antenna chromophore energy levels originates dynamically within individual sites as well as between different sites. Protein dynamics thus plays a significant role in increasing the bandwidth of Chl absorption.
Our calculations have correctly reproduced the absorption spectrum of the PSII core complex, confirmed the site energy levels of the RC chromophores recently derived analytically (51), and correctly identified the lowest-energy Chl in PSII (61). These results lend strong support to our MD/QM calculated assignments for the site energies of each individual PSII antenna chromophore. Our simulation identifies the lowest-energy chlorophyll in PSII as Chl coordinated by CP47 His-114. This is the same Chl that was found experimentally, by site-directed mutagenesis (61), to be the source of the long-wavelength fluorescence emission (695 nm) in PSII. Our calculations place the excited state of active (D1) pheophytin higher than the inactive (D2) pheophytin. At first glance, this is in contradiction to experimental assignments that have been made based upon spectroscopic changes induced by reduction of Pheo and/or QA. In that work, the energy levels of the D1/D2 pheophytins were found to lie at 681/670–672 nm (48) in isolated PSII reaction centers, at 685.6/669.3 nm in PSII core complexes of cyanobacteria Synechocystis (49), and at 685.3/670 nm in plant PSII core complexes (50). However, recent detailed calculations of PSII reaction-center optical properties based on its x-ray structure have shown that the site energy of D1 Pheo should actually be higher than the energy of the D2 Pheo to account for a wealth of spectroscopic data (51). These calculations place the D1 Pheo energy at 672 nm. The authors attribute the discrepancy in data interpretation to a mixture of excitonic and electrochromic effects that result in a bleaching at 681–685 nm upon reduction of D1 Pheo even though its site energy is at 672 nm (51). The results of our study are in agreement with the latter identification of site energies of pheophytins, as well as all other PSII RC chromophores (Table 1). Our results also confirmed similar site energies for both P680 monomers and ChlZ/D. The higher site energy for the accessory Chl D1 compared to its D2 analog was also reproduced.
|
,63). The conformational change observed in bacterial RC x-ray structures is similar to the conformational change between two QB binding-site conformations that emerged in our MD simulation of the PSII reaction-center core (Fig. 6). The headgroup of QB was displaced in the same direction and the aliphatic tail exhibited the same propeller-like motion observed in the bacterial x-ray structures. In our simulation, the system was in the state equivalent to the dark-adapted state (no charge on QA). We found QB to be in the distal binding pocket most of the time, corresponding to the dark-adapted bacterial reaction-center structure and only 10% of the time in the state resembling the light-adapted x-ray conformation. Compared to the conformation of the light-adapted bacterial x-ray structure, the proximal QB conformation observed in our MD simulation was intermediate between the light-adapted and dark-adapted x-ray structures. These findings strongly suggest the possibility of a light-induced conformational gating in the PSII reaction center, which may be a common feature for Type II photosynthetic reaction centers.
Our MD study has revealed relationships between protein dynamics, the configuration of the light-harvesting antenna system, and electronic properties of chromophores in the PSII core complex. This information, combined with kinetic modeling of the photochemical reactions, has allowed us to quantify the contribution of different dynamic changes to the efficiency of primary photosynthetic energy conversion. Our MD calculations thus offer novel insight into the evolution of a variety of processes, operating at a number of different levels that contribute to the photochemical efficiency of PSII. Selection pressures operating on a variety of parameters, sometimes in competition with each other, must be considered when looking at the optimization of reaction-center efficiency. For example, although efficient light harvesting requires a dense packing of chromophores, a minimum separation must be maintained to prevent aggregation and formation of quenching centers. Evolutionary solutions to this dilemma appear limited, as all reaction centers containing core-antenna chromophores share a homologous antenna binding polypeptide motif. We have previously shown that this common chromophore binding motif results in the majority of PSII antenna chromophores having highly positionally conserved counterparts in PSI (14). Interestingly, not all of these conserved chromophores are orientationally optimized for energy transfer. Clearly, other selection pressures are operating and other factors (i.e., protein folding constraints related to maximizing pigment density and minimizing the formation of quenching centers) are involved. Our MD/QM calculations of chromophore site energies show that protein-chromophore interactions influence the absorption of light by antenna chromophores. Protein dynamics increase the effective absorption bandwidth of all individual chromophores. From an evolutionary point of view, this will confer an advantage by increasing the potential of the antenna to capture light. In addition, individual sites have different average peak wavelengths, which means the localization of spectral forms must be considered when assessing photosynthetic efficiency. The localization and extent of dynamic variation of individual chromophore spectra will strongly influence the efficiency of light energy transfer from core antenna to the reaction center, between other auxiliary antenna systems and the core antenna, between monomers in PSII dimers, and even between PSII and PSI. This competing array of selection pressures still does not include explicit consideration of any of the photoregulatory mechanisms known to affect light harvesting by PSII. The most significant effect on photosynthetic efficiency we observed in our MD calculations arose from the movements of the two quinone electron acceptors, QA and QB. The positional changes we observed in QB compared to those previously observed between dark-adapted and light-adapted bacterial reaction centers suggests that conformational changes affecting electron transport efficiency triggered by light may be a common factor in Type II reaction centers (62). In summary, our MD calculations show a strong influence of protein dynamics on energy transfer and electron-transport efficiency in the PSII core. Our work also provides an additional framework for understanding the evolution of photosynthetic reaction centers and underlines the importance of considering protein dynamics in reaction-center structure and function.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Submitted on October 7, 2005; accepted for publication January 18, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
2. Barber, J., and B. Andersson. 1994. Revealing the blueprint of photosynthesis. Nature. 370:31–34.[CrossRef]
3. Rhee, K. H., E. P. Morris, J. Barber, and W. Kuhlbrandt. 1998. Three-dimensional structure of the plant photosystem II reaction centre at 8 Å resolution. Nature. 396:283–286.[CrossRef][Medline]
4. Schubert, W.-D., O. Klukas, W. Saenger, H.-T. Witt, P. Fromme, and N. Krauß. 1998. A common ancestor for oxygenic and anoxygenic photosynthetic systems: a comparison based on the structural model pf photosystem I. J. Mol. Biol. 280:297–314.[CrossRef][Medline]
5. Barber, J. 2003. Photosystem II: the engine of life. Q. Rev. Biophys. 36:71–89.[CrossRef][Medline]
6. Diner, B. A., and F. Rappaport. 2003. Structure, dynamics, and energetics of the primary photochemistry of photosystem II of oxygenic photosynthesis. Annu. Rev. Plant Biol. 53:551–580.[CrossRef]
7. Krauß, N., W.-D. Schubert, O. Klukas, P. Fromme, H.-T. Witt, and W. Saenger. 1996. Photosystem I at 4 Å resolution represents the first structural model of a joint photosynthetic reaction centre and core antenna system. Nat. Struct. Biol. 3:965–973.[CrossRef][Medline]
8. Klukas, O., W.-D. Schubert, P. Jordan, N. Krauß, P. Fromme, H.-T. Witt, and W. Saenger. 1999. Photosytem I, an improved model of the stromal subunits PsaC, PsaD, and PsaE. J. Biol. Chem. 274:7351–7360.
9. Jordan, P., P. Fromme, H.-T. Witt, O. Klukas, W. Saenger, and N. Krauß. 2001. Three-dimensional structure of cyanobacterial photosystem I at 2.5 Å resolution. Nature. 411:909–917.[CrossRef][Medline]
10. Ben Shem, A., F. Frolow, and N. Nelson. 2003. Crystal structure of plant photosystem I. Nature. 426:630–635.[CrossRef][Medline]
11. Zouni, A., H.-T. Witt, J. Kern, P. Fromme, N. Krauß, W. Saenger, and P. Orth. 2001. Crystal structure of oxygen evolving photosystem II from Synechococcus elongatus at 3.8 Å resolution. Nature. 409:739–743.[CrossRef][Medline]
12. Kamiya, N., and J.-R. Shen. 2003. Crystal structure of oxygen-evolving photosystem II from Thermosynechococcus vulcanus at 3.7-Å resolution. Proc. Natl. Acad. Sci. USA. 100:98–103.
13. Ferreira, K. N., T. M. Iverson, K. Maghlaoui, J. Barber, and S. Iwata. 2004. Architecture of the photosynthetic oxygen-evolving center. Science. 303:1831–1838.
14. Vasil'ev, S., and D. Bruce. 2004. Optimization and evolution of light harvesting in photosynthesis: the role of antenna chlorophyll conserved between photosystem II and photosystem I. Plant Cell. 16:3059–3068.
15. Fyfe, P. K., M. R. Jones, and P. Heathcote. 2002. Insights into the evolution of the antenna domains of Type-I and Type-II photosynthetic reaction centres through homology modelling. FEBS Lett. 530:117–123.[CrossRef][Medline]
16. Heathcote, P., P. K. Fyfe, and M. R. Jones. 2002. Reaction centres: the structure and evolution of biological solar power. Trends Biochem. Sci. 27:79–87.[CrossRef][Medline]
17. Vasil'ev, S., C.-I. Lee, G. W. Brudvig, and D. Bruce. 2002. Structure-based kinetic modeling of excited-state transfer and trapping in His-tagged PSII core complexes from Synechocystis. Biochemistry. 41:12236–12243.[CrossRef][Medline]
18. Vasil'ev, S., J.-R. Shen, N. Kamiya, and D. Bruce. 2004. The orientations of core antenna chlorophylls in photosystem II are optimized to maximize the quantum yield of photosynthesis. FEBS Lett. 561:111–116.[CrossRef][Medline]
19. Sener, M. K., D. Y. Lu, T. Ritz, S. Park, P. Fromme, and K. Schulten. 2002. Robustness and optimality of light harvesting in cyanobacterial photosystem I. J. Phys. Chem. B. 106:7948–7960.
20. Byrdin, M., P. Jordan, N. Krauss, P. Fromme, D. Stehlik, and E. Schlodder. 2002. Light harvesting in photosystem I: Modeling based on the 2.5-angstrom structure of photosystem I from Synechococcus elongatus. Biophys. J. 83:433–457.
21. Damjanovic, A., H. M. Vaswani, G. R. Fleming, and P. Fromme. 2002. Chlorophyll excitations in Photosystem I of Synechoccus elongatus. J. Phys. Chem. B. 106:10251–10262.
22. Hasegawa, K., and T. Noguchi. 2005. Density functional theory calculations on the dielectric constant dependence of the oxidation potential of chlorophyll: implication for the high potential of P680 in photosystem II. Biochemistry. 44:8865–8872.[CrossRef][Medline]
23. Ishikita, H., B. Loll, J. Biesiadka, W. Saenger, and E. W. Knapp. 2005. Redox potentials of chlorophylls in the photosystem II reaction center. Biochemistry. 44:4118–4124.[CrossRef][Medline]
24. Kale, L., R. Skeel, M. Bhandarkar, R. Brunner, A. Gursoy, N. Krawetz, J. Phillips, A. Shinozaki, K. Varadarajan, and K. Schulten. 1999. NAMD2: greater scalability for parallel molecular dynamics. J. Comput. Phys. 151:283–312.[CrossRef]
25. Humphrey, W., A. Dalke, and K. Schulten. 1996. VMD: visual molecular dynamics. J. Mol. Graph. 14:33–38.[CrossRef][Medline]
26. Cornell, W. D., P. Cieplak, C. I. Bayly, I. R. Gould, K. M. Merz, D. M. Ferguson, D. C. Spellmeyer, T. Fox, J. W. Caldwell, and P. A. Kollman. 1995. A second generation force field for the simulation of proteins, nucleic acids, and organic molecules. J. Am. Chem. Soc. 117:5179–5197.[CrossRef]
27. Basma, M., S. Sundara, D. Calgan, T. Vernali, and R. J. Woods. 2001. Solvated ensemble averaging in the calculation of partial atomic charges. J. Comput. Chem. 22:1125–1137.[CrossRef][Medline]
28. Ceccarelli, M., P. Procacci, and M. Marchi. 2003. An ab initio force field for the cofactors of bacterial photosynthesis. J. Comput. Chem. 24:129–142.[CrossRef][Medline]
29. Perdew, J. P. 1986. Density-functional approximation for the correlation-energy of the iInhomogeneous electron-gas. Phys. Rev. B. 33:8822–8824.[CrossRef]
30. Becke, A. D. 1986. Density functional calculations of molecular-bond energies. J. Chem. Phys. 84:4524–4529.[CrossRef]
31. Bayly, C. I., P. Cieplak, W. D. Cornell, and P. A. Kollman. 1993. A well-behaved electrostatic potential based method using charge restraints for deriving atomic charges: the RESP model. J. Phys. Chem. 97:10269–10280.[CrossRef]
32. Mercer, I. P., I. R. Gould, and D. Klug. 1999. A quantum mechanical/molecular mechanical approach to relaxation dynamics: calculation of the optical properties of solvated bacteriochlorophyll a. J. Phys. Chem. B. 103:7720–7727.
33. Damjanovic, A., I. Kosztin, U. Kleinekathofer, and K. Schulten. 2002. Excitons in a photosynthetic light-harvesting system: a combined molecular dynamics, quantum chemistry, and polaron model study. Phys. Rev. E. 65:031919.
34. Knox, R. S. 2003. Dipole and oscillator strengths of chromophores in solution. Photochem. Photobiol. 77:492–496.[CrossRef][Medline]
35. Knox, R. S., and B. Q. Spring. 2003. Dipole strengths in the chlorophylls. Photochem. Photobiol. 77:497–501.[CrossRef][Medline]
36. Shipman, L. L., T. M. Cotton, J. R. Norris, and J. J. Katz. 1976. An analysis of the visible absorption spectrum of chlorophyll-a monomer, dimer, and oligomers in solution. J. Am. Chem. Soc. 98:8222–8230.[CrossRef][Medline]
37. Laible, P. D., R. S. Knox, and T. G. Owens. 1998. Detailed balance in Forster-Dexter excitation transfer and its application to photosynthesis. J. Phys. Chem. B. 102:1641–1648.
38. Roelofs, T. A., C.-H. Lee, and A. R. Holzwarth. 1992. Global target analysis of picosecond chlorophyll fluorescence kinetics from pea chloroplasts. Biophys. J. 61:1147–1163.
39. Schatz, G. H., H. Brock, and A. R. Holzwarth. 1988. Kinetic and energetic model for the primary processes in photosystem II. Biophys. J. 54:397–405.
40. Leibl, W., J. Breton, J. Deprez, and H.-W. Trissl. 1989. Photoelectric study on the kinetics of trapping and charge stabilization in oriented PSII membranes. Photosynth. Res. 22:257–275.[CrossRef]
41. Gobets, B., I. H. M. van Stokkum, F. van Mourik, J. P. Dekker, and R. van Grondelle. 2003. Excitation wavelength dependence of the fluorescence kinetics in photosystem I particles from Synechocystis PCC 6803 and Synechococcus elongatus. Biophys. J. 85:3883–3898.
42. Vasil'ev, S., G. W. Brudvig, and D. Bruce. 2003. The X-ray structure of photosystem II reveals a novel electron transport pathway between P680, cytochrome b(559) and the energy-quenching cation, Chl(Z)(+). FEBS Lett. 543:159–163.[CrossRef][Medline]
43. Vasil'ev, S., P. Orth, A. Zouni, T. G. Owens, and D. Bruce. 2001. Excited-state dynamics in photosystem II: insights from the x-ray crystal structure. Proc. Natl. Acad. Sci. USA. 98:8602–8607.
44. Moser, C. C., C. C. Page, R. Farid, and P. L. Dutton. 1995. Biological electron-transfer. J. Bioenerg. Biomembr. 27:263–274.[CrossRef][Medline]
45. Marcus, R. A. 1996. Electron transfer reactions in chemistry. Theory and experiment. In Protein Electron Transfer. D. S. Bendall, editor. Bios Scientific, Oxford, UK.
46. Murata, N., S.-I. Higashi, and Y. Fujimura. 1990. Glycerolipids in various preparations of photosystem II from spinach chloroplasts. Biochim. Biophys. Acta. 1019:261–269.
