Plants and photosynthetic bacteria convert solar
energy into chemical energy with a quantum yield of almost 100 %. The initial
stage of the conversion occurs at photosynthetic reaction center (PRC),
where the system transfers electrons generated by solar light toward a
determined direction along a determined route with a high speed of 4 ps
for 20 A at room temperature. Since Deisenhofer et al. reported the structural
data of Rhodopseudomonas viridis in 1984 , many studies have treated
the electron transfer (ET) reaction, and left several problems unsolved;
the most important issue is unidirectonality of charge separation.
At the PRC, the protein consists of C-, L-, M-,
and H-subunits, and the pigments include a pair of bacteriochlorophyll
b dimer (P), two molecules of accessory bacteriochlorophyll b (B), two
molecules of bacteriopheophytin b (H), and two quinone (Q) molecules. We
call the pigment line as Qm-Hm-Bm-P-Bl-Hl-Ql.
The Qm stands Q in the M-subunit, and the other definitions
are also same manner. Electrons go through from P to Ql exclusively
in spite of both possibility of L- and M-subunit form consideration on
only pigments. The ratio is about 100:1 (P to Hl is preferable).
Because this mechanism suppresses the electron recombination at the radical
pair of the PRC, this system can convert solar light to chemical energy
very efficiently. Many researchers consider protein controls the ET, and
we have investigated this point from theoretical point of view .
In order to study the ET problem, we adopted charge
equilibration method (QEq) by Rappe and Goddard  to calculate atomic
partial charges depending on the structure, and generalized this method
(charge separation) . These theories solve non- symmetric and dense
linear equation with a dimension of total numbers of atoms for a few iterations,
and determine the partial charges so that we can reproduce quantum chemistry
calculation for the system.
Although the PRC leads us to the linear equation
with a dimension of over twenty thousands, we
can solve this problem with less than 2 minutes per one iteration on HITACHI
SR8000 16 node (1 node is 8 GFLOPS, 8G) of Tsukuba Advanced Computing
Center, Agency of Industrial Science and Technology.
We constructed PRC structure based on X-ray structure
data , and calculated electrostatic energy of PRC by QEq before ET and
QEq-CS after ET. The initial state of ET the photoexcited P, P*, and this
state is higher by 1.25 eV than the ground P state . As summarized in
Table 1, all four ET (P* to Bl, Bm, Hl,
and Hm) are exothermic and the P* to Hl is most favorable.
Because energy expression of QEq and QEq-CS is a sum of one- and two-bodies
for atomistic suffixes, this table includes both terms. Although the sudden
photoexcited ET makes the charge separation system unstable via the one-body
term, the two-body term stabilized the charge separation system. The traditional
simulations use the fixed charges during ET, they cannot express the stabilization
by the two-body term.
In the QEq-CS method, we define both donor and
acceptor regions. We calculated the difference of partial charge along
each ET and investigated the contents for each case. When the donor region
loses one electron, the neighbor region increases the electron density
a little bit. At the same time, the acceptor region gains an electron,
and the neighbor region decreases the electron density. These reorganizations
depend on the specific position of several atoms within each region; our
method can describe the realistic change. On the bases of QEq and QEq-CS,
we checked the partial charge distribution of the PRC. These noticed us
minus charge distribution just above P, and plus charge distribution beside
Hl. With this contents, we summarized the minus charge distribution
pushes electrons, and the plus charge distribution pulls the electrons,
and proposed the "push-pull mechanism" . We consider
this mechanism converts solar light to chemical energy with a high efficiency
at the earliest stage of photosynthetic reaction, and gives some ideas
to artificial photosynthesis materials.
All values are based on the energies from P* to All (except
We have done this study under "COE project on Photoreaction
Control and Photofunctional Materials" and "Promote Research Projects for
High Performance Computing, Tsukuba Advanced Computing Center, Agency of
Industrial Science and Technology". We thank Prof. William A. Goddard for
giving O.K. an original QEq code, and Drs. Osamu Tatebe, Satoshi Sekiguchi,
Akira Okada, and Nobuaki Miura for valuable comments.
|Table 1. Change of electrostatic
energy along each ET (at X-ray structure; in eV)
|P* to Bl
|P* to Bm
|P* to Hl
|P* to Hm
 J.Deisenhofer, O.Epp, K.Miki, R.Huber, and H.Michel, J.Mol.Biol.,
180, 385 (1984).
 O.Kitao, H.Ushiyama, and N.Miura, J.Chem.Phys. , 110,2936(1999).
 A.K.Rappe and W.A.Goddard III, J.Phys.Chem., 95,3358(1991).
 O. Kitao, N.Miura, and H.Ushiyama, Theochem, 461-462,239-247(1999).
 O.Kitao and K.Aoki, submitted for Japanese Journal of Chem.Info.
and Comp. Science.
 J.Breton, Biochim.Biophys.Acta, 810,235(1985).
 O.Kitao, K.Aoki, and T.Ogawa, to be submitted.