1 Introduction

In anoxygenic and oxygenic photosynthesis, the absorption of light results in the conversion of the light energy into chemical energy through a series of electron and proton transfer reactions in pigment–protein complexes. In anoxygenic photosynthetic bacteria, light energy is initially absorbed by antenna complexes and the energy is transferred to the reaction center, where the primary photochemistry occurs, namely, the creation of a charge-separated state [1, 2]. In purple bacteria, absorption of light by the reaction center results in excitation of a bacteriochlorophyll (BChl) dimer followed by the transfer of an electron through a series of electron acceptors, a BChl monomer, a bacteriopheophytin (BPhe) monomer, and the primary quinone, until the secondary quinone is reduced. After reduction of the oxidized BChl dimer by an exogenous cytochrome c 2 or a bound tetraheme cytochrome, light can be absorbed again, leading to the transfer of a second electron to the secondary quinone in a process that is coupled to the uptake of two protons. After the second electron transfer, the quinol carries the electrons and protons to the cytochrome bc 1 complex in a cycle that generates the proton gradients needed for the creation of energy-rich compounds.

In oxygenic photosynthesis, two pigment–protein complexes, photosystem I and II, absorb light energy and perform electron transfer reactions [3]. Unlike the cyclic pathway found in purple bacteria, cyanobacteria, algae, and plants make use of the Z scheme, which has terminal electron donors and acceptors. In this scheme, light excitation of photosystem II results in oxidation of the primary electron donor, P680, which is reduced by a redox active tyrosine, YZ, followed by reduction of YZ •+ by the site of water oxidation, namely, the Mn4Ca cluster. As described by the S cycle, after four photons of light have been absorbed and four electrons have been transferred, two water molecules bound to the Mn4Ca cluster are converted into molecular oxygen. The use of water as an electron donor in this four-electron, four-proton process places special constraints on the energetics of the manganese cluster. In this review, the energetics of both the bacterial reaction center and photosystem II are discussed, with an emphasis on the differences and similarities of the primary electron donors. Also discussed are efforts to modify the bacterial reaction center such that it gains functional cofactors corresponding to YZ and the Mn4Ca cluster.

In photosynthetic complexes, the light-driven reactions are able to proceed with essentially every photon producing useful reactions, corresponding to a quantum efficiency of near unity. In addition to efficiently performing these light-induced forward reactions, unfavorable side reactions and undesired back reactions are minimized [4]. This balance is achieved by fine-tuning the properties of the cofactors through interactions with the protein in which they are embedded. Theoretical treatments have identified the aspects of electron transfer that the protein surrounding the cofactors of the reaction center can modulate: the energetics, the coupling, and the protein dynamics [5]. This chapter reviews how interactions with the protein can alter the energetics of the cofactors, in particular the oxidation–reduction midpoint potential. For this analysis, the focus is on how interactions with the protein can alter the energetics of the tetrapyrroles of reaction centers and photosystem II, including how these interactions influence the energetics of cofactors to make them highly oxidizing and capable of performing water oxidation.

2 The Structures of Bacterial Reaction Centers and Photosystem II

In the purple bacterium Rhodobacter (Rb.) sphaeroides, the reaction center consists of three protein subunits, termed the L, M, and H subunits, and ten cofactors: four BChl a molecules, two BPhe a molecules, two quinones, a non-heme iron, and a carotenoid. The structures of the bacterial reaction centers from Rb. sphaeroides and Blastochloris (Bl.) viridis have been determined by X-ray diffraction [6–12]. These structures have shown that the reaction center consists of a core domain that is formed by the L and M subunits, which each have five transmembrane helices related to each other by a twofold symmetry axis (Fig. 9.1). The H subunit consists largely of a cytoplasmic domain with only one transmembrane helix. The core domain contains the cofactors that are divided into two branches also related by a twofold symmetry axis, with only one branch being active in performing electron transfer in wild-type reaction centers. The core domain can be biochemically isolated and is active, although the presence of the H subunit is required for stability of the complex [13].

Fig. 9.1
figure 1

Three-dimensional structures of the core domain of the bacterial reaction center (RC) and photosystem II (PSII) showing the cofactors involved in the electron transfer. The view is perpendicular to the twofold symmetry axis that relates the two branches of cofactors (red) and transmembrane helices of the core subunits (L (yellow) and M (blue) subunits of the reaction center, and D1 (yellow) and D2 (blue) subunits of photosystem II). The coordinates are from Allen and coworkers [8] (PDB file 4RCR) and Ferreira and coworkers [15] (PDB file 1S5L)

Full size image

Photosystem II is a much larger complex as it contains over 20 protein subunits and approximately 80 cofactors, most of which are chlorophylls (Chls) [3]. The large size reflects the dual function of photosystem II in harvesting light and performing electron transfer reactions. The three-dimensional structures of photosystem II from the cyanobacteria Thermosynechococcus elongatus and Thermosynechococcus vulcanus [14–18] show that the complex can be thought to have an outer domain that harvests light and surrounds a core domain that performs the primary electron transfer reactions and contains the cofactors involved in the primary photochemistry. The core domain is formed by two protein subunits, D1 and D2, that are evolutionarily related to the L and M subunits of the reaction center and exhibit the same pattern of five transmembrane helices arranged about a twofold symmetry axis (Fig. 9.1). These two subunits encase four Chls, two pheophytins (Phes), two quinones, and a non-heme iron whose positions show a remarkable structural similarity with the corresponding cofactors of bacterial reaction centers. In addition, the core contains some cofactors not present in reaction centers, two Chls, YZ, and the Mn4Ca cluster. The core domain of photosystem II can be biochemically isolated although full activity requires the presence of the additional subunits [19].

In reaction centers from Rb. sphaeroides, the primary electron donor, P865, is a BChl dimer with the two tetrapyrroles overlapping at the ring A position with a separation of ~3 Å (Fig. 9.2). Photosystem II has two Chls at the equivalent position that presumably represent the primary electron donor, P680 (Fig. 9.2), at least in the final oxidized state [20, 21]. For both P865 and P680, the central magnesium atoms are coordinated by histidine residues, L173 and M202 in reaction centers and D1-198 and D2-197 in photosystem II, that are found at similar locations in the structures. While the general orientations of the two Chls are similar to the corresponding BChls in reaction centers, the relative angles between the tetrapyrroles are slightly different, resulting in a longer separation distance of 3.3 Å in photosystem II [18].

Fig. 9.2
figure 2

Three-dimensional structures of P865 from the bacterial reaction center (RC) and P680 from photosystem II (PSII). For the reaction center, the coordinating ligands of the BChls, His L173 and His M202, are shown. For photosystem II, the coordinating ligands of the Chls, His D1-198 and His D2-197, are shown. The view is the same as shown in Fig. 9.1. The coordinates are from Allen and coworkers [8] (PDB file 4RCR) and Ferreira and coworkers [20] (PDB file 1S5L)

Full size image

3 Transition Energies of Tetrapyrroles

In photosynthesis, energy and electron transfer is initiated by the absorption of light that results in the excitation of an electron from the ground state to an excited state. The energetics of these transitions are primarily determined by the chemical composition of the pigments. The tetrapyrroles of reaction centers from Rb. sphaeroides are BChl a and BPhe a with the corresponding cofactors of photosystem II being Chl a and Phe a. The chemical composition of Chl a compared to BChl a differs with a vinyl group rather than an acetyl group at the ring A position as well as a difference in the hydration at the ring B position [22]. The presence of the central magnesium in BChl a and Chl a distinguishes these tetrapyrroles from BPhe a and Phe a, respectively. When in monodisperse solutions, BChl a and BPhe a have optical spectra with a Qy band at 772 nm and 749 nm, respectively, while Chl a and Phe a have transitions at the shorter wavelengths of 662 nm and 667 nm, respectively. In addition to the Qy bands, the optical spectra have absorption bands due to the Qx and Soret transitions at shorter wavelengths. In some organisms the optical absorption bands are significantly shifted because of incorporation of cofactors with alternate substituents; for example Bl. viridis makes use of BChl b that has a Qy transition at 794 nm.

One of the major effects of the incorporation of tetrapyrroles into the photosynthetic complexes is an alteration of their transition energies compared to those measured for the isolated tetrapyrroles. For reaction centers from Rb. sphaeroides, the optical spectrum has Qy bands at 760, 802, and 865 nm. The peaks at 760 nm and 802 nm are associated with the Bphe and BChl monomers, respectively, while the 865 nm peak is assigned to P865. In photosystem II, the optical absorption band associated with P680 is not resolved from the contributions of the other Chls as the Qy peaks are all centered near 680 nm.

The observed spectral shifts of the pigments in the proteins compared to the isolated pigments arise from a combination of protein interactions and pigment–pigment interactions. In certain instances, shifts in the optical spectrum can be attributed to changing interactions between a tetrapyrrole and a specific residue. For example, the position of the Qx transition of the BPhe monomer on the active branch can be shifted from 546 to 534 nm by removing a hydrogen bond from Glu L104 to the keto group [23, 24]. However, changing the hydrogen bonds to the conjugated system of a tetrapyrrole is not generally correlated with a shift in the absorption peak. In general, theoretical modeling of the excited states of tetrapyrroles has proven to be difficult because of uncertainties in key parameters such as accounting for the contribution of electron–vibrational coupling [25, 26].

New techniques are being developed to experimentally determine these key parameters needed to describe the excited states. These methodologies make use of the general approach of two-dimensional transient optical spectroscopy, where the couplings between pigments are found by analysis of the off-diagonal peaks, similar to procedures used in two-dimensional nuclear magnetic resonance spectroscopy [27, 28]. One of the best-characterized systems using this technique is the BChl-a containing FMO protein, or Fenna-Matthews-Olson protein, which is a light-harvesting protein in green bacteria. The three-dimensional structures of this protein from Prosthecochloris aestuarii, Chlorobaculum tepidum, and Pelodictyon phaeum have been determined, showing a conserved arrangement of eight BChls within two β-sheets, with one BChl being largely disordered in the structures [29–34] (Fig. 9.3). The optical spectrum of the BChl cofactors of the FMO protein shows a broad peak centered near 810 nm associated with the unresolved Qy transitions of the BChl cofactors. Measurements of the FMO protein using this optical technique identified not only the contributions of the individual BChls to the optical spectrum but also specific energy transfer pathways among the BChl cofactors (Fig. 9.3).

Fig. 9.3
figure 3

The three-dimensional structure of the FMO protein from Pelodictyon phaeum and model for energy couplings. Shown is one subunit of the FMO protein (wheat) that is a trimer in the cell. Each protein subunit is composed of two β sheets that enclose seven BChls, with an eighth BChl being located at the protein–protein interface of the trimer (atom type). The couplings between the BChls are illustrated by ellipses with the direction of energy flow shown by arrows. Figure adapted from Brixner and coworkers [27] and Larson and coworkers [34] using coordinates from the PDB file 3OEG

Full size image

4 Oxidation–Reduction Midpoint Potentials of Tetrapyrroles

After light excitation, an electron is transferred, resulting in oxidation of the electron donor and reduction of the electron acceptor. The ability of a cofactor to serve as either an electron donor or acceptor is determined by the oxidation–reduction midpoint potential of the cofactor. The potential is sensitive to the chemical nature of the cofactor, for example the potential for BChl a typically is 0.1–0.2 V lower in value than the potential of Chl a, while Chl d typically has a slightly higher potential by ~0.1 V compared to the Chl a potential [35]. The oxidation–reduction midpoint potentials of isolated tetrapyrroles are also sensitive to the particular solvent, with typical values being +0.81 V and 1.14 V for Chl a and Phe a in acetonitrile [36].

The oxidation–reduction midpoint potentials of most of the tetrapyrroles bound to the bacterial reaction center and photosystem II are largely unknown. Experimental measurements using chemical or electrochemical titrations have not been successful because of the lack of equilibrium between the cofactors and the potential of the solution, a problem that is often found for cofactors buried within proteins. The oxidation–reduction midpoint potentials are sometimes inferred based upon electron transfer measurements. For example, the lack of electron transfer along the inactive branch of the cofactors in bacterial reaction centers has been interpreted as arising from those cofactors having unfavorable energetics for electron transfer due to higher potentials compared to the corresponding cofactors on the active branch [37]. In contrast, a wealth of experimental data exists concerning the oxidation–reduction midpoint potential of primary electron donors; therefore, we focus on the effect of protein interactions on those pigments in detail below.

5 Oxidation–Reduction Midpoint Potential of P865

In anoxygenic bacteria, the primary electron donor must efficiently transfer an electron from the excited state of the BChl dimer to the nearby electron acceptors as well as be reduced by secondary electron donors, all of which part of the overall electron transfer pathway. In Rb. sphaeroides, the excited donor P865* transfers an electron through a BChl monomer to the BPhe monomer on the A branch. The difference in energy between the P865*BPheA excited state and the P865•+ BPheA •− charge-separated state has been estimated to be 0.20–0.26 eV using transient optical spectroscopy [37]. Although estimates of energy differences between transient states are made difficult by the complex nature of the kinetic decays, the energetics can be modeled as arising from a time-dependence of the relative energies due to dynamical motion of the protein [38].

The P865/P865•+ midpoint potential of wild-type reaction centers from Rb. sphaeroides is 0.50 V [39–42]. The oxidized donor P865•+ is reduced by an exogenous cytochrome c 2 having an oxidation–reduction midpoint potential of 0.35 V [43]. With this difference in the oxidation–reduction midpoint potentials, the energy differences are favorable for both secondary electron transfer from cytochrome c 2 and forward electron transfer [44]. For reaction centers from Bl. viridis that have a bound tetraheme cytochrome, the midpoint potentials of the BChl dimer and the heme closest to the BChl dimer have been measured to be 0.50 V and 0.38 V, respectively [45, 46]. Thus, the energetics for the primary and secondary electron donors are very comparable despite the structural differences between a small cytochrome that binds transiently compared to a large bound cytochrome.

6 Electrostatic Interactions of P865 with Ionizable Amino Acid Residues

The energy of P865•+ is sensitive to electrostatic interactions with charged amino acid residues with a dependence that is determined by the distance and effective dielectric constant. The dielectric constant provides a measure of the screening of the medium between two charges with values ranging from 4 for hydrophobic protein environments to 80 for charges on the protein surface and exposed to water. The effect of electrostatic interactions between P865 and the protein on the properties of P865 was investigated by alterations that either inserted or removed ionizable residues at several different positions located approximately 10–15 Å from P865 (L135, L155, L170, L247, M164, and M199) [47–49]. Several of these mutants exhibited a pronounced pH dependence of the P865/P865•+ midpoint potential compared to wild type, demonstrating the effect of an electrostatic interaction between the altered amino acid residue and P865•+. The P865/P865•+ midpoint potential was generally found to decrease up to 60 mV due to the introduction of a negative charge located approximately 10 Å from the donor or increase up to 50 mV due to the introduction of a positive charge. These changes in the midpoint potential show that the introduced charges are screened with a bulk dielectric constant having a value of ~20 that can be more explicitly modeled using an exponential dependence for the dielectric constant. In general, the changes of the midpoint potential of P865 measured for these mutants are in agreement with electrostatic models, provided that the charges due to the altered amino acid residues are largely screened [48–50].

7 Hydrogen Bonding to the Conjugated Macrocycles of P865

BChl a has two positions, at the acetyl group of ring A and the keto carbonyl of ring E, that are part of the conjugated macrocycle and can serve as proton acceptors. To investigate how hydrogen bonds influence the electronic structure of the dimer, mutants were constructed in which the number of hydrogen bonds to these positions was altered (Fig. 9.4). Wild-type reaction centers have one hydrogen bond between His L168 and the acetyl group of the A side of P865, and this hydrogen bond can be removed by a His to Phe mutation, while the complementary mutation, Phe to His at M197 at the symmetry-related residue, introduces a hydrogen bond to the acetyl group of the B side of P865 [51–54]. Histidine residues introduced at Leu L131 and Leu M160 form hydrogen bonds with the keto carbonyl groups of the A side and B side of P865, respectively [40]. By constructing mutants with different combinations of these alterations, the number of bonds was decreased to zero or increased to four [42]. The gain or loss of a hydrogen bond at each position was measured by use of Fourier transform infrared (FTIR) spectroscopy. For the mutants that were designed to introduce new hydrogen bonds to the keto carbonyls of P865 by the substitutions Leu to His at L131 and Leu to His at M160, large frequency downshifts of the vibrational bands assigned to the keto carbonyl groups were observed [55]. The gain of a hydrogen bond to the acetyl substituent of the B side of P865 in the Phe to His at M197 mutant and the loss of the hydrogen bond to the acetyl substituent of the A side of P865 in the His to Phe at L168 mutant were clearly evident in the infrared spectra of the primary donor obtained using Fourier transform Raman spectroscopy [56] and X-ray diffraction [57].

Fig. 9.4
figure 4

Three-dimensional structure of P865 and nearby amino acid resides Leu L131, His L168, Leu M160, and Phe M197. In wild type, there is one hydrogen bond to His L168. Hydrogen bonds were introduced by substitutions of His at L131, M160, and M197 in different combinations. Substitution of Phe to His at 168 results in loss of the existing hydrogen bond. The coordinates are from Allen and coworkers [8] (PDB file 4RCR)

Full size image

When the single hydrogen bond between His L168 and the acetyl group of the A side of P865 in wild type is removed by replacement of His with Phe, the optical absorption band at 865 nm shifts slightly to shorter wavelengths, and the midpoint potential of P865 decreases [52–54]. Structural studies show that the substitution of His L168 with Phe does not alter the overall structure of the protein, and the only significant changes other than the mutation are a 20°–27° rotation of the acetyl group and a small displacement of the A side of P865 [53, 58]. When several different amino acid residues were substituted at L168, systematic shifts were observed in the decrease in the P865/P865•+ midpoint potential and the shift of the 865 nm band, with His forming the strongest bond [53].

The most striking effect that the alteration of the hydrogen bonds has on P865 is the pronounced change in the midpoint potential with the number of hydrogen bonds. Wild-type reaction centers have one hydrogen bond and a P865/P865•+ midpoint potential of 505 mV (Table 9.1). Removal of this hydrogen bond with the His to Phe mutation at L168 results in P865 having no hydrogen bonds and a 95 mV decrease in the midpoint potential. The addition of a single hydrogen bond, resulting in a total of two hydrogen bonds from His to P865, increases the P865/P865•+ midpoint potential by 60–125 mV. When P865 has three or four hydrogen bonds, the P865/P865•+ midpoint potential increases even further, to a maximum of 765 mV for a mutant that has a total of four hydrogen bonds. The mechanism by which the number of hydrogen bonds controls the P865/P865•+ midpoint potential can be understood in terms of a Hückel molecular orbital model as described in the following section.

Table 9.1 Effect of changing hydrogen bonds on the P865/P865•+ midpoint potential and unpaired spin density ratio of P865•+
Full size table

8 Modeling the Electronic Structure of P865

Many of the properties of P865 can be characterized in terms of a Hückel molecular orbital model. In this model, P865 is represented by two BChl molecules, A and B, corresponding to the L and M sides, that are coupled together as described by the parameter β (Fig. 9.5). The energies of the molecular orbitals of each BChl are considered to be energetically inequivalent because of different interactions with the surrounding protein. This inequivalence is modeled by poising the two molecules at energies εA and εB that differ by Δα. The energy difference between the two molecular orbitals of the coupled system, ΔE, is then given by:

Fig. 9.5
figure 5

Hückel model of P865. In wild type, the molecular orbitals are split by the inequivalence in the energies of the BChls on the L side (or equivalently the A side) and M side (or equivalently the B side) and their coupling according to Eq. (9.1). The E m value of P865/P865•+ corresponds to the energy difference between the highest molecular orbital and the continuum. The introduction of a hydrogen bond to the M side of P865 stabilizes the energy of that BChl resulting in a larger E m value and a more asymmetric dimer. The introduction of a hydrogen bond to the L side of the dimer also results in a larger E m value but a more symmetric dimer. Modified from Williams and Allen [5]

Full size image
$$ \Delta E=\sqrt{\Delta {\alpha}^2+4{\beta}^2} $$
(9.1)

The model predicts that the electron spin densities, ρA and ρB, of the two BChls of P865 will be inequivalent according to the ratio:

$$ {{\rho}_A}\!\left/ \!{{\rho}_B}\right.={\left[{\left(\Delta \alpha +\Delta E\right)}\!\left/ \!{2\beta }\right.\right]}^2 $$
(9.2)

The P865/P865•+ midpoint potential is determined by the energy of the highest occupied molecular orbital, and the model predicts that the midpoint potential E m (P865/P865•+) is directly related to the value of the spin density ratio ρA/ρB according to:

$$ {E}_m\left( P865/ P{865}^{\bullet +}\right)=-{\epsilon}_B-\beta \sqrt{{{\rho}_A}\!\left/ \!{{\rho}_B}\right.} $$
(9.3)

The Hückel model provides a platform to understand a number of features of the electronic structure of P865 that can be measured experimentally and related to the ΔE, β, and Δα parameters [49, 59–61]. For example, the FTIR spectrum of wild-type reaction centers has a broad band centered at 2,600 cm−1 that has been assigned as an intervalence charge-transfer band, yielding a value of 320 meV for ΔE [60]. Similarly, the asymmetry of the unpaired electron spin density over P865•+, which has a spin density ratio for ρA/ρB of 2.09 for wild-type reaction centers as determined by electron nuclear double resonance spectroscopy, yields an estimate of 0.37 for the Δα/β ratio [59].

The Hückel model can be used qualitatively to explain the trends evident in the spin density ratios of the hydrogen-bonding mutants, which have a range of values from 0.28 to 4.94 (Table 9.1) [62]. In this model, the energy of the BChl near M160 in wild type is lower compared to the energy of the BChl near L131 (Fig. 9.5). Hydrogen bonds to M160 stabilize the nearby BChl of P865, making the energies of the two halves more asymmetric, whereas hydrogen bonds to L131 make the dimer more symmetric. Because the spin density reflects this asymmetry, the M160 mutants have an increase in the ρA/ρB ratio, while the L131 mutants show a decrease in this ratio. In both cases stabilization results in a higher E m (P865/P865•+), that is, a larger amount of energy is required to remove an electron from P865 to the continuum. The effects on the ρA/ρB ratio are additive, for example in the mutant with hydrogen bonds introduced to the keto groups of both BChls of P865 by changes at both L131 and M160, the energies of both BChls change. Because the change in energy on both sides is similar, the relative asymmetry is unchanged, resulting in essentially the same spin density ratio as measured in wild type. Another set of mutants was tested in which a series of amino acid residues was substituted at L131 and M160 [63, 64]. For each mutant, electrochemical titrations and electron nuclear double resonance measurements were performed, and the resulting set of data demonstrates a well-defined correlation between ρA/ρB and E m (P865/P865•+) [64, 65].

The mutations at L168 or M197 change the hydrogen bonding to the acetyl groups of P865, in a different position than L131 and M160, which are near the keto group of ring E of the BChls (Fig. 9.4). The introduction of a hydrogen bond at M197 does not alter the spin density ratio significantly, while an increase in the spin density ratio is observed when a hydrogen bond is added at L168. Mutations at these two positions are more difficult to interpret, since the changes may result in rotation of the acetyl groups in the mutants relative to wild type, with a consequent change in the conjugation of the macrocycle. Structural rearrangements may contribute to the observed effects in addition to the contributions from the change in energy due to the hydrogen bond.

The model provides estimates of energies for the heterodimer mutants, which have changes at His L173 or His M202, the two ligands to P865. Substitution with Leu results in incorporation of a BChl-BPhe dimer in place of P865 [66–71]. The heterodimer mutants have many spectral and electron-transfer changes although the directionality of transfer along the active branch is maintained. Notably, the oxidation–reduction midpoint potential of the heterodimer is 0.64 V compared to 0.50 V for wild type [72]. This increase can be understood using the Hückel model, as the energy associated with the BPhe side of the heterodimer is increased compared to the corresponding BChl of P865 because of the higher midpoint potential of BPhe compared to BChl. The increase in the difference in energies of the two macrocycles causes the spin density distribution to become more localized on the BChl side. In addition, the optical bands associated with the heterodimer have a significant increase in their broadness compared to P865 because of the contribution of charge-transfer states [73].

While the model provides a simple qualitative description of the effects of different mutations on the properties of P865, more complete descriptions of the electronic states are required in order to accurately describe the effects of protein interactions. For example, the reorganization energy associated with charge transfer, λ, is the energy change needed to move a charge from one side of the dimer to the other [64, 65]. Incorporating this term into the model generates a modified relationship between the midpoint potential and spin density in which large values of this term compared to ΔE would significantly alter the localization of charge and asymmetry of P865•+. For the reaction center mutants with alterations at residues L131 and M160, resulting in changes in the hydrogen bonds to P865 at the two keto positions, the revised model provides estimates that the energy associated with the introduction of a hydrogen bond to stabilize a BChl is approximately 100 meV, the coupling, β, is 120–160 meV, and the reorganization energy, λ, is 100–200 meV [64]. To model the broadness of the optical bands in the near infrared region, it is necessary to include the contribution of vibrational states [48]. Also contributing to the IR band at 2,600 cm−1 is a second electronic transition between the second highest occupied molecular orbital and the highest occupied molecular orbital, which is only partially filled in the P865•+ state [65]. Combining the contributions of the two transitions with the use of several vibrational modes results in a greatly improved correspondence between the calculated and experimental optical spectra [65] including the P865•+ Stark spectrum [74]. The resulting values of the coupling, β, and reorganization energy, λ, are 126 meV and 139 meV, respectively, with a value of 69 meV for Δα in wild type. Thus, the Hückel model can be extended with the additional parameters to provide an accurate, quantitative description of the electronic structure of P865.

9 Achieving High Oxidation–Reduction Midpoint Potentials for Water Oxidation

In order to oxidize water, P680 must have a high oxidation–reduction midpoint potential of at least +0.82 V at pH 7 and +0.93 V at pH 5, which is the pH range of the thylakoid lumen. Because of its high value, the P680/P680•+ midpoint potential cannot be directly measured, but it has been estimated to be approximately 1.1–1.3 V based upon electron transfer measurements, making P680 the strongest known biological oxidant [3, 20, 75]. The effect of the interactions with the surrounding protein environment on the P680/P680•+ midpoint potential has been calculated for photosystem II [76, 77]. Based upon these computational calculations of the electrostatic contributions, the dipolar and charged amino acid residues result in a significant increase in the P680/P680•+ midpoint potential, with residues D1-176 to D1-195 and D2-176 to D2-194 having the largest influence. The presence of the positive charges from the Mn4Ca cluster on the P680/P680•+ midpoint potential has been estimated to increase the P680/P680•+ midpoint potential by 0.1–0.2 V, although experimental measurements of electron transfer rates suggest that the potential is higher when the Mn4Ca cluster is removed [20].

A number of factors contribute to the increase in potential of ~0.7 V for P680 compared to P865, including the difference of ~0.2 V due to the presence of Chl a instead of BChl a. As discussed above, the P865/P865•+ midpoint potential can be increased by the addition of hydrogen bonds to the macrocycle, raising the potential to above 0.75 V. However, the presence of the vinyl group on Chl a rather than the acetyl group found in BChl a limits the possible hydrogen bonds to P680 to those involving the two keto groups. Such hydrogen bonds are weak according to FTIR measurements [20] and their influence on the high midpoint potential of P680 has not been established. Another factor is the weaker coupling of P680 compared to P865 due to their structural differences. A preferential localization of the unpaired electron of P680•+ is observed that is attributed to an asymmetry in the energies of the two Chls of P680 [77, 78]. This asymmetry is calculated to be partially due to differences in the electrostatic interactions with the surrounding protein with residues Asp D1-61, Asn D1-181, Asn D1-298, His D2-61, Arg D2-180, and Arg D2-294 having critical contributions. The Hückel model predicts that the P680/P680•+ midpoint potential is correlated to the electron spin distribution and that the weaker coupling of P680 should increase the potential by ~0.1 V.

The high P680/P680•+ midpoint potential is required to oxidize water, which occurs through the sequential transfer of four electrons from the Mn4Ca cluster, which becomes systematically oxidized during the S cycle. A key property of the Mn4Ca cluster is that as it becomes more oxidized the oxidation–reduction midpoint potentials associated with the Mn4Ca cluster remain relatively constant, achieved by coupling electron transfer with proton transfer. This coupling maintains a sufficient driving force for electron transfer to P680 through YZ while minimizing unfavorable reactions that would be driven by the presence of highly oxidizing states [79]. None of the oxidation–reduction midpoint potentials associated with the different oxidation states of the Mn4Ca cluster have been experimentally established although electron transfer considerations suggest values near 1 V [20, 79].

The ability to manipulate the bacterial reaction center such that it becomes highly oxidizing has provided the opportunity to redesign the protein to have a new Mn cofactor that is capable of performing oxidation–reduction reactions. A binding site for Mn was created in the highly oxidizing reaction centers at a location analogous to that of the Mn4Ca cluster of photosystem II (Fig. 9.6). As an initial step, the introduction of two carboxylates, at M168 and M288, with a tyrosine at M164, produced a site capable of binding a mononuclear Mn cofactor [57, 80]. Several residues near the binding site in the reaction center are conserved including Glu M173, His M193, and Tyr M164 that correspond to Asp D1-170, His D1-190, and Tyr D1-161, respectively, in photosystem II. Electron transfer measurements demonstrate that the Mn cofactor serves as a rapid secondary electron donor to P865. The Mn2+/Mn3+ midpoint potential was experimentally determined to be 0.63 V by measuring the equilibrium between the Mn cofactor and P865•+ for a series of mutants with different P865/P865•+ midpoint potentials [81].

Fig. 9.6
figure 6

Three-dimensional structure of the Mn-binding bacterial reaction centers. (Top) The mononuclear Mn cofactor (purple sphere) and several surrounding amino acid residues, Tyr M164, Glu M168, Glu M173, His M193, and Asp M288 (atom type). (Bottom) The Mn-binding site is located approximately 10 Å from P865 at a location analogous to the position of the Mn4Ca cluster of photosystem II. Shown are the L (yellow) and M (blue) subunits, P865 (red), and the amino acid residues forming the Mn-binding site (red). Several residues near the binding site in the reaction center, Glu M173, His M193, and Tyr M164, have counterparts in photosystem II, namely, Asp D1-170, His D1-190, and Tyr D1-161, respectively. The three-dimensional structure of the mutant with the bound Mn was determined by X-ray diffraction [57] (PDB file 1Z9J)

Full size image

The ability of the Mn cofactor incorporated in bacterial reaction centers to participate in oxidation–reduction reactions was shown by its ability to catalyze oxygen production from superoxide [82]. Light excitation of the Mn-binding reaction centers leads to the Mn2+ P865•+ Q− charge-separated state that quickly converts to Mn3+ P865 Q−. In the presence of superoxide, the Mn3+ is reduced to Mn2+ and the superoxide is converted into molecular oxygen following the equivalent reactions of Mn-superoxide dismutase. However, Mn-binding reaction centers are not capable of the metal-oxidizing reaction of converting superoxide into hydrogen peroxide, as found in Mn-superoxide dismutase, despite favorable energetics for this process, presumably because of the necessity of coupling this reaction with the transfer of two protons. The continuing investigation of properties of the Mn-binding mutants coupled with intensive studies of the Mn4Ca cluster should provide an insight into how the midpoint potentials of Mn-cofactors are influenced by protein interactions, in particular how the protein environment poises the Mn4Ca cluster at the potentials needed to efficiently oxidize water.