Photosystem II core antenna complex

In oxygenic photosynthesis, Photosystem I1 is the antenna/RC supercomplex that carries out light-induced charge separation across the membrane by doubly reducing plastoquinone on the acceptor side, and the water-splitting reactions in the oxygenevolving complex attached to the donor side (Sauer, 1979; Barber and Santini et al., 1994; Kuhlbrandt, 1999). The intermediate-resolution X-ray structures shown in Fig. 2.11 (p. 86) at 3.8 A resolution have recently been reported for the core PSII complex of the cyanobacterium S. elongatus; this occurs in its native form as a homodimer in the membrane (Zouni et al., 2001; Kamiya and Shen, 2003). The monomeric unit of the PSII core complex is made up of at least 21 protein subunits, 18 of them located within the photosynthetic membrane. In contrast to the PSI supercomplex, the RC pigments and the antenna pigments of the PSII core complex are bound to different protein subunits. The RC pigments, 4 Chls and 2 pheophytins are located on the D1 @sbA) and D2 @sbD) subunits (Fig. 2.12, p. 87). These RC subunits are highly homologous to the L and M subunits of the purple bacterial RC complex and to the cofactor positioning.
 

The subunits carrying the antenna pigments are CP43 (PsbC) and CP47 (PsbB), which are arranged around a local pseudo-C2 axis in the monomer. Each has 6 transmembrane helices arranged as a pseudo-trimer. This structure is very similar to the six amino-terminal transmembrane-helices of PsaA and PsaB, the large core subunits in the PSI RC, suggesting that the two types of reaction centre had a
common ancestor.
 

According to the X-ray structures, CP43 and CP47 bind 12 and 14 Chl a molecules respectively. As in the PSI RC complex, the Chl pigments are arranged in two layers close to the two membrane surfaces, most of them presumably being bound via histidine ligands to the protein (Fig. 2.13, p. 88). The range of centre-tocentre distances for the Chls is 8.5-13 A, within the range expected for rapid energy transfer. Compared with the purple bacterial RC, the D1 and D2 subunits each carry one additional Chl, Chl Z ~anId Chl zD2, respectively. These Chls, despite being bound to the RC subunits, are relatively far distant (about 30 A) from the RC pigments. They may act partly as linker Chls in energy transfer to the core RC. The nearest antenna Chls of the CP43 and CP47 subunits are somewhat closer, at a distance of 20 A, but this is still a relatively large distance for energy transfer to the RC.
 

The energy-transfer steps have been studied in isolated CP43 and CP47 complexes, and overall energy trapping in the intact core complex. In the isolated complexes, energy equilibration is very fast, mostly in the time range of 0.2-0.4 ps, with some slower 2-3 ps contributions (de Weerd et al., 2002a and 2002b). The fast transfer steps, which have mainly been assigned to equilibration in the Chl layer near the stromal side of the membrane, imply average single-step transfer times of -100 fs between pairs of Chls. Such fast rates are expected from the relatively close packing of the Chls, and are in reasonable agreement with theoretical calculations (de Weerd et al., 2002b). The slower transfer step has been assigned to transfer from the lumenal to the stromal layer of Chls, and to the transfer to the lowest-energy excitonic state in each system at low temperature. This final transfer step may however be substantially faster at room temperature. De Weerd et al. (2002b) have concluded that the rates of energy transfer within the isolated CP43 and CP47 complexes are fast enough not to limit the overall trapping of excitation energy in Photosystem 11.
 

The initial trapping of energy by charge separation in the RC of PSII core complexes occurs in about 40 ps in cyanobacterial complexes of S. elongatus (Schatz et a/., 1987 and 1988) and increases up to -200-250 ps for the large PSII antenna/RC particles of higher plants (Holzwarth and Roelofs, 1992). From a comparison of the trapping lifetimes and rates with open and closed RCs, it had been concluded before the X-ray structure determination that kinetics in the PSII core are mostly trap-limited (Schatz et al., 1987 and 1988). Since the antenna and the RC are essentially isoenergetic at room temperature, the 40 ps lifetime and the 32 pigments per RC would, according to the simple picture of trapping presented in Section 2.1.2, imply a primary charge-separation rate of about (1.3-1.5 ps)-l within a trap-limited model.
 

However, recent modelling based on the structural data has led to widely deviating conclusions (Vasiliev et al., 2001 and 2002). According to these workers, the fastest transfer steps from the antenna Chls into the RC are slower than 5 ps because of the large antenna-to-RC distances, which may be necessary for avoiding oxidation of the antenna Chls by oxidised P680. They thus advocated a transfer-to-trap-limited model for energy transfer in the PSII core complex, arguing that electron transfer in the RC is much faster than transfer from the antenna to the RC.
 

Many assumptions about unknown spectral properties are involved in such modelling, and as yet it is not clear whether the transfer-to-trap-limited model is correct since there are several unexplained contradictions with other data. For example, it is well known that the fluorescence yield and lifetime of PSII critically depend on whether the RC is open or closed (see, for example, Holzwarth and Roelofs, 1992). This cannot easily be explained within a transfer-to-trap-limited model, but it is easily understood within an essentially trap-limited model (Holzwarth and Roelofs, 1992): if one assumes at the extreme a situation where energy equilibration between antenna and RC occurs at about the same rate as charge separation in the RC, one would still have an essentially trap-limited model which would at the same time be consistent with the RC redox-state sensitivity of the fluorescence lifetime and the antenna size dependence of the PSII lifetime. We thus believe that the models of Vassiliev et al. (2002) and Dekker and van Grondelle (2000) are too extreme, and further work is necessary finally to clarify the transfer steps in the PSII core.
 

A recent detailed study of the dependence of trapping kinetics on antenna size in various PSII particles also came to the conclusion that indeed there is a shallow equilibrium established in PSII prior to charge separation (Barter et al., 2001), consistent with the early model and conclusions of Schatz et al. (1988) and Holzwarth and Roelofs (1992). Barter et al. (2001) concluded that a shallow equilibrium between the antenna and reaction centre in Photosystem I1 would facilitate regulation via, for example, non-photochemical quenching, and went on to propose that Photosystem II is optimised for regulation rather than for efficiency. However, the efficiency of PSII in the absence of quenching is very high (usually better than 92%), which implies that regulation capabilities and high efficiency are not in any way mutually exclusive. In conclusion, at present it seems likely that Photosystem I1 kinetics are essentially trap-limited rather than transfer-to-trap-limited.

Photosystem I

The PSI antenna/RC complex appears as a native trimeric unit in the membrane of cyanobacteria, in contrast to the monomeric PSI complex occurring in higher plant species. Despite substantial differences in the number of total polypeptides and differences in structural details, including the different macroorganisation of PSI in the different organisms, there probably exist many similarities in the structure of their core antenna, as judged on the basis of the sequence homology. No detailed structure is as yet available from a higher-plant PSI complex, although a 4.5 A X-ray structure of LHI-PSI is in press at the time of writing; see also Kargul et al. (2003).
 

The PSI complex of S. elongatus consists of 12 protein subunits and contains 96 Chls (95 Chl a and one Chl a’ ’, which is located in the ‘special pair’ of the RC), 22 carotenoids (mostly pcarotene), two phylloquinones and the [4Fe-4S] centres involved in electron transfer as electron acceptors.
 

Figure 2.4 (p. 83) shows the structure of Photosystem I (only one monomer of the trimeric structure is shown for clarity) to a resolution of 2.5 8, (Fromme and Witt, 1998; Fromme et al., 2001; Jordan et al., 2001) at near atomic resolution. These studies have for the first time revealed the orientations of all the chlorin ring systems, enabling more rigorous theoretical calculations of energy-transfer properties based on the distances and transition-dipole orientations of the Chls. A salient feature of the PSI complex is that a large number of the antenna pigments (89 Chls), as well as the 6 RC Chls, are bound to the same two polypeptides, namely the highly homologous psaA and psaB units forming the core of the structure. These two polypeptides are related by a pseudo-C2 axis located at the centre of the PSI monomer. The organic cofactors of the electron-transfer chain of RCs are arranged in two branches along this pseudo-C2 axis.
 

Figure 2.5 (p. 84) shows the arrangement of the pigments in one monomer of the Photosystem I trimer. Viewed from above (upper diagram), the antenna Chls form an elliptically distorted, cylindrical ring structure around the six central RC Chls. Except for the two ‘linking Chls’, which may possibly connect energetically to the antennae and the RC pigments, the distance between the nearest antenna pigments and the RC pigments is relatively large (>18 A). The antenna Chls have centre-to-centre distances of 7-16 A, with a maximum in the distance distribution around 10 A, well within the range allowing ultrafast energy transfer. The side view of the pigment arrangement (lower diagram of Fig. 2.5) shows that in the greater part of the PSI antenna, except close to the RC, the Chls are arranged in two layers located near the two membrane surfaces, with large distances between the two. Thus energy transfer is expected to occur preferentially within the layers, while layer-layer transfer will occur only near the RC. Thus a large part of the PSI antenna is quasi-two-dimensional with respect to the arrangement of the Chls.
 

There are three regions in the antenna structure (highlighted in red in Fig. 2.5) where close stacking of two or three Chls occurs. These arrangements should lead to a substantial excitonic coupling, conferring special spectroscopic properties on these Chls. In all probability, these stacking regions contain the so-called special ‘red Chls’ that are present in all PSI complexes to various extents. This term denominates those groups of Chls that absorb beyond 700 nm, above the absorption maximum of the RC, which is located around 700 nm. These ‘red Chls’ are particularly prominent in cyanobacterial PSI and are believed to play a special limiting role in the overall energy-transfer process to the special pair.
 

Energy transfer in PSI core complexes of cyanobacteria and green algae has been extensively studied, both experimentally and theoretically, but no general agreement has so far been reached as to the rates of the various energy-transfer steps and ratelimiting processes. Most experimental studies have been performed on the core complexes of either cyanobacteria or green algae and higher plants, while relatively few detailed studies are available for intact higher-plant PSI complexes that also carry the light-harvesting I complexes (LHCI) of the outer antenna. These studies have been interpreted either qualitatively or in more detail in terms of so-called compartment models. In these simpler models, energy and electron-transfer processes between groups of pigments are analysed, rather than those between each pair of pigments in each complex.

Figure 2.6 summarises various compartment models for the energy-transfer processes in the PSI core. These models differ in the relative rates of energy transfer from the core antenna to the RC and the effective charge-separation rates. Figure 2 . 6 ~ shows the so-called trap-limited scheme, where the energy transfer from the core antenna to the RC is much faster than the charge-separation lifetime. This model has been adopted by several groups (for reviews see Karapetyan et al., 1999 and Melkozernov, 2001). A contrasting model is the so-called transfer-to-trap limited model, which is primarily based on data from cyanobacterial core complexes (for reviews, see Gobets and van Grondelle, 2001 and Gobets et al., 2001). A common theme in these models, shown in Figs. 2.6b-d, is the overall transfer time from the core antenna pigments to the RC, which is the rate-limiting step, having a lifetime of -20 ps. The models differ, however, in the charge-separation lifetimes within the RC, with variations from 1 ps to 10 ps. Individual energy-transfer steps between groups of pigments are very rapid, in the range of 100-200fs, leading to a very fast subpicosecond energy equilibration within the core antenna. A complication in the kinetics arises from the presence of the special ‘red pigments’ in PSI of cyanobacteria, which generally slow down energy transfer to the RC if they are located in the antenna.
 

Figure 2.7 gives the absorption spectra of some PSI and PSII preparations, showing the much larger red tail in the absorption of the PSI as compared with the PSII particles. The tail is particularly pronounced for Spirulina platensis PSI because of the extreme content of red pigments. The various cyanobacterial PSI complexes are believed to contain different numbers of red pigment molecules, ranging from about 2 in Synechocystis to an extreme of about 6-8 in Spirulina platensis. They also differ in their spectral signatures, ranging from 708 nm absorption (Ch1708) in Synechocystis, through 7 18 nm in Synechococcus elongatus, to 735 nm in Spirulina. The latter seems to contain the longest-wavelength absorbing Chls of any PSI complex, giving rise to 760 nm fluorescence at low temperature (Karapetyan et al., 1999). The red pigments are believed to exchange energy with the core antenna in about 5-10 ps (Gobets and van Grondelle, 2001). This relatively slow energy exchange slows down the overall energy-transfer rate to the RC and severely complicates the observed kinetics. The time constant of the overall energy-trapping process (as measured by the kinetics of formation of the charge-separated states) in cyanobacterial core complexes at room temperature ranges from about 23 ps for Synechocystis, a species with minimal redpigment content, to about 35 ps in s. elongatus with an intermediate amount of red pigments, to a maximum of about 50 ps in Spirulina platensis. This slowing down in the overall trapping rate has been ascribed to the effects of the red pigments (Gobets et al., 2001).

A problem for all these models at present is that until recently no detailed spectra of the intermediate species involved in the kinetics have been resolved. Based on such data, Muller er ul. (2003) have recently proposed the new model shown in Fig. 2.8. This does not so far take into account the effects of the very long-wavelength red pigments, but is limited to PSI particles with low red pigment content. such as the green algae Cltlnrnyrloiiioiius reitlhurdtii and the cyanobacterium Synechocystis (Milller EI nl.. 2003). It is essentially a trap-limited model where energy equilibration within the core is subpicosecond, the transfer between the core antenna and the RC is very rapid (at most a few ps), and the effective charge-separation step is still fast (about 6-9 ps) but significantly slower than the energy equilibration between the core antenna and RC. Energy equilibration within the RC itself is also ultrafast, typically about 200 fs. In this model, the intrinsic charge-separation step in the RC is estimated to occur on a sub-picosecond time scale. This model is the first time to provide consistent spectra of the various intermediates, along with a description of the dynamics of the core antenna processes of PSI. Eventually it will have to be extended lo include the effects of the more extreme red pigments.
 

These data imply that PSI is the fastest RC known, featuring an intrinsic initial charge-separation step taking about 0.5-0.8 ps, which is a factor of five faster than bacterial RCs. with their charge-separation time of about 3 ps. Despite disagreements in the literature about the relative rates of antenndRC equilibration and charge separation, there is general agreement that energy equilibration within the PSI core, barring the slower transfer to/from the small number of red pigments, is very fast, typically in the time range of 200-600 fs (Melkozernov et al., 2000; Kennis et al., 2001; Gibasiewicz et al., 2001; Muller et al., 2003). Such fast equilibration is also in agreement with the detailed structure-based modelling studies mentioned above.
 

We have already noted that S. elongatus contains several close-lying groups of Chls that could be assigned to the special red Chl pigments. This is borne out by data showing that these red pigments derive their bathochromic shift from excitonic coupling rather than a special protein environment. Several groups have tried to model the spectral and kinetic properties of the S. elongatus PSI in detail, based on crystallographic information (Beddard, 1998; Byrdin et al., 2002; Sener et al., 2002; Damjanovic et al., 2002). However, even a very precise high-resolution structure does not provide the exact spectroscopic properties of a specific Chl in a protein complex. This energetic position is determined by the detailed interaction of the pigments with the environment, which can be obtained only by a full quantum-mechanical calculation based on an exact structure. PSI is particularly variable in this respect: the core antenna contains only Chl a, but the absorption maxima of the various Chls range from about 640nm up to 735 nm, indicating a wide range of pigment environments. Much of this range is due to pigment-protein interactions, but pigment-pigment interactions such as charge-transfer and excitonic interaction also have an effect. Thus theoretical descriptions usually treat the spectral properties of the individual pigments as a fitting parameter ((Byrdin et al., 2002; Sener et al., 2002).
 

One attempt has been made (for PSI) to calculate the energetic locations of the pigments quantum-mechanically, taking into account the interaction of each specific Chl with its environment (Damjanovic et al., 2002). At present, the conclusions of these studies, based on the information available for the PSI complex of S. elongatus, vary substantially. More work is needed to arrive at final conclusions about the specific location of the red pigments as well as many other details of the spectral and kinetic properties of PSI of S. elongatus. However, irrespective of the outcome of such calculations, it can already be stated that the exact details of the pigment arrangement in the antenna and the distribution of spectral forms across the antenna do not have any decisive influence on the overall kinetics, because of the quasistatistical averaging of pigment properties that occurs in such a large antenna array. Thus the functioning of the PSI antenna system seems to be fairly robust against even relatively drastic changes in spectral distribution and other properties. By contrast, the particular spectral and kinetic properties of the reaction centre, as well as its electrontransfer rate, seem to be much more decisive for the overall trapping kinetics and the total yield of charge separation. This is not surprising since these parameters directly

influence the key steps in the energy equilibration between antenna and RC and the overaIl energy flow toward the RC. This can be understood in more detaiI from the scheme in Fig. 2.8. Photosynthetic organisms have probably used these parameters to fine-tune PSI function in different organisms.
 

The PSI core antemst of higher plants and green algae binds various amounts of LHCI light-harvesting ncomplexes, whose structures may be similar to those of LHCU (discussed below) but are not known in any structura1 detail. These peripheral antenna complexes in higher plants also contain red pigments (Melkozernov, 2001). The average energy of the Chls in the LHCI complexes is above 700 nm, i.e. above the RC absorption maximum. The overall energy equilibration within the PSI antenna systems of higher plants seem however to be very rapid, with lifetimes from a few ps to about 12 ps, depending on the amount and type of peripheral antenna complexes present. The overall trapping times in the intact complex range from -5Ops to -100 QS at room temperature: see Melkozernov (2001) for a detaiIed review. Much longer lifetimes may be observed at low temperatures because of trapping on longwavelength pigments Iocated far from the RCs. Thus the overall trapping times in PSI of higher plants are generaliy dower than in cyanobacterial PSI by a factor of 2-5. However, the ovem1l trapping times are still fast, allowing a high yield of >90% for charge separation. The slower trapping in higher-plant PSI is mainly a consequence of the larger number of pigment moIecuIes per RC but partly due to the distant location of red pigments in the peripheral antenna. Thus in the PSI compIexes of higher plants with associated peripheral light-harvesting complexes, there may be a mixed situation. with trap-Iimited kinetics in the core, but diffusion-limited kinetics in at least part of the peripheral LHC compkxes (Jennings ef al., 1998; Croce ef al., am).

The contribution of red pigments to the kinetics of the cyanobacterial core complexes could also perhaps be described using such a mixed model. Some structural details of the arrangement of the peripheral antenna complexes in PSI of higher plants and green algae have recently been obtained by electron microscopy (Boekema et al., 2001b; German0 et al., 2002; Kargul et al., 2003). It was found that the LHCI complex binds on one side of the PSI core only. There are, however, substantial differences in the size and arrangement of the outer antenna complexes between different organisms. Figure 2.9 (p. 84) shows the arrangements of the various PSI complexes in the green algae Chlamydomonas reinhardtii, spinach and the cyanobacterial trimeric PSI core. In contrast to PSII (discussed below), there does not seem to be high symmetry in the antenna PSI complexes of green algae and higher plants.
 

As regards the carotenoids, about 60 of the antenna chlorin heads in the PSI structure of Synechococcus are in van der Waals contact with the 22 carotenoids (Jordan et al., 2001), which should provide excellent conditions for light harvesting through the carotenoids as well as photoprotection of the antenna by Chl triplet quenching. Little experimental information is available at present on carotenoid-to- Chl energy transfer in the core antenna of PSI as compared with PSII. However, one can conclude from simple fluorescence excitation spectra that the carotenoids contribute significantly (typically above 50% yield, but there are exceptions: see below) to the light-harvesting function in the wavelength range 450-650 nm where the Chls do not absorb well.
 

A very interesting case of regulation involving a major restructuring of the PSI antenna has recently been found in cyanobacteria (Bibby et al., 2001a; Boekema et al., 2001a) and Prochlorococcus (Bibby et al., 2001b), the most abundant photosynthetic organism in the oceans. Iron deficiency, which is often the limiting factor for growth of photosynthetic organisms in aquatic ecosystems, leads to the induction of additional proteins around the PSI core such as IsiA in cyanobacteria and a similar protein in Prochlorococcus. IsiA has been implicated in chlorophyll storage, energy absorption and protection against excessive light. However, it has now been shown that a PSI-IsiA supercomplex is abundant under conditions of iron limitation. Electron microscopy has revealed that this supercomplex consists of trimeric PSI surrounded by a giant closed ring of 18 IsiA proteins binding about 180 chlorophyll molecules (Bibby et al., 2001a and 2001b; Boekema et al., 2001a; Nield et al., 2003). Figure 2.10 (p. 85) shows the structure of this giant ring.
 

Energy transfer within this supercomplex has been recently studied by timeresolved absorption and emission spectroscopy (Melkozernov ef al., 2003), showing that the ring is energetically tightly coupled to the core, and energy equilibration  within the ring occurs on a sub-picosecond time scale. One should realise that this time constant probably does not reflect energy transfer through the whole giant ring, but simply energy equilibration between directly neighbouring subunits. Energy transfer along the outer ring cannot be resolved by time-resolved spectroscopy since it occurs among spectrally identical subunits. Energy transfer from the outer ring to the nearest Chl molecules in the central core occurs with a time constant of -1.7 ps, while overall energy transfer from the ring to the core takes -10 ps.

Structural and functional basis for light absorption and harvesting

In plants, green algae, cyanobacteria and a few other oxygen-evolving photosynthetic organisms, the primary steps of photosynthesis occur in two membrane-bound protein supercomplexes, Photosystem I (PSI) and Photosystem 11 (PSII), introduced in Chapter 1. Recently the structures of several of the antenna systems of oxygenic photosynthetic organisms have been determined to various degrees, some to atomic resolution. The best characterised of these complexes is the PSI antenna/RC complex of the cyanobacterium Synechococcus elongatus. The structures of the PSII core/RC complex and most of the peripheral light-harvesting complexes have so far only been determined to significantly lower resolution.

Chlorophylls and carotenoids

The most important functional parts of a photosynthetic antenna are the chromophores. All photosynthetic organisms use either chlorophylls (mainly Chl a and Chl b, but also a few other Chls) or bacteriochlorophylls (e.g. BChl a and BChl b) as pigments. The chromophore moiety of the Chls and BChls are magnesium-containing cyclical tetrapyrroles, the so-called chlorins and bacteriochlorins. Figure 1 . 4 ~sh ows the structures of the most common Chls and BChls. All Chls also contain a long ester side chain, which help in anchoring them in their surrounding protein by hydrophobic interactions. In addition all photosynthetic antennae contain carotenoids, which play a dual function as energy donors and photoprotective pigments (for an overview, see Smith, 1991).
 

Because the pigments in antenna systems must be held at a certain range of optimal distances and orientations to each other and to the RCs, all known antenna systems (with one notable exception-the chlorosome structures discussed below) are pigment-protein complexes. The proteins both hold the pigments in a defined position and provide the possibility of fine-tuning their absorption properties by specific pigment-protein interactions. All Chls and BChls, as well as the carotenoids, are attached non-covalently to the proteins by way of either weak ligand-metal interactions of the central Mg of the Chls to amino acid side chains (often histidine), or by other non-covalent interactions such as hydrogen bonding, En-interactions or hydrophobic interactions with their long ester side chains (Smith, 1991).

A few families of photosynthetic organisms make use of additional chromophores and/or other interaction principles in their antennae. A notable, and indeed in all respects exotic, antenna system is employed by the green sulphur bacteria and green gliding bacteria in their chlorosomal antennae. As briefly explained in Chapter 1 and shown in Fig. 1.8, these are extramembranous antennae, located on the inner surface of the cytoplasma membrane, which do not contain proteins in their central part. Instead these chlorosomal antennae consist of huge (several tens to hundred thousands of Chls each) pigment aggregates formed by self-organisation of the special bacteriochlorophylls BChl c, d, e. As Fig. 2.3 shows, these special BChls of green bacteria are characterised by their 3'-OH groups. These are not found in any of the usual Chls or BChls and they enable the molecules to interact closely with each other. Actually these 'bacteriochlorophyl1s'-as they are traditionally called-are chlorophylls, rather than true bacteriochlorophylls in terms of their electronic structures (Smith, 199 1).
 

Another class of photosynthetic pigments is found in the phycobiliproteins of cyanobacteria, red algae and cryptomonads. These are open-chain tetrapymoles which are covalently bound to their apoproteins, typically via S-cysteine linkages (Stanier, 1974; Wildman and Bowen, 1974; Gantt, 1975; Zuber, 1978; Scheer, 1981; Holzwarth, 1986; Holzwarth, 1991; Schaffner et al., 1991; MacColl and Guard-Friar, 1987).

Coherent exciton motion

A t the opposite end of the scale of pigment interactions is the strong-coupling or socalled exciton coupling mechanism, which also occurs through the Coulomb interaction of pigments. In this case the coupling energy is sufficiently strong, typically greater than the vibrational quantum energy h v, that it influences the shape of the absorption spectrum of the pigments involved, replacing the excited states of the individual pigments by a new set of excited states characteristic of the coupled system. Excitation energy is no longer located on a single pigment, but is delocalised over the ensemble of pigments involved in the excitonic coupling. In the simplest case of an excitonically coupled dimer made up from a pair of identical molecules, the new states are obtained by solving the Schrodinger equation for the dimer

where H1,2 are the Hamilton operators for the two pigments, V12 is the coupling energy, E the excited-state energy of the coupled system and Y the excitonic wavefunction. Solution of eq. 2.6 gives the energies and wavefunctions of the excitonically coupled system. This leads to the two excitonic excited states energies

where D is the (ground-state) energy shift due to the change in environment for each pigment (going from the gas phase into the solution environment) and is the excited-state transition energy of the uncoupled pigment. Thus excitonic coupling leads to two new excited states separated by an energy 2VI2. The result is a modification of the absorption spectrum of the dimer as compared to the monomers. The dipole moments dl,z of the two new delocalised exciton states are oriented perpendicularly to each other and depend on the orientations of the transition moments of the monomers according to

 

where do is the dipole strength of the uncoupled monomer and 8 the angle between the transition-dipole moments of the two monomers, thus conserving the total dipole strength of the combined monomers. Where more than two molecules are involved in the excitonic coupling, the relevant equations can be solved either analytically when symmetry prevails, or numerically in the general case (Pearlstein, 1982b and 1982~). While the exciton coupling determines the shape of the absorption, linear dichroism and CD spectra (Pearlstein, 1982b and 1982c), the exciton states do not live for very long at room temperature since the phase relationship of the electronic wavefunctions between the two molecules making up the coupled dimer are disturbed by thermal motions. Typically, dephasing occurs within 1 ps or less. This dephasing leads to a localisation of the excited-state energy on one of the monomers. From then on, Fiirster-type hopping transfer of energy between the two molecules may occur

 

However, during the lifetime of the excitonic states, the energy-transport dynamics are controlled by entirely different equations than for the Forster hopping mechanism, which is particularly important for larger assemblies of excitonically coupled systems. (For a modern in-depth treatment of the exciton concept in photosynthetic systems, see van Amerongen et al., 2000). It turns out that the major part of the energy transfer processes in photosynthetic systems can be quite well described using the Forster mechanism, if appropriate adaptations are made for the calculation of coupling strengths, spectral properties etc. (Yang and Fleming, 2002; Konermann et al., 1997). A rigorous excitonic description only seems necessary for larger assemblies such as the LHI or LHII complexes of purple photosynthetic bacteria, the FMO antenna complex (discussed in Section 2.4.7) and the supramolecular pigment aggregates in the chlorosomes of green sulphur bacteria (Prokhorenko et al., 2000).

 

Forster energy transfer

The light energy that is absorbed in a particular antenna pigment molecule must be transported non-radiatively over relatively large distances of the order of hundreds Angstroms from the location of initial absorption to the RC. As we have already noted, this energy transfer must be very rapid in order to compete with intramolecular deactivation of excited states by processes such as internal conversion and intersystem crossing, which occur in the time range of a few ns. Several mechanisms that in principle allow energy migration through the antenna complexes on this timescale exist. The best known of these is the so-called very weak dipole interaction mechanism, better known as the Forster mechanism (Forster, 1965; Forster, 1959) (see Pearlstein, 1982a and 1982b for early reviews on photosynthetic energy transfer). The Forster mechanism is also known as the ‘hopping process’ of excitation energy transfer, since an excited state migrates through the system in a type of random walk process, albeit one which may be directionally biased depending on the amount of energetic funnelling built into the system architecture. The Forster mechanism is the most abundant mechanism of energy transfer in photosynthesis, being relevant in all photosynthetic systems, except at the shortest time scales where other mechanisms (see below) may apply.
 

The basis for the Forster mechanism is the coupling energy between the transition dipoles of the donor and acceptor molecules and an energetic restriction that is dealt with by the so-called overlap integral in the Forster equation. The advantage of the Forster formulation of energy-transfer rates is that it is directly based on easily accessible experimental quantities such as the absorption coefficient of the donor, the emission spectrum I F of the acceptor, the distance R between them, and the relative orientations of the donor and acceptor as reflected in the so-called orientation factor K* of the transition dipole moments of the donor and acceptor. The rate constant kET of a single energy-transfer step is given by

where n is the refractive index of the medium, 4~ the fluorescence yield, rF the lifetime of the donor molecule without energy transfer, and v the frequency. The part of the equation under the integral is the spectral overlap factor. Due to the inverse R6 dependence of the energy-transfer rate constant, the Forster mechanism allows energy transfer over relatively large distances, well up to 100 8, and beyond, depending on the molecular transition dipoles involved. However, for the fast transfer required in photosynthetic antenna, where single-step transfer times should typically be shorter than 1 ps, the typical maximum allowed distances R for Chl pairs are up to 15-20 A, depending somewhat on orientation.
 

The Forster mechanism assumes that both the donor and the acceptor molecules are in thermal equilibrium with the environment and that the interaction energy between the dipoles is very small compared with the energy of a typical molecular vibration. This means that the coupling does not influence the absorption spectra of the involved pigments in any appreciable fashion as compared with the uncoupled system. The mechanism allows electronic singlet-to-singlet, and also triplet-to-singlet, energy transfer. Transfer from a singlet to a triplet state is, however, not possible because the overlap integral is negligible (the ground-state-to-triplet absorption probability lying close to zero).
 

The kinetics of energy migration in the case of FiSrster transfer in the very weak coupling limit2 are described by the master equation

where pi is the probability of the excitation being located on pigment i, k"' is the rate of loss processes other than energy transfer, k, is the Forster rate of energy hopping from pigment i to pigment j , kRC is the rate of charge separation at the RC and PRC the probability of the RC being excited. Using this equation and the known Forster rate constants, the overall energy-transfer dynamics in a complex antenna system can in principle be fully calculated. However, the problem with using eq. 2.5 is the uncertainties in the Forster rate constants, which depend critically on the often-

 

unknown spectral properties of the individual pigments and on the distances and relative orientations of the pigments.
 

Due to the very large number of antenna chromophores comprising a particular antenna system, it is often not practical or possible to take into account all the individual chromophores. Rather, groups of pigments or chromophores with similar spectroscopic and kinetic properties are lumped together in a pseudo-pigment complex, called a 'compartment' when interpreting antenna energy-transfer spectra and kinetics. Such a 'compartment model', representing the general photosynthetic unit shown in Fig. 2.1, is given in Fig. 2.2. The 11 chromophores present in the antenna are lumped together into three different antenna compartments, which would differ in their spectra and kinetics. Very often such a compartment can be identified with a particular biochemical subunit of the antenna, although this is not necessarily always the case.

Why are aruema syslerris necessary?

One may well ask why photosynthetic antenna systems are necessary for the functioning of photosynthesis. After all, the reaction centres themselves contain several chromophores that absorb visible light, and they are in principle fully capable of performing the processes of light absorption and charge separation by themselves. However, all photosynthetic organisms have developed photosynthetic antennae, thereby dramatically increasing the effective absorption cross-section of their RCs. The reason is that the complex machinery of an RC would not be able to work at an optimal rate and yield under typical sunlight conditions (we are thinking here of higher plants located at Earth’s surface: conditions for most photosynthetic bacteria would be even worse). A bare RC would have a light-limited turnover rate of 1- 10 min-I, but it is capable of operating at turnover rates of up to a few hundreds per second. Thus an RC without its antenna would use only a small fraction of its photosynthetic capacity. Moreover, the limitation of the absorption cross-section would further reduce turnover rates by additional quantum losses in the intermediate charge-separated states, which have a finite lifetime before recombining to the ground state. This can only be prevented if a second photon, leading to another turnover, is absorbed within a short time. Thus an increase in the light-limited turnover rate by a factor of at least several hundred as compared with the bare RCs would be optimal. This has been achieved by the development of light-harvesting antennae that increase the effective absorption cross-section of the RC by factors of up to several hundred. Moreover, antennae systems coupled to RCs can harvest a larger bandwidth of the solar spectrum by combining pigments that absorb at different energies.
 

One might conclude that very large antenna sizes would be the most desirable. However, there are several limitations on maximal antenna size, because it is essentially limited by the relative rates of energy migration through the antenna to the RC and the rate of charge separation in the RC. The larger the antenna system, the longer the average time for the arrival of excitation energy at the RC (this is often called the ‘first passage time’ of the antenna), a:id thus the larger the energy lost through competing processes in the antenna such as fluorescence and radiationless decays. Furthermore, larger antennae (other than the so-called diffusion-limited PSUs) are typically associated with longer average times for charge separation, which also increases the probability that loss processes will occur.
 

Let us consider some basics in order to gain insight into the principles. We can distinguish two extreme cases of antenna kinetics: ( I ) The so-called eriergy-trurzsferlimited case, where the overall rate-limiting step is energy transfer through the antenna to the RC. In this case, the intrinsic electron-transfer step in the RC is very fast (faster than the first passage time); and (2) the so-called trap-limited case, where energy transfer through the antenna is faster than charge separation in the RC. This leads to quasi-equilibrium between excited antenna and excited RC chromophores, such that energy migrates back and forth between antenna and RC several times For case (l), the average time zET for energy transfer from a regular lattice of identical chromophores to the RC can be estimated asbefore charge separation occurs.

where N is the number of antenna chromophores (the antenna size), and z,, the time of a single pairwise energy-transfer step. For a typical z,, of -100 fs, the overall transfer time is -50-100 ps for antenna sizes N of 200-300 and antenna pigments with excited-state lifetimes of -1-3 ns. Under these conditions the yield of the photosynthetic process is limited to a maximum of about 90%; the rest of the excitation energy would be dissipated in the antenna as heat and would be lost. However, the yield is well above 90% in most photosynthetic systems. Thus for the transfer-limited case the maximal antenna size is expected to be -200 chromophores per RC, or somewhat higher in the case of energetically heterogeneous antennae.
 

A similar limitation on antenna size arises, albeit for different reasons, in the other extreme case, case (2), that of trap-limited kinetics. In this case, energy transfer is assumed to be very fast (although there are limits on the maximal rates), but the excitation migrates through the system from pigment to pigment in a random, hopping process. If we assume that the excited-state energy of the RC is identical to that of the antenna pigments, the probability of the excitation being located on the RC becomes smaller as the antenna size increases for statistical reasons. This reduces the effective rate k,, of charge separation in the system, according to

Again using typical values for the deactivation processes in the antenna, it becomes clear that k,, must be at least 2 ns-', corresponding to an overall charge separation lifetime of about 500 ps, in order to achieve a quantum yield of more than 90%. This again limits the maximal antenna size to about 200 chromophores if severe losses in overall quantum yield are to be avoided. For heterogeneous antenna systems with chromophore energies higher than the RC energy, the maximal antenna sizes are somewhat higher.

 

The numbers resulting from these simple considerations agree very well with actual antenna sizes of higher plants, which range from about 100 up to a maximum of 250-300 chromophores per RC. However, in organisms which contain highly heterogeneous antenna systems, such as green photosynthetic bacteria, the number of chromophores per RC can increase to several thousands, because excitation is funnelled to the RC rather than encountering it by random hops.

The photosynthetic unit

Before delving into the details of specific light-harvesting systems in photosynthetic organisms, we will first consider the general concept of a photosynthetic unit (PSU). As Fig. 2.1 shows, a PSU consists of an array of light-absorbing chromophores

arranged around a reaction centre. The antenna chromophores may be arranged regularly or irregularly in space, as we shall see later when discussing specific systems. There exist two essential requirements for the functioning of an antenna system. These are: (1) the distances between the chromophores must be short enough in order to allow sufficient electronic interaction for fast, efficient transfer of energy to occur between individual chromophores; and (2) the ordering of their excited-state energies must be such that there is sufficient energetic overlap between adjacent chromophores for energy transfer to occur at physiological temperatures. While there exist additional limitations for an efficient antenna system, the presently known structures tell us that these essential requirements can be satisfied by a very wide range of spatial arrangements of pigments, pigment compositions erc. As we shall see when we discuss specific antennae systems, typical interchromophore distances are in the range 7-15 A, and the number of antenna chromophores in a photosynthetic antenna unit ranges from a minimum of 32 in the LHI antenna of purple photosynthetic bacteria to several hundred per reaction centre in higher plants and algae, and possibly even higher in green photosynthetic bacteria.

Energy and chemicals from biomass

Biomass was one of humanity’s earliest energy sources, and it remains an important resource today and for the future. Traditional plant biomass, particularly fuelwood, currently supplies a significant part of the energy needs of the developing world. Used and regrown sustainably, such energy sources could become an important component in a future COz-neutral energy economy. In the USA, biomass provides nearly 4% of final energy consumption. In the EU, biomass provided 7% of total primary production in 1997. Methane produced by landfill wastes is used to generate electricity as a matter of good practice in many countries.
 

Plant biomass resources include wood and wood wastes, agricultural crops and their residues, and aquatic plants and algae. Energy can be derived from biomass in three main ways:

• by direct combustion to provide heat and light as such, or to raise steam and hence generate electricity;  
• by gasification to provide ‘biogas’, a combustible gas mixture predominantly consisting of H2, CO and C02, that can be used for heating or electricity generation or converted to useful chemicals such as methanol;
• by fast pyrolysis at temperatures around 480-550 C, giving a high yield of ‘biooil’, an espresso-coffee-like liquid that can substitute for conventional fuel oil in transport and static applications

Wood is substantially the largest source of biomass energy and indeed of renewable energy, providing more than twice the contribution of hydroelectricity worldwide. Energy crops (herbaceous plants and forest plantations grown specifically for their energy content) and biofuels'' (solid, liquid or gaseous fuels made from
biomass) are established or emerging features in several parts of the world. At the moment, biofuels and electricity made from biomass are expensive compared with conventional alternatives, and their use generally needs to be stimulated by subsidy or regulation. Future cost improvements should come from volume production, the development of more efficient chemical processes and the production of value-added chemicals.
 

The chemical composition of dry biomass varies somewhat with species but is roughly 75% carbohydrates or sugars and 25% lignin, with the empirical formula C3H4O2. The main carbohydrates are cellulose-a polymer of glucose and the single most abundant product of photosynthesis-and hemicellulose. Because of their oxygen content, biomass and biofuels have a much lower calorific value (- 18 GJ tonne-' @ 100% dry matter) than conventional fuel oils (4244 GJ tonne-'). Biomass generally has low (<0.1%) sulphur content. Its nitrogen content depends on the protein content, which should be kept low to minimise the emission of NO, on combustion.
 

Wood and straw are widely used as fuels. However, most other plant products are not suitable to be directly burned because they are too wet. Even fresh wood contains a considerable amount of water (30-50% by weight) and is best dried before use. Charcoal, produced by slow pyrolysis of wood in limited air, is used as a domestic and industrial fuel. In Brazil, the world's largest charcoal producer and consumer, charcoal is used in heavy industries such as pig-iron, steel making and cement manufacture.
 

Mike Bullard discusses energy crops in Chapter 9. These are (generally perennial) species that grow rapidly and can be harvested for their biomass annually or every few years. In northern Europe, the most promising energy crops are perceived to be coppiced willow or poplar, often referred to as short rotation coppice (SRC) or arable energy coppice (AEC). Other energy crops include conventional arable crops, novel annual crops such as sweet sorghum and perennial species such as elephant grass (Miscanthus). Although a native of Asia and Africa, Miscanthus grows well in temperate climates. Its C4 photosynthetic pathway permits biomass yields of up to 55 t ha-' in the UK (compare wheat, for example, with a theoretical maximum UK yield of 33 t ha-'). Field trials of such energy crops have been underway for some years in Europe and the USA. In developed countries, the economic viability of energy crops is generally marginal, or even poor at times of low energy prices. However, the growth of energy crops on idle or set-aside farmland could represent a major future opportunity if robust supply chains and markets can be established and costs held down.
 

Tony Bridgwater and Kyriakos Maniatis discuss biofuels in Chapter 10. The most widely used biofuel is ethanol, made by fermentation of starch crops. Brazil and USA have pioneered large-scale ethanol fuel programmes. Nearly 1.5 billion gallons of ethanol are produced from corn annually in the USA to blend with gasoline and reduce air pollution. Brazil began its Pro-Alcohol programme, making ethanol from sugar cane, to counter the oil price hikes of the 1970s and its dependency on imported oil. It is by far the world's major producer of cane alcohol, producing 10-15 million cubic metres annually and with significant export opportunities now opening up.
 

Biodiesel is another significant biofuel. Chemically, this consists of a range of fatty acid alkyl esters. It can be made from any vegetable oil, produced from any oilbearing crop or microalga, by esterification with ethanol or methanol. The properties of biodiesel are quite similar to those of conventional diesel oil, enabling it to be used neat or in blends in conventional diesel engines. Rape methyl ester (RME) produced from oilseed rape is the main form of biodiesel in Europe and Canada. The USA produced about 5 billion gallons of biodiesel in 2000 from recycled cooking oils and soy oil.
 

Hydrogen, an important energy vector of the future, can be produced biologically under certain growth conditions by a range of photosynthetic microorganisms that contain either the hydrogenase enzyme or the closely related nitrogenase enzyme. Some strains of cyanobacteria have developed intracellular protective mechanisms that enable them to do this in air. Photosynthetic bacteria and algae have not developed protective systems, and can produce H2 under anaerobic but not aerobic conditions. Historically, these cultures have been easily poisoned and short-lived, but recent cyanobacterial cultures are encouragingly robust and other tricks are being used to enhance H2 evolution from green algae. Boichenko, Greenbaum and Seibert discuss progress in Chapter 8.
 

Many other chemicals and products can be derived from biomass. Cellulose fibres from wood are used to make paper and textiles, natural pharmaceuticals can be extracted from many plants. In Chapter 7, Rosa Martinez and Zvi Dubinsky discuss the range of products that can be derived from algae.
 

Statistics on the use and trends in use of biomass for energy generation are needed in energy and COz emissions modelling, particularly where developing countries are moving from traditional biomass use to fossil fuel use. Unfortunately, reliable data do not generally exist. Traditional biomass (fuelwood and crop and animal residues) is either collected by the user or traded in highly informal markets. The volume of these transactions is not metered and can only be roughly estimated from spot consumer surveys. Statistics on commercial use of biomass in forestry, paper and sugar industries and large-scale CHP and power generation are better.
 

Table 1.1 showed what are probably still the most reliable statistics on global and regional biomass use, derived from two workshops held by the IEA in 1997 and 1998. According to this survey, biomass accounts for -1 1% of world total primary energy supply (TPES) and -14% of world total final consumption (TFC). Nearly all of this consumption is in non-OECD (developing) countries, where biomass comprises -20% of TPES and -27% of TFC. (Most biomass is simply burned in such countries, rather than being converted to derived biofuels or used to generate electricity, so the proportion of biomass in final energy use is higher than in primary supply.) The main sources of national biomass data are the UN’s Food & Agriculture Organisation (wwwf.a o.org) and Energy Statistics Division and, for Asia, the Center for Energy-Environment Research and Development (www.ceerd.ait.uc.tW). The FA0 (www.fao. org/waicent/faoinfo/forestry/energy/) has made considerable efforts to gather statistics on forest woodfuel, through its series of publications Wood Energy Today for Tomorrow (FAO, 1997-1999). Other useful sources of information are the latest biennial survey of the World Energy Council (WEC, 2001), the world biomass scenarios of Hoogwijk et al. (2001) and the final paper by that most indefatigable of champions of biomass, the late Professor David Hall of King’s College, London (Hall et al., 2000).

Efficiencies achieved in wild and cultivated crops

The time-average energy-storage efficiency of green-plant photosynthesis is much lower than the instantaneous maximum values of -5% calculated in the previous section, for a number of obvious reasons. For example, the growing season is limited, the plant canopy does not intercept all incident sunlight, light levels may be too low or too high for maximum photosynthetic efficiency, and plant growth may be inhibited by factors such as water or thermal stress.
 

The global energy-storage efficiency of photosynthesis can be calculated from knowledge of net primary production (NPP), the mass of carbon fixed annually by photosynthesis. Global NPP is -100 GtC yr-I, about half on land and half in the oceans (see Section 6.5). The molar free energy of the photosynthesis reaction (eq. 1.6) is 496 kJ mol-I, which corresponds to a specific energy-storage capacity of 41.3 kJ g-’ fixed carbon. Global NPP of 100 GtC yr-’ thus corresponds to -4 x 10” J of chemical energy stored in photosynthetic biomass per year. J yr-’. Thus the net efficiency of photosynthesis averaged over Earth’s surface (land and oceans) is -0.15%. About half of the incoming solar energy is in the PAR range of 360-720 nm, so the energy efficiency of PAR utilisation is -0.3%. Modest as these efficiencies are, the amount of energy stored annually by photosynthesis is about ten times greater than current world energy consumption.
 

Agriculture-the cultivation of plants for food-arose in the Fertile Crescent about 10,500 years ago. Traditional methods of plant breeding by the selection and crossing of species with favourable characteristics (for example, hardiness and plant size) have hugely improved food crop yields since then. Energy -storage efficiencies of 0.5-1.0% on an annual basis are typical in modern food crops, and short-term yields can be as high as -4%. Cd plants, with their modified C02 fixation pathway, have considerably higher efficiencies than C3 plants, especially in tropical and subtropical areas where their growth rate is less likely to saturate under high light levels. In future, global warming may extend the geographic range of some crops and trees to higher latitudes, and increased CO1 levels may increase growth rates.

It is not widely appreciated that traditional plant-breeding techniques were effectively an imprecise form of genetic engineering: plants that have favourable characteristics and are therefore selected for breeding have favourable genotypes, which are therefore selectively replicated in future generations of plants. About half of past improvements in yields of rice, wheat and maize is due to genetic inputs. Recently, plant-breeding methods have been extended by two new ‘genetic’ techniques: tissue culture, which allows the crossing of favourable genotypes at cellular level to form new cultivars, and genetic modification, which involves the incorporation of individual genes directly into plant genomes. Despite the current outcry about GMOs (genetically modified organisms), a significant part of future improvement is likely to come from transgenically improved plants. Denis Murphy describes the enormous potential of ’agbiotech’-the application of genetic techniques to improve food and non-food crop traits and yields in Chapter 13.

Net eflciency allowing for respiration

The calculated gross energy-storage efficiency of -9% of a green plant can never be achieved in real life because all photosynthetic organisms must constantly consume a portion of their stored energy in the process of respiration to obtain the energy to stay alive. Respiration effectively reverses oxygenic photosynthesis, and so reduces the energy-storage efficiency to a net value below the gross value.
 

There are two types of respiration-dark respiration and photorespiration, the latter occurring only in the light. The rate of dark respiration in green leaves lies in the range 0.5-4.0 mg C02 drn-2 hr-l at 25 C (Zelitch, 1971). This has only a small effect on the efficiency in bright sunlight, subtracting perhaps -0.2% from the gross efficiency. Photorespiration, on the other hand, is responsible for a much more serious loss of fixed carbon. In plants using the normal C3 carbon fixation cycle, photorespiration occurs at such a rate that some 30% (at 25 C) to 40% (at 35 C) of the gross yield of photosynthesis is lost. Plants using the C4 cycle lose rather less.

The combination of dark respiration and photorespiration reduces the calculated maximum net efficiency of photosynthesis to a value between 5.3% at 35 C and 6.2% at 25 C (Bolton, 1979). This agrees well with other estimates: 5.3% (Bassham, 1976); 5.5% (Hall, 1977); 5% (Boardman and Larkum, 1975). This is the expected instantaneous maximum efficiency for a healthy leaf growing in optimal conditions. There are a number of factors, explored in the next section. that reduce time-average values well below this, although short-term values can approach 5%.

Figure 1.12 shows in another way how the upper bound of -5% on the energystorage efficiency of a green plant comes about. Nearly half (47%) of the solar energy incident on a plant is lost because it lies outside the photosynthetically active range of 40&700nm. A further 16% is lost by incomplete absorption of PAR (Photosynthetically Active Radiation) or by its absorption by components other than the chloroplast. A further 9% is lost by thermalisation-the degradation to heat of the ‘excess‘ energy of absorbed photons of wavelength below 700 nm. that is, energy above 1.77 eV, which is the threshold or ‘bandgap’ energy Us of P700. A further substantial loss of 19% arises because the synthesis of D-glucose stores only the fraction (AH/8U8) of the energy of eight thermalised P700* states. That leaves only -5% of the incident solar energy to be stored as chemical energy. This is still a formidable value; the instantaneous energy-storage capability of a green leaf in the Sun leaves most artificial molecular photoconverters of solar energy in the shade.

Gross efficiency ignoring respiration

Green plants contain carbon in a reduced state, mainly as carbohydrates. The chemical energy stored in these compounds is released when they are metabolised in the living plant, or when the plant biomass is burned or otherwise oxidised to C02 and H2O. Since plant biomass is mainly comprised of D-glucose polymers, the energy

released when biomass is oxidised to C02 and H20 is roughly equal to the enthalpy of synthesis of D-glucose from COz and H20. Under normal atmospheric conditions, we can write this reaction as

where C6H1206 is D-glucose. For this reaction, AH = 467 kJ and AG = 496 kJ at 298 K; the standard values are AH0 = 467 kJ and AGO = 480 kJ (Bolton, 1979). Neglecting for the moment the inevitable loss of some biomass in a living plant by the process of respiration, we can define the photosynthetic energy (enthalpy) storage efficiency %s of a green plant, acting as a photoconverter of solar energy, as

To calculate 77ps for sunlight of a given spectral distribution, Schneider (1973) and Bolton (1979) rewrote eq. 1.7 as

where N A is the Avogrado constant and, for the wavelength band h to h + dh, j: is the incident solar spectral photon flux (photons m-2 s-' nm-I), E: (W m-2 nm-l) is the incident solar spectral irradiance ah is the spectral absorptivity (fraction of light absorbed by the plant), 41. is the quantum yield for the production of 0 2 or consumption of COz, AH is the enthalpy of the photosynthesis reaction, and h,,,in and A,,, are the minimum and maximum wavelengths that effect the reaction,. Using experimental data at 10 nm intervals for a k , j ; , E: and 4, Bolton (1979) calculated rps from eq. 1.8 and obtained a value of (9.2 k 0.8)%. This is the upper bound on the gross efficiency of energy storage in a healthy growing leaf, ignoring respiration. In using experimental values of ah and 4, it takes account of the dip in ah values in the green and 4 values towards the red. If 4 were to maintain the 'ideal' value of 0.125 (or 8 photons per O2 molecule evolved) from 360 nm to 700 nm and then drop off as the experimental values from 700 nm to 720 nm, the gross efficiency would rise even higher, to 13.3%.

Carbohydrates

The immediate end product of photosynthesis is the monosaccharide D-glucose. Fig. l.lla shows its structure in its usual cyclic form. Both plants and animals break down this simple sugar to obtain energy via the metabolic process known as glycolysis. The end product of this depends on the nature of the organism and whether oxygen is present. In green plants in normal, aerobic conditions, glycolysis proceeds (via the citric acid cycle) to form the fully oxidised products C02 and H20, and theenergy released by this process, when coupled with respiratory electron flow, drives the synthesis of 36 molecules of ATP (adenosine triphosphate), the energy-carrying molecule that is found in the cells of all living organisms.
 

Glucose that is not immediately required by a photosynthetic organism is polymerised, to provide both the oligosaccharides often associated with lipids and proteins and the polysaccharides that constitute the main structural materials and nutritional reservoirs of plants. Cellulose (Fig. 1.1 lb) is the most abundant structural material, constituting about half the cell-wall material of wood and higher plants, and accounting for over half of all the fixed carbon in the biosphere. It is a linear polymer of up to 15,000 D-glucose residues held by hydrogen bonds in a rigid assembly of great strength. Glycogen (Fig. 1.1 lc), the chief food reserve of photosynthetic bacteria and animals, is a branched polymer of D-glucose residues. Starch, which is the main food reserve of plants as well as a major nutrient for herbivorous animals, is a mixture of the two polysaccharides a-amylose (an isomer of cellulose) and amylopectin (similar to, but more highly branched than, glycogen).

Energy storage efficiency of photosynthesis

Photosynthesis is the only natural process able to store a significant amount of solar energy as chemical energy in biomass: terrestrial plants, particularly trees, are the main repositories. However, nature has not entered photosynthesis for any energy efficiency awards-the imperative for any photosynthetic organism is replication, not the accretion of biomass. Nonetheless, if the energy-storage process were not adequately efficient. it would not serve this primary purpose.
 

In our context, the energy-storage efficiency of photosynthesis is of course of great interest. It determines the flux of energy into the biosphere, the land area required to produce a given number of food calories, and the biomass yield from a given area of an energy crop plantation. In this section, we first look at the structures of the carbohydrates that are the main energy-storage compounds of photosynthesis, and then the maximum gross and net efficiencies permitted by the characteristics of the photosynthesis reaction, and finally the energy-storage efficiencies actually achieved in the wild and in cultivated crops.

The dark reactions of photosynthesis

The sequences of reactions by which C02 is reduced to carbohydrate are sometimes referred to as the 'dark' reactions of photosynthesis because C02 can be fixed in the dark by a leaf or photosynthetic organism if the appropriate reagents are available. The dark reactions take place separately from the light-driven reactions in the stroma or cytoplasm, as indicated in Fig. 1.10. Electrons from the light-driven process in the thylakoid membrane reduce either nicotinamide adenine dinucleotide (NAD') or its phosphorylated form (NADP'), and the reduced forms NADH or NADPH provide the reducing power for CO2 fixation, with the help of some additional free energy in the form of adenosine triphosphate (ATP) generated by photosynthetic phosphorylation. There are several mechanisms of C02 reduction, characteristic of different photosynthetic species. The reductive pentose cycle or C3 cycle (so called because the

first 'stable' product of C02 reduction is a three-carbon compound) is the commonest mechanism, operating in algae and most plants. Some plants, especially those indigenous to hot climates, such as corn (maize) and sugar cane, operate the C4 cycle. Edwards and Walker discuss these and other carbon fixation cycles such as the CAM cycle in Chapter 4; Blankenship (2002) provides a full account.

Reaction centre structures

Figure 1.9 shows how PSI and PSII are functionally coupled with cyt 6 6 and ATP synthase in the thylakoid membrane. We now know for certain from x-ray crystallographic studies that all reaction centres are characterised by a pseudo-2 fold symmetry axis that relates the cofactors and the proteins that bind them. In Type 11 RCs, this symmetry gives rise to a redox-active branch and an inactive branch, as shown for PSII in Fig. 1.9. Despite intense studies on the purple bacterial RC it is still not clear how Type I1 centres are able to differentiate their active and inactive branches. However, this property has distinct advantages when the terminal acceptor (i.e. QB) requires two electrons to be fully reduced. In Type I reaction centres, where a single, centrally located iron-sulphur centre Fx is the electron acceptor (as for PSI in Fig. 1.9), it seems possible that primary charge separation occurs with similar probability up either branch; this must be so in green sulphur bacterial RCs, which are homodimeric while in the case of PSI the situation is less clear. The two protein subunits that constitute the RCs of PSI, PSII and purple bacteria are not identical, as in the case of green sulphur bacteria, but form a heterodimer. In purple bacteria, the two subunits are called L and M, while in PSII the closely related

subunits are known as the D1 and D2 proteins (see Figs. 3.3 and 3.10). All four proteins show considerable homologies, and all have five transmembrane helices related to each other in their reaction centres by the same pseudo-2 fold axis that relates the cofactors.
 

The two proteins that make up Type I reaction centres, PsaA and PsaB, are also arranged around the pseudo-2 fold axis that relates the cofactors (see Fig. 3.8), but in this case they have eleven transmembrane helices. Interestingly the five transmembrane helices at the C-terminal ends of these Type I RC proteins are arranged in a similar, but not identical, manner as in Type I1 RCs (Rhee et al., 1998; Schubert et al., 1998).
 

The structural details briefly described above have emerged from X-ray crystallographic studies which began with the elucidation of the structure of a Type I1 RC isolated from the purple bacterium R. viridis in the 1980s by Deisenhofer, Huber, Michel and colleagues (Deisenhofer et al., 1984, 1985) and have recently advanced to the determination of the structure of PSI at 2.5 (Jordan et al., 2001) and PSI1 at resolutions ranging from 3.8 8, to 3.5 8, (Zouni et al., 2001; Kamiya and Shen, 2003; Ferreira et al., 2004). As discussed in detail in Chapter 3, these studies have given a structural basis for the interpretation of data obtained by a variety of spectroscopic techniques and for developing general theories of electron transfer in proteins. Moreover, the determination of the structures of the cytochrome bc (Iwata et al., 1998; Zhang et al., 1998) and ATP synthase (Abrahams et al., 1994) complexes of the respiratory membranes allows realistic structural extrapolations to the corresponding complexes of photosynthesis.

Energetics of electron transfer processes in reaction centres

Before discussing the structural and functional properties of the reaction centres of different types of photosynthetic organisms, it is necessary to appreciate their specific electron-transfer pathways in terms of redox potentials. Figure 3.2 compares the Type I and Type I1 reaction centres of anoxygenic and oxygenic organisms. In purple photosynthetic bacteria (specifically R. sphaeroides), the primary donor is called P870, because the long wavelength absorption peak of its special pair of bacteriochlorophylls is at 870 nm. Similar notation is used for other primary donors e.g. P840 (green sulphur bacteria), P870 (green non-sulphur bacteria), P700 (PSI) and P680 (PSII). However, as hinted in Section 1.1.4, P680 differs from the other primary electron donors in that it seems not to be a special pair (Barber and Archer, 2001).
 

As we noted in Section 1.3.1, when excitation arrives at the RC from the LH system, primary charge separation occurs and this is followed by secondary electron flow to a terminal electron acceptor, ferredoxin (Fd) in the case of green sulphur bacteria and PSI (Type I RCs), or quinone (QB), in the case of purple bacteria and PSII (Type I1 RCs). In the case of anoxygenic bacteria, some of the reducing potential is used to convert NAD' to the NADH needed for C 0 2 fixation and some is utilised in cyclic electron flow, whereby the reductant indirectly reduces the oxidised primary donor. This cyclic electron flow involves the cytochrome bc complex, which is also embedded in the chromatophore membrane and which couples the electron flow to the vectorial movement of protons across the membrane, as shown in Fig. 3.6. The resulting pH and electrical gradients are then used to drive the conversion of ADP to ATP in accordance with the chemiosmotic mechanism of Peter Mitchell (1966), a contribution for which he received the Nobel Prize for Chemistry in 1978.
 

As already mentioned and shown diagrammatically in Figs. 1.4, 1.7 and 3.2, Photosystem I and Photosystem I1 work together in oxygenic photosynthetic organisms to oxidise water and reduce ferredoxin. PSII functions as the waterplastoquinone oxidoreductase while PSI is a plastocyanin-ferredoxin oxidoreductase. The redox coupling between the two reaction centres is accomplished by a cytochrome bc complex rather like that found in anaerobic photosynthetic bacteria but called, for historical reasons, the cytochrome b6f complex. An important feature of this scheme is that two photons are used to drive one electron from water to ferredoxin. The cytochrome b6f complex acts as a plastoquinol-plastocyanin oxidoreductase and, like its counter part in photosynthetic bacteria, facilitates the maintenance of the electrochemical potential gradient of protons across the thylakoid membrane needed to convert ADP to ATP. In oxygenic photosynthesis, the reduced ferredoxin is used to convert NADP' to NADPH, which together with ATP is required to convert C02 to carbohydrate.
 

Green sulphur bacteria also use reduced ferredoxin in the same way as PSI except that they, like purple bacteria, use non-phosphorylated nicotinamide adenine dinucleotide (NAD') rather than NADP'. The similarity in the redox properties and electron transport pathways of the Type I and Type I1 RCs is evident in Fig. 3.2 except for the important fact that P680' is a much stronger oxidant (with a midpoint potential of -1 V) than P700', P840'and P870'(-0.4 V). This is because P680'must be sufficiently oxidising to remove electrons from water, which is a very stable molecule and difficult to oxidise compared with the substrates oxidised by other reaction centres. This oxidation reaction involves a cluster of 4 Mn atoms and the transfer of electrons and protons from the substrate water molecules is facilitated by a redox-active tyrosine, named Yz, positioned between the (Mn)4-cluster and P680 (see Fig. 1.7). As the production of dioxygen from water is a four-electron process

and a dioxygen molecule is produced at a single PSII reaction centre, the Mn cluster must accumulate four oxidising equivalents. This is why the evolution of 0 2 oscillates with a period of four when oxygenic organisms are subjected to single turnover flashes of light, as discovered by Pierre Joliot and colleagues in 1969. This discovery caused Kok et al. (1970) to propose the S-state cycle, whereby the absorption of four successive photons drives the series of reactions

When S4 is formed, dioxygen is released and the cycle resets itself to the So-state. Although the precise chemical mechanism of the S-state cycle is unknown, it is generally believed that the two water substrate molecules bind at the So-state and that H' and electrons are extracted before arriving at the S4-state. The late Jerry Babcock and colleagues (Tommos and Babcock, 2000) have suggested an attractive 'hydrogenatom abstraction' hypothesis for the water oxidation mechanism.
 

Not surprisingly, the high redox potential of P680' and the possibility of forming reactive oxygen species during the water-splitting reaction give rise to oxidative damage of the PSII RC. This manifests itself as rapid degradation and regular replacement of protein, as Godde and Bornman describe in Chapter 5. Plants and other oxygenic organisms have evolved a range of protective strategies that reduce the frequency of photoinduced PSI1 damage and allow the repair process to cope under normal conditions. The effect of this intrinsic and detrimental property of PSII is, however, observed when organisms are exposed to environmental stress, when the rate of repair does not match the rate of damage and photoinhibition occurs. When this happens, the efficiency of photosynthesis and biomasskrop productivity decline.

Photosynthetic membranes

The reaction centres of purple and green sulphur bacteria are localised in membranes, often called chromatophore membranes, which lie close to or include the outer cell membrane. In purple photosynthetic bacteria. the LH proteins are also intrinsic to the chromatophore membrane. However, in green bacteria the very large LH chlorosome, packed with many thousands of molecules of bacteriochlorophyll, is stacked into rodlike structures attached to the cytoplasmic side of the photosynthetic membrane, which does not invaginate as it does in purple bacteria (Fig.l.8)

In oxygenic photosynthetic organisms, the photosynthetic apparatus involved in light reactions is embedded in the specialised thylakoid membrane (see Fig.l.3). In cyanobacteria, the thylakoid membranes tend to form concentric rings within the cytoplasm and are characterised by the presence of the large LH phycobilisomes attached to their surfaces, which induces a considerable spacing between them. In the green oxyphotobacteria (prochlorophytes), the same concentric rings are present but the membranes lie more closely together because of the absence of bulky phycobilisomes. The presence of phycobilisomes in the chloroplast of red algae leads to a thylakoid membrane organisation reminiscent of cyanobacteria. In striking contrast, the thylakoid membranes of higher plant chloroplasts, and to a lesser extent those of green algae, are arranged in stacked (grana) and unstacked regions (see Fig.l.3). The granal thylakoids are highly enriched in PSII, while PSI is found in the unstacked regions. However, this extreme lateral separation does not seem to occur in the thylakoid membranes of cyanobacteria and many forms of algae and therefore cannot be an absolute requirement for oxygenic photosynthesis to occur.

Light harvesting system

As Alfred Holzwarth explains in detail in the next chapter, photosynthetic organisms have evolved light-harvesting (LH) antenna systems that service photosynthetic reaction centres so that they can operate efficiently under relatively low light intensities. The nature of these LH systems varies considerably according to the type of organism, but all function to intercept light and transfer the excitation energy rapidly to the reaction centre. The process is efficient, so the overall transfer rate must be faster than the singlet lifetimes of the pigments, which are typically in the nanosecond time domain. In fact, overall transfer times of energy migration from the

LH system to the RC are in the sub-nanosecond time domain, and in most cases transfer seems to occur by the Forster resonance mechanism. This requires good overlap between the absorption and emission spectra of the pigments, location of each pair of energy donor and acceptor pigment molecules to be close (typically within 10- 15 8, centre-to-centre) and with appropriate orientations. To achieve these properties, the pigment molecules are bound to a protein scaffold and these pigment-proteins associate with the RC. The number of light-harvesting pigmcnt molecules servicing an RC varies according to the type of organism and the growth conditions, from 50 (in some purple photosynthetic bacteria) to many thousands (as in the case of the chlorosome of green sulphur bacteria). In plants and algae, the number is around 250 pigment molecules per reaction centre.
 

The LH system and RC together comprise the photosynthetic unit. In the case of higher plants and green algae, the pigments bound to LH proteins are chlorophyll a, chlorophyll b and carotenoids. In addition to chlorophyll a and carotenoids, red algae contain the phycobilin pigments that covalently bind to protein to form the phycobilisomes. large macromolecular structures that attach to the outer (stromnl) surface of the photosynthetic membrane. Cryptomonads also contain phycobilins but they do not associate to form phycobilisomes and are located on the other sidc of the membrane.

Like red algae, brown algae, dinoflagellates and diatoms do not contain chlorophyll b, but differ again in that they contain chlorophyll c as an LH pigment as well as chlorophyll a and carotenoids. Cyanobacteria also do nut contain chlorophyll 6. but like red algae they contain phycobiliproteins that assemble into phycobilisomes. However, as mentioned in Section 1.2, there are related prokaryotic organisms (oxyphotobacteria) known as prochlorophytes that do not contain phycobilins but instead have an LH system composed of chlorophyll a and chlorophyll 6. In contrast, the purple and green sulphur bacteria contain different forms of bacteriuchlorophyll and carotenoids.
 

The photosynthetic unit is a marvellously tuned sunlight-gathering apparatus. The different spectral properties of the wide range of LH pigments. coupled with finetuning of the IR spectra by interactions with the proteins to which they bind, allow photosynthetic organisms to absorb at all the wavelengths available in the solar spectrum at the Earth’s surface (35C-I000 nm).