738

Femtosecond spectroscopy of photosynthetic light-harvesting systems Graham R Fleming* and Rienk van Grondellet

Observing the elementary steps of light-harvesting in real time is now possible using femtosecond spectroscopy. This, combined with new structural data, has allowed a fairly complete description of light-harvesting in purple bacteria and substantial insights into higher plant antenna systems.

Addresses ~Department of Chemistry, University of California, Berkeley, CA 94720-1460, USA; e-mail: fieming@cchem.berkeley.edu tDepartment of Physics and Astronomy, Vrije Universiteit, De Boelelaan 1081, NL-1081 HV, Amsterdam, The Netherlands; e-mail: rienk@nat.vu.nl

Current Opinion in Structural Biology 1997, 7:738-748

http://biomednet.com/elecref/O959440XO0700738

O Current Biology Ltd ISSN 0959-440X

Abbreviations 3PEPS three-pulse photon echo peak shift BChl bacteriochlorophyll CD circular dichroism Chl chlorophyll RC reaction center

Introduction In order to harvest solar light, photosynthetic organisms are equipped with a light-harvesting antenna system. Photons absorbed by the antenna pigments are transferred to the photosynthetic reaction center (Re) with great speed. Once absorbed by the RC, the excitation energy is efficiently converted into a stable charge separation.

Since the basic description of energy transfer and trapping processes by Duysens [1,2] in the early 1950s, it has been clear that the elementary steps of the light-harvesting process are extremely rapid. Only recently, however, has it become possible to make direct experimental observations on the timescale of individual energy transfer steps; and currently energy migration can be investigated in the range of tens of femtoseconds to many nanoseconds. Application of femtosecond laser spectroscopy has been greatly stimulated by the remarkable successes in struc- ture determination of several important light-harvesting complexes in recent years. For example, the peripheral light-harvesting complex (LHCII) of green plants [3], the peripheral light-harvesting (LH2) complex of Rhodopseu- domonas (Rps.) acidophila [4,5°°], the LH2 of RhodospiHllum (Rs.) molischianum [6°°], and the core of Photosystem 1 (PSI) of cyanobactcria [7 "°] have all been determined. All these are intrinsic membrane proteins, and, together with the known structures of the bacteriochlorophyll (BChl) a protein of green sulphur bacteria [8], the phycobilipro- teins [9], and the recently resolved structure of the

peridinin-carotenoid protein of dinoflagellates [10 °°] - -a l l membrane-attached light-harvesting sys tems--we now have a multitude of structures available which exhibit amazing variation which will allow us to greatly extend our knowledge of the process of excitation energy transfer and the underlying physics.

In this review, we describe the considerable recent progress in understanding the purple bacterial antenna system and outline the current views on green plant and cyanobacterial systems, for which the structural data do not yet allow for fully detailed modeling.

D i s o r d e r e d v e r s u s o r d e r e d l i g h t - h a r v e s t i n g s y s t e m s Although the various structures now known exhibit a wide spread in organizational motifs, one striking aspcct stands out. Comparing bacterial and plant light-harvesting systems, the bacterial peripheral, LH2, and core, LH1, antenna are structures with a ve~ high degree of sym- metry (see Figure 1), whereas LHCII and even more so PSI appear spatially (i.e. positionally and orientationally) much more disordered. One of the major reasons for this variation is, of course, the size of the elementary building block. In LH1 and LH2, this is a pair of small transmembrane polypeptides, ot and 13, which carries two and three BChls, respectively. Assembly into a larger system will always lead to a structure with a high degree of symmetry.

In contrast, the PS1 core consists of a single pair of large polypeptides, the PsaA and PsaB gene products, which, together with a large number of smaller subunits, forms the PSI core that binds - 100 chlorophylls. LHCII seems to be an intermediate case. Although monomeric, LHCI1 still appears quite disordered in comparison with the LH1 and LH2 rings; the basic unit of LHCII is a trinaer of an -25 kDa protein that exhibits perfect C3 symmetry. The other apparent difference between plant and bacterial light-harvesting systems is the pigment density. In LH1 of purple bacteria, the density is two BChl ot-polypeptides per 12kDa; in LHCII of green plants, 12-14 Chls occur per 2 5 k D a - - a factor of three more. A similar variation applies in PSI. Thus, in plant light-harvesting systems, the various opportunities to bind Chl molecules have been exploited optimally.

Finall-y, although PS1 differs greatly from the LH1-RC core of purple bacteria, it also has a fundamental similarity. As should be apparent from Figure 1, in the LH1-RC core the rate-limiting step for trapping photons is energy transfer from any of the light-harvesting pigments to the

Femtosecond spectroscopy of photosynthetic light-harvesting systems Fleming and van Grondelte 739

Figure 1

B8o0

,c 1997 Current Opinion in Structural Biology

A model for the light-harvesting and photon trapping machinery in the photosynthetic membrane of a purple bacterium. The view is along the membrane plane and only the bacteriochlorophyll pigments are shown. The primary electron donor (the special pair, P) of the RC is indicated by an arrow. The LH2 (smaller gray rings) and RC structures are based on crystallography. The LH1 structure (large b~ack ring) is modeled truing the size of the RC protein (not shown) and the c~ I] unit of the LH2 structure. Note that no special orientation requirements are needed for effective transfer from LH2 to LH1. Carotenoids present in both structures are not shown.

special pair (the primal" electron donor, consisting of a pair of strongly interacting bacteriochlorophyll molecules) in the RC. In this scenario, the trapping efficiency is highest when as many antenna sites as possible are able to transfer to the RC, and this is clearly optimized within a ring. In PS1, crudely speaking, the pigments are organized in a band around the electron transfer chain, and, on average, they are all at a distance o f - 2-3 nm; however, within this structure, the number of contact sites has also been optimized, leading to efficient trapping. The symmetry itself is not important, rather the avoidance of quenching centers (e.g. stacked dimers) and the location of the maximum number of pigments close to the site where the primary charge separation occurs are important. In order to avoid undesirable oxidation or reduction of the antenna by the primary electron donor, however, the bulk of the antenna molecules are kept at a distance > 2 n m from the components involved in the electron transfer.

An additional feature of all Chl and BChl antenna complexes resolved to date is the presence of carotenoid molecules. These serve photo-protective, light-harvesting and often structural roles.

B a c t e r i a l a n t e n n a s Energy transfer in the peripheral, LH2, and core, LH1, of photosynthetic purple bacteria The recently resolved structures of LH2 of Rps. acidophila [4] and Rs. molischianum [6 °°] have revealed the highly symmetric pigment-protein ring, displaying C9 symmetry in the case of Rps. acidophila and C8 symmetry for Rs. molischianum. Although only a low-resolution structure is available for LH1 [11], it is evident that LH1 is also organized as a ring, most probably with 16-fold symmetry. The RC structure can be nicely fitted into the proposed LH1 ring [11,12"].

Both for LH1 and LH2, the basic building block of the structure is a heterodimer of two small (5-6kDa) polypeptides, a and 13. Both consist of a single trans- membrane helix with a highly conserved histidine that ligates the BChl approximately one third of the way along the a-helical stretch. Thus, in the LH1 and LH2 pigment-protein rings, the basic element is a BChl dimer. For LH1, the o~13-BChl 2 subunit can be purified; it is called B820 after its absorption maximum and retains many of the essential spectral properties of LH1 [13]. For LH2, such

740 Biophysical methods

a subunit cannot be obtained. In LH2, and most probably in LH1, the heterodimeric subunits associate into a ring with the 0t-polypeptides on the inside, the [3-polypeptides at the outside, and the pigments sandwiched between the two concentric rings of polypeptides. As a result of the formation of the ring, the absorption shifts to -870 nm for LH1, and to - 8 5 0 n m for LH2, although in the latter case other absorption maxima are also found for some species (820 nm, 830 nm), depending on the presence or absence of hydrogen bonds [14,15]. The [~-polypeptide of LH2 binds a second pigment, nearer to the cytosolic side of the complex, and in LH2 of Rps. acidophila these BChls form a nine-membered ring which absorbs at - 8 0 0 n m and is positioned at a distance of - 1.7 nm from the B850 ring.

Electronic structure Within the B850 ring of dimers, all distances between the pigments are very s imilar- -somewhat less than 1 rim. Nevertheless, within the ctl3 subunit, the electron density seems continuous, whereas electron density due to the two neighbouring BChls on adjacent subunits is discontinuous. The reason for this is that the ]3-polypeptide BChl is clearly bent. One further important point is that within the ct[3 subunit, overlap between the two BChls occurs between chlorin tings I, as in the special pair of the RC, where as overlap between BChls on adjacent subunits is between chlorin rings III. As a consequence, one may view LH2 and most probably LH1 as 'rings of interacting directs'. This concept is supported by many experimental observations (see below; and e.g. [16°]). The BChls in the B850 ring of LH2 and in the B870 ring of LH1 all have their Qy transition dipole almost parallel and their Qx transition dipole perpendicular to the membrane plane. The estimated excitonic coupling between BChls in a subunit is - 2 5 0 c m -1 [17"',18"]; the coupling is somewhat less between BChls on adjacent subunits. In contrast, the pigments in the B800 ring ate 'monomeric', the distance between two neighbours is -2.1 nm, and the corresponding dipole-dipole coupling is - 2 0 c m -1. The interaction between pigments in the B800 ring and the pigments in the B850 ring is of a similar magnitude. The B800 rings in LH2 are almost flat in the plane of the membrane, those in the LH2 of Po. molischianum are tilted away from the membrane plane by - 3 0 °. Finallx; the LH2 rings of Rps. acidophila and Rs. mo/ischianum contain two carotenoids per ot]3 subunit.

LH1 and LH2 have been subjected to a large num- ber of spectroscopic studies, notably, polarized light spectroscopy (circular dichroism [CD], linear dichroism) [19], a variety of line-narrowing techniques (holeburning [20,21,22"',23"1, site-selective fluorescence [24,25]), and infrared and Raman spectroscopies. In addition, using structural information, several of the spectroscopic features have been modeled [17"',26°',27°']. From these studies, two opposing views have emerged which we will discuss.

In the first view, LH1 and LH2 are considered to be rings of interacting directs, in which many of the essential

spectroscopic features, including the dramatic red shift, largely originate from within a dimer. The excitonic interaction between neighbouring subunits within the ring is considered to be a relatively weak perturbation, that is, relative to the intradimer cxcitonic interaction, the possible (and so far unknown) contribution from electron exchange arising from ring I overlap, the intrinsic energetic disorder and the electron-phonon or electron-vibration coupling. The spectra of all photosynthetic pigment- proteins are now known to be strongly inhomogeneously broadened, and estimates of the amount of inhomoge- neous broadening range from 200-500cm -1. In addition, the electron-phonon coupling is estimated to be of the same order of magnitude. The general idea behind the 'ring of dimers' model is that, following excitation, any phase relation between excitations on different dimers is rapidly destroyed, either dynamically, because of the coupling to vibrations or phonons, or as a consequence of the interference of the pure eigenstates due to energetic disorder.

In the alternative view, the spectroscopic features are totally determined by the set of excitonic eigenstates of the full ring. In this model, the excitonic interaction between adjacent BChls is the dominant term that completely determines the red shift observed upon formation of the ring. The lowest state of the exciton manifold is almost optically forbidden, because of the inplane orientation of the Qy transition dipoles, and all the oscillator strength is equally divided between two orthogonal transitions slightly above the lowest one. Experimental evidence to support this model includes holeburning experiments [21,22"°,23"], the interpretation of the low-temperature absorption spectrum [17"'], and estimates of the absorption cross-section of the major transition at 850nm or 870nm [28°',29]. We are of the opinion that the latter view is less accurate, mainly because it ignores all the nonexcitonic contributions that all have the effect of destroying the fully delocalized coherent states. In addition, as we show below, the ring of dimers model provides a simple and elegant explanation for many of the dynamic results.

Intraring energy transfer From a variety of spectroscopic studies (for a review, see van Grondelle eta/. [30]), the energy transfer dynamics within LH1 and LH2 have been concluded to be ultrafast. With the advent of femtosecond laser spectroscopy, in particular using Ti:Sapphire lasers, many of the elementary energy transfer steps have been resolved in time. The B800--+B850 energy transfer at room temperature takes -700-800fs for LH2 from Rb. sphaeroides, and this time constant is not very species dependent [31-35,36°',37"]. The energy transfer time is only weakly dependent on temperature, being - 1 ps at 77K and - 2 p s at 4K. The B800--+B850 energy transfer has been modeled in terms of a F6rster process [31,33,34]. The weak temperature

Femtosecond spectroscopy of photosynthetic light-harvesting systems Fleming and van Grondelle 741

dependence of this energy transfer step suggests the involvement of some vibronic level of B850 or possibly of the higher excitonic states of the B850 ring [22"°,37°]. Previously, efficient energy transfer was concluded to occur within the B800 ring from fluorescence polarization experiments [19]. More recently, a time constant of -0 .5-1 ps has been estimated for energy transfer between neighbouring B800 rings from polarized pump-probe spectroscopy [34,35,38°°].

For energy transfer within the B850 and B875 rings, single-site lifetimes of the order of a few hundred femtoseconds have been estimated, for example, from an analysis of the efficiency of singlet-singlet annihilation [39]. In addition, the observation that within a few picoseconds transient absorption changes were almost fully depolarized has been interpreted as subpicosecond energy transfer among B850 rings in LH2 and B870 rings in LH1 [40,41]. The energy migration in B850 and B870 has been recorded directly using fluorescence depolarization [42,43°°], and using equilibration of the transient absorption spectrum [44]. Both studies used a similar interpretation based on hopping between dimers in a ring. The site energies of the dimers were taken at random from an inhomogeneous distribution o f - 4 0 0 cm -1 width, and average hopping times o f - 1 0 0 f s were obtained. The fluorescence anisotropy decays faster in LH2 than in LH1, and in this model this arises simply from the smaller ring size of LH2 (larger angle change per hop). The model could be extended to low temperatures where the site energy variation impedes the energy transfer over more than a few sites [45°°]. Remarkably, Chachisvilis et al. [46] and Bradforth et al. [42] found that oscillations at 105 cm -I, assigned to vibrational wavepacket motion, dephased significantly slower than the observed depolarization timescale, suggesting vibrational coherence transfer [47] in the energy transfer process.

As discussed above, the extent of exciton delocalization in LH1 and LH2 has been extensively debated. Key quantities are the electronic coupling between the BChls, the electron-phonon coupling (reorganization energy and timescale), the temperature, and the disorder. From the difference in position between the pump-induced bleaching (ground state to one-exciton state) and pump- induced absorption (one-exciton state to two-exciton state), Sundstrtim and coworkers [48°,49 °] estimate a delocalization length of 4+2 molecules in LH1 and LH2, more or less independent of temperature. A measurement of the superradiance in LH1 and LH2 gave an even smaller number [50°°]. On the other hand, an ultrafast decay in the transient absorption and emission of LH2 was taken as an indication for relaxation between fully delocalized states [51°]. An incisive discussion of how delocalization influences different observables has been given by Leegwater [52 °] and more recently by Meier et al. [53°°].

In an attempt to provide experimental characterization of the electron-phonon coupling, Jimenez et al. [54 °°] carried out three-pulse photon echo peak shift (3PEPS) measurements on LH1 and LH2. They concluded that on a 50fs timescale fluctuations in the environment and vibrations lead to the dynamic localization on a dimeric subunit of LH1 and LH2. A similar conclusion was drawn from the ultrafast reorganization, as observed by the formation of the Stokes' shift in a few tens of femtoseconds [55°]. In the peak shift decay, this initial phase was followed by an exponential phase that was interpreted as a loss in the rephasing capability of the system due to energy transfer. During the energy transfer, the system samples all the various environments that contribute to the inhomogeneous broadening, and, as a consequence, the information about the original environment is lost. Again, the model that assumes hopping on an inhomogeneously broadened ring of dimers gave a fit to the results. The homogeneous broadening was estimated to be - 2 0 0 c m -1, the inhomogeneous b r o a d e n i n g - 5 0 0 c m -I, and the hopping time -100fs . This interpretation is very much supported by a 3PEPS experiment on the LH1 subunit, B820, in which the 100 fs phase in the peak shift decay ascribed to energy transfer was absent and replaced by a nondecaying component arising from inhomogeneous broadening [56°°]. The striking similarity between the 3PEPS of LH1 and B820 further supports the idea that excitations in these antenna complexes are delocalized over only a dimer unit.

Carotenoids Energy transfer from carotenoid to BChl in LH2 of Rb. sphaeroides can occur on a timescale of a few 100 fs [57]. Recent fluorescence upconversion experiments demonstrated that, in LH1 and LH2 of Rb. sphaeroides, the $2 lifetime of sphaeroidene is shortened to - 5 5 f s for the former and 80fs for the latter. This should be compared with a 150-250 fs internal conversion time from $2---)S 1, dependent on the solvent [58°]. For a B800-830 complex of Chromatium purpuratum, Gillbro and coworkers [59 °°] report an S2--~BChl (Qx) transfer time of 100fs and S1---)BChl (Qy) transfer times of 3.8 ps and 0.5 ps for the carotenoid that transfers to B830 and the carotenoid that transfers to B800, respectively.

Interring transfer In the intact bacterial photosynthetic unit, the energy transfer from one LH2 to another, and from LH2 to LH1, takes place on a timescale of a few picoseconds [60-62,63°]. Assuming Ftirster energy transfer between BChls on neighbouring rings, this would imply a closest distance o f - 3 nm between the two pigment-protein rings. Long before the crystal structure of LH2 became available, it was realized that the rate-limiting step in excitation trapping was the step from the L f t l pigments to the special pair in the RC. A transfer time o f - 3 5 p s was obtained for the LH1 to special pair energy transfer fL64],

742 Biophysical methods

and this was interpreted as a distance of 4.5 nm between the special pair and the LH1 ring, in good agreement with models suggested for the LH1-RC core, assuming that LH1 is organized as a ring of 16 ¢~-BChl 2 subunits with a structure as in LH2 of Rps. acidophila [12",65*°]. A summary of the timescales is given in Figure 2.

Green plant and cyanobacter ia antennas LHCII L H C I I is the major light-harvesting pigment-protein of higher plants and algae and is responsible for the binding o f - 5 0 % of all Chl on earth. It serves to feed excitation energy into the minor light-harvesting complexes, CP29, CP26, CP24, and into the core of Photosystem 2 (PS2) which eventually is used for charge separation. L H C 2 is a member of a family of light-harvesting complexes which includes the various forms of L H C I I and the minor light-harvesting complexes. T he basic unit of all these complexes is a membrane protein o f - 2 5 kDa, which is known to fold into a structure with three transmembrane helices: A, B and C.

In LHCII , the monomeric subunit binds 7-8 Chl a, 5-6 Chl b, 2 luteins, 1 neoxanthin and substoichiometric amounts of violaxanthin. In its native form L H C I I is a trimer, and in 1994 the structure of the trimer of L H C I I b was resolved to a resolution of 3 - 4 a by K0hlbrandt and coworkers [3] using cryoelectron microscopy. At the

current resolution, Chl a and Chl b are indistinguishable. In addition, the phytol tails of the Chls cannot be observed, and, as a consequence, the orientation of the Qx and Qy transition dipoles within each of the chlorin planes is not known. In the proposed model, the assignment of the Chl as and Chl bs is based on the following argument. After excitation, there is a small but finite chance that a triplet is formed selectively on one of the Chl as because of the assumed fast Chl b to Chl a energy transfer. As one role of the carotenoids is to quench these Chl a triplets with high efficiency to prevent the formation of harmful oxygen radicals, the Chl as must be positioned in van der Waals contact with the luteins. T h e seven Chls in the core of LHCII , which all make close contact with the luteins, have been therefore assigned to the seven Chl as: the remaining Chls to Chl b. In view of recent reconstitution experiments with LHCII , it may be possible that some of the binding sites are promiscuous and can be occupied by either a Chl a or a Chl b. L H C I I exhibits intense CD spectra, indicative of Chl a-Chl a and Chl b--Chl b cxcitonic interactions; the interaction between Chl as and Chl bs is most probably weak.

A variety of picosecond and femtosecond studies have been performed to explore the dynamics of energy transfer within LHCII . In a pioneering fluorescence upconversion study by Eads et al. [66], the dominant time constant for energy was estimated to be 0.5+0.2 ps. In a low-intensity

Figure 2

B875

100 fs 3 ps B850

0.7 ps

B800

LH1 8o fs L H 2

~c 1997 Current Opinion in Structural Biology

A summary of the timescales for energy transfer in purple bacteria. Note the slow, final step (35 ps) from LH1 to the special pair in the RC. Not shown are carotenoid to BOhl transfer times which are of the order of 100fs in LH2. The B875 and B850 molecules are shown as dimers (ovals), whereas the B800 molecules are shown as monomers (diamonds).

Femtosecond spectroscopy of photosynthetic light-harvesting systems Fleming and van Grondelle 743

pump-probe study, a slower Chl b--+Chl a energy transfer time in the range of a few picoseconds was obtained in addition to the ultrafast process [67]. Transient absorption with shorter pulses [68] revealed Chl b--+Chl a energy transfer times of 160fs and also the slow process of 5 + 2 p s similar to the results of Kwa et al. [67]. In a fluorescence upconvetsion study by D u e t a/. [69], two lifetimes in the rise in presumably Chl a fluorescence were detected upon excitation at 650 nm, + 250 fs and 5 ps, in rather good agreement with the pump-probe results. On the other hand, P~lsson et al. [70] using one-colour pump-probe, detected a major 500 fs and a minor 2-3 ps Chl b-+Chl a transfer time, and, despite the superior time resolution, they could not distinguish any transfer component faster than 500 fs. More recently, Visser et al. [71"], and later Connelly et al. [72 °] resolved all three phases in the Chl b--+Chl a energy transfer: 180fs, 600fs and - 5 p s , with a relative anaplitude ratio of 40%, 40% and 20%, respectivel'~: A study on L H C I I monomers demonstrated that all three decay times are associated with Chl b--+Chl a energy transfer within a monomeric unit of L H C I I (FJ Kleima et al., unpublished data). According to Visser e ta / . [7l*'], in the trimer all the energy transfer occurred to the major red absorbing species at 676nm. Connelly eta/. [72"] concluded from their data, which was obtained with an excellent signal-to-noise ratio, that the 175 fs component probably partly reflected energy transfer between 'blue ' and ' red ' Chl bs [73]. Measurements of singlet-singlet and singlet-triplet annihilation suggest that intermonomer energy transfer occurs on a timescale of 10-20 ps [68,71"].

"Very recently, Gradinaru et a/. (unpublished data) have studied the Chl b--+Chl a transfer in one of the minor light-harvesting complexes, CP29. In CP29, six of the Chl a and two of the Chl b binding sites are conserved [74"], suggesting a pigment stoichiometry of six Chl a : t w o Chl b : one lutein : one neoxanthin : one violaxanthin. The kinetics in CP29 contain many components similar to those in L H C I I and most probably reflect the same energy transfer processes. Specifically; Gradinaru e t a / . (unpublished data) could assign a slow, 2-3ps energy transfer phase to a Chl b absorbing at 650 nm - - most probably Chl b5 in the L H C I I a s s i gnm e n t - - a nd a fast, 0.2-0.3 ps energy transfer phase to a Chl b3.

Carotenoid to Chl energy transfer in L H C I I is highly efficient. Recently, two conflicting reports appeared on the dynamics and pathway of carotenoid to Chl a transfer. Peterman et al. [75 *°] argued that no direct carotenoid to Chl b transfer occurred, while carotenoid to Chl a energy transfer took place in - 2 2 0 fs. In contrast, Connelly et al. [76"'] claimed that the carotenoids exclusively transferred energy to Chl b, followed by Chl b--+Chl a energy transfer. The latter would be inconsistent with the assignment by Ktihlbrandt and coworkers [3], where only close contacts between Chl as and carotenoids exist.

PS1 The core of PS1 is the most complex photosynthetic light-harvesting plus electron transfer system for which a structure is now available [7"']. The functional unit of the PSI core of Synechococcus elongatus consists of 11 subunits, including the two major subunits PSaA and PSaB, each having a molecular weight o f - 8 0 k D a with known sequcnce, and each binding - 1 0 0 Chl as, 10-25 carotenoids and three FeS clusters. The structure has been resolved to 4 ~, and the positions o f - 90 Chls have been determined. As in the case of LHCII , no information about the direction of the Qv and Qx transition dipoles is available. The core of P~;1 is characterized by 22 transmembrane helices, 11 for each large subunit, which exhibit C2 symmetry around an axis that passes through the centrally located special pair of the electron transfer chain, P700, and the FeS cluster E The electron transfer chain is embedded in a structure of ten transmembrane helices, five from each subunit, the arrangement of which is strongly reminiscent of that of the L and M subunits in the purple bacterial RC. All the other 90 antenna Chl as are dispersed in a band around this core and, for a large part, are associated with the remaining six transmembrane helices on each of the large subunits. For all the Chls except two, the distance to any of the pigments in the electron transfer chain exceeds 1.6nm, making energy transfer slow (10-20 ps). Two chlorophylls are found that seem to connect the antenna with the second and third pair of Chls of the electron transfer system, and it has been suggested that these form a special entry for excitation energy. A remarkable sequence analogy exists between the antenna part of the large PSI subunits and the core proteins of PS2, CP47 and CP43, and for that reason it has been suggested that the pigment-protein arrangement of the six outer transmembrane helices and their associated Chls may be a good model for the PS2 core.

Trapping in PSI is fast (20-25 ps), and charge separation is essentially irreversible [30,77,78]. Using ultrafast fluo- rescence depolarization, Du et al. [79] estimated that the major hopping process within PS1 occurs on a timescale of 100fs. On a timescale of a few picoseconds, the excitation energy is seen to equilibrate between a pool of very red pigments, absorbing at - 7 2 0 - 7 3 0 n m , and the major PSI core pigments. The process of energy transfer to P700 must occur at the same rate as this equilibration between core and red pigments as, even at very low temperatures where escape from the red states is impossible, a reasonably high quantum yield for charge separation is still observed upon excitation of the core pigments [80,81"]. This has led to a model in which essentially all sites within the PSI core are more or less equally efficient in transferring their energy to P700 (or any other pigment of the electron transfer chain) and which may be viewed as the 3D version of the 2D ring to special pair energy transfer model that seems to operate for purple bacteria [82]. In our view, it is highly unlikely that

744 Biophysical methods

the two Chls that were proposed to act as a special entry for excitation energy indeed have that role. They have more rapid energy transfer to the pigments in the electron transfer chain but are simply outnumbered by all the other Chls. A simple simulation of the trapping kinetics in PSI shows that leaving out the pair of connecting Chls hardly changes the trapping time.

The process of energy migration and charge separation cannot be experimentally separated in PSI. Kumazaki et

al. [83•], Trinkunas and Holzwarth [84 °] and White et al.

[85 •°] have used modeling to extract the intrinsic electron transfer rate. In a very recent study using a PSI mutant, which seemed to affect the special pair P700 but not the antenna spectra or dynamics, it was observed that the excited state lifetime approximately doubled [86"]. This was taken by Melkozernov et al. [86 °'] as evidence for a model in which the charge separation rate by P700 is the rate limiting step, in contrast to the 'transfer-to-the-trap' limited model discussed above.

Conclusions The combination of high-resolution structural data and uhrafast spectroscopy has enabled the development of a fairly complete picture of the light-harvesting process in purple bacteria. The efficiency of the overall process is based on individual energy transfer steps of 80-100fs. In LH1, the core antenna surrounding the Re, several hundred energy transfer steps occur before the final transfer to the special pair (35ps) and the initiation of charge separation. The observation of a sub 100fs energy transfer, along with the retention of coherence and the enhanced radiative rates in LH1 and LH2, raises many challenging issues which will provide stimulus for theory and experiment for years to come. Despite the high symmetry and potential for strong intermolecular coupling, it does not appear that extensive electronic delocalization is necessary for achieving the near unit efficiency of the light-harvesting process.

In green plant and cyanobacterial antennas, the structural information is not yet sufficient for the most detailed molecular modeling of energy migration. Enough is known, however, to reveal both striking similarities and differences with the purple bacteria. In particular, antenna molecules are held away from close contact with the primary electron donor, and efficiency is achieved by using large numbers of antenna molecules with roughly similar transfer rates to perform the final transfer step to the primary donor.

References and recommended reading Papers of particular interest, published within the annual period of review, have been highlighted as:

• of special interest • • of outstanding interest

Duysens LMN: Transfer of excitation energy in photosynthesis [PhD Thesis]. Utrecht, The Netherlands: Utrecht University; 1952.

2. Duysens LMN: Photosynthesis. ProgrBiophys 1964, 14:1-104.

3. KOhlbrandt W, Wang DN, Fujiyoshi Y: Atomic model of plant l ight-harvesting complex by electron crystallography. Nature 1994, 367:614-621.

4. McDermott G, Prince SM, Freer AA, Hawthornthwaite-Lawless AM, Papiz MZ, Cogdell PJ, Isaacs NW: Crystal structure of an integral membrane l ight- harvesting complex from photosynthetic bacteria. Nature 1995, 374:51 ?-521.

5. Freer AA, Prince S, Sauer K, Papiz MZ, Hawthornthwaite- o, Lawless AM, McDermott G, Cogdell R J, Isaacs NW:

Pigment-pigment interaction and energy transfer in the antenna complex of the photosynthetic bacterium Rps. acidophile. Structure 1996, 4:449-462.

This paper discusses the pathways of excitation energy transfer in the LH2 peripheral antenna complex of Rps. acidophi/a, in the light of the recently obtained high-resolution structure [4]. The FSrster dipole-dipole resonance coupling is concluded to dominate the energy transfer from the B800 to the B850 ring. Within the B850 ring, strong interactions exist between nearest neighbour BChls partly because of a close to optimal alignment of their transition moments, suggesting that delocalized excitonic states play a role in the energy transfer. The orientations and distances of the rhodopin molecules, the BS00 and B850 BChls, suggest that singlet-singlet energy transfer from carotenoid to BChl involves mainly transfer to B850 and occurs predominantly from the S 2 state of the carotenoid to the Qx state of the BChl.

6. Koepke J, Hu X, Muenke C, Schulten K, Michel H: The crystal • . structure of the light-harvesting complex II (B800-850) from

Rhodospirillum molischienum. Structure 1996, 4:581-597. The crystal structure of LH2 of Rs. mo/ischianum is obtained via a molecular replacement method at a resolution of 2.4]k. It is an (c(~) 8 complex with 16 B850 and 8 B800 BChls in an eightfold symmetric ring. The 16 B850 BChls, sandwiched between the two polypeptide rings, are in a ring with a radius of 2.3 nm. The BSO0 BChls are situated between the 13-polypeptides in a ring with a diameter of 2.88 nm. The B6OOs are bound to Asp6 of the c(-polypeptide. The eight lycopenes span the membrane and are held in place by aromatic sidechains. The B800 chlorin planes are rotated by - 90" relative to their position in LH2 of Rps. acidophila and are very much tilted away from the plane of the membrane. Nevertheless, in this structure the B800 Q./ transition dipoles are more or less parallel to the B850 Qy transition dipoles.

7. Krauss N, Schubert W-D, Klukas O, Fromme P, Witt HT, • . Saenger W: Photosystem I at 4A resolution represents the first

structural model of a joint photosynthetic reaction centre and core antenna system. Nat Struct Bio/1996, 3:965-973.

The structure of PS1 from Synechococcus e/ongatus is determined to 4 A resolution using X-ray crystallographic methods. The arrangement of the 22 transmembrane and 4 surface helices displays a twofold symmetry for the large subunits PsaA and PsaB. The central part of the structure, which is proposed to carry the electron transfer chain, including PTO0, the acceptor A 0, and the three FeS clusters, shows a striking resemblance to the LM core of the bacterial reaction center. The 90 densely packed antenna Chls form an oval clustered net, relatively distant from the heart of the complex and only continuous with the electron transfer chain via the second and third Chl pairs of the electron transfer system. This suggests a dual role for these Chl as both in excitation transfer and electron transfer.

8. Fenna RE, Matthews BW: Chlorophyll arrangement in a bacteriochlorophyll protein from Chlorobium limicola. Nature 1975, 258:573-577.

9. Brejc K, Ficner R, Huber R, Steinbacher S: Isolation, crystallization, crystal structure analysis and refinement of allophycocyanin from the cyanobacterium Spirulina platensis at 2.3 A resolution. J Mol Bio/1995, 249:424-440.

10. Hoffmann E, Wrench PM, Sharpies FP, Hiller RG, Welte W, • . Diederichs K: Structural basis of light-harvesting by

carotenoids: peridinin-chlorophyll*protein from Amphinidium carterae. Science 1996, 272:1 788-1791.

The structure of the peridinin-chlorophyll-protein (PCP) is solved to a resolu- tion of 2.0/~ using X-ray diffraction. PCP is a water-soluble light-harvesting complex, which has a blue-green absorbing carotenoid as its major pigment and which is present in most photosynthetic dinoflagellates. The fold of the N-terminal and C-terminal domains of each polypeptide is related by a twofold symmetry axis, and it surrounds a hydrophobic cavity filled with two lipid, eight peridinin and two Chl a molecules. The structural basis for efficient energy transfer from peridinin to Chl is found in the clustering of peridinins at van der Waals distances around the Chls.

11. Karrasch S, Bullough PA, Ghosh R: The 8.5/~ projection map of the light-harvesting complex I from Rhodospirillum rubrum reveals a ring composed of 16 subunits. EMBO J 1995, 14:631-638.

Femtosecond spectroscopy of photosynthetic light-harvesting systems Fleming and van Grondelle 745

12. Papiz MZ, Prince SM, Hawthornthwaite-Lawless AM, • McDermott G, Freer AA, Isaacs NW, Cogdell PJ: A model for the

photosynthetic apparatus of purple bacteria. Trends P/ant Sci 1996, 1:198-206.

A model is produced for the whole photosynthetic unit of purple bacteria based on the crystal structure of the LH2 peripheral light-harvesting com- plex of Rps. acidophila. To model the (x16~1 s projection map of Karrasch et aL [11 ] the (x9139 structure of LH2 is increased to produce the larger ring. The known structure of the reaction center is found to fit in this LH1 ring. Precisely six LH2s can be fitted in a circle around this LH1 core, thereby reproducing the typical stoichiometry of LH1 to LH2.

13. Visschers RW, Chang MC, van Mourik F, Parkes-Loach PS, Heller BA, Loach PA, van Grondelle R: Fluorescence polarization and low-temperature absorption spectroscopy of a subunit form of light-harvesting complex I from purple photosynthetic bacteria. Biochemistry 1991, 30:5734-5?42.

14. Fowler GJS, Visschers RW, Grief GG, van Grondelle R, Hunter CN: Genetically modified photosynthetic antenna complexes with blue-shifted absorbance bands. Nature 1992, 355:848-850.

15. Fowler GJS, Sockalingum GD, Robert B, Hunter CN: Blue shifts in becteriochlorophyll ebsorbance correlate with changed hydrogen bonding patterns in light-harvesting LH2 mutants of Rhodobacter spheeroides with alterations at c(Tyr44 and 45. Biochem J 1 g94, 299:695-700.

16. Beekman LMP, Steffen M, Stokkum IHM, van Olsen JD, Hunter • CN, Boxer SG, van Grondelle R: Characterization of the light

harvesting antennas of photosynthetic purple bacteria by Stark spectroscopy. 1. LH1 antenna complex and the B820 subunit from Rhodospiri/lum rubrum. J Phys Chem 1997, in press.

The response of the optical absorption spectrum to an externally applied electrical field (Stark effect) is measured for LH1, the B820 dimeric subunit of LH1, and the reconstituted LH1 of purple photosynthetic bacteria. The Stark effect for LH1 is strongly reminiscent of that measured for the special pair of the purple bacterial reaction center, it is dominated by a large change in polarizability between ground and excited states-most probably due to the mixing of charge transfer states into the lower excited states of a dimeric subunit of LH1.

17. Sauer K, Cogdell R J, Prince, SM, Freer A, Isaacs NW, Scheer H: °• Structure-based calculations of the optical spectra of the LH2

bacteriochlorophyll-protein complex from Rhodopseudomonas acidophila. Photochem Photob~o11996, 64:564-576.

The molecular structure of the LH2 complex of Rps. acidophila is used to provide orientations and distances for the 27 BChls in the complex as the input parameters for an excitonic hamiltonian. Assuming dipole-dipole interactions among all the chromophores, the ring of 18 closely coupled ¢hromophores is assigned to B850, whereas the parallel ring of nine weakly ¢oup4ed BChls is assigned to B800. The pairwise excitonic interactio,1 be- tweetn the B850s is estimated to be - 2 5 0 - 3 0 0 cm -1. The general trends obaervable in the CD spectrum of LH2 are tentatively explained. The cal- culations predict strong CD features at - 7 9 0 nm which await experimental verification.

18. Sturgis JN, Robert B: The role of chromophore coupling in • tuning the spectral properties of the peripheral light-harvesting

protein of purple bacteria, Photosynth Res 1996, 50:5-10. The authors calculate the excitonic contribution to the absorption spectrum for LH2 of photosynthetic purple bacteria and conclude that this contribution is not very sensitive to small variations in the LH2 structure. For that reason, they propose that in LH2 of Rps. acidophi/a the redshift to 850 nm originates roughly equally from pigment-pigment and pigment-protein interactions. In BS00-B820, a related but slightly different LH2, the redshift is largely due to pigment-pigment interaction. As a consequence, the total amount of ex- citonic interaction between neighbouring pigments can not be much larger than 200 cm -1.

19. Kramer HJM, van Grondelle R, Hunter CN, Westerhuis WHJ, Amesz J: Pigment organization of the B800-850 antenna complex of Rhodopseudomonas sphaeroides. Biochim Biophys Acta 1984, 765:156-165.

20. De Caro C, Visschers RW, van Grondelle R, VSIker S: Inter- and intraband energy transfer in LH2-antenna complexes of purple bacteria. A fluorescence line-narrowing and hole-burning study. J Phys Chem 1994, 98:10584-10590.

21. Reddy NRS, Picorel R, Small G J: B896 and B870 components of the Rhodobacter sphaeroides antenna: a hole-burning study. J Phys Chem 1992, 96:6458-6464.

22. Wu HM, Savikhin S, Reddy NRS. Jankowiak R, Cogdell R J, • • Struve WS, Small G J: Femtosecond and hole-burning studies

of B8OO's excitation energy relaxation dynamics in the LH2

antenna complex of Rhodopseudomonas ecidophila (strain 10050), J Phys Chem 1996, 100:12022-12033.

The transfer of excitation energy in the LH2 complex of Rps. aci- dophila is studied using holeburning and femtosecond laser spectroscopy. BS00~B850 energy transfer is observed to occur with monophasic kinet- ics with a time constant ranging from 1.6 ps ( lgK) to 1.1 ps (130K), while holeburning at 4.2K yields 1.8 ps. Holeburning with applied pressure shows no effect on the energy transfer rate. Together with the weak temperature dependence, this suggests that the B800 emission overlaps with a weak vibronic band of B850. Time-domain and holebuming spectroscopy show that there is an additional relaxation channel for B800 excitations when the excitation is to the blue of the B800 band. Two possible processes, intra-B800 transfer and coupling with the quasi degenerate upper exciton manifold of B850, are discussed.

23. Wu HM, Reddy NRS, Cogdell RJ, Muenke C, Michel H, Small G J: • A comparison of the LH2 antenna complex of three purple

bacteria by hole-burning and absorption spectroscopies. Mo/ Cryst Liq Cryst 1996, 291:163-173.

The LH2 complexes of Rps. acidophila, Rb. sphaeroides and Rs. molischi- anum are investigated using holeburning. B800--)B850 energy transfer at 4.2K takes 1.9 ps, very similar to the time constant observed for the same process in the other species. The absorption spectra of Rs. molischianum and Rps. acidophila undergo a dramatic redshift and thermal narrowing upon cooling from room temperature to 4.2K; for LH2 of Rb. sphaeroides, these effects are much smaller. This is interpreted as a smaller excitonic coupling in the latter species.

24. Van Mourik F, Visschers RW, van Grondelle R: Energy transfer and aggregate size effects in the inhomogeneously broadened core light-harvesting complex of Rhodobacter sphaeroides. Chem Phys Lett 1992, 193:1-7.

25. Monshouwer RM, Visschers RW, van Mourik F, Freiberg A, van Grondelle R: Low-temperature absorption and site- selected fluorescence of the light-harvesting antenna of Rhodopseudomonas v/r/dis. Evidence for heterogeneity. Biochim Biophys Acta 1995, 1229:373-380.

26. Alden RG, Johnson E, Nagarajan V, Parson WW, Law C J, • o Cogdell R J: Calculations of spectroscopic properties of

the LH2 bacteriochlorophyll-protein antenna complex from Rhodopseudomonas acidophila. J Phys Chern B 199'7, 101:4667-4680.

This paper describes the calculation of absorption and CD spectra of a photosynthetic bacterial antenna complex based on the crystal structure of the LH2 complex from Rps. acidophila. Molecular orbitals for the three differ- ent BChl structures in the complex are obtained by semiempirical quantum mechanical calculations. Exciton and charge transfer interactions are intro- duced at the level of configuration interactions. Absorption bandshapes are treated with vibronic parameters as obtained from holeburning experiments, whereas inhomogeneous broadening is included by a Monte Carlo method. Calculations reproduce the measured absorption and CD spectra. The re- suits support the idea that excitations are rather delocalized in LH2.

27. Koolhaas MHC, van der Zwan G, Frese RN, van Grondelle R: • - The red shift of the zero-crossing in the CD spectra of the

LH2 antenna complex of Rhodopseudomonas acidophila: a structure based study. J Phys Chem 1997, in press.

The published crystal structure of LH2 of Rps. acidophi/a is used to calculate absorption and CD spectra of the complex. It is shown that the relative position of the CD zero crossing with respect to the absorption maximum is an important parameter that is rather sensitive to structural changes. It is demonstrated that the experimentally observed CD spectrum can only be explained if the whole ring is considered and if the o(-polypeptide and I~-polypeptide BChls are allowed to have different excitation energies.

28. Leupold D, Stiel H, Teuchner K, Nowak F, Sandner W, 0cker B, °° Scheer H: Size enhancement of transition dipoles to one and

two-exciton bands in a photosynthetic antenna. Phys Rev Lett 1996, 77:4675-46"78.

From a measurement of the nonlinear absorption, the differential absorp- tion density spectrum, and the fluorescence for LH2 from Rb. sphaeroides, the dipole moments associated with the ground state-->one-exciton and one-exciton-->two-exciton transition are estimated to be 25.5 D and 21.5 D, respectively. These values are seen as an indication that in LH2 the exciton is delocalized over 16 +4 BChl molecules, corresponding to the full physical length of the circular aggregate.

29. Novoderezhkin VI, Razjivin AP: Exciton dynamics in circular aggregates: application to antenna of photosynthetic purple bacteria. Biophys J 1995, 68:1089-1100.

30. van Grondelle R, Dekker JP, Gillbro T, SundstrSm V: Energy transfer and trapping in photosynthesis. Biochim Biophys Acta ~994, 1187:1-65.

746 Biophysical methods

31. van Grondelle R, Kramer HJM, Rijgersberg CP: Energy transfer in the B800-850-carotenoid l ight-harvesting complex of various mutants of Rhodopseudomonas sphaeroides and of Rhodopseudomonas capsulatus. Biochim Biophys Acta 1982, 682:208-215.

32. Trautman JK, Shreve AP, Violette CA, Frank HA, Owens TG, Albrecht AC: Femtesecond dynamics of energy transfer in B800-850 light-harvesting complexes of Rhodobacter sphaeroides. Proc Nat/Acad Sci USA 1990, 87:215-219.

33. Laan H, Schmidt Th, Visschers RW, Visscher KJ, van Grondelle R, VSIker S: Energy transfer in the B800-850 antenna complex of the purple bacterium Rhodobacter sphaeroides: a study by spectral hole-burning. Chem Phys Lett 1990, 170:231-238.

34. Monshouwer R, Ortiz de Zarate I, van Mourik F, van Grondelle R: Low-intensity pump-probe spectroscopy on the B800 to B850 transfer in the light harvesting 2 complex of Rhodobacter sphaeroides. Chem Phys Lett 1995, 246:341-346.

35. Hess S, ~kesson E, Cogdell R J, Pullerits T, Sundstr6m V: Energy transfer in spectrally inhomogeneous light-harvesting pigment- protein complexes of purple bacteria. Biophys J 1995, 69:2211- 2225.

36. Joe 1", Jia Y, Yu, J-Y, Jonas DM, Fleming GR: Dynamics in isolated • • bacterial light harvesting antenna (LH2) of Rhodobacter

sphaeroides at room temperature. J Phys Chem 1996, 100:2399-2409.

Transient absorption, transient grating and photon echoes are measured on the B800 band of LH2 of Rb. sphaeroides using 30fs pulses. B800--~B850 energy transfer occurs in 800fs. The three-pulse photon echo peak shift experiment identifies important contributions to the 6800 lineshape and thereby the dynamics of the system involved: low-frequency intramotecular vibrations; ultrafast bath (solvent plus protein) responses; and static inho- mogeneity on a timescale longer than the energy transfer time. The transient absorption is observed to decay nonexponentially. The authors argue that the fast phase is vibrational relaxation within the B800 band.

3?. Ma Y-Z, Cogdell R J, Gillbro T: Energy transfer and exciton • annihilation in the B800-850 antenna complex of the

photosynthetic bacterium Rhodopseudomonas ecidophila (strain 10050). A transient femtosecond absorption study. J Phys Chem B 1997, 101:1087-1095.

This paper describes femtosecond pump-probe experiments on LH2 of Rps. acidophi/a. At room temperature, B800--~B850 energy transfer is -0 .8 ps; at 77K, -1 .3ps. Anisotropy kinetics measured within 6800 band indicate that relaxation within the B800 band is wavelength dependent. A fast depolarization time of 210fs observed at 77K is thought to originate from exciton relaxation. From a dramatic energy dependence of the B800 kinetics, it is speculated that several high-lying excitonic states of B850 exist that show good spectral overlap with the B800 band and thus could serve as excellent accepters for the energy transfer from B800 to B850.

38. Monshouwer R, van Grondelle R: Excitations and excitons in • • bacterial l ight-harvesting complexes. Biochim Biophys Acta

1996, 1275:70-75 The localization versus delocalization of excitations in bacterial light- harvesting complexes is discussed. It is argued that a 'ring-of-dimers' model is adequate to explain most of the spectroscopic and time-resolved data. Furthermore, two-colour pump-probe experiments in the B800 band of Rb. sphaeroides reveal 'blue-to-red' energy transfer on a timescale of 400fs within the B800 band.

39. Bakker JGC, van Grondelle R, den Hollander WTF: Trapping, loss and annihilation of excitations in a photosynthetic system. II. Experiments with the purple bacteria Rhodospirillum rubrum and Rhodopseudomonas capsulatus. Biochim Biophys Acta 1 g83, 725:508-518.

40. Sundstr6m V, van Grondelle R, BergstrSm H, ,~kesson E, Gillbro T: Excitation-energy transport in the bacteriochlorophyll antenna systems of Rhodospirillum rubrum and Rhodobacter sphaeroides studied by low-intensity picosecond absorption spectroscopy. Biochim Biophys Acta 1986, 851:431-446.

41. van Grondelle R, Bergstr6m H, SundstrSm V, Gillbro T: Energy transfer within the bacteriochlorophyll antenna of purple bacteria at 77K studied by picosecond absorption recovery. Biochim Biophys Acta 1987, 894:313-326.

42. Bradforth SE, Jiminez R, van Mourik F, van Grondelle R, Fleming GR: Excitation transfer in the core l ight-harvesting complex (LH-1) of Rhodobacter sphaeroides: an ultrafast fluorescence depolarization and annihilation study. J Phys Chem 1995, 99:16179-16191.

43. Jiminez R, Dikshit SN, Bradforth SE, Fleming GR: Electronic • . excitation transfer in the LH2 complex of Rhodobactar

sphaeroides. J Phys Chem 1996, 100:6825-6834. Fluorescence upconversion experiments on LH-2 yield two time constants in the anisotropy decay: 50 -90 fs and 400-500fs . 6800-->B850 transfer oc- curs within 650fs. Depolarization is modeled using a model with inhomoge- neous broadening (250 cm -1) and a dimer-to-dimer hopping time of 100fs. A calculation of the spectrum using an intradimer coupling of 230cm -1, ~and coupling between adjacent chromophores on different subunits of 110 cm -1 yields an average deloealization length of - 5 molecules in the middle of the band.

44. Visser HM, Somsen OJG, van Mourik F, Lin S, van Stokkum IHM, van Grondelle R: Direct observation of sub-picosecond equilibration of excitation energy in the light-harvesting antenna of Rhodospirillum rubrum. Biophys J 1995, 69:1083- 1099.

45. Visser HM, Somsen OJG, van Mourik F, van Grondelle R: Excited- , , state energy equilibration via sub-picosecond energy transfer

within the inhomogeneously broadened light-harvesting antenna of the LH-1 only Rhodobacter sphaeroides mutant M2192 at room temperature and 4.2K. J Phys Chem 1996, 100:18859-18867.

The spectral evolution following an ultrafast laser flash is studied for the core antenna of photosynthetic purple bacteria. The authors conclude that ultra- fast energy transfer occurs, resulting in a net shift of the excitation distribution to the low-energy pigments. At room and low temperature, the major part of the spectral shift takes less than a picosecond. The energy transfer dynamics at room temperature are fitted assuming a hopping time of - 100 fs between dimers on a ring and an inhomogeneous width of 400cm -1. At 4K hopping is somewhat slower, -0 .4 ps, and the inhomogeneous width has decreased to - 200cm -1.

46. Chachisvilis M, Pullerits T, Jones MR, Hunter CN, Sundstr~m V: Vibrational dynamics in the light harvesting complexes of the photosynthetic bacterium Rhodobacter sphaeroides. Chem Phys Lett 1994, 224:345-351.

47. Jean JM, Fleming GR: Competition between energy and phase relaxation in electronic curve crossing processes. J Chem Phys 1995, 103:2092-2101.

48. Pullerits 1", Chaehisvilis M, Sundstr6m V: Exciton delocalization • length in the B850 antenna of Rhodobacter sphaeroides.

J Phys Chem 1996, 100:10787-10792. Excitation transfer dynamics in LH2 of Rb. sphaeroides is investigated using transient absorption spectroscopy. In LH2, the anisotropy decays in 130 fs, whereas the isotropic decay occurs in ?0fs. For a ninefeld symmetric ring, with the orientation of the transition dipoles as in LH2, a factor of 3 between the two time constants is expected. As this is not observed, the authors con- clude that the energy migration is (partially) coherent. From an analysis of the transient absorption difference spectrum, they conctude that the excitation is delocalized over 4_+2 monomers.

49. Chachisvilis M, K(Jhn O, Pullerits T, Sundstr6m V: Excitons in • photosynthetic purple bacteria. Wavelike motion or incoherent

hopping? J Phys Chem 1997, in press. From a comparison of isotropic and anisotropic pump probe signals meas- ured for LH1 and LH2 of Rb. sphaeroides, the anisotropy decay is observed to be always about twofold slower. Modeling using a hopping model predicts a much more dramatic difference between isotropic and anisotropy decays. As a consequence, the authors believe that the excitation transfer is partly coherent. From a fit of the pump-probe spectrum, they conclude that the delocalization length is - 4 BChls and is not stongly dependent on temper- ature,

50. Monshouwer R, Abrahamsson M, van Mourik F, van Grondelle R: • • Superradiance and exciton delocalisation in bacterial

photosynthetic l ight-harvesting systems. J Phys Chem 1997, in press.

The radiative rate of LH1 and LH2 of Rb. sphaeroides are measured as a function of temperature. For LH2, the radiative rate is about threefold larger compared with that of monomeric BChl a and is independent of tempera- ture. LH1 is very similar to LH2 at room temperature, but the radiative rate increases about 2.4-fold upon lowering the temperature to 4K. The results are interpreted in terms of a model that includes both the coupling between all the pigments and the inhomogeneous broadening and suggests that the ratio between coupling and inhomogeneous broadening is -2 . The results suggest that in LH2 the excitation is rather localized at all temperatures. The degree of delocalization in LH1 may be somewhat larger, certainly at low temperature.

51. Nagarajan V, Alden RG, Williams JC, Parson WW: Ultrafast • exciton relaxation in the B850 antenna complex of

Femtosecond spectroscopy of photosynthetic light-harvesting systems Fleming and van Grondel le 747

Rhodobacter sphaeroides. Proc Nat/Acad Sci USA 1996, 93:13774-13779.

The femtosecond response of LH2 of Rb. sphaeroides at room temperature is measured. The authors observe a 35fs relaxation phase in absorption and emission spectra of the excited state, and a 20fs anisotropy decay. They ascribe these dynamics to interlevel relaxation and dephasing, respectively, of extensively delocalized exciton states of the circular bacteriochlorophyll aggregate.

52. Leegwater JA: Coherent versus incoherent energy transfer and • trapping in photosynthetic antenna complexes. J Phys Chem

1996, 100:14403-14409. A model is described which has an arbitrary ratio of homogeneous broad- ening versus interaction energy. This allows the study of the crossover from hopping dynamics to exciton dynamics. For the survival time, the hopping approach is shown to yield a surprisingly accurate estimate, even when the dynamics is excitonic. For LH1, it is estimated that the excitation is, on the average, delocalized over two dimers. The excitation is localized by phonons.

53. Meier T, Chernyak V, Mukamel S: Multiple exciton coherence • . sizes in photosynthetic antenna complexes viewed by

pump-probe spectroscopy. J Phys Chem 199"7, in press. From an analysis of the pump-probe signal from the LH2 light-harvesting antenna of purple bacteria, the localization size is determined to be 15 at 4.2K. The analysis of the difference in frequency between positive and neg- ative peaks in the pump-probe spectrum yields an estimate for the exciton mean free path (or the exciton dephasing lengthscale) of - 11.

54. Jimenez R, van Mourik F, Fleming GR: Three pulse echo • . peak shift measurements on LH1 and LH2 complexes of

Rhodobacter sphaeroides: a nonlinear spectroscopic probe of energy transfer. J Phys Chem 1997, in press.

Three pulse photon echo peak shift measurements are performed on the B875 and B850 bands of detergent-isolated LH1 and LH2 complexes at room temperature. The peak shifts are rnuch larger and decay much faster than those typically observed for dye molecules in solution. The peak shift decay is simulated on the basis of the optical frequency correlation function, M(t), which includes contributions from rapid fluctuations of the protein, vi- brational motion and energy transfer. The 90 fs and 130 fs exponential com- ponents in M(t) observed for LH1 and LH2, respectively, are ascribed to energy transfer. A simulation based on a model that assumes hopping in a ring of dimers, with each dimer randomly selected from an inhomogeneous distribution, explains the results.

55. Kumble R, Palese S, Visschers RW, Dutton PL, Hochstrasser RM: • Ultrafast dynamics within the B820 subunit f rom the core

(LH-1) antenna complex of Rs. rubrum. Chem Phys Let/1996, 261:396-401.

This paper describes polarized femtosecond transient absorption experi- ments on B820, the oc~-BChl 2 subunit of LH1, and on its reaggregated form, B873. In B820, the timescale of the Stokes' shift is sub 50fs, as reflected by a shift of the nuclear wavepacket and a fast component in the anisotropy decay. In similar experiments on reassociated B873, the anisotropy is ob- served to decay from 0.24 to 0.07 with two time constants: 70fs and 400 fs. The authors suggest that, following fast scattering, the excitation transfer proceeds between states of dimeric subunits.

56. Yu J-Y, Nagasawa Y, van Grondelle R, Fleming GR: Three pulse • . echo peak shift measurements on B820 subunit of LH1 of

Rhodospirillum rubrum. Chem Phys Let/1997, in press. This paper describes the measurement cf the three pulse photon echo peak shift for the LH1 subunit, B820, a protein bound BChl dimer. The major difference between B820 and LH1 is the absence of the 100fs exponential phase that was ascribed to energy transfer in LH1 and is now present as a nondecaying component. The experiment strongly supports the idea that, in LH1, the excitation is also localized on s BChl dimer.

5?. Shreve AP, Trautman JK, Frank HA, Owens TG, Albrecht AC: Femtosecond energy-transfer processes in the B800-850 l ight-harvesting complex of Rhodobacter sphaeroides 2.4.1. Biochim Biophys Acta 1991, 1058:280-288.

58. Ricci M, Bradforth SE, Jiminez R, Fleming GR: Internal • conversion and energy transfer dynamics of spheroidene in

solution and in the LH-1 and LH-2 light-harvesting complexes. Chem Phys Lett 1996, 259:381-390.

The lifetime of the 1Bu+ state of spheroidene is measured in vitro in vari- ous polar and nonpolar solvents and in LH1 and LH2 of Rb. sphaeroides. The 1Bu+-->2A - internal conversion time varies from 150fs to 250fs in g . the solvents studied and depends on the polarizability of the surrounding environment. Estimated internal conversion time is 150fs within LH2, and 170fs within LH1. The 1Bu + lifetime inside LH1 and LH2 is 60fs and 80fs, respectively, and suggests fast energy transfer from the 1Bu + state.

59. Andersson PO, Cogdell ILl, Gillbro T: Femtosecond dynamics • • of carotenoid-to-bacteriochlorophyll a energy transfer in the

l ight-harvesting antenna complexes from the purple bacterium Chromatium purpuratum. Chem Phys 1996, 210:195-217.

Energy transfer from the carotenoid okenone to BChl a in the light-harvesting complex B800-830 and in chromatophores of Chromatium purpuratum was studied by steady state fluorescence and femtosecond transient absorp- tion spectroscopy. The overall efficiency for energy transfer from okenone to BChl a is 95+5%. There is a fast (<200fs) transfer from at least one carotenoid to B830, and this occurs from the S 2 state of the carotenoid to the ex transition dipole of BChl a, probably employing the F~rster mecha- nism. On a longer timescale, the okenone S 1 state transfers energy to B830 in 3.8 ps, whereas a second carotenoid transfers its energy via the S t state to B800 in - 0.5 ps.

60. Freiberg A, Godik VI, Pullerits T, Timpmann K: Directed picosecond excitation transport in purple photosynthetic bacteria. Chem Phys 1988, 128:227-235.

61. Zhang FG, van Grondelle R, Sundstr6m V: Pathways of energy flow through the light-harvesting antenna of the photosynthetic purple bacterium Rhodobacter spheeroides. Biophys J 1992, 61:911-920.

62. Hess S, Chachisvilis M, Timpmann K, Jones MR, Fowler GJC, Hunter CN, Sundstr6m V: Temporally and spectrally resolved subpicosecond energy transfer within LH2 and from LH2 to LH1 in photosynthetic purple bacteria. Proc Nat/Acad Sci USA 1995, 92:12333-1233'7.

63. Nagarajan V, Parson WW: Excitation energy transfer between • the B850 and B875 antenna complexes of Rhodobacter

sphaeroides. Biochemistry 199'7, 36:2300-2306. In membrane of Rb. sphaeroides, energy transfer from B850 (LH2) to B875 (LH1) proceeds with two time constants, 4.6 ps and 26.3 ps, but a significant fraction of the excitations remain in B850 for considerably longer times. The fast step is ascribed to hopping from LH2 to LH1, the slow step to migration within the LH2 pool. Back transfer from LH1 to LH2 could not be detected.

64. Beekman LMP, van Mourik F, Jones MR, Visser HM, Hunter CN, van Grondelle R: Trapping kinetics in mutants of the photosynthetic purple bacterium Rhodobacter sphaeroides: influence of the charge separation rate and consequences for the rate-limiting step in the l ight-harvesting process. Biochemistry 1994, 33:3143-3147.

65. Pullerits T, Sundstr~m V: Photosynthetic l ight-harvesting o. p igment-proteins: toward understanding how and why. Acc

Chem Res 1996, 29:381-389. The energy transfer and trapping dynamics in photosynthetic purple bacteria is reviewed: a model is proposed for the LH2-LH1 association based on the measured LH2--~LH1 energy transfer time of 3.3ps at room temperature. Modeling this step gives a 3 nm distance of closest approach between the LH2 and LH1 rings.

66. Eads DD, Castner EW Jr, Alberte RS, Mets L, Fleming GR: Direct observation of energy transfer in a photosynthetic membrane: chlorophyll b to chlorophyll a transfer in LHCII. J Phys Chem 1990, 93:8271-8275.

67. Kwa SLS, van Amerongen H, Lin S, Dekker J P, van Grondelle R, Struve WS: Ultrafast energy transfer in LHC-II tr imers from the Chl a/b light-harvesting antenna of Photosystem II. Biochim Biophys Acts 1992, 1102:202-212.

68. Bittner T, Irrgang KD, Renger G, Wasielewski MR: Ultrafast excitation energy transfer and exciton-exciton annihilation processes in isolated light harvesting complexes of Photosystem II (LHCII) from spinach. J Phys Chem 1994, 98:11821-11826.

69. Du M, Xie X, Mets L, Fleming GR: Direct observat ion of ultrafast energy-transfer processes in l ight-harvesting complex II. J Phys Chem 1994, 98:4736-4741.

70. P~lsson LO, Spangfort MD, Gulbinas V, Gillbro T: Ultrafast chlorophyll ~-chlorophyll a excitation energy transfer in the isolated l ight harvesting complex, LHCII, of green plants. Implications for the organization of chlorophylls. FEBS Let/ 1994, 339:134-138.

71. Visser HM, Kleima FJ, van Stokkum IHM, van Grondelle R, • , van Amerongen H: Probing the many energy-transfer processes

in the photosynthetic light-harvesting complex II at 77K using energy-selective sub-picosecond transient absorption spectroscopy. Chem Phys 1996, 210:297-312.

The ultrafast energy transfer dynamics in LHCII is measured using transient absorption spectroscopy. Three phases in the Chl b--~Chl a energy transfer are resolved: 200 fs, 600fs and - 5 ps, with relative amplitude ratios of 40%,

748 Biophysical methods

40% and 20%, respectively. In the trimer, all the energy transfer occurs to the major red absorbing species at 676 nm. Measurements of singlet-singlet and singlet-triplet annihilation suggest that intermonomer energy transfer occurs on a timescale of 10-20 ps.

72. Connelly JP, M~ller MG, Hucke M, Gatzen G, Mullineaux CW, • Ruban AV, Horton P, Holzwarth AR: Ultrafast spectroscopy of

trimeric light-harvesting complex II from higher plants. J Phys Chem B 1997, 101:1902-1909.

These authors perform a highly sensitive transient absorption experiment to resolve all three phases in the Chl b-->Chl a energy transfer and find time constants of 175 fs, 600 fs and - 5 ps. Furthermore, they conclude from their data that, probably, the 175fs component partly reflected energy transfer between 'blue' and 'red' Chl bs.

73. Trinkunas G, Connelly JP, Mi.iller MG, Valkunas L, Holzwarth AR: A model for the excitation dynamics in the l ight-harvesting complex II from higher plants. J Phys Chern B 1997, in press.

74. Giuffra E, Cugini D, Croce R, Bassi R: Reconstitution and • pigment-binding properties of recombinant CP29. Eur J

Biochern 1996, 238:112-120. This paper describes the reconstitution of the minor chlorophyll a/b-binding protein CP29, overexpressed in Escherichia coil. The recombinant pigment- protein shows biochemical and spectroscopic properties identical to the native CP29 complex, with a Chl a:Chl b ratio ef three. Also other stoi- chiometries yielded stable complexes.

75. Peterman EJG, Monshouwer R, van Stokkum IHM, • - van Grondelle R, van Amerongen H: Ultrafast singlet excitation

transfer from carotenoids to chlorophylls via different pathways in light-harvesting complex II of higher plants. Chern Phys Lett 199"7, 264:279-284.

Energy transfer from the xanthophylls to the chlorophylls is studied in trimeric LHCII at 77K. No evidence for direct energy transfer from the xanthophylls to Chl b is found, whereas efficient xanthophyll to Chl a energy transfer occurs in 220fs. With preferential violaxanthin excitation (514 nrn) relative to lutein excitation (500 nm), energy transfer to Chl as absorbing at - 6 7 0 nm is more pronounced, compared with transfer to Chl as absorbing at - 6 7 6 nm. Fol- lowing 514 nm excitation, the transfer from 670 to 676 nm occurs in 2.1 ps.

76. Connelly JP, MiJIler MG, Bassi R, Croce R, Holzwarth AR: • . Femtosecond transient absorption study of carotenoid to

chlorophyll energy transfer in the light-harvesting complex II of Photosystem II. Biochemistry 10gT, 36:281-287.

Energy transfer from the xanthophylls to the chlorophylls is studied in LHCII trimers from Arabidopsis thafiana at room temperature. At 475 and 490 nm excitation, energy transfer is mainly from xanthophyll to Chl b.

7"7. Holzwarth AR, Schatz G, Brock H, Bittersmann E: Energy transfer and charge separation kinetics in Photosystem I. Part 1: Picosecond transient absorption and fluorescence study of cyanobacterial Photosystem I particles. Biophys J 1993, 64:1813-1822.

78. Hastings G, Hoshina S, Webber AN, Blankenship RE: Universality of energy and electron transfer processes in Photosystem I. Biochemistry 1995, 34:15512-15522.

79. Du M, Xie X, Jia Y, Mets L, Fleming GR: Direct observat ion of ultrafast energy transfer in PS1 core antenna. Chern Phys Lett 1993, 201:535-542.

80. Gobets B, van Amerongen H, Monshouwer R, Kruip J, R6gner M, van Grondelle R, Dekker JP: Polarized site-selection spectroscopy of isolated Photosystem 1 particles. Biochim Biophys Acta 1994, 1188:75-85.

81. P&lsson LO, Dekker JP, Schlodder E, Monshouwer R, • van Gronclelle R: Polarized site-selective fluorescence

spectroscopy of the long-wavelength emitting chlorophylls in isolated Photosystem I particles of Synechococcus elongatus. Photosynth Res 1996, 48:239-246.

Isolated trimeric PS1 complexes of Synechococcus e/ongatus have been studied in absorption and polarized fluorescence. Two types of long wave- length pigments are distinguished: C708 and C719. Their contribution to the absorption spectrum corresponds to ~ 4 -5 C708 and 5 -6 C? 19 per PT00. From low-temperature energy-selective polarized fluorescence experiments, it is concluded that at ultra low temperatures C708 still is able to transfer ex- citation energy to C719 and furthermore that energy transfer among C719s occurs.

82. Valkunas L, Liuolia V, Dekker JP, van Grondelle R: Description of energy migration and trapping in Photosystem I by a model with two distance scaling parameters. Photosynth Res 1995, 43:149-154.

83. Kumazaki S, Ikegami I, Yoshihara K: Excitation and electron • t ransfer f rom selectively excited primary donor chlorophyll

(P700) in a Photosystem I reaction center. J Phys Chem A 1997, 101:597-604.

Primary processes in a Photosystem 1 reaction center are studied using sub- picosecond fluorescence upconversion. In these enriched reaction centers there are - 14 Chl as per PTO0. The 1 ps fluorescence anisotropy decay fol- lowing selective PT00 excitation is ascribed to equilibration between P?00 and the surrounding antenna Chls. In the isotropic fluorescence decay, at least two components can be distinguished: 2.2 ps (35%) and 15 ps (55o/o). The fast and slow phases are interpreted in terms of charge separation before and after full equilibration of the excited state, respectively. From kinetic modeling, the intrinsic time constant for charge separation from PT00 is concluded to be < 4 ps.

84. Trinkunas G, Holzwarth AR: Kinetic modeling of exciton • migration in photosynthetic systems. 3. Application of genetic

algorithms to simulations of excitation dynamics in three- dimensional Photosystem I core antenna/reaction center complexes. Biophys J 1996, 71:351-364.

This paper describes calculations of energy transfer and trapping in Pho- tosystem 1 using a genetic algorithm. Various 3D models for the pigment arrangement and the corresponding energy transfer dynamics are tested for Photosystem 1. It is concluded that: the red pigments never are close to P700; the red pigments are also never far away from P700 and tend to cluster; the charge separation time is shorter than 1.2 ps; and the total energy transfer time within the main antenna pool is < 1 ps.

85. White NTH, Beddard GS, Thorne JRG, Feehan TM, Keyes TE, o- Heathcote P: Primary charge separation and energy transfer

in the Photosystem I reaction center of higher plants. J Phys Chern 1996, 100:12086-12099.

A detailed analysis is given of the Photosystem 1 trapping kinetics in a Photosystem 1 core particle from plants. The A0--A 0 difference spectrum takes 3 ps to form upon ?08 nm excitation reflecting the equilibration time between the first charge separated state and the antenna excitation. The equilibrated state decays in 20-20 ps. From extensive modeling, the intrinsic rate of electron transfer is concluded to be 0.7 ps -1.

86. Melkozernov AN, Su H, Lin $, Bingham S, Webber AN, • ° Blankenship RE: Specific mutation near the primary donor

in Photosystem 1 from Chlamydamonas reinhardtii alters the trapping time and spectroscopic properties of P7oo- Biochemistry 1997, 36:2898-2907

Time-resolved absorption and fluorescence spectroscopy are used to inves- tigate the energy and electron transfer processes in the detergent-isolated Photosystem I core particles from the site directed mutant of Ch/arnyda- rnonas reinhardtiiwith the His656 of PsaB replaced by asparagine. There is no indication that the mutation affects the spectral distribution in the antenna; however, the excited state lifetime increases from - 3 0 ps to - 6 5 ps. It is proposed that the excited state decay is limited by charge separation.