Light-induced spin polarization in type I photosynthetic reaction centres

Biochimica et Biophysica Acta 1507 (2001) 212^225 www.bba-direct.com

Review

Light-induced spin polarization in type I photosynthetic reaction centres Arthur van der Est Department of Chemistry, Brock University, 500 Glenridge Avenue, St. Catharines, ON, Canada L2S 3A1 Received 11 December 2000; received in revised form 20 April 2001; accepted 18 June 2001

Abstract The use of light-induced spin polarization to study the structure and function of type I reaction centres is reviewed. The absorption of light by these systems generates a series of sequential radical pairs, which exhibit spin polarization as a result of the correlation of the unpaired electron spins. A description of how the polarization patterns can be used to deduce the relative orientation of the radicals is given and the most important structural results from such studies on photosystem I (PS I) are summarized. Quinone exchange experiments which demonstrate the influence of protein^cofactor interactions on the polarization patterns are discussed. The results show that there are significant differences between the binding sites of the primary quinone acceptors in PS I and purple bacterial reaction centres and suggest that Z^Z interactions probably play a more important role in PS I. Studies using spin-polarized EPR transients and spectra to investigate the electron transfer pathway and kinetics are also reviewed. The results from PS I, green-sulphur bacteria and Heliobacteria are compared and the controversy surrounding the role of a quinone in the electron transfer in the latter two systems is discussed. ß 2001 Elsevier Science B.V. All rights reserved. Keywords: Electron spin resonance; Electron transfer; Quinone; Radical pair; Iron^sulphur center

1. Introduction The emerging picture of the structure and function of photosystem I (PS I) documented in the articles in this volume is the result of many years of spectroscopic investigation. X-ray crystallography, optical spectroscopy and electron paramagnetic resonance spectroscopy (EPR) have played a predominant role in these studies, each providing unique information. The determination of the X-ray structure represents an enormous achievement. However, the structure alone is not su¤cient to understand the function of PS I. The kinetics and pathway of the light-induced energy and electron transfer have been

E-mail address: [email protected] (A. van der Est).

studied extensively using optical methods but such data give relatively few structural details. Time resolved EPR provides a link between these two types of data by allowing both the kinetics and orientation of the redox active cofactors to be investigated simultaneously. In addition, essentially all important forms of the protein complex are accessible to transient EPR, from whole cells of cyanobacteria under physiological conditions to single crystals of isolated PS I reactions centres at cryogenic temperatures. Photosynthetic reaction centres (RCs) have been described as a `playground' for the EPR spectroscopist because of the many signals which can be measured. Standard EPR and electron nuclear double resonance (ENDOR) can be used to examine the electronic structure of the cofactor radical ions generated by steady state illumination. Pulsed EPR

0005-2728 / 01 / $ ^ see front matter ß 2001 Elsevier Science B.V. All rights reserved. PII: S 0 0 0 5 - 2 7 2 8 ( 0 1 ) 0 0 2 0 4 - 3

BBABIO 45078 12-10-01

A. van der Est / Biochimica et Biophysica Acta 1507 (2001) 212^225

methods can be applied to the same species to give information about their spin dynamics from which conclusions about the motion and binding of the cofactors can be deduced. When combined with pulsed light excitation, transient EPR and pulsed EPR can be used to study the light-induced radical pairs, which are the functional states of the RC. Alternatively, the triplet state of the primary electron donor can be generated by charge recombination when the forward electron transfer is blocked. This review will focus on transient EPR studies of light-induced electron spin polarization (ESP) in type I RCs. ESP is a characteristic feature of light-induced processes which occurs because the pathway of the reaction is spin dependent and/or because the correlation of the electron spins in the initial state of the system persists for a considerable time after the reaction has taken place. As a result, the spin states of the products are populated selectively. In photolysis reactions in solution, the spin polarization is referred to as CIDEP (chemically induced dynamic electron polarization) and the early literature also refers to the spin polarization in reaction centres as CIDEP. However, this terminology was dropped when it became clear that the signals from RCs are due to coupled radical pairs with a ¢xed geometry as opposed to the uncoupled randomly oriented radicals observed in solution. In RCs, light-induced spin polarization occurs because the electron is transferred from the excited singlet state of the donor, in which the electrons are spin paired. This leads to the spatial separation of a pair of correlated particles with selective population of the spin states with singlet character in the resulting radical pairs. The spin polarization manifests itself as very intense, short-lived EPR signals which can be either absorptive or emissive. In addition, coherence e¡ects such as transient nutations, quantum beats and nuclear modulations are observed. Apart from providing unambiguous evidence that light-induced electron transfer occurs, transient EPR spectroscopy is sensitive to the orientation and geometry of the radicals involved because the magnetic interactions of the spins are tensorial and anisotropic. Moreover, the data depend on the dipolar coupling and exchange interaction, which are measures of the distance between the two radicals and

213

the overlap of their electronic wavefunctions, respectively. The time dependence of the light-induced EPR signals can be used to study the electron transfer kinetics and the interactions of the cofactors with the surrounding protein are re£ected in the g-factors and hyper¢ne couplings of the radicals. Several previous reviews have included discussions of electron spin polarization in RCs [1^8] in the broader context of photosynthesis and electron spin resonance. This article is narrower in scope and focuses primarily on PS I and the di¡erences between the quinone acceptors in type I and type II RCs. A signi¢cant development in the past several years in this area has been the extension of such studies to single crystals [9] and to high magnetic ¢elds/microwave frequencies [8]. The microwave frequency dependence of the spectra not only provides more data on which the parameters can be ¢xed but also allows the frequency dependent Zeeman interaction to be varied independently of the other magnetic interactions. Thus, even a moderate increase in frequency to 24 or 35 GHz (K-band or Q-band) is often su¤cient to distinguish between the various contributions to the spectra. The availability of time-resolved high¢eld EPR at 95 GHz (W-band) has lead to a huge increase in the spectral resolution and the ease with which the data can be interpreted. Before discussing speci¢c questions which have been addressed by transient EPR, a brief summary of the early literature and relevant background information will be presented. 2. Transient EPR of PS I Reaction centres fall into two classes, based on the nature of the terminal acceptors, referred to as type I and type II in reference to PS I and PS II in oxygenic organisms. Type I reaction centres have a series of iron^sulphur clusters, whereas type II RCs contain a non-haem iron £anked by two quinones. This di¡erence has important consequences for transient EPR. The non-haem iron in type II RCs is paramagnetic (S = 2), and interacts magnetically with the unpaired electrons generated by the light-induced electron transfer. As a result, the spectra depend on three spins and are di¤cult to interpret. More importantly, the large zero-¢eld-splitting of the iron broadens the

BBABIO 45078 12-10-01

214

A. van der Est / Biochimica et Biophysica Acta 1507 (2001) 212^225

spectra and leads to rapid relaxation of the spin polarization. The net result is that without either chemical modi¢cation of the iron or low temperature, no spin polarization is observed from type II reaction centres. Type I RCs, on the other hand, do not contain a non-haem iron and show strong spin-polarized EPR signals which are readily interpreted. An important consequence of the di¡erences between the two RC-types is that transient EPR can be used to study PS I in intact cells or chloroplasts without interference from PS II. 2.1. Room temperature signals Light-induced spin polarization from PS I at room temperature was ¢rst reported by Blankenship et al. [10]. The species giving rise to the signals was identi¢ed as P‡c 700 [11,12]. However, the interpretation of the spin polarization using CIDEP theory for uncoupled radicals remained inconclusive. Some of these early experiments and their initial interpretation are reviewed in [13]. Spectra measured at higher microwave frequency [14,16] showed signal contributions which were attributed to the quinone acceptor. With improvements in the time resolution two se-

quential spin-polarized EPR spectra were resolved [17,18]. An example is shown in Fig. 1 for PS I particles from the cyanobacterium Synechococcus elongatus. The upper part of Fig. 1 shows the complete time/¢eld data set constructed by collecting the transient responses to pulsed (10 ns) laser excitation at a series of ¢xed magnetic ¢elds. Positive signals represent microwave absorption (A) and negative signals are emission (E). The transition from an E/A/E pattern at early times to a predominantly emissive pattern at late times is evident in the dataset. The transient corresponding B0 = 339.0 mT is shown in the lower left part of Fig. 1. The absorptive contribution 3c at early times is due to the state P‡c 700 A1 and the ‡c 3 emissive part is due to P700 FeS , where FeS3 is one of the three iron^sulphur centres. Decay associated spectra corresponding to the two states are shown in the lower right of Fig. 1. The spectra are extracted by ¢tting the transients with a kinetic model and plotting the amplitudes of the decay components as a function of magnetic ¢eld. The emissive spectrum represents the P‡c 700 contribution to the spec3 trum of P‡c FeS and corresponds to the spectra 700 reported in [10^13]. An important point is that the decay of the signals is in£uenced by both forward

Fig. 1. Spin-polarized transient EPR data from PS I particles at room temperature. (Top) Complete time/¢eld data set. (Bottom left) Transient taken near the middle of the data set. (Bottom right) Decay-associated spectra extracted from the data set.

BBABIO 45078 12-10-01

A. van der Est / Biochimica et Biophysica Acta 1507 (2001) 212^225

215

Fig. 2. Spin-polarized transient EPR data from PS I particles at 150 K. (Top) Complete time/¢eld data set. (Bottom left) Transient taken near the middle of the dataset. (Bottom right) Boxcar spectrum extracted from the data set.

electron transfer and spin-lattice relaxation of the spin polarization. Thus, while the transition from the absorptive to emissive signal in Fig. 1 is caused by the electron transfer from A3c 1 to the iron sulphur centres, the decay of the emissive signal is primarily due to spin-lattice relaxation. 2.2. Low temperature signals In PS I the charge separation from P700 to the terminal iron^sulphur centres FA /FB is irreversible at low temperature and photo-accumulation of the 3 state P‡c occurs. Because transient EPR 700 (FA /FB ) measurements require a cyclic photo-induced process, at ¢rst sight low temperature experiments would seem unfeasible. Fortunately, charge recombination from A3c 1 occurs in a fraction of the centres [19,20]. Below 77 K stable reduction of FA /FB occurs in only about 35% of the centres and forward electron transfer past A3c 1 is blocked in the remaining 65%. Under 3c these conditions the lifetime of P‡c 700 A1 is t1=2 W200 Ws [21]. Fig. 2 shows a typical data set for PS I taken at 150 K. As can be seen, there is no emissive contribution due to P‡c (FeS)3 , and only the E/A/E pat3c tern due to P‡c 700 A1 is observed. The signal decay is

governed by the relaxation of the spin polarization and is also dependent on the microwave power. The data shown in Figs. 1 and 2 have been measured using `direct detection'. This term is used to describe detection without the usual ¢eld modulation and is something of a misnomer because the signal is actually detected indirectly using a resonator. Alternatively, light and/or magnetic ¢eld modulation techniques or pulsed EPR have also been used to detect spin polarization in PS I (see, e.g., [2,22,23]). Indeed, the experiments at K-band and Q-band [14^16] which ¢rst revealed the contribution from the quinone to the spin-polarized signals were detected us3c ing light modulation. A typical spectrum of P‡c 700 A1 detected using ¢eld modulation at X-band is shown in Fig. 3. Note that the ¢rst derivative of the absorption spectrum with respect to the ¢eld is observed. 3. The kinetics and pathway of electron transfer As can be seen from Fig. 1, transient EPR data contain information about the electron transfer from A3c 1 to the iron^sulphur centres. The advantage of EPR over optical methods for studying this process

BBABIO 45078 12-10-01

216

A. van der Est / Biochimica et Biophysica Acta 1507 (2001) 212^225

3c Fig. 3. The spin-polarized EPR spectrum of P‡c 700 A1 measured using ¢eld modulation detection at 150 K. The spectrum is the average signal intensity in a time window between 0 and 200 Ws following the laser £ash.

is that it can be applied to a whole range of di¡erent samples ranging from isolated reaction centres to whole cells. In addition, the signals are free of interference from background signals because of the strong spin polarization and can be measured on completely intact systems. By comparison, optical methods have a higher sensitivity and better time resolution. Initial determinations of the electron transfer rate from A3c 1 to the iron^sulphur centres in PS I gave con£icting results. While transient EPR experiments gave a value of t1=e = 290 ns [18] in agreement with the optical results of Brettel [24], Mathis and Se¨tif [25] found a lifetime of t1=e = 36 ns. Later measurements showed that this discrepancy was due to di¡erences in the preparations used and that both rate constants could be observed in spinach preparations, whereas the slower rate was predominant in cyanobacterial PS I particles [26]. The origin of the biphasic electron transfer was also investigated using transient EPR to study a wide range of PS I preparations. Although, the 36 ns phase is not detected directly, it has a clear e¡ect on the relative amplitudes of the contributions to the transients. This is demonstrated in Fig. 4 for two spinach preparations. The transient in the upper part of the ¢gure is from chloroplasts with a small amount of the fast component, whereas the lower transient is from PS IL particles which have a large fraction of centres showing fast electron transfer (see [27] for preparation details). This di¡erence is re£ected in the very di¡erent relative amplitudes of

the absorptive and emissive parts of the transients. Thus, the relative proportions of the two kinetic phases can be determined from the EPR data. More importantly, this ratio can be determined in intact preparations, without interference from PS II. The rate constant for the slower phase is found to be k31 et = 290 þ 50 ns for all preparations [28] and evidence for the fast phase was only observed in isolated particles. In particular, the use of the detergent Triton X100 to solubilize the reaction centre complexes leads to a large amount of fast phase [26]. Yang et al. [29] have shown that solubilization of thylakoids with Triton can lead to the loss of PsaE and PsaF. Moreover in mutants devoid of PsaE and/ or PsaF, particles isolated using Triton X-100 displayed dramatically di¡erent photoreduction behaviour under reducing conditions compared to particles isolated with L-dodeclymaltoside. These results clearly suggest that the fast phase could be an artefact brought about by the in£uence of the detergent. On the other hand, Joliot and Joliot [30] recently

Fig. 4. Comparison of spin-polarized EPR transients from two spinach PS I samples. The top trace is from chloroplasts before detergent treatment. The lower trace is from PS I-L particles isolated using Triton X-100. The di¡erence absorptive part of 3c the signal is due to P‡c the emissive part is due to 700 A1 ‡c 3 P700 FeS . The relatively weak absorptive contribution in the lower transient is due to large fraction of RCs in which fast electron transfer to the iron^sulphur clusters occurs.

BBABIO 45078 12-10-01

A. van der Est / Biochimica et Biophysica Acta 1507 (2001) 212^225

Fig. 5. The e¡ect on selected room temperature spin-polarized EPR transients of successive removal of the iron^sulphur clusters by urea treatment.

reported equal amounts of the two kinetic phases in whole cells of PS II-depleted mutants, which indicates that although the use of Triton in£uences the relative amounts of the two phases, the fast phase is an inherent feature of intact PS I. One interpretation of the biphasic behaviour is that it represents bidirectional electron transfer. Indeed, recent studies of site directed mutants in the two branches [31] add considerable weight to this argument. Although this is an appealing explanation, it con£icts with much of the existing evidence mainly from EPR experiments. Recent ENDOR studies [33] show that the spin density on P‡c 700 is localized on the chlorophyll which is axially coordinated by His(B656). If electron transfer to both quinones were to occur, this would lead to signi¢cantly di¡erent geometries for the resulting two radical pairs. However, no clear evidence of this is observed in any of the spin polarization patterns, echo modulation curves, etc. Moreover, studies of subunit deletion mutants [29] show e¡ects only when subunits PsaE and PsaF, which lie closer to

217

the PsaA side of the reaction centre, are removed. Indeed, in [31] it is reported that the EPR spectrum A3c 1 of changes only when point mutations are made near the quinone binding site in PsaA. This is also consistent with the ¢ndings of Purton et al. [32], who observed changes in the EPR and ENDOR signals of 3c photoaccumlated A3c 1 and the kinetics of the A1 to FX electron transfer when tryptophan PsaA W693 in Chlamydomonas reinhardtii is changed to either a histidine or a leucine. If the localization of the spin density on the primary donor is used as a criterium for drawing an analogy between PS I and bRC, the active branch of acceptors again corresponds to PsaA. Thus, there is both evidence for bi-directional electron transfer as well as uni-directional electron transfer along the PsaA branch. Clearly, further studies are needed to resolve this issue. In all of the early studies of the electron transfer past A1 it was unclear which of the three iron^sulphur centres, FA , FB or FX was acting as the electron acceptor. This question was investigated simultaneously and independently by three groups using transient EPR [28], pulsed EPR [34] and transient absorption spectroscopy [35] by successive removal of the three iron^sulphur clusters. The transient EPR results are shown in Fig. 5. The fact that (i) the emissive contribution to the transients is present in all preparations containing FX but absent when FX is removed, and (ii) the rate of electron transfer past A3c 1 is unchanged in the absence of FA and FB indicates that it proceeds from A1 to FX . However, it is important to point out that the rate of the subsequent electron transfer from FX to (FA /FB ) is still an open question. Photovoltage measurements [36] suggest that it proceeds from FX to FA and that the rate of this step is faster than the transfer from A3c 1 to FX . In this case, the spin-polarized EPR spec3 trum observed at late time would be due to P‡c 700 FA . The di¡erences in the spin polarization patterns ex‡c 3 ‡c 3 3 pected for P‡c 700 FX , P700 FA and P700 FB were investigated in [37] and found to be too subtle to interpret in the experimental spectra with any con¢dence. 4. Structural information In addition to obtaining the electron transfer rate from A3c 1 to the iron^sulphur centres, the light-in-

BBABIO 45078 12-10-01

218

A. van der Est / Biochimica et Biophysica Acta 1507 (2001) 212^225

duced spin polarization in PS I can be used to de3c termine the orientation of P‡c 700 and A1 . The description of the charge separated state as a correlated coupled radical pair (CCRP) with a singlet precursor was ¢rst introduced by Closs et al. [38] and Buckley et al. [39] and later generalized by Norris et al. [40]. These studies showed that the polarization pattern of such a system depends on the ¢ve angles needed to describe the relative orientation of the g-tensors of the two radical partners and the distance vector between them. These angles are di¤cult to extract if only X-band data are used because of the extreme overlap of the g-tensor components and the line broadening caused by unresolved hyper¢ne splittings. 3c By measuring the spectrum of P‡c 700 A1 in fully deuterated whole cells of Synechococcus sp. at K-band (24 GHz), Stehlik et al. [41] obtained greatly improved spectral resolution and showed that A3c 1 is oriented with its carbonyl bonds directed towards P‡c 700 (within 20³ of the PA1 vector). The spectra of ‡c 3c 3c P‡c 700 A1 were compared to those of P865 QA from ZnbRC at X- and K-band by Fu«chsle et al. [42]. An important conclusion from this study was that the orientation of A1 is quite di¡erent from that of QA in purple bacterial reaction centres. This is illustrated very clearly in the more recent comparison of ‡c 3c 3c P‡c 700 A1 and P865 QA at three microwave frequencies [43] shown in Fig. 6. The W-band (95 GHz) spectra, in particular, show a marked di¡erence on the low ¢eld side. This region of the spectra is dominated by the x and y components of the respective quinone gtensors and the sign of the polarization is determined by the sign of the dipolar coupling when a given gtensor axis is parallel to the magnetic ¢eld. Thus, the di¡erences in the polarization patterns in Fig. 6 are due to the di¡erent quinone orientations in the respective RCs. It is important to emphasize, however, that to go beyond a qualitative analysis of the polarization patterns, additional information is required. The values of the magnetic parameters, g-tensors and hyper¢ne tensors, can be measured independently by cw-EPR and ENDOR. The orientation of the tensor axes relative to the respective molecular frames presents a more di¤cult problem. In the case of the chlorophyll donor, they have come high ¢eld EPR studies of P‡c 865 [44] and P‡c [45] in single crystals. For the quinone 700 acceptor, theoretical considerations [46,47] are used

Fig. 6. Comparison of the spin-polarized boxcar spectra of ‡c 3c 3c P‡c 700 A1 in PS I and P865 QA in Zn-bRCs at three microwave frequencies [43].

to ¢x the g-tensor axes. Recently, we have proposed a simpli¢ed strategy for obtaining the geometry of weakly coupled radical pairs with a singlet precursor 3c [48]. Under the conditions which hold for P‡c 700 A1 ‡c and P865 Q3c A at least one angle must be ¢xed if a unique geometry is to be obtained from the polarization patterns of unoriented samples. In addition, to the spin polarization patterns outof-phase echo experiments [49,50] have been used to 3c determine the distance between P‡c 700 and A1 (see the contribution from R. Bittl in this volume for a detailed discussion). Rutherford and co-workers have studied the photoaccumulated [51,52] acceptors using a various EPR techniques to determine structural features of PS I. Using the spectra of oriented mem-

BBABIO 45078 12-10-01

A. van der Est / Biochimica et Biophysica Acta 1507 (2001) 212^225

branes and angles of the carbonyl bonds relative to the PA1 vector obtained from the spin-polarized spectra [43], they were able to estimate the orientation of A3c 1 relative to the membrane normal [51]. 4.1. Coherence e¡ects The correlation between the electron spins in the initial singlet state manifests itself in coherence effects observed in the time dependence of the spinpolarized EPR signals. Both transient nutations due precession of the magnetization about the microwave ¢eld [16] and quantum beats [53] due to free motion of the spin system have been observed and these e¡ects also depend on the geometry of the radical pair. Gierer et al. [54] showed that the frequency of the transient nutations is dependent on the strength of the dipolar coupling and is thus orientation dependent. Kothe and co-workers [55] analysed the 3c quantum beats from P‡c 700 A1 to obtain a prediction of the geometry of radical pair similar to that obtained from the spectra [41]. However, the analysis of these e¡ects is very complex and yields the same angles which are more easily obtainable from the spin polarization patterns. In addition to the transient nutations and quantum beats, coherent oscillations resulting from the nuclear spins also occur and were ¢rst observed by Bittl et al. [56] in bacterial reaction centres. Kothe and co-workers have studied these e¡ects extensively in PS I [57^59] and have analysed the transients to obtain hyper¢ne couplings in the radical partners. 4.2. Oriented samples The data discussed above yield structural parameters which are internal to the radical pairs. However, if the orientation of the radicals relative to the protein is to be determined, it is necessary to orient the sample macroscopically. The ¢rst spin polarization patterns from PS I single crystals were reported by Kamlowski et al. [9] and in a parallel study Bittl et al. [60] measured out-of-phase spin echo modulations on the same samples. Together, the transient EPR and out-of-phase echo data ¢x both the orientation and position of A1 within the reaction centre. These results were then used by Klukas et al. [61] to con¢rm the position and orientation of A1 found in a

219

î resolution electron density map. Recently, Bert4A hold et al. [62] were able to orient PS I reaction centres using the magnetic ¢eld of a W-band spectrometer. From an analysis of the orientation dependence of the spin polarization patterns, they were able to reproduce the geometry of the radical pair 3c P‡c 700 A1 found in the more extensive single crystal data. 4.3. Quinone exchange experiments The nature of the A1 binding site has also been investigated using transient EPR. One approach has been to use spin-polarized EPR data from quinone exchanged samples to study the kinetics, electronic and structural properties of foreign quinones in the A1 site. The majority of these studies involve extraction of the native phylloquinone using organic solvents followed by reconstitution with a non-native electron acceptor. (See also the contribution by S. Itoh in this volume.) Experiments of this type were used by Sieckmann et al. [63] and Synder et al. [64] to provide de¢nitive evidence that A1 is indeed phylloquinone. Rustandi et al. [65] used the presence or absence of radical pair spin polarization to investigate the ability of a wide range of quinones and other compounds to accept electrons from A0 . The fact that the native quinones in both bacterial reaction centres and PS I can be exchanged opens the possibility of comparing the in£uence of the respective binding sites on a given quinone. Comparisons of this type were made in [66] and it was shown that the di¡erent protein-cofactor interactions in the two RCs leads to a shift of the x-component of the quinone g-tensor. This is demonstrated for naphthoquinone and duroquinone in Fig. 7. The shift to larger g-values (lower ¢eld) implies weaker H-band and/or stronger Z^Z interactions in PS I. However, it is dif¢cult to draw de¢nite conclusion from such e¡ects because of the complex relationship between the gtensor and the quinone environment. A surprising result of such comparisons is that various small quinones in the A1 binding site take on very di¡erent orientations whereas the same quinones in the QA site all have essentially the same orientation. This is demonstrated in Fig. 8. The PS I spectra in the top part of the ¢gure show very strong di¡erences on the low-¢eld side indicative of

BBABIO 45078 12-10-01

220

A. van der Est / Biochimica et Biophysica Acta 1507 (2001) 212^225

avoids these di¤culties was developed by Chitnis and Golbeck [67]. By deleting genes for enzymes involved in the biosynthesis of phylloquinone, they were able to generate mutant strains of Synechocystis PCC6803 devoid of phylloquinone. Using a range of spectroscopic results, including the spin polarization patterns, Zybailov et al. [68] showed that in the absence of phylloquinone, the organisms incorporate plastoquinone into the A1 site. Moreover, kinetic studies [69] again involving transient EPR, show that forward electron transfer to the iron^sulphur clusters is maintained but with a reduced rate in the mutants. 5. Quinone binding site

Fig. 7. Comparison of the spin-polarization patterns of P‡c Q3c in PS I and purple bacterial reaction centres containing naphthoquinone and duroquinone. Note the shift of the low-¢eld features due to a larger g-anisotropy of the quinones in the A1 site.

large di¡erences in the quinone orientation. In contrast, the corresponding spectra in bacterial reaction centres (bottom) are all very similar. An analysis of the spectra for PS I shows that naphthoquinone is oriented with its CNO bonds perpendicular to the P^Q distance vector while for native phylloquinone, the carbonyl bonds are parallel to the distance vector. Such a large di¡erence in quinone orientation precludes the same H-bonding arrangement in both cases. Thus, the implication of this result is that Hbonds to the CNO groups are not the dominant factor in determining the orientation of the quinone headgroup in PS I. However, because the orientation of the molecular plane of the quinone is the same in both cases and Z^Z interactions depend on this orientation, it is conceivable that they determine the orientation of the quinone. A valid criticism of such studies is that the extraction procedures are very harsh and an unknown amount of damage to the RC occurs. Recently, a more elegant method of quinone exchange which

Using the structural information from the single crystal EPR data [9,60] together with the positions î resolution X-ray structure of the helices from the 4 A [70], Kamlowski et al. [71] constructed a detailed set of possible models of the A1 binding site. In all of the constructed models a nearly coplanar arrangement of the quinone headgroup and the aromatic

Fig. 8. Comparison of the spin-polarization patterns of various quinones in PS I (top) and purple bacterial reaction centres (bottom). Notice the di¡erences in the low-¢eld sides of the PS I spectra compared to the very similar spectra for Zn-bRC.

BBABIO 45078 12-10-01

A. van der Est / Biochimica et Biophysica Acta 1507 (2001) 212^225

tryptophan side chain of TrpA697 (or TrpB677) was found, which is consistent with the role of Z^Z interactions as suggested by the large g-tensor anisotropy of A3c observed in transient EPR experiments 1 [43,66]. However, the interpretation of quinone gtensor anisotropy in not straightforward and more direct experimental evidence for such interactions has come from ESEEM experiments on 15 N-labelled PS I [72] in which it was shown that the unpaired electron on A3c 1 is coupled to the nitrogen of a tryptophan. On the other hand, ENDOR data on photoaccumulated A3c 1 [73^76] suggest the presence of hydrogen bonds between A1 and the protein. The comparatively weak couplings to exchangeable protons, the large g-anisotropy of A3c 1 [43,45,51] and the drastic di¡erences in the orientation of the exchanged quinones all suggest that such H-bonds are weaker in PS I than in bRCs. A comparison [71] of the modelled A1 binding site and the structure of the QA site in bRCs indicated that while a coplanar tryptophan-quinone arrangement exists in both RCs, the histidine residue to which QA is strongly H-bonded has no analogue in PS I. Moreover, the corresponding oxygen on the quinone in PS I does not appear to be H-bonded. For QA it is the carbonyl in the meta position relative to the phytyl tail which is strongly H-bonded. For A1 the large hyper¢ne coupling to the phylloquinone methyl protons indicates an opposite asymmetry in the spin density distribution, suggesting that one-sided H-bonding at the carbonyl adjacent to the phytyl tail may be a feature of the quinone binding. Thus, from the EPR data and modelling studies it can be concluded that while Z stacking with a tryptophan is present in the quinone binding site in both RCs, the hydrogen bonding situation is very di¡erent. These features have now been con¢rmed by the most recent X-ray structure [77] which shows the predicted Z stacking arrangement of the two phylloquinones with TrpA697 and TrpB677, respectively. The plane-to-plane distances range from î , suggesting a stronger interaction than in 3.0 to 3.5 A the comparable stacking in QA site in bRC. In agreement with the observed spin density distribution for A1 one-sided H-bonding is observed between the carbonyl oxygen adjacent to the phytyl chain of the respective quinones and the backbone NH groups of LeuA722 and LeuB706. The results of the quinone exchange experiments

221

suggest that there is a subtle balance of forces holding the quinone in place. Apart from the Z stacking and H-bonding, the phytyl tail and methyl group also appear to play a role. In the absence of the these two side groups, the quinone is rotated by 90³ about the normal to the aromatic plane. In this orientation, the H-bond to the backbone NH group of LeuA722 (LeuB706) cannot be maintained without a signi¢cant rearrangement of the protein. Thus, it is probably not involved in binding naphthoquinone, which implies that the phytyl tail and/or methyl group are necessary for the formation of the H-bond. 6. Heliobacteria and green-sulphur bacteria These two very primitive organisms have type 1 reaction centres similar to PS I; however, they have fewer polypeptides and a considerably lower molecular mass. Despite this simplicity, the details of the electron transfer are not ¢rmly established. In particular, it is not known whether a quinone analogous to A1 in PS I acts as an acceptor between the primary chlorophyll acceptor A0 and the ¢rst iron^sulphur centre FX . This is a point of controversy with transient absorptions and photovoltage measurements indicating that the electron transfer from A0 proceeds directly to FX with a time constant of dW600^700 ps [78,79]. In apparent contradiction to this, a semiquinone radical is photo-accumulated [80^83] when electron transfer to the iron^sulphur centres is blocked. In light of the detailed information obtained for the quinone acceptors in PS I and bRCs, transient EPR is ideally suited for investigating the role of the quinone as an electron acceptor in RCs from Heliobacteria and green-sulphur bacteria. Initial attempts at these types of experiments showed that spin-polarized transient EPR signals could be measured. However, their interpretation was unclear. The ¢rst such experiments were reported by Heathcote and Warden [84] using an X-band spectrometer with V10 Ws time resolution and a spectrum which was thought to be similar to that of PS I was observed for green-sulphur bacteria. More recently [2], it was pointed out that the spectrum from green-sulphur bacteria in fact resembles that observed for some purple bacterial samples. However, the data in [84,2] are di¤cult to interpret because of the low time resolution and the

BBABIO 45078 12-10-01

222

A. van der Est / Biochimica et Biophysica Acta 1507 (2001) 212^225

fact that only one frequency band has been measured. In [85], X- and K-band spectra of Heliobacteria and green-sulphur bacteria taken with V20 ns time resolution at room temperature and 100 K were studied. In contrast to PS I, electron transfer from A1 to FX is not observed in the time range accessible by transient EPR. The comparison of the spin polarization observed during the ¢rst 100 ns after the laser £ash is shown in 3c Fig. 9. Unlike the spectra of P‡c 700 A1 in PS I in the top part of the ¢gure, the spectral width of the greensulphur bacteria and Heliobacteria spectra does not increase at higher microwave frequency. Thus, the spectra are assigned to the nearly isotropic contribution from P‡c to P‡c F3 X . This assignment is consistent with optical measurements [78,79], which indicate that the electron transfer to the iron^sulphur clusters is faster than in PS I. Because green-sulphur bacteria are strictly anaerobic, they are di¤cult to manipulate biochemically

and the photoactivity of isolated RCs must be tested. An important result of the measurements in [85] is that signi¢cant photoactivity is found for samples from which no quinone could be extracted using organic solvents suggesting that a quinone is not necessary for forward electron transfer. The green-sulphur bacteria and Heliobacteria spectra in Fig. 9 show a net absorptive spin-polarization indicative of singlet^triplet mixing in the precursor [86]. Thurnauer and co-workers have studied this e¡ect extensively [40,86^88] in purple bacterial RCs and analysed its e¡ect on the spin-polarized spectra of 3c P‡c 865 QA . Kandrashkin et al. [37] derived an analytical solution for the polarization patterns of sequential radical pairs in the limiting cases of short-lived and long-lived precursors. The spectra for both PS I and Heliobacteria were calculated using known values from EPR and X-ray data to ¢x all of the parameters and excellent agreement with the experimental results was obtained. In the case of the Heliobacteria, direct transfer from A0 to FX with a lifetime of d = 600^ 700 ps was assumed. Thus, the transient EPR results give no evidence for the involvement of a quinone acceptor. However, this must be reconciled with the fact that a semiquinone radical can be photo-accumulated. One possibility is that the distance between the quinone and FX is very short so that the reduced form of the quinone is very short-lived and thus, has little or no e¡ect on the spin polarization and optical transients. 7. Conclusions and outlook

Fig. 9. Comparison of the X- and K-band spin-polarization patterns of PS I (top) and green-sulphur bacteria and Heliobacteria (middle and bottom).

The structural and functional data obtained from transient EPR discussed above provide the groundwork on which a detailed understanding of the protein-cofactor interactions in the two classes of RCs can be built. With the improved structural detail provided by the latest X-ray data, it is now possible to rationalize some of the observed di¡erences in the properties of the quinone acceptors in bRCs and PS I on a molecular level. However, the consequences of these di¡erences for the electron transfer are not yet clear. The ultimate goal is to understand why these di¡erences in the various photosystems have evolved and how the protein-cofactor interactions control the electron transfer. At present there are two key ques-

BBABIO 45078 12-10-01

A. van der Est / Biochimica et Biophysica Acta 1507 (2001) 212^225

tions that remain open. (i) Which of the two quinones found in type I reaction centres are actively involved in electron transfer? (ii) What properties of the quinone determine its function? Experiments in which the properties of the quinones and their binding sites are varied, particularly using molecular biological techniques, are beginning to provide insight in this direction [67^69]. Site selective mutation experiments such as those in [31] can be expected to provide the clearest answer to the question of bidirectional versus unidirectional electron transfer, since mutations in the active branch should have a much larger e¡ect on the function that those in the inactive branch. If the mutations are made in the quinone-binding region, transient EPR is ideally suited for screening for their e¡ect on the structure and function because structural information for the functional state P‡ Q3 is obtained under physiological conditions. The obvious ¢rst candidates for such studies are the tryptophans TrpA697 and TrpB677 which are clearly involved in binding A1 . Another novel and less invasive approach is to provide an observer electron spin on one side of the reaction centre as discussed in [89]. This third spin can have a dramatic e¡ect on the spin polarization pattern at distances which are expected to leave the function of the system unperturbed. Spin-labelling techniques combined with mutagenesis should allow selective placement of observer spins at a variety of positions in the PS I reaction centre and make a de¢nitive determination of activity of the two quinones possible. Acknowledgements I wish to express my thanks to Dr Motoko AsanoSomeda and the Tokyo Institute of Technology whose great hospitality I enjoyed while preparing the manuscript. I also want to thank the many collaborators and students who contributed to the data presented in this review. In particular I extend my gratitude to Dietmar Stehlik and John Golbeck for many years of support and extremely fruitful collaboration. This work was supported by the Natural Sciences and Engineering Research Council, the Canada Foundation for Innovation and the Ontario Innovation Trust.

223

References [1] M. Bixon, J. Fajer, G. Feher, J.H. Freed, A.J. Ho¡, H. Levanon, K. Mo«bius, R. Nechustai, J.R. Norris, A. Scherz, J.L. Sessler, D. Stehlik, Israel J. Chem. 32 (1992) 369^ 518. [2] S.W. Snyder, M.C. Thurnauer, in: J.R. Norris, J. Deisenhofer (Eds.), The Photosynthetic Reaction Center, Vol. II, Academic Press, New York, 1993, Chapter 11. [3] A. Angerhofer, R. Bittl, Photochem. Photobiol. 63 (1996) 11^38. [4] D. Stehlik, A. van der Est, A. Kamlowski, Mol. Phys. Rep. 13 (1996) 21^36. [5] D. Stehlik, A. van der Est, A. Kamlowski, Ber. Bunsenges. Phys. Chem. 100 (1996) 2028^2035. [6] H. Levanon, K. Mo«bius, Annu. Rev. Biophys. Biomol. Struct. 26 (1997) 495^540. [7] D. Stehlik, K. Mo«bius, Annu. Rev. Phys. Chem. 48 (1997) 745^784. [8] K. Mo«bius, Chem. Soc. Rev. 2 (2000) 129^139. [9] A. Kamlowski, S.G. Zech, P. Fromme, R. Bittl, W. Lubitz, D. Stehlik, J. Phys. Chem. B 102 (1998) 8266^8277. [10] R. Blankenship, A. McGuire, K. Sauer, Proc. Natl. Acad. Sci. USA 72 (1975) 4943^4947. [11] G.C. Dismukes, A. McGuire, R. Blankenship, K. Sauer, Biophys. J. 21 (1978) 239^256. [12] G.C. Dismukes, K. Sauer, Biochim. Biophys. Acta 504 (1978) 431^445. [13] G.C. Dismukes, A. McGuire, R. Friesner, K. Sauer, Rev. Chem. Intermed. 3 (1979) 59^88. [14] R. Furrer, M.C. Thurnauer, FEBS Lett. 153 (1983) 399^403. [15] M.C. Thurnauer, P. Gast, Photobiochem. Photobiophys. 9 (1985) 29^38. [16] C.H. Bock, D. Stehlik, M.C. Thurnauer, Israel J. Chem. 28 (1988) 177^182. [17] H. Manikowski, A.R. MacIntosh, J.R. Bolton, Biochim. Biophys. Acta 765 (1984) 68^73. [18] C.H. Bock, A.J. van der Est, K. Brettel, D. Stehlik, FEBS Lett. 247 (1989) 91^96. [19] P. Se¨tif, P. Mathis, T. Va«nngard, Biochim. Biophys. Acta 767 (1994) 404^414. [20] E. Schlodder, K. Brettel, K. Falkenberg, M. Gergeleit, in: P. Mathis (Ed.), Proceedings of the Xth Intl. Congr.on Photosynthesis, Vol. II, Kluwer Academic Publishers, Dordrecht, 1995, pp. 107^110. [21] K. Brettel, Biochim. Biophys. Acta 1318 (1997) 322^373. [22] M.C. Thurnauer, J.R. Norris, Chem. Phys. Lett. 76 (1980) 557^561. [23] R.R. Rustandi, S.W. Snyder, L.L. Feezel, T.J. Michalski, J.R. Norris, M.C. Thurnauer, J. Biggins, Biochemistry 29 (1990) 8030^8032. [24] K. Brettel, FEBS Lett. 239 (1989) 93^98. [25] P. Mathis, P. Se¨tif, FEBS Lett. 237 (1989) 65^68. [26] P. Se¨tif, K. Brettel, Biochemistry 32 (1993) 7846^7854. [27] B. Lagoutte, P. Se¨tif, J. Duranton, FEBS Lett. 174 (1984) 24^29.

BBABIO 45078 12-10-01

224

A. van der Est / Biochimica et Biophysica Acta 1507 (2001) 212^225

[28] A. van der Est, C. Bock, J. Golbeck, K. Brettel, P. Se¨tif, D. Stehlik, Biochemistry 33 (1994) 11789^11797. [29] F. Yang, G. Shen, W.M. Schluchter, B. Zybailov, A.O. Ganago, I.R. Vassiliev, D.A. Bryant, J.H. Golbeck, J. Phys. Chem. 102 (1998) 8288^8299. [30] P. Joliot, A. Joliot, Biochemistry 38 (1999) 11130^11136. [31] M. Guergova-Kuras, B. Boudreaux, A. Joliot, P. Joliot, K. Redding, Proc. Natl. Acad. Sci. USA 98 (2001) 4437^4442. [32] A. Purton, D.R. Stevens, I.P. Muhiuddin, M.C.W. Evans, S.E. Carter, J. Rigby, P. Heathcote, Biochemistry 40 (2001) 2167^2175. [33] L. Krabben, E. Schlodder, R. Jordan, D. Carbonera, G. Giacometti, H. Lee, A.N. Webber, W. Lubitz, Biochemistry 39 (2000) 13012^13025. [34] P. Moe«nne-Loccoz, P. Heathcote, D.J. Maclachlan, M.C. Berry, I.H. Davis, M.C.W. Evans, Biochemistry 33 (1994) 10037^10042. [35] J. Lu«neberg, P. Fromme, P. Jekow, E. Schlodder, FEBS Lett. 338 (1994) 197^202. [36] A. Quintana, A. Diaz, W. Leibl, H. Bottin, P. Se¨tif, Biochemistry 37 (1998) 3429^3439. [37] Yu.E. Kandrashkin, K.M. Salikhov, A. van der Est, D. Stehlik, Appl. Magn. Res. 15 (1998) 417^447. [38] G.L. Closs, M.D.E. Forbes, J.R. Norris, J. Phys. Chem. 91 (1987) 3592^3599. [39] C.D. Buckley, D.A. Hunter, P.J. Hore, K.A. MacLaughlan, Chem. Phys. Lett. 135 (1987) 307^312. [40] J.R. Norris, A.L. Morris, M.C. Thurnauer, J. Tang, J. Chem. Phys. 92 (1990) 4239^4249. [41] D. Stehlik, C.H. Bock, J. Petersen, J. Phys. Chem. 93 (1989) 1612^1619. [42] G. Fu«chsle, R. Bittl, A. van der Est, W. Lubitz, D. Stehlik, Biochim. Biophys. Acta 1142 (1993) 23^35. [43] A. van der Est, T. Prisner, R. Bittl, P. Fromme, W. Lubitz, K. Mo«bius, D. Stehlik, J. Phys. Chem. B 101 (1997) 1437^ 1443. [44] R. Klette, J. To«rring, M. Plato, B. Bo«nigk, W. Lubitz, K. Mo«bius, J. Phys. Chem. 97 (1993) 2015^2020. [45] S.G. Zech, W. Hofbauer, A. Kamlowski, P. Fromme, D. Stehlik, W. Lubitz, R. Bittl, J. Phys. Chem. B 104 (2000) 9728^9739. [46] A.J. Stone, Proc. R. Soc. A 271 (1963) 424. [47] A.J. Stone, Mol. Phys. 6 (1963) 509. [48] Yu. Kandrashkin, A. van der Est, Spectrochim. Acta A 57 (2001) 1697^1709. [49] R. Bittl, S.G. Zech, J. Phys. Chem. B 101 (1997) 1429^1436. [50] S.A. Dzuba, H. Hara, A. Kawamori, M. Iwaki, S. Itoh, Yu.D. Tsvetkov, Chem. Phys. Lett. 264 (1997) 361^366. [51] F. MacMillan, J. Hanley, L. van der Weerd, M. Knupling, S. Un, A.W. Rutherford, Biochemistry 36 (1997) 9297^9303. [52] Y. Deligiannakis, J. Hanley, A.W. Rutherford, Biochemistry 37 (1998) 3329^3336. [53] G. Kothe, S. Weber, R. Bittl, E. Ohmes, M.C. Thurnauer, J.R. Norris, Chem. Phys. Lett. 186 (1991) 474^480. [54] M. Gierer, A. van der Est, D. Stehlik, Chem. Phys. Lett. 186 (1991) 238^247.

[55] G. Kothe, S. Weber, E. Ohmes, M.C. Thurnauer, J.R. Norris, J. Phys. Chem. 98 (1994) 2706^2712. [56] R. Bittl, A. van der Est, A. Kamlowski, W. Lubitz, D. Stehlik, Chem. Phys. Lett. 226 (1994) 349^358. [57] S. Weber, T. Berthold, E. Ohmes, M.C. Thurnauer, J.R. Norris, G. Kothe, Appl. Magn. Reson. 11 (1996) 461^469. [58] S. Weber, G. Kothe, J.R. Norris, J. Chem. Phys. 106 (1997) 6248^6261. [59] G. Kothe, M. Bechtold, G. Link, E. Ohmes, J.-U. Weidner, Chem. Phys. Lett. 283 (1998) 51^60. [60] R. Bittl, S.G. Zech, P. Fromme, H.T. Witt, W. Lubitz, Biochemistry 36 (1997) 12001^12004. [61] O. Klukas, W.D. Schubert, P. Jordan, N. Krauss, P. Fromme, H.T. Witt, W. Saenger, J. Biol. Chem. 274 (1999) 7361^7367. [62] T. Berthold, M. Bechtold, U. Heinen, G. Link, O. Poluektov, L. Utschig, J. Tang, M.C. Thurnauer, G. Kothe, J. Phys. Chem. B 103 (1999) 10733^10736. [63] I. Sieckmann, A. van der Est, H. Bottin, P. Se¨tif, D. Stehlik, FEBS Lett. 284 (1991) 98^102. [64] S.W. Snyder, R.R. Rustandi, J. Biggins, J.R. Norris, M.C. Thurnauer, Proc. Natl. Acad. Sci. USA 88 (1991) 9895^9896. [65] R.R. Rustandi, S.W. Snyder, J. Biggins, J.R. Norris, M.C. Thurnauer, Biochim. Biophys. Acta 1101 (1992) 311^320. [66] A. van der Est, I. Sieckmann, W. Lubitz, D. Stehlik, Chem. Phys. 194 (1995) 349^359. [67] T.W. Johnson, G.Z. Shen, B. Zybailov, D. Kolling, R. Reategui, S. Beauparlant, I.R. Vassiliev, D.A. Bryant, A.D. Jones, J.H. Golbeck, P.R. Chitnis, J. Biol. Chem. 275 (2000) 8523^8530. [68] B. Zybailov, A. van der Est, S.G. Zech, C. Teutlo¡, W. Johnson, G.Z. Shen, R. Bittl, D. Stehlik, P.R. Chitnis, J.R. Golbeck, J. Biol. Chem. 275 (2000) 8531^8539. [69] A.Yu. Semenov, I.R. Vassilev, A. van der Est, M. Mamedov, B. Zybailov, G. Shen, D. Stehlik, B. Diner, P. Chitnis, J. Golbeck, J. Biol. Chem. 275 (2000) 23429^23438. [70] W.-D. Schubert, O. Klukas, N. KrauM, W. Saenger, P. Fromme, H.T. Witt, J. Mol. Biol. 272 (1997) 741^769. [71] A. Kamlowski, B. Altenberg-Greulich, A. van der Est, S. Zech, R. Bittl, P. Fromme, W. Lubitz, D. Stehlik, J. Phys. Chem. B 102 (1998) 8278^8287. [72] J. Hanley, Y. Deligiannakis, F. MacMillan, H. Bottin, A.W. Rutherford, Biochemistry 36 (1997) 11543^11549. [73] F. MacMillan, Ph.D. Thesis, Freie Universita«t Berlin, 1993. [74] P. Heathcote, S.E.J. Rigby, M.C.W. Evans, in: P. Mathis (Ed.), Photosynthesis: From Light to Biosphere, Proceedings of the Xth Intl. Congr. on Photosynthesis, Vol. II, Kluwer Academic Publishers, Dordrecht, 1995, pp. 163^166. [75] S.E.J. Rigby, M.C.W. Evans, P. Heathcote, Biochemistry 35 (1996) 6651^6656. [76] C. Teutlo¡, Master's Thesis, Technische Universita«t Berlin, 1997. [77] P. Jordan, P. Fromme, H.T. Witt, O. Klukas, W. Saenger, N. KrauM, Nature 411 (2001) 909^917. [78] S. Lin, H.-C. Chiou, R.E. Blankenship, Biochemistry 34 (1995) 12761^12767.

BBABIO 45078 12-10-01

A. van der Est / Biochimica et Biophysica Acta 1507 (2001) 212^225 [79] K. Brettel, W. Leibl, U. Liebl, Biochim. Biophys. Acta 1363 (1998) 175^181. [80] W. Nitschke, U. Feiler, W. Lockau, G. Hauska, FEBS Lett. 218 (1987) 283^286. [81] I.P. Muhiuddin, S.E.J. Rigby, M.C.W. Evans, P. Heathcote, in: P. Mathis (Ed.), Photosynthesis: From Light to Biosphere, Vol. II, Kluwer Academic Publishers, Dordrecht, 1995, pp. 159^162. [82] B. Kj×r, N.-U. Frigaard, F. Yang, B. Zybailov, M. Miller, J.H. Golbeck, H.V. Scheller, Biochemistry 32 (1998) 3237^ 3242. [83] I.P. Muhiuddin, S.E.J. Rigby, M.C.W. Evans, J. Amesz, P. Heathcote, Biochemistry 38 (1999) 7159^7167.

225

[84] P. Heathcote, J.T. Warden, FEBS Lett. 140 (1982) 277^ 281. [85] A. van der Est, C. Hager-Braun, W. Leibl, G. Hauska, D. Stehlik, Biochim. Biophys. Acta 1409 (1998) 87^98. [86] A.L. Morris, S.W. Snyder, Y. Zhang, J. Tang, M.C. Thurnauer, P.L. Dutton, D.E. Robertson, M.R. Gunner, J. Phys. Chem. 99 (1995) 3854^3866. [87] L.M. Utschig, S.R. Green¢eld, J. Tang, P.D. Laible, M.C. Thurnauer, Biochemistry 36 (1997) 8548^8558. [88] J. Tang, L.M. Utschig, O. Poluetkov, M.C. Thurnauer, J. Phys. Chem. B 103 (1999) 5145^5150. [89] K.M. Salikhov, A.J. van der Est, D. Stehlik, Appl. Magn. Res. 16 (1999) 101^134.

BBABIO 45078 12-10-01