Radical pair state in photosystem II

Proc. Natl Acad. Sci. USA Vol. 79, pp. 7283-7287, December 1982 Biophysics

Radical pair state in photosystem II (photosynthesis/electron paramagnetic resonance/electron transfer/pheophytin/donors)

A. W. RUTHERFORD*t AND M. C. THURNAUERt *Department of Physiology and Biophysics, University of Illinois, Urbana, Illinois 61801; and *Chemistry Division, Argonne National Laboratory,

Argonne, Illinois 60439

Communicated by Gerhard L. Closs, July 19, 1982

ABSTRACT A stable light-induced EPR signal is reported in photosystem IH particles and in chloroplasts at 5 K. Characteristic spectral features indicate that the signal arises from dipole-dipole interactions of a radical pair triplet state. From its dependence on potential, its relationship to the spin-polarized triplet state, and the redox state of the pheophytin acceptor (Ph) and because it is present in Tris-washed chloroplasts but not in untreated chloroplasts, we conclude that the signal is formed when the reaction center is in the state D+P60Ph- (P680 is the primary chlorophyll donor and D+ is an oxidized donor to P680). The low-temperature photochemical sequence is thought to occur as follows. (i) Donation from D to the Pmo'Ph- state occurs at liquid helium temperature in low quantum yield; this reaction is reversible at temperatures above 5 K. (ii) In normal chloroplasts, donation occurs to the D+P6WPh- state, but this does not occur in Tris-washed chloroplasts or in the photosystem H particles at 77 K or lower. (iii) 1llumination, at 200 K, ofphotosystem particles or Tris-washed chloroplasts results in donation to the D+P680Ph- state from another donor. From experiments in the absence of redox mediators and the temperature dependence of the splitting of the signal, it is suggested that the state D+P680Ph- itself may be the origin of the radical pair triplet signal. The signal has been simulated by assuming the presence of at least two distinct radical pairs that differ slightly in the distance separating the radicals of the pairs. The distance between the radicals of the pair is calculated to be 6-7

MATERIALS AND METHODS PSII particles were prepared from pea chloroplasts as described (9). Some biophysical properties of these particles have been reported (7, 10). Untreated and Tris-washed chloroplasts were used fresh or after being stored at 77 K in SHN buffer (0.4 M sucrose/20 mM Hepes, pH 8.0/15 mM NaCl). Washing with Tris was done by using the method described in ref. 11 (Tris at pH 8.8, 1 hr). Oxygen evolution was measured with a Clarktype oxygen electrode to determine the effectiveness of the treatment. Oxidation-reduction potentiometry and EPR sample preparation were exactly as described (12). EPR spectra were obtained by using a Bruker ER-200D X band spectrometer and an Oxford Instruments cryostat and variable temperature system. Illumination at 200 K and illumination in the EPR cavity were done as described (7). EPR spectra were simulated on a Sigma V computer (Chemistry Division, Argonne National Laboratory) with a modified version of a standard program (13) to calculate the EPR spectrum of triplet molecules randomly oriented in a rigid medium. RESULTS AND DISCUSSION Description and interpretation of the light-induced signal When PSII particles poised at low redox potential (-500 to -620 mV) were illuminated at liquid helium temperature, an EPR signal was induced (Fig. 1). At 5 K the photoinduced signal did not decay when the light was switched off. The shape ofthe signal was better resolved (Fig. ic) when the dark spectrum (Fig. la) was subtracted from that recorded after illumination (Fig. lb). The signal exhibits two intense peaks split by approximately 10 mT (100 gauss). Each of these intense peaks has a marked shoulder. The splitting between the two shoulders is approximately twice that ofthe intense peaks. This shape of the signal and magnitude of the splitting suggest that it may be due to magnetic dipole-dipole interactions between two radicals of a pair (Am = + 1 transitions) (14). § The signal shape could be simulated (Fig. ld) by assuming the presence of a minimum of two distinct radical pairs, with zero field splitting parameters of |DI = 0.0120 cm-1, |E = 0.0 cm-' and |DI = 0.00944 cm-1, |El = 0.0 cm-1. Application of the point dipole approximation (15) to these values gives 6-7 A as the average distance between the two unpaired electrons on the two radicals of the pairs.

A.

The current view is that the primary reactions in photosystem II (PSII) are similar to those occurring in purple photosynthetic bacteria (for a review, see ref. 1). Upon photoexcitation, a charge separation takes place between a special chlorophyll donor, P680 (2), and an intermediary acceptor, probably pheophytin (Ph) (3). The electron on Ph- is rapidly transferred to a quinone acceptor, Q (4), which is probably associated with a ferrous iron atom (5, 6). If Q is reduced prior to illumination, the light-induced radical pair P6w'Ph- recombines. This recombination is believed to give rise to characteristic changes of chlorophyll fluorescence yield observed at room temperature (3) and to a spin-polarized triplet state of P680 observed by EPR at liquid helium temperature (7). The extra oxidizing power generated on the donor side of PSII (1.1 V) might be expected to result in a situation different from that occurring in bacteria (0.450 V). Indeed, kinetic optical and EPR measurements of secondary donation events in PSII have resulted in a complex picture and analogy to the bacterial system does not help in understanding it (for a review of the donor side of PSII, see ref. 8). In this paper we report an EPR signal, photoinduced at cryogenic temperatures in PSII particles, that can be used to study donor reactions.

Abbreviations: PSII, photosystem II; Ph, pheophytin; Q, quinone acceptor. t Present address: Service de Biophysique, D6partment de Biologie, Centre d'Etudes Nucleares de Saclay, 91191 Gif-sur-Yvette, Cedex, France. § In general, a radical pair will have a singlet (S = 0) and triplet (S = 1) state. The singlet-triplet separation is given by the isotropic exchange interaction. Because the triplet state of the pair is the EPR active state we refer throughout the paper to the radical pair triplet

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact.

state. 7283

7284

Biophysics: Rutherford and Thurnauer

I

I

I

I

Proc. Natl. Acad. Sci. USA 79 (1982)

I

320 330 340 350 320 330 340 350 Field, mT

FIG. 1. EPR spectra of reduced PSH1 particles at 5 K. The sample was poised at -603 mV in 100 mM KCl/100 mM K glycinate, pH 11, containing, each at 100 uM, benzyl viologen, methyl viologen, triquat (1,1'-trimethylene-2,2'-dipyridinium dibromide), and tetraquat (1,1'tetramethylene-2,2'-dipyridinium dibromide). The potential was decreased by small additions of 20% (wt/vol) sodium dithionite in 100 mM K glycinate at pH 11. (a) Sample frozen in the dark. (b) Same sample as a, recorded in the dark after 10 min of illumination in the EPR cavity. (c) Subtraction of a from b. EPR conditions were: microwave power, 18 dB down from 200 mW; modulation amplitude, 1.0 mT. (d) Computer simulation of the spectrum. The best fit was obtained by adding two triplet spectra [37% of the intensity with D = 0.0120 cm' , E = 0.0 cm-' and 63% of the intensity with D = 0.00944 cm-', E = 0.0 cm11. For both cases the calculated stick spectra for a given orientation were broadened by the derivative of a gaussian lineshape with a linewidth of 1.2 mT.

A signal at g 4.04 was photoinduced at the same time as the signal at g 2.00 (see Fig. 5). This signal occurs at the field position expected for the half-field transition ("Am = 2"). The observation of a half-field transition is a strong indication that the signal is due to a spin-coupled system (14, 16). =

=

Relationship of the radical pair triplet to photosynthetic electron transport Temperature Dependence. The radical pair triplet signal exhibited a distinct temperature dependence (Fig. 2). The signal was largest after illumination at the lowest temperatures available (approximately 4.2 K), and it was not easily saturated by microwave power (31 mW gave optimal signal-to-noise ratio on this instrument). As the temperature was raised from 4.2 K, the signal amplitude decreased, but a signal was still observable at temperatures higher than 80 K. When the temperature was lowered again to 5 K, the signal intensity remained reduced (Fig. 2, d compared to a). Illumination again at 5 K restored the signal to its full extent. Thus, the decrease of the signal amplitude as the temperature was raised was due to reactive loss of the radical pair triplet state rather than to a temperature dependence of the signal. This is presumably due to a back reaction occurring at temperatures above 5 K and leading to loss of the light-induced state. The splitting of the radical pair triplet signal also changed with temperature. The splitting of the signal that appeared to be stable at 80 K (Fig. 2 c and d) was narrower (-8.5 mT between intense peaks) than the signal that was photoinduced at 5 K (Fig. 2a; splitting =10 mT). The signal with the smaller splitting and stable at 80 K (Fig. 2d) was more easily saturated

C

d

/

320

rerun at 5 K

330 340 Field, mT

350

FIG. 2. Temperature dependence of the photoinduced EPR signal. A sample was poised at approximately -600 mV under conditions as in Fig. 1. The sample was frozen in the dark and illuminated for approximately 10 min at 5 K. All spectra are of the same sample and were recorded in the dark without further illumination. The temperature of the sample was varied as follows. After illumination at 5 K, a spectrum was recorded (a); the temperature was raised to 40 K and then to 80 K and spectra were recorded at each temperature (b and c, respectively); the temperature was then lowered to 5 K and a further spectrum was recorded (d). EPR conditions were as in Fig. 1 except that power was 8 dB down from 200 mW.

with microwave power than was the signal with the larger splitting (Fig. 2a). These observations fit well with the necessity to simulate the lowest temperature signal by using more than one set of zero field splitting parameters (Fig. ld). In fact, the smaller values= 0.00944 cm-1, IE = 0.0-give a reasonable fit to the spectrum in Fig. 2d (low-temperature spectrum after warming irradiated sample). It is of note that the signal with the smaller zero field splitting (greater average distance between the radicals of the pair) was more stable than the signal exhibiting the larger zero field splitting (smaller average distance between the radicals). This may be taken as an indication that both of the component radicals ofthe pair are products of the photoreaction rather than being a single photoinduced radical interacting with a radical already present in the dark. Potential Dependence. The photoinduced radical pair triplet signal was only observable at potentials lower than -500 mV (Fig. 3). Under these conditions the primary quinone acceptor was already reduced (17). Illumination at low temperature results in the formation of the Pm+ Ph- radical pair which recombines to form the spin-polarized triplet state of P680 before returning to the ground state (7). It is possible that donation to the Pm+ Ph- radical pair occurs at helium temperature, trap-

|DI

Biophysics: Rutherford and Thurnauer


mV

7285

4 min

-450 -

a

-550'

-600

-650 i c

320

330

340 Field, mT

On

350

FIG. 3. Potential dependence of the photoinduced EPR signal. The redox potential of each sample was poised at the potential indicated by the methods described in Fig. 1. Each sample was illuminated for 10 min at 5 K in the EPR cavity. EPR conditions were as in Fig. 1.

ping Ph in its reduced state. The radical pair triplet may reflect the state D+P68OPh-. Because the results imply the presence of at least two distinct radical pairs, D+ could be either two different donor species or the same donor with two or more different orientations in the system. At -650 mV the extent of light-induced signal was considerably diminished. This corresponds to the chemical reduction of Ph which occurs with an Em of approximately -610 mV (12, 18). The redox change occurring between -500 and -600 mV is more difficult to assign because no known PSII component undergoes reduction in this range. In bacterial reaction centers, Q undergoes a further reduction step in this potential range to form the fully reduced quinol (19, 20). An analogous redox change in PSII could be responsible for the potential dependence of the photoinducibility of the radical pair triplet signal. When Ph- is photoinduced in the presence of singly reduced Q, the so-called split I- signal is observed both in bacteria (21) and in PSII (5-7, 10, 22). This signal has been attributed to interaction between the semiquinone iron complex and Ph-. These interactions could either prevent formation ofthe radical pair triplet state or, more likely, prevent observation of such a state. The possibility that one ofthe radicals ofthe pair is a mediator radical with an Em in this redox range has not been ruled out and is discussed below. Kinetics of Formation. If the radical pair triplet signal represents the formation of the D+P680Ph- state, then a simultaneous loss in the extent of the spin-polarized triplet signal would be predicted. This was found to be the case (Fig. 4). The radical pair triplet state (trace a) was formed slowly and irreversibly at 5 K whereas the spin-polarized triplet (trace b) was formed immediately (within our time resolution) upon illumination and appeared to decay as soon as the light was switched off. The

00400 Off On Off

FIG. 4. Kinetics of formation of the radical pair triplet signal compared with those of the formation of the spin-polarized triplet signal. Spectra were recorded at 5 K in similar samples of PSI1 particles poised at approximately -600 mV and frozen in the dark. (a) Trace obtained by setting the magnetic field at the high field trough of the radical pair triplet signal (343 mT). EPR conditions were as in Fig. 1. (b) Trace obtained by setting the magnetic field at the high field maximum of the spin-polarized triplet signal (370 mT). EPR conditions were: microwave power, 36 dB down from 200 mW (0.05 mW); modulation amplitude, 2.0 mT. (c) Trace recorded off resonance (375 mT) under conditions identical to those described for b. A 1,000-watt projector lamp, providing illumination of the samples in the cavity, was switched on or off as marked. Response time of the instrument was 2 sec in all traces.

intensity of the spin-polarized triplet signal was larger on the first illumination than on subsequent illuminations, and during illuminations a decrease ofamplitude occurred with a decay rate similar to that of the formation of the radical pair triplet. Trace c was recorded at a field set off resonance under the same EPR conditions used for trace b. The slow rate of formation of the radical pair triplet signal and simultaneous decay of the spinpolarized triplet signal indicate that donation to the P68o0Phradical pair occurs with a low quantum yield relative to the formation of the spin-polarized triplet. The fraction of the spin-polarized triplet signal that did not decay under illumination presumably represents those centers in which the donor is absent or unable to donate to the P680'Phradical pair. At temperatures higher than 5 K the kinetics of formation of the radical pair triplet signal showed an irreversible and a reversible portion. This behavior would be predicted from the temperature-dependence data described above. The decrease in the amplitude of the spin-polarized triplet signal was also observed at redox potentials higher than -500 mV, indicating that donation to the P680'Ph- radical pair occurred outside the potential range where the radical pair triplet was observable. Effect of Illumination at 200 K. It had been demonstrated previously that illumination of reduced PSII particles at 200 K results in the trapping ofthe Ph in its reduced form (10, 22) and consequently in the loss of the ability to photoinduce the spinpolarized triplet (7). When PSII particles poised between -500

Biophysics: Rutherford and Thumauer

7286

and -620 mV were illuminated at 200 K, the radical pair triplet signal was not photoinduced (Fig. 5c). Illumination at 5 K (or 80 K) of sample that had previously been illuminated at 200 K resulted in the formation ofonly a very small radical pair triplet signal (Fig. 5d). When the sample was thawed under oxygenfree nitrogen and refrozen in the dark, the maximum amplitude of the radical pair triplet signal could be regenerated by illumination at 5 K (not shown). Illumination at 200 K had an effect upon the half-field signal identical to that described for the signal around g = 2 (Fig. 5). This phenomenon can be explained by assuming that at 200 K another donor, Dwo, donates to D+ (or directly to P6w') resulting in a state that does not give rise to the radical pair triplet signal. [It is possible that D2w is sodium dithionite because donation from this reductant has been observed when photosynthetic bacteria are illuminated under similar redox conditions (unpublished data). ] Because donation from D" only occurs at 200 K and the state produced does not readily back react Frozen dark

Illuminated at 200 K

a

Proc. Nad Acad. Sci. USA 79 (1982) at this temperature, it is likely that D2w is further from the reaction center than is D. Radical Pair Triplet Signal in Chloroplasts. Because the intensity of the radical pair triplet signal was relatively strong in PSII particles it seemed possible that the same signal might be observable in chloroplasts. However, under conditions of redox poise and temperature identical to those described above (i.e., Fig. 1) no such signal could be observed in chloroplasts. One possible explanation for this was that in chloroplasts the donor side was too "intact" and that at helium temperature the charge on D was being translocated away, resulting in the absence of the D'P6wPh- state and consequently the absence of the rad-

ical pair triplet signal. When the experiment was repeated with

chloroplasts that had been washed in alkaline Tris buffer, a procedure that inhibits some of the donation reactions in PSII (23), large light-inducible radical pair triplet signals could be observed. The radical pair triplet signal in Tris-washed chloroplasts was similar to that in PSII particles with regard to spectral features, temperature dependence, redox potential dependence, and kinetics of formation (spectra will be published elsewhere). Is there a mediator radical involved in the radical pair

dark

triplet state?

To investigate the possible involvement of a mediator radical in the formation of the radical pair triplet state, a series of poising experiments was carried out with Tris-washed chloroplasts in the absence of mediators or in the presence of individual mediators. The maximal extent of the signal could be observed only in the presence of the mediator triquat. In the absence of mediators and after several hours of incubation in excess sodium dithionite, a small signal (approximately 25% of maximal amplitude) could be photoinduced at 5 K (Fig. 6). This signal was formed stably at 5 K but decayed at 80 K. It had a peak position at the same place as the high-field intense peak of the radical pair triplet but the linewidth was narrower than that formed in the presence of triquat. This signal may be the radical pair trip-

b off

310 320 330 340 350 320 330 340 350 360

dark

C

a

dark dark

b

d off

off off

---4

160 165 170 175 160 165 170 175 Field, mT FIG. 5. Effect of illumination at 200 K on the radical pair triplet signal. (Upper) EPR spectra in the g = 2 region (EPR conditions as in Fig. 1). (Lower) Region of the half-field signal (g 4.00) (EPR conditions as in Fig. 1 but microwave power was 8 dB and the gain was increased 10-fold) recorded from a similar sample. Samples were poised at -600 mV as described in Fig. 1 and frozen in the dark (a). The samples were illuminated at 5 K for 10 min and a spectrum was recorded in the dark after illumination (b). The samples were then illuminated at 200 K for 30 min and cooled to liquid nitrogen temperature and then to liquid helium temperature in the dark. Spectra were recorded in the dark at 5 K (c). The samples were then illuminated at 5 K for 10 min and spectra were recorded in the dark after this illumination period =

(d).

320

330

340

350

360

Field, mT FIG. 6. Irreversible light-induced signal in Tris-washed chloroplasts in the absence of mediators. Tris-washed chloroplasts were in-

cubated anaerobically in the presence of sodium dithionite for 8 hr. EPR spectra were recorded at 5 K after freezing in the dark and after illuminating at 5 K for 10 min. EPR conditions were as in Fig. 1.

Biophysics: Rutherford and Thumauer let state, the low-field peak being masked by the large peak at g 2.05 from the photosystem I iron-sulfur centers. The narrow linewidth of the signal may be due to the fraction of centers accessible to dithionite all having the donor in the same position relative to the reaction center. From these results no unequivocal answer emerges concerning the involvement of a mediator in the radical pair triplet state. It is difficult to conceive of an adequate explanation involving triquat as a photoreactant. A role for reduced triquat as a paramagnetic probe forming a radical pair with D+ may be conceivable. The radical pair triplet state in such circumstances would originate from the triquat(red)D+P68oPh- state. This would explain the specificity of triquat and the potential dependence of formation of the radical pair signal. However, the chemical similarities between triquat and the other pyridinium dyes (methyl viologen, benzyl viologen, -and tetraquat) make such a specificity seem unlikely. A role for triquat as a mediator required to reduce a component (Q-/Q--?) that is virtually inaccessible to dithionite alone and that has an Em between -500 and -600 mV also can be visualized. Again this would explain the specificity of triquat since it is the only mediator with an Em in this range. The presence of a small signal which may be the radical pair triplet in the absence of mediators may be taken as evidence that a mediator is not actually involved in formation of the radical pair triplet. The relationship of the distance between the radicals to the stability of the photoinduced change (discussed above) provides circumstantial evidence that the products of the photoreaction themselves form the observed radical pair triplet state. Thus, overall the data appear to favor the formation of the radical pair from D+P68oPh- rather than from triquat(red)D+P6wPh- although the latter cannot be definitely ruled out. =

CONCLUDING REMARKS The photochemical reactions in reduced PSII particles and Triswashed chloroplasts take place at low temperatures as follows: 5-200 K

D20OD'P68oPhSpecies responsible for the radical

I pair triplet state

In untreated chloroplasts illuminated at liquid helium temperature, donation probably occurs to D+ from either the donor designated D200 or another donor. This donation step is sensitive to washing with Tris and the detergent treatment used to prepare particles. This reaction scheme is in agreement with


7287

the suggestion by Nugent et al that photoreduction of PS11 Ph at 4 K can explain an increase in a free radical signal in chloroplasts (24) and their observation of a split signal in PSII particles (6). The data here indicate that at least two distinct radical pairs with marked differences in stability are formed under the experimental conditions. This observation may be the result of conformational differences (of a single donor) present at room temperature which are "frozen in" under the conditions of the experiment. However, other explanations involving more than one donor or heterogenicity of centers cannot be ruled out. Finally, if the radical pair triplet signal is indeed due to D+P 'Ph-, then the signal provides a measurement of the average distance between the two unpaired electrons on the component radicals. The distance of 6-7 A places some tight restrictions on the structure of the reaction center. A.W.R thanks A. R. Crofts for useful discussion and continued support. We also thank J. E. Mullet for preparation of PSI1 particles used in the early part of this work and also for useful discussions, L. A. Oimoen for expert technical assistance, and M. C. W. Evans for giving us the low-potential mediator dyes. M.C.T. acknowledges many discussions with J. R. Norris concerning radical pair interactions. This work was performed under the auspices of the Office of Basic Energy Sciences, Division of Chemical Science, U.S. Department of Energy under Contract W-31-109-ENG-38. The work in this paper that was done at the University of Illinois was supported by National Science Foundation Grant PCM 78-16574. 1. Sauer, K. (1979) Annu. Rev. Phys. Chem. 30, 155-178. 2. Doring, A., Renger, G., Vater, J. & Witt, H. T. (1969) Biochim. Biophys. Acta 347, 439-442. 3. Klimov, V. V., Klevanick, A. V., Shuvalov, V. A. & Krasnovskii, A. A. (1977) FEBS Lett. 82, 183-186. 4. Van Gorkom, H. J. (1974) Biochim. Biophys. Acta 347, 439-442. 5. Klimov, V. V., Dolan, E., Shaw, E. R. & Ke, B. (1980) Proc. Natl. Acad. Sci. USA 77, 7227-7231. 6. Nugent, J. H. A., Diner, B. A. & Evans, M. C. W. (1981) FEBS Lett. 124, 241-244. 7. Rutherford, A. W., Paterson, D. R. & Mullet, J. E. (1981) Biochim. Biophys. Acta 635, 205-214. 8. Bouges Boquet, B. (1980) Biochim. Biophys. Acta 594, 85-103. 9. Mullet, J. E. & Arntzen, C. J. (1981) Biochim. Biophys. Acta 635, 236-248. 10. Rutherford, A. W., Mullet, J. E., Paterson, D. R., Robinson, H. H., Arntzen, C. J. & Crofts, A. R. (1981) Proc. Int. Congr. Photosynth. Res. 3, 919-928. 11. Inoue, Y., Yamashita, T., Kobayashi, Y. & Shibata, K. (1977) FEBS Lett. 82, 303-306. 12. Rutherford, A. W., Mullet, J. E. & Crofts, A. R. (1981) FEBS Lett. 123, 235-237. 13. Norris, J. R. (1968) Dissertation (Washington Univ., St. Louis, MO). 14. Lebedev, Y. S. (1969) Radiat. Eff. 1, 213-227. 15. Hirota, N. .& Weissman, S. I. (1969)J. Am. Chem. Soc. 86, 25382545. 16. Wertz, J. E. & Bolton, J. R. (1972) Electron Spin Resonance Theory and Practical Applications (McGraw-Hill, New York), pp. 223-256. 17. Knaff, D. B. (1975) FEBS Lett. 60, 331-335. 18. Klimov, V. V., Allakhverdiev, S. T., Demeter, S. & Krasnovskii, A. A. (1980) DokL Akad. Nauk. 249, 227-230. 19. Rutherford, A. W., Heathcote, P. & Evans, M. C. W. (1979) Biochem. J. 182, 515-523. 20. Rutherford, A. W. & Evans, M. C. W. (1980) FEBS Lett. 110, 257-261. 21. Tiede, D. M., Prince, R. C., Reed, G. H. & Dutton, P. L. (1976) FEBS Lett. 65, 301-304. 22. Klimov, V. V., Dolan, E. & Ke, B. (1980) FEBS Lett. 112, 97100. 23. Yamashita, T. & Butler, W. (1969) Plant Physiol. 44, 1342-1346. 24. Nugent, J. H. A., Moller, B. L. & Evans, M. C. W. (1980) FEBS Lett. 121, 355-357.