PhotosynthesisResearch 48: 239-246, 1996. (~) 1996 KluwerAcademicPublishers. Printedin the Netherlands. Regular paper

P o l a r i z e d site-selective fluorescence s p e c t r o s c o p y o f the l o n g - w a v e l e n g t h emitting c h l o r o p h y l l s in isolated P h o t o s y s t e m I particles o f Synechococcus

elongatus

Lars-Olof P~lsson 1, Jan E Dekker 1, Eberhard Schloddera, Ren6 Monshouwer I & Rienk van Grondelle 1 l Department of Physics and Astronomy, Vrije Universiteit, De Boelelaan 1081, 1081 HV Amsterdam, The Netherlands; 2Max-Volmer lnstitut far Biophysikalische und Physikalisehe Chemie, Technische Universittit Berlin, Strasse des 17. Juni 135, 10623 Berlin, Germany Received 13 November1995; acceptedin revised form 14 February 1996

Key words: anisotropy, fluorescence, long-wavelength chlorophylls, low temperature spectroscopy, Photosystem I, Synechococcus elongatus

Abstract

Isolated trimeric Photosystem I complexes of the cyanobacterium Synechococcus elongatus have been studied with absorption spectroscopy and site-selective polarized fluorescence spectroscopy at cryogenic temperatures. The 4 K absorption spectrum exhibits a clear and distinct peak at 710 nm and shoulders near 720, 698 and 692 nm apart from the strong absorption profile located at 680 nm. Deconvoluting the 4 K absorption spectrum with Gaussian components revealed that Synechococcus elongatus contains two types of long-wavelength pigments peaking at 708 nm and 719 nm, which we denoted C-708 and C-719, respectively. An estimate of the oscillator strengths revealed that Synechococcus elongatus contains about 4-5 C-708 pigments and 5-6 C-719 pigments. At 4 K and for excitation wavelengths shorter than 712 nm, the emission maximum appeared at 731 nm. For excitation wavelengths longer than 712 nm, the emission maximum shifted to the red, and for excitation in the far red edge of the absorption spectrum the emission maximum was observed 10-11 nm to the red with respect to the excitation wavelength, which indicates that the Stokes shift of C-719 is 10-11 nm. The fluorescence anisotropy, as calculated in the emission maximum, reached a maximal anisotropy of r = 0.35 for excitation in the far red edge of the absorption spectrum (at and above 730 nm), and showed a complicated behavior for excitation at shorter wavelengths. The results suggest efficient energy transfer routes between C-708 and C-719 pigments and also among the C-719 pigments.

Abbreviations: Chl - chlorophyll; FWHM - full width at half maximum; PS I - Photosystem I Introduction

Most photosynthetic systems contain one or more chromophores in their light-harvesting antennae with an excited state energy level lower than that of the primary electron donor (van Grondelle et al. 1994). These chromophores could have a focusing function to concentrate excitations in the vicinity of the reaction center, but it is also possible that they serve to increase the

absorption cross section and thereby enhance the efficiency of the photosynthetic unit (Trissl et al. 1993). The low-energy traps of the antenna of Photosystem I (PS I) are among the deepest considering the various photosynthetic systems. The spectral bandwidth of these chromophores in P S I of Synechocystis PCC 6803 was found to be large, which is probably due to specific pigment-pigment interactions (Gobets et al. 1994). This results in a relatively large spectral overlap

240 between P700 and the long-wavelength chromophores, which could help to explain the efficient uphill energy transfer at higher (physiological) temperatures. At very low temperatures (e.g. at and below 77 K), the thermal energy is usually not sufficient for the uphill energy transfer from the long-wavelength chlorophylls to P700, due to which the long- wavelength pigments act as traps of excitation energy. These traps will be highly fluorescent (relative to P700) and therefore dominate the steady-state emission spectrum. The long-wavelength chlorophyll a pigments of P S I core complexes of Synechocystis PCC 6803 and of higher plants give rise to emission spectra peaking near 720 nm (see van Grondelle et al. 1994 and references therein). However, trimeric P S I core complexes from Spirulina show low-temperature emission maxima near 760 nm (Shubin et al. 1992), intact P S I complexes from higher plants (i.e. those containing the chlorophyll a/b-binding peripheral antenna LHCI) show maxima near 735 nm, and those of Chlamydomonas near 718 nm (see van Grondelle et al. 1994 and references therein), which indicates that there is quite some variability in the spectral characteristics of the long-wavelength chromophores of P S I of the different organisms. Low-energy traps have been included in compartmental and lattice models describing the energy migration within the PS I unit (Werst et al. 1992; Wittmershaus et al. 1992; Turconi et al. 1993; Trinkunas and Holzwarth 1994; Laible et al. 1994; Van Grondelle et ai. 1994; Valkunas et al. 1995). The models were largely based on kinetic data obtained with time-resolved spectroscopy, picosecond fluorescence and absorption experiments. However, these models were to a large extent built without much knowledge of the number of long-wavelength chromophores, their average energy levels and their inhomogenous and homogenous linewidths. In a previous study the properties of the longwavelength emission of thylakoid membrane and PS I core preparations of Synechocystis PCC 6803 were investigated (Gobets et al. 1994). This study showed that there is basically only one long-wavelength pigment within this organism which was designated C708 and which shows the properties of a chlorophyll a dimer. At 4 K it carries a phonon side-band with a FWHM of about 170 cm -1 and shows an inhomogenous distribution of about 215 cm - l (FWHM). In the present study we continue to investigate the long-wavelength emission in P S I and discuss the properties of the long-wavelength emission of P S I core

complexes of the thermophilic cyanobacterium Synechococcus elongatus by using absorption and polarized site-selective fluorescence spectroscopy at cryogenic temperatures. More detailed knowledge about the spectroscopic properties of the long-wavelength pigments of this organism could also be important in view of the detailed structural information on P S I of this organism (Krauss et al. 1993). We conclude that the P S I complexes of Synechococcusand Synechocystis show considerable differences in the amounts and energies of the long-wavelength chlorophylls. Some of the results have been presented at the Xth International Congress on Photosynthesis (P~lsson et al. 1995).

Materials and methods

Trimeric complexes (--~110 Chl/P700) were isolated from the thermophilic cyanobacterium Synechococcus elongatus as described in Witt et al. (1987), except that P S I was extracted with 0.6% (w/w) n-dodecyl-/3,Dmaltoside and purified by the use of a Mono Q column (Pharmacia). For elution a MgSO4 gradient (25-200 mM) in 20 mM Mes pH 6.4 and 0.02% n-dodecylfl,D-maltoside was applied. The absorption spectrum was recorded on a Cary 219 spectrophotometer with a spectral bandwidth of 0.5 nm and using an Oxford flow cryostat (CF 1204). The background was subtracted from the absorption spectra. The background spectrum was recorded at 4 K for a cuvette with buffer/glycerol solution only. Spectral decomposition using Gaussian components was performed using a home written program. The quality of the fit was judged by the value of the standard deviation. The site-selective fluorescence experiments were performed using a CW Ti:Sapphire laser (Coherent 890) pumped by a CW Argon-ion laser (Coherent Innova 300). The bandwidth of the laser light was typically around 1 c m - I (FWHM) and the excitation intensity corresponded to ,-~ 1 mW/cm 2. A Glan Thompson polarizer was used to ensure vertical polarization of the excitation light. In some fluorescence experiments, a broadband 150 W tungsten lamp was used as the excitation source. In the latter case, the excitation wavelength was selected by a 590 nm interference filter with 10 nm spectral bandwidth (FWHM). The detection system consisted of an imaging CCD (charge coupled device) camera (Chromex). The fluorescence spectra were recorded using slitwidths of 100 #m which corresponds to wavelength resolution of ,.~ 0.1 nm. The detection wavelength of the CCD camera was calibrat-

241 ed using a He-Ne laser. The anisotropy was measured using a Rochon polarizer which separated the vertical and horizontal polarizations of the emission in space. The images were then focused to different parts of the CCD camera. The anisotropy was calculated according to the standard relation,

i

e

i

i

~t

~ - G.I± where G is the correction factor which was determined in the region of the fluorescence maximum of the PS I particles. This factor was found to be G = 1.05 -40.01 and essentially independent of wavelength. The anisotropy was determined after setting the fluorescence intensity at 670 nm to zero. In the fluorescence experiments an Utreks liquid Helium cryostat was used. The sample material was diluted in 20 mM Mes pH 6.5, 10 m M CaC12, 10 mM MgCl=, 0.05% n-dodecyl-/~,D-maltoside and 70% w/v glycerol and confined in an acrylic cuvette. The OD in the absorption maximum at 680 nm was around --~ 1.0. This high OD causes no problem in the site-selective fluorescence experiments since the emission spectrum is well separated from the absorption spectrum.

eso

~o

e~o do Wavelength (]am)

7~o

~so

Figure 1. The absorption spectrum of trimeric PSI particles of Synechococcus at 4 K (solid line), 100 K (dashed line) and 175 K (dotted line). o~.

i

i

i

I

Synechococcus o~-

~.~-

T=4K .~-708

~Results Figure 1 (solid line) displays the 4 K absorption spectrum of trimeric P S I particles of Synechococcus in the region 650 to 750 nm. Below 690 nm, the spectrum shows peaks at 673 nm and 679 nm and a clear shoulder near 688 nm, and is very similar to the 4 K absorption spectrum ofSynechocystis PCC 6803 (Gobets et al. 1994). At longer wavelengths, however, the 4 K spectrum of Synechococcus shows more intensity than that of Synechocystis with a pronounced peak at 710 nm and clear shoulders near 720 and 698 nm. Absorption spectra at several higher temperatures up to 250 K were also recorded (see the dashed and dotted line in Figure 1 for the spectra at 100 and 175 K, respectively). These reveal that below 150 K a distinct peak is observable at about 710 nm which gradually transforms into the smooth red wing of the major 680 nm absorption profile at T > 150 K. In order to obtain more information on the spectral composition of the red part of the absorption spectrum, we deconvoluted the 4 K absorption spectrum between 690 and 750 nm with Gaussian bands. We are well aware of the limitations of this procedure, since a unique solution is rarely found and since there

°e~0

70:~

d4 7~e Wavelength (nm)

"~e

750

Figure2. The absorption spectrum of Synechococcustrimers at 4 K in the 690-750 nm interval. The result of the decomposition with Gaussian components is shown as well (see Table 1 for details). is no good evidence that the absorption bands have Gaussian shapes. However, as we will show below, the red-most spectral components are characterized by very broad inhomogeneous distributions, which can be described very well by Gaussian band shapes because of the statistical origin of this broadening (V~lker 1989). Between 690 and 750 nm, the spectrum could be fitted well with a minimum of six different spectral components (Figure 2 and Table 1). In this fit, the three red-most spectral components appeared with very similar linewidths and amplitudes in every fit result. The other components, on the other hand, could vary significantly with respect to position, bandwidth and amplitude from fit to fit without substantially reducing

242

Table 1. Results of the spectral decompositionof the 4 K absorption spectrum of trimeric PSI particles from Synechococcuswith Gaussian bands. The fourth column contains amplitudes in arbitrary units Band no.

Position[nm]

Bandwidth[nm] Amp.[a.u.]

1 2 3 4 5 6

680.0 684.9 692.6 697.6 708.3 719.1

8.96 10.0o 5.00 6.90 10.70 18.70

1.05 0.63 0.12 0.15 0.11 0.08

the fit quality. The essential results of the spectral composition as obtained here are markedly different from those obtained in Synechocystis (Gobets et al. 1994). The red-most spectral form in Synechocystis, denoted C-708 (Gobets et al. 1994), corresponds very well to the 708 nm component obtained in this study, both with respect to its position and to its bandwidth. But in Synechococcus there is an additional spectral component located at 719 nm which has a bandwidth of almost 20 nm (FWHM) and which from here on we denote C-719. We also found a component peaking at 698 nm (C-698), which was not observed in Synechocystis. However, a similar absorption band was observed in PS 1-200 (Gobets et al. 1994), where a similar pronounced shoulder in the 4 K absorption spectrum appeared. It is likely that part of this absorption band is caused by P700. In addition to the absorption bands at 698, 708 and at 719 nm we also found an absorption band at 692 nm. This position is very similar to the position of a pronounced absorption band in Synechocystis (Gobets et al. 1994). The fluorescence maximum of Synechococcus trimers at 4 K and with excitation wavelength 696 nm is located at 731 nm (Figure 3). We also found this maximum upon broad band excitation at 590 nm and upon increasing the temperature up to 270 K (Figure 4). Upon excitation at wavelengths shorter than 670 nm we observed a shoulder around 670 nm, which we attribute to a small population of free or partly disconnected chlorophylls, transferring energy with very low efficiency. Figure 4 shows the emission spectrum obtained by non-selective excitation at 590 nm as a function of temperature. The strong band peaking at 731 nm greatly diminishes upon increasing the temperature, but is still observable at 300 K. At that temperature, the spectrum is characterized by two broad bands peaking near 685 and 730 nm. We note that this spectrum differs

somewhat from the spectra reported by Holzwarth et al. (1993) on P S I particles from the same organism, which may be caused by the different detergent conditions in the preparations (Triton X-100 in the preparations analyzed by Holzwarth et al. vs. dodecylmaltoside in the preparations analyzed by us). In Figure 5 the fluorescence quantum yield of PS I of Synechococcus (determined from the integrated fluorescence spectrum over the range 650-800 nm) as a function of temperature is shown. The fluorescence yield was found to be strongly temperature dependent. The yield remains relatively constant in the interval 4 - 5 0 K but decreases when going towards higher temperatures to a value of about 10% (of the 4 K yield) at 300 K. We note that a small part of the emission could be caused by a small population of free or partly disconnected chlorophylls. The emission of such chlorophylls was shown to depend only very weakly on the temperature (Groot et al. 1995), due to which the ratio of intensities of the connected pigments at low and high temperatures should be regarded as the lower limit. The emission anisotropy was calculated at the wavelength maximum of the emission band (Figure 3). In Figure 6 the anisotropy as function of excitation wavelength is shown. At the blue-most excitation there is very little anisotropy, while in the 704-712 nm region there is an anisotropy of 0.10. At longer excitation wavelengths the anisotropy gradually increases to a second plateau value of about 0.20 in the 7 2 0 725 nm region, while the anisotropy increases again above 725 nm to reach a value of -,~ 0.35 for excitation at wavelengths longer than 730 nm. This appears to be the maximal anisotropy since it no longer increases and remains on this level for excitation wavelengths longer than 730 nm. This excitation wavelength dependence of the anisotropy in Synechococcus differs considerably compared to what has been observed for PS I complexes from Synechocystis and PS 1-200 from spinach since the maximal anisotropy is obtained for much longer excitation wavelengths than in these two systems (Gobets et al. 1994). For excitation at shorter wavelengths than 712 nm the emission maximum appears at 731 nm (Figure 7) while for longer excitation wavelengths the emission maximum shifts with excitation wavelength. It was observed for excitation at 726 nm and longer wavelengths that the emission maximum is approximately 11 nm red-shifted (Figure 7). We remark that this shift is very similar to the shift observed for C-716 in PS 1-200 from spinach (Gobets et al. 1994). The data also

243 3000.0

3000.0

Synechococcus

Synechococcu.$

c= 696 nm 2000.0

.

Xexc= 730 nm 2000.0

bt ~ 1000.0

1000.0

°°75o ~o0 r:~o ~ o

Wavelength [rim]

~0

0.0 67S,0

800.0

700.0 725.0 7SO.O 778.0 800.0 Wavelength [nm]

Figure 3. Site-selective fluorescence spectra, laser-excited at 696 nm (left) or at 730 nm (fight). The laser excitation light was vertically polarized, while a Rochon polarizer in the detection beam separated the vertically polarized light (upper curves) from the horizontally polarized light (lower curves).

250000.0

0.40 T- 4 K

Synechococcu~

.

,

.

,

.

,

Synechococcus

200000.0

~o.3o

I) 150000.0

Tffi4K

-

,

.

_

/

: E.x.pe~-ne.ntal

~ 1~00.0 50000.0 o.o 670.0

~

690.0

710.0 730.0 Wavelcn~.h [nml

0.10

750.0

Figure 4. Fluorescence spectra of Synechococcous trimers at various temperatures upon unselective excitation at 590 nm.

°'~.0

7oo.o

7~ .0

~,0.0

739.0

74o.o

Excitation wavelength [nm] 1.0

SyncchococcuJ 0.8 P~

0.6

~

Figure 6. The fluorescence anisotropy in the emission maximum of trimeric P S I particles of Synechococcus trimers as a function of excitation wavelength at 4 K (filled circles). See Figure 7 for the values of the emission maxima at the different excitation wavelengths. Also displayed is the result of the simulated fluorescence anisotropy (open squares - see Discussion for details).

0.4

0.2

0"00.0

100.0

~.0

TzmPer~um IK]

300.0

Figure 5. The fluorescence yield of trimeric PS I particles of Synechococcus as a function of temperature. Excitation was at 590 nm. The emission yield was normalized to 1 at 4 K.

suggest that the C-719 pigment pool is mainly inhomogenously broadened since the slope of Aem over Aexc is ,-, 0.9 (upper line in Figure 7). If the slope of the line is equal to 1 this corresponds to a situation where homogeneous broadening is negligible (Gobets et al. 1994). This situation also reminds us of that observed for C-716 of PS 1-200, and differs somewhat from that

observed for C-708 of Synechocystis, where the homogeneous linewidth (the width of the phonon side wing) is not much smaller than the inhomogeneous distribution.

Discussion

Spectral characteristics of the long-wavelength chlorophylls The spectral decomposition of the 4 K absorption spectrum revealed that there are, in trimeric P S I complexes from Synechococcus, two different long-wavelength spectral components peaking near 708 and 719 nm,

244 74S.0

.

Synechococcus T=4K 740.0

..~ G 0 " ~ 735.0

730.0 . 690.0

. . . . . . . . . . . . . . . . . 700.0 710.0 720.0 730.0

740.0

Excitation wavelength [nm]

Figure7. The shift of the emission maximumas a function of excitation wavelength for trimeric PSI particles of Synechococcusat 4 K. which we designated C-708 and C-719, respectively. It was found that C-719 has a relatively large bandwidth (FWHM) of almost 20 nm (370 cm - l ) and that the spectral overlap with C-708 is substantial in the 705 to 720 nm region. The spectroscopic properties of C-708 of Synechococcus were found to be similar to those of the long-wavelength absorbing pigment C-708 of Synechocystis, but in Synechococcusit is not the emitting species. At 4 K all emission originates from C-719. By comparing the total integrated absorption profile with the integrated individual C-708 and C-719 bands, an estimate of the number of pigments can be done through their oscillator strengths. We base this estimate on a total number of 110 Chl per P S I monomer. This number was calculated from P700+/P700 absorbancedifference measurements and is based upon a recent re-evaluation of the extinction coefficient of P700 (E. Schlodder, unpublished observation). This estimate suggests that there are 4-5 C-708 pigments and 5-6 C-719 pigments per PS I monomer (or P700 dimer), provided that the oscillator strengths are similar to the strengths of the bulk chlorophylls. Tht~, the number of long-wavelength pigments in Synechococcus is much higher than in Synechocystis, where only ,-~ 2 C-708 pigments were found (Gobets et al. 1994). We note that the deconvolution of the red part of the absorption spectrum into two types of pigment pools (C-708 and C-719) represents the minimal sufficient description of the present data, but that we cannot exclude that in fact these pools should have to be divided into two or more sub-pools when more data (such as linear and circular dichroism data) become available. For the time being, however, we restrict

ourselves to the most simple interpretation of the data, which is based on two inhomogeneously broadened pigment pools peaking at 708 and 719 nm. The pronounced differences between the amounts and energies of the long-wavelength chlorophylls in Synechococcus and Synechocystis are expected to influence the equilibration kinetics in the two systems. Recently, Hastings et al. (1995) attributed a decay phase of 2.7-4.3 ps to equilibration in P S I particles from Synechocystis, noted that this time is 2-3 times shorter than the corresponding process observed previously by Holzwarth et al. (1993) in P S I particles from Synechococcus, and suggested that differences in the long-wavelength chlorophylls could have caused the observed differences in the equilibration kinetics. This is somewhat surpprising since the opposite should be expected, i.e. more long-wavelength pigments should result in a faster equilibration time between bulk chlorophylls and the long-wavelength pigment pool. The fact that a longer equilibration time for Synechococcus relative to that of Synechocystis is observed could indicate that there is a very special arrangement of the long-wavelength pigments in Synechococcus. It should also be mentioned that the depth in energy of the long-wavelength pigment pools could have an impact on the equilibration. Excitation at 4 K at the far red edge of the absorption spectrum with 730 nm light resulted in a maximal anisotropy of the emission (Figure 6). This means that upon this type of excitation, chlorophylls are excited which at 4 K no longer transfer their energy to neighboring chlorophylls. In other words, at and above 730 nm the probability is very high that the red-most pigment within the connected pool of inhomogeneously broadened C-719 pigments is excited, which at 4 K will not transfer energy to the other (higher energy) chlorophylls. An inspection of Figure 3 (right trace) reveals that the emission peaks at about 740 nm upon this excitation. The observed Stokes shift (,--200 c m - 1) is of the same order as found previously in PS 1-200 (,-,240 cm -1) and almost two times larger than found for C-708 in Synechocystis(,-- 120 c m - 1) (Gobets et al. 1994). These shifts are much larger than observed in uncoupled Chl (Kwa et al. 1994) or in LHC-II (Reddy et al. 1994) and were earlier attributed to enhanced pigment-pigment interactions (Gobets et al. 1994). Taken together, the spectral properties of C-719 of Synechococcus resemble those of C-716 of PS 1-200 of spinach and to a smaller extent also those of C-708 of Synechocystis.

245

Energy transfer Since in Synechococcus there is more than one C-719 pigment per PS I monomer, it is in principle possible that there is energy transfer from C-719 pigments on the high-energy side of the inhomogenous distribution to C-719 pigments on the low-energy side. A consequence of energy transfer is a loss of anisotropy (unless the dipole moments are aligned parallel). If, on the other hand, a pigment within the inhomogenous distribution can not transfer its energy, a high polarization of the emission will prevail (van der Lee et al. 1993; Wittmershaus et al. 1992). The probability that in a cluster of N pigments a pigment excited at Ae is the red-most one is given by van Mourik et al. (1992)

]

N-I

P ()%[)%) =

A ()~) d)~

where A(A) is the absorption. From this approach the residence probability or the trapped fraction r/(0 < ~ < 1) can be estimated. The trapped fraction is then related to the anisotropy of a certain excitation wavelength according to the relation r ()%) = 0.4. r/()~e) By applying these relations we simulated the fluorescence anisotropy as a function of excitation wavelength for a certain width of the inhomogeneous (and homogeneous) distributions and for an arbitrary number of C-719 long-wavelength pigments. It was accordingly found that the simulated anisotropy profile which best coincides with the experimental data requires an inhomogenous distribution of 15 nm (340 cm -1) and consists of 4 long-wavelength pigments, as is shown in Figure 6 (open squares). The simulation furthermore required that the inhomogeneous broadening dominates over homogeneous broadening. The data of Figure 7, in particular those regarding the slope of the emission maximum as a function of excitation wavelength, show that this is indeed the case. The energy transfer between C-719 chlorophylls is more likely to occur within a P S I monomer than between monomers, because the latter process will probably be inefficient since this will be an donordonor type of energy transfer process over relatively long distances. On the other hand, if the analogy between the long-wavelength pigments of Synechococcus and Synechocystis holds, then also the long-wavelength pigments of Synechococcus could be

arranged as dimers (in Synechocystis the two C-708 chlorophylls behaved as a N = 1 system, due to which they were proposed to be organized as a chlorophyll d i m e r - Gobets et al. 1994). If the 5-6 C-719 chlorophylls of Synechococcus are arranged as three dimers, then a N = 3 system is expected. The simulation in Figure 6 can then be explained by some energy transfer between C-719 pigments on different monomers. On the other hand, there is as yet no firm evidence in favour of a dimeric configuration. Experiments on monomeric P S I complexes of Synechococcus may shed more light on this issue. Another interesting feature is the rise of the anisotropy as a function of excitation wavelength (Figure 6). The rise of the anisotropy is not continuous, which differs from the situation in the Synechocystis particles and in PS 1-200 of spinach where smooth curves were observed (Gobets et al. 1994). Furthermore, the peak maximum of the emission exhibits a more complicated wavelength dependence in Synechococcus than in Synechocystis (Figure 7). These findings can be explained by the results of the spectral decomposition of the absorption spectrum into Gaussian components (Figure 2, Table 1). The first plateau value in the anisotropy excitation profile between 704 and 714 nm (Figure 6) corresponds to the absorption maximum of C-708 and indicates that on the average there is an angle between the absorbing C-708 dipole and the emitting C-719 dipole. The second plateau value between 718 and 723 nm (Figure 6) coincides with the absorption maximum of C-719 and suggests that the C-719 dipoles are not randomly distributed in a plane, but confined within a relatively small angle. In the intermediate region of 714-719 nm the spectral bands overlap, giving rise to a complex anisotropy behavior. The energy difference AE between the absorption maxima of C-719 and P700 is about 400 cm - l . Accordingly, at 4 K AE > > kBT. Even at 90 K AE > kBT(kBT ,.~ 65 cm -1 at 90 K), but as can be seen in Figures 4 and 5, the emission quantum yield is clearly less than at 4 K. This suggests that at this temperature the rate of uphill energy transfer from C-719 to P700, which perhaps occurs via the C-708 pigment is no longer negligible. At room temperature only very little emission is observed, suggesting very efficient uphill energy transfer. There may be several reasons why the uphill energy transfer is rather efficient at higher temperatures. Both P700 and C-719 are characterized by large homogeneous bandwidths; the Stokes shift of C-719 is about 200 cm -1, which points to a very pro-

246 nounced phonon side band and which is in fact similar to the width of the phonon side band of P700 (Gillie et al. 1989). The broad homogeneous bandwidths will increase the spectral overlap between C-719 and P700, and therefore may enlarge the rate of energy transfer. In addition, the homogeneous bandwidths of C-719 and P700 will further increase upon raising the temperature. The inhomogeneous distribution of the energy levels of C-719 and P700 is also considerable (,-~370 cm -1 - see above for C-719 and Gillie et al. 1989 for P700), due to which in some complexes the energy gap will be considerably smaller. As a last possible explanation we mention a possible short distance between C-719 and P700 (with perhaps C-708 as an intermediate), which will increase the energy transfer rates between these pigments and may increase the possibility that the excitation is trapped by P700.

Acknowledgements This project was supported by the Netherlands Foundations for Chemical Research (SON) and Life Sciences (SLW). LOP is supported by a fellowship within the Human Capital & Mobility program project ERBCHBGCT 930361 supported by the European Union. LOP is also grateful to Erwin J.G. Peterman for the introduction to high resolution spectroscopy.

References Gillie JK, Lyle PA, Small GJ and Golbeck JH (1989) Spectral hole burning of the primary electron donor state of Photosystem I. Photosynth Res 22:233-246 Gobets B, van Amerongen H, Monshouwer R, Kruip J, R6gner M, van Grondelle R and Dekker JP (1994) Polarized site-selected fluorescence spectroscopy of isolated Photosystem I particles. Biochim Biophys Acta 1184:75-85 Groot M-L, Peterman EJG, van Stokkum IHM, Dekker JP and van Grondelle R (1995) Triplet and fluorescing states of the CP47 antenna complex of Photosystem II studied as a function of temperature. Biophys J 68:281-290 Hastings G, Reed LJ, Lin S and Blankenship RE (1995) Excited state dynamics in Photosystem I: Effects of detergent and excitation wavelength. Biophys J 69:2044-2055 Holzwarth AR, Schatz G, Brock H and Bittersman E (1993) Energy transfer and charge separation in Photosystem I. Part I: Picosecond transient absorption and fluorescence study of cyanobacterial Photosystem I particles. Biophys J 64:1813-1826 Krauss N, Hinrichs W, Witt I, Fromme P, Pritzkow W, Dauter Z, Betzel C, Wilson KS, Witt HT and Saenger W (1993) Threedimensional structure of system I of photosynthesis at 6 ,~ resolution. Nature 361:326-331

Kwa SLS, V61ker S, Tilly NT, van Grondelle R and Dekker JP (1994) Polarized site-selection spectroscopy of chlorophyll a in detergent Photochem Photobiol 59:219-228 Laible PD, Zipfel W and Owens TG (1994) Excited state dynamics in chlorophyll-based antennae: The role of transfer equilibrium. Biophys J 66:844-860 P~sson L-O, Schlodder E, van Grondelle R and Dekker JP (1995) Properties of the long-wavelength emitting chlorophylls in isolated Photosystem I particles of Synechococcus sp. In: Mathis P (ed) Photosynthesis: From Light to Biosphere, Vol II, pp 195198. Kluwer Academic Publishers, Dordrecht, The Netherlands Reddy NRS, van Amerongen H, Kwa SLS, van Grondelle R and Small GJ (1994) Low energy exciton level structure and dynamics in LHC-II trimers from the Chl a/b antenna complex of Photosystem II. J Phys Chem 98:4729-4735 Shubin VV, Bezsmertnaya IN and Karapetyan NV (1992) Isolation from Spirulina membranes of two Photosystem I-type complexes, one of which contains chlorophyll responsible for the 77 K fluorescence band at 760 nm. FEBS Lett 309:340-342 Trinkunas G and Holzwarth AR (1994) Kinetic modelling of exciton migration in photosynthetic systems. 2. Simulations of exciton dynamics in two-dimensional Photosystem I core antenna/reaction center complexes. Biophys J 66:415-429 Trissl H-W and Wilhelm C (1993) Why do thylakoid membranes from higher plants form grana stacks? Trends Biochem Sci 18: 415-419 Turconi S, Schweitzer G and Holzwarth AR (1993) Temperature dependence of picosecond fluorescence kinetics of a cyanobacterial Photosystem I particle. Photochem Photobiol 57:113-119 Valkunas L, Liuolia V, Dekker JP and van Grondelle R (1995) Description of energy migration and trapping in Photosystem I by a model with two distance scaling parameters. Photosynth Res 43:149-154 Van der Lee J, Bald D, Kwa SLS, van Grondelle R, R6gner M and Dekker JP (1993) Steady-state polarized light spectroscopy of isolated Photosystem I complexes. Photosynth Res 35:311-321 Van Grondelle R, Dekker JP, Gillbro T and Sundstr6m V (1994) Energy transfer and trapping in photosynthesis. Biochim Biophys Acta 1187:1-65 Van Mourik F, Visschers RW and van Grondelle R (1992) Energy transfer and aggregate size effects in the inhomogenously broadened core light-harvesting complex ofRhodobacter sphaeroides. Cbem Phys Lett 195:1-7 V61ker S (1989) Hole-burning spectroscopy. Ann Rev Phys Chem 40:499-530 Werst M, Jia Y, Mets L and Fleming GR (1992) Energy transfer and trapping in the Photosystem I core antenna. A temperature study. Biophys J 61:868-878 Witt I, Witt HT, Gerken S, Saenger W, Dekker JP and ROgner M (1987) Crystallization of reaction center I of photosynthesis. Low-concentration crystallization of photoactive protein complexes from the cyanobacterium Synechococcus sp. FEBS Lett 221: 260-264 Wittmershaus BP, Woolf VM and Vermaas WFJ (1992) Temperature dependence and polarization of fluorescence from Photosystem I in the cyanobacterium Synechocystis sp. PCC 6803. Photosynth Res 31:75-87