Photosynthesis Research 24: 201-208, 1990. © 1990 KluwerAcademic Publishers. Printed in the Netherlands Regular paper

Detection of photosynthetic energy storage in a photosystem I reaction center preparation by photoacoustic spectroscopy Thomas G. Owens I, Robert Carpentier 2 & Roger M. Leblanc 2 1Section of Plant Biology, Cornell University, Ithaca, N Y 14853-5908, USA (for correspondence and reprints); 2Centre de Recherche en Photobiophysique, Universitd du Quebec fi Trois-Rivikres, C.P. 500, Trois-Rivikres, Quebec, G9A 5H7 Canada Received 3 July 1989; acceptedin revised form 18 December1989

Key words:

photosynthesis, thermal emission, P700, quantum yield, energy conversion

Abstract

Thermal emission and photochemical energy storage were examined in photosystem I reaction center/ core antenna complexes (about 40 Chl a/P700) using photoacoustic spectroscopy. Satisfactory signals could only be obtained from samples bound to hydroxyapatite and all samples had a low signal-to-noise ratio compared to either PSI or PS II in thylakoid membranes. The energy storage signal was saturated at low intensity (half saturation at 1.5 W m -2) and predicted a photochemical quantum yield of >90%. Exogenous donors and acceptors had no effect on the signal amplitudes indicating that energy storage is the result of charge separation between endogenous components. Fe(CN)63 oxidation of P700 and dithionite-induced reduction of acceptors FA-F B inhibited energy storage. These data are compatible with the hypothesis that energy storage in PSI arises from charge separation between P700 and Fe-S centers FA-F B that is stable on the time scale of the photoacoustic modulation. High intensity background light (160Wm -2) caused an irreversible loss of energy storage and correlated with a decrease in oxidizable P700; both are probably the result of high light-induced photoinhibition. By analogy to the low fluorescence yield of PS I, the low signal-to-noise ratio in these preparations is attributed to the short lifetime of Chl singlet excited states in PS 1-40 and its indirect effect on the yield of thermal emission.

Abbreviations: F F T - fast F6urier transform; H A - hydroxyapatite; 150- half saturation intensity for energy storage; P A - photoacoustic; P S - photosystem; PS 1-40- photosystem I reaction center/core antenna complex containing about 40 Chl a/P700, ~b"- photoacoustic energy storage signal; S / N signal-to-noise

Introduction

The fate of singlet Chl excited states in photosynthetic systems is determined by the competition among the useful processes of singlet energy transfer and photochemical charge separation with the wasteful processes of intersystem crossing (triplet formation), thermal and fluorescence emission. Under optimal conditions, once an

excited state reaches a reaction center, photochemistry proceeds with a quantum yield that approaches unity (Wraight and Clayton 1973, Hiyama 1985). In real photosynthetic systems the overall quantum yield of photochemistry may be diminished as the result of several processes including: 1. the finite lifetime of the excited state in the antenna and

202 2. the time required for secondary electron transport to regenerate an "open" trap. Because fluorescence and thermal emission are competing with photochemistry for excited state energy, fluorescence and photoacoustic techniques provide powerful tools for analysis of photosynthetic systems (Butler 1977, Malkin et al. 1981). For thermal emission, the detection of photosynthetic energy storage (also called photochemical loss) is of particular utility (Malkin and Cahen 1979, Carpentier et al. 1983b, 1984). In measurement of photosynthetic energy storage using PA techniques, a distinction must be made between the total PA signal and the energy storage signal. The total PA signal is the sum of all thermal decay processes whose half-times are fast compared to the effective PA modulation period. The energy storage signal is the difference between the maximum PA signal (obtained by closing all traps with continuous background light) and the signal obtained without background light (some fraction of the traps open). This decrease in thermal emission with open traps is associated with the formation of stable (on the time scale of the modulation frequency) redox equivalents stored in electron transport components. Mathematical treatments of energy storage have been previously presented (Malkin and Cahen 1979, Carpentier et al. 1989a). Recently, we have used simultaneous measurements of thermal and fluorescence emission in uncoupled spinach thylakoids to demonstrate that energy storage in PS II is the result of stable charge separation between the oxygen evolving complex and plastoquinone pools (Carpentier et al, 1989b). Analogous studies with PSI are complicated by the fact that PSI exhibits little or no variable fluorescence (Ikegami 1976, Butler 1977). Picosecond fluorescence decay studies with isolated PSI reaction center complexes (PS 1-40, about 40 Chl a/P700) showed that the lifetime of Chl excited states in the complexes did not vary between open and closed (Fe(CN)63-oxidized) samples (Owens et al. 1988, Holzwarth et al. 1989). This lack of variable fluorescence in PSI could result in substantially higher thermal emission and PA signals in closed PS I traps in comparison to PS II. Using PA spectroscopy, energy

storage in PSI has been used to measure the action spectrum for PSI in thylakoid membranes (Carpentier et al. 1984). Energy storage during cyclic electron transport around PSI was also measured in heterocysts isolated from cyanobacteria (Carpentier et al. 1986). Here we present the first measurements and analysis of energy storage in isolated PSI reaction center complexes.

Materials and methods

Photosystem I reaction center/core antenna complexes (PS 1-40) were isolated from selected market spinach according to the procedure of Owens et al. (1988). Briefly, the procedure involves extraction of thylakoid membranes in 1% Triton X-100 followed by hydroxyapatite (HA) chromatography. After about 50% of the nonPSI pigment was washed from the HA column, the pigment-containing HA was removed from the column and the remaining washes were accomplished in a slurry of phosphate buffer followed by brief centrifugation. This procedure permits detemination of the Chl a/P700 ratio from small aliquoits of the HA prior to bulk elution of the sample, and is necessary to minimize sample contamination with protein-bound core Chls that are uncoupled in energy transfer from P700 (Owens et al. 1987, 1988). PS 1-40 samples were eluted in 0.2 M P O 4 buffer pH 7.0, 0.05% Triton X-100. The resulting complex contains a doublet of polypeptides near 60 kDa and the five low molecular weight polypeptides characteristic of PS I core complexes, but completely lacks polypeptides of the cytochrome b6/f complex (Owens unpublished data). Absorption spectra were measured on a SLM Aminco DW-2000 spectrophotometer. The ratio of Chl/P700 was determined from ferricyanideoxidized minus ascorbate-reduced spectra (Markwell et al. 1980) using a differential extinction coefficient of 64mM -1 (Hiyama and Ke 1972). Samples for photoacoustic analysis were prepared by gentle aspiration of washed HA (containing the PS 1-40 complex) or eluted complexes on nitrocellulose filters (Millipore Corp. type AA, 25 mm diameter, 0.8 mm pore size). HA-

203 bound PS 1-40 samples were 0.3-0.5 mm thick and contained 150 nmol Chl a~ filter. Filters were cut to the proper dimensions for introduction into the photoacoustic cell (Ducharme et al. 1979). Photoacoustic measurements were made on a laboratory-constructed spectrometer as previously described (Carpentier et al. 1983b). The PA measuring beam (680 nm) was modulated by a mechanical chopper at 35 Hz resulting in an effective PA sampling time of 4.5 ms (=l/2~-f) (Carpentier et al. 1985). The intensity of the modulated beam was attentuated using a calibrated variable slit on the excitation side of the monochromator. Non-modulated background illumination (380-750 nm, 220 W m - maximum) was provided by a quartz-halogen fiber optics illuminator and attenuated with neutral density filters. Photosynthetic energy storage was calculated as 05[ = 100x ( Q ~ , - Qc)/(Qm where Qm and Qc are the PA signals in the presence and absence of non-modulated background light, respectively (Carpentier et al. 1984). A single measurement of 05; was made on each sample and data are reported as averages over 3 to 8 replicates. The output of the lock-in amplifier was plotted on a chart recorder with an effective time constant of 3 s. The recorder traces were digitized (at 0.5 s resolution) and the major noise contributions determined by FFT analysis. The S/N ratio was improved using appropriate digital lowpass and notch filters. Finally, signal levels at Qc and Qm were determined by averaging 30s windows centered between changes in sample illumination.

Results and discussion

Because of the lack of variable fluorescence in PS I, we initially expected that the PA signal per unit Chl of our PS 1-40 preparations might be large compared to those of PS II under conditions when traps were closed. However, measurement of detergent-solubilized PS 1-40 preparations aspirated onto membrane filters produced unusually low signal-to-noise and poor reproducibility in estimation of 05;. Varying the amount of Chl on the filter between 0.15 and 0.3 ~mol did not affect the signal-to-noise. As this amount of Chl is comparable to previous

measurements on algae, thylakoids and PS II particles (Carpentier et al. 1983b, 1989a,b), we conclude that there must be other effects (potentially aggregation of the complexes or interaction with the filter) that alter the signal amplitude. Measurement of Chl/P700 ratios on the filter by chemical oxidation with ferricyanide gave results that were identical to the soluble PS 1-40 (not shown) indicating that the P700 reaction center is not damaged by aspiration on the filter. Ultrafiltration/dialysis of the soluble PS 1-40 extract on PM30 membranes (Amicon) against 0.2 M PO 4 buffer without detergent did not improve the signal. Subsequently, we used samples in which the PS 1-40 preparation remained bound to the HA column material from the final step of purification. (Soluble PS 1-40 samples were the same material eluted from the HA in PO 4 buffer containing 0.05% Triton X-100.) An example of the PA signal measured in the absence of added donors or acceptors is shown in Fig. 1A. Addition of high intensity, nonmodulated background light closes the PSI traps and induces an increase in the PA signal to the Qm level. Typical signal-to-noise of the unprocessed lock-in output at Qm was about 10 at 9.3 W m z modulated intensity and increased to 25 at 20.2 W m -2. The S/N ratio in measurement of 05" was about 2 for all modulated intensities. Despite the improvement in S/N with the use of HA-bound PS 1-40, the S/N is small compared with thylakoids or PS II particles (Carpentier et al. 1989a,b). However, the errors in estimation of &" were substantially improved (+8% standard deviation on five replicate samples) by digital filtering of the digitized recorder traces (not shown). The effects of exogenous donors and acceptors on the PA signals and 05" from HA-bound PS 140 are summarized in Table 1. Addition of ascorbate or methyl viologen had no significant effect on either the PA signals or energy storage. In solution, ascorbate and methyl viologen act as donor and acceptor to PS 1-40 with 71/2 of several seconds or more (Hiyama and Ke 1972) and frequently enhance the ability to detect lightinduced oxidation of P700 (Shiozawa et al. 1974). The lack of effect here may be due in part to reduced accessibility to PS 1-40 when bound to HA and aspirated on filters. Oxidation of P700

204 2

2

3

Qm

,vVV, ,

I I Vyvv ' v'v'vv'v' v

Qe

I

O.l m V

I

I I rain

1

1

Fig. I. Photoacoustic signals from a HA-bound PS 1-40 preparation before digital processing. Modulated (35 Hz) intensity = 9.3 W m -2, background intensity = 160 W m -2. A: no additions; B: + Fe(CN)2 a. Arrows indicate changes in sample illumination as follows: (1) modulated light on; (2) background light on; (3) background light off. (Qc) control level of PA signal; (Qm) maximum level of PA signal.

Table 1. The effects of exogenous donors, acceptors and redox reagents on the PA signals measured on hydroxylapatite-bound PSI-40 samples. Modulated excitation was 9.3 W m-2 at 680 nm, background intensity was 1 6 0 W m -2. Chl a/P700 ratio of sample was 43. Reagents were added to the hydroxylapatite slurry in 10raM PO 4 buffer pH7.0 and mixed for 2 min prior to filtration and measurement. Units of Qc and Qm are 10 -4 V. Data are reported as averages from 3 to 5 replicates per treatment (standard deviation of 9% for 5 replicates) Treatment

Qc

Qm

~ rt ( O~)

No additions Ascorbate MV Ascorbate + MV Ferricyanide Dithionite

38 40 38 40 41 45

32 44 42 45 41 45

9.4 9.1 9.6 10 0.0 0.0

by ferricyanide (Fig. 1B) or reduction of ironsulfur centers FA-F a by dithionite both produced maximum PA signals in the absence of background illumination and a total loss of energy storage, indicating that energy storage in PS 140 is likely the result of charge separation between endogenous electron transport components. The detection of energy storage in PA spectroscopy requires that the products of the photochemical reaction be stable on a time scale de-

termined by the modulation frequency. For PA spectroscopy with 35Hz modulated excitation, this time scale is 1/2~-f or 4.5 ms (Carpentier et al. 1983a). The inhibition of energy storage by ferricyanide or dithionite suggests that charge separation between P700 and Fe-S centers F AF B may be involved in the energy storage signal. Since plastocyanin (the endogenous donor to P700 +) and ferredoxin (the acceptor from (F AFb)- ) are absent in our PS 1-40 preparation and exogenous compounds do not affect the energy storage signal, the stability of charge separation will be determined by the recombination rate between P700 ÷ and (FA-FB)-. In a similar PS I preparation, Golbeck and Cornelius (1986) measured a rl/2 for charge recombination between P700 ÷ and (FA-Fs)- of 30 ms. Removal of FA-F B with detergent or its reduction with dithionite yield room temperature ~'1/2 values for the back reaction between P700 ÷ and F x ( = A 2) of 1.2 ms and 250 ms, respectively (Golbeck and Cornelius 1986). These temporal data suggest that PA energy storage in PS 1-40 is the result of charge separation between P700 and Fe-S centers FA-F B. In previous studies, variations in ~b" with modulated excitation intensity have been attributed to partial saturation of the photochemical reac-

205 is determined by the rate of charge recombination in the absence of endogenous donors and acceptors (Carpentier et al. 1985, 1989a). In contrast, the rate of PS ii trap reopening, resulting from whole chain electron transport in subsaturating light, is sub-ms (Mathis and Rutherford 1986) and yields a higher forward electron transport rate and 150 (Carpentier et al. 1989a). The data in Fig. 2 also provide an estimate for the maximum photochemical quantum yield in this PS 1-40 preparation. The value of the yintercept, 1/tb'0, is the maximum energy storage yield (=58%) and is related to the photochemical quantum yield (~be) by:

tion by the modulated light (see Carpentier et al. 1989 for a detailed discussion). A similar treatment of HA-bound PS 1-40 (Fig. 2) shows a typical hyperbolic decline in ~b" with increasing modulated intensity that is qualitatively similar to measurements of PS II in thylakoid membranes (Carpentier et al. 1989b) and in PS II particles (Carpentier et al. 1989a). A plot of 1/~br' against modulated intensity provides an estimate for I50, the half-saturation constant of energy storage, of 1.5 W m -2. This is similar to I50 values determined in PSII particles (1.7Win -2) (Carpentier et al. 1989a) but less than those for PS II energy storage in thylakoid membranes ( 8 . 5 W m -2) (Carpentier et al. 1989b). Previous studies have shown that I50 is closely related to the rate of electron transfer t while the maximum energy storage (~b,0) is dependent only on the intermediates in which energy is stored (Carpentier et al. 1989a). In PS 1-40, at a given modulated intensity the forward electron transfer rate is determined by the fraction of open traps and is related to the Qc level of the PA signal. The fraction of open traps is, in turn, determined by the forward electron transport rate (trap closing) and the recombination rate between reduced FA-F B and P700 ÷ (reopening traps). This recombination process has a ~1/2 of about 30 ms at room temperature (Golbeck and Cornelius 1986). A similar description applies to PS II particles in which the rate of trap reopening (and thus the forward electron transport rate)

12

I

,

.

i\

O-

,

6"0 =

[aGA/Nhc]

where [AEp A / N h c ] represents the ratio of energy stored in photochemical products to the absorbed light energy at wavelength A (Malkin and Cahen 1979). The term aEp is equal to the energy conserved in the redox products of the photochemical reaction detected in the PA signal. For whole chain electron transport in thylakoids, the value of AEp depends on the working potentials of P700/P700 ÷ and (FA-FB)/ (FA-FB)- and are likely to deviate substantially from their standard midpoint potentials (Arcelay et al. 1988). In PS 1-40, assuming that forward charge separation and charge recombination are the principle reactions determining the redox state of the sample, AEp will be independent of

t

~

~

,

i

o\\

,

[~ 0.35

~

0.50

I "S. 6 ÷

4

+o.15T ,, I

---Q---io.lo

20 -5

o.2o "ic" 0.05 "-"

, vwq .... , " '," 0 5 MODULATED

5 % ,--~° 10

,

~ 15

INTENSITY

,

, 20

,

4-0.00 25

(W m - 2 )

Fig. 2. The effect of modulated beam intensity on the energy storage signal (th") in HA-bound PS 1-40 (Chl a/P700 = 51).

Background intensity = 160W m-2. Open circles and dashed line: ~b"; filled circles and solid line: 1/4>"• Solid line is a regression of 1/4," vs. intensity for determination of 150and maximum energy storage yield, 4~r'0;dashed line is the fit of ~b" vs. intensity derived from the 1/6" regression.

206 the fraction of traps open and equal to 1.04 eV. This predicts a maximum ratio of energy stored to energy absorbed (at 680 nm) of 0.57 and a photochemical quantum yield of unity. Previous investigations of thylakoid membranes (Avron and Ben-Hayyim 1969) and large PS I particles (Hiyama 1985) have measured PSI photochemical quantum efficiencies approaching unity. However, measurements on PS 1-40 preparations (Owens et al. 1987) indicate a maximum quantum yield of about 90%. Time-resolved fluorescence measurements show that there are two contributions to the lower quantum efficiency in PS 1-40. First, the PS 1-40 preparations contain a small amount (5-10%) of protein-bound core antenna Chls that are not functionally coupled in energy transfer to P700 (Owens et al. 1988). After correction for this loss, the intrinsic efficiency of functional complexes was 85-95% (Owens et al. 1987) suggesting that other losses may result from the detergent isolation procedure. The fate of Chl excited states in uncoupled PSI core antenna pigments complicates calculation of ~bp and ~br'0. If the excited states in the uncoupled Chls decay via thermal emission, the long lifetime of these excited states (Owens et al. 1987, 1988) would produce thermal emission that was 100 times larger than that from coupled PS I Chls (see below). The fact that PA energy storage in PS 1-40 exhibits a dependence on modulated light intensity that is similar to PS I or PSII in thylakoids (Carpentier et al. 1983a, 1989b) indicates that either the uncoupled Chls are absent in this preparation or that their decay is not detected under our experimental conditions. Because fluorescence and thermal emission are competing for singlet excited state energy in photosynthetic systems, the lack of variable fluorescence associated with PSI trap closing suggested that the PA signal from closed PSI traps might be large in comparison to PS II. This expectation was not fulfilled as the PS 1-40 samples exhibited a lower S/N ratio than PS II at similar Chl concentrations. The yield of thermal emission (for a fixed rate of light absorption and all traps open) depends directly on the relative rate constants for fluorescence (kl), thermal emission (kd), triplet formation (kt) and the effective rate constant for photochemistry (in-

cluding antenna transfer and trapping contributions, kp). For both PSI and PSII, the high photochemical quantum yields show that kp must be large compared with the other competing processes. Under these conditions, fluorescence and thermal emissions do not significantly deplete the Chl excited state population and the amount of these emissions will strongly depend on the Chl singlet excited state lifetime in each photosystem. The measured fluorescence lifetimes for PSI are shorter than those of PS II with the result that PSI contributes only about 5% to total steady state fluorescence emission (Gulotty et al. 1985, Owens et al. 1987). Assuming comparable values of k r and k a between PSI and PS II, thermal emission from PSI must be similarly low. For PS 1-40 (40 Chl a/P700), this lifetime is shortened by a factor of 3 compared to PSI in thylakoids (120 Chl a/P700) because the lifetime of excited states in PSI is linearly related to the core antenna size (Owens et al. 1987). Thus the short Chl singlet lifetimes characteristic of PS I, and in PS 1-40 in particular, are likely to be major factors in the low S/N ratios observed in these samples. In addition to the low S/N, the determination of Qm and ~br' in HA-bound PS 1-40 samples also differed from previous measurements on PS II and PSI in that the rise to Qm induced by background light was not reversible upon removal of the background light for all background intensities examined (75-200 W m-2). However, when the HA-bound PS 1-40 is pretreated with ferricyanide to oxidize P700, the PA signal rises immediately to the Qm level of the untreated sample and the effect of the background light is lost (Fig. 1B). These data suggest that irreversibility of Qm reflects an inability to reduce P700 + after illumination with background light in the HA-bound sample. This was investigated by measuring the Chl a/P700 ratio of PS 1-40 eluted from the HA before and after illumination in the photoacoustic cell. The results (Table 2) show that the low intensity modulated illumination (