Dark inactivation of ferredoxin-NADP reductase and cyclic electron flow under far-red light in sunflower leaves

Abstract

The oxidation kinetics under far-red light (FRL) of photosystem I (PSI) high potential donors P700, plastocyanin (PC), and cytochrome f (Cyt f) were investigated in sunflower leaves with the help of a new high-sensitivity photometer at 810 nm. The slopes of the 810 nm signal were measured immediately before and after FRL was turned on or off. The same derivatives (slopes) were calculated from a mathematical model based on redox equilibrium between P700, PC and Cyt f and the parameters of the model were varied to fit the model to the measurements. Typical best-fit pool sizes were 1.0–1.5 μmol m⁻² of P700, 3 PC/P700 and 1 Cyt f/P700, apparent equilibrium constants were 15 between P700 and PC and 3 between PC and Cyt f. Cyclic electron flow (CET) was calculated from the slope of the signal after FRL was turned off. CET activated as soon as electrons accumulated on the PSI acceptor side. The quantum yield of CET was close to unity. Consequently, all PSI in the leaf were able to perform in cycle, questioning the model of compartmentation of photosynthetic functions between the stroma and grana thylakoids. The induction of CET was very fast, showing that it was directly redox-controlled. After longer dark exposures CET dominated, because linear e⁻ transport was temporarily hindered by the dark inactivation of ferredoxin-NADP reductase.

Introduction

The rate and role of the cyclic electron transport around PSI has been an enigma for about a half of a century, mainly because it is difficult to measure the relatively slow cycling of electrons on the background of the fast linear flow. To bypass the problem, conditions have been chosen where the cyclic flow is maximized, e.g., under light saturation and stress conditions (Clarke and Johnson 2001; Joët et al. 2002; Golding and Johnson 2003; Miyake et al. 2005a, b) and linear flow is minimized, e.g., during the dark-light induction of photosynthesis (Joliot and Joliot 2002; Joliot et al. 2004). Despite of some methodical uncertainties, there is general consensus that cycling of electrons around PSI is accelerated as soon as the linear flow is restricted. The cycle is thought to function mainly around PSI located in the stroma-exposed thylakoids (Albertsson 1995; Joliot et al. 2004). The proposed pathway involves ferredoxin (Fd), donating electrons to the high-potential Cyt b_H of the Cyt b₆f complex either directly or via the newly discovered Cyt c’, and further to plastoquinone (PQ) at the Q_I site. The cycling electrons enter the Q-cycle, being coupled with H⁺ transport.

In this work, following Joliot and Joliot (2005), we have taken the approach to induce CET by illuminating the leaf with far-red light (FRL) exciting mainly PSI. Under this condition the linear flow is very slow, facilitating the measurement of CET. The experimental signal used in this work—leaf transmittance changes at 810 nm—indicate the reduction state of electron carriers on the donor and acceptor side of PSI (Harbinson and Woodward 1987). The signal is complex, containing components from PC⁺, P700⁺, and Fd⁻ (Klughammer and Schreiber 1991). Under conditions where the accumulation of Fd⁻ is limited and fluxes are slow enough not to induce redox disequilibrium, the signal has been deconvoluted into the components of PC⁺ and P700⁺ on the assumption of redox equilibrium between P700, PC and Cyt f (Oja et al. 2003). This method of oxidative titration of PSI donors is based on the time kinetics the 810 nm signal during the gradual oxidation by FRL of the dark-reduced PSI donor side carriers. Already our first experiments (Oja et al. 2003), as well as those of Joliot and Joliot (2005) showed the speed of oxidation by FRL decreased after longer pre-exposures in the dark, as well as during the continuing electron transport under FRL. The reasons for this could be either the transported electrons accumulated on the acceptor side, blocking further electron transport by charge recombination, or the transported electrons returned to the donor side via a pathway of CET. Intuitively, the assumption of CET has been used to explain the apparently low quantum efficiency of FRL (Asada et al. 1992, 1993; Joliot and Joliot 2005), neglecting the possibility of charge recombination. In this work we measured the time kinetics of electron arrival at the PSI donor side with an aim to see whether the PSI electron transport efficiency was controlled by charge recombination or cyclic electron transport. We constructed a new modulated spectrophotometer for the high-resolution measurement of small changes in leaf transmittance, significantly improving the quality of the oxidative titration measurements. The method of oxidative titration was further elaborated by considering the simultaneous inflow of electrons during the oxidation of the PSI donors under FRL. We showed that the quantum yield of PSI remained high under FRL, supporting the cyclic pathway with the quantum yield close to unity.

Materials and methods

Plant material and growth conditions

Sunflower, Helianthus annuus L. plants were grown in a growth chamber in 4-l pots in fertilized peat-soil mixture at 25/20° and a 14/10 h day/night regime. The incident PFD from high-pressure mercury vapor lamps at the top of the plants was 450 μmol quanta m⁻² s⁻¹. Fully expanded attached leaves were used in experiments.

Gas exchange measurements

In the two-channel fast-response photosynthesis measurement system (Fast-Est, Tartu, Estonia) (Laisk and Oja 1998; Laisk et al. 2002) a part of a leaf was enclosed in a sandwich-type chamber (diameter 31 mm, height 3 mm) where the gas flow rate was 0.5 mmol s⁻¹. For the sake of temperature stabilization and mechanical fixation the upper epidermis of the leaf was sealed with starch gel to the glass window of the chamber, washed by thermostatted water from the other side. The bath temperature was 22°C and leaf temperature did not exceed 22.5°C at the highest illumination. Gas exchange proceeded through the lower epidermis. Humidity was controlled by passing an adjustable fraction of the gas over water at 50°C and was measured by a psychrometer referenced to dry air. The vapor pressure deficit of the flushing gas was 1.7 kPa. Stomatal opening was indicated by the transpiration rate. Measurements were carried out only when stomata were sufficiently open (leaf gas phase resistance was <0.4 s mm⁻¹). The used O₂ concentration of 210 mmol mol⁻¹ and CO₂ concentration of 360 μmol mol⁻¹ were obtained by mixing pure gases at controlled flow rates.

Leaf illumination

Plastic fibers (1 mm, Toray Polymer Optical Fiber, PF series, from Laser Components, Gröbenzell/München, Germany) were individually arranged to produce uniform illumination of the chamber-enclosed leaf surface from the following three sources. A Schott KL 1500 tungsten halogen light source equipped with a heat-reflecting filter (Optical Coating Laboratory, Inc., Santa Rosa, CA) provided white actinic illumination (WL), used for preconditioning the leaf for experiments (opening of stomata, activation of enzymes). Another Schott KL1500 equipped with a heat-reflecting filter provided saturating multiple turnover pulses of 10000 μmol quanta m⁻² s⁻¹, 20 ms length, for the maximum oxidation of PSI donor side. Computer-operated electro-pneumatic shutters (Fast-Est, Tartu, Estonia) mounted in the slide filter holders of the KL1500 light sources controlled WL and saturation pulses with flying times of 1.3 ms.

Far-red light was produced by a LED-based source LS-720A (Fast-Est, Tartu, Estonia), especially designed for the oxidative titration of PSI donors. About 60 LEDs (L-A720H, Applied Optotech, Cork, Ireland) were arranged in a battery focused on the entrance of the light guide (diameter 17 mm). A long-pass filter KS 19 (LOMO, St. Petersburg) cut off the short-wave tail of the LED emission. The emission of the LED battery was feedback-stabilized by a circuit involving a light-measuring diode (OPT 201, Texas Instruments, Dallas, Texas) and a LED-controlling amplifier. The intensity of FRL was set manually, but the on/off switchings were computer-controlled. The PFD of the FRL was measured with the help of the spectroradiometer PC-2000 (Ocean Optics, Dunedin, FL), calibrated against a standard lamp. White actinic light was measured with a LI-190SB sensor (LiCor, Lincoln, NE).

Leaf absorption coefficients for WL and FRL were measured with an integrating sphere and the spectroradiometer PC-2000. Rates of photon absorption in WL and FRL were calculated by integrating the product of the absorption coefficient of the leaf times PFD over the spectrum. The absorption coefficient of FRL by different leaves was in the range from 44 to 47% for the low-intensity FRL of 50 μmol quanta m⁻² s⁻¹, but decreased to 23% due to the long-shifted emission wavelength for FRL of 400 μmol quanta m⁻² s⁻¹. The rate of PSII excitation was measured from the steady-state O₂ evolution under FRL (Oja et al. 2003; Pettai et al. 2005). Total absorbed FRL minus that exciting PSII was considered exciting PSI.

Leaf transmission at 810 nm

A single-beam spectrophotometer FS-810A (Fast-Est, Tartu, Estonia) was especially designed for the measurement of leaf transmittance changes at 810 nm. An 810 nm LED (type ELD 810-525, Roithner Lasertechnique, Vienna, Austria) was driven with rectangular pulses of 5 μs length at 100 kHz frequency by a quartz-stabilized generator. The measuring beam was transmitted to the front of the leaf chamber via a plastic fiber guide that illuminated 1 cm² measurement area. Another bundle of fibers collected the transmitted 810 nm radiation from the rear of the leaf chamber and guided to the sensor PIN diode (type S3590-01, Hamamatsu, Japan), protected by a Schott RG 780 filter against Chl fluorescence and other non-modulated radiation. The photocurrent of the PIN diode was amplified by a current-feedback controlled primary amplifier, holding a constant potential on the PIN diode independent of modulated or non-modulated irradiation. The 100 kHz modulated output signal of the primary amplifier was filtered by a LC filter, further amplified, and rectified to a DC signal of >2 V. A constant offset of 2 V was applied and the difference DC signal exceeding the offset was further amplified by a factor of 100, for convenient computer-recording by the 12 bit A/D converter board. The precise stabilization of the 100 kHz generator and the current-feedback controlled primary amplifier ensured a stable signal with the noise level of 0.1% of the typical full redox signal, sufficient for successful single-trace recording of most of the P700 reduction transients. The output signal of FS 810A was insensitive to the non-modulated saturation pulses produced by the filtered KL 1500 light source.

Operation of the system and data recording

The system was operated and data were recorded using an A/D converter board ADIO 1600 (Kontron, San Diego, CA) and a system-operation and data-recording program RECO (Fast-Est, Tartu, Estonia). The program mimics both a multi-channel chart recorder for slow processes and a recording oscilloscope for fast processes, and also operates gas mixers, light sources, and shutters. A preloaded sequence of textual meta-language commands executed the computer-actuated measurement routine.

Results

Improved method of oxidative titration of PSI donors

The oxidative titration of PSI donors is based on the recording of the 810 nm signal during the process of oxidation of the PSI donor side electron carriers under FRL (Oja et al. 2003). During the redox equilibrium between the high-potential carriers, Cyt f is being oxidized first, followed by PC and, finally, P700. The 810 nm signal change per e⁻ is largely variable over the oxidation span. The signal is the most sensitive when P700 is being oxidized, less sensitive when PC is being oxidized and no signal is generated when only Cyt f is being oxidized. The transition between the oxidized substrates is the more mixed and smoother the smaller is the equilibrium constant between the electron carriers (ΔE _m). In order to consider this, the time course of the 810 nm signal during the oxidation by FRL was mathematically modeled and the model parameters were varied to fit the modeled trace to the measured data. The obtained best-fit model parameters were the pools of P700, PC and Cyt f and the equilibrium constants between these compounds. This approach allowed us to non-destructively determine the pools of electron carriers at the PSI donor side.

The approach of Oja et al. (2003) contained an implicit assumption that electrons transferred to the acceptor side did not return to the donor side via the cyclic pathway. However, already our first experiments indicated that the quantum efficiency of PSI tended to decrease during the process of oxidation, indicating either PSI acceptor side closure or cyclic electron transport. In order to minimize this effect, the titration curve was linked together from short sections, but it was not clear that such an integrative approach deprived the method of the influence of the cyclic electron flow. In this work we further developed the oxidative titration procedure, measuring the rate of electron inflow in the dark immediately before the measurement of the net rate of electron outflow under FRL. The difference of the two measurements represented the gross (true) rate of electron transport through PSI under FRL. This approach of “differential oxidative titration” allowed us to carry out more reliable conversion of the 810 nm signal into the number of electrons in the PSI high-potential donors. Afterwards the conversion was applied to measure the rate of electron cycling around PSI under different experimental conditions.

During the automated experimental procedure, the high-potential e⁻ carriers were first pre-oxidized under steady-state FRL, as previously described (Oja et al. 2003). Then FRL was turned off and the high-potential carriers were allowed to reduce to some extent, before FRL was turned on again (Fig. 1). Differently from the earlier routine, the rate of decrease (slope toward reduction) of the 810 nm signal immediately before the application of FRL and the rate of increase (slope toward oxidation) of the signal immediately after the application of FRL were measured. After a certain time interval the FRL was turned off and the slopes of the recordings before and after the cessation of the illumination were also measured. The difference of the slopes of the signal after and before the break-point reflected the true (gross) rate of electron transport by FRL. The maximum signal obtained in the beginning of a 20 ms saturation pulse of white light applied on top of FRL was considered to indicate the complete oxidation. The level the signal approached in the dark after the saturation pulse (after the relaxation of Fd⁻ generated by the pulse) indicated the complete reduction. These two boundary levels were used to calculate the redox levels of P700, PC and Cyt f at any intermediate position of the signal.

Fig. 1

The automatic procedure of oxidative titration on a sunflower leaf. The routine was repeated in the following sequence: FRL on for 50 s (stabilization under FRL, not shown), FRL off for increasing periods from 0 s to 50 s (shown traces begin at the end of each dark period), FRL on at the sampled data point # 55, FRL off at point # 105, saturation pulse of 10000 μmol quanta m⁻² s⁻¹ on at point # 170, saturation pulse off at point # 190. Sampling interval was 2 ms (except 1 ms for the curve of 0 s darkness). The dark level of the signal obtained after each saturation pulse (after the relaxation of the Fd⁻ signal) was set to zero (full reduction). Full oxidation was assumed to be at the peak of the saturation pulse applied without the dark exposure

The titration curve was computer-modeled assuming redox equilibrium between the PSI donor side carriers Cyt f, PC, and P700 (Fig. 2). The Rieske FeS + Cyt f were considered as one pool denoted Cyt f here. In the dark the rate constants k ₃ for P700 oxidation and k ₄ for e⁻ arrival from PSII were zero and the system became reduced through reactions 5, a still not well understood pathway causing the spontaneous reduction of the high-potential inter-photosystem carriers in the dark. This rate is represented by the slope of the transmission signal in the dark, before FRL was admitted at point # 55 (Fig. 1). After FRL was admitted, electrons left the system via P700 at a rate determined by PAD of FRL and the reduction state of P700 (rate constant k ₃). Simultaneously with the oxidation, the reduction continued from stromal reductants at the rate detected immediately before FRL. In addition to this, electrons were also generated by PSII under FRL at the rate k ₄, and soon the flux of CET joined the stromal reduction pathway k ₅. For correct interpretation, the slope of the signal had to be measured before these additional fluxes arrived at the PSI donor side (within about 5 ms). The removal of every e⁻ from P700 by FRL was followed by immediate redistribution of the remaining e⁻ according to the equilibrium conditions (equilibrium constants K _C and K _P) between the high-potential donors of PSI. The Rieske FeS and Cyt f became oxidized first due to their more negative E _m values, then PC, and finally P700. Two compounds in this chain, P700⁺ and PC⁺ generated the 810 nm signal, but the optical efficiency (cross-section) of P700⁺ was considered to be 6.4 times the cross-section of PC⁺ (Kirchhoff et al. 2004).

Fig. 2

The structure of the mathematical model. Cytochrome f plus probably the Rieske FeS protein (together denoted Cyt f), plastocyanin (PC), and P700 are assumed to be in redox equilibrium with constants K _C and K _P. The equilibrated system is reduced by stromal reductants in the dark (rate k ₅). Under FRL the system is reduced by stromal reductants + CET and by PSII (rate k ₄) and is oxidized by PSI (rate k ₃). The 810 nm signal is generated by P700⁺ (stronger) and PC⁺ (weaker)

The model was solved like in Oja et al. (2003), but not with respect to the 810 nm signal, but with respect to the derivative (slope) of the signal (Appendix). The changes of the slope were measured at different redox states of the high-potential donor system and presented as data points (Fig. 3). The model-calculated slopes were then fitted to experimental data points by varying the pool sizes of the carriers and their equilibrium constants.

Fig. 3

The rate of response of the 810 nm signal to FRL on/off during the titration routine (Fig. 1). Empty diamonds, FRL on; filled diamonds, FRL off; squares, steady state under FRL; circle, maximum oxidation. Solid line, the model-calculated signal response; dotted line (right ordinate), calculated 810 nm signal change per e⁻. For this leaf the best-fit values were P700 = 1.41 μmol m⁻², PC = 3.3 per P700; Cyt f = 0.8 per P700, K _C = 2.2; K _P = 15

To the right from the maximum the rate of oxidation by FRL decreased due to the increased oxidation of P700. Data points measured under steady-state FRL remained apart from the complete oxidation, because the spontaneous reduction from stromal reductants, the FRL-induced PSII electron flow and CET reduced the PSI donors simultaneously with the oxidation by FRL. The complete oxidation was achieved in the beginning of saturating pulse of white light, superimposed over FRL. The left (reduced) side of the titration curve approached the ordinate axis at a certain rate of the oxidation signal (about 1200 AD s⁻¹ in Fig. 3). This reflected the immediate response of the oxidation signal to the admitting of FRL. If there were no Cyt f in equilibrium with PC, the first electrons were removed mainly from PC and the initial rate of the oxidation signal would have represented the strength of the optical signal per e⁻ removed from PC. Actually the initial rate of the oxidation signal was different in different leaves, indicating that some e⁻ were removed from a compound equilibrated with PC, but not generating the 810 nm signal. The variable initial rate of oxidation of PC immediately after FRL was admitted after longer dark exposures was the experimental basis for the evaluation of the Cyt f pool equilibrated with PC. However, there was an interfering phenomenon that complicated the modeling of the Cyt f pool: after too long dark exposures (but just necessary for the complete reduction of PC and Cyt f), the initial re-oxidation rate decreased, indicating an apparent decrease of the PSI electron transport efficiency (e.g., the two data points right on the ordinate axis of Fig. 3). Alternatively, similar apparent slow-down of the oxidation was expected if plastoquinone (PQ), having considerably lower E _m than Cyt f, became partially reduced during the extended dark exposures. Therefore, the Cyt f pool could be more or less reliably modeled only if the length of the dark exposure was a compromise choice, such that Cyt f became reduced, but not yet PQ.

This oxidative titration procedure was carried out for every leaf and the 810 nm signal change per e⁻ (or actually the reciprocal, e⁻ in PSI donors per bit of signal change) was applied to calculate the rate of electron arrival at the PSI donors under different experimental conditions. In the example leaf of Fig. 3 the total of 5.1 e⁻/PSI accumulated in the high-potential donors in the dark (1 e⁻/PSI in P700, 3.3 e⁻ PSI in PC and 0.8 e⁻/PSI in Cyt f, PSI density was 1.41 μmol m⁻²).

Cyclic flow or inactivation of PSI electron transport?

In experiments of Oja et al. (2003) self-consistent titration curves could be produced only if the dark exposure time was no longer than about 1 min. After longer dark exposures the following oxidation under FRL was slower than anticipated. After the two longest dark exposures the rate of oxidation under FRL decreased also in the experiment of Fig. 3. Below we further investigated the time dependence and the possible mechanism of such an apparent dark inactivation of PSI electron transport.

In the experiment of Fig. 4 the leaf was exposed under FRL and different dark intervals were applied, after which the FRL was turned on again. Dark exposures of shorter than 15 s were insufficient to completely reduce the PSI donors containing altogether 5.2 e⁻/PSI in this leaf (1 e⁻ in P700, 3.3 e⁻/P700 in PC, and 0.9 e⁻/P700 in Cyt f). After these short dark exposures FRL caused fast re-oxidation of the PSI donors to the same steady-state level of about 0.2 e⁻/PSI as before the dark exposure (only traces corresponding to 15 s and longer dark exposures are shown in Fig. 4). As soon as the length of the dark exposure exceeded 75 s the following oxidation curve became multiphasic. There was a short initial faster phase during which about 0.5 e⁻/PSI were transported, then the rate slowed down, accelerated again after 0.4 s under FRL, but still proceeded more slowly, and the high-potential carriers remained more reduced than after the shorter dark exposures. For example, after the longest dark exposure of 500 s more than 3 e⁻/PSI remained in the donor side carriers, instead of 0.2 e⁻/PSI after the short dark exposures. In order to reveal the reasons why the oxidation by FRL became so inefficient, FRL was turned off at 0.95 s. The initial slopes of the following dark re-reduction traces represented the rate of electron inflow into the PSI donors. This re-reduction rate was 1.6 (e⁻/PSI) s⁻¹ after the short dark exposures, but increased to about 10 (e⁻/PSI) s⁻¹ after long dark exposures. This indicated cycling of electrons around PSI was the immediate reason of the decreased oxidative efficiency of FRL.

Fig. 4

Time courses of oxidation of PSI donors under the incident FRL of 50 μmol quanta m⁻² s⁻¹ after dark exposures of 15 (filled squares), 75 (empty diamonds) 150 (filled circles), 250 (empty triangles), and 500 s (crosses). FRL was turned on at time = 0 and off at 0.95 s. The initial rates of the post-FRL re-reduction of PSI donors are indicated at the curves (e⁻/PSI) s⁻¹

In order to obtain a more detailed picture about the development of the cyclic electron flow during the oxidation process, the experiment of Fig. 4 was repeated with the dark exposure of 150 s, but FRL was interrupted after different time intervals (Fig. 5). The measurement of the post-FRL re-reduction rate was difficult immediately after FRL was turned on, because there was too little oxidation, but after a few e⁻/PSI were transferred the measurements indicated the re-reduction rate of 4 (e⁻/PSI) s⁻¹. The re-reduction rate steeply increased after more than 3 e⁻/PSI were transferred to the acceptor side, approaching the rate of 9 (e⁻/PSI) s⁻¹. Then the oxidation by FRL and re-reduction by CET became equal.

Fig. 5

Re-oxidation of PSI donors by FRL and induction of CET after a dark exposure of 150 s. The experiment was repeated interrupting FRL after different time intervals under FRL and the following re-reduction rate was measured (cyclic rate, left ordinate). The rate of oxidation of PSI donors under FRL was calculated from the model (right ordinate)

In the above experiments we investigated the cycling when electrons pre-accumulated on the donor side were transferred to the acceptor side of PSI. In the following experiment we investigated CET during the transport of electrons generated by PSII. A 5 μs xenon flash was applied on the background of FRL. The flash first caused complete oxidation of PSI donor side, then electrons arriving from PSII caused a temporary PSI donor side reduction transient, but were removed by FRL soon (Fig. 6). Still, there was a noticeable slow tail in the re-oxidation trace, indicating the slow-down of the net e⁻ transport through PSI.

Fig. 6

Transient reduction of PSI donor side carriers after a single-turnover xenon flash. At time of zero a 5 μs xenon flash was applied on top of FRL, causing the following temporary reduction of PSI donors. The experiment was repeated interrupting FRL at different phases of the post-flash transient—without the flash (curve a); immediately after the flash (b) and after longer time intervals as seen from the traces. Electron transport rates (insert) were calculated from the initial slopes of the post-FRL curves (empty squares), or as the difference of the slopes after and before FRL was turned off (empty diamonds), or as the running slope of the trace b. PSI density was 1.17 μmol m⁻², absorbed FRL was 19.3 μmol quanta m⁻² s⁻¹

In order to visualize electron cycling the experiment of Fig. 6 was repeated interrupting FRL at different phases of the re-oxidation process. The slopes of the post-FRL re-reduction traces represented the rate of electron arrival at the PSI donor side. When FRL was interrupted without the flash, the initial re-reduction rate was 2.4 (e⁻/PSI) s⁻¹, as seen from the initial slope of the trace a. From this rate, 1.4 (e⁻/PSI) s⁻¹ were arriving from PSII, as detected from O₂ evolution rate under FRL (data not shown, see Pettai et al. 2005). Under steady-state FRL the cyclic rate was thus about 1 (e⁻/PSI) s⁻¹. When FRL was turned off immediately after the flash (trace b), the initial rate of arrival of electrons generated by PSII was 60 e⁻ s⁻¹ PSI⁻¹. All these PSII electrons were transferred through Cyt b₆f within about 100 ms (Fig. 6, insert, the initial rate of curve b outscaled at 60 (e⁻/PSI) s⁻¹).

When FRL was interrupted later at different phases of the following transient, the initial slope showed the rate of electron inflow into the PSI donors at that time moment. Shortly after the flash the post-FRL electron arrival rates rather well matched the rate derived from the running slope of the post-flash transient, showing that the linear e⁻ flow dominated, but a slowly decaying component of electron inflow of about 4 (e⁻/PSI) s⁻¹ remained after PQ had donated all its electrons. The cyclic rate had rapidly increased by about three times as a result of the flash-induced electron transport. This cyclic rate slowly decreased, finally approaching the steady-state level under FRL (dotted line, Fig. 6 insert). The insert of Fig. 6 also displays the total (gross) electron transport rate of PSI calculated as the difference of the post-FRL re-reduction slope and the downward slope of the transient before FRL was interrupted. This electron transport rate was compared with the rate of FRL quanta hitting the PSI centers with reduced P700 (Fig. 9) and the quantum efficiency of PSI with reduced P700 was found to be very close to unity (about 0.95 in this experiment).

Dark inactivation of ferredoxin-NADP reductase

The experiments of Figs. 4 and 5 showed that cycling of electrons back to the oxidized donor side carriers was a significant component of PSI electron transport under FRL. The cycling of electrons increased after the leaf was kept in the dark, but only temporarily; after some time FRL still oxidized the PSI donors. In order to investigate the time course how the dark-inactivated PSI electron transport was reactivated, the FRL exposure time was extended and the incident FRL intensity was increased to 400 μmol quanta m⁻² s⁻¹. Under this high PFD of FRL the initial oxidation of the PSI donor side rapidly reversed to increasing reduction, peaking after about 1.3 s from admitting FRL (Fig. 7). This temporary reduction was retrieved after 3 s. Following the longest dark exposures of 200, 250, and 500 s another temporary reduction peak was observed after about 15–20 s under FRL (not shown). This secondary reduction peak was lower than the primary peak, but still significantly distinguishable from the steady-state oxidation level under the intense FRL. The transient of CET was investigated in another similar experiment, where FRL of 50 μmol quanta m⁻² s⁻¹ was interrupted at different phases of the oxidation process. The detected rate of the cyclic electron flow increased during the reduction peak and decreased as soon as the steady-state level was established (Fig. 8).

Fig. 7

Time courses of oxidation of PSI donors under the strong FRL of 400 μmol quanta m⁻² s⁻¹ after the dark exposures of 15 (filled diamonds), 50 (filled squares), 100 (filled triangles), 200 (crosses), 250 (empty squares), and 500 s (empty circles)

Fig. 8

The experiment of Fig. 7 (dark exposure 150 s, FRL 50 μmol quanta m⁻² s⁻¹, solid line, left ordinate) was repeated interrupting the oxidation process after different time intervals in order to measure the post-FRL re-reduction rate at different phases of the oxidation process (filled diamonds, right ordinate)

The number of electrons accumulated on the PSI acceptor side before the first reduction maximum indicates the possible site of inhibition of electron flow. First, four e⁻/PSI were transferred by FRL from the donor side and accommodated in the acceptor side carriers without significant problems. Electrons generated by PSII continued to arrive under FRL, causing the increasing reduction of the donor and acceptor side pools. It is known that the FRL of 720 nm can excite PSII with the efficiency of about 6%, as indicated by O₂ evolution under FRL (Pettai et al. 2005). In the experiment of Fig. 7, 23% of the incident FRL of 400 μmol quanta m⁻² s⁻¹ was absorbed, causing the total excitation of PSI and PSII at the rate of 83 μmol m⁻² s⁻¹. Most of this excited PSI, but PSII was still excited at the rate of about 5 μmol quanta m⁻² s⁻¹. Assuming also some (fluorescence-indicated) losses, the PSII electron transport rate was 4 μmol m⁻² s⁻¹ under the FRL of 400 μmol m⁻² s⁻¹. Considering the titrated PSI density of 1.26 μmol m⁻², 3.2 (e⁻/PSI) s⁻¹ arrived from PSII. The steepest slope in Fig. 7 indicated 2 (e⁻/PSI) s⁻¹ accumulated on the donor side during the phase of increasing reduction at FRL of 400 μmol quanta m⁻² s⁻¹ (at lower FRL intensities this slope was proportionally smaller). Correspondingly, 3.2 − 2 = 1.2 (e⁻/PSI) s⁻¹ accumulated in the acceptor side carriers. The estimate of the total number of electrons transported from the donor to the acceptor side before the reduction maximum in Fig. 7 was thus 4 e⁻/PSI from the dark-reduced donors plus about 1.8 e⁻/PSI from PSII during 1.5 s under FRL, together about 6 e⁻ /PSI. A similar estimate suggests about 48 e⁻/PSI were released by PSII and transferred through PSI before the secondary reduction peak at 15 s from the admitting of the intense FRL. These estimates suggest, the first reduction peak was caused by the dark inhibition of an enzyme very close to PSI, e.g., ferredoxin-NADP reductase (FNR).

PSI electron transport efficiency

The above experiments raised a question about the actual quantum efficiency of PSI electron transport, considering the relatively fast cycling of electrons. In our experiments the intensity and the absorptance by the leaf of FRL were measured, allowing us to quantify the quantum yield of PSI electron transport. The reduced fraction of P700 was calculated from the redox equilibrium model at any signal value between the reduced and oxidized borders. The abscissa of Fig. 9 represents the number of FRL quanta s⁻¹ absorbed in PSI with reduced P700. The positioning of the data points on the abscissa is proportional to the reduction states of P700 established during the experiments (absorbed FRL intensity was about constant in different experiments). The ordinate represents electron flow rate per PSI, calculated as the sum of the rate of oxidation plus the rate of reduction (mainly CET). Data points measured during the standard titration procedure indicate the quantum yield of unity. Data points measured turning FRL off in the presence of relatively fast cyclic flow indicate quantum yields reduced by up to 30%. The data of Fig. 9 confirm that the quantum yield of PSI was close to unity during the oxidative titration as well as during the transport of single-flash electrons, where the cyclic electron flow did not approach its maximum rate. The PSI quantum yield was lower in experiments of Figs. 5 and 8, where the cyclic flow considerably accelerated as a result of dark inactivation of FNR.

Fig. 9

Quantum efficiency of PSI with reduced P700. The rate of oxidation of PSI donors was plotted against the excitation rate of PSI with reduced P700 in different experiments; empty symbols; FRL on, filled symbols, FRL off. Diamonds, the standard titration with short dark intervals (Fig. 1); squares, oxidation after long dark intervals (Fig. 4); triangles, from the experiment of Fig. 5; circles, from the experiment of Fig. 8; crosses, from the experiment of Fig. 6. Solid line indicates the quantum efficiency of 1.0

Discussion

Oxidative titration of PSI donors

The assumption of redox equilibrium between the high-potential PSI donor side electron carriers has been questioned for the situations of fast electron transport (Kirchhoff et al. 2004), but during the slow electron transport under FRL, rather, intraleaf inhomogeneities are more important (Joliot and Joliot 2005). In a thylakoid of a certain chloroplast the PSI electron transport chain is still close to the redox equilibrium, but there are diffusionally disconnected compartments with different PC/P700 ratio in the leaf. During the slow oxidative titration the PC/P700 redox equilibrium could advance with different speed in compartments having different PC/P700 ratio, as a result of what the measured 810 nm signal could be mixed and smoothed over the transition phase between PC and P700, but clean at the extreme redox situations. In order to minimize the inhomogeneity, our procedure started with the slow dark re-reduction, free at least of inhomogeneities caused by the gradient of FRL in the leaf. Subsequently under FRL only the first about 5 ms were considered for the determination of the oxidation speed, during which the developing oxidation was the least inhomogeneous. Nevertheless, the best-fit equilibrium constant between P700 and PC obtained from the oxidative titration was lower than 40–100 derived from the standard E _m values of these electron carriers. In (Oja et al. 2003) the average equilibrium constant of 25 satisfied all experiments with leaves of different species, including sunflower. In this work the best-fit K _P value between PC and P700 was 15 for sunflower, still indicating inhomogeneities. Still, though there were inhomogeneities, the apparent equilibrium constant averaged and to a large extent took into account these inhomogeneities when the model was reversed later to calculate the PSI donor side electron pools under different PC/P700 transition situations.

The oxidative titration curve (Fig. 3) was calculated on an assumption of the quantum yield of 1.0. Assuming a lower quantum yield (e.g., 0.95, Hiyama 1985) would not change the titration curve, except that the best-fit absolute PSI density would decrease proportionally and the calculated electron signal strength would increase reciprocally proportionally to the assumed PSI yield. Thus, the oxidative titration by FRL is a relative method relating the pools of Cyt f and PC to P700, but the obtained absolute value of PSI density (μmol m⁻² of P700) is dependent on the quantum yield of PSI (as well as on the correct quantification of absorbed FRL). The reductive titration method, relating PSI density to the amount of oxygen, μmol m⁻², evolved per single-turnover flash, is an absolute method for the determination of the PSI density (Oja et al. 2004). Numerous comparative measurements in our laboratory have shown the ratio of PSI densities obtained from reductive versus oxidative titrations equals to 1.0 (R ² = 0.87) in sunflower, tobacco, and potato leaves (H. Eichelmann, unpublished data). Thus 1.0 ± 0.15 is the estimate of the absolute quantum yield of PSI in leaves during the oxidative titration by FRL.

Cyclic electron flow around PSI under FRL

First indication of the cyclic electron transport under FRL was obtained from experiments where leaves were illuminated with single- or multiple turnover flashes given on top of FRL and the following transient changes in the 810 nm signal indicated too many FRL quanta were needed to transfer the generated electrons through PSI (Asada et al. 1992, 1993). The slow rate of oxidation of the PSI donor side by FRL has been assumed showing CET rather than PSI acceptor side closure also in recent works (Joliot and Joliot 2005; Breyton et al. 2006). Using photo-acoustic spectroscopy it has been shown that cyclic electron flow can be activated under FRL in C₃ plants, but special poising conditions are necessary, e.g., such created by low O₂ concentration (Joët et al. 2002).

Our present experiments confirmed the presence of fast CET under FRL. During the fast cycling the quantum yield of PSI generally remained close to unity, but partial acceptor side closure was detectable in experiments where FNR was inactivated after prolonged darkness (Fig. 9). The observed maximum rate of electron cycling of 8–10 (e⁻/PSI) s⁻¹ was not limited by the capacity of the cyclic pathway, but by the excitation rate of PSI under FRL, because the quantum yield of PSI electron transport remained at least at 0.7, but in several cases close to unity. Such a high quantum yield of PSI supporting CET has important implications. It shows that all PSI in the leaf were able to perform in cycle, questioning the model according to which PSI located in the stroma-exposed thylakoids are preferentially functioning in cycle, but PSI in the grana margins are carrying out linear electron transport (Albertsson 1995, 2001; Anderson et al. 1995). No characteristic dual-phase kinetics of oxidation (Joliot and Joliot 2005) were observed under FRL, to evidence the presence of two populations of PSI with different cyclic rates. The necessity to apply the apparent equilibrium constant of 15 between PC and P700 instead of 40–100 calculated from in vitro titrated redox potentials of these compounds, was the only evidence of inhomogeneity that could reflect the differences of stroma and grana thylakoids. The induction of CET was very fast, e.g., its rate increased by three times within 100 ms after the xenon flash (Fig. 6). This fast reaction showed CET was activated not around PSI in the stroma thylakoids, placed far from PSII, but around PSI that could rapidly receive electrons from PSII—in those placed at the grana margins. The result is in good accordance with the observations of Breyton et al. (2006), showing that CET is redox-controlled in all PSI in the leaf. No structural changes—such as the re-complexing of FNR from PSI to Cyt b₆f—are necessary to induce CET.

Dark inactivation of FNR and cyclic electron flow

When strong FRL was turned on after longer dark exposures, about six electrons were transferred from the donor to the acceptor side before further electron transport was blocked (Fig. 7). In spinach chloroplasts the ratios of Fd to FNR to PC approximated 5:3:4 per P700, as determined by immuno-electrophoresis (Böhme 1978). Thus there are about eight reducible sites in Fd + FNR, in good accordance with the hypothesis that the first partially dark-inhibited step was ferredoxin-NADP reductase (EC 1.18.1.2). A similar conclusion has been drawn earlier from the observation of a similar three-phase curve of Cyt f oxidation (Ruhle et al. 1987) and from a similar transient in the 820 nm signal during the induction of photosynthesis (Harbinson and Hedley 1993). We assume that the second, slower reduction peak corresponded to inactivated GAP-dehydrogenase, considering the pre-accumulated pool of 48 e⁻ or 24 NADPH/PSI. Measurements have shown about 10 NADP/NADPH per PSI (Gerst et al. 1994), but the coincidence is satisfactory considering different plants and methodical difficulties of these measurements. The mechanism of the dark inactivation of FNR is not related to its movements, because decomplexing of FNR from PSI was not observed (Breyton et al. 2006), but presumably is a kind of redox control of the activity of the enzyme.

Appendix

Redox equilibrium-based modeling of the 810 nm signal

For convenience, we defined the following denotations:

$$ \varepsilon _{{{\text{PC}}}} \; = \;\frac{{{\text{extP}}700}} {{{\text{extPC}}}}\;\;\;\;{\text{(ratio of extinction coefficients for the two compounds),}} $$

$$ g_{{{\text{PC}}}} \; = \;\frac{{{\text{P}}700_{{{\text{ox}}}} /{\text{P}}700_{{{\text{red}}}} }} {{{\text{PC}}_{{{\text{ox}}}} /{\text{PC}}_{{{\text{red}}}} }}\;\;\;\;( = 1/K_{{\text{P}}} {\text{ in Fig}}{\text{. }}2), $$

$$ g_{{\text{f}}} \; = \;\frac{{{\text{P}}700_{{{\text{ox}}}} /{\text{P}}700_{{{\text{red}}}} }} {{{\text{Cyt f}}_{{{\text{ox}}}} /{\text{Cyt f}}_{{{\text{red}}}} }}\;\;\;\;( = 1/(K_{{\text{P}}} \; \cdot \;K_{{\text{C}}} ){\text{ in Fig}}{\text{. }}2), $$

where g denote the reciprocal redox equilibrium constants between the specified compound and P700. The system of the redox-equilibrated compounds is resolved with the help of an assistant variable x, reflecting the oxidation state of P700:

$$ x\; = \;\frac{{P700_{{{\text{ox}}}} }} {{P700_{{{\text{red}}}} }}. $$

Since the optical signal is generated by oxidized fractions, the model is based on the oxidized fractions of the redox-equilibrated compounds. Using these denotations, the sum of all redox-equilibrated oxidized fractions is

$$ e_{x} \; = \;P700_{{\text{T}}} {\left( {\frac{x} {{1\; + \;x}}\; + \;{\text{PC}}\frac{x} {{g_{{{\text{PC}}}} \; + \;x}}\; + \;{\text{Cyt f}}\frac{x} {{g_{{\text{f}}} \; + \;x}}} \right)},\;{\text{$\upmu$mol}}\;{\text{missing}}\;{\text{e}}^{-} \,{\text{m}}^{{ - 2}} . $$

(1)

Here P700_T denotes the only involved total pool in μmol m⁻², other donor side electron carriers PC and Cyt f are expressed relative to P700_T. The 810 nm optical signal T is generated by oxidized P700 and PC as follows:

$$ T\; = \;{\text{extP}}700\; \cdot \;{\text{P}}700_{{\text{T}}} {\left( {\frac{x} {{1\; + \;x}}\; + \;\frac{{{\text{PC}}}} {{\varepsilon _{{{\text{PC}}}} }}\; \cdot \;\frac{x} {{g_{{{\text{PC}}}} \; + \;x}}} \right)}, $$

(2)

where extP700 is the absolute extinction coefficient of P700, the fractional transmittance change per μmol m⁻² of oxidized P700. The maximum signal is expressed from (2) by letting x → ∞, as follows

$$ T_{{\max }} \; = \;{\text{extP}}700\; \cdot \;{\text{P}}700_{{\text{T}}} {\left( {1\; + \;\frac{{{\text{PC}}}} {{\varepsilon _{{{\text{PC}}}} }}} \right)}. $$

(2a)

Under FRL the rate of oxidation of all equilibrated PSI donors is

$$ \frac{{de_{x} }} {{dt}}\; = \;\eta \; \cdot \;{\text{FRL}}\; \cdot \;{\text{P}}700\; = \;\frac{{{\text{FRL}}\; \cdot \;\eta }} {{1\; + \;x}},\;\;\upmu {\text{mol}}\;{\text{e}}^{ - } \;{\text{m}}^{{ - 2}} \,{\text{s}}^{{ - 1}} , $$

(3)

where η is the quantum efficiency of PSI electron transport under FRL (assumed to be unity), FRL is the flux density absorbed by PSI and P700 is now denoting the reduced fraction, $P700\; = \;1/(1\; + \;x)$.

Before we can express the speed of signal change versus time, we express the change versus the P700 oxidation parameter x by differentiating (1) and (2) with respect to x:

$$ \frac{{de_{x} }} {{dx}}\; = \;{\text{P}}700_{{\text{T}}} {\left[ {\frac{1} {{{\left( {1\;+\;x} \right)}^{2} }}\;+\;{\text{PC}}\frac{{g_{{{\text{PC}}}} }} {{{\left( {g_{{{\text{PC}}}} \;+\;x} \right)}^{2} }}\;+\;{\text{Cyt}}\;{\text{f}}\frac{{g_{{\text{f}}}}} {{{\left( {g_{{\text{f}}}\;+\;x} \right)}^{2} }}} \right]} $$

(4)

$$ \frac{{dT}} {{dx}}\; = \;{\text{extP}}700\; \cdot \;{\text{P}}700_{{\text{T}}} {\left[ {\frac{1} {{{\left( {1\; + \;x} \right)}^{2} }}\; + \;\frac{{{\text{PC}}}} {{\varepsilon _{{{\text{PC}}}} }}\; \cdot \;\frac{{g_{{{\text{PC}}}} }} {{{\left( {g_{{{\text{PC}}}} \; \cdot \;{\text{PC}}\; + \;x} \right)}^{2} }}} \right]} $$

(5)

The change of the signal per change of electrons in the equilibrated compounds is expressed as

$$ \frac{{dT}} {{de_{x} }}\; = \;\frac{{dT/dx}} {{de_{x} /dx}} $$

(6)

where the numerator is from Eq. 5 and denominator from Eq. 4. Considering that

$$ \frac{{dT}} {{dt}}\; = \;\frac{{dT}} {{de_{x} }}\; \cdot \;\frac{{de_{x} }} {{dt}}, $$

(7)

we obtain, considering Eq. 3

$$ \frac{{dT}} {{dt}}\; = \;{\text{extP}}700\; \cdot \;{\text{FRL}}\; \cdot \;\eta \frac{{\frac{1} {{{\left( {1\; + \;x} \right)}^{2} }}\; + \;\frac{{{\text{PC}}}} {{\varepsilon _{{{\text{PC}}}} }}\; \cdot \;\frac{{g_{{{\text{PC}}}} }} {{{\left( {g_{{{\text{PC}}}} \; + \;x} \right)}^{2} }}}} {{{\left( {1\; + \;x} \right)}{\left( {\frac{1} {{{\left( {1\; + \;x} \right)}^{2} }}\; + \;{\text{PC}}\frac{{g_{{{\text{PC}}}} }} {{{\left( {g_{{{\text{PC}}}} \; + \;x} \right)}^{2} }}\; + \;{\text{Cyt f}}\frac{{g_{{\text{f}}} }} {{{\left( {g_{{\text{f}}} \; + \;x} \right)}^{2} }}} \right)}}}. $$

(8)

In Eq. 8 extP700 is presented as a fraction of transmittance change per μmol m⁻² of P700, FRL is the absorbed flux density, μmol quanta m⁻² s⁻¹ and η is dimensionless net efficiency of PSI electron transport. For the oxidative titration, x was varied and the corresponding abscissa value of Fig. 3 was calculated as the signal normalized to the maximum signal:

$$ \frac{T} {{T_{{\max }} }}\; = \;\frac{{\frac{x} {{1\; + \;x}} + \frac{{{\text{PC}}}} {{\varepsilon _{{{\text{PC}}}} }}\; \cdot \;\frac{x} {{g_{{{\text{PC}}}} \; + \;x}}}} {{1\; + \;\frac{{{\text{PC}}}} {{\varepsilon _{{{\text{PC}}}} }}}} $$

(9)

The corresponding ordinate was calculated from Eq. 8. The whole modeled curve was obtained as a series of pairs obtained from Eqs. 8 and 9 by varying x from 0 to ∞. Model parameters P700, PC and Cyt f, as well as equilibrium constants g _PC and g _f were varied until the calculated curve best fitted to the data points. Later, when the corresponding number of electrons was calculated from a certain optical signal, the value of x corresponding to the particular measured T/T _max value was expressed from Eq. 9 and the corresponding number of electrons in PSI donors was calculated from Eq. 1, separately in Cyt f, PC and P700, as well as their sum.

The solution of Eq. 9 with respect to x results in a quadratic equation

$$ x\; = \;\frac{{ - B\; - \;{\sqrt {B^{2} \; - \;4AC} }}} {{2A}}, $$

(10)

where

$$ A\;=\;a\;- \;1\; - \;\frac{{{\text{PC}}}} {{\varepsilon_{{{\text{PC}}}} }};\;\;B\;=\;a\; + \;a\;\cdot \;g_{{{\text{PC}}}}\;-\;g_{{{\text{PC}}}}\;-\;\frac{{{\text{PC}}}} {{\varepsilon _{{{\text{PC}}}} }};\;\;C\;=\;a\;\cdot\;g_{{{\text{PC}}}} , $$

where $a = T/T_{{\max }} \; \cdot \;(1\; + \;(PC/\varepsilon _{{{\text{PC}}}})).$