Introduction

Oxygen is evolved by splitting water as a by-product of the linear electron flow (LEF) from PSII via cytochrome b 6 f and PSI to NADP+, by which light energy is converted to chemical energy in the form of NADPH and ATP. The chemical energy is used in the Calvin–Benson cycle to assimilate CO2. However, for example, when CO2 is in short supply, NADPH production exceeds its consumption in the Calvin–Benson cycle. In this situation, electrons drain from PSI to O2 in a process known as the Mehler reaction or pseudocyclic electron flow (Mehler 1951; Miyake 2010) which is the first step in the so-called water–water cycle (Asada 2000). In addition, O2 is also reduced by photorespiration and chlororespiration (recently reviewed by Bauwe et al. 2010). While perhaps viewed as wasteful processes, these reactions may help protect the photosynthetic apparatus from damage due to strong light; they serve as an electron sink under environmental stress conditions such as drought and high temperatures (Ort and Baker 2002). They also balance the ATP/NADPH ratio (Rumeau et al. 2007). Understanding O2-dependent electron flow capacity and its competition with other protective alternative electron flows is of great interest because of its relation to photosynthetic regulatory mechanisms (Kono et al. 2014; Allahverdiyeva et al. 2015; Yamamoto et al. 2016).

However, it is not easy to measure the activity associated with O2-related reactions. Membrane inlet mass spectrometry and O2 stable isotopes have been used to quantify them. These techniques identify (i) O2 uptake in photorespiration plus the Mehler reactions and (ii) O2 evolution in the oxygen-evolving complex in photosystem II (PSII) (Ruuska et al. 2000; Beckmann et al. 2009). Using membrane inlet mass spectrometry, Badger et al. (2000) convincingly demonstrated that O2-dependent electron flow is greater in algae and cyanobacteria than in the chloroplasts of higher plants (such as tobacco). On the other hand, Schreiber et al. (1995) used simple chlorophyll (Chl) fluorescence method to show that a 1-s illumination with actinic light (450 µE m−2 s−1) induces O2-dependent electron flow in Chlamydomonas and Synechocystis. Evidence for this induction was the fast dark relaxation of Chl fluorescence. O2-dependent electron flow, however, was not induced in spinach chloroplasts.

Recently, Shirao et al. (2013) reported that the measurement routinely used to estimate the widely used fluorescence parameter Fv/Fm (maximal photochemical efficiency of PSII) can be easily applied to characterize the O2-dependent electron flow capacity. This was based on the established relationship between the halftime of Chl fluorescence decay after 1-s saturating light pulse (SP) and the O2-dependent electron flow capacity directly measured by mass spectrometry (Shirao et al. 2013). The O2-dependent electron flow capacity was much higher in gymnosperms than in angiosperms in a survey of 101 outdoor plant species (Shirao et al. 2013). However, there are 295,383 described angiosperm species (~95.8% of all higher plants), whereas there are only about 11,850 fern species (3.8%) and 1,079 gymnosperm species (0.35%) (Christenhusz and Byng 2016). Therefore, previous work on gymnosperms reported on a trait found in only a small proportion of higher plants. To understand alternative electron flows in higher plants better, it is important to know why the O2-dependent electron flow capacity is low in angiosperms (Shirao et al. 2013), the biggest constituent of higher plants. One plausible reason could be the absence of Flavodiiron proteins (Flv) in angiosperms, while they are conserved in gymnosperms during evolution, as in cyanobacteria (Peltier et al. 2010; Allahverdiyeva et al. 2015; Yamamoto et al. 2016). Moreover, cyanobacteria and gymnosperms have higher O2-reduction capacity (Badger et al. 2000; Shirao et al. 2013) and Flv mediate the O2-reduction to water (Mehler-like reactions) in cyanobacteria (Helman et al. 2003; Allahverdiyeva et al. 2011, 2013). Further, the lower capacity in angiosperms may be an evolutionarily determined consequence, because oxygen can cause oxidative damage to living cells and photosynthetic organisms have developed antioxidant systems to avoid adverse effects of oxygen from the era of ancestral Oxyphotobacteria, the first organism with oxygenic photosynthesis in Earth’s history (Christenhusz and Byng 2016). Therefore, it is important to determine whether gymnosperms and angiosperms differ in their capacities for other alternative electron flows (such as CEF-PSI) in addition to O2-dependent electron flow.

In CEF-PSI, part of the LEF electrons, which reached the PSI acceptor side, are branched back to the plastoquinone (in the PQ pool and cytochrome b 6 f complex) via two pathways (for a recent review, e.g., Yamori and Shikanai 2016): (a) PGR5/PGRL1-dependent (Munekage et al. 2002; DalCorso et al. 2008) and (b) NDH-dependent (Rumeau et al. 2005; Gotoh et al. 2010) pathways. PGR5 stands for Proton Gradient Regulation 5 protein, PGRL1 for PGR5-like photosynthesis phenotype 1, and NDH for chloroplast NADH dehydrogenase-like complex. In both cases, the electrons from PSI are donated via the stromal acceptor of PSI, ferredoxin (Fd). Despite significant progress in the characterization of the alternative electron transport pathways (for recent reviews, e.g., Finazzi and Johnson 2016; Yamori and Shikanai 2016; Fan et al. 2016; Shikanai and Yamamoto 2017), comparison of the relative capacity of O2-dependent electron flow and CEF-PSI in gymnosperms and angiosperms is lacking, probably because of the absence of single method for simultaneous quantification of different alternative electron flows after PSI.

The established high-sensitivity part of the Chl fluorescence decay curve (up to about 3 s) after 1-s SP to O2-dependent electron flow capacity (Shirao et al. 2013) prompted us to further develop this method for answering the above questions. Generally, the Chl fluorescence decay after a single-turnover µs-flash (e.g., Deák et al. 2014) or 1-s SP (Bukhov et al. 2001) is not a direct measure of alternative electron transport pathways. It reflects the re-oxidation kinetics of the fully reduced (during the saturating flash or pulse) primary quinone acceptor of PSII, QA −, via several electron transport processes, which depend on the redox state of the electron transport chain (PQ pool) (Renger et al. 1995; Vass et al. 1999; Bukhov et al. 2001; Deák et al. 2014 and references therein; Laisk et al. 2015) (see “Discussion”). Under certain experimental conditions, CEF-PSI was inferred to modify the slow fluorescence decay part by dark reduction of the PQ pool via stromal reductants (Yamamoto et al. 2006; Deák et al. 2014), while Laisk et al. (2015) have shown that the slow decay part (2.9–250 s) decays monoexponentially and its halftime was sensitive to the O2 concentration. Concomitant analysis of the fast and slow parts of fluorescence decay after a 1-s SP application is therefore necessary to understand both the differences and the relationships between all the decay parameters (i.e., amplitudes and halftimes) of the respective electron transport processes, providing pathways for QA − re-oxidation. If these decay parameters are also sensitive to the alternative electron transport processes, one may reveal simultaneously their relative activities and interplay. In contrast to the widely applied multicomponent analysis of fluorescence decay curves after a single saturating µs-flash (e.g., Refs. in Deák et al. 2014; Bukhov et al. (2001) were the first to perform detailed multicomponent analysis of Chl fluorescence decay after a 1-s SP application on dark-adapted barley leaves. However, they used mechanical shutter for overcoming the fitting problems with the SP tail and they did not address the alternative electron flows. As far as we know, a well-defined approach for multicomponent analysis of the routinely recorded Chl fluorescence decay curves after 1-s SP excitation is still not reported.

In the present work, we elaborated such approach. It allowed us to carry out reliable comparative analyses of the Chl fluorescence decay parameters in various species and samples under different treatment conditions. In addition to the outdoor plant species, used in the previous work (Shirao et al. 2013), we included additional laboratory-grown gymnosperms and angiosperms, cyanobacterial cells, and mutants affected in the different alternative electron transport pathways. The mutants used were as follows: the Arabidopsis mutants pgr5, which is defective in PGR5-dependent CEF-PSI (Munekage et al. 2002), and ndhM, which lacks NDH-dependent CEF-PSI (Rumeau et al. 2005; Gotoh et al. 2010), as well as cyanobacterium Synechocystis sp. PCC 6803 Δflv1 mutant, which lacks Flv-mediated O2-reduction (Helman et al. 2003; Allahverdiyeva et al. 2013; Mustila et al. 2016). Using only this simple Chl fluorescence method, we were able to provide new knowledge on the relative importance of Flv-mediated O2-photoreduction, PGR5-dependent CEF, and NDH-dependent CEF pathways, which may be of use in future studies related to evolution of higher plants and their ability to survive under dynamic photosynthetic conditions.

Materials and Methods

Plant material

Arabidopsis thaliana WT (Columbia gl1 background) was used along with its mutants pgr5 (Col-gl1 background) (Munekage et al. 2002) and ndhM (Col-0 background) (Rumeau et al. 2005; Gotoh et al. 2010). An Arabidopsis transgenic line overexpressing the gene encoding PGR5 protein (35S::PGR5, Col-gl1 background) (Okegawa et al. 2007), Nicotiana tabacum (tobacco), and seedlings of the gymnosperms Chamaecyparis obtusa, Cryptomeria japonica, Picea abies, and Pinus thunbergii were also used. Plants were raised in a growth chamber (PFD 50–60 µE m−2 s−1, 23 °C, 50–70% RH). An 8-h light/16-h dark cycle was used for Arabidopsis and tobacco for 11–14 weeks and a 16-h light/8-h dark cycle was used for the gymnosperms for 15 months. Data obtained from plants grown in outdoor conditions (Shirao et al. 2013) were re-analyzed in this study. Synechocystis sp. PCC 6803 cells (Synechocystis) were cultured in a growth chamber (PFD 40 µE m−2 s−1, 30 °C) on BG-11 medium shaken at 90 rpm and equilibrated with ambient CO2 on a 16-h light/8-h dark cycle. When the medium had an OD750 of about 0.9, the cells were harvested for experimental use. The Chl concentration was 3.4–3.7 µg ml−1.

Chl fluorescence measurements

Leaf and needle Chl fluorescence was measured with a pulse-amplitude-modulated MINI-PAM or PAM 101 fluorometer (Walz, Germany) as described by Shirao et al. (2013). Fluorescence was induced with modulated red measuring light from a light-emitting diode (650 nm; <0.04 µE m−2 s−1). A saturating white-light pulse (1 s, 1700 µE m−2 s−1) and a far-red light pulse (720 nm; 1 s, 70 µE m−2 s−1) were provided using a halogen lamp. A xenon-PAM fluorometer (Walz, Germany) was employed to measure Chl fluorescence in Synechocystis cells. Light from a xenon-flash lamp was passed through a BG39 blue-glass Schott broadband filter (center wavelength = 500 nm) and neutral density filters. The measuring beam output (0.75 µE m−2 s−1) was modulated at 64 kHz to record Chl fluorescence decay after 1-s SP (white light, 6600 µE m−2 s−1). The fluorometer detector was shielded by a RG 645 red-glass Schott filter and a R65 long-pass Balzers filter. Pure O2, CO2, and N2 were mixed in a gas blender with GB-3C mass flow controllers (KOFLOC, Japan) to provide air with the desired O2 concentration. Leaves were dark-adapted for 1 h before use. A 20 µM MV (1,1′-dimethyl-4,4′-bipyridinium dichloride) solution was vacuum infiltrated into the leaves in the dark. The leaves were left in darkness for 30 min in vacuo and for another 1.5 h at atmospheric air pressure in darkness. Synechocystis cells were left in darkness for 1 h before use, and MV treatment (100 µM) was performed for another 20 min in darkness.

Multicomponent Chl fluorescence decay analysis after 1-s saturating pulse

The relaxation kinetics of Chl fluorescence, which reflects the QA − re-oxidation kinetics after the termination of illumination with a 1-s SP (i.e., the Chl fluorescence decay part) of the dark-adapted leaf, was fitted to an exponential decay function with two or three components (OriginPro 8 SR0) (Fig. S1). Before the analysis, the curves were double-normalized (between 1 and 0) to Fo and Fm levels, that is, in the form of the relative variable fluorescence \({V}\left( {t} \right){ }={ }({F}\left( {t} \right) - {Fo})/({Fm} - {Fo})\) (Strasser et al. 2000; Stirbet and Govindjee 2012). F(t) is the fluorescence intensity at time t, and Fo is the minimal fluorescence level of dark-adapted leaves recorded with very low measuring light (QA is fully oxidized; PSII centers are “open”, V(t) = 0). Fm is the maximal fluorescence level of dark-adapted leaves reached with SP illumination (QA is fully reduced, PSII centers are “closed”, V(t) = 1). The various curves were first adjusted to one and the same time point, t = 0, corresponding to the starting point of SP (1 s, 1700 µmol m−2 s−1) (Fig. S1). The fitted time range was 35 s but similar results were obtained using 70 s. The fit started from 85% from the maximum fluorescence level after the end (at t = 1 s) of the SP maximum (Fig. S1). In this way, the initial non-exponential decay part of the Chl fluorescence decay, which is distorted by the instrument response function (IRF), was excluded without mathematical deconvolution of the IRF from the pure decay signal (Ma et al. 2005). To ensure reliable comparative analysis of the calculated decay parameters for various fluorescence decay curves, we strongly recommend the fit to start from the fixed intensity level, but not a certain time after turning off the SP. This is because our experience with numerous decay curves analyses has shown that at certain fixed time after turning off the SP, the intensity level is different, especially for different leaf origin, when the O2 concentration is different or other leaf treatments are applied. The x0 fixed option of the Origin software was applied using x0 = 1 s, which is the time when the SP is ended. The calculated y0 value in the decay curve fitting equation (Fig. S1) corresponds to the non-decaying centers with the estimated decay parameters. However, their relative contribution to the total number of participants in the decaying processes was below 0.1%. Therefore, they were not included in the normalization of the sum of the amplitudes Ai to 100%. The goodness-of-fit was judged by the statistical criteria and the residual function supplied by the Origin software. The Chl fluorescence decay of dark-adapted (80 min) cyanobacterial cells in BG-11 was measured with xenon-PAM and analyzed in the same way as that of leaves and needles, but in a shorter time range—up to 15 s after turning off the SP.

Construction of the Δflv1 mutant of Synechocystis sp. PCC 6803

The Flv1 gene (sll1521) was disrupted by a cassette of the kanamycin-resistant gene (Km) via homologous recombination (Shimakawa et al. 2015 and references therein). The Km was amplified by PCR using the primers Km-f (TAGACTGGGCGGTTTTATGG) and Km-r (ATTCCGAAGCCCAACCTTT) from the pUC4-KIXX vector to which Km had been transformed from the pACYC184 vector. The up- and downstream regions of the Flv1 gene were amplified by PCR using the primers up-f (GCCCCGGCAATTTCTGTTTG), up-r (AAAGGTTGGGCTTCGGAATAACAAATTTCGGGATGCGGG), dn-f (CCATAAAACCGCCCAGTCTAACTTGGTCACCACAAACGTG), and dn-r (TTTAGTGAACTGGCTCCCCC). The front ends of up-r and dn-f complement the Km-r and Km-f, respectively. To obtain the disruption cassette, the three fragments (Km, up- and downstream regions of Flv1) were linked by successive PCR. The DNA thus formed was mixed with the WT cells and the mutant was selected on 0.5% (w/v) BG11 agar plates containing 20 µg ml−1 kanamycin. This procedure was repeated four times. Segregation of the Flv1 gene was checked by means of a PCR using the up-f and dn-r primers.

Results

Different Chl fluorescence decay curves after 1-s saturating light pulse

Figure 1a, b shows the changes in Chl fluorescence intensity induced by illuminating WT leaves with 1-s SP (white light, 1700 µE m−2 s−1). At 21% O2, Chl fluorescence decay was slower in the angiosperm Arabidopsis WT (Fig. 1a) than in the gymnosperm Pinus thunbergii (Fig. 1b). In both plants, the decay slowed down in the absence of O2 (Fig. 1a, b). Such differences in the QA − re-oxidation kinetics were previously quantified by the decay halftime of the fast part of the curves (up to 3 s after SP), and it was found that the shorter decay halftime in gymnosperms (~0.3 s) than that of angiosperms (~0.7 s) reflected the higher O2-dependent electron flow capacity of gymnosperms (Shirao et al. 2013). Figure 1c, d shows the impact of defects in CEF-PSI on the fluorescence decay curves of Arabidopsis mutants. In pgr5 mutant, the slow decay part was suppressed (Fig. 1c) relative to that in WT (Fig. 1a). On the other hand, in the transgenic line overaccumulating PGR5 protein (35S::PGR5) the slow decay part was larger than that of the WT (Fig. S3). In ndhM mutant, the fluorescence decay curve (Fig. 1d) resembles that of the WT at 21% O2 (Fig. 1a). As for the WT, at 0% O2, Chl fluorescence decay slowed down in both the pgr5 and ndhM mutants (Fig. 1c, d).

Fig. 1
figure 1

Relative variable Chl fluorescence, \({V}\left( {t} \right){ = (F}\left( {t} \right){-Fo)/(Fm-Fo)}\), induced by a 1-s saturating light pulse. Measurements were performed in leaves of Arabidopsis thaliana WT (a), Pinus thunbergii (b), Arabidopsis pgr5 mutant (c), and Arabidopsis ndhM mutant (d), either at 21% O2 (black) or 0% O2 (blue). In (c), WT at 21% O2 was also shown (green). Results from 4 to 5 independent experiments were averaged. The CO2 concentration was set to 400 ppm

Full size image

Effects of O2-dependent electron flow on the decay parameters of Chl fluorescence in WT plants

The decay parts of the curves for the four angiosperms (Arabidopsis WT, pgr5, ndhM, and tobacco) and the four gymnosperms (C. obtusa, C. japonica, P. abies, and P. thunbergii) were each fitted separately to an exponential decay function with two or three components. Table 1 lists the amplitudes (Ai) and halftimes (t 1/2i) of the decay components at decreasing O2 concentrations from 21 to 0% (see also Fig. S2). The halftime of the fast component (t 1/21) was in the region of 170–260 ms, that of the middle component (t 1/22) was 1.0–2.3 s, and that of the slow component (t 1/23) was above 4.2 s.

Table 1 Oxygen concentration dependence of amplitudes (Ai) and halftimes (t 1/2i) of decay components of exponential curves fitted to chlorophyll fluorescence decay after application of 1-s saturating pulse (white light, 1700 µE m−2 s−1) on 1-h dark-adapted leaves from Arabidopsis thaliana wild-type and tobacco, Arabidopsis mutants, and gymnosperms
Full size table

In Arabidopsis WT at 21% O2, the decay was two-exponential with a fast (A1, t 1/21) and a slow (A3, t 1/23) component (Table 1). The amplitudes were ~84% (A1) and ~16% (A3) at all O2 concentrations tested. In contrast, the halftimes t 1/21 and t 1/23 increased with decreasing O2 concentration. The results in tobacco were essentially the same as those in Arabidopsis WT.

In gymnosperms at 21% O2, the decay was also two-exponential, but it was three-exponential at lower O2 concentrations (Table 1). The appearance of the middle component was somewhat species specific: at 10% O2 in C. obtusa and P. abies and at 2–3% O2 in C. japonica and P. thunbergii. At 21% O2, the gymnosperm A1 was 94–97%, which is 10–13% greater than for Arabidopsis and tobacco. The t 1/21 was smaller by about 30 and 90 ms in the gymnosperms than in Arabidopsis and tobacco, respectively (Table 1). Whereas A1 strongly decreased with O2 concentration (from 94–97% at 21% O2 to 67–73% at 0% O2), both A2 and A3 increased at the expense of A1, and, in fact, the increase in A2 was more than two times greater than that in A3 (Table 1; Fig. S2). In contrast, the t 1/21 increased in the same way as that of the angiosperms. Interestingly, the t 1/22 was virtually independent (within the error limits) of the O2 concentration (Table 1). The gymnosperm t 1/23 was lower and was less dependent on O2 concentration than that of the WT angiosperms. Taken together, these results showed significant differences between gymnosperm and angiosperm decay parameters at all O2 concentrations tested, which could not be seen in the previous study (Shirao et al. 2013), because only the fast decay part was analyzed therein.

The MV treatment of the WT Arabidopsis leaves at 21% O2 almost eliminated the differences in decay curves and decay parameters between angiosperms and gymnosperms (Fig. 2; Table 2). MV treatment in Arabidopsis significantly increased its A1 from 84 to 97% and decreased its t 1/21 with 40 ms. These results align with the report of Bukhov et al. (2001), which showed that the MV treatment of barley increased their fast component amplitude at the expense of the others (see “Discussion”). As shown in Fig. 2b, the curves of the control and MV-treated Gnetum gnemon leaves coincided. The MV treatment did not change the Chl fluorescence decay parameters of the gymnosperm leaves (Table 2). The amplitude of the slow component A3 ~ 3%, which is equal in both gymnosperm and angiosperm decay curves (Table 2), should not be connected with PSI, because MV opens the PSI acceptor side (Bukhov et al. 2001); most probably, A3 reflects the contribution of the inactive PSII centers, as described in Bukhov et al. (2001). MV accepts electrons from Fe-S center (FA/FB) in PSI (Takahashi and Katoh 1984) and as MV radical rapidly reacts with O2 (Golbeck and Cornelius 1986). MV treatment is also used to abolish CEF-PSI (Jia et al. 2008). Therefore, the significant effect of MV treatment in angiosperms and the lack of the same in gymnosperms (Table 2) clearly indicated that the differences in decay parameters between untreated gymnosperms and angiosperms (Table 1) can be explained by differences in the relative distributions of electrons beyond the acceptor side of PSI, that is, to electron flow for O2-photoreduction and to CEF-PSI pathways for dark electron recycling.

Fig. 2
figure 2

Effects of methyl viologen (MV) on the relative variable Chl fluorescence, \({V}\left( {t} \right){ }={ }({F}\left( {t} \right) - {Fo})/({Fm} - {Fo})\), in leaves of Arabidopsis thaliana WT (a) and Gnetum gnemon (b) induced by a 1-s saturating light pulse. Measurements were performed in air containing 21% O2 and 400 ppm CO2 in either the absence or the presence of 20 μM MV (red). In the absence of MV, leaves were treated with distilled water (blue). Results from intact leaves were shown (black). Results from 4 to 5 independent experiments were averaged

Full size image
Table 2 Effect of MV treatment on the decay parameters of Chl fluorescence
Full size table

Effects of Flv-mediated O2-photoreduction on the Chl fluorescence decay parameters

Since both cyanobacteria (Badger et al. 2000) and gymnosperms (Shirao et al. 2013) have high capacity of O2-dependent electron flow, we used a mutant of Synechocystis deficient in Flv1 gene (Helman et al. 2003) to quantify the maximal possible decay parameter changes in the presence and absence of Flv-mediated O2-reduction. Although both Flv1 and Flv3 are required for the Flv-mediated O2-reduction in cyanobacteria (Helman et al. 2003; Allahverdiyeva et al. 2011, 2013), the single mutant Δflv1 we used lacks Flv-mediated O2-reduction, because in this mutant the Flv3 accumulation was reduced (Mustila et al. 2016). Figure 3 shows the comparison of the effect of presence and absence of Flv-mediated O2-reduction on the Chl fluorescence decay. Fm is lower in Synechocystis (Fm/Fo ~ 1.9) than in higher plants (Fm/Fo > 5.0) because fluorescence from phycobilisomes increases Fo (Schuurmans et al. 2015). Nevertheless, the comparison of the decay curves (Fig. 3, black lines) and their multicomponent analysis (Table 3) provide direct evidence for the significant impact of Flv-mediated O2-photoreduction on Chl fluorescence decay.

Fig. 3
figure 3

Effects of O2-dependent electron flow mediated by Flv1 in Synechocystis on the relative variable Chl fluorescence, \({V}\left( {t} \right){ }={ }({F}\left( {t} \right) - {Fo})/({Fm} - {Fo})\). Measurements were performed in WT (a) and Δflv1 mutant (b) in either the absence (black) or the presence of 100 µM MV (red). Results from 4 to 5 independent experiments were averaged

Full size image
Table 3 Effect of MV treatment on Chl fluorescence decay parameters in Synechocystis cells at 21% O2
Full size table

It can be seen from Fig. 3 (black lines) that Chl fluorescence decay slowed down in Δflv1, thus confirming the high sensitivity of the decay curves to O2-dependent electron flow in cyanobacteria as well. Both cyanobacteria (Fig. 3) and gymnosperms (Fig. 1) have faster decaying curves than their respective counterparts (Δflv1 and angiosperms). As shown in Table 3, the Δflv1 mutant had lower A1 and higher A3 than the WT, and a middle component appeared. This effect of the lack of Flv-mediated O2-reduction is just as that of low O2 concentrations in gymnosperms (Table 1). The Flv-mediated O2-photoreduction increased the A1 by 11% (Table 3). This difference is as large as that between A1 values in gymnosperms and angiosperms at 21% O2 (10–13%) (Table 1). This experiment thus indicated that such difference in A1 between gymnosperms and angiosperms could be attainable if Flv in gymnosperms mediate the O2-photoreduction. In this study, pilot measurements of the Arabidopsis WT decay curves at 80% O2 ruled out the possibility that excess O2 could alter A1 (data not shown).

The appearance of the middle component in Δflv1 (Table 3) reflected an increase in the reduction of the PSI acceptor side because of the lack of Flv-mediated electron drain from PSI to O2 (see “Discussion”). Since the appearance of the middle component in Δflv1 resembled that in gymnosperms upon decreasing O2 concentration (Table 1), this result also provided the reason for the appearance of middle component in gymnosperms—the decrease in O2 (the substrate of the O2-reduction reaction) leads to a decrease in the electron drain from PSI to O2, which increases the reduction of the PSI acceptor side.

The MV treatment accelerated the decay in Δflv1 to near the WT level but the WT was insensitive to MV treatment (Fig. 3; Table 3), thus demonstrating that MV was able to affect the decay kinetics in Synechocystis only when Flv-mediated O2-dependent electron flow was missing. MV treatment in Δflv1 equalized its A1 with that of WT (Table 3), just as the MV treatment in Arabidopsis WT, which lacks Flv, rendered its A1 equivalent with that of gymnosperms (Table 2). MV treatment in Synechocystis WT (Table 3) as in gymnosperms (Table 2) affected neither their A1 nor their t 1/2 1. Taken together, these results demonstrated the virtually equal extent of maximal changes in the decay parameters induced by the lack of Flv-mediated O2-reduction in cyanobacteria and that observed between the decay parameters of gymnosperms and angiosperms.

Effects of PSI cyclic electron flows on the decay parameters of Chl fluorescence

In contrast to Arabidopsis WT, the decay was three-exponential in pgr5 mutant at all O2 concentrations tested (Table 1). The middle component A2 appeared at the expense of the A3, because the A3 of pgr5 was about half that of the WT, while the corresponding values for A1 were similar. When O2 decreased, the A1, A2, and A3 of pgr5 did not change. Although the t 1/23 of pgr5 was slightly less than that of the WT over the range of O2 concentrations, t 1/23 also increased with decreasing O2 and the largest difference was detected at 0% O2 (18.5 s in pgr5 vs. 54.4 s in the WT). Interestingly, the t 1/21 of pgr5 was longer than that of WT with ~30 ms. As was the case for gymnosperms, the pgr5 t 1/22 was virtually independent of O2 concentration. These results suggested that a PGR5-dependent CEF-PSI participates in all three Chl fluorescence decay components. PGR5 accelerated the fast decay component kinetics, eliminated the middle component (its A2 was transmitted to the A3), and decelerated the slow decay component kinetics. While the effect of O2 above 10% in gymnosperms resembles that of PGR5 on the three decay components, there is strict difference in the elimination of the middle component—in gymnosperms A2 was transmitted to A1. This comparison clearly indicates that A1 involves the relative parts of electrons branched at PSI stroma side in the O2-photoreduction pathway, while A3 involves the relative part of electrons branched in the CEF-PSI pathway. There was a very weak shoulder in the Chl fluorescence for the WT and ndhM at 21% O2 (arrows in Fig. 1a, d). The shoulder became taller when the O2 decreased and shifted to longer times: from 3–4 s (at 21% O2) to 8–10 s (at 0% O2) after SP ended. Sometimes, at the lowest O2 concentrations, the fit did not converge. In such cases, we fixed the y0 (see “Materials and Methods”) to the value calculated after a longer-range fit until the curve was fully relaxed, e.g., 70 s or more. When there was a shoulder, the decay was always two-exponential, because when three exponents were fitted, two of them had equal decay times. Similar shoulders have been reported for the fluorescence decay curve of transplastomic tobacco (which overexpresses Fd) after 1-s SP illumination (Yamamoto et al. 2006). This response was attributed to the fact that it had higher CEF-PSI activity than the WT (Yamamoto et al. 2006). A shoulder was also observed with cyanobacteria after single-turnover saturating flash application (Deák et al. 2014) and was ascribed to CEF-PSI mediated by the cyanobacterial NDH-1 complex. In the present study, the shoulder was strictly absent in pgr5 leaves (Fig. 1c), which suggested that it can be accounted for by PGR5-dependent CEF-PSI. The shoulder was also strictly absent in all laboratory (see P. thunbergii in Fig. 1b) and outdoor (see below) gymnosperms implying lower capacity of PGR5-dependent CEF-PSI than in angiosperms under the present experimental conditions. In the ndhM mutant, the middle component was absent as it was in the WT, but the parameters indicated that the curves decayed slightly faster than in WT (Fig. 1a, d; Table 1). In ndhM, t 1/23 at 0% O2 was 27.5 s whereas it was 54.4 s for the WT. This result suggested that NDH-dependent CEF-PSI is involved in decelerating the slow component, although less significantly than PGR5-dependent one.

To further examine the origin of the shoulder in WT and ndhM, but its absence in pgr5 (Fig. 1), we used a far-red-light pulse (FR pulse) (1 s, about 70 µE m−2 s−1 at >700 nm). The FR light preferentially excites PSI: the typical Fm/Fo values under FR pulse were about half (~3.0 ± 0.2) than those under white SP (~5.5 ± 0.2) and induces CEF-PSI (Desai et al. 1983; Katsumata et al. 2008 and Refs therein). It can be seen from Figure 4 that upon FR pulse application the shoulders in the WT (Fig. 4a) and ndhM (Fig. 4b) curves were more prominent than under white SP. Nevertheless, even under far-red, the pgr5 curve lacked any shoulder (Fig. 4c), which confirmed the fact that the shoulder was due to PGR5-dependent CEF-PSI.

Fig. 4
figure 4

Relative variable Chl fluorescence, \({V}\left( {t} \right){ }={ }({F}\left( {t} \right) - {Fo})/({Fm} - {Fo})\) induced by 1-s far-red pulse. Measurements were performed in leaves of Arabidopsis thaliana WT (a), ndhM (b), and pgr5 (c) in nitrogen gas containing 21% O2 and 400 ppm CO2. Measurements were performed under far-red pulse (red) and white SP (black; data from Fig. 1). Results from 4 to 5 independent experiments were averaged

Full size image

Chl fluorescence decay analysis in various plants grown under field conditions

Chl fluorescence decay analyses on laboratory plants (Fig. 1) were cross-validated in various outdoor species (Table S1). The decay was fitted to either two or three exponentials in angiosperms and essentially to two exponentials in gymnosperms (Table S1). The results were similar to those obtained for laboratory plants (Table 1) except that a middle component was often detected in the angiosperms (23/50 species) and less often in the gymnosperms (2/27 species). A1 was larger in the outdoor gymnosperms (average 97 ± 3%) than the angiosperms (average 76 ± 7%) and t 1/21 was also smaller in the gymnosperms (average 163 ± 22 ms) than the angiosperms (average 213 ± 47 ms). Therefore, gymnosperms generally have higher O2-dependent electron flow capacities than angiosperms. The wider spread in A1 between outdoor gymnosperms and angiosperms (Table S1) than that between the laboratory plants (Table 1) would be due to the different SP intensity used (5500 vs. 1700 µmol m−2 s−1), respectively; the latter was chosen for the laboratory plants because the former produced slightly lower (with few %) Fm/Fo values. In 44% of the angiosperms, a shoulder was detected in the Chl fluorescence decay curve at ~2–5 s after SP termination. In contrast, the shoulder was not detected at all in the gymnosperms. The middle component in part of outdoor angiosperms might be linked with species-specific limitation of PGR5-dependent CEF-PSI since it appeared in the absence of PGR5 (Table 1) and/or to some species-specific restrictions at the PSI acceptor side under these experimental conditions (see “Discussion”). In the two gymnosperm species where the middle component was detected, the acceptor side of PSI may have been reduced by the limitation of the electron flow to O2 similar to that which occurs in gymnosperms at low O2 (Table 1).

The results for the ferns resembled those for the gymnosperms. Nevertheless, only two fern species were tested and this sample size does not suffice to determine a general tendency for this plant group.

Discussion

Assignment of the Chl fluorescence decay components to different electron transport processes

In the present study, a multicomponent analysis was conducted on the relative variable Chl fluorescence relaxations induced by the illumination of dark-adapted leaves with a 1-s SP. The time resolution of the commercial PAM apparatus used in this study was low (time step 10 ms), and no mechanical shutter was implemented (Bukhov et al. 2001) for overcoming the problems with mathematical deconvolution of the pure decay function from the IRF. Nonetheless, an approach was developed to realize robust comparative analyses of the Chl fluorescence decay curves from various species and samples under different conditions.

Multicomponent analysis of Chl fluorescence decay after the application of single-turnover saturating flash of µs-range duration is widely used to characterize the effects of stressors on electron transport processes (Renger et al. 1995; Vass et al. 1999; Deák et al. 2014 and references therein). Three decay components are usually resolved: fast (~300–500 μs), middle (~5–15 ms), and slow (several seconds) (see, e.g., Deák et al. 2014 and references therein). In general, the fast component is related to the re-oxidation of QA − by QB, the middle one to the re-oxidation of QA − by QB in centers that had an empty QB site during excitation, and the slow one to the recombination between QA − and the S2(3) state of the oxygen-evolving complex. Deák et al. (2014) exposed cyanobacteria (150 µs–100 s fitting time range) to 20 µs single saturating flash and found that the fast component correlated to the forward electron flow and that of 100 ms to 1–2 s to electron transport chain restrictions caused by microanaerobic conditions plus inhibitors like DCMU or DBMIB. Bukhov et al. (2001) subjected barley to 1-s SP with a mechanical shutter. They showed that the fast component (6–8 ms) indicates a rapid QA − oxidation during the electron flow to PQ. It is commensurate with maximum LEF through the electron transport chain. The middle (60–80 ms) and slow (650–680 ms) components were observed under restricted electron flow and excessive electron carrier reduction. They were ascribed to the recombination of QA − with an oxidized species, Yz or tyrosine-161 of the D1 polypeptide on the donor side of PS II, and to the charge recombination between QA − and the S2 state of the oxygen-evolving complex, respectively (Bukhov et al. 2001). The slowest component in their 10-s time fitting range accounted for 4–5% of the maximum variable fluorescence. It had a lifetime of several seconds and was suggested to represent a minor population of inactive PSII centers. Laisk et al. (2015) studied the slow decaying part (2.9–250 s) of the Chl fluorescence relaxation in sunflower (an angiosperm) after a 0.6-s SP exposure. They fitted it with a single exponential function, and based on its O2 dependence, they attributed the long component (~13 s) to PQ re-oxidation in darkness by the plastid terminal oxidase (PTOX).

In the present work, the fast halftime (~170–260 ms, Table 1) is longer than those obtained from exposures to µs saturating flash (300–500 μs,) or to 1-s SP with a mechanical shutter (6–8 ms). Nevertheless, A1 resolved in the present work indicates almost the same relative number of electrons participating in the fast QA − re-oxidation process as that found using Chl fluorescence relaxation induced by single µs-flash or 1-s SP. The reason is that the amplitudes (pre-exponential factors) in general and the A1 in particular represent the relative number of participants at time zero in the fluorescence relaxation processes, which decay with different decay halftimes. To find a better correspondence between the halftimes estimated in Bukhov et al. (2001) and those we report here, we calculated the decay halftime of the SP in the same way as we have analyzed the decay curves (Fig. S2). The SP decay halftime was t E ~ 70 ms. The corrected (pure) t 1/21 can be estimated using the empirical formula of Teuchner (1979): \({t_{{1}/{2}}}{1 }={t_{{exp}}} - {t_{E}}\), where t exp is the calculated t 1/21 given in Table 1. For the species listed in Table 1, t /21 approximates the fast (6–8 ms) + middle (60–80 ms) halftimes of reported by Bukhov et al. (2001). Moreover, the fact that MV decreased t 1/21 and increased A1 (see Table 2) indicates that the fast component is equivalent to the sum of the fast and middle components resolved by Bukhov et al. (2001): the pure fast halftime t 1/21 of MV-treated angiosperm decreases with the same value (~40 ms, Table 2) as that of the middle component of MV-treated angiosperm in Bukhov et al. (2001) (~50 ms). The amplitudes of the middle (60–80 ms) and slow (650–680 ms) components in Bukhov et al. (2001) were significantly diminished by the MV treatment. This effect was attributed to MV opening the PSI acceptor side. For these reasons, Bukhov et al. (2001) reported that restrictions of electron flow on the acceptor side of PSI account for the middle (60–80 ms) and slow (650–680 ms) components. Based on the above comparisons and our results for A1, t 1/2, and MV treatment effects in all samples, our fast component can be used as a relative measure of QA − re-oxidation via forward electron transport, that is, the LEF passing through the PSI acceptor side.

The middle component in outdoor plants (lowest t 1/22 = 540 ms, Table S1) was, in some cases, as low as the slow component of Bukhov et al. (2001) (650–680 ms). In this study, the middle component (1–2.3 s, Table 1) is thought to correspond to the slow component resolved by Bukhov et al. (2001) and the middle component (1–2 s) in cyanobacteria treated with electron transport chain inhibitors (Deák et al. 2014). Moreover, our results, demonstrating that the middle component appeared in pgr5 (Table 1), Δflv1 (Table 3), and gymnosperms at low O2 (Table 1) and reflected limitations at the PSI acceptor side, support this assignment. Following the above authors, its molecular origin could be related to charge recombination between QA − and S2 state of the oxygen-evolving complex.

The slow component of this study (>4.2 s) resembles that (~13 s) resolved by Laisk et al. (2015) and the long component (~10 s) in the µs-flash experiments of Deák et al. (2014). This component is not resolved in the study of Bukhov et al. (2001), because of the shorter (10 s) fitting range used. Moreover, our slow component is sensitive to CEF-PSI (Fig. 1; Table 1), similar to the afterglow band in the thermoluminescence curves (Ducruet 2003; Havaux et al. 2005); the afterglow band is due to charge recombination between QB − and the S2 or S3 states (Ducruet 2003), for which electrons are fed from stromal reductants to PQ and QB by CEF-PSI (Havaux et al. 2005).

Taken together, these assignments show that the applied analysis of the Chl fluorescence decay curves can be successfully used to gain the same knowledge on photosynthetic electron transport pathways as with more specialized methods using µs-flash or 1-s SP with mechanical shutter. Moreover, the decay parameters obtained in the present study are sensitive to the alternative electron transport processes and reveal their relative activities as discussed in the following sections.

The larger amplitude A1 and the smaller halftime t 1/21 indicate the larger capacity for O2-dependent electron flow from PSI

The multicomponent analysis of Chl fluorescence decay at 21% O2 shows that the most significant differences in the decay parameters (Ai, t 1/2i) between angiosperms and gymnosperms are found in A1 and t 1/21 (Table 1, Table S1). The observation that these differences vanish when the angiosperm leaves are treated with MV (Fig. 2; Table 2) convincingly demonstrates that the electrons used for the O2-photoreduction from PSI are part of the forward linear electron flow upstream of PSI, responsible for the fast QA − re-oxidation. Therefore, A1 is correlated with the relative number of electrons directed from PSI to O2 (Mehler reaction) and this electron transfer pathway has a higher rate coefficient (~1/t 1/21) than that of CEF-PSI quenched by MV. Further, A1 in Synechocystis is larger when the Mehler-like reaction is present (~98%) but smaller in its absence (~87%) (Fig. 3; Table 3). Based on these results, it is concluded that the higher capacity of the Mehler-like reaction is expressed by both a larger A1 and a smaller t 1/21.

The experiments in which O2 was decreased revealed that whereas A1 did not change in angiosperms, it significantly decreased in gymnosperms. On the other hand, t 1/21 similarly increased (Table 1). Therefore, the larger the O2-dependent electron flow capacity, the larger the impact on the fast component. It is thought that in gymnosperms, both A1 and t 1/21 are affected by the large decreases in O2-dependent electron flow that result from decreasing O2. In angiosperms, only t 1/21 is affected by the O2 decrease since O2-dependent electron flow is small or negligible. MV, however, increased A1 in Arabidopsis (Fig. 2; Table 1) because it significantly increases O2-dependent electron flow. A1, then, is altered only when the O2-dependent electron flow capacity is large.

Indirect evidence for Mehler-like reaction in gymnosperms

The genes encoding proteins homologous to Flv1 and Flv3 from bacteria and Archaea are conserved in mosses, lycophytes, and gymnosperms (Allahverdiyeva et al. 2015; Yamamoto et al. 2016). Several candidate genes were also confirmed here for the Flv homologs in the gymnosperms P. sylvestris, P. taeda, P. sitchensis, P. glauca, and C. japonica (Fig. S4). Angiosperms, however, do not conserve Flv-encoding genes (Zhang et al. 2009; Peltier et al. 2010; Yamamoto et al. 2016). Cyanobacteria and algae have high electron flow activity dependent on the Mehler-like reaction (Schreiber et al. 1995; Badger et al. 2000), whereas the Mehler reaction has low activity in angiosperms (Badger et al. 2000; Shirao et al. 2013). This difference has raised the question as to whether the Flv in gymnosperms account for the fact that O2-dependent electron flow capacity is higher in them than in angiosperms (Shirao et al. 2013).

The present work provides the answer to this question. The strong similarity established here between cyanobacteria and gymnosperms in terms of the extent of changes in the characteristics reflecting their higher O2-dependent electron flow capacity, larger A1, smaller t 1/21, and the absence of MV effect (Tables 1, 2, 3) relative to the respective counterparts (Δflv1 and angiosperms) represents the first convincing evidence that the conserved Flv in gymnosperms mediate a Mehler-like reaction resembling that in cyanobacteria. This indirect evidence also clarifies the reason for the higher O2-dependent electron flow activity in gymnosperms than in angiosperms. In future studies, direct evidence for this mechanism could be obtained using Flv-depleted gymnosperm mutants. Therefore, we conclude that the most likely reason for the higher O2-dependent electron flow capacity in gymnosperms than in angiosperms is the conserved Flv in gymnosperms.

Differences between PGR5-dependent and NDH-dependent CEF-PSI

This study showed that under non-steady-state photosynthetic conditions, the absence of PGR5 has a much stronger impact on the decay parameters than the absence of NDH-M (Table 1). These results are in line with the domination of PGR5-dependent over NDH-dependent CEF-PSI in Arabidopsis (Yamori and Shikanai 2016 for a review). In addition, however, this study revealed new knowledge regarding the mechanism by which the defects in the two CEF-PSI pathways (absence of either PGR5 or NDH-M) affect the fluorescence parameters.

First, our results, showing that at 21% O2 t 1/21 in pgr5 is longer by about 30 ms than those in WT and ndhM, while A1 values are similar (Table 1), clearly demonstrate that PGR5 is required for a higher rate of electron acceptance at the PSI acceptor side. It is concluded that the presence of PGR5 accelerates LEF (lowers t 1/21) but does not change the relative number of electrons in it (A1). The method used reveals this property of PGR5 by temporally distinguishing the electron flux between LEF and CEF-PSI because their rate constants differ significantly (Laisk et al. 2005; Hald et al. 2008; Fisher and Kramer 2014). Since PGR5 accelerates LEF kinetics but NDH does not, in strongly illuminated dark-adapted samples PGR5 should accelerate electron drain from PSI and therefore interact more frequently with it than the NDH complex. Recall that in gymnosperms, t 1/21 at 21% O2 is on average ~50 ms less than that for WT angiosperms (Table 1). Therefore, based on these values, at 21% O2, the channeling of electrons from PSI to O2 is more efficient in gymnosperms than that from PSI to CEF-PSI in angiosperms. The least efficient channeling of electrons is from PSI to CEF-PSI pathway in the absence of PGR5.

Second, the lack of PGR5 (but not of NDH-M) results in the appearance of the middle component (Table 1), which reflects restrictions of electron flow at the acceptor side of PSI as described above. Since PGR5 is missing, P700 and therefore the acceptor side of PSI would be reduced (Iwai et al. 2010). This effect results in electron accumulation in the transport chain (Tsuyama and Kobayashi 2009), which may drive reverse electron transfer process and cause the middle component to appear (A2, t 1/22, Table 1). Its amplitude A2 represents the relative parts of electrons, which are neither involved in forward electron transport at the PSI side (A1) nor recycled as part of the CEF-PSI (A3). It is as yet unknown whether this backward electron transfer starts at PSI or a PQ pool. However, an important point is that t 1/22 is practically constant upon decrease in O2 (Table 1). This means that the kinetics of the fluorescence relaxation process (characterized by 1/t 1/22) is independent of O2 availability. Since t 1/21 and LEF depend on O2 concentration but t 1/22 does not (Table 1), we could propose that the reverse electron flux resulting in fluorescence decay described by the middle component (A2, t 1/22) is either a charge recombination in PSI or involves PQ pools not participating in LEF.

Third, this study showed the following: (i) the absence of PGR5 significantly diminishes A3 (16–17% in WT and ndhM, 6% in pgr5), whereas the absence of NDH-M does not (Table 1); (ii) as O2 decreases, t 1/23 in pgr5 and ndhM increase in a similar way but much less than t 1/23 in WT (Table 1); (iii) PGR5 overexpression increases t 1/23 to almost twice that of WT at 21% O2 but does not change the other decay parameters (Fig. S3). Based on these findings, a mechanism by which the absence of PGR5 or NDH-M affects fluorescence decay is proposed. During SP, electrons that depend on the PGR5 interaction with PSI (because only PGR5, but not NDH, can decrease t 1/21) photoaccumulate on the PSI acceptor side, while neither CEF-PSI is active during SP since their turnover is >1 s (Laisk et al. 2005; Hald et al. 2008; Fisher and Kramer 2014). These photoaccumulated electrons contribute to ~2/3 of the A3 in WT and ndhM. After SP cessation, in darkness, both CEF-PSI pathways help reduce the PQ pool because the absence of either decreases t 1/23. Based on A3 and t 1/23, the slow component indicates the impact of CEF-PSI. The PGR5-dependent CEF-PSI contributes far more to the relative number of recycled electrons (up to 2/3) than does NDH-dependent CEF-PSI. The presence of the shoulder that accompanies the slow decay component supports this conclusion (Figs. 1c, 4c). The shoulder is large when the activity of PGR5-dependent CEF-PSI is high.

Simultaneous quantification of the relative parts of electrons branched in cyclic and pseudocyclic pathways

Based on the established absence of O2-dependent change in A1 and A3 of angiosperms (Table 1), the relative part of electrons branched at PSI for O2-photoreduction is below 1–2% (within error limits), while for CEF-PSI it is around 13% (A3 corrected with 3% for the A3 value after MV treatment). For gymnosperms, the difference between A1 values of gymnosperms and angiosperms at 21% O2 (10–13%) represents the relative part of electrons branched in the Flv-mediated O2-photoreduction pathway, while that in CEF-PSI is 0–3% (A3 corrected with 3% as above). Therefore, we can conclude that the cyclic electron flow dominates over pseudocyclic in angiosperms, while the opposite occurs in gymnosperms.

Conclusion

Chl fluorescence decay in dark-adapted gymnosperm and angiosperm leaves after a 1-s saturating light pulse can be fitted to an exponential decay function with two or three components: fast, middle, and slow. O2-dependent electron flow enhances the amplitude and decreases the halftime of the fast component in both gymnosperms and Synechocystis. The middle component appears when the PSI acceptor side is reduced and was observed in the pgr5 and Δflv1 mutants and gymnosperms at lower O2 concentrations. The slow component was attributed to electron recycling via CEF-PSI pathways. At 21% O2, gymnosperms have higher O2-dependent electron flow (~10–13% of the total electron flow originating from PSII) than angiosperms (below 1–2%), and their O2-dependent electron flow dominates over CEF-PSI (~0–3%). In contrast, CEF-PSI is the favored alternative electron flow (~13%) in angiosperms. This reversal is due to a Mehler-like reaction in gymnosperms, evidenced here by the strong similarities between gymnosperms and cyanobacteria in terms of decay characteristics sensitive to O2-dependent electron flow and MV treatment. In dark-adapted, non-stressed higher plants, Chl fluorescence decay after 1-s SP is determined from the relative capacities of LEF (including O2-dependent electron flow) and CEF-PSI. The significant effects of O2-dependent electron flow and CEF-PSI on Chl fluorescence decay after SP may serve as easy and rapid indicators of their relative capacities in plants subjected to a variety of environmental conditions and stress factors. They can also be of interest in future studies on the fastest regulatory mechanisms of dynamic photosynthesis.

Note added in proof

After submission of this manuscript we became aware of studies, showing that the Flv mediate fast oxidation of the P700, the primary electron donor of PSI, during dark-to-light transition using Flv-deletion mutants of Synechocystis, Physcomitrella (a moss) (Ilík et al. 2017), and Marchantia (a liverwort) (Takagi et al. 2017). Our results are consistent with those of Ilik et al. (2017) and Takagi et al. (2017), because fast P700 oxidation accompanies fast decay of Chl fluorescence after SP application (Shirao et al. 2013). An advantage of our fluorescence decay-based approach is that it also provides concomitant quantification of the different alternative electron flows.