Photoinhibition is the light-induced down-regulation of photosynthetic efficiency, the primary target of which is photosystem II (PSII). Currently, there is no clear consensus on the exact mechanism of this process. However, it is clear that inhibition can occur through limitations on both the acceptor- and donor side of PSII. The former mechanism is caused by electron transport limitations at the PSII acceptor side. Whilst, the latter mechanism relies on the disruption of the oxygen-evolving complex. Both of these mechanisms damage the PSII reaction centre (RC). Using a novel chlorophyll fluorescence methodology, RC photoinactivation can be sensitively measured and quantified alongside photoprotection in vivo. This is achieved through estimation of the redox state of QA, using the parameter of photochemical quenching in the dark (qPd). This study shows that through the use of PSII donor-side inhibitors, such as UV-B and Cd2+, there is a steeper gradient of photoinactivation in the systems with a weakened donor side, independent of the level of NPQ attained. This is coupled with a concomitant decline in the light tolerance of PSII. The native light tolerance is partially restored upon the use of 1,5-diphenylcarbazide (DPC), a PSII electron donor, allowing for the balance between the inhibitory pathways to be sensitively quantified. Thus, this study confirms that the impact of donor-side inhibition can be detected alongside acceptor-side photoinhibition using the qPd parameter and confirms qPd as a valid, sensitive and unambiguous parameter to sensitively quantify the onset of photoinhibition through both acceptor- or donor-side mechanisms.
Photoinhibition of photosynthesis has been long characterised as the harmful, light-induced down-regulation of photosynthetic efficiency [1,2]. The main target of this being photosystem II (PSII) [3,4], a multi-protein water-plastoquinone oxidoreductase composed of a homodimeric core and a peripheral antenna system . In higher plants, the PSII subunits are chlorophyll-binding proteins which act as a ‘programmed solvent’ to effectively deliver excitation energy to the PSII reaction centre (RC) [6,7]. The RC is located within the D1/D2 heterodimer , with the catalytic site of oxygen evolution within the RC consisting of a tetramanganese-calcium cluster, Mn4CaO5, called the oxygen-evolving complex (OEC), which sits on the lumenal face of the PSII core [9,10]. Once the excitation energy from a captured photon is delivered into the RC, charge separation of the P680 chlorophyll species generates a redox potential of around +1.21 V . This state is re-reduced within nanoseconds through electron transfer from both the donor- and acceptor-sides of PSII [12,13]. On the acceptor side, the pheophytin (Phe) electron is transferred to the QA site, and then to the QB plastoquinone, which diffuses away into the mobile plastoquinone pool within the lipid phase of the membrane [4,14]. On the PSII donor side, the D1 Tyrz extracts an electron from the OEC, which then undergoes a series of changes in redox state through a process known as the S-state cycle. Within this process, water is oxidised, protons are released into the thylakoid lumen, and an electron is transferred from the water back into the OEC [15–17].
However, due to the high oxidative potential generated within the RC , any limitations in electron transfer at either the donor or acceptor side of PSII can lead to photodamage. The exact mechanisms of photodamage of PSII are still widely debated and there is no clear consensus [3,4]. Under conditions of acceptor-side limitation, photochemistry outpaces processes downstream of PSII, and electron transport via the plastoquinone pool becomes the principal rate-limiting step . Consequently, this leads to the presence of semi-stable and stable reduced states of QA . The collapse of the Phe- P680+ charge-separated state leads to the formation of 3P680 and singlet oxygen and other ROS within the RC [18,19]. Within this model, the allocation of excitation energy regulates the potential for damage, with dynamic regulatory processes compensating for changes in the light environment . Limitations at the PSII donor side have been classified into two categories: ‘statistical’ photoinhibition, and Mn-dependent photoinhibition . The former is built on the direct in vitro observations that after chemical inactivation of the OEC through hydroxylamine-washing of samples [22–24], tris washing [25,26], or Cd2+ treatment [27–29], the lifetime of the P680+ species is lengthened and has also been hypothesised in vivo . The latter Mn-dependent mechanism has been proposed due to a perceived linearity between the rate of photoinhibition and light intensity . In this model, direct excitation of Mn causes loss of functionality of the OEC, as the first photoinhibitory step [3,21]. This has been directly observed on in vitro samples under blue and UV wavelengths (i.e. λ < 500 nm) . Whilst the OEC has been widely accepted as a locus of damage in PSII under UV wavelengths, photoinhibition under visible light in vivo is much more complex . For example, some observations have shown that the loss of the OEC is a much later step in the photoinhibitory pathway under visible light  and that the OEC is unlikely to be directly excited under visible wavelengths .
To mitigate photodamage, however, nature has evolved a number of photoprotective measurements. Many of which rely on the strategies of either managing light absorption or managing the absorbed energy itself [34,35]. Nonphotochemical quenching (NPQ), which forms and relaxes on a timescale of seconds to minutes, is an example of the latter. Through this process, the excess excitation energy is harmlessly dissipated as heat, reducing the density of excitation pressure on PSII  and has been shown to be vital to crop productivity in the field . NPQ is a heterogeneous process and can be broken down into multiple major components: qE, energy-dependent quenching; qI, photoinhibitory quenching; and qZ, zeaxanthin-dependent quenching . qE relies on acidification of the thylakoid lumen, is modulated by the PsbS protein and the deepoxidation of the xanthophyll pigment violaxanthin into zeaxanthin, and is the major photoprotective component of NPQ . Whilst qI, traditionally prescribed to photoinhibitory components at the level of the RC , has also been shown to also be associated with long-term protective mechanisms [40,41]. Due to the functional heterogeneity in widely used parameters, such as qE, qI, and Fv/Fm, using these as markers of light tolerance, photoprotection, and photoinhibition often leads to ambiguous results . Furthermore, other in vitro methods for estimating photoinhibition, such as D1 immunoblotting can lead to ambiguity as they give no indication of functionality and often have low accuracy with respect to other methods [42,43]. To overcome this, a novel pulse-amplitude-modulated (PAM) fluorescence procedure was recently developed to monitor the protective effectiveness of NPQ against the photoinactivation of RCs in vivo [43,44]. Using the parameter of photochemical quenching in the dark (qPd), the redox state of QA, and therefore the early onset of RC inactivation can be sensitively monitored alongside photoprotective mechanisms (i.e. NPQ). qPd has also been shown to decline linearly with the rates of oxygen evolution , showing that it is a valid parameter to accurately estimate overall photoinactivation of PSII. For a recent detailed review of the procedure, see .
After light treatment, when the acceptor side is photoinhibited, there is a rise in the minimum fluorescence yield (Fo) , due to the presence of the stable QA− state . This will be directly measured by the qPd parameter. However, under donor-side inhibition, there is little to no change in Fo . Currently, how donor-side photoinhibition is reflected in the qPd parameter is unclear. Within this study, we have investigated whether donor-side inhibition can be detected by the qPd parameter through the treatment of wild-type Arabidopsis thaliana leaves with donor-side inhibitors, such as UV-B light and Cd2+.
Materials and methods
Plant material and growth conditions
Wild-type Arabidopsis thaliana (Col-0) was used within this work. Seeds were sterilised in 50% ethanol and 0.1% Triton-X 100 and stored at 4°C for 72 h prior to sowing. Plants were grown on a 6 : 6 : 1 ratio of John Innes No. 3 soil, Levington M3 compost and Perlite (Scotts U.K., Ipswich, U.K.). Light intensity (190 µmol m−2 s−1), photoperiod (10 h light/14 h darkness) and temperature (22°C) were controlled, with water being added into the trays twice a week. All measurements were carried out on 5–6-week-old plants.
All measurements were carried out using a DUAL-PAM-100 fluorimeter (Walz, Effeltrich, Germany), fitted with a red (peak λ = 620 nm) actinic light source. Using a series of nineteen 2 min long steadily increasing actinic light intensities, from 0 to 2636 µmol photons m−2 s−1. NPQ levels were quantified via a saturating pulse at the end of each light phase (0.6 s, 4000 µmol photons m−2 s−1). Each light phase was proceeded by 7 s of far red light, 3 s of darkness, and another saturating pulse, from which qPd was quantified. An example trace can be seen in Supplementary Figure S1. Population light tolerance curves were constructed using the percentage of a population of leaves that were photoinhibited at each light intensity (i.e. qPd < 0.98). These data were fitted with the Hill equation: I% = (%max·ALb)/(I50%b + ALb), where %max is the total percentage of photoinactivated leaves over the range of light intensities (typically 100%), AL is the actinic light intensity at each step, b is the Hill parameter, and I50% is the light intensity at which 50% of the population of leaves have qPd < 0.98. The I50% was used as a comparison of light tolerance between datasets.
For leaf infiltration experiments, whole leaves were removed and vacuum infiltrated with a control buffer (330 mM sorbitol, 20 mM HEPES, pH 7) either with or without cadmium (1 mM Cd(NO3)2), a PSII donor-side inhibitor. The concentration of Cd2+ was chosen as to specifically affect the PSII donor side, as in previous studies [28,29]. For experiments using UV light, detached leaves were floated on water and exposed to either 10 µmol photons m−2 s−1 of UV-B (peak λ = 312 nm) or white light for 30 min, leaves were then dark-adapted for 15 min. To identify what portion of photoinactivation seen was as a result of direct damage to the OEC, after UV-B treatment, leaves were infiltrated with the earlier control buffer with 1,5-diphenylcarbazide (0.2 mM DPC), a PSII electron donor. DPC has been shown to donate electrons to PSII in the absence of a functional OEC .
To determine the effects of donor-side photoinhibition on the qPd parameter, leaves were treated with the specific donor-side inhibitors, Cd2+ [27–29] and UV-B . Fv/Fm for control and cadmium-treated leaves was 0.810 ± 0.003 and 0.812 ± 0.002. The relationship between qPd, the actual and theoretical yields of PSII, and NPQ for control and cadmium-treated leaves is shown in Figure 1. The cadmium-treated leaves show an earlier decline in the qPd parameter relative to the control leaves, with the final qPd value being 0.854 ± 0.012 in control leaves, and 0.812 ± 0.015 in cadmium-treated leaves (although these values were not significantly different (P > 0.05)).
The relationship between NPQ and Yield PSII/qPd, taken from the 19-step pNPQ procedure.
Figure 2 shows the measure of RC openness and amount of NPQ with respect to the actinic light intensity. The gradient of photoinactivated RCs vs. NPQ and light intensity is indicated. Here, it is clear that the formation of NPQ is inhibited in the cadmium-treated leaves, with total NPQ reaching 1.97 ± 0.07 and 1.612 ± 0.08 (P < 0.01) in both control and cadmium-treated leaves, respectively. However, despite the decline in total NPQ, the gradient of photoinactivation is enhanced in the cadmium-treated sample, even at similar NPQ levels. For instance, after treatment at a light intensity of 124 µmol photons m−2 s−1, there is no significant difference between the NPQ attained (P > 0.05), but the average qPd value is significantly lower being 1.000 ± 0.000 for control leaves (i.e. no photoinactivation) and 0.973 ± 0.001 for cadmium-treated leaves (P < 0.01). Thus, this indicates that the decline in qPd is due to direct weakening of the OEC, rather than due to a diminished NPQ response.
The measure of openness of RCs, represented in three dimensions.
Figure 3 shows population light tolerance curves (see Materials and Methods for more details). Each point on these curves represents the percentage of probed leaves with photoinactivated leaves (i.e. qPd < 0.98) at each corresponding light intensity. Thus, at any light intensity out of the population of leaves (n = 5), the percentage of leaves with qPd < 0.98 are calculated and are shown via the symbols in Figure 3A. Curves have been fitted with the Hill equation, as shown with the black lines. Grey lines show 95% confidence intervals. The I50%, the light intensity at which 50% of the population of leaves display photoinactivated RCs (i.e. qPd < 0.98), for control leaves is 280 ± 6.6 µmol photons m−2 s−1 and for the cadmium-treated leaves is 104 ± 2.6 µmol photons m−2 s−1 (P < 0.001). Overall, cadmium treatment of leaves causes a decrease in the light tolerance of PSII by 62%, with respect to control leaves.
Effect of cadmium on the light tolerance of photosystem II in leaves.
Another well-characterised inhibitor of the OEC is UV-B light , despite this, some have proposed that UV wavelengths are not as highly specific in their locus of damage as other chemical inhibitors [51,52]. To control for this, the PSII-specific electron donor, which donates electrons into the PSII donor side in the absence of a functional OEC , was used. The Fv/Fm was 0.817 ± 0.003 for control leaves, 0.790 ± 0.009 for UV-B-treated leaves, and 0.807 ± 0.003 for UV-B- and DPC-treated leaves. For this dataset, the relationship between qPd, the actual and theoretical yields of PSII, and NPQ for control and cadmium-treated leaves is shown in Figure 4. As with the cadmium-treated leaves, the UV-B-treated leaves show a decline in qPd at much lower NPQ levels. Once again, there are no significant differences between the final qPd in the control, UV-B-treated, or the UV-B- and DPC-treated leaves (P > 0.05). However, the final qPd value is 0.840 ± 0.008 for control leaves, 0.816 ± 0.015 for UV-B-treated leaves, and 0.833 ± 0.013 for the UV-B- and DPC-treated leaves.
The relationship between NPQ and Yield PSII/qPd, taken from the 19-step pNPQ procedure.
Once again, inhibition of the OEC also causes a decrease in the levels of NPQ, as shown in Figure 5, with levels of 2.03 ± 0.03 for the control leaves, and 1.77 ± 0.08 for the UV-B-treated leaves (P < 0.05). Interestingly, the control levels of NPQ are restored once the UV-B-treated leaves were infiltrated with DPC, with NPQ = 2.29 ± 0.21 (P > 0.05 vs. control, P < 0.05 vs. UV-B treated). Once again, despite the changes in the NPQ values, the gradient of photoinactivation is steeper in the leaves with the weakened donor side. For example, after treatment at 124 µmol photons m−2 s−1, there is no significant difference between the NPQ attained (P > 0.05), but the average qPd value is significantly lower being 0.997 ± 0.001 for control leaves (i.e. no photoinactivation) and 0.977 ± 0.006 for UV-B-treated leaves (P < 0.05). As with the cadmium-treated leaves, this shows that the steeper decline in qPd seen is due to the impaired donor side.
The measure of openness of RCs, represented in three dimensions.
Figure 6 shows population light tolerance curves for the control, UV-B-treated, and the UV-B- and DPC-treated leaves (see Materials and Methods for more details). The I50%, the light intensity at which 50% of the population of leaves display photoinactivated RCs (i.e. qPd < 0.98), for the control leaves is 235 ± 2.2 µmol photons m−2 s−1, and for the UV-B-treated leaves is 112 ± 2.6 µmol photons m−2 s−1 (P < 0.001). Similar to the NPQ values, the light tolerance of UV-B-treated leaves is partially restored upon DPC infiltration, with the I50% for this dataset being 169 ± 6.4 µmol photons m−2 s−1 (P < 0.001, with respect to both control and UV-B-treated). The light tolerance of PSII drops by 52% upon UV-B treatment relative to the control, DPC infiltration restores the light tolerance to levels corresponding to 72% of the control tolerance.
Effect of UV light on the light tolerance of photosystem II in leaves.
Currently, the mechanism by which photoinhibition occurs in the RC has generated much discussion, and there is no clear consensus [3,4]. The qPd parameter sensitively monitors the redox state of QA and can accurately estimate subtle percentage changes in the early onset of photoinactivation . Here, the purpose of the study was to assess the impact of donor-side photoinhibition on the measurement of the functional activity and resistance to photoinactivation using the qPd parameter. Tracking the decline of qPd against the onset of NPQ has so far allowed sensitive detection of true photoinactivation of PSII, as shown by the sensitivity of its recovery to lincomycin , and the linear correlation between the decline of qPd and the decline of oxygen evolution .
qPd estimates photoinactivation through the divergent change between Fo′act and Fo′calc, causing qPd < 1. The rise in Fo′act, with respect to Fo′calc, is caused by a stabilised QA− state at the acceptor side of PSII, therefore sensitively monitoring acceptor-side photoinhibition directly [18,48]. This raises important questions with regard to the validity of qPd as a true measure of photoinactivation, especially as there is some evidence of the OEC being a primary locus of PSII inhibition, particularly under blue and UV wavelengths [21,32]. Using the chemical OEC inhibitor, Cd2+ [27–29], oxygen evolution can be impaired through competition between Cd2+ and Ca2+ in the high-affinity site within the OEC . Whereas, UV-B light causes direct damage to the thylakoid and electron transfer cofactors within . Through the use of both Cd2+ and UV-B, the function of the OEC can be weakened, and the effects on the qPd parameter can be seen. We have shown, using these independent methods, that destabilisation of the OEC causes a decline in the qPd parameter, with a corresponding decline in the light tolerance of PSII. This is due to the ability of the equation of Oxborough and Baker  to sense changes not only at the level of differences in the quenching of Fo′act, but in Fm′ also, which is quenched regardless of inhibitory mechanism . The mechanism-dependent changes in Fm and Fo have been summarised in Figure 7. In essence, as protective NPQ resides in the PSII light-harvesting antenna (LHCII) , quenching of Fm′ in the antenna competes with the RC, when the RCs are open. The open RCs cause a decrease in Fo as long as the PSII acceptor side is without photoinhibition. However, as acceptor-side photoinhibition occurs, there is a well-characterised increase in Fo [18,48]. The formula of Oxborough and Baker  takes this into account; therefore, there is a readily detectable decline in qPd. However, donor-side photoinhibition does not cause any significant change in Fo′act . Under these conditions, alongside the contribution of protective NPQ, there is a much greater quenching of Fm, and Fv (calculated as Fm − Fo) disappears. This leads to an underestimation of Fo′calc. Hence, as qPd detects disparity between Fo′act and Fo′calc, donor-side photoinhibition-mediated underestimation of Fo′calc will cause a divergence between Fo′act and Fo′calc, which will also manifest in the decline of qPd. Hence, as acceptor-side photoinhibition causes a rise in Fo′act and donor-side photoinhibition causes underestimation of Fo′calc, there will be a disparity between Fo′act and Fo′calc in both cases and will be sensitively detected by qPd.
Schematic representation of changes in chlorophyll fluorescence parameters under acceptor-side photoinhibition, donor-side photoinhibition, and NPQ states.
Therefore, our results show that qPd is capable of detecting true photoinactivation whether by solely acceptor-side photoinhibition [18,20], a ‘statistical’ donor-side mechanism , or an Mn-dependent inhibitory mechanism , due to the sensing of inhibitory changes at both the level of Fo′ and Fm′. It is also worth noting that it is highly unlikely for any of these mechanisms to occur in isolation within in vivo systems , and that even conditions that predominantly favour donor-side inhibition are likely to have a photosensitisation effect at the acceptor side of PSII, in agreement with work done by Murata and co-workers . As shown in this study, all of these effects on PSII will be detected holistically within the qPd parameter. This is particularly shown within the experiments using UV-B light; light tolerance here is only partially restored by the PSII electron donor DPC. Despite strong evidence of UV-B as a key inhibitor of the OEC [3,21], the OEC is not the only locus of UV-B damage, but the PSII acceptor side also [51,52]. We have shown here that both acceptor- and donor-side damage can be accounted for with qPd, and using DPC, the balance between the two can be seen also.
As a result of the weakened PSII donor side, it is also shown here that NPQ is diminished under both Cd2+ and UV-B treatment. Within the UV-B experiments, control levels of NPQ are restored upon addition of DPC. As water oxidation in PSII is one key factor responsible for acidification of the thylakoid lumen , it is likely if proton translocation around PSII was disrupted, there would also be a weaker NPQ response, as the formation of ΔpH is a key trigger for the qE scenario . Restoration of NPQ upon DPC addition, whilst unlikely to boost ΔpH at the level of PSII, would likely restore native electron transport, allowing normal or increased functionality of downstream proton pumps . Furthermore, when acclimated to UV-B, the green alga Chlamydomonas reinhardtii has been shown to have an up-regulated qE response, due to greater amounts of PsbS . The PsbS up-regulation here could be in response to the diminished lumen acidification, or as a mitigating factor against the sensitisation of the acceptor side as a result of a weakened donor side, as seen in our study and in others .
Overall, our results clearly show that by using two independent, established methods to inhibit electron transfer from the PSII donor side, that any damage here can be readily accounted for using the qPd chlorophyll fluorescence parameter. Therefore, this study confirms qPd as a robust and unambiguous parameter for functional, physiological measurements of photoinactivation via both acceptor- and donor-side photoinhibition.
A.V.R. and S.W. designed the experiments, and S.W. acquired and analysed the data, and wrote the manuscript.
This work was supported by The Royal Society Wolfson Research Merit Award, a BBSRC grant [BB/L019027/1] to A.V.R. and Queen Mary Principal's research studentship to S.W.
The Authors declare that there are no competing interests associated with the manuscript.