Phylloquinone is the principal Mehler reaction site within photosystem I in high light

Abstract

Photosynthesis is a vital process, responsible for fixing carbon dioxide, and producing most of the organic matter on the planet. However, photosynthesis has some inherent limitations in utilizing solar energy, and a part of the energy absorbed is lost in the reduction of O₂ to produce the superoxide radical (⁠${\mathrm{O}}_2^{•-}$⁠) via the Mehler reaction, which occurs principally within photosystem I (PSI). For decades, O₂ reduction within PSI was assumed to take place solely in the distal iron–sulfur clusters rather than within the two asymmetrical cofactor branches. Here, we demonstrate that under high irradiance, O₂ photoreduction by PSI primarily takes place at the phylloquinone of one of the branches (the A-branch). This conclusion derives from the light dependency of the O₂ photoreduction rate constant in fully mature wild-type PSI from Chlamydomonas reinhardtii, complexes lacking iron–sulfur clusters, and a mutant PSI, in which phyllosemiquinone at the A-branch has a significantly longer lifetime. We suggest that the Mehler reaction at the phylloquinone site serves as a release valve under conditions where both the iron–sulfur clusters of PSI and the mobile ferredoxin pool are highly reduced.

Introduction

Under high irradiance, oxygenic phototrophs use molecular oxygen (O₂) as an alternative sink for surplus electrons within the photosynthetic apparatus (Curien et al., 2016). Unfortunately, uncontrolled leakage of electrons to O₂ could decrease the quantum yield of photosynthesis and produce deleterious reactive oxygen species. As a result, the photosynthetic apparatus appears to have evolved strong regulation of such reactions in an aerobic environment (Kozuleva et al., 2020).

In 1951, Alan Mehler reported that under illumination, spinach (Spinacia oleracea) thylakoid membranes produced hydrogen peroxide (H₂O₂) via the reduction of dioxygen, that is, O₂ served as a Hill reactant (Mehler, 1951). Since then, O₂ photoreduction by the photosynthetic electron transfer (ET) chain with the production of H₂O₂ is called the Mehler reaction. Later, the superoxide radical (⁠${\mathrm{O}}_2^{•-}$⁠) was shown to be the primary product of this reaction (Allen and Hall, 1973). Despite the long history of the issue, there is still no consensus about the principal sites generating ${\mathrm{O}}_2^{•-}$ in the photosynthetic electron transport (ET) chain under different conditions. Some evidence indicates that photosynthetic ${\mathrm{O}}_2^{•-}$ production can take place in photosystem II (PSII; Pospíšil, 2012), the plastoquinone pool (Khorobrykh and Ivanov, 2002), cytochrome b₆/f complex (Baniulis et al., 2013), ferredoxin (Fd; Misra and Fridovich, 1971; Allen, 1975a; Kozuleva et al., 2016), Fd-NADP⁺ oxidoreductase (FNR; Miyake et al., 1998), and PSI. There is a common agreement, which is based on direct experimental evidence (Fork and Heber, 1968; Takahashi and Asada, 1988; Hormann et al., 1993) as well as on indirect considerations (Ogawa et al., 1995; Kozuleva et al., 2011), that PSI is the major site of ${\mathrm{O}}_2^{•-}$ generation in chloroplasts (Asada, 1999; Badger et al., 2000; Rutherford et al., 2012; Baniulis et al., 2013; Kozuleva et al., 2020).

The PSI complex mediates a light-driven charge separation that is accompanied by plastocyanin (Pc) oxidation and Fd reduction (Figure 1A). Two membrane subunits, PsaA and PsaB, together with an extrinsic subunit PsaC coordinate the ET cofactors comprising six chlorophyll a (Chl) molecules, two phylloquinone (PhQ) molecules, and three [4Fe–4S] clusters. With the exception of the [4Fe–4S] clusters, cofactors are organized within two active asymmetric branches named A- and B- (Guergova-Kuras et al., 2001). One of the most critical differences between branches is related to the redox properties of PhQs, where a higher potential quinone is located in the A-branch, while the B-branch has a lower potential quinone (Figure 1A). Notably, the environment influences the properties of PhQs, so that the reduction potentials and lifetimes of the phyllosemiquinones (PhQ^•−) are modulated via the identity of nearby amino acids. The most drastic effect published so far was due to the replacement of PsaA-Phe₆₈₉ with Asn (PsaA-F₆₈₉N mutation), which resulted in an approximately two orders of magnitude longer lifetime of the PhQ^•− in the A-branch (⁠${\text{PhQ}}_{\mathrm{A}}^{•-}$⁠), from ∼0.25 to 17 µs (Santabarbara et al., 2015), probably as a result of a ∼125-mV increase in the PhQ_A/${\text{PhQ}}_{\mathrm{A}}^{•-}$ reduction potential (Figure 1B).

Figure 1

Diagram of electron transfer in PSI with lifetimes and mid-point redox potentials (E_m) of cofactors. A, For the WT according to Ptushenko et al. (2008). B, For PsaA-F₆₈₉N mutant taking into account for the effect of mutation on the E_m of PhQ_A/${\text{PhQ}}_{\mathrm{A}}^{•-}$ according to Santabarbara et al. (2015); blue lines represent E_m of O₂/${\mathrm{O}}_2^{•-}$ in water (−160 mV) and dimethylformamide (−500 to −600 mV; Wardman, 1990). P₇₀₀, the dimer of Chl a molecules in PSI; A₀, the primary electron acceptor in PSI; PhQ, phylloquinone, a secondary electron acceptor in PSI; F_X, a 4Fe–4S cluster, a secondary electron acceptor in PSI; F_A and F_B, 4Fe–4S clusters, the terminal electron acceptors in PSI; Fd, ferredoxin, the mobile electron acceptor. Modified from Kozuleva and Ivanov (2016).

Photoreduction of O₂ to ${\mathrm{O}}_2^{•-}$ in PSI is generally assumed to occur in the F_A/F_B clusters. This common belief stems from the fact that F_A/F_B are the terminal cofactors in isolated thylakoid membranes, and could therefore reduce O₂ under steady-state illumination in the absence of Fd (Asada et al., 1974; Allen, 1975b; Furbank and Badger, 1983; Kozuleva and Ivanov, 2010). However, the primary site of O₂ photoreduction in thylakoid membranes was shown to be the heterodimer PsaA/PsaB, rather than PsaC carrying F_A/F_B clusters (Takahashi and Asada, 1988). A contribution of PhQs to the Mehler reaction was first suggested by Kruk et al. (2003) and the first experimental evidence for the involvement of PhQs in O₂ photoreduction under steady-state illumination was provided by Kozuleva et al. (2014). However, the contribution of PhQ in the A and B branches, as well as the contribution of [4Fe–4S] clusters, especially the intermediate F_X cluster, are still open questions. The mechanisms underlying ${\mathrm{O}}_2^{•-}$ production within PSI occurring concomitantly with ET to Fd and then to NADP⁺ under varying irradiance are also unclear.

The aim of the present study was to identify where and how one-electron photoreduction of O₂ occurs within PSI. Toward that goal, we measured the apparent second-order rate constant of O₂ photoreduction by PSI complexes (k₂) as a function of irradiance. We investigated intact fully mature PSI complexes isolated from a model organism, the green alga Chlamydomonas reinhardtii, PSI complexes lacking either F_A/F_B (F_X-core), or all three [4Fe–4S] clusters (A₁-core), and PSI complexes harboring the PsaA-F₆₈₉N mutation. Our results strongly support the conclusion that PhQs, especially PhQ_A, are the major site of O₂ photoreduction within PSI under high irradiance.

Results

Under continuous illumination, O₂ photoreduction is the rate-limiting step of electron transfer

The identification of the sites of O₂ photoreduction within PSI under steady-state illumination requires that the step of ET to O₂ from the PSI is the rate-limiting step of the overall process of ET from the ultimate donor to O₂. This can be accomplished by providing an efficient electron donor to ${\mathrm{P}}_{700}^{+}$⁠. Here, we used excess Pc to reduce ${\mathrm{P}}_{700}^{+}$ in fully mature intact PSI complexes isolated from C. reinhardtii. It was important to use mature PSI complexes, since immature complexes are deficient in the PsaF subunit (Marco et al., 2018), which is required for proper binding of Pc (Hippler et al., 1996). We verified that Pc was saturating using our previously described method (Petrova et al., 2018). Since increasing the concentration of Pc added to PSI did not affect the O₂ photoreduction rate in the absence of other acceptors (Figure 2, squares), we conclude that the O₂ photoreduction is indeed the rate-limiting step of the overall reaction. The same result was observed in our system when Pc was replaced by N,N,N′,N′-tetramethyl-p-phenylenediamine (TMPD) as an electron donor (Supplemental Figure S1). Although the rate constant for TMPD (3.0 × 10⁴ M⁻¹ s⁻¹; Fujii et al., 1990) is approximately four orders of magnitude lower than that of Pc, the rates of O₂ photoreduction were the same. In contrast, when methyl viologen (MV), an efficient redox mediator from PSI to O₂, was added to intact PSI complexes, the rate of O₂ photoreduction did increase with increasing Pc concentration (Figure 2, circles). Thus, we can conclude that the electron donation from Pc to ${\mathrm{P}}_{700}^{+}$ was the rate-limiting step only in the presence of efficient electron acceptors, such as MV.

Figure 2

Effect of Pc concentration on the O₂ photoreduction rate in the absence or the presence of MV. Mature WT PSI complexes from C. reinhardtii, 15 nM; Asc, 10 mM; initial O₂ concentration, 250 µM; irradiance, 2,200 µmol photons m⁻² s⁻¹; where indicated MV, 50 µM. Values are mean for three to four replicates; ±se.

Our working hypothesis is that under steady-state illumination, there are multiple oxygen photoreduction sites in PSI, each catalyzing a mechanistically independent reaction associated with a discrete rate constant. The accumulation of electrons on these sites changes in response to irradiance, due to the competition between forward electron transfer to the site (driven by light) and electron transfer from the site either to the next cofactor or back to ${\mathrm{P}}_{700}^{+}$ (charge recombination). In other words, each site reaches maximal activity at a different specific irradiance. Therefore, if the apparent second-order rate constant of O₂ photoreduction by PSI (k₂) is not the same at all light levels, but is rather a function of irradiance, it means that multiple O₂ photoreduction sites are involved. If there are few electron acceptors, electrons will be accumulated in the system, with the terminal sites filling up first. Hence, electrons will start to accumulate at the F_A/F_B centers under low irradiance and then, with increasing irradiance, they will begin to occupy the F_X site and then A₁ sites.

Determination of the second-order rate constant of O₂ photoreduction in PSI

The exact rate constant of the Mehler reaction is not currently known and estimates for the kinetic parameter for O₂ photoreduction by PSI range from ∼10³ M⁻¹ s⁻¹ (Khorobrykh and Tyystjärvi, 2018) to 10⁷ M⁻¹ s⁻¹ (Asada and Nakano, 1978). The median value (7 × 10⁴ M⁻¹ s⁻¹) was obtained by kinetic modeling of charge recombination kinetics observed after a light flash (Milanovsky et al., 2017). We measured the apparent k₂ for intact fully mature PSI complexes purified from the wild-type (WT) strain (PSI^WT) of C. reinhardtii, under conditions where O₂ was the principal electron acceptor. In this case, k₂ was denoted as k₂^°2. The apparent k₂^°2 varied slightly between different PSI preparations and was estimated to be ∼3.5 × 10⁵ M⁻¹ s⁻¹ at the highest irradiance tested (2,200 µmol photons m⁻² s⁻¹). Figure 3A shows the dependency of the apparent k₂^°2 on irradiance as measured in a typical PSI preparation. The apparent k₂^°2 clearly increased with irradiance up to 2,200 µmol photons m⁻² s⁻¹. The O₂ photoreduction rate measured at atmospheric pressure of O₂, that is, at the saturating O₂ level, as a function of irradiance resembles that of k₂^°2 (Figure 3A, dashed line), as expected.

Figure 3

Effect of irradiance on the apparent rate constant of O₂ photoreduction (k₂) by fully mature intact PSI complexes from WT strain of C. reinhardtii. A, In the absence of MV; the rate of O₂ consumption observed at atmospheric pressure of O₂ (dashed line) is shown for comparison. B, In the presence of MV at 50 µM (dark blue line, closed circles) or at 4 µM (light blue line, open circles); the curve observed in the absence of MV from (A) (green line, closed squares) is shown for comparison. Mature WT PSI complexes, 10 nM; Asc, 10 mM; catalase, 500 U mL⁻¹; Pc, 5 µM; initial O₂ concentration, 10–30 µM. Values are mean for 7–10 replicates; ±se.

Saturating concentrations of MV de facto produce a single O₂-reducing site

To test the photoreduction of O₂ in the presence of an auto-oxidizable PSI electron acceptor, we added MV at different concentrations. In the presence of MV at a saturating concentration (50 µM), the F_A/F_B clusters can be oxidized completely, which should suppress the direct O₂ photoreduction by PSI. In this case, the measured constant of O₂ photoreduction was denoted as k₂^MV, since MV was the principal electron acceptor from PSI and was constantly recycled by reduction of O₂. Indeed, the apparent k₂ and O₂ uptake rates were higher in the presence of MV (Figures 2, 3B), reflecting higher PSI turnover. In contrast to behavior of k₂^°2, the values of the apparent k₂^MV reached a plateau at 220 µmol photons m⁻² s⁻¹ (Figure 3B, dark blue circles), demonstrating that in the presence of saturating MV, there is de facto a single O₂ photoreduction site. In order to negate the possibility that the saturation behavior of k₂^MV at 50-µM MV is due to limitations in re-reduction of ${\mathrm{P}}_{700}^{+}$⁠, we replaced Pc by TMPD, a less efficient electron donor than Pc. In the presence of TMPD, the apparent k₂^MV was lower but still reached a plateau at the same irradiance (Supplemental Figure S2), confirming that the irradiance dependence of the apparent k₂^MV is not influenced by processes taking place at the donor side of PSI.

In order to mimic the situation in which O₂ photoreduction could be carried out by both PSI and its acceptor, MV was added at 4 µM, which is near the K_M(MV) value determined in separate experiments (Supplemental Figure S3). In this condition, k₂^MV was lower than at 50-µM MV and increased with increasing irradiance, but it did not reach a plateau (Figure 3B, light blue circles). In this regard, the behavior of k₂^MV at 4-µM MV resembled that of k₂^°2 (Figure 3B, green squares). Since O₂ reduction by MV would account for only about half of the maximal rate when present at concentrations around the K_M, we conclude that electrons can still accumulate at PSI cofactors in this condition and are available for direct reduction of O₂.

Stripping PSI of its [4Fe–4S] clusters enhances O₂ photoreduction at high light

The hypothesis that there are multiple sites within PSI for O₂ photoreduction (i.e. F_X, PhQ_A, and PhQ_B) can be tested by sequential removal of the [4Fe–4S] clusters. This should remove the competition for electrons and favor reduction of O₂ by upstream cofactors. Initially, the PsaC subunit containing the F_A/F_B clusters was removed, resulting in a PSI complex in which the terminal cofactor is F_X (F_X-core complexes). The behavior of the apparent k₂^°2 in the F_X-core complexes was somewhat similar to that of the intact complexes (Figure 4A, yellow circles versus green squares), especially at high irradiance. Removal of PsaC actually decreased O₂ reduction at low irradiances, which supports the contention that the F_A/F_B clusters play a role in O₂ reduction in low light. This makes sense, as one would expect accumulation of electrons primarily on the F_A/F_B clusters at low irradiances. However, the lack of effect upon O₂ photoreduction at high irradiances indicates that the F_A/F_B clusters are not responsible for the additional O₂ photoreduction by intact PSI seen in these conditions. One would expect that in the condition of high light and saturating electron donation to ${\mathrm{P}}_{700}^{+}$⁠, the preceding cofactor (F_X cluster) would accumulate electrons. If the F_X cluster was responsible for the additional O₂ photoreduction at high irradiances, then this should be unleashed at lower irradiances after removal of the F_A/F_B clusters. The fact that this was not observed seems to disprove the hypothesis that F_X is responsible for much (or all) of the additional O₂ photoreduction at high irradiances.

Figure 4

Effect of irradiance on the apparent rate constant of O₂ photoreduction (k₂) by PSI preparations from C. reinhardtii challenging PhQs contribution to O₂ photoreduction. A, F_X- and A₁-core complexes; B, The PsaA-F₆₈₉N mutant; the curve observed for intact WT PSI complexes in the absence of MV from Figure 3A (green line, closed squares) is shown on both panels for comparison. The F_X-core, A₁-core, PSI^AFN complexes, 10 nM; TMPD, 1 mM (A), or Pc, 5 µM (B); Asc, 10 mM; catalase, 500 U mL⁻¹; initial O₂ concentration, 10–30 µM. Values are mean for 7–10 replicates; ±se.

In order to test that hypothesis more directly, we next removed the F_X cluster, leaving the bound PhQ molecules at the A₁-sites as the terminal cofactors (A₁-core complexes). Increasing the concentration of electron donor did not affect the rate of O₂ photoreduction when all the [4Fe–4S] clusters were removed (Supplemental Figure S4), although the basal rate was higher than in intact PSI complexes (Figure 4A). This shows that the ET to O₂ is still the rate-limiting step of the overall ET. In contrast to the results obtained with the F_X-core, the A₁-core complexes showed markedly higher rates of O₂ photoreduction at irradiances above 350 µmol photons m⁻² s⁻¹ (Figure 4a, blue triangles versus green squares). Interestingly, O₂ photoreduction in the A₁-core complexes in low light is faster than in the F_X-core complexes. Taken together, these experiments definitely demonstrate that F_X is a poor reductant of O₂. Moreover, none of the [4Fe–4S] clusters is essential for photoreduction of O₂ by PSI.

Furthermore, the results with the F_X-core and A₁-core complexes indicate that the PhQs (presumably in the semiquinone state) are the moieties responsible for O₂ photoreduction in intact PSI in high light. It should be noted, however, that in intact PSI under low light, phyllosemiquinone would be present in very low amounts at steady state, given the high rate of forward ET from PhQ to F_X/F_A/F_B (Byrdin et al., 2006). Thus, we hypothesize that in intact PSI the F_A/F_B clusters are the principal site(s) of O₂ reduction at low light, but as irradiance increases the PhQs become the principal sites of O₂ reduction.

Dramatically enhanced O₂ photoreduction by a mutant PSI with a longer ${\text{PhQ}}_{\mathrm{A}}^{•-}$ lifetime

The removal of the [4Fe–4S] clusters was an artificial way to accumulate semiquinone in PSI. In order to test our hypothesis in intact PSI, it would be necessary to slow ET from PhQ to F_X, thus allowing greater accumulation of phyllosemiquinone at steady state. Accordingly, we made use of the PsaA-F₆₈₉N mutant, in which ${\text{PhQ}}_{\mathrm{A}}^{•-}$ has a lifetime about two orders of magnitude longer than in the WT (Santabarbara et al., 2015). The behavior of charge recombination kinetics in PSI complexes from this mutant (PSI^AFN) more closely resembled that of WT A₁-core complexes than intact PSI^WT (Supplemental Figure S5; Petrova et al., 2021), with a lifetime on the order of hundreds of microseconds rather than the tens of milliseconds. The faster charge recombination kinetics have been seen previously in other mutants that stabilize ${\text{PhQ}}_{\mathrm{A}}^{•-}$ (Boudreaux et al., 2001). We used mature PSI^AFN complexes capable of reducing NADP⁺ in the presence of Fd and FNR under steady-state illumination (Supplemental Table S1), demonstrating that they retained all of the [4Fe–4S] clusters. However, the rates of NADP⁺ reduction in PSI^AFN complexes were two to five times lower than in PSI^WT. Moreover, the rate of O₂ uptake in the presence of MV was much lower in PSI^AFN than in PSI^WT complexes (Supplemental Figure S6). We can explain this by a diminished steady-state population with reduced F_A/F_B; thus, fewer PSI^AFN complexes would be available to reduce Fd or MV at a given light intensity than PSI^WT.

Our hypothesis predicted faster rates of O₂ reduction in intact PSI^AFN complexes than in PSI^WT complexes, due to greater steady-state amounts of ${\text{PhQ}}_{\mathrm{A}}^{•-}$⁠. Indeed, at intensities above 350 µmol photons m⁻² s⁻¹, the values of the apparent k₂^°2 in intact PSI^AFN were markedly higher than those of intact PSI^WT (Figure 4B, pink circles versus green squares) and the curve resembled that of WT A₁-core complexes (Figure 4a, blue triangles) more than intact PSI^WT. Thus, we can state with confidence that ${\text{PhQ}}_{\mathrm{A}}^{•-}$ is an effective reducer of O₂. Moreover, these results deal another blow to the hypothesis that only the [4Fe–4S] clusters can reduce O₂, as the lower steady-state amount of (F_A/F_B)⁻ in PSI^AFN complexes due to rapid charge recombination could only have decreased the O₂ photoreduction rate; this hypothesis cannot explain the increase at higher irradiances. Finally, these experiments should dispel concerns that the results with the A₁-core complexes were merely due to unspecified damage or conformational changes caused by the removal of the [4Fe–4S] clusters, as the intact PSI^AFN complexes retained all [4Fe–4S] clusters.

O₂ photoreduction does not compete with NADP⁺ reduction

One might question the relevance of these in vitro results to the situation in vivo, where electron acceptors, such as Fd are available to PSI. In order to assess the impact of such acceptors upon O₂ photoreduction by PSI^WT, we repeated the experiment with the addition of Fd, FNR, and NADP⁺, that is, under conditions, which simulate the situation in vivo. The presence of FNR and excess NADP⁺ provided an efficient electron sink for Fd, as evidenced by the measured NADP⁺ photoreduction rate (Figure 5B). Furthermore, Fd also contributes to O₂ reduction, especially in the absence of NADP⁺. Hence, the apparent k₂ in the presence of Fd (k₂^Fd) should be the sum of k₂ for ET from PSI directly to O₂ and indirectly via Fd to O₂. Indeed, in the presence of Fd alone, the values of the apparent k₂^Fd were higher than k₂^°2 (Figure 5A), suggesting a strong contribution of Fd to the total O₂ reduction under such conditions. However, in the presence of FNR and NADP⁺, we observed similar rates of O₂ photoreduction at high irradiance as in the absence of Fd/FNR/NADP⁺ (Figure 5A, violet circles versus green squares). We interpret this as a decrease in Fd-dependent O₂ reduction due to rapid electron flow from Fd to NADP⁺ lowering the steady-state amount of reduced Fd. Interestingly, the presence of Fd/FNR/NADP⁺ slows down O₂ photoreduction at low irradiance, indicating that NADP⁺ reduction decreases the accumulation of electrons on the F_A/F_B clusters, but has little effect at high irradiance. The similarity in behavior of k₂^Fd in the presence of FNR/NADP⁺ and k₂^°2 at high irradiance strongly suggests that PhQs at the A₁ sites will represent the major site of O₂ photoreduction in PSI in vivo under high light.

Figure 5

Effect of irradiance on the apparent rate constant of O₂ photoreduction (k₂) by fully mature WT PSI complexes from C. reinhardtii concomitantly reducing NADP⁺. A, Effect of irradiance on apparent k₂ in the presence of Fd (orange triangles) or Fd, FNR, NADP⁺ (purple circles); the curve observed for intact WT PSI complexes in the absence of Fd from Figure 3A (green squares) is shown for comparison; B, Effect of irradiance on the rate of NADP⁺ reduction. Mature WT PSI complexes, 10 nM; Pc, 5 µM; Asc, 10 mM; catalase, 500 U mL⁻¹; Fd, 5 µM; where indicated: FNR, 200 nM, and NADP⁺, 1 mM; initial O₂ concentration, 10–30 µM. Values are mean for 7–10 replicates for k₂ and five replicates for rates; ±se.

Discussion

F_A/F_B is not the sole O₂ photoreduction site within PSI, nor is it the major site in high light

In this study, we measured the rates of O₂ photoreduction by various PSI complexes and demonstrated that the apparent rate constant of this reaction depends on irradiance (Figure 3A). This observation could be explained by either (1) the presence of a single O₂ reduction site (i.e. F_A/F_B), that attains full activity under high irradiance, or (2) the presence of multiple O₂ reduction sites within PSI, which are activated under different light regimes. Our data are much more consistent with the latter model.

The first piece of evidence in favor of the multisite model is the shape of the curve itself (Figure 3A). If there was a single O₂ reduction site in PSI, the O₂ photoreduction curve would exhibit saturation as irradiance increased. Instead, the curve looks more like the additive effect of sites with different light saturation behaviors. The observation that the removal of PsaC slows the O₂ reduction rate at low irradiances in consistent with F_A/F_B being the major site for the process under low light (Figure 4A), consistent with previous expectations.

The results obtained following MV addition to the intact PSI complexes (Figure 3B) demonstrate that there is unlikely to be a single O₂ photoreduction site in PSI. The rate constant of the reaction between the reduced MV and O₂ is 8 × 10⁸ M⁻¹ s⁻¹ (Farrington et al., 1973), which is close to diffusion limit and almost two orders of magnitude higher than the rate constant of MV reduction by PSI (10⁷ M⁻¹ s⁻¹ (Hiyama and Ke, 1971)). The fast recycling of MV by O₂ makes it such an efficient electron acceptor from PSI and enable us to use MV to simulate the behavior of a single site of O₂ reduction. With saturating concentrations of MV, k₂^MV reached a plateau at an irradiance of 220 µmol photons m⁻² s⁻¹. As MV accepts electrons exclusively from F_A/F_B in intact PSI, reduction of O₂ by F_A/F_B should saturate at that light intensity. In other words, if k₂ increases any further as irradiance increases in the absence of MV, it would indicate reduction of O₂ by upstream cofactor(s). Moreover, the contribution of upstream cofactor(s) to O₂ reduction must depend on the rate of electron withdrawal from PSI by available acceptors, since a k₂^MV plateau was no longer observed when lower concentrations of MV were provided (Figure 3B).

PhQ_A is a potential major site of O₂ photoreduction in PSI under high irradiance

In order to localize the site(s) of O₂ photoreduction in PSI, we removed the [4Fe–4S] clusters sequentially. Experiments with stripped complexes simulate the situation in intact PSI when forward ET is slowed and the electron flow to O₂ competes primarily with charge recombination. The lifetime of reduced F_X is longer in the F_X-core complexes since the half-time of charge recombination from F_X is 1 ms (Brettel and Leibl, 2001), while the half-time of forward ET from F_X to F_A/F_B in intact PSI is < 50 ns. This should increase the probability of O₂ reduction by F_X in the F_X-core complexes. Another factor that affects the probability of O₂ photoreduction is the change in the redox potential. The removal of F_A/F_B slightly increases the E_m of F_X in the F_X-core complexes (Ishikita et al., 2006); however, it is still more negative than that of F_A/F_B in intact PSI complexes (Figure 1). Therefore, the driving force of the reaction between F_X and O₂ should still be greater than that between F_A/F_B and O₂. However, despite all these factors favoring O₂ reduction by F_X in the F_X-core complexes, stripping PSI of F_A/F_B lowered O₂ photoreduction rates at low irradiances and had little effect at high irradiances (Figure 4A), which can be explained by a minor role for the F_X cluster in O₂ reduction.

In the A₁-core complexes, the half-time of charge recombination from PhQs is ∼10–100 µs, which is much greater than the half-time of direct ET from PhQs to F_X in intact PSI (20–250 ns). Therefore, the lifetime of phyllosemiquinones in the A₁-core complexes is longer than that of in intact PSI and F_X-core complexes. The removal of [4Fe–4S] clusters could cause significant structural changes, which might change the local concentration of O₂ near PhQs in the A₁-core complexes and increase E_m of PhQs in the A₁-core complexes. The later should decrease the probability of O₂ photoreduction by PhQs in the A₁-core complexes. In our experiments, the additional removal of F_X resulted in a substantial increase in k₂^°2 values in high light (Figure 4A) that could be explained by longer lifetime of phyllosemiquinones.

The experiments using the mature intact PSI particles from PsaA-F₆₈₉N mutant, in which the lifetime of ${\text{PhQ}}_{\mathrm{A}}^{•-}$ is increased from ∼0.2 to ∼17 µs (Santabarbara et al., 2015), were aimed to increase the steady-state level of PhQ^•− without removal of [4Fe–4S] clusters. At high irradiances, the k₂^°2 values for PSI^AFN complexes were significantly larger than those measured for the PSI^WT complexes (Figure 4B). Note that the competing hypothesis—that [4Fe–4S] cluster(s) are the primary site(s) of O₂ reduction at high irradiance—would have predicted the opposite result, as the population with reduced [F_A/F_B] clusters is diminished in this mutant due to the ∼1,000-fold increase in charge recombination rate (Supplemental Figure S5). This is consistent with our observation of a two to five times lower rate of NADP⁺ photoreduction under steady-state illumination (Supplemental Table S1) as well as a decrease in the MV-driven O₂ reduction rates (Supplemental Figure S6) by PSI^AFN than by PSI^WT complexes.

Our interpretation of the results with PSI^AFN complexes is that the longer lifetime of ${\text{PhQ}}_{\mathrm{A}}^{•-}$ increases the chance of this semiquinone reacting with O₂. Santabarbara reported that the lifetime of ${\text{PhQ}}_{\mathrm{B}}^{•-}$ in the PsaA-F₆₈₉N mutant is the same as in WT (Santabarbara et al., 2015), which is about an order of magnitude shorter than the lifetime of ${\text{PhQ}}_{\mathrm{A}}^{•-}$ in WT, consistent with theoretical computations of their potentials (Ptushenko et al., 2008; Kawashima and Ishikita, 2017). There is an interesting interplay between reduction potential and lifetime that should be considered. As the potential of the PhQ/PhQ^•− couple is raised, the lifetime of the semiquinone increases, but it also becomes a poorer electron donor. From our experiments with A₁-core complexes and PSI^AFN complexes, the effect upon lifetime appears to be more important for O₂ photoreduction, assuming that the potential is not raised so high that it exceeds that of the O₂/${\mathrm{O}}_2^{•-}$ couple. Considering all of these points, we hypothesize that ${\text{PhQ}}_{\mathrm{A}}^{•-}$ is the principal site for O₂ photoreduction in high light. However, the asymmetry of charge separation between A and B branches, redox potential of quinones, and other factors should be taken into account to elucidate the role of ${\text{PhQ}}_{\mathrm{B}}^{•-}$ into O₂ reduction.

One might be surprised by the reaction of O₂ with a cofactor buried within an integral membrane protein. The presence of water cavities in close vicinity of the A₁ sites in cyanobacterial PSI (Milanovsky et al., 2017) may provide a pathway for diffusion of O₂ and ${\mathrm{O}}_2^{•-}$⁠. It should also be kept in mind that O₂ is a fairly nonpolar molecule and has in fact been found to partition from aqueous solutions into liposomes with a partition coefficient of ∼4 (Möller et al., 2005). Moreover, computational simulations suggest that small apolar molecules like O₂ can penetrate proteins fairly easily, making use of the proteins’ internal thermal motions (Cohen et al., 2006). Thus, the reaction of O₂ with a semiquinone buried in a hydrophobic protein should not be difficult.

O₂ photoreduction by PhQs is concomitant with NADP⁺ reduction

The physiological significance of the Mehler reaction in angiosperms is well recognized (Makino et al., 2002; Ort and Baker, 2002; Pérez‐Torres et al., 2007). In other green lineages the role of the H₂O₂-producing Mehler reaction is more vague due to the high significance of other O₂-reducing pathways with no H₂O₂ produced (flavodiiron proteins, terminal plastid oxidase). Nevertheless, the accumulation of H₂O₂ in the C. reinhardtii chloroplasts was shown to occur when cells were subjected to high light (Roach et al., 2015). As the composition of the ET cofactors in PSI is highly conserved through all green lineages, we suggest that O₂ photoreduction by PhQs shown here for PSI complexes from C. reinhardtii represents a universal mechanism of the Mehler reaction in PSI.

We tested our hypothesis about PhQ role in the Mehler reaction under conditions simulating physiological conditions, that is, when Fd and NADP⁺ reduction takes place. The addition of Fd alone significantly increased the rates of O₂ reduction (Figure 5A). This imitates the situation in vivo when the stromal NADP⁺ pool is in the reduced state (e.g. due to a retarded Calvin cycle at limiting CO₂). However, since we used Fd at much higher level relative to PSI than present in vivo (500:1 ratio versus 5:1 ratio in spinach chloroplast; Böhme, 1978), care should be taken when extrapolating this result to the in vivo situation. It is likely that the addition of FNR and NADP⁺ minimized the Fd-dependent O₂ reduction by decreasing the accumulation of reduced Fd. These observations are in good agreement with previously reported conclusions (Kozuleva and Ivanov, 2010; Kozuleva et al., 2016). The results of our experiments indicated saturation of NADP⁺ photoreduction at high irradiances (Figure 5B). This provides insights into the in vivo situation, as these are the conditions when Mehler reaction plays a physiological role in protecting the photosynthetic apparatus from overreduction of the ET chain and when the PhQs pass electrons to O₂. This suggests a crucial role for PhQs, and specifically PhQ_A (see above), in O₂ photoreduction in PSI in vivo in high light.

Conclusion

In summary, our results demonstrate that PhQs at the A₁-sites (primarily PhQ_A) are the principal site for O₂ photoreduction in Chlamydomonas PSI at high irradiances (≥300 µmol photons m⁻² s⁻¹). Under such light regimes, we propose that the Mehler reaction starts to function as a safety valve that prevents overreduction of the photosynthetic ET chain.

Methods

Strains, growth, thylakoid membrane and PSI isolation

Chlamydomonas reinhardtii strains with His₆-tag added to the N-terminus of the PsaA protein (PBC1 pKR152 for WT and PBC1 pKR411 for the strain bearing a mutation in codon 689 of psaA exon 3 [psaA-3]) were grown in TAP medium at 25°C with constant air flow under continuous white light (∼70 µmol photons m⁻² s⁻¹). The culture was harvested, and thylakoid membranes and mature PSI complexes were purified as described in (Marco et al., 2018). The fraction corresponding to the mature PSI complexes was collected, concentrated, and stored at −80°C. The maturity of the complexes was confirmed (Supplemental Table S1).

Expression and purification of recombinant proteins

The recombinant C. reinhardtii Pc, Fd, and FNR were heterologously expressed in Escherichia coli BL21 cells, and purified as described in Marco et al. (2018) and Ben Zvi and Yacoby (2016).

Determination of ${\mathrm{P}}_{700}^{+}$ concentration

The concentration of fully photo-oxidized ${\mathrm{P}}_{700}^{+}$ was determined by measuring light minus dark absorption changes at 698 nm (ε = 100mM⁻¹× cm⁻¹; Witt et al., 2003). The assay medium contained 30-µg Chl mL⁻¹ PSI, 20-mM NaCl, 5-mM MgCl₂, 10-mM Asc, 10-nM Pc, 50-µM MV, 0.03% (w/v) N-dodecyl β-D-maltoside (β-DDM), 20-mM Tricine-NaOH (pH 7.5) was placed in spectrophotometer (Cary 50, Varian) and illuminated with saturating white light (Intralux 5000, Volpi).

Measurement of O₂ concentration changes using Membrane Inlet Mass Spectrometer

Changes in O₂ (the stable isotope ¹⁶O₂) concentration were measured with a Membrane Inlet Mass Spectrometer (MIMS, QMS 200 M1; Pfeiffer Vacuum) according to Liran et al. (2016); the data collection frequency was 1 s. Measurements were conducted under constant stirring at 21°C. The signal was calibrated by calculating the difference in values obtained in air saturated buffer (containing 250-µM O₂) and in O₂-depleted buffer after removal of O₂ by glucose (10 mM), glucose oxidase (30 U mL⁻¹), and catalase (500 U mL⁻¹). A suspension of PSI complexes in buffer containing 20-mM NaCl, 5-mM MgCl₂, 0.03% (w/v) β-DDM, 20-mM Tricine-NaOH (pH 7.5), and electron donors and acceptors as indicated in figure legends, was placed in a sealed cuvette and illuminated by red light provided by a Dual-PAM (Walz GmbH).

In preliminary experiments, using the approach described in Kozuleva et al. (2014, 2015), we confirmed that light-dependent O₂ consumption resulted from direct ET from PSI cofactors to O₂ with production of ${\mathrm{O}}_2^{•-}$ as the primary product (Supplemental Tables S2, S3). The light-induced rate of O₂ consumption (V_O2) was calculated by subtraction of the dark slope from the light slope.

In experiments aimed for determination of k₂, the O₂ concentration in the reaction medium was decreased to 10–30 µM by N₂ flushing of the headspace over the suspension in the sealed vessel, followed by transfer of the suspension by syringe to the air-depleted sealed cuvette containing the MIMS inlet probe and stir bar. It was necessary to restrict the O₂ concentration because k₂ can only be determined if the rate of the reaction depends on the concentration of O₂ (Supplemental Figure S7). In these experiments, catalase (500 U mL⁻¹) was added to prevent accumulation of H₂O₂, which is decomposed by [4Fe–4S] clusters to give a highly reactive hydroxyl radical capable of inhibiting PSI. In addition, accumulated ${\mathrm{H}}_2^{16}$O₂ molecules (mass 34) could potentially interfere with the ¹⁶O₂ (mass 32) signal and introduce errors.

Calculation of the apparent rate constant of O₂ photoreduction by PSI

The reduction of O₂ by PSI (⁠$V_{{\mathrm{O}}_2}$⁠) is a second-order reaction:

$$V_{{\mathrm{O}}_2}=k_2\times \left[\text{PSI}\right]\times \left[{\mathrm{O}}_2\right]V_{{\mathrm{O}}_2}=k_2\times \left[\text{PSI}\right]\times \left[{\mathrm{O}}_2\right]$$

(1)

Since the concentration of PSI, which was equal to concentration of the maximal concentration of photo-oxidized ${\mathrm{P}}_{700}^{+}$⁠, did not change, k₂ × [PSI] was expressed as a pseudo first-order rate constant (k₁′):

$$V_{{\mathrm{O}}_2}=k_1\prime \times \left[{\mathrm{O}}_2\right]$$

(2)

The k₁′ was calculated as follows:

$$k_1\prime =\frac{1}{t}\times \ln \frac{{\left[{\mathrm{O}}_2\right]}^0}{{\left[{\mathrm{O}}_2\right]}^t}$$

(3)

where, t represents the duration of illumination (s), [O₂]⁰ and [O₂]^t represent the concentrations of oxygen at the onset and at t timepoint during the illumination, respectively. It was necessary to correct [O₂]^t to account for nonspecific O₂ concentration changes. For that, the rate of O₂ concentration changes in dark was subtracted from the rate of O₂ uptake measured in light, that is, the light-induced $V_{{\mathrm{O}}_2}$ was calculated. It was also necessary to correct the rate for the addition of catalase, which released one O₂ per two H₂O₂ molecules produced via O₂ photoreduction by PSI (Supplemental Table S2), thereby decreasing V_O2 by a factor of two. Therefore, the [O₂]^t values in Equation 3 were based on calculating Equation 4:

$${\left[{\mathrm{O}}_2\right]}^t={\left[{\mathrm{O}}_2\right]}^0–t\times 2\times {\mathrm{V}}_{\mathrm{O}2}$$

(4)

Preparation of F_X-core and A₁-core complexes

The F_X-core complexes were obtained by incubating PSI complexes in buffer containing 6.8 M urea (Parrett et al., 1989). The A₁-core complexes were obtained by incubating F_X-core complexes in buffer containing 3.4-M urea and 5-mM K₃[Fe(CN)₆]. The treatment efficiency was followed by monitoring the acceleration of charge recombination kinetics from ∼50 ms in the intact PSI complexes to 1 ms and 10 µs in F_X-core and A₁-core, respectively (Supplemental Figure S5).

NADPH measurement

Light induced NADPH production was detected by a Dual-PAM equipped with DUAL-ENADPH and DUAL-DNADPH modules. The signal was calibrated by adding a known concentration of NADPH.

Chlorophyll extraction and measurement

Chl concentration was determined spectrophotometrically after extraction in 96% ethanol according to Lichtenthaler (1987).

Accession numbers

Sequence data can be found in the GenBank/EMBL data libraries under accession numbers ferredoxin (PetF), P07839; ferredoxin-NADP+ oxidoreductase (FNR), P53991; Plastocyanin (PCY1), A8JH68; Photosystem I P700 chlorophyll a apoprotein A (PsaA), P12154; Photosystem I iron–sulfur center (PsaC), Q00914.

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. Effect of concentration of Pc (green) or TMPD (orange) as ${\mathrm{P}}_{700}^{+}$ electron donor on the O₂ photoreduction rate.

Supplemental Figure S2. Effect of irradiance on the apparent k₂^MV in the presence of Pc (5 µM; dark blue lines) or TMPD (1 mM; orange line) as electron donor to ${\mathrm{P}}_{700}^{+}$⁠.

Supplemental Figure S3. Rate of light-induced O₂ uptake in a PSI suspension as a function of methyl viologen concentration.

Supplemental Figure S4. Effect of TMPD concentration on the O₂ photoreduction rate in the A₁-core complexes.

Supplemental Figure S5. Charge recombination in the intact PSI, F_X-core, and A₁-core complexes from WT and intact mature PSI from the PsaA-F₆₈₉N mutant.

Supplemental Figure S6. Effect of MV addition on the rate of O₂ photoreduction in PSI complexes from WT and PsaA-F₆₈₉N mutant.

Supplemental Figure S7. Rate of light-induced O₂ uptake in a PSI suspension as a function of O₂ concentration.

Supplemental Table S1. Characterization of intact PSI complexes purified from WT and the mutant PsaA-F₆₈₉N.

Supplemental Table S2. Effect of superoxide dismutase and catalase on the rate of O₂ uptake in a suspension of fully mature WT PSI complexes.

Supplemental Table S3. Effect of sodium ascorbate (Asc) on the rate of O₂ uptake in a suspension of intact PSI complexes.

M.K. conceived the original research plans; I.Y. supervised the experiments; M.K. isolated photosystem I complexes and proteins and performed the experiments using MIMS and A.P. prepared stripped samples and performed charge recombination measurements; Y.M. provided technical assistance to M.K. in terms of MIMS measurements; K.R. provided the His-tagged photosystem I complexes of the wild-type and the mutant and designed the experiment with the mutant complexes; M.K. and B.I. analyzed MIMS data and A.S. analyzed charge recombination kinetics; M.K. and I.Y. wrote the article with contributions of all the authors.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plphys/pages/general-instructions) is: Iftach Yacoby ([email protected]).

Acknowledgments

M.K. is thankful to Dr. Syed Lal Badshah, Dr. Oren Ben-Zvi, and Dr. Pini Marco for help with protein purifications, Patricia Baker for strain construction and help with alga culture maintenance, and to Dr Ilya Naydov, Dr Maria Borisova-Mubarakshina, and Dr Georgy E. Milanovsky for valuable discussion.

Funding

This work was supported by the ISF (Israel Science foundation) 1646/16, BSF (US Israel binational science foundation) 201666, The Ministry of Science and Higher Education of the Russian Federation, State Scientific Program no. 121040500121-3; the investigation of F_X-core and A₁-core complexes was supported by Russian Science Foundation (RSF) grant #19-14-00366. Fulbright Visiting Scholar Program supported M.K. visit to K.R.’s lab.

Author notes

Senior authors.