The wavelength of the incident light determines the primary charge separation pathway in Photosystem II

Charge separation is a key component of the reactions cascade of photosynthesis, by which solar energy is converted to chemical energy. From this photochemical reaction, two radicals of opposite charge are formed, a highly reducing anion and a highly oxidising cation. We have previously proposed that the cation after far-red light excitation is located on a component different from PD1, which is the location of the primary electron hole after visible light excitation. Here, we attempt to provide further insight into the location of the primary charge separation upon far-red light excitation of PS II, using the EPR signal of the spin polarized 3P680 as a probe. We demonstrate that, under far-red light illumination, the spin polarized 3P680 is not formed, despite the primary charge separation still occurring at these conditions. We propose that this is because under far-red light excitation, the primary electron hole is localized on ChlD1, rather than on PD1. The fact that identical samples have demonstrated charge separation upon both far-red and visible light excitation supports our hypothesis that two pathways for primary charge separation exist in parallel in PS II reaction centres. These pathways are excited and activated dependent of the wavelength applied.

energy lower than 690 nm. Interestingly, this also holds for PS II in green algae 19 and higher plants 20,21 containing a Chl a/b antennae without any Chl d. Spectroscopic investigation of the different electron transfer reactions in PS II, defined the far-red limit for water-oxidizing photochemistry in PS II to 780 nm (single flash) up to 800 nm (photoaccumulation) 22 . Based on these findings, an alternative charge separation pathway was proposed to occur upon far-red light excitation. When this pathway is activated, a state denoted X* that is lower in energy compared to P 680 * (formed in visible light) is formed in PS II. Despite its lower energy, X* is able to trigger the charge separation, reduce Pheo and oxidize tyrosine Z (Y Z ). This is an important evolutionary finding as it has redefined the required threshold energy for oxygenic photosynthesis, since X * apparently works with an energetic threshold which is lower than P 680 * 22 . The existence of such a state was suggested previously, but never shown to induce charge separation at room temperature or at wavelengths as high as 800 nm. Although oxygen evolution and variable chl a fluorescence have been observed in sunflower (Helianthus annuus) and bean (Phaseolus vulgaris) leaves using wavelengths up to 780 nm 21 , the maximum long wavelength with which successful charge separation was achieved was 730 nm, as demonstrated by spectral hole-burning experiments on PS II core complexes at cryogenic temperatures (1.5 K) 23,24 . This range has now been extended to ca 780-800 nm where complete water oxidation can be achieved at room temperature in different PS II preparations 22 .
Mokvist et al., 2014 provided further insight into the electron transfer reactions in PS II, induced by far-red photochemistry 25 . It was shown that at 5 K, Y Z was the preferred secondary electron donor in far-red light. This is different from the situation with visible light where the Cyt b 559 /Chl Z /Car D2 secondary pathway was equally important 25 .
This observation led to the proposal of a different first stable charge pair denoted as P x + Q A − being formed under far-red light, as compared to the normal P D1 + Q A − under visible light excitation. The proposed electron hole in P x + was suggested to be residing on the Chl D1 molecule in the PS II reaction centre at 5 K. The observation also further supported the proposed existence of a low-energy threshold charge separation pathway, where the primary donor is Chl D1 rather than P D1 25 . (See Fig. 1 which shows the redox active components in the PS II reaction centre from ref. 26 ).
In this study, we have used the spin polarized triplet EPR signal from 3 P 680 , to further probe the photochemistry following visible or far-red illumination. Our results strengthen the hypothesis that the stable cation formed by the charge separation resides at Chl D1 when the excitation energy is provided by long wavelength (far-red) light, rather than on P D1 of the P D1 / P D2 Chl pair, which is the normal situation in visible light excitation.

Results
Functionality of over-reduced PS II. The experimental protocol to achieve reliable and quantitative double reduction of Q A in PS II enriched membranes of our type (so called BBY particles), is complicated and has several potential pitfalls. The double reduction protocol involves reduction with sodium dithionite using benzyl viologen as mediator for several hours and demands exclusive handling under anaerobic and dark conditions [27][28][29] . Also the re-oxidation protocol is difficult and could potentially lead to sample damage 27 . It was therefore important to test for the integrity of the critical electron transfer reactions both after double reduction and after the re-oxidation of the samples used here.
Status of the charge separation reaction. The most important reaction in the current study is the primary charge separation reaction. In PS II centres with Q A H 2 (double reduced and protonated Q A ) or samples lacking Q A , the primary charge separation can be followed by photo-accumulation of the Pheo − radical under reducing conditions [30][31][32] . This reaction is also known to be functional using far-red illumination 22 . Figure 2 shows the EPR signal from the Pheo − radical in PS II enriched membranes where Q A was doubly reduced. The radical EPR spectrum formed by either visible or far-red (732 nm) light was 13G wide with g = 2.0035 [30][31][32] . Both parameters are indicative that the signal originates from the Pheo − radical. Thus, the mere formation of the signal is an indication that PS II was able to perform the charge separation reaction despite being exposed to the dithionite/benzyl viologen reduction protocol. The size of the EPR spectrum from Pheo − can be compared to the size of the EPR spectrum from the Y D • radical in the corresponding intact PSII sample (non-reduced) that amounts to 1 radical/PSII reaction center 33,34 . In this particular experiment, exposure to white light resulted in Pheo − radical formation in 70% of PS II, while the illumination at 732 nm resulted in Pheo − radical formation in 55% of PS II. We can thus conclude that ≥70% of the PS II centres can perform the primary charge separation after the reduction treatment. In addition, it seems that 732 nm is nearly as efficient as white light, the small difference probably reflecting the weaker light source used at 732 nm.
Status of the Mn-cluster. It is generally thought that extensive reduction of PS II results in inactivation of the OEC through removal of the CaMn 4 -cluster. This also holds for our samples that showed typical characteristics of samples lacking the CaMn 4 -cluster. After the re-oxidation procedure, no O 2 evolution could be detected. Corroborating this, neither the Split S 1 EPR signal nor the S 2 state multiline EPR signal, both involving a functional CaMn 4 -cluster, could be detected (EPR spectra not shown).

Status of Tyrosine-D oxidation and the quinone acceptors Q A and Q B .
Chemical double reduction of Q A is most probably followed by protonation, forming Q A H 2 [27][28][29] . Presumably this species could leave the Q A -binding pocket irreversibly during the very long double reduction procedure. However, this does not seem to be the case, at least not in the majority of the PSII centres. In the re-oxidized samples, Y D • could be formed in a major fraction of the PS II centres (>55%) (Fig. 3, red trace). Y D oxidation involves both primary photochemistry in PSII (charge separation between P 680 and Pheophytin) and the transfer of, at least, one electron to the quinones on the acceptor side of PS II. Thus the formation of Y D • unequivocally shows both that charge separation worked (Figs 2 and 3, red trace) and that Q A remained bound in a majority of the samples despite the harsh double reduction/re-oxidation protocols.
We also performed flash induced fluorescence measurements in the re-oxidized samples. Our results show that a large fraction of PSII was able also to perform secondary electron transfer. After the flash, the immediate fluorescence induction is indicative of reduction of Q A , when bound to its binding pocket. Q A − then decays with different kinetics dependent on the integrity of PSII. The dominating intermediate decay phase in our samples (calculated half-life = 13 ms) (Fig. 4, black trace), is similar to the Y Z • Q A − recombination phase observed in for example Tris-washed PS II membranes 35 . In Tris-washed PSII membranes the OEC is absent, corroborating our conclusion that our samples after double reduction and re-oxidation had lost the CaMn 4 -cluster. We can also conclude that, not surprisingly, a substantial amount of Q A − remained reduced also after the re-oxidation protocol. This is shown by the high F 0 (0.63) in our fluorescence measurement (Fig. 4, black trace). Furthermore, there seems to remain very few PS II centres with bound Q B after the re-oxidation procedure. This can be concluded from comparison of the kinetic traces in the presence and absence of the inhibitor DCMU. Addition of DCMU, which inhibits electron transfer from Q A to Q B , removed a small fast component from the fluorescence decay curve (Fig. 4, red trace). This fast phase most likely reflects electron transfer from Q A to a remaining bound Q B . Normally this fast decay phase dominates in PSII and its small amplitude in our samples clearly indicates that there is very little Q B remaining after the double reduction/re-oxidation treatments. The rest of the decay is similar to the decay observed in presence of DCMU and again reflects Y Z Induction of the spin polarized 3 P680 under white and far-red light excitation. The spin polarized triplet state 3 P 680 is a useful probe to the physical environment and photochemistry of P 680 and the primary  radical pair in PSII 36 . In intact PSII, 3 P 680 is not observable with EPR spectroscopy 37 . However, when Q A is either removed (in for example the D1/D2/Cyt b 559 preparation) 37,38 or double reduced either by chemical treatment [27][28][29] or extensive illumination under anaerobic conditions 39 , the spin polarized EPR signal from 3 P 680 , can be observed as a result of recombination of the P680 + Pheo − charge pair. Here, we have investigated the formation of 3 P 680 in samples where Q A has been chemically double reduced. Figure 5 depicts the "light minus dark" difference EPR spectra from the spin-polarized triplet state of 3 P 680 obtained from PSII membranes with double reduced Q A , when illuminated by continuous white light (spectrum (a)) at 5 K. Our chemical reduction procedure converts Q A to Q A 2− , which is then protonated to form Q A H 2 27,29 . After excitation with light, no forward electron transfer from Pheo − is possible. Instead, the charge-separated state P 680 + Pheo − will decay through recombination. However, it is long-lived enough to allow spin dephasing, thereby allowing formation of the triplet form of the radical pair, the so called spin-polarized 3 P 680 state as follows 36,40 : Hence, at low temperatures (5 K), the triplet state of the primary donor, 3 P 680 , is formed with high yield 27,29,38,41 . From orientation studies of the 3 P 680 EPR signal it was concluded that it is localized on a Chl molecule (unidentified at that time) most likely to be Chl D1 (in the present structural nomenclature, see Fig. 1) 38,42,43 . It is known that the product of the primary charge separation in visible light is P D1 + Pheo − . Therefore, the localization of the 3 P 680 on another Chl molecule than P D1 is unexpected. The phenomenon has been explained by two alternative mechanisms. Either the triplet state migrates from P D1 to Chl D1 after its formation or it is the Chl cation (P D1 + ) that moves to Chl D1 (forming Chl D1 + ) prior to 3 P 680 formation ( 38 and refs. therein). As shown in Fig. 5, spectrum (a), the characteristic spin-polarized EPR signal from 3 P 680 was observed under white light illumination. The spectrum is in agreement with previously published EPR spectra of 3 P 680 27,29,36-38 . The triplet EPR spectrum depicts a low-field to high-field AEEAAE (Absorption = A, Emission = E) polarization pattern spanning ~650 G. X, Y and Z (D > 0, E < 0) indicate the field positions of the respective canonical orientations for the zero field splitting tensor of the triplet state.
We have also intensively searched for the formation of 3 P 680 in an identical sample after illumination in the EPR cavity with far-red light. Interestingly, as depicted in Fig. 5, spectrum (b), no spin polarized EPR signal from 3 P 680 could be detected under these conditions. This obvious difference between visible and far-red light excitation is intriguing as our control measurements clearly indicate that: i) our samples were indeed able to form 3 P 680 in visible light (Fig. 5, spectrum a) and ii) were indeed able to perform charge separation and a multitude of electron transfer reactions on both the donor and the acceptor side of PS II in both visible and far-red light (Figs 2-4).
Accumulated Spectra of spin polarized 3 P 680 . It is possible that the lack of the 3 P 680 EPR signal under far-red light reflects the complete absence of 3 P 680 formation. However, it could also be caused by a very fast 3 P 680 decaying species not detectable under standard CW EPR conditions. To test this, we therefore accumulated  Figure 6 shows the flash-associated transients recorded in the low-field region of the 3 P 680 spectrum (2850-3150 G) upon 532 nm (a), 610 nm (b), 689 nm (c) and 730 nm (d) laser flash excitation. The intensity of the spin-polarized 3 P 680 features is high under 532 nm and 610 nm excitation and decreased by more than 50% under 689 nm excitation. Ultimately, there is no EPR signal from 3 P 680 detected even with our fastest available time resolution and after extensive illumination with far-red laser flash excitation (730-790 nm) (Fig. 7). It is also important to point out that the decay kinetics of respectively the X and Z peaks are identical irrespective of the induction wavelength between 532 nm and 689 nm (Fig. 7a,b). The decay kinetics are, however, faster for the X peak than for the Z peak similar to earlier observations 36 . At ≥730 nm no peaks were observed.

Discussion
The understanding of how low energy excited states contribute to PSII photochemistry continues to be a challenging task 44 . This is especially true if specific charge transfer bands, not detectable by conventional spectroscopy, are responsible for this type of photochemistry 44,45 . In this study we are analysing these phenomena, by using EPR based, comparative product analysis (see also ref. 45 ) using the characteristic EPR signal of spin polarized 3 P 680 as a probe for the primary charge separation at low temperatures.
In our first publication we showed, for the first time, that PSII operates much further to the far-red region (by 100 nm) than was believed before. We have shown that efficient secondary electron transfer reactions take place under these conditions, including turnover of the OEC, reduction of the primary and secondary quinones, oxidation of the secondary donor Tyr Z and reduction of the primary donor, Phe, using far-red (up to ≥800 nm) photons. This was an interesting observation and in an earlier study we proposed an alternative charge separation reaction under far-red conditions involving a low energy primary donor, denoted X * , which was able to drive both Pheo − reduction and Y Z oxidation 22 .
This suggestion was further substantiated in a study of low temperature (5 K) electron transfer on the donor side of PS II. To elaborate, we found that far-red light preferentially promoted electron transfer from Y Z to the primary donor. In contrast, visible light promoted almost equal electron transfer from either Y Z or the Cyt b 559 / Chl Z /Car D2 secondary pathway. The partition ratio between the two electron donor pathways was thus wavelength dependent, Y Z being the better donor after far-red excitation 25 .
An important implication of this work, discussed in 22,25 is that the far-red photochemistry is triggered by weakly absorbing charge transfer states among the core pigments in PSII and not by the well-known bands at 680 nm (P D1 ) or 684 nm (Chl D1 ).
In our present manuscript, we are getting even closer to the primary charge pair, P680 + Pheo − , showing that the primary donor is indeed different. Thus, our, step wise refined, product analysis technique provides a unique approach to study the far-red photochemistry in PSII and to cast further light on this phenomenon by studying the primary electron donor in PS II, P 680 , directly. Our result is surprising. As described previously, the spin polarized EPR signal from 3 P 680 can be observed from PS II centres where Q A has been doubly reduced (or it is absent). It has earlier been shown that its formation using white light is almost quantitative 27,36 . In our hands, the 3 P 680 EPR signal is large and fairly consistent in amplitude, when PS II is exposed to light in the visible region. However, in the far-red, at wavelengths ≥730 nm the illumination does not induce any observable 3 P 680 EPR signal at all. There can be several reasons for the above phenomenon. To find the answers and formulate our hypothesis, certain questions must be addressed: (i) Is the electron hole on the primary donor not forming at all under far-red light illumination? (ii) If it was formed in far-red light, would the 3 P 680 decay too fast and escape detection? (iii) Do we trigger an alternative charge separation pathway by far-red light, hence creating a different, low energy, state as suggested in ref. 20 ? (iv) Is the cation located at a different Chl if compared to the visible light as hypothesized in ref. 23 ? If yes, then the recombination between the primary donor and Pheo − occurs too fast to allow 3 P 680 formation after far-red illumination.
First, we can exclude that this particular type of PS II sample (sodium dithionite/benzyl viologen reduced) is not able to carry out any charge separation in the far-red at all, since we in fact can photoaccumulate the Pheo − radical in the majority of PS II under illumination with 730 nm light. Therefore, an electron hole on the P 680 entity is formed after excitation by far-red light. In that case, recombination between Pheo-and P 680 + will take place, either via or without the formation of 3 P 680 .
Second, in far-red light we observe no spin polarized 3 P 680 EPR signal. However, in the event where 3 P 680 is anyway formed by the far-red light, it is possible that it decays faster than we can detect it, in contrast to the situation in the visible part of the spectrum. An attempt has been made to investigate this possibility further by our kinetic measurements of 3 P 680 induced by different light wavelengths. Here we report identical decay kinetics between the different light regimes where we could observe the 3 P 680 in the visible light spectrum. However, we were not able to observe any 3 P 680 formation above 700 nm, despite our sub-millisecond time resolution. We argue that there is little reason to suggest that 3 P 680 formed by far-red light should decay too fast to allow its detection. Instead, we propose that the far-red light illumination, although it efficiently drives charge separation and creation of a primary radical pair [P 680 + Phe − ], does not result in the formation of the spin polarized EPR signal from 3 P 680 . Since Pheo − is the same in both cases, it is the nature of the Chl cation, (Chl + ) that differs after excitation with visible or far-red light.
This observation also has important implications with respect to the earlier orientation studies of the spin-polarized 3 P 680 EPR signal formed by visible light (see above) 38,42,43 . From our results, we can rule out that the triplet state is formed after migration of the cation from one monomeric Chl to another (from P D1 to Chl D1 ). In this case we would have observed the 3 P 680 EPR signal under both illumination regimes since the triplet would have been formed from the same Chl D1 + Pheo − charge pair. Instead, detection of the signal under visible light only strongly indicates involvement of different charge pairs under the two illumination regimes with the triplet state only being formed by the P D1 + Pheo − charge pair recombination. Third, we have recently demonstrated preferential electron transfer from Y Z in comparison to very weak electron transfer from the Cyt b 559 /Chl Z /Car D2 side donors under far-red light illumination at 5 K 25 . To explain this, we hypothesized the existence of an alternative charge separation pathway under far-red light involving the oxidized primary donor after charge separation residing on Chl D1 rather than P D1 25 . Thus, we suggest that upon far-red excitation, a new pathway is activated. We denote this as the the Chl D1 pathway (Fig. 1). Here, a lower energy excited state [Chl D1 Pheo] * is formed, most probably corresponding to the proposed X * in ref. 20 , which would then lead to a different radical pair after the charge separation, [Chl D1 + Pheo − ] (Fig. 1). Lastly, the lack of observable triplet signal implies that the recombination from Pheo − is too fast to allow spin dephasing. This suggests that the electron hole is presumably residing on Chl D1 instead of P D1 . Chl D1 is much closer to Pheo than P D1 (5 Å vs 8.1 Å) 26 . This shorter distance would probably facilitate faster recombination of the charge pair Chl D1 + Pheo − . Hence, we propose that the far-red light preferentially drives charge separation to Pheo from Chl D1 in the reaction center (Fig. 1). Since Chl D1 + and Pheo D1 − are in close proximity, we propose that recombination occurs faster than between the radical pair [P D1 + Pheo − ] found in visible light. Thus, the [Chl D1 + Pheo − ] radical pair probably recombines back to the ground state prior to spin dephasing can occur. Consequently no spin polarized 3 P 680 is formed at any of the PSII centres.
It is generally thought that at low temperatures the cation in the primary radical pair is stabilized on P D1 (Fig. 1) 40,46-50 , which is not far from neither Y Z nor the Cyt b 559 /Chl Z /Car D2 secondary pathway. Therefore they compete, and both are oxidized in substantial yield at very low temperatures 25 . Contrary, the localization of the primary electron hole on Chl D1 rather than P D1 under far-red light was discussed in ref. 23 . It was proposed that with the formation of the primary electron hole on Chl D1 , the location of Chl D1 being relatively closer to Y Z but much more distant from the Cyt b 559 /Chl Z /Car D2 secondary pathway, could explain at least qualitatively, the preference for oxidation of Y Z over Cyt b 559 /Chl Z /Car D2 .
The existence of parallel charge separation pathways in PSII under different light regimes and under physiological and cryogenic conditions has been suggested and discussed before [50][51][52][53][54][55] . In 2009 Thapper et al. 22 suggested the formation of a low-energy state denoted X * in PS II after far-red light excitation at room temperature. Although functioning at a lower energetic threshold compared to P 680 * , it was shown to effectively trigger the charge separation, reduce Pheo and oxidize Y Z up to ~800 nm. More recently, transient absorption spectroscopy at 77 K, suggested the existence of two different charge separation pathways in PS II 50 . These were dependent on the light induced protein configuration in PSII core complexes. In one path, denoted the Chl D1 path, the charge separation events were proposed to involve an interaction between Chl D1 and Pheo, and follow the sequence (Chl D1 Pheo)* ⇒ (Chl D1 + Pheo − ) ⇒ (P D1 + Pheo − ). In the second pathway, denoted the P D1 path, a charge transfer state P D1 P D2 is excited and the charge separation events were suggested to follow an alternative sequence (P D1 P D2 Pheo)* ⇒ (P D1 52,54,55 . The proposal that the two pathways are actually of an antagonistic nature and the effectiveness of each of the pathways essentially depends on recognition of energetic disorder, fits with our observations. When a low energy photon in the far-red is used for excitation, the subsequent energetic disorder causes the preferential excitation to be strongly localized at Chl D1 , thus allowing for charge separation to take place with the first charge separated state being Chl D1 + Pheo − . This localized state should be energetically lower than the P D1 P D2 charge transfer state, which is in agreement with the results of the present study. Thus, we propose that two Scientific REPORTS | (2018) 8:2837 | DOI:10.1038/s41598-018-21101-w pathways for primary charge separation exist in parallel in PS II reaction centres and their excitation and activation is wavelength dependent (Fig. 1).

Concluding Remarks
Here, we have provided further insight into the location of the primary charge separation upon far-red light excitation of PS II, using the EPR signal of the spin polarized 3 P 680 as a probe. We demonstrate that, under far-red light illumination, the spin polarized 3 P 680 is not formed, although we show that the primary charge separation is still occurring at these conditions. We propose that this is because under far-red light excitation, the primary electron hole is localized on Chl D1 , rather than on P D1 (Fig. 1). The formation of the lower energy excited state Chl D1 Pheo * by far-red light, leads to the [Chl D1 + Pheo − ] primary charge pair. The close proximity of these two cofactors, allows fast recombination to the ground state before spin dephasing can occur, hence the 3 P 680 cannot be formed.
The fact that identical samples have demonstrated charge separation upon both far-red and visible light excitation further supports our hypothesis that two pathways for primary charge separation exist in parallel in PS II reaction centres. They are excited and activated dependent of the wavelength applied. In the visible part of the spectrum the first product of the charge separation is normally considered to be P D1 + Pheo − . In stark contrast, we hypothesize that far-red illumination results in the Chl D1 + Pheo − radical pair. The latter recombines without formation of the 3 P 680 , thereby explaining the results in the present contribution. When far-red light is applied to intact PS II however, the oxidizing electron hole, presumably Chl D1 + , preferentially drives electron transfer from Y Z and consequently the CaMn 4 cluster. As described earlier, the Yz-CaMn 4 pathway is preferred over the Cyt b 559 /Chl Z /Car D2 pathway at low temperatures due to the favourable location of Chl D1 vs Y Z over Chl Z 21 .

Materials and Methods
PSII membrane preparation. Spinach (Spinacia oleracia) was grown hydroponically as described previously at 20 °C under cool white fluorescent light (Osram Powerstar HQI-400W/DV dysprosium lamp, intensity 300 μEm −2 s −1 ), with light-dark periods of 12 h 33 . Oxygen evolving PSII enriched membranes (BBY-type) were prepared according to previously published procedures 56,57 . The membrane particles were re-suspended in a final buffer containing 400 mM Sucrose, 15 mM NaCl, 3 mM MgCl 2 and 25 mM MES-NaOH pH 6.1, and frozen as beads at −80 °C, at a Chl concentration of 6 mg/ml. Chemical reduction and re-oxidation of PS II. In order to obtain reaction centers with doubly reduced Q A , PS II membranes were exposed to reducing conditions as described in 27,28 with modifications as in 29 .
Specifically, upon addition of 40 mM sodium dithionite, 100 μM benzyl viologen and 3 mM EDTA under anaerobic conditions, dark-adapted PS II membranes at a Chl concentration of 6 mg/ml, were incubated in the dark for 5 hrs to achieve the double reduction of Q A to Q A H 2 . All incubations were carried out in argon flushed EPR tubes. Upon completion of the incubation time, the EPR samples were frozen within 1 sec in a 200 K dry ice/ethanol bath and subsequently transferred into N 2 (l) before the measurements.
"Reversed" samples where Q A H 2 was re-oxidized, were prepared according to previously published procedures 27 . The doubly reduced samples were initially washed three times (15 000 × g, 15 min cycle) with argon flushed buffer containing 400 mM Sucrose, 15 mM NaCl, 3 mM MgCl 2 and 25 mM MES-NaOH pH 6.1, to remove sodium dithionite and benzyl viologen. They were then re-oxidized with 5 mM K 3 Fe(CN) 6 , which was subsequently removed by repeating the washing procedure three times with final buffer, as described above. All washing steps were performed in complete darkness and at 4 °C.
Illumination procedures. The spin polarized 3 P 680 EPR signal was induced by illumination of the samples in the EPR cavity with white or far-red light at 5 K. For continuous illumination, white light, filtered through a 5 cm-thick copper sulphate solution heat filter, was provided by an 800 W halogen projector lamp. Far-red light illumination was achieved by the use of a custom made LED module, emitting continuous light centred at 732 nm. Appropriate long pass filters were used so as to ensure blocking of transmission of any stray UV light (CC4) as well as transmission of light <725 nm (RG9). Transient formation and accumulation of the spin polarized 3 P 680 EPR signal was studied by applying laser flashes into the EPR cavity, from a Spectra Physics PRO-290 Q-switched Nd:YAG laser (6 ns flash, 5 Hz flash frequency) equipped with a Spectra Physics Quanta Ray MOPO 730 optical parametric oscillator.
The Split S 1 signal was obtained by white or far-red continuous wave light illumination at 5 K, as in 58 .
For the measurements of the Pheo − radical, doubly reduced PS II membranes (6.0 mg Chl/ml) were subjected to white or far-red (732 nm) continuous wave light illumination as described earlier [30][31][32] , at 20 °C, for 6 and 10 minutes respectively at room temperature. Thereafter, the samples were frozen within 1 sec in a 200 K dry ice/ethanol bath and subsequently transferred into N 2 (l) before the measurements. The size of the EPR spectrum of the formed Pheo − radical was compared by double integration to the size of the EPR spectrum from Y D • in a corresponding sample prior to double reduction, allowing us to quantify the function of the primary charge separation in the double reduced samples.
Fluorescence measurements. Flash-induced fluorescence decay measurements were performed at room temperature at a sample concentration of 20 µg Chl/ml. PS II membranes were dark adapted for 5 min and measurements were performed in the presence or absence of 20 µM DCMU. The variable fluorescence decay traces were recorded with a FL3000 double modulated fluorometer (PSI Photon Systems Instruments, Czech Republic) according to 59 . The first data point was taken 100 μs after the actinic flash. Measuring flashes were then applied logarithmically eight times per decade in a time range up to 100 s 60 .
Scientific REPORTS | (2018) 8:2837 | DOI:10.1038/s41598-018-21101-w EPR spectroscopy. X-band EPR measurements were performed with an Elexsys E580 (Bruker BiosSpin) equipped with a standard cavity (ST 4102). All measurements were performed at low temperature that was achieved with the use of a helium flow cryostat and an ICT-4 temperature controller (Oxford Instruments, UK). The time-resolved EPR (TREPR) measurements were performed with the use of the ADF fast digitizer board (2 MHz fixed sampling rate). Signal acquisition was synchronized with 5 Hz laser flashes using a LC880 TTL pulse generator (100 MHz internal clock) (LabSmith, Livermore, California). An accessory transformer was used to amplify the TTL pulses to the 5 V amplitude required for triggering the laser lamp and Q-switch. Analysis of the EPR spectra was carried out with the Bruker Xepr 2.1 software.