Impact of energy limitations on function and resilience in long-wavelength Photosystem II

Article activity feed

1. 

   Version published to 10.7554/elife.79890 on eLife

   Jul 19, 2022
2. 

   eLife

   Jul 12, 2022

   Author Response

   Reviewer #1 (Public Review):

   Viola et. al. compared the electron transfer efficiency of two types of oxygenic far-red photosystem II (PSII) with the "conventional" PSII and analyzed how these far-red PSII use the limited energy from infrared photons to proceed photosynthesis. Oxygenic photosynthesis is an energy-intensive process, and a large headroom is also needed for preventing harmful back-reactions from occurring, which can produce singlet oxygen. This research investigated how the far-rad PSII managed to do their work with limited energy.

   The authors measured and compared the forward reactions of different kinds of PSII (Chl-a-PSII, Chl-d-PSII and Chl-f-PSII), including the flash-induced chlorophyll fluorescence decay and S-states turnover. These results led to a conclusion that the forward reaction quantum …

   Author Response

   Reviewer #1 (Public Review):

   Viola et. al. compared the electron transfer efficiency of two types of oxygenic far-red photosystem II (PSII) with the "conventional" PSII and analyzed how these far-red PSII use the limited energy from infrared photons to proceed photosynthesis. Oxygenic photosynthesis is an energy-intensive process, and a large headroom is also needed for preventing harmful back-reactions from occurring, which can produce singlet oxygen. This research investigated how the far-rad PSII managed to do their work with limited energy.

   The authors measured and compared the forward reactions of different kinds of PSII (Chl-a-PSII, Chl-d-PSII and Chl-f-PSII), including the flash-induced chlorophyll fluorescence decay and S-states turnover. These results led to a conclusion that the forward reaction quantum efficiency was not changed between "conventional" PSII and far-red PSII. However, the back-reactions of three types of PSII are different based on the measurements of the prompt fluorescence decay, delayed luminescence decay, and thermoluminescence band locations. The authors concluded that the two far-red PSII (Chl-d-PSII and Chl-f-PSII) have a different strategy for utilizing infrared light. Indeed, the authors showed that Chl-d-PSII containing cyanobacteria produced more singlet oxygen than other types, and this result was explained by the energy profile in the electron transfer chain.

   The major strength of this research is the authors made a direct comparison of different far-red PSII under the same conditions. It's exciting to have a side-by-side comparison between two types of far-red PSII. In addition, the authors also measured the singlet oxygen produced from all types of PSII which clearly showed the differences in the routes of recombination.

   We thank the reviewer for the interest demonstrated in our work and for the thoughtful comments, that we have addressed below.

   However, there are some concerns:

   1. The flash-induced fluorescence decay, thermoluminescence, delayed luminescence and S-states turnovers of the Chl-d-PSII and Chl-f-PSII have been characterized before (ref 5, 26, 39), but from intact cells compared to isolated membranes in this study, and similar conclusions have been achieved. The authors mentioned four reasons (lines 115-120, see the manuscript for the authors' arguments "i." to "iv.") why it's important to use isolated membranes. However, in my opinion, these reasons are not sufficiently strengthened:

   i. The transmembrane potentials from cells can be collapsed by adding uncouplers;

   ii. The authors mentioned the quinone pool in the cells is uncontrollable, but the authors didn't actually measure or manipulate the quinone pool in the membrane (e.g., the ratio of QB/QB-/empty-pocket in the samples);

   iii. The phycobilisomes can be controlled by different conditions through state transitions;

   iv. The isolation of membranes may not remove membrane-related quenching mechanisms (e.g., PSII quenching in State II, spillover, etc.).

   We do not agree with the reviewer on this point. We consider the use of membranes (or isolated PSII) as being the best solution to limit the effects listed at the end of the Introduction and to provide consistency between the different measurements, some of which cannot be performed in intact cells (i.e., the UV absorption measurements). More specifically:

   i) The effectiveness of uncouplers in dissipating the membrane potential is likely to vary between species (e.g., Chroococcidiopsis cells form aggregates incapsulated by a protective layer of excreted polymers) and should be assessed by directly measuring the membrane potential. ElectroChromic Shift-based measurements of the membrane potential in cyanobacteria have only been demonstrated in Synechocystis sp. PCC6803 and Synechococcus elongatus sp. PCC7942 (Viola et al. 2019, https://doi.org/10.1073/pnas.1913099116) and still need to be adapted to the far-red species used here. Additionally, commonly used uncouplers such as CCCP and FCCP are ADRY reagents, that interfere with PSII water splitting by directly reducing TyrZ (Ghanotakis et al. 1982, https://doi.org/10.1016/0005-2728(82)90115-3), and would affect all the measurements presented in this work.

   ii) In the dark, the redox state of the PQ pool in cyanobacterial cells has been observed to be kept in a highly reduced state by respiration, with potential consequences on the QB/QB- ratio. This could well vary between species, based on their different physiologies and growth conditions. In isolated cyanobacterial membranes and PSII, the QB/QB- ratio is expected to be around 50% after a short dark adaptation. This seems to be the case in our samples, based on the flash-dependent oscillations of the S2QB- and S3QB- thermoluminescence shown in Appendix 2 compared to the literature (Rutherford et al. 1982, https://doi.org/10.1016/0005-2728(82)90061-5), assuming an initial ~75% S1 population, as confirmed by the flash-dependent oxygen evolution and UV absorption. This is now mentioned in Appendix 2.

   iii) The control of state transitions requires specific illumination regimes incompatible with the conditions required for our experiments. Moreover, state transitions remain largely uncharacterised in the far-red species used in the present work. In some of these species, the situation is further complicated by the presence of both visible and far-red light-absorbing phycobilisomes that have a different spatial distribution in the cell (MacGregor-Chatwin et al. 2022, https://doi.org/10.1126/sciadv.abj4437).

   iv) Non-photochemical energy quenching in cyanobacteria seems to occur in phycobilisomes, due to the action of the Orange Carotenoid Protein (OCP). Both OCP and the phycobilisomes, if present in cyanobacterial cells (and that depends on the strains), are removed when membranes are isolated. It’s been proposed that direct quenching of the PSII core occurs in Synechococcus elongatus 7942 cells in state II (Choubeh et al. 2018, https://doi.org/10.1016/j.bbabio.2018.06.008), but since the mechanism has not been elucidated, no conclusion can be made on whether this could occur in membranes. The same is true for spill-over. Additionally, neither of the two mechanisms could be better controlled in cells than in membranes, so there would be no advantage here from working in vivo.

   In addition, the authors reached a conclusion that the Chl-f-PSII containing species should suffer from fluctuation light-induced membrane potential spikes, but don't actually measure this in physiologically relevant preparations. It will be more beneficial to use intact cells instead of an isolated membrane. I suggest the authors either restrict their conclusions to what the isolated membranes clearly show or make measurements in intact cells.

   The proposal that the far-red forms of PSII (both Chl-d-PSII and Chl-f-PSII) should suffer from increased charge recombination induced by spikes of membrane potential in fluctuating light is not new (see for example Nürnberg et al. 2018, https://doi.org/10.1126/science.aar8313), and is based on the observations made in plant PSII (Davis et al. 2016, https://doi.org/10.7554/eLife.16921) and assumed to be universal in oxygenic photosynthesis. In PSII, the transfer of electrons from the primary donor chlorophyll to QA occurs vectorially in the membrane, against the trans-membrane electric field, thanks to these electron transfer steps being exergonic. Spikes in the electric field due to sudden intensity fluctuations increase the probability of backward electron transfer. If the overall drop in the energy of the electron from the primary donor to QA is smaller (in a long wavelength PSII), it should result in a higher probability of backward transfer for a given trans-membrane electric field, and therefore a greater susceptibility to spikes in the electric field. We did not measure these effects and we do not claim to have done so. As already mentioned in the answer to point i) above, doing so would require the development of ElectroChromic Shift-based measurements of the membrane potential in the cyanobacterial species containing far-red photosystems. This is a separate research project beyond the scope of the present work.

   In conclusion, we believe that our statement justifying the use of isolated membranes at the end of the Introduction is valid.

   1. The authors measured the fluorescence decays as part of the evidence to show the stability of S2QA-. I have several concerns about these measurements:

   i. In figure 2B, the WL C. thermalis (blue) trace has a unique decay phase with a lifetime of about 0.2s, which the authors denoted as S2QA- recombination. Could the author elaborate on how this phase was assigned to this state?

   All decay kinetics in presence of DCMU are bi-phasic (with an additional faster phase in the WL and FR C. thermalis samples, attributed to a small fraction of centres where DCMU did not bind). In the manuscript we did originally assign both phases as arising from S2QA- recombination, but it is true that the middle phase, that is slightly faster in WL C. thermalis, is too fast to originate from that. This phase can rather be ascribed to TyrZ•(H+)QA- recombination occurring in a fraction of intact PSII centres before the full stabilization of charge separation, as shown in Debus et al. 2000 (https://doi.org/10.1021/bi992749w), or in centres lacking a Mn-cluster. We have now modified the paragraph regarding the fluorescence decay in presence of DCMU accordingly (L. 142-145): “The shorter lifetime (~0.22-1 s) of the middle decay phase (amplitude 15-20%) was compatible with it originating from TyrZ•(H+)QA- recombination occurring either in centres lacking an intact Mn-cluster (24) or in intact centres before charge separation is fully stabilised, as proposed in (23).”.

   A luminescence decay phase with a similar lifetime was initially ascribed, incorrectly, only to TyrZ•(H+)QA- recombination occurring in centres devoid of an intact Mn-cluster, in Appendix 5. This has now been rectified.

   ii. In figure S1 (the full version of 2B), all the fluorescence traces seem to rise at the end of the measurements. Could the authors check whether the measuring light intensity was actinic?

   This rise is significant only in the A. marina dataset (now Figure 2-figure supplement 1), and given the low signal to noise ratio in the last points of the fluorescence curve, we consider this small anomaly to be a measuring artefact. The rise is absent in the other traces in Figure 2- figure supplement 1 and in Figure 2B, except for the last point of the A. marina dataset in Fig. 2B. The corresponding Source data provided, shows that a rise in the last point of the measurements is only present in one of the three A. marina replicates (#2), while the non-decaying fluorescence is present in all A. marina samples and discussed in the text. Except for this last anomalous point, the decay curves of the A. marina replicate #2 do not differ significantly from the other two replicates. This clearly suggests an artefact, and is not consistent with the measuring light being actinic. A clarifying sentence has been added in the legend of Figure 2- figure supplement 1.

   iii. In figure S2, it seems to me that the fluorescence decay of Synechocystis + DCMU (Green open squares) was slower than the WL C. thermalis and is similar to the FRL C. thermalis in figure 2B. If the Synechocystis + DCMU is indeed similar to FR C. thermalis, would that be consistent with the authors' conclusions?

   When fitting the Synechocystis+DCMU fluorescence decay kinetics (in what is now Appendix 1-figure 1), we obtain two decay phases with, respectively: an amplitude of ~12% and lifetime of ~0.22 s, and an amplitude of ~81% and lifetime of ~7.9 s. These values are similar to those reported for WL C. thermalis in Table 1, with an overall fluorescence decay faster than in FR C. thermalis. Nonetheless, because of the limited number of Synechocystis biological replicates, we limit ourselves to a qualitative comparison. The luminescence decay kinetics are also faster in Synechocystis (as in WL C. thermalis) than in FR C. thermalis (now Figure 5- figure supplement 2).

   These data are consistent with our conclusions: the energy gap between QA- and Phe in Chl-f-PSII is at least as large as in Chl-a-PSII, or could even be larger, as suggested by the slower S2QA- recombination measured by fluorescence (Figure 2) and luminescence (Figure 3) decay.

   iv. It's known that DCMU will alter the redox potential of QA/QA- in plants. Would it have similar effects to the PSII studied in this research? If so, it will be meaningful to include these effects in the energy diagram in fig 7.

   Yes, we do expect DCMU to change the QA/QA- redox potential in our samples, as it does in plants and other cyanobacteria, although the actual effect in different PSII types would need to be measured. The energy gap values in now Figure 8 are only estimates based on literature values and on the relative changes reported here, they are not calculated from any of our data and do not specifically refer to the experimental conditions we used, including the use of DCMU. For this reason, we think that adding the effects of DCMU in the diagram would not be particularly useful and could be confusing.

   1. The authors didn't use WL C. thermalis for measuring oxygen evolution and the authors claimed that the PSII content in WL C. thermalis is too low. Is that a technical issue (e.g., cannot purify PSII enriched membranes) or a biological issue (i.e., white light condition produced less PSII)? In Fig S9C, the oxygen generated from WL C. thermalis is comparable to FR C. thermalis. Could the author explain how they reached the conclusion that PSII in WL C. thermalis was low? In addition, the author should also provide evidence showing that the samples of WL C. thermalis do not have significant PSII activity under far-red light.

   We did measure the flash dependence of oxygen evolution in WL C. thermalis membranes, and we did observe oscillations with visible flashes (but not with far-red flashes, as expected). However, the data were not good enough to be able to perform any significant analysis. Unfortunately, in the case of WL C. thermalis, we have not been able to isolate O2-evolving cores, as stated in L. 194-195. The WL C. thermalis data have now been added in Figure 3- figure supplement 1, together with the non-normalised traces of all other samples (following the suggestion by reviewer #3), and the text has been modified accordingly. The data in Figure 3- figure supplement 1 also provide evidence that the samples of WL C. thermalis do not have significant PSII activity under far-red light (although this was already clearly demonstrated in Nürnberg et al. 2018).

   We do have evidence that the PSII content per chlorophyll is lower in WL C. thermalis than in FR C. thermalis, based on fluorescence emission spectra, yield of isolated PSII and PSI from purification procedures, and O2 evolution per chlorophyll, as can be seen for example in Figure 3- figure supplement 1. The levels of PSII accumulation depend on the growth stage (among other factors) in model species such as Synechocystis. Since C. thermalis cells grow more slowly than other cyanobacteria species and their physiology has not been studied in detail yet, it is difficult to control the levels of PSII accumulation. This explains the inter-sample variability in the rates of O2 evolution per chlorophyll measured with the Clark electrode, that have now been added in Appendix 6-table 1.

   1. The authors used an indirect method, which used chemical trap histidine and oxygen consumption, for measuring the production of singlet oxygen from different types of PSII. I have several concerns about this approach.

   i. Why not use a probe that reacts directly with singlet oxygen probes like SOSG or EPR probes to unambiguously confirm the production of singlet oxygen? The difficulties of not using SOSG mentioned in Rehman et al (SI Ref#22) should be no longer problems when isolated membranes were used. The advantage would be a validation of the results and perhaps increased sensitivity.

   Although SOSG or EPR probes could also be used to detect singlet oxygen production, these other methods seem to be significantly less sensitive than histidine trapping. For example, Fufezan et al. 2007 (https://doi.org/10.1074/jbc.M610951200) used the EPR spin trap TEMPO and needed 30 minutes of illumination. Extended illumination (up to 1 hour) has also been used to detect singlet oxygen using SOGE (Flors et al 2006, https://doi.org/10.1093/jxb/erj181).With the histidine trapping method used here, less than 2 minutes of illumination were required to measure the singlet oxygen production rates. This allowed potential problems of prolonged illumination (e.g. a loss of intact PSII centres due to photodamage) to be minimised, and allowed us to confirm the results obtained in isolated membranes with those obtained in intact cells.

   As shown in now Figure 6- figure supplement 1E, the histidine-dependent oxygen consumption was suppressed by the singlet oxygen quencher sodium azide, as also shown in Rehman et al. 2013 (https://doi.org/10.1016/j.bbabio.2013.02.016). We also independently confirmed that the singlet oxygen generated by illumination of the dye Rose Bengal can be efficiently detected with the histidine trapping method and suppressed by the addition of sodium azide (Figure 6- figure supplement 1F). For these reasons, we are confident that what we measure with the histidine trapping method is singlet oxygen production.

   ii. In Rehman et al (SI Ref#22), wild-type Synechocystis cells showed significant production of singlet oxygen in the presence of DCMU and His (Figure 3A in SI Ref#22), however, the amount of singlet oxygen measured from the membranes in this study seemed to be less (Fig S10E). Could the authors provide some explanations?

   Fig. 3A in Rehman et al. showed that the production of singlet oxygen was about 10% with respect to the oxygen evolution activity in absence of additions (open squares). The light saturation curves in Fig. 4B of the same paper also show that at saturating light intensity the singlet oxygen production rate is about 10% compared to the O2 evolution rate. The traces we show in Figure 6-figure supplement 1 are only representative. The comparison should be made between the results in Rehman et al. and the averages of biological replicates that we show in Fig. 6 (membranes) and Appendix 6-figure 4A (cells). For WL and FR C. thermalis, we measure singlet oxygen production rates that are about 20% of the O2 evolution rates, slightly higher than those measured in Synechocystis in Rehman et al. Considering the variability between biological replicates, we consider our values in line with those in Rehman et al.

   iii. Can the presented results distinguish the production of singlet oxygen from recombination or other sources (e.g., antenna, free chlorophyll)? Some key controls are needed to strengthen the authors' claims.

   This is difficult to demonstrate unequivocally, but we have different lines of evidence that support the conclusion that the increase in singlet oxygen production in A. marina originates from differences in PSII charge recombination with respect to the other samples:

   i) The high levels of singlet oxygen production are observed in intact cells as well as in membranes. In neither of these samples do we expect to have significant amounts of damaged PSII or free chlorophyll, so these seem highly unlikely as the main sources of the singlet oxygen in our measurements. This is now stated more explicitly in L. 305 and Appendix 6.

   ii) According to the data in Appendix 6-figure 1B, singlet oxygen production in A. marina membranes shows a similar light saturation to that of maximal O2 evolution. This suggests that the singlet oxygen production we measure is related to PSII photochemistry. We have now stated this explicitly in L. 288-290.

   iii) Our thermoluminescence and delayed luminescence results indicate that in Chl-d-PSII the energy gap between Phe and QA is smaller than in Chl-a-PSII, as already suggested in the literature, and Chl-f-PSII. Therefore, this indicates more charge recombination going via repopulation of Phe- in Chl-d-PSII, with a consequent increase of singlet oxygen production.

   The antenna chlorophylls could form triplets under high light, by inter-system crossing, but in intact antennas the chlorophyll triplets are expected to be mostly quenched by nearby carotenoids (see https://www.jstor.org/stable/24030848 for a review on the subject). The generation of antenna triplet states in non-photoinhibitory conditions has been demonstrated in plant and algal thylakoids (Santabarbara et al 2002, 2007 doi: 10.1021/bi0201163, doi: 10.1016/j.bbabio.2006.10.007). Yet, these signals, which are attributed to a small population of damaged antennas, are small compared to those of triplets generated by charge recombination. Due to its apparently stochastic nature, the generation of antenna triplets by inter-system crossing is not expected to be significantly different between the different PSII complexes investigated in this study.

   On the other hand, it is generally recognised that in the PSII reaction centre, the carotenoid on the D1 side is not close enough to ChlD1 to directly quench its triplet state, when formed (see Telfer et al. 1994, https://doi.org/10.1016/S0021-9258(17)36825-4). The singlet oxygen produced in the reaction centre could disrupt the coupling between chlorophylls and carotenoids in the antenna, resulting in singlet oxygen production also from the antenna, in a cascade effect. This can happen with prolonged strong illumination (Fufezan et al. 2002, https://doi.org/10.1016/S0014-5793(02)03724-9).

   iv. I could not fully understand the singlet oxygen production experiments with tris-washed samples. In my opinion, the Mn-cluster depleted PSII should have accelerated charge recombination (100 ms between the YZ/QA, vs ~ 5 sec between the S2/QA), which should lead to an increase in singlet oxygen production. Correct me if I'm wrong about this, but if my reasoning is correct then how do the authors explain the discrepancy?

   Our rationale for performing the tris-washing experiment was indeed to see if this would lead to an increase in singlet oxygen production, thus implying that the high production in the A. marina samples could arise from a higher fraction of PSII centres without the Mn-cluster, as explained both in the main text and in Appendix 6. The fact that the treatment did not increase the singlet oxygen production suggests that this does not specifically arise from PSII lacking the Mn-cluster.

   The lack of singlet oxygen increase following tris-washing is not necessarily controversial, as the fact that TyrZ•QA- recombination is faster than S2QA- recombination does not necessarily imply that more of it occurs via backward electron transfer from QA- to Phe. The removal of the Mn-cluster could decrease the production of singlet oxygen by charge recombination, since it causes an increase in the redox potential of QA and, therefore, of the energy gap between Phe and QA, thus decreasing the probability of charge recombination going via the repopulation of Phe-. This is proposed to be a mechanism to protect PSII during photoactivation of the Mn-cluster (see Johnson et al 1995, https://doi.org/10.1016/0005-2728(95)00003-2).

   Our data show that the singlet oxygen production in A. marina is not specifically related to PSII lacking the Mn-cluster and are not in conflict with what is expected based on our knowledge of PSII energetics.

   v. The y-axes in Figure S10 should either contain "delta" (Δµmol O2 ml-1) or use the measured absolute oxygen concentration. I'd suggest the latter, since the reaction is oxygen consuming, it's good to show that all the samples started with similar amounts of dissolved oxygen. Low O2 levels could decrease 1O2 production, though this would be more of an issue with cells than membranes.

   The y-axis labels in the figures (now Figure 6-supplementary figure 1 and Appendix 6-figures 1D and E, 2, 3 and 4A) have been changed to Δµmol O2 ml-1. We prefer to show the traces after subtraction of the baseline recorded in the dark (now explicitly indicated in the corresponding figure legends) for a better visual comparison. All samples were left to equilibrate with air (stirred) before starting the measurements, so all started with similar levels of dissolved oxygen. This is especially important when measuring PSI-dependent oxygen consumption (Appendix 6-figure 3), because the addition of ascorbate and TMPD leads to a transient drop in oxygen concentration in the sample, which leads to artefacts in absence of the equilibration step. This information has been added to the corresponding Materials and Methods section (4.5). Additionally, when using Rose Bengal to generate singlet oxygen, the histidine-dependent oxygen consumption was about 10 times higher than in any of the measurements done with biological samples, and still we did not observe saturation of the signal in the illumination time used (added panel F in Figure 6- figure supplement 1). Therefore, we are confident that the singlet oxygen measurements in membranes and cells were not skewed by limiting oxygen concentrations in the measuring chamber.

   The y-axis labels of what is now Appendix 6-figure 1B and C have also been corrected (as ml-1 was used instead of h-1).

   Reviewer #3 (Public Review):

   In this manuscript, Viola and co-authors address the question of how far-red-light-adapted (FRL) Photosystem II (PSII) is able to bypass the "red limit", or the minimum photon energy/frequency for charge separation to proceed effectively. They attempt to do so primarily by measuring the consequence of failure to overcome the red limit: charge recombination. From this work they have concluded that FRL PSIIs are able to achieve similar efficiency of flash-induced water-oxidizing complex turnover to those adapted to standard visible light. However, they conclude that FRL PSII which uses chlorophyll-d is significantly more susceptible to charge recombination and singlet oxygen formation, leading to increased sensitivity to high-light conditions. FRL PSII which uses chlorophyll-f, however, is adapted to be more resistant to photodamage. These strategies are differentiated by the number and type of far-red chlorophyll used and tuning of redox potentials of cofactors in PSII.

   The methods employed are well-chosen to present complementary evidence to address the questions posed. The authors have supported themselves using polarography, fluorescence decay, absorption, luminescence and thermoluminescence, and spectrometry, all of which are employed in a manner well-established in the quantification of processes in standard PSII preparations. The results, however, have some loss of data such as total yields which would be useful in interpretation as the authors have chosen to extensively normalize data for ease of visual comparison of certain features.

   Overall, the authors have adequately achieved their aims and their conclusions are well-supported. The authors also clearly state their own expectations of the impact of their work at the end of the Discussion; thanks to these results, we can better understand the ecological niche of each type of FRL-PSII and how these significantly disparate systems may be used in future agricultural research and development.

   We thank the reviewer for the positive evaluation of our work.

   Following the reviewer’s suggestions, the total yields (on a chlorophyll basis) of the flash-dependent oxygen evolution have been provided in Figure 3- figure supplement 1. These include the flash-dependent oxygen evolution data measured in WL C. thermalis membranes, that were previously omitted because of the unsatisfactory quality, and are still omitted from Figure 3 (normalised data and fits) for the same reason. The S-state distributions calculated from the fits of the flash-dependent oxygen evolution have been added in Table 2.

   Additionally, the non-normalised oxygen evolution and consumption rates used for Figure 6A and Appendix 6-figure 4 are now provided in Appendix 6-table 1.

   Read the original source
3. 

   eLife

   Jun 24, 2022

   Evaluation Summary:

   The work describes the energetic constraints and preferred operating conditions of these "strategies" in particular on how nature has solved the problem of low energy "headroom'" required to prevent deleterious back reactions while maintaining efficient energy storage. The differences between the species are quite interesting and show that nature has evolved multiple solutions to fundamental limitations. Given the importance of understanding and improving the efficiency of photosynthesis, and the new insights revealed, the work will be of interest to a broad audience.

   (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 and Reviewer #2 agreed to share their names with the …

   Evaluation Summary:

   The work describes the energetic constraints and preferred operating conditions of these "strategies" in particular on how nature has solved the problem of low energy "headroom'" required to prevent deleterious back reactions while maintaining efficient energy storage. The differences between the species are quite interesting and show that nature has evolved multiple solutions to fundamental limitations. Given the importance of understanding and improving the efficiency of photosynthesis, and the new insights revealed, the work will be of interest to a broad audience.

   (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 and Reviewer #2 agreed to share their names with the authors.)

   Read the original source
4. 

   eLife

   Jun 24, 2022

   Reviewer #1 (Public Review):

   Viola et. al. compared the electron transfer efficiency of two types of oxygenic far-red photosystem II (PSII) with the "conventional" PSII and analyzed how these far-red PSII use the limited energy from infrared photons to proceed photosynthesis. Oxygenic photosynthesis is an energy-intensive process, and a large headroom is also needed for preventing harmful back-reactions from occurring, which can produce singlet oxygen. This research investigated how the far-rad PSII managed to do their work with limited energy.

   The authors measured and compared the forward reactions of different kinds of PSII (Chl-a-PSII, Chl-d-PSII and Chl-f-PSII), including the flash-induced chlorophyll fluorescence decay and S-states turnover. These results led to a conclusion that the forward reaction quantum efficiency was not …

   Reviewer #1 (Public Review):

   Viola et. al. compared the electron transfer efficiency of two types of oxygenic far-red photosystem II (PSII) with the "conventional" PSII and analyzed how these far-red PSII use the limited energy from infrared photons to proceed photosynthesis. Oxygenic photosynthesis is an energy-intensive process, and a large headroom is also needed for preventing harmful back-reactions from occurring, which can produce singlet oxygen. This research investigated how the far-rad PSII managed to do their work with limited energy.

   The authors measured and compared the forward reactions of different kinds of PSII (Chl-a-PSII, Chl-d-PSII and Chl-f-PSII), including the flash-induced chlorophyll fluorescence decay and S-states turnover. These results led to a conclusion that the forward reaction quantum efficiency was not changed between "conventional" PSII and far-red PSII. However, the back-reactions of three types of PSII are different based on the measurements of the prompt fluorescence decay, delayed luminescence decay, and thermoluminescence band locations. The authors concluded that the two far-red PSII (Chl-d-PSII and Chl-f-PSII) have a different strategy for utilizing infrared light. Indeed, the authors showed that Chl-d-PSII containing cyanobacteria produced more singlet oxygen than other types, and this result was explained by the energy profile in the electron transfer chain.

   The major strength of this research is the authors made a direct comparison of different far-red PSII under the same conditions. It's exciting to have a side-by-side comparison between two types of far-red PSII. In addition, the authors also measured the singlet oxygen produced from all types of PSII which clearly showed the differences in the routes of recombination.

   However, there are some concerns:

   1. The flash-induced fluorescence decay, thermoluminescence, delayed luminescence and S-states turnovers of the Chl-d-PSII and Chl-f-PSII have been characterized before (ref 5, 26, 39), but from intact cells compared to isolated membranes in this study, and similar conclusions have been achieved. The authors mentioned four reasons (lines 115-120, see the manuscript for the authors' arguments "i." to "iv.") why it's important to use isolated membranes. However, in my opinion, these reasons are not sufficiently strengthened:
   i. The transmembrane potentials from cells can be collapsed by adding uncouplers;
   ii. The authors mentioned the quinone pool in the cells is uncontrollable, but the authors didn't actually measure or manipulate the quinone pool in the membrane (e.g., the ratio of QB/QB-/empty-pocket in the samples);
   iii. The phycobilisomes can be controlled by different conditions through state transitions;
   iv. The isolation of membranes may not remove membrane-related quenching mechanisms (e.g., PSII quenching in State II, spillover, etc.).

   In addition, the authors reached a conclusion that the Chl-f-PSII containing species should suffer from fluctuation light-induced membrane potential spikes, but don't actually measure this in physiologically relevant preparations. It will be more beneficial to use intact cells instead of an isolated membrane. I suggest the authors either restrict their conclusions to what the isolated membranes clearly show or make measurements in intact cells.

   2. The authors measured the fluorescence decays as part of the evidence to show the stability of S2QA-. I have several concerns about these measurements:
   i. In figure 2B, the WL C. thermalis (blue) trace has a unique decay phase with a lifetime of about 0.2s, which the authors denoted as S2QA- recombination. Could the author elaborate on how this phase was assigned to this state?
   ii. In figure S1 (the full version of 2B), all the fluorescence traces seem to rise at the end of the measurements. Could the authors check whether the measuring light intensity was actinic?
   iii. In figure S2, it seems to me that the fluorescence decay of Synechocystis + DCMU (Green open squares) was slower than the WL C. thermalis and is similar to the FRL C. thermalis in figure 2B. If the Synechocystis + DCMU is indeed similar to FR C. thermalis, would that be consistent with the authors' conclusions?
   iv. It's known that DCMU will alter the redox potential of QA/QA- in plants. Would it have similar effects to the PSII studied in this research? If so, it will be meaningful to include these effects in the energy diagram in fig 7.

   3. The authors didn't use WL C. thermalis for measuring oxygen evolution and the authors claimed that the PSII content in WL C. thermalis is too low. Is that a technical issue (e.g., cannot purify PSII enriched membranes) or a biological issue (i.e., white light condition produced less PSII)? In Fig S9C, the oxygen generated from WL C. thermalis is comparable to FR C. thermalis. Could the author explain how they reached the conclusion that PSII in WL C. thermalis was low? In addition, the author should also provide evidence showing that the samples of WL C. thermalis do not have significant PSII activity under far-red light.

   4. The authors used an indirect method, which used chemical trap histidine and oxygen consumption, for measuring the production of singlet oxygen from different types of PSII. I have several concerns about this approach.
   i. Why not use a probe that reacts directly with singlet oxygen probes like SOSG or EPR probes to unambiguously confirm the production of singlet oxygen? The difficulties of not using SOSG mentioned in Rehman et al (SI Ref#22) should be no longer problems when isolated membranes were used. The advantage would be a validation of the results and perhaps increased sensitivity.
   ii. In Rehman et al (SI Ref#22), wild-type Synechocystis cells showed significant production of singlet oxygen in the presence of DCMU and His (Figure 3A in SI Ref#22), however, the amount of singlet oxygen measured from the membranes in this study seemed to be less (Fig S10E). Could the authors provide some explanations?
   iii. Can the presented results distinguish the production of singlet oxygen from recombination or other sources (e.g., antenna, free chlorophyll)? Some key controls are needed to strengthen the authors' claims.
   iv. I could not fully understand the singlet oxygen production experiments with tris-washed samples. In my opinion, the Mn-cluster depleted PSII should have accelerated charge recombination (100 ms between the YZ/QA, vs ~ 5 sec between the S2/QA), which should lead to an increase in singlet oxygen production. Correct me if I'm wrong about this, but if my reasoning is correct then how do the authors explain the discrepancy?
   v. The y-axes in Figure S10 should either contain "delta" (Δµmol O2 ml-1) or use the measured absolute oxygen concentration. I'd suggest the latter, since the reaction is oxygen consuming, it's good to show that all the samples started with similar amounts of dissolved oxygen. Low O2 levels could decrease 1O2 production, though this would be more of an issue with cells than membranes.

   Read the original source
5. 

   eLife

   Jun 24, 2022

   Reviewer #2 (Public Review):

   The authors have presented a thorough investigation aimed to understand the photochemical efficiency and resilience to photodamage of two kinds of cyanobacterial photosystem II (PSII) systems that use low-energy absorbing Chl-d and Chl-f molecules (with respect to the typical higher-energy absorbing Chl-a found in the majority of oxygenic photosynthetic organisms) to perform the energy demanding water splitting and quinone reduction processes. To that end, the authors have comprehensively studied a collection of PSII systems isolated from different organisms and grown under different light conditions, and have applied a vast compendium of spectroscopic and electrochemical methods. On the systems side, the "systems of interest" are Chl-d-PSII from Acaryochloris marina, Chl-f-PSII from Chroococcidiopsis …

   Reviewer #2 (Public Review):

   The authors have presented a thorough investigation aimed to understand the photochemical efficiency and resilience to photodamage of two kinds of cyanobacterial photosystem II (PSII) systems that use low-energy absorbing Chl-d and Chl-f molecules (with respect to the typical higher-energy absorbing Chl-a found in the majority of oxygenic photosynthetic organisms) to perform the energy demanding water splitting and quinone reduction processes. To that end, the authors have comprehensively studied a collection of PSII systems isolated from different organisms and grown under different light conditions, and have applied a vast compendium of spectroscopic and electrochemical methods. On the systems side, the "systems of interest" are Chl-d-PSII from Acaryochloris marina, Chl-f-PSII from Chroococcidiopsis thermalis (grown under far-red light); as well as a variety of "control" samples: Chl-a-PSII from Chroococcidiopsis thermalis (grown under white light), from Synechocystis, and from Thermosynechococcus elongatus. On the methods side, the techniques are fluorescence, thermoluminescence and luminescence, measurement of oxygen evolution and consumption rates, determination of flash-dependent oxygen evolution with Joliot electrode, and UV transient absorption. In my view, the strength of this work resides in the combination of samples and methods, the authors do not restrict themselves to a single sample and/or method, they look at their systems from many different angles to obtain a complete physico-chemical picture of how these low-energy PSII systems work. Taken all together, the authors have presented a careful and elegant piece of work where each portion of information is fully supported by the data presented in the main text, and additionally, further confirmed in the extensive dataset presented in the supplemental information. Here it is worth noting the large number of experimental replicates presented, another sign of the great care the authors are taking to present consistent data. In addition, the authors guide the reader through all the results with clear and concise explanations, providing different possible explanations whenever the results do not provide an unambiguous answer. This is especially relevant considering that the authors are dealing with extremely complex systems. In this respect, the concise and illustrative figures presented in figures 1 and 7 are particularly useful to follow the information provided in the text. I consider that this work is highly relevant for the photosynthesis research field since it provides ample information on how these low-energy-Chl PSII systems are still capable of splitting water and reducing quinone (something that was long thought to be only possible with the higher-energy Chl-a molecule), and on which is the price these systems have to pay to perform these energy demanding processes with lower energy. Furthermore, this work teaches us valuable lessons on how to balance energy gaps and redox potentials to achieve efficient charge separation and avoid photodamage. All this information has far-reaching implications since it should be taken into account when designing artificial systems for solar-to-electricity and solar-to-fuel conversion, and it can be employed for agricultural and biotechnological applications in environments depleted of high-energy photons but rich in low-energy photons such as canopies.

   Read the original source
6. 

   eLife

   Jun 24, 2022

   Reviewer #3 (Public Review):

   In this manuscript, Viola and co-authors address the question of how far-red-light-adapted (FRL) Photosystem II (PSII) is able to bypass the "red limit", or the minimum photon energy/frequency for charge separation to proceed effectively. They attempt to do so primarily by measuring the consequence of failure to overcome the red limit: charge recombination. From this work they have concluded that FRL PSIIs are able to achieve similar efficiency of flash-induced water-oxidizing complex turnover to those adapted to standard visible light. However, they conclude that FRL PSII which uses chlorophyll-d is significantly more susceptible to charge recombination and singlet oxygen formation, leading to increased sensitivity to high-light conditions. FRL PSII which uses chlorophyll-f, however, is adapted to be more …

   Reviewer #3 (Public Review):

   In this manuscript, Viola and co-authors address the question of how far-red-light-adapted (FRL) Photosystem II (PSII) is able to bypass the "red limit", or the minimum photon energy/frequency for charge separation to proceed effectively. They attempt to do so primarily by measuring the consequence of failure to overcome the red limit: charge recombination. From this work they have concluded that FRL PSIIs are able to achieve similar efficiency of flash-induced water-oxidizing complex turnover to those adapted to standard visible light. However, they conclude that FRL PSII which uses chlorophyll-d is significantly more susceptible to charge recombination and singlet oxygen formation, leading to increased sensitivity to high-light conditions. FRL PSII which uses chlorophyll-f, however, is adapted to be more resistant to photodamage. These strategies are differentiated by the number and type of far-red chlorophyll used and tuning of redox potentials of cofactors in PSII.

   The methods employed are well-chosen to present complementary evidence to address the questions posed. The authors have supported themselves using polarography, fluorescence decay, absorption, luminescence and thermoluminescence, and spectrometry, all of which are employed in a manner well-established in the quantification of processes in standard PSII preparations. The results, however, have some loss of data such as total yields which would be useful in interpretation as the authors have chosen to extensively normalize data for ease of visual comparison of certain features.

   Overall, the authors have adequately achieved their aims and their conclusions are well-supported. The authors also clearly state their own expectations of the impact of their work at the end of the Discussion; thanks to these results, we can better understand the ecological niche of each type of FRL-PSII and how these significantly disparate systems may be used in future agricultural research and development.

   Read the original source
7. 

   eLife

   Jun 24, 2022

   Reviewer #4 (Public Review):

   Viola et al. report various spectroscopic results whose interpretations are used to describe the differences in energy dynamics between cyanobacterial PSII systems using different pigment types. They convincingly show (1) that all three systems are similarly efficient, (2) that PSII containing primarily Chl d broadens its absorption cross-section at the expense of increased photodamage, and (3) that the PSII system containing only a few Chl f/d sites is tuned to avoid back reactions which allows for robustness similar to systems containing only Chl ¬a, but this also limits its ability to implement more far-red pigments that would enhance the absorption cross-section. The work is interesting, thoughtful, and of great importance for future efforts in enhancing crops that have improved shade tolerance.

   Read the original source
8. 

   Version published to 10.1101/2022.04.05.486971v2 on bioRxiv

   Apr 30, 2022
9. 

   Version published to 10.1101/2022.04.05.486971v1 on bioRxiv

   Apr 6, 2022