Two dominant boreal conifers use contrasting mechanisms to reactivate photosynthesis in the spring

Boreal forests are dominated by evergreen conifers that show strongly regulated seasonal photosynthetic activity. Understanding the mechanisms behind seasonal modulation of photosynthesis is crucial for predicting how these forests will respond to changes in seasonal patterns and how this will affect their role in the terrestrial carbon cycle. We demonstrate that the two co-occurring dominant boreal conifers, Scots pine (Pinus sylvestris L.) and Norway spruce (Picea abies), use contrasting mechanisms to reactivate photosynthesis in the spring. Scots pine downregulates its capacity for CO2 assimilation during winter and activates alternative electron sinks through accumulation of PGR5 and PGRL1 during early spring until the capacity for CO2 assimilation is recovered. In contrast, Norway spruce lacks this ability to actively switch between different electron sinks over the year and as a consequence suffers severe photooxidative damage during the critical spring period.

1. The authors well discussed the mechanism of Norway spruce coping with springtime photodamage, i.e. alternative electron flow (AEF) under down-regulated photosynthetic capacity. The reviewer recommends the authors to discuss the strategy of Norway spruce during the critical spring recovery phase as well, from the view point of forest succession. As the authors describe, Scots pine and Norway spruce is classified as pioneer and late successional climax species, respectively. As late successional species is often found in the forest understory, photosynthetic carbon gain during early spring before the flush of canopy trees (especially in the case of deciduous tree species such as European Beech) is substantially important for growth and survival of the forest-floor seedlings. In this context, relatively higher An in Norway spruce observed early in the spring (Fig. 1a, b, e) may be a strategy of this species for efficient carbon gain. Furthermore, shade-tolerance or shade-preference of this species may be a way of circumventing photooxidative damage without enhanced AEF.
2. Budbreak, i.e. the onset of new shoot development, is physiologically relevant for evergreen conifer trees, since pre-existing shoots act as a sink of photosynthate before budbreak, but as a source of new shoot development after budbreak; starch content changes drastically in both species around budbreak (Egger et al., 1996;Wyka et al., 2016). Therefore, the authors should indicate the date of budbreak for each species in the related figures.
Minor comments 1. [Line 320] Light intensity (150 µmol m-2 s-1) for the growth chamber seedlings was substantially low (≈10 % of full sunlight). In this case, it should be noted that needles from mature trees were sun-acclimated, but those from growth chamber seedlings were shadeacclimated. Overall, I found that the work is novel both for how these co-occurring species deal with seasonal changes in climate, but also for its broad implications for how these two globally common species respond to changes in temperature. The data are comprehensive, spanning anatomy, fluorescence and gas exchange, biochemistry and gene expression results from a suite of interconnected field and chamber experiments. The paper is well written and the arguments are clear and convincing.
Having said that, I do have some comments that would clarify the analyses. The statistical approach is quite mixed -some data have no stats, which is unacceptable. While many of the results are obviously significant, all of them need to be backed with stats. Figure 3 -in panel b, there's a marked and unusual disjunct in the ACi data around 250 ppm Ci. Why is this? If it's a break caused by the measurements protocol (i.e from moving down in CO2 and then moving back up to high CO2), then I would be concerned about the quality of the data. If it's a real biological phenomenon, then I think this needs some explanation. The matching data in panel c does not have this issue, which gives me confidence that it's more likely biological than protocol-based.
The authors discuss some of the differences in the ecology of these species. Are there other data pointing to differences in how they might cope with environmental conditions? Is the "superior" strategy of pine indicative of a greater capacity to cope with environmental change in general, or is its advantage limited to seasonal responses of photosynthesis?
While I think that the novel contrasting strategies outlined here are important, photosynthetic recovery and performance are not growth. As such, the legend in Figure 6 need to be toned down.

Response to Reviewer 1
It is very interesting that the amount of PGR5 is much lower in spruce than in pine. However, it is not convincingly shown that this is due to cyclic electron flow. A convincing determination of cyclic electron flow is difficult. It is heavily debated if the ratio between Y(I) and Y(II) can be interpreted as a sign of cyclic electron flow. One has to keep in mind that Y(I) is overestimated with the Dual-PAM. Additional measurements like electrochromic shift are needed to show that it is indeed cyclic electron flow that is responsible for the observed differences. To measure only Y(I) and Y(II) is insufficient. It may well be that something else is changed in the properties of PSI and that, for example, charge recombination reactions within PSI are increased in pine and not in spruce. Or that Mehler reaction/activity of flavodiiron proteins is higher in pine than in spruce. This would give the same effect on the measured parameters. The effect of animycin is not very convincing. The decrease in Y(I) by antimycin is rather small and even smaller in pine than in spruce. The effect of the recovery from antimycin treatment is hard to see (Fig. 5d). The signals in spruce are much larger than for pine (and therefore easier to see) and it seems that only a part of the signal is recovered after 24h. At least for the early time point (2-7 min), the effect seems to be similar in spruce and pine. It is an interesting observation that the amount of PGR5 is lowered after room temperature recovery (Fig. 5 e). -Thank you for these insightful comments. We have addressed them to the best of our abilities. We must note that the main message of the work presented is how differential regulation of PSI function and CO 2 assimilation capacity underpin the photosynthetic recovery process (winter to spring and summer) in spruce and pine. Our intention was not to present evidence for the molecular bases of cyclic electron transport (CET). We might have mis-lead the reader to get this impression by overstating some of the differences between pine and spruce and we sincerely apologize for that. We have now carefully gone through the manuscript to correct for such phrasings in the text.
Specifically, our findings indicate that the regulation of PGR5 expression in pine is one of the central aspects that provides this species with a better protection against PSI acceptor-side limitation compared to spruce. The other aspect that contributes to this capacity is the presence of flavodiiron proteins (FLVs) which also provides a large electron sink at the PSI acceptor-side. However, FLV levels remain stable during the whole period of the study according to our data ( Fig. 5a and 5b). The coordinate function of PGR5 and FLV is the mechanism behind the dynamic capacity observed in pine to avoid the redox imbalance as shown by the fluctuating light experiment (Fig. 4) Although the induction of PGR5 strongly influences AEF during winter to spring transition, we cannot attribute the better performance of pine to cyclic electron transport (CET) -as you correctly pointed out, the molecular bases of CET still remains elusive. In addition, as you acknowledged, there is no efficient method to quantify CET, not even by electrochromic shift (ECS) in the presence of FLVs. The measurement of pmf and their components ΔpH and ΔΨ cannot provide reliable support for CET. Yamamoto and coworkers have used ECS to evaluate the contribution of FLV in the pgr5 and wild-type backgrounds and concluded that both PGR5 and FLV contribute to pmf (Yamamoto et al., 2016; Shikanai and Yamamoto, 2017). We did not claim that "ratio between Y(I) and Y(II) can be interpreted as a sign of cyclic electron flow", rather as an indicator of AEF to photosynthetic LEF. Recently, Grebe and co-workers (Grebe et al., 2019) revealed the existence of a large subpopulation of PSI, named PSI*, with different antenna compositions. The authors proposed that this PSI* subpopulation can generate a photoprotective mechanism including FLV and/or as a supplementary structure under demanding conditions. Immunoblots (Fig. 5a, b): The question arises why PGR5 levels are still very high in June. Fig. 5a: It would be nice to show data also on PGRL1 to exclude that the antibodies directed against PGR5 don't recognize well the spruce protein. In addition, the level of the NDH complex should be checked. It may also differ between the two species and depending on the time of the year. Why PGR5 levels are still very high in June: In the publication by Suorsa et al from 2012 it was shown that high light and fluctuating light could induce the accumulation of the PGR5 protein, which was also suggested to indicate a protective function of PGR5 under such stress conditions. In Northern Sweden, the days are very long in June (about 20 hours light per day) and the temperature could still occasionally be low also at this time of year, this could potentially generate conditions to induce the accumulation of PGR5 resulting in maintaining rather high levels of the protein also after the critical spring period.

The result presented by Grebe et al is in line with our view that the ratio Y(AEF)= Y(I)-Y(II) is not a specific indicator of CEF, but an indicator of alternative electron fates at the PSI acceptor-side. Although other variables (e.g. specific PSI structures, different FLV-PSI affinity constants or species-specific variations in Mehler reaction) might contribute to the differences in PSI function between pine and spruce, the combined effect of PGR5 and FLV on the PSI redox status can be considered as major contributors to the photoprotective mechanisms during the winter-spring transition. The Antimycin
Antibodies directed against PGR5 don't recognize well the spruce protein: The PGR5 sequence alignment between pine and spruce as was shown in the previous Figure  S9 (now Figure S11) demonstrated that PGR5 from the two conifer species are highly conserved. They are also both very similar to the Arabidopsis PGR5. The same is true for PGRL1, and the PGRL1 sequence alignment for pine and spruce has now been included in the new Fig. S11.
It would be nice to show data also on PGRL1: Thank you for the good suggestion to investigate also the levels of the PGRL1 protein during the spring summer transition. Figure L. 193,194. In Fig. 5d, the Y(I) in pine is 0.6 and in spruce 0.8. How to explain this compared with Fig. 4d? -The samples in Fig. 4d and Fig. 5d were collected in 2017 and 2018, respectively. In the spring 2018 the recovery of photosynthesis in Norway spruce was observed slightly earlier compared to Scots pine. All samples were collected in April, but due to marginally warmer temperatures during this period (which is also seen in the earlier bud break 2018 compared to 2017) slightly higher Y(I) was observed in spruce 2018 compared to 2017. The difference between the years for Y(I) was only observed in Norway spruce under low light and we feel that the broad reproducibility of these findings from field collected samples across two different growing seasons strengthens the findings we report.
Response to Reviewer 2 1. The authors well discussed the mechanism of Norway spruce coping with springtime photodamage, i.e. alternative electron flow (AEF) under down-regulated photosynthetic capacity. The reviewer recommends the authors to discuss the strategy of Norway spruce during the critical spring recovery phase as well, from the view point of forest succession. As the authors describe, Scots pine and Norway spruce is classified as pioneer and late successional climax species, respectively. As late successional species is often found in the forest understory, photosynthetic carbon gain during early spring before the flush of canopy trees (especially in the case of deciduous tree species such as European Beech) is substantially important for growth and survival of the forest-floor seedlings. In this context, relatively higher An in Norway spruce observed early in the spring (Fig. 1a, b, e) may be a strategy of this species for efficient carbon gain. Furthermore, shade-tolerance or shade-preference of this species may be a way of circumventing photooxidative damage without enhanced AEF. -Thank you for this very interesting comment. We have included a section in the discussion to bring up these important points. The new text is high-lighted in the document and can be found in the last paragraph of the discussion. (Egger et al., 1996;Wyka et al., 2016). Therefore, the authors should indicate the date of budbreak for each species in the related figures. -Thank you for this suggestion, we have included data for budbreak for both species from the experimental site in Figure 1b and (150 µmol m-2 s-1) for the growth chamber seedlings was substantially low (≈10 % of full sunlight). In this case, it should be noted that needles from mature trees were sun-acclimated, but those from growth chamber seedlings were shade-acclimated.

[Line 333] Y(NA) and Y(ND) should be explained here.
-Thank you, we corrected this.

[Line 352]
Please explain why using Ca of 800 µmol mol -1 , twice higher than the ambient air.
-We used 800 µmol mol -1 to assess the full capacity in the needles for CO 2 assimilation during the spring and summer period as we were interested in potential differences in the enzymatic capacities in the Calvin cycle etc. between the two species during the recovery phase. Our data showed the CO 2 assimilation rate was very low (below zero) in the early spring samples (March and April) under 400 µmol mol -1 . However, we have now included the data collected for 400 µmol mol -1 as the new supplementary Figure S2 and refer to it in the text. Figure 6] Based on Figure 1c, no significant difference seems to be observed in ETR(II) of Scots pine between spring and summer. Please check the blue lines through PSII. Figure S10, the ETR(II) in pine in summer samples was higher compared to the spring samples under different light intensities, especially from 100 to 500 µmol m -2 s -1 the differences were significant (p<0.05). In Figure 1c, the ETR(II) was calculated under saturated light intensity of 1292 µmol m -2 s -1 . Due to the large variation of ETR(II) under high light, the differences became less pronounced as presented in Figure 1c.

Response to Reviewer 3
The statistical approach is quite mixed -some data have no stats, which is unacceptable. While many of the results are obviously significant, all of them need to be backed with stats.
-Thank you for pointing this out. It was a mistake on our part. We have now included proper and uniform statistical analysis using ANOVA for Figure 2 Figure 3e shows the stomatal conductance values (g s ), which were 0.01 and 0.03 mol CO 2 m -2 s -1 for Pine and Spruce field samples, respectively, and 0.06 mol CO 2 m -2 s -1 for both recovered species. The response to the warm temperature was indeed slow and would explain the observed a gap during the measurement. Moreover, at low concentrations of CO 2 , most A N values are near zero. This suggests that Rubisco, Calvin and Krebs cycles, and diffusion (aquaporins) etc., needs to be activated in response to the warm temperatures. Before each measurement, we placed the needles in the leaf chamber of the IRGA under 1400 µmol m -2 s -1 light intensity and 25 °C to equilibrate for 15 minutes. The field samples need longer time to pre-accumulate to the conditions used for the measurements. However, we also wanted to limit the recovery of photosynthesis to reflect the photosynthetic abilities in the field. We have included a note in the manuscript about the disjunct in the ACi data referring to the g s , line 168 and onwards.
The authors discuss some of the differences in the ecology of these species. Are there other data pointing to differences in how they might cope with environmental conditions? Is the "superior" strategy of pine indicative of a greater capacity to cope with environmental change in general, or is its advantage limited to seasonal responses of photosynthesis? -It was recently shown that in response to increased seasonal temperatures and elevated CO 2 Scots pine is more able compared to Norway spruce to acclimate photosynthesis and respiration. Pine demonstrated increasing growth in response to temperature increases as high as +8 °C, whereas spruce showed minimal capacity to acclimate energy metabolism and suffered growth losses at elevated seasonal temperatures (Kurepin et al, 2018). We feel that pine generally is the more responsive of the two species and better able to acclimate to environmental fluctuations and we have included this information in the discussion. The new text is high-lighted in the word file and found in the last paragraph of the discussion. While I think that the novel contrasting strategies outlined here are important, photosynthetic recovery and performance are not growth. As such, the legend in Figure  6 need to be toned down.
-Thank you, you are correct. We have re-written the legend of Figure 6 to focus primarily on photosynthesis.    Figure 3 -low gs will not generate a break in the ACi curve, it will just shift the Ci values down for a given Ca. So that text can be deleted from the paper (line 167 and on). Instead, the break appears to be where the measurements went from decreasing Ci below 400 ppm to raising it above that Ca level. That instead indicates that the continued exposure to warmer temperatures stimulated photosynthetic processes over the tie that the curves were made. Provided that the Vcmax data are generated from only low Ci data and the Jmax data from high CI data, this may be OK. But there should still be more data points at low Ci -are data points missing? 3. Figure 1 -I would argue that needles from a single tree are not independent replicates. I suggest rephrasing. Thank you, we have rephrased this. Figure 3 -low gs will not generate a break in the ACi curve, it will just shift the Ci values down for a given Ca. So that text can be deleted from the paper (line 167 and on). Instead, the break appears to be where the measurements went from decreasing Ci below 400 ppm to raising it above that Ca level. That instead indicates that the continued exposure to warmer temperatures stimulated photosynthetic processes over the time that the curves were made. Provided that the Vcmax data are generated from only low Ci data and the Jmax data from high CI data, this may be OK. But there should still be more data points at low Ci -are data points missing? The reviewer is correct, the low gs could cause the lower Ci values for a given Ca. We have deleted the text indicated by the reviewer.

4.
The reviewer is also correct in that the continued exposure to warmer temperatures could have stimulated photosynthetic processes especially in the field samples. The procedure for ACi curve measurement was set to decrease from 400 ppm to 50 ppm, and the warm temperature in the cuvette relative to the field temperatures may have affected photosynthetic processes by the time the 50ppm measurements were made. In addition, in these measurements the maximum carboxylation rate (V cmax ) and maximum electron transport rate (J max ) were estimated according to Sharkey et al. 51 , and as noted by these authors "Data at very low [CO2] can be limited by Rubisco deactivation and it may be useful to exclude them from the analysis.". With both these issues in mind, we removed the 50ppm data point and only included data from the 100, 150 and 200ppm measurements for our calculations of Vcmax. However, for comparison purposes in the table below we compare the calculated values of Vcmax including the 50ppm data point with those that we report that did not utilize this measurement point. These data clearly show that including the 50ppm data point does not alter the outcome from the comparisons we report nor our final conclusions. While we could include these data if the Editor wishes us to, we would prefer to not change the Vcmax calculations because the seedling data is also calculated from a dataset that does NOT include 50ppm measurements -according to the recommendations of Sharkey et al -and we feel the dataset is more coherent if all model outputs are from datasets using a similar range of CO2 concentrations.