Introduction

O2 in the Earth’s atmosphere is generated by photosynthetic organisms growing in water and on land. Boreal forests cover 14% of Earth’s land (1.9 billion hectares) and account for 33% of Earth’s total forests, thereby contributing significantly to the global carbon balance and O2 production1. Photosynthetic O2 evolution from H2O splitting is carried out by a penta-μ-oxo bridged tetra-manganese calcium cluster (Mn4CaO5) in the oxygen-evolving complex (OEC) of photosystem (PS) II2 during the light reactions. The electrons extracted from H2O is further transferred to PSI through several redox carriers and subsequently accepted by NADP+ to produce NADPH3 in the photosynthetic electron transfer chain (PETC). Later, in the so-called dark reaction, CO2 assimilation occurs involving NADPH in the Calvin-Benson-Bassham (CBB) cycle4. The redox imbalance between light and dark reactions often leads to reactive oxygen species (ROS) formation, which can damage photosystems5. Hence, several protection mechanisms6, such as non-photochemical quenching (NPQ) in PSII7, PTOX-mediated oxidation (chlororespiration) of the plastoquinone (PQ)8, cyclic electron flow (CEF)9,10, Mehler-reaction11 around PSI and photorespiration via RuBisCO in chloroplast stroma12, have evolved in plants. PTOX, a chloroplastic non-heme diiron quinol oxidase, oxidizes the PQ pool via consumption of O213, whereas in CEF, the low pH generated across the thylakoid membrane enhances NPQ14 and controls the excess electron flow towards PSI15. The Mehler-reaction consumes electrons from PSI by reducing O2 to H2O with H2O2 (hydrogen peroxide) as intermediate11, thereby protect PSI. In photorespiration, RuBisCO fixes O2 instead of CO2 and releases CO2 through an inter-mitochondrial and inter-peroxisomal shuttle12. Cyanobacteria16, algae17, and mosses18 have an additional flavodiiron (Flv) protein-mediated pathway utilizing excess electrons to reduce O2 directly to H2O at the acceptor-side of PSI. More recently, a similar mechanism was also predicted to exist in gymnosperms (including conifers)19,20,21, but never experimentally verified.

In boreal forests, most plant species overwinter without exposing their green parts to the light, however, conifers needles face extremely high oxidative stress in early spring when solar radiation is high but photosynthesis is constrained by low temperatures5. We recently demonstrated the molecular basis of the ‘sustained NPQ’ mechanism22 that protects PSII in winter/ES in Scots pine23 and Norway spruce24. Even though this quenching is extremely efficient in protecting PSII, it is hard to completely prevent light-driven ROS production under such low temperatures25. Therefore, questions remain about the protection mechanism of PSI. Previous reports have suggested that PTOX13,26 and CEF may be involved27. However, the absence of thylakoid NDH complexes (NADH dehydrogenase-like)28 in gymnosperms and the potential presence of Flv proteins19,20 makes the situation complex. Moreover, photorespiration has been shown not to be the major electron sink under low temperatures29,30. Obviously, conclusions made from studies of angiosperms may not hold true for conifers31.

In the present study, we measured light-induced O2 exchange in isolated thylakoids from summer (S) and ES in Scots pine and Norway spruce needles. Using different PETC inhibitors, such as DCMU{3-(3,4-dichlorophenyl)−1,1-dimethylurea} (blocks QB site at PSII)32 and mercuric chloride (HgCl2; profoundly affects electron transfer via plastocyanin)33, we obtained direct evidence that O2 photoreduction around PSI is much stronger than PSII-related O2 evolution, as PSII remained extremely quenched and PSI activity was higher than PSII in early spring23. In combination with P700 absorbance and immunodetection, we demonstrate that Flv-dependent O2 consumption is the major functional electron sink alleviating the over-reduction of the PQ pool by stromal metabolic reductants in early spring when plants are exposed to the combined stresses of cold temperatures and high irradiance. This mechanism could remove internal O2, prevent over-reduction of the acceptor side of PSI, and may protect both photosystems against photooxidative stress in early spring better than other suggested pathways. Our study provides functional evidence of this kind of photoreduction of O2 around PSI and its seasonal variations in vascular plants.

Results

Distinct dynamics of O2 evolution in thylakoid membranes of conifers and angiosperms

Isolated thylakoid membranes are devoid of stromal components, their illumination rapidly leads to over-reduction of the PQ pool, which lowers the rate of H2O splitting. Hence, to measure the “pure” O2-evolving activity of PSII without the influence of other components of the PETC34, we performed O2 evolution measurements in the presence of PPBQ and FeCy, which accept electrons from the QB site on the electron-acceptor-side of PSII (Supplementary Figs. 1 and 2). In intact thylakoid preparations, luminal acidification can quickly suppress O2 evolution, hence, uncouplers need to be employed to dissipate the pH gradient under illumination. This was not required here since we performed experiments with frozen thylakoids that are leaky to protons. First, we measured O2 exchange with a Clark-type oxygen electrode (‘Clark-electrode’ hereafter) (Supplementary Fig. 1a, b). Second, we performed time-resolved membrane-inlet mass spectrometry (TR-MIMS) assays with 10% of 18O-enriched water (H218O) (Supplementary Fig. 2) to discriminate between oxygen production (mainly 16O18O) and consumption (mainly as 16O2) reactions, monitored at m/z 34 and m/z 32 signals, respectively34.

Clark-electrode O2 measurements showed that in spinach thylakoids, the O2 yield was independent of the O2 level in the medium as O2 exchange was similar in both air-saturated and O2-free buffer (Fig. 1a). In contrast, O2 evolution in S pine thylakoid membranes was strongly dependent on the O2 level in the medium (Fig. 1b, c green line). O2 evolution of S pine thylakoid in air-saturated buffer was approximately half of spinach thylakoids. Moreover, the O2 yield reached a maximum after the first 30 s (at 60th second) and then decreased slowly over the next 30 s (60th to 90th second) of the illumination period (Fig. 1b green line). In O2-free buffer, O2 evolution of S thylakoids increased for the first 50 s and then plateaued for the last 10 s (Fig. 1c green line). This data indicates that, in vitro, the O2 production of the S pine thylakoids is inhibited by a photoinactivation mechanism35 that is dependent on the O2 level in the sample cuvette. Interestingly, spinach membranes in air-saturated buffer without PPBQ and FeCy supplementation did not show any O2 exchange (Supplementary Fig. 1c), whereas ES and S pine thylakoids showed only O2 consumption (Supplementary Fig. 3).

Fig. 1: O2 yields in spinach and pine thylakoid membranes.
figure 1

a O2 yields in spinach thylakoid membranes in air-saturated (+Air) and O2-free buffer (−O2); b O2 yields in summer (S) and early spring (ES) pine thylakoid membranes in air, and c in O2-free buffer measured with a Clark-type electrode. d 16,18O2 and 16O2 yields in spinach thylakoid and e 16,18O2 and f 16O2 yields in S and ES pine thylakoid membranes measured with TR-MIMS under partially degassed conditions. Gray-shaded regions indicate dark periods before and after illumination of the thylakoid membranes with 800 (Clark-electrode)/1200 (TR-MIMS) µmol of photons m−2 s−1 for 60 s (offsets between start of illumination and O2 production/consumption are intrinsic to the respective technique and do not affect the measurements). In all measurements, thylakoid membranes were supplemented with exogenous electron acceptors, 250 µM PPBQ, and 500 µM FeCy that are sufficient for sustaining electron flow through PSII for 60 s (see spinach data, panels A and D). The transient O2 exchange data for pine thylakoids are thus a consequence of O2 consumption reactions. In spinach, S and ES pine thylakoid membranes data represent O2 exchange corresponding to the same chlorophyll content (50 µg) for both Clark-type electrode and MIMS measurements. For measurements in O2-free buffer, the thylakoid suspension buffer was bubbled with a continuous flow of N2 in the Clark-electrode chamber until the O2 yield became zero prior to the addition of the thylakoid membranes (for the experimental design, see Supplementary Figs. 1 and 2;). Colored shaded regions around O2 yield curve (mean) indicate ±SEM (n = 3) in a–c, where n = biological replicates. For the TR-MIMS (with H218O enrichment of 10%), one representative trace out of 3–5 independent measurements is shown. Signals processed as described in Supplementary Fig. 2A; note that due to background subtraction 0 mV does not correspond to the absence of O2. O2 exchange rates for the Clark-electrode measurements are given in Supplementary Fig. 11). Source data are provided as a Source Data file.

Full size image

TR-MIMS data of spinach thylakoids were similar to the Clark-electrode measurements and both 16O2 and 16,18O2 exhibited similar kinetics (Fig. 1d). As the 16,18O2-traces at the beginning of illumination are much less sensitive to O2 reduction (due to the initially low 16,18O2 concentration in the measuring buffer), this suggests that in spinach thylakoids O2 reduction is of minor importance. The spinach measurements also demonstrate that sufficient artificial electron acceptors have been added to sustain O2 evolution for 60 s under our experimental conditions. In line with the Clark-electrode measurements, for pine S thylakoids (Fig. 1e, f green line) only a transient O2 production is observed for both oxygen species, even at the reduced O2 levels in the MIMS cell. In addition, the O2 consumption showed a transient behavior, possibly indicating a light-induced effect of PSI (see below).

O2 consumption dominates in early spring pine thylakoids

As the photosynthetic activity differs significantly in S and ES pine needles23 we compared the light-dependent O2 exchange between S and ES thylakoid membranes using a Clark-electrode. ES samples in air-saturated buffer showed strong O2 consumption (Fig. 1b, blue line) compared to S. In O2-free buffer, O2 evolution in ES very slowly increased until the end of the illumination period, but the overall amplitude of ES samples was ~40% of the S samples (compare blue and green traces in Fig. 1c). This suggests that although light-dependent O2 evolution was present in ES, light-dependent O2 consumption was much stronger from the beginning of illumination, resulting in net consumption of O2. ES samples in air-saturated buffer without electron acceptors supplementation showed ~2.5 times higher consumption than those from S (Supplementary Fig. 3). To understand the dynamics of simultaneous O2 evolution and consumption from a physiological perspective, we performed TR-MIMS also on ES samples, which showed that 16,18O2 evolution in ES was much lower than in S (Fig. 1f). However, unlike in S, 16O2 was strongly consumed immediately after illumination and consumption reached a maximum after 40 s of illumination in ES (Fig. 1e). PSII showed 16,18O2 evolution immediately after illumination in both ES and S but with a much smaller amplitude in ES.

Water oxidation is not the main source of electrons for photoreduction of O2 in ES thylakoids

Earlier studies have suggested that high oxidative stress imposed on conifer needles in ES could lead to photoinhibition of PSII via photodamage36,37. Mechanisms of photodamage involve a strong reduction in O2 evolution38 (Supplementary Fig. 4 and the discussion there in) and suggested to form Mn-depleted PSII reaction centers that could consume O232. Our previous study of ES pine needles suggested minor photoinhibition in PSII23, but it was far from severe under the conditions tested. In the present study, ES thylakoids evolved ~40% O2 compared to S thylakoids (Fig. 1c). Hence, PSII was not severely photodamaged. This makes it unlikely that photodamaged PSII centers could be responsible for the O2 consumption in ES and agrees with an O2 reduction site(s) downstream of PSII in the PETC.

All reduction reactions in the PETC require electrons, which are typically supplied by oxidation of water in PSII. To determine whether electrons involved in the photo-consumption of O2 were supplied from PSII, we added DCMU to the thylakoid suspension. In the presence of DCMU, the 16,18O2 yield diminished completely, suggesting that H2O oxidation in PSII was completely blocked in both S and ES (Fig. 2b). Surprisingly, both S and ES samples consumed 16O2, but the negative amplitude of 16O2 in ES was stronger than S (Fig. 2a). This showed that H2O oxidation in PSII did not supply electrons for the photoreduction of O2 neither in S nor ES. Hence, we conclude that in pine thylakoid membranes, simultaneous O2 evolution and consumption were two independent processes occurring in both S and ES but with different magnitudes. This difference resulted in contrasting net O2 yield patterns between S and ES and the O2 consumption was not PSII dependent.

Fig. 2: Electron donation for O2 photoreduction in early spring thylakoid membranes.
figure 2

a 16O2 yields b 16,18O2 (note that O2 consumption in the mixed labeled O2 was small due to the extremely low natural abundance of 16,18O2 compared to 16O2) in S and ES pine thylakoid membranes supplemented with 25 µM DCMU measured by TR-MIMS under partially degassed conditions in the presence of 250 µM PPBQ and 500 µM FeCy. Gray-shaded regions indicate dark periods before and after illumination of the thylakoid membranes with 1200 µmol of photons m−2 s−1 for 60 s. For TR-MIMS (10% H218O enrichment), one representative spectrum out of 3–5 independent measurements is shown. c Intersystem e-/P700 pool measured as a ratio between area under MT and area under ST flash, applied under 200 µmol of photons m−2 s−1 constant FR-light (720 nm) illumination (n = 3) from S and ES intact needles. d Chlorophyll fluorescent signal recorded from S (n = 2) and ES (n = 3) intact needles under constant actinic red light (320 µmol of photons m−2 s−1) illumination for 5–6 min followed by 120–180 s of dark relaxation (Time scale is set to ‘0’ upon actinic light off). c, d n = number of measurements where needles were pooled from 5 trees. e Redox state of the PQ pool presented as the ratio of plastoquinol (PQH2) to plastoquinone (PQ) (n = 9) from dark-adapted S and ES thylakoids where n defines independent measurement replication (n = 3) of individual biological replicates (n = 3). Statistically significant mean differences were calculated t test (p < 0.05) (Supplementary Table 5). f O2 yields in S and ES thylakoid membranes in air-saturated buffer measured by a Clark-electrode with 1 mg ml−1 HgCl2 supplementation in the absence of PPBQ and µM FeCy (-A). The colored shaded regions around O2 yield curve (mean) indicates ±SEM (n = 3) where n = biological replicates. Data represents O2 exchange corresponding to the same chlorophyll content (50 µg) for both Clark-type electrode and MIMS measurements. O2 exchange rates from Clarke-electrode are given in Supplementary Fig. 11. c, e The box bounds indicate ±SEM, minima/maxima indicates ±SD, middle line in the box indicates mean and the dots indicate data points. Source data are provided as a Source Data file.

Full size image

Non-photochemically reduced PQ pool supplies electrons to PSI for the photoreduction of O2

PQ, a common biological redox mediator, is present predominantly within the thylakoid lipid bilayer and plastoglobuli39. In higher plants, ~30% of the total PQ is photochemically active40 and the rest is modulated via non-photochemical processes. O2 consumption in the presence of DCMU indicated that the electrons for photoreduction of O2 came from an already stored e- pool in the thylakoid, which could originate from non-photochemical sources in intact needles in the absence of LEF during early spring23. Hence, we first measured the intersystem e- pool size on intact needles, which suggested that the e−/P700 was 3 times higher in ES compared to S (Fig. 2c). Moreover, post illumination Fo’ rise measurements suggested that non-photochemical dark reduction of the PQ pool indeed occurred in ES intact needles (Fig. 2d). This was also confirmed by comparing the changes in the P700 re-reduction upon switching off FR-light illumination, in intact needles with thylakoid samples (Supplementary Figs. 5a–d). To decipher if our isolated thylakoid samples also had an increased non-photochemical electron pool, we analyzed the prenylquinones by LC/QTOF mass spectrometry. In dark-adapted thylakoids, the PQ pool in ES thylakoids was predominantly reduced: the PQ/PQH2 ratio in ES was ~40% lower than in S (Fig. 2e, Supplementary Table 5). In addition, the ubiquinone pool (UQ) was heavily reduced (Supplementary Fig. 5g). Interestingly, the stromal e- pool was previously reported to be 5 times higher in ES pine needles than S41. Nevertheless, this suggests that reduced PQ in ES thylakoids upon illumination supplied electrons either to PTOX for photoreduction of O2 to H2O8 or to PSI through PETC. To distinguish between these two pathways, we measured O2 exchange in air-saturated buffer with/without the addition of mercuric chloride (HgCl2). HgCl2 affects electron transfer via plastocyanin (PC) and hinders electron flow from cytochrome b6f (Cyt b6f) to PSI33. PPBQ and FeCy were not added during these measurements as they can take electrons upstream of Cyt b6f, and hence influence the effect of HgCl2. We found that O2 consumption in both S and ES was diminished (Fig. 2f) compared to without HgCl2 supplementation (Supplementary Fig. 3). Taken together, these results show that O2 consumption in the ES samples was a photoreduction process occurring around PSI with electrons supplied from the non-photochemically reduced PQ pool and PTOX was not the major site of O2 photoreduction.

O2 consumption in ES thylakoids is accompanied by higher P700 oxidation and a higher abundance of flavodiiron proteins

O2 consumption around PSI is known to have profound effects on the redox state of PSI42. In line with previous reports23,27, in our study, P700 became more oxidized (Y[ND] increased) and acceptor-side limitation (Y[NA]) decreased in intact ES needles upon increasing irradiance, compared to S samples (Fig. 3a). Interestingly, when we compared P700 oxidation (P700+) from the fast kinetics of the saturating-pulse (SP) between spinach leaves and S pine needles (Supplementary Fig. 6), we found that in S pine needles, P700 did not show a biphasic re-reduction but instead a transient dip. We interpreted this as follows: P700 first reached maximum oxidation upon the SP trigger, then the oxidation state decreased as electrons were supplied from reduced PQ, but after 100 ms P700 re-oxidized back to maximum as the electrons were consumed from the acceptor side. In ES pine, we detected only a very small transient dip in P700 after 100 ms (Supplementary Fig. 6B) even though e−/P700 was higher (Fig. 2c) and PQ/PQH2 was lower (Fig. 2e). This suggests that electrons were taken much faster from the acceptor-side of PSI in ES needles, thereby oxidizing P700. While higher P700 oxidation could in principle be due to CEF27, recent reports showed that a reduced intersystem PQ pool (Fig. 2c, e) hinders CEF43. Therefore, to completely block CEF we used Antimycin A (known CEF blocker)44. One earlier report suggested that PGR5 in Pinus tedea is resistant to Antimycin A45 (at 10 µM) when it was expressed in Arabidopsis. However, a very recent report clearly demonstrated that Antimycin A effectively blocks the PGR5-CEF pathway in intact needles of Pinus sylvestris (at 200 µM)27. Hence, we used 3 times higher Antimycin A concentration than reported previously45 and at this concentration, Antimycin A was effective in modulating P700 re-reduction kinetics (Supplementary Figs. 5c–f). Nevertheless, we found a similar overall trend in SP measurements, on both thylakoids with Antimycin A and intact needles without Antimycin A, i.e., with increasing irradiance, Y(ND) was higher, and Y(NA) was lower in ES than S (compare Fig. 3a and 3b).

Fig. 3: Photoreduction of O2 around PSI.
figure 3

a Changes in PSI donor-side limitation [Y(ND)] and acceptor-side limitation [Y(NA)] measured by the SP method with increasing PAR (µmol of photons m−2 s−1) in S and ES pine needles (n = 3) (without antimycin A), and b in S (n = 4) and ES (n = 5) isolated thylakoid membranes (100 µg/ml) (with 30 µM antimycin A). Data in a, b indicate the mean ± SEM and n = number of measurements where needles/thylakoids were pooled from 3 to 5 biological replicates. c Schematic of SP-induced maximum P700 oxidation (Pm) followed by steady-state P700 oxidation (P) achieved by intermittent FR illumination. The effective P700 oxidation state was considered as the difference between the Pm and P signal at the 6th cycle of FR illumination. d Changes in P700 absorbance in S and ES pine thylakoid membranes (corresponding to 100 µg/ml chlorophyll) measured with six cycles of intermittent FR illumination in the presence of 30 µM antimycin A (n = 3). e O2 yields in S and ES thylakoid membranes in air-saturated buffer measured by a Clark- electrode with 1000 u ml−1 catalase supplementation in the absence of 250 µM PPBQ and 500 µM FeCy (-A). The colored shaded regions around the O2 yield curve (mean) indicate ±SEM (n = 3) in the Clark-electrode measurements. S and ES pine thylakoid membranes data represent O2 exchange corresponds to the same chlorophyll content (50 µg). d, e n = biological replicates. f Relative abundance of flavodiiron A protein in S and ES pine thylakoid membranes. Two summer samples (S1, S2) and three early spring samples (ES1, ES2, ES3) (Supplementary Table 1) corresponding to 4 µg of chlorophyll (100%) were loaded in separate lanes and 1 µg and 2 µg of chlorophyll (25% and 50%) of S1 were loaded in the first two lanes as quality controls, and the membrane was immunoblotted against anti-FlvA antibody. (For the relative quantitation of Flv proteins in S and ES, see Supplementary Table 4). A Coomassie-stained membrane is shown in the bottom panel. Similar immunoblotting results were obtained in four independent experiments. Source data are provided as a Source Data file.

Full size image

Recently, the SP method was shown not to be fully reliable measure for a possible acceptor-side limitation of PSI as the steady-state concentration of P700+ may be influenced by electrons contributed from the PSII via the PQ pool46,47. Therefore, to determine steady-state P700+, we exposed S and ES thylakoid membranes to six intermittent cycles of far-red (FR) light48 in the presence of Antimycin A (Fig. 3c). This method does not reflect the maximum oxidized P700 in terms of Y(ND) like the SP method but only indicates steady-state P700+ changes with increasing cycles of illumination as a consequence of electron flow within PSI (Fig. 3c). The PSI population in S and ES may be different due to different abundance of PSI23,24. Therefore, we normalized the P700 signal to the maximal P700 oxidation level, achieved by the initial SP (Pm) prior to the start of the FR cycles. We found that at the 6th cycle of FR illumination in S samples, the steady-state P700 signal (P) decreased by ~50% compared to Pm, whereas in ES, it only decreased by ~15% (Fig. 3d). Moreover, the relaxation of the P700 signal during the dark periods (after FR illumination) decreased faster and further in ES than in S (also seen in Supplementary Fig. 5). The dark relaxation of P700 is mainly caused by electron donation from reduced PQ. Thus, the results suggested that the non-photochemically reduced PQ in ES donated more electrons to P700 than in S. Overall, these findings indicated that in ES, a strong electron sink apart from CEF at the acceptor-side of PSI consumes the excess electrons donated to P700 from the reduced PQ.

The Mehler-reaction scavenges electrons from the acceptor-side of PSI consuming O2 and producing H2O2, subsequently metabolized through the action of (per)oxidases. In isolated thylakoid membranes, stromal components are lacking, and the Mehler-reaction requires external supply of enzymes that could convert H2O2 to O2 and H2O. To quantify Mehler-reaction-driven O2 consumption in ES thylakoids we performed O2 measurements by the Clark-electrode in air-saturated buffer (without PSII acceptors, PPBQ, and FeCy) supplemented with 1000-unit ml−1 catalase. O2 consumption did not change in S thylakoids, but in ES consumption decreased ~34% (Fig. 3e) compared to no catalase addition (Supplementary Fig. 4). This suggested that the Mehler-reaction could explain only a fraction (~1/3rd) of the O2 consumption in ES thylakoid membranes, and that another acceptor at the PSI site accounted for the bigger part (~2/3rd).

As the evidence showed that none of the other possible mechanisms (damaged PSII, PTOX or photorespiration, Mehler-reaction) could explain the major fraction of the O2 consumption in ES thylakoids, we considered the involvement of Flv proteins—as the genes coding for Flv proteins are present in conifer genomes19. Spruce FlvA and FlvB are most similar to type 3 Flv proteins in Physcomitrella sp. and Chlamydomonas sp. (Supplementary Figs. 8b, c), which are known to be associated with photoreduction of O2 at PSI19,49. Photoreduction of O2 at the acceptor-side of PSI has recently been demonstrated and suggested to be related to Flv proteins that can scavenge excess electron from PSI and keep P700 in an oxidized state (P700+) in conifers19. Therefore, we quantified the abundance of Flv proteins in S and ES pine thylakoids and found that FlvA accumulation was ~3 times higher in ES compared to S (Fig. 3e, Supplementary Table 4) when samples were loaded based on equal chlorophyll amounts.

O2 photoreduction by Flavodiiron protein is also triggered under artificially simulated semi-early spring conditions in the climate chambers

Thylakoid membranes of pine saplings from climate-chamber acclimated to subzero (−8 °C) temperatures demonstrated stronger O2 consumption compared to control plants (Supplementary Fig. 7a). In presence of HgCl2 most of the O2 consumption disappeared (Supplementary Fig. 7b) but addition of catalase did not change the O2 uptake in any of the samples (Supplementary Fig. 7c), unlike in samples collected from the field in ES (Fig. 3e). P700 oxidation (Supplementary Fig. 7d) and accumulation of FlvA (Supplementary Fig. 7f) were also higher in thylakoid samples from subzero acclimated needles compared to control. This suggests that O2 consumption phenomena in climate-chamber acclimated samples was similar, if not identical, to the natural environment during ES. This indicated that even though in climate chamber the conditions were less harsh than in the field, similar mechanism might have been evoked.

Flv proteins may mediate consumption of O2 in early spring also in Norway spruce

To elucidate whether photoreduction of O2 at PSI is a general response of conifers to winter conditions, we further explored the possibility of higher O2 consumption and correlation of Flv protein accumulation in Norway spruce thylakoids during ES (Supplementary Figs. 8–10). Norway spruce is a more challenging study system than Scots pine regarding both its biochemistry and physiology. In our study, biochemical preparations of Norway spruce varied more and the “winter states” were less stable during and after preparations than those of Scots pine. However, our data showed that FlvA protein again accumulated in Norway spruce thylakoid membranes, and amounts were ~60% higher in ES compared to S (Supplementary Fig. 8a). Because the accumulation was lower than in pine, the magnitude of changes in other parameters was expected to be lower. This was indeed the case when ES thylakoid membranes were compared to those in S (Supplementary Fig. 9a): the steady-state P700+ signal decreased by ~30% after the 6th cycle of FR illumination compared to the first cycle. The pattern of fast SP kinetics of P700 oxidation in ES was similar in pine and spruce intact needles (Supplementary Fig. 9b). Clark-electrode measurements with spruce samples did not work in our hands as the thylakoids stuck to the membrane hindering gas exchange. However, in our TR-MIMS assays in the presence of DCMU, ES spruce samples showed stronger 16O2 consumption but no change in 16,18O2 (Supplementary Fig. 10) compared to S samples. These results suggest that Norway spruce behaved similarly to Scots pine and that Flv proteins were most likely involved in P700 oxidation. However, compared to pine, the magnitude was lower in spruce. This suggests that the Flv-mediated O2 consumption under stressful conditions like in ES is a common phenomenon among certain groups of conifers.

Discussion

Conifers constitute a large fraction of terrestrial biomass, but in comparison with angiosperms, algae, and cyanobacteria, they are difficult to study mainly because genetic tools are not sufficiently developed. Photosynthesis studies in conifers are also challenging as their photosynthetic tissue—the needles—are significantly different in morphology and chemical composition than that in other plants. However, we recently demonstrated that conifer needles in the winter elicit a protection mechanism23 that involves direct energy transfer from PSII to PSI, a mechanism whose existence in angiosperms is still a matter of discussion. This protection mechanism may also be associated with phosphorylation of PsbS and triple-phosphorylation of Lhcb1, which trigger thylakoid destacking24. Here, we present evidence that an additional protection mechanism that thus far was described in detail only for lower green organisms, namely Flv protein-dependent O2 photoreduction, also contributes significantly to the winter survival of conifer needles by providing protection for PSI. Our study thus provides important experimental support for earlier predictions regarding the involvement of Flv in the photoprotection of gymnosperms19,20,21.

Since reverse genetic tools are not available for conifers, we used different PETC inhibitors and measured O2 exchange in thylakoids isolated from S and ES needles. Furthermore, we correlated O2 exchange with protein abundance to understand how conifer needles under some conditions exhibit substantial light-dependent net O2 consumption instead of O2 production as during normal photosynthesis. Four mechanisms—photorespiration, PTOX-mediated chlororespiration, the Mehler-reaction, and Flv pathway—could explain this phenomenon. By using PETC inhibitors, we employed an ‘elimination approach’ to distinguish between these four possible pathways. Among them, photorespiration in general remains lower in gymnosperms compared to angiosperms, specifically at low temperatures30; Moreover, it was easily excluded since we observed O2 consumption in isolated thylakoid membranes. Secondly, Lodgepole Pine in low temperatures has been shown to possess an O2-dependent excess energy dissipation capability that was assigned to PTOX29. However, when we added HgCl2 to block electron transfer through plastocyanin, O2 consumption dropped significantly suggesting that the consumption occurred around PSI and not directly from the heavily reduced PQ pool via PTOX (Fig. 2f). Finally, the Mehler-reaction can only take a minor fraction of the total electrons in the PETC when CO2 assimilation is restricted50 (such as during winter in conifers27). Moreover, the Mehler-reaction was previously shown to not be involved in the winter sustainability of conifers29 and in line with this, experiments with catalase indicated that (Fig. 3e) the contribution of the Mehler-reaction to net O2 consumption in our ES thylakoid membranes was low.

Therefore, the only remaining mechanism that could explain our data is Flv-mediated O2 photoreduction. We acknowledge the fact that we reach this conclusion by eliminating of other possible mechanisms. A critical proof—demonstration of decreased O2 consumption and photoprotection in mutant lacking Flv proteins—is still not possible to obtain from a conifer; transformation/regeneration protocols and genome editing tools have still not been developed enough. But the fact that genes coding for Flv proteins are among the very few photosynthesis-related genes that are up-, not down-regulated, in Norway spruce in the winter51,52 which is also reflected in the higher abundance of the protein levels (Fig. 3e, Supplementary Fig. a),—is direct evidence supporting our assumption.

Oxygen reduction via a Mehler-like Flv pathway is known to oxidize P700 by accepting electrons from PSI in cyanobacteria and algae16,18,48,53,54, and the presence of genes coding for Flv proteins in conifer genomes has been noted by others19. The exact site of Flv interaction with PSI remains unclear, but our data suggested that in both examined conifers, FlvA accumulated highly in ES thylakoid membranes (Fig. 3e, Supplementary Fig. 8a). Flv proteins are believed to be soluble proteins, but in pine and spruce, they were retained, at least partially, in conifer thylakoid preparations. Spruce FlvA and FlvB are most similar to type 3 Flv proteins in Physcomitrella sp. and Chlamydomonas sp. (Supplementary Fig. 8b), which are known to associate with Mehler-like reactions for photoreduction of O2 by PSI49. Therefore, our data suggest that Flvs in conifers act as an electron sink for PSI, readily consuming electrons from the acceptor-side during light stress and rapidly reducing O2 to H2O (Fig. 1c, e, f). In this way, the acceptor-side remains in a sufficiently oxidized state to accept electrons from reduced P700, and P700 upon illumination can readily oxidize by donating electrons to FX (Fig. 3d, Supplementary Fig. 9a). In addition, O2 consumption would also reduce the risk of ROS production under conditions where the capacity for ROS detoxification mechanisms is low by creating a lower oxygenic environment around the photosystems. Thus, Flv proteins may have a dual protective function. In the winter, and under other severe stress conditions, needle gas exchange is very low due to the thick cuticle and closed stomata. Hence, any O2 produced by H2O oxidation would accumulate, leading to an increased risk of photooxidative stress. Whether this O2 consumption could give anaerobic conditions in vivo is hard to estimate, but according to our Clark-electrode experiments, 10–15 min of illumination was enough to consume all oxygen in the chamber.

It has been shown that during winter P700 in conifers remain in a donor-side limited condition—meaning P700 is mostly in its oxidized state (P700+)23,24,27. However, we found that in ES thylakoids, PQ remains in a highly reduced state (Fig. 2e) concomitant with three times higher intersystem e- pool (Fig. 2c) and 5 times higher stromal e- pool in the intact pine needles41,55. In absence or extremely low LEF, this non-photochemical reduction of the PQ (Fig. 2d, Supplementary Fig. 5a–d) might be linked to higher NAD(P)H reduction via chloroplastic NDH II activity as predicted previously41,51. How can P700 remain oxidized when the PQ is predominantly reduced? An influx of electrons from the luminal side is passed to the primary stable electron-acceptor (FX). However, in the presence of Flv activity, electrons could be passed to oxygen, preventing FA, FB from becoming over-reduced, which would otherwise promote ROS formation33,56, potentially causing damage57. Hence, Flvs could contribute to photoprotection in conifers, in particular in ES, by keeping P700 in an oxidized state to avoid irreversible PSI damage (Fig. 3a–c).

CEF can also lower reduction pressure on PSI by accepting electrons from the PETC45. Most conifers lack all plastidial ndh genes58 known to be involved in CEF. We found that P700 remained highly oxidized in ES compared to S (Fig. 3a, c), even when the Prg5/Prgl-CEF pathway was blocked by reduced PQ43 (Fig. 3d, Supplementary Fig. 9a). Therefore, CEF is unlikely to be a dominant pathway for scavenging electrons and oxidation of P700 in ES thylakoids. However, as PGR5/PGRL1 was reported to be more abundant in ES than S27, along with higher abundance of ATP in winter acclimated pine needles55, hence, we speculate that PGR5-CEF may contribute to ATP production via proton gradient formation59. In ES when intersystem e- pool is much higher29,41, over reduction of the PQ could prevent CEF43 and thereby hinder pmf generation59 which could lower ATP production. However, electron scavenging by Flvs from the acceptor-side of PSI through O2 photoreduction upon illumination could shift the PQ/PQH2 balance and restore CEF. This would maintain the pmf and continue ATP generation even if LEF was restricted23,55.

A schematic of the possible electron flow in ES and S conifer thylakoid membranes is shown in Fig. 4. Upon illumination of S thylakoid membranes, electrons generated from H2O splitting to O2 reduce the PQ and further pass through Cyt b6f to PC and then to P700. Upon receiving electrons, P700 becomes reduced and then under illumination becomes re-oxidized by donating electrons to the electron acceptors (FX to FA to FB), which are then taken up by Fd for the forward reactions. In ES (A), although the H2O splitting reaction slows down, the PQ pool remains in a highly reduced state (B) (most likely with contribution from ndh-2 mediated stromal reduction), which contentiously donates electrons to P700. Upon illumination (C), P700 readily becomes oxidized by donating electrons to acceptors (FX, FA and FB) (D). As a result, the acceptors (FX, FA, and FB) become highly reduced as electron demand from the forward reaction is limited due to the down-regulation of the CBB (E). However, FX, FA and FB remain in an oxidized state as the electrons are taken up by Flv proteins (F). Higher Flv activity results in stronger O2 consumption, resulting in net O2 consumption and donor-side limited PSI.

Fig. 4: Comparative schematic of light-induced electron flow between summer (green arrows) and early spring (insets with blue arrows) thylakoid membranes.
figure 4

Insets show possible alterations in the electron flow in early spring compared to summer as directly measured by our experimental setups. Each step of altered electron flow in early spring is indicated sequentially from A to F. PDB structures used in the model are PSII (Pisum sp) 5XNL79; Cyt b6f (Mastigocladus laminosus) 4H0L80; plastocyanin-PSI-ferredoxin (Pisum sp) 6YEZ81. Note, that the PSII cofactors in contrast to PSI cofactors are not shown in the model as their redox state was not measured experimentally. Electron flow from Cyt b6f to PSI is mediated through plastocyanin that upon binding to Cyt b6f accepts electrons and diffuses to PSI followed by interaction with PSI and electron donation to P700. However, due to the non-availability of Cyt b6f bound plastocyanin PDB structure plastocyanin is only shown in association with PSI. The position of Flv on PSI does not represent any empirical binding site nor define the exact site of electron transfer from PSI acceptor side to Flv. Bold text in the schematic indicates direct experimental evidence of the phenomena.

Full size image

The Flv pathway in cyanobacteria and algae was previously considered ‘futile’ and consumes a large fraction of electrons from PETC to reduce O2, which may lower the photosynthetic yield19. However, it was recently shown that Flvs do not compromise CO2 assimilation60. Perhaps Flvs are useful for conifers as a mechanism that can be rapidly invoked when conditions get worse, and dismantled, when they get better. In general, having many parallel mechanisms to protect the precious photosynthetic apparatus would be advantageous, which could be the case for conifers. Although we suggest that Flv-mediated oxygen consumption is substantial in conifer needles in ES, other mechanisms for O2-reduction can be involved in parallel, namely the Mehler-reaction (Fig. 3e). Conifers and angiosperms have evolved from a common ancestor that grew in a suitable but light limited environment61. So, why have conifers retained but angiosperms lost the Flv proteins? Under low light conditions, carelessly dissipating the reducing power by routing electrons to O2 would be an evolutionary disadvantage25. However, conifers have instead lost the type I NDH-mediated CEF and only retained Pgr5/Pgrl158. It is possible that these different evolutionary trajectories were adaptive as conifers in general adapted to harsh environments but typically competed less well with angiosperms in richer ecosystems. Perhaps Flv proteins provide better protection, but NDH-mediated CEF gives better energy economy under a more favorable environment. Hence, Flvs could be a part of a ‘better safe than sorry’ evolutionary strategy in conifers.

Methods

Plant material harvesting

Fully developed needles (10-15 gm) were harvested 15 times from the south facing branches of 3-5 mature (40+ years old, 8–10-meter tall) trees of Pinus sylvestris (Scots pine) and Picea abies (Norway spruce) (during 2017-2020 (Sampling dates are provided in Supplementary Table 1). Needles were immediately transferred to the lab (≤5 min) and either subjected to intact needle measurements or thylakoid isolation as described in62 with slight modifications. After isolation thylakoid membranes were suspended in storage buffer (50 mM HEPES-KOH (pH 7.5), 5 mM MgCl2, 100 mM sorbitol) (previously named as buffer B3 in ref. 62) and flash frozen in 40–50 µl aliquots by liquid N2 and stored in −80 °C until further use. Before use chlorophyll concentration of each aliquot was determined following63.

Forclimate-chamber experiments, small 2–3 years old pine saplings (~1 meter tall) were acclimated to control condition with 120 µmol of photons m−2 s−1 at 18 °C for 10–14 days, and needles were harvested from 12 saplings and subjected thylakoid isolation. Then 15 saplings were moved subzero temperature (−8 °C) and acclimated for 10–12 days and needles were harvested and subjected to thylakoid isolation. In all climate chamber experiments n = number of biological replicate groups, where one group comprised of thylakoids pulled from 4 plants in control and 5 plants in −8 °C conditions, hence, total 12 plants pulled into 3 groups (i.e., n = 3) in control 273 and 15 plants pulled into 3 groups (i.e., n = 3) in −8 °C conditions.

Clark-type electrode O2 measurements

Light-induced O2 evolution/consumption was measured in a Clark-type electrode (Hansatech instruments, England) with 800 µmol of photons m−2 s−1 white LED light for 60 s (see Supplementary Fig. 1 for more details). First background signal was determined by only adding 500 µl of B3 buffer in the Clark-type electrode chamber. Later, for all O2 measurements by the Clark-type electrode, thylakoid membranes were resuspended (thawed in dark on ice in case of frozen samples for 30 min) in storage buffer supplemented with/without 250 µM PPBQ (2-phenyl-p-benzoquinone) and 500 µM FeCy (potassium ferricyanide) as exogenous electron acceptors for PSII64 to a final volume of 500 µl and measurements were performed immediately (as described in Supplementary Fig. 1A). The temperature was maintained at 10–12 °C during all the measurements. For inhibition of PSII activity, DCMU (3-(3,4-dichlorophenyl)−1,1-dimethylurea) (dissolved in DMSO) was added at a final concentration of 25 µM. For blocking electron transfer through plastocyanin, mercuric chloride (HgCl2) was added at a final concentration of 1 mg ml−1. For catalase activity catalase (Sigma) was added at a final concentration of 1000-unit ml−1 from a stock concentration of 5000-unit ml−1 in water. In case of measurements in O2-free buffer (N2 was bubbled in the Clark-type electrode chamber), first 500 µl of storage buffer supplemented with 250 µM PPBQ and 500 µM FeCy was added in the Clark-type electrode chamber and the chamber was closed. N2 was supplied in the chamber through a needle from a pressure-controlled gas tap. Once the O2 signal reached zero (no O2 left in the chamber), thylakoid samples were added which caused an overshoot in the O2 signal. N2 bubbling was continued for 8–10 s further to bring the O2 signal back to zero. Illumination started once O2 yield reached Zero (Supplementary Fig. 1B). The data shown in all Clark-type electrode measurements represent total O2 exchange from thylakoid membranes corresponding to 50 µg of chlorophyll content.

Time-resolved membrane-inlet mass spectrometry (TR-MIMS) measurements

Gas exchange was measured in thylakoid suspension (in Buffer B3) by the TR-MIMS setup as previously described65. The MIMS setup contained an isotope ratio mass spectrometer (Delta V Plus; Thermo Fischer Scientific) connected to an in-house built gas-tight membrane-inlet chamber (200-μL) and a cooling trap. Prior to the measurements, H218O (97%; Larodan Fine Chemicals AB) was added (to a final enrichment of 10%). Analysis of O2 reactions was performed on the m/z 32 (16O2) and m/z 34 (16,18O2) signals, with Faraday cup amplification of 3 × 108 and 1 × 1011 (in the present study, the signal m/z 36 (18O2) was not analyzed due to its’ low amplitude at the employed H218O enrichment and its overlap with the 36Ar (m/z 36) signal). Continuous saturating Illumination (1200 µmol of photons m−2 s−1) was provided by external white light source (S2 High-Intensity White Light Source; Hansatech Instruments Ltd.). The measurements were performed at 20 °C under continuous stirring of the sample suspensions diluted with the storage buffer containing electron acceptors, 250 µM PPBQ and 500 µM FeCy. Recording of the MIMS signals was started immediately after the samples were loaded in the MIMS cell and covered with the plunger. Light was switched after 120-s incubation of the samples in the cell. The final O2 exchange curve was obtained after correction of the gas consumptions by the mass spectrophotometer (Supplementary Fig. 2). The correction was performed by fitting the traces recorded in the dark period before the illumination period using Origin Pro 2021 for each individual mass signal (see Supplementary Fig. 2 for further details of the data analysis). A lag time (of a few seconds) observed between the onset of sample illumination and monitored O2 evolution/consumption represents an offset caused by the instrumental setup (described in ref. 31). Such offset did not affect the monitored differences between ES and S samples.

P700 and fluorescence measurements

P700 measurements were performed either on intact needles or on thylakoid membranes with saturating-pulse (SP) or on thylakoid membranes with intermittent cycles FR illumination by Dual PAM100 (Walz). First, needles or thylakoid membranes were dark-adapted from 15–20 minutes. For SP measurements on intact needles P700 absorbance was recorded on a FR-light background followed by application of a 600 ms SP of 4000 µmol of photons m−2 s−142. For SP measurements on thylakoid membranes, P700 absorbance was recorded on a FR-light background followed by application of a 50 ms SP of 1000 µmol of photons m−2 s−1. For measurements on needles PAR (photosynthetically active radiation) was increased step by step up to 1600 µmol of photons m−2 s−1 whereas for measurements on thylakoid PAR was increased step by step up to 800 µmol of photons m−2 s−1.

For intermittent cycles of FR-light measurement on thylakoids, first, maximum P700 oxidation was determined by a 50 ms SP of 1000 µmol of photons m−2 s−1, followed by 30 s of dark interval, and then P700 signal was monitored for six cycles of 10 s FR illumination (intensity was 250 µmol of photons m−2 s−1) followed by 30 s of dark interval between each FR illumination48. In this measurement thylakoid samples were dark incubated in the presence of 30 µM Antimycin A for blocking Prg5/Prgl1 mediated cyclic electron flow.

The transient reduction of FR-light-induced (200 µmol of photons m−2 s−1, 720 nm, 5 min) steady-state P700+ signal caused by single turnover (ST) and multiple (MT) saturating flashes of white light was used for estimation of the apparent intersystem electron (e−/P700) pool size in vivo in both S and ES pine needles66,67. MT saturating flashes (50 ms) and ST saturating flashes (half peak width 14 ms) were applied by the power/control units of Dual PAM100 (Heinz Walz GmbH, Effeltrich, Germany). The complementary areas between the transient reduction/re-oxidation of P700 after ST and MT applications and the steady-state level of P700+ under FR illumination were used for estimation of the functional pool size of intersystem electrons on a P700 reaction center basis as described in ref. 67. For calculating the area under the curve, the curves were integrated with minimum and maximum P700 signals as the limits by using Origin Pro 2021.

For P700 re-reduction measurement, first P700 oxidation was induced via FR-light for 3 minutes, then FR-light was switched off and the re-reduction rate was calculated in terms of time constant Tau (s−1) by using a mono-exponential decay fit as done previously in refs. 68,69. Under FR illumination PSI was preferentially excited, and thereby drew electrons from the PQ pool that were present in the intersystem e- pool prior to the FR illumination. In addition, FR-light activates PSI-dependent CEF. With time, under FR-light illumination, PSI kept draining electrons from the intersystem e- pool and reached a steady-state (after 25–45 s). When this measurement was performed on intact needles, where all stromal components are intact, the re-reduction kinetics of P700 after the FR-light is switched off is used as a reliable measure of both PSI-driven CEF and/or electron flow(s) reducing the PQ pool non-photochemically in darkness via different stromal reductant pathways. In the presence of Antimycin A which effectively blocks PGR5-dependent CEF in pine27, P700 re-reduction would reflect only the electron flow(s) through the combination of different non-photochemical pathways.

In vivo Pulse/Amplitude chlorophyll fluorescence measurements of both S and ES pine needles were performed as described earlier23. The reduction state of plastoquinone (PQ) was assessed following the post-illumination increase of chlorophyll fluorescence at the Fo’ level70,71,72. Chlorophyll fluorescence of both S and ES pine needles was measured in vivo using Dual PAM100 chlorophyll fluorescence measuring system (Heinz Walz GmbH, Effeltrich, Germany). After 20 minutes of dark adaptation, the intact needles were subjected to a constant actinic red light (AL) of 320 µmol of photons m−2 s−1 illumination for 5–6 minutes until a steady-state level of Fs was reached as reported previously in intact pine needles55. Then the actinic light was switched off and the transients from Fs to Fo’ were recorded for 60 s. Post illumination rise of Fo’ is not influenced due to induction of artificial quenching by AL light intensity23. The fluorescence signals of both S and ES needles were normalized between Fs and Fo’ for direct comparison of post-illumination rise of Fo’ and the time scale was set to ‘0’ when the AL was switched off.

SDS-PAGE and immunoblotting

After chlorophyll concentration was determination by ref. 63, thylakoid membranes containing equal amounts of chlorophyll were separated on 4–20% TGX PAGE gels (Bio-Rad Criterion TGX Any kD™ precasted gels (Bio-Rad, #5678124)23. Thylakoid membranes samples (n = 3–5) were solubilized in Laemmli sample buffer73 supplemented with 100 µM DDT. After SDS-PAGE, proteins were electrophoretically transferred on a nitrocellulose membrane (Merck) by wet transfer at 15 V for 6 h at 4 °C and blocked with 2% skimmed milk at room temperature for 2 hours. After blocking, membranes were incubated with specific primary antibody against FlvA (1:2000 dilution, provided by Prof. E.-M. Aro) overnight at 4 °C and then incubated with secondary anti-rabbit antibody (AgriseraAB, Vännäs, Sweden—AS101461 at 1:15000 dilution) for 2 hours at room temperature. The antibody complexes were detected using ECL reagent (#AS16 ECL-S-N, AgriseraAB). Images were captured in Azure imaging system.

Prenylquinones in thylakoid membranes of pine

Three aliquots of the isolated thylakoid samples (~50 µg of thylakoid in 10 µl) were kept in dark (ambient state, n = 9). Thylakoids were pelleted down by centrifugation (2700 × g), sample buffer was removed and three glass beads (3 mm diameter) and 100 µl of ice-cold 30:70 (chloroform:methanol) containing labeled internal standards (Tocopherol) were immediately added to the samples. Samples were extracted for 5 min in a multi-vortex and centrifuged at 4 °C at 18,000 × g for 5 min. Supernatant was transferred into an insert in a vial and samples were analyzed by liquid chromatography-mass spectrometry (LC-MS) with a method adopted from ref. 39. Samples were directly injected in the extraction solvent without drying the aliquots to prevent changes in the redox state of quinones. In addition, quality control (QC) samples were prepared by combining an aliquot of each sample group and treatment. Based on the QC samples that were run between samples sets, the state of quinone pool did not change during the whole analysis period. Furthermore, an aliquot of QC sample was mixed with high concentration of ascorbic acid (in methanol, 50 mM) to artificially reduce the metabolite pools and to identify redox active quinones in the thylakoid samples.

The LC-MS system consisted of Agilent 1290 Infinity LC and 6546 LC/QTOF mass spectrometer equipped with atmospheric pressure chemical ionization (APCI) source and DAD (diode array detector) (Agilent Technologies). Metabolites (1 µl sample) were separated with Acquity UPLC BEH C18 column (Waters, 100 × 2.1 mm, 1.7 µm particle size) combined with an Acquity UPLC BEH C18 VanGuard pre-column (Waters, 5 × 2.1 mm, 1.7 µm particle size. Column temperature was set to 60 °C, autosampler temperature to 10 °C, and flow rate to 0.5 ml/min. Mobile phase (A = water, B = Methanol) composition at the start was 80% B increasing to 100% B in 1 min, maintained at 100% B for 6.5 min followed by a re-equilibration at 80% B for 1.0 min. The mass spectrometer parameters were optimized with α-tocopherol to increase the sensitivity and to detect mainly molecular ions. The samples were run with positive and negative mode. Data were acquired with a scan time of 4 scans/s and with a range of m/z 120–950. The corona current was 4 µA in positive and 16 µA in negative mode and nebulizer was 45 psi in positive and 40 psi in negative mode. Drying gas flow was 6 l/min, gas temperature 300 °C, vaporizer temperature 350 °C and capillary voltage 3000 V. Fragmentor was set to 120 V, skimmer to 65 V, and octopole RF peak to 750 V. In addition, MS/MS spectra were acquired with positive and negative APCI mode for QC samples over a range of collision energy 10, 20, and 40 V. Prenylquinones (11 compounds) were identified based on the standard compounds (α-tocopherol), accurate mass measurements and MS/MS fragmentation patterns (See Supplementary. Table 3). Data were processed with Mass Hunter Profinder (version B.08.00, Agilent Technologies). Prenyllipid levels were expressed relative to the total metabolite pool and the redox state based on the ratio between the oxidized (quinone) and reduced (quinol) forms.

Statistical analyses

The effects of season variation on quinone redox status (ratio of oxidized and reduced forms) were tested with t test (IBM SPSS statistics version 27) (See Supplementary Table 5). Statistically significant differences of FlvA accumulation between samples collected on different dates of summer and early spring were calculated by one-way ANOVA (SPSS) and the effect of season by t-test (SPSS) (See Supplementary Table 4). Pairwise comparisons were performed with Fisher’s Least Significant Difference test (LSD) (See Supplementary dataset 2). In all cases, P value < 0.05 was considered significant.

Bioinformatic analyses of FlvA and FlvB in spruce

FlvA and FlvB protein sequences were generated from the FlvA and FlvB gene model obtained from the Congenie database (https://congenie.org/) by using Expasy translate tool (https://web.expasy.org/translate/)74 and was aligned with other flavoproteins in other species (See Supplementary dataset 1) by using Clustal Omega multiple sequence alignment tool (https://www.ebi.ac.uk/Tools/msa/clustalo/)75. The unrooted phylogenetic tree was constructed by using IQ-TREE 276 from the aligned sequences. Spruce FlvA and FlvB domains were predicted by using NCBI CDD search tool (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi)77 and Expasy PROSITE tool (https://prosite.expasy.org/)78.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.