Regulation of chloroplast NADH dehydrogenase-like complex by NADPH-dependent thioredoxin system

ABSTRACT

Linear electron transport in the thylakoid membrane drives both photosynthetic NADPH and ATP production, while cyclic electron flow (CEF) around photosystem I only promotes the translocation of protons from stroma to thylakoid lumen. The chloroplast NADH-dehydrogenase-like complex (NDH) participates in one CEF route transferring electrons from ferredoxin back to the plastoquinone pool with concomitant proton pumping to the lumen. CEF has been proposed to balance the ratio of ATP/NADPH production and to control the redox poise particularly in fluctuating light conditions, but the mechanisms regulating the NDH complex remain unknown. We have investigated the regulation of the CEF pathways by the chloroplast NADPH-thioredoxin reductase (NTRC) in vivo by using an Arabidopsis knockout line of NTRC as well as lines overexpressing NTRC. Here we present biochemical and biophysical evidence showing that NTRC activates the NDH-dependent CEF and regulates the generation of proton motive force, thylakoid conductivity to protons and redox homeostasis between the thylakoid electron transfer chain and the stroma during changes in light conditions. Further, protein-protein interaction assays suggest a putative TRX-target site in close proximity to the ferredoxin binding domain of NDH, thus providing a plausible mechanism for regulation of the NDH ferredoxin:plastoquinone oxidoreductase activity by NTRC.

INTRODUCTION

In their natural habitats, plants face constant fluctuation of light intensity, including both seasonal changes in photoperiod and daily fluctuations according to environmental conditions. Optimization of photosynthesis in plant leaves requires strict balancing between conversion of light energy to chemical energy in photosynthetic light reactions and the energy-consuming reactions of chloroplast metabolism. Multiple regulatory and photoprotective mechanisms have evolved in photosynthetic organisms to cope with fluctuating light conditions and to prevent the photodamage of both Photosystem (PS) II and PSI (Tikkanen et al., 2012; Tikkanen and Aro, 2014; Tiwari et al., 2016; Townsend et al., 2017). Regularly occurring light variations induce long-term acclimatory changes in the photosynthetic machinery via signaling mechanisms, while temporary fluctuation of light within a day transiently activates short-term regulatory mechanisms (Bailey et al., 2001; Grieco et al., 2012; Kono and Terashima, 2014; Armbruster et al., 2014). The short-term mechanisms include non-photochemical quenching (NPQ), photosynthetic control of electron flow between PSII and PSI, state transitions (ST), cyclic electron flow (CEF), and activation of photosynthetic enzymes both in light and carbon fixation reactions (Demmig-Adams et al., 2012; Tikkanen and Aro, 2014; Balsera et al., 2014; Yamori et al., 2016; Gollan et al., 2017). Discovery of redox-regulated activation of Calvin Benson Cycle (CBC) enzymes by chloroplast thioredoxins was the first indication of direct regulatory coupling between photosynthetic light reactions and carbon fixation reactions (reviewed by (Buchanan, 2016). Later on, thioredoxins have been shown to be involved also in the regulation of thylakoid electron flow via redox-control of the ATP synthase (McKinney et al., 1978; Nalin and McCarty, 1984), the STN7 kinase (Rintamäki et al., 2000), and NPQ (Hall et al., 2010; Brooks et al., 2013; Hallin et al., 2015; Naranjo et al., 2016), suggesting that thioredoxins are important members of the chloroplast regulatory network controlling photosynthetic redox balance in fluctuating light conditions (Nikkanen and Rintamäki, 2014).

Light drives the electron flow from water through PSII, plastoquinone (PQ), cytochrome b6f, plastocyanin (PC) and PSI to ferredoxin and ultimately to NADP⁺, producing NADPH. These photosynthetic electron transfer reactions are coupled to ATP synthesis via translocation of protons to the thylakoid lumen, generating a proton gradient over the thylakoid membrane (ΔpH), which together with membrane potential (ΔΨ) constitutes the proton motive force (pmf) (Junesch and Gräber, 1991; Armbruster et al., 2017). ΔpH also contributes to induction of NPQ, a photoprotective mechanism that dissipates excess excitation energy from the electron transfer chain (Niyogi and Truong, 2013; Ruban,2016),and maintains photosynthetic control at Cyt b6f (Joliot and Johnson, 2011; Johnson, 2011). Other regulatory mechanisms include the reversible rearrangements of light harvesting complexes to balance the excitation of PSII and PSI known as state transitions (Tikkanen et al., 2006; Ruban and Johnson, 2009; Rochaix, 2011) as well as cyclic electron flow around PSI (CEF), a process where electrons are transferred from ferredoxin back to the PQ pool. CEF contributes to generation of ΔpH and therefore to production of ATP, and has been suggested to adjust the ATP/NADPH ratio in chloroplasts according to the needs of the CBC (for a recent review, see (Yamori and Shikanai, 2016)). Moreover, CEF provides an alternative electron valve to relieve stromal over-reduction that is needed to protect the photosystems from damage during early developmental stages of chloroplasts (Allorent et al., 2015; Suorsa, 2015) and during excess illumination or fluctuating light conditions (Miyake et al., 2004; Suorsa et al., 2012; Yamori and Shikanai, 2016; Yamori et al., 2016). CEF has also been shown to be important for regulation of the proton motive force (Wang et al., 2015; Shikanai and Yamamoto, 2017),and during induction of photosynthesis (Joliot and Joliot, 2002; Fan et al., 2007). Fan et al. (2007) calculated that CEF contributes a maximum of 68% of total electron flux after 30 s illumination of spinach leaves with red and far red light.

Two distinct and partially redundant pathways of CEF have been suggested to exist in plant chloroplasts (Munekage et al., 2004). One CEF pathway involves the chloroplast NADH dehydrogenase-like complex (NDH), an orthologue of mitochondrial respiratory complex I (Shikanai, 2016; Peltier et al., 2016). However, unlike complex I, which is reduced by NADH, the chloroplast NDH complex is reduced by ferredoxin (Yamamoto et al., 2011; Yamamoto and Shikanai, 2013). It has been suggested recently in several studies that CEF via the NDH complex is essential for photosynthesis in low light conditions (Yamori et al., 2015; Kou et al., 2015; Martin et al., 2015) as well as for the tolerance of drought (Horvath et al., 2000) and low temperature (Yamori et al., 2011). The antimycin A–sensitive CEF pathway depends on the proteins PROTON GRADIENT REGULATION 5 (PGR5) (Munekage et al., 2002) and PGR5-LIKE 1 (PGRL1) (DalCorso et al., 2008), and has been suggested to constitute the hypothetical ferredoxin-plastoquinone reductase (FQR) (Hertle et al., 2013). However, controversy still exists over the molecular identity of FQR and the physiological function of PGR5 (Leister and Shikanai, 2013; Tikkanen and Aro, 2014; Kanazawa et al. 2017). In general, the CEF activity is highly dependent on stromal redox state (Breyton et al., 2006). In fact, both the PGR-dependent pathway (Hertle et al., 2013; Strand et al., 2016a) and the NDH pathway (Courteille et al., 2013) have been proposed to be subject to thiol-regulation by stromal thioredoxins.

In chloroplasts of higher plants, two thioredoxin systems function in parallel. The Ferredoxin-thioredoxin system depends on photosynthetically reduced ferredoxin to supply electrons to the Ferredoxin-thioredoxin reductase (FTR), which in turn reduces several thioredoxins, namely TRX-f1 and f2, four isoforms of TRX-m, TRX-x as well as TRX-y1 and y2 (Schürmann and Buchanan, 2008; Yoshida and Hisabori, 2017). The other system consists of a single enzyme, NADPH-thioredoxin reductase (NTRC) that contains both a reductase and a thioredoxin domain (Serrato et al., 2004). NTRC is reduced by NADPH, which is produced, besides in the light reactions, also in the oxidative pentose phosphate pathway (OPPP) in darkness. Both chloroplast TRX systems are essential for normal development and growth of plants (Serrato et al., 2004; Wang et al., 2014). The ntrc knockout has a stunted and low chlorophyll phenotype, which is particularly severe in plants grown under short photoperiods (Perez-Ruiz et al., 2006; Lepistö et al., 2009; Lepistö et al., 2013). The mutant suffers from impaired ability to activate the ATP synthase and CBC enzymes as well as elevated non-photochemical quenching (NPQ) (Nikkanen et al., 2016; Carrillo et al., 2016; Naranjo et al., 2016; Thormählen et al., 2017). In contrast, the NTRC overexpression line (OE-NTRC), with 15–20 times higher NTRC content compared to WT, shows enhanced vegetative growth and increased activation of the ATP synthase and CBC enzymes particularly in darkness and low light (Toivola et al., 2013; Nikkanen et al., 2016). NTRC has a less negative midpoint redox potential than FTR (Hirasawa et al., 1999; Yoshida and Hisabori, 2016) and plays an important regulatory role under low irradiance, while the FTR-dependent system probably requires more extensive illumination to be fully activated (Thormählen et al., 2017; Geigenberger et al., 2017; Nikkanen et al., 2016). Recent studies have revealed significant functional overlap and crosstalk between the two chloroplast TRX systems, and indicated that they cooperatively regulate ATP synthesis, the CBC, starch synthesis and scavenging of reactive oxygen species (ROS) (Thormählen et al., 2015; Nikkanen et al., 2016; Pérez-Ruiz et al., 2017; Geigenberger et al., 2017). Moreover, redox-regulation of both CEF pathways has been previously reported (Courteille et al., 2013; Hertle et al., 2013; Strand et al., 2016a). The physiological roles of each CEF pathway and TRXs involved in the regulation are nevertheless still unclear.

Here we have used the ntrc knockout mutant as well as NTRC overexpression lines of Arabidopsis thaliana to investigate the potential role of the NTRC system in regulating CEF. Our results emphasize the important role of thioredoxins in the chloroplast regulatory network, particularly controlling the photosynthetic redox balance under fluctuating light conditions. NTRC plays a crucial role in activation of the NDH-dependent CEF in darkness (chlororespiration) and during dark to light transitions. Overexpression of NTRC, on the other hand, maintains constant NDH-CEF activity leading to elevated pmf and improved utilization of light energy under fluctuating light conditions. Our results also suggest that NTRC does not activate the PGR-dependent CEF, but contributes to the PGR5-dependent downregulation of thylakoid membrane proton conductivity upon transitions to high light intensity. Through control of both CEF and the activity of the ATP synthase, NTRC plays a pivotal role in adjusting the proton motive force and photosynthetic redox poise in Arabidopsis chloroplasts.

RESULTS

NTRC is an active reductant in dark and in low light conditions

NADPH produced in the oxidative pentose phosphate pathway (OPPP) has been proposed to maintain NTRC partially reduced, and thus active in darkness and when low irradiance limits photosynthesis (Perez-Ruiz et al., 2006; Geigenberger et al., 2017). To confirm this hypothesis, we analyzed the in vivo redox state of NTRC by a mobility shift assay using the WT or OE-NTRC protein extracts alkylated with methoxypolyethylene glycol maleimide (MAL-PEG). The assays indicated that NTRC redox state remains fairly constant in all light intensities and during dark-to-light transitions, with a significant proportion of the enzyme pool in fully or partially reduced form (Fig. 1A and 1C). This is also the case in OE-NTRC, despite the increase in NTRC content of leaves (Fig. 1B, Suppl. Fig. 1A). These results are in agreement with the hypothesis that NTRC acts as a thiol regulator of photosynthesis and chloroplast metabolism in darkness and low light conditions (Nikkanen et al., 2016; Carrillo et al., 2016; Thormählen et al., 2017).

Figure 1.

In vivo redox state of NTRC in dark-adapted and illuminated leaves.(A) and (B) Total protein extract was isolated from WT (A) and OE-NTRC (B) leaves incubated in darkness (D), or illuminated for 2h in low light (LL, 40 μmol photons m^-2 s^-1), growth light (GL, 200 μmol photons m^-2 s^-1) or high light (HL, 800 μmol photons m^-2 s^-1). Free thiols of proteins were blocked with NEM, disulfides reduced with DTT and newly formed thiols alkylated with MAL-PEG. The in vivo-reduced form of NTRC therefore migrates faster in SDS-PAGE than the in vivo oxidized forms. -DTT stands for the unlabeled control sample where DTT was not added after incubating the leaf extracts in a buffer containing NEM. Protein content of samples has only been equalized based on the amount of starting leaf material, and the apparent differences in band intensity should not be taken as indication of differences in NTRC content between light treatments. For an analysis of the origin of different MAL-PEG labelled bands see Suppl. Fig. 1.(C) NTRC redox state in WT during a transition from dark to growth light. Samples were taken from darkness (2h) (D) and 15, 30, 45 and 60 seconds after onset of illumination.

NDH-dependent CEF is enhanced by overexpression of NTRC

In order to determine the effect of altered chloroplast thiol-redox state on the activity of NDH-dependent CEF, we measured the post-illumination rise of chlorophyll a fluorescence (PIFR). The PIFR has been suggested to represent electron flow from stromal reductants via the NDH-complex to the plastoquinone (PQ) pool upon cessation of illumination (Shikanai et al., 1998; Gotoh et al., 2010). The OE-NTRC line showed a significantly larger PIFR after pre-illumination with low intensity white light than WT, suggesting increased CEF activity (Fig. 2A). In agreement with previous reports, no PIFR was detected in the ndho mutant, which is lacking a functional NDH complex (Rumeau et al., 2005), while a diminished PIFR was observed in the pgr5 line, which is deficient in PGR-dependent CEF (Munekage et al., 2002) (Fig. 2B). These results suggest that NTRC contributes to activation of NDH-dependent CEF. In order to confirm that the increased PIFR in OE-NTRC derives from the activity of the NDH complex, we generated an NTRC overexpression line in the ndho mutant background (OE-NTRC ndho). PIFR was indeed fully eliminated in that line (Fig. 2C). The level of NTRC overexpression in OE-NTRC ndho plants was confirmed by immunoblotting and found to be similar to the OE-NTRC line (Suppl. Fig. 1B).

Figure 2.

Post-illumination fluorescence rise (PIFR) in dark-adapted leaves.(A) and (B) PIFR was measured from WT, OE-NTRC, ntrc (A), pgr5 and ndho (B) leaves. The smaller windows show magnifications of the ∼100 s of the PIFR. The cyan bars indicate exposure to a 480 nm measuring light of 0.28 μmol photons m^-2 s^-1, the white bar depicts illumination with 67 μmol photons m^-2 s^-1 white light and the red bar shows the duration of a pulse of far red light. The dashed lines indicate the F0 values of the lines. The curves are averages from 3-7 measurements of individual leaves.(C) PIFR in WT and OE-NTRC ndho. Only the post-illumination phase of the experiment is shown in the figure. The curves are averages from 3-4 measurements of individual leaves.

The ntrc knockout exhibited a slower initial PIFR response after 500 s of light, but the PIFR did not decline after 15–20 s as in WT or OE-NTRC, but instead continued to rise throughout the duration of the dark phase of the experiment (Fig. 2A). A brief pulse of far red (FR) light quenched the fluorescence, but after cessation of the FR light, fluorescence quickly rose back to its high pre-FR level. The F0 level was elevated in dark-adapted ntrc leaves compared to WT, OE-NTRC or other mutant lines (Fig. 2A). The abnormal fluorescence pattern in Fig. 2A may also be due to the highly pleiotropic phenotype of the ntrc mutant, particularly when grown under a short day photoperiod. To clarify whether the differences in PIFR were caused indirectly by metabolic disturbances due to impaired growth in a short photoperiod, we repeated the experiment with plants grown in a 12h/12h photoperiod. PIFR mostly disappeared in 12h photoperiod grown ntrc, but remained similar to 8h photoperiod grown plants in WT and OE-NTRC (Suppl. Fig. 2A).

NTRC increases dark-reduction of the plastoquinone pool

As overexpression of NTRC increased NDH-dependent CEF in darkness after illumination (Fig. 2), we analyzed the phosphorylation level of LHCII proteins in dark-adapted and illuminated leaves to determine if the higher CEF activity alters the redox state of the PQ pool. Reduction of the PQ pool induces phosphorylation of LHCII by activating the STN7 kinase through interaction with the Cyt b6f complex (Vener et al., 1997; Bellafiore et al., 2005; Shapiguzov et al., 2016). In WT LHCII proteins were mostly non-phosphorylated in darkness, maximally phosphorylated in low light, moderately phosphorylated in growth light and mostly de-phosphorylated in high light (Fig. 3H), in agreement with earlier studies (Rintamäki et al., 1997; Tikkanen et al., 2010). The stn7 mutant was unable to phosphorylate LHCII (Bellafiore et al., 2005). In contrast to WT, LHCII was phosphorylated in darkness in OE-NTRC (Fig. 3H). As in WT, only a small amount of phosphorylated LHCII was present in thylakoids isolated from dark-adapted leaves of ntrc. Interestingly, de-phosphorylation of LHCII proteins in high light was slightly impaired in ntrc (Fig. 3A). This is in accordance with earlier studies showing that phosphorylation of LHCII is highly dependent on stromal thiol redox state in high light (Rintamäki et al., 2000; Martinsuo et al., 2003).

Figure 3.

Redox state of the PQ pool in dark-adapted leaves.(A-F) PQ pool redox state in darkness was determined from measurement of the OJIP transients in dark-adapted leaves (black) and leaves pre-illuminated with far red light (red). Data is presented as averaged curves from 5-10 individual measurements on a logarithmic time scale from WT (A), OE-NTRC (B), ntrc (C), pgr5 (D),ndho (E) and OE-NTRC ndho (F).(G) Proportions of reduced QA calculated as (FJ-FJ ox)/ (Fm-FJ ox). Values are averages from 5-10 measurements from individual leaves ± SE. * indicates statistically significant difference to WT according to Student’s T test (P<0.05). All values are normalized to dark-adapted Fm.(H) Determination of phosphorylation status of LHCII proteins in WT, ntrc, OE-NTRC andstn7 after 2 h of darkness and in low light (40 μmol photons m^-2 s^-1), growth light (200 μmol photons m^-2 s^-1) and high light (600 μmol photons m^-2 s^-1). 0.4 μg of thylakoid extracts were separated with SDS-PAGE and detected with a Phosphothreonine-specific antibody. Coomassie Brilliant Blue staining of LHCII on the membrane (CBB-LHCII) was used as loading control.

To further investigate the effect of NTRC on PQ pool reduction in darkness, we measured the kinetics of Chl a fluorescence OJIP transients in dark-adapted leaves and in leaves pre-illuminated with far red light (FR) to fully oxidize the PQ pool. The difference in F/Fm at the J phase of the transient (FJ, 3 ms after onset of illumination) between dark-adapted and pre-illuminated leaves is an indicator for the redox state of the PQ pool in darkness (Toth et al., 2007; Stirbet et al., 2014). The results suggesteda significantly larger proportion of reduced PQ in darkness in the OE-NTRC line when compared to WT (Fig. 3A–B, Suppl. Table S1). The ndho mutant had a more oxidized PQ pool in darkness than WT, while there was no significant difference between pgr5 and WT (Fig. 3G), suggesting that the NDH complex is the main CEF pathway contributing to dark-reduction of the PQ pool. Attribution of the increased dark-reduction of PQ in OE-NTRC to enhanced activity of the NDH complex was further supported by the observation that redox state of the PQ pool in darkness was more oxidized in the OE-NTRC ndho line in comparison to WT and OE-NTRC (Fig. 3G). OJIP transients also showed increased dark-reduction of PQ in ntrc mutant when compared to WT (Fig. 3C), but it must be noted that the overall kinetics of the OJIP transient in ntrc differed considerably from the other lines, which may affect interpretability of the results from the pleiotropic ntrc knockout line.

Differences in PQ redox state could be caused by altered content of PSII or PSI complexes, the NDH complex, Cytochrome b6f or plastid terminal oxidase (PTOX). No significant differences were detected in the amounts of the PSII core protein D1, Cyt b6f subunit Cyt f or NDH subunits NhdS and NdhH between the studied lines, while a decrease in the amount of PGR5 and the PSI core protein PsaB, and an elevated PTOX content were detected in ntrc (Fig. 4).

Figure 4.

Content of proteins functioning in the photosynthetic electron transfer chain (PETC), cyclic electron flow (CEF) and chlororespiration in WT, ntrc, OE-NTRC, pgr5 andndho.(A) Representative immunoblots showing the content of D1, PsaB, Cytf NdhH, NdhS, PGR5, PGRL1 and PTOX. Appropriate amount of thylakoid extract was separated with SDS-PAGE and probed with specific antibodies.(B) Averages of quantified protein content ±SE in 3-5 biological replicates. Statistically significant differences to WT according to Student’s T-tests (P<0.05) are marked with *. Equal loading was confirmed by and protein quantification normalized according to staining with Li-Cor Revert Total Protein Stain.

NTRC enhances the formation of the proton motive force during dark-to-light transitions

Reduction of plastoquinone by the thylakoid NDH complex is known to be coupled to translocation of protons to the lumen, which contributes to the formation of pmf and enhances ATP synthesis (Strand et al., 2017a). We therefore investigated whether formation of the pmf during dark-to-light transitions and in plants illuminated with different light conditions is affected by deficiency or overexpression of NTRC. The pmf, conductivity of the thylakoid membrane to protons (gH+) and proportions of the pmf components ΔpH and ΔΨ were determined by measuring absorbance changes at 515–550 nm, also known as the electrochromic shift (ECS) (Cruz et al., 2005). Upon onset of illumination at growth light intensity pmf was transiently elevated in WT, peaking after 15–20 seconds (Fig. 5A) and coinciding with a decrease in gH+ (Fig. 5C). The initial decrease in gH+ occurred despite rapid reduction of the gamma subunit of the ATP synthase (CF₁γ), as after 20 s under growth light CF₁γ was already fully reduced (Fig. 5G) (Kramer et al., 1990). Under growth light intensity, another slight rise in pmf was observed after ca. 30–60 seconds in light, coinciding with P700 oxidation (Fig. 5A and 6D). Subsequently pmf slowly decreased to a steady state value.

Figure 5.

Generation of the proton gradient during dark-to-light transitions.(A-B) Proton motive force(pmf) at specific time points during transitions from dark to 166 μmol photons m^-2 s^-1 actinic light (AL) in dark-adapted leaves of WT Col-0, OE-NTRC, ntrc, ndho and OE-NTRC ndho (A) and in WTCol-gl1, OE-NTRC, pgr5 and OE-NTRC pgr5 (B). The pmf was measured as light-induced change in the ECS signal(ECST) and normalized with the magnitude of ECS induced by a 20 μs saturating single turnover flash administered prior to the onset of AL (ECSST). The small window in the bottom left corner of (A) shows representative P700 oxidation kinetics measured with the Dual-PAM spectrometer in synchronization with the ECS data. Saturating pulses were administered with 15 s intervals during the P700 measurement. Values are averages from 4-16 individual measurements ±SE.(C-D) Conductivity of the thylakoid membrane to protons (gH+), calculated as the inverse of the time constant of a first order fit to the decay of ECS during 250 ms dark intervals.(E) Totalpmf and its partitioning to ΔpH and ΔΨ after 3 min illumination with low, growth or high light.(F) Data in (E) represented as percentages of total pmf. Values are averages from 4-16 individual measurements ±SE, and statistically significant differences to WT according to Student’s T-tests (P<0.05) are marked with *.(G) Mobility shift assays with MAL-PEG labelled protein extracts to determine the in vivo redox state of CF₁γ during first 60 seconds of dark-to-growth light transitions. * marks an unspecific band of unknown origin.

Figure 6.

Photosynthetic parameters during dark-to-light transitions.(A-E) Induction of non-photochemical quenching (A), photosystem II quantum yield (B), redox state of the PQ pool (1-qL) (C), P700 oxidation (Y(ND)) (D) and PSI acceptor side limitation (Y(NA)) (E) were calculated from Chl a fluorescence and 870-820 nm absorbance changes during transitions from dark to 166 μmol photons m^-2 s^-1 actinic light in dark-adapted WT Col-0, ntrc, OE-NTRC,ndho, OE-NTRCndho, WT Col-gl1, pgr5 and OE-NTRC pgr5 leaves. The graphs are averages from 4-9 individual measurements ±SE.

In OE-NTRC the initial pmf increase occurred already in a few seconds after the onset of illumination, reached a higher level and decreased more slowly than in WT (Fig. 5A). This initial pmf increase coincided with a transient peak in P700 oxidation both in OE-NTRC and in WT (Fig. 5A). While gH+ was drastically elevated in dark-adapted OE-NTRC leaves, it rapidly decreased to a level comparable to WT (Fig. 5C). The pmf was higher in OE-NTRC in all light intensities when compared to WT, and the differences were mainly due to elevated ΔpH (Fig. 5E). There was no significant difference between OE-NTRC and WT in gH+ under growth light level illumination apart from the enhanced conductivity in dark-adapted leaves (Fig. 5C) as shown previously (Nikkanen et al., 2016). Moreover, the PSII quantum yield (Fig. 6B) increased only slightly during early photosynthetic induction and slightly but insignificantly decreased at steady state illumination in comparison to WT, while P700 oxidation was significantly enhanced (Fig. 6D). These results strongly suggested that the increase in pmf in OE-NTRC derives from CEF. Despite a higher ΔpH under growth light intensity in OE-NTRC, we observed elevated NPQ only during the first minute after the dark-to-light transition (Fig. 6A). At steady state OE-NTRC had similar NPQ to WT.

In ndho, generation of pmf followed similar kinetics as in WT, but its level was diminished (Fig. 5A). In comparison to WT, thylakoid proton conductivity was increased during the 30–60 s time period (Fig. 5C) and NPQ induction was slightly delayed (Fig. 6A). However, overexpression of NTRC in ndho background resulted in elevation of pmf during transitions from dark to growth light similarly to OE-NTRC, except for a time period between 15–40 s after onset of illumination, where pmf was lowered in OE-NTRC ndho in comparison to OE-NTRC (Fig. 5A). Upon dark-to light transitions the high initial thylakoid proton conductivity decreased like in OE-NTRC, but after 10 s of illumination gH+ again rose more rapidly than in OE-NTRC (Fig. 5C). As in ndho, NPQ induction was slightly delayed in OE-NTRC ndho (Fig. 6A). In the absence of the NDH complex, NTRC overexpression was also unable to enhance P700 oxidation during dark-to-growth light transitions, and OE-NTRC ndho suffered from increased PSI acceptor side limitation in comparison to WT (Fig. 6D and 6E). This data indicates that the enhanced capacity of the stroma in OE-NTRC to pull electrons from PSI during dark-to-light transitions is dependent on the NDH-complex.

Increased activation of the NDH complex is not, however, sufficient to fully explain the elevated pmf in OE-NTRC, especially immediately after dark-to-light transitions and at steady state. Therefore, we also generated an NTRC overexpression line in the pgr5 mutant background (OE-NTRC pgr5) whose NTRC expression level was comparable to OE-NTRC (Suppl. Fig. 1B). In the pgr5 mutant, pmf generation and NPQ induction at the onset of illumination were impaired (Fig. 5B and Fig. 6A). The secondary pmf rise was missing and the steady state level of pmf was lower than in WT, in part due to increased gH+ (Fig. 5B and 5E). Absence of PGR5 did not alter the kinetics of pmf induction in plants overexpressing NTRC, but the magnitude of pmf remained lower (Fig. 5B). At steady state, pmf in the OE-NTRC pgr5 plants did not differ from pgr5 plants (Fig. 5B), while proton conductivity of the thylakoid membrane was drastically increased even in comparison to pgr5 (Fig. 5D). Moreover, PSI acceptor side limitation was only slightly alleviated in OE-NTRC pgr5 in comparison to pgr5 (6E). Interestingly, a high initial gH+ value and rapid generation of a pmf peak after onset of illumination were observed in OE-NTRC, OE-NTRC ndho and OE-NTRC pgr5 plants but missing in WT, ndho and pgr5 plants (Fig. 5A and 5B). This demonstrates that the first rapidly-induced pmf peak in plants overexpressing NTRC is not caused by increased CEF.

Upon onset of illumination at growth light intensity, the initial pmf increase in ntrc occurred with similar kinetics to WT, but had a lesser magnitude (Fig. 5A). The secondary pmf increase was absent in ntrc leaves. At steady state the strenght of pmf was comparable to WT, but contribution of ΔpH to total pmf was slightly diminished (Fig. 5E–F). NPQ was however elevated (Fig. 6A), implying active ΔpH-independent upregulation of NPQ. Decreased thylakoid conductivity to protons was observed in growth light intensity in ntrc (Fig. 5C). Reduction of CF₁γ is impaired in ntrc under low light but not in growth light (Nikkanen et al., 2016), implying that thylakoid proton conductivity is inhibited by other means. In ntrc, pgr5 and OE-NTRC pgr5 QA was maintained in a highly reduced state during dark-to-growth light transitions (Fig. 6C). In pgr5 and OE-NTRC pgr5 this was due to impaired induction of NPQ (Fig. 6A), lack of photosynthetic control at Cyt b6f and the consequential inability to oxidize PSI (Suorsa et al., 2012) (Fig. 6D). In ntrc, however, high donor side limitation of PSI was measured in growth light despite high excitation pressure in the PQ pool (Fig. 6D) and decreased PsaB content (Fig. 4), suggesting inhibition of electron transfer between the PQ pool and PSI, possibly at Cyt b6f. Lower and higher content of PGR5 and PTOX, respectively, (Fig. 4) may assist to relax excitation pressure in the PQ pool of ntrc plants.

In fluctuating light NTRC overexpression enhances PSI yield in low light and represses thylakoid conductivity to protons upon transitions from low to high irradiance

As it has been suggested in recent reports that NTRC is particularly important under low and fluctuating light conditions (Nikkanen et al., 2016; Carrillo et al., 2016; Thormählen et al., 2017), we next investigated pmf formation and photosynthetic electron transfer under these conditions. During transitions from dark to low light intensity, increased pmf formation was again observed in OE-NTRC, but the difference to WT was less dramatic than in growth light (Fig. 7A). PSII yield was enhanced in OE-NTRC during the transition from dark to low light (Fig. 8A), contributing to the pmf increase. P700 oxidation was also enhanced during dark to low light transitions (Fig. 9B), while NPQ was decreased, despite higher ΔpH (Fig. 8B). A high PSI yield was maintained throughout the low light periods in OE-NTRC due to low acceptor side limitation (Fig. 9). Overexpression of NTRC in ndho background reverted these changes observed in OE-NTRC plants to levels comparable to ndho knockout plants, except for the increased PSII yield during dark-to-low light transitions (Fig. 7, 8 and 9), suggesting that enhanced activation of NDH-mediated CEF contributes to photosynthetic performance of OE-NTRC during dark to light transition and under low light.

Figure 7.

Formation and regulation the proton motive force during changes in light conditions.(A-B) The pmf (A) and proton conductivity of the thylakoid membrane(gH+) (B) at specific time points during transitions from darkness to low actinic light (39 μmol photons m^-2 s^-1) and from low to high light (825 μmol photons m^-2 s^-1) in dark-adapted leaves of WT Col-0, OE-NTRC, ntrc, ndho and OE-NTRC ndho. The pmf and gH+ were measured and calculated as explained in the legend for Figure 5. The graphs shown are averages from 4-13 individual measurements ±SE.(C-D) The pmf and gH+ in dark-adapted leaves of WT Col-g/7, OE-NTRC, pgr5 and OE-NTRC pgr5. The graphs shown are averages from 3-13 individual measurements ±SE.

Figure 8.

Analysis of Chlorophyll a fluorescence in fluctuating light.(A-C) PSII yield (Y(II) (A), non-photochemical quenching (NPQ) (B) and redox state of the PQ pool (1-qL) (C) in light conditions fluctuating between periods of low actinic light (LL, 39 μmol photons m^-2 s^-1) and high light (HL, 825 μmol photons m^-2 s^-1) in WT Col-0, OE-NTRC, ntrc, ndho, OE-NTRC ndho, WT Col-g/7, pgr5 and OE-NTRC pgr5. Five weeks old plants were dark-adapted for 30 min before measuring fluorescence from detached leaves. All values are averages of 3-10 individual measurements ±SE.

Figure 9.

Analysis of P700 oxidation in fluctuating light.(A-C) PSI yield (Y(I) (A), P700 oxidation (Y(ND)) (B) and PSI acceptor side limitation (Y(NA)) (C) in light conditions fluctuating between periods of low actinic light (LL, 39 μmol photons m^-2 s^-1) and high light (HL, 825 μmol photons m^-2 s^-1) in WT Col-0, OE-NTRC, ntrc, ndho, OE-NTRC ndho, WT Col-g/7, pgr5 and OE-NTRC pgr5. Five weeks old plants were dark-adapted for 30 min before measuring fluorescence from detached leaves. All values are averages of 3-10 individual measurements ±SE.

Both pgr5 and ndho showed impaired pmf generation during transitions from dark to low light, with most severe impairment occurring in pgr5 30–60 seconds after the onset of low illumination. (Fig. 7A and B). Notably, this is the time frame where transient NPQ and P700 oxidation are induced in WT but not in pgr5 (Fig. 8B and 9B). In contrast, pmf generation in OE-NTRC pgr5 during dark-to-low and low-to-high light transitions was recovered to WT levels (Fig. 7A) despite elevated gH+ especially in high light (Fig. 7D), most likely due to enhanced activity of NDH-CEF. Slightly improved tolerance to light fluctuation was also observed in OE-NTRC pgr5 in comparison to pgr5 as the PSI yield was better maintained in low light following the high light periods (Fig, 9A). Importantly, overexpression of NTRC improved the ability to oxidize PSI in low light even in the pgr5 background (Fig. 9B, 9C).

Upon the switch from low to high light intensity, a transient pmf spike was observed in all lines (Fig. 7A), likely deriving from a sudden increase in the electric field component of pmf (Δψ) (Davis et al., 2016). In WT, the proton conductivity of the thylakoid memrane was downregulated (Fig. 7B and 8B). The downregulation was not due to oxidation of CF₁γ, as it remained fully reduced in high light conditions (Suppl. Fig. 3B). The downregulation of gH+ was even stronger in OE-NTRC upon the shift from low to high light (Fig. 7B). Despite elevated pmf (Fig. 7A), NPQ induction was attenuated (Fig. 8B). Overexpression of NTRC in ndho background decreased pmf generation during transitions from low to high light in comparison to the OE-NTRC line (Fig. 7A), suggesting that enhanced activation of NDH-mediated CEF contributes to the high pmf in OE-NTRC in these conditions. OE-NTRC ndho showed increased steady-state gH+ under high irradiance similarly to ndho (Fig. 7B). It is therefore likely that the elevated gH+ values in ndho and OE-NTRC ndho (Fig. 5 and Fig. 7) are at least partly caused by lack of NDH-mediated proton influx to the lumen.

In high light the pgr5 mutant was unable to oxidize P700 (Fig. 9B) or to inhibit thylakoid conductivity (Fig. 7D), and the high gH+ led to a loss of pmf (Fig. 7C). The strong decrease of gH+ observed in OE-NTRC in high light (Fig. 7B) disappeared in OE-NTRC pgr5, which lacks PGR5 (Fig. 7B), suggesting that PGR5 contributes to the downregulation of thylakoid proton conductivity under these conditions. Recovery of pmf in high light through NTRC overexpression (Fig. 7C) was not sufficient to induce NPQ (Fig. 8B) or to control excess electron flow to PSI in pgr5 background (Fig. 9B). This supports the hypothesis that the PGR5 protein is directly required to induce photosynthetic control at Cyt b6f (Suorsa et al., 2013; Tikkanen et al., 2015).

In ntrc, high steady state pmf under low light intensity (Fig. 7A) was likely caused by impaired activation of the chloroplast ATP synthase and the Calvin-Benson cycle as previously reported (Nikkanen et al. 2016, Carrillo et al. 2016). Partitioning of the pmf to ΔpH and ΔΨ revealed that the contribution of ΔΨ to total pmf was slightly increased in ntrc and decreased in OE-NTRC in both low and high light (Fig. 5F). Exceptionally high NPQ was recorded in the ntrc line, especially at low light (Fig. 8B).

Concluding from Figures 5–9, it is evident that both the knockout and overexpression of NTRC had a distinct influence on the formation and regulation of pmf as well as on the induction of NPQ during transitions from dark to light and from low to high light through regulation of CEF activity and ATP synthase conductivity.

Identification of CEF-related NTRC target proteins

Distinct effects of NTRC overexpression or deficiency on the post-illumination fluorescence rise (Fig. 2), the dark-reduction level of the plastoquinone pool (Fig. 3) and formation of pmf as well as on the regulation of thylakoid conductivity during dark/light transitions and low/high light transitions (Fig. 5 and 7) suggested that NTRC may either directly or indirectly regulate CEF. In order to screen for potential targets of direct NTRC-mediated regulation, we performed co-immunoprecipitation (Co-IP) assays with an antibody against NTRC, and analyzed eluates from WT, ntrc and OE-NTRC total leaf extracts by mass spectrometry (MS). A full list of identified peptides is provided in Supplemental Dataset 1.

The Co-IP/MS analysis revealed that among peptides that were detected in WT and OE-NTRC eluates but were absent in the eluates from ntrc were several proteins involved in CEF around PSI (Table 1). Most notably, five subunits of the thylakoid NDH complex; NdhH, Ndh48, NdhS, NdhU and NdhO, (in order of abundance) as well as PGR5 were identified as potential NTRC interactors. Furthermore, although Ndh45 and NdhJ were detected in ntrc eluates, they were notably enriched in WT and/or OE-NTRC eluates. Intriguingly, all of the NDH subunits identified are located close proximity to the proposed ferredoxin binding and oxidation site on the stromal side of the NDH complex (Yamamoto et al., 2011; Yamamoto and Shikanai, 2013; Peltier et al., 2016; Shikanai, 2016).

Table 1.

Identification of putative NTRC target proteins in CEF pathways by Co-IP/MS.Only proteins of which at least two unique peptides were detected, and which were absent from ntrc eluates or clearly enriched in WT / OE-NTRC are included. + indicates the presence and - the absence of at least two unique peptides from the protein. MW (kDa) indicates molecular weight, and #Cys the number of conserved cysteine residues (see Suppl. Tables 2-6). For a description of experimental procedures see Materials and methods and for a full list of detected peptides see Suppl. Dataset 1.

The potential interactions of NTRC with NdhS and PGR5 were further supported by positive results in bimolecular fluorescence complementation tests (BiFC). Co-expression of NTRC with both NdhS and PGR5 in Nicotiana benthamiana leaves resulted in YFP fluorescence that was strictly co-localized with chlorophyll autofluorescence, suggesting that it originated from the thylakoid membranes (Fig. 10). TRX-m1 interacted with PGRL1, while no interaction capability was detected between PGRL1 and NTRC (Fig. 10).

Figure 10.

Bimolecular fluorescence complementation (BiFC) tests ofin planta interactions between chloroplast TRXs and potential CEF target proteins.The left panel shows yellow fluorescent protein (YFP) fluorescence in green, the middle panel Chlorophyll a autofluorescence in red and the right panel a merged image of YFP, chlorophyll and brightfield images. YFP-N and YFP-C indicate expression of a fusion proteins including the N-terminal and C-terminal parts of YFP, respectively, in tobacco(Nicotiana benthamiana) leaves.

To assess the potential of the CEF-related proteins identified by Co-IP/MS (Table 1) and BiFC (Figure 10) to be targets of redox-regulation by TRXs, their amino acid sequences were analysed for conserved cysteine residues. Of the NDH subunits identified as putative NTRC targets by Co-IP/MS, NdhS, NdhH, Ndh48, NdhJ and Ndh45 contain cysteine residues that are conserved in angiosperms (Table 1, Suppl. Tables 2–6), and therefore could in theory be subject to thiol regulation. NdhO and NdhU do not contain conserved cysteine residues, but they likely co-precipitate with the NTRC antibody because of their interactions with NdhH and NdhS subunits of the NDH complex, respectively (Shikanai, 2016).

PGR5 has been shown to form redox-dependent heterodimers with PGRL1 which have been proposed to be required for acceptance of electrons from ferredoxin and for reduction PGRL1 (Hertle et al., 2013; Leister and Shikanai, 2013). The mature PGR5 polypeptide contains a single highly conserved cysteine residue (Munekage et al., 2002), which could hypothetically form an intermolecular disulfide with PGRL1 or some other partner, or be a target for S-nitrosylation or glutathionylation (Couturier et al., 2013; Zaffagnini et al., 2016). As direct determination of PGR5 redox state with the alkylation method was not feasible, we investigated if the redox state of PGRL1 is affected by NTRC deficiency or overexpression. PGRL1 contains six conserved thiol-sensitive cysteine residues that form inter- and intramolecular disulfides (Petroutsos et al., 2009; Hertle et al., 2013). We observed that PGRL1 was mostly oxidized in dark-adapted leaves, but underwent a transient reduction during approximately 60 seconds of illumination with growth light (Suppl. Fig. 3). This corresponds with the timescale of NPQ induction (Fig. 6B), as well as with the transient increase in pmf and decrease in gH+ during dark-to-light transitions (Fig. 5A and 5B). No significant difference to WT in PGRL1 reduction or protein content was detected in OE-NTRC (Suppl. Fig. 3, Fig. 4). PGR5 content of thylakoid membranes was, however, decreased in ntrc by 40% in comparison to WT (Fig. 4), in line with observations in a previous study (Yoshida and Hisabori, 2016).

DISCUSSION

The role of CEF around PSI in plant acclimation to fluctuating light conditions has attracted great attention during the past 10 years. Importance of CEF likely relies in its capacity to maintain redox homeostasis in chloroplasts upon fluctuations in light intensity and during dark-to-light transitions (Yamori et al., 2016; Suorsa et al., 2016; Strand et al., 2016b). Plastidial thioredoxin systems, on the other hand, are crucial regulators of chloroplast metabolic reactions in the stroma. It has, however, remained unclear whether the thioredoxin-related redox regulation is also involved in the achievement CEF-mediated redox homeostasis upon exposure of plants to changing light intensities. To this end, we applied here an in-vivo approach to investigate whether NTRC contributes directly to regulation of CEF pathways in chloroplasts.

In-depth analyses of NTRC overexpression lines in respect to thylakoid functional properties provided strong evidence that NTRC indeed is involved in the regulation of CEF in the thylakoid membrane. Compelling support for NTRC-induced activation of the NDH complex was obtained by analyzing thylakoid CEF-related functions in NTRC overexpression lines made on the backgrounds of ndho (OE NTRC ndho) and pgr5 (OE NTRC pgr5), incapable of performing NDH- and PGR-dependent CEF, respectively. Distinct effect of NTRC overexpression on the post-illumination rise of chlorophyll a fluorescence, the redox state of the plastoquinone pool in darkness as well as on generation of the pmf and oxidation of P700 upon dark-to-light transitions and sudden increases in light intensity demonstrated the control of NDH-dependent CEF by NTRC (Figs. 2, 3, 5, 6, 7). Furthermore, evidence for a direct effect of NTRC on the NDH activity was obtained by identification of NDH subunits in a close proximity of the ferredoxin binding site as potential NTRC interactors (Table 1, Fig. 10, Suppl. Dataset 1, Suppl. tables 2–6). Although several NDH subunits were detected by Co-IP/MS, most likely only one or few of these subunits are genuine NTRC targets. The others likely co-precipitate with the NTRC antibody due to reciprocal interactions of the NDH subunits on the stromal side of the thylakoid membrane (Shikanai, 2016; Peltier et al., 2016). Existence of a thiol-regulated component in the ferredoxin binding site would provide a mechanism for dynamic control of the ferredoxin:plastoquinone oxidoreductase activity of the complex in response to fluctuations in light conditions. Redox-regulation of the NDH complex would allow rapid adjustment of pmf and non-photochemical quenching as well as a maintenance of redox balance between the electron transfer chain and the electron sink capacity of stromal acceptors, most importantly the CBC. In high light, less active NDH could prevent the reverse function of the complex (i.e. oxidization of PQ to reduce ferredoxin and transfer of H+ from lumen to stroma) in conditions of high ΔpH and a reduced PQ pool (Strand et al., 2017a).

Notably, the overexpression of NTRC also affects the function of PGR5. Nevertheless, the effect is not necessarily related to the putative role of PGR5 in CEF. Our results, more likely, support the hypothesis (Avenson et al., 2005; Tikkanen et al., 2015; Kanazawa et al., 2017; Armbruster et al., 2017) that PGR5 controls the proton conductivity of the thylakoid membrane that, consequently, affects the generation of pmf.

The photosynthetic parameters measured for the ntrc knockout plants were not always in line with the results obtained with NTRC overexpressing plants (Fig. 3). In short photoperiod the ntrc plants have a highly pleiotropic phenotype (Fig. 4, Supl. Fig. 2) (Kirchsteiger et al., 2009; Pulido et al., 2010; Thormählen et al., 2015; Naranjo et al., 2016 Nikkanen et al., 2016; Pérez-Ruiz et al., 2017) that complicates the interpretation of results from the ntrc line. Thus the OE-NTRC line, whose visible phenotype and development are not considerably dissimilar to WT, provides a more reliable platform to examine the direct effects of NTRC on specific plastidial processes.

Regulation of the pmf and redox homeostasis via NDH and NTRC during changes in light conditions

Overexpression of NTRC caused elevated pmf under all light conditions, while no significant changes were observed in PSII electron transfer rate or thylakoid proton conductivity in comparison to WT (Figures 5–8). These results strongly suggest that the elevation of pmf derives from enhanced CEF. Increased P700 oxidation during dark-to-light transitions in OE-NTRC was fully reverted in OE-NTRC ndho, and a lack of NDH also delayed the ability to oxidize P700 during the high light phases in fluctuating light (Fig. 6, 9). It is therefore evident that the NDH complex regulates the trans-thylakoid pmf as well as the redox balance between the electron transfer chain and the stroma, and this regulation is under the control of the stromal TRX systems, with our results suggesting a specific role for the NTRC system.

While our results demonstrate enhancement of the NDH-dependent CEF by NTRC overexpression (Fig. 2), earlier studies have revealed an inhibition of NDH-dependent CEF upon TRX-m4 overexpression and, conversely, an enhancement in trxm4 mutants (Courteille et al., 2013). Thus, it is conceivable that the two chloroplast TRX systems regulate CEF in an antagonistic way, although it remains to be elucidated how such regulation might be mechanistically accomplished. We propose that in low light and upon sudden changes in the light intensity, NTRC is crucial for activation of the NDH-dependent CEF, while the TRX-m4-dependent inhibition of NDH-CEF requires higher light intensity or longer duration of illumination. Sustained moderate to high-light illumination is required to fully activate the FTR-dependent TRX system (reviewed in Geigenberger et al. 2017) that possibly contributes to downregulation of NDH. In OE-NTRC the NDH-dependent CEF is constitutively active in light, which contributes to elevated pmf in all light intensities. Upon transition from dark to low light, there is less difference between OE-NTRC and WT in terms of pmf formation (Fig. 7), because in those conditions the NTRC-mediated activation of NDH occurs similarly in WT and OE-NTRC.

The NDH complex translocates protons from the stroma to the lumen not only via Cyt b6f, but also functions as a proton pump with a 2 H+/ e^- stoichiometry (Strand et al., 2017a). NDH-mediated CEF therefore contributes relatively more to ΔpH generation and consequently to ATP synthesis and NPQ induction than the PGR-dependent pathway. It has been postulated that the NDH complex is unlikely responsible for CEF during the early induction phase of photosynthesis, due to a low concentration of the complex in thylakoids in relation to the total PSI content (Joliot and Joliot, 2002). However, the NDH complex forms functional CEF-supercomplexes with PSI in stroma thylakoids (Peng et al., 2008), and a single NDH complex can bind up to six PSI complexes (Yadav et al., 2017), indicating that even a relatively low NDH content may have a significant impact on pmf generation.

Redox-regulation of the PGR-dependent CEF pathway and ATP synthase

Increased activation of NDH-CEF is not alone sufficient to explain all observed changes of pmf in OE-NTRC plants. When compared to WT, pmf remained elevated in OE-NTRC ndho during the first seconds of photosynthetic induction and at steady state in growth and high light (Fig. 5A and 7A). These results could be explained by activation of PGR-dependent CEF as well in plants overexpressing NTRC. Stromal thiol redox state has been previously suggested to control the PGR-dependent CEF by a component that has a midpoint redox potential of -310 (Hertle et al., 2013; Strand et al., 2016a). It has also been proposed that m-type TRXs, with redox potentials between -357 and -312 mV (Collin et al., 2003; Yoshida et al., 2015), reduce an intermolecular disulfide in PGRL1 homodimers, and subsequently, the released monomeric PGRL1 may function as the ferredoxin-plastoquinone reductase (Hertle et al., 2013). Here we confirm the previously reported transient reduction of PGRL1 during dark-to-light transitions (Hertle et al., 2013), but NTRC overexpression does not intervene in the reduction (Suppl. Fig. 3). Moreover, TRX-m1 but not NTRC interacts with PGRL1 in BiFC (Fig. 10). Our results thus support the hypothesis that TRX-m is a primary reductant of PGRL1. Crosstalk between NTRC and FTR-dependent systems (Toivola et al., 2013; Thormählen et al., 2015; Nikkanen et al., 2016), and the interaction of NTRC with TRX-m1 in BiFC (Nikkanen et al., 2016), further support the interpretation that the activation of PGR-dependent CEF is indirectly increased in NTRC-overexpressing plants through enhancement of TRX-m reduction. This would also be in line with the steady state pmf increase observed in OE-NTRC ndho in comparison to WT (Fig. 5A and 7A).

Alternatively, NTRC overexpression may affect the function of PGR5 in a way that is independent of its involvement in CEF. Redox regulation of PGR5 may occur to control its association with the ATP synthase during dark-to-light and low-to-high light transitions, and thereby inhibit the conductivity of the ATP synthase in an unknown mechanism, as suggested earlier (Avenson et al., 2005; Tikkanen et al., 2015; Kanazawa et al., 2017; Armbruster et al., 2017). Such a mechanism would result in acidification of the lumen and induction of NPQ, allowing dissipation of excess excitation energy from the electron transfer chain until CBC is activated. This hypothesis is supported by the impaired abilities of pgr5 and OE-NTRC pgr5 to control thylakoid conductivity at early stages of dark-light transitions and upon transitions to high light intensities (Avenson et al., 2005, Fig. 5 and Fig. 7). Furthermore, the elevated NTRC content in leaves accelerated the decline of the thylakoid proton conductivity upon increases in light intensity (Fig. 7B), suggesting that NTRC controls the PGR5-dependent down-regulation of proton efflux from lumen. This is supported by the identification of PGR5 as a potential NTRC interactor (Table 1, Fig. 4, Fig. 10).

However, the absence of the NDH complex, albeit to a lesser extent than the absence of PGR5, also causes an impaired ability to decrease thylakoid proton conductivity in high light (Fig. 7). Thus, it is likely that proton influx to the lumen caused by proton pumping in CEF decreases the measured value of gH+. This could be expected to have a particularly significant impact during dark-to-light and low-to-high light transitions when the fraction of CEF from total electron flow is high (Joliot and Joliot, 2002; Fan et al., 2007), and in mutants, where CEF capacity is altered (Livingston et al., 2010; Wang et al., 2015; Suorsa et al., 2016). Therefore, the deficiency of the NDH and PGR-dependent pathways of CEF in ndho and pgr5 backgrounds, respectively, may contribute to the high gH+ values measured in those lines. Therefore, a hypothesis in which the rapid repression of gH+ in OE-NTRC derives, at least partially, from the enhanced activity of the CEF pathways, cannot be discarded.

The initial strong pmf increase in OE-NTRC after onset of growth light illumination was evident also in both the OE-NTRC ndho and OE-NTRC pgr5 plants (Fig. 5A and B) indicating that this pmf peak is not caused by CEF. More likely, the initial pmf results from dark-activation of the CBC enzymes in plants overexpressing NTRC (Nikkanen et al. 2016), which provides an enhanced ability of the stroma to accept electrons from the PETC upon dark-to light transition and consequently enhances proton pumping to the lumen.

Cooperative regulation of photosynthetic electron transfer and carbon fixation by chloroplast thioredoxin systems

Light-dependent reductive activation of the ATP synthase, the CBC and the NADP-malate dehydrogenase (NADP-MDH) by TRXs has been well established for several decades (reviewed in (Buchanan, 2016). More recently, knowledge of TRX-mediated control has been extended to various regulatory and photoprotective mechanisms of photosynthesis, including regulation of state transitions (Rintamäki et al., 2000; Shapiguzov et al., 2016), NPQ (Hall et al., 2010; Brooks et al., 2013; Naranjo et al., 2016; Da et al., 2017) and CEF (Courteille et al., 2013; Hertle et al., 2013; Strand et al., 2016a). We propose here a model, comprising a cooperative function of the two chloroplast TRX systems with distinct reductants and redox potentials that allows the maintenance of redox homeostasis between the two photosystems and stromal metabolism during fluctuations in light conditions. This is achieved through dynamic regulation of the activities of the ATP synthase, NPQ, the NDH complex, PGRL1/PGR5 as well as the LHCII kinase STN7 by reversible thiol modifications. We propose a specific role for NTRC in regulating NDH-CEF, the ATP synthase and CBC enzymes in low light, dark-to-light transitions and during sudden increases in light intensity, as schematically depicted in Fig. 11.

Figure 11.

A schematic model of the role of chloroplast TRX systems in regulating CEF and the pmf during dark-to-light transitions and fluctuations in light conditions.(A-C) Dark-adapted leaves (A), transition from dark to low light (B) and transition from low to high light (C). Blue color indicates the level of reduction, green and red arrows represent the activating and inhibitory effects, respectively, while orange represents the thiol regulation by NTRC and dark yellow by the Fd-TRX system. Thicker lines depict stronger effect than thin and dotted lines. For details see the text.

In darkness, NDH is partially activated by NTRC, and moderate chlororespiration from NDH to PTOX occurs. Due to inactivity of the ATP synthase and proton pumping activity of NDH, a weak proton gradient over the thylakoid membrane is established. Redox-regulated CBC enzymes are inactive, causing PC and P700 to be reduced due to lack of acceptors. In OE-NTRC, chlororespiration via NDH to the PQ pool is increased due to enhanced activation of NDH. This leads to increased protonation of the lumen and reduction of the PQ pool. Overexpressed NTRC reduces CF₁γ, partially activating the ATP synthase and making it leaky to H+. CBC enzymes are also partially activated by NTRC.

Upon transition from dark to low light, the ATP synthase is fully and CBC enzymes partially reduced by NTRC in WT plants. This causes, however, a lag phase where a lack of stromal acceptors is limiting electron transfer, leading to reduction of the PQ pool. NTRC contributes to activation of NDH-dependent CEF, which alleviates electron pressure at PSI and transiently increases ΔpH and induces NPQ. In OE-NTRC P700 and PC are effectively oxidized upon onset of low illumination, as the acceptor side limitation is negligible due to fully active NDH-dependent CEF, ATP synthase and redox-activated CBC enzymes. This results in an elevated ΔpH and faster induction of NPQ in comparison to WT at the initial phase of illumination. At steady state NPQ is lower than in WT despite of high ΔpH, suggesting downregulation of NPQ by thioredoxin via a ΔpH-independent mechanism, as reported previously by Brooks et al. (2013).

When a leaf shifts from low to high irradiance, both TRX systems become fully active, and the CBC enzymes as well as PGR-dependent CEF are fully activated. NTRC affects PGR5-dependent down-regulation of thylakoid conductivity to protons, possibly through direct PGR5-mediated inhibition of the ATP synthase, which contributes to accumulation of protons in the lumen. Consequently, NPQ and inhibition of electron transfer at Cyt b6f are induced. Electrons are effectively pulled from PSI, and the donor side is limiting electron transfer. In OE-NTRC, increased reduction of PGR5 likely leads to stronger inhibition of thylakoid proton conductivity. This, together with proton pumping by constantly active NDH and possibly through increased TRX-m-mediated activation of the PGR-dependent pathway, results in high ΔpH. NPQ is however lower than in WT due to ΔpH-independent downregulation of NPQ by overexpressed NTRC.

METHODS

Plant material and growth conditions

Experiments have been done with Arabidopsis thaliana wild type (WT) lines of the Columbia ecotype (Col-0 and Col-gl1), and with the following transgenic lines: NTRC overexpression line (Toivola et al., 2013), T-DNA knockout mutants of NTRC (At2g41680, SALK_096776) (LepistÖ et al., 2009), ndho (At1g74880, SALK_ 068922) (Rumeau et al., 2005) and STN7 (AT1G68830, SALK_073254) (Bellafiore et al., 2005) as well as the pgr5 mutant (AT2G05620) (Munekage et al., 2002). The plants were grown in a photoperiod of 8 h light / 16 h darkness at 23 °C under 200 μmol of photons m^-2 s^-1 for all experiments except for the measurements shown in Suppl. Fig. 2, for which plants were grown in a 12 h /12 h photoperiod under 130 μmol m^-2 s^-1. Wild type tobacco (Nicotiana benthamiana) plants used in BiFC tests were grown under 130 μmol photons m^-2 s^-1 at 23 °C in a 16h light/8 h dark photoperiod. The OE-NTRC ndho and OE-NTRC pgr5 lines were generated by Agrobacterium tumefaciens and floral dipping–mediated transformation of the ndho knockout and pgr5 mutant lines, respectively, with the NTRC overexpression construct as described previously (Toivola et al., 2013). The OE-NTRC ndho and OE-NTRC pgr5 plants used in the experiments were heterozygous T2 generation plants that were selected on agar plates with 0.5X Murashige-Skoog medium (MS) (Murashige and Skoog, 1962) and 50 μg/ml kanamycin. The plants were subsequently transferred to soil and grown in a 8 h light / 16 h darkness photoperiod at 23 °C under 200 μmol of photons m^-2 s^-1 for four weeks before usage in the experiments. As control, OE-NTRC plants were similarly selected on kanamycin-containing plates while WT Col-0 and WT Col-gl1 (ecotype of the pgr5 mutant) plants were grown on 0.5X MS-agar plates without antibiotics for an equivalent time.

Measurement of Chlorophyll a fluorescence and P700 oxidation changes

The post-illumination chlorophyll a fluorescence rise (PIFR) was measured from detached leaves with the Multicolor-PAM fluorometer (Walz). A 480 nm measuring beam at an intensity of 0.2 μmol photons m^-2 s^-1 was used to measure fluorescence changes after illumination of dark-adapted (30 min) leaves with 67 μmol photons m^-2 s^-1 of white actinic light for 500 seconds, with saturating pulses in the beginning and at 400 s to determine Fm and Fm’. The actinic light was then switched off and the changes in chlorophyll a fluorescence in the dark were observed for 300 s. A brief pulse of far red light was then given to fully oxidize the PQ pool, and the subsequent re-reduction PQ pool was detected through a rise in Chl fluorescence.

The OJIP transients were recorded with the Multicolor-PAM from dark-adapted (30 min) leaves and from leaves pre-illuminated with far red light (intensity setting 15) for 6 s, according to the method described by (Toth et al., 2007). A saturating pulse of 3000 μmol photons m^-2 s^-1 and measuring light at 440 nm were used in the measurements.

The Dual-PAM-100 was used to simultaneously record the Chl a fluorescence and 870–820 nm absorbance changes during transitions from dark to 166 μmol photons m^-2 s^-1 (Fig. 6) and during a light regime, where a 620 nm AL fluctuates between 39 and 825 μmol photons m^-2 s^-1 (Fig. 8 and Fig. 9). Saturating pulses were administered at 10 or 15 s intervals for the measurements in Fig. 6 and at 15 s intervals for the first minute after onset of illumination, and at 20 s intervals thereafter for Fig. 8 and Fig. 9. Because ntrc leaves are very small in size and low in chlorophyll content, it was in some cases necessary to record from two or three leaves simultaneously to obtain a P700 signal of sufficient quality. The parameters shown were calculated with the Dual-PAM-100 software and are based on (Kramer et al., 2004).

Measurement of electrochromic shift (ECS)

In order to measure the strength and formation kinetics of the proton motive force (pmf), changes in the electrochromic shift (ECS, P515) signal were recorded with the Dual-PAM-100 and the P515/535 accessory module (Walz) (Schreiber and Klughammer, 2008). A dual beam difference signal at 550-515 nm was used to avoid distortion of results by scattering effects. For the dark-to-light and low-to-high light transition measurements in Figures 5 and 7, plants were first dark-adapted for 30 min. A single-turnover (20 μs) saturating flash of 14000 μmol photons m^-2 s^-1 was then applied to obtain ECSST, a maximum absorbance change value that was used to normalize all results to account for differences in leaf thickness and chlorophyll content between individual leaves and lines (Kramer and Crofts, 1989). The obtained values of ECSST were in good correlation with the chlorophyll content in OE-NTRC and ntrc lines reported previously (Toivola et al., 2013). In order to distinguish the light-induced ECS change (ECST) from signal drift and baseline change attributable to zeaxanthin formation (Klughammer et al., 2013), frequent dark intervals of 250 ms were applied during AL illumination.

ECST was calculated as the difference between total ECS in light and an Y0 value obtained from the first-order exponential fit to the decay kinetics of the ECS signal during a dark interval. Total pmf was then calculated as ECST/ECSST. The gH+ parameter, describing thylakoid membrane conductivity to protons, was calculated as the inverse of the time constant of a first-order exponential fit to ECS decay kinetics during a dark interval (Cruz et al., 2001; Avenson et al., 2005; Cruz et al., 2005). A green measuring light at a 2000 Hz pulse frequency was used in all ECS measurements. Partitioning of total pmf to its components ΔpH and ΔΨ was performed as described by (Cruz et al., 2001).

Protein extraction, alkylation of thiols and SDS-PAGE

Proteins and thylakoids were isolated as previously described (Lepistö et al., 2009), while chlorophyll content was determined according to (Porra et al., 1989) and protein content with the Bio-Rad Protein Assay kit. For determination of the redox states of TRX-regulated proteins, leaf proteins were precipitated with trichloroacetic acid (TCA) and free thiols in proteins alkylated with N-ethylmaleimide (NEM, Sigma-Aldrich). After alkylation protein disulfides were reduced with dithiothreitol (DTT, Sigma-Aldrich) and subsequently produced thiols were alkylated with methoxypolyethylene glycol maleimide Mn 5000 (MAL-PEG, Sigma-Aldrich) as described earlier (Nikkanen et al., 2016). Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting was performed as reported in (Nikkanen et al., 2016). For running the MAL-PEG samples pre-cast 4-20% Mini-PROTEAN TGX gels (Bio-Rad) were used, except for the gel in Fig. 1B, Suppl. Fig. 1A and Suppl. Fig. 3A, where a 12% polyacrylamide gel was used. PVDF membranes were probed with antibodies raised against NTRC (Lepistö et al., 2009), D1 (Research Genetics, Inc (Thermo Fisher)), PsaB (Agrisera, AS10 695), Cyt f (kindly provided by L. Zhang), PTOX (kindly provided by M. Kuntz), NdhH (Agrisera), NdhS (Agrisera), CF₁γ (Agrisera, AS08 312), PGRL1 (Agrisera, AS10 725), PGR5 (Agrisera) or phosphothreonine (P-Thr) (New England Biolabs). Membranes were then treated with a horseradish peroxidase (HRP)-conjugated goat anti-rabbit secondary antibody (Agrisera, AS09 602) for 2 h. All immunoblots shown are representative of at least 3 replicate experiments. Quantifications of protein content shown in Fig. 4 were performed using the ImageJ software (Schneider et al., 2012) and normalized according to the intensity of Li-Cor Revert Total Protein Stain. Statistical significance was determined using two-tailed Student’s T-tests for unequal variances with p-values below 0.05 interpreted as statistically significant.

Co-immunoprecipitation and Mass spectrometry

Co-immunoprecipitation (Co-IP) was performed from WT, ntrc and OE-NTRC total leaf extracts using the Pierce Co-IP kit (Thermo-Fisher) and an NTRC-specific antibody as described previously (Nikkanen et al., 2016). Co-IP eluates where denatured and purified by SDS-PAGE in a 6% acrylamide gel with 6 M urea, subjected to in-gel tryptic digestion and the extracted peptides analyzed with the Q Exactive Hybrid Quadruple-Orbitrap mass spectrometer (Thermo-Fisher Scientific) in DDA mode as previously described (Trotta et al., 2016). MS/MS spectra were analyzed with an in-house installation of Mascot (v.2.4) (Matrix Science) search engine and analyzed with Proteome Discoverer (v.1.4) Software (Thermo Scientific), restricting the search to the non-redundant database TAIR10 supplemented with most common laboratory contaminants (Trotta et al., 2016). Peptides were validated by Decoy Database Search, with target false discovery rates (FDR) set to be below 0.01 (strict) or below 0.05 (relaxed).

BiFC tests

Bimolecular fluorescence complementation tests (BiFC) were performed as described in (Nikkanen et al., 2016). For the current study, coding sequences of PGR5, PGRL1a and NdhS obtained from Arabidopsis Biological Resource Center (ABRC) were cloned into pSPYNE-35S and pSPYCE-35S binary vectors (Walter et al., 2004), and the resulting constructs were checked by sequencing. Primer sequences used for cloning are listed in Suppl. Table 7. Imaging of YFP and chlorophyll autofluorescence from N. benthamiana leaves infiltrated with Agrobacterium tumefaciens strain GV3101 carrying the appropriate binary vectors was performed with a Zeiss LSM780 laser scanning confocal microscope at 3 days after infiltration. The negative result between PGRL1:YFP-N and NTRC:YFP-C also serves as a negative control.

Multiple alignment of amino acid sequences

Amino acid sequences of NdhH, Ndh48, NdhS, NdhJ and Ndh45 in Arabidopsis thaliana and, as available, in Populus trichocarpa, Vitis vinifera, Glycine max, Solanum lycopersicum, Oryza sativa, Sorghum bicolor, Brachypodium distachion, Physcomitrella patens, Selaginella moellendorffii and Synechocystis PCC 6803 were obtained from the UniProtKB database and aligned with the Clustal Omega 1.2.4 online alignment tool (Sievers et al., 2011) using default settings.

Accession Numbers

The Arabidopsis Genome Initiative locus identifiers (AGI) used in this paper are listed in Table 1, Suppl. Tables 2–7 and Suppl. Dataset 1.

Supplemental Data

Supplemental Figure 1. NTRC redox state in different light conditions in OE-NTRC and level of NTRC expression in NTRC overexpression lines.

Supplemental Figure 2. Post-illumination fluorescence rise in dark-adapted WT, OE-NTRC and ntrc plants grown in a 12 h/12 h photoperiod.

Supplemental Figure 3.In vivo redox states of PGRL1 and CF₁γ during changes in light conditions.

Supplemental Table 1. Parameters determined from OJIP transients.

Supplemental Table 2. Multiple alignment of NdhS amino acid sequences.

Supplemental Table 3. Multiple alignment of NdhH amino acid sequences.

Supplemental Table 4. Multiple alignment of Ndh48 amino acid sequences.

Supplemental Table 5. Multiple alignment of NdhJ amino acid sequences.

Supplemental Table 6. Multiple alignment of Ndh45 amino acid sequences.

Supplemental Table 7. Primers used for cloning of BiFC constructs.

Supplemental Dataset 1. MS/MS identification of peptides in Co-IP eluates.

AUTHOR CONTRIBUTIONS

L.N. and E.R. designed the research, L.N., J.T., and A.T. performed the research, L.N., J.T., A.T., M.T., and E.R. analyzed the data, L.N. and E.R. wrote the article with input from E-M.A., M.T., A.T., M.G.D., and J.T.

ACKNOWLEDGEMENTS

We thank Jesse Ojala for assistance with the experiments, Esa Tyystjärvi for the invaluable advice for PIFR measurements, Sari Järvi and Marjaana Suorsa for antibodies and advice on their use, Alexandrina Stirbet for expert advice on measurement of OJIP transients, and Mika Keränen, Kurt Ståhle and Tapio Ronkainen for technical assistance. This work was funded by the Academy of Finland Grants 276392 (to E.R.) and 307335 (the Center of Excellence in Molecular Biology of Primary Producers to E-M.A.) and by the Doctoral Program in Molecular Life Sciences in the University of Turku Graduate School (to L.N.).