The evolutionary conserved iron-sulfur protein TCR controls P700 oxidation in photosystem I.

In natural habitats, plants have developed sophisticated regulatory mechanisms to optimize the photosynthetic electron transfer rate at the maximum efficiency and cope with the changing environments. Maintaining proper P700 oxidation at photosystem I (PSI) is the common denominator for most regulatory processes of photosynthetic electron transfers. However, the molecular complexes and cofactors involved in these processes and their function(s) have not been fully clarified. Here, we identified a redox-active chloroplast protein, the triplet-cysteine repeat protein (TCR). TCR shared similar expression profiles with known photosynthetic regulators and contained two triplet-cysteine motifs (CxxxCxxxC). Biochemical analysis indicated that TCR localizes in chloroplasts and has a [3Fe-4S] cluster. Loss of TCR limited the electron sink downstream of PSI during dark-to-light transition. Arabidopsis pgr5-tcr double mutant reduced growth significantly and showed unusual oxidation and reduction of plastoquinone pool. These results indicated that TCR is involved in electron flow(s) downstream of PSI, contributing to P700 oxidation.

Introduction

As an important component of life on Earth, plants and other photosynthetic organisms convert light energy to chemical energy through photosynthesis, which sustains almost all life activities on Earth (Blankenship, 2002). Photosystem (PS) II captures the photon energy to induce water splitting, and the electrons generated in this process are transferred sequentially to plastoquinone (PQ), cytochrome bf (Cyt bf), plastocyanin (PC), PSI, ferredoxin (Fd), and, finally, Fd-NADP+ reductase (FNR), which reduces NADP+ to NADPH (Govindjee et al., 2017). The electron transfer, called photosynthetic linear electron transfer (LET), generates NADPH and the proton-motive force across the thylakoid membranes that is required for ATP synthesis. NADPH and ATP are then used in the Calvin-Benson-Bassham (CBB) cycle to fix CO2 (Blankenship, 2002).

After charge separation in the special pair of the chlorophyll dimer in PSI (P700), electrons are transferred to the primary electron acceptor A0 followed by F, F, and F. The oxidized P700 is then re-reduced by PC (Blankenship, 2002). Fd receives electrons directly from F, which is the main electron transfer route downstream of PSI. When P700 is highly reduced due to the limitations of the electron transfer downstream of PSI, accumulated electrons produce reactive oxygen species (ROS), which damage iron-sulfur clusters (i.e., F, F, and F) and cause PSI photoinhibition (Tiwari et al., 2016). The PSI photoinhibition critically and negatively impacts plant growth, because it requires several days or weeks to replace the damaged PSI (Kudoh and Sonoike, 2002; Zhang and Scheller, 2004; Zivcak et al., 2015). Furthermore, the damaged PSI simultaneously induces PSII photoinhibition by downregulating the LET (Somersalo and Krause, 1989; Zhang and Scheller, 2004; Zivcak et al., 2015). Thus, maintaining a high level of P700+ is essential in avoiding ROS generation and protecting PSI and PSII from photoinhibition. To date, several mechanisms involved in the P700 oxidation have been elucidated, including the downregulation of Cyt bf activity, chlororespiration, photorespiration, cyclic electron transfer (CET), and Mehler reaction (water-water cycle) (reviewed by Shimakawa and Miyake, 2018a).

The downregulation of Cyt bf activity is activated through the acidification of the thylakoid lumen (Nishio and Whitmarsh, 1993; reviewed by Tikhonov, 2013), and reduction of the PQ pool (Shaku et al., 2016; Shimakawa et al., 2018). Nigericin-induced inhibition of luminal acidification leads to photoinhibition of 69% PSI and 29% PSII after a 20-min high-intensity light treatment (Joliot and Johnson, 2011), indicating that Cyt bf activity downregulation critically protects the photosystems from photodamage. Moreover, PQ pool reduction inhibits the Q-cycle in the Cyt bf (Shaku et al., 2016; Shimakawa et al., 2018). Once the Cyt bf activity is downregulated, the electron flow from Cyt bf to PC and then to the PSI acceptor side is limited, consequently suppressing electron donation to P700+.

Chlororespiration is a light-independent O2 consumption process in chloroplasts, wherein O2 is reduced by electrons from PQ through the plastid terminal oxidase (PTOX) (Bennoun, 1982). Several studies on algae, Ostreococcus and Chlamydomonas elucidated the critical function of PTOX for photosynthetic regulation through P700 oxidation (Cardol et al., 2008; Houille-Vernes et al., 2011; Nawrocki et al., 2019a, 2019b, 2019c). However, the role of PTOX in land plants, including Arabidopsis thaliana, Nicotiana tabacum, and Solanum lycopersicum, is less clear, because PTOX over-expression causes increased light sensitivity (Ahmad et al., 2012; Heyno et al., 2009; Rosso et al., 2006). On another note, photorespiration, that is, the oxygenation of ribulose-1,5-biphosphate (RuBP) by RuBP carboxylase/oxygenase, plays a major role in maintaining P700 oxidation during a steady photosynthetic state (Cornic and Fresneau, 2002; Wada et al., 2018), because most PSI-derived electrons are transferred to FNR for NADPH formation for the CBB cycle and the photorespiratory pathway (Kannchen et al., 2020).

The CET is also critical for proper P700 oxidation. The CET was originally discovered by Arnon et al. as an electron transfer from the PSI acceptor side to the PQ pool through a putative Fd:PQ reductase (FQR) (Arnon et al., 1954). Together with the LET, the CET contributes to the pumping of H+ from the stroma to the lumen in chloroplasts through Cyt bf, to generate the proton gradient across the thylakoid membranes (ΔpH) that is required for ATP synthesis as well as proper downregulation of the Cyt bf activity (reviewed by Munekage and Shikanai, 2005; Yamori and Shikanai, 2016). Using genetic approaches, PGR5 (Munekage et al., 2002), PGRL1 (DalCorso et al., 2008), and the NAD(P)H dehydrogenase-like (NDH) complex (Burrows et al., 1998; Shikanai et al., 1998) were identified as the main components of the CET chains (reviewed by Shikanai, 2007). Specifically, these components have been proposed to control two independent CET pathways including the PGR5/PGRL1-dependent CET (PGR-CET) and the NDH-complex-dependent CET (NDH-CET) (reviewed by Yamori and Shikanai, 2016). Nevertheless, there is a growing body of evidence even for functional models of the NDH-CET and the PGR-CET. Specifically, PGR5 was recently suggested to control P700 oxidation through regulation of the LET, but not CET (Takagi and Miyake, 2018). PGR5 and PGRL1 cannot fulfill efficient FQR activity (Nandha et al., 2007; Nawrocki et al., 2019a, 2019b, 2019c); the PGR5/PGRL1 complex has been proposed as an indirect regulator of CET (Joliot and Johnson, 2011; Nawrocki et al., 2019a, 2019b, 2019c). In fact, PGR5 does not contain any motifs to ensure an electron-transfer capability (Munekage et al., 2002). Regarding the NDH-CET, its contribution is very small because of the low NDH-complex:PSI stoichiometry (1:100) in C3 plants; therefore, the NDH-CET may not support the ΔpH formation and ATP synthesis required for CO2 fixation (reviewed by Nawrocki et al., 2019a; 2019b; 2019c), although H+ pumping rate in the complex has not been directly measured. Thus, the specific mechanisms underlying the CET chains still remain unclear, although the importance of CET for oxidation of P700 has been revealed.

Using a reverse-genetics screening, we identified a previously uncharacterized chloroplast redox-active protein, named triplet-cysteine repeat protein (TCR). TCR is a redox active [3Fe-4S] iron-sulfur cluster-binding protein localized in chloroplasts, and tcr mutant showed similar photosynthetic activity with those of the NDH complex mutant ndho, suggesting that TCR functions as an electron carrier involved in an alternative electron flow around PSI, which in turn controls the oxidation of P700.

Results

Identification of TCR

We aimed to identify molecular components contributing to the precise regulation of photosynthesis in oxygenic phototrophs. Recently, with the development of rapid advanced technologies (e.g., microarrays and RNA sequencing), the gene expression status can be identified systematically in the interaction with other genes (van Dam et al., 2018). Genes showing similar expression profiles may be involved in similar biological processes; thus, co-expressed genes of known function(s) in a co-expressed gene network can provide insights into the functions of other genes with unknown function(s) (Eisen et al., 1998; Walker et al., 1999; Aoki et al., 2016). By use of the tools, we previously screened for A. thaliana genes that have the following properties: (1) co-expressed with known photosynthetic regulatory genes, (2) predicted to encode a chloroplast protein, and (3) conserved among oxygenic phototrophs (Mai Duy Luu et al., 2018; Sato et al., 2017). In addition to FLAP1 (Mai Duy Luu et al., 2018; Sato et al., 2017) and DLDG1 (Harada et al., 2019), the details of which were reported previously by us, we here found another previously uncharacterized Arabidopsis gene at the open reading frame At1g78995 in a gene cluster in the STRING data resources (Szklarczyk et al., 2015), which include NDH-complex genes such as NdhO, NdhL, NdhN, PnsB2, PnsB3, PnsB4, PnsB5, PnsL1, PnsL2, PnsL3, CRR3, and CRR7 (Figure 1A). This gene, abbreviated as TCR, encodes a small protein (159 amino acids) with an N-terminal predicted plastid-transit peptide, as assessed using the ChloroP server (Emanuelsson et al., 1999) (Figure 1B); the gene product was detected in the chloroplast fraction as examined by proteomics analysis of A. thaliana (Ferro et al., 2010). Based on the ATTED-II database (Aoki et al., 2016), TCR is co-expressed with NPQ1 (which encodes violaxanthin de-epoxidase), NDH-complex genes (PnsL3, PnsB3, CRR7, and CRR42), carbohydrate metabolic pathways (FBA5 and FBA7), and Fd-dependent electron transfer genes (FTRA2 and FNR2). The amino acid sequence alignment of TCR homologs (Figure 1B) and a phylogenetic tree of TCR (Figure S1) showed that TCR was specifically conserved in the photosynthetic green lineage, which includes green algae, liverworts, ferns, monocots, and dicots, but not found in red algae and cyanobacteria. Note that the amino acid sequence of TCR does not show any similarity to known proteins, such as Fds.

Figure 1

Identification of TCR by the reverse genetics screening

(A) TCR-related co-expression network. This network was built up by the STRING webserver (Szklarczyk et al., 2015).

(B) Deduced amino acid sequence alignment of the TCR homologs. The TCR homologs from A. thaliana (TCR), Zea mays (Z), Glycine max (G), Physcomitrella patens (P), Selaginella moellendorffii (S), and Chlamydomonas reinhardtii (C) were aligned. The red arrow indicates the predicted cleavage site of the chloroplast-transit peptide. The alignment was done by the T-Coffee webserver (Di Tommaso et al., 2011; Notredame et al., 2000).

See also Figure S1.

Among the TCR homologs, amino acid sequence similarities were higher at the C terminus versus the N terminus (Figure 1B). More strikingly, we identified an evolutionarily conserved triplet cysteine motif (CxxxCxxxC) that was repeated twice near the C terminus of the protein (Figure 1B). According to the structural prediction server Phyre2 (Mezulis et al., 2015), TCR secondary structure contains five helices, and the two helices contain the triplet cysteine motifs with high confidence (Figure S2A). The Phyre2 could not predict a reliable tertiary structure of TCR, because no similar protein to TCR was identified in the protein databank (PDB) library (Bairoch et al., 2005). Hence, ab initio modeling QUARK (Xu and Zhang, 2012, 2013) was utilized to build the predicted tertiary structure of TCR. Top five obtained models of TCR structure showed that two triplet cysteine motifs positioned in two distinct helices (Figure S2B), which is consistent with the data obtained by Phyre2 (Figure S2A), and a disulfide bond was formed between the two helices of the first obtained model of TCR structure (Figure S2B). The remaining four cysteine residues of this predicted model (Figure S2B) seemed to coordinate an iron-sulfur cluster, as shown in several iron-sulfur proteins designed de novo (Dizicheh et al., 2017; Nanda et al., 2016); however, in most cases, the iron-sulfur cluster(s) is bound in loop region(s) in natural proteins. These unique characteristics of TCR prompted us to proceed with the characterization of TCR function in relation to photosynthetic regulation.

TCR is an iron-sulfur protein

The N-terminal His-tagged TCR from A. thaliana was expressed in Escherichia coli, and purified by Ni-affinity resin under anaerobic conditions (Figure 2A). Under anaerobic conditions, purified TCR had a dark-brown color (Figure 2B), which was not clearly observed for TCR purified under ambient air (data not shown) indicating that TCR might be a ligand-binding protein and the ligand might be sensitive to ambient air (containing ∼21% O2). In fact, the loss of dark-brown color was accompanied by the aggregation and precipitation of purified protein. The absorption spectrum of TCR had a broad peak at 420 nm (Figure 2C), which is a typical feature of iron-sulfur-cluster-binding proteins (Freibert et al., 2018; Hoppe et al., 2011; Kennedy et al., 1984; Li et al., 2019; Nakamaru-Ogiso et al., 2002; Yabe et al., 2004). The 420 nm absorption (A) is attributed to ligand-to-metal (Cys-S/Fe2+) charge transfer in the iron-sulfur clusters (Hoppe et al., 2011; Lippard and Berg, 1994), suggesting that TCR binds an oxygen-sensitive iron-sulfur cluster. The absorption spectrum of oxidized TCR is similar to those of iron-sulfur proteins containing a [4Fe–4S] or [3Fe–4S] cluster (Hoppe et al., 2011; Ibrahim et al., 2020; Lippard and Berg, 1994; Nakamaru-Ogiso et al., 2002). Dithionite reduces TCR, which results in decreased A (Figure 2C, blue line) and diminished color intensity (Figure 2B). In contrast, A was modestly increased by the 1-min exposure to ambient air (Figure 2C, dashed red line), indicating that most of the purified TCR contained an oxidized iron-sulfur cluster and that the remaining reduced form can be oxidized by O2. Given that we expressed TCR in aerobically grown E. coli, the iron-sulfur cluster in the expressed TCR was oxidized in E. coli cells, although it is labile to ambient air after breaking the cells.

Figure 2

Biochemical properties of TCR

(A) SDS-PAGE profile of the purified TCR expressed in E. coli. The arrow indicates the His-tagged TCR band.

(B) Color change of the purified TCR by oxidation and reduction of a disulfide bond and a putative iron-sulfur cluster. The black arrows indicate specific positions showing color change. Dithio, dithionite.

(C) Absorption spectra of purified TCR in the presence and absence of dithionite. Spectra were acquired under anaerobic conditions, with the exception of that of TCR exposed to ambient air for 1 min.

(D) Absorption spectra of purified TCR in the presence and absence of dithionite and/or DTT. Spectra were acquired under anaerobic conditions.

(E) Second-derivative absorption spectra of the spectra shown in (D) and (E). Dithio, dithionite.

Next, we treated TCR with dithiothreitol (DTT) under anaerobic conditions. In contrast to dithionite that specifically reduced iron-sulfur clusters in proteins without reducing disulfide bonds (Yin et al., 2013), DTT can efficiently reduce disulfide bonds and iron-sulfur clusters having relatively high redox potentials (Daltrop et al., 2002). Strikingly, the visible TCR spectrum absorption was increased by DTT treatment (Figure 2D, green line), accompanied by the appearance of a new peak at 480 nm, as clearly observed in the second-derivative absorption spectrum (Figure 2E, green line), indicating that the iron-sulfur cluster in TCR could not be reduced by DTT (−0.33V), and its absorption property was altered, possibly, due to breakage of the disulfide bond near the cluster (Figure S2B). As expected, adding dithionite to the DTT-treated TCR resulted in decreased levels of both A and A due to the reduced iron-sulfur cluster (Figures 2D and 2E, yellow lines). Notably, DTT addition to the dithionite-reduced TCR induced an increased A, but not A (Figure 2E, dashed blue line), supporting the hypothesis that the increased A and the appearance of A upon DTT treatment (Figure 2D) may have resulted from TCR's conformational change through the breakage of a disulfide bond (Figure S2B). The A to A ratios of purified TCR with or without DTT were 0.35 and 0.22, which were a little smaller than those of other reconstituted [3Fe–4S] iron-sulfur proteins (0.3–0.4) (Ibrahim et al., 2020), suggesting that the purified sample contained a certain amount of apo-proteins.

To determine the structure of the iron-sulfur cluster in TCR, electron spin resonance (ESR) measurements were conducted with purified TCR. Air-oxidized TCR showed an ESR signal at g = 2.028, which is a typical feature of the electronic spin state (S = 1/2) of irons in oxidized [3Fe–4S] clusters (Figure 3A) (Ibrahim et al., 2020; Surerus et al., 1989; Telser et al., 2000). This signal must have originated from the metal center, because the ESR signal was not saturated at high microwave powers at cryogenic temperature (Figure 3B). In other words, the recorded ESR signal, herein, reflects the status of the electron spin of irons in the iron-sulfur cluster. Notably, the symmetrical ESR spectrum of TCR (Figure 3A) indicated the small g-anisotropy, which suggests that three Fe ions are located in a symmetrical coordinate. Given that a reduced [3Fe–4S] is ESR-silent (Telser et al., 2000), the weaker ESR signal observed in the dithionite-treated TCR sample (Figure 3A) indicated that dithionite cannot fully reduce TCR. In fact, the absorption spectrum of the dithionite-treated TCR sample still had a small absorption peak at 420 nm (Figure 2C). This raised a possibility that TCR has a very low redox-potential and cannot be completely reduced by dithionite.

Figure 3

ESR spectra of the TCR recombinant protein

(A) ESR spectra of the oxidized (red) and reduced (blue) TCR. Microwave frequency, 9.46 GHz. Microwave power, 100 mW. Field modulation, 4G. Temperature, 6K.

(B) The microwave power saturation the ESR signal of the oxidized TCR. Microwave frequency, 9.46 GHz. Field modulation, 4G. Temperature, 6K.

TCR localizes in chloroplast

To determine whether TCR localizes in chloroplasts, as predicted by ChloroP (Emanuelsson et al., 1999), we transiently expressed the recombinant green fluorescent protein (GFP) fused with a predicted signal peptide of TCR (64 N-terminal amino acids) in onion epidermal cells. The recombinant GFP was localized in plastids, as observed for the plastid-localizing control CYO1-GFP (Shimada et al., 2007), but not for the negative control (GFP alone) (Figure 4A), indicating that the N terminus of TCR functions as the plastid-signaling peptide. We also expressed GFP fused with the full-length TCR in A. thaliana using the 35S promoter of the cauliflower mosaic virus and found that GFP fluorescence overlapped with chlorophyll autofluorescence in mesophyll chloroplasts (Figure 4B), indicating that mature TCR localizes in chloroplasts.

Figure 4

Subcellular localization of TCR

(A) GFP fluorescence in onion epidermal cells transiently expressing GFP, CYO1-GFP, and TCR(1–64 a.a.)-GFP. Scale bars, 30 μm.

(B) GFP and chlorophyll fluorescence from transgenic A. thaliana expressing 35Spro::TCR-GFP. WT (Col 0) was used as a negative control. Scale bars, 10 μm.

(C) Western blotting using anti-GFP, anti-PSBO, and anti-D1 antibodies. T, total chloroplast; M, thylakoid membrane fraction after four washes with 10 mM Na-pyrophosphate-NaOH, pH 7.8; S, soluble-protein fractions (collected in the supernatant at every washing step [first, second, third, and fourth]).

(D) The thylakoid membrane fraction shown in (C) was used as the original fraction (Origin), which was treated with different salts and applied to western blotting. Subsequently, the PVDF membranes were stained with Coomassie blue (right). RbcL, Rubisco large subunit.

Next, we characterized the localization of TCR at the sub-organelle level. Unfortunately, anti-TCR antibodies used for immunoblotting cannot be produced, although six trials to produce anti-TCR were achieved with short peptides and full-length TCR. Alternatively, we used the TCR-GFP-expressing line to characterize the sub-cellular localization. We isolated total chloroplasts from the transgenic A. thaliana plants expressing TCR-GFP described above and separated them into the thylakoid membrane and soluble fractions. To obtain the membrane fractions, thylakoid membranes were washed with a low-osmotic buffer four times, followed by the confirmation that most soluble proteins, including the Rubisco large subunit (RbcL), were washed out (Figure 4C). Notably, the thylakoid membranes were broken in the low-osmotic washing buffer, so that the soluble proteins in the lumen were also washed out. Although we detected TCR-GFP mostly in soluble fractions, a small amount of TCR-GFP was detected in membrane fractions, even after the four washing steps, suggesting that TCR localizes in both thylakoid membranes and soluble fractions, as observed for the PSII-associated protein PSBO (Figure 4C). The washed thylakoid membranes were then treated with different salts and analyzed using western blotting. We used chaotropic salt solutions containing 2 M NaBr or 2 M NaSCN for disrupting the hydrogen bonding network between water molecules and weakening the hydrophobic effect within and/or between protein molecules. Alkaline solution (0.1 M Na2CO3) will affect ionic bonds within and/or between protein molecules and the stability of lipid bilayer membranes. Signal intensity of membrane-bound TCR-GFP was increased after NaBr, NaSCN, or Na2CO3 treatment perhaps due to the reduction of overlapped proteins on the polyvinylidene difluoride (PVDF) membranes. The relative amount of TCR-GFP in soluble fractions, compared with those in membrane fractions, was clearly increased after the salt treatment (Figure 4D). A similar phenomenon was observed for PSBO. The thylakoid-membrane-spanning PSII subunit D1 was stably maintained in the thylakoid membrane fractions. Taken together, these data demonstrated that TCR is soluble, some of which associate with thylakoid membranes, probably, via an interaction with an unknown membrane protein or complexes; this finding is consistent with a previous proteomics analysis of A. thaliana chloroplasts that showed that TCR was present in the isolated thylakoid membranes (Ferro et al., 2010). High accumulation of TCR-GFP in soluble fractions may have resulted from higher TCR-GFP expression than that of the endogenous TCR, resulting in TCR binding site saturation on thylakoid membranes.

Green tissue-specific expression of TCR

To investigate the location and timing of TCR expression in A. thaliana, we performed a histochemical analysis using a TCR promoter-fused GUS expression construct. The putative promoter region (2,000 bp upstream of the start codon) of TCR drove GUS expression at very early stages of plant development. Specifically, intensive GUS activity was observed in cotyledons, mature leaves, and hypocotyls at the seedling stage (from day 1–3; Figures 5A and 5B). These high GUS activity patterns were observed over 17 days on MS agar plates (Figures 5A–5E). GUS expression was observed at the tips of emerging leaves (Figures 5C and 5E). Notably, the GUS expression patterns were not uniform (Figure 5C), as higher activity was detected in mesophyll cells near stomata and veins of the leaves of 5-day-old cotyledon (Figure 5G). The 17-day-old transgenic plants were transferred to soil and further grown until flowering and producing fruits. In reproductive tissues, GUS expression was observed only in green regions, such as sepals, green seeds, and fruit coats (29-day-old plant; Figure 5F), indicating that the TCR promoter is active during seed development after pollen fertilization. Moreover, the wounding of transgenic plants led to the observation of GUS expression specifically at the damaged areas, especially in the mesophyll cells and root hairs that surrounded the wounded areas (Figure S3). These data indicate that TCR is expressed mainly in green tissues at very early stages of plant growth and at the seed-development stage, and that wounding induced the local expression of TCR, even in the roots.

Figure 5

Histochemical analysis of TCR-promoter-GUS expression

(A) Germinating seeds after 1-day illumination.

(B) 2-day seedling.

(C) 5-day seedling.

(D) 7-day seedling.

(E) 17-day seedling.

(F) Flowers and fruits from a 29-day transgenic plant.

(G) Leaf tissue of 5-day cotyledons. The yellow circles indicate the stomatal position on epidermal layers. Observations were focused on the epidermal layer, to determine the position of stomata; subsequently, the focus shifted to the mesophyll layer, to the detect mesophyll cells that surrounded stomata, which showed strong GUS expression. Scale bars, 0.1 mm (1 day), 1 mm (3, 5, and 29 days), and 100 μm in (G) (white) and 10 μm in (G) (yellow).

See also Figure S3.

We found that the putative TCR promoter contained binding sites of many transcription factors, such as Dof2, Dof3, HMG-1, MNB1A, PBF, ARR10, SOC1, SEP3, and MYC3, as revealed through the CircadiaNET web server (http://viridiplantae.ibvf.csic.es/circadiaNet/). MYC3, a member of basic-helix-loop-helix (bHLH) transcription factor family, regulates genes involved in tryptophan, jasmonic acid, and other stress responses (Cheng et al., 2011; Fernández-Calvo et al., 2011; Schweizer et al., 2013; Smolen et al., 2002). SOC1 and SEP3 regulate genes involved in the flowering time control (Lee et al., 2008; Liu et al., 2008) and floral organogenesis (Honma and Goto, 2001; Pelaz et al., 2000). These results suggested that TCR expression is controlled by different environmental changes through several transcription factors.

tcr mutant is sensitive to light stress conditions

Because T-DNA insertional knockout lines of TCR are not available, we knocked out TCR using the CRISPR/Cas9 system. The diagram provided in Figure S4A depicts the experimental process. We isolated one frameshift mutant that carried a large deletion (83 amino acid) and replacement (23 amino acid) in the TCR-coding region (Figure S4B). The frameshift mutant caused 52% loss of amino acid residues of TCR peptide with the remaining C-terminal (without counting the predicted chloroplast transit peptide) containing only 21 alternative amino acid residues, suggesting that TCR is completely knocked out in the tcr mutant. After backcrossing with wild-type (WT), a homozygous line was used as a knockout mutant of tcr.

No significant phenotypic differences were observed between WT and tcr plants regarding growth (Figure 6A), pigment content (Figure 6B), chloroplast ultrastructure (Figure 6C), and protein levels of some photosynthetic complexes (Figure 6D and 6E). The protein levels of other photosynthetic complexes were also checked by mass-spectrometry-based quantitation. The mass spectrometric data show that there are no differences in the photosynthetic complexes in WT and tcr lines (Table S1).

Figure 6

Characterization of the phenotype of the tcr mutant

(A) Shoot phenotype of plants grown under short-day conditions for 70 days. Scale bars, 1 cm.

(B) Pigment content of the plants shown in (A).

(C) Ultrastructure of chloroplasts, as observed using transmission electron microscopy. Leaf samples were collected from 19-day-old plants. Yellow scale bars, 2 μm (WT) and 1 μm (tcr); blue scale bars, 500 nm (WT) and 200 nm (tcr).

(D) Analysis of photosynthetic protein complexes via large-pore Blue Native PAGE.

(E) Western blotting using specific antibodies. Loading: 1, 3.1 μg/well; 2, 1.55 μg/well.

(F) Maximum quantum yield at PSII before and after 8 h of high-light stress. Detached leaves from 50-day-old plants were used. All plants were grown under short-day conditions. Data are the mean ± SD (n = 4). Statistically significant differences were analyzed as shown in the figure for the different types of plants using the two-tailed Student's t test with a 95% confidence interval (∗∗p < 0.01, ∗∗∗p < 0.001).

(G) Maximum P700 oxidation at PSI before and after 8 h of high-light stress. Detached leaves from 50-day-old plants were used. All plants were grown under short-day conditions. Data are the mean ± SD (n = 4).

(H) Plants were grown under continuous strong light (1,200 μmol photons m−2 s−1) from germination until reproduction (flowering). Scale bar, 1 cm.

(I) Phenotypic observation of plants grown under fluctuating light conditions for 4 weeks with or without FR illumination (λmax = 740 nm). The fluctuating light (FL) cycle was high light (HL, 700 μmol photons m−2 s−1, red + blue LEDs) for 2 min, followed by low light (LL, 30 μmol photons m−2 s−1, fluorescent lamps) for an additional 2 min. Plants were treated with FL during the light phase of the 8 h/16 h light/dark cycle. Far red (FR) light was induced with an intensity of 50 μmol photons m−2 s−1 during the HL phase and an intensity of 10 μmol photons m−2 s−1 during the LL phase of FL. FR source: IR LED Stick (NAMOTO). Scale bar, 1 cm.

(J) Shoot phenotype of WT, pgr5, crr4-2pgr5, and pgr5tcr plants. Scale bar, 1 cm.

(K) Shoot phenotypes of pgr5 and pgr5tcr plants. Plants were grown under short-day conditions for 19 days on 1/2 MS agar. Scale bar, 1 cm.

(L) The mean shoot mass of pgr5 and pgr5tcr plants grown under the conditions indicated for (K). Data are the mean ± SD (n = 5, 10–16 plants per measurement). Statistically significant differences were analyzed as shown in the figure for the different types of plants using the two-tailed Student's t test with a 95% confidence interval (∗∗∗p < 0.001). F.W., fresh weight.

See also Figure S5.

To obtain additional insights into the physiological importance of the TCR, the photosynthetic activities of tcr plants were analyzed under different light stress conditions. We first characterized the high-light sensitivity of the tcr mutant. Detached leaves were exposed to intense light (1,500 μmol photons m–2 s–1) for 8 h, and PSII photoinhibition was analyzed. A more pronounced photoinhibition was observed in tcr versus WT plants, and the maximum quantum yield of PSII (F/F) was lower in tcr versus WT plants (Figure 6F). This phenotype was complemented partly in a complementing line Compl-1 and was complemented fully in another complementing line Compl-2 (Figure 6F). Notably, real-time PCR analysis with mRNA isolated from whole shoot tissues indicated that TCR mRNA levels in Compl-1 and Compl-2 were similar to and higher than those in WT, respectively (Figure S4C). Given that the TCR expression in leaves is not uniform (Figure 5C), TCR expression in Compl-1, driven by the constitutive 35S promoter, may be lower than that in WT at specific areas, which may be a reason why Compl-1 could not fully complement the tcr phenotype. After the high-light treatment, no significant change in Pm values was observed in tcr compared with WT plants (Figure 6G), suggesting that the tcr mutation does not cause photoinhibition at PSI upon the high-light treatment. We also grew plants under high-light conditions (1,200 μmol photons m–2 s–1) from germination to the reproductive stage. As shown in Figure 6H, all plants accumulated anthocyanins during the vegetative growth stage; however, when the plants grew to the reproductive stage, plant growth of tcr lines was reduced relative to WT and Compl-2 plants, indicating the importance of TCR function for the high-light stress response.

Because fluctuating light potentially induces PSI photoinhibition because of the over-reduction of P700 (Shimakawa and Miyake, 2018b), next we checked the phenotype of tcr and WT plants under long-term (4 weeks) fluctuating-light treatment. In this condition, both WT and tcr leaves accumulated a large amount of anthocyanin (Figure 6I, left), suggesting a strong ROS accumulation caused by the over-reduction of P700 (Xu et al., 2017; reviewed by Li et al., 2009). However, continuous far-red (FR) light (λmax = 740 nm) illumination, which specifically oxidizes P700 (Kono et al., 2017), during the fluctuating-light treatment rescued the phenotype of the photo-oxidative stress in both WT and tcr plants, and anthocyanin did not accumulate in leaves (Figure 6I, middle), indicating the importance of FR light for the reduction of ROS generation and PSI photoinhibition (Kono et al., 2017), and, consequently, PSII photoprotection. Interestingly, when plants were grown under fluctuating light without FR for 2 weeks, followed by FR illumination for an additional 2 weeks, tcr leaves accumulated higher anthocyanin than did WT leaves (Figure 6I, right). It is possible that FR illumination during the last 2 weeks of the treatment rescued PSI from photoinhibition to a greater extent in WT versus tcr plants, suggesting the importance of TCR for P700 oxidation and PSI photoprotection.

To further characterize the tcr mutant, we crossed the pgr5 and tcr lines to produce the pgr5tcr double mutant and proceeded to characterize its phenotype under ambient air. Under short-day conditions, the pgr5tcr double mutant showed delayed growth on soil (Figures 6J and S5A) and MS agar medium (Figure 6K) compared with pgr5 plants; the average shoot weight of pgr5tcr was ∼50% that of the pgr5 line under short-day (Figures 6L and S5B) and long-day (Figures S5C and S5D) conditions. Also, growth retardation could be observed in crr4-2pgr5 double mutants, which is stronger than those observed in pgr5tcr double mutants (Figure 6J). Notably, crr4-2pgr5 was reported as a typical mutant in which both NDH-CET and PGR-CET are inhibited due to the lack of NDH-complex and PGR5 protein (Munekage et al., 2004; Yamamoto et al., 2011).

tcr mutation affects the redox states of the PQ pool

Next, we investigated the effects of the tcr mutation on the redox states of the PQ pool. When actinic-light illumination was turned off, chlorophyll fluorescence was transiently increased in the dark (Figures 7A and 7B). This phenomenon, called the post-illumination chlorophyll fluorescence transient (PIFT), is caused by the reduction of the PQ pool by the Fd-dependent CET from PSI (Gotoh et al., 2010). In fact, the specific excitation of the PSI by FR illumination resulted in a decrease in the fluorescence (Figures 7A and 7B), which supported the notion presented above. Although no significant PIFT changes were detected between WT and tcr plants (Figures 7A and 7B), we observed a significant decrease in the PIFT signal in pgr5tcr compared with pgr5 specimens when plants were shifted to the dark after 2 and 5 min of actinic-light illumination (Figures 7C and 7D, respectively). Additionally, full trace of the chlorophyll fluorescence (Figure S6) showed that the second rise of fluorescence after actinic light was turned on in pgr5tcr occurred faster than in pgr5, indicating that the PQ pool in pgr5tcr was reduced faster than in pgr5. These results implied that TCR accepted electrons from PSI, and the tcr mutant decreased the stromal electron sink during illumination. Consequently, TCR contributed to PQ pool(s) oxidation and reduction at least under post-illumination conditions, the importance of which is increased when the PGR5-dependent CET and/or LET are impaired.

Figure 7

Post-illumination chlorophyll fluorescence transient (PIFT) analysis of WT and mutant plants

(A–D) PIFT analysis of WT and mutant plants after AL pre-illumination for 2 min (A and C) or 5 min (B and D). The fluorescence was normalized to the point at which AL was turned to OFF. ↑, Light ON; ↓, light OFF; ML, measuring light; AL, red actinic light; FR, far-red light; ST, saturating pulse. All plants were grown under short-day conditions for 49–56 days on soil.

See also Figure S6.

Effect of the tcr mutation on P700 oxidation

To investigate further the possibility that TCR is involved in electron flows from PSI, next we characterized the P700 oxidation/reduction kinetics upon actinic-light illumination. As expected, the relaxation of Y(NA) and induction of Y(ND) were delayed in tcr with WT plants at the induction stage (<2 min) upon turning on the actinic light (Figures 8A and 8B), indicating that electron acceptance from PSI is limited and that the oxidation of P700 is lowered in the tcr line. In steady-state conditions of actinic-light illumination (>2 min), Y(I) was slightly lower in tcr than it was in WT plants (Figure 8C), indicating that the electron flow from PSI to Fd and/or PC to PSI is limited in tcr specimens, which results in lowered PSI activity. Notably, the tcr phenotypes were partially and fully complemented in the complement lines Compl-1 (Figure S7) and Compl-2 (Figures 8D, 8E, and 8F), respectively.

Figure 8

PSI activities and oxidation/reduction kinetics of P700

(A and D) Measurement of the acceptor side limitation at PSI. ∗p < 0.05, ∗∗∗p < 0.001 (one-way ANOVA followed by Tukey's test with a 95% CI). Plants were grown on soil under short-day conditions for 56 days. AL = 150 μmol photons m−2 s−1, P700 ML = 10, RT = 20°C. Dark adaptation time = 30 min. Data represent the mean ± SD (n = 3).

(B and E) Measurement of the donor side limitation at PSI. ∗∗∗p < 0.001 (one-way ANOVA followed by Tukey's test with a 95% CI). Plants were grown on soil under short-day conditions for 56 days. AL = 150 μmol photons m−2 s−1, P700 ML = 10, RT = 20°C. Dark adaptation time = 30 min. Data represent the mean ± SD (n = 3).

(C and F) Measurement of the PSI quantum yield. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 (one-way ANOVA followed by Tukey's test with a 95% CI). Plants were grown on soil under short-day conditions for 56 days. AL = 150 μmol photons m−2 s−1, P700 ML = 10, RT = 20°C. Dark adaptation time = 30 min. Data represent the mean ± SD (n = 3).

(G and H) Measurement of P700 in the presence of a dominant LET or a dominant CET, respectively, using a Joliot-type spectrometer. The asterisk indicates an unusual signal that was not observed in WT plants. ↑, Light ON; ↓, light OFF; FR, far-red light. All plants were grown on soil under short-day conditions for 49–56 days.

(I and J) Measurement of P700 after the adaptation of plants to the dark and light, respectively, using a DUAL-PAM-100 system. The FR was kept ON during the measurement as shown in the panel (L). The P700 signals in (K) were double-normalized using the lowest (P) and the highest (P′) values. Dashed line, P. The P700 signals of WT and tcr or pgr5 and pgr5tcr plants in (L) could not be double-normalized using the point at which AL was turned to OFF or to the lowest value of P700. Bars, regions showing differences between WT and mutant plants. ↑, Light ON; ↓, light OFF; AL, red actinic light; FR, far-red light. Plants were grown on soil under short-day conditions for 49–56 days.

See also Figure S7.

We also investigated the induction kinetics of P700 oxidation upon FR light illumination using a Joliot-type spectrophotometer. We characterized the P700 oxidation/reduction kinetics of WT and tcr plants that were pre-treated with two distinct light conditions; one was used to monitor specifically LET in which electron transfer from PSI to NADP+ would be dominant (plants were pre-illuminated with red light for 10 min before measurements), whereas the other treatment was used to monitor CET in which electrons from PSI to Cyt bf would be dominant (plants were kept in dark for 10 min, followed by irradiation with a saturating pulse for 100 ms before measurements) (Ghysels et al., 2013; Joliot and Joliot, 2006; Talts et al., 2007). When plants were pre-treated to monitor LET, the WT and tcr lines showed FR-induced P700 oxidation without any intermediates (Figure 8G); in contrast, when WT plants were pre-treated to monitor CET, two-step oxidation of P700 was observed, i.e., FR illumination resulted in rapid P700 oxidation, followed by re-reduction of P700 2 s after the FR illumination and re-oxidization of P700, as reported previously (Figure 8H, black line) (Ghysels et al., 2013; Nagarajan et al., 2014; Talts et al., 2007). Conversely, we observed an abnormal three-step P700 oxidation in tcr when the CET was dominant; a small shoulder was additionally observed at 5 s after turning on the FR illumination (Figure 8H, red line; asterisk). This suggests that TCR function is related to the P700 oxidation process, especially when the CET is dominant, although the exact reason for the abnormal three-step P700 oxidation kinetics observed in tcr plants is unknown.

We also monitored the oxidation kinetics of P700 at a relatively faster range using the DUAL-PAM-100 system. When dark-adapted plants were illuminated with FR light, the P700 oxidation kinetics in tcr, pgr5, and pgr5tcr plants was faster than that detected in the WT counterparts (Figure 8I, red, green, and blue lines, respectively; marked bar), suggesting that the stromal electron sink in the dark is smaller in the mutants versus WT plants, as reported previously for the pgr5 mutant (Okegawa et al., 2007). We also monitored the P700 reduction/oxidation kinetics upon turning off the actinic light under FR-light illumination. Upon turning off the actinic light, sudden P700 reduction was observed in WT and tcr plants, which was significantly diminished in pgr5 and pgr5tcr plants (Figure 8J). This implies that, in pgr5 background, P700 is already highly reduced compared with the WT background under actinic-light illumination conditions, because of the inhibition of PGR5-dependent CET and/or LET. Two seconds after turning off the actinic light, P700 was gradually re-oxidized by FR light in WT plants (Figure 8J; marked bar), which was slowed down in tcr lines, suggesting that P700 oxidation is reduced in the latter via limited electron acceptance from PSI. The decreased P700 oxidation kinetics observed upon tcr mutation was not observed in the pgr5 background (Figure 8J). Taken together, these results suggest that TCR plays a role in the maintenance of the P700 oxidation/reduction status, which seems to be important when the stromal electron sink is limited.

We next tested P700+ reduction kinetics by the PC-dependent electron transfer. Specifically, leaves were illuminated by FR light for 30 s, and then P700+ was completely oxidized to the greatest level (Pm) by supplying saturating light for 3 ms. After turning off the saturating light, P700+ signal was decreased, which reflected the P700+ reduction by PC (Figure S8). No difference was observed between the recorded P700+ kinetics of tcr, Compl-2, and WT plants (Figures S8A and S8B), indicating that electron transfer from Cyt bf to the PSI donor side is normal in tcr mutant. This further suggests that alteration of P700 oxidation/reduction status caused by the tcr mutation (Figures 8I and 8J) may result from limitation of the electron flow downstream of the acceptor side of PSI.

Given Fd is the primary electron acceptor of PSI, TCR may receive electrons from Fd. To test this possibility, we performed in vitro reduction of TCR by Fd (Figure 9). In the presence of a large amount of NADPH, electrons were transferred to FNR and Fd, which then potentially reduce TCR (Figure 9A). It is noteworthy that, in the reaction mixture, the concentration of Fd and FNR was ∼400-fold lower than that of TCR. Upon addition of a large amount of NADPH (1 mM), decrease in A was observed (Figure 9B), indicating the electron transfer from NADPH to TCR via FNR and Fd. TCR reduction could be observed even without Fd (Figure 9C), although the TCR reduction rate without Fd was slower than that observed with Fd (Figure 9D). We confirmed that the TCR reduction was not observed without the Fd system for 4 h (Figure 9E, dashed lines). These results indicated that Fd catalyzes TCR reduction. In turn, A was increased 1 min after exposure to the ambient air of the reaction solution that had been incubated with NADPH for 4 h (Figure 9F, solid blue line) and was fully recovered after 3 min of exposure (Figure 9F, solid purple line). These observations confirmed that the decreased A reflected TCR reduction. These data demonstrated that TCR is redox active in vitro.

Figure 9

Fd-dependent reduction of TCR

(A) Schematic representation of the experiment performed to reduce TCR by NADPH, FNR, and Fd in vitro.

(B) Time-lapse observation of the reduction of TCR in vitro by NADPH, FNR, and Fd.

(C) Time-lapse observation of the reduction of TCR in vitro by NADPH and FNR.

(D) Time dependency of absorbance change at 420 nm (A420nm) due to the reduction of iron-sulfur center in TCR in (B) and (C).

(E) Absorption change in the purified TCR 5 min and 4 h after the addition of buffer only (dashed lines) or NADPH, FNR, and Fd (solid lines, + Fd system) under anaerobic conditions.

(F) Absorption change of the reduced TCR (by the Fd system) after exposure to ambient air for 1 and 3 min

To assess further the possibility that TCR is involved in the Fd-dependent electron transfer from PSI, we performed an in vitro Fd-dependent PQ pool reduction assay (Munekage et al., 2002; Okegawa et al., 2008). In the absence of antimycin A (AA), thylakoid membranes isolated from WT, tcr, and NDH-complex-mutant crr4-2 lines showed fluorescent increments upon Fd addition, indicating the presence of proper Fd-dependent PQ reduction in the dark (Figure S9). In the presence of AA, although chlorophyll fluorescence was inhibited in all samples, the WT and tcr lines exhibited a similar fluorescence increment, and only the crr4-2 line showed significant fluorescence impairment (Figure S9). This suggests that TCR is not involved in the main CET flows, although this experiment was recently questioned in that the fluorescent increment may not reflect Fd-dependent PQ reduction (see Discussion for more details) (Fisher and Kramer, 2014).

tcr shares similar photosynthetic parameters with ndho under fluctuating light

We next monitored the photosynthetic activities of PSI and PSII under fluctuating-light conditions in WT, tcr, pgr5, pgr5tcr, and NDH-complex null-mutant ndho plants. All mutants exhibited altered PSI activity compared with the WT (Figure 10). Specifically, the tcr mutant showed higher Y(NA) than did WT plants from the second-round shift from high light to low light (Figure 10B). This phenotype was also observed (even more significantly) in pgr5tcr versus pgr5 plants upon the shift from high light to low light (Figure 10B). Moreover, slight changes in Y(ND) levels were observed in tcr compared with WT plants, especially when the light intensity was changed from low light to high light (Figure 10C). An increase in Y(NA) accompanied the decrease in PSI quantum yield, Y(I), in tcr compared with WT (Figure 10D). The ndho mutant exhibited similar phenotypes to those observed in the tcr mutant under fluctuating light conditions, suggesting that the physiological function and impact of the TCR-dependent electron transfer are similar to those of NDH-CET. Notably, the ndho phenotypes were similar to those reported previously (Nikkanen et al., 2018; Shimakawa and Miyake, 2018b; Suorsa et al., 2012).

Figure 10

Photosynthetic activity under fluctuating light conditions

(A) PSII quantum yield.

(B) Electron limitation at the acceptor side of PSI.

(C) Electron limitation at the donor side of PSI.

(D) PSI quantum yield. Plants were grown on soil under short-day conditions for 50–57 days. Data represent the mean ± SD (n = 3). LL, 33 μmol photons m−2 s−1; HL, 920 μmol photons m−2 s−1; dark acclimation time = 30 min; RT = 20°C.

Discussion

TCR is a plastidial iron-sulfur protein

TCR is a low-copy gene specifically found in oxygenic phototrophs and its expression level is strictly controlled in A. thaliana (Dunand et al., 2011), suggesting its importance for photosynthesis regulation. Similar to that reported for PSBO, TCR-GFP was detected in both thylakoid membranes and soluble fractions (Figures 4C and 4D), suggesting that TCR functionally interacts with thylakoid-membrane-anchored protein(s), although its exact target(s) remain unclear. We determined that TCR contained an iron-sulfur cluster based on the observation that it exhibited absorption at around 420 nm (Figure 2C), with an ESR signal at g = 2.028 (Figure 3A), both of which are typical features of 3Fe–4S cluster proteins (Hoppe et al., 2011; Lippard and Berg, 1994; Nakamaru-Ogiso et al., 2002). Furthermore, TCR may contain a redox-active disulfide bond (Figure S2B), and its reduction induced alterations in the redox potential and/or structure of the protein (Figures 2D and 2E). These results established that TCR is a plastidial iron-sulfur protein involved in photosynthetic electron transfer.

TCR expression patterns, as examined by the promoter-GUS fusion analysis (Figure 5), were similar to those of other photosynthetic regulatory genes, such as Pisum sativum thioredoxin f1 (PsTRXf1), thioredoxin m1 (PsTRXm1), cucumber VDE (CsVDE), and A. thaliana PGR5 (AtPGR5) (Barajas-López et al., 2007; Li et al, 2013, 2015; Long et al., 2008). Specifically, all these genes are highly expressed in cotyledons, mature leaves, and other green tissues, supporting the hypothesis that TCR is involved in photosynthetic control. Furthermore, as shown for TCR (Figures 5 and S3), AtPGR5 (Long et al., 2008) and CsVDE (Li et al., 2015) expression patterns were variable, and PsTRXf1, PsTRXm1, CsVDE, and AtPGR5 were expressed in non-photosynthetic organs, such as flower sepals (PsTRXf1, PsTRXm1, CsVDE, and AtPGR5), green seeds (PsTRXm1), fruit coats (CsVDE, PsTRXf1, and PsTRXm1), hypocotyls (CsVDE, PsTRXf1, and AtPGR5), vasculatures, and stems (CsVDE and AtPGR5) (Barajas-López et al., 2007; Li et al, 2013, 2015; Long et al., 2008). Interestingly, AtPGR5 expression was induced upon wounding (Long et al., 2008), similar to TCR (Figure S3). Aside from the photosynthetic control discussed in this study, these results suggested TCR's additional functions related to germination, plant reproduction, and abiotic stress responses, as discussed previously for TRXf, TRXm (Barajas-López et al., 2007), VDE (Li et al., 2015), and PGR5 (Long et al., 2008).

TCR controls electron transfer around PSI

TCR is redox active in vitro (Figure 9), and mutational loss of TCR affected the P700 oxidation kinetics (Figure 8) as well as the redox status of PQ pools (Figure 7), suggesting that TCR controls the electron transfer around the PSI acceptor side for P700 oxidation. Because the purified TCR was oxidized upon exposure to ambient air in vitro (Figure 2C), and had a relatively low redox potential (Figures 2C and 3A), TCR may also contribute to the water-water cycle by reducing oxygen with electrons derived from Fd/PSI.

CO2 assimilation is limited when the CBB cycle cannot be fully activated. This is observed in nature when plants suffer drought-stress-induced stomatal closure, cold-stress-induced metabolic changes, high-light-stress-induced photorespiration, and nutrient deprivation (reviewed by Demmig-Adams and Adams, 1992; Asada, 1999). Short-term limitation of CO2 assimilation is observed when dark-adapted plants are exposed to light (Ptushenko et al., 2019). When CO2 assimilation is low, CET and other alternative electron transfer systems are enhanced, to transfer electrons from PSI to the PQ pool and/or from PSI to O2 through photorespiration and the water-water cycle. In fact, when plants had just shifted from dark to light conditions, we observed significant differences between WT and tcr plants at photosynthetic parameters Y(I), Y(NA), Y(ND) (Figures 8A, 8B, and 8C), and P700+ (Figures 8H and 8J). Also, tcr mutant showed clear phenotypes only when plants were treated with high-light and/or fluctuating-light stresses (Figures 6F, 6H, and 6I). As mentioned above, under such conditions, CO2 assimilation was inhibited. Therefore, we propose that TCR function is especially required when CO2 assimilation is limited and photorespiration is not fully activated, which supports P700 oxidation to avoid PSI photoinhibition, as the excess electrons accumulated in PSI lead to ROS generation (Tiwari et al., 2016; reviewed by Shimakawa and Miyake, 2018b; Sonoike, 2011). In fact, because the P700 oxidation system receives contributions from various mechanisms (reviewed by Sonoike, 2011; Shimakawa and Miyake, 2018b), PSI photoinhibition is rarely observed, even under natural strong-light illumination (2,000–3,000 μmol photons m−2 s−1) (Shimakawa and Miyake, 2018b); hence, PSI photoinhibition in tcr was not different from that in WT under strong-light conditions (1,500 μmol photons m−2 s−1, Figure 6G).

TCR-dependent electron transfer affects CET activity

We observed many similarities between the tcr and NDH-CET mutants (ndho, crr2-1, and/or crr4-2). First, neither tcr nor ndho plants exhibited any visible phenotype under controlled conditions (Figures 6A–6C) (Suorsa et al., 2012). Second, most of the photosynthetic parameters for PSII and PSI activities were similar between tcr and ndho plants (Figure 10). Third, both the pgr5tcr and crr4-2pgr5 mutants showed growth retardation compared with the pgr5 mutant (Figures 6J–6L and S5) (Munekage et al., 2004; Yamamoto et al., 2011). Fourth, high-light treatment on either tcr or NDH-CET mutants could cause PSII photoinhibition (Figure 6F) (Endo et al., 1999). Finally, the expression profiles of TCR and NDH-complex genes, such as PnsL2, PnsL3, CRR7, NdhN, and NdhL, were similar (Figure 1A) (Aoki et al., 2016; Szklarczyk et al., 2015). These observations suggest that TCR-dependent electron transfer affects CET in A. thaliana. In fact, the PIFT signal, which was abolished in NDH mutants (Kamruzzaman Munshi et al., 2005; Takabayashi et al., 2009), was normal in the tcr mutant (Figures 7A and 7B), suggesting that TCR-dependent electron transfer is independent of NDH-CET.

Based on the structural characterization of photosynthetic protein complexes, an alternative pathway for CET, other than PGR-CET and NDH-CET, has been proposed in which Fd transfers electrons to PQ through a putative protein super-complex embedded in the thylakoid membranes (Johnson, 2011; Nawrocki et al., 2019a, 2019b, 2019c). The putative CET super-complex (FQR complex) may include PSI, Cyt bf, FNR, PGRL1, PGR5, ANR1, and CAS; thus the FQR-complex-dependent CET is independent from the NDH-CET, but may partially overlap with the PGR-CET (Nawrocki et al., 2019a, 2019b, 2019c). Although experimental evidence is still lacking, the existence of the FQR complex was supported by the isolation of the Cyt bf-PSI super-complex (Steinbeck et al., 2018; Yadav et al., 2017), the Cyt bf-FNR complex (Clark et al., 1984; Zhang et al., 2001), and the CAS-ANR1-PGRL1 complex (Terashima et al., 2012), as well as by physical interactions between PGRL1 and PSI, Fd, Cyt bf, FNR, and PGR5 (DalCorso et al., 2008). In fact, Fd and FNR are encoded by multiple gene copies in the genome, and FNR iso-proteins have various abilities to interact with thylakoid membrane complexes (Goss and Hanke, 2014). Given that Fd and FNR were also involved in LET, and TCR has an affinity with both Fd and FNR to receive electrons (Figures 9B and 9C), TCR may assist to fine-tune electron divergence between the LET and the CET, which are specifically controlled by cell types (e.g., bundle sheath and mesophyll cells) and/or specific physiologic conditions. Given TCR has redox-active cysteine residues (Figures 2D and S2B), oxidation and reduction of the disulfide bond may have a role in the TCR function. Electrons derived from Fd are transferred to the thioredoxin system that affects CBB cycle activation as well as CET (Nikkanen et al., 2018; Okegawa and Motohashi, 2020), suggesting that TCR also functionally interacts with the thioredoxin system, through disulfide-bond exchange, to control electron transfer downstream of PSI.

The isolated thylakoid membranes of tcr plants showed a WT-like Fd-dependent fluorescent increment both in the presence and absence of AA (Figure S9), suggesting that TCR is not involved in the Fd-dependent CET. However, this experiment was recently questioned in that it may not reflect CET-dependent PQ reduction; rather, it may reflect QA reduction in PSII by Fd in vitro (Fisher and Kramer, 2014). Alternatively, it is possible that TCR was inactivated during membrane isolation, because the Fe-S center in TCR is sensitive to oxygen (Figure 2C, mentioned above in the Results section).

TCR appeared after plastid endosymbiosis

According to the endosymbiotic theory, photosynthetic eukaryotes originated from two endosymbiosis events: (1) the endosymbiosis of an ancestor of α-proteobacteria that became mitochondria and (2) the endosymbiosis of an ancestor of cyanobacteria that became plastids (reviewed by Timmis et al., 2004). As TCR homologs exist in most oxygenic phototrophs, including green algae, liverworts, ferns, monocots, and dicots, but not in cyanobacteria (Figure S1), TCR might have appeared after the plastid endosymbiosis and after the separation of the green and red lineages, as was the case for the PGRL1A, PGRL1B (DalCorso et al., 2008), NPQ4 (reviewed by Niyogi and Truong, 2013), NPQ1 (Li et al., 2016), and FdC1 (Guan et al., 2018; Voss et al., 2011) genes. Moreover, the expansion of the content of redox-sensitive cysteines in the proteins of eukaryotes after the first plastid acquisition is believed to coincide with the growing demands of the eukaryotic cells regarding the balancing of the cellular redox states, resulting in protection from oxidative stresses (Woehle et al., 2017). In plants, a large number of cysteine-rich proteins (e.g., thioredoxins, glutaredoxins, peroxiredoxins, and several protein-disulfide isomerases) play essential functions in the hierarchical systems that are used for poising the redox status in stroma and protecting PSI from ROS (Scheibe et al., 2005). Therefore, TCR was established during the evolution of oxygenic phototrophs and contributed to the control of the redox balance in stroma and to the protection of PSI from photodamage.

Limitations of the study

Although this study provides evidences showing that TCR is involved in an alternative electron flow downstream of PSI, further investigation is needed to fully understand the mechanism of the proposed function of TCR. Specifically, this work has not yet fully uncovered the direct electron donor and acceptor of TCR. Furthermore, the stoichiometric ratio between PSI and TCR is still missing due to the lack of the TCR-specific antibody for detecting TCR and the difficulties in identifying the TCR peptide by mass spectrometric analysis.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Shinji Masuda (shmasuda@bio.titech.ac.jp).

Materials availability

Materials generated in this study are available from the lead contact with a completed materials transfer agreement.

Data and code availability

Mass-spectroscopy-based proteome data have been deposited to jPOST (https://jpostdb.org) with the accession number JPST000936 or PXD020752.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.