Two distinct redox cascades cooperatively regulate chloroplast functions and sustain plant viability

Significance

A fundamental challenge for the plant life cycle is how to manage fluctuating light conditions. The thiol-based redox regulation in chloroplasts is a significant system for controlling chloroplast functions in response to light signals. Although chloroplasts likely possess a complex redox network supported by multiple redox-mediator proteins and target enzymes, the whole organization and biological significance have not been fully resolved. We performed biochemical and reverse-genetic studies and concluded that (i) two major redox systems via the ferredoxin-thioredoxin reductase/thioredoxin and NADPH-thioredoxin reductase C pathways differentially but cooperatively drive chloroplast redox regulation and (ii) their concerted action is critical for plant autotrophic growth. The regulatory circuits presented in this work contribute to the understanding of how plants survive in continuously changing environments.

Abstract

The thiol-based redox regulation system is believed to adjust chloroplast functions in response to changes in light environments. A redox cascade via the ferredoxin-thioredoxin reductase (FTR)/thioredoxin (Trx) pathway has been traditionally considered to serve as a transmitter of light signals to target enzymes. However, emerging data indicate that chloroplasts have a complex redox network composed of diverse redox-mediator proteins and target enzymes. Despite extensive research addressing this system, two fundamental questions are still unresolved: How are redox pathways orchestrated within chloroplasts, and why are chloroplasts endowed with a complicated redox network? In this report, we show that NADPH-Trx reductase C (NTRC) is a key redox-mediator protein responsible for regulatory functions distinct from those of the classically known FTR/Trx system. Target screening and subsequent biochemical assays indicated that NTRC and the Trx family differentially recognize their target proteins. In addition, we found that NTRC is an electron donor to Trx-z, which is a key regulator of gene expression in chloroplasts. We further demonstrate that cooperative control of chloroplast functions via the FTR/Trx and NTRC pathways is essential for plant viability. Arabidopsis double mutants impaired in FTR and NTRC expression displayed lethal phenotypes under autotrophic growth conditions. This severe growth phenotype was related to a drastic loss of photosynthetic performance. These combined results provide an expanded map of the chloroplast redox network and its biological functions.

To preserve the integrity and efficiency of photosynthesis and other metabolic reactions, chloroplast enzymes need to be flexibly and appropriately controlled in response to changes in light environments. The photosynthetic electron transport chain in the chloroplast thylakoid membrane converts light energy into chemical energy, which is captured and stored in ATP and NADPH. These molecules are primarily consumed by the Calvin–Benson cycle in the stroma, but part of the reducing power is used for other metabolic reactions in chloroplasts (e.g., nitrogen and sulfur metabolism). One pathway for the reducing power is the redox cascade, mediated by ferredoxin-thioredoxin reductase (FTR) and thioredoxin (Trx). In this system, FTR receives reducing power from the light-driven photosynthetic electron transport chain via ferredoxin and then donates the reducing power to Trx. A reduced form of Trx subsequently transfers reducing power to target proteins through a dithiol–disulfide exchange reaction, allowing the targets to modulate the enzymatic activities. The redox cascade via the FTR/Trx pathway provides the basis for thiol-based redox regulation in chloroplasts and ensures light-responsive control of chloroplast functions (1, 2).

The FTR/Trx pathway has been considered the only pathway regulating redox reactions in chloroplasts. Information about its target proteins has also been limited to a specific set of light-activated enzymes, such as some enzymes involved in the Calvin–Benson cycle. However, plant genome sequences indicated that as many as five Trx subtypes (f, m, x, y, and z) are targeted to chloroplasts (3, 4). Other putative redox-mediator proteins, including several Trx-like proteins and glutaredoxins, were also reported to be localized in chloroplasts (5, 6). Furthermore, recently developed proteomic techniques and methodologies for screening Trx target proteins have identified potential redox-regulated proteins associated with a broad spectrum of chloroplast functions (7–11). These advances suggest that chloroplasts have a complex redox network composed of divergent pathways rather than a simple one-directional cascade. How does the highly organized redox network control a wide range of critical chloroplast functions in flexible and sophisticated ways? These questions must be addressed because the whole organization and biological significance of the chloroplast redox network are currently unresolved.

Among multiple proteins involved in the chloroplast redox network, NADPH-Trx reductase C (NTRC) is emerging as an important enzyme. NTRC is a unique redox-mediator protein in terms of harboring both an NADPH-Trx reductase (NTR) domain and a Trx domain in a single polypeptide (12). NTRC uses NADPH as a source of reducing power, and is thus able to work in the redox regulation independent of the FTR/Trx pathway. NTRC-deficient plants display obvious phenotypes with pale-green leaves (12, 13), indicating that NTRC-dependent redox regulation in chloroplasts has an important role in plants. Although target proteins (13–16) and catalytic mechanisms (17, 18) for NTRC have been partly reported to date, its physiological importance, especially in the context of functional differences from the FTR/Trx system, is largely uncharacterized.

Here we show that the FTR/Trx and NTRC pathways have distinct regulatory roles in the chloroplast redox network. We also present evidence that the cooperative functions of these two redox pathways are essential for plant autotrophic growth. Our results shed light on how and why chloroplasts organize a divergent composition of the redox network, and thus contribute to a better understanding of plant strategies for adjusting chloroplast functions under fluctuating environmental conditions.

Results

NTRC and Trx-f Show Distinct Target Protein Profiles.

For the first step in this study, we explored NTRC target proteins in chloroplasts. For this purpose, Trx-affinity chromatography was applied as described previously (7), except that the bait was changed to an Arabidopsis NTRC mutant protein in which the second Cys residue in the Trx domain active site was substituted with Ser (NTRC_C457S). Chloroplast soluble proteins prepared from spinach leaves were loaded onto an NTRC_C457S-immobilized affinity chromatography column. We successfully captured some proteins that interact with NTRC_C457S via the mixed-disulfide bond (Fig. 1A, lane 4). These potential NTRC targets were identified by mass spectrometry, N-terminal sequence analysis, and/or immunoblotting. The results revealed that NTRC can associate with several currently known Trx target candidates (11). An unexpected and notable NTRC target was Trx-z; Trx-z bound to NTRC_C457S, whereas Trx-f and Trx-m did not. It should be noted that Trx-z emerged as a strong signal in DTT eluates (Fig. 1A, lane 4), although it was below the detection limit in an original sample (Fig. 1A, lane 1). These results indicate that Trx-z bound to NTRC_C457S with high affinity and specificity, because the chromatography method used here enables target candidate proteins to be selectively concentrated from the original sample (7).

Fig. 1.

Screening of NTRC target candidate proteins from spinach chloroplasts. (A) Protein elution profiles of NTRC-affinity chromatography. NTRC mutant protein with a modified Trx domain (TRXd) active site (NTRC_C457S) was used as bait. Lane 1, original sample; lane 2, flow-through; lane 3, NaCl wash just before DTT elution; lane 4, DTT elution. (B) Comparison of NTRC and Trx-f1 target profiles. Chromatography-based target screening was performed using NTRC mutant protein with a modified NTR domain (NTRC_C217S) and Trx-f1 mutant protein (Trx-f1_C102S) as baits (Fig. S1). Target protein profiles of NTRC_C217S (lane 1), NTRC_C457S (lane 2), and Trx-f1_C102S (lane 3) are shown. (A and B) Proteins were visualized by silver staining or immunoblotting. Proteins indicated with arrows were identified by mass spectrometry (including peptide mass fingerprinting and MS/MS analysis) and/or N-terminal sequence analysis. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GSAT, Glu-1-semialdehyde aminotransferase; RbcS, Rubisco small subunit; RCA, Rubisco activase.

We repeated the experiment using NTRC mutant protein with a modified NTR domain (NTRC_C217S) and Trx-f1 mutant protein (Trx-f1_C102S) as baits (Fig. S1), followed by comparison of each target protein profile. As demonstrated in Fig. 1B, several proteins showed differential binding efficiencies to NTRCs and Trx-f1: 2-Cys peroxiredoxin (2-Cys Prx) and Mg-chelatase I subunit (CHLI) preferably bound to NTRC_C217S and NTRC_C457S, whereas fructose-1,6-bisphosphatase (FBPase), NADP-malate dehydrogenase (NADP-MDH), and ADP-glucose pyrophosphorylase (AGPase) preferentially bound to Trx-f1_C102S. Interestingly, Trx-z interacted specifically with NTRCs but not with Trx-f1. These results suggest that NTRC and Trx-f have distinct selectivity for their target proteins.

Fig. S1.

Screening of NTRC and Trx-f1 target candidate proteins from spinach chloroplasts. (A) Protein elution profiles of NTRC-affinity chromatography. NTRC mutant protein with a modified NTR domain (NTRd) active site (NTRC_C217S) was used as bait. (B) Protein elution profiles of Trx-f1–affinity chromatography. Trx-f1 mutant protein with a modified active site (Trx-f1_C102S) was used as bait. Lane 1, original sample; lane 2, flow-through; lane 3, NaCl wash just before DTT elution; lane 4, DTT elution. Proteins were visualized by silver staining or immunoblotting.

NTRC and Five Trx Subtypes Transfer Reducing Power to Target Proteins with Different Efficiencies in Vitro.

To obtain biochemical insights into crosstalk among NTRC, the Trx family, and their target proteins, we analyzed the efficiencies in reducing power transfer from NTRC and five Trx subtypes to some targets. For this purpose, we prepared recombinant Arabidopsis proteins of NTRC, Trxs, and several targets. NTRC was confirmed to possess NADPH-dependent dithiol–disulfide exchange reaction activity (Fig. S2). First, we focused on the target proteins 2-Cys Prx and CHLI, which were preferentially captured by NTRCs (Fig. 1B).

Fig. S2.

Measurement of recombinant NTRC protein activity. SDS/PAGE profile (Left) and DTNB reduction activity (Right) are shown. For DTNB reduction activity, the time course of absorbance changes at 412 nm (Abs₄₁₂) is shown. NTRC was added at the indicated concentrations. NADPH (0.2 mM) was added at the time point indicated by the arrow.

The target protein 2-Cys Prx is a component of the antioxidant defense system in chloroplasts and has a role in detoxification of hydrogen peroxide (H₂O₂) (19). Two isoforms of 2-Cys Prx, 2-Cys Prx A and B, are conserved in Arabidopsis and share high amino acid sequence homology (96–97% in the mature protein region). The oxidized form of 2-Cys Prx exists as a dimer stabilized by intermolecular disulfide bonds. Fig. 2 A and B show the redox shift pattern of 2-Cys Prx A from an oxidized dimeric form to a reduced monomeric form when incubated with various concentrations of NTRC or Trxs. As the source of reducing power, 0.5 mM NADPH (for NTRC) or 0.5 mM DTT (for Trxs) was supplied; each reductant could adequately reduce NTRC or Trxs, respectively (see below), but did not directly affect the redox state of 2-Cys Prx A (Fig. 2 A and B). We thus analyzed the efficiencies of NTRC and Trxs in reducing 2-Cys Prx A using NADPH/NTRC and DTT/Trx, respectively. Trx-f1, Trx-m2, and Trx-x had relatively high efficiencies among the five Trx subtypes tested, but greater than 2 μM was required for full reduction of 2-Cys Prx A. By contrast, NTRC achieved full reduction of 2-Cys Prx A at the substantially lower concentration of 0.2 μM. These results indicate that NTRC exerts a much higher efficiency in reducing 2-Cys Prx A than the five Trx subtypes.

Fig. 2.

Biochemical characterization of reducing power transfer from NTRC and Trxs to 2-Cys Prx A or CHLI1_C102S/C193S. (A and B) Redox shift visualization of 2-Cys Prx A. (A) Oxidized 2-Cys Prx A (2 μM) was incubated with NADPH/NTRC (0.5 mM NADPH and 0–1 μM NTRC) or DTT/Trx (0.5 mM DTT and 0–5 μM Trx). Free thiols were modified with N-ethylmaleimide, and proteins were subjected to nonreducing SDS/PAGE. (B) 2-Cys Prx A reduction-level dependency on NTRC and Trx concentrations. (C–F) Redox shift visualization of CHLI1_C102S/C193S. (C) Oxidized CHLI1_C102S/C193S (2 μM) was incubated with NADPH/NTRC (0.5 mM NADPH and 0–1 μM NTRC) or NADPH/EcNTR/Trx-m2 (0.5 mM NADPH, 0.1 μM EcNTR, and 0–1 μM Trx-m2). (D) CHLI1_C102S/C193S reduction-level dependency on NTRC and Trx-m2 concentrations. (E) Oxidized CHLI1_C102S/C193S (2 μM) was incubated with various concentrations of DTT (0–500 μM) in the presence or absence of 1 μM Trx. (F) CHLI1_C102S/C193S reduction-level dependency on DTT concentration in the presence or absence of Trx. (C and E) Free thiols were modified with maleimide-PEG₁₁-biotin, and proteins were subjected to nonreducing SDS/PAGE. (B, D, and F) Reduction level was quantified as the ratio of the reduced form to the total. Each value represents the mean ± SD (three or four independent experiments). Ox, oxidized form; Red, reduced form.

CHLI is involved in the chlorophyll biosynthetic pathway as an ATP-hydrolytic component of the Mg-chelatase complex (20). Arabidopsis chloroplasts contain two CHLI isoforms, CHLI1 and CHLI2; the former was reported to be the major isoform with abundant expression in plants (21). Arabidopsis CHLI1 has four Cys residues (Cys¹⁰², Cys¹⁹³, Cys³⁵⁴, and Cys³⁹⁶) that form two intramolecular disulfide bonds (Cys¹⁰²–Cys¹⁹³ and Cys³⁵⁴–Cys³⁹⁶) (22). The C-terminal disulfide bond (Cys³⁵⁴–Cys³⁹⁶) is expected to be critically involved in controlling the ATPase activity of CHLI1. To analyze the efficiencies of NTRC and Trxs in cleaving this disulfide bond, a CHLI1 mutant protein was constructed in which the N-terminal Cys pair (Cys¹⁰² and Cys¹⁹³) was substituted with Ser (CHLI1_C102S/C193S). Molecular mass shift-based determination of the redox state using a thiol-modifying reagent (7, 23, 24) indicated that CHLI1_C102S/C193S was efficiently reduced only by 0.5 mM DTT even in the absence of Trxs (Fig. S3A, lanes 2–5). Therefore, it appeared to be difficult to evaluate the reduction efficiency for CHLI1_C102S/C193S using DTT/Trx. To only reduce Trxs without directly affecting the CHLI1_C102S/C193S redox state, we prepared recombinant NTR from Escherichia coli (EcNTR). Redox-state determination (Fig. S3B) and activity measurement (Fig. S3C) indicated that EcNTR efficiently reduced Trx-m2 but not other Trxs. Consequently, CHLI1_C102S/C193S was reduced in the presence of NADPH, EcNTR, and Trx-m2 (Fig. S3D). We then compared the efficiency in reducing power transfer from NTRC and Trx-m2 to CHLI1_C102S/C193S using NADPH/NTRC and NADPH/EcNTR/Trx-m2, respectively, with changing NTRC and Trx-m2 concentrations (Fig. 2 C and D). The results showed that NTRC had higher reduction efficiency than Trx-m2 for CHLI1_C102S/C193S. We also evaluated the CHLI1_C102S/C193S-reduction efficiencies of five Trx subtypes with varying concentrations of DTT (Fig. 2 E and F). By comparing the reduction pattern of CHLI_C102S/C193S, Trx-y1 was determined to be the most efficient electron donor to CHLI1_C102S/C193S, but Trx-m2 also had comparatively high reduction efficiency.

Fig. S3.

Biochemical characterization of reducing power transfer of NTRC, EcNTR, Trxs, and CHLI1_C102S/C193S. (A) Redox shift visualization of CHLI1_C102S/C193S. Oxidized CHLI1_C102S/C193S (2 μM) was incubated with NADPH/NTRC (0.5 mM NADPH and 0–5 μM NTRC; lanes 6–9) or DTT/Trx (0.5 mM DTT and 1 μM Trx-f1 or Trx-m2; lanes 3–5). (B) Redox shift visualization of Trxs. Each oxidized Trx (1 μM) was incubated with 0.5 mM DTT or NADPH/EcNTR (0.5 mM NADPH with or without 0.1 μM EcNTR). (C) Monitoring electron transfer from NADPH via EcNTR and Trxs to DTNB. The time course of absorbance changes at 412 nm (Abs₄₁₂) is shown. Both EcNTR and Trx were added at 1 μM. NADPH (0.2 mM) was added at the time point indicated by the arrow. (D) Redox shift visualization of CHLI1_C102S/C193S. Oxidized CHLI1_C102S/C193S (2 μM) was incubated with 0.5 mM DTT or NADPH/EcNTR/Trx (0.5 mM NAPDH, 0.1 μM EcNTR, and 1 μM Trx). (A and D) Free thiols were modified with maleimide-PEG₁₁-biotin, and proteins were subjected to nonreducing SDS/PAGE. (B) Free thiols were modified with 4-acetamido-4′-maleimidylstilbene-2,2′-disulfonate, and proteins were subjected to nonreducing SDS/PAGE. Ox, oxidized form; Red, reduced form.

The efficiencies of reducing power transfer to 2-Cys Prx A and CHLI1_C102S/C193S were further investigated with changing experimental conditions. Fig. S4 A and B show the time-dependent change in 2-Cys Prx A redox state after incubation with NADPH/NTRC or DTT/Trx. NTRC reduced 2-Cys Prx A more rapidly than Trxs, and achieved full reduction of 2-Cys Prx A within 10 min. CHLI1_C102S/C193S was also reduced more rapidly by NADPH/NTRC than NADPH/EcNTR/Trx-m2 (Fig. S4 C and D). Compared with the responses at pH 7.5 (Fig. 2), NTRC- and Trx-dependent reductions of both 2-Cys Prx A and CHLI1_C102S/C193S were generally suppressed at pH 8.5, but the trend of higher reduction efficiency of NTRC compared with those of Trxs was true under these conditions as well (Fig. S4 E–H). Taken together, we concluded that NTRC serves as a dominant reducer of 2-Cys Prx and CHLI, at least within the range of physiological pH.

Fig. S4.

Biochemical characterization of reducing power transfer from NTRC and Trxs to 2-Cys Prx A or CHLI1_C102S/C193S. (A and B) Time course of NTRC- or Trx-dependent 2-Cys Prx A reduction. Oxidized 2-Cys Prx A (2 μM) was incubated with NADPH/NTRC (0.5 mM NADPH and 1 μM NTRC) or DTT/Trx (0.5 mM DTT and 1 μM Trx) for the indicated times. (C and D) Time course of NTRC- or Trx-m2–dependent CHLI1_C102S/C193S reduction. Oxidized CHLI1_C102S/C193S (2 μM) was incubated with NADPH/NTRC (0.5 mM NADPH and 1 μM NTRC) or NADPH/EcNTR/Trx-m2 (0.5 mM NADPH, 0.1 μM EcNTR, and 1 μM Trx-m2) for the indicated times. (E and F) 2-Cys Prx A reduction-level dependency on NTRC and Trx concentrations at pH 8.5. Oxidized 2-Cys Prx A (2 μM) was incubated with NADPH/NTRC (0.5 mM NADPH and 0–1 μM NTRC) or DTT/Trx (0.5 mM DTT and 0–5 μM Trx). (G and H) CHLI1_C102S/C193S reduction-level dependency on NTRC and Trx-m2 concentrations at pH 8.5. Oxidized CHLI1_C102S/C193S (2 μM) was incubated with NADPH/NTRC (0.5 mM NADPH and 0–1 μM NTRC) or NADPH/EcNTR/Trx-m2 (0.5 mM NADPH, 0.1 μM EcNTR, and 0–1 μM Trx-m2). (A and E) Free thiols were modified with N-ethylmaleimide, and proteins were subjected to nonreducing SDS/PAGE. (C and G) Free thiols were modified with maleimide-PEG₁₁-biotin, and proteins were subjected to nonreducing SDS/PAGE. (B, D, F, and H) Reduction level was quantified as the ratio of the reduced form to the total. Each value represents the mean ± SD (three or four independent experiments).

We performed additional biochemical analyses with other target proteins, including peroxiredoxin Q (PrxQ), FBPase, sedoheptulose-1,7-bisphosphatase (SBPase), and NADP-MDH (Fig. 3). PrxQ is classified as an antioxidant enzyme along with 2-Cys Prx (19). Real-time monitoring of electron transfer (Fig. 3A) and redox state determination (Fig. 3B) indicated that PrxQ could be reduced by NTRC. However, its efficiency was low compared with those of Trxs; PrxQ was strongly reduced by 1 μM Trx-f1 or 1 μM Trx-m2 (Fig. 3B, lanes 2–4), whereas it was only partially reduced by NTRC even at 5 μM (Fig. 3B, lanes 5–8). It should be noted that other Trxs (Trx-x, Trx-y, and Trx-z) have higher PrxQ-reduction efficiency than Trx-f and Trx-m (24). FBPase, SBPase, and NADP-MDH are classically known Trx target enzymes. The former two enzymes constitute the Calvin–Benson cycle, whereas the latter is involved in the malate valve, which plays a key role in exporting excess reducing power from chloroplasts (25). FBPase, SBPase, and NADP-MDH were reduced by Trx-f1 and Trx-m2 with different efficiencies (Fig. 3C, lanes 2–4), in agreement with our previous report (24). By contrast, NTRC failed to reduce these proteins (Fig. 3C, lanes 5–8). These results indicate that Trxs, but not NTRC, exclusively function in the redox regulation process of FBPase, SBPase, and NADP-MDH, although these target proteins were captured by NTRCs with an affinity comparable to (SBPase) or lower than (FBPase and NADP-MDH) that of Trx-f1 in the target screening experiments (Fig. 1). In this regard, it is worth considering that affinity chromatography-based screening often identifies nonspecific targets, and individual biochemical studies are needed to confirm the validity of the redox regulation (26).

Fig. 3.

Biochemical characterization of reducing power transfer from NTRC and Trxs to PrxQ, FBPase, SBPase, or NADP-MDH. (A) Monitoring electron transfer from NADPH via NTRC and PrxQ to H₂O₂. The time course of absorbance changes at 340 nm (Abs₃₄₀) is shown. NTRC was added at the indicated concentrations. H₂O₂ (1 mM) was added at the time point indicated by the arrow. (B) Redox shift visualization of PrxQ. (C) Redox shift visualization of FBPase, SBPase, and NADP-MDH. (B and C) Each oxidized protein (2 μM) was incubated with NADPH/NTRC (0.5 mM NADPH and 0–5 μM NTRC; lanes 5–8) or DTT/Trx (0.5 mM DTT only, or 0.5 mM DTT and 1 μM Trx-f1 or Trx-m2; lanes 2–4). Free thiols were modified with 4-acetamido-4′-maleimidylstilbene-2,2′-disulfonate, and proteins were subjected to nonreducing SDS/PAGE.

NTRC Preferably Reduces Trx-z but Not Other Trx Subtypes.

A notable target candidate protein of NTRC was Trx-z (Fig. 1). Therefore, we investigated whether NTRC transfers reducing power to each of the five Trx subtypes (Fig. 4A). Trx-f (Trx-f1 and Trx-f2) was slightly reduced by NADPH/NTRC, whereas Trx-m (Trx-m2 and Trx-m4), Trx-x, or Trx-y was not reduced at all. In contrast to these four Trx subtypes, Trx-z was reduced by NADPH/NTRC with substantial efficiency. Further experiments with lower concentrations of NTRC (Fig. 4B) and on the time-dependent changes in redox states (Fig. 4C) indicated that NTRC did reduce Trx-z more efficiently than Trx-f1. Redox titration of NTRC indicated that the midpoint redox potentials of the conserved Cys pairs in the NTR domain (Cys²¹⁷–Cys²²⁰) and Trx domain (Cys⁴⁵⁴–Cys⁴⁵⁷) were −275 ± 2 mV and −274 ± 1 mV at pH 7.5, respectively (Fig. S5). These values are comparable to the midpoint redox potential of Trx-z (–276 mV) and less negative than those of other Trxs (Fig. 4D) (24). The surface electrostatic potentials in the 3D protein structure model indicate that Trx-z has more negative surface charge than any other Trxs (27). The NTRC Trx domain is positively charged; therefore, efficient electron transfer from the NTRC Trx domain to Trx-z seems reasonable in terms of electrostatic interactions. These combined results suggest that, out of the five Trx subtypes, only Trx-z receives reducing power from NTRC.

Fig. 4.

Biochemical characterization of reducing power transfer from NTRC to Trxs. (A) Redox shift visualization of Trxs. Each oxidized Trx (2 μM) was incubated with 0.5 mM DTT (lane 2) or NADPH/NTRC (0.5 mM NADPH and 0–5 μM NTRC; lanes 3–6). (B) Trx reduction-level dependency on NTRC concentration. Oxidized Trx-f1 or Trx-z (2 μM) was incubated with NADPH/NTRC (0.5 mM NADPH and 0–1,000 nM NTRC). (C) Time course of NTRC-dependent Trx reduction. Oxidized Trx-f1 or Trx-z (2 μM) was incubated with NADPH/NTRC (0.5 mM NADPH and 1 μM NTRC) for the indicated times. (A–C) Free thiols were modified with 4-acetamido-4′-maleimidylstilbene-2,2′-disulfonate, and proteins were subjected to nonreducing SDS/PAGE. (B and C) Reduction level was quantified as the ratio of the reduced form to the total. Each value represents the mean ± SD (three independent experiments). (D) Comparison of midpoint redox potentials of NTRC [NTR domain (NTRd) and Trx domain (TRXd)] with those of Trxs. Midpoint redox potentials of Trxs were plotted according to data from our previous study (24).

Fig. S5.

Determination of midpoint redox potentials of the NTR domain (NTRd) and Trx domain (TRXd) in Arabidopsis NTRC. (A) Proteins were equilibrated with various [reduced DTT (DTT_red)]/[oxidized DTT (DTT_ox)] ratios of redox buffers. Free thiols were modified with maleimide-PEG₁₁-biotin (for NTRd) or 4-acetamido-4′-maleimidylstilbene-2,2′-disulfonate (for TRXd), and proteins were subjected to nonreducing SDS/PAGE. (B) Reduction level was quantified as the ratio of the reduced form to the total and plotted against the redox potential of DTT buffer (pH 7.5). Data were fitted to the Nernst equation. Midpoint redox potentials (E_m) were determined from three independent experiments; each value represents the mean ± SD.

NTR and Trx Domains Are Both Necessary for NTRC-Dependent Redox Regulation.

Three target proteins preferably reduced by NTRC (2-Cys Prx, CHLI, and Trx-z) bound to NTRC_C217S and NTRC_C457S with almost equal affinity (Fig. 1B, lanes 1 and 2), implying that these proteins can receive reducing power directly from the NTR domain without passing through the Trx domain. In fact, a previous study reported that the Trx domain active site is dispensable for NTRC-dependent 2-Cys Prx reduction in a thermophilic cyanobacterium (28). We validated this possibility using a series of NTRC variant proteins, in which redox-active Cys pairs conserved in the NTR domain and/or Trx domain were substituted with Ser (NTRC_C217S/C220S, NTRC_C454S/C457S, and NTRC_{C217S/C220S/C454S/C457S}). In all cases examined, only wild-type NTRC could transfer reducing power to target proteins (Fig. S6). These results clearly indicate that electron transfer from NADPH via the NTR domain to the Trx domain is essential for NTRC-dependent redox regulation of 2-Cys Prx, CHLI, and Trx-z in Arabidopsis.

Fig. S6.

Biochemical characterization of reducing power transfer from NTRC wild-type (WT) and mutant proteins to 2-Cys Prx A, CHLI1_C102S/C193S, or Trx-z. (A) Redox shift visualization of 2-Cys Prx A. Oxidized 2-Cys Prx A (2 μM) was incubated with 0.5 mM NADPH with or without 1 μM NTRC. Free thiols were modified with N-ethylmaleimide, and proteins were subjected to nonreducing SDS/PAGE. (B) Monitoring electron transfer from NADPH via NTRC and 2-Cys Prx A to H₂O₂. The time course of absorbance changes at 340 nm (Abs₃₄₀) is shown. NTRC was added at 1 μM. H₂O₂ (1 mM) was added at the time point indicated by the arrow. (C) Redox shift visualization of CHLI1_C102S/C193S. Oxidized CHLI1_C102S/C193S (2 μM) was incubated with 0.5 mM NADPH with or without 1 μM NTRC. Free thiols were modified with maleimide-PEG₁₁-biotin, and proteins were subjected to nonreducing SDS/PAGE. (D) Redox shift visualization of Trx-z. Oxidized Trx-z (2 μM) was incubated with 0.5 mM NADPH with or without 1 μM NTRC. Free thiols were labeled with 4-acetamido-4′-maleimidylstilbene-2,2′-disulfonate, and proteins were subjected to nonreducing SDS/PAGE.

Arabidopsis Double Mutants Impaired in FTR and NTRC Expression Display Severe Growth Phenotypes.

In vitro biochemical studies identified distinct regulatory properties for Trxs and NTRC. To gain physiological insights into FTR/Trx and NTRC pathway functions in planta, we obtained Arabidopsis T-DNA insertion mutants in the FTRB gene encoding the FTR catalytic subunit (ftrb) and the NTRC genes (ntrc-1 and ntrc-2) (Fig. S7 A and B) and cultivated the plants in soil (Fig. S7C). In RT-PCR analyses, the NTRC transcripts were undetected in the ntrc mutants, whereas the FTRB transcript was largely decreased but still detected in the ftrb mutant (Fig. S7D), possibly due to the T-DNA insertion in an intron of the FTRB gene (Fig. S7A). Quantitative (q)RT-PCR and immunoblotting analyses indicated that the FTRB transcript and FTR catalytic subunit protein levels in ftrb were 14 ± 5% and 23 ± 13%, respectively, of those of the wild-type plants (mean ± SD, n = 3) (Fig. S7 D and E). The FTR variable subunit was also lowered in ftrb to ∼70% of that of wild type (Fig. S7E). In soil-grown plants, the FTR catalytic subunit was elevated in ntrc to approximately twofold greater than that of wild type (Fig. S7E), implying partial functional redundancy between the FTR/Trx and NTRC pathways under certain conditions (Discussion). Although growth was impaired in ftrb and ntrc, their phenotypes were distinguishable; shoot fresh weight was more remarkably lowered in ntrc than in ftrb (Fig. S7F), and ntrc, but not ftrb, developed pale-green leaves due to lower chlorophyll content (Fig. S7 C and G).

Fig. S7.

Characterization of Arabidopsis WT and mutant plants. (A) The T-DNA insertion site in the FTRB and NTRC genes in mutant plants. Each insertion site was determined by sequencing the DNA fragment containing the region bordered by T-DNA. Fw, forward; Rev, reverse. (B) PCR-based genotyping of WT and mutant plants. (C) Plants were cultivated in soil for 4 wk. (D) Transcript-level analyses. (Left) RT-PCR analyses. (Right) Quantitative RT-PCR analyses of FTRB transcript levels. Relative values to the WT level are shown. Each value represents the mean ± SD (three biological replicates). (E) Immunoblotting analyses. The same amount of leaf total protein was loaded into each lane. As a loading control, Rubisco large subunit (RbcL) was stained with CBB. FTR-c and FTR-v, FTR catalytic and variable subunits. (F) Shoot fresh weight (FW). (G) Chlorophyll (Chl) content and a/b ratio. (F and G) Each value represents the mean ± SD (12 biological replicates). Different letters denote significant differences among plants (P < 0.01, Tukey–Kramer multiple comparison test).

We further generated ftrb ntrc double mutants, which displayed severe growth phenotypes (Fig. 5). Under autotrophic growth conditions [soil or half-strength Murashige-Skoog (MS) medium in the absence of sucrose], the ftrb ntrc double mutants displayed a lethal phenotype after germination (Fig. 5 A and B). The ftrb ntrc double mutants were viable in MS medium supplemented with 2% (wt/vol) sucrose, but their growth was drastically retarded (Fig. 5C). These results strongly suggest that both the FTR/Trx and NTRC pathways participate in the redox regulation of chloroplast functions in vivo and cooperatively support plant autotrophic growth.

Fig. 5.

Growth phenotypes of Arabidopsis wild-type (WT) and mutant plants. Plants were cultivated in soil (A) or half-strength MS medium in the absence (B) or presence (C) of 2% sucrose (Suc) for the indicated days.

Photosynthesis Is Disrupted in ftrb ntrc Double Mutants.

We characterized physiological parameters related to photosynthesis using rosette leaves developed in MS medium with 2% sucrose (Fig. 5C). Chlorophyll contents of ntrc and ftrb ntrc were 57–64% and 28–36%, respectively, of that of wild type (Fig. 6A). The chlorophyll a/b ratio was specifically lowered in ftrb ntrc. Immunoblotting analyses indicated that the accumulation levels of several proteins were extensively altered, particularly in ftrb ntrc (Fig. 6B and Fig. S8 A–C). In particular, proteins involved in photosynthetic electron transport, including photosystem (PS)II/I core proteins and light-harvesting complex proteins, were strongly decreased in ftrb ntrc (Fig. 6B and Fig. S8A). Photosynthetic electron transport was evaluated by chlorophyll fluorescence measurements. The results suggested that ftrb ntrc was more susceptible to PSII photoinhibition and unable to efficiently perform photosynthetic electron transport (Fig. 6C).

Fig. 6.

Photosynthesis-related physiological parameters of Arabidopsis WT and mutant plants grown in half-strength MS medium supplemented with 2% sucrose. (A) Chlorophyll (Chl) content and a/b ratio. Each value represents the mean ± SD (five or six biological replicates). Different letters denote significant differences among plants (P < 0.01, Tukey–Kramer multiple comparison test). (B) Accumulation levels of photosynthetic electron transport proteins. Relative values to the WT level are shown. Each value represents the mean ± SD (three or four biological replicates). Raw immunoblotting images used for the quantification are shown in Fig. S8A. (C) Chlorophyll fluorescence measurements. Maximal quantum yield of PSII (F_v/F_m) and relative electron transport rate (ΦPSII × light intensity) are shown. Each value represents the mean ± SD (four to six biological replicates). Different letters denote significant differences among plants (P < 0.01, Tukey–Kramer multiple comparison test).

Fig. S8.

Protein analyses of Arabidopsis WT and mutant plants grown in half-strength MS medium supplemented with 2% sucrose. (A and B) Immunoblotting analyses of chloroplast proteins. The same amount of leaf total protein was loaded into each lane. Quantification data for the proteins shown in A are described in Fig. 6B. (C) Quantification data for the proteins shown in B. Relative values to the WT level are shown. Each value represents the mean ± SD (three or four biological replicates). CET-PSI, cyclic electron transport around PSI; Cyt b₆, cytochrome b₆; FNR, ferredoxin-NADP⁺ reductase; GluTR, glutamyl-tRNA reductase; GSAT, Glu-1-semialdehyde aminotransferase; LHCs, light harvesting complexes; PC, plastocyanin; PORC, protochlorophyllide oxidoreductase C; RbcS, Rubisco small subunit. N.S., nonspecific protein determined by the comparison between wild-type and trxz mutant plants (Fig. S9). (D) Visualization of in vivo redox states of redox-regulated proteins and Trxs. Experiments were performed under light conditions of 0, 70, and 660 μmol photons⋅m⁻²⋅s⁻¹. The same amount of leaf total protein was loaded into each lane. As a loading control, RbcL was stained with Ponceau S. Quantification data for the reduction levels are described in Fig. 7. Due to the detection of plural protein bands, the reduction level of Trx-f1 was not quantified.

We next examined the in vivo photoreduction dynamics of several redox-regulated proteins and Trxs (Fig. 7 and Fig. S8D). In accordance with our previous results (24, 29), each protein showed distinct redox shift patterns upon illumination; in the wild-type plants, the ATP synthase CF₁-γ subunit was fully reduced even under low-light conditions, whereas FBPase and SBPase were gradually reduced concomitantly with increasing light intensity. Redox states of the antioxidant enzymes PrxQ and 2-Cys Prx were essentially stably maintained irrespective of the light conditions. All Trxs examined here (Trx-f1, Trx-m2, Trx-m4, and Trx-x) showed uneven redox shifts upon illumination. The redox behaviors in mutant plants were significantly altered from those in wild type. Photoreduction responses of FBPase and SBPase were partially restricted in the ftrb and ntrc single mutants, with more marked inhibitory effects in ntrc. It is noteworthy that the basal reduction level of 2-Cys Prx was lowered in ntrc. More drastic changes in photoreduction behaviors were observed in the ftrb ntrc double mutants. The saturating reduction level of CF₁-γ was lowered specifically in ftrb ntrc. Photoreduction responses of FBPase and SBPase were also severely impaired in ftrb ntrc. The relationship between the redox state and the enzymatic activity of these proteins has been well-documented (1, 2, 30, 31). Taken together, our results suggest that light-responsive activation of ATP synthase and the Calvin–Benson cycle enzymes is repressed in ftrb ntrc. Furthermore, photoreduction responses of Trx-f1, Trx-m2, Trx-m4, and Trx-x were impaired specifically in ftrb ntrc, indicating that the chloroplast redox regulation system itself was less driven in these mutants. Considering all the results of this study, we conclude that cooperative redox regulation via the FTR/Trx and NTRC pathways is required for optimal chloroplast functions, including photosynthesis.

Fig. 7.

In vivo photoreduction dynamics of redox-regulated proteins and Trxs. Experiments were performed under light conditions of 0, 70, and 660 μmol photons m⁻² s⁻¹. Reduction level was quantified as the ratio of the reduced form to the total. Each value represents the mean ± SD (three biological replicates). Raw immunoblotting images used for the determination of reduction level are shown in Fig. S8D.

Discussion

Genomics and proteomics analyses have identified an extensive catalog of redox-mediator proteins and their target enzymes, suggesting that chloroplasts contain a complex redox network. However, the whole organization and biological importance of the redox network were unresolved. Our study has elucidated unrevealed aspects of the chloroplast redox network, which provides a more comprehensive map of chloroplast function and regulation.

NTRC and Five Trx Subtypes Differentially Recognize Target Proteins.

NTRC is one of the key redox-mediator proteins, first reported in 2004 (12). NTRC target proteins include 2-Cys Prx (13), AGPase (14), and Mg-chelatase subunits (and perhaps other proteins involved in tetrapyrrole metabolism) (15, 16). Nevertheless, the specific regulatory function of NTRC has remained elusive, because NTRC targets reported to date largely overlap with Trx targets (11). Although there are a few comparative studies addressing the functional difference between NTRC and specific Trx subtypes (16, 32–34), the results are partly inconsistent and, thereby, insufficient to verify distinct roles for NTRC from those for the Trx family.

In this study, we determined specific or preferred NTRC targets. Using NTRC-affinity chromatography, we identified several NTRC target candidates in spinach chloroplasts (Fig. 1A). Although most of these were already listed as Trx target candidates (11), a comparison of the NTRC target profile with that of Trx-f revealed an intriguing fact: Several targets had distinct interaction efficiencies with NTRC and Trx-f (Fig. 1B). For example, AGPase, a key enzyme of starch synthesis, had greater binding to Trx-f than to NTRC. AGPase is activated by both NTRC- and Trx-f–mediated reduction (14, 35). A recent study demonstrated that combined deficiency of NTRC and Trx-f1 strongly inhibited starch accumulation in Arabidopsis, which may indicate cooperative activation of AGPase by NTRC and Trx-f1 in vivo (36), although it was suggested that AGPase redox regulation is not a major determinant of starch accumulation (37). Our present results suggest that Trx-f has higher affinity for AGPase than NTRC, at least in terms of thiol-based physical interaction. By contrast, 2-Cys Prx and CHLI preferentially interacted with NTRC rather than Trx-f (Fig. 1B). Biochemical assays using recombinant proteins provided evidence that 2-Cys Prx and CHLI actually receive reducing power from NTRC more efficiently than almost all Trx subtypes (Fig. 2 and Fig. S4). We also clarified that NTRC is totally inefficient in reducing FBPase, SBPase, or NADP-MDH, which are well-known Trx target proteins (Fig. 3C). These combined results were integrated to generate a redox network model with distinct target selectivity for NTRC and the Trx family (Fig. 8). To simplify the model, we omitted depicting the functional diversity within the Trx family, which was clarified in part by previous studies (24, 38, 39). In this regard, we further found a previously undescribed Trx selectivity in redox regulation: Trx-y is the most efficient reducer of CHLI (Fig. 2 E and F). Therefore, Trx-y may have a major role in Mg-chelatase activation in addition to NTRC, but differences in their contributions should be investigated in the future.

Fig. 8.

Proposed model of the chloroplast redox network. Distinct target selectivity between NTRC (red arrows) and the Trx family (blue arrows) is shown. For common targets (2-Cys Prx, CHLI, and PrxQ), the difference in reducing power transfer efficiencies is roughly represented as the thickness of the arrows. Possible redox pathways around Trx-z suggested by this study and other previous studies (dotted arrows) are also shown. ETC, electron transport chain. See main text for details.

Consequences of the NTRC/Trx-z Redox Cascade.

As a notable finding of this study, Trx-z was characterized as a target of NTRC (Figs. 1 and 4). Interplay between NTRC and Trxs has been addressed so far, but a consistent picture has not been drawn. For example, Bohrer et al. concluded that NTRC does not have a significant capacity to reduce Trxs (40), whereas Rintamäki and colleagues proposed that NTRC has crosstalk with some Trxs (27, 41); there is a clear discrepancy in their conclusions. Using a chromatography-based method, we found that NTRC associates with Trx-z with high affinity (Fig. 1). The subsequent biochemical assay clearly indicated that NTRC transfers reducing power to Trx-z (Fig. 4 A–C). Although we also observed NTRC-dependent reduction of Trx-f, its efficiency was quite low compared with that of Trx-z (Fig. 4 A–C). Indeed, Trx-f was not trapped by NTRC-affinity chromatography (Fig. 1A and Fig. S1A). Inefficiency of reducing power transfer from NTRC to Trx-f was supported by the difference in midpoint redox potentials (Fig. 4D). We therefore concluded that Trx-z is a unique Trx targeted by NTRC (Fig. 8).

An Arabidopsis mutant deficient in Trx-z (trxz) displays an albino phenotype (Fig. S9) (42, 43). Arsova et al. demonstrated that Trx-z binds to two fructokinase-like proteins (FLNs) in a thiol-dependent manner, which is essential for plastid-encoded RNA polymerase (PEP)-dependent gene expression and proper chloroplast development (42). In accordance with this report, Trx-z and FLNs were identified as PEP complex components by proteomic studies (44, 45). All these results let us hypothesize that the NADPH/NTRC/Trx-z redox cascade acts as a redox-based switch of PEP-dependent gene expression in chloroplasts, although Trx-z likely receives reducing power from other redox-mediator proteins (Fig. 8; see below for details).

Fig. S9.

Growth phenotypes of Arabidopsis WT and trxz mutant plants. (A) Plants were cultivated in half-strength MS medium supplemented with 2% sucrose for 12 d. (B) PCR-based genotyping of trxz. WT Trx-z–specific (P1, Trx-z Fw and Trx-z Rev) and T-DNA–inserted Trx-z–specific (P2, Trx-z Fw and Salk-LB2R) primer pairs were used (Table S1). (C) Immunoblotting analyses of Trx-z and NTRC. The same amount of leaf total protein was loaded into each lane. Proteins were also stained with CBB. N.S., nonspecific protein.

On the other hand, recent work showed that disrupted FLN redox regulation may not be the leading cause of the drastic phenotype of trxz (46). If so, other proteins indispensable for chloroplast development must be present as alternative Trx-z targets. Therefore, systematic determination of Trx-z target proteins is an attractive research subject. Besides FLNs, several antioxidant enzymes are known to be the target of Trx-z so far (24, 47), which is, however, unlikely to account for the trxz phenotype. In addition to the present finding of the NADPH/NTRC/Trx-z redox cascade, future identification of its downstream targets will further expand the chloroplast redox network beyond current expectations.

Cooperative Redox Regulation by the FTR/Trx and NTRC Pathways Is Essential for Plants.

What is the significance of chloroplast redox regulation in vivo? To address this question, a number of studies using FTR (48, 49), Trx (24, 35, 50–53), or NTRC (15, 16, 27, 54) mutant plants have been reported even during the last few years. Most of these studies focused on the in vivo function of individual redox-mediator proteins. In the current study, we gained insights into the biological importance of cooperative redox regulation by the FTR/Trx and NTRC pathways. Critical growth phenotypes with lethality under autotrophic conditions in the ftrb ntrc double mutants clearly indicate the requirement for concerted redox regulation by these pathways for plants (Fig. 5). The ftrb ntrc double mutants also exhibited pleiotropic phenotypes of photosynthesis-related parameters (Figs. 6 and 7 and Fig. S8). Therefore, retarded growth in ftrb ntrc appears to result from inactivation of multiple layers of chloroplast biological processes.

In particular, impaired biosynthetic processes of the photosynthetic machineries, including chlorophyll synthesis and construction of the electron transport complexes, are thought to be the primary factors for the severe phenotype of ftrb ntrc. The pale-green leaf phenotype was observed in ntrc and ftrb ntrc but not in ftrb (Fig. 5 and Fig. S7). These results suggest that NTRC predominantly functions in the redox regulation of chlorophyll synthesis. Our biochemical data showing efficient CHLI reduction by NTRC support this possibility (Fig. 2 C and D and Fig. S4 C, D, G, and H). Other proteins involved in tetrapyrrole metabolism may be specifically redox-regulated by NTRC (55). However, given the Trx ability to reduce CHLI in vitro (Fig. 2 E and F) (22) and the physical interaction between Trx-f and CHLI in vivo (33), the FTR/Trx pathway is thought to contribute to CHLI redox regulation possibly in a redundant manner. It is highly plausible that the severely chlorotic phenotype of ftrb ntrc is a consequence of disrupted redox regulation of chlorophyll biosynthesis via both the FTR/Trx and NTRC pathways. A similar scenario can be considered for regulation of chloroplast gene expression. NTRC was determined to serve as a reducer of Trx-z that is required for PEP-dependent chloroplast gene expression (Figs. 1 and 4). However, a less severe phenotype of ntrc than that of trxz (Fig. 5 and Fig. S9) suggests that NTRC is unlikely to be the sole electron donor to Trx-z. In accordance, Trx-z can be reduced by FTR directly or indirectly via other Trxs in vitro (Fig. 8) (40, 47). Therefore, additional loss of FTR in the ntrc mutant background might critically impair Trx-z reduction and, in turn, PEP-dependent gene expression. This hypothesis is in line with the drastic declines in D1 and PsaA protein accumulation levels in ftrb ntrc (Fig. 6B and Fig. S8A); these proteins are encoded by PsbA and PsaA genes, respectively, which are under PEP-dependent regulation (42). In contrast to other Trxs (Fig. 7 and Fig. S8D), the Trx-z redox state in vivo could not be determined in this study, at least partly due to the low expression level in mature leaves (40, 52, 56) and the smeared nature of the protein band in the immunoblotting analysis (Figs. S8B and S9C). Elucidation of the Trx-z redox dynamics is highly worth trying, especially with a focus on the chloroplast development phase (see also below).

The above-described regulation is important during plastid differentiation, but the chloroplast redox regulation system is also dynamically working in mature leaves and enables light-responsive control of chloroplast functions (29). Unexpectedly, photoreduction responses of Trxs and redox-regulated proteins were still largely maintained in the ftrb mutant (Fig. 7 and Fig. S8D). FTR proteins partially accumulated in ftrb (Figs. S7E and S8 B and C), which might be enough to achieve substantial reducing power transfer from the electron transport chain to Trxs and their target proteins. Impaired NTRC expression affected the photoreduction behaviors of redox-regulated proteins (Fig. 7 and Fig. S8D). Most remarkably, photoreduction responses of FBPase and SBPase were drastically impaired in the ftrb ntrc double mutants. These results raise the possibility that NTRC is involved in light-responsive redox regulation of FBPase and SBPase in vivo. Very recently, Thormählen et al. suggested a similar possibility, based on their observation of lowered reduction level and activity of FBPase in Arabidopsis double mutants defective in Trx-f1 and NTRC (36). However, these data should be interpreted with caution. The ntrc mutants (including ntrc single and ftrb ntrc double mutants) exhibited multiple concurrent phenotypic traits, as represented by lowered chlorophyll content and impaired photosynthetic electron transport (Fig. 6). These changes directly affect the conversion of light energy to reducing power. It is therefore highly possible that impaired photoreduction responses observed in the ntrc mutants resulted from suppression of reducing power generation itself rather than its transfer to the targets. Our biochemical studies showing that NTRC (even at high concentrations) fails to reduce FBPase or SBPase (Fig. 3C) underpin this idea.

Concluding Remarks

We have clarified that two redox systems via the FTR/Trx and NTRC pathways cooperatively control a range of chloroplast functions by fulfilling differential regulatory roles, which is critical for plant survival. Environmental and developmental factors are likely to change the in vivo activities of these two pathways. The FTR/Trx pathway is exclusively dependent on light-driven photosynthetic electron transport, whereas NTRC can function even under dark conditions by using metabolically generated NADPH (Fig. 8). NTRC activity independent of electron transport is evident from observation of the in vivo redox state of 2-Cys Prx, a preferred target of NTRC; 2-Cys Prx was partly present in the reduced form in wild-type plants under dark conditions, whereas it was nearly completely oxidized in ntrc mutants (Fig. 7 and Fig. S8D). This characteristic of NTRC is likely to be important at an early phase of chloroplast development, because the FTR/Trx pathway is unable to function at this stage before electron transport system construction. Therefore, it seems physiologically relevant that NTRC efficiently transfers reducing power to CHLI (involved in chlorophyll synthesis; Fig. 2 C and D and Fig. S4 C, D, G, and H) and Trx-z (involved in chloroplast gene expression; Fig. 4). After chloroplast maturation, the FTR/Trx pathway contribution to overall redox regulation is likely to become more dominant, because this pathway allows reductive activation of metabolic enzymes such as FBPase, SBPase, and NADP-MDH, whereas NTRC does not (Fig. 3C). Furthermore, the glutathione system is likely to participate in the chloroplast redox network in vivo, as reported in the cytosol (57–59). For example, the stable redox state of PrxQ even in ftrb ntrc (Fig. 7 and Fig. S8D) may be linked to a functional backup by this system, as PrxQ is a target candidate of the glutathione/glutaredoxin system (60). Other factors such as the variation and/or severity of environmental stresses possibly affect the manner of chloroplast redox regulation. Future studies from an ecophysiological perspective would provide valuable insights into chloroplast redox network dynamics.

Materials and Methods

Preparation of Expression Plasmids.

Total RNA was isolated from Arabidopsis thaliana as described previously (61) and used as a template for RT-PCR. Gene fragments encoding the mature protein region (determined by Edman sequencing or predicted by TargetP; www.cbs.dtu.dk/services/TargetP/) of NTRC (At2g41680), 2-Cys Prx A (At3g11630), and CHLI1 (At4g18480) were cloned into the pET-23c expression vector (Novagen). The NTRC and CHLI1 plasmids were designed to express His-tagged proteins at the N and C termini, respectively. A plasmid for EcNTR was designed to express His-tagged protein at the C terminus and constructed using the pET-23c expression vector and DNA from E. coli strain BL21 (DE3) as a template for PCR. Plasmids for other proteins used in this study were prepared previously (24). All point mutations were introduced using the PrimeSTAR Mutagenesis Basal Kit (Takara) according to the manufacturer’s instructions. Primers used for plasmid construction and site-directed mutagenesis are listed in Table S1.

Table S1.

Oligonucleotides used in this study

Name	Sequence, 5′ to 3′	Purpose
NTRC Fw (NdeI)*	AACTGCAGCATATGCACCACCACCACCACCACGCCACC-	Plasmid construction
	GCCAATTCTCCGTC
NTRC Rev (BamHI)*	CTGCAGGGATCCTCATTTATTGGCCTCAATGAATTCTCGG	Plasmid construction
2-Cys Prx A Fw (NdeI)*	AACTGCAGCATATGGCCGATGATCTTCCACTGGTTG	Plasmid construction
2-Cys Prx A Rev (EcoRI)*	GCGAATTCAAATAGCTGAGAAGTACTCT	Plasmid construction
CHLI1 Fw (NdeI)*	AACTGCAGCATATGGTTATGAATGTAGCCACTGA	Plasmid construction
CHLI1 Rev (XhoI)*	AATCGGCTCGAGGCTGAAAATCTCGGCGAACT	Plasmid construction
NTRC NTRd Rev (BamHI)*	CTGCAGGGATCCTCAAAATTCAACAAGAAGATTGT	Plasmid construction
NTRC TRXd Fw (NdeI)*	AACTGCAGCATATGCACCAGCCTCAAACTGAAGAG	Plasmid construction
EcNTR Fw (NdeI)*	AACTGCAGCATATGGGCACGACCAAACACAGTAAAC	Plasmid construction
EcNTR Rev (XhoI)*	AATCGGCTCGAGTTTTGCGTCAGCTAAACCATC	Plasmid construction
NTRC C217S Fw†	AGTGCTTCTGCTATCTGTGATGGAGCT	Site-directed mutagenesis
NTRC C217S Rev†	GATAGCAGAAGCACTTATCCCCCTACT	Site-directed mutagenesis
NTRC C217S/C220S Fw†	GCTTCTGCTATCTCTGATGGAGCTTCGCCTTTA	Site-directed mutagenesis
NTRC C217S/C220S Rev†	AGAGATAGCAGAAGCACTTATCCCCCTACTCCA	Site-directed mutagenesis
NTRC C454S Fw†,‡	CCAACATCTGGCCCCTCTAGGACTCTG	Site-directed mutagenesis
NTRC C454S Rev†	GGGGCCAGATGTTGGTGAAGTGTATAG	Site-directed mutagenesis
NTRC C457S Fw†	GGCCCCTCTAGGACTCTGAAGCCTATT	Site-directed mutagenesis
NTRC C457S Rev†	AGTCCTAGAGGGGCCACATGTTGGTGA	Site-directed mutagenesis
Trx-f1 C102S Fw†	GGTCCATCTAAAGTGATTGCCCCTAAA	Site-directed mutagenesis
Trx-f1 C102S Rev†	CACTTTAGATGGACCACACCATTGAGT	Site-directed mutagenesis
CHLI1 C102S Fw†	AAGTTATCTCTTTTGTTGAATGTTATT	Site-directed mutagenesis
CHLI1 C102S Rev†	CAAAAGAGATAACTTCATCTCATCTTG	Site-directed mutagenesis
CHLI1 C193S Fw†	AGAGTTTCTGGAACCATCGATATCGAA	Site-directed mutagenesis
CHLI1 C193S Rev†	GGTTCCAGAAACTCTATCTTCTGTTGC	Site-directed mutagenesis
FTRB Fw	TGGCGAGCATTGGTTTCTGG	Genomic PCR, RT-PCR
FTRB Rev	TTTCGTGTAAAGTTGGATCC	Genomic PCR, RT-PCR
GK-LB	ATATTGACCATCATACTCATTGC	Genomic PCR
NTRC Fw	GCTGCGTCTCCCAAGATAGG	Genomic PCR, RT-PCR
NTRC Rev	GTTGTTTTGTAAAATCTTAAAGC	Genomic PCR, RT-PCR
Trx-z Fw	TCTAAAACCCCTCAAAAGCT	Genomic PCR
Trx-z Rev	AATGACGTCATATCTCTCTC	Genomic PCR
Salk-LB2R	GACCGCTTGCTGCAACTCTCTCA	Genomic PCR
FTRA1 Fw	GAATCAGTTTTATTCCACTTGC	RT-PCR
FTRA1 Rev	AAGATCAAGCCAATGAATCAAATC	RT-PCR
FTRA2 Fw	TTCTCTCACACTTTCTTAAG	RT-PCR
FTRA2 Rev	AAAGTACATACAAACAAACC	RT-PCR
18s rRNA Fw	CTGCCAGTAGTCATATGCTT	RT-PCR
18s rRNA Rev	ACTACGGTTATCCGAGTAGT	RT-PCR
FTRB qPCR Fw	TCGCTCTGGGACTTACTTCTG	qPCR
FTRB qPCR Rev	GGCACTCCTTCCTCTCTCTC	qPCR
RPS15aA qPCR Fw	GAAGCACGGTTACATTGGTG	qPCR
RPS15aA qPCR Rev	TGTCTAGAAGGGAGCAAACG	qPCR

^*Restriction sites for the enzyme shown in parentheses are underlined.

^†Mutated bases are underlined.

^‡This primer was designed for construction of the NTRC_C454S/C457S plasmid using the NTRC_C457S plasmid as template.

Protein Expression and Purification.

Each expression plasmid was transformed into E. coli strain Rosetta (DE3) (for NTRC and CHLI1) or BL21 (DE3) (for 2-Cys Prx A and EcNTR). Transformed cells were cultured at 37 °C. Expression was induced by the addition of 0.5 mM isopropyl β-d-1-thiogalactopyranoside followed by overnight culture at 21 °C. Cells were disrupted by sonication. After centrifugation (125,000 × g for 40 min), the resulting supernatant was used to purify the protein. His-tagged NTRC, CHLI1, and EcNTR proteins were purified using the Ni-nitrilotriacetic acid–affinity column as described previously (29). 2-Cys Prx A protein was purified by a combination of anion-exchange chromatography and hydrophobic-interaction chromatography as described previously (24). Each mutant protein was prepared in a similar manner to the wild type. All purification procedures were performed at 4 °C. Protein concentration was determined with a BCA protein assay (Pierce).

Target Protein Screening of NTRC and Trx-f1 from Spinach Chloroplasts.

The method of Trx-affinity chromatography (7) was applied to screen for targets of NTRC or Trx-f1. Chloroplast soluble proteins were obtained by collecting intact chloroplasts from spinach leaves and disrupting these by osmotic stress. Subsequent procedures were similar to those described previously (62).

Visualization of Protein Redox Shifts Using Thiol-Labeling Reagents.

The redox state of a protein was determined by discriminating thiol status with the use of thiol-modifying reagents as described previously (7, 23, 24). The thiol-modifying reagents 4-acetamido-4′-maleimidylstilbene-2,2′-disulfonate (molecular mass 536.44) and maleimide-PEG₁₁-biotin (molecular mass 922.09) lower protein mobility on nonreducing SDS/PAGE, allowing the determination of redox state as the band shift. Oxidized proteins of interest were incubated with the indicated redox-mediator proteins (NTRC, EcNTR, and/or Trx) and reductants (0.5 mM NADPH or 0.5 mM DTT) in medium containing 50 mM Tris⋅HCl (pH 7.5 unless otherwise specified) and 50 mM NaCl. After incubation for 30 min (unless otherwise specified) at 25 °C, proteins were precipitated with 10% (vol/vol) trichloroacetic acid and then washed with ice-cold acetone. Precipitated proteins were labeled with thiol-modifying reagents as described in each figure. Then, proteins were subjected to nonreducing SDS/PAGE and stained with Coomassie Brilliant Blue R-250 (CBB).

Spectrophotometric Monitoring of Reducing Power Transfer.

The NADPH-dependent 5,5′-dithiobis(2-nitrobenzoic acid) (DTNB) reduction activities of NTRC and EcNTR (±Trx) were monitored as the increase in absorbance at 412 nm due to generation of 5-thio-2-nitrobenzoic acid (TNB) in medium containing 100 mM Tris⋅HCl (pH 7.5), 50 mM NaCl, 1 mM EDTA, 5 mM DTNB, and indicated concentrations of NTRC or EcNTR (±Trx). Reactions were initiated by the addition of 0.2 mM NADPH. Electron transfer from NADPH to H₂O₂ via NTRC and 2-Cys Prx A or PrxQ was monitored as the decrease in absorbance at 340 nm due to NADPH oxidation in medium containing 50 mM Tris⋅HCl (pH 7.5), 50 mM NaCl, 0.2 mM NADPH, indicated concentrations of NTRC, and 20 μM 2-Cys Prx A or PrxQ. Reactions were initiated by the addition of 1 mM H₂O₂. All spectrophotometric measurements were performed at 25 °C.

Determination of the Midpoint Redox Potential of NTRC.

To determine the midpoint redox potential of individual Cys pairs conserved in the NTR and Trx domains of NTRC, recombinant proteins of truncated NTRC (NTR domain: Ala⁶⁸–Phe⁴⁰⁰; Trx domain: His⁴⁰¹–Lys⁵²⁹) were prepared as described above. Primers used for plasmid construction are listed in Table S1. The procedures for determining the midpoint redox potential were described previously (24).

Plant Materials.

A. thaliana wild-type plants (Col-0) and homozygous T-DNA insertion mutants in FTRB (At2g04700), NTRC (At2g41680), or Trx-z (At3g06730) genes (ftrb, GK-686B09; ntrc-1, Salk_114293C; ntrc-2, Salk_012208C; trxz, Salk_028162C) were used in this study. Each homozygous mutant was backcrossed to the wild type and isolated again from the F₂ generation. The ftrb ntrc double mutants (ftrb ntrc-1 and ftrb ntrc-2) were obtained by crossing each single mutant and screening from the F₂ generation. Genotyping was performed by genomic PCR using the primers listed in Table S1. Plants were grown in soil or half-strength MS medium in the absence or presence of 2% (wt/vol) sucrose in a controlled growth chamber (70 μmol photons m⁻² s⁻¹, 22 °C, 60% relative humidity, 16 h day/8 h night).

Transcript-Level Analyses.

Total RNA was extracted from Arabidopsis wild-type and mutant plants, and each cDNA was obtained from the extracted RNA by reverse transcription as described above. FTRB transcript level was measured using the Mx3000P qPCR System (Agilent Technologies). THUNDERBIRD SYBR qPCR Mix (Toyobo) was used for PCR reactions according to the manufacturer’s instructions. RPS15aA (At1g07770) was used for the internal control gene. Primers used for transcript-level analyses are listed in Table S1.

Immunoblotting.

Leaf total protein was extracted as described previously (63). Proteins were separated by SDS/PAGE and transferred to a PVDF membrane. Antibodies against the FTR catalytic subunit, FTR variable subunit, NTRC, and 2-Cys Prx were newly prepared using each recombinant protein as the antigen. Antibodies against proteins involved in tetrapyrrole metabolism and cyclic electron transport around PSI were kindly provided by Tatsuru Masuda, The University of Tokyo, Tokyo, and Toshiharu Shikanai, Kyoto University, Kyoto, respectively. Other antibodies used in this study were prepared previously (24, 29) or were commercially available (Agrisera).

Chlorophyll Content Measurements.

Chlorophyll content and chlorophyll a/b ratio were determined spectrophotometrically after extraction with 80% (vol/vol) acetone according to the method of Porra et al. (64).

Chlorophyll Fluorescence Measurements.

Chlorophyll fluorescence was measured with a pulse-amplitude modulation (PAM) fluorometer (MINI-PAM; Walz). Actinic light intensity was elevated from low to high levels in a stepwise manner. Chlorophyll fluorescence parameters were calculated as described previously (65).

Determination of in Vivo Protein Redox State.

Plants were placed at the indicated light intensities for 15 min at 25 °C and used for the determination of in vivo protein reduction level as described previously (29).