Thioredoxin Redox Regulates ATPase Activity of Magnesium Chelatase CHLI Subunit and Modulates Redox-Mediated Signaling in Tetrapyrrole Biosynthesis and Homeostasis of Reactive Oxygen Species in Pea Plants^,,,

Abstract

The chloroplast thioredoxins (TRXs) function as messengers of redox signals from ferredoxin to target enzymes. In this work, we studied the regulatory impact of pea (Pisum sativum) TRX-F on the magnesium (Mg) chelatase CHLI subunit and the enzymatic activation of Mg chelatase in vitro and in vivo. In vitro, reduced TRX-F activated the ATPase activity of pea CHLI and enhanced the activity of Mg chelatase reconstituted from the three recombinant subunits CHLI, CHLD, and CHLH in combination with the regulator protein GENOMES UNCOUPLED4 (GUN4). Yeast two-hybrid and bimolecular fluorescence complementation assays demonstrated that TRX-F physically interacts with CHLI but not with either of the other two subunits or GUN4. In vivo, virus-induced TRX-F gene silencing (VIGS-TRX-F) in pea plants did not result in an altered redox state of CHLI. However, simultaneous silencing of the pea TRX-F and TRX-M genes (VIGS-TRX-F/TRX-M) resulted in partially and fully oxidized CHLI in vivo. VIGS-TRX-F/TRX-M plants demonstrated a significant reduction in Mg chelatase activity and 5-aminolevulinic acid synthesizing capacity as well as reduced pigment content and lower photosynthetic capacity. These results suggest that, in vivo, TRX-M can compensate for a lack of TRX-F and that both TRXs act as important redox regulators of Mg chelatase. Furthermore, the silencing of TRX-F and TRX-M expression also affects gene expression in the tetrapyrrole biosynthesis pathway and leads to the accumulation of reactive oxygen species, which may also serve as an additional signal for the transcriptional regulation of photosynthesis-associated nuclear genes.

Thioredoxins (TRXs) are small (approximately 12 kD), ubiquitous oxidoreductases that mediate the dithiol-disulfide exchange of Cys residues, thereby modulating the function and stability of their target proteins (Schürmann and Buchanan, 2008). The short peptide motif WC(G/P)PC, with its two conserved Cys residues, has been characterized as the conserved redox-active site of TRXs (Jacquot et al., 1997; Buchanan and Balmer, 2005; Meyer et al., 2008). In chloroplasts, the disulfide bonds of TRXs can be reduced by ferredoxin-TRX reductase receiving the receipt of electrons from the photoreduced ferredoxin. Subsequently, the reduced TRX acts to reduce the disulfide groups of its target proteins (Dai et al., 2000, 2007; Schürmann and Buchanan, 2008). The five different classes of chloroplast TRXs have been characterized as TRX-F, TRX-M, TRX-X, TRX-Y, and TRX-Z in Arabidopsis (Arabidopsis thaliana). The two classes of TRX-M and TRX-F consist of four and two separate members, respectively, and are well-known and important modulators of the activity of several chloroplast enzymes involved in the Calvin-Benson cycle (Zhang and Portis, 1999; Balmer et al., 2001; Miginiac-Maslow and Lancelin, 2002; Sparla et al., 2002; Samra et al., 2006). TRX-X has recently been reported as a reductant of chloroplast 2-Cys peroxiredoxins (2-Cys Prx; Collin et al., 2003; Pulido et al., 2010), a central element of the antioxidant defense system and the dithiol-disulfide redox-regulatory network of the plant cell (Pulido et al., 2010; Dietz, 2011). TRX-Y exists in two isoforms that are involved in oxidative stress responses (Collin et al., 2004), and TRX-Z regulates plastid-encoded RNA polymerase-dependent transcription (Arsova et al., 2010).

The magnesium (Mg) chelatase CHLI subunit has been reported to be one of the chloroplast-located target proteins controlled by TRX (Balmer et al., 2003; Ikegami et al., 2007). Jensen et al. (2000) analyzed the individual subunits of Mg chelatase for their sensitivity to the thiol modifier N-ethylmaleimide. Modification of the Cys residues in CHLI and CHLH results in enzyme inactivation, whereas the Cys residues in CHLD are not essential for Mg chelatase activity (Jensen et al., 2000). Moreover, the Mg chelatase activity of recombinant Synechocystis subunits and of protein extracts of Arabidopsis chloroplasts was enhanced by dithiothreitol (DTT; Jensen et al., 2000; Ikegami et al., 2007), suggesting that Mg chelatase is likely redox regulated in chloroplasts. A recent study indicated that spinach (Spinacia oleracea) TRX-F reduces the internal disulfide bond (Cys₃₅₄–Cys₃₉₆) of Arabidopsis CHLI1, which is the main CHLI isoform, and stimulates its ATPase activity (Ikegami et al., 2007).

In plants, tetrapyrroles are the most abundant biomolecules, serving as cofactors for many apoproteins essential for plant development (Moulin and Smith, 2005). Chlorophyll is the predominant tetrapyrrole end product and is essential for photoautotrophic growth. By capturing solar energy and transforming it into a charge separation in the reaction center, chlorophyll initiates the redox reactions of the photosynthetic electron transport chain. The tetrapyrrole biosynthesis pathway begins with activated Glu in the form of glutamyl-tRNAglu, which is processed to the common precursor of all tetrapyrroles, 5-aminolevulinic acid (ALA). Eight ALA molecules form a porphyrin, which is introduced into either the Mg or the iron branch, leading to chlorophyll or heme synthesis, respectively. The ATP-dependent heterotrimeric Mg chelatase is the first enzyme of the chlorophyll biosynthesis branch and consists of the three subunits CHLI, CHLD, and CHLH (Gibson et al., 1995; Jensen et al., 1996a; Petersen et al., 1998; Luo et al., 1999). GENOMES UNCOUPLED4 (GUN4) has recently been identified as an activator of Mg chelatase that interacts with CHLH, binds protoporphyrin IX and Mg-protoporphyrin IX, and controls the flow of substrates into the Mg branch (Larkin et al., 2003; Wilde et al., 2004; Peter and Grimm, 2009; Adhikari et al., 2011). Moreover, our latest study revealed that the C-terminal residues of GUN4 are required to activate the CHLH subunit (Zhou et al., 2012).

To further examine the TRX-dependent regulation of Mg chelatase, we analyzed the effects of pea (Pisum sativum) TRX-F on CHLI and the activation of Mg chelatase in vitro and provided experimental evidence that the Mg chelatase activity is redox regulated by chloroplast TRXs in vitro and in vivo. In addition, the physiological consequences of the reduced content of chloroplast TRXs in the tetrapyrrole biosynthesis pathway and reactive oxygen species (ROS) signaling were explored by means of virus-induced gene silencing (VIGS).

RESULTS

Enzymatic Enhancement of Pea TRX-F for the ATPase Activity of Pea CHLI and the Activity of Reconstituted Mg Chelatase

After oxidation with 50 μm CuCl₂, CHLI loses its potential to hydrolyze ATP (50 μm CuCl₂, Fig. 1A), whereas upon reduction with 5 mm DTT, CHLI retains its strong capacity for ATP hydrolysis (5 mm DTT, Fig. 1A). Incubation of CHLI with 10 μm DTT results in a lower ATPase activity (10 μm DTT, Fig. 1A). In the presence of TRX-F, the ATPase activity of CHLI is elevated, suggesting that TRX-F preserves CHLI to a greater extent in the reduced state (10 μm DTT + 3 μm TRX-F, Fig. 1A).

Figure 1.

TRX-F enhances the ATPase activity of CHLI and Mg chelatase activity. A, The ATPase activity of pea CHLI was elevated in vitro with pea TRX-F. One micromolar recombinant pea CHLI was oxidized with 50 μm CuCl₂ or reduced by different concentrations of DTT with and without 3 μm pea TRX-F. B, TRX-F enhances the reconstituted Mg chelatase in vitro. Recombinant pea CHLI (0.5 μm) was untreated (Untreated), oxidized by 50 μm CuCl₂, and reduced by different concentrations of DTT with and without 1.5 μm pea TRX-F and then combined with recombinant rice CHLD (0.125 μm), CHLH (3 μm), and GUN4 (3 μm) to reconstitute the Mg chelatase activity containing 2 mm ATP together with a ATP regenerated system (2 unit creatine kinase and 10 mm phosphocreatine) and 30 mm MgCl₂ as indicated by a stopped fluorometric assay (Guo et al., 1998). The error bars represent nine repeats from a total of three independent experiments. -GUN4, pea CHLI, and TRX-F together with rice CHLD and CHLH containing 1 mm DTT are reconstituted in the Mg chelatase assay.

As pea TRX-F stimulates pea CHLI ATPase activity, we assayed the activity of reconstituted Mg chelatase with recombinant pea CHLI, rice (Oryza sativa) CHLD, rice CHLH, rice GUN4, and pea TRX-F proteins purified from Escherichia coli extracts. High Mg chelatase activity was detectable upon subjecting these recombinant Mg chelatase subunits and GUN4 to the enzyme assay in the presence of 1 mm DTT (catalytic activity: 3.298 ± 0.211 pmol Mg-deuteroporphyrin IX min⁻¹, 1 mm DTT; Fig. 1B). However, the reconstituted Mg chelatase was entirely inactivated upon oxidation with 50 μm CuCl₂ (catalytic activity: 0.015 ± 0.008 pmol Mg-deuteroporphyrin IX min⁻¹, 50 μm CuCl₂; Fig. 1B). The basal Mg chelatase activity, without any redox-dependent changes in our assays, was 0.151 ± 0.022 pmol Mg-deuteroporphyrin IX min⁻¹ (Untreated, Fig. 1B). Treatment with 10 μm DTT caused the activity of the same amount of recombinant proteins to increase by a factor of 2.5 (0.368 ± 0.029 pmol Mg-deuteroporphyrin IX min⁻¹, 10 μm DTT; Fig. 1B). Importantly, however, upon the addition of 1.5 μm TRX-F, a 12-fold increase in Mg chelatase activity was achieved (1.77 ± 0.139 pmol Mg-deuteroporphyrin IX min⁻¹, 10 μm DTT + 1.5 μm TRX-F; Fig. 1B). It is notable that the combination of recombinant pea and rice subunits leads to significant enzyme activity.

The Interaction between Pea TRX-F and Pea CHLI in Vivo

We aimed to verify the protein-protein interaction between TRX-F and CHLI by employing the yeast two-hybrid system. Cotransformation of the yeast (Saccharomyces cerevisiae) strain AH109 with pGBKT7-CHLI and pGADT7-TRX-F resulted in a positive protein-protein interaction between these two proteins, as demonstrated by visualized α-galactosidase activity and growth on synthetic dropout/-Leu/-Trp/-His/-Ade plates (Fig. 2A). TRX-F neither interacted with the other two subunits of the Mg chelatase nor with GUN4 (Fig. 2A).

Figure 2.

Analysis of the interaction between pea TRX-F and CHLI. A, Physical interaction was shown between TRX-F and CHLI in yeast-two-hybrid assays. p53, Murine p53; T, SV40 large T-antigen. p53 and large T-antigen interact in a yeast-two-hybrid assay (positive control). TRX-F, CHLI, and CHLD are from pea and CHLH, GUN4 are from rice. B, Visualization of protein-protein interaction between pea TRX-F and CHLI in chloroplast by the BiFC assay. Tobacco leaves were transformed by Agrobacterium-mediated infiltration of gene constructs encoding the fusion proteins. Chloroplasts were isolated from leaves after transient expression of transgenes and YFP fluorescence of chloroplasts was visualized by confocal laser-scanning microscopy. The right sections (Merge) indicate an overlay of the YFP (representative of the interaction that occurs between the two proteins) and chlorophyll autofluorescence images. Each image is representative of at least three independent experiments.

An additional experiment providing evidence of the pairwise interaction of TRX-F and CHLI in chloroplasts was performed by means of a bimolecular fluorescence complementation (BiFC) assay. The leaf infiltration method was applied for transient expression of gene constructs after Agrobacterium-mediated infiltration of 4- to 6-week-old Nicotiana benthamiana plants. The chloroplasts were isolated from N. benthamianat plant leaf cells infiltrated with construct combinations of pDEST-CHLI-^GWVYNE and pDEST-TRX-F-^GWVYCE or pDEST-CHLI-^GWVYCE and pDEST-TRX-F-^GWVYNE. These cells transiently expressed Venus^N/C-fused full-length TRX-F and CHLI, and the interaction of the two proteins was indicated by yellow fluorescent protein (YFP) fluorescence (Fig. 2B). These results support the interaction of CHLI and TRX-F that was demonstrated by the yeast two-hybrid system and further validate the previously published findings regarding the TRX-dependent reduction of CHLI (Balmer et al., 2003; Ikegami et al., 2007).

Pale-Green Phenotype of VIGS-TRX-F/TRX-M Plants, But No Phenotypic Difference in VIGS-TRX-F Plants

By means of the VIGS method using the pea early browning virus (PEBV) vector (Constantin et al., 2004), we silenced the TRX-F gene in pea plants. A 390-bp cDNA fragment of pea TRX-F (160–549 bp) was inserted into the pCAPE2 vector (pCAPE2-TRX-F; Supplemental Fig. S1). After infection, all leaves in the VIGS-TRX-F plants (Fig. 3, B and D) and VIGS-GFP plants (a negative control, see Fig. 3, A and D) remained green. Analysis of the expression by quantitative real-time (RT)-PCR revealed that the TRX-F gene expression was reduced by approximately 90% in the VIGS-TRX-F plants (Fig. 3E). The reduced TRX-F expression had no negative impact on chlorophyll biosynthesis or accumulation (Table I). The Mg chelatase activity of the VIGS-TRX-F plants was not attenuated relative to the activity observed in the VIGS-GFP plants (Table I). Thus, as the absence of TRX-F does not impair Mg chelatase activity in vivo, we hypothesized that TRX-F function could be complemented under our growth conditions. This idea was supported by previously published reports suggesting the CHLI protein as a putative target protein of TRX-F and TRX-M (Balmer et al., 2003) and demonstrating slight activation of the in vitro ATPase activity of CHLI by TRX-M (Ikegami et al., 2007). Therefore, we aimed to examine the role of TRX-M in the redox control of CHLI in vivo by simultaneous gene silencing of TRX-F and TRX-M. The TRX-F and TRX-M cDNA fragments were inserted into the pCAPE2 vector (pCAPE2-TRX-F/TRX-M; Supplemental Fig. S1) prior to the Agrobacteria-mediated infiltration of pea plants with gene constructs. A phenotype with pale-green leaves (Fig. 3, C and D) was the result of the parallel inactivation of TRX-M and TRX-F. In the VIGS-TRX-F/TRX-M plants, more than 90% of both TRX-F and TRX-M expression was inhibited (Fig. 3E).

Figure 3.

Phenotype of the pea VIGS-TRX-F and VIGS-TRX-F/TRX-M plants. A, The VIGS-GFP plants represent a negative control for the effect of the virus infection. B, The VIGS-TRX-F plants do not show a significant phenotype and resembled the VIGS-GFP control plants. C, A phenotype with pale-green leaves and stem was observed on the VIGS-TRX-F/TRX-M plants. All the plants indicated three independent infiltrations and were observed 3 weeks after infiltration. D, Representative leaflets from VIGS plants. E, The suppression rate of the mRNA level of TRX-F in VIGS-TRX-F plants and, simultaneously, of TRX-F as well as TRX-M in VIGS-TRX-F/TRX-M plants were analyzed by real-time PCR, respectively.

Table I. Pigment contents, F_v/F_m, Mg chelatase activity, and ALA synthesizing capacity in VIGS plants.

FW, Fresh weight.

Pigment Contents, Photosynthesis Efficiency, and Enzyme Activities	VIGS-GFPa	VIGS-TRX-Fa	VIGS-TRX-F/TRX-Ma,b
Chlorophyll, a + b (mg g−1 FW)	3.67 ± 0.21	3.70 ± 0.23	1.35 ± 0.15***
Chlorophyll, a/b	2.88 ± 0.06	2.87 ± 0.09	2.08 ± 0.18***
Carotenoids, x + c (mg g−1 FW)	0.73 ± 0.03	0.74 ± 0.04	0.42 ± 0.04***
(a + b)/(x + c)	5.02 ± 0.20	5.03 ± 0.13	3.18 ± 0.15***
Fv/Fm	0.82 ± 0.01	0.82 ± 0.01	0.68 ± 0.03***
Mg chelatase activity (pmol Mg-Deutero min−1 mg−1 protein)	4.31 ± 0.22	4.34 ± 0.19	2.90 ± 0.14***
ALA synthesizing capacity (nmol ALA h−1 g−1 FW)	194.00 ± 22.21	181.90 ± 15.30	153.00 ± 7.76***

^aData represent the mean ± sd of nine to12 plants from three independent infiltrations.  ^b*** P < 0.0001 by Student’s t test.

Redox State of the CHLI Protein in the Leaves of VIGS Plants

To explore the effects of TRX-F silencing and the simultaneous silencing of TRX-F and TRX-M on the redox regulation of CHLI in vivo, we investigated the in vivo redox state of CHLI in the VIGS plants. Total proteins, which were extracted from unopened leaves at the top of the VIGS plants, were labeled with 4-acetoamido-4’-maleimidylstilbene-2,2’-disulfonate (AMS; Invitrogen). Immunoanalysis of CHLI after SDS-PAGE revealed that entirely reduced CHLI (red) and partially oxidized CHLI (ox1) with elevated molecular mass were detectable in all of the VIGS plants as a result of the AMS-alkylated sulfhydryl groups of CHLI (Fig. 4A). Considerably higher amounts of CHLI in the ox1 form were observed in the VIGS-TRX-F/TRX-M plants than in the other two VIGS plants (ox1, Fig. 4A). In addition, a CHLI protein band corresponding to the entirely oxidized form was detectable in the protein extract of the VIGS-TRX-F/TRX-M plants (ox2, Fig. 4A). We quantified the intensity of the immune-reacting CHLI bands presented in Figure 4A. The relative content of the ox1 and ox2 forms of CHLI is approximately 22% of the total CHLI amount in the VIGS-TRX-F/TRX-M plants. These results indicate that pea plants with simultaneous silencing of TRX-F and TRX-M fail to completely maintain CHLI in the reduced form. Furthermore, total proteins from the VIGS-TRX-F/TRX-M plants were oxidized by 1 mm CuCl₂ and subsequently labeled with AMS to indicate the protein band of the oxidized CHLI (ox1 and ox2, 1 mm CuCl₂, Fig. 4A). However, we found that entirely reduced CHLI was still observed in the CuCl₂-treated sample (red, 1 mm CuCl₂, Fig. 4A). In light of these results and the observation that most of the CHLI proteins remained in the entirely reduced form (red) in the TRX-F/TRX-M-silenced plants (VIGS-TRX-F/TRX-M, Untreated, Fig. 4A), it is reasonable to assume that other proteins of the redox-regulatory network in chloroplasts could auxiliarily contribute to the redox regulation of CHLI in vivo.

Figure 4.

In vivo redox state of CHLI in VIGS plants and the effect of NTRC on Mg chelatase. A, In vivo redox states of CHLI in VIGS plants. Total proteins extracted from leaves of VIGS plants, which grew for 3 weeks after infiltration, were labeled by AMS. The in vivo CHLI was detected by western-blot analysis using the anti-CHLI antibody. red, The entirely reduced form; ox1, the partially oxidized form; ox2, the entirely oxidized form. Each lanes were loaded 50 μg proteins. B, Examination of the interaction between pea CHLI and Arabidopsis NTRC by yeast-two-hybrid assays. p53, Murine p53; T, SV40 large T-antigen. p53 and large T-antigen interact in a yeast-two-hybrid assay (positive control). C, Detection of the effect of NTRC on Mg chelatase activity. The recombinant pea CHLI (0.5 μm) was untreated (Untreated), treated with 1.5 μm Arabidopsis NTRC containing 0.5 mm NADPH (NTRC) and 1 mm DTT (1 mm DTT), respectively. Combined with recombinant rice CHLD (0.125 μm), CHLH (3 μm), and GUN4 (3 μm), the treated CHLI was applied to reconstitute the Mg chelatase activity in vitro by a stopped fluorometric enzyme assay (Guo et al., 1998). The error bars represent nine repeats from three independent experiments. [See online article for color version of this figure.]

In general, the main redox-regulatory network in chloroplasts consists of NADPH-dependent TRX reductase C (NTRC), all plastidal TRXs, plastidal peroxiredoxins, and the glutathione/glutaredoxin system (Kirchsteiger et al., 2009; Pulido et al., 2010; Dietz and Pfannschmidt, 2011; Tovar-Méndez et al., 2011). Recently, the comprehensive analysis of an Arabidopsis knockout mutant for NTRC revealed that NTRC may redox regulate several enzymes and thus modulate the expression of genes involved in chlorophyll synthesis (Stenbaek et al., 2008; Lepistö et al., 2009). Stenbaek and Jensen (2010) reported in their review that NTRC can stimulate the ATPase activity of CHLI in vitro. Therefore, we examined the physical interaction between NTRC and CHLI using the yeast two-hybrid assay and investigated the effect of NTRC on Mg chelatase activity. However, no interaction between NTRC and CHLI was observed (Fig. 4B), and NTRC did not influence the Mg chelatase activity in vitro (Fig. 4C); thus, CHLI may not be a primary target protein of NTRC.

We do not entirely exclude the possibility that residual amounts of TRX-F and TRX-M may be sufficient in the VIGS-silenced plants to largely, but not completely, maintain CHLI in a reduced state. Alternatively, it is proposed that other proteins of the chloroplast redox-regulatory network may engage in the redox control of CHLI in the absence of TRX-F and TRX-M.

Reduction of Mg Chelatase Activity, ALA Synthesizing Capacity, Pigment Contents, and Photosynthetic Capacity in VIGS-TRX-F/TRX-M Plants

The pale-green phenotype of the leaves from the VIGS-TRX-F/TRX-M plants indicates a disruption of chlorophyll accumulation and/or synthesis. In previous studies, the ATPase activity of CHLI was reported to be essential for the activity of Mg chelatase (Ikegami et al., 2007). Our data has demonstrated that pea TRX-F activates reconstituted Mg chelatase in vitro, and VIGS-TRX-F/TRX-M plants exhibited a compromised redox state of CHLI. We thus assayed the Mg chelatase of our VIGS plants. Intact chloroplasts were isolated from developing leaflets at the top of the shoots of VIGS plants and subjected to activity measurements. The Mg chelatase activity was significantly decreased in the VIGS-TRX-F/TRX-M plants compared with the VIGS-GFP control plants (Table I).

Active Mg chelatase is essential for chlorophyll biosynthesis. Impaired expression of genes encoding one of the subunits of Mg chelatase leads to reduced enzyme activity and to a pale-green phenotype (Jensen et al., 1996b; Papenbrock et al., 2000b; Sawers et al., 2006; Zhang et al., 2006). The chlorophyll content in plants that experience a parallel silencing of TRX-F and TRX-M was reduced by 60%, whereas no significant difference in the chlorophyll content was observed in the TRX-F-silenced plants in comparison to the control plants. In addition to this observed reduction in chlorophyll content, the VIGS-TRX-F/TRX-M plants exhibited carotenoid contents that were diminished by one-third compared with the VIGS-TRX-F and VIGS-GFP plants (Table I). In addition, the ALA synthesizing capacity, the rate-limiting step of chlorophyll biosynthesis, is decreased in the VIGS-TRX-F/TRX-M plants compared with the other two VIGS plants (Table I). Chlorophyll fluorescence measurements revealed that the F_v/F_m ratio is diminished in the VIGS-TRX-F/TRX-M plants compared with the control plants. All of these results are consistent with the in vivo redox state of CHLI in the VIGS plants (Fig. 4A), implying that a deficient expression of TRX-F and TRX-M reduces Mg chelatase activity, although TRX-M and other members of the chloroplast redox-regulatory network may supplement the function of TRX-F in VIGS-TRX-F plants. Therefore, we propose that both TRX isoforms are responsible for the reduction of CHLI and thus contribute to the regulation of Mg chelatase activity. Future studies will clarify the extent to which other enzymatic steps of chlorophyll synthesis are directly redox regulated by TRX.

Accumulation of ROS in VIGS-TRX-F/TRX-M Plants

TRX-F and TRX-M serve as the electron donors for Cys-containing proteins and contribute to the reduction of their disulfide bonds. Among the target proteins for these TRX isoforms is 2-Cys Prx, which is involved in peroxide detoxification in the chloroplast during the dark phase (König et al., 2002; Pulido et al., 2010; Dietz, 2011; Tovar-Méndez et al., 2011). TRX-F and TRX-M are able to reduce oxidized 2-Cys Prx (Chi et al., 2008). Therefore, we assumed that silencing of the expression of TRX genes may disturb the metabolic balance of ROS in plant cells. The results of histochemical staining demonstrated that more O₂⁻ and hydrogen peroxide (H₂O₂) accumulate in the leaves of VIGS-TRX-F/TRX-M plants in comparison to our control plants (Fig. 5), confirming that TRX-F and TRX-M not only regulate photosynthesis and chlorophyll synthesis but are also involved in ROS metabolism.

Figure 5.

Visualization of H₂O₂ and O₂⁻ in leaves of VIGS plants by histochemical staining. The VIGS plants grown for 3 weeks after infiltration at normal condition were taken for histochemical staining using 3, 3′-diaminobenzidine (H₂O₂) and NBT (O₂⁻), respectively. The brownish and bluish color represents the accumulated H₂O₂ and O₂⁻, respectively.

The Expression of Photosynthesis-Associated Nuclear Genes in Response to Silencing TRXs

The expression of genes involved in chlorophyll binding, tetrapyrrole biosynthesis, and photosynthesis was explored in the leaves of pea VIGS plants. The top and developing leaves of the VIGS plants were harvested during the first hour of light exposure after the transition from dark to light. From these samples, RNA was isolated and protein extracts were prepared to perform quantitative real-time PCR and western blotting, respectively. The mRNA levels of the examined genes in the VIGS-TRX-F plants did not change, with the exception of a reduced level of ferrochelatase (FECHI) mRNA and a small increase in Mg-protoporphyrin IX methyltransferase (CHLM) transcript content (Fig. 6). By contrast, the transcription of almost all examined genes, except Glutamyl-tRNA reductase (HEMA1), was down-regulated in the VIGS-TRX-F/TRX-M plants (Fig. 6).

Figure 6.

Determination of the mRNA level of nuclear genes in VIGS plants by real-time PCR. The mRNA level of examined genes was normalized to VIGS-GFP plants using the 2^−ΔΔCt method. The data represent the mean ± sd of three independent infiltrations. Transcripts encoding the following enzymes are depicted: CHLD, CHLH, and CHLI, the three subunits of Mg chelatase; CHLM, Mg-protoporphyrin IX methyltransferase; FECHI, ferrochelatase; HEMA1, glutamyl-tRNA reductase; LHCB3, light-harvesting chlorophyll-binding proteins 3 of PSII; RBCS, Rubisco small subunit.

In the plants with silenced TRX-F expression, the levels of most of the analyzed proteins, which contribute to photosynthesis, chlorophyll binding, and tetrapyrrole biosynthesis, were not significantly altered, although GUN4, Glutamyl-tRNA reductase (GLUTR), Chlorophyll a oxygenase (CAO), light-harvesting chlorophyll-binding protein 1 of photosystem I (LHCA1), and CHLM levels were increased while Heme oxygenase (HO1) levels were decreased in the VIGS-TRX-F plants relative to the control VIGS plants (Fig. 7). In the VIGS-TRX-F/TRX-M plants, the content of chlorophyll-binding proteins was reduced, especially for the three analyzed LHCB proteins (Fig. 7). Moreover, in the VIGS-TRX-F/TRX-M sample, most of the examined proteins, such as protoporphyrinogen IX oxidase (PPO), phytochromobilin synthase (HY2), CHLD, CHLI, CHLM, Mg-protoporphyrin IX monomethylester (oxidative) cyclase (CHL27), NADPH-protochlorophyllide oxidoreductase (POR), and Ribulose-1,5-bisphosphate carboxylase oxygenase (Rubisco) small subunit (SSU) displayed a decrease in content relative to the control VIGS plants, whereas the GLUTR, CHLH, and CAO levels were significantly increased (Fig. 7).

Figure 7.

Determination of the contents of photosynthesis-related proteins in VIGS plants by western-blot analysis. Proteins are depicted as follows: LHCA1, light-harvesting chlorophyll-binding proteins 1 of PSI; LHCB1, LHCB5, and LHCB6, light-harvesting chlorophyll-binding proteins 1, 5, and 6 of PSII, respectively. CHLP, Geranylgeranyl reductase; SSU, Rubisco small subunit; CHLI, CHLD, and CHLH, the three subunits of Mg chelatase; CHLM, Mg-protoporphyrin IX methyltransferase; CHL27, subunit of Mg-protoporphyrin IX monomethylester (oxidative) cyclase; POR, NADPH-protochlorophyllide oxidoreductase; CAO, Chlorophyll a oxygenase; GLUTR, glutamyl-tRNA reductase; GSAAT, Glu 1-semialdehyde aminotransferase; PPO, protoporphyrinogen IX oxidase; HO1, heme oxygenase; HY2, phytochromobilin synthase; LSU, Rubisco large subunit. The relative levels of examined proteins in VIGS-TRX-F and VIGS-TRX-F/TRX-M were both normalized to those in VIGS-GFP plants and showed below each protein bands. Each lane was loaded with 30 μg of total proteins. [See online article for color version of this figure.]

DISCUSSION

Activity Regulation of CHLI and Mg Chelatase by TRX-F in Vitro

Ikegami et al. identified two disulfide bonds in Arabidopsis CHLI1, and the bond between Cys₃₅₄ and Cys₃₉₆ was found to contribute substantially to the redox-regulated ATPase activity of CHLI1 (Ikegami et al., 2007). Both cysteines are conserved in the pea CHLI homolog. It is conceivable that the corresponding cysteines of pea CHLI are targets of the TRX-mediated redox regulation. Our in vitro data indicate that recombinant pea TRX-F can substantially enhance the ATPase activity of the recombinant pea CHLI (Fig. 1A).

To our best knowledge, a direct involvement of TRX in the activation of Mg chelatase activity has not previously been reported (Ikegami et al., 2007; Stenbaek and Jensen, 2010). Our results (Fig. 1B) provide direct evidence that TRX-F contributes to activating Mg chelatase in vitro. Moreover, in accordance with our previously published results (Zhou et al., 2012), we found that GUN4 is also essential for the in vitro activity of plant Mg chelatase, which was reconstituted with pea CHLI instead of rice CHLI in this study (-GUN4, Fig. 1B).

In this study, the yeast two-hybrid system was applied to demonstrate the interaction of pea TRX-F with pea CHLI (Fig. 2A). Furthermore, the BiFC assay substantiates this result and confirms the interaction of TRX-F and CHLI in the chloroplasts of tobacco leaf cells (Fig. 2B). These data are also consistent with previous reports on the accumulation of both proteins in the stroma of chloroplasts (Gibson et al., 1996; Miller et al., 2004). However, TRX-F does not interact with CHLD, CHLH, or GUN4 in the yeast two-hybrid assays. It is therefore suggested that CHLI is the sole target protein of TRX-F among the Mg chelatase subunits and GUN4.

In Vivo Regulation of CHLI and Mg Chelatase by Chloroplast TRXs

In higher plants, the plastidic TRXs contain a conserved peptide motif as the redox-active site, which consists of the amino acid sequence WCGPC (Supplemental Fig. S2). TRX-F and TRX-M share several common target proteins, but many target proteins of TRX-F interact less efficiently with TRX-M (Balmer et al., 2003). Similarly, both TRX-F and TRX-M stimulate the ATPase activity of CHLI in vitro, but TRX-F is more efficient in this process (Ikegami et al., 2007). Consistent with the in vitro activity analysis, our data revealed that in chloroplasts of the VIGS-TRX-F plants, CHLI was maintained in the reduced state (VIGS-TRX-F, Fig. 4A). Our in planta results substantiate the hypothesis that TRX-M can compensate for TRX-F deficiencies through redox regulation of CHLI in chloroplasts. Based on the in vivo redox state of CHLI (Fig. 4A) and the phenotypes of the VIGS-TRX-F and VIGS-TRX-F/TRX-M plants (Fig. 3), we proposed that TRX-F and TRX-M preferentially redox regulate CHLI, but they are apparently not the sole redox regulators of CHLI. Other members of the chloroplast redox-regulatory network also supplement the redox control of CHLI in plants with reduced TRX-F and TRX-F/TRX-M contents. Therefore, silencing of the expression of both TRX-F and TRX-M genes in pea plants results in approximately 22% of oxidized CHLI with ox1 and ox2 forms in vivo (Fig. 4A), and a 30% decrease in Mg chelatase activity relative to the enzyme activity of the control plants (Table I). It is notable, however, that the reduced Mg chelatase activity of the VIGS plants with silenced TRX-F and TRX-M expression also correlates with a significant decrease in the amounts of CHLI and CHLD (Fig. 7). These results indicate that the TRX-mediated redox regulation of CHLI and the posttranslational control of the CHLD and CHLI stability affect the Mg chelatase activity in vivo.

Modulation of Tetrapyrrole Biosynthesis and Photosynthesis Capacity by Both TRX-F and TRX-M

The essential pathway of tetrapyrrole biosynthesis requires a tight network of regulatory mechanisms. The importance of posttranslational control of tetrapyrrole biosynthesis has been emphasized (Stenbaek and Jensen, 2010; Czarnecki and Grimm, 2012). Our data revealed that TRX-F and TRX-M not only redox regulate the Mg chelatase subunit CHLI but also serve as key regulators that exert a remarkable influence on the gene expression and stability of other proteins in the tetrapyrrole biosynthesis pathway (Figs. 6 and 7). In addition to the proteins involved in photosynthesis, such as the ribulose-1,5-bisphosphate carboxylase oxygenase small subunit (SSU) and light-harvesting chlorophyll-binding proteins (LHC, LHCA1, LHCB1, LHCB5 and LHCB6), the decreased protein contents of CHLD, CHLI, and CHLM correlate with reduced mRNA levels of the corresponding genes (Fig. 6). These results indicate that either TRXs directly regulate the tetrapyrrole biosynthesis pathway at the transcriptional level in the nucleus through plastid-mediated retrograde signaling (Dietz and Pfannschmidt, 2011) or that such regulation can occur as an indirect effect of the lower activity of either CHLI or Mg chelatase resulting from reductions in TRX activity. Interestingly, CHLH accumulates even with reduced levels of corresponding mRNA. Similar results were reported for the Arabidopsis chlm mutant (Pontier et al., 2007). In addition, VIGS-CHLD and VIGS-CHLI plants with reduced Mg chelatase activity also accumulate more CHLH than is observed in the control plants (T. Luo, J. Yu, W. Araújo, H. Schlicke, M. Rothbart, T. Fan, Y. Liu, A. Fernie, B. Grimm, and M. Luo, unpublished data). These findings suggest that CHLH stability inversely correlates to the catalytic rate of Mg chelatase. Nevertheless, the complex control of CHLI and Mg chelatase requires further investigation, as different regulatory mechanisms, such as retrograde signaling and ABA signaling, contribute to transcriptional and posttranslational control of Mg chelatase (Mochizuki et al., 2001; Shen et al., 2006; Wu et al., 2009; Shang et al., 2010; Tsuzuki et al., 2011).

ALA synthesis is the first committed step in tetrapyrrole biosynthesis and is considered to be a rate-limiting step controlling the influx into the entire pathway (Beale, 1999; Papenbrock et al., 2001; Tanaka and Tanaka, 2007). Inactivated expression of CHLI or CHLH in tobacco plants decreased not only Mg chelatase activity but also ALA synthesizing capacity (Papenbrock et al., 2000a, 2000b). In the VIGS-TRX-F/TRX-M plants, the ALA synthesizing capacity was decreased in parallel with the reduced Mg chelatase activity (Table I), which is consistent with observations of feedback-regulated ALA synthesis capacity by Mg chelatase (Papenbrock et al., 2000a). We also found that the protein contents of the ALA biosynthesis enzymes (GLUTR, Fig. 7) were altered in TRX-F/TRX-M-silenced plants. Balmer et al. reported that the Glu-1-semialdehyde aminotransferase catalyzing the final transamination step of ALA synthesis was a TRX-interacting protein (Balmer et al., 2003). Therefore, we suggest that ALA synthesis can also be redox controlled at the posttranslational level by chloroplast TRXs.

Finally, in the VIGS-TRX-F/TRX-M plants, decreased Mg chelatase activity results in reduced amounts of chlorophyll, carotenoids (Table I), and LHC proteins (Fig. 7), thus explaining the pale-green phenotype (Fig. 3C). Consequently, as an indication of compromised photosynthesis capacity, the maximum quantum yield of PSII was lower in the VIGS-TRX-F/TRX-M plants than in the VIGS-GFP plants (Table I).

Down-Regulation of Photosynthesis-Associated Nuclear Genes Transcription by Chloroplast Redox Signals in Response to Simultaneously Silenced TRX-F and TRX-M

The chloroplast is the site of photosynthesis and multiple biochemical reactions. Many of these reactions are controlled by redox signals that are mediated by the photosynthetic electron transport chain (Pfannschmidt et al., 2001; Wagner et al., 2004; Dietz and Pfannschmidt, 2011). These redox signals are generated by photosynthesis-coupled redox-active compounds, such as TRX or glutathione, which are regarded as emitters of retrograde signals that affect the expression of photosynthesis-associated nuclear genes (PhANGs; Fey et al., 2005; Nott et al., 2006; Ruckle et al., 2007; Dietz and Pfannschmidt, 2011). Retrograde signaling represents an important feedback control that couples the expression of PhANGs to the functional state of the chloroplast (Nott et al., 2006; Ruckle et al., 2007). Upon silencing of TRX-F and TRX-M in pea plants, ROS accumulated in the leaf tissue (Fig. 5). Consequently, we assumed that the redox-dependent control was hampered as a result of the TRX deficiency. Additionally, the formation of ROS also appears to be involved in the down-regulation of PhANGs transcription in the VIGS-TRX-F/TRX-M plants (Fig. 6).

In conclusion, the chloroplast TRX-dependent redox system consisting of chloroplast TRXs and TRX reductases provides an important control mechanism for plastidic metabolic pathways. At present, a number of metabolic enzymes are already known to be controlled by this regulatory network. Our findings indicate that CHLI is the primary redox-regulated subunit of the Mg chelatase and, in particular, that CHLI is redox controlled by chloroplast TRXs. Both TRX-F and TRX-M are responsible for maintaining the reduced state of CHLI and the Mg chelatase activity in vivo. Thus, chloroplast TRX-F and TRX-M act as important regulators for the redox control of Mg chelatase. They contribute to the maintenance of steady-state ROS levels and modulate the expression of PhANGs. A proposed mechanism describing the regulatory role of TRX-F and TRX-M is illustrated in Figure 8, with a further explanation in the figure legend.

Figure 8.

A proposed mechanism of the regulator role of chloroplast TRXs. In light, Fd is reduced in the photosynthetic electron transport chain of PSI. Chloroplast TRXs are then reduced by FTR that receives electrons from Fd_red. Thereby, the reduced TRXs facilitate reduction of the CHLI subunit to activate the ATPase activity of CHLI. Subsequently, activated Mg chelatase catalyzes the Mg²⁺ insertion into protoporphyrin IX and chlorophyll biosynthesis continues. If the chloroplast TRXs are absent or inactive in plants, the redox state of CHLI is affected, resulting in decreased Mg chelatase activity. Feedback regulation leads to reduced ALA synthesizing capacity. Under the light, the photosynthetic electron transport chains in thylakoid membranes employ diverse redox cofactors such as iron-sulfur clusters, quinones, and excitable systems in photosynthesis that can generate ROS. These ROS can be scavenged by antioxidant system. But in plants with missing or inactive chloroplast TRXs, the redox state of some enzymes of the oxidative stress defense system, such as 2-Cys Prx, could be modified, leading to the imbalanced ROS metabolism. The chlorophyll biosynthesis pathway that is impaired by lack of chloroplast TRXs, also causes accumulation of ROS. The additional ROS accumulation in the chloroplasts might contribute also to retrograde signaling, which coordinates the expression of PhANGs. ABI4, An AP2-like transcription factor; CHLD, CHLH, and CHLI, Mg chelatase subunits CHLD, CHLH, and CHLI; Fd, ferredoxin; Fd_red, reduced ferredoxin; FTR, ferredoxin-TRX reductase. [See online article for color version of this figure.]

MATERIALS AND METHODS

Plant Material and Growth Conditions

Pea (Pisum sativum ‘Torsdag’; JI992) seeds were cleaned and soaked in tap water for 24 h and then germinated on moist filter paper in the dark for 2 d before being planted in soil. The uninfected plants and the VIGS plants were grown in growth chambers (20°C, 65% relative humidity, 250 μmol m⁻² s⁻¹, 14/10 h light/dark photoperiod).

Expression and Purification of Recombinant Proteins

The pea CHLI (JN198382) and TRX-F (X63537) and the Arabidopsis (Arabidopsis thaliana) NTRC (NM_129731) cDNA sequences encoding the mature proteins (without the putative transit peptides) were amplified by PCR using primers with restriction sites (Supplemental Table S1) and cloned into the EcoRI/SalI, EcoRI/XhoI, and BamHI/SalI sites of the pET28a expression vector to yield the constructs pET28a-CHLI, pET28a-TRX-F, and pET28a-NTRC, respectively. These constructs and the previously published vectors expressing the rice (Oryza sativa) CHLD, CHLH, and GUN4 (Zhou et al., 2012) were introduced into the Escherichia coli strain BL21 by electroporation. Expression was induced by adding 0.1 mm isopropylthio-β-galactoside at 28°C overnight. All of the recombinant proteins were purified through Ni⁺ affinity chromatography according to the manufacturer’s protocol (Bio-Rad), and imidazole in the elution buffer was eliminated by fast ultrafiltration using 10,000 nominal molecular weight limit Amicon Ultra-15 centrifugal filter units (Millipore).

Determination of Enzyme Activity and ALA Synthesizing Capacity

The ATPase activity of CHLI was measured as previously described (Ikegami et al., 2007). For the measurement of Mg chelatase activity, intact chloroplasts were isolated from leaflets in the pea VIGS plants, and Mg chelatase activity was determined by a stopped fluorometric assay as described by Guo et al. (1998). For analysis of the ALA synthesizing capacity (Richter et al., 2010), leaf discs with a diameter of 7 mm from different VIGS plants were incubated in 50 mm Tris-HCl, pH 7.2, and 40 mm levulinic acid for 3 h in growing light.

Yeast Two-Hybrid Assay

The cDNA sequences encoding pea CHLI, CHLD, and TRX-F as well as rice CHLH, GUN4, and Arabidopsis NTRC were generated by PCR amplification using the primers described in Supplemental Table S1 and inserted behind the coding sequence of GAL4-AD or BD domains within the plasmid pGADT7/prey or pGBKT7/bait, respectively. The yeast (Saccharomyces cerevisiae) strain AH109 was cotransformed by the lithium acetate method as described by the manufacturer (Clontech, protocol PT3247-1) with both bait and prey plasmids. The cotransformants were first selected on plates containing double-dropout medium (lacking Leu and Trp) and then scraped onto high-stringency plates containing quadruple-dropout medium (lacking Leu, Trp, His, and Ade) supplemented with 40 mg/L X-a-Gal to screen the interactions.

BiFC Assay

The entire coding regions of pea CHLI and TRX-F were first introduced into the attB1 and attB2 sites, respectively, using two-step PCR with the aforementioned primers (Supplemental Table S1) and then inserted into pDONR-221 using B/P-Clonase (Invitrogen). To generate the BiFC constructs, the entry clones of CHLI and TRX-F in pDONR-221 were used in an gateway attL/attR sites recombination reaction with a Gateway system-compatible version (Invitrogen) of the BiFC vector pDEST-^GWVYNE (Venus amino acids 1–173) and pDEST-^GWVYCE (Venus amino acids 156–239; Gehl et al., 2009). The constructs were introduced into the leaves of Nicotiana benthamiana by Agrobacterium infiltration as described by Gehl et al. (2009). All of the plants were kept in the dark for 2 to 5 d, and transformed cells were analyzed by confocal laser-scanning microscopy (TCS SP2; Leica).

VIGS Assay

The plasmids pCAPE1, pCAPE2-PDS, and pCAPE2-GFP are VIGS vectors based on PEBV. To construct the pCAPE2-TRX-F vector, a 390-bp fragment of pea TRX-F cDNA was generated by RT-PCR with the same primers used to generate the expression vectors in E. coli. To construct the pCAPE2-TRX-F/TRX-M plasmid, a 403-bp fragment of pea TRX-F cDNA linked to a 429-bp fragment of TRX-M was generated by conducting RT-PCR with the primers described in Supplemental Table S1.The pCAPE1 and pCAPE2 derivatives were inoculated into the pea plants by Agrobacterium infiltration as described by Constantin et al. (2004).

Determination of the in Vivo Redox State of CHLI Protein

The tops of unopened leaves from the VIGS plants were ground in liquid nitrogen and then resuspended in 100 mm Tris-HCl (pH 7.5). Then, TCA (final concentration 5%) was added to denature and precipitate the proteins, which were collected by centrifugation, washed with acetone, and dissolved in 1% SDS and 100 mm Tris-HCl (pH 7.5). Total leaf proteins were labeled with AMS (Invitrogen) at 25°C for 30 min followed by a 37°C incubation for 30 min, as previously described by Kobayashi and Ito (1999). Labeled proteins were separated by 12.5% SDS-PAGE without using any reducing agents and were then electrophoretically blotted onto an Immobilon membrane filter (GE Healthcare). The filter was subsequently incubated with anti-CHLI rabbit polyclonal antibodies visualized using the ECL detection kit (GE Healthcare).

Determination of Pigment Contents and F_v/F_m Ratios

To measure the pigment contents, leaf discs were ground in 0.5 mL of 100% acetone at 4°C for 24 h in the dark. Chlorophyll and carotenoids were determined spectrophotometrically according to the Lambert-Beer law (Lichtenthaler, 1987). All of the spectrophotometrical analyses were performed using a Beckman DU-730 spectrophotometer. The maximum quantum yield of PSII photochemistry (calculated as the ratio F_v/F_m = [F_m − F₀]/F_m) was determined as described by Richter et al. (2010).

Measurement of ROS in Leaves

The in situ accumulation of O₂⁻ and H₂O₂ was examined using histochemical staining with nitroblue tetrazolium as described by Hoffmann et al. (2005) and with 3, 3′-diaminobenzidine as described by Thordal-Christensen et al. (1997), respectively.

Gene Expression Analysis

Total RNA was extracted from 40 to 100 mg of leaf tissue using TRIzol reagent (Invitrogen). The RNA was treated with RQ1 RNase-free DNase (Promega). First-strand cDNAs were synthesized from 2 μg of total RNAs with oligo(dT)₁₈ primer using the first-strand cDNA synthesis kit ReverTra Ace-α (TOYOBO). RT was performed at 42°C for 30 min, followed by 30 min at 37°C. Quantitative real-time PCR was performed using Bio-Rad CFX manager. Each 20-μL PCR reaction contained 1 μL of RT reaction buffer, 10 μL of 2×SYBR Green Master Mix (TOYOBO), and 0.5 pm of each primer (Supplemental Table S1). The samples were heated to 95°C for 3 min, followed by 45 cycles of 10 s at 95°C, 10 s at 58°C, and 10 s at 72°C. The transcript levels were quantitatively normalized to the transcript level of pea EF-1α (X96555), which encodes the elongation factor 1-α.

The total leaf proteins were electrophoresed on SDS-polyacrylamide gels and transferred onto a polyvinylidene difluoride membrane (GE Healthcare). After incubation with the primary antibody and HRP-conjugated goat anti-rabbit IgG (Thermo Scientific), the filter was visualized using the ECL detection kit (GE Healthcare). The antibody data are provided in Supplemental Table S2.

Sequence data from this article can be found in the GenBank/EMBL data libraries under accession numbers JN198382 to JN198385.

Supplemental Data

The following materials are available in the online version of this article.

• Supplemental Figure S1. Pea VIGS binary vectors based on PEBV.

• Supplemental Figure S2. Alignment of the amino acid sequences of chloroplast TRXs from Arabidopsis (At) and pea (Ps).

• Supplemental Table S1. PCR primers designed for the experiments in this manuscript.

• Supplemental Table S2. Antibodies applied in western-blot analyses described in this manuscript.

Glossary

TRXs

chloroplast thioredoxins

TRX

chloroplast thioredoxins

Mg

magnesium

2-Cys Prx

2-Cys peroxiredoxins

DTT

dithiothreitol

ALA

aminolevulinic acid

ROS

reactive oxygen species

VIGS

virus-induced gene silencing

BiFC

bimolecular fluorescence complementation

PEBV

pea early browning virus

RT

reverse transcription

AMS

4-acetoamido-4’-maleimidylstilbene-2,2’-disulfonate

NTRC

NADPH-dependent thioredoxin reductase C

H₂O₂

hydrogen peroxide

PhANGs

photosynthesis-associated nuclear genes