Thioredoxin Selectivity for Thiol-based Redox Regulation of Target Proteins in Chloroplasts*

Background: Thioredoxin (Trx) plays a pivotal role in the redox regulation of target proteins. Results: Functional diversity of chloroplast Trxs was determined by observing Trx-dependent redox shifts of several thiol-modulated enzymes in vitro and in vivo. Conclusion: Novel insights into the chloroplast redox network were provided. Significance: Our results shed light on the molecular basis of the light-responsive adjustment of chloroplast functions. Redox regulation based on the thioredoxin (Trx) system is believed to ensure light-responsive control of various functions in chloroplasts. Five Trx subtypes have been reported to reside in chloroplasts, but their functional diversity in the redox regulation of Trx target proteins remains poorly clarified. To directly address this issue, we studied the Trx-dependent redox shifts of several chloroplast thiol-modulated enzymes in vitro and in vivo. In vitro assays using a series of Arabidopsis recombinant proteins provided new insights into Trx selectivity for the redox regulation as well as the underpinning for previous suggestions. Most notably, by combining the discrimination of thiol status with mass spectrometry and activity measurement, we identified an uncharacterized aspect of the reductive activation of NADP-malate dehydrogenase; two redox-active Cys pairs harbored in this enzyme were reduced via distinct utilization of Trxs even within a single polypeptide. In our in vitro assays, Trx-f was effective in reducing all thiol-modulated enzymes analyzed here. We then investigated the in vivo physiological relevance of these in vitro findings, using Arabidopsis wild-type and Trx-f-deficient plants. Photoreduction of fructose-1,6-bisphosphatase was partially impaired in Trx-f-deficient plants, but the global impact of Trx-f deficiency on the redox behaviors of thiol-modulated enzymes was not as striking as expected from the in vitro data. Our results provide support for the in vivo functionality of the Trx system and also highlight the complexity and plasticity of the chloroplast redox network.

Redox regulation based on the thioredoxin (Trx) system is believed to ensure light-responsive control of various functions in chloroplasts. Five Trx subtypes have been reported to reside in chloroplasts, but their functional diversity in the redox regulation of Trx target proteins remains poorly clarified. To directly address this issue, we studied the Trx-dependent redox shifts of several chloroplast thiol-modulated enzymes in vitro and in vivo. In vitro assays using a series of Arabidopsis recombinant proteins provided new insights into Trx selectivity for the redox regulation as well as the underpinning for previous suggestions. Most notably, by combining the discrimination of thiol status with mass spectrometry and activity measurement, we identified an uncharacterized aspect of the reductive activation of NADP-malate dehydrogenase; two redox-active Cys pairs harbored in this enzyme were reduced via distinct utilization of Trxs even within a single polypeptide. In our in vitro assays, Trx-f was effective in reducing all thiol-modulated enzymes analyzed here. We then investigated the in vivo physiological relevance of these in vitro findings, using Arabidopsis wild-type and Trx-f-deficient plants. Photoreduction of fructose-1,6-bisphosphatase was partially impaired in Trx-f-deficient plants, but the global impact of Trx-f deficiency on the redox behaviors of thiol-modulated enzymes was not as striking as expected from the in vitro data. Our results provide support for the in vivo functionality of the Trx system and also highlight the complexity and plasticity of the chloroplast redox network.
Thioredoxin (Trx) 2 is a small ubiquitous protein that plays a crucial role in the thiol-based redox regulation of a specific set of biological systems. Trx has a conserved WCGPC motif at an active site, enabling a dithiol-disulfide exchange reaction with target proteins. In chloroplasts, Trx receives reducing equivalents from the light-driven photosynthetic electron transport chain through ferredoxin-Trx reductase. The reduced form of Trx subsequently transfers reducing equivalents to specific disulfide bonds on target proteins, allowing modulation of enzymatic activities. It has been well documented that this redox cascade ensures a light-responsive control of chloroplast functions (1).
Our understanding of chloroplast thiol-modulated enzymes has grown dramatically since the development of proteomicsbased methodologies for determining Trx-interacting proteins in 2001 (2,3). In addition to classically known light-activated enzymes such as Calvin cycle enzymes (e.g. fructose-1,6-bisphosphatase (FBPase) and sedoheptulose-1,7-bisphosphatase (SBPase)) and NADP-malate dehydrogenase (NADP-MDH) (1), a number of chloroplast proteins have been newly described as Trx target candidates to date (4 -6). Some of them have been further investigated in detail by biochemical and/or reverse genetic studies, providing important indications that the chloroplast Trx system engages in crosstalk with a diverse array of chloroplast functions, including starch synthesis (7), tetrapyrrole metabolism (8,9), and the antioxidant defense system (10 -12). The chloroplast Trx system in plants is also characterized by a large Trx family composed of Trx-f, -m, -x, -y, and -z (13). Trx-f and Trx-m were initially identified as effective activators of FBPase and NADP-MDH, respectively (14,15). Trx-x, Trx-y, and Trx-z were discovered with the increased availability of plant genome information and reported to be efficient in donating reducing equivalents to antioxidant defense systems (10,11,16). Trx-z was also shown to act as a decisive regulator for plastidial transcription (17).
Given the localization of divergent Trx subtypes and Trx target proteins in chloroplasts, it is readily conceivable that chloroplasts host a complex redox network for elaborate regulation of many chloroplast functions. Although functional specificity and redundancy of different Trx subtypes have been partly reported as mentioned above, little consistent description has been achieved. For example, Trx-m was originally defined as the most favorable partner for NADP-MDH activation (14,15), but it was later shown that Trx-f can also activate NADP-MDH with an efficiency comparable with or even higher than that of Trx-m (10,18,19). Such ambiguity may reflect the reality that almost all hypotheses in this field have been based solely on the monitoring of enzymatic activity, and not on the direct observation of the redox state. Furthermore, very few attempts have been made to clarify in vivo Trx selectivity for the redox regulation of target proteins. Consequently, there seems to be a critical gap in the current knowledge of the chloroplast redox network.
To fill this gap, direct observation of Trx-dependent reduction of chloroplast thiol-modulated enzymes is desirable in both biochemical and physiological contexts. In this study, we have clarified some novel aspects of the functional diversity of Trxs and of in vivo redox dynamics in chloroplasts, providing an important step toward advancing the understanding of the chloroplast redox network and of plant strategies for adapting to fluctuating light environments.
Protein Expression and Purification-Each expression plasmid was transformed into Escherichia coli strain BL21 (DE3). The transformed cells were cultured at 37°C. Expression was induced by the addition of 0.5 mM isopropyl-1-thio-␤-D-galactopyranoside followed by further culture at 21°C overnight. The cells were disrupted by sonication. After centrifugation (125,000 ϫ g for 40 min), the resulting supernatant was used to purify the protein of interest. Each protein except for Trx-f2 was purified by a combination of anion exchange chromatography, using a DEAE-Toyopearl 650M column (Tosoh) and Q-Toyopearl 600C column (Tosoh), and hydrophobic interaction chromatography, using a butyl-Toyopearl 650M column (Tosoh). Purification was performed in a medium containing 25 mM Tris-HCl (pH 7.5-8.1), 1 mM EDTA, and 0.5 mM DTT, but EDTA and DTT were removed by dialysis after purification. The His-tagged Trx-f2 was purified as described previously (21). All the purification procedures were performed at 4°C. Trx-h1 and Trx-o1 proteins were prepared in our previous studies (22,23). Protein concentration was determined with a BCA protein assay (Pierce).
Insulin Reduction Assay of Trxs-The dithiol-disulfide exchange ability of Trxs was measured as the change in turbidity of an insulin solution due to the precipitation of the free insulin B chain by reduction (24). The assay mixture contained 50 mM Tris-HCl (pH 7.5), 50 mM NaCl, 1 mM EDTA, 230 M bovine insulin, and 0.5 mM DTT. Change in turbidity was monitored at 650 nm at 25°C.
Determination of the Midpoint Redox Potential of Trxs-The midpoint redox potential (E m ) of Trxs was determined as described previously (25) with modifications. Each recombinant Trx (0.6 M) was incubated in 25 mM Tris-HCl (pH 7.5), 50 mM oxidized DTT, and various concentrations of reduced DTT (0.001-50 mM). After incubation for 3 h at 25°C, Trx was precipitated with 10% (v/v) TCA and washed with ice-cold acetone. Precipitants were then suspended by SDS sample buffer (62.5 mM Tris-HCl (pH 6.8), 2% (w/v) SDS, 7.5% (v/v) glycerol, and 0.01% (w/v) bromphenol blue) containing the thiol-modifying reagent 4-acetamido-4Ј-maleimidylstilbene-2,2Ј-disulfonate (AMS). After labeling for 1 h at room temperature, protein samples were subjected to nonreducing SDS-PAGE. Proteins were stained with Coomassie Brilliant Blue R-250 (CBB). The E m value of Trxs was calculated by fitting the titration data of the reduction level of Trxs to the Nernst equation. A value of Ϫ357 mV was used as the E m of DTT at pH 7.5, which was given by the observation that the E m of DTT is Ϫ327 mV at pH 7.0 and shows the linear pH dependence of Ϫ59 mV/pH (26,27).
Determination of the Redox State of Thiol-modulated Enzymes in Vitro-FBPase, SBPase, and NADP-MDH proteins were prepared in almost completely oxidized form, whereas approximately half of PrxQ was present in reduced form. PrxQ was accordingly completely oxidized with 0.1 mM H 2 O 2 before the reduction assay. H 2 O 2 was removed by dialysis after the oxidation treatment.
Peptide Mapping Analysis-After separation on nonreducing SDS-PAGE, CBB-stained protein bands were excised from the gel and fully destained with 50 mM NH 4 HCO 3 and 50% (v/v) acetonitrile. The gel slice was dried completely and then incubated with 50 mM NH 4 HCO 3 containing 20 ng/l trypsin at 37°C overnight. Tryptic peptides were extracted from the gel with 0.1% (v/v) TFA with 50% (v/v) and 75% (v/v) acetonitrile, continuously. Whole extracts were concentrated with a centrifugal concentrator and desalted using ZipTip C18 (Millipore). The peptide sample was spotted onto the matrix (␣-cyano-4-hydroxycinnamic acid) and air-dried on a MALDI plate (MTP 384 target plate ground steel BC, Bruker Daltonics). MALDI mass spectra were obtained using an ultrafleXtreme-TK2 spectrometer (Bruker Daltonics). Results were queried with the Mascot search engine from Matrix Science to identify matched peptides.
Activity Measurement of NADP-MDH-After incubation with Trx as described earlier, NADP-MDH activity was monitored as a decrease in absorbance at 340 nm due to NADPH oxidation. The molar extinction coefficient for NADPH of 6200 M Ϫ1 cm Ϫ1 was used for calculation of the amounts of oxidized NADPH. Assays were performed in a medium containing 50 mM Tris-HCl (pH 7.5), 50 mM NaCl, indicated concentrations of oxaloacetic acid, indicated concentrations of NADPH, and 10 nM NADP-MDH at 25°C. Only Trx-f2 was prepared as the His-tagged protein, which resulted in higher M r (13,817.0) than that of Trx-f1 (13,119.2). B, monitoring of insulin reduction activity of Trxs. Time course of absorbance change at 650 nm (Abs 650) is shown. C and D, determination of the E m of Trxs. C, Trxs were equilibrated with various (reduced DTT (DTT red ))/(oxidized DTT (DTT ox )) ratios of redox buffers. Proteins were TCA-precipitated, labeled with AMS, loaded on nonreducing SDS-PAGE, and stained with CBB. Red, reduced form; Ox, oxidized form. D, the reduction level quantified as the ratio of the reduced form to the total was plotted against the redox potential of DTT buffer. The data were fitted to the Nernst equation, and each E m value was determined. The E m values are the mean of two independent titrations with deviations of Ͻ3 mV.
Light-dependent Redox Behaviors of Trxs and Thiol-modulated Enzymes in Vivo-Plants were placed at the indicated light intensities (0 -660 mol photons m Ϫ2 s Ϫ1 ) for 15 min at 25°C and used for the determination of the in vivo reduction level of Trxs and thiol-modulated enzymes as described previously (21). Antibodies against Trx-f1, Trx-m4, Trx-x, Trx-y1, Trx-z, and SBPase were newly prepared using each recombinant protein as the antigen. For PrxQ immunodetection, a commercially available polyclonal antibody (Agrisera) was used.

Results
Trx Selectivity for Redox Regulation of FBPase, SBPase, and PrxQ in Vitro-In Arabidopsis, Trx-f, Trx-m, Trx-x, Trx-y, and Trx-z are encoded by two, four, one, two, and one nuclear genes, respectively (13). Isoforms of each Trx subtype generally show high homology of amino acid sequence in the mature protein region (e.g. Trx-f1 and Trx-f2, 88%; Trx-m1 and Trx-m2, 81%), suggesting that they have similar properties. Then we transformed some of these genes into E. coli and prepared recombinant proteins for five Trx subtypes as shown in Fig. 1A. Except for Trx-z, they were confirmed by the insulin reduction assay to be capable of dithiol-disulfide exchange reaction (Fig.  1B). Redox titration of Trxs with reduced and oxidized DTT indicated that E m of Trx-z is less negative than that of other Trxs (Fig. 1, C and D), which may relate to the lack of insulin reduction activity in Trx-z. However, Trx-z possessed an ability of the dithiol-disulfide exchange reaction, as a specific protein was reduced depending on this Trx (see below). Recombinant proteins of Arabidopsis chloroplast thiol-modulated enzymes including FBPase, SBPase, NADP-MDH, and PrxQ were also prepared. These recombinant proteins were used for comparing the efficiency of Trxs in reducing each thiol-modulated enzyme in vitro.
To gain clues for performing this experiment under nearphysiological conditions, we roughly estimated the in vivo stoichiometry of Trxs to thiol-modulated enzymes. Using a dilution series of recombinant proteins and Arabidopsis leaf extracts, we successfully determined the in vivo amounts of some proteins (Fig. 2, A and B). Amounts of Trxs were almost within the same order of magnitude as those of thiol-modulated enzymes, although they varied depending on the protein species. Note that the in vivo amounts of some Trx isoforms were possibly overestimated because of cross-reaction with other isoforms (e.g. the Trx-f1 antibody also reacted with Trx-f2 with lower affinity ; Fig. 2C). On the basis of these preliminary tests, we evaluated the in vitro Trx selectivity for redox regulation, under conditions where the molar ratio of Trx to thiol-modulated enzyme was 1:2 (M) unless specified.
The redox state of thiol-modulated enzymes was determined by discriminating thiol status with the use of the thiol-modifying reagent AMS (see "Experimental Procedures"). As shown in Fig. 3A, only Trx-f assisted in the shift of FBPase from the oxidized to reduced forms. In contrast, Trx-f and Trx-m showed the capacity to reduce SBPase, but Trx-f showed higher efficiency (Fig. 3B). PrxQ plays a role in the detoxification of reactive oxygen species and is accordingly known as a component of the antioxidant system in chloroplasts (29). To fill this role, PrxQ must receive reducing equivalents from Trx (or other sources). The ability of each Trx for PrxQ reduction was evaluated with varying concentrations of Trx (0.1 or 1 M) and DTT (0 -500 M) because it was anticipated that PrxQ is reduced by 0.5 mM DTT even in the absence of Trx (2). When each Trx was incubated at 0.1 M with PrxQ, only Trx-y apparently promoted PrxQ reduction (Fig. 4). When incubated at 1 M, all Trxs promoted PrxQ reduction with different efficiencies; Trx-x, Trx-y, and Trx-z converted PrxQ to the reduced form even under lower (5-10 M) concentrations of DTT. This result also confirmed the Trx-z capability for the dithiol-disulfide exchange reaction, although Trx-z did not show insulin reduction activity (Fig. 1B).
Two-step Reductive Activation Mechanism of NADP-MDH-NADP-MDH plays a key role in the export of excess reducing equivalents from chloroplasts (30). As shown in Fig. 5A, NADP-MDH adopted three distinct redox states, depending on the type of Trx co-incubated. This finding may reflect that two redox-active Cys pairs, Cys 77 -Cys 82 and Cys 418 -Cys 430 (Arabidopsis numbering), are conserved in the N-and C-terminal extensions of NADP-MDH, respectively (31). NADP-MDH showed the largest AMS-derived shift of molecular weight (M r ) when incubated with Trx-f. In contrast, Trx-m caused only a partial M r shift of NADP-MDH. These results suggest that Trx-f can reduce both of the disulfide bonds, whereas Trx-m can reduce either the N-terminal or the C-terminal disulfide bond alone. Trx-x, Trx-y, or Trx-z did not cause the M r shift of NADP-MDH, indicating that these Trxs can reduce neither N-terminal nor C-terminal disulfide bonds.
One disadvantageous point for AMS-based discrimination of thiol status is that, because of the small molecular mass of AMS (536.44 Da), it is difficult to determine the precise number of thiol groups involved in the redox regulation. To overcome this problem, we recently developed a new thiol-labeling reagent DNA-Mal (28). DNA-Mal causes a larger change in SDS-PAGE mobility (ϳ9 kDa per incorporated DNA-Mal molecule) of proteins and is, therefore, applicable for precise determination of the number of free thiols in the protein molecule. Using this reagent, we determined the number of free thiols that emerged in NADP-MDH by Trx-f and Trx-m, respectively. The molecular mass shifts caused by Trx-f and Trx-m were estimated to be 36.5 kDa (corresponding to four free thiols) and 18.7 kDa (two free thiols), respectively (Fig. 5B). These results strongly support the above mentioned hypothesis of distinct NADP-MDHreducing abilities of Trx-f and Trx-m.
Our next question addressed the detailed mechanism of NADP-MDH reduction by Trx-m; of the two Cys pairs harbored in N-and C-terminal extensions, which one was reduced by Trx-m? To answer this question, the fully reduced, partially reduced, and oxidized forms of NADP-MDH shown in Fig. 5A were in-gel digested with trypsin, and mass spectra of the resulting peptides were compared (Fig. 6A). The theoretical M r values of the tryptic peptides containing the N-terminal Cys  pair (Glu 76 to Lys 88 ) and C-terminal Cys pair (Cys 418 to Val 443 ) were 1553.8 and 2661.0, respectively (calculated using the ExPASy ProtParam tool). A specific peak of m/z 1551.7 was observed in the peptides prepared from the partially reduced and oxidized NADP-MDH, implying that this peak is associated with the peptide of Glu 76 to Lys 88 in which Cys 77 -Cys 82 forms the disulfide bond. Indeed, when a gel slice excised from the partially reduced form was fully reduced with DTT and then alkylated with iodoacetamide, the peak of m/z 1551.7 disappeared and a new peak of m/z 1667.7 emerged (Fig. 6B). This new peak corresponded to the M r of Glu 76 to Lys 88 (containing carbamidomethyl Cys) in the Mascot search engine. These results clearly indicate that Trx-m efficiently reduces the C-terminal Cys pair (Cys 418 -Cys 430 ) but not the N-terminal one (Cys 77 -Cys 82 ).
We further investigated the Trx-dependent change in NADP-MDH activity (Fig. 6C). NADP-MDH showed the highest activity when incubated with Trx-f, whereas Trx-m could not fully activate NADP-MDH. NADP-MDH showed no activity in the absence of Trx or in the presence of Trx-x, Trx-y, or Trx-z. Taken together, we conclude that (i) NADP-MDH exerts maximal activity upon cleavage of both disulfide bonds at Nand C-terminal extensions by Trx-f, (ii) NADP-MDH is only partially activated upon cleavage of the C-terminal disulfide bond by Trx-m, and (iii) NADP-MDH is completely inactive when both of the Cys pairs form disulfide bonds.

In Vivo Redox Behaviors of Trxs and Thiol-modulated Enzymes in Chloroplasts in Arabidopsis Wild-type and Trx-fdeficient Plants-Our in vitro assays
showed that Trx-f is widely used for the reduction of FBPase, SBPase, NADP-MDH, and PrxQ (Figs. [3][4][5]. We finally asked whether these in vitro findings are functionally significant in vivo. For this purpose, Arabidopsis T-DNA insertion mutants in Trx-f1 (trxf1-1, trxf1-2, and trxf1-3) and Trx-f2 (trxf2) genes were prepared (Fig. 7A). The trxf1 trxf2 double mutants were also generated by crossing each single mutant. RT-PCR analysis confirmed that Trx-f1 and Trx-f2 transcript levels are severely decreased or undetected in each corresponding mutant (Fig. 7B).
Wild-type and trxf mutant plants showed similar growth phenotypes (Fig. 7A). Immunoblotting analysis indicated that neither Trx-f1 nor Trx-f2 proteins were detected in trxf1 trxf2 double mutants, whereas other Trxs were accumulated at the same level as in the wild-type plant (Fig. 7C). Despite the dele-  tion of Trx-f2 transcript (Fig. 7B), Trx-f2 protein appears to be normally accumulated in the trxf2 single mutant. This might be due to the cross-reaction of Trx-f2 antibody with Trx-f1 protein (Fig. 2C). This expectation is supported by the observation that Trx-f1 is the major form of Trx-f in Arabidopsis leaves (7).
Using these plants, we characterized the in vivo redox behaviors of Trxs and thiol-modulated enzymes in chloroplasts and evaluated the effects of Trx-f deficiency. In accordance with our previous data (21), Trx-f itself did not show a drastic redox shift upon illumination (Fig. 7D). In contrast, Trx-m, Trx-x, and Trx-y showed photoreduction responses, which were not uniform across the Trx family (Fig. 7, E and F). Chloroplast thiolmodulated enzymes including ATP synthase CF 1 -␥ subunit, FBPase, SBPase, and NADP-MDH also showed distinct lightresponsive redox behaviors (Fig. 7, E and F). CF 1 -␥ was very efficiently reduced even under low light conditions. FBPase, SBPase, and NADP-MDH were shifted to the reduced form more gradually with increasing light intensity, although it was difficult to quantitatively evaluate the reduction level of NADP-MDH. The reduction level of PrxQ was maintained at a stable level (ϳ70%) irrespective of light conditions. Photoreduction of FBPase was partially impaired in trxf1 and trxf1 trxf2 mutants (Fig. 7, E and F). For example, the FBPase reduction level at 660 mol photons m Ϫ2 s Ϫ1 in trxf1 trxf2 mutants was 25% lower than that in the wild-type plant. It was thus evident that Trx-f is directly involved in the redox regulation of FBPase in vivo. However, it should be noted that more than half of FBPase could be reduced even in trxf1 trxf2 mutants. Furthermore, other thiol-modulated enzymes examined here did not show a clear change in photoreduction patterns in trxf mutants. These unexpected consequences suggest the possibility that, together with the Trx system, other redox pathways cooperatively function in the chloroplast redox regulation in illuminated leaves.

Discussion
It is generally acknowledged that the chloroplast redox regulation system serves to transmit light signal from the photosynthetic electron transport chain to thiol-modulated enzymes and thereby ensures light-responsive control of a variety of chloroplast functions (1, 4 -6). This scenario is certainly operative in plants, as shown by our recent study revealing the lightresponsive redox dynamics of chloroplast thiol-modulated enzymes (21). It remains, however, to be uncovered how the underlying redox network is orchestrated in chloroplasts. To gain insights into this critical question, we addressed Trx selectivity for the redox regulation by directly observing Trx-dependent change in the redox state of target proteins.
It has been demonstrated by site-directed mutagenesis studies (32) and structural analyses (33) that FBPase forms a regulatory disulfide bond between Cys 153 and Cys 173 (or Cys 178 upon mutation of Cys 173 ; pea numbering described in Refs. 32 and 33). Our results indicated that Trx-f plays a critical role in the cleavage of this disulfide bond, whereas any other Trxs fail to reduce it (Fig. 3A). These results agree strikingly with previous findings that only Trx-f can substantially enhance the enzymatic activity of FBPase (10,14,15,19). In the case of SBPase, the redox-active Cys residues were identified (34), but the Trx selectivity responsible for the redox regulation has been little characterized. We have reported here that SBPase is reduced primarily by Trx-f and less efficiently by Trx-m (Fig. 3B). Similar trends of Trx efficiency were exemplified in the activation of glyceraldehyde-3-phosphate dehydrogenase and phosphoribulokinase (35). These results indicate that (i) Trx-f plays the predominant role in the reductive activation of Calvin cycle enzymes, (ii) Trx-m is also involved to a lesser extent, and (iii) other Trxs are inefficient.
All types of Trx were capable of reducing PrxQ, but their efficiencies were not equivalent (Fig. 4). Trx-y was the most efficient reducer of PrxQ, followed by Trx-x and Trx-z. This result correlates with the study showing the distinct effects of Trxs on peroxidase activity of PrxQ (11). It has also been shown that Trx-x possesses high ability to reduce 2-Cys Prx (10,11) and that Trx-z can act as an electron donor for several antioxidant enzymes (16). Notably, a recent study using several Trx mutants in Arabidopsis showed that the activity of methionine sulfoxide reductase was lowered in Trx-y-deficient plants, suggesting crosstalk between Trx-y and methionine sulfoxide reductase in vivo (12). These findings, together with our present The transcript levels of Trx-f genes were examined. As a control, 18s rRNA was also examined. C, immunoblotting analysis of chloroplast Trxs. The same amount of leaf total protein was loaded into each lane. As a loading control, Rubisco large subunit (RbcL) was stained with CBB. D and E, photoreduction of Trxs and thiol-modulated enzymes in chloroplasts. Experiments were performed under indicated light conditions (0 -660 mol photons m Ϫ2 s Ϫ1 ). The same amount of leaf total protein was loaded into each lane. As a loading control, Rubisco large subunit (RbcL) was stained with Ponceau S. Light int., light intensity. Red, reduced form; Ox, oxidized form. F, the reduction level of Trxs and thiol-modulated enzymes in chloroplasts. The reduction level was quantified as the ratio of the reduced form to the total. Each value represents the mean Ϯ S.D. (three biological replicates).
results, suggest that Trx-x, Trx-y, and Trx-z can be regarded as compatible partners in various antioxidant systems for supplying reducing equivalents.
The complex mechanism of NADP-MDH redox regulation has been well described based on accumulated biochemical and structural studies (36 -43). Using the discrimination of thiol status combined with mass spectrometry and activity measurement, we further found that N-and C-terminal disulfide bonds harbored in NADP-MDH were reduced via distinct Trx selectivity even within a single polypeptide. Although Trx-f could fully activate NADP-MDH by cleaving both of two disulfide bonds, Trx-m allowed only partial activation due to the inability to cleave the N-terminal one (Figs. 5 and 6). Other Trxs could reduce neither N-terminal nor C-terminal disulfide bonds of NADP-MDH, keeping this enzyme completely inactive (Figs. 5 and 6). These findings coincide with the currently accepted model of NADP-MDH redox regulation: N-terminal disulfide bond formation is involved in a change in the whole NADP-MDH structure, whereas C-terminal disulfide bond formation results directly in blocking of the pathway for the substrate (31).
These results allow us to draw an outline of a redox network composed of five Trxs and four Trx target proteins in chloroplasts (Fig. 8). An important question is how the distinct efficiencies of Trxs in reducing each target protein are defined. The difference in E m among Trxs does not seem to be the major determinant because there was little correspondence between this factor (Fig. 1, C and D) and Trx selectivity for redox regulation (Figs. 3-5). An alternative possible factor is the surface electrostatic potential on Trxs, Trx target proteins, and their interrelationships. Three-dimensional modeling of the Trx structure indicated that the proximal region to the active site of Trx-f is more positively charged than that of Trx-m (10, 44). Geck et al. (19) reported that the replacement of Lys (positively charged) located near the active site of Trx-f with Glu (negatively charged) drastically lowers the efficiency of FBPase activation. Involvement of surface electrostatic potential has also been implicated by studies using ATP synthase CF 1 -␥ subunit mutants; deletion or replacement of negatively charged residues close to the redox-active Cys pair in CF 1 -␥ strongly affects its property of Trx-dependent redox regulation (45)(46)(47). However, further studies are needed to gain deeper insights into the involvement of surface electrostatic potential in redox regulation. In this respect, structure determination of Trx and its target protein co-crystals will directly reveal the critical residues required for charge-charge interaction. According to the model of redox pathways proposed on the basis of the in vitro data (Fig. 8), Trx-f is likely to play the predominant role in the redox regulation of FBPase, SBPase, and NADP-MDH. CF 1 -␥ was also reported to be favorably reduced by Trx-f (48). Are these scenarios physiologically relevant in plants? Characterization of the in vivo redox behaviors of chloroplast thiol-modulated enzymes indicated that FBPase photoreduction was partially impaired in trxf1 (lacking the major isoform of Trx-f) and trxf1 trxf2 mutants (Fig. 7, E and F). This result provides direct evidence that Trx-f mediates the transfer of reducing equivalents to FBPase in vivo. An unexpected but intriguing finding is that FBPase could still be reduced under conditions even where Trx-f was completely lacking. Furthermore, redox behaviors of other proteins that we investigated showed no clear changes in any trxf mutants. Thus, other pathways, not addressed in this study, for transferring reducing equivalents must participate together in the redox regulation in chloroplasts. In other words, a chloroplast redox network may extend beyond the ferredoxin-Trx reductase/Trx system described in current textbooks. The NADPH/NADPH-Trx reductase C system and/or glutathione/glutaredoxin system seem to be strong candidates. Some NADPH/NADPH-Trx reductase C target proteins in chloroplasts have been reported to date (49 -51), but information about them is still limited. Glutaredoxin target proteins were also screened by a proteomics-based procedure (52), but most of them are still putative. More importantly, the in vivo working dynamics and biological significance of these systems remain to be elucidated. These potential systems may confer more flexible and sophisticated regulatory ways on chloroplasts. Our results are valuable in highlighting the complexity and plasticity of the chloroplast redox network and providing an important step toward its clarification.