New thioredoxin targets in the unicellular photosynthetic eukaryote Chlamydomonas reinhardtii

Abstract

Proteomics were used to identify the proteins from the eukaryotic unicellular green alga Chlamydomonas reinhardtii that can be reduced by thioredoxin. These proteins were retained specifically on a thioredoxin affinity column made of a monocysteinic thioredoxin mutant able to form mixed disulfides with its targets. Of a total of 55 identified targets, 29 had been found previously in higher plants or Synechocystis, but 26 were new targets. Biochemical tests were performed on three of them, showing a thioredoxin-dependent activation of isocitrate lyase and isopropylmalate dehydrogenase and a thioredoxin-dependent deactivation of catalase that is redox insensitive in Arabidopsis. In addition, we identified a Ran protein, a previously uncharacterized nuclear target in a photosynthetic organism. The metabolic and evolutionary implications of these findings are discussed.

Our knowledge on redox regulation of various physiological processes through thiol-disulfide interchange with proteins of the thioredoxin (TRX) superfamily is rapidly progressing as genome-wide approaches and functional genomics studies are expanding. In the plant biology domain, thioredoxin-dependent regulation has been first discovered for chloroplastic enzymes involved in carbon assimilation (1). The completion of the sequencing of the Arabidopsis genome revealed that TRXs constituted a multigene family and led to the assumption that numerous TRX targets were still to be discovered (2-4). Thus, proteomic approaches aimed at identifying the largest possible number of TRX targets have been developed. The first attempts took advantage of the reaction mechanism of TRXs with their targets, in which a transient heterodisulfide forms between the reduced TRX and the oxidized protein, followed by the release of the reduced target due to the attack of the mixed disulfide by the second cysteine of TRX. Model studies showed that when this second cysteine was mutated, the heterodisulfide was stabilized (5, 6). This property was applied in vivo by transforming a TRX-null mutant yeast with a monocysteinic TRX and allowed isolating a peroxiredoxin (PRX) (7). Subsequently, monocysteinic TRX mutants were used in affinity chromatography to trap interacting proteins (8-11). Native TRX affinity columns also allowed retaining some targets by electrostatic interaction (12). Another technique was also developed that consists of a reduction of a crude soluble protein extract with reduced TRX followed by a separation of the proteins on a 2D gel after derivatization of the newly appeared thiols either with a fluorescent (13) or with a radioactive (14) reagent. In one case, both techniques have been applied in parallel (15), yielding similar results.

Most of the disulfide proteomics have been led with higher plants: spinach (8, 10, 12), Arabidopsis (11, 14), or cereal grains (12). A single study was led with cyanobacteria and revealed targets mostly different from those of higher plants (16). The unicellular eukaryotic green alga Chlamydomonas reinhardtii possesses a TRX system in chloroplasts and cytosol (17) and very likely in mitochondria (3). The cytosolic TRX h1 of C. reinhardtii (CrTRXh1) is able to reduce chloroplastic targets, although with a lower efficiency (17). Having set up an affinity column with the monocysteinic mutant of this TRX (9), we identified the retained proteins by mass spectrometry and investigated the effect of TRX on the activity of three of them. Similar to the situation in cyanobacteria, many TRX targets could be identified in Chlamydomonas, and many of them appear specific for the green alga and different from higher plants, drawing an interesting evolutionary picture. In addition, the first nuclear target ever found in a photosynthetic organism could be identified.

Materials and Methods

Chlamydomonas and Arabidopsis Cultures and Extracts. Chlamydomonas cw15 cell wall-less strain was grown in Tris-acetate phosphate medium prepared as described (18), with up to 5 × 10⁶ cells per ml at 150 μE·m^-2·sec^-1 [1 E (einstein) = 1 mol of photons] and at 25°C. Arabidopsis thaliana ecotype Wassilewskija were cultured on soil in a greenhouse. Total proteins were extracted in 30 mM Tris·HCl, pH 7.9/1 mM EDTA as described for Chlamydomonas (9) or Arabidopsis leaves (14). Protein concentration was determined by using the bicinchoninic acid protein assay.

Affinity Chromatography. Site-directed mutagenesis of Chlamydomonas TRX h1 C39S and recombinant protein purification were carried out as described (6). Affinity chromatography and sample preparation were performed as described (9), except that protein extracts were prepared from 3 liters of culture in Tris-acetate phosphate medium and that 150 mM NaCl was added during chromatography.

Tricarboxylic Acid (TCA) Precipitation and 2D Gel Electrophoresis. Proteins were precipitated by 10% TCA (final concentration), centrifuged, and successively washed with 10% TCA (three times), 5 mM NH₄HCO₃, and distilled water. The pellets were vacuumdried and resolubilized in 600 μl of urea buffer [8 M urea/4% (wt/vol) 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate/20 mM DTT/0.2% ampholytes/0.001% (wt/vol) bromophenol blue]. Isoelectric focusing was performed on dry immobilized pH gradient strips (pH range, nonlinear 3-10; 17 cm long) as recommended by the manufacturer (Bio-Rad). In the second dimension, proteins were separated in 12% polyacrylamide gels with a Protean II XL system (Bio-Rad) and detected by staining with Coomassie brilliant blue R-250.

Protein Identification. Spots were excised and destained, and the proteins were digested with trypsin. Peptides were extracted as described (14) and submitted to matrix-assisted laser desorption ionization time-of-flight analysis on a Voyager DE-STR mass spectrometer (PerSeptive Biosystems, Framingham, MA). Proteins were identified by using search programs peptident and ms-fit (http://us.expasy.org/tools/peptident.html and http://prospector.ucsf.edu/ucsfhtml4.0/msfit.htm, respectively). SwissProt/TrEMBL (Translated European Molecular Biology Laboratory) and NCBI (National Center for Biotechnology Information) were referred as protein databases, respectively. When identification could not be achieved with protein databases, translated EST databases were queried. The positive EST sequences were used to identify the corresponding full coding sequence in the Chlamydomonas EST contig database (http://www.biology.duke.edu/chlamy_genome) and the Chlamydomonas genome (http://genome.jgi-psf.org/chlre1). Comparison between the theoretical mass spectra of the deduced proteins and the experimental spectra then allowed confirmation of the identifications. The putative functions of the proteins were assigned based on sequence identities with known proteins by using BLAST alignments.

Enzyme Assays. Isocitrate lyase (ICL) was assayed according to Rua et al. (19). The reaction mixture contained 50 mM Hepes-Na, pH 7.2, 10 mM MgCl2, 4 mM threo DL-isocitrate, 4 mM phenylhydrazine, and enzyme in a final volume of 1 ml. Isocitrate cleavage was measured at 30°C by the change in absorbance at 324 nm associated with the formation of glyoxylate phenylhydrazone. Isopropylmalate dehydrogenase activity was determined at 30°C in a 750-μl standard assay mixture containing 50 mM Hepes-Na, pH 8.0, 200 mM KCl, 5 mM MgCl₂, 2 mM NAD, and 1 mM isopropylmalate (Wako Pure Chemicals, Osaka) by following the absorbance increase at 340 nm. Catalase activity was measured by following the absorbance decrease at 240 nm of an 88 mM solution of hydrogen peroxide. Crude extracts were previously desalted on a PD-10 column (Amersham Pharmacia).

Results and Discussion

CrTRXh1 mutated on its buried active-site cysteine (C39S mutant) was bound to activated Sepharose and used to trap potential algal targets. CrTRXh1 contains only the two cysteines of the active site, which excludes artifactual disulfides with another cysteine(s). Eluted targets were separated by large-pH-range bidimensional electrophoresis and identified by peptide mass fingerprint. About 70 distinct spots were observed (Fig. 4, which is published as supporting information on the PNAS web site). The experiment was done in triplicate by using two separate columns and protein batches, yielding identical results. Because Chlamydomonas proteins deduced from genome annotation are not available in protein databases, identification was performed in different ways: 23 proteins (43%) were identified directly by peptide mass fingerprint by using protein databases. For the others, EST databases and Chlamydomonas genome sequences were used. In two cases [ferredoxin (Fd) and cytosolic 2-cys PRX] sequencing by Edman degradation was also performed. Only six proteins could not be identified. Finally, 55 different proteins involved in various processes could be proposed as TRX targets (Table 1).

Table 1. TRX targets identified in C. reinhardtii.

Proteins	Cys	Loc	Acc	Proteins	Cys	Loc	Acc
Calvin cycle				LytB	5/4/3	P	GG 1092.1
Rubisco large subunit*†	12/7	P	P00877	Oxidative stress response
Rubisco small subunit*	4/2	P	P00873	2-cys PRX (chloroplastic)*	2/2/2	P	Q9FE86
SBPase*	10/4	P	P46284	2-cys PRX (cytosolic)	3/2/2	C	1879
Phosphoribulokinase*	5/4	P	P19824	PRX type II-E*	3/2/2	P	3393
Ribose 5P isomerase*	4/2/0	P	7591	Peptide methionine sulfoxide reductase*	4/4/4	C	6191
Fructose bisphosphate aldolase*	6/2/1	P	Q42690
Photosynthesis				Glutathione peroxidase	3/3/3	C	O22448
OEE1*	3/2	P	P12853	Catalase*	5/2/0	M	022472
Fd*	7/5	P	P07839	Energy, ATP metabolism Inorganic pyrophosphatase
Glycolysis					4/0/0	P	Q93Y52
Enolase*	9/2/2	C	P31683	Soluble inorganic pyrophosphatase	4/2/2	C	Q949J1
TCA cycle/carbon metabolism
Aconitase*	9/3/7	C	6157	Vacuolar ATP synthase alpha	10/5/4	C	592
Dihydrolipoamide succinyltransferase (oxoglutarate dehydrogenase complex su E2)	2/0/0	M	GW730.2.1	Vacuolar ATP synthase E	3/1/0	C	4771
				Chloroplastic ATP synthase alpha	2/1	P	P26526
				Adenylate kinase*	3/1/0	C?	8234
Acetate assimilation/glyoxylate cycle				Folding
ICL	4/1/2	?	Q39577	Hsp70A	6/5/3	C	P25840
Acetyl-CoA synthetase	16/8/4	?	5544	Hsp70B*	3/2/0	P	Q39603
Nitrogen metabolism				Cpn60 (alpha)*†	3/0/0	P	Q42694
Chloroplastic glutamine synthetase*	7/3/3	P	Q42689	Cpn20	2/0/0	P	GG 173.13
Cytosolic glutamine synthetase*	10/3/5	C	Q42688	Translation
Sulfur metabolism				RB60 PDI	4/4/4	P	O48949
S-Adenosyl methionine synthetase	9/7/3	C	O22350	Elongation factor 1	7/1/2	C	6042
5′ adenylylsulfate reductase	9/7	P?	Q8S2V8	Elongation factor 2*	15/9/6	C	97
Aminoacid biosynthesis				Elongation factor Tu*†	1/1/1	P	P17746
Argininosuccinate synthase*†	4/1/0	C	4999	Nuclear transport
Dihydroxyacid dehydratase (Val, lle)*	10/8/4	C	468 + 1541	Ran	5/5/3	N/C	8597
Diaminopimelate epimerase (Lys)	8/7/5	C	4040	Degradation
3-isopropylmalate dehydrogenase (Leu)	6/2/0		520	Serine protease-like	6/2/1	?	2790
Ketol acid reductoisomerase (Val, lle)	7/2/0	P	5044	26S proteasome particle 12*	2/1/0	C	5688
Threonine synthase (Thr)	8/8/1	P	7464	Cytoskeleton
Fatty acid biosynthesis				Actin	5/3/3	C	P53496
Malonyl-CoA ACP transacylase	8/3/1	P?	4681	Unknown functions
Purine biosynthesis				Protein containing rhodanese domain	4/1/1	?	5391
AIR synthase	1/0/0	P/M	2236
Thiamine biosynthesis				Aldose-1-epimerase-like	8/2/1	?	Q9M6D7
Thiazole biosynthetic enzyme*	3/3/1	P	2889	Reversibly glycosylated peptide*	9/8	C?	6839
Isoprenoid biosynthesis				Hypothetical protein	3	?	9270

Cys, total number of cysteines in the protein/conserved cysteines in plants/conserved cysteines in animals and/or bacteria. Loc, putative subcellular localization: P, plastidial; M, mitochondrial; C, cytosolic; N, nuclear, ?, uncertain. Acc, accession numbers. SwissProt accessions are preceded by a letter, other numbers correspond to contigs from the 20021010 EST assembly or to gene predictions in C. reinhardtii genome (URL in Material and Methods) when the number is preceded by GW (genewise) or GG (greengenie). Targets already identified are italicized. Identified targets in higher plants (*) and in Synechocystis (†) are shown.

Calvin Cycle and Photosynthesis. Several key enzymes of the Calvin cycle were the first TRX targets identified in photosynthetic eukaryotes (1). We have identified six Calvin-cycle enzymes as possible TRX targets. Phosphoribulokinase and sedoheptulose-1,7-bisphosphatase are well known TRX-regulated enzymes, and Rubisco small and large subunits and fructose bisphosphate aldolase were found to bind to TRX affinity columns in higher plant extracts (8, 10). Ribose 5-phosphate isomerase is a new target that we also found in Arabidopsis (14). This target is implicated in the Calvin cycle and in the oxidative pentose phosphate pathway. A redox regulation of this enzyme would control the partitioning of pentose phosphates between these two pathways. Surprisingly, two well known photosynthetic enzymes, oxygen-evolving enhancer 1 (OEE1) and Fd, were also identified in Arabidopsis by using a derivatization method, thereby excluding an artifact of the affinity column (14). OEE1 is a subunit of the oxygen evolving enhancer located on the luminal side of photosystem II. The recently reported TRX activity of this subunit in green algae (20) might explain its binding to the column, although the functional role of a redox regulation of OEE1 remains to be determined. Fd is reduced by the photosynthetic electron transfer and donates electrons to many acceptors. TRX reduction being mediated by Fd thioredoxin reductase, a direct regulation of Fd by TRX is most improbable. SoxR, a sensor/transducer of oxidative stress and nitric oxide in E. coli, contains a [2Fe-2S] cluster. It was shown that TRX promotes the assembly of this cluster into apo-SoxR (21). A similar role of chloroplastic TRXs in the assembly of the [2Fe-2S] cluster into apo-Fd might explain the binding of Fd to the TRX column.

Glycolysis. Enolase is a glycolytic enzyme that had already been shown to bind to TRX columns in higher plants (10). It has been previously suggested that some higher plant enolases might be redox-regulated (22).

TCA Cycle and Carbon Metabolism. Dihydrolipoamide succinyltransferase is the E2 subunit of the oxoglutarate dehydrogenase complex. The regulation of this mitochondrial multienzymatic complex by TRX has been reported in mammals (23). Aconitase, a [4Fe-4S] enzyme, which reversibly isomerizes citrate to isocitrate, was recently proposed as a possible target in higher plants (15). In eukaryotes two isoforms are found, one in mitochondria and the other in cytoplasm. The isoform we identified is probably cytosolic. In mammals, aconitase can be converted into an iron-regulatory protein able to bind RNA after exposure to nitric oxide (NO). It has been reported that this binding is TRX-activated (24). As plant aconitase activity is also modulated by NO (25), a role of TRX in the NO signaling pathway through iron-regulatory protein/aconitase would be worth investigating.

Acetate Assimilation. Acetyl-CoA synthetase (ACS) converts acetate into acetyl-CoA by using ATP and CoA. Acetyl-CoA is involved in a variety of physiological processes, including the TCA cycle and fatty acid, amino acid, and isoprenoid metabolisms. It also provides carbon for the glyoxylate pathway. ACS is thus a key enzyme for growth on acetate. Its presence in the TRX column eluate, although showing its TRX dependence, is consistent with the fact that the extracts were made from cells grown on acetate. A TRX regulation of this enzyme has not been reported, but, interestingly, ACS from Penicillium chrysogenum is activated by DTT (26). ICL catalyzes the reversible cleavage of isocitrate to glyoxylate and succinate. It is part of the glyoxylate pathway, which enables microorganisms to grow on two carbon compounds. In higher plants, ICL is localized in glyoxysomes and converts lipids to carbohydrates during the germination of oil-rich seeds. Chlamydomonas is thought to be devoid of glyoxysomes; thus, the subcellular localization of ICL remains unknown. The redox regulation of ICL activity was tested on cell-free extracts (Fig. 1A). ICL was fully activated within 20 min, with a 6- to 10-fold increase in activity, with either 5 mM DTT or 56 μM TRX reduced with NADPH-thioredoxin reductase (NTR) and NADPH. Activation was slower with 30 μM TRX. ICL eluted from the affinity column was also 3- to 4-fold more active in the presence of DTT or enzymatically reduced TRX (Fig. 1B). All these data confirm that ICL from C. reinhardtii is a target of TRX. ICL from E. coli was not activated by DTT (data not shown), although the sequence of C. reinhardtii ICL shares more identity with its bacterial counterparts (61% with E. coli) than with the protein from A. thaliana (30% identity). ICLs from higher plants are 18 kDa larger than the proteins from Chlamydomonas or bacteria because of a 100-aa insertion (Fig. 5, which is published as supporting information on the PNAS web site). Sequence alignments indicate that only the cysteine belonging to the active site (Cys 195, E. coli numbering) is conserved among all sequences.

Fig. 1.

Regulation of ICL activity by TRX. (A) Activation kinetics in crude extracts. Protein extracts (80 μg) were incubated at room temperature with no addition (•) or in the presence of 5 mM DTT alone (▾) or NTR-reduced A. thaliana TRX h3 [30 μM (○) or 56 μM (▿)] with 0.5 μM A. thaliana NTR and 2 mM NADPH. (B) Activities of C. reinhardtii ICL eluted from affinity column. After elution by DTT, the eluate was dialyzed to remove DTT and then concentrated. DTT (5 mM) or TRX h3 (30 μM) reduced with NADPH and NTR as above were used. ICL activity was assayed as described in Material and Methods.

Nitrogen and Sulfur Metabolism. Glutamine synthetase is a key enzyme of nitrogen assimilation, catalyzing the synthesis of glutamine from ammonium and glutamate. The chloroplastic isoform is already known to be regulated by TRX (27) and was bound on TRX affinity columns in higher plants (8, 10). The identification of the cytosolic isoform also found in Arabidopsis (14) suggests that nitrogen assimilation is fully under TRX control. Sulfur assimilation is known to be regulated by TRX at the level of 5′-adenylylsulfate reductase, which catalyzes the reduction of activated sulfate to sulfite for the synthesis of cysteine and glutathione (28). S-adenosyl methionine (SAM) synthetase catalyzes the formation of SAM from ATP and methionine. SAM is involved in a myriad of biochemical pathways, including transmethylation reactions and the biosynthesis of ethylene, biotin, and polyamines (29).

Amino Acid Biosynthesis. A redox regulation of amino acid biosynthesis in chloroplast has been suggested (30), and the involvement of TRX in this process has been proposed recently, when the first enzyme in the biosynthesis of aromatic amino acids was shown to be regulated by TRX f in higher plants (31). In C. reinhardtii we found six potential targets involved in biosynthesis of arginine (argininosuccinate synthase), threonine (threonine synthase), lysine (diaminopimelate epimerase), branched amino acids valine and isoleucine (dihydroxyacid dehydratase and ketolacid reductoisomerase), and leucine (isopropylmalate dehydrogenase). Synechocystis argininosuccinate synthase (16) and Arabidopsis dihydroxyacid dehydratase (14) have already been proposed as TRX targets. The four other proteins are new targets. We focused on isopropylmalate dehydrogenase (IPMDH) catalyzing the third step of the leucine biosynthetic pathway. IPMDH contains two cysteines conserved in higher plants (Fig. 6, which is published as supporting information on the PNAS web site). As shown in Fig. 2, no activity was detected in Chlamydomonas extracts in absence of reductant. Addition of DTT or reduced TRX was required for activity, thus confirming that Chlamydomonas IPMDH is activated by TRX.

Fig. 2.

Effect of reduction on isopropylmalate dehydrogenase activity. Crude extracts (80 μg) were incubated for 15 min at room temperature in the presence of various reductants, and isopropylmalate dehydrogenase activity was assayed as described in Material and Methods. TRX h3 was used at 30 μM.

Fatty Acid Biosynthesis. Acetyl-CoA carboxylase, which catalyzes the formation of malonyl-CoA from acetyl-CoA, was shown to be regulated by TRX (32). We identified malonyl-CoA:acyl carrier protein transacylase, which catalyzes the formation of malonyl-ACP, the key building block for de novo fatty acid biosynthesis.

Purine, Isoprenoid, and Thiamine Biosynthesis. Enzymes involved in additional biosynthetic pathways, not previously known to be redox-regulated, were also identified. Phosphoribosyl-aminoimidazole synthase catalyzes the fifth step of de novo purine biosynthesis. A dual targeting of this enzyme to mitochondria and chloroplasts was recently demonstrated in cowpea (33). LytB catalyses the last step of the non-mevalonate terpenoid biosynthesis pathway. This pathway was discovered very recently and is only present in bacteria, plants, and protozoa (34). Thiazole biosynthetic enzyme is involved in the biosynthesis of the thiamine precursor thiazole and was also identified as a possible TRX target in higher plants (10).

Oxidative Stress. TRXs are known to play a major role in oxidative stress responses, either through redox signaling or more directly as electron donors to the numerous members of the PRX family (35). The trapping of three PRXs (chloroplastic and cytosolic 2-Cys PRX and type II-E PRX) on TRX affinity columns is thus not very surprising. Chloroplastic 2-Cys PRX was one of the first TRX targets identified in Chlamydomonas and higher plant extracts (8, 9, 14). Type II-E PRX was identified more recently (11). In addition, we identified cytosolic 2-Cys-PRX in the present study. The identification of a glutathione peroxidase suggests that TRX could substitute for glutathione as a reductant, as demonstrated for two plant glutathione reductase homologues (36). Peptide methionine sulfoxide reductase is a well characterized TRX target recently confirmed in poplar (37). The presence of catalase is more surprising, considering that this enzyme has been extensively studied in higher plants and that the redox regulation of its activity has never been reported. However catalase has been found among TRX targets in potato mitochondrial extracts (15). We have thus measured the effect of DTT with or without TRX on the catalase activity of Chlamydomonas and Arabidopsis extracts (Fig. 3A). Catalase activity in Chlamydomonas extracts was decreased by 50% after 10- to 15-min treatments in both conditions; it was unaffected in Arabidopsis extracts even after 30 min, or when DTT concentration was increased to 100 mM (data not shown). When TRX was reduced enzymatically with NADPH and NTR, a similar 50% inhibition was observed, whereas Arabidopsis activity remained stable. Only one catalase gene is apparently present in Chlamydomonas genome although earlier biochemical studies had suggested the presence of three distinct enzymes (38). In the latter case, the 50% inhibition observed on total extracts might represent a 50% inhibition of the major catalase or a more important inactivation of an isoform accounting for part of the total activity. However, when redox regulation of catalase activity was tested on TRX column eluates (Fig. 3B), a similar 50% inhibition was found, suggesting that the TRX-dependent catalase is the major or only catalase in Chlamydomonas. Chlamydomonas catalase contains five cysteines, two of which are conserved in higher plants (Fig. 7, which is published as supporting information on the PNAS web site); this suggests that the TRX dependence might reside among the three additional cysteines.

Fig. 3.

Regulation of Chlamydomonas catalase activity by TRX. (A) Kinetics of catalase inactivation by reductants in total extracts from Chlamydomonas (circles) and Arabidopsis leaves (triangles). Open symbols, DTT. Closed symbols, DTT + TRX. (B) Activity of C. reinhardtii catalase eluted from affinity column. Reductants are as in Fig. 1, except that 50 μM CrTRXh1 was used.

The difference in catalase regulation between algae and higher plants might be linked to the subcellular localization of catalase. Indeed, in higher plants, catalases are localized in peroxisomes, where no TRX are thought to be present, whereas the algal enzymes were shown to be mitochondrial (38). The purpose of this regulation is consistent if TRXs are considered to be sensors of the redox state.

Energy and ATP Metabolism. Adenylate kinase, which catalyses the formation of ADP from AMP and ATP, was recently identified as a possible TRX target in mitochondria (15). The regulation of chloroplastic ATP synthase by TRX has been well characterized (39); however, it involves the γ subunit, and we bound the α subunit. We also identified two isoforms of vacuolar ATP synthase. Several subunits of the mitochondrial ATP synthase were also recently identified as putative TRX targets in higher plants (15). Finally, two inorganic pyrophosphatases, putatively localized in the chloroplast and cytosol respectively, appear as new potential TRX targets. Altogether, these results suggest that ATP synthesis and metabolism are tightly controlled by TRX in the cell. Such a regulation might be necessary to coordinate the ATP pool with the reducing power in all subcellular compartments.

Protein Folding. Four molecular chaperones were found in the TRX column eluate. Chloroplastic Hsp70 and Cpn60 (GroEL) had already been linked to TRX in higher plants and Synechocystis (10, 11, 15, 16). This work adds the cytosolic Hsp70 as well as Cpn20, the chloroplast homologue of bacterial GroES, to the list of known putative TRX-dependent chaperones. A recent study has demonstrated the formation of a complex between an Arabidopsis TRX-like protein and yeast Hsp70 that is released under oxidative stress (40). Another study demonstrated the existence of a redox-regulated molecular chaperone network (41). The target of redox regulation is the Hsp33 chaperone, but the network also includes Hsp70 homologues and the GroES and GroEL proteins. These data indicate that all of the molecular chaperones identified as TRX targets on affinity columns are directly or indirectly redox controlled.

Translation. Elongation factor Tu, a chloroplast protein encoded by the nucleus in higher plants and by the chloroplast in algae, had previously been identified among TRX targets in higher plants and cyanobacteria (10, 16). The mitochondrial elongation factor Tu was also identified as a potential TRX target (15). The Rb60 protein is a chloroplastic protein disulfide isomerase involved in the control by light of psbA mRNA translation in Chlamydomonas (42). The identification of Rb60 among TRX targets confirms that the light control of psbA translation involves TRX, as previously suggested (43). The identification of cytosolic elongation factors 1 and 2 suggests the implication of TRXs in the control of translation in both chloroplast and cytosol.

Nuclear Transport. Ran constitutes the first nuclear target of TRX identified in a photosynthetic organism. Ran is a small ras-like protein involved in nuclear transport, mitosis spindle formation, and nuclear envelope assembly (reviewed in 44). It is localized in the nucleus, mainly in a GTP-bound form, and in the cytosol, mainly in a GDP-bound form. The gradient of Ran-GTP across the nuclear envelope, because of asymmetric distribution of hydrolase and nucleotide exchange activities, is the key regulator of the directionality of nuclear transport across nuclear pore complexes. Nuclear transport has been less studied in plants than in mammals and yeast but basic mechanisms, including the importance of Ran-GTP gradient, appear to be conserved (45). Plant Ran proteins contain two additional conserved cysteines absent in nonphotosynthetic eukaryotes (Fig. 8, which is published as supporting information on the PNAS web site), suggesting a specific interaction between Ran and TRX in photosynthetic eukaryotes. Several transcription factors involved in oxidative stress signaling in yeast and mammals have been shown as being TRX-regulated (46). They are imported to and exported from the nucleus through a Ran-dependent process. Such transcription factors have not yet been identified in plants. The lack of such identification might be due to the very low abundance of these proteins or to the existence of TRX independent oxidative stress signaling pathways in plants. However, the identification of Ran as a target of TRX in the present study might indicate that TRX could regulate transcription factors by controlling their nucleocytoplasmic translocation.

Protein Degradation. A serine protease-like and 26S proteasome non-ATPase-dependent subunit 12 were identified. Several subunits of 26S proteasome were previously identified as possible TRX targets in other organisms, along with other types of proteases (10, 11, 14, 47). In yeast, animals, and plants, a close relationship between the 26S proteasome and the nuclear envelope has been found and is thought to play a role in monitoring the protein traffic between the cytosol and the nucleus (reviewed in ref. 48). The regulation of the 26S proteasome might thus be linked to the regulation of Ran.

Actin and Proteins of Unknown Functions. The identification of actin as a TRX target is surprising. However, the polymerization of actin is already known to be redox-regulated through glutathionylation (49). Reversibly, glycosylated polypeptides are involved in cell-wall biosynthesis and/or starch synthesis (50) and were recently identified as possible TRX targets in cereal starchy endosperm (47). The other identified proteins have unknown functions but sometimes harbor known domains, such as rhodanese or aldose-1-epimerase domains.

Cysteine Conservation. Many chloroplastic enzymes regulated by TRX have been shown to contain additional conserved cysteine residues compared with homologous enzymes from nonphotosynthetic organisms (51). All identified putative TRX targets contain at least one cysteine. The presence of a single cysteine (AIR synthase, Elongation factor Tu) suggests a regulation through reduction of an intersubunit disulfide. Sequence alignments with homologous sequences from diverse species allow estimating the conservation of the cysteine residues (Table 1). In many cases, plant sequences contain two additional conserved cysteines compared to bacteria and/or animals. The sequence alignment of Ran proteins provides a nice example of such a situation (Fig. 8). A second situation corresponds to conservation of cysteine residues in most sequences, and such is the case for PRXs that are indeed TRX-dependent in all species examined so far. In some cases, like aconitase, the number of conserved cysteines is higher in the Chlamydomonas sequence and its animal homologues than in plant sequences. In these cases, a higher identity between the algal and animal sequences compared with plant sequences is observed. This observation suggests a mechanism present in algae and animals but not in plants. An example of such a situation is the recently reported presence of selenoproteins in Chlamydomonas and animals but not in plants (52). In contrast, the absence of cysteine conservation might indicate a TRX regulation restricted to Chlamydomonas or algae, as in the case of inorganic pyrophosphatase. However, the conservation of cysteines only provides hints about possible sites of redox regulation. Hence, all of the proteins identified as putative TRX targets will have to be confirmed biochemically.

Conclusion. The number of putative TRX targets is growing steadily with the development of proteomic approaches. While the importance of redox signaling becomes widely recognized, the emerging picture points to the diversity of the regulatory mechanisms within the various phyla. TRXs seem to be involved in numerous metabolic pathways, but the redox-regulated enzymes are not always the same. For example, ICL seems to be redox-regulated only in Chlamydomonas (or possibly other green algae). Thus, disulfide proteomes of other organisms would deserve to be examined, given that an extrapolation based on the available results can hardly be done. In particular, no disulfide proteome is available for mammals or yeast. The culture conditions might also modify the general picture. In the present case, the fact that Chlamydomonas was grown on acetate in mixotrophic conditions certainly helped finding the enzymes of acetate assimilation that are probably expressed at a higher level. The finding of a nuclear protein in Chlamydomonas extracts suggests that working with isolated nuclei would lead to the discovery of new redox signaling elements.

Finally, the finding of a certain number of targets that are common to different species, like PRXs that seem to be TRX-dependent in all living organisms, makes one confident about the validity of the investigation methods. A study of the redox-triggered modifications of the biochemical properties of the identified new targets along with reverse genetics approaches will be necessary to ascertain the physiological significance of their regulation.