Chloroplast cyclophilin is a target protein of thioredoxin. Thiol modulation of the peptidyl-prolyl cis-trans isomerase activity.

Chloroplast cyclophilin has been identified as a potential candidate of enzymes in chloroplasts that are regulated by thioredoxin (Motohashi, K., Kondoh, A., Stumpp, M. T., and Hisabori, T. (2001) Proc. Natl. Acad. Sci. U. S. A. 98, 11224-11229). In the present study we found that the peptidyl-prolyl cis-trans isomerase activity of cyclophilin is fully inactivated in the oxidized form. Reduction of cyclophilin by thioredoxin-m recovered the isomerase activity. Two crucial disulfide bonds were determined by disulfide-linked peptide mapping. The relevance of these cysteines for isomerase activity was confirmed by the mutagenesis studies. Because four cysteine residues in Arabidopsis thaliana cyclophilin were conserved in the isoforms from several organisms, it appears that this redox regulation must be one of the common regulation systems of cyclophilin.

Cyclophilin (CyP) 1 is a member of immunophilin superfamily and a target of the immunosuppressive drug cyclosporin A (1). CyP shows peptidyl-prolyl cis-trans isomerase (PPIase) activity and functions as a catalyst in protein folding, facilitating the slow isomerization around Xaa-Pro peptide bonds, which is a general rate-limiting step of protein folding (2,3). The second major role of CyP is its function as a component in cellular signaling. CyP is essential for the regulation of immunosuppression in mammalian T-cells (4,5). In plants CyP was first reported by Gasser et al. (6). Breiman et al. (7) found cyclosporin A-sensitive PPIase in chloroplasts. The corresponding gene for this enzyme was cloned in the nuclear genome of Arabidopsis thaliana as a single gene targeted to chloroplast (8). Members of the CyP family were identified in various organelles of plants (8 -12), and the A. thaliana genome project identified at least 10 isoforms of typical cyclophilins (13). However, the physiological significance of CyP in plant cytoplasm and in chloroplasts has been obscure thus far (9,14,15). The important function of CyP in the process of T-cell activation for the regulation of immunosuppression in mammalian cells suggests that CyP also has an important physiological role in the plant cell, e.g. in signal transduction. Although CyPs of some eukaryotic organisms have several conserved cysteines, there is still no information available on the role of these cysteines. However, we found that CyP is the potential target protein of chloroplast thioredoxin (Trx) (16).
Trx is a small, ubiquitous, disulfide oxidoreductase with two redox active cysteines at the active center (17)(18)(19). The active site sequence (-Trp-Cys-Gly-Pro-Cys-) of Trx is well conserved regardless of the lower overall sequence homology. Trx induces a conformational change in the target protein via exchange of the disulfide bond and thereby modulates the activity of these enzymes. The mechanism for reduction of target proteins by Trx has been studied in vitro (20,21). The first Cys in the active site sequence of Trx probably forms a mixed disulfide intermediate with the target protein. Then the established disulfide bond between the two redox partners is attacked by another cysteine of Trx. Consequently the reduced form of the target protein is released from Trx, and Trx itself is oxidized.
In the chloroplasts of higher plants two Trx isoforms, designated f-type (Trx-f) and m-type (Trx-m) based on their first identified target proteins, are well known (18,22,23). The various chloroplast enzymes are regulated by reduction of their internal disulfide or reoxidation of thiols. Calvin cycle enzymes, glyceraldehyde-3-phosphate dehydrogenase, fructose-1,6-bisphosphatase, sedoheptulose-1,7-bisphosphatase, and phosphoribulokinase, are regulated by their redox states, and their activities are enhanced in the reduced enzymes (24). Recently several approaches to identify the target proteins of Trx have been challenged (16,25,26). Within the captured candidate proteins, we identified CyP as an unreported Trx target in chloroplasts (16).
In the present study, we revealed that CyP in chloroplasts is an actual target protein of Trx and that the PPIase activity of CyP is regulated by the reduction of the internal disulfide bond by Trx-m. In addition, cysteine residues involved in this redox regulation were identified.

EXPERIMENTAL PROCEDURES
Preparation of Trx-m-The recombinant Trx-m was expressed in Escherichia coli (27) and was purified as follows. The E. coli cells were suspended in 25 mM Tris-HCl (pH 7.5) containing 0.5 mM dithiothreitol (DTT), disrupted by a French pressure cell (5501-M, Ohtake Works, Tokyo), and centrifuged at 100,000 ϫ g for 40 min at 4°C. The supernatant was applied to a DEAE-Toyopearl 650 M column (Tosoh, Tokyo) and then eluted with a 0 -150 mM linear gradient of NaCl in 25 mM Tris-HCl (pH 7.5) and 0.5 mM DTT. The peak fraction containing Trx-m was collected, and solid ammonium sulfate was added to be the final concentration of 1.6 M. The solution was then applied to a butyl-Toyopearl 650 M column, and eluted with a 1.6 to 0 M inverse gradient of ammonium sulfate. The protein concentration was calculated from the A 278 using the published molar absorption coefficient value for Trx-m, 20,500 M Ϫ1 ⅐cm Ϫ1 (28).
Preparation of CyP-E. coli BL21(DE3) cell carrying CyP-pET23c, the plasmid for the wild-type CyP (CyP WT ), was cultured at 37°C, and the desired protein expression was induced by the addition of isopropyl-␤-D-thiogalactopyranoside (0.5 mM) at 25°C. CyP WT was obtained as soluble protein and purified as described previously (16).
Preparation of the Mutant CyP-To prepare the mutant CyPs, CyP C53S/C170S , CyP C128S/C175S , and CyP C53S/C128S/C170S/C175S (CyP NoCys ), site-directed mutagenesis was performed with mega-primer PCR method (29) using KOD DNA polymerase (Toyobo, Osaka). The sequences of the resultant plasmids were confirmed by DNA sequencing (Prism 310, Applied Biosystems). When the desired proteins were expressed in E. coli BL21(DE3), CyP C128S/C175S was obtained as soluble protein and could be purified by the method for CyP WT . However, CyP C53S/C170S and CyP NoCys were insoluble proteins and were purified as follows. After induction, the cells were suspended in 25 mM Tris-HCl (pH 7.5) and 5 mM EDTA, disrupted by sonication, and centrifuged at 11,000 ϫ g for 10 min. The inclusion bodies obtained were suspended in 25 mM Tris-HCl (pH 7.5), 5 mM EDTA, and 2% (v/v) Triton X-100, incubated for 30 min at room temperature, and centrifuged at 11,000 ϫ g for 10 min. This washing step was repeated twice. Then the precipitate was washed twice with 25 mM Tris-HCl (pH 7.5) and 1 mM EDTA and dissolved finally in 25 mM Tris-HCl (pH 7.5), 1 mM EDTA, 8 M urea, and 0.5 mM DTT. The insoluble fraction was removed by centrifugation, and the supernatant was diluted 50 -100-fold with 25 mM Tris-HCl (pH 7.5), 1 mM EDTA, and 0.5 mM DTT. Solid ammonium sulfate was added to the diluted solution up to 1.3 M and the solution was applied to a butyl-Toyopearl 650 M column, which was equilibrated with 25 mM Tris-HCl (pH 7.5) containing 1.3 M ammonium sulfate. The protein was eluted from the column with a 1.3-0 M inverse gradient of ammonium sulfate. The peak fraction containing CyP was collected and stored at Ϫ80°C. The protein concentration of CyP was determined by the method described previously (30).
In Vitro Reduction of CyP Assisted by Trx-The recombinant CyPs (both wild-type and mutants) were obtained in reduced form, and the oxidized form of CyP (CyP ox ) was prepared as described (16). Oxidized CyP (3.0 M) was incubated with DTT (50 M) in the presence or absence of Trx-m (3 M) for 1 h at 25°C. To assess the redox state of CyP, 4-acetamido-4Ј-maleimidyl-stilbene-2,2Ј-disulfonate (AMS) (Molecular Probes), a maleimidyl reagent that specifically modifies cysteine residues, was used. AMS labeling was carried out as described previously (16), and CyP ox and reduced CyP (CyP red ) were separated by 15% (w/v) SDS-PAGE.
PPIase Activity of the Recombinant CyP-The PPIase activity of CyP was measured as described (31)(32)(33) with minor modifications. An assay mixture containing 35 mM HEPES-NaOH (pH 8.0), 50 M N-succinyl-Ala-Ala-Pro-Phe-4-methylcoumaryl-7-amide (Peptide Institute) was incubated at 10°C. ␣-Chymotrypsin (final concentration 20 M) was added to initiate the reaction. CyP was added after 15 s. Absorbance at 360 nm was monitored by UV spectrophotometry. First-order rate constants (k obs ) were derived by a curve fit to a first-order rate equation (A 360 ϭ A 1 ϩ A 0 e Ϫkt , where k is the rate constant). The k cat /K m values were calculated according to the equation where k 0 is the first-order rate constant for spontaneous cis-trans isomerization (31).

RESULTS
The Oxidized CyP Is Inactive as PPIase-In the previous study, we reported that chloroplast CyP is a possible target protein for Trx-m and that CyP ox could be reduced by Trx-m in vitro (16). CyP, known as a member of the immunophilin superfamily, promotes the isomerization from cis-form to transform in peptidylproline. Therefore we examined whether the PPIase activity of CyP is affected by reduction or oxidation. We measured the PPIase activity using the model substrate N-succinyl-Ala-Ala-Pro-Phe-4-methylcoumaryl-7-amide and monitored the change of the absorbance at 360 nm derived by the release of the methylcoumaryl moiety from the artificial model peptide when the trans-form peptide was digested by ␣-chymotrypsin. A slow increase in absorbance was observed even in the absence of the catalyzing enzyme, indicating that the model substrate was gradually transferred from the cis-form to the trans-form irrespective of PPIase (Fig. 1A, none). The addition of CyP ox did not change the rate of increase in absorbance at 360 nm (Fig. 1A, CyP ox ). Thus, PPIase activity of CyP must be suppressed when CyP is present in the oxidized form. In contrast, the model substrate was rapidly shifted to the trans-form and digested by ␣-chymotrypsin when CyP red was added to the assay mixture (Fig. 1A, CyP red ). The K obs value of PPIase activity was measured in the presence of various concentrations of CyP red or CyP ox , and the k cat /K m values were calculated as described under "Experimental Procedures" (Fig. 1B and Table I). The activity of CyP ox was less than 1% of the CyP red activity, suggesting that CyP is inactive in the oxidized form.
CyP ox Can Be Reduced, and the PPIase Activity Is Reactivated by Trx-m-We examined whether CyP ox could be reduced and the PPIase activity reactivated by Trx-m. When CyP ox was incubated with Trx-m only, PPIase activity was not observed (Fig. 1B). Weak PPIase activity was detected when 50 M DTT was used instead of Trx-m. In contrast, PPIase activity recovered to 55% of the rate observed in the reduced state when Trx-m was added together with 50 M DTT ( Fig. 1B and Table  I). Inactivation and partial recovery of PPIase activity was dependent on the redox conditions following the redox state of CyP, which was visualized by AMS labeling (Fig. 2).
Identification of Cysteine Pairs Involved in Disulfide Bond Formation in CyP ox -CyP from A. thaliana contains four cysteine residues allowing six different combinations of disulfide bonds. To specify the cysteine pairs involved in disulfide formation that are responsible for the regulation of the PPIase activity, CyP ox was digested by proteases, and the resultant peptide fragments were separated by reversed-phase HPLC after incubation under non-reduced or reduced conditions (Fig.  3). We identified a single redox-responsive peptide fragment in trypsin-digested and in chymotrypsin-digested fragments of CyP ox , respectively (Fig. 3A, TO1, and Fig. 3C, CTO1). Under reduced conditions, TO1 emerged with two specific peaks (Fig.  3B, TR1 and TR2). In the case of the reduced fragments after chymotrypsin digestion, CTO1 disappeared and three specific peaks emerged (Fig. 3D, CTR1, CTR2, and CTR3). The elution profiles of the peptide fragments from CyP red were very similar to those of CyP ox proteolytic fragments after reduction (data not shown). These redox-specific peptide fragments were analyzed by N-terminal peptide sequencing and mass spectrometry (Table II). TO1 obtained by trypsin digestion of CyP ox was composed of two peptides, which were recovered as two peaks, TR1 and TR2, under reduced conditions. The TR1 peptide contains Cys 175 , whereas TR2 contains Cys 128 . Therefore these two cysteines should form the disulfide bond. In the case of chymotrypsin cleavage, CTO1 was composed of two peptides containing Cys 53 and Cys 170 , respectively. The peptide containing Cys 53 of CyP was recovered as CTR1 in the reduced frag- ments. The peptide containing Cys 170 was not identified in the fragments under reduced conditions. In contrast, CTR2, which possibly contains Cys 53 , and CTR3, containing Cys 175 , were obtained only from the reduced fragments. The failure to detect additional peaks corresponding to cysteine containing fragments under reduced and oxidized conditions might be attributed to an incomplete digestion by chymotrypsin or the insufficient separation of the peptides by the reversed-phase HPLC. As a consequence, we were able to determine two disulfide bond pairs, Cys 53 -Cys 170 and Cys 128 -Cys 175 , in CyP ox (Fig. 4A). To confirm the formation of these two disulfide bonds in CyP, the number of free sulfhydryl groups in oxidized or reduced CyP was quantified with 5,5Ј-dithio-bis(2-nitrobenzoic acid). According to these measurements CyP ox contained 0.2 mol of thiol/mol of CyP, and CyP red contained 4.0 mol of thiol/mol of CyP.
Both Disulfide Bonds in CyP ox Inactivate the PPIase Activity and Are Reduced by Trx-m-A CyP NoCys mutant, containing no cysteines, and two cysteine mutants, CyP C53S/C170S and CyP C128S/C175S , which cannot form either of the internal disulfide bonds identified above, were constructed. The PPIase activities of these mutants were examined under oxidizing and reducing conditions to determine the redox-responsive disulfide bonds in the molecule. Although the PPIase activities of three mutants were affected by the number of deleted cysteines, redox sensitivity of the PPIase activity was clearly maintained in the case of CyP C53S/C170S and CyP C128S/C175S (Table III). Redox sensitivity of the internal disulfide bonds in these mutants was confirmed by AMS-labeled CyP mobility on SDS-PAGE (Fig. 4B). The PPIase activity of CyP NoCys was remarkably lower than that of other mutants and was not affected by the reduction or oxidation (Table III). This insensitivity was also confirmed by AMS labeling (Fig. 4B). As observed in Fig. 4B, CyP C128S/C175S was greatly reduced by Trx-m and DTT, although activity was not recovered very much by reduction (Table III). This apparent discrepancy might be due to the instability of the mutant protein after the redox treatment. Taken together, both Cys 53 -Cys 170 and Cys 128 -Cys 175 disulfide bonds were redox-sensitive and involved in redox regulation of the PPIase activity of CyP.

DISCUSSION
CyP Is a Trx-Regulated Enzyme-In the present study, we observed that the PPIase activity of CyP is modulated by the redox state of the molecule and revealed that CyP ox is an inactive PPIase, whereas CyP red is an active one (Fig. 1). A disulfide bond in CyP ox was definitely reduced by Trx-m, and the PPIase activity of CyP was recovered by the reduction of disulfide bonds in CyP. We identified two critical disulfide bonds involved in the thiol modulation of PPIase activity, Cys 53 -Cys 170 and Cys 128 -Cys 175 .   b Obs., observed mass (m/z) estimated by matrix-assisted laser desorption ionization/time-of-flight (MALDI-TOF) mass spectrometry. c Calc., calculated monoisotopic mass of a single-charged molecule (͓MϩH͔ ϩ ) based on the assigned peptide sequence. d These fractions were analyzed by MALDI-TOF mass spectrometry after reduction with DTT. e Because of signal suppression by matrix molecules in matrix-assisted laser desorption ionization, the signals corresponding to the peptides were not detected.
f Peptide sequencing of this fraction indicated that this fragment was digested by chymotrypsin just after Tyr 59 .
Cyclophilin 3 from Caenorhabditis elegans has four cysteine residues, which correspond to the residues in A. thaliana. Their amino acid sequence homology is very high; the identity of the amino acids was 65%. The crystal structure of cyclophilin 3 has been reported (36), and the distance between C ␣ -C ␣ in each disulfide bond pair of CyP ox in A. thaliana could be estimated based on the coordinates of the three-dimensional structure of cyclophilin 3. The distances between C ␣ -C ␣ in Cys 53 -Cys 170 and Cys 128 -Cys 175 was estimated as 16.49 and 18.91 Å, respectively. When Dornan et al. (36) reported the structure, they suggested the possible disulfide bond formation in cyclophilin 3 because the distance between the sulfur atoms of Cys 40 and Cys 168 (corresponding to Cys 53 and Cys 175 of CyP from A. thaliana) was close enough (5.38 Å) to form the bond. However, disulfide bond formation between Cys 53 and Cys 175 has not been found in our case (Table II and Fig. 4). As the calculated distances between C ␣ -C ␣ in Cys 53 -Cys 170 and Cys 128 -Cys 175 were too great to form the intramolecular disulfides, a remarkable conformational change may occur when CyP is oxidized, which may be the cause of the loss of PPIase activity.
Physiological Role of Redox Regulation of CyP-In the thylakoid lumen, a 40-kDa cyclophilin-like protein named TLP40 had been reported (37). TLP40 shows PPIase activity and affects the dephosphorylation of several key proteins of photosystem II, probably as a constituent of a transmembrane signal transduction chain (38). Another cyclophilin is CyP40, which has a cyclophilin-like PPIase domain in the N terminus and has been identified as a regulatory factor in the vegetative phase of A. thaliana (39). Thus cyclophilin-related molecules seem to have significant roles in various compartments of the plant cell and in chloroplasts. So far, the role of CyP in chloroplasts is uncertain. Here we found that CyP in chloroplasts is redox-sensitive and its PPIase activity is modulated by the redox states via Trx-m. Light initiates many physiological phenomena in chloroplasts including biogenesis of proteins. Therefore, the accurate activation of the PPIase activity of CyP under reducing conditions must be physiologically meaningful, as CyP is a member of the protein folding catalyst proteins.
Recently, Lee et al. (40) have reported that a mammalian cyclophilin, CyP-A, binds to 1-Cys peroxiredoxin and supports the peroxidase activity as an electron donor. This finding allowed proposing the cascade from Trx to 1-Cys peroxiredoxin via cyclophilin in the cell. However we could not yet identify the counterpart proteins of CyP in chloroplasts under the physiological conditions, and further studies will be required to understand the physiological significance of the redox transmission network involving CyP in chloroplast.
Perspective-It appears that the reduction/oxidation cascade via Trx in chloroplasts must be an important regulation network for the metabolic pathways in chloroplasts. Actually various metabolic enzymes are suggested to be involved in this network (16,26). In the present study, we have clearly shown the redox regulation of the PPIase activity of CyP. Although we do not have information on the physiological contribution of CyP in chloroplasts at the moment, and we do not know why CyP-PPIase activity was affected by their redox states, our finding should be important in revealing the role of CyP in chloroplasts. Further studies in vivo are necessary to clarify the physiological significance of redox regulation of CyP to-gether with structure analysis to understand the redox regulation of PPIase activity.