Plant Organelles Contain Distinct Peptidylprolyl cis , trans-Isomerases *

Peptidylprolyl cis,trans-isomerase (PPIase) activity was detected in the cytosol, mitochondria, and chloroplast of pea plants. Cyclosporin A inhibited the activity largely localized to the mitochondrial matrix while rapamycin inhibited the PPIase activity associated with the mitochondrial membranes. Differential inhibition by the two immunosuppressive drugs, the specific binding of these drugs to different mitochondrial fractions, and the immunological detection of a putative 25-kDa rapamycin-binding protein (RBP) in mitochondrial extracts attests to the presence in plant mitochondria of both cyclophilin and RBP classes of PPIases. Cyclosporin A-sensitive PPIase detected in the chloroplast was mostly localized to the thylakoids, which is suggestive of its function in the folding of membranal proteins. PPIase associated with the chloroplast stroma and the thylakoids was not inhibited by rapamycin nor was any cross-reactive RBP detected in chloroplast extracts. These results demonstrate the presence of distinct classes of PPIases in the mitochondria and the chloroplasts of plants.

lyzes slow isomerization of peptide bonds in oligopeptides as well as cellular proteins involving the amino acid proline (Gething and Sambrook, 1992;Freedman, 1989). Protein disulfide isomerase activity has generally been found associated specifically with the endoplasmic reticulum of lower eukaryotes and mammals while proteins with PPIase activity have a wider distribution and are shown to be present in bacteria to mammals (Freedman, 1989;Fischer and Schmid, 1990;Liu and Walsh, 1990;Schonbrunner et al., 1991).
In plants, relatively more information is available on the involvement of chaperones in the biosynthesis, oligomerization, and maturation of proteins than on the occurrence and/ or role of the protein-folding enzymes. Protein disulfide isomerase activity has been detected in plant tissues, but very little is known about PPIase in plants (Holmgren, 1985;Gasser et al., 1990). PPIases in other tissues have been characterized and found to be easily distinguishable from one another by selective inhibitory effects of the immunosuppressive agents, cyclosporin A (CsA) and FK506 (Schrieber, 1991). Proteins that bind CsA are called cyclophilins while structurally distinct drugs such as FK506 and rapamycin bind to proteins called FKBP (FK506-binding proteins) or rapamycin-binding proteins (RBP) (Handschumacher et al., 1984;Harding et al., 1989;Siekierka et al., 1989;Schonbrunner et al., 1991). These proteins show a wider distribution, being present in the cytosol (Koletsky et al., 1986), periplasmic space (Hultsch et al., 1991), and the mitochondria (Tropschug et al., 1988) of different organisms. Although their role(s) and physiological substrates remain to be determined, PPIases have been termed "conformases" (Fischer and Schmid, 1990) or "rotamases" (Schreiber, 1991) because they catalyze slow steps in the initial folding/rearrangement of proteins. The mechanism of this catalysis around specific peptide bonds is also unknown.
We have undertaken an investigation toward finding PPIases in pea plant organelles, the mitochondria and the chloroplasts. We show here for the first time that both the chloroplast and the mitochondria contain CsA-sensitive PPIases while a putative rapamycin-binding protein is found exclusively associated with pea mitochondria.

EXPERIMENTAL PROCEDURES
Plant Material-Pea seedlings (cv. Alaska) were grown for 10 days under white light in a growth chamber at 25 "C, and the leaves were harvested for chloroplast preparation. For mitochondrial preparations etiolated pea seedlings grown for 6 days at 25 "C were harvested.
Pea Chloroplast Stroma and Membrane Isolation-Intact pea chloroplasts were isolated by the method of Bartlett et al. (1982), resuspended in a lysis buffer containing 50 mM NaCl, 50 mM Tris-HCI, pH 7.4, 20 mM MgC12, 0.1 mM EDTA, 10% glycerol, 2 mM phenylmethylsulfonyl fluoride, 2 pg/ml leupeptin, and 5 mM P-mercaptoethanol, and then vortexed. The broken chloroplasts were pelleted at 7,000 X g for 10 min and the supernatant used as stroma. The chloroplast membranes were washed two times in the lysis buffer supplemented with 300 mM NaCI. Finally the membranes were washed twice in 10 mM Tris-HCI, pH 8.0.
Localization and Partial Purification of Cyclosporin A Binding Activity from Thylakoids-Thylakoids were isolated according to Marder et al. (1982) and fractionated into grana and stroma lamellae as previously described (Mattoo and Edelman, 1987;Callahan et al., 1989). For studying CsA binding, whole thylakoids (1 mg of chlorophyll/ml) were solubilized in 2% Triton X-100 for 30 min at 4 "C. The insoluble material was removed by centrifugation at 2,000 X g 21293 for 10 min and the supernatant applied to a DEAE-Toyopearl650 S column equilibrated with 50 mM Tris-NaOH, pH 7.2, and 0.2% Triton X-100. The column was washed with 10 ml of the same buffer and the bound material eluted with 20 ml of 0.5 M NaC1. Fractions (0.5 ml) were collected and analyzed for chlorophyll content and CsA binding.
Determination of Chlorophyll and Protein Concentrations-Chlorophyll was determined in leaf extracts prepared in 80% acetone (Arnon, 1949). Protein content was determined by the Bradford method (Bradford, 1976).
Mitochondria, Matrix, and Membrane Isolations-Mitochondria were isolated as previously described (Breiman, 1987) and fractionated into the matrix and membranal fractions (Hack et al., 1991). The mitochondria were resuspended in sucrose phosphate buffer (SP) (0.3 M sucrose, 20 mM sodium phosphate, 0.25 mM EDTA, and 0.25 mM EGTA, pH 7.2) and sonicated 6 times for 5 s at 25-5 intervals with a microprobe of a Virsonic 300 (VirTis) at 60% of full power and the tubes kept in a mixture of methanol/ice. The unbroken mitochondria and aggregated material were pelleted by centrifugation a t 15,600 X g for 10 min. The supernatant was then centrifuged at 230,000 X g for 70 min to obtain the matrix (soluble) fraction. The pellet was resuspended in S P buffer, divided into two portions, and centrifuged at 230,000 X g for 70 min. One portion was washed twice with SP buffer, and the second membrane portion was resuspended in 0.1 M sodium carbonate (pH 11) to remove peripheral proteins (Fujiki et al., 1982). The tubes were incubated for 40 min on ice and then centrifuged at 230,000 X g for 70 min. The membrane pellets were resuspended in 10 mM Tris-HC1, pH 7.2, 150 mM NaC1, 1 mM phenylmethylsulfonyl fluoride, 1.8% N-octyl glucoside and incubated on ice for 60 min. The solubilized mixture was centrifuged at 15,600 X g for 5 min to remove the insoluble material. For CsA binding assay and PPIase activity, the mitochondrial fraction was solubilized in 1.8% N-octyl glucoside for 60 min on ice and centrifuged at 15,600 X g for 10 min to remove the insoluble material.
Peptidylprolyl &,trans-Isomerase Assay-PPIase activity was measured in a coupled assay with chymotrypsin (Boehringer Mannheim) using a Shimadzu UV-160 spectrophotometer essentially by the method of Fischer et al. (1989) with the following exceptions. The test peptide N-succinyl-Ala-Ala-Pro-Phe-p-nitroanilidine (Sigma) at 60 p~ final concentration was added to a solution of the assay buffer (40 mM Hepes pH 8.0, 0.015% Triton X-100) and the plant extract in a final volume of 1.5 ml. The reaction was initiated by the addition of chymotrypsin to a final concentration of 20 p~. Immediately following the addition of chymotrypsin the change in absorbance at 390 nm (A390) was monitored for 100 s. First order rate constants were calculated from semilog plots derived for each curve. The protein concentration of each plant extract tested is indicated in the appropriate figure legend. One unit of PPIase activity is expressed as nmol (substrate hydrolyzed) s" mg protein". The difference between the catalyzed and uncatalyzed first order rate constants, derived from the kinetics of the absorbance change at 390 nm, was multiplied by the amount of substrate in each reaction. Under equilibrium conditions, 88% of the peptide substrate is present in the trans form. The remaining 12% is present in the cis form, which is cleaved upon enzymatic conversion to the trans form by PPIase.
CsA Binding Assay-Binding of [3H]CsA to the plant components was studied using Sephadex LH-20 columns as previously described (Handschumacher et al., 1984). The reactions were carried out in a 100-pl final volume in a solution containing 10 mM Hepes pH 8, 100 mM NaCI, 0.015% Triton x-100, 5 mM 0-mercaptoethanol, 0.5 p M CsA, ['HH]CsA (lo6 cpm), and plant extract. In some reactions, different concentrations of nonradioactive CsA were added to study specific dilutions of bound I3H]CsA. The reactions were incubated for 60 min a t 25 "C and then applied to 1.7-ml LH20 columns. The columns were washed with 400 p1 of 10 mM Hepes, pH 8.0, 100 mM NaCl, 0.015% Triton X-100, 5 mM 0-mercaptoethanol, and the bound material was eluted with an additional 500 p1 of the same buffer. The binding of CsA was calculated from the amount of [3H]CsA bound to the fraction that eluted in the void volume. In order to demonstrate specific binding, binding was also determined with different specific activities of [3H]CsA. The amount of specific binding is given as the difference in the binding at 500 nM and that at 5 nM.
Gel Electrophoresis and Western Blotting-Proteins were fractionated by SDS-polyacrylamide gel electrophoresis on 10-20% gradient acrylamide gels and visualized by staining with Coomassie Brilliant blue R (Laemmli, 1970). For Western blot analysis, proteins from an identical gel were transferred electrophoretically to nitrocellulose paper (Schleicher & Schuell) at 30 V overnight in a liquid transfer apparatus (Callahan et al., 1989). Immunodetection was carried out by the method of Burnette (1981) using the antibody raised against the conserved region of 12-26 amino acids of the yeast rapamycinbinding protein (Koltin et al., 1991). Antigen-antibody complexes were visualized by reaction with horseradish peroxidase conjugated to anti-rabbit antibodies (Bio-Rad) as described (Callahan et al., 1989).

RESULTS AND DISCUSSION
Pea Mitochondrial PPIase-Solubilized pea and mitochondrial extracts were found to contain a bona fide PPIase activity; the enzyme activity exhibited first order kinetics (Fig. lA), was linearly correlated to the amount of the mitochondrial extract (Fig. lB), and was inhibited close to 90% in the presence of CsA (14 pM) or to 16% in the presence of rapamycin (24 p~) (Fig. l . 4 and Table I). The distribution of this activity between the soluble (matrix) and membrane fractions of purified mitochondria was next determined. Results (Table I) showed a 13-fold higher specific activity of PPIase in the matrix as compared with the mitochondrial membranes, suggestive of an enrichment of this activity in the soluble fraction. Of the total enzyme units present in the two mitochondrial fractions, 78% were recovered in the matrix fraction (not shown). PPIase activity was peripherally associated with washed mitochondrial membranes because it could be easily removed by treatment with 0.1 M sodium carbonate (Table I).
Both mitochondrial fractions strongly bound CsA, which resulted in the near total inhibition of their PPIase activity (Table I) (Fig. 2). In washed membranes treated with 0.1 M sodium carbonate, which resulted in the removal of PPIase activity from the membranes (see Table I), this cross-reactive protein band was not detected (Fig. 2, lane 4). These data are consistent with the interpretation that the putative 25-kDa RBP is peripheral to pea mitochondrial membranes and that its removal was possibly the cause for our inability to detect PPIase activity in membrane fractions treated with sodium carbonate ( Table  I)  ' The specific binding of CsA is given as a difference in the binding a t 500 nM and that at 5 nM.

. In contrast, rapamycin inhibited PPIase activity of only the mitochondrial membranes. However, when an antibody against yeast RBP was used to detect cross-reactive protein bands in different mitochondrial fractions, a 25-kDa protein was detected in both matrix and membranes
' This value is from another experiment where PPIase activity in the total extract was 7.02 f 0.97 nmo1.s-l mg", which in the presence of 'This value is from a different experiment where the matrix PPIase activity in the absence of CsA was 4.0 f 0.54 nmol-s" mg".
-24 p~ rapamycin was reduced to 5.85 f 0.28 nmol-s" .mg".  , 1992). The results presented here indicate the presence in pea mitochondria of two distinct activities of PPIase. The association of the majority of the CsA binding activity with the pea mitochondrial matrix is in agreement with previous observations made with Neurospora and rat liver mitochondria (Tropschug et al., 1988, Halestrap andDavidson, 1990). Chloroplast PPZmes-Chloroplast is a special organelle in green plants which houses most of the biosynthetic reactions and a distinct, highly organized and regulated genome. The photosynthetic membranes (thylakoids) of chloroplasts are comprised of stacked (granal) membranes, enriched in photosystem I1 components, interconnected with non-appressed (stromal) membranes, enriched in photosystem I components (see Callahan et al., 1989). Thylakoid proteins are classifiable into two types. One class represents those proteins that translocate within the thylakoids and thus are located on both membrane types. The second class constitutes those proteins that do not translocate and are exclusive to one of the membrane types (Callahan et al., 1989). Correct protein folding is therefore highly desirable in this photosynthetic organelle, as is apparent from the discovery and involvement of protein chaperones in the assembly of some of the multisubunit complexes in this organelle (Goloubinoff et al., 1989;Ellis, 1991). We sought to look for the presence of other classes of protein foldases in the chloroplasts.
An active PPIase was present in the soluble (stroma) fraction of the purified chloroplasts, specific activity of which was lower than that in the mitochondrial matrix (compare Tables   I and 11). The chloroplast stroma PPIase was inhibited completely at 14 p~ CsA (Table 11). Relative to the ease with which PPIase activity could be determined with both of the mitochondrial fractions as well as the chloroplast stroma, chloroplast membranes (thylakoids) proved initially to be intractable. This inability to detect PPIase in the thylakoids was found linked to the presence of chlorophyll in these membranes, which interfered with the PPIase activity assay. To circumvent this problem, we solubilized the thylakoids and fractionated the proteins on the DEAE-Toyopearl column  Elution profile of solubilized thylakoids on DEAE-Toyopearl 650 S column. Thylakoids were solubilized in 2% Triton X-100 as described under "Experimental Procedures." Two-mg eq of chlorophyll were applied to a 2-cm column previously equilibrated with 50 mM Tris-NaOH, pH 7.2, containing 0.2% Triton X-100. The column was washed with 10 ml of the equilibration buffer, and then the bound proteins were eluted with 20 ml of 0-1 M NaCl gradient. Fractions (0.5 ml) were collected and analyzed for chlorophyll content (0) and CsA binding (A).

FIG. 4.
A, kinetics of the hydrolysis of N-succinyl-Ala-Ala-Pro-Phe-p-nitroanilidine were followed in the absence (0) or presence (0) of partially purified thylakoids (50 pg of protein/assay) as described under "Experimental Procedures." Thylakoid proteins were fractionated on a DEAE-Toyopearl650 S column before determining PPIase activity. The effect of 14 p~ CsA is also shown (+). B, dependence of the rate constant as a function of protein concentration. PPIase activity kinetics were followed as in A in triplicate each at 0,25,50, and 75 pg of protein of DEAE-purified thylakoid fraction/ assay. (Fig. 3) as described under "Experimental Procedures." The solubilized thylakoidal PPIase thus obtained exhibited first order kinetics (Fig. 4A), showed a linear increase in activity with increasing concentrations of the protein (Fig. 4B), and was strongly inhibited by CsA (Table 11).
CsA bound relatively stronger with the chloroplast fractions than the mitochondrial ones (Tables I and 11); this binding was to both the stroma and the thylakoids. However, in contrast to the peripheral association of the PPIase with the mitochondrial membranes, the binding associated with the thylakoids could not be removed by washing at alkaline pH (Table 11).
Further, a clear distinction was apparent between the two organelles in their abilities to bind rapamycin. While a crossreactive 25-kDa RBP was identified in the mitochondria and the mitochondrial PPIase activity was partially inhibited by rapamycin ( Fig. 2 and Table I), such was not the case with the chloroplast fractions. We were not able to detect any cross-reactive chloroplast RBP nor was the chloroplast PPIase inhibited by rapamycin (data not shown). It is possible that our inability to detect a cross-reactive RBP in the chloroplast fractions could be due to epitope differences and, thereby, non-recognition of any chloroplast RBP by an antibody against a heterologous, yeast RBP. Our data are consistent with the absence of a chloroplast RBP activity, but more direct results are awaited to prove this unequivocally.
In conclusion, these results demonstrate for the first time that both plant organelles share similar but distinct CsAbinding PPIases and that a putative RBP (rapamycin-sensitive PPIase) is selectively localized to the mitochondria. These data further add to the diversity of PPIases in nature and should lead to the isolation and identification of the corresponding genes encoding the organellar PPIases in plants. In this regard, the cloning of the plant cytosolic cyclophilin is encouraging (Gasser et al., 1990). We speculate that PPIases are involved in the correct folding of newly synthesized proline-containing soluble and membrane proteins in plant organelles in a fashion similar to that demonstrated with the protein chaperone, the rubisco-binding protein (Ellis, 1990), for the assembly of the rubisco protein complex.