Oxidative Stress Causes Rapid Membrane Translocation and in Vivo Degradation of Ribulose- 1 ,&bisphosphate Carboxylase/Oxygenase*

We have studied the turnover of an abundant chlo- roplast protein, ribulose-1,s-bisphosphate carboxylaseloxygenase (Rbu-Pz carboxylaseloxygenase), in plants (Spirodela oligorrhiza and Triticum aeetivum L.) and algae (Chlamydomonas reinhardtii and C. moewusii) induced to senesce under oxidative conditions. Rbu-Pz carboxylase/oxygenase activity and stability were to resulting cross- linking of large subunits by disulfide bonds within the holoenzyme, rapid and specific translocation of the soluble enzyme complex to the chloroplast membranes, and finally protein degradation. The redox state of Cys-247 in Rbu-Pz carboxylase/oxygenase large subunit sensitivity inactivation cross-linking. degradation Rbu-P, carboxylaseloxygenase senescence.

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisemnt" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The first two authors contributed equally to this paper. ll Present address: Dept. of Life Sciences, Ben Gurion University, Beer Sheva, Israel.
Although much insight into transcriptional, translational, and post-translational regulation of synthesis, assembly, and activity of Rbu-Pz carboxylase/oxygenase has been gained (1-8), our understanding of other important aspects of its function is not well defined. For instance, the protein accounts for about 4040% (w/w) of soluble chloroplast protein that accumulates during leaf expansion, mainly because of high rates of its synthesis, with minimal, almost unmeasurable, degradation. But soon after leaf expansion ceases and senescence ensues Rbu-P, carboxylase/oxygenase is rapidly degraded concomitant with a marked decrease in CO, assimilation rates (9-13). The onset of degradation of Rbu-Pz carboxylase/oxygenase and other proteins during leaf senescence has been speculated to provide nitrogen in the form of amino acids to young, developing leaves for growth (14). Thus, the senescing leaf has been used as a model for studying protein turnover (12). The recognition that protein turnover can occur within the chloroplast in opposition to the "lysosome" concept involving vacuolar proteases has shifted the emphasis from the vacuole to the photosynthetic organelle for studying Rbu-P2 carboxylase/oxygenase degradation (15). However, little is known about the pathway or signals that trigger Rbu-P2 carboxylase/oxygenase degradation. Studying Rbu-Pp carboxylase/oxygenase degradation during whole plant senescence is problematic, in part because of the long duration of the process and difficulties in analyzing pulse-chase experiments in the face of contributions from changing developmental processes. Detached leaves or leaf discs have frequently been used in place of whole plants, but results obtained from using such material confounds analysis because of the additional effects of injury. Alternatively, senescence is enhanced by subjecting plants to environmental or chemical stresses that, in turn, lead to inactivation or degradation of Rbu-Pz carboxylase/oxygenase (10, 16).
During chemical stress imposed by relatively high concentrations of cupric ions, higher plants undergo rapid physiological/biochemical changes comparable to those observed during normal senescence (17). We have used cupric ion-induced senescence in intact Spirodela oligorrhiza and wheat (Triticum aestiuum L.) plants and purified wheat chloroplasts to analyze Rbu-P2 carboxylase/oxygenase metabolism. We report here that oxidative conditions result in cross-linking of Rbu-Pz carboxylase/oxygenase via disulfide bridges, translocation of the protein to chloroplast membranes, and rapid degradation of the protein. We also show that the redox state of Cys-247 Ribulose-P2 Carboxylase/Oxygenase Turnover and Senescence 281 1 in the large subunit contributes to the sensitivity of Rbu-Pi carboxylase/oxygenase to oxidative inactivation and crosslinking and that this process is not coupled to membrane translocation or degradation of Rbu-P2 carboxylase/oxygenase in vivo.

EXPERIMENTAL PROCEDURES
Plant Material and Treament with Cupric Ions-Axenic cultures of S. oligorrhiza (Kurtz) Hegelm. were grown phototrophically (15-20 pmol m-* s-', 400-700 nm light, 25 "C) for 10-15 days in halfstrength Hutner's mineral medium (18) containing 1% sucrose. Chlamydomonas reinhardtii and Chylamydomonas moewwii were grown to midlog phase in TAP medium (19) at room temperature, bubbled continuously with house air under fluorescent light (70-80 pmol m-* s-I). Wheat (T. aestiuum L.) seeds were planted directly in wet vermiculite. Seedlings were grown for 5-6 days in the dark in a growth chamber at 22 "C, and prior to use they were held for 12 h in the light followed by 12 h in darkness. In the case of Spirodela and Chlamydomonas, a stock solution of copper sulfate was directly diluted in the growth media and the organisms further incubated. Wheat plants were removed from the vermiculite, roots carefully cleaned with distilled water, and the plantlets were placed in a beaker containing water or 10 mM copper sulfate. These plantlets were incubated for 19 h at 25 "C under 15-20 pmol. m-*.s", 400-700 nm light. Isolated wheat chloroplasts were washed and pelleted as described below and the pellets resuspended in sorbitol/Hepes, pH 8, with or without 1 mM copper sulfate.
Isolation of Wheat Chloroplasts-Intact wheat chloroplasts were isolated as described (20) using a Percoll gradient of 10-90%. The chloroplasts were resuspended in 330 mM sorbitol and 50 mM Hepes/ KOH, pH 8.0. Wheat chloroplast stroma and membranes were isolated as follows. After each treatment the chloroplasts were pelleted by letting the centrifuge accelerate to 3,500 X g followed immediately by deceleration. The intact chloroplast pellet was resuspended in the homogenization buffer and vortexed. The stroma fraction was separated from the membranes by centrifugation at 6,750 X g for 10 min. The membranes were washed as described below for total Spirodela membranes.
In Viuo Radiolabeling-Prior to incubation with [3H]leucine (143.7 Ci/mmol; Du Pont-New England Nuclear) for radiolabeling proteins, Spirodela plants were transferred for 24 h to the mineral medium without sucrose. Conditions for protein labeling and in vivo chase of radiolabeled proteins are described in the legend to Fig. 3.
Isolation of Soluble and Membrane-associated Proteins-Spirodela fronds and wheat leaves were homogenized in a medium containing 50 mM NaCl, 50 mM Tris-HC1, pH 7.4, 20 mM MgC12, 0.1 mM EDTA, 10% glycerol, 2 mM phenylmethylsulfonyl fluoride, 2 pg/ml leupeptin, and 5 mM 8-mercaptoethanol. Soluble and membrane fractions were isolated as described previously (21). Membrane fractions were washed two times in the homogenization medium containing 300 mM NaCl followed by two washes in 10 mM Tris-HC1, pH 8.0, and 10 mM Tris-HC1, pH 7.5. C. reinhardtii and C. moewusii cells were pelleted by centrifugation at 2,000 X g for 5 min, washed twice in ice-cold TAP medium, and then resuspended to obtain a chlorophyll concentration of 65 pglrnl. After treatment with 1 mM CuSO, in TAP medium, the cells were pelleted, resuspended in the homogenization medium described above, and broken in an ice-cold French press at 20,000 p s i . Soluble and membrane fractions were separated and washed as described above. Protein content in soluble fractions was determined by the dot blot assay (22) or Bio-Rad color reaction using bovine serum albumin as a standard. Membrane fractions were equalized on the basis of chlorophyll content determined by the standard method (23).
Gel Electrophoresis and Immunoblotting-Proteins were fractionated using nondenaturing polyacrylamide (7%) gels and denaturing SDS-polyacrylamide 10-20% gradient gels. Each sample was applied to the gels on either equal protein (soluble proteins) or equal chlorophyll (membrane proteins) or equal radioactivity bases as indicated in the legend to each figure. A part of each gel was stained with Coomassie Blue or fluorographed and the remaining parts electrotransferred onto nitrocellulose for immunodetection (24) with antisera against specific proteins as described in the appropriate figure legend.
Each gel and immunoblot were repeated a minimum of four times. Reproducible results were obtained each time, and typical results are presented here.

Inactive Rbu-P, CarboxylaselOxygenase Associates
with Membranes-Spirodelu, wheat plants, and wheat chloroplasts were incubated with or without copper sulfate for different times, following which soluble and membrane-associated proteins were isolated and fractionated by SDS-PAGE. The results, shown in Fig. 1, indicate a steady decline in the levels of soluble proteins fractionating at about 55 kDa and 14 kDa in plants treated with cupric ions concomitant with an increased accumulation of similar size proteins in the membrane fraction ( Fig. lA). Since large and small subunits of Rbu-P2 carboxylase/oxygenase, respectively, fractionate at 55 and 14 kDa during SDS-PAGE, immunodetection using anti-Rbu-P2 carboxylase/oxygenase antibodies was carried out. It is clear from Fig. 1B that the 55-and 14-kDa proteins in the soluble pool which are markedly affected by cupric ion toxicity indeed represent large and small subunits of Rbu-Pz carboxylase/ oxygenase, respectively. These results are consistent with previous observations of increased in vitro proteolysis of large subunit of Rbu-P2 carboxylase/oxygenase under oxidative treatments (25). The presence of lower molecular weight polypeptides immunoreactive with antibodies against Rbu-Pp carboxylase/oxygenase, particularly in isolated wheat chloroplasts, is further evidence of degradation of large subunit of Rbu-P2 carboxylase/oxygenase under these conditions (Fig. lC, compare lanes 6-8 with lune 5 ) . Inability to detect discrete breakdown products of Rbu-P2 carboxylase/oxygenase in Spirodela or intact wheat plants may be caused by faster rates of degradation and/or loss of immunological epitopes in the breakdown products. However, association of these proteins with membranes was somewhat surprising. Therefore, to rule out nonspecific adhesion of Rbu-P2 carboxylase/oxygenase to membranes, the membrane fractions were thoroughly washed with 0.3 and 1.5 M NaCl as well as with low ionic strength buffers at pH values of 7.5 and 8.0. None of these treatments was effective in dissociating the immunoreactive Rbu-P, carboxylase/oxygenase subunits from the membranes.
The disappearance of soluble Spirodela and wheat Rbu-P2 carboxylase/oxygenase subunits closely correlated with a corresponding decrease in the Rbu-Pp carboxylase/oxygenase carboxylation activity which was, however, not recovered with the membranes (data not shown), suggesting that membraneassociated Rbu-P2 carboxylase/oxygenase did not represent an active enzyme.  (Fig. 1B). The quantified data for the membrane translocation kinetics of large subunit of Rbu-P2 carboxylase/oxygenase in Spirodelu and wheat chloroplasts are presented in Fig. 2. A relatively small, but measurable, amount of Rbu-P2 carboxylase/oxygenase was found consistently associated with the membranes from con- To determine if the changes observed in the steady-state levels of proteins were caused by differential degradation rates, pulse-chase experiments were carried out. Plants were pulse labeled for 3 h with [3H]leucine, and then radioactivity in proteins was chased for different time periods with or without 1 mM cupric sulfate in the mineral medium containing nonradioactive leucine (1 mM). Soluble and membrane proteins were isolated and fractionated by SDS-PAGE. Fluorographs of these gels are presented in Fig. 3.

Steady-state Level of Soluble Rbu-Pp CarboxylaselOxygenase Decreases during Cupric Ion Toxicity-The
Pulse-chase experiments showed that soluble proteins turn in particular the turnover of Rbu-P2 carboxylase/oxygenase was more striking (Fig. 3, Soluble). However, some amount of radiolabeled Rbu-P2 carboxylase/oxygenase was found associated with membranes only in samples incubated for 8 and 18 h (Fig. 3, Membrane). Clearly, the membrane-associated radiolabeled Rbu-P, carboxylase/oxygenase also turned over appreciably in the 18-h sample. These results are different from those of the steady-state distribution of Rbu-P2 carboxylase/oxygenase during cupric ion treatment (Fig. lA), indicating that membrane association of Rbu-P2 carboxylase/ oxygenase is transient and that membrane-associated Rbu-P:! carboxylase/oxygenase undergoes turnover. Also evident from data in Fig. 3 is that, among other changes, the degradation of light harvesting chlorophyll a/b apoprotein, and the D l  (22). The positions of large subunit ( L S ) and small subunit ( S S ) of Rbu-P, carboxylase/oxygenase, the 32-kDa protein of the photosystem I1 (32), and the light-harvesting chlorophyll a/b apoprotein (LHCP) are indicated. Samples applied on the gels contained 20,000 cpm (soluble fraction) or 80,000 cpm (membrane jruction) of hot trichloroacetic acid-precipitable radioactivity.
Steady-state membrane protein levels have been reported to change at different rates during senescence (26, 27). Similarly, cupric ions cause inactivation of chloroplast photosystems of some (28,29) but not all plants (30). However, in contrast to the effect on Rbu-P1 carboxylase/oxygenase demonstrated above, the steady-state levels of several chloroplast membrane proteins did not change as markedly within the time frame of cupric ion treatment of intact 'plants, results that are consistent with previous observations using intact spinach plants (30). This is particularly evident from immunoblots for the extrinsic 33-kDa oxygen-evolving complex protein, light-harvesting chlorophyll a/b apoprotein, subunit 1 of the photosystem I, 32-kDa D l protein of the photosystem 11, and plastocyanin (24; Fig. 4). The p-subunit of ATPase, on the other hand, declined in abundance as the treatment of Spirodela with cupric ions increased to 4 h. Similar instability of CF1-ATPase was shown to occur during senescence of wheat leaves (31) but not in oat leaves (26). Overall, these results indicate that the majority of photosynthetic membrane proteins remain more or less unaffected during the early period (several hours) of cupric ion-induced senescence. Thus, one of the earliest and most profound consequences of this phenomenon is the instability of Rbu-P, carboxylase/oxygenase and its translocation from the soluble fraction to the membrane fraction of the chloroplast.

Dimerization of Rbu-P, Carboxyylase/Oxygenase Involving Sulfhydryl Group(s) Occurs in Concert with Its Membrane
Translocation-What triggers the disappearance and membrane translocation of Rbu-Pz carboxylase/oxygenase in cupric ion-treated plants? Cupric ions, as transition elements, are strong oxidants (32) and can change oxidation-reduction potential in a biological environment. This in turn might adversely affect macromolecules such as proteins, the degree of damage being dependent upon the micro milieu and the presence of, yet to be defined, sensitive amino acid residues (33). Since large subunits of Rbu-P, carboxylase/oxygenase have been shown to cross-link in vitro (34,35), we sought to check the possibility that oxidative conditions caused by cupric ions might result in the in vivo cross-linking of Rbu-P, carboxylase/oxygenase molecules involving cysteine resi-

dues.
Spirodela, wheat plants, and isolated wheat chloroplasts were incubated with or without CuS04, and soluble proteins were isolated. Samples were then boiled with the sample application buffer with or without P-mercaptoethanol and fractionated by SDS-PAGE. Results are presented in Fig. 5. The elimination of P-mercaptoethanol from a parallel set of samples was expected to maintain sulfhydryl groups in an oxidized configuration and thus enable cross-linked proteins to electrophorese more slowly under nonreducing but denaturing conditions. Indeed, we found a protein band of -110-120 kDa in cupric ion-treated Spirodela and wheat chloroplast samples electrophoresed under nonreducing conditions, the appearance of which occurred concomitant with the disap- pearance of the large subunit at -55 kDa (Fig. 5). In the treated intact wheat samples, both the protein bands appeared to be absent. This was attributed to longer time period used in this experiment which resulted in the degradation of both the protein forms, since in short term experiments with intact wheat plantlets, we did observe a trend similar to that in Spirodela and wheat chloroplasts. The -110-120-kDa protein, representing a large subunit dimer, was either absent or present in very low levels in the nontreated controls. When identical cupric ion-treated samples were fractionated under reducing conditions, the doublet disappeared (cf. Fig. 1). The concentration of the Rbu-P, carboxylase/oxygenase small subunit appeared not to change under these conditions (data not shown). These results suggest that large subunits in the holoenzyme are cross-linked in vivo under oxidative conditions.
If large subunits of the Rbu-Pz carboxylase/oxygenase holoenzyme were oxidized and then cross-linked via disulfide bonds, we surmised that the oxidized holoenzyme might be separable from the unoxidized form under nondenaturing conditions. Thus, soluble proteins from control and cupric ion-treated Spirodela plants were fractionated on 7% polyacrylamide gels under nondenaturing conditions and either stained with Coomassie Blue (Fig. 6A, lanes 1-5) or immunoblotted (Fig. 6A, lanes 6-10). A perceptible shift in the mobility of Rbu-P, carboxylase/oxygenase was apparent as the period of incubation with cupric ions was increased. Also, the level of the Rbu-P, carboxylase/oxygenase holoenzyme appeared to decrease (Fig. 6A), suggesting that the holoenzyme concentration in the soluble compartment was negatively affected by cupric ions. To confirm the identity of the stained protein bands, these were excised, reelectrophoresed under reducing and nonreducing conditions on denaturing SDS-polyacrylamide gels, and subjected to Western blot analysis using antibodies against Rbu-P, carboxylase/oxygen-  1 and 6 ) . Soluble proteins were isolated, electrophoresed on native PAGE (12 pgllane), and either stained with Coomassie Blue (lanes 1-5) or electrotransferred onto nitrocellulose paper and immunoreacted with anti-Rbu-P? carboxylase/oxygenase antibodies (lanes 6-10). The immunoblot was overdeveloped to ensure the presence of the slow moving protein form and is shown for qualitative comparison only. B, the stained bands corresponding to Rbu-P2 carboxylase/oxygenase protein in A were excised from the gel and then rerun on SDS-PAGE in the presence (lanes 1-5) or absence (lanes 6-10) of 100 mM DTT. Note that in the absence of DTT, dimers of Rbu-P2 carboxylase/oxygenase large subunit were formed, at -110-120 kDa, only in samples from plants incubated with CuSO,.
ase. Results obtained confirmed the presence of both the large (Fig. 6B) and small subunits (data not shown) of Rbu-P:! carboxylase/oxygenase in the parent bands. Further, it became evident that dimerization of Rbu-Pz carboxylase/oxygenase via disulfide bonds in the cupric ion-treated samples occurred between large subunits in assembled Rbu-P, carboxylase/oxygenase. This was evidenced by the presence, particularly in cupric ion-treated samples, of an immunodetectable Rbu-P, carboxylase/oxygenase holoenzyme that was electrophoretically less mobile than the parent enzyme (Fig. 6A, Immunoblot).
Antibody data with anti-large subunit-binding protein (not shown) indicated that the mobility shift in Rbu-P:! carboxylase/oxygenase on native gels was not caused by the association of the binding protein. It is of interest to note that the steady-state level of the large subunit-binding protein also decreased in plants under oxidation stress, with the &subunit being more unstable than the a-subunit (data not shown).
Membrane Association of Rbu-Pz Carboxylase/Oxygene Is Independent of -S-S-Cross-linking-The above results substantiate previous in uitro reconstitution studies with chemically disassembled wheat Rbu-P2 carboxylase/oxygenase (34), and x-ray diffraction studies of recombinant Synechococcus Rbu-Pa carboxylase/oxygenase expressed in E. coli (35), in which dimerization of the large subunits was shown to occur via -S-Scross-linking. We thought it possible that the same single cysteine residue, identified in spinach Rbu-P, carboxylase/oxygenase as 36), was involved in Rbu-P, carboxylase/oxygenase cross-linking in the organisms tested here since the spatial arrangement of this particular cysteine residue is such that its disulfide cross-linking with another large subunit does not disturb the assembled but oxidized Rbu-P, carboxylase/oxygenase.
The generality of disulfide cross-linking in Rbu-P2 carboxylase/oxygenase implies that Cys-247 and its micro milieu may have special characteristics that make the protein highly sensitive to oxidative conditions. In fact, Cys-247 is conserved in almost all the Rbu-P, carboxylase/oxygenase large subunits sequenced thus far (37). An interesting exception is the C. moewusii protein, in which it is replaced by serine (38). It was, therefore, of interest to investigate if, in this alga, Rbu-P, carboxylase/oxygenase is refractory to oxidative damage and if disulfide bridge formation is linked to membrane translocation of Rbu-Pz carboxylase/oxygenase.
Phototrophic cultures of C. reinhardtii and C. moewusii were incubated with or without 1 mM CuS04 for 0.5-4 h, and cells were then harvested and lysed. Soluble and membrane proteins were isolated, fractionated on SDS-polyacrylamide gels under reducing and nonreducing conditions, and immunoblotted using a mixture of antibodies against Rbu-PZ carboxylase/oxygenase large and small subunits.
Under reducing conditions, cupric ion treatment of both algal cultures resulted in a time-dependent loss of the Rbu-Pz carboxylase/oxygenase subunits from the soluble pool and their translocation to the membranes (Fig. 7A). These data are consistent with other results presented above for Spirodela and wheat. Moreover, when the same set of soluble samples in Fig. 7A was electrophoresed under nonreducing conditions (i.e. in the absence of DTT) and reacted with the anti-Rbu-P2 carboxylase/oxygenase antibodies, it became evident that Rbu-P, carboxylase/oxygenase from C. reinhardtii was reversibly cross-linked via disulfide bonds (Fig. 7B, lanes 17-20) whereas Rbu-P, carboxylase/oxygenase from C. moewusii was not and appeared impervious to oxidation stress in this regard (Fig. 7B, lanes 7-10). Nonetheless, even in the absence of cross-linking, C. moewusii Rbu-P, carboxylase/oxygenase was found to translocate to membranes under these conditions (Fig. 7A) (lanes 2, 7,12,  and 17), 1 h (lanes 3,8,13, and 18), 2 h (lanes 4 , 9 , 1 4 , and 19) and 4 h (lanes 5, 10, 15, and 20) with (+) or without (-) 1 mM CuSO,. After incubation, cells were washed and lysed. Soluble (Sol) and membrane (Memb) proteins were isolated, fractionated under reducing conditions by SDS-PAGE, immunoblotted, and probed with antibodies against Rbu-P2 carboxylase/oxygenase. Lanes I , 6, 11, and 16 correspond to control samples incubated without CuSO, for 4 h. B, soluble proteins from the two algal cultures were isolated as in A and electrophoresed in the presence (+) or absence (-1 of 100 mM DTT. Cross-linked dimer (upper arrow) is seen only in the samples from C. reinhardtii incubated with cupric ions and electrophoresed in the absence of DTT (lanes [17][18][19][20]. Lower arrows indicate the position at which the large subunit electrophoresed. Equal amounts of soluble protein (12 pg) or membrane protein (4 pg) were applied on the gels. Cys-247 is not required for Rbu-P, carboxylase/oxygenase translocation to membranes. Thus, the two processes seem to occur independent of each other in vivo. DISCUSSION We have demonstrated that Rbu-P, carboxylase/oxygenase, an abundant chloroplast protein, is highly sensitive to oxidative stress. As a result, Rbu-P, carboxylase/oxygenase undergoes in vivo -S-Scross-linking (most probably at Cys-247 of large subunit in the assembled holoenzyme), inhibition of enzyme activity, membrane translocation, and, finally, degradation. Further, our data show that membrane translocation of Rbu-Pz carboxylase/oxygenase occurs independent of disulfide cross-linking. Thus, oxidative stress affects Rbu-P2 carboxylase/oxygenase stability in more than one way.
The results described here may have relevance to the precipitous enhancement in the inactivation and degradation of Rbu-P, carboxylase/oxygenase protein during plant senescence. Since the redox state of Cys-247 seems to determine the sensitivity of Rbu-P, carboxylase/oxygenase to inactivation and cross-linking ( Fig. 7; 25), a highly reduced environment must be maintained by the plastid to ensure the stability of Rbu-P, carboxylase/oxygenase during normal growth. When these protective mechanisms in the cell break down, for instance, during senescence and stress, a change in redox to a more oxidized state might result in the instability of Rbu-Pz carboxylase/oxygenase. Indeed, indirect evidence has been presented to show that during senescence more oxidative conditions exist in the chloroplast (39). Furthermore, recently it has been reported that removing fruit from soybean plants causes formation of insoluble Rbu-P, carboxylase/oxygenase in leaf extracts (40). Thus, the oxidation-reduction state of the chloroplast stroma appears closely associated with shifts in the leaf from a normal growth/maturation stage to senescence and death.
Translocation of oxidized Rbu-P, carboxylase/oxygenase to the chloroplast membranes may be a mechanism for the regulation of its turnover, particularly during senescence. In its oxidized and membrane-associated conformation Rbu-P, carboxylase/oxygenase may be more prone to proteolysis. The nature and type of the protease that specifically recognizes Rbu-P, carboxylase/oxygenase in vivo during senescence and degrades it have remained puzzling questions about Rbu-P, carboxylase/oxygenase biology. Many studies have demonstrated involvement of proteases in the degradation of Rbu-P2 carboxylase/oxygenase in vitro. However, none of these i n vitro studies shows the specificity expected of an in vivo proteolytic system for Rbu-P, carboxylase/oxygenase (41-43). It is possible that it is the oxidized form of Rbu-P, carboxylase/oxygenase that is the actual/natural substrate for its specific protease, and the inability to find a Rbu-P, carboxylase/oxygenase-specific protease may be because the substrate tested in all such studies was the reduced protein. In this context, it is of interest that in vitro studies using general proteases, uiz. trypsin, chymotrypsin, proteinase K, and papain, have shown enhanced degradation of the oxidized form compared with reduced Rbu-P, carboxylase/oxygenase as the substrate (25).
Difficulties encountered in the isolation of a Rbu-P, carboxylase/oxygenase-specific protease may also be linked to yet another possibility raised by the data reported here. Since oxidative conditions result in membrane translocation of Rbu-P2 carboxylase/oxygenase prior to its degradation in vivo, it may be that Rbu-P, carboxylase/oxygenase degradation is in fact catalyzed by a membrane-associated protease rather than a stroma one. Alternatively, membrane association of oxidized Rbu-P, carboxylase/oxygenase may act as a scaffold providing the right conformation for a stroma or membrane protease to act. Implicit in such a possibility is the involvement of a membrane binding step for oxidized (damaged/modified) Rbu-P, carboxylase/oxygenase during its degradation. If this is the case, then oxidized Rbu-Pz carboxylase/oxygenase will have higher affinity for the membrane than will the native, undamaged protein. Indeed, in our preliminary reconstitution experiments using salt-washed chloroplast membranes and gel-purified radiolabeled Rbu-P2 carboxylase/oxygenase, we have found a higher affinity of oxidized, rather than reduced, Rbu-P2 carboxylase/oxygenase protein for chloroplast membranes.