Partial proteolysis as a probe of the conformation of the gamma subunit in activated soluble and membrane-bound chloroplast coupling factor 1.

Treatments that enhance the latent ATPase activity of the chloroplast coupling factor (CF1) also induce hypersensitivity of the gamma subunit toward trypsin. A number of different gamma subunit cleavage products are formed (Moroney, J. V., and McCarty, R. E. (1982) J. Biol. Chem. 257, 5910-5914). We have compared the gamma cleavage products of membrane-bound and isolated CF1, activated either by reduction of the gamma disulfide bond or by removal of the epsilon subunit. The gamma subunit of isolated CF1 lacking the epsilon subunit was cleaved to a 27,000-Da species. The same cleavage site became exposed following energy-dependent conformational changes in the membrane-bound enzyme. Activation by reduction of the gamma disulfide bond also exposed this site. However, the gamma subunit of reduced CF1 was cleaved rapidly at an additional site and trypsin treatment gave rise to a 25,000-Da gamma species. The small peptide generated by the second cleavage contains one of the cysteinyl residues of the reduced disulfide bridge of gamma. This peptide dissociates from the enzyme and can be isolated by gel filtration. The close proximity of the trypsin cleavage sites to the disulfide bond of gamma is discussed with respect to the effects of tryptic cleavage on the ATPase activity of CF1. The data indicate that structural changes in a limited region of the gamma subunit strongly influence the catalytic properties of both soluble and membrane-bound CF1.

Partial Proteolysis as a Probe of the Conformation of the y Subunit in Activated Soluble and Membrane-bound Chloroplast Coupling Factor l* (Received for publication, April 23, 1985) Jiirgen SchumannS, Mark L. Richter Treatments that enhance the latent ATPase activity of the chloroplast coupling factor (CFI) also induce hypersensitivity of the y subunit toward trypsin. A number of different y subunit cleavage products are formed (Moroney, J. V., and McCarty, R. E. (1982) J.
We have compared the y cleavage products of membrane-bound and isolated CF,, activated either by reduction of the y disulfide bond or by removal of the t subunit. The y subunit of isolated CF, lacking the t subunit was cleaved to a 27,000-Da species. The same cleavage site became exposed following energy-dependent conformational changes in the membranebound enzyme. Activation by reduction of the Y disulfide bond also exposed this site. However, the 7 subunit of reduced CF1 was cleaved rapidly at an additional site and trypsin treatment gave rise to a 25,000-Da 7 species. The small peptide generated by the second cleavage contains one of the cysteinyl residues of the reduced disulfide bridge of y. This peptide dissociates from the enzyme and can be isolated by gel filtration. The close proximity of the trypsin cleavage sites to the disulfide bond of y is discussed with respect to the effects of tryptic cleavage on the ATPase activity of CF,. The data indicate that structural changes in a limited region of the y subunit strongly influence the catalytic properties of both soluble and membranebound CF,.
Chloroplast coupling factor 1 catalyzes the light-dependent formation of ATP from ADP and Pi (1,2). CF,' is comprised of five different polypeptide subunits with a proposed stoichiometry of (3). The two larger polypeptides, a and p, are thought to contain the catalytic site(s) (4). The y subunit is involved in regulation of catalysis and may also function in regulating the flow of protons through the hydrophobic proton channel (CFo) portion of the coupling factor complex (1). Both the 6 and t subunits are required for the tight coupling *This work was supported by Grant PCM 83-06747 from the National Science Foundation and by Hatch funds. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18U.S.C. Section 1734 solely to indicate this fact.
$ Recipient of a fellowship from the Deutsche Forschungsgemeinschaft.
of ATP synthesis to proton translocation ( 5 ) . The t subunit is a potent inhibitor of the ATPase activity of the isolated enzyme (6).
Four cysteinyl residues have been identified in the y subunit (7). One residue reacts with SH-directed probes in membranes kept in the dark and is therefore called the "dark-accessible site." Another residue reacts with probes only under energized conditions ( i e . in the light) and is called the "light-accessible site." The other two cysteinyl residues form a disulfide bond within y and are only accessible to labeling after reduction, for example by dithiothreitol.
Although illuminated thylakoids catalyze rapid rates of ATP synthesis, the rates of ATP hydrolysis in the dark are very low unless the thylakoids are preilluminated in the presence of thiol reagents (8). Isolated coupling factors of mitochondria and bacteria are active per se in ATP hydrolysis (9), while isolated CF, is inactive and may be activated by a variety of methods (1,2).
Activation of the ATPase of CF, in solution by thiol reagents results from reduction of the disulfide bond in the y subunit (lo), whereas heat (5, ll), detergents (12), and alcohols (6) all cause dissociation of the t subunit. Activation of the ATPase of oxidized CF, by proteases was originally thought to result from digestion o f t (13). However, activation appears to correlate well with digestion of the a subunits (14).
Pretreatment of CF1 with dithiothreitol or with heat both induce a hypersensitivity of the y subunit toward proteolytic cleavage by trypsin (14). Cleavage of y under these conditions results in a further increase of the ATPase activity of the enzyme. A similar hypersensitivity of y toward trypsin was observed with membrane-bound CF, during illumination of thylakoids (15). We have investigated the tryptic cleavage products of the y subunit in further detail. Two trypsin cleavage sites have been identified. One of these sites is exposed by removal of the E subunit from isolated CF,. The same site also becomes exposed in CF1 attached to the thylakoid membrane during illumination. Reduction of the y disulfide bond leads to a conformation of CF, in which both cleavage sites become exposed to trypsin with either the isolated or the membranebound enzyme.

MATERIALS AND METHODS
Chloroplast thylakoids were prepared from fresh market spinach (16). CF, was prepared by a modification (17)  Dithiothreitol activation of the isolated enzyme was carried out by incubation in 50 mM Tris-HCI (pH 8 ) , 2 mM EDTA for 2 h at room temperature in the presence of 50 mM dithiothreitol for CF, (10) or 15 min at room temperature in the presence of 20 mM dithiothreitol for c-free CF,? Dithiothreitol was removed by gel filtration. For membrane-bound CF,, thylakoids (equivalent to 0.1-0.2 mg of chlorophyll/ml) were illuminated for 5 min with white light (2000 watts/ m2) in a reaction mixture containing 50 mM Na+-Tricine (pH 8.0), 50 mM NaCI, 5 mM MgC12, 25 p M pyocyanine (illumination medium), and 5 mM dithiothreitol. Thylakoids were collected by centrifugation and resuspended in the illumination medium.
For trypsin treatment of membrane-bound CFI, thylakoids (equivalent. to 0.1-0.2 mg of chlorophyll/ml) were incubated for 5 min at room temperature in the illumination medium with trypsin (10-20 Gg/ml) either in the dark or in the light. Soybean trypsin inhibitor (60 pg/ml) was added to stop digestion. For trypsin treatment of isolated CFI or t-free CF,, 2 pg of trypsin were added per 100 fig of enzyme preparation in 50 mM Tris-HCI (pH 8.0), 2 mM EDTA. The reaction was terminated by addition of soybean trypsin inhibitor (50 pglml) or trichloroacetic acid (0.5 mM).
For the identification of cysteine-containing peptides, the SHreactive probe 6-acryloyl-2-dimethylaminonaphthalene (Acrylodan, Ref. 23) was used as a fluorescent label. This probe was described as mainly SH-directed (23), and CF1 labeled with Acrylodan under various conditions behaves like CF1 labeled with fluorescent maleimides with respect to distribution of label on y and c and to trypsin digestion patterns. Pretreatment of CF, with N-ethylmaleimide reduces the amount of incorporated Acrylodan by more than 90%, indicating that labeling of amino groups is low under the conditions employed. The probe was dissolved in dimethyl formamide and was stable for several weeks at 4 "C. The final concentration of dimethyl formamide in reaction mixtures was less than 1% (v/v). Acrylodan exhibits a light absorption maximum at 380 nm with an extinction coefficient of 16 X lo3 cm2/mmol at pH 8 (23). The fluorescence emission maximum (excitation 380 nm) of the probe reacted with mercaptoethanol in water is at 540 nm. Reaction products of Acrylodan with CF1, however, show emission maxima over the range of 475-485 nm, indicative of a hydrophobic environment (23). The disulfide cysteinyl residues of the y subunit of membrane-bound CFI were labeled as follows. Reactive sulfhydryls were blocked by incubation of thylakoids (equivalent to 1 mg of chlorophyll/ml) with Nethylmaleimide (10 mM) for 15 min at room temperature. Thylakoids were diluted with illumination medium containing 5 mM dithiothreitol to a final concentration equivalent to 0.16 mg of chlorophyll/ ml and further incubated for 5 min in the light. The membranes were washed once with 50 m M Na+-Tricine (pH 8), 50 mM NaC1, 5 mM M g Q , and resuspended in the illumination medium. Acrylodan was added to a final concentration of 10 +M and the thylakoids were incubated for 5 min at room temperature. Dithiothreitol (0.1 mM) was added to react with any remaining free Acrylodan.
Caz+-dependent ATP hydrolysis was measured for 2-5 min at 37 "C in the presence of 50 mM Tris-HC1 (pH 8), 5 mM ATP, 5 mM CaC12, and 2 pg/ml of enzyme preparation. Pi was determined spectrophotometrically (24). SDS-gel electrophoresis was carried out using published procedures (25, 26). Absorbance measurements were recorded using a Beckman DU-7 spectrophotometer. Fluorescence measurements were made using a Farrand fluorescence spectrophotometer. Tosylphenylalanyl chloromethyl ketone-treated trypsin was purchased from Worthington, soybean trypsin inhibitor from Sigma, and Acrylodan from Molecular Probes, Inc. All other chemicals were high quality commercial grades and all aqueous solutions were prepared with distilled, deionized water.

RESULTS
Cleavage of Membrane-bound CF, by Trypsin-CF1 was isolated from thylakoid membranes treated with trypsin in the dark or in the light. The thylakoids were either illuminated in the presence of dithiothreitol or kept in the dark in the absence of dithiothreitol prior to trypsin treatment. After isolation, ATPase activities were measured (Table I) and samples were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Fig. 1). Consistent with a previous report (151, CFl from membranes treated with trypsin in the dark had only a slightly higher Ca2+-ATPase activity than control CF, ( Table I). The amount of y in this preparation was not reduced and no degradation products were seen (Fig.  1, lane B ) . Trypsin treatment in the light, however, resulted in much higher ATPase activities and most of the y subunit was degraded to a smaller polypeptide of about 27 kDa ( Fig.  1

, lane C).
Pretreatment of thylakoids with dithiothreitol in the light, which causes reduction of the disulfide bond in the y subunit (27), and subsequent incubation with trypsin either in the dark or in the light resulted in CF1 preparations with Ca2+-ATPase activities similar to that of CF, isolated from thylakoids trypsin-treated in the light (Table I). Again, the y subunit of these preparations was clipped, but this time the major cleavage product was a polypeptide of approximately 25 kDa (lanes D and E ) . The amount of y remaining was much higher in the sample from thylakoids treated with trypsin in the dark (lane D ) than in that from thylakoids exposed to trypsin in the light (lane E ) . It should be pointed out that all samples were exposed to a high concentration of dithiothreitol prior to electrophoresis. This eliminated the possibility that the disulfide bond in oxidized CF, altered the migration of the 27-kDa species. Trypsin cleavage of y to either the 25-kDa or 27-kDa species also resulted in a small peptide visible at the gel front. This species, as judged by electrophoresis on higher percentage gels, is about 6 to 8 kDa and is probably the same fragment previously shown to contain the dark-accessible cysteinyl residue of y (28). Cleavage of Isolated CF1 by Trypsin-The y subunit of oxidized CF, is relatively stable during short-term incubation (<5 min) with low concentrations of trypsin (14). Longer incubations with trypsin result in cleavage of all five subunits with a concomitant increase in the Ca2+-dependent ATPase activity of the enzyme to a maximum value after about 60-   1 (left). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of CF, purified from trypsintreated thylakoids. Thylakoids were treated with trypsin (10 pg/ml) for 5 min in the dark or in the light. Soybean trypsin inhibitor (50 pglml) was added and the CF, was isolated. Each lane contained 30 pg of CF1. Samples were electrophoresed on 12% polyacrylamide gels and stained with Coomassie blue. Lane A, control CF, without trypsin treatment. B, CF, from thylakoids trypsin-treated in the dark. C, CF1 from thylakoids trypsintreated in the light. D, CFI from dithiothreitol-treated thylakoids trypsin-treated in the dark. E, CF, from dithiothreitol-treated thylakoids trypsin-treated in the light. The numbers indicate molecular masses in kilodaltons of fragments derived from the y subunit. FIG. 2 (center). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of trypsin-treated CF1 and t-free CFI. Isolated CF, was incubated with trypsin (2 pg/lOO pg of CF,) for 5 min and precipitated with trichloroacetic acid. Each lane contained 50 pg of the enzyme preparation. Lane A , control CF,; B, control CF, + trypsin; C, dithiothreitol-treated CF, + trypsin; D, e-free CF,; E, e-free CF1 + trypsin; F, dithiothreitol-activated c-free CF, + trypsin; G, CF, treated with trypsin on illuminated thylakoids (see Fig. 1, lane C). Only y, 8, and e subunits are shown. The numbers indicate molecular masses in kilodaltons. FIG. 3 (right). Trypsin digestion of Acrylodan-modified CF,. The disulfide cysteines of isolated CF, were modified with Acrylodan as described under "Materials and Methods." Trypsin (2 pg/100 pg of CFJ was added and aliquots were taken at the times indicated and precipitated with trichloroacetic acid. Each lune contained 10 pg of enzyme preparation. Electrophoresis was carried out on a 14% polyacrylamide gel. Lane 1, CF, labeled with Acrylodan after sodium dodecyl sulfate treatment; ssu, small subunit of ribulose-bisphosphate carboxylase. The arrows point to fluorescent fragments derived from they subunit; numbers indicate molecular masses in kilodaltons.

TABLE I1
Ca2+-ATPase activities of isolated CF, CFI and e-free CFI, with or without dithiothreitol activation, were treated with trypsin (2 pg/100 pg of CF1) for 5 min, soybean trypsin inhibitor was added (50 pg/lOO pg of CF,), and Ca2+-ATPase activities were determined.

CF, preparation Ca*+-ATPase activity
Control CF, +Trypsin (5 min) +Trypsin (60 min) +Trypsin (5 min) +Trypsin (5 min (14). Pretreatment of CF1 with dithiothreitol(l4) or removal of the t subunit (Ref. 14 and Footnote 2) results in a rapid cleavage of the y subunit by trypsin that is complete within a few minutes. The cleavage products of reduced CF, and CF,(-t) (Fig. 2) and Ca2+-ATPase activities (Table 11) of the preparations were compared. Close to the maximum Ca2+-ATPase activity attainable with this preparation of CF, (25 pmol of Pi/min/mg) was reached after incubation of oxidized CF, with trypsin for 60 min (Table 11). A similar rate was obtained for both dithiothreitol-treated CF, and t-free CF, after only 5 min of incubation with trypsin. No further activation was observed in either case after longer incubation with trypsin (not shown). Consistent with earlier reports, the rapid further activation of reduced CF, and t-free CF, by trypsin coincided with a rapid cleavage of the y subunit (Fig. 2, lune C). The major y cleavage product of reduced CF, was the 25-kDa species also produced by tryptic cleavage of reduced CF, on the membrane (Fig. 1). Interestingly, the y subunit of CF,(-t) was rapidly cleaved to a 27-kDa species similar to that obtained from trypsin cleavage of oxidized CF, bound to illuminated thylakoid membranes (Fig. 2, lune E; Fig. 1, lune C).
Trypsin rapidly cleaved the y subunit of reduced CFl(-t) to the 25-kDa species (Fig. 2, lune F ) . Treatment of reduced CF,(-t) with trypsin had no effect on the ATPase activity since the enzyme was already maximally activated (Table 11). Regardless of whether a 27-kDa or 25-kDa species was formed, the 6-8-kDa y fragment appeared at the front.
Trypsin Cleavage of Dithiothreitol-activated CFl-A fluorescent maleimide (anilinonaphthylmaleimide) was previously used to investigate the trypsin digestion pattern of isolated CF, labeled either on the dark-accessible or lightaccessible y sulfhydryls (28). A similar approach was carried out to examine the fate of the sulfhydryls that participate in the y disulfide using 6-acryloyl-2-dimethylaminonaphthalene (Acrylodan), a new fluorescent probe with sulfhydryl, but little amine reactivity (23). The accessible sulfhydryls of the y and t subunits of isolated CF, were blocked by incubation with N-ethylmaleimide. After reduction of the y disulfide Partial Proteolysis of the y Subunit of CF, bond with dithiothreitol, the resulting free sulfhydryls were reacted with Acrylodan and the modified enzyme was treated with trypsin. A time course of the digestion by trypsin is shown in Fig. 3. Within the first minute after addition of trypsin, the fluorescence of y decreased while several new fluorescent bands appeared. The most prominent bands were the 27-kDa and 25-kDa species described above. A diffuse band was also formed with a molecular mass of about 11-11.5 kDa; at longer incubation times, these fragments formed a sharp band of 11 kDa which is probably the 11-kDa species found as a cleavage product of the 25-kDa fragment (14). The 14-kDa species formed 3-5 min after trypsin addition exhibited a low fluorescence, presumably due to a small amount of label incorporated into the "light"-accessible y sulfhydryl. Another diffuse band running below the free dye band appeared after 1-2 min. The amount of fluorescence in this band was difficult to estimate. However, it seemed to increase until y and the 27-kDa species were nearly completely missing. In addition, the amount of fluorescence in the 25-kDa species after digestion for 5 min appeared to be lower than that of the original y band. Essentially the same digestion pattern was obtained with CF1 labeled with fluorescein maleimide on the disulfide cysteines; a 27-kDa species was formed which was further cleaved into the 25-kDa and 2-kDa fluorescent species. Therefore, the two labeled cysteines of the disulfide bond are both likely to be present in the 27-kDa species. An additional clip creates a 25-kDa fragment, as well as a small peptide with a molecular mass of about 2000 Da. Both the 25-kDa and 2-kDa fragments contain a disulfide cysteine. The faint fluorescent band above the free dye band probably originated from a slight unspecific labeling of the a and p subunits.
Isolation of the Small Fluorescent Peptide by Gel Filtration-A fluorescent band near the front was detected when disulfidelabeled, trypsin-treated CF, was subjected to acrylamide gel electrophoresis in the absence of sodium dodecyl sulfate. This result suggests that the 2-kDa fragment dissociates from the remainder of the enzyme. To isolate this species, CFl was treated with trypsin for 4 min and was passed through a Sephadex G-75 column. Two peaks of Acrylodan fluorescence and of AZT7 and A380 were observed (Fig. 4). The first peak eluted with the void volume (fractions [8][9][10][11][12]; its fluorescence emission maximum was at 475 nm which is identical with that of labeled undigested CF1. The second peak eluted in fractions 18-30; the relationship between the absorption at 380 nm (the absorption maximum of Acrylodan) and 277 nm varied throughout the peak indicating that several peptides were present. The absorption at 380 nm matched well with the fluorescence. The emission maximum was at 525 nm in the second peak samples, indicating that the probe was in a more polar environment than in peak I. The molecular mass of peptides eluted in peak I1 (deduced from the separation range of Sephadex G-75) is 5 kDa or less. As judged by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, peak I1 contained only fluorescent and Coomassie blue staining material that migrated near the front (not shown). Samples of the first peak contained the slightly modified a and p subunits, together with t, and the 25-kDa and 6-8 kDa fragments of the y subunit. 6 was completely digested in accordance with its high sensitivity to trypsin (14). Fractions 18 to 30 were combined and applied to a small DEAE-cellulose column (0.5 X 5 cm). Free label and other fluorescent material eluted with the void volume. The small fluorescent peptide was eluted with 0.2-0.3 M NaC1. CF1 labeled on the dark site with Acrylodan, reduced by dithiothreitol, and treated with trypsin in the same way as described above showed no second fluorescent peak upon Sephadex G-75 separation (data not shown).
Identification of Acrylodan-labeled Tryptic Peptides-A detailed analysis of cysteine-containing tryptic peptides of the y subunit was carried out previously by modification of sulfhydryl groups with 4-vinylpyridine followed by total tryptic digestion and separation of peptides by HPLC (7). The two cysteines forming the disulfide bond and the light-and darkaccessible cysteines eluted on four different peptides, S1 to S4 (7). A similar analysis was carried out using Acrylodanlabeled CFI to identify the y peptide released from the enzyme by trypsin. The peptides were detected by their fluorescence since the sensitivity of this method is much higher than detection by UV absorption (7).
When isolated oxidized CFI is treated with Acrylodan, most of the label was incorporated into the y subunit (about 80%) and a small amount into the t subunit. After total tryptic digestion of the labeled protein, two major fluorescent peaks were separated by HPLC (Fig. 5A). The second peak with a retention time of 34 min corresponds to the peptide containing the dark-accessible site (S4) and the first peak (retention time 31 min), to the cysteine-containing peptide of t. However, the tryptic peptide of S3 (containing the light-accessible y sulfhydryl) eluted at nearly the same position. It was found that FIG. 5. HPLC elution profile of the total trypsin digest of Acrylodan-modified CFI. Acrylodan-modified CFI (0.5-1.5 mg/ ml) was treated with trypsin (2 pg/100 pg of CF,) in the presence of CaC12 (3 mM) for 1-3 h a t 37 "C. T h e samples were then heated in boiling water for 10-15 s. After a second addition of trypsin (2 pg/ 100 pg of CF,), the samples were incubated a t 37 "C overnight. The solutions were centrifuged in a Beckman Airfuge for 5-10 min to remove small amounts of precipitate and 20-pl aliquots were applied to HPLC. Peptides were eluted using gradients formed by mixing 0.1% phosphoric acid in HZ0 (solvent A) and phosphoric acid in acetonitrile (solvent B). The elution program was composed of two linear gradients from 20-34% solvent B for 15 min followed by a 34-45% solvent B linear gradient for 40 min at a constant flow rate of 2 ml/min. A, CF,; B, dithiothreitol-activated CF,. this peak varies in height in different preparations and may contain S3 as well as the t peptide.
CF1 labeled with Acrylodan after pretreatment of the enzyme with dithiothreitol showed two additional peaks with retention times of 18 and 25 min (Fig. 5 B ) , corresponding to the two disulfide sulfhydryls of y. Therefore, the four peptides S1 to S4 appear to elute in the same sequence as the peptides labeled with 4-vinylpyridine (7).
Treatment of isolated CF, with N-ethylmaleimide prior to dithiothreitol incubation, and subsequent labeling with Acrylodan, decreased the amount of label incorporated into t and the accessible y sulfhydryl. Consequently, HPLC peaks c/S3 and 5 4 were reduced (Fig. 6 A ) . CF, specifically labeled with Acrylodan on the disulfide sulfhydryls was subjected to a brief exposure to trypsin and the fluorescent products separated by Sephadex G-75 chromatography (cf. Fig. 4). Polypeptides in the two fluorescent peaks were digested with trypsin overnight and examined by HPLC. The first peak (void volume fractions) (Fig. 6 B ) contained only S1, while S2 was nearly exclusively present in the second G-75 peak (Fig.  6C). The overnight trypsin digestion of the S2-containing fractions did not affect the retention time of S2. This result indicates that the S2 fragment is released from the enzyme in its final form rather than as a larger precursor.  Fig. 4 were analyzed by HPLC as in Fig. 5. A , CF, labeled with Acrylodan on the y disulfide cysteines (accessible SH-groups were blocked with N-ethylmaleimide before reduction of the disulfide bond); B , Fraction 10 ( Fig. 4); C, Fraction 23 (Fig. 4).
Modification of the disulfide sulfhydryls with Acrylodan is not a prerequisite for trypsin cleavage and release of S2 because trypsin treatment of the reduced, unmodified CFI gave the same cysteine-containing peptide. In this case, released S2 could be identified by addition of Acrylodan to the late fractions of Sephadex G-75 chromatography and subsequent HPLC separation (data not shown).
Release of S2 from Membrane-bound CFl-Membranebound CF1 was labeled with Acrylodan on the disulfide sulfhydryls and the release of fluorescent peptides into the medium during incubation with trypsin in the light or the dark was followed. After trypsin treatment, the thylakoids were precipitated by centrifugation and the resulting supernatants were passed through small DEAE-cellulose columns. Fluorescent peptides were eluted and identified by HPLC without further trypsin digestion. A single, small peak was present in the supernatants derived from membranes treated with trypsin either in the dark or in the light (Fig. 7, A and B ) . The identity of this peak as S2 was confirmed by co-chromatography with authentic S2 derived from isolated CF, (Fig. 7 C ) . Thylakoids were labeled with Acrylodan after incubation with N-ethylmaleimide and preillumination in the presence of dithiothreitol as described under "Materials and Methods." To 60 ml of illumination medium containing Acrylodan-labeled thylakoids (equivalent to 6 mg of chlorophyll), trypsin was added (20 pg/ml) and duplicate samples were either illuminated for 5 min or kept in the dark. Soybean trypsin inhibitor (50 pglml) was added and thylakoids were removed by centrifugation. The supernatants were applied to DEAE-cellulose columns (0.5 X 5 cm) equilibrated with Tris/EDTA buffer. Fluorescent peptides were eluted with 0.4 M NaCl in Tris/ EDTA buffer and were subjected directly to HPLC without further trypsin treatment. A , supernatant of thylakoids treated with trypsin in the dark; B, supernatant of thylakoids treated with trypsin in the light; C, 15 g1 of supernatant B plus 5 g1 of authentic Acrylodanlabeled S2 peptide isolated from CF,.

DISCUSSION
We have shown that limited proteolysis can be used to probe active conformations of both soluble and membranebound CF1. Two major sites of trypsin cleavage of the y subunit were identified. Although the y subunit of oxidized CF, in solution is resistant to attack by trypsin, removal of the c subunit causes a fast cleavage of y to a 27-kDa species. Reduction of the y disulfide also causes increased cleavage of y, but, in this case, a 25-kDa protein is formed. The y polypeptide bearing the S2 cysteinyl residue is released from the enzyme. Remarkably, similar results were obtained with membrane-bound CF,. Energization of thylakoids by illumination permits the rapid cleavage of the y subunit to a 27-kDa species. Reduction of the disulfide bond in y, followed by trypsin treatment, causes the liberation of the S2 polypeptide from y and the generation of a 25-kDa y species. Illumination enhances trypsin cleavage of y of reduced, membranebound CFI. These results suggest that the structure of the activated forms of membrane-bound CF1 resemble those of activated CF1 in solution. Oxidized CF, in illuminated membranes resembles CFl(-t) with respect to its trypsin sensitivity and reduced membrane-bound CFI is similar to its soluble counterpart.
Illumination of thylakoid membranes thus induces conformational changes that cause the region of y that bears the disulfide bond to become accessible to trypsin. These changes also influence the ease of reduction of the y disulfide. The reduction by dithiothreitol of the disulfide in either soluble CF, or in membrane-bound CF, (29) in the dark requires incubation for 1-2 h with high dithiothreitol concentrations. In contrast, the disulfide is reduced within a few minutes at much lower dithiothreitol concentrations when the membranes are illuminated. Removal of the t subunit also markedly enhances the rate of reduction of the y disulfide.' Previously, the trypsin cleavage products of y in reduced CF1 that bear the dark (S4) and light site (S3) sulfhydryls were identified. The dark site sulfhydryl was present on a fragment of about 6 kDa. S3 was initially found on the 25-kDa fragment which was further degraded to a 14-kDa peptide (28). In this paper, we describe the fate of the two disulfide sulfhydryls of y following trypsin digestion. The results are summarized in the degradation scheme shown in Fig. 8. Both sulfhydryls are present on a precursor of the 25-kDa fragment having an apparent molecular mass of 27 kDa. This fragment also contains the light site y sulfhydryl residue. The dark site is found on a 6-8-kDa polypeptide which appears without any apparent precursor. The scheme of Fig. 8 assumes that no fragments other than the 27-and 6-8-kDa species are formed by the first trypsin clip.
Upon formation of the 25-kDa species from the 27-kDa fragment, one of the labeled cysteines is lost and appears on a 2-kDa peptide. This labeled molecule freely dissociates from the CF, complex and can be isolated by gel filtration. HPLC showed that this peptide probably is the tryptic peptide S2 (7) containing 11 amino acid residues, including one of the y disulfide sulfhydryls.
The 25-kDa species is further cleaved to two fragments of 14 and 11 kDa, containing the light site and the other disulfide cysteinyl residue, respectively. Although the arrangement of these two fragments within the y subunit was not confirmed, we assume that the 14-kDa species, containing the light site cysteine, is the N-terminal portion of y because in the oxidized enzyme the 27-kDa species is sometimes clipped a t trypsin sites I and I11 only giving a slightly larger precursor of the 11-kDa species (see Fig. 3). Cleavage at the second trypsin site is very slow with the oxidized enzyme indicating that the disulfide bridge keeps this region of y in a conformation inaccessible to trypsin. Removal of t from oxidized CF, exposes only one of the trypsin sites ( I in Fig. 8). This might indicate either that t is bound to this region of y or that t binds to an "allosteric" site and keeps the coupling factor in a conformation in which the trypsin sites are inaccessible.
The latent ATPase of CF, is activated by reduction of the y disulfide bond or further activated by trypsin cleavage at a site or sites which appear to be close to the disulfide bond. All of these effects appear, therefore, to be related in that they involve changes in a limited region of the y polypeptide. The y disulfide bond may be involved in keeping CF, in a constrained, inactive conformation. Reduction, or trypsin cleavage of y, could result in a higher flexibility of the modified complex in which rapid ATP hydrolysis is possible. Removal of t from CF, may invoke similar changes in the y subunit. Although the effect of t removal is only partial, reduction of the disulfide bond in t-free CFl results in complete activation. Together with the effect of t removal on accessibility of trypsin or dithiothreitol to y, such results have been taken' to indicate a close physical and functional relationship between c and y. A critical region of the y subunit Cleavage of y at the first trypsin site (I) generates a 27-kDa and a 6-8-kDa fragment (with the dark site cysteine). When the disulfide bond in y (S-S) is reduced, a second site (11) is exposed and trypsin cleavage leads to a 25-kDa fragment and the 1.3-kDa peptide containing a disulfide cysteine (S2 cysteine). The 25-kDa fragment is slowly cleaved at a third site (ZIZ) giving a 14-kDa and an 11-kDa fragment. may, therefore, play a central role in the process of activation of CFI both on and off the thylakoid membrane.