The 55-kDa polypeptide released from spinach thylakoid membranes with 1 M LiCl is not the beta subunit of chloroplast F1.

It was reported by Frasch et al. (Frasch, W. D., Green, J., Caguiat, J., and Mejia, A. (1989) J. Biol. Chem. 264, 5064-5069) that washing spinach thylakoid membranes with 1 M LiCl caused the release of the beta subunit of chloroplast F1 (CF1) which, existing as 180-kDa complexes of beta 3, retained considerable ATPase activity. We repeated their procedures and confirmed that a CF1 beta-like 55-kDa polypeptide was a major constituent of the 1 M LiCl-washed extract. However, the extract contained another polypeptide of which the Mr was 14,000, and these two polypeptides comprised a complex with approximate Mr 550,000 that had the same mobility in native polyacrylamide gel electrophoresis as that of ribulose-1,5-bisphosphate carboxylase. Only very low ATPase activity, less than 1% of the reported value, was detected for the extract and the purified complex. Antibody against the beta subunit of F1 from a thermophilic bacterium PS3 showed a clear cross-reactivity with the CF1 beta subunit but not with the 55-kDa polypeptide. Analysis of the N-terminal amino acid sequences of the 55- and 14-kDa polypeptides and the whole complex revealed that the complex was ribulose-1,5-bisphosphate carboxylase and that the 55- and 14-kDa polypeptides were its large and small subunits, respectively.

that washing spinach thylakoid membranes with 1 M LiCl caused the release of the @ subunit of chloroplast F1 (CF,) which, existing as 180-kDa complexes of Ba, retained considerable ATPase activity.
We repeated their procedures and confirmed that a CFI &like 55-kDa polypeptide was a major constituent of the 1 M LiCl-washed extract.
However, the extract contained another polypeptide of which the M, was 14,000, and these two polypeptides comprised a complex with approximate M, 550,000 that had the same mobility in native polyacrylamide gel electrophoresis as that of ribulose-1,5-bisphosphate carboxylase.
Only very low ATPase activity, less than 1% of the reported value, was detected for the extract and the purified complex. Antibody against the B subunit of F1 from a thermophilic bacterium PS3 showed a clear cross-reactivity with the CFl@ subunit but not with the 55-kDa polypeptide.
Analysis of the N-terminal amino acid sequences of the 55-and 14-kDa polypeptides and the whole complex revealed that the complex was ribulose-1,5-bisphosphate carboxylase and that the 55-and 14-kDa polypeptides were its large and small subunits, respectively.
H+-ATP synthase couples a transmembrane proton flow to ATP synthesis, and F1' . 1s the water-soluble part of this enzyme (1). F1 has ATPase activity, and its M, is 380,000-400,000. The subunit structure of F1 is a&y&, and the catalytic sites reside most likely on the fl subunits. However, the isolated p subunit by itself shows only 0.1% of ATPase activity of F, (2-4). The ATPase-active complex with the simplest subunit composition so far found is the (Y& complex in which the activity is about 20% of F1 (3,4 Frasch et al. (5) reported recently the isolation of a further simple & complex.
It is known that the F1 fi subunit is removed from chromatophores of Rhodospirillum rubrum by 2 M LiCl (6). Similarly, washing chloroplast thylakoids by 2 M LiCl or LiBr results in the release of some or all subunits of chloroplast F1 (CFI) from membranes (7,8). A reconstitutively active CF, ,f3 subunit which restored ATPase activity to R. rubrum chromatophores lacking the /3 subunit has been purified from isolated spinach CFI by freeze-thaw treatment (9). Frasch et al. (5) washed spinach chloroplast thylakoid membranes with 1 M LiCl and found that the major constituent of the washed extract was a polypeptide that has the same mobility as that of the ,8 subunit of CF1 in polyacrylamide gel electrophoresis in the presence of SDS. They proved by Western immunoblotting that this protein (55-kDa protein) cross-reacted with anti-CF1 p subunit antibody. From the results of gel permeation chromatography using a Bio-Gel P-150 column, they suggested that the 55-kDa protein existed as a NO-kDa complex, a trimer of the p subunit. Moreover, this complex had rather strong ATPase activity, more than 10% of that of CF1, in the presence of 40 mM octyl glucoside. Since the finding of the ATPase-active /33 complex appears to be an important contribution for studies of F1, we attempted to obtain the & complex according to the procedures of Frasch et al. (5 On the fi3 Complex of CF, a gas-phase peptide sequenator (Applied Biosystem 470A). The purified complex was dialyzed against pure water and analyzed directly. For the analysis of the 55-and 14-kDa polypeptides, the purified complex was subjected to SDS-polyacrylamide gel electrophoresis, and protein bands were blotted to a polyvinylidene difluoride membrane. The blotted membrane was stained by Coomassie dye, and the 55-and 14-kDa bands were cut out. Pieces of membrane thus obtained were analyzed with a sequenator (13).

ATPase
Assays and Other Methods-Since the ATPase activity of CF, is usually latent and octyl glucoside is known to activate it (14), reaction mixtures for ATPase assays contained 50 mM Tricine-NaOH (pH 8.0), 5 mM ATP, 5 mM MgCl*, and 40 mM octyl glucoside. The reaction mixture was preincubated for 3 min at 37 "C, and the reaction was initiated by addition of the enzyme. After 30 min of incubation at 37 "C, the reaction was terminated by the addition of 0.25 ml of 2% perchloric acid, and the amount of Pi produced was measured as described previously (15). One unit of activity is defined as the activity liberating-l pmol of P,/min. Protein concentrations were assayed by the Bradford nrocedure (16). Sninach CF, was nurified as described in Ref. 17. Tdis preparation of'CF1 was &t&d to contain a small amount of ribulose-1,5-bisphosphate carboxylase (ribulose-Pz carboxylase) by native polyacrylamide gel electrophoresis and by reaction with anti-ribulose-Pa carboxylase antibody. The TF, /3 subunit used as an antigen was purified using the overexpression system in Escherichia coli (18).

Sephacryl S-400 Chromatography of the 1 M LiCkwashed
Extract-In order to obtain the & complex we carried out the procedures reported by Frasch et al. (5). We confirmed that washing thylakoid membranes with 1 M LiCl caused the specific release of a protein complex which contained a polypeptide with M, 55,000 (Fig. la, left two lanes). As shown in Fig. 2, when the washed extract was applied to Sephacryl S-400 chromatography, most of the protein in the washed extract was eluted as a single peak. The constituents of the fractions around the protein peak were analyzed by SDSpolyacrylamide gel electrophoresis (Fig. la). It was evident that the protein complex eluted at around fraction 39 contained a polypeptide of which the M, was estimated to be Experimental details are described under "Experimental Procedures." Inset, the molecular weight of the purified complex was analyzed by gel permeation HPLC on a G~OOOSWXL column (Tosoh Co., Japan).
Approximatly 20 pg of the purified complex were injected into the column. Rabbit muscle lactate dehydrogenase (LDH) (M,, 140,000), rabbit muscle pyruvate kinase (PK) (M,, 240,000), and spinach CFI (M,, 400,000) were used as M, standards. The column was eluted with 20 mM Tricine-KOH (pH 7.4) buffer containing 10% glycerol and 50 mM NaCl at a flow rate of 0.5 ml/min at room temperature. The elution was monitored by absorbance at 280 nm. Each sample was analyzed separately. 55,000. The electrophoretic mobility of this 55-kDa polypeptide in SDS-polyacrylamide gel was the same as or very similar to that of the CFI /3 subunit. As shown already by Frasch et al. in Fig. 2 of their paper (5), another two polypeptides, one below the band of the CF1 c subunit and the other at almost the same position of the CFI y subunit, were also seen in the gels (Fig. la). These results are essentially the same as those reported by Frasch et al. (5).

Native Polyacrylamide
Gel Electrophoresis of the Washed Extract and Fractions of Chromatography-It is evident from Fig. lb that a protein complex moving slower than CF, in the native polyacrylamide gel electrophoresis is almost the only constituent of the washed extract (left two lanes). Consequently, it is also the case for the eluted fractions from Sephacryl S-400 chromatography (right lanes). It is noteworthy that the electrophoretic mobility of the protein complex contained in the washed extract and eluted fractions is the same as that of ribulose-Ps carboxylase which is contained in our preparation of CF1 (Fig. lb, left two lanes). Since the report by Frasch et al. (5) did not include results of native polyacrylamide gel electrophoresis of their complex, we cannot compare our results with theirs.

Gel Permeation HPLC of the Purified
Complex-The molecular weight of the protein complex was estimated by gel permeation HPLC using a G4000SWx~ column. As shown in Fig. 2 (inset), the elution of the complex was earlier than that of CFI, and the M, of the complex was estimated to be 550,000. This value is much larger than the reported value, 180,000 (5).
Subunits of the Protein Complex-As described above, there is a protein band running faster than the CF, c subunit in the SDS-polyacrylamide gel electrophoresis. The molecular weight of this small polypeptide was estimated to be 14,000 from another SDS-polyacrylamide gel electrophoresis with M, standards. Judging from the relative staining intensity of this 14-kDa band to the 55-kDa band and from the fact that the 14-kDa polypeptide was always copurified with the 55-kDa polypeptide, we concluded that the 14-kDa polypeptide, as well as the 55-kDa polypeptide, was also a subunit of the protein complex. The faint band running at almost the same mobility as the CR y subunit was eluted only at the latter part of the protein peak of Sephacryl S-400 chromatography (Fig. la)  The complex (70 rg) purified with Sephacryl S-400 chromatography using a buffer without 20% glycerol was directly analyzed with a gas phase sequenator. Three PTH-derivatives were always found at each cycle, except cycles 1 and 2 where two PTH-derivatives were detected. PTH-derivatives found in the N-terminal region of the small subunit (Cl) and large subunit (0, 0) of spinach ribulose-Pz carboxylase are connected by lines. The left two lanes (1, a and b) are stained with Coomassie dye, and the right two lanes (2, a and b) show the immunoreaction with anti-TFi @ antibody. Approximately 10 pg of proteins were separated on 12.5% SDS-polyacrylamide gels, and protein bands were blotted to polyvinylidene difluoride membranes. The blotted membranes were incubated with the blocking medium and then with the solution of the anti-TF1 p antibody. After washing the membranes, the second antibody, goat anti-rabbit IgG antibody conjugated to alkaline phosphatase, was allowed to bind the primary antibodies on the membranes. Localization of bound antibodies were visualized by the addition of nitroblue tetrazolium and 5-bromo-4-chloro-3indolyl phosphate.
purified complex using an antibody against the TF, /3 subunit is shown in Fig. 3. The p subunit of the authentic CFi showed a clear cross-reaction with anti-TF, p antibody (lane 2~). Stained bands below the CFi /3 subunit correspond to a small amount of the degraded CF, 0 subunit contained in our CFi preparation. Contrary to this, the 55-kDa polypeptide of the complex did not react with anti-TF1 /3 antibody (lane 2b).
ATPase Activity of the Complex-ATPase activities of the washed extract and the purified complex were examined under the conditions described under "Experimental Procedures." However, the maximum activities we were able to detect in the presence of octyl glucoside were only 0.0030 and 0.0075 units/mg protein for the washed extract and the purified complex, respectively. When octyl glucoside was omitted from the reaction mixtures, these activities were further reduced to 0.0020 and 0.0010 units/mg, respectively. These values are less than 1% of those reported in Ref. 5. Analysis of N Terminus of the 55-kDa Polypeptides-The results described above indicated that the complex we had purified was ribulose-Pn carboxylase rather than the CR p3 complex. In order to obtain exclusive evidence, we analyzed the N-terminal amino acid sequences of the 55-and 14-kDa polypeptides and the complex. It is known that the N terminus of the large subunit of spinach ribulose-Pn carboxylase is blocked by acetylated proline (19). Consistently, when the 55-kDa polypeptide was analyzed by five cycles of Edman degradation, no meaningful PTH-derivative was liberated at each cycle. Neither was the sequence of the N-terminal region of the spinach CFI /3 subunit (20) detected.

N-terminal
Amino Acid Sequence of the 14-kDa Polypeptide--Analysis of 20 amino acids from the N terminus of the 14-kDa polypeptide gave a sequence, SKVWPTQNMKHYE-TLSYLPP. As shown below, this sequence is homologous, if not identical, to the published N-terminal amino acid sequence of the small subunit of spinach ribulose-Pz carboxylase, of which amino acids at the first and second cycle of Edman-dansyl reactions were not determined unequivocally (21).
. . ylase small subunit: Identical amino acids between the two sequences designated by asterisks are also members of the conserved amino acids for almost all ribulose-Pz carboxylase small subunits from various sources (22). Therefore, we concluded that the 14-kDa polypeptide was the ribulose-Pp carboxylase small subunit. Disagreement at positions 6,7,8,9,11, and 12 can be explained by the possible presence of multiple genes of ribulose-P* carboxylase small subunits in spinach chromosomes as established in other plants (23,24) or by the difference of strains of market spinach.

N-terminal
Amino Acid Sequences of the Complex-When the N-terminal amino acid sequence of the intact complex was directly analyzed, only the sequence of the ll-kDa polypeptide appeared. When 10 amino acids from the N terminus of the complex prepared by a Sephacryl S-400 column using a buffer without glycerol were analyzed, three PTH-derivatives appeared in most of each cycle of degradation, indicating three kinds of polypeptides were contained in the complex (Fig. 4). One of them is derived from the 14-kDa polypeptide. The sequences of the remaining two polypeptides were found in the N-terminal region of the published sequence of the large subunit of spinach ribulose-P2 carboxylase, one starting from Ser-8 and the other from 25)  Although the N-terminus of the large subunit of spinach ribulose-Pp carboxylase is blocked by acetylated proline (19), the N-terminal region of the ribulose-P2 carboxylase large subunit is highly susceptive to proteolysis (26). Apparently these two polypeptides were generated by cleavage of the peptide bonds of the intact 55-kDa polypeptides at two different sites during chromatography without glycerol. The sum of the amount of PTH-derivatives produced at each cycle from these two polypeptides is roughly equal to the amount of PTH-derivative derived from the ll-kDa polypeptide. From these results we concluded that the complex we purified is spinach ribulose-Pz carboxylase and that the 55-and 14-kDa polypeptides are its large and small subunit, respectively. DISCUSSION We tried to isolate the /33 complex from spinach thylakoid membranes by the procedures reported by Frasch et al. (5) and confirmed that the protein complex containing a CFI /3like polypeptide was a main constituent of the 1 M LiCl- any, and shows the same mobility as that of ribulose-P2 carboxylase in native polyacrylamide gel electrophoresis. Antibody against the TF1 /3 subunit reacted clearly with the p subunit of authentic CF1 but not with any polypeptides contained in our complex. Analysis of the N-terminal amino acid sequences of the complex and the 55-and 14-kDa polypeptides proved that the complex is ribulose-Pz carboxylase and that the 55-and 14-kDa polypeptides are the large and small subunits of ribulose-P, carboxylase, respectively. Ribulose-P2 carboxylase is composed of eight large and eight small subunits. The molecular weights of spinach ribulose-P2 carboxylase and its large and small subunits are 532,000,52,600, and 13,900, respectively (19,25), showing good agreement with the values of the complex we obtained.
It is known that spinach ribulose-Pz carboxylase is a major contaminant of the CFi preparation (27), and the ribulose-P2 carboxylase large subunit and CFI /3 subunit are practically indistinguishable from each other in SDS-polyacrylamide gel electrophoresis. Thus, it is obvious that we obtained ribulose-P2 carboxylase instead of trimers of the CF1 /3 subunit, even though we followed the procedures described in Ref. 5. We could not confirm the presence of the CF1 /3 subunit in our preparation by Western immunoblotting or by N-terminal analysis. These results are reproducible in our laboratory since we repeated the experiments five times and obtained essentially the same results each time. The reason for the discrepancy between Frasch's results and ours is not known. However the following points should be considered.
Patterns of SDS-polyacrylamide gel electrophoresis of the 1 M LiCl-washed extract and fractions of Sephacryl S-400 column chromatography ( Fig. la) were very similar to those presented in Fig. 2, A  Therefore, the band found in their gels just below the position of the CFi t subunit, which the major 55-kDa polypeptide always accompanied, is probably the same polypeptide as the ll-kDa polypeptide of our preparation, that is the small subunit of ribulose-P2 carboxylase.
The M, of our complex was estimated to be about 550,000 using a HPLC gel permeation column, G~OOOSWXL, by which large proteins and protein complexes up to M, 700,000 can be analyzed (28). Frasch et al. (5) estimated the M, of their complex to be 180,000 from Bio-Gel P-150 chromatography.
However, in general, proteins having a M, of more than 150,000 are eluted out at void volume from a Bio-Gel P-150 column and hence cannot be analyzed (29). Therefore, their value could be an underestimation, and the real M, of the complex may be larger than 180,000.