Eur. J . Biochem. 92, 9-14 (1978)
Characterization of the Dicyclohexylcarbodiimide-Binding Protein Isolated from Chloroplast Membranes Kristine SIGRIST-NELSON, Hans SIGRIST, and Angelo AZZI Medizinisch-chemisches Institut und Biochemisches Institut der Universitit Bern (Received August 17, 1978)
Characterization of a butanol-solubilized protein isolated from chloroplast membranes is reported. The proteolipid, which specifically and covalently binds dicyclohexylcarbodiimide, has an apparent molecular weight of 8000 in dodecylsulfate electrophoresis. The minimum molecular weight calculated from amino acid analysis data is 7700. N-Formyl-methionine was determined to be the N-terminal amino acid. Glycine, alanine and leucine were present in elevated amounts, resulting in a polarity of 29 %. Cysteine and histidine were lacking. In high-voltage electrophoresis the peptide appeared as a single homogenous spot which migrated, at pH 6.5, with the relative mobility of glycine. At concentrations where dicyclohexylcarbodiimide inhibited ATPase activity maximally (20 nmol per mg membrane protein), 0.17 nmol dicyclohexylcarbodiimide was covalently bound per nmol isolated proteolipid, indicating that one out of six molecules of proteolipid was labeled.
Dicyclohexylcarbodiimide has long been known to be an inhibitor of oxidative phosphorylation in mitochondria and bacteria as well as photosynthetic phosphorylation in chloroplasts. Various experiments have suggested that the inhibitor's mode of action is to block sites participating in proton translocation [l- 31. Dicyclohexylcarbodiimide has been found to bind selectively to a small polypeptide located in the H +-translocating ATPase complexes present in mitochondrial, bacterial and chloroplast membranes [4,5]. Recently the dicyclohexylcarbodiimide-binding protein was purified from chloroplast membranes and reconstituted. Evidence was presented that the protein was able to function as a dicyclohexylcarbodiimidesensitive proton channel [6]. Heightened interest has been aroused in this protein due to the implications that it may possess biologically important ionophoric activity. An understanding of the mode of functioning of such a protein is only possible when its structural parameters are well defined. To this purpose the purification and chemical characterization of the carbodiimide-reactive component of chloroplasts is reported. MATERIALS AND METHODS Preparation and labeling of lettuce chloroplast membranes with dicyclohexyl['4C]carbodiimide was Abbreviution. Dansyl, S-dimethylaminonaphthalene-l-sulphonyl.
carried out as described elsewhere [6]. Unless otherwise stated all organic solvent systems are given in volume ratios. Pro ttolipid Isolation
The isolation procedure was performed essentially as previously reported [7], except that the butanol solubilizate was additionally filtered through a Satorius filter (11305, 0.6 pM) to ensure complete removal of precipitated protein. The final ether-precipitated proteolipid fraction was washed in three subsequent washes with water, butan-1-01 and cold diethyl ether. Published procedures were followed for the determination of chlorophyll [8] and protein [9]. Chloroplast membrane protein was determined after precipitation of the membrane protein with acetone. The anthrone procedure was used for determination of sugar residues in estimating glycolipid content [lo]. Polyacrylamide Gel Electrophoresis
Two methods of electrophoretic analysis were performed. Sodium dodecyl sulfate/polyacrylamide gel electrophoresis was carried out in accordance with Weber and Osborn [ l l ] . The electrophoresis was performed in 0.1 % sodium dodecyl sulfate, 0.05 M sodium phosphate buffer pH 7.2 using 10% polyacrylamide gels. Gels were stained with Coomassie brilliant blue and destained with 7 acetic acid. Addi-
10
tionally, the sodium dodecyl sulfate/urea system of Swank and Munkres [12] was utilized. The protein was dissolved in 1 % sodium dodecyl sulfate, 8 M urea, 1% mercaptoethanol and 0.01 M sodium phosphate buffer pH 7.2. The samples were run in 10% polyacrylamide gels containing 0.1 % sodium dodecyl sulfate, 8 M urea, 0.1 M sodium phosphate buffer pH 7.2. For molecular weight determination bovine serum albumin ( M , 68000), ovalbumin ( M , 43000), egg white lysozyme ( M , 14300) and bovine pancreatic insulin ( M , 5700) served as standards. When dicyclohe~yl['~C]carbodiimide-labeledsamples were electrophoresed, identical samples were applied to separate gels. One gel was stained, the duplicate gel was immediately cut into 1-mm slices. Individual gel slices were extracted overnight with 10 mM Triton-X-100. Scintillation fluid was added and the slices were counted for radioactive content. Thin-Layer Chroma tography
Silica gel plates without fluorescent indicator were used (Merck no. 5721). A two-dimensional chromatography was performed. The first solvent system was chloroform/methanol/water (65/24/4), the second was chloroform/methanol/isopropylamine/ammonia (65/ 35/0.5/5). Plates were exposed to iodine and then either sprayed with ninhydrin to localize amino-positive lipids and protein or with naphthalindiol for sugar identification. The naphthalindiol spray was composed of 100 mg 1,3-naphthalindiol dissolved in 49 ml of ethanol with 1 ml concentrated sulfuric acid added. When determination of the dicyclohexyl['4C]carbodiimide radioactivity was desired the thin-layer plates were exposed to iodine and radioactive content of the lipid-positive spots determined.
Proteolipid Characterization
organic phase of butanol/acetic acidlwater (41115). The dansylated proteolipid, present at the origin as a green fluorescent spot, was eluted from the silica gel by phenol/acetic acidlwater (2/1/1, w/v/v). The eluted proteolipid was dried under high vacuum and subjected to acid hydrolysis. The dansylated amino acids were analyzed by microscale thin-layer chromatography using polyamide plates with the following solvent systems : 1.5 % formic acid (first dimension), benzene/acetic acid (9/l, second dimension), ethylacetate/methanol/acetic acid (20/1/1, third solvent) or 1 M ammonium hydroxide/ethanol ( l / l , third solvent). Dansylated amino acids served as reference. Removal of the N-formyl group from the terminal amino acid was carried out by mild acid hydrolysis [14]. The washed, ether-precipitated proteolipid was mixed with 10 volumes of methanol16 M HCl (6011) and incubated for 2 h at 37 "C. The proteolipid was recovered by ether precipitation and dansylated by the standard procedure. Amino Acid Analysis
The isolated proteolipid was hydrolyzed with 6 M HCl in an evacuated, sealed tube for 24 h at 100 "C. Cysteine was determined as cysteic acid, methionine as methionine sulfone after performic acid oxidation ~51. ATPase Assay
Light-induced Mg2 -ATPase activity of chloroplasts was assayed according to McCarty and Racker [2]. Preincubation of the chloroplasts with dicyclohexylcarbodiimide was performed for 30 min at 25 "C prior to the ATPase assay.
High- Voltage Electrophoresis
Materials
The proteolipid fraction was dissolved in a small volume of chloroform/methanol (2/1) and applied to Whatman no. 1 paper. The electrode buffer consisted of pyridine/water/acetic acid (100/900/4) pH 6.5. The electrophoresis was carried out for 50 min at 2000 V (50 Vjcm). The strip was then cut into 1-cm sections and the radioactive content determined. Protein was localized with ninhydrin. The amino acids lysine, glycine and glutamic acid served as mobility markers.
N,N '-Dicyclohexyl[14C]carbodiimide (50 Ci/mol) was the generous gift of Dr H. R. Kaback (Roche Institute of Molecular Biology, Nutley). All other chemicals and reagents were of the highest quality commercially available.
N- Terminal Determination N-terminal amino acid determination was carried out by dansylation and thin-layer chromatographic identification of dansyl derivatives [13]. In a first step, thin-layer chromatography on silica gel was used for separation of reaction by-products from the dansylated proteolipid. The solvent system employed was the
RESULTS Isolation
The technique previously utilized for isolating proteolipid from sarcoplasmic reticulum membranes and the inner mitochondria1 membrane [7] has proven successful for isolating the dicyclohexylcarbodiimidebinding proteolipid from chloroplast membranes. The procedure involves injection of the membranes into butan-1 -01, followed by separation of the precipitated protein from the butanol. The butan-1-01 solubilizate
11
K. Sigrist-Nelson, H. Sigrist, and A. Azzi Table 1, Isolution of chloroplast proteolipid Sugar was determined by the anthrone method as described in Methods. D-GalaCtOSe served as standard
A Fraction
Chloroplast membrane Butanol: supernatant precipitate Ether: supernatant precipitate
Protein recovered mg
%total
10
100
-
-
9.9 -
99 -
0.1
1
B Ether precipitate
Ratio of protein to glycolipid ~
20 30 40 Slice number
llgilrg sugar Unwashed Washed with ether Washed with butan-1-01
0.94 1.03 8.04
is then combined with diethylether, resulting in precipitation of the proteolipid fraction. Table 1 A shows the distribution of protein throughout the isolation procedure. Of the total membrane protein, 1 % is present in the ether precipitate. It is to be noted that complete removal of the butanol-precipitated protein is required to prevent contamination of this fraction. This is best accomplished by filtration of the butanol prior to ether addition. In addition to protein, lipid is also present in the ether precipitate. The amount of glycolipid may be, however, reduced by washing the etherprecipitated fraction with butan-1-01. While butan1-01 does not serve to solubilize ether-precipitated protein, it does reduce the glycolipid content of this fraction (Table 1B). Washing of the precipitated fraction with ether does not significantly affect the protein to glycolipid ratio. Additionally, the ether precipitate was routinely washed with water to reduce the amount of salts which are soluble to a limited extent in butanol and precipitate upon addition of ether. Solvents other than butan-1-01 which solubilized the proteolipid include methanol, ethanol, propanol, acetone and chloroform/methanol (2/1). A mixture of chloroform/methanol/diethylether (2/1/12) served to extract the major portion of chlorophyll and lipid while leaving the proteolipid unsolubilized. Chloroplast membranes were incubated with dicyclohexyl[14C]carbodiimide.Electrophoretic characterization of the labeled membranes and proteolipid fraction obtained during the isolation process is shown in Fig. 1. The radioactivity was covalently associated with one Coomassie-blue-staining band in the membrane fraction (Fig. 1A). The dicyclohexylcarbodiimide-binding protein was present only in
50
60
B
0
10
20
30 40 Slke number
50
60
Fig. 1, Electrophoretic analysis o j fractions obiuined during proteolipid isolution. Dicyclohexyl['4C]carbodiimide-labeled chloroplast membranes (10 mg protein/ml) were injected into butan-1-01, 0.2 m1/40 ml butan-1-01. The butanol precipitate was removed and the supernatant was combined with 5 volumes o f cold diethyl ether. The resulting ether precipitate was collected. (A) Di~yclohexyI['~C]carbodiimide-labeled chloroplast membranes (70 pg); (B) proteolipid fraction (ether precipitate, 10 pg)
greatly reduced amounts in the butanol-precipitated protein (not shown). In contrast, the dicyclohexylcarbodiimide-binding proteolipid was present as a single Coomassie-blue-staining band, with associated radioactivity in the ether precipitate (Fig. 1 B). The apparent molecular weight of the proteolipid in the sodium dodecyl sulfate system was 8000 [6]. Electrophoretic analysis of the proteolipid fraction was performed with gels containing sodium dodecyl sulfate and 8 M urea (Fig. 1B). The protein pattern with associated radioactivity is identical to that of [6].
Thin-Layer Chromatography Two-dimensional thin-layer chromatography of the ether-precipitated dicyclohexyl[14C]carbodiimidelabeled proteolipid fraction was performed (Fig.2).
12
Proteolipid Characterization
1029 @Start
Fig. 2. Two-dimensional thin-layer chromatogruphy of the proteolipid fraction. The washed (see Methods) proteolipid fraction (3 pg protein) was dissolved in chloroform/methanol (2/1) and applied to the thin-layer plate. Solvent systems and developmental procedures are described in the Methods section. Vertical bars indicate ninhydrin-positive spots; horizontal bars, naphthalindiol-positive spots. Radioactivity (presented as counts/min) was only associated with one spot
The chromatogram revealed the presence of lipids in addition to the proteolipid which was present as a single ninhydrin-positive spot remaining at the origin. Radioactivity was associated only with the ninhydrinpositive spot. Of the four lipid spots present, three reacted positively with naphthalindiol indicating the presence of carbohydrate moieties.
High- Voltage Electrophoresis When the dicycl~hexyl[’~C]carbodiimide-labeled chloroplast proteolipid was subjected to high-voltage electrophoresis only a single ninhydrin spot was detectable (Fig. 3). All of the radioactivity co-migrated with the protein. The relative mobility of the proteolipid was identical to glycine at pH 6.5 and was not altered by the presence of bound dicyclohexylcarbodiimide. N- Terminal Determinat ion Additional evidence for chemical homogeneity of the dicyclohexylcarbodiimide-bindingproteolipid was obtained by end-group analysis. The proteolipid was dansylated as described. Before acid hydrolysis was performed, preparative thin-layer chromatography was used to separate reaction by-products from the dansylated proteolipid. Three fluorescent bands were in evidence on the silica gel plates, a green band at the origin, a blue band with an RF of 0.4 and a pinkyellow band with an RFof 0.7. The proteolipid present at the origin could only be successfully eluted by either acidified chloroform/methanol (2/1) or phenol/ acetic acid/HzO (2/1/1, w/v/v). Other solvents which were tried but were unsuccessful include butan-1-01,
chloroform/methanol (2/1), 1.5 ”/, and 10 % formic acid, butan-1-ol/formic acid (50/1) and H2O/NH3 (5/l>. Following acid hydrolysis of the eluted proteolipid, the dansylated amino acids were analyzed. The only dansylated amino acids detectable in the hydrolyzate were &-N-dansyl-lysineand 0-dansyl-tyrosine, indicating that the N-terminal amino acid is either sterically hidden or chemically blocked. Treatment of the proteolipid with mild acid hydrolysis, prior to dansylation, resulted in the appearance of N-dansylmethionine. The N-terminal amino acid was thus judged to be N-formyl-methionine.
Amino Acid Composition and Molecular Weight The amino acid analysis of the chloroplast proteolipid is reported in Table 2. To be noted is the high preponderance of the nonpolar amino acids glycine, alanine and leucine. A polarity of only 29% was calculated using the formula of Capaldi and Vanderkooi [16]. After subjecting the results from the amino acid analysis to the search procedures of Hoy et al. [17] a coplot ofj; an indicator of goodness of fit to integral numbers against the molecular weight was carried out. A pronounced minimum was obtained at 7700.
Inhibition of ATPase Activity and Binding of’ Dicyclohexyl[ ‘‘C]carbodiimide to the Chloroplast Proteolipid Whole chloroplasts were incubated for 30 min with different amounts of dicyclohexyl[’4C]carbodiimide relative to the chloroplast protein content. Maximal inhibition of activity was obtained at a ratio of 20 nmol dicyclohexylcarbodiimide per mg chloroplast membrane (Fig. 4). The quantity of dicyclohexylcarbodiimide bound by the proteolipid varied in proportion to the amount of dicyclohexylcarbodiimide added to the membranes. The proteolipid bound a maximum of 1 nmol dicyclohexylcarbodiimide per nmol proteolipid when the reagent was present at concentrations of 150 nmol per mg chloroplast membrane protein or greater. Interestingly, however, maximal inhibition of ATPase activity was achieved when only 0.17 nmol dicyclohexylcarbodiimide was bound per nmol proteolipid. Approximately 60 inhibition of ATPase activity was observed when 0.084 nmol of dicyclohexylcarbodiimide was bound per nmol proteolipid.
DISCUSSION A polypeptide soluble in butan-1-01 and originating from chloroplast membranes has been charac-
13
K. Sigrist-Nelson, H. Sigrist, and A. AzLi Glutamic
acid
I Glycine
Lysine
1
0
0
Fig. 3. High-voltage electrophoresis of the dieyclohexylcarbodiimide-binding proteolipid. 10 pg of dicycl~hexyl['~C]carbodiimide-labeled proteolipid were dissolved in chloroform/methanol (2/1) and applied to Whatman no. 1 paper. Electrophoresis conditions are presented in Methods. Lysine, glycine and glutamic acid served as electrophoretic mobility markers. The proteolipid and amino acids were localized with ninhydrin staining. Distribution of radioactivity on the electropherogram is displayed
Table 2. Amino acid composition of' the chloroplast dicyclohexylcarbodiimide-reactiveproteolipid The results are the average of ten separate hydrolysatcs. Cysteine was analyzed as cysteic acid and methionine as methionine sulfone, following performic acid oxidation. Tryptophan was not determined. Results are means S.D.
*
Amino acid
Amount present mol/mol
Aspartic acid Threonine Serine Glutamic acid Proline Glycine Alanine Cysteine Valine Methionine Isoleucine Leucine Tyrosine Phenylalanine Lysine Histidine Arginine Tryptophan
3.81 k 0.30 3.25 f 0.35 4.92 0.39 5.72 0.29 5.16 k 0.72 9.72 0.82 12.04 k 1.02 0.20 k 0.25 5.98 k 0.42 1.11 f 0.21 4.43 f 0.43 9.77 0.81 1.26 k 0.15 3.90 k 0.35 1.49 k 0.19 0.22 f 0.28 1.97 k 0.10 n.d.
*
*
x Polarity
29
terized. The proteolipid which specifically and covalently binds dicyclohexylcarbodiimide has an apparent molecular weight of 8000 in dodecylsulfate electrophoresis. Thin-layer chromatography of the proteolipid fraction revealed the presence of four lipids, three containing carbohydrate moieties. Both high-voltage electrophoresis and N-terminal amino acid analysis indicated the presence of only one polypeptide in the isolated fraction. Similar to the
0
25 50 75 100 125 150 175 200 225 Dicyclohexylcarbodiimide added ( n m o l i mg membrane)
0
Fig.4. Inhihition of ATPuseactivityandbindin~qfdicyclohexyl['4Cjcurbodiimide to the chloroplast proteolipid at varied dicyclohexylcarbodiimide :chloroplast membraneprotein ratios. Chloroplast membranes were incubated with the varied amounts of dicyclohexylcarbodiimide indicated and ATPase activity was assayed (a).The proteolipid was isolated and the amount of bound inhibitor was determined (-1
dicyclohexylcarbodiimide-binding proteolipid isolated from Saccharomyces cerevisiae [18] and Aspergillus nidulans [ 191, N-formyl-methionine was present as the N-terminal amino acid. The chloroplast dicyclohexylcarbodiimide-binding protein displayed a high content of nonpolar amino acids. The absence of cysteine and histidine in the chloroplast proteolipid composition accentuated its correspondence with proteins isolated from other systems (Escherichia coli, [5] ; Saccharomyces cerevisiae and Neurospora crassa [20]). The calculated polarity for the chloroplast proteolipid was 29 %, indicating a rather marked hydrophobic nature. Other published values for dicyclohexylcarbodiimide-bindingproteins include Aspergillus nidulans, 38 % [19], Escherichia coli, 16 % [ 5 ] and Neurosporu crassa, 29 % [20]. Similar to observations with mitochondria, complete inhibition of ATPase activity was observed
14
K. Sigrist-Nelson, H. Sigrist, and A. Azzi: Proteolipid Characterization
when chloroplast membranes were incubated for short periods of time with dicyclohexylcarbodiimide at a ratio of 20 nmol/mg membrane protein [21]. At this ratio, 0.17 nmol of the inhibitor was bound per nmol of isolated proteolipid. One nmol of dicyclohexylcarbodiimide was bound covalently per six nmol of proteolipid, assuming that no loss of label occurred. On the basis of yield of the proteolipid, it may be calculated that there are some four to six proteolipid molecules per ATPase (CFI moiety) in chloroplast membranes. Covalent labeling by dicyclohexylcarbodiimide of one out of six proteolipids results in complete inhibition of ATPase activity. Proton translocation may be a cooperative process. Carbodiimide binding to one proteolipid out of' a hexamer results in complete inhibition. The structural organization of the dicyclohexylcarbodiimide-binding protein and its participation in possible ion-translocation activity remains to be further investigated. The key to delineation of possible functional activities lies id a defined picture of the proteins conformation in the bilayer. With the information gained through the chemical characterization of the peptide, labeling agents may now be used to advantage. Modification of membrane components may be executed through utilization of hydrophobic labeling agents. Both probes which nonspecifically label, such as the photoactive iodonaphthylazide [22] and group-specific reagents, such as phenylisothiocyanate [23] are currently being utilized in our laboratories and will most certainly prove fruitful in helping define how the proteolipid is situated in the chloroplast membrane. The authors gratefully acknowledge the excellent technical assistance of Miss Elizabeth Kislig and fruitful discussions with Prof. Peter Zahler. These investigations were supported by grants
from the Swiss National Science Foundation (grant no. 3.228-0.77), the €mil Barrel1 Stifttung and the Clark Joller Fond.
REFERENCES 1. Racker, E. & Horstman, L. (1967) J. Biol. Chem. 242, 25472551. 2. McCarthy, R. & Racker, E. (1967) J . Biol. Chem. 242, 34353439. 3. Patel, L., Schuldiner, S. & Kaback, H. (1975) Proc. Natl Acad. Sci. U.S.A. 72, 3387-3391. 4. Cattell, K., Lindop, C., Knight, I. & Beechey, R. (1971) Biochem. J . 125, 169-177. 5. Fillingame, R. (1976) J . B i d . Chem. 251, 6630-6637. 6. Nelson, N., Eytan, E., Notsani, B., Sigrist, H., Sigrist-Nelson, K. & Gitler, C. (1977) Proc. Nutl Acad. Sci. U.S.A. 74, 2375 - 2378. 7. Sigrist, H., Sigrist-Nelson, K. & Gitler, C. (1977) Biochem. Biophys. Res. Commun. 74, 178- 183. 8. Arnon, D. (1959) Plant Physiol. 23, 1 - 15. 9. Lowry, 0.. Rosebrough, N., Farr, A. & Randall, R. (1951) J. Biol. Chem. 193, 265-275. 10. Spiro, R. (1966) Methods Enzymol. 8, 3-26. 11. Weber, K. & Osborn, M. (1969) J . Biol. Chem. 244,4406-4412. 12. Swank, R. & Munkres, K. (1971) Anal. Biochem. 39,462-477. 13. Gray, W. (1972) Methods Enzymol. 25, 121. 14. Sheehan, J. & Yang, D. (1958) J. Am. Chem. Soc. 80, 11541158. 15. Hirs, C. (1967) Methods Enzymol. 25, 59-62. 16. Capaldi, A. & Vanderkooi, G. (1972) Proc. Nutl Acad. Sci. U.S.A.69,930-932. 17. Hoy, T., Ferdinand, W. & Harrison, P. (1974) Int. J. Peptide Protein Res. 6, 121- 129. 18. Sebald, W., Graf, Th. & Wild, G. (1976) in Genetics and Biogenesis of Chloroplast and Mitochondria (Bucher, Th., Nenpert, W., Sebald, W. & Werner, S., eds) pp. 167-174, Elsevier, Amsterdam. 19. Marahiel, M., Imam, G., Nelson, P., Pieniazek, N., Stepien, P. & Kiintzel, H. (1977) Eur. J . Biochem. 76, 345-354. 20. Sebald, W. (1977) Biochim. Biophys. Acta, 463, 1-21. 21. Beechey, R., Roberton, A., Holloway, C. & Knight, I. (1967) Biochemistry, 6, 3867 - 3879. 22. Sigrist-Nelson, K., Sigrist, H., Bercovici, T. & Gitler, C. (1977) Biochim. Biophys. Acta, 468, 163- 176. 23. Sigrist, H. & Zahler, P. (1978) FEBS Lett. in press.
K. Sigrist-Nelson, H. Sigrist, and A. Azzi, Medizinisch-Chemisches Institnt der Universitat Bern, Buhlstrasse 28, CH-3012 Bern, Switzerland
Characterization of the Dicyclohexylcarbodiimide-Binding Protein Isolated from Chloroplast Membranes Kristine SIGRIST-NELSON, Hans SIGRIST, and Angelo AZZI Medizinisch-chemisches Institut und Biochemisches Institut der Universitit Bern (Received August 17, 1978)
Characterization of a butanol-solubilized protein isolated from chloroplast membranes is reported. The proteolipid, which specifically and covalently binds dicyclohexylcarbodiimide, has an apparent molecular weight of 8000 in dodecylsulfate electrophoresis. The minimum molecular weight calculated from amino acid analysis data is 7700. N-Formyl-methionine was determined to be the N-terminal amino acid. Glycine, alanine and leucine were present in elevated amounts, resulting in a polarity of 29 %. Cysteine and histidine were lacking. In high-voltage electrophoresis the peptide appeared as a single homogenous spot which migrated, at pH 6.5, with the relative mobility of glycine. At concentrations where dicyclohexylcarbodiimide inhibited ATPase activity maximally (20 nmol per mg membrane protein), 0.17 nmol dicyclohexylcarbodiimide was covalently bound per nmol isolated proteolipid, indicating that one out of six molecules of proteolipid was labeled.
Dicyclohexylcarbodiimide has long been known to be an inhibitor of oxidative phosphorylation in mitochondria and bacteria as well as photosynthetic phosphorylation in chloroplasts. Various experiments have suggested that the inhibitor's mode of action is to block sites participating in proton translocation [l- 31. Dicyclohexylcarbodiimide has been found to bind selectively to a small polypeptide located in the H +-translocating ATPase complexes present in mitochondrial, bacterial and chloroplast membranes [4,5]. Recently the dicyclohexylcarbodiimide-binding protein was purified from chloroplast membranes and reconstituted. Evidence was presented that the protein was able to function as a dicyclohexylcarbodiimidesensitive proton channel [6]. Heightened interest has been aroused in this protein due to the implications that it may possess biologically important ionophoric activity. An understanding of the mode of functioning of such a protein is only possible when its structural parameters are well defined. To this purpose the purification and chemical characterization of the carbodiimide-reactive component of chloroplasts is reported. MATERIALS AND METHODS Preparation and labeling of lettuce chloroplast membranes with dicyclohexyl['4C]carbodiimide was Abbreviution. Dansyl, S-dimethylaminonaphthalene-l-sulphonyl.
carried out as described elsewhere [6]. Unless otherwise stated all organic solvent systems are given in volume ratios. Pro ttolipid Isolation
The isolation procedure was performed essentially as previously reported [7], except that the butanol solubilizate was additionally filtered through a Satorius filter (11305, 0.6 pM) to ensure complete removal of precipitated protein. The final ether-precipitated proteolipid fraction was washed in three subsequent washes with water, butan-1-01 and cold diethyl ether. Published procedures were followed for the determination of chlorophyll [8] and protein [9]. Chloroplast membrane protein was determined after precipitation of the membrane protein with acetone. The anthrone procedure was used for determination of sugar residues in estimating glycolipid content [lo]. Polyacrylamide Gel Electrophoresis
Two methods of electrophoretic analysis were performed. Sodium dodecyl sulfate/polyacrylamide gel electrophoresis was carried out in accordance with Weber and Osborn [ l l ] . The electrophoresis was performed in 0.1 % sodium dodecyl sulfate, 0.05 M sodium phosphate buffer pH 7.2 using 10% polyacrylamide gels. Gels were stained with Coomassie brilliant blue and destained with 7 acetic acid. Addi-
10
tionally, the sodium dodecyl sulfate/urea system of Swank and Munkres [12] was utilized. The protein was dissolved in 1 % sodium dodecyl sulfate, 8 M urea, 1% mercaptoethanol and 0.01 M sodium phosphate buffer pH 7.2. The samples were run in 10% polyacrylamide gels containing 0.1 % sodium dodecyl sulfate, 8 M urea, 0.1 M sodium phosphate buffer pH 7.2. For molecular weight determination bovine serum albumin ( M , 68000), ovalbumin ( M , 43000), egg white lysozyme ( M , 14300) and bovine pancreatic insulin ( M , 5700) served as standards. When dicyclohe~yl['~C]carbodiimide-labeledsamples were electrophoresed, identical samples were applied to separate gels. One gel was stained, the duplicate gel was immediately cut into 1-mm slices. Individual gel slices were extracted overnight with 10 mM Triton-X-100. Scintillation fluid was added and the slices were counted for radioactive content. Thin-Layer Chroma tography
Silica gel plates without fluorescent indicator were used (Merck no. 5721). A two-dimensional chromatography was performed. The first solvent system was chloroform/methanol/water (65/24/4), the second was chloroform/methanol/isopropylamine/ammonia (65/ 35/0.5/5). Plates were exposed to iodine and then either sprayed with ninhydrin to localize amino-positive lipids and protein or with naphthalindiol for sugar identification. The naphthalindiol spray was composed of 100 mg 1,3-naphthalindiol dissolved in 49 ml of ethanol with 1 ml concentrated sulfuric acid added. When determination of the dicyclohexyl['4C]carbodiimide radioactivity was desired the thin-layer plates were exposed to iodine and radioactive content of the lipid-positive spots determined.
Proteolipid Characterization
organic phase of butanol/acetic acidlwater (41115). The dansylated proteolipid, present at the origin as a green fluorescent spot, was eluted from the silica gel by phenol/acetic acidlwater (2/1/1, w/v/v). The eluted proteolipid was dried under high vacuum and subjected to acid hydrolysis. The dansylated amino acids were analyzed by microscale thin-layer chromatography using polyamide plates with the following solvent systems : 1.5 % formic acid (first dimension), benzene/acetic acid (9/l, second dimension), ethylacetate/methanol/acetic acid (20/1/1, third solvent) or 1 M ammonium hydroxide/ethanol ( l / l , third solvent). Dansylated amino acids served as reference. Removal of the N-formyl group from the terminal amino acid was carried out by mild acid hydrolysis [14]. The washed, ether-precipitated proteolipid was mixed with 10 volumes of methanol16 M HCl (6011) and incubated for 2 h at 37 "C. The proteolipid was recovered by ether precipitation and dansylated by the standard procedure. Amino Acid Analysis
The isolated proteolipid was hydrolyzed with 6 M HCl in an evacuated, sealed tube for 24 h at 100 "C. Cysteine was determined as cysteic acid, methionine as methionine sulfone after performic acid oxidation ~51. ATPase Assay
Light-induced Mg2 -ATPase activity of chloroplasts was assayed according to McCarty and Racker [2]. Preincubation of the chloroplasts with dicyclohexylcarbodiimide was performed for 30 min at 25 "C prior to the ATPase assay.
High- Voltage Electrophoresis
Materials
The proteolipid fraction was dissolved in a small volume of chloroform/methanol (2/1) and applied to Whatman no. 1 paper. The electrode buffer consisted of pyridine/water/acetic acid (100/900/4) pH 6.5. The electrophoresis was carried out for 50 min at 2000 V (50 Vjcm). The strip was then cut into 1-cm sections and the radioactive content determined. Protein was localized with ninhydrin. The amino acids lysine, glycine and glutamic acid served as mobility markers.
N,N '-Dicyclohexyl[14C]carbodiimide (50 Ci/mol) was the generous gift of Dr H. R. Kaback (Roche Institute of Molecular Biology, Nutley). All other chemicals and reagents were of the highest quality commercially available.
N- Terminal Determination N-terminal amino acid determination was carried out by dansylation and thin-layer chromatographic identification of dansyl derivatives [13]. In a first step, thin-layer chromatography on silica gel was used for separation of reaction by-products from the dansylated proteolipid. The solvent system employed was the
RESULTS Isolation
The technique previously utilized for isolating proteolipid from sarcoplasmic reticulum membranes and the inner mitochondria1 membrane [7] has proven successful for isolating the dicyclohexylcarbodiimidebinding proteolipid from chloroplast membranes. The procedure involves injection of the membranes into butan-1 -01, followed by separation of the precipitated protein from the butanol. The butan-1-01 solubilizate
11
K. Sigrist-Nelson, H. Sigrist, and A. Azzi Table 1, Isolution of chloroplast proteolipid Sugar was determined by the anthrone method as described in Methods. D-GalaCtOSe served as standard
A Fraction
Chloroplast membrane Butanol: supernatant precipitate Ether: supernatant precipitate
Protein recovered mg
%total
10
100
-
-
9.9 -
99 -
0.1
1
B Ether precipitate
Ratio of protein to glycolipid ~
20 30 40 Slice number
llgilrg sugar Unwashed Washed with ether Washed with butan-1-01
0.94 1.03 8.04
is then combined with diethylether, resulting in precipitation of the proteolipid fraction. Table 1 A shows the distribution of protein throughout the isolation procedure. Of the total membrane protein, 1 % is present in the ether precipitate. It is to be noted that complete removal of the butanol-precipitated protein is required to prevent contamination of this fraction. This is best accomplished by filtration of the butanol prior to ether addition. In addition to protein, lipid is also present in the ether precipitate. The amount of glycolipid may be, however, reduced by washing the etherprecipitated fraction with butan-1-01. While butan1-01 does not serve to solubilize ether-precipitated protein, it does reduce the glycolipid content of this fraction (Table 1B). Washing of the precipitated fraction with ether does not significantly affect the protein to glycolipid ratio. Additionally, the ether precipitate was routinely washed with water to reduce the amount of salts which are soluble to a limited extent in butanol and precipitate upon addition of ether. Solvents other than butan-1-01 which solubilized the proteolipid include methanol, ethanol, propanol, acetone and chloroform/methanol (2/1). A mixture of chloroform/methanol/diethylether (2/1/12) served to extract the major portion of chlorophyll and lipid while leaving the proteolipid unsolubilized. Chloroplast membranes were incubated with dicyclohexyl[14C]carbodiimide.Electrophoretic characterization of the labeled membranes and proteolipid fraction obtained during the isolation process is shown in Fig. 1. The radioactivity was covalently associated with one Coomassie-blue-staining band in the membrane fraction (Fig. 1A). The dicyclohexylcarbodiimide-binding protein was present only in
50
60
B
0
10
20
30 40 Slke number
50
60
Fig. 1, Electrophoretic analysis o j fractions obiuined during proteolipid isolution. Dicyclohexyl['4C]carbodiimide-labeled chloroplast membranes (10 mg protein/ml) were injected into butan-1-01, 0.2 m1/40 ml butan-1-01. The butanol precipitate was removed and the supernatant was combined with 5 volumes o f cold diethyl ether. The resulting ether precipitate was collected. (A) Di~yclohexyI['~C]carbodiimide-labeled chloroplast membranes (70 pg); (B) proteolipid fraction (ether precipitate, 10 pg)
greatly reduced amounts in the butanol-precipitated protein (not shown). In contrast, the dicyclohexylcarbodiimide-binding proteolipid was present as a single Coomassie-blue-staining band, with associated radioactivity in the ether precipitate (Fig. 1 B). The apparent molecular weight of the proteolipid in the sodium dodecyl sulfate system was 8000 [6]. Electrophoretic analysis of the proteolipid fraction was performed with gels containing sodium dodecyl sulfate and 8 M urea (Fig. 1B). The protein pattern with associated radioactivity is identical to that of [6].
Thin-Layer Chromatography Two-dimensional thin-layer chromatography of the ether-precipitated dicyclohexyl[14C]carbodiimidelabeled proteolipid fraction was performed (Fig.2).
12
Proteolipid Characterization
1029 @Start
Fig. 2. Two-dimensional thin-layer chromatogruphy of the proteolipid fraction. The washed (see Methods) proteolipid fraction (3 pg protein) was dissolved in chloroform/methanol (2/1) and applied to the thin-layer plate. Solvent systems and developmental procedures are described in the Methods section. Vertical bars indicate ninhydrin-positive spots; horizontal bars, naphthalindiol-positive spots. Radioactivity (presented as counts/min) was only associated with one spot
The chromatogram revealed the presence of lipids in addition to the proteolipid which was present as a single ninhydrin-positive spot remaining at the origin. Radioactivity was associated only with the ninhydrinpositive spot. Of the four lipid spots present, three reacted positively with naphthalindiol indicating the presence of carbohydrate moieties.
High- Voltage Electrophoresis When the dicycl~hexyl[’~C]carbodiimide-labeled chloroplast proteolipid was subjected to high-voltage electrophoresis only a single ninhydrin spot was detectable (Fig. 3). All of the radioactivity co-migrated with the protein. The relative mobility of the proteolipid was identical to glycine at pH 6.5 and was not altered by the presence of bound dicyclohexylcarbodiimide. N- Terminal Determinat ion Additional evidence for chemical homogeneity of the dicyclohexylcarbodiimide-bindingproteolipid was obtained by end-group analysis. The proteolipid was dansylated as described. Before acid hydrolysis was performed, preparative thin-layer chromatography was used to separate reaction by-products from the dansylated proteolipid. Three fluorescent bands were in evidence on the silica gel plates, a green band at the origin, a blue band with an RF of 0.4 and a pinkyellow band with an RFof 0.7. The proteolipid present at the origin could only be successfully eluted by either acidified chloroform/methanol (2/1) or phenol/ acetic acid/HzO (2/1/1, w/v/v). Other solvents which were tried but were unsuccessful include butan-1-01,
chloroform/methanol (2/1), 1.5 ”/, and 10 % formic acid, butan-1-ol/formic acid (50/1) and H2O/NH3 (5/l>. Following acid hydrolysis of the eluted proteolipid, the dansylated amino acids were analyzed. The only dansylated amino acids detectable in the hydrolyzate were &-N-dansyl-lysineand 0-dansyl-tyrosine, indicating that the N-terminal amino acid is either sterically hidden or chemically blocked. Treatment of the proteolipid with mild acid hydrolysis, prior to dansylation, resulted in the appearance of N-dansylmethionine. The N-terminal amino acid was thus judged to be N-formyl-methionine.
Amino Acid Composition and Molecular Weight The amino acid analysis of the chloroplast proteolipid is reported in Table 2. To be noted is the high preponderance of the nonpolar amino acids glycine, alanine and leucine. A polarity of only 29% was calculated using the formula of Capaldi and Vanderkooi [16]. After subjecting the results from the amino acid analysis to the search procedures of Hoy et al. [17] a coplot ofj; an indicator of goodness of fit to integral numbers against the molecular weight was carried out. A pronounced minimum was obtained at 7700.
Inhibition of ATPase Activity and Binding of’ Dicyclohexyl[ ‘‘C]carbodiimide to the Chloroplast Proteolipid Whole chloroplasts were incubated for 30 min with different amounts of dicyclohexyl[’4C]carbodiimide relative to the chloroplast protein content. Maximal inhibition of activity was obtained at a ratio of 20 nmol dicyclohexylcarbodiimide per mg chloroplast membrane (Fig. 4). The quantity of dicyclohexylcarbodiimide bound by the proteolipid varied in proportion to the amount of dicyclohexylcarbodiimide added to the membranes. The proteolipid bound a maximum of 1 nmol dicyclohexylcarbodiimide per nmol proteolipid when the reagent was present at concentrations of 150 nmol per mg chloroplast membrane protein or greater. Interestingly, however, maximal inhibition of ATPase activity was achieved when only 0.17 nmol dicyclohexylcarbodiimide was bound per nmol proteolipid. Approximately 60 inhibition of ATPase activity was observed when 0.084 nmol of dicyclohexylcarbodiimide was bound per nmol proteolipid.
DISCUSSION A polypeptide soluble in butan-1-01 and originating from chloroplast membranes has been charac-
13
K. Sigrist-Nelson, H. Sigrist, and A. AzLi Glutamic
acid
I Glycine
Lysine
1
0
0
Fig. 3. High-voltage electrophoresis of the dieyclohexylcarbodiimide-binding proteolipid. 10 pg of dicycl~hexyl['~C]carbodiimide-labeled proteolipid were dissolved in chloroform/methanol (2/1) and applied to Whatman no. 1 paper. Electrophoresis conditions are presented in Methods. Lysine, glycine and glutamic acid served as electrophoretic mobility markers. The proteolipid and amino acids were localized with ninhydrin staining. Distribution of radioactivity on the electropherogram is displayed
Table 2. Amino acid composition of' the chloroplast dicyclohexylcarbodiimide-reactiveproteolipid The results are the average of ten separate hydrolysatcs. Cysteine was analyzed as cysteic acid and methionine as methionine sulfone, following performic acid oxidation. Tryptophan was not determined. Results are means S.D.
*
Amino acid
Amount present mol/mol
Aspartic acid Threonine Serine Glutamic acid Proline Glycine Alanine Cysteine Valine Methionine Isoleucine Leucine Tyrosine Phenylalanine Lysine Histidine Arginine Tryptophan
3.81 k 0.30 3.25 f 0.35 4.92 0.39 5.72 0.29 5.16 k 0.72 9.72 0.82 12.04 k 1.02 0.20 k 0.25 5.98 k 0.42 1.11 f 0.21 4.43 f 0.43 9.77 0.81 1.26 k 0.15 3.90 k 0.35 1.49 k 0.19 0.22 f 0.28 1.97 k 0.10 n.d.
*
*
x Polarity
29
terized. The proteolipid which specifically and covalently binds dicyclohexylcarbodiimide has an apparent molecular weight of 8000 in dodecylsulfate electrophoresis. Thin-layer chromatography of the proteolipid fraction revealed the presence of four lipids, three containing carbohydrate moieties. Both high-voltage electrophoresis and N-terminal amino acid analysis indicated the presence of only one polypeptide in the isolated fraction. Similar to the
0
25 50 75 100 125 150 175 200 225 Dicyclohexylcarbodiimide added ( n m o l i mg membrane)
0
Fig.4. Inhihition of ATPuseactivityandbindin~qfdicyclohexyl['4Cjcurbodiimide to the chloroplast proteolipid at varied dicyclohexylcarbodiimide :chloroplast membraneprotein ratios. Chloroplast membranes were incubated with the varied amounts of dicyclohexylcarbodiimide indicated and ATPase activity was assayed (a).The proteolipid was isolated and the amount of bound inhibitor was determined (-1
dicyclohexylcarbodiimide-binding proteolipid isolated from Saccharomyces cerevisiae [18] and Aspergillus nidulans [ 191, N-formyl-methionine was present as the N-terminal amino acid. The chloroplast dicyclohexylcarbodiimide-binding protein displayed a high content of nonpolar amino acids. The absence of cysteine and histidine in the chloroplast proteolipid composition accentuated its correspondence with proteins isolated from other systems (Escherichia coli, [5] ; Saccharomyces cerevisiae and Neurospora crassa [20]). The calculated polarity for the chloroplast proteolipid was 29 %, indicating a rather marked hydrophobic nature. Other published values for dicyclohexylcarbodiimide-bindingproteins include Aspergillus nidulans, 38 % [19], Escherichia coli, 16 % [ 5 ] and Neurosporu crassa, 29 % [20]. Similar to observations with mitochondria, complete inhibition of ATPase activity was observed
14
K. Sigrist-Nelson, H. Sigrist, and A. Azzi: Proteolipid Characterization
when chloroplast membranes were incubated for short periods of time with dicyclohexylcarbodiimide at a ratio of 20 nmol/mg membrane protein [21]. At this ratio, 0.17 nmol of the inhibitor was bound per nmol of isolated proteolipid. One nmol of dicyclohexylcarbodiimide was bound covalently per six nmol of proteolipid, assuming that no loss of label occurred. On the basis of yield of the proteolipid, it may be calculated that there are some four to six proteolipid molecules per ATPase (CFI moiety) in chloroplast membranes. Covalent labeling by dicyclohexylcarbodiimide of one out of six proteolipids results in complete inhibition of ATPase activity. Proton translocation may be a cooperative process. Carbodiimide binding to one proteolipid out of' a hexamer results in complete inhibition. The structural organization of the dicyclohexylcarbodiimide-binding protein and its participation in possible ion-translocation activity remains to be further investigated. The key to delineation of possible functional activities lies id a defined picture of the proteins conformation in the bilayer. With the information gained through the chemical characterization of the peptide, labeling agents may now be used to advantage. Modification of membrane components may be executed through utilization of hydrophobic labeling agents. Both probes which nonspecifically label, such as the photoactive iodonaphthylazide [22] and group-specific reagents, such as phenylisothiocyanate [23] are currently being utilized in our laboratories and will most certainly prove fruitful in helping define how the proteolipid is situated in the chloroplast membrane. The authors gratefully acknowledge the excellent technical assistance of Miss Elizabeth Kislig and fruitful discussions with Prof. Peter Zahler. These investigations were supported by grants
from the Swiss National Science Foundation (grant no. 3.228-0.77), the €mil Barrel1 Stifttung and the Clark Joller Fond.
REFERENCES 1. Racker, E. & Horstman, L. (1967) J. Biol. Chem. 242, 25472551. 2. McCarthy, R. & Racker, E. (1967) J . Biol. Chem. 242, 34353439. 3. Patel, L., Schuldiner, S. & Kaback, H. (1975) Proc. Natl Acad. Sci. U.S.A. 72, 3387-3391. 4. Cattell, K., Lindop, C., Knight, I. & Beechey, R. (1971) Biochem. J . 125, 169-177. 5. Fillingame, R. (1976) J . B i d . Chem. 251, 6630-6637. 6. Nelson, N., Eytan, E., Notsani, B., Sigrist, H., Sigrist-Nelson, K. & Gitler, C. (1977) Proc. Nutl Acad. Sci. U.S.A. 74, 2375 - 2378. 7. Sigrist, H., Sigrist-Nelson, K. & Gitler, C. (1977) Biochem. Biophys. Res. Commun. 74, 178- 183. 8. Arnon, D. (1959) Plant Physiol. 23, 1 - 15. 9. Lowry, 0.. Rosebrough, N., Farr, A. & Randall, R. (1951) J. Biol. Chem. 193, 265-275. 10. Spiro, R. (1966) Methods Enzymol. 8, 3-26. 11. Weber, K. & Osborn, M. (1969) J . Biol. Chem. 244,4406-4412. 12. Swank, R. & Munkres, K. (1971) Anal. Biochem. 39,462-477. 13. Gray, W. (1972) Methods Enzymol. 25, 121. 14. Sheehan, J. & Yang, D. (1958) J. Am. Chem. Soc. 80, 11541158. 15. Hirs, C. (1967) Methods Enzymol. 25, 59-62. 16. Capaldi, A. & Vanderkooi, G. (1972) Proc. Nutl Acad. Sci. U.S.A.69,930-932. 17. Hoy, T., Ferdinand, W. & Harrison, P. (1974) Int. J. Peptide Protein Res. 6, 121- 129. 18. Sebald, W., Graf, Th. & Wild, G. (1976) in Genetics and Biogenesis of Chloroplast and Mitochondria (Bucher, Th., Nenpert, W., Sebald, W. & Werner, S., eds) pp. 167-174, Elsevier, Amsterdam. 19. Marahiel, M., Imam, G., Nelson, P., Pieniazek, N., Stepien, P. & Kiintzel, H. (1977) Eur. J . Biochem. 76, 345-354. 20. Sebald, W. (1977) Biochim. Biophys. Acta, 463, 1-21. 21. Beechey, R., Roberton, A., Holloway, C. & Knight, I. (1967) Biochemistry, 6, 3867 - 3879. 22. Sigrist-Nelson, K., Sigrist, H., Bercovici, T. & Gitler, C. (1977) Biochim. Biophys. Acta, 468, 163- 176. 23. Sigrist, H. & Zahler, P. (1978) FEBS Lett. in press.
K. Sigrist-Nelson, H. Sigrist, and A. Azzi, Medizinisch-Chemisches Institnt der Universitat Bern, Buhlstrasse 28, CH-3012 Bern, Switzerland
