Structural and Functional Properties of ... - Wiley Online Library

Eur J Biochem. 131, 97- 103 (1983) h) FEBS 1983

Structural and Functional Properties of Cytochrome c Oxidase from Bacillus subtilis W23 Wim DE VRIJ, Angelo AZZI and Wit N. KONINGS Department of Microbiology, University of Groningen; and Medizinisch-Chemisches Institut, University of Bern (Received July 9/November 4, 1982)

-

EJB 5745

The terminal component of the electron transport chain, cytochrome c oxidase (ferrocytochrome c : oxygen oxidoreductase) was purified from Bacillus subtilis W23. The enzyme was solubilized with alkylglucosides and purified to homogeneity by cytochrome c affinity chromatography. The enzyme showed absorption maxima at 414 nm and 598 nm in the oxidized form and at 443 nm and 601 nm in the reduced form. Upon reaction with carbon monoxide of the reduced purified enzyme the absorption maxima shifted to 431 nm and 598 nm. Sodium dodecylsulfate polyacrylamide gel electrophoresis indicated that the purified enzyme is composed out of three subunits with apparent molecular weights of 57000, 37000 and 21 000. This is the first report on a bacterial aa,-type oxidase containing three subunits. The functional properties of the enzyme are comparable with those of the other bacterial cytochrome c oxidases. The reaction catalyzed by this oxidase was strongly inhibited by cyanide, azide and monovalent salts. Furthermore a strong dependence of cytochrome c oxidase activity on negatively charged phospholipids was observed. Crossed immunoelectrophoresis experiments strongly indicated a transmembranal localization of cytochrome c oxidase.

Eukaryotes and many prokaryotes contain cytochrome c oxidase as the terminal component of the electron-transport chain [I]. The enzyme plays a fundamental role in aerobic oxidation and energy metabolism. The most important functional property is the transfer of electrons from reduced cytochrome c to oxygen. The free energy of this reaction is conserved as an electrochemical proton gradient [2]. Bacterial aa,-type oxidases initially received attention from a viewpoint of evolution of heme proteins [3]. A general difficulty in the study of bacterial cytoplasmic membranebound proteins, is their low content in these membranes [4]. Bacteria are capable of adjusting to various environmental conditions and of synthesizing individual redox components in response to environmental changes [5 - 71. The structural and functional properties of cytochrome c oxidases (aa,-type) from several different bacteria have recently been determined [S, 91. A general structural characteristic of bacterial cytochrome c oxidases is their relative simple polypeptide composition compared to eukaryotic oxidases. Previous studies about the energy-transducing properties of this membrane-bound protein, suggested that an electrochemical proton gradient is generated across the membrane, only by proton consumption on the matrix (mitochondria) or cytoplasmic (bacteria) side of the membrane [lo]. More recent studies, however, indicate that the enzyme also acts as an outwardly directed proton pump [I 1 - 131. Except for Paracoccus denitriffl'cans cytochrome oxidase, conclusive evidence for a proton-pumping activity of bacterial cytochrome c oxidases has not yet been presented [I 41.

It is, therefore, of interest to investigate whether proton pumping is a more general property of eukaryotic or bacterial cytochrome c oxidases. A number of functional and structural properties of respiratory chain proteins of the gram-positive bacterium Bacillus subtilis W23 have already been investigated [15 - 191. In this investigation we describe the isolation of B. subtilis W23 cytochrome c oxidase. We used cytochrome c affinity chromatography to purify this oxidase. This is a more simple and less destructive technique than previously described isolation procedures for cytochrome c oxidases [20]. With this technique a successful isolation of several bacterial and eukaryotic cytochrome c oxidases has been achieved [21]. Several structural and functional characteristics of this enzyme have been determined. MATERIALS A N D METHODS Cell Growth and Preparation of Membrane Vesicles

Bacillus suhtilis W23 was grown at 37 C with vigorous aeration in a medium containing 0.8 % (w/v) trypton (Difco), 0.5 (w/v) NaCl, 25 mM KCI and 150 pl/l micronutrient solution. This micronutrient solution contained; MnCI, (2.2 %), ZnS0,.7 H,O (0.05%), H,BO, (0.5%), CuSO, 5 H,O (0.01 6 %), Na,MoO,. 2 H,O (0.025 %), Co(NO,), ' 6 H,O (0.46%) and 0.5 m1/100 ml solution concentrated H2S04. The medium was supplemented with CaCI, . 2 H,O (50 mg/l) and Abbreviations. Hepes 4-(2-hydroxyethyl-l-piperazineethanesulfonic FeSO, ' 7 H,O (50 mg/l). Logarithmically growing cells were harvested at an absoracid; I,,, median inhibitory concentration. bance at 660 nm of 0.8 - 1.0. Membrane vesicles were prepared Enzyme. Cytochrome c oxidase o r ferrocytochrome c: oxygen oxidoas described by Bisschop and Konings [I61 and finally rereductase (EC 1.9.3.1).

98

suspended in a medium containing 50 mM Tris/HCI, p H 7.4 and 1 mM EDTA. Soluhilization of Cytochrome c Oxidase Cytochrome c oxidase was solubilized by incubating membrane vesicles with 1 %, (w/v) laurylmaltoside (dodecyl P-D-maltoside) in 50 mM Tris/HCl, pH 7.4 and 1 mM EDTA for 30 min at 2 0 ' C with gentle stirring (2mg membrane protein/mg laurylmaltoside). The mixture was centrifuged (30 min, 48000 x g at 4°C). If not immediately used, the supernatant was rapidly frozen in liquid nitrogen and stored at - 80 "C. Laurylmaltoside has been synthesized and purified according to Rosevear et al. [22]. Purification Procedure qf Cytochrome c Oxidase Yeast-cytochrome-c thiol-Sepharose 4B gel was prepared essentially as described by Bill et al. [21]. Saccharomyces cerevisiae cytochrome c (25 mg) was bound to approximately 6 ml swollen activated thiol-Sepharose 4B gel. Thiol groups of the gel, which did not react with cytochrome c were blocked by washing with 50 ml 50 mM acetate buffer, pH 4.5, containing 1.5 mM 2-mercaptoethanol. Under these conditions, no covalently bound cytochrome c was removed from the gel. The gel was poured into a column (1 x 10cm). The column was washed with 100 ml 50 mM Tris/HCI, pH 7.4, containing 1 % (w/v) Triton X-100, 1 mM ferricyanide and 1 M KCI, to remove non-covalently bound cytochrome c and to oxidize fully cytochrome c. The column was finally washed with 200 rnl 50 mM Tris/HCI, pH 7.4 and 1 mM EDTA and subsequently equilibrated with 50 mM Tris/HCl, pH 7.4, 1 mM EDTA and 0.1 % (w/v) laurylmaltoside, unless stated otherwise. A diluted laurylmaltoside membrane extract was loaded on the affinity column (final detergent concentration 0.2 %,, w/v). In a typical experiment 10 - 15 nmol heme a was loaded on the column at a flow rate of 7 ml/h. Subsequently the column was washed with 50 mM Tris/HCI, p H 7.4,l mM EDTA and 0.1 "/, (w/v) laurylmaltoside. Fractions (5 ml) were collected and analyzed spectrophotometrically for cytochromes using an Aminco-DW2a spectrophotometer (American Instrument Company, Silver Spring, MD, USA). If no heme h and heme c could be detected any more in the fractions, the elution was continued with 50 mM Tris/HCl, p H 7.4, 1 mM EDTA, 0.1 % (w/v) laurylmaltoside and 100 mM KCl to remove heme a from the column. Fractions containing the highest heme a to protein ratios were combined and loaded on a 2 ml Amberlite CG-50 column in order to remove contaminants of S. cerevisiae cytochrome c. This column was equilibrated with 5 0 m M Tris/HCI, pH 7.4 and 0.1 % (w/v) laurylmaltoside. Due to the high isoelectricpoint (1 0.5) this cytochrome c binds under these conditions to the column material. Elution with 5 0 m M Tris/HCl, p H 7.4 and 0.1 (w/v) laurylrnaltoside results in a separation of heme u and S. cerevisiae cytochrome c. Fractions containing spectrophotometrically pure heme a were pooled and loaded on a 1 ml DEAE-Biogel column, which was equilibrated with 50 mM Tris/HCl, pH 7.4,O.l o/, (w/v) laurylmaltoside and 100 m M KCI. The DEAE-Biogel can be used as an excellent final step in the isolation procedure to concentrate heme a and to remove final contaminants. Heme a was eluted from the column using 10 ml 50 mM Tris/HCl, p H 7.4, 1 mM EDTA, 0.1 (w/v) laurylmaltoside and 200 mM KCI at a flow rate of 7 ml/h. If necessary the samples were further concentrated with Amicon-B15 ultrafilters. All actions were carried out at 4 'C.

Protein Determination Protein was determined by the method of Wang and Smith [23] or by the method of Lowry et al. [24], using bovine serum albumin as a standard. Spectral Analysis Spectrophotometric determinations were performed with an Aminco DW-2a spectrophotometer. The concentrations of heme a, heme b and heme c + c I were calculated using absorption coefficients (reduced minus oxidized) of 16.5mM-' cm-' (601-630nm), 2 0 m M - ' cm-' (559577 nm) and 19.0 mM-' cm-' (553 -538 nm), respectively, assuming that the values are similar to those of mammalian components reported by Yonetani I251 and Chance and Williams [26].

~

Polyacrylumide Gel Electrophoresis Sodium dodecylsulphate polyacrylamide gel electrophoresis was performed according to Laemrnli [27] in vertical slab gels or tube gels (4 mm x 10 cm). Native polyacrylamide gel electrophoresis was performed as described by Pentilla et al. [28] in tube gels (3 mm x 10 cm) using Tween-80 as non-ionic detergent. The gels were loaded with 50- 100 pg purified cytochrome aa, supplemented with 40% saccharose in 40 mM Tris/glycinate p H 8.3, 0.01 "/, (w/v) sodium dodecylsulphate and 0.1 (w/v) Tween-80, in order to reach a higher buoyant density. Electrophoresis was carried out at 4 ° C with 100 V for 6 h. Following electrophoresis the gel was stained for cytochrome c oxidase activity or heme as described in cytochrome c oxidase assays. Stained bands were cut out of the gel and incubated in Laemmli sample buffer for 4 h. Finally the gel pieces were transferred to tubes (4 mm x 10 cm) containing 10 % gels as described by Laemmli 1271. Electrophoresis was carried out at room temperature with 100 V for 8 h. The gels were stained with Coomassie brilliant blue and scanned in tubes at 570 - 530 nm in a special scanner attached to the Aminco DW-2a spectrophotometer [29]. Bovine serum albumin, carbonic anhydrase, myoglobin and cytochrome c were used as molecular weight standards. Cytochrome c Oxidase Assays For detection of cytochrome c oxidase activity the following staining procedure was used. N,N,N',N-Tetramethyl-pphenylenediamine (100 mg) was dissolved in 50 ml 50 mM Tris/HCI, p H 7.4. The gels were soaked in this solution. After 10 min a violet colour appears. Heme was stained by soaking the gel for 30 min in a solution composed of 10 ml methanol containing 5 mg tetramethylbenzidine and 23 ml 0.25 M sodium acetate, pH 5.0. After addition of 100 p1 of 30 %, H,Oz a green colour appeared in 10- 15 min. Determination of Enzymatic Activity Cytochrome c oxidase activity was measured at 37 C by following the decrease in absorbance of the a-peak of cytochrome c, using an absorption coefficient (reduced minus oxidized) of 19.5 mM-' cm-' (550-540 nm) as described by Yonetani [30]. Ferrocytochrome c was prepared by adding a few grains of sodium dithionite to 0.5 ml of a solution

99

containing 20 mg ferricytochrome c in 100 mM Hepes and 40 mM Tris/HCl, pH 7.4. To remove excess dithionite the solution was applied to a column (15 cm x 1 cm) of Sephadex G-25 (course grade) eluted with 100 mM choline chloride. The concentration of eluted ferrocytochrome c was approximately 1 mM, as measured from the absorbance change at 550540 nm. The reaction mixture contained 1.4 ml 50 mM Tris/HCl, pH 7.4, 1 mM EDTA and different concentrations of reduced cytochrome c. An appropriate amount of cytochrome c oxidase was finally added to start the reaction. The reaction rate was calculated from the initial slope and expressed as the molecular activity [mol of ferrocytochrome c oxidized (mole oxidase) x s '1, unless indicated otherwise. Oxygen consumption was measured polarographically in 3.4-1111 reaction mixture, using a Clark-type electrode (Yellow Springs Instrument Co.). Enzyme was assayed at 37 "C in the presence of 40 mM Hepes p H 7.4, 30 mM sodium ascorbate and 40 pM horse heart cytochrome c or 100 pM N,N,N',N'-tetramethyl-p-phenylenediamine.

dimension agarose gel. If not immediately used, the supernatants were stored at 4 "C.

Materials S. cerevisiae cytochrome c (type VIII), horse heart cytochrome c (type VI) and asolectin were obtained from Sigma (St Louis, MO, USA). Activated thiol-Sepharose 4B was obtained from Pharmacia (Uppsala, Sweden). DEAE-Biogel was obtained from Bio-Rad (Richmond, CA, USA). Phosphatidylserine used here was from Lipid Products (South Nutfield, England).

~

~

Isolation of Anti- Vesicle Immunoglobulins Anti-(glucose vesicles) immunoglobulins were obtained as described [15].

Solubilization of Membrane Vesicles jiw Crossed Immunoebctrophoresis

For use in the first dimension of crossed immunoelectrophoresis, membrane vesicles were solubilized with a Triton X-100 to a protein ratio of 5: 1 (w/w). Maximal solubilization was obtained by incubating for 1 h at pH 7.5 and 20 "C under occasional shaking. After centrifugation (1 h, 48000 x g at 4°C) the supernatants were stored in liquid nitrogen, if not immediately used. With this method about 80% of the membrane proteins were solubilized. Crossed Immunoelectrophoresis Crossed immunoelectrophoresis was performed as described by Smyth et al. [31]. The electrophoresis was performed as described [I 51. Cytochrome c oxidase activity and succinate dehydrogenase activity of the immunoprecipitates were detected as described by Owen and Salton 1321 and Burstone [33). Peak areas were determined with a Hewlett Packard integrator.

RESULTS

Isolation The cytochrome content of Bacillus subtilis W23 cells varies with the growth conditions and the growth stage. Cytoplasmic membranes from B. subtilis W23, obtained as described in the Materials and Methods, contain 0.4-0.5 nmol heme a/mg membrane protein, 0.7 - 0.8 nmol heme b/mg membrane protein and varying small amounts of other types of heme c. Spectral characteristics of these types of cytochromes are shown in Fig. 2A. These cytochromes can effectively be solubilized with laurylmaltoside. With a laurylmaltoside to membrane protein ratio of 0.5 (w/w) approximately 70 % of the membrane proteins, including the different types of cytochromes, can be solubilized. In this study a laurylmaltoside to membrane protein ratio of 0.5 was used. After centrifugation the protein concentration in the supernatant was about 14 mg protein/ml; the concentration of heme a and heme b varied from 8 - 10 pM and 14- 16 pM, respectively. This cytoplasmic membrane extract was diluted five times and applied on a cytochrome c affinity column to which approximately 330 nmol Saccharomyces cerevisiae cytochrome c/ml swollen gel were bound. The maximum binding capacity of the column was about 6 pmol heme alnmol cytochrome c. The binding of heme b and heme a is strongly dependent on the concentration of the detergent in the elution buffer. At detergent concentrations of 0.1 % (w/v) the binding of heme b, but not of heme a, is weak. Heme b can be bound at lower concentrations of

Localization Experiments The localization of the enzyme in the vesicle membranes was determined with the immunoabsorption technique described by Owen and Kaback [34]. Membrane vesicles were resuspended in 50mM Tris/HCI, pH 7.4 to a final protein concentration of 5 mg/ml. The samples were supplemented with 5 Triton X-100 (final concentration, v/v) or with the same volume of buffer. The suspensions containing Triton X-100 were incubated at 25 "C under frequent shaking for 1 h, while the other samples were kept at 4 "C for the same period. Absorption experiments were performed by adding 0.25 ml IgG solution (containing 17 mg/ml anti-vesicle immunoglobulins) to 0 - 85 pI intact or solubilized membrane vesicle suspension. Tris/HCl, 50 mM, p H 7.4 was added to give a final volume of 0.335 ml. After incubation, under frequent shaking for 1 h at 25 "C, the vesicles and precipitated immunoglobulins were removed by centrifugation (30 min, 48000 x g at 4 "C) and 280 pl of the supernatant was incorporated in the second

2

4

6

8

10

12

FRRCTION NUMBER

Fig. 1. Elution proyile of a cytochrome c ctJfitzity chro~nutography.A 6-ml column, equilibrated as described, was loaded with 5 ml 0.2% (w/v) laurylmaltoside cytoplasmic membrane extract, containing 10.5 nmol cytochrome au3,at 4 " C with a flow rate of 7 ml/h. Elution continued using 50 mM Tris/HCl, pH 7.4, 1 mM EDTA and 0.1 % laurylmaltoside to remove heme b from the column. If no heme b was detectable in the fractions, elution was continued with 100 mM KCI added to the elution buffer (arrow). Fractions of 5 ml were collected and analyzed spectrophoheme a (nmol) tometrically for cytochromes. ( 0 )Heme b (nmol); (0)

w 0 z 4

m a: 0 Ul

i

m 4

]A002

500

30

WAVELENGTH(nm1

L

600

I WAVELENGTHlnrn) 500

LOO

600

Fig. 2. Absorption spectra of solubilized membranes and purified cytochrome c oxidase of Bacillus subtilis W23. (A) Dithionite reduced minus oxidized spectrum of a laurylmaltoside cytoplasmicmembrane extract, containing 0.8 nmol cytochrome aa,/ml. (B) Dithionite reduced minus oxidized spectrum of purified cytochrome au,, containing 0.12 nmol/ml. Baselines (oxidized minus oxidized) are given by the dashed lines. Spectra were recorded at room temperature using an Aminco DW-2a spectrophotometer at a scanspeed of 2 nmjs and bandwidth of 3 nm

G 3

443 1

w

z u

u w

4

z

m

4

(L

m

v1 0

n L” 0

m a

m

’\

U

I

u5 J

do

I

500

I

I

00

600

WAVELENGTH(nm1

500

I

600

WAVELENGTH (nml

Fig. 3. Absorption spectra ofpur+edBacillus subtilis W23 cytochrome c oxidase. (A) Absolute spectra of purified cytochrome aa,, containing 0.12 nmo!/ml. ( .....)Absolute spectrum, oxidized enzyme; (--) absolute spectrum, reduced with dithionite; (----)baseline. (B) Carbon monoxide spectra of purified cytochrome aa,, containing 0.85 nmol/ml. The sample was reduced with dithionite and CO was bubbled through the solution for 2 min. (......) Difference absolute spectrum; (-- --) baseline spectrum, CO reduced minus reduced; (--)

protein ratios (varying from 10 - 13 nmol/mg protein) were pooled and applied on an Amberlite CG-50 column. A purification of heme a with the cytochrome c affinity column of 15- 20-fold could be achieved. After the cationexchange column the highest heme a to protein ratios (nmoljmg) reached were 15 - I 7 which was spectroscopically pure cytochrome c oxidase. If necessary, a DEAE-Bio-Gel column could be used to concentrate the fractions containing cytochrome c oxidase. J

Spectral Properties

MIGRATION4

Fig. 4. Electrophoreticprofile ofBacillus subtilis W23 cytochrome c oxiduse. 10 sodium dodecylsulphate/polyacrylamidegel electrophoresis pattern of 0.3 nmol purified cytochrome c oxidase. After electrophoresis the gel was stained with Coomassie brilliant blue RC250. The stained gcl was scanned in a special scanner attached to the Aminco DW-2a spectrophotometer at 570-530 nm

detergents (0.02 %, wjv). In Fig. 1a typical elution profile with 0.1 ”/, (w/v) laurylmaltoside is shown. Heme b and some heme a eluted first from the column. If heme b was not further detectable in the fractions, elution was performed with the same buffer containing 100 mM KCl. At this ionic strength heme a and some S. cerevisiae cytochrome c eluted simultaneously from the column. Fractions containing the highest heme a to

The purified enzyme showed spectral characteristics of an aa,-type oxidase (Fig. 2b, 3). N o other cytochromes could be detected. The enzyme showed absorption peaks at 414 nm and 598 nm in the oxidized form and at 443 nm and 601 nm in the reduced form (Fig. 3a). The presence of cytochrome a, was demonstrated by the absolute and difference spectra of the COsaturated enzyme (Fig. 3b). The complex of CO and cytochrome c oxidase (in the reduced form) showed peaks at 431 nm and 598 nm. Molecular Weight Electrophoretic analysis of the fractions with the highest heme a to protein ratios showed four main bands. These fractions were contaminated with small amounts of cyto-

101 Table 1. Measurements of oxygen consumption of purified oxidase and oxidase in the cytoplasmic membranes Polarographic measurements were performed as described at 37 "C. The enzyme activity was expressed as molecular activity (mol of ferrocytochrome c oxidized (mol oxidase)-' x s - l Assay

Enzyme activity cytoplasmic membranes

mol x mol-' x s-' __ ~ 86

Ascorbate/cytochrome c

_

purified oxidase in the presence of 0.1 % laurylmaltoside

purified oxidase in the presence of 0.1 Tween-80

_ 38

66

~

Ascorbate/N,N,N,N-tetramethyl-p-phenylenediamine

32

64

20

40 60 UOLUVE ( p l )

80

100

Fig. 5. Peak areas ( A ) of immunoprecipitates of cytochrome c oxidase activity and succinate dehydrogenase activity after immunoudsorption of immunoglobulins with various concentrations of membrane antigens. Absorption of cytochrome c oxidase was performed with 4.25 mg anti(membrane vesicle) immunoglobulins and 0- 90 pi of intact vesicle suspension (0) or membrane vesicles solubilized in 5 % Triton X-100 (M), both at 5 mg protein/ml. Absorption of succinate dehydrogenase was performed as described above with an intact vesicle suspension (0)or membrane vesicles solubilized in 5 %Triton X-I00 (0).The immunoprecipitates were detected by zymogram staining techniques as described. Peak areas are expressed as cmz, and plotted as the reciprocal

chrome c. The lowest molecular weight component ( M , 13000) could effectively be removed with an Amberlite CG-50 column, and therefore, could be ascribed to cytochrome c. The three remaining bands are polypeptide subunits of cytochrome c oxidase. Values for the apparent molecular weights of subunits I, I1 and I11 were 57000, 37000 and 21 000, respectively (Fig. 4).

Enzymut ic Properties

In native gelelectrophoresis of purified cytochrome c oxidase one single band stained was present after staining with Coomassie brilliant blue. This band also showed zymogram staining for N,N,N,N'-tetramethyl-p-phenylenediamineoxidase activity, as well as for heme. This protein showed again a three subunits pattern on sodium dodecylsulfate-polyacrylamidegels. A 'fuzzy' appearance of the largest molecular weight component was also observed here. Cyanide blocked cytochrome c oxidase activity of the purified enzyme at very low concentrations (Iso FZ 1 IM), whereas much higher concentrations of azide were necessary to block this activity (I,o z 80 pM). It is of interest that 10-fold higher concentrations of both inhibitors are required to reach the same level of inhibition of cytochrome c oxidase activity in

47

cytoplasmic membranes. Due to the ionic character of interaction of cytochrome r oxidase and its substrate, cytochrome c, low concentrations of monovalent salts inhibit cytochrome c oxidase activity (Iso = 20 mM KCI). High cytochrome c oxidase activity could be obtained at relative high p H (7.5-8.5) B. subtilis W23 cytochrome c oxidase has enzymatic properties which are comparable to those of other bacterial cytochrome c oxidases (Table 1) [8, 91. The enzyme activity varied dramatically with the assay conditions. Remarkable is the ascorbate/N,N,N,N-tetramethyl-p-phenylenediamineoxidase activity of the purified enzyme in the absence of cytochrome c. Apparently N,NJ"'N-tetramethyl-p-phenylenediamine can donate its electron directly to the cytochrome c oxidase without interference of cytochrome c. This N,N,N',N'tetramethyl-p-phenylenediamine-oxidaseactivity was sensitive to cyanide, but insensitive to salt. Purified B. subtilis cytochrome c oxidase possesses a high affinity ( K , = 2pM) for S. cerevisiae cytochrome c. A similar observation is made for horse heart cytochrome c. The V,,, values of the purified enzyme in the presence of 0.05% laurylmaltoside in the spectroscopic assay varied from 4- 5 rnol cytochrome c x sx rnol oxidase-'. Comparison of the spectroscopic assay and the polarographic assay shows an enormous discrepancy between the maximum activities of cytochrome c oxidase.

'

Role of Phospholipids

The activity of the purified enzyme is strongly dependent on the presence of negatively charged phospholipids. In the presence of asolectin the maximum activity of the purified enzyme in the spectroscopic assay varied from 5 - 7 rnol cytochroine c x s - ' x mol oxidase-I. When phosphatidylserine was used a maximum activity of 3 - 4 rnol cytochrome c s - l x mol oxidase-' could be obtained. Localization in the Cytoplasmic Membrane

Crossed immunoelectrophoresis of the purified cytochrome c oxidase against anti-(B. subtilis W23 membrane vesicle)

immunoglobulins showed one precipitation line after staining with Coomassie brilliant blue. This immunoprecipitate also stained for N,N,N',N-tetramethyl-p-phenylenediamine oxidase activity and heme (Fig. 6A). Quantitative information about the localization of cytochrome c oxidase in the cytoplasmic membrane can be obtained using a technique described by Owen and Kaback [34]. The accessibility of an enzyme, known to be located at the inner surface of the cytoplasmic membrane, succinate dehydrogenase, is compared with the accessibility of

102

Fig. 6. Crossed immunoelectrophoresispattern ofpurified cytochrome c oxiduse against anti-(membrane vesicleJ immunoglobulins.(A) In the first dimension 3 pg purified oxidase was used. In the second dimension 4.1 mg anti-vesicle immunoglobulins were incorporated in the gel. After electrophoresis the immunoplate was stained with Coomassie brillant blue. With the oxidase zymogram staining technique the same precipitation line was found. (B, C) Localization of cytochrome c oxidase immunoprecipitate in total pattern of B. subtilis W23 membrane proteins. In the first dimension 40 pg protein of a 5 "/o (w/v) Triton X-100 membrane extract was used. In the second dimension 5.1 mg anti-(membrane vesicle) immunoglobulins were incorporated in the agarose gel (B). Immunoglobulins obtained after absorption of 3 pg purified cytochrome c oxidase with 5.1 mg anti-(membrane vesicle) immunoglobulins were incorporated in the second dimension as described in Materidis and Methods. (C) Arrows indicate the cytochrome c oxidase immunoprecipitate

the cytochrome c oxidase [I 51. Linear relationships between I / A and V(volume) were found when 4.25 mg anti-(membrane vesicle) immunoglobulins were titrated with 0 - 0.45 mg membrane vesicle proteins (Fig. 5). The fractions accessible from the outer surface of the vesicles were calculated to be 18% for succinate dehydrogenase and 65 % for cytochrome c oxidase. From other studies it was already known that at least 80 % of the cytoplasmic membranes had the right-side-out orientation [15]. The accessibility of the cytochrome c oxidase, therefore, strongly suggests that cytochrome c oxidase possesses antigenic sites at both sides of the membrane. With purified oxidase the precipitate of cytochrome c oxidase could be identified in the total pattern of membrane bound proteins (Fig. 6B, C).

DISCUSSION Bacillus subtilis W23 cytochrome c oxidase can be purified to homogeneity with cytochrome c affinity chromatography. It appeared to be essential to solubilize cytochrome c oxidase with alkylglucosides, such as laurylmaltoside or octylglucoside. With other non-ionic detergents the binding capacity of the affinity column decreased drastically. A disadvantage of laurylmaltoside is the low critical micelle concentration and the large micellar weight. This property makes an exchange of this detergent with other detergents or phospholipids difficult [22]. Cytochrome c forms reversible associations with both its reductase and oxidase, mediated by complementary charge interactions. The cytochrome c affinity column can also be used for the purification of B. subtilis W23 cytochrome c reductase, heme b. Aspecific binding of B. suhtilis membrane proteins to the column material is limited, possible as a result of their basic nature. The binding of the oxidase and reductase to cytochrome c can be disturbed by high ionic strength. Another way to release the binding is elution with cytochrome c [35]. A disadvantage of the affinity column is the low binding capacity. The spectral properties of purified aa,-type oxidase in eukaryotes and prokaryotes are very similar but the structural properties are distinctly different [8, 361. Eukaryotic oxidases possess at least seven subunits [36]. B. subtilis W23 cytochrome c oxidase contains only three subunits. The heme a to protein ratio and the sum of the apparent molecular weights of the subunits indicates the presence of one of each subunit per molecule. Therrnus thermophilus HB8 is the only other bacterium of which a three-subunit composition has been shown. This oxidase contains, just as most thermophilic bacteria,

heme a and heme c as prosthetic groups (9). A simple polypeptide composition seems to be a general property of bacterial cytochrome c oxidases. From an evolutionary point of view it is of interest that B. subtilis oxidase contains three subunits. The three largest subunits of the yeast enzyme are synthesized by mitochondria1 ribosomes, and coded by the mitochondrial genome [37]. Immunological studies showed that the largest molecular weight subunits of the yeast enzyme are closely related to bacterial cytochrome c oxidase subunits. In general the molecular activity of the bacterial oxidases is lower than that of the eukaryotic oxidases [8, 361. Two explanations can be advanced for this phenomenon: (i) for a maximal rate of electron transfer the corresponding combination of cytochrome c and cytochrome c oxidase should be used; (ii) the intramolecular electron transfer rate is slower in bacterial cytochrome c oxidase. There is an enormous difference in enzyme activity depending on the assay used. In the polarographic assay the highest turnover rates are obtained at low ionic strength and relatively high pH, thus favouring the tight binding of cytochrome c to the oxidase and permitting rapid reduction of the cytochrome c-cytochrome c oxidase complex by ascorbate or possibly N,N,N',N'-tetramethyl-pphenylenediamine. In the spectrophotoinetric assay in the absence of ascorbate or N,N,N',N'-tetramethyl-p-phenylenediamine, dissociation is required of cytochrome c from the oxidase and lower rates of oxidation of cytochrome c are measured. The detergent, in which the enzyme is dispersed, influences strongly the state of aggregation and the functional and structural properties of the enzyme. Therefore it seems likely that the altered activities observed in some detergents, might simply reflect less effective dispersion of the enzyme. In contrast to eukaryotic oxidases, where high activities can be obtained in the presence of alkylglucosides, the B. subtilis oxidase shows decreased activities upon addition of these detergents [22]. As shown in Results the purified cytochrome c oxidase has a high ascorbate - N,N,N',N'-tetramethyl-pphenylenediamine oxidase activity. This is also reported for the T. thevmophilus HB8 and Rhodopseudornonas .sphaeroides oxidases [9,35]. This N,N,N',N'-tetramethyl-p-phenylenediamine activity of the B. subtilis enzyme is, just as the Rps. sphueroides enzyme, insensitive to salt [35]. For reconstituted beef-heart oxidase, energy transduction, as well as the proton-translocating function has been established [I 1 - 131. Recently it was reported that Paracoccus denitrificans oxidase conserves also these functional properties [14]. These functions can be performed only if cytochrome c oxidase is a transmembranal

103

protein. Crossed immunoelectrophoresis experiments showed that B. .subrilis oxidase possesses antigenic sites located on the inner and outer side of the membrane, indicating its transmemb r a d localization. This study has been made possiblc by financial support from the Dutch Organization of Pure Scientific Research (Z.W.0.); the grant no. 3 .739-0.80 I-om Sch IW izcvischr N o rionu/fiinds, a grant lro m Sii n d o z St i Itung 10 A.A. and a short-term fellowship to W.d.V. from thc Europcan Molecular Biology Organization (t.M.13.0.) for visiting the laboratory of A.A. We are thankful to Dr S. Salardi for the synthesis oflanrylmaltosidc.

REFERENCES 1. Lcmberg, M. R. (1969) f'hysiol. Rev. 49, 48-121. 2. Hinkle, P. C., Kim, J . J. & Racker, E. (1972) J . Bid. Chem. 247, 13381339. 3. Yamanaka, T. & Fukomori, Y . (1977) FEBS Let/. 77, 155-158. 4. Smith, L. (1978) Methods Enzymol. 53, 202-212. 5. Haddock, B.A. & Jones, C. W . (1977) Bacteriol. Rev. 41, 47-99. 6. Tochikubo, K. (1971) J . Bacteriol. IOS, 652-661. 7. Chaix, P. & Petit, J . F. (1956) Biochim. Biophys. Acru, 22, 66-71, 8. Ludwig, B. (1 980) Biochim. 5iophy.r. Ac/a, 5Y4, 177 - 189. 9. Hon-nami, I(. & Oshima, T. (1980) Biochem. Biophys. Res. Comniun. 92, 1023-1029. 10. Mitchell, 1'. & Moyle, 1. (1970) in Electron Trunsport and Energy Conservution (Tager, J. M., Papa. S., Quagliariello, E. & Slatcr, E. C., eds), pp. 575- 587, Adriatica Edictrice, Bari. 11. Wikstrom, M. & Krab, K. (1979) Biodiini. Biophys. Acta, 549, 177-222. 12. Sigel, F. & Carafoli, F. (1979) J . Biol. Chem. 239, 586-590. 13. Casey, R. P., Chappell, J. B. & Azzi, A. (1979) Biochem. J . 182, 149-156. 14. Solioz, M., Carafoli, E. & Ludwig, B. (1982) J . Bid. Chem. 257, 15791582. 15. Bergsma, J., Strijker, R., Alkema, J . Y . k,., Seyen. H. G. & Konings. W . N . (1981) Eur. J . Biochem. 120, 599-606.

16. Bisschop, A. & Konings, W . N . (1976) Eur. J . Biocl7em. 67, 357-365. 17. Bisschop, A., Bergsma, J. & Konings, W. N . (1979) Eur. J . Biochem. 93, 369 - 374. 18. Bergsma. J., Meihuizen, K. E., van Oeveren, W. & Konings, W.N. (1982) Eur. J . Biocl7em. 125, 651 -657. 19. Bergsma. J., van Ilongen, M. B. M. & Konings, W. N. (1 982) Eur. J . Bioehcn7. 128, 151 - 157. 20. Hartzell, R. C., Beinert, H., van Gelder, B. F. & King, T. E. (1978) Mi,rh(id.s Enzymol. 53, 54- 66. 21. Bill, K., Casey, R. P., Broger, C. & Azzi, A. (1980) FEBS Lett. 120, 248 - 250. 22. Rosevear, P., Vanaken, T., Baxter, J. & Ferguson-Miller, S. (1980) Biochefnisiry 19, 4108-4115. 23. Wang. C . & Smith, R. L. (1975) Anul. Biochem. 63, 414-417. 24. Lowry, 0.H., Rosebrough, N. J., Farr, A. L & Randall, R. J. (1951) J . Bid. Chem. 193, 265 - 275. 25. Yonetani, T. (1959) J . Biochem. ( T o k y o ) 46, 917. 26. Chance, B. & Williams, G. R. (1955) J . Bicit. CI7em. 2 i 7 , 395. 27. Laemmli, U.K. (1970) Nature (Land.) 277, 680-685. 28. Pentilli. T., Saraste, M. & Wikstrom M . (1979) FEBSLeti. 101,295300. 29. Broger, C., Allemann, P. & Azzi, A. (1979) J . Appl. Biochem. I , 455 459. 30. Yonetani, T. (1965) J . B i d . Chem. 236, 1680- 1688. 31. Smyth, C. J., Sigel, J. Salton, M. R. J. &Owen, P. (1978) J . Bueteriol. 133, 306- 319. 32. Owen, P. & Salton, M . R. J. (1975) P r o , . N u / / Arad. S(.i, U S A , 72, 3711-3715. 33. Burstone, M.S. (1959) J . Histochem. Cytochem. 7, 112-122. 34. Owen, P. & Kaback, H. R. (1978) Proc. Null Acud. Sci. U S A , 75, 3148- 31 52. 35. Gennis, R. B., Casey, R. P., Azzi, A., & Ludwig, B. (1982) Eur. J . Biochem. 125, 189 - 195. 36. Azzi, A. (1980) Biochit??. Biophys. A d a , 594, 231 -252. 37. Capaldi, R. A , , Bell. R. L. & Brunchek, T. (1977) Biochem. Bioph,vs. Rcs. C'oinmun. 74. 425- 433. -

W. de Vrij and W. N . Konings Laboratorium voor Microbiologie, Rijksuniversiteit (te) Groningen, Biologisch Centrum, Kerklaan 30, NL-9751 -NN Haren, The Netherlands A. Azzi Medizinisch-Chemisches lnstitut der Universitit Bern, Postfxh, CH-3000 Bern 9, Switzerland