Bacillus subtilis expresses two kinds of haem-A-containing terminal oxidases

Eur. J. Biochem. 197,699-705 (1991) 0FEBS 1991 0014295691002866

Bacillus subtilis expresses two kinds of haem-A-containing terminal oxidases Marko LAURAEUS, Tuomas HALTIA, Matti SARASTE and Mirten WIKSTROM Helsinki Bioenergetics Group, Department of Medical Chemistry, University of Helsinki, Finland (Received October 29, 1990)

-

EJB 90 1276

The expression of two different aa3-type cytochrome oxidases is demonstrated in Bacillus subtilis. One of them (denoted caa3-605), was predicted by DNA-sequencing of Bacillus cytochrome oxidase genes, but has not been found previously. It contains covalently bound haem C in subunit I1 and is very similar to the enzyme previously described in the thermophilic bacterium PS3. The other oxidase (denoted aa3-600) deviates from most known oxidases of aa3 type, and is probably identical with the oxidase described by de Vrij et al. [de Vrij, W., Azzi, A. & Konings, W. N. (1983) Eur. J. Biochem. 131, 97-1031. It shows no immunological cross-reactivity to the PS3 enzyme and differs from this spectroscopically; it contains no CuA and does not oxidise cytochrome c despite of its haem-A chromophores. It catalyses oxidation of quinols, which is proposed to be its physiological function.

Various bacterial terminal oxidases have been purified and characterized [l]. The best known are the aa3-type cytochrome-c oxidases having two haem-A units (haem a and a 3 )and two redox-active copper ions (CuAand CuB), and the cytochrome-bo-type quinol oxidases having two b-type haems, but only one redox-active copper. Cytochrome bo from Escherichia coli is structurally strongly related to the aa3-type cytochrome-c oxidases [2,3], but it lacks the CuAcentre typical of the latter [4-61. These oxidases probably bind both haems and one copper (Cu,) to the largest subunit, whilst CuA is probably bound to subunit I1 in the aa3-type oxidases [7, 81. One main difference between aa3-type cytochrome oxidases described so far is that in some thermophilic bacteria (many of which are Bacilli) cytochrome-c is fused to the subunit-I1 polypeptide of the enzyme. This has been thought to be linked to the thermal resistance of these bacteria [9-111. An aa3-type cytochrome oxidase has been previously purified to homogeneity from B. subtilis. This enzyme was reported to have low cytochrome-c oxidation activity and a peculiar blue-shifted CI band when compared with other wellcharacterized cytochrome-c oxidases [12]. It did not have covalently bound cytochrome c. In contrast to this, recent DNA sequence analyses of the B. subtilis cytochrome-c-oxidase genes indicated strong similarity to its counterpart from the thermophilic Bacillus PS3 [13] and suggested that a cytochrome-c domain may be fused with subunit 11. In this study we report that B. subtilis expresses two different aa3-type oxidases, the relative abundance of which varies with different culture conditions. One of these (caa3-605) resembles the cytochrome-c oxidases from PS3 [9,14] and Bacillus stearothermophilus [l 11, having cytochrome c fused to subCorrespondence 20 M. Wikstrom, Department of Medical Chemistry, University of Helsinki, Siltavuorenpenger 10 A, SF-00170 Helsinki, Finland Abbreviations. HOQNO, hydroxyquinoline-N-oxide; TMPD, tetramethyl-p-phenylenediamine. Enzyme. Cytochrome c : oxygen oxidoreductase (EC 1.9.3.1).

unit I1 and a normal CuA metal center. The other oxidase (aa3-600) exhibits little or no cytochrome-c oxidase activity, catalyzes oxidation of quinols and does not have the CuAEPR signal. It thus somewhat resembles cytochrome o, but has haem A chromophores. MATERIALS AND METHODS Media Glucose medium: 25 g/1 glucose, 30 g/1 potassium monohydrogen phosphate, 10 g/l ammonium sulphate, 1 g/1 sodium citrate, 2.5 g/1 tryptone, 2.5 g/1 yeast extract, 1 ml/l salt solution, pH 7.5 [15]. Succinate medium: 20 g/1 succinic acid, 8 g/1 ammonium sulphate, 20 g/1 potassium monohydrogen phosphate, 1 g/l sodium citrate, 0,5 g/l tryptone, 1 ml/l salt solution, pH 7.5 ~51. Growth conditions B. subtilis strain 168 was used in all experiments. Large scale cultures were grown aerobically at 35°C using a 400-1 fermenter and a 5-1 starter culture. The bacterial mass was harvested at the stationary growth phase when grown on glucose. When the bacteria were grown in succinate medium particular attention had to be exercised to maximize aeration of the culture. In this case the bacterial mass was harvested before reaching the stationary growth phase. Preparation of membranes and solubilization The bacterial mass was suspended in a small volume of 50 mM Tris/HCl pH 7.4, containing 50 mM NaCl and stored frozen at - 20 "C. After thawing, 0.25 mM phenylmethylsulfonyl fluoride and 0.1 mM benzamidine were added to inhibit proteolysis, and 5 mg/ml lysozyme was added. This mixture was kept at 35 "C for 1 - 1.5 h and gently stirred frequently. Ten volumes of cold water was then added to cause

700 osmotic lysis. 0.2 mg/l DNAase I (Sigma) was added, and the suspension homogenised using an ultra-turrax mixer. The mixture was kept at 35"C, usually for 15 min to minimize viscosity caused by released DNA. The membranes were pelleted by centrifugation for 20 min at 21 000 rpm (Sorvall SS34 rotor or equivalent). Membranes from glucose-grown cells were mixed with 150 mM NaCl, 50 mM Tris/_HCl, pH 8.0, at 20 mg protein/ ml and 3% (mass/vol.) Triton X-100 was added. Insoluble material was removed by centrifugation at about I00000 x g for 30 min using a Ti-60 rotor. Membranes from succinate-grown cells were mixed with 1 M NaCl, 50 mM Tris/HCl, pH 8.0, containing 3% (mass/ vol.) sodium cholate, pH 7.6, at a protein concentration of 20 mg/ml. The cholate-treated membranes were pelleted by centrifugation as above and solubilized with 150 mM NaCl, 50 mM Tris/HCl, pH 8.0, containing 5% (massivol.) Triton X-100, at a protein concentration of 10 mg/ml. Insoluble material was removed as above. The extract was dialysed twice against 20 volumes of 50 mM Tris/HCl, pH 8.0, containing 0.5% (massivol.) Triton X-100. Purijication of aa3-600 A 2.6 x 27 cm DEAE-Sepharose CL-6B column (Pharmacia) was equilibrated with 50 mM Tris/HCl, pH 8.0, containing 150 mM NaCl and 0.2% (mass/vol.) Triton X-100 and membrane extract from glucose-grown cells was applied to the column at room temperature. The column was washed with equilibration buffer until no chromophores were eluted. Elution was then continued by increasing the salt concentration in 50-mM steps from 150mM NaCl. The chromatography was performed at room temperature at a flow rate of 2.5 ml/min. Green fractions eluted at 0.3 M NaCl, and were collected and concentrated using an Amicon ultrafiltration cell and YM-100 membranes. For fast-liquid chromatography, the concentrated enzyme preparation was diluted two times by 50 mM Tris/HCI, pH 8.0. It was applied to a Mono Q HR lOjl0 column (Pharmacia), which was equilibrated with 50 mM Tris/HCl, pH 7.4, containing 0.25% (mass/vol.) Triton X-100. Protein was eluted at room temperature at a flow rate of 1 ml/min with a linear gradient of 0.35 0.6 M NaCl in 50 mM Tris/HCl, pH 7.4, containing 0.25% (massivol.) Triton X-100. Green fractions were concentrated as before. Purfication of caa,-605 The solubilized membranes from succinate-grown cells were applied to a 3.1 x 35 cm column of DEAE-Sepharose CL-6B that had been equilibrated with 50 mM Tris/HCl, pH 8.0, containing 0.5% (massjvol.) Triton X-100, at 4°C. The column was washed with the same buffer, which additionally contained 25 mM NaCl, until no chromophores were eluted. The enzyme was then eluted by increasing the NaCl concentration to 50 mM at a flow rate of 2.5 ml/min. Redgreenish fractions were collected and applied to a 3.0 x 22 cm poly-L-lysine agarose column (Sigma) equilibrated with the same buffer used for the DEAE-column. Cytochrome caa3605 studied here did not bind to the column and the redgreenish fractions were collected and concentrated as above for aa3-600. Optical spectroscopy Oxidase samples were dissolved in 0.2 M Hepes (potassium salt), pH 7.4 and spectra were recorded using a Shimadzu UV-

3000 spectrophotometer at room temperature. Dithionitereduced enzyme was taken as the baseline for the CO difference spectrum of reduced aa3-600; then CO was bubbled through the cuvette for 2 min in the dark and the difference spectrum was recorded at room temperature. Haem determination was made using the pyridine haemochrome technique 1161. E P R spectroscopy Oxidase samples were dissolved in 0.1 M Hepes (potassium salt), pH 7.4 and were concentrated using Amicon Centricon 100 microconcentrators. The final concentrations in samples of Fig. 5 A and C were 81 pM for aa3-600 and 88 pM for aa3605, and their volume about 0.3 ml. EPR spectroscopy was performed with a Bruker ESP-300 X-band spectrometer equiped with an Oxford Instrument ESR-900 liquid-helium cryostat. The conditions were: modulation amplitude, 19.8 G ; modulation frequency, 100 kHz; scanning speed, 5.96 G/s, time constant, 41 ms; microwave frequency, 9.44 GHz; microwave power, 1 mW; temperature, 12 K. The copper spectrum in Fig. 5B was recorded at 20 K, modulation amplitude, 9.9 G ; scanning rate, 26.8 G/s; microwave power, 2 mW.

Electrophoresis, immunoblotting and haem-staining SDSjPAGE was carried out using the buffer system of Laemmli [ 171in 12- 22% gradient acrylamide gels containing 5 M urea. Electrotransfer of proteins to nitrocellulose filters was performed according to Towbin et al. [18] with the addition of 0.1 % SDS to the blotting buffer. The Trans-blot cell was placed in ice and the current was kept between 0.4 - 0.5 A for 4 h. After blotting, the filter was incubated in 150 mM NaCl, 10 mM Tris/HCl, 0.05% (massivol.) Triton X-100 in the presence of 1YO(mass/vol.) bovine serum albumin for 12 h at 4"C, then with antisera raised against the PS3 caa3 [19] and the B. subtilis aa3 [12] (the enzyme was a generous gift by Dr. de Vrij), diluted 1 :400 in the same medium. For colour detection we used anti-(rabbit IgG) antibodies coupled to alkaline phosphatase, supplied by Promega Biotech. Haem staining was carried out according to [20].

Polarographic activity measurements Oxygen consumption was measured using a Clark-type oxygen electrode at 25°C. The medium contained 50 or 100 mM monopotassium phosphate buffer, pH 7.4, 1 mM EDTA (sodium salt), and 20 mg/ml asolectin lipids (Sigma). Enzymatic oxygen consumption was determined by subtracting cyanide-insensitive activity. Duroquinone, menadione, docylbenzoquinone and 2-hydroxy-l,4-napthoquinone were reduced by borohydride as described by White et al. [21]. All quinones were dissolved in methanol as a 50 mM stock solution, except tetrachlorohydroquinone, which was dissolved in ethanol. In all measurements the amount of oxidase is calculated from reduced-minus-oxidized spectra, using specific absorptivities of 25.2 mM-' . cm-' for caa3-605 at 605-630 nm and 26.6 mM-' . cm-' for aa3-600 at 600630 nm. These were determined from pyridine hemochromeminus-hemichrome difference spectra using the recommended 25.02 mM-' . cm-' absorptivity at the 588-620 nm wavelength pair [16].

701

Fig. 1. SDSjPAGE analysis and haem staining of purified oxidases from B. subtilis. (A) Lane 1, caa3-605 purified from succinate-grown cells; lane 2, PS3 cytochrome-c oxidase (a generous gift of Dr N. Sone); lane 3, au3-600 purified from glucose-grown cells; lane 4, molecular mass markers. (B) The SDSjPAGE gel was haem stained as described in Material and Methods. The lanes are defined as in A

Fig. 2. Immunoblots of B. subtilis oxiduses. The SDSjPAGE gel seen in Fig. 1 was immunoblotted onto nitrocellulose filters as described in Materials and Methods. (A) The filter is stained using antisera raised against PS3 oxidase. Lane 1, caa3-605; lane 2, PS3 oxidase; lane 3, aa3-600. (B) Immunoblotting as in A, but staining using antiserum against cytochrome au3 purified by de Vrij et al. [I21

strongly to DEAE-Sepharose, which is therefore useful in separating the two oxidases from one another. Purification of aa3-600

Miscellaneous Protein concentrations were measured according to [22]. Triton X-100 was changed to 2% (mass/vol.) sodium cholate before SDS/PAGE, according to Penefsky [23].

RESULTS Purification of caa3-605 Treatment of the bacterial membranes with cholate at high ionic strength is an effective purification method that based on protein determinations extracted 80% of the membrane proteins. Especially cytochrome c and some b-type cytochromes were removed, which otherwise easily contaminate the final oxidase preparation. An alkaline treatment described by Berry and Trumpower [24] as well as a cholate/deoxycholate wash [15] were also tried, but the former method did not remove cytochrome c and the release of membranebound proteins was ineffective.The latter method also extracts caa3-605 from the membranes. A high-Triton-X-lOO/protein ratio was necessary for good purification. At a low-detergent/ protein ratio there is a tendency for caa3-605 to form a complex with a protein having a reduced-minus-oxidized absorbance maximum at 562 nm. At high concentrations Triton X-100 has an inhibitory effect on electron transfer and most of the caa3-605 is kept in a reduced form. However, the reduced enzyme was readily re-oxidizable in air-saturated media by lowering the Triton X-100 concentration, or by changing the detergent to dodecyl maltoside. The poor binding of caa3-605 to DEAE-Sepharose is apparently typical for caa3-type cytochrome-c oxidases [9, 111. After passing through the DEAE-Sepharose the caa3-605 was still contaminated by b-type cytochrome(s) and SDSjPAGE revealed several bands unrelated to caa3-605 (not shown). This necessitated further purification (see below). However, aa3-600, which is also present in succinate-grown cells, binds

Cytochrome aa3-600 was purified from glucose-grown cells, where it is abundant. We found no significant amounts of caa3-605 in such cells. This made is possible to simplify the isolation procedure by leaving out the cholate washing step. aa3-600 is fully oxidised even at a high-Triton-X-100 concentration and has a much higher affinity for DEAE-Sepharose than caa3-605. The DEAE chromatography was performed at room temperature because better separation between aa3600 and a b-type haemoprotein was achieved under such conditions. Mono-Q chromatography was used as a final step to remove a few contaminating polypeptides that were left after DEAE-Sepharose chromatography, as judged by SDSjPAGE (not shown). The main part of aa3-600 is eluted at 0.5 M NaCl as two peaks very close to one another. We have not detected any differences in activity, spectrum or polypeptide content between these peak fractions. Therefore, the two were routinely combined for further characterization. Polypeptide composition and immunological reactivity Fig. 1A compares the subunit composition of caa3-605 and aa3-600, as isolated here, with cytochrome-c oxidase isolated from PS3. Subunits I and I1 of caa3-605 (lane 1) and PS3 oxidase (lane 2) migrate similarly in the SDSjPAGE gel, indicating the same or similar molecular masses of these polypeptides. Subunits I11 and IVB are predicted by the gene structure of caa3-605 [13], but they are not present in our preparation of the enzyme. Subunit I11 is easily lost from oxidases of other organisms (see [25]). Subunit IVB was only recently found in an isolated oxidase preparation from PS3 [26]. Also the structurally related cytochrome o of E. coli has been isolated as a two-subunit and four-subunit enzyme [27, 281. Subunits I and I1 of aa3-600 (lane 3) migrate differently in comparison with both PS3 oxidase and caa3-605. A weak

702

Fig. 3 . Reduced-minus-oxidized spectra of B. subtilis oxiduses. (A) uu3-600 from glucose-grown cells. (B) cau3-605 from succinate-grown cells. (C) shows the CO-reduced-minus-reduced difference spectrum of ua3-600

band, possibly corresponding to subunit 111, is observed slightly above the position of the 20.1-kDa marker protein. Fig. 1B shows haem staining of the gel in Fig. 1A. Subunits I1 of the caa3-605 and PS3 oxidases are both stained, indicating that they contain covalently bound haem C. This suggests that in both enzymes subunit I1 is fused with a cytochrome-c domain, which increases the molecular mass of this subunit. No subunit of aa3-600 exhibits haem staining. Fig. 2A and B show immunoblots with antisera raised against oxidases isolated from PS3 by Sone and Yanagita [9] and from B. subtilis by de Vrij et al. [12], respectively. They show that subunit I1 of caa3-605 is recognised by the antiserum raised against the PS3 enzyme. Several proteins in the latter relatively impure preparation are recognized by the antiserum raised against it. However, this antiserum does not recognise any subunit of aa3-600 and only the subunits of aa3-600 are stained when antiserum raised against B. subtilis aa3 isolated by de Vrij is used (see Materials and Methods). These results suggest that subunits I1 from caa3-605 and PS3 oxidase are related. They share a common epitope, which is lacking from the aa3-600 enzyme. Optical spectra of isolated aa3-600 and caa3-605 Fig. 3 shows the reduced-minus-oxidised spectra of the isolated aa3-600 and caa3-605. caa3-605 exhibits an optical spectrum similar to that of cytochrome-c oxidases isolated from thermophilic bacteria. It has CI bands at 605 nm and 551 nm, the latter narrow band arising from the cytochrome c. The p band of the C-type haem is seen at 520 nm. In the Soret region there is the typical 445-nm band and in addition a 417-nm band from cytochrome c ; the Soretla band ratio is about 6: 1. The spectrum of aa3-600 is similar to that described by de Vrij et al. [12, 291. There is no cytochrome c band and the a band is blue-shifted to 600 nm as compared to most other bacterial aa3-type cytochrome-c oxidases [l]. In addition, there is a shoulder near 570 nm, which is atypical in comparison with other bacterial aa3-type oxidases. In the Soret region the spectrum resembles that of other cytochrome oxidases with a 444-nm maximum and a 415-nm through. The latter indicates that one of the two haemes is in a high-spin ferric state in the enzyme as isolated. The Soretla band ratio is high, of the order of 15 : 1. Fig. 1C shows the CO-difference

spectrum of the reduced cytochrome aa3-600. It is virtually identical to the corresponding well-known spectrum of mitochondrial cytochrome aa3 with the CI peak near 590 nm and a fi peak near 545 nm (see [12, 29, 301). Fig. 4 shows the pyridine haemochrome spectra of aa3600 and caa3-605. The 589-nm absorption maximum typical of haem A is found in both enzymes. The caa3-605 shows, in addition, a band at 550 nm, indicating the presence of haem C. Activity and electron donor specgicity Oxygen consumption activities of aa3-600 and caa3-605 are reported in Table 1. aa3-600 does not exhibit significant activity with tetramethyl-p-phenylenediamine (TMPD) or various cytochromes c as electron donors, indicating that it may not be a cytochrome-c oxidase. With quinol analogues, however, oxygen consumption activity was found, suggesting that this enzyme may be a quinol oxidase in vivo. The highest activities were reached with tetrachlorohydroquinol and duroquinol. The quinol substrates shown in Table 1 are all structurally related to benzoquinol. This might cause lowered activity because B. subtilis membranes contain only menaquinone-7 [31], which is a naphtoquinone-like molecule. Two tested naphtoquinols, menadiole and 2-hydroxy-l,4naphtoquinol, were both highly auto-oxidizable under our measuring conditions, which prevented their assessment as electron donors. The possibility that a quinol is the true electron donor to aa3-600 is supported by the hydroxyquinolineN-oxide (HOQNO) inhibition data (Table 1). HOQNO is a known inhibitor of quinol-oxidising enzymes such as the cytochrome-bc, complex, cytochrome d and cytochrome bo [32, 331. In contrast, em3-605 had no detectable activity with quinols as electron donors, but exhibited reasonably high activity with TMPD plus ascorbate, especially in the presence of cytochrome c from Candida krusei. Cytochrome c from horse heart or from Saccharomyces cerevisiae did not increase activity above that with TMPD. Cytochrome-c oxidase from the thermophilic bacillus PS3 [9] is also known to prefer cytochrome c from C. krusei. The activity was not increased further when Triton X-100 was changed to dodecyl maltoside; activity was abolished when the detergent was changed to sodium cholate.

703 Table 1. Oxygen consumption by cytochromes aa3-60Uand caa3-6U5 The polarographicassay is described in Materials and Methods. 5 mM ascorbate was present in all but the two last assays. Values shown within brackets are percentage inhibition caused by HOQNO. DTT, dithiotreitol; TCQHZ, tetrachlorobenzoquinol; DQH2, duroquinol; DBH2, decyl benzoquinol 1401; Cyt, cytochrome; UQ-1, ubiquinone 5 Conditions

- Oxygen consumption activity au3-600

CUU,-605

electrons/ua3/s

0.310 mM TMPD