Streptomyces reticuli produces a heme-containing homodimeric enzyme (160 kDa), the catalase-peroxidase CpeB, which is processed to the enzyme CpeC during prolonged growth. CpeC contains four subunits of 60 kDa each that do not include the C-terminal portion of the progenitor subunits. A genetically engineered cpeB gene encodes a truncated subunit lacking 195 of the C-terminal amino acids; four of these subunits assemble to form the enzyme CpeD. Heme binds most strongly in CpeB, least in CpeD. The catalase-peroxidase CpeB and its apo-form (obtained after extraction of heme) catalyze the peroxidation of Mn(II) to Mn(III), independent of the presence or absence of the heme inhibitor KCN. CpeC and CpeD, in contrast, do not exhibit manganese-peroxidase activity.
The data show for the first time that a bacterial catalase-peroxidase has a heme-independent manganese-peroxidase activity, which depends on the presence of the C-terminal domain.
Catalases promote the disportionation of hydrogen peroxide, contain heme as a prosthetic group and are homotetramers, the subunits of which are about 60 kDa in size but in a few cases are larger . Some identified catalases, however, lack a heme group and therefore are not inhibited by cyanide or azide ions. These pseudo-catalases have been found in Thermus thermophilus, Thermoleophilum album and Lactobacillus plantarum and contain a binuclear manganese binding site .
Peroxidases use H2O2 to oxidize a number of compounds . Bifunctional catalase-peroxidases comprise varying ratios of two enzymatic activities. They are mostly dimeric or tetrameric enzymes, but in rare cases are monomeric . The known and studied bacterial catalase-peroxidases usually contain 726–753 amino acids per subunit and consist of two highly homologous halves, each of which shares significant amino acid identity with eukaryotic monomeric peroxidases, including cytochrome c peroxidase (CCP), but not with typical catalases . Using comparative alignments of catalase-peroxidases and the crystal structure of the yeast CCP as a model , the following predictions have been made: the N-terminal half of the bacterial enzymes can, contrary to the C-terminal one, bind one heme group. The key residues of CCP, which play a part in catalysis, folding and structural stability, are either invariant or highly conserved in bacterial catalase-peroxidases .
Heme-containing fungal lignin peroxidases and manganese peroxidases (MnPs) are only distantly related to CCP. From the crystal structure of MnP , it was shown that three acidic amino acids and the heme proprionate are involved in Mn2+ binding. Spectroscopical studies suggest that a binding site for Mn2+ within the catalase-peroxidase KatG from Mycobacterium smegmatis corresponds to that of fungal MnP .
Streptomyces reticuli produces a mycelia-associated, dimeric, heme-containing enzyme that exhibits catalase and peroxidase activity with broad substrate specificity and was thus named CpeB. The gene product (740 amino acids) from the S. reticuli cpeB gene  shares a high number of identical amino acids with several other bacterial catalase-peroxidases: KatG from Caulobacter crescentus, KatG Mycobacterium tuberculosis[13,14], KatP from Escherichia coli, and PerA from Bacillus stearothermophilus.
In this report we reveal that the S. reticuli CpeB enzyme exhibits, in addition to catalase and peroxidase activities, heme-independent manganese-peroxidase activity, and that the presence of the C-terminal domain is required for this activity. It is thus shown for the first time that a catalase-peroxidase has a heme-independent manganese-peroxidase activity.
Materials and methods
Bacterial strains and plasmids
S. reticuli Tü45 (H. Zähner, Institute of Microbiologie, Tübingen, Germany) and S. lividans 66 (D. A. Hopwood, John Innes Institute, Norwich, UK) were used. The plasmids pWHM3  and pUC18/pUC19  were gifts from C. R. Hutchinson (School of Pharmacy, Madison, USA) and J. Messing (Waksmen Institute, Rutgers, USA), respectively.
Media and culture conditions
For cultivation of the Streptomyces strains, complete and minimal media were used . Minimal media were supplemented with glucose or Avicel (1%). Depending on the purpose of the experiments, cultures were grown in baffled Erlenmeyer flasks containing 5–200 mL on a rotary shaker for 2–8 days. E. coli strains (DH5α or XL1-blue) were cultivated in Luria–Bertani medium at 37 °C .
Chemicals and enzymes
Chemicals for SDS gel electrophoresis were obtained from Serva. Molecular mass markers, monochlorodimedon, o-dianisidine, 4-chloro-1-naphthol, and 2,2′-azino-bis (3-ethylbenzthiazoline-6-sulfonic acid) were supplied by Sigma. Columns for FPLC were purchased from Pharmacia. Hydrogen peroxide (30% w/v) was bought from Merck.
Isolation of total DNA and plasmids
The E. coli plasmid pUC18 was isolated using the alkaline lysis method , and the multicopy Streptomyces/E. coli vector pWHM3 was obtained by a modified alkaline lysis method [19,20].
Cleavage of DNA, ligation and agarose gel electrophoresis
DNA was cleaved with various restriction enzymes according to the suppliers’ instructions. Ligation was performed with T4 ligase . Gel electrophoresis was carried out in 0.8–1% agarose gels using Tris/borate/EDTA buffer. Fragments were visualized under UV after staining with ethidium bromide .
E. coli was transformed with plasmid DNA using the CaCl2 method , or by electroporation . S. lividans 66 protoplasts were transformed and regenerated as described . Transformants were selected using an overlay of 0.4% agarose containing 500 µg·mL−1 thiostrepton .
The protein solution was obtained by washing the mycelia with 0.1% Triton X-100, as described , and then chromatographed. Anion exchange chromatography was performed on a DEAE column (Pharmacia HR 16/60) previously equilibrated in 20 mm sodium acetate buffer, pH 5.5, and eluted with an NaCl gradient (0–0.5 m in 80 mL) at a flow rate of 2 mL·min−1. MonoQ chromatography (Pharmacia HR 5/5) was carried out in a 30 mL gradient (0–0.5 m NaCl) at a flow rate of 1 mL·min−1 using 20 mm sodium acetate buffer, pH 5.5. Hydrophobic chromatography (Pharmacia HR 5/5) was carried out using 20 mm sodium phosphate buffer, pH 7.0, with an (NH4)2SO4 gradient (1.7–0 m in 35 mL).
Determination of N-terminal amino acids
N-terminal amino acids from purified proteins were determined by Edman degradation (Wüster).
Native and SDS/PAGE
For electrophoresis of the native enzyme, 10% (w/w) of a polyacrylamide gel, pH 7.5, was used . SDS/PAGE was performed in the presence of 0.1% SDS .
Detection of catalase activity
The native PA gel (10%) was washed three times for 15 min with distilled H2O, suspended in a solution of 0.01 mL 30% H2O2 in 100 mL H2O, and gently rocked for 10 min. The H2O2 was aspirated and the gel quickly rinsed in H2O. A freshly prepared mixture of 30 mL each of 2% ferric chloride and 2% potassium ferricyanide, both in H2O, was poured into a fresh staining pan, and the rinsed gel transferred to the ferricyanide mixture . The gel tray was gently but steadily rocked by hand over a light box. As soon as a green colour began to appear in the gel itself, the ferricyanide mixture was rapidly aspirated and replaced with water. The gel was washed twice with water.
Catalase activity was determined spectrophotometrically by following the disappearance of H2O2 at 240 nm, taking E240 as 43.6 m−1. The reaction mixture contained 100 mm potassium phosphate buffer, pH 7.0, 20 mm H2O2 and the enzyme. The reaction was run for 2 min at room temperature, and the initial linear rate was used to calculate the activity.
Test for peroxidase activity
Samples were loaded onto a native 10% polyacrylamide gel. After the run, the gel was washed twice with acetate buffer (20 mm, pH 5.5), and activity staining was carried out with 4-chloro-1-naphthol (10 mg·100 mL−1) and 5 mm H2O2.
Alternatively, the peroxidase activity was monitored spectrophotometrically at room temperature in a reaction mixture containing 50 mm sodium acetate buffer (pH 5.5), 2 mm H2O2, and 1 mmo-dianisidine by following the rate of oxidation at 460 nm (ε460 = 11.3 × 103 mm·cm−1). One unit of peroxidase activity is defined as the disappearance of 1 µmol of substrate.
Measurement of MnP activity
MnP activity was estimated by the formation of the Mn3+ tartrate complex (ε238 = 6.5 mm−1·cm−1) during the oxidation of 0.1 mm MnSO4 in 0.1 m sodium tartrate buffer, pH 5.5, with 0.1 mm H2O2, or indirectly by the oxidation of 0.1 mmo-dianisidine in the presence of 1.0 mm H2O2, 0.1 mm Mn2+ and 10 mm azide in 100 mm of citrate/sodium phosphate buffer, pH 7.5 .
Extraction of heme from the native CpeB
CpeB was converted into apo-CpeB by extraction of heme with distilled methylethyl ketone at 0 °C, pH 2.0  and after vacuum evaporation of the solvent.
Preparation of antibodies against the C-terminal domain of CpeB
CpeD (400 µg) was coupled to 0.5 g of CNBr-activated Sepharose 4B, according to the manufacturer's instructions (Pharmacia), and placed onto a column (0.5 cm × 3 cm). After equilibration with 20 mm sodium phosphate buffer, pH 7.0, 2 mL of anti-CpeB serum was loaded onto the column, and the bound protein was eluted with 0.1 m glycine/HCl, pH 2.7. Antibodies specific to the C-terminal part of CpeB do not bind to the matrix coupled with CpeD and are thus found in the protein portion not binding to the column. The procedure was repeated six times until no more protein from the anti-CpeB serum bound to the column.
Characteristics of the enzyme CpeB
S. reticuli secretes the previously identified dimeric enzyme CpeB, a catalase-peroxidase consisting of two subunits of 82 kDa each, which has catalase and peroxidase activities with broad substrate specificity . The catalase and the peroxidase activities have a pH maximum of 7.0 and 4.5, respectively (Table 2).
The peroxidase activity could be inhibited by low concentrations of the three typical heme protein inhibitors, KCN, NaN3 and NH2OH. On native gels, peroxidase activity was detectable in the presence of 10 mm KCN; it increased upon addition of MnSO4(Fig. 1A). With o-dianisidine as electron donor, the peroxidase activity had a broad pH optimum (Fig. 1B). It showed highest activity at pH 5.5, and it still retained 41% of the maximum activity at pH 7.5. At this pH, the addition of KCN and NaN3 barely affected the activity level. Addition of 50 µm, 100 µm or 200 µm MnSO4, however, led to an increase of 51.9%, 58.6% or 23%, respectively. The total peroxidase activity could only be enhanced by adding MnSO4. Other ions reduced this activity to a slight (Ca2+, Mg2+, Ni2+) or moderate (Co2+, Zn2+, Cu2+) extent. The data suggest that the increase in peroxidase activity by Mn(II) can be attributed to heme-independent MnP activity. In the tartrate buffer, at pH 5.5, the enzyme could directly oxidize Mn(II) to form the Mn(III)-tartrate complex. Using this method, the MnP activity was determined to be 0.45 U·mg−1 protein. This value corresponds to the one obtained with the indirect method by the manganese-dependent oxidation of o-dianisidine (Fig. 1B). Further evidence for manganese binding to the catalase-peroxidase enzyme was supplied by spectroscopic investigation. The addition of Mn(II) to the enzyme solution could increase the intensity of the Soret band at 406 nm (Fig. 1C), which is similar to that of the Phanerochaete chrysosporium MnP .
The genetically well characterized Streptomyces strain S. lividans lacks the cpeB gene. When transformed with the plasmid pWKS30, it produces CpeB, the catalase and peroxidase  and MnP activities of which are similar to those of CpeB isolated from S. reticuli (data not shown).
Processing of the enzyme CpeB
In the course of prolonged growth of S. reticuli, the decrease of the mycelia-associated 82-kDa protein correlated with the appearance of one protein of distinct size (60 kDa) cross-reacting with antibodies raised against CpeB. It had a predominant catalase activity (see below) and was therefore named CpeC. Within native gels, the mobility of the native CpeC enzyme was reduced compared to the native dimeric CpeB enzyme (Fig. 2). After purification (see below), it was shown by gel filtration that CpeC assembles as a tetramer. CpeC bound, like CpeB , to the surface of the mycelia. It was purified as described in Materials and methods. Highly purified CpeC (2.1 mg) was obtained from mycelia grown in 7 L (Table 1), and a portion was subjected to Edman degradation. The experimentally determined N-terminal sequence (TENHDA) corresponded to that previously determined for the CpeB protein . These data suggest that the purified protein was derived from CpeB by proteolytic processing, leading to a truncation of its C-terminal part.
Table 1. Purification of CpeC and CpeD. During the purification procedure, catalase and peroxidase activities were scored. Here the values for specific catalase activity are given.
Total protein (mg)
Total activity (U)
Specific activity (U·mg−1)
Wash fraction of mycelia (obtained from a 7 l-culture)
Wash fraction of mycelia (obtained from a 2 l-culture)
Mono Q chromatography
Processing of CpeB occurred in a similar fashion in S. lividans pWKS30 carrying the cloned S. reticuli cpeB gene (see above).
Characteristics of CpeC
The molecular mass of the native CpeC ascertained by gel filtration (data not shown) was 240 kDa; the denatured form was 60 kDa (in SDS/PAGE, Figs 2, 3). Native CpeC is consequently a homotetramer. Spectroscopic studies showed that native CpeC and CpeB have Soret bands at 406 nm. In the presence of cyanide, these are shifted to 421 nm (Fig. 4). Like CpeB, CpeC is a heme-containing enzyme and the spectra of CpeB and CpeC differed only slightly at longer wavelength (450–650 nm). The specific catalase and peroxidase activities of CpeC are about seven times higher and 130 times lower, respectively, than the ones of CpeB (Table 2). Unlike CpeB, CpeC lacks MnP activity, but similar to CpeB it is not inhibited by 3-amino-1,2,4-triazole. The characteristics of CpeC isolated from the original S. reticuli host or from the S. lividans pWKS30 strain were identical (see above).
Table 2. Comparison of CpeB, CpeC and CpeD. The measurements are described in Materials and methods. The MnP activity ascertained in test 1 is due to the direct oxidation of Mn(II), then the Mn3+ tartrate complex was analyzed. MnP test 2 indicates the use of the indirect method for determining the enzyme activity with o-dianisidine as electron donor. No activity (–), not tested (NT).
Molecular mass (kDa)
+ 10 mm KCN
+ 1.0 mm DTT
The ratio of A406 : A280
Specific activity (U·mg−1)
MnP test 1
MnP test 2
9.62 × 10−6
2.8 × 10−6
6.25 × 10−6
8.44 × 10−5
6.2 × 10−7
6.47 × 10−5
9.02 × 10−5
5.1 × 10−7
3.45 × 10−6
5.0 × 10−7
2.9 × 10−6
5.50 × 10−5
3.3 × 10−7
3.1 × 10−6
5.48 × 10−6
1.24 × 10−5
1.8 × 10−5
5.67 × 10−5
Generation and cloning of a truncated cpeB gene
Considering the size of CpeC, the data suggest that this truncated form differs from CpeB by 22 kDa, corresponding to 195 amino acids of the C-terminal region of CpeB . A precisely tailored gene was constructed to obtain a protein whose exact number of amino acids could be predicted. Therefore an experiment was designed to generate a truncated cpeB gene matching in size to the proteolytically processed form CpeB. Using the construct pUKS30  as a template, and the primers 5′-ACTGAGCTCAGATCTCGAAGCCGGCTTCCTTGG-3′ and 5′-TCACTCGAGGACTGGCCAGGCGAGAGCCGTC-3′, both of which contain an introduced SstI site, a 6.69 kb fragment including the vector pUC18 (Fig. 5) was generated by PCR. After its restriction by SstI and ligation, it was transformed into E. coli DH5α. Having propagated the transformants, several were found to contain the correct construct pUD7. The insert of pUD7 (4.06 kb of an EcoRI/HindIII-fragment) was cloned into the bifunctional Streptomyces/E. coli vector pWHM3  containing a thiostrepton resistance gene (for selection in Streptomyces) and the ampicillin resistance gene (for selection in E. coli). E. coli transformants that carried the resulting pWD7 construct were obtained. Sequencing of the insert showed that two original codons determining Ala95 (GCG) and Thr276 (ACC) were changed to the codons GCC and ACT, respectively, although both of these still encoded the original amino acids and did not lead to a change in protein.
Production and purification of CpeD
S. lividans was transformed with the plasmid pWD7 (Fig. 5). Thiostrepton-resistant S. lividans pWD7 colonies were shown to harbour the correct plasmid, including the designed truncated cpeB gene. Unlike the host S. lividans, the S. lividans pWD7 transformants showed mycelia-associated catalase peroxidase activity (Table 1). The CpeD enzyme was purified in the same fashion as CpeC (Table 1). As the cloned gene was present on a high-copy plasmid, about 1 mg of pure CpeD was obtained from a 2 L culture of the transformant. After SDS/PAGE, a protein-containing band corresponding to about 60 kDa cross-reacted with antibodies previously raised against CpeB .
Comparative characteristics of the enzymes
The molecular masses of the native (240 kDa, determined by gel filtration) and denatured (60 kDa on SDS/PAGE) forms of CpeD corresponded to those of CpeC. CpeD displayed highest peroxidase activity in sodium acetate buffer and less than 40% of this activity in the other buffers. The catalase activity of CpeD was, like that of CpeC, barely affected by the buffer composition (data not shown). Also like CpeC, but in contrast to CpeB, its catalase activity was independent of pH in the range 5.5–9.5. The peroxidase activity of CpeD has a narrow optimum at pH 5.0 (Fig. 6) and is 0.12 U·mg−1; the catalase activity of the purified enzyme is 240 U mg−1 so both activities are much lower than those of CpeC (Table 2).
Contrary to CpeB and CpeC, even solutions containing high concentrations (up to 500 µg·mL−1) of CpeD lack a pronounced ferric high-spin spectrum (Fig. 4).
Considering also that the catalase and peroxidase activities of CpeD are considerably low compared to CpeB and are also affected by KCN, NaN3 and NH2OH, the results suggest that only a small portion of the purified CpeD enzyme contains a heme group. Monitoring for the presence of the heme group (using A406/A280 ratios, data not shown) and the enzymatic activity (Table 1) revealed that a large part of it was already lost during the first purification step of DEAE chromatography (as compared to CpeC, Table 1). Consequently, the enzymatic activity of CpeD decreased rapidly during the course of purification.
The addition of heme to purified CpeD led to a significant increase in both catalase and peroxidase activities (Table 3). As a control, the heme group was released from native CpeB (see Materials and methods), to give apo-CpeB, which lacks catalase and peroxidase activities (Table 3). In the presence of KCN, catalase and peroxidase activities of CpeC and CpeD were not ascertainable; the same had been found for CpeB (Tables 2,3). The data clearly show that the catalase and peroxidase activities of all three enzymes are heme-dependent.
Table 3. Comparison of enzyme activities.
CpeB + 1 mm KCN
CpeC + 1 mm KCN
CpeD + 1 mm KCN
CpeD + hemin
Role of the C-terminal domain from CpeB
In a previous study, antibodies that were raised against CpeB (82 kDa)  also cross-reacted with CpeD (Fig. 3). CpeD was used to titrate antibodies that still recognize CpeB, but not CpeD out of the pool of polyclonal anti-CpeB antibodies (see Materials and methods). These selected antibodies did not cross-react with CpeC and were thus specific to the 22-kDa C-terminal domain (Fig. 3). They were designated as anti-(C-terminus) antibodies. The results proved that a domain corresponding to the C-terminal region (22 kDa) of CpeB is missing both in CpeC and CpeB.
CpeB exhibited MnP activity in the presence and absence of KCN, but neither CpeC nor CpeD shared this characteristic. In the presence or absence of additional heme, CpeD, like CpeC, could not oxidize Mn2+ (Table 3). The apo-form of CpeB, however, catalyzed the oxidation of Mn2+. Taken together, the data clearly demonstrate that MnP activity is independent of the heme group, but requires the C-terminal domain of CpeB.
S. reticuli and S. lividans pWKS30 carrying the cloned S. reticuli cpeB gene produce a mycelia-associated, heme-containing catalase-peroxidase (CpeB) with broad substrate specificity  that catalyzes the peroxidation of Mn(II) to Mn(III). This oxidation reaction has been best elucidated for the heme-containing MnP from the fungus Phanerochaete chrysosporium. This 46-kDa MnP is a heme-containing glycoprotein that catalyzes the H2O2-dependent and Mn2+-dependent oxidation of phenols and other substrates. On the basis of biochemical studies and the known crystal structure of this fungal MnP, it could be shown that Mn(II) binds to two glutamates (Glu35 and Glu39), one aspartate residue (Asp179), one heme proprionate and two water molecules . The wild-type yeast CCP does not bind Mn(II). Using site-directed mutagenesis, three codons were exchanged within the ccp gene, leading to the formation of a triple mutant of CCP (Gly41→Glu, Val45→Glu, His181→Asp), which had a functional Mn(II) site closely related that of the fungal MnP. The S. reticuli CpeB enzyme contains amino acid residues in positions that differ from those within the N-terminal part of the mutant CCP (see above). CpeB oxidizes Mn(II) equally well in the presence or absence of CN; the apo-form of CpeB (lacking the heme group) catalyzes the same reaction. The data suggest that the heme group of CpeB is not essential for Mn(II) binding or in Mn(II)/(III) peroxidation, as proven for the fungal MnP. The influence of Mn2+ on the Soret band 406 nm of CpeB (Fig. 1) indicates an increase of six coordinate high spin; therefore, the heme must be located close to the Mn(II) binding site. Recent titration experiments with Mn(II) have revealed that a change of intensity of the Soret band peak of a solution containing the catalase-peroxidase KatG from M. smegmatis and a reduction of Mn(III) by isoniazide can be monitored . Up to now, however, it has not been determined whether the KatG enzyme from M. smegmatis or other mycobacteria has a heme-independent MnP activity resembling that of S. reticuli CpeB.
Physiological, biochemical and immunological studies performed during growth of S. reticuli or S. lividans pWKS30 showed that the CpeB enzyme is processed to a truncated protein (CpeC) lacking about 22 kDa of the C-terminal part. CpeC retains heme as a prosthetic group and this is required for its catalase and peroxidase activity. Unlike CpeB, CpeC does not catalyze the oxidation of Mn(II) to Mn(III).
The genetically engineered cpeB gene induced in S. lividans pWD7 the formation of a truncated protein (CpeD) that lacks 195 C-terminal amino acids of CpeB. Most of the CpeD molecules lost the heme group during the purification procedure. Consequently, the catalase and peroxidase activities dependent on the presence of the prosthetic heme group, as well as the ratio of these two functions, were reduced considerably compared to wild-type CpeB. In this context it is interesting that the heme content of various bacterial dimeric and tetrameric catalase-peroxidases ranges from 0.54 to 0.49 per subunit of about 80 kDa. The tetrameric catalase-peroxidase from Rhodopseudomonas capsulata, however, contains only 0.395 hemes per 60 kDa subunit . Data obtained through analysis of the cpeA gene from Rhodobacter capsulatus indicated that the CpeA subunit has a molecular mass of 60 kDa and that its amino acid composition corresponds to that of CpeD; it also lacks a portion corresponding to the C-terminal region in CpeB. Addition of the heme group (extracted from the wild-type S. reticuli CpeB) to CpeD caused a considerable increase in its heme-dependent catalase and peroxidase activities (Table 3). The data indicate that lack of the C-terminal part leads to a differently folded protein, in which the amino acids involved in heme binding are either deeply buried and thus not accessible, or are oriented in such a way that a firm attachment of the heme group does not occur. Thus it can be assumed that the C-terminal part of CpeB is essential for folding of the protein in a fashion such that heme can be firmly integrated. The assembly of heme in cytochrome b6 in Chlamydomonas reinhardtii requires a transacting protein , the role of which could correspond to that of the C-terminus of CpeB.
Like CpeC, but contrary to CpeB, CpeD is devoid of MnP activity. These findings, together with the spectroscopic data (Fig. 4, see also above), indicate that the heme-independent MnP activity requires the presence of the C-terminal region, possibly to stabilize a configuration of CpeB that permits Mn(II) interaction.
Various deduced mycobacterial (M. smegmatis, M. tuberculosis) KatG proteins (Fig. 7) share a high number of identical amino acids with the S. reticuli CpeB enzyme. In this context it is interesting that neither CpeB nor KatG (from M. tuberculosis and M. smegmatis) contain closely spaced glutamic acid residues (corresponding to E35 and E39 in MnP, see above) within their N-terminal regions. This supports our finding that the N-terminal part of CpeB is not involved in Mn(II)-dependent catalysis. Several other closely spaced glutamic acid residues (Fig. 7) are found scattered within CpeB, including its C-terminal part. Whether they play a part in the newly identified heme-independent MnP activity remains to be determined.
Some bacterial enzymes that do not contain heme still exhibit manganese catalase activity. The 266 amino acids deduced from the gene encoding one subunit of the Lactobacillus plantarum manganese catalase lack a motif that can be detected by sequence alignments within the E. coli manganese superoxide dismutase , other manganese-containing proteins or the S. reticuli CpeB. The low-resolution structure of the T. thermophilus manganese catalase  suggested the presence of four antiparallel α helices for two metal ions; corresponding computer-predicted helices are absent in CpeB. Mitochondrial cytochrome oxidases generally contain a binding site for the nonredox-active manganese ion at the interphase of subunits I and II. In the cytochrome oxidase of Paracoccus denitrificans, manganese occupies this site and acts as a reporter group for a redox-linked conformational change. Two acidic amino acids (Glu and Asp) of subunit II play a part in Mn2+ binding . From the proteins analyzed so far, the amino acids Asp, Glu, His and Lys are possible ligands for manganese. It will be interesting to gain more detailed insights into the binding sites for manganese that are required for heme-independent manganese-peroxidase activity of the S. reticuli enzyme CpeB.
We thank M. Lemme for supporting the writing of the manuscript. The work was financed by the Sonderforschungsbereich (SFB 171/C14).