Identification of a copper-repressible C-type heme protein of Methylococcus capsulatus (Bath) A member of a novel group of the bacterial di-heme cytochrome c peroxidase family of proteins Odd A. Karlsen 1 , Louise Kindingstad 1 , Solveig M. Angelska ˚ r 1 , Live J. Bruseth 1 , Daniel Straume 1 , Pa ˚ l Puntervoll 2 , Anne Fjellbirkeland 1 , Johan R. Lillehaug 1 and Harald B. Jensen 1 1 Department of Molecular Biology, University of Bergen, Norway 2 Computational Biology Unit, Bergen Centre for Computational Science, Norway Copper plays a very significant role in the physiology of the methanotrophic bacterium Methylococcus capsul- atus (Bath). The availability of this metal ion regulates expression of the two forms of the methane-oxidizing enzyme methane monooxygenase (MMO) the bacter- ium possesses and formation of an extensive intracyto- plasmic membrane network [1–4]. When copper is scarce, at a low copper-to-biomass ratio, a soluble cyto- plasmic MMO (sMMO) is responsible for the oxidation of methane. At high copper-to-biomass ratios the parti- culate MMO (pMMO) is expressed and there is no detectable sMMO expression. Furthermore, copper also influences the expression of at least two of the four M. capsulatus formaldehyde dehydrogenases [5–7]. Keywords cell surface exposed; copper regulated; cytochrome c peroxidase; methanotrophs; Methylococcus capsulatus Correspondence O. A. Karlsen, Department of Molecular Biology, University of Bergen, HIB, Thormøhlensgt. 55, 5020 Bergen, Norway Fax: +47 555 89683 Tel: +47 555 84372 E-mail: odd.karlsen@mbi.uib.no (Received 27 July 2005, revised 7 October 2005, accepted 17 October 2005) doi:10.1111/j.1742-4658.2005.05020.x Genomic sequencing of the methanotrophic bacterium, Methylococcus cap- sulatus (Bath), revealed an open reading frame ( MCA2590 ) immediately upstream of the previously described mopE gene (MCA2589). Sequence analyses of the deduced amino acid sequence demonstrated that the MCA2590-encoded protein shared significant, but restricted, sequence simi- larity to the bacterial di-heme cytochrome c peroxidase (BCCP) family of proteins. Two putative C-type heme-binding motifs were predicted, and confirmed by positive heme staining. Immunospecific recognition and bioti- nylation of whole cells combined with MS analyses confirmed expression of MCA2590 in M. capsulatus as a protein noncovalently associated with the cellular surface of the bacterium exposed to the cell exterior. Similar to MopE, expression of MCA2590 is regulated by the bioavailability of cop- per and is most abundant in M. capsulatus cultures grown under low cop- per conditions, thus indicating an important physiological role under these growth conditions. MCA2590 is distinguished from previously character- ized members of the BCCP family by containing a much longer primary sequence that generates an increased distance between the two heme-bind- ing motifs in its primary sequence. Furthermore, the surface localization of MCA2590 is in contrast to the periplasmic location of the reported BCCP members. Based on our experimental and bioinformatical analyses, we suggest that MCA2590 is a member of a novel group of bacterial di-heme cytochrome c peroxidases not previously characterized. Abbreviations BCCP, bacterial cytochrome c peroxidase; CCP, cytochrome c peroxidase; ECL, enhanced chemiluminescence; MADH, methylamine dehydrogenase; MALDI, matrix-assisted laser desorption ionization; MauG, methylamine-utilizing protein G; MeDH, methanol dehydrogenase; MMO, methane monooxygenase; Mop, M. capsulatus outer membrane protein; NMS, nitrate mineral salt; ORF, open reading frame; pMMO, particulate methane monooxygenase; SACCP, surface-associated cytochrome c peroxidase; sMMO, soluble methane monooxygenase. 6324 FEBS Journal 272 (2005) 6324–6335 ª 2005 The Authors Journal compilation ª 2005 FEBS We have previously described the surface-associated protein, MopE, from which an N-terminally truncated form, MopE*, is released into the culture medium [8,9]. Furthermore, it was recently shown that MopE responds to changes in copper concentration during growth, and was most abundant when copper was lim- ited, indicating an important physiological role at low copper-to-biomass ratios [10]. Genomic sequencing of M. capsulatus revealed an open reading frame (ORF, denoted MCA2590) immediately upstream of the mopE gene [11]. In this study, we show that the MCA2590- encoded protein is expressed by M. capsulatus, and, similarly to MopE, is located on the cellular surface of the bacterium. Furthermore, expression of MCA2590 also responded to the availability of copper, and was abundant when copper was scarce in the growth med- ium. The MCA2590 primary sequence shares signifi- cant but restricted sequence similarity to the bacterial di-heme cytochrome c peroxidase (BCCP) family of proteins and contains two conserved heme-binding motifs. We also demonstrate that the MCA2590 pro- tein contains C-type heme; this was in line with our predictions from the primary sequence. Extensive bio- informatical analyses suggest that MCA2590 is a member of a novel group of the BCCP family. Results Sequence analyses An ORF of 2322 nucleotides (MCA2590) was predic- ted upstream and in the same orientation as mopE in the M. capsulatus genome (Fig. 1A) [11]. Twenty-six nucleotides separated the predicted stop codon for the putative protein MCA2590 and the translation start site of MopE. Sequence analyses including the MCA2590, mopE and upstream nucleotides of MCA2590, revealed a candidate promoter region 5¢ of the potential start codon of the putative MCA2590 protein (Fig. 1B). In addition, a transcription termin- ation site consisting of a GC-rich palindrome sequence followed by an AT-dominant stretch of nucleotides was found downstream of the mopE gene [9]. The lack of significant predictions of promoter and terminator regions immediate 5¢ of mopE indicates an MCA2590⁄ mopE operon. Furthermore, two putative ribosomal binding sites could be predicted upstream of both the MCA2590 and the mopE gene (Fig. 1). Our attempts to detect transcripts overlapping the MCA2590 ⁄ mopE transition have so far been unsuccessful. The putative MCA2590-encoded protein consists of 773 amino acids (Fig. 1B). Bioinformatical analyses using signalp and psort predicted a leader peptide with a putative cleavage site between Ala41 and His42 (Fig. 1B). N-Terminal processing would lead to a mature protein of 732 amino acids with a theoretical molecular mass of 78 kDa. A search in the PROSITE database of protein families and domains [12] with MCA2590 revealed two regions matching the c-type cytochrome superfamily profiles (PS51007). Both regions contain the c-type cytochrome motif (CxxCH; Fig. 1B), suggesting that MCA2590 binds two hemes. Sequence similarity searches (BLASTp) revealed signi- ficant similarity to hypothetical proteins of Photobacte- rium profundum (UniProt: Q6LQ47), Pseudomonas fluorescens (GenBank: ZP_00262397.1), and Nostoc punctiforme (GenBank: ZP_00109402.2), with expect (E)-values of 1e )116 ,1e )98 and 1e )69 , respectively. However, neither confirmation of expression nor any functions have been assigned to these putative pro- teins. Also proteins with known functions were revealed in the BLASTp searches, e.g. the cyto- chrome c peroxidase (CCP) of Nitrosomonas europaea (UniProt: P55929, E ¼ 4e )5 ) and the methylamine util- ization protein MauG of Paracoccus denitrificans (UniProt: Q51658, E ¼ 4e )4 ). Both these proteins are members of the BCCP family of proteins [13] (pfam 03150), but are representatives of two functionally dis- tinct subsets of the BCCP family [14]. To further explore the relationship between MCA2590 and the CCP and MauG proteins, members of each group were collected by performing several BLASTp searches (see Experimental procedures). Pairwise comparisons revealed that the hypothetical MCA2590-related sequences from P. profundum, P. fluorescens and N. punctiforme are 40–44% identical to the MCA2590, in contrast to the CCP and MauG sequences that are < 30% identical. A multiple sequence alignment including the hypothetical MCA2590-related sequences, and the CCP and MauG sequences, was constructed (alignment of representative members from each group of sequences are shown in Fig. 2). Figure 2 shows that there are conserved segments throughout the alignment and that these segments coincide well with secondary structure elements obtained from the resolved structures of the Pseudomonas aeruginosa and N. europaea CCPs. Strikingly, residues of both the heme-binding sites (CxxCH), in addition to the amino acids coordinating the calcium ion present in the interface domain of the solved CCPs are positionally conserved in MCA2590. It is also evident that the MCA2590 and the hypothetical MCA2590-related sequences, owing to the near double length, introduce large gaps in the CCP and MauG sequences when sequence similarities are aligned. Importantly, almost all of these gaps were introduced between secondary O. A. Karlsen et al. Heme protein identification in M. capsulatus FEBS Journal 272 (2005) 6324–6335 ª 2005 The Authors Journal compilation ª 2005 FEBS 6325 structure elements, leaving b sheets and a-helical struc- tures intact. It is also interesting to notice that MCA2590 and the hypothetical MCA2590-like sequences share significant sequence similarity in some of the long segments that remained unaligned to the CCPs and MauGs, indicating a common secondary structure ⁄ fold for the hypothetical MCA2590-related sequences in these regions. To examine the similarity between the MCA2590 and the BCCP sequences more closely, the multiple sequence alignment (Fig. 2) was visualized in the context of the N. europaea CCP struc- ture (Fig. 3). This analysis showed that positions that are identical or display a high degree of similarity between the MCA2590 and the CCP sequences are located in the core of the CCP structure (Fig. 3A). Furthermore, all of the additional insertions of the extended MCA2590 sequence are introduced in loop regions on the surface of the structure, thereby not interrupting secondary structure elements in the N. europaea CCP (Fig. 3B). The MCA2590 insertions were mainly dispersed around the entire surface. Taken together, these observations strongly suggest that the MCA2590 and hypothetical MCA2590-related sequences are homologous to the BCCP family of pro- teins, and form a separate group with a similar fold and core structure. However, the members of this new A MCA2590 mopE B Fig. 1. Genomic orientation (A), nucleotide and amino acid sequence (B) of MCA2590. (A) MCA2590 is oriented in the M. capsula- tus genome immediate upstream of the mopE gene separated by 26 nucleotides. (B) Amino acids are indicated below the nucleo- tide sequence. The underlined promoter region was predicted using the Neural Net- work Promoter Prediction (NNPP) and was estimated to have a probability of 0.97. The previously predicted termination loop is indi- cated by black arrows [9]. The two putative ribosomal binding sites predicted upstream of MCA2590 and mopE, respectively, are enlarged in the nucleotide sequence. The signal peptide predicted by SIGNALP is col- oured red in the MCA2590 amino acid sequence. The two putative C-type heme- binding sites revealed by SCANPROSITE are shown in yellow. The amino acid sequence used for construction of the anti-MCA2590 serum is blue. The peptides of MCA2590 that were revealed in the MALDI MS ⁄ MS analysis (Table 1) are boxed. Heme protein identification in M. capsulatus O. A. Karlsen et al. 6326 FEBS Journal 272 (2005) 6324–6335 ª 2005 The Authors Journal compilation ª 2005 FEBS Fig. 2. Multiple sequence alignment of MCA2590 and selected members of the BCCP family of proteins. The alignment was constructed from five MCA2590-like sequences, 26 CCP sequences and seven MauG sequences as identified in BLASTp searches (see Experimental procedures for details). Secondary structures obtained from the resolved structures of the CCP proteins from N. europaea (CCP_Ne; PDB:1IQC) and P. aeruginosa (CCP_Pa; PDB:1EB7) are indicated above the alignment (a, a helix; b , b strand; g,3 10 -helix; T: turn). Only the region of the alignment fully covered by the N. europaea CCP sequence is shown. The four cysteine residues that covalently bind two hemes and the two histidines that functions as heme–iron ligands are marked with triangles. The calcium ligand residues are indicated with circles. Only selected members of CCPs and MauGs are shown (accession numbers are given: GB, Gen- Bank; UP, UniProt). CCP members (group 1): CCP_Ne: N. europaea (UniProt: P55929); CCP_Pa: P. aeruginosa (UniProt: P14532); Q60BW8_Mc: M. capsulatus (UniProt: Q60BW8). MauG members (group 2): MauG_Pd: P. denitrificans (UniProt: Q51658); MauG_Mf: M. flagellatum (UniProt: Q50426); MauG_Mm: M. methylotrophus (UniProt: Q50233); M. extorquens (UniProt: Q49128). Hypothetical MCA2590-like sequences (group 3): 109402_Np: N. punctiforme (GenBank: ZP_00109402); Q6LQ47_Pp: P. profundum (UniProt: Q6LQ47); 262397_Pf: P. fluores- cens (GenBank: ZP_00262397); MCA2590_Mc: M. capsulatus (UniProt: Q6LQ47); U74385_Ma: M. album (GB:U74385 – fragment: missing N- and C-termini are marked with *). Visualiza- tion was made using ESPRIPT: Positions that are identical between all groups are marked with black boxes and similar positions are marked with white boxes. Positions that are similar or identical within one group are shown in bold. The annotation is based on the full alignment. O. A. Karlsen et al. Heme protein identification in M. capsulatus FEBS Journal 272 (2005) 6324–6335 ª 2005 The Authors Journal compilation ª 2005 FEBS 6327 group will contain longer loops and possibly additional secondary structure elements outside the CCP-similar core. A similarity search against the translated GenBank nucleotide database (tBLASTn) revealed significant similarity between MCA2590 and an unannotated ORF of Methylomicrobium album (GenBank: U74385 – nucleotides 3279–4205; E-value 1e )67 ) [15]. Interest- ingly, this ORF is located immediately downstream of the corA gene, which encodes the only protein in the databases that shows significant sequence similarity to MopE [9,15]. The deduced amino acid sequence of the M. album U74385 ORF was added to the multiple sequence alignment (Fig. 2), revealing that it is 50% identical to the N-terminal half of the MCA2590 pro- tein. Because the fully sequenced U74385 ORF is not available in the databases, it remains to be elucidated if the sequence similarity of MCA2590 and the M. album U74385 ORF extends even further. Localization and identification of the mature MCA2590 protein Enriched fractions of M. capsulatus-soluble proteins, inner membrane proteins and outer membrane proteins were obtained as described previously [8]. The resulting fractions were assessed by SDS⁄ PAGE, demonstrating the presence of the large subunit of the methanol dehy- drogenase (MeDH) in the soluble fraction and the outer membrane proteins MopB and MopE in the A B Fig. 4. SDS ⁄ PAGE (A) and protein immunoblot analyses (B) of pro- teins obtained during the fractionation of M. capsulatus. Samples of each step during the fractionation procedure were collected and comparable samples (10 lL of each 1 mL fraction) were analysed. (A) Coomassie Brilliant Blue (CBB) R-250 stained 10% polyacryl- amide gel. (A, B) Lane 1, whole cells (W); lane 2, soluble fraction (S); lane 3, total membrane fraction (TM); lane 4, Triton X-100 soluble membranes (enriched inner membrane fraction, IM); lane 5, Triton X-100 insoluble membranes (enriched outer membrane fraction, OM). MopE in addition to the OmpA related MopB [38] are indicated. (B) Protein immunoblot of (A) using the anti- MCA2590 serum. Molecular mass markers are indicated to the left of both subfigures. Fig. 3. The sequence similarity between MCA2590 and the CCP proteins visualized on the N. europaea CCP structure. Bound ligands are shown in green; the heme groups are shown as sticks, and the Ca 2+ ion is shown as a sphere. (A) The surface of the structure is shown in transparent white. Residues that are identical between the MCA2590 and CCP sequences in the alignment are shown in red, and similar residues are shown from yellow to red with increasing similarity. (B) A cartoon representation of the structure. Positions in the structure where gaps were introduced in the alignment are marked with blue spheres. The N- and C-termini are marked. Heme protein identification in M. capsulatus O. A. Karlsen et al. 6328 FEBS Journal 272 (2005) 6324–6335 ª 2005 The Authors Journal compilation ª 2005 FEBS Triton X-100 insoluble fraction (enriched outer mem- brane fraction) (Fig. 4A). When protein immunoblots of the enriched fractions were probed with a constructed antibody to a short peptide derived from the MCA2590 sequence, an anti-MCA2590 immunoreactive polypep- tide was found to cofractionate with the outer mem- brane from cells grown at a low copper-to-biomass ratio (Fig. 4B). The polypeptide migrated with a relative molecular mass of  80 kDa, which is close to the mass predicted for the mature MCA2590 (78 kDa). To further establish the cellular localization of MCA2590, M. capsulatus whole cells were treated with a high ionic strength buffer to extract proteins associ- ated with the cell surface. Treatment of whole cells with high concentrations of NaCl disrupts electro- static ⁄ ionic bonds between biomembranes and pro- teins, thus leaving surface-associated polypeptides in the resulting supernatant upon centrifugation [16,17]. The NaCl extraction was performed in two steps with increasing NaCl concentrations. Analyses of the resulting fractions by SDS ⁄ PAGE displayed a complex crude cell-extract protein pattern in the untreated cells and in cells treated with 0.5 and 1 m NaCl (Fig. 5A, lanes 1, 4 and 6). Importantly, the large 60 kDa sub- unit of the MeDH was seen only in these fractions. MeDH is a periplasmic protein [18], hence indicating that the cells remained undisrupted during the extrac- tion procedure and thereby avoiding leakage of pro- teins from the periplasm. Protein immunoblot using antibodies raised against MCA2590 revealed this protein in the 0.5 m NaCl-extracted fraction (Fig. 5B, lane 3). This observation suggests a specific association of MCA2590 to the outer membrane and that it is exposed to the cell exterior. As expected and in line with previous reports, a similar extraction behavior was observed for MopE [9] (Fig. 5B). The integral outer membrane OmpA-related MopB protein was not extracted from the outer membrane by the NaCl treat- ment, demonstrating the selectivity of the extraction procedure (Fig. 5B). To ensure that the localization of MCA2590 is at the external surface and not at the periplasmic inter- face, we performed two additional experiments. M. capsulatus whole cells were immobilized on a nitro- cellulose membrane and treated with the anti- MCA2590 serum (Fig. 6A). The MCA2590-specific staining obtained is consistent with SACCP being sur- face exposed. As another demonstration of surface exposure, we labelled intact cells grown in flask cul- tures with biotin (Fig. 6B). Biotin is too large to penet- rate the outer membrane and will only label polypeptides that have a lysine-containing sequence accessible at the cell surface. The biotinylated cells AC B Fig. 5. SDS ⁄ PAGE and protein immunoblot analyses of protein frac- tions obtained during the NaCl extraction of surface proteins. (A) 10% polyacrylamide gel stained with CBB R-250. The periplasmic methanol dehydrogenase (MeDH) and the surface associated MopE are indicated. (A, B) Lane 1, cells resuspended in 5 mL buffer with low ionic strength (20 m M Tris ⁄ HCl) (10 lL sample applied); lane 2, 20 m M Tris ⁄ HCl wash (20 lL sample applied); lane 3, 0.5 M NaCl extract (20 lL sample applied); lane 4, whole cells treated with 0.5 M NaCl (10 lL sample applied); lane 5, 1 M NaCl extract (20 lL sample applied); lane 6, whole cells treated with 1 M NaCl (10 lL sample applied). (B) Protein immunoblots of (A) using anti- MCA2590, anti-MopE and anti-MopB sera, respectively. Molecular mass markers are indicated to the left of (A) and (B). (C) Concentra- ted (selective centrifugation, Amicon 10 kDa cut-off) 0.5 M NaCl extracted fraction separated by SDS ⁄ PAGE and stained with CBB R-250. The polypeptide migrating with a corresponding molecular mass to the anti-MCA2590 immunogenic band is indicated. A B Fig. 6. Dot-blot and biotinylation analyses of whole cells. (A) Dot- blot analyses of M. capsulatus cells (0.5 · 10 8 ) immobilized onto nitrocellulose membranes. (1) Cells treated with the anti-MCA2590 serum. (2) Negative control; cells treated exclusively with secon- dary HRP-conjugated antibody. (B) Protein immunoblott of biotin- labelled M. capsulatus cells. (1) Biotin-labelled proteins visualized by treatment with streptavidine biotinylated HRP complex. (2) The same protein immunoblot as (1) but treated with the anti-MCA2590 serum. O. A. Karlsen et al. Heme protein identification in M. capsulatus FEBS Journal 272 (2005) 6324–6335 ª 2005 The Authors Journal compilation ª 2005 FEBS 6329 were analyzed with protein immunoblotting using the anti-MCA2590 serum, followed by removal of bound antibodies and staining of all biotinylated proteins using streptavidin-biotinylated horseradish peroxidase complex on the very same membrane. By using mark- ers on the nitrocellulose membrane we observed an exact comigration of a biotinylated protein with the immunoreactive MCA2590 band, strongly suggesting that MCA2590 has been biotinylated. The major peri- plasmic protein, MeDH, was not labelled by biotin, demonstrating that only surface-exposed proteins were labelled during the procedure. A 0.5 m NaCl-extracted fraction of M. capsulatus whole cells was concentrated by selective centrifugation (Amicon, 10 kDa cut-off) and the polypeptide migra- ting with an apparent molecular mass of 80 kDa in the SDS ⁄ PAGE analysis of the concentrate was excised from the gel and subjected to MS analyses (Fig. 5C). MALDI MS ⁄ MS analyses revealed three sequenced peptides (Table 1), which identified the excised poly- peptide as the gene product of MCA2590. Based on these findings and our bioinformatical analyses, we have named the MCA2590-encoded protein-surface- associated cytochrome c peroxidase (SACCP). Expression The expression of MopE is influenced by the avail- ability of copper [10]. Thus, it was of great interest to explore if MCA2590 and mopE were concomit- antly expressed if organized in a single transcrip- tional unit. M. capsulatus was grown at high and low copper-to-biomass ratios in batch cultures (0, 0.8 and 5 lm copper included in the growth medium) and expression was analysed by protein immunoblots using the anti-MCA2590 serum (Fig. 7). The subunits of pMMO (Fig. 7A), as well as detection of sMMO activity [19], were used as markers for copper-to-bio- mass ratios in M. capsulatus cultures. The protein immunoblots revealed that the expression of SACCP was altered by the different copper concentrations, SACCP being most strongly expressed in cells grown in low copper media (Fig. 7B). Specific PCR prim- ers were designed for MCA2590, and differential expression in cells cultured under high- or low- copper conditions was verified by RT-PCR analyses (Fig. 7C). Detection of C-type heme Because of the sequence similarity of SACCP to mem- bers of the BCCP family of proteins and the predic- tion of heme-binding motifs in the primary sequence, it was of interest to assay for the presence of C-type heme in SACCP. However, because our SACCP pre- paration contains other proteins that may also express C-type heme-specific peroxidase activity, we designed an assay that would distinguish SACCP from other proteins with similar activity. NaCl-extracted proteins separated by SDS ⁄ PAGE were transferred to a nitro- cellulose membrane and directly stained for heme using enhanced chemiluminescence (ECL). This method takes advantage of the intrinsic peroxidase activity of the covalently bound heme group of dena- tured c-type cytochromes, and should, in principle, detect any hemeprotein that retains heme after such treatment as described above [20,21]. The activity measurement of the NaCl extract resulted in four Table 1. MALDI MS ⁄ MS analyses of anti-MCA2590 immunogenic band. The localization of the identified peptides in the MCA2590 amino acid sequence is shown in Fig. 1. m ⁄ z (measured) Peptide sequence 1453.80 NTPTVINAALFHR 1593.80 VRDEAAPFDHPALR 2348.10 SVSDADGDGLADDEFLEIPAVGR A B C Fig. 7. Copper regulation analyses of MCA2590 (SACCP). (A) SDS ⁄ PAGE analyses of M. capsulatus cells grown at 5 (lane 1), 0.8 (lane 2), and  0 l M copper included in the growth medium, respectively. The 10% polyacrylamide gel was stained with CBB R-250. The PmoA, PmoB and PmoC components of the particulate methane monooxygenase are indicated (observed in lane 1 and 2). sMMO activity in the cell cultures are shown by either (+) or (–) below the lanes. (B) Protein immunoblot of (A) using anti-MCA2590 and anti-MopE serum, respectively. Molecular mass markers are indicated to the left of both (A) and (B). (C) RT-PCR analyses of the MCA2590, mopE and mopB transcripts prepared from the same cultures as used in (A) and (B). Heme protein identification in M. capsulatus O. A. Karlsen et al. 6330 FEBS Journal 272 (2005) 6324–6335 ª 2005 The Authors Journal compilation ª 2005 FEBS distinct bands, all of which could be traced to corres- ponding bands on a parallel CBB-stained polyacryl- amide gel (Fig. 8). Most importantly, a distinct band of molecular mass corresponding to the migration of SACCP was identified, suggesting that SACCP contained bound heme. A protein immunoblot using antibodies directed against SACCP on the same nitro- cellulose membrane showed that the peroxidase-activ- ity-stained band and the SACCP immunogenic band colocalized (Fig. 8). Using MS, The two additional high molecular mass bands, which also stained positive for heme, were identified as the proteins encoded by the genes MCA0421 and MCA0423, annotated as a cytochrome c 553o family protein and cytochrome c 553o , respectively (data not shown) [11,22]. Discussion In this study, we show that the predicted ORF (MCA2590) upstream of the mopE gene in M. capsula- tus encodes a C-type heme protein (SACCP) whose expression is strongly increased in bacteria cultured at low-copper concentrations. The protein is located on the surface of the bacterium and is noncovalently attached to the outer membrane. The extracellular localization is in accordance with the prediction of a signal peptide in a primary translation product. Bioinformatical analyses revealed that SACCP shares characteristics with previously described mem- bers of the BCCP family of proteins by both having significant sequence similarity and containing two con- served C-type heme-binding motifs [13]. This family constitutes the bacterial di-heme cytochrome c peroxid- ases. Similar to other CCPs, they reduce hydrogen per- oxide to water using C-type heme as cofactor. The BCCP family also includes the MauG proteins, whose similarity to di-heme CCPs has previously been recog- nized [23,24]. However, the MauG subset is distinct from the CCPs, and MauGs are proposed to function in the oxidation of methylamine in facultative methylo- trophs [23,24]. Importantly, SACCP significantly differ from the BCCP family of proteins in its much longer amino acid sequence. The regions of similarities to the CCPs and MauGs are therefore restricted, and in mul- tiple alignments distributed as conserved segments throughout the SACCP amino acid sequence. As in the BCCPs, two putative heme-binding motifs were found conserved N- and C-terminally of the SACCP sequence, indicating the binding of heme-groups anal- ogous to the low- and high-potential heme present in both CCPs and in the MauG proteins. Furthermore, the regions in SACCP with highest sequence similarity to the BCCP family of proteins coincided nicely with the known secondary structure elements of both the P. aeruginosa and N. europaea CCPs [25], indicating a native fold of SACCP which resembles the structure of CCPs. This supposition was substantially supported by structural visualization of the multiple alignments between the SACCP sequence and its close homo- logues, and the BCCP sequences (Fig. 3). These analy- ses showed that the additional SACCP-specific amino acid sequences were all located on the protein surface, and in-loop regions not interrupting secondary struc- ture elements. Furthermore, the majority of the con- served residues were buried amino acids, thereby maintaining the structural integrity of an N. europaea CCP-similar fold of SACCP, and thus bringing the hemes in close proximity to each other in contrast to increased interheme distance found in the primary sequence. We also showed that SACCP stained posit- ively for peroxidase activity typical of the covalently bound heme groups of denatured c-type cytochromes [20,21]. The C-type heme-associated peroxidase activity is in accordance with the presence of heme-binding motifs in the primary sequence. Interestingly, in the course of this study, two other surface-associated proteins that also stained positive for C-type heme were identified. Both of these proteins have previously been described as multi-c-heme cytochromes (cyto- chrome c 553o and cytochrome c 553o family protein) of M. capsulatus, and the respective genes are clustered in the genome [11,22]. We have now provided novel evi- dence for these proteins being noncovalently attached to the cellular surface. It is interesting to notice that 57 putative c-type cytochrome proteins are annotated in the M. capsulatus genome, of which five are members AB C Fig. 8. Detection of C-type heme in MCA2590 (SACCP). (A–C) Lane 1, 0.5 M NaCl extract from M. capsulatus; lane 2, bovine serum albumin. (A) 10% CBB R-250 stained polyacrylamide gel obtained from SDS ⁄ PAGE analysis. (B) Samples corresponding to (A) trans- ferred to a nitrocellulose membrane and directly stained for C-type heme peroxidase activity. (C) Protein immunoblot of a membrane corresponding to (A) using the anti-MCA2590 serum. Molecular mass markers are indicated to the left of (A). O. A. Karlsen et al. Heme protein identification in M. capsulatus FEBS Journal 272 (2005) 6324–6335 ª 2005 The Authors Journal compilation ª 2005 FEBS 6331 of the cytochrome c 553o family and four are annotated as putative cytochrome c peroxidases [11]. In this study, three of these have been localized to the cellular surface. As shown previously for MopE [10], the expression of SACCP was found regulated by the copper concen- tration in the growth medium and was found most abundant in cells grown in copper-depleted medium. RT-PCR analyses revealed that this regulation takes place at the transcriptional level. The concomitant regulation of SACCP and MopE is also in accordance with the possibility that the MCA2590 and mopE genes constitute an operon, but evidence of this assumption has not yet been provided. Furthermore, a potential promoter was predicted with high significance both by BPROM and NNPP immediate 5¢ of MCA2590. Very interestingly, the following nucleotide sequence was identified within the predicted promoter region: 5¢-TTGAGN(5)ATCGA-3¢. This nucleotide sequence closely resembles the consensus binding site of the transcriptional factor Fnr, 5¢-TTGATN(4)ATCAA-3¢, introducing only two mismatched nucleotides. FNR- type regulators are known to regulate aerobic ⁄ anaer- obic-dependent gene expression in c-proteobacteria, and the gene product of the fnrA gene in Pseudomonas stutzeri controls the expression of cythochrome cbb 3 - type terminal oxidase, cytochrome c peroxidase and the oxygen-independent coproporphyrinogen III oxidase [26]. M. capsulatus harbours one copy of a Fnr-type transcriptional regulator (MCA2120) [11]. However, further studies are necessary to elucidate the regulatory mechanisms for MCA2590 and mopE. The biological function of SACCP remains to be elucidated. However, upregulation of SACCP in M. capsulatus when grown at a low copper-to-biomass ratio indicates an important physiological role under these growth conditions. Furthermore, the concomitant expression, in addition to both the genomic and cellu- lar colocalization, of SACCP and MopE, indicates that these proteins may cooperate and have linked func- tions. In general, CCPs are thought to play a protect- ive role in the periplasm by reducing peroxides generated in oxidative metabolism [27]. MauG seems to have a more specific function, as demonstrated in P. denitrificans, by being involved in the maturation of the tryptophan tryptophylquinone cofactor of methyl- amine dehydrogenase (MADH) [14,28]. This enzyme is responsible for the oxidative deamination of methyl- amine when facultative methylotrophs are grown on methylamine as the sole source for carbon and energy [24]. The strong indications of SACCP having a core structure that resembles the CCPs may point toward similar enzymatic mechanisms. However, the increased size and cellular localization of SACCP open the possi- bility of other physiological functions distinct from what has been described for CCPs and MauGs. In conclusion, we have described a novel C-type heme protein located to the cellular surface of the methanotrophic bacterium M. capsulatus. This protein shares characteristics with the members of the BCCP family, but separates itself from the described CCPs and MauGs in both amino acid sequence and cellular location. Significant sequence similarity to three hypo- thetical proteins found in the prokaryotes P. profun- dum, P. fluorescens and N. punctiforme, respectively, was observed, in addition to a partial sequenced ORF of the methanotroph, M. album. Based on the amino acid sequence, experimental and bioinformatical data, we propose that SACCP belongs to a novel class of the bacterial di-heme cytochrome c peroxidases. Inter- estingly, the M. album ORF was found in close prox- imity to the CorA encoding gene (corA) which shares sequence similarity to MopE of M. capsulatus.It remains to be seen if this colocalization is a unique feature of methanotrophs. However, the colocalization may indicate a linked and unique function of these proteins in bacteria possessing both genes. Experimental procedures Growth of Methylococcus capsulatus (Bath) M. capsulatus NCIMB 11132 was grown in batch cultures at 45 °C while shaking in an atmosphere of CH 4 ,CO 2 , and O 2 (45 : 10 : 45) in nitrate mineral salt (NMS) medium as described previously [29]. Most analyses were performed using cells grown at a low copper-to-biomass ratio in ‘copper-free’ medium (no copper added). The cultures were screened for soluble methane monooxygenase (sMMO) activity by the naphthalene assay described by Brusseau et al. [19] to ensure that a low copper-to-biomass ratio was achieved. When analysing differential expression of MCA2590, cells grown at high copper-to-biomass ratios were included in the experiment (0.8 and 5 l m copper in the growth medium). Batch cultures of M. capsulatus were grown to a cell density of  10 8 cellsÆmL )1 before harvesting. Fractionation of cells Cells were harvested by centrifugation at 5000 g for 10 min. Enriched fractions of the soluble proteins, inner membrane proteins and outer membrane proteins were obtained as described by Fjellbirkeland et al. [8]. 20 mm Tris ⁄ HCl, 1 mm CaCl 2 , pH 7.4, was used as buffer throughout the fractionation procedure. Heme protein identification in M. capsulatus O. A. Karlsen et al. 6332 FEBS Journal 272 (2005) 6324–6335 ª 2005 The Authors Journal compilation ª 2005 FEBS Extraction of cell-surface proteins M. capsulatus cells were harvested by centrifugation at 5000 g for 10 min, and resuspended in a small volume of 20 mm Tris ⁄ HCl, 1 mm CaCl 2 , pH 7.4. The cell suspen- sion was centrifuged (as described above), and the super- natant was collected for SDS ⁄ PAGE analyses. Pelleted cells were treated with Tris ⁄ HCl buffer of high ionic strength (20 mm Tris ⁄ HCl, pH 7.4, 1 mm CaCl 2 , 0.5 m NaCl) to disrupt noncovalent bonds between surface-asso- ciated proteins and components of the outer membrane [16,17]. Cells were incubated with rotation in this buffer for 1–2 h at 4 °C. Cells were subsequently collected by centrifugation, and the resulting supernatant contained the 0.5 m NaCl-extracted proteins. Treated cells were re- suspended in 20 mm Tris ⁄ HCl, 1 mm CaCl 2 ,1m NaCl, pH 7.4, thereby increasing the ionic strength of the extraction buffer. Cells were thereafter incubated with rotation and harvested as outlined above, resulting in the 1 m NaCl extracted protein fraction. Cells were resus- pended in 20 mm Tris ⁄ HCl, 1 mm CaCl 2 , pH 7.4 after centrifugation. Biotinylation of cell-surface proteins Biotinylation of whole cells was performed with EZ-Link Sulfo-NHS-biotin according to the manufacturer’s instruc- tions (Pierce, Rockford, IL). Biotinylated proteins were visualized on nitrocellulose membranes by treatment with streptavidin-biotinylated horseradish peroxidase complex as described by the manufacturer (Amersham Biosciences, Little Chalfont, UK). SDS ⁄ PAGE and Western blotting SDS ⁄ PAGE was performed as described by Laemmli et al. [30] using 10% (w ⁄ v) running gels and 3% (w ⁄ v) stacking gels. Protein immunoblotting was carried out as described previously [8]. Rabbit polyclonal peptide-specific antibodies against MCA2590 were produced by Sigma Genosys. The immunogen correspond to the amino acids 742–755 of the MCA2590 primary sequence (Fig. 1B). The specificity of the produced antibody was confirmed by indirect ELISA. Horseradish peroxidase-linked anti-rabbit sera were pur- chased from Bio-Rad (Hercules, CA). Protein immunoblots that were incubated with anti-MopE or anti-MopB sera were developed using the colour reagent HRP CDR (Bio- Rad). Anti-MCA2590-treated membranes were developed using ECL (Amersham Biosciences). Detection of C-type heme Proteins separated by SDS ⁄ PAGE were transferred to a nitrocellulose membrane by electroblotting and directly stained for C-type heme with the ECL assay as described previously [20,21]. The ECL reagent was supplied from Amersham Biosciences. MS identification of proteins MS analyses were performed at the PROBE facility at the University of Bergen, Norway. Isolation of total RNA M. capsulatus flask cultures were harvested by centrifugation at 16 300 g for 1 min. Pellets were resuspended in Tris ⁄ EDTA buffer (0.1 mm Tris ⁄ HCl, 0.1 mm EDTA, pH 8.0) and submerged directly into liquid nitrogen and stored at )80 °C until used. RNA isolation and DNase treatment were carried out using the QIAGEN RNeasy Mini Kit as described in the supplied ‘RNAprotect Bacteria Reagent Handbook’ with an additional DNase treatment on the column. The RNA was dissolved in 50 lL of RNase-free water and the A 260 and A 280 values were determined. The RNA quality and quantity was determined with the Agilent 2100 Bioanalyser and the 2100 expert software. RT-PCR First-strand synthesis was carried out with 1 lg of total RNA. Superscript TM II RNase H – Reverse Transcriptase (Invitrogen, Carlsbad, CA) was used for first strand synthe- sis according to the manufacturer’s protocol. RNasin (Promega, Madison, WI) was used in the mixture to inhibit any RNAse activity. The cDNA synthesis was carried out using an elongation gradient ranging from 35 to 55 °C, where the temperature was raised 5 °C every 10 min. The primers used in the synthesis were 40 oligomers calculated to anneal to all ORFs in the M. capsulatus genome. The primers were defined using the program ‘Genome directed primers’ as described by Talaat et al. [31]. PCR was performed according to standard procedures [32]. The reverse transcription reaction (1 lL) was subjec- ted to PCR amplification using 1 U of DyNAzyme TM EXT DNA Polymerase (Finnzymes), in a 25-lL reaction volume with 0.2 mm dNTP, 0.5 lm of each specific primer, 1.5 mm MgCl 2 5% dimethylsulfoxide and the appropriate buffer. The ‘housekeeping’ genes encoding ribosome binding protein B, rplB, and M. capsulatus outer membrane protein B, mopB, were amplified. The amounts of PCR products were determined using the Fuji FLA-2000 phosphoimager, and the amount of template of first-strand cDNA was corrected to give an equal amount of amplified PCR prod- ucts for each ‘housekeeping’ gene. The differential expres- sion of the sapE and mopE transcripts was determined using the primers: sapE–F sapE–R and mopE–F mopE–R (Table 2). O. A. Karlsen et al. Heme protein identification in M. capsulatus FEBS Journal 272 (2005) 6324–6335 ª 2005 The Authors Journal compilation ª 2005 FEBS 6333 [...].. .Heme protein identification in M capsulatus O A Karlsen et al Table 2 Primers used in the RT-PCR analyses given in 5¢fi3¢ direction mopE–F 2589 F mopE–R 2589 R sapE–F 2590 F sapE–R 2590 R mopB–F 3103 F mopB–R 3103 R GGCAACGAGCAAGGTCCGAAG AAGTCGTTGCAATCGGCGTCG GGCAACGAGCAAGGTCCGAAG GACGTCGTGAGTGCCTCCGTG CGACGTGCAGTATTACTTTTCTAGGG AGTATCAAACCGTGCTGGTCTCC Bioinformatic analyses Sequence similarity searches... heme- containing proteins by chemiluminescence Anal Biochem 209, 219–223 21 Vargas C, McEwan AG & Downie JA (1993) Detection of c- type cytochromes using enhanced chemiluminescence Anal Biochem 209, 323–326 22 Bergmann DJ, Zahn JA & DiSpirito AA (1999) Highmolecular-mass multi -c -heme cytochromes from Methylococcus capsulatus Bath J Bacteriol 181, 991–997 23 Chistoserdov AY, Chistoserdova LV, McIntire... (2003) The surfaceassociated and secreted MopE protein of Methylococcus capsulatus (Bath) responds to changes in the concentration of copper in the growth medium Appl Environ Microbiol 69, 2386–2388 11 Ward N Larsen E Sakwa J Bruseth L Khouri H Durkin AS Dimitrov G Jiang L Scanlan D Kang KH et al (2004) Genomic insights into methanotrophy: the complete genome sequence of Methylococcus capsulatus (Bath). .. proteins of Methylococcus capsulatus (Bath) Arch Microbiol 168, 128–135 9 Fjellbirkeland A Kruger PG Bemanian V Hogh BT Murrell JC & Jensen HB (2001) The C- terminal part of the surface-associated protein MopE of the methanotroph Methylococcus capsulatus (Bath) is secreted into the growth medium Arch Microbiol 176, 197–203 10 Karlsen OA Berven FS Stafford GP Larsen O Murrell JC Jensen HB & Fjellbirkeland A. .. sequence alignment was subjected to visualization and analysis in the context of available structures using the ESPript service [35] Briefly, similarity scores for each position in the alignment between the four full-length MCA2590-like sequences and the 26 CCP sequences were calculated with ESPript using the Risler matrix, and these scores were used to replace the B factors of the N europaea CCP structure... Johnston SA (2000) Genomedirected primers for selective labeling of bacterial transcripts for DNA microarray analysis Nature Biotechnol 17, 679–682 32 Sambrook J, Fritsch EF & Maniatis T (1989) Molecular Cloning: A Laboratory Manual Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 33 Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W & Lipman DJ (1997) Gapped BLAST and PSI-BLAST: a new... Nielsen AK, Gerdes K, Degn H & Murrell JC (1996) Regulation of bacterial methane oxidation: transcription of the soluble methane mono-oxygenase operon of Methylococcus capsulatus (Bath) is repressed by copper ions Microbiology 142, 1289–1296 2 Nielsen AK, Gerdes K & Murrell JC (1997) Copperdependent reciprocal transcriptional regulation of methane monooxygenase genes in Methylococcus capsulatus and Methylosinus... from comparative proteome analysis Bioinformatics 21, 617–623 38 Fjellbirkeland A, Bemanian V, McDonald IR, Murrell JC & Jensen HB (2000) Molecular analysis of an outer membrane protein, MopB, of Methylococcus capsulatus (Bath) and structural comparisons with proteins of the OmpA family Arch Microbiol 173, 346–351 FEBS Journal 272 (2005) 6324–6335 ª 2005 The Authors Journal compilation ª 2005 FEBS 6335... Genetic organization of the mau gene cluster in Methylobacterium extorquens AM1: complete nucleotide sequence and generation and characteristics of mau mutants J Bacteriol 176, 4052–4065 24 van der Palen CJ, Slotboom DJ, Jongejan L, Reijnders WN, Harms N, Duine JA & van Spanning RJ (1995) Mutational analysis of mau genes involved in methylamine metabolism in Paracoccus denitrificans Eur J Biochem 230,... Ridout CJ, Greenwood C & Hajdu J (1995) Crystal structure of the di-haem cytochrome c peroxidase from Pseudomonas aeruginosa Structure 3, 1225– 1233 Heme protein identification in M capsulatus 26 Vollack KU, Hartig E, Korner H & Zumft WG (1999) Multiple transcription factors of the FNR family in denitrifying Pseudomonas stutzeri: characterization of four fnr-like genes, regulatory responses and cognate . GGCAACGAGCAAGGTCCGAAG mopE–R 2589 R AAGTCGTTGCAATCGGCGTCG sapE–F 2590 F GGCAACGAGCAAGGTCCGAAG sapE–R 2590 R GACGTCGTGAGTGCCTCCGTG mopB–F 3103 F CGACGTGCAGTATTACTTTTCTAGGG mopB–R. Identification of a copper-repressible C- type heme protein of Methylococcus capsulatus (Bath) A member of a novel group of the bacterial di -heme cytochrome