The Wolinella succinogenes mcc gene cluster encodes an unconventional respiratory sulphite reduction system


  • Melanie Kern,

    1. Institute of Microbiology and Genetics, Department of Biology, Technische Universität Darmstadt, Schnittspahnstr. 10, 64287 Darmstadt, Germany.
    Search for more papers by this author
  • Martin G. Klotz,

    1. Department of Biology, University of North Carolina, 9201 University City Boulevard, Charlotte, NC 28223, USA.
    Search for more papers by this author
  • Jörg Simon

    Corresponding author
    1. Institute of Microbiology and Genetics, Department of Biology, Technische Universität Darmstadt, Schnittspahnstr. 10, 64287 Darmstadt, Germany.
      E-mail; Tel. (+49) (0)6151 165203; Fax (+49) (0)6151 162956.
    Search for more papers by this author

E-mail; Tel. (+49) (0)6151 165203; Fax (+49) (0)6151 162956.


Assimilatory and dissimilatory sulphite reductions are key reactions in the biogeochemical sulphur cycle and several distinct sirohaem-containing sulphite reductases have been characterized. Here, we describe that the Epsilonproteobacterium Wolinella succinogenes is able to grow by sulphite respiration (yielding sulphide) with formate as electron donor. Sulphite is reduced by MccA, a prototypical member of an emerging new class of periplasmic cytochrome c sulphite reductases that, phylogenetically, belongs to a multihaem cytochrome c superfamily whose members play crucial roles in the global sulphur and nitrogen cycles. Within this family, MccA represents an unconventional octahaem cytochrome c containing a special haem c group that is bound via two cysteine residues arranged in a unique CX15CH haem c binding motif. The phenotypes of numerous W. succinogenes mutants producing MccA variants underlined the structural importance of this motif. Several open reading frames of the mcc gene cluster were individually inactivated and characterization of the corresponding mutants indicated that the predicted iron-sulphur protein MccC, the putative quinol dehydrogenase MccD (a member of the NrfD/PsrC family) as well as a peptidyl-prolyl cis-trans isomerase, MccB, are essential for sulphite respiration. MccA synthesis in W. succinogenes was found to be induced by sulphite (but not by thiosulphate or sulphide) and repressed in the presence of fumarate or nitrate. Based on the results, a sophisticated model of respiratory sulphite reduction by the Mcc system is presented.


Many microorganisms reduce sulphite in assimilatory and/or dissimilatory metabolism, thereby contributing to the biogeochemical sulphur cycle on Earth (Crane and Getzoff, 1996; Dahl and Friedrich, 2008). Typical cytoplasmic assimilatory and dissimilatory sulphite reductases (aSir/dSir) contain a coupled sirohaem-[4Fe-4S] cluster cofactor and catalyse the six-electron reduction of sulphite to sulphide. Dissimilatory sulphite reduction is a key step in sulphate reducing Bacteria and Archaea, for example in Desulfovibrio and Archaeoglobus species. The isolated dSir of these organisms is a DsrAB complex that reduces sulphite to sulphide and generates thiosulphate and trithionate as by-products. Several high-resolution structures of dSir enzymes have been determined, which helped to largely elucidate the reaction mechanism of sulphite reduction (Oliveira et al., 2008; 2011; Schiffer et al., 2008; Hsieh et al., 2010; Parey et al., 2010). In sulphate-reducing bacteria, electron transport from membranous quinol molecules to DsrAB most likely involves a membrane-bound DsrMKJOP complex and DsrC, which contains two conserved cysteine residues proposed to constitute a redox-active centre (Pereira et al., 2011). Non-DsrAB dissimilatory sulphite reductases (AsrABC from Salmonella enterica and Fsr from, for example, Methanocaldococcus jannaschii) have also been described (Huang and Barrett, 1990; Johnson and Mukhopadhyay, 2005). These enzymes also contain a sirohaem cofactor.

In addition to sirohaem-dependent enzymes, the respiratory cytochrome c nitrite reductase (NrfA), a periplasmic pentahaem cytochrome c, from several Gamma-, Delta- and Epsilonproteobacteria has been described to possess sulphite reductase activity although the physiological significance of this activity is not clear (Pereira et al., 1996; Stach et al., 2000; Clarke et al., 2006; Lukat et al., 2008; Kemp et al., 2010; Simon et al., 2011). NrfA catalyses the six-electron reduction of nitrite to ammonium, which is isoelectronic to sulphite reduction yielding sulphide. Electrons are transferred from the menaquinol pool to NrfA either via the membrane-bound tetrahaem cytochrome c NrfH (in Delta- and Epsilonproteobacteria) or via the NrfDCB proteins (in Gammaproteobacteria) (Simon et al., 2000; 2001; 2002; Gross et al., 2005; Clarke et al., 2007). Furthermore, the MccA (SirA) protein from the Gammaproteobacterium Shewanella oneidensis MR-1 was recently described as a sulphite-reducing enzyme (see Discussion) (Shirodkar et al., 2011).

The non-fermentative Epsilonproteobacterium Wolinella succinogenes is a metabolically versatile organism that is known to grow either by microaerobic or anaerobic respiration using fumarate, nitrate, nitrite, nitrous oxide, dimethyl sulphoxide or polysulphide (but not thiosulphate or sulphate) as possible terminal electron acceptors (Schumacher et al., 1992; Kröger et al., 2002; Simon, 2002; Klimmek et al., 2004; Simon et al., 2004; Kern and Simon, 2009a). Hydrogen gas, formate and sulphide were described as electron donor substrates (Macy et al., 1986; Kröger et al., 2002). Many of the corresponding electron transport chains have been investigated in detail and several quinone-reducing and quinol-oxidizing electron transfer modules have been identified that mediate electron transfer from primary dehydrogenases to terminal reductases (Simon et al., 2000; 2003; 2008; Gross et al., 2004; Kern and Simon, 2008; 2009a). These include many multihaem cytochromes b and c (some of which are quinone-reactive) as well as iron-sulphur proteins (Simon and Kern, 2008). W. succinogenes is genetically tractable and its genome sequence is known (Baar et al., 2003). More recently, the organism was also described to be suitable for heterologous production of complex metalloproteins (Mileni et al., 2006; Kern and Simon, 2011). The W. succinogenes genome neither encodes a dSir enzyme nor any other obvious sirohaem containing protein. Nonetheless, we describe in this study the capability of W. succinogenes to grow by sulphite respiration and characterize the corresponding electron transport chain that employs the periplasmic octahaem cytochrome c MccA as a terminal sulphite reductase. Previously, MccA was purified from overproducing W. succinogenes mutant cells and extensively characterized using a variety of biochemical and biophysical methods (Hartshorne et al., 2007). All the same, the function of the protein had not been elucidated. MccA represents an unusual cytochrome c as one of its haem c groups is attached via both cysteine residues of a CX15CH haem c binding motif whereas the other seven haem c groups are attached via conventional haem c binding motifs (CX2CH). The biogenesis of MccA was shown to require a dedicated cytochrome c haem synthase (CcsA1; also designated as cytochrome c haem lyase) encoded in the mcc gene cluster (Hartshorne et al., 2006; 2007; Kern et al., 2010a,b). Here, site-directed modification of MccA as well as the individual inactivation of other mcc genes was performed by constructing appropriate W. succinogenes mutant strains. The phenotypes of these cells led us to propose a model of the electron transport chain from the quinol pool to the active site of MccA. Given the analogy between the six-electron reduction of sulphite to sulphide and nitrite to ammonium, we also investigated the evolutionary relationship of MccA proteins to other multihaem cytochromes c that play important roles in the biogeochemical sulphur and nitrogen cycles. Overall, the results firmly characterized a novel cytochrome c sulphite reduction system found in many phylogenetically distant Proteobacteria that represents an alternative to sirohaem-dependent sulphite reduction.


W. succinogenes cells grow by sulphite respiration in the absence of other electron acceptors

In anoxic medium containing formate (100 mM) as electron donor substrate, succinate (5 mM) as carbon source, ammonium (10 mM) and sulphate (5 mM) as assimilatory nitrogen and sulphur source, respectively, cells of W. succinogenes did not grow unless sodium sulphite was added. In the presence of 10 mM sulphite, modest cell growth with a cell doubling time of ∼13 h was observed following a lag phase of about 1 day (Fig. 1A). Growth was accompanied by a pH change from 7.5 to ∼ 8.5 and stopped after the complete consumption of sulphite that was coupled to the oxidation of ∼30 mM formate (Fig. 1B). Both sulphide (∼7 mM) and thiosulphate (∼1 mM) were present in the aqueous phase of the culture in the stationary growth phase. Thiosulphate was probably produced in an abiotic reaction from sulphite and sulphide because a comparable thiosulphate concentration was also found in the absence of cells when sulphide was provided. Conversion of thiosulphate by W. succinogenes cells was not observed under any known growth condition. Initial sulphite concentrations lower than 10 mM correlated linearly with decreased final optical densities of the cultures (not shown). On the other hand, growth in the presence of sulphite concentrations above 10 mM proved to be increasingly detrimental to cell growth.

Figure 1.

Growth by sulphite respiration of W. succinogenes wild-type cells. A. Growth curves of W. succinogenes batch cultures using the medium described in Experimental procedures. Cells of the wild-type (filled squares) and the Δmcc::kan mutant (open squares) were incubated in the presence of 10 mM sodium sulphite. Wild-type cells did not grow when sulphite was left out of the medium (triangles). The medium was inoculated using fumarate-grown cells. B. Traces of formate (squares), sulphite (triangles), sulphide (circles) and thiosulphate (diamonds) concentration in medium of W. succinogenes wild-type cells grown in the presence of sulphite (as shown in A).

Notably, sulphite was not converted in the presence of either nitrate (5 mM) or fumarate (20 mM), which are both potent alternative electron acceptors of anaerobic respiration. However, when these substrates had been fully consumed (yielding either ammonium or succinate), sulphite reduction started in the culture after a lag phase in the range of 23–26 h (Figs S1 and S2).

Specific rates of formate and sulphite consumption were measured using wild-type cell suspensions (Table 1). The results confirmed a formate/sulphite turnover ratio of three indicating that sulphite was completely reduced in a six-electron step to hydrogen sulphide [which equilibrates with sulphydryl ions, pKa (H2S/HS-) = 7.0]. Therefore, it is suggested that growth was driven by a proton motive force (pmf)-generating formate-dependent sulphite reduction according to the following reaction: 3 HCO2- + SO32− +  4H+ → 3 CO2 + HS- + 3 H2O, ΔG0′ = − 57 kJ mol−1 formate.

Table 1.  Specific rates of formate, sulphite and sulphide turnover in cell suspensions of W. succinogenes strains.
StrainTerminal acceptor of anaerobic respirationFormate consumption (nmol min−1 mg protein−1)Sulphite consumption (nmol min−1 mg protein−1)Sulphide production (nmol min−1 mg protein−1)
  1. Mean values and standard deviations from at least three independent experiments are shown.

Wild-typeSulphite179 ± 759 ± 447 ± 5
Wild-typeFumarate< 0.1< 0.1< 0.1
Wild-typeNitrate1.3 ± 0.10.43 ± 0.030.37 ± 0.03
Δmcc::kanFumarate< 0.1< 0.1< 0.1
Δmcc::kanNitrate1.2 ± 0.10.46 ± 0.070.40 ± 0.06
Δmcc::kanΔnrfHA::catFumarate< 0.1< 0.1< 0.1
Δmcc::kanΔnrfHA::catNitrate< 0.1< 0.1< 0.1
Pfrd-mccFumarate153 ± 852 ± 544 ± 4
MccA 2xStrepFumarate149 ± 551 ± 747 ± 6
Pfrd-mccΔccsA1Fumarate< 0.1< 0.1< 0.1
Pfrd-mccΔmccBFumarate< 0.1< 0.1< 0.1
Pfrd-mccΔmccCFumarate< 0.1< 0.1< 0.1
Pfrd-mccΔmccDFumarate< 0.1< 0.1< 0.1
MccA A575SFumarate149 ± 746 ± 341 ± 3
MccA C562A A575CFumarate< 0.1< 0.1< 0.1
Δws384/5::kanSulphite177 ± 759 ± 551 ± 4
Pfrd-mccΔws384/5::kanFumarate151 ± 950 ± 540 ± 3

Sulphite respiration requires the octahaem cytochrome c MccA

A homogenate of sulphite-grown W. succinogenes wild-type cells was subjected to denaturing SDS polyacrylamide gel electrophoresis and subsequently stained for covalently bound haem. This procedure revealed the presence of an 80 kDa cytochrome c that is absent from fumarate-grown wild-type cells (Fig. 2). This cytochrome c was previously identified as the octahaem cytochrome c MccA (Hartshorne et al., 2007). MccA identification by haem staining is unequivocal as the W. succinogenes genome does not encode any other cytochrome c of a similar size (Kern et al., 2010a). In the absence of fumarate or nitrate, sulphite concentrations as low as 60 µM were sufficient to fully induce MccA production (not shown). In contrast to sulphite, the addition of sodium thiosulphate (10 mM) or sodium sulphide (10 mM; non-toxic to cells) did not give rise to detectable formation of MccA in haem staining experiments (Fig. 2). Furthermore, the production of MccA was not observed in cells grown with nitrate, nitrite, nitrous oxide or polysulphide as electron acceptor (not shown). When sulphite (10 mM) was present in addition to fumarate or nitrate, MccA was produced only after the consumption of fumarate or the complete reduction of nitrate to ammonium (Figs S1 and S2). In these cases, the appearance of MccA coincided with the start of sulphite conversion.

Figure 2.

Haem stain analysis of cells from different W. succinogenes strains. Cells were grown by sulphite or fumarate respiration as indicated (lanes A, D and E). To check the effect of the presence of either sulphide or thiosulphate (10 mM each; lanes B and C), fumarate-grown cells were incubated in the same medium but without fumarate for several days before harvesting. 150 µg of protein was applied to each lane of an SDS polyacylamide gel that was subsequently stained using the 3,3′-dimethoxybenzidine haem stain assay. Molecular masses given on the right refer to the size marker shown between lanes D and E.

The experiments described above are the first report on MccA production in wild-type W. succinogenes cells. Previously, MccA was purified from cells of mutant Pfrd-mccA that expressed the mcc gene cluster under the control of the fumarate reductase promoter (Table 2; Hartshorne et al., 2007). Fumarate-grown cells of this mutant were found to produce MccA amounts almost identical with those of sulphite-grown wild-type cells and commensurate specific rates of formate and sulphite turnover were determined indicating an intact electron transport chain from formate to sulphite (Fig. 2; Table 1).

Table 2.  Strains of W. succinogenes used in this study.
StrainConstruction and/or relevant propertiesaReference
  1. a. CmR and KmR denote resistance against chloramphenicol or kanamycin.

 1. Wild-typeType strain DSMZ 1740.DSMZ
 2. Pfrd-mccAOverproduction of MccA; KmR.Hartshorne et al. (2007)
 3. Δmcc::kanDeletion-insertion mutant lacking the genes mccA, mccB, mccC, mccD and ccsA1; KmR.Kern et al. (2010b)
 4. Pfrd-mccDerivative of strain 3 containing a restored mcc gene cluster transcribed from the fumarate reductase promoter; CmR.Kern et al. (2010b)
 5. MccA 2xStrepSimilar to strain 4 but encoding an MccA protein containing two consecutive Strep-tags at the C-terminus; CmR.This work
 6. ΔnrfHA::catDeletion-insertion mutant lacking the cytochrome c nitrite reductase genes nrfHA; CmR.Simon et al. (2003)
 7. ΔmccA::kanΔnrfHA::catDerivative of strain 3 that lacks the nrfHA genes coding for the subunits of the membrane-bound cytochrome c nitrite reductase complex NrfHA; CmR, KmR.This work
 8. Pfrd-mccΔccsA1Similar to strain 5 but lacking the ccsA1 gene; CmRThis work
 9. Pfrd-mccΔmccBSimilar to strain 5 but lacking the mccB gene; CmR.This work
10. Pfrd-mccΔmccCSimilar to strain 5 but lacking the mccC gene; CmR.This work
11. Pfrd-mccΔmccDSimilar to strain 5 but lacking the mccD gene; CmR.This work
12. Δws0384/5::kanDeletion-insertion mutant lacking the genes encoding Ws0384 and Ws0385; KmR.This work
13. Pfrd-mccΔws0384/5::kanSimilar to strain 5 but lacking the genes encoding Ws0384 and Ws0385; CmR, KmR.This work
14. MccA C562ASimilar to strain 5 but encoding the C562A variant of MccA; CmR.This work
15. MccA C562SSimilar to strain 5 but encoding the C562S variant of MccA; CmR.This work
16. MccA A575CSimilar to strain 5 but encoding the A575C variant of MccA; CmR.This work
17. MccA A575SSimilar to strain 5 but encoding the A575S variant of MccA; CmR.This work
18. MccA C578ASimilar to strain 5 but encoding the C578A variant of MccA; CmR.This work
19. MccA C578SSimilar to strain 5 but encoding the C578S variant of MccA; CmR.This work
20. MccA H579ASimilar to strain 5 but encoding the H579A variant of MccA; CmR.This work
21. MccA C562A A575CSimilar to strain 5 but encoding the C562A A575C variant of MccA; CmR.This work
22. MccA C562A A575SSimilar to strain 5 but encoding the C562A A575S variant of MccA; CmR.This work
23. MccA C562S A575CSimilar to strain 5 but encoding the C562S A575C variant of MccA; CmR.This work
24. MccA C562S A575SSimilar to strain 5 but encoding the C562S A575S variant of MccA; CmR.This work

Previously, a W. succinogenes mutant (strain Δmcc::kan) was constructed that lacks the mcc gene cluster consisting of the mccA, -B, -C, -D and ccsA1 genes (Table 2). The mcc gene cluster was restored in this mutant by insertion of a suitable plasmid, which contained the fumarate reductase promoter for the control of mcc expression, and the resulting strain was designated W. succinogenes Pfrd-mcc (Table 2; see Fig. 1 in Kern et al., 2010b). The Δmcc::kan mutant did not grow by sulphite respiration (Fig. 1A). On the other hand, mutant Pfrd-mcc produced MccA in amounts similar to those of mutant Pfrd-mccA (Fig. 3B). The consumption rates of formate and sulphite in a cell suspension of this strain amounted to more than 85% of sulphite-grown wild-type cells (Table 1).

Figure 3.

Modification of the unconventional haem c binding motif in MccA. A. Primary structure of the CX15CH haem c binding motif (top). Bold residues are strictly conserved throughout the MccA family (see Fig. S3 for an alignment). The numbered residues have been replaced in this study and the corresponding substitutions are given below (see also Table 2). B. Detection of MccA and NrfA in cells of the indicated W. succinogenes strains. Formate/fumarate medium was inoculated with nitrate-grown cells, incubated at 37°C and samples were taken at the given incubation times. 150 µg protein was applied to each lane of an SDS polyacylamide gel, blotted onto a PVDF membrane and subjected to haem staining using the chemoluminescence assay.

In the absence of detectable amounts of MccA, nitrate-grown wild-type cells contained a residual activity of formate-dependent sulphite consumption that was less than 1% of the activity found in a sulphite-grown wild-type cell suspension (Table 1). In contrast, sulphite turnover was not observed in fumarate-grown wild-type cells (Table 1). In nitrate-grown cells of the Δmcc::kan mutant the residual rate of sulphite consumption was still present; however, it was absent from cells of strain W. succinogenesΔmcc::kanΔnrfHA::cat that lacked the cytochrome c nitrite reductase complex NrfHA in addition to MccA (Table 1). These results suggest that NrfA is responsible for the low sulphite reductase activity in strain Δmcc::kan, which, however, is irrelevant to sulphite respiration. The results are in line with the fact that the amount of NrfA in fumarate-grown cells is much lower than that in nitrate-grown cells (Lorenzen et al., 1993; Kern and Simon, 2009a). They also confirm the previously determined sulphite reductase activity of NrfA (Lukat et al., 2008).

MccA functions as respiratory sulphite reductase

Specific sulphite reductase activities using reduced methyl viologen as electron donor were determined in cell homogenates and cell fractions of different W. succinogenes strains (Table 3). Overall, the activity values correlate very well with the sulphite consumption rates presented in Table 1, thus indicating that MccA is the dominant sulphite reductase. The residual specific sulphite reductase in cells of strain Δmcc::kan is caused by NrfA as nitrate-grown cells of the ΔnrfHA::cat mutant completely lacked sulphite reductase activity (Table 3). When grown by sulphite respiration, the ΔnrfHA::cat mutant contained wild-type amounts of sulphite reductase activity confirming that the contribution of NrfA to total sulphite reduction is very low (Table 3). Such cells, on the other hand, did not contain any nitrite reductase activity determined with benzyl viologen as electron donor despite the presence of MccA (not shown). This result indicates that MccA is not capable to reduce nitrite.

Table 3.  Specific activities of reduced methyl viologen (MVred)-dependent sulphite reduction in cell homogenates and cell fractions of W. succinogenes strains harvested in the early stationary growth phase.
StrainTerminal acceptor of anaerobic respirationSpecific sulphite reductase activity MVred→ sulphite (U mg protein−1)
Cell homogenateSoluble fractionMembrane fraction
  1. Mean values and standard deviations from at least three independent experiments are shown.

Wild-typeSulphite0.69 ± 0.061.44 ± 0.080.01 ± 0.01
Wild-typeFumarate0.02 ± 0.010.02 ± 0.010.01 ± 0.01
Wild-typeNitrate0.04 ± 0.020.03 ± 0.010.05 ± 0.03
Δmcc::kanFumarate0.03 ± 0.010.02 ± 0.010.02 ± 0.01
Δmcc::kanNitrate0.04 ± 0.020.02 ± 0.010.04 ± 0.02
ΔnrfHA::catFumarate< 0.01< 0.01< 0.01
ΔnrfHA::catNitrate< 0.01< 0.01< 0.01
Pfrd-mccFumarate0.51 ± 0.031.16 ± 0.050.01 ± 0.01
MccA 2xStrepFumarate0.55 ± 0.041.24 ± 0.080.01 ± 0.01
Pfrd-mccΔmccCFumarate0.50 ± 0.011.17 ± 0.070.01 ± 0.01
Pfrd-mccΔmccDFumarate0.58 ± 0.021.21 ± 0.060.01 ± 0.01
Δws384/5::kanSulphite0.70 ± 0.071.41 ± 0.070.01 ± 0.01
Pfrd-mccΔws384/5::kanFumarate0.56 ± 0.011.24 ± 0.040.01 ± 0.01

A Strep-tagged variant of MccA was produced in cells of W. succinogenes MccA 2xStrep (strain 5 in Table 2). This strain showed substrate conversion rates and specific sulphite reductase activities similar to W. succinogenes Pfrd-mcc (Tables 1 and 3). Strep-tagged MccA was purified to homogeneity (not shown) and its UV/Vis absorption spectrum proved to be indistinguishable from previously purified wild-type MccA (Fig. 3 in Hartshorne et al., 2007). A typical preparation of Strep-tagged MccA protein (yield > 60%) had a specific sulphite reductase activity of 135 U mg protein−1 and was thus enriched by a factor of about 260.

The W. succinogenes MccA protein contains a CX12AKGCH sequence and both cysteine residues of this motif were previously attributed to covalent binding of a special haem c group (Fig. 3A) (Hartshorne et al., 2007). The attachment of haem c to this motif was suggested to depend on the dedicated cytochrome c haem synthase CcsA1 encoded within the mcc gene cluster and a W. succinogenesΔccsA1 mutant was shown previously to be incapable to produce stable MccA (Hartshorne et al., 2006; 2007; Kern et al., 2010a,b). The CX12AKGCH sequence therefore represents a most unusual haem c binding motif that is conserved in each of the 42 known primary structures of MccA proteins (consensus sequence CX15-17CH or, more precisely, CA/GRTX3DX7-9GCH; Fig. S3). Derivatives of strain Pfrd-mcc were constructed encoding MccA variants that had one or two of the crucial amino acids residues within the special haem c binding motif replaced (Cys-562, Ala-575, Cys-578, His579 of pre-MccA) (Fig. 3A; strains 14 to 24 in Table 2). Out of 11 mutants, only the one producing the A575S variant of MccA was able to convert sulphite and to synthesize stable MccA in amounts similar to that of strain Pfrd-mcc (Table 1; Fig. 3B). In the stationary growth phase, the other mutants did not contain MccA in amounts detectable by haem staining indicating that the absence of the cysteines or the histidine residue of the special haem c binding motif prevented production of stable MccA (not shown). Notably, mutant MccA C562A A575C transiently formed haem c-containing MccA during growth and this behaviour was similar to that of a Pfrd-mccΔccsA1 mutant (Fig. 3B). None of the variants that lacked either one of the cysteines or the histidine residue of the special haem c binding motif was stably produced. The results imply that neither the CX12CKGCH nor the A/SX12CKGCH motif (i.e. motifs that contain the conventional CX2CH haem c binding motif) allowed haem c attachment. These findings support previous observations arguing that CcsA1 needs specific structural features in addition to the intact unconventional haem c binding motif (Kern et al., 2010a). It is not known why the cytochrome c haem synthase CcsA2 (the enzyme that recognizes conventional CX2CH haem c binding motifs; Kern et al., 2010a,b) is unable to attach haem to the introduced CX12CKGCH and A/SX12CKGCH motifs.

The role of MccB, MccC, MccD, Ws0384 and Ws0385 in sulphite respiration

The genetic system outlined above allowed the construction of in-frame deletion mutants that lacked either mccB, mccC, mccD or ccsA1 (Table 2). In addition, two mutants lacking the open reading frames ws0384 and ws0385 located immediately downstream of ccsA1 were constructed. In these mutants, the mcc gene cluster was expressed either from the wild type or from the fumarate reductase promoter (strains 12 and 13 in Table 2). It is uncertain to date whether the ws0384 and ws0385 genes are co-transcribed with genes of the mcc gene cluster whose transcription might terminate downstream of the ccsA1 gene. The hypothetical proteins Ws0384 (putatively membrane-bound) and Ws0385 (possibly located in the periplasm) are present downstream of the mcc gene clusters of some Epsilonproteobacteria but do not possess conserved motifs or homologues whose function has been examined (Fig. S4).

The mccC and mccD genes encode a putative membrane-bound electron transport complex that is envisaged to couple quinol oxidation (catalysed by MccD) to MccA reduction using MccC as an intermediate electron carrier that contains four [4Fe-4S] cluster binding motifs (Hartshorne et al., 2007; Simon and Kern, 2008). This system is reminiscent of the NrfCD module required for ubiquinol oxidation and electron transport to the pentahaem cytochrome c NrfB in respiratory nitrite ammonification of enteric bacteria like Escherichia coli (Clarke et al., 2007). Both the Pfrd-mccΔmccC and Pfrd-mccΔmccD mutants produced active MccA (Table 3) and the amount of MccA in these strains was found to be similar to those of the Pfrd-mcc mutant (not shown). In contrast, cultures of the Pfrd-mccΔmccC and Pfrd-mccΔmccD mutants did not convert sulphite suggesting that electron transport to the MccA sulphite reductase was abolished (Table 1).

The mccB gene encodes a putative peptidyl-prolyl cis-trans isomerase of unknown function. A W. succinogenesΔmccB mutant was constructed but MccA was not detected in this strain and, consequently, no sulphite consumption was observed (Table 1). This phenotype supports the view that MccB is required for MccA maturation, which was anticipated as all mcc clusters from a diverse range of bacteria contain a mccB homologue (see Discussion and Fig. S4). Furthermore, eight strictly conserved proline residues are present in all 42 MccA sequences and two of them are located close to haem c binding motifs within CX2CHXP arrangements.

Both mutants devoid of the genes ws0384 and ws0385 showed properties that corresponded to those of the respective parent strains (Tables 1 and 3), thus preventing us to assign a function to the encoded proteins in sulphite respiration.


Sulphite respiration in W. succinogenes

It has been demonstrated here for the first time that W. succinogenes cells are able to grow by anaerobic sulphite respiration. This capacity adds to other dissimilatory reactions of W. succinogenes using sulphur compounds, namely polysulphide respiration with formate or hydrogen gas as electron donor and fumarate respiration using sulphide as electron donor (Macy et al., 1986; Schumacher et al., 1992; Klimmek et al., 2004). In contrast to Campylobacter jejuni, W. succinogenes was not reported to grow by microaerobic respiration using sulphite as electron donor and the corresponding sulphite:cytochrome c oxidoreductase (sor) genes are absent from its genome (Myers and Kelly, 2005).

Most likely, the octahaem cytochrome c MccA functions as the sole terminal reductase of sulphite respiration (see the corresponding model in Fig. 4). MccA was detected in sulphite-grown W. succinogenes wild-type cells for the first time since the protein is absent from cells grown with alternative electron acceptors indicating tight regulation of mccABCD-ccsA1 expression. This finding argues against a role of MccA in sulphite detoxification. The nature of the assumed sulphite-sensor and the corresponding signal transduction mechanism has not been investigated as yet. In this context it is notable that a two-component regulatory system named MccR/MccS is encoded immediately upstream of the mccABCD-ccsA1 gene cluster that might be involved in the synthesis of MccA (Fig. 4) (Hartshorne et al., 2006). The mccRS genes were only found in the genomes of W. succinogenes, Sulfurospirillum deleyianum and some, but not all, Campylobacter species (Fig. S4), whereas all other MccA-encoding bacteria including many Epsilonproteobacteria of the genus Campylobacter lack these genes.

Figure 4.

Model of the electron transport chain catalysing electrogenic formate oxidation by sulphite in W. succinogenes. The left part shows the putative electron transport chain that catalyses the six-electron reaction of sulphite reduction by formate and comprises the electrogenic heterotrimeric formate dehydrogenase complex (FdhABC), the quinone (Q)/quinol (QH2) pool, the MccCD quinol dehydrogenase and MccA (see text for further details). Formate and sulphite are drawn in their protonated forms to illustrate the bioenergetics of sulphite respiration based on the coupled turnover of protons and electrons (underlined). While formate-dependent quinone reduction is electrogenic, quinol oxidation by sulphite is assumed to be electroneutral. The right part depicts the distinct functions of two cytochrome c haem lyases (CcsA1 and CcsA2) in periplasmic MccA maturation. Both haem lyases contain ten transmembrane segments and are assumed to transport haem b across the membrane (Kern et al., 2010b). The W. succinogenes mcc gene cluster is shown at the bottom. Mo, molybdenum cofactor; R, mccR gene.

MccA represents a novel sulphide-generating cytochrome c sulphite reductase distinct from functionally similar sirohaem-containing enzymes. Isolated W. succinogenes MccA was shown previously to be inactive towards reduction of fumarate, nitrate, nitrite, thiosulphate, tetrathionate and Fe(III) complexes (Hartshorne et al., 2007). Therefore, MccA appears to be more specific than NrfA, which reduces nitrite and sulphite as well as nitric oxide, hydroxylamine and hydrogen peroxide (Lukat et al., 2008; Kern et al., 2011; Simon et al., 2011). Like MccA, NrfA also contains a special haem c group as the active site haem c is bound by a CX2CK haem c binding motif that donates a proximal lysine ligand to the haem c iron (Einsle et al., 2000; Pisa et al., 2002).

It was shown previously that MccA forms a periplasmic homotrimer under physiological conditions that contains 24 haem c groups (Hartshorne et al., 2007). One of the eight haem c groups of the MccA monomer was clearly distinct in showing a red-shifted absorption maximum in its reduced form. Furthermore, spectropotentiometric titration of MccA revealed that this haem c had a considerably more positive midpoint potential (+ 17 mV) than any other MccA haem c group (Hartshorne et al., 2007). These features are indicative of a special haem c group that is possibly involved in sulphite reduction and it seems to be an attractive idea that this haem c group is the one that is attached to the special CX15CH haem c binding motif (Fig. 4). On the other hand, previous EPR and UV/Vis spectroscopy did not suggest the presence of a high-spin haem c, which, however, might have been masked by strong electronic coupling (Hartshorne et al., 2007). The maturation of MccA is a complex process that involves the Sec system for apo-MccA secretion, the cytochrome c biogenesis system II (involving the distinct cytochrome c haem synthases CcsA2 and CcsA1) and probably also the assumed peptidyl-prolyl cis-trans isomerase MccB (Fig. 4) (Hartshorne et al., 2006; 2007; Kern et al., 2010a,b; Simon and Hederstedt, 2011).

Which components are involved in the set-up of a pmf generating electron transport chain catalysing formate oxidation by sulphite? In our current model (Fig. 4), the well-known electrogenic heterotrimeric formate dehydrogenase complex (FdhABC) couples formate oxidation to quinone reduction (Simon et al., 2008; Kern and Simon, 2009a). The quinol-reducing part of the electron transport chain is less clear but might minimally involve the putative quinol dehydrogenase module MccCD and MccA as the terminal reductase. MccD, a PsrC/NrfD homologue with eight putative membrane-spanning segments, might catalyse quinol oxidation and electron transfer to the iron-sulphur protein MccC, which is assumed to work as a specific redox mediator between MccD and MccA (Simon and Kern, 2008). It appears that MccC and MccD can not be functionally replaced by other proteins in W. succinogenes sulphite respiration under the tested conditions. At present, it is unknown which type of quinone is involved in W. succinogenes sulphite respiration. W. succinogenes is known to contain menaquinone6 as an essential quinone under fumarate- and nitrate-respiring conditions but produces methyl-menaquinone6 during polysulphide respiration (Dietrich and Klimmek, 2002). Methyl-menaquinone6 was shown to be involved in electron transport from formate to polysulphide and was proposed to bind to the membrane anchor subunit PsrC (a MccD homologue) of the polysulphide reductase complex PsrABC (Dietrich and Klimmek, 2002; Klimmek et al., 2004). This finding is also supported by the only known structure of a member of the PsrC/NrfD family (Jormakka et al., 2008). However, it is presently unknown whether methyl-menaquinone6 plays a role in sulphite respiration. In the context of the low midpoint potential of the redox pair sulphite/sulphide (– 116 mV) it is reasonable that a low-potential menaquinone derivative like methyl-menaquinol might be the electron donor to MccD. However, the capacity of the FdhABC complex to reduce methyl-menaquinol6 has not been investigated. In Fig. 4, a proton per electron ratio of 1 is assumed for the electron transport from formate to sulphite via the respiratory Fdh and Mcc systems. According to this model, formate oxidation coupled to quinone reduction is electrogenic using the redox loop mechanism while quinol oxidation by sulphite is not (Simon et al., 2008). Given the small amount of available free energy, a proton per electron ratio of less than 1 cannot be excluded, similar to what was experimentally suggested for polysulphide respiration (Klimmek et al., 2004).

Function of MccA and organization of mcc gene clusters in other bacteria

Apart from W. succinogenes, the Gammaproteobacterium S. oneidensis MR-1 is the best-studied bacterium with respect to MccA function. The transcription of the S. oneidensis mcc gene cluster was shown to be upregulated in the presence of thiosulphate (Beliaev et al., 2005). Because this organism is able to reduce thiosulphate to yield sulphite, it is conceivable that the observed regulatory effect might have been mediated by sulphite (Burns and DiChristina, 2009). In line with this hypothesis, S. oneidensis MccA (recently renamed SirA) was shown to be involved in respiratory sulphite reduction and SirA was identified as the sulphite reductase on the basis of activity staining in a gel by Shirodkar et al. (2011). These authors also reported that the presence/absence of mcc (sir) gene clusters in different Shewanella species correlated with their ability/disability to reduce sulphite. To our knowledge, W. succinogenes and S. oneidensis are the only organisms for which physiological and/or biochemical data on MccA function are available to date.

The 42 known mccA (sirA) genes are encoded in one or two copies as part of mcc gene clusters in genomes of the epsilonproteobacterial genera Wolinella (1 single mccA gene copy), Sulfurospirillum (1 single copy) and Campylobacter (11 single copies); the gammaproteobacterial genera Edwardsiella (3 single copies), Ferrimonas (2 gene copies in one genome) and Shewanella (19 genes; 11 single copies; 4 double copies); the deltaproteobacterial genus Anaeromxyobacter (3 single copies); and in two betaproteobacterial genomes in the order Burkholderiales (single copies) (Fig. S4). Each locus contains a MccB-type peptidyl-prolyl cis-trans isomerase and a cytochrome c haem synthase (either of the CcsA1-type or the NrfEFG-type depending on the presence of the cytochrome c biogenesis system I or II) (Hartshorne et al., 2007; Kranz et al., 2009). A S. oneidensis mutant that lacked the nrfG homologue (sirG) was unable to reduce sulphite suggesting that the corresponding protein is involved in MccA maturation similar to W. succinogenes CcsA1 (Shirodkar et al., 2011). Of the mcc loci, 26 encode an MccCD quinol dehydrogenase module. Constitutive expression of the mccCD (sirCD) genes in S. oneidensis was recently shown to compensate the lack of CymA, which acts as a versatile membrane-bound tetrahaem cytochrome c quinol dehydrogenase that directly or indirectly delivers electrons to various terminal reductases of anaerobic respiration (Cordova et al., 2011). This result indicated that MccCD (SirCD) is functional and supports the finding that S. oneidensis sulphite respiration is independent of CymA (Shirodkar et al., 2011). Bacteria lacking the mccCD/sirCD genes might functionally compensate this deficit by using alternative proteins encoded elsewhere on their genomes. In principal, the Mcc/Sir systems of W. succinogenes and S. oneidensis are assumed to have equivalent functions. However, the precise role of the MccC and MccD homologues and the nature of the quinone species involved in electron transport of MccA needs to be elucidated for both bacteria in the future. Apart from the genes discussed above, there are several more genes that are present only occasionally in mcc gene clusters (Fig. S4). The function of these genes is not known and corresponding mutant strains have not been described.

Evolutionary history of MccA

Phylogenetic analysis of MccA protein sequences in context with sequences of other multihaem cytochromes c revealed a surprisingly close relationship with a rather diverse family of octa- and tetrahaem cytochromes c encoded in the genomes of predominantly Deltaproteobacteria as well as species of Veruccomicrobia, Planctomycetes and Beta- and Gammaproteobacteria (Fig. 5A and B). The tetrahaem cytochromes c of the latter three are known as cytochrome c554 proteins. Studied cytochrome c554 proteins function as catabolic redox shuttles in nitrification but have also been shown to possess nitric oxide reductase activity associated with the haem c group bound to the second haem c binding motif (Upadhyay et al., 2006; Klotz and Stein, 2011). An overwhelming number of these cytochrome c554 proteins, encoded in deltaproteobacterial genomes, have not been characterized. Interestingly, cytochrome c554-encoding genes appear to be absent from epsilonproteobacterial genomes.

Figure 5.

Phylogenetic relationships within the superfamily of multihaem cytochromes c. A. Consensus tree constructed after Bayesian inference of phylogeny from the ClustalX alignments of 285 NrfA, Onr, Hao/Hzo, cytochrome c554, MccA and Otr protein sequences. The N- and C-terminal sequences outside the trihaem core haem c binding motifs were eliminated from the final alignment that was subjected to Bayesian inference of phylogeny using the BEAST package (see Experimental procedures). An unrooted 50% majority rule consensus phylogram was displayed as a circle tree, for which posterior probability values of the main nodes are shown. Mean branch lengths are characterized by a scale bar indicating the evolutionary distance between the proteins (changes per amino acid position). The branches are annotated with labels indicating the source organism and the protein sequence accession number. Clusters of sequences were categorized as tetrahaem (TCC; cytochrome c554 in teal), pentahaem (PCC; NrfA in green) or octahaem cytochromes c (OCC; Onr in brown, Hao/Hzo in red/gold, OTR in black, cytochrome c554 in teal and MccA in magenta) with clades of sequences from Epsilonproteobacteria shown in blue. B. Star tree configuration of (A) indicating that the clades coalesce in a node that separates a superclade of MccA/cytochrome c554 and Otr sequences from the Hao/Hzo, Onr and NrfA clades. C. Alignment of haem c binding motifs in multihaem cytochrome c families indicating the trihaem core (see text for details).

Other multihaem cytochromes c that are reactive with sulphur and/or nitrogen compounds comprise the octahaem tetrathionate reductase (Otr), members of the octahaem hydroxylamine oxidoreductase/hydrazine oxidoreductase (Hao/Hzo) family as well as pentahaem cytochrome c (PCC) nitrite reductase (NrfA, see Introduction) and the related octahaem cytochrome c nitrite reductase (Onr) (Simon et al., 2011). A scenario for the evolution of this superfamily of multihaem cytochromes was reported recently by Klotz et al. (2008) based on NrfA, Onr and Hao/Hzo sequences and a putative evolutionary relationship between Hao and cytochrome c554 proteins had been suggested (but not tested) based on protein structure and haem c arrangements (Hooper et al., 2005). Inclusion of an extensive number of MccA, cytochrome c554 and Otr protein sequences in an updated phylogenetic analysis of the NrfA, Onr and Hao/Hzo proteins resulted in a phylogenetic tree that separated three major clades (the MccA/cytochrome c554 proteins, the Otr proteins and the Hao/Hzo/Onr/NrfA proteins), which are connected at the level of octahaem cytochrome c (OCC) proteins (Fig. 5B). All known Hao/Hzo, MccA and Otr proteins are OCC proteins. Interestingly, the detailed tree (Fig. 5A) reveals that the history of the cytochrome c554 family included three independent events of gene/protein truncation from OCC to tetrahaem cytochrome c (TCC) thereby creating several subfamilies whose emergence was likely tied to functional pressures. While the function of the three deltaproteobacterial tetrahaem cytochrome c554 protein subfamilies is experimentally untested, the tetrahaem cytochrome c554 protein subfamily represented in aerobic ammonia-oxidizing bacteria is known to be part of catabolic electron flow. In contrast, one of the two cytochrome c554 proteins encoded in anaerobic ammonia-oxidizing bacteria (CAJ75043), the genomic background in which nitrogen-based catabolic electron flow probably emerged (Klotz and Stein, 2011 and references therein), is a OCC protein. This apparent scenario of reduction in participating haems within the MccA/cytochrome c554 branch of the superfamily led us to propose a slightly revised scenario for the evolutionary relationships. While not contradicting previous findings that led to a proposal of an evolutionary direction from PCC (NrfA) to OCC (Onr, Hao), the extended dataset suggests that the great diversity of extant proteins originated from an ancestral OCC protein that very likely was able to reduce a sulphur compound like sulphite or tetrathionate (as in extant MccA and Otr proteins). Divergence over evolutionary time has then led to new characteristics whereby modified and de novo-evolved active sites successfully reduced nitrogen oxides and/or oxyanions (cytochrome c554; Hao/Hzo, Onr, NrfA and Otr). Across the entire superfamily, it appears that only a subset of the Hao/Hzo family acquired the ability to oxidize hydroxylamine or hydrazine, utilizing disproportionation chemistry enabled by a covalently bonded circular symmetric trimer configuration (Hooper et al., 2005; Kostera et al., 2008). None of the enzymes in the Hao/Hzo family is known to oxidize sulphur compounds. The functional driver in the evolution of the MccA protein family might have been the ability to provide additional respiratory capacity paired with enriching the pool of reduced sulphur. Apparently, this development occurred exclusively in Proteobacteria and prevented emergence of the capacity to reduce nitrite. As illustrated in Fig. 5C, all members of the multihaem cytochrome c superfamily share a core of three haem c binding motifs including the one that binds the active site haem group in NrfA (haem 1), Onr (haem 4), Hao/Hzo (haem 4) and Otr (haem 2). This motif aligns with haem 2 of MccA and cytochrome c554. It is currently unclear which haem c group of MccA might be situated close to the active of sulphite reduction or which role the CX15CH-ligated haem c group of MccA (haem 8) might play apart from being essential for protein integrity.

Conclusion and perspective

The respiratory Mcc system has been established as a proteobacterial periplasmic sulphite reduction module that is clearly distinct from the cytoplasmic Sir enzymes found in various sulphate reducing Deltaproteobacteria and in some Archaea. Both systems rely on different cofactors (sirohaem or haem c) and their respective mechanisms of sulphite respiration are assumed to differ. Important open questions concern the structure–function relationships of MccA proteins and the detailed role of the putative MccCD complex in quinol oxidation and electron transport to MccA.

Experimental procedures

Growth of W. succinogenes cells

Strains of W. succinogenes used in this study are listed in Table 2. The cells were grown by sulphite respiration in medium (pH 7.5) containing Tris (50 mM), sodium formate (100 mM), sodium sulphite (10 mM), ammonium sulphate (5 mM), sodium succinate (5 mM), K2HPO4 (1 mM), MgCl2 (1 mM), CaCl2 (0.2 mM), glutamic acid (0.68 mM) and trace element solution SL-7 (0.2 ml l−1) as used by Pfennig and Trüper (1981). Media were degassed and flushed with nitrogen gas several times to reduce the oxygen content. Brain–heart infusion broth (0.1 or 0.5%, w/v), kanamycin (25 mg l−1) and/or chloramphenicol (12.5 mg l−1) was added to the medium where appropriate. Routinely, cells were grown in liquid media containing formate and either fumarate or nitrate as energy substrates (Kröger et al., 1994; Kern and Simon, 2009b).

Determination of formate, sulphite, sulphite and thiosulphate

Formate was measured as described (Kern et al., 2007). Sulphite was determined according to Pachmayer (1960) using a slightly modified protocol: 100 µl of fuchsine reagent (40 mg fuchsine dissolved in 87.5 ml deionized water and 12.5 ml concentrated H2SO4) was mixed with 400 µl sample (containing up to 0.1 mM sulphite) and 400 µl water. After 10 min of incubation at room temperature 10 µl formaldehyde was added. The optical density at 570 nm was determined after 90 min against a fuchsine reagent blank. Thiosulphate was determined by measuring the discoloration of methylene blue (Quentin and Pachmayr, 1964). Sulphide was quantified following methylene blue formation according to King and Morris (1967). Control experiments with mixed solutions of sulphide and thiosulphate verified that the development of methylene blue was not affected by the presence of thiosulphate and that the bleaching of the dye in the thiosulphate assay was not affected by sulphide.

Determination of turnover rates

Wolinella succinogenes cells were harvested in the late exponential growth phase, washed twice in anoxic buffer (pH 8) containing 50 mM Tris/HCl and 0.3 M sucrose and resuspended (15–20 g protein l−1) in an anoxic buffer (pH 8.5) containing 150 mM Bicin and 0.3 M sucrose. Aliquots of the cell suspension were incubated for 5 min at 37°C before the addition of sodium sulphite (10 mM final concentration) and sodium formate (50 mM). Samples were taken after various time intervals.

Cell fractionation and determination of sulphite reductase activity

To separate soluble and membrane fractions, W. succinogenes cells harvested in the late exponential growth phase were suspended (10 g protein l−1) in an anoxic buffer (pH 8.0) containing 50 mM Tris/HCl. The suspension was passed through a high-pressure cell disruption system (Constant Systems) at 135 MPa and centrifuged for 10 min at 3500 g. The resulting cell homogenate was centrifuged for 1 h at 100 000 g to yield the membrane fraction (sediment) and the soluble fraction. Sulphite reductase activity was measured under anoxic conditions at 37°C by photometrically recording the oxidation of methyl viologen (MV) at 578 nm. The test solution contained 50 mM potassium phosphate (pH 7.5) and 1 mM MV reduced by the addition of titanium(III)citrate (Zehnder and Wuhrmann, 1976). The reaction was started by the addition of 10 mM sodium sulphite. Activities were calculated using an extinction coefficient of ε = 19.6 mM−1 cm−1. One unit of enzyme activity is defined as the oxidation of 2 µmol MV per min. Protein was determined using the Biuret method with KCN (Bode et al., 1968).

Cytochrome c detection

Haem staining of SDS polyacrylamide gels was performed using 3,3′-dimethoxybenzidine according to the method of Francis and Becker (1984). To detect cytochromes c by chemoluminescence, samples were subjected to SDS-PAGE, transferred to PVDF membrane by Western blotting and detected using the SuperSignal West Pico chemoluminescence substrate (Thermo Scientific). X-ray films (CL-Xposure film, Thermo Scientific) were exposed for 15 min.

Purification of Strep-tagged MccA

Strep-tagged MccA was purified from the soluble cell fraction of strain W. succinogenes MccA 2xStrep using a Strep-Tactin MacroPrep cartridge (1 ml bed volume, flow rate 1 ml min−1) (IBA GmbH). The protein was eluted with 10 ml buffer (100 mM Tris/HCl, 150 mM NaCl, 2.5 mM desthiobiotin, pH 8.0) in 500 µl fractions. Cytochrome c containing fractions were pooled and concentrated via a Roti-Spin MIDI-10 column (Roth) with a molecular weight cut-off of 10 kDa.

Construction of W. succinogenes mutants

Standard genetic procedures were used (Sambrook et al., 1989). Genomic DNA was isolated from W. succinogenes using the DNeasy Tissue Kit (Qiagen). Plasmid DNA and PCR fragments were purified using the GeneJet Plasmid Mini or PCR-Purification Kit (Fermentas). PCR was carried out using Phusion High Fidelity DNA polymerase (Finnzymes) for cloning procedures or Biotaq Red DNA polymerase (Bioline) for mutant and plasmid screening with standard amplification protocols. Site-directed mutagenesis was performed using the QuikChange XL Site-Directed Mutagenesis Kit (Agilent Technologies) or the Phusion Site-Directed Mutagenesis Kit (Finnzymes) with specifically synthesized primer pairs.

Wolinella succinogenes mutants producing MccA variants (strains 5 and 14 to 24 in Table 2) were obtained from W. succinogenesΔmcc::kan upon integration of pPfrd-mcc cat derivatives, resulting in a restored mcc operon (Kern et al., 2010b). The pPfrd-mcc cat derivative encoding Strep-tagged MccA was constructed by inserting a nucleotide stretch encoding two consecutive Strep-tags separated by a linker region at the 3′ end of mccA. The Strep-tag/linker fragment was amplified from pMK2 (Kern and Simon, 2011) using the primer pair 5′-AGCGCGTGGAGCCATCC and 5′-TTATTATTTCTCGAACTGAGGGTGGCTCC and blunt-end ligated into a linear vector fragment obtained from pPfrd-mcc cat with the primer pair 5′-GGCCAAAGGTGTCGCTTTGAG and 5′-AAGGCTCTTTTTCATGATTCTCTGCGC. Site-directed mutagenesis of mccA aiming at replacements of amino acid residues in MccA was performed with pPfrd-mcc cat as template and a pair of complementary primers (Table S1). Transformation of W. succinogenesΔmcc::kan with constructed plasmids was performed by electroporation as described previously (Simon et al., 1998) Transformants were selected in the presence of chloramphenicol (12.5 mg l−1) and the intended double homologous recombination event was verified by PCR. The desired amino acid exchange as well as the integrity of DNA stretches involved in recombination events was confirmed by sequencing suitable PCR products.

Wolinella succinogenes mutants expressing the mcc gene cluster under the control of the frd promoter, but containing in-frame deletions of mccB, mccC, mccD or ccsA1 (strains 8 to 11 in Table 2) were constructed upon integration of pPfrd-mcc cat derivatives in W. succinogenesΔmcc::kan. The corresponding plasmids were constructed by amplifying linear PCR fragments from pPfrd-mcc cat as template, thereby eliminating the respective target gene, followed by phosphorylation and blunt-end ligation. The sequences of the corresponding primer pairs are given in Table S1.

Wolinella succinogenes Δws0384/5::kan and W. succinogenes Pfrd-mcc Δws0384/5::kan (strains 12 and 13 in Table 2) were constructed through homologous recombination between the genome of the wild type or W. succinogenes Pfrd-mcc with the deletion plasmid pΔws0384/5kan designed to replace the consecutive genes ws0384 and ws0385 by the kanamycin resistance gene cassette (kan). For homologous recombination, the plasmid contained kan flanked by two DNA segments obtained by PCR that were identical to appropriate regions on the W. succinogenes genome. The two PCR fragments were synthesized using the following primer pairs: 5′-GCGAATTCGAGACTTGGTCGTATGTCTCGATTC-3′ and 5′-GCGGATCCACCTCCTTAGATAGAGTTTTCAAATCG-3′ for amplifying the upstream fragment and 5′-GCGGATCCAGAAAGGATTTATGAATGTTGAAGAAG-3′ and 5′-CGGTCGACCTCTGTTCACGCTTGCTTTTGATTC-3′ for the downstream fragment. Primers carried EcoRI, BamHI or SalI restriction sited (underlined) for cloning. Both fragments as well as kan (obtained by BamHI excision from pUC4K) were consecutively inserted in pBR322 using appropriate restriction enzymes. PCR analysis confirmed that the plasmid contained kan in the same orientation as ws0384/5. Transformants of W. succinogenes or W. succinogenes Pfrd-mcc were obtained by electroporation as described above. Desired deletions were confirmed by PCR and the integrity of DNA stretches involved in recombination events was verified by sequencing suitable PCR products.

Phylogenetic analysis

Sequence-related tetrahaem (TCC), pentahaem (PCC) and octahaem (OCC) cytochrome c protein sequences, retrieved from public databases and embargo genome projects using the blast algorithm (Altschul et al., 1997), were aligned using ClustalX version 1.83 (Thompson et al., 1997). Based on these alignments, distance neighbour-joining trees were constructed with the BioNJ function in PAUP* version 4.10b (Swofford, 1999), which were used for comparison with previously published results from phylogenetic analyses (Bergmann et al. 2005; Klotz et al., 2008) and for manual refinement of the ClustalX alignments. Following the rationale in the discussion of the evolution of this multihaem cytochrome c protein superfamily introduced by Klotz et al. (2008), both structural and protein sequence-analytical features guided us to use an in-gap alignment (of haem c binding motifs 5 and 6 in MccA) and the identification of a trihaem protein core based on haem c binding motifs as shown in Fig. 5C. If this hypothesis of a core ancestral to the OCC superfamily were to be correct, then a phylogenetic analysis of the respective protein sequences should accurately reflect the evolutionary relationships between the extant representatives in the included six protein families. To test this, the N- and C-terminal sequences outside the three core haem c binding motifs in the NrfA, Onr, Hao/Hzo, cytochrome c554, MccA and Otr protein sequences were eliminated from the final alignment that was subjected to a Bayesian inference of phylogeny using the BEAST package [BEAUti v1.5.3, BEAST v1.5.3, TreeAnnotator v1.5.3, FigTree v.1.3 (Drummond and Rambaut, 2007)]. By utilizing unique sites, tree likelihoods (ignoring ambiguities) were determined for the alignment by creating a Monte-Carlo Markov Chain (10 000 000 generations) in three independent runs. The searches were conducted assuming an equal or a gamma distribution of rates across sites, sampling every 1000th generation and using the WAG empirical amino acid substitution model (Whelan and Goldman, 2001). The resulting 10 000 trees (omitting the first 350 trees as burn-in) were used to construct an unrooted phylogenetic consensus tree (Fig. 5B) that was used as the basis to discuss the evolution of the MccA protein family.


The authors thank David Richardson (Norwich), Ben Berks (Oxford) and Alan Hooper (Minneapolis) for long-term insightful discussions and Anke Mager for experimental support. This work was funded by the Deutsche Forschungsgemeinschaft (Grants SI 848/1–2 and 2–1 to J.S.). M.G.K. was supported by NSF grants EF-0412129 and MCB-0948202 as well as incentive funds from the College of Liberal Arts & Sciences and the Charlotte Research Institute, UNC-Charlotte.