Siro(haem)amide in Allochromatium vinosum and relevance of DsrL and DsrN, a homolog of cobyrinic acid a,c-diamide synthase, for sulphur oxidation


  • Editor: Jörg Simon

  • Present address: Hyung-Sun Youn, Department of Fermented Food Science, Seoul University of Venture & Information, 1603-54 Seocho-dong, Seocho-gu, Seoul 137-070, South Korea.

Correspondence: Christiane Dahl, Institut für Mikrobiologie & Biotechnologie, Rheinische Friedrich-Wilhelms-Universität Bonn, Meckenheimer Allee 168, D-53115 Bonn, Germany. Tel.: +49 228 732119; fax: +49 228 737576; e-mail:


In the purple sulphur bacterium Allochromatium vinosum, the prosthetic group of dissimilatory sulphite reductase (DsrAB) was identified as siroamide, an amidated form of the classical sirohaem. The genes dsrAB are the first two of a large cluster of genes necessary for the oxidation of sulphur globules stored intracellularly during growth on sulphide and thiosulphate. DsrN is homologous to cobyrinic acid a,c diamide synthase and may therefore catalyze glutamine-dependent amidation of sirohaem. Indeed, an A. vinosumΔdsrN in frame deletion mutant showed a significantly reduced sulphur oxidation rate that was fully restored upon complementation with dsrN in trans. Sulphite reductase was still present in the ΔdsrN mutant. DsrL is a homolog of the small subunits of bacterial glutamate synthases and was proposed to deliver glutamine for sirohaem amidation. However, recombinant DsrL does not exhibit glutamate synthase activity nor does the gene complement a glutamate synthase-deficient Escherichia coli strain. Deletion of dsrL showed that the encoded protein is absolutely essential for sulphur oxidation in A. vinosum.


Sulphite reductases are important enzymes in both assimilatory and dissimilatory sulphur metabolism. They catalyze the six-electron reduction of sulphite to sulphide. While assimilatory sulphite reductases are widespread, dissimilatory sulphite reductases are restricted to the prokaryotes. Sulphite reductase is assumed to be operating in the ‘reverse’, i.e. sulphide-oxidizing direction in the chemolithotrophic sulphur oxidizer Thiobacillus denitrificans (Schedel & Trüper, 1979) and the photolithotrophic sulphur oxidizer Allochromatium vinosum (Schedel et al., 1979). In A. vinosum, the enzyme is encoded by the first two genes of a large cluster (dsrABEFHCMKLJOPNRS). These genes are essential for the oxidation of sulphur deposited intracellularly as an intermediate of the oxidation of sulphide and thiosulphate to the final product sulphate (Pott & Dahl, 1998; Dahl et al., 2005).

In most cases, the prosthetic group of sulphite reductases has been described to be sirohaem-[Fe4S4]. Sirohaem is an iron tetrahydroporphyrin with eight carboxylic acid side chains (Murphy et al., 1973). The iron-free form is termed sirohydrochlorin. Some dissimilatory sulphite reductases from Desulfovibrio species contain siroamide, a sirohaem, in which one of the acetate chains is amidated (Matthews et al., 1995). Sirohaem and cobalamin are related tetrapyrrolic structures formed from the precursor urophorphyrinogen III. Cobyrinic acid a,c-diamide synthase (CobB) catalyzes an important step in cobalamin biosynthesis, the ATP- and glutamine-dependent amidation of cobyrinic acid to cobyrinic acid a,c-diamide (Debussche et al., 1990). Notably, the dsrN-encoded proteins from A. vinosum, the green sulphur bacterium Chlorobaculum tepidum (formerly Chlorobium tepidum) (Eisen et al., 2002; Imhoff, 2003) and several eubacterial sulphate reducers resemble CobB and the idea has been put forward that DsrN might catalyze the dependent amidation of sirohaem to siro(haem)amide (Larsen et al., 2000, 2001; Dahl et al., 2005).

On the basis of glutamine as a possible amino group donor for sirohaem amidation, it appeared noteworthy that DsrL from A. vinosum is homologous to the small subunits of bacterial glutamate synthases (Dahl et al., 2005). These enzymes usually consist of two different subunits and catalyze the formation of two glutamate molecules via reductive transfer of the glutamine amide group to 2-oxoglutarate (Vanoni & Curti, 2005). One exception might be the DsrL homologous protein GltA from Thermococcus kodakarensis that has been reported to act as glutamate synthase without the presence of a second subunit (Jongsareejit et al., 1997).

Taking the possible functions of DsrN as a glutamine-dependent sirohaem-amidating enzyme and of DsrL as a glutamate synthase providing glutamine for the former reaction as our starting point, we assessed the presence of siro(haem)amide in A. vinosum cell extracts and in highly enriched sulphite reductase. Furthermore, we studied the importance of DsrN and DsrL for sulphur oxidation in A. vinosum by construction, characterization and complementation of in frame deletion mutants. The possibility was assessed that DsrL functions as a glutamate synthase.

Materials and methods

Bacterial strains, plasmids, media and growth conditions

The strains and plasmids used in this study are listed in Table 1. Allotchromatium vinosum was grown and harvested as described earlier (Dahl, 1996). Antibiotics were used at the following concentrations (in μg mL−1): for Escherichia coli, ampicillin, 100; kanamycin, 50; for A. vinosum, kanamycin, 10; rifampicin, 50.

Table 1.   Bacterial strains, plasmids and PCR primers
Strains, plasmids, primersGenotype or phenotypeSource or reference
Escherichia coli strains
 DH5αFΦ80dlacZΔM15Δ(lacZYA-argF)U169 recA1 endA1 hsdR17 (rK mK+) supE44λthi-1 gyrA relA1Hanahan (1983)
 S17-1294 (recA pro res mod+) Tpr Smr (pRP4-2-Tc::Mu-Km::Tn7)Simon et al. (1983)
 BL21(DE3)FompT hsdSB (rB mB) gal dcm met (DE3)Novagen
 EB5001F, psiQ39::Mud1I-1734 (gltD::Mud1I-1734); wild type W3110Goss et al. (2001)
 K12Wild typeDSM 498
Allochromatium vinosum strains
 Rif50Rifr, spontaneous rifampicin-resistant mutant of A. vinosum DSM 180TThis work
 ΔdsrNRifr, ΔdsrN (deletion: 1281 bp of the dsrN gene)This work
 ΔdsrLRifrdsrL (deletion: 1443 bp of the dsrL gene)This work
 pET15bApr, His-Tag (N-terminal)Novagen
 pK18mobsacBKmr, lacZ′, sacB, Mob+Schäfer et al. (1994)
 pBBR1MCS-2Kmr, lacZ′, Mob+Kovach et al. (1995)
 pBAD22AKmr, araC, PBAD-promoterGuzman et al. (1995)
 pET11a-NApr, NdeI–BamHI fragment of PCR-amplified dsrN in NdeI–BamHI of pET11aThis work
 pET11a-NPdsrApr, XbaI–NdeI fragment of PCR-amplified dsr promoter in XbaI–NdeI of pET11a-NThis work
 pET15b-LApr, NdeI–BamHI fragment of PCR-amplified dsrL in NdeI–BamHI of pET15bThis work
 pET15b-LPdsrApr, XbaI–NdeI fragment of PCR-amplified dsr promoter in XbaI–NdeI of pET15b-LThis work
 pK18mobsacBΔdsrNKmr, XbaI fragment of PCR-amplified genome region around dsrN with deletion of 1281 bp of the dsrN sequenceThis work
 pK18mobsacBΔdsrLKmr, XbaI fragment of PCR-amplified genome region around dsrL with deletion of 1443 bp of the dsrL sequenceThis work
 pBBR1MCS2-NKmr, XbaI–HindIII fragment of PCR-amplified dsr promoter and dsrN in XbaI–HindIII of pBBR1MCS2This work
 pBBR1MCS2-LKmr, XbaI–HindIII fragment of PCR-amplified dsr promoter and dsrL in XbaI–HindIII of pBBR1MCS2This work
 pBAD22A-LKmr, NcoI–XbaI fragment of PCR-amplified dsrL in NcoI–XbaI of pBAD22AThis work
PCR primers
 dsrL for cloning in pBAD 22A5′-AACGATTGCCATGGCGACTTCC-3′This work
 Dsr promoter+dsrN5′-GATGGCGATCTAGACTGACTTCATGG-3′This work
 Dsr promoter+dsrL5′-GATGGCGATCTAGACTGACTTCATGG-3′This work

Construction of in frame deletions in dsrN and dsrL

All general molecular genetic techniques were described earlier (Dahl, 1996; Pott & Dahl, 1998). PCR amplicons containing internal deletions of dsrN or dsrL without altering the reading frames were obtained by gene SOEing PCR (Horton, 1995) with modified primers (Table 1). The PCR amplicons were cloned into the XbaI site of the mobilizable suicide vector pK18mobsacB. Heterogenote A. vinosum recombinants were identified via the kanamycin resistance encoded by the plasmid. After growth on nonselective liquid RCV medium for several generations homogenote recombinants were enriched as sucrose-resistant survivors by plating on solid RCV medium containing 10% sucrose. The genotypes of the resulting mutants were confirmed by PCR and Southern hybridization.

Construction of complementation plasmids

NcoI and XbaI restriction sites were introduced at the 5′ and 3′ ends of dsrL by PCR (Table 1). The product was cloned into the arabinose-inducible vector pBAD22A and used for complementation of E. coli EB5001. For complementation of A. vinosumΔdsrN and ΔdsrL, the putative dsr promoter (Pott & Dahl, 1998) was amplified by PCR, cut with NdeI and XbaI and placed immediately upstream of the dsrN or dsrL gene in plasmids pET11a-N and pET15b-L, respectively (Table 1). To verify amplification of the correct dsr promoter fragment, the nucleotide sequence of four products arising from independent reactions was determined (Sequiserve, Vaterstetten). All yielded an identical sequence; however, this sequence differed in nine positions from that originally deposited by us at GenBank (Dahl et al., 2005). Necessary changes included the insertion of one nucleotide each at four different positions (updated sequence accession number U84760). Sequences of derived proteins are not affected. The dsr promoter and the dsrN or dsrL gene were amplified together from pET11a-NPdsr or pET15b-LPdsr and cloned between XbaI and HindIII of pBBRMCS-2 (Table 1). The latter plasmid can be transferred into A. vinosum by conjugation and is autonomously replicated in this bacterium.

Complementation of E. coli EB5001

Complementation of gltD-deficient E. coli EB5001 with dsrL from A. vinosum was studied in W4 minimal medium supplemented with 0.2% glucose, 0.4% arabinose and 0.2%l-arginine as the sole nitrogen source. A 250 mL flask with 100 mL W4 medium containing 100 μg ampicillin mL−1 and 50 μg kanamycin mL−1 (to select for the Mud1I-1734 insert) was inoculated with a 5 mL overnight culture of E. coli EB 5001 containing plasmid pBAD22A-L (washed twice in W4 before inoculation). The cells were cultured at 37°C and 180 r.p.m. and growth was followed by OD600 nm.

Purification of recombinant DsrL

DsrL was overproduced with an amino-terminal His-tag in E. coli BL21(DE3) containing plasmid pET15b-L (Table 1). The cells were cultured in 500 mL LB medium containing 100 μg ampicillin mL−1 at 37°C and 180 r.p.m. At an OD600 nm of 0.6 1 mM IPTG was added. Incubation at 30°C was continued for 2 h. After harvesting, the cell pellet (washed twice) was redissolved in buffer A (50 mM NaH2PO4, 300 mM NaCl, 10 mM imidazole, 1 mM TCEP, 10% glycerol, pH 7.5). Cells were disrupted by sonication (1.5 min mL−1, Cell Disruptor B15, Branson), followed by centrifugation (25 000 g, 30 min, 4°C). The supernatant was chromatographed on a His-Bind Resin column (Novagen, Madison, WI; 3.5 mL; flow rate: 1 mL min−1) according to the manufacturer. After removal of weakly bound proteins, DsrL was eluted by a linear gradient of 50–500 mM imidazole in a volume of 30 mL. The combined fractions containing DsrL were immediately dialyzed against buffer B (25 mM HEPES, 1 mM TCEP, 1 mM EDTA, 10% glycerol, pH 7.5) and loaded onto MonoQ HR 5/5 equilibrated with the same buffer. Proteins were eluted with a linear gradient from 0–400 mM NaCl. DsrL eluted at 250–400 mM NaCl and was stored at −70°C.

Purification of sulphite reductase from A. vinosum

Purification of sulphite reductase was followed by the absorbances at 395 and 595 nm and by Western blot analysis with specific antibodies directed against sulphite reductase from Archaeoglobus fulgidus (Dahl et al., 1993). Thawed cells (150 g) were resuspended in 50 mM potassium phosphate buffer pH 7.5 containing some grains of DNase at a ratio of 3 mL buffer (g wet weight)−1, disrupted by ultrasonic treatment, centrifuged and ultracentrifuged (145 000 g, 3 h, 4°C). The supernatant was brought to 1.7 M (NH4)2SO4 saturation. After incubation on ice for 10–16 h, precipitated protein was removed by centrifugation (25 000 g, 30 min, 4°C). The supernatant was subjected to a low-substitution phenyl-Sepharose matrix (Amersham Pharmacia Biotech, Uppsala, Ø 2.6 × 15 cm) equilibrated with 1.7 M (NH4)2SO4 in 50 mM potassium phosphate buffer, pH 7.5. The column was washed with at least two volumes of equilibration buffer until all unbound proteins were removed. Bound protein was eluted by decreasing the (NH4)2SO4 concentration in a linear gradient of 350 mL (flow rate: 2 mL min−1). DsrAB eluted at 0.5–0.25 M (NH4)2SO4. The respective fractions were combined, dialyzed against 10 mM Tris/HCl, pH 7.5 and loaded onto MonoQ HR 5/5 equilibrated with the same buffer. The column was washed with 10 mM Tris/HCl containing 300 mM NaCl and proteins were eluted with a linear gradient from 300 to 600 mM NaCl. DsrAB eluted at 460–500 mM NaCl and the respective fractions were combined, frozen at −70°C and lyophilized. A total of 4.5 mg of c. 80% pure protein were obtained.

Isolation of sirohydrochlorin-based tetrapyrroles

A soluble cell extract of A. vinosum was prepared from 250 g wet weight of cells resuspended at a ratio of 1 mL buffer (g wet weight)−1 in 50 mM Tris/HCl pH 8. The cells were disrupted, centrifuged and ultracentrifuged as described above. The supernatant was frozen at −70°C and lyophilized. Published procedures were followed to isolate tetrapyrroles from this material as well as from purified sulphite reductase and to convert them to methyl ester derivatives (Matthews et al., 1995). Special care was taken to exclude oxygen during isolation steps: All solvents were deoxygenated by prolonged purging with high purity nitrogen gas and kept under a nitrogen atmosphere. Chromatography fractions were dried under vacuum or by a gentle stream of dry nitrogen gas. In between purification steps, samples were stored under nitrogen at −25°C. Exposure to direct sunlight was never allowed and the time samples were exposed to normal fluorescent lab light was minimized. In past experience with sirohydrochlorin-like tetrapyrroles, these precautions have always been sufficient to minimize the formation of oxidized derivatives such as the lactone (Fig. 1, structure 4). However, small amounts of oxidized derivatives were inevitably produced, since even the most rigorous protocols cannot absolutely guarantee the elimination of all traces of oxygen. Esterified material was partially purified by chromatography on low resolution columns (c. 1 × 10 cm of preparative grade silica gel, Aldrich Chemical Co. grade 951) and high-performance liquid chromatography [HPLC; Whatman 5 μm silica gel in a column of 4.6 × 250 mm eluted at 1 mL min−1 with 98 : 2 ethyl acetate : methanol containing 0.025% (v/v) pyridine]. Samples that had excess impurity peaks in their mass spectra were rechromatographed by HPLC to improve purity. Fractions were analyzed by thin layer chromatography on silica gel plates (EM Science Silica Gel 60, 250 μm thick) developed with 9 : 1 ethyl acetate : methanol. Mass spectra were obtained on a Bruker Apex III matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometer under conditions reported previously for tetrapyrroles (Youn et al., 2002). The acidic conditions during the applied isolation procedure are sufficient to de-metallate haem forms.

Figure 1.

 Relevant structures. 1, siroamide; 2, stereo isomer of siroamide; 3, hydroxylated derivative of siroamide; 4, lactone derivative of siroamide. Important differences in the structures are highlighted by boxes. Atoms C3 and C8 are numbered.

Protein techniques

Immunoblot analysis and sodium dodecyl sulphate polyacrylamide gel electrophoresis were performed as described earlier (Dahl et al., 2005). Glutamate synthase activity of recombinant DsrL was assayed spectrophotometrically by measuring the initial rate of NADPH oxidation at 340 nm as described in Jongsareejit et al. (1997). UV/Vis spectra of samples in 1 mL quartz cuvettes were recorded using an Agilent Technologies (Böblingen, Germany) 8453 diode array spectrophotometer.

Growth experiments

Photolithoautotrophic growth of A. vinosum wild-type and mutant strains was followed in batch culture under continuous illumination essentially as described by Prange et al. (2004) in a medium containing carbonate and sulphide as the sole sulphur compound (Pfennig & Trüper, 1992). One liter of a photoheterotrophically grown stationary-phase culture was harvested (5900 g, 20 min) and the cell material was used to inoculate 1.5 L of sulphide-containing medium in a thermostatted fermenter. Sulphur compounds were determined as described in Prange et al. (2004).

Results and discussion

Detection of siroamide in soluble extracts and in highly purified sulphite reductase from A. vinosum

Soluble extract of A. vinosum grown photolithoautotrophically on sulphide and thiosulphate was analyzed for the presence of siroamide. Formation of ferredoxin-dependent assimilatory sulphite reductase (CysI) is avoided under the chosen growth conditions because cysI is strictly negatively regulated by reduced sulphur compounds (Neumann et al., 2000). When examined by thin layer chromatography with fluorescence detection the isolated material behaved similar to pure siroamide (structure 1, Fig. 1) from Desulfovibrio vulgaris (Matthews et al., 1995) except that the low resolution chromatograms indicated multiple closely spaced peaks for the A. vinosum sample. Increased resolution obtained by HPLC (Fig. 2) indicated the possibility that the A. vinosum material contained multiple isomers of siroamide ester. Note that siroamide is more polar than sirohydrochlorin methyl ester (Matthews et al., 1995); it is chromatographically well resolved from sirohydrochlorin ester and no peak in Fig. 2 is sirohydrochlorin ester. Fractions 2–7 all had visible spectra indistinguishable from D. vulgaris siroamide ester. The retention time of fraction 4 matched D. vulgaris siroamide (structure 1, Fig. 1) (Matthews et al., 1995). The sirohydrochlorin tetrapyrrole undergoes a spontaneous epimerization at either one or both of C3 and C8 during its isolation (see Burkhalter & Timkovich, 1997 and reference therein to the earlier literature) to produce isomers such as that shown in structure 2 (Fig. 1). Different isomers exhibit slightly different retention time in chromatographic systems. In this regard, fraction 2 of the A. vinosum material matched an epimer of D. vulgaris siroamide (Matthews et al., 1995).

Figure 2.

 HPLC chromatogram of the siroamide material isolated from the soluble cell extract of Allochromatium vinosum (bottom trace). Fractions labeled 2–7 had visible absorbance spectra typical of siroamide (Matthews et al., 1995), while other earlier fractions did not have tetrapyrrole-like visible spectra. The upper trace is siroamide isolated from Desulfovibrio vulgaris analyzed under the same conditions (Matthews et al., 1995).

The crude yield of total siroamide was 69 nmol from 250 g wet mass of A. vinosum cells, estimated by visible spectroscopy. Desulfovibrio vulgaris yielded 70 nmol g−1 wet cell mass (Matthews et al., 1995). The relatively low yield from A. vinosum probably results from the removal of membranes before isolation of siroamide. While this step was imperative because of the high bacteriochlorophyll-content of membranes of this phototrophic bacterium, a substantial part of sulphite reductase was lost as it is associated with membrane-bound proteins in vivo (Dahl et al., 2005).

Confirmation of the identification of siroamide ester was accomplished by MALDI-TOF mass spectrometry in a matrix of dihydroxybenzoic acid that inevitably produces the (M+H)+ ion as the parent. Siroamide is unusual among tetrapyrroles in that it follows the odd-nitrogen rule: an organic molecule with an odd number of trivalent nitrogen atoms has an odd integer mass, so, the protonated molecular ion has an even mass. In contrast, most common tetrapyrroles have an odd mass for (M+H)+. Fractions 2, 3, and 4 all showed the characteristic mass of 960 for siroamide heptamethyl ester (Matthews et al., 1995) as shown in Fig. 3.

Figure 3.

 MALDI-TOF mass spectrum of Allochromatium vinosum siroamide. A cluster of ion peaks differing by 1 Da are typically seen in this type of spectrum . The (M+H)+ peak is always the most intense in this matrix of dihydroxybenzoic acid. Higher mass peaks reflect the presence of 13C.

An unexpected result was that fractions 5–7 demonstrated a higher mass of 976 for (M+H)+. The even mass and the characteristic chromatographic behaviour indicate the presence of the amide functional group. A reasonable hypothesis is that fractions 5–7 are isomers of a hydroxylated derivative of siroamide as shown in structure 3 (Fig. 1). It is well known (Battersby et al., 1977) that sirohydrochlorin exposed to air during its isolation from natural sources can be oxidized at C3 and/or C8 to form an alcohol that then spontaneously undergoes dehydration to form the lactone 4 (Fig. 1), in an entropically favoured reaction due to the fortuitous presence of the neighbouring acetate group. Lactones are more difficult to form and hence the A. vinosum fractions may stop at the alcohol-amide stage. The isolation of the A. vinosum material was performed under anoxic conditions, but in our hands it is still typical to encounter small amounts of lactone forms while isolating sirohydrochlorin. The presence of multiple isomers of A. vinosum siroamide may therefore be due to spontaneous epimerization and accidental hydroxylation via oxidation. Another possibility is that sirohaem amidation does not occur rigorously stereospecific, but that amides are formed at the other acetates or potentially even at propionate side chains.

To confirm that the observed siroamide is the prosthetic group of the dsrAB-encoded dissimilatory sulphite reductase we analyzed highly enriched protein for siroamide. Indeed, 3 nmol of siroamide ester were identified in a sample of 4.5 mg protein. Yield of siroamide was relatively low as 45.7 nmol had been expected on the basis of UV/Vis spectroscopy (Young & Siegel, 1988). While isolating non-covalently bound haems from haemoproteins, low yields can occur from precipitation of the protein in the heterogeneous extraction mixture and physical occlusion of otherwise soluble tetrapyrroles. The main tetrapyrrole fraction from the protein extract had an HPLC retention time close to fractions 5/6 of the cell extract and a mass of 976 for (M+H)+.

A widespread arrangement of dsr genes in sulphate/sulphite-reducing bacteria is dsrABDN (Zverlov et al., 2005) and the frequent close connection of dsrN to dsrAB supports the postulated function of DsrN as sirohaem amidase. It may furthermore be noteworthy that other sirohaem biosynthesis genes including sirohaem synthase (uroporphyrinogen III methyltransferase) and uroporphyrinogen III synthase are located in the vicinity of dsrAB encoding sulphite reductase in several instances (Eisen et al., 2002; Mussmann et al., 2005; Beller et al., 2006).

Construction, characterization and complementation of A. vinosum strain with in frame deletions of dsrN and dsrL

Allochromatium vinosum strains with in frame deletions in dsrN and dsrL were constructed. In accordance with earlier studies on dsr interposon mutants (Pott & Dahl, 1998) the sulphide oxidation rate of the mutants was not influenced in media containing 2–7 mM sulphide (not shown). However, the ΔdsrL mutant was completely unable to oxidize sulphur which is formed by the wild type as an intracellular intermediate during the oxidation of sulphide to sulphate. (Fig. 4). The phenotype of the ΔdsrN mutant was different as it was still able to oxidize sulphur (Fig. 5) and to produce sulphate (not shown) albeit with a significant delay. In media containing an initial sulphide concentration of 2 mM, the sulphur oxidation rate of the ΔdsrN mutant [1.2±0.2 nmol min−1 (mg protein)−1] was reduced to ∼15% of the oxidation rate of the wild type [7.4±0.5 nmol min−1 (mg protein)−1]. The same effect was observable on media containing higher (7 mM) sulphide concentrations [28±1 vs. 3.7±0.5 nmol sulphur oxidized min−1 (mg protein)−1 in wild type and mutant, respectively]. Growth under photoorganoheterotrophic conditions was not influenced by either one of the mutations (data not shown).

Figure 4.

 Sulphur accumulation and oxidation by Allochromatium vinosum wild type (▪) and A. vinosumΔdsrL (○) and A. vinosumΔdsrL+dsrL (◊). Cells were grown photolitoautotrophically in batch culture in the presence of 2 mM sulphide. Protein concentrations at the onset and end of the experiments 140–170 μg mL−1. Representatives of three independent growth experiments for each strain are shown.

Figure 5.

 Sulphur accumulation and oxidation by (a) Allochromatium vinosum wild type (▪) and A. vinosumΔdsrN (○) and (b) A. vinosum wild type (▪) and A. vinosumΔdsrN+dsrN (◊). Cells were grown photolitoautotrophically in batch culture in the presence of (a) 2 mM or (b) 4 mM sulphide. Protein concentrations at the onset and end of the experiments 130–160 μg mL−1. Representatives of three independent growth experiments for each strain are shown.

Complementation of dsrN and dsrL in trans restored the mutant sulphur oxidation rates to those of the wild type confirming that the observed phenotypes were indeed a consequence of the specific loss of dsrN and dsrL, respectively (Figs 4, 5). A short delay was observed in sulphur accumulation for the complemented dsrN mutant that cannot be explained at present. In A. vinosum DsrN and DsrL are soluble 50 and 70 kDa proteins, respectively, as predicted from the nucleotide sequence (Dahl et al., 2005 and Fig. 6). Western blot analysis with antisera against DsrN and DsrL demonstrated the absence of the proteins in the respective A. vinosum mutants and their presence in the wild-type and the complemented strains (Fig. 6). We analyzed A. vinosum DNA for the presence of further dsrN homologous genes by low stringency Southern hybridization. No other dsrN gene was detected. This is additional evidence that the deletion of dsrN is the cause of the detected phenotype.

Figure 6.

 Immunological detection of DsrN (lanes 1–4) and DsrL (lanes 5–7) in Allochromatium vinosum wild-type and mutant strains. Lanes: 1, fraction after hydrophobic interaction chromatography of soluble extract from A. vinosum wild type; lanes 2–7, crude extracts (200 μg protein per lane) of A. vinosum wild type (lane 2), A. vinosumΔdsrN (3) A. vinosumΔdsrN+dsrN (4), A. vinosum wild type (5), A. vinosumΔdsrL (6), A. vinosumΔdsrL+dsrL (7). Both antisera were raised against oligopeptides comprising a highly immunogenic epitope deduced from the nucleotide sequence and their specific reaction with DsrN and DsrL produced in Escherichia coli was proven (Dahl et al., 2005).

Sulphite reductase was partially purified from the ΔdsrN mutant. Fractionation patterns during chromatographic steps and final yield of ∼80% pure protein were virtually identical to those of the wild type. Furthermore, sulphite reductase isolated from the mutant showed the characteristic sirohaem spectrum (Schedel et al., 1979). It should be noted however, that the presence of the additional amide-group in the siro(haem)amide molecule is not visible in UV/Vis spectra (Matthews et al., 1995). It would of course be highly desirable to obtain direct evidence for or against an amidated haem in cell material or sulphite reductase of the ΔdsrN mutant. However, with regard to the already low yields of siroamide from wild-type cells and enzyme, negative findings for the mutant would not have provided any really conclusive evidence.

Does DsrL function as a glutamate synthase in A. vinosum?

Pure, recombinant DsrL was tested for glutamate synthase activity, however, the protein was unable to catalyze the reaction. Complementation of the glutamate synthase-deficient E. coli strain EB5001 with dsrL was a second approach for assessing the potential glutamate synthase activity of DsrL. Escherichia coli EB5001 is unable to grow on L-arginine as the sole nitrogen source because the cells starve for glutamate (Goss et al., 2001). Although the DsrL protein was clearly detectable by Western blotting in E. coli EB5001 containing dsrL in the pBAD22a vector after induction with arabinose (data not shown), the cells did not regain their ability to grow on minimal medium with arginine.

The dsrL gene product is indispensable for sulphur oxidation in A. vinosum. It therefore appears unlikely that DsrL provides substrate for DsrN, a protein that is itself not absolutely essential for this metabolic process. Taking together all currently available experimental evidence we conclude that the A. vinosum dsrL gene product is not functional as a glutamate synthase. An involvement in siroamide biosynthesis by delivering glutamine for the amidation reaction is highly improbable.

It should be noted that proteins homologous to the small subunits of glutamate synthases are often misannotated as glutamate synthases (Stutz & Reid, 2004). Furthermore, the supposed monomeric glutamate synthase of T. kodakarensis (Jongsareejit et al., 1997) is homologous to the small subunit of the NADPH-dependent ferredoxin oxidoreductase of Pyrococcus furiosus. On the basis of these findings and an analysis of its amino acid sequence, DsrL more likely works as a NAD(P)H:acceptor oxidoreductase (Dahl et al., 2005).


We provide evidence that the prosthetic group of the dissimilatory sulphite reductase from A. vinosum is siro-(haem)amide as it has been found for the dissimilatory sulphite reductase from Desulfovibrio species (Matthews et al., 1995). This is indirect proof for the proposed function of DsrN as siro(haem)amidase. Allochromatium vinosum mutants lacking dsrN show a severe impairment in the oxidation of intracellularly stored sulphur. The presence of DsrN in sulphate/sulphite-reducing bacteria and in sulphur-oxidizing bacteria as well as the pronounced effect of the dsrN deletion in A. vinosum point at an important function of the DsrN protein in dissimilatory sulphur metabolism. While DsrL is absolutely essential for sulphur oxidation, it does not appear to be involved in siroamide biosynthesis. We suggest siro-(haem)amide as the general prosthetic group of dissimilatory sulphite reductases, irrespective of whether the enzymes catalyze the reductive step from sulphite to sulphide or operate in the opposite direction. The prosthetic group of assimilatory sulphite reductases is definitely sirohaem (Matthews et al., 1995), indicating that siro(haem)amide plays a special role in dissimilatory sulphur metabolism.


This work was supported by grants Da 351/3-1 and Da 351/3-3 from the Deutsche Forschungsgemeinschaft. We thank Hans G. Trüper for continuing interest and support. The group of Prof. Bender, Department of Biology, University of Michigan, Ann Arbor, is gratefully acknowledged for providing E. coli EB5001.


This study is dedicated to Prof. Dr. Dr. Dr. h.c. Hans G. Trüper on the occasion of his 70th birthday.