A Sulfite Respiration Pathway from Thermus thermophilus and the Key Role of Newly Identified Cytochrome c₅₅₀

Abstract

Sulfite, produced for instance during amino acid metabolism, is a very reactive and toxic compound. Various detoxification mechanisms exist, but sulfite oxidoreductases (SORs) are one of the major actors in sulfite remediation in bacteria and animals. Here we describe the existence of an operon in the extreme thermophilic bacterium Thermus thermophilus HB8 encoding both a SOR and a diheme c-type cytochrome. The in vitro analysis clearly showed that the newly identified cytochrome c₅₅₀ acts as an acceptor of the electrons generated by the SOR enzyme during the oxidation of sulfite. The electrons are then rapidly shuttled via cytochrome c₅₅₂ to the terminal ba₃- and caa₃-type oxidases, thereby unveiling a novel electron transfer pathway, linking sulfite oxidation to oxygen reduction in T. thermophilus: sulfite → SOR_HB8 → cytochrome c₅₅₀ → cytochrome c₅₅₂ → ba₃ oxidase/caa₃ oxidase → O₂. The description of the complete pathway reveals that electrons generated during sulfite oxidation by the SOR are funneled into the respiratory chain, participating in the energy production of T. thermophilus.

INTRODUCTION

Naturally appearing in the environment, sulfite (SO₃²⁻) can also be generated within living cells. As the main intermediate in the oxidation of sulfur compounds to sulfate, sulfite plays a key role in oxidative sulfur metabolism (24). Sulfite is a highly reactive compound, and its ability to react with biomolecules such as proteins or DNA makes it very toxic for living organisms. Detoxification of sulfite is crucial for living cells, and both prokaryotic and eukaryotic organisms have developed efficient mechanisms to neutralize this compound (24).

In bacteria, sulfite oxidation is known to occur either by indirect, AMP-dependent oxidation via the intermediate adenylylsulfate (24) or by direct oxidation, usually utilizing a molybdenum-containing sulfite:acceptor oxidoreductase (SOR) (12). The direct oxidation route is much more prevalent, but the two mechanisms are present simultaneously in many organisms, including Betaproteobacteria and Gammaproteobacteria. SORs, the main actors of the direct oxidation, have been identified in mammals (5), birds (27), plants (10), and prokaryotes (23). They play a key role in sulfur detoxification, notably in the final step of the degradation pathway of sulfur-containing amino acids. In bacteria and archaea, they can also utilize sulfur compounds as an electron source for chemotrophic, lithotrophic, and phototrophic growth (13, 32). SORs catalyze the two-electron oxidation of sulfite to sulfate (SO₃²⁻ + H₂O → SO₄²⁻ + 2 H⁺ + 2 e⁻), where the final electron acceptor is O₂ and/or heme iron ions (26). Most SORs possess a heme relay, shuttling electrons from the molybdenum cofactor to the subsequent acceptor, while in plants, for instance in Arabidopsis thaliana, no heme relay is present and O₂ is the direct and exclusive final electron acceptor (10). In mammals and birds, the heme relay is an integral part of the SOR (1, 28, 38). In bacteria, however, the relay, if present, consists of an additional c-type cytochrome subunit (23).

To date, bacterial SORs have been classified in two main groups. Group 1 SORs consist of a complex of two subunits, one containing the molybdenum cofactor and the other containing the heme center, while group 2 SORs contain only the molybdenum cofactor (22).

The most representative SOR from group 1 is the extensively characterized SorAB sulfite dehydrogenase (SDH) from Starkeya novella (23). Both SorA and SorB subunits of SorAB are encoded by a single operon, where the cytochrome c-like electron acceptor is located directly downstream of the sulfite oxidoreductase (22, 23). In vitro, SorAB utilizes horse heart cytochrome c as well as ferricyanide as an electron acceptor, but no direct association of this enzyme with a dissimilatory sulfur oxidation pathway has been described.

To date, group 2 SORs (containing solely the molybdenum cofactor) have been identified in Cupriavidus necator (7), Sinorhizobium meliloti (39), Deinococcus radiodurans (8), and Delftia acidovorans (7). For those so-called “atypical” SORs, it has been shown that horse heart cytochrome c, routinely used as an electron acceptor for characterization of sulfite oxidoreductases, cannot be efficiently utilized (8, 39), a finding attributed presumably to the lack of a heme relay.

Similar to SOR members of group 1, group 2 SORs from C. necator, D. acidovorans and S. meliloti are also encoded by an operon containing putative cytochrome c-encoding genes. These enzymes, however, are isolated without an additional c-type cytochrome subunit (6, 39). Conversely, Denger and coauthors demonstrated that the group 2 SOR from Cupriavidus necator H16 is coexpressed with the membrane-anchored c-type cytochrome SorB, encoded within the same operon and likely acting as the final electron acceptor of sulfite oxidation (6).

For a long time, the ability of Thermus thermophilus to metabolize sulfur was unrevealed. Recently, however, Di Salle and coauthors isolated and characterized a SOR from T. thermophilus (9). This enzyme belongs to group 2, as it contains only the molybdenum cofactor. Like other members of this group, the SOR from T. thermophilus cannot reduce horse heart cytochrome c, but it can actively reduce ferricyanide. Interestingly, its operon contains a gene annotated as a putative c-type cytochrome, located downstream of the SOR-encoding gene. This finding suggests a possible role for this putative cytochrome c as an electron acceptor from SOR. This function, so far not demonstrated, if it exists would shed light on the mechanism and role of the SOR from T. thermophilus. In the present study, the function of a novel c-type cytochrome from T. thermophilus was investigated by biochemical and biophysical means, and its key role in coupling sulfite oxidation to cell respiration in this extremophile was demonstrated. This newly revealed pathway unfolds the complexity of the electron transfer resulting from sulfite oxidation, which could be extended to sulfur metabolism in other organisms.

MATERIALS AND METHODS

Fermentation.

Fermentation of T. thermophilus HB8 was performed at the Helmholtz Centre for Infection Research, Braunschweig, Germany. Cells were grown at 70°C in 100 liters of Castenholtz medium (4) in a stainless steel jar fermentor under restricted oxygen conditions (0.05 volumes of air per volume of medium per minute), yielding ∼500 g biomass. Under these conditions, the bacteria expressed the highest yield of the ba₃ oxidase and cytochrome c₅₅₂ in addition to the caa₃ oxidase and other cytochromes. Cells were harvested in the early to mid-exponential growth phase and stored at −80°C.

Isolation of cytochrome c₅₅₀ from the periplasmic space of T. thermophilus HB8.

Aliquots (100 g) of frozen cells were thawed in 500 ml of 100 mM Tris-HCl (pH 7.6) containing 100 mM KCl and were homogenized for 1 min at full speed in a Braun (type MX-32) homogenizer. After 30 min of stirring, the suspension was centrifuged for 30 min at 17,700 × g at 4°C. The supernatant was diluted with H₂O to a conductivity lower than 2 mS and loaded onto an S-Sepharose column (GE Healthcare, Germany) equilibrated with 10 mM Tris-HCl (pH 7.6; buffer A). After washing with 200 ml of buffer A, proteins were eluted with a linear salt gradient of buffer A (0 to 500 mM NaCl). Fractions containing cytochrome c₅₅₀ were pooled, concentrated (Centricon 10 kDa; Millipore), and purified by anion exchange chromatography on a Fractogel TMAE 650(S) apparatus (Merck, Germany). The protein was bound to the anion exchanger equilibrated with buffer A. The column was washed again with 10 column volumes with buffer A, and the protein was then eluted with a linear 0-to-200 mM NaCl gradient in buffer A at a flow rate of 2 ml/min over 30 min. The last purification step was performed by gel filtration on Superdex 75 (GE Healthcare, Germany). The protein solution was concentrated and applied to the column, which was equilibrated with 50 mM Tris-HCl (pH 7.6). The cytochrome c₅₅₀ fractions were collected, concentrated by ultrafiltration, and used for further biochemical characterization.

Purification of ba₃ and caa₃ cytochrome c oxidases and cytochrome c₅₅₂.

Native ba₃- and caa₃-type cytochrome c oxidases were isolated from T. thermophilus HB8 cells according to previously published procedures (16, 34). Native cytochrome c₅₅₂ was purified according to procedures described by Soulimane and coauthors (35). Purified proteins were concentrated by ultrafiltration, fast frozen in liquid nitrogen, and stored at −80°C.

Determination of cytochrome and cytochrome oxidase concentrations.

Absorption spectra were recorded with a Perkin-Elmer Lambda 5 UV/VIS spectrophotometer. Heme c concentrations were obtained from the dithionite reduced-minus-oxidized spectrum at 550 nm using the molar absorption coefficient ε of 18,000 M⁻¹ cm⁻¹ determined from spectrophotometric and amino acid analysis data. Similarly, absorption coefficients (ε) of 21,000 M⁻¹ cm⁻¹ for cytochrome c₅₅₂, 6,300 M⁻¹ cm⁻¹ for ba₃ oxidase, and 24,000 M⁻¹ cm⁻¹ for caa₃ oxidase were used.

Amino acid analysis and Edman degradation.

Quantitative amino acid determination was performed on a Biotronik amino acid analyzer LC 5001 equipped with a fluorescence detector for the detection of o-phtalaldehyde derivatives. The protein sample was hydrolyzed in 5.7 N HCl containing 0.2% thioglycolic acid for 24 h or 72 h at 108°C. Tryptophan was determined after hydrolysis in 5.7 N HCl containing 5% thioglycolic acid for 24 h at 108°C. Cysteine contents were determined as carboxymethylcysteine after reduction with dithioerythritol and by oxidation of the entire protein to cysteinsulfonic acid with performic acid before hydrolysis. Automated Edman degradation was carried out in a Knauer 910 gas/liquid-phase protein sequencer on polyvinylidene difluoride (PVDF) membranes with autoconversion and with on-line high-performance liquid chromatography (HPLC) identification of the phenylthiohydantoin amino acids.

Cleavage with cyanogen bromide.

Methionyl bonds were cleaved with cyanogen bromide by using the method of Gross and Morell (17). Forty nanomoles of protein was dissolved in formic acid and mixed with the 400 M excess of cyanogen bromide, and the reaction proceeded in the dark for 3 h. The reaction mixture was then evaporated under reduced pressure. Peptides were purified on a reversed-phase Synchropak C₄ column (Agilent Technologies) using a Hewlett-Packard 1050 HPLC system with a multiple wavelength detector. Gradients were formed with the aid of a microprocessor-controlled valve by low-pressure mixing of the following solutions: (A) 5% formic acid, 5% acetonitrile, 90% water (vol/vol); (B) 20% formic acid, 80% acetonitrile (vol/vol).

Construction of cytochrome c₅₅₀ expression plasmid.

The sequence encoding the mature cytochrome was amplified from T. thermophilus HB8 genomic DNA by PCR using the primers 5′-ATCTGACCATGGCTCAGACCACCCTCCCCGAG-3′, containing a NcoI restriction site (underlined), and 5′-CAGTGACTCGAGTCATGGCAGTTTGAGGCCTTGGCGGAG-3′, containing a XhoI restriction site (underlined). The product was digested with NcoI and XhoI and inserted into the expression vector pET22b+ (Invitrogen), digested by the same restriction enzymes, to yield pET22bC550 vector. This construct permits the expression of the recombinant cytochrome fused with the PelB leader sequence for optimal translocation to the periplasmic space in Escherichia coli.

Expression and purification of recombinant cytochrome c₅₅₀.

The expression vector and pEC86 (2) were cotransformed in BL21(DE3) E. coli strain cells. pEC86 encodes the cytochrome maturation gene cluster from E. coli necessary for the production of cytochrome c in this bacterium under aerobic conditions (19). The cytochrome was produced in LB medium containing ampicillin (100 μg/ml) and chloramphenicol (34 μg/ml) at 37°C for 24 h under shaking without induction. Periplasmic proteins were prepared from fresh biomass. Cells were washed in phosphate-buffered saline (PBS) buffer (20 mM phosphate, 135 mM NaCl, 1 mM KCl [pH 7.4]) and spun down at 8,000 × g for 20 min at 4°C. The pellet was resuspended in 100 mM Tris-HCl (pH 8) buffer containing 0.75 M sucrose. Osmotic shock was induced by slowly adding 2 volumes of ice-chilled 1 mM EDTA. Following 10 min of incubation at room temperature, spheroplasts were prepared by incubation with 1 mg/ml lysozyme for 45 min at room temperature under gentle shaking. Spheroplasts were returned in suspension by addition of 25 mM MgCl₂, and DNase I was added at a final concentration of 50 μg/ml to reduce the viscosity of the extract. Intact spheroplasts were removed by centrifugation at 8,000 × g for 10 min at 4°C. The supernatant containing the cytochrome c was extensively dialyzed against 5 mM Tris-HCl (pH 8.0) and then loaded on a Fractogel TMAE 650(S) column (Merck, Germany) equilibrated at 4°C with the same buffer. The cytochrome was eluted with a gradient of NaCl (0 to 250 mM), the fractions containing the protein were pooled and concentrated, and the sample was desalted using a PD10 column (GE Healthcare, Germany) equilibrated with 5 mM Tris-HCl (pH 8.0). The eluate was then purified on a CaptoQ XL anion exchange column (GE Healthcare, Germany) equilibrated with the same buffer and eluted with a gradient of NaCl (0 to 150 mM). Fractions containing the cytochrome were pooled, concentrated, and finally purified by gel filtration on a Superdex 75 column (GE Healthcare, Germany) at 4°C with 5 mM Tris-HCl (pH 8.0). Cytochrome c₅₅₀ was concentrated by ultrafiltration, fast-frozen in liquid nitrogen, and stored at −80°C.

Construction of SOR expression plasmid.

The sequence encoding the mature TTHA1325 was amplified from T. thermophilus HB8 genomic DNA by PCR using the primers 5′-CAGCAAGCCCCCACGGCCGACCAGTTG-3′ and 5′-GAAGATCTGGCCACGGTGAACTTGACCCGCATG-3′, containing a BglII restriction site (underlined). The expression vector pQE-60 (Qiagen, Germany) was digested by BglII and NcoI, and the latter was blunt ended by the use of the DNA blunting kit (Takara Bio Inc., Japan). The PCR product was digested with BglII and inserted into pQE-60 to yield the pQE-60-SO2 vector. This construct permits the expression of the recombinant His-tagged SOR in the cytoplasm of E. coli.

Expression and purification of the recombinant SOR.

Expression was conducted in E. coli TP1000 cells (36) transformed with pQE-60-SO2. Cells were grown in LB medium containing ampicillin (100 μg/ml), kanamycin (25 μg/ml), and Na₂MoO₄ (1 mM) at 37°C. At an optical density (OD) of 0.5, the temperature was lowered to 22°C and expression was induced with 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) for 3 h. The cells were harvested, and proteins were extracted in buffer A (20 mM phosphate, 100 mM NaCl [pH 7.4]) with the addition of 5% (vol/vol) CelLytic reagent (Sigma-Aldrich), 1 mg/ml lysozyme (Sigma-Aldrich), and 50 μg/ml DNase I (Sigma-Aldrich)/ml. After centrifugation at 12,000 × g for 30 min at 4°C, the solution was filtered, and NaCl and imidazole were added up to 500 mM and 10 mM, respectively. The extract was loaded onto a Superflow Ni²⁺ immobilized metal affinity chromatography column (GE Healthcare, Germany). The resin was then washed with buffer A containing 500 mM NaCl and imidazole from 20 to 500 mM, and fractions containing eluted SOR were pooled. After concentration, the protein was finally purified by gel filtration on a Hiload Superdex 200 column (GE Healthcare, Germany) at 4°C with 10 mM Tris-HCl (pH 8). Purity was assessed by SDS-PAGE and Western blotting, using antibodies directed against the His tag. The SOR was further concentrated, fast-frozen in liquid nitrogen, and stored at −80°C.

Electron transfer activity.

O₂ consumption assays were carried out at 37°C in 100 mM Tris-HCl (pH 8.0), 100 μM EDTA, 0.1% DDM (n-dodecyl β-d-maltoside) (vol/vol), using a high-resolution respirometer (Oxygraph-2k; Oroboros Instruments) equipped with two 1.5-ml chambers with thermostats. Stopped-flow experiments were carried out with a thermostat instrument (DX.17MV; Applied Photophysics, Leatherhead, United Kingdom) equipped with a 1-cm-path-length observation chamber. Reactions were investigated by monitoring the absorption changes at selected wavelengths. Kinetic measurements were carried out in 5 mM bis-Tris (pH 7.0); the buffer was supplemented with 0.1% DDM in experiments with ba₃ or caa₃ oxidase. When necessary, ionic strength was adjusted by addition of KCl, and the buffer was degassed with vacuum/N₂ cycles. Data were analyzed using the software MATLAB (Mathworks, South Natick, MA). The kinetics of electron transfer between cytochrome c₅₅₀ and cytochrome c₅₅₂ was assayed under non-pseudo-first-order conditions, i.e., at comparable concentrations of the two proteins. Observed rate constants (k_obs) were therefore obtained by fitting the experimental time courses to the equations described in reference 30 for the analysis of bimolecular reactions assayed under second-order conditions. According to reference 20, the forward (k_F) and reverse (k_R) second-order rate constants for the electron transfer reaction from cytochrome c₅₅₀ to cytochrome c₅₅₂ were determined from the dependence of k_obs on the concentration of reduced cytochrome c₅₅₀, using the equation in which c₅₅₀ and c₅₅₂ are the concentrations of the two c-type cytochromes in the experiment.

RESULTS

Genetic organization of a sulfite-oxidizing system from T. thermophilus HB8.

Examination of the T. thermophilus HB8 genome sequence led to the identification of the gene TTHA1325, putatively encoding a SOR (Fig. 1 A). Based on the predicted sequence, the encoded SOR_HB8 protein comprises 407 residues (mass, 49,963 Da; pI 9.33) and shows 51% sequence identity with the SORs from T. thermophilus AT62 (SOR_AT62) and HB27 (SOR_HB27) (9) and 33% identity with the well-characterized SOR from S. novella. The relatively low identity between SOR_HB8 and SOR_HB27 was unexpected, given that the average protein sequence identity between the T. thermophilus HB27 and HB8 strains is 97.6% (3). Motif analysis using the PFAM program (Protein Families database of Alignments and HMMs; http://www.sanger.ac.uk) revealed the presence of a molybdopterin cofactor (MoCo) binding domain, from residues 77 to 255, and a MoCo oxidoreductase dimerization domain, from residues 287 to 407. The TTHA1325 sequence did not show a heme-binding site, in line with SOR_HB27 or SOR_AT62 and all the bacterial SORs described to date.

Fig. 1.

Genomic arrangement of the putative SOR genes in T. thermophilus HB8 and multiple sequence alignment of cytochromes c₅₅₀. (A) The presence of a promoter and a terminator putative region suggests that the genes form an operon. Promoter, predicted −35 and −10 regions as well as the start of transcription (shown in bold). Terminator, U-rich region (depicted in bold); the predicted hairpin loop is shown. (B) Multiple sequence alignment of the cytochrome c₅₅₀ from T. thermophilus HB8 and HB27, the cytochrome c₅₅₂ from T. thermophilus HB8, and the sulfite:cytochrome c oxidoreductase subunit B Cj0004c from C. jejuni NCTC 11168 (CC). The predicted leader sequence is underlined. Residues determined by internal sequencing are shown in bold. Residues of the heme-binding sites are indicated by an asterisk, while methionine residues potentially involved in heme coordination are indicated by triangles.

Similar to T. thermophilus HB27, the gene TTHA1326 from T. thermophilus HB8, located downstream of the putative SOR gene, encodes a putative c-type cytochrome in which two heme-binding motifs can be easily identified (Fig. 1B). Interestingly, the c-type cytochromes from T. thermophilus HB27 and HB8 also exhibited a relatively low identity (41%) (Fig. 1B), far below the predicted average protein sequence identity between those two strains (3). Based on BLAST analysis, the TTHA1326 cytochrome c protein sequence as a whole showed poor similarity to known proteins. Nevertheless, when each heme-binding motif was analyzed separately, both N- and C-terminal domains of the protein exhibited high similarity to known c-type cytochromes. The N-terminal region, covering the first 106 residues and including the first heme-binding motif, showed 33% sequence identity to the subunit SorB Cj0004c of the aforementioned sulfite:cytochrome c oxidoreductase (SorA) from Campylobacter jejuni NCTC 11168. Based on this finding, the N-terminal domain of cytochrome c₅₅₀ may be expected to act as an electron acceptor for SOR_HB8. The remaining part of the sequence (two-thirds of the protein), hosting the second heme-binding site, showed ∼50% sequence identity (Fig. 1B) with the well-characterized cytochrome c₅₅₂ from T. thermophilus, the substrate of ba₃ and caa₃ cytochrome c oxidases (21, 35, 37), the two respiratory terminal oxidases of this organism (11, 40).

Downstream of this locus, a third gene is present (TTHA1327) that encodes a protein with a putative conserved MoCo sulfurase C-terminal domain (Fig. 1A). This β-strand-rich domain, identified in the molybdenum cofactor sulfurase, is predicted to be a sulfur carrier that receives sulfur abstracted by the pyridoxal phosphate-dependent NifS-like enzymes on its conserved cysteine and delivers it for the formation of diverse sulfur-metal clusters. The role of TTHA1327 and its potential role in the posttranslational modification of the SOR TTHA1325 are yet to be elucidated.

A unique promoter upstream of TTHA1325 and a unique terminator region downstream of TTHA1327 were identified using BPROM and FindTerm (Softberry), respectively. Those findings showed that the three genes are most likely organized in a single operon and suggested the coexpression of the c-type diheme cytochrome and SOR as an enzyme-acceptor pair.

Subcellular localization and purification of cytochrome TTHA1326.

Since SOR_AT62 was found to be a periplasmic protein (9), it is likely that SOR_HB8 is also located in the same compartment along with its putative diheme cytochrome c electron acceptor. Consistently, using the SignalP 3.0 program (http://www.cbs.dtu.dk/services/SignalP/), a signal peptide sequence was predicted both in SOR_HB8 and cytochrome TTHA1326, with cleavage sites between residues 28 and 29 and residues 21 and 22, respectively. This finding suggests that both proteins are likely to be translocated into the periplasmic space.

In order to confirm the subcellular localization of the cytochrome TTHA1326, the T. thermophilus periplasmic fraction, prepared as described in Materials and Methods, was subjected to sequential cation/anion exchange chromatographic steps. While the cytochrome c₅₅₂ was found to bind to the cation exchanger, a small quantity of a second colored protein corresponding to the novel c-type cytochrome could be identified in the periplasmic extract. A purified fraction of this cytochrome was obtained after removing a large amount of ferredoxin with two subsequent anion exchange chromatographic steps on Fractogel TMAE resin. The cytochrome was further purified by gel filtration, resulting in only one major peak of Gaussian distribution with an elution peak corresponding to a molecular mass of ca. 20 kDa. Due to the large amount of ferredoxin (29) partly coeluting with the c-type cytochrome, this purification procedure yields maximally only about 5 mg of purified protein from a 500-g biomass. SDS-PAGE of the purified protein showed a single red band at about 23 kDa, in agreement with the theoretical molecular mass (22,897 Da) of the mature TTHA1326 product. When the newly purified cytochrome was subjected to Edman degradation, no PTH amino acids were detected after several cleavage cycles, demonstrating that the protein applied on the PVDF membrane was N-terminally blocked. In parallel experiments, an equivalent aliquot was used for amino acid content determination to verify that the analyzed sample contained protein. For further sequence analysis, the cytochrome was cleaved with cyanogen bromide. The fragments were chromatographed by HPLC, yielding two major peptides, and N-terminal sequencing of both fragments allowed identification of the 81 amino acids depicted in bold in Fig. 1B, thereby confirming that the isolated protein was the cytochrome encoded by the ORF TTHA1326.

The localization of the TTHA1326 cytochrome c in the periplasmic space of T. thermophilus HB8, where SOR_HB8 is also likely to be expressed, further supports the hypothesis that this c-type cytochrome serves as the SOR's electron acceptor.

Characteristics of cytochrome c₅₅₀.

The absorption spectrum of the oxidized cytochrome c exhibits a maximum at 410 nm in the Soret band, whereas the spectrum of the dithionite-reduced protein shows a Soret band centered at 415 nm and a broad α-band with a maximum at 550 nm. Consequentially, the novel cytochrome will be herein referred to as cytochrome c₅₅₀.

As mentioned above, the cytochrome c₅₅₀ polypeptide contains two heme c-binding motifs (Fig. 1B). The purified protein shows a heme c/protein ratio of 94.8 nmol/mg, a value that is close to the theoretical value (95.3 nmol/mg) expected for two heme cofactors per protein molecule. Based on its high sequence identity to cytochrome c₅₅₂, whose three-dimensional crystal structure is available (35, 37), Met171 could be identified as the sixth heme c axial ligand for the second heme-binding site, whereas most likely Met64 is the sixth heme c axial ligand for the first heme c (Fig. 1B).

Recombinant production of cytochrome c₅₅₀.

Due to the low purification yield of the native protein, cytochrome c₅₅₀ was expressed recombinantly in E. coli BL21(DE3) under the control of T7 promoter. The protein was directed to the periplasmic space of E. coli by fusion with the PelB signal sequence, and its maturation was ensured using the plasmid pEC86, carrying cytochrome c maturation genes from E. coli (2). Induction of expression of the cytochrome with IPTG led to accumulation of large amounts of inclusion bodies in the cytoplasm. The protein was therefore expressed in E. coli without addition of IPTG, taking advantage of the basal expression from the T7 promoter for a prolonged period of time. After 16 h at 37°C, the periplasmic extract was prepared, and the processed cytochrome c₅₅₀ was purified by two anion exchange chromatographic steps and an additional gel filtration step. Size exclusion chromatography yielded two forms of the recombinant protein of ∼20 and ∼40 kDa. As confirmed by SDS-PAGE, after denaturation, the protein appeared as a homogeneous band at 23 kDa (data not shown), in accordance with the theoretical molecular mass. This band displays a brownish color typical of c-type cytochromes. The recombinant cytochrome is produced in E. coli in a monomeric and dimeric form. A similar phenomenon has been described during the expression of T. thermophilus cytochrome c₅₅₂ in E. coli (25), as Keightley and coworkers showed that a conformational isomer of the cytochrome is produced with a strong propensity to dimerize. Since the native cytochrome c₅₅₀ was purified from T. thermophilus HB8 as a monomer, only the fractions corresponding to the 23-kDa product were further investigated. The purified recombinant cytochrome c₅₅₀ expressed in E. coli exhibits the same UV-visible absorption spectrum of the native cytochrome.

Recombinant production and spectral properties of SOR_HB8.

In order to verify if SOR_HB8 and cytochrome c₅₅₀ can exchange electrons, the gene encoding the sulfite oxidase was cloned into the pQE-60 vector and the protein was expressed recombinantly in E. coli TP1000, essentially as described previously (8, 36). The enzyme was purified by immobilized metal affinity chromatography and an additional gel filtration step. The latter resulted in a single peak of Gaussian distribution corresponding to a homogeneous protein with a molecular mass of ∼80 kDa, twice the theoretical mass of the enzyme, suggesting that the isolated SOR_HB8 forms a homodimer (data not shown).

The UV-visible absorption spectrum of the oxidized protein showed a maximum at 350 nm and a broad shoulder between 479 nm, which can be assigned to the electronic transitions of molybdenum(VI)-cysteine and enedithiolate-molybdenum charge transfer (15), respectively.

Electron transfer from sulfite oxidoreductase to cytochrome c₅₅₀.

In order to investigate the kinetics of electron transfer between SOR_HB8 and cytochrome c₅₅₀, a stopped-flow analysis was carried out (Fig. 2). In this experiment, oxidized cytochrome c₅₅₀ was anaerobically mixed at 45°C with an excess of SOR, in the presence or absence of 2 mM sulfite, and the cytochrome reduction was monitored at 418 nm. Regardless of sulfite being present, cytochrome c₅₅₀ was found to be slowly reduced by SOR_HB8 according to a biphasic time course (Fig. 2) that, at first glance could be tentatively attributed to the presence of two heme cofactors in cytochrome c₅₅₀. Although a diheme cytochrome c serving as an electron acceptor for SOR has been also identified in Paracoccus pantotrophus (33), the role of the two heme moieties is yet unknown. The electron transfer from SOR_HB8 to cytochrome c₅₅₀ is slow (tens of seconds) but, interestingly, remarkably faster than the direct reduction of cytochrome c₅₅₀ by a large excess of sulfite. The relatively low electron transfer rate measured could possibly result from a greater intermetallic distance between the MoCo center in SOR_HB8 and the electron acceptor heme(s) in cytochrome c₅₅₀; however, we ruled out that the slow kinetics is due to the low ionic strength in the experiment, because even slower rates were measured upon increasing the ionic strength up to ∼100 mM (data not shown).

Fig. 2.

Reduction of cytochrome c₅₅₀ by SOR_HB8 in the presence of sulfite. Absorption changes collected after anaerobically mixing oxidized cytochrome c₅₅₀ (1 μM) with 2 mM sulfite in the absence (dashed line) or in the presence (continuous line) of ∼6 μM SOR_HB8. Temperature, 45°C.

Identification of the cytochrome c₅₅₀ electron acceptor.

Electrons derived from sulfite could in principle be funneled into the respiratory chain in order to sustain the high energy-yielding dioxygen reduction catalyzed by the terminal cytochrome c oxidases of T. thermophilus. With this regard, notably cytochrome c₅₅₀ shares a high sequence identity with cytochrome c₅₅₂, which is recognized as the physiological substrate of the ba₃ and caa₃ cytochrome c oxidases. This prompted us to test if cytochrome c₅₅₀ can act directly as an electron donor for these oxidases. The hypothesis was tested by stopped-flow mixing at 25°C of an anaerobic solution of cytochrome c₅₅₀, prereduced by ascorbate, with air-equilibrated buffer containing either ba₃ or caa₃ cytochrome c oxidase (Fig. 3, dotted traces). Under these conditions, the oxidation of cytochrome c₅₅₀, as monitored at 418 nm, was very slow; despite its high sequence similarity with cytochrome c₅₅₂, we thus conclude that cytochrome c₅₅₀ is a poor electron donor for both T. thermophilus terminal oxidases. Surprisingly, addition of catalytic concentrations (≤10 nM) of cytochrome c₅₅₂ resulted in a remarkably faster reoxidation of cytochrome c₅₅₀ by ba₃ or caa₃ oxidase (Fig. 3). The stimulatory kinetic effect was shown to be proportional to the concentration of cytochrome c₅₅₂ and more pronounced with caa₃ oxidase than with ba₃ oxidase, possibly reflecting a faster electron transfer from cytochrome c₅₅₂ to caa₃ oxidase.

Fig. 3.

Cytochrome c₅₅₂ mediates cytochrome c₅₅₀ oxidation by ba₃ or caa₃ oxidase, as shown in the time course of cytochrome c₅₅₀ oxidation by ba₃ (A) or caa₃ (B) oxidase in the absence (dashed traces) or presence (solid traces) of increasing concentrations of cytochrome c₅₅₂. Temperature, 25°C; λ, 418 nm. Prior to the experiment, cytochrome c₅₅₀ was reduced with 300 μM ascorbate. Concentrations after mixing: [c₅₅₀] = 500 nM; [ba₃] and [caa₃] = 100 nM; [c₅₅₂] = 0, 0.2, 0.4, 1.0, 2.0, 5.0, and 10 nM (A) or 0, 0.2, 0.4, 1.0, 2.0, and 5.0 nM (B), from right to left. (Insets) Linear dependence on cytochrome c₅₅₂ concentration of the reciprocal of the reaction half-time.

In these experiments the measured cytochrome c₅₅₀ oxidation time course did not follow a single exponential decay, but it rather seemed to obey zero-order kinetics (constant rate), as if steady-state-like conditions were settled. All together, these results suggest that cytochrome c₅₅₂ is able to rapidly shuttle electrons between cytochrome c₅₅₀ and the terminal ba₃- and caa₃-type oxidases, thereby unveiling a novel electron transfer pathway, putatively linking sulfite oxidation to oxygen reduction in T. thermophilus: sulfite → SOR → cytochrome c₅₅₀ → cytochrome c₅₅₂ → ba₃ oxidase/caa₃ oxidase.

Taking advantage of the thermal stability of the isolated proteins, the efficiency of electron transfer along the c₅₅₀ → c₅₅₂ → ba₃ oxidase/caa₃ oxidase pathway was evaluated over a wide temperature range; as expected, in these experiments, regardless of ba₃ or caa₃ oxidase being present, a stepwise temperature increase from 5°C to 60°C overall resulted in a significant (∼8-fold) increase in the oxidation rate of cytochrome c₅₅₀ (data not shown).

The kinetics of electron transfer between cytochromes c₅₅₀ and c₅₅₂ was also directly investigated by stopped-flow spectroscopy in 5 mM bis-Tris (pH 7.0) (Fig. 4). In these experiments, ascorbate-reduced cytochrome c₅₅₀ was mixed anaerobically with oxidized cytochrome c₅₅₂; the reaction was monitored at 416 nm, where the highly similar reduced-minus-oxidized absorption spectra of the two c-type cytochromes exhibited maximal differences. As shown in Fig. 4A, despite the low experimental temperature (4.5°C), the reaction was fast (<100 ms) and sensitive to ionic strength, with slower rates measured at higher salt concentrations. Since the reaction had to be investigated with comparable concentrations of the two cytochromes (c₅₅₀ and c₅₅₂), i.e., under non-pseudo-first-order conditions, experimental time courses were analyzed as described in references 20 and 30 (see Materials and Methods). In order to estimate the forward (k_F) and reverse (k_R) second-order reaction rate constants, the experiment was repeated at two different ionic strengths (12 mM and 152 mM) by systematically varying the cytchrome c₅₅₀ concentration. According to references 20 and 30, from the dependence of the observed rate constant (k_obs) on cytochrome c₅₅₀ concentration, k_F and k_R were estimated to be on the order of 10⁷ M⁻¹ s⁻¹ and 10⁶ M⁻¹ s⁻¹, respectively, with higher values measured at lower ionic strengths (Fig. 4B). These high rates suggest that cytochrome c₅₅₀ acts physiologically as an electron donor for cytochrome c₅₅₂.

Fig. 4.

Kinetics of electron transfer from cytochrome c₅₅₀ to cytochrome c₅₅₂. (A) Absorption changes measured at 416 nm after mixing 5.5 μM oxidized cytochrome c₅₅₂ with 3.4 μM cytochrome c₅₅₀ prereduced with 1 mM ascorbate. Temperature, 4.5°C. Ionic strength, 12 mM. The experimental trace (dashed) is shown together with its best fit (solid line) (see Materials and Methods). (Inset) Ionic strength dependence of the observed rate constant. (B) Observed rate constant measured as a function of cytochrome c₅₅₀ concentration at [c₅₅₂] = 2.75 μM (•) or 2.05 μM (○). Second-order rate constants estimated according to the methods described in reference 30: k_F = 5.5 × 10⁷ M⁻¹ s⁻¹ and k_R = 0.5 × 10⁷ M⁻¹ s⁻¹ at the lower ionic strength (•); k_F = 2.0 × 10⁷ M⁻¹ s⁻¹ and k_R = 0.2 × 10⁷ M⁻¹ s⁻¹ at the higher ionic strength (○).

The electron pathway from sulfite to oxygen in T. thermophilus.

To provide direct evidence for an electron transfer pathway coupling sulfite oxidation to oxygen reduction in T. thermophilus, the proposed pathway was reconstituted in vitro and oxygen consumption was measured polarographically at 37°C in the presence of sulfite. As shown in Fig. 5 (trace A), after addition of SOR_HB8 to an air-equilibrated solution containing excess sulfite in the presence of cytochrome c₅₅₀, cytochrome c₅₅₂, and ba₃ oxidase, the rate of O₂ consumption increased significantly. The increase was proportional to the concentration of SOR_HB8. The observed O₂ consumption rate, normalized to the SOR_HB8 concentration, was low (≈1 min⁻¹). The bottleneck in the electron transfer pathway resided uphill of cytochrome c₅₅₀, since addition of ascorbate and N,N,N′,N′-tetramethyl-p-phenylenediamine (TMPD) greatly increased the apparent rate of O₂ consumption (Fig. 5, trace A). In agreement with the results presented above, no consumption of oxygen was recorded when the novel cytochrome c₅₅₀ was omitted in the assay mixture (Fig. 5, trace B). This result shows that (i) cytochrome c₅₅₂ is unable to accept electrons directly from the sulfite oxidoreductase and (ii) the protein encoded by TTH1325, similar to all the bacterial SORs described to date (12), is not a sulfite oxidase per se, but rather a sulfite:ferricytochrome c oxidoreductase/sulfite dehydrogenase. As expected, identical results were obtained when ba₃ oxidase was replaced by caa₃ oxidase in the oxygraphic assays described above (data not shown).

Fig. 5.

Sulfite respiration. O₂ consumption by ba₃ cytochrome oxidase in the presence of SOR and the cytochromes c₅₅₀ and c₅₅₂, with sulfite serving as the primary electron donor. Values refer to the observed rate of O₂ consumption, expressed in μM O₂ min⁻¹. Concentrations: [sulfite] = 2 mM; [c₅₅₂] and [c₅₅₀] = 0.5 μM; [ba₃] = 0.3 μM; [Asc] (ascorbate) = 6.7 mM; [TMPD] = 0.2 mM. In trace A, the two SOR additions were 1.8 μM and 3.6 μM, respectively; in trace B, 5.6 μM SOR was present. Temperature, 37°C.

Altogether, the results presented support the existence in T. thermophilus of a metabolic pathway (sulfite → SOR_HB8 → cytochrome c₅₅₀ → cytochrome c₅₅₂ → ba₃ oxidase/caa₃ oxidase), in which the novel cytochrome c₅₅₀ here characterized couples sulfite oxidation to cell respiration. By this novel pathway, electrons generated by oxidation of sulfur compounds can enter the respiratory chain at the level of cytochrome c₅₅₂, thereby possibly sustaining bacterial bioenergetics.

DISCUSSION

Several microorganisms are able to utilize sulfur as the sole energy source. However, the metabolism of sulfur compounds, including those generated during the catabolism of sulfur-containing amino acids, is usually a means of detoxification. Recently it has been shown that T. thermophilus is no longer an exception. Annotation of the genomes available from T. thermophilus HB8 (NCBI GPID 13202) and HB27 (NCBI GPID 10617) (18) indeed revealed the presence of numerous genes potentially involved in sulfur metabolism, including a soxC gene encoding a putative sulfite dehydrogenase, associated with other cytochrome and flavoprotein genes. In addition, a sulfite oxidase was purified from T. thermophilus AT62 and characterized (9), although the physiological electron acceptor for this enzyme and hence its integration in a metabolic pathway have not been determined.

To gain insight into the physiological role of the SOR from T. thermophilus HB8, the vicinity of the gene encoding this enzyme (TTHA1325) was analyzed. The TTHA1325 gene was found to belong to an operon, along with the TTHA1326 and TTHA1327 genes encoding cytochrome c₅₅₀ and a protein with a putative MoCo sulfurase domain, respectively. The operon structure strongly suggests that these genes are coexpressed. Our results suggest that the novel c-type cytochrome (TTHA1326) and the sulfite-oxidizing enzyme (TTHA1325) are both produced in the periplasmic space of T. thermophilus HB8, the compartment where sulfite oxidation takes place in Gram-negative bacteria (6).

The genes encoding SorA from S. novella, SorT from S. meliloti, and the sulfite-oxidizing enzymes (SOE) from C. jejuni, C. necator, and D. acidovorans are all associated with a cytochrome c-encoding gene (7, 31). Nevertheless, these enzymes present a high diversity in their quaternary structures. The c-type cytochromes themselves are diverse: the cytochromes from S. novella, S. meliloti, C. jejuni, and D. acidovorans are monoheme proteins, while those from C. necator and T. thermophilus are diheme proteins. Both cytochrome c₅₅₀ and the SOR from T. thermophilus HB8 were isolated without associated proteins, suggesting that the cytochrome is not an integral part of the SOR. Similarly, the SOR_AT62 was isolated by Di Salle et al. (9) as a monomeric protein, and no associated proteins could be identified. Further in vitro analysis using gel filtration chromatography confirmed that the purified SOR and cytochrome c₅₅₀ do not form a stable complex.

Sequence analysis of the gene TTHA1326 revealed the presence of two heme-binding sites, and the presence of the two cofactors was verified in the purified protein. The N-terminal region of cytochrome c₅₅₀ exhibits a high sequence identity with the oxidoreductase subunit B Cj0004c from C. jejuni. Such a finding suggests that this domain may act as an electron acceptor for SOR_HB8. This hypothesis is supported by the fact that the C-terminal region of cytochrome c₅₅₀ is highly similar to the well-characterized cytochrome c₅₅₂ from the same organism and that the latter cannot accept electrons from the sulfite oxidase, as we demonstrated above.

Gel filtration chromatography has shown that the cytochrome is expressed as a monomer, while the SOR is expressed as a homodimer. The presence of a MoCo dimerization domain, however, does not account by itself for the dimeric structure of SOR_HB8, because such a domain has been also identified in monomeric SORs, like the enzymes from T. thermophilus A62 and C. necator (7, 9).

SOR_HB8 is unable to use O₂ as a direct final electron acceptor and therefore should not be referred to as a sulfite oxidase, but rather as a sulfite:ferricytochrome c oxidoreductase. Cytochrome c₅₅₀ was indeed identified as an electron acceptor for this enzyme. Based on the high sequence similarity between the C-terminal domain of cytochromes c₅₅₀ and c₅₅₂, the substrate of the ba₃ and caa₃ terminal oxidases, we hypothesized that cytochrome c₅₅₀ could be an intermediary between sulfite oxidation and the respiratory chain. Although the novel cytochrome c₅₅₀ proved to be a very poor substrate for both oxidases, it bridges the two pathways as it shuttles electrons to cytochrome c₅₅₂ at high rates (k_F, >10⁷ M⁻¹ s⁻¹ at 4.5°C). Formerly, such a link between sulfur oxidation and the respiratory chain was shown with C. jejuni, where an electron transfer was shown to occur from the heme-containing subunit of the group 1 SOR to the cytochrome cb oxidase via cytochrome c₅₅₂, however, using cell extracts and not purified proteins (31). In the present study, for the first time such a complete pathway has been demonstrated for a heme-lacking (group 2) SOR. Based on sequence analysis, it is tempting to speculate that cytochrome c₅₅₀ comprises two heme domains in redox equilibrium with possibly distinct functions: one heme domain accepting electrons from the SOR and the other one involved in electron delivery to cytochrome c₅₅₂. Although existence of a diheme cytochrome c serving as an electron acceptor for SOR has been identified in P. pantotrophus (33), the specific role of each heme in the electron transfer reaction is yet unknown and is being further investigated.

Figure 6 illustrates the proposed pathway of sulfite oxidation in T. thermophilus. It postulates cytochrome c₅₅₀ to be an electron shuttle between the SOR and the ba₃/caa₃ cytochrome oxidases via cytochrome c₅₅₂. In this work, for the first time we have identified a complete pathway of sulfite respiration in Thermus spp. Our results show that SOR_HB8 belongs to the second group of bacterial sulfite-oxidizing enzymes. It cannot transfer electrons to molecular oxygen but, in contrast to the enzyme from S. novella, it can reduce a soluble c-type cytochrome, without need for a heme relay.

Fig. 6.

Proposed mechanism of sulfite metabolism via the SOR system in T. thermophilus. Electrons generated during sulfite oxidation are captured by cytochrome c₅₅₀ and transferred to the terminal oxidases via cytochrome c₅₅₂ to finally reach the ultimate acceptor, molecular oxygen. The electron transfer from the bc₁ complex to the terminal oxidases via cytochrome c₅₅₂ is also illustrated. PDB IDs: cytochrome c oxidase ba₃, 1EHK; cytochrome c₅₅₂, 1C52.

It is yet unknown if the proposed sulfite-metabolizing pathway in Thermus is expressed under particular growth conditions, such as exposure to sulfite. In soil or water, sulfite is more stable in low-oxygen niches than under aerobic conditions; the SOR system in T. thermophilus may therefore be prevalent under microaerobic conditions. Further investigations are needed to clarify whether the novel sulfite oxidation pathway elucidated in T. thermophilus is of bioenergetic relevance for the bacterium (at least under specific growth conditions) or if it is rather a means of detoxification. The presence in T. thermophilus of another putative (SoxCD-like) sulfur oxidation pathway has been revealed by genomic analysis. Such systems have been shown, notably for P. pantotrophus, to be responsible for the metabolism of environmental inorganic sulfur (14). Under limited oxygen availability, the SOR system could contribute to energy conservation via the respiratory chain. To confirm that, further work is needed to understand how the expression of the SOR system is regulated in T. thermophilus.

Recently a new heme-lacking SOR was identified in D. radiodurans (8) which was unable to transfer electrons to horse heart cytochrome c. Those authors speculated that this could be due to the lack of a heme relay and that therefore O₂ could be the direct electron acceptor in vivo. Based on our findings, it appears that the direct physiological electron acceptor could be a yet-unidentified cytochrome c, even if no cytochrome c-encoding genes were identified in the vicinity of this SOR gene (8). In the absence of an obvious link to the respiratory chain, our results show that one cannot exclude a putative coupling of sulfite oxidation with the latter. The revealed pathway in T. thermophilus could apply indeed to other organisms. Intermediaries, such as cytochrome c₅₅₂, could ensure the transport of electrons generated by the SOR and their contribution to the high-energy-yielding dioxygen reduction.