|Home | About | Journals | Submit | Contact Us | Français|
Thiosulfate-oxidizing sox gene homologues were found at four loci (I, II, III, and IV) on the genome of Bradyrhizobium japonicum USDA110, a symbiotic nitrogen-fixing bacterium in soil. In fact, B. japonicum USDA110 can oxidize thiosulfate and grow under a chemolithotrophic condition. The deletion mutation of the soxY1 gene at the sox locus I, homologous to the sulfur-oxidizing (Sox) system in Alphaproteobacteria, left B. japonicum unable to oxidize thiosulfate and grow under chemolithotrophic conditions, whereas the deletion mutation of the soxY2 gene at sox locus II, homologous to the Sox system in green sulfur bacteria, produced phenotypes similar to those of wild-type USDA110. Thiosulfate-dependent O2 respiration was observed only in USDA110 and the soxY2 mutant and not in the soxY1 mutant. In the cells, 1 mol of thiosulfate was stoichiometrically converted to approximately 2 mol of sulfate and consumed approximately 2 mol of O2. B. japonicum USDA110 showed 14CO2 fixation under chemolithotrophic growth conditions. The CO2 fixation of resting cells was significantly dependent on thiosulfate addition. These results show that USDA110 is able to grow chemolithoautotrophically using thiosulfate as an electron donor, oxygen as an electron acceptor, and carbon dioxide as a carbon source, which likely depends on sox locus I including the soxY1 gene on USDA110 genome. Thiosulfate oxidation capability is frequently found in members of the Bradyrhizobiaceae, which phylogenetic analysis showed to be associated with the presence of sox locus I homologues, including the soxY1 gene of B. japonicum USDA110.
Bradyrhizobium japonicum exhibits two life styles: as a free-living soil bacterium and as a bacteroid that fixes atmospheric nitrogen in soybeans (2, 16, 17). The genomic sequence of B. japonicum USDA110 (20) is unusual and complex, containing numerous copies of genes for degradation, transport, C1 metabolism, transcriptional regulation, and respiratory chains (12), which might allow the bacterium to adapt to diverse and fluctuating environments (17). However, B. japonicum USDA110 expresses only a portion of the complicated gene sets for the degradation of aromatics and C1 compound metabolism in the presence of vanillate, a naturally occurring aromatic compound (17, 42).
Biological oxidation of reduced inorganic sulfur compounds is one of the major reactions in the global sulfur cycle (9, 26) and is the major reaction under extreme conditions such as those in deep-sea hydrothermal vents (31), solfataras (23), and volcanic environments (22). The oxidation reactions in these ecosystems are carried out by prokaryotes of the domains Archaea and Bacteria (9, 23). In the Bacteria, sulfur is oxidized by chemotrophic and phototrophic bacteria, which are phylogenetically and physiologically diverse (9).
A sulfur-oxidizing (Sox) enzyme system has been found and characterized in Paracoccus versutus and Paracoccus pantotrophus GB17 through the pioneering work of research groups associated with Lu et al. (28) and Friedrich et al. (8, 9, 46). Genes homologous to those encoding the Sox enzyme have been found from genomes of the members of the domain Bacteria but not in the domain Archaea (8, 9). However, the functional aspects of the Sox system based on genome information remain to be answered (9).
A survey of the B. japonicum USDA110 genome revealed multiple homologues of sox genes; however, it was not clear if these homologues were functional for the oxidation of inorganic sulfur compounds and chemolithotrophic growth in this bacterium (7). The sox gene cluster in P. pantotrophus GB17 comprises at least 15 genes, soxTRSVWXYZABCDEFGH, which are responsible for thiosulfate oxidation and chemolithotrophic growth (9). Thiosulfate has been regarded as crucial among the reduced inorganic compounds in Sox biochemistry (7) and the geochemical sulfur cycle (18).
B. japonicum is known to grow chemolithoautotrophically using the gaseous electron donors H2 and CO (6, 13, 27). H2 uptake-positive (Hup+) strains of B. japonicum USDA122 (13) and USDA110 (6) carrying hup genes can grow chemolithoautotrophically in mineral salts medium with H2, CO2, and low concentrations of O2. B. japonicum USDA110 is also reportedly able to grow chemolithotrophically with carbon monoxide as an electron donor (27). Thus, it is possible that B. japonicum USDA110 may utilize an inorganic sulfur compound, thiosulfate, as an electron donor and CO2 as a carbon source for chemolithoautotrophic growth.
With this background, the present study addressed the following questions: (i) Does B. japonicum USDA110 grow chemolithoautotrophically with thiosulfate as an energy source? (ii) Which of the sox genes could be responsible for thiosulfate oxidation in B. japonicum? (iii) Do these features extend to other members of the Bradyrhizobiaceae?
Using the amino acid sequences for the sox genes of P. pantotrophus GB17 and Chlorobium tepidum TLS (4), the BLASTP program was used to search the Bradyrhizobium section of Rhizobase (http://bacteria.kazusa.or.jp/rhizobase/Bradyrhizobium/index.html) for matching amino acid sequences of sox gene homologues belonging to B. japonicum USDA110. The amino acid sequences of the sox genes of P. pantotrophus GB17 and C. tepidum TLS were obtained from the National Center for Biotechnology Information (NCBI) database (http://www.ncbi.nlm.nih.gov/) and Cyanobase (http://genome.kazusa.or.jp/cyanobase/), respectively. The signal peptide and transmembrane helices of the sox genes were predicted by the SignalP (version 3.0), TatP, and TMHMM programs available from the Center for Biological Sequence Analysis at the Technical University of Denmark (http://www.cbs.dtu.dk/services/). Molecular weights of sox genes were calculated by Genetyx, version 5.1, for Windows (Genetyx, Tokyo, Japan). We searched for conserved domains of sox genes with the InterProScan Sequence Search available from the European Bioinformatics Institute of the European Molecular Biology Laboratory (EMBL-EBI; http://www.ebi.ac.uk/Tools/InterProScan/). The amino acid sequences of soxY1 genes were aligned using the ClustalX program (http://www.clustal.org/).
The bacterial strains and plasmids used are listed in Table Table1.1. Cultures of B. japonicum, Bradyrhizobium elkanii, Bradyrhizobium sp., and Agromonas oligotrophica were precultured aerobically at 30°C in HM salt medium (1) supplemented with 0.1% (wt/vol) arabinose and 0.025% Difco yeast extract (Becton, Dickinson and Co., Sparks, MD). Rhodopseudomonas palustris CGA009 (24) was precultured aerobically at 30°C in Difco nutrient broth (Becton, Dickinson and Company). For growth experiments, these strains were grown aerobically at 25°C in Taylor medium (43), which is a minimal medium for thiosulfate-oxidizing bacteria, supplemented with sodium thiosulfate (Na2S2O3) or succinate (Na salt). Taylor medium contained the following components dissolved in 1 liter of distilled water (pH 6.8): 1.0 g of NH4Cl, 2.0 g of KH2PO4, 0.8 g of MgSO4·7H2O, and 1 ml of a trace metal solution consisting of 5.0 g of Na2EDTA, 2.2 g of ZnSO4·7H2O, 7.3 g of CaCl2·2H2O, 2.5 g of MnCl2·4H2O, 0.5 g of CoCl2·6H2O, 0.5 g of (NH4)6Mo7O2·7H2O, 5.0 g of FeSO4·7H2O, and 0.2 g of CuSO4·5H2O. Antibiotics were added to the medium for growing B. japonicum USDA110 as follows: tetracycline (Tc), spectinomycin (Sp), streptomycin (Sm), and kanamycin (Km) at 100 μg ml−1 and polymyxin B at 50 μg ml−1. For Escherichia coli the concentrations were 15 μg ml−1 Tc, 50 μg ml−1 Sp, 50 μg ml−1 Sm, 50 μg ml−1 Km, and 100 μg ml−1 ampicillin.
Cells were washed twice in Taylor medium before inoculations. For plate assays of thiosulfate oxidation, phenol red was added to the medium at 20 μg liter−1 as a pH indicator. The color was monitored for a change from red to yellow because thiosulfate oxidation causes a drop in pH (46). The cells were grown at 25°C for 20 days on Taylor agar plates supplemented with various concentrations of thiosulfate. For chemolithotrophic growth experiments in liquid cultures, the cells were grown aerobically in 100 ml of Taylor medium in 500-ml Erlenmeyer flasks supplemented with 4 mM sodium thiosulfate for 28 days at 25°C. The number of CFU was monitored as a growth index by serial dilution of the culture with sterilized water and plating on HM salt medium. For mixotrophic growth conditions, the cells were grown aerobically in 100 ml of Taylor medium in 500-ml Erlenmeyer flasks supplemented with 4 mM sodium thiosulfate and 0.1% (wt/vol) sodium succinate for 6 days at 25°C. The turbidity (optical density at 660 nm [OD660]) of cultures was measured with a UV-1200 spectrophotometer (Shimadzu, Kyoto, Japan) as an indicator of cell growth.
Thiosulfate concentrations in the culture supernatants were measured by iodometric titration (21). Sulfate was determined gravimetrically (10). Fifty milliliters of the culture supernatant was acidified with 1 ml of 5 N HCl. Subsequently, sulfate was precipitated with 4 ml of 0.5 M BaCl2. The precipitation reaction was allowed to proceed for 2 h, after which the BaSO4 was collected by filtration and dried overnight at 90°C.
Isolation of plasmids, DNA ligation, and transformation of E. coli were performed as described by Sambrook et al. (37). Total bacterial DNA was isolated from cultured cells as described previously (30). Southern hybridization was carried out as described previously (39).
A 4.8-kb SalI DNA fragment containing soxVWX1Y1Z1A1B1 was isolated from brp14182, a clone from the pUC18 library for the sequences of B. japonicum USDA110, and inserted into the SalI site of pK18mob. The resulting plasmid, pSAC20, was double digested by AccIII and StuI. The Tcr cassette was isolated from p34S-Tc at the SmaI site (3) and inserted into pSAC20 at the AccIII and StuI sites, yielding pSAC21. pSAC21 was introduced into B. japonicum USDA110 by triparental mating using pRK2013 as a helper plasmid (39).
A 4.6-kb BamHI DNA fragment containing soxX2Y2Z2A2B2 was isolated from brp01133 and inserted into the BamHI site of pK18mob, resulting in the vector pSAC22. pSAC22 was double digested by BsiWI and BstPI. The Ω-cassette was isolated from pHP45Ω at the SmaI site (32) and inserted into pSAC22 at the BsiWI and BstPI sites, resulting in the vector pSAC23. pSAC23 was introduced into B. japonicum USDA110 by triparental mating. Double crossover events of soxY1 and soxY2 deletion mutations were verified by PCR.
The O2 uptake rate of whole cells was determined polarographically with a Clark-type O2 electrode (5300 Biological Oxygen Monitor; Yellow Springs Instrument Co. Inc., Tokyo, Japan). The cells were harvested by centrifugation and then washed twice and resuspended in 50 mM potassium phosphate buffer containing 2.5 mM MgCl2 (pH 7.0) (13). The assay mixture contained the potassium phosphate-MgCl2 buffer, about 10 mg of cells (wet weight), and 30 mM thiosulfate at 25°C to start the reaction (13, 46). The amount of dissolved oxygen in air-saturated water was calculated on the basis of 258 nmol of O2 ml−1 in air-saturated water at 25°C. The rate of thiosulfate-dependent oxygen consumption was calculated using the rates of thiosulfate oxidation and endogenous respiration and expressed as nanomoles of O2 consumed per minute per gram of wet cells (nmol O2 min−1 g of wet weight−1).
NaH14CO3 (specific activity, 2.18 GBq mmol−1) was obtained from GE Healthcare UK Limited (GE Healthcare, Tokyo, Japan). The CO2 fixation assay was conducted using the method of Lepo et al. (25) and Maier (29). After the cells had been grown chemolithotrophically for 12 days at 25°C, NaH14CO3 (1.91 μmol) was added to 100 ml of culture in Taylor medium in 500-ml Erlenmeyer flasks. One milliliter of the culture was periodically transferred to scintillation vials and acidified with 0.3 ml of 60% trichloroacetic acid. CO2 that had not been fixed was allowed to dissipate for 48 h into a fume hood. Fifteen milliliters of scintillation cocktail (ACS II; Amersham Biosciences, Tokyo, Japan) was added to each vial, and the radioactivity was determined by liquid scintillation counting (Aloka LSC-5100; Tokyo, Japan). As a negative control, cells were killed by placing the culture flask in a boiling water bath for 10 min just before the addition of NaH14CO3.
Resting cells were prepared to examine whether CO2 fixation depends on thiosulfate. Cells were harvested from 1 liter of culture that had been grown chemolithotrophically for 12 days, washed twice, and suspended in 20 ml of Taylor medium without thiosulfate. The 14CO2 uptake assay was started by the addition of NaH14CO3 (370 kBq) at a final concentration of 20 mM NaHCO3 and 4 mM thiosulfate to cells at a density of (4.6 ± 0.7) × 109 CFU ml−1 at 25°C.
For the analysis of CO2 concentrations in the flask headspace, 1-ml gas samples were injected into a GC-2014 gas chromatograph ([GC] Shimadzu), equipped with a thermal conductivity detection (TCD) detector (Shimadzu) and Parapak Q column (80/100 mesh; 0.3-mm diameter by 2-m length). The flow rate of the carrier gas (He) was 30 ml min−1. Throughout the GC analysis, the temperatures of injection, column, and detector were 100, 40, and 100°C, respectively. Peak areas were calculated from chromatograms by an integrator (ChromatoPack C-R18; Shimadzu).
We found four regions of the genomic sequence of B. japonicum USDA110 (20) that are homologous to well-known sox genes that mediate thiosulfate oxidation (9). In the present study, these regions are designated sox loci I (soxTSRVWX1Y1Z1A1B1C1D1), II (soxX2Y2Z2A2B2), III (soxFY3Z3A3), and IV (soxC2D2) (Fig. (Fig.1A).1A). The amino acid sequence analyses of the proteins deduced from these sox gene homologues suggest that most of them carry Tat- and Sec-dependent signal peptides so that they are excreted into the periplasm of the bacterium (see Table S1 in the supplemental material), where the Sox system functions (8).
As shown in Fig. Fig.1B,1B, the gene organization at sox loci I, II, and III in B. japonicum USDA110 was very similar to the organization of photosynthetic stem-nodulating bradyrhizobia, Bradyrhizobium sp. BTAi1 and Bradyrhizobium sp. ORS278 (11). The homology of the amino acid sequences was high (identities of 67% to 92%) between USDA110 and photosynthetic bradyrhizobia (see Table S2 in the supplemental material). R. palustris CGA009 (24) carried sox genes that were homologous to those at sox locus I in B. japonicum USDA110 (identities of 44% to 73%) (see Table S2). Beyond the above members (Bradyrhizobium and Rhodopseudomonas spp.), the deduced amino acid sequences for loci I and II were highly homologous to those of well-characterized P. pantotrophus GB17 (44% to 73% identities) (7) and the green sulfur bacterium C. tepidum (44% to 57% identities) (4), respectively (see Table S2).
Plate assays showed thiosulfate-oxidation capability in B. japonicum USDA110 at thiosulfate concentrations below 4 mM (Table (Table2;2; see also Fig. S1 in the supplemental material). The presence of 20 mM sodium thiosulfate inhibited its oxidation, as indicated by color changes on the Taylor plate (Table (Table2;2; see also Fig. S1). Thus, we examined thiosulfate oxidation by B. japonicum USDA110 under chemolithotrophic conditions with 4 mM thiosulfate as the sole energy source (Fig. (Fig.2A).2A). The numbers of CFU increased with time over 28 days; thiosulfate concentrations (S2O32−) in the culture decreased along with the cell growth (Fig. (Fig.2A).2A). During chemolithotrophic cultivation, 3.5 ± 0.1 mM (average ± standard deviation [SD]) thiosulfate was consumed while 7.6 ± 0.1 mM sulfate was produced. Thus, the stoichiometry of thiosulfate oxidation is approximately 2 mol of sulfate generated by 1 mol of thiosulfate. No growth was observed when thiosulfate was not added to the medium (data not shown). These results indicate that B. japonicum USDA110 can grow chemolithotrophically using thiosulfate as a sole electron donor.
B. japonicum USDA110 also oxidized thiosulfate (4 mM) in the presence of 0.1% (wt/vol) succinate as an energy source (Fig. (Fig.2B),2B), indicating that the addition of succinate does not inhibit thiosulfate oxidation (Fig. 2A and B). Under these conditions, 1 mol of thiosulfate produced 1.9 mol of sulfate during growth (4.0 ± 0.0 mM [average ± SD] thiosulfate was consumed while 7.4 ± 0.1 mM sulfate was produced). B. japonicum USDA110 was unable to grow with 20 mM sodium thiosulfate under the mixotrophic condition (data not shown). This growth inhibition was not due to toxicity of the counter ion (Na+) in the sodium thiosulfate added because 40 mM NaCl did not affect cell growth of USDA110 under the mixotrophic condition (data not shown).
SoxY, a subunit of the SoxYZ complex, plays a key role in thiosulfate oxidation in Sox systems (9). Sulfur-SoxY complexes are covalently bound to a cysteine residue (Cys110Y is on the carboxy terminus of SoxY in P. pantotrophus GB17) located within a characteristic GGCGG pentapeptide at the C terminus of the SoxY subunit of the SoxYZ complex. C-terminal glycine residues are conserved in SoxY orthologs and are presumably essential (33, 40). In the genome of B. japonicum USDA110, the GGCGG motif was not conserved in the C terminus of soxY3 in locus III but was well conserved in the C terminus of soxY1 and soxY2 in loci I and II, respectively (see Fig. S2 in the supplemental material). Therefore, for mutation analysis of the B. japonicum USDA110 genome, we focused on soxY1 and soxY2 carrying the GGCGG motif. Deletion mutants for soxY1 and soxY2 were constructed as described in Materials and Methods (Fig. (Fig.1B1B).
The soxY1 mutant culture showed no chemolithotrophic growth (expressed as CFU), and thiosulfate concentrations did not decrease over 28 days (Fig. (Fig.2C).2C). Under mixotrophic conditions, the soxY1 mutant was unable to oxidize thiosulfate even when growth was observed (Fig. (Fig.2D).2D). The soxY2 mutant was able to oxidize thiosulfate and grow under both chemolithotrophic and mixotrophic conditions (Fig. 2E and F), similar to wild-type USDA110 (Fig. 2A and B). These results indicate that soxY1 is responsible for thiosulfate oxidation in B. japonicum USDA110.
It is likely that B. japonicum USDA110 can grow chemolithotrophically using thiosulfate as an electron donor (Fig. (Fig.2A).2A). If this is true, then addition of thiosulfate could enhance O2 respiration under conditions deficient in electron donors. To test this, USDA110 cells from mixotrophic growth conditions with thiosulfate and succinate were subjected to a respiration assay with an O2 electrode. To minimize endogenous respiration, cells were prepared from cultures in late log phase. The O2 uptake rate appeared to increase just after the addition of thiosulfate to the cell suspensions (Fig. (Fig.3A,3A, arrowhead). Five independent assays yielded 3.2 ± 0.1 (average ± SD) μmol of O2 min−1 g of wet cells−1 as the thiosulfate-dependent O2 uptake rate for USDA110 cells. In contrast, the soxY1 mutant showed no thiosulfate-dependent O2 uptake (0.0 ± 0.1 μmol of O2 min−1 g of wet cells−1) (Fig. (Fig.3B).3B). In cultures of the soxY2 mutant, the O2 uptake rate also increased just after the thiosulfate addition (2.8 ± 0.1 μmol of O2 min−1 g of wet cells−1) (Fig. (Fig.3C).3C). These results show that B. japonicum USDA110 uses O2 as an electron acceptor during thiosulfate oxidation and that the soxY1 gene in B. japonicum USDA110 is required for this to occur. The molar ratio of thiosulfate added and O2 consumed was 2.1 ± 0.2 in four independent determinations, suggesting that 2 mol of O2 are stoichiometrically consumed during oxidation of 1 mol thiosulfate.
We examined the fixation of 14CO2 by USDA110 under chemolithotrophic growth conditions. NaH14CO3 was added into the culture flasks at a concentration of 1.91 μmol (equivalent to 0.087 ppm CO2 in the flasks) 12 days after inoculation. The average ambient CO2 concentration in the flasks was 691 ppm. 14CO2 uptake by chemolithotrophically grown USDA110 cells increased linearly with time (Fig. (Fig.4A,4A, live cells), whereas 14CO2 uptake by heat-treated cells was not detected (Fig. (Fig.4A).4A). The CO2 fixation rate was calculated as 7.7 μmol h−1 108 CFU−1 on the basis of the 14CO2 uptake rate (18.2 Bq h−1 flask−1) (Fig. (Fig.4A),4A), the CO2 concentration in the flask (691 ppm), and the 14C specific activity of NaH14CO3 (2.18 GBq mmol−1). The resting cells for this experiment (4.6 × 109 CFU ml−1) were prepared from chemolithotrophic growth culture 12 days after inoculation. When thiosulfate was added to the cell suspensions (final concentration, 4 mM), the CO2 fixation rate increased significantly compared with cells incubated without thiosulfate (Fig. (Fig.4B4B).
When Southern hybridization was performed using two long probes of sox locus I (soxTRSV and soxWX1Y1Z1A1B1C1D1) from B. japonicum USDA110, the positive signals were detected in only eight strains: B. japonicum USDA122, NC6, NC4, and NK2; A. oligotropica S58; photosynthetic bradyrhizobia BTAi1 and ORS278; and R. palustris CGA009 (Table (Table2);2); this suggests that the sox locus I genes are conserved in these strains. However, no signal was detected in four strains of B. japonicum (USDA124, T7, T9, and USDA6T), Bradyrhizobium sp. (G14130, HWK12, and HW13), and B. elkanii USDA76T.
To determine whether the eight strains belonging to the Bradyrhizobiaceae that carry homologues of the B. japonicum sox locus I genes express thiosulfate-oxidizing activity, they were subjected to a plate assay for this activity. These strains consistently produced yellow halos on the medium containing less than 4 mM thiosulfate (Table (Table2).2). The remaining strains (B. japonicum USDA124, T9, T7, USDA6T, Bradyrhizobium sp. G14130, HW12, HWK13, and B. elkanii USDA76T) did not produce yellow halos on any medium containing thiosulfate. These results show that Bradyrhizobiaceae carrying the sox locus I gene homologues of USDA110 consistently oxidize thiosulfate at concentrations of less than 4 mM.
When a phylogenetic tree was constructed on the basis of 16S rRNA gene sequences (Fig. (Fig.5A),5A), cluster BJ1 of B. japonicum (16) and a group of photosynthetic bradyrhizobia (PB) including A. oligotrophica S58 consistently carried sox locus I genes and expressed thiosulfate oxidation activity (Table (Table22).
To examine how the Sox system is distributed in the phylum Alphaproteobacteria, 50 amino acid sequences homologous to the B. japonicum soxY1 gene with E-values lower than 5e−20 were collected from the NCBI database. Among these sequences, the functionally important cysteine residue and GGCGG motif (33, 40) were well conserved (see Fig. S3 in the supplemental material). A phylogenetic analysis of these SoxY sequences (Fig. (Fig.5B;5B; see also Fig. S4) showed that major members of the order Rhodobacterales, including P. pantotrophus GB17, formed a single cluster in terms of SoxY homology (Fig. (Fig.5B,5B, gray triangle). A group including B. japonicum USDA110 and photosynthetic bradyrhizobia BTAi1 and ORS278 and another group including five strains of R. palustris formed respective compact clusters with high bootstrap values (Fig. (Fig.5B,5B, black triangles), which are distant from the well-characterized Rhodobacterales cluster.
Herein, we demonstrated that B. japonicum USDA110 can grow chemolithotrophically with low concentrations of thiosulfate as an electron donor (Fig. (Fig.2A)2A) and with oxygen as an electron acceptor. In addition, 14C experiments showed that USDA110 cells fixed ambient CO2 under chemolithotrophic conditions, indicating that the cells used ambient CO2 as a carbon source (Fig. (Fig.4A).4A). This chemolithoautotrophic growth required the functional soxY1 gene (Fig. (Fig.2C).2C). The members of Bradyrhizobiaceae with homologues of the sox locus I genes of USDA110 also demonstrated thiosulfate oxidation (Table (Table22 and Fig. Fig.5A5A).
Generally, there are two different end products of bacterial oxidation of thiosulfate: sulfate and tetrathionate (14). In B. japonicum USDA110, 1 mol of thiosulfate was converted to approximately 2 mol of sulfate and consumed approximately 2 of mol O2. Thus, the stoichiometry of the reaction for thiosulfate (S2O32−) oxidation in B. japonicum USDA110 should be the following: S2O32− + 2O2 + H2O → 2SO42− + 2H+ (ΔG′0 = −818.29 kJ).
This reaction indicates that sulfate is a product of thiosulfate oxidation in B. japonicum USDA110, yielding energy for growth which is identical to that reported for P. pantotrophus GB17 (10).
B. japonicum is known to grow chemolithoautotrophically using the gaseous electron donors H2 and CO (6, 13, 27). Thus, the present study demonstrates that B. japonicum USDA110 can utilize an inorganic sulfur compound, thiosulfate, as a nongaseous electron donor for chemolithoautotrophic growth.
Previously, thiosulfate-oxidizing bacteria such as aerobic sulfur-oxidizing bacteria have been isolated from sulfur-rich environments using enrichment culture. Therefore, high concentrations of thiosulfate (20 mM) have been added to the medium for cultivation and for biochemical and ecological analyses of these sulfur-oxidizing bacteria. Our results make it clear that these copiotrophic procedures would fail to detect strains capable of bacterial thiosulfate oxidation only at low thiosulfate concentrations, such as B. japonicum USDA110. Therefore, it is possible that past studies have overlooked such phenotypes of sulfur-oxidizing bacteria in nature. In fact, this study has shown that several members of the Bradyrhizobiaceae such as B. japonicum, A. oligotrophica, Bradyrhizobium sp. (photosynthetic bradyrhizobia), and R. palustris can oxidize thiosulfate only at low concentrations (Table (Table2).2). A phylogenetic tree based on SoxY (Fig. (Fig.5B;5B; see also Fig. S4 in the supplemental material) shows unique clusters that are distant from the well-characterized Rhodobacterales cluster. Taken together, it is likely that some members of the Bradyrhizobiaceae are a new type of thiosulfate-oxidizing bacteria that may adapt to oligotrophic sulfur environments.
A careful homology search for sox genes (Fig. (Fig.1;1; see also Table S1 and Fig. S2 in the supplemental material) initially suggested that sox loci I and II would be functional for thiosulfate oxidation. Subsequently, our deletion mutant analyses revealed that the soxY1 gene is essential for thiosulfate oxidation (Fig. (Fig.11 and and2).2). Among the duplicate copies of oxygenase genes for aromatic compound degradation in B. japonicum, only one set of genes that included pcaG1H1 was functional for vanillate catabolism (17, 42). Therefore, it is likely that the B. japonicum genome carries one functional gene for specific biochemical traits among redundant copies.
In the present study, we frequently found thiosulfate oxidation capability in members of Bradyrhizobiaceae, which was associated with the presence of homologues for sox locus I genes of B. japonicum. So far, we do not know the phylogenetic or ecological implications of the coexistence of Sox+ and Sox− strains among Bradyrhizobiaceae and even within B. japonicum strains (Table (Table22 and Fig. Fig.5A).5A). However, it is interesting that the strains showing thiosulfate-oxidizing capability (Sox+) were clustered with those with hupSL genes encoding uptake hydrogenase, and nosZ genes encoding N2O reductase for denitrification (Fig. (Fig.5A).5A). One possible explanation is that the genomic regions containing sox, hup, and nos genes behave similarly during the evolution of the Bradyrhizobiaceae owing to ecological advantages in low-nutrient environments. In particular, B. japonicum lineages may diverge into the BJ1 cluster and others for these inorganic metabolisms. However, this possibility cannot be explained by a single event of horizontal gene transfer because these genes (sox, hup, and nos) are scattered on different loci of the USDA110 genome (Fig. (Fig.11).
The possession of alternative respiration systems provides flexibility to bacteria for adapting to fluctuating environments (15, 34). It will be interesting to clarify the ecological role of Bradyrhizobiaceae that exhibit sulfur oxidation in soil environments such as paddy fields.
This work was supported in part by a Grant-in-Aid for Scientific Research on Priority Area Comparative Genomics, by Grants-in-Aid for Scientific Research (number 17380046), and by PROBRAIN.
We thank C. Harwood (University of Washington) for providing R. palustris CGA009 and E. Giraud (French National Institute for Agricultural Research) and M. Sadowsky (University of Minnesota) for providing Bradyrhizobium sp. strains ORS278 and BTAi1, respectively.
Published ahead of print on 19 February 2010.
†Supplemental material for this article may be found at http://aem.asm.org/.