ABSTRACT
Metagenomic analyses of surface seawater reveal that genes for sulfur oxidation are widespread in bacterioplankton communities. However, little is known about the metabolic processes used to exploit the energy potentially gained from inorganic sulfur oxidation in oxic seawater. We therefore studied the sox gene system containing Roseobacter clade isolate Phaeobacter sp. strain MED193 in acetate minimal medium with and without thiosulfate. The addition of thiosulfate enhanced the bacterial growth yields up to 40% in this strain. Concomitantly, soxB and soxY gene expression increased about 8-fold with thiosulfate and remained 11-fold higher than that in controls through stationary phase. At stationary phase, thiosulfate stimulated protein synthesis and anaplerotic CO2 fixation rates up to 5- and 35-fold, respectively. Several genes involved in anaplerotic CO2 fixation (i.e., pyruvate carboxylase, propionyl coenzyme A [CoA], and crotonyl-CoA carboxylase) were highly expressed during active growth, coinciding with high CO2 fixation rates. The high expression of key genes in the ethylmalonyl-CoA pathway suggests that this is an important pathway for the utilization of two-carbon compounds in Phaeobacter sp. MED193. Overall, our findings imply that Roseobacter clade bacteria carrying sox genes can use their lithotrophic potential to gain additional energy from sulfur oxidation for both increasing their growth capacity and improving their long-term survival.
INTRODUCTION
Heterotrophic bacteria in the surface ocean typically rely on dissolved organic carbon compounds for obtaining energy and cellular building blocks. Nevertheless, metagenomic inventories of the world's oceans reveal a substantial diversity and abundance of genes, allowing the use of alternative energy sources, such as hydrogen gas, CO, and inorganic sulfur, in oxic surface waters (1, 2). For example, using soxB as a marker gene, about 10% of the bacteria in the Sargasso Sea were identified to carry the sulfur oxidation (Sox) enzyme system (1). Also, community expression analysis in surface waters at the ALOHA Station in the North Pacific subtropical gyre showed that several of the genes in the sox cluster were actively expressed (3). Considering the abundance of sox genes and the fact that reduced inorganic sulfur compounds are usually found in very low concentrations in surface seawater, if at all, it remains unclear what the substrates for sox-encoded proteins are.
The sulfur-oxidizing enzyme system encoded by sox genes is widely present in heterotrophic, phototrophic, lithotrophic, and methylotrophic bacteria. The complete sox gene cluster contains 15 genes, which oxidize thiosulfate to two molecules of sulfate without any intermediates. The number of sox genes differs between organisms, from a complete set in, e.g., Paracoccus pantotrophus and Rhodobacter palustris, to partial sets in Chlorobium tepidum and Allochromatium vinosum (4). The core set of genes that allows this process is soxXYZABCD, which encode the dimeric proteins SoxXA, SoxYZ, and SoxCD and the nondimeric protein SoxB, which are functional in the periplasm. The sox enzyme system was first reported in the alphaproteobacterium P. pantotrophus (5), after which individual components have been further genetically and biochemically characterized (4, 6). Mutational analysis of soxB and soxY in the phototrophic purple sulfur bacterium A. vinosum and Bradyrhizobium japonicum revealed that these genes are essential for thiosulfate oxidation (7, 8). The dimeric SoxYZ does not hold metals or cofactors and covalently binds with sulfur of various oxidation states (9). The dinuclear manganese-containing SoxB, proposed to function as a sulfate thiol esterase or sulfate thiol hydrolase (9), indicates the presence of the sox enzyme system (10). The central importance of soxB and soxY in the sox gene system suggests that the study of the expression of these genes could be useful for monitoring the regulation of inorganic sulfur oxidation across different stages of growth in bacteria.
Thiosulfate oxidation plays a major role in the biological sulfur cycle, where it yields energy for bacterial cellular metabolism and growth in a variety of anaerobic marine environments (11–14), such as sediments or the interface between anoxic and oxic waters, where reduced sulfur compounds abound (7, 15–17). Still, few studies have investigated the ecological benefit and/or the potential function of sox genes of marine bacteria in aerobic seawater. In the Roseobacter clade member Ruegeria pomeroyi, thiosulfate oxidation increases the growth yield by up to 52% in acetate-grown aerobic batch culture compared to that in cultures without thiosulfate (1). Also, in Pseudovibrio, growth in aerobic cultures was stimulated by thiosulfate (16). In various contexts, it has been suggested that additional energy sources, such as reduced inorganic sulfur compounds, could benefit heterotrophic bacteria by supplying energy (1, 18–20). Thus, analysis of sox gene transcription patterns could potentially provide information on how this enzyme system is used by marine bacteria.
Although rather unexplored in the field of marine microbial ecology, anaplerotic reactions, including anaplerotic CO2 fixation, are a conserved feature in the metabolism of all living organisms (21, 22). Thus, most heterotrophic bacteria assimilate CO2 into biomass through various carboxylase reactions (23–25). These include, for example, anaplerotic enzymes, such as malic enzyme, phosphoenolpyruvate (PEP) carboxykinase, pyruvate carboxylase, and PEP carboxylase, that assimilate CO2 and thereby replenish carbon molecules in the tricarboxylic acid cycle when cycle intermediates are used for biosynthesis. Such anaplerotic CO2 fixation depends on the availability of substrate, the species composition in bacterial communities, and the metabolic state of the organisms (26–29). Still, in contrast to the fixation of inorganic carbon by ribulose-1,5-bisphosphate carboxylase/oxygenase or other autotrophic CO2 fixation pathways, anaplerotic CO2 fixation does not result in net CO2 fixation in heterotrophic bacteria. Recent work on natural bacterioplankton communities from northwestern U.S. Pacific coastal waters, using stable isotope probing, shows that representatives of a very wide variety of major taxa have the ability to carry out CO2 fixation through alternative pathways (26), including different members of the SAR11 and Roseobacter clades. These findings extend a previous report of significant bicarbonate uptake by prokaryotes at stationary phase in Arctic seawater cultures incubated in the dark (30).
Few studies have explored the potential for anaplerotic CO2 fixation in model species of marine heterotrophic bacteria. In their genome analysis of Roseobacter denitrificans, Swingley et al. (19) showed the presence of a set of genes encoding anaplerotic enzymes that could be involved in anaplerotic CO2 fixation. Thereafter, Tang et al. (20) revealed that R. denitrificans could fix CO2 by pyruvate and/or phosphoenolpyruvate carboxylase. These studies thus support earlier suggestions of the CO2-fixing role of enzymes, such as pyruvate carboxylase (1, 29). Also in the proteorhodopsin-containing organism Polaribacter sp. strain MED152, anaplerotic enzymes were suggested to account for the observed stimulation of CO2 fixation in cultures exposed to light compared to the levels of CO2 fixation of cultures in the dark (18). Another anaplerotic reaction to fix inorganic carbon is through reductive carboxylation in the ethylmalonyl coenzyme A (CoA) pathway, mediated by the enzymes crotonyl-CoA carboxylase/reductase and propionyl-CoA carboxylase, as originally described in the purple sulfur bacterium Rhodobacter sphaeroides (24). The genes encoding these carboxylases are present in all the organisms suggested to use the ethylmalonyl-CoA pathway for acetate assimilation (31). Taken together, these findings imply that experimental work establishing links between bacterial growth, anaplerotic CO2 fixation, and the expression of genes involved in these processes in marine bacteria would be important.
Bacteria in the Roseobacter clade are widespread inhabitants of surface seawater, but they are also found in other marine environments, including the deep sea and anaerobic sediments (17, 32–34). While they are most abundant in coastal waters, Roseobacter clade bacteria are also frequent in open-ocean environments (35–37). The metabolism of roseobacters is highly versatile (21, 37, 38), with roseobacters encompassing bacteria with the capability for aerobic anoxygenic photosynthesis; anaplerotic CO2 fixation; aromatic compound degradation; and H2, CO, and sulfur oxidation (1, 21, 39–42). Genes involved in sulfur and nutrient transformations and anaplerotic CO2 fixation are among those highly expressed in oxic marine environments (3, 43, 44). Still, compared to the large amount of knowledge on the role of this group in organic sulfur compound transformations, e.g., dimethylsulfoniopropionate, that has been gained (35, 45), little or no information about the growth and transcriptional response of Roseobacter clade members upon exposure to reduced inorganic sulfur compounds in oxic marine environments is available.
The aim of the present study was to gain insights into how inorganic sulfur oxidation impacts the metabolism of marine bacteria. We therefore carried out experiments with the Roseobacter clade isolate Phaeobacter sp. MED193 in minimal medium with and without thiosulfate. Phaeobacter sp. MED193 has a sox gene cluster comprising 11 genes, including soxXYZABCD, essential for thiosulfate oxidation (soxEFG and soxH are absent). We investigated the temporal dynamics in growth responses, protein synthesis, and anaplerotic CO2 fixation rates from early exponential phase to late stationary phase. Further, to characterize how these processes are regulated, we quantified the gene expression profiles of sulfur oxidation genes (soxB and soxY) and anaplerotic CO2 fixation genes over time.
MATERIALS AND METHODS
Bacterial isolate and culture conditions.Phaeobacter sp. strain MED193 was originally isolated from surface water (0.5-m depth) in the northwestern Mediterranean Sea at the Blanes Bay Microbial Observatory (41°40′N, 2°48′E). Contextual information on the isolation of the strain in relation to other bacterioplankton has been reported by Lekunberri et al. (46). MED193 was obtained by cultivation on ZoBell agar plates and stored at −80°C in marine broth (Difco) with glycerol at a final concentration of 25%. The GenBank accession number of the MED193 genome sequence is NZ_AANB00000000. The alignment used to construct the phylogenetic tree was generated using the MUSCLE program (47) and trimmed using trimAl software (48) to eliminate highly diverged regions (Fig. 1). The maximum likelihood tree was inferred with the RAxML program (49), using the GTR model with gamma distribution of rates (implemented as GTRGAMMA). Bootstrap analysis was performed with 100 replicates.
Maximum likelihood phylogenetic tree based on nearly complete 16S rRNA gene sequences depicting Phaeobacter sp. MED193 to be closely related to roseobacters. Grouping into clade 1 to 5 roseobacters follows the nomenclature by Newton et al. (38). For taxa with full Latin names, 16S rRNA gene sequences were from type strains of taxonomically validly described bacterial species. GenBank accession numbers are in parentheses. The sequence of Rhodobacter sphaeroides (GenBank accession number X53853) served as an outgroup (not shown). Numbers at the nodes indicate bootstrap values larger than 50%. The scale bar denotes the number of substitutions per nucleotide position.
Isolate MED193 was grown in minimal medium with 5 mM acetate as the carbon source. The minimal medium contained 750 ml of artificial seawater (28.5 g of artificial sea salt per liter; Sigma Chemical Co., St. Louis, MO) and was supplemented with 150 ml of Tris-HCl solution (containing 1 M Tris-HCl with the pH adjusted to 7.5, 87 mg K2HPO4, and 1.5 g NH4Cl). The medium was also supplemented with 50 ml of 0.1 mM FeEDTA solution (Sigma Chemical Co., St. Louis, MO), 0.1% vitamin solution (which contains 2 mg of biotin and folic acid, 10 mg pyridoxine-HCl, and 5 mg each of riboflavin, thiamine, nicotinic acid, pantothenic acid, and p-aminobenzoic acid with 0.1 mg cyanocobalamin per liter), and 0.1% mineral solution [3.92 mg CuSO4·5H2O, 13.63 mg ZnSO4·7H2O, 1.3 mg NiCl2, 3.82 mg Co(NO3)2·6H2O, 3.82 mg H3BO3 per liter]. Also, the vitamin levels of yeast extract were added (50 mg per liter; this concentration does not in itself support growth).
Bacterial growth experiments.Three sets of growth experiments were done with Phaeobacter sp. MED193. For each experiment, cells were first grown overnight in marine broth, after which for each treatment and replicate 1 ml of an overnight culture (optical density [OD] at 600 nm [OD600], 1.13) was transferred into 300 ml of minimal medium with 5 mM acetate. All treatments and controls were done in three biological replicates, and the cultures were incubated in the dark at 22°C on shakers at 142 rpm. Growth was measured by determining the OD600 (Beckman DU 640 spectrophotometer).
In experiment 1, the acetate medium was complemented with four concentrations of thiosulfate (10, 20, 40, and 100 mM), while the controls received no thiosulfate. Growth was measured through determination of the OD. In experiment 2, to determine thiosulfate utilization during bacterial growth, acetate medium with 10 mM thiosulfate was used (the controls received no thiosulfate). Growth was measured through determination of the OD, and the concentration of thiosulfate was monitored by the iodometric/thiosulfate titration method (50). Five-milliliter samples from each triplicate culture were collected at 0, 4, 6, 10, 14, 24, 30, and 36 h of incubation. Each sample received 500 μl of formaldehyde (37%) and 1 ml of sodium acetate buffer (containing 120 g/liter of sodium acetate and 10 ml/liter concentrated acetic acid [1%, vol/vol]) with 600 μl of iodine solution (0.05 M). Thereafter, the mixture was titrated with 0.01 M standard thiosulfate solution until the pale yellow color disappeared. The concentration of thiosulfate ([S2O32−]) in the sample was calculated by using the following equation: [(2 · D · 0.05) − (F · 0.01)]/C, where D is the volume of iodine (in ml) added to sample, F is the volume of titrant (in ml) used in the sample, and C is the volume of sample (in ml). To control for the chemical stability of thiosulfate in the medium used, thiosulfate concentrations were also determined in separate triplicate bottles with medium only (no bacterial inoculum was added).
In experiment 3, the acetate medium was supplemented with 10 mM thiosulfate, while the controls received no thiosulfate. In this experiment, in addition to determination of the OD, we determined protein synthesis rates (through leucine incorporation rate measurements) and anaplerotic CO2 fixation rates (through bicarbonate incorporation rate measurements) and collected RNA samples for gene expression analyses at five time points.
Leucine incorporation rates.Protein synthesis was estimated by determination of the level of incorporation of leucine into bacterial protein (51). Triplicate samples of 1.2-ml bacterial cultures were incubated with 40 nM [3H]leucine (specific activity, 172 Ci mmol−1; Amersham). Samples were incubated in the dark at 22°C for 0.45 to 2.50 h, and controls were killed with 120 μl of 50% cold trichloroacetic acid (TCA). The incubation was stopped by adding 120 μl of 50% cold TCA to each tube. Samples were stored at −20°C and then processed by the centrifugation method (52). The radioactivity in the samples was measured by liquid scintillation counting (Win spectral 1414; Wallac, Finland).
Anaplerotic CO2 fixation.Bicarbonate uptake rates were measured essentially as previously described (18). At five time points, duplicate 20-ml samples of bacterial culture were collected. Controls received 2 ml of formaldehyde (final concentration, 2%). Twelve microliters of [14C]bicarbonate (final concentration, 3 μCi; DHI, Denmark) was added to the controls and the samples, followed by incubation in the dark at 22°C for 3 to 4 h. Incubation was terminated by addition of 2 ml of formaldehyde solution to the samples. Subsequently, samples were filtered through 0.2-μm-pore-size polycarbonate membrane filters (diameter, 25 mm; Pall), and the filters were exposed to fumes of concentrated HCl for 12 to 24 h. Then, the filters were placed in scintillation vials with 4 ml of scintillation cocktail (Perkin-Elmer), stored in the dark for at least 24 h, and counted by liquid scintillation. Bicarbonate uptake rates were calculated on the basis of the standard radioactive carbon assimilation technique (53). For both anaplerotic CO2 fixation and leucine incorporation rates, statistical analyses were done by one-way analysis of variance (ANOVA), followed by Tukey's post hoc honestly significant difference (HSD) test, using GraphPad Prism (v6.0) software (GraphPad Software, San Diego, CA). Differences with P values of <0.05 were regarded as significant.
RNA isolation.Samples for RNA isolation were collected from each replicate culture during experiment 2 after 4, 7, 10, 24, and 26 h of incubation. One milliliter of bacterial culture was stabilized by addition of 2 ml of RNAprotect bacterial reagent (Qiagen, Valencia, CA) to the culture. Cells were thereafter harvested by centrifugation at 15,000 × g for 10 min, the supernatant was discarded, and the pellet was immediately frozen at −80°C. Subsequently, the pellets were resuspended in RNA collection tubes and total RNA extraction was performed using an RNeasy minikit according to the manufacturer's instructions (Qiagen, Valencia, CA). In the final step, RNA was eluted in 50 μl of nuclease-free water (Sigma). RNA was treated with DNase (Turbo DNA-free kit; Ambion) to remove genomic DNA. The total quantity of RNA and RNA purity were determined on a NanoDrop ND 2000 spectrophotometer (NanoDrop Technologies, Wilmington, MA), and samples were stored at −80°C.
Gene expression analysis by qRT-PCR.Quantitative real-time reverse transcription-PCR (qRT-PCR) was performed on an ABI StepOne real-time PCR system (Applied Biosystems). Prior to qRT-PCR, the total RNA was adjusted to a final concentration of 1 ng μl−1 with nuclease-free water. The primers listed in Table 1 were designed using the Primer3Plus web interface (54) on the basis of the genome sequence. The SYBR green-based qRT-PCR was carried out by using a one-step RT-PCR kit (Applied Biosystems), and each reaction was performed in a total volume of 10 μl consisting of 1 μl of RNA as the template, 0.08 μl of reverse transcription reagents (which include MultiScribe reverse transcriptase [50 U/ml] and RNase inhibitor [20 U/ml]), and 5 μl of 2× SYBR green master mix with 0.45 μM each primer. The one-step RT-PCRs were performed under the following conditions: 48°C for 30 min and 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. Contamination of the reaction mixtures with genomic DNA and melting curve analyses were done to confirm RNA purity and target gene amplification in the reaction. All qRT-PCRs were performed in three technical replicates for each of three biological replicate samples per treatment along with negative control PCRs (in which the reaction mixtures did not contain RNA).
Nucleotide sequences of PCR primers used to assay gene expression by qRT-PCR
To obtain adequate reference genes for normalization of gene expression levels, the 16S rRNA, rpoB, rpoD, recA, and gyrB genes were evaluated. qRT-PCRs were carried out as described above. According to the minimum information for publication of quantitative real-time PCR experiment (MIQE) guidelines, three reference genes were selected for analysis (55). The expression stability of the reference genes was determined with the Normfinder (56) and Best Keeper (57) software tools. The expression levels of the genes encoding RNA polymerase subunit D (rpoD), recombinase A (recA), and DNA gyrase (gyrB) remained stable in the treatments mentioned above; therefore, these were selected as reference genes.
The relative gene expression ratios of different genes were calculated by the comparative threshold cycle (2−ΔΔCT) method, which combines gene quantification and normalization into a single calculation, incorporating the PCR efficiencies of the target and reference genes to correct for differences between the assays (58). Statistical analyses were done by one-way ANOVA followed by Tukey's post hoc HSD test using GraphPad Prism (v6.0) software. Differences with P values of <0.05 were regarded as significant.
RESULTS
Experiment 1: impact of thiosulfate on growth.To study the effect of thiosulfate on the growth of marine heterotrophic bacteria, the Roseobacter clade strain MED193 was grown in minimal acetate medium with and without thiosulfate. Phylogenetically, strain MED193 is a member of the genus Phaeobacter (Fig. 1), which, together with, for example, Ruegeria pomeroyi DSS-3, is part of the clade 1 roseobacters, as defined by Newton et al. (38). In Phaeobacter sp. MED193, the sox gene cluster consists of the 11 genes soxXYZABCDVSRW, and this set of genes is highly conserved among distinct Roseobacter clade members (Fig. 2). In addition, several roseobacters, including Ruegeria pomeroyi (clade 1), contain the soxEF genes downstream of soxD. The Roseobacter litoralis genome further contains the soxHG genes and thus contains the full complement of sox genes (Fig. 2).
Arrangement of sox cluster genes in Phaeobacter sp. MED193 compared with that in other roseobacters. Clade designations below the strain names denote the Roseobacter clades to which the strains were assigned by Newton et al. (38) (bacteria currently assigned to clade 4 do not have sox genes). Gene homologues are indicated by colors. Genes without color in R. litoralis are genes for two conserved hypothetical proteins. Sequence information for the strains was retrieved from GenBank.
Addition of thiosulfate at 10, 20, and 40 mM concentrations improved the maximum growth yields about 30% after 24 h, while enrichment with 100 mM thiosulfate increased the growth yield nearly 40% compared to that for the unamended controls (Fig. 3A). Moreover, in the 100 mM treatment, the optical density did not decrease in the stationary phase like it did in the cultures with lower thiosulfate concentrations, but it remained stable for the last 48 h of the experiment (Fig. 3A).
Growth of MED193 in minimal medium with acetate and thiosulfate. (A) Growth in medium with 5 mM acetate and different concentrations of thiosulfate during experiment 1. (B) Bacterial growth and thiosulfate concentrations in medium with 5 mM acetate and 10 mM thiosulfate or without thiosulfate during experiment 2; white columns, thiosulfate concentrations in medium free of bacteria. Error bars denote the standard deviations from three biological replicates; if not visible, error bars are hidden in the symbols.
Experiment 2: consumption of thiosulfate.To determine the ability for consumption of thiosulfate, we analyzed the concentration of thiosulfate in the medium during different time points of bacterial growth (Fig. 3B). The concentration of thiosulfate in the medium was stable at 10 mM during early stages of growth. The concentrations of thiosulfate started decreasing at 10 h and onwards, reaching 1 mM after 36 h of incubation and well into stationary phase.
Experiment 3: impact of thiosulfate on leucine incorporation and anaplerotic CO2 fixation.A time course experiment to monitor the temporal development of metabolic activities was done by growing MED193 in cultures with 5 mM acetate with thiosulfate (10 mM) and without thiosulfate (Fig. 4A). In agreement with the findings in experiment 1, after 24 h of incubation, thiosulfate increased the growth yield 28% compared to that for the unamended control. As expected due to the active growth, leucine incorporation rates in exponential phase were ca. 8-fold higher than those in stationary phase (F = 16.95, P < 0.05). Similarly, anaplerotic CO2 fixation rates were >10-fold higher during exponential phase (F = 20.57, P < 0.05). Leucine incorporation rates during the exponential phase were, on average, 67% higher for cultures with thiosulfate than for the controls (F = 18.32, P < 0.05), while no differences in anaplerotic CO2 fixation between cultures with thiosulfate and the controls were observed at these times (F = 0.91, P > 0.05). Although the overall metabolic activities were much lower in stationary phase (Fig. 4B), the presence of thiosulfate at this phase caused an approximately 5-fold stimulation of leucine incorporation compared to that for the controls (F = 46.94, P < 0.05). The thiosulfate-induced response of anaplerotic CO2 fixation rates in stationary phase was even more pronounced, with an increase of up to 35-fold being detected in cultures with thiosulfate compared to that in control cultures (F = 21.29, P < 0.05) (Fig. 4C).
Dynamics of growth and metabolic activities of MED193 with and without thiosulfate during experiment 3. (A) Growth in acetate minimal medium in the presence (10 mM) and absence of thiosulfate. (B, C) Temporal development of leucine incorporation rates, i.e., protein synthesis (B) and CO2 fixation (C), with thiosulfate (dark columns) and without thiosulfate (open columns). (Inset) The difference in anaplerotic CO2 fixation levels between the cultures with and without thiosulfate after 24 and 30 h of growth. Error bars denote the standard errors from three biological replicates; if not visible, error bars are hidden in the symbols. Asterisks denote significance levels (*, P < 0.05; **, P < 0.01; ***, P < 0.001) by one-way ANOVA and Tukey's post hoc HSD test.
Experiment 3: impact of thiosulfate on gene expression patterns.The relative gene expression levels of the soxB and soxY genes were similar to those for the controls after 4 h. Then, from 7 h onwards, expression levels were nearly 5-fold higher in cultures with thiosulfate than in cultures without thiosulfate. The soxB gene was upregulated in cultures with thiosulfate after 7 h to 26 h, and relative gene expression levels peaked at 7 to 10 h and started declining toward stationary phase (one-way ANOVA, Tukey's post hoc test, P < 0.05) (Fig. 5A). The soxY gene expression level was even more strongly stimulated by thiosulfate from 7 h onwards, and transcript abundance declined only slowly into stationary phase (F = 27.13, P < 0.05) (Fig. 5B).
Relative gene expression patterns in MED193 grown in the presence (dark columns) and absence (open columns) of thiosulfate during experiment 3. The relative expression profiles of genes involved in sulfur oxidation (A, B) and anaplerotic CO2 fixation (C to F) are shown. Quantitative real-time PCR was performed using RNA from cultures in different growth phases from the experiment whose results are shown in Fig. 4. Relative levels of gene expression were obtained after normalization to the levels of expression for the housekeeping genes rpoD, recA, and gyrB. Error bars denote the standard deviations from three biological replicates. Asterisks denote significance levels (*, P < 0.05; **, P < 0.01; ***, P < 0.001) by one-way ANOVA and Tukey's post hoc HSD test.
For anaplerotic enzyme (CO2 fixation) genes, the relative expression level of the pyruvate carboxylase (pyc) (Fig. 5C) gene was about 1.0 in the exponential growth phase and decreased to about 0.4 at stationary phase. The phosphoenolpyruvate carboxykinase (pckA) expression level increased from a low of about 0.1 at 4 h to a peak of 1.5 at 10 h in the controls, while the peak was delayed until 24 h with thiosulfate (Fig. 5D). The highest level of crotonyl-CoA carboxylase/reductase (ccr) relative gene expression (nearly 1.5) was observed after 7 h in the controls, and it remained higher in the controls than in the cultures with thiosulfate throughout the experiment (Fig. 5E). Similarly, the expression level of propionyl-CoA carboxylase gene (Fig. 5F) peaked at about 1.5 after 7 h in the controls and then decreased to a level of about 0.2 in stationary phase.
DISCUSSION
We set out to study the impact of thiosulfate on growth, protein synthesis rates, and anaplerotic CO2 fixation, together with analysis of expression patterns of genes related to these processes, in Phaeobacter sp. MED193. The increase in growth yields of MED193 in response to thiosulfate enrichment (up to 40%) is in agreement with the finding that inorganic sulfur oxidation enhances biomass production up to 52% in the model Roseobacter clade species R. pomeroyi under similar experimental conditions (1). Consistent with these levels of growth stimulation, energy yields for thiosulfate oxidation calculated for Thiobacillus species growing under aerobic conditions range from 1.4 to 3.0 mol ATP per mol of thiosulfate (59), while energy yields from acetate are about 5 mol ATP per mol of acetate. Our finding that the most pronounced consumption of thiosulfate in MED193 occurred from mid-exponential phase onwards is highly consistent with the detailed measures of thiosulfate consumption in Thiobacillus sp. strain A2 in a mixed acetate and thiosulfate batch culture (14). We did not measure the acetate concentrations in our cultures, but it could be noted that in the Thiobacillus cultures the consumption of thiosulfate was initiated when about half of the available acetate had been consumed. These findings indicate that the ability to efficiently use thiosulfate as an energy source for growth under oxic conditions is conserved among related members of the Roseobacter clade containing a similar complement of sox genes.
Consistent with the observed stimulation of growth by thiosulfate, our results showed that the expression levels of the soxY and soxB genes were high from mid-exponential phase onwards. The expression patterns of these genes merit some further examination. First, in agreement with their close location in the genome (soxB is three genes upstream of soxY), the overall soxB and soxY expression patterns were similar, although soxY expression was slightly more dynamic. Second, we note that both genes were expressed at low but detectable levels through all growth phases even when thiosulfate was not added to the medium; i.e., they showed constitutive expression at low levels. These expression patterns are consistent with previous observations that thiosulfate stimulated sulfur oxidation gene expression in the phototrophic purple sulfur bacterium A. vinosum during anaerobic growth (60). Third, in the cultures with thiosulfate, soxB and soxY expression was as low as that in the control cultures in early exponential phase (4 h; OD, 0.07), while expression levels increased nearly 3- and 21-fold until late exponential phase (10 h; OD, about 0.20). It is possible that these low values at early exponential phase were due to a temporal delay in the sensing of thiosulfate in the medium. Still, considering that the biomass at this early time point had already doubled and that, for example, the genes for anaplerotic CO2 fixation were highly expressed at this time, we find it more likely that at this early time the cell metabolism of MED193 was entirely directed toward organoheterotrophy, and only when acetate consumption had progressed did the cells sense a change in resource availability that triggered an upregulation of lithoheterotrophy through the sulfur oxidation system. Interestingly, since the upregulation occurred only in the cultures with thiosulfate, this implies that the cells have multiple levels of regulation of the sox system, allowing them to integrate the sensing of general depletion of the medium and the presence of thiosulfate. Fourth, surprisingly, sox gene expression remained high well into stationary phase, potentially explaining the differential activity responses at this time (see below).
Metabolic activities and gene expression at exponential phase.During exponential growth, addition of thiosulfate did not influence anaplerotic CO2 fixation and gene expression rates, although it did affect the rate of bacterial leucine incorporation, which is a standard measure of protein synthesis. Moreover, our study showed that not only bacterial leucine incorporation rates but also anaplerotic CO2 fixation rates were directly dependent on the growth phase, with the rates being up to 20-fold higher during exponential growth compared to those during stationary phase. Very few studies have so far explored the relation between growth and anaplerotic CO2 fixation in heterotrophic marine bacteria (18, 29). Working with Black Sea isolates, Sorokin (29) likewise found the highest bicarbonate uptake rates at exponential growth phase. Moreover, incubation experiments with natural seawater showed elevated rates of anaplerotic CO2 fixation in stationary-phase cells, as detected by microautoradiography coupled with catalyzed reporter deposition-fluorescent in situ hybridization (fixation rates were not measured during the active growth phase) (30). Heterotrophic bacteria could assimilate significant amounts of CO2 into cell material using anaplerotic CO2 fixation pathways via various carboxylases (28, 61). Roseobacters are known to contain genes encoding anaplerotic enzymes that could be involved in anaplerotic CO2 fixation, such as pyruvate carboxylase (19, 38). Experimental work with R. denitrificans showed that while a majority of the cellular carbon during late exponential phase was derived from organic matter, between 10 and 15% came from CO2 fixed by anaplerotic enzymes (20). We did not directly measure how much the anaplerotic CO2 fixation in MED193 contributed to its cellular carbon budget. However, a calculation from the increase in abundance and the measured anaplerotic CO2 fixation rates during active growth would result in at least 2 to 5% of the cell carbon in MED193 being provided by anaplerotic CO2 fixation. These findings suggest that anaplerotic CO2 fixation could be more important than has hitherto been recognized during active growth of heterotrophic marine bacteria.
Considering the large changes in anaplerotic CO2 fixation from exponential to stationary phase in our experiments and the recognition of pyruvate and PEP carboxylases as key anaplerotic enzymes (19, 20), we expected to observe a shift in the expression levels of pyruvate carboxylase in isolate MED193 across growth phases (the genome of MED193 does not encode PEP carboxylase). However, pyruvate carboxylase gene expression remained relatively stable over time (although there was a tendency to somewhat higher levels of expression in exponential phase). Another anaplerotic enzyme encoded by the MED193 genome and potentially involved in anaplerotic CO2 fixation was PEP carboxykinase. This enzyme is involved in either gluconeogenic or anaplerotic reactions, depending on the organism (25, 62–65). Tang et al. (20) suggested that PEP carboxykinase in R. denitrificans OCh 114 was involved in anaplerotic CO2 fixation. The highly varied patterns of expression of this enzyme in our experiment gave no direct clues to its role in MED193. Nevertheless, the lack of other enzymes in this bacterium carrying out the initial step in gluconeogenesis suggests a gluconeogenic role of PEP carboxykinase in this bacterium.
In addition to pyruvate carboxylase and PEP carboxykinase, genome analysis of MED193 showed the presence of genes involved in the ethylmalonyl-CoA pathway, which allows bacteria lacking the glyoxylate shunt to grow on carbon compounds like acetate (22, 31, 66, 67). An important step in this pathway is the fixation of 1 molecule of CO2 and 1 molecule of bicarbonate per 3 molecules of acetate, whereby it fulfills an important anaplerotic reaction (24). In our study, the gene expression level of the two main genes in the ethylmalonyl-CoA pathway, i.e., those encoding crotonyl-CoA carboxylase/reductase and propionyl-CoA carboxylase, was up to 5-fold higher in the exponential phase than in stationary phase, coinciding with elevated levels of anaplerotic CO2 fixation during active growth. This suggests that MED193 uses the ethylmalonyl-CoA pathway for growth on acetate and, further, that the activity of the carboxylases in this pathway could account for important parts of the increased level of anaplerotic CO2 fixation during active growth.
Metabolic activity and gene expression at stationary phase.Although the rates of leucine incorporation and anaplerotic CO2 fixation were much lower in stationary phase than in exponential phase, we found a very strong positive impact of thiosulfate on the metabolic activities of Phaeobacter sp. MED193 in stationary phase. Ultimately, the maintenance of higher levels of metabolism during starvation could lead to prolonged survival, which was also indicated by the more stable OD values in stationary phase with 100 mM thiosulfate treatment than with treatments with lower concentrations. The ecological significance of such prolonged survival could be expressed either through the provision of additional energy to all cells by thiosulfate or through the selection of surviving cells and manifested through the resulting high per cell activity. Our gene expression analysis showed that the sox genes were highly expressed across growth phases from mid-exponential phase and well into stationary phase. We are unaware of studies detailing the progression from transcription of the sox genes to the maturation of the Sox enzyme system and how long it remains functional. Nonetheless, the continued sox gene expression indicates that the cells continued the synthesis of the Sox enzyme system, even though it should become progressively more costly the further that the cells advanced into stationary phase. Thus, our results suggest that, in addition to the importance of thiosulfate for providing energy for increased growth yields in marine bacteria, as observed here and elsewhere (1, 68), thiosulfate could play a largely unrecognized role in providing energy for maintaining the activity of cell metabolism under nutrient-depleted conditions and thereby lead to prolonged survival.
With regard to how energy from thiosulfate oxidation is used in stationary phase, the present data do not allow detailing of whether the pathway for anaplerotic CO2 fixation in stationary phase was the same as that in exponential phase. However, the gene expression levels of the ethylmalonyl-CoA pathway decreased, while those of the pyruvate carboxylase pathway remained stable, suggesting that anaplerotic CO2 fixation through the latter pathway may have increased in relative importance over time. In general, fatty acid and amino acid biosyntheses in heterotrophic bacteria are among the processes that require anaplerotic reactions. In Escherichia coli, the entry into stationary phase implies major rearrangements in protein and lipid composition, and lipid synthesis at this stage has been shown to require an increased amount of anaplerotic CO2 fixation (69). This supports our contention that thiosulfate could provide energy for cell metabolism in stationary-phase cells that require complementary anaplerotic CO2 fixation.
Conclusion.The present study showed that the utilization of thiosulfate influenced bacterial metabolism during the active growth phase as well as in stationary phase. This implies that lithotrophy in this bacterium positively affects the two major fitness components, reproduction and survival. From an ecological perspective, it remains an intriguing issue to determine under which environmental conditions either of the two is more beneficial to the bacteria. The relative benefit of using thiosulfate for obtaining higher abundances or improved survival is likely to vary depending on the stability of the resource supply, the availability of organic matter compared to that of reduced inorganic substrates, and maybe also the inorganic nutrient concentrations. Overall, these results suggest that the genetic potential for sulfur oxidation via the sox enzyme system improves the metabolic versatility of marine bacteria facing ever-changing environmental conditions, potentially allowing a more efficient utilization of organic carbon to enhance growth and to contribute to cell energetics during starvation.
ACKNOWLEDGMENTS
We thank Jeremy Forsberg, Johanna Sjöstedt, Joakim Palovaara, Mark Dopson, Christopher Abin, and Sabina Arnautovic for skillful technical assistance and helpful advice during the experiments and for providing important food for thought.
This research was supported by the Crafoord Foundation, European Science Foundation/Swedish Research Council EuroEEFG project MOCA, BONUS project BLUEPRINT, and the research program EcoChange (to J.P.) and by grant MarineGems (CTM2010-20361) from the Spanish Ministry of Science and Innovation (to J.M.G.).
FOOTNOTES
- Received 19 June 2014.
- Accepted 27 August 2014.
- Accepted manuscript posted online 29 August 2014.
- Copyright © 2014, American Society for Microbiology. All Rights Reserved.
