OUP user menu

Dimethylsulfide is an energy source for the heterotrophic marine bacterium Sagittula stellata

Rich Boden, J. Colin Murrell, Hendrik Schäfer
DOI: http://dx.doi.org/10.1111/j.1574-6968.2011.02349.x 188-193 First published online: 1 September 2011


Dimethylsulfide (DMS) is a volatile organosulfur compound, ubiquitous in the oceans, that has been credited with various roles in biogeochemical cycling and in climate control. Various oceanic sinks of DMS are known – both chemical and biological – although they are poorly understood. In addition to the utilization of DMS as a carbon or a sulfur source, some Bacteria are known to oxidize it to dimethylsulfoxide (DMSO). Sagittula stellata is a heterotrophic member of the Alphaproteobacteria found in marine environments. It has been shown to oxidize DMS during heterotrophic growth on sugars, but the reasons for and the mechanisms of this oxidation have not been investigated. Here, we show that the oxidation of DMS to DMSO is coupled to ATP synthesis in S. stellata and that DMS acts as an energy source during chemoorganoheterotrophic growth of the organism on fructose and on succinate. DMS dehydrogenase (which is responsible for the oxidation of DMS to DMSO in other marine Bacteria) and DMSO reductase activities were absent from cells grown in the presence of DMS, indicating an alternative route of DMS oxidation in this organism.

  • chemoorganoheterotrophy
  • mixotrophy
  • energy metabolism
  • marine microbiology
  • bioenergetics
  • chemostat kinetics


Dimethylsulfide (DMS) is a volatile organosulfur compound ubiquitous in marine environments that has been implicated in playing major roles in both climate control and in the biogeochemical cycling of sulfur (Charlson et al., 1987; Bentley & Chasteen, 2004). Chemical and biological transformations serve as major sinks for DMS in the oceans, although the mechanisms and organisms responsible for the biological transformations are poorly understood (reviewed in Schäfer et al., 2010).

The biological production of dimethylsulfoxide (DMSO) in the environment has been well documented in the literature, particularly for marine systems, and is associated with both Eukarya and Bacteria (Hatton, 2002; del Valle et al., 2007, 2009), although the exact mechanism of the oxidation remains unknown. Various hypotheses have been put forward regarding the oxidation of DMS to DMSO by marine Bacteria, although the purpose of the oxidation is, to date, unknown. Light-stimulated DMSO production has led to the hypothesis that phototrophic Bacteria may use DMS as an energy source in the environment as observed in pure cultures (reviewed in Hatton, 2002). It is also possible that the oxidation of DMS to DMSO is chemically mediated by oxygen-free radicals (Snow et al., 1976) and as such DMS could act as an antioxidant within living systems, as has been found with other sulfur compounds in Spirillum winogradskii (Podkopaeva et al., 2005). It has been shown recently (Green et al., 2011) that a number of marine Bacteriodetes isolates are capable of oxidizing DMS to DMSO during growth on glucose, with some increase in the amount of biomass formed during growth. Muricauda sp. DG1233 was studied in batch cultures and was shown to exhibit small increases in the amount of biomass formed; although DMSO production was monitored, glucose consumption was not, and so it is not possible to determine the increase in yield from these data. It was suggested by Green . (2011) that the increase in biomass production in the presence of DMS could be due to the organism harnessing electrons from the DMS to DMSO oxidation and passing them onto the respiratory chain. This was not further investigated, nor was the role of DMS as an antioxidant ruled out.

Photoorganoautotrophic Bacteria (such as Rhodovulum sulfidophilum) can use DMS as an energy source, producing DMSO in a pure culture. This has been shown to be catalyzed by DMS dehydrogenase, which has been purified and characterized from R. sulfidophilum (McDevitt et al., 2002). The oxidation of DMS to DMSO (without assimilation of DMS-carbon) in nonphototrophic Bacteria has been reported previously during the heterotrophic growth of Delftia acidovorans DMR-11 (previously ‘Pseudomonas acidovorans DMR-11’; Zhang et al., 1991) and in Sagittula stellata (González et al., 1997), but the purpose of this oxidation and the mechanisms behind it are not known. The aim of this study was to determine the role of DMS oxidation during the growth of S. stellata.

Materials and methods

Organisms and materials

Sagittula stellata DSM 11524T (E37T) was obtained from the Deutsche Sammlung von Mikroorganismen und Zellkulturen (Braunschweig, Germany). Hyphomicrobium sulfonivorans S1T was a gift from Dr Ann P. Wood (King's College London, UK). Rhodovulum sulfidophilum SH1 was a gift from Dr Ben Berks (University of Oxford, UK). All reagents were obtained from Sigma-Aldrich and used without prior purification, with the exception of NADH, which was first washed to remove traces of ethanol according to Boden . (2010).

Analytical techniques

DMS was quantified by GC according to Schäfer (2007). DMSO was quantified after reduction to DMS. One volume of sample was treated with nine volumes of 0.1M stannous chloride in concentrated hydrochloric acid at 90°C for 2h. Vials were then cooled before the determination of headspace DMS (Li et al., 2007). ATP was extracted and quantified as described (Boden et al., 2010). Succinate was quantified using the K-SUCC Succinate Assay Kit (Megazyme, Bray, Eire); fructose was quantified using the FA20 Fructose Assay Kit (Sigma-Aldrich), both according to the manufacturers' instructions.

Growth and harvesting of cells

Continuous-flow chemostat cultures using marine ammonium mineral salts medium for the cultivation of S. stellata were operated essentially as described by Boden . (2010), with the exception that the rate of agitation was 350r.p.m. Carbon sources were dissolved in the medium and DMS supplied from 10mM stock solutions at one tenth of the medium flow rate to yield an effective final concentration of 1mM. Hyphomicrobium sulfonivorans S1T was grown in a batch culture on dimethylsulfone as described previously by Boden . (2011) and R. sulfidophilum was grown photoorganoautotrophically on DMS according to McDevitt . (2002).

Sagittula stellata was grown in steady-state chemostats at a range of dilution rates (D) between 0.01 and 0.15h−1 on fructose (12mM) and between 0.01 and 0.10h−1 on succinate (2mM) with or without the addition of DMS (1mM). Kinetic parameters were determined as described previously (Boden et al., 2010). Five volume changes at each steady state occurred before the kinetic parameters were determined.

Cells were harvested for enzyme assays by centrifugation at 13000g for 30min at 4°C. Cells were washed and resuspended in 50mM PIPES-HCl, pH 7.4, containing 50mM magnesium sulfate. If not used immediately, cells were snap-frozen in liquid nitrogen and stored at −80°C.

Enzyme assays

Spectrophotometric enzyme assays were routinely conducted at 30°C in an Ultrospec 3100pro UV/Visible Spectrophotometer (Amersham). Each reaction was conducted in a sevenfold replicate against a blank. Cell-free extracts were prepared by three passages through a French pressure cell (120MPa), with debris removed by centrifugation (13000g, 30min, 4°C). Protein was quantified using the method of Bradford (1976).

DMS dehydrogenase activity was assayed using a modification of the method of McDevitt . (2002). Two milliliters of 500mM Tris-HCl, pH 8.0, 300μL of 35mM phenazine methosulfate and 300μL of 100μM 2,6-dichlorophenolindophenol (DCPIP) were placed in a 3mL modified Thunberg cell (Baumberger, 1933) and degassed by bubbling with oxygen-free nitrogen for 10min before adding 100μL cell-free extract (containing 5–10-mg protein). Sixty microliters of 100mM DMS solution in ethanol was placed in the bulb of the side-arm and the cell was assembled. The cell was evacuated on ice for 10min before sealing and the reaction was initiated by pouring the contents of the side-arm into the main chamber. The A600nm was monitored and the rate of reduction of DCPIP was determined using the millimolar extinction coefficient for the oxidized form of 21.5mM−1cm−1. Cell-free extracts prepared from R. sulfidophilum SH1 grown photoorganoautotrophically with DMS as an energy source were used as a positive control (McDevitt et al., 2002). An alternative assay was performed using 300μL of 3mM potassium ferricyanide in place of DCPIP solution and reduction was monitored at 420nm with a millimolar extinction coefficient of 1.0mM−1cm−1.

DMSO reductase was assayed in the same way using a reaction mixture comprising 150μL of 1.0M Tris-HCl, pH 7.6, 20μL of cell-free extract and 1.03mL of MilliQ water in the main chamber of the cell. These were degassed in situ before adding 1.5mL of 2mM dithionite-reduced methyl viologen (MV) in 50mM Tris-HCl, pH 7.6. Three hundred microliters of 50mM DMSO was placed in the bulb of the side arm and was then used to initiate the reaction. The oxidation of MV was monitored by the decrease in A600nm and the rate of oxidation was determined using the millimolar extinction coefficient of the reduced form, being 1.13mM−1cm−1 (Kelly & Wood, 1994). Cell-free extracts prepared from H. sulfonivorans S1T grown heterotrophically on dimethylsulfone were used as the positive control.

Monitoring of DMS-dependent ATP production

ATP production experiments were performed essentially as described previously (Boden et al., 2010) using 1mM DMS as an energy source in place of thiosulfate.

Results and discussion

Chemostat kinetics

The kinetic parameters derived from the growth of S. stellata in chemostat culture on fructose (12mM) or succinate (2mM) are given in Table 1. The maximum yield coefficient (Ymax) increased in the presence of DMS, which was oxidized stoichiometrically to DMSO without assimilation into biomass. No DMS was detected in the cultures in a steady state. Upon the addition of DMS to a succinate or a fructose-limited chemostat, there was no immediate perturbation of the steady state and the dissolved oxygen concentration did not begin to decrease for approximately 6h in the case of fructose or 3h in the case of succinate, independent of the dilution rate. The delay in oxygen consumption in the presence of DMS would indicate that the enzyme system for DMS oxidation was not constitutively expressed and the culture essentially underwent a lag phase while expression was induced. While the Ymax increased, it should be noted that the maintenance coefficient (mS) remained constant in the case of both carbon sources used. This was also the case when thiosulfate was used to support the chemolithoheterotrophic growth of Methylophaga thiooxydans (Boden et al., 2010) and mixotrophic growth of Acidithiobacillus thiooxidans (Mason & Kelly, 1988). As stated previously, it is not possible to compare these data with those of Green . (2011) owing to insufficient data being available from their paper to calculate Y– i.e. without quantifying substrate disappearance, Y cannot be calculated.

View this table:
Table 1

Growth of Sagittula stellata in fructose or succinate-limited chemostats with or without DMS as a supplementary substrate

Limiting substrateSupplementary substrateYmax (gdrybiomassmol−1substrate)Ymax (gdrybiomassmol−1substratecarbon)mS (mmolcarbonsourceg−1h−1)μmax (h−1)
Fructose (12mM)None64.
Fructose (12mM)DMS (1mM)
Succinate (2mM)None33.
Succinate (2mM)DMS (1mM)
  • Ymax, Maximum growth yield; mS, maintenance energy coefficient; μmax, maximum specific growth rate. At steady state, neither the limiting substrates nor DMS could be detected in the culture, although DMSO was present at concentrations of 0.90 ± 0.03mM (fructose/DMS, n=3) and 0.90 ± 0.02mM (succinate/DMS, n=3). DMSO was not detected during growth on fructose or succinate alone.

The theoretical Ymax for growth on succinate is 37.1g drybiomassmol−1succinate (9.23gdrybiomassmol−1substrate carbon), calculated using the assumption that 32% of succinate carbon is assimilated to biomass, as per the determinations performed by Anthony (1982) in a range of organisms. The experimental Ymax for succinate was found to be 33.6gdrybiomassmol−1succinate (8.4gdrybiomassmol−1substrate carbon), which increased in the presence of DMS to 38.9gdrybiomassmol−1succinate (9.7gdrybiomassmol−1substrate carbon) – this is higher than the theoretical Ymax and a 16% increase on the Ymax in the absence of DMS.

The theoretical Ymax for growth on fructose dissimilated to 3-phosphoglycerate via the Entner–Doudoroff pathway is 73.7gdrybiomassmol−1fructose (12.3gdrybiomassmol−1 substrate carbon), whereas that for fructose dissimilated via the Embden–Meyerhof–Parnas pathway is 122.4gdrybiomassmol−1fructose (20.4gdrybiomassmol−1substrate carbon). Genes encoding the enzymes required for both pathways of glycolysis are present in the genome of S. stellata. The Ymax observed during fructose-limited growth (64.2gdrybiomassmol−1fructose) is closer to that predicted for dissimilation via the Enter–Doudoroff pathway, suggesting that it is probably in use in this case; however, it must be noted that many bacteria use multiple pathways of glycolysis simultaneously during growth on hexoses (Wood & Kelly, 1977). The Ymax in the presence of DMS (73.2gdrybiomassmol−1fructose, a 14% increase in Ymax from growth on fructose alone) is closer to the theoretical Ymax, indicative of a tighter coupling of fructose oxidation to growth in the presence of DMS, with less dissimilation to carbon dioxide to meet the energy requirements of growth and maintenance.

Enzymes of DMS metabolism

The oxidation of DMS to DMSO is catalyzed by DMS dehydrogenase in R. sulfidophilum (McDevitt et al., 2002): Embedded Image

The subunits of DMS dehydrogenase have been shown to be encoded by the operon ddhABDC (McDevitt et al., 2002). Searching the S. stellata genome using the blastp algorithm reveals predicted proteins with >55% identity to DdhABC, clustered together and annotated as components of a nitrate reductase NarYZV (EBA07058–EBA07060). It is worth noting that González . (1997) did not observe nitrate reduction during heterotrophic growth of S. stellata under anoxic conditions.

Additionally, genes annotated as a DMSO reductase-like molybdopterin-containing dehydrogenase are also present in the genome of S. stellata (EBA06368–EBA06370); a tblastx search against the GenBank database confirms the annotation. The oxidation of DMS to DMSO could potentially be catalyzed by this enzyme performing its reverse reaction (Adams et al., 1999).

Enzyme assays were conducted for DMS dehydrogenase and DMSO reductase on cell-free extracts prepared from cells obtained from succinate-limited chemostats (D=0.03h−1) grown with DMS (Table 2). It can be seen that DMS dehydrogenase (DCPIP or ferricyanide linked) activity was absent, although it could be assayed in the control organism R. sulfidophilum; DMSO reductase activity was also absent, but could be assayed in the positive control (H. sulfonivorans). It is, of course, possible that a DMS dehydrogenase is present in S. stellata grown under these conditions, but is either too unstable in cell-free extracts to assay or does not couple to DCPIP or ferricyanide in vitro. Additional assays were conducted in the presence of 1mM NAD+ and NADP+, but no activity was observed.

View this table:
Table 2

Specific activities of enzymes of DMS oxidation in cell-free extracts prepared from cells of Sagittula stellata obtained from a succinate-limited chemostat [dilution rate (D)=0.03h−1, [succinate]0=2mM] with DMS (1mM) as an auxiliary energy source

OrganismSpecific activity (nmolsubstratereducedmin−1mg−1protein)
DMS dehydrogenase (DCPIP)DMS dehydrogenase (ferricyanide)DMSO reductase (MV)
Sagittula stellata1 (± 1)00
Rhodovulum sulfidophilum24 (± 4)36 (± 3)0
Hyphomicrobium sulfonivorans0ND57 (± 3)
  • Positive controls are Rhodovulum sulfidophilum grown photoorganoautotrophically in a batch culture with DMS as an energy source (McDevitt et al., 2002) for DMS dehydrogenase and Hyphomicrobium sulfonivorans grown heterotrophically in batch culture on DMSO (Borodina et al., 2002) for DMSO reductase. ND, Not determined. Numbers in parentheses are SE (n=7).

DMS-stimulated ATP formation

The ATP content of whole cells obtained from a succinate-limited chemostat (D=0.03h−1) grown in the presence of DMS was monitored over time after the addition of DMS to 1mM and the results are shown in Fig. 1. It can be seen that ATP is produced in the presence of DMS by cells of S. stellata, with approximately 0.2mgATPmg−1drybiomass being formed per mole of DMS oxidized to DMSO (which is in the same order of magnitude as that produced during thiosulfate oxidation by M. thiooxydans [0.13mg, Boden . (2010)]. It is interesting to note that the production of ATP here apparently follows an exponential rather than a logarithmic pattern – as observed in M. thiooxydans and Halothiobacillus neapolitanus during thiosulfate oxidation (Kelly & Syrett, 1964; Boden et al., 2010). There is also a slight lag as ATP formation begins, suggesting that the oxidation of DMS is not immediate and that DMS must first be transported into the cells – possibly by active transport. Alternatively, this lag could be due to a high ATP demand of the cells for example, to fuel motility. This is in contrast to the immediate ATP formation during thiosulfate oxidation in M. thiooxydans and H. neapolitanus, which is thought to occur in the periplasm.

Figure 1

ATP production by cells of Sagittula stellata obtained from a succinate-limited chemostat (D=0.03h−1, [succinate]0=2mM) grown in the presence of 1mM DMS. ATP production experiments were conducted as described previously (Boden et al., 2010) using 1mM DMS in place of thiosulfate as the energy source. Hollow circles/broken line represent control incubations without exposure to DMS; solid circles/solid line represent experimental incubations with 1mM DMS. Error bars indicate SE (n=7).

The oxidation of DMS to DMSO alone provides 2mol of electrons per mole of DMS oxidized. This is not sufficient to provide the 14–16% increases in Ymax observed here. The same amount of electrons from thiosulfate oxidation in M. thiooxydans provides only a 9% increase in Ymax during growth on methanol (Boden et al., 2010). This could indicate that, in addition to providing electrons to the respiratory chain, the oxidation affects some other system within the cell that generates an increased yield of reducing equivalents that are responsible for a larger conservation of carbon into biomass. More complex radiorespirometric or metabolomic studies are required to fully investigate the pathway of DMS-dependent energy metabolism in S. stellata; however, we have demonstrated that DMS acts as an energy source for the chemoorganoheterotrophic growth of this organism on different carbon sources and that the oxidation of DMS to DMSO is coupled to ATP synthesis.

Few data are available on the kinetics and growth yields in mixotrophic bacteria – particularly those capable of chemoorganoheterotrophy – and the data we present here add to this understudied area of bacterial physiology. The regulation and environmental significance of mixotrophic Bacteria are unknown, although the substrates and products of their energy-yielding oxidations can be compounds of global biogeochemical significance – such as DMS and DMSO, which we report here. Further work is required to better the understanding of these mixed metabolic modes, their use by Bacteria in the environment and their contribution to the flux of compounds through biogeochemical cycles.


We thank Don Kelly for many stimulating discussions on growth kinetics and Gez Chapman is thanked for technical support. We thank the Natural Environment Research Council (UK) for funding via a studentship to R.B. and fellowships to H.S. (NE/B501404/1 and NE/E013333/1). Ann P. Wood and Ben Berks are thanked for the kind donation of strains.


  • Editor: Wilfrid Mitchell


View Abstract