APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Sept. 1987, p. 2026-2032 Vol. 53, No. 90099-2240/87/092026-07$02.00/0Copyright X 1987, American Society for Microbiology

Oxidation of Dimethyl Sulfide to Dimethyl Sulfoxide byPhototrophic Purple Bacteria

JOSEF ZEYER,* PETRA EICHER, STUART G. WAKEHAM,t AND RENE P. SCHWARZENBACHSwiss Federal Institute for Water Resources and Water Pollution Control (EAWAG), 6047 Kastanienbaum, Switzerland

Received 1 December 1986/Accepted 3 June 1987

Enrichment cultures of phototrophic purple bacteria rapidly oxidized up to 10 mM dimethyl sulfide (DMS)to dimethyl sulfoxide (DMSO). DMSO was qualitatively identified by proton nuclear magnetic resonance. Byusing a biological assay, DMSO was always quantitatively recovered from the culture media. DMS oxidationwas not detected in cultures incubated in the dark, and it was slow in cultures exposed to full daylight. Underoptimal conditions, the second-order rate constant for DMS oxidation was 6 day-' mg of protein-1 ml-'. Therate constant was reduced in the presence of high concentration of sulfide (>1 mM), but was not affected bythe addition of acetate. DMS was also oxidized to DMSO by a pure strain (tentatively identified as a Thiocystissp.) isolated from the enrichment cultures. DMS supported growth of the enrichment cultures and of the purestrain by serving as an electron source for photosynthesis. A determination of the amount of protein producedin the cultures and an estimation of the electron balance suggested that the two electrons liberated during theoxidation of DMS to DMSO were quantitatively used to reduce carbon dioxide to biomass. The oxidation ofDMS by phototrophic purple bacteria may be an important source of DMSO detected in anaerobic ponds andmarshes.

Global mass balances for sulfur indicate that dimethylsulfide (DMS) is the most important volatile biogenic com-pound involved in the transfer of sulfur from the ocean (2, 3,6), from marshes (15, 25), and from soils (20) to the atmo-sphere (4, 16, 21, 36). Although estimated emission ratesvary considerably, the total biogenic sulfur flux to theatmosphere is estimated to be about 100 x 1012 g of S year-'and, therefore, approximately equal to the anthropogenicflux of sulfur dioxide (SO2) (4, 21, 36). Some 50% of the totalbiogenic sulfur flux is believed to be due to DMS, whilecontributions from hydrogen sulfide (H2S), carbonyl sulfide(COS), carbon disulfide (CS2), dimethyl disulfide (DMDS),methylmercaptan (CH3SH), and biogenic SO2 are compar-atively minor (4, 25, 36). It has been calculated that >70% ofthe DMS evolved to the atmosphere originates from theoceans (4, 25).The main source of DMS in the ocean is phytoplankton,

which uses the DMS precursor dimethyl propiothetin as anosmoticum (3, 6, 7, 27). Dimethyl propiothetin is enzymati-cally cleaved to DMS and acrylic acid (10). Some grassspecies such as Spartina alterniflora can also producedimethyl propiothetin and release DMS (15, 25). DMS is alsoliberated from soils and in freshwater lakes during thedecomposition of plants (20), algae (7, 39), and especially theamino acids methionine and cysteine (17). One additionalsource of DMS is the biological reduction of dimethylsulfoxide (DMSO) (9, 12, 35, 37), a physiological product (1)and widely used organic solvent.

Interestingly, there is only a limited amount of informationon DMS sinks. Once in the atmosphere, DMS has beenreported to be photochemically converted to oxidized sulfurspecies such as DMSO and SO2 (5, 13, 36). Proposed sinks inthe aquatic environment include the anaerobic degradationof DMS to methane (CH4), carbon dioxide (CO2), and H2S

* Corresponding author.t Present address: Skidaway Institute of Oceanography, Savan-

nah, GA 31416.

(19, 38) and aerobic metabolism ofDMS to products such asformaldehyde and CH3SH (12, 18, 24).

In previous reports, we described the spatial and temporaldistribution ofDMS in coastal Salt Pond (30, 31) and in GreatSippewissett Marsh (15), both near Woods Hole, Mass. Inthe summer months, Salt Pond is a well-stratified lake withan H2S-rich hypolimnion harboring a dense population ofanaerobic phototrophic bacteria. DMS concentration pro-files in the pond showed 10 to 60 nM DMS in the epilimnionand only <2 nM in the hypolimnion (30, 31). We postulatedthat some of the anaerobic phototrophic bacteria metabolizenot only H2S but also DMS. Initial experiments with enrich-ment cultures obtained from both Great Sippewissett Marshand Salt Pond confirmed our assumption (31). We nowdescribe the conditions, rates, and products of this metabo-lism and demonstrate that anaerobic phototrophic bacteriacan utilize DMS as an electron source.

MATERIALS AND METHODSChemicals. DMS, CH3SH, dimethyl sulfone (DMSO2), and

2-bromoethanesulfonic acid sodium salt were obtained fromFluka AG (Buchs, Switzerland), Na2S 9H20 and DMSOwere from Merck AG (Darmstadt, Federal Republic ofGermany), and CH4 and N2/CO2 (90:10, vol/vol) were fromPan Gas (Lucerne, Switzerland).

Organisms. Soil and sand samples containing a densepopulation of phototrophic purple bacteria a few millimetersbelow the surface were collected from Sippewissett Marsh(15, 25). Water samples containing phototrophic green andpurple bacteria were collected at a depth of 3 to 5 m fromSalt Pond (30, 31). All samples were collected in June andJuly 1985 and mixed in the absence of molecular oxygen.This mixture was then used to inoculate enrichment cul-tures. Escherichia coli HB101, which was used for theDMSO bioassay, was a gift from J. Frey, University ofBerne, Berne, Switzerland.Media. The anaerobic marine basal medium was prepared

by the method of Widdel and Pfennig (33). It consisted of thefollowing (grams per liter of distilled water): KH2PO4, 0.20;

2026

on Novem

ber 25, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

OXIDATION OF DMS TO DMSO 2027

NH4Cl, 0.25; NaCl, 30.0; MgCl2. 6H20, 2.20; KCl, 0.50;CaCl2. 2H20, 0.15; NaHCO3, 2.52; trace element solutionSL10 (32, 33), 0.5 ml/liter; vitamin solution (33), 1.0 ml/liter.The pH of the medium was adjusted to 7.2 to 7.3. The basalmedium was supplemented with sulfide (from a 0.5 MNa2S 9H20 stock solution previously neutralized with HCl[33]), acetate (from a 1.0 M sodium acetate stock solution),and DMS (from a 0.1 M stock solution) as indicated inResults.

Culture conditions and absorption spectra. Cultures wereincubated under strictly anaerobic conditions at 22°C inserum flasks (57 ml) sealed with butyl rubber stoppers,which allowed sampling by syringes. The flasks contained 50ml of culture and a 7-ml headspace of N2/CO2 (90:10). Unlessindicated otherwise, the cultures were incubated withoutagitation under an incandescent lamp at a light intensity of 7to 12 microeinsteins m-2 s-1. Absorption spectra of cultureswere measured by suspending the cells in a 4 M sucrosesolution by the method of Truper and Pfennig (26).

Isolation and tentative identification of pure strains. Purestrains of phototrophic bacteria were isolated from theenrichment cultures by using the agar shake dilution methoddescribed by Pfennig and Truper (23). A microscopic exam-ination of the isolated colonies revealed three morphologi-cally different types of phototrophic purple bacteria, referredto as strains A, B, and C throughout this paper. Cultures ofthe three strains were purple-red (strains A and B) or

green-brown (strain C) and all had a distinct absorptionmaximum at 590, 800 to 805, and 855 to 870 nm, whichindicates the presence of bacteriochlorophyll a (26). Strain Aconsisted of spheric motile cells and was tentatively identi-fied as a Thiocystis sp. Strain B consisted of spiral motilecells and was tentatively identified as a Rhodospirillum sp.Strain C consisted of spheric nonmotile cells.

Analysis of H2S and protein. H2S and protein concentra-tions were determined in 1-ml culture samples previouslyremoved with a syringe. H2S was analyzed by a colorimetricassay, using N,N-dimethyl-p-phenylenediamine and ferricchloride (11). Total protein was quantified according to a

modified Lowry method (14). In the presence of interferingchemicals such as sulfide, proteins were precipitated withsodium deoxycholate and trichloroacetic acid (8) prior totheir quantitative determination.

Analysis of volatile compounds. Volatile compounds suchas DMS, CH3SH, DMDS, and CH4 were analyzed byinjecting 10 ,ul of the culture headspace (withdrawn with a

gas-tight syringe) into a gas chromatograph (Carlo Erba,model 2101) equipped with a glass capillary column (40 m by0.32-mm inside diameter) coated with OV-1701. The carriergas was H2 (0.3 atm) and the oven temperature was 70°C.Under these conditions, the retention times of standardcompounds were as follows: 2.8 (CH4), 3.4 (CH3SH), 4.1(DMS), and 11.8 (DMDS) min. For calibration, standardsolutions of the corresponding compound in identical serum

flasks were used. DMDS could be obtained for qualitativeanalysis by exposing an aqueous solution of CH3SH to air; a

quantitative DMDS standard, however, was not available.Analysis of DMSO. A qualitative determination of DMSO

was performed by proton nuclear magnetic resonance ('H-NMR). For the 'H-NMR analysis, 100 ml of a culture ofpurple phototrophic bacteria grown on basal medium plus 2mM acetate, 2 mM H2S, and 10 mM DMS until no DMS was

detected in the headspace was filtered, and the filtrate was

concentrated 10-fold by evaporating the water by vacuum

distillation. During the evaporation, salt precipitation was

observed; thus, the final sample was saturated with sodium

chloride. The sample was diluted (1:1) with D20 (deuteratedwater), and TSP (sodium salt of 3-trimethylsilyl-tetrade.-teropropionic acid) was added as an internal standard. The'H-NMR spectra were recorded in a 5-mm-outside diameterNMR tube on a Bruker WP-200 SY Fourier Transform NMRspectrometer (200.13 MHz; spectral width, 2,994 Hz).

Quantitative tests for DMSO are generally not based ondirect analysis but rather on a chemical reduction of DMSOto DMS which is then analyzed by gas chromatography (1).In this study, DMSO was biologically converted to DMS byE. coli HB101, a strain that has been reported to reduceDMSO to DMS under oxygen-limiting conditions with amolybdate-requiring enzyme (9). A standard assay for thequantitative determination of DMSO in the enrichment cul-tures was developed: 30 ml of complex medium [1.0%tryptone (Difco Laboratories, Detroit, Mich.) plus 0.5%yeast extract (Difco) plus 10 ,uM (NH4)6Mo17024 4H20 inwater, pH 7.2] plus 15 ml of culture to be analyzed forDMSO, plus 2 ml of freshly grown culture of E. coli HB101(precultured on complex medium) were mixed aerobically ina serum flask (57 ml). This flask was sealed with a butylrubber stopper, the headspace was replaced by N2, and theculture was incubated at 37°C on a rotary shaker. Theculture rapidly became anaerobic. The headspace was ana-lyzed for DMS after an incubation time of about 4 days.Under these conditions, E. coli HB101 quantitatively con-verted up to 10 mM DMSO to DMS. DMS was not detect-able in sterile controls or in cultures incubated with E. coliHB101 but without DMSO. E. coli HB101 was unable tometabolize DMS or DMSO2 under the assay conditionsused.

RESULTS

Metabolism of DMS in enrichment cultures. Anaerobicmedia supplemented with various concentrations of acetate(0 to 2 mM), H2S (1 to 2 mM), and DMS (0 to 10 mM) wereinoculated with a mixture of anaerobic phototrophic bacteriataken from Salt Pond and Great Sippewissett Marsh (31; see

above). During the first week, no disappearance of DMS wasobserved and growth of the cultures was slow, especially inmedia containing 5 and 10 mM DMS. During the secondweek, however, dense cultures of phototrophic bacteriadeveloped, and a significant consumption of DMS wasobserved (31). All cultures were distinctly purple in colorand exhibited absorption maxima at 590 nm, in the range of708 to 718 nm, and at 803 and 864 nm. In particular, theabsorption maxima at 590, 803, and 864 nm are characteristicof the phototrophic purple bacteria (bacteriochlorophyll a

[26]), and these peaks were especially dominant in culturessupplemented with acetate. After 1 month, 5-ml samples ofall cultures were transferred into fresh media (45 ml) toenrich for organisms able to rapidly metabolize DMS. Thistransfer procedure was repeated monthly, and growth andDMS turnover increased continuously. After three transfers,the apparent rate of DMS metabolism no longer increased,indicating that no further enrichment of DMS-convertingorganisms took place. The concentrations ofDMS recoveredin the fully enriched cultures after 14, 20, and 29 days are

listed in Table 1. Addition of 2 mM acetate to the basalmedium clearly enhanced DMS metabolism, whereas in-creasing the concentration of H2S from 1 to 2 mM onlypartially stimulated DMS turnover (only at low DMS con-

centrations). In sterile controls and cultures incubated in thedark, no DMS conversion was observed, suggesting thatDMS metabolism is a biological, light-dependent process.

VOL. 53, 1987

on Novem

ber 25, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

2028 ZEYER ET AL.

TABLE 1. Metabolism of DMS in enrichment cultures ofphototrophic purple bacteria

Supplement (mM) DMS (% of initial) recovered DMSO (mM)added to basal medium in cultures after given accumulatedincubation timea after 36-day

Acetate H2S DMS 14 days 20 days 29 days incubation"

0 1 2 36 12 <1 2.20 1 5 39 14 <1 5.10 1 10 54 31 21 9.82 1 2 <1 <1 <1 1.82 1 5 6 <1 <1 5.12 1 10 15 <1 <1 9.90 2 2 22 4 <1 2.10 2 5 34 10 <1 4.8O 2 10 61 52 54 8.2c2 2 2 <1 <1 <1 2.22 2 5 8 <1 <1 5.02 2 10 26 <1 <1 9.5

Dark controlsd >95 >95 >95 <0.1Sterile controlsd >95 >95 >95 <0.1

a DMS concentrations were determined in cultures of the fourth to ninthtransfer series. The concentrations indicated represent average values of atleast four independent measurements. The H2S concentrations were alsodetermined and found to be below the detection limit (0.05 mM) in all cultures.

I Cultures of the 10th transfer series were used for the DMSO determina-tion. Samples with known concentrations of DMSO were also analyzed andused to calculate the DMSO concentrations in the unknown samples.

c Gas chromatography performed before the DMSO assay showed that thisculture also contained some residual DMS.

d Medium: Basal medium plus 2 mM acetate, 2 mM H2S, and 5 mM DMS.

The dependence of DMS consumption on light intensity,acetate, and H2S concentrations was subsequently examinedin more detail, since these factors generally dominate growthand metabolism of the anaerobic phototrophic bacteria (26,28, 29).

a)-._

0

0-

2LightDark

Light dependence of DMS metabolism. Both growth of theenrichment cultures and elimination of DMS depended onlight intensity (Fig. 1). In the dark and in full daylight,growth and DMS turnover were significantly inhibited, dem-onstrating that phototrophic purple bacteria, which usuallyexhibit optimal growth under dim light conditions (26, 28),are involved in this metabolism. A protein concentration of60 to 90 ,ug ml-', which was even detected in culturesincubated in the dark or in full daylight, may have been dueto chemotrophic growth on acetate. A light intensity of 7 to12 microeinsteins m-2 s-1 was successfully used in thisstudy to select for DMS-converting phototrophic organisms.Figure 1 shows, however, that the enriched phototrophiccultures had their optimum growth at much greater lightintensities. Similarly, Veldhuis and van Gemerden (29) re-ported that low light intensities favored the development ofblooms of anaerobic phototrophic bacteria in an anaerobiclake, but that pure isolates from this habitat showed opti-mum growth at rather high light intensities.

Influence of H2S and acetate concentration on DMS turn-over rates. Cultures containing acetate grew well and metab-olized DMS rapidly. An increased concentration of H2S,however, had only a moderate effect on DMS metabolism(Table 1). To evaluate whether the observed variations inDMS consumption reflected only differences in the amountof biomass or individual effects of H2S or acetate, rateconstants of DMS turnover were determined. Cultures sup-plemented with 5 mM DMS and varying amounts of acetateand H2S were incubated, and concentrations of biomass (asprotein) and DMS were measured daily. After 7 days, theprotein content reached 60 to 90% of its final value, while theDMS concentration decreased following a logarithmic orlogistic curve. A pseudo-first-order rate constant (consider-ing DMS concentration versus time only) and a second-orderrate constant (also considering the biomass of the cultures)were calculated for each culture (Table 2). The pseudo-first-

4 10 20 40 100 200 400intensity [pE(m2.secy)- I

2

-o

o

a

C

DaylightFIG. 1. Light dependence of growth and DMS metabolism of enriched purple bacteria. Several flasks containing 4i ml of basal medium

plus 2 mM acetate, 2 mM H2S, and 5 mM DMS were inoculated with a 5-ml culture from the 10th transfer series grown on an identical medium.The cultures (protein concentration at time zero, <10 ,ug ml-') were incubated at various distances from a constant light source, in the dark,or in the daylight (24-h cycle). Symbols: 0, biomass expressed as protein after an incubation time of 4.5 days (relative standard deviation,<35% for all values); 0, DMS recovered (percent of initial) after 4.5 days; A, DMS recovered after 20 days. ,uE, Microeinsteins.

APPL. ENVIRON. MICROBIOL.

on Novem

ber 25, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

OXIDATION OF DMS TO DMSO 2029

TABLE 2. Rates of DMS turnover as a function of acetateand H2S concentration

Supplement (mM) Measurement and calculation of kineticadded to basal mediuma parameters after 7-day incubationb

Pseudo- Second-orderAcetate/ DMS Protein first-order rate constantAcetate H2S . recovered produced rate (day-', mgH25 ratio (% of initial) (p.g ml-') constant of protein'

(day-') ml-l)

0 1 0 80 32 0.11 3.40 2 0 77 34 0.13 3.70 4 0 80 102 0.11 1.12 4 0.5 80 163 0.10 0.61 1 1 59 41 0.22 5.42 2 1 55 94 0.24 2.52 1 2 43 77 0.32 4.24 2 2 58 141 0.23 1.64 1 4 26 111 0.67 6.1

a In addition to H2S and acetate, all media received an initial supplement of5 mM DMS. All media (45 ml) were inoculated with a 5-ml culture of the 10thtransfer series grown on basal medium plus 1 mM acetate, 2 mM H2S, and 5mM DMS.

b To calculate the rate constants of DMS degradation on day 7, the coursesof the protein increase and the DMS degradation within the first 8 days ofincubation were considered. The pseudo-first-order rate constant (k) wascalculated by the formula, k = -(In DMS2 - In DMS1)/(t2 - tl), where DMS,and DMS2 are the DMS concentrations in the culture at times t1 and t2. Thesecond-order rate constant was obtained by dividing k by the proteinconcentration.

order rate constants generally increased with increasingconcentration ratios of acetate/H2S which is in agreementwith the data presented in Table 1. The second-order rateconstants differed by an order of magnitude, suggesting thatthe DMS turnover rate depended on the composition of themedium rather than on the biomass concentrations. The data

120

100

E 80

A

60a

aa

40

0

20.

0 2 5 10DMS [mM I

FIG. 2. Biomass produced in enrichment cultures of pho-totrophic purple bacteria. The protein concentrations were deter-mined in samples from the fourth to ninth transfer series after thecultures reached the stationary growth phase. The concentrationsindicated represent average values of at least five independentmeasurements. The relative standard deviation was <25% for allvalues. Symbols: x, basal medium + 1 mM H2S; 0, basal medium+ 2 mM H2S; El, basal medium + 1 mM H2S + 2 mM acetate; A,basal medium + 2 mM H2S + 2 mM acetate. All media were

supplemented with DMS as indicated.

TABLE 3. Evolution of CH4 in enrichment cultures ofphototrophic purple bacteria

Supplement (mM) added to Evolution of CH4 (mM)" in culturesbasal mediuma after given incubation time

H2S DMS 14 days 20 days 29 days

1 0 <0.01 <0.01 <0.011 2 0.01 0.03 0.041 5 0.08 0.13 0.121 10 0.03 0.20 0.262 0 <0.01 <0.01 <0.012 2 <0.01 0.01 0.012 5 0.03 0.15 0.132 10 0.23 0.79 1.08

a No cultures supplemented with acetate showed CH4 evolution.I CH4 was determined in the headspace of the cultures (see Materials and

Methods), but the data received have been expressed as millimolar CH4 in theliquid phase. The data represent average values of at least four independentmeasurements in cultures of the sixth to ninth transfer series.

indicate that increasing concentrations of H2S in the mediumfavored growth but inhibited DMS metabolism. The second-order rate constants were in the range of 3.4 to 6.1 at 1 mMH2S, but only 1.6 to 3.7 at 2 mM H2S and 0.6 to 1.1 at 4 mMH2S. Acetate also favored growth but had no distinct effecton the second-order rate constant of the DMS consumption.

Yield of protein in enrichment cultures. Growth of allcultures listed in Table 1 and of controls incubated in theabsence of DMS was followed by periodically measuringprotein concentrations. Concentrations reached a plateauafter an incubation time of about 2 to 3 weeks. Figure 2demonstrates that the level of this plateau depended on notonly the supplements of H2S and acetate, but also theconcentration of DMS initially added. The average yields ofprotein per 1 mM substrate were calculated to be 18 ug ml-'for H2S, 22 ,ug ml-' for acetate, and 6 ,ug ml-' for DMS. Thedata suggested that DMS supported growth of thephototrophic purple bacteria; however, an identification ofthe transformation products of DMS is required to determinewhether DMS was used directly as an electron source orwhether it was first degraded to H2S by methanogenes (19)and thus supported growth indirectly.

Products of DMS metabolism. The enrichment cultureslisted in Table 1 were periodically assayed for possible DMStransformation products. CH3SH and DMDS were neverdetected in the culture headspace, and the H2S concentra-tion decreased rapidly after inoculation. Although thesethree compounds have been reported to be products ofmicrobial DMS metabolism (12, 19, 38), they are apparentlynot end products of DMS conversion mediated by theenrichment cultures. Methane, however, was detected incultures supplemented with H2S and DMS (Table 3). Thetotal amount of CH4 accumulated depended on the initialconcentration of DMS, and no CH4 was found in culturesincubated in the absence of DMS (Table 3) or in the presenceof acetate (data not shown). We therefore expected CH4 tobe a product of the metabolism of DMS. Experimentsconducted to clarify the role of CH4, however, subsequentlyrevealed that CH4 was not a direct metabolite of DMS. Thefollowing findings led to this conclusion. (i) DMS metabolismgenerally began in all media listed in Table 3 within a fewdays of inoculation, and >20% of DMS was consumed after10 days. CH4 production, however, only started after about10 days. (ii) Even in cultures that converted 10 mM DMS,the total amount of CH4 produced did not exceed 1.08 mM(Table 3). A stoichiometric liberation of CH4 is, therefore,

VOL. 53, 1987

on Novem

ber 25, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

H20/HDO A

J2.74ppm

x20

H20/HDO B

2.73ppm

Tspx20

H20/HDO c

2.72 ppm

TSPx20

6 5 4 3 2 1 0ppm

FIG. 3. 'H-NMR spectra of a DMS-consuming enrichment cul-ture of phototrophic purple bacteria. The procedures for the samplepreparation are given in Materials and Methods. (A) Concentratedculture filtrate supplemented with D20 and TSP. (B) Sample Asupplemented with an aliquot of a 1 M DMSO solution in distilledwater. (C) Sample B supplemented with an aliquot of a 1 M DMSO2solution in distilled water.

unlikely. (iii) Addition of 10 mM 2-bromoethanesulfonic acidsodium salt, a known inhibitor of methanogenesis (34), didnot affect DMS metabolism in any of the cultures butstopped all CH4 evolution. Such an uncoupling of DMSconsumption and CH4 formation suggests that CH4 is not adirect metabolite of DMS. (iv) Calculation of electron andprotein balances (see Discussion) suggest that two electronsare liberated per molecule of DMS and, consequently,DMSO, the oxidized species, rather than CH4, the reducedspecies, was likely to be the major metabolite of DMS.

Determination of DMSO in enrichment cultures. A 1H-NMR analysis of a DMS-degrading culture clearly demon-strated the presence of DMSO (Fig. 3). The culture sampleshowed a distinct signal at 2.74 ppm relative to TSP (Fig.3A). A sample supplemented with additional DMSO had asignal at the same position but with a higher intensity (Fig.3B). Addition ofDMS02 yielded a signal at 3.15 ppm relativeto TSP (Fig. 3C), and since this signal was not detected in theculture sample, DMS02 is unlikely to be a metabolite ofDMS. The chemical shifts ofDMSO and DMS02 determinedin distilled water-D20 were 2.71 and 3.13 ppm, respectively.

TABLE 4. Metabolism of DMS by pure strains ofphototrophic purple bacteriaa

Growth on basal medium

+ 1mM H2S + 2mMH2S + 2mMStrain acetate

DMS Protein DMS Proteinrecovered produced recovered produced

(% of initial) (pg ml-') (% of initial) (j,g ml-')

A 25 25 <1 110B 15 37 37 95C >95 34 >95 130a All media were supplemented with 5 mM DMS, and the incubation time

was 20 days.

The slight decrease in ionic strength upon adding smallaliquots of the DMSO and DMS02 solutions to the culturesample may explain the small differences in the chemicalshift of DMSO in the three different spectra.A quantitative biological assay for DMSO was used to

examine the enrichment cultures listed in Table 1 for theirDMSO concentration after an incubation time of 36 days.Essentially all of the DMS initially supplied to the cultureswas oxidized and could quantitatively (82 to 110%) berecovered as DMSO (Table 1). Cultures incubated in theabsence of DMS never contained any detectable levels ofDMSO. These data suggest that the enrichment culturesoxidize DMS to DMSO and that DMS is used as an electronsource.Metabolism ofDMS by pure cultures of phototrophic purple

bacteria. Three different strains of phototrophic purple bac-teria (strains A, B, and C; see Materials and Methods) wereisolated from the enrichment cultures, and their metabolicactivities towards DMS were determined (Table 4). Strain A,which was tentatively identified as a Thiocystis sp., grewwell on basal medium supplemented with 2 mM H2S and 2mM acetate and completely metabolized 5 mM DMS within20 days. On basal medium supplemented with only 1 mMH2S, however, growth of strain A was poor and DMSelimination was incomplete. Strain B transformed DMS onlypartially and strain C was inactive towards DMS on bothmedia.To evaluate whether DMS supported growth of strain A,

this organism was incubated in the presence of differentconcentrations of DMS and the biomass was determinedafter the stationary growth phase was reached (Table 5).More biomass was produced with increasing initial concen-tration of DMS. The average yield of protein per 1 mM DMSwas calculated to be about 5 p.g ml-', which is in goodagreement with the value previously determined in the

TABLE 5. Oxidation of DMS to DMSO and biomassproduction in cultures of strain Al

Analysis of culture after 20-day incubationInitial DMS DMS DMSO Proteinconcn (mM) recovered accumulated produced

(% of initial) (mM) (,ug ml1)

0 <0.1 951 <1 1.2 962 <1 2.2 1115 <1 5.3 117

10 <1 9.1 137a Strain A was incubated on basal medium plus 2 mM H2S, 2 mM acetate,

and DMS.

2030 ZEYER ET AL.

on Novem

ber 25, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

OXIDATION OF DMS TO DMSO 2031

TABLE 6. Electron balance of the growth ofphototrophic purple bacteria

Electrons Theoretical ObservedSubstrate Possible liberated biomass (mM biomass (mMproduct(s) (meq) CH20)a CH20)

H2S So 2 0.5 1.3S042- 8 2.0

Acetate 2CO2 8 2.0 1.4DMS H2S + 2CH4 (+4)C

CH3SH + CH4 (+ 2)cDMSO 2 0.5 0.4DMS02 4 1.0S + CO2 + CH4 6 1.5S042- + 2CO2 20 5.0

a Based on the assumption that a liberation of 4 meq of electrons results inthe reduction of 1 mM CO2 to 1 mM CH2O (28).

b The protein yields previously determined (Fig. 2 and Table 5) wereconverted to biomass (expressed as CH2O) with the assumption that theorganisms contain 50% protein.

c This reaction consumes rather than liberates electrons and is, therefore,unlikely to support growth of phototrophic purple bacteria.

enrichment cultures (Fig. 2). The DMS oxidized by strain Awas quantitatively recovered as DMSO (Table 5), whichsuggests that DMS supported growth of strain A by servingas an electron source for photosynthesis.

DISCUSSION

The basal medium used in this work was supplementedwith H2S, acetate, and DMS. Assuming that the photo-trophic purple bacteria can oxidize these substrates and usethe liberated electrons to reduce CO2 to biomass (expressedas CH2O [28]), an electron balance may be calculated (Table6). Phototrophic purple bacteria are known to oxidize H2S toS0 or even to S042- (22, 26, 28), thereby liberating 2 to 8electrons depending on the final oxidized species. Thisresults in a theoretical biomass production of 0.5 to 2.0 mMCH2O (28). The observed concentration of biomassamounted to 1.3 mM CH2O, which suggests that part of theH2S is oxidized to SO and part is oxidized to s042-. Foracetate, the theoretical and observed concentrations ofbiomass were also comparable (Table 6). We are aware thatpart of the acetate was probably converted directly tobiomass, although this would not affect the electron balancepresented in Table 6. Theoretically, DMS may be convertedto various metabolites, ranging from fully reduced species(H2S and CH4) to fully oxidized species (SO42- and CO2).Only the oxidation of DMS to DMSO, which liberates twoelectrons, however, is in agreement with the observedconcentration of biomass (Table 6). These calculations, inaddition to the analytical data, strongly suggest that DMSserved as an electron donor for the phototrophic purplebacteria.Enrichment cultures supplemented with H2S and high

concentrations of DMS only, but without acetate, exhibiteda low capacity for DMS metabolism (Table 1), a highevolution of CH4 (Table 3), and a low yield of biomass (Fig.2). The culture supplemented with 2 mM H2S and 10 mMDMS, in particular, almost completely terminated its metab-olism of DMS after 2 weeks and evolved up to 1 mM CH4.The yield of biomass in this culture was about 30% below theexpected value. Calculations of second-order rate constantsconfirmed that the H2S interfered with DMS turnover. Thismay be due to a competitive inhibition at an enzymatic levelor perhaps to an alteration in the mixed culture composition

(i.e., an H2S-induced selection of organisms unable to me-tabolize DMS). The species composition of anaerobicphototrophic cultures is known to be influenced by theconcentrations of H2S and acetate (22, 26, 29), and avariation in the absorption spectra of the cultures dependingon the medium composition was also detected in this study.Although the mechanism of the H2S interaction reported inthis paper is presently unknown, the observed effects (poorDMS conversion, evolution of CH4, and low yield ofbiomass) may well have a common explanation. Electronsliberated during the oxidation of H2S, DMS, or previouslyproduced biomass (<2 weeks of incubation) may no longerbe used to support light-dependent autotrophic growth but,rather, be consumed in the reduction of CO2 to CH4 and ofDMSO to DMS. The electron-accepting function of DMSOunder anaerobic conditions is well established (9, 35, 37),and such a continuous recycling of DMSO to DMS wouldexplain the low apparent net DMS elimination rate in certaincultures. We are currently investigating whether or not sucha sulfur cycle involving DMS and DMSO is active in some ofthe enrichment cultures.The number of anaerobic phototrophic bacteria in the

hypolimnion of Salt Pond is about an order of magnitudelower than in our enrichment cultures (31). In addition, onlya small fraction of the bacterial population in the pond maybe capable of oxidizing DMS. The high turnover ratesreported in this paper, however, suggest that even a limitednumber of organisms may easily be able to keep the DMSconcentration in the anaerobic hypolimnion of Salt Pondvery low by converting it to DMSO. It is noteworthy tomention that the DMSO concentration in the hypolimnion ofSalt Pond exceeds the DMS concentration (31). It remains tobe investigated whether or not the oxidation of DMS byanaerobic phototrophic bacteria is a major source of DMSOdetected in many aquatic habitats (1, 31).

ACKNOWLEDGMENTS

This study was initiated during a summer course ("Microbiology:Molecular Aspects of Cellular Diversity") at the Marine BiologicalLaboratory, Woods Hole, Mass. J.Z. and R.P.S. are grateful toR. S. Wolfe, E. P. Greenberg, and B. Schink, who guided usthroughout this stimulating course and with whom we had manyvaluable discussions. We thank D. Welti, Swiss Federal Institute ofTechnology, Zurich, for recording the 1H-NMR spectra and N.Pfennig, University of Constance, Constance, Federal Republic ofGermany, for assistance in identifying the phototrophic organisms.We are indebted to P. J. Colberg for reviewing the manuscript andto T. Walker for secretarial assistance.

This research was partially supported by National Science Foun-dation grant OCE 84-16203 to S.G.W.

LITERATURE CITED1. Andreae, M. 0. 1980. Dimethylsulfoxide in marine and fresh-

waters. Limnol. Oceanogr. 25:1054-1063.2. Andreae, M. O., and W. R. Barnard. 1983. Determination of

trace quantities of dimethyl sulfide in aqueous solution. Anal.Chem. 55:608-612.

3. Andreae, M. O., and W. R. Barnard. 1984. The marine chem-istry of dimethylsulfide. Mar. Chem. 14:267-279.

4. Andreae, M. O., and H. Raemdonck. 1983. Dimethyl sulphide inthe surface ocean and the marine atmosphere: a global view.Science 221:744-747.

5. Atkinson, R., J. N. Pitts, Jr., and S. M. Aschmann. 1984.Tropospheric reactions of dimethyl sulfide with NO3 and OHradicals. J. Phys. Chem. 88:1584-1587.

6. Barnard, W. R., M. 0. Andreae, W. E. Watkins, H. Bingemer,and H. -W. Georgii. 1982. The flux of dimethylsulfide from theoceans to the atmosphere. J. Geophys. Res. 87:8787-8793.

VOL. 53, 1987

on Novem

ber 25, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

7. Bechard, M. J., and W. R. Rayburn. 1979. Volatile organicsulfides from freshwater algae. J. Phycol. 15:379-383.

8. Bensadoun, A., and D. Weinstein. 1976. Assay of proteins in thepresence of interfering materials. Anal. Biochem. 70:241-250.

9. Bilous, P. T., and J. H. Weiner. 1985. Dimethyl sulfoxidereductase activity by anaerobically grown Escherichia coliHB101. J. Bacteriol. 162:1151-1155.

10. Cantoni, G. L., and D. G. Anderson. 1956. Enzymatic cleavageof dimethylpropiothetin by Polysiphonia lanosa. J. Biol. Chem.222:171-173.

11. Cline, J. D. 1969. Spectrophotometric determination of hydro-gen sulfide in natural waters. Limnol. Oceanogr. 14:454-458.

12. De Bont, J. A. M., J. P. van Dijken, and W. Harder. 1981.Dimethyl sulphoxide and dimethyl sulphide as a carbon, sulphurand energy source for growth of Hyphomicrobium S. J. Gen.Microbiol. 127:315-323.

13. Graedel, T. E. 1979. Reduced sulfur emission from the openoceans. Geophys. Res. Lett. 6:329-331.

14. Herbert, D., P. J. Phipps, and R. E. Strange. 1971. Determina-tion of protein with the Folin-Ciocalteu reagent, p. 249-252. InJ. R. Norris and D. W. Ribbons (ed.), Methods in microbiology,vol. SB. Academic Press, Inc., New York.

15. Howes, B. L., J. W. H. Dacey, and S. G. Wakeham. 1985. Effectsof sampling technique on measurements of porewater constitu-ents in salt marsh sediments. Limnol. Oceanogr. 30:221-227.

16. Ivanov, M. V., and J. R. Freney. 1983. The global biogeochemi-cal sulfur cycle. John Wiley & Sons, Inc., New York.

17. Kadota, H, and Y. Ishida. 1972. Production of volatile sulfurcompounds by microorganisms. Annu. Rev. Microbiol. 26:127-138.

18. Kanagawa, T., and D. P. Kelly. 1986. Breakdown of dimethylsulfide by mixed cultures and by Thiobacillus thioparus. FEMSMicrobiol. Lett. 34:13-19.

19. Kiene, R. P., R. S. Oremland, A. Catena, L. G. Miller, and D. G.Capone. 1986. Metabolism of reduced methylated sulfur com-pounds in anaerobic sediments and by a pure culture of anestuarine methanogen. Appl. Environ. Microbiol. 52:1037-1045.

20. Lewis, J. A., and G. C. Papavizas. 1970. Evolution of volatilesulfur-containing compounds from decomposition of crucifers insoil. Soil Biol. Biochem. 2:239-246.

21. Lovelock, J. E., and R. J. Maggs. 1972. Atmospheric dimethylsulphide and the natural sulphur cycle. Nature (London)237:452-453.

22. Pfennig, N. 1975. The phototrophic bacteria and their role in thesulfur cycle. Plant Soil 43:1-16.

23. Pfennig, N., and H. G. Truper. 1981. Isolation of members of thefamilies Chromatiacea and Chlorobiacea, p. 279-289. In M. P.Starr, H. Stolp, H. G. Truper, A. Balows, and H. G. Schlegel(ed.), The prokaryotes, a handbook on habitats, isolation andidentification of bacteria. Springer Verlag, New York.

24. Sivela, S., and V. Sundman. 1975. Demonstration of Thiobacil-lus-type bacteria which utilize methyl sulphides. Arch. Micro-biol. 103:303-304.

25. Steudler, P. A., and B. J. Peterson. 1984. Contribution of

gaseous sulphur from salt marshes to the global sulphur cycle.Nature (London) 311:455-457.

26. Truper, H. G., and N. Pfennig. 1981. Characterization andidentification of the anoxygenic phototrophic bacteria, p.299-312. In M. P. Starr, H. Stolp, H. G. Truper, A. Balows, andH. G. Schlegel (ed.), The prokaryotes, a handbook on habitats,isolation and identification of bacteria. Springer Verlag, NewYork.

27. Vairavamurthy, A., M. 0. Andreae, and R. L. Iverson. 1985.Biosynthesis of dimethylsulfide and dimethylpropiothetin byHymenomonas carterae in relation to sulfur source and salinityvariations. Limnol. Oceanogr. 30:59-70.

28. van Gemerden, H., E. Montesinos, J. Mas, and R. Guerrero.1985. Diel cycle of metabolism of phototrophic purple sulfurbacteria in lake Ciso (Spain). Limnol. Oceanogr. 30:932-943.

29. Veldhuis, M. J. W., and H. van Gemerden. 1986. Competitionbetween purple and brown phototrophic bacteria in stratifiedlakes: sulfide, acetate, and light as limiting factors. FEMSMicrobiol. Ecol. 38:31-38.

30. Wakeham, S. G., B. L. Howes, and J. W. H. Dacey. 1984.Dimethyl sulphide in a stratified coastal salt pond. Nature(London) 310:770-772.

31. Wakeham, S. G., G. L. Howes, J. W. H. Dacey, R. P.Schwarzenbach, and J. Zeyer. 1987. Biogeochemistry ofdimethylsulfide in a seasonally stratified coastal salt pond.Geochim. Cosmochim. Acta 51:1675-1684.

32. Widdel, F., G.-W. Kohring, and F. Mayer. 1983. Studies ondissimilatory sulfate-reducing bacteria that decompose fattyacids. III. Characterization of the filamentous gliding De-sulfonema limicola gen. nov. sp. nov., and Desulfonema mag-num sp. nov. Arch. Microbiol. 134:286-294.

33. Widdel, F., and N. Pfennig. 1981. Studies on dissimilatorysulfate-reducing bacteria that decompose fatty acids. I. Isola-tion of new sulfate-reducing bacteria enriched with acetate fromsaline environments. Description of Desulfobacter postgateigen. nov., sp. nov. Arch. Microbiol. 129:395-400.

34. Wolfe, R. S., and I. J. Higgins. 1979. Microbial biochemistry ofmethane-a study in contrasts, p. 267-353. In J. R. Quale (ed.),Microbial biochemistry, vol. 21. University Park Press, Balti-more.

35. Yen, H.-C., and B. Marrs. 1977. Growth ofRhodopseudomonascapsulata under anaerobic dark conditions with dimethylsulfoxide. Arch. Biochem. Biophys. 181:411-418.

36. Zehnder, A. J. B., and S. H. Zinder. 1980. The sulfur cycle, p.105-145. In 0. Hutzinger (ed.), The handbook of environmentalchemistry, vol. 1, part A. Springer Verlag, Berlin.

37. Zinder, S. H., and T. D. Brock. 1978. Dimethyl sulphoxidereduction by micro-organisms. J. Gen. Microbiol. 105:335-342.

38. Zinder, S. H., and T. D. Brock. 1978. Production of methane andcarbon dioxide from methane thiol and dimethyl sulphide byanaerobic lake sediments. Nature (London) 273:226-228.

39. Zinder, S. H., W. N. Doemel, and T. D. Brock. 1977. Productionof volatile sulfur compounds during the decomposition of algalmats. Appl. Environ. Microbiol. 34:859-860.

2032 ZEYER ET AL.

on Novem

ber 25, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from