Polysulfides as Intermediates in the Oxidation of Sulfide to Sulfate byBeggiatoa spp.

Jasmine S. Berg,a Anne Schwedt,b* Anne-Christin Kreutzmann,b Marcel M. M. Kuypers,a Jana Miluckaa

‹Department of Biogeochemistry, Max Planck Institute for Marine Microbiology, Bremen, Germanya; Department of Microbiology, Max Planck Institute for MarineMicrobiology, Bremen, Germanyb

Zero-valent sulfur is a key intermediate in the microbial oxidation of sulfide to sulfate. Many sulfide-oxidizing bacteria produceand store large amounts of sulfur intra- or extracellularly. It is still not understood how the stored sulfur is metabolized, as themost stable form of S0 under standard biological conditions, orthorhombic �-sulfur, is most likely inaccessible to bacterial en-zymes. Here we analyzed the speciation of sulfur in single cells of living sulfide-oxidizing bacteria via Raman spectroscopy. Ourresults showed that under various ecological and physiological conditions, all three investigated Beggiatoa strains stored sulfuras a combination of cyclooctasulfur (S8) and inorganic polysulfides (Sn

2�). Linear sulfur chains were detected during both theoxidation and reduction of stored sulfur, suggesting that Sn

2� species represent a universal pool of bioavailable sulfur. Forma-tion of polysulfides due to the cleavage of sulfur rings could occur biologically by thiol-containing enzymes or chemically by thestrong nucleophile HS� as Beggiatoa migrates vertically between oxic and sulfidic zones in the environment. Most Beggiatoaspp. thus far studied can oxidize sulfur further to sulfate. Our results suggest that the ratio of produced sulfur and sulfate variesdepending on the sulfide flux. Almost all of the sulfide was oxidized directly to sulfate under low-sulfide-flux conditions, whereasonly 50% was oxidized to sulfate under high-sulfide-flux conditions leading to S0 deposition. With Raman spectroscopy we couldshow that sulfate accumulated in Beggiatoa filaments, reaching intracellular concentrations of 0.72 to 1.73 M.

Sulfur storage in bacteria was first documented in the 19th cen-tury when Winogradsky (1) proposed the concept of chemo-

lithotrophy from observations of sulfur globules in Beggiatoa fil-aments. Since then, taxonomically diverse microorganisms havebeen reported to store zero-valent sulfur either intracellularly(e.g., Beggiatoaceae and Chromatiaceae) or extracellularly (e.g., Ec-tothiorhodospiraceae, Chlorobiaceae, and Thiobacillus) as an inter-mediate in the oxidation of reduced sulfur compounds (2). Inter-nal sulfur, which is contained within invaginations of theperiplasm, is often surrounded by a protein envelope thought tobe purely structural in function (3, 4). External sulfur globules arenot enclosed in a membrane; rather, they feature a hydrophilicsurface of, e.g., polythionates (5, 6). It has been noted that biolog-ical sulfur is characterized by a lower density and different crystal-line structure than orthorhombic sulfur (7, 8). However, our un-derstanding of the chemical nature of stored sulfur is still limitedby conventional methods of extraction and analysis that can in-duce artificial changes in the chemistry of biogenic sulfur com-pounds. In transmission electron micrographs, for example, thepresence of sulfur in cells must be inferred from empty vesiclesresulting from the dissolution of sulfur in ethanol during the de-hydration step in resin embedding (see, e.g., reference 9). Onlyrecently has the novel application of spectroscopic methods tobiological samples indicated that the chemical species of storedsulfur may differ across groups of ecologically and physiologicallydistinct bacteria (see, e.g., references 6, 10, and 11). This has im-plications for the metabolism of stored sulfur, as enzyme-S0 inter-actions may be highly specific for the sulfur species utilized.

Sulfur atoms readily catenate to form linear or cyclic moleculesand can form bonds with both organic and inorganic end groups.The resulting sulfur compounds are often extremely pH and redoxsensitive and thus difficult to measure. The most thermodynami-cally stable form of elemental sulfur under standard biologicalconditions is orthorhombic �-sulfur, which consists of puckered

S8 rings (12), but more than 180 allotropes have been described(13). Thus far, a variety of species other than cyclooctasulfur havebeen identified in sulfur-storing bacteria, including inorganicpolysulfides (Sn

2�), polythionates (�O3SOSnOSO3�), and long-

chain organosulfanes (ROSnOR) (6, 11, 14). The solubilities ofthese S0 species in water are very different, ranging from the nearlyinsoluble S8 rings to extremely soluble inorganic polysulfides. Sul-fur-oxidizing and sulfur-reducing bacteria growing on solid sulfurpreferentially take up the charged and soluble fraction, leavingbehind the insoluble rings (15–17). In Beggiatoa spp., oxidation ofelemental sulfur to sulfate is mediated by the cytoplasmic enzymereverse dissimilatory sulfite reductase (rDSR) (18). It is not yetunderstood how extracellularly and periplasmically stored S8

rings are accessed by the Dsr system, as this implies transmem-brane transport. It has been hypothesized that cyclooctasulfur iseither reduced to H2S or taken up via an unknown persulfide(ROSOSH) carrier molecule across the cytoplasmic membrane(19). Furthermore, enzyme active sites probably do not accom-modate large molecules of cyclooctasulfur, and recent studies in-dicate that at least the archaeal sulfur oxygenase reductase fromAcidianus ambivalens contains an elongated active-site pocketwhich can accommodate only linear polysulfides of up to 8 atoms

Received 23 August 2013 Accepted 4 November 2013

Published ahead of print 8 November 2013

Address correspondence to Jasmine S. Berg, jberg@mpi-bremen.de.

* Present address: Anne Schwedt, Department of Biogeochemistry, Max PlanckInstitute for Marine Microbiology, Bremen, Germany.

Supplemental material for this article may be found at http://dx.doi.org/10.1128/AEM.02852-13.

Copyright © 2014, American Society for Microbiology. All Rights Reserved.

doi:10.1128/AEM.02852-13

January 2014 Volume 80 Number 2 Applied and Environmental Microbiology p. 629 – 636 aem.asm.org 629

on April 27, 2019 by guest

http://aem.asm

.org/D

ownloaded from

in length (20). The enzymes involved in sulfur utilization in bac-teria are still poorly characterized, and therefore mechanisms ofsulfur activation and transport remain speculative.

Published studies on the chemistry of stored sulfur frequentlyreport contradictory findings: the same organism appears to con-tain different sulfur species depending on the culture conditions,sample preparation, and analytical techniques used. Moreover,many of these investigations were performed on dead bacteria atthe risk of altering the sulfur chemistry. We tried to resolve someof the conflicting results while investigating the link between sul-fur species and enzymatic pathways of sulfur utilization by com-paring three strains of Beggiatoa under various growth conditions.In this study, sulfur compounds in single cells of living bacteriawere characterized, and their distribution was mapped using con-focal Raman microscopy, a technique which relies on the inelasticscattering of light to provide a unique molecular-structural fin-gerprint of a compound, allowing for its identification. The non-invasive, nondestructive nature of Raman spectroscopy makes itideal for biological applications and especially for measuring pH-and redox-sensitive sulfur species.

MATERIALS AND METHODSCulturing. Beggiatoa sp. strain 35Flor was cultivated in agar-stabilizedsulfide gradient medium prepared in glass test tubes (1.4 by 14.5 cm) asdescribed by Schwedt et al. (21). Each tube contained an �2-cm-highbottom layer of solid agar (1.5% [wt/vol] agar) and an �5-cm-high toplayer of semisolid agar (0.25% [wt/vol] agar), which allowed Beggiatoa toglide through the medium. Agar layers were prepared by mixing sepa-rately autoclaved sulfate-free artificial seawater (24.34 g NaCl, 8.34 gMgCl2 · 6H2O, 0.66 g CaCl2 · 2H2O, and 1.02 g KCl per 1 liter Milli-Qwater) and triple-washed agar solutions. The bottom agar was amendedwith a sterile 1 M sulfide solution to a final concentration of 4, 12, or 16mM, referred to as low-, medium-, and high-sulfide-flux conditions, re-spectively. The top agar received additional trace elements, vitamins, min-erals (555 mg K2HPO4, 28.72 mg Na2MoO4, 750 mg Na2S2O5, and 29 mgFeCl3 · 6H2O per liter), and HCO3

� (2 mM final concentration) as spec-ified by Schwedt et al. (21). Culture tubes were loosely capped, allowinggas exchange with the atmosphere. Oxygen gradients were allowed todevelop overnight before inoculation with 100 �l filament suspensionapproximately 1 cm below the agar surface. Beggiatoa alba strains B15LDand B18LD were grown in freshwater-based agar-stabilized gradient me-dium (22) with a bottom agar sulfide concentration of 8 mM.

For Raman spectroscopy, 10 to 50 �l of each live bacterial culture waspipetted onto a 20- by 20- by 1-mm glass coverslip coated with 0.1%poly-L-lysine solution (Sigma-Aldrich). Samples were protected with asecond coverslip and then sealed with electrical tape to minimize oxygendiffusion and prevent drying. At least 15 different filaments from eachtime point and culture condition were analyzed, and representative pointspectra are shown in Results.

Reference compounds. The following analytical grade (purity,�99%) compounds were used as references: elemental sulfur (RothGmbH), sodium sulfate (Sigma-Aldrich), potassium sulfate (Merck),magnesium sulfate (Merck), and calcium sulfate (Applichem). A polysul-fide solution was prepared from 5.06 g Na2S · 9H2O and 5.8 g elementalsulfur per 100 ml H2O, with a final pH of 9.5 and sulfide concentration of210 mM. Spectra for aqueous polysulfides were recorded under anoxic tooxic conditions by transferring a 100-�l drop to a glass coverslip andimmediately measuring in live mode, with a point spectrum taken everysecond. A subsample of the polysulfide solution was dried under a 90:10N2-CO2 atmosphere for comparison. Crystalline standards, such as so-dium sulfate, were measured directly on glass coverslips.

Raman spectroscopy. All measurements were carried out with anNTEGRA Spectra confocal spectrometer (NT-MDT, Eindhoven, Nether-lands) coupled to an inverted Olympus IX71 microscope. The excitation

light from a 532-nm solid-state laser was focused on the sample throughan Olympus 100� (numerical aperture [NA], 1.3) oil immersion objec-tive. The pinhole aperture was maintained at 55 �m, corresponding to aspatial resolution of 250 to 300 �m. Raman scattered light was dispersedwith a 150-line · mm�1 grating and collected by an electron-multiplyingcharge-coupled device (EMCCD) camera (Andor Technology, Belfast,Northern Ireland) cooled to �70°C. To verify resonance Raman spectra,a 785 nm diode laser coupled to an IR-CCD camera (Andor Technology,Belfast, Northern Ireland) was used, and the pinhole aperture was set at 80�m. Exposure times varied between 0.2 and 6 s, and the Raman spectrawere recorded between 0 and �4,500 cm�1 with a spectral resolution of0.2 cm�1. The laser power at the sample was checked with a laser powermeter and did not exceed 1 mW. The instrument was also equipped witha motorized piezostage for x-y scanning. Full-spectrum maps were takenwith a maximum 1.2-s dwelling time per point and 0.3-mW laser power.

Data processing. The software Nova_Px 3.1.0.0 (NT-MDT, Eind-hoven, Netherlands) was used for all spectral analyses. Spectra were nor-malized to an arbitrary unit of intensity on the y axis, maintaining relativepeak heights, for easier comparison. Noisy spectra were smoothed using aGaussian function with a tau value of 3. In order to map specificcompounds, relevant peak areas were first background corrected, andimages were then overlaid and brightness/contrast adjusted in AdobePhotoshop CS2.

For the Raman quantification of sulfate, spectra of 8 different Beggia-toa sp. 35Flor filaments grown on sulfate-free medium were baseline cor-rected using a modified average fit curve with a period of 20 (an exampleis provided in Fig. S1A in the supplemental material). Although peak areais generally used for Raman quantification, peak height can also be used asan accurate measure of intensity (see, e.g., reference 23). In our study,standard curves were generated from the height of the SAO peak at 997cm�1 in spectra from a 0.2 to 2.0 M dilution series of Na2SO4 in Milli-Qwater and resulted in a good fit with an r2 value of �0.9309 for the threedifferent measurement conditions used (see Fig. S1B in the supplementalmaterial). The detection limit for sulfate under the applied conditions onthe instrument used was �25 mM.

Chromatographic sulfate measurements. Sulfate was analyzed on a761 Compact ion chromatograph (Metrohm) with a Metrosep A SUPP 5column. A carbonate buffer, prepared with 3.2 mmol liter�1 Na2CO3 and1 mmol liter�1 Na2CO3, was used as the eluent. The duration of a run was14 min, with sulfate eluting at �11.5 min. Sodium sulfate standards of 1 to400 �M and IAPSO standard seawater were used as references.

The intracellular sulfate content of Beggiatoa sp. 35Flor was deter-mined by picking individual filaments with a glass needle and transferringthem to 3 ml of 10 mM HCl. Triplicates of samples containing 25, 50, or100 filaments were freeze-thawed twice to ensure cell lysis. The resultingsolution was passed through a 0.45-�m syringe filter into ion chromatog-raphy (IC) vials before analysis. Because some liquid adhering to filamentsis inevitably transferred with each sample, single liquid droplets from theinoculated medium were also measured in 10 mM HCl. Sulfate concen-trations in these control samples were so close to the detection limits ofour instrument that single-droplet effects were considered negligible here.

Extracellular sulfate concentrations were determined by extractingpore water from the agar medium with 2-cm-long microrhizones (Rhizo-sphere Research Products, Wageningen, Netherlands) inserted verticallyinto the culture tubes. It is possible that differential pressures occurredalong the rhizone, resulting in unequal sampling of the upper 2 cm ofmedium. Parallel cultures with different bottom agar sulfide concentra-tions (4 mM, 12 mM, or 16 mM) were inoculated at the same time andsampled over a period of 14 days. Uninoculated cultures containing 4 mMsulfide were monitored for the chemical formation of sulfate. For this, onetube from each culture condition was sacrificed at every time point. Allliquid samples were diluted in 10 mM HCl and filtered through a 0.2-�mfilter before analysis.

For sulfate depth profiles, gradient media with a sulfide concentrationin the bottom agar of 4 mM were prepared in triplicate in open-bottom

Berg et al.

630 aem.asm.org Applied and Environmental Microbiology

on April 27, 2019 by guest

http://aem.asm

.org/D

ownloaded from

glass tubes (1.4 by 14.5 cm) plugged from below with a tapered rubberstopper. After approximately 2 weeks of incubation, the lower end of eachtube was immersed in liquid nitrogen up to a depth of 2 cm, leaving thesemisolid top agar nonfrozen, and the rubber stopper was removed. Theagar core was then mounted on a micromanipulator, which allowed forsectioning of the top agar in 2-mm steps. Each section of semisolid agarwas gently passed through a 0.45-�m syringe filter and diluted in 3 ml of10 mM HCl for analysis by ion chromatography. Sulfate production wascalculated from the change in the measured sulfate concentrations overtime and adjusted to the 2-mm-high Beggiatoa mat.

Microsensor measurements. Gradients of pH and H2S were recordedin triplicate, and the culture tubes were subsequently sectioned for sulfateprofiling as described above. The profiles were measured in 250-�m in-tervals from the agar surface to a depth of 1 cm. Microsensors for pH(pH-10) and H2S (H2S-10) were purchased from Unisense A/S (Aarhus,Denmark) and calibrated as described by Schwedt et al. (21). Total sulfideconcentrations were calculated as described by Kühl et al. (24) using a pk1

value of 6.569. Sulfide fluxes were calculated using Fick’s first law of dif-fusion with a diffusion coefficient D of 1.52 � 10�9 m2 s�1 for H2S/HS�

at 20°C (25). All microsensors were calibrated immediately before use andthen after the final measurement to check for a possible drift.

RESULTSSulfur species during sulfur oxidation and reduction in a ma-rine Beggiatoa strain. Beggiatoa was grown in sulfide gradienttubes simulating the environmental conditions in which Beggiatoacan migrate vertically between oxic and anoxic sediment zones. Ingradient medium with a low sulfide flux, Beggiatoa sp. 35Flor fil-aments were present in a dense mat ca. 0.5 cm below the air-agarinterface at 16 days after inoculation. Within this mat, the Beggia-toa completely oxidized upward-diffusing sulfide. A peak in thesulfate concentration profile indicated that sulfate was formed asan end product (Fig. 1). The Beggiatoa mat was whitish in colordue to intracellular deposition of sulfur globules (Fig. 1 and 2A),as demonstrated by Schwedt et al. using high-performance liquidchromatography of methanol extracts from filaments (21). UsingRaman spectroscopy, we could confirm presence of zero-valentsulfur in these inclusions of living Beggiatoa. In Fig. 2B, Ramanspectra of these sulfur inclusions are shown together with spectraof two sulfur standards for comparison. The spectrum of the cy-clooctasulfur standard was characterized by two strong peaks at

152 and 216 cm�1 corresponding to the bending and stretchingmodes of the 8-fold ring and a third strong peak at 473 cm�1

corresponding to the vibration of the SOS bond (26). It is impor-tant to note that not only the absolute positions of peaks but alsotheir relative heights are significant in a Raman spectrum and thatthe three peaks of S8 are similar in height. The spectrum of aglobule from 5-day-old culture matched that of S8, but spectrafrom 8- and 16-day cultures exhibited a weakening of the 152- and216-cm�1 peaks relative to the 473-cm�1 SOS peak, a ratio whichis typical of aqueous polysulfides as visible in the Sn

2� standardspectrum (see also references 27 and 28). Thus, globules in fila-ments from a later growth phase were composed of a mixture ofpolysulfides and cyclooctasulfur, with the ratio of Sn

2� to S8 ap-parently increasing over time.

The distribution of sulfur within the cells was mapped usingthe characteristic SOS vibration peak at 473 cm�1 (Fig. 2C). Apeak at 757 cm�1, which is depicted in the uppermost spectrum inFig. 2B, was tentatively attributed to cytochrome c and used in thespectral map as a proxy for the Beggiatoa cell membrane. The757-cm�1 peak cooccurred with peaks at 979, 1,136, 1,321, and1,584 cm�1 which are the characteristic resonance Raman bandsfor cytochrome c at an excitation wavelength of 532 nm (29). Thephenomenon of resonance Raman scattering occurs when the ex-citation frequency is close to an electronic transition of a moleculeand thus is specific to the wavelength of the laser used. The factthat the peaks at 757, 979, 1,136, 1,321, and 1,584 cm�1 were notobserved under excitation with a 785-nm diode laser (see Fig. S2 inthe supplemental material) provides further support for the as-signment of these peaks to cytochrome c.

In cultures grown with a high sulfide flux, filaments at the

FIG 1 Example of H2S (Œ), total sulfide (�), and SO42� (�) profiles in a

Beggiatoa sp. 35Flor culture after 16 days of incubation with a low sulfide flux.The Beggiatoa mat was located at a depth of 0.5 to 0.7 cm.

FIG 2 (A) Optical micrograph of a Beggiatoa sp. 35Flor filament with a lowsulfide flux. (B) Point spectra of sulfur inclusions measured 5, 8, and 16 daysafter inoculation using constant 180-�W laser power and 1-s exposure time,compared to sulfur standards (Std). A spectrum of cytochrome c (cyt c)iden-tified in the Beggiatoa membrane region is also shown. (C) Raman map of afilament, showing the 473-cm�1 sulfur peak in yellow and autofluorescence ofthe cell in green.

Polysulfide Intermediates in Bacterial S0 Metabolism

January 2014 Volume 80 Number 2 aem.asm.org 631

on April 27, 2019 by guest

http://aem.asm

.org/D

ownloaded from

sulfide-oxygen interface accumulated so much sulfur (Fig. 3A andB) that some eventually burst. Presumably to avoid this fate, asubpopulation of filaments migrated downwards and establisheda second bacterial mat in the anoxic, sulfidic zone. In this subpop-ulation, a depletion of internal sulfur was observed coincidingwith the production of sulfide (21). It has been hypothesized thatthe anoxic subpopulation of Beggiatoa sp. 35Flor reduces intracel-lular sulfur with stored polyhydroxyalkanoates (PHAs), whichwere identified using the nonspecific lipid stain Nile Red (21).Although PHA is a strong Raman scatterer, we were not able todetect this particular carbon storage compound in Beggiatoa sp.35Flor filaments under any of the investigated growth conditions.The remaining sulfur globules in filaments of this lower Beggiatoamat were composed of both S8 rings and linear Sn

2� species (Fig.3C), as indicated by the dominant SOS peak at 473 cm�1. Addi-tionally, two weak S8 ring vibration peaks and a 272-cm�1 peakoccurred in the SOSOS bending region. The 272-cm�1 peak ischaracteristic of long-chain (n � 8) polymeric sulfur (30).

Sulfate, a strong Raman scatterer with a signature peak belong-ing to the SAO stretch at 997 cm�1 (Fig. 4A), was also identifiedin Beggiatoa sp. 35Flor filaments but not in the surrounding me-

dium. Porewater sulfate concentrations as measured by ion chro-matography were below the �25 mM detection limit of the Ra-man microscope. In Beggiatoa filaments, sulfate appeared to beconcentrated in particular regions of the cell, in some cases inassociation with the sulfur globules (Fig. 4B). The intracellularsulfate was further quantified by Raman spectroscopy using theheight of the SAO peak as an indicator of sulfate concentration(see Fig. S1 in the supplemental material). Assuming that sulfatealone contributed to the 997-cm�1 peak, the estimated concentra-tions of intracellular sulfate were between 0.72 and 1.73 M, withan average of 1.14 M. Ion chromatography of individual, hand-picked filaments confirmed the presence of molar concentrationsof sulfate in this marine Beggiatoa strain. We then monitored theproduction of sulfate from biological sulfide oxidation in the porewater of cultures grown in medium prepared with sulfate-freeseawater. The pore water sulfate concentrations increased linearlyduring the first 2 weeks of incubation in all cultures (Fig. 5). Thesulfate production increased proportionally to the sulfide flux,resulting in concentrations of 3.2, 5.1, and 9.9 mM in the 2-cm-thick layer the around the Beggiatoa mat in 14-day-old cultureswith initial bottom agar sulfide concentrations of 4, 12, and 16mM, respectively. No abiotic sulfate production was detected inthe uninoculated controls (Fig. 5).

In order to relate the observed sulfate production to sulfide

FIG 3 (A) Optical micrograph of Beggiatoa sp. 35Flor from the upper mat of a culture grown with a high sulfide flux. (B) Corresponding Raman map showingthe occurrence of the 473-cm�1 sulfur peak in yellow/red and autofluorescence of the cell in green. (C) Representative point spectra of sulfur inclusions fromupper and lower mat filaments together with polysulfide and cyclooctasulfur standards (Std).

FIG 4 (A) From top to bottom: (i) a background-corrected spectrum of liveBeggiatoa sp. 35Flor grown on sulfate-free medium exhibiting a sulfate peak at997 cm�1 and (ii) sulfate detected in a dried filament in comparison to (iii) anNa2SO4 standard. (B) Raman map of a dried Beggiatoa sp. 35Flor filament,with the 473-cm�1 SOS peak shown in yellow/red, the 997-cm�1 SAO peakof sulfate in blue, and autofluorescence of the cell in green.

FIG 5 Porewater sulfate concentrations in the upper 2 cm of agar mediumfrom cultures grown with different sulfide concentrations in the bottom agar:uninoculated control (�), 4 mM (�), 12 mM (Œ), and 16 mM (Œ).

Berg et al.

632 aem.asm.org Applied and Environmental Microbiology

on April 27, 2019 by guest

http://aem.asm

.org/D

ownloaded from

consumption, we estimated sulfide consumption rates by inte-grating the sulfide fluxes in Beggiatoa sp. 35Flor cultures measuredby Schwedt et al. (21) over the height of the Beggiatoa mat. Theresulting sulfide oxidation rates (� the standard deviation) forcultures with 16 mM sulfide were 13.65 � 2.5 and 8.55 � 1.8 mol ·m�3 · day�1, after 7 and 13 days, respectively (Table 1). These rateswere 30 to 40% higher than the corresponding sulfate productionrates of 7.9 and 5.8 mol · m�3 · day�1. The sulfide flux in cultureswith 4 mM sulfide did not fluctuate significantly over the 13-dayperiod, which can be explained by a continuous downward migra-tion of the Beggiatoa mat, adjusting the oxygen-sulfide interface tothe decreasing flux. The average sulfide oxidation rate was 2.88 �0.6 mol · m�3 · day�1, and the sulfate production rate was 2.14 mol· m�3 · day�1, showing that at a low sulfide flux, sulfide may bedirectly oxidized to sulfate.

Although the sulfate production rate remained relatively con-stant within the first 2 weeks after inoculation, sulfide consump-tion decreased after this period, implying that the ratio of H2S andS0 oxidation rates varied over time. To better compare the ratio ofsulfide oxidation products, sulfate and sulfide fluxes were mod-eled from the respective chemical profiles (Fig. 1) measured inlow-sulfide-flux cultures after 16 days of incubation. The calcu-lated average sulfate production rate was 2.84 � 0.02 mol · m�3 ·

day�1, more than four times higher than the sulfide consumptionrate of 0.8 � 0.25 mol · m�3 · day�1. Visual inspection of thefilaments at this time point confirmed that most of refractive sul-fur globules had disappeared.

Sulfur species in freshwater Beggiatoa strains oxidizing sul-fide incompletely to sulfur. Beggiatoa alba B18LD is a hetero-trophic, freshwater strain capable of oxidizing sulfide to elementalsulfur but not to sulfate (31). Instead, the strain can reduce inter-nally stored sulfur to sulfide, presumably by degrading carbon-rich compounds such as PHAs during periods of anoxia (31). Incontrast to the case for the autotrophic Beggiatoa sp. 35Flor, poly-hydroxybutyrate (PHB) inclusions (Fig. 6A) were easily identifiedin B. alba B18LD based on their characteristic Raman bands at906, 1,357, 1,431, 1,733 and �2,940 cm�1 (32). After 4 days ofgrowth in gradient cultures with a medium sulfide flux, highlyrefractive intracellular inclusions were visible under the opticalmicroscope (Fig. 6B). These were identified as sulfur globulescomposed of a mixture of S8 rings and polysulfides by their char-acteristic Raman spectra (Fig. 6A). The wave numbers of the SOSvibration varied in different point spectra from 466 to 486 cm�1

(data not shown). Because the vibration frequency of the SOSbond depends on the number of sulfur atoms in a chain (28), theshift in peak position may reflect Sn

2� species of different chainlengths (n) coexisting in a polysulfide “pool” within the sulfurglobules. The characteristic sharp SAO peak of sulfate at 997cm�1 was not detected in this freshwater Beggiatoa strain.

Beggiatoa alba B15LD appears to be a special case in the familyBeggiatoaceae, members of which have thus far all been reported todeposit sulfur intracellularly (33). Surprisingly, no sulfur could bedetected inside B. alba B15LD filaments, but instead abundant,refractive spherules, composed of a mixture of S8 rings and poly-sulfides (Fig. 6A), were found in the medium surrounding the cells(Fig. 6C). Previously, this strain was observed to deposit intracel-lular sulfur (34), and therefore it is possible that a mutation caus-ing the sulfur globules to be formed extracellularly occurred. Inuninoculated controls the abiotically formed sulfur precipitatedas crystals rather than globules, suggesting that the extracellularsulfur globules associated with the Beggiatoa alba filaments might

TABLE 1 Sulfide consumption and sulfate production rates for marineBeggiatoa sp. 35Flor cultures

Sulfide in bottomagar (mM)

Time(days)

Avg rate (mol · m�3 · day�1) �SD

Sulfideconsumption

Sulfateproduction

16 (high flux) 7 13.65 � 2.5a 7.913 8.55 � 1.8a 5.8

4 (low flux) 7 2.88 � 0.6a 2.1413 2.88 � 0.6a 2.1416 0.8 � 0.25 2.84 � 0.02

a Value calculated from fluxes measured by Schwedt et al. (21).

FIG 6 (A) From top to bottom: (i) representative point spectra of sulfur inclusions from B. alba B18LD, (ii) extracellular sulfur from B. alba B15LD, (iii) apolysulfide standard, (iv) cytochrome c, and (v) PHB. (B and C) Optical micrographs of strain B18LD (B) and strain B15LD (C) grown with 8 mM sulfide in thebottom agar. (D and E) Raman map of a B15LD filament, with the 2,900-cm�1 COH peak in pink and the 757-cm�1 cytochrome c resonance peak in green (D),and the corresponding optical image (E). The red square outlines the Raman map.

Polysulfide Intermediates in Bacterial S0 Metabolism

January 2014 Volume 80 Number 2 aem.asm.org 633

on April 27, 2019 by guest

http://aem.asm

.org/D

ownloaded from

be of biological origin. Indeed, these sulfur globules behaved dif-ferently from chemically produced crystalline sulfur; for example,exposure to strong laser radiation caused burning of the sulfurglobules, suggesting that globules were more amorphous in struc-ture. Full-spectrum Raman mapping of a B15LD filament after 4days growth in sulfide gradient medium revealed PHB stored asinclusions within the cells and c-type cytochromes associated withthe cell membrane (Fig. 6D and E). After 11 days, PHB could nolonger be detected in Beggiatoa alba B15LD filaments.

DISCUSSION

All of the Beggiatoa strains that we studied deposited sulfur glob-ules as a mixture of S8 rings and linear polysulfides, with the rela-tive abundances of the different sulfur species depending on thegrowth conditions. Although only cyclooctasulfur was detected inBeggiatoa sp. 35Flor cultures at the earliest growth stage, it is pos-sible that minor amounts of polysulfides were also present butcould not be distinguished due to the overlap with the strongRaman signal of predominant S8 rings. This observation is consis-tent with X-ray absorption near edge structure (XANES) spectros-copy of Allochromatium vinosum, which demonstrated a correla-tion between intracellular concentrations of polysulfides (amongother sulfur species) and the bacterial growth phase (17). In cul-tures of Beggiatoa sp. 35Flor, an increase of the polysulfide fractionunder both sulfur-oxidizing and -reducing conditions suggests thatSn

2� species represent a universal pool of activated sulfur utilized inboth the oxidative and reductive parts of the sulfur cycle. Polysulfidescan be derived from cyclic sulfur species by enzymatic cleavage ofSOS bonds with membrane-bound thiol groups or glutathione, aspostulated for, e.g., Thiobacillus thiooxidans, which oxidizes solidelemental sulfur (35). Alternatively, sulfur rings can be openedchemically by strong nucleophiles such as HS� according to thefollowing reaction: n/8 S8 HS� ↔ Sn1

2� H (36). Polysul-fide anions are also strong nucleophiles, and therefore their for-mation could have an autocatalytic effect on SOS bond cleavage.It has been shown that the addition of sodium sulfide to cultures ofsulfur-reducing bacteria enhances the solubilization and reduc-tion of crystalline sulfur (15, 37). Thus, the cleavage of sulfur ringsby HS� or Sn

2� could provide a fortuitous pool of linear sulfurspecies, which, in contrast to large sulfur ring structures, are ac-cessible to enzymes involved in sulfur oxidation and sulfur reduc-tion.

In the case of the Beggiatoa strains that we investigated, it couldnot be determined whether the activation of cyclooctasulfur topolysulfide was biologically or chemically regulated. It is, none-theless, tempting to speculate that bacteria may harness the chem-ical reaction instead of expending energy to cleave sulfur rings. Inthe presence of free sulfide, bacteria might not be able to preventthe decomposition of elemental sulfur to polysulfides. This couldexplain the presence of polysulfides in both intra- and extracellu-lar sulfur globules of two different Beggiatoa alba strains that can-not further oxidize zero-valent sulfur to sulfate. In the environ-ment, Beggiatoa can migrate vertically between oxic and anoxicsediment zones, thus experiencing dramatic fluctuations in thesulfide supply. Periodic sulfide starvation would require the sul-fur-oxidizing Beggiatoa to enzymatically activate S8 rings in orderto utilize stored sulfur. Thus, a combination of environmental andphysiological factors may influence the species of stored sulfurcompounds in these sulfide-oxidizing bacteria.

It is important to note that the speciation of sulfur as a combi-

nation of S8 rings and Sn2� described here is discussed only for

inclusions in neutrophilic, microaerophilic bacteria. Sufficientconcentrations of polysulfides (�10% of total dissolved HS�) ex-ist in solution to support bacterial growth at circumneutral toalkaline pH, but polysulfides undergo hydrolysis in acid (35). Theacidophile Acidithiobacillus ferrooxidans produces extracellularglobules thought to be composed of a cyclooctasulfur core andsurrounded by a hydrophilic layer of polythionates, which arestable only at extremely low pH (6, 38). Although the ultrastruc-ture of sulfur globules in Beggiatoa spp. could not yet be resolved,it is possible that polysulfides contribute to the hydrophilic behav-ior of the biogenic sulfur, analogous to the function of polythio-nates in Acidithiobacillus sp. sulfur globules. A proteinaceousmembrane which encloses intracellular sulfur of Beggiatoa spp.(39) may protect other cellular components from the highly reac-tive Sn

2� ions. With Raman spectroscopy, we were unable to re-solve the exact localization of the sulfur globules within the cell.However, we could observe associations of the sulfur inclusionswith the cell membrane, identified by the presence of c-type cyto-chromes. These observations further confirm previous reports ofunusually high cytochrome c contents in many Beggiatoa strains(39, 40), which in some cases causes pink or orange coloration ofthe bacteria (41, 42).

Our data suggest that elemental sulfur is temporarily stored asan electron reserve, and the activated polysulfide fraction is fur-ther oxidized to sulfate or reduced to sulfide by Beggiatoa. Ourstudy of sulfate production by Beggiatoa sp. 35Flor provides fur-ther insight into the complete biological pathway of sulfide oxida-tion. Unlike other sulfur bacteria such as Thiobacillus denitrificans,which oxidize stored sulfur only after other reduced sulfur com-pounds such as thiosulfate have been depleted (43), Beggiatoa spp.oxidize sulfide and sulfur simultaneously. Our data suggest thatsulfide is completely oxidized to sulfate by Beggiatoa when thesestrains are grown with a low sulfide flux. However, it appears thatunder high sulfide fluxes, the rate of sulfur oxidation cannot keepup with sulfide oxidation, resulting in up to half of the oxidizedsulfide being deposited as sulfur and in some cases causing cells torupture (21). Conversely, when the sulfide supply has been de-pleted, the sulfur oxidation rate substantially exceeds the sulfideoxidation rate. This excess sulfate production can be attributed tothe oxidation of stored sulfur. To date it has been shown thatBeggiatoa spp. store phosphate, nitrate, carbon compounds (suchas PHA), and elemental sulfur intracellularly. Our results showthat Beggiatoa filaments can also contain molar concentrations ofsulfate. Similarly, sulfate has been found inside the related gam-maproteobacterium Thioploca in concentrations up to 100 to1,000 times higher than that in its freshwater environment (44).Storing sulfate offers no apparent advantage to marine bacteria,and it is therefore likely that sulfate is not stored per se but accu-mulates inside the cell as a product of the sulfate-evolving enzymecomplex APS reductase/ATP sulfurylase or another sulfite-oxidiz-ing enzyme, depending on the organism (40, 42, 45). In fact, sul-fate appears to be localized in regions around degraded sulfurglobules (Fig. 4B) where it is produced. Not much is known aboutthe mechanism of sulfate export in Beggiatoa. A gene for the H-driven sulfate permease sulP was identified in the genome of Beg-giatoa sp. 35Flor (M. Winkel, S. Verena, T. Woyke, H. Schulz-Vogt, M. Richter, and M. Mußmann, unpublished data). Theactivity of this symporter may be limited by the energy requiredfor ion export or by the low surface-to-volume ratio of these large

Berg et al.

634 aem.asm.org Applied and Environmental Microbiology

on April 27, 2019 by guest

http://aem.asm

.org/D

ownloaded from

sulfur bacteria, which restricts substrate-product exchange withthe environment.

With our in vivo Raman analyses presented here, we attemptedto resolve some of the discrepancies surrounding the discussionsof the chemical species composition of bacterial sulfur globules. Inthe literature there are conflicting reports of sulfur rings (10, 46),polysulfides (11), or polythionates (5, 6) in these inclusions. Basedon our results, we hypothesize that even within the same species ofbacteria, the speciation of stored sulfur varies under differentecophysiological conditions.

ACKNOWLEDGMENTS

We thank Martina Meyer, Kirsten Imhoff, and Daniela Franzke for tech-nical support and Sandra Havemeyer for kindly providing Beggiatoa albastrains B15LD and B18LD. We are grateful to Heide Schulz-Vogt for help-ful discussions and for providing cultures of Beggiatoa sp. 35Flor.

Funding was provided by the International Max Planck ResearchSchool of Marine Microbiology, the Max Planck Society, and the Deut-sche Forschungsgemeinschaft (through the MARUM Center for MarineEnvironmental Sciences).

REFERENCES1. Winogradsky S. 1887. Uber Schwefelbakterien. Bot. Zeitung 45:489 – 610.2. Dahl C, Prange A. 2006. Inclusions in prokaryotes, p 21–51. Springer,

Berlin, Germany.3. Nicolson GL, Schmidt GL. 1971. Structure of the Chromatium sulfur

particle and its protein membrane. J. Bacteriol. 105:1142–1148.4. Brune DC. 2004. Anoxygenic photosynthetic bacteria, p 847– 870.

Springer, Houten, Netherlands.5. Steudel R, Holdt G, Göbel T, Hazeu W. 1987. Chromatographic sepa-

ration of higher polythionates SnO62� (n 3. . .22) and their detection in

cultures of Thiobacillus ferroxidans; molecular composition of bacterialsulfur secretions. Angew. Chem. Int. Ed. Engl. 26:151–153. http://dx.doi.org/10.1002/anie.198701511.

6. He H, Xia J-L, Jiang H-C, Yan Y, Liang C-L, Ma C-Y, Zheng L, ZhaoY-D, Qiu G-Z. 2010. Sulfur species investigation in extra- and intra-cellular sulfur globules of Acidithiobacillus ferrooxidans and Acidithioba-cillus caldus. Geomicrobiol. J. 27:707–713. http://dx.doi.org/10.1080/01490451003597655.

7. Trüper HG, Hathaway JC. 1967. Orthorhombic sulfur formed by pho-tosynthetic sulphur bacteria. Nature 215:435– 436. http://dx.doi.org/10.1038/215435a0.

8. Mas J, Gemerden H. 1987. Influence of sulfur accumulation and com-position of sulfur globule on cell volume and buoyant density of Chroma-tium vinosum. Arch. Microbiol. 146:362–369. http://dx.doi.org/10.1007/BF00410937.

9. Vetter RD. 1985. Elemental sulfur in the gills of three species of clamscontaining chemoautotrophic symbiotic bacteria: a possible inorganic en-ergy storage compound. Mar. Biol. 88:33– 42. http://dx.doi.org/10.1007/BF00393041.

10. Pasteris JD, Freeman JJ, Goffredi SK, Buck KR. 2001. Raman spectro-scopic and laser scanning confocal microscopic analysis of sulfur in livingsulfur-precipitating marine bacteria. Chem. Geol. 180:3–18. http://dx.doi.org/10.1016/S0009-2541(01)00302-3.

11. Prange A, Chauvistré R, Modrow H, Hormes J, Trüper HG, Dahl C.2002. Quantitative speciation of sulfur in bacterial sulfur globules: X-rayabsorption spectroscopy reveals at least three different species of sulfur.Microbiology 148:267–276.

12. Roy AB, Trudinger PA. 1970. The biochemistry of inorganic compoundsof sulphur. Cambridge University Press, Cambridge, United Kingdom.

13. Meyer B. 1964. Solid allotropes of sulfur. Chem. Rev. 64:429 – 451. http://dx.doi.org/10.1021/cr60230a004.

14. Prange A, Arzberger I, Engemann C, Modrow H, Schumann O, TrüperHG, Steudel R, Dahl C, Hormes J. 1999. In situ analysis of sulfur in thesulfur globules of phototrophic sulfur bacteria by X-ray absorption nearedge spectroscopy. Biochim. Biophys. Acta 1428:446 – 454. http://dx.doi.org/10.1016/S0304-4165(99)00095-1.

15. Blumentals II, Itoh M, Olson GJ, Kelly RM. 1990. Role of polysulfides in

reduction of elemental sulfur by the hyperthermophilic archaebacteriumPyrococcus furiosus. Appl. Environ. Microbiol. 56:1255–1262.

16. Engel AS, Lichtenberg H, Prange A, Hormes J. 2007. Speciation of sulfurfrom filamentous microbial mats from sulfidic cave springs using X-rayabsorption near-edge spectroscopy. FEMS Microbiol. Lett. 269:54 – 62.http://dx.doi.org/10.1111/j.1574-6968.2006.00600.x.

17. Franz B, Gehrke T, Lichtenberg H, Hormes J, Dahl C, Prange A. 2009.Unexpected extracellular and intracellular sulfur species during growth ofAllochromatium vinosum with reduced sulfur compounds. Microbiology155:2766 –2774. http://dx.doi.org/10.1099/mic.0.027904-0.

18. Mussmann M, Hu FZ, Richter M, de Beer D, Preisler A, Jørgensen BB,Huntemann M, Glöckner FO, Amann R, Koopman WJ. 2007. Insightsinto the genome of large sulfur bacteria revealed by analysis of singlefilaments. PLoS Biol. 5:e230. http://dx.doi.org/10.1371/journal.pbio.0050230.

19. Pott AS, Dahl C. 1998. Sirohaem sulfite reductase and other proteinsencoded by genes at the dsr locus of Chromatium vinosum are involved inthe oxidation of intracellular sulfur. Microbiology 144:1881–1894. http://dx.doi.org/10.1099/00221287-144-7-1881.

20. Kletzin A. 2008. Microbial sulfur metabolism, p 184 –201. Springer, Ber-lin, Germany.

21. Schwedt A, Kreutzmann AC, Polerecky L, Schulz-Vogt HN. 2011. Sulfurrespiration in a marine chemolithoautotrophic Beggiatoa strain. Front.Microbiol. 2:276. http://dx.doi.org/10.3389/fmicb.2011.00276.

22. Strohl WR, Larkin JM. 1978. Enumeration, isolation, and characteriza-tion of Beggiatoa from freshwater sediments. Appl. Environ. Microbiol.36:755–770.

23. Barletta RE, Gros BN, Herring MP. 2009. Analysis of marine biogenicsulfur compounds using Raman spectroscopy: dimethyl sulfide and meth-ane sulfonic acid. J. Raman Spectrosc. 40:972–981. http://dx.doi.org/10.1002/jrs.2211.

24. Kühl M, Steuckart C, Eickert G, Jeroschewski P. 1998. A H2S microsen-sor for profiling biofilms and sediments—application in an acidic lakesediment. Aquat. Microb. Ecol. 15:201–209. http://dx.doi.org/10.3354/ame015201.

25. Jørgensen BB, Revsbech NP, Blackburn TH, Cohen Y. 1979. Diurnalcycle of oxygen and sulfide microgradients and microbial photosynthesisin a cyanobacterial mat sediment. Appl. Environ. Microbiol. 38:46 –58.

26. Harvey PD, Butler IS. 1986. Raman spectra of orthorhombic sulfur at 40K. J. Raman Spectrosc. 17:329 –334. http://dx.doi.org/10.1002/jrs.1250170407.

27. Janz GJ, Downey JR, Jr, Roduner E, Wasilczyk GJ, Coutts JW, EluardA. 1976. Raman studies of sulfur-containing anions in inorganic polysul-fides. Sodium polysulfides. Inorg. Chem. 15:1759 –1763. http://dx.doi.org/10.1021/ic50162a004.

28. Khan SA, Hughes RW, Reynolds PA. 2011. Raman spectroscopic deter-mination of oxoanions in aqueous polysulfide electrolyte solutions. Vib.Spectrosc. 56:241–244. http://dx.doi.org/10.1016/j.vibspec.2011.03.001.

29. Pätzold R, Keuntje M, Theophile K, Müller J, Mielcarek E, NgezahayoA, Anders-von Ahlften A. 2008. In situ mapping of nitrifiers and anam-mox bacteria in microbial aggregates by means of confocal resonance Ra-man microscopy. J. Microbiol. Methods 72:241–248. http://dx.doi.org/10.1016/j.mimet.2007.12.003.

30. Eckert B, Steudel R. 2003. Molecular spectra of sulfur molecules and solidsulfur allotropes. Top. Curr. Chem. 231:181–191. http://dx.doi.org/10.1007/b13181.

31. Schmidt TM, Arieli B, Cohen Y, Padan E, Strohl WR. 1987. Sulfurmetabolism in Beggiatoa alba. J. Bacteriol. 169:5466 –5472.

32. De Gelder J, Willemse-Erix D, Scholtes MJ, Sanchez JI, Maquelin K,Vandenabeele P, De Boever P, Puppels GJ, Moens L, De Vos P. 2008.Monitoring poly(3-hydroxybutyrate) production in Cupriavidus necatorDSM 428 (H16) with Raman spectroscopy. Anal. Chem. 80:2155–2160.http://dx.doi.org/10.1021/ac702185d.

33. Strohl WR, Howard KS, Larkin JM. 1982. Ultrastructure of Beggiatoaalba strain B15LD. J. Gen. Microbiol. 128:73– 84.

34. Strohl WR, Geffers I, Larkin JM. 1981. Structure of the sulfur inclusionenvelopes from four Beggiatoas. Curr. Microbiol. 6:75–79. http://dx.doi.org/10.1007/BF01569007.

35. Suzuki I. 1965. Oxidation of elemental sulfur by an enzyme system ofThiobacillus thiooxidans. Biochim. Biophys. Acta 104:359 –371. http://dx.doi.org/10.1016/0304-4165(65)90341-7.

36. Schauder R, Müller E. 1993. Polysulfide as a possible substrate for sulfur-reducing bacteria. Arch. Microbiol. 160:377–382.

Polysulfide Intermediates in Bacterial S0 Metabolism

January 2014 Volume 80 Number 2 aem.asm.org 635

on April 27, 2019 by guest

http://aem.asm

.org/D

ownloaded from

37. Zöphel A, Kennedy MC, Beinert H, Kroneck PMH. 1988. Investigationson microbial sulfur respiration. Arch. Microbiol. 150:72–77. http://dx.doi.org/10.1007/BF00409720.

38. Steudel R. 1989. On the nature of the ‘elemental sulfur’ (S0) produced bysulphur-oxidizing bacteria—a model for S0 globules, p 289 –303. In Schle-gel HG, Bowien B (ed), Autotrophic bacteria. Springer Verlag, Berlin,Germany.

39. Larkin JM, Strohl WR. 1983. Beggiatoa, Thiothrix, and Thioploca. Annu.Rev. Microbiol. 37:341–367. http://dx.doi.org/10.1146/annurev.mi.37.100183.002013.

40. Schmidt TM, DiSpirito AA. 1990. Spectral characterization of c-typecytochromes purified from Beggiatoa alba. Arch. Microbiol. 154:453– 458.http://dx.doi.org/10.1007/BF00245227.

41. Prince RC, Stokley KE, Haith CE, Jannasch HW. 1988. The cytochromesof a marine Beggiatoa. Arch. Microbiol. 150:193–196. http://dx.doi.org/10.1007/BF00425161.

42. MacGregor BJ, Biddle JF, Siebert JR, Staunton E, Hegg EL, MatthysseAG, Teske A. 2013. Why orange Guaymas Basin Beggiatoa spp. are orange:

single-filament-genome-enabled identification of an abundant octahemecytochrome with hydroxylamine oxidase, hydrazine oxidase, and nitritereductase activities. Appl. Environ. Microbiol. 79:1183–1190. http://dx.doi.org/10.1128/AEM.02538-12.

43. Schedel M, Trüper HG. 1980. Anaerobic oxidation of thiosulfate andelemental sulfur in Thiobacillus denitrificans. Arch. Microbiol. 124:205–210.

44. Kojima H, Nakajima T, Fukui M. 2007. Carbon source utilization andaccumulation of respiration-related substances by freshwater Thioplocaspecies. FEMS Microbiol. Ecol. 59:23–31. http://dx.doi.org/10.1111/j.1574-6941.2006.00201.x.

45. Kappler U, Dahl C. 2001. Enzymology and molecular biology of prokary-otic sulfite oxidation. FEMS Microbiol. Lett. 203:1–9. http://dx.doi.org/10.1016/S0378-1097(01)00304-4.

46. George GN, Gnida M, Bazylinski DA, Prince RC, Pickering IJ. 2008.X-ray absorption spectroscopy as a probe of microbial sulfur biochemis-try: the nature of bacterial sulfur globules revisited. J. Bacteriol. 190:6376 –6383. http://dx.doi.org/10.1128/JB.00539-08.

Berg et al.

636 aem.asm.org Applied and Environmental Microbiology

on April 27, 2019 by guest

http://aem.asm

.org/D

ownloaded from