A bacterial isolate from the Black Sea oxidizes sulfide with manganese(IV) oxide

Abstract

Mn is one of the most abundant redox-sensitive metals on earth. Some microorganisms are known to use Mn(IV) oxide (MnO₂) as electron acceptor for the oxidation of organic compounds or hydrogen (H₂), but so far the use of sulfide (H₂S) has been suggested but not proven. Here we report on a bacterial isolate which grows autotrophically and couples the reduction of MnO₂ to the oxidation of H₂S or thiosulfate (S₂O₃²⁻) for energy generation. The isolate, originating from the Black Sea, is a species within the genus Sulfurimonas, which typically occurs with high cell numbers in the vicinity of sulfidic environments [Y. Han, M. Perner, Front. Microbiol. 6, 989 (2015)]. H₂S and S₂O₃²⁻ are oxidized completely to sulfate (SO₄²⁻) without the accumulation of intermediates. In the culture, Mn(IV) reduction proceeds via Mn(III) and finally precipitation of Ca-rich Mn(II) carbonate [Mn(Ca)CO₃]. In contrast to Mn-reducing bacteria, which use organic electron donors or H₂, Fe oxides are not observed to support growth, which may either indicate an incomplete gene set or a different pathway for extracellular electron transfer.

In stratified basins, for example the Black Sea, in between the oxygenated surface waters and sulfidic bottom waters a suboxic zone lacking oxygen (O₂), H₂S, and mostly also nitrate (NO₃⁻) has been frequently reported (1). Despite the absence of electron acceptors, high bacterial CO₂ fixation rates at the border with sulfidic waters were measured, without a known energy metabolism which could fuel growth under these environmental conditions (2, 3). Thermodynamically, a suitable electron acceptor for H₂S oxidation at this depth would be Mn. Even though Mn concentrations are low (Fig. 1A), the oxidized form MnO₂ is particulate and is therefore transported much faster than dissolved electron acceptors to the sulfidic waters by gravitational sinking (4). Nevertheless, so far all attempts to cultivate microorganisms which oxidize H₂S with MnO₂ have failed.

Fig. 1.

(A) Water column profile of the Black Sea suboxic zone: O₂ (blue), NO₃⁻ (green), H₂S (yellow), and particulate (part.) Mn (black triangles). Red arrow indicates sampling for cultivation. Data from ref. 19. (B) Scanning electron microscopy (SEM) picture of the isolate ‘S. marisnigri.’ (C and D) Bottle photo and SEM picture of the particulate Mn phase before (C, Insert) and after (D, Insert) growth. (Scale bars: B, 1 $\mathrm{\mu }$M; C and D, 10 $\mathrm{\mu }$m.)

During an expedition on the research vessel Maria S. Merian in November 2013 we sampled the water column of the Black Sea, focusing on the suboxic zone. We took a water sample at the depth of highest abundance of Epsilonbacteraeota [12 to 15% (2)] and transferred it into a gas-tight serum bottle containing MnO₂ (Fig. 1A, red arrow). At this depth H₂S was detectable and neither O₂ nor NO₃⁻ was present. After a first enrichment with daily additions of H₂S resulting in ∼10 to 20 μM concentrations, we transferred a small volume into an artificial medium. From here on we used mainly S₂O₃²⁻ for cultivation instead of H₂S as it is a more convenient electron donor, nontoxic even in higher concentrations, and nonreactive with the MnO₂ used in our study. A single strain was isolated by repeated series of dilutions-to-extinction transfers.

The isolate belongs to the genus Sulfurimonas and its closest cultured relative is Sulfurimonas gotlandica, known for the oxidation of H₂S with NO₃⁻ in the pelagic redoxcline of the Baltic Sea (5) with a 3% difference in the full 16S rRNA gene sequence. We propose calling the strain ‘Sulfurimonas marisnigri,’ in reference to the Latin notation of Pontus Euxı̄nus, meaning Sulfurimonas from the Black Sea. The cells are slightly curved, with lengths of 1 to 4 $\mathrm{\mu }$m and widths of 200 to 300 nm (Fig. 1B). ‘S. marisnigri’ grows autotrophically with doubling times of 9 to 13 h during the exponential growth phase and reaches a final cell density of 3 to 6 $\times$ 10⁷ cells per mL after 7 to 10 d (Fig. 2A). Toward the end of the growth phase, the medium turns from black to brownish-gray (Fig. 1 C and D, Inserts), due to the reduction of MnO₂ and precipitation of Mn(Ca)CO₃ (Fig. 2B). Even though this may be an artifact due to the cultivation conditions, this particular mineral phase was reported in exceptional amounts from the anoxic basins of the Baltic Sea, but the mechanism of its formation is still under debate (6). Cultivation of ‘S. marisnigri’ with NO₃⁻ and successive additions of H₂S resulted in growth and undetectable H₂S levels in the culture, indicating a principal ability to use H₂S directly as electron donor. In contrast to Shewanella oneidensis and Geobacter metallireducens, attempts to cultivate ‘S. marisnigri’ with amorphous FeOOH, goethite ($\alpha$-FeOOH), Fe₂O₃, and FeCl₃ were unsuccessful, leading to the conclusion that Fe(III) was not a viable electron acceptor under these conditions. This may be due to the absence of a critical protein component for the reduction of Fe oxides or could indicate that extracellular electron transfer onto MnO₂ might function in a different manner.

Fig. 2.

Growth of ‘S. marisnigri’ with MnO₂ and S₂O₃²⁻ (A and B) and MnO₂ and H₂S (C). (A) Increasing cell counts (crosses, dashed line) concurrent with the complete oxidation of S₂O₃²⁻ (black solid circles) to SO₄²⁻ (black open circles). (B) Occurrence of different Mn species during growth: particulate (part.) Mn [MnO₂ and later Mn(Ca)CO₃, black solid circles], dissolved (diss.) Mn (black open circles, solid line), and diss. reactive Mn (Mn³⁺, black open circles, dashed line). The reduction of part. MnO₂ results in an equal increase of diss. Mn, almost exclusively Mn³⁺, until day 6. Molar ratio of Ca to Mn in the particulate phase (bars) indicates the precipitation of Mn(Ca)CO₃ which was used to calculate the amount of reduced Mn in the particulate phase (blue solid circles, dashed line). (C) Cultivation of ‘S. marisnigri’ with MnO₂ and constant addition of 3 $\mathrm{\mu }$L$\cdot$min⁻¹ (until day 12) and after that 7.5 $\mathrm{\mu }$L$\cdot$min⁻¹ of an ∼10 mM Na₂S stock solution results in growth (crosses, dashed line), accumulation of SO₄²⁻ (black squares), and low concentration of S⁰ (black solid circles). In the sterile control (blue), S⁰ (solid blue circles) is accumulating and concentrations of SO₄²⁻ (blue squares) increase just slightly.

Growth of ‘S. marisnigri’ with MnO₂ and a constant supply of H₂S resulted in accumulation of SO₄²⁻ as cell numbers increased, and negligible concentrations of elemental sulfur (S⁰). In the sterile control, however, S⁰ accumulated and SO₄²⁻ increased just slightly (Fig. 2C). Likewise, in the central gyres of the Black Sea S₂O₃²⁻ and sulfite (SO₃²⁻) were undetectable (7) and S⁰ occurred in nanomolar concentrations (8). Growth with MnO₂ and S₂O₃²⁻ was accompanied by the complete oxidation of S₂O₃²⁻ to SO₄²⁻ with a stoichiometry of 1:2 (Fig. 2A) and following Eq. 1:

$$\mathrm{4}\hspace{0.17em}\mathrm{M}\mathrm{n}{\mathrm{O}}_2+{\mathrm{S}}_2{{\mathrm{O}}_3}^{2-}+6\hspace{0.17em}{\mathrm{H}}^{+}\hspace{0.17em}\to \hspace{0.17em}4\hspace{0.17em}\mathrm{M}{\mathrm{n}}^{2+}+2\hspace{0.17em}\mathrm{S}{{\mathrm{O}}_4}^{2-}+3\hspace{0.17em}{\mathrm{H}}_2\mathrm{O}.$$ [1]

As with H₂S, no detectable accumulation of S⁰ and SO₃²⁻ was observed with S₂O₃²⁻ as electron donor. Growth was observed concurrent with the reduction of Mn(IV) to Mn(III) from day 3 to 6 and continued with the reduction of Mn(III) to Mn(II), leading to the precipitation of Ca-rich particles (Fig. 2 A and B), in contrast to S. oneidensis, where growth is only supported by the reduction of Mn(III) to Mn(II) (9). Total inorganic carbon (TIC) analysis of the particulate fraction at the end of the growth phase confirmed the formation of carbonate. The Ca-to-Mn ratio of single Mn(Ca)CO₃ precipitates was determined using SEM and energy dispersive X-ray microanalysis (EDX). The mean ratio of 0.19 with an SD of 0.03 in cube-shaped precipitates was used to estimate the proportion of reduced Mn in the particulate phase (Fig. 2B, blue dashed line). With this approach, we can show that MnO₂ was almost completely reduced and transformed to Mn(Ca)CO₃.

We excluded disproportionation of S₂O₃²⁻as an alternate explanation for our findings. Cultivation of ‘S. marisnigri’ solely on S₂O₃²⁻ as an energy source did not result in growth and formation of H₂S or SO₄²⁻. This is supported by cultivation of ‘S. marisnigri’ with MnO₂ and a surplus of S₂O₃²⁻, in which the concentration of S₂O₃²⁻ remained constant after the depletion of MnO₂. Finally, disproportionation in the presence of Fe oxides would have led to the precipitation of FeS and FeS₂, which was not the case.

The isolate belongs to the group of Epsilonbacteraeota which is reported to be highly abundant in the redox trasition zones of, for example, the Black Sea (12% of the total bacterial community), the Baltic Sea (21%), and the Cariaco Basin (27%) (2, 10). In these systems, Epsilonbacteraeota can be responsible for up to 100% of the dark CO₂ fixation activity in the absence of O₂ and often NO₃⁻. Therefore, Jost et al. (3) and Taylor et al. (11) already suggested a potentially MnO₂-dependent H₂S oxidation by autotrophic bacteria.

In addition to its presence in pelagic environments, the genus Sulfurimonas is globally abundant in redox transition environments such as hydrothermal vents and marine sediments (12). In sediments, the addition of MnO₂ is thought to promote the production of SO₄²⁻, apparently depending on microbial activity and leading to the precipitation of Mn(Ca)CO₃ (13). In those experiments, addition of FeOOH did not stimulate SO₄²⁻ production. Those findings fit remarkably well to our observations in pure culture, suggesting that an organism with a physiology similar to that of ‘S. marisnigri’ may have been responsible for the observed activity.

A further indication that bacterial H₂S oxidation with MnO₂ may be of more general importance is the formation of intermediate Mn³⁺ in our cultures, which we detected indirectly as dissolved reactive Mn (14) (Fig. 2B). Mn³⁺ was reported to be a major constituent of the marine Mn cycle both in sediments (15) and in the water column of the Black Sea (16) in the absence of O₂ and H₂S (17). So far, known processes mediating Mn³⁺ formation are the oxidation of organic matter with MnO₂ reduction, enzymatic oxidation of Mn²⁺, and the abiotic reaction of MnO₂ with Fe(II) and H₂S (15, 18). Our study adds another biologically mediated process via lithotrophic MnO₂ reduction, which can promote the buildup of Mn³⁺, as observed both in marine sediments and across pelagic redoxclines. In conclusion, we suggest that this bacterial metabolism, which we prove here in pure culture, may be widespread in pelagic redoxclines and to a minor extent in marine sediments where H₂S is produced and Mn is present in sufficient amounts with important consequences for Mn and S cycling.

Materials and Methods

Gases and Nutrients.

Gases and nutrients in the water column were measured as reported in Schulz-Vogt et al. (19).

Cultivation.

Water samples were taken at 44°−16.7586 N and 36°−18.9567 E at the depth indicated in Fig. 1. Culture purity was ensured by sequencing, lack of growth in organic-rich media, and by microscopy. Medium for cultivation and experiments was prepared anaerobically following the technique described by Widdel and Pfennig (20) based on a SO₄²⁻-free artificial seawater with a salinity of 21 and a pH of 7.5. The following sterile and anoxic components were added (milliliters per liter): 1 M NaHCO₃, 30; 1 mM NH₄Cl, 20; Pfennig’s trace elements solution SL7, 1; 10 mM Na₂HPO₄, 1; and vitamin solution, 0.42. Vitamin solution contained (milligrams per milliliter) the following: B₁₂, 0.1; inositol, 0.1; biotin, 0.1; folic acid, 0.1; PABA, 1; nicotinic acid, 10; d-pantothenate, 10; and thiamine, 20. Technical MnO₂ provided as electron acceptor was purchased from Merck and additionally grinded with an agate ball mill.

Molecular Analysis and Bioinformatics.

DNA was extracted using QIAamp DNA Mini Kit (51306) following the manual. For PCR we used Thermo Fisher Scientific Kit EP0072 according to the manufacturer’s protocol description and primers 27f -1492r. PCR products were purified with the Agentcourt AMpure XP (Beckman Coulter GmbH) magnetic beads and cloned into the vector pSC-A-amp/kan (StrataClone PCR Cloning Kit, competent cells Strataclone SoloPack) following the manual. Clones were sequenced (LGC) forward and reverse using vector primers, trimmed, manually corrected, assembled and deposited at NCBI GenBank (accession nos. MF563385–MF563475).

Cell Counting.

A glutaraldehyde-fixed sample (2.5% final concentration) was treated with at least five times the volume of a hydroxylamine solution (1.5 M NH₂OH$\cdot$HCl dissolved in 0.25 M HCl) to dissolve Mn particles. The mixture was treated with ultrasonic for 5 min and subsequently filtered onto a 0.2-$\mathrm{\mu }$m polycarbonate filter and embedded in DAPI-containing oil. Cells were enumerated with an epifluorescence microscope at 1,000× magnification.

S₂O₃²⁻ and SO₃²⁻.

We added 0.5 mL 0.2-$\mathrm{\mu }$m-filtered subsample to 25 $\mathrm{\mu }$L of Hepes/EDTA buffer (200 mM/50 mM in MilliQ) and 25 $\mathrm{\mu }$L of the monobrombimane solution (48 mM in acetonitrile) and incubated at room temperature for 30 min in the dark. Derivatization was stopped by adding 25 $\mathrm{\mu }$L of methanesulfonic acid (324 mM in MilliQ water). Samples were diluted with MilliQ water and measured daily. Samples and calibration series were analyzed using a BioTek HPLC System with pump 525 (1 mL$\cdot$min⁻¹), oven 582 (40 °C), column LiChrospher 60 RP-Select B (5 $\mathrm{\mu }$m) 125 × 4, and a Jasco FP 1520 detector (excitation at 380 nm, emission at 480 nm). Data were analyzed with the program Geminyx III Version 1.10.3.7. Eluent A contained 0.25% acetic acid and eluent B 100% methanol. The gradient protocol was as follows: 0 min 100% A, 2 min 100% A, 5.5 min 92% A, 8 min 68% A, 12 min 68% A, 13 min 0% A, 18 min 0% A, 19 min 100% A, and 23 min 100% A. With these adjustments, S₂O₃²⁻ peak appeared after 10 min and SO₃²⁻ peak after 8 min.

S⁰.

A 900-$\mathrm{\mu }$L sample was added to 100 $\mathrm{\mu }$L of zinc acetate solution (5% wt/vol). Then, 500 $\mathrm{\mu }$L chloroform were added and intensively mixed for 2 min. One hundred microliters of the chloroform phase were diluted with 400 $\mathrm{\mu }$L methanol and measured with HPLC using BioTek HPLC System with pump 525 (1 mL$\cdot$min⁻¹), oven 582 (25 °C), column LiChrospher 100 RP-18 B (5 $\mathrm{\mu }$m) 125 × 4, and DAD 545V detector (wavelength 265 nm and 6-nm bandwidth). Data analysis was as described above. An isocratic gradient with 100% methanol was applied. With these adjustments, the peak appeared after 4.4 min.

Mn and Ca.

Part. Mn and Ca dissolved in hydroxylamine solution as well as total diss. (<0.2 $\mathrm{\mu }$m) Mn and diss. reactive Mn [dMn_react, mainly comprising Mn³⁺ (14)] were measured by ICP-OES (iCAP 7400 Duo; Thermo Fisher Scientific) using an external calibration and Sc as internal standard. Precision and accuracy were checked by international reference materials (SGR-1b for the part. and SLEW-3 for the diss. fraction) and were below 2%.

SEM-EDX.

A Zeiss Merlin Compact SEM (variable pressure, in-lens SE and BSE detector) equipped with EDX (Oxford Instruments) was used to identify Mn(Ca)CO₃ precipitates and to directly quantify the Ca-to-Mn ratios. The sample preparation was done as described elsewhere (19). Reduced Mn in the part. phase was calculated with Eq. 2:

$$\mathrm{p}\mathrm{a}\mathrm{r}\mathrm{t}.\mathrm{r}\mathrm{e}\mathrm{d}.\mathrm{M}\mathrm{n}\hspace{0.17em}=\mathrm{p}\mathrm{a}\mathrm{r}\mathrm{t}.\hspace{0.17em}\mathrm{M}\mathrm{n}\hspace{0.17em}\mathrm{*}\hspace{0.17em}\frac{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}\hspace{0.17em}\mathrm{p}\mathrm{a}\mathrm{r}\mathrm{t}.\hspace{0.17em}\mathrm{C}\mathrm{a}/\mathrm{M}\mathrm{n}\hspace{0.17em}\mathrm{b}\mathrm{y}\hspace{0.17em}\mathrm{I}\mathrm{C}\mathrm{P}-\mathrm{O}\mathrm{E}\mathrm{S}}{\mathrm{s}\mathrm{p}\mathrm{o}\mathrm{t}\hspace{0.17em}\mathrm{C}\mathrm{a}/\mathrm{M}\mathrm{n}\hspace{0.17em}\mathrm{b}\mathrm{y}\hspace{0.17em}\mathrm{S}\mathrm{E}\mathrm{M}-\mathrm{E}\mathrm{D}\mathrm{X}\hspace{0.17em}\mathrm{i}\mathrm{n}\hspace{0.17em}\mathrm{M}\mathrm{n}\left(\mathrm{C}\mathrm{a}\right)\mathrm{C}{\mathrm{O}}_3}.$$ [2]

TIC.

Dried material was treated with 40% H₃PO₄ and the released CO₂ was analyzed by an IR detector (multi-EA 4000; Analytic Jena). Pure standard CaCO₃ (12.0% TIC) was used for calibration.