ABSTRACT
A lithotrophic freshwater Beggiatoa strain was enriched in O2-H2S gradient tubes to investigate its ability to oxidize sulfide with NO3− as an alternative electron acceptor. The gradient tubes contained different NO3− concentrations, and the chemotactic response of the Beggiatoa mats was observed. The effects of the Beggiatoa sp. on vertical gradients of O2, H2S, pH, and NO3− were determined with microsensors. The more NO3− that was added to the agar, the deeper the Beggiatoa filaments glided into anoxic agar layers, suggesting that the Beggiatoa sp. used NO3− to oxidize sulfide at depths below the depth that O2 penetrated. In the presence of NO3−Beggiatoa formed thick mats (>8 mm), compared to the thin mats (ca. 0.4 mm) that were formed when no NO3− was added. These thick mats spatially separated O2 and sulfide but not NO3− and sulfide, and therefore NO3− must have served as the electron acceptor for sulfide oxidation. This interpretation is consistent with a fourfold-lower O2 flux and a twofold-higher sulfide flux into the NO3−-exposed mats compared to the fluxes for controls without NO3−. Additionally, a pronounced pH maximum was observed within the Beggiatoa mat; such a pH maximum is known to occur when sulfide is oxidized to S0 with NO3− as the electron acceptor.
Beggiatoa spp. are gliding, filamentous, colorless sulfur bacteria (22). These multicellular bacteria can occur in dense mats at the surface of sulfide-rich sediments in many freshwater and marine habitats (2, 10, 11, 21). The filaments of bigger marine species of Beggiatoa can be more than 120 μm wide (2) and >1 cm long, are white, and are visible with the naked eye; even single filaments of narrow freshwater Beggiatoa species whose filaments are ca. 3 μm wide (14, 21) can be observed with a stereomicroscope. Beggiatoa spp. are sulfide-oxidizing bacteria that have an important effect on the benthic sulfur cycle (4, 6). The presence of Beggiatoa mats at the sediment surface prevents toxic sulfide from diffusing into the water column, because biological sulfide oxidation is much more rapid and efficient than chemical sulfide oxidation (13).
In addition, Beggiatoa spp. can have a great effect on the aquatic nitrogen cycle when they use NO3− anaerobically as an alternative electron acceptor in place of O2. The ability of freshwater and marine Beggiatoa spp. to oxidize sulfide anaerobically with NO3− has been studied for some time (11, 19, 20, 21), especially because large marine species contain a vacuole in which NO3− can be stored at concentrations up to 160 mmol/liter (11). This enables the filaments to penetrate into anoxic sediment layers and perform anaerobic sulfide oxidation. However, anaerobic sulfide oxidation by freshwater Beggiatoa species has not been unequivocally documented, and the impact of freshwater Beggiatoa species on the nitrogen cycle is unclear (5, 11). Therefore, there is significant interest in obtaining more information about possible anaerobic sulfide oxidation with NO3− by freshwater Beggiatoa species.
The freshwater Beggiatoa strain that was used in this study was sustained for more than 2 years in highly enriched O2-H2S gradient tubes (12). Using microsensors to measure changes in the O2 contents, H2S contents, pH, and NO3− contents in these gradient tubes, the position of the Beggiatoa filaments in the transparent agar could be optically related to high-resolution chemical gradients. This experimental approach was used to address the following questions. (i) Does the freshwater Beggiatoa sp. exhibit a chemotactic response to the presence of different NO3− and H2S concentrations? (ii) Does a Beggiatoa mat use NO3− as an alternative electron acceptor in place of O2? (iii) Do the Beggiatoa filaments alter the vertical O2, H2S, and pH gradients differently when they are exposed to NO3− in addition to O2?
MATERIALS AND METHODS
Sampling site and cultivation.Samples of Beggiatoa sp. with a filament width of 3 μm were collected in 2003 from the NO3−-rich stream Giber Aa, south of Aarhus, Denmark. Here, mats of Beggiatoa were found on the mud around outlets for primary treated sewage.
The Beggiatoa filaments were enriched in lithotrophic agar gradient tubes, modified as described by Nelson and Jannasch (12). These gradient tubes contained two layers of agar, a layer of dense bottom agar (1.5% Bacto Agar [Difco Laboratories]) containing a high ΣH2S concentration ([ΣH2S] = [H2S] + [HS−] + [S2−]) overlaid by a layer of softer top agar (0.25%) without ΣH2S, which led to opposing gradients of ΣH2S and O2 in the top agar. The composition of the medium is shown in Table 1. The pH was adjusted to approximately 7.0 with NaOH. The gradients were prepared in screw-cap tubes (length, 150 mm; inside diameter, 14 mm). The tubes were filled with 4 ml of autoclaved bottom agar and 8 ml of top agar. Unless indicated otherwise, the bottom agar was prepared with 4 mmol/liter Na2S. The top agar also contained 150 μl of a sterile vitamin solution (Table 1), 4 mmol/liter NaHCO3, and, unless indicated otherwise, 50 μmol/liter NaNO3, 50 μmol/liter NH4Cl, and 50 μmol/liter sodium acetate. The screw caps on the tubes were left loose to permit exchange of the headspace gas with the atmosphere. To allow gradient development, the agar was aged for at least 2 days before inoculation. For the different experiments, Beggiatoa filaments were taken from existing gradient tubes, pooled, and mixed, and identical subsamples of enriched Beggiatoa biomass were inoculated approximately 5 mm below the agar surface. All cultures were grown at room temperature in the dark.
Compositions of medium, micronutrient solution, and vitamin solution
Vertical position of the Beggiatoa mats.For determination of the NO3−- and ΣH2S-dependent vertical positions of the Beggiatoa mats, the agar was prepared with 0, 100, 200, 400, and 600 μmol/liter NaNO3 and with 4 and 8 mmol/liter Na2S, respectively (n = 3). The mat positions within the gradient system were determined using the tip of a microsensor dummy as a pointer. The dummy was mounted vertically on a micromanipulator, which was attached to a heavy stand. Via its motor drive, the micromanipulator allowed slow, small-scale insertion of the microsensor dummy into the agar down to the Beggiatoa mat, while the tip was viewed through the side of the gradient tube with a stereomicroscope (magnification, ×10 to ×20). The meniscus of the agar surface was defined as a depth of 0 μm, from which the position of the clearly visible upper boundary of the Beggiatoa mat was measured. The mat position was determined 1 to 6 days after inoculation.
Chemical microgradients.The O2 concentrations, H2S concentrations, pH values, and NO3− concentrations in the gradient tubes were measured with microsensors. Agar was prepared with 0 and 600 μmol/liter NaNO3, and profiles were determined 2 and 4 days after inoculation; profiles in uninoculated tubes that were the same age were also determined.
The microsensors were either purchased from Unisense A/S (Aarhus, Denmark) or manufactured at the Max Planck Institute for Marine Microbiology (Bremen, Germany). The O2 microsensors with a guard cathode (17) had tip diameters of 10 to 15 μm and 90% response times of <5 s. They were calibrated with air- and N2-flushed medium used for agar preparation (100 and 0% air saturation, respectively). The glass-type pH microsensors (18) had tip diameters of <12 μm and 90% response times of <20 s and were calibrated with commercial buffer solutions (pH 4.0, 7.0, and 9.2; Mettler-Toledo, Switzerland). The pH microsensors were used together with homemade reference electrodes, which consisted of a chlorinated Ag wire (length, 30 mm; diameter, 0.5 mm) that was inserted into one end of a glass capillary. The capillaries (length, 100 mm; inside diameter, 1 mm) were filled with 1% agar prepared in 3-mol/liter KCl and thus served as a salt bridge. The H2S microsensors (3) had tip diameters of 10 μm and 90% response times of <10 s. They were calibrated with deoxygenated PO4 buffer (200 mmol/liter K2HPO4/KH2PO4, pH 7.5) to which Na2S was added stepwise to obtain final concentrations of approximately 0 to 400 μmol/liter (9). The precise ΣH2S concentration of each calibration solution was determined spectrophotometrically by the method of Pachmeyer (16). The concentrations of free H2S in the calibration solutions were calculated as follows: $$mathtex$$\[[\mathrm{H}_{2}\mathrm{S}]{=}[{\sum}\mathrm{H}_{2}\mathrm{S}]/[1{+}(10^{\mathrm{pH}}/10^{\mathrm{pK}_{1}})]\]$$mathtex$$(1) where pK1 = 7.027 is the negative logarithm of K1, the first dissociation constant of the sulfide equilibrium system (pK2 can be neglected at pH <9). From these data, the calibration curve for the H2S microsensor was plotted. ΣH2S gradients in the tubes were calculated as follows: $$mathtex$$\[[{\sum}\mathrm{H}_{2}\mathrm{S}]{=}[\mathrm{H}_{2}\mathrm{S}]{\times}[1{+}(10^{\mathrm{pH}}/10^{\mathrm{pK}_{1}})]\]$$mathtex$$(2) using the [H2S] and the pH gradients measured with microsensors.
LIX-type NO3− microsensors (1) with tip diameters of 5 to 10 μm and 90% response times of <30 s were prepared on the day before use to improve the signal stability. NO3− microsensors were used together with homemade reference electrodes (see above). Calibration was performed using uninoculated gradient tubes in which the NaNO3 concentration was adjusted to 0, 15, 30, 60, 150, 300, or 600 μM. All sensors were calibrated before and after measurement at room temperature. One microsensor at a time was mounted on a motorized micromanipulator that was operated by the software Profix (Unisense A/S, Aarhus, Denmark). The microsensor was positioned in the center of the tube cross section and then lowered toward the agar surface (depth, 0 μm [see above]). Starting at this depth, vertical profiles were recorded at increments of 100, 200, or 400 μm down to 30 mm. The O2, pH, H2S, and NO3− profiles were determined at the same spot of the same tube whenever possible and were related to the position and thickness of the Beggiatoa mat in the inoculated enrichment culture (for mat position designations see above). The lower boundary of the mat was defined as the position where filaments were present more than just sporadically.
Flux calculations.The amounts of O2 and ΣH2S that flowed across a unit of area per unit of time (flux) were determined for uninoculated controls as well as for the tubes that were inoculated with the Beggiatoa enrichment. Assuming steady-state conditions, Fick's first law of diffusion was used: $$mathtex$$\[J{=}{-}D({\delta}C/{\delta}x)\]$$mathtex$$(3) where J is the flux (in nmol cm−2 s−1), D is the diffusion coefficient (in cm2 s−1), C is the concentration (in nmol cm−3), and x is the depth (in cm). The diffusion coefficients for O2 and ΣH2S (in agar at room temperature) were 2.03 × 10−5 and 1.57 × 10−5 cm2 s−1, respectively (13). For the uninoculated controls, the linear regions of the concentration gradients above and below the O2-ΣH2S overlap zone were used for δC/δx (13); for the Beggiatoa-containing gradient tubes, the linear regions above and below the Beggiatoa mat were used.
RESULTS
Mat position experiments.The experiments showed that the mat position depended on three factors: the concentrations of NO3− and ΣH2S and the length of incubation (Fig. 1). Generally, the mat position was deeper when the NO3− concentration was higher. This effect was less pronounced when 8 mmol/liter Na2S was used instead of 4 mmol/liter Na2S. In all treatments Beggiatoa mats moved upward with time (12). Three-way analysis of variance with NO3− and ΣH2S concentrations as between-subject factors and with time as a within-subject factor revealed that the dependence of the mat position on all three factors (for NO3−, F4,19 = 478 and P < 0.001; for ΣH2S, F1,19 = 529 and P < 0.001; and for time, F = 1,229, df = 5, and P < 0.001) was highly significant.
Mean depth (in mm) of the upper boundary of the Beggiatoa mat, depending on the NO3− and ΣH2S concentrations in the gradient tubes over time (days [d]). (A) Bottom agar prepared with 4 mmol/liter Na2S. (B) Bottom agar prepared with 8 mmol/liter Na2S. Symbols: ×, no NO3−; ⋄, 100 μmol/liter NO3−; ▵, 200 μmol/liter NO3−; ○, 400 μmol/liter NO3−; □, 600 μmol/liter NO3−. Some of the error bars, which indicate standard deviations (n = 3), are smaller than the symbols.
O2 and ΣH2S microgradients.Without NO3− addition, the vertical O2 and ΣH2S gradients were steeper in the Beggiatoa gradient tubes than they were in the uninoculated controls (Fig. 2A to D). Correspondingly, the O2 and ΣH2S fluxes into the Beggiatoa mats were greater than those into the O2-ΣH2S overlap zone (Table 2). Furthermore, the O2 and ΣH2S gradients became steeper with time, which resulted in upward movement of both the O2-ΣH2S overlap zone (uninoculated controls) and the Beggiatoa mat (Fig. 2A to D; cf. Fig. 1). The Beggiatoa mat in the experiment without added NO3− was approximately 0.4 mm thick and was slightly above the O2-ΣH2S overlap zone. NO3− addition to Beggiatoa tubes had a strong effect on the O2 and ΣH2S microgradients, on the mat position, and on the thickness of the mat, which increased to >8 mm (Fig. 2E and F). The NO3− effect was most pronounced 2 days after inoculation. An approximately 4-mm gap appeared between the O2 and ΣH2S profiles (Fig. 2E). Additionally, the corresponding O2 microgradient was considerably less steep, resulting in a flux of 3.6 pmol cm−2 s−1, which was only one-half the value obtained for the uninoculated control and less than one-fourth the value obtained for the treatment without NO3− (Table 2). In contrast, the ΣH2S flux was about twofold higher than that in the Beggiatoa gradient tube without NO3− and about fourfold higher than that in the uninoculated control (Table 2). The NO3− effect was less pronounced after 4 days; the O2 profile in the NO3−-containing Beggiatoa enrichment culture became steeper, and the ΣH2S profile became less steep (Fig. 2F).
Microprofiles of O2 (gray circles) and ΣH2S (black circles) and positions of the upper (dark gray lines) and, where applicable, lower (light gray lines) boundaries of the Beggiatoa mats. (A and B) Uninoculated gradient tubes. (C and D) Beggiatoa gradient tubes without NO3−. (E and F) Beggiatoa gradient tubes with an initial NO3− concentration of 600 μM. The incubation times were 2 days (A, C, and E) and 4 days (B, D, and F) after inoculation. The shaded areas within the boundaries of the Beggiatoa mats (E and F) indicate that filaments were more abundant in the upper mat regions. Gray and black circles overlap in some panels.
O2 and ΣH2S fluxes in uninoculated controls and in Beggiatoa-enriched gradient tubes without NO3− and with an initial NO3− concentration of 600 μmol/litera
NO3− microgradients.The NO3− microsensor measurements for the uninoculated control (Fig. 3A) and the Beggiatoa enrichment culture after 2 and 4 days (Fig. 3B and C) illustrate that the NO3− concentrations decreased in the presence of Beggiatoa sp. during incubation. The mean NO3− concentration in the upper 30-mm agar layer decreased from the initial concentration (600 μmol/liter) to 86 μmol/liter after 2 days and to 54 μmol/liter after 4 days. Furthermore, the profiles show that all of the NO3− diffused from the small upper agar volume into the mat, whereas some NO3− was still diffusing upward from the much larger volume of agar below the mat that also contained a larger total amount of NO3−. In contrast to O2 and ΣH2S, which were spatially separated after 2 days in the NO3−-containing treatment, NO3− and ΣH2S overlapped in the Beggiatoa mat (Fig. 2E and 3B).
Microprofiles of NO3− (circles) and positions of the upper (dark gray lines) and lower (light gray lines) boundaries of the Beggiatoa mats. (A) Uninoculated gradient tube. (B and C) Beggiatoa gradient tubes 2 days (B) and 4 days (C) after inoculation. The initial NO3− concentration was 600 μM. The shaded areas within the boundaries of the Beggiatoa mats (B and C) indicate that filaments were more abundant in the upper mat regions. Circles overlap in some panels.
pH microgradients.In the uninoculated control, the pH was 7.8 at the agar surface and increased to 8.3 at a depth of 30 mm due to the increasing ΣH2S concentration (Fig. 4A). In the Beggiatoa enrichment culture without NO3−, the pH profile showed that the minimum pH was close to the Beggiatoa mat (Fig. 4B). In contrast, in the Beggiatoa enrichment culture with NO3− the pH profile had a completely different shape and there was a pronounced maximum pH in the Beggiatoa mat (Fig. 4C).
Microprofiles of pH (circles) and positions of the upper (dark gray line) and, where applicable, lower (light gray line) boundaries of the Beggiatoa mats. (A) Uninoculated gradient tube. (B and C) Beggiatoa gradient tubes without NO3− (B) and with an initial NO3− concentration of 600 μM (C). The incubation time was 2 days. The shaded area within the boundaries of the Beggiatoa mat (C) indicates that filaments were more abundant in the upper mat regions. Circles overlap in some panels.
DISCUSSION
The hypothesis that the freshwater Beggiatoa strain investigated is able to oxidize ΣH2S anaerobically with the alternative electron acceptor NO3− originated from observations made during the mat position experiments; at higher NO3− concentrations the Beggiatoa mats moved deeper into the agar toward the electron donor ΣH2S (Fig. 1). This hypothesis was supported by microsensor profiles and flux calculations, which demonstrated that the Beggiatoa filaments indeed moved into anoxic, NO3−-rich agar layers and could oxidize even more ΣH2S if NO3 was available (Fig. 2C to F and Table 2). Furthermore, the O2 flux into the Beggiatoa mat exposed to NO3− was much lower than the O2 fluxes in the tubes without NO3− and the uninoculated control tubes after 2 days (Table 2). This finding can be explained by the missing O2-ΣH2S overlap zone in the NO3−-amended Beggiatoa tubes (Fig. 2E). Because of the spatial separation of O2 and ΣH2S, neither chemical nor biological ΣH2S oxidation with O2 could take place. The effect of the initial NO3− concentration on Beggiatoa sp. became less pronounced over time (Fig. 1 and 2C to F), which is explained by the finding that NO3− limitation occurred as incubation progressed (Fig. 3). It is likely that not all NO3− was immediately used for anaerobic ΣH2S oxidation and that an unknown fraction of NO3− was assimilated or stored intracellularly (11, 23). Vacuoles in freshwater Beggiatoa have not been detected so far (22), but cytoplasmic storage of NO3− is another possibility. This could explain the finding that more NO3− was taken up during the first 2 days of incubation than during the second 2 days (Fig. 3).
Beggiatoa oxidizes ΣH2S first to S0, which can be stored as intracellular globules, and subsequently to SO42− (22, 24). When O2 is used as the electron acceptor, the oxidation of H2S to S0 is pH neutral (if HS− is used as the electron donor, its oxidation to S0 is moderately alkaline; S2− can be neglected at pH <9), whereas the oxidation of S0 to SO42− is acidogenic. In total, the aerobic oxidation of ΣH2S to SO42− is acidogenic, which explains the pH profile found in the Beggiatoa enrichment culture without NO3−, in which the minimum pH largely coincided with the position of the Beggiatoa mat (Fig. 4B) (7, 13). When NO3− is used as the electron acceptor, the oxidation of ΣH2S to S0 increases the pH, while the oxidation of S0 to SO42− decreases the pH (20). This was visible in the pH profiles that were determined for the NO3−-containing treatments; after 2 days of incubation, the maximum pH was 8.7 in the lower region of the Beggiatoa mat (Fig. 4C), which must have resulted from the oxidation of ΣH2S to S0 with NO3−. Toward the upper region of the Beggiatoa mat, where less ΣH2S was available, the pH decreased. However, the pH in this layer did not decrease to values lower than those in the uninoculated control (Fig. 4A and C). Therefore, there was no indication that oxidation of S0 to SO42− took place in the upper region of the Beggiatoa mat. However, if oxidation of S0 to SO42− occurred at all, NO3− rather than O2 must have been used as the electron acceptor, because the O2 flux into the Beggiatoa mat was extremely low. The measured pH profiles are consistent with the results of a recent study of Sayama et al. (20), in which these authors found similar pH profiles in marine sediment colonized with Beggiatoa spp. It was hypothesized that the oxidation of H2S to S0 occurred with NO3− and was not necessarily spatially coupled to the oxidation of S0 to SO42−.
Furthermore, Sayama et al. (20) demonstrated that the marine Beggiatoa spp. investigated reduce NO3− to NH4+ under anoxic conditions (dissimilatory nitrate reduction to ammonium). This metabolic pathway was also hypothesized to occur in other marine sulfur bacteria (19) and is known to occur in large marine Thioploca spp. (15) that are close relatives of large marine Beggiatoa spp. (22). Another possibility for anaerobic ΣH2S oxidation with NO3− is denitrification, which was discussed by Sweerts at al. (21) for freshwater Beggiatoa spp. To date, this study is the only study in which anaerobic ΣH2S oxidation with NO3− was postulated for freshwater Beggiatoa spp., but questions about contamination of the Beggiatoa filaments with unicellular denitrifying bacteria have been raised by other authors (5, 11). The Beggiatoa enrichment culture used in our study also contained unicellular bacteria. Despite numerous trials, a pure culture could not be obtained, suggesting that this Beggiatoa strain is not able to grow without associated bacteria, which is a well-known phenomenon for other bacteria (8). However, the visibility of the Beggiatoa filaments in the transparent agar can be used. Using a stereomicroscope, it was observed that NO3− had an effect on the filaments because the Beggiatoa mat position and thus the chemotactic response of the filaments to O2 and ΣH2S were indeed changed. Alternatively, the movement of the Beggiatoa filaments may have resulted from an intimate association with unicellular NO3− reducers, which were directly responsible for the ΣH2S oxidation, and because of an absolute dependence of the Beggiatoa sp. on these reducers, the Beggiatoa sp. followed the movement of the NO3− reducers in the gradient tubes. However, this seems unlikely because in this case the Beggiatoa sp. would have had to disassociate from the energetically favorable electron acceptor O2. Hence, the changed chemotactic response of the Beggiatoa sp. strongly suggests that the freshwater Beggiatoa filaments themselves were chiefly responsible for the anaerobic ΣH2S oxidation with NO3−.
ACKNOWLEDGMENTS
L. P. Nielsen is gratefully acknowledged for providing the Beggiatoa sp. from his sewage outlet, as well as for fruitful discussions. A.-T. Henze and H. Plattner are thanked very much for valuable help. G. Eickert and M. Schubert provided technical support.
This study was funded by grant SCHU1416/2-1 from the Deutsche Forschungsgemeinschaft (German Research Foundation) and by the Max Planck Society, Germany.
FOOTNOTES
- Received 21 January 2006.
- Accepted 26 April 2006.
- Copyright © 2006 American Society for Microbiology
