Transition from Anaerobic to Aerobic Growth Conditions for the Sulfate-Reducing Bacterium Desulfovibrio oxyclinae Results in Flocculation

Continuous culture experiment. D. oxyclinae was initially grown in the growth vessel of the chemostat in an anaerobic batch culture mode under nitrogen gassing until sulfate was completely depleted from the medium. At this point, the cell concentration reached 6.7 × 10⁸ ml⁻¹ and the protein level was 33.8 mg liter⁻¹. Growth yield of the batch culture was 3.6 g of protein per mol of sulfate consumed. Medium flow was then set at a dilution rate of 0.05 h⁻¹. The anaerobic chemostat was maintained for 436 h under gassing with nitrogen. At this point, 1% oxygen was added to the incoming gas. Activities ofD. oxyclinae under anaerobic conditions and after exposure to oxygen are presented in Fig. 1; average values for the different stages are given in Table1. The sequence of activities of this organism during anaerobic growth and in the transition from anoxic to oxic conditions can be divided into four stages according to the calculated apparent efficiency of sulfate reduction and the overall appearance of the culture. These stages differ also in the calculated oxygen consumption and measured rates of oxygen uptake and thiosulfate reduction.

• Open in new tab
• Download powerpoint

Fig. 1.

Growth parameters of D. oxyclinae in anaerobic chemostat culture and under 1% oxygen in the incoming gas grown on 20 mM lactate. (Upper) Protein concentration (■), cell numbers (□), dissolved oxygen (DO; ○, solid line); (Lower) concentrations of sulfide (□), sulfate (▴), and dissolved oxygen (○). Arrowheads indicate initiation of oxygen supply.

MPN counts demonstrated that D. oxyclinae persisted in the growth vessel throughout the entire experiment at the growth rate of 0.05 h⁻¹. The culture was axenic during all the stages of the experiment. Tests for aerobic contamination performed at each sampling point were all negative.

During the initial 122 h of the anaerobic growth stage, sulfate concentration decreased from 10 mM to the residual concentration of 0.2 mM and remained stable during the next 314 h of steady-state chemostat culture (stage I [Fig. 1]). The concentration of dissolved sulfide was low as a result of intense nitrogen gassing at 0.5 1 min⁻¹. Lower gassing rates were found to result in suppression of sulfate reduction, presumably due to inhibition by sulfide (data not shown). During stage I, protein concentration and cell numbers also remained stable. At this stage, 2.4 ± 0.1 g of protein was produced per mol of sulfate consumed in the chemostat, corresponding to the calculated sulfate reduction rate of (2.0 ± 0.2) × 10⁻¹¹ μmol of sulfate cell⁻¹min⁻¹. The rates of thiosulfate reduction and of aerobic respiration measured in washed cell suspensions in the respiration chamber were 1.2 × 10⁻¹¹ and 1.9 × 10⁻¹¹ μmol min⁻¹ cell⁻¹, respectively (Fig. 1; Table 1). No cell aggregates were observed.

When at the end of the steady-state anaerobic stage the chemostat was bubbled with 1% O₂ in the incoming gas mixture, D. oxyclinae was not washed out. During the first 42-h transition phase, stage II, the calculated sulfate reduction rate decreased by 20%, the residual sulfate concentration increased from 0.2 to 4.6 mM, and the sulfide concentration decreased twofold. Cell numbers nevertheless remained stable (Fig. 1, upper panel), and the culture remained macroscopically homogeneous. At this oxygen flux of 45 μmol min⁻¹, most of the oxygen was utilized; only a residual oxygen concentration of 0.7 ± 0.2 μM was detected in the growth vessel. The calculated growth yield increased to 3.4 ± 0.8 g of protein per mol of sulfate consumed. This increase may indicate the use of intermediate sulfur compounds at a considerably lowerK_m than that of sulfate (32a).

A marked increase in sulfate reduction indicated the onset of stage III, which continued for the following 97 h. Elevated sulfate reduction may be a result of the utilization of the intermediate sulfur electron acceptors generated during stage II, probably by chemical oxidation of sulfide. Residual sulfate concentration decreased sharply and later stabilized at 0.6 mM (Fig. 1, lower panel). Although this was evidence of sulfate reduction activity restored to a level similar to that found at stage I, residual sulfide concentration did not change. Protein concentration increased steadily from 23.8 to 58.3 mg liter⁻¹, and MPN counts also increased. The yield increased further to 4.4 ± 1.2 g of protein per mol of sulfate consumed. The calculated in situ rate of oxygen consumption in the chemostat increased threefold, to (9.7 ± 1.2) × 10⁻¹¹ μmol of O₂ cell⁻¹min⁻¹ compared to stage II, while the rate of sulfate reduction did not change significantly, remaining at (1.7 ± 0.5) × 10⁻¹¹ μmol cell⁻¹min⁻¹. The rates of both oxygen consumption and thiosulfate reduction measured on washed cell suspensions in the respiration chamber increased 1.8-fold compared to stage I, to 3.4 × 10⁻¹¹ and 2.2 × 10⁻¹¹ μmol cell⁻¹ min⁻¹, respectively (Table 1). The culture remained macroscopically homogeneous throughout this stage.

The onset of stage IV was indicated by formation of cell aggregates of up to 3 mm, with no marked wall growth. This heterogeneous culture was no longer a true chemostat, and therefore repetitive sampling became variable. As a result, the protein concentration ranged from 29 to 98.2 mg liter⁻¹. This nonhomogeneous culture was maintained for 167 h (Fig. 1). Compared to the previous stage, the sulfate reduction rate did not change significantly and remained at (1.4 ± 0.5) × 10⁻¹¹ μmol cell⁻¹min⁻¹. The average protein concentration was about twice as high as in the anaerobic growth mode, and both cell counts and MPN increased substantially. The calculated apparent growth yield of 6.5 ± 2.3 mg of protein per mol of sulfate consumed was the highest ever recorded for D. oxyclinae. The calculated in situ rate of oxygen consumption in the chemostat decreased to (6.0 ± 2.3) × 10⁻¹¹ μmol of O₂cell⁻¹ min⁻¹ compared to stage III, while the rate of sulfate reduction did not change significantly, remaining at (1.4 ± 0.5) × 10⁻¹¹ μmol cell⁻¹min⁻¹ (Table 1). The rate of oxygen consumption measured on washed cell suspensions in the respiration chamber decreased by 50%, to 1.7 × 10⁻¹¹ μmol O₂cell⁻¹ min⁻¹, while the rate of thiosulfate reduction increased drastically (2.9-fold) compared to stage III, to 6.3 × 10⁻¹¹ μmol cell⁻¹min⁻¹ (Table 1). The concentration of residual dissolved oxygen throughout this phase of the experiment fluctuated in the range of 0.1 to 1.7 μM.

The high rate of gas flow applied in this experiment made quantitative estimations of sulfide production impossible. However, the ratio of residual dissolved sulfide per millimole of sulfate reduced may be used to estimate the portion of sulfide not removed by gassing. In the anaerobic coculture of D. oxyclinae withMarinobacter sp. strain MB this ratio was 43.5 μmol of sulfide per mmol of sulfate reduced (29). As the result of increased gassing rate in axenic culture of D. oxyclinaeunder anaerobic continuous (stage I) this ratio was much lower, 13.6 ± 2.1 μmol of sulfide per mmol of sulfate reduced. During stage II the ratio decreased by 14%, probably due mostly to chemical oxidation of sulfide by oxygen. During stage III it decreased further, to 7.4 ± 1.2 μmol of sulfide per mmol of sulfate reduced. Since the rate of oxygen supply did not change, this decrease can be attributed to biological sulfide oxidation. The subsequent increase of this ratio to 10.6 ± 1.5 μmol of sulfide per mmol of sulfate reduced during stage IV is probably the result of intensified sulfate reduction inside bacterial aggregates.

In this range of oxygen concentrations, the rate of aerobic respiration previously reported for D. oxyclinae was 260 nmol of O₂ min⁻¹ mg of protein⁻¹(17). In the present experiment, the potential respiration rates measured in the respiration chamber using washed cell suspensions during the anoxic stage were somewhat higher (1.9 × 10⁻¹¹ μmol of O₂ min⁻¹cell⁻¹, which corresponds to 360 nmol of O₂min⁻¹ mg of protein⁻¹), and those obtained during stage III were considerably higher, 3.4 × 10⁻¹¹ μmol of O₂ min⁻¹cell⁻¹, corresponding to 620 to 700 nmol of O₂min⁻¹ mg of protein⁻¹ (Table 1). The rate of oxygen consumption by aerobic respiration at stage III was therefore calculated to be (9.7 ± 1.2) × 10⁻¹¹ μmol of O₂ min⁻¹ cell⁻¹. During stage IV the measured respiration rate was lower, 1.7 × 10⁻¹¹μmol of O₂ min⁻¹ cell⁻¹, and the calculated rate of oxygen consumption was (6.0 ± 2.3) × 10⁻¹¹ μmol of O₂ min⁻¹cell⁻¹. In the transition between stages III and IV, the measured respiration rate decreased 2-fold whereas the calculated respiration rate decreased only 1.6-fold. This discrepancy may be caused by flocculation. During stage IV, only a fraction of the cells in the growth vessel were in contact with oxygen, resulting in lower calculated respiration rates per cell (Table 1). Both values were well below the oxygen flux of 45 μmol of O₂ min⁻¹applied in this experiment.

The measured values of respiration rates were always less than the calculated ones. This may be partly the result of the shock received by the cells during washing and resuspension. On the other hand, some portion of oxygen in the growth vessel was bound chemically due to sulfide oxidation.

The yield of 2.4 g of protein per mol of sulfate obtained in the anaerobic continuous culture of D. oxyclinae was relatively low. Higher yields of 3.4 g of protein per mol of sulfate consumed were previously reported for this organism in an anaerobic continuous culture (17). The same yield was observed in this study during the batch culture stage. Much higher yields of 9.9 ± 1.6 g (dry weight) per mol of lactate (28) and 5.3 g of protein per mol of lactate (33) were reported forD. desulfuricans grown in batch culture on lactate and sulfate. These values correspond to approximately 10 g of protein per mol of sulfate. The yields reported for otherDesulfovibrio species grown on lactate and sulfate vary from 3.7 to 7.8 g of protein per mol of sulfate (calculated using data from reference 14).

During stages II and III, the apparent yield of protein per sulfate consumed was greater than in the anaerobic mode of growth. Higher protein concentrations in the presence of oxygen implied lower calculated sulfate reduction rates per milligram of protein. Indeed, the calculated rates of sulfate reduction decreased from (2.0 ± 0.2) × 10⁻¹¹ μmol cell⁻¹min⁻¹ under anaerobic conditions (stage I) to (1.7 ± 0.5) × 10⁻¹¹ μmol cell⁻¹min⁻¹ at stage III. In contrast, the rate of thiosulfate reduction as measured in washed cell suspensions in the four-electrode chamber increased during the same period from 1.2 × 10⁻¹¹ μmol cell⁻¹ min⁻¹ at stage I to 2.2 × 10⁻¹¹ μmol cell⁻¹min⁻¹ at stage III.

D. oxyclinae has a low affinity for sulfate. Sulfate reduction occurred only when sulfate concentrations were over 30 μM, and the rate of sulfate reduction at this low sulfate concentration was 20-fold less than at 30 mM (17). This does not hamper growth of D. oxyclinae in its natural, sulfate-rich environment where sulfate levels of 90 to 120 mM prevail (16). Thiosulfate, on the other hand, was reduced at high rates even at 3 μM. Moreover, unlike sulfate reduction, thiosulfate reduction was found to proceed even in the presence of oxygen, provided the oxygen concentration was below 2 μM. It was previously suggested that thiosulfate was an important natural substrate for D. oxyclinae in environments where both sulfate reduction and sulfide oxidation occur (17). It is postulated that in the presence of oxygen, the initial steps of sulfate reduction, namely, formation of adenosine sulfate (APS) and reduction to bisulfite, are impaired to a greater degree than the consequent steps (30).

Although D. oxyclinae was demonstrated to oxidize sulfide to sulfate (17), we may speculate that under the experimental conditions when sulfide was in excess, some of sulfide was oxidized incompletely only to thiosulfate or to other intermediate sulfur compounds. This process was expected to result in the formation of free thiosulfate and possibly other sulfur intermediates in the beginning of the transition period upon the initial introduction of oxygen to the anaerobic steady state (stage II). Thiosulfate was indeed found in the growth vessel during stage II at concentrations of about 0.1 mM and was not detected in the growth vessel at later stages. During stages III and IV, thiosulfate was below the detection limit, probably as a result of its rapid recycling.

While the growth rate remained constant, protein content per cell increased. Protein content per cell at stage III was 1.4-fold higher than during the anoxic stage. The enhanced yield can be due to the ability of D. oxyclinae to oxidize sulfide aerobically. This process is most efficient when oxygen concentrations are below 30 μM (17). The combination of incomplete oxidation of sulfide to thiosulfate and subsequent thiosulfate reduction may well contribute to the observed increase in yields.

Cell aggregation.Formation of cell aggregates by D. oxyclinae under oxygen supply, as well as profuse wall growth was reported previously (17). In the course of this experiment, wall growth was less pronounced and macroscopic cell aggregates ofD. oxyclinae were observed only after 140 h of exposure to oxygen. Smaller microscopic aggregates were formed earlier, although these micrometer-sized aggregates were easily disrupted during sampling and dilution, as evidenced by the good correlation between cell counts and MPN values.

To determine the initial stage of flocculation, the effect of oxygen on turbidity of the cell suspension was studied. Under anaerobic conditions the culture was homogeneously suspended as single cells, as evidenced by the stable turbidimetric readings. When a single pulse of 87.3 μM oxygen was introduced, initial flocculation was observed after 45 min, as evidenced by increased turbidity fluctuations (Fig.2a). Continuous input of 5% oxygen resulted in more rapid aggregation of D. oxyclinae, but at a smaller range of turbidity fluctuations (Fig. 2b).

• Open in new tab
• Download powerpoint

Fig. 2.

Initial cell aggregation in a response to exposure to oxygen, based on changes in turbidity in cell suspension of D. oxyclinae after addition of oxygen to an anaerobically grown culture. Introduction of oxygen is marked by arrowheads. (a) Response to a single pulse of 87.3 μM oxygen; (b) response to continuous input of 5% oxygen. AU, arbitrary units.

In an attempt to estimate the size of the initial flakes, shaken batch cultures of D. oxyclinae were incubated in sealed bottles for 24 h with 2% O₂ in the gas phase and compared to an anaerobic control. Protein content of the culture and of the fraction filtered through 8-μm-pore-size Nuclepore filters was determined. During the first 10 h of incubation, the fractions of cells in aggregates larger than 8 μm in the presence of oxygen and in the anoxic control were the same. However, after 24 h of incubation with 2% oxygen, there were 19% more aggregates than in the anoxic control (data not shown).

It was previously demonstrated that oxygen was a preferred electron acceptor for anaerobically grown cells of D. oxyclinae. In the respiration chamber, sulfate reduction stopped immediately after the introduction of oxygen and did not resume until all oxygen was reduced (17). In the chemostat culture, however, D. oxyclinae continued to reduce sulfate in the presence of dissolved oxygen. One possible explanation to this apparent discrepancy is the need of new enzymes for sulfate reduction in the presence of oxygen. Only when these enzymes are induced in the chemostat can “aerobic” sulfate reduction be demonstrated. In the case of flocculated culture, however, only the outer layers of cell aggregates are exposed to oxygen. Depth of oxygen diffusion through the interface when the rate of oxygen consumption is constant is described by the equation X = (2C₀D/A)^1/2, where A is the rate of oxygen consumption per unit volume per unit time, C₀ is the concentration of oxygen at the biofilm surface and D is the diffusion coefficient of oxygen (4). Considering the high stirring rate, we can substitute for C₀ the ambient oxygen concentration in solution. Assuming D = 3 × 10⁻⁵cm² s⁻¹ (A. Wieland, D. van Dusschoten, L. R. Damgaard, D. de Beer, M. Kühl, and H. Van As, unpublished data), C₀ = 1 μM, a respiration rate of 260 nmol mg⁻¹ min⁻¹(17), and cell density in the aggregates of 1 μm⁻³, it is possible to calculate that oxygen penetrated only to the depth of 1.5 μm into the aggregates. Hence, in aggregates of 3 μm and larger, the internal cells could perform sulfate reduction in an anaerobic microenvironment.

Aggregation of SRB in a cyanobacterial mat.Flocculation may be one of the mechanisms enabling some SRB to survive under periodic exposure to oxic conditions in their natural habitats. In situ hybridization revealed numerous clusters of SRB in the oxic zone of the Solar Lake cyanobacterial mats. A representative sample taken from 500 μm below the surface, where oxygen concentrations of up to 1.5 mM were detected by oxygen microelectrode at noontime, showed both clustered and individual SRB cells among the filaments of the cyanobacterium Microcoleus chthonoplastes (Fig.3). Similar results for aggregates of SRB were obtained when fluorescein-labeled probe 687, targeted to desulfovibrios, was used (21). Diurnal measurements of sulfate reduction activity in a Solar Lake cyanobacterial mat showed the highest activity at the O₂-H₂S chemocline at noontime (11). These measurements underestimated sulfate reduction in the presence of oxygen because of rapid sulfide reoxidation. Two-dimensional mapping of sulfate reduction by the Na₂ ³⁵SO₄-coated silver foil technique allows the detection of sulfate reduction even in the presence of high oxygen concentrations because of the very low solubility of Ag₂ ³⁵S (Y. Helman, unpublished data). Figure 4 shows an example of phosphor imaging of the two-dimensional pattern of sulfate reduction obtained by the silver foil technique correlated to profiles of oxygen, oxygenic photosynthesis rate, sulfide, and pH obtained by specific microelectrodes. Because of changes in specific activity of the radiolabeled sulfate during the experiment, these results cannot be translated to quantitative rates of sulfate reduction; nevertheless, they clearly show the heterogeneity of sulfate reduction activity under oxic conditions. This qualitative mapping shows that active sulfate reduction activity is detected well above the chemocline. Most of this activity is associated with several heterogeneously distributed distinct patches. The combined observations of SRB clumps by in situ hybridization and the heterogeneity of sulfate reduction mapping in the oxic zone of a cyanobacterial mat indicate the existence of anoxic microniches with active SRB in this environment.

• Open in new tab
• Download powerpoint

Fig. 3.

Confocal laser scanning in situ hybridization with fluorescein-labeled probe SRB385 (green) and autofluorescence of cyanobacteria (purple) in a horizontal transect of a Solar Lake cyanobacterial mat at a depth of 500 μm. (A) Clump of unicellular SRB; (B) filamentous Desulfonema-like SRB. Note additional scattered unicellular SRB.

• Open in new tab
• Download powerpoint

Fig. 4.

Phosphor imaging of a two-dimensional pattern of sulfate reduction obtained in a Solar Lake cyanobacterial mat by the silver foil technique, correlated to profiles of oxygen (○), sulfide (●), pH (•), and oxygenic photosynthesis rate (bars) as obtained by specific microelectrodes. The horizontal line indicates area of O₂ and H₂S overlap.

Concluding remarks.Oxygen flux of 45 μmol min⁻¹ and oxygen uptake of 9.4 ± 0.5 μmol min⁻¹ resulted in low concentrations of dissolved oxygen up to 10 μM, as was the case in the chemostat experiment (Table 1). Under these conditions, anoxia can develop in the center of a 10-μm aggregate of D. oxyclinae. In aerated environments, anoxia was reported to develop inside the millimeter-sized marine fecal pellets (1), in activated sludge (27), and in nitrifying bacterial aggregates 100 μm below the surface (10), i.e., in relatively large aggregates containing bacteria with high respiration rates at high concentrations of oxygen.

The two SRB aggregates shown in Fig. 3 are small, 10 and 30 to 40 μm. SRB aggregates of similar size were found in the upper oxic layers of wastewater biofilms under aeration (24). Even assuming high respiration rates, these aggregates are too small to become anoxic under 1.5 mM dissolved oxygen as is the case in cyanobacterial mats. Moreover, D. oxyclinae has a high respiration rate of 260 nmol mg⁻¹ min⁻¹ only when oxygen concentrations are not higher than 30 μM (17). Anoxic microniches therefore could not be formed by D. oxyclinaealone.

In its natural environment, D. oxyclinae occurs in close association with oxygen-scavenging aerobes. Such consortia were repeatedly isolated from Solar Lake mats (29, 31). One of these organisms, Marinobacter sp. strain MB, is an aerobe growing under high levels of aeration. A chemostat coculture ofD. oxyclinae and Marinobacter sp. strain MB was growing and reducing sulfate under various levels of oxygen input up to gassing with air. Unlike the axenic culture of D. oxyclinae, macroscopic cell aggregates were not observed in this coculture (29). It can therefore be suggested that SRB in oxic environments occur mostly in the form of relatively small aggregates surrounded with a layer of oxygen-scavenging aerobes. The actual interaction between D. oxyclinae and Marinobactersp. strain MB in mat environments is presently being investigated.

Cell aggregates of tens of micrometers cannot be detected by oxygen microelectrodes as anoxic microniches. Likewise, such environments were not found in a study of two-dimensional O₂ distribution using planar optodes with spatial resolution of 130 μm, although spatial heterogeneity of oxygen concentrations was reported, yet patches of high oxygen net consumption were found to be surrounded by zones of oxygen net production (13). These can be interpreted as clusters of oxygen-scavenging bacteria surrounded by cyanobacterial matrix similar to those shown in Fig. 3. The size of anoxic zones in the upper layer of a cyanobacterial mat must therefore be less than 130 μm.

This is the first report of growth of an SRB under exposure to oxygen. Both sulfate reduction and oxygen respiration occurred in the homogeneous axenic culture, followed by later clumping of the cells. Clumping may be a result of the lack of mechanisms to deal with oxygen radicals in D. oxyclinae. Active sulfate reduction by clumps of SRB even in the presence of very high concentrations of dissolved oxygen was demonstrated within the oxic zone of a cyanobacterial mat community.