Assessing the Impact of Denitrifier-Produced Nitric Oxide on Other Bacteria

Abstract

A series of experiments was undertaken to learn more about the impact on other bacteria of nitric oxide (NO) produced during denitrification. The denitrifier Rhodobacter sphaeroides 2.4.3 was chosen as a denitrifier for these experiments. To learn more about NO production by this bacterium, NO levels during denitrification were measured by using differential mass spectrometry. This revealed that NO levels produced during nitrate respiration by this bacterium were in the low μM range. This concentration of NO is higher than that previously measured in denitrifiers, including Achromobacter cycloclastes and Paracoccus denitrificans. Therefore, both 2.4.3 and A. cycloclastes were used in this work to compare the effects of various NO levels on nondenitrifying bacteria. By use of bacterial overlays, it was found that the NO generated by A. cycloclastes and 2.4.3 cells during denitrification inhibited the growth of both Bacillus subtilis and R. sphaeroides 2.4.1 but that R. sphaeroides 2.4.3 caused larger zones of inhibition in the overlays than A. cycloclastes. Both R. sphaeroides 2.4.3 and A. cycloclastes induced the expression of the NO stress response gene hmp in B. subtilis. Taken together, these results indicate that there is variability in the NO concentrations produced by denitrifiers, but, irrespective of the NO levels produced, microbes in the surrounding environment were responsive to the NO produced during denitrification.

During complete denitrification, nitrate is sequentially reduced to dinitrogen gas through the generation of the intermediates nitrite, nitric oxide (NO), and nitrous oxide (N₂O) (44). Each reductive step is catalyzed by a separate nitrogen oxide reductase. The concerted activity of the four nitrogen oxide reductases typically maintains a steady flow of intermediates (15, 44). It is important that bottlenecks be avoided, since some of these intermediates are reactive. The most reactive intermediate is NO, and several in vitro studies demonstrate that the denitrifiers Paracoccus denitrificans, Pseudomonas stutzeri, and Achromobacter cycloclastes maintain steady-state NO concentrations in the low nM range when respiring either nitrate or nitrite (14, 15). Studies with denitrifying wastewater and soil samples also measured NO concentrations at low nM levels (33, 37). This efficient reduction of NO is most likely due to the high substrate affinity and catalytic activity of the NO reductase expressed by denitrifiers (15, 17, 34, 44).

NO is a free radical that can potentially interact with an array of targets in both prokaryotic and eukaryotic cells (25). When exposed to NO, many organisms induce a set of genes whose products provide protection against the deleterious effects of NO (32). The variety of genes induced subsequent to NO exposure indicates that NO can impact a range of cell functions and that most organisms have specific mechanisms for mitigating NO toxicity. Even though NO is an obligate intermediate during denitrification, the impact of the NO produced during denitrification on other microbes has not been systematically studied. This is surprising, given that denitrifiers are widespread in the environment and that many denitrifiers form close associations with eukaryotes (1, 16, 44).

To learn more about the impact of the denitrification-derived NO on other organisms, Rhodobacter sphaeroides 2.4.3 was chosen as a representative denitrifier for a series of coculture experiments. Since NO levels during denitrification have not previously been measured for this bacterium, NO production by R. sphaeroides 2.4.3 was measured using differential mass spectrometry (DMS). The values measured for this organism during nitrate respiration were higher than those measured for most other denitrifiers. Therefore, A. cycloclastes was included in the coculture experiments for comparative purposes.

The abilities of various strains of 2.4.3 as well as of wild-type A. cycloclastes to inhibit growth of selected target bacteria were assessed using bacterial overlays, and a nitrogen oxide-dependent inhibition was observed. In a more direct test of how denitrifier-derived NO affects the NO stimulon of a nondenitrifier, the expression of the NO-inducible gene hmp in Bacillus subtilis was measured during exposure to denitrifying 2.4.3 and A. cycloclastes. Both denitrifiers were found to be very effective at inducing hmp expression. These experiments demonstrate that the NO produced during denitrification is toxic to organisms in the surrounding environment and that the NO stimulons of these bacteria are sensitive to the levels of NO produced during denitrification.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

R. sphaeroides 2.4.3 and 2.4.1 are the wild-type denitrifying and partially denitrifying strains, respectively. Strains 15.12 (20), 11.10 (20), and R125 (39) are Nor-deficient, Nir-deficient, and NnrR-deficient strains, respectively, of 2.4.3. Strain R438 is a 2.4.3 strain containing the plasmid pWLNIR, which is pRK415 containing the nitrite reductase structural gene nirK from 2.4.3 fused to the promoter of the rrnB gene resulting in the constitutive expression of Nir (22). AK13 is an NnrR-deficient strain of 2.4.1 (21). Other denitrifying strains used were A. cycloclastes (ATCC 21921), P. stutzeri (ATCC 14405), and P. denitrificans (ATCC 19367). Bacillus subtilis strain CU1065 (W168 trpC2 attSPβ) was used for bacterial overlays and mixed-culture experiments.

All of the Rhodobacter strains as well as P. stutzeri and P. denitrificans were grown in Sistrom's medium at 30°C (23). Antibiotics were added as previously described (38). A. cycloclastes was grown in Sistrom's medium containing 5.0 g liter⁻¹ tryptone and 2.5 g liter⁻¹ yeast extract at 30°C. Denitrifying R. sphaeroides and A. cycloclastes were grown in the presence of 12 mM nitrate as previously described (38). B. subtilis and Escherichia coli were grown in Luria-Bertani medium at 30°C. Cells to be used in DMS experiments were grown for approximately 18 h in 100 ml of culture medium. When harvested, the cultures contained, on average, 2.0 × 10⁹ cells ml⁻¹. The entire culture was then concentrated by centrifugation at 12,000 × g for 10 min, resuspended in 50 ml of Sistrom's medium, and recentrifuged. After centrifugation, cells were resuspended in Sistrom's medium to a cell density of 2.5 × 10¹⁰ cells ml⁻¹. Cells were placed on ice until use.

Differential mass spectrometry.

DMS is a variation of the differential electrochemical mass spectrometry technique (5, 41). The apparatus used in this work has been described in detail by Smith et al. and consists of a quadrupole mass spectrometer and a Channeltron electron multiplier/Faraday cup detector with a sensitivity of 100 A torr⁻¹ (35). The response time is typically in the millisecond range. The sample cell used in the DMS experiments consisted of a Teflon block (4.8 cm × 3.2 cm × 4.4 cm) with two small holes drilled on the sides where tubes were connected in order to keep the liquid level above 0.5 ml, so as to avoid large fluctuations in pressure. A closed, ultrahigh-vacuum flange with a hole drilled through the center, into which a stainless steel frit was inserted, served as the bottom of the cell. The frit supports a membrane (Scimat 200/40/60) which is made of an ethylene-tetrafluoroethylene copolymer and has a mean thickness of 60 μm, a mean pore size of 0.17 μm, and a porosity of 50% and separates the cell from the vacuum system. The analysis chamber pressure was approximately 8.0 × 10⁻⁵ torr for all experiments. The mass spectrometer scan rate was set at 1.4 s scan⁻¹ (2.33 min per 100 scans).

Standard curve experiments and quantification of results.

NO was generated from acidified nitrite using a variation of an existing protocol (15). Various concentrations of sodium nitrite were added to nitrogen-sparged, crimp-sealed serum vials containing 1 M sulfuric acid plus 12 mM KI. Nitrite was added to these vials with a gas-tight syringe to final concentrations of 10, 50, 100, 250, 750, and 1000 μM. The vials were incubated at 25°C for 10 min, and then 0.5 ml of each solution was added to the sample cell containing 0.5 ml of nitrogen-sparged water. Four separate samples at each NO concentration were assayed, each for 200 scans. The area under the curve from the time of injection to the time of flushing was calculated and plotted against NO quantity. A linear fit was then generated to yield a standard curve that was linear in the low to mid-μM range (Fig. 1). For determination of NO concentrations, it was assumed that similar amounts of NO were lost for experiments with equivalent scan times. Since the shapes of the curves were essentially the same both for cumulative assays (see below) and for NO generated by acidified sodium nitrite, the integrated area under a defined region of the curve can be used to estimate the NO concentration from the biological sample.

FIG. 1.

Relationship between integrated area of the m/z 30 peak and concentration of NO in the gaseous phase.

Cumulative DMS assays.

Assays of NO produced by cells incubated with nitrate in closed vessels for defined periods of time were termed cumulative assays. For these assays, 500 μl of cells was added to 500 μl Sistrom's medium in a 1.5-ml Eppendorf tube. Samples were then supplemented with 12 mM sodium nitrate and incubated at 30°C for 30 min if required. After incubation, the entire sample was pipetted into the sample cell of the DMS instrument. Sistrom's medium (0.5 ml) was present in the Teflon cell before sample addition, making the total assay volume 1.5 ml.

Bacterial overlays.

Cells of R. sphaeroides strains or of A. cycloclastes were spotted onto Sistrom's agar supplemented with 2 mM sodium nitrite and grown aerobically at 30°C for 2 days. Overlays consisted of 0.4% Sistrom's agar, pH 6.9, supplemented with 2 mM sodium nitrite and bacteria in a final volume of 10 ml. R. sphaeroides cells used in the overlays were grown aerobically in Sistrom's medium at 30°C, and 3 ml of a stationary-phase culture was used in the overlay suspension. B. subtilis cells used in overlays were grown aerobically in LB broth, and 100 μl of stationary-phase cells was used in the overlay suspension. Overlaid plates were incubated microaerobically at 30°C for 18 h.

RNA isolation and monitoring of hmp expression in mixed cultures.

For coculture experiments, 10 ml of an aerobically grown, stationary-phase B. subtilis culture was added to 100 ml of an overnight culture of either R. sphaeroides or A. cycloclastes grown under various conditions. The mixed cultures were then incubated with agitation at 30°C either aerobically or under oxygen-limited conditions. Control cultures consisted of 10 ml of a B. subtilis overnight culture incubated micro-oxically in 100 ml of nitrogen-sparged Sistrom's medium amended with 12 mM nitrate or 2.0 mM sodium nitroprusside (SNP). All cultures were set up in triplicate, and 10-ml aliquots were extracted and pooled before RNA extraction. Samples were taken from each coculture after 5 min.

Total RNA was extracted from bacterial samples as previously described (6). In order to normalize the amount of B. subtilis RNA in mixed-culture experiments, a slot blot prepared according to the manufacturer's instruction with a Bio-Dot SF slot blot apparatus (Bio-Rad) and a Zeta-Probe blotting membrane (Bio-Rad) was probed with the ³²P-radiolabeled B. subtilis-specific 16S rRNA oligonucleotide 5′-GTTCCCCAGTTTCCAATGACCC-3′ (8). Equivalent amounts of RNA from each extraction were determined based on 16S rRNA hybridization intensities and then blotted for hmp expression analysis. The hmp expression probe was generated via PCR with the oligonucleotides hmpF (5′- TACCTGAATTCAAGCAGG GC-3′) and hmpR (5′-GATACAAGCTTATGCATTGCCG-3′) and labeled as previously described (6). The slot blot was then hybridized and washed (6). The hmp probe and 16S rRNA oligonucleotide did not cross hybridize with R. sphaeroides RNA (data not shown).

RESULTS

Detection of denitrification products produced by R. sphaeroides 2.4.3.

To test if DMS could detect the denitrification products generated by R. sphaeroides 2.4.3, gas production by 2.4.3 suspensions was monitored using cumulative-assay DMS. Figure 2 shows a typical response curve. The first sharp feature that appears during the assay occurs at the time of sample injection (Fig. 2, beginning at scan 165). This feature represents the detection of gaseous denitrification products of m/z 30. The very sharp maximum is due to the convection created in the solution inside of the sample cell by the injection of the sample, which increases the transport rate to the membrane. A plot of the intensity versus t^−1/2 of this feature, where t is time, exhibited a linear dependence (r = 0.99965), as would be anticipated for a diffusion-controlled process (data not shown). The second sharp peak occurs during the injection of water to flush the sample solution out of the cell (Fig. 2, beginning at scan 275). Following the decay of this transient peak, the m/z 30 signal generally returned to baseline.

FIG. 2.

Characterization of R. sphaeroides 2.4.3 NO production by cumulative-assay DMS. Cells used in this assay were incubated for 30 min at 30°C in Sistrom's medium supplemented with 12 mM sodium nitrate. The output signal corresponds to the detection of ions of m/z 30. Arrows indicate addition of (i) 2.4.3 grown microaerobically with nitrate and (ii) cells grown aerobically. The DMS cell was flushed with water three times between additions of samples.

When supplemented with nitrate, denitrifying 2.4.3 samples produced a readily detectable signal at m/z 30 (Fig. 2). Aerobically grown cells of strain 2.4.3, which lack Nir activity, did not generate a signal at m/z 30 when incubated with nitrate. Furthermore, the 2.4.3 Nir-deficient strain 11.10 did not generate a peak at m/z 30 (data not shown). These results demonstrate that the expression of nirK is necessary for the production of the m/z 30 signal (Fig. 2).

Quantification of NO production by R. sphaeroides 2.4.3.

NO and N₂O generate overlapping signals at m/z 30, since 31% of the total N₂O fragments into an ion of m/z 30 (24). Thus, the m/z 30 signal shown in Fig. 2 does not represent NO alone. In order to estimate the amount of NO that contributes to the m/z 30 signal, peak areas were integrated for the m/z 12, 30, and 44 signals. Since CO₂ and N₂O have the same mass of 44, the area of the unique ion m/z 12, which represents 6% of the total CO₂, was used to calculate the contribution of CO₂ to the m/z 44 peak area as previously described (9). The estimated contribution of N₂O to the m/z 30 area was then subtracted from the total m/z 30 area to yield the contribution of NO to the m/z 30 signal (18). These calculations revealed that N₂O was the major component of the m/z 30 signal, which is somewhat surprising, since R. sphaeroides 2.4.3 has been shown to contain N₂O reductase (26).

A total of 22 separate cumulative assays from three independent cultures of R. sphaeroides 2.4.3 were analyzed by DMS, each for 200 scans. A 1-ml sample with approximately 2.5 × 10¹⁰ cells of strain 2.4.3 incubated with nitrate for 30 min was used for each assay. The average contribution of NO to the area of the m/z 30 peaks for these experiments was 0.396 ± 0.3 μA (Table 1). Using the standard curve generated by acid nitrite production of NO, it is estimated that this area corresponds to a concentration of NO in the aqueous phase ([NO_aq]) of 7.5 μM (Fig. 1 and Table 1). NO measured in individual samples ranged from about 1 to 10 μM. This level of NO accumulation is several orders of magnitude greater than that previously observed in studies with model denitrifiers (13-15, 34, 44).

TABLE 1.

Calculated NO contributions to m/z 30 from three independent cultures (A, B, and C)

Culturea	Replicatesb	Area	SD	[NOaq] (μM)c
A	7	9.28E-8	3.37E-8	2.7
B	8	3.07E-7	9.98E-8	5.3
C	7	7.97E-7	7.98E-8	15.0

^aA, B, and C were three separate denitrifying cultures of 2.4.3.

^bIdentical replicates from each culture.

^cCalculated from the standard curve (Fig. 1) and the ratio [NO_g]/[NO_aq] = 20 (13), where [NO_g] is the concentration of NO in the gaseous phase. Final [NO]_aq values were obtained by multiplying by 1.5 to adjust for volume differences.

NO production by the denitrifying bacteria P. stutzeri, P. denitrificans, and A. cycloclastes was also assayed by cumulative DMS. m/z 30 peak areas from cumulative assays with denitrifying P. stutzeri, P. denitrificans, and A. cycloclastes ranged from 1.32 nA to 16.4 nA (data not shown). For these denitrifiers, the m/z 30 signal yielded areas that were below the limit of detection of the standard curve, which was approximately 500 nM. This result is consistent with previous measurements (15).

Growth inhibition by NO produced during denitrification.

To assess the impact on other bacteria of the NO produced during denitrification, bacterial overlays were performed to compare the abilities of denitrifying A. cycloclastes and R. sphaeroides 2.4.3 to inhibit growth of bacteria that lack an NO-producing Nir (Fig. 3). A similar approach has successfully been used to compare the cyanide production levels of various pseudomonads (2). The bacterial strains used in the overlays were R. sphaeroides 2.4.1, which contains Nor but lacks Nir, and the nondenitrifier B. subtilis CU1065. R. sphaeroides 2.4.1 was used because it has been previously shown to be more sensitive to NO derivatives than 2.4.3 (42). B. subtilis was chosen because it a nondenitrifying soil microbe whose NO stimulon has been described previously (28). Concordant with their levels of NO production, 2.4.3 inhibited the growth of the overlaid bacteria more than A. cycloclastes. For example, the inhibition distance of B. subtilis CU1065 was sevenfold greater when grown over 2.4.3 than when grown over A. cycloclastes (Fig. 3). R. sphaeroides 2.4.1 was the least affected by R. sphaeroides 2.4.3 colonies. In contrast, the 2.4.1 strain AK13, which lacks Nor activity due to inactivation of the gene encoding NnrR (21), was more susceptible to denitrifying 2.4.3 colonies than 2.4.1 was, providing evidence that NO is involved in inhibiting growth and that Nor is responsible for the increased resistance observed in 2.4.1 (Fig. 3). There is some variability in the results of these experiments that is likely due to the difficulty of standardizing the number of cells and colony size of the NO producer between experiments.

FIG. 3.

Antimicrobial activity of various R. sphaeroides strains and A. cycloclastes. Values in mm indicate the distance between the outer edge of the NO-producing colony and the point at which visible growth was observed in the bacteria used in the overlay. NO-producing strains: 2.4.3, R. sphaeroides 2.4.3; 11.10, R. sphaeroides 2.4.3 Nir-deficient mutant; 15.12, R. sphaeroides 2.4.3 Nor-deficient strain; R438, R. sphaeroides 2.4.3 Nir-constitutive strain; Ac, A. cycloclastes. Overlaid recipient strains: white bars, R. sphaeroides 2.4.1; gray bars, R. sphaeroides 2.4.1 Nor-deficient strain AK13; hatched bars, B. subtilis strain CU1065. The results shown are averages of four individual experiments. Error bars represent 1 standard deviation.

The size of the zones of growth inhibition could be altered by using strains in which the copy number of nirK or norB was changed either by deletions or by inclusion of plasmids to increase gene copy number (Fig. 3). The Nor-deficient mutant strain 15.12 significantly increased the zone of inhibition, since NO accumulation is not abated by Nor in this strain. R438, in which nirK copy number and transcription levels have been increased (22), also caused greater growth inhibition than 2.4.3. As expected, the Nir-deficient mutant strain 11.10 failed to cause any zones of growth inhibition due to its inability to produce NO.

Induction of hmp expression in B. subtilis by denitrifying R. sphaeroides 2.4.3 and A. cycloclastes.

As a further test of the capacity of denitrification-derived NO to affect nondenitrifiers, the expression of hmp in B. subtilis was monitored during coincubations with denitrifiers. hmp is a key component of the NO stress response in several bacteria whose transcription has been shown to be induced by NO and nitrosative stress (11, 12, 28, 29, 36). hmp transcription was monitored during the coincubation of B. subtilis CU1065 with either R. sphaeroides 2.4.3 or A. cycloclastes under denitrifying conditions by RNA slot blots (Fig. 4). Denitrifying 2.4.3 induced hmp transcription at levels four- to fivefold higher than those seen when CU1065 was coincubated with 2.4.3 under nondenitrifying conditions or when CU1065 was incubated with nitrate alone (Fig. 4, lane 1 versus lanes 3, 4, and 5). The effectiveness of denitrification-derived NO in activating hmp expression was further demonstrated by the observation that coincubation with denitrifying 2.4.3 increased hmp transcription 3.5-fold above that of cells incubated with the NO-generating agent SNP (Fig. 4, lanes 1 and 6). Coincubation with denitrifying A. cycloclastes also induced high levels of hmp expression (Fig. 4, lane 2). Induction of hmp expression during coincubation with either denitrifier was rapid, with an increase in expression observed within 2 min (data not shown).

FIG. 4.

Effect of NO produced by denitrification on expression of hmp in B. subtilis. Slots were loaded either with 0.25 or 0.125 μg of total RNA from cocultures or with pure culture of B. subtilis incubated for 5 min under various conditions. The blot was hybridized with a ³²P-labeled hmp probe. Nitrate and SNP were added to final concentrations of 12 mM and 2 mM, respectively. RNA samples were extracted from the following: (1) coculture of B. subtilis + R. sphaeroides + nitrate under limiting oxygen, (2) coculture of B. subtilis + A. cycloclastes + nitrate under limiting oxygen, (3) coculture of B. subtilis + R. sphaeroides under aerobic conditions, (4) B. subtilis incubated aerobically, (5) B. subtilis + nitrate under limiting oxygen, and (6) B. subtilis + SNP under limiting oxygen.

DISCUSSION

In previous studies, NO levels produced in 23 to 40 min by P. stutzeri, P. denitrificans, and A. cycloclastes ranged from 0.5 to 4.0 nM in a reaction using 4.0 × 10⁹ cells supplemented with 8.0 mM nitrate (15). In this study, 2.5 × 10¹⁰ R. sphaeroides 2.4.3 cells generated a concentration of 7.5 μM NO in 30 min in samples supplemented with 12 mM nitrate. Similar levels of NO were measured during nitrate respiration by an unidentified isolate; however, no additional work has been done with this bacterium (10). Taken together, these results indicate that the level of NO produced by denitrifiers is variable and not necessarily constrained to the low nM levels measured in previous studies.

The cause of the relatively high concentrations of NO produced by 2.4.3 during denitrification is unknown. Previous studies with this bacterium indicate that Nor is highly transcribed and active during denitrification (3). Nir and Nor from 2.4.3 appear to have activities and active sites that are similar to those of other denitrifiers (27, 30). Nevertheless, the level of NO accumulation during denitrification suggests that the relative balance of the activities of Nir and Nor in 2.4.3 is different from the balance of these two activities in previously characterized denitrifiers. This high level of NO produced by R. sphaeroides most likely carries some cost to the cell, which may explain why most R. sphaeroides strains have lost Nir but retain Nor as a potential protection mechanism against exogenous NO (21).

Since the immune response of animals and plants typically produces NO as a broad-spectrum antibiotic, studies assessing the impact of exogenous NO on growth and survival are most frequently carried out on pathogenic bacteria. Not surprisingly, most pathogens can effectively mitigate NO toxicity (32). However, many saprophytic bacteria, such as B. subtilis, have large NO stimulons that are induced in the presence of NO-generating compounds and encode proteins such as flavohemoglobin, whose only known function is to reduce NO to less-reactive compounds (28, 29). It has been difficult to explain the presence of such a response system in a bacterial species that is unlikely to ever encounter NO produced by the immune response. The experiments described here demonstrate that denitrifiers are a likely source of the NO that a bacterium such as B. subtilis might encounter in its natural environment. The finding that coincubation with A. cycloclastes was as effective at inducing hmp expression as that with 2.4.3 suggests that NO-inducible stimulons respond to the nM concentrations of NO typically found in the environment (33, 37). In addition, this result also demonstrates that biologically produced NO is as good or better at inducing expression of genes in the NO stimulon than the NO-generating compounds used in most studies (Fig. 4). This supports the view that the NO encountered by B. subtilis in its natural environment is produced by processes such as denitrification.

The experiments described in this work demonstrate that NO levels produced during denitrification vary between denitrifiers. Different levels of NO production likely correlate with factors that are unique to each denitrifier's particular environmental niche. Once produced, this NO is freely diffusible and will impact adjacent bacteria. NO is an effective inhibitor of respiration and will bind to the metalloenzymes found in electron transport chains at low nM concentrations (40). Thus, its presence may influence the metabolic activity of a microbial community under favorable conditions. NO may play a role in the denitrification-dependent inhibition of sulfate reduction and other anaerobic processes that has been previously observed (19, 31). While denitrifiers are likely important sources of NO in the environment, other groups of bacteria, including nitrifiers and enteric bacteria, can produce NO (4, 7). Additional work needs to be done to more critically assess the impact of NO produced by these different processes on microbial communities.