Nitrite Control over Dissimilatory Nitrate/Nitrite Reduction Pathways in Shewanella loihica Strain PV-4

Abstract

Shewanella loihica strain PV-4 harbors both a functional denitrification (NO₃⁻→N₂) and a respiratory ammonification (NO₃⁻→NH₄⁺) pathway. Batch and chemostat experiments revealed that NO₂⁻ affects pathway selection and the formation of reduced products. Strain PV-4 cells grown with NO₂⁻ as the sole electron acceptor produced exclusively NH₄⁺. With NO₃⁻ as the electron acceptor, denitrification predominated and N₂O accounted for ∼90% of reduced products in the presence of acetylene. Chemostat experiments demonstrated that the NO₂⁻:NO₃⁻ ratio affected the distribution of reduced products, and respiratory ammonification dominated at high NO₂⁻:NO₃⁻ ratios, whereas low NO₂⁻:NO₃⁻ ratios favored denitrification. The NO₂⁻:NO₃⁻ ratios affected nirK transcript abundance, a measure of denitrification activity, in the chemostat experiments, and cells grown at a NO₂⁻:NO₃⁻ ratio of 3 had ∼37-fold fewer nirK transcripts per cell than cells grown with NO₃⁻ as the sole electron acceptor. In contrast, the transcription of nrfA, implicated in NO₂⁻-to-NH₄⁺ reduction, remained statistically unchanged under continuous cultivation conditions at NO₂⁻:NO₃⁻ ratios below 3. At NO₂⁻:NO₃⁻ ratios above 3, both nirK and nrfA transcript numbers decreased and the chemostat culture washed out, presumably due to NO₂⁻ toxicity. These findings implicate NO₂⁻ as a relevant modulator of NO₃⁻ fate in S. loihica strain PV-4, and, by extension, suggest that NO₂⁻ is a relevant determinant for N retention (i.e., ammonification) versus N loss and greenhouse gas emission (i.e., denitrification).

INTRODUCTION

Two major dissimilatory pathways determine the fate of nitrate (NO₃⁻) in anoxic environments: denitrification and respiratory ammonification (1, 2). In denitrification, NO₃⁻ is stepwise reduced via nitrite (NO₂⁻), nitric oxide (NO), and nitrous oxide (N₂O) to dinitrogen (N₂). In respiratory ammonification, NO₃⁻ is reduced via NO₂⁻ to ammonium (NH₄⁺). Nitrite is the common intermediate of the two pathways and the branching point for both dissimilatory pathways. In the presence of NH₄⁺, NO₂⁻ also can serve as the electron acceptor for anaerobic ammonium oxidation (anammox) (3); however, in soil environments this process appears to be less relevant than the other two processes (4). The fate of NO₂⁻ via denitrification or respiratory ammonification has great environmental impact. Denitrification forms gaseous products (N₂O, N₂), which are emitted from the soil, resulting in N loss, whereas respiratory ammonification generates NH₄⁺ leading to N retention (5, 6). Atmospheric N₂O is a potent greenhouse gas and an ozone-depleting agent (7, 8). Therefore, knowledge of the environmental factors that control these NO₂⁻ reduction pathways is needed to estimate, predict, and possibly manipulate N loss versus N retention.

The major sources of NO₂⁻ in the environment include NH₄⁺ oxidation performed by nitrosifiers and NO₃⁻ reduction. In these conversions, NO₂⁻ accumulates when its production is kinetically faster than NO₂⁻ consumption (9, 10). Nitrite accumulation had been thought to occur rarely in the environment (11); however, recent observations suggested that NO₂⁻ formation can occur as a result of NO₃⁻ reduction and NH₄⁺ oxidation under conditions that favor NO₂⁻ production over consumption (12–16). For example, high pH and abundance of NH₄⁺ and hydroxylamine (NH₂OH) affect NO₂⁻ oxidizers, causing NO₂⁻ accumulation from nitrosification (15), while oxygen intrusion and/or electron donor limitations may cause NO₂⁻ accumulation from denitrification (12). In pure-culture studies, several denitrifiers and respiratory ammonifiers were found to reduce NO₃⁻ at a higher rate than NO₂⁻, causing dynamic changes of the NO₂⁻:NO₃⁻ ratios in the medium (10). Additional NO₂⁻ may be generated by NO₃⁻-to-NO₂⁻ reducers sensu stricto, which generate NO₂⁻ as an end product (12). For example, in activated sludge, the activity of NO₃⁻-to-NO₂⁻ reducers leads to NO₂⁻ formation; however, the contribution of NO₃⁻-to-NO₂⁻ reducers to NO₂⁻ accumulation in natural environments is uncertain (17, 18).

Even though NO₂⁻ formation occurs in diverse environments (12, 19), information regarding the impact of NO₂⁻ on dissimilatory NO₃⁻/NO₂⁻ reduction pathways is scarce. Increased NO₂⁻ concentrations and respiratory ammonification activity have been observed in river sediments (20); however, causality was not established. A recent study reported that denitrification was the dominant pathway in a chemostat mixed culture derived from a tidal flat sediment when NO₂⁻ instead of NO₃⁻ was provided as the electron acceptor (21).

Based on genome analyses, at least three different bacterial species harbor both the denitrification and respiratory ammonification pathways: Shewanella loihica strain PV-4, Opitutus terrae strain PB90-1, and Marivirga tractuosa strain DSM 4126 (22). In Shewanella loihica strain PV-4, the functionality of both the denitrification and the respiratory ammonification pathways was confirmed (23). Recent efforts demonstrated the role of C:N ratios on pathway selection in S. loihica strain PV-4; however, the effects of NO₂⁻:NO₃⁻ ratios were not explored. Since fluctuating NO₂⁻ concentrations are expected in many environmental systems, this work explored the effects of NO₂⁻:NO₃⁻ ratios on pathway selection in a defined, tractable experimental system.

MATERIALS AND METHODS

Culture conditions.

Phosphate-buffered basal salts medium for the batch and chemostat experiments was prepared as described previously (24). For batch experiments, 100 ml of anoxic medium was dispensed into 160-ml serum bottles. Throughout the medium preparation procedure, anoxia was maintained by using the Hungate technique. The serum bottles were sealed with black butyl-rubber stoppers (Geo-Microbial Technologies, Inc., Ochelata, OK). The autoclaved medium was amended with vitamins (25) prepared as an anoxic, filter-sterilized, 200-fold concentrated stock solution. Lactate (60% lactate syrup; Sigma-Aldrich, St. Louis, MO) was added as the electron donor from an anoxic, autoclaved 0.5 M stock solution. A maximum of 2.0 mM lactate was required to reduce the total amount of N oxyanions added to the batch culture systems, and excess lactate (i.e., 5 mM) was added to avoid electron donor limitations (24). Various amounts of NO₃⁻ and NO₂⁻ (see below) were added to achieve a total final N oxyanion concentration of 1.0 mM. NH₄⁺ was added to a final concentration of 0.1 mM as a source of biomass N. Sodium nitrate (≥99%; Fisher Scientific, Pittsburg, PA), sodium nitrite (≥99%; JT Baker, Phillipsburg, NJ), and ammonium chloride (≥99%; Fisher Scientific, Pittsburg, PA) were dissolved in distilled water to prepare 0.1 M NO₃⁻, NO₂⁻, and NH₄⁺ stock solutions, which were degassed and autoclaved. After preparation of the culture vessels, 10% of the headspace (6 ml) was withdrawn and 6 ml of acetylene gas (99.6%; Airgas, Knoxville, TN) was added to block N₂O turnover to N₂ and measure N₂O production as a proxy for denitrification activity (26). For the chemostat experiments, higher phosphate (25 mM) and ammonium chloride (0.5 mM) concentrations were used to increase the buffering capacity and provide reactive N for assimilation, respectively. Trace metals were added from 200-fold stock solutions to the chemostat reactor after autoclaving to avoid precipitate formation (23). All experiments were performed at room temperature (21°C).

Analytical procedures.

Analytical measurements followed established protocols (23, 24). NO₃⁻ and NO₂⁻ were measured with a Dionex ICS-2100 system (Sunnyvale, CA), and NH₄⁺ was measured with a Dionex ICS-1100 system. For both ion chromatography systems, the limits of detection for the target analytes was ∼50 μM. N₂O was quantified in headspace gas with an Agilent 3000A Micro gas chromatograph (MicroGC; Palo Alto, CA) with a limit of detection of ∼100 ppm by volume (ppmv). A dimensionless Henry's constant of 1.53 was calculated for N₂O using equation 1 and parameters provided by Sander (http://www.henrys-law.org/henry-3.0.pdf).

$$k_H=\frac{1}{k_H^{\theta }\hspace{0.17em}RT\times \exp \hspace{0.17em}[\frac{d\hspace{0.17em}\ln \hspace{0.17em}{k}_H}{d\hspace{0.17em}(1/T)}\times (\frac{1}{T}-\frac{1}{T^{\hspace{0.17em}\theta }}\left)\right]}$$ (1)

where k_H is the dimensionless Henry's constant, $k_H^{\theta }$ is Henry's constant under standard conditions, expressed as moles liter⁻¹ atm⁻¹, R is the gas constant (0.082 liters atm K⁻¹ mol⁻¹), T is temperature (in Kelvins), and T^θ is temperature under standard conditions (298.15 K).

This Henry's constant was adjusted for the ionic strength of the solution (27, 28). Using the corrected Henry's constant of 1.751, the aqueous-phase concentration of N₂O was calculated from the headspace concentration.

Effect of NO₂⁻:NO₃⁻ ratios on product formation.

NO₂⁻:NO₃⁻ ratios of 0.33, 0.5, 1, and 3 were established by adding NO₃⁻ and NO₂⁻ to achieve a combined concentration of 1 mM. In addition, cultures were grown with 1 mM NO₃⁻ and 1 mM NO₂⁻ as the sole electron acceptors. The vessels were inoculated with 0.5 ml of an S. loihica strain PV-4 culture grown with 2.0 mM lactate and 1.0 mM NO₃⁻. The initial measurements of NO₂⁻, NO₃⁻, NH₄⁺, and N₂O were taken immediately after inoculation, and the final measurements were taken after 5 days, when all of the initial amounts of electron acceptor(s) had been consumed. The quantitative assessment of the amounts of NH₄⁺ and N₂O determined the activity of respiratory ammonification and denitrification, respectively, under the conditions tested. Additional experiments determined that shifting NO₂⁻:NO₃⁻ ratios, rather than changes in absolute NO₂⁻ concentration, affected denitrification and ammonification activities. S. loihica strain PV-4 batch cultures were amended with 5 mM lactate and 0.25, 0.5, 0.75, and 1.0 mM NO₂⁻, and N₂O and NH₄⁺ amounts were determined following complete NO₂⁻ consumption.

In addition to batch cultures, chemostat experiments were performed to allow the control of the NO₂⁻:NO₃⁻ ratios during growth. The chemostat design (460 ml total volume, 200 ml medium) and general operational parameters have been described previously (23). Anoxic medium was prepared in 2-liter glass bottles with a C:N ratio of 4.5 by adding 3.0 mM lactate and a 2.0 mM total concentration of NO₃⁻ and NO₂⁻. This medium was supplied to the continuously stirred tank reactor at a flow rate of 20 ml h⁻¹. NO₂⁻:NO₃⁻ ratios of the medium were adjusted to 0 (no NO₂⁻), 0.33, 1, and 3. To initiate the chemostat culture, lactate and NO₃⁻ were added to final concentrations of 2 mM and 1 mM, respectively, along with the vitamin and trace metal solutions. After flushing the reactor vessel with N₂ gas for 1 h, S. loihica strain PV-4 cells were inoculated. The reactor was incubated for 2 days with N₂ flushing before operation of the chemostat system began. The reactor headspace was flushed with a constant stream of N₂ gas to prevent O₂ intrusion and CO₂ and N₂O buildup. For all experimental conditions examined, steady-state conditions, indicated by stable dissolved carbon and nitrogen species concentrations, was reached within 72 h. In a steady-state chemostat system, the consumption of the reactants and the formation of products are quantified as rates of change. The steady-state concentration of NO₂⁻ and NO₃⁻ were maintained below the detection limit of ∼0.05 mM, and consumption rates in the reactor were calculated by multiplying the flow rate by the influent medium concentrations. The rates of NH₄⁺ production in the reactor were calculated from the steady-state NH₄⁺ concentrations, influent NH₄⁺ concentrations, and the flow rate (equation 2) and served as a measure of respiratory ammonification activity.

$$P_{{\text{NH}}_4^{+}}=Q(C_{{\text{NH}}_4^{+}}-C_{{\text{NH}}_4^{+},0})$$ (2)

where $P_{{\text{NH}}_4^{+}}$ is the rate of NH₄⁺ production (micromoles hour⁻¹), Q is the flowrate of the chemostat (milliliters hour⁻¹), $C_{{\text{NH}}_4^{+}}$ is the steady-state ${\text{NH}}_4^{+}$ concentration in the reactor (millimolar), and $C_{{\text{NH}}_4^{+},0}$ is the ${\text{NH}}_4^{+}$ concentration of the influent medium (millimolar).

Denitrification activity was quantified by measuring N₂O production rates in the headspace in the presence of acetylene. The N₂ gas flow was stopped, and 26 ml (10%) of the headspace N₂ was replaced with acetylene gas. N₂O concentrations in the headspace then were measured four times in 40- to 50-min intervals. Since the dissolved N₂O exits the system with the effluent, the loss was calculated for accurate measurement of the N₂O production rates according to equation 3.

$$L_{{\mathrm{N}}_{2}O}=Q\hspace{0.17em}{\int }_{t_1}^{t_2}f(t)\hspace{0.17em}dt$$ (3)

where L_N2O is the loss of N₂O with the effluent (micromoles), t₁ is the time of initial N₂O measurement (hours), t₂ is the time of final N₂O measurement (hours), and f(t) is the regression equation for the change in aqueous N₂O concentration.

Using this approach, the NH₄⁺ and N₂O production rates could be directly compared. After each culture suspension sampling event, the reactor volume was adjusted to 200 ml by operating the chemostat with the outflow valve closed. Independent replicate chemostat experiments were performed for each NO₂⁻:NO₃⁻ ratio tested. After each experiment, the reactor was disassembled, thoroughly cleaned, and sterilized.

Growth yield determination.

Dry weight measurements determined the biomass produced in defined medium amended with 5 mM lactate and 1.0 mM NO₂⁻ or 1.0 mM NO₃⁻ (22). The MicroGC was used for fast (1-min run time) headspace CO₂ concentration measurements, and the cessation of CO₂ formation served as a proxy for complete NO₂⁻/NO₃⁻ consumption. Forty ml of the culture suspension was passed through a preweighed 0.22-μm membrane filter (Millipore, Billerica, MA). The filters with biomass were dried at 100°C for 24 h and cooled in a desiccator until weight consistency was achieved. For each growth condition, triplicate culture vessels were examined. The dry weight data were corrected for the weight of salts associated with the culture medium and introduced onto the membrane filter.

Nucleic acid extraction and analyses.

For cell quantification and gene expression analyses, 15-ml samples were collected from the reactor under steady-state conditions with different NO₂⁻:NO₃⁻ ratios. The samples for expression analyses were prepared by immediately mixing 0.5-ml aliquots with 1.0 ml of RNA Protect bacterial reagent (Qiagen, Germantown, MD) and collecting the biomass by centrifugation for 10 min at 5,000 × g. The pellets were stored immediately at −80°C. The samples for cell enumeration by quantitative PCR (qPCR) were prepared by centrifuging 1.5-ml aliquots at 16,000 × g, and the pellets were immediately stored at −80°C. The extraction and purification of total RNA and reverse transcriptase quantitative PCR (RT-qPCR) followed an established protocol (23). Immediately prior to extraction, 1 μl luciferase control mRNA (Promega, Madison, WI) diluted to 10¹⁰ copies μl⁻¹ was added to each frozen cell pellet. The recovery of the control mRNA was used to account for RNA loss during the extraction and reverse transcription procedures (29, 30). The cell pellets then were disrupted with an Omni Bead Ruptor 24 homogenizer (Omni, Kennesaw, GA) in 350 μl buffer RLT provided with the RNeasy Mini kit (Qiagen, Germantown, MD), and total RNA was extracted from the cell pellets according to the protocol provided with the kit. Extracted mRNA was purified using an RNA MinElute kit (Qiagen) after DNA was digested with the RNase-free DNase set kit (Qiagen). An 11-μl aliquot of purified RNA solution was subjected to reverse transcription using Superscript III reverse transcriptase (Invitrogen, Carlsbad, CA) according to the manufacturer's recommendations, and the remaining 7 μl of RNA extract was tested for DNA contamination using qPCR (23).

Genomic DNA for cell quantification was extracted using a DNeasy blood and tissue kit (Qiagen, Santa Clarita, CA) as described previously (24). PCR primers used in this study are summarized in Table 1. S. loihica strain PV-4 cells were enumerated with qPCR targeting 16S rRNA genes in the genomic DNA, and gene expression analyses were performed with qPCR targeting nirK, nrfA, and luc cDNA templates. In silico PCR analysis (31) suggested that the Slo16Sf and Slo16Sr primer set amplified all nine 16S rRNA copies of S. loihica strain PV-4. Strain PV-4 possesses two different nrfA gene copies, and previous findings indicated that nrfA₈₄₄ responds to N oxyanions (23). Transcripts of nrfA₀₅₀₅ were detected, but they were found to be of at least two orders of magnitude lower abundance than those for nrfA₀₈₄₄ under all growth conditions tested. qPCR was performed with an ABI ViiA7 real-time PCR system (Life Technologies) equipped with ViiA7 software v1.1 using Power SYBR green detection chemistry (Life Technologies, Carlsbad, CA) (23, 24). Plasmids carrying the strain PV-4 16S rRNA gene nirK and nrfA inserts were used to construct qPCR calibration curves. These plasmids were prepared by inserting PCR amplicons of the respective target genes into the PCR2.1 vector using the TOPO TA cloning kit (Invitrogen). Tenfold dilution series with concentrations ranging from 10⁸ to 10¹ copies/μl were prepared and used as qPCR templates. The amplification efficiencies ranged from 94.5% to 103.2%. The standard deviations in the quantification cycle (C_q) values increased up to 5-fold (up to 0.71) between the template DNA dilutions of 10 and 100 copies μl⁻¹, indicating a detection limit of about 10 copies μl⁻¹ for all three target genes. No amplification was observed in controls without target DNA/cDNA, and the amplicons of each target exhibited consistent and expected melting curve profiles, indicating specific amplification. Both cDNA and genomic DNA samples were prepared from triplicate samples, and triplicate qPCR assays were performed for each DNA and cDNA sample. The RT-qPCR data were adjusted based on the luc control mRNA recovery, which ranged from 7.0% to 23.8% and was within the expected range (30). The corrected expression data were normalized to the cell abundance.

TABLE 1.

Primers used for RT-qPCR analyses and qPCR calibration curve parameters

Primer	Sequence	Target gene (locus tag)	Amplicon length (bp)	Slope	y intercept	Amplification efficiency (%)	R2	Reference
Slo_nirK853f	AAGGTGGGTGAGTCTGTGCT	nirK (Shew_3335)	188	−3.393	35.64	97.1	0.999	23
Slo_nirK1040r	GGCTGGCGGAAGGTGTAT
Slo_nrfA1083f	GGATATCCGTCACGCTCAAT	nrfA (Shew_0844)	226	−3.299	34.93	101.0	0.992	24
Slo_nrfA1308r	GTCCATACCCAATGCAGCTT
Slo_16Sf	CACACTGGGACTGAGACACG	16S rRNA genes	191	−3.461	34.27	94.5	1.000	23
Slo_16Sr	TGCTTCTTCTGCGAGTAACG
luc_refA	TACAACACCCCAACATCTTCGA	Luciferase gene control mRNA	67	−3.247	35.46	103.2	0.999	29
luc_refB	GGAAGTTCACCGGCGTCAT

Statistical analyses.

Statistical analyses for both phenotypic and expression analysis data were performed with SPSS 22.0 software for Mac (IBM Corp., Armonk, NY, USA). Unless otherwise mentioned, triplicate data sets were collected for each of the batch and chemostat experiments, and the Student's t tests were performed to test the null hypotheses. The RT-qPCR data were transformed to logarithmic scales for statistical analyses. P values below 0.05 were regarded as significant.

RESULTS

Batch cultivation with NO₃⁻ or NO₂⁻.

S. loihica strain PV-4 batch cultures reduced NO₃⁻ or NO₂⁻ to different products, depending on which substrate was provided (Fig. 1). An initial amount of 100.0 ± 1.9 μmol of NO₃⁻ was reduced to 90.3 ± 6.6 μmol of N₂O-N after a 68-h incubation period, indicating that ∼90% of the NO₃⁻ was denitrified. Some NH₄⁺ also was produced, and the initial amount of 10.3 ± 0.4 μmol provided in the medium increased to 14.0 ± 3.5 μmol. Cultures that received NO₂⁻ but no NO₃⁻ produced exclusively NH₄⁺, and no N₂O was observed (i.e., N₂O below the detection limit of ∼100 ppmv, or 4.15 μmol liter⁻¹ dissolved-phase concentration at equilibrium). Of the 98.9 ± 1.1 μmol of NO₂⁻ added to the cultures, 86.9 ± 1.5 μmol (87.8% ± 1.8%) was recovered as NH₄⁺. The lack of complete recovery probably was due to NH₄⁺-N assimilation into biomass. These batch culture studies suggested a role of NO₂⁻:NO₃⁻ ratios for controlling NO₂⁻ fate via the respiratory ammonification or denitrification pathways.

FIG 1.

NO₃⁻/NO₂⁻ reduction in batch cultures of S. loihica strain PV-4. All vessels received 10% acetylene (C₂H₂) gas in the headspace for blocking N₂O reduction to N₂. Cultures were amended with 5.0 mM lactate as the electron donor and 1.0 mM NO₃⁻ (A) or 1.0 mM NO₂⁻ (B) as the electron acceptor. The change in the amounts of NO₃⁻ (■), NO₂⁻ (●), NH₄⁺ (□), and N₂O-N (○) in the vessels was monitored until all NO₃⁻ and NO₂⁻ in the medium was reduced to NH₄⁺ and N₂O. The error bars represent the standard deviations from triplicate cultures.

S. loihica strain PV-4 cultures amended with different NO₂⁻ concentrations (0.25, 0.5, 0.75, and 1 mM) produced predominantly NH₄⁺, independent of the initial amount of NO₂⁻ provided (Table 2). With NO₂⁻ as the sole electron acceptor, no more than 5% of the total NO₂⁻ provided was reduced to N₂O. This finding suggests that the absolute NO₂⁻ concentration (in the absence of NO₃⁻) did not affect pathway selection in batch cultures.

TABLE 2.

Reduced product distribution in S. loihica batch cultures amended with various NO₂⁻ concentrations

Initial NO2− concn (mM)	NH4+ produced (μmol/vessel)a	% recovered as NH4+	N2O-N produced (μmol/vessel)a	% recovered as N2O-N
0.25	22.9 (3.3)	91.6	1.0 (0.1)	4.0
0.5	41.7 (1.4)	83.4	2.3 (0.1)	4.6
0.75	64.8 (0.5)	86.4	2.6 (0.6)	3.5
1.0	86.1 (1.7)	86.1	1.5 (0.1)	1.5

^aNumbers in parentheses are standard deviations from triplicate samples.

Under batch cultivation conditions with 99.6 ± 0.8 NO₃⁻ as the electron acceptor, a biomass yield of 14.9 ± 1.5 μg (dry weight) per μmol NO₃⁻ reduced was determined (Table 3). When amended with 99.3 ± 0.3 μmol NO₂⁻, the biomass yield was 16.2 ± 2.2 μg (dry weight) per μmol NO₂⁻ reduced. Therefore, S. loihica strain PV-4 had statistically similar (P > 0.05) growth yields from denitrification of NO₃⁻ to N₂O and respiratory ammonification of NO₂⁻ to NH₄⁺ per amount of electron acceptor consumed.

TABLE 3.

Biomass yield of S. loihica strain PV-4 upon growth on 5.0 ml lactate and 1.0 mM NO₃⁻ or 1.0 mM NO₂⁻

Electron acceptora		Producta				ΔG0′ (kJ/mol of e− accepter)	fsc
Type	Amt (μmol/vessel)	N2O-N (μmol/vessel)	NH4+ (μmol/vessel)	Biomass (μmol C5H7O2N/vessel)	Biomass yieldb (mg/μmol of e− acceptor)
NO3−	99.6 (0.8)	87.5 (1.5)	2.8 (1.7)	13.1 (1.4)	14.9 (1.5)	−342.3	0.458
NO2−	99.3 (0.3)	1.4 (0.4)	84.0 (3.8)	14.5 (1.9)	16.2 (2.2)	−397.3	0.407

^aNumbers in parentheses are standard deviations from triplicate samples.

^bBiomass yield was determined with the assumption that the amounts of the minor products were negligible compared to those of the major products.

^cf_s values were calculated with the assumption that NO₃⁻ and NH₄⁺ contributed equally to biomass (24 e⁻ eq/mol C₅H₇O₂N on average) (24). The following equation was used to calculate f_s: f_s = (e⁻ eq to biomass)/[(e⁻ eq to biomass) + (e⁻ eq to e⁻ acceptors)].

Effects of NO₂⁻:NO₃⁻ ratios on pathway selection.

Varying the initial inputs of NO₂⁻ and NO₃⁻ in batch cultures affected the relative abundance of reduced products. The net production of NH₄⁺ increased and the production of N₂O decreased with increasing NO₂⁻:NO₃⁻ ratios (Fig. 2). At NO₂⁻:NO₃⁻ ratios below 1, N₂O was the major product, suggesting that the denitrification pathway was favored. At a NO₂⁻:NO₃⁻ ratio of 0.33, 83.5 ± 0.1 μmol of N₂O-N and 11.4 ± 2.8 μmol of NH₄⁺ were recovered from 94.6 ± 0.7 μmol of N oxyanions. At a NO₂⁻:NO₃⁻ ratio of 1 (0.5 mM NO₂⁻ and 0.5 mM NO₃⁻), 94.2 ± 0.1 μmol of N oxyanions was reduced to 59.2 ± 12.8 μmol N₂O-N and 29.5 ± 13.7 μmol NH₄⁺ (Fig. 2). At a NO₂⁻:NO₃⁻ ratio of 3, significantly more NH₄⁺ (70.2 ± 6.0 μmol) than N₂O (21.6 ± 0.7) was produced from reduction of NO₂⁻ and NO₃⁻ (P < 0.05), indicating that higher NO₂⁻:NO₃⁻ ratios shifted NO₃⁻/NO₂⁻ reduction toward respiratory ammonification. These batch experiments demonstrated that the relative proportion of NO₂⁻ to NO₃⁻ affected pathway selection and reduced product formation; however, the dynamic change in the relative proportions of NO₃⁻ and NO₂⁻ (i.e., changing NO₂⁻:NO₃⁻ ratios) (Fig. 1A) made the interpretation of the results ambiguous. For example, NO₃⁻ consumption rates exceeded NO₂⁻ utilization, causing an increase in the NO₂⁻:NO₃⁻ ratio over time in the NO₃⁻-fed batch cultures (Fig. 1A).

FIG 2.

Net production of N₂O (white bars) and NH₄⁺ (hatched bars) over a 5-day incubation period at 21°C in S. loihica strain PV-4 cultures amended with 5.0 mM lactate and different NO₂⁻:NO₃⁻ ratios to achieve 1.0 mM total N-oxyanion concentration. Each bar represents the averages from triplicate samples, with the error bars representing their standard deviations.

To better evaluate NO₂⁻ effects on pathway selection, chemostat experiments were performed that allowed cultivation under constant NO₂⁻:NO₃⁻ ratios. Since the C:N ratio dominates the control of the N pathway products, chemostat experiments were conducted at a C:N ratio of 4.5 (i.e., 3 mM lactate and 2 mM NO₃⁻ plus NO₂⁻ in the feed medium), which was shown not to favor one pathway or the other (23). These C:N conditions in the reactor system revealed the effects of various NO₂⁻:NO₃⁻ ratios on pathway selection (Fig. 3). When the electron acceptor pool consisted entirely of NO₃⁻, N₂O and NH₄⁺ were produced at rates of 8.9 ± 0.5 μmol h⁻¹ (17.8 ± 1.0 μmol h⁻¹ N₂O-N) and 8.8 ± 0.2 μmol h⁻¹, respectively, indicating that approximately twice as much NO₃⁻ was reduced to N₂O than to NH₄⁺. Higher NO₂⁻:NO₃⁻ ratios increased the net NH₄⁺ production but decreased the N₂O production, leading to the formation of significantly larger (P < 0.05) amounts of NH₄⁺ and significantly smaller (P < 0.05) amounts of N₂O. At a NO₂⁻:NO₃⁻ ratio of 0.33 (feed rates of 29.3 ± 0.1 μmol NO₃⁻ h⁻¹ and 9.1 ± 0.1 μmol NO₂⁻ h⁻¹), the NH₄⁺ production rate exceeded the N₂O production rate on a per-N-atom basis by 1.36-fold ± 0.03-fold. At a NO₂⁻:NO₃⁻ ratio of 3, NH₄⁺ was produced at a rate of 26.8 ± 0.1 μmol h⁻¹ and N₂O formation was not observed. When 2.0 mM NO₂⁻ was provided as the sole electron acceptor in the influent medium, biomass washout occurred (i.e., the reactor failed), suggesting toxicity when NO₂⁻ was continuously supplied at a rate of 40 μM h⁻¹. The chemostat results were consistent with the outcome of the batch culture experiments and indicated that increasing NO₂⁻:NO₃⁻ ratios favor respiratory ammonification.

FIG 3.

Net production of N₂O (white bars) and NH₄⁺ (hatched bars) in chemostat cultures of S. loihica strain PV-4 with NO₂⁻:NO₃⁻ ratios varying from 0 to 3 in the feed solution. The feed solution contained 5 mM lactate and combinations of NO₂⁻ and NO₃⁻ with a total N oxyanion concentration of 2.0 mM. The N oxide(s) were supplied to the reactor at ∼40 μmol h⁻¹. Each bar represents the averages from two samples taken from separate reactors, with error bars representing the standard deviations.

nrfA and nirK expression analysis.

The RT-qPCR analyses of nrfA and nirK transcripts supported the phenotypic data collected from the batch and chemostat experiments (Fig. 4). The cell numbers were statistically indistinguishable at all NO₂⁻:NO₃⁻ ratios tested (P > 0.05 by one-way analysis of variance [ANOVA]). The number of nirK transcripts was highest (1.3 ± 0.4 transcripts cell⁻¹) when 2.0 mM NO₃⁻ was the sole electron acceptor in the feed medium and lowest (3.5 × 10⁻² ± 2.6 × 10⁻² transcripts cell⁻¹) at a NO₂⁻:NO₃⁻ ratio of 3. The numbers of nirK transcripts detected per cell were not significantly different (P > 0.05) in cells maintained in chemostats receiving medium with a NO₂⁻:NO₃⁻ ratio of 0.33 or with NO₃⁻ as the sole electron acceptor; however, nirK expression at a NO₂⁻:NO₃⁻ ratio of 1 (1.8 × 10⁻¹ transcripts cell⁻¹) was significantly lower than expression at lower NO₂⁻:NO₃⁻ ratios. Different expression patterns were observed with nrfA, and transcription levels with NO₃⁻ as the sole electron acceptor (9.6 transcripts cell⁻¹) were statistically indistinguishable from the transcription levels at NO₂⁻:NO₃⁻ ratios of 0.33 and 1 (P > 0.05). The transcription of nrfA was downregulated at a NO₂⁻:NO₃⁻ ratio of 3, with 0.6 ± 0.4 transcripts cell⁻¹ compared to 11.5 ± 7.3 transcripts cell⁻¹ at a NO₂⁻:NO₃⁻ ratio of 1. This significant decrease in nrfA expression may be attributed to the repression of metabolic activity due to NO₂⁻ toxicity. As the expression of nrfA remained relatively constant, the decrease in expression of nirK at high NO₂⁻:NO₃⁻ ratios provides an explanation for the observed decrease in denitrification activity in strain PV-4.

FIG 4.

RT-qPCR analysis of nirK (white bars) and nrfA (hatched bars) transcripts in S. loihica strain PV-4 chemostat cultures under various NO₂⁻:NO₃⁻ ratios. The feed solution contained 5 mM lactate and combinations of NO₂⁻ and NO₃⁻ with a total N oxyanion concentration of 2.0 mM. Each bar represents the average copies of transcript measured in triplicate samples subjected to independent extraction procedures. The error bars represent the standard deviations from triplicate independent samples.

DISCUSSION

Both the phenotypic and the transcriptional analyses indicated that the concentration of NO₂⁻ relative to that of NO₃⁻ (i.e., the NO₂⁻:NO₃⁻ ratio) is a controlling factor influencing the fate of these N oxyanions. Both the batch and chemostat experiments performed with various NO₂⁻:NO₃⁻ ratios demonstrated that as the proportion of NO₂⁻ increases, respiratory ammonification activity supersedes denitrification activity. Several pure-culture studies report the effects of NO₃⁻ or NO₂⁻ on pathway selection and the fate of NO₃⁻/NO₂⁻ (32–35). For example, decreased nrfA expression and subsequent accumulation of NO₂⁻ were observed in Escherichia coli chemostat cultures, with steady-state NO₃⁻ concentrations exceeding 0.2 μM (34). The experiments with S. loihica strain PV-4 demonstrated that the initial NO₂⁻ concentration did not affect pathway selection (Table 2) and suggested that the NO₂⁻:NO₃⁻ ratio is a critical parameter determining pathway selection. Low NO₃⁻ concentrations (0.2 mM) were shown to induce denitrification pathway enzymes, including NirK, in Agrobacterium tumefaciens (35); however, S. loihica strain PV-4 batch cultures amended with 0.25 mM NO₃⁻ and 0.75 mM NO₂⁻ (i.e., a NO₂⁻:NO₃⁻ ratio of 3) exhibited predominantly ammonification (Fig. 2). This observation also supports that the NO₂⁻:NO₃⁻ ratio, rather than the absolute N oxyanion concentration, is the relevant control parameter. Further, the effects of NO₂⁻:NO₃⁻ ratios were observed in strain PV-4 chemostat cultures with very low (below the limits of detection) steady-state NO₃⁻ and NO₂⁻ concentrations, emphasizing the relevance of the NO₂⁻:NO₃⁻ ratios for pathway selection. Increased input of NO₃⁻ to chemostat cultures did not affect nrfA expression, suggesting NO₃⁻ does not inhibit nrfA gene activity, as was observed in E. coli (34). Therefore, a plausible explanation for the observations is that the NO₂⁻:NO₃⁻ ratio, rather than the absolute concentration of NO₂⁻ or NO₃⁻, determines pathway selection in S. loihica strain PV-4.

The reasons why S. loihica strain PV-4 prefers respiratory ammonification under high NO₂⁻:NO₃⁻ ratios may be based on thermodynamics. S. loihica strain PV-4 produced similar amounts of biomass per micromole of NO₃⁻ denitrified to N₂O and per micromole of NO₂⁻ reduced to NH₄⁺ in batch cultures. Less biomass would be produced if S. loihica strain PV-4 denitrified NO₂⁻, as denitrification from NO₂⁻ to N₂O would involve transfer of only two electrons per N atom compared to a four-electron transfer from NO₃⁻ to N₂O. NO₃⁻ reduction to NO₂⁻ is the most efficient step in terms of energy conservation in the denitrification pathway (36). Energetic considerations (37) also suggest that respiratory ammonification is more favorable under high NO₂⁻:NO₃⁻ ratios. With lactate as the electron donor (oxidized to acetate), the ΔG^0′ of NO₃⁻-to-N₂O reduction is −398.5 kJ/mol NO₃⁻, and less free-energy change is associated with NO₂⁻-to-N₂O reduction (−231.1 kJ/mol NO₂⁻). Respiratory ammonification of NO₂⁻ with a ΔG^0′ of −450.1 kJ/mol NO₂⁻ is energetically more favorable, providing a rationale for the preference of respiratory ammonification at high NO₂⁻:NO₃⁻ ratios.

The preference of denitrification over respiratory ammonification at low NO₂⁻:NO₃⁻ ratios, on the other hand, cannot be explained in terms of energetics and associated growth yields, as respiratory ammonification of NO₃⁻ would yield more biomass and energy than NO₃⁻ denitrification. Possible explanations include more efficient electron transfer in denitrification than in respiratory ammonification or the kinetic properties of the enzymes involved. For example, NO₂⁻ toxicity may be a reason to favor a pathway that minimizes intermediate NO₂⁻ accumulation (10). Depending on the electron donor provided, 20 to 60% of the initial NO₃⁻ was observed as NO₂⁻ during denitrification in strain PV-4 cultures (24). Previous experiments with S. oneidensis strain MR-1 demonstrated that NO₂⁻ reduction to NH₄⁺ proceeded only after NO₃⁻ had been stoichiometrically converted to NO₂⁻ (10). In S. loihica strain PV-4, the NO-forming NO₂⁻ reductase NirK may have faster kinetics than the NH₄⁺-forming NO₂⁻ reductase NrfA at physiologically relevant NO₂⁻ concentrations; thus, it would prevent NO₂⁻ toxicity. When NO₂⁻ exceeds NO₃⁻ concentrations in the medium, this advantage is lost and the NO₃⁻/NO₂⁻ reduction pathway is determined by energetics. Unfortunately, these S. loihica enzyme systems have not been characterized and kinetic data are not available to corroborate this hypothesis.

Somewhat unexpectedly, the expression levels of both nrfA and nirK decreased when the NO₂⁻:NO₃⁻ ratio was raised to 3. Higher NO₂⁻:NO₃⁻ ratios resulted in reactor failure, although NO₂⁻ concentrations in the reactor vessel never exceeded 2.0 mM. These observations suggest that the continuous supply of NO₂⁻ exerted stress on S. loihica strain PV-4 cells. The inhibitory effect of elevated NO₂⁻ concentrations on strain PV-4 may be the reason why the organism initially was identified as a nondenitrifier (38). NO₂⁻ toxicity was observed previously in Shewanella oneidensis strain MR-1, and no growth occurred with 4 mM NO₂⁻ provided as the electron acceptor (10). NO₂⁻ was found to be toxic to other microbial groups, including methanogens and phosphate-accumulating bacteria, at concentrations below 1 mM (39, 40), and 3 mM NO₂⁻ inhibited respiratory ammonification in Wolinella succinogenes (36). Although NO₂⁻ did not result in cell death in Agrobacterium tumefaciens, NO₂⁻ impacted O₂ and NO₂⁻ metabolism (35). These findings indicate that microorganisms vary in their susceptibility to NO₂⁻ toxicity. Therefore, a continuous supply of NO₂⁻ without NO₃⁻ may select for organisms tolerant to elevated NO₂⁻ concentrations, and this selection process may bias the outcome of the competition between denitrifiers and respiratory ammonifiers. Kraft et al. (21) observed increased denitrification activity in a tidal flat mixed community when NO₂⁻ was replaced with NO₃⁻ as the electron acceptor. In the chemostat seeded with the tidal flat mixed community, phylogenetically diverse populations contributed to denitrification and respiratory ammonification (21). These populations may vary in their susceptibility to NO₂⁻ toxicity, and continuous NO₂⁻ feeding without NO₃⁻ may have selected for organisms with greater NO₂⁻ tolerance. A mixture of organic compounds and/or sulfide were provided as electron donors to the tidal flat sediment chemostat (21), suggesting that a diversity of organisms was supported. In pure-culture studies of denitrifiers and respiratory ammonifiers, growth yields varied greatly depending on the bacterial strain examined and the electron donors used (36). It is possible that the dominant denitrifying population in the mixed-culture chemostat community utilized NO₂⁻ more efficiently than NO₃⁻, as observed in pure-culture studies with P. stutzeri (36), and has an advantage over ammonifying strains when grown with NO₂⁻.

Prior chemostat experiments with strain PV-4 and NO₃⁻ as the sole electron acceptor demonstrated that the simultaneous production of NH₄⁺ and N₂O occurred only when the C:N ratio of the influent medium was 4.5 (23). The effects of various NO₂⁻:NO₃⁻ ratios on denitrification versus respiratory ammonification pathway selection were apparent only in chemostats operated at a C:N ratio of 4.5. Apparently, the C:N ratio control dominates over the NO₂⁻:NO₃⁻ ratio for pathway selection (Fig. 5). Respiratory ammonification dominates at C:N ratios above 4.5, and denitrification dominates at C:N ratios below 4.5 regardless of the NO₂⁻:NO₃⁻ ratio the culture is experiencing. These observations suggest that the role of NO₂⁻:NO₃⁻ ratios in determining the fate of NO₃⁻/NO₂⁻ is secondary to the role of the C:N ratio. Nevertheless, NO₂⁻ may play a relevant role as a modulator of denitrification versus respiratory ammonification activities under conditions where neither electron donors nor N oxyanions are available in excess. As the results from the batch experiments suggested, NO₂⁻:NO₃⁻ ratios also become important when organic electron donor and N oxyanions are provided in nonlimiting amounts, regardless of the C:N ratio. Such conditions may occur following N fertilizer application events in agricultural soils. NO₃⁻ reduction may exceed NO₂⁻ consumption and subsequently increase the NO₂⁻: NO₃⁻ ratio, leading to a condition favoring respiratory ammonification.

FIG 5.

Hierarchical effects of C:N and NO₂⁻:NO₃⁻ ratios on the fate of NO₂⁻/NO₃⁻ via denitrification and/or respiratory ammonification in Shewanella loihica strain PV-4 chemostat cultures.

Certain environmental conditions may favor high rates of aerobic NH₄⁺ oxidation and lower rates of NO₂⁻ oxidation to NO₃⁻, which can lead to elevated NO₂⁻ concentrations. For example, elevated NO₂⁻ concentrations have been observed in an alkaline fluvo-aquic loam soil for >200 h after application of an ammonium sulfate solution (21). The transient formation of up to 67.2 mg NO₂⁻ kg soil⁻¹ was observed after application of 100 mg urea kg⁻¹ (14). In semiarid pine forest soils, first winter rains caused transient formation of NO₂⁻, and the NO₂⁻ concentrations greatly exceeded those of NO₃⁻ (13, 14, 41). Our results suggest that the fluctuating NO₂⁻:NO₃⁻ ratios in such environments can influence the activities of microbial community members involved in NO₃⁻ and NO₂⁻ reduction. Further investigations of NO₃⁻ and NO₂⁻ reduction in environmental systems experiencing fluctuating NO₂⁻:NO₃⁻ ratios is warranted to more comprehensively understand the parameters controlling the fate of N oxyanions in anoxic environments.

In summary, the growth experiments with various NO₂⁻/NO₃⁻ supplies to S. loihica strain PV-4 revealed that NO₂⁻ affects denitrification and respiratory ammonification pathway selection. In both batch and chemostat cultures, the increase in NO₂⁻ content in the electron acceptor pool favored respiratory ammonification over denitrification, and transcriptional analyses of nirK and nrfA genes supported the phenotypic observations. However, the chemostat experiments demonstrated that respiratory ammonification dominated even at low NO₂⁻:NO₃⁻ ratios when sufficient electron donor was available, indicating that the C:N ratio has higher-level control over NO₂⁻/NO₃⁻ fate. Pure-culture studies have limitations to predict the behavior of natural microbial assemblages. Nevertheless, the results obtained with S. loihica strain PV-4 add to our understanding of environmental parameters governing the fate of NO₃⁻ and NO₂⁻ in anoxic ecosystems. The pure-culture experiments allowed the observation of the isolated effect of various NO₂⁻:NO₃⁻ ratios on pathway selection (i.e., denitrification and respiratory ammonification) in the absence of competition.