The medium for cultivation of S. loihica strain PV-4 was prepared as described previously (Yoon et al., 2013). Medium containing 20 g NaCl, 0.233 g KH₂PO₄, 0.46 g K₂HPO₄ and 2 mL trace metal solution (Myers and Nealson, 1990) per liter was boiled with N₂ flushing. pH of the medium was adjusted to 6.0, 7.0, or 8.0 with 5.0 N HCl or 5.0 N NaOH before boiling. After 100-mL aliquots of the degassed medium were distributed to N₂-flushed 160 mL serum bottles, the bottles were capped with black butyl stoppers (Geo-Microbial Technologies, Inc., Ochelata OK, United States) and aluminum crimp seals and autoclaved. Wolin vitamin solution was added from a filter-sterilized 200X stock solution after autoclaving (Wolin et al., 1964). The medium was amended with sodium lactate, KNO₃ (or KNO₂) and NH₄Cl from sterilized anoxic stock solutions to final concentrations of 560.3 mg/L (5 mM), 101.1 mg/L (1 mM), and 10.7 mg/L (0.2 mM), respectively. The inoculated serum bottles were incubated in dark at 25°C without shaking. After each experiment, final pH was measured to confirm that pH was maintained constant (within ± 0.1 of the initial value).

Concentrations of NO₃^–, NO₂^–, NH₄⁺ and N₂O were monitored using established analytical procedures. N₂O concentration was measured using HP 6890 Series gas chromatograph equipped with a HP-PLOT Q column (30 m × 0.320 mm diameter, 20 μm film thickness) and an electron capture detector (Agilent, Palo Alto, CA, United States). Headspace N₂O concentration was measured by manually injecting 100 μL headspace sample into the gas chromatograph. The detection limit was approximately 2 ppmv. The total amounts of N₂O in the vessels were calculated from the headspace concentrations using Equation 1.

where C_h is the headspace concentration in moles/L, V_a and V_h are the volumes of the aqueous phase and the headspace, respectively, in liters and H is the dimensionless Henry’s constant of N₂O. The dimensionless Henry’s constant at 25°C was calculated as previously described (Yoon et al., 2015b). After correcting for the effect of high salt concentration in the medium, the dimensionless Henry’s constant was calculated to be 1.82.

Aqueous NO₂^– and NH₄⁺ concentrations were determined colorimetrically using HS-NO₂ (N)-L and HS-NH₃ (N)-L kits (Humas, Daejeon, Korea), respectively, according to the protocols provided by the manufacturer. Aqueous NO₃^– concentration was determined with Metrohm 863 Basic IC plus ion chromatography system (Metrohm, Riverview, FL, United States) equipped with a Metrosep A Supp 4-250/4.0 anion exchange column. In presentations of the experimental data, the quantities of the nitrogen species were expressed in μmoles/bottle for convenience in mass balance calculations.

Precultures of S. loihica strain PV-4 cells were prepared with 5.0 mM lactate and 1.0 mM NO₃^– as the electron donor and the electron acceptor, respectively. These precultures were incubated until NO₃^– and NO₂^– were depleted, and 0.5 mL of the precultures were inoculated to serum bottles with fresh media adjusted to the desired pH (6.0, 7.0, or 8.0) and amended with 5.0 mM lactate and 1 mM NO₃^– or NO₂^–. Immediately after inoculation, 0.5 mL of headspace N₂ was aseptically replaced with >99.999% N₂O (Samoh Corporation, Daejeon, Korea). With time intervals determined from preliminary experiments, headspace and aqueous phase samples were extracted for monitoring of NO₃^–, NO₂^–, NH₄⁺ and N₂O concentrations and cell densities. For determination of N₂O concentration, 100 μL of headspace gas was extracted using a 1700-series gastight syringe (Hamilton Company, Reno, NV, United States) and manually injected into the gas chromatograph. Immediately after headspace sampling, 1.5 mL of the aqueous phase was extracted using a disposable 3-mL syringe (Becton Dickinson, Franklin Lakes, NJ, United States). In order to avoid pressure loss in the vessels, 1.6 mL of N₂ gas was injected before extraction of the liquid samples. After OD_{600 nm} of the extracted cell suspension was measured using Genesys 30 visible spectrophotometer (Thermo Scientific, Waltham, MA, United States), the suspension was centrifuged and the concentrations of NO₃^–, NO₂^–, NH₄⁺ and N₂O in the supernatant were measured. Incubation was carried out until no further change in the concentrations of the nitrogen species was observed.

The samples for transcription analyses of the genes involved in denitrification and respiratory ammonification were extracted at multiple time points during batch cultivation of S. loihica strain PV-4 prepared and carried out identically as described above. For each batch culture amended with NO₃^– and N₂O, N₂O concentration was monitored to determine three sampling time points (Supplementary Figure S1). The pH 6.0 cultures were sampled at t = 24, 48, and 60 h, the pH 7.0 cultures were sampled at t = 11, 24, and 27 h, and pH 8.0 cultures were sampled at t = 19, 45, and 52 h. As no growth was observed at pH 6.0 when incubated with NO₂^– and N₂O, samples were extracted only from the cultures incubated at pH 7.0 and 8.0. These cultures were sampled twice, before the onset of the exponential phase (t = 17 and t = 20, respectively, for pH 7.0 and pH 8.0 cultures) and during the mid-exponential phase (t = 49 h and t = 45 h, respectively, for pH 7.0 and pH 8.0 cultures). One-half milliliter samples were collected from three biological replicates upon each sampling event. One milliliter of RNA Protect Bacteria Reagent (Qiagen, Hilden, Germany) was immediately added to each of the aliquots and the mixture was immediately centrifuged for 10 min at 5,000 × g. The cell pellets were stored at -80°C until further processing.

An established protocol was used for extraction and purification of total RNA with a few modifications (Park et al., 2017). The cell pellets were thawed in ice and were subjected to disruption with Omni Bead Ruptor 12 Homogenizer (Omni International, Kennesaw, GA, United States) after addition of 350 μL buffer RLT solution and 30 mg of 0.1 mm glass beads (Omni International). Total RNA was extracted using RNeasy Mini Kit (Qiagen) according to the protocol provided by the manufacturer and resulting RNA was eluted with 60 μL RNase-free water. Remaining DNA in the eluent was digested using RNase-free DNase Set Kit (Qiagen) as previously described (Park et al., 2017) and the reaction mix was purified using RNase-free DNase Kit (Qiagen). Reverse transcription was performed with 11 μL of 20 μL eluent using Superscript^TM III Reverse Transcriptase (Invitrogen, Carlsbad, CA, United States). The remaining eluent was used later in a quantitative polymerase chain reaction (qPCR) to confirm the absence of contaminating DNA. After reverse transcription, each sample was treated with RNase H (Invitrogen) to remove traces of RNA. The cDNA solution was diluted five-fold with nuclease-free water and stored at -20°C until analyzed using qPCR.

The quantities of nirK, nrfA, nosZ, and recA genes in the cDNA solutions were determined using the qPCR technique. nirK and nosZ genes were present in single copies in S. loihica strain PV-4 genome and only nrfA₀₈₄₄ of the two nrfA-like genes was targeted, as the expression of nrfA₀₅₀₅ did not correlate with respiratory ammonification activity (Yoon et al., 2015a). The primers used in this study are listed in Table 1. The primers specifically targeting nrfA, nirK, and recA genes of S. loihica strain PV-4 were designed de novo using Primer 3 software (Rozen and Skaletsky, 2000) and a previously designed primer was used for amplification of nosZ (Yoon et al., 2015a). The target specificities of the primer sets were tested with ordinary PCR and the PCR products were used to construct calibration curves for absolute quantification of target genes. The PCR products were inserted into PCR2.1 vectors and the resulting plasmids were extracted using QIAprep Spin Miniprep Kit (Qiagen). The copy numbers of the extracted plasmids were calculated from the nucleic acid concentration measured with Nanodrop 2000 UV-Vis spectrometer (Thermo Fisher Scientific) and the expected molecular weights of the plasmids. Dilution series of the plasmids ranging from 1 to 10⁸ copies/μL were prepared and used for calibration curve construction. qPCR was performed with QuantStudio^TM 3 Real-Time PCR System (Thermo Fisher Scientific) using SYBR Green detection chemistry. 2X Power SYBR Green PCR Master Mix Solution (Applied Biosystems) was used to prepare the reaction mix and a two temperature-cycle procedure was used for qPCR, with 95°C denaturation step (15 s) followed by 60°C elongation step (60 s) in each of 40 cycles. The amplification efficiencies ranged between 94.0 and 99.8% and the R² values of the calibration curves were no less than 0.998 (Table 1). No amplification was observed in the negative controls without target DNA or cDNA. The qPCR assays for all four target genes were reliable down to 10¹ copies/μL, and the preparations with 1 copy/μL did not yield consistent results. Consistent melting curves confirmed the specificity of the qPCR reactions.

The copy numbers of nirK, nrfA, and nosZ in each cDNA sample were normalized with the copy number of recA, the housekeeping gene that encodes DNA recombination/repair protein RecA. Transcription levels of recA genes were previously observed to be relatively stable under different growth conditions and growth stages in diverse groups of bacteria including Streptococcous agalactiae and Lactobacillus plantarum (Marco and Kleerebezem, 2008; Florindo et al., 2012). Thus, recA was selected as the most suitable target gene for normalization of the RT-qPCR data to account for the differences in cell densities and overall metabolic activities, as well as mRNA loss during extraction, purification, and reverse transcription procedures. The copy numbers of the nirK, nrfA, and nosZ genes in the cDNA samples were divided by the copy number of the recA gene and the nirK/recA, nrfA/recA, and nosZ/recA values were compared across the samples collected from different incubation conditions.

Statistical analyses for the RT-qPCR results were performed using SPSS Statistics 24 software (IBM Corp. Armonk, NY, United States). The RT-qPCR reactions were performed with the samples collected from triplicate reaction vessels and independently processed through extraction, purification and reverse transcription procedures. Statistical analyses were performed with the data transformed to a logarithmic scale.

pH conditions determined whether S. loihica strain PV-4 reduced NO₃^– to NH₄⁺ (respiratory ammonification) or to N₂ via N₂O (through denitrification) and also, whether the batch system functioned as a sink or a source of N₂O during NO₃^– reduction (Figure 1). At all three pH conditions tested, all of NO₃^– added to a nominal concentration of 1.0 mM was consumed after 28 – 72 h after inoculation. The strain PV-4 culture incubated at pH 7.0 had the shortest lag period (∼8 h), while pH 6 and pH 8 cultures both had longer lag phases, as significant decreases in NO₃^– concentrations and increases in the cell densities were observed 35 and 22 h after inoculation, respectively. The maximum observed rates of NO₃^– reduction (pH 6.0: 14.1 μmoles h^-1 mL^-1 OD_{600 nm}^-1 at t = 54.5 h; pH 7.0: 19.1 μmoles h^-1mL^-1 OD_{600 nm}^-1 at t = 13.5 h; pH 8.0: 6.4 μmoles h^-1 mL^-1 OD_{600 nm}^-1 at t = 36 h) and exponential growth rates (0.18, 0.33, and 0.15 h^-1 at pH 6, 7, and 8, respectively) indicated that neutral pH was optimal for S. loihica strain PV-4, but also that overall cellular function of S. loihica strain PV-4 was not substantially compromised by the shift of pH within the examined range.

The strain PV-4 cultures incubated at different pH conditions were distinguished by the magnitudes of transient NO₂^–accumulation. Accumulation of NO₂^– was not observed in the culture incubated at pH 6.0 and the concentration of NO₂^– remained below the detection limit throughout the experiment, indicating that potential NO₂^– reduction rate was at least as high as the rate of NO₃^– reduction. Contrastingly, at pH 7.0 and 8.0, NO₂^– accumulations up to 61.7 ± 2.6 μmoles/bottle and 82.6 ± 9.0 μmoles/bottle were observed, respectively, indicating that the rates of NO₂^– reduction were slower than the rates of NO₃^– reduction.

At pH 7.0 and 8.0, S. loihica strain PV-4 functioned as a net sink of N₂O, as N₂O concentrations decreased consistently throughout incubation periods (Figure 1). No transient N₂O accumulation was observed in the cultures incubated at pH 7.0 and 8.0. A clearly distinct trend was observed in the strain PV-4 culture incubated at pH 6.0, as a transient accumulation of N₂O up to 72.3 ± 4.5 μmoles N₂O-N/bottle (37.4 μmoles N₂O-N/bottle higher than initially added N₂O-N) occurred before NO₃^– depletion. Although N₂O accumulation was observed, N₂O reduction was active throughout the experiment. The maximum N₂O reduction rate of 12.2 μmoles N₂O-N h^-1 mL^-1 OD_{600 nm}^-1 was calculated at t = 54.5 h. This rate was comparable to the calculated maximum N₂O reduction rate of 14.2 μmoles N₂O-N h^-1 mL^-1 OD_{600 nm}^-1 at pH 7.0 at t = 23.5 h. At pH 8.0, the maximum observed N₂O reduction rate (279 nmoles N₂O-N h^-1mL^-1 OD_{600 nm}^-1 at t = 40 h) was substantially lower than those observed at pH 6.0 and 7.0. N₂O reduction stopped after 60 h, when the dissolved N₂O concentration was lowered to 0.75 μM and no other electron acceptor was available in the medium. Energy gained from N₂O reduction may not be sufficient per se to support growth at the alkaline pH. In fact, no significant growth or N₂O consumption was observed for 200 h when strain PV-4 was incubated with N₂O as the sole electron acceptor at pH 8.0 (data not shown).

pH was also a determinant of NO₃^– fate in the strain PV-4 cultures. Nitrogen mass balance was used to estimate the magnitude of denitrification activity, assuming that >90% of NO₃^–/NO₂^– was dissimilated to either NH₄⁺ or denitrification products (Yoon et al., 2015a). At pH 6.0, consumption but not production of NH₄⁺ was observed, as the amount of NH₄⁺ decreased from 21.1 ± 0.5 μmoles/bottle to 4.2 ± 2.1 μmoles/bottle after cultivation, indicating that denitrification was the dominant NO₃^– reduction pathway. The respiratory ammonification pathway was mostly switched off, although a statistically insignificant increase (3.4 ± 1.9 μmoles/bottle) in the amount of NH₄⁺ was observed after t = 66.5 h. At pH 7.0, an increase in the NH₄⁺ concentration was observed between t = 25 h and t = 27 h and ∼9.6% (9.7 ± 2.3 μmoles) of initially added 101.6 ± 2.1 μmoles NO₃^– was reduced to NH₄⁺; however, NH₄⁺ was still a minor product. The distribution of products from NO₃^– reduction at pH 8.0 was distinctively different from the results observed at pH 6.0 and 7.0. NH₄⁺ was the major product of NO₃^– reduction, as 83.5 ± 1.2 μmoles of NO₃^– was reduced to NH₄⁺. Respiratory ammonification dominated NO₃^– reduction at the alkaline pH and shutdown of the NirK-mediated NO₂^– reduction activity could be inferred from the mass balance. With diminished denitrification activity, S. loihica strain PV-4 cultures functioned as an N₂O sink at pH 8.0 despite of reduced N₂O reduction activity.

When S. loihica strain PV-4 cultures were amended with NO₂^– instead of NO₃^–, the effects of pH on NO₂^– reduction pathways were obscured by the effect of NO₂^– (Figure 2). At pH 6.0, neither cell growth or significant change in concentrations of NO₂^–, NH₄⁺, or N₂O was observed for >300 h. Lowering of the initial NO₂^– concentration to 0.1 mM did not result in cell growth or metabolic activity (data not shown), precluding the possibility that the toxicity of HNO₂ at the acidic pH was the cause of growth inhibition. The experiments at pH 7.0 and pH 8.0 yielded statistically indistinguishable results (p > 0.05), as NH₄⁺ was the predominant product (78.9 ± 8.5 μmoles and 85.7 ± 5.5 μmoles recovered as NH₄⁺ at pH 7.0 and 8.0, respectively) regardless of pH. No significant change in the concentration of N₂O was observed at either pH, indicating that the pathways leading to the production and consumption of N₂O were inactive when S. loihica strain PV-4 was incubated with NO₂^–. These results suggested that the pH effect on nitrogen-oxide dissimilation was eclipsed by the NO₂^– effect.

Transcription analyses were performed using RT-qPCR technique with samples extracted at different time points during incubation of S. loihica strain PV-4 (Figures 3, 4). To account for the differences in the cell densities and overall metabolic activities at different culturing conditions and growth stages, the nosZ transcription data were normalized with the recA transcription data. At pH 6.0 and pH 7.0, nirK transcription levels were at least five-fold higher than those of nrfA throughout the incubation periods except at t = 60 h at pH 6.0. The reduced transcription of nirK at this time point may be due to the depletion of NO₃^–. At pH 8.0, the differences in transcription of nrfA and nirK were insignificant (p > 0.05) throughout incubation (Figure 3). These transcription profiles explained the dominance of denitrification activity at pH 6.0 and 7.0 and the predominance of respiratory ammonification activity at pH 8.0. As expected from the sustained N₂O reduction activity at pH 6.0, nosZ transcription was not adversely affected by the acidic pH (Figure 3). Transcription of nosZ was significantly more active at pH 6.0 and 7.0 than at pH 8.0, as the maximum nosZ transcripts / recA transcript values of 7.28 ± 2.59, 3.15 ± 1.50, and 0.62 ± 0.16 were recovered from samples extracted from the mid-exponential-phase cultures incubated at pH 6.0 (at t = 48 h), 7.0 (at t = 27 h), and 8.0 (at t = 52 h), respectively.

In the samples extracted from S. loihica strain PV-4 cultures grown with NO₂^– and N₂O as the electron acceptors, distinguishable shifts in nirK, nrfA, and nosZ transcription levels from the lag phase (t = 17 h for pH 7.0 and t = 20 h for pH 8.0) to the mid-exponential phase (i.e., at t = 49 h for pH 7.0 and t = 45 h for pH 8.0) were observed (Figure 4). At the mid-exponential phase when active NO₃^–/NO₂^– reduction occurred, nrfA transcription levels were an order of magnitude higher than nirK transcription levels. The nirK transcription-to-nrfA transcription ratios were statistically indistinguishable between pH 7.0 and pH 8.0 (p > 0.05). These observations were in agreement with the predominance of respiratory ammonification. The diminished transcription of the genes responsible for N₂O-producing reactions (nirK) and N₂O-consuming reaction (nosZ) explained the absence of N₂O production or consumption in the cultures amended with NO₂^– and N₂O.