Denitrification Biokinetics: Towards Optimization for Industrial Applications

Denitrification is a microbial process that converts nitrate (NO3–) to N2 and can play an important role in industrial applications such as souring control and microbially enhanced oil recovery (MEOR). The effectiveness of using NO3– in souring control depends on the partial reduction of NO3– to nitrite (NO2–) and/or N2O while in MEOR complete reduction of NO3– to N2 is desired. Thauera has been reported as a dominant taxon in such applications, but the impact of NO3– and NO2– concentrations, and pH on the kinetics of denitrification by this bacterium is not known. With the goal of better understanding the effects of such parameters on applications such as souring and MEOR, three strains of Thauera (K172, NS1 and TK001) were used to study denitrification kinetics when using acetate as an electron donor. At low initial NO3– concentrations (∼1 mmol L–1) and at pH 7.5, complete NO3– reduction by all strains was indicated by non-detectable NO3– concentrations and near-complete recovery (> 97%) of the initial NO3-N as N2 after 14 days of incubation. The relative rate of denitrification by NS1 was low, 0.071 mmol L–1 d–1, compared to that of K172 (0.431 mmol L–1 d–1) and TK001 (0.429 mmol L–1 d–1). Transient accumulation of up to 0.74 mmol L–1 NO2– was observed in cultures of NS1 only. Increased initial NO3– concentrations resulted in the accumulation of elevated concentrations of NO2– and N2O, particularly in incubations with K172 and NS1. Strain TK001 had the most extensive NO3– reduction under high initial NO3– concentrations, but still had only ∼78% of the initial NO3-N recovered as N2 after 90 days of incubation. As denitrification proceeded, increased pH substantially reduced denitrification rates when values exceeded ∼ 9. The rate and extent of NO3– reduction were also affected by NO2– accumulation, particularly in incubations with K172, where up to more than a 2-fold rate decrease was observed. The decrease in rate was associated with decreased transcript abundances of denitrification genes (nirS and nosZ) required to produce enzymes for reduction of NO2– and N2O. Conversely, high pH also contributed to the delayed expression of these gene transcripts rather than their abundances in strains NS1 and TK001. Increased NO2– concentrations, N2O levels and high pH appeared to cause higher stress on NS1 than on K172 and TK001 for N2 production. Collectively, these results indicate that increased pH can alter the kinetics of denitrification by Thauera strains used in this study, suggesting that liming could be a way to achieve partial denitrification to promote NO2– and N2O production (e.g., for souring control) while pH buffering would be desirable for achieving complete denitrification to N2 (e.g., for gas-mediated MEOR).

− concentrations (∼1 mmol L −1 ) and at pH 7.5, complete NO 3 − reduction by all strains was indicated by non-detectable NO 3 − concentrations and near-complete recovery (> 97%) of the initial NO 3 -N as N 2 after 14 days of incubation. The relative rate of denitrification by NS1 was low, 0.071 mmol L −1 d −1 , compared to that of K172 (0.431 mmol L −1 d −1 ) and TK001 (0.429 mmol L −1 d −1 ). Transient accumulation of up to 0.74 mmol L −1 NO 2 − was observed in cultures of NS1 only. Increased initial NO 3 − concentrations resulted in the accumulation of elevated concentrations of NO 2 − and N 2 O, particularly in incubations with K172 and NS1. Strain TK001 had the most extensive NO 3 − reduction under high initial NO 3 − concentrations, but still had only ∼78% of the initial NO 3 -N recovered as N 2 after 90 days of incubation. As denitrification proceeded, increased pH substantially reduced denitrification rates when values exceeded ∼ 9. The rate and extent of NO 3 − reduction were also affected by NO 2 − accumulation, particularly in incubations with K172, where up to more than a 2fold rate decrease was observed. The decrease in rate was associated with decreased transcript abundances of denitrification genes (nirS and nosZ) required to produce enzymes for reduction of NO 2 − and N 2 O. Conversely, high pH also contributed to the delayed expression of these gene transcripts rather than their abundances in strains NS1 and TK001. Increased NO 2 − concentrations, N 2 O levels and high pH appeared to cause higher stress on NS1 than on K172 and TK001 for N 2 production. Collectively,

INTRODUCTION
Denitrification is a major nitrate (NO 3 − ) reduction pathway that contributes to biochemical gas (NO, N 2 O and N 2 ) production in anoxic environments. It is a microbially facilitated process performed by facultative anaerobic bacteria capable of reducing NO 3 − to nitrite (NO 2 − ), ultimately producing di-nitrogen (N 2 ) via other gas species (NO and N 2 O; Korom, 1992). The complete reduction of NO 3 − to N 2 requires the presence of sufficient amounts of organic (for organoheterotrophic denitrification; like organic acids, hydrocarbons) and/or inorganic (for lithotrophic denitrification; like sulfur containing compounds, transition metals) electron donors (Spormann and Widdel, 2000;Chakraborty and Coates, 2004;Zhu and Getting, 2012;Capua et al., 2019). The transient accumulation of intermediates NO,and N 2 O) can be observed at the onset of or during denitrification, under optimum growth and metabolic conditions. On the other hand, accumulated NO 2 − and N 2 O can persist at sub-optimal and extreme environmental conditions (e.g., high temperatures, acidic or alkaline pH conditions, high salinities; Liu et al., 2014;Fida et al., 2016;An et al., 2017). Limited availability of electron donors can also drive partial denitrification, but many natural denitrifier habitats contain rich sources of electron donors.
Denitrification is widespread in nature and the production of NO 2 − , N 2 O and/or N 2 as end products is required in many NO 3 − -dependent practical uses. This metabolic process can play an important role in controlling sulfide formation and microbially enhanced oil recovery (MEOR) in oil reservoirs. Souring control can be achieved by NO 2 − inhibition of dissimilatory sulfite reductase (Dsr) required for the reduction of sulfite to sulfide by sulfate-reducing bacteria (SRB; Gieg et al., 2011;Callbeck et al., 2013). Nitrous oxide (N 2 O) in its dissolved form can also be beneficial to limiting sulfide production since it is toxic to SRB and many other microbial species by raising the redox potential in its ambient environment (Sørensen et al., 1980;Londry and Suflita, 1999). However, the liberation of excess N 2 O to the atmosphere is a concern because of its high potency for ozone depletion in the stratosphere and contribution to greenhouse gas levels (Ravishankara et al., 2009;Montzka et al., 2011). Nitrite can also cause toxicity to other organisms. Thus, the complete reduction of produced NO 2 − and N 2 O to N 2 is preferred in many industrial applications, including NO 3 − removal from contaminated waters and gas-mediated enhanced oil recovery (Soares, 2000;Rocca et al., 2007;Lazar et al., 2007;Safdel et al., 2017;Huno et al., 2018).
The production of denitrification intermediates and N 2 is dependent on physico-chemical growth conditions, enzyme kinetics and differential gene expression patterns among denitrifying species (Philippot et al., 2001;Liu et al., 2013;Fida et al., 2016). The majority of denitrifying bacteria are distributed within the classes α-, β-, γ-, and εof phylum Proteobacteria though they can be members of other phyla such as Firmicutes, Actinobacteria, Bacteroidetes and Planctomyces (Zumft, 1997). Several genera of these bacteria have been enriched and isolated from different environments to determine rates of denitrification and their abilities to express denitrification genes. Their phenotypic responses vary more with diverse regulatory networks than the genes controlling denitrification outcomes (Zumft, 1997;Van Spanning et al., 2007). Under mesophilic conditions, members of the genus Thauera within the Betaproteobacteria have been reported to be dominant in NO 3 − -impacted environments (Agrawal et al., 2012;Shen et al., 2018). Some strains of this genus were isolated and characterized for denitrification products and their denitrification regulatory phenotypes (DRPs; Liu et al., 2013;Fida et al., 2016). Thauera can use volatile fatty acids (VFAs such as acetate, propionate and butyrate), monoterpenes, heterocyclic aromatic compounds and hydrocarbons as electron donors for organoheterotrophic NO 3 − reduction (Foss and Harder, 1998;Mechichi et al., 2002). Two distinct types of DRPs have been distinguished in Thauera, based on the accumulation of NO 2 − as an intermediate when NO 3 − reduction occurs with acetate oxidation. These include rapid, complete onset (RCO) or progressive onset (PO) of denitrification genes. The regulatory control of N 2 O was observed to be less stringent than the other intermediates among all the Thauera strains irrespective of their DRPs. Although the N 2 O intermediate remained as a major end product in the incubations of one of these strains (T. phenylacetica; Liu et al., 2013), it was not clear whether this difference was related to its phylogeny, specific DRP, or physico-chemical growth conditions. This particular study (Liu et al., 2013) did not include strains like T. aromatica that are often the dominant denitrifiers in different environments, and which have been found to be important mediators of NO 3 − -dependent sulfide control and MEOR (Evans et al., 1991;Anders et al., 1995;Mechichi et al., 2002;Agrawal et al., 2012;Suri et al., 2019).
To further extend the knowledge on regulatory phenotypes of other Thauera strains and the factors affecting NO 3 − reduction to N 2 production in the interest of application to industrial processes, Thauera strains isolated from organic carbon-rich environments were selected for the current study. The known optimum growth temperature and pH range for these strains is 28-30 • C and 7.0-7.5, respectively. Incomplete to complete reduction of 1 to 5 mmol L −1 NO 3 − using VFAs, esters and alkylbenzenes as electron donors by these strains was shown in previous studies under these optimal conditions (Mechichi et al., 2002;Fida et al., 2016;Suri et al., 2017). The partial reduction of NO 3 − to NO 2 − , and prolonged NO 2 − accumulation was observed at NO 3 − concentrations higher than 2 mmol L −1 , but the role of N 2 O and N 2 cannot be assessed since their concentrations were not reported. In addition, the effect of important growth parameters (e.g., pH) on their metabolic abilities to produce N 2 was not addressed (Fida et al., 2016;Suri et al., 2017;Suri et al., 2019). The variable production of NO 2 − , N 2 O and N 2 by denitrifiers is desired in different industrial applications, which can potentially be customized by modification of their growth environments. Keeping the published data in view, the impact of initial NO 3 − concentrations, intermediate denitrification products (NO 2 − , N 2 O), and culture pH values on the rate and extent of denitrification were evaluated for the growth of representative Thauera strains. In addition, NO 2 − toxicity to the Thauera stains was evaluated by using it as an electron acceptor.

Denitrifying Bacterial Strains
Three strains of Thauera were used in this study ( Table 1). Two of these namely, T. aminoaromatica TK001 and T. aromatica NS1, were isolated from enrichment cultures obtained using produced water from a NO 3 − treated heavy oil reservoir (Fida et al., 2016;Suri et al., 2019). The third strain, T. aromatica K172, originated from activated sludge and was purchased from DSMZ culture collection, Braunschweig, Germany 1 .

Culture Medium and Growth Conditions
All strains were grown in CSBK medium (a minimal salts medium; Suri et al., 2017) for comparison of their denitrification characteristics. After autoclaving, the medium was allowed to cool to room temperature while being flushed with 99.9% helium (He) to remove dissolved oxygen. The cooled medium was amended with pre-sterilized stock solutions of trace elements and tungstate and selenite (Widdel and Bak, 1992). Sodium bicarbonate (0.03 M) was added as a buffer and the initial pH of medium was adjusted to 7.5 by addition of 1 M hydrochloric acid (HCl).
Microcosms were prepared by dispensing 50 mL anoxic medium into 120 mL serum bottles using the Hungate technique (Löffler et al., 2005). The sterilized anaerobic stock solutions of sodium acetate (1 M) and sodium nitrate (1 M) were prepared aseptically with ultrapure (milliQ R ) water. Microcosms were prepared using two different acetate and NO 3 − concentrations (0.5 and 1 mmol L −1 ), with the same concentrations of acetate and NO 3 − (1:1) added to each individual microcosm. The serum bottles were sealed with butyl rubber stoppers and crimped with aluminum seals.
Frozen bacterial stocks of Thauera strains were initially used for cultivation in CSBK with 0.5 mmol L −1 of NO 3 − and acetate. Incubations were conducted at 30 • C until the NO 3 − and reduced intermediates were not detectable in microcosms (using the analytical methods described in sections "Biochemical Analyses" and "Headspace Gas Measurements"). The bacterial cultures were centrifuged, the collected cell pellets were resuspended in CSBK, and were then used to inoculate duplicate or triplicate microcosms containing 1 mmol L −1 of NO 3 − and acetate (1:1) to monitor denitrification kinetics. Prior to inoculation, the ∼70 mL headspace of each bottle was flushed with He and equilibrated to laboratory barometric pressure (∼0.9 atm) by piercing the butyl rubber stopper with a sterile syringe needle.
Bacterial cells were added to microcosms to an optical density (OD 600 ) of approximately 0.005 in each bottle to attain similar initial cell densities for each of the three strains. The microcosms were then incubated at 30 • C, and bacterial growth and changes in concentrations of acetate, NO 3 − , NO 2 − , N 2 O and N 2 were monitored by withdrawing aqueous and gas phase samples periodically from the bottles during incubation using He-flushed syringes following procedures outlined in sections "Biochemical Analyses" and "Headspace Gas Measurements."

Biochemical Analyses
Culture sub-samples (0.5-1 mL) were collected during the incubation period to assess bacterial growth and substrate consumption rates. The aqueous samples were centrifuged in microcentrifuge tubes at 13,000 rpm for 10 min, and the collected supernatants were analyzed for acetate, NO 3 − , and NO 2 − concentrations using High Performance Liquid Chromatography (HPLC; Waters 600E; Suri et al., 2017;Okpala and Voordouw, 2018). The remaining cell pellets were washed with sterilized distilled water twice and re-suspended in 1 mL of water. The bacterial cell concentrations were subsequently estimated by measuring optical density of cell suspensions at 600 nm (OD 600 ), using distilled water as a blank. Acetate concentrations were measured by HPLC using 300 µL of the supernatant aliquots that were acidified with 20 µL of 1 M H 3 PO 4 prior to being eluted through a Prevail organic acid (OA) 5 µ column (250 mm × 4.6 mm, Alltech, Guelph, ON, Canada) using 25 mM KH 2 PO 4 (pH 2.5) at a flow rate of 1 mL min −1 . Supernatants (100 µL) mixed with 400 µL of acetonitrile buffer were eluted through an IC-PAK anion column (4 × 150 mm, Waters) with 24% (v/v) acetonitrile, 2% butanol and 2% borate-gluconate buffer at a flow rate of 2 mL min −1 for measuring NO 3 − and NO 2 − . The peaks detected using a Waters 2487 UV detector at 210 and 220 nm were compared with known standards to obtain acetate, NO 3 − and NO 2 − concentrations, respectively. The pH of aqueous phase samples was measured using a Thermo Scientific Inc., Orion model 310 pH meter (VWR International, Mississauga, ON, Canada) calibrated to a pH range of 4-10. Alkalinity was determined by titration of aqueous phase samples with phthalate buffer in the presence of bromophenol blue to an end point pH of 3.5. Absorbance of bromophenol blue complex measured at 600 nm (A 600 ) compared to the known standards was used as a measure of total alkalinity expressed as millieq/L. An automated pre-calibrated titrator equipped with a spectrophotometer (Thermoscientific) was used for these measurements.

Headspace Gas Measurements
Headspace gases (CO 2 , N 2 O and N 2 ) were quantified using gas chromatography (GC; Hewlett Packard 5890A). Prior to sampling, the needle of an empty syringe was first pierced into rubber stoppers to release excess pressure in the headspace to barometric pressure (Supplementary Figure 1). The volume that the piston moved to was recorded as additional headspace volume. Samples (1 mL) withdrawn from the headspace using a gas tight syringe (Agilent) were injected into a GC equipped with a RT-Msieve 5A column (Restek- RT-19722; 30 m × 0.32 mm) and pulsed discharge detector-HIID mode (PDD-HIID). The injector, oven, and detector temperatures were set to 28 • C, 30 • C, and 33 • C, respectively. Helium ( > 99.99% purity; Praxair) served as the carrier and makeup gas. Standard curves generated in the range of 0.001 to 99.9% (equivalent to 0.002 to 20 mmol L −1 ) were used to calculate headspace concentrations of gases in the culture bottles. Aqueous concentrations of gases were calculated from the headspace concentrations using equation 1 and dimensionless Henry's constants of 0.726, 0.528, and 0.014 for CO 2 , N 2 O and N 2 , respectively at 30 • C. The constants were calculated using parameters provided by Sander (2015): where C aq , C g , and H are the aqueous concentration (in mmol L −1 ), the headspace concentration (in mmol L −1 ) and dimensionless Henry's constant, respectively. The Henry's constants were further adjusted for the ionic strengths of the medium and culture conditions (Schumpe et al., 1982;Schumpe, 1993).

Calculation of % Mole N-Species Remaining
In some instances, the data are presented as the % of a particular N-species remaining (e.g., NO 3 − , NO 2 − , N 2 O or N 2 ). These values were estimated as follows (using the percent N 2 produced from NO 3 − as an example, where NO 3 − INITIAL , NO 3 − FINAL and N 2 produced are the initial NO 3 − concentration, final NO 3 − concentration and final N 2 concentration, respectively, each expressed as mmol L −1 )

Batch Microcosm Tests
Batch microcosms were prepared using 120 mL pre-sterilized serum bottles flushed with He, closed with butyl rubber stoppers and crimped with aluminum seals. In each case, sterilized solutions of acetate and N-species were added aseptically at different initial concentrations to anoxic CSBK medium and the pH was adjusted using 1 M HCl or 1 M NaOH. The headspaces of the bottles were subsequently flushed with He and equilibrated to barometric pressure prior to being inoculated with Thauera strains at initial cell densities ca. OD 600 ∼ = 0.005. Four sets of iterative batch tests were prepared using the techniques above (sets 1 to 4; Table 2), with N-species from the three electron acceptors that participate in denitrification (i.e., NO 3 − , NO 2 − , or N 2 O), and in one case with varying initial pH values (7.5 to 10). In each case acetate was added as an electron donor in equimolar concentrations to the initial N-species (i.e., in 1:1 ratio). Two to four replicate microcosms were used for each of the three denitrifying strains in each set of batch microcosm tests. After preparation and inoculation, microcosms were incubated at 30 • C for either 14, 30, or 90 days (depending on the experiment). Samples for initial aqueous (acetate, NO 3 − and NO 2 − ) and headspace concentrations (CO 2 , N 2 O and N 2 ) were collected immediately after inoculation and for the final measurements at the end of the incubation period. Denitrification intermediates, alkalinity and pH were also measured with time and with sampling conducted during the incubation period as described above.
Effect of Initial pH on NO 3 − Reduction to N 2 Anoxic CSBK medium containing 1 mmol L −1 of acetate and NO 3 − (1:1) in sealed serum bottles was amended with different ratios of pre-sterilized solutions of sodium bicarbonate (1 M) and sodium carbonate (1 M). The range of initial pH in these bottles was 7.5 to 10 (set 5; Table 2). The bottles were inoculated with Thauera isolates at similar initial cell densities (OD 600 ∼ = 0.005). Headspaces of these bottles were flushed and equilibrated with He to barometric pressure prior to incubation at 30 • C. The reduction of NO 3 − , production of reduced intermediates (NO 2 − , N 2 O) and N 2 upon denitrification was measured periodically in the aqueous  5 Effect of initial pH and expression of denitrification genes as a function of pH NO 3 − ∼1, with NaHCO 3 and Na 2 CO 3 added at variable ratios to reach initial pH values that ranged from 7.5 to 10 Figure 6 and Supplementary Figure 6 phase and headspace of bottles during 14 days of incubation. The time of incubation was chosen based on the observations from initial denitrification experiments conducted at pH 7.5 ( Figure 1). Considering the amount of N 2 produced in cultures under this pH equivalent to 1, the fold change was calculated using equation 3 and measured concentrations of N 2 in the bottles. For values lower than 1, the fold decrease was calculated using equation 4. (3) Fold decrease = − 1/Fold change (4)

Quantitative PCR (qPCR) Assay
Culture subsamples (1 mL) from incubations with optimum and higher pH (section "Effect of Initial pH on NO 3 − Reduction to N 2 ") were transferred to 1.5 mL sterilized RNase free microcentrifuge tubes during early and late exponential phases and were centrifuged at 12,000 rpm for 30 min to collect cell pellets. The collected cell pellets were frozen immediately at −80 • C. RNA was extracted from these frozen cell pellets within 2 weeks of collection using the RNeasy kit (Qiagen). For extraction, the cell pellets were first re-suspended in 100 µL of TE buffer (10 mM Tris, 1 mM EDTA) containing 15 mg/mL lysozyme (Sigma) and the tubes were incubated at room temperature on a shaker at 300 rpm for 10 min. Buffer RLT (Qiagen) was then added (350 µL) to the tubes followed by vigorous vortexing. The resulting mixture was transferred to a 2 mL tube containing 25-50 mg of acidwashed glass beads (212-300 µm; Sigma) and the cells were disrupted in a Tissuelyser LT (Qiagen) at 50 Hz for 5 min. The lysate obtained was centrifuged at maximum speed for 5-10 s, the supernatant was transferred to a clean tube, and then mixed with 220 µl of 100% ethanol. The extracted RNA was purified from the mixture using RNeasy spin columns according to the manufacturer's instructions and quantified using a nanophotometer (Implen).
Total RNA (100-500 ng) was reverse transcribed to obtain cDNA using the SuperScript IV VILO Master Mix (Invitrogen) according to the manufacturer's instructions. The cDNA concentrations of all the samples were normalized to 1 ng/µl with PCR grade water. Real time qPCR assays were performed with primers specific for 16S rRNA, nirS (nitrite reductase) and nosZ (nitrous oxide reductase) genes using PowerUP SYBR Green Master Mix (Applied Biosystems) on a Quanstudio 3 Real Time PCR System (Thermofisher) running Quantstudio Design and Analysis v1.5.1 Software. The primer sequences used for PCR amplification of these genes are summarized in Table 3. The 16S rRNA gene primers were designed based on the similarity of aligned 1400-1500 bp 16S rRNA sequences of Thauera strains used in this study. The PCR cycling conditions were 95 • C for 2 min followed by 45 cycles of 95 • C for 1 s and 60 • C for 30 s. The specificity of the PCR reaction was confirmed using melt curve analysis. Standard curves were performed using control cDNA from representative samples that yielded a dynamic linear range of 0.005 to 50 ng sample per reaction and efficiencies ranging from 96.8 to 102.6% (Bustin et al., 2009). The data was analyzed using the Thermofisher relative quantification application on the Thermofisher cloud that employs the 2 − CT method for determining relative gene expression (Schmittgen and Livak, 2008).

Gene Sequence Analysis and Nucleotide Accession Numbers
The cDNA amplicons from representative end-point qPCR reactions for the 16S rRNA, nirS and nosZ genes were purified using PCR spin columns (Sigma). The gene sequences were obtained by Sanger dideoxy sequencing at the Core DNA Services Laboratory of the University of Calgary with both forward and reverse primers ( Table 3). The resulting sequences were assembled into contigs using SnapGene software 2 . The gene identity of assembled sequences was confirmed using NCBI Nucleotide BLAST search using 16S rRNA gene sequences as control  sequences for the three strains 3 . The identified partial nirS and nosZ sequences of strains of Thauera obtained were submitted to NCBI GenBank database (accession numbers MT186692 to MT186694 and MT186689 to MT186691, respectively).

Denitrification Kinetics of Thauera Strains
All of the tested Thauera strains rapidly reduced NO 3 − to N 2 using acetate as an electron donor upon incubation under identical initial conditions (∼ 0.9 mmol L −1 NO 3 − , pH 7.5; Figure 1). Incubations were performed in four replicates via two separate experiments (both included in Set 1; Table 2) conducted with two replicates in each. The rates of denitrification, and end products were similar within the replicates and the experiments for each strain. Nitrate concentrations were reduced to below detection levels ( < 0.001 mmol L −1 ) in all the cultures within two days.
The denitrification kinetics in cultures with K172 and TK001 were similar, with complete reduction of NO 3 − to N 2 occurring within three days of incubation for both cultures (with denitrification rates of 0.431 and 0.429 mmol L −1 d −1 , respectively; Table 4). On the other hand, NS1 took 14 days for complete conversion of NO 3 − to N 2 , with transient accumulation of NO 2 − concentrations that peaked within two days at 0.74 mmol L −1 and declined over the course of incubation. The denitrification rate (0.071 mmol L −1 d −1 ) of NS1 was also lower as compared to the other two strains. No significant accumulation of N 2 O was detected 3 https://www.ncbi.nlm.nih.gov/ in any of these cultures, and the final concentrations of 0.42 to 0.45 mmol L −1 of N 2 indicated that 97.3 to 98.9 % of the initial NO 3 − -N was reduced to N 2 (Figure 1 and Table 4).
The microcosms that were prepared using denitrification intermediates as electron acceptors (NO 2 − and N 2 O) instead of NO 3 − also gave similar rates of N recovery as N 2 , with no N 2 O detected as an intermediate in NO 2 − amended cultures (Supplementary Figure 2). However, the rates of reduction of these intermediates were lower than those with NO 3 − . There was a lag period of 0.5 to 3 days before the Thauera strains initiated reduction of NO 2 − and N 2 O. On average, 0.42 ± 0.01 mmol L −1 of N 2 was produced in these cultures by the end of the incubation period (Supplementary Figure 2).
The bacterial cell growth characteristics and product yield coefficients for the three Thauera strains are summarized in Table 4. The strain K172 cultures consumed 0.86 mmol L −1 NO 3 − and grew to a highest maximum OD 600 of 0.119. In comparison, NS1 and TK001 grew to a lower maximum OD 600 of 0.095 and 0.042, respectively after similar amounts of NO 3 − consumption (i.e., concentrations of 0.85 and 0.91 mmol L −1 , respectively; Table 4). The rates of denitrification observed in the microcosms were inversely correlated with the exponential growth rates of these strains. The rate of NO 3 − reduction was highest in TK001 cultures with lowest exponential growth rate of 0.321 d −1 . Strain NS1 showed lowest rate of NO 3 − reduction but had higher exponential growth rate of 0.592 d −1 . Although the K172 cultures had a rates of NO 3 − reduction similar to that of TK001, their exponential growth rates were highest among the three strains (0.697 d −1 ; Table 4 and Supplementary Figure 3). The product yield coefficients from the cultures were 0.42 to 0.43 (Table 4). 4 | Microcosm parameters, microbial growth characteristics (first order growth rate), denitrification rate and N 2 production by denitrifying Thauera strains amended with NO 3 − and acetate in CSBK medium in 14-day microcosm incubations at 30 • C (Figure 1). a Numbers in parentheses are standard errors for three to four replicates. b % N 2 recovered = (Total N 2 produced in culture bottles/(0.5*Initial nitrate))*100. c Y P/S = (P t −P 0 )/(S 0 −S t ) where P 0 and P t are product concentrations at time 0 and final, respectively; S 0 and S t are substrate concentrations at time 0 and final substrate concentrations, respectively.

Effect of Initial NO 3 − Concentrations on Denitrification Products
All three bacterial strains were further cultivated on two to fivefold higher NO 3 − concentrations (1.90 to 4.60 mmol L −1 ; set 2; Tables 2, 5) with equimolar initial acetate concentrations. Varying the initial inputs of NO 3 − in culture bottles affected the denitrification rate and abundance of intermediate products.
Complete reduction of NO 3 − to N 2 was observed in all three cultures amended with 2.0 mmol L −1 initial NO 3 − . The percent N 2 recovered from these bottles ranged from 96.6 to 110.9 % ( Table 5). Denitrification was complete after 5 (K172), 60 (NS1) and 11 (TK001) days of incubation. Nitrite accumulated at higher concentrations (up to 1.46 mmol L −1 ) and persisted longer in the NS1 cultures as compared to the other two strains. The formation of N 2 O upon reduction of NO 2 − was not detected in cultures with K172 and TK001 while up to 0.23 mmol L −1 N 2 O formed upon reduction of NO 2 − in the NS1 cultures (Figure 2). On average, 0.98 ± 0.06 mmol L −1 N 2 was produced upon reduction of 1.88 ± 0.03 mmol L −1 NO 3 − in all cultures ( − was completely reduced and 0.44 ± 0.00 mmol L −1 N 2 O remained with the formation of 0.80 ± 0.18 mmol L −1 N 2 in all cultures ( Table 5).
The percentage of NO 3 − -N recovered as N 2 at the end of a 90-day incubation decreased as initial NO 3 − concentrations increased. The decrease was similar in cultures with K172 and TK001. A four to five-fold higher input of NO 3 − (e.g., than 1 mmol L −1 ) in these cultures gave 14.38 to 19.42 % lower N 2 ( Table 5 and Figures 1, 2). A larger decrease in N 2 production occurred with strain NS1, ranging from 40.74 to 57.85 %, and the decrease was observed at comparatively lower fold increase in initial NO 3 − inputs. Average initial NO 3 − concentrations of 3.65 and 4.59 mmol L −1 NO 3 − , produced NO 2 − , N 2 O and N 2 at the end of 90-day incubation in cultures with K172 and NS1. In the TK001 cultures amended with similar NO 3 − concentrations, only N 2 O and N 2 were formed (Table 5 and Figures 2C,F). Overall, the reduction of NO 3 − to N 2 was less complete when higher NO 3 − concentrations were added in the tested Thauera cultures.

Nitrite Inhibition of Thauera Strains as a Function of NO 2 − Concentrations
High concentrations of NO 2 − can be toxic to bacteria, thus inhibiting their metabolic activities. In order to investigate the effects of NO 2 − concentrations on the denitrification potential of the Thauera strains, two sets of kinetic experiments (both included in Set 3; Table 2) were conducted using NO 2 − as the sole electron acceptor when added at different concentrations (1 to 5 mmol L −1 ) and with equimolar initial concentrations of acetate as an electron donor (i.e., in 1:1 ratio). The patterns of NO 2 − reduction observed were similar in both the experiments. However, the rates of NO 2 − reduction were distinct among the three Thauera strains and different starting NO 2 − concentrations in batch cultures. With ∼1 mmol L −1 of initial NO 2 − , the reduction started after 2 (K172), 1 (NS1) and 0.5 (TK001) days of incubation of cultures (Supplementary Figure 2). The lag period before the initiation of NO 2 − reduction in these cultures further increased with the increase in NO 2 − concentration. For instance, it took 5 (K172), 3 (NS1) and 0.7 (TK001) days for Thauera strains to start reducing NO 2 − in cultures with five-fold higher NO 2 − (∼5 mmol L −1 ; Supplementary Figure 4). The NO 2 − concentrations had minimal effects on the NO 2 − reduction potential of strain TK001 in the beginning of incubation as compared to the other two Thauera strains. However, after 15 days of incubation there were no significant changes in the amounts of NO 2 − reduced in these cultures (Supplementary Figure 4C). Similarly, there was not much NO 2 − reduction in cultures with strain K172 after a similar incubation period (18 days; Supplementary Figure 4A). In comparison, levels of NO 2 − became more constant in cultures with strain NS1 in a shorter time of incubation (after 10 days; Supplementary Figure 4B). Continued incubation of all the Thauera cultures for 30 days did not show any additional substantive changes in NO 2 − reduction after these time periods.
Nitrite was completely reduced to N 2 in all the cultures with initial NO 2 − concentrations of ∼1 mmol L −1 a Numbers in parentheses are standard errors for three to four replicates. b % N 2 recovered = (Total N 2 produced in culture bottles/(0.5*Initial nitrate))*100. ND, not detected; detection limit for NO 3 − , NO 2 − and N 2 O is 0.001 mmol/L.

FIGURE 2 | Time series of NO 3
− , NO 2 − , and measured N 2 during denitrification with three different denitrifying Thauera strains (K172, NS1 and TK001) in batch cultures amended with two different initial NO 3 − concentrations of ∼ 2 mmol L −1 (N2; upper graphs) or ∼ 5 mmol L −1 (N4; lower graphs). Error bars represent the standard errors for three to four replicates. The dashed line indicates the N 2 concentrations calculated assuming decreases in initial NO 3 -N concentrations are completely reduced to N 2 . Figure 2). However, in cultures with higher initial NO 2 − concentrations complete reduction of NO 2 − to N 2 was observed only in cultures with strain K172 and 2 mmol L −1 NO 2 − (Supplementary Figure 4A). The percent reduction of NO 2 − decreased with an increase in the initial NO 2 − concentrations with strain NS1 being most affected (Figure 3A). With a 5-fold increase in NO 2 − concentrations, the percent NO 2 − reduction in cultures decreased to 49.21% (K172), 41.39% (NS1) and 43.93% (TK001). The reduction of NO 2 − produced only N 2 in the cultures with NS1 (Figures 3B,C). Nitrous oxide was not detected at any time point during incubation of these cultures. However, both N 2 O and N 2 were produced as reduced end products of NO 2 − reduction in cultures with K172 and TK001. More N 2 O (0.56 to 0.67 mmol L −1 ) was produced in  Table 4) produced through acetate oxidation coupled to denitrification by Thauera strains in batch cultures. Error bars represent standard errors for three to four replicates.

Effect of High pH on Completeness of Denitrification
pH is a measure of the hydrogen ion concentration in the aqueous phase and alkalinity is a measure of the capacity of aqueous phase in a contained system to neutralize acids. In anoxic closed carbonate system such as the culture bottles in our experiments, changes in dissolved inorganic species such as bicarbonates (HCO 3 − ) and carbonates (CO 3 2− ) upon acetate oxidation to CO 2(g ) cause changes in alkalinity and pH unless there are mineral buffers. At the end of batch experiments that were initially amended with ∼1 mmol L −1 acetate and NO 3 − , a decrease of 0.72 ± 0.03 mmol L −1 in acetate concentration was observed. Based on the redox stoichiometry (equation 5), 0.55 ± 0.02 mmol L −1 (76.4 ± 4.7%) of acetate was oxidized for complete reduction of NO 3 − to N 2 and the remaining 0.18 ± 0.06 mmol L −1 (25.0 ± 7.0 %) was presumably incorporated into bacterial biomass and other by-products (Figure 1 and Supplementary Figure 5; Chen et al., 2017).
Although carbon dioxide (CO 2(g) ) was not detected or detected at negligible concentrations by GC analysis, the CO 2 loss to the atmosphere is negligible given the sealed culture bottles used. This indicates that all CO 2 produced remained in the aqueous phase, where it would partition into the dissolved inorganic species. Accordingly, the pH of these cultures increased from 7.50 to 8.82 ± 0.08 (Figure 4). Using the measured pH values and dissociation constants, the total alkalinity was calculated to increase from 2.51 to 3.00 meq/L (Supplementary Material section 3). The measured alkalinity values (3.26 ± 0.16 meq/L) in these cultures upon completion of denitrification by Thauera strains were very similar to the calculated values and the reaction is well constrained for the lower initial NO 3 − concentrations (∼1 mmol L −1 ). The increase in pH and alkalinity of cultures with increasing initial NO 3 − concentrations is shown in Figure 4. The cultures amended with higher initial NO 3 − concentrations tended to not produce as high alkalinity as equation 5 would predict (Figure 4). The difference was highest at ∼ 5 mmol L −1 initial NO 3 − concentrations especially in the cultures with strains K172 and NS1. In comparison, at the similar initial NO 3 − concentrations, the difference was lower in the cultures with TK001 that showed comparatively more complete denitrification ( Table 4). This could be due to incomplete denitrification caused by high pH (up to 9.56), and less production of CO 2 in NS1 and K172 cultures.

N 2 O Reduction Kinetics as a Function of Increasing pH
The direct effect of initial pH on growth and reduction of N 2 O by Thauera strains was tested in batch cultures with 0.04 to 1 mmol L −1 N 2 O as the sole electron acceptor and 0.5 mmol L −1 acetate as the electron donor (set 4; Table 2). The average amounts of remaining N 2 O in the cultures after 15 days of incubation are summarized in Figure 5. The percent N 2 O is plotted against the initial N 2 O concentrations added to the cultures. The reduction of N 2 O to N 2 was not much affected at pH values of ∼8.1 where N 2 O was completely reduced to N 2 at all concentrations in cultures with Thauera strains except in cultures with 1.05 mmol L −1 of initial N 2 O and inoculated with strain NS1. Minimal amounts of N 2 O (0.02 ± 0.04 %) remained in these cultures ( Figure 5B). With the increase in initial pH, the remaining N 2 O that was not reduced to N 2 increased in the cultures when 0.79 to 1 mmol L −1 initial N 2 O was provided. The percent N 2 O remaining was 32.73 ± 10.81, 24.72 ± 5.14 and 68.97 ± 7.63 % in NS1 cultures at pH 9.00, 9.40, and 9.80, respectively. Nitrous oxide added at lower concentrations (0.04 to 0.10 mmol L −1 ) to these cultures was not detected at any of the pH conditions tested and was completely reduced to N 2 . At the highest tested pH 9.8, the percent N 2 O not utilized increased from 22.49 ± 7.45 to 68.97 ± 7.63 % with the increase in initial added N 2 O from 0.35 to 0.93 mmol L −1 (Figure 5B). Similarly, at pH 9.8, the percent N 2 O not utilized increased with the increase in initial N 2 O concentrations in cultures with strains K172 and TK001, however, the amounts of increases were lower as compared to those in cultures with strain NS1. The increase in remaining %N 2 O cultures was from 0.29 to 36.97 % (K172) and 0.58 to 21.18 % (TK001) at initial N 2 O concentrations ranging from 0.35 to 0.83 mmol L −1 (Figures 5A,C). No significant amounts of remaining N 2 O were detected in cultures with K172 and TK001 at pH 9.4 at any of the added N 2 O levels. However, up to 41.29 ± 12.15 % (K172) and 12.80 ± 4.63 % (TK001) N 2 O remained in cultures at pH 9.0 and initial N 2 O concentrations of more than 0.40 mmol L −1 . Overall, at higher pH, the percent reduction of high concentrations of N 2 O to N 2 was lower than at pH 7.5 and 8.1 (Figure 5).

Abundance of Denitrification Genes and N 2 Production Under High pH Conditions
The expression of denitrification genes (nirS and nosZ; equation 6) was tested using reverse transcription qPCR (RT-qPCR) analysis to determine the role of these genes in prolonged transient accumulation of denitrification intermediates (NO 2 − and N 2 O) under high pH conditions.

(6)
The reduction of NO 3 − , production of N 2 and expression of these genes was quantified in a separate experiment conducted at pH values identical to those measured in the cultures of denitrifying Thauera strains (Figures 4, 6). After 15 days of incubation, NO 3 − was completely reduced to NO 2 − in all the cultures except K172 cultures with pH of 9.42 and 9.87 (Supplementary Figure 6). The NO 2 − formed was further reduced completely in cultures with TK001 but not in cultures with K172 and NS1. The remaining NO 2 − concentrations were 5 to 6 fold lower (0.02 to 0.05 mmol L −1 ) in cultures inoculated with K172 than those with NS1 (0.13 to 0.26 mmol L −1 ). Nitrous oxide was not detected in any of the TK001 cultures throughout the incubation period, while minimal amounts of N 2 O ( < 0.001 mmol L −1 ) were produced in cultures with NS1 and pH 9.88. Comparatively higher amounts of N 2 O (0.02 to 0.24 mmol L −1 ) accumulated in cultures with K172. The concentrations of NO 2 − and N 2 O in cultures of both these strains (K172 and NS1) correlated with their pH values. Increased pH values were associated with increased concentrations of these intermediates detected in these cultures.
For quantification of nirS and nosZ genes, the 16S rRNA gene was used as a control and their relative abundances at higher pH values were compared to those at optimum pH conditions using the 2 − CT method (Schmittgen and Livak, 2008). The standard curves for qPCR calibration were linear (R 2 = 0.99) for all the three genes with amplification efficiencies between 96.8 to 102.6 % ( Table 3). The amounts of N 2 produced decreased with an increase in pH of cultures with the three Thauera strains; up to a 2.40, 1.66 and 1.48-fold decrease was observed for cultures inoculated with K172, NS1 and TK001, respectively as compared to the cultures with optimum pH (Figure 6). The decrease was most evident in the cultures exhibiting pH values between 9.11 and 9.90. The RT-qPCR analysis of nirS and nosZ transcripts supported this decrease of N 2 production in batch cultures with K172 ( Figure 6A). The nirS transcription levels in these cultures were 5.79 to 7.72-fold lower than the control cultures except for the cultures having pH values of 9.42 and 9.87. FIGURE 6 | Changes in production of N 2 and expression of denitrification genes (NirS and NosZ) as a function of increasing pH in batch cultures of denitrifying Thauera strains (K172, NS1 and TK001) amended with 1 mmol L −1 of NO 3 − and acetate. The red star (*) indicates optimum pH-7.5 taken as equivalent to 1 for comparison with high pH conditions. Error bars represent standard errors for three replicates.
The abundance of nirS in these cultures increased by 1.40 to 1.59fold as compared to its abundance at optimum pH conditions. The higher pH levels more greatly impacted the transcription levels of nosZ as its abundance decreased by 7.13 to 20.38-fold in these cultures ( Figure 6A). Note that the samples for qPCR from K172 cultures were taken during the exponential growth phases; t = 2 and t = 5 to 15 days for optimum pH and higher pH conditions, respectively.
The bacterial cells for RT-qPCR were collected at the same time, t = 5 and t = 15 days, for cultures inoculated with strains TK001 and NS1, respectively. Both nirS and nosZ were highly expressed in the cultures with NS1 and TK001 at all high pH conditions despite less N 2 production (Figures 6B,C). For culture NS1, the nirS transcription levels were 1.77 to 6.86-fold higher in the cultures containing increased pH compared to the optimum pH control. The nosZ transcription levels also followed the similar pattern except in cultures having pH values of 7.77 and 8.56. In fact, the abundance of nosZ decreased by 1.10 to 1.39-fold in these cultures ( Figure 6B). In all TK001 cultures incubated under all conditions, nosZ was more abundantly transcribed than nirS. However, the abundance of both these genes was higher under higher pH conditions than at the control pH. The transcription levels of nirS and nosZ were 1.61 to 2.44 and 1.83 to 3.30-fold higher, respectively, in these cultures ( Figure 6C). The observed difference in the expression patterns and transcription levels of nirS and nosZ in cultures with three Thauera strains may be attributed to differences in their denitrification kinetics, sensitivity to pH stress conditions, and sampling times for analysis.

Type and Concentration of N-Sources Affect Denitrification Regulatory Phenotypes
This study focused on the effects of physico-chemical parameters of the microbial growth environment on the metabolic activities of denitrifying bacterial isolates affiliated to the genus Thauera, namely T. aromatica K172, T. aromatica NS1 and T. aminoaromatica TK001. The biokinetic experiments conducted at 30 • C and under anoxic conditions showed denitrification differences between these strains. Strains of Thauera previously studied were grouped into two categories based on the observed differences in their denitrification regulatory phenotypes (DRPs; Liu et al., 2013). The strains were characterized either by rapid, complete onset (RCO) or by progressive onset (PO) of denitrification genes linked to NO 3 − reduction and accumulation of intermediates like NO 2 − (Liu et al., 2013). At lower NO 3 − concentrations (∼1 mmol L −1 ), strains K172 and TK001 displayed an RCO type of DRP with nondetectable NO 2 − and N 2 O accumulation during denitrification coupled with acetate oxidation (Figures 1A,C). However, strain S2 studied by Liu et al. (2013), and similar to strain TK001, was reported to have a PO type of DRP upon reduction of NO 3 − (2 mM per 50 mL) coupled to acetate oxidation. The observed differences between the two T. aminoaromatica strains (TK001 and S2) in these two independent studies seem to be less related to similar culture conditions used than their reported metabolic activities. Even though there are high 16S rRNA gene sequence similarities (99% sequence identity), considerable phenotypic differences are known to exist between these two strains (Mechichi et al., 2002;Fida et al., 2016). The type strain K172 that showed similar denitrification phenotypic traits as TK001 in this study was also revealed to be genotypically different in its protein profile compared to strain S2 (Mechichi et al., 2002). The variation in observed DRPs of these strains could possibly be related to the expression patterns of proteins and abundance of denitrification genes (Mechichi et al., 2002;Liu et al., 2013). In contrast, strain NS1 exhibited similarities to a PO type of DRP and transient accumulation of NO 2 − resulting from reduction of NO 3 − ( Figure 1B). This strain is phylogenetically more related to strains K172 and S100 ( > 90% 16S rRNA gene similarity; Suri et al., 2019) but unexpectedly showed lower denitrification rates as compared to the type strain K172 under identical denitrifying culture conditions. At higher NO 3 − concentrations ( > 2 mmol L −1 ), the denitrification kinetics of the three Thauera strains became more similar showing a PO type of DRP and prolonged accumulation of NO 2 − and N 2 O (Figure 2 and Table 4), even at an ideal temperature (30 • C). Suri et al. (2017) reported incomplete NO 3 − reduction and persistence of relatively high concentrations of NO 2 − formed through reduction of similarly high NO 3 − concentrations under these conditions. The toxicity posed by produced NO 2 − to dominant Thauera and Pseudomonas species was thought to be the reason but was not evaluated. In the present study, the NO 2 − accumulated to maximum concentration of 3.2 mmol L −1 in microcosms ( Table 4). The exposure of Thauera strains to NO 2 − concentrations (0.9 to 5.2 mmol L −1 ), even higher than the accumulated concentrations (0.01 to 3.19 mmol L −1 ), did not appear to be inhibitory (Supplementary Figure 4). Instead, the denitrification rates were slower and ceased after reduction of 1.9 mmol L −1 NO 2 − . The use of NO 2 − as compared to NO 3 − as a sole initial electron acceptor had a variable impact on the DRPs. Strains TK001 and NS1 behaved similarly showing a RCO type of DRP but K172 showed a PO type of DRP. Strain K172 only partially reduced NO 2 − to N 2 producing N 2 O as a product in addition to N 2 (Supplementary Figures 2, 4 and Figure 3). In contrast, Liu et al. (2013) showed N 2 O produced as the only end product by T. phenylacetica characterized to have PO type of DRP. Considerable genotypic differences exist between these tested species, namely T. aromatica and T. phenylacetica, that appeared to be a reason for observed differences in their DRPs (Suri et al., 2019).

Increased pH Affects Denitrification End Products
The impaired reduction of NO 3 − and the produced intermediate NO 2 − by Thauera strains under the influence of increasing temperature and decreasing electron donor concentrations has previously been shown in the laboratory (Fida et al., 2016;Suri et al., 2017). At temperatures above 45 • C and at NO 3 − concentrations equivalent to the 1 mmol L −1 concentration used in this study, denitrification halts at NO 2 − (Reinsel et al., 1996;Greene et al., 2003;Fida et al., 2016;Okpala et al., 2017). However, subsequent complete reduction of NO 2 − to N 2 occurs at temperatures below 45 • C provided sufficient amounts of preferable electrons donors are available (Voordouw et al., 2009;Fida et al., 2016;Okpala et al., 2017). Similar patterns of NO 2 − persistence during NO 3 − reduction can also occur with an increase in salinity from 0.5 to 2.5 M . Here, we additionally demonstrate that pH changes, particularly alkalinity, can cause substantive impact on denitrification. The assays performed demonstrate that the denitrification product ratios (NO 2 − : N 2 O: N 2 ) can vary with the variation in pH (Figure 4 and Table 5).
Higher pH conditions (pH ≥ 9) had a greater impact on the production and reduction of NO 2 − and N 2 O compared to the use of high NO 3 − concentrations ( > 2 mmol L −1 ) by Thauera species. Such a pH effect has also been reported in other studies. At a pH of 8 and higher, the growth of Dechloromonas aromatica strain RCB was inhibited and there was accumulation of substantial amounts of N 2 O during NO 3 − reduction (Han et al., 2019). In contrast, transient to permanent accumulation of NO 2 − and N 2 O have been observed at neutral to acidic pH conditions. For instance, rates of denitrification in Thauera, Pseudomonas and Paracoccus cultures were lower, with increasing accumulated concentrations of NO 2 − and N 2 O at pH values of 6.0 to 7.5 (Bergaust et al., 2010;Suri et al., 2017;Kim et al., 2017). In this present study, the accumulation of NO 2 − at pH 7.5 was observed only for strain NS1 but was further reduced to N 2 (Figure 1).
Microbial metabolisms vary with the type and concentration of electron donor, which can substantially affect pH and alkalinity (Gallagher et al., 2012). Acetate is a highly degradable electron donor readily available in many natural environments and its use in this study contributed to pH increase (Figure 4). Denitrifiers like Thauera and Pseudomonas also possess abilities to use other organic electron donors for driving their metabolism and energy processes (Agrawal et al., 2012;Suri et al., 2017;An et al., 2017). The effect of pH increase on NO 3 − reduction could be mitigated by using electron donors that do not cause pH shifts upon oxidation. Nonetheless, the use of acetate in the current study helped to gain initial knowledge on the overall variable effect of pH on the outcomes of denitrification.

Increased pH Affects Denitrification Enzyme Activity
The key enzymes that catalyze the reduction of NO 2 − to N 2 O and of N 2 O to N 2 are nitrite reductase and nitrous oxide reductase, respectively. These enzymes are encoded by the nirS (nitrite reductase) and nosZ (nitrous oxide reductase) genes in many Thauera species and are intracellular but can undergo considerable changes in response to fluctuations in pH and other environmental conditions (Zumft, 1997;Wilks and Slonczewski, 2007;Fida et al., 2016). The mechanisms can include altered enzyme synthesis, inhibition of enzyme function and/or reduced or delayed transcription of denitrification genes (Bergaust et al., 2010;Liu et al., 2010;Fida et al., 2016). Our results indicated a correlation of high pH with reduced transcription of the nirS (at pH 7.8-9.0) and nosZ (at pH 7.8-9.9) genes resulting in less complete denitrification and lower N 2 production in cultures of strain K172. Unexpectedly, even though the transcriptional levels of nirS were higher at pH ≥ 9, the amount of N 2 O produced from reduction of NO 2 − was much lower than that produced at pH ≤ 9 ( Figure 6A, and Supplementary Figure 6A). Nitrate was not completely reduced in these cultures. Since NO 3 − is a preferred electron acceptor compared to NO 2 − and N 2 O, one possible explanation for our observation is that nitrate reductase (encoded by the nar gene and not quantified in this study) was still active and can compete more efficiently for electrons than nirS, thus causing partial denitrification (Almeida et al., 1995;Liu et al., 2013). The protons from the inside of cytoplasmic membrane are used for NO 3 − reduction while those from the periplasmic side of cytoplasmic membrane are used for NO 2 − reduction. At pH ≥ 9, protons may be comparatively scarce in the periplasm than at pH ≤ 9, thus resulting in inhibition of enzyme activity of nitrite reductase even though nirS transcriptional levels were high (Meijer et al., 1979;Glass and Silverstein, 1998). In contrast, the transcriptional levels of nirS increased only after depletion of NO 3 − in cultures of Thauera terpenica (Liu et al., 2013).
The transcriptional levels of nirS and nosZ were similar in cultures of strain TK001 at a tested pH, which is characterized to have the RCO type of DRP (Figures 6C, 1). The levels and expression patterns of the nirS and nosZ genes in cultures of Thauera sp. 63 with an RCO type of DRP were also shown to be similar during denitrification at pH 7.5 (Liu et al., 2013). The transcriptional levels of both these genes increased similarly at the beginning of denitrification, reached similar maximum levels and then decreased at the end of denitrification, however, the abundances of nosZ transcripts were sustained at higher levels than nirS transcripts, similar to our experiments (Liu et al., 2013; Figure 6C). Both nirS and nosZ were expressed simultaneously at each pH; however, the timing of maximum transcription was delayed with the increase in pH of cultures. This observation suggested that a decrease in the rates of denitrification at pH > 7.5 were caused by slower enzyme activity in comparison to that observed at pH 7.5, as was evident from lower N 2 production at the incubation time assayed (t = 5 days; Figure 6C). Similar patterns of increased transcriptional levels of nirS and nosZ with the increase in pH were observed in NS1 cultures. Nitrite accumulated in these cultures, an expected effect of the PO type of DRP of this strain (Supplementary Figures 1, 6B and Figure 6). The temporary shut down of electron flow to nitrite reductase, deficiency of co-factors such as Cu required for its activity, inhibition of translation of its mRNA and/or repression of its activity at high pH and high NO 2 − concentrations can also cause NO 2 − to accumulate (Stouthamer, 1991;Baumann et al., 1997;Granger and Ward, 2003). Since there was sufficient electron donor (acetate) present, the observed inhibited NO 2 − reduction can possibly be explained by inhibition of nitrite reductase activity and DRP of the strain. The minimal N 2 O accumulation observed can be explained by the DRP of this strain, with lower transcription levels of nosZ and less N 2 O production by NO 2 − reduction (Figure 1 and Supplementary Figure 6). It has been suggested in previous studies that the accumulation of N 2 O during denitrification depends on the difference in rates of its production and reduction, but pH increase only lowered its reduction rate without any significant effect on its production upon NO 2 − reduction (Betlach and Tiedje, 1981;Liu et al., 2010;Han et al., 2019). In our study, the pronounced decrease in N 2 O reduction at concentrations, 0.69 to 0.93 to mmol L −1 and pH 9.8 can explain the observed accumulation of N 2 O (Figures 3, 5). In contrast, transient to permanent N 2 O accumulation at slightly acidic conditions (pH 6.0) has been repeatedly demonstrated for denitrifiers like Paracoccus denitrificans and Shewanella loihica (Bergaust et al., 2010;Liu et al., 2014;Kim et al., 2017). Variable patterns of reduced affinities of nosZ to N 2 O and reduced copy numbers of nosZ can explain the relevant decrease in its reduction rates. The transcription of nosZ was unaffected at optimal pH while it was lower at alkaline pH during denitrification in other studies (Liu et al., 2014;Kim et al., 2017;Han et al., 2019). Nitric oxide, not analyzed in our studies, has known toxicity to many microorganisms and can inhibit the activity of key enzymes including nitrous oxide. Based on the results obtained in this study, the observed decrease in the rates of denitrification by Thauera strains can be overall explained by inhibition of denitrification enzyme activities caused by lower or delayed transcription of genes encoding these enzymes and reduced affinity of these enzymes to substrates at tested high pHs.

Implications of Increased pH on Industrial Applications
Other researchers have shown the importance of thermophilicity, salinity and electron donors in NO 3 − -mediated industrial applications (Fida et al., 2016;An et al., 2017;Chen et al., 2017). This study provides information on the impact of N sources, their initial concentrations, and pH. The observed accumulation of NO 2 − and N 2 O upon NO 3 − reduction at increased pH is desirable in control of sulfide formation in soured oil fields. Although the results from our experiments with NO 2 − as a sole electron acceptor suggested that the use of NO 2 − to inhibit SRB would be preferable to the use of NO 3 − , this may not be favorable in other applications, such as wastewater treatment. In such cases, this reaction can readily cause damage to other beneficial microorganisms and aquatic life through increasing NO 2 − concentrations over time or discharge of NO 2 − rich waters in different aquifers (Philips et al., 2002;Zhou et al., 2011). Also, in applications such as MEOR, the reduction of NO 2 − to N 2 is required for enhanced oil production (Gassara et al., 2015;Suri et al., 2019).
Based on our findings, we hypothesize that NO 3 − injections require careful adjustment of pH (e.g., liming) of oil reservoir waters to an alkaline pH range that could maintain a longer presence of denitirificaiton intermediates (NO 2 − and/or N 2 O) for more effective souring control even at mesophilic conditions. On the other hand, maintaining optimal pH levels to facilitate complete reduction of NO 3 − to N 2 for activating maximum dissolved and free phase gas pressure for applications like MEOR could be managed by careful selection of N-sources and/or pH buffering.
The microbial communities in these applications are often complex and consist of actively interacting and competing bacterial species. Since the knowledge obtained here is based on studies of a single genus and a limited number of model strains, the suggested roles of these parameters in practical applications needs to be tested rigorously in laboratory to obtain a further proof-of-concept. More information is needed on the regulatory controls on exhibition of denitrification outcomes by wide range of denitrifiers from different ecosystems. Our studies may not be able to account for variations in denitrification outcomes as a response to changed physiological conditions in these scenarios. However, results of this work provide a solid basis for future research and long-term evaluations of correlations between N-source, alkalinity, pH, and end products of denitrification in various industrial systems.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

AUTHOR CONTRIBUTIONS
NS planned and conducted the experiments, collected, analyzed, and interpreted the data, drafted and revised the manuscript. YZ helped in conducting experiments, analyzing data and revising the manuscript. LMG supervised the work through ideas and discussions and revised the manuscript. MCR supervised the work, discussed obtained data, revised and approved the manuscript to be published. All authors contributed to the article and approved the submitted version.

This work was supported by a Mitacs-Kerui grant awarded to MCR and Global Research Initiative-Canada First Research
Excellence Award awarded to the University of Calgary. An NSERC Collaborative Research and Development grant awarded to LMG also partially supported this research.