Influence of copper on expression of nirS, norB and nosZ and the transcription and activity of NIR, NOR and N2 OR in the denitrifying soil bacteria Pseudomonas stutzeri

Summary Reduction of the potent greenhouse gas nitrous oxide (N2O) occurs in soil environments by the action of denitrifying bacteria possessing nitrous oxide reductase (N2 OR), a dimeric copper (Cu)‐dependent enzyme producing environmentally benign dinitrogen (N2). We examined the effects of increasing Cu concentrations on the transcription and activity of nitrite reductase (NIR), nitric oxide reductase (NOR) and N2 OR in Pseudomonas stutzeri grown anaerobically in solution over a 10‐day period. Gas samples were taken on a daily basis and after 6 days, bacterial RNA was recovered to determine the expression of nirS, norB and nosZ encoding NIR, NOR and N2 OR respectively. Results revealed that 0.05 mM Cu caused maximum conversion of N2O to N2 via bacterial reduction of N2O. As soluble Cu generally makes up less than 0.001% of total soil Cu, extrapolation of 0.05 mg l−l soluble Cu would require soils to have a total concentration of Cu in the range of, 150–200 μg g−1 to maximize the proportion of N2O reduced to N2. Given that many intensively farmed agricultural soils are deficient in Cu in terms of plant nutrition, providing a sufficient concentration of biologically accessible Cu could provide a potentially useful microbial‐based strategy of reducing agricultural N2O emissions.


Introduction
Since the industrial revolution, global agricultural intensification and the use of artificial fertilizers has increased the amount of reactive nitrogen (N) in the natural environment by an order of magnitude (Galloway et al., 2003;Richardson et al., 2009;Taylor and Townsend, 2010). This has resulted in a reorganization of the global N cycle causing a number of environmental problems including increased nitrous oxide (N 2 O) emissions (Taylor and Townsend, 2010;Magalhaes et al., 2011).
Nitrous oxide accounts for~10% of the total greenhouse gas emissions and is produced as a by-product of bacterial and fungal respiration pathways in soil. Both denitrification and nitrification respiratory pathways emit N 2 O with rates increasing with the addition of N fertilizer (Taylor and Townsend, 2010;Magalhaes et al., 2011). Denitrification occurs when oxygen (O 2 ) is in limited supply and bacteria with denitrifying capability can switch to respiring nitrate (NO 3 À ), converting NO 3 À to nitrite (NO 2 À ) and the gases nitric oxide (NO) and N 2 O and finally dinitrogen (N 2 ) ( Fig. 1). This process requires four enzymes to sequentially reduce NO 3 À to N 2 with each of these enzymes requiring a redox metal cofactor (Fig. 1). Denitrifying soil bacteria such as Pseudomonas stutzeri generate N 2 O via the reduction of NO, an endogenous cytotoxin, by reducing N 2 O via the enzyme nitric oxide reductase (NOR), hence bacteria deficient in NOR cannot grow through denitrification (Zumft, 2005a,b). Because so much N 2 O is produced from soils carrying out bacterial denitrification, it implies that the bacterial enzyme nitrous oxide reductase (N 2 OR), or the bacterial population as a whole, do not always carry out the reduction of N 2 O to N 2 efficiently or in synchrony with pathways upstream (i.e. nitrifier-denitrification) (Richardson et al., 2009). Thus, managing N 2 O emissions requires consideration of the factors affecting the production of N 2 O at both the molecular and soil microbial ecology levels. Nitric oxide reductase is a key enzyme in the production of N 2 O and is responsible for catalysing the reduction of NO to N 2 O (Fig. 1, equation 3). Under intensive pasture management, N 2 O is the predominant endproduct during the denitrification process, which is a main contributor to global N 2 O emissions through the agricultural nitrogen cycle (Mosier et al., 1998). The structure of NOR has been well studied and is known to be a member of the haem-copper oxidase superfamily . The homology analogue of this protein taken from Paracoccus denitrificans, Pseudomonas aeruginosa, Bacillus azotoformans and Ralstonia eutropha were also comprehensively characterized (DeBoer et al., 1996;Pohlmann et al., 2000;Suharti et al., 2001) P. aeruginosa. NOR is expressed by P. stutzeri under anaerobic conditions (Korner, 1993) and is found exhibiting as a complex consisting of several components including high-spin cytochrome b, low-spin Fe(III) haem centre and a cytochrome c subunit or short-chain NOR acting as the electron donor at the active centre (Zumft, 2005a,b). Although the majorities of NORs reported consist of iron-binding site, remarkable structural similarities was found between cbb3 cytochrom oxidase and NOR which contains a high-spin haem-copper (CuB) active site (Vanderoost et al., 1994). Furthermore, there is emerging evidence that NOR from Bacillus azotoformans to be a hybrid between copper-containing cytochrome oxidases and NOR found in Gram-negative bacteria (Suharti et al., 2001;Al-Attar and de Vries, 2015). Due to the similarities between cbb3 cytochrom oxidase and NOR, and lack of crystallography structure evidence, it can be questionable whether NOR from P. stutzeri is also regulated by copper metal cofactor. Moreover, Cu-dependency of NOR from P. stutzeri has not been investigated before.
The reduction of N 2 O to N 2 is strongly exergonic (Fig. 1, equation 4) (ΔG°' = À339.5 kJ mol À1 ), and electron delocalization stabilizes the molecule and leads to an activation energy barrier of 250 kJ mol À1 (Tolman, 2010). However, N 2 OR, first identified in 1982 , is the only known enzyme capable of reducing N 2 O to N 2 during the denitrification process (Zumft and Kroneck, 2007). The N 2 OR enzyme is a Cu-dependent enzyme with a three-dimensional structure revealing a multi-Cu sulfide centre Cu z [4: Cu:2S] where the two-electron reduction of N 2 O takes place (Richardson et al., 2009). The Cu requirement for the active dimeric form of N 2 OR requires the bacterium to have an adequate supply of Cu. An absence of Cu in some culture studies has resulted in a rise in N 2 O emissions (Granger and Ward, 2003). While a Cu-deficient denitrifying bacterial community can still remain viable, it is likely that they will release much higher levels of N 2 O. Several studies have documented the central role of Cu in enzyme structure and function Farrar et al., 1998;Zumft, 2005a,b;Pomowski et al., 2010); however, the effects of Cu on the expression of these genes nirS, norB and nosZ to increase Cu bioavailability have yet to be described. We hypothesize that denitrifying bacteria have the ability to sense Cu availability and adjust the synthesis of N 2 OR in response to environmental conditions. In this study, we examine the effects of increasing Cu concentrations on the expression and activity of nitrite reductase (NIR), NOR and N 2 OR in P. stutzeri and the potential implications for molecular and stoichiometric influences on N 2 O emissions from agricultural soils.

Results and discussion
The effect of increasing copper concentrations on nitrous oxide accumulation and conversion into dinitrogen in anaerobic cultures of Pseudomonas stutzeri Growth of P. stutzeri from anaerobic cultures over 7-day period showed no significant difference throughout all Cu concentration, except 5.00 and 20.00 mM of Cu (Fig. 1A), and the growth reached stationery phase at day 3 with ≤ 1.00 mM of Cu. Colony counts of P. stutzeri cultures containing 5.00 mM Cu were significantly reduced compared with those in Cu concentrations < 5.00 mM with a delay in growth after day 3 (Fig. 1A). This could indicate the onset of toxicological effects, although denitrification was occurring at relatively high rate (Fig. 2B). As expected, given P. stutzeri the highest concentration of Cu (20.00 mM) inhibited the growth completely (result not shown). Daily measurements revealed production of N 2 O and subsequent reduction to N 2 varied significantly with changes in Cu concentrations (Fig. 2B). Interestingly, the Cu-deficient broth (control, 0.00 mM) did not produce the highest concentrations of N 2 O compared with the Cu-containing treatments as has been suggested in a previous study to be the most likely outcome (Granger and Ward, 2003). Furthermore, a comparison between the total amount of N 2 produced per Cu treatment revealed no significant differences. In fact, the highest amount of N 2 O produced was observed in the culture containing 0.15 mM, which corresponded Over the 10-day sampling period (apart from 0.50 mM Cu treatment), treatments ≥ 0.15 mM Cu resulted in a significantly higher (P < 0.001) total yield of N 2 O than N 2 . Studies that have investigated the effect of Cu addition on the conversion rate of N 2 O to N 2 concluded that the absence of Cu resulted in an accumulation of N 2 O (Granger and Ward, 2003;Manconi et al., 2006;Felgate et al., 2012). However in these studies, the Cu concentration at which N 2 OR maintained an optimum activity and where the lowest proportion of N 2 O to N 2 levels were generated, were not evaluated in further detail. Additionally, the presence of sulfide, especially H 2 S in a Cu-deficient environment has been known to affect the reduction of N 2 O to N 2 (Manconi et al., 2006;Pan et al., 2013). Given that in our study, S 2 À and H 2 S could have been generated during anaerobic incubation of the broth from CuSO 4 (source of the Cu) and cysteine, this may have also contributed to the accumulation of N 2 O.
The effect of copper availability on the expression of nirS, norB and nosZ From the Cu treatments it emerges that a Cu concentration of 0.15 mM is the level at which nirS, norB and nosZ expression are at their highest (Fig. 4), although N 2 O level and the mean N 2 O:(N 2 +N 2 O) ratio was at the highest of all Cu concentration given. Our results show inhibition of nosZ transcription from a Cu concentration of 0.50 mM onwards. Interestingly, even though the nirS enzyme contains a haem-iron cofactor, it appears the transcription of nirS is responsive to changes in Cu con-centrations ( Fig. 4). Furthermore, norB expression was also declined with increased Cu concentration. Similarly, in a previous study conducted by Magalhães et al. (2011), a pronounced inhibition by Cu on the transcription of nosZ and nirS was detected. In the treatment containing 0.15 mM of Cu, the N 2 O:(N 2 +N 2 O) ratio was elevated as well as the observation of comparatively high levels of denitrifying genes expressed. This suggests that there could exist an alternative pathway for P. stutzeri to synthesize N 2 O at higher levels of Cu; thus, the level of N 2 O produced due to NOR activity could be overestimated in this study. X-ray crystallographic structures of the periplasmic membrane protein cbb3 cytochrome oxidase of P. stutzeri revealed a highspin haem-copper (CuB) binuclear centre of which the catalytic active site was also found reducing NO to N 2 O during denitrification process via the proton pathway through K-channel (Forte et al., 2001;Buschmann et al., 2010). Increased Cu concentration might also contribute in elevating the N 2 O level via NO reduction by cbb3 cytochrome oxidase in this study. Therefore, expression of genes involved in cbb3 cytochrome oxidase (ccoQ-NOS) should further be investigated. This may provide a more detailed insight of N 2 O formation regulated by Cu from an alternative pathway. The enzyme N 2 OR is considered as the only protein catalysed by Cu due to the multi-Cu-sulfide redox centres (Pomowski et al., 2010). Surprisingly, expression of nosZ did not respond to the incensement of Cu concentration in this study, instead, the transcriptional level of nosZ decreased tremendously at 0.50 mM of Cu onward (Fig. 4). This suggested high level of Cu could be displaying inhibitory effect. This result differs with a study conducted by Felgate et al. (2012) where transcription of nosZ of P. denitrificans was found upregulated at high Cu concentration (13 lmol l À1 which is equivalent to 2 mM). the Cu concentrations used were an order of magnitude lower than the lowest concentrations used in this study and atypical of bioavailable Cu concentrations found in agricultural soils (Black et al., 2011), results demonstrated that if sufficient NO 3 À was present when Cu was depleted, P. denitrificans would maintain biomass, but release N 2 O at a rate > 1000 times the rate of the corresponding Cu replete cultures. It was also noted that the N 2 O electron acceptor lost under Cu-depleted conditions was compensated for by increased (%20%) consumption of NO 3 À compared with cultures replete in Cu. This observation of N 2 O being produced more quickly but consumed more slowly is mirrored when comparing the control and 0.02 mM Cu treatment concentration ( Fig. 2A and B), although nitrate consumption did not increase. The N 2 O:(N 2 O+N 2 ) trend observed in this study reflected observations made in a denitrifying community sourced from river sediments which were enriched in Cu due to agriculture run-off (Magalhaes et al., 2011). The suggested reason for the observed pattern was the different sensitivities of each of the enzymes that catalyse the first three steps of denitrification (NO 3 À ? NO 2 À , NO 2 À ?NO À and NO À ? N 2 O) (Magalhaes et al., 2011). Similarly, the authors found that denitrifcation rates were highly affected by Cu  concentrations, moreover they also observed a decrease in the diversity of nirK, nirS and nosZ and the corresponding transcribed enzymes with increasing Cu. Results from this study suggest that Cu bioavailability can influence the expression and activity of N 2 OR and NIR, as well as the growth rates of P. stutzeri and imply that it may be possible to use stoichiometry to manage N 2 O emissions from agricultural soils. This approach has been previously suggested as a possible strategy to mitigate N 2 O emissions by providing an adequate source of essential micronutrients particularly Cu, Mo and Fe for redox reactions (Richardson et al., 2009). Although the reduction of N 2 O to N 2 is not energetically favoured under optimal conditions, this reduction does occur at no loss of energy requirements to the bacteria. Moreover, our results confirm that there indeed exists an optimal Cu concentration threshold for P. stutzeri with respect to maximizing N 2 O consumption.

Conclusion
In summary, our findings do support a key role for Cu in the regulation of N 2 O emissions, by demonstrating a Cu concentration gradient effect in the production and consumption of N 2 O. While total soil Cu is still the most common soil measure of bioavailability, bioavailable Cu is typically < 0.001% of total soil Cu (Bolan et al., 2003;McLaren et al., 2010;Black et al., 2011). Thus, values of 0.02-0.15 Cu mM used in this study are more realistic representations of biologically available Cu in agricultural soil environments to consider for potential management strategies of N 2 O emissions (Richardson et al., 2009). Extrapolating the Cu concentrations used in this study into an agricultural setting imply in some situations, current soil levels of Cu maybe deficient to allow this enzymatic pathway to operate at an optimal level. Calculating back from total soil Cu concentrations to biologically available Cu (0.05-0.10 mM Cu in solution) that is required to attain maximum conversion of N 2 O to N 2 , equates to a total amount of soil Cu in the range 150-200 lg g À1 (Black et al., 2011). Soils containing higher concentrations of Cu may result in inhibition of denitrification via a shift in denitrifier community composition (Manconi et al., 2006;Taylor and Townsend, 2010;Felgate et al., 2012) and correspondingly, Cu-deficient soils may result in N 2 O production exceeding N 2 O conversion to N 2 .
Using one species of denitrifying bacteria in a simple basal salt solution has provided some insight into the effects of increasing Cu bioavailability on the transcription and activity of N 2 OR and NIR to reduce N 2 O to N 2 . We observed that adequate bioavailable Cu concentrations (0.15 mM) resulted in the greatest transcription of the nirS, norB and nosZ, which is not the optimal consumption of N 2 O to produce N 2 . However, the level of N 2 O may be overestimated due to another possible alternative pathway of NO reduction by cbb3 cytochrome oxidase. Based on our experience with measuring amounts of bioavailable soil Cu, extrapolation of this soluble Cu concentration to an agricultural soil environment equates to total soil Cu concentrations in the range of 150-200 lg g À1 . Furthermore, a majority of soils globally under intensive agriculture, are now considered too deficient in Cu to complete the last stages of the denitrification process to N 2 (Skrbic and Durisic-Mladenovic, 2010). Therefore, in addition to other mitigation methods, supplying adequate Cu would contribute to an overall effective management of N 2 O emissions in increasing N intensification systems. Future studies will need to investigate if similar Cu concentrations may apply to other soil denitrifying bacteria, including investigating the possibility of adopting a Cu strategy for whole soil microbial communities. However, any strategy employed would have to consider any potential toxicity issues for grazing animals.

Experimental procedures
Cultivation conditions of P. stutzeri (ATCC 17588) The denitrifying bacteria P. stutzeri (ATCC 17588) was chosen due to the presence of nirS (Fe-cofactor) instead of nirK (Cu-cofactor). Bacteria were first grown in cultures under aerobic conditions. The basal salt solution of the aerobic medium comprised the following (mM): 1.5 KH 2 PO 4 , 5.6 NH 4 Cl, 0.5 MgCl 2 Á6H 2 O, 0.7 CaCl 2 Á2H 2 O, 0.4 MgSO 4 Á7H 2 O and 30 KNO 3 . The solution was autoclaved and after cooling 99% D-Na lactate (Sigma Aldrich, St Louis, Missouri, United States) was added to a final concentration of 31 mM in 100 ml volume (Johnsson et al., 2006). The aerobic culture was grown in 100 ml solution contained per 250 ml Erlenmeyer flasks for 4 days before 1 ml of culture-containing solution, containing approximately 10 6 bacterial cells, was transferred into the anaerobic Cu medium experiment.
The basal salt solution of the anaerobic medium comprised (mM) 0.06 KH 2 PO 4 , 5.6 NH 4 Cl, 0.5 MgCl 2 Á6H 2 O, 0.8 0.7 CaCl 2 Á2H 2 O, 0.4 MgSO 4 Á7H 2 O and 11 mM of 15 N-enriched KNO 3 (10 atom%). The pH was adjusted to 8 using 1 M NaOH, with sufficient KH 2 PO 4 in solution to provide buffering capacity (Johnsson et al., 2006). Cu sulfate (CuSO 4 ) was added to each gas-tight medical boston bottle at concentrations equivalent to 0.00, 0.02, 0.05, 0.15, 0.50, 1.00, 5.00, 20.00 mM of Cu. The basal solution (100 ml per bottle) with Cu treatments was autoclaved and 99% D-Na lactate and cysteine-HCl (Sigma Aldrich) were added to produce final concentrations of 4.40 and 4.50 mM respectively. Copper treatments were performed in triplicate and each treatment replicate was inoculated with 1 ml of 10 6 bacterial cells before being fitted with gas-tight Teflon â septa and purged using 99.99% Argon (Ar) (remaining 0.01% comprising CO 2 < 5 ppm, N 2 < 25 ll l À1 , O 2 < 10 ll l À1 , H 2 O < 10 ll l À1 , CO, C 2 H 6 , CH 4 < 1 ll l À1 ) to create an anaerobic environment with minimum N 2 . Non-inoculated controls were run simultaneously. The total headspace volume of each bottle was 150 ml. Cultures were grown in a pressurized Ar atmosphere at room temperature (23°C) and were agitated on an orbital shaker for 10 days. Colony-forming units were used to estimate the number of viable cells for each replicated treatment each day using 10-fold dilution (up to 10 À9 ) with 100 ll of diluted cells plated out onto solidified Luria-Bertani agar (Oxoid Thermo Fisher Scientific Inc. Waltham, MA, United States) and incubated for 3 days at room temperature (24°C) for 3 days.

Gas sampling and analyses
Gas samples were taken on a daily basis for 10 days (2 days after inoculation) where 1 ml of headspace was extracted and added to 11 ml of helium (99.99% pure) contained in 12 ml glass Exetainers â . Sample volume removed was replaced with 1 ml of Ar. Samples were analysed for Ar, O 2 , N 2 , N 2 O and 15 N using a Continuous Flow Isotope Ratio Mass Spectrometer (PDZ Europa TGII/20-20). The concentration of 15 N-N 2 was below the limit of quantification and thus not reported. Accuracies of Cu concentrations per treatment were verified at AE 10% at the end of the experiment using inductively coupled plasma optical emissions spectrophotometer (ICP-OES).

Anion and cation analyses
After 10 days, 10 ml of the growth medium was filtered using a 0.22 lm cellulose filter (DISMIC, Advantec, Tokyo, Japan) and analysed for NO 2 À and NO 3 À concentrations using a Dionex DX-120 Suppressed Ion Transcripts of nirS, norB and nosZ encoding the cytochrome cd 1 haem NIR, NOR and N 2 OR, respectively, were quantified by qRT-PCR. Specific primers for three reference genes fdxA, ropD and gyrB, and one denitrification gene, norB, were designed from Pseudomonas stutzeri ATCC 17588 genome sequence (accession: PRJNA68131) using the web-based program Primer-BLAST (Ye et al., 2012) to generate~200 bp DNA sequences (Table 1). The cDNA products were amplified with an ABI Prism 7000 (Applied Biosystems, Mulgrave Australia) using SensiFAST TM SYBR â Hi-ROX Kit (Bioline, London, UK) according to supplier's instructions. Quantitative RT-PCR was performed with an initial denaturation at 95°C for 2 min, followed by 40 cycles of 95°C for 5 s, annealing at 60°C for 10 s and extension at 72°C for 20 s, with exception to the nirS which was amplified with an extension step for 30 s at 72°C. Raw data were analysed based on Pfaffl et al. (2004) mathematical model to determine the relative quantitation of the target genes (nirS, norB and nosZ) that are normalized by three non-regulated reference genes (fdxA, ropD and gyrB). High resolution melting analysis was performed to determine PCR integrity and primer dimers at the end of each run. Each qRT-PCR efficiency is within the value between 1.60 and 2.10 and a R 2 value of > 0.980. The qRT-PCR was performed in triplicate, and the means and standard errors were calculated.

Statistical analysis
Statistical analysis on gas results, ion chromatography for broth chemistry and colony-forming units were performed on log-transformed data using one-way ANOVA in GenStat (GenStat14; VSN International, Hemptstead UK). Non-linear regression to evaluate the relationships between ratios of N 2 O:N 2 +N 2 O per treatment over the 10-day period was performed in GenStat (GenStat 16.1; VSN International).