Denitrifying metabolism of the methylotrophic marine bacterium Methylophaga nitratireducenticrescens strain JAM1

Background Methylophaga nitratireducenticrescens strain JAM1 is a methylotrophic, marine bacterium that was isolated from a denitrification reactor treating a closed-circuit seawater aquarium. It can sustain growth under anoxic conditions by reducing nitrate (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}${\mathrm{NO}}_{3}^{-}$\end{document}NO3−) to nitrite (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}${\mathrm{NO}}_{2}^{-}$\end{document}NO2−). These physiological traits are attributed to gene clusters that encode two dissimilatory nitrate reductases (Nar). Strain JAM1 also contains gene clusters encoding two nitric oxide (NO) reductases and one nitrous oxide (N2O) reductase, suggesting that NO and N2O can be reduced by strain JAM1. Here we characterized further the denitrifying activities of M. nitratireducenticrescens JAM1. Methods Series of oxic and anoxic cultures of strain JAM1 were performed with N2O, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}${\mathrm{NO}}_{3}^{-}$\end{document}NO3− or sodium nitroprusside, and growth and N2O, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}${\mathrm{NO}}_{3}^{-}$\end{document}NO3−, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}${\mathrm{NO}}_{2}^{-}$\end{document}NO2− and N2 concentrations were measured. Ammonium (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}${\mathrm{NH}}_{4}^{+}$\end{document}NH4+)-free cultures were also tested to assess the dynamics of N2O, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}${\mathrm{NO}}_{3}^{-}$\end{document}NO3− and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}${\mathrm{NO}}_{2}^{-}$\end{document}NO2−. Isotopic labeling of N2O was performed in 15NH4+-amended cultures. Cultures with the JAM1ΔnarG1narG2 double mutant were performed to assess the involvement of the Nar systems on N2O production. Finally, RT-qPCR was used to measure the gene expression levels of the denitrification genes cytochrome bc-type nitric oxide reductase (cnorB1 and cnorB2) and nitrous oxide reductase (nosZ), and also nnrS and norR that encode NO-sensitive regulators. Results Strain JAM1 can reduce NO to N2O and N2O to N2 and can sustain growth under anoxic conditions by reducing N2O as the sole electron acceptor. Although strain JAM1 lacks a gene encoding a dissimilatory \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}${\mathrm{NO}}_{2}^{-}$\end{document}NO2− reductase, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}${\mathrm{NO}}_{3}^{-}$\end{document}NO3−-amended cultures produce N2O, representing up to 6% of the N-input. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}${\mathrm{NO}}_{2}^{-}$\end{document}NO2− was shown to be the key intermediate of this production process. Upregulation in the expression of cnorB1, cnorB2, nnrS and norR during the growth and the N2O accumulation phases suggests NO production in strain JAM1 cultures. Discussion By showing that all the three denitrification reductases are active, this demonstrates that M. nitratireducenticrescens JAM1 is one of many bacteria species that maintain genes associated primarily with denitrification, but not necessarily related to the maintenance of the entire pathway. The reason to maintain such an incomplete pathway could be related to the specific role of strain JAM1 in the denitrifying biofilm of the denitrification reactor from which it originates. The production of N2O in strain JAM1 did not involve Nar, contrary to what was demonstrated in Escherichia coli. M. nitratireducenticrescens JAM1 is the only reported Methylophaga species that has the capacity to grow under anoxic conditions by using \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}${\mathrm{NO}}_{3}^{-}$\end{document}NO3− and N2O as sole electron acceptors for its growth. It is also one of a few marine methylotrophs that is studied at the physiological and genetic levels in relation to its capacity to perform denitrifying activities.


Introduction
The complete denitrification pathway describes the successive reduction of nitrate (NO 3 − ) to nitrite (NO 2 − ), nitric oxide (NO), nitrous oxide (N 2 O), and nitrogen (N 2 ) (1). This process is used by bacteria for respiration in environments with low oxygen concentrations and with NO 3 − as an electron acceptor. The process is driven by metalloenzymes NO 3 − reductase, NO 2 − reductase, NO reductase, and N 2 O reductase (2). As a facultative trait, denitrification occurs frequently across environments and is performed by bacteria of diverse origins (3). However, numerous bacterial strains have been isolated with incomplete denitrification pathway, meaning that at least one reductase-encoding gene cluster is missing. As proposed by Zumft (3), the four steps of reduction from NO 3 − to N 2 could be seen as a modular assemblage of four partly independent respiratory processes that respond to combinations of different external and internal signals. This could explain the vast diversity of bacteria with incomplete denitrification pathway that can sustain growth with one of the four nitrogen oxides as electron acceptor. Another purpose of the incomplete pathway is related to detoxification, as nitrite and NO are deleterious molecules (4)(5)(6)(7).
Methylophaga nitratireducenticrescens JAM1 is a marine methylotrophic gammaproteobacterium that was isolated from a naturally occurring multispecies biofilm that has developed in a methanol-fed, fluidized denitrification system that treated recirculating water of the marine aquarium in the Montreal Biodome (8,9). This biofilm is composed of at least 15 bacterial species and of numerous protozoans (10,11), among which Methylophaga spp. and Hyphomicrobium spp. compose more than 70% of the biofilm (12). Along with the denitrifying bacterium Hyphomicrobium nitrativorans NL23, M. nitratireducenticrescens JAM1 was shown to be the representative of the Methylophaga population in the biofilm (8).
M. nitratireducenticrescens JAM1 is considered as a nitrate respirer as it can grow under anoxic conditions through the reduction of NO 3 − to NO 2 − , which accumulates in the culture medium (8). This trait is correlated with the presence of two gene clusters encoding dissimilatory nitrate reductases (narGHJI, referred as Nar1 and Nar2) in the genome of M. nitratireducenticrescens JAM1, which we showed that both contribute to NO 3 − reduction during strain JAM1 growth (13). Anaerobic growth by strain JAM1 is a unique among Methylophaga spp. that were described as strictly aerobic bacteria (14). Genome annotation revealed that stain JAM1 seems to maintain an incomplete denitrification pathway with the presence of gene clusters encoding two putative cytochrome bc-type complex NO reductase (cNor) (cnor1: norQDBCRE and cnor2: norCBQD) and one putative dissimilatory N 2 O reductase (N 2 OR) (nosRZDFYYL), but lacks gene encoding a dissimilatory copper-(NirK) or cytochrome cd1-type (NirS) NO 2 − reductase. These gene clusters have been shown to be transcribed (13). These data suggest that M. nitratireducenticrescens JAM1 has other respiratory capacities by performing NO and N 2 O reduction.
In this study, we aimed to assess the denitrification capacities of M. nitratireducenticrescens JAM1. Our results show that strain JAM1 can reduce NO to N 2 O and then to N 2 . It can use N 2 O as a source of energy for its growth under anoxic conditions. Through our investigation, we found that strain JAM1 cultured with NO 3 − under anoxic and oxic conditions generates a small amount of N 2 O, despite the absence of gene encoding NirK or NirS. NO 2 − was found to be a key intermediate of this production process. By using the JAM1ΔnarG1-narG2 double mutant, we showed that the two Nar were not involved in N 2 O production via NO. We analyzed at the gene expression level the succession of N 2 O production and consumption with the denitrification genes cnorB (cnorB1 and cnorB2) and nosZ, and also nnrS that encodes a NO-sensitive regulator. We found that gene expression level of cnorB1 and nnrS increased during the N 2 O production phase, which suggest the presence of NO.   In N 2 O-amended cultures, N 2 O decrease was apparent from the start and consumption continued for 48 hours (Fig. 1A)

N 2 O production in NO 3 --amended cultures
During the first assays to test the capacity of strain JAM1 to reduce N 2 O, cultures were performed with N 2 O (3.5 mg-N vial -1 ) but with the addition of NO 3 -(20 mg-N vial -1 ) to make sure that growth would occur. Although N 2 O was completely consumed within 24 h, a net production of N 2 O was observed after 48 h (data not shown). The production of N 2 O by strain JAM1 is puzzling, as its genome does not contain NirS or NirK. To further investigate this observation, strain JAM1 was cultured under anoxic conditions with NO 3 − , and NO 3 − , NO 3 − and N 2 O were measured (Fig.  3A). Complete NO 3 − reduction (19.3 ± 0.3 mg-N vial -1 ) was performed within 55 h. The nitrite level reached 17.5 ± 0.2 mg-N vial -1 over this period and decreased slowly to 15.9 ± 0.5 mg-N vial -1 . N 2 O production initiated when NO 3 − was nearly reduced and reached 0.70 ± 0.21 mg-N vial -1 after 55 h of incubation (Fig. 3A). N 2 O was completely reduced after 127 h. In parallel, for cultures in which the headspace was flushed with argon, N 2 production was also measured. The corresponding results show an increase of N 2 in the headspace (Fig. 3A)  The original 1403 medium recommended by the ATCC for culturing Methylophaga spp. contains 20.9 mg-N vial -1 NH 4 Cl and 0.1 mg-N vial -1 ferric ammonium citrate (see Material and Methods). Therefore, NO 3 − transformation in NH 4 + should not be necessary in Methylophaga metabolism for nitrogen assimilation in biomass (Fig. S1). For the next set of experiments, we aimed to determine the effect of the absence of NH 4 + on the dynamics of NO 3 − and NO 2 − consumption, and N 2 O production and consumption. We hypothesized that that forcing strain JAM1 to reroute some NO 3 − for N assimilation would affect denitrification and thus growth rates. Strain JAM1 was cultured with ca. 20 mg-N vial -1 NO 3 − under anoxic or oxic conditions in NH 4 Cl-free medium ( Fig. 4A and B). Growth pattern observed under anoxic conditions was similar between the regular and NH 4 Cl-free cultures, as also the growth pattern under oxic conditions between the regular and NH 4 Cl-free cultures. Under anoxic NH 4 Cl-free conditions, full nitrate reduction (19.1 ± 0.6 mg-N vial -1 ) occurred within 48 h (Fig. 4A). The N 2 O production and consumption profile found was similar to that observed in regular cultures (Fig. 3A), though lower N 2 O concentrations were detected during the accumulating phase. The nitrite level reached 18.5 ± 0.8 mg-N vial -1 after 24 h and then slowly decreased to 12.8 ± 0.5 mg-N vial -1 after 96 h. Nitrogen assimilation by the biomass and the production of N 2 O and its reduction to N 2 could account for the difference in nitrogen mass balance (32.7 ± 2.5%).
In conjunction with NO 3 − reduction, NO 2 − levels stopped accumulating at 13.0 ± 2.6 mg-N vial -1 after 24 h. N 2 O was observed after 48 h of incubation ( Fig. 4A), after which it slowly accumulated and reached a concentration of 0.043 ± 0.048 mg-N vial -1 . This level is 7 times lower than that of the regular culture medium (Fig. 3B). Nitrogen assimilation by the biomass under oxic conditions could account for the difference in nitrogen mass balance (25.5 ± 4.3%) found between the initial concentration of NO 3 − and residual concentrations of NO 3 − , NO 2 − and N 2 O (13.9%, 60.4% and 0.20% of the N input, respectively). In addition, N 2 O production and consumption could have reached an equilibrium and loss of nitrogen would occur by N 2 production.
To assess whether N 2

NO reduction by M. nitratireducenticrescens JAM1
To verify NO reduction by strain JAM1, N 2 O generation was monitored in cultures without NO 3 − and supplemented with sodium nitroprusside hypochloride (SNP) used as an NO donor (Fig. 5). Because N 2 O is quickly reduced under anoxic conditions but accumulates under oxic conditions, these assays were performed under oxic conditions. N 2 O started to accumulate in both 2 mM and 5 mM SNP-supplemented media after 24 h of incubation, reaching 7.9 ± 0.5 µg-N vial -1 and 14.5 ± 0.4 µg-N vial -1 , respectively, after 168 h. No N 2 O production was observed in strain JAM1 cultures without SNP or in the controls with non-inoculated culture medium supplemented with SNP or inoculated with autoclaved biomass. Strain JAM1 was cultured under oxic conditions without NO3 − and with 2 mM (square), with 5 mM (triangle), or with no (circle) sodium nitroprusside. N2O concentrations were measured over different time intervals. Controls with 5 mM SNP in non-inoculated culture medium (reverse triangle) and in culture medium inoculated with autoclaved biomass (diamond) were also performed. Data represent mean values ± SD (n=3).

Role of Nar systems in NO/N 2 O production
In the absence of NirK or NirS, N 2 O could have been generated via NO by the Nar system (see discussion). We used the JAM1ΔnarG1narG2 double mutant, which lacks functional Nar-type nitrate reductases and which cannot grow under anoxic conditions (13). Strain JAM1 and the JAM1ΔnarG1narG2 were cultured with 16.8 mg-N vial -1 NO 3 − under oxic conditions. The growth of strain JAM1 and the mutant was similar (13). After 96 h of incubation, strain JAM1 completely reduced NO 3 − to NO 2 − and produced 0.14 mg-N vial -1 of N 2 O ( Table 1). As was expected, NO 3 − was not reduced, and NO 2 − was not produced by JAM1ΔnarG1narG2. Contrary to the wild type strain, the mutant did not produce N 2 O.

Indirect detection of NO production by M. nitratireducenticrescens JAM1
We assessed whether variations in the expression levels of denitrification genes correlate with the N 2 O production and consumption cycles of strain JAM1 cultures. Strain JAM1 was cultured in NH 4 Cl-free medium with 22 mg-N vial -1 NO 3 − under anoxic conditions. RNA was extracted from cells harvested over four different phases (Fig. 4) : 1) at T0 for the pre-cultures, 2) during the growth phase with nitrate reduction and no N 2 O production, 3) during the N 2 O-production phase, and 4) during the N 2 O-consumption phase. The relative transcript levels of cnorB1, cnorB2 and nosZ, which encode the catalytic subunits of the corresponding NO and N 2 O reductases, and nnrS, were measured by RT-qPCR. nnrS encodes a NO-sensitive regulator and was used as an indicator of the presence of NO in the cultures. cnorB1 and nnrS expression patterns observed were similar, with the highest expression levels observed during the N 2 O-production phase (Fig. 6), which suggests the presence of NO. The cnorB2 expression level remained stable except during the initial phase, when it was significantly lower. With the exception of that of the pre-cultures, the cnorB2 transcript level was always lower than those detected for cnorB1. No significant changes in nosZ expression levels were observed over the four phases.   Figure 1B under the same conditions with regular 1403 medium. Samples were drawn from the pre-culture and during the growth phase (no N2O production), N2O-production phase, and N2O-consumption phase (see Figure 4D), from which total RNA was extracted. Gene expression levels of cnorB1, cnorB2, nnrS and nosZ were measured by RT-qPCR and were reported as the gene copy number per copy of dnaG (reference gene). Student's t-tests were performed for each phase to draw comparisons to the pre-culture phase. *: 0.05<P<0.01; **: 0.01<P<0.001. The results are derived from triplicate cultures from different inoculums. Data represent mean values ± SD (n=3).

Discussion
Our results show that M. nitratireducenticrescens JAM1 can consume NO and N 2 O via the mechanism of reduction of NO to N 2 O and then to N 2 as predicted by the genome sequence (Fig. S1) (9,13). The N 2 O-amended cultures yielded equivalent biomass results to those of the NO 3 − -amended cultures as predicted by the respiratory electron transport chains of the denitrification pathway (15). Therefore, in addition of reducing nitrate, strain JAM1 has another respiratory capacity under anoxic conditions by reducing N 2 O for its growth. As observed with nitrate reduction, NO and N 2 O reduction can occur under oxic conditions, reinforcing the lack of a functional oxygen regulation response in strain JAM1 (see Discussion in 13). N 2 O production was observed in NO 3 −amended cultures either under oxic or anoxic conditions when NO 2 − was accumulating. This production represented up to 6% of N-input in the anoxic cultures, and NO 2 − was shown to be the key element of this production process. Because, we showed that the NO reductase activities were carried out in strain JAM1 cultures, the N 2 O could originate from NO production despite the absence of gene encoding NirS or NirK. Intermediate NO creates problems as this molecule is highly toxic to microorganisms, inducing nitrosative stress in cells (5). Reducing NO is a key step in denitrification and is closely regulated by various sensors and regulators. NnrS is involved in cell defense against nitrosative stress and is positively regulated by the presence of NO (16)(17)(18). Therefore, nnrS expression reflects NO concentrations in a medium and was used as a marker of NO presence. The higher expressions of nnrS found during N 2 O production strongly suggest that NO is produced during this phase. This correlates with higher expressions of cnorB1, which can also be regulated by NO-sensitive regulators such as NnrS or NorR (19). Moreover, the increase in cnorB1 expression found can be directly linked to observed N 2 O production levels. During the N 2 O-consumption phase, only cnorB1 expression decreased and could have changed the balance between N 2 O production and consumption. As nosZ expression is mainly reduced by the presence of O 2 , the stable expression of this gene under constant anoxic conditions was expected (19). Interestingly, cnorB2 expression was not affected by the presence of NO, unlike cnorB1. This suggests a putatively different regulation mechanism for this gene like those observed for narG1 and narG2 in a previous study (13). Finally, the succession of different phases of N 2 O production/consumption correlates with the presence of NO through an increased expression of nnrS, which strongly suggests that NO is an intermediate in N 2 O production in strain JAM1.
Other nitrate respiring bacteria that lack NirK or NirS have been shown to be N 2 O producers (20)(21)(22). For instance, Bacillus vireti contains three denitrification reductases (Nar, qCu A Nor, N 2 OR) and lacks, like M. nitratireducenticrescens JAM1, gene encoding NirK or NirS (23). This bacterium also produces NO and N 2 O in anaerobic, NO 3 − -amended TSB cultures during NO 2 − accumulation. NO was shown to originate from chemical decomposition of NO 2 − (6) and from an unknown biotic reaction. In our study, the abiotic control of the Methylophaga 1403 medium amended with NO 3 − and NO 2 − did not show N 2 O production. Furthermore, no N 2 O was detected in this medium inoculated with autoclaved biomass (Fig. 5). These results rule out NO/N 2 O production by chemical decomposition of NO 2 − in strain JAM1 cultures. The possible biotic source of NO in absence of NirS or NirK has been studied in Escherichia coli (see review by Vine and Cole (24). There are supporting evidence that NO is generated in E. coli as a side product during nitrite reduction (i) by the cytoplasmic, NADH-dependent nitrite reductase (NirBD), (ii) by the nitrite reductase NrfAB, and (iii) by NarGHI. Vine et al. (25) showed, with mutants defective in these reductases, that NarGHI is the major enzyme responsible of NO production. However, a small production of NO was still occurring in narG mutant, suggesting the involvement of another molybdoprotein. In M. nitratireducenticrescens JAM1, the double-knockout mutant JAM1ΔnarG1narG2, which lacks the two dissimilatory NO 3 − reductases, was still able to produce N 2 O under oxic conditions at the same level of the wild type when NO 2 − was added to the cultures. These results suggest the two Nar systems are not involved in NO production. The genome of strain JAM1 did not reveal gene encoding NrfAB, but contain a gene cluster encoding a cytoplasmic, NADH-dependent nitrite reductase (CP003390.3; Q7A_2620 and Q7A_2621), which may be the source of NO (Fig. S1).
The significance of maintaining an incomplete pathway by M. nitratireducenticrescens JAM1 is unclear and may depend upon the original habitat and environment, here the denitrifying biofilm. While M. nitratireducenticrescens JAM1 serves as an important actor among the microbial community of the marine biofilm in performing optimal denitrifying activities (10,26), it was thought to participate uniquely in the reduction of NO 3 − to NO 2 − . It was previously proposed that NO 2 − reduction to N 2 is carried out by Hyphomicrobium nitrativorans NL23, the second most represented bacterium in the biofilm (12,27). Its capacity to reduce NO and N 2 O and to grow on N 2 O suggests that M. nitratireducenticrescens JAM1 may participate in the reduction of NO and N 2 O during denitrification in the biofilm. Although our culture assays were performed with high levels of NO 3 − (37 mM), which is rarely exceeds a value of 0.7 mM in natural environments (28), similar levels can be reached in closed-circuit systems like the seawater aquarium tank located in the Montreal Biodome, where NO 3 − levels reached up to 14 mM (29). Rissanen et al. (30) observed also the combination of Methylophaga spp. and Hyphomicrobium spp. in the fluidized-bed type denitrification reactors treating the recirculating seawater of the public fish aquarium SEA LIFE at Helsinki, Finland. Although, this study provided no indication of the denitrification pathway in these Methylophaga and Hyphomicrobium, it reinforces the importance of the natural combination of these two genera in marine denitrification environment.

Conclusions
M. nitratireducenticrescens JAM1 is one of few isolated marine methylotrophic bacterial strains to exhibit anaerobic respiratory capacities by reducing NO 3 − to NO 2 − and, as reported here, by reducing N 2 O to N 2 . It can also generate N 2 O via NO by an unknown biotic system. Very few marine denitrifying bacteria have been isolated from recirculating marine systems (31)(32)(33)(34). No previous studies have generated genetic information related gene arrangement or expression on these bacteria. Based on substantial data accumulated on the genome, gene arrangement and gene expression of denitrification and on methylotrophy, M. nitratireducenticrescens JAM1 can serve as a model for studying such activities in marine environments. Finally, our results enable a better understanding of the ecophysiological role of M. nitratireducenticrescens JAM1 in the original biofilm developed in the denitrification reactor of a closed-circuit marine aquarium.

Materials and Methods
Bacterial growth conditions M. nitratireducenticrescens JAM1 and the JAM1ΔnarG1narG2 double mutant were cultured in the American Type Culture Collection (ATCC, Manassas, VA, USA) Methylophaga medium 1403 (9,13 [4] where A N2O : the N 2 O mixing ratio measured in the headspace (ppmv/10 6 ; no unit); P: 1 atm; V 1 and V g : volume of the aqueous (0.04 or 0.06 L vial -1 ) and gaseous phases (0.68 or 0.66 L vial -1 ), respectively; and V n : molar volume [RT (gas constant): 0.08206 L atm K -1 mol -1 * 303K = 24.864 L mol -1 ]. K H30sw is the corrected Henry's constant for seawater at 30°C (0.01809 mol L -1 atm -1 ) according to Weiss and Price (1980). X N2O was then converted (eq.5) in an mg-N vial -1 for an easier calculation of mass balances using the other nitrogenous compounds: The reproducibility of the N 2 was assessed before each set of measurements was made by a repeated analysis of N 2 (purity >99.99%, Praxair) diluted in a 720 mL gastight bottle (0 and 500 ppmv) flushed with argon (purity >99.99%, Praxair). The total quantity of N 2 in the culture bottles was only considered for the headspace, as the quantity of dissolved N 2 in the aqueous phase was considered to be negligible in our experimental design based on Henry's constant (0.0005 mol L -1 atm -1 ) and was thus calculated according to Eq. 6.

RNA extraction
Anoxic cultures of strain JAM1 were created in an NH 4 Cl-free 1403 medium supplemented with 22 mg-N vial -1 NO 3 − . Cells were harvested at specific times, and RNA was immediately extracted using the PureLink RNA mini kit (Ambion Thermo Fisher Scientific, Burlington, ON, Canada). RNA extracts were treated twice with TurboDNase (Ambion), and RNA quality was verified by agarose gel electrophoresis. The absence of remaining DNA was checked via the end-point polymerase chain reaction (PCR) amplification of the 16S rRNA gene using RNA extracts as the template.

Gene expression
cDNAs samples were generated from the RNA using hexameric primers and the Reverse Transcription System developed by Promega (Madison, WI, USA) with 1 μg of RNA. Real-time quantitative PCR (qPCR) assays were performed using the Faststart SYBR Green Master (Roche Diagnostics, Laval, QC, Canada) according to the manufacturer's instructions. All reactions were performed in a Rotor-Gene 6000 real-time PCR thermocycler (Qiagen Inc. Toronto, ON, Canada), and each reaction contained 25 ng of cDNA and 300 nM of primers ( Table 2). Genes tested included cnorB1, cnorB2, nosZ and nnrS. PCR began with an initial denaturation step of 10 min at 95°C followed by 40 cycles of 10 s at 95°C, 15 s at 60°C, and 20 s at 72°C. To confirm the purity of the amplified products, a melting curve was performed by increasing the temperature from 65°C to 95°C at increments of 1°C per step with a pause of 5 s included between each step. All genes for each sample and standard were tested in a single run. The amplification efficiency level was tested for each set of primer pairs by qPCR using a dilution of strain JAM1 genomic DNA as the template. The amplification efficiencies for all primer pairs varied between 0.9 and 1.1. The copy number of each gene was calculated according to standard curves using dilutions of strain JAM1 genomic DNA. To normalize the gene expression of the different growth phases, results were expressed as copy numbers per dnaG copy numbers for each sample. In accordance with previous studies (13), dnaG generated the least variability of the reference genes tested (dnaG, rpoD and rpoB) in strain JAM1. dnaG encodes for a DNA primase and is present in one copy in strain JAM1 genomes. RNA extraction and qPCR were performed with three independent biological replicates. The significance of differential expression levels was tested for each phase against the pre-culture phase via Student's t-test.