nosX is essential for whole-cell N2O reduction in Paracoccus denitrificans but not for assembly of copper centres of nitrous oxide reductase

Nitrous oxide (N2O) is a potent greenhouse gas that is produced naturally as an intermediate during the process of denitrification carried out by some soil bacteria. It is consumed by nitrous oxide reductase (N2OR), the terminal enzyme of the denitrification pathway, which catalyses a reduction reaction to generate dinitrogen. N2OR contains two important copper cofactors (CuA and CuZ centres) that are essential for activity, and in copper-limited environments, N2OR fails to function, contributing to rising levels of atmospheric N2O and a major environmental challenge. Here we report studies of nosX, one of eight genes in the nos cluster of the soil dwelling α-proteobaterium Paraccocus denitrificans. A P. denitrificans ΔnosX deletion mutant failed to reduce N2O under both copper-sufficient and copper-limited conditions, demonstrating that NosX plays an essential role in N2OR activity. N2OR isolated from nosX-deficient cells was found to be unaffected in terms of the assembly of its copper cofactors, and to be active in in vitro assays, indicating that NosX is not required for the maturation of the enzyme; in particular, it plays no part in the assembly of either of the CuA and CuZ centres. Furthermore, quantitative Reverse Transcription PCR (qRT-PCR) studies showed that NosX does not significantly affect the expression of the N2OR-encoding nosZ gene. NosX is a homologue of the FAD-binding protein ApbE from Pseudomonas stutzeri , which functions in the flavinylation of another N2OR accessory protein, NosR. Thus, it is likely that NosX is a system-specific maturation factor of NosR, and so is indirectly involved in maintaining the reaction cycle of N2OR and cellular N2O reduction.


INTRODUCTION
Nitrous oxide is a potent greenhouse gas which has rapidly increased in the atmosphere over the past century [1]. The rise in N 2 O coincides with the introduction and application of anthropogenic nitrogen species in agriculture, to improve crop yield and ultimately feed the growing global population [2,3]. Of the total N 2 O released, 40 % is produced by soil bacteria [4]. Soil dwelling denitrifying micro-organisms such as Paracoccus denitrificans consume nitrate as an alternative electron acceptor during anaerobic growth conditions. N 2 O is an intermediate substrate in the denitrification pathway; it is reduced to N 2 by the copper enzyme nitrous oxide reductase (N 2 OR). N 2 OR-containing bacteria can be separated into two clades, and a feature that distinguishes the clades is the ability of the micro-organism to produce and consume, or only consume, N 2 O [5,6]. Clade-I members are complete denitrifiers with the nitrite reductase genes nirS or nirK present in their genome. In contrast, about half of the clade-II members are non-denitrifying N 2 O reducers, and are therefore N 2 O sinks [5]. Ammonia-oxidizing bacteria (AOB) are another microbial source of N 2 O in coastal ecosystems, through a process named 'nitrifier denitrification' . However, they do not harbour genes encoding N 2 O reduction activity [7]. Environmental factors such as soil pH, Cu content, and moisture impact on N 2 O emissions from soil [8][9][10]. In order OPEN ACCESS to identify N 2 O mitigation strategies, we are trying to understand the optimal genetic components needed to biologically remove N 2 O. Nitrous oxide reductase (N 2 OR) is a homo-dimeric, ~120 kDa, multi-Cu protein. Each monomer contains two Cu cofactors, the Cu A and Cu Z centres, responsible for electron transfer and the catalytic reduction of N 2 O, respectively. The Cu A centre is a bis-thiolate-bridged di-nuclear Cu centre, accommodated within a cupredoxin fold domain, similar to that of subunit II of cytochrome c oxidase. The Cu Z centre is a unique [Cu-S] cluster ligated by seven conserved histidine residues within a β-barrel domain. It comprises four Cu atoms and one or two sulphur atoms, depending on the purification method [11][12][13]. Notably, the subunits of the active homodimer are orientated in a head to tail configuration, with one Cu A centre in close proximity to the Cu Z centre of the other monomer. N 2 OR is encoded by the nosZ gene, which, in denitrifying organisms such as Paracoccus denitrificans and Pseudomonus stutzeri, is translated and exported through the twin-arginine transport [14] pathway to the periplasm, as a folded apo-protein, before acquiring its Cu cofactors. Consistent with this, a TAT signal leader sequence mutant accumulated unprocessed, dimeric, apo-protein in the cytoplasm of the cell [15]. In contrast, the N 2 OR of clade-II members are transported through the Sec pathway [16]. The functional significance of this is currently unknown.
The nosZ gene is found among the nos gene cluster (NGC), which comprises eight genes in P. denitrificans: nosCRZD-FYLX. The nosC and nosR genes are copper responsive in P. denitrificans and function in the regulation of nosZ transcription. During Cu limitation, nosCR transcription is increased, whilst nosZ transcription is reduced [17]. In Pseudomonas stutzeri, NosR is a cytoplasmic membrane protein with two soluble domains located at either side of the membrane: the N-terminal periplasmic domain covalently binds a flavin mononucleotide, while the C-terminal cytoplasmic domain binds two [4Fe-4S] clusters [18]. The P. denitrificans homologue (44.3 % identical) is predicted to have similar features. The function of NosR is not well understood; in addition to the regulatory role mentioned above, it is important for whole-cell N 2 O reduction [17,18], with evidence indicating that it is not involved in the assembly of the Cu centres of N 2 OR, but may be the physiological electron donor to NosZ [18].
nosDFY encode a cytoplasmic membrane spanning ABC-type transporter that functions in the maturation of the Cu Z centre of N 2 OR, as illustrated by an insertional mutation in P. stutzeri nosD, which produced an N 2 OR without the key spectroscopic signal of the Cu Z centre [19]. Similarity to mitochondrial ABC transporters that export a sulphur species to the cytoplasm for iron-sulphur cluster biogenesis suggests a role for NosDFY in providing the essential sulphur atoms of the catalytic Cu Z centre [20]. The nosL gene is well conserved across NGCs and is essential for whole-cell N 2 O reduction in P. denitrificans. NosL is a Cu-binding lipoprotein, putatively anchored to the outer membrane of the cell. The properties of N 2 OR purified from a PdΔnosL strain revealed that Cu-binding NosL is a component of the Cu Z maturation apparatus under Cu replete conditions and, more importantly, is an essential maturation factor for both Cu centres during Cu limitation [21].
The nosX gene is predominantly found in αand β-proteobacterial NGCs in clade I but does not feature among γ-proteobacteria nor clade-II NGCs (Fig. 1). NosX is a soluble protein of ~30 kDa, which is exported to the periplasm by the Tat pathway. Previously, it was reported that insertional mutagenesis of P. denitrificans nosX resulted in wild-type-like growth [22]. Interruption of both nosX and the homologue nirX did, however, present a Nos-negative (Nos -) phenotype, leading to the conclusion that NosX and NirX are functional homologues [22]. Furthermore, this study demonstrated that the nosX nirX double mutant strain contained N 2 OR that was deficient in its Cu A centre, implicating these proteins in copper cofactor assembly [22].
The γ-proteobacterium P. stutzeri, which does not feature nosX in its NGC, instead contains a NosX homologue encoded elsewhere on the genome. The protein, called ApbE, was shown to be a FAD-binding flavinyl transferase that serves as a flavin donor to NosR, which in turn activates N 2 OR [23]. PsApbE and PdNosX share 32 % amino acid homology, in particular the conservation of key amino acid residues associated with flavin binding suggest that their roles are similar while their genetic context implies they may differ in system specificity.
Here, we present a re-examination of the role of NosX in P. denitrificans, through the analysis of full nosX deletion in P. denitrificans, in terms of cell growth and the properties of N 2 OR purified from an unmarked mutant background. The data show that NosX is essential for N 2 OR activity and cannot be substituted by NirX. Furthermore, NosX plays no role in assembly of the NosZ Cu cofactors, nor does it have a major function in the regulation of nosZ expression. Instead, the role of NosX is consistent with a system-specific maturation factor for NosR to support the activity of NosZ in vivo.

Construction and complementation of a nosXdeficient strain of P. denitrificans
A double allelic exchange method was employed to generate a whole nosX gene deletion strain (Table S1, available in the online version of this article), as described previously [17,21]. Briefly, the suicide plasmid pK18mobsacB containing DNA regions that flank the nosX gene (pSPBN4) was conjugated into PD1222 using the E. coli helper plasmid pRK2013. Single cross-over recombination events resulted in Spec R / Km R transconjugants, from which a double cross over mutant (Spec R ), named PD2502, was generated. The mutated region was PCR amplified and confirmed by sequencing.
PdΔnosX (PD2502) was complemented in trans using pSPBN5, which contains the coding sequence of Pden_4214. The gene was synthesized by Genscript with flanking 5′ NdeI and 3′ EcoRI restriction sites and subcloned into a taurine inducible modified pLMB509 derivative with gentamycin resistance (20 µg ml −1 ) to generate pSPBN5. The complementation plasmid was conjugated into the mutant strain using the helper E. coli pRK2013 strain, with successful conjugants identified as Spec R /Gm R . Expression of nosX from the plasmid was induced by adding 1 mM taurine to the medium at the start of growth.

Growth and phenotypic analysis of cultures
Anaerobic minimal media batch cultures (400 ml) were grown in sealed Duran flasks fitted with a septum seal to allow for gas-tight sample extraction. Minimal media consisted of: 30 mM succinate, 20 mM nitrate, 11 mM dihydrogen orthophosphate, 29 mM di-sodium orthophosphate, 0.4 mM magnesium sulphate, 1 mM ammonium chloride, pH 7.5. The minimal media was supplemented with a 2 ml l −1 Vishniac and Santer trace element solution [24] where copper sulphate was present (Cu-sufficient, 12.8 µM) or excluded (Cu-limited, <0.5 µM) from the original recipe. Media were inoculated using a 1 % inoculum from a starter culture to give a starting OD 600 nm of ~0.02 and incubated at 30 °C. Samples of the liquid culture were taken in 1 ml aliquots and OD 600 nm measured. The 3 ml gas samples were removed from the headspace of the cultures and stored in pre-evacuated 3 ml Exetainer vials. A 50 µl gas sample was injected into a Clarus 500 gas chromatograph (PerkinElmer) equipped with an Elite-PLOT Q (30 m×0.53 mm internal diameter) and an electron capture detector. Carrier gas was N 2 , make-up gas was 95 % (v/v) argon, 5 % (v/v) methane. Standards containing N 2 O at 0.4, 5, 100, 1000, 5000 and 10 000 p.p.m. (Scientific and Technical Gases) were measured and total N 2 O was determined as previously described [17]. The core nosZDFYL genes encode the nitrous oxide reductase polypeptide (NosZ), an ABC transporter complex (NosDFY) that is essential for Cu Z centre maturation, and a Cu centre maturation factor (NosL). The nosR and nosX genes are less conserved across the two clades. NosR is a transmembrane iron-sulflur cluster containing protein with an FMN moiety, which is obtained from an ApbE-type flavinyltransferase (proposed as NosX here). Together the function of these proteins may involve supplying electrons to N 2 OR for catalytic turnover and, where absent in the NGC, a homologue is likely to be found elsewhere in the genome.

Purification and characterization of affinity-tagged N 2 OR from P. denitrificans strains
Plasmid pMSL002, which encodes NosZ (N 2 OR) with a C-terminal Strep-tag II, was conjugated into wild-type (PD1222), PdΔnosZ (PD2303) and PdΔnosX (PD2502) strains using the E. coli pRK2013 helper strain. Strep-tagged N 2 OR was overproduced and purified as previously described [21]. Briefly, this involved applying the soluble portion of cell lysates to a Hi-Trap HP Strep II affinity column (5 ml, GE Healthcare) and eluting with 20 mM HEPES, 150 mM NaCl and 2.5 mM desthiobiotin, pH 7.2, before exchanging into 20 mM HEPES, 150 mM NaCl, pH 7.2. Sample purity was confirmed using SDS-PAGE analysis and LC-MS. Protein concentrations were determined using the Bradford assay (BioRad) [25] and bovine serum albumin as a protein standard.
UV-visible absorbance spectra of N 2 OR-Strep-tag II from different backgrounds were recorded on a Jasco V-550 spectrophotometer. Samples were made anaerobic by sparging with nitrogen gas for 5 min and oxidized or reduced with 5 mg ml −1 stocks of potassium ferricyanide and sodium dithionite, respectively, in 20 mM HEPES, 150 mM NaCl, pH 7.5, by titrating concentration equivalents. Total copper content of the protein was determined using a colorimetric bathocuproinedisulfonic acid (BCS) assay as previously described [21].
Activities of N 2 OR-Strep-tag II isolated from different backgrounds were determined using an adapted methyl viologen assay [26,27] in which samples were pre-incubated with a 500-fold excess of reduced methyl viologen for 150 min. Reaction was initiated by adding N 2 O saturated buffer and the oxidation of blue (reduced) methyl viologen to its oxidized colourless form was followed at 600 nm as a function of time and data converted to specific activity using ε 600 nm =13 600 M −1 cm −1 for the reduced methyl viologen cation radical [27].

RNA isolation, cDNA synthesis and qRT-PCR experiments
Expression of the nosZ gene was determined by qRT-PCR, using an AriaMx Real-Time PCR System G9930A (Agilent Technologies). The nosX mutant and PD1222 wild-type strains were cultivated under anoxic conditions as mentioned above for 12 h, reaching final OD 600 nm of 0.6. Total RNA extraction, RNA quality and integrity assays, and RNA quantification were performed using the methodology previously described [17]. Briefly, 2 µg of total RNA were used for cDNA synthesis using RevertAid First Strand cDNA synthesis kit (Thermo Scientific) and random hexamers following the supplier's instructions. qRT-PCR reactions were run in triplicate in a total volume of 20 µl containing 10 µl of SensiFAST SYBR No-ROX Mix (Bioline), 0.7, 7 or 70 ng of cDNA and 2 µM of each primer. Melting curves were generated to verify the specificity of each amplification reaction. Expression of nosZ gene was determined using the oligonucleotide pair nosZ2F/ nosZ2R [17] and normalized against the housekeeping gene gapA (glyceraldehyde-3-phosphate dehydrogenase; GAPDH1F/GAPDH1R [17]). The changes in gene expression were analysed accordingly to Plaffl methodology [28]. The data presented correspond to the average of three independent biological replicates.

NosX is essential for whole-cell N 2 O reduction in P. denitrificans
Wild-type P. denitificans (PD1222), ΔnosZ (PD2303, missing the gene encoding N 2 OR) and ΔnosX (PD2502, missing the gene Pden_4214) were grown in batch culture, in minimal medium, under Cu-sufficient and limited conditions. The wild-type culture produced a small amount of N 2 O (~1 mM) in Cu-deficient conditions, but this was no longer detected as the culture moved into the stationary phase of growth Fig. 2. A N 2 OR-negative phenotype (Nos -), in terms of growth and N 2 O production, was observed in the ΔnosZ strain under both Cu regimes. For the ΔnosX strain, growth was affected both under Cu-sufficient and limited conditions, and N 2 O levels were similar to those of the ΔnosZ strain, demonstrating the absence of a functioning enzyme.
The Nosphenotype of the ΔnosX strain was almost fully complemented under both Cu regimes by a plasmid-borne nosX gene copy (pSPBN5) expressed in trans from a taurine inducible promoter (Fig. 3), demonstrating that the Nosphenotype is associated with the absence of nosX and not a downstream effect of the deletion. The data demonstrate that the nosX deletion mutant strain of P. denitrificans is unable to catalyse N 2 O reduction. This is in contrast to a previous study by Saunders and co-workers [22] involving a marked nosX deletion, where it was concluded that that NosX and NirX are functionally redundant, such that only one is required for N 2 O reduction.

NosX is not involved in maturation of either Cu cofactor in N 2 OR
Three possible explanations for the Nosphenotype in the ΔnosX mutant are apparent: the incomplete maturation/ assembly of copper centres of N 2 OR; the failure to activate N 2 OR catalytic activity, for example through disruption of supply of electrons; or, the severe down-regulation of nosZ transcription. To investigate this further, a C-terminal strep II-tagged N 2 OR was purified from the ΔnosX mutant strain and the properties of the N 2 OR analysed with respect to the status of the Cu A and Cu Z centres.
Aerobically purified N 2 OR, also known as the pink form of N 2 OR, has been spectroscopically well characterized and all oxidized spectra were normalised to ε 580 nm 5000 M −1 cm −1 per monomer, as described by Rasmussen et al. [13]. Absorbance spectra of N 2 OR enzymes isolated from cultures grown under Cu-sufficient conditions are shown in Fig. 4a. Spectra of N 2 OR from wild-type cells and ΔnosX and ΔnosZ mutants have features at 480, 540 and 640 nm, in agreement with the previous literature on N 2 OR from P. denitrificans [21], P. pantotrophus (PpN 2 OR) [13], Pseudomonas stutzeri (PsN 2 OR) [29], Pseudomonas nautica (PnN 2 OR) [30], Achromobacter cycloclastes (AcN 2 OR) [31] and Marinobacter hydrocarbonoclasticus (MhN 2 OR) [32]. Features in the absorption spectrum at these wavelengths arise from S 2to Cu(II) charge-transfer bands and additional optical bands due to interactions between the Cu(I) and Cu(II) ions of the centres [13]. Spectra of N 2 OR isolated from wild-type cells have lower extinction coefficients than those from the mutant strains, suggesting that it contains lower levels of Cu cofactors.
The Cu content of all isolated N 2 OR enzymes was determined ( Table 1), confirming that enzymes isolated from ΔnosZ and ΔnosX mutants are replete with Cu, while that from wildtype cells contains slightly lower amounts, consistent with the absorption spectra. N 2 OR activity was measured using a methyl viologen assay in which the reduced MV extinction coefficient, ε 600 nm = 13 600 M −1 cm −1 [27], was used to quantify activity, and N 2 OR was pre-incubated with a 500-fold excess reduced methyl viologen (MV) before initiating the reaction with N 2 O (Table 1). Each N 2 OR sample was active, with values for the enzyme from the wild-type and ΔnosZ  strains consistent with those previously reported [21,26,30]. Activity for N 2 OR from the ΔnosX mutant was similar to that from wild-type, even though it contained significantly more Cu, suggesting that the enzyme from the ΔnosX mutant has a slightly lower activity.
Reduction of N 2 OR samples with dithionite leads to reduction of the Cu A centre to a [Cu 1+ :Cu 1+ ] diamagnetic species, which is colourless and thus does not contribute in the visible region of the absorbance spectrum. Thus, in Fig. 4b, bands at 480, 540 and 900 nm are lost to leave a Cu Z * signature, consisting of a peak at 640 nm, in agreement with the literature for pink N 2 OR [13]. The oxidized minus reduced difference spectrum, Fig. 4c, revealed the spectrum due to the Cu A centre. The close similarity of spectral form and absorption extinction coefficients for N 2 OR from ΔnosZ and ΔnosX mutants demonstrate that the assembly of the Cu cofactors of N 2 OR is not affected by the nosX deletion when grown under Cu sufficiency [13].
An equivalent spectroscopic analysis of N 2 OR enzymes isolated from cultures grown under Cu limitation (Fig. 4d-f) revealed spectra similar to those of Fig. 4a-c for enzymes from ΔnosZ and ΔnosX mutants, but with lower extinction coefficients, suggesting lower incorporation of Cu. Spectra for enzyme isolated from wild-type cultures, however, indicate very low levels of Cu incorporation. Determination of Cu content (Table 1) revealed that N 2 OR from ΔnosZ and ΔnosX mutants contain ~4 Cu per N 2 OR monomer, while that recovered from wild-type cells contains <1 Cu per monomer, consistent with absorbance data (Fig. 4d-f). As above, the close similarity between N 2 OR enzymes isolated from ΔnosZ and ΔnosX mutants demonstrate that NosX does not play a role in assembly of the Cu cofactors of N 2 OR under Cu-limited conditions.

NosX has a minor effect on transcription of nosZ under Cu-sufficient conditions
The data presented in Fig. 4, Table 1 revealed some variability in the extent to which Cu Z centres are assembled in enzymes isolated from different strains and grown under different conditions; specifically, plasmid-encoded strep-tagged N 2 OR isolated from wild-type cells contained fewer Cu Z centres than that from the two mutants. Thus, the ΔnosX mutant behaves similarly to the ΔnosZ mutant, in which chromosomal nosZ is missing. This suggests that there may be fewer chromosomally encoded versions of N 2 OR in the nosX mutant than in wild-type cells, as would be expected if nosZ expression is perturbed in the nosX mutant. Less chromosomally encoded N 2 OR would provide less competition for Cu, leading to greater incorporation of Cu into the plasmid-encoded N 2 OR.
To investigate this, qRT-PCR experiments were performed to determine the differential expression of nosZ in the nosX mutant compared to wild-type cells. Under Cu-sufficient conditions, a twofold decrease (1.9±0.2) in expression of nosZ was measured in ΔnosX compared to wild-type cells (Fig. 5). This likely contributes to the observed increased incorporation of Cu into step-tagged N 2 OR isolated from the ΔnosX mutant compared to that from wild-type cells. However, no significant difference in expression of nosZ was detected between ΔnosX and wild-type grown under Cu-limiting conditions. In both cases, the nosZ expression in Cu-limiting conditions was ~15-fold lower than that under Cu-sufficient conditions (Fig. 5), consistent with previous report on the effect of Cu on nosZ expression in wild-type cells [17]. Thus, effects on nosZ expression do not account for the very low incorporation of Cu into strep-tagged N 2 OR in wild-type cells compared to in the ΔnosX mutant.

DISCUSSION
The nosX gene is conserved across the NGC of αand β-proteobacteria, but not among γ-or clade-II members of N 2 O-reducing bacteria. Here, we have demonstrated a Nosphenotype for a nosX deletion mutant in P. denitrificans (PD2502), which was complemented in trans using a functional nosX plasmid-borne gene copy under taurine . †N 2 O reductase activity was determined for enzymes isolated from cultures grown under Cu-sufficient conditions using a reduced methyl viologen assay (µmol N 2 O min −1 mg −1 enzyme). Proteins were pre-incubated with a 500fold excess reduced methyl viologen for 150 min prior to activity assay. All reactions were carried out in triplicate and sd is shown. nd. The data show that even though the ΔnosX strain has a Nosphenotype, N 2 OR isolated from it is fully or close to fully active in an in vitro assay. inducible control. NosX is a member of the AbpE protein family, which bind flavin adenine dinucleotide [33,34]. Some AbpE proteins are flavinyl transferases, functioning in the post-translational maturation of another flavin-requiring protein. For example, Vibrio cholera ApbE transfers a flavin mononucleotide (FMN) to a threonine residue in NqrC [35]. In P. denitrificans there are three abpE homologues: nosX, encoded by pden_4214, nirX (pden_2485) and pden_3291. NosX and NirX are exported to the periplasm via the Tat pathway while Pden_3291 is predicted to be cytoplasmic.
An earlier study of an antibiotic cassette insertion mutation in the P. dentrificans nosX gene reported no effect on N 2 OR activity [22]. This led to the proposal that the nirX gene in P. denitrificans is a functional homologue of nosX, such that mutation of both genes are required in order to observe a Nosphenotype. This previous conclusion is clearly at odds with the data presented here. One possibly important observation is that the previous mutagenesis study did not involve full nosX deletion. Conserved residues within the putative FAD binding pocket in NosX are now known, including Ser68, Tyr70, Thr174 and Gly256, based on sequence similarities with the SeApbE (Fig. S1) [34]. If these residues are important for NosX function, then the previous mutation strategy for P. dentrificans nosX, in which a kanamycin insertion was made 469 bp into the gene, would not have disrupted the conserved Ser68 and Tyr70 residues. The resulting truncated NosX may have retained some function, which would account for why a clear phenotype was not observed in the single nosX insertional mutant. We note that the requirement for nosX in N 2 O reduction has also been demonstrated in Sinorhizobium meliloti. In that case, a Tn5-mediated mutation 31 nucleotides into the total 966 nucleotide sequence downstream of nosDFYL, a region now recognized as nosX, abolished N 2 OR activity [36].
In the earlier report of a double nosXnirX mutant of P. denitrificans, it was reported that the N 2 OR present in unfractionated periplasm from this mutant was deficient in the Cu A centre, leading to the conclusion that NosX and NirX play a role in assembly of this cofactor [22]. However, subsequent studies of anaerobically purified N 2 OR from the double nirXnosX mutant and a single nirX mutant indicated that the absence of NosX resulted in N 2 OR with both Cu cofactors assembled, but with Cu Z exhibiting a spectroscopically distinct from, termed pink Cu Z *, that is normally only observed upon reaction with O 2 [37]. This Cu Z form is not catalytically active, but is proposed to represent a catalytically relevant intermediate oxidation state of the Cu Z centre ([4CuS] 3+ ), which binds N 2 O and proceeds through a state denoted as Cu Z 0 [38].
Here, to determine the effect of the absence of nosX/NosX alone on N 2 OR, we utilized a previously reported plasmidencoded Strep-tagged N 2 OR that can be readily isolated from different background strains and characterized in terms of its Cu cofactor content and spectroscopic properties. These experiments demonstrated unequivocally that the assembly of the Cu A and Cu Z centres was unaffected in the absence of nosX. Thus, the phenotype exhibited by the mutant does not arise because of a deficiency in the insertion of Cu into N 2 OR. We note that the spectroscopic properties of N 2 OR from the ΔnosX mutant strain are the same as those of the Cu Z * centre from purified from the nosXnirX mutant. This may suggest the Cu Z centre was purified in a catalytically inactive redox state. However, the pink form reported in this work was generated by aerobic purification, with nirX remaining in the genome and under conditions where we expect to observe the Cu Z centre is this pink Cu Z * form, as demonstrated by the control experiments with N 2 OR isolated from the wild-type strain.
ApbE from the N 2 O-reducing bacterium P. stutzeri is a monomeric FAD-binding protein [23]. In the absence of nosX in the NGC of P. stutzeri, AbpE functions as a flavin donor, catalysing the covalent flavinylation of a threonine residue of NosR [23]. Importantly, the post-translationally modified, FMN-bound NosR is proposed to be the electron donor to N 2 OR, such that in the absence of NosR N 2 OR is not functional. Our data indicate that N 2 OR Cu cofactor maturation is unaffected by the loss of NosX, and we conclude that in P. denitrificans it most likely functions as the main system-specific maturation factor for NosR, and thus as an indirect activator of N 2 OR. If this is the case, then a Nosphenotype would be expected for a ΔnosR strain. This was recently demonstrated: a P. denitrificans ΔnosR strain exhibited a vastly decreased capacity to reduce N 2 O, irrespective of the levels of Cu in the cell [17]. However, we note that the ΔnosR strain did retain some ability to reduce N 2 O, whereas the nosX mutant investigated here did not, and so the nosX phenotype is actually more dramatic than the nosR phenotype. Why this is the case is not clear. One possibility is that NosX does not only mature NosR, such that in the absence of NosX, there is a further effect on NosZ activity. Alternatively, having a non-flavinylated NosR present might somehow inhibit NosZ more severely than having no NosR present at all. We also note that the previously reported transcription data revealed the loss of Cu-responsive transcription of nosZ in the nosR deletion strain [17], suggesting that NosR itself may be multifunctional, or that its absence leads to pleiotropic effects, some of which may be indirect. Clearly, further studies are needed to investigate directly the role of NosX in NosR maturation, and more generally other possible roles of NosX and the function(s) of NosR.
An intriguing observation reported here is the lower levels of Cu cofactor incorporation observed under Cu-limited conditions for the Strep-tagged N 2 OR from wild-type cells compared to that from the nosZ and nosX mutants. One possibility that we examined was that nosX/NosX is involved in the regulation of nosZ, such that in the absence of nosX/ NosX, lower amounts of chromosomally encoded N 2 OR were present, perhaps leading to less competition for copper and higher incorporation of Cu into the plasmid-encoded Streptagged form. While the absence of nosX did result in a twofold reduction of nosZ expression under Cu-sufficient conditions, no significant difference between the wild-type and nosX mutant strains was detected under Cu-limited conditions where the incorporation of Cu was most pronounced. The very low expression of the chromosomal nosZ gene under Cu-limited conditions suggests that a simple competition between chromosomal-and plasmid-encoded N 2 OR enzymes for Cu is unlikely. A further possibility is that the presence of the Strep-tag required for rapid recovery and biochemical analysis of NosZ results in modest perturbation of Cu cofactor assembly factor interactions such that the wild-type enzyme is a preferred substrate, an effect that only becomes apparent under very low Cu conditions. Clearly, further studies are needed to explore this possibility.
In summary, the data presented here show that nosX is essential for whole-cell N 2 O reduction in the α-proteobacterium P. denitrificans, and that the nosX and nirX gene products are not functionally redundant under our experimental conditions. The function of NosX is not associated with the assembly of the Cu cofactors of N 2 OR. Instead, based on homology between NosX and ApbE proteins, and the recent demonstration of an essential role for an ApbE family flavin transferase in the maturation of NosR in P. stutzeri, it is likely that NosX is involved in indirectly maintaining the reaction cycle of N 2 OR through the flavinylation of another accessory protein, NosR.

Conflicts of interest
The authors declare that there are no conflicts of interest.