The Anaerobic Regulatory Network Required for Pseudomonas aeruginosa Nitrate Respiration

Abstract

In Pseudomonas aeruginosa, the narK₁K₂GHJI operon encodes two nitrate/nitrite transporters and the dissimilatory nitrate reductase. The narK₁ promoter is anaerobically induced in the presence of nitrate by the dual activity of the oxygen regulator Anr and the N-oxide regulator Dnr in cooperation with the nitrate-responsive two-component regulatory system NarXL. The DNA bending protein IHF is essential for this process. Similarly, narXL gene transcription is enhanced under anaerobic conditions by Anr and Dnr. Furthermore, Anr and NarXL induce expression of the N-oxide regulator gene dnr. Finally, NarXL in cooperation with Dnr is required for anaerobic nitrite reductase regulatory gene nirQ transcription. A cascade regulatory model for the fine-tuned genetic response of P. aeruginosa to anaerobic growth conditions in the presence of nitrate was deduced.

The most efficient way for the gram-negative bacterium Pseudomonas aeruginosa to generate energy in the absence of oxygen is through denitrification. During this process, molecular oxygen is replaced by nitrate as the terminal electron acceptor. Nitrate (NO₃⁻) is reduced in four consecutive steps, via nitrite (NO₂⁻), nitric oxide (NO), and nitrous oxide (N₂O) to dinitrogen (N₂). This process is vital for growth and survival under microaerobic and anaerobic conditions as found in biofilms and microcolonies of infectious P. aeruginosa (1, 25a). The majority of earlier investigations focused on the enzymology and regulation of nitrite (NO₂⁻)-to-dinitrogen (N₂) conversion (27). Here, the regulatory network for the onset of nitrate respiration under oxygen-limiting conditions was elucidated using reporter gene fusions, strains carrying mutated regulatory genes, and site-directed mutagenesis of potential regulator binding sites.

Importance of narGHJI, narXL, anr, and dnr for anaerobic growth of P. aeruginosa.

In order to confirm the importance of the nitrate reductase genes narGHJI and the regulatory genes anr, dnr, and narXL for the anaerobic growth of P. aeruginosa, knockout mutants were characterized concerning their growth behavior. P. aeruginosa Anr is the oxygen-sensing regulatory protein homologue to Escherichia coli Fnr (19, 26). Dnr of P. aeruginosa belongs to the Crp-Fnr superfamily of transcriptional regulators and was reported to activate transcription of the genes nir, nor, and nos (6, 9). In Pseudomonas stutzeri, DnrD was shown to detect NO (13, 22). NarXL is a nitrate-responding two-component regulatory system (14). All investigated P. aeruginosa mutant strains failed to grow under anaerobic nitrate respiratory conditions (data not shown). They did not reveal any growth phenotype when tested under aerobic conditions (data not shown). These experiments identify narL, anr, dnr, and narG as key players in the anaerobic growth of P. aeruginosa.

Transcriptional control of the nar locus is mediated by the narXL-narK₁ intergenic region.

In E. coli and P. stutzeri, Fnr- and NarXL-dependent transcription of the narGHJI operon is mediated by the narG upstream region (8, 15, 23, 24). In contrast to these observations, inspection of the 200-bp 5′ upstream region of narG in P. aeruginosa revealed no obvious binding motifs for the Fnr homologue Anr or the nitrate response regulator NarL. Using the two PnarG1 and PnarG2 reporter gene fusions, containing 100 bp and 411 bp of the upstream region of the narG gene fused to the reporter gene lacZ in the pQF50 plasmid, respectively, we showed that these DNA fragments did not mediate transcriptional activation of PnarG-lacZ under any of the tested aerobic and anaerobic growth conditions (Table 1). Information regarding primer sequences, construction of all reporter gene fusions, mutated promoter constructs, and various strains can be provided upon request. We failed to detect a transcriptional start site using primer extension experiments upstream of narG (data not shown). Additionally, the intergenic region between narK₂ and narG was successfully amplified from cDNA synthesized from mRNA extracted from anaerobically grown P. aeruginosa PAO1 cells, confirming that narG is cotranscribed with the narK₁K₂ genes located upstream. Consequently, the 173-bp DNA fragment localized between narX and narK₁ harbors two divergently oriented promoters (Fig. 1A). This was also recently detected using a promoter predictor program and mutational analysis (19a). The 5′ end of narK₁K₂GHJI mRNA was localized at 29 bp upstream of the translational start codon of narK₁, and the 5′ end of the narX mRNA was localized at 45 bp upstream of the translational start codon of narX, by using primer extension analysis on an ALF model DNA sequencer (Pharmacia) (16).

TABLE 1.

Regulation of the narK₁K₂GHJI transcription by Anr, Dnr, IHF, and NarL

P. aeruginosa strain (genotype)	Promoter constructc	Gene in transb	β-Galactosidase activity (Miller units)a
			+ O2 + nitrate	+ O2 + nitrite	+ O2	− O2 + nitrate	− O2 + nitrite	− O2 fermentative
PAO1	PnarG1	—	3	4	5	11	10	7
PAO1	PnarG2	—	9	8	6	10	10	8
PAO1	PnarK1	—	11	10	12	720	390	159
PAO6261 (Δanr)	PnarK1	—	6	6	5	8	12	7
PAO6261 (Δanr)	PnarK1	dnr	—	—	—	134	60	53
PAO1	PnarK1	dnr	—	—	—	761	399	329
RM536 (Δdnr)	PnarK1	—	11	10	9	256	195	104
RM536 (Δdnr)	PnarK1	anr	—	—	—	678	161	268
PAO1	PnarK1	anr	—	—	—	846	200	441
PAO1	PnarK1ΔAnr	—	8	6	5	19	16	21
CHA-A2 (ΔihfA)d	PnarK1	—	4	5	5	72	53	37
PAO1	PnarK1ΔIHF	—	8	7	6	299	152	183
PAO9104 (ΔnarL)	PnarK1	—	4	4	5	14	10	22
PAO9104 (ΔnarL)	PnarK1	narL	9	8	10	630	378	126
PAO1	PnarK1ΔNarL1	—	10	9	10	32	33	30
PAO1	PnarK1ΔNarL2	—	11	8	11	147	107	91
PAO1	PnarK1ΔNarL3	—	2	—	—	138	—	—

^aAll values are results of three independent experiments performed in triplicate. Bacterial strains were grown aerobically and anaerobically in LB medium supplemented with 10 mM sodium nitrate or 1 mM sodium nitrite as described previously (7). Arginine fermentation conditions were achieved as outlined previously (21). The β-galactosidase activities are given in Miller units and are calculated as described previously (18). Standard deviations were between 3 and 13% of given values. —, not measured.

^bThe plasmids pHA411 (anr) and pHA541Ω (dnr) provided the anr and dnr genes in trans as described previously (4, 5). Vector without an insert served as background control. —, no gene in trans.

^cDetails for the construction of tested reporter gene fusions and P. aeruginosa mutants will be provided upon request.

^dCHA-A2 is not isogenic to PAO1. Due to the clear cut results obtained with this strain, the data obtained are presented.

FIG. 1.

(A) Schematic representation of the nar locus in Pseudomonas aeruginosa. The narXL and narK₁K₂GHJI operons share a divergently oriented promoter region covering 173 bp. The 5′ end of the narXL mRNA was localized at 45 bp upstream of the translational start of narX. The narK₁K₂GHJI genes are transcribed as an operon starting at 29 bp upstream of the narK₁ translational start. (B) The currently elucidated regulatory network for the onset of denitrification in P. aeruginosa. The major initial signal to turn on the denitrification pathway in P. aeruginosa is low-oxygen tension. This signal is measured by the Fe-S clusters attached to Anr (27). Anr increases the transcription of the narXL operon encoding a two-component regulatory system responding to the presence of nitrate. Anr and NarL cooperatively activate the dnr gene for the third involved regulatory system, Dnr, responding to NO (5). The fourth regulatory system, NirQ, in turn requires Dnr and NarL for its formation (4). Under low-oxygen tension conditions and in the presence of nitrate, Anr and NarL activate in concert with the DNA bending protein IHF, the narK₁K₂GHJI operon encoding nitrate/nitrite transporters, and the structural genes for the respiratory nitrate reductase. The enzyme converts nitrate into nitrite. Now, both Dnr and NirQ, most likely responding to N-oxides, are essential for the gene regulatory scenario required for the formation of the three enzyme complexes that catalyze the conversion of nitrite into N₂.

Regulation of the narK₁K₂GHJI promoter.

A reporter gene fusion, PnarK₁-lacZ containing 435 bp upstream of the translational start codon of narK₁ was constructed. The reporter gene activity was found to be induced 65-fold when P. aeruginosa was grown under anaerobic conditions in the presence of nitrate compared to that under identical but aerobic conditions. A 39-fold induction was detected when nitrate was replaced with nitrite, and still a 13.3-fold induction was present during arginine fermentative conditions (Table 1). Subsequently, the PnarK₁ construct was introduced into the two P. aeruginosa regulatory mutant strains PAO6261 (Δanr) and RM536 (dnr::Tc). Aerobic narK₁ promoter activity in the wild-type and in the two mutant strains was comparably low (Table 1). No obvious narK₁ promoter activity was detected in the anr mutant strain under any of the tested anaerobic conditions. Deletion of the dnr gene resulted in 2.8-fold lower promoter activity levels than that in the wild-type strain under anaerobic growth conditions in the presence of nitrate. A 2-fold and 1.5-fold reduction of wild-type reporter gene activity was observed for anaerobic growth in the presence of nitrite and under fermentative growth conditions, respectively. Thus, Anr is essential, and Dnr enhances narK₁ activation (Fig. 1B). A regulatory cascade was proposed where Anr activates the expression of dnr, which in turn activates target promoters (5). To determine whether both regulators act in a coordinated manner, gene complementation assays were performed. The anr mutant strain carrying the PnarK₁-lacZ promoter reporter gene fusion was transformed with the dnr gene on an expression plasmid (pHA541Ω) in trans. Reporter gene expression increased significantly compared to that of the anr mutant without dnr in trans. However, the values obtained were below those of reporter gene expression in the wild-type strain or the wild-type strain carrying a dnr expression plasmid (Table 1). A minor but significant role for Dnr in the anaerobic activation of the narK₁ promoter was indicated, in which Dnr function was independent of Anr. Introducing an anr expression plasmid (pHA411) into the dnr mutant led to narK₁ promoter activity levels that were almost the same as those of the wild type (Table 1). Consequently, the narK₁ promoter represents another example of the dual actions of Anr and Dnr, besides those of the P. aeruginosa hemF and hemN promoters (Fig. 1B) (18). A highly conserved potential Anr/Dnr binding site (5′-TTGATTCCTATCAA-3′; conserved nucleotides in the Anr binding site are underlined) centered at −40.5 bp upstream of the 5′ end of the narK₁ mRNA, was detected using Virtual Footprint of PRODORIC software (www.prodoric.de). Three potential NarL binding sites, NarL1 (5′-TACCTCT-3′) at −108 bp, NarL2 (5′-TACGGCT-3′) at −113 bp, and NarL3 (5′-TACCTCC-3′) at −208 bp with respect to the narK₁ mRNA 5′ end, were found. One putative binding site for the DNA bending protein IHF (5′-CAATAATTTCAGCCG-3′) was proposed at −119 bp upstream of the narK₁ mRNA 5′ end.

Next, nucleotide exchanges (5′-TTGATTCCTATCAA-3′ to 5′-TCGATTCCTACTTA-3′) were introduced into the Anr box consensus motif (PnarK₁ΔAnr). Putative regulator binding sites were mutated using a QuikChange mutagenesis kit (Stratagene, Amsterdam, The Netherlands) or via crossover PCR (12) (details can be provided on request). No obvious promoter activity was detected under any of the tested conditions (Table 1). These results confirm the importance of the putative Anr binding site centered at −40.5 bp upstream of the transcriptional start for anaerobic induction of narK₁. The PnarK1-lacZ reporter gene fusion was next introduced into the IHF (CHA-A2) and narL (PAO9104) mutant strains. The β-galactosidase activities were found to be significantly reduced (10-fold and 51-fold, respectively) for both mutants, indicating the involvement of both proteins in narK₁ transcription (Table 1 and Fig. 1B). Complementation of the narL mutant with a narL expression plasmid (pRK-LM) nearly restored wild-type-level promoter activity (Table 1). To verify the importance of a predicted IHF binding motif within the narK₁ promoter, the motif was mutated from 5′-CAATAATTTCAGCCG-3′ to 5′-GGGGAATTTCAGCCG-3′ (PnarK₁ΔIHF). The 2.4-fold decrease observed for the promoter activity of wild-type P. aeruginosa strain PAO1 carrying a PnarK₁ΔIHF-lacZ fusion indicated the importance of the predicted IHF binding motif in narK₁ promoter activation (Table 1). The failure to completely eliminate narK₁ promoter activity with the introduced mutations might be due to residual IHF binding capacity of the mutated site or to an additional as-yet-unknown second IHF binding site in the narK₁ promoter. Additionally, the IHF protein can be replaced by less binding sequence-specific HU proteins (11, 17). The potential NarL binding sites NarL1 and NarL3 were correctly oriented for narK₁ activation; NarL2 was oriented in the opposite direction. Mutagenesis of the NarL1 binding site (PnarK₁ΔNarL1) resulted in a nearly total loss of reporter gene activities under denitrifying conditions (Table 1). Mutagenesis of NarL2 (PnarK₁ΔNarL2) decreased narK₁ promoter activity down to a level which was in the range of that of the IHF mutant. Since the NarL2 and the IHF binding motifs overlap, secondary effects of the NarL2 mutagenesis on the IHF binding site cannot be excluded. Mutagenesis of the third NarL binding site, NarL3 (PnarK₁ΔNarL3), abolished reporter gene activities, as shown in Table 1. Consequently, NarL1 and NarL3 are at least important for the anaerobic induction of the narK₁ promoter.

P. aeruginosa possesses a second dissimilatory nitrate-reducing system localized in the periplasm, encoded by the napEFDABC operon. No obvious influence of tested regulators, oxygen tension and nitrate or nitrite availability on napEFDABC promoter activity was detected using a napE promoter reporter gene fusion (Table 1). Clearly, napEFDABC expression is not coregulated with the onset of denitrification.

Regulation of the narXL promoter.

To study the regulation of the narXL promoter, a PnarX-lacZ fusion carrying 206 bp of the narX promoter region (PnarX) was constructed. A moderate 1.7- to 2.0-fold promoter induction was observed during anaerobic nitrate and nitrite respiratory and fermentative conditions (Table 2). The aerobic constitutive narXL transcription was independent of all regulators tested (Table 2). Moreover, P. aeruginosa NarXL did not autoregulate its own gene expression and did not require IHF for expression (8, 20). However, P. aeruginosa narX promoter activity was slightly reduced under any of the tested anaerobic conditions in the case of a missing Anr or Dnr regulator (Table 2 and Fig. 1B). To distinguish between cascade regulation and dual activities by Anr and Dnr, the dnr gene was expressed in trans from a plasmid in the anr mutant PAO6261 carrying the PnarX-lacZ reporter gene fusion. As a control, anr was expressed in the dnr mutant RM536 harboring the same fusion. In both experiments, full anaerobic expression was not restored, indicating a dual function for both regulators at the narXL promoter. For weaker Anr-dependent promoters in P. aeruginosa, Anr half-site reactivity has been reported (14, 25). Therefore, both half sites of the potential Anr binding site centered at −60.5 bp upstream of the narX mRNA 5′ end were mutated independently, from 5′-TTGATTCCTATCAA-3′ to 5′-TGAATTCCTATCAA-3′ in PnarXΔAnr1 and to 5′-TTGATTCCTAAGAA-3′ in PnarXΔAnr2. Decreased reporter gene activities of both mutated narX reporter gene fusions were at the levels observed for the reporter gene fusions with the intact narXL promoter tested with the regulatory mutants PAO6261 (Δanr) and RM536 (dnr::Tc), respectively (Table 2). These results demonstrated that both half sites of the Anr/Dnr binding sequence are involved in anaerobic narX transcriptional activation. The location of the potential Anr/Dnr binding site at −60.5 bp might provide an explanation for the moderate 1.7-fold induction of the narX promoter under anaerobic conditions.

TABLE 2.

Anaerobic induction of the napE and narXL promoters by NarL, IHF, Anr, and Dnr and NarL influence on the expression of Pdnr-lacZ, PnirQ-lacZ, Panr-lacZ, PnirS-lacZ, and PnorC-lacZ

P. aeruginosa strain	Promoter constructc	Gene in transb	β-Galactosidase activity (Miller units)a
			+ O2 + nitrate	+ O2 + nitrite	+ O2	− O2 + nitrate	− O2 + nitrite	− O2 fermentative
PAO1	PnapE	—	352	410	361	291	305	310
PAO6261 (Δanr)	PnapE	—	402	576	408	350	361	338
RM536 (Δdnr)	PnapE	—	238	253	262	286	284	239
PAO9104 (ΔnarL)	PnapE	—	322	313	292	400	396	320
PAO1	PnarX	—	293	269	251	504	539	519
PAO9104 (ΔnarL)	PnarX	—	220	225	216	523	426	580
CHA-A2 (ΔihfA)d	PnarX	—	300	310	285	526	549	530
PAO6261 (Δanr)	PnarX	—	246	259	241	367	239	336
PAO6261 (Δanr)	PnarX	anr	272	—	—	622	—	—
PAO6261 (Δanr)	PnarX	dnr	272	—	—	372	—	—
RM536 (Δdnr)	PnarX	—	259	280	257	259	291	282
RM536 (Δdnr)	PnarX	dnr	257	—	—	582	—	—
RM536 (Δdnr)	PnarX	anr	298	—	—	398	—	—
PAO1	PnarXΔAnr1	—	194	216	200	280	320	316
PAO1	PnarXΔAnr2	—	241	269	228	331	393	365
PAO1	Pdnr	—	7	8	5	210	132	122
PAO9104 (ΔnarL)	Pdnr	—	8	6	7	12	14	15
PAO1	PnirQ	—	17	22	12	327	427	60
PAO9104 (ΔnarL)	PnirQ	—	8	6	8	32	53	24
PAO9104 (ΔnarL)	PnirQ	dnr	11	12	10	5	12	7
PAO1	Panr	—	103	114	105	233	293	322
PAO9104 (ΔnarL)	Panr	—	81	93	92	191	206	268
PAO1	PnirS	—	12	23	29	470	852	268
PAO9104 (ΔnarL)	PnirS	—	15	14	15	321	627	331
PAO9104 (ΔnarL)	PnirS	dnr	13	—	—	521	—	—
PAO1	PnorC	—	18	28	25	1053	1642	493
PAO9104 (ΔnarL)	PnorC	—	22	17	18	480	1627	307
PAO9104 (ΔnarL)	PnorC	dnr	25	—	—	880	—	—

^aAll values are results of three independent experiments performed in triplicate. Bacterial strains were grown aerobically and anaerobically in LB medium supplemented with 10 mM sodium nitrate or 1 mM sodium nitrite as described previously (7). Arginine fermentation conditions were achieved as outlined previously (21). The β-galactosidase activities are given in Miller units and are calculated as described previously (18). Standard deviations for the experiment shown were between 3 and 12%. —, no gene in trans.

^bThe anr and dnr genes were cloned in a vector compatible with the lacZ fusion vector pQF50 and cotranscribed where indicated. Vector without an insert served as background control. —, not measured.

^cDetails for the construction of tested reporter gene fusions and P. aeruginosa mutants will be provided upon request.

^dCHA-A2 is not isogenic to PAO1. Due to the clear cut results obtained with this strain, the data obtained are presented.

The NarL regulator and dnr, nirQ, anr, nirS, and norC transcription.

Transcriptional lacZ fusions of the regulatory genes dnr, nirQ, and anr and the structural genes nirS (encoding nitrite reductase) and norC (encoding nitric-oxide reductase cytochrome c subunit) were introduced into P. aeruginosa PAO1 and the narL mutant strain PAO9104. In agreement with previous findings, the Pdnr-lacZ fusion was found to be induced by Anr under anaerobic conditions (5). Moreover, dnr transcription was highly dependent on NarL (Table 2 and Fig. 1B). In agreement, highly conserved NarL binding sites were localized, centered at 189 bp and 114 bp upstream of the translational start point. The ATP-binding protein NirQ from P. aeruginosa activates respiratory nitrite reductase activity (10). The expression of the PnirQ-lacZ fusion was found to be strongly induced under anaerobic conditions. Like Pdnr-lacZ expression, PnirQ-lacZ transcription was found to be dependent on NarL (Table 2 and Fig. 1B). Arai et al. described Anr- and Dnr-dependent nirQ expression (4). Since our experiments showed that expression of dnr in turn is also dependent on NarL, we investigated whether the observed NarL dependence of the nirQ promoter was caused by an indirect effect via a NarL-dependent decrease of Dnr levels. The nirQ promoter activities were still found to be low even in the presence of a high Dnr concentration, which confirmed a direct effect of NarL on nirQ expression (Fig. 1B). In agreement with these observations, the nirQ promoter harbored potential NarL binding sites centered at positions 97 bp and 4 bp upstream of the translational start site. Clearly, Dnr- and NarL-dependent NirQ formation provides the regulatory link between nitrate and nitrite respiration. We did not observe NarL dependence of anr expression (Table 2). The nitrite reductase nirS gene expression occurred under anaerobic conditions and was induced by nitrate and nitrite (3). Anr mediates this transcriptional activation indirectly via the induction of dnr expression in the regulatory cascade, as outlined previously (2, 4). Anaerobic PnirS-lacZ induction in the narL mutant in the presence of nitrate or nitrite was only about 1.4-fold lower than that in wild-type PAO1 (Table 2). Therefore, the effect of NarL on nirS transcription seemed to be of an indirect nature, most likely via the missing induction of dnr (Fig. 1B). Consequently, dnr expression in trans complemented the observed expression phenotype. Additionally, the narL mutant, expressing the structural genes for the nitrate reductase at a low level, failed to efficiently convert nitrate into nitrite. However, high nitrite levels are important for nirS induction as well. Almost identical results were obtained for the analysis of a PnorC-lacZ fusion in a narL mutant. Transcription of norC is also dependent on Anr and Dnr and the presence of nitrite (2). The observed strong nitrite-dependent induction is not completely understood. Either dnr transcription is activated via formed N-oxides, or an additional as-yet-unknown system is involved. Obtained results clearly showed that norC transcription is not directly regulated by NarL (Table 2 and Fig. 1B). Again, an indirect effect via NarL-dependent dnr expression was concluded.