Nitrate-responsive NarX-NarL represses arginine-mediated induction of the Pseudomonas aeruginosa arginine fermentation arcDABC operon

In Pseudomonas aeruginosa anaerobic metabolism is important for biofilm growth and for the infection of the cystic fibrosis lung involving biofilm-like P. aeruginosa microcolonies (Barraud et al., 2006; Filiatrault et al., 2006; Hassett et al., 2002; Palmer et al., 2007; Platt et al., 2008; Sauer et al., 2002; Van Alst et al., 2007; Worlitzsch et al., 2002; Xu et al., 1998). In the absence of oxygen P. aeruginosa grows either by denitrification or by arginine fermentation (Carlson & Ingraham, 1983; Vander Wauven et al., 1984; Zumft, 1997). The expression of genes which encode enzymes for denitrification in P. aeruginosa are controlled by the regulatory proteins Anr and Dnr, as well as the nitrate-responsive two-component system NarX-NarL (Arai et al., 1997; Schreiber et al., 2007; Ye et al., 1995). Enzymes required for arginine fermentation are encoded by genes which are organized in the operon arcDABC (Luthi et al., 1990). Anaerobic expression of arcDABC has been shown to be Anr-dependent and further stimulated by the arginine-responsive regulator ArgR (Lu et al., 1999). The global oxygen-sensing regulator Anr is an Escherichia coli Fnr homologue and is essential for fermentative and denitrifying growth in the absence of oxygen (Ye et al., 1995). Moreover, Anr controls the expression of dnr, which encodes the nitrogen-oxide-sensing regulator Dnr (Arai et al., 1997). Dnr finally induces expression of the genes encoding the nitrite, nitric oxide and nitrous oxide reductases required for denitrification (Arai et al., 1999, 2003). Mutant and gene expression studies in P. aeruginosa and Pseudomonas stutzeri showed that the two-component system NarX-NarL is additionally required for denitrifying growth (Härtig et al., 1999; Schreiber et al., 2007). The P. aeruginosa NarX protein is a sensor kinase and shares 38 % and 34 % sequence identity with NarX from P. stutzeri and E. coli, respectively (Härtig et al., 1999). The NarL protein is the corresponding response regulator to NarX and shares amino acid identity of 63 % and 58 % with NarL from P. stutzeri and E. coli, respectively (Härtig et al., 1999). In addition to NarX-NarL, E. coli harbours a second two-component system named NarQ-NarP (Chiang et al., 1992; Rabin & Stewart, 1993). However, so far no obvious narQ homologue has been functionally identified in the P. aeruginosa genome (Stewart, 2003). Recently, the NarL regulon of E. coli was reassessed in a comprehensive microarray experiment (Constantinidou et al., 2006). This approach revealed the activation of 51 operons and the repression of 41 operons by NarL. In E. coli the NarX-NarL system is involved in the coordination of anaerobic energy metabolism by inducing nitrate respiration and repressing less efficient modes of energy metabolism including fermentation processes (Constantinidou et al., 2006; Stewart & Rabin, 1995). In pseudomonads only a few genes have been experimentally confirmed to be targets of NarL activation. In P. stutzeri these are narG, narK and dnrE and in P. aeruginosa hemA, narK, nirQ and dnr (Härtig et al., 1999; Krieger et al., 2002; Schreiber et al., 2007; Vollack et al., 1999). However, so far no gene has been shown to be repressed by NarL. The effect of NarL on the transcription of genes involved in other, less favourable energy-generating systems than denitrification is unknown. In this study we investigated the effect of NarL on arginine fermentation in more detail using a combination of 2D gel electrophoresis, Northern blot analysis and lacZ reporter gene fusion experiments. Bacterial strains, plasmids and media.
The bacterial strains and plasmids used in this study are listed in Table 1. Bacteria were grown in LB medium (per litre: 10 g tryptone, 5 g yeast extract and 5 g NaCl) or modified AB minimal medium (Heydorn et al., 2000; Schreiber et al., 2006) as indicated. For anaerobic growth 50 mM KNO₃ was added. If appropriate, the antibiotic tetracycline was added at final concentrations of 5 µg ml^–1 for E. coli and 100 µg ml^–1 for P. aeruginosa. All incubations were carried out at 37 °C.

Table 1. Strains and plasmids used in this study

Proteome analysis.
P. aeruginosa cells were grown anaerobically in LB containing 50 mM KNO₃. Cultures in the exponential growth phase were mixed with twice the volume of ice-cold potassium phosphate buffer (100 mM; pH 7.4) and allowed to cool for 5 min. Cells were centrifuged at 8000 g for 20 min at 4 °C and washed twice with potassium phosphate buffer. Sedimented P. aeruginosa cells were resuspended in a small volume of potassium phosphate buffer and protein isolation was done as described previously (Schreiber et al., 2006). The protein concentration was determined in sample buffer using the PlusOne 2D Quant kit (Amersham Biosciences). The 2D gel electrophoresis was performed using immobilized pH gradient (IPG) strips of 17 cm length covering the pH range 5–8 (Bio-Rad). The IPG strips were rehydrated overnight in rehydration buffer containing 500 µg protein. Isoelectric focusing (IEF) was carried out at 20 °C under mineral oil in a PROTEAN IEF Cell (Bio-Rad) for a total of 110 kV h. The focused IPG strips were reduced for 15 min in an SDS equilibration solution containing 15 mM DTT and afterwards alkylated twice for 15 min in the same buffer containing 150 mM iodacetamide prior to SDS-PAGE. The IPG strips were transferred to 10 % polyacrylamide gels (25.5x20.5 cm). SDS-PAGE was performed at a constant temperature of 20 °C with 2 W per gel for approximately 20 h. All gels were stained with ruthenium(II)-tris(bathophenanthroline disulfonate) (RuBPS), as described before (Schreiber et al., 2006). Gels were documented with an FX-Scanner (Bio-Rad). Analysis and quantification of differential protein spot patterns was performed using the Software Z3 (Compugen). Two independent replicates were used for analysis. Proteins were identified by mass spectrometry as described previously (Schreiber et al., 2006).

Northern blot analysis.
For RNA extraction, cells were grown under oxygen-limited conditions in LB medium at 37 °C to an OD₅₇₈ of 0.2. RNA was prepared as described by Boes et al. (2006). Ten micrograms of RNA was separated electrophoretically on a 1 % agarose gel containing 5 % formaldehyde and then transferred to a nylon membrane. For hybridization, a digoxigenin-labelled arcA probe was used and the typical arcABC, arcAB and arcA transcripts were detected (Gamper et al., 1992). The arcA probe was generated using the primer pair arcA-NO-for (5'-CTGACCGAGACCATCCAGAA-3') and arcA-NO-rev (5'-CTAATACGACTCACTATAGGGAGACAGCAGGGTGTTGGTGTAGG-3'). DNA was labelled using the Digoxigenin RNA Labelling kit (Roche). CDP-Star (NEB) was used for detection.

Construction of chromosomal arcD promoter-lacZ reporter gene fusions and corresponding reporter gene assay.
The complete arcD promoter region (Fig. 3) was fused to E. coli lacZ. A 1005 bp PCR product was generated using the primer pair ArcDfor_A (5'-AACTGCAGGCTGCCGTGGCTCATGAT-3') and ArcDrev_B (5'-CGGGATCCTTTGCGGGAGGGAGAAGA-3'); the recognition sequences for PstI and BamHI are underlined. The resulting DNA fragment was digested with PstI and BamHI and cloned into the PstI/BamHI-digested mini-CTX-lacZ vector to generate pDH11 (Table 1). This plasmid was integrated into the attB site of the P. aeruginosa PAO1 genome and the corresponding narL mutant strain to generate strains BB43 and BB45, respectively (Table 1). Transfer of plasmids into P. aeruginosa was carried out by a diparental mating as described before (Schreiber et al., 2006). In the mutant strains BB43, BB45 and BB46, parts of the mini-CTX-lacZ containing the tetracycline-resistance gene were deleted using a FLP recombinase encoded on the pFLP2 plasmid (Hoang et al., 1998). The β-galactosidase activities of the strains carrying lacZ reporter gene fusions were determined in the early exponential phase at an OD₅₇₈ of 0.2 and are given in Miller units (MU) (Schreiber et al., 2006). Data are the result of at least three independent experiments.

(19K): Fig. 3. Nucleotide sequence of the arcDABC promoter region. The previously determined transcriptional start site (Galimand et al., 1991), the putative –10 region and the start of the arcD coding region are marked. Binding sites for the regulatory proteins ArgR, NarL and Anr are boxed. Bases identical to published consensus sequences (Galimand et al., 1991; Lu et al., 1999; Spiro & Guest, 1990) are marked by an asterisk. Compared to the ArgR box, the orientation of the NarL-binding site is in the opposite direction to the transcriptional start site (TACTCAA). The published consensus sequence used for NarL (Tyson et al., 1993) reads TACC/TNA/CT and differs in one position from that detected in the arc promoter.

Mutation of the putative NarL-binding site.
A potential NarL-binding site in the arcD promoter (pDH11) was mutated using the crossover PCR technique (Ho et al., 1989). The putative binding motif in the arcD promoter region for the regulatory protein NarL was detected using the Virtual Footprint tool of the PRODORIC database (Münch et al., 2005). Mutation of the NarL box (TACTCAA→CATTCAA) was based on the published NarL consensus binding sequence (Tyson et al., 1993). The mutation was introduced using the primer pair oBB42 (5'-AATAGCTTCCCATTCAAAGTAATTAGAT-3') and oBB43 (5'-ATCTAATTACTTTGAATGGGAAGCTATT-3'; the mutated NarL-binding site is underlined) to generate pBB20. Nucleotide exchanges were verified by DNA sequence determination. The corresponding vector pBB20 was integrated into the attB site of the P. aeruginosa strain PAO1 chromosome to generate strain BB46 (Table 1). Proteome analysis of the narL mutant strain
2D gel electrophoresis of water-soluble proteins was employed to monitor the NarL-dependent production of proteins in P. aeruginosa. We compared the protein patterns of the P. aeruginosa narL mutant PAO9104 with that of the wild-type PAO1. Seven proteins were found differentially synthesized in the narL mutant (Fig. 1, Table 2). We detected increased protein concentrations of NarH in wild-type cells. The NarH protein is the soluble, cytoplasmic subunit of the membrane-bound nitrate reductase, encoded by the narK₁K₂GHJI operon (Palmer et al., 2007; Schreiber et al., 2007; Sharma et al., 2006). This is in agreement with earlier investigations, which described that in P. aeruginosa and P. stutzeri the NarX-NarL two-component system activates transcription of the nitrate reductase operon (Härtig et al., 1999; Schreiber et al., 2007). The second protein found in increased concentrations in wild-type cells was porin E1 (OprE). The porin protein E1 gene oprE has been described before to be expressed in response to anaerobiosis (Yamano et al., 1993). Furthermore, the two-component class II ribonucleotide reductase subunits NrdJa and NrdJb showed an increased concentration in wild-type cells, suggesting a NarL-dependent induction of the corresponding genes (Torrents et al., 2005). Both proteins were recently found induced under anaerobic conditions in the presence of nitrate (Platt et al., 2008). Only three proteins with increased concentration in the narL mutant compared to PAO1 wild-type were found, suggesting a repression of the corresponding genes by NarL. MALDI-TOF analysis identified these proteins as arginine deiminase (ArcA), catabolic ornithine carbamoyltransferase (ArcB) and carbamate kinase (ArcC) (Table 2). All three proteins are encoded by genes of the arcDABC operon and represent the arginine fermentation pathway for ATP generation.

Table 2. Two representative gel images from three replicate gels are shown. Proteins were stained with RuBPS (see Methods).

Table 2. Identified proteins differentially synthesized in the P. aeruginosa wild-type compared to the narL mutant

Northern blot analysis of the arcDABC mRNA
No gene repression by NarL has previously been described for pseudomonads. Consequently, we focused our investigation on the observed decrease in ArcA, ArcB and ArcC protein concentration. In order to determine whether the regulation occurs on the mRNA or on the protein level, we performed Northern blot analyses using a digoxigenin-labelled arcA probe. The detection of three abundant transcripts, arcABC, arcAB and arcA, has been described before (Gamper et al., 1992). First, we compared RNA levels prepared from P. aeruginosa wild-type cells grown under oxygen limitation in LB medium either with or without 50 mM nitrate. Strong signals corresponding to the arcABC, arcAB and arcA transcripts were detected only for RNA prepared from wild-type cells grown without nitrate (see Fig. 2). We did not detect an arcDABC transcript. However, its low abundance was described before (Gamper et al., 1992). Interestingly, the probing of RNA prepared from the narL mutant strain PAO9104 grown with 50 mM nitrate resulted in almost identical signal intensities. These results clearly indicate that NarL represses arginine fermentation on the transcriptional level.

(75K): Fig. 2. Northern blot analysis of transcripts from the arcDABC operon with RNA prepared from P. aeruginosa wild-type PAO1 grown under oxygen limitation in LB medium with 50 mM nitrate (lane 1), the wild-type PAO1 grown without nitrate addition (lane 2) and the P. aeruginosa narL mutant PAO9104 incubated in LB medium supplemented with 50 mM nitrate (lane 3). Total RNA (10 µg) was isolated after 3 h incubation at 37 °C. The mRNAs were detected using an arcA-specific probe.

The arcDABC operon is repressed by NarL
We used transcriptional P_arcD-lacZ fusions to investigate the influence of NarL on the arcDABC promoter. We determined β-galactosidase activities of strains BB43 (wild-type) and BB45 (narL mutant) under identical conditions as described for the proteome analysis experiments. β-Galactosidase activities of P_arcD-lacZ increased more than threefold in the narL mutant (BB45, 1975 MU) compared to the wild-type (BB43, 550 MU) under anaerobic conditions in the presence of nitrate. This observation suggests that the absence of the NarL regulator leads to an increased transcription of arcD. Consequently, NarL is responsible for arcDABC repression.

Identification of the NarL-binding site in the arcDABC promoter
Next, we were interested to know if the observed decrease in the arcDABC mRNA level and corresponding in vivo promoter activity is mediated by direct repression of the arcDABC promoter via NarL. Earlier investigations showed that the arcDABC operon is expressed in response to oxygen limitation in an Anr-dependent manner (Galimand et al., 1991) and that the presence of arginine increases transcription via the ArgR regulator (Lu et al., 1999). The arcDABC promoter region contains an Anr-binding site at –41.5 nt (Gamper et al., 1991) and an ArgR-binding site, which spans from –94 nt to –53 nt relative to the transcriptional start site (Lu et al., 1999). We searched the arcD promoter for the presence of a NarL-binding site using the Virtual Footprint tool from the PRODORIC database (Münch et al., 2005). This analysis revealed one conserved heptameric NarL-binding site located –60 nt upstream of the published transcriptional start site shown in Fig. 3 (Gamper et al., 1991). The position of the NarL heptameric binding site at –60 nt overlaps with the ArgR-binding site and suggests that NarL interferes with ArgR but not with the Anr regulator. In order to functionally confirm the bioinformatics prediction, the putative heptameric NarL-binding site was mutated (TACTCAA→CATTCAA) based on the published NarL consensus binding sequence from E. coli (TAC^C/_TN^A/_CT) (Tyson et al., 1993). Since the NarL box partly overlaps with the ArgR box we carefully selected positions for mutagenesis in order to avoid the change of nucleotides known to be important for ArgR binding (see Fig. 3 for details) (Lu et al., 1999). We determined β-galactosidase activities of strain BB46 containing the mutated P_arcDΔNarL-lacZ fusion in the wild-type strain. The activities of BB46 (1696 MU) were similar to those obtained for the P_arcD-lacZ in the narL mutant BB45 (1975 MU). This result confirms a direct repression of the arcDABC promoter via NarL.

Arginine-dependent ArgR activation of arcDABC is repressed by nitrate-dependent NarL
To elucidate whether NarL interferes with the Anr or ArgR regulator at the arcD promoter, we extended our reporter gene experiments for the separate addition of nitrate and arginine to defined growth medium. The results are given in Table 3. The highest β-galactosidase activities of the P_arcD-lacZ fusion in the wild-type (BB43) and the narL mutant strain (BB45) were measured for anaerobic conditions in the presence of 20 mM arginine. The addition of nitrate decreased β-galactosidase activities by 55 % in the wild-type, but by only 24 % or 19 % in the narL mutant strain or the wild-type strain containing the P_arcDΔNarL-lacZ fusion with the mutated NarL-binding site in the arcD promoter region (Table 3). These results clearly show a NarL-dependent repression of the arginine-mediated activation of the arcD promoter via the proposed NarL-binding site.

Table 3. Expression of the ParcD–lacZ reporter gene fusions in the P. aeruginosa wild-type and the narL mutant

Three enzymes of the arginine deiminase pathway, ArcA, ArcB and ArcC, were produced in higher amounts under anaerobic denitrifying conditions in the P. aeruginosa narL mutant strain, indicating a repression of the corresponding genes by NarL. It has been shown previously that the specific activity of the catabolic ornithine carbamoyltransferase (ArcB) of the arginine deiminase pathway is repressed by nitrate (Mercenier et al., 1980). However, it was unknown whether this repression was mediated by the NarL regulator itself. Our Northern blot analysis and experiments with P_arcD-lacZ reporter gene fusions of the PAO1 wild-type and the PAO9104 mutant strains clearly indicated a repression at the transcriptional level.

A bioinformatics analysis of the arcD promoter region identified a putative heptameric NarL-binding site at –60 nt to the transcriptional start site. This putative binding site (TACTCAA) differs only in one position from the published E. coli NarL consensus binding sequence (TAC^C/_TN^A/_CT) (Tyson et al., 1993). This putative NarL-binding site overlaps with the binding site of the ArgR regulator, which complicated mutagenesis of the NarL-binding site. However, we carefully selected positions for mutagenesis in order to avoid the change of nucleotides known to be important for ArgR binding (see Fig. 3 for details) (Lu et al., 1999). The consensus sequence of the ArgR-binding site has been deduced from DNase I footprinting studies (Lu et al., 1999, 2004). Control experiments in AB minimal medium with arginine confirmed that the mutation left the ArgR-binding site functional. It showed that β-galactosidase activities of strain BB45, which carries the reporter gene fusion with the mutated NarL-binding site, increased fivefold in response to arginine compared to the reporter gene fusion with the intact NarL-binding site. The position of the putative NarL-binding site, as well as the results of the reporter gene fusion experiments, suggests the following model. The NarX-NarL regulatory system of P. aeruginosa is employed for activation of nitrate respiration and downregulation of arginine fermentation under anaerobic conditions (Fig. 4). However, NarL does not completely abolish the expression of the arcDABC operon. In the presence of nitrate and arginine, NarL binding most likely prevents interaction of the arginine-dependent ArgR activator with its overlapping binding site and represses the further induction of arcDABC transcription by ArgR. No repression of arcDABC transcription by NarL was observed in the absence of arginine (Table 3). This finding is in agreement with the concept of a double role for NarL as an activator for nitrate reductase formation as well as a repressor of energetically less effective fermentative pathways.

(14K): Fig. 4. Schematic representation of NarX-NarL-dependent regulation. The major regulator for anaerobic growth is Anr, which activates arcDABC even in the absence of arginine and nitrate. In the presence of arginine the ArgR regulator stimulates transcription of the arcDABC operon. If arginine and nitrate are available, the NarL regulator represses the ArgR-dependent arcDABC stimulation by binding to the corresponding promoter region. On the other hand, expression of the nitrate reductase operon narK1K2GHJI, which catalyses the first step in denitrification, is induced by the nitrate-response system NarX-NarL.

Moreover, our findings are also in agreement with data recently published by Platt et al. (2008). The authors observed expression of the arginine deiminase pathway under anaerobic denitrifying conditions in complex medium containing arginine. Under these conditions NarL prevents additional activation of arcDABC via ArgR, but not the basic expression mediated by the Anr regulator.

Our proteomics approach revealed four additional proteins with decreased concentration in the narL mutant strain. Only the corresponding gene of NarH, which is a part of the narK₁K₂GHJI operon, has been shown to be under direct control of the NarL regulator (Schreiber et al., 2007). We did not identify any highly conserved NarL-binding sites in the putative promoter regions of the genes for the remaining three proteins (data not shown). Therefore, a direct control via NarL seems unlikely. Moreover, proteins homologous to OprE, NrdJa and NrdJb are not part of the E. coli NarL regulon (Constantinidou et al., 2006).

We thank H. P. Schweizer (University of Colorado, USA) for providing the mini-CTX-lacZ and pFLP2 plasmids and Dana Heldt from our laboratory for the construction of pDH11. The investigation was founded by the Deutsche Forschungsgemeinschaft, the German Research Centre for Biotechnology and the Fonds der Chemischen Industrie. K. S. was supported by the DFG-European Graduate College 653.

Edited by: M. A. Kertesz

Footnotes

†Present address: Institute of Biochemistry and Biotechnology, Technical University Braunschweig, Spielmannstr. 7, D-38106 Braunschweig, Germany.