The ArgR Regulatory Protein, a Helper to the Anaerobic Regulator ANR during Transcriptional Activation of the arcD Promoter in Pseudomonas aeruginosa

Abstract

Pseudomonas aeruginosa, when deprived of oxygen, generates ATP from arginine catabolism by enzymes of the arginine deiminase pathway, encoded by the arcDABC operon. Under conditions of low oxygen tension, the transcriptional activator ANR binds to a site centered 41.5 bp upstream of the arcD transcriptional start. ANR-mediated anaerobic induction was enhanced two- to threefold by extracellular arginine. This arginine effect depended, in trans, on the transcriptional regulator ArgR and, in cis, on an ArgR binding site centered at −73.5 bp in the arcD promoter. Binding of purified ArgR protein to this site was demonstrated by electrophoretic mobility shift assays and DNase I footprinting. This ArgR recognition site contained a sequence, 5′-TGACGC-3′, which deviated in only 1 base from the common sequence motif 5′-TGTCGC-3′ found in other ArgR binding sites of P. aeruginosa. Furthermore, an alignment of all known ArgR binding sites confirmed that they consist of two directly repeated half-sites. In the absence of ANR, arginine did not induce the arc operon, suggesting that ArgR alone does not activate the arcD promoter. According to a model proposed, ArgR makes physical contact with ANR and thereby facilitates initiation of arc transcription.

The arginine deiminase (ADI) pathway catabolizes l-arginine to l-ornithine, with concomitant formation of ATP from ADP; three enzymes are involved: ADI, catabolic ornithine carbamoyltransferase, and carbamate kinase (17). The ADI pathway can provide Pseudomonas aeruginosa (17, 27) and other bacteria (33) with energy under anaerobic conditions in the absence of terminal electron acceptors such as oxygen or nitrate. In P. aeruginosa, the structural genes (arcABC) for the three enzymes are preceded by a gene (arcD) encoding an arginine-ornithine antiporter. The four genes are organized as an operon whose nucleotide sequence has been determined (2, 14, 15). Consistent with its function, the ADI pathway is strongly induced by oxygen limitation (17). The transcriptional regulator ANR (for anaerobic regulation of arginine catabolism and nitrate reduction), a homologue of the FNR protein of Escherichia coli, mediates this induction by acting at the −40 region of the arcD promoter (4, 5, 29, 32). Mutants defective in the arcDABC operon (27) or in ANR (4) cannot grow anaerobically with arginine as the only energy source.

In addition to anaerobic control by ANR, exogenous arginine can also induce the ADI pathway (1, 15, 17). Recent work (21, 22) has characterized a regulatory protein, ArgR, that represses the arginine biosynthesis carAB and argF genes encoding carbamoylphosphate synthetase and anabolic ornithine carbamoyltransferase, respectively. ArgR also mediates induction of the aruCFGDBE operon, which encodes enzymes of the arginine succinyltransferase pathway (8, 22). This pathway is considered to be the major route for arginine catabolism in P. aeruginosa under aerobic conditions (6, 11) and provides the cell with carbon, nitrogen, and energy. Furthermore, ArgR was shown to positively control the aotJQMOP-argR operon, which encodes a system for the uptake of arginine and ornithine as well as ArgR itself (19).

The mechanism by which arginine induces the ADI pathway in P. aeruginosa has not been studied previously. We have considered two hypotheses. When terminal electron acceptors are absent, arginine—being the only energy source—stimulates the synthesis of macromolecules and cell growth. Hence, arginine could have a general stimulatory effect on enzyme induction, including that of the ADI pathway. Alternatively or additionally, arginine might specifically enhance the expression of the arc operon. Our present data support the second hypothesis and establish that the arginine effect is essentially mediated by ArgR, which binds to a conserved sequence located upstream of the ANR binding site in the arcD promoter.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

P. aeruginosa strains used include PAO1 (wild type), PAO501 (argR::Gm^r) (21), PAO6251 (ΔarcDABC) (5), and PAO6261 [anrΔ(PvuII)] (31). The plasmids used are listed in Table 1. P. aeruginosa was grown in minimal medium OS (20) containing 25 mM l-glutamate and 25 mM l-arginine, but no (NH₄)₂SO₄. This medium was supplemented with carbenicillin at 100 μg/ml to maintain recombinant plasmids and, when PAO501 was the host strain, with gentamicin at 10 μg/ml. Oxygen limitation was achieved in tightly closed bottles as previously described (4).

TABLE 1.

Characteristics of the plasmids used in this study

Plasmid	Characteristics	Source or reference
pAC100	Ap, arc regulatory region amplified by PCR from pME190, cloned in the HindIII-BamHI sites of pUC19	This study (Fig. 1)
pAC101	Ap, arc regulatory region amplified from pME3731, with a mutated ArgR binding sequence, cloned in the HindIII-BamHI sites of pUC19	This study (Fig. 1)
pAC102	Same as pAC100, except for a mutated ArgR binding sequence	This study
pME190	Ap Cb, IncQ replicon carrying the arcDABC operon	15
pME336	Ap Cb, IncQ replicon carrying an arcD′-′lacZ fusion	5
pME3731	Ap Cb, pME190 derivative containing a 16-bp substitution (with an EcoRI site) in the arc promoter	5
pME3731-1	Same as pME3731, except for a mutated ArgR binding sequence from pAC101	This study
pME6306	Same as pME190, except for a 413-bp SacII deletion in arcC	This study
pUC19	Ap, ColE1 replicon	30

Plasmid constructions.

The regulatory region of the arcDABC operon carried by pME190 (15) was subcloned after PCR amplification employing two oligonucleotide primers designed to generate the inborn HindIII site upstream of the arc promoter (5) and a novel BamHI site at the 5′ end of the arcD gene (Fig. 1). The amplified fragment was digested with HindIII and BamHI and ligated into pUC19 cleaved with the same enzymes. The resulting plasmid was designated as pAC100. The recombinant PCR procedure described by Higuchi (7) was employed to modify the ArgR binding sequence in the arc regulatory region carried by pME190 and its derivative, pME3731. The amplified PCR fragments were cloned into HindIII and BamHI sites of pUC19, and the resulting plasmids were designated pAC102 and pAC101, respectively. Modification of the 5′-GCGTCA-3′ sequence (positions −79 to −74) (Fig. 1) to 5′-ATACTG-3′ in each of these two plasmids was confirmed by nucleotide sequencing. The modified sequence of pAC101 was used to replace the ArgR binding sequence in pME3731 by digestion of the plasmid with HindIII and EcoRI, treatment with alkaline phosphatase, and ligation with the purified HindIII-EcoRI fragment of pAC101; this produced pME3731-1. Plasmid pME6306 was constructed by deletion of the internal 413-bp SacII fragment in the arcC gene. The mutated arcC gene was then used to replace the arcC⁺ gene in pME190.

FIG. 1.

(a) Schematic drawing of the genetic organization of the arc operon in the recombinant plasmids used. pME190 carries the entire wild-type operon, and pME6306 carries a large deletion in arcC, as described in Table 1. (b) Nucleotide sequence of the arc regulatory region in pME190 and its derivatives. The operator sites for ArgR and ANR are indicated by rectangular and oval boxes, respectively. The HindIII site upstream of the arc promoter occurs naturally (5). The BamHI site at the 3′ end was introduced by PCR to construct pAC100 from pME190. Conserved nucleotides are shown in boldface.

For plasmid transformation into E. coli DH5α (Bethesda Research Laboratory) and P. aeruginosa strains, the method described by Chung et al. (3) for one-step preparation of competent cells was followed.

Enzyme assays.

ADI was assayed by measurements of citrulline production as described previously (15). Specific activities are expressed as units (micromoles of citrulline formed per hour) per milligram of protein. β-Galactosidase activity was determined by the Miller method (18).

DNA footprinting and gel retardation experiments.

A DNA fragment carrying the regulatory region of the arc operon was obtained from pAC100 by digestion with HindIII and BamHI endonucleases and was radioactively labeled with DNA polymerase Klenow fragment, with either [α-³²P]dATP at the HindIII site for the bottom strand or with [α-³²P]dGTP at the BamHI site for the top strand.

DNase I footprinting was carried out with homogeneous ArgR protein as previously described (22). The reaction mixture (200 μl) contained 2.5 × 10⁻¹⁰ M operator DNA, 2.5 to 20 nM ArgR protein, 50 mM Tris-HCl (pH 7.5), 50 mM KCl, 10 mM MgCl₂, 1 mM dithiothreitol, 1 μg of sheared salmon sperm DNA, and 10 μg of bovine serum albumin. After incubation for 30 min at 25°C, pancreatic DNase I (0.2 μg; Boehringer) was added. The digestion was allowed to proceed for 2 min and then terminated by the addition of 20 μl of 3 M sodium acetate, 10 μg of yeast tRNA, and 600 μl of ethanol. After precipitation with ethanol at −70°C, the pellet was dissolved in 20 μl of formamide-dye mixture, and the reaction products were analyzed on a 6% denaturing polyacrylamide sequencing gel against a G sequencing ladder (16).

For gel retardation experiments, the radioactively labeled DNA probe (5 × 10⁻¹² M) was allowed to interact with different concentrations of ArgR in 20 μl of 20 mM Tris-HCl (pH 7.6), 50 mM KCl, 1 mM EDTA, 5% (vol/vol) glycerol, and 50 μg of bovine serum albumin per ml. The reaction mixtures were allowed to equilibrate for 20 min at 25°C, the reaction was terminated by the addition of an excess of cold DNA probe (10⁻¹⁰ M), and then the mixtures were applied to a 5% polyacrylamide gel while the gel was running (13). The apparent equilibrium binding constant, defined as the protein concentration required for half-maximal binding, was determined from a plot of the percentage of DNA bound versus the protein concentration as previously described (13).

RESULTS

Induction by arginine can occur in the absence of ATP formation from arginine.

Strain PAO6251, in which the entire arcDABC operon is deleted, cannot generate ATP from arginine, unless this mutant is complemented with a plasmid carrying the intact arc operon (5), such as the IncQ recombinant plasmid pME190 (Fig. 1). The effect of exogenous arginine on induction of the arc operon was examined by measurements of ADI, the arcA product, in strain PAO6251 carrying various plasmids and grown under oxygen limitation and in the absence or presence of arginine. The results (Table 2) show that exogenous arginine induces ADI twofold in cells harboring pME190, which carries the intact arc operon. A similar level of induction is observed in cells harboring pME6306, a derivative of pME190 containing a large deletion in the arcC gene (Fig. 1). This deletion blocks the function of the ATP-regenerating enzyme, carbamate kinase. Furthermore, a translational arcD′-′lacZ fusion on pME336 (5) was induced approximately threefold by arginine. These results indicate that induction by extracellular arginine is not merely due to an extra supply of ATP produced from arginine. Therefore, we tested the hypothesis that arginine could directly affect the expression of the arc operon.

TABLE 2.

Arginine-dependent induction of the arc operon in PAO6251 (ΔarcDABC)

Plasmid (genotype)	Supplementa	Sp actb
		ADI (U/mg)	β-Galactosidase (Miller units)
pME190 (arcDABC+)	Glutamate	177 ± 47
	Glutamate + arginine	358 ± 51
pME6306 (arcDAB+ ΔarcC)	Glutamate	160 ± 40
	Glutamate + arginine	332 ± 73
pME336 (arcD′-′lacZ)	Glutamate		10,600 ± 1,300
	Glutamate + arginine		34,000 ± 7,000

^aCells were grown at 37°C under oxygen-limiting conditions in 25 mM glutamate minimal medium with or without 25 mM arginine to a final density of about 5 × 10⁸ cells per ml. 

^bSpecific activities are mean values ± standard deviations from six experiments. 

ArgR protein and an ArgR binding motif in the arc promoter are required for arginine induction.

Centered at −73.5 nucleotides from the transcription start (5), the arc promoter contains a sequence motif (Fig. 1) resembling the ArgR binding sites in the aotJ, aruC, carA, and argF promoters (Fig. 2). The ArgR binding site consists of two half-sites in a direct repeat arrangement, with the consensus sequence of 5′-TGTCGCN₈AA-3′ (22). The most-conserved nucleotides, 5′-GCGTCA-3′, which are located on the complementary strand at positions −79 to −74 in the arc promoter, were mutated to 5′-ATACTG-3′ (Fig. 1). To introduce this mutation into the arc promoter, we used a plasmid (pME3731) containing an artificial EcoRI site in the −30 region of the arc promoter (Fig. 1). We have shown before that the introduced EcoRI site does not interfere with anaerobic, ANR-dependent induction of the arc operon (5). The plasmid containing the mutated ArgR binding motif was designated pME3731-1.

FIG. 2.

Sequence alignment of ArgR binding sites. The sequences were obtained from the results of DNase I footprintings and aligned by using the Clustal W program (25). The first and second halves of the binding sites (I and II) are depicted by arrows. The consensus sequence was deduced from the second half-sites, which are more conserved (19, 21, 22). Nucleotides identical to those of the consensus site are shaded. ς⁷⁰-RNP, RNA polymerase holoenzyme; −35 and −10, positions of the promoter recognized by ς⁷⁰. The sequences shown for arcD and argF are for the complementary strands and are in the opposite orientation from the direction of transcription.

ADI activity in the wild-type strain PAO1 and its argR and anr derivatives was measured. The results (Table 3) clearly establish that induction by exogenous arginine during oxygen limitation is abolished in either derivative. Similarly, in PAO1 harboring pME3731, ADI was inducible in response to arginine, but to a lower extent, likely reflecting the titration of one or both activators under multicopy conditions (The IncQ vector used has 20 to 40 copies in P. aeruginosa) (12). In contrast, no arginine induction was detected in strain PAO1/pME3731-1 (Table 3), in which the ArgR binding site was modified. In the argR mutant PAO501, both pME3731 and pME3731-1 were not induced by arginine (Table 3). In the anr mutant PAO6261 carrying pME3731 or pME3731-1, ADI expression was low and was not influenced by extracellular arginine (Table 3). On the one hand, these data confirm that ANR is the major transcriptional regulator of the arc operon (5, 32). On the other hand, the data suggest that the presence of ANR is necessary for ArgR to be an activator of the arc operon.

TABLE 3.

ArgR- and ANR-dependent induction of the arc operon

Strain	Plasmid	ADI sp act (U/mg)a
		Without l-arginine	With 25 mM l-arginine
PAO1	None	11 ± 2	43 ± 4
	pME3731	343 ± 43	548 ± 99
	pME3731-1	357 ± 82	351 ± 42
PAO501 (argR)	None	9 ± 2	12 ± 2
	pME3731	287 ± 47	365 ± 63
	pME3731-1	263 ± 27	318 ± 21
PAO6261 (anr)	None	1 ± 1	1 ± 1
	pME3731	33 ± 3	31 ± 4
	pME3731-1	66 ± 17	67 ± 11

^aCells were grown at 37°C under oxygen-limiting conditions in glutamate minimal medium with or without an arginine supplement to a final density of about 5 × 10⁸ cells per ml. Specific activities are mean values ± standard deviations from six experiments. 

ArgR binds to the −70 region of the arc promoter in vitro.

In order to test whether the ArgR binding motif in the arc promoter is a functional ArgR binding site, we carried out gel retardation experiments with homogeneous ArgR protein and a 232-bp HindIII-BamHI fragment carrying the control region of the arc operon (Fig. 1). The ArgR protein was found to bind specifically to this arcD regulatory region (Fig. 3), with an apparent dissociation constant of 4 × 10⁻¹¹ M as determined from a plot of the percentage of bound DNA against the concentration of ArgR (data not shown). When the arc control region containing the mutated ArgR binding site (see above) was used as a control, no ArgR-dependent band shift was observed (Fig. 3).

FIG. 3.

Gel retardation experiments. The radioactive ³²P-labeled arc operator DNA fragments were incubated with various ArgR concentrations in the absence of l-arginine. Lanes 1 to 5 contained the wild-type promoter fragment (from pAC100) and ArgR at 0, 25, 50, 100, and 200 pM, respectively. Lanes 6 to 10 contained the promoter fragment mutated in the ArgR binding site (from pAC102) and ArgR at 0, 25, 50, 100, and 200 pM, respectively.

DNase I footprinting analysis was used to define the extent of the ArgR binding site. Binding of ArgR protected a 45-bp region against nuclease digestion on both strands (Fig. 4). This region was found at the predicted site immediately upstream of the ANR box (Fig. 1).

FIG. 4.

DNase I footprintings. The DNA fragment was labeled with [³²P]dATP at the 3′ end of the HindIII site on the bottom strand of the arc operator (Fig. 1). Lanes: 1 and 2, DNase I digestion in the absence of ArgR and in the presence of 20 nM ArgR, respectively; 3 and 4, ′G′ Maxam-Gilbert sequencing ladder (16); 5 and 6, analogous to lanes 2 and 1, but the DNA fragment was labeled with [³²P]dGTP at the 3′ end of the BamHI site on the top strand. The nucleotide sequence of the ArgR-protected region is shown below the sequencing ladder, with italicized characters corresponding to the most conserved bases of the ArgR binding sites, TGTCGCN₈AA.

DISCUSSION

The major signal governing the expression of the arcDABC operon in P. aeruginosa is oxygen limitation (17). Such conditions allow the ANR protein to activate transcription (4, 32) by binding to the −40 region of the arc promoter (5, 29). However, it has been clear from the early studies of Mercenier et al. (17) and Abdelal et al. (1) that exogenous arginine can also act as an inducing signal, although the induction factor is significantly smaller than that brought about by lowering the oxygen tension.

Several lines of evidence reported here show that arginine induction of the arc operon is mediated by ArgR. (i) Induction by exogenous arginine is abolished in a derivative of PAO1 in which argR was inactivated by gene replacement, by using a gentamicin cassette. (ii) Gel retardation experiments (Fig. 3) show that ArgR binds specifically to a DNA fragment carrying the control region of the arc operon and that this binding is abolished when conserved bases in the ArgR binding site are modified. (iii) When the arc promoter contains the modified ArgR binding site, ADI is no longer induced by exogenous arginine (Table 3).

DNase I footprinting experiments (Fig. 4) show that ArgR protects a segment of 45 bases located upstream of the arc promoter. The identified binding site was compared with the ArgR binding sites for the argF, car, aru, and aot promoters (Fig. 2). This alignment shows that a conserved sequence (TGACGC) in the ArgR binding site of the arc operon differs by only 1 base from the consensus sequence, TGTCGC, which was shown by premethylation and depurination experiments (21) to be important for ArgR binding. Apparently, the nonconserved base (A) found in the arc promoter does not prevent ArgR from binding.

In the arc promoter, the ArgR and ANR binding sites are adjacent, suggesting that the ArgR and ANR proteins may physically interact with each other during activation of transcription. Interestingly, the ArgR binding sequence resembling the consensus sequence is on the bottom strand of the arc regulatory region. This arrangement differs from that found for other arginine-inducible operons (aot and aru) (22), but it occurs in the argF promoter, which is repressible by ArgR (Fig. 2). It will be interesting to determine if this spatial arrangement is a prerequisite for ArgR and ANR interactions. In the model presented in Fig. 5, ANR is assumed to interact with RNA polymerase in exactly the same way as FNR does with RNA polymerase of E. coli (23). Considering the high degree of sequence conservation between FNR and ANR (28, 29, 32), this assumption appears justified. In oxygen-limited cells, ANR can activate transcription without arginine and ArgR (15, 17, 29) (Table 2). Arginine and ArgR boost the ANR-dependent induction (Table 3). In well-aerated cells (17) or in an ANR-negative mutant (Table 3), arginine does not induce the arc operon, suggesting that ArgR alone cannot activate arc transcription. This situation is reminiscent of nitrate respiration in E. coli in which the anaerobic activator FNR is assisted by the nitrate- and nitrite-responsive response regulators NarL and NarP at the nir and nrf promoters (24, 26).

FIG. 5.

Model depicting the interaction between ArgR and ANR at the arc promoter of P. aeruginosa. Given the overlapping specificity of FNR and ANR and the functional conservation of the two regulators (28, 29, 32), this model is based on the structural data elaborated for FNR and RNA polymerase of E. coli (23). The ς⁷⁰, β, and β′ subunits of the RNA polymerase were labeled accordingly. αCTD and αNTD represent the carboxy-terminal domain and the amino-terminal domain of the α subunits, respectively. The possible contact between ANR and the ς⁷⁰ subunit is represented by a solid square.

Recently, Maghnouj et al. (15a) reported the cloning and characterization of the arc operon from Bacillus licheniformis. While anaerobiosis is a prerequisite for arginine induction of the arc operon of P. aeruginosa, as outlined above, both anaerobiosis and arginine, as well as a functional ArgR, were required for induction of the arc operon in B. licheniformis (15a). ArgR of B. licheniformis was shown to bind a sequence homologous to the Arg boxes that characterize operator sites for the arginine repressors from enteric bacteria and B. subtilis. A nucleotide sequence feature similar to the E. coli Crp and B. subtilis Fnr consensus sequences was found between the ArgR binding site and the promoter, suggesting that a second regulatory protein from the Crp or Fnr family might function as the anaerobic activator of this system. It should be noted that ArgR of B. licheniformis is a hexamer of a 17-kDa polypeptide, as is the case for the arginine regulatory proteins from enteric bacteria and B. subtilis; these proteins also contain highly conserved domains (15a). In contrast, the ArgR of P. aeruginosa is a dimer of 37 kDa that belongs to the AraC-XylS family and does not exhibit sequence similarity to the regulatory proteins of enteric bacteria or B. subtilis (21, 22).

In vitro, the presence of arginine did not enhance binding of ArgR to the arc promoter from P. aeruginosa (data not shown). Previous work (19, 21, 22) has shown that arginine had no effect on the affinity of ArgR for the car, argF, aru, and aot promoters. Two interpretations can be offered. Arginine might facilitate the interaction between ArgR and the ANR-RNA polymerase complex, without influencing the binding of ArgR to its operator sequence. Alternatively, intracellular arginine might not be the relevant signal. Instead, extracellular arginine might be sensed by a sensor protein, which, via a response regulator, would activate the expression of the aot-argR operon. This hypothesis is supported by the recent identification of the ArgSU two-component system in P. aeruginosa. Mutational inactivation of the sensor ArgS results in an arginine-utilization-negative phenotype, presumably because of loss of ArgR synthesis (9), and in an argU-negative background, an aot-lacZ fusion is not expressed (10).