Pseudomonas aeruginosa AlgR Controls Cyanide Production in an AlgZ-Dependent Manner

Abstract

Pseudomonas aeruginosa is an opportunistic pathogen that causes chronic infections in individuals suffering from the genetic disorder cystic fibrosis. In P. aeruginosa, the transcriptional regulator AlgR controls a variety of virulence factors, including alginate production, twitching motility, biofilm formation, quorum sensing, and hydrogen cyanide (HCN) production. In this study, the regulation of HCN production was examined. Strains lacking AlgR or the putative AlgR sensor AlgZ produced significantly less HCN than did a nonmucoid isogenic parent. In contrast, algR and algZ mutants showed increased HCN production in an alginate-producing (mucoid) background. HCN production was optimal in a 5% O₂ environment. In addition, cyanide production was elevated in bacteria grown on an agar surface compared to bacteria grown in planktonic culture. A conserved AlgR phosphorylation site (aspartate at amino acid position 54), which is required for surface-dependent twitching motility but not alginate production, was found to be critical for cyanide production. Nuclease protection mapping of the hcnA promoter identified a new transcriptional start site required for HCN production. A subset of clinical isolates that lack this start site produced small amounts of cyanide. Taken together, these data show that the P. aeruginosa hcnA promoter contains three transcriptional start sites and that HCN production is regulated by AlgZ and AlgR and is maximal under microaerobic conditions when the organism is surface attached.

Pseudomonas aeruginosa is a gram-negative opportunistic pathogen that can cause acute infections in patients suffering from cancer (6), human immunodeficiency virus/AIDS (27, 51), and severe burns (9). P. aeruginosa can also cause chronic infections in patients affected with the genetic disorder cystic fibrosis (CF) (8, 31). Mutation of the gene that encodes the CF transmembrane conductance regulator, a chloride ion channel found in epithelial cells, results in a dehydrated layer of mucus in the airways of the lungs. The impaired mucociliary clearance entraps bacteria and allows for the establishment of a chronic infection (65, 85). Pulmonary colonization by P. aeruginosa is the leading cause of morbidity and mortality of CF patients (8).

During chronic CF infection, P. aeruginosa strains acquire mutations that cause them to undergo a phenotypic switch from a nonmucoid to a mucoid phenotype. The mucoid phenotype results from the production of copious amounts of the capsular exopolysaccharide alginate (23). The presence of a mucoid strain in the CF lung is associated with a worsening clinical prognosis (22, 24). Production of alginate aids the organism in the evasion of phagocytosis and in resistance to killing by neutrophils and macrophages (11, 64, 74), quenching of oxygen intermediates (46, 76), and antibiotic resistance (1). Alginate biosynthesis is regulated by multiple factors (34, 69) including the transcriptional regulator AlgR, a response regulator belonging to the bacterial two-component signaling system superfamily (21, 35). AlgR is required for virulence in an acute septicemia model (48) and regulates a variety of other virulence factors, including twitching motility (80, 81), hydrogen cyanide (HCN) production (13), and biofilm formation and quorum sensing (58).

The effects of cyanide on various cell types have been extensively tested, and depending on cell type and level of oxidative stress, cyanide causes either apoptosis or necrosis (66, 67). P. aeruginosa was one of the first bacteria shown to undergo cyanogenesis (18). Growing evidence suggests that HCN production may be an important virulence factor. The presence of cyanide in burn wounds was directly associated with P. aeruginosa infections (33). More recently, the ability of P. aeruginosa to paralyze and kill Caenorhabditis elegans in an in vitro model was shown to be dependent on hcnC (30). HCN production has also proven to be clinically relevant in the CF lung as (i) P. aeruginosa CF isolates produced higher levels of HCN than did laboratory strains (13), (ii) cyanide has been detected in the sputum of P. aeruginosa-infected CF and bronchiectasis patients (12, 26, 40, 71), and (iii) detection of HCN in sputum and bronchoalveolar lavage fluid directly correlated with a worsening prognosis (12). The reported HCN concentrations in the sputum of infected patients ranged from 1.6 to 3.25 times the lethal blood concentration (71).

HCN is a secondary metabolite of P. aeruginosa produced from glycine in a poorly understood oxidative reaction catalyzed by HCN synthase (15, 82). HCN is not produced by P. aeruginosa under anaerobic conditions (15) but is expressed under low oxygen tensions (14). Optimal production requires the anaerobic regulator ANR and the quorum sensing regulators LasR and RhlR, which bind the hcnA promoter at high cell densities during late exponential growth (62). There are two reported transcriptional start sites for hcnA in P. aeruginosa (62) and one in Pseudomonas fluorescens (45). P. aeruginosa T1 is regulated by the rhl and las quorum sensing systems, while T2 relies on a synergistic effect of LasR, RhlR, and ANR. GacA and RsmA function as posttranscriptional regulators of HCN production through juxtaposition of the hcnA ribosomal binding site and indirectly by positively and negatively regulating amounts of N-acylhomoserine, respectively (43, 61, 70). Additionally, AlgR directly regulates HCN production in P. aeruginosa through control of T1 and T2 of the hcnA promoter. Moreover, AlgR has recently been shown to repress HCN production in nonmucoid strains and to activate production in mucoid strains, demonstrating AlgR's ability to switch from a repressor to an activator (13).

The purpose of this study was to examine the role of AlgZ, the putative sensor for AlgR in P. aeruginosa HCN production. Two independent research groups identified the open reading frame (PA5262) that encodes AlgZ almost simultaneously (80, 89). The first group identified this gene while screening a transposon library for mutants that lost twitching motility and termed the gene fimS (80). The second group identified this same open reading frame while examining alginate production and called it AlgZ (89). The algZ or fimS gene is located directly upstream of algR and shares homology with other two-component sensors. PA3385 has also been referred to as algZ in the literature (2-4, 86) but has since been renamed amrZ (3). In addition to identifying a previously unreported hcnA transcriptional start site, the present study expands the role of AlgZ and AlgR by providing evidence that this putative sensor/regulator pair control HCN production through a surface attachment-specific mechanism that requires the conserved AlgR aspartate at position 54. These data suggest that AlgZ functions as a kinase in regulating HCN production in P. aeruginosa.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

The strains and plasmids used in this study are listed in Table 1. Escherichia coli was grown on Luria-Bertani (LB) agar (Difco). P. aeruginosa was grown on Pseudomonas isolation agar (PIA; Difco). Both were incubated at 37°C overnight. LB agar was supplemented with kanamycin (25 μg/ml) or ampicillin (50 μg/ml), and PIA was supplemented with tetracycline (300 μg/ml), streptomycin (10 μg/ml), or gentamicin (10 μg/ml) for selection, plasmid maintenance, or genetic manipulation.

TABLE 1.

Strains used in this study

Strain or plasmid	Description	Mucoid phenotypea	Reference or source
P. aeruginosa strains
PAO1	Wild type	NM	38
PAZ-1	ΔalgZ	NM	This study
PAO6886	algZ::Tcr	NM	89
PSL317	ΔalgR	NM	48
PSL317Z	algZ::Tcr ΔalgR	NM	This study
PAO568	mucA2	M	29
PDO300	mucA22	M	54
PAZ568	mucA2 ΔalgZ	M	This study
PAO6884	mucA2 algZ::Tcr	M	89
PAR568	mucA2 ΔalgR	NM	13
PAO6885	mucA2 algZ::GmralgR::Tcr	NM	89
TUMC92	Clinical isolate	NM	13
TUMC92Z	ΔalgZ	NM	This study
TUMC92R	ΔalgR	NM	13
FRD-1	mucA	M	32
FRD-1Z	mucA ΔalgZ	M	This study
FRD-1R	mucA algR::Smr	NM	50
WFPAPAO1	Wild type	NM	50
WFPA8	algR(D54N)	NM	50
WFPA13	algR(D85N)	NM	50
WFPA16	algR(D54N D85N)	NM	50
WFPA8A	mucA algR(D54N)	NM	This study
WFPA13A	mucA algR(D85N)	M	This study
WFPA16A	mucA algR(D54N D85N)	M	This study
WFPA8Z	mucA algR(D54N)	NM	This study
WFPA13Z	mucA algR(D85N)	M	This study
WFPA16Z	mucA algR(D54N D85N)	M	This study
WFPA8AZ	mucA ΔalgZ algR(D54N)	M	This study
WFPA13AZ	mucA ΔalgZ algR(D85N)	M	This study
WFPA16AZ	mucA ΔalgZ algR(D54N D85N)	M	This study
PAO1 vfr	Δvfr	NM	This study
PAK	Hyperpiliated	N	39
PAK mucA	mucA22	M	87
PAK algR	algR::Gmr	NM	87
PAK mucA algR	mucA22 algR::Gmr	NM	87
PA14	Wild type	NM	68
PA103	Wild type	NM	47
19660	Clinical isolate	NM	83
X13273	Clinical isolate	NM	83
S54485	Clinical isolate	NM	83
CF27	Clinical isolate	M	83
PAK pscC	ΔpscC	NM	84
PAK HCN KI	ΔpscC; PAO1 hcnA promoter	NM	This study
Plasmids
pCR2.1	TA cloning vector; Apr Kanr		Invitrogen
pCVD442	Gene replacement vector; TcrsacB+		26
pZKO442	algZ in-frame deletion vector; Apr		This study
pUCP18	Expression vector; Tcr		78
pUCP18Z	algZ expression vector for complementation		This study
pRK2013	Helper plasmid for conjugation, tra functions		28
pEXGmGW	Gateway converted pEX18Gm		83
phcnAKI	pEXGmGW with hcnA PAO1 promoter		This study

^aM, mucoid; NM, nonmucoid.

DNA manipulations.

The oligonucleotides used in this study are listed in Table 2. The plasmid pZKO442, a derivative of pCVD442 (25), was used to create a 907-bp in-frame deletion of the algZ gene in multiple strains. The deletion plasmid was constructed by a two-step crossover PCR (37) with oligonucleotide primers algZkoXbaF, algZkoR′2, algZkoF′2, and algZkoSstR using the PAO1 chromosome as the template DNA. Oligonucleotides algZkoXbaF and algZkoSstR were used to engineer XbaI and SstI restriction enzyme sites to the 5′ and 3′ ends of the final PCR product, respectively. The resulting 1,743-bp fragment was cloned into pCVD442 using the restriction enzymes XbaI and SstI. The plasmid pUCP18Z was used to complement the algZ deletion strains. To construct this plasmid, a 2,326-bp fragment was amplified by PCR with oligonucleotides algZko1 and algZko4 to obtain the algZ open reading frame and promoter region from the PAO1 chromosome. This fragment was cloned into pCR2.1 (Invitrogen) and subcloned into the XbaI and KpnI restriction sites of pUCP18.

TABLE 2.

Oligonucleotides used in this study

Name	Oligonucleotide sequence	Gene	Locationa
algZkoXbaF′	CCCTCTAGACCCAGCAACTGGCAGGTGATCTCG	algZ	−889 to −866 (+)
algZkoR′2	ACGCCTGTCGCGTGCCTTTCGACGCTTGAATCGGATAGGCATCGACAGAG	algZ	−8 to +17 (−)
algZkoF′2	CGTCGAAAGGCACGCGACAGGCGT	algZ	+925 to +948 (+)
algZkoSstR′	CCCGAGCTCGAGGCCTTTCAGGTAGAGCTGGA	algZ	+1736 to +1758 (−)
algZKO1F	TGTCTTCCTGGTTGTCCTTGTTGTA	algZ	−1212 to −1188 (+)
S1algRhindIII	CCCAAGCTTCCAGAGGTTCGTCATCGACAATCAG	algZ	+1090 to +1114 (−)
algZKO4R	CGAGTGGCGATGCGGATT	algZ	+1950 to +1967 (−)
hcnAprimetext	GTGTTGACGACGTTCAAGAAGGTGCAT	hcnA	−495 to +24 (−)
hcnAS1BglI	GCAAAGCCGGCTCACCCGTGGAGG	hcnA	−3166 to −3137
PhcnA KI 5′	TACAAAAAAGCAGGCTCCGCTTCCGACCGCCCCAGGCGGCTGA	hcnA	−1622 to −1600 (+)
PhcnA KI 3′	TACAAGAAAGCTGGGTCCGCAGTTTTTCCAGCATCCGGGTCTA	hcnA	+533 to +508 (−)
ArgHF	ATATATGAGCTCGGACCTGTCCGACCTGTTCC	argH	−482 to −426 (+)
HemCR	ATATATGAGCTCGGCTGGCGTAGGTGTTCGAG	hemC	+1197 to +1177 (−)

^aLocation numbering is relative to the translational start site of gene. (+), coding strand; (−), antisense strand.

DNA manipulations.

In-frame deletions of algZ were created in laboratory and clinical P. aeruginosa strains by introducing the algZ deletion plasmid (pZKO442 [ΔalgZ]) through triparental conjugation (42) and selection of carbenicillin-sensitive and sucrose-resistant double recombinants using the helper plasmid pRK2013 (28). Deletion of algZ was confirmed by PCR and Southern blotting with a digoxigenin-labeled probe (Roche) using oligonucleotides algZko1 and algZko4 (data not shown). Western blot analysis with a monoclonal anti-AlgR antibody, 3H9, confirmed that there were no polar effects on AlgR expression in algZ deletion strains. In-frame deletion of the algR gene was confirmed by PCR with oligonucleotides ArgHF and HemCR and Western blot analysis with anti-AlgR antibody (20). The PAO1 ΔalgR algZ::Tc^r mutant was created by electroporating fragmented chromosomal DNA from PAO6886 (algZ::Tc^r) into PSL317 and selection of tetracycline-resistant recombinants, as described by Choi et al. (17). Deletion of algR and insertional inactivation of algZ was confirmed by Southern blot analysis using digoxigenin-labeled probes (Roche) created with primers ArgHF and HemCR or algZko1 and algZko4, respectively (data not shown).

The entire hcnA promoter region from PAO1 was amplified from genomic DNA with primers PhcnA KI 5′ and PhcnA KI 3′ (Table 2). The resulting PCR fragment (700 bp) was engineered with AttB1 and AttB2 sequences at the 5′ and 3′ ends, respectively, and cloned into pEXGmGW (84) via cloning to create the hcnA knock-in (KI) vector phcnAKI. This 700-bp PAO1 hcnA promoter fragment was introduced into an attenuated PAK strain (PAKΔpscC), which lacks a functioning type III secretion system, as previously described (84). Positive hcnA KI mutants were identified by PCR and confirmed by sequencing.

S1 nuclease protection assays.

RNA for S1 nuclease protection assays was isolated from PAO1, PSL317 (ΔalgR), and PAZ (ΔalgZ) grown in LB to stationary phase using CsCl as previously described (52). An S1 nuclease protection assay was performed as previously described using 100 μg of RNA with the following modifications (51). The 519-bp region of the hcnA promoter ranging from −495 to +24 (numbering relative to the translational start site) was cloned into M13mp18. Single-stranded phage DNA was isolated and used as a template for the uniformly labeled ([α-³²P]dCTP; NEN DuPont) single-stranded DNA probe. Two probes were constructed using either of the oligonucleotides hcnAprimetext and hcnAS1BglI. Probes constructed with hcnAprimetext were digested with BglI, and probes constructed with hcnAS1BglI were cut with BglI. Digested probes were purified on a 5% polyacrylamide gel. Total RNA (100 μg) was hybridized to each probe at 67°C for 1 h. Reactions were then digested with S1 nuclease for 30 min and then precipitated. S1 reaction mixtures were electrophoresed through an 8% acrylamide-urea gel adjacent to a sequencing ladder generated using the same oligonucleotide as the probe.

HCN production assay of P. aeruginosa strains.

HCN was measured using a modification of the protocols described by Zlosnik and Williams (91) and Gallagher and Manoil (30). Laboratory and clinical P. aeruginosa isolates were grown on PIA plates aerobically for 16 h at 37°C. The plates were then inverted and individually sealed with lids containing 1 ml of 4 M NaOH. The inverted plates were then placed inside a trigas incubator under microaerobic conditions (5% O₂, 5% CO₂, and 90% N₂) at 30°C for 4 h. After a 4-hour incubation at 30°C and after 20 h total, the NaOH/NaCN was collected and measured directly using a cyanide micro-ion-selective electrode (Lazar Research Laboratories) connected to an Accument Basic AB15 pH meter (Fisher Scientific) set to read mVs. Values recorded from known concentrations of a KCN standard ranging from 1 × 10⁻⁵ to 1 M were used to generate a line of regression on a semilog graph. Unknown values were calculated using the slope and y intercept of this line. For the time course, cells were incubated with the NaOH for 4 hours prior to the collection time indicated. Bacterial cells were collected from the plates, total protein was determined by the Bradford assay (Bio-Rad) as previously described (13), and HCN production was reported as μmol HCN/mg protein.

Pseudomonas isolation liquid medium (2% peptone, 0.14% MgCl₂, 1% K₂SO₄, 0.0025% Irgasan) broth cultures were grown in a sidearm flask with an attached enclosed chamber with shaking at 150 rpm under temperature and oxygen conditions identical to those of the plate-grown cultures. Four hours prior to the collection time point 1 ml of 4 M NaOH was added to the sidearm collection chambers to act as a cyanide trap. The micro-ion-selective electrode was used to measure the cyanide concentration in the NaOH/NaCN collected in the attached chamber and broth directly, and the concentrations were added together to obtain the total amount of cyanide produced in the broth culture. The cells were harvested by centrifugation, and total protein was determined as described above. Six biological replicates were performed for each strain. One-way analysis of variance (ANOVA) and the Tukey-Kramer multiple test were used for statistical analysis.

Optimization of oxygen concentration for P. aeruginosa HCN production.

PAO1 was grown on PIA plates aerobically for 16 h at 37°C. The plates were then inverted, individually sealed with lids containing 1 ml of 4 M NaOH, and incubated for 4 additional hours at 30°C in a trigas incubator under various oxygen conditions (1%, 3%, 5%, 7%, 10%, and 15% O₂). The NaOH/NaCN and total protein were collected and measured as described above.

RESULTS

P. aeruginosa HCN production is surface attachment specific.

HCN is produced by P. aeruginosa during the transition from logarithmic to stationary growth phase at low oxygen tensions and high cell densities due to regulation of the hcnA promoter by the synergistic effects of the anaerobic regulator ANR and the quorum sensing regulators LasR and RhlR (14, 16, 62). Cyanide production is also under the control of the transcriptional regulator AlgR (13). RhlA and RhlI have recently been shown to be controlled by AlgR through an attachment-specific mechanism (58). A recent report showed increased cyanide production from P. aeruginosa grown in a static glass bead biofilm system (72). In order to determine if HCN production is increased during surface growth, HCN production by PAO1 was measured following growth of cells in broth and on agar media. Planktonic cultures of P. aeruginosa PAO1 produced a maximal amount of 0.94 μmol HCN/mg protein during the transition from mid-log to stationary phase (8 h) consistent with previous findings (14, 16) (Fig. 1A). Cultures grown on agar plates produced various amounts of HCN over the time course with an increase at 20 h to 1.5 μmol HCN/mg protein (Fig. 1B). These results show that surface-grown P. aeruginosa in stationary phase produces more HCN than do planktonically grown cells.

FIG. 1.

PAO1 HCN production in liquid and plate media and under different oxygen tensions. (A) HCN production (solid squares and solid line) by PAO1 in liquid Pseudomonas isolation broth and bacterial CFU/ml (solid triangles and dotted line) over time. (B) HCN production (solid squares and solid line) by PAO1 grown on PIA plates and bacterial CFU/ml (solid triangles and dotted line) over time. (C) HCN production under various oxygen tensions. Statistical significance was determined by one-way ANOVA and the Tukey-Kramer test. ***, P < 0.001.

Optimal oxygen tension for P. aeruginosa HCN production.

HCN is produced at low oxygen tensions during planktonic growth (14, 16), but the optimal oxygen concentration for surface-attached cells is unknown. To determine the range of oxygen concentrations that permit HCN production, P. aeruginosa was grown at various oxygen levels and HCN production was measured. PAO1 produced the maximum amount of HCN in a 5% oxygen environment (Fig. 1C). Consistent with previous studies, HCN production was greatly reduced as oxygen tensions neared anaerobic or aerobic conditions (14, 16).

Identification of previously unidentified hcnA transcriptional start site.

AlgR directly regulates HCN production in P. aeruginosa by controlling expression of hcnA from the T1 and T2 transcriptional start sites (13). The effects of algZ deletion on the transcription of the hcnA promoter were examined under microaerobic growth conditions by an S1 nuclease protection assay. This assay showed that AlgZ, like AlgR, is required for repression of hcnA transcription in a nonmucoid background (Fig. 2A). Interestingly, the assay also revealed an additional transcriptional start site (designated T3) upstream of the two previously identified sites (13, 62). The intensity of this band of protection indicates that the amount of transcript produced from the T3 site is equal to or greater than that of the T1 and T2 transcripts. The T3 start site is a G residue located −221 bp from the translational initiation codon.

FIG. 2.

The hcnA promoter is controlled by AlgZ and AlgR. (A) S1 nuclease protection assay of the hcnA promoter. Previously mapped T1 and T2 transcriptional start sites (indicated by respective labels and arrows) of the nonmucoid strain PAO1 and ΔalgR and ΔalgZ derivatives. The open triangle with the adjacent black line denotes the beginning of the common deletion region found upstream of the hcnA promoter. (B) S1 nuclease protection assay identifying T3 transcriptional start site. The arrow indicates the band of protection indicating the +1 start site. An asterisk denotes the first nucleotide in the antisense probe. (C) Sequence of the hcnA promoter region. The shaded sequence denotes the exoY deletion region. Binding sites are underlined for AlgR (AlgR Binding Site), ANR (ANR Box), RhlR and LasR (lux α and lux β), and ribosome (RBS). +1, HcnA translational start site; boldface nucleotides, hcnA open reading frame.

A naturally occurring deletion upstream of hcnA results in low cyanide production.

Previously, Wolfgang et al. identified a region upstream of hcnA in PAO1 (termed the exoY deletion region) that is absent in several other laboratory and clinical strains (83). The hcnA T3 transcriptional start site and previously identified AlgR and Lux β binding sites are all within this deleted region (Fig. 2B) (13, 62). Measurement of HCN production on the laboratory strain PAK, which features this chromosomal deletion, showed low levels of cyanide production (∼0.3 μmol/mg protein) in comparison to PAO1 (∼1.4 μmol/mg protein) (Fig. 3A) when measured under microaerobic conditions. Mutations in PAK algR, mucA2, or both did not affect the amount of cyanide produced by these isogenic strains (Fig. 3A). Sequence analysis of the hcnA promoter region in strains PA14 (HCN production = 0.4 μmol/mg protein) and PA103 (HCN production = 0.1 μmol/μg protein) revealed that they did not contain the hcnA T3 promoter either (data not shown). Insertion of the PAO1 hcnA promoter onto the PAK chromosome by homologous recombination restored HCN production to PAO1 wild-type levels in the PAK background (Fig. 3A). Moreover, cyanide measurements of blood isolate X13273, urinary tract infection isolate S54485, eye isolate 19660, and clinical isolate CF27, all previously shown to contain the same deletion region (83), revealed that these strains also produce very little HCN (Fig. 3B).

FIG. 3.

Strains with the exoY chromosomal deletion region upstream of hcnA produced decreased amounts of cyanide. (A) HCN production from nonmucoid wild-type PAO1 and laboratory strains with the exoY deletion region (PA14, PA103, and PAK), the PAK algR deletion (PAK algR::Gm), the PAK mucA22 mutation, the PAK mucA22 algR deletion, the PAK pscC deletion (PAK ΔpscC), and the PAK pscC deletion with the PAO1 hcnA promoter (PAK HCN KI). (B) HCN produced from PAO1, CF isolate CF27, eye isolate 19660, urinary tract infection isolate S54485, and blood isolate X13273. Statistical significance was determined by one-way ANOVA and the Tukey-Kramer test. **, P < 0.01, and ***, P < 0.001, compared to the isogenic parent.

AlgZ activates and represses HCN production in P. aeruginosa.

Previous analysis of HCN production in P. aeruginosa strains indicated that mucoid strains produce approximately sevenfold more HCN than do their nonmucoid parental strains. In-frame algR deletions in nonmucoid strains produced ∼3-fold more HCN and ∼1.5-fold less HCN in a mucoid background, showing that AlgR switches from a repressor to an activator of the hcnA promoter depending upon mucoid status (13). Both AlgR and AlgZ are involved in alginate production (19, 89) and required for twitching motility (5, 80, 81). An S1 nuclease protection assay indicated that AlgZ also plays a role in HCN regulation (Fig. 2A). Cyanide production levels from in-frame algZ deletion and insertional algZ mutants in both nonmucoid and mucoid backgrounds were measured under microaerobic conditions on agar plates. Mutations in algR, algZ, or both resulted in a ∼3-fold increase in HCN production in nonmucoid laboratory strains (Fig. 4A) compared to the wild type. In mucoid laboratory strains, mutations in algR, algZ, or both resulted in an ∼3.5-fold decrease in HCN production (Fig. 4B). In order to determine if the results observed with laboratory strains and their isogenic mutants are consistent with those for clinical isolates, HCN production was measured in mucoid and nonmucoid CF isolates and their isogenic algR and algZ mutants. Consistent with the results for the laboratory strains, mutations in algR or algZ resulted in an ∼1.5-fold increase in HCN production in nonmucoid clinical isolates (Table 3). In mucoid clinical isolates, mutations in algR or algZ resulted in a decrease in HCN production (Table 3). Additionally, complementation of the algZ mutant with the wild-type allele restored HCN production to wild-type levels in all strains (Fig. 4A and B). These results provide evidence that AlgZ works through AlgR to regulate HCN production.

FIG. 4.

AlgZ and AlgR control HCN production. HCN production for nonmucoid (A) and mucoid (B) algZ and algR mutants was measured as for Fig. 3. (A) PAO1, PSL317 (ΔalgR), PAZ (ΔalgZ), PAO6886 (algZ::Tc), PAO6886 harboring pUCP18Z (algZ::Tc + algZ), and PSL317Z (ΔalgR algZ::Tc). (B) PAO568, PAR568 (ΔalgR mucA2), PAZ568 (ΔalgZ mucA2), PAO6884 (algZ::Tc mucA2), PAO6884 harboring plasmid pUCP18Z (algZ::Tc mucA2 + algZ), and PAO6885 (algR::Gm algZ::Tc mucA2). Statistical significance was determined as for Fig. 3.

TABLE 3.

Effects of algR and algZ mutations in clinical isolates on HCN production

Strain	Description or genotype	HCN production (mean μmol/μg protein ± SD)
TUMC92	Clinical	5.64 ± 0.64
TUMC92R	ΔalgR	9.30 ± 0.32
TUMC92Z	ΔalgZ	8.02 ± 0.69
FRD-1	Clinical	7.60 ± 0.62
FRD-1R	ΔalgR	2.38 ± 0.47
FRD-1Z	ΔalgZ	2.89 ± 0.08

AlgZ and AlgR regulation of HCN production is surface attachment specific.

AlgZ and AlgR are both required for fimbrial biogenesis (80, 81) and affect alginate production (55, 56, 88). Both of these phenotypes are associated with biofilm formation and surface colonization (60, 78). Moreover, AlgR control of the rhlI and rhlA promoters has recently been shown to be surface attachment specific (58). In order to determine if AlgZ and AlgR regulation of HCN production was dependent on surface growth, cyanide assays were performed on nonmucoid and mucoid strains both grown in broth or on agar plates in stationary phase, 8 h for broth-grown strains and 20 h for plate-grown strains under microaerobic conditions (see Materials and Methods) (Fig. 1A and B). In agreement with our previous data, strains grown in broth produced approximately half (Table 4) the amount of HCN observed when the same strains were grown on plates (Fig. 4). In broth, neither algR nor algZ mutation had an effect on HCN production in either background (Table 4) whereas the same strains grown on agar plates resulted in threefold changes in HCN production (Fig. 4). Taken together, these results indicate that AlgZ/AlgR-dependent cyanide production was regulated through a surface-attachment-specific mechanism.

TABLE 4.

Effects of algR and algZ mutations in broth-grown strains on HCN production

Strain	Genotype	HCN production (mean μmol/μg protein ± SD)
PAO1	Wild type	0.63 ± 0.05
PSL317	PAO1 ΔalgR	0.55 ± 0.01
PAZ	PAO1 ΔalgZ	0.55 ± 0.01
PAO568	mucA2	0.55 ± 0.01
PAR568	PAO568 ΔalgR	0.57 ± 0.01
PAZ568	PAO568 ΔalgZ	0.63 ± 0.01

AlgR D54 is critical for HCN production.

The aspartate at position 54 of AlgR is highly conserved among two-component regulators and is able to accept phosphate from E. coli CheA in vitro (20, 50). An additional aspartate at position 85 is conserved in the AlgR homologues Staphylococcus aureus LytR and E. coli YehT but is not capable of accepting phosphorylation. AlgR D85 has been implicated in controlling the amount of pilin subunits exported (81). To determine the role of these conserved aspartates in HCN production, site-directed AlgR mutants containing D54N, D85N, or D54N D85N chromosomal alleles (81) were tested for their ability to produce HCN under microaerophilic conditions on PIA. These strains were made mucoid by mutating the mucA gene, and HCN production was measured from all isolates. Though the parental strain of these mutants, a PAO1 strain obtained from the Wozniak laboratory, produced less cyanide than did our PAO1 strain, the same trends for HCN production were observed for an algR mutant and an algZ mutant as with our PAO1 strain (data not shown). The AlgR D54N and AlgR D54N D85N mutants produced increased HCN levels in nonmucoid strains (∼2-fold [Fig. 5A]) while producing lower levels in mucoid (mucA::Tc^r) strains (∼3-fold [Fig. 5B]). Similar to the results obtained from assays of twitching motility (81) and alginate production (50), the D85N mutant had no effect on HCN production. Deletion of algZ resulted in the same effects on HCN production as those found with the AlgR D54N mutant (Fig. 5). These data show that repression of HCN production in nonmucoid strains and activation in mucoid strains are both dependent on AlgZ and AlgR D54. Taken together, these data strongly suggest that phosphorylation of AlgR D54 is required for HCN regulation and that AlgZ is required for phosphorylation of AlgR. However, these data also indicate that phosphorylation of AlgR is not the mechanism responsible for its ability to switch from a repressor in nonmucoid P. aeruginosa to an activator in mucoid isolates for HCN production.

FIG. 5.

AlgR D54 is critical for HCN. (A) HCN production was measured from WFPAO1, WFPAO1 with chromosomal algR mutation D54N (AlgR D54N), WFPAO1 with chromosomal algR mutation D85N (AlgR D85N), WFPAO1 with double chromosomal algR mutations D54N and D85N (AlgR D54N D85N), WFPAO1 containing chromosomal algR mutation D54N and an algZ deletion (D54N ΔalgZ), and WFPAO1 containing chromosomal mutations in algR D54N and D85N and an algZ deletion (D54N D85N ΔalgZ). (B) HCN production from mucoid derivatives (mucA::Tc) of the same site-directed algR and algZ mutants as those in panel A. Statistical significance was determined as described for Fig. 3.

DISCUSSION

This study shows that mutations in algZ or algR affect HCN production in P. aeruginosa and that these effects are partially due to the phosphorylation state of AlgR. In addition, results shown here demonstrate that the hcnA promoter contains three transcriptional start sites with the most upstream T3 promoter being required for maximal activity. The hcnA promoter is controlled at transcriptional and posttranscriptional levels by many different regulators including ANR (62, 90), LasR and RhlR (62), GacA (7, 44, 61, 70), RsmA (10, 36, 44), and AlgR (13, 49). Two hcnA promoters have been previously reported for P. aeruginosa, termed T1 and T2 (62). T1 is regulated by quorum sensing alone, while T2 relies on a synergistic effect of LasR, RhlR, and ANR (62). In P. fluorescens, there are three RsmA/RsmE recognition elements that function as posttranscriptional regulators of the hcnA promoter by directly binding to the 5′ leader region (44). The chromosome of P. aeruginosa also contains rsmA, and deletion of this gene increased HCN production, suggesting that RsmA may function similarly in P. aeruginosa (63, 70). GacA also indirectly controls HCN production by positively regulating the rhlI promoter and hence amounts of N-butanoyl-homoserine-lactone (63, 70). The effects of algZ and algR deletion on hcnA transcription could be partially indirect, as AlgR represses rhlI transcription when P. aeruginosa is attached (58). However, AlgR has been shown to bind directly to the hcnA promoter in vitro, indicating that it controls HCN production directly (13). The most plausible explanation is that AlgR is interacting with other regulators (RhlR, ANR, or others) to control hcnA. Further studies are under way to determine the effects of AlgR, ANR, RsmA, LasR, and RhlR interactions on the hcnA promoter.

S1 nuclease protection assays (Fig. 2) established that there is one additional transcriptional start site upstream of the two previously reported. The hcnA′-′lacZ translational fusion used in a previous study that contained the T3 transcriptional start site showed higher β-galactosidase activity than did those fusions that lacked this promoter (62). The new start site is located in the exoY deletion region and is found in several laboratory and clinical strains (83). Results from Fig. 3 indicate that T3 may be at least as important as T1 and T2, as the absence of this 5′ region of the hcnA promoter resulted in extremely low levels of HCN production. These data suggest that the T3 promoter is responsible for a considerable amount of hcnA expression in P. aeruginosa. Interestingly, with the exception of PAK and CF27, most of the strains examined with this deletion region are exoU positive and exhibit increased type III secretion activity compared to PAO1, suggesting that only one of these virulence factors may be featured in a given strain (83).

It has been proposed that AlgZ acts as the sensor to AlgR in a two-component regulatory system (80, 89). The gene encoding AlgZ is directly upstream of algR, and both proteins share homology to two other sensor/regulator pairs, S. aureus lytS and lytR and E. coli yehU and yehT. In fact, AlgR, LytR, and AgrA compose a unique branch of the two-component system family, called the LytTR family (59). In this study we provide additional evidence that AlgZ acts as the AlgR sensor. Deletion of algZ resulted in the same phenotype as did an algR deletion and the site-directed algR(D54N) strain with respect to HCN production, strongly indicating that AlgZ acts through AlgR to control this promoter.

The AlgR response regulator responds to different environmental conditions depending upon the promoter examined. For instance, AlgR responds to nitrate concentrations on the algD promoter (57) and to surface contact or perhaps lowered oxygen tensions in control of the rhlI and rhlA promoters (58). HCN is not produced by P. aeruginosa under aerobic or anaerobic conditions (15) but is expressed under low oxygen tensions that range from 0.1 μM to 180 μM dissolved oxygen in liquid medium (14). This broad range of HCN production was not observed here (Fig. 1); however, the conditions used in this study were differing concentrations of atmospheric oxygen and not dissolved oxygen. The results presented here show that HCN production was maximal at 5% (1.5 μM O₂) and is within the range previously reported. Part of the oxygen dependence can be explained by the requirement for the anaerobic regulator ANR for hcnA transcription (62, 90). However, it is not clear how or if AlgR and ANR interact directly. The current data support a model where AlgR is repressing hcnA and ANR is activating hcnA, in a nonmucoid background. In the mucoid background, the data indicate that there may be some synergy between these two transcriptional regulators to activate hcnA, as both are required for maximal hcnA expression.

As with AlgZ and AlgR control of the rhlI and rhlA promoters (58), changes in HCN production were not observed in liquid cultures of P. aeruginosa in strains containing mutations in algR or algZ. These results together suggest that AlgR may respond to either a microaerobic environment or surface attachment. Recently, Ryall et al. have shown that bacteria of the Burkholderia cepacia complex are cyanogenic under biofilm conditions (a glass bead system and agar plate) but not in liquid cultures (73). The results shown in Fig. 1A are in agreement with those studies. Cultures grown on agar have been described as open-air biofilms (41) and share many characteristics, including surface attachment, high cell densities, and coordinated cellular behavioral patterns (41, 75). Type IV pili could directly mediate either interaction and simultaneously act as a sensor for attachment and have been shown to affect gene transcription of lipC (53). Alternatively, surface growth could have a nonspecific effect. For example, surface growth could result in high local concentrations of signaling molecules (N-acylhomoserine lactones) or alterations in the local available O₂ concentration, all things known to affect HCN production (14, 62, 63). In order to address this possibility, HCN was measured in a pilA strain that resulted in a slightly lower but not statistically significant amount of HCN compared to the amount produced by PAO1 (data not shown). Further studies are being performed to elucidate this potentially important mechanism.

P. aeruginosa has been proposed to grow in a biofilm in the CF lung, and this environment also contains the low oxygen tension and high cell density required for HCN production (77, 85). A steep oxygen gradient is present within secondary bronchi, reported as more than 70 times lower than the partial pressure of oxygen (pO₂) found in the primary bronchi (85). Recently, reports have shown that HCN is detectable in the sputum and bronchoalveolar lavage fluid (12, 40, 71) of CF patients and is associated with impaired lung function. These reports are consistent with our previous data showing that 86% of CF isolates produced HCN (13) and with the data of others who have measured volatile compounds from CF sputum samples and detected HCN in vitro (12). Altogether, these data indicate that HCN may be an important virulence determinant for P. aeruginosa in CF.

Previous studies show that AlgZ and AlgR are both activators of twitching motility, with AlgZ acting in an AlgR-dependent manner. It has also been shown that AlgZ is a repressor of alginate production in the mucoid strain PAO568 (88), while AlgR acts as an activator (19, 55, 56). The conserved AlgR aspartate at position 54 does not play a role in alginate production in P. aeruginosa (50); however, it is required for twitching motility (81). The different phenotypes observed in alginate and twitching motility with respect to AlgR D54N suggest both a phosphorylation-dependent and a phosphorylation-independent mechanism for AlgR regulation. Analysis of HCN production by P. aeruginosa strains demonstrated that AlgR switches from a repressor to an activator of the hcnA promoter depending upon mucoid status (13). This study demonstrates that AlgR control of HCN production requires AlgR D54, indicating that some regulation occurs through the phosphorylation-dependent mechanism. In contrast, the AlgR switch from repressor to activator is still observed in the AlgR D54N site-directed mutants in comparisons of the nonmucoid to the mucoid background. These data provide evidence that phosphorylation is not the mechanism by which AlgR switches from a repressor to an activator in the mucoid background. Furthermore, these data suggest that AlgZ acts as a kinase in the microaerobic environment necessary for HCN production. Further investigation into the protein-protein interactions and posttranslational modification of AlgR is required to explain the repressor/activator switch.