Functional Analysis of Genes for Biosynthesis of Pyocyanin and Phenazine-1-Carboxamide from Pseudomonas aeruginosa PAO1

Bacterial strains and plasmids.

All of the bacterial strains and plasmids used in this study are described in Table 1. Strains of P. fluorescens, P. aeruginosa, P. chlororaphis, P. aureofaciens, B. cepacia, B. phenazinium, and B. iodinum were grown at 28°C in Luria-Bertani (LB) broth (2). Escherichia coli strains were grown at 28 or 37°C in LB medium. To examine phenazine production, P. aeruginosa strains were grown on PIA plates (Difco Laboratories, Detroit, Mich.) or in PB medium (containing [per liter of distilled water] 20 g of Bacto Peptone [Difco Laboratories], 1.4 g of MgCl₂, and 10 g of K₂SO₄) (18) for 1 to 3 days at 28 or 37°C. Strains of P. fluorescens were tested for phenazine production in LB medium supplemented with 2% glucose. The antibiotics used in this study were gentamicin (300 μg/ml) and carbenicillin (500 μg/ml) in experiments with mutant derivatives of P. aeruginosa; tetracycline (15 to 20 and 100 μg/ml) in experiments with isolates of P. fluorescens and P. aeruginosa, respectively; and ampicillin (80 μg/ml) and tetracycline (12.5 μg/ml) in experiments with E. coli.

DNA manipulations.

Standard methods were used for plasmid DNA isolation, restriction enzyme digestion, agarose gel electrophoresis, ligation, and transformation (2). Pseudomonad cells were electroporated with a Gene Pulser system (Bio-Rad Laboratories, Hercules, Calif.) by using the procedure of Enderle and Farwell (17). Total DNA from P. aeruginosa was isolated and purified by using a cetyltrimethylammonium bromide miniprep procedure (2).

A 6.4-kb DNA probe containing the entire phz locus from P. fluorescens 2-79 (GenBank accession number L48616) was generated by PCR performed with oligonucleotide primers PHZ-UP and PHZ-LOW (Table 1). The amplification was carried out by using a 50-μl reaction mixture containing 1× eLONGase buffer (Life Technologies, Inc., Rockville, Md.), 2 mM MgSO₄, 3.0% dimethyl sulfoxide (Sigma Chemical Co., St. Louis, Mo.), 200 μM (each) dGTP, dATP, dTTP, and dCTP (Perkin-Elmer, Norwalk, Conn.), 10 pmol of each primer, 0.7 μl of eLONGase enzyme mixture (Life Technologies, Inc.), and 20 ng of purified genomic DNA from strain 2-79. All amplifications were performed with a PTC-200 thermal cycler (MJ Research, Inc., Watertown, Mass.). The cycling program included a 30-s initial denaturation step at 94°C, followed by 30 cycles of 94°C for 30 s, 64°C for 30 s, and 72°C for 7 min. The PCR product was gel purified and labeled with a random primer biotin labeling kit (NEN Life Science Products Inc., Boston, Mass.). Other PCR amplifications were carried out by using 25-μl reaction mixtures that contained 1× Taq DNA polymerase buffer (Life Technologies, Inc.), 1.5 mM MgCl₂, 200 μM (each) dGTP, dATP, dTTP, and dCTP, 20 pmol of each primer, 1.2 U of Platinum Taq DNA polymerase (Life Technologies, Inc.), and 20 ng of target DNA.

DNA samples used for Southern hybridization were digested with restriction endonucleases, separated by electrophoresis in 0.8% agarose gels, and transferred to BrightStar-Plus nylon membranes (Ambion, Inc., Austin, Tex.) in 0.4 M NaOH, and there was subsequent cross-linking of the DNA by exposure of the membranes to UV radiation (2). Hybridization analyses were carried out as described previously (13). DNA-DNA hybrids were detected with a BrightStar nonisotopic detection kit (Ambion, Inc.) by using the manufacturer's protocol.

Insertional inactivation of phzM and phzS genes.

P. aeruginosa PAO1 phzM and phzS knockout mutants were generated by using a gene replacement strategy previously described by Schweizer (50). To mutagenize phzM, the gene was amplified by PCR with oligonucleotide primers METHYL1 and METHYL2 (Table 1) from cosmid 1G5 DNA of P. aeruginosa. The 1.26-kb PCR product was digested with EcoRI and BamHI, cloned into pNOT19, and inactivated by insertion of an 879-bp aacC1 cassette from pUCGM into a unique EcoRV site (Fig. 1A). The resulting plasmid, pNOT-ORF1-Gm, was digested with NotI and ligated with the 5.3-kb MOB3 sacB cassette (50).

To inactivate phzS, the gene was amplified from cosmid 1G5 target DNA with oligonucleotide primers OXY1 and OXY2 (Table 1). The 1.24-kb PCR product was digested with XbaI and SphI, gel purified, and cloned into pNOT-19. The aacC1 cassette was then inserted into a unique ScaI site within phzS (Fig. 1A). The resulting plasmid, pNOT-ORF2-Gm, was ligated to the MOB3 sacB cassette and used for mutagenesis.

Plasmids containing the sacB cassette and the inactivated phzM or phzS gene were mobilized in P. aeruginosa PAO1 from E. coli S17-1. Following selection for double crossovers, isolates were screened by PCR for the presence of plasmid-borne bla and sacB genes in the genome. The β-lactamase gene was detected by PCR performed with primers BLA1 and BLA2 (Table 1), which amplified a 744-bp fragment of bla. The sacB cassette was detected by PCR performed with primers SAC1 and SAC2 (Table 1), which amplified a 1.05-kb DNA fragment. The cycling program included a 1-min initial denaturation step at 94°C, followed by 25 cycles of 94°C for 45 s, 53°C for 45 s, and 72°C for 1.25 min. The presence of mutated alleles of phzS and phzM was confirmed by PCR performed with oligonucleotide primers MET1 and MET2 and oligonucleotide primers ORF2UP and ORF2LOW (Table 1), respectively.

DNA sequencing and analysis.

DNA was sequenced by using an ABI PRISM dye terminator cycle sequencing kit (Perkin-Elmer, Norwalk, Conn.). All oligonucleotides were obtained from Operon Technologies, Inc. (Alameda, Calif.). Sequence data were compiled and analyzed with the Genetics Computer Group package (22) and the OMIGA 2.0 software package (Oxford Molecular Ltd., Oxford, United Kingdom). A database search for similar protein sequences was carried out with OMIGA's BLAST tool. A probable domain similarity search was performed by using the PROSITE (European Molecular Biology Laboratory, Heidelberg, Germany) (3) and ISREC ProfileScan (Swiss Institute for Experimental Cancer Research, Epalinges, Switzerland) (4) web servers. The significance of the similarity of a predicted protein to known proteins was determined by calculating the binary comparison score (Z score) as described elsewhere (13). Similarities with Z scores grater than 9 were considered significant.

Multiple sequence alignments were constructed with OMIGA's ClustalW. Phylogenetic analyses were carried out with PHYLIP 3.57c software (20). Distance matrices were generated with PROTDIST. Trees were inferred from the distances by using NEIGHBOR and FITCH with global rearrangement. Protein parsimony analyses were carried out with the PROTPARS program of the PHYLIP package. PHYLIP's SEQBOOT program was used for bootstrap analyses.

The present study was greatly facilitated by the use of the P. aeruginosa genome web site resources (http://www.pseudomonas.com).

Phenazine transformation assays.

Cultures of E. coli JM109 bearing pUCP26, pUCP-M, pUCP-S, pUCP-H, or pUCP-phnAB (Table 1) were grown in 2× YT (2) supplemented with tetracycline and were induced with 0.5 mM isopropyl-β-d-thiogalactopyranoside (IPTG) at an optical density at 600 nm of 0.6. Immediately after induction, PCA was added to a final concentration of 0.3 mg/ml from a 25 mM stock solution in 5% (wt/vol) NaHCO_3, and the cultures were grown for 6 h. Samples were extracted with chloroform at 3-h intervals and analyzed for phenazine composition by reverse-phase high-performance liquid chromatography (HPLC).

Cross-feeding experiments.

Overnight cultures of E. coli JM109 bearing pUCP26-based plasmids were diluted 1:100 in fresh 2× YT broth supplemented with tetracycline, grown with shaking to an optical density at 600 nm of 0.6, and induced with 0.5 mM IPTG. Immediately after induction, PCA was added to a concentration of 0.3 mg/ml, and the cultures were grown for 6 h to allow phenazine metabolites to accumulate. Finally, E. coli cells were removed by centrifugation and passage through a 0.22-μm-pore-size filter, and the filtrates were added to broth cultures of E. coli JM109 bearing pUCP26, pUCP-S, or pUCP-M. After 6 h of bacterial growth, phenazines were extracted with chloroform and analyzed by reverse-phase HPLC and mass spectrometry.

Analyses of phenazine compounds.

Phenazine compounds were extracted from bacterial cultures with chloroform, filtered (pore size, 0.2 μm), and subjected to C₁₈ reverse-phase HPLC. Analyses were performed by using a Waters HPLC Integrity system consisting of an Alliance 2690 separation module, a 996 photodiode array detector, and an electron ionization ThermaBeam mass detector. Phenazine compounds were separated on a Symmetry C₁₈ column (5 μm; 3.0 by 150 mm). The solvent flow rate was 350 μl min⁻¹, and the flow consisted of 2 min of 8% acetonitrile–25 mM ammonium acetate, followed by a 25-min linear gradient to 80% acetonitrile–25 mM ammonium acetate. A filtered 1 M ammonium acetate solution was used to prepare all solvents. HPLC gradient profiles were monitored at spectral peak maxima of 257.0 and 313.0 nm, which are characteristic of PCA and pyocyanin in the solvent system used. Mass spectrometry total-intensity chromatogram analyses were performed from m/z 100 to 355 at a rate of 1 scan s⁻¹. The nebulizer, ion source, and expansion region temperatures were 84, 220, and 80°C, respectively, with a helium flow rate of 15 lb/in². The UV spectra and mass spectral characteristics of the major phenazine compounds detected are summarized in Fig. 2A.

Reference samples of pyocyanin and PCA were purchased from Colour Your Enzyme (Bath, Ontario, Canada). Mass spectrometry analyses of protonated (molecular weight, 211) and unprotonated (molecular weight, 210) forms of synthetic pyocyanin revealed the presence of an unknown m/z 224 ion peak [m/z (relative intensity) 224(13) 211(58) 210(100, M⁺) 197(11) 196(14) 182(60, M⁺-CO) 181(44, M⁺-·CHO) 168(18) 167(29, M⁺-NCHO) 149(14) 102(7) 91(16, C₇H₇⁺) 79(13) 78(8) 77(16, C₆H₅⁺) 76(10) 75(5)] that was not detected (Fig. 2) when the same reference samples were analyzed with a Voyager matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) mass spectrometer (PerSeptive Biosystems, Framingham, Mass.), in which no heating of the sample occurred. The predominant peak detected was at m/z 211, which corresponded to the protonated form of pyocyanin. The MALDI-TOF analyses were carried out in dihydrobenzoic and α-cyano-4-hydroxycinnamic acids with an acceleration voltage of 2,500 V (positive ion), a grid voltage of 60%, a guide wire voltage of 0.03%, and a delay of 150. We concluded that the m/z 224 peak was the product of a thermally promoted intermolecular reaction in the ThermoBeam mass spectral detector, which is consistent with the observations of Watson et al. (60).

Phenazine compounds extracted from P. chlororaphis ATCC 17809 and ATCC 17411 were used as a source of reference material for PCN because the synthetic compound was not available. Both of these strains are well characterized and are known to produce large amounts of PCN (52). The retention time, UV spectra, and mass spectral characteristics of the major phenazine compounds produced by each of these strains were identical to those of the other and to those of the putative PCN peak observed in extracts from P. aeruginosa or E. coli expressing phzH from PAO1 in the presence of PCA (data not shown).

Nucleotide sequence accession numbers.

The GenBank accession numbers for the nucleotide sequence data for the phzA1B1C1D1E1F1G1 and phzA2B2C2D2E2F2G2 operons from P. aeruginosa PAO1 are AF005404 and AE004616, respectively.

Detection, cloning, and sequence analysis of phzA1B1C1D1E1F1G1 and phzA2B2C2D2E2F2G2.

The phzA1B1C1D1E1F1G1 genes were initially detected in P. aeruginosa PAO1 by Southern hybridization of total DNA restriction digests with a probe containing the phzCD genes from P. fluorescens 2-79. The analysis identified a 2.4-kb PstI-EcoRI DNA fragment that hybridized strongly to the 2-79 phzCD probe (data not shown). The fragment was cloned into pUC19, and two positive clones, plasmids p137 and p143, were identified by colony hybridization with the phenazine-specific probe (Fig. 1). Sequence analysis indicated that the cloned DNA fragment contained a complete sequence similar to phzD and partial flanking sequences similar to phzC and phzE of 2-79. The DNA from plasmid p143 was then used as a probe to screen a P. aeruginosa PAO1 genomic library (a membrane with arrayed cosmid DNA was a kind gift from Stephen Lory). Four positive cosmid clones were identified, and clone 1G5 was chosen for further study. After restriction mapping, a 10-kb BglII-NotI DNA fragment containing the phenazine biosynthetic genes was subcloned into pUCP26, and the resulting plasmid, pUCP-1G5, was used to determine the complete nucleotide sequence of the phenazine biosynthetic locus phzA1B1C1D1E1F1G1 from P. aeruginosa PAO1 (Fig. 1A).

The complete DNA sequence of the operon (GenBank accession number AF005404) is identical to the corresponding region of the P. aeruginosa PAO1 genome that was described by Stover et al. (53). Computer analyses of the DNA sequence of clone 1G5 revealed seven open reading frames, designated phzA1 through phzG1, located in the P. aeruginosa PAO1 genome at positions 4,713,795 to 4,720,062 (Fig. 1A). A well-conserved ribosome binding site precedes each gene. The stop and start codons of open reading frames phzC1, phzD1, and phzE1 overlap, possibly reflecting translational coupling. A 16-bp region of imperfect dyad symmetry (5′-CTACCAGATCTTGTAG-3′) located 372 bp upstream of phzA1 is similar to putative lux boxes found upstream of the phzA gene in P. fluorescens 2-79 and P. aureofaciens 30-84 (36, 43), suggesting that expression of the gene cluster is regulated in a cell density-dependent manner. Sequence comparisons indicated that phzA1B1C1D1E1F1G1 from P. aeruginosa is structurally and functionally homologous to the phenazine biosynthetic loci of P. fluorescens 2-79 (36), P. aureofaciens 30-84 (43), and P. chlororaphis PCL1391 (10) (data not shown).

Similarity searches performed with the PAO1 genome revealed a second copy of the phenazine cluster containing seven genes, designated phzA2B2C2D2E2F2G2 (Fig. 1B). As in the first copy, each gene is preceded by a well-conserved ribosome binding site, and the stop and start codons of phzC2, phzD2, and phzE2 overlap. A putative transcriptional terminator consisting of an 18-bp region of dyad symmetry (5′-ATAACCGCAAGCGGTTAT-3′) was identified 30 bp downstream of the phzG2 stop codon. The phzA2B2C2D2E2F2G2 gene cluster is located approximately 2.6 Mb from phzA1B1C1D1E1F1G1 at positions 2,070,685 to 2,076,985 of the P. aeruginosa PAO1 genome. The two phz operons are 98.3% identical at the DNA level. Essentially all of the sequence divergence occurs in the first pair of genes, phzA1B1 and phzA2B2; the regions containing phzC through phzG are nearly identical. Notably, the lux box in the putative promoter region upstream of phzA1 was not found upstream of phzA2. A cosmid clone bearing the full-length phzA2B2C2D2E2F2G2 gene cluster, clone 1H4, was identified by comparative Southern hybridization analyses performed with a 3.1-kb ScaI fragment containing phzE1, phzF1, and phzG1 (data not shown).

Identification of phzM, phzS, and phzH genes in the P. aeruginosa PAO1 genome database.

We identified two potential phenazine-modifying genes in P. aeruginosa PAO1 by analysis of the genomic region containing phzA1B1C1D1E1F1G1. This operon is flanked by genes designated PA4209 and PA4217 in the PAO1 database (53). PA4209, referred to in this paper as phzM, is preceded by a putative ribosome binding site, GAGAGA, and spans positions 4,713,098 to 4,712,094 in the PAO1 genome. The phzA1B1C1D1E1F1G1 operon and phzM are transcribed divergently and are separated by 695 bp (Fig. 1A). phzM encodes a 334-residue protein with a calculated molecular mass of 36.4 kDa that most closely resembles O-demethylpuromycin-O-methyltransferase (DmpM) from Streptomyces anulatus (NCBI accession number P42712; 30.1% identity; Z score, 72.4), a putative O-methyltransferase (SC5C11.09c) from Streptomyces coelicolor (accession number CAB76315; 30.9% identity; Z score, 64.4), O-methyltransferase (MmcR) from Streptomyces lavendulae (accession number AAD32742; 29.6% identity; Z score, 58.2), and carminomycin 4-O-methyltransferase (DauK) from Streptomyces sp. (accession number AAB16938; 28.9% identity; Z score, 44.9). A ProfileScan database search revealed a generic methyltransferase motif (residues 234 to 274) and a conserved S-adenosyl-l-methionine (SAM) binding domain (residues 164 to 273) within PhzM.

PA4217, referred to in this paper as phzS, is located 236 bp downstream from phzG1 and spans positions 4,720,300 to 4,721,508 of the P. aeruginosa PAO1 genome. phzS is preceded by a well-conserved ribosome binding site, AAGGAA, and encodes a 402-amino-acid protein with a molecular mass of 43.6 kDa. PhzS is similar to bacterial monooxygenases, including salicylate hydroxylase (NahW) from Pseudomonas stutzeri (NCBI accession number AAD02157; 27.7% identity; Z score, 38.9), a putative salicylate hydroxylase (SCE29.14c) from S. coelicolor (accession number T36193; 37.5% identity; Z score, 50.6), and putative salicylate hydroxylases from Bordetella pertussis (accession number AAC46266; 32.1% identity; Z score, 48.2) and Acinetobacter sp. (accession number AAC27110; 28.7% identity; Z score, 43.4). A Pfam database search revealed a monooxygenase motif (residues 157 to 366) and a putative N-terminal NAD binding site (residues 7 to 36) within PhzS.

A third potential phenazine-modifying gene was identified in the P. aeruginosa genome by using the sequence of the recently described phzH gene from P. chlororaphis PCL1391 (10) as a query sequence in a BLAST search (10). This search identified the PA0051 gene, which exhibits 80.3% identity to phzH and spans positions 66,303 to 68,135 in the P. aeruginosa genome (53). This gene, referred to in this paper as phzH, is preceded by a well-conserved ribosome binding site, GAGG, and encodes a 610-amino-acid protein with a calculated molecular mass of 69.5 kDa. The deduced amino acid sequence encoded by phzH most closely resembles the sequence of phenazine-modifying enzyme PhzH from P. chlororaphis (NCBI accession number AAF17502; 80.3% identity; Z score, 556.5) and is similar to the sequences of bacterial asparagine synthetases, including TcsG from Streptomyces aureofaciens (accession number BAB12569; 46.1% identity; Z score, 267.9), PA2084 from P. aeruginosa (accession number AAG05472; 45.4% identity; Z score, 291.6), and AsnO from Bacillus subtilis (accession number O05272; 40.2% identity; Z score, 273.3). Two well-conserved Pfam protein motifs commonly associated with asparagine synthetases were found in PhzH: the glutamine amidotransferase class II signature (residues 43 to 153) and an asparagine synthetase signature (residues 199 to 608).

Functional analyses of the core phenazine loci and the phnAB genes from P. aeruginosa PAO1.

To determine the function of the phenazine genes from P. aeruginosa, each of the seven-gene operons was cloned in pUCP26 under control of the lac promoter. The resulting plasmids, pUCP-A1G1 and pUCP-A2G2 (Fig. 1), and pUCP-1G5 were expressed in P. fluorescens M4-80, which does not produce phenazines, and in P. fluorescens 2-79, which produces a single phenazine compound, PCA. Extracts from each plasmid-bearing strain were analyzed by HPLC and mass spectrometry, and the data were compared to elution profiles and spectra generated with synthetic pyocyanin or extracts from wild-type P. fluorescens 2-79 and P. aeruginosa PAO1. Extracts from P. aeruginosa PAO1 contained a mixture of pyocyanin, 1-OH-PHZ, and PCA. Only PCA was detected in extracts from strain M4-80 containing either pUCP-A1G1 or pUCP-A2G2 (Table 2). Similar results were obtained for derivatives of strain 2-79, in which the presence of either plasmid led to increased levels of PCA accumulation (data not shown) but did not change the range of compounds synthesized (Table 2). In contrast, strain 2-79 bearing pUCP-1G5 produced large amounts of both PCA and the derivative 1-OH-PHZ (Table 2). Plasmid pUCP-1G5 contained the complete phzA1B1C1D1E1F1G1 operon and 3.3 kb of additional downstream DNA.

Previous studies of tryptophan biosynthesis in P. aeruginosa identified the phnAB genes(18), which encode an anthranilate synthase homologue that complements trpE and trpE(G) mutants of E. coli. PhnAB also contributes to pyocyanin synthesis in P. aeruginosa, and pyocyanin production by phnA mutants is greatly reduced (18). To further evaluate the role of phnAB in pyocyanin synthesis, the genes were excised from M13-10116 phage (the phage DNA was a kind gift from M. G. Bangera) by digestion with SacI and KpnI, treated with T4 DNA polymerase, and inserted into the SmaI site of pUCP26. The resulting plasmid, pUCP-phnAB, was introduced into P. fluorescens M4-80R and 2-79 by electroporation, and transformants were assayed for phenazine production by HPLC. In no case was pyocyanin produced, and strain 2-79 produced only PCA (Table 2) in quantities comparable to those detected in 2-79 containing pUCP26.

We also conducted phenazine transformation assays to determine whether phnAB conferred the ability to transform PCA, a precursor of pyocyanin and 1-OH-PHZ, into other phenazine compounds. This approach has proven to be useful in characterizing other genes encoding phenazine-modifying enzymes (13). Cultures of E. coli JM109 containing pUCP-phnAB or the control plasmid pUCP26 were induced with IPTG and incubated for 9 h in the presence of 0.3 mg of PCA ml⁻¹. HPLC analyses revealed that the extracts from E. coli JM109(pUCP-phnAB) were identical to those from the control strain, and no conversion of PCA to pyocyanin or any other phenazine derivative was detected (data not shown).

Functional analyses of phzM, phzS, and phzH

To determine whether phzM, phzS, and phzH encode phenazine-modifying enzymes, gene function was analyzed in E. coli transformed with pUCP-M, pUCP-S, pUCP-H, and pUCP-MS (Table 1). These plasmids were constructed by amplifying phzM, phzS, and phzH from PAO1 genomic DNA with oligonucleotide primers METHYL1 and METHYL2, oligonucleotide primers OXY1 and OXY2, and oligonucleotide primers phzHup and phzHlow, respectively (Table 1). The PCR products were digested with restriction endonucleases, gel purified, and cloned into pUCP26 under control of the lac promoter (Fig. 1). To construct pUCP-MS, pUCP-M was digested with XbaI and SphI, and the PCR fragment containing phzS was cloned immediately downstream of phzM. All plasmids were single-pass sequenced to confirm the integrity of cloned gene(s).

Induced cultures of E. coli bearing pUCP-MS converted PCA to pyocyanin within 6 h, whereas no such conversion occurred in control cultures harboring only pUCP26 (data not shown). We did not detect derivatives of PCA in supernatants of E. coli JM109(pUCP-M) cultures incubated under similar conditions, although these cultures were an intense dark red. Induced cultures of E. coli JM109 bearing pUCP-S and E. coli bearing pUCP-H converted PCA to 1-OH-PHZ and PCN, respectively (data not shown). We were able to restore the complete pyocyanin biosynthetic pathway in E. coli by mixing induced cultures containing either pUCP-A1G1 or pUCP-A2G2 with a culture containing pUCP-MS. HPLC analysis revealed that in both cases the mixed cultures produced small but detectable amounts of pyocyanin (Fig. 2B).

Plasmid pUCP-MS also was introduced into P. fluorescens 2-79. Overnight broth cultures of 2-79(pUCP-MS) were intensely blue, which was consistent with synthesis of pyocyanin, and the results of HPLC and mass spectrometry analyses (Fig. 2B) confirmed that large amounts of this compound were present.

Effect of phzM and phzS mutations on pyocyanin production.

The gene replacement approach was used to evaluate the role of the phenazine-modifying genes phzM and phzS in pyocyanin biosynthesis in P. aeruginosa PAO1. Mutant derivatives with mutations in phzM (strain PAOmxM) or phzS (strain PAOmxS) exhibited unusual pigment phenotypes when they were cultivated on PIA. Whereas cultures of wild-type PAO1 were blue due to production of pyocyanin, cultures of P. aeruginosa PAOmxM were yellow, and P. aeruginosa PAOmxS produced large amounts of a dark red water-soluble pigment (Fig. 3). Chloroform extracts prepared from cultures of these mutants grown in PB medium were analyzed by HPLC and mass spectrometry (Table 2). Both mutants were able to synthesize PCA and PCN but did not produce pyocyanin. In fact, the phzM mutant produced and secreted more PCN than the parent strain (data not shown). PAOmxM produced 1-OH-PHZ, although it produced markedly less than the parental strain, while no 1-OH-PHZ was detected in extracts from PAOmxS. Finally, complementation of PAOmxM and PAOmxS with pUCP-M and pUCP-S, respectively, restored production of pyocyanin (Table 2).