Ferripyochelin uptake genes are involved in pyochelin-mediated signalling in Pseudomonas aeruginosa

Iron is essential for most living organisms, including bacteria. However, despite its abundance on earth, iron is not freely available to micro-organisms under aerobic conditions, as it forms poorly soluble ferric hydroxides in the environment or is tightly bound to transport and storage proteins in mammalian hosts (Andrews et al., 2003). To acquire ferric ions, bacteria have evolved a number of strategies; the most common involves high-affinity iron-chelating molecules, termed siderophores (Guerinot, 1994; Wandersman & Delepelaire, 2004).

In Gram-negative bacteria, iron uptake is mediated by specific transport systems consisting of an outer-membrane receptor, which is energized by the TonB-ExbB-ExbD system (Moeck & Coulton, 1998), and an inner-membrane permease, which often belongs to the family of periplasmic binding protein-dependent ABC transporters (Köster, 2001). The entire ferrisiderophore complex may be transported to the cytoplasm, where iron is released from the chelator, e.g. in the case of iron uptake by ferrichrome (Köster, 1997). Alternatively, iron release can occur in the periplasm and the siderophore does not cross the inner membrane, e.g. during ferric citrate transport (Hussein et al., 1981).

Biosynthesis of siderophores and their cognate uptake systems is tightly regulated to ensure that they are produced only when needed and to avoid accumulation of iron, which can be deleterious to the cell (because free ferrous iron can catalyse the generation of hydroxyl radicals through the Fenton reaction: Andrews et al., 2003). Under iron-rich conditions, the Fur protein represses, directly or indirectly, the expression of siderophore biosynthesis and uptake genes (Escolar et al., 1999; Hantke, 2001; Prince et al., 1993; Vasil & Ochsner, 1999). Fur-mediated repression is alleviated when iron becomes limiting, allowing a basal level of gene expression to occur. Full expression often requires the presence of the siderophore. By a variety of different mechanisms involving extracellular cytoplasmic function (ECF) sigma/anti-sigma factors, two-component regulatory systems or AraC-type regulators, the siderophore induces the expression of genes necessary for its uptake and, in certain cases, also for its biosynthesis (Poole & McKay, 2003; Visca et al., 2002). In this type of regulation, also known as siderophore-mediated signalling (Lamont et al., 2002), perception of the siderophore can occur either at the cell surface, in the periplasm, or in the cytoplasm; examples are provided by pyoverdine-, enterobactin- and pyochelin-mediated signalling in the Gram-negative bacterium Pseudomonas aeruginosa. This opportunistic human pathogen produces two siderophores, pyoverdine and pyochelin (Cox, 1980; Cox & Adams, 1985; Meyer & Abdallah, 1978; Rinehart et al., 1995), but can promote iron uptake also with a variety of heterologous siderophores of fungal and bacterial origin (Poole & McKay, 2003). Pyoverdine is perceived at the cell surface, where interaction of the ferripyoverdine complex with the outer-membrane receptor FpvA transmits a signal to the anti-sigma factor FpvR. This inner-membrane-spanning protein then activates two ECF-sigma factors, PvdS and FpvI, which are required for the transcription of pyoverdine biosynthesis and uptake genes, respectively (Beare et al., 2003; Lamont et al., 2002; Redly & Poole, 2003). Perception of the heterologous Escherichia coli siderophore enterobactin occurs in the periplasm by the sensor kinase PfeS, which activates its cognate response regulator PfeR by phosphorylation such that PfeR becomes able to upregulate the transcription of the outer-membrane receptor gene pfeA (Dean & Poole, 1993; Dean et al., 1996). Pyochelin sensing occurs in the cytoplasm of P. aeruginosa. We have shown recently that pyochelin, possibly in its iron-loaded form (Michel et al., 2005), is the intracellular effector required by the AraC-type regulator PchR (Heinrichs & Poole, 1993, 1996) to activate the expression of the two pyochelin biosynthesis operons pchDCBA (Serino et al., 1997) and pchEFGHI (Reimmann et al., 1998, 2001), and of the fptA gene, encoding the outer-membrane ferripyochelin receptor (Ankenbauer & Quan, 1994). (Note that until the iron status of the PchR effector has been confirmed by additional experiments, the more general term pyochelin will be used here.) As addition of pyochelin to the growth medium triggers expression of these target genes in pyochelin-negative mutants, it can be concluded that the siderophore needs to be translocated to the cytoplasm in order to act as a PchR effector. Genes involved in pyochelin-mediated iron uptake could thus be involved in pyochelin-mediated signalling and hence might also affect pyochelin production.

The fptA gene, located immediately downstream of the pyochelin biosynthesis genes (Fig. 1), is followed by three contiguous ORFs, PA4220 (=fptB), PA4219 (=fptC) and fptX, which was reported recently to encode an inner-membrane permease required for growth with pyochelin as an iron source (Ó Cuív et al., 2004). Here we demonstrate that these genes form a ferripyochelin transport operon and we evaluate their importance in pyochelin utilization, signalling, and pyochelin production.

Table 1) is designated by Ω. Deletions are indicated by Δ and filled triangles represent the location of the vectors lac promoter. Arrows indicate the direction of transcription or the transcriptional units. Restriction sites used for subcloning experiments are indicated (note that only the locations relevant to this study are shown); artificially introduced sites are marked by an asterisk.

Bacterial strains, plasmids and growth conditions.
Bacterial strains, plasmids and cosmids are listed in Table 1. Bacteria were routinely grown on nutrient agar and in nutrient yeast broth (Stanisich & Holloway, 1972) at 37 °C. For β-galactosidase assays, P. aeruginosa strains were cultivated in GGP medium (Carmi et al., 1994), in which limited iron availability allows the expression of pyochelin biosynthesis and uptake genes. Pyochelin-utilization assays were performed with M9 minimal medium (Sambrook & Russell, 2001) using 0.5 % glycerol as a carbon source. Antibiotics were added to the growth media at the following concentrations: ampicillin (Ap) 100 µg ml¹, kanamycin (Km) 25 µg ml¹, spectinomycin (Sp) 50 µg ml¹ and tetracycline (Tc) 25 µg ml¹ for E. coli; carbenicillin (Cb) 250 µg ml¹, Sp 1000 µg ml¹ and Tc 125 µg ml¹ for P. aeruginosa. To monitor β-galactosidase expression, X-Gal was incorporated into solid media at a final concentration of 0.02 %. To counterselect E. coli donor cells in matings with P. aeruginosa, chloramphenicol (Cm) was used at a concentration of 10 µg ml¹; mutant enrichment was performed with Tc at a final concentration of 20 µg ml¹ and Cb at a final concentration of 2000 µg ml¹, as described previously (Ye et al., 1995).

Table 1. Strains, plasmids and cosmids

DNA manipulations and cloning procedures.
Small- and large-scale preparations of plasmid DNA were made with the cetyltrimethylammonium bromide method (Del Sal et al., 1988) and the Jetstar kit (Genomed), respectively. DNA fragments were purified from agarose gels with the Geneclean II kit (Bio 101) or the MinElute Gel Extraction Kit (Qiagen). DNA manipulations were performed according to standard procedures (Sambrook & Russell, 2001). Transformation of E. coli and P. aeruginosa was carried out by electroporation (Farinha & Kropinski, 1990). All constructs involving PCR techniques were verified by sequence analysis. Sequencing was performed with the BigDye Terminator Cycle Sequencing Kit and an ABI-PRISM 373 automatic sequencer (Applied Biosystems) or was carried out at Microsynth (). Sequences were compared with the DNA sequence available from the P. aeruginosa sequencing project (). Database searches were conducted at the National Center for Biotechnology Information () with BLAST algorithms, and protein sequences were analysed with the Expert Protein Analysis System ().

Construction of plasmids used for complementation (Fig. 1).
Plasmid pME7034, which carries fptA, was generated as follows. A 1.2 kb EcoRIEcoRV fragment containing the promoter and the 5' region of fptA was excised from pMO012405 and cloned into pUCPSK, giving pME7033. The resulting plasmid was digested with EcoRI and SmaI and the remaining part of fptA was added on a pMO012405-derived 1.7 kb EcoRIXhoI fragment, made blunt at its XhoI end by T4 DNA polymerase. To construct pME7036, which carries the entire fptABCX operon, a 4.8 kb EcoRINruI fragment from pMO012405 was cloned into pUK21 between EcoRI and StuI, giving pME7032. The 4.8 kb insert was then excised with EcoRI and SpeI and cloned into pME7033, cleaved with the same enzymes. Expression of fptA and fptABCX in pME7034 and pME7036, respectively, occurs from the fptA promoter as the vector-encoded lac promoter is located at the end of the cloned genes. Plasmid pME7204, expressing the fptX gene under P_lac control, was generated by inserting the 1.6 kb NotISpeI fragment from pME7032 into pUCPSK, linearized with the same enzymes.

Generation of translational lacZ fusions to fptB, fptC and fptX (Fig. 1).
The fptB''lacZ fusion carried by pME7518 was constructed as follows. First, a 0.2 kb fragment containing the 3' part of fptA and the first 2 codons of fptB was PCR-amplified from chromosomal DNA of PAO1 with primers yfpB-1 (GGCGTGAGCATGCGCCAGG; SphI-tagged) and fptB-1 (ACGTCTGCAGCGGCATCAGAACGCGCCCCG; PstI-tagged), cleaved with SphI and PstI and cloned into pME7215. From the resulting plasmid, a 2.8 kb fragment was excised with BglII and PstI and ligated to BamHI and PstI-linearized pME6015. An fptC''lacZ fusion was obtained in a similar way. A 0.5 kb PCR fragment, amplified from PAO1 chromosomal DNA with primers yfpB-1 and yfpB-2 (ACGTCTGCAGCGCCACTTCAACCGCGCCCC; PstI-tagged), was trimmed with SphI and PstI and cloned into pME7215. A 3.1 kb BglIIPstI fragment, carrying fptA, fptB and the first 2 codons of fptC, was excised from the resulting plasmid and cloned into pME6015 to give pME7517. To generate a translational lacZ fusion to fptX, a 2 kb PCR fragment was amplified with yfpB-1 and 4218-1 (ACGTCTGCAGAAGCATGGTGGTCTCCGGTG; PstI-tagged), cleaved with SphI and PstI and cloned into pME7215 as described above. Digestion with BglII and PstI yielded a 4.6 kb fragment containing fptA, fptB, fptC and the first two codons of fptX, which was cloned into pME6015 to generate pME7520. Plasmids pME7522, pME7521 and pME7523 lacking the fptA promoter region were generated by removing the 1.2 kb EcoRIEcoRI fragment from pME7518, pME7517 and pME7520, respectively.

P. aeruginosa mutant constructions.
Gene replacement mutants were generated as described previously (Ye et al., 1995) using suicide plasmids constructed as follows. To create in-frame fptA deletions, plasmid pME7041 (Fig. 1) was cleaved with SphI and BclI, treated with T4 DNA polymerase and religated, thus removing codons 54659 (note that the second BclI site present on pME7041 had not been cleaved during this experiment). The remaining 1.65 kb insert was excised with HindIII and BamHI and cloned into the suicide vector pME3087 to give pME7158. This plasmid was then used to delete the fptA gene in P. aeruginosa PAO1 and PAO6297, generating the corresponding mutants PAO6428 and PAO6429, respectively.

Chromosomal in-frame deletions in fptB were constructed as follows. First, a 0.97 kb XmaIIIXmaIII fragment containing fptB was excised from pME7041 and cloned into pBLS II KS linearized with NotI. The resulting plasmid served as template in an inverse PCR reaction with the BglII-tagged primers fptB-2 (ACGTAGATCTGGCGGGGCGCGGTTGAAG) and fptB-3 (GATCAGATCTAAGCCCGACTGGCGCGGCATC). The amplified fragment was cleaved with BglII and religated, giving a plasmid containing fptB with a deletion of codons 788 (ΔfptB). This ΔfptB gene, together with flanking DNA, was inserted on a 0.73 kb SacIHindIII fragment into pME3087 to give the suicide construct pME7208, which was used to generate the ΔfptB mutants PAO6423 and PAO6424.

The construction of the in-frame fptC deletion mutants PAO6387 and PAO6388 required several steps as well. A 0.54 kb XhoIXmaI fragment and a 0.4 kb NotIBamHI fragment, both originating from pME7032 and treated with T4 DNA polymerase at their XmaI and NotI ends, respectively, were ligated together and inserted into pBLS II KS between the XhoI and BamHI sites. This generated an fptC deletion derivative (ΔfptC) lacking codons 128445. The 0.94 kb XhoIBamHI fragment containing ΔfptC was then excised with KpnI and BamHI and cloned into pME3087 to yield the suicide construct pME7043, which was used to mutate fptC in strains PAO1 and PAO6297.

In strains PAO6368 and PAO6396, the fptX gene was mutated by insertion. A 1 kb ApaIPstI fragment originating from pME7032 and carrying the 5' part of fptX was first cloned into pBLS II KS, excised as a KpnIPstI fragment and inserted into pME3087. The 2 kb Ω-Sp/Sm cassette from pHP45Ω was then cloned into the BamHI site located 80 codons downstream of the fptX start codon. This generated the suicide plasmid pME7038 used to mutate fptX in strains PAO1 and PAO6297.

To generate pyoverdine-negative mutants, in-frame deletions were constructed in the pvdF gene as follows. Primers pvdF-1 (ACGTAGATCTTGCCCGGTATTTAGCGGC; BglII-tagged) and pvdF-2 (TCGAAAGCTTCAGAGCTTCTCGGCGAC; HindIII-tagged) were used to PCR-amplify the pvdF gene with its promoter region from chromosomal DNA of PAO1. The 1 kb PCR product was digested with BglII and HindIII and cloned into pUK21. A 0.22 kb XmaI fragment was removed from this plasmid to create an in-frame deletion in the pvdF ORF and the remaining insert was excised with BglII and PstI, and cloned into the suicide vector pME3087, cleaved with the same enzymes. The resulting plasmid pME7152 was then mobilized from E. coli S17-1 to P. aeruginosa PAO6297, PAO6429, PAO6424, PAO6388 and PAO6396, and chromosomally integrated, with selection for tetracycline resistance. Excision of the vector via a second crossing-over was obtained by enrichment for tetracycline-sensitive cells (Ye et al., 1995), generating the mutants PAO6383, PAO6541, PAO6540, PAO6390 and PAO6397, respectively. Deletion mutants were generally identified either by Southern blotting or by PCR analysis, while Ω-insertion mutants could also be screened for by their Sp/Sm-resistant phenotype.

Identification of salicylate, dihydroaeruginoate (Dha) and pyochelin in culture supernatants of P. aeruginosa.
P. aeruginosa strains were grown in GGP medium for the time indicated. For HPLC analysis, ethyl acetate extracts of acidified culture supernatants were dried by evaporation, dissolved in 60 % (v/v) methanol/10 mM H₃PO₄ and injected into an HPLC system as described previously (Reimmann et al., 1998). Compounds were identified by their retention times and UV spectra. Dha and salicylate were quantified at 256 and 237 nm, respectively. Pyochelin, which exists as a mixture of two interconvertible isomers, pyochelin I and pyochelin II (Rinehart et al., 1995), was quantified at 258 nm and 254 nm, respectively.

Pyochelin-utilization assays.
Utilization of pyochelin as an iron source was measured with liquid growth assays as follows. Erlenmeyer flasks (50 ml) with 15 ml M9-glycerol minimal medium containing, or not, the iron chelator 2,2'-dipyridyl at 500 µM were inoculated to OD₆₀₀ 0.02 with precultures grown in M9-glycerol medium. HPLC-purified pyochelin was added at 20 µM final concentration and growth at 37 °C and 220 r.p.m. was recorded for 150 h.

β-Galactosidase assays.
Fifty millilitre Erlenmeyer flasks containing 15 ml GGP medium were inoculated with 0.15 ml portions of precultures grown in the same medium. Incubation was at 37 °C and 220 r.p.m. for 12 h. When required, pyochelin, which was isolated from P. aeruginosa PAO1 and purified by HPLC (Reimmann et al., 1998), was added at a final concentration of 20 µM. β-Galactosidase activities were determined by the method of Miller (Sambrook & Russell, 2001) using cells permeabilized with 5 % (v/v) toluene.

RT-PCR.
Total RNA was extracted from a stationary-phase P. aeruginosa PAO1 culture in GGP medium by using the hot-phenol extraction method (Leoni et al., 1996). Residual DNA was digested with 40 U RNase-free DNase I (Roche). Four micrograms of total RNA was added to a 38 µl reaction mixture (Omniscript RT kit, Qiagen) containing primer fptB-bw (GACCACGCGCCAGCAACCCG), fptC-bw (GGATGTCGACGCCTTCCTCG) or fptX-bw (CGACCCAGGGTGCCCAGAGG), respectively, but lacking reverse transcriptase. Reaction mixtures were divided in two equal parts to which either 1 µl (=4 U) of reverse transcriptase was added (positive reaction) or RNase-free water (negative control). Incubation was for 2 h at 37 °C for reactions using fptB-bw (RT-1) or fptC-bw (RT-2), and at 42 °C when fptX-bw was used (RT-3). Subsequent PCR reactions were performed on 2 µl of each RT mixture, using primer pairs fptA-fw (GACTACAGCGTCGACTACCG) and fptB-bw (for RT-1), fptB-fw (CGACCGCCAGCGGCTATCTG) and fptC-bw (for RT-2), and fptC-fw (GCGGATTGCTCGGCGTCGCC) and fptX-bw (for RT-3).

fptB, fptC and fptX are transcribed from the fptA promoter
PA4220 (Fig. 1), originally described as ORF2 (Ankenbauer & Quan, 1994), seems to be translationally coupled to fptA as its predicted start codon overlaps with the fptA stop codon. The nucleotide sequence of PA4220 predicts a 93 amino acid cytoplasmic membrane protein of 9569 Da of unknown function. The following ORF, PA4219 [ORF 3 (Ankenbauer & Quan, 1994)], may be larger than originally proposed if translation initiates at a GTG codon located immediately downstream of PA4220. In this case, PA4219 would encode a protein of 501 amino acids with a molecular mass of 53 460 Da, predicted to be located in the cytoplasmic membrane as well. The fptX gene overlaps with PA4219 (Fig. 1) by a few bases and encodes a 414 amino acid protein of 43 151 Da, which belongs to a new family of inner-membrane permeases involved in siderophore transport (Ó Cuív et al., 2004). To test the possibility that PA4220, PA4219 and fptX could form an operon together with fptA, translational lacZ fusions to the second codon of each of these three genes were constructed, with or without the fptA promoter. β-Galactosidase activities were measured under iron-limiting growth conditions in P. aeruginosa PAO1. As shown in Table 2, all three fusions containing the fptA promoter (pME7518, pME7517, pME7520) were well expressed whereas expression dropped 100010 000-fold in the three constructs pME7522, pME7521 and pME7523 lacking the fptA promoter. Note that the expression of fptX was much stronger than the expression of PA4220 and PA4219. This is likely due to the fact that fptX is preceded by a plausible ribosome-binding site (GGAGA) whereas this is not the case for PA4220 and PA4219. The GTG initiation codon of PA4219 may also contribute to the low expression of this gene. From these results we conclude that (i) the two ORFs of unknown function, PA4220 (which we name now fptB) and PA4219 (now named fptC), are transcribed and translated; (ii) FptC is indeed larger than reported earlier and that translation is initiated at the proposed GTG codon located immediately downstream of fptB; and (iii) fptB, fptC and fptX form an operon together with fptA. The operon structure was confirmed in addition by an RT-PCR analysis, which revealed co-transcription of fptA with fptB, fptB with fptC, and fptC with fptX (Fig. 2).

Table 2. Expression of fptB, fptC and fptX requires the fptA promoter Cultures of P. aeruginosa PAO1 carrying the plasmids indicated were grown at 37 °C and 220 r.p.m. for 12 h in 50 ml Erlenmeyer flasks containing 15 ml GGP medium inoculated with 0.15 ml portions of precultures grown in the same medium. β-Galactosidase activities represent the means±SD of three parallel experiments.

(58K): Fig. 2. Transcriptional analysis of the fptABCX genes. RNA extracted from strain PAO1 grown under iron limitation was used in reverse transcription reactions containing (+), or not () reverse transcriptase. Reaction products generated with primers fptB-bw, fptC-bw, or fptX-bw, respectively, were used in subsequent PCR reactions to amplify specific DNA fragments of 420 bp, 421 bp and 400 bp, respectively, corresponding to transcripts between fptAB (primer pairs fptA-fw and fptB-bw), fptBC (primer pairs fptB-fw and fptC-bw) and fptCX (primer pairs fptC-fw and fptX-bw). No amplification occurred in control experiments performed without reverse transcriptase. M, 1 kb ladder (Promega).

Utilization of pyochelin as an iron source requires fptA and fptX but not fptB and fptC
We evaluated the role of the fptABCX genes in pyochelin utilization in strains defective for the production of pyoverdine and pyochelin. As shown in Fig. 3(a), the ΔpvdF, ΔpchBA mutant PAO6383 was unable to grow in iron-limited M9 medium but growth could be restored to a large extent when the medium was supplemented with 20 µM pyochelin as an iron source. Similar results were obtained with the PAO6383 derivatives affected in fptB (PAO6540, Fig. 3b) or fptC (PAO6390, Fig. 3c), indicating that these two genes were not required for pyochelin utilization. In contrast, mutations in fptA (PAO6541, Fig. 3d) or fptX (PAO6397, Fig. 3f) strongly interfered with pyochelin-mediated growth promotion, thus confirming the proposed functions for FptA and FptX in pyochelin-mediated iron uptake (Ankenbauer & Quan, 1994; Ó Cuív et al., 2004). Complementation of PAO6541 and PAO6397 with pME7036 (fptABCX) and pME7204 (fptX), respectively, fully restored pyochelin-mediated growth promotion (Fig. 3e, g).

(20K): Fig. 3. Role of the fptABCX genes in pyochelin-mediated growth promotion of iron-starved P. aeruginosa strains. M9 minimal medium containing (), or not () the iron chelator 2,2'-dipyridyl (500 µM) was inoculated with PAO6383 (ΔpvdF, ΔpchBA; a), PAO6540 (ΔpvdF, ΔpchBA, ΔfptB; b), PAO6390 (ΔpvdF, ΔpchBA, ΔfptC; c), PAO6541 (ΔpvdF, ΔpchBA, ΔfptA; d), PAO6541 carrying pME7036 (fptABCX; e), PAO6397 (ΔpvdF, ΔpchBA, fptX : : Ω; f), and PAO6397 carrying pME7204 (fptX; g). Pyochelin was added at 20 µM to M9 medium containing 2,2'-dipyridyl (•). OD600 values represent the means and standard deviations of three parallel experiments.

fptA and fptX are important for pyochelin-mediated signalling
We have previously shown that expression of the ferripyochelin receptor gene and of all pyochelin biosynthesis genes is strongly reduced in pyochelin-negative mutants but that expression is restored to wild-type levels when the growth medium is supplemented with pyochelin (Michel et al., 2005; Reimmann et al., 1998). To test whether pyochelin-mediated signalling involves genes of the fptABCX operon, we measured the expression of translational lacZ fusions to fptA and pchD in PAO1, PAO6297 (ΔpchBA), and its derivatives PAO6429 (ΔpchBA, ΔfptA), PAO6424 (ΔpchBA, ΔfptB), PAO6388 (ΔpchBA, ΔfptC) and PAO6396 (ΔpchBA, fptX : : ΩSp/Sm), in both the presence and the absence of exogenously added pyochelin. As shown in Table 3, both fusions were poorly expressed in all pyochelin-negative mutants but addition of 20 µM pyochelin restored expression of fptA''lacZ and pchD''lacZ to wild-type levels in PAO6297, PAO6424 and PAO6388, suggesting that fptB and fptC are not required for pyochelin-mediated signalling. In contrast, addition of pyochelin to strains PAO6429 and PAO6396 stimulated fptA and pchD expression to a certain extent, but expression remained below wild-type level. fptA and fptX thus play an important role in pyochelin-mediated signalling without being absolutely essential.

Table 3. Effect of fptA, fptB, fptC and fptX mutations on the expression of pyochelin biosynthesis and uptake genes P. aeruginosa strains carrying pME7153 (fptA''lacZ) or pME7160 (pchD''lacZ) were grown at 37 °C and 220 r.p.m. for 12 h in 50 ml Erlenmeyer flasks containing 15 ml GGP medium inoculated with 0.15 ml portions of precultures grown in the same medium. Pyochelin (Pch) was added to the culture medium at a final concentration of 20 µM. β-Galactosidase activities represent the means±SD of three parallel experiments.

fptA and fptX were also required for maximal expression of fptA''lacZ and pchD''lacZ in pyochelin-producing strains (Table 4), indicating that pyochelin needs to be released to the extracellular medium and then taken up again before it can interact with PchR.

Table 4. Impact of fptA and fptX mutations on signalling in pyochelin-producing strains P. aeruginosa strains carrying pME7153 (fptA''lacZ) or pME7160 (pchD''lacZ) were grown at 37 °C and 220 r.p.m. for 12 h in 50 ml Erlenmeyer flasks containing 15 ml GGP medium inoculated with 0.15 ml portions of precultures grown in the same medium. β-Galactosidase activities represent the means±SD of three parallel experiments.

Plasmid pME7204, which expresses fptX from the vector's lac promoter, fully restored the expression of fptA''lacZ in the fptX mutant PAO6396. By contrast, the fptA mutation in strain PAO6429 was complemented only partially by plasmid pME7034, which expresses fptA from its natural, iron-regulated promoter (Table 5). Full complementation was achieved only when the entire fptABCX operon was provided in trans on plasmid pME7036. It thus seems that the deletion in fptA, although in-frame, exerts a polar effect on the expression of the downstream genes. Indeed, plasmid pME7204 was able to partially complement the mutation in PAO6429 as well (Table 5).

Table 5. Complementation of fptA and fptX mutants P. aeruginosa strains carrying pME7153 (fptA''lacZ) and, where indicated, pME7204, pME7034, or pME7036, respectively, were grown at 37 °C and 220 r.p.m. for 12 h in 50 ml Erlenmeyer flasks containing 15 ml GGP medium inoculated with 0.15 ml portions of precultures grown in the same medium. Pyochelin (Pch) was added to the culture medium at a final concentration of 20 µM. β-Galactosidase activities represent the means±SD of three parallel experiments. The β-galactosidase activities of the uncomplemented mutants are included in Table 3.

Impact of the fptABCX operon on the formation of salicylate, Dha and pyochelin
During pyochelin biosynthesis, chorismate is converted to salicylate, which is then coupled to two molecules of cysteine by a thiotemplate mechanism (Serino et al., 1995; Gaille et al., 2002; Reimmann et al., 1998; Quadri et al., 1999). Dha, a byproduct of the pathway, is formed by P. aeruginosa in only small amounts and consists of a single cysteine moiety coupled to salicylate (Serino et al., 1997). We evaluated the importance of the fptABCX genes on the production of these metabolites (Table 6). Compared to strain PAO1, the production of Dha and pyochelin was reduced about threefold in the fptA mutant PAO6428 and by about 20 % in the fptX mutant PAO6368. Mutations in fptB and fptC did not affect the production of Dha and pyochelin and, as in the wild-type, salicylate was not detected as this compound is converted quantitatively to Dha and pyochelin under these experimental conditions. Small amounts of salicylate accumulated only in PAO6428, where a reduced pyochelin formation slowed down the entire biosynthetic pathway.

Table 6. Effects of fptA, fptB, fptC and fptX mutations on salicylate, Dha and pyochelin formation GGP medium (30 ml) was inoculated with 0.3 ml portions of precultures grown in the same medium. After incubation at 37 °C and 220 r.p.m. for 35 h, supernatants were extracted and analysed for their amount of salicylate, Dha and pyochelin by HPLC. The values given represent the means±SD obtained for three parallel experiments.

Taken together, these data show that the fptA and fptX genes are not only involved in pyochelin-mediated iron uptake as shown previously (Ankenbauer & Quan, 1994; Ó Cuív et al., 2004), but are also important for pyochelin-mediated signalling, which ultimately affects the production of the siderophore itself. We have shown in this work that the genes fptB, fptC and fptX are transcribed from the PchR-dependent (Heinrichs & Poole, 1996) and Fur-regulated (Ankenbauer & Quan, 1994) fptA promoter, thus forming an fptABCX operon (Table 2 and Fig. 2). Given the results of Table 2, additional, operon-internal promoters are unlikely. It is not clear, however, why the in-frame deletion in fptA affected fptX expression (Table 5) and it is thus possible that the regulation of the fptABCX operon is more complex. BLAST searches with microbial genomes reveal similar fptABCX operons in Burkholderia pseudomallei K96243, Burkholderia sp. 383 and Rhodospirillum rubrum. Whereas pyochelin production is documented in Burkholderia sp. (Darling et al., 1998; Visser et al., 2004) and pyochelin biosynthetic genes are highly conserved in the genomes of B. pseudomallei K96243 and Burkholderia sp. 383, there is no evidence for pyochelin formation by R. rubrum. The presence of an operon similar to fptABCX suggests, however, that R. rubrum might be able to utilize pyochelin as a heterologous siderophore for iron uptake.

Previous work has established that fptA and fptX are both involved in pyochelin-mediated iron uptake (Ankenbauer & Quan, 1994; Ó Cuív et al., 2004). We have shown here that these genes are also important in pyochelin-mediated signalling in pyochelin-deficient and in pyochelin-producing strains (Tables 3 and 4) and, as a consequence, affect the production of pyochelin (Table 6). This is consistent with the fact that pyochelin, possibly in its iron-loaded state, acts as an effector for the cytoplasmic PchR regulator (Michel et al., 2005). After FptA-promoted translocation of ferripyochelin across the outer membrane, the inner-membrane permease FptX is believed to be necessary for subsequent ferripyochelin uptake into the cytoplasm. FptX belongs to a new family of single-subunit siderophore transporters which differs from classical, binding-protein-dependent ABC proteins, such as FhuBCD, which is required for ferric hydroxamate uptake in E. coli (Braun & Killmann, 1999). Members of this new family seem to lack associated proteins that function in energy coupling or as periplasmic binding proteins (Ó Cuív et al., 2004). The proteins encoded by the fptB and fptC genes are not expected to have such a role and we have shown here that they are not essential for pyochelin utilization (Fig. 3), signalling (Tables 3 and 4) or pyochelin production (Table 6). Given their co-regulation with fptA and fptX and their conservation in Burkholderia sp. and R. rubrum, it is difficult to believe that fptB and fptC should not be involved in these processes at all. We cannot exclude that their functions are redundant in P. aeruginosa such that fptB and fptC mutants do not show a distinct phenotype.

FptX seemed less important for signalling than FptA (Tables 3 and 4) and the fptX mutation did not entirely abolish, but rather delayed, pyochelin utilization (Fig. 3). While this can be explained at least in part by the polar nature of the fptA mutation (Table 5) it also indicates that, in the absence of fptX, ferripyochelin may enter the cytoplasm by an alternative permease, as transport across the inner membrane exhibits less specificity than at the outer membrane. The FhuBCD permease, for instance, facilitates the uptake of several hydroxamate siderophores, each of which requires its own receptor at the outer membrane (Köster, 2001). Some hydroxamate siderophores can also be transported via a heterologous permease related to FptX (Ó Cuív et al., 2004), illustrating that classical ABC-type permeases and single-subunit siderophore transporters may replace each other in some cases.

Mutations in ferripyochelin uptake genes also affected metabolite production (Table 6) although the effect was less pronounced than on gene expression (Tables 3 and 4). This is likely due to the fact that under the experimental conditions used, gene expression is not the only parameter which determines the amount of product formed. Indeed, we have shown previously that pyochelin formation could be significantly increased when the growth medium was supplemented with cysteine (Gaille et al., 2003), indicating that the intracellular cysteine pool is another important factor.

In conclusion, we have shown here that pyochelin-mediated signalling (and hence pyochelin production) involves the ferripyochelin uptake functions FptA and FptX. These results thus confirm and extend our previous work demonstrating that pyochelin, possibly in its iron-loaded form, acts as an intracellular effector of the AraC-type regulator PchR (Michel et al., 2005). In the absence of ferripyochelin uptake functions, pyochelin-mediated signalling is, however, not entirely abolished, indicating that target gene induction may occur by an alternative signalling pathway as well.

We thank Catherine Gaille for pyochelin purification, Nicolas González for help with mutant construction and Dieter Haas for helpful suggestions and for carefully reading the manuscript. This work was supported by the Swiss National Science Foundation for Scientific Research (project 31-102174).

Edited by: P. Cornelis

Footnotes

†Present address: BIOMERIT Research Centre, Department of Microbiology, National University of Ireland, Cork, Ireland.