The Pseudomonas aeruginosa Orphan Quorum Sensing Signal Receptor QscR Regulates Global Quorum Sensing Gene Expression by Activating a Single Linked Operon

To identify possible binding sites of QscR within the genome of PAO1, we performed a chromatin immunoprecipitation assay using a purified anti-QscR antibody, followed by next-generation sequencing of the DNA bound to QscR. QscR is produced at very low levels in strain PAO1, and we were unable to reliably detect QscR-DNA complexes from the wild type using the ChIP-seq approach. Therefore, we expressed qscR under the control of an arabinose-inducible promoter in a lasR rhlR qscR triple deletion (3R) mutant and compared DNA sequences captured by an anti-QscR antibody from this mutant (at an optical density at 600 nm [OD₆₀₀] of 1.8) to those captured from a QscR-null mutant containing the vector only.

Our ChIP-seq analysis revealed a single region of the genome to which QscR appeared to bind, in the intergenic region between qscR and the adjacent PA1897 gene (P = <1 × 10⁻¹⁰, P score = >145) (Fig. 1). This promoter region has previously been identified as a target of QscR (10) with a QscR-binding site defined at −54 to −29 bp from the PA1897 transcription start site. We did not identify PA5351 in this ChIP-seq experiment, possibly reflecting the previous finding that the level of PA5351 promoter binding was substantially lower than that of PA1897 (10). Our finding of only a single QscR genome target was surprising; however, false-negative results are a consideration in any ChIP-seq analysis.

To probe for additional QscR targets not revealed by the ChIP-seq analysis, we performed electrophoretic mobility shift assays (EMSAs) with purified QscR and several promoters that have been shown to be repressed in a QscR transcriptomic analysis (12). We selected from this list those genes that were most likely to have a binding site for QscR, based on previous descriptions of LuxR-homologue binding targets (17). In addition, we asked if QscR could bind to the promoter of the LasR-regulated genes lasI (which encodes the 3OC12-HSL signal synthase) and lasB (elastase). As expected and predicted by the ChIP-seq results, addition of QscR to the promoter of PA1897 resulted in a DNA gel shift, consistent with binding of QscR to this promoter target, as previously reported (10) (Fig. 2). We did not observe a change in the electrophoretic mobility of DNA fragments containing other promoters (Fig. 2), supporting the idea that QscR does not affect transcription of these genes, including the LasR-regulated lasI and lasB genes, through direct binding to their promoters.

The ChIP-seq and EMSA results led us to hypothesize that the QscR counterregulation of QS occurs by activation of genes transcribed from the PA1897 promoter. Although transcription of QscR itself has been reported to be activated through this promoter (18), we asked if a PA1897 gene deletion, or a deletion of the operon containing PA1897, might phenocopy a QscR mutant.

The genomic region around PA1897 contains seven genes, PA1891-1897. PA1891-1894 are in a reading frame different from PA1895-1897, and there is a separate transcription start site for PA1894 (determined in P. aeruginosa PA14) (19) (Fig. 1). Whether or not they are in the same operon, PA1894, PA1895, and PA1897 have been reported to be regulated by QS (13). PA1895-1897 code for unknown proteins, although PA1895 and PA1897 have Pfam domains that suggest a role in fatty acid synthesis (Fig. 1).

Initially, we asked if a PA1897-null mutant or a PA1895-1897 deletion mutant (constructed to preserve the PA1894 transcriptional start site and promoter; see Table S1 in the supplemental material) would have the same phenotype of early QS gene activation as a QscR mutant. Deletion of qscR accelerates production of 3OC12-HSL and C4-HSL and the phenazine pyocyanin (9), although the final concentrations of all three were similar to those seen with the wild type. We found that both the PA1897 and PA1895-1897 mutants had enhanced pigmentation in mid-logarithmic-phase cultures (Fig. 3A) but the mutations did not affect growth. We observed a similar phenomenon with AHL signal production: both the PA1895-1897 and PA1897 deletion mutants exhibited early 3OC12-HSL and C4-HSL production compared to the wild type (Fig. 3B and C). Expression of a plasmid-borne PA1895-1897 operon with its native promoter was sufficient to complement the operon deletion mutant; however, expression of the PA1897 gene alone could not complement the PA1895-1897 mutation (Fig. 3), strongly supporting the idea that expression of all three genes in the PA1895-1897 operon, and not expression of PA1897 alone, is required for QscR to retard QS gene activation.

The ChIP-seq data and the data shown in Fig. 3 together suggest that QscR antiactivation of QS is mediated by the PA1895-1897 operon. To test this idea, we transformed a QscR mutant with plasmids containing either the arabinose-inducible PA1895-1897 genes or PA1897 alone. We used an arabinose-inducible promoter because the native QscR-responsive promoter would be ineffective in a QscR mutant background. Consistent with the idea that QscR activation of PA1895-1897 is the primary means by which this transcription factor impedes QS gene activation, arabinose-induced expression of PA1895-1897 was able to fully complement the QscR mutant. Expression of PA1895-97 in a QscR-null mutant reduced pyocyanin production at 12 h to the low, wild-type level (Fig. 4). Finally, and similarly to the experiments described above, pyocyanin production in the qscR mutant was not complemented by expression of PA1897 alone (Fig. 4). Together, these findings are consistent with the idea that the antiactivation phenotype of QscR relies upon its regulation of PA1895-1897.

A prior study showed that the QscR transcriptome substantially overlaps the LasR and RhlR regulons (12). Where LasR and RhlR activate many genes, QscR represses many of them (12, 13). Another study suggested that QscR acted to modulate LasI, although the mechanism was unresolved (9). One possible explanation, consistent with these studies and our findings, is that the PA1895-1897 gene products affect 3OC12-HSL levels, either by interfering with synthesis or by modifying the synthesized signal. LasR is highly specific for the 3OC12-HSL signal. Even small modifications to the chemical structure of the signal (including but not limited to reduction of the oxo-group or chain length changes) result in little LasR-dependent activity (10). Little is known about the products of PA1895, PA1896, and PA1897 (Fig. 1). We have not identified homologous proteins in any organism. Results of a Pfam search indicate that PA1897 possibly contains a fatty acid hydroxylase domain and that PA1895 may have a fatty acid desaturase domain. There are no putative functions for PA1896.

One of the simplest mechanisms by which the PA1895-1897 operon could mediate a delay in QS gene activation would be through AHL signal modification or degradation, comparably to the lactonase AiiA from Bacillus thuringiensis (20) or to the stationary-phase-expressed acylases of P. aeruginosa (21). To test the hypothesis that PA1895-1897 gene products modify 3OC12-HSL such that it no longer functions as a signal for LasR, we incubated P. aeruginosa cells expressing the PA1895-1897 operon with ¹⁴C-radiolabeled AHL (¹⁴C-AHL) signals (synthesized as described in Materials and Methods). After 2 h of incubation, the ¹⁴C-AHLs were subjected to solvent extraction, separated by high-performance liquid chromatography (HPLC), and quantitated by scintillation counting. Even though cells expressing PA1895-1897 were delayed in pyocyanin production relative to the vector controls (confirming that the expressed gene products were functional), there was no difference observed in ¹⁴C-AHL HPLC elution behavior (Fig. 5). These data suggest that neither 3OC12-HSL nor C4-HSL is degraded by the PA1895-1897 gene products, although we cannot exclude the possibility that they cause subtle modifications in the AHL compounds that do not alter the HPLC elution behavior.

The bacterial strains and plasmids used in this study are listed in Table S1 in the supplemental material. For all experiments, either P. aeruginosa or E. coli was grown at 37°C with shaking (250 rpm). Unless otherwise indicated, Luria-Bertani (LB) broth buffered with 50 mM 3-(N-morpholino)propanesulfonic acid (MOPS) was used. Media were supplemented with antibiotics for selection as follows: for E. coli, ampicillin at 100 µg ml⁻¹ and gentamicin at 10 µg ml⁻¹; for P. aeruginosa, carbenicillin at 300 µg ml⁻¹ and gentamicin at 100 µg ml⁻¹. Deletion mutants of P. aeruginosa were constructed by homologous recombination as previously described (31). To generate plasmids for homologous recombination, we used the method of Kostylev et al. (32). This technique requires the production of overlapping PCR products, including that of a plasmid, which are assembled in vivo by E. coli DH5α. Primers used for the generation of these constructs are listed in Table S2.

Bacterial strains used in this experiment include a P. aeruginosa PAO1 qscR deletion mutant and the P. aeruginosa lasR rhlR qscR triple deletion (3R) mutant. The qscR overexpression plasmid pJN105-QscR (10) was introduced into the 3R mutant. PAO1 ΔqscR and 3R-pJN105-QscR were incubated with 2 µM 3OC12-HSL in LB-MOPS (pH 7.0) at 37°C with shaking. For the 3R strain containing pJN105-QscR, 20 µg/ml gentamicin, with or without 0.01% arabinose, was added for plasmid maintenance. Growth was monitored as optical density at 600 nm.

A chromatin immunoprecipitation assay was performed for P. aeruginosa as previously described for LasR (33). Bacteria were grown in 40 ml of LB-MOPS to an OD₆₀₀ of 1.8. DNA was cross-linked to transcription factors with 1% formaldehyde and fragmented using a Branson microtip sonicator (three times for 10 s each time at an output of 0.4). The size of the fragments was confirmed to be in the 300-bp to 1,000-bp range by agarose gel electrophoresis. Fragmented DNA was immunoprecipitated using a 50% protein A Sepharose slurry (nProtein A 4 Fast Flow; Amersham Biosciences) with purified anti-QscR antibody (see below). Cross-links between the protein and DNA were reversed by incubating the eluate at 65°C overnight. DNA was purified using phenol-chloroform-isoamyl alcohol extraction and ethanol precipitation and was then further purified using a Monarch PCR and DNA cleanup kit (New England Biolabs). DNA was sequenced using an Illumina MiSeq system.

ChIP-seq libraries were prepared according to Illumina protocol and sequenced in a single MiSeq flow cell. ChIP-seq analysis of FASTQ output was performed using the Strand NGS (Bangalore, India) v 3.1.1 ChIP-seq pipeline. Single-end 150-bp reads were aligned to the PAO1 reference genome (RefSeq accession no. NC_002516). Peaks were identified using MACS (model-based analysis of ChIP-seq) peak detection (34) set to default parameters with the PAO1 qscR deletion mutant alignment as the control.

Purified recombinant QscR was prepared as previously described (35). A polyclonal antiserum against QscR was obtained from rabbits immunized with purified recombinant QscR. An affinity column, in which up to 9 mg of QscR was immobilized, was prepared using an AminoLink Plus immobilization kit (Thermo Fisher Scientific Company). Antibody was preadsorbed with a PAO1 qscR deletion mutant (9) grown in 50 ml of LB-MOPS at 37°C overnight. Cells were harvested by centrifugation, suspended in 2 ml of Tris-buffered saline (pH 7.2), and lysed by sonication. After insoluble material was removed by ultracentrifugation at 150,000 × g for 30 min, the cleared cell extract was mixed with an equal volume of antiserum and incubated for 4 h on ice. Insoluble aggregates were removed by ultracentrifugation, and the resulting supernatant was subjected to affinity purification with the prepared column. Binding, washing, and elution were performed according to the manufacturer’s instructions. The quality and specificity of the purified antibody were assessed by Western blotting.

We used Regulatory Sequence Analysis Tools (RSAT; http://embnet.ccg.unam.mx/rsat/) to search for promoter sequences with a conserved CT-N12-AG binding motif for LuxR-type transcription factors (36). Primers (Table S2) were designed to amplify 150-bp to 300-bp DNA fragments as probes to assess QscR binding in EMSAs. These probes were generated by PCR amplification with PAO1 genomic DNA as a template. The PCR-amplified probes (40 ng) were mixed with purified QscR protein (0 to 100 ng) in a 10-µl reaction mixture containing 20 mM Tris (pH 7.5), 50 mM KCl, 10% glycerol, 1 mM dithiothreitol, and 20 µM 3OC12-HSL. After 20 min at room temperature, the samples were separated by electrophoresis on a 5% nondenaturing acrylamide-Tris-glycine-EDTA gel in Tris-glycine-EDTA buffer at 4°C. Gels were soaked in 10,000-fold-diluted SYBR green I nucleic acid stain (catalog no. S7585; Invitrogen, Eugene, OR, USA), and DNA was visualized under UV light at 300 nm.

Pyocyanin assays were performed with P. aeruginosa grown in a pyocyanin production medium (PPM), which was composed of a pancreatic digest of gelatin (20.0 g), magnesium chloride (1.4 g), and potassium sulfate (10.0 g) in 1 liter of purified water containing 10 g of glycerol per liter (37). Pyocyanin was extracted from 5 ml culture fluid with 2 ml chloroform and was then extracted from the chloroform with 1 ml 0.2 M hydrochloric acid–water. The absorbance was measured at 520 nm, and this value was converted to milligrams per milliliter of pyocyanin by multiplying the optical density at 520 nm (OD₅₂₀) by 17.072 (38).

Concentrations of 3OC12-HSL were measured using a biological assay as described previously (10). To measure C4-HSL concentrations, we developed an mCherry reporter strain. This strain contains rhlR under the control of the tac promoter and a PrhlA-mCherry fusion. The reporter strain was grown overnight and was subcultured to an OD₆₀₀ of 0.025 into minimal A medium (5) with extracted AHLs. After 18 h, the samples were loaded into a 96-well plate (Costar assay; Corning Incorporated, Kennebunk, ME) and fluorescence was measured in a BioTek Synergy H1 plate reader (excitation, 587 nm; emission, 620 nm). Synthetic 3OC12-HSL and C4-HSL (Cayman Chemical) were used to generate standard curves.

To obtain ¹⁴C-labeled 3OC12-HSL and C4-HSL compounds, we used a radiolabeling protocol that was described previously (39, 40). P. aeruginosa PAO1 cells were grown with shaking at 37°C in Jensen’s medium–0.3% glycerol (41) to mid-log phase. A 10-ml aliquot of cells was incubated (90 min) with 10 µCi of l-[1-¹⁴C]methionine (American Radiolabeled Chemicals, St. Louis, MO), a portion of which is converted to S-adenosylmethionine, a substrate for AHL synthesis. To harvest ¹⁴C-AHLs, cells were removed by centrifugation and the cell-free supernatant was extracted twice with equal volumes of acidified ethyl acetate (EtAc). To test if the PA1895-1897 gene products modified ¹⁴C-AHLs, the ΔPA1895-1897 mutant harboring arabinose-inducible plasmid pJN105.PA1895-97 was inoculated into fresh PPM plus gentamicin and 0.05% arabinose and grown for 6 h (to an OD₆₀₀ of between 0.7 and 0.8) at 37°C with shaking. A 5-ml culture aliquot was transferred to a 15-ml polypropylene tube containing 1/20 of the total ¹⁴C-AHL extract (prepared as described above; solvent removed under a gentle stream of N₂ gas) and incubated at 37°C with shaking for 2 h. ¹⁴C-AHLs were reextracted twice with equal volumes of EtAc and evaporated to dryness. Approximately 30% of this extract was separated by C₁₈ reverse-phase HPLC in a 10% to 100% methanol gradient with a flow rate of 0.75 ml/min as described previously (42). Fractions were collected at 1-min intervals and mixed with 4 ml of counting cocktail (3a70b; Research Products International), and radioactivity was determined by liquid scintillation counting. For comparison, the same protocol was performed using the ΔPA1895-1897 mutant with pJN105 vector control and PPM broth.

Sequencing data from this study were deposited in the NCBI Sequence Read Archive under accession number SRP149739 and in BioProject under accession number PRJNA472067.