Phenylacetyl Coenzyme A, Not Phenylacetic Acid, Attenuates CepIR-Regulated Virulence in Burkholderia cenocepacia

Creation and phenotypic characterization of a paaK double mutant.

Previously, we demonstrated a link between the attenuation of B. cenocepacia K56-2 ΔpaaABCDE and the release of PAA (18). However, the interruption of the pathway at this step could also result in an accumulation of PAA-CoA because PAA is converted to PAA-CoA by a PaaK ligase (Fig. 1). To determine the role of each of these molecules in the attenuation of virulence, we characterized a paaK mutant that should produce only PAA and compared it to the paaABCDE mutant. Similar to the reference strain B. cenocepacia J2315 (22), K56-2 harbors two similar but not identical phenylacetyl-CoA ligases, PaaK1 and PaaK2 (23). We created markerless deletion mutants of paaK1 (BCAL0404) and paaK2 (BCAM1711). Knockout mutants of the other paaK gene were then made in each background, resulting in two independently made double knockouts, which we have named the ΔpaaK1 ΔpaaK2 and ΔpaaK2 ΔpaaK1 strains.

To confirm the interruption of the PAA pathway, K56-2 and mutants were grown on lysogeny broth (LB) medium and M9 minimal medium with glucose or PAA as a sole carbon source. No growth defect was observed in LB or glucose medium (Fig. 2A and B). Both paaK single knockout mutants were also able to grow on PAA as a sole carbon source to wild-type levels, indicating that PAA can be used as a substrate for both ligases (Fig. 2B). While mutants were able to grow normally on glucose as a sole carbon source, the paaK double knockout mutants (ΔpaaK1 ΔpaaK2 and ΔpaaK2 ΔpaaK1 strains) were unable to grow on PAA (Fig. 2B). Growth was complemented with a plasmid containing a rhamnose-inducible paaK2 (BCAM1711) (see Fig. S2 in the supplemental material), confirming that the interruption of the pathway in these mutants was caused by the deletion of paaK. Therefore, mutants with full interruption of the pathway (ΔpaaABCDE, ΔpaaK1 ΔpaaK2, and ΔpaaK2 ΔpaaK1 strains) were unable to grow on PAA as a sole carbon source. Single deletion mutants of the paaK ligases (ΔpaaK1 and ΔpaaK2 strains) grew to wild-type levels on PAA, indicating that the pathway is not interrupted in these strains.

The PaaABCDE mutant has been shown to accumulate extracellular PAA due to the interruption of the pathway (18), but the concentrations of intracellular PAA and PAA-CoA were unknown. Intracellular and extracellular concentrations of both PAA and PAA-CoA were measured with ion-pair reversed-phase ultrahigh-performance liquid chromatography tandem mass spectrometry (IP-RP-UHPLC-MS/MS) using the differential method (24–26). In general, high levels of PAA were found extracellularly rather than intracellularly. The average extracellular concentrations of PAA in the wild-type, ΔpaaABCDE, ΔpaaK1 ΔpaaK2, and ΔpaaK2 ΔpaaK1 supernatants were 37.3, 357.0, 107.3, and 93.9 μM, respectively (Fig. 2C), whereas the average values for intracellular PAA in the same strains were 6.0, 2.0, 1.7, and 8.0 μM (Fig. 2D). When we examined the concentrations of PAA-CoA, we found that the average extracellular levels of PAA-CoA were below or near the limit of detection for all strains, as expected for a CoA molecule that cannot transverse cell membranes. However, PAA-CoA accumulated intracellularly to levels slightly above the limit of detection (Fig. 2E). The intracellular levels of PAA-CoA (Fig. 2F) were higher in the ΔpaaABCDE mutant (0.94 μM) than in the PaaK mutants (0.54 μM and 0.38 μM). However, the low-micromolar amounts detected and the high variability between biological replicates made it challenging to establish clear differences between the PAA-CoA content of the wild type and the PAA degradation mutants.

To determine how interruption of each of the two consecutive steps of PAA degradation could affect QS-regulated virulence traits, the killing ability of PAA pathway mutants was compared to that of the wild type in slow-killing assays, which quantify the survival of gut-infected nematodes fed on the respective mutants (Fig. 3A). As previously seen, the ΔpaaABCDE mutant-exposed nematodes had a higher survival rate than those fed on K56-2 (P < 0.001). Surprisingly, the nematodes exposed to the ΔpaaK1 ΔpaaK2 mutant showed only 31% (±19%) survival on day 2. This was an increase in pathogenicity compared to that of K56-2, which had 66% (±10%) survival on day 2 (P < 0.0001). This virulent phenotype was also observed in nematodes fed with the ΔpaaK2 ΔpaaK1 mutant (Fig. S3A) but not in those fed with the single paaK deletion mutants (Fig. S3B).

To further investigate the opposite virulence phenotypes of the ΔpaaABCDE and ΔpaaK1 ΔpaaK2 mutants, we measured the exoprotease activity of both strains, a well-studied virulence trait of B. cenocepacia that is regulated by CepIR (10, 27). To ensure that PAA-related metabolites could be produced, we added the precursor phenylalanine to the 2% skim milk agar used for the exoprotease assay, as previously performed (18). In the absence of phenylalanine, the proteolytic activity of the ΔpaaABCDE mutant was similar to that of K56-2. However, when phenylalanine was added to the medium, the proteolytic activity of the ΔpaaABCDE mutant was abolished (Fig. 3B). On the other hand, the proteolytic activities of the ΔpaaK1 ΔpaaK2 and ΔpaaK2 ΔpaaK1 strains were higher than the activity of K56-2, regardless of the addition of phenylalanine (Fig. 3B and Fig. S3C). This is in agreement with the results from the slow-killing assays where only the ΔpaaABCDE mutant was attenuated for virulence while the ΔpaaK1 ΔpaaK2 and ΔpaaK2 ΔpaaK1 mutants were more pathogenic than the wild type. Together, we observed no correlation between extracellular levels of PAA and pathogenicity (Fig. S4).

PAA-CoA has no effect on CepR and CepI function in vitro.

Because PAA-CoA synthesis is interrupted in the paaK mutants but not in the paaABCDE mutant, we hypothesized that PAA-CoA was responsible for attenuating virulence. As the ΔpaaABCDE mutant is attenuated for virulence traits that are under the control of the CepIR quorum sensing system (28), we focused our attention on a possible direct effect of PAA-CoA on CepIR. In addition, the ΔpaaABCDE mutant had decreased cepI and cepR promoter activity and AHL signaling (18). Specifically, we asked whether PAA-CoA could (i) inhibit the formation of CepR:AHL complexes, (ii) inhibit the enzymatic activity of CepI, and/or (iii) indirectly affect an upstream regulator.

To address the first question, we built a reporter system similar to that commonly used to identify quorum sensing inhibitors (29). E. coli W14 was selected as our reporter strain as it lacks the PAA pathway genes (30, 31), ensuring that the exogenously added PAA-related compounds would not be degraded. Using arabinose-inducible expression of cepR (BCAM1868), we created a reporter strain that responds to CepR:AHL complexes by activating the cepI promoter, which controls the expression of luminescence genes luxCDABE (Fig. 4A). Expression of luminescence is decreased in the presence of an inhibitor of the formation of CepR:AHL complexes. However, PAA-CoA cannot be added exogenously to cultures because it is too large to cross the membrane. Therefore, we created a plasmid that contains the E. coli W paaK ligase gene (ECW_RS07845) under the control of a rhamnose-inducible promoter. In the presence of rhamnose, PaaK converts exogenously added PAA to PAA-CoA, as confirmed by IP-RP-UHPLC-MS/MS (Fig. 4B). To ensure that any effects we saw were not due to an inhibition of the arabinose-inducible promoter or luminescence itself, we created two control strains that contain either the arabinose-inducible P_BAD promoter or a constitutive promoter, dhfr, controlling the luxCDABE genes.

Using this reporter system (W14/pTL7/pTL5/pTL22 is referred to from here on as the CepR reporter strain) (Fig. 4A), we observed that although there was inhibition of luminescence in the presence of PAA-CoA, the level of inhibition was similar to the levels of the control strains (Fig. 4C). Similarly, PAA itself also had no effect on the system (Fig. 4D). In summary, these results show that PAA-CoA and PAA do not directly affect CepI activity or CepR:C₈-HSL formation.

To analyze if PAA or PAA-CoA inhibits the enzymatic activity of CepI, we used recombinant CepI in the presence of dichlorophenolindophenol (DCPIP) to spectrophotometrically measure the activity of CepI (32). CepI catalyzes the formation of C₈-HSL from octanoyl-acyl carrier protein (C₈-ACP) and S-adenosyl methionine (SAM). During this reaction DCPIP is reduced, resulting in a color change that can be used to measure activity of CepI. Enzyme activity showed no change in the presence of 200 μM PAA-CoA or PAA (Table 1), suggesting that neither molecule directly inhibits the activity of CepI.

As our data did not show a direct effect of PAA-CoA on the CepIR QS system but cepI and cepR promoter activity was decreased in the PaaABCDE mutant (18), we hypothesized that PAA-CoA may indirectly affect an upstream regulator of the CepIR QS system. If this was the case, we predicted that the differences in pathogenicity between the ΔpaaABCDE and the ΔpaaK1 ΔpaaK2 mutant strains would be explained by differential expression of cepI and cepR. Using stationary-phase cultures, we measured the transcript levels of cepI and cepR and the levels of C₈-HSL in the PAA pathway mutants and compared them to the levels in the wild type. Surprisingly, the ΔpaaABCDE and the ΔpaaK1 ΔpaaK2 mutants had wild-type levels of cepI and cepR transcription (Fig. 5A and Fig. S5A) despite the striking differences in their virulence phenotypes and contradicting our previous results where reporter systems measured decreased cepI and cepR promoter activity and decreased AHL signaling (18). This suggested that the attenuation of the ΔpaaABCDE mutant is independent of the transcription of the CepIR QS system. To confirm the wild-type transcription levels of our mutants, we measured the levels of AHL molecules in each strain. Similar to the transcription levels, the levels of C₈-HSL, taken from stationary-phase samples, did not correlate with the decreased virulence of our PAA pathway mutants (Fig. 5B and Fig. S5B). However, both the ΔcepI and ΔcepR mutants had no measurable C₈-HSL, which is consistent with their lack of virulence (Fig. 5B).

The CepR-independent virulence of PaaK knockout mutants is mediated by the presence/absence of PAA-CoA.

Transcription of cepI and cepR is not decreased in PAA pathway mutants, and C₈-HSL is present in the mutants, yet the virulence factors attenuated in the ΔpaaABCDE mutant (pathogenicity in Caenorhabditis elegans and exoprotease activity) are known to be under the control of the CepIR system. To further understand how the PAA pathway may regulate QS-related virulence without affecting cepIR transcription, we created markerless deletions of cepI and cepR in the PAA pathway deletion mutant backgrounds. We reasoned that if PAA-related regulation of virulence was completely independent of CepI and CepR activity, then deletion of cepI and cepR in a ΔpaaABCDE strain background should show an additive attenuation effect. Similarly, deletion of cepI and cepR in a ΔpaaK1 ΔpaaK2 mutant background should show the same degree of attenuation as the cepI and cepR deletions in a wild-type background. Once we had created knockout mutants of cepI and cepR for each PAA pathway mutant in parallel with cepI and cepR deletions in wild-type K56-2, we analyzed their growth and found no defect when strains were grown in LB medium over 24 h (Fig. S6). Next, we used virulence assays to determine if the CepIR QS system was needed for the regulation of their QS. The virulence levels of the cepI and cepR knockout mutants in wild-type and PAA mutant backgrounds were compared in slow-killing assays (Fig. 6A and B and Fig. S7A and B). Both cep QS mutants in the ΔpaaABCDE background were more attenuated for virulence than the ΔpaaABCDE strain (Fig. S7A). However, they failed to show an additive effect (Fig. S7A). Moreover, while the ΔpaaK1 ΔpaaK2 ΔcepI mutant was as attenuated as ΔcepI, the ΔpaaK1 ΔpaaK2 ΔcepR mutant was not, showing that the pathogenic phenotype was independent of CepR (Fig. 6A and B). This was mirrored in the ΔpaaK2 ΔpaaK1 ΔcepR mutant (Fig. S7B). This is surprising considering that the cepR deletion in the wild type and the paaABCDE mutant has a severe attenuation (Fig. S7A).

Corroborating the results of the C. elegans assay, the ΔpaaK1 ΔpaaK2 ΔcepR mutant had wild-type levels of proteolytic activity whereas the ΔpaaK1 ΔpaaK2 ΔcepI mutant had none (Fig. 6C and Fig. S7C). The only difference between the ΔpaaK1 ΔpaaK2 ΔcepR mutant and the cepR mutant is the presence of PAA. If PAA were responsible for the elimination of the cepR attenuation, the ΔpaaABCDE ΔcepR mutant should be pathogenic, but our results show that it is not.

To further understand the virulence of the ΔpaaK1 ΔpaaK2 ΔcepR mutant, we examined the expression levels of cepI and cepR and the levels of C₈-HSL in the mutants. While transcription of cepR is abolished in ΔcepI, it is restored to wild-type levels when the deletion of cepI occurs in a ΔpaaK1 ΔpaaK2 or ΔpaaABCDE mutant background (Fig. 6D; see also Fig. S7D for ΔpaaK2 ΔpaaK1). Therefore, PAA, which is present in both PAA pathway mutants, restores cepR transcription but not pathogenicity because there is no CepI and no C₈-HSL (Fig. 6F). When cepI transcript levels were analyzed in ΔcepR in the wild type and PAA degradation mutant backgrounds, the transcription of cepI was downregulated in the CepR mutant, as expected. Similarly, cepI transcription was reduced in the attenuated ΔpaaABCDE ΔcepR mutant and in the pathogenic ΔpaaK1 ΔpaaK2 ΔcepR mutant (Fig. 6E; see also Fig. S7E for the ΔpaaK2 ΔpaaK1 ΔcepR mutant). The lack of C₈-HSL in both the ΔpaaK1 ΔpaaK2 ΔcepI and ΔpaaK1 ΔpaaK2 ΔcepR mutants (Fig. 6F) suggests that the pathogenicity of the ΔpaaK1 ΔpaaK2 ΔcepR mutant is independent of cepI. This was mirrored in the ΔpaaK2 ΔpaaK1 ΔcepR mutant (Fig. S7F).

Deletions of the cepI and cepR genes in the PAA pathway mutant backgrounds allowed us to determine that the PAA pathway has an effect on CepIR-regulated virulence that is evident when CepI or CepR is absent. In this context, our results suggest that there is an alternative signaling pathway to activate virulence in the ΔpaaK1 ΔpaaK2 ΔcepR mutant. The opposite effects of the PaaK and PaaABCDE mutations with regard to pathogenicity and the striking differences that paaK and paaABCDE gene deletions have in the cepR mutant background suggest that PAA-CoA, or a derivative, is the central molecule implicated in this alternative signaling pathway.

Strains, medium preparation, and growth conditions.

Bacterial strains, listed in Table 2, were grown at 37°C with agitation at 230 rpm in lysogeny broth (LB) Lennox medium (without glucose) (43), unless specified otherwise. All cultures were started from isolated colonies grown 48 h on LB agar at 37°C, unless specified otherwise. Plasmids used are listed in Table 3. For E. coli strains the following antibiotics and concentrations were used: ampicillin (100 μg/ml; 200 μg/ml for W14 strain), chloramphenicol (30 μg/ml), kanamycin (30 μg/ml; 40 μg/ml for MM290 strain), trimethoprim (50 μg/ml; 100 μg/ml for W14 strain), and tetracycline (20 μg/ml). For B. cenocepacia strains the following antibiotics and concentrations were used: trimethoprim (100 μg/ml), tetracycline (100 μg/ml), and gentamycin (50 μg/ml). Stock solutions of C₈-HSL (Sigma-Aldrich) were suspended in dimethyl sulfoxide (DMSO) and stored at −20°C. For assays with exogenous addition of C₈-HSL, cultures were grown in LB medium supplemented with 50 mM 3-(N-morpholino)propanesulfonic acid (MOPS) to prevent pH-dependent lactonolysis (44). Stock solutions of l-arabinose (Alfa Aesar; Fisher Scientific) and l-rhamnose (Sigma-Aldrich) were prepared at a concentration of 20% (wt/vol) in deionized water and filter sterilized. Stock solutions of phenylalanine and PAA (Sigma-Aldrich) were prepared at 100 mM in deionized water, filter sterilized, and stored at room temperature. Stock solutions of PAA-CoA (Sigma-Aldrich) were prepared at 3.5 mM in DMSO and stored at –80°C.

Molecular biology techniques.

Genetic manipulation of B. cenocepacia K56-2 was performed via triparental matings with E. coli MM290/pRK2013 as a helper strain. Plasmids generated were maintained in E. coli SY327 λpir or DH5α Z-competent cells depending on their origin of replication. PCRs were performed with an Eppendorf Mastercycler EP gradient S thermocycler with either HotStar HiFidelity Taq polymerase (Qiagen) or Taq DNA polymerase (Qiagen). Conditions were optimized for each primer set; primers used are listed in Table 4. Restriction enzymes and T4 DNA ligase (New England Biolabs) were used as recommended by the manufacturer. PCR products and digested DNA were purified by a QIAquick Gel or PCR purification kit (Qiagen) and plasmids were isolated using a QIAprep Miniprep kit (Qiagen).

Construction of unmarked deletion mutants in B. cenocepacia.

Unmarked deletion mutants were created using the system developed by Flannagan et al. (45). Flanking regions of the target gene were synthesized as a custom gene fragment with XbaI and SmaI restriction sites (Integrated DNA Technologies, Inc.) (see Table S1 in the supplemental material for sequences). The fragment with regions flanking BCAL0404 (paaK1) and pGPI-SceI were digested and ligated to create pTL17, and pTL18 was created by insertion of the fragment with regions flanking BCAM1711 (paaK2). Each plasmid was introduced into wild-type K56-2 by triparental mating and were integrated into the genome by homologous recombination. Colonies were picked onto LB medium supplemented with trimethoprim and onto Burkholderia cepacia selective agar (BCSA) (46). Conjugates with a dull phenotype on BCSA were selected. The integration of the suicide plasmids was confirmed with primers 900/901 for pTL17 and primers 902/903 for pTL18. Then, pDAI-SceI, expressing I-SceI endonuclease, was introduced into K56-2::pTL17 and K56-2::pTL18 by triparental mating. The double-stranded break caused by the I-SceI endonuclease triggered a recombination event that resulted in a reversion of the conjugate to a wild-type allele or a deletion of the fragment between the two flanking regions. Potential conjugates were tetracycline resistant and trimethoprim sensitive. The deletion event was confirmed by PCR using primer set 980/981 for paaK1 and 982/983 for paaK2. Positive clones were cured of the pDAI-SceI plasmid by sequential subculturing in LB medium. Loss of the plasmid was confirmed by the loss of tetracycline resistance using the replica plate method and PCR with primer set 920/921 for the absence of pDAI-SceI.

Once cured, the mutants were tested for growth defects by washing overnight cultures twice in phosphate-buffered saline (PBS), diluting to an optical density at 600 nm (OD₆₀₀) of 0.04 in 200 μl of LB medium in a clear 96-well plate in triplicate. Plates were incubated at 37°C in a Biotek Synergy II plate reader with continuous shaking for 24 h, and the OD₆₀₀ was measured every hour. Deletions of cepI and cepR were performed in the wild-type and PAA pathway mutant backgrounds using the pDelCepI and pDelCepR plasmids created by Aubert et al. (47). Integration of the plasmids was confirmed using primer set 1082/154 for cepI deletions and 1080/154 for cepR. pDAI-SceI was introduced, and potential conjugates were selected on appropriate antibiotics; the deletion event was confirmed by PCR with primer sets 1112/1113 and 1080/1081 for cepI and cepR, respectively. Positive clones were cured of the pDAI-SceI plasmid and screened for the loss of the plasmid as described above.

Plasmid construction.

To create pKF2, paaK2 (BCAM1711) was amplified from genomic B. cenocepacia K56-2 DNA using primers 1055/1056 and restricted with XbaI and NdeI before ligation into similarly digested pSCrhaB2 (48). The plasmid construct was verified using primers 1055/848. To create the E. coli reporter system, cepR (BCAM1868) was amplified using primers 703/704, and the amplicon and pBAD30 were restricted with EcoRI and XbaI and then ligated to form pTL5. To construct a plasmid expressing the luxCDABE operon, the cepI promoter region including the first 15 codons of cepI was amplified, and an in-frame stop codon was added using primers 712 and 713. Plasmid pBBRlux and the amplicon were digested with SpeI and BamHI, and ligation then resulted in pTL7. Plasmids pTL12 and pTL13 were created in the same way with primers 768/769 for the P_BAD promoter and regulatory araC and primers 770/771 for the dhfr promoter region of pTnMod-STp. To create a plasmid that expressed PaaK and was compatible with the reporter strain, we had to use the pSC101 ori. pTL9 was created by digesting pTnMod-STp with BglII to remove the Tn5 transposase, and then the rhamnose-inducible promoter and its regulatory genes were added to create pTL21 using primers 724 and 947. Using primers 725 and 726, the paaK ligase from E. coli W was amplified from pAAD (30) and restriction cloned into pTL21 to create pTL22.

Growth of B. cenocepacia strains on various carbon sources.

One milliliter of culture from a 5-ml overnight culture in LB with appropriate antibiotics was washed in PBS twice, and the OD₆₀₀ was adjusted to 0.08 in 1 ml of M9 with 25 mM glucose or M9 with 5 mM PAA. A 100-μl aliquot was added to each well of a clear 96-well plate together with 100 μl of the appropriate medium for a starting OD₆₀₀ of 0.04 in triplicate. Plates were incubated at 37°C with aeration at 230 rpm, and the OD₆₀₀ was read after 48 h of incubation. OD₆₀₀ values were converted to 1-cm path length by prior calibration with a GeneQuant III 4283, version 4283V1.6.

Quantification of intracellular and extracellular levels of PAA and PAA-CoA by IP-RP-UHPLC-MS/MS.

B. cenocepacia cultures were diluted 1:100 into 10 ml of LB in a 50-ml Erlenmeyer flask and incubated at 37°C with agitation at 230 rpm for 6 h until late exponential phase. Using the differential method (24, 25), 1 ml of culture was immediately transferred into weighed tubes containing 5 ml of 60% (vol/vol) methanol in water held at –40°C in an acetonitrile–dry-ice bath, vortexed, and returned to –40°C. Two milliliters of culture was drawn into a syringe equipped with a 0.45-μm-pore-size filter, and 1 ml of filtrate was immediately transferred into 5 ml of 60% (vol/vol) methanol in water held at –40°C in an acetonitrile–dry-ice bath, vortexed, and returned to –40°C. Tubes were weighed, and 500 μl was transferred to a fresh tube and held at –40°C. Five milliliters of 75% (vol/vol) ethanol was boiled at 95°C for 4 min and then poured into the tube with 500 μl of sample, mixed, and placed at 95°C for 3 min. Extracted samples were then put back at –40°C and evaporated in a SpeedVac (Savant SC110 with refrigerated condensation trap RT100) until dry. Care was taken to ensure that the samples were not overdried. All measurements were carried out using ion-pair reversed-phase ultrahigh-performance liquid chromatography tandem mass spectrometry (IP-RP-UHPLC-MS/MS) as described previously (26), with slight modifications. Briefly, no splitter or makeup flow was used, and the chromatographic separation was performed on a narrower column (1 by 100 mm) with a lower eluant flow (0.075 ml/min).

For the E. coli W14 CepR reporter assay strains, cultures were diluted 1:100 in 5 ml of LB supplemented with 50 mM MOPS and appropriate antibiotics and grown at 30°C and 230 rpm to an OD₆₀₀ of ∼0.3. Cultures were provided 0.5 mM PAA and 0.01% rhamnose to induce the expression of paaK to convert PAA to PAA-CoA. Cultures were incubated again until an OD₆₀₀ of 0.6 and diluted back to an OD₆₀₀ of 0.3 before incubation for 1.5 h to simulate the conditions of the reporter assay. Samples were immediately prepared using the differential method described above.

Slow-killing assays in Caenorhabditis elegans.

Worms were propagated and maintained on nematode growth medium (NGM) agar (2.5 g peptone/liter) seeded with a lawn of E. coli OP50 at 16°C. Slow-killing assays were performed as described previously (18). Briefly, 60-mm petri plates filled with 10 ml of a modified NGM agar (3.5 g peptone/liter) were seeded with 50 μl of a stationary-phase culture adjusted to an OD₆₀₀ of 1.7. Plates were incubated overnight at 37°C to allow formation of a bacterial lawn. Twenty to 40 hypochlorite-synchronized C. elegans DH26 nematodes at larval stage L4 were inoculated onto each plate and subsequently incubated at 25°C for the duration of the assay. Worms were scored as live or dead for a period of 7 days.

Exoprotease assays.

Cells from an overnight culture were washed with PBS twice and adjusted to an OD₆₀₀ of 1.0. Three microliters of culture was spotted on 2% skim milk agar plates or 2% skim milk agar supplemented with 2 mM phenylalanine. Plates were incubated at 37°C for 48 h. Plates were prepared in duplicate, and three biological replicates were performed. Proteolytic activity was measured as a zone of clearing surrounding the colony. Images were taken at 48 h, and the area of the zone of clearance was analyzed using ImageJ software (49). A scale was applied to the images, and a threshold analysis was used to determine the zone of clearance. The area of this zone was then calculated (minus the area of the colony).

CepI enzymatic assay.

Heterologous expression and purification of cepI were performed according to the conditions described previously (32). CepI enzyme activity was determined spectrophotometrically by measuring holo-ACP formation with dichlorophenylindophenol (DCPIP; ε = 19,100 M⁻¹ cm⁻¹) using a previously described assay (32, 50). Reaction mixtures contained 50 mM HEPES (pH 7.5), 0.005% Nonidet P-40, 0.13 mM DCPIP, 70 μM C₈-ACP, and 4 μM CepI. After 10 min of preincubation of the mixture, the reaction was initiated by addition of 40 μM S-adenosyl methionine. The compounds were assayed in triplicate at a concentration of 200 μM.

Luminescence assays with a CepR reporter strain.

The reporter assay was performed based on the assay described by McInnis and Blackwell (51), with slight modifications. Briefly, overnight cultures were prepared from glycerol stocks in LB with appropriate antibiotics and incubated overnight at 30°C with agitation at 230 rpm. The culture was diluted 1:100 into 5 ml of fresh LB medium supplemented with 50 mM MOPS and appropriate antibiotics. This subculture was incubated at 30°C with shaking at 230 rpm until it reached an OD₆₀₀ of 0.6. One hundred microliters of culture was plated into wells with 100 μl of the appropriate medium containing C₈-HSL, PAA, and arabinose as needed, resulting in a starting OD₆₀₀ of 0.3. For the reporter strain with pTL22 (rhamnose-inducible paaK), the culture was grown and subcultured as described above. This subculture was incubated at 30°C with shaking at 230 rpm until it reached an OD₆₀₀ of ∼0.3. At this time 0.5 mM PAA and 0.01% rhamnose were added to induce paaK expression, resulting in the conversion of PAA to PAA-CoA. The cultures were incubated at 30°C and 230 rpm until an OD₆₀₀ of 0.6. One hundred microliters of culture was plated into wells with 100 μl of appropriate medium containing C₈-HSL, PAA, and arabinose and rhamnose as needed, resulting in a starting OD₆₀₀ of 0.3. Arabinose and rhamnose were used at final concentrations of 0.2% (vol/vol) and 0.01% (vol/vol), respectively. C₈-HSL was used at a final concentration of 20 nM, which was the 50% effective concentration (EC₅₀) of C₈-HSL for the conditions of this assay as determined using a dose-response curve (see Fig. S1 in the supplemental material). PAA was tested at several concentrations (100 μM, 500 μM, 1 mM, and 5 mM), and PAA was added at 500 μM in the presence of 0.01% rhamnose to produce PAA-CoA. White 96-well microplates with clear bottoms were used, and the lids were treated with a Triton X-100–ethanol solution. Plates were incubated at 30°C with agitation in a Biotek Synergy 2 plate reader. The OD₆₀₀ and relative luminescence were measured at 30-min intervals for 4 h. Inhibition was measured at the time point with peak signal intensity (2 h for PAA and 3 h for PAA-CoA).

Reverse transcription-quantitative PCR (RT-qPCR).

Since quorum sensing is associated with cell density, the expression levels of quorum sensing genes are often measured during stationary phase (52–54). To study the expression of cepI and cepR, B. cenocepacia cells were harvested from stationary-phase cultures (OD₆₀₀ of ∼7). RNA was purified and DNase treated using a RiboPure bacteria kit (Ambion) with the DNase treatment extended to 2 h. RNA quality was confirmed by running the sample on a 2% agarose gel. cDNA was synthesized using an iScript reverse transcriptase kit (Bio-Rad) and quantified using iQ SYBR green Mastermix (Bio-Rad) performed on a CFX96 Touch real-time PCR detection system (Bio-Rad). Primer efficiency was calculated for each primer set, and efficiencies between 95% and 105% were deemed acceptable. Data were analyzed using the comparative threshold cycle (C_T) method (55, 56). Genes were normalized to the RNA polymerase sigma factor gene rpoD (BCAM0918; formerly known as sigE) as a reference gene (primers 750 and 751) (7, 57, 58).

LC-MS/MS quantification of AHLs.

For AHL extraction, 5 ml of B. cenocepacia strain culture was harvested at the same time as the samples used for RT-qPCR (see the paragraph above). Samples were extracted and analyzed as described previously (59), and 5,6,7,8-tetradeutero-4-hydroxy-2-heptylquinoline (HHQ-d₄) was used as an internal standard at a concentration of 0.2 ppm.