Transcriptional Responses of Pseudomonas aeruginosa to Inhibition of Lipoprotein Transport by a Small Molecule Inhibitor

A key set of lipoprotein transport components, LolCDE, were inhibited by both a small molecule as well as genetic downregulation of their expression. The data show a unique signature in the Pseudomonas aeruginosa transcriptome in response to perturbation of outer membrane biogenesis. In addition, we demonstrate a transcriptional response in key genes with marked specificity compared to several antibiotic classes with different mechanisms of action. As a result of this work, we identified genes that could be of potential use as biomarkers in a cell-based screen for novel antibiotic inhibitors of lipoprotein transport in P. aeruginosa.

In Gram-negative bacteria, a substantial fraction of the lipoproteins are bound to the OM. A dedicated lipoprotein localization machinery directs their targeting to the OM, which includes extraction from the inner cytoplasmic membrane (IM), transport across the periplasm, and incorporation into the OM in a functional form (2,12). In gammaproteobacteria, the lipoprotein transport pathway consists of a LolCDE ATP-binding cassette transporter responsible for the recognition and release of the OM-targeted lipoproteins from the IM and directing them into a complex with the periplasmic molecular chaperone LolA. In turn, LolA transports the nascent lipoprotein across the periplasm (3,4,13). The final step in lipoprotein biogenesis is their transfer from LolA into the OM; this process is facilitated by the OM lipoprotein LolB (14,15). Recent evidence has emerged suggesting that, at least in Escherichia coli, an alternative pathway utilizes LolCDE but not LolA and LolB (16).
Previously, we described the transcriptional responses of Escherichia coli to inhibition of lipoprotein transport to the OM (17) by a novel, small molecule inhibitor referred to as compound 2 (18). This inhibitor interacts with E. coli LolCDE, and amino acid substitution mutants displaying high-level resistance are located in LolC or LolE. In transcriptome sequencing (RNA-seq) transcription studies in E. coli, the primary responses to the inhibition of LolCDE were in the CpxA/R, S , and regulator of capsular synthesis (RCS) envelope stress response systems (17,(19)(20)(21). Neither the RpoE ( E )-, RpoH ( 32 )-, nor BaeSR-controlled genes were upregulated by compound 2 inhibition (22).
In a study of lipoprotein transport in Pseudomonas aeruginosa, we found that the LolCDE system from E. coli could replace the native Pseudomonas LolCDE (23). Placing the E. coli lolCDE genes in the P. aeruginosa chromosome at the ctx phage site and subsequently deleting the native lolCDE genes resulted in cells that showed growth kinetics comparable to those of the wild type, had normal cell morphology, and were found to localize several tested proteins to the correct membrane compartments (23). In addition, the E. coli lolCDE genes inserted into the P. aeruginosa ctx site were under the control of the arabinose promoter, and this strain (P. aeruginosa PAO1 ΔmexAB-oprM ΔlolCDE PAO1 ctx::lolCDE E.coli ) was arabinose dependent for growth. The successful substitution of the E. coli lolCDE genes for the P. aeruginosa orthologues was somewhat unexpected as previous studies suggested that these two bacteria employ different sorting signals for lipoproteins destined for the IM or OM, and these were recognized by the LolCDE complex (24,25).
In the case of wild-type P. aeruginosa, neither compound 2 nor a more potent derivative, compound 2A (26) (see Fig. S1 in the supplemental material), was lethal. This observation was not unexpected as there are important differences between E. coli and P. aeruginosa in their respective LolCDE amino acid sequences, and key differences are located at the sites of some of the LolCDE compound 2-resistant E. coli mutants. When the above-described LolCDE replacement strain was used, the P. aeruginosa strain became susceptible to the inhibitor due to the reliance on the compound-sensitive E. coli lolCDE genes for viability.
The observations that the E. coli LolCDE system could support P. aeruginosa growth and normal lipoprotein transport, along with the ability of compound 2A to inhibit the replacement E. coli system, suggested that it would be possible to define the P. aeruginosa transcriptional responses to inhibition of Lol-mediated lipoprotein transport by compound 2A, employing the strain expressing the susceptible E. coli LolCDE. In this report, we describe the effect of compound 2A on the P. aeruginosa transcriptome, where we observe a transcriptional response that is different from that in E. coli treated with the same compound. The expression of a set of genes in this P. aeruginosa strain expressing heterologous lolCDE was the same as that in a strain where the levels of its native LolCDE were reduced by limiting the expression of the corresponding genes. In addition, we found several genes that appear to specifically respond to compound 2A inhibition of LolCDE while remaining unperturbed by several antibiotics with differing mechanisms of action.

RESULTS
Transcriptional responses of Pseudomonas aeruginosa to LolCDE inhibition by a small molecule inhibitor. P. aeruginosa is not naturally susceptible to either compound 2 or 2A due to overall modest amino acid sequence identity with LolC and LolE from E. coli (LolC, 38.9% identity; LolE, 35.6% identity), including key residues found at the sites of compound 2-resistant E. coli mutants (17,18). However, the strain described previously where the wild-type E. coli lolCDE genes can function in P. aeruginosa lipoprotein transport becomes susceptible to compounds 2 and 2A (23). Compound 2A is similar to compound 2, and it is more potent; however, it is a substrate for the MexAB-OprM efflux pump. Consequently, a ΔmexAB-oprM strain was employed in this study. The MIC of this P. aeruginosa strain with the E. coli lolCDE genes for compound 2A is 16 g/ml. Figure 1 shows the growth curves of the P. aeruginosa PAO1 ΔmexAB-oprM ΔlolCDE PAO1 ctx::lolCDE E.coli strain at 3ϫ (48 g/ml) and 6ϫ (96 g/ml) the MIC of compound 2A along with the untreated control. Two duplicate cultures were used, and the time of sampling for RNA isolation after compound addition was at 45 min, approximately 1 doubling at an optical density (OD) of 0.5, as indicated in Fig. 1. Following extended incubation, the cultures exposed to the compound underwent cell death and lysis at both 3ϫ and 6ϫ MIC.
Following RNA extraction with hot acid phenol, DNase treatment, and rRNA depletion, the RNA was converted to cDNA and sequenced on an Illumina NextSeq 500 platform. The number of reads for the individual samples ranged from 23,387,502 to 27,335,904 (see Table S1 in the supplemental material). Tables 1 and 2 show genes significantly upregulated or downregulated compared to the control at the two concentrations of compound 2A exposure. Table 1 lists the transcripts that were upregulated by 10-fold or more with at least one of the two compound exposure concentrations. Most prominently upregulated are numerous genes associated with alginate biosynthesis, osmotic regulation, and lipotoxin F and a number of transcripts annotated as encoding hypothetical proteins. There were fewer transcripts downregulated by 3-fold or more in Table 2, and the genes encoding the components of flagella and type IV pili are well represented in this class. The complete sets of RNA-seq transcriptional data are available in Tables S2 and S3 in the supplemental material.
Examining the data (Table 3) for the known two-component regulator systems of virulence and antibiotic responses in P. aeruginosa, it was possible to identify changes in the regulation of systems associated with alginate (upregulated) and type IV pili and flagella (both downregulated). These are very consistent with the other transcript expression changes in genes associated with these regulators upon compound 2A addition.
Ϫ1. Alginate production in response to compound 2A. A notable finding in the RNA-seq analysis was an increase, in response to compound 2A, in the level of transcripts from the cluster of genes responsible for the biosynthesis, modification, and export of alginate. Because treatment with compound 2A is lethal, the alginate product could not be detected by the growth of mucoid colonies. Antibodies to alginate were available, offering the possibility of examining alginate production (Fig. 2). P. aeruginosa PAO1 ΔmexAB-oprM ΔlolCDE Pa ctx::lolCDE E.coli cells were treated for 1 or 2 h with compound 2A, and normalized lysates were prepared and analyzed by dot blot immunoassays with antialginate as well as antibodies against the outer membrane protein OprF. The results clearly show that alginate production was increased in response to treatment of the bacteria with compound 2A. A matching control using a strain with the algU gene deleted did not exhibit increased levels of alginate (Fig. 2). These results indicate that inhibition of lipoprotein transport by compound 2A affects alginate production via a regulatory mechanism acting through the AlgU-mediated control of the alginate biosynthetic operon.
Specificity of the response to compound 2A. As mentioned above, the native P. aeruginosa LolCDE is refractory to the effects of compound 2A due to key amino acid differences, and it is only by substituting E. coli LolCDE that P. aeruginosa becomes susceptible to this inhibitor of lipoprotein transport. Therefore, by comparing P. aeruginosa responses to compound 2A with the two LolCDE versions, it can be established that the responses observed were through susceptible LolCDE inhibition. Again, the transcripts of genes exhibiting the most responses to compound 2A were tested using reverse transcription-quantitative PCR (RT-qPCR) after exposure to 1ϫ MIC of compound 2A. Figure 3 illustrates the results for 3 different time intervals. It is clear that only in the case of P. aeruginosa cells with the susceptible E. coli version of LolCDE did we observe changes in transcript levels. This indicates that the transcriptional changes are a direct consequence of the effect of compound 2A on the susceptible E. coli LolCDE. Change magnitudes mostly peaked at 1 h of compound treatment, diminishing with prolonged exposure time due to subsequent growth impairment and viability loss.
As an additional check, RT-qPCR analysis of selected genes was performed in P. aeruginosa in which the native lolCDE operon of P. aeruginosa was placed under the control of the arabinose promoter. By growing the bacteria in Luria-Bertani (LB) medium supplemented with 0.2% arabinose and then reducing (0.05%) or removing the arabinose, LolCDE levels would diminish over time. The expectation would be that similar changes in expression would be observed in response to compound 2A inhibition of LolCDE function. Levels of RNA were measured by RT-qPCR for several of the identified genes that exhibited significant expression changes upon treatment with compound 2A. As shown in Fig. 4 and Fig. S2, the expression changes of these genes confirmed the results seen in cells treated with compound 2A; namely, the loss of LolCDE function leads to a conserved transcriptional response. Similarly, the levels of alginate measured by dot blot immunoassays were found to increase substantially with the removal of the arabinose inducer and subsequent growth dilution of the native P. aeruginosa LolCDE (Fig. 5).
Comparison of the effects of compound 2A and antibiotics. Several of the key genes whose transcript levels were significantly altered by the exposure of PAO1 ΔmexAB-oprM ΔlolCDE PAO1 ctx::lolCDE E.coli to compound 2A were reconfirmed using RT-qPCR assays. At the same time, we also wished to determine the specificity of the response in relation to several antibiotics with different, known mechanisms of action. MICs were measured in LB medium (used for all experiments) and are presented in duplicate experiments on different days are indicated by vertical bars. Panels A through I show genes that were upregulated more than 5-fold, and panels J and K below the line, show genes identified as being downregulated. Changes tended to peak at around 1 h and subsequently declined, most likely due to decreased viability from compound 2A. In all cases, the no effects of compound 2A were observed with native P. aeruginosa lolCDE-containing cells. Table S4. The P. aeruginosa strain with ΔmexAB-oprM and the E. coli lolCDE genes in place of its native lolCDE was grown for two independent sets of experiment. The cultures were treated for 45 min at an OD at 600 nm (OD 600 ) of 0.5 with 3ϫ MIC of compound 2A or with 3ϫ MIC of the indicated antibiotics. The RNAs were extracted and prepared for RT-qPCR. Primers were designed to amplify key genes whose expressions were perturbed by compound 2A. The proC gene was employed as a housekeeping gene, whose level of transcription should not vary significantly from that for the control cells, to determine the efficiency and variability of the RNA extractions in the different experiments using cells treated with the LolCDE inhibitor or with various antibiotics. Figure 6 presents the average results from the two independent RT-qPCR experiments for several key upregulated genes. As far as responses to all the inhibitors, algD, osmC, PA2171 (hypothetical unknown), and PA3404 (probable OM protein precursor) gave the clearest responses to induction exclusively by compound 2A, and these represent biomarkers of inhibition of lipoprotein transport. In the cases of PA2146, PA2167, PA1471, PA1323, PA2562, PA2414, PA2176, PA5212 (all encoding hypothetical unknown proteins), and the annotated cpo and lptF genes, while the transcriptional responses were strongly induced by compound 2A, additional responses were observed in cells exposed to meropenem, polymyxin B, or both antibiotics. Open reading frame (ORF) PA0737 (hypothetical unknown) had equivalent responses to compound 2A and polymyxin B.
A less specific response was observed among genes downregulated in the RNA-seq experiments. When a selected group of these transcripts was quantified by RT-qPCR studies, compound 2A was the most potent in reducing their cellular levels; however, several of the other antibiotics had similar effects (Fig. 7). The most notable exception was ciprofloxacin and, to a lesser extent, chloramphenicol. In these cases, the responses to ciprofloxacin matched the untreated control values, while treatment with chloramphenicol showed xphA and fleS transcripts to be at nearly control levels, and flgF and pilQ mRNA concentrations were reduced.

DISCUSSION
Our previous work (23) substituting the E. coli lolCDE operon for the native P. aeruginosa version laid the groundwork for the present study of transcriptional responses to LolCDE inhibition by compound 2A. By employing RNA-seq, a number of substantive changes were found to occur in gene expression that could be attributed to the activity of compound 2A on the LolCDE complex; the response very likely reflects cellular sensing of a block in lipoprotein transport and disruption of OM biogenesis. Compared to a similar study in E. coli (17), we find little commonality in the transcriptomes of compound 2-treated cells. Some may arise from the absence of certain genes in each organism; however, where orthologues exist, there was no apparent accumulation or reduction in these mRNAs, indicating that bacteria adapt to specific niches by unique responses.
Most prominent among the genes whose transcript levels were increased in response to compound 2A were those in the alginate pathway. The regulation of alginate synthesis in P. aeruginosa is complex (27). The mucA gene forms part of the algU-mucA-mucB-mucC operon, which is homologous to the rpoE-rseA-rseB-rseC operon in E. coli (28). The alg genes are under the control of the stress-related extracytoplasmic function sigma factor (ECF) AlgU; however, it is inactive due to its sequestration by the anti-sigma factor MucA. alg operon expression is activated by the degradation of MucA by the AlgW protease or through the acquisition of mutations in MucA, frequently encountered among isolates from   Table S4 in the supplemental material.
individuals with the genetic disease cystic fibrosis (29). Free AlgU directs the expression of alg genes by binding to the algD promoter, the first gene of the operon (30,31). Additional regulatory elements control alginate expression, including genes for the two-component regulators AlgR/FimS and AlgB/KinB (32-37). These genes indirectly regulate the expression of algU through algD. Control may also be exerted at both MucA degradation/AlgU Journal of Bacteriology sequestration as well as the induction of the above-described regulators. Since the different alginate regulatory inputs converge on AlgU, and the effect of compound 2A was AlgU dependent (Fig. 2), the activation of any one of the stress-responsive regulators by a defect in lipoprotein transport could lead to the observed increase in alginate production. The transcription of genes associated with osmotic stress (e.g., osmC and osmE) increased as well. Both of these genes encode lipoproteins, and their transcription may be increasing due to both osmotic stress and their failure to be transported to their outer membrane locations (38). The lipotoxin F gene lptF, also markedly upregulated, is also controlled by the AlgU/T system (39). The chloroperoxidase gene has also been observed to be upregulated by cell wall stress (30). Interestingly, several of the hypothetical genes (PA2167 to PA2173) that showed an upregulated response clustered together, suggesting that an operon response to cell envelope stress may be located in this region. Further work would clarify this possibility.
In addition to the alginate pathway, a number of other transcripts were affected by treatment with compound 2A, including those encoding determinants of flagellum and pilus formation, both of which were substantially downregulated. These effects may be indirectly attributable to compound 2A-induced membrane stress resulting in the downregulation of additional two-component systems, namely, fleR-fleS and pilG-pilH, resulting in negative effects on the expression of motility and pilus genes, respectively (40,41). Moreover, the assembly of the type IV pilus utilizes one outer membrane  (42,43). A lack of their transport and, consequently, the formation of defective organelles may be sensed by the regulatory machinery controlling the transcription of genes encoding flagellar or pilus components. This response appears to be specific for flagellum and pilus genes since compound 2A caused an opposite effect on alginate expression; the transport of this polysaccharide also requires a lipoprotein (AlgK), while other transcripts of molecular machineries, such as type II, type III, and type VI secretion systems, each containing the lipoproteins HxcQ (type II), ExsB, PscJ (type III), and TagQ (type IV), were unaffected (44)(45)(46).
Although there is strong evidence that compound 2A functions by specifically inhibiting LolCDE, we could not exclude the possibility that the observed changes in transcript levels following treatment with this compound and analyzed by RNA-seq were due to an indirect effect on bacterial physiology and not the consequence of inhibition of lipoprotein transport. After identifying genes with changes in mRNA levels above 10-fold, we confirmed the effect of compound 2A in P. aeruginosa ΔmexAB-oprM lacking native lolCDE but expressing the E. coli orthologues by RT-qPCR. We then used RT-qPCR to determine the mRNA levels in P. aeruginosa ΔmexAB-oprM expressing the native P. aeruginosa lolCDE genes under the control of the pBAD promoter and AraC. By reducing the amount of P. aeruginosa LolCDE by limiting the concentration of the inducer (arabinose) in the growth medium, we observed increases in the levels of the same transcripts as those seen in P. aeruginosa lolCDE E.coli following treatment with compound 2A. In addition, using antibodies to alginate, we have shown that alginate production was upregulated by P. aeruginosa LolCDE depletion. These data reinforce the case that the gene expression changes are an authentic response to interference with LolCDE function.
Several antibiotics with distinct mechanisms of action were tested along with compound 2A for gene expression changes upon exposure. The antibiotics were tested for a limited time (45 min) and at concentrations that minimized extensive cell inhibition and killing. The objective was to identify genes whose expression was changed largely or solely by LolCDE inhibition. These genes could then serve as reporters for the presumptive identification of novel candidate molecules or chemical modifications of compound 2A that would inhibit the native P. aeruginosa LolCDE. Promising candidates in terms of specificity were identified in the responses of algD, osmC, and ORFs PA2171 and PA3404. Several groups have considered the various steps during the biogenesis of the Gram-negative cell envelope as novel antibiotic targets (18,(47)(48)(49)(50)(51)(52). For the screening of lipoprotein transport in P. aeruginosa for a potential therapeutic inhibitor, the development of promoter-reporter constructs with one or more of these genes, employing, for example, ␤-galactosidase, fluorescent proteins, or luciferase, would serve as a first line of identification in a cell-based screening system for P. aeruginosa LolCDE inhibitors (18,49,61). It was interesting to note that antibiotics such as meropenem and polymyxin B, which gave some gene responses overlapping those of the LolCDE inhibitor, are also known to affect aspects of bacterial cell envelope biogenesis. Meropenem has an affinity for Pseudomonas penicillin-binding proteins (PBPs) 2, 3, and 4, whereas the other ␤-lactam antibiotic, cefepime, has a PBP binding profile distinctly different from that of meropenem (53,54). Disruption of the OM by polymyxins has been well documented (55).
The effects of LolCDE inhibition in both E. coli and P. aeruginosa have now been described (17,23). In both cases, the two bacteria cease growth and undergo cell lysis in response to LolCDE inhibition. In contrast to the well-defined envelope stress systems of E. coli (19)(20)(21)(22), the responses described here for P. aeruginosa are still in the process of being delineated. Clearly, one well-understood aspect is the massive upregulation of alginate biosynthesis. In addition, a number of genes of unknown function were involved in the responses observed. How these changes are connected to the disruption of cell envelope lipoprotein transport will be a topic for future exploration.

MATERIALS AND METHODS
Bacterial strains and culture conditions. P. aeruginosa and E. coli genotypes and plasmids are listed in Table 4. The bacteria were cultured for all experiments in Luria-Bertani (LB) medium at 37°C with shaking at 300 rpm. Antibiotics for genetic selection were used at the following concentrations: tetracycline (Tc) at 30 g/ml and gentamicin (Gm) at 75 g/ml for P. aeruginosa and tetracycline at 10 g/ml, ampicillin (Amp) at 100 g/ml, and gentamicin at 15 g/ml for E. coli. Compound 2A (26) (see Fig. S1 in the supplemental material) was obtained in a powder form from AChemtek (Worcester, MA), dissolved in dimethyl sulfoxide to obtain a 5-mg/ml stock, and stored at Ϫ20°C. MIC determination. The MIC values for ΔmexAB-oprM P. aeruginosa strains with either P. aeruginosa PAO1 or E. coli lolCDE genes at the ctx site against compound 2A were determined in microtiter plates (LB broth with 0.2% L-arabinose) at 5 ϫ 10 5 CFU/ml. The MIC values of gentamicin, meropenem, cefepime, polymyxin B, chloramphenicol, fosfomycin, and ciprofloxacin were determined using Etest strips (bioMérieux Inc.) on LB agar plates with 0.2% L-arabinose and an inoculum of 5 ϫ 10 5 CFU/ml. MIC values on Mueller-Hinton (MH) agar were within 1 dilution of the LB broth values. LolCDE replacement. The lolCDE genes from P. aeruginosa or E. coli were cloned into the EcoRI/SpeI sites of pSW196 under the control of the arabinose-inducible P BAD promoter. Plasmids pSW196-lolCDE PAO1 and pSW196-lolCDE E.coli were conjugated into PAO1 using triparental mating with helper plasmid pRK2103. Tetracycline-resistant transconjugants were checked for the genomic insertion of the lolCDE genes at the CTX site (56) by PCR with primers flanking the insertion site and subsequent DNA sequencing. P. aeruginosa lolCDE deletion. Following the introduction of either the P. aeruginosa or E. coli lolCDE genes into the CTX site, for the deletion of the lolCDE genes at their original genome locus, ϳ500 bp of upstream and downstream regions flanking the native Pseudomonas lolCDE genes were cloned into pEXG2 in E. coli. The resulting plasmid, pEXG2 ΔlolCDE, was conjugated into the PAO1 strains that have lolCDE from P. aeruginosa or E. coli inserted at the CTX site. Transconjugants with a deletion of the native genomic lolCDE alleles were selected on medium containing 6% sucrose and 0.5% L-arabinose. Resolved strains were tested for gentamicin sensitivity, and the deletion of the native lolCDE locus was confirmed by sequencing a PCR product using primers for the upstream and downstream genes flanking the native lolCDE operon. The lolCDE deletion strains were dependent on arabinose for viability. Deletions of mexAB-oprM and algU were done in a similar fashion.
Transcriptome analysis by RNA-seq. For transcriptome sequencing (RNA-seq), P. aeruginosa mexAB-oprM ΔlolCDE PAO1 ctx::lolCDE E.coli was grown overnight in LB broth with 0.2% arabinose with shaking at 37°C. The next morning, a 1:200 dilution was made in 100 ml of LB broth plus 0.2% arabinose, and the bacteria were grown at 37°C with shaking until they reached an optical density at 600 nm (OD 600 ) of 0.5. The culture was then split into six portions of 10 ml each that were placed into six flasks, with bacteria in two flasks being used as biological replicate controls, each of two replicate flasks receiving 48 g/ml of compound 2A, and two replicate flasks receiving 96 g/ml of the same inhibitor.
After 45 min of exposure to compound 2A, 800 l of the culture from each flask was placed directly into 800 l of a prewarmed (65°C) lysis mix-acid phenol solution. Lysis mix consisted of 320 mM sodium acetate, 8% SDS, and 16 mM EDTA (all from Ambion and Thermo Fisher) in nuclease-free water. One hundred microliters of the above-described lysis mix was combined with 700 l of acid phenolchloroform (Ambion) in 2-ml tubes. The cells and lysis mix-acid phenol were rapidly mixed on a vortex mixer and kept at 65°C with periodic vortexing for 5 to 10 s every minute for 10 min. RNA isolation, purification, concentrations, and DNase treatment were performed as previously described (17). Ribosome integrity numbers (RINs) were determined with an Agilent Bioanalyzer 2100 instrument and an Agilent RNA 6000 Nano kit. The RINs ranged from 9.8 to 10 for all samples. rRNA depletion was carried out with a RiboMinus kit (Thermo-Fisher). Library preparation for Illumina sequencing was performed as previously described (17). RNA-seq was carried out on an Illumina NextSeq 500 platform. Analysis of the data was performed using CLC Bio Genomics Workbench software, with the reads being mapped to the genome sequence of P. aeruginosa PAO1. Replicates were highly comparable in plots against each other and divergent in plots of controls versus compound treatment (Fig. S3).
RT-qPCR. For the determination of RNA levels by reverse transcription-quantitative PCR (RT-qPCR), P. aeruginosa strains were grown with the indicated concentrations of either arabinose (depletion experiments) or antibiotics for either 4 or 6 h. RNA was then prepared from these cultures using the hot acid phenol procedure as described above. Primers were designed by the use of the GenScript real-time PCR primer design tool. cDNA was synthesized with a SuperScript III first-strand synthesis system for reverse transcription-PCR (Invitrogen) and random hexamer primers. RT-qPCR was carried out using PerfeCTa SYBR green FastMix (Quanta Biosciences) in a Mastercycler Realplex2 system from Eppendorf. Changes in transcript levels relative to the levels in the untreated control cultures were calculated. Growth and compound exposures for RNA extractions for the RT-qPCR experiments were performed twice (biological replicates) on different days.
Alginate antibody dot blots. A sample of the culture corresponding to 400 l at an OD 600 of 1.0 was centrifuged at 12,000 ϫ g for 1 min. The supernatant was discarded, the cell pellet was resuspended in algU deletion construct 40 l of 2ϫ Laemmli sample buffer (Bio-Rad) and boiled for 5 min, and 2 l was pipetted onto a nitrocellulose membrane. The membrane was allowed to air dry (ϳ30 min), blocked with Tris-buffered saline-Tween 20 (TBST) with 5% skim milk for 1 h, washed three times for 10 min with TBST, and incubated with primary antibody overnight at 4°C. The next day, the filter was washed three times for 10 min each with TBST, incubated with secondary antibody (horseradish peroxidase [HRP] conjugated) for 1 h at room temperature, washed three times for 10 min with TBST, and incubated with ECL chemiluminescent solution for 5 min, and luminescence was detected with X-ray film. The alginate antibody was obtained from Sigma (monoclonal, anti-mouse), and OprF was detected by anti-rabbit antibody, raised in-house.

ACKNOWLEDGMENT
This work was supported by NIH grant R01 AI136789. S.L. is on the scientific advisory board of and holds equity in X-Biotix Therapeutics, Inc.