The Pseudomonas aeruginosa Two-Component Regulator AlgR Directly Activates rsmA Expression in a Phosphorylation-Independent Manner

Previous studies suggested that AlgU and AlgR control rsmA expression (28, 31). However, whether both AlgZ and AlgR were involved in the control of rsmA expression was not tested. We constructed and assayed an rsmA transcriptional fusion (Fig. 1A) that contains both rsmA promoters (rsmATF1-lacZ). The rsmATF1-lacZ fusion was analyzed in the wild-type strain P. aeruginosa PAO1, ΔalgR and algZ mutants, and in the corresponding mucA22 mutants. The algZ mutant has a mutation in the conserved histidine residue, which prevents AlgZ-mediated phosphorylation of AlgR (19, 32) but does not disrupt the internal algR promoter (33). As shown in Fig. 1B, there was a slight increase in rsmATF1-lacZ activity in a ΔalgR mutant. When the rsmATF1-lacZ fusion was assayed in an algZ mutant strain, there was also a slight increase in reporter activity (Fig. 1B). The modest increase in reporter activity suggests that AlgZ and AlgR have a minor role in rsmA regulation in strain PAO1.

As previous studies also indicated a role for AlgU in regulating rsmA expression, the rsmATF1-lacZ transcriptional fusion was assayed in a mucA22 strain, where AlgU is most active. Reporter activity was increased ∼3-fold in a mucA22 strain, as previously described (31) (Fig. 1B). However, in the mucA22 ΔalgR double mutant, there was a drastic decrease (∼3-fold) in rsmATF1-lacZ activity (Fig. 1B). The mucA22 algZ mutant strain had rsmATF1-lacZ activity almost identical to that of the mucA22 mutant strain (Fig. 1B). These data implicate AlgR but not AlgZ in rsmA regulation in a mucA22 mutant strain and suggest that AlgR phosphorylation is not necessary for rsmA activation.

To confirm that AlgR affected rsmA expression, an epitope-tagged rsmA allele was introduced into the wild-type strain (PAO1), mucA22 mutant, and respective algR mutant strains and analyzed by Western blot analysis. As shown in Fig. 1C, a slight decrease in RsmA levels in the ΔalgR mutant strain was detected compared to the wild-type strain PAO1. As reported previously (31), a mucA22 mutant strain had drastically increased RsmA levels (Fig. 1C). The mucA22 ΔalgR mutant strain had significantly decreased RsmA compared with the mucA22 mutant strain (Fig. 1C). The Western blot analysis confirmed the transcriptional fusion analysis and supported a significant role for AlgR activating rsmA in the mucoid mucA22 mutant strain but not in the nonmucoid strain PAO1.

To further investigate the role of AlgR phosphorylation in the regulation of rsmA, algR site-directed mutants were constructed that mimic either the unphosphorylated (D54N) or phosphorylated form of AlgR (D54E). The algRD54N and algRD54E mutant alleles were also constructed in the mucA22 mutant background to confirm a phosphorylation-independent mode of AlgR activation of rsmA. Both the mucA22 algRD54N and the mucA22 algRD54E mutant strains were mucoid (data not shown), supporting the notion that phosphorylation is not necessary for alginate production and that the mutant AlgR proteins produced in these strains were still functional. The mucA22D54N mutant had increased rsmA reporter activity compared to a mucA22 mutant strain (Fig. 2). The mucA22 algRD54E strain had decreased rsmATF1-lacZ activity compared to the mucA22 mutant (Fig. 2). Because the mucA22 algRD54N and the mucA22 algZ mutant strains had elevated and similar levels, respectively, of rsmATF1-lacZ activity (Fig. 2 and 1B, respectively) compared to the mucA22 mutant strain, this suggests that phosphorylation of AlgR is not required for rsmA activation.

Previous work determined that rsmA has two promoters (31). An RNase protection assay was performed using a probe spanning the upstream region of rsmA (Fig. 3A) to determine which rsmA message AlgR affected. As we previously reported (31), two rsmA messages were seen in both the wild-type strain PAO1 and the mucA22 mutant strain, with the longer transcript increased in the mucA22 mutant strain (Fig. 3B). There was little difference in the rsmA transcripts between PAO1 and the ΔalgR mutant (Fig. 3B). In contrast, there was a substantial decrease in the longer rsmA transcript in the mucA22 ΔalgR mutant strain compared to the mucA22 parent strain (Fig. 3B). The specificity of the rsmA probe was confirmed using the probe incubated with yeast tRNA (+RNase) and using a ΔrsmA mutant (Fig. 3B). These results suggest that AlgR is most important for rsmA control in the mucA22 background and that AlgR controls the distal rsmA promoter that is also under control by AlgU.

To confirm that AlgR regulates the distal rsmA promoter, transcriptional fusions containing individual rsmA promoters were used (Fig. 3A; see also Fig. S1A in the supplemental material). The transcriptional fusion rsmATF2-lacZ contains only the proximal rsmA promoter (Fig. S1A). The transcriptional fusion rsmATF3-lacZ (Fig. 3A) contains only the distal promoter that has been shown to be controlled by AlgU (31). The deletion of algR had no effect in the rsmATF2-lacZ transcriptional fusion in the wild-type or a mucA22 background (Fig. S1B). The transcriptional fusion rsmATF3-lacZ had increased reporter activity in a mucA22 mutant strain compared to that in PAO1 (Fig. 3C). Compared to PAO1, no change in rsmTF3-lacZ activity was seen in a ΔalgR mutant strain (Fig. 3C). However, algR deletion in a mucA22 mutant strain (mucA22 ΔalgR) resulted in a significant decrease in rsmATF3-lacZ activity (Fig. 3C), supporting the RNase protection assay. Taking these results together, we conclude that AlgR activates the distal rsmA promoter, and this further suggests that AlgU and AlgR both are required to activate rsmA transcription from the distal promoter.

We hypothesized that AlgR directly activated the distal rsmA promoter. In support of this idea, two potential AlgR-binding sites are located upstream of rsmA. One of the potential AlgR-binding sites is 2 nucleotides away from the AlgU −35 consensus promoter (Fig. 4A). Purified AlgR was tested in gel shift studies with biotinylated rsmA promoter fragments (Fig. 4A). A biotinylated pscF gene, a component of the T3SS (34, 35), was used as a negative control and did not shift when incubated with purified AlgR (Fig. 4B to D). A PCR fragment representing the entire rsmA upstream region containing both promoters was dramatically shifted using purified AlgR (Fig. 4B), suggesting that AlgR directly binds to at least one of the AlgR-binding sites upstream of rsmA. To further define the region upstream of rsmA that was bound by AlgR, PCR amplicons that corresponded to the transcriptional fusion constructs containing individual rsmA promoters (Fig. 4A) were biotinylated and tested via gel shift analysis. A biotinylated probe containing the proximal promoter (TF2) and upstream sequence to the more distal transcriptional start site did not shift when incubated with purified AlgR (Fig. 4B). This result suggested that AlgR does not bind the proximal rsmA promoter. As shown in Fig. 4B, AlgR bound the distal promoter sequence fragment, TF3, containing the AlgU-dependent promoter and potential AlgR-binding sites. The TF3 fragment containing the AlgU promoter was tested in a competition assay using increasing concentrations of AlgR and unlabeled probe. A 1:1 ratio of labeled to unlabeled probe was capable of decreasing the shift of the TF3 probe (Fig. 4C). Competition with the unlabeled specific probe in a 1:5 or 1:10 ratio was able to compete for AlgR binding, as indicated by severe reduction of the probe shift (Fig. 4C). These data suggest that AlgR specifically binds the AlgU-dependent promoter region and strongly suggests that AlgR binds at least one of the AlgR consensus sequences located in this promoter region.

AlgR has previously been shown to bind an 11-bp consensus sequence, with nucleotides 2 to 10 being the most conserved (36). To determine if one, or both, of the putative AlgR-binding sites were required for AlgR binding, we used annealed primers upstream of the distal promoter in gel shift analyses. Approximately 30-bp primers were annealed and biotinylated that contained the predicted AlgR-binding sites. AlgR was able to shift the annealed biotinylated primers containing the two predicted AlgR-binding sites (Fig. 4D). The first putative AlgR-binding site, CCTTTTGTC, was mutated, and this mutation resulted in a loss of AlgR binding (Fig. 4D), suggesting that the first AlgR-binding site is required for AlgR binding to the rsmA promoter. The second putative AlgR-binding site, CCGTTTGGC, is located 7 bp downstream and directly adjacent to the AlgU −35 consensus sequence. When the second AlgR-binding site was mutated, AlgR was no longer able to bind to the annealed primers (Fig. 4D). These data suggest that AlgR binds both of the consensus sequences in order to activate rsmA expression.

A previous study indicated that rsmY and rsmZ expression was also increased in a mucA mutant background and required AlgR (28). The transcriptional start site of rsmY was determined in Pseudomonas fluorescens using 5′ rapid amplification of cDNA ends (RACE), and this information was used to deduce the transcriptional start site in P. aeruginosa (30, 37). To confirm the P. aeruginosa rsmY transcriptional start site, we performed primer extension. As indicated in Fig. 5A, a single extension product was obtained after reverse transcription of total RNA that confirmed the rsmY transcriptional start site assigned using P. fluorescens.

A transcriptional fusion, rsmYTF1-lacZ (Fig. 5B), was constructed and assayed in algR and algZ mutants in both the PAO1 (wild-type) and the mucA22 background. When a ΔalgR or algZ mutant was tested for rsmY reporter activity, there was no decrease compared to PAO1 (Fig. 5C). There was a >3-fold increase in rsmYTF1-lacZ activity in a mucA22 mutant strain compared to PAO1 (Fig. 5C). When tested in a mucA22 ΔalgR mutant strain, the increased activity in the mucA22 mutant strain was reduced to PAO1 levels (Fig. 5C). In the case of the mucA22 algZ mutant strain, there was a slight but statistically significant decrease in rsmYTF1-lacZ activity (Fig. 5C). Overall, the fusion results suggest that AlgR affects rsmY expression, but only in the mucA22 background.

To confirm that the rsmY levels decreased, Northern blotting was performed on the same strains. Two bands hybridized to the rsmY probe, as has been seen previously (Fig. 5D) (30, 38). Little difference was seen between a ΔalgR mutant strain and the wild-type PAO1 (data not shown). However, an increase in both hybridizing RNAs was seen in a mucA22 mutant strain (Fig. 5D). The mucA22 ΔalgR and mucA22 algZ mutant strains had decreased hybridizing fragments (Fig. 5D). The loading control used (5S rRNA) was consistently uneven upon repeated attempts, but the Northern blotting does confirm the transcriptional fusion analysis. A second probe using proC was also used and gave similar results. A ΔrsmY strain had no hybridizing RNA, confirming the specificity of the rsmY probe. Overall, these results support AlgR control of rsmY expression in the mucA22 strain.

Both rsmY and rsmZ contain upstream activating sequences (UAS) that GacA controls (39). A potential AlgR-binding site is located 23 nucleotides upstream of the UAS for rsmY. An additional AlgR-binding site is present further upstream on the opposite strand (Fig. 5B). Two additional transcriptional fusions, rsmYTF2-lacZ and rsmYTF3-lacZ (Fig. 5B and S2A), were constructed to determine the sequences required for AlgR control. The rsmYTF2-lacZ transcriptional fusion deletes the AlgR-binding sites but retains the UAS sequence (Fig. 5B). When assayed in the algR mutant strains in both PAO1 and mucA22 backgrounds, a difference was only seen in the mucA22 background (Fig. 5E). This result suggests that the potential AlgR-binding sites are not necessary for increased rsmY reporter activity.

An additional rsmY reporter, rsmYTF3-lacZ, was constructed that does not contain the UAS sequence. This fusion had very low activity in both PAO1 and mucA22 (Fig. S2B). This result suggests that the UAS is required for increased rsmY reporter activity in a mucA22 mutant strain as well as PAO1. Overall, the data suggest that AlgR and AlgU indirectly affect rsmY expression.

The regulation of rsmZ is more complex than rsmY and includes regulators other than GacA (38, 40, 41). Primer extension analysis confirmed the location of the transcriptional start site in P. aeruginosa that was predicted from P. fluorescens (Fig. 6A) (42, 43). A transcriptional fusion, rsmZTF1-lacZ, was assayed in the algZ and algR mutants (Fig. 6B). There was no difference in the rsmZ reporter in the ΔalgR or algZ mutant compared to PAO1 (Fig. 6C). There was a 5-fold increase in rsmZTF1-lacZ activity in the mucA22 mutant strain compared to PAO1 (Fig. 6C). In a mucA22 ΔalgR mutant strain, there was a significant decrease in rsmZTF1-lacZ activity. In a mucA22 algZ mutant strain, there was a significant but slight decrease in rsmZTF1-lacZ activity compared to the mucA22 mutant (Fig. 6C). In all, these results suggest that AlgR plays a major role and AlgZ plays only a minor role, if any, in terms of rsmZ expression in the mucA mutant background.

To confirm the transcriptional fusion results, we again wanted to monitor the actual RNA levels. Using Northern blot analysis, a trend was seen similar to that in the transcriptional fusions. As shown in Fig. 6D, there were two bands detected, with the larger fragment being the most intense. Previous studies utilizing Northern blotting suggested that the smaller hybridizing fragment represents the small RNA lacking the stem-loop found at the 3′ end of the small RNA (30). Therefore, the smaller hybridizing bands may represent the small RNAs lacking a transcriptional terminator. A ΔrsmZ mutant was used as a negative control. The mucA22 mutant strain had increased RsmZ compared to PAO1 (Fig. 6D). The mucA22 ΔalgR mutant strain had the most drastic decrease in RsmZ levels, and there was a slight decrease in RsmZ levels in the mucA22 algZ mutant (Fig. 6D). Altogether, these results suggest that AlgR plays a role in increasing the RsmZ small RNA in the mucA22 background.

We hypothesized that like rsmY expression, AlgR indirectly affects rsmZ expression. An additional transcriptional fusion, rsmZTF2-lacZ (Fig. S3A), was constructed and assayed in the wild-type strain PAO1 and a mucA22 mutant strain, which lacks the UAS sequence. Both strains had drastically decreased reporter activity (Fig. S3B). Interestingly, the mucA22 mutant strain had significantly decreased activity compared to PAO1 (Fig. S3B). Further support for an indirect mechanism of AlgR activation was obtained using gel shift analysis of a PCR amplicon of the rsmZ upstream region (Fig. S3A). Purified AlgR was not able to shift the rsmZ upstream region tested (Fig. S3C), suggesting AlgR does not directly activate rsmZ expression. These results suggest that the UAS is required for increased rsmZ expression, and we conclude that AlgR affects rsmZ expression indirectly.

To ascertain whether AlgR control of rsmA was significant in a biological context, the direct RsmA targets hcnA and tssA1 were used to assess RsmA activity (29). A new integrating vector was constructed that contains the lacUV5 promoter and lacks a ribosome-binding site or a start codon for lacZ. When tested in the wild-type strain PAO1, there was no activity of the vector alone.

The hcnA gene encodes part of the hydrogen cyanide synthase enzyme that is implicated in virulence (44, 45). RsmA was shown to previously negatively affect a leader fusion using the tac promoter and the ribosome-binding site region of hcnA (45). An hcnA leader fusion was constructed by annealing 33 nucleotides, including one of the predicted RsmA-binding sites, the ribosome-binding site, and the translational start codon of hcnA. As shown in Fig. 7A, the lacUV5 hcnA-lacZ fusion had activity in strain PAO1. When assayed in a mucA22 mutant strain, there was an ∼2.5-fold decrease in reporter expression (Fig. 7A), consistent with increased RsmA levels in the mucA22 mutant strain. If rsmA expression requires AlgR and AlgU, mutating algR and algU should result in increased fusion activity in the mucA22 background due to decreased rsmA expression. Both the mucA22 ΔalgR and the mucA22 ΔalgU mutant strains had at least a 3-fold increase in reporter expression compared to the mucA22 strain (Fig. 7A). The mucA22 ΔrsmA mutant strain was assayed and had a 3.6-fold increase in reporter activity compared to the mucA22 mutant (Fig. 7A). Overall, these results are consistent with the conclusion that RsmA is active in the mucA22 strain and that AlgU and AlgR are important for the decreased posttranscriptional regulation of hcnA.

Another RsmA target, tssA1, was also analyzed. The tssA1 gene encodes a portion of the T6SS and is directly regulated by RsmA (29, 46). Previous studies have used a lacUV5 tssA1 construct containing 227 bp of tssA1 upstream sequence that had extremely low activity (28, 29, 31). However, this fusion was found to contain additional transcriptional controls that resulted in low activity (data not shown). To more carefully assess RsmA activity using tssA1, we constructed a new leader fusion by annealing 31 nucleotides, including the predicted RsmA-binding site, the ribosome-binding site, and the translational start codon of tssA1. The lacUV5 tssA1-lacZ fusion had approximately 200 times the activity as the previously constructed lacUV5 tssA1-lacZ fusion containing 227 bp of upstream sequence in the wild-type strain PAO1 (Fig. S4). There was no activity when this fusion was assayed in a mucA22 mutant strain (Fig. 7B and S4). The lacUV5 tssA1-lacZ fusion confirmed our previous result (31), demonstrating increased activity in the mucA22 ΔalgU mutant strain, validating the use of this construct. When a mucA22 ΔalgR mutant was tested, the fusion was also increased from the mucA22 mutant strain to levels similar to those of the mucA22 ΔalgU mutant strain (Fig. 7B). A mucA22 ΔrsmA mutant strain also had statistically significant activity compared to that of the mucA22 mutant (Fig. 7B). These results demonstrate that AlgR and AlgU are necessary for posttranscriptional regulation of tssA1. From these data, we conclude that AlgU and AlgR are both required for increased rsmA expression in the mucA mutant background and that this leads to posttranscriptional regulation on two known RsmA targets, most likely through RsmA.

The strains used in this study are presented in Table 1. Escherichia coli strains were maintained on LB (Difco) plates or broth without or with antibiotics as appropriate. Pseudomonas aeruginosa strains were grown on Pseudomonas isolation agar (PIA), LB, or Vogel-Bonner minimal medium (64) and supplemented with the appropriate concentration of antibiotics. For E. coli, antibiotics were used at the following concentrations when appropriate: 10 μg/ml tetracycline, 15 μg/ml gentamicin, 100 μg/ml ampicillin, and 35 μg/ml kanamycin. For Pseudomonas strains, antibiotics were used at the following concentrations: 150 μg/ml gentamicin, 50 μg/ml tetracycline, and 300 μg/ml carbenicillin. For allelic exchange, sucrose was supplemented at 10% in YT (1% tryptone and 0.5% yeast extract) medium.

All PCR products were amplified from P. aeruginosa PAO1, unless otherwise noted, using Q5 polymerase (New England BioLabs). Crossover PCR (SOE'ing) (65) was used to construct deletion mutations and to clone into the suicide vector pEX18Tc or pEX18Gm (66). All cloned constructs were confirmed via sequencing. P. aeruginosa strains were conjugated with E. coli as a donor strain and the pRK2013-containing helper strain (67). Conjugations were performed overnight on LB plates at 30°C, and conjugations were plated for single-crossover mutants on the appropriate selective media. Merodiploids were grown without selection and then screened for sucrose sensitivity on YT–10% sucrose plates. Mutations were confirmed using PCR with primers containing the suffix intF and intR shown in Table 1 and sequencing of the resulting PCR fragment. Hemagglutinin (HA) tagging of proteins was accomplished using primers containing the HA tag at the 3′ end of the gene and introduced as described above using the suicide vector pEX18Gm.

The wild-type algR genomic region was amplified using Q5 (NEB) and primers algRXbaIR and algRHindIIIF and cloned into pEX18Tc. Site-directed mutagenesis was performed using the algRD54XbaIF/algRD54NR primers for the D54N allele or the algRD54EF/algRD54ER primer pair for the D54E mutation. The algZ mutant was constructed using the algZHSDMF/algZHSDMR primers. The primers were phosphorylated and used in site-directed mutagenesis, in accordance with the manufacturer's instructions, using Q5 (NEB). Constructs were analyzed by restriction enzyme analysis and sequencing. Mutant strains were constructed using homologous recombination, as described above, and were checked using PCR and the algRintF/algRintR primer pair and digestion with the appropriate restriction enzyme. Additional PCR amplicons were sequenced to confirm the mutation in each strain using the same primers. Further confirmation of mutants was done using phenotypic assays.

Upstream DNA fragments containing promoter regions were generated by using primers listed in Table 1 in conjunction with Q5 polymerase (New England BioLabs). PAO1 genomic DNA was used as the template. PCR products were cloned into pMiniT (NEB) and then subcloned into miniCTXlacZ using the restriction enzymes HindIII and BamHI, HindIII and EcoRI, or KpnI and BamHI (NEB). The rsmA transcriptional fusions have been previously described (31). To construct rsmY and rsmZ transcriptional fusions, the rsmYTFFEcoRI/rsmYTFR and rsmZTFF/rsmZTFR primer pairs, respectively, were used. PCR products were purified, cut with restriction enzymes, and inserted into the EcoRI and BamHI sites of miniCTXlacZ using T4 DNA ligase (NEB). The leader fusion vector was constructed by annealing the lacUV5NotI/lacUV5BamHI primer pair together and cloning into the NotI/BamHI site of CTXCP (31). Translational/leader fusions were constructed using the tssA1annealF/R and hcnAannealF/R primer pairs (Table 1) and were cloned into the ScaI/BamHI site of the leader fusion vector. Fusion constructs were confirmed by sequencing and conjugated into P. aeruginosa strains by triparental conjugation. Strains were selected for tetracycline resistance and then conjugated with pFLP2 to remove vector sequences (66). Strains were selected for carbenicillin resistance, grown overnight without selection, and plated on YT medium with 10% sucrose to select for the loss of pFLP2. Individual colonies were patch-plated onto VBMM CB300 and PIA to ensure the loss of pFLP2. To confirm the presence of the fusion constructs, PCR was performed using the forward primer used to construct the fusion and the reverse primer lacZRforTF (Table 1). β-Galactosidase activity was determined by incubating cell extracts with o-nitrophenyl-β-d-galactopyranoside (ONPG) (4 mg/ml), as described by Miller (68). A strain carrying the empty vector miniCTXlacZ was also conjugated into PAO1 and assayed, and this background (28 Miller units) was subtracted from all transcriptional fusions. The translational/leader fusion backbone CTXCPlacUV5 had no background activity. All mucoid strains were confirmed mucoid at the end of each experiment by plating on PIA plates to ensure all colonies were mucoid. Three biological replicates were reproduced for all assays.

The algR gene was PCR amplified using Q5 (NEB) and PAO1 chromosomal DNA using oligonucleotides algRSal1F and algRNot1R (Table 1). A 754-bp SalI/NotI fragment was cloned into pGEX-4T-3 (Novagen) to be expressed as a glutathione S-transferase–AlgR (GST-AlgR) fusion protein. The resulting plasmid (pGEX-4TAlgR) was transformed into E. coli BL21(DE3) (NEB) cells and incubated overnight. The colonies from this transformation were collected, inoculated into LB supplemented with 100 μg/ml ampicillin, and grown to an optical density at 600 nm (OD₆₀₀) of 0.6; 0.2 mM isopropyl-β-d-galactopyranoside (IPTG) was added to induce AlgR expression for 4 h at 15°C. The cells were collected by centrifugation (6,740 × g, 15 min), washed once in 20 ml of phosphate-buffered saline (PBS), and resuspended in 1 ml of PBS containing protease inhibitors (Thermo Fisher). AlgR was purified from this supernatant using the GST spin purification kit (Thermo Fisher). After binding of the fusion protein and washing the column, AlgR was cleaved away from GST with 10 U of thrombin/column overnight at 22°C. Purified protein was dialyzed using a Slide-A-Lyzer (Thermo Fisher) and storage buffer (20% glycerol, 20 mM Tris [pH 7.5], 5 mM MgCl₂, and 1 mM dithiothreitol [DTT]) overnight at 22°C. The identity of the purified protein was determined by mass spectrometry. The purity of AlgR was visually determined in a Coomassie-stained 7.5% electrophoresis gel (SDS-PAGE).

Purified AlgR was used in electrophoretic mobility shift assays (EMSAs) using either PCR amplicons or annealed primers using the LightShift kit, in accordance with the manufacturer's instructions (Thermo Fisher Scientific). For rsmA, the rsmAPE3F/rsmAPE3R, rsmAPE1F/rsmAEcoRIR, or rsmAPE3F/rsmAEcoRIR primer pair was used to produce PCR amplicons using Taq polymerase (NEB). Amplicons were gel extracted and biotinylated using the 3′ biotinylation kit (Thermo Fisher Scientific). Annealed primers were heated to 95°C for 5 min and cooled to room temperature. The annealed primers were biotinylated and incubated with purified AlgR. Purified AlgR was incubated at increasing concentrations to determine the suitable concentrations to be used. The nonspecific competitor poly(dI-dC) at 25 ng/μl was used for all gel shift reactions. DNA and protein were electrophoresed through 5% or 10% native polyacrylamide gels, transferred to a nylon membrane using a semidry apparatus (Hoefer), UV cross-linked, developed using the chemiluminescence detection kit (Thermo Fisher Scientific), and visualized using a charge-coupled-device (CCD) camera (ProteinSimple). The gel shift assays were repeated at least two times.

P. aeruginosa strains were grown in LB at 37°C for 8 h. The bacteria were collected by centrifugation, resuspended in sterile PBS, and lysed by sonication using a Branson sonifier. Total protein concentrations were determined by the Bradford protein assay (Bio-Rad, Carlsbad, CA). Cell extracts (5 to 10 μg) were separated by 15% SDS-PAGE gels and transferred to a polyvinylidene difluoride membrane (Bio-Rad). The membranes were blocked and probed using a 1:2,500 dilution of anti-HA monoclonal antibody in blocking buffer (Thermo Fisher, Pittsburgh, PA), followed by a 1:30,000 dilution of horseradish peroxidase-conjugated goat anti-rabbit antibody (69). Densitometry of Coomassie-stained SDS-PAGE gels was used to standardize the Western blots using total protein analysis, as described previously (70, 71). Detection was performed using the ECL Plus kit (Thermo Fisher) and chemiluminescence detection (ProteinSimple, Santa Clara, CA). The blots and Coomassie-stained SDS-PAGE were analyzed using ImageJ analysis. All Western blotting was repeated at least four times.

Total RNA (15 μg) from P. aeruginosa was isolated using the PureLink RNA minikit (Life Technologies, Grand Island, NY). Radiolabeled primers rsmYPE1 and rsmZPE1 (see Table 1) were used in a reverse transcription reaction using ThermoScript (Life Technologies). A temperature of 42°C was used for the extension. The same primer used for primer extension was also used in a sequencing reaction with the rsmY or rsmZ upstream sequence in pGEM-T (Promega, Madison, WI) using the Sequenase 7-deaza-deoxy-GTP (7-deaza-dGTP) kit (Affymetrix, Cleveland, OH). Both primer extensions and sequencing reactions were run on a 6% acrylamide-8 M urea gel. The gel was dried and extension products detected using autoradiography.

The rsmA, rsmY, and rsmZ probes for RNase protection assay (RPA) were synthesized using a PCR product generated with primers that incorporated a T7 promoter sequence at the 5′ end and using the MAXIscript T7 kit (Life Technologies) (Table 1). The reverse primer containing the T7 promoter sequence contained a nonhomologous sequence to discriminate between the full-length probe and protected fragments. The primers rsmASDMAlgRF and rsmARPAR were used to produce the PCR product for rsmA. The rsmYRPAT&R/rsmYRPAF primer pair for rsmY or rsmZRPAT7RPAF/rsmZRPAF primer pair for rsmZ was used to produce probes for the appropriate target. Probes were labeled by biotin-11-UTP (Life Technologies) at a ratio of 4:6 with UTP and were gel purified after in vitro transcription. Five micrograms of total RNA from each P. aeruginosa strain was precipitated with 800 pg of labeled probe and resuspended in hybridization buffer [80% formamide, 400 mM NaCl, 40 mM piperazine-N,N′-bis(2-ethanesulfonic acid) (PIPES) (pH 6.4), 1 mM EDTA], heated 95°C for 5 min, and incubated overnight at 42°C. Negative controls using the probe incubated with 10 μg of yeast RNA were also included. Digestion of unhybridized RNA was performed using RNase A at 25 μg/ml and RNase T1 at 10 U/ml. A biotinylated RNA ladder (Kerafast) was also used to determine the size of protected probe fragments. The reactions were run on a 5% acrylamide-8 M urea gel for rsmA and a 10% acrylamide-8 M urea gel for rsmY and rsmZ. RNA was transferred to a positively charged membrane (Hybond N+; Amersham Biosciences) using a semidry blotting system and 0.5× Tris-borate-EDTA (TBE; Hoefer, Holliston, MA). Nucleic acid was UV cross-linked and probes detected using the chemiluminescent nucleic acid detection module (Thermo Scientific). After washing, the blot was incubated for 5 min and developed using a FluorChem M (ProteinSimple). RNase protection assays were performed at least three times.

Total RNA was extracted from strains using the PureLink RNA minikit (Thermo Fisher). RNA was quantitated using a NanoDrop (Thermo Fisher). Two to 5 μg of total RNA was electrophoresed through a 10% acrylamide-8 M urea gel using formaldehyde loading buffer. After electrophoresis, RNA was transferred using a semidry blotter (Hoefer, Holliston, MA) at 300 mAmp for 45 min and the membranes cross-linked using UV light (Bio-Rad). Probes produced by in vitro transcription were biotin labeled and hybridized to the membrane using ULTRAhyb buffer (Thermo Fisher) at 65°C overnight. The primers used to make probes for rsmY and rsmZ are listed in Table 2. For rsmY, the rsmYRPAF/rsmYT7RPAR primer pair was used. For rsmZ, the rsmZRPAF/rsmZRPART7 primer pair was used. Membranes were washed with two 5-min low-stringency washes at room temperature and 2 high-stringency washes for 15 min at 65°C. Blots were developed using the chemiluminescence detection kit (Thermo Fisher), in accordance with the manufacturer's instructions. For normalization, 5S RNA probes (Table 2) made by in vitro transcription were used using primers 5SrRNAI and 5SrRNA2T7. Northern blotting was repeated more than three times.

Statistical analyses were performed using Prism 6.0 (GraphPad software, La Jolla, CA). Transcriptional and translational fusions and densitometry analysis were compared using a one-way analysis of variance (ANOVA) with Tukey's posttest.