|Home | About | Journals | Submit | Contact Us | Français|
AlgR controls numerous virulence factors in Pseudomonas aeruginosa, including alginate, hydrogen cyanide production, and type IV pilus-mediated twitching motility. In this study, the role of AlgR in biofilms was examined in continuous-flow and static biofilm assays. Strain PSL317 (ΔalgR) produced one-third the biofilm biomass of wild-type strain PAO1. Complementation with algR, but not fimTU-pilVWXY1Y2E, restored PSL317 to the wild-type biofilm phenotype. Comparisons of the transcriptional profiles of biofilm-grown PAO1 and PSL317 revealed that a number of quorum-sensing genes were upregulated in the algR deletion strain. Measurement of rhlA::lacZ and rhlI::lacZ promoter fusions confirmed the transcriptional profiling data when PSL317 was grown as a biofilm, but not planktonically. Increased amounts of rhamnolipids and N-butyryl homoserine lactone were detected in the biofilm effluent but not the planktonic supernatants of the algR mutant. Additionally, AlgR specifically bound to the rhlA and rhlI promoters in mobility shift assays. Moreover, PAO1 containing a chromosomal mutated AlgR binding site in its rhlI promoter formed biofilms and produced increased amounts of rhamnolipids similarly to the algR deletion strain. These observations indicate that AlgR specifically represses the Rhl quorum-sensing system during biofilm growth and that such repression is necessary for normal biofilm development. These data also suggest that AlgR may control transcription in a contact-dependent or biofilm-specific manner.
The opportunistic pathogen Pseudomonas aeruginosa is the major cause of morbidity and mortality in patients with cystic fibrosis (CF) (28). The factors that enable P. aeruginosa to predominate and persist in the CF lung despite aggressive antimicrobial therapy are numerous and include alginate production (27), antimicrobial resistance mechanisms (20, 66), and secreted factors (41, 56). Furthermore, several studies suggest that P. aeruginosa persists in the CF lung as organized communities known as biofilms (14, 75). Biofilms are composed of many individual bacteria in various stages of development and contain self-generating diversity to produce insurance effects (4, 37). Bacterial biofilms are encased in an extracellular polymeric substance (40) and are intrinsically more resistant than planktonic organisms to innate immune defense mechanisms and antimicrobial therapy (8, 20, 46).
To date, three exopolysaccharides associated with P. aeruginosa biofilms, alginate (12), the product of psl genes (85), and the product of pel genes (25), have been identified. Biofilms formed by mucoid P. aeruginosa contain significant amounts of alginate, and alginate production in mucoid strains influences biofilm architecture (29, 54). However, others have shown that alginate is not the predominant polysaccharide present in nonmucoid P. aeruginosa biofilms cultured in vitro (85) and is not required for biofilm development (76). Evidence from the existing literature indicates that alginate is most likely an exopolysaccharide produced under stress by P. aeruginosa (5, 79, 84). The conversion of nonmucoid P. aeruginosa to the alginate-overproducing mucoid phenotype is a critical step in the pathogenesis of CF disease and coincides with a worsening prognosis for the CF patient (28). Thus, the activation of the alginate biosynthetic pathway and biofilm development in P. aeruginosa both represent a critical juncture in CF pathology.
One of the molecular mechanisms for the constitutive expression of the exopolysaccharide alginate has been discovered and involves the alternative sigma factor, AlgU (47, 48) (also known as AlgT ). Upon activation through mutations acquired in mucA (48), alginate is produced in copious amounts by transcriptional activation of the regulatory protein AlgR and its subsequent upregulation of the 12 alginate biosynthetic genes (algD through algA) (28, 49, 68-70). The transcriptional regulator AlgR is required for algD transcription by binding to three sites within the algD promoter (RB1, RB2, and RB3) (51-53). Mucoid P. aeruginosa strains in which algR is disrupted are no longer able to produce alginate (18).
AlgR has been shown to regulate several other P. aeruginosa processes, including hydrogen cyanide (HCN) production (7) and twitching motility (44, 82, 83), suggesting a more global role for AlgR in P. aeruginosa pathogenesis. In support of this, AlgR is required for full virulence in both the acute septicemia and pneumonia murine infection models (43). However, the genes involved in the global affects observed in the virulence studies have not yet been identified.
In this study, the AlgR regulon of the nonmucoid laboratory strain PAO1 was examined during biofilm growth using flow chamber and static biofilm technology. These findings expand the role of AlgR as a regulator of virulence in P. aeruginosa by demonstrating that AlgR directly represses the Rhl quorum-sensing circuit in a biofilm-specific manner. Furthermore, these findings support the hypothesis that AlgR may utilize contact-dependent or biofilm-specific mechanisms of gene regulation that may account for its differential regulation of alginate production, twitching motility, and biofilm maturation.
The bacterial strains and plasmids used in this study are listed in Table Table1.1. Plasmid CTXlacZ490ScaI was constructed using oligonucleotides lacZhinD and CTXXho (Table (Table2)2) to amplify the first 490 nucleotides of lacZ from plasmid pRS415 (74). The PCR product was digested with HinDIII and XhoI and ligated into plasmid mini-CTX-1 (34) digested with the same restriction enzymes. The resulting plasmid, CTXlacZ490, was subjected to site-directed mutagenesis using the Stratagene Quick Change II protocol with oligonucleotides lacZScaF’ and lacZScaR’ (Table (Table2)2) to introduce an in-frame ScaI restriction endonuclease site into the lacZ open reading frame to create translational fusions. Plasmid pCR2.1 rhlI was constructed by ligation of the PCR product of oligonucleotides rhlIgsF and rhlIgsR (Table (Table2)2) into Invitrogen's pCR2.1 vector. Plasmid pCR2.1 rhlA was constructed by cloning the PCR product of oligonucleotides rhlAgsF and rhlAgsR into pCR2.1.
Flowthrough biofilms were grown in a one-flowthrough model using Pseudomonas putida minimal medium supplemented with glutamate (1.6 mM) as sole carbon source as described previously (67). Briefly, 3 ml of an overnight culture of the same medium used for the biofilm was inoculated into a flowthrough biofilm system with a flow rate of approximately 0.4 ml/min and grown for 6 days in minimal medium at room temperature (~26°C). The resulting biofilm was collected in RNALater (Ambion) for RNA isolation.
P. aeruginosa strains were grown in FAB (73) with 1.6 mM glucose, glutamate, or succinate medium solidified with 0.5% Noble agar. Plates were inoculated by using a sterilized platinum wire with log-phase cells (optical density at 600 nm [OD600] of 0.6) grown in the respective carbon source overnight and incubated at 30°C for 24 h. The zones of migration from the point of inoculation were measured in triplicate for each condition.
P. aeruginosa strains PAO1 and PSL317 were grown as biofilms using the flowthrough model described above. RNA was isolated using a CsCl gradient as previously described (71) and analyzed with an Agilent 2100 Bioanalyzer to determine the RNA integrity (see Fig. S1 in the supplemental material). Ten micrograms of total RNA was used for cDNA synthesis, fragmentation, and labeling according to the Affymetrix GeneChip P. aeruginosa genome array expression protocol. Briefly, random hexamers (Invitrogen) were added to 10 μg of RNA along with in vitro-transcribed Bacillus subtilis control spikes. cDNA was synthesized using Superscript III (Invitrogen) and the following conditions: 25°C for 10 min, 37°C for 60 min, and 70°C for 10 min. RNA was removed by alkaline treatment and subsequent neutralization. The cDNA was purified by using a QIAquick PCR purification kit (QIAGEN) and eluted in 40 μl of elution buffer (QIAGEN). The cDNA was then fragmented by using 0.6 U DNase I (Amersham) per μg cDNA at 37°C for 10 min. The fragmented cDNA was end labeled with biotin-ddUTP by using a BioArray terminal labeling kit (Enzo) per the manufacturer's instructions. A gel shift mobility assay was performed using NeutrAvadin (Pierce) on a 5% polyacrylamide gel stained with SYBR green (Roche) to ensure complete fragmentation and labeling. Samples were hybridized, washed, stained, and scanned as described in the Affymetrix GeneChip P. aeruginosa genome array expression analysis protocol.
The absolute expression transcript levels were normalized for each chip by globally scaling all probe sets to a target signal intensity of 500. Three statistical algorithms (detection, change call, and signal log ratio) were used to identify differential gene expression in experimental and control samples. The decision of a present, absent, or marginal identification for each gene was determined by using MicroArray Suite software (version 5.0; Affymetrix). Those transcripts that received an “absent” designation were removed from further analysis. A t test was used to isolate those genes whose transcriptional profile was statistically significant (P < 0.05) between the control and experimental conditions. Pair-wise comparisons between the individual experimental and control chips were done by batch analyses using MicroArray Suite to generate a change call and signal log ratio for each transcript. A positive change was defined as a call whereby more than 50% of the transcripts increased or marginally increased for up-regulated genes or decreased or marginally decreased for down-regulated genes. Lastly, the median value of the signal log ratios for each comparison was calculated and only transcripts that had a value greater than or equal to 1 for up-regulated and less than or equal to 1 for down-regulated genes were placed on the final list of transcripts whose profile had changed. The signal-log ratio was converted and expressed as the change (n-fold).
P. aeruginosa PAO1, PSL317, PSL317 (pVDtacPIL), PSL317 (pVDZ'2R), and PAO1rhlImut were grown in a flowthrough biofilm as described above and imaged with the aid of an image chamber (Stovall, Inc.). An overnight culture of 3 ml of the individual strains grown in Pseudomonas minimal medium (see above) was inoculated into a flowthrough biofilm system with a flow rate of approximately 0.4 ml/min. The biofilms were grown for 1, 3, or 6 days in minimal medium supplemented with 130 mg/liter of glutamate as a carbon source. The bacteria were stained with LIVE/DEAD BacLight (Molecular Probes). Z-section images were collected on a Zeiss Axioplan II microscope (step size, 0.1 to 0.2 μm; magnification, ×630) using Slidebook 4.0 as the imaging software (Intelligent Imaging Inc., Denver, CO). Postacquisition images were processed using Volocity software (Improvision, Ltd., Lexington, MA). Quantitative analysis of the flow cell-grown biofilms was performed with the COMSTAT image analysis software package (31).
The 96-well biofilm assay was performed as previously described (23) with the following modifications. Briefly, biofilm formation was assayed by the ability of cells to adhere to the wells of 96-well microtiter plates (Becton Dickinson Labware). Overnight cultures grown in the minimal medium supplemented with glutamate (1.6 mM) used for the continuous-flow biofilm of PAO1, PSL317, PSL317 (pVDtacPIL), and PSL317 (pVDZ'2R) were diluted 1:100 in fresh minimal medium and inoculated into the 96-well plate. The plates were incubated at 25°C for 24 h to allow for biofilm formation. After 24 h, the plates were washed once in ddH2O and then a solution of 1% crystal violet was added to stain the cells. The plates were set aside for 10 min and washed three times to remove any residual crystal violet. A solution of 33% acetic acid was added to each well to lyse the bacterial cells and solubilize the crystal violet. The absorbance was determined at 580 nm in a μQuant microtiter plate reader (Biotek Instruments, Inc.). The assays were performed in triplicate with five technical replicates (wells) for each replicate.
A PCR product generated using oligonucleotides rhlAF1 and rhlAR1 (Table (Table2)2) encompassing −1864 to +9 bp relative to the translational start of rhlA was amplified with SmaI and SacI restriction sites and directionally cloned into CTXscaI-lacZ to generate a translational fusion. The final construct, CTXrhlA::lacZ, was sequenced before use in P. aeruginosa. Escherichia coli DH5α harboring the plasmid CTXrhlA::lacZ was used in triparental conjugations with either PAO1, PSL317, or PSL317 (pVDZ'2R) and the helper strain, DH5α pRK2013 (22). Conjugants were screened by PCR for each promoter and the plasmid backbone was removed by pFlp2-mediated excision (33). Integration at the attB site was confirmed by PCR using attB-specific primers and Southern blotting. PAO1 containing the rhlI::lacZ chromosomal fusions was generously provided by Daniel Hassett. The deletion of algR in PAO1 containing the chromosomal fusion genes was performed as previously described (44) and confirmed by Southern blotting.
β-Galactosidase assays were performed as described by Miller (50), with slight modification. All assays were performed on P. aeruginosa strains grown for 6 days as continuous culture biofilms in the flow cell model or to stationary phase in P. putida minimal medium broth at room temperature at 250 rpm. After the indicated incubation period, bacteria were removed by scraping and resuspended in 200 μl phosphate-buffered saline. An amount of 800 μl Z-buffer (60 mMNa2HPO4, 40 mM NaH2PO4, 10 mM KCl, 1 mM MgSO4, and 50 mM β-mercaptoethanol) was added, and the OD600 of each suspension was measured. Amounts of 20 μl chloroform and 20 μl 0.1% sodium dodecyl sulfate were added to each tube and vortexed for 10 s. An amount of 200 μl of orthonitrophenyl galactoside (4 mg/ml in H2O) was added, and the reaction was allowed to proceed for 1 to 3 min (determined empirically for each promoter). The reaction was stopped by the addition of 250 μl 1 M NaHCO4. Samples were centrifuged to pellet the cells and chloroform, and the supernatant was measured at 420 nm. P. aeruginosa strains harboring a promoterless lacZ gene in the attB site were used as negative controls in every experiment to determine background β-galactosidase activity.
The concentrations of N-3-oxododecanoyl-l-homoserine lactone (3-oxo-C12-HSL) and N-butyryl homoserine lactone (C4-HSL) were assayed in each sample in triplicate using previously described bioassays (60, 64).
The concentrations of rhamnolipids in the effluents of biofilm-grown or supernatants of broth-cultured P. aeruginosa PAO1, PSL317 (ΔalgR), or PSL317 harboring pVDZ'2R were determined by the orcinol method (78). Briefly, 500 μl of three independently grown biofilm or broth cultures was extracted twice with 1 ml of diethyl ether. The ether fractions were pooled and evaporated to dryness. One milliliter of a 0.19% orcinol (in 53% H2SO4) solution was added to each sample. The samples were heated to 80°C for 30 min and cooled at room temperature for 15 min, and the absorption was measured at 421 nm by UV spectrophotometer. The concentration of rhamnolipids was calculated by comparing the data with 0 to 50 μg/ml rhamnose standards. Standards, blanks, and unknowns were analyzed in triplicate from three independent experiments.
The binding of AlgR to the rhlA and rhlI promoter regions was examined by using recombinant AlgR, expressed as previously described (51). A 175-bp DNA fragment of the rhlA promoter (−847 to −1022 in relation to the translational start site) was excised from pCRrhlA by using EcoRI and gel purified. A 174-bp DNA fragment of the rhlI promoter (−19 to −196 in relation to the translational start site) was excised from pCRrhlI by using EcoRI and gel purified. The RB1 promoter fragment of algD (52) was utilized as a positive control. The fragments were end labeled with [γ-32P]ATP (6,000 Ci/mmol; NEN Dupont) using T4 polynucleotide kinase (Invitrogen, Carlsbad, CA). The probes were purified by being passed through a G-25 Sephadex microspin column (Amersham Pharmacia Biotech, Inc., Piscataway, NJ). Binding reactions were carried out as described previously, with some modifications (51). Briefly, the probes were mixed with 200 pmol of purified AlgR containing 25 mM Tris-HCl (pH 8.0), 0.5 mM dithiothreitol, 20 mM KCl, 0.5 mM EDTA, 5% glycerol, 10 μg salmon sperm DNA, and an additional 0.25 μg of poly(dI-dC) per ml as nonspecific competitor DNA. Competition assays were performed by the addition of 1, 5, and 10 μg of unlabeled rhlA or rhlI fragments or 10 μg of poly(dI-dC) to determine the specificity of AlgR. In addition, mutagenesis of the AlgR consensus sequence (CCGTTCGTCC) in rhlA RB1 (−920 to −939 relative to the translational start), rhlA RB2 (−936 to −955), and rhlI RB1 (−134 to −115) was performed according to Mohr et al. (52) using a QuikChange II mutagenesis system (Stratagene) and oligonucleotides rhlaNBFor and rhlaNBRev (Table (Table2)2) for rhlA and rhlINBFor and rhlINBRev (Table (Table2)2) for rhlI to generate plasmids pCRrhlARB1M, pCRrhlARB2M, pCRrhlARB1&2M, and pCRrhlIRB1M (Table (Table1).1). The mutation of the rhlA AlgR consensus sequence CCGTTCGTCC to TTACTCGTCC was confirmed by DNA sequencing of the gel shift fragments. After incubation for 10 min at room temperature, the samples were separated by electrophoresis on a 5% native polyacrylamide gel with Sharp's buffer (6.7 mM Tris-HCl [pH 8.0], 3.3 mM sodium acetate, 1.0 mM EDTA) used as running buffer for approximately 1.5 h at 30 mA. Subsequently, the gel was dried and bands were visualized by autoradiography.
The construct for mutation of the rhlI chromosomal AlgR binding site was constructed in vitro by crossover PCR (32). Two initial PCRs were performed, the first using forward primer RP1 and reverse primer RP2 and the second using forward primer RP3 and reverse primer RP4 (Table (Table2).2). The products of these reactions have complementary sequences (the 5′ end of primer RP2 is complementary to primer RP3) and contain the desired mutation TTAC. These products were used as the template for a subsequent crossover PCR using primers RP1 and RP4. This resulting product was digested with SacI and XbaI and cloned directionally into the suicide vector pCVD442 (19) to create plasmid pCVD442rhlImut. The plasmid pCVD442rhlImut was introduced into PAO1 by triparental conjugation. Single recombinants were selected by screening for carbenicillin resistance. The allelic exchange (second) recombination event was induced by selection for sucrose resistance. Several clones were selected for DNA sequencing to confirm the desired mutation.
Statistics on the lacZ reporter assays, autoinducer quantitations, rhamnolipid determinations, and elastolytic assays were performed with one-way analysis of variance (ANOVA) with Tukey's correction. Statistics on biofilm key variables were done with COMSTAT (31).
Previous work indicated that an algR mutant was deficient in biofilm formation in a static biofilm model up to the 8-h time point (83). In order to further examine this phenotype, wild-type PAO1 and its isogenic algR deletion strain PSL317 (Table (Table1)1) were grown for 6 days in a continuous-flow system and the biofilm development was imaged at days 1, 3, and 6. In the flow chamber biofilm system, P. aeruginosa was grown under hydrodynamic conditions with a continuous nutrient supply and glutamate as the carbon source (31, 38, 83). As shown in Fig. 1A and B, wild-type PAO1 and the algR mutant strain (PSL317) formed similar biofilms after 24 h of culture. However, by day 3, the PSL317 biofilm was greatly reduced in biomass and thickness compared to those of wild-type PAO1 (Fig. 1C, D, E, and F) as measured by COMSTAT (Table (Table3).3). By day 6, PAO1 formed the characteristic column-like macrocolonies surrounded by fluid-filled channels (Fig. 2A and B). In contrast, the algR mutant contained sparsely distributed microcolonies (Fig. 2C and D). Furthermore, day 3 and day 6 ΔalgR biofilms were significantly decreased in total biomass (P < 0.001), average thickness (P < 0.001), and maximum thickness (P < 0.001; day 6 only) compared to those of PAO1 biofilms as determined by COMSTAT analysis (Table (Table33).
It has been previously demonstrated that algR mutants are defective in type IV pilus-mediated twitching motility (82, 83) and that AlgR activates twitching motility via the fimTU-pilVWXY1Y2E operon (44). In order to determine if the defect in twitching motility in the algR mutant could account for the defects observed in biofilm formation, we complemented PSL317 in trans with the fimTU-pilVWXY1Y2E operon which restored twitching motility (44). It has also been shown that twitching and flagellar motility are required for swarming motility (39) and proper biofilm formation (73). The ΔalgR mutant is defective for swarming motility, and complementation with the plasmid pVDtacPIL harboring the fimTU and pilVWXY1Y2E genes restored normal swarming activity (Table (Table4).4). However, complementation with pVDtacPIL did not restore the ΔalgR static or flowthrough biofilms to the wild-type phenotype (Fig. 2E and F and and3;3; Table Table3),3), whereas complementation with algR did (Fig. 2G and H and 3A and B; Table Table3).3). It has also been shown that biofilm phenotypes are affected by the composition of nutrients used to grow the biofilm (38, 58, 73). However, at least for the static biofilm assay, the same phenotype for the algR biofilm was observed in LB as in minimal medium with glutamate (Fig. (Fig.3B).3B). This suggests that AlgR controls genes in addition to those involved in twitching motility and alginate production which may be responsible for the defects in biofilm formation in algR-deficient strains.
In order to determine which AlgR-dependent genes were responsible for the defective biofilm phenotype observed, global transcriptional analyses were performed on wild-type (PAO1) and algR deletion (PSL317) strains in the day 6 biofilms. Microarray analysis of PAO1 and PSL317 grown as continuous culture biofilms identified 765 genes that were differentially regulated by at least twofold and whose differences were statistically significant (P < 0.05; see the supplemental material). The most highly up-regulated genes in the ΔalgR biofilm were classified as phage/transposon genes (42 genes), putative enzymes (33 genes), and secreted factors (24 genes). In contrast, genes involved in motility and attachment (18 genes) and transcriptional regulation (36 genes) were largely down-regulated (Fig. (Fig.4).4). Interestingly, of the 76 genes represented in the quorum-sensing regulon by three independent studies (30, 72, 81), 44 (58%) were regulated by AlgR during biofilm growth (Fig. (Fig.55 and Table Table5).5). Moreover, with the exception of three genes, the quorum-sensing genes identified in the ΔalgR biofilm transcriptional profile were not present in three separate AlgR-profiling experiments that utilized planktonic growth conditions (44), suggesting that AlgR regulation of these quorum-sensing genes may be biofilm specific (Fig. (Fig.5).5). Most of the genes belonging to the AlgR and quorum-sensing regulons remain to be characterized, while 23 of the genes have been well described (Table (Table5)—for5)—for example, the two most highly regulated quorum-sensing genes identified, rhlA and rhlB, which were up-regulated 111- and 79-fold, respectively, in the algR deletion strain and are tightly regulated by the RhlR-RhlI tandem (55). The LasR-LasI-dependent genes lasA and lasB (80) were also significantly elevated in PSL317 (4- and 10-fold, respectively). In addition, the HCN synthesis genes hcnABC (65) were significantly upregulated (three-, seven-, and fourfold, respectively) in the ΔalgR strain. rsaL, the negative regulator of lasI (15), was significantly down-regulated, by fivefold, in the algR deletion mutant (Table (Table5).5). The vast number of quorum-sensing genes that were differentially regulated in the ΔalgR biofilm, but not in algR mutants grown planktonically, indicated that AlgR may regulate quorum sensing in a biofilm-specific manner.
Table Table55 lists 44 genes identified in the array analysis that also belong to the quorum-sensing regulon according to three independent analyses (30, 72, 81). Of the 23 genes that are well characterized, three genes (pqsA, rsaL, and phzB) and two operons (hcnABC and rhlAB) contain putative AlgR binding sites in their promoter regions. Interestingly, the genes involved in rhamnolipid biosynthesis (rhlA and rhlB) demonstrated the highest severalfold changes in the transcriptome analysis (111- and 79-fold, respectively). Several studies have demonstrated that the levels of rhamnolipids can influence biofilm architecture and composition (3, 10, 36). Therefore, it was hypothesized that AlgR repression of rhlAB may be required for normal biofilm development in the continuous-culture model. In order to confirm the array data, an rhlA::lacZ translational fusion was constructed and integrated as a single copy on the chromosome of the ΔalgR and wild-type strains. No difference in β-galactosidase activity was observed when PAO1 and PSL317 were grown in minimal medium broth (Fig. (Fig.6A).6A). However, rhlA expression was significantly increased in PSL317 compared to its expression in PAO1 when the strains were cultured for 6 days in continuous culture biofilms (P < 0.05) (Fig. (Fig.6A).6A). Both LasR-3-oxo-C12-HSL and RhlR-C4-HSL can regulate rhlAB expression. Therefore, the ability of AlgR to regulate either of these regulatory genes was also explored. Although rhlI was not identified in the transcriptome analysis, its promoter region contains a perfect AlgR consensus sequence located at −133 to −117 relative to the translational start site (Fig. (Fig.7E).7E). Therefore, the expression of a rhlI::lacZ transcriptional fusion (as well as lasI::lacZ expression, for comparison) in PAO1 and PSL317 was examined to determine the effect of AlgR on the autoinducer synthases. The expression of rhlI in PSL317 was significantly increased (P < 0.01) under biofilm growth conditions but not in broth (Fig. (Fig.6B).6B). In contrast, no difference between PAO1 and PSL317 in lasI transcription was observed when grown as a biofilm or planktonically (data not shown). These data suggest that AlgR represses Rhl but not Las quorum sensing in a biofilm-specific manner.
In order to confirm the results of the lacZ promoter fusion expression studies, the levels of C4-HSL and 3-oxo-C12-HSL produced by PAO1 and PSL317 grown in aerobically shaken minimal medium broth and in continuous-flow biofilms were compared. The amount of C4-HSL was increased nearly fivefold (P < 0.01) in PSL317 compared to the amount in PAO1 in 6-day biofilm effluents (Fig. (Fig.6D).6D). These differences were less apparent in planktonically grown P. aeruginosa. PSL317 produced approximately twofold more C4-HSL (P < 0.05) than PAO1 in minimal medium broth. No significant difference in 3-oxo-C12-HSL levels was observed between strains PAO1 and PSL317 during biofilm or planktonic growth (data not shown). These results indicate that AlgR represses the production of the Rhl autoinducer C4-HSL, but not 3-oxo-C12-HSL, during biofilm growth.
Since rhlA and rhlB were the most significantly increased quorum-sensing genes in ΔalgR biofilms, the concentrations of rhamnolipids in day 6 biofilm effluents and in supernatants of broth-grown P. aeruginosa were measured. Rhamnolipid concentrations were significantly higher (P < 0.001) in PSL317 than in PAO1 when grown as biofilms. In contrast, no differences in rhamnolipid production were observed when PAO1 and PSL317 were cultured in minimal medium broth (Fig. (Fig.6C).6C). These results were in agreement with the elevated levels of C4-HSL observed in ΔalgR biofilms and indicate that AlgR represses the Rhl quorum-sensing cascade in continuous-culture biofilms.
The gene encoding elastase, lasB, was also upregulated (10-fold) in the ΔalgR mutant in a biofilm. This suggested that AlgR either directly or indirectly (i.e., via Rhl) regulated elastase production. Therefore, the amounts of elastolytic activity in effluents from PAO1 and PSL317 continuous-culture biofilms were measured after 6 days of growth. Elastolytic activity was twofold higher in the ΔalgR mutant (P < 0.001; data not shown). The elastolytic activity was also tested in the supernatants of PAO1 and PSL317 grown in minimal medium broth overnight. Supernatants from planktonic PSL317 had 1.5-fold-higher elastolytic activity than the wild-type or complemented strains, but the differences were less significant than those from biofilms (P < 0.01; data not shown). These results suggest that AlgR represses lasB gene transcription and downstream elastase production, possibly via its repression of the Rhl quorum-sensing circuit (62).
Since rhlA and rhlI transcription was repressed by AlgR in the biofilms, the promoter regions of each gene were examined for putative AlgR binding sites. The rhlI promoter region contains a nearly perfect AlgR recognition sequence located at −129 to −120 relative to the translational start site (Fig. (Fig.7E).7E). The rhlA promoter region contains two putative overlapping AlgR binding sites, RB1 located at −925 to −934 and RB2 at −940 to −949 relative to the translational start. Therefore, the ability of AlgR to bind each of these sites in an in vitro gel mobility shift assay was tested. As shown in Fig. 7A and B, specific complexes were formed when AlgR was added to the radioactive rhlA and rhlI DNA fragments. The addition of nonradioactive specific competitor reduced the amount of probe shifted by AlgR in a dose-dependent manner (Fig. 7A and B; lanes 2 to 4), indicating the specificity of AlgR for both the rhlA and rhlI promoter region. Furthermore, mutation of the AlgR consensus sequence of conserved nucleotides CCGT (52) to TTAC within the rhlI promoter completely abolished AlgR binding (Fig. (Fig.7D).7D). When the same site-directed mutagenesis was performed on the two putative AlgR binding sites in the rhlA promoter, mutation of RB1 but not RB2 resulted in a loss of mobility shift (Fig. (Fig.7C).7C). These results provide in vitro evidence that AlgR binds specifically to the rhlI and rhlA promoters and lend support to the hypothesis that AlgR directly controls rhlA and rhlI expression.
In order to determine if AlgR repression of the Rhl quorum-sensing system was required for normal biofilm development under the conditions examined, the AlgR binding site within the rhlI promoter was mutated in vitro and allelic exchange was used to place it on the chromosome of PAO1. Mutation of the wild-type sequence CCGT to TTCA was confirmed by DNA sequencing. Gel mobility shift assays confirmed that AlgR did not bind this mutated sequence (Fig. (Fig.7D),7D), and therefore, AlgR repression of rhlI should be abolished in the mutated strain. When PAO1 containing the mutated rhlI promoter (PAO1rhlImut) was grown for 6 days as a continuous-culture biofilm, the resulting phenotype was similar to that of the algR mutant strain. PAO1rhlImut biofilms were significantly reduced in total biomass, average thickness, and maximum thickness and were significantly rougher than wild-type biofilms (Fig. (Fig.88 and Table Table3).3). In addition, this strain produced elevated amounts of rhamnolipids compared to the levels in PAO1, similar to the algR mutant (Fig. (Fig.6C).6C). These results demonstrate that AlgR regulation of rhlI is direct and confirm that AlgR repression of the Rhl quorum-sensing circuit during biofilm growth is essential for proper biofilm maturation.
In this report, evidence is presented demonstrating that the P. aeruginosa virulence regulator AlgR controls biofilm maturation by repressing the Rhl quorum-sensing system in P. aeruginosa PAO1. An indication that AlgR may be required for biofilm initiation in the static biofilm model was previously demonstrated when an algR mutant formed one-third the biomass of wild-type P. aeruginosa after 8 h of static culture (83). In contrast to the reported observations in the 8-h static biofilm system (83), we report that the P. aeruginosa algR deletion mutant displayed adherence and initial colonization of the flowthrough cell similar to that of the wild type up to 24 h, but abnormal biofilm formation was observed after 3 days of culture.
Possible explanations for the altered biofilm phenotype observed for the algR mutant include a twitching motility defect, altered swarming motility, and increased rhamnolipid secretion. Type IV pilus-mediated twitching plays a role in biofilm development by enabling P. aeruginosa microcolonies to spread over the substratum (38, 59). Mutants that were defective in twitching motility were not impaired in the early stages of biofilm development but eventually formed biofilms distinguishable from those of wild-type organisms in a flow chamber system like the one used in this study (31, 38, 59). AlgR controls twitching motility, and this control is likely dependent upon phosphorylation (82, 83). In addition, the expression of the fimTU-pilVWXY1Y2E operon alone can complement the algR twitching motility defect, strongly indicating that AlgR controls this operon and that lack of its expression results in loss of twitching motility (1, 44). Under the conditions tested, the biofilm formed by the algR deletion strain displayed one-third less biomass and an inability to form the columnar architecture typical of mature wild-type biofilms (11). When twitching motility was restored by the introduction of the fimTU-pilVWXY1Y2E operon in trans, normal biofilm maturation was not restored after 6 days of culture. Therefore, twitching motility likely did not play a significant role in the biofilm defect observed for the algR mutant. It has also been shown that quorum sensing affects biofilm formation in a nutritionally dependent fashion through swarming motility (73). Since the algR biofilm phenotype was not dependent upon nutrition, as rich and minimal medium biofilms resulted in the same phenotype (Fig. (Fig.3),3), and complementation of swarming motility with the fimTU-pilVWXY1Y2E operon (Table (Table4)4) did not complement the algR biofilm defect (Fig. 2E and F), the deletion of algR increased rhamnolipid production (Fig. (Fig.6C)6C) which resulted in the altered biofilm phenotype observed.
Microarray analyses were utilized to determine which AlgR-controlled genes did affect biofilm formation, and a large number of known quorum-sensing genes were identified. Based upon these results, the possibility that coordinate regulation of quorum-sensing circuits and the AlgR regulatory network is necessary for successful biofilm development was explored. Interestingly, one previous report has indicated that there may be a connection between AlgR and the Rhl quorum-sensing system, when the culturing of an rhlI mutant resulted in compensatory mutations in algR (2). P. aeruginosa currently has three identified types of quorum-sensing systems: (i) the Las system which produces 3-oxo-C12-HSL (60); (ii) the Rhl system which produces C4-HSL (61); and (iii) the PQS system which produces 2-heptyl-3-hydroxy-4-quinolone (63). Disruption or alteration of the quorum-sensing regulatory cascade in P. aeruginosa has been shown to interfere with normal biofilm architecture and development (10, 14, 36). However, others have reported no difference between wild-type biofilms and biofilms defective in the Las quorum-sensing system (31, 77). These inconsistencies may be attributed to differences in P. aeruginosa strains and culture conditions. One recent report indicates that the expression of the Las and Rhl quorum-sensing systems is clearly dependent upon growth conditions (21). The Rhl quorum-sensing system has also been studied extensively in biofilms (10, 42). Rhamnolipid surfactants under the control of the Rhl quorum-sensing system are essential for proper maintenance of water channels and biofilm architecture (10), and there is evidence that rhamnolipid expression occurs after stalks have formed but before the capping in of the mushroom-like structures (42).
The ability of AlgR to directly regulate Rhl quorum sensing most likely explains the biological properties of the ΔalgR strain during biofilm growth. Such dysregulation of quorum sensing could also explain the attenuated phenotypes of both algR deletion and AlgR overexpression strains during murine infection (43), as disruption of all three quorum-sensing systems has resulted in loss of virulence (6, 16, 26). Interestingly, the overall atypical characteristics of ΔalgR biofilms resembled those of P. aeruginosa biofilms overexpressing rhamnolipids (10). AlgR regulation of rhamnolipid production has been previously suggested by the involvement of AlgC in rhamnolipid synthesis (57). The product of the algC gene, which is involved in alginate production through its phosphomannomutase activity (86, 87) and in lipopolysaccharide synthesis through its phosphoglucomutase activity (9), also participates in rhamnolipid production (57). In addition, the overproduction of rhamnolipids by P. aeruginosa, as observed in ΔalgR biofilms, inhibits the maintenance of the biofilm infrastructure (10). Furthermore, the addition of exogenous rhamnolipids to established P. aeruginosa and Burkholderia biofilms causes detachment and dispersion of bacteria from the biofilm (3, 36). Thus, numerous investigators have established that excess levels of rhamnolipids interfere with normal biofilm architecture. The inability of AlgR to repress rhlAB transcription or the upstream positive regulator rhlI can account for the increased levels of rhamnolipids observed in ΔalgR biofilms. Moreover, the mutation of the chromosomal rhlI promoter AlgR binding site in strain PAO1rhlImut resulted in elevated rhamnolipid production and a biofilm phenotype identical to that of the algR deletion strain. Taken together, these results strongly support the notion that dysregulation of the Rhl quorum-sensing system by deletion of algR leads to overproduction of rhamnolipids and the altered biofilm phenotype. These studies also suggest that the signal to which AlgR responds in vivo may be involved in the dispersion of the biofilm. It is relatively easy to imagine that some physiologically relevant molecule binds to AlgR or AlgZ and causes the derepression of rhlI, resulting in increased rhamnolipid production and, hence, dispersion of the biofilm.
The regulatory mechanisms of AlgR are complex in that both phosphorylation-dependent and -independent mechanisms of activation have been demonstrated (45, 83). When the predicted phosphorylated residue of AlgR, aspartate 54, was mutated to asparagine, alginate production was still activated (45). However, the same mutation abolished AlgR activation of twitching motility (83). Furthermore, AlgR can switch from a repressor of HCN production in the nonmucoid background to an activator of HCN production in mucoid strains (7). In the current study, it appears that AlgR only controls the Rhl quorum-sensing system when it is attached in a biofilm, further complicating the sensory input required for AlgR regulation. A similar contact dependence has been reported for another AlgR-controlled gene, algC, when Davies and Geesey (13) showed that bacterial attachment induces the transcription of specific genes, including algC. This would imply that attachment to a surface may provide a signal to the bacterium to stimulate the AlgR regulon. The signaling requirements for AlgR activation and repression of target genes such as those involved in alginate production, HCN production, and twitching motility may parallel changes in the bacterium's environmental state.
In addition to rhlAB and rhlI, a significant number of transcripts that encode secreted factors also appear to be repressed by AlgR in nonmucoid PAO1 biofilms (Fig. (Fig.4).4). In earlier work, 17 proteins uniquely expressed in an algR mutant were identified by two-dimensional sodium dodecyl sulfate-polyacrylamide gel electrophoresis (43). Furthermore, AlgR represses hcnA in nonmucoid P. aeruginosa and activates its expression in mucoid strains (7). Taken together, these findings further support the idea that AlgR plays a previously unrecognized role as a repressor of gene expression in nonmucoid P. aeruginosa. In addition, the concept that AlgR switches from being a repressor to an activator of virulence products during mucoidy is intriguing and is under further investigation.
In conclusion, our results indicate that AlgR represses the Rhl quorum-sensing system in nonmucoid P. aeruginosa during continuous culture biofilm growth and that such repression is necessary for proper biofilm maturation. Furthermore, the ability of AlgR to repress rhlI and rhlAB during biofilm growth, but not during planktonic culture, suggests that AlgR may utilize a contact-dependent or biofilm-specific mode of regulation. Further insight into the coordinate regulation of the AlgR- and Rhl-dependent pathways during biofilm growth will enhance our understanding of P. aeruginosa pathogenesis in CF disease.
The work was supported by NIH grants RO1AI50812 to M.J.S. and R37AI37713 to B.H.I., NIH training grant 5T32AI07285 to V.E.W., and NIH grant R01-40541 and a Cystic Fibrosis Foundation grant to D.J.H.
Published ahead of print on 31 August 2007.
†Supplemental material for this article may be found at http://jb.asm.org/.