• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of aacPermissionsJournals.ASM.orgJournalAAC ArticleJournal InfoAuthorsReviewers
Antimicrob Agents Chemother. Sep 2005; 49(9): 3858–3867.
PMCID: PMC1195439

Molecular Basis of Azithromycin-Resistant Pseudomonas aeruginosa Biofilms


Pseudomonas aeruginosa biofilms are extremely recalcitrant to antibiotic treatment. Treatment of cystic fibrosis patients with azithromycin (AZM) has shown promise. We used DNA microarrays to identify differentially expressed transcripts in developing P. aeruginosa biofilms exposed to 2 μg/ml AZM. We report that transcripts for multiple restriction-nodulation-cell division (RND) efflux pumps, known to be involved in planktonic antibiotic resistance, and transcripts involved in type III secretion were upregulated in the resistant biofilms that developed in the presence of AZM. Interestingly, the MexAB-OprM and MexCD-OprJ efflux pumps, but not type III secretion, appear to be integral to biofilm formation in the presence of AZM, as evidenced by the fact that a mutant deleted in both mexAB-oprM and mexCD-oprJ was unable to form a biofilm in the presence of AZM. A mutant deleted in type III secretion was still able to form biofilms in the presence of drug. Furthermore, single mexAB-oprM- and mexCD-oprJ-null mutants were able to form a biofilm in the presence of drug, indicating that either of the pumps can confer resistance to AZM during biofilm development. In contrast to planktonically grown cells, where no mexC expression was detectable regardless of the presence of AZM, biofilms exhibited induction of mexC expression from the outset of their formation, but only in the presence of AZM. mexA, which is constitutively expressed in planktonic cells, was uniformly expressed in biofilms regardless of the presence of AZM. These data indicate that the MexCD-OprJ pump acts as a biofilm-specific mechanism for AZM resistance.

Pseudomonas aeruginosa is a ubiquitous gram-negative opportunistic pathogen and a prevalent organism implicated in nosocomial infections (12). This bacterium has also been shown to be highly problematic in immunocompromised individuals (1, 8, 26). A serious and significant manifestation of this pathogen is its ability to cause chronic pulmonary infections in people with cystic fibrosis (CF), the most common lethal genetically inherited disease among Caucasians (22). Once chronic infection is established, P. aeruginosa is extremely difficult, if not impossible, to eradicate using conventional antibiotic treatments, in large part due to its propensity to form biofilms within the CF lung (7, 39). P. aeruginosa, living as a biofilm, has been shown to be as much as 1,000 times more resistant to antibiotic treatment than its planktonic counterpart (31).

One antibiotic that has shown promise in treatment against chronic P. aeruginosa infections in CF is the macrolide azithromycin (AZM) (21). CF patients treated with AZM have shown improvement based on increased lung function and body weight (11, 21, 36, 48). The treatment is controversial, because the exact mechanism of killing of P. aeruginosa is unknown (36). Elucidation and further understanding of the mechanism by which AZM acts against P. aeruginosa infections in the CF lung are needed to optimize treatment efficacy.

An antibiotic resistance mechanism that has undergone a great deal of study in P. aeruginosa is the transport of antimicrobials via efflux pump systems, especially those of the resistance-nodulation-cell division (RND) family (38). RND efflux pumps have been shown to cause the efflux of a multitude of substrates, including most classes of antibiotics (29, 34, 35). These tripartite RND pumps require an inner membrane efflux transporter, a periplasmic membrane fusion protein, and an outer membrane protein channel for function (32). Sequencing of the P. aeruginosa genome has revealed at least 12 efflux pumps of the RND family (43). To date, the most extensively characterized efflux pump is MexAB-OprM, the only one known to be constitutively expressed in P. aeruginosa PAO1 (16, 35). MexCD-OprJ has been shown to play a role in antibiotic resistance; however, this pump is thought to be expressed only when its corresponding negative regulator, encoded by nfxB, is disrupted (33, 35). This pump was shown to be upregulated in multidrug-resistant strains of P. aeruginosa when they were exposed to various fluoroquinolones due to nfxB mutations (33). MexEF-OprN, another inducible pump, has been shown to be expressed when its negative regulator, encoded by nfxC, is disrupted (23). MexEF-OprN expression also requires the functional activator protein MexT (27). Finally, MexXY has been shown to cause efflux of aminoglycosides and to be present in panaminoglycoside-resistant clinical isolates (40). Curiously, although contributions of RND efflux pumps to antibiotic resistance in biofilms have been investigated, no correlation has yet been shown (10).

In the present study, we used a combination of DNA microarray analysis of biofilm transcript expression and established microscopic and molecular techniques to investigate AZM resistance mechanisms in P. aeruginosa biofilms. We show that mexAB-oprM and/or mexCD-oprJ is essential for biofilm development in the presence of sub-MIC AZM concentrations. Furthermore, when mexCD-oprJ is expressed during sub-MIC AZM exposure, it remains hyperexpressed even in the absence of AZM, but only when P. aeruginosa is grown as a biofilm. Importantly, hyperexpression of mexCD-oprJ in an AZM-resistant biofilm phenotype does not confer increased resistance to other antibiotics, including tobramycin, ciprofloxacin, cefepime, meropenem, and tetracycline.


Bacterial strains and plasmids.

All bacterial strains utilized in this study were derived from P. aeruginosa PAO1 (Table (Table1).1). Plasmid pTdK, which contains a lac promoter fusion to gfp-mut3 (10), was introduced into wild-type (WT) and mutant strains via electroporation according to established protocols. Strains PAO200 [Δ(mexAB-oprM)], PAO238 [Δ(mexAB-oprM) Δ(mexCD-oprJ)], PAO255 [Δ(mexAB-oprM) Δ(mexEF-oprN)], and PAO280 [Δ(mexAB-oprM) Δ(mexXY)] were derived from PAO1 as previously described (5). The unmarked strain PAO461 [Δ(mexCD-oprJ)] was derived from PAO1 using the previously described Flp/FRT recombinase technology (19). The source for the Δ(mexCD-oprJ) mutant allele was pPS1094, a pEX18Ap (19) derivative containing a BamHI fragment from pKMJ002 (33) on which a deletion of a 6,138-bp region encompassing three ClaI fragments is replaced by the Gmr marker from pPS856 (19). Although the nfxB gene is still present in the mutant, it is neither transcribed nor translated, because the ClaI deletion removes both the nfxB promoter and the Shine-Dalgarno sequence. For complementation of Δ(mexCD-oprJ) mutants, a mini-Tn7 transposon system was used (4). The 9,923-bp BamHI fragment from pKMJ002 encompassing the nfxB+-mexC+D+-oprJ+ region was cloned into the BamHI site of pUC18-mini-Tn7T-Gm to form pPS1450. The mini-Tn7T-Gm-nfxB+-mexC+D+-oprJ+ region was then transposed into the attTn7 sites of the PAO238 and PAO461 chromosomes, followed by Flp-mediated excision of the Gmr marker in PAO238 to yield strain PAO438.

Strains and plasmid used in this study

Strain PAO430, harboring a chromosomally integrated mexC-GFPmut3 fusion, was constructed as follows. pBS.mexC391 (laboratory collection), a plasmid harboring the nfxB-mexC intergenic region, was digested with BamHI and KpnI, and a gel-purified 450-bp BamHI-KpnI fragment was cloned between the same sites of pPS742 to create pPS1451. pPS742 (laboratory collection) is pUCP20 with gfp-mut3 (6) cloned in the same orientation as the lac promoter. Next, a 1,219-bp HindIII-KpnI fragment harboring the mexC′-gfp-mut3 fusion from pPS1451 was cloned between the same sites in the mini-Tn7 delivery vector pUC18-mini-Tn7T-Gm to form pPS1455. The mini-Tn7 insert was then transposed to the PAO1 chromosome by coelectroporation with the helper plasmid pTNS1 (4), Gmr transformants were selected on LB medium with 30 μg/ml gentamicin, and one of these was purified and retained as PAO428. PAO430 was then obtained by Flp-mediated excision of the Gmr marker according to an established procedure (19).

Stack reactor biofilm studies.

A stack reactor was used to cultivate biofilms for RNA extraction (see Fig. S1 in the supplemental material) as described previously (45). Briefly, biofilms were cultivated in the stack reactor in a once-flowthrough system at 37°C. Three glass slides placed within the separator plates of the stack reactor were used for the biofilm substratum. A continuous flow rate of 0.78 ml/min was maintained in the stack reactor using a peristaltic pump (Cole-Parmer, Vernon Hills, IL), with bubble traps located between the pump and the stack reactor. Each chamber of the stack reactor was fed from an individual influent line of modified FAB medium (9). When specified, AZM, a generous gift from Pfizer (Groton, CT), was added to a final concentration of 2 μg/ml.

Each chamber of the stack reactor was inoculated with 1.0 ml of a PAO1-pTdK culture at an optical density at 660 nm of 1.3 by using a syringe. Flow commenced after a 2-h quiescent period. Biofilm samples, either exposed or not exposed to 2 μg/ml AZM, were cultivated in three independent biological replicates. Biofilms cultivated in the absence of AZM were harvested 4 days after inoculation. Biofilms grown in the presence of AZM typically had a 2-day lag in accretion compared to controls grown in the absence of AZM and thus were harvested 6 days after inoculation to provide equivalent numbers of cells.

RNA extraction from biofilm samples.

Biofilms cultivated on the glass slides of the stack reactor were physically scraped using a sterile razor blade and placed directly in a 1.5-ml sterile Eppendorf tube containing 700 μl of RNAwiz (Ambion, Austin, TX). Each side of the glass slide was scraped into a separate tube. After scraping, the sample was immediately vortexed for 15 seconds and placed on ice. This process was repeated for each glass slide until the biofilms from all glass slides in the stack reactor had been harvested.

Each biofilm sample was divided into two 350-μl aliquots and mechanically disrupted using the RiboPure Bacteria kit (Ambion) according to the manufacturer's instructions, except that aliquots were recombined during the filtration step. Biofilm nucleic acid samples were further combined such that one sample contained all the biofilm nucleic acid from a single stack reactor prior to ethanol precipitation overnight at −80°C using 0.1× volume 3 M sodium acetate and 2.5× volume 100% ethanol. Following ethanol precipitation, RNA was purified with DNase I for 1 h at 37°C by using a DNA-free kit (Ambion) according to the manufacturer's instructions. RNA quantity was assessed via UV spectrophotometer (Beckman) according to standard methods. RNA purity was assessed by PCR (model 9700; Applied Biosystems, Foster City, CA) (30 cycles of 94°C for 60 s, 65°C for 45 s, and 72°C for 60 s, with a final elongation at 72°C for 5 min) using a SuperTaq kit (Ambion) with primers specific for groEL (sense primer, 5′-CGCTCGCAAGAAAATGCTGGTC-3′; antisense primer, 5′-CGACGGACAGTTCGTTTTCCAG-3′). The quality of the total RNA was determined using a model 2100 Agilent (Palo Alto, CA) Bioanalyzer and a Lab-on-a-Chip kit according the manufacturer's instructions.

Microarray analysis.

Synthesis of cDNA, target hybridization, staining, and scanning using the Affymetrix GeneChip system were performed at the Microarray Core Facility in the Functional Genomics Center at the University of Rochester as previously described (44). Microarray data were analyzed using Microarray Suite Software v. 5.0. Reverse transcriptase PCR (RT-PCR) was performed on biofilm RNA samples to validate microarray analysis by using a One-Step RT-PCR kit (Invitrogen, Carlsbad, CA) in a Sprint thermal cycler (ThermoElectron, Waltham, MA). Reverse transcription parameters were 50°C for 40 min and 94°C for 4 min, followed by PCR parameters of 35 cycles of 94°C for 40 s, 60°C for 45 s, and 72°C for 60 s, with a final elongation of 72°C for 7 min. Transcript targets used for RT-PCR were as follows: mexA (sense primer, 5′-CAGCAGCTCTACCAGATCGAC-3′; antisense primer, 5′-GTATTGGCTACCGTCCTCCAG-3′), mexC (sense primer, 5′-CTGATTTGCGTGCAATAGGAAG-3′; antisense primer, 5′-ATCGATCTGGAACAGCAGGT-3′), nfxB (sense primer, 5′-GAGACCGTACTGAACCAGATCAT; antisense primer, 5′-GTGATGAACAGTTCGGTGAACA-3′), and exsA (sense primer, 5′-TTCCATTATCTGCCCAGTTTCTA-3′; antisense primer, 5′-TGAGGTAGTGCTTCTCCATGAAT-3′).

Sequence analysis of nfxB by PCR.

Total chromosomal DNAs of PAO1 and the PAO1 biofilm variant (PAO1-BV) were prepared from overnight cultures grown in LB using a DNA isolation kit (QIAGEN) according to the manufacturer's instructions. nfxB was amplified by PCR using a SuperTaq (Ambion) kit with primers as described above. A 50-μl PCR mixture included 10 ng of chromosomal DNA, 0.5 μM of each primer, 1× SuperTaq buffer (Ambion), 0.2 μM deoxynucleoside triphosphates, and 1 U of SuperTaq polymerase (Ambion). PCR was performed using 30 cycles of 1 min at 96°C, 45 s at 55°C, and 1.5 min at 72°C, with a final 5-min elongation at 72°C. PCR products were purified with the QIAGEN PCR purification kit and sequenced by the NucleoCore Sequencing Facility, University of Rochester, Rochester, NY. A sequence comparison was then done between PAO1 and PAO1-BV as well as the published nfxB sequence available from www.pseudomonas.com.

Flow cell biofilm studies.

PAO1, efflux pump mutants, and a type III secretion system (TTSS) mutant, PAO1 exsA::Ω (14), were analyzed for their biofilm-forming abilities in the presence and absence of sub-MIC AZM levels using flow cells at 37°C and modified FAB as previously described (9). When specified, AZM was added to a final concentration of 2 μg/ml. All biofilms were nondestructively imaged using a scanning confocal laser microscope (SCLM) equipped with a dual photon laser at 488 nm (Leica Lasertechnik, GmbH, Heidelberg, Germany). During the day zero imaging time, five areas were randomly chosen per flow cell for monitoring throughout the duration of the experiment. Spatial positioning (x and y coordinates) was recorded for each of the five areas imaged by securing the flow cell to the microscope stage and recording the position using the left/right and forward/back graduations on the microscope stage, with each corresponding area imaged daily for 6 days. Scans were taken through the biofilm that had accumulated on the glass surface at varying depths from the substratum to the biofilm surface. Typically, optical sections were taken every 0.5 μm. For each of the image stacks collected, multiple quantitative measurements were calculated using COMSTAT (18) coupled to MatLab, v. 6.5, software (Mathworks, Springfield, MA). Parameters chosen for analysis included (i) total biomass, defined as μm3/μm2, (ii) average thickness, in μm, (iii) the surface roughness coefficient, and (iv) maximum thickness, in μm. Following calculation, data were averaged and standard deviations calculated for the five image areas analyzed per sample.

For the flow cell biofilm study of PAO238, bacteria were stained with BacLight live/dead stain (Molecular Probes, Eugene, OR) to image bacteria on the surface (25), since no green fluorescent protein (GFP)-fluorescing bacteria were noted.

Temporal expression studies.

For determination of mexC temporal expression during development of AZM-resistant biofilms, PAO430 or PAO430-BV biofilms were cultivated in flow cells as described above. Flow cells were stained with a propidium iodide/SYTO 85 (Molecular Probes, Corvallis, OR) emulsion prior to imaging (10). Images were collected from five randomly selected areas using an SCLM. Image stacks were then analyzed using MetaMorph Image Analysis software (Universal Imaging, Germany) to determine the percent mexC-gfp fluorescing bacteria (green) versus total bacteria (red). Data were plotted as percent mexC-gfp expression versus biofilm thickness.

Antibiotic susceptibility testing.

To determine the susceptibility of PAO1-BV versus PAO1 to other antibiotics, planktonic MIC susceptibility testing was carried out using CLSI (formerly NCCLS) standard clinical MIC methods (30). Antibiotics tested included AZM, tobramycin (Sigma), ciprofloxacin and cefepime (USP, Rockville, MD), meropenem (Zeneca, Wilmington, DE), and tetracycline (Sigma). Minimum bactericidal concentrations (MBCs) for biofilm were also determined as described previously (25).


RND efflux genes and type III secretion genes are upregulated in nascent PAO1 biofilms developed during sub-MIC AZM exposure.

Most studies regarding antibiotic resistance mechanisms in biofilms involve exposure of mature biofilms. In a separate study, we report the development of an AZM-resistant PAO1 biofilm phenotype during sub-MIC exposure to AZM throughout biofilm development (PAO1-BV) (15). As a result, we hypothesize the existence of mechanisms occurring at the cellular level in nascent films that may contribute to the AZM-resistant phenotype. To test this hypothesis, we analyzed the global transcriptional response of PAO1 biofilms versus PAO1 biofilms developed in the presence of sub-MIC AZM levels by using Affymetrix GeneChip P. aeruginosa genome arrays. This analysis of these 4-day biofilms revealed 274 genes promoted or repressed at least fivefold in PAO1-BV (PAO1 developed in the presence of sub-MIC AZM concentrations) biofilms compared to PAO1 biofilms (see Table S1 in the supplemental material). Interestingly, RND efflux pump transcripts nfxB, mexC, mexD, oprJ, oprN, and mexX, as well as multiple TTSS transcripts, were upregulated fivefold or more in the PAO1-BV biofilms (Fig. (Fig.11 and and2).2). Expression of a recently identified suppressor of type III secretion, ptrA (PA2808), was downregulated 16-fold in PAO1-BV (see Table S1 in the supplemental material) (17). While mexAB-oprM genes were not markedly upregulated in PAO1-BV (Fig. (Fig.1),1), these genes are constitutively expressed at relatively high levels in PAO1 (38). Our microarray data were validated by RT-PCR for selected transcripts (Fig. (Fig.33).

FIG. 1.
Differential expression of efflux pump transcripts in P. aeruginosa biofilms developed in the presence of azithromycin (2 μg/ml) (PAO1-BV) versus control PAO1 biofilms not exposed to AZM.
FIG. 2.
Differential expression of known type III secretion operons and genes that were up- or downregulated more than fivefold in P. aeruginosa growing in biofilms in response to azithromycin (2 μg/ml) (PAO1-BV) versus control PAO1 biofilms not exposed ...
FIG. 3.
Validation of microarray data by RT-PCR. Lanes: M, 1 kb Plus marker; 1, RNA from biofilm not exposed to AZM; 2, RNA from biofilm developed in the presence of 2 μg/ml AZM; 3, master mix without template; 4; PAO1 DNA.

Either mexAB-oprM or mexCD-oprJ is essential for biofilm development during exposure to sub-MIC AZM levels.

Given the implications of MexAB-OprM, MexCD-OprJ, and MexXY efflux pumps in macrolide efflux (28, 29, 41) and the induction of the type III secretion operons implicated in P. aeruginosa virulence (13) (Fig. (Fig.2),2), we chose to investigate the abilities of selected mutants to form biofilms. Each strain harbored the GFP-expressing pTdK plasmid to assist in microscopic imaging. Interestingly, PAO238 [Δ(mexAB-oprM) Δ(mexCD-oprJ)] was unable to form a biofilm in the presence of sub-MIC AZM levels during the 6-day duration of the experiment (Fig. (Fig.4).4). As shown in Fig. 5A and B, the lack of biofilm formation for PAO238 was evident in the five randomly selected areas per biofilm. To verify that the glass substratum of the flow cell did not contain cells not expressing GFP, cells were stained with the BacLight Live/Dead stain on day 6. Although some surface-associated bacteria were still evident, nearly all stained red, an indication of nonviable cells (data not shown).

FIG. 4.
Scanning confocal laser micrographs of representative image areas of biofilms cultivated in flow cells in the presence and absence of 2 μg/ml AZM. (A) PAO1; (B) PAO238 [Δ(mexAB-oprM)Δ(mexCD-oprJ)]; (C) PAO438 [Δ(mexAB-oprM ...
FIG. 5.
Biofilm characteristics of wild-type PAO1 and PAO1 mutants formed in the presence (dashed lines) or absence (solid lines) of 2 μg/ml AZM. (A) Total biomass (μm3/μm2); (B) average thickness (μm). Each data point is the average ...

These results suggest a role for either mexAB-oprM, mexCD-oprJ, or both in biofilm resistance to AZM. PAO200 [Δ(mexAB-oprM)] was able to form a biofilm in the presence of sub-MIC AZM levels (Fig. (Fig.44 and and5),5), suggesting that mexCD-oprJ was integral to the formation of a biofilm variant. PAO461 [Δ(mexCD-oprJ)] was also able to form a biofilm in sub-MIC AZM concentrations (Fig. (Fig.44 and and5),5), suggesting a role of mexAB-oprM in the formation of a biofilm variant. Partial complementation of PAO238 with nfxB+-mexC+D+-oprJ+ (PAO438) restored the ability to form a biofilm in the presence of AZM (Fig. 5A and B). The differences in characteristics of PAO438 biofilms compared to WT biofilms are not clear but may be an artifact of the complementation due to the chromosomal location of the complementing insert. This would indicate that mexCD-oprJ was transcribed at sufficient levels in the absence of mexAB-oprM to allow the formation of a biofilm in the presence of AZM. Although oprN, mexX, and multiple type III secretion transcripts were also upregulated in PAO1-BV biofilms (see Table S1 in the supplemental material), mutants PAO255 [Δ(mexAB-oprM) Δ(mexEF-oprN)] and PAO280 [Δ(mexAB-oprM) Δ(mexXYZ)], as well as the type III secretion mutant PAO1 exsA::Ω, were able to form biofilms in the presence of sub-MIC AZM levels (Fig. 5A and B). This suggested that these genes are not required for the development of an AZM-resistant biofilm variant. Furthermore, neither mexEF-oprN nor mexXY appears to be able to act independently of either mexAB-oprM or mexCD-oprJ, as evidenced by the lack of biofilm-forming ability of PAO238 (Fig. (Fig.4).4). From these data, we concluded that either mexAB-oprM or mexCD-oprJ is necessary and sufficient for the development of a resistant biofilm variant in the presence of sub-MIC AZM levels.

mexA and mexC are expressed in biofilms developed in the presence of sub-MIC AZM levels.

The biofilm studies using various RND efflux pump mutants (described above) demonstrate the significance of mexAB-oprM and mexCD-oprJ for the development of a biofilm in the presence of sub-MIC AZM levels. Temporal and spatial expression of both mexA and mexC during biofilm development was measured using appropriate transcriptional fusions to GFP. Not unexpectedly, mexA-gfp promoter fusions yielded no difference between biofilms developed in the presence or absence of AZM (see Fig. S2 in the supplemental material), due to the constitutive expression of mexAB-oprM in P. aeruginosa (35). In contrast, studies examining WT PAO1 with a chromosomally integrated mexC-gfp fusion (PAO430) revealed that, almost at the outset, microcolonies formed in the presence of 2 μg/ml AZM expressed mexC-gfp (Fig. (Fig.6A).6A). By day 4, the majority of the microcolonies were expressing mexC-gfp. Colocalization analyses of the SCLM images indicate an even more impressive increase in temporal expression. At day 2, approximately 60% of the bacteria were expressing mexC; however, nearly all the bacteria expressed mexC by day 4 (Fig. (Fig.6B).6B). Spatially, mexC expression levels were highest near the middle of the biofilms and decreased near the substratum and the top. Although more pronounced with the thicker biofilms at the day-4 time point, this correlation also held true at the day-2 time point. In the control experiment, PAO430 biofilms not exposed to AZM had a mexC expression level near zero, with values typical of background fluorescence. Quantitative analyses of the SCLM images collected at day 4 yielded total biomass values of 13.5 ± 2.4 μm3/μm2 for PAO430 biofilms exposed to AZM and 15.1 ± 1.0 μm3/μm2 for PAO430 biofilms not exposed to AZM. Furthermore, the average thickness was calculated to be 15.7 ± 3.4 μm for AZM-exposed PAO430 biofilms and 17.5 ± 1.5 μm for unexposed PAO430 biofilms. The average microcolony size at the substratum was calculated and determined to be 271 ± 216 μm2 for the AZM-exposed PAO430 biofilms, while for PAO430 biofilms not exposed to AZM the average size was 125 ± 89 μm2. From these data, we conclude that there is no difference in biofilm characteristics between PAO430 biofilms that developed after 4 days in the absence or presence of AZM.

FIG. 6.
Scanning confocal micrographs of biofilms formed by PAO1 (mexC-gfp) with or without continuous exposure to 2 μg/ml AZM. (A) Biofilms were examined to identify cells expressing mexC-gfp (green signal) compared to total cells (red signal). (B) The ...

The AZM-resistant phenotype (PAO1-BV) constitutively expresses mexC in biofilms.

In another study, we report that PAO1-BV is stable upon repeated subculturing (15). We wanted to determine whether the expression of mexC in the AZM-resistant biofilm phenotype was constitutive, which may account for the stability of the phenotype. Consequently, we evaluated the temporal expression patterns of mexC-gfp in biofilms of PAO430 that had developed in the presence of sub-MIC AZM levels (PAO430-BV). PAO430-BV was inoculated into flow cells and cultivated in the presence or absence of sub-MIC AZM levels, and biofilm development was monitored as previously described. As shown in Fig. Fig.7A,7A, mexC appears to be hyperexpressed in PAO430-BV biofilms, regardless of exposure to sub-MIC AZM levels. In order to quantify this observation, we analyzed the image stacks collected. No statistical difference was noted between the expression patterns of the biofilms cultured in the presence or absence of AZM (Fig. (Fig.7B).7B). This would indicate that once mexC is expressed, it remains expressed in a biofilm, regardless of the presence of drug.

FIG. 7.
Scanning confocal micrographs of PAO1-BV (mexC-gfp) biofilms after 4 days of cultivation in the absence or presence of 2 μg/ml AZM. (A) Biofilms were examined to identify cells expressing mexC-gfp (green signal) compared to total cells (red signal). ...

mexC is only expressed in PAO1-BV biofilms.

In PAO1-BV, once mexCD-oprJ is expressed, it appears to stay upregulated in a biofilm, regardless of the presence or absence of AZM (Fig. (Fig.7B).7B). To evaluate whether the mexCD-oprJ efflux pump is also expressed planktonically during AZM exposure in either the wild type or the biofilm variant, respective strains containing mexC-gfp were exposed to 2 μg/ml AZM and fluorescence was monitored hourly by a fluorimeter. Over the 8-h period, neither strain exhibited GFP fluorescence, with fluorescence values unchanged compared to those of control cultures without AZM (data not shown). The two strains grew equally well based on optical densities in the presence of the drug, thus verifying that the lack of fluorescence was not due to lack of growth in the presence of the antibiotic. Consequently, we conclude that mexC is not expressed planktonically at fluorimetrically detectable levels in the presence of sub-MIC AZM concentrations, even in the AZM-resistant biofilm variant.

Upregulation of mexCD-oprJ is not due to mutation in the negative regulator, nfxB.

It has been shown that expression of mexCD-oprJ can be due to mutation of nfxB, a known negative regulator of this efflux pump (33). mexC was expressed in PAO1-BV biofilms, yet its corresponding negative regulator, nfxB, was also expressed at significant levels (Fig. (Fig.1;1; see also Table S1 in the supplemental material), which might indicate that PAO1-BV contained a mutation in nfxB. Therefore, the sequence of nfxB in PAO1-BV was compared to the sequence in WT PAO1. No differences in the sequence were found, indicating that PAO1-BV nfxB has a normal wild-type sequence. The fact that nfxB is expressed at higher levels in PAO1-BV, which is also hyperexpressing the mexCD-oprJ pump, is as yet unexplained; however, its upregulation is not sufficient to prevent hyperexpression of mexCD-oprJ.

The AZM-resistant phenotype does not confer resistance to other antibiotics.

As is commonly reported, prolonged antibiotic treatment can select for multiple-antibiotic-resistant bacteria. To determine if treatment with sub-MIC AZM levels caused a multiple-resistant phenotype, we analyzed a select group of antibiotics, including the antipseudomonal drugs ciprofloxacin, cefepime, meropenem, and tobramycin, as well as tetracycline. Although there was some variance in MICs for planktonically grown bacteria, the differences were not significant (Table (Table2).2). Additionally, MBC testing in a biofilm did not reveal increased resistance of PAO1-BV to these antibiotics (Table (Table3).3). Specifically, PAO1-BV exhibited stable resistance to tetracycline and increased susceptibility to ciprofloxacin. This was surprising, since fluoroquinolones are substrates of the MexCD-OprJ pump, and we would have expected an increase in ciprofloxacin resistance. Similarly, reintroduction of the wild-type mexCD-oprJ operon into a Δ(mexAB-oprM) Δ(mexCD-oprJ) double mutant augmented resistance to azithromycin but made the complemented strain more susceptible to ciprofloxacin. While at present we cannot explain these findings, these changes in MICs of antibiotics other than AZM suggest that besides MexCD-OprJ upregulation, there are perhaps other changes in PAO1-BV that contribute to the observed drug resistance profiles. Furthermore, although all strains tested contained the pTdK plasmid, which has a carbenicillin resistance marker, this did not affect resistance to cephalosporins. Cephalosporins are not hydrolyzed by the TEM-type beta-lactamase present on pTdK but only by sequential TEM lactamase mutants or metallo-beta-lactamases (46).

Individual susceptibilities of planktonically grown P. aeruginosa PAO1 and PAO1 efflux pump mutant strains to selected antibiotics
Individual susceptibilities of P. aeruginosa PAO1 and PAO1 efflux pump mutant-strain biofilms to selected antibiotics


The contribution of RND efflux pumps to planktonic antibiotic resistance is well established (35), but their contributions in biofilms are more speculative. Previous reports examining gene expression in mature P. aeruginosa biofilms exposed to tobramycin (47) or sub-MIC imipenem (2) suggest no significant efflux pump induction. In contrast to these previous reports, we examined the effects of the macrolide AZM on developing biofilms. In a previous study, we reported the development of an AZM-resistant P. aeruginosa biofilm (PAO1-BV) in the presence of sub-MIC AZM levels (15). Using microarrays to analyze the Pseudomonas transcriptome, we found increases in mexCD-oprJ, nfxB, oprN, and mexX as well as in multiple TTSS transcripts in the AZM-resistant biofilm (PAO1-BV) relative to WT PAO1. However, not all of these upregulated genes appear to be involved in the AZM-resistant biofilm phenotype, since neither exsA, mexEF, nor mexXY mutations affect sensitivity to AZM in biofilms. Further assessment using efflux pump mutants revealed that either mexAB-oprM or mexCD-oprJ is sufficient for emergence of AZM resistance in biofilms. Thus, we found a requirement for a double mexAB-oprM mexCD-oprJ null mutant for AZM sensitivity, which emphasizes the phenotypic plasticity of P. aeruginosa and the utilization of redundant systems in this organism. Our findings may also explain the inability of several studies, which examined single null mutants, to find a role for various efflux pumps in biofilm antibiotic resistance (3, 10, 37).

To further assess the contributions of mexAB-oprM and mexCD-oprJ during the development of an AZM-resistant biofilm, temporal and spatial expression was monitored using appropriate fusions. In contrast to mexA, which was expressed uniformly throughout the biofilm in the presence or absence of AZM (see Fig. S2 in the supplemental material), mexC expression was significantly expressed only in the presence of AZM and then appeared to be highest near the center of the biofilm (Fig. (Fig.6).6). In contrast, once a resistant biofilm variant was selected in the presence of AZM, mexC was expressed regardless of the presence of AZM, provided the cells were grown in a biofilm (Fig. (Fig.7).7). The observation that mexCD-oprJ expression is detected only during biofilm development in the presence of AZM raises the exciting possibility that we have identified a biofilm-specific, efflux pump-directed antibiotic resistance mechanism. The temporal expression profiles reported here differ from earlier reports of efflux pumps in biofilms. de Kievit et al. (10) showed that both mexA and mexC were expressed maximally near the substratum in 4-day biofilms exposed to carbenicillin. Differences in the abilities of these chemically dissimilar antibiotics to penetrate the biofilms (42), as well as efflux pump substrate specificities, may account for these profile disparities.

The mechanism underlying the upregulation of mexCD-oprJ in the AZM-resistant biofilm variant remains unclear. Regulation of RND-inducible efflux pumps in P. aeruginosa is thought to occur through a corresponding negative regulator, with single or multiple point mutations in nfxB shown to contribute to overexpression of mexCD-oprJ in laboratory-selected antibiotic-resistant isolates (28, 33). The sequence of nfxB from PAO1-BV was shown to be identical to the WT sequence, indicating that the upregulation of mexCD-oprJ reported in the present study cannot be attributed to a mutation in nfxB. Others have speculated that the regulation of inducible RND efflux pumps may be more complex and include other factors not previously reported (35, 40). One could speculate that exposure to sub-MIC AZM levels may select for a mutation in at least one unknown regulator that may be able to override the effect of nfxB.

Masuda and coworkers (28) reported that a mutation conferring resistance to ofloxacin also conferred resistance to multiple antibiotics. We report that, although we selected an AZM-resistant biofilm variant during sub-MIC exposure to AZM, this strain did not have increased resistance to other antibiotics. While at present we do not know the reasons for this observation, there may perhaps be other changes besides MexCD-OprJ upregulation in PAO1-BV that contribute to the observed drug resistance profiles. For example, it has been documented that physiological changes in biofilms, such as synthesis of periplasmic glucans (25), contribute to the increased drug resistance of biofilms. Despite constitutive low-level expression of mexAB-opM and upregulation of mexCD-oprJ, such changes, together with the conditions employed here, may not allow for increased resistance to the other antibiotics tested.

For the first time, we have shown, using three independent methods (DNA microarrays, RND efflux pump mutants, and transcriptional fusions), that the mexAB-oprM and mexCD-oprJ efflux pumps are required for formation of AZM-resistant biofilms of P. aeruginosa. Recent clinical trials have shown AZM to improve CF patient outcome by as yet unknown mechanisms (11, 36, 48). In an earlier study, we reported that AZM can retard, but does not inhibit, biofilm formation (15). The fact that mexCD-oprJ is critical to AZM antibiotic resistance in developing biofilms may have significance for efforts to inhibit new bacterial colonization in the CF lung. Recently, an efflux pump inhibitor was shown to increase the sensitivity of P. aeruginosa to levofloxacin (24). One could envision a similar combinatorial approach, using inhibitors of the MexAB-OprM and MexCD-OprJ efflux pumps in conjunction with AZM, to potentially decrease colonization of P. aeruginosa in the CF lung or decrease advancement into new regions of the lung. Further work in this area is needed to examine the possibilities of this approach, which might give patients and clinicians another option for treatment of these extremely recalcitrant infections.

Supplementary Material

[Supplemental material]


We thank M. Filiatrault and L. Passador for critical review of the manuscript. We are grateful to D. Frank for providing PAO1 exsA::Ω. We also thank B. Pitts and P. Stewart for sharing their image analysis software.

This work is supported in part by National Institutes of Health (NIH) research grant R37AI37713, Cystic Fibrosis Foundation research grants IGLEWS00V0 and IGLEWS00G0 (to B.H.I.), and NIH training grant 5T32A107362 (to R.J.G.). Further support was provided to H.P.S. by NIH research grant AI051588.


Supplemental material for this article may be found at http://aac.asm.org/.


1. Afessa, B., and B. Green. 2000. Bacterial pneumonia in hospitalized patients with HIV infection: the Pulmonary Complications, ICU Support, and Prognostic Factors of Hospitalized Patients with HIV (PIP) Study. Chest 117:1017-1022. [PubMed]
2. Bagge, N., M. Schuster, M. Hentzer, O. Ciofu, M. Givskov, E. P. Greenberg, and N. Hoiby. 2004. Pseudomonas aeruginosa biofilms exposed to imipenem exhibit changes in global gene expression and beta-lactamase and alginate production. Antimicrob. Agents Chemother. 48:1175-1187. [PMC free article] [PubMed]
3. Brooun, A., S. Liu, and K. Lewis. 2000. A dose-response study of antibiotic resistance in Pseudomonas aeruginosa biofilms. Antimicrob. Agents Chemother. 44:640-646. [PMC free article] [PubMed]
4. Choi, K.-H., J. B. Gaynor, K. G. White, C. Lopez, R. R. Karkhoff-Schweizer, and H. P. Schweizer. 2005. A Tn7-based broad-range bacterial cloning and expression system. Nat. Methods 6:443-448. [PubMed]
5. Chuanchuen, R., K. Beinlich, T. T. Hoang, A. Becher, R. R. Karkhoff-Schweizer, and H. P. Schweizer. 2001. Cross-resistance between triclosan and antibiotics in Pseudomonas aeruginosa is mediated by multidrug efflux pumps: exposure of a susceptible mutant strain to triclosan selects nfxB mutants overexpressing MexCD-OprJ. Antimicrob. Agents Chemother. 45:428-432. [PMC free article] [PubMed]
6. Cormack, B., R. Valdivia, and S. Falkow. 1996. FACS-optimized mutants of the green fluorescent protein (GFP). Gene 173:33-38. [PubMed]
7. Costerton, J., P. Stewart, and E. Greenberg. 1999. Bacterial biofilms: a common cause of persistent infections. Science 284:1318-1322. [PubMed]
8. Crouch Brewer, S., R. G. Wunderink, C. B. Jones, and K. V. Leeper, Jr. 1996. Ventilator-associated pneumonia due to Pseudomonas aeruginosa. Chest 109:1019-1029. [PubMed]
9. de Kievit, T. R., R. Gillis, S. Marx, C. Brown, and B. H. Iglewski. 2001. Quorum-sensing genes in Pseudomonas aeruginosa biofilms: their role and expression patterns. Appl. Environ. Microbiol. 67:1865-1873. [PMC free article] [PubMed]
10. de Kievit, T. R., M. D. Parkins, R. J. Gillis, R. Srikumar, H. Ceri, K. Poole, B. H. Iglewski, and D. G. Storey. 2001. Multidrug efflux pumps: expression patterns and contribution to antibiotic resistance in Pseudomonas aeruginosa biofilms. Antimicrob. Agents Chemother. 45:1761-1770. [PMC free article] [PubMed]
11. Equi, A., I. M. Balfour-Lynn, A. Bush, and M. Rosenthal. 2002. Long term azithromycin in children with cystic fibrosis: a randomised, placebo-controlled crossover trial. Lancet 360:978-984. [PubMed]
12. Feldman, R. B., D. C. Rhew, J. Y. Wong, R. A. Charles, and M. B. Goetz. 2003. Azithromycin monotherapy for patients hospitalized with community-acquired pneumonia: a 3 1/2-year experience from a veterans affairs hospital. Arch. Intern. Med. 163:1718-1726. [PubMed]
13. Frank, D. W. 1997. The exoenzyme S regulon of Pseudomonas aeruginosa. Mol. Microbiol. 26:621-629. [PubMed]
14. Frank, D. W., G. Nair, and H. P. Schweizer. 1994. Construction and characterization of chromosomal insertional mutations of the Pseudomonas aeruginosa exoenzyme S trans-regulatory locus. Infect. Immun. 62:554-563. [PMC free article] [PubMed]
15. Gillis, R. J., and B. H. Iglewski. 2004. Azithromycin retards Pseudomonas aeruginosa biofilm formation. J. Clin. Microbiol. 42:5842-5845. [PMC free article] [PubMed]
16. Guan, L., M. Ehrmann, H. Yoneyama, and T. Nakae. 1999. Membrane topology of the xenobiotic-exporting subunit, MexB, of the MexAB-OprM extrusion pump in Pseudomonas aeruginosa. J. Biol. Chem. 274:10517-10522. [PubMed]
17. Ha, U. H., J. Kim, H. Badrane, J. Jia, H. V. Baker, D. Wu, and S. Jin. 2004. An in vivo inducible gene of Pseudomonas aeruginosa encodes an anti-ExsA to suppress the type III secretion system. Mol. Microbiol. 54:307-320. [PubMed]
18. Heydorn, A., A. T. Nielsen, M. Hentzer, C. Sternberg, M. Givskov, B. K. Ersboll, and S. Molin. 2000. Quantification of biofilm structures by the novel computer program COMSTAT. Microbiology 146:2395-2407. [PubMed]
19. Hoang, T., R. Karkhoff-Schweizer, A. Kutchma, and H. Schweizer. 1998. A broad-host-range Flp-FRT recombination system for site-specific excision of chromosomally-located DNA sequences: application for isolation of unmarked Pseudomonas aeruginosa mutants. Gene 212:77-86. [PubMed]
20. Holloway, B. W., V. Krishnapillai, and A. F. Morgan. 1979. Chromosomal genetics of Pseudomonas. Microbiol. Rev. 43:73-102. [PMC free article] [PubMed]
21. Jaffe, A., J. Francis, M. Rosenthal, and A. Bush. 1998. Long-term azithromycin may improve lung function in children with cystic fibrosis. Lancet 351:420. [PubMed]
22. Knowles, M. R., M. J. Stutts, A. Spock, N. Fischer, J. T. Gatzy, and R. C. Boucher. 1983. Abnormal ion permeation through cystic fibrosis respiratory epithelium. Science 221:1067-1070. [PubMed]
23. Kohler, T., C. van Delden, L. K. Curty, M. M. Hamzehpour, and J. C. Pechere. 2001. Overexpression of the MexEF-OprN multidrug efflux system affects cell-to-cell signaling in Pseudomonas aeruginosa. J. Bacteriol. 183:5213-5222. [PMC free article] [PubMed]
24. Lomovskaya, O., M. S. Warren, A. Lee, J. Galazzo, R. Fronko, M. Lee, J. Blais, D. Cho, S. Chamberland, T. Renau, R. Leger, S. Hecker, W. Watkins, K. Hoshino, H. Ishida, and V. J. Lee. 2001. Identification and characterization of inhibitors of multidrug resistance efflux pumps in Pseudomonas aeruginosa: novel agents for combination therapy. Antimicrob. Agents Chemother. 45:105-116. [PMC free article] [PubMed]
25. Mah, T. F., B. Pitts, B. Pellock, G. C. Walker, P. S. Stewart, and G. A. O'Toole. 2003. A genetic basis for Pseudomonas aeruginosa biofilm antibiotic resistance. Nature 426:306-310. [PubMed]
26. Mansouri, E., S. Blome-Eberwein, J. Gabelsberger, G. Germann, and B. U. von Specht. 2003. Clinical study to assess the immunogenicity and safety of a recombinant Pseudomonas aeruginosa OprF-OprI vaccine in burn patients. FEMS Immunol. Med. Microbiol. 37:161-166. [PubMed]
27. Maseda, H., K. Saito, A. Nakajima, and T. Nakae. 2000. Variation of the mexT gene, a regulator of the MexEF-oprN efflux pump expression in wild-type strains of Pseudomonas aeruginosa. FEMS Microbiol. Lett. 192:107-112. [PubMed]
28. Masuda, N., N. Gotoh, S. Ohya, and T. Nishino. 1996. Quantitative correlation between susceptibility and OprJ production in NfxB mutants of Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 40:909-913. [PMC free article] [PubMed]
29. Masuda, N., E. Sakagawa, S. Ohya, N. Gotoh, H. Tsujimoto, and T. Nishino. 2000. Substrate specificities of MexAB-OprM, MexCD-OprJ, and MexXY-oprM efflux pumps in Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 44:3322-3327. [PMC free article] [PubMed]
30. National Committee for Clinical Standards. 2002. Performance standards for antimicrobial susceptibility testing. 12th informational supplement. M100-S12. National Committee for Clinical Laboratory Standards, Wayne, Pa.
31. Nickel, J. C., I. Ruseska, J. B. Wright, and J. W. Costerton. 1985. Tobramycin resistance of Pseudomonas aeruginosa cells growing as a biofilm on urinary catheter material. Antimicrob. Agents Chemother. 27:619-624. [PMC free article] [PubMed]
32. Nikaido, H. 1994. Prevention of drug access to bacterial targets: permeability barriers and active efflux. Science 264:382-388. [PubMed]
33. Poole, K., N. Gotoh, H. Tsujimoto, Q. Zhao, A. Wada, T. Yamasaki, S. Neshat, J. Yamagishi, X. Z. Li, and T. Nishino. 1996. Overexpression of the mexC-mexD-oprJ efflux operon in nfxB-type multidrug-resistant strains of Pseudomonas aeruginosa. Mol. Microbiol. 21:713-724. [PubMed]
34. Poole, K., K. Krebes, C. McNally, and S. Neshat. 1993. Multiple antibiotic resistance in Pseudomonas aeruginosa: evidence for involvement of an efflux operon. J. Bacteriol. 175:7363-7372. [PMC free article] [PubMed]
35. Poole, K., and R. Srikumar. 2001. Multidrug efflux in Pseudomonas aeruginosa: components, mechanisms and clinical significance. Curr. Top. Med. Chem. 1:59-71. [PubMed]
36. Saiman, L., B. C. Marshall, N. Mayer-Hamblett, J. L. Burns, A. L. Quittner, D. A. Cibene, S. Coquillette, A. Y. Fieberg, F. J. Accurso, and P. W. Campbell III. 2003. Azithromycin in patients with cystic fibrosis chronically infected with Pseudomonas aeruginosa: a randomized controlled trial. JAMA 290:1749-1756. [PubMed]
37. Sanchez, P., J. F. Linares, B. Ruiz-Diez, E. Campanario, A. Navas, F. Baquero, and J. L. Martinez. 2002. Fitness of in vitro selected Pseudomonas aeruginosa nalB and nfxB multidrug resistant mutants. J. Antimicrob. Chemother. 50:657-664. [PubMed]
38. Schweizer, H. P. 2003. Efflux as a mechanism of resistance to antimicrobials in Pseudomonas aeruginosa and related bacteria: unanswered questions. Genet. Mol. Res. 2:48-62. [PubMed]
39. Singh, P. K., A. L. Schaefer, M. R. Parsek, T. O. Moninger, M. J. Welsh, and E. P. Greenberg. 2000. Quorum-sensing signals indicate that cystic fibrosis lungs are infected with bacterial biofilms. Nature 407:762-764. [PubMed]
40. Sobel, M. L., G. A. McKay, and K. Poole. 2003. Contribution of the MexXY multidrug transporter to aminoglycoside resistance in Pseudomonas aeruginosa clinical isolates. Antimicrob. Agents Chemother. 47:3202-3207. [PMC free article] [PubMed]
41. Srikumar, R., T. Kon, N. Gotoh, and K. Poole. 1998. Expression of Pseudomonas aeruginosa multidrug efflux pumps MexA-MexB-OprM and MexC-MexD-OprJ in a multidrug-sensitive Escherichia coli strain. Antimicrob. Agents Chemother. 42:65-71. [PMC free article] [PubMed]
42. Stewart, P. S. 1996. Theoretical aspects of antibiotic diffusion into microbial biofilms. Antimicrob. Agents Chemother. 40:2517-2522. [PMC free article] [PubMed]
43. Stover, C. K., X. Q. Pham, A. L. Erwin, S. D. Mizoguchi, P. Warrener, M. J. Hickey, F. S. Brinkman, W. O. Hufnagle, D. J. Kowalik, M. Lagrou, R. L. Garber, L. Goltry, E. Tolentino, S. Westbrock-Wadman, Y. Yuan, L. L. Brody, S. N. Coulter, K. R. Folger, A. Kas, K. Larbig, R. Lim, K. Smith, D. Spencer, G. K. Wong, Z. Wu, I. T. Paulsen, J. Reizer, M. H. Saier, R. E. Hancock, S. Lory, and M. V. Olson. 2000. Complete genome sequence of Pseudomonas aeruginosa PA01, an opportunistic pathogen. Nature 406:959-964. [PubMed]
44. Wagner, V. E., D. Bushnell, L. Passador, A. I. Brooks, and B. H. Iglewski. 2003. Microarray analysis of Pseudomonas aeruginosa quorum-sensing regulons: effects of growth phase and environment. J. Bacteriol. 185:2080-2095. [PMC free article] [PubMed]
45. Wagner, V. E., R. J. Gillis, and B. H. Iglewski. 2004. Transcriptome analysis of quorum-sensing regulation and virulence factor expression in Pseudomonas aeruginosa. Vaccine 22(Suppl. 1):S15-S20. [PubMed]
46. Walsh, C. 2003. Antibiotics: actions, origins, resistance. ASM Press, Washington, D.C.
47. Whiteley, M., M. G. Bangera, R. E. Bumgarner, M. R. Parsek, G. M. Teitzel, S. Lory, and E. P. Greenberg. 2001. Gene expression in Pseudomonas aeruginosa biofilms. Nature 413:860-864. [PubMed]
48. Wolter, J., S. Seeney, S. Bell, S. Bowler, P. Masel, and J. McCormack. 2002. Effect of long term treatment with azithromycin on disease parameters in cystic fibrosis: a randomised trial. Thorax 57:212-216. [PMC free article] [PubMed]

Articles from Antimicrobial Agents and Chemotherapy are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...