Tetracycline Rapidly Reaches All the Constituent Cells of Uropathogenic Escherichia coli Biofilms

We have developed a method for visualizing Escherichia colicells that are exposed to tetracycline in a biofilm, based ona previous report that liposomes containing the E. coli TetR(B)protein fluoresce when exposed to this antibiotic. By our method,cells devoid of TetR(B) also exhibited tetracycline-dependentfluorescence. At 50 µg of tetracycline ml^-1, planktoniccells of a uropathogenic E. coli (UPEC) strain developed maximalfluorescence after 7.5 to 10 min of exposure. A similar behaviorwas exhibited by cells in a 24- or 48-h UPEC biofilm, as examinedby confocal laser microscopy, regardless of whether they linedempty spaces or occupied densely packed regions. Further, acomparison of phase-contrast and fluorescent images of correspondingbiofilm zones showed that all the cells fluoresced. Thus, allthe biofilm cells were exposed to tetracycline and there wereno pockets within the biofilm where the antibiotic failed toreach. It also appeared unlikely that niches of reduced exposureto the antibiotic existed within the biofilms.

INTRODUCTION

Top

Introduction

Diseases involving bacterial biofilms are generally chronicand difficult to treat, because bacteria in biofilms are moreresistant to antimicrobial agents than their planktonic (free-living)counterparts. Biofilms contribute to several serious diseases,such as endocarditis, cystitis, cystic fibrosis, and infectionsof prosthetic devices (4-6a).

It is not known why biofilms possess greater resistance. Thepossibility that this might be due to a lack of penetrationof antibiotics has received considerable attention. Althoughseveral studies suggest that biofilms permit antibiotic penetration(1, 6, 8, 9, 10, 12, 14), it was not ascertained in these studiesif this involved exposure of all constituent cells to the antibiotic.

We have developed a method to directly visualize cells comingin contact with tetracycline, based on the report that in thepresence of the TetR(B) protein in liposomes (see http://biosafety.ihe.be/AR/Tetracycline/Menu_Tetfor tetracycline resistance determinant nomenclature), thisantibiotic fluoresces when activated by UV light (13). TetR(B)is constitutively synthesized in bacteria containing Tn10 andregulates the expression of the tetA(B) gene that encodes thetetracycline efflux pump (2). We report here that the biofilmsof uropathogenic Escherichia coli (UPEC) and K-12 strains ofE. coli (11) are readily permeable to tetracycline and permitexposure of all the constituent cells to this antibiotic.

MATERIALS AND METHODS

Top

Materials and Methods

Bacterial strains and plasmids. A clinical isolate of uropathogenic E. coli (AMG1), obtainedfrom the Microbiology Section of the Infectious Diseases Departmentof the Stanford Medical Center, and our laboratory E. coli strain,AMS6 (11), were employed. A tetracycline-resistant strain ofAMG1, referred to as AMG2, containing the TetR(B) protein wasgenerated using the Tn10 transposon, and the " ${lambda}$ hop" technique;the latter employs a temperature-sensitive phage ( ${lambda}$ 561) as theTn10 vehicle. Transposition was achieved as described previously(15), using ${lambda}$ ym plates supplemented with 15 µg of tetracyclineml^-1. The selected mutant (AMG2) exhibited normal morphologicaland growth characteristics.

A recombinant pUC18 plasmid was used for TetR(B) overexpression.The plasmid was cleaved with BamHI and KpnI and ligated to aPCR product containing the full tetR(B) gene, a truncated tetA(B)gene encoding the first six amino acids of TetA(B), and thecomplete intergenic region (2). The PCR product was generatedusing DNA from the Tn10-bearing ${lambda}$ 561 phage and the followingprimers: 5'-GCCGGGTACCTTAAGACCCACTTTCACATTT-3' and 5'-GCCGTCTAGAAGCGATCTTTGTCGAACTATT-3'.The resulting recombinant plasmid was electroporated into AMG1,generating the strain AMG3, which overproduced TetR(B). Luria-Bertani(LB) broth, with or without ampicillin (50 µg ml^-1), wasused for cell growth (37°C); tetracycline was purchasedfrom Sigma.

Fluorescence microscopy. Planktonic cell suspensions (4.5 µl; 0.1 A₆₆₀) were mixedwith 0.5 µl of phosphate-buffered saline (PBS) solutionof tetracycline at room temperature on a glass slide to givea tetracycline concentration of 5, 10, 50, or 100 µg ml^-1,as specified. The suspension was overlaid with a no. 1 coverslip,and observed immediately in an Olympus BX-60 upright fluorescentmicroscope (100x oil immersion phase-contrast lens; total magnification,x1,000). Conditions for optimal tetracycline-dependent fluorescencewere determined empirically. For each time point, the cell-boundtetracycline was photoactivated by a short pulse of 360-nm light,followed by excitation of the resulting fluorophore at 490 nmand measurement of the emission at 510 nm; 360- and 490-nm lightswere generated using a mercury arc lamp and DAPI (4',6'-diamidino-2-phenylindole)and fluorescein isothiocyanate filters, respectively. Imageswere acquired with a Hamamatsu C4742CCD digital camera and ImagePro Plus software (version 4.1.0.0), using a 1-s capture timeand a gain value of 70. Phase-contrast images were obtainedunder identical conditions. Some 300 cells were observed foreach experimental condition.

Confocal microscopy. Biofilms were grown at 37°C in Plexiglas flow cells (16),with interior dimensions of 4 by 36 by 1 mm. These were sterilizedwith 70% ethanol, filled with LB broth or (for strain AMG2)LB broth plus ampicillin, and inoculated with overnight culturesto an initial A₆₆₀ of 0.1. After 1 h of incubation, the mediumwas pumped at a flow rate of 2.8 ml h^-1 for 24 or 48 h. Biofilms(24 h) attained a thickness of 15 µm.

To determine tetracycline diffusion through the biofilms, theflow cell was detached from the medium feed assembly and mountedon a Zeiss 510 confocal microscope. Tetracycline in PBS at roomtemperature was added to the specified concentration, and thebacterial cells were imaged at a magnification of x1,000 withan inverted (Pan Neofluar) 100x oil immersion differential interferencecontrast lens. Photoactivation of the cell-bound tetracyclineand the excitation and emission light settings were as above;a scanning argon laser was used to generate the excitation light.Sagittal sections (Z series; vertical slices) were obtainedusing a software positioning system (Zeiss 510 LSM) to markthe top (biofilm-medium interface) and bottom (biofilm-coverslipinterface) of the biofilm. A distance of 0.5 µm betweeneach x-y plane was used. Reconstruction of fluorescence alongthe z axis necessitated some compromise on the resolution ofindividual cells; nevertheless, the approximate location ofthe individual could be discerned.

A 1- and 5-s 360-nm light exposure before each imaging of theplanktonic and biofilm cells, respectively, permitted a snapshotof the cell-complexed tetracycline at a given time. Exposureof the planktonic cells for over 1 s (up to 5 s) of this lightdid not expedite fluorescence development, but it caused cellbleaching. Such bleaching was not observed upon a 5-s exposureof the biofilm cells (presumably because of the biofilm matrixand higher cell density) and permitted faster development offluorescence in the upper layers of the biofilm (see Results).Fluorescence was seen only when both the cells and tetracyclinewere present, and not with either alone.

The image files were opened in Adobe Photoshop (version 6.01)and, after analyzing the histograms, an adjustment level correctionwas created and saved. This adjustment, which enhanced the contrast,was applied to each file with the exception of Fig. 2C, to whicha lower-level correction was applied to prevent excessive brightness.

View larger version (65K):
[in this window]
[in a new window]
FIG. 2. (A) Time-dependent development of tetracycline-induced fluorescence in a 24-h AMG1 biofilm cells in the x-y plane; panels from left to right represent 3, 7.5, 10, and 15 min of exposure, respectively. The images reproduced here are from about the middle of the biofilm but represent the situation throughout the biofilm, except in the uppermost layers (see Results). (B) Biofilm fluorescence in reconstructed z axis; panels from top to bottom represent 3, 7.5, and 10 min of exposure to the antibiotic, respectively. The bottom of the figure represents the biofilm base (coverslip end of the biofilm through which light was beamed). (C) Biofilm fluorescence in reconstructed composite z axis; panels from top to bottom represent 7.5 and 10 min of exposure to the antibiotic, respectively. The bottom of the figure represents the biofilm base, as in Fig. 2B.

Tetracycline sensitivity. Overnight planktonic cultures were diluted to an A₆₆₀ of 0.1in LB medium, allowed to grow to an A₆₆₀ of 0.8, and mixed withtetracycline (50 µg ml^-1); controls did not receive theantibiotic. At indicated intervals, samples were removed andsubjected to serial dilution and viability determination, asdescribed previously (10a). Biofilms (24 h old) also received50 µg of tetracycline ml^-1. At indicated intervals biofilmswere recovered by scraping and vortexing at high speed in 400µl of 0.1 M phosphate buffer (pH 7.0); a Maxi Mix II Vortexmixer (Barnstead/Thermolyne, Dubuque, Iowa) (1) was used. Thesuspensions were serially diluted, and 50 µl of appropriatedilutions were grown on plates with 30 ml of LB agar for viabilitydetermination. As this entailed a 10⁶ or greater dilution, theantibiotic carryover effect was negligible. Controls were treatedidentically except that they were not exposed to the antibiotic.In planktonic and biofilm control samples, the cell count continuedto increase.

RESULTS

Top

Results

Tetracycline sensitivity. At 50 µg of tetracycline ml^-1, the biofilm-grown cellsof the UPEC strain (initial CFU count, 3.2 x 10⁸ ml^-1) weremore resistant, with nearly 90% survival after 1 h of exposurecompared to 2% survival for the planktonic counterparts (initialCFU count, 5.9 x 10⁹ ml^-1). At subsequent time points also,the biofilm cells showed greater survival: for example, therewas 18 and 0.01% survival for biofilm and planktonic cells,respectively, after 4 h of exposure. These data represent anaverage of two independent experiments; variation for individualtime points was within 3%. The MIC of tetracycline at which50% of isolates tested were inhibited for planktonic and biofilmscells of this strain were 100 and 400 ng ml^-1, as determinedby the method of Das et al. (7).

Tetracycline-induced fluorescence of planktonic cells. We determined optimal conditions for detecting cell contact-dependenttetracycline fluorescence by using planktonic cells of strainAMG2 [TetR(B) positive]. Using the range of tetracycline concentrationsspecified above, we found that a concentration of 50 µgml^-1 produced optimal results. Fluorescence was dim below thisconcentration and not enhanced above it (data not shown). Fluorescencedevelopment was time dependent at this concentration, beingdetectable at 3 min and maximal at 7.5 to 10 min (Fig. 1). Phase-contrastimages of the same samples indicated that all cells fluoresced.While the fluorescence was more intense at the cell membrane,the cell interior also appeared to fluoresce (Fig. 1 and 2A).Exponential- and stationary-phase cells produced about equalfluorescence (not shown).

View larger version (7K):
[in this window]
[in a new window]
FIG. 1. Time-dependent development of tetracycline-induced fluorescence in AMG2 planktonic cells. Panels from left to right represent fluorescence after 1, 3, 7.5, or 10 min of exposure to tetracycline, respectively; the last panel is a phase-contrast image of the corresponding region to visualize all the cells in the field (and not the fluorescent ones only). Magnification, x2,000.

Effect of the absence of TetR(B). As stated in the introduction, the presence of TetR(B) is requiredfor tetracycline-induced fluorescence of liposomes (13). Wefound, however, that the planktonic cells of strains AMG1 andAMS6, which are devoid of TetR(B) (data not shown), exhibiteda fluorescence pattern identical to that of AMG2 (Fig. 1); moreover,strain AMG3, which overproduces TetR(B), also generated similarfluorescence (not shown).

Tetracycline diffusion in biofilms. Time-lapse imaging of tetracycline diffusion through the biofilmswas performed by monitoring fluorescence development, usingconfocal microscopy. When tetracycline was added in 400 µlof PBS at 50 µg ml^-1 to 24- or 48-h biofilms, fluorescencedeveloped rapidly as shown by cross (x-y plane) sections. Within3 min, fluorescent cells could be seen throughout the biofilms(except the uppermost layers; see below), indicating thoroughpenetration (Fig. 2A). As was the case with the planktonic cells,maximal fluorescence developed in ca. 10 min, with no furtherincrease at 15 or 20 min (not shown). Similar results were seenwhen tetracycline was added at 500 µg ml^-1 in 30 or 50µl of PBS at the inflow end, except for a ca. 1-min lagin fluorescence development at biofilm distal end. When cellswere exposed to PBS without tetracycline, no fluorescence developed.

x-y plane images from the same biofilm regions were used toconstruct sagittal (x-z plane) images showing fluorescence alongthe z axis (Fig. 2B); from several z sections taken at constantintervals along a horizontal plane, a composite image of fluorescencealong the z axis was obtained (Fig 2C). The biofilm cells adjacentto the glass coverslip (biofilm base) fluoresced more rapidlythan those located at the top, where the antibiotic was added.We suspect that this effect arose from the inability of the360-nm light, which is essential for the development of thefluorophore and is beamed through the coverslip end, to adequatelypenetrate the entire biofilm. That shorter exposures slowedand longer ones expedited fluorescence development in the upperlayers is consistent with this interpretation and with the conclusionthat tetracycline rapidly penetrated the entire biofilm. Figure3 is a representative micrograph of comparison of the phase-contrastand fluorescent images of the above biofilms at correspondingregions, and shows that all the cells, regardless of their density,fluoresced and were exposed to the antibiotic.

View larger version (87K):
[in this window]
[in a new window]
FIG. 3. Representative micrograph of comparison of the phase-contrast and fluorescent images of the above biofilms at corresponding regions. The images reproduced here are from about the middle of the biofilm but represent the situation throughout the biofilm, except the uppermost layers (see Results).

The data presented are for a 24-h biofilm of strain AMG1. Identicalresults were obtained for a 48-h biofilm of this strain and24- or 48-h biofilms of the other strains studied.

DISCUSSION

Top

Discussion

The UPEC biofilm cells were more resistant to tetracycline thantheir planktonic counterparts. We determined whether this increasedresistance was due to the inability of tetracycline to reachall the cells within the biofilm. Although biofilms as a wholehave been shown to allow certain antibiotics to penetrate (1,6, 8, 12, 14), it remains to be determined if this results inexposure of all the cells to the antibiotic. As the polymericsubstance that surrounds biofilms is likely to be heterogeneous,it may be impenetrable in some regions while permitting passagein others. Furthermore, the channels contained in a biofilm(6) may also transport antibiotics, confining exposure mainlyto cells lining the channels.

We developed a method for direct visualization of cells exposedto tetracycline. Although the original report of UV-inducedtetracycline-dependent fluorescence suggested that the TetR(B)protein is required for this phenomenon (13), we found thatE. coli strains devoid of TetR(B) also fluoresced upon tetracyclineexposure. We could thus examine tetracycline penetration inthe wild-type, tetracycline-sensitive, UPEC strain. The observedfluorescence of TetR(B)-deficient strains could be (i) becauseour method differed from that of Sigler et al. (13), (ii) becausethe strains we used contain homologues of TetR(B) (3), (iii)because the fluorescence is due to tetracycline interactionwith the cell membrane, and/or (iv) because the tetracycline-30S ribosomal complex fluoresces.

At 50 µg ml^-1 tetracycline, both the planktonic and biofilmcells developed maximal fluorescence within 10 min; in the caseof the biofilms, this was irrespective of whether the cellslined empty spaces or occupied dense regions. A comparison ofphase contrast and fluorescent images of corresponding regionsindicated that all the cells fluoresced. Thus, there were nopockets within the biofilms where the antibiotic failed to reach.The fact that the time-dependence of fluorescence developmentwas the same in planktonic and biofilm cells may suggest thatat 50 µg ml^-1 tetracycline, the biofilm and the planktoniccells were exposed to the antibiotic to an equal degree, andthat the increased biofilm resistance is not due to niches ofdecreased exposure. But firm conclusions in this latter regardrequire a more quantitative method.

ACKNOWLEDGMENTS

This work was supported by a NASA grant (NAG2-1); G.F. and P.W.were supported by an NIH training grant (NRSA 5T32 AI0732801)and a Dean's fellowship from Stanford University School of Medicine,respectively.

We thank MaryAnn Wijtman for help with image processing.

FOOTNOTES

* Corresponding author. Mailing address: Department of Microbiology and Immunology, Stanford University School of Medicine, Sherman Fairchild D317, 299 West Campus Dr., Stanford, CA 94305. Phone: (650) 725-4745. Fax: (650) 725-6757. E-mail: a.matin{at}stanford.edu.

REFERENCES

Top