![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Charles T. Campbell Laboratory of Ophthalmic Microbiology, Department of Ophthalmology, UPMC Eye Center, Pittsburgh, Pennsylvania 15213,1 Department of Molecular Genetics and Biochemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15261,2 Department of Pharmacology and Genetics, Dartmouth Medical School, Lebanon, New Hampshire 03756,3 Center for Biological Imaging, Department of Cell Biology and Physiology, University of Pittsburgh, Pittsburgh, Pennsylvania 152614
Received 2 June 2007/ Accepted 26 July 2007
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Recent studies of S. marcescens biofilms have implicated a number of genes as having an important role in attachment to abiotic surfaces and biotic surfaces in a quorum sensing-dependent or -independent manner (30, 31). The genes important for attachment to hydrophilic abiotic surfaces include the luxI homolog swrI, the two-component regulator genes rssA and lipB, a type I secretion system gene, and two quorum-sensing-controlled genes that are important for biofilm structure and exopolysaccharide production, bsmA and bsmB (30, 31). The same study found a role for certain genes, including bsmA, bsmB, and the type I fimbrial adhesin gene fimA, in the attachment of S. marcescens MG1 to immortalized corneal epithelial cells (31). Other studies have revealed the importance of quorum sensing regulation in biofilm maturation and biofilm-dependent protection against predation by protozoans (30, 50, 51). While these reports have provided important insights into S. marcescens biofilm formation and have defined some of the genes required for wild-type biofilm formation, it is likely that other important biofilm factors await discovery.
OxyR is reported to play a role in the oxidative stress response and pathogenicity in several genera of bacteria (4, 15, 19, 25, 26, 32, 34, 45, 52). There is substantial evidence that OxyR also responds to the thiol-disulfide status in Escherichia coli (46, 60). In E. coli, OxyR is posttranscriptionally modified through the formation of a disulfide bond that leads to an altered transcriptional profile. Also in E. coli, fimbria assembly-associated disulfide bridge formation can cause a net reduction of the cell, leading to OxyR-dependent inhibition of the surface-associated antigen 43 protein (55).
Antigen 43 (encoded by the flu gene) in E. coli creates a Velcro-like adhesive protein that mediates cell-cell interactions and is a positive factor in biofilm formation (13, 58). In E. coli, OxyR is a negative regulator of flu, such that oxyR mutants become hyper-biofilm formers and exhibit elevated levels of autoaggregation (13, 58). OxyR is also a negative regulator of biofilm formation in Burkholderia pseudomallei, but it appears to be a positive regulator of biofilm formation in Neisseria gonorrhoeae; however, the mechanism responsible for these phenotypes has not been determined (34, 56).
In addition to antigen 43, fimbriae and fimbria-like adhesins are important in biofilm formation in E. coli and other organisms (5, 6, 12, 14, 20, 33, 49, 55, 59). These large surface structures also play a key role in attachment to epithelial cells and pathogenesis (11, 31, 40). These appendages are also important in pellicle formation and impart growth advantages in some environmental niches (16, 42). A previous study showed that while OxyR is important for regulation of antigen 43, it apparently is not important for the expression of type I fimbriae in E. coli (21).
Here, we identified an oxyR homolog in a screen for genes that contribute to S. marcescens biofilm formation. Similar to what is observed with N. gonorrhoeae, contrary to what is reported for E. coli and B. pseudomallei, S. marcescens oxyR protein exhibit reduced biofilm formation and decreased autoagglutination (cell-cell interactions). The S. marcescens OxyR protein was found to contribute to both oxidative stress survival and the production of functional fimbriae.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
pir bearing the mariner-based transposon delivery plasmid pBT20 (29) was mated with a red-pigmented wild-type S. marcescens strain. Briefly, 0.5-ml portions of overnight cultures of both strains were mixed in a Microfuge tube, spun down, resuspended in 200 µl Tris-EDTA (TE), incubated at 42°C for 10 min, spotted onto an LB plate, and incubated at 30°C for 3 to 12 h. A portion of the cells was taken up with a toothpick and suspended in 1 ml of TE. Aliquots were plated onto LB plates supplemented with gentamicin (10 µg/ml) to select for transposon recipients and with tetracycline (10 µg/ml) to select against E. coli growth. Plates were incubated for 2 to 3 days at 30°C. Pigmented colonies were picked and placed into individual wells of a flat-bottom 96-well dish (Corning Costar catalog no. 3595) in 0.1 ml of LB and incubated at 37°C for 16 to 18 h. A 96-prong multiwell transfer device (Dan-Kar catalog no. MC96) was used to transfer aliquots of mutant libraries into individual wells of round-bottom 96-well polyvinylchloride (PVC) dishes (Falcon catalog no. 3911) containing 85 µl LB, which were incubated at 37°C for 6 or 24 h (11,000 mutants were incubated for 6 h, and 3,000 separate mutants were tested after 24 h). Biofilm formation was then assessed using previously described protocols (43). Transposon insertion sites were mapped using arbitrary PCR (44).
Biofilm assays. Static microplate assays were performed as described by O'Toole and Kolter, with modifications (43). Cells at an A600 of 0.1 were incubated in LB for 6 h (as described below) before biofilm formation was assessed (43). Relative biofilm formation was determined by solubilizing crystal violet with 30% glacial acetic acid and determining the absorbance at 590 nm. Biofilm assays were performed using at least four replicates from triplicate independent cultures for each genotype and were repeated at least two times on different days.
Primary attachment assay.
Bacteria were attached to tissue culture-treated polystyrene (Costar catalog no. 3513) as follows. Bacterial cultures were grown to saturation with aeration at 30°C and washed with phosphate-buffered saline (PBS), and the A600 was adjusted to 1.0 (
5 x 109 cells/ml). One milliliter of LB was added to each well of a 24-well dish, and then 100 µl of the washed culture was added to each well and incubated at room temperature for 20 min. LB and nonattached cells were then removed, and wells were washed five times with 1 ml of PBS. One milliliter of PBS was added, and bacterium-surface interactions were observed by phase-contrast microscopy at a magnification of x400. Digital micrographs were taken, and the number of attached bacteria per field was determined using Metamorph software. At least 14 fields from three independent cultures per strain were used, and the experiment was repeated twice.
Growth curves. The growth rates of the wild type and an isogenic oxyR mutant strain were determined in 5 ml of LB incubated at 30°C with rotation at high speed on a TC-7 tissue culture roller (New Brunswick, Edison, NJ). Triplicate 100-µl aliquots were removed at various time points, and the culture turbidity (A600) was determined using a plate reader (Synergy HT; Bio-Tek, Winooski, VT). This experiment was performed twice on different days.
Oxidative stress sensitivity assays.
Disk diffusion assays were done by spreading 5 x 108 bacteria on an LB agar plate and placing a sterile 6-mm paper disk (catalog no. BB31039; BBL) on the plate, to which 10 µl of 30% H2O2 (Fisher Scientific catalog no. H325) was added, and incubating the preparation overnight at 30 and 37°C. Four replicates from four independent cultures were used per experiment, and each experiment was repeated at least three times on different days.
Hexidium iodide is a quantitative fluorescent DNA dye that stains cells with compromised membranes and is used as a "dead stain." Cultures that were exposed to oxidative stress were spun down, resuspended in saline (0.85% NaCl) with hexidium iodide using a Bac-Light Live/Dead staining kit (Molecular Probes, Eugene, OR), and incubated at room temperature for 15 min. Red fluorescence and the A600 were determined using a plate reader (Synergy 2; Bio-tek, Winooski, VT). Numbers of relative fluorescence units (RFU) were determined by dividing the red fluorescence units by the A600.
Fimbria activity assays. Timed agglutination of yeast cells was performed by placing 25 µl of bacteria that had been washed and resuspended in PBS and normalized (A600, 2.0) onto a glass microscope slide (Corning, Corning, NY). Simultaneously, 25 µl of Saccharomyces cerevisiae (Sigma, product no. YSC2, 2% [wt/vol] in PBS) was added to the slide along with the bacteria, and a timer was started. Concurrently, a second researcher turned the orbital shaker on to mix the bacteria and yeast cells (Lab-line model 2309; setting 10). The researcher with the timer determined the moment at which visible aggregates appeared and was unaware of the genotype of the experimental strain. For each genotype, individual single colonies were used to make three cultures that were tested. The experiment was performed at least three times on separate days for each genotype.
Yeast agglutination was measured spectrophotometrically by mixing 1.5 ml PBS with 500 µl of yeast cells (as described above) and 400 µl of bacteria in PBS (A600, 1.0). The cultures were vortexed at moderate speed for 5 s and allowed to stand at room temperature for 10 min. Three 100-µl samples were then removed from the upper aqueous phase of each tube, and the A600 was determined; then the tubes were vortexed for 30s at high speed, three 100-µl samples were removed from the upper aqueous phase of each tube, and the A600 was determined. The percentage of agglutination was determined as follows: 100 x (1 – A600 before vortexing/A600 after vortexing). For each genotype, single colonies were used to make three cultures that were tested. The experiment was performed at least three times on separate days for each genotype.
Plasmid construction and genetic manipulations. Chromosomal DNA preparations were made using a kit (Gentra catalog no. D5500A). DNA was amplified using a high-fidelity polymerase (TripleMaster; Eppendorf), except where noted below, and was cloned using either T4 DNA ligase (Quick Ligase; New England Biolabs) or yeast in vivo cloning (57).
Targeted mutagenesis of oxyR was carried out using a single-crossover strategy. A 426-bp internal fragment of oxyR was amplified using Taq DNA polymerase (New England Biolabs, Beverly, MA), cloned into pCR2.1 (Invitrogen), and subcloned into pMQ118, a suicide vector in S. marcescens (R. M. Q. Shanks and G. A. O'Toole, unpublished results). The resulting plasmid, pRMQS133, was introduced into S. marcescens via conjugation as described above, except that kanamycin was used as the selectable marker. pRMQS133 chromosomal integrations were verified by PCR. The resulting oxyR-2 mutation has an internal duplication and has pMQ118 inserted at bp 435 (the total length of oxyR is 936 bp).
A wild-type revertant of oxyR-2 was created by growing the oxyR-2 mutant to saturation four times in the absence of antibiotic selection to allow a recombination event to loop out the integrated plasmid. The bacterial culture was streaked to obtain single colonies on LB plates, and replicates were generated on LB plates and LB plates containing kanamycin with a replica block. The wild-type oxyR status of kanamycin-sensitive isolates was confirmed by PCR and phenotype. One isolate was designated RSS4 and used as a control in further experiments (Table 1).
pRMQS140 is a medium-copy-number shuttle plasmid bearing a wild-type copy of oxyR and the adjacent gene prgL. The S. marcescens DNA was amplified using a high-fidelity polymerase and was cloned into pMQ131 using yeast in vivo cloning (57).
pRMQS143, a p15a-based shuttle plasmid bearing a wild-type oxyR open reading frame under control of the PBAD promoter, was made using yeast in vivo cloning (57). The oxyR gene was amplified from the wild-type S. marcescens WT-D strain and recombined into pMQ125 under control of the PBAD promoter.
pRMQS145 and pRMQS147 are similar to pRMQS143 but contain versions of oxyR with point mutations that result in C205S and H204R mutations, respectively. These plasmids were made using yeast in vivo cloning techniques by recombining two PCR amplicons with the pMQ125 vector. The amplicons were made using primers that were engineered with the desired point mutation. These constructs were verified by sequencing.
pRMQS169 was designed to place the chromosomal fimABCD operon under control of Plac through an integration event. A 335-bp region of fimA including 20 bp upstream of the start codon was cloned 30 bp downstream of the Plac promoter in pMQ118 using techniques described above.
Microscopy. Phase-contrast microscopy and fluorescent microscopy were performed using a Nikon Eclipse TE2000-U microscope equipped with a CoolSnap HQ charge-coupled device camera (Photometrics, Tucson, AZ), and images were acquired using either RS Image or Metamorph software. Confocal laser scanning microscopy was performed with bacteria grown under flow conditions, using the Kadori System 2 method as previously described (35) except that we used LB diluted 10-fold in water, glass coverslips were inserted in the wells, and the flow rate was 16.5 ml per h. The inoculum was 108 bacteria from a stationary-phase culture. The growth medium was removed at 24 h, and coverslips were resuspended in PBS and stained with a Bac-Light kit, as described above. Coverslips were dipped three times in PBS and were mounted on hanging drop microscope slides filled with PBS to maintain hydration. Images were acquired with an Olympus FV1000 confocal system equipped with an Olympus IX81 microscope, 405/458/488/515/543/633 nm lasers, and Fluoview software. The initial acquisition was at 12 bits (512 x 512; dwell time, 10 ms/pixel) with an NA 0.85 x 20 oil objective and a z spacing of 1.47 µm between stacked images. Three-dimensional reconstructions were generated in Fluoview at 256 x 256 resolution.
Bacterial samples were prepared for transmission electron microscopy (TEM) in one of two ways. Test tubes (25 by 150 mm) containing 5 ml of LB were inoculated with bacteria and incubated at 30°C. The resulting cultures were classified as "agitated" or "static" cultures based on the growth conditions. Agitated culture tubes were incubated on a TC-7 tissue culture roller (New Brunswick Scientific, New Jersey) rotating at maximum speed, and samples were taken at 20 to 24 h, whereas static cultures were placed vertically in a rack and incubated for 48 to 72 h without agitation. Samples were washed and resuspended in PBS to a final A600 of 1.0. Cells were incubated with a Formvar-coated copper grid, negatively stained with 1% uranyl acetate, and observed with a Jem-1210 electron microscope.
RNA isolation. Strains were grown to stationary phase overnight and then subcultured 1:100 and grown until the A600 was 1. Each 5-ml liquid culture was then mixed with 15 ml of TriReagent LS (Molecular Research Center, Inc.) and frozen at –80°C until RNA was precipitated. RNA was isolated using the TriReagent LS protocol. After a 5-min incubation at room temperature, 4 ml of chloroform was added to each tube, which was then shaken vigorously and incubated at room temperature for 5 min. The mixtures were transferred into 15-ml gel phase lock tubes (Eppendorf) and centrifuged. Each aqueous layer was then mixed with isopropanol and glycogen prior to overnight precipitation at –80°C. The tubes were thawed and centrifuged at 1,900 x g for 60 min at 4°C. The pellets were washed with 80% ethanol and dried prior to suspension in nuclease-free water (USB). The quantity of RNA was determined using a DU800 spectrophotometer (Becton Dickenson), and the quality was determined using an Agilent 2100 Bioanalyzer. The RNA was treated with DNase to remove any contaminating DNA using a Turbo DNA-free kit (Ambion). In short, 50 µg of nucleic acid was mixed with 0.1 volume of 10x Turbo DNase buffer, 1 µl of DNase, and nuclease-free water to obtain a final volume of 50 µl. The mixture was incubated at 37°C for 30 min, and then an additional 1 µl of DNase was added and the mixture was incubated for 30 min at 37°C. The reaction mixture was inactivated by using the manufacturer's protocol, with additional centrifugation to ensure complete removal of the inactivation reagent. The remaining nucleic acids were precipitated with isopropanol and ammonium acetate. The resulting pellet was washed with 80% ethanol and dried before it was suspended in nuclease-free water. The quantity and quality were assessed as described above.
PAGE analysis of surface proteins. Surface proteins from cells grown overnight on LB plates at 30°C were prepared as previously described (31). Trichloroacetic acid-precipitated proteins were washed twice with cold acetone, dried, resuspended in 10 µl of TE and 15 µl of polyacrylamide gel electrophoresis (PAGE) sample buffer, boiled, and run on an 8 to 16% gradient gel. Identification of proteins by mass spectroscopy was performed by the University of Pittsburgh Genomics and Proteomics Core Laboratories.
Real-time PCR.
DNase-treated RNA (1 µg) was used in the cDNA synthesis reaction. Duplicate RNA samples were incubated with random primers (catalog no. 48190-011; Invitrogen) and deoxynucleoside triphosphates for 5 to 10 min at 70°C and cooled on ice. To one of the duplicates, SuperScript III reverse transcriptase (RT) (Invitrogen), 0.1 M dithiothreitol, and 5x First Strand buffer were added according to the manufacturer's directions. For the other duplicate, water was substituted for the SuperScript III RT. The samples were incubated for 60 to 120 min at 50°C. The cDNA was then diluted with nuclease-free water (UBS) to obtain a final dilution of 1:5,000 for use in the real-time PCR. Diluted cDNA template (5 µl) was added to a 96-well PCR plate (Bio-Rad). Separate master mixtures were prepared for each primer set according to manufacturer's protocol (iQ SYBR green Supermix; Bio-Rad). A two-step program was used with a Bio-Rad iQ5 real-time PCR machine. The data are the average ± standard error fold changes relative to the wild type for three biological replicates. To calculate the fold change, the average cycle threshold (CT) for the control gene (rplU) was subtracted from the average CT for the experimental gene (fimC or fimA), which resulted in the 
CT. The wild-type expression CT was subtracted from the mutant strain CT to determine the CT, which was transformed (
), giving the average fold change relative to the wild type. The primer sequences are as follows: F-rplU, GCTTGGAAAAGCTGGACATC; R-rplU, TACGGTGGTGTTTACGACGA; F-fimA-RT, ACTACACCCTGCGTTTCGAC; R-fimA-RT, GCGTTAGAGTTTGCCTGACC; F-fimC-RT, AAGATCGCACCGTACAAACC; and R-fimC-RT, TTTGCACCGCATAGTTCAAG.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
OxyR is required for S. marcescens biofilm formation and initial attachment. Complementation analysis of the oxyR-1 mutant was performed using a plasmid bearing a wild-type copy of oxyR (Table 1). The biofilm defect of the oxyR-1 mutant was fully complemented by wild-type oxyR added in trans (Fig. 2A and Table 2). This result is consistent with the hypothesis that the transposon mutation of oxyR is responsible for the biofilm defect of strain 50A3 (oxyR-1) (Table 2). Similar patterns of biofilm formation were obtained using tissue culture-treated polystyrene as a substrate rather than PVC for the wild-type, oxyR-1, and oxyR-3 strains (Table 2). Tissue culture-treated polystyrene is a hydrophilic surface, whereas PVC is relatively hydrophobic. A persistent biofilm defect on both PVC and polystyrene suggests that an OxyR-dependent adhesive factor(s) mediates attachment to diverse substrates.
|
Confocal microscopic analysis of the oxyR-1 mutant and the wild type grown under flow conditions confirmed the dramatic reduction in biofilm formation by the oxyR-1 mutant relative to the wild type seen with the microtiter plate assay. At 24 and 48 h wild-type cells produced very large microcolonies, whereas the oxyR-1 mutant exhibited no biofilm architecture whatsoever, and only scattered cells were associated with the surface (Fig. 2B and data not shown). Similar results were obtained with epifluorescence microscopy of biofilms grown under static conditions on both PVC and glass coverslips (data not shown). The decrease in attached cells suggests that there is a defect in the initial attachment of the bacteria to the surface.
The initial attachment of the wild-type, oxyR-1, and oxyR-3 strains and the oxyR-1 strain bearing pRMQS140 (pMQ131 plus oxyR) to polystyrene was determined after 20 min of incubation with a surface. After nonattached cells were washed away, the number of bacteria attached per microscopic field was reduced approximately 1,000-fold for the oxyR mutants. By supplying wild-type oxyR in trans, full levels of attachment were restored, suggesting that OxyR contributes to biofilm formation through promoting the initial attachment of bacteria to surfaces (Fig. 2C). We noted that whereas cells of the wild-type and complemented strains were associated with the surface along their longitudinal axis, the few cells observed for the oxyR mutant strains were polarly attached and spinning as if they were attached by a flagellum. The lack of longitudinal attachment of oxyR mutants compared to the wild type is characteristic of a defect in the "reversible-to-irreversible" attachment step of biofilm formation, as exhibited by lapA, sadB, and certain pho regulon mutants of Pseudomonas species (9, 22, 39). Sauer and colleagues first characterized the "reversible-to-irreversible" attachment switch as a developmental step in early biofilm formation after primary attachment to a substrate, which may represent bacteria sensing the environment (53). This phenomenon was quantified by determining the ratio of the number of nonmotile, longitudinally attached cells per field to the total number of cells per field using microscopy. oxyR-1 mutants exhibited a significant reduction in irreversibly attached cells (23.3% ± 8.7% compared with 96.9% ± 0.5% for the wild type; n > 700 cells for each strain). A similar reluctance for longitudinal surface association with the surface was exhibited by the oxyR-1 mutant under flow conditions at 24 h; less than 9% of surface-associated oxyR-1 cells were longitudinally attached, whereas the wild-type strain coated the surface.
oxyR gene is required for hydrogen peroxide resistance. To determine whether the S. marcescens OxyR homolog has a conserved function, its role in oxidative stress was assessed using a disk diffusion assay. Hydrogen peroxide was employed as an oxidative stress response agent. The oxyR-1 mutant exhibited elevated sensitivity to H2O2 compared to the isogenic wild-type strain (Table 3). The oxyR-2 and oxyR-3 mutants were also hypersensitive to H2O2. Similar trends were seen at 30 and 37°C, suggesting that OxyR is important at both environmental and body temperatures (Table 3).
|
Mutation of oxyR does not result in reduced growth rates or a defect in swimming motility. The reduced level of biofilm formation exhibited by oxyR-1 mutants could be a result of a lower growth rate. The growth rates of the oxyR-1 mutant and the isogenic wild-type strain were assessed. No difference in the growth rates was observed (Fig. 3A).
|
The oxyR-1 and oxyR-2 strains were assessed to determine their surface motility (swarming), a group activity akin to biofilm formation in which populations of bacteria move in groups to form finger-like projections. A radical difference was observed. Whereas the wild-type strain was able to cover much or all of a plate in 48 h, the isogenic oxyR-1 strain did not form finger-like swarming projections under our experimental conditions (Fig. 3C and data not shown). In a sample experiment, after 24 h of incubation at room temperature on swarming plates, the wild type swarmed a maximum of 13.5 ± 2.0 mm from the point of inoculation, compared to the absence of any swarming for the oxyR-1 mutant (n = 5 plates per strain).
OxyR is important for fimbria-associated phenotypes. In the biofilm screen, several mutations leading to biofilm-defective phenotypes mapped to predicted fimbrial genes. Most of these mutations resulted in severe defects in biofilm formation similar to those seen in the oxyR-1 strain (Table 2 and data not shown). Fimbriae have previously been described as structures that play a role in biofilm formation on abiotic surfaces and primary attachment in other organisms, such as E. coli (7, 49, 59). Recently, phenotypic variants arising from S. marcescens biofilms have been isolated that display elevated levels of piliation and increased abilities to form biofilms, correlating the presence of surface pili to biofilm formation in this species (27). While OxyR has not been described as a protein that plays a role in regulating fimbrial expression in E. coli, it regulates the surface protein antigen 43 (21). Based on these observations, we hypothesized that S. marcescens OxyR positively regulates fimbrial expression, resulting in reduced biofilm formation in the oxyR mutant strains.
Fimbria activity is commonly measured through agglutination of eukaryotic cells, such as S. cerevisiae cells or erythrocytes (24, 37, 38). The oxyR-1 mutant was found to be severely deficient in the ability to agglutinate budding yeast cells (Fig. 4A). An assay that determined the time of appearance of agglutinated cells on a glass slide was performed. We noted that whereas visible aggregates appeared with the wild-type strain in 4.3 ± 0.6 s, no aggregates appeared with the mutant strain during the period of observation (60 s; n 
6) (Fig. 4A). The oxyR-1 mutant with pRMQS140 (poxyR), which provides wild-type oxyR in trans, was able to form aggregates rapidly (2.7 ± 0.4 s; n = 6), indicating that the agglutination defect is linked to oxyR. A similar deficiency was observed with type I fimbrial mutants (fimB and fimC) (no agglutination at 60 s; n 6). The oxyR-2 and oxyR-3 mutants also exhibited a severe yeast agglutination phenotype with no agglutination at 60 s, and the oxyR-2 mutant restored to the wild-type strain (RSS4) agglutinated yeast in 2.9 ± 0.7 s (n 6).
|
Autoagglutination (bacterium-bacterium interactions) of bacterial cells can also be dependent upon type I fimbriae and other pili (10, 41). It was noted that wild-type cultures grown in PG medium became flocculent, whereas the oxyR-1 mutant produced homogeneous, turbid cultures. The extent of autoagglutination in PG medium was determined spectrophotometrically after tubes were allowed to stand without agitation for 1 h. In a representative experiment done in triplicate using three different single colonies, the wild-type culture exhibited 48.14% ± 4.95% autoagglutination and the oxyR-1 mutant exhibited 6.56% ± 6.0% autoagglutination (P < 0.001). Similar results were obtained with bacteria grown in brain heart infusion medium (data not shown).
Electron microscopy of fimbriae and oxyR-1 mutants reveals a lack of surface pili. TEM was utilized to directly determine the effect of fimbriae and oxyR-1 mutations on S. marcescens surface structures. Aliquots of agitated overnight cultures were observed using TEM. The wild-type strain was covered with pilus-like surface projections (60.3% of cells; n = 189) (Fig. 5A), whereas the oxyR-1 mutant strain had no surface structures other than flagella (0.7% of cells; n = 140) (Fig. 5B and C), similar to the fimC (50E2) mutant (2.7% of cells; n = 109) (Fig. 5E). Moreover, oxyR-1 mutants bearing a high-copy-number plasmid with oxyR (pRMQS140) exhibited a high frequency of piliation (95.7% of cells; n = 254) (Fig. 5D). These data suggest that overexpression of oxyR leads to hyperpiliation, while the absence of oxyR leads to the absence of piliation. The appearance of the fimbriae was similar to the appearance of previously reported Serratia fimbriae (2).
|
To assess the importance of OxyR in fimbria production under the most stringent conditions, bacteria were grown in static conditions in which fimbria production is highly stimulated in E. coli (42). Almost all cells from wild-type static cultures were covered with multitudinous fimbriae (98%; n = 101) (Fig. 6A), whereas only 2.6% of the cells from an oxyR-1 culture exhibited a surface projection that resembled a fimbria-like structure (n = 40) (Fig. 6B). The pRMQS140 plasmid was able to complement the fimbria-deficient phenotype conferred by the oxyR-1 mutation, as described above (100% fimbria positive; n = 31) (Fig. 6C and D). Together, these TEM data suggest that oxyR acts as a positive regulator of fimbria expression.
|
OxyR mediates expression of type I fimbriae. The surface-associated proteins of S. marcescens were observed by PAGE analysis to further assess the effect of mutation of oxyR on fimbria production. The surface protein migration pattern exhibited by the wild-type strain used in this study is very similar to that recently reported for strain MG1 (31). A major band migrating at about 20 kDa, which corresponds to the predicted mass of the FimA subunit (18.1 kDa), was observed in the wild-type strain but was absent in the oxyR-1 mutant (Fig. 7A). The band from the wild type was excised and identified by mass spectroscopy to be the FimA protein. Addition of poxyR to the oxyR-1 mutant reproducibly restored the presence of FimA (as verified by mass spectroscopy) (Fig. 7A). The predicted protein was absent in a fimC mutant, as expected for a fimbrial subunit, as the FimA protein would not be secreted in the absence of an usher protein (Fig. 7A). A band corresponding to flagellin (37 kDa) was observed in the tested strains, as expected, and was identified by mass spectroscopy (Fig. 7A).
|
OxyR-independent fimABCD expression restores biofilm formation to oxyR-1 mutants. In has not been determined whether expression of type I fimbriae is sufficient to suppress the attachment-deficient phenotypes of the oxyR-1 mutant. The chromosomal fimABCD operon was placed under control of the Plac promoter to make its expression independent of OxyR through integration of plasmid pRMQS169 (Fig. 8A). To confirm that this construct conferred a Fim+ phenotype to the oxyR-1 strain, we utilized the yeast agglutination assay (Fig. 8B). As a control we integrated pRMQS169 into the chromosome of a fimC mutant. Plac-mediated expression of fimABCD in this strain should not yield functional fimbriae due to the transposon mutation in fimC. The oxyR-1 strain with Plac-fimABCD was able to agglutinate yeast cells, but Plac-fimABCD did not rescue the fimC mutant defects (Fig. 8B and data not shown). Consistent with a model indicating that oxyR mutants are defective in biofilm formation due to severely reduced fimbria production, Plac expression of fimABCD restored wild-type levels of biofilm formation to the oxyR-1 mutant (Fig. 8 C-D).
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The role of an oxidative stress response regulator in biofilm formation is consistent with the importance of environmental context to the choice between biofilm formation and planktonic lifestyles. The environmental signals could originate from the status of plant and animal defense systems, tissue type, competing microorganisms, and a range of other conditions, including sensing of the air-liquid interface (13, 17, 50, 55). Expression of stress response genes has been shown to be upregulated in biofilms, which makes sense as biofilms are highly heterogeneous and some cells are likely to find themselves in inhospitable niches (7). In the case of OxyR, stress signals likely occur before attachment, leading to altered expression of surface adhesins, attachment to a surface, and an increased chance of survival. Consistent with the hypothesis that stress responses are important in biofilm formation, we have also genetically identified three different predicted oxidoreductases and a pspA (phage shock protein) homolog as mediators of S. marcescens biofilm formation (data not shown). The addition of hydrogen peroxide to biofilms did not induce an increase in biofilm formation under our experimental conditions (data not shown). The expression of fimbriae in response to various forms of oxidative stress is currently being studied.
Many of the genes required for biofilm formation in other organisms were not identified in this screen. Flagella are important in the initial steps of biofilm formation in many gram-negative organisms, and the many genes involved in making functional flagella make this a large target when random mutagenesis is performed. Although only 14,000 different candidates were screened, it is likely that a gene important for flagellum function would have been disrupted, as numerous genes are required for building functional flagella. This suggests either that the flagellum machinery was not mutated in our screen or that under our experimental conditions S. marcescens does not require the flagella for biofilm formation.
Another gene expected to be found in our biofilm screen is a homolog of the gene encoding antigen 43. In E. coli, antigen 43 is a major biofilm factor and is negatively regulated by OxyR. A search of the S. marcescens strain Db11 genome at the Sanger Center did not reveal a gene similar to the gene that codes for antigen 43 or antigen 43-related domains. An obvious antigen 43 homolog gene was also absent in the Serratia proteamaculans 568 genome. Recently, it has been shown that oxyR mutants of E. coli have antigen 43-dependent defects in flagellum motility, whereas S. marcescens oxyR mutants are proficient in flagellum-based movement (58). Additionally, oxyR mutants of E. coli exhibit enhanced biofilm formation due to overexpression of antigen 43, whereas S. marcescens oxyR mutants are defective for biofilm formation (13). Together, these data are consistent with the absence of an antigen 43-like factor in S. marcescens. In E. coli, antigen 43 provides a mechanism for cell-cell interactions. Here we provide evidence that fimbriae are important in cell-cell interactions as the oxyR mutant is defective in both cell-cell interactions and fimbria production and the presence of oxyR on a multicopy plasmid stimulates both cell-cell interactions and fimbria production.
S. marcescens biofilm formation likely increases the opportunities of this organism to infect humans through attachment to abiotic surfaces (catheters, contact lens cases, etc.). OxyR is a good candidate for further study because it is important for both the attachment of bacteria to abiotic and biotic surfaces and the ability of the bacteria to withstand hostile environments, giving it a double role in successful pathogenesis.
|
ACKNOWLEDGMENTS |
|---|
This work was supported by the Campbell Lab of Ophthalmic Microbiology, the UPMC Department of Ophthalmology, and an NEI Core Grant for Vision Research (grant EY08098).
|
FOOTNOTES |
|---|
Published ahead of print on 3 August 2007.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
