- Email: [email protected]
Biofilm formation on human airway epithelia by encapsulated Neisseria meningitidis serogroup B
Microbes and Infection 11 (2009) 281e287 www.elsevier.com/locate/micinf
Original article
Biofilm formation on human airway epithelia by encapsulated Neisseria meningitidis serogroup B R. Brock Neil a, Jian Q. Shao b, Michael A. Apicella a,* a
Department of Microbiology, 3-401 BSB, The University of Iowa, 51 Newton Rd., Iowa City, IA 52242, USA b Central Microscopy Research Facility, 85 EMRB, The University of Iowa, Iowa City, IA 52242, USA Received 26 August 2008; accepted 2 December 2008 Available online 10 December 2008
Abstract Neisseria meningitidis is the etiologic agent of meningococcal meningitis. We compared 48-h biofilm formation by N. meningitidis serogroup B strains NMB, MC58, C311 and isogenic mutants defective in capsule formation on SV-40 transformed human bronchial epithelial (HBE) cells in a flow cell. We demonstrated that strains NMB and NMB siaA-D were defective in biofilm formation over glass, and there was a partial rescue of biofilm growth for strain NMB on collagen-coated coverslips at 48 h. We demonstrated all three serogroup B strains form biofilms of statistically equivalent average height on HBE cells as their isogenic capsular mutants. Strain NMB also formed a biofilm of statistically equivalent biomass as the NMB siaA-D mutant on HBE cells at 6 and 48 h. These biofilms are significantly larger than biofilms formed over glass or collagen. Verification that strain NMB expressed capsule in biofilms on HBE cells was demonstrated by staining with 2.2.B, a monoclonal antibody with specificity for the serogroup B capsule. ELISA analysis demonstrated that strains MC58 and C311 also produced capsules during biofilm growth. These findings suggest that encapsulated meningococci can form biofilms on epithelial cells suggesting that biofilm formation may play a role in nasopharyngeal colonization. Ó 2009 Elsevier Masson SAS. All rights reserved. Keywords: Neisseria meningitidis; Capsular polysaccharides; Biofilms; Airway epithelial cells
1. Introduction Neisseria meningitidis is commonly isolated from the human nasopharynx in approximately 10% the healthy human population [1]. N. meningitidis infection can result in sepsis and meningitis which have a mortality rate of 10e15% with antibiotic treatment [2]. The high mortality rate and epidemic nature of the illness have led to the development of effective capsular and capsular-conjugate vaccines for the prevention of this infection [3,4]. These vaccines contain four of the five most common capsular polysaccharides found in N. meningitidis strains. The fifth capsule type, serogroup B, has a polysaccharide structure that mimics human neural cell adhesion molecules and a successful, broadly used vaccine has not been made to this serogroup [5]. * Corresponding author. Tel.: þ1 (319) 335 7807; fax: þ1 (319) 335 9006. E-mail address: [email protected] (M.A. Apicella). 1286-4579/$ - see front matter Ó 2009 Elsevier Masson SAS. All rights reserved. doi:10.1016/j.micinf.2008.12.001
The capsule is required for N. meningitidis resistance to complement-mediated killing and for intracellular survival during infection of tissue culture cells [6,7]. However, it was suggested that capsule might be down-regulated or not expressed due to phase variation during colonization in order for the organism to attach to host tissue [8]. Studies of N. meningitidis attachment and invasion to tissue culture cells frequently utilize unencapsulated bacteria because capsule confers a strong negative charge and hydrophilicity to the organism and may physically occlude Neisserial adhesins from gaining proximity to the host cell surface [9]. Carriage studies demonstrated encapsulated N. meningitidis to be isolated repeatedly over periods of 6 months from the same subjects even though the infected carrier shows no clinical symptoms [10]. A logical explanation for the persistence of N. meningitidis in the nasopharynx could be the presence of a biofilm. However, it was previously reported that presence of the capsular polysaccharide inhibited biofilm
282
R. Brock Neil et al. / Microbes and Infection 11 (2009) 281e287
formation by N. meningitidis in plastic tubing, polystyrene plates, and over glass in flow chambers while unencapsulated strains were able to form biofilms on these surfaces [11,12]. Our studies examined encapsulated N. meningitidis strains for the ability to form a biofilm over human airway cells. Our results showed that wild-type encapsulated N. meningitidis form biofilms on human tissue culture cells in a flow chamber and that the organisms within the biofilm remain encapsulated.
centered on the ends of the chamber at 2 mm depth. A 1 mm thick rubber gasket the size of the acrylic block with a 17 mm 45 mm section removed from the center was placed on top of the chamber bottom. The top piece was acrylic with the same 17 mm 45 mm middle section removed. The underside of the top piece has a 22 mm 50 mm section machined out to the same depth as the coverslip to accommodate the tissue culture cells. Ten screws held the apparatus together.
2. Materials and methods
2.5. Confocal microscopy and in situ immuno-labeling
2.1. Bacteria strains and culture conditions
Tissue culture cells were labeled with cell tracker orange (Molecular Probes, Eugene, OR) prior to chamber inoculation. Use of the Nikon Eclipse 80I confocal microscope, lasers, Comstat analysis, and Volocity image rendering were previously described [17]. For each analysis, we performed a minimum of 3 separate experiments using a minimum of 6 individual biofilms for each growth condition. To label the biofilm in situ, the pump was stopped and media flasks disconnected and bubble traps evacuated. The bubble traps were sequentially filled pumped out with 4% paraformaldehyde [m/ v], 5 ml 1% normal goat serum [v/v] in 1 PBS pH 7.4 to block, 5 ml blocking solution and 1:40,000 antibody 2.2.B to the N. meningitidis serogroup B capsule (a gift from Dr. Wendell Zollinger, Walter Reed Army Institute of Research) and the antibody was allowed to incubate 2 h or overnight. Next, the bubble trap was filled with two washes of PBS, then 5 ml blocking solution with 1:40,000 Cy5 labeled goat anti-mouse (GAM) IgM (Jackson ImmunoResearch, West Grove, PA) and allowed to incubate for 30 min and last washed with two rounds of PBS. The flow rate was 150 ml per min as previously described for Neisseria and all procedures were performed at 37 C [18].
N. meningitidis strains NMB and NMB siaA-D (ST4609 ST-8 complex, cluster A4), MC58 and MC58 siaD (ST-32 ET5), and C311 and C311 siaD (unassigned MLST) were described previously [13e15]. All strains contained the pLES98 plasmid containing the green fluorescence protein (GFP) reporter gene as a marker for the bacterium [16]. Strains were cultivated from frozen stock as described previously with 5 mg/ml chloramphenicol (Sigma, St. Louis, MO) [13]. 2.2. Tissue culture conditions HBE cells were cultured in MEM tissue culture medium plus 10% heat inactivated fetal calf serum (FCS) [v/v] and 2 mM L-glutamine at 37 C in the presence of 5% CO2 (Gibco, Grand Island, NY). HBE cells were trypsinized and placed onto 22 mm 50 mm coverslips coated with 0.5 mg/ml bovine collagen for use in biofilm chambers once growth matured to at least 90% confluency as determined by visual examination. 2.3. Biofilms on SV-40 transformed human bronchial epithelial (HBE) cells N. meningitidis was cultured for 16 h as above then suspended in 0.1 HBE tissue culture medium (MEM þ 10% FCS þ 2 mM L-glutamine) diluted in 1 phosphate buffered saline (PBS) pH 7.4 supplemented with isovitalex [1:100 v/v] and 100 mM sodium nitrite. Cultures were vortexed for 15 s at high speed then shaken at 150 RPM for 30 min to disperse the bacteria (as verified by microscopic inspection) before adjustment to an absorbance of 0.3 at 600 nm. Bacterial suspensions were added to biofilm chambers containing the HBE cell monolayers and allowed to incubate at 37 C for 1 h before initiation of media flow. The flow rate was 150 ml per minute. Biofilms were allowed to mature for either 6 or 48 h before microscopy and Comstat analysis. 2.4. Biofilm apparatus To assess biofilm growth on tissue culture cells, we constructed a flow-through chamber that permitted the insertion of a coverslip coated with live HBE cells. The chamber was composed of three pieces. The bottom was a 35 mm 70 mm 4 mm piece of acrylic with a 17 mm wide, 45 mm long, and 2 mm deep chamber machined out of the block. Media inflow and outflow ports were
2.6. ELISA analysis Three biofilms from each experimental condition were washed from the tissue culture cells and suspended in 4% paraformaldehyde [m/v] for 48 h at 4 C. The paraformaldehyde was removed from the suspension by centrifugation and resuspended in PBS at an absorbance of 0.1 at 600 nm, and 100 ml of this suspension distributed to 96 well plates for ELISA and processed with antibody 2.2.B against capsule as previously described [19]. To determine degree of piliation at the time of chamber inoculation, we coated the ELISA plates with inoculum as described above and probed with a rabbit polyclonal antibody against PilE (a gift from Michael Jennings, University of Queensland, Brisbane, Australia). Each encapsulated strain was compared to its unencapsulated isogenic mutant as well as a pilE mutant of strain C311. Control ELISA analysis against H8 with monoclonal antibody 2C3 was utilized to demonstrate equal loading between strains. 2.7. Statistical analysis For each biofilm condition, 14e18 Z-series were taken on the LSCM and analyzed by Comstat software for average
R. Brock Neil et al. / Microbes and Infection 11 (2009) 281e287
283
height and total biomass (MathWorks, Natick, MA) [20]. Comstat analysis is a comprehensive digital analysis of a Zseries of images from a confocal microscope. It allows statistical analysis of the biomass and average height of the biofilm across the entire Z-series of images. To accomplish this, each Z-series photomicrograph is saved as a series of tiff images that are converted into 8-bit grayscale images in Comstat for pixel analysis [20]. GraphPad Prism (San Diego, CA) was used to calculate unpaired Student’s t-tests and generate graphs from a minimum of 14 Z-series. A P value of 0.05 was considered significant. We performed Student’s t-test on ELISA data and also determined percent survival of HBE cells through trypan blue exclusion counts. 3. Results 3.1. Biofilm formation by N. meningitidis on HBE cells Previous research suggested encapsulated N. meningitidis did not form biofilms on glass [11,12]. To test if a different coated surface would rescue the reported biofilm defect on glass, NMB and NMB siaA-D were inoculated onto collagencoated coverslips. The encapsulated NMB formed a modest biofilm on collagen-coated glass but not on glass alone (P value 0.05) while the unencapsulated strain showed no statistical difference between growth on glass compared to growth on collagen. When comparing biofilm growth of NMB and NMB siaA-D on glass, the unencapsulated strain formed a biofilm of significantly greater average height than the wildtype strain. However, both encapsulated and unencapsulated strains formed biofilms on monolayers of HBE cells whose average height and biomass were not significantly different from each other. Biofilms formed on HBE cells had statistically more biomass and height than the biofilms grown on collagen or glass (Fig. 1). The confocal images in Fig. 2 of biofilms at 6 and 48 h on HBE cells show the bacteria prefer to localize to HBE cells. Trypan blue exclusion staining indicated 98% of HBE cells were alive at the time of inoculation and 82% were alive after 48 h in negative control chambers. For unencapsulated NMB biofilms 73% were alive at 48 h while for the encapsulated strain 60% of the HBE cells were alive at 48 h. The 48-h images in Fig. 2 also demonstrate that HBE cells had lifted off the coverslip when compared to the 6-h biofilm images. Further changes in structure of infected and uninfected HBE cells after 48 h in the flow cell in Supplemental Fig. 1. Comstat analysis of all images in this study for the red channel (HBE) showed no significant difference in the biomass of the HBE cells in chambers inoculated with or without N. meningitidis when biomass of the HBE cells was compared by Student’s t-test. This indicates the each group of Z-stack images were taken with similar amounts of HBE cells even though the infected conditions had a higher percentage of HBE cells that had perished as described above. Comstat analysis of the 6-h biofilms demonstrated no statistically significant difference in biomass or average height of the bacteria (Fig. 3A, B). We repeated 48-h biofilms on HBE cells with N. meningitidis strains MC58, C311 and their isogenic unencapsulated siaD
Fig. 1. Average height (A) and biomass (B) of biofilms formed on HBE cells, collagen, or glass of 48 h biofilms calculated by Comstat analysis of Z-series from confocal microscopy. Error bars are standard error of the mean. (*P value 0.05, **P value 0.005, ***P value 0.0005).
mutants to ensure the biofilm results on HBE cells were not strain specific. Piliation in N. meningitidis was previously shown to contribute to microcolony formation [11], therefore we determined the degree of piliation for each pair of strains (encapsulated and unencapsulated). Piliation of strains NMB, MC58, and C311 and their isogenic capsular mutants was found to be equally piliated when analyzed by ELISA against the whole bacteria for the presence of PilE. Piliation levels were significantly above the background of the pilE mutant of strain C311 (data not shown). As a further control serial dilutions of the inoculum for the biofilm experiments demonstrated an average of 9.6 107 colony forming units per ml of inoculum, which is consistent with biofilm literature. Dapi stained images revealed an average of 9.84 5.0 bacteria per cell for the capsule positive organisms while the Dapi images of the capsule negative bacteria indicated an average of 22.6 8.8 bacteria per cell after the 1-h incubation at the start of the media flow which demonstrates the 1 h static incubation did not significantly burden the HBE cells prior to initiation of media flow. The 2.5 increase in adherence of the unencapsulated bacteria to the HBE cells is not a surprising result, however, Fig 3A, B demonstrated there was no difference in biofilm biomass and average height at 6 h revealing the initial difference in
284
R. Brock Neil et al. / Microbes and Infection 11 (2009) 281e287
Fig. 2. Confocal microscopy of 6-h (A and B) and 48-h biofilms (C and D) from strains NMB (A and C) and NMB siaA-D (B and D) on HBE cells. Each unit represents 64 mm.
adherence did not result in long term differences in biofilm formation. Next, confocal analysis was performed on encapsulated and unencapsulated strains to compare each strain’s ability to form a biofilm for 48 h. The encapsulated strains formed biofilms whose average height was not statistically different from the isogenic unencapsulated mutant when analyzed as described above (Fig. 3C). The unencapsulated strains produced on average more biomass than the encapsulated strains, however this was only statistically significant for strains C311 and MC58 (Fig. 3D). 3.2. Immuno-label of capsule Previous research on meningococcal biofilms indicated that capsule might be a hindrance to biofilm formation over glass [11,12]. We demonstrated similar results when biofilms were grown over glass or collagen. However, the encapsulated and unencapsulated N. meningitidis formed mature biofilms on
HBE cells. N. meningitidis siaD is phase-variable and resulting biofilms could have been formed by organisms that were phase ‘‘off’’ for capsule gene expression. To determine if encapsulated bacteria were responsible for biofilm formation on HBE cells by wild-type bacteria, we performed in situ labeling of strain NMB and NMB siaA-D biofilms with monoclonal antibody 2.2.B. Images in Fig. 4AeD demonstrate specific labeling with capsular antibody 2.2.B to strain NMB when compared to the secondary only images. Strain NMB siaA-D demonstrates no specific binding of anti-capsular antibody 2.2.B as seen in Fig. 4EeH. Biofilms collected from strains MC58, C311, and their siaD mutants showed significant 2.2.B antibody binding to the wild-type strains but not the isogenic capsular mutants by ELISA (Table 1). We cannot exclude the possibility that some members of the encapsulated population were phase-varied ‘‘off’’ but our studies indicated that the predominate phenotype of the encapsulated meningococci in biofilms was phase-varied ‘‘on’’ for capsule production.
R. Brock Neil et al. / Microbes and Infection 11 (2009) 281e287
285
Fig. 3. Average height and biomass (AeB) of 6-h biofilms formed by encapsulated and unencapsulated strain NMB and average height and biomass (CeD) of 48-h biofilms formed by encapsulated and unencapsulated strains of NMB, MC58, and C311 on HBE cells. Error bars denote standard error of the mean. (*P value 0.05, **P value 0.005).
4. Discussion Previous studies concluded N. meningitidis strains that produce capsule do not make biofilms on abiotic surfaces [11,12]. This research used several different strains, including NMB and MC58, which were used in this study, and allude to the possibility that capsular negative phase-variants likely initiate colonization and biofilm formation [11,12]. Our studies demonstrated encapsulated and unencapsulated N. meningitidis form biofilms on HBE cells. For other bacterial species, the literature differs on the role of capsule in biofilm formation over surfaces such as plastic, glass and human extracellular matrix [21,22]. The differences reported for the role of capsule in biofilm formation among bacterial species may be due to capsule composition or due to the surface on which the biofilm forms. This is the first demonstration that N. meningitidis can form biofilms on transformed human airway epithelial cells in a flow chamber system. It is important to note that the HBE cells are SV-40 transformed in which the virus inhibits G1 arrest and apoptosis mechanisms, which allows for persistent cellular proliferation. This is a mixed airway epithelial cell population consisting of a number of different airway cell
types expressing receptors commonly found on the airway surface such as the platelet activating factor receptor and ICAM-1. We saw 30e40% cell death in the N. meningitidis infected biofilm chambers and 20% death in the uninfected chambers during the 48-h period of study. Similar studies with Neisseria gonorrhoeae in these same chambers on transformed human cervical epithelial cells demonstrate essentially no cellular death. At this time it is unknown if the cell death evident with the HBE cells is a function of media composition, the effect of N. meningitidis biofilm on the epithelial cells, or factors related to the chamber environment. In conclusion, encapsulation did not appear to impede biofilm formation in the presence of airway epithelial cells. The formation of biofilms on HBE cells compared to glass or collagen in Fig. 1 emphasizes the need for interaction of encapsulated N. meningitidis with cellular surfaces and/or matrices to form a biofilm. It was previously reported that N. meningitidis capsule aids in intracellular survival and unencapsulated bacteria are significantly more susceptible to complement-mediated killing than encapsulated bacteria, [6,7]. A biofilm of capsule expressing organisms on respiratory tissue would provide a nidus of a relatively large
R. Brock Neil et al. / Microbes and Infection 11 (2009) 281e287
286
Fig. 4. Biofilms formed on HBE cells by strains NMB carrying the GFP plasmid (A and C) and strain NMB siaA-D (E and G) and stained with mouse monoclonal IgM antibody 2.2.B against serogroup B capsule (B and F) and GAM Cy5 secondary antibody (B, D, F, H) imaged in blue. Each unit is 64 mm. [For interpretation of color in this figure legend the reader is referred to web version of the article].
MC58
MC58 siaD
C311
C311 siaD
population of N. meningitidis fit to spread from person to person. This would enhance the ability of the meningococcus to traverse into the intracellular environment and survive in the bloodstream under the right circumstances of host immunity. This study has demonstrated that encapsulated N. meningitidis can form a biofilm on human airway epithelial cells.
0.8677 0.1803 0.015a
0.0560 0.0111
0.9667 0.1765 0.011a
0.0693 0.0137
Acknowledgments
Table 1 ELISA data from 0.1 OD600 biofilm culture dried to EIA plates and probed with antibody 2.2.B against meningococcal serogroup B capsule at a dilution of 1:500.
Average Standard Deviation Paired t-test P-value a
Statistically significant difference.
This work was supported by AI045728 and AI007511.
R. Brock Neil et al. / Microbes and Infection 11 (2009) 281e287
Appendix. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.micinf.2008.12.001. References [1] K.A. Cartwright, J.M. Stuart, D.M. Jones, N.D. Noah, The Stonehouse survey: nasopharyngeal carriage of meningococci and Neisseria lactamica, Epidemiology and Infection 99 (1987) 591e601. [2] A. Sharip, F. Sorvillo, M.D. Redelings, L. Mascola, M. Wise, D.M. Nguyen, Population-based analysis of meningococcal disease mortality in the United States, The Pediatric Infectious Disease Journal 25 (2006) 191e194. [3] D.B. Fishbein, K.R. Broder, L. Markowitz, N. Messonnier, New, and some not-so-new, vaccines for adolescents and diseases they prevent, Pediatrics 121 (2008) S5e14. [4] M.D. Snape, K.P. Perrett, K.J. Ford, T.M. John, D. Pace, L.-M. Yu, J.M. Langley, S. McNeil, P.M. Dull, F. Ceddia, A. Anemona, S.A. Halperin, S. Dobson, A.J. Pollard, Immunogenicity of a tetravalent meningococcal glycoconjugate vaccine in infants: a randomized controlled trial, JAMA 299 (2008) 173e184. [5] J. Holst, Strategies for development of universal vaccines against meningococcal serogroup B disease, Human Vaccines 3 (2007) 290e294. [6] G.A. Jarvis, N.A. Vedros, Sialic acid of group B Neisseria meningitidis regulates alternative complement pathway activation, Infection and Immunity 55 (1987) 174e180. [7] M.R. Spinosa, C. Progida, A. Tala, L. Cogli, P. Alifano, C. Bucci, The Neisseria meningitidis capsule is important for intracellular survival in human cells, Infection and Immunity 75 (2007) 3594e3603. [8] A.-E. Deghmane, D. Giorgini, M. Larribe, J.-M. Alonso, M.-K. Taha, Down-regulation of pili and capsule of Neisseria meningitidis upon contact with epithelial cells is mediated by CrgA regulatory protein, Molecular Microbiology 43 (2002) 1555e1564. [9] A. Kolb-Maurer, A. Unkmeir, U. Kammerer, C. Hubner, T. Leimbach, A. Stade, E. Kampgen, M. Frosch, G. Dietrich, Interaction of Neisseria meningitidis with human dendritic cells, Infection and Immunity 69 (2001) 6912e6922. [10] A. Wilder-Smith, T.M.S. Barkham, S. Ravindran, A. Earnest, N.I. Paton, Persistence of W135 Neisseria meningitidis carriage in returning Hajj Pilgrims: risk for early and late transmission to household contacts, Emerging Infectious Diseases 9 (2003) 123e126.
287
[11] M. Lappann, J.A.J. Haagensen, H. Claus, U. Vogel, S. Molin, Meningococcal biofilm formation: structure, development and phenotypes in a standardized continuous flow system, Molecular Microbiology 62 (2006) 1292e1309. [12] K. Yi, A.W. Rasmussen, S.K. Gudlavalleti, D.S. Stephens, I. Stojiljkovic, Biofilm formation by Neisseria meningitidis, Infection and Immunity 72 (2004) 6132e6138. [13] D.M. Post, M.R. Ketterer, N.J. Phillips, B.W. Gibson, M.A. Apicella, The msbB mutant of Neisseria meningitidis strain NMB has a defect in lipooligosaccharide assembly and transport to the outer membrane, Infection and Immunity 71 (2003) 647e655. [14] M. Virji, K. Makepeace, I.R. Peak, D.J. Ferguson, M.P. Jennings, E.R. Moxon, Opc- and pilus-dependent interactions of meningococci with human endothelial cells: molecular mechanisms and modulation by surface polysaccharides, Molecular Microbiology 18 (1995) 741e754. [15] R.A. Stabler, G.L. Marsden, A.A. Witney, Y. Li, S.D. Bentley, C.M. Tang, J. Hinds, Identification of pathogen-specific genes through microarray analysis of pathogenic and commensal Neisseria species, Microbiology 151 (2005) 2907e2922. [16] J.L. Edwards, M.A. Apicella, I-domain-containing integrins serve as pilus receptors for Neisseria gonorrhoeae adherence to human epithelial cells, Cellular Microbiology 7 (2005) 1197e1211. [17] K.H. Lim, C.E. Jones, R.N. vanden Hoven, J.L. Edwards, M.L. Falsetta, M.A. Apicella, M.P. Jennings, A.G. McEwan, Metal binding specificity of the MntABC permease of Neisseria gonorrhoeae and its influence on bacterial growth and interaction with cervical epithelial cells, Infection and Immunity 76 (2008) 3569e3576. [18] L.L. Greiner, J.L. Edwards, J. Shao, C. Rabinak, D. Entz, M.A. Apicella, Biofilm Formation by Neisseria gonorrhoeae, Infection and Immunity 73 (2005) 1964e1970. [19] R.E. Mandrell, F.H. Azmi, D.M. Granoff, Complement-mediated bactericidal activity of human antibodies to poly alpha 2 / 8 N-acetylneuraminic acid, the capsular polysaccharide of Neisseria meningitidis serogroup B, Journal of Infectious Diseases 172 (1995) 1279e1289. [20] A. Heydorn, A.T. Nielsen, M. Hentzer, C. Sternberg, M. Givskov, B.K. Ersboll, S. Molin, Quantification of biofilm structures by the novel computer program COMSTAT, Microbiology 146 (Pt 10) (2000) 2395e2407. [21] M. Allegrucci, K. Sauer, Characterization of colony morphology variants isolated from Streptococcus pneumoniae biofilms, Journal of Bacteriology 189 (2007) 2030e2038. [22] J.D. Boddicker, R.A. Anderson, J. Jagnow, S. Clegg, Signature-tagged mutagenesis of Klebsiella pneumoniae to identify genes that influence biofilm formation on extracellular matrix material, Infection and Immunity 74 (2006) 4590e4597.
Original article
Biofilm formation on human airway epithelia by encapsulated Neisseria meningitidis serogroup B R. Brock Neil a, Jian Q. Shao b, Michael A. Apicella a,* a
Department of Microbiology, 3-401 BSB, The University of Iowa, 51 Newton Rd., Iowa City, IA 52242, USA b Central Microscopy Research Facility, 85 EMRB, The University of Iowa, Iowa City, IA 52242, USA Received 26 August 2008; accepted 2 December 2008 Available online 10 December 2008
Abstract Neisseria meningitidis is the etiologic agent of meningococcal meningitis. We compared 48-h biofilm formation by N. meningitidis serogroup B strains NMB, MC58, C311 and isogenic mutants defective in capsule formation on SV-40 transformed human bronchial epithelial (HBE) cells in a flow cell. We demonstrated that strains NMB and NMB siaA-D were defective in biofilm formation over glass, and there was a partial rescue of biofilm growth for strain NMB on collagen-coated coverslips at 48 h. We demonstrated all three serogroup B strains form biofilms of statistically equivalent average height on HBE cells as their isogenic capsular mutants. Strain NMB also formed a biofilm of statistically equivalent biomass as the NMB siaA-D mutant on HBE cells at 6 and 48 h. These biofilms are significantly larger than biofilms formed over glass or collagen. Verification that strain NMB expressed capsule in biofilms on HBE cells was demonstrated by staining with 2.2.B, a monoclonal antibody with specificity for the serogroup B capsule. ELISA analysis demonstrated that strains MC58 and C311 also produced capsules during biofilm growth. These findings suggest that encapsulated meningococci can form biofilms on epithelial cells suggesting that biofilm formation may play a role in nasopharyngeal colonization. Ó 2009 Elsevier Masson SAS. All rights reserved. Keywords: Neisseria meningitidis; Capsular polysaccharides; Biofilms; Airway epithelial cells
1. Introduction Neisseria meningitidis is commonly isolated from the human nasopharynx in approximately 10% the healthy human population [1]. N. meningitidis infection can result in sepsis and meningitis which have a mortality rate of 10e15% with antibiotic treatment [2]. The high mortality rate and epidemic nature of the illness have led to the development of effective capsular and capsular-conjugate vaccines for the prevention of this infection [3,4]. These vaccines contain four of the five most common capsular polysaccharides found in N. meningitidis strains. The fifth capsule type, serogroup B, has a polysaccharide structure that mimics human neural cell adhesion molecules and a successful, broadly used vaccine has not been made to this serogroup [5]. * Corresponding author. Tel.: þ1 (319) 335 7807; fax: þ1 (319) 335 9006. E-mail address: [email protected] (M.A. Apicella). 1286-4579/$ - see front matter Ó 2009 Elsevier Masson SAS. All rights reserved. doi:10.1016/j.micinf.2008.12.001
The capsule is required for N. meningitidis resistance to complement-mediated killing and for intracellular survival during infection of tissue culture cells [6,7]. However, it was suggested that capsule might be down-regulated or not expressed due to phase variation during colonization in order for the organism to attach to host tissue [8]. Studies of N. meningitidis attachment and invasion to tissue culture cells frequently utilize unencapsulated bacteria because capsule confers a strong negative charge and hydrophilicity to the organism and may physically occlude Neisserial adhesins from gaining proximity to the host cell surface [9]. Carriage studies demonstrated encapsulated N. meningitidis to be isolated repeatedly over periods of 6 months from the same subjects even though the infected carrier shows no clinical symptoms [10]. A logical explanation for the persistence of N. meningitidis in the nasopharynx could be the presence of a biofilm. However, it was previously reported that presence of the capsular polysaccharide inhibited biofilm
282
R. Brock Neil et al. / Microbes and Infection 11 (2009) 281e287
formation by N. meningitidis in plastic tubing, polystyrene plates, and over glass in flow chambers while unencapsulated strains were able to form biofilms on these surfaces [11,12]. Our studies examined encapsulated N. meningitidis strains for the ability to form a biofilm over human airway cells. Our results showed that wild-type encapsulated N. meningitidis form biofilms on human tissue culture cells in a flow chamber and that the organisms within the biofilm remain encapsulated.
centered on the ends of the chamber at 2 mm depth. A 1 mm thick rubber gasket the size of the acrylic block with a 17 mm 45 mm section removed from the center was placed on top of the chamber bottom. The top piece was acrylic with the same 17 mm 45 mm middle section removed. The underside of the top piece has a 22 mm 50 mm section machined out to the same depth as the coverslip to accommodate the tissue culture cells. Ten screws held the apparatus together.
2. Materials and methods
2.5. Confocal microscopy and in situ immuno-labeling
2.1. Bacteria strains and culture conditions
Tissue culture cells were labeled with cell tracker orange (Molecular Probes, Eugene, OR) prior to chamber inoculation. Use of the Nikon Eclipse 80I confocal microscope, lasers, Comstat analysis, and Volocity image rendering were previously described [17]. For each analysis, we performed a minimum of 3 separate experiments using a minimum of 6 individual biofilms for each growth condition. To label the biofilm in situ, the pump was stopped and media flasks disconnected and bubble traps evacuated. The bubble traps were sequentially filled pumped out with 4% paraformaldehyde [m/ v], 5 ml 1% normal goat serum [v/v] in 1 PBS pH 7.4 to block, 5 ml blocking solution and 1:40,000 antibody 2.2.B to the N. meningitidis serogroup B capsule (a gift from Dr. Wendell Zollinger, Walter Reed Army Institute of Research) and the antibody was allowed to incubate 2 h or overnight. Next, the bubble trap was filled with two washes of PBS, then 5 ml blocking solution with 1:40,000 Cy5 labeled goat anti-mouse (GAM) IgM (Jackson ImmunoResearch, West Grove, PA) and allowed to incubate for 30 min and last washed with two rounds of PBS. The flow rate was 150 ml per min as previously described for Neisseria and all procedures were performed at 37 C [18].
N. meningitidis strains NMB and NMB siaA-D (ST4609 ST-8 complex, cluster A4), MC58 and MC58 siaD (ST-32 ET5), and C311 and C311 siaD (unassigned MLST) were described previously [13e15]. All strains contained the pLES98 plasmid containing the green fluorescence protein (GFP) reporter gene as a marker for the bacterium [16]. Strains were cultivated from frozen stock as described previously with 5 mg/ml chloramphenicol (Sigma, St. Louis, MO) [13]. 2.2. Tissue culture conditions HBE cells were cultured in MEM tissue culture medium plus 10% heat inactivated fetal calf serum (FCS) [v/v] and 2 mM L-glutamine at 37 C in the presence of 5% CO2 (Gibco, Grand Island, NY). HBE cells were trypsinized and placed onto 22 mm 50 mm coverslips coated with 0.5 mg/ml bovine collagen for use in biofilm chambers once growth matured to at least 90% confluency as determined by visual examination. 2.3. Biofilms on SV-40 transformed human bronchial epithelial (HBE) cells N. meningitidis was cultured for 16 h as above then suspended in 0.1 HBE tissue culture medium (MEM þ 10% FCS þ 2 mM L-glutamine) diluted in 1 phosphate buffered saline (PBS) pH 7.4 supplemented with isovitalex [1:100 v/v] and 100 mM sodium nitrite. Cultures were vortexed for 15 s at high speed then shaken at 150 RPM for 30 min to disperse the bacteria (as verified by microscopic inspection) before adjustment to an absorbance of 0.3 at 600 nm. Bacterial suspensions were added to biofilm chambers containing the HBE cell monolayers and allowed to incubate at 37 C for 1 h before initiation of media flow. The flow rate was 150 ml per minute. Biofilms were allowed to mature for either 6 or 48 h before microscopy and Comstat analysis. 2.4. Biofilm apparatus To assess biofilm growth on tissue culture cells, we constructed a flow-through chamber that permitted the insertion of a coverslip coated with live HBE cells. The chamber was composed of three pieces. The bottom was a 35 mm 70 mm 4 mm piece of acrylic with a 17 mm wide, 45 mm long, and 2 mm deep chamber machined out of the block. Media inflow and outflow ports were
2.6. ELISA analysis Three biofilms from each experimental condition were washed from the tissue culture cells and suspended in 4% paraformaldehyde [m/v] for 48 h at 4 C. The paraformaldehyde was removed from the suspension by centrifugation and resuspended in PBS at an absorbance of 0.1 at 600 nm, and 100 ml of this suspension distributed to 96 well plates for ELISA and processed with antibody 2.2.B against capsule as previously described [19]. To determine degree of piliation at the time of chamber inoculation, we coated the ELISA plates with inoculum as described above and probed with a rabbit polyclonal antibody against PilE (a gift from Michael Jennings, University of Queensland, Brisbane, Australia). Each encapsulated strain was compared to its unencapsulated isogenic mutant as well as a pilE mutant of strain C311. Control ELISA analysis against H8 with monoclonal antibody 2C3 was utilized to demonstrate equal loading between strains. 2.7. Statistical analysis For each biofilm condition, 14e18 Z-series were taken on the LSCM and analyzed by Comstat software for average
R. Brock Neil et al. / Microbes and Infection 11 (2009) 281e287
283
height and total biomass (MathWorks, Natick, MA) [20]. Comstat analysis is a comprehensive digital analysis of a Zseries of images from a confocal microscope. It allows statistical analysis of the biomass and average height of the biofilm across the entire Z-series of images. To accomplish this, each Z-series photomicrograph is saved as a series of tiff images that are converted into 8-bit grayscale images in Comstat for pixel analysis [20]. GraphPad Prism (San Diego, CA) was used to calculate unpaired Student’s t-tests and generate graphs from a minimum of 14 Z-series. A P value of 0.05 was considered significant. We performed Student’s t-test on ELISA data and also determined percent survival of HBE cells through trypan blue exclusion counts. 3. Results 3.1. Biofilm formation by N. meningitidis on HBE cells Previous research suggested encapsulated N. meningitidis did not form biofilms on glass [11,12]. To test if a different coated surface would rescue the reported biofilm defect on glass, NMB and NMB siaA-D were inoculated onto collagencoated coverslips. The encapsulated NMB formed a modest biofilm on collagen-coated glass but not on glass alone (P value 0.05) while the unencapsulated strain showed no statistical difference between growth on glass compared to growth on collagen. When comparing biofilm growth of NMB and NMB siaA-D on glass, the unencapsulated strain formed a biofilm of significantly greater average height than the wildtype strain. However, both encapsulated and unencapsulated strains formed biofilms on monolayers of HBE cells whose average height and biomass were not significantly different from each other. Biofilms formed on HBE cells had statistically more biomass and height than the biofilms grown on collagen or glass (Fig. 1). The confocal images in Fig. 2 of biofilms at 6 and 48 h on HBE cells show the bacteria prefer to localize to HBE cells. Trypan blue exclusion staining indicated 98% of HBE cells were alive at the time of inoculation and 82% were alive after 48 h in negative control chambers. For unencapsulated NMB biofilms 73% were alive at 48 h while for the encapsulated strain 60% of the HBE cells were alive at 48 h. The 48-h images in Fig. 2 also demonstrate that HBE cells had lifted off the coverslip when compared to the 6-h biofilm images. Further changes in structure of infected and uninfected HBE cells after 48 h in the flow cell in Supplemental Fig. 1. Comstat analysis of all images in this study for the red channel (HBE) showed no significant difference in the biomass of the HBE cells in chambers inoculated with or without N. meningitidis when biomass of the HBE cells was compared by Student’s t-test. This indicates the each group of Z-stack images were taken with similar amounts of HBE cells even though the infected conditions had a higher percentage of HBE cells that had perished as described above. Comstat analysis of the 6-h biofilms demonstrated no statistically significant difference in biomass or average height of the bacteria (Fig. 3A, B). We repeated 48-h biofilms on HBE cells with N. meningitidis strains MC58, C311 and their isogenic unencapsulated siaD
Fig. 1. Average height (A) and biomass (B) of biofilms formed on HBE cells, collagen, or glass of 48 h biofilms calculated by Comstat analysis of Z-series from confocal microscopy. Error bars are standard error of the mean. (*P value 0.05, **P value 0.005, ***P value 0.0005).
mutants to ensure the biofilm results on HBE cells were not strain specific. Piliation in N. meningitidis was previously shown to contribute to microcolony formation [11], therefore we determined the degree of piliation for each pair of strains (encapsulated and unencapsulated). Piliation of strains NMB, MC58, and C311 and their isogenic capsular mutants was found to be equally piliated when analyzed by ELISA against the whole bacteria for the presence of PilE. Piliation levels were significantly above the background of the pilE mutant of strain C311 (data not shown). As a further control serial dilutions of the inoculum for the biofilm experiments demonstrated an average of 9.6 107 colony forming units per ml of inoculum, which is consistent with biofilm literature. Dapi stained images revealed an average of 9.84 5.0 bacteria per cell for the capsule positive organisms while the Dapi images of the capsule negative bacteria indicated an average of 22.6 8.8 bacteria per cell after the 1-h incubation at the start of the media flow which demonstrates the 1 h static incubation did not significantly burden the HBE cells prior to initiation of media flow. The 2.5 increase in adherence of the unencapsulated bacteria to the HBE cells is not a surprising result, however, Fig 3A, B demonstrated there was no difference in biofilm biomass and average height at 6 h revealing the initial difference in
284
R. Brock Neil et al. / Microbes and Infection 11 (2009) 281e287
Fig. 2. Confocal microscopy of 6-h (A and B) and 48-h biofilms (C and D) from strains NMB (A and C) and NMB siaA-D (B and D) on HBE cells. Each unit represents 64 mm.
adherence did not result in long term differences in biofilm formation. Next, confocal analysis was performed on encapsulated and unencapsulated strains to compare each strain’s ability to form a biofilm for 48 h. The encapsulated strains formed biofilms whose average height was not statistically different from the isogenic unencapsulated mutant when analyzed as described above (Fig. 3C). The unencapsulated strains produced on average more biomass than the encapsulated strains, however this was only statistically significant for strains C311 and MC58 (Fig. 3D). 3.2. Immuno-label of capsule Previous research on meningococcal biofilms indicated that capsule might be a hindrance to biofilm formation over glass [11,12]. We demonstrated similar results when biofilms were grown over glass or collagen. However, the encapsulated and unencapsulated N. meningitidis formed mature biofilms on
HBE cells. N. meningitidis siaD is phase-variable and resulting biofilms could have been formed by organisms that were phase ‘‘off’’ for capsule gene expression. To determine if encapsulated bacteria were responsible for biofilm formation on HBE cells by wild-type bacteria, we performed in situ labeling of strain NMB and NMB siaA-D biofilms with monoclonal antibody 2.2.B. Images in Fig. 4AeD demonstrate specific labeling with capsular antibody 2.2.B to strain NMB when compared to the secondary only images. Strain NMB siaA-D demonstrates no specific binding of anti-capsular antibody 2.2.B as seen in Fig. 4EeH. Biofilms collected from strains MC58, C311, and their siaD mutants showed significant 2.2.B antibody binding to the wild-type strains but not the isogenic capsular mutants by ELISA (Table 1). We cannot exclude the possibility that some members of the encapsulated population were phase-varied ‘‘off’’ but our studies indicated that the predominate phenotype of the encapsulated meningococci in biofilms was phase-varied ‘‘on’’ for capsule production.
R. Brock Neil et al. / Microbes and Infection 11 (2009) 281e287
285
Fig. 3. Average height and biomass (AeB) of 6-h biofilms formed by encapsulated and unencapsulated strain NMB and average height and biomass (CeD) of 48-h biofilms formed by encapsulated and unencapsulated strains of NMB, MC58, and C311 on HBE cells. Error bars denote standard error of the mean. (*P value 0.05, **P value 0.005).
4. Discussion Previous studies concluded N. meningitidis strains that produce capsule do not make biofilms on abiotic surfaces [11,12]. This research used several different strains, including NMB and MC58, which were used in this study, and allude to the possibility that capsular negative phase-variants likely initiate colonization and biofilm formation [11,12]. Our studies demonstrated encapsulated and unencapsulated N. meningitidis form biofilms on HBE cells. For other bacterial species, the literature differs on the role of capsule in biofilm formation over surfaces such as plastic, glass and human extracellular matrix [21,22]. The differences reported for the role of capsule in biofilm formation among bacterial species may be due to capsule composition or due to the surface on which the biofilm forms. This is the first demonstration that N. meningitidis can form biofilms on transformed human airway epithelial cells in a flow chamber system. It is important to note that the HBE cells are SV-40 transformed in which the virus inhibits G1 arrest and apoptosis mechanisms, which allows for persistent cellular proliferation. This is a mixed airway epithelial cell population consisting of a number of different airway cell
types expressing receptors commonly found on the airway surface such as the platelet activating factor receptor and ICAM-1. We saw 30e40% cell death in the N. meningitidis infected biofilm chambers and 20% death in the uninfected chambers during the 48-h period of study. Similar studies with Neisseria gonorrhoeae in these same chambers on transformed human cervical epithelial cells demonstrate essentially no cellular death. At this time it is unknown if the cell death evident with the HBE cells is a function of media composition, the effect of N. meningitidis biofilm on the epithelial cells, or factors related to the chamber environment. In conclusion, encapsulation did not appear to impede biofilm formation in the presence of airway epithelial cells. The formation of biofilms on HBE cells compared to glass or collagen in Fig. 1 emphasizes the need for interaction of encapsulated N. meningitidis with cellular surfaces and/or matrices to form a biofilm. It was previously reported that N. meningitidis capsule aids in intracellular survival and unencapsulated bacteria are significantly more susceptible to complement-mediated killing than encapsulated bacteria, [6,7]. A biofilm of capsule expressing organisms on respiratory tissue would provide a nidus of a relatively large
R. Brock Neil et al. / Microbes and Infection 11 (2009) 281e287
286
Fig. 4. Biofilms formed on HBE cells by strains NMB carrying the GFP plasmid (A and C) and strain NMB siaA-D (E and G) and stained with mouse monoclonal IgM antibody 2.2.B against serogroup B capsule (B and F) and GAM Cy5 secondary antibody (B, D, F, H) imaged in blue. Each unit is 64 mm. [For interpretation of color in this figure legend the reader is referred to web version of the article].
MC58
MC58 siaD
C311
C311 siaD
population of N. meningitidis fit to spread from person to person. This would enhance the ability of the meningococcus to traverse into the intracellular environment and survive in the bloodstream under the right circumstances of host immunity. This study has demonstrated that encapsulated N. meningitidis can form a biofilm on human airway epithelial cells.
0.8677 0.1803 0.015a
0.0560 0.0111
0.9667 0.1765 0.011a
0.0693 0.0137
Acknowledgments
Table 1 ELISA data from 0.1 OD600 biofilm culture dried to EIA plates and probed with antibody 2.2.B against meningococcal serogroup B capsule at a dilution of 1:500.
Average Standard Deviation Paired t-test P-value a
Statistically significant difference.
This work was supported by AI045728 and AI007511.
R. Brock Neil et al. / Microbes and Infection 11 (2009) 281e287
Appendix. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.micinf.2008.12.001. References [1] K.A. Cartwright, J.M. Stuart, D.M. Jones, N.D. Noah, The Stonehouse survey: nasopharyngeal carriage of meningococci and Neisseria lactamica, Epidemiology and Infection 99 (1987) 591e601. [2] A. Sharip, F. Sorvillo, M.D. Redelings, L. Mascola, M. Wise, D.M. Nguyen, Population-based analysis of meningococcal disease mortality in the United States, The Pediatric Infectious Disease Journal 25 (2006) 191e194. [3] D.B. Fishbein, K.R. Broder, L. Markowitz, N. Messonnier, New, and some not-so-new, vaccines for adolescents and diseases they prevent, Pediatrics 121 (2008) S5e14. [4] M.D. Snape, K.P. Perrett, K.J. Ford, T.M. John, D. Pace, L.-M. Yu, J.M. Langley, S. McNeil, P.M. Dull, F. Ceddia, A. Anemona, S.A. Halperin, S. Dobson, A.J. Pollard, Immunogenicity of a tetravalent meningococcal glycoconjugate vaccine in infants: a randomized controlled trial, JAMA 299 (2008) 173e184. [5] J. Holst, Strategies for development of universal vaccines against meningococcal serogroup B disease, Human Vaccines 3 (2007) 290e294. [6] G.A. Jarvis, N.A. Vedros, Sialic acid of group B Neisseria meningitidis regulates alternative complement pathway activation, Infection and Immunity 55 (1987) 174e180. [7] M.R. Spinosa, C. Progida, A. Tala, L. Cogli, P. Alifano, C. Bucci, The Neisseria meningitidis capsule is important for intracellular survival in human cells, Infection and Immunity 75 (2007) 3594e3603. [8] A.-E. Deghmane, D. Giorgini, M. Larribe, J.-M. Alonso, M.-K. Taha, Down-regulation of pili and capsule of Neisseria meningitidis upon contact with epithelial cells is mediated by CrgA regulatory protein, Molecular Microbiology 43 (2002) 1555e1564. [9] A. Kolb-Maurer, A. Unkmeir, U. Kammerer, C. Hubner, T. Leimbach, A. Stade, E. Kampgen, M. Frosch, G. Dietrich, Interaction of Neisseria meningitidis with human dendritic cells, Infection and Immunity 69 (2001) 6912e6922. [10] A. Wilder-Smith, T.M.S. Barkham, S. Ravindran, A. Earnest, N.I. Paton, Persistence of W135 Neisseria meningitidis carriage in returning Hajj Pilgrims: risk for early and late transmission to household contacts, Emerging Infectious Diseases 9 (2003) 123e126.
287
[11] M. Lappann, J.A.J. Haagensen, H. Claus, U. Vogel, S. Molin, Meningococcal biofilm formation: structure, development and phenotypes in a standardized continuous flow system, Molecular Microbiology 62 (2006) 1292e1309. [12] K. Yi, A.W. Rasmussen, S.K. Gudlavalleti, D.S. Stephens, I. Stojiljkovic, Biofilm formation by Neisseria meningitidis, Infection and Immunity 72 (2004) 6132e6138. [13] D.M. Post, M.R. Ketterer, N.J. Phillips, B.W. Gibson, M.A. Apicella, The msbB mutant of Neisseria meningitidis strain NMB has a defect in lipooligosaccharide assembly and transport to the outer membrane, Infection and Immunity 71 (2003) 647e655. [14] M. Virji, K. Makepeace, I.R. Peak, D.J. Ferguson, M.P. Jennings, E.R. Moxon, Opc- and pilus-dependent interactions of meningococci with human endothelial cells: molecular mechanisms and modulation by surface polysaccharides, Molecular Microbiology 18 (1995) 741e754. [15] R.A. Stabler, G.L. Marsden, A.A. Witney, Y. Li, S.D. Bentley, C.M. Tang, J. Hinds, Identification of pathogen-specific genes through microarray analysis of pathogenic and commensal Neisseria species, Microbiology 151 (2005) 2907e2922. [16] J.L. Edwards, M.A. Apicella, I-domain-containing integrins serve as pilus receptors for Neisseria gonorrhoeae adherence to human epithelial cells, Cellular Microbiology 7 (2005) 1197e1211. [17] K.H. Lim, C.E. Jones, R.N. vanden Hoven, J.L. Edwards, M.L. Falsetta, M.A. Apicella, M.P. Jennings, A.G. McEwan, Metal binding specificity of the MntABC permease of Neisseria gonorrhoeae and its influence on bacterial growth and interaction with cervical epithelial cells, Infection and Immunity 76 (2008) 3569e3576. [18] L.L. Greiner, J.L. Edwards, J. Shao, C. Rabinak, D. Entz, M.A. Apicella, Biofilm Formation by Neisseria gonorrhoeae, Infection and Immunity 73 (2005) 1964e1970. [19] R.E. Mandrell, F.H. Azmi, D.M. Granoff, Complement-mediated bactericidal activity of human antibodies to poly alpha 2 / 8 N-acetylneuraminic acid, the capsular polysaccharide of Neisseria meningitidis serogroup B, Journal of Infectious Diseases 172 (1995) 1279e1289. [20] A. Heydorn, A.T. Nielsen, M. Hentzer, C. Sternberg, M. Givskov, B.K. Ersboll, S. Molin, Quantification of biofilm structures by the novel computer program COMSTAT, Microbiology 146 (Pt 10) (2000) 2395e2407. [21] M. Allegrucci, K. Sauer, Characterization of colony morphology variants isolated from Streptococcus pneumoniae biofilms, Journal of Bacteriology 189 (2007) 2030e2038. [22] J.D. Boddicker, R.A. Anderson, J. Jagnow, S. Clegg, Signature-tagged mutagenesis of Klebsiella pneumoniae to identify genes that influence biofilm formation on extracellular matrix material, Infection and Immunity 74 (2006) 4590e4597.