The impact of the competence quorum sensing system on Streptococcus pneumoniae biofilms varies depending on the experimental model

Abstract

Background: Different models for biofilm in Streptococcus pneumoniae have been described in literature. To permit comparison of experimental data, we characterised the impact of the pneumococcal quorum-sensing competence system on biofilm formation in three models. For this scope, we used two microtiter and one continuous culture biofilm system.

Results: In both microtiter models the competence system influences stability and structure of biofilm in the late attachment phase and synthetic competence stimulating peptide (CSP) restored wild type phenotypes in the comC mutants unable to produce the peptide. Early attachment of single cells to well bottoms was found for both systems to be competence independent, while later phases, including microcolony formation correlated to an intact competence system. The continuous culture biofilm model was not affected by mutations in the competence locus, but deletion of capsule had a significant impact in this model.

Conclusions: Since biofilm remains a largely uncharacterised multi-parameter phenotype it appears to be advisable to exploit more than one model in order to draw conclusion of possible relevance of specific genotypes on pneumococcal physiology.

Figure 1

Characteristics of the biofilm model based on exponentially growing cells. Pneumococcal attachment to 96-well microtiter wells after 1:100 dilution in TSB medium was evaluated by viable counts following detachment of cells by sonication. Prior to sonication 18 hour biofilm was washed three times with medium. Panel A reports the effect of the duration of sonication (2 sec. white; 5 sec. light grey; 10 sec dark grey; 30 sec. black) on detachment and survival of pneumococcal cells. Panel B reports biofilm formation of TIGR4 (open bar), FP184 (mutated for comD response regulator; grey bar) and FP218 (mutant of response regulator of the BLP system; black bar) in media supplemented with CSP2, its allelic variant CSP1, BLPTIGR4 or its allelic variant BLPR6. Panel C shows a time course experiment with simultaneous evaluation of turbidity of the planktonic culture (closed circle; OD values of TIGR plotted on right axis) and biofilm counts using encapsulated TIGR4 (square) and its rough isogenic mutant FP23 (triangle). Experiments were performed in TSB supplemented with CSP2 (open symbols) or plain TSB (closed symbols). Turbidity data are form strain TIGR4. Data are from quadruplicate experiments (the small SD are not visible due to log scale of the graph)

Figure 2

Dynamics of biofilm formation in the model based on exponentially growing cells. Biofilm formation in comC and comD mutants in different genetic backgrounds. Biofilm formation in microtiter plates was evaluated in the presence (closed symbols) and absence of CSP (open symbols). Rough wt pneumococci (squares), the mutants for comC encoding CSP (circles) and for comD encoding the CSP-receptor histidine kinase (triangles) were assayed in parallel in a time course experiment. Panel A: Biofilm formation induced by CSP2 in strains derived from strain TIGR4 (comC2, comD2). Mutants assayed were FP23 (non-capsulated TIGR4) and its derivatives FP64 (comC mutant) and FP184 (comD mutant). Panel B: Biofilm formation induced by CSP1 in strains derived from D39 (comC1, comD1). Mutants assayed were RX1 (non-capsulated mutant) and its derivatives FP5 (comC mutant) and FP48 (comD mutant). Data of the twelve time course experiments are from one representative series; repetition showed comparable results.

Figure 3

Impact of competence in the stationary phase type microtiter biofilm model. In this model, biofilm formation was evaluated by both crystal violet staining and analysis at the spectrophotometer. The FP23 strain (non-capsulated TIGR4) was compared with its isogenic mutants in comD (FP231) and comC (FP259). The comC mutant FP259 was also assayed with addition of synthetic CSP to the medium (striped bars). The experiment was performed in BHI and read after 24 (panel A) or 48 hours (panel B) of incubation at 37°C. The differences in biofilm formations between the wt and the comC and comD mutants and between FP259 with and without CSP were statistically significant (p < 0.005). Data are from triplicate experiments.

Figure 4

Microscopy of cells in the stationary phase microtiter biofilm model. The images (40 × magnification) show the attachment of pneumococci to the surface of microtiter plates after 24 hours of incubation. The wt strain TIGR4 (panel A), the comD mutant (panel B), the comC mutant (panel C) and the comC mutant with the addition of CSP2 (panel D) were compared. Biofilm images are taken on crystal violet stained cells observed in bright filed using the 40 × objective of a Leica DM1000 Microscope and a DFC digital camera.

Figure 5

Biofilm formation on coupons in the continuous culture biofilm model. Continuous culture biofilm was analysed for TIGR4 (closed square), its rough mutant FP23 (open square) and the comD mutant FP184 (closed triangle). Bacterial counts in flow through (panel A) and on the coupon (panel B) are from a single experiment while data on biomass (panel C) and the surface area of the biofilm (panle D) are from 15 measurements at each timepoint. Biofilm samples grown on polycarbonate disks were collected at 12, 24, 36, and 48 hours and fixed in formaldehyde. Biofilm was stained with Sybr Green I, a general double stranded DNA stain, and examined with a Zeiss epifluorescence microscope with an ApoTome attachment. Image stacks through the depth of the biofilm were obtained (Z-stacks) from each disk, with 5 stacks per disk and 3 disks per time point and per strain. Biomass in each image stack was enumerated in the COMSTAT image analysis program. Data was transformed by multiplying each point by 10,000 and obtaining the log (base 10) value.