C-di-GMP regulates Pseudomonas aeruginosa stress response to tellurite during both planktonic and biofilm modes of growth

Stress response plays an important role on microbial adaptation under hostile environmental conditions. It is generally unclear how the signaling transduction pathway mediates a stress response in planktonic and biofilm modes of microbial communities simultaneously. Here, we showed that metalloid tellurite (TeO32–) exposure induced the intracellular content of the secondary messenger cyclic di-GMP (c-di-GMP) of Pseudomonas aeruginosa. Two diguanylate cyclases (DGCs), SadC and SiaD, were responsible for the increased intracellular content of c-di-GMP. Enhanced c-di-GMP levels by TeO32– further increased P. aeruginosa biofilm formation and resistance to TeO32–. P. aeruginosa ΔsadCΔsiaD and PAO1/plac-yhjH mutants with low intracellular c-di-GMP content were more sensitive to TeO32– exposure and had low relative fitness compared to the wild-type PAO1 planktonic and biofilm cultures exposed to TeO32–. Our study provided evidence that c-di-GMP level can play an important role in mediating stress response in microbial communities during both planktonic and biofilm modes of growth.

Scientific RepoRts | 5:10052 | DOi: 10.1038/srep10052 increase synthesis of EPS matrix, resulting in biofilm formation 10,11 . In contrast, biofilm cells increase their motility and disperse from biofilms when the intracellular c-di-GMP content is low 12,13 . C-di-GMP signaling can be induced by stress conditions such as antimicrobial exposure 14,15 . The impact of c-di-GMP on mediating stress response by microbial communities during both planktonic and biofilm modes of growth remains unclear.
Anthropogenic activities have resulted in serious metal(loid) pollution, especially in industrialized countries and regions. The natural ecosystems are often direct or indirect recipients of toxic metal(loid)s such as TeO 3 2− . Many environmental bacteria including Pseudomonas aeruginosa are capable of surviving in the presence of TeO 3 2− at low concentrations by reducing TeO 3 2− to Te(0) nanomaterials, as a result of either detoxification, redox maintenance or respiration [16][17][18][19] . Although the toxic effects of metal(oild)s on environmental microorganisms at individual cell levels have been extensively studied 20 , little is known about the impacts of metal(loid)s on bacterial social behaviours 21 .
In the present study, we investigated the role of c-di-GMP in mediating stress responses by the opportunistic pathogen Pseudomonas aeruginosa to a toxic metalloid, tellurite (TeO 3 2-). TeO 3 2− is highly toxic to most microbes and had been previously described by Alexander Fleming as an antimicrobial agent 22 . Bacterial cells take up TeO 3 2and subsequently reduce it to tellurium nanoparticles, which can be easily tracked by the black precipitates on the bacterial cell surface. Quantification of intracellular c-di-GMP and proteomic analysis indicated that c-di-GMP levels were induced by TeO 3 2exposure, which enhanced P. aeruginosa TeO 3 2resistance and biofilm formation. SadC and SiaD were found to be important in the induction of c-di-GMP by TeO 3 2exposure. We showed that mutants with low intracellular c-di-GMP content could be outcompeted by the wild-type strain from biofilm and planktonic cultures under metalloid stress condition. 3 2− induced c-di-GMP signaling. Cultivation of different bacterial species in the presence of sub-lethal concentrations of antimicrobial agents is a widely employed method to investigate their stress responses [23][24][25] . The MIC of P. aeruginosa to TeO 3 2− is 100 μg/ml in ABTGC medium. Large aggregates (approximately 1-3 mm) were formed when P. aeruginosa was grown in ABTGC media containing 10 μg/ml TeO 3 2− at 37 °C (Fig. 1a). Further analysis of the TeO 3 2− -induced aggregates by field emission scanning electron microscopy (FE-SEM) and energy-dispersive X-ray spectroscopy (EDX) revealed the presence of tellurium-containing precipitates around the bacterial cells (Fig. 1b,c). No tellurium-containing precipitates were observed for P. aeruginosa cells growing in medium without TeO 3 2− . Thus, the tellurium-containing precipitates might generate conditions of membrane-associated stress for P. aeruginosa cells.

Stress responses of P. aeruginosa to TeO
TeO 3 2− and oxyanions such as selenate/selenite are well known to exert their toxic effects on microorganisms via generation of reactive oxygen species (ROS) 26,27 . We measured the generation of ROS by P. aeruginosa cells exposed to sub-lethal concentrations of TeO 3 2− as well as SeO 3 2− and SeO 4 2− by using the OxiSelect ™ in vitro ROS/RNS assay kit. As anticipated, exposure of P. aeruginosa cells to the TeO 3 2− , SeO 3 2− and SeO 4 2− significantly increased their cytoplasmic ROS content regardless of the nutrient conditions (Fig. 1d).

Proteomic analysis of TeO 3
2− stressed P. aeruginosa cells. Oxidative stress response by P. aeruginosa leading to aggregate formation, recently reported to resemble the biofilm physiology 28 has not been documented. We thus investigated the overall impact of TeO 3 2− on P. aeruginosa cells using a comparative proteomic approach for cells cultivated with and without 10 μg/ml TeO 3 2− . Using a p-value cut-off of 0.05 and a fold change cut-off of 5 (as described in the Materials and Methods), 129 proteins were significantly affected by TeO 3 2− exposure with 64 proteins upregulated (Supplementary table 1) and 65 proteins being down-regulated (Supplementary table 2).
The expression of several of outer membrane associated proteins was induced by TeO 3 2− treatment, including OprQ (PA2760, 28.8-fold), OprI precursor (PA2853, 15-fold), probable outer membrane protein precursor (PA2391, 10.9-fold), OprM (PA0427, 10.5-fold), OprL precursor (PA0973, 9.8-fold), OprD precursor (PA0958, 9.8-fold), OprB (PA3186, 9.7-fold) and OprC (PA3790, 8.1-fold) (Supplementary  table 1). The membrane transporter CdrB of the large extracellular protein CdrA 29 was induced 25.8-fold by exposure to TeO 3 2− (Supplementary table 1). CdrAB expression has been used as a c-di-GMP indicator 30 and reported to promote biofilm formation and auto-aggregation in a Psl polysaccharide dependent manner 29 , and co-immunoprecipitation experiments have clearly shown that CdrA binds to Psl 29 . HPLC analysis showed that P. aeruginosa PAO1 cultivated in ABTGC medium with 10 μg/ml TeO 3 2− treatment had a higher relative intracellular c-di-GMP concentration compared to untreated control samples (approximately 2.5-fold) (Fig. 1e). 3 2− . CdrAB belongs to a family of bacterial proteins secreted by the two-partner secretion system 31 . Recently, two other members of this family, XacFhaB from Xanthomonas axonopodis pv. Citri and FHA from Bordetella pertussis have also been implicated in biofilm formation 32,33 . These large inter-bacterial adhesins may play a key role in establishing structured biofilm communities under stress conditions. The cdrA promoter is positively regulated by the c-di-GMP concentration, and the expression of P cdrA -gfp has been recently used as a biosensor of the intracellular content of c-di-GMP in P. aeruginosa 30 . We tested the expression of the P cdrA -gfp reporter in P. aeruginosa cultures with and without the presence of TeO 3 2− and found that TeO 3 2− exposure significantly increased the expression of fluorescence in a dose dependent manner (Fig. 2a). This result is in accordance with our HPLC quantification and indicates that TeO 3 2− exposure increases the intracellular content of c-di-GMP and that TeO 3 2− induced aggregates might carry physiological traits similar to those of biofilms.

SadC and SiaD contribute to c-di-GMP induction by TeO
Recently, both SadC and SiaD, were shown to be able to transduce an extracellular signal generated by the toxic detergent SDS and catalyze synthesis of c-di-GMP for facilitating biofilm formation by P. aeruginosa 34,35 . The defect environmental signaling Δ sadC and Δ siaD mutants were severely impaired in expression of the P cdrA -gfp reporter in the presence of TeO 3 2− (Fig. 2a). SiaD appears to be more important than SadC for P cdrA -gfp induction by TeO 3 2− since the Δ sadC mutant still displayed a slight induction of P cdrA -gfp by TeO 3 2− (Fig. 2a). Exopolysaccharides are the major EPS components of P. aeruginosa biofilms and are well known to be induced by high intracellular c-di-GMP content in P. aeruginosa. We examined the expression of a lacZ-based biosensor of the Pel synthesis operon (mini-CTX-pel-lacZ 36 ) in P. aeruginosa strains under TeO 3 2− stress. As with P cdrA -gfp fusion, the expression of the pel-lacZ fusion was induced by TeO 3 2− treatment, with SiaD essential for this induction (Fig. 2b). However, there was a slight induction of the pel-lacZ fusion by tellurite even in the Δ sadCΔ siaCD double mutant (Fig. 2b). Consistent with our observation of TeO 3 2− -induced aggregation, P. aeruginosa grown in the presence of TeO 3 2− formed more biofilms than cells grown without TeO 3 2− (Fig. 3). The induction of biofilm formation was dependent on the presence of Pel and Psl polysaccharides (Fig. 3).
Induction of c-di-GMP confers a growth advantage under tellurite exposure during planktonic cultures. Since c-di-GMP signaling was induced by TeO 3 2− exposure, we examined whether induction of c-di-GMP signaling would confer a growth advantage of P. aeruginosa under TeO 3 2− exposure. There was no growth defect of Δ sadC, Δ siaD and Δ sadCΔ siaD mutants under normal growth condition as compared to PAO1 control (Fig. 4a). However, the P. aeruginosa Δ sadC, Δ siaD single or double mutants were more sensitive to TeO 3 2− (Fig. 4b). Similarly, the PAO1/p lac -yhjH mutant, which Relative intracellular c-di-GMP content of PAO1 cultures in ABTGC medium with and without 10 μg/ml TeO 3 2− was quantified by HPLC (e). Means and standard deviations of three replicates are shown. Student's t-test was performed for testing differences between groups. * P < 0.05. contains a PBBRMCS-2 plasmid with a constitutively expressed phosphodiesterase gene yhjH fused to and expressed by the lac promoter and thus has a low intracellular of c-di-GMP content 12 , was also more sensitive to TeO 3 2− (Fig. 4). These results showed that intracellular c-di-GMP content determines the tolerance of P. aeruginosa to TeO 3 2− exposure during planktonic cultures.

Low intracellular c-di-GMP mutants lose fitness under stress during both planktonic and biofilm modes of growth.
When cfp-tagged PAO1 and yfp-tagged Δ sadCΔ siaD mutant strains were combined 1:1 (vol/vol) for planktonic co-cultivation experiments, the wild-type showed higher survival rates and gained a higher level of relative fitness than the Δ sadCΔ siaD mutant in the presence of TeO 3 2− than without TeO 3 2− (Fig. 5a). Since diverse phenotypic and genotypic variants coexist in bacterial biofilms 37,38 , we tested whether TeO 3 2− exposure-induced biofilm formation by high c-di-GMP containing cells would lead to protection of mutants with low intracellular c-di-GMP content in co-cultures. Here, PAO1 displayed a higher relative fitness than the Δ sadCΔ siaD mutant in biofilm co-cultures with and without the presence of TeO 3 2− (Fig. 5b). However, the relative fitness of Δ sadCΔ siaD compared to PAO1 in biofilm co-cultures was slightly higher with the presence of TeO 3 2− than in its absence (Fig. 5b). This suggests TeO 3 2− could potentially induce expression of other DGC harboring proteins in the Δ sadCΔ siaD mutant and partly restore the intracellular c-di-GMP levels and biofilm formation. When we mixed cfp-tagged PAO1 and yfp-tagged PAO1/p lac -yhjH strains 1:1 (vol:vol) for planktonic co-cultivation experiments, the wild-type PAO1 strain gained a higher level of relative fitness than the c-di-GMP depleted PAO1/p lac -yhjH strain with and without exposure to TeO 3 2− (Fig. 6a). Moreover, PAO1/p lac -yhjH was fully outcompeted by PAO1 in biofilm co-cultures supplemented with TeO 3 2− (Fig. 6b). These results suggest that variants with low intracellular c-di-GMP content are unlikely to be protected and maintained by both P. aeruginosa planktonic and biofilm communities when c-di-GMP is required for stress response.

Discussion
Bacterial cells face various types of stress during the colonization of natural environments and hosts. A series of stress response mechanisms has evolved in bacteria to cope with these harmful conditions. One well characterized stringent stress response mechanism is SpoT-mediated ppGpp accumulation, which can be provoked by nutritional stress caused by harmful conditions such as antibiotic treatment and UV irradiation 39 . ppGpp is able to bind directly to the bacterial RNA polymerase and further regulate transcriptional activity of many genes.
In addition to the stringent stress response, bacteria employ a wide range of social behaviors for surviving under unfavorable environmental conditions and these responses also contribute to bacterial pathogenesis 40 . For example, the Staphylococcus aureus agr quorum-sensing system is involved in the oxidative stress response 41 . Biofilm formation is evoked as a stress response mechanism by a wide range of bacteria 42 . It involves encasing bacterial cells inside the densely packed EPS matrix components and attaching firmly to biotic and abiotic surfaces. Biofilms are up to 1,000 times more resistant to antimicrobial agents compared to their planktonic counterparts 43 .
Recently, bacteria were found to form floating biofilm-resembling aggregates that are resistant to antimicrobials and phagocytosis 28 . Our work here showed that TeO 3 2− exposure can elevate the c-di-GMP level in P. aeruginosa and lead to the formation of floating aggregates. TeO 3 2− -induced floating aggregate formation requires Pel and Psl polysaccharides as well as extracellular DNA (eDNA) (Fig. S1), in accordance with the Psl polysaccharide-eDNA interaction enabling the formation of skeleton of P. aeruginosa biofilms 44 . In addition to serving as matrix scaffolds, the polysaccharides could also induce synthesis of iron siderophore pyoverdine via the Gac/Rsm pathway in the floating aggregates, as we had previously demonstrated 45 . The formation of stress-induced biofilm-resembling aggregates might contribute to the dissemination of infection in the host.
The results presented here demonstrate that P. aeruginosa mutants with low c-di-GMP content were more sensitive to TeO 3 2− exposure in planktonic cultures and thus their growth was negatively affected by TeO 3 2− exposure, as compared to c-di-GMP containing wild-type strain (Fig. 4). Consistent with this finding, a recent study on biodegradation of 3-chloroaniline by Comamonas testosteroni reported that, compared with the wild type, the strain with an elevated c-di-GMP level exhibited a better growth on the toxic substrate at high concentrations 46 . In addition to TeO 3 2− , the detergent Na-dodecylsulfate (SDS) 35 also raised the c-di-GMP levels and caused aggregation of P. aeruginosa. In accordance with the TeO 3 2− findings, the Δ siaD mutant with low intracellular c-di-GMP content was more sensitive to SDS during planktonic growth 35 . Together, these studies highlight that c-di-GMP signaling is involved in multiple stress response mechanisms, which might due to multiple DGCs and PDEs being encoded by many bacterial species.
Finally, we found that wild-type PAO1 strain biofilms prevented the attachment of mutants with low intracellular c-di-GMP content in both normal and TeO 3 2− stress co-cultures. Our previous study revealed that the polysaccharides in P. aeruginosa biofilms could not be shared, for structural or functional benefits, by mutants that are defective in their synthesis 38 . These latter findings corroborate with the results presented here, and c-di-GMP mediated synthesis of polysaccharides may form another strategy to repress the proliferation and maintenance of c-di-GMP defective variants in biofilms. Considering . P. aeruginosa PAO1, Δ sadC, Δ siaD, Δ sadCΔ siaD, and PAO1/p lac -yhjH strains were cultivated in ABTGC medium at 37 °C with shaking for growth measurement. For TeO 3 2− tolerance assay, P. aeruginosa PAO1, Δ sadC, Δ siaD, Δ sadCΔ siaD, and PAO1/p lac -yhjH strains were cultivated in ABTGC medium with the presence of 20 μg/ml TeO 3 2− overnight followed by CFU determination. Means and standard deviations of three replicates are shown.
that polysaccharides with similar structure to the P. aeruginosa polysaccharides are widely distributed in natural bacterial species, our results might reflect a conserved strategy employed by a range of bacterial species to repress the spreading of variants which cannot respond to environmental conditions by moderating their own c-di-GMP levels.

Methods
Bacterial strains and growth medium. The bacterial strains, plasmids, and primers used in this study are listed in Table 1. Escherichia coli DH5α strain was used for standard DNA manipulations. LB medium 47 was used to cultivate E. coli strains. Batch cultivation of P. aeruginosa was carried out at 37 °C in ABT minimal medium 7 supplemented with 5 g glucose l -1 (ABTG) or 2 g glucose l -1 and 2 g casamino acids l -1 (ABTGC). For plasmid maintenance in E. coli, the medium was supplemented with 100 μg ampicillin (Ap) ml −1 , 15 μg gentamicin (Gm) ml −1 , 15 μg tetracycline (Tc) ml −1 , or 8 μg chloramphenicol (Cm) ml −1 . For marker selection in P. aeruginosa, 30 μg Gm ml −1 , 50 μg Tc ml −1 , and 200 μg carbenicillin (Cb) ml −1 were used, when appropriate. Antibiotics were not added to P. aeruginosa cultures for c-di-GMP, stress response and biofilm assays as the plasmids we used were highly stable for these short-term experiments.
Construction of P. aeruginosa mutants. The Δ pelA, Δ pslBCD and Δ pelAΔ pslBCD mutants defective for Pel and/or Psl polysaccharide biogenesis were constructed by allelic displacement as previously described 48 . The Δ sadC, Δ siaD and Δ sadCΔ siaD mutants defective for SadC and/or SiaD diguanylate cyclase were constructed by allelic displacement as previously described 34 . Quantification of static biofilms. The microtitre tray biofilm formation assay was performed as described by O'Toole & Kolter 49 . Briefly, overnight cultures grown in ABTG medium were diluted to OD 600 = ~0.001 with fresh ABTG medium and transferred to the wells of polystyrene 96-well microtitre trays (200 μl per well) and incubated for 24 h at 37 °C. Liquid culture was removed from each well and the wells were washed twice with 0.9% NaCl followed by staining with 0.1% crystal violet and washing twice with 0.9% NaCl. The crystal violet-stained biofilms were then resuspended in 96% ethanol, and the absorbance of biofilm-associated dye was measured at 600 nm. Experiments were performed in triplicate, and the results are shown as the mean ± sd. Figure 6. Relative fitness of PAO1/p lac -yhjH mutant to PAO1 in planktonic co-cultures and biofilm cocultures in ABTGC medium with and without the presence of 10 μg/ml TeO 3 2− (a). Means and standard deviations of three replicates are shown. Student's t-test was performed for testing differences between groups. * P < 0.05. CLSM images of biofilm co-cultures formed by cfp-tagged P. aeruginosa PAO1 and yfptagged PAO1/p lac -yhjH mutant in ABTGC medium with and without the presence of 10 μg/ml TeO 3 2− (b). Representative image from triplicate experiments was shown for each condition. Bars, 50 μm. Field emission scanning electron microscopy (FE-SEM) and energy-dispersive X-ray spectroscopy (EDX). The aggregates were dried and coated with platinum (Pt) using a vacuum electric sputter coater (JEOL JFC-1300, JEOL Asia Pte Ltd, Singapore). SEM images were taken using a field emission scanning electron microscope (FE-SEM, JSM-7600, JEOL Asia Pte Ltd, Singapore) at a voltage of 2.0-5.0 kV and EDX spectra were obtained using an energy-dispersive X-ray spectroscope (AZtecEnergy, Oxford Instruments, Oxfordshire, UK) as previously described 50 . Experiments were performed in triplicate, and representative images were shown. iTRAQ-based proteomics analyses. P. aeruginosa PAO1 was grown in ABTG medium with and without 10 μg/ml TeO 3 2− at 37 °C with shaking until stationary phase was reached. Cells were harvested and iTRAQ-based proteomics analyses were carried out as previously described 12 . Determination of minimal inhibitory concentration (MIC). The MIC assays employed a microtiter broth dilution method as previously described in the NCSLA guidelines 51 . Briefly, fresh ~16 h cultures of P. aeruginosa were diluted in ABTG medium. For determination of MIC, potassium tellurite was dissolved in water at a concentration 10 times higher than the required range by serial dilutions from a stock solution. 10 μl of each concentration were added to each corresponding well of a 96-well microtiter plate (polypropylene, Costar Corp.) and 90 μl of bacterial culture (~1 × 10 5 cells) in ABTG medium were added. The plate was incubated at 37 °C for 16-18 h. MIC was taken as the lowest concentration where no visual growth (based on OD 600 ) of bacteria was detected. Experiments were performed in triplicate and representative results were shown. Beta-galactosidase activity assay. A classical β-galactosidase assay 52 was used to measure expression of the P pel -lacZ fusion in P. aeruginosa strains transformed with the mini-CTX-pel-lacZ fusion 36 , which carries the pel promoter fused to the E. coli lacZ gene. Experiments were performed in triplicate, and the results are shown as the mean ± sd. Student's t-test was performed for testing differences between groups.
Gfp reporter fusion assay. The expression of the c-di-GMP P cdrA -gfp biosensor 30 in P. aeruginosa strains in the presence and absence of TeO 3 2− was monitored by using a Tecan Infinite 2000 Microplate Reader. Monitoring strains were cultivated in 24-well microtiter plate with ABTGC medium with different concentrations of TeO 3 2− at 37 °C with shaking. OD 600 and GFP fluorescence (in relative fluorescence units, rfu) were measured every hour until the culture reach stationary growth phase. Experiments were performed in triplicate, and the results are shown as the mean ± sd. Student's t-test was performed for testing differences between groups.
Quantification of c-di-GMP concentration. Extraction of c-di-GMP was conducted as previously described 45 . 10 ml of P. aeruginosa cells in the early stationary phase from the ABTGC medium with and without 10 μg/ml TeO 3 2− were washed twice with 1 mM ammonium acetate. Cells were lysed with 0.6 M HClO 4 on ice for 30 min. Cell debris was removed by centrifugation and supernatant was neutralized to pH 6.0 with the addition of 2.5 M KHCO 3 . The precipitated KClO 4 was removed by centrifugation and the supernatant was used for relative quantification of c-di-GMP. The concentration was measured by High Performance Liquid Chromatography (HPLC), the injection volume is 20 µl with 254 nm as detection wavelength. Reverse-phase C18 Targa column (2.1 x 40 mm, 5 μm) (catalog number: TR-0421-C185) was used with solvent A (10 mM ammonium acetate in water) and solvent B (10 mM ammonium acetate in methanol) at a flow rate of 0.2 ml min-1. Eluent gradient is as follows: 0 to 8 min, 1% B; 8 to 14 min, 15% B; 14 to 16 min, 19% B; 16 to 24 min, 100% B; 24 to 32 min, 100% B; 32 to 40 min, 1% B; 40 to 42 min, 1% B. The retention time of c-di-GMP is around 14.0 min. The c-di-GMP concentration was normalized by total protein concentration. The relative c-di-GMP concentrations of cells treated with 10 μg ml −1 tellurite against cells in ABTGC only were shown. Experiments were performed in triplicate, and the results are shown as the mean ± sd. Student's t-test was performed for testing differences between groups.
Competition assay. Competition assays were performed in both planktonic and biofilm co-cultures.
In planktonic co-cultures, cfp-tagged wild-type PAO1 was mixed 1:1 (vol/vol) with yfp-tagged PAO1/ p lac -yhjH (or yfp-tagged Δ sadCΔ siaD) and the mixtures inoculated into fresh ABTGC medium with and without the presence of 10 μg/ml TeO 3 2− . For relative fitness calculation, co-cultures were plated in LB agar plates after 24 h cultivation at 37 °C with shaking. Colony-forming units (CFUs) N i were determined from three individual experiments and the number of PAO1 and PAO1/p lac -yhjH (or Δ sadCΔ siaD) colonies were determined based on their specific fluorescence at times t = 0 and at t = T. Relative fitness was determined as r ij = [N i (T)-N i (0)]/[N j (T)-N j (0)] as previously described with modification 53 , resulting in a fitness of '1' when competing organisms are equally fit. Experiments were performed in triplicate, and the results are shown as the mean ± sd. Student's t-test was performed for testing differences between groups.
In biofilm co-cultures, cfp-tagged wild-type PAO1 cells were mixed with yfp-tagged PAO1/p lac -yhjH (or yfp-tagged Δ sadCΔ siaD) cells at 1:1 (vol/vol) and the mixtures were inoculated into fresh ABTGC medium with and without the presence of 10 μg/ml TeO 3 2− . Static biofilms were cultivated on cover slides at 37 °C for 24 h as previously described 54 . Biofilms were imaged with a Zeiss LSM780 confocal laser scanning microscope (CLSM) equipped with detectors and filter sets for monitoring of Cfp and Yfp fluorescence. Images were obtained using a 40 × /1.4 objective. Simulated three-dimensional images and sections as well as biovolumes were generated using the Imaris software package (Bitplane AG) 8 . The biovolume V i of each strain in the biofilm mode was determined from three individual experiments based on their fluorescence at times t = 0 and at t = T. Relative fitness was determined as r ij = [V i (T)-V i (0)]/ [V j (T)-V j (0)] as previously described with modification 53 , resulting in a fitness of '1' when competing organisms are equally fit. Experiments were performed in triplicate, and the results are shown as the mean ± sd. Student's t-test was performed for testing differences between groups.