Salmonella enterica Serovar Typhimurium Swarming Mutants with Altered Biofilm-Forming Abilities: Surfactin Inhibits Biofilm Formation

S. enterica serovar Typhimurium SJW1103 (a fliC stable derivative of serovar Typhimurium LT2) and its transposon mutants have been described (18). The bacteria were grown in Luria-Bertani (LB) medium. Surfactin, sodium dodecyl sulfate (SDS), and Tween 80 were purchased from Sigma Chemical Co. Rhamnolipid (extracted from Pseudomonas aeruginosa) was a gift from Jeneil Biosurfactant Co. Inc. PVC 96-well microtiter plates were from Costar (2596; nontreated). Clear vinyl urethral catheters (14Fr) were from Kendall Co., Mansfield, Mass.

The quantification assay was performed as previously described (14). Typically, 10 μl of an overnight culture was used to inoculate PVC microtiter wells containing 90 μl of LB without NaCl but with 2% glucose. The covered microtiter dish was sealed with Parafilm during incubation at 30°C. Cultures were removed to determine the optical density at 630 nm (OD₆₃₀), and the wells were rinsed with distilled water. After drying at room temperature for 15 min, 200 μl of crystal violet (1%) was added to the wells for 20 min. The stained biofilms were rinsed several times with distilled water, allowed to dry at room temperature for 15 min, and extracted twice with 200 μl of 95% ethanol. The OD₅₅₀ was estimated using a Beckman DU-640B spectrophotometer after adjusting the volume to 1 ml with distilled water.

An overnight culture of S. enterica (10 μl) was inoculated into 500 μl of medium and injected into clear vinyl urethral catheters. The catheters were capped at both ends and incubated at 30°C overnight. Media and growth conditions were as described above for PVC wells. Cultures were removed to determine the OD₆₃₀, and the catheters were rinsed with distilled water. After drying at room temperature for 15 min, 700 μl of crystal violet (1%) was added to the catheters for 20 min. The stained biofilms were rinsed several times with distilled water and allowed to dry at room temperature for 15 min before examination.

The biofilm assay used in this study monitors the ability of S. enterica to attach to the wells of microtiter dishes (14). The biofilm forms at the interface between the air and the liquid medium and is quantitated by staining with crystal violet as described in Materials and Methods. Initial experiments with different abiotic materials (PVC, polystyrene, and borosilicate glass) showed that the wild-type strain SJW1103 forms the best biofilms on PVC, in LB broth without NaCl but with 0.2% glucose, and at 30°C; these conditions were used throughout. SJW1103 did not form a detectable biofilm in M9 minimal medium, consistent with the results reported forEscherichia coli (15). Figure1 shows the kinetics of biofilm formation by wild-type S. enterica. The exponential phase of biofilm formation coincided with that of cell growth. Biofilm formation began to slow around 13 h, decreased up to 17 h, and then leveled off, coincident with entry of the culture into stationary phase.

Fifty-nineS. enterica swarming mutants (18), all completely defective in swarming, were tested for their ability to form biofilms. A subset of these is shown in Fig.2. Also included in this analysis were four mutants with mutations that did not affect swarming (luxS [18]; 1, 2, and 3 are random Tn10dCm insertions in SJW1103). Biofilms were harvested when cultures reached an OD₆₃₀ of approximately 0.2. The data (arranged in increasing biofilm formation efficiency) are shown in Fig. 2. Overall, a continuous spectrum of biofilm formation was observed in the mutants, from 10 to 156% of the wild-type level. Except for the waaP and waaL mutants, all the LPS mutants clustered at the far end of this spectrum. Mutants not affected in swarming showed biofilm formation similar to that of wild-type cells (Fig. 2, arrowheads and arrow, respectively). Specific mutants are discussed below.

As shown previously with E. coli (15), bothmotA and motB mutants were significantly defective in biofilm formation, reaching only about 10% of the wild-type level, while the che mutants were less defective, attaining between 38 and 62% of wild-type levels. Biofilm formation by mutants not defective in swarming (Fig. 2, arrowheads) was within 10 to 20% of the wild-type level. Among the swarming mutants, a whole range of biofilm formation levels was observed. We note that the crystal violet staining technique is only a measure of adherence and that this analysis only measures gross biofilm formation. For example, in thelasI mutant of P. aeruginosa, total biofilm formation was unaffected but the structure was physically different and, in contrast to the wild-type parent, could be dislodged with 0.2% SDS (4). Similarly, a colanic acid mutant of E. coli showed normal crystal violet staining, but when examined under the microscope, the mutant did not display the complex pillars of cells typical of the wild-type strain (3). Thus, it is possible that mutants that appear to be unaffected by this assay could have different structures to their biofilms. If we consider only the very low or very high end of the biofilm formation spectrum, then of the uncharacterized two-component signaling mutations, baeSis at the low end and yfhk and yojN are towards the middle. At the high end there are a cluster of LPS mutations,waaK, waaC, wbaP, wzx, andddhC. The kinetics of biofilm formation for one of these mutants (ddhC) was determined to be similar to the wild-type kinetics in the initial phase of growth; however, the mutant continued to accumulate more biofilm after the wild type had begun to slow (data not shown). None of the LPS mutants (or the wild type) formed biofilm on a hydrophilic surface such as glass under these growth conditions (data not shown). Attempts to measure cell surface hydrophobicity of the wild type and the mutants by measuring adherence to hydrocarbons (17) were not successful.

We have recently shown that the swarming defect of the LPS mutants could be rescued by addition of the biosurfactant surfactin isolated from Bacillus subtilis (18). To analyze the effect of surfactin on biofilm formation, either the PVC wells were coated with surfactin or surfactin was included in the growth medium. Figure3 shows that increasing amounts of surfactin led to a decrease in the amount of biofilm formed by the wild-type strain and that 5 μg of surfactin was more than sufficient to completely abolish biofilm formation. Bacterial growth was unaffected under all surfactin concentrations tested (data not shown). Surfactin had a similar inhibitory effect on biofilm formation by theddhC mutant (data not shown).

To test if surfactin would dislodge a preformed biofilm, surfactin was added to the PVC wells after the culture had reached an OD₆₃₀ of ≈0.15 to 0.2 (Fig.4A). The OD₅₅₀ of the surfactin-treated sample decreased at a faster rate than that of the untreated sample for the initial sloughing phase of biofilm formation, resulting in an approximately 85% decrease in total biofilm by the end of the experiment at 22 h (Fig. 4B). To determine if other surfactants would have similar effects on preformed biofilms, we tested SDS (ionic surfactant), Tween 80 (anionic surfactant), and rhamnolipid. While all the chemicals tested dispersed the preformed biofilm, SDS and rhamnolipid decreased the OD₆₃₀of the culture as well (Fig. 4A). (Surfactin concentrations in this and the rest of the experiments were maintained at 100 μg in order to compare its activity to that of the biosurfactant rhamnolipid, which affected biofilm formation at higher concentrations.)

To investigate the biofilm-forming ability of bacteria known to produce surfactants, we tested Serratia marcescensand B. subtilis as well as mutants defective in surfactant production. S. marcescens mutants defective in production of the surfactant serrawettin are unable to swarm (9, 13), as are surfactin mutants of B. subtilis (obtained from P. Zuber, Oregon Graduate Institute of Science and Technology; our unpublished data). Mutants of S. marcescens that were defective in serrawettin synthesis made approximately threefold more biofilm that their wild-type parent (Fig. 5). These results are consistent with the notion that absence of the biosurfactant promotes biofilm formation. To test if surfactin would inhibit biofilm formation by the serrawettin mutants, surfactin was included in parallel wells (Fig. 5). A significant reduction in biofilm formation was observed for all the strains in the presence of surfactin. Purified serrawettin (gift from T. Matsuyama, University of Niigata, Niigata, Japan), had an inhibitory effect on biofilm formation but was not as effective as surfactin (data not shown). We were unable to observe biofilm formation by wild-type B. subtilis or its surfactin mutants under these conditions (data not shown).

To test the effect of surfactin on medically relevant objects, we grewS. enterica in clear vinyl urethral catheters. The stained catheters are shown in Fig. 6. The biofilm formed by S. enterica was dispersed all along the growth surface. Surfactin eliminated formation of the biofilm. Precoating the catheters by running the surfactin solution through them prior to inoculation with medium was just as effective as including surfactin in the growth medium (data not shown). Among other surfactants tested on S. enterica, Tween 80 (0.25%) was as effective as surfactin, while rhamnolipid seemed to be only half as effective (data not shown). Given the importance of opportunistic infections with Salmonella species, including urinary tract infections of AIDS patients (7), these results have the potential for practical applications.

To test the effect of surfactin on biofilm formation by other medically relevant organisms, we grew E. coli, Proteus mirabilis, and Pseudomonas aeruginosa in urethral catheters. E. coli and Proteus mirabilis formed a biofilm mainly at the air-liquid interface, while the biofilm formed byP. aeruginosa, like that formed by S. enterica, was dispersed all along the catheter (Fig. 6). Surfactin inhibited biofilm formation (but not growth) by all organisms except P. aeruginosa. Similar inhibition profiles were obtained with Tween 80 (data not shown). We note that there are earlier reports on inhibition of biofilms formed by uropathogens and yeasts on silicone rubber by biosurfactants produced by Streptococcus thermophilus and lactobacilli (1, 19).

The main conclusions from this work can be summarized as follows. At least one difference between adherent and moving biofilms may lie in the surfactant composition of the slime, since the absence of surfactants inhibits swarming but promotes biofilm formation and vice versa. O-antigen mutants of S. enterica, which are defective in swarming (18), were generally more proficient in biofilm formation than the wild type (Fig. 2). O-antigen has been postulated to provide a surfactant or wettability function that allows spreading of the swarmer colony (18). Surfactin, a cyclic lipopeptide from B. subtilis which was shown to rescue the swarming defect of the LPS mutants (18), inhibited biofilm formation (Fig. 3). Mutants ofS. marcescens defective in serrawettin production (serrawettin is an analog of surfactin [10]) are defective in swarming (9, 13) and make better biofilms (Fig. 5). Surfactin (as well as the chemical surfactant Tween 80) dispersed preformed biofilms without affecting cell growth (Fig. 4) and also prevented biofilm formation in urethral catheters by organisms such as S. enterica, E. coli, and P. mirabilis (Fig. 6). The nonbactericidal nature of these compounds would be advantageous in preventing the selection of resistant organisms if used for inhibiting biofilm formation. Biofilms formed byP. aeruginosa were not affected by the surfactants tested in this study (Fig. 6). Our results suggest the need for additional research into the use of this genre of compounds for biofilm containment. Since P. aeruginosa did not form biofilms on a hydrophilic glass surface under our growth conditions, manipulating the hydrophilicity of the growth surface might provide another avenue to the control of this class of biofilms.