ABSTRACT
Stenotrophomonas maltophilia WR-C is capable of forming biofilm on polystyrene and glass. The lipopolysaccharide/exopolysaccharide-coupled biosynthetic genes rmlA, rmlC, and xanB are necessary for biofilm formation and twitching motility. Mutants with mutations in rmlAC and xanB display contrasting biofilm phenotypes on polystyrene and glass and differ in swimming motility.
Microorganisms can develop biofilms or clogging mats, causing the failure of septic tanks, systems for on-site wastewater disposal. If water in these clogged systems were contaminated by pathogens, it would pose a threat to human health. We isolated Stenotrophomonas maltophilia strain WR-C from a clogged septic tank system that consistently formed biofilms on sand grains, produced exopolysaccharides (EPS), and caused clogging in sand columns (33). S. maltophilia is a gram-negative, rod-shaped, and obligate aerobic bacterium with polar flagella in the γ-β subdivision of Proteobacteria (3, 14). It is found in various environments and recently has emerged as an important human pathogen. Very little is known about the mechanisms of biofilm formation by S. maltophilia. It has been shown to adhere to HEp-2 cells as well as abiotic surfaces, such as plastic, glass, and Teflon (7-10, 16).
To identify genes that are involved in biofilm formation by S. maltophilia WR-C, about 4,500 transposon mutants generated with the EZ::TN < R6Kγori/KAN-2 > Tnp transposome (Epicenter, Madison, Wis.) were screened for defects in biofilm formation in 96-well polystyrene plates (Becton-Dickinson Labware, Franklin Lakes, N.J.) by using a modified microtiter plate assay (11, 29). Briefly, wells containing 200 μl Trypticase soy (TS) broth were inoculated with overnight cultures and incubated at 30°C, 50 rpm for 2 days. Biofilm cells were stained with 0.1% crystal violet and washed, the stain remaining in the cells was solubilized with 70% ethanol, and the optical density at 590 nm was determined. Three mutants, TPH7, TPH11, and TPH13, whose growth was similar to that of the wild type but which were deficient in biofilm formation, were selected for further characterization.
The transposon flanking regions were rescued by “rescue cloning” as described by the manufacturer and sequenced by using the primers KAN-2 FP-1 and R6KAN-2 RP-1 (Table 1). The transposon-inserted genes are homologous to three genes involved in the biosynthesis of nucleotide sugar precursors of lipopolysaccharide (LPS) and EPS in Xanthomonas campestris pv. campestris ATCC 33913 (GenBank accession no. AE008922) (6). The genes are rmlC (for TPH7, 74% identity), rmlA (for TPH11, 78% identity), and xanB (for TPH13, 80% identity), which are located in the rml and xan operons of X. campestris pv. campestris (19, 20). As no genome sequence is available and the gene cluster involved in LPS/EPS biosynthesis has not been reported for S. maltophilia, the DNA segments flanking the transposon in TPH7, TPH11, and TPH13 were sequenced. The complete sequences of the rmlBACD and xanAB operons and their flanking genes were determined. The genetic organization of these genes is different from that of X. campestris pv. campestris (Fig. 1). The xanA sequence from S. maltophilia WR-C shares 75% identity with xanA in X. campestris pv. campestris and 72% identity with spgM (GenBank accession no. AY179964) in S. maltophilia K1014 (24).
Genetic organization of LPS biosynthetic genes and their flanking genes in X. campestris pv. campestris (Xcc) ATCC 33913 (GenBank accession no. AE008922) and S. maltophilia WR-C (GenBank accession no. AY956411). The genes depicted are not proportional to their respective sequence lengths.
Oligonucleotides used in this study
To determine whether phenotypic differences that were observed in TPH7, TPH11, and TPH13 are due to mutations generated in the rml and xan operons, rmlBACD, xanAB, and their respective predicted promoters were cloned into pJN105 to generate prmlBACD and pxanAB for complementation. The plasmids were electroporated into the wild type and TPH7, TPH11, and TPH13 to generate the respective complemented strains as listed in Table 2. Providing prmlBACD in trans in TPH7 and TPH11 and pxanAB in TPH13 restored all of the phenotypes tested to those of the wild type (Table 3 and described below), while the introduction of pJN105 into the mutants had no effect. The presence of pJN105, prmlBACD, or pxanAB in the wild type did not affect the phenotypes tested (not shown).
Bacterial strains and plasmids
Phenotypes of S. maltophilia WR-C
To determine whether rmlA, rmlC, and xanB in S. maltophilia WR-C play a role in LPS biosynthesis, Tris-Tricine sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis of the LPS was performed with proteinase K-digested whole-cell lysates (13). The wild type produced both the diffuse O-antigen-containing LPS and the LPS core and lipid A (Fig. 2); these results were similar to what has been observed with X. campestris pv. campestris (20, 39). No O-antigen-containing LPS was observed in the rmlC, rmlA, and xanB mutants. The xanB mutant produced a truncated core and lipid A. These results suggest that rmlC and rmlA are necessary for the biosynthesis of the LPS O-antigen and xanB for the O-antigen and LPS core. Nonmucoid morphology on TS agar and autoagglutination in TS broth were observed in the xanB mutant but not the rmlC and rmlA mutants and wild type (not shown). These phenotypical alterations are correlated with defective LPS in other gram-negative bacteria (4, 19, 21, 27). In addition to LPS biosynthesis, rml and xanB orthologues in Pseudomonas aeruginosa, Escherichia coli, and X. campestris pv. campestris are required for the biosynthesis of exopolysaccharides (19, 32). The xanB, rmlA, and rmlC mutants all produced lower amounts of EPS when assayed by the phenol-sulfuric acid method for total carbohydrates (13) (not shown), suggesting that the three genes are also involved in EPS biosynthesis in S. maltophilia WR-C. Whether S. maltophilia produces an EPS similar to xanthan or alginate is unknown.
SDS-PAGE analysis of LPS from proteinase K-digested whole-cell lysates of S. maltophilia. The rmlC, rmlA, and xanB mutants were defective in biosynthesis of LPS O-antigen. The xanB mutant was also defective in LPS core. The wild-type (WT) LPS profile was restored in the complemented strains. M, LPS standard from Salmonella enteritidis serovar Typhimurium (12.5 μg; Sigma).
Biofilm formation by the rmlC, rmlA, and xanB mutants was significantly decreased on polystyrene plates (a relatively hydrophobic surface) compared to that by the wild type (P < 0.05; Student's two-sample t test and one-way analysis of variance [ANOVA]) (Fig. 3A). Interestingly, the rmlA and rmlC mutants developed more biofilm in glass tubes (a relatively hydrophilic surface) than did the wild type and xanB mutant (P < 0.05) (Fig. 3B). Differential attachment was also reported for O-antigen-deficient mutants of Pseudomonas fluorescens (41) and P. aeruginosa (22) relative to the corresponding wild-type strains and was attributed in part to a differences in cell surface hydrophobicity. By the bacterial adhesion to hexadecane assay (34), we did not observe any difference in the relative cell surface hydrophobicities displayed by the S. maltophilia wild type and its mutants (not shown), possibly because the strains were too hydrophilic (16) to be evaluated by this method. Alternatively, other unknown factors or mechanisms that are associated with alterations in the LPS or EPS may be involved.
Biofilm formation by S. maltophilia. (A) Formation in a 96-well polystyrene microtiter plate. (B) Formation in a borosilicate glass tube. All experiments were performed at least three times. Results represent the mean and standard deviation (error bars) of a representative experiment. *, significant difference (P < 0.05, analyzed by Student's two-sample t test and one-way ANOVA) relative to biofilm formation by other strains. WT, wild type; OD590, optical density at 590 nm.
Flagellum-mediated swimming is involved in biofilm development (28, 36). Alterations in LPS were reported to interfere with the production, export, or assembly of flagella in E. coli, Helicobacter pylori, P. aeruginosa, and Salmonella enterica serovar Typhimurium (2, 12, 18, 30). The swimming motility of our strains was determined on TrA (1% tryptone, 0.5%NaCl, and 0.25% Bacto agar) (35). Overnight cultures (5 μl) were spotted on TrA plates (60 by 15 mm, containing 8 ml TrA) and incubated at 30°C. Only the xanB mutant was less motile than the wild type (Fig. 4A). One possible cause for the impairment in swimming motility in the xanB mutant could be a lack of flagellar formation as a result of the truncated LPS core. Cells of all strains were harvested from the edge of a 2-day swimming culture on TrA and stained with crystal violet by using a method described by Mayfield and Inniss (23) and modified by Kearns and coworkers (17). Flagella were observed consistently in the wild type and rmlC and rmlA mutants (Fig. 5), all of which existed predominantly as single cells. In contrast, most cells of the xanB mutant were linked into chains of three or more bacteria that were devoid of flagella. This likely contributed to the autoagglutination by the xanB mutant that was observed in liquid medium (not shown). Occasionally single flagellated cells were observed.
Motility of S. maltophilia. (A) Swimming motility. Only xanB mutant had reduced swimming motility. (B) Twitching motility. The rmlC, rmlA, xanB mutants were all defective in twitching motility. The arrow indicates the edge of twitching movement away from the colony center. WT, wild type.
Flagellar formation by S. maltophilia. The xanB mutant formed long chains devoid of flagella. The complemented strain was mostly dispersed as flagellated single cells as in other strains. WT, wild type. Bar, 2 μm.
Type IV pilus-mediated twitching motility is involved in biofilm formation by P. aeruginosa (36). Defects in O-antigen polysaccharides affected twitching motility in P. aeruginosa (40) and prevented the export and assembly of the toxin-coregulated pilus subunit, a type IV pilin of Vibrio cholerae O1 (15). The involvement of rml and xan operons in twitching motility has not been examined. Twitching motility was determined in plates containing 1% TS and 1% Noble agar (36). Overnight cultures (5 μl) were stabbed to the bottom of the agar and incubated at 30°C. All three mutants were defective in twitching motility (Fig. 4B). Twitching in S. maltophilia is positively correlated with biofilm development in polystyrene microtiter plates, which is in agreement with that observed in P. aeruginosa by Chiang and Burrows (5) and O’Toole and Kolter (28) but is in contrast to the results of Singh et al. (37).
LPS is important as a barrier to antimicrobial agents and neutral detergents (31, 38) and for maintaining cell function and integrity (1, 25, 38). Growth was slightly decreased to similar extents (62 to 74%) in the wild type and the rmlA and rmlC mutants when 0.01% SDS, an anionic detergent, was added; however, the xanB mutant was particularly susceptible (Table 3). The alteration of LPS affected the type II secretory system in P. aeruginosa (26). We observed that protease activity was significantly reduced in the xanB mutant, while hemolytic and lecithinase activities were not affected in the rmlA, rmlC, and xanB mutants (not shown). LPS biosynthetic genes in S. maltophilia may have effects on the secretory machinery or expression of certain putative virulence factors by this opportunistic pathogen.
rmlA, rmlC, and xanB in S. maltophilia WR-C contribute to an LPS/EPS-coupled biosynthetic pathway. The formation of S. maltophilia WR-C biofilm on a polystyrene surface requires EPS and intact LPS. Alteration in LPS caused by the rmlAC and xanB mutations may contribute to changes in outer membrane components and apparatuses, such as flagella and type IV pili; this in turn interferes with motility, attachment, and biofilm formation. Whether the differences in biofilm phenotype of rmlAC and xanB mutants on glass is due to differences in polysaccharide compositions remains to be determined.
Nucleotide sequence accession numbers.
The sequences obtained in this study were deposited in the GenBank database (accession no. AY956411).
ACKNOWLEDGMENTS
This research was supported by the Wisconsin Small Scale Waste Management Project and the College of Agricultural and Life Sciences, University of Wisconsin-Madison.
We thank C. Fuqua for his generous gift of the plasmid pJN105 and A. Charkowski for E. coli strain DH5α λpir.
FOOTNOTES
- Received 23 November 2005.
- Accepted 3 February 2006.
- Copyright © 2006 American Society for Microbiology
