Disintegration of Staphylococcus epidermidis Biofilms under Glucose-Limiting Conditions Depends on the Activity of the Alternative Sigma Factor σ^B

Abstract

To evaluate the role of the polysaccharide intercellular adhesin as an energy-storage molecule, we investigated the effect of nutrient limitation on S. epidermidis biofilms. The stability of established biofilms depends on σ^B activity; however, the slow decay of biofilms under conditions of nutrient limitation reveal its use as an energy-storage molecule to be unlikely.

Biofilm formation is the major pathogenic mechanism involved in infections caused by Staphylococcus epidermidis (5) and is associated with an increased antibiotic resistance (14). Cell-to-cell adhesion within S. epidermidis biofilms is mediated by the polysaccharide intercellular adhesin (PIA) (18, 19). PIA is synthesized by the gene products of the icaADBC gene locus (4, 6). PIA synthesis and methicillin resistance are regulated by the gene rsbU via the alternative sigma factor σ^B (8, 11, 12, 21). The σ^B activity in staphylococci is regulated by at least three regulatory genes located upstream of sigB (15, 23, 24, 26). The genes rsbU and rsbV encode negative regulators, whereas rsbW encodes a positive regulator of σ^B activity. Interestingly, in σ^B-negative mutants PIA synthesis could be restored under ethanol stress (11), as observed in the majority of icaADBC-positive clinical isolates (9). In addition to these regulatory pathways of biofilm formation, additional regulatory gene loci are involved in the regulation of PIA synthesis (2, 13, 17, 20).

Implanted medical devices have only been used in the last few decades, indicating that the role of PIA as a virulence factor plays only a secondary role, whereas the primary physiologic function of PIA synthesis is still poorly understood. The observation of sugar-dependent biofilm formation in staphylococci (3, 10) has led to the hypothesis that PIA could act as an energy storage molecule (7). To further evaluate this possible function of PIA, we investigated the stability of S. epidermidis biofilms under conditions of glucose limitation in the present study. The possibility to restore biofilm formation in mutants with dysfunctional σ^B by ethanol induction (8, 11) allowed us to characterize the influence of the alternative sigma factor σ^B on the stability of the established biofilms under different environmental conditions. Mutants of clinical isolates S. epidermidis 1457 and 8400 with inactivation of the regulatory genes rsbU, rsbW, the entire regulatory cascade rsbUVW, the sigB gene, and the entire sigma B operon were investigated (Table 1).

TABLE 1.

Strains used in this study

S. epidermidis strain	Reference	Comments
1457	22	Isolate from infected central venous catheter
1457rsbU	11	rsbU::erm derivate from S. epidermidis 1457, generated by allelic gene replacement
1457rsbW	11	rsbW::erm derivate from S. epidermidis 1457, generated by allelic gene replacement
1457sigB	11	sigB::erm derivate from S. epidermidis 1457, generated by allelic gene replacement
1457rsbUVW	11	rsbUVW::erm derivate from S. epidermidis 1457, generated by allelic gene replacement
1457rsbUVWsigB	11	rsbUVWsigB::erm derivate from S. epidermidis 1457, generated by allelic gene replacement
8400	22	Isolate from infected central venous catheter
8400rsbU	11	rsbU::erm derivate from S. epidermidis 8400, derived by phage transduction
8400rsbW	11	rsbW::erm derivate from S. epidermidis 8400, derived by phage transduction
8400sigB	11	sigB::erm derivate from S. epidermidis 8400, derived by phage transduction
8400 rsbUVW	11	rsbUVW::erm derivate from S. epidermidis 8400, derived by phage transduction
8400rsbUVWsigB	11	rsbUVWsigB::erm derivate from S. epidermidis 8400, derived by phage transduction

Biofilms were grown in 96-well tissue culture plates (Nunc, Roskilde, Denmark), and biofilm formation was measured quantitatively after staining with gentian violet as described previously (1, 16). To induce biofilm formation for the primary cultures Trypticase soy broth (TSB; Becton Dickinson, Cockeysville, Md.) was supplemented with 3% ethanol for all strains. For investigations of biofilm stability, the culture supernatant was replaced daily with TSB, TSB without glucose (TSB_wo; 1.7% tryptone, 0.3% soy peptone [Oxoid, Basingstoke, England], 0.5% NaCl, and 0.25% dipotassium phosphate [Merck, Darmstadt, Germany]), or phosphate-buffered saline (PBS). The amount of remaining biofilms was quantified daily.

Established biofilms were stable over at least 6 days in all mutants when TSB was used to replace the culture supernatants, indicating that σ^B activity is not required to maintain established S. epidermidis biofilms under energy-rich conditions in these strains (Fig. 1). In TSB_wo and PBS used for replacement of the culture supernatants, biofilms of strains with deletions of rsbU, as well as of sigB and the entire σ^B operon, in S. epidermidis 8400 disintegrated within 2 to 3 days, whereas in S. epidermidis 1457 the respective mutants displayed even in PBS, after an initially rapid decrease, still substantial biofilm. In contrast, biofilms of strains with mutations of the negative regulator RsbW or the entire regulatory cascade rsbUVW displayed a high stability over at least 6 days in TSB_wo. In wild-type S. epidermidis 8400 a slight decay of the primarily established biofilm was observed. However, even after this time period the strains would be classified as strongly biofilm positive (9). The biofilm of S. epidermidis 1457 remained stable under these conditions.

FIG. 1.

Stability of biofilms under nutrient limitation. The wild-type strains S. epidermidis 1457 (A) and 8400 (B), as well as their respective mutants with inactivation of rsbU, rsbW, sigB, the regulatory cascade rsbUVW, and the entire σ^B operon, were investigated. After primary biofilm formation in TSB supplemented with 3% ethanol culture, the supernatants were replaced daily with fresh TSB, TSB lacking glucose (TSB_wo), or phosphate-buffered saline for 6 days. All experiments were performed three times in quadruplicate. The displayed bars represent mean values of all results obtained in three experiments. Error bars represent the standard error.

In PBS the rsbW and rsbUVW mutants, as well as both wild-type strains, displayed a slow decay of primary biofilms over 6 days; however, no complete disintegration of biofilms was observed, and the remaining levels of biofilms were higher compared to the mutants with dysfunctional σ^B activity. Similar phenotypic results were obtained for wild-type strains and mutants with inactivation of rsbW or rsbUVW if TSB without ethanol was used for primary biofilm formation (data not shown), indicating that ethanol induction has no influence on biofilm stability under the investigated conditions.

The relation of biofilm stability with PIA stability was investigated for representative mutants of strain 8400 by confocal microscopy using SYTO9 to stain cells and TRITC (tetramethyl rhodamine isothiocyanate)-labeled wheat germ agglutinin (Molecular Probes, Eugene, OR), a lectin which binds to PIA (25), to stain the extracellular polysaccharide matrix. Even after 6 days of energy limitation, similar amounts of PIA could be detected in strains still displaying a strong biofilm-positive phenotype (Fig. 2) indicating that the stability of PIA is responsible for the observed stability of biofilms in mutants with intact σ^B activity. Interestingly, only a few intact cells could be detected after 6 days, whereas large masses of PIA with significantly altered structure were still detectable (Fig. 2). For mutants with dysfunctional σ^B activity, thin biofilms could be detected at day 4 in TSB_wo which also displayed a significantly altered structure (Fig. 3). A thin layer with almost no detectable PIA was located at the top with ca. 5 to 10 μm distance to the basal layer of attached cells at the bottom of the chambers only fixed by a few PIA-containing clusters (Fig. 3).

FIG. 2.

Horizontal sections of biofilms in representative strains displaying a stable biofilm-positive phenotype under conditions of energy limitation. Strains 8400 (A, B, and C) and 8400rsbUVW (D, E, and F) were grown in coverglass cell culture chambers (Nunc, Roskilde, Denmark) in TSB containing 3% ethanol (A and D), and culture supernatants were replaced daily with TSB_wo (B and E) or PBS (C and F) for 6 days. Chambers were washed four times with PBS, and remaining biofilms were stained with SYTO9 and TRITC labeled wheat germ agglutinin to stain cells (green) and PIA (red), respectively. Compared to day 0 (A and D), both strains displayed a similar structure with slight reduction of cells in TSB_wo and a similar amount of PIA at day 6 (B and E). In PBS both strains displayed a significantly altered structure of the biofilm. Almost no intact cells could be detected, and PIA-containing fibers (arrows) were observed connecting the remaining PIA clusters. For strains with dysfunctional σ^B activity, no remaining biofilms could be detected at day 6 (data not shown). Displayed scales indicate 10 μm.

FIG. 3.

Vertical sections of biofilms. Strains 8400 (A), 8400rsbU (B), and 8400rsbUVWsigB (C) were grown in coverglass cell culture chambers in TSB containing 3% ethanol, and culture supernatants were replaced daily with TSB_wo for 4 days. Chambers were washed four times with PBS, and remaining biofilms were stained with SYTO9- and TRITC-labeled wheat germ agglutinin to stain cells (green) and PIA (red), respectively. Strain 8400 displayed a compact cell mass with single cells enclosed by PIA (A), as it was observed for all strains at day 0 and for strains with a strong biofilm-positive phenotype in TSB_wo after 6 days (data not shown). Due to the high density of the biofilm cells of the upper layers are insufficiently illuminated. In contrast, strains 8400rsbU and 8400rsbUVWsigB displayed a thin cell layer with almost no detectable PIA (B and C) that was located at the top with about 5 to 10 μm distance to the basal layer of attached cells at the bottom of the chambers only fixed by a few PIA-containing clusters (arrows). In contrast to the biofilm observed for strain 8400, the cells embedded in the thin layer displayed a more elastic cell-to-cell contact (see movies in the supplemental material) as observed for a protein dependent biofilm (J. K.-M. Knobloch and H. Rohde, unpublished data). Displayed scales indicate 20 μm.

The data presented here suggest that the stability of S. epidermidis biofilms under nutrient limitation depends on σ^B activity. The slightly different observations in two investigated S. epidermidis isolates indicate that additional factors could influence biofilm stability under conditions of energy limitation. However, the slow reduction of wild-type strain biofilms and the stability of biofilms of the rsbW and rsbUVW mutants in TSB lacking glucose indicates that it is unlikely that PIA is used as an energy storage molecule. The regulation of biofilm formation by σ^B, which is a major inductor of genes required to cope with environmental stress, suggests that the primary role of PIA synthesis could be interpreted as a kind of sporulation equivalent that is protecting biofilm-embedded cells from environmental stresses.