Oxygen-Mediated Regulation of Biofilm Development Is Controlled by the Alternative Sigma Factor σ^B in Staphylococcus epidermidis

Abstract

Using a modified rotating-disk reactor to sparge oxygen to Staphylococcus epidermidis cultures, we found that oxygen negatively regulates biofilm development by influencing the activity of σ^B. Under anaerobic conditions, increased σ^B activity activates icaADBC, which encodes enzymes responsible for polysaccharide intercellular adhesin synthesis, by repressing transcription of the negative regulator icaR.

Coagulase-negative staphylococci and Staphylococcus aureus are the most common causes of device-related infections and are known to cause 50 to 70% of intravenous-catheter-related infections (1). Biofilm development by Staphylococcus epidermidis and S. aureus on the surfaces of implanted devices can give rise to persistent and difficult-to-treat infections. In both S. aureus and S. epidermidis, polysaccharide intercellular adhesin (PIA) is an important component of the staphylococcal biofilm (10). Synthesis of PIA requires enzymes encoded by the intercellular adhesion (ica) operon icaADBC (15, 17). A number of studies have indicated that the ica locus may be a useful marker for distinguishing between significant and contaminating isolates (6, 7, 14, 21). In S. epidermidis strains carrying the ica locus, upregulation of ica operon expression and PIA production are also required for biofilm formation. Environmental triggers such as ethanol and salt stress, excess glucose, and subinhibitory antibiotic concentrations can also activate biofilm development (2, 3, 4, 11, 18, 19). A number of important ica operon regulators have now been identified. The icaR gene, which is located adjacent to the ica operon, encodes a transcriptional repressor of the ica locus (3, 11). Expression of icaR is in turn repressed indirectly by the alternative stress-responsive sigma factor σ^B (13). Thus, environmental conditions that activate σ^B result in repression of the icaR gene and deregulation of the icaADBC operon (2, 12, 16, 20).

Biofilm development by S. epidermidis 1457 (12) and its isogenic mutant M15, which contains a Tn917 transposon insertion in the rsbU gene at the sigB locus (16), was investigated under anaerobic conditions. Biofilm development on polycarbonate coupons was measured in a modified rotating-disk reactor (RDR; Biosurface Technologies Corp., MT), in which a sparger extending from the lid to below the liquid level was employed to enable precise control of dissolved oxygen concentrations. Profiles of oxygen utilization of S. epidermidis 1457 within the RDR were performed using oxygen sensor spots (PreSens GmbH, Regensburg, Germany) for the extremes of oxygen concentrations tested (Fig. 1). These profiles show that the rate of oxygen utilization by the cells is higher than the supply. It is important to note at this stage, however, that though the profiles look identical from 7 h onwards (not shown), the cells are in completely different conditions. Cells sparged with 0% oxygen are forced to grow anaerobically, whereas cells sparged at 21% consume oxygen at the rate supplied.

FIG. 1.

Dissolved oxygen profiles for S. epidermidis 1457 planktonic cultures grown in a modified RDR sparged with different oxygen concentrations. Profiles are means of two independent runs.

Biofilm development in the reactor was examined at oxygen concentrations of 0%, 7%, 14%, and 21% with a gas flow rate of 0.5 liters min⁻¹ in all cases. Each run initially involved a 24-h batch phase, followed by a 24-h continuous phase in which quarter-strength brain heart infusion (BHI) medium (Oxoid) was fed at 90 ml h⁻¹ using a peristaltic pump (Watson Marlow, United Kingdom). Diluted concentrations of BHI medium are commonly used by others to culture biofilms (2, 9, 25), and our previous work (not shown) revealed quarter strength to be the optimum nutrient concentration for biofilm growth by strain 1457 in our studies. Total RNA was extracted from biofilm cells as described previously (2) with the following modifications. Immediately on termination of an RDR run, the rotating disk was aseptically removed and washed. Biofilms were scraped from the coupons and the Viton surface of the disk and resuspended in RNAlater (Ambion, Austin, TX) to maintain the biofilm mRNA expression profile. Prior to RNA purification, RNAlater was removed and the biofilm was dispersed by treatment with 100 μl 0.2 M sodium metaperiodate (Sigma-Aldrich, Germany) for 5 min as described previously (18). Reverse transcription-PCR (RT-PCR) analysis, using the OneStep RT-PCR kit (Qiagen, United Kingdom), of ica operon transcript levels in RNA extracted from biofilm cells involved an RT step at 55°C for 30 min followed by 31 amplification cycles of 90°C for 20 s, 50°C for 20 s, and 72°C for 20 s. Similar reaction conditions were used to measure ica operon expression in RNA extracted from planktonic cells, but only 26 amplification cycles were required. RT-PCR analysis of transcript levels of asp23, which is a known target gene of σ^B (8), involved an RT step at 55°C for 30 min followed by 12 amplification cycles of 90°C for 20 s, 50°C for 20 s, and 72°C for 20 s. 16S rRNA yields were compared by agarose gel electrophoresis, and the constitutively expressed gyrB gene (2) was used as an internal standard in all RT-PCR experiments.

Oxygen availability in the infection environments of implanted medical devices, which typically involve biofilms, is likely to vary, and environments may in some instances be anoxic (22). Our data, obtained while tightly controlling oxygen concentrations, revealed enhanced S. epidermidis biofilm formation under anaerobic conditions, with a statistical difference evident between biofilms grown at 0% and 21% oxygen (P < 0.05) (Fig. 2). Examination of planktonic cells grown at 0% oxygen and 21% oxygen within the reactor (data not shown) revealed that differences in biofilm CFU counts could be directly attributed to enhanced biofilm development. RT-PCR analysis revealed a substantial increase in icaA expression under anaerobic conditions compared to that under aerobic (21% oxygen) conditions (5) in both biofilm (Fig. 3) and planktonic (Fig. 4) cells. It is important that ica transcripts are still detected at low levels in samples grown at 21% oxygen and that the icaADBC transposon mutation abolishes biofilm formation by 1457 cells (12). Concomitant with the activation of the ica operon, expression of the icaR gene was also substantially higher in cells grown at 21% oxygen, explaining, at least in part, why more biofilm was consistently formed under anaerobic conditions. These data suggest that activation of the ica locus under anaerobic conditions is the result of icaR repression. To investigate the possible mechanism of icaR repression by oxygen, we examined the impact of oxygen on the activity of the alternative sigma factor σ^B by measuring transcription of asp23. Significantly, asp23 expression was dramatically activated under anaerobic conditions, indicating that σ^B activity is increased in the absence of oxygen (Fig. 4). Under anaerobic conditions in the modified RDR biomass, yields of the rsbU mutant, S. epidermidis M15 (13), were significantly lower than those of the wild-type strain (Fig. 5). These data strongly suggest that activation of σ^B activity under anaerobic conditions increases icaADBC expression, and accordingly biofilm formation, by repressing transcription of the icaR gene. Under aerobic conditions, biomass yields of M15 were similar to those of the wild-type strain. These results, which contrast with previously published data (13), may be explained by differences in the growth environment between the modified RDR and 96-well plates. For example, localized anoxic regions are more likely to occur in the latter, where the specific oxygen transfer rate can be expected to be lower than in the RDR. Importantly, ethanol also activates the ica operon and biofilm development by repressing icaR transcription, but in a σ^B-independent manner (2, 12, 13). Using the RDR system, we confirmed that biofilm formation by M15 was restored to wild-type levels under anaerobic conditions in BHI medium supplemented with 4% ethanol (Fig. 5). These data support the existence of two separate pathways for ica locus activation in S. epidermidis and further reveal that anaerobic activation of ica operon expression and biofilm development is dependent on the σ^B regulatory pathway. It appears that σ^B is less important for biofilms under high-oxygen conditions (21% oxygen) than under low-oxygen conditions (0% oxygen) as the wild type (1457) and the mutant exhibit similar biofilm phenotypes. In S. aureus, the staphylococcal respiratory response regulator SrrA directly activates ica transcription under anaerobic conditions and does not modulate icaR expression (23). A potential role for SrrA in S. epidermidis has yet to be investigated, but amino acid sequence alignments suggest that no SrrA homologue exists in S. epidermidis. These findings may suggest that σ^B is less important for oxygen-dependent biofilm regulation in S. aureus than previously thought and are consistent with previous studies indicating that σ^B plays a more important role in biofilm regulation in S. epidermidis than in S. aureus (5, 12, 13, 24).

FIG. 2.

Biofilm development by S. epidermidis 1457 on polycarbonate coupons in a modified RDR after 48 h of growth in quarter-strength BHI medium at 21%, 14%, 7%, and 0% dissolved oxygen concentrations. Error bars represent standard errors from three independent experiments.

FIG. 3.

Comparative measurement of icaA, icaR, and gyrB (control) transcription by RT-PCR in biofilm biomass of S. epidermidis 1457 grown in a modified RDR in quarter-strength BHI medium at 21%, 14%, 7%, and 0% sparged oxygen concentrations. Comparative intensities of 16S rRNA bands after agarose gel electrophoresis are also shown. These experiments were performed three times, and representative results are shown.

FIG. 4.

Comparative measurement of icaA, icaR, asp23, and gyrB (control) transcription by RT-PCR in 48-h planktonic cultures of an S. epidermidis 1457 biofilm grown in a modified RDR in quarter-strength BHI medium at 21% and 0% sparged oxygen concentrations. Comparative intensities of 16S rRNA bands after agarose gel electrophoresis are also shown. The cells were harvested from the reactor waste. These experiments were performed three times, and representative results are shown.

FIG. 5.

Comparison of biofilm development of S. epidermidis 1457 and its isogenic rsbU mutant M15 on polycarbonate coupons in a modified RDR after 48 h of continuous growth in quarter-strength BHI medium or BHI medium supplemented with 10% ethanol. CFU for biofilms grown under aerobic (21% oxygen) and anaerobic (0% oxygen) conditions are shown. Error bars represent standard errors from three independent experiments. EtOH, ethanol.