Subinhibitory Concentrations of Mupirocin Stimulate Staphylococcus aureus Biofilm Formation by Upregulating cidA

Previous studies have shown that the administration of antibiotics at subinhibitory concentrations stimulates biofilm formation by the majority of multidrug-resistant Staphylococcus aureus (MRSA) strains. Here, we investigated the effect of subinhibitory concentrations of mupirocin on biofilm formation by the community-associated (CA) mupirocin-sensitive MRSA strain USA300 and the highly mupirocin-resistant clinical S. aureus SA01 to SA05 isolates.

ABSTRACT

Previous studies have shown that the administration of antibiotics at subinhibitory concentrations stimulates biofilm formation by the majority of multidrug-resistant Staphylococcus aureus (MRSA) strains. Here, we investigated the effect of subinhibitory concentrations of mupirocin on biofilm formation by the community-associated (CA) mupirocin-sensitive MRSA strain USA300 and the highly mupirocin-resistant clinical S. aureus SA01 to SA05 isolates. We found that mupirocin increased the ability of MRSA cells to attach to surfaces and form biofilms. Confocal laser scanning microscopy (CLSM) demonstrated that mupirocin treatment promoted thicker biofilm formation, which also correlated with the production of extracellular DNA (eDNA). Furthermore, quantitative real-time PCR (RT-qPCR) results revealed that this effect was largely due to the involvement of holin-like and antiholin-like proteins (encoded by the cidA gene), which are responsible for modulating cell death and lysis during biofilm development. We found that cidA expression levels significantly increased by 6.05- to 35.52-fold (P < 0.01) after mupirocin administration. We generated a cidA-deficient mutant of the USA300 S. aureus strain. Exposure of the ΔcidA mutant to mupirocin did not result in thicker biofilm formation than that in the parent strain. We therefore hypothesize that the mupirocin-induced stimulation of S. aureus biofilm formation may involve the upregulation of cidA.

INTRODUCTION

Methicillin-resistant Staphylococcus aureus (MRSA) is one of the major causes of hospital-acquired infections, which include bloodstream infections, pneumonia, nosocomial infections, and surgical-site bacterial colonization (1). The major risk factor for these nosocomial infections is the extensive use of central venous catheters, artificial lenses, heart valves, and prosthetic joints (2). Bacteria colonize the nasal cavity and the skin, which serve as reservoirs (1). Biofilms formed during MRSA infection are difficult to treat, enabling the bacteria to resist the host immune response and antibacterial agents. It has been reported that the effect of antimicrobial agents on bacteria embedded in biofilms can be 10 to 1,000 times weaker than that when when the same organisms assume a planktonic state (3). Furthermore, treatment may require different doses and combinations of antibiotics, which can result in bacterial antibiotic resistance.

Biofilms are a group of microbial cells that attach to the surface of living or inanimate objects and coat themselves in a polysaccharide matrix (1). The capacity to form biofilms increases the risk of hospital-acquired infections. Biofilms are frequently associated with human acute and chronic disease and serve to protect bacteria by decreasing susceptibility to antimicrobial agents and the host’s immune system. Various studies have shown that the biofilm matrix of many clinical S. aureus isolates is composed of substantial quantities of extracellular DNA (eDNA) (2, 3), which is released through a process called autolysis. The activation of murein hydrolases helps the release of eDNA in S. aureus biofilms (4, 5). The Cid/Lrg holin-antiholin system is associated with cell lysis and the release of eDNA during planktonic growth to facilitate adhesion and biofilm formation (6, 7). In addition, the cidA-encoded protein increases the activity of murein hydrolases, promoting the detachment of bacteria from the biofilm and their spread to new infection sites in the dispersal phase (6–8). Moreover, many studies have reported that phenol-soluble modulins (PSMs) are important biofilm matrix components that stabilize S. aureus biofilms; furthermore, owing to their surfactant-like properties, these proteins also play a role in biofilm dispersal (9–11). The Agr quorum-sensing system regulates the transcription of the psmα genes. Moreover, it has been demonstrated that the ica locus, which is composed of the icaR (regulatory) and icaADBC (biosynthetic) genes, is necessary for biofilm formation (12–14). Moreover, the proteins encoded by this set of genes appear to coordinate precise biofilm formation during secondary infection.

Mounting evidence suggests that S. aureus specifically responds to subinhibitory concentrations of antibiotic by inducing biofilm formation and changing its biofilm matrix composition. For example, Brunskill and Bayles (15) and Mlynek et al. (16) showed that low doses of β-lactam antibiotics such as amoxicillin and methicillin stimulate biofilm formation in most MRSA strains by inducing autoaggregation and eDNA release. Furthermore, Schilcher and colleagues showed that subinhibitory concentrations of clindamycin stimulate S. aureus biofilms by modulating the biofilm matrix (17). However, limited studies have evaluated the effect of mupirocin on S. aureus biofilms. Mupirocin (pseudomonic acid A) is a topical antibiotic that is produced by Pseudomonas fluorescens (18). Its antibacterial effect is achieved via the inhibition of bacterial protein synthesis by reversibly binding to isoleucyl-tRNA synthetase, which limits the availability of isoleucine. This drug shows bacteriostatic and bactericidal activity at low and high concentrations, respectively, and it is widely used for the decolonization of MRSA carriers (19). Mupirocin was successfully used as part of a decolonization protocol in the United Kingdom MRSA control program (20). Furthermore, in a previous well-characterized study of intensive care unit (ICU) patients, mupirocin treatment proved the most effective strategy for the significant reduction of MRSA-positive clinical cultures and bloodstream infections (P < 0.05) (21–23).

The present study provides a more thorough investigation of the modulation and regulation of S. aureus biofilms in response to subinhibitory concentrations of mupirocin, a topical antibiotic commonly used in the clinical setting. The ΔcidA mutant, which is characterized by the deletion of the cidA gene by allelic replacement, was generated following initial transcriptomic and reverse transcription-PCR (RT-PCR) analyses of the mupirocin-sensitive S. aureus strain USA300. This ΔcidA mutant was subsequently used to investigate the effect of subinhibitory mupirocin concentrations on biofilm formation.

RESULTS

The effect of subinhibitory concentrations of mupirocin on mupirocin-sensitive and highly mupirocin-resistant MRSA biofilm formation.

Biofilm formation by the community-associated (CA) mupirocin-sensitive MRSA strain USA300, and by the highly mupirocin-resistant clinical S. aureus isolates SA01 to SA05, was assessed in 96-well polystyrene microtiter plates containing tryptic soy broth (TSB) supplemented with 0.5% glucose. The mupirocin MIC for the SA01 to SA05 isolates was ≥1,024 μg/ml (SA01, SA02, and SA05 MIC = 1,024 μg/ml; SA03 and SA04 MIC = 2,048 μg/ml), demonstrating a high level of mupirocin resistance. In contrast, the MIC for USA300 was 4 μg/ml, evidencing mupirocin sensitivity. The details of the strains used in this study are listed in Table 1.

TABLE 1.

The S. aureus strains used in this study

Bacterial strain(s) or plasmid	Descriptiona	Source
Bacterial strains
SA01 to SA05	Clinical high-level mupirocin-resistant MRSA isolates, biofilm negative	The First Affiliated Hospital of Wenzhou Medical University
USA300	CA-MRSA isolate, biofilm negative	Fudan University
ΔcidA mutant	cidA deletion mutant obtained using USA300 as the parent strain	This study
c-ΔcidA mutant	ΔcidA mutant complemented with plasmid PRB473 harboring the cidA gene	This study
Δpsmα mutant	psmα deletion mutant obtained using USA300 as the parent strain	This study
DH5α and DC10B	Escherichia coli isolates	The Renji Hospital of Shanghai Jiaotong University School of Medicine
Plasmids
PKOR1	Temp-sensitive E. coli (AmpR)- Staphylococcus (CmR) shuttle vector	The Renji Hospital of Shanghai Jiaotong University School of Medicine
PRB473	An inducible shuttle plasmid, CmR	The Renji Hospital of Shanghai Jiaotong University School of Medicine

^aCm^R, chloramphenicol resistance; Amp^R, ampicillin resistance.

As shown in Fig. 1a, enhanced biofilm formation was observed when USA300 was cultured for 24 h in TSB supplemented with 0.125 to 0.5 μg/ml mupirocin. However, when cultured with <0.125 μg/ml mupirocin, no difference in biofilm formation was observed in relation to the untreated group (cultured without mupirocin). Therefore, the 0.5 μg/ml mupirocin concentration was used in further investigations. Bacterial biofilm formation was monitored at 24 h on microtiter plates stained with crystal violet, and the optical density at 600 nm (OD₆₀₀) was read. As shown in Fig. 1c, the mature biofilm of the mupirocin-treated (0.125 μg/ml-0.5 μg/ml) USA300 strain (OD₆₀₀, 1.315 to 2.378) was significantly increased compared to that of the untreated strain (OD₆₀₀, 0.045 ± 0.06) after 24 h of incubation. However, the concentrations of mupirocin below 0.125 μg/ml had little effect on biofilm formation. Biofilm formation by the clinical strains is shown in Fig. 1b and c. Because of the different genetic backgrounds of clinical strains, the biofilm formation is not as same as that of the USA300 strain. In general, there is a trend that subinhibitory concentrations of mupirocin can indeed stimulate S. aureus biofilm formation. The mature biofilm of the mupirocin-treated (4 μg/ml-64 μg/ml) SA01 to SA05 strains (OD₆₀₀, 1.524 to 2.872) was significantly increased compared with that of the untreated strain (OD₆₀₀, 0.085 ± 0.36) after 24 h of incubation. However, the concentrations of mupirocin below 4 μg/ml had little effect on biofilm formation. It indicated that not all subinhibitory concentrations of mupirocin have an impact on S. aureus biofilms.

FIG 1.

Biofilm formation by S. aureus in the presence of subinhibitory concentrations of mupirocin. (a) Biofilm formation by mupirocin-sensitive MRSA strain USA300 (0.01 to 0.5 μg/ml). (b) Biofilm formation by five clinical mupirocin-resistant MRSA isolates, SA01 to SA05, in the presence of subinhibitory concentrations of mupirocin (2 to 64 μg/ml). Photographs of crystal violet-stained microtiter plate wells (in triplicate) are shown. (c and d) Biofilm formation by mupirocin-treated and untreated USA300 strain and SA01 to SA05 strains on microtiter plates. Glacial acetic acid (33%) was used to release the biofilms into solution, and the absorbance at 600 nm (A₆₀₀) was recorded. NO-MUP, mupirocin-untreated strain. Each experiment was repeated three times, and the data represent means ± standard deviations (SDs). **, P < 0.01; *, P < 0.05.

CLSM analysis of biofilms exposed to subinhibitory concentrations of mupirocin.

Confocal laser scanning microscopy (CLSM) was used to evaluate the biofilm thickness in the mupirocin-treated and untreated groups. Similar to the biofilm formation results, the thickness of mature biofilms increased (by approximately 14.42 μm) when bacteria were cultured in the presence of 0.5 μg/ml mupirocin (Fig. S1). Interestingly, the biofilms were also more robust, dense, and compact in the presence of antibiotic. Additionally, the live cell ratio in the biofilm of the mupirocin-treated group (89.51%) was 5.48-fold higher than that in the biofilm of the untreated group (16.32%).

The quantification of PIA and eDNA.

To investigate the effect of mupirocin on biofilm matrix production, the release of polysaccharide intercellular adhesin (PIA) and eDNA was determined. As shown in Fig. 2a, when the PIA antigen was diluted 10,000-fold, PIA production remained similar between the mupirocin-treated and untreated strains, as determined semiquantitatively with wheat germ agglutinin (WGA)-horseradish peroxidase (HRP) conjugate using a dot blot 96 system. It is indicated that the stimulation of S. aureus biofilm formation by mupirocin is not achieved by upregulating the expression pathway of the ica operon. However, the eDNA concentration (quantified by the level of gyrB transcription) in the 48-h-old biofilm belonging to the mupirocin-treated group was 3.67-fold higher than that in the untreated group (Fig. 2b). These results imply that S. aureus biofilm formation induced by subinhibitory concentrations of mupirocin is not dependent on the production of PIA, but rather involves eDNA secretion and the acceleration of bacteria autolysis.

FIG 2.

Effect of subinhibitory mupirocin concentrations on extracellular matrix biosynthesis by the USA300 strain. (a) PIA biosynthesis was semiquantified using a dot blot assay with whole-genome sequence alignment (WGA). Compared with the mupirocin-untreated group, there was no significant change in PIA content between the two groups when the PIA was diluted to 1:10,000, suggesting that mupirocin-induced MRSA biofilm formation is not PIA dependent. (b) The release of eDNA was quantified by quantitative PCR (qPCR) targeting the chromosomal locus gyrB. The expression of gyrB in the mupirocin-treated group was significantly higher than that in the untreated group (P < 0.05). (c) Comparison of the autolytic abilities of mupirocin-treated and untreated strains. Triton X-100 induced 3.16-fold higher levels of autolysis in the presence of mupirocin than that in the untreated control. Each result was derived from three independent experiments, and the data represent means ± SDs. *, P < 0.05.

The effect of mupirocin on the activity of USA300 proteolytic enzyme.

Triton X-100 was used to investigate the effect of mupirocin on autolysis in USA300. As shown in Fig. 3c, we found that Triton X-100 induced 3.16-fold higher levels of autolysis in the presence of mupirocin compared to that in the untreated control.

FIG 3.

Reverse transcription-PCR (RT-PCR) quantification of the effect of subinhibitory mupirocin concentrations on gene expression by the USA300 strain. The USA300 strain was cultured with 0.5 μg/ml mupirocin or in the absence of the antibiotic. The results are expressed as n-fold differences in the atlA to gyrB, aap to gyrB, bap to gyrB, fnbA to gyrB, fnbB to gyrB, icaA to gyrB, icaC to gyrB, cidA to gyrB, agrA to gyrB, and psmα to gyrB ratios in the presence of mupirocin, relative to the same gene expression ratios in the absence of mupirocin. Each independent experiment was performed in triplicate, with the data presented as mean ± SD. **, P < 0.01; *, P < 0.05.

Transcriptome comparison of the mupirocin-treated and untreated USA300 strain.

To compare the transcriptional profiles of the mupirocin-treated and untreated strains, RNA samples were extracted from bacteria following 10 h of culture and subjected to transcriptome sequencing (RNA-Seq). The sequencing libraries were prepared in triplicate for the mupirocin-treated and untreated groups.

A volcano plot of differences in gene expression between mupirocin-treated USA300 strain and mupirocin-untreated USA300 strain is shown in Fig. S2. A gene with a false discovery rate (FDR)-adjusted P value of <0.05 (t test), a q value of <0.05 (q value indicates the difference between the two sets of data), and an |log₂(fold change)| of >1 in the transcript level between the treatment groups was considered to be significantly differentially expressed. We identified 670 differentially expressed genes (DEGs) between mupirocin-treated and untreated strains; among these, 352 genes were upregulated, and 318 were downregulated. We selected 66 of the DEGs for validation by reverse transcription-quantitative PCR (RT-qPCR). The transcription data for the 58 DEGs in the mupirocin-treated strain were consistent with our RNA-Seq results.

Among the DEGs, transcription of the two-component system agr gene was upregulated 24.60-fold (2^4.62-fold) in the untreated strain. The expression of cidA, an operon encoding a holin-like protein that is essential for regulating bacterial growth, was also upregulated 10.48-fold (2^3.39-fold). In addition, another anti-holin-like protein-encoding gene, lrgB (operon encoded), was downregulated 5.28-fold (2^2.40-fold) relative to its expression in the untreated strain. Transcription of the δ-hemolysin (Hld) virulence factor-encoding gene hld was upregulated in the mupirocin-treated strain, while the gene encoding the major leukocyte lysis toxin (PSMs), psmα, was significantly upregulated (Table 2). PSMs are surfactant-like proteins that contain five α-peptides (PSMα1 to PSMα4 and δ-toxin) and two β-peptides (PSMβ1 an PSMβ2). PSMα proteins are encoded by the psmα operon, and the δ-hemolysin is encoded by hld, which is contained within the RNAIII transcript. The agr operon is a regulator of psmα and hld. As expected, compared to the untreated group that was not treated with mupirocin, the expression of the agr operon genes, such as agrA, agrB, and agrC, was upregulated in the mupirocin-treated group. Transcriptome sequencing revealed that the formation of biofilms was associated with the upregulated expression of cidA, psmα, agrA, agrB, and agrC genes in the mupirocin-treated S. aureus strain; in contrast, the lrgB gene was downregulated (Table 2).

TABLE 2.

Gene expression changes associated with biofilm formation during mupirocin treatment

Gene identifier	log2 fold change in expression	P value	q value	Description of gene product
Biofilm-related genes
SAR_RS13760	3.3893	5.28E−06	5.25E−06	Holin-like protein CidA
SAR_RS01295	−2.4027	1.14E−53	4.99E−53	Anti-holin-like protein LrgB
SAR_RS14885	9.5994	0	0	Phenol-soluble modulin alpha 3 peptide
SAR_RS11025	10.875	0	0	δ-Hemolysin
Membrane protein genes
SAR_RS15320	11.142	0	0	Membrane protein
SAR_RS05030	2.7013	1.84E−69	1.04E−68	Membrane protein
Transcriptional regulator genes
SAR_RS11030	2.1844	1.54E−32	4.68E−32	Accessory gene regulator agrB
SAR_RS11040	4.5872	5.49E−291	1.59E−289	agrC
SAR_RS11045	4.6152	0	0	agrA
SAR_RS02365	2.6596	2.89E−57	1.34E−56	GntR family transcriptional regulator

Comparison of biofilm-associated DEGs in the mupirocin-treated and untreated strains.

Next, we evaluated the expression of candidate genes in the mupirocin-treated and untreated strains in more detail. As shown in Fig. 3, when S. aureus USA300 was cultured with mupirocin, the expression level of the atlA gene was 3.46-fold higher (P < 0.05). However, the expression levels of icaA, icaC, bap, fnbA, and fnbB were not significantly altered (P > 0.05). Differences in eDNA concentration between the mupirocin-treated and untreated strains led us to additionally investigate the expression of cidA. In agreement with the transcriptome data, cidA expression levels significantly increased 35.52-fold (P < 0.05) on mupirocin administration, confirming that subinhibitory concentrations of mupirocin affected the S. aureus holin-/antiholin-like protein CidA. Moreover, in order to explore the transcriptome analysis results in depth, we investigated the expression of the major S. aureus virulence regulatory locus (agr and sarA) under the same conditions. We found that the levels of psmα and agrA expression significantly increased by 3.57- and 6.81-fold (P < 0.05), respectively. It is thus likely that subinhibitory concentrations of mupirocin activate the expression of the virulence of regulatory genes such as agrA and psmα. In the clinical S. aureus strains, cidA expression levels significantly increased (6.05-fold on average; P < 0.05) when cultured with mupirocin (Fig. S3a1 to a5). Furthermore, when cultured with mupirocin, the level of lrgB was significantly reduced (6.25-fold on average; P < 0.05); Fig. S3c and b1). In addition, the levels of atlA, psmα, and agrA expression were significantly increased by 2.67-, 4.08-, and 4.21-fold (P < 0.05) on average, respectively. These data suggest that cidA, lrgB, psmα, atlA, and agrA also play important roles in the stimulation of the biofilm formation in highly mupirocin-resistant clinical MRSA isolates treated with mupirocin.

Construction of ΔcidA and Δpsmα mutant strains and the complemented strains.

To identify the mechanism by which subinhibitory concentrations of mupirocin induced S. aureus biofilm formation, the USA300 mutants ΔcidA and Δpsmα (lacking cidA and psmα, respectively), were constructed using the temperature-sensitive shuttle plasmid pKOR1. Deletion of the target genes was verified by RT-qPCR and sequencing. The complemented ΔcidA strain, c-ΔcidA, was generated using the pRB473 shuttle vector.

Assessing the biofilm formation capacities of the ΔcidA and Δpsmα deletion mutants, the c-ΔcidA mutant, and the parent strain.

We subsequently evaluated the effect of cidA and psmα knockout on the biofilm formation capacity of S. aureus under mupirocin treatment by comparing the ΔcidA and Δpsmα mutants with the c-ΔcidA and parent USA300 strains. In comparison with the parent strain, the ΔcidA mutant displayed a marked reduction in biofilm formation in the presence of mupirocin (Fig. 4a). In contrast, the c-ΔcidA mutant formed biofilms similar to those of the parent strain when the strains were incubated under the same conditions. The Δpsmα mutant, however, exhibited no differences in biofilm formation compared with the parent strain (Fig. 4b). The semiquantification results for biofilms are shown in Fig. 4c and d. The thickness of the mature biofilm of the mupirocin-treated ΔcidA mutant was significantly decreased compared to that of the parent strain (25.69-fold, P < 0.01) under mupirocin treatment after 24 h of incubation. Moreover, the density of the c-ΔcidA mutant biofilm was significantly increased compared with that of the mupirocin-treated ΔcidA mutant. However, the density of the mature biofilm of the mupirocin-treated Δpsmα mutant was similar to that of the parent strain.

FIG 4.

Biofilm formation capacities of the ΔcidA and Δpsmα deletion mutants, the c-ΔcidA mutant, and the parent strain. (a) Biofilm formation by the ΔcidA deletion mutant versus the c-ΔcidA mutant and the parent USA300 strain. The ΔcidA mutant displayed a marked reduction in biofilm formation in the presence of mupirocin compared to that of the parent strain. The c-ΔcidA mutant formed biofilms similar to those of the parent strain when the strains were incubated under the same conditions. (b) Biofilm formation by the Δpsmα deletion mutant and the parent strain. Photographs of crystal violet-stained microtiter plate wells (in triplicate) are shown. The Δpsmα mutant exhibited no differences in biofilm formation compared with that of the parent strain. NO-MUP, mupirocin-untreated strain. (c and d) Glacial acetic acid (33%) was used to release the biofilms into solution, and the absorbance at 600 nm (A600) was recorded. Each experiment was repeated three times, with the data presented as mean ± SD. *, P < 0.05.

Autolysis assessment of the ΔcidA and Δpsmα deletion mutants, c-ΔcidA, and the parent strain.

The autolytic capacity of the parent strain, the ΔcidA and Δpsmα mutants, and the c-ΔcidA mutant was assessed by Triton X-100 induction. After 6 h of incubation with mupirocin, all strains were adjusted to an OD₆₀₀ of 1.0 and treated with 0.05% Triton X-100 for 3 h. As shown in Fig. 5a, the autolysis rate of the parent strain reached 88% after 3 h of Triton X-100 induction, which was significantly higher than that of the ΔcidA mutant (46%). Moreover, the autolysis rate of the c-ΔcidA mutant reached 67%, while the autolysis rate of Δpsmα (43%) was similar to that of the parent strain (Fig. 5b).

FIG 5.

Autolysis assessment of the ΔcidA and Δpsmα deletion mutants, the c-ΔcidA mutant, and the USA300 parent strain. (a) Comparison of autolytic ability of the mupirocin-treated and untreated ΔcidA deletion mutant, parent strain, and c-cidA mutant. The autolysis rate of the parent strain reached 88% after 3 h of Triton X-100 induction, which was significantly higher than that of the ΔcidA mutant (46%). (b) Comparison of the autolytic ability of the mupirocin-treated and untreated Δpsmα deletion mutant and parent strain. (c) The release of eDNA by the ΔcidA deletion mutant, parent strain, and the c-cidA mutant in the presence or absence of mupirocin was quantified by qPCR (targeting the chromosomal gyrB locus). The relative concentration of eDNA in the 48-h-old biofilms of the mupirocin-treated parent strain was 10.12-fold higher (in terms of the level of gyrB transcription) than that of the ΔcidA mutant strain (P < 0.05). (d) The release of eDNA by the Δpsmα deletion mutant and parent strain, either treated with mupirocin or untreated. Each result was derived from three independent experiments; the data represent means ± SDs. *, P < 0.05.

Quantification of eDNA released by the ΔcidA and Δpsmα deletion mutants, c-ΔcidA, and the parent strain.

The release of eDNA by the parent strain, the ΔcidA and Δpsmα mutants, and the c-ΔcidA strain was determined as previously described. As shown in Fig. 5c, the relative concentration of eDNA in the 48-h-old biofilms of the mupirocin-treated parent strain was 10.12-fold higher (in terms of the level of gyrB transcription) than that of the ΔcidA mutant strain (P < 0.05). The relative concentration of eDNA in the c-ΔcidA mutant was also much higher (5.69-fold) than that of the ΔcidA mutant strain. However, eDNA release by the Δpsmα mutant was similar to that of the parent strain (Fig. 5d).

DISCUSSION

MRSA, which was first reported in 1961 (24), is the primary cause of hospital-acquired infections. Biofilm formation is associated with many types of refractory infection (25), and biofilm-forming bacteria have a higher resistance to antibiotics than that of planktonic bacteria. Furthermore, biofilms can adhere to the surface of medical devices such as artificial lenses, central venous catheters, heart valves, and prosthetic joints (26). As a result, biofilm-forming bacteria act as secondary invaders that can cause bloodstream infections. In addition, the biofilm-encased bacteria, with negatively charged exopolysaccharides, bind antimicrobial agents and restrict their diffusion into the biofilms, making their eradication especially challenging (27).

Mupirocin is an effective antibiotic against most Gram-positive bacteria, and particularly against MRSA. However, a study in the United States showed that the resistance rate of MRSA to mupirocin was as high as 31.3% (28). Several epidemiological survey and studies have also shown that the effect of mupirocin for decolonization of patients with recurrent MRSA skin and soft tissue infections was reduced compared to that in previous years (18, 28). Moreover, high-level mupirocin resistance has been associated with treatment and decolonization failure (29). What is more, due to the pharmacokinetics of antibiotics, the germicidal efficacy often decreases with decreasing available concentration, shortening contact time in tissue and organs. Therefore, subinhibitory concentrations of agents may occur in vivo. In a previous in vitro investigation, Ishikawa et al. (30) studied the effect of a subinhibitory concentration of mupirocin (256 μg/ml, below the MIC) on biofilm formation in a Pseudomonas aeruginosa model. This subinhibitory antibiotic concentration significantly reduced biofilm formation and the production of biofilm glycocalyx (P < 0.05), but it failed to lower viable P. aeruginosa numbers (30). However, several studies have shown that low-dose β-lactam antibiotics stimulate biofilm formation in most MRSA strains (16, 17). Not only that, but Rao et al. demonstrated that antibacterial agents like β-lactamases can stimulate the expression of S. aureus virulence factors and may affect the prognosis of severe infections (31). It indicates that, in the process of clinical standardization, an antibacterial drug with subinhibitory concentration will have an opposite effect at the infected site. What is more, for MRSA infections, subinhibitory concentrations of agents not only lead to invalid or ineffective treatment but also possibly enhance the virulence and colonization of MRSA, contributing to poor outcomes. This is the clinical relevance and significance of this research. Based on these investigations, we decided to investigate the effect of subinhibitory concentrations of mupirocin on biofilm formation in highly mupirocin-resistant MRSA strains.

In this study, we found that under subinhibitory concentrations of mupirocin, biofilm formation by one mupirocin-sensitive CA-MRSA strain, USA300, and five different clinical highly mupirocin-resistant MRSA isolates was increased. Furthermore, phenotypic CLSM analysis confirmed the appearance of mature and viable bacteria within these biofilms. Advantages of the CLSM technique include the ability to analyze bacterial cell structures at high magnification without damaging the cells. Live/dead staining has been used as indicator of cell viability, as determined by the integrity of the cell wall membrane, in many bacterial populations, including biofilms. The results here revealed that S. aureus biofilms grown in subinhibitory concentrations of mupirocin were much thicker (by approximately 14.42 μm) than those grown in the absence of the antibiotic. We next demonstrated that the mupirocin-induced S. aureus biofilm formation was related to the production of eDNA and not to that of PIA. Although complex extracellular matrixes within S. aureus biofilms are mainly composed of PIA, cell wall-associated proteins, and eDNA, the regulatory networks involved in biofilm formation are intricate. Previous investigations have shown that biofilm formation is controlled by related genes such as icaA, icaC, fnbAB, and cidA (7–9) and by a set of regulatory genes, agr (28–30) and sarA (17, 30), which appear to precisely coordinate biofilm production during chronic infection. This implies that both CidA and PSMs are important biofilm matrix components for the stabilization of S. aureus biofilms (6, 10, 32). In this study, the level of cidA expression significantly increased by 35.52-fold in response to subinhibitory concentrations of mupirocin; this increase was the most likely cause of the observed increase in eDNA concentrations within the biofilm. Previous studies have reported that the cidAB operon encodes a puncturing needle-like protein that increases extracellular wall proteolytic activity (6, 7, 33). In contrast, the S. aureus lrgAB operon, which encodes an anti-acupoint-like protein, can inhibit extracellular wall proteolytic enzyme activity (34–36). As a result, when cultured with mupirocin, the level of lrgB was significantly reduced by 6.25-fold on average (P < 0.05). Although there was no difference in sarA expression, we performed RT-PCR to measure the downregulation of lrgB and the upregulation of agrA, atlA, and psmα expression in response to subinhibitory doses of mupirocin. Of note, cidA and psmα are positively regulated by the quorum-sensing Agr system, which was shown to be upregulated in response to mupirocin (37, 38).

Mupirocin was shown to upregulate cidA and agrA expression; accordingly, we speculated that additional biofilm factors, such as cell wall proteins (e.g., FnAPs, FnBps, and Bap) and PIA (regulated by the icaADBC locus) were also upregulated. However, we found that subinhibitory mupirocin concentrations did not stimulate S. aureus biofilm formation in a PIA- or cell wall protein-dependent manner. To corroborate this, we used RT-PCR to investigate the differences in fnbA, fnbB, bap, icaA, and icaC expression levels between mupirocin-treated and untreated strains. However, we observed no significant changes in the expression of fnbA, fnbB, bap, icaA, and icaC (P > 0.05); this suggested that the S. aureus holin-/antiholin-like proteins encoded by the cidA gene were responsible for the mupirocin-induced changes in the biofilm matrix. This conclusion is in contrast with the findings of a previous study that assessed the effect of subinhibitory concentrations of ceftaroline on MRSA biofilms (39); in this previous work, it was found that ceftaroline enhanced biofilm formation but did not upregulate the expression of agrA, icaA, or sarA (39). Previous studies showed that eDNA is a major biofilm matrix adhesin in S. aureus biofilms cultured in the absence of antibiotic treatment (5). The present findings show that biofilm formation in the presence of subinhibitory concentrations of mupirocin is also a result of cell lysis mediated by the autolysin system (21). It is thus possible that subinhibitory concentrations of mupirocin induce S. aureus biofilm formation by also upregulating atlA expression and downregulating lrgB expression, either directly or via the modulation of other regulatory proteins. Furthermore, previous studies have reported that the quorum-sensing Agr system can regulate S. aureus lrgB expression (34). Yamaki et al. (40) and Veringa and Verhoef (41) found that the production and expression of PSMs were significantly increased in the S. aureus response to subinhibitory concentrations of tigecycline and clindamycin, respectively. Our work is consistent with these previous findings, which are of interest in terms of broadening our understanding of the mechanisms implicated in biofilm formation, delineating the signaling pathways involved in the response to cell wall stress, and predicting the responses of MRSA strains to subinhibitory concentrations of mupirocin. The results show that after challenge with mupirocin, highly mupirocin-resistant MRSA strains upregulate their Agr system and express genes (such as cidA or psmα) implicated in biofilm formation, allowing them to exhibit resistance to antimicrobial agents.

Additionally, in order to further investigate the mechanism underlying biofilm formation in response to mupirocin by S. aureus, the ΔcidA and Δpsmα USA300 mutants were constructed. Compared to the untreated group or the mupirocin-treated parent strain, the ΔcidA mutant did not exhibit thicker biofilm formation capacity; furthermore, psmα deletion had little effect on biofilm formation following mupirocin treatment. These findings suggest that the upregulation of psmα may be associated with the Agr system response to mupirocin.

To further explore the involvement of cidA in biofilm formation, we generated the c-ΔcidA mutant. After induction with mupirocin, c-ΔcidA exhibited the same levels of biofilm formation as the parent strain. Comparison of the mupirocin-treated and untreated ΔcidA mutants revealed that the induction of biofilm formation under subinhibitory mupirocin concentrations in S. aureus was mainly due to CidA, a regulator of murein hydrolase and cell death. In brief, our results showed that deletion of cidA (USA300 ΔcidA) failed to stimulate biofilm formation under mupirocin treatment (Figure. 4a), whereas USA300 Δpsmα showed biofilm formation comparable to that of its wild-type strain, indicating that mupirocin-induced biofilm formation in MRSA is probably controlled by cidA.

In conclusion, this work focused on the function and regulation of cidA in response to the induction of subinhibitory concentrations of mupirocin. Mupirocin-induced MRSA biofilm formation is not PIA dependent, and upregulation of cell lysis after mupirocin treatment is directly controlled by the holin-like protein CidA, thereby possibly contributing to promotion of S. aureus colonization and infection. Our data support the recommendation to clinicians regarding the prudent usage of subinhibitory concentrations of antimicrobials agents, which may possibly contribute to poor prognosis of MRSA infections. However, there are some limitations in this study. Further investigations should be performed to clarify how mupirocin regulates cidA expression through some global regulators, and the mechanism of action of other implicated regulatory genes and the interactions between these genes in clinical high-level mupirocin-resistant S. aureus isolates warrant further investigation.

MATERIALS AND METHODS

Bacterial strains, plasmids, and antibiotics.

All bacterial strains and plasmids used in the study are shown in Table 1. The five clinical mupirocin-resistant S. aureus strains (SA01 to SA05) were isolated from the First Affiliated Hospital of Wenzhou Medical University (Wenzhou, China) between 2014 and 2016. Identification and antimicrobial susceptibility testing of the isolates were carried out using a Vitek 2 microbiology analyzer (bioMérieux, Marcy l’Etoile, France) in accordance with the manufacturer’s instructions. S. aureus strain USA300, Escherichia coli strain DC10B, and plasmids PKOR1 and PRB473 were provided by Li Min, based at the Shanghai Renji Hospital, Shanghai Jiaotong University School of Medicine. All antibiotics used in this study were purchased from the Sigma Chemical Company (Los Angeles, CA,).

MIC determination.

Samples (100 μl) were added to 96-well microfilter plates (BD Biosciences, New Jersey, USA) containing mupirocin (Sigma-Aldrich, St. Louis, MO, USA) (1 to 1,024 μg/ml) and 100 μl of tryptic soy broth (TSB). Control bacteria were cultured in the absence of mupirocin in accordance with Clinical and Laboratory Standards Institute (CLSI) breakpoints. After 24 h cultivation at 37°C, we used a microplate reader (Bio-Rad, USA) to monitor bacterial growth by recording the OD₆₀₀ value. The MIC was defined as the lowest concentration of mupirocin that completely inhibited S. aureus growth. The quality control strain for antimicrobial susceptibility test was S. aureus ATCC 29213.

Biofilm formation.

Bacteria were grown in 96-well microfilter plates in tryptic soy broth (TSB) containing 0.5% glucose (Sigma-Aldrich, St. Louis, MO, USA) without mupirocin or with mupirocin at different concentrations. Following 24 h of culture at 37°C, 50 mM phosphate-buffered saline (PBS; pH 7.2) was used to wash the wells three times. Methanol (99.5%) was used to stabilize biofilms. Then biofilms were stained with 0.5% crystal violet dye for 15 min. Distilled water was used to wash wells gently 3 times (to remove planktonic cells). After drying, 33% glacial acetic acid was used to release the biofilms into solution. Absorbance at 600 nm (A₆₀₀) was recorded.

Confocal laser scanning microscopy.

Strains were incubated in 4-well microchambers (ibidi GmbH, Martinsried, Germany) without shaking for 48 h. The assays were performed in the presence of 0.5 μg/ml of mupirocin or without mupirocin. Strains incubated in the absence of antibiotic were used as a positive control. After 48 h of incubation, the wells were washed with 0.85% normal saline. Next, 300 μl of SYTO-9 (0.02%; component A of the Live/Dead BacLight bacterial viability kit [Molecular Probes, USA]) and propidium iodide (PI; 0.067%) were added for 30 min in the dark at room temperature. A series of optical sections were observed with a Leica TCS SP5 CLSM, using a 63 × 1.4-numerical aperture (NA) oil immersion objective. Images were rendered in 3-dimensional (3D) mode using Imaris 7.4.2 software (Bitplane). This assay was performed in triplicate, with similar results.

Autolysis of the mupirocin-treated strains and mupirocin-untreated strains induced by Triton X-100.

Autolysis assays were performed as described by Brunskill and Bayles (15). Strains were cultured in 50 ml TSB containing 1 M NaCl and 0.5 μg/ml mupirocin or without mupirocin until they reached the early exponential phase (3 h). Following centrifugation (4,000 × g, 4°C, 20 min), cells were washed twice with ice-cold water, resuspended in lysis buffer (Tris-HCl; pH 7.2) containing 0.05% (vol/vol) Triton X-100, and shaken at 220 rpm for 3 h. The optical density at 600 nm was measured at 30-min intervals.

Quantitation of PIA in biofilms.

Polysaccharide intercellular adhesion (PIA) in the biofilms of mupirocin-treated strains and untreated strains was semiquantified by a dot blot assay with a wheat germ agglutinin (WGA)-horseradish peroxidase (HRP) conjugate, as described previously (43). Overnight cultures of the strains were incubated in 6-well plates at 37°C for 24 h. A cell scraper was used to scrape the cells from the surface, and the cells were transferred to a 1.5-ml microcentrifuge tube. Then sediments were resuspended in 0.5 M EDTA and centrifuged (13,000 × g for 10 min) after they were heated with boiling water for 5 min. Proteinase K (20 μl at 20 mg/ml) was added to the supernatant at 37°C for 3 h and inactivated at 100°C for 5 min. Serial dilutions of the PIA assay extract (diluted 10-, 100-, 1,000-, 5,000-, and 10,000-fold, respectively) were transferred to a nitrocellulose membrane (Millipore, Billerica, MA) using a 96-well dot blot device (Biometra GmbH, Gottingen, Germany). After drying, 5% skim milk was used as blocking reagent to block the membrane for 2 h at 25°C, and then washed 3 times. Subsequently, 1 μl WGA-HRP conjugate was added in 1 ml 5% skim milk, and incubated for 1 h (Lectinotest Laboratory, Lviv, Ukraine). After washing, enhanced chemiluminescence (ECL; Thermo Fisher Scientific, Waltham, MA) was used to visualize the quantitation of PIA.

Isolation and quantification of extracellular DNA (eDNA).

eDNA isolation from biofilms was performed as previously described by Mlynek et al. (16). Strains were cultured in 2 ml TSB without or supplemented with mupirocin in 4-well polystyrene plate at 37°C. After 48 h of incubation, TSB was removed, and the wells were washed with 0.85% normal saline. Next, 1 ml of TEN buffer (10 mM Tris, 1 mM EDTA, pH 8.0, 1 mM NaCl) was added to the wells, and biofilms were chilled at 4°C for 1 h. After cooling, a cell scraper was used to scrape the cells from the surface, and the cells were transferred to a 1.5-ml microcentrifuge tube. The tubes were centrifuged at 13,000 rpm for 1 min, and the supernatant was removed. EDTA was added at a final concentration of 2.5 mM. After measurement of the OD₆₀₀ of the unwashed biofilm (biofilm biomass), 500 μl eDNA extraction solution (50 mM Tris-HCl, 10 mM ETDA, 500 mM NaCl [pH 8.0]) was added to the wells. The biofilms were scraped off and centrifuged (4,000 × g) for 10 min at 4°C. The eDNA in the supernatant was extracted with phenol-chloroform-isoamyl alcohol (25:24:1), precipitated with absolute alcohol, and resuspended in Tris-EDTA (TE) buffer (PH 7.2). The amount of eDNA was quantified using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA). Moreover, the method for quantification of eDNA was performed as described by Wang et al. with minor modifications (42). The amount of eDNA was quantified using the SYBR Premix Ex Taq kit (TaKaRa, Tokyo, Japan), using primers specific for the gyrase B gene (gyrB). Relative eDNA secretion was calculated as the total amount of eDNA (in nanograms) divided by the extent of biofilm formation (OD₆₀₀).

RNA isolation and sequencing.

Bacteria (4 ml) were cultured in the presence or absence of mupirocin at 37°C with overnight shaking and collected as previously described (12). Briefly, each solution was centrifuged at 12,000 × g for 2 min and combined with 1 ml of 0.1 mm zirconia-silica beads in tubes filled with normal saline, making sure to avoid air bubble formation. The tubes were centrifuged at 4,000 rpm for 1 min using a Mini-Beadbeater (BioSpec, Bartlesville, OK), followed by 1 min of cooling on ice. This procedure was repeated five times. Total RNA was purified using an RNeasy Plus minikit (Qiagen, Berlin, Germany) in accordance with the manufacturer’s protocol. Purified RNAs (1 μg of each) were used as the templates for cDNA production, using a PrimeScript RT reagent kit (TaKaRa, Tokyo, Japan) according to the manufacturer’s protocol.

RNA-Seq was performed in accordance with the Illumina RNA sequencing sample preparation guide. Three biological replicates were prepared for each of the S. aureus strains. To prevent contamination with genomic DNA, RNase-free DNase I (TaKaRa, Tokyo, Japan) was added to the mupirocin-treated and untreated RNA samples. A 2100 Bioanalyzer system (Agilent Technologies Deutschland GmbH) was used to evaluate the RNA quality. After the quality of sample was tested, the Ribo-Zero kit was used to remove the rRNA and enrich the mRNA. Subsequently, the fragmentation buffer was used to break the mRNA into short fragments. Using the mRNA as a template, a single-stranded cDNA was synthesized using random hexamers, and then double-stranded cDNA was synthesized by adding buffer, deoxynucleoside triphosphates (dNTPs), DNA polymerase I, and RNase H. The double-stranded cDNA was purified by AMPure XP beads, and then the second strand of the cDNA, containing U, was degraded by the USER enzyme. The purified double-stranded cDNA was first end repaired, A tailed, and ligated to the sequencing linker, and the fragment size was selected using AMPure XP beads. Finally, PCR amplification was performed, and the PCR product was purified using AMPure XP beads to obtain the final RNA-Seq library. The original sequence was obtained by sequencing alone, which contains low-quality reads with connectors. In order to ensure the quality of information analysis, raw reads must be filtered to obtain clean reads, and subsequent analysis is based on clean reads. A |log₂(fold change)| of >1 and a q value of <0.005 were used to deem a gene differentially expressed.

RNA-Seq data analysis and validation by RT-qPCR.

The resultant cDNA was amplified using a SYBR Green Premix mix (TaKaRa, Tokyo, Japan) and analyzed with Bio-Rad CFX96 Manager software. Primers are listed in Table 3. Cultures of mupirocin-untreated stains were used as positive controls (relative expression = 1), with gyrB expression representing an endogenous control. RNA transcript levels were calculated using the threshold cycle formula, 2^−ΔΔCT. Each reaction was performed in triplicate.

TABLE 3.

Primers used in this study

Primer	Sequence (5′→3′)	Notea
cidA-up-F	GGGGACAAGTTTGTACAAAAAAGCAGGCTATTCTTACCGATGCAGGTCAA	attb1
cidA-up-R	CGGAATTCCGGACCCACGTAATGATGCGCCT	EcoRI
cidA-down-F	CGGAATTCCGAATCCAGCATTAATTGCATCT	EcoRI
cidA-down-R	GGGGACCACTTTGTACAAGAAAGCTGGGTTTGGCATCATACATACCATTT	attb2
psmα-up-F	GGGGACAAGTTTGTACAAAAAAGCAGGCTTATGGCTAGCACAATACAATC	attb1
psmα-up-R	CGGAATTCCGATTCTCAGGCCACTATACC	EcoRI
psmα-down-F	CGGAATTCCGAGATTACCTCCTTTGCTTATG	EcoRI
psmα-down-R	GGGGACCACTTTGTACAAGAAAGCTGGGTAGACTCACGTGGCACTTTCC	attb2
gyrB-F	GGTGGCGACTTTGATCTAGC
gyrB-R	TTATACAACGGTGGCTGTGC
RT-agrA-F	TCCAGCAGAATTAAGAACTCG
RT-agrA-R	ATATCATCATATTGAACATACACT
RT-SarA-F	AAACCCTGAATTTGAATG
RT-SarA-R	GATATTACATCTGCTCCT
RT-icaA-F	CTTGGATGCAGATACTATCG
RT-icaA-R	GCGTTGCTTCCAAAGACCTC
RT-icaC-F	GGGAATCCAATTTCTCTTT
RT-icaC-R	ATACAATGACAGCAGACT
RT-cidA-F	TGTACCGCTAACTTGGGTAGAAGAC
RT-cidA-R	CGGAAGCAACATCCATAATACCTAC
RT-atlA-F	GGTGACACTCGTGCTAAT
RT-atlA-R	AGGGCATGTGAGATAAGA
RT-fnbA-F	ATGGAACGAATGGAACAATAAC
RT-fnbA-R	GTGTGGTAATCAATGTCAAGAG
RT-fnbB-F	CTGACGGTGCCGTAGTTC
RT-fnbB-R	GTCGCCTTTAGCGTTATT
RT-lrgB-F	CATCGGAGGTATTGGTATCG
RT-lrgB-R	GTAGTTGCTGCTTGAGGTAA
RT-bap-F	ATATGAACGCCAAGTATC
RT-bap-R	ACAACGGTATCTATATTAGTAA
RT-psmα-F	GTCCTCCTGTATGTTGAT
RT-psmα-R	AATTCACTGGTAAGTAAGTTAT

^aThe underlined sequences represent the BP reaction sites or restriction enzyme sites.

Construction of ΔcidA and Δpsmα mutants and complemented strains.

The cidA and psmα deletion mutants of the USA300 strain were constructed by allelic replacement using the temperature-sensitive plasmid pKOR1, as described by Bae and Schneewind (44). Briefly, the upstream and downstream fragments of cidA and psmα were amplified by PCR and separately cloned into the pKOR1 vector, resulting in pKOR1-ΔcidA and pKOR1-Δpsmα. The recombinant plasmids pKOR1-ΔcidA and pKOR1-Δpsmα were successively transferred into E. coli DH5α and DC10B, respectively. The cidA and psmα deletion mutants were verified by PCR, RT-qPCR, and direct sequencing and are referred to as the ΔcidA and Δpsmα mutants. The complemented ΔcidA mutant strain was constructed using a shuttle vector, pRB473. Plasmid pRB473-cidA was transformed by electroporation into the ΔcidA mutant, forming the complemented ΔcidA strain (c-ΔcidA). The primers used in this assay are listed in Table 2.

Statistical analysis.

All data analyses were performed using the SPSS statistical software package (version 19; IBM SPSS Statistics). The one-way factorial analyses of variance (ANOVA) tests were used to compare the CLSM data obtained for the strain treated with subinhibitory concentrations of mupirocin and the untreated group. The RT-qPCR and autolysis assay results were analyzed with Prism 7.0 software (GraphPad, San Diego, CA). P values of <0.05 were considered statistically significant.