The bacterial strains and plasmids used or constructed in this study are listed in Tables 1, 2. Luria broth (LB) and LB agar plates were used throughout for growth of Escherichia coli and brain heart infusion (BHI) broth and BHI agar plates for S. aureus. Chloramphenicol was used at 10 μg/ml and ampicillin was used at 100 μg/ml for plasmid selection. Unless otherwise stated, all cultures were grown aerobically at 37°C with shaking, and growth was monitored at 600 nm with a HITACHI U-2800 (Hitachi, Japan) spectrophotometer.

Standard methods for DNA manipulation, preparation, and analysis were employed as described previously (Qazi et al., 2001). Restriction enzymes and T4 DNA ligase were purchased from New England Biolabs (NEB, China) and used in accordance with the manufacturer's instructions. PCR primers (Table 3) were purchased from (Life technologies, Hong Kong). DNA fragments were isolated from electrophoresis agarose (Lonza, Swiss) using gel extraction kit (Qiagen, German). PCR was performed in an ABI thermal cycler 7900 in 50-μL reaction volumes with PrimeSTAR DNA polymerase (Takara, Japan) in accordance with manufacturer's instructions. E. coli Top10 cells (Life technologies, Hong Kong) were transformed by heat shock operation. S. aureus cells were transformed according to published method (Qazi et al., 2001).

To construct the gfp-lux dual-reporter plasmid, the luxABCDE operon was excised from pAL2 as an EcoRI/PstI fragment and inserted into the EcoRI and PstI serial digested plasmid pALC2084 (Bateman et al., 2001), followed by the addition of an adaptor (5′- AATTCTTGCTAGCTTAGATCTTTGCGGCCGCTTGTTTAA ACT-3′) to generate a multiple cloning sites (MCS) with NheI, NotI, and NheI cleavage sites. The GFP fragment digested from pALC2084 with EcoRI was ligated in the upstream of the lux genes without a promoter to generate plasmid pGL (Figure 1). To generate the promoter::gfp-luxABCDE expression vector, PCR primers (Table 3) were used to amplify the different promoters from S. aureus USA300 FPR3757 chromosomal DNA. Each amplicon was restricted with NheI/NotI double-digestion or NheI single digestion and ligated with pGL digested with corresponding restriction enzyme(s). The direction of ligation product from single digestion was confirmed by PCR. All of the constructions and PCR products were verified by sequencing. The schematic representation of the construction of the plasmids is illustrated in Figure 1.

The preparation of total RNA from S. aureus was performed using RNA protection reagent according to the manufacturer's instructions (Qiagen, Germany). Briefly, total RNA was prepared by lysostaphin extraction using 5 × 10⁸ CFU of bacteria at each time point, followed by further purification with an RNeasy kit (Qiagen, Germany) according to the manufacturer's instructions. The quality and quantity of total RNA were confirmed by agarose electrophoresis and UV spectrophotometry, respectively.

Contaminating chromosomal DNA was removed by DNase treatment (Life technologies, Hong Kong). Purified S. aureus RNA was reverse transcripted into cDNA by SuperScript® III First-Strand Synthesis SuperMix (Life technologies, Hong Kong) and then subjected to real−time PCR analysis using an ABI 7500 thermocycler (Life technologies, Hong Kong) using Fast SYBR® Green Master Mix (Life technologies, Hong Kong). The relative quantification of S. aureus transcripts was determined by the ratio of expression of target transcripts relative to gyrB (housekeeping or calibration gene). The sequences of primers for real-time PCR experiments are provided in Table 3.

For the correlation of luxA and virulence gene expression, the equation is as follow:

For the compound treatment, the equation is as follow:

For quantification of GFP fluorescence and bioluminescence, overnight bacterial cultures were diluted in appropriate media containing 10 μg/ml chloramphenicol. Samples (100 μL) were transferred into microtiter plate from culture tubes, and fluorescence was measured by using DTX 800/880 multimode plate reader (Beckman). Bacteria with pGL plasmid were included as control to allow correction for background fluorescence.

For bacteria growth curve monitoring, samples (100 μL) with 10⁶ S. aureus were separated into aliquots in triplicate into clear-bottom 96-well microtiter plates and incubated at 37°C. The optical density at 620 nm (OD₆₂₀), the fluorescence (GFP), and the bioluminescence were measured every 30min in DTX 800/880 multimode plate reader (Beckman).

MIC was determined by inoculating 5 × 10⁴ S. aureus cells in 100 μl BHI media in 96-well plates with a serial dilution of antibiotics. The MIC was defined as the minimum concentration resulting in a cell density less than 0.01 OD at 620 nm (Ohlsen et al., 1998; Ji, 2007), which corresponded to no visible growth, after incubating for 18 h at 37°C.

A single colony of bioluminescent S. aureus from BHI agar was resuspended in 200 μl of sterile water, diluted to 75 ml 0.7% (w/v) soft agar (375-fold dilution) and overlaid onto BHI plates. Antibiotic disks (Becton Dickinson, Mississauga, ON, Canada; Difco, Detroit, MI, USA) were placed on the overlay and the plates incubated at 37°C. After 20 h, inhibition zones were measured and luminescence was detected with a luminograph LB980 photon camera (Berthold, Oak Ridge, TN, USA) and Xenogen IVIS 100 in vivo imaging system (Xenogen, Alameda, CA).

The overlay assay was performed with 24.5 cm × 24.5 cm plates (Corning) with 100 paper discs per plate (Mesak et al., 2010). Each paper disc was loaded with 20 μL of different testing samples and incubated at 37°C for 24 h. The luminescence signals were recorded by luminograph LB980 photon camera. We screened 204 crude extracted samples from different sources, such as lichen, tree, moss and Traditional Chinese Medicine (TCM). A total of 9 virulence factors promoters, srtA, clfA, hla, spa, pvl, psm, fnbB, cap5, and cap8 were used for this screen. The screen was repeated twice. Extracts that inhibit the light emission of more than 3 promoters were assigned as hits. Similar screens were also carried out with in-house chemical compounds.

Overnight bacterial culture with or without ampicillin and/or compound treatment were washed 3 times with PBS (pH 7.4) and then diluted to 10⁷ CFU/ml with MEM medium before inoculation (defined as the original bacterial CFU). A549 cells were seeded onto a 24-well tissue culture plate (Greiner) at a concentration of 2 × 10⁵/ml in MEM for counting bacterial adherence and invasion ratio. Briefly, A549 cells were grown overnight at 37°C in 5% CO₂ to form confluent monolayers. The medium was removed in the following morning and A549 cells were washed twice with 1 ml of PBS, followed by infection with 1 ml of the prepared bacterial inoculum. For invasion assay, after infection of A549 cells at 37°C for 2 h, the supernatants from the wells were collected for total bacterial count (defined as the total bacterial CFU). A549 cells were then washed twice with PBS followed by incubation with MEM containing gentamicin (100 μg/ml; Sigma) and lysostaphin (10 μg/ml; Sigma) for 1 h at 37°C; all wells were then washed 3 times with 1 ml PBS. Subsequently, wells were trypsinized with 150 μl of 0.25% trypsin-EDTA for 5 min, the cells in each well were carefully collected into tubes, and then 400 μl of ice-cold 0.025% Triton X-100 was added to the tubes and put on ice. The numbers of bacterial CFU released from the lysed epithelial cells were determined by plating lysates on BHI agar plates (defined as the invaded bacterial CFU). For adherence assay, after infection of A549 cells at 37°C for 1 h, the medium was removed and A549 cells washed 3 times with 1 ml PBS. Subsequently, a total number of adhered and invaded bacteria released from the lysed epithelial cells was defined as the adhered bacterial CFU.

The bacterial adhesion in each well was determined as the CFU that adhered to and invaded into the cells and is expressed as a percentage of the CFU in the inoculum. The controls were wells pretreated with medium alone (MEM) considered to have 100% adhesion. Adhesion and invasion were then normalized against controls according to the equations.

Each experiment was repeated three times, and all of the relative adhesion and invasion values were calculated and statistically analyzed by Student's t-test, using SigmaPlot software 11.0. P < 0.05 were considered significant (Ji, 2007).

For confocal microscopic analysis of internalized bacteria, A549 cells were seeded onto glass coverslips and incubated at 37°C with 1 ml of the bacterial inoculum for 2 h. After the removal of external bacteria, image capture was done using a Zeiss LSM 700 Inverted Confocal Microscope (Carl Zeiss, Jena, Germany). In order to perform quantitative analysis, a minimum of 5 fields per slide were examined.

Figure 1 is a schematic illustration of the construction of plasmid pGL confirmed by restriction enzyme digestion and sequencing. In this construct, gfp and lux genes were cloned in an operon and co-expressed from the promoters of interest. GFP amplified from pALC2084 was used here for investigating the accumulation of gene products expressed. The lux gene cluster from pAL2 was used for monitoring the real-time gene expression. Transformation frequency of pGL in S. aureus RN4220 and S. aureus USA300 was 1.3 × 10⁴/μg and 8.1 × 10³/μg respectively. Antibiotic resistance markers used were ampicillin in E. coli and chloramphenicol in S. aureus. Promoters of 18 regulators and virulence factors (Table 4), map promoter and ami promoter from pAL2 were amplified and ligated into plasmid pGL. All constructs were confirmed by DNA sequencing.

The plasmids were successfully constructed and introduced into S. aureus strains by electroporation. The transformants were confirmed by plasmid isolation and fluorescence and bioluminescence readings in liquid culture.

When bioluminescence was monitored in S. aureus USA300-pGLspa grown in liquid culture, gene expression driven by spa promoter peaked during exponential phase and diminished when reaching post-exponential phase (Figure 2), similar to what was reported earlier (Gao and Stewart, 2004). Consistent with the study reported by Qazi et al. (2001), GFP fluorescence was detected later during the growth of the bacteria, and the fluorescence signal increased for a longer period of time (Figure 2). This observation may due to a lag period needed for GFP folding and the well-known stability of GFP (>24 h). Our data were in accordance with the results from Qazi (Qazi et al., 2001) that Lux may serve as a real-time reporter of promoters while GFP may be used for the observation of accumulation effects of gene expression. The combination of both reporters in a dual-reporter system will be appropriate for carrying out high-throughput screening for unknown repressors or inducers of promoters.

Taking protein A as an example, expression driven by the spa promoter showed a similar phase-dependent effect as reported (Gao and Stewart, 2004), the correlation of luminescence signal, LuxA gene expression, and spa expression was confirmed in our system. Using the house-keeping gyrB gene as a standard, the relative ratio of luxA and spa expression was calculated, plotted in Figure 3 and correlated with luminescence signal normalized by CFU. As shown in Figure 3, the luminescence signal fully represented the expression pattern of luxA and spa gene, which were all under the control of spa promoter. The correlation among relative expressions of luxA, virulence gene, and normalized luminescence indicated that the luminescence signal can represent the endogenous virulence gene expression.

After confirming the relationship between luminescence and gene expression, a well-studied environmental factor CO₂ (Ohlsen et al., 1997) was used to analyze the CO₂ effects on the expression of 8 genes (agr, sarA, hla, spa, saeP1, saeP3, sarS, and RNAIII) closely related to S. aureus virulence (Schmidt et al., 2003). Results showed that CO₂ affected the expression level of protein A by repressing the activity of the spa promoter and inducing the expression of hla gene (Figure 4A); as reported by Ohlsen et al. (1997). As sarA and agr regulate hla and spa (Schmidt et al., 2003), in this pathway, up-regulated sarA may repress spa expression and induce hla expression directly and indirectly through inducing agr (RNAIII) to repress sarT expression, which may further induce hla expression and repress spa expression (Figures 4A–D). Furthermore, consistent with the report of Herbert et al. (1997), activities of type 5 promoter was reduced in the presence of CO₂ (Figure S1). The platform may be used for the illustration of the interplay between regulators and virulence factors.

It has been shown that sub-inhibitory concentrations of β-lactam antibiotics strongly induce hla expression (Ohlsen et al., 1998). Applying this platform, 5 different β-lactam antibiotics were tested on different virulence factors' promoters (Figure 4E). Meropenem, cefotaxime, cefaclor and cefoxitin target penicillin-binding protein 1 (PBP1), PBP2, PBP3, and PBP4 respectively and have shown induction effect on hla, psm, clfA, fnbA, cap5, and cap8 at subinhibitory concentrations. The induction effect was also monitored on ampicillin. We observed different induction level of these antibiotics on cap5 promoter and different induction level of one antibiotic on different promoters' activities. The induction of clfA and fnbA may lead to induced adherence and internalization. The platform may be used for the investigation of modulating effects of extrinsic factors on virulence gene expression.

As ampicillin was shown to induce nearly all of the tested virulence-related promoters (Figures 4B, 7B and Table 5), we tested if sub-inhibitory concentrations of ampicillin increase invasiveness of S. aureus in epithelial cells. We prepared USA300 overnight culture with or without ampicillin treatment to infect A549 cells (human lung epithelial cells) so as to evaluate the cell adherence and cellular invasion of the bacteria. To illustrate the enhanced intracellular localization of the bacteria by ampicillin in the invasion assay, GFP fluorescence from the engineered bacteria was used to trace the localization of the bacteria. Images taken from confocal fluorescence microscope indicated that the bacteria were internalized into the A549 cells and ampicillin treatment apparently increased the number of bacteria internalized (Figure 5A). After calculating the ratio of bacteria adherence and internalization of S. aureus cells (Ji, 2007), the bacteria adherence ratio and bacterial invasion ratio of ampicillin-treated bacteria were considerably higher than that without the ampicillin treatment (Figure 5B). The induction effect on adherence factors at subinhibitory concentrations of ampicillin has led to induced adherence and invasion of the bacteria.

The multiplex promoter reporter platform was used directly for screening potential virulence repressors from a natural products library. We used COL with plasmid pGLcap5, strain AE052 with plasmid pGLcap8 and strain USA300 with other plasmids to screen compound libraries. For the crude extracted samples from different sources, such as lichens, tree, mosses and TCM, used in the screening (Figure 6), 20 samples were found to reduce luminescence signals of more than 3 (out of the 7) different promoters. Five hits out of 208 samples were showing the suppressing effect on more than 4 promoters and 13 hits showing the suppressing effect on all the 9 promoters. We then selected 3 samples with different repression profiles showing their repression effects on 7 selected virulence-related promoters (hla, spa, pvl, psm, fnbB, cap5, and cap8, Figures 7A,B). The data showed that sample 15 extracted from Abies grandis (grand fir) reduced the activity of 5 promoters except cap5 and cap8; sample AE62D extracted from Sphaerophorus globosus (coral lichen) reduced expression of all 7 promoters tested and sample AE63 extracted from Usnea filipendula (beard lichen) reduced exotoxins, for example, hla, pvl, and psm but induced luminescence signal of cell surface-associated virulence factors like spa, fnbB, cap5, and cap8 (Table 5).

Among the 5 hits showing inhibition effect on more than 4 promoters, 3 of them were extracts from Usnea species, namely Usnea subfloridana, Usnea filipendula, Usnea rigida, and the other two were from Sciadopitys verticillata and Cryptomeria japonica. The effective patterns of promoters activities were shown in Table 6. As 4 of them were from usnea, and usnic acid is uniquely found in lichens and is especially abundant in genera such as Alectoria, Cladonia, Usnea, Lecanora, Ramalina, and Evernia, usnic acid may be the common effective component in these samples. The effect of Usnic acid was also monitored compared with crude extracts from lichens. As shown in Figure 8, usnic acid nearly replicated the relative activity of difference extract. They also showed repression effects on psm, hla, pvl, and clfA, while induction effects on spa and cap8. These results indicated that the main effective component in these lichens may be usnic acid.

We also monitored the growth of USA300 with different concentration of usnic acid, and 12.5 μM usnic acid did not affect bacterial growth while the MIC was around 25 μM (Figure 9A). By analyzing the promoter activity monitored by luminescence signal, we found that usnic acid mainly reduced expression of psm (Figure 9B). This was confirmed by real time-PCR results in Figure 9C. In all the time points, the expression of psm was reduced for more than 100-fold, while spa and fnbA were induced for more than 100-fold. In some of them, hla and pvl were repressed, indicating a regulating network among these virulence factors.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.