INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Staphylococcus aureus is the etiologic agent of numerous infections in humans and domesticated animals and has been implicated in a multitude of diseases, ranging from minor wound infections to more serious diseases, including endocarditis, osteomyelitis, and septic shock (reviewed by Projan and Novick [34]). The expression of many S. aureus virulence factors is under the control of the accessory gene regulator (agr) which, on entering post-exponential phase, downregulates the production of cell surface-associated proteins and upregulates the expression of secreted toxins and extracellular enzymes (28, 33, 38). The role of the agr regulon is supported by in vivo studies, which show that agr mutants are greatly attenuated in several animal models, including intramammary infections (13), arthritis in mice (1), and endocarditis in rabbits (7). The agr locus is a quorum-sensing-regulated system activated by autoinducing peptide pheromone (AIP) (21, 25). The agr locus consists of two divergent transcriptional units, RNAII and RNAIII, which are under the control of the P2 and P3 promoters, respectively (reviewed by Novick and Muir [30]). RNAII is a polycistronic mRNA that encodes the agrB and agrD genes required for the synthesis of the AIP and also the two component signal transduction proteins, AgrA and AgrC, which are responsible for sensing and responding to the AIP. RNAIII is the effector molecule in the agr regulon acting primarily at the level of gene transcription. Different S. aureus strains produce AIPs with distinct structures, and strains can be grouped on this basis since they will activate the agr response of strains within the same group and inhibit the agr response of strains from different groups by competitive inhibition (21, 30). This inhibitory action of AIPs has identified them as potential novel therapeutic and anti-infective agents for S. aureus.

S. aureus is primarily known as an extracellular pathogen; however, it has been shown that endothelial cells can act as nonprofessional phagocytes and promote the uptake of S. aureus (11, 15, 31). Other groups have shown that S. aureus is able to internalize and survive in a wide variety of mammalian cells (2, 5, 19, 45). S. aureus invades nonprofessional phagocytes via a mechanism that requires a specific interaction between fibronectin-binding proteins and the host cell. This subsequently leads to host cell signal transduction through protein tyrosine kinases and cytoskeletal rearrangements (12, 24, 32, 41) and uptake of the bacteria into an endosome. Bayles et al. (5) have shown that S. aureus is able to escape from this endosome, leaving the bacterial cells to survive and possibly multiply within the cytoplasm; however, the mechanism by which this endosomal membrane is breached has not been elucidated. Internalization experiments using a pulmonary epithelial cell line have demonstrated the ability of internalized S. aureus to replicate intracellularly (22). It is now believed that intracellular replication plays an important role in the frequency and persistence of invasive staphylococcal infections, perhaps by providing protection against both host defenses and antibiotic treatment.

Based on observations that agr mutants and also cells in exponential phase are internalized more efficiently, Wesson et al. (47) proposed a model for the function of agr-mediated quorum sensing in staphylococcal invasion of cells: in an extracellular environment, levels of AIP are low due to dilution into surrounding fluids and S. aureus expresses the cell wall-associated factors that promote binding to host cell surfaces and subsequent internalization. The expression of these surface proteins is known to be repressed by induction of the agr regulon. Upon internalization, bacteria are surrounded by an endosomal membrane, the presence of which may allow concentrations of AIP to rapidly accumulate and trigger expression of agr and lead to repression of the cell wall-associated proteins and the production of agr-regulated exoproteins, such as the hemolysins, which then facilitate bacterial release.

One way in which bacterial localization, movement, and gene expression can be studied is through the use of reporter genes. There are a number of reporters available for such investigations, e.g., green fluorescent protein (GFP) and bacterial luciferase (lux), which have their own particular advantages and disadvantages as reporter and/or marker genes. The major disadvantages of lux are that it uses reduced flavin mononucleotide as an energy source and therefore requires live cells for signal generation (17, 27), so gene expression cannot be assessed on fixed samples and spatial resolution by photon-counting microscopy is poor (18). However, the short half-life of Lux proteins allows gene expression and promoter kinetics to be monitored in "real time," and signal can be detected with great sensitivity with little background luminescence interfering with signal detection. GFP, on the other hand, suffers from poor sensitivity and high-background fluorescence problems and cannot give a real-time representation of promoter kinetics. This is due both to the time taken for the chromophore to fold and generate fluorescent protein (8, 16) and to the long half-life of the protein once it has formed (44). Clearly, the utilization of both of these reporters represents an opportunity to capitalize on the complementary strengths of each so that gene expression can be assessed both in "real-time" in vivo by luminometry and/or photonic imaging and also "retrospectively" on fixed samples by fluorescence microscopy or fluorometry.

In this study we describe the construction and evaluation of lux-gfp dual operons expressed in S. aureus. The use of a growth-phase-dependent promoter linked to bioluminescence measurement has led to the development of a novel, noninvasive technique for monitoring S. aureus internalization and subsequent replication inside eukaryotic cells. Induction of the agr regulon in both broth culture and during internalization by bovine mammary epithelial (MAC-T) cells was also monitored and has facilitated determination of agr expression during this process.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. The bacterial strains and plasmids used or constructed during this study are listed in Table 1. Throughout this study a red-shifted gfp variant, gfp mutant 3 (gfp3 [10]) was utilized. Except where stated, Luria broth and Luria plates (40) were used throughout for growth of Escherichia coli and S. aureus. Chloramphenicol was used at 7 µg/ml and ampicillin was used at 50 µg/ml for plasmid selection, as appropriate. Unless otherwise stated, all cultures were grown aerobically at 37°C, and growth in liquid culture was monitored at 600 nm (Cecil 2000 series spectrophotometer).

Preparation, manipulation, and analysis of DNA. Standard methods were performed as described by Ausubel et al. (4) using enzymes supplied by Boehringer-Mannheim in accordance with the manufacturer's instructions. PCR primers (Table 2) were supplied by Genosys Biotechnologies (Europe), Ltd. T4 DNA ligase (Promega) was used for ligations. DNA fragments were isolated from low-melting-point agarose (FMC Bioproducts) via a freeze-thaw extraction method (37). PCR (39) was performed in a Techne Progene thermal cycler in 50-µl reaction volumes with Taq DNA polymerase (Advanced Biotechnologies, Ltd.) in accordance with manufacturer's instructions. E. coli JM109 cells were transformed by electroporation as described by Sambrook et al. (40), S. aureus cells were transformed according to the method of Augustin and Gotz (3).

Construction of dual reporter expression vectors. The PCR primers are described in Table 2, and the plasmid pBluelux (Table 1) was used as a template; a luxAB amplicon was restricted with EcoRI/KpnI and inserted into pHG327 (42) to give pSB2023. A luxCD amplicon was restricted with KpnI/BamHI and a luxE amplicon with BamHI/PstI. These were ligated into pSB2023 restricted with KpnI/PstI. Recombinant plasmids (pSB2024) were selected by using ampicillin and screened for bioluminescence in the absence of exogenous aldehyde by using a Hamamatsu VIM3 intensified video camera (Hamamatsu Photonics United Kingdom, Ltd.). The modified luxABCDE operon was then subcloned into the superlinker plasmid pSL1190 (Pharmacia), which had been restricted with MunI/PstI to give plasmid pSB2025. The luxABCDE cassette was then placed downstream of gfp in pSB2019 (36) as a SalI/PstI fragment to generate the growth-dependent reporter plasmid pSB2030. To generate the agr P3::gfp,luxABCDE expression vector, PCR primers 7 and 8 (Table 2) were used to amplify the P3 promoter from S. aureus 8325-4 chromosomal DNA. The amplicon was restricted with EcoRI/SmaI and ligated with pSB2019 (36) that had been restricted with EcoRI/SmaI to excise P_xylA to create plasmid pSB2031. The luxABCDE operon was excised from pSB2025 (SalI/PstI) and inserted downstream of gfp in pSB2031, generating a dual reporter designated pSB2035.

Gene expression measurement of bacterial cultures. GFP was detected using a Nightowl CCD camera system with integrated fluorescence excitation (Perkin Elmer Instruments) or by eye using a blue LED for excitation of GFP. For quantification of GFP, overnight bacterial cultures were diluted 1/100 into prewarmed medium containing the necessary antibiotics. A 1.5-ml sample was centrifuged at 13,000 × g for 2 min, washed twice in an equal volume of phosphate-buffered saline (PBS), and then concentrated 10-fold in PBS. Samples (150 µl) were transferred into microtiter plate wells, and fluorescence was measured by using Victor 1420 multilabel counter (Perkin-Elmer Instruments). A control sample of nontransformed bacteria was included to allow correction for background fluorescence.

Preparation of cells for agr induction experiments. Bacteria harboring agr P3 expression vectors were grown overnight in broth containing chloramphenicol. Cells were centrifuged (5,000 × g) and then washed with an equal volume of fresh medium to remove accumulated AIPs. Bacteria were diluted 1/20 into fresh medium and grown for 2 h before the culture was again diluted 1/20 into fresh medium and grown for a further 2 h. Finally, these bacterial cultures were diluted 1/50 into fresh medium to produce cells in the mid-exponential phase of growth without significant accumulation of AIP. To investigate the response of the agrP3 reporter to exogenous AIPs, a crude preparation of the group I AIP was prepared as filtered spent overnight culture supernatants of RN6390 and added to a final concentration of 10% (vol/vol). Alternatively, either the activating AIP (group I peptide [21, 25]) or inhibitory AIP (S. lugdunensis) synthesized as described by McDowell et al. (26) was added, and the reporter gene activity was monitored (as described above) over a specific time period.

Cell invasion assays. The bacterial inoculum was prepared as described for the agr assay. For the final growth cycle, bacterial cells were washed twice in an equal volume of Dulbecco modified Eagle medium (DMEM; Sigma) and finally resuspended in 1/10 volume DMEM and inoculated at a 1/400 dilution into medium (HEPES-buffered DMEM plus 10% [vol/vol] RPMI). These were then grown for ca. 4 h until an OD₆₀₀ value of 0.1 to 0.2 was reached.

Bacterial internalization assays for microscopic analysis. MAC-T cells that had been seeded onto glass coverslips were incubated at 37°C with 1 ml of the bacterial inoculum. After invasion, the monolayer was washed three times with PBS and then incubated with lysostaphin (10 µg/ml) in DMEM and incubated for 20 min at 37°C before three washes with PBS. Immunostaining was carried out as described by Sambrook et al. (40) with the modification that anti--tubulin Cy3 conjugate (1/25 in PBS; Sigma) was used to visualize the microtubule network. During the last 10 min of the staining procedure, DAPI (4',6'-diamidino-2-phenylindole; 50 µg/ml; Sigma) in PBS was added for visualization of the GFP-negative bacteria and eukaryotic DNA. Epifluorescent microscopy was carried out with a Zeiss Axiovert 135TV fluorescence microscope equipped with a Prior motorized stepper stage. Excitation was done with a polychrome II monochromator (T. I. L. L. Photonics) with triple-pass dichroic filter and single-pass emission filters (Omega Optical) mounted in a Biopoint filter wheel. Image capture was done with a Hamamatsu ORCA-2 cooled CCD controlled by Openlab software (Improvision). Images were captured as 1-µm Z-stacks that were deconvolved by using the Openlab volume deconvolution algorithms and merged for presentation. A minimum of 20 fields per slide were examined for qualitative microscopic analysis.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction of gfp-lux dual operon and expression in S. aureus The native luxCDABE operon of Photorhabdus luminescens is expressed very poorly in S. aureus when linked to a gram-positive promoter (35). To modify the lux operon for high expression in gram-positive bacteria, enhanced translational signals (46) were introduced in front of luxA, luxC, and luxE by using PCR primers incorporating these sequences (Table 2). These primers were used to amplify luxAB (primers 1 and 2), luxCD (primers 3 and 4), and luxE (primers 5 and 6). The purified amplicons were restricted using the restriction sites introduced via the PCR primers (Table 2) and cloned into pSL1190 (Pharmacia) in the gene order luxABCDE (see Materials and Methods for details). Recombinant plasmids were identified by their bioluminescent property in the absence of exogenous aldehyde substrate and designated pSB2025. To create a dual reporter operon, the luxABCDE operon was excised from pSB2025 as a SalI/PstI fragment and inserted into the gram-positive gfp reporter plasmid SB2019 (36), downstream of the gfp gene and under the control of the xylA promoter from Bacillus megaterium to generate pSB2030.

View larger version (18K): [in this window] [in a new window]   FIG. 1.   Growth-dependent expression of the gfp-luxABCDE dual reporter from PxylA. S. aureus RN6390(pSB2030) was grown in 1-ml volumes in a 24-well microtiter plate in DMEM supplemented with 10% RPMI. Samples were incubated at 37°C in a Victor 2 Multilabel Counter, and growth (OD600; ), luminescence (counts/second, panel A; ), and fluorescence (panel B; ) were measured for 10 h.

Expression of the dual gfp-lux operon from agrP3. To construct a dual expression vector to study agr gene expression, PCR primers 7 and 8 (Table 2) were used to amplify the P3 promoter from S. aureus 8325-4, and this was used to replace the xylA promoter in the gram-positive gfp plasmid pSB2019 (35); luxABCDE was then inserted SalI/PstI downstream of gfp, creating plasmid pSB2035. To confirm that this new reporter construct accurately reflected P3 activity, pSB2035 was introduced into S. aureus 8325-4 (agr group I). Cells were grown to mid-log phase after repeated subculture and washing to remove naturally produced AIP from the culture supernatant. Synthetic group I AIP (Fig. 2A) was added to these cells, and bioluminescence was measured as a reporter of P3 induction. The results of these experiments indicate that the agrP3 promoter fusion is activated by the synthetic group I AIP (Fig. 3A). The plasmid-encoded agr-P3 promoter exhibits a dose-dependent response to the activator molecule, with lower levels of the AIP activator leading to a lower luminescent output. When no activator was added to the bacterial culture, the levels of luminescence observed were ca. 10-fold lower than the induction seen after addition of even the lowest concentrations of AIP. Since S. aureus 8325-4 is not an agr mutant, it is able to produce its natural group I AIP; hence, some expression of P3 is expected in the absence of exogenous AIP. These data indicated that the bioluminescence genes are effective reporters of agrP3 induction.

View larger version (12K): [in this window] [in a new window]   FIG. 2.   Structures of S. aureus group I AIP (25) (A) and S. lugdunensis AIP (26) (B).

View larger version (25K): [in this window] [in a new window]   FIG. 3.   S. aureus 8325-4(pSB2035) response to activating and inhibitory AIPs. S. aureus 8325-4(pSB2035) was grown at 37°C in 96-well plates in an Anthos Lucy 1 photoluminometer. OD600 and luminescence readings (in relative light units [RLU]) were taken over a period of 20 h. These data are plotted as specific luminescence (RLU/OD) versus time to normalize changes in cell density. (A) Cells were grown in Luria broth alone () or Luria broth supplemented with group I AIP at 5 µM (), 25 µM (), 50 µM (), 75 µM (), or 100 µM (). (B) Cells were grown in Luria broth supplemented with 10% spent culture supernatant () or also supplemented with S. lugdunensis AIP at 1 µM (), 5 µM (×), 10 µM (), or 50 µM (+).

Development of a staphylococcal internalization assay. The strain of S. aureus chosen for internalization assays was RN6390 since it has previously been shown to be both virulent in several animal models (6, 7) and internalized successfully by MAC-T cells (5, 47). To determine whether we could use the reporter genes to monitor S. aureus growth during the invasion of MAC-T cells, S. aureus RN6390(pSB2030) was used. Bioluminescence from all samples peaks at 120 min, regardless of the presence of MAC-T cells, and then decreases (Fig. 4). In the wells containing MAC-T monolayers, bioluminescence increases again after 200 min. This is in marked contrast to the bacteria incubated in wells without MAC-T cells, where bioluminescence decreases to background levels (Fig. 4). The luminescence from the cultures seen over the first 120 min represents the growth of S. aureus in the tissue culture medium, followed by the expected decrease in bioluminescence when these bacteria enter stationary-phase growth and downregulate P_xylA. The observed increase in bioluminescence after 200 min when MAC-T cells are present is believed to be due to replication of S. aureus on the surface of, or within, MAC-T cells. To test this hypothesis, cytochalasin D, which inhibits F-actin polymerization in the MAC-T cells and compromises their ability to internalize S. aureus (5, 22), was added to the infected MAC-T cells prior to and during infection. In this case, the level of bioluminescence in the second peak is lower than that in wells with no cytochalasin D, commensurate with a reduction in the number of internalized bacteria (Fig. 4).

View larger version (32K): [in this window] [in a new window]   FIG. 4.   Invasion of MAC-T cells by S. aureus RN6390(pSB2030). S. aureus RN6390(pSB2030) were used to inoculate a 24-well plate containing MAC-T monolayers, in DMEM supplemented with 10% RPMI with (×) or without () 1 µg of cytochalasin D/ml. As controls, bacteria were also inoculated into wells containing medium alone with () or without () 1 µg of cytochalasin D/ml. All samples were incubated at 37°C, and luminescence readings (in counts/second [cps]) taken over a 10-h period in a Victor 2 Multilabel Counter.

View larger version (18K): [in this window] [in a new window]   FIG. 5.   Measurement of intracellular growth of S. aureus RN6390(pSB2030). S. aureus RN6390(pSB2030) were used to inoculate a 24-well plate containing MAC-T monolayers, in DMEM supplemented with 10% RPMI with (×) or without () 1 µg of cytochalasin D/ml. As controls, bacteria were also inoculated into wells containing medium alone with () or without () 1 µg of cytochalasin D/ml. After a 2-h infection period, all samples were treated with lysostaphin, and then the medium was replaced with DMEM. The plate was incubated at 37°C, and luminescence readings (in counts/second [cps]) were taken over a 10-h period in a Victor 2 Multilabel Counter.

Analysis of patterns of agrP3 expression by S. aureus upon MAC-T invasion. To examine agr expression in an intracellular environment, MAC-T cell invasion assays were performed with the agrP3 reporter strain S. aureus RN6390(pSB2035). As above, MAC-T monolayers were incubated with bacteria both in the presence and in the absence of cytochalasin D with lysostaphin treatment after a 2-h infection period. As controls, bacteria were incubated in assay wells alone, with or without cytochalasin D. In these control experiments no induction of either promoter was seen (Fig. 6). In the MAC-T internalization assay, bioluminescence from S. aureus RN6390(pSB2035) was at a maximum after 100 min (Fig. 6). In parallel infection experiments performed with S. aureus RN6390(pSB2030) (P_xylA), luminescence did not peak until ~370 min (Fig. 6), a finding which indicates bacterial replication (Fig. 1). It is known that internalized S. aureus are surrounded by an endosomal membrane (5, 19), from which they then escape by lysis (5). Our data now suggest that agr is induced to high levels while the bacteria are within the endosome (as illustrated by high luminescence output during the first 100 min of internalization) and is followed by replication on release into the cytoplasm. Hence, it is likely that production of agr-regulated exoproteins results in endosomal lysis.

View larger version (25K): [in this window] [in a new window]   FIG. 6.   Growth and expression of agr by S. aureus RN6390 in MAC-T cell invasion assay. To measure agrP3 expression, S. aureus RN6390(pSB2035) was inoculated into a 24-well plate containing MAC-T monolayers, in DMEM supplemented with 10% RPMI with () or without () 1 µg of cytochalasin D/ml. As controls, bacteria were also inoculated into wells containing medium alone with () or without () 1 µg of cytochalasin D/ml. To monitor staphylococcal replication, S. aureus RN6390(pSB2030) was used to inoculate wells containing MAC-T monolayers with (×) or without () 1 µg of cytochalasin D/ml. After a 2-h infection period, all wells were treated with lysostaphin and then the medium was replaced with DMEM. The plate was incubated at 37°C, and luminescence readings were taken over a 10-h period in a Victor 2 Multilabel Counter. The data are plotted as the percent maximum luminescence for each reporter construct to compensate for variations in promoter strength.

View larger version (78K): [in this window] [in a new window]   FIG. 7.   Microscopic evaluation of growth and agrP3 expression by S. aureus RN6390 invading MAC-T cells. MAC-T monolayers on coverslips were inoculated with S. aureus RN6390(pSB2035) in DMEM supplemented with 10% RPMI. The specimens were treated with lysostaphin after 2 h and stained with anti--tubulin-Cy3 conjugate (red) and DAPI (blue) at 2 h (a and b), 4 h (c), and 6 h (d) postinfection. The specimens were visualized by fluorescence microscopy.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The lux operon from P. luminescens has been reconstructed to contain enhanced translational signals for genes whose sequence analysis suggested might be poorly translated. Three genes in this operon, luxA, -C, and -E, were engineered to contain optimized gram-positive translational initiation sequences. This was combined with gfp that had been similarly modified (36) to construct a dual reporter operon and expressed in S. aureus to allow studies of bacterial growth and gene expression during cell invasion.

The gfp-lux dual reporter downstream of the B. megaterium xylA promoter, which is expressed in a growth-dependent manner, has been used to investigate ex vivo the intracellular replication of S. aureus in MAC-T cells. When MAC-T monolayers were incubated with the bacteria for the duration of the invasion assay, two peaks of bioluminescence were observed. The first corresponds with luminescence emitted from bacteria incubated in the absence of MAC-T cells and represents bacterial replication in the cell culture medium. The second peak of bioluminescence was shown to be due to replicating S. aureus both on and within the eukaryotic cells. To measure the replication of intracellular S. aureus alone, the assay was further modified by the inclusion of a lysostaphin treatment after a 2-h invasion period. In this improved assay, only wells containing MAC-T monolayers inoculated with S. aureus in the absence of cytochalasin D exhibited a bioluminescent output since cytochalasin D inhibits the uptake of S. aureus by MAC-T cells (5). Thus, by using the new dual-reporter operons, we have developed an assay to study intracellular replication of S. aureus, monitoring the event in real time by using the lux genes and in fixed samples by using the GFP signal. When this latter reporter is used, time is needed to allow the posttranslational modification required to produce the functional chromophore; however, we have demonstrated its effectiveness in the microscopic analysis of cell invasion.

For S. aureus to replicate within MAC-T cells, the bacteria must first escape from the encapsulating endosomal membrane (5, 19). agr expression has previously been reported to be important for S. aureus intracellular survival (47), although the specific role of agr in either endosomal escape or intracellular replication has not been elucidated. The use of the new nondestructive dual reporters to measure agr expression in our novel cell invasion assay system has now allowed this question to be addressed. Previously, Wesson et al. (47) demonstrated that the agr mutant, RN6911, was internalized at higher efficiencies in MAC-T cells than the wild-type strain RN6390. This was not unexpected, since the agr mutants were known to produce increased levels of the cell surface binding proteins such as the fibronectin-binding proteins. However, although this agr mutant was internalized, it was unable to replicate intracellularly; in fact, the numbers of viable bacteria decreased over time. Wesson et al. hypothesized that agr-regulated products are necessary for escape from the endosome and further growth in the cytoplasm (47). Furthermore, these authors suggested that, due to the confined space within the endosome, the accumulation of AIP is rapid and lysis of the endosome may be due to the induction of agr-regulated exoproteins. This hypothesis has now been substantiated by our data with the agrP3 dual reporter in which bioluminescence from the agr reporter is induced within the first 100 min of internalization prior to bacterial replication, as monitored by the expression of P_xylA, which is not induced until the cells are released into the host cell cytoplasm. Recently, it has been found that the S. aureus strains from the NCTC 8325 lineage (including RN6390 and 8325-4) are downregulated in sigB expression due to a defect in rsbU (14), which may in turn affect the level of agr expression if internalization is perceived as a stress condition. To address this, we carried out preliminary studies using a clinical isolate (WCUH29) with an intact sigB locus, and in this background the promoter kinetics are in agreement with those seen with RN6390.

The involvement of agr in internalization and subsequent replication in MAC-T cells has also been demonstrated microscopically by utilizing the GFP protein in combination with fluorescent dyes. By using the agr dual reporters, it has been shown that agr expression is low upon internalization, as indicated by low gfp expression. Cell surface-associated proteins are highly expressed under these conditions, as illustrated by the binding of the antitubulin conjugate to protein A in the bacterial cell wall, which gives a phenotypic indicator of agr expression to further validate the reporter data. At both 4 h and 6 h postinfection, the number of GFP⁺ staphylococci increases significantly. At these time points, no detectable levels of protein A were observed by immunofluorescence, a finding again indicative of upregulation of agr and a corresponding decrease in cell surface-associated proteins.

The data gathered during the course of this study show that the novel dual-reporter gene system is a powerful, nondestructive tool for gene expression studies ex vivo and potentially in vivo. For example, lux expression coupled with low-light imaging can allow the visualization of bacterial gene expression in complex environments such as whole animals (9), providing information about the temporal and spatial regulation of particular genes. Coupling this approach with histologic studies using the GFP signal to give a "retrospective" measure of which bacterial genes were expressed in which tissues or cells would give a more detailed picture of staphylococcal pathogenesis.