INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Staphylococcus aureus, a major pathogen both in the community and in hospitals (27), has a highly invasive nature. Once the organism gains access to the bloodstream, patients are at risk of developing serious diseases such as endocarditis and other metastatic complications (4). Despite the use of newer antimicrobial agents, the morbidity and mortality from serious S. aureus infections remain high (11). The recent emergence of vancomycin-resistant strains in Japan and subsequently in the United States further underscores the importance of identifying alternative strategies for the development of novel antimicrobial strategies to manage invasive S. aureus infections.

It is generally recognized that the pathogenesis of S. aureus infections is complex and involves the coordinate expression of multiple gene products (22). However, the majority of the data on S. aureus virulence have evolved from in vitro studies of bacterial cells at a particular growth phase, usually in nutrient-rich liquid medium. A major impediment to directly applying these data to in vivo conditions is the finding that this organism can significantly alter its phenotypes in response to changing microenvironments (22, 24). These in vitro studies also ignore the interplay of the organism with important tissues and host defense mechanisms, including host proteins and phagocytes to which the organism is exposed. Recognizing that host factors likely modulate the expression of microbial virulence determinants, we wanted to characterize the in vivo expression of sar, a global regulatory locus of S. aureus that up-regulates the expression of both extracellular virulence determinants (e.g., hemolysins) and cell wall-associated virulence determinants (e.g., fibronectin-binding proteins) in vitro. In this study, we describe the use of a green fluorescence protein (GFP) reporter gene system to examine sar promoter activation in vivo, using a model of invasive S. aureus infection (rabbit endocarditis).

Genetic analyses have indicated that the sar locus is composed of three overlapping transcripts, with a common 3' end but initiated from three distinct promoters designated P1, P3, and P2 (2). In broth cultures, the promoters P1 and P2 are activated during the exponential phase and expressed less as cells move toward the stationary phase. In contrast, the P3 promoter, being dependent on the alternative sigma factor, SigB (13, 20), is maximally expressed during the postexponential phase. Using XylE transcriptional fusions, we recently showed that the P1 promoter is the strongest, with ~30-fold more activity than P2 and P3 (20). To examine and compare the activation of individual sar promoters in vitro and in vivo, we recently constructed an Escherichia coli-S. aureus shuttle vector containing a promoterless gfp_UV reporter gene (a gfp derivative [Clontech, Palo Alto, Calif.] optimized for expression in prokaryotes) preceded by a polylinker region. By linking individual sar promoters to gfp_UV, we found that the expression of GFP_UV in this vector system is dependent on the strength of the upstream promoter. To examine the activation of the sar promoters in vivo, S. aureus strains containing individual sar promoter-gfp_UV transcriptional fusions were injected into rabbits that had been catheterized to induce endocarditis. Upon sacrifice, fluorescence microscopy revealed that the sar P1 promoter was active both in vivo and in vitro (on agar plates), while the P3 promoter was silent in both scenarios. Remarkably, the P2 promoter was silent in vitro but became active in vivo, and it appeared to be differentially expressed within different parts of the infected tissues.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains and plasmids. Bacterial strains and plasmids are listed in Table 1. Phage 11 was used as the transducing phage for S. aureus strains. CYGP, 0.3GL media (21) and tryptic soy broth were used for the growth of S. aureus strains, while Luria-Bertani medium was used for growing E. coli. Antibiotics were used at the following concentrations: tetracycline, 5 µg/ml; chloramphenicol, 10 µg/ml; and ampicillin, 50 µg/ml.

Cloning strategies. The plasmid pGFPuv (Clontech) contains a gfp derivative that has been optimized for expression in an E. coli host. To optimize expression of GFP_UV in an S. aureus host, we introduced a sarA ribosomal binding site (9) upstream of gfp_UV, thereby enhancing translation in S. aureus. Expermentally, we amplified by PCR a ~750-bp fragment encompassing the gfp_UV gene preceded by the sarA ribosomal binding site with the following primers containing restriction sites: 5'-ACGCGTCGAC(SalI)-TAGGGAGAGGTTTTAAAC-²⁸⁹ATGAGTAAAGGAGAAGAACTT³⁰⁹-3' (the fragment containing the sarA ribosomal binding site is in boldface) and 5'-AACTGCAC(PstI)-¹⁰⁰⁵TTATTTGTAGAGCTCATCCAT⁹⁸⁵-3' (numbers indicate nucleotide positions in gfp_UV). This fragment was first cloned into the PCR cloning vector pCR2.1 (Invitrogen, Carlsbad, Calif.). The recombinant vector was cut with SalI, blunted with T4 polymerase, cleaved with PstI, and analyzed in a 1% SeaPlaque gel (FMC Inc., Rockland, Maine). The ~750-bp fragment was gel purified and ligated to the HincII/PstI site of the polylinker region of the E. coli-S. aureus shuttle vector pSK236, a chimera of pUC19 and pC194 (12). The recombinant plasmid containing the promoterless gfp_UV reporter gene, designated pALC1434, was amplified in E. coli and verified by restriction analysis and DNA sequencing.

Genetic manipulation of S. aureus. Shuttle plasmids were transformed into S. aureus RN4220 by electroporation as described by Schenk and Laddaga (25). Transformants were selected on tryptic soy agar containing chloramphenicol. For transduction, phage 11 was used to produce a phage lysate of strain RN4220 containing the recombinant shuttle plasmids. The phage lysate was then used to infect the parental strain RN6390 as described previously (8). Transductants were selected on chloramphenicol-containing agar.

Isolation of RNA and Northern analysis. Total cellular RNAs of S. aureus strains were obtained from bacterial cultures grown at 37°C with the FastPrep system (BIO101, Vista, Calif.) as previously described (6). Ten micrograms of RNA was electrophoresed through a 1.2% agarose-0.66 M formaldehyde gel in MOPS (morpholinepropanesulfonic acid) running buffer (20 mM MOPS, 10 mM sodium acetate, 2 mM EDTA, pH 7.0). RNA was transferred onto a Hybond N+ membrane (Amersham) under mild alkaline conditions by using a Turboblotter system (Schleicher and Scheull, Keene, N.H.), fixed to the membrane by baking (80°C for 1 h), hybridized under aqueous conditions at 65°C with -³²P-labeled gel-purified DNA fragments, washed, and autoradiographed (8). Band intensities were quantitated by densitometric scanning with SigmaGel software (Jandel Scientific, San Rafael, Calif.); these values are presented as integrated area units.

Immunoblot analysis of GFP expression. Cell extracts were prepared from S. aureus strains. In brief, the bacterial pellet was resuspended in 1 ml of TEG buffer (25 mM Tris, 5 mM EGTA, pH 8), and cell extracts were prepared from lysostaphin-treated cells as described by Mahmood and Khan (19). Proteins from cell extracts of S. aureus strains were resolved and transferred onto nitrocellulose membranes as described previously (26). Rabbit anti-GFP polyclonal antibodies (Clontech), diluted 1:5,000, were allowed to incubate with the membrane for 1 h, followed by an additional hour of incubation with a 1:10,000 dilution of goat anti-rabbit-alkaline phosphatase conjugate (Jackson ImmunoResearch, West Grove, Pa.). Immunoreactive bands were detected as previously described (3). SeeBlue prestained protein standards (Novex, San Diego, Calif.) were used for molecular weight estimation.

Spectrofluorimetry. Spectrofluorimetry was conducted to quantitate in vitro expression of GFP_UV by various sar promoter-gfp_UV constructs. Protein concentrations in cell extracts from S. aureus strains were assayed by the Bradford dye-binding procedure with bovine serum albumin as the standard (5). The emission and excitation spectra of diluted cell lysates (6.66 µg/ml) were analyzed in an Aminco-Bowman fluorescent spectrophotometer (SLM Instruments, Rochester, N.Y.).

Transcriptional fusion assay with the xylE reporter gene. We have previously constructed transcriptional fusions of P1, P3, and P2 promoters of sar with xylE as a reporter gene in shuttle vector pLC4 (20). To determine the XylE (catechol 2,3-dioxygenase) activities of these sar promoter fragments, 10 to 50 ml of cell culture grown overnight was pelleted. The cells were washed twice with 1 ml of ice-cold 20 mM potassium phosphate buffer (pH 7.2), resuspended in 500 µl of 100 mM potassium phosphate buffer (pH 8.0) containing 10% acetone and 25 µg of lysostaphin per ml, incubated for 15 min at 37°C, and then iced for 5 min. Extracts were centrifuged at 20,000 × g for 50 min at 4°C to pellet cell debris. The XylE assays were determined spectrophotometrically at 30°C in a total volume of 3 ml of 100 mM potassium phosphate buffer (pH 8.0) containing 100 µl of cell extract and 0.2 mM catechol as described previously (28), with readings of optical density at 375 nm taken at 2, 5, 15, and 25 min. One milliunit is equivalent to the formation of 1.0 nmol of 2-hydroxymuconic semialdehyde per min at 30°C. Specific activity is defined as milliunits per milligram of cellular protein (28).

Rabbit model of endocarditis. To assess the activation of these individual sar promoters in vivo, S. aureus RN6390 strains containing various sar promoter-gfp_UV constructs were tested in the rabbit endocarditis model. Briefly, overnight bacterial plate cultures were harvested by centrifugation (2,000 × g for 10 min), washed twice in sterile normal saline, and resuspended to an optical density at 620 nm of 1.6 (10⁹ CFU/ml). Dilutions of the bacterial suspension in phosphate-buffered saline were prepared, and the final infecting inoculum was confirmed by plate counting. For in vivo gene expression, RN6390-derived strains containing recombinant pALC1434 with individual sar promoters (P1, P3, and P2) were used to induce endocarditis on the aortic valve of New Zealand White rabbits as described previously (7). In brief, rabbits were anesthetized by intramuscular injections of ketamine chloride at 35 mg/kg and xylazine at 1.5 mg/kg. Forty-eight hours after the introduction of a transaortic valve polyethylene catheter (inner diameter, 0.86 mm) (to induce sterile thrombotic lesions on the valve), animals (three each) were challenged intravenously with 3 × 10⁸ CFU. This inoculum was chosen, based on pilot studies, to ensure adequate numbers of bacteria within vegetations for routine and fluorescence microscopic visualization. Catheters remained in place until animals were sacrificed by lethal injection of sodium pentobarbital (100 mg/kg) at 24 h after bacterial challenge. Aortic valve vegetations were removed and placed into O.C.T.-Tek holding solution for subsequent tissue processing for routine and fluorescence microscopy. Several vegetations were also removed, homogenized, and quantitatively cultured to ensure induction of infection.

Fluorescence microscopy. To evaluate the fluorescence status of various sar-gfp_UV constructs prior to animal challenge, direct fluorescence of bacterial colonies on overnight agar plates was performed with hand-held long range UV light (365 nm). Additionally, an aliquot of each of the challenge inocula was applied to a slide, air dried, and examined by epifluorescence microscopy.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

In vitro expression of GFP with transcriptional fusions containing sar promoters. To evaluate the feasibility of GFP as a reporter system in vitro and to assess in vivo activation of individual sar promoters, we constructed a pSK236-derived shuttle vector containing the gfp_UV gene preceded by an S. aureus ribosomal binding site (of sarA) and a polylinker region. This recombinant vector, pALC1434, was then used to test activation of the sar promoters in vitro and in vivo. Using the multiple cloning site within the pUC19 portion of pALC1434, we first cloned individual sar promoters into the vector in E. coli. These recombinant shuttle vectors were then introduced into S. aureus RN6390.

View larger version (35K): [in this window] [in a new window]   FIG. 1.   Excitation of RN6390-derived clones containing sar promoters with a long-range UV light (maximal excitation at 365 nm). Strains ALC1435, ALC1436, and ALC1437 contain pALC1434 with the sar P1, P3, and P2 promoters, respectively. The control strain ALC1440 contains only the vector pALC1434.

View larger version (30K): [in this window] [in a new window]   FIG. 2.   Northern blot of gfp transcripts of pALC1434 driven by the sar P1, P3, and P2 promoters. The control vector with no promoter upstream yielded no transcripts (data not shown). stat, stationary.

View larger version (43K): [in this window] [in a new window]   FIG. 3.   Immunoblot of cell lysates of RN6390 clones with GFPUV driven by sar promoters. The anti-GFP antibody was used at a 1:5,000 dilution.

Expression of GFP in vivo in the rabbit endocarditis model. To explore whether the observed in vitro patterns of sar promoter activation were similar in vivo, we examined vegetation tissues of animals with experimental infective endocarditis induced by each of the sar-gfp_UV constructs by routine and fluorescence microscopy. In animals challenged with the three sar-gfp_UV constructs, there were no significant differences in bacterial vegetation densities achieved at 24 h after infection (data not shown). The sar P1 promoter, as seen with fluorescence microscopy in vitro (Fig. 1), was activated in infected vegetations both on the lesion surface (Fig. 4B) and deep within the vegetations (Fig. 4A). The location of these fluorescent bacteria in the vegetations was confirmed by Giemsa staining (Fig. 4C). To avoid the possibility that the fluorescence activity associated with the P1 promoter was a carryover from the bacterial culture (in which P1 was activated), we conducted a parallel study in which harvested bacteria of the sar P1 construct in vitro were resuspended in RPMI 1640 at 4°C for 6 days to completely bleach the fluorescence activity. At the end of this incubation period, epifluorescence microscopy revealed that >90% of the bacteria were nonfluorescent. These bleached bacteria, upon injection into the rabbits, were able to fluoresce within 24-h-old vegetations in a manner similar to that of the nonbleached control, thus suggesting that the fluorescence activity was directly attributed to in vivo activation of P1. In contrast to the case for the sar P1 promoter, fluorescence activities were not detected in vegetations containing bacteria with the P3 promoter fusion even though numerous colonies could easily be demonstrated with Giemsa stain throughout the vegetations (data not shown). Despite its quiescent activity in vitro (Fig. 1), the sar P2 promoter was highly activated in vivo as revealed by fluorescence microscopy. In particular, the P2 promoter was active in bacterial cells located at the periphery of the vegetative lesion (Fig. 4E) while remaining inactive in cells in the center of the vegetation (Fig. 4D). These results supported the notion that individual sar promoters can be differentially activated in vivo within the animal host. More importantly, the pattern of activation in vivo can be different from that in vitro.

View larger version (63K): [in this window] [in a new window]   FIG. 4.   (A) Twenty-four-hour-old vegetations infected with ALC1435 (RN6390 containing the sar P1-gfpUV construct), in which the infecting strain was preincubated in RPMI 1640 medium for 6 days at 4°C to turn off fluorescence (the half-life of GFPUV is 24 h [10b]). These bacteria were then injected into catheterized rabbits to induce endocarditis. As seen, bacterial colonies deep within the vegetation now exhibit uniformly bright green fluorescence. Magnification, ~×300. The area deep within the vegetation is farther from the area of blood flow. (B) Same experiment as in panel A, except that bacterial colonies are on the surfaces of the vegetations. Note brightly fluorescent bacterial colonies (arrowhead). Magnification, ~×300. This area is proximal to the region of blood flow. (C) Same preparation as in panel B stained with Giemsa stain to show bacterial colonies on the vegetation surface. Magnification, ~×225. (D) Twenty-four-hour-old vegetations infected with ALC1437 (sar P2-gfpUV construct). The arrowhead denotes a colony of nonfluorescent bacteria upon excitation with a UV light source. Note that the yellow-stained background represents infected tissue, not green fluorescent bacteria. Magnification, ~×200. (E) Same experiment as in panel D, except that bacterial colonies are found on the vegetation surface. Note fluorescent P2-gfpUV-expressing cells (arrowhead) attached to the vegetation surface. Magnification, ~×300. (F) Same preparation as in panel E stained with Giemsa stain to show bacterial colonies on the vegetation surface. Magnification, ~×225.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Temporal expression of many of the extracellular and cell wall virulence determinants in S. aureus has been shown to be under the control of the global regulatory locus sar. The sar locus is composed of three overlapping transcripts in a parallel array, with the sarA open reading frame (ORF) present in all three transcripts (2). As documented previously (20), transcriptional assays with an xylE reporter gene in vitro revealed that the P1 promoter is at least 30-fold more active than P2 and P3. Activation of these promoters is also growth cycle dependent (2), with the P1 and P2 promoters more active in mid- to late log phase and P3 maximally expressed in the postexponential phase (20).

Recent studies disclosed that the SarA protein level is critical to agr activation (10a). In particular, SarA protein levels are positively influenced by the smaller ORFs, designated ORF3 and ORF4, encoded within the largest sarB transcript initiated from the sar P2 promoter. Activation of the P2 promoter, in conjunction with the P1 promoter, will thus likely lead to a SarA expression level higher than that from the P1 promoter alone (20). Recognizing that SarA binds directly to the agr promoter region (10), we also found that increased SarA levels correlated with a higher degree of agr activation in vitro (10a).

As the sar locus controls an assortment of virulence determinants in S. aureus (7) that might be influenced by microenvironmental factors in vivo which are absent in vitro (e.g., host proteins and phagocytic cells), we wanted to evaluate if the pattern of distinct sar promoter activations in vitro paralleled those in vivo. For this purpose, we constructed a shuttle vector containing a promoterless gfp_UV gene preceded by a polylinker site. By inserting a ribosomal binding site of S. aureus upstream of the gfp_UV gene, we attempted to optimize the translation of the GFP_UV reporter protein of pALC1434 upon activation by an appropriate promoter in an S. aureus host. Transcription and immunoblot analyses of the gfp gene product (Fig. 2 and 3) driven by different sar promoters paralleled sar activation data obtained by utilizing a XylE reporter fusion assay (20). Accordingly, P1 was the most active promoter in both assays, with significantly higher activity than either P2 or P3. The growth phase dependency of the sar promoters in liquid culture was also evident in the sar promoter-gfp_UV fusion. More specifically, the gfp_UV transcript initiated from the P1 and P2 promoters was maximally transcribed during the mid- to late log phase and activity tapered during the stationary phase, whereas the P3-initiated gfp_UV transcript began in the late log phase, with activity peaking during the stationary phase (Fig. 2). These data are consistent with those obtained with the XylE reporter fusion in vitro (20). Although both GFP and the XylE gene products are relatively stable, the GFP reporter system offers unique advantages because quantitative fluorescence of GFP does not require developing substrate, nor does it necessitate cell lysis. Taken together, our data indicate that the shuttle vector pALC1434 containing a promoterless gfp_UV gene is highly useful for assaying in vivo promoter activity of S. aureus.

To assess the activation of sar promoters in vivo, we cloned various sar-gfp_UV fusions into S. aureus RN6390, a prototypic isolate which we have previously used in virulence studies in the rabbit endocarditis model (7). This animal model has several advantages for studying in vivo gene activation, including (i) high achievable bacterial densities within the vegetation for routine and fluorescence microscopy; (ii) wide bacterial distribution throughout the vegetations, thus allowing assessment of specific gene activation in distinct anatomic regions within the same lesion; and (iii) the presence of host proteins (e.g., fibrinogen and fibronectin) and host cells (e.g., platelets) within the vegetations. By harvesting infected cardiac vegetations 24 h after intravenous injection, the activity from individual sar promoters can be detected directly in situ by UV fluorescence in infected tissue. Predictably, the P1 promoter was active both in vitro and in vivo. By bleaching the fluorescence activity of the P1 promoter construct prior to intravenous challenge, we were able to show that the fluorescence activity associated with the P1 promoter within the vegetation was not a carryover from in vitro promoter activation in culture. In contrast to the P1 promoter, the P3 promoter was not activated under either in vitro or in vivo conditions. However, since the sar P3 promoter is SigB dependent (13, 20) and is activated late in the growth cycle in vitro, we cannot rule out the possibility that the P3 promoter may be activated much later than P1 as the infected vegetation matures (e.g., at 4 to 6 days postinfection). Most remarkably, the P2 promoter, which was not activated under in vitro conditions, became highly active in vivo (Fig. 4). The P2 promoter was active in the periphery of the vegetative lesion, where the bacteria are likely to be active metabolically (14), but not in the center of the lesion, where the organisms are more metabolically quiescent (14). Presumably, the bacteria at the periphery of the vegetations are more rapidly dividing and hence require additional nutrients. One potential means to acquire more nutrients is to lyse host cells with hemolytic enzymes (22), thus requiring activation of specific genes involved in such a cytolytic pathway (e.g., -hemolysin gene activation via sar or agr) (15).

We have recently shown that the elaboration of normal levels of -hemolysin (hla gene) is essential for S. aureus persistence within vegetations of animals with infective endocarditis (1). One plausible mechanism that the organism can deploy to maximize activation of hla is to optimize SarA expression via simultaneous activation of P1 and P2. In a recent study (20), we demonstrated by transcriptional analysis with an XylE reporter system that a combined P2-P3-P1 promoter was more potent in its activity than the P1 promoter alone. As the P3 promoter has a down-regulatory effect on the P1 promoter activity (i.e., the P3-P1 promoter is weaker than P1) (20), it is likely that combined P2 and P1 promoter activities of the sar locus would result in higher transcriptional activity. As the sarA gene is the major ORF within the sar locus, it is reasonable to presume that a higher level of SarA protein expression would ensue. An increased amount of the SarA protein, by virtue of its binding to the agr promoter, would in turn activate RNAII and RNAIII transcription to a higher level than that from the P1 promoter alone, thereby leading to increased -hemolysin production (8). Accordingly, the promoter analysis in vitro, combined with our in vivo finding of simultaneous activation of sar P1 and P2 promoters at the periphery (but not at the center) of the lesion, implied that such selective gene activations may provide the organism with a distinct survival advantage (perhaps via enhanced but tightly controlled -hemolysin production). To verify this scenario in vivo, experiments are planned to examine activation of hla promoter-gfp_UV fusions in different areas of the vegetations in the rabbit endocarditis model.

Utilizing the sar promoter-gfp_UV reporter fusion, our studies here clearly demonstrate the selective and differential gene activation of S. aureus in vivo in a relevant animal model system. More importantly, gene expression in vivo, as revealed by the sar promoter systems, is likely to differ from that in vitro. By delineating promoter activation of specific virulence genes at particular anatomic sites within infectious lesions (e.g., peripheries versus centers of lesions in different target sites, such as kidneys and spleens) in a relevant animal system (e.g., rabbit infective endocarditis), we will have a unique opportunity to scrutinize serially over time the complex pathogenic process of S. aureus infections with respect to specific gene activation.