Regulation of Staphylococcus aureus Pathogenesis via Target of RNAIII-activating Protein (TRAP)*

Staphylococcus aureus can cause disease through the production of toxins. Toxin production is autoinduced by the protein RNAIII-activating protein (RAP) and by the autoinducing peptide (AIP), and is inhibited by RNAIII-inhibiting peptide (RIP) and by inhibitory AIPs. RAP has been shown to be a useful vaccine target site, and RIP and inhibitory AIPs as therapeutic molecules to prevent and suppress S. aureus infections. Development of therapeutic strategies based on these molecules has been hindered by a lack of knowledge of the molecular mechanisms by which they activate or inhibit virulence. Here, we show that RAP specifically induces the phosphorylation of a novel 21-kDa protein, whereas RIP inhibits its phosphorylation. This protein was termed target of RAP (TRAP). The synthesis of the virulence regulatory molecule, RNAIII, is not activated by RAP in the trap mutant strain, suggesting that RAP activates RNAIII synthesis via TRAP. Phosphoamino acid analysis shows that TRAP is histidine-phosphorylated, suggesting that TRAP may be a these

In culture, the bacteria produce toxic exomolecules only when in higher densities, at the post-exponential phase of growth. In the early exponential phase, when in lower densities, the bacteria express surface molecules, such as fibronectin-binding proteins and fibrinogen-binding protein, that allow the bacteria to adhere to and colonize host cells. The ability of the bacteria to switch between expression of surface adhesion molecules and toxin exomolecules (1) is regulated primarily by an RNA molecule termed RNAIII (6 -8). It is hypothesized that RNAIII enables the bacteria to adhere to host cells when in low numbers but to disengage and spread when too crowded, thus allowing dissemination and establishment of the infection.
RNAIII is encoded by the agr locus (6) and regulates at least 15 genes coding for potential virulence factors. agr mutants are nonpathogenic and show a decreased synthesis of extracellular toxins and enzymes, such as ␣-, ␤-, and ␦-hemolysin, leucocidin, lipase, hyaluronate lyase, and proteases, and at the same time an increased synthesis of adhesion molecules, coagulase, and protein A (2,9). The agr locus contains two divergent transcription units, RNAII and RNAIII, driven by the promoters P2 and P3, both of which are active only from the mid-exponential phase of growth (9). RNAII contains four open reading frames: agrA, agrB, agrC, and agrD. The agrA and agrC genes encode a classical two-component signal transduction pathway composed of the AgrC signal receptor and the AgrA response regulator. The agrD gene product is a propeptide that is processed and secreted through AgrB, which is an integral membrane protein. The resultant mature autoinducing peptide (AIP) 1 (10) is the ligand that binds to and activates the phosphorylation of AgrC (11), which in turn is thought to phosphorylate AgrA, leading to up-regulation of RNAIII synthesis (12).
The synthesis of RNAIII is regulated by a quorum sensing mechanism (13). Molecules produced and secreted by the bacteria (autoinducers) accumulate, and when they reach a threshold concentration, RNAIII is synthesized. The autoinducers of RNAIII that have been described to date are the RNAIII-activating protein (RAP) (14 -16) and the agr-encoded AIPs (10,17,18). RAP is a ϳ38-kDa protein containing the NH 2 -terminal sequence IKKYKPITN (16). The AIPs are octapeptides encoded by the agr, are processed from AgrD, and activate RNAIII by inducing the phosphorylation of their receptor AgrC. Interestingly, AIPs produced by some S. aureus strains inhibit the expression of agr in other strains, and the amino acid sequences of peptide and receptor (AgrC) are markedly different between such strains, suggesting a hypervariability-generating mechanism (18). Biochemical analysis of AIPs has suggested that they contain an unusual thiol esterlinked cyclic structure, which is absolutely necessary for full biological activity (10).
RNAIII synthesis can be inhibited by antibodies directed against RAP. Mice vaccinated with RAP were shown to be protected from a S. aureus infection. The protection level correlated with the titer of anti-RAP antibodies, suggesting that RAP is a promising vaccine candidate (15). RNAIII synthesis can be inhibited by AIPs of nonself (10,18) and by RNAIIIinhibiting peptide (RIP) (14 -16, 20). The RIP is produced by coagulase negative staphylococcus (suggested to be Staphylococcus warnerii or Staphylococcus xylosus) (16,19) and has the sequence YSPXTNF, where X can be a cysteine, a tryptophan, or a modified amino acid. Both native RIP and a synthetic analogue YSPWTNF are extremely effective in inhibiting RNAIII synthesis in vitro and in suppressing S. aureus infections in vivo (15,20). RIP (native or synthetic) has been shown to prevent S. aureus SD cellulitis in mice (15). Synthetic RIP has been shown to prevent keratitis (tested in rabbits against S. aureus 8325-4), osteomyelitis (tested in rabbits against S. aureus MS), mastitis (tested in cows against S. aureus Newbould 305, AE-1, and environmental infections), and septic arthritis (tested in mice against S. aureus LS-1) (20). These findings strongly evidence the potential value of RIP as a therapeutic agent. The therapeutic potential of inhibiting RNAIII synthesis was confirmed by Mayville et al. (10), who demonstrated that peptides (AIPs) that inhibit RNAIII synthesis in vitro do in fact inhibit S. aureus infections in vivo.
Although the therapeutic potential of RAP, RIP, and inhibitory AIPs is not in dispute, many questions remain unanswered and hinder the development of vaccines and therapeutics. These include understanding the detailed mechanism by which RAP, RIP, and AIPs regulate RNAIII synthesis; the precise mutual interactions between RAP, RIP and AIPs; and the signal transduction pathway that leads to RNAIII synthesis and concludes in virulence.
Because of the sequence similarity between the NH 2 -terminal sequence of RAP and RIP (YKPITN as compared with YSPXTN), and because RIP has been shown to compete with RAP on the activation of RNAIII synthesis, we hypothesized that RAP and RIP may bind to the same receptor, one as an agonist (RAP) and the other as an antagonist (RIP). To test this hypothesis, RIP derivatives were synthesized according to the putative NH 2 -terminal sequence of RAP and tested for their ability to inhibit RNAIII synthesis in vitro and for their ability to prevent S. aureus cellulitis in vivo. The results of these experiments indicate that the peptides most successful in inhibiting RNAIII synthesis and cellulitis were those that most resembled the NH 2 terminus of RAP and contained the sequence YKPITN (16). These results further suggest that RAP and RIP may in fact act as an agonist (RAP) and an antagonist (RIP) to the same receptor.
Here, we show that RAP activates and RIP inhibits the phosphorylation of a 21-kDa protein. We termed this protein target of RAP (TRAP). Amino acid sequence analysis of TRAP indicates that it is a 167-amino acid polypeptide that is unique to S. aureus. RAP does not activate RNAIII synthesis in a trapmutant, suggesting that RAP activates RNAIII synthesis via TRAP. We also show here that the phosphorylation of TRAP is inhibited by AIP of self, which uses the agr signal transduction system to activate RNAIII synthesis. Taken together, our results indicate that the trap and the agr signal transduction systems interact with one another, resulting in up-regulation of RNAIII synthesis and in a coordinated production of virulence factors.

EXPERIMENTAL PROCEDURES
Bacterial Strains-The bacterial strains used were as follows: wild type S. aureus strain RN6390B (ATCC 55620); agr-null S. aureus mutant strain RN6911 (6); RIP-producing, coagulase negative Staphylococcus strain RN833 (ATCC 55619) (14 -16, 19); S. aureus strain RN4220, a mutant of the wild type S. aureus strain 8325-4 that is capable of accepting foreign DNA (21); and S. aureus strain OU20, containing a disrupted trap, grown at 42°C in the presence of 10 g/ml erythromycin. Unless mentioned otherwise, all bacteria were grown in CY broth supplemented with ␤-glycerophosphate (21) at 37°C with shaking from early exponential phase of growth.
Preparation of RAP-To purify RAP, RN6390B cells were grown to the post-exponential phase of growth. Growth culture was centrifuged at 6000 ϫ g for 10 min at 4°C. The supernatant was collected and filtered through a 0.22 m filter to remove residual cells. The supernatant was lyophilized (FlexiDry MP lyophilizer) and resuspended in water to one-tenth of the original volume (total, 10ϫ). 15 ml of 10ϫ supernatant was applied to a 10-kDa cutoff membrane (Centriprep 10 (Amicon)). This enabled us both to concentrate the material further and to remove material that was smaller than 10 kDa. 1 ml of concentrated material greater than 10 kDa was washed twice in PBS by resuspending it in 15 ml of PBS and reconcentrating it on the Centriprep 10, and the material greater than 10 kDa was collected (Ͼ10); this material is usually at a concentration of 25ϫ-40ϫ. This material contained no AIP and was used for RAP/RIP or RAP/AIP competition assays (see below), as well as for the further purification of RAP. To further purify RAP, 600 l of Ͼ10 were fractionated on a gel filtration column (Superose 12, Amersham Pharmacia Biotech) in 1 mM phosphate buffered saline, pH 7.2 (0.1ϫ PBS), at a flow rate of 0.5 ml/min, and 1-ml fractions were collected. Fractions were concentrated to one-tenth of their original volume by lyophilization and tested for RNAIII by Northern blotting, as described below. Active gel filtration fraction (1 ml) was collected, and RAP was further purified by anion exchange chromatography (HPLC SynchroPak Q300, Keystone Scientific, Inc.) in water, pH 7.2. Bound material was eluted by a salt gradient of 0 -1 M NaCl in water in 1-ml fractions. 38-kDa RAP eluted at 0.75 M NaCl. Active fraction was lyophilized and resuspended in 100 l of water (10ϫ active fraction, or RAP). To test for activity, 50 l of 10ϫ active fraction (RAP) was applied to 450 l of early exponential cells (containing about 2 ϫ 10 9 cells) as described below.
Preparation of AIP and RIP-To partially purify AIP from RN6390B post-exponential culture supernatants or native RIP from RN833 postexponential supernatants, cells were grown to the post-exponential phase of growth. Growth culture was centrifuged at 6000 ϫ g for 10 min at 4°C. The supernatant was collected and filtered through a 0.22 m filter to remove residual cells. The supernatant was lyophilized (Flexi-Dry MP lyophilizer) and resuspended in water to one-tenth of the original volume. 15 ml of 10ϫ concentrated supernatant was applied to a 3-kDa cutoff membrane (Centriprep 10 (Amicon)), and material smaller than 3 kDa (flow-through) was collected and used to test for activity. To test for activity, 50 l of the flow-through was applied to 450 l of early exponential cells (containing about 2 ϫ 10 9 cells) as described below.
In Vivo Phosphorylation Assays-S. aureus RN6390B cells (4 ml) were grown in CY/GP from log phase of growth (A 650 of 0.03) until early exponential phase of growth of A 650 of about 0.2 (equivalent to about 1 ϫ 10 9 cells/ml). Cells were collected by centrifugation at 3000 ϫ g for 30 min or at 12,000 ϫ g for 2 min. Supernatants were discarded, and the cell pellet was resuspended in 0.9 ml of phosphate-free buffer (PFB) (20 mM KCl, 80 mM NaCl, 20 mM NH 4 Cl, 0.14 mM Na 2 SO 4 , 100 mM Tris, pH 7.4, 2.5 mM MgCl 2 , 0.1 mM CaCl 2 , 2 M FeCl 2 , 0.4% glucose, 9 g/ml thiamine, 0.8 mM potassium phosphate buffer, pH 7.4, 0.25 mM Larginine, 0.21 mM L-histidine, 0.62 mM L-lysine, 0.13 mM L-glutamic acid, 0.056 mM glycine, 0.32 mM L-alanine, 0.46 mM L-valine, 0.36 mM L-isoleucine, 0.82 mM L-proline, 0.016 mM L-phenylalanine, 0.50 mM L-serine, 0.34 mM L-threonine, 0.016 mM L-tyrosine, L-0.27 mM cystein, 0.21 mM L-methionine, 0.13 mM L-asparagine, 0.029 mM nicotinic acid, 0.13 mM L-glutamine) and 28 Ci of radiolabeled orthophosphate ( 32 P) (ICN Biochemicals). Cells were grown with shaking for 40 min at 37°C in the presence of one of the following (prepared as described above): 50 l of total post-exponential supernatant (total 10ϫ, containing 80% AIP and 20% RAP (16)), 50 l of AIP, 50 l of RAP, and 50 l of native RIP, with synthetic RIP (10 g RIP/4 ϫ 10 6 cells) or with 50 l control buffer. For growth phase experiments, cells in PFB and 32 P were grown for the times indicated. Cells were collected by centrifugation for 2 min at 12,000 ϫ g, supernatants were removed, and cells pellets were washed once in PBS to remove unincorporated 32 P. Radiolabeled cells were resuspended in 20 l of 50 g/ml lysostaphin in 10 mM Tris, pH 8.0, 1 mM EDTA for 10 min at room temperature, Laemmli sample buffer was added (without boiling), and the sample (total cell homogenate) was separated by 15% SDS-PAGE. The gel was autoradiographed, and the density of the bands was determined. The gel was then stained in Coomassie to ensure that equal amounts of protein were in fact loaded on the gel.
For RAP/RIP or RAP/AIP competition experiments, 450 l of cells in PFB and 32 P prepared as described above were incubated with 50 l of RAP (40ϫ Ͼ10), which was partially purified from post-exponential supernatants (see above), together with 50, 25, and 12.5 l of native RIP or AIP (prepared as described above), and the volume was adjusted to 550 l with CY.
Activation of RNAIII Synthesis-Early exponential 6390B (450 l) was grown with 50 l of the sample in question (prepared as described above: RAP, AIP, total 10ϫ, native RIP, CY, or PBS) for 30 min at 37°C. Cells were collected by centrifugation (2 min at 12,000 ϫ g), and RNA was purified as described below. If cells were incubated with synthetic RIP, the amount of peptide used was 10 g/4 ϫ 10 6 cells.
Detection of RNAIII and TRAP Transcript: RNA Purification and Northern Blotting-Equal number of cells were resuspended in 50 l of lysostaphin in TES buffer (200 g/ml lysostaphin (Sigma) in 100 mM Tris, pH 7.2, 1 mM EDTA, 20% sucrose) and incubated for 10 min at room temperature. 50 l of 2% SDS containing proteinase K (100 g/ml) was added and vigorously vortexed for 1 min followed by 10 min of incubation at room temperature. The sample was frozen and thawed twice. 15 l of RNA sample was mixed with 11% deionized glyoxal, 16 mM phosphate buffer, pH 7.0, and 55% Me 2 SO (final concentrations) and incubated for 1 h at 65°C. RNA loading buffer (Ambion) was added, and a sample (of about 5 ϫ 10 8 cells) was applied to a 1% agarose gel in 10 mM phosphate buffer, pH 7.0, supplemented with 5 mM iodoacetic acid (Sigma). Gel was Northern blotted by dry transfer and membranestained in methylene blue to view RNA and ensure that the same amounts of RNA were transferred. The membrane was prehybridized using Rapid-Hyb (Amersham Pharmacia Biotech) followed by hybridization with PCR-radiolabeled RNAIII-specific DNA (6) or with PCRradiolabeled 3Ј trap (nt 400 -550). The gels were autoradiographed, and the intensity of the band determined using a quantitative analysis program (Molecular Analyst).
Phosphorylated TRAP (TRAP-P) Purification-A cell pellet of 500 ml of early exponential S. aureus cells (grown as described above) was resuspended in 30 ml of PFB, 560 Ci of 32 P, and 3 ml of RAP and grown for 1 h with shaking at 37°C. Cells were collected by centrifugation, washed in PBS, resuspended in 2.5 ml of water containing 100 g/ml lysostaphin (Sigma), and incubated on ice for 20 min. 7.5 ml of water was added, and cells were further incubated on ice for 20 min and disrupted by extensive sonication (Sonic Dismembrator, Fisher Scientific, microtip probe) (three times for 10 s each time on ice). Sonicated material was centrifuged (10 min at 12,000 ϫ g), and soluble material was collected. Soluble material was concentrated by lyophilization (Savant, Speedvac Plus SC110A) to 800 l. Material was fractionated on a gel filtration HPLC column (Bio-Sil SEC-125 300 ϫ 7.8 mm, Bio-Rad) in 1 mM phosphate-buffered saline, pH 8.0, at a flow rate of 1 ml/min. 1-ml fractions were collected and tested for radioactivity in a scintillation counter (Beckman, LS600 multipurpose scintillation counter). Samples containing high cpm in comparison with the rest of the fractions were selected, separated by 15% SDS-PAGE, and gel autoradiographed. The fraction containing the radiolabeled 21-kDa protein (phosphorylated TRAP) was used to further purify TRAP by anion exchange chromatography (HPLC SynchroPak Q300, Keystone Scientific, Inc.) in 0.1ϫ PBS, pH 7.5. Bound material was eluted by a salt gradient of 0 -1 M NaCl in 0.1ϫ PBS, pH 7.5. Fractions containing high cpm were separated by 15% SDS-PAGE and gel autoradiographed to determine the fraction containing TRAP. Phosphorylated TRAP eluted at 0.75 M NaCl. The 21-kDa radioactive band was cut and submitted for amino acid sequencing (see below).
Amino Acid Sequence Analysis-The anion exchange TRAP-containing fraction was applied to SDS-PAGE, and the gel was stained in Coomassie and autoradiographed. The protein band corresponding to phosphorylation was cut and NH 2 -terminally sequenced or subjected to tryptic digestion for acquiring internal sequences. Specifically, the gel was dried, rehydrated in 50 mM ammonium bicarbonate, pH 7.8, and incubated with 0.5 g of trypsin overnight at 37°C. Peptides were extracted by 70% acetonitrile/5% formic acid and fractionated on a C18 FIG. 1. RNAIII synthesis is activated by RAP and inhibited by RIP. Early exponential S. aureus cells were incubated for 40 min together with synthetic RIP (lane 1), with PBS as a control (lane 2), or with RAP (lane 3) as described under "Experimental Procedures" ("Activation of RNAIII Synthesis"). RNA was purified, equal amount of RNA were applied to the gel, and the gel was Northern blotted. RNAIII was detected using radiolabeled RNAIII-specific DNA as a probe. The membrane was autoradiographed, and the density of the bands was determined.  4 -6). After 60 min, cells were collected, the total cell homogenate was applied to SDS-PAGE, and the gel was autoradiographed.
HPLC (Vydac, 4.6 ϫ 25 cm) in 0.1% TFA. Peptides were eluted on a 115-min gradient of 0 -70% acetonitrile. Peptides were collected and amino acid sequenced commercially by Edman degradation chemistry (ABI 477 sequencer, Protein Structure Laboratory, UC Davis). Sequences were compared with the S. aureus Genome Sequencing Project data base, and the sequence of TRAP was determined. Primers corresponding to the 5Ј-and 3Ј-ends of the gene were constructed, trap was amplified by PCR, and the DNA sequence was confirmed.
Inactivation of trap-An internal 317-bp fragment (73-390) of the trap gene was amplified by PCR using the following primers: 1) 5Ј-CGCGCGGATCCCAACTATTCCAATTTTCAG-3Ј (containing the BamHI site), and 2) 5Ј-CGCGAAGCTTCTTAAAGTCTTCGTATG-3Ј (containing the HindIII site). The amplified PCR fragment was cloned into the BamHI/HindIII sites of the pAUL-A vector (kindly provided by S. Del-Cardayre), which is a shuttle vector between S. aureus and Escherichia coli. This plasmid contains an erythromycin resistance marker and carries a temperature-sensitive mutation at the S. aureus origin of replication, and therefore it is capable of replication in S. aureus cells only at the permissive temperature of 30°C. S. aureus strain RN4220 cells (a restriction-deficient derivative of strain 8325-4 that is therefore capable of accepting foreign DNA (21)) were transformed with the above construct by electroporation as described (22), and transformants were grown on NYE agar (23) in the presence of 10 g/ml erythromycin at the permissive temperature of 30°C overnight. Transformants were grown on NYE agar in the presence of 10 g/ml erythromycin at the restrictive temperature of 42°C overnight. Colonies were analyzed for integration of the plasmid into the chromosome at the trap site via a Campbell insertion process. The analysis was done by PCR, employing primers that are homologous to the plasmid region (universal reverse primer 5Ј-GTAAAACGACGGCCAGT-3Ј) and absent in the chromosome and a primer that is homologous to the pre-5Ј-end of the trap gene and is not present on the plasmid construct (5Ј-GTGG-TAATGACTAGTTTATCATCGT-3Ј (nucleotides -59 to -34). A DNA fragment of 500 bp was generated using the above primers, indicating the integration of the plasmid and disruption of trap. The S. aureus containing the disrupted trap gene was termed OU20. Of note is the fact that the trap gene was also inactivated in S. aureus 8325-4 and termed YG1. The phenotype of S. aureus, YG1, had a similar phenotype to that of OU20. 2 Phosphoamino Acid Analysis-Purified radiolabeled TRAP-P was applied to SDS-PAGE. Slices of acrylamide, containing labeled TRAP-P, were excised and submerged in 3 N KOH at 105°C for 5 h. The resulting hydrolysate was diluted 25-fold with water containing internal standards of phosphoserine and phosphotyrosine. Phosphoamino acids were separated by ion-exchange chromatography (24). O-Phthalaldehyde was added to the eluate, and the resulting fluorescence was detected on-line (24). Radioactivity was quantified by liquid scintillation counting. Phospholysine and phosphohistidine were synthesized as described previously (24). All other standards were purchased from Sigma. In this system, phosphoarginine and phospholysine elute before phosphoserine, phosphothreonine elutes close to phosphoserine, and phosphohistidine elutes between phosphoserine and phosphotyrosine (24).

RAP Activates and RIP Inhibits RNAIII Synthesis and the
Phosphorylation of a 21-kDa Protein-Early exponential wild type S. aureus cells were incubated for 40 min in the presence of RAP, synthetic RIP (Genemed Synthesis, Inc. CA), or PBS only as a control. Cells were collected, RNA purified, and Northern blotted, and membranes were incubated with radiolabeled RNAIII-specific DNA as a probe. As previously demonstrated (14,16) and as shown in Fig. 1, RAP activates and RIP inhibits RNAIII synthesis. The pathway by which the autoinducer RAP activates and the peptide RIP inhibits RNAIII synthesis was not known, but it was hypothesized that RAP and RIP interact with the same receptor, one as an agonist (RAP), the other as an antagonist (RIP). Therefore, it seemed reasonable to assume that, like other quorum sensing molecules, they would regulate a bacterial two component system by phospho-2 Y. Gov, unpublished data.  4. A and B, amino acid sequence of TRAP (A) and DNA sequence of trap (B). C, secondary structure prediction of TRAP generated by the PHD package (27). Cylinders and arrows denote ␣-helix and ␤-sheet, respectively. rylation (25). To identify the signal transduction pathway regulated by RAP and RIP, in vivo phosphorylation assays were performed. Early exponential wild type S. aureus were incubated in phosphate-free buffer supplemented with radiolabeled orthophosphate, together with RAP in PBS, with PBS, or with RIP (native or synthetic). After a 40-min incubation period, the cells were collected by centrifugation and treated with lysostaphin followed by the addition of sample buffer; without boiling, total cell homogenate was applied to both 7.5 and 15% SDS-PAGE, and the gel was stained in Coomassie or autoradiographed. As shown in Fig. 2, RAP activates and RIP inhibits the specific phosphorylation of a 21-kDa protein that we termed TRAP. Endogenous RAP is produced as the cells grow (14), probably contributing to the positive signal in the control PBS group (Fig. 2, lane 1). To determine whether RIP competes with RAP on TRAP phosphorylation, cells were incubated with RAP together with increasing amounts of native RIP and in vivo phosphorylation assays were carried out. As demonstrated in Fig. 3, the higher the amount of RIP present, the lower the amount of TRAP phosphorylation, suggesting that RIP competes with RAP on the phosphorylation of TRAP.
To determine whether RAP induces the synthesis of TRAP and not only its phosphorylation, the in vivo phosphorylation assays were carried out in the presence of 100 g/ml chloramphenicol, to inhibit potential translation processes. The results of these experiments (not shown) indicate that RAP activates TRAP phosphorylation also in the presence of chloramphenicol, suggesting that RAP activates TRAP phosphorylation and not synthesis.
Structure of TRAP-To purify TRAP, wild type early exponential S. aureus cells were in vivo phosphorylated, cells were disrupted by extensive sonication, and soluble material containing TRAP-P was fractionated on an HPLC gel filtration column. Positive fractions (determined by peak radioactivity and confirmed by separating a sample by SDS-PAGE) were applied to an HPLC anion exchange column, and bound material was eluted by a 0 -1 M NaCl gradient. The positive fraction containing TRAP-P eluted at ϳ0.75 M NaCl.
To determine the amino acid sequence of TRAP, purified TRAP was internally digested by trypsin, and peptide digests were amino acid sequenced. Acquired sequences were compared with the S. aureus data base, and the sequence of TRAP was determined to be a 167-amino acid polypeptide (Fig. 4) (GenBank TM accession number AF202641). The sequence of TRAP (Fig. 4, A and B) is unique to S. aureus and shows no significant sequence homology to known proteins or genes but for 5Ј-end of the Bacillus subtilis penicillin-binding protein gene (pbpF), with which it shares 28% identity (26). Twodimensional proton NMR spectra (not shown) reveal a folded protein made up of both ␣-helices and ␤-sheet secondary structure elements, in agreement with sequence and threading analysis (Fig. 4C) generated by the PHD package and threading analysis (27).
RAP Does Not Activate RNAIII Synthesis in the trap Mutant Strain-The trap gene was inactivated by gene disruption. An internal 317-bp fragment of the trap gene lacking regions of about 100 bp from the 5Ј-and 3Ј-ends of the gene was cloned into pAUL (Fig. 5A), a temperature-sensitive shuttle vector (kindly provided by S. Del Cardayre), and plasmid was used to transform S. aureus RN4220 cells. Transformants were analyzed for integration of the plasmid into the chromosome at the trap site via a Campbell insertion process. The analysis was  ϩ). RNA was purified, equal amounts of RNA (10 g) were applied to the gel, and the gel was Northern blotted. RNAIII was detected using radiolabeled RNAIII-specific DNA as a probe, and the membrane was autoradiographed. D, the production of RNAIII is reduced in the trapstrain. Early exponential (1 ϫ 10 9 cells/ml) S. aureus trapand trap ϩ cells were grown for 1-4 h. Equal number of cells were collected, RNA was extracted, the RNAIII and TRAP transcript was tested by Northern blotting, and the membrane was autoradiographed. Lanes 1-4, cells were grown for 1-4 h, respectively. done by PCR, employing primers that are homologous to the vector and to the 5Ј sequences of the pre-trap gene not present on the plasmid construct. A DNA fragment of about 500 bp was generated, indicating the integration of the plasmid and the disruption of trap, resulting in a trapmutant strain (S. aureus OU20). Sequence analysis of OU20 indicated the replacement of the 3Ј-end of the trap gene (from nt 390) with pAUL DNA. S. aureus OU20 was in vivo phosphorylated in the presence of RAP, total cell homogenate was applied to SDS-PAGE, and TRAP-P was detected by autoradiography. As shown in Fig. 5B, RAP activated the phosphorylation of TRAP in the trap ϩ strain (lane 1), but it did not activate phosphorylation in the trap -OU20 strain (lane 2), suggesting that in fact the trap gene was disrupted.
To test whether RAP activates the synthesis of RNAIII via TRAP, early exponential S. aureus trap ϩ and trapcells were incubated together with RAP; after 40 min, cells were collected, RNA was extracted, and RNAIII was analyzed by Northern blotting using radiolabeled RNAIII-specific DNA as a probe. As shown in Fig. 5C, although RNAIII synthesis was activated by RAP in the parent trap ϩ strain, it was not activated in the trapstrain, suggesting that the presence of an intact trap gene is necessary for RAP to activate RNAIII synthesis. To test whether RNAIII can be synthesized in the absence of TRAP during bacterial growth, trap ϩ and trapcells were grown for several hours from the early exponential phase of growth, and RNAIII and TRAP tested by Northern blotting. As shown in Fig. 5D, RNAIII synthesis was greatly reduced in the trapmutant strain but was not abolished. These results suggest that trap is important for the activation of RNAIII synthesis but that the synthesis of RNAIII can nevertheless be activated at a later stage in the absence of TRAP, possibly by alternate pathways, such as sar (28,29). As also shown in Fig. 5D, trap transcription is in fact absent in the trapstrain and is constitutive in the trap ϩ strain. Of note is the fact that the translation of TRAP also appears to be constitutive in the wild type trap ϩ strain (data not shown). The fact that trap is constitutively transcribed and translated while its phosphorylation is regulated further supports our results indicating that RAP regulates TRAP phosphorylation and not synthesis.
TRAP Is Histidine-phosphorylated-Two-component systems act through phosphorylation of the substrate domain of the sensor protein and subsequent transfer of the phosphate to an aspartate residue in the regulator protein. The initial phosphorylation is catalyzed by a protein histidine kinase domain in the sensor protein and results in an N-phosphorylated histidine residue, which is stable in alkaline conditions but not in acidic conditions (24).
To test whether TRAP may be histidine-phosphorylated, we tested the sensitivity of phosphorylated TRAP to acidic and basic conditions. Phosphorylated TRAP was incubated at pH ranging from 1 to 10 for 10 min at room temperature. The mixture was then applied to SDS-PAGE, and the gel was autoradiographed. As shown in Fig. 6A, phosphorylation of TRAP was found to be stable at pH greater than 8.0 but labile at lower pH values, consistent with a possible N phosphorylation of a histidine. Phosphoamino acid analysis indicates that in fact TRAP-P contains phosphohistidine. Purified radiolabeled TRAP-P was applied to SDS-PAGE. The gel band containing TRAP-P was subjected to alkaline hydrolysis followed by chromatography (24). The labeled phosphoamino acid eluted at the position of phosphohistidine, which is distinct from phosphoarginine, phospholysine, phosphothreonine, phosphoserine, or phosphotyrosine (Fig. 6B). Histidine phosphorylation indicates that TRAP may in fact be a sensor of RAP.
Regulation of TRAP Phosphorylation-RNAIII is produced only from the mid-exponential phase of growth, whereas TRAP is continuously transcribed (Fig. 5D). If RAP activates RNAIII via TRAP phosphorylation, it seemed reasonable to assume that TRAP phosphorylation and RNAIII synthesis should be coupled. To determine when TRAP is phosphorylated during bacterial growth, wild type S. aureus were grown from early to late logarithmic phase of growth in the presence of 32 P. Cells were collected at time intervals and assayed both for TRAP phosphorylation and for RNAIII synthesis. As shown in Fig. 7,  A and B, peak phosphorylation of TRAP was reached at the mid-exponential phase of growth. Peak phosphorylation of TRAP directly correlates with RNAIII synthesis, supporting FIG. 6. A, TRAP-P is stable in alkaline conditions. Phosphorylated TRAP was incubated for 10 min at room temperature at increasing pH values, applied to SDS-PAGE, and autoradiographed, and the density of the bands was determined. Results are presented as percentage of maximum phosphorylation observed (% of max). B, TRAP-P contains phosphohistidine. An alkaline hydrolysate of radiolabeled TRAP-P was analyzed by chromatography. The labeled phosphoamino acid eluted at the position of phosphohistidine (P-His), distinct from phosphoarginine, phospholysine, phosphothreonine, phosphoserine (P-Ser), or phosphotyrosine (P-Tyr). our hypothesis that RAP regulates RNAIII synthesis via TRAP phosphorylation.
As shown in Fig. 7, A and B, TRAP reached its peak phosphorylation by the mid-exponential phase of growth but was dephosphorylated by the late logarithmic phase of growth. The RNAIII gene, on the other hand, once activated, remained up-regulated throughout growth (Fig. 7B). The fact that TRAP was dephosphorylated by late log indicates that TRAP phosphorylation is necessary only for the induction of the RNAIII gene but not for its ongoing transcription.
AIP Activates RNAIII Synthesis but Inhibits TRAP Phosphorylation-RNAIII production has been shown to be autoinduced also by AIP, an octapeptide encoded by the agr itself. AIP activates RNAIII synthesis by inducing the phosphorylation of a two-component system, also encoded by the agr. Specifically, once the agr is activated in the mid-exponential phase of growth, an octapeptide is produced (processed from AgrD), inducing the phosphorylation of a 46-kDa protein, AgrC (11), which is hypothesized to phosphorylate AgrA (12), leading to up-regulation of RNAIII synthesis. To determine the interac- FIG. 7. A and B, TRAP reaches peak phosphorylation at the mid-exponential phase of growth. Wild type S. aureus was grown from early to late log phase of growth in PFB together with 32 P. Cells were collected at time intervals and tested both for RNAIII (by Northern blotting) and for TRAP phosphorylation (by SDS-PAGE). A, cells were resuspended in sample buffer and applied to 15% SDS-PAGE, and the gel was autoradiographed. Approximate molecular mass is indicated in kDa. B, TRAP phosphorylation versus RNAIII and cell number (cell #). Results are presented as percentage of maximum phosphorylation, RNAIII, or cell number observed (% of max). C, TRAP is phosphorylated in agr-null strains. Mutant agr-null S. aureus cells RN6911 were grown from early to late log phase of growth in PFB together with 32 P. Cells were collected, the total cell homogenate was applied to SDS-PAGE, and the gel was autoradiographed.
tion of the agr and the TRAP signal transduction systems, we tested whether TRAP can be phosphorylated also in an agr-null strain.
An agr-null S. aureus strain RN6911 (a mutant strain that contains a tetM gene instead of agr (6)) was grown from early to late logarithmic phase of growth in the presence of 32 P. Cells were collected at time intervals, applied to SDS-PAGE, and autoradiographed. As shown in Fig. 7C, TRAP is phosphorylated also in the agr-null strain. As in the wild type, peak phosphorylation was reached at the mid-exponential phase of growth. However, unlike in the wild type strain, TRAP was not dephosphorylated by the late logarithmic phase of growth, suggesting that the agr itself, once activated in the mid-exponential phase, produces a factor that down-regulates TRAP phosphorylation.
To determine whether AIP is the dephosphorylating, agrencoded factor, we incubated wild type cells with AIP, with RAP, or with culture supernatants containing both RAP and AIP in a RAP:AIP ration of 20:80 (16). TRAP phosphorylation was tested by in vivo phosphorylation assays, and RNAIII synthesis was tested by Northern blotting. As shown in Fig. 8, A and B, whereas RAP activates RNAIII synthesis and activates TRAP phosphorylation, AIP activates RNAIII synthesis but inhibits TRAP phosphorylation.
To test whether AIP competes with RAP on TRAP phosphorylation, cells were grown in the presence of RAP together with increasing amounts of AIP. As shown in Fig. 8C, the level of TRAP-P is dependent on the ratio between the two autoinducers. The more AIP in the culture supernatant as compared with RAP, the less TRAP phosphorylation occurred. These results can explain why TRAP is dephosphorylated from the midexponential phase of growth, which is when agr is activated and when AIP is produced.
RAP and AIP Activate RNAIII Synthesis via Different Signal Transduction Pathways-RAP does not activate RNAIII synthesis in a trapstrains (Fig. 5C), suggesting that RAP activates RNAIII via TRAP phosphorylation. To test whether AIP and RAP activate RNAIII synthesis by interacting with the same signal transduction pathway, we tested whether AIP can activate RNAIII synthesis in a trapstrain. trap ϩ and trapstrains were grown in the presence of RAP or AIP and tested for RNAIII synthesis. As shown in Fig. 8D, RAP did not activate RNAIII in the trapstrain, whereas AIP activated RNAIII synthesis both in the trap ϩ and trapstrain. These results suggest that AIP does not activate RNAIII via TRAP and that RAP and AIP activate RNAIII synthesis via different signal transduction pathways. Whereas AIP activates RNAIII synthesis via the agr system (phosphorylation of AgrC (11)), RAP activates RNAIII synthesis via the TRAP system (phosphorylation of TRAP). DISCUSSION Our work demonstrates that the autoinducer of virulence RAP activates and the inhibitor of virulence RIP inhibits the phosphorylation of a 21-kDa protein termed TRAP. Amino acid sequence analysis of TRAP indicates that the 167 amino acid polypeptide is unique to S. aureus. RAP does not activate RNAIII synthesis in a S. aureus strain containing a disrupted trap, suggesting that an intact trap gene is necessary for the activation of RNAIII synthesis by RAP. Phosphoamino acid analysis of TRAP-P indicates that TRAP is histidine-phosphorylated, indicating that TRAP may be a sensor of RAP. Secondary structure predictions suggest it, however, to be globular ␣-helix ␤-sheet protein that does not have a transmembrane region that is classical for a histidine kinase. TRAP may therefore either be part of a nonclassical histidine kinase (30) or be bound to a yet to be identified, membrane-associated molecule (Fig. 9).
TRAP reaches its peak phosphorylation by the mid-exponential phase of growth but is dephosphorylated by late logarithmic phase of growth. The gene for RNAIII, on the other hand, once activated, remains up-regulated throughout. The fact that TRAP is dephosphorylated by late log indicates that TRAP FIG. 8. The phosphorylation of TRAP is activated by RAP and inhibited by AIP. A, wild type early exponential S. aureus cells were incubated for 1 h in PFB, 32 P, together with RAP (40ϫ, Ͼ10) containing no AIP, with post-exponential total supernatant (total, containing 20% RAP and 80% AIP (16,19)), with AIP (containing no RAP), or with PBS as a control. Cells were collected and applied to SDS-PAGE, and the gel was autoradiographed. The autoradiogram was scanned, and the density of the bands was determined. B, in parallel, cells (in CY) were incubated in the presence of RAP, PBS, AIP, and total supernatant for 40 min, and cells were assayed for RNAIII by Northern blotting. The density of the bands was determined, and results are presented as percentage of maximum RNAIII observed (% of max). C, in vivo phosphorylation: RAP/AIP competition experiment. Wild type early exponential cells in PFB and 32 1 (lane 3). After 40 min, cells were collected and applied to SDS-PAGE, the gel was autoradiographed, and the density of the bands was determined. D, RNAIII synthesis is activated by AIP but not by RAP in a trap mutant strain. Early exponential (1 ϫ 10 9 cells/ml) S. aureus trapand trap ϩ cells were grown for 40 min in the presence of RAP, PBS, or AIP. Equal numbers of cells were collected, RNA was extracted, RNAIII was tested by Northern blotting, and the membrane was autoradiographed. phosphorylation is necessary only for the induction of the RNAIII gene but not for its ongoing transcription.
TRAP is phosphorylated also in the agr-null strain. As in the wild type, peak phosphorylation is reached at the mid-exponential phase of growth. However, unlike in the wild type strain, TRAP is not dephosphorylated by late log, suggesting that the agr itself, once activated in the mid-exponential phase, produces, or regulates the production of, a factor, which downregulates TRAP phosphorylation.
One of the factors produced by the agr is the octapeptide AIP that also activates RNAIII synthesis. However, the agr locus is temporally regulated, and therefore the AIP is only produced from the mid-exponential phase of growth (17). We show here that whereas RAP activates RNAIII synthesis as well as TRAP phosphorylation, the AIP activates RNAIII synthesis but inhibits TRAP phosphorylation. Furthermore, whereas RAP does not activate RNAIII synthesis in a trapstrain, AIP does upregulate RNAIII synthesis in a trapstrain, suggesting that RAP and AIP activate RNAIII synthesis via different signal transduction pathways. The fact that TRAP phosphorylation is down-regulated in the presence of the AIP may explain why TRAP is dephosphorylated at the mid-exponential phase of growth, coinciding with AIP production. The interplay between the two signal transduction pathways suggests that TRAP and the agr gene products could be part of a phosphorelay system. The phosphorelay is an extended and more complex version of the two component system, involving multiple activation signals processed by at least two histidine kinases (30,31). We propose that whereas RAP (signal A) activates the TRAP signal transduction (kinase A), the AIP (signal B) activates the agr system (kinase B), which probably leads to activation of a phosphatase and to the dephosphorylation of TRAP (Fig. 9).
It has been suggested that RNAIII synthesis is only regulated by the peptides encoded by the agr (AIPs) and that RIP is part of the AIP family of peptides (19). Although this may be the case, several experimental data do not support this hypothesis (16). 1) AIPs must contain a thiolactone structure to be active, whereas the RIPs are synthesized without a cysteine and a thiolactone structure and are active as linear peptides. 2) Both RIP and AIP of self inhibit TRAP phosphorylation, but whereas RIP inhibits RNAIII synthesis, AIP of self activates RNAIII synthesis. 3) AIP activates RNAIII synthesis in a trapstrain, whereas RAP does not, suggesting that AIP and RAP activate RNAIII synthesis via different signal transduction pathways. Furthermore, RIP prevents infections cause by various strains of S. aureus in different infection models, suggesting that unlike the AIPs, RIP is not strain-specific in its inhibitory activity.
In summary, we propose ( Fig. 9) that autoinduction of virulence occurs in a two-step process. As the colony multiplies, the autoinducer RAP accumulates and induces the phosphorylation of its target molecule TRAP, resulting in up-regulation of agr to produce RNAII. 2 Once agr is activated (in the midexponential phase of growth), AIP and its receptor AgrC are produced. AIP up-regulates the phosphorylation of its receptor, AgrC (11), leading to phosphorylation of AgrA, to up-regulation of RNAIII synthesis (9), and to down-regulation of TRAP phosphorylation. Production of RNAIII, in parallel with up-regulation of sar (32) and sae (33), causes the expression of toxic exomolecules and the suppression of adhesion molecules, resulting in dissemination and in disease. In the presence of anti-RAP antibodies, RIP, or inhibitory AIPs, RNAIII is not produced, and the pathogenic potential of the bacteria is greatly reduced (10,15,16,20). FIG. 9. Interaction of TRAP and AgrC. As the colony multiplies, the autoinducer RAP accumulates and induces the phosphorylation of its target molecule TRAP (A), resulting in the production of RNAII (B). Once agr is activated (in the mid-exponential phase of growth), AIP and its receptor AgrC are produced (9). AIP down-regulates TRAP phosphorylation and up-regulates the phosphorylation of its receptor, AgrC (B) (11), which is hypothesized to phosphorylate AgrA (C), which then acts as a transcription activator to activate P3 (12), leading to the production of RNAIII. Production of RNAIII, in parallel with up-regulation of sar and sae, causes the expression of toxic exomolecules and the suppression of adhesion molecules (C) (6,32,33), resulting in dissemination and in disease.