Identification of the Putative Staphylococcal AgrB Catalytic Residues Involving the Proteolytic Cleavage of AgrD to Generate Autoinducing Peptide*

The P2 operon of the staphylococcal accessory gene regulator (agr) encodes four genes (agrA, -B, -C, and -D) whose products compose a quorum sensing system: AgrA and AgrC resemble a two-component signal transduction system of which AgrC is a sensor kinase and AgrA is a response regulator; AgrD, a polypeptide that is integrated into the cytoplasmic membrane via an amphipathic α-helical motif in its N-terminal region, is the propeptide for an autoinducing peptide that is the ligand for AgrC; and AgrB is a novel membrane protein that involves in the processing of AgrD propeptide and possibly the secretion of the mature autoinducing peptide. In this study, we demonstrated that AgrB had endopeptidase activity, and identified 2 amino acid residues in AgrB (cysteine 84 and histidine 77) that might form a putative cysteine endopeptidase catalytic center in the proteolytic cleavage of AgrD at its C-terminal processing site. Computer analysis revealed that the cysteine and histidine residues were conserved among the potential AgrB homologous proteins, suggesting that the Agr quorum sensing system homologues might also exist in other Gram-positive bacteria.

The Agr quorum sensing system encoded by the staphylococcal accessory gene regulator (agr) 1 is one of the two-component signal transduction systems that involved in the regulation of virulence gene expression (1)(2)(3). The agr locus consists of two major transcripts: RNAII and RNAIII, that are transcribed divergently from the two agr promoters, P2 and P3, respectively (3). RNAII encodes four genes (agrA, -B, -C, and -D) whose products constitute a quorum sensing system: AgrC and AgrA are the sensor kinase and the response regulator of the Agr two-component signal transduction system, respectively (3,4); AgrD is the propeptide of the autoinducing peptide (AIP) that is secreted from the bacteria (5) and functions as the ligand for AgrC (6); and AgrB is a membrane protein that is involved in the processing of AgrD (7).
AgrC is a membrane protein with its N-terminal half integrated into the cytoplasmic membrane that contains the AIP binding site (6,8), and with its C-terminal half located in the cytoplasm that possesses histidine kinase activity (6). Among the identified AgrCs so far from various species of staphylococci, the N-terminal halves are divergent and the C-terminal halves are highly conserved (9,10). This reflects the fact that the AgrCs are activated only by their cognate AIPs but are inhibited by heterologous AIPs (9,(11)(12)(13)(14)(15). Based on the AIP cross-activation and cross-inhibition activities, four specificity groups of Staphylococcus aureus (9,14) and three groups of Staphylococcus epidemidis (10,16,17) have been identified. Upon the binding of AIP, AgrC is autophosphorylated (6). It has been proposed that the phosphoryl group of the phosphorylated AgrC is transferred to AgrA, and the phosphorylated AgrA then interacts with the P2 and P3 promoters to activate the transcription of both RNAII and RNAIII (3,4,6). RNAIII is the actual regulator that activates the expression of genes encoding secreted virulence factors and represses those encoding cell surface-associated proteins (1,18).
AgrD is a membrane protein anchored in the inner leaflet of the cytoplasmic membrane via an amphipathic ␣-helix formed by its N-terminal region (19). AgrD sequences from various staphylococcal species are remarkably divergent with only 4 identical amino acids (1). The mature AIPs isolated so far from a number of staphylococcal species are 7 to 9 amino acids in length, and all are thiolactone molecules containing a 5-amino acid ring linked by a thioester bond formed between the sulfhydryl group of a conserved cysteine residue and the carboxyl group of the C-terminal amino acid, except the Staphylococcus intermedius AIP, a lactone molecule that contains a ester bond formed between the hydroxyl group of a serine residue (in place of the cysteine residue that is absolutely conserved among other AIPs) and the carboxyl group of the C-terminal residue (11,12,15,20,21). The AIP sequence is in the middle of the AgrD sequence that is preceded by the N-terminal amphipathic ␣-helix and followed by a highly hydrophilic C-terminal region. The processing of AgrD to generate mature AIP involves the proteolytic cleavages at two processing sites, the thioester (or ester) bond formation, and the secretion of the mature AIP.
AgrB is a membrane protein with six transmembrane segments including four transmembrane ␣-helices and two highly hydrophilic regions (22). Like AgrD, the AgrBs sequenced from various staphylococcal species are also divergent, except the N-terminal region located in the cytoplasm and the two highly hydrophilic regions that are proposed to be in the membrane according to the AgrB topological model (22). Computer analyses show similar hydropathy profiles among the AgrB sequences (22). It is likely that all AgrBs are structurally and functionally similar and the mechanisms of processing AgrD and of secreting the mature AIP by AgrBs are the same or similar even though the AgrD propeptides are different and the interaction between AgrB and AgrD is specific (23). In this study, we identified two amino acid residues in both S. aureus and S. intermedius AgrBs that were involved in the proteolytic cleavages of AgrD, and proposed that the staphylococcal AgrB protein was a putative cysteine endopeptidase.

EXPERIMENTAL PROCEDURES
Bacterial Strains and Growth Conditions-S. aureus RN6390B is a derivative of NCTC8325 and is our standard laboratory group I strain. RN6911 is a derivative of RN6390B in which the agr locus is replaced by the tetM gene (18). SA502A is our standard S. aureus group II strain (9) and GJ2035 contains plasmid pI254 carrying an inducible bla operon in RN6911 (22). S. intermedius ATCC29663 was obtained from the American Type Culture Collection (Manassas, VA). Escherichia coli JM109 was used for cloning and BL21(DM3) was used for protein expression. The plasmids used in this study are listed in Table I. S. aureus cells were grown in CYGP broth (24), supplemented with either 5 g/ml chloramphenicol or 5 g/ml erythromycin or both when necessary. E. coli cells were grown in LB medium (25). Cell growth was monitored with a Klett-Summerson colorimeter with a green (540 nm) filter (Klett, Long Island City, NY). S. aureus cells expressing AgrB or AgrD or both under the control of the staphylococcal P bla promoter were induced with 0.5 g/ml methicillin. E. coli cells expressing the Agr protein(s) under the T7lac promoter control were induced with 1 mM isopropyl 1-thio-␤-D-galactopyranoside.
Plasmid Constructions-Plasmid pLZ5001 was constructed by cloning a ClaI DNA fragment containing the S. aureus group I agrB-His 6 gene under the control of the P bla promoter from pLZ2004 (22) into the ClaI site of pRN6441. pLZ5002 was made by ligating two NcoI-digested PCR products amplified from pLZ2004 using T4 polynucleotide kinasephosphorylated primers: GJ number 50, 5Ј-GTCTTAGCTAAAAATAT-AGG-3Ј (in agrB, nt 1878 -1897 of the S. aureus group I agr, agr-1; GenBank TM accession number X52543) and GJ number 45, 5Ј-GTA-AATGAAGTCCATGGAATAATAG-3Ј (around the NcoI site of pRN5548), and primers GJ number 51, 5Ј-CAATTTTACACCACTCTC-CTC-3Ј (in agrB, nt 1758 -1778 in agr-1) and GJ number 44 5Ј-CTAT-TATTCCATGGACTTCATTTAC-3Ј (complementary to GJ number 45), respectively. pLZ5003 was constructed the same way as pLZ5002 except the primer pairs used were GJ number 58 5Ј-AAGTGCACCAT-CACCCTTCTTCTTTTTGGT-3Ј (in agrB, nt 1996 -2025 in agr-1; changed nt, underlined) and GJ number 45, and GJ number 59 5Ј-AAGTGCACATGATGATGTCTTCTTATTAAATAAAAT-3Ј (in agrB, nt 1970 -2005 in agr-1; changed nt, underlined) and GJ number 44, and the PCR products were digested with HgiAI/NcoI. Plasmids pLZ5004 to pLZ5006 were constructed by ligating an appropriate NcoI/PstI DNA fragment of pLZ2004 and a NcoI/PstI-digested PCR product generated from pLZ2004 using primers GJ number 61 5Ј-AACTGCAGCAGCAGC-TACTGAGATTACACCTAAAG-3Ј (in agrB, nt 2122-2154 in agr-1; changed nt, underlined; PstI site, italic) and GJ number 44 (for pLZ5004), or LZ number 22 5Ј-GTTGCTGCAGGAGCAGCTACT-GAGAT-3Ј (in agrB, nt 2133-2158 in agr-1; changed nt, underlined; PstI site, italic) and GJ number 44 (for pLZ5005), or LZ number 23 5Ј-TGCTCCTGCAGCAACTGCTGCTAAGCCCAT-3Ј (in agrB, nt 2144 -  (20) was amplified by PCR using oligonucleotides WP12, 5Ј-CCATCACCAATGTGATGATG-3Ј (P3 promoter region), and SINT11, 5Ј-CGGCTCTCCTCCTTGTTT-3Ј (P2 promoter region), as primers and chromosomal DNA as template. The PCR product was cloned into an E. coli cloning vector pGEM-T (Promega). The resulting plasmid pGEM-T-P2P3 was digested with SpeI and PstI, and ligated to a XbaI/PstI DNA fragment prepared from pRN6441-gfp, which was made by cloning the green fluorescent protein (gfp) gene from pEBD166 (26) (kindly provided by Dr. Kane at Uniformed Services University of the Health Sciences) plus the SD sequence from the sarA gene of S. aureus (27). An EcoRI/PstI DNA fragment of pGEM-T-P2P3gfp was then cloned into the EcoRI/PstI sites of pRN5543. Plasmid pWP1003 was constructed by cloning the XbaI/PstI DNA fragment of pRN6441-gfp into the XbaI/PstI sites of pRN5543. To construct pWP1103, a PCR product was prepared using oligonucleotides SINT3, 5Ј-GCGAATTCACATGAGAATTTTAGAAG-3Ј (5Ј of S. intermedius agrD (agrD-Si, GenBank accession number AY557375), EcoRI site, underlined) and SINT4, 5Ј-GCGATTCATGATTAATGATGATGATGAT-GATGTTTTTCCTCTTCTAACAACTCAGC-3Ј (3Ј end of agrD-Si; 6 histidine codons and a stop codon, italic; BspHI site, underlined) as primers and S. intermedius chromosomal DNA as a template. The EcoRI/ BspHI-digested PCR product was then cloned into the EcoRI/BspHI sites of pLZ4012 (19). Plasmid pWP1104 was made by cloning a PCR product amplified from S. intermedius chromosomal DNA using primers SINT1 and SINT5 into the XbaI site of pWP1103. Plasmid pWP1104 (switch) was prepared as follows. PCR product A was generated using oligonucleotide GJ number 14, 5Ј-GCTCTAGATCGTATAAT-GACAG-3Ј (before the SD sequence of the S. aureus group I agrB, XbaI site, underlined) and SINT24 5Ј-GGTGTTAATCACGACAACCTG-CATCCCTAATCGTAC-3Ј (amino acid residues 35-40 codons of S. intermedius agrB (agrB-Si), italic; amino acid residues 29 -34 codons of S. aureus group I agrB, underlined) as primers and pRN6397 (the S. aureus group I agr cloned into E. coli vector pUC19 (3)) as a template. PCR product B was made by using oligonucleotides SINT23 (complementary to SINT24) and SINT5 as primers, and S. intermedius chromosomal DNA as a template. The PCR product A and B mixture (1:1) was used as the template for a PCR with primers GJ number 14 and SINT5. The final PCR product was then cloned into the XbaI site of pWP1103. Plasmid pWP1201 was constructed by cloning an NheI DNA fragment of pLZ4012 into the NheI site of the E. coli expression vector, pET11a. To make plasmid pWP1202, an PCR product using oligonucleotides GJ number 14 5Ј-GCTCTAGATCGTATAATGACAG-3Ј (before the SD sequence of agrB, XbaI site, underlined) and GJ number 15-1, 5Ј-GCGTCTAGATCATTTTAAGTCCTCC-3Ј (3Ј of agrB, italic; XbaI site, underlined) as primers, and pRN6852 (5) as a template. After digestion with XbaI, the PCR product was cloned into the XbaI site of pWP1201. Site-directed mutagenesis of agrB-Si was done using PCR primers containing the desired mutations and the ExSite PCR-based site-directed mutagenesis method according to the manufacturer's instruction (Stratagene), resulting in plasmids pWP1111 to pWP1120 (Table I). Plasmid pWP1105 was constructed by cloning an XbaI DNA fragment containing the H77A mutation in agrB-Si prepared from pWP1118 into the XbaI site of pWP1102. The nucleotide sequences of the cloned wild type and mutated genes in the constructed plasmids were confirmed by DNA sequencing.
Membrane Vesicle Preparation-S. aureus cells were grown in CYGP broth to 70 Klett units, and induced with 0.5 g/ml methicillin at 37°C for 3-5 h. Cells were harvested by centrifugation, washed with 1ϫ sucrose-sodium maleate-MgCl 2 (1 ϫ SMM) (24), and suspended in 1ϫ SMM containing 10 g/ml lysostaphin. After 60 min incubation at 37°C, protoplasts were prepared and washed with 1ϫ SMM plus 5 mM EDTA. The protoplasts were then lysed by the addition of ice-cold buffer A (20 mM HEPES, pH 7.2, 5% glucose, 5 mM EDTA) followed by brief sonication. After centrifugation at 12,000 ϫ g for 10 min at 4°C to remove unlysed cells, the total cell lysates were centrifuged at 100,000 ϫ g for 90 min at 4°C to separate the cell membrane vesicles and the cytoplasmic fractions. The membrane vesicles were suspended in ice-cold buffer A for immediate use or snap-frozen in liquid nitrogen and stored at Ϫ80°C until use.
In Vitro Processing of AgrD-Membrane fusion experiments were performed using the freeze-thaw method as described (28). Equal volumes of membrane fractions prepared from S. aureus cells expressing AgrB or AgrD were mixed, and the mixtures were snap-frozen in liquid nitrogen followed by thawing at room temperature or at 4°C three times to fuse the membranes. After incubation at 37°C, the mixtures were centrifuged at 100,000 ϫ g for 60 min at 4°C. The supernatants were used for AIP activity assays, and the membrane pellets were suspended in 1ϫ T7 bind/wash buffer (Novagen) plus 2% sodium cholate and protease inhibitor mixture set II (1:17 final dilution) (Calbiochem). The T7 epitope-tagged intact AgrD, processing intermediate(s), and final product(s) were purified by affinity chromatography using T7 monoclonal antibody agarose (Novagen) and suspended in 1ϫ SDS-PAGE sample buffer (25).
Western Blot Hybridization-Samples were incubated at 42°C for 30 min. Proteins were separated on Tris/Tricine SDS-PAGE (29), and the separated proteins were then transferred to polyvinylidene difluoride membrane (Millipore). After incubation at 4°C overnight, or at room temperature for 1 h in TBS buffer (25) plus 0.05% Tween 20 and 5% bovine serum albumin, anti-T7 tag monoclonal antibody (1:8000 dilutions; Novagen) was added, and the membranes were incubated for 1 h at room temperature. The membranes were washed extensively with TBS plus 0.05% Tween 20, and probed with horseradish peroxidaseconjugated goat anti-mouse Ig antibody (Amersham Biosciences). The immunoblots were detected with ECL plus Western blotting detection system (Amersham Biosciences) followed by exposure onto Hyperfilm TM ECL (Amersham Biosciences).
GFP Fluorescence Assays-S. aureus cells containing pWP1002 or pWP1003 were grown in CYGP to 40 Klett units at 37°C. Various concentrations of protease inhibitors were then added to the culture. After incubation at 37°C overnight, cells were washed and suspended in 1ϫ phosphate-buffered saline (25). Cell suspensions (100 l) were transferred into the wells of a 96-well plate (Corning Inc.). GFP was measured using the MFX Microtiter Plate Fluorometer (Dynex Technologies).

Identification of AgrD Processing
Intermediates-In a previous report (22), we showed that the S. aureus AgrB was involved in the proteolytic cleavage of AgrD to generate mature AIP. However, the identity of the potential AgrD processing intermediate(s) could not be determined because the AgrD used in that study was His 6 double-tagged at both the N and C termini, so the processing products containing either the Nterminal or the C-terminal portion of AgrD were not distinguishable. Subsequently, we made a new S. aureus AgrD in which a T7 tag was fused at its N terminus and a His 6 tag at its C terminus (TLDH) (19). This double-tagged AgrD was used in this study to identify the possible AgrD processing intermediate(s) by AgrB in S. aureus. In the absence of AgrB, no mature AIP was detected from S. aureus cells expressing TLDH alone as determined by AIP activity assay using reporting cells (Fig.  1B), and TLDH was expressed as a single band with an apparent molecular mass of ϳ9 kDa corresponding approximately to the calculated molecular mass of TLDH (9,182 Da) as detected by Western blot analysis with either an anti-T7 tag monoclonal antibody (Fig. 1A) or an anti-penta-His monoclonal antibody as a probe (data not shown). In the presence of AgrB, AIP activity was detected (Fig. 1B), and two additional bands with apparent molecular masses of ϳ6.5 and 5.5 kDa, respectively, were detected by Western blot analysis with anti-T7 tag antibody (Fig.  1A), but not with anti-penta-His antibody (data not shown). These two additional bands were also detected by Western blot analysis of the purified samples from S. aureus cells expressing both TLDH and AgrB using T7 tag antibody-agarose affinity chromatography but not nickel-nitrilotriacetic acid-agarose beads (Fig. 1C). These results indicated that the two additional bands were AgrD processing intermediate peptides B and C (Fig. 1D). Attempts to confirm the identities of these two peptides by mass spectrometry have not been successful due to the failure of obtaining enough purified materials because both peptides were membrane bound (data not shown). We have also not been able to detect peptide D both in the cytosol and the culture supernatant (data not shown), suggesting that peptide D might not be stable.
Accessory Protein(s) for AgrB?-We clearly demonstrated that AgrB was involved in the processing of AgrD to generate mature AIP. We then asked whether AgrB alone or together with other protein(s) carried out these processes because the processing of AgrD involved two proteolytic cleavages, the thioester (or ester) bond formation and the secretion of the mature AIP. To address this question, we first used 3,3Ј-dithiobis(succinimidyl propionate) as a cross-linking reagent in an attempt to identify protein(s) that might be associated with AgrB. However, no AgrB accessory proteins were found using such an approach. We reasoned that if there were additional proteins that might be involved in the processing of AgrD and the secretion of the mature AIP, it was unlikely that these proteins would exist in both S. aureus and E. coli because no AgrD homologues were identified from E. coli by computer analysis. Subsequently, we made two constructs in which the tagged agrD or agrB genes were cloned into an E. coli expression vector. One construct carried DNA sequence encoding the T7 epitope fused at the 5Ј end of agrD and sequence encoding 6 histidine residues and a stop codon fused at the 3Ј end of agrD. The other contained both the double-tagged agrD and wild type agrB genes. In both constructs, the genes were under the control of the inducible T7lac promoter. Whole cell lysates were prepared from E. coli cells expressing the double-tagged AgrD with or without AgrB, and the proteins were separated onto SDS-PAGE and detected by Western blot using an anti-T7 monoclonal antibody as a probe. As shown in Fig. 2A, a band with an apparent molecular mass of approximate 8 kDa corresponding to the calculated mass of the double-tagged AgrD (8,223 Da; peptide A* in Fig. 2C) was detected in lanes containing lysates from induced E. coli cells expressing the doubletagged AgrD alone, but not in lanes containing lysates from uninduced cells or control cells (cells carrying the cloning vector alone), indicating that the double-tagged AgrD were well expressed in E. coli. In the presence of AgrB, an additional anti-T7 antibody reactive band with an apparent molecular mass of 5.5 kDa was detected ( Fig. 2A), and the size of this band (Fig. 2B, lane 1) suggested that it likely corresponds to the AgrD processing intermediate peptide B* (Fig. 2C). We note that the sizes of the double-tagged AgrD in S. aureus and E. coli were different because the length of the amino acid residues preceded the T7 epitope sequence, so the tagged AgrD and its potential processing products in E. coli were labeled as A*, B*, and C*, to reflect their size differences from those in S. aureus. Interestingly, an additional band with an apparent molecular mass of about 25 kDa was also detected. This band was only seen in lanes containing lysates from E. coli cells expressing both the double-tagged AgrD and AgrB (Fig. 2A). The true identity of this band remains unknown. However, we suspected that this band might represent an AgrD processing intermediate covalently bound to AgrB because AgrD was T7 tagged but not AgrB. Repeated attempts to detect the AgrD processing intermediate peptide C* and AIP activity failed. It is possible that in E. coli the peptide C* generated is degraded rapidly and the mature AIP that is presumably secreted into the periplasm is either degraded or inactivated by the destruction of the thioester bond because the thioester bond is absolutely required for its agr activating activity (5,12). It is equally possible that in E. coli AgrB can only proteolytically cleave AgrD at its C-terminal processing site. Taken together, although we could not rule out the possibility that other pro-FIG. 1. The AgrD processing intermediates in S. aureus. S. aureus cells were grown, induced with methicillin, and samples were taken at various induction times as indicated. After centrifugation, cell pellets were used to prepare total cell lysates followed by Western blot analysis and the culture supernatants were subjected to AIP activity assays. Cells carrying the cloning vector were used as controls. A, Western blot hybridization of total cell lysates separated by 16% polyacrylamide Tris/Tricine SDS-PAGE with an anti-T7 tag monoclonal antibody. B, AIP activities produced from cells expressing TLDH in the presence or absence of AgrB. Values are means from three independent experiments with S.E. as indicated. C, Western blot analyses of purified TLDH and its processing products. S. aureus cells expressing AgrB or TLDH, or both were grown and induced for 5 h. Total cell lysates were prepared and proteins were then purified either by anti-T7 monoclonal antibody-agarose beads or by nickel-nitrilotriacetic acid (Ni-NTA) bead chromatography. The purified proteins were separated by 21% polyacrylamide Tris/Tricine SDS-PAGE followed by Western blot analysis using an anti-T7 monoclonal antibody as a probe. D, a schematic diagram of TLDH and its possible processing intermediates. The tags and the AIP region are indicated. The intact TLDH and its processing intermediates and their predicted molecular weights are shown (A-D). teins were involved in the processing of AgrD to generate mature AIP, we confidently concluded that the cleavage of AgrD at its C-terminal processing site was carried out by AgrB alone.
AgrB Is a Putative Endopeptidase-We used several protease inhibitors in an attempt to block the mature AIP production to determine whether AgrB had peptidase activity. S. aureus cells expressing either double-tagged AgrD (TLDH) or TLDH plus AgrB were grown in CYGP broth to ϳ70 Klett units at 37°C. Methicillin (0.5 g/ml) together with 10 mM 4-(2-aminoethyl)benzenesulfonyl fluoride-HCl (AEBSF, serine protease inhibitor), 50 M 1-trans-epoxysuccinyl-L-leucylamido-(4-guanidino)butane (E-64, cysteine protease inhibitor), 50 M [(2S,3R)-3-amino-2-hydroxy-4-phenylbutanoyl]-Leu (Bestatin, aminopeptidase B, and leucine aminopeptidase inhibitor), or 5 M isovaleryl-Val-Val-4-amino-3-hydroxy-6-methylheptanoyl-Ala-4-amino-3-hydroxy-6-methylheptanoic acid (pepstatin A, aspartic protease inhibitor) were then added. After incubation at 37°C for 4 h, the cultures were centrifuged. The supernatants (conditioned media) were used for AIP activity assays and the cell pellets were subjected to Western blot analysis using an anti-T7 monoclonal antibody as a probe. AEBSF slightly inhibited cell growth and significantly inhibited the production of AIP (data not shown). In contrast, other inhibitors tested had no effects on either cell growth or AIP production (data not shown). However, AEBSF also inhibited the ␤-lactamase activities (data not shown) that were used as an indicator for AIP activity (5,9). Subsequently, we made two new constructs: pWP1003 contained a promoterless green fluorescent protein (gfp) gene, and pWP1002 had the P3 promoter fused with the gfp gene. These plasmids were then transformed into the wild type S. aureus RN6390B. Cells were grown in CYGP media to ϳ40 Klett units, various concentrations of AEBSF (1.25, 2.5, 5, or 10 mM) or E-64 (15, 30, 66, or 100 M) were then added. After 15 h incubation at 37°C, the cultures were subjected to GFP assays. E-64 had no effect on the production of GFP (data not shown). However, AEBSF (10 mM) significantly inhibited GFP production in cells carrying the P3-gfp construct, whereas cells containing the promoterless gfp had no detectable GFP produced (data not shown). These results suggested that AgrB might be a putative protease.
To clearly demonstrate that the processing of AgrD by AgrB can indeed be inhibited by protease inhibitor(s), an in vitro AgrD processing system will be needed. Attempts to use purified double-tagged AgrD (TLDH) and purified membrane vesicles from cells expressing AgrB to develop such a system failed because of the fact that both AgrB and AgrD were membrane proteins (19,22) and the purified AgrD might not be able to integrate into the membrane properly (data not shown). So the membrane fusion method using purified membrane vesicles from S. aureus cells expressing either TLDH or AgrB-His 6 was employed. Equal volumes of the purified membrane vesicles were mixed and the mixtures were then frozen and thawed three times in liquid nitrogen and at room temperature according to the method described (28,30). We note that the majority of the fused membrane vesicles by the freeze-thaw method are sealed (30). After incubation, the mixtures were separated onto SDS-PAGE and detected by Western blot hybridization with an anti-T7 monoclonal antibody as a probe. Only TLDH proteins were detected in lanes with the mixture of TLDH containing membrane vesicles plus control membrane vesicles prepared from S. aureus agr-null mutant GJ2035 carrying the cloning vector pRN5548. In the presence of AgrB containing membrane vesicles, an additional band was detected (Fig. 3A). This band had an apparent molecular mass of 6.5 kDa corresponding to AgrD processing intermediate peptide B (Fig. 1D). We have not been able to detect either the AgrD processing intermediate peptide C or AIP activity from this reaction mixture (data not shown). It is possible that further processing of peptide B to generate peptide C is significantly less efficient in vitro so that peptide C and the AIP generated are not detectable by our assay methods, or that the processing of peptide B, the thioester bond formation, and the secretion of the mature AIP are coupled so that the secretion of the mature AIP cannot be accomplished by our current method. It is also possible that further processing of peptide B requires additional factor(s) even though the identification of AgrB accessory protein(s) have not be successful and the addition of cell extracts from FIG. 2. AgrD processing by AgrB in E. coli. The S. aureus agrB and double-tagged agrD genes were separately cloned into E. coli expression vectors in which the cloned gene was under the control of the T7lac promoter. A, cells expressing the tagged AgrD in the presence or absence of AgrB were grown in LB broth to ϳ70 Klett units at 37°C and induced with 1 mM isopropyl 1-thio-␤-D-galactopyranoside for various times as indicated. After centrifugation, cell pellets were mixed with SDS-PAGE sample buffer and the proteins were separated by 16% polyacrylamide Tris/Tricine SDS-PAGE followed by Western blot using an anti-T7 monoclonal antibody as a probe. B, cell lysates prepared from E. coli cells expressing T7-AgrD-His 6 alone (induced with isopropyl 1-thio-␤-D-galactopyranoside for 1 h) or together with AgrB (induced with isopropyl 1-thio-␤-D-galactopyranoside for 3 h), and from S. aureus cells expressing TLDH and AgrB (induced with methicillin for 4 h) were separated on 16% polyacrylamide Tris/Tricine SDS-PAGE, and Western blot analysis was performed using an anti-T7 monoclonal antibody as a probe. C, a schematic diagram of the tagged AgrD and its processing intermediate in E. coli is shown with the tags, AIP region, predicted AgrD processing intermediates, and their predicted molecular weights (A*, B*, and C*) as indicated. In panels A and B, the intact T7-AgrD-His 6 and its processing product in E. coli are indicated as A* and B*, respectively. In addition, the presumed AgrB/AgrD complex formed in E. coli (? in panel A) and the intact TLDH and its processing products in S. aureus (A, B, and C in panel B; see Fig. 1D) are labeled.
S. aureus agr-null mutant RN6691 to our in vitro system had no effects (data not shown).
The AgrB Conserved Regions That Are Required for Its Function-The AgrBs sequenced so far from S. aureus, S. intermedius, Staphylococcus lugdunesis, and S. epidermidis were considerably different as revealed by computer analysis (Fig. 4A). Interestingly, the N-terminal region, two regions that are highly hydrophilic and are proposed to be two transmembrane segments (22) were highly conserved. In fact, the N-terminal 34-amino acid residues were identical among the four groups of S. aureus AgrBs. To determine the role of these conserved regions in AgrBs in the processing of their cognate AgrDs to generate their corresponding mature AIPs, we first made an S. aureus AgrB mutant in which the N-terminal 34 amino acid residues were deleted. S. aureus cells co-expressing this AgrB mutant and wild type AgrD did not generate any detectable mature AIPs as determined by agr activation using RN6390B(pRN6683) (group I S. aureus reporter) and agr inhibition using SA502A(pRN6683) (group II S. aureus reporter) assays (Fig. 4, C and D) (5,9), although this mutant was located in the membrane (data no shown) and its expression was much stronger than that of the wild type AgrB as shown by Western blot hybridization analysis using an anti-penta-His monoclonal antibody as a probe (Fig. 4B). Similarly, mutations in the conserved region located in the first transmembrane hydrophilic region (Fig. 4A) totally eliminated AgrB activity (Fig. 4, C and D). These results indicated that the N-terminal and the AHGAHA conserved regions were critical for AgrB function. In contrast, mutations in the second hydrophilic transmembrane region had no effects on its activity (Fig. 4, C  and D).
Identification of the Putative Catalytic Residues in Staphylococcal AgrB-Because the N-terminal region of AgrB was required for its activity but was not a leader peptide involved in the translocation of AgrB into the membrane, it was possible that this region was involved in the proteolytic cleavage of AgrD. To explore this possibility, we first made a chimeric AgrB in which the N-terminal 34 amino acids of AgrB-Si was replaced by that of S. aureus group I AgrB. The chimeric AgrB could process AgrD-Si as determined by AIP-Si activation assays using S. intermedius reporter cells (Fig. 5A), AIP-Si inhibition assays using S. aureus group II reporter cells (Fig. 5B), and Western blot hybridization analysis using an anti-T7 monoclonal antibody as a probe (Fig. 5C). We note that AgrD-Si used in these experiments was double-tagged with a T7 tag at the N terminus and a His 6 tag at the C terminus, which would facilitate the detection of AgrD-Si as well as its potential processing intermediate(s) by Western blot analyses using commercially available antibodies, and the addition of these tags had no effects on AgrD-Si processing and AIP-Si secretion (data not shown). As shown in Fig.  5C, a band with a molecular mass of ϳ9 kDa corresponding to the calculated molecular mass of the double-tagged AgrD-Si (9,205 daltons) was detected in the lane containing a lysate of cells lacking agrB-Si and producing the tagged AgrD-Si. In the presence of either the wild type AgrB-Si or the chimeric AgrB, an anti-T7 antibody responding band with a molecular mass of ϳ6 kDa was detected. Whether this band was an AgrD-Si processing intermediate or a final product has not yet been determined. These results suggested that the conserved residues in the N-terminal region of AgrB might be critical for its activity. However, subsequent site-directed mutagenesis analyses of these conserved residues revealed that these mutants (D5A, N20A, H23A, R30A, Q34A, N39A, and K42A) had no significant effects on the AgrB-Si activity (Fig. 5). These results indicated that the N-terminal region of AgrB was not involved in the proteolytic cleavage of AgrD. It is possible that this region is required for the initial binding of AgrD to AgrB that has not yet been determined experimentally.
From the AgrB alignment analysis shown in Fig. 4A, we noticed that three residues (histidine 77, serine 81, and cysteine 84) were identical among AgrBs. It is possible that these amino acids are putative catalytic residues that are involved in the proteolytic processing of AgrD. Accordingly, we replaced these residues with alanines in S. aureus group I AgrB. Interestingly, changing the histidine residue at position 77 (H77) or   FIG. 3. The effects of peptidase inhibitors on the AgrD processing in vitro. A, kinetics of AgrD processing. Equal volumes of membrane vesicles prepared from S. aureus cells expressing TLDH, or AgrB-His 6 , or none were mixed. After freeze and thaw three times, the mixtures were incubated at 37°C for various times as indicated. Sample buffer was then added and the mixtures were incubated at 42°C for 30 min. Proteins were separated on 21% polyacrylamide Tris/Tricine SDS-PAGE followed by Western blot analysis with an anti-T7 monoclonal antibody as a probe. Total cell lysate of S. aureus cells expressing both TLDH and AgrB-His 6 was used as control. B, the effects of protease inhibitors on AgrB. Membrane vesicles prepared from S. aureus cells expressing AgrB-His 6 were treated with 10 mM AEBSF, 1.2 mM VFK-CMK, 1.2 mM TPCK, 100 M E-64, 100 M pepstatin A, or 100 M bestatin at 37°C for 30 min. The mixtures were ultracentrifuged and the membrane fractions were then suspended in buffer A (20 mM HEPES, pH 7.2, 5% glucose, 5 mM EDTA) and mixed with equal volumes of TLDH containing, or control (non-Agr proteins), membrane vesicles. After fusing the two membranes, the mixtures were incubated at 37°C for 2 h, and the proteins were separated on 21% polyacrylamide Tris/Tricine SDS-PAGE followed by Western blot using an anti-T7 monoclonal antibody as a probe. The intact TLDH (peptide A) and its processing intermediate (peptide B) were indicated.
the cysteine residue at position 84 (C84) totally eliminated its ability to process AgrD as shown by AIP activity assays (Fig. 6,  A and B), Western blot hybridization analysis of total cell lysates (Fig. 6C), and in vitro AgrD processing by membrane fusions (Fig. 6D), whereas mutation of serine 81 (S81) had little effect on the function of AgrB (Fig. 6). Additionally, changing histidine 74 or serine 80 to an alanine residue had no significant effect on AgrB activity. Of note, the total cell lysates and  (20). B, expression of AgrB-His 6 and mutants. Total cell lysates prepared from S. aureus cells were separated by 12% SDS-PAGE followed by Western blot using an anti-penta-His monoclonal antibody as a probe. Lane 1, control, GJ2035 containing cloning vector pRN5548; lane 2, AgrB-His 6 ; lane 3, AgrB-His 6 (⌬N); lane 4, AgrB-His 6 (⌬Bsi); lane 5, AgrB(⌬Pst)-His 6 . C and D, AIP activity assays. Conditioned media were prepared from cultures of S. aureus cells expressing wild type AgrD in the absence (control) or presence of AgrB-His 6 (WT) or AgrB mutants. AIP activation activities (C) using S. aureus group I reporter cells and AIP inhibition activities (D) using S. aureus group II reporter cells were measured as described (5,9). Values are means from three independent experiments with S.E. as indicated.
the in vitro membrane fusion mixtures were also subjected to Western blot analyses using an anti-penta-His monoclonal antibody as a probe. As shown in Fig. 6, the AgrB-His 6 or mutant proteins were expressed approximately equally from cells used for cell lysate preparations (Fig. 6E), and membrane vesicles containing equal amounts of AgrB-His 6 proteins were used in the in vitro experiments (Fig. 6F). These results indicated that both His 77 and Cys 84 were required for proteolytic processing of AgrD. Similar results were obtained using AgrB-Si H77A, S81A, and C84A mutants (data not shown).
To eliminate the possibility that cysteine84 was involved in the polymerization of AgrB, we made a new construct (pWP1105) that contained a copy of the wild type agrB-Si gene, a copy of a mutated gene (agrB-Si(H77A)) and the agrD-Si gene. As shown in Fig. 7, S. aureus cells expressing AgrD-Si and both wild type and mutant AgrB-Si produced mature AIP-Si in an amount comparable with that of cells expressing AgrD-Si and the wild type AgrB-Si only. Similarly, no signifi-cantly different AIPs were produced from S. aureus cells harboring plasmid carrying the S. aureus group I wild type of agrB and agrD genes with or without a copy of the agrB(C84S) gene (data not shown).

DISCUSSION
Quorum sensing systems in Gram-positive bacteria predominately use peptides as the signals that are the ligands of membrane receptors (histidine kinases) or that interact with other protein(s) which in turn regulates the activity of the response regulator of the signal transduction pathways. The peptide signal molecules identified so far are processed from FIG. 5. The conserved residues of the AgrB N-terminal region have no effects on its endopeptidase activity. S. aureus GJ2035 cells expressing both the double-tagged S. intermedius AgrD-Si (TLD-SiH) and the wild type S. intermedius AgrB-Si or AgrB-Si switch mutant (the N-terminal 34 residues was replaced by the same region of S. aureus group I AgrB), or AgrB-Si mutants were grown and induced with 0.5 g/ml methicillin. Cells were centrifuged, and the culture supernatants were used as the conditioned media to perform AIP-Si activation activity assays using S. intermedius containing pWP1004 as reporter cells (A), and AIP-Si inhibition assays using S. aureus SA502A(pRN6683) as reporter cells (B). Values are means from three independent experiments with S.E. as indicated. For Western blot hybridization analysis (C), S. aureus cells were grown, induced, and the cell cultures were mixed with ethanol/acetone (1:1). The mixtures were centrifuged and the cells were washed and lysed. Whole cell lysates were separated on 16% polyacrylamide Tris/Tricine SDS-PAGE and transferred to nitrocellulose membranes. The membranes were then probed with an anti-T7 tag monoclonal antibody.
FIG. 6. The catalytic residues in AgrB endopeptidase. S. aureus cells expressing TLDH (control) and AgrB-His 6 (WT) or AgrB-His 6 mutants were grown and induced. The culture supernatants were then used as conditioned media and the AIP activation activities (A) using S. aureus group I reporter cells and inhibition activities (B) using S. aureus group II reporter cells were measured. Values are means from three independent experiments with S.E. as indicated. The cell pellets were washed and lysed. The total cell lysates were then separated on 21% polyacrylamide Tris/Tricine SDS-PAGE followed by Western blot analysis using an anti-T7 monoclonal antibody as a probe (C) or on 10% polyacrylamide Tris/glycine SDS-PAGE followed by Western blot analysis using an anti-penta-His monoclonal antibody as a probe (E). In vitro AgrD processing by mixing membrane vesicles prepared from S. aureus cells expressing TLDH and non-AgrB (control), or AgrB-His 6 , or AgrB-His 6 mutant (D) were performed using the method as described under "Experimental Procedures." The mixtures were also separated on 10% polyacrylamide Tris/glycine SDS-PAGE and the amounts of AgrB proteins in each mixture were determined by Western blot analysis using an anti-penta-His monoclonal antibody as a probe (F). WT, wild type.
ribosomally synthesized precursor peptides and in some cases are subjected to post-translational modification. Among the known peptide-mediated regulatory systems, two mechanisms of the precursor processing to generate mature active peptides have been proposed. One mechanism involves the cleavage of a leader peptide containing precursor either by a dedicated peptidase followed by the secretion of the mature peptide via an ABC transporter (e.g. the Lactococcus lactis Nis system that involves the production of lantibiotic nisin (33,34)), or by a presumed peptidase domain of an ABC transporter preceded secretion via the same transporter (e.g. the Streptococcus pneumoniae Com system that regulates the competence (35,36)). Another one, which has only been described for the Phr system of Bacillus subtilis that is required for competence and sporulation (37)(38)(39)(40), involves the cleavage of a leader peptide during secretion via the SecA-dependent transport system. The secreted peptide, a processing intermediate, is subjected to further proteolytic cleavage extracellularly by a peptidase that has not yet been identified and the mature active peptide is then transported into the cytoplasm by the oligopeptide-permease and negatively regulate the Rap-phosphatases activities.
The staphylococcal Agr quorum sensing system has unique characteristics among the peptide-mediated quorum sensing systems studied so far: first, its signal molecule is a thiolactone or lactone molecule; second, the processing of the AgrD precursor involves two proteolytic cleavages and the AIP sequence is in the middle, which is different from other systems in that the precursors are subjected to only one cleavage and the C-terminal portions are served as the signal molecules; and third, the AgrB protein, which has been proposed as transporter for the secretion of the mature AIP (22), does not have the characteristics of an ABC transporter.
Previously, we showed that AgrD was proteolytically cleaved in the presence of the AgrB protein, however, it was not clear whether the reaction was carried out by AgrB or AgrB-associated accessory protein(s) (22). In this study, we showed that AgrB was a putative endopeptidase whose activity could be inhibited by protease inhibitors. We clearly demonstrated that AgrB was involved in the cleavage of AgrD at the C-terminal processing site and identified two amino acid residues (histidine 77 and cysteine 84) in both S. aureus group I AgrB and S. intermedius AgrB that were required for their endopeptidase activities. It is possible that these two amino acids are the putative catalytic residues that form a catalytic center. We note that these two amino acids are at the same positions and conserved among all AgrBs identified so far, and are located juxtapositionally on the inner surface of the cytoplasmic membrane (Fig. 4A) (22), suggesting that the putative catalytic activities are conserved among these AgrBs. Cysteine and histidine are the two most common conserved residues that form the catalytic centers of cysteine peptidases (41), it is possible that AgrB is a putative cysteine endopeptidase although it has no sequence homologies with any known cysteine peptidases. However, why AgrB activity was not sensitive to cysteine protease inhibitors we tested is unknown. One possibility is that the cysteine protease inhibitors cannot reach the AgrB catalytic center because of their membrane permeability, or other amino acid residues surround the AgrB catalytic center and prevent these inhibitors from interacting with their targeted amino acids. Another is that those protease inhibitors cannot reach their inhibitory concentrations in the cytosol. The reason that certain serine protease inhibitors could inhibit AgrB peptidase activity might be because of their inhibitory specificities. We note that there is only one conserved serine residue in AgrBs and replacement of that serine to an alanine residue had no effect on AgrB activity, a finding strongly suggesting that AgrB was not a serine peptidase.
Other Gram-positive bacteria may use a similar mechanism with that of AgrD processing by AgrB to produce mature autoinducing peptide from precursor proteins. It is interesting to note that the recently identified Fsr quorum sensing system in Enterococcus faecalis (42) is remarkably similar to that of the Agr system. The FsrB (AgrBfs), an AgrB homologue, has been proposed to be the protein responsible for the production of an 11-amino acid residue lactone molecule (gelatinase biosynthesis-activating pheromone) that is likely to be the ligand for FsrC of the FsrC/FsrA signal transduction system (42-44). However, no agrD homologous gene is found in the fsr operon, instead, the gelatinase biosynthesis-activating pheromone sequence is in the C-terminal region of FsrB. It has been proposed that the possible mechanisms of processing and secretion of gelatinase biosynthesis-activating pheromone are similar to those of the AgrB/AgrD (45).