Evidence That VirS Is a Receptor for the Signaling Peptide of the Clostridium perfringens Agr-like Quorum Sensing System

C. perfringens beta toxin (CPB) is essential for the virulence of type C strains, a common cause of fatal necrotizing enteritis and enterotoxemia in humans and domestic animals. Production of CPB, as well as several other C. perfringens toxins, is positively regulated by both the Agr-like QS system and the VirS/R two-component regulatory system. This study presents evidence that the VirS membrane sensor protein is a receptor for the AgrD-derived SP and that the second extracellular loop of VirS is important for SP binding. Understanding interactions between SP and VirS improves knowledge of C. perfringens pathogenicity and may provide insights for designing novel strategies to reduce C. perfringens toxin production during infections.

Using the same cDNA samples, agrD and virS transcript levels were also determined by qRT-PCR. The results showed that the agrD transcript level pattern was very similar to the cpb expression pattern (Fig. 1C, left). However, virS gene transcript levels were maximally expressed at 1 h postinoculation and then decreased thereafter (Fig. 1C, right). The patterns of agrD or virS expression were very similar in both C. perfringens strains examined.
Sequence differences between the virS gene in CN1795 versus that in CN3685. Our lab previously prepared agrB mutants of type B strain CN1795 and type C strain CN3685 (25,30). These mutants have an intron insertion in their agrB gene that inactivates expression of both the agrB and agrD genes, which are located in the same operon (24,25,30). We also showed previously (27) that those CN1795 and CN3685 agrB mutants respond to SP-based synthetic peptides ( Fig. 2A) named 5R (a 5-aminoacid thiolactone ring that likely corresponds to the natural C. perfringens SP [27]), but only the CN3685 agrB mutant responds to 8R, which is 5R plus a 3-amino-acid tail. The current study first confirmed those previous observations using newly synthesized 5R and 8R peptides. The results, shown in Fig. 2A, indicated that the CN1795 agrB null mutant only responded to the 5R peptide. In contrast, the CN3685 agrB null mutant responded to both the 5R and 8R peptides, although there was stronger signaling (more CPB production) with the 5R versus the 8R peptide, consistent with our previous study (27).
Since it has been hypothesized (9) that the VirS membrane sensor protein of the VirS/R TCRS is a receptor for the AgrD signaling peptide, and the two agrB null mutant strains respond differently to the 5R or 8R peptides, we compared the VirS proteins made by these two strains using virS sequencing. Differences detected between the deduced VirS sequences of these two strains are shown in Table 1. Those differences The predicted topology of VirS protein produced by two wild-type strains. The predicted CN1795 VirS structure is shown in black numbers (total of 440 amino acids). The predicted CN3685 VirS structure is shown in red numbers (total of 446 amino acids). Asterisks (*) with different colors show the differences in this VirS region between these two strains. Amino acid (Aa) differences (indicated by asterisks) between the two strains are shown on the right side, with amino acid sequence numbers for CN1795 shown in black and those of CN3685 shown in red.

Li and McClane
® included 3 single amino acid differences at residues 81, 360, and 382 (relative to the CN1795 VirS sequence). However, the most striking difference noted between the virS open reading frames (ORFs) of these two strains is that the virS ORF in CN3685 encodes a 6-amino-acid insertion at the 152nd amino acid, producing a VirS protein of 446 amino acids compared to the 440-amino-acid VirS protein made by CN1795.
The transmembrane prediction algorithm (TMHMM) predicts that the VirS protein has 4 extracellular regions that might interact with SP if VirS is an SP receptor (Fig. 2B). Furthermore, the TMHMM program predicts that the sequence differences detected between CN3685 versus CN1795 impact VirS structure, as shown in Fig. 2B. Of particular interest for this study, these sequence differences are predicted to result in an extracellular loop 2 (ECL2) of 14 amino acids for the CN1795 VirS protein versus an ECL2 of 19 amino acids for the CN3685 VirS protein (Fig. 2B, blue asterisk).
Preparation of virS agrB double-null mutant strains of CN1795 or CN3685 and swapping of the virR/S operon expressed by those double-null mutant strains. Given the differences in 8R signaling sensitivity observed between CN1795 and CN3685 ( Fig. 2A; see also reference 27) and the predicted structural differences in their VirS proteins (Fig. 2B), we inactivated the virS gene in our existing agrB mutants and then complemented those double mutants to swap which VirS protein they expressed in order to test if this swap affects 8R sensitivity, as might be anticipated if VirS is an SP receptor.
For this purpose, a Clostridium-modified TargeTron insertional mutagenesis method (31) was employed to introduce a targeted intron into the virS gene of CN1795agrBKO and CN3685agrBKO to create virS agrB double-null mutants that do not express either the virS gene or the operon encoding agrB/D. The identity of those putative double mutants, named CN1795DKO and CN3685DKO, was first demonstrated by PCR using primers specific for internal agrB open reading frame (ORF) or virS ORF sequences. Compared to the PCR products amplified from the agrB or virS genes in wild-type strains, PCR of DNA from the double mutants amplified larger bands from both genes due to insertion of an intron into the wild-type agrB and virS genes (Fig. 3A).
To create complementing strains producing a swapped VirS protein (e.g., CN1795 producing the VirS protein of CN3685), the virR/S operons of CN1795 or CN3685 were separately cloned into the pJIR750 shuttle plasmid. The plasmid carrying the CN1795 virR/S operon was then electroporated into CN3685DKO, and the plasmid carrying the CN3685 virR/S operon was electroporated into CN1795DKO, creating (respectively) two complementing strains named CN3685DKOc1795virR/S and CN1795DKOc3685virR/S that now expressed the virR/S operon of the other strain (Fig. 3A). PCR results confirmed that, in these two complementing strains, the presence of a wild-type virS gene was restored. Furthermore, an intron-specific Southern blot analysis (Fig. 3B) demonstrated that, while only one intron is present in the agrB single mutant strains, the double mutant strains now carry two intron insertions. RT-PCR results (Fig. 3C) demonstrated that, in the double-null mutant strains, neither the agrB nor the virS gene was expressed, while the swapped virR/S complementing strains exhibited restored virS gene expression. Western blot results (Fig. 3D) confirmed that CPB, which is regulated by both the Agr-like QS system and the VirS/R TCRS, was not expressed in either double-null mutant strain, even when the 5R signaling peptide was added.
The main purpose of constructing these complementing strains producing a swapped VirS was to test their responsiveness to signaling by peptides 5R or 8R, as assessed by CPB Western blot analysis (Fig. 4). The CN1795 complementing strain  double-null mutant construction by targeted intron-based mutagenesis and genetic complementation of those mutants with a swapped virR/S operon. Using DNA isolated from wild-type strains, a PCR assay amplified ϳ500-bp products using internal agrB primers (upper) or internal virS primers (lower). After targeted insertion of a 900-bp intron, the same PCR assays amplified an ϳ1.5-kb product using DNA isolated from the double-null mutant (DKO) strains. After complementation of the double mutants to carry the swapped virR/S operon (creating CN1795DKOc3685virR/S and CN3685DKOc1795virR/S), agrB PCR products remained the large size indicative of an intron insertion, but the virS PCR products were the smaller size present in wild-type strains, confirming genetic virS complementation. A 1-kb molecular ladder (Fisher Scientific) was used, with the size in bp shown at left. (B) Southern blot hybridization of an intron-specific probe with DNA from wild-type strains, the agrB single mutants, or agrB virS doublenull mutants. DNA from each strain was digested with EcoRI and electrophoresed on a 1% agarose gel prior to blotting and hybridization with an intron-specific probe. Size of DNA fragments, in kb is shown at left. (C) RT-PCR analyses for expression of 16S RNA (top), the agrB gene (middle), or the virS gene (bottom) by wild-type CN1795 (left) or CN3685 (right), their agrB virS double-null mutants (DKO), or complementing strains of those double mutants with a swapped virR/S operon. Those samples did not amplify a product when subjected to PCR without reverse transcription, indicating that the RNA preparations from all strains were free from DNA contamination (data not shown). (D) Western blot analyses of CPB production by wild-type strains, agrB virS double-null mutants, or those mutants cultured in the presence of 100 M 5R or 8R peptides. Size of proteins in kDa is shown at left. All experiments were repeated three times, and representative results of three repetitions are shown.
production did not increase further when supplied the 8R peptide. We also tested CPB production by this complementing strain when incubated with or without 8R for different time points (3 h, 5 h, 8 h, or overnight), but the presence of 8R did not increase CPB production compared to the absence of 8R at any time point (data not shown).
Evidence that a biotin-labeled 5R peptide (B-5R) can bind to the His 6 -tagged VirS protein of CN3685. Attempts to prepare a VirS antibody for use in a biotin-labeled 5R peptide (B-5R) pulldown experiment were unsuccessful. Therefore, we used a CN3685 strain producing His 6 -tagged VirS and a His 6 antibody Western blot to determine, in a pulldown assay, whether VirS can physically bind with B-5R, as would be expected if VirS is an SP receptor in C. perfringens.
For this purpose, the virS gene in CN3685 was inactivated by insertion of an intron using the Clostridium-modified TargeTron insertional mutagenesis method (31). Construction and characterization of this mutant, named CN3685::virS, is shown in Fig. 5. Specifically, PCR analyses (Fig. 5A) demonstrated the insertion of an ϳ900-bp intron into the virS gene, since the PCR product amplified from this mutant strain DNA was 900 bp larger than the PCR product amplified from the wild-type CN3685 virS gene. An intron-specific Southern blot analysis (Fig. 5B) confirmed the presence of only one intron in this mutant.
Complementing strains named CN3685::virSc3685virR/S and CN3685::virSc1795virR/S were prepared that produce, respectively, the VirS of either CN3685 or CN1795. As expected, DNA from these complementing strains supported amplification of a PCR product of the same size as the wild-type virS gene (Fig. 5A). RT-PCR (Fig. 5C) analyses detected no virS gene expression by the mutant, while the two virR/S complementing strains showed virS gene expression. Western blot analyses (Fig. 5D) detected no CPB production by the CN3685 virS null mutant strain, while both complementing strains did produce CPB, confirming that their complemented VirS expression enabled functional signaling. At the same time, we similarly prepared a CN1795 null mutant (CN1795::virS) with an intron insertion in virS and a complementing strain producing the CN1795 VirS for studies (described later in Fig. 8).

FIG 4
Western blot analyses of CPB production by wild-type, agrB virS double-null mutant, and swapped complementing strains, grown in the presence or absence of 100 M 5R or 8R peptide. Cultures were grown for 5 h at 37°C in TY medium, and the OD 600 values of each culture were adjusted to the same density. Supernatants were then collected, and equal volumes were subjected to CPB Western blotting. Size of proteins in kDa is shown at left. All experiments were repeated three times, and results representative of three repetitions are shown.

C. perfringens VirS Is a Signaling Peptide Receptor
® For use in the pulldown experiment, we also transformed the CN3685 virS null mutant to produce VirS protein with a C-terminal His 6 tag, creating CN3685::virS c3685virR/Shis; the only difference between this complementing strain and CN3685:: virSc3685virR/S is that a His 6 tag was added to the VirS C terminus using a 5=-to-3= PCR primer. For CN3685::virSc3685virR/Shis, CPB Western blotting detected similar CPB production to that observed for CN3685::virSc3685virR/S (data not shown), confirming that the presence of the His 6 tag did not interfere with VirS function.
Before using a biotin-labeled 5R peptide (B-5R) in a pulldown experiment to test for its interactions with VirS, it was important to evaluate whether this biotin-modified peptide retains biologic activity, i.e., signaling function. Western blot analysis showed that B-5R induced CPB production by both the CN1795 and CN3685 agrB null mutant strains (Fig. 6A). null mutants created by intron-based mutagenesis or complementing strains carrying a virR/S operon from CN1795 or CN3685 (indicated by c1795virR/S or c3685virR/S). Using DNA isolated from wild-type strains, an ϳ500-bp product was PCR amplified using internal virS primers. However, after targeted insertion of a 900-bp intron into virS, the same PCR assay amplified an ϳ1.5-kb product using DNA from the null mutant strains. Using DNA from the complemented strains, the same-size virS PCR products were amplified as when using DNA from wild-type strains. A 1-kb molecular ladder (Fisher Scientific) was also electrophoresed, and the size in bp is shown at left. (B) Southern blot hybridization of an intron-specific probe with DNA from wild type or the virS null mutants of CN1795 or CN3685. DNA from each strain was digested with EcoRI and then electrophoresed on a 1% agarose gel prior to blotting and hybridization with an intron-specific probe. Size of DNA fragments in kb is shown at right. (C) RT-PCR analyses of 16S RNA (top) and virS (bottom) transcription in wild-type CN1795 (left) or CN3685 (right), their virS null mutants, or complementing strains carrying the virR/S operon of either strain. To show that the RNA preparations from both strains were free from DNA contamination, the samples were also subjected to PCR without reverse transcription, but no products were amplified (data not shown). (D) Western blot analyses of CPB expression by wild type, virS null mutants, and complementing strains. Size of proteins in kDa is shown at left. All experiments were repeated three times, and results representative of three repetitions are shown.
With positive results from those control experiments, we then performed a pulldown experiment to evaluate if VirS can physically bind to SP. This involved preincubating streptavidin-coated beads with B-5R or 5R before mixing those beads with B-PER buffer cell extracts from 4-h TY cultures of CN3685::virSc3685virR/Shis cells (Fig. 6B, lane 1). When those extracts were Western blotted with His 6 -specific antibody prior to incubation with the B-5R-or 5R-treated beads, a major immunoreactive band was detected that matched the ϳ50-kDa size of VirS, although higher molecular mass species were also present, including an ϳ100-kDa species that is likely a VirS dimer. After incubation of those extracts with the B-5R-coated streptavidin beads, much of the His 6 antibody-immunoreactive material, particularly the apparent dimer, remained unbound in the supernatant (lane 2).
As a specificity control, Western blotting with His 6 tag antibody detected (lane 3) no immunoreactive material in the crude extract of a control complementing strain (CN3685::virSc3685virS/R) that does not produce His 6 -tagged VirS, confirming that the immunoreactive material present in CN3685::virSc3685virR/Shis extracts was attributable to His 6 -tagged VirS. In addition, no immunoreactive material was detected by His 6 antibody using extracts of CN3685::virSc3685virR/S after pulldown by beads preincubated with either 5R or B-5R (lanes 6 and 7). However, an ϳ50-kDa immunoreactive . After those extracts were incubated with streptavidin beads preincubated with B-5R or 5R, unbound supernatant material is shown in lanes 2 and 4. Pulldowns of extracts from complementing strains using beads preincubated with 5R or B-5R are shown in lanes 5 to 9. Note that in these lanes, using CN3685::virSc3685virR/Shis-containing extracts, but not CN3685::virSc3685virR/S-containing extracts, incubation of beads pretreated with B-5R, but not with 5R, resulted in pulldown of a protein that reacted with His 6 tag antibody and was ϳ50 kDa, the size of VirS. Size of protein markers in kDa is shown at left. All experiments were repeated three times, and results shown are representative of three repetitions.
C. perfringens VirS Is a Signaling Peptide Receptor ® species was pulled down from extracts of CN3685::virSc3685virR/Shis using beads preincubated with B-5R, which contains biotin and binds well to streptavidin-coated beads, but not from extracts of those cells preincubated with 5R, which lacks biotin and thus does not bind well to streptavidin-coated beads. This Western blot result analysis supports the ability of C. perfringens VirS to physically bind SP.
A synthetic peptide corresponding to the predicted VirS second extracellular loop inhibits SP-induced CPB production. The ability of B-5R-coated beads to pulldown VirS supports the binding of SP to VirS as a receptor. To confirm binding interactions between 5R and VirS by an independent approach and to discern a region of VirS involved in SP binding, we used the VirS structural model (Fig. 2B) to identify predicted extracellular regions of the VirS protein that could interact with SP. Since the predicted VirS ECL2 of CN1795 and CN3685 have significant size differences and those strains respond differently to signaling by 8R (Fig. 2), we hypothesized that this VirS loop is involved in SP binding. To test this hypothesis, we synthesized a 14-amino-acid peptide named KIGK ( Table 2) that corresponds to the sequence of the predicted ECL2 of CN1795 VirS, as well as a random sequence control peptide with a molecular weight similar to that of KIGK.
To test if 5R can bind to KIGK, as might be anticipated if Agr-like QS signaling involves SP binding to the predicted ECL2 of VirS, the 5R and KIGK peptides were preincubated together before their addition to cultures of CN1795 or CN3685 agrB null mutant strains. CPB Western blot analysis (Fig. 7A) showed that neither agrB null mutant strain produces CPB in the absence of 5R, as expected from the results shown in Fig. 2. However, consistent with the hypothesis that 5R binds to the second ECL of VirS, preincubation of 5R with KIGK blocked Agr-like QS signaling for the agrB null mutants, i.e., CPB production was significantly reduced. In contrast, similar preincubation of 5R with the control peptide did not block AIP signaling to agrB null mutant strains, i.e., strong CPB production was detected. To further confirm the specificity of the binding interaction between 5R and KIGK, a peptide named KIGK_D was prepared where the N residue located in the middle of the predicted VirS second extracellular loop was switched to D ( Table 2). Preincubation of 5R with this peptide had no effect on subsequent AIP signaling to agrB mutants, i.e., CPB production was not reduced (Fig. 7B).
To further test our hypothesis that SP binding involves the ECL2 of VirS, we applied KIGK (500 M or 1 mM) to CN1795 and CN3685 to determine if this peptide can also block Agr-like QS signaling by those two wild-type strains. Results (Fig. 7C) showed that the presence of 1 mM KIGK peptide efficiently blocked CPB production by both wild-type strains, while the same dose of the KIGK_D peptide had no effect on CPB production. A 500 M concentration of KIGK can partially block CPB production by these strains (data not shown).
Cross-talk regulation of expression between the virR/S operon and the agrB/D operon. In S. aureus, sensing of the AIP causes the receptor AgrC to undergo histidine autophosphorylation and then phosphorylate AgrA, which initiates a positive feedback loop to make more AgrB and AgrD (21). Since the results shown in Fig. 3 to 7 provided substantial evidence supporting VirS as an SP receptor, we tested whether a positive feedback loop also functions in C. perfringens to help regulate expression of the Agr-like QS system and the VirS/R TCRS. This first involved using virS null mutant strains to study if the presence of VirS, an SP receptor, affects agrD expression. Based upon the kinetics for agrD expression determined in Fig. 1, expression levels of the agrD gene in 5-h TY medium cultures were compared between the two wild-type strains CN1795 and CN3685, as well as cultures of their virS null mutants and their virR/S operon-complemented strains. Results from these qRT-PCR analyses revealed significantly higher agrD transcript levels in the wild type and in its virS/R-complemented strain than in the virS null mutant strains (Fig. 8A).
To investigate whether SP production affects virS expression, the CN1795 and CN3685 agrB null mutant strains were used. Note that since agrB and agrD (and two upstream genes) are located within the same operon, the agrD gene is not expressed in null mutant strains with an insertion in their agrB gene (23,24). Because virS expression levels peak early, i.e., at 1 to 2 h (Fig. 1), transcript levels of the virS gene were compared in 2-h TY medium cultures for both wild-type strains, their agrB null mutants, and those mutants complemented with the agr operon. Those qRT-PCR analyses detected significantly higher virS transcript levels in the wild type and its agr operon-complemented strains compared to those in the agrB null mutant strains (Fig. 8B).

DISCUSSION
Producing toxins such as CPB is critical for C. perfringens pathogenicity (20,32). Previous studies established that regulation of toxin production, including CPB production (19,25), in C. perfringens often involves both the Agr-like QS system and the VirS/R TCRS. However, major gaps have remained in understanding of how those systems regulate C. perfringens toxin production, e.g., do these systems work cooperatively or independently? In comparison to the agr operon of the S. aureus Agr system, the C. perfringens agr-like operon encodes a different SP and does not encode a homolog of AgrC (11,23,24), which is the SP receptor in the S. aureus Agr QS system (11). Instead, since both the Agr-like QS system and VirS/R TCRS positively regulate production of several C. perfringens toxins, it has been proposed that VirS may be an SP receptor in C. perfringens (9). However, this important hypothesis had never been directly tested.
Given that CPB production is positively regulated by both the Agr-like QS system and the VirS/R TCRS (19,25,33), the current study used production of this toxin as a readout for SP signaling. Consequently, initial experiments in this study compared the timing of cpb, agrD, and virS expression in C. perfringens. Consistent with our previous reports that CPB is most highly produced during late log-phase growth by type B or C isolates in TGY medium (28,29), the current study determined that, in TY cultures of either type B strain CN1796 or type C strain CN3685, cpb transcript levels increased from 1 h until they reached a peak at 5 h, a time corresponding to the late log phase-early stationary phase. Some transcription of agrD was observed even during early growth, but this expression also peaked at ϳ5 h, consistent with previous studies showing that CPB production is controlled by the Agr-like QS system (25). In contrast, the virS gene was most strongly expressed early during growth, as would be expected if significant

Li and McClane
® amounts of VirS membrane sensor already need to be available when the Agr-like QS maximally signals to increase CPB production.
We then conducted a series of studies to test directly whether VirS could be an SP receptor involved in regulating CPB production by C. perfringens type B or C strains. Since previous results (27) showed that an agrB mutant of type C strain CN3685 upregulates CPB production in response to both the 5R and 8R synthetic peptides but that an agrB mutant of type B strain CN1795 upregulates CPB production only in response to 5R, we reasoned that if VirS is a major SP receptor, there should be sequence differences between the VirS membrane sensor histidine kinase of these two strains. Sequencing did identify several differences between the deduced VirS amino acid sequences of these two strains, as shown in Table 1. When computer modeling was performed to predict the structural impact of those sequence variations, several structural differences were predicted, including differences in the predicted VirS ECL2.
Results of experiments performed in this study then supported a role for the second ECL loop in SP binding. For example, preincubation of a synthetic peptide corresponding to ECL2 with 5R, the likely natural SP, inhibited signaling to increase CPB production in agrB mutants of CN3685 or CN1795.
This second ECL loop is predicted by computer modeling to include 14 amino acids (extending from amino acids 111 to 124) in the VirS made by CN1795 but to be 19 amino acids long (extending from amino acids 110 to 128) in the VirS made by CN3685. We speculate that, relative to the VirS made by CN1795, the larger size of its ECL2 allows the CN3685 VirS to functionally dock and accommodate the larger 8R (as well as the likely natural 5R) for signaling. Our results with ECL2 peptides (discussed in further detail below) implicated the Asp residue present at residue 117 in the CN1795 VirS as being important for SP binding. Detailed analysis of this loop should be performed in the future, since it represents a potential therapeutic target.
The VirS variant made by CN3685 is unusual. GenBank analysis of deduced VirS sequences in 53 other strains representing many C. perfringens types and origins indicated (data not shown) that only 2 of 53 strains resemble CN3685 in possessing a 6-amino-acid insertion at amino acid 152. This analysis also revealed that the KINSLVNVSELLGK ECL2 sequence of the CN3685 and CN1795 VirS was reasonably conserved among all 53 strains, with only 3 strains having 4 mutations and 26 strains having a single mutation. More common ECL2 sequence variations in these 26 strains were 16 strains producing an ECL2 with an S (versus a V) at position 6 and 10 strains producing an ECL2 with a D (versus an E) at position 10. Whether those ECL2 variations affect SP signaling should be investigated in the future.
Those findings also extend understanding of the VirS structure versus function relationship. Consistent with the modeling shown in Fig. 2, previous modeling work had also reported that the VirS sensor histidine kinase likely contains seven transmembrane (TM) domains and a C-terminal tail (13). That previous study (13) using random mutagenesis did not identify a putative SP binding site. However, it did determine that the predicted TM4 domain of the VirS N-terminal region is important for PFO production, which is regulated by both the VirS/R TCRS and the Agr-like QS system (9,16,23,24,34). Since modeling predicts that TM4 is adjacent to the second ECL, our new findings may suggest that the TM4 domain perturbations that altered PFO production in that previous study (13) affected the second ECL structure and thus altered SP binding and signaling. The previous random mutagenesis study (13) also identified regions of the C-terminal tail of VirS that are important for signaling, including the H255 residue that is an autophosphorylation site and G and N boxes that are involved in ATP binding and catalysis. The current results now add to that previous VirS structure versus function information by implicating the second ECL in SP binding and signaling.
Considerable progress was achieved toward the main goal of the study, i.e., testing the hypothesis that VirS is an SP receptor in C. perfringens. Several lines of evidence were obtained that now directly support VirS as an SP receptor. First, we exploited the previous observation (27) that both 5R and 8R can signal agrB null mutants of CN3685 while 5R (but not 8R) signals agrB null mutants of CN1795, which suggested that there may be differences between the VirS of CN3685 versus CN1795 if VirS is a major SP receptor. As discussed earlier, sequencing confirmed significant differences between the second ECL of VirS of CN3685 versus CN1795; therefore, we reasoned that, if VirS is an SP receptor, then swapping expression of these virS variants between virR/S and agrB/D double-null mutants of CN3685 versus those of CN1795 should change their sensitivity to 8R signaling. This hypothesis was supported when the CN1795 double mutant complemented to express the CN3685 VirS variant was shown to increase its CPB production in the presence of 8R, while the CN3685 double mutant complemented to express the CN1795 VirS variant lost the ability to increase CPB production in the presence of 8R. There was some CPB production by these complementing strains even in the absence of 8R; as noted previously, this effect is likely due to overexpression of VirR from the multicopy plasmid used for complementation (13).
A second approach to assess whether VirS is an SP receptor used a pulldown approach to evaluate directly if VirS can physically bind SP. Because attempts to prepare a VirS antibody were unsuccessful, this pulldown experiment instead used streptavidin beads pretreated with biotin-labeled 5R, which retains signaling activity. When those beads were reacted with extracts from CN3685 expressing a His 6 -tagged VirS, Western blotting with an His 6 antibody detected specific pulldown of His 6 -tagged VirS using the biotin-labeled 5R beads. This result supported the ability of 5R to physically bind with VirS.
A third and final line of evidence supporting VirS as an SP receptor came from experiments using synthetic peptides corresponding to either the predicted ECL2 of VirS or to this ECL sequence with a single N to D substitution. The rationale behind this experiment was that if the second ECL of VirS is important for SP binding, as suggested by the VirS sequencing results described earlier, then preincubating this ECL2 peptide with 5R may cause ECL2:5R binding and thus inhibit subsequent Agr-like QS signaling by agrB mutants. This hypothesis was verified when preincubation of 5R with KIGK, the peptide corresponding to the wild-type ECL2 sequence of the CN1795 VirS, inhibited the ability of 5R to signal the agrB mutants of either CN3685 or CN1795, as evidenced by a reduction in 5R-induced CPB production. Furthermore, the presence of the KIGK peptide in wild-type CN3685 or CN1795 cultures also reduced CPB production. These effects were specific, since KIGK_D, which is the peptide corresponding to a single N to D substitution in KIGK, did not reduce the 5R-induced increase in CPB production by the same wild-type strains or their agrB mutants. In addition to supporting SP binding to VirS, these results (as mentioned earlier) also implicate the second ECL of VirS in this binding. Collectively, these ECL2 peptide results not only further support VirS:SP binding being a mechanistic basis for interactions between the Agr-like QS system and VirS/R TCRS during C. perfringens pathogenesis but could also be instructive for developing peptide therapeutics to inhibit QS signaling and reduce toxin production to control type C or type B diseases in which CPB is, respectively, proven or likely to be important for virulence.
The current findings coupling the Agr-like QS system and VirS/R TCRS also hold broader relevance for understanding C. perfringens pathogenesis beyond diseases caused by type B or C strains. Both the Agr-like QS system and VirS/R TCRS positively regulate production of CPA and PFO (9,16,23,24), which are the toxins causing gas gangrene (35), and the importance of the Agr-like QS system for gas gangrene has been directly demonstrated (36). Similarly, both the Agr-like QS system and VirS/R control NetB toxin production, which is required for type G strains to cause avian necrotic enteritis, and the virulence of NetB-producing type G strains requires the Agr-like QS (17,26). Therefore, coupling the Agr-like QS system and VirS/R TCRS provides C. perfringens with a single, versatile regulatory pathway for controlling production of several important toxins produced during vegetative growth. However, the Agr-like QS and VirS/R systems do not universally regulate all C. perfringens toxin production during vegetative growth. For example, agrB null mutants of type B strains CN1793 and CN1795 still produce wild-type levels of ETX (30). Regulatory control of ETX production is poorly understood and requires further study. Coupling of the Agr-like QS and VirS/R TCRS may also be important for regulating production of toxins produced during C. perfringens sporulation, since the Agr-like QS system is an important positive regulator of sporulation and C. perfringens enterotoxin (CPE) production by type F strains (37). However, it has not yet been assessed whether this regulation involves VirS/R.
While this study provides compelling support for VirS as an SP receptor, it remains possible that C. perfringens possesses one or more additional receptors for the SP of the Agr-like QS, particularly since this bacterium possesses ϳ20 different TCRS (38). Supporting that possibility, silver staining of gels in the pulldown experiments revealed an ϳ75-kDa band that was not reactive with His 6 antibody and that had a larger molecular mass (ϳ75 kDa) than VirS (ϳ50 kDa) (data not shown). Whether that band reflects nonspecific binding or a second SP receptor will require further study.
In the S. aureus Agr system, SP signaling upregulates expression of the agr operon encoding itself and the AgrC receptor in a positive feedback loop (7,8). Therefore, having obtained strong evidence supporting VirS as an SP receptor, we used qRT-PCR to examine whether the VirR/S TCRS and the Agr-like QS system upregulate expression of genes encoding each other. Results indicated that, in both CN3685 and CN1795 backgrounds, the presence of a functional virS gene results in significantly higher expression of the agrD gene and vice versa. This positive feedback effect was not attributable to growth differences between the mutants versus their wild-type parents. These observations suggest a model where small amounts of AgrD present early during growth may help to upregulate VirS production, which then results in the availability of more receptors to further amplify AgrD production and Agr-like QS signaling later during the growth cycle. This effect could contribute to toxin production and C. perfringens pathogenicity.

MATERIALS AND METHODS
Bacteria, media, and reagents. C. perfringens wild-type, null mutant, and complementing strains used in this study are listed in Table 3. All isolates were stored in cooked meat medium (CMM) at Ϫ20°C. Fluid thioglycolate medium (FTG; Difco Laboratories), TY broth (3% tryptic soy broth [Becton, Dickinson], 1% yeast extract [Becton, Dickinson], and 0.1% sodium thioglycolate [Sigma-Aldrich]), and TGY broth (TY broth supplemented with 2% glucose [Sigma-Aldrich]) were used for broth cultures. After inoculation, brain heart infusion (BHI) agar (Research Products International) plates containing 15 g · ml Ϫ1 chloramphenicol (Sigma-Aldrich) were incubated at 37°C under anaerobic growth conditions in GasPak jars to screen the knockout mutants constructed in this study. A CN3685::agrB null mutant named BMJV10, a CN1795::agrB null mutant, and complementing strains of both mutants had been previously constructed and characterized (25,30). All chemical reagents used in this study were purchased from Fisher Scientific, Sigma-Aldrich, or Bio-Rad Laboratories. Construction of CN1795 and CN3685 virS single-null mutants and complemented strains or double-null mutants of those strains with inactivated virS and agrB genes. The virS gene was disrupted in CN1795 or CN3685 to generate single-null mutant strains that do not produce VirS. The virS gene was also inactivated in existing CN1795::agrB or CN3685::agrB strains (25,30) to create double-null mutants with inactivated virS and agrB genes. Disruption of the virS gene in these mutants was achieved by specifically inserting, in the antisense orientation, a group II intron (ϳ900 bp) into the virS gene, generating virS single-null mutants or virS agrB double-null mutants of CN1795 or CN3685. For this purpose, the intron donor plasmid pJIR750virSi, which carries a virS-targeted intron, was electroporated into CN1795, CN3685, CN1795::agrB, or CN3685::agrB. Transformants were selected by plating onto BHI agar plates containing 15 g · ml Ϫ1 of chloramphenicol, followed by overnight anaerobic growth in a GasPak jar. Colony PCR was carried out for screening using internal virS primers virSKOF and virSKOR (Table 4), which amplify a PCR product of ϳ370 bp using DNA from a wild-type strain but amplify an ϳ1,300-bp product using DNA from virS null mutants due to the insertion of an ϳ900 bp intron. Each virS gene null mutant was subcultured daily in FTG medium over 10 days to cure the intron-carrying plasmid, creating a CN1795 virS null mutant (CN1795::virS), a CN3685 virS null mutant (CN3685::virS), a CN1795 virS agrB double-null mutant (CN1795DKO), and a CN3685 virS agrB double-null mutant (CN3685DKO). Each mutant was then further characterized by PCR, RT-PCR, and Southern blotting analyses, as described below.
VirS complementing strains of the single and double mutants were prepared by transformation with pJIR750virR/S1795comp or pJIR750virR/S3685comp. Transformants were selected on chloramphenicol, as described earlier.
Measurement of C. perfringens growth. For analysis of C. perfringens vegetative growth, a 0.2-ml aliquot of an overnight FTG culture of a wild-type or null mutant strain was inoculated into 10 ml of TY medium. The cultures were incubated at 37°C; thereafter, at 0-, 1-, 3-, 5-, 8-, and 24-h culture times, 1 ml of each culture was removed for measurement of optical density at 600 nm (OD 600 ) using a Bio-Rad Smart spectrometer.
For the two wild-type strains, another 1-ml aliquot of culture was removed and centrifuged at 15,000 rpm for 3 min. Equal volumes of each resultant culture supernatant were then mixed with 5ϫ SDS loading buffer and boiled for 5 min. An aliquot (30 l) of each boiled sample was electrophoresed on a 10% SDS-PAGE gel and then subjected to a CPB Western blot analysis (see "Western blot analyses of CPB production"). Using the same cultures, total RNA was isolated from the pellets at each time point, and qRT-PCR were performed as described in "C. perfringens RNA isolation, RT-PCR, and qRT-PCR analyses."