The Penicillin-Binding Protein PbpP Is a Sensor of β-Lactams and Is Required for Activation of the Extracytoplasmic Function σ Factor σP in Bacillus thuringiensis

The bacterial cell envelope is the target for numerous antibiotics. Many antibiotics target the synthesis of peptidoglycan, which is a central metabolic pathway essential for bacterial survival.

Penicillin-binding proteins (PBPs) are some of the enzymes responsible for peptidoglycan synthesis. In the cytosol, dimers of NAG-NAM with pentapeptide side chains are synthesized and then flipped outside the cell membrane. These dimers are added to the growing peptidoglycan polymer by transglycosylation, which results in strands of repeating NAG-NAM subunits. These strands are cross-linked by transpeptidation of their pentapeptide side chains in a reaction carried out by PBPs. There are two types of high-molecular-weight PBPs. Type a PBPs have both transglycosylase activity and transpeptidase activity. Type b PBPs have only transpeptidase activity but work in concert with monofunctional SEDS (shape, elongation, division, sporulation) family transglycosylases to synthesize peptidoglycan (2,4). The activities of type a PBPs and type b PBPs are required for cell viability (5)(6)(7).
b-Lactam and cephalosporin antibiotics inhibit peptidoglycan synthesis by forming a covalent bond with the transpeptidase active-site serine of PBPs (5,8,9). This inhibition prevents cross-linking of the peptide side chains, which results in peptidoglycan instability and lysis during cell growth (10). Resistance to b-lactams and cephalosporins is a growing problem that complicates the treatment of bacterial infections. Resistance to b-lactams is usually due to the secretion of b-lactamases, which destroy the antibiotic by cleaving the b-lactam ring, or mutations that lead to modification of the transpeptidase active sites of PBPs and prevent b-lactam binding (11,12).
In response to stresses like antimicrobial peptides or antibiotics, many bacteria utilize alternative s factors to regulate subsets of genes required for the stress response. The extracytoplasmic function (ECF) s factor family is the largest and most diverse group of alternative s factors and represents the "third pillar" of bacterial signal transduction (13)(14)(15). ECF s factors are part of the s 70 family but contain only region 2 and region 4.2 of s 70 . These regions bind to the 210 and 235 regions of promoters, respectively (13,16). Many ECF s factors are held inactive by anti-s factors (13,17,18). The activation of these ECF s factors requires release from their cognate anti-s factors to allow the transcription of specific stress response genes.
A recent study identified .150 different families of ECF s factors (15). The roles of the vast majority of these s factors remain poorly understood; however, of the studied ECF s factors, the mechanisms of ECF s factor activation are diverse (18)(19)(20). One common mechanism known to control ECF s factor activation is the proteolytic destruction of the anti-s factor (18,21). Among those ECF s factor systems that use proteolytic destruction of the anti-s factor, the mechanisms controlling the initiation of this proteolytic cascade are diverse (21). In Escherichia coli, the activation of s E is controlled by the binding of misfolded outer membrane proteins to the site 1 protease DegS and lipopolysaccharide (LPS) binding to RseB (a negative regulator of s E activation) (22)(23)(24)(25). These binding events lead to the cleavage of the anti-s factor RseA at site 1 by DegS (26). In Bacillus subtilis, the activation of s V by lysozyme is controlled by the direct binding of the anti-s factor RsiV to lysozyme and then cleavage of RsiV at site 1 by signal peptidase (27)(28)(29)(30)(31)(32).
In Bacillus anthracis, Bacillus cereus, and Bacillus thuringiensis, resistance to penicillin and other b-lactam antibiotics is dependent upon s P , an ECF s factor (33,34). s P was originally classified as a member of the ECF01 group of ECF s factors but was recently reclassified to the ECF265 group, the members of which are primarily found in Firmicutes (14,15). Approximately 50% of ECF265 s factors are associated with an antisigma factor that contains a single transmembrane helix (15). Little is known about how the activity of the ECF265 s group is controlled, and s P could represent a model to understand the activation of this subclass of ECF s factors.
s P activity is inhibited by the anti-s factor RsiP, which contains a single transmembrane helix. The activation of s P results in the expression of at least two genes that encode b-lactamases and are involved in resistance to penicillin, ampicillin, and other b-lactam antibiotics. s P also activates the expression of its operon, thus controlling the expression of sigP and rsiP (33,34). We previously demonstrated that s P is activated in the presence of a subset of b-lactams, ampicillin, methicillin, cefoxitin, cephalothin, and cefmetazole, but not other cell envelope stresses (34). We also identified a subset of b-lactams that do not activate s P : piperacillin, cefsulodin, and cefoperazone (34). In response to the activating b-lactams, RsiP is destroyed by a cascade of proteases, resulting in s P activation (34). An unidentified site 1 protease initiates the proteolytic cascade by cleaving RsiP at site 1, which is then followed by cleavage at site 2 by RasP, the highly conserved site 2 protease (34) (Fig. 1B). Here, we demonstrate that b-lactam activation of s P is dependent on the PBP HD73_3488 (also known as HD73_RS17405), which we have named PbpP. Our data indicate that PbpP is required for site 1 cleavage of RsiP in response to b-lactams, but PbpP is likely not the site 1 protease. Our data suggest that PbpP likely functions as a sensor of b-lactams by directly binding b-lactams and triggering s P activation by promoting site 1 cleavage of RsiP.

RESULTS
PbpP is required for r P activation. Bacillus thuringiensis, B. cereus, and B. anthracis contain two open reading frames in the sigP region that encode predicted penicillinbinding proteins (PBPs). In Bacillus thuringiensis subsp. kurstaki HD73, these PBPs are called pbpP (HD73_3488) and bt3491 (HD73_3491). We also identified a third open reading frame that appears to be found only in Bacillus thuringiensis subsp. kurstaki HD73, called bt3487 (HD73_3487) (Fig. 1A). Although they are not located in the same operon as sigP and rsiP, we hypothesized that they may play a role in the response of s P to b-lactams because PBPs have been well characterized as targets of b-lactam antibiotics (9,35). Additionally, genes involved in the same signaling system are often located in the neighboring regions. To determine if BT3487, PbpP, and BT3491 were required for the response of s P to b-lactams, we generated strains with in-frame deletions of each of the genes and measured the effect on ampicillin resistance. We found In the presence of b-lactams, RsiP is sequentially cleaved by an unknown site 1 protease (S1P) (dark blue) and RasP (light blue) (34). (C) PbpP is required for the activation of s P . All strains contain the reporter P sigP -lacZ. The relevant genotypes of the tested strains included WT (THE2549), DsigP-rsiP (EBT232), and DpbpP (EBT151). Cells were grown to mid-log phase (OD of 1.0 to 1.4), washed, and resuspended in LB medium and LB medium plus cefoxitin (Cef) (0.02 to 2 mg/ml). b-Galactosidase (b-Gal) activities were calculated as described in Materials and Methods. Experiments were performed in technical and biological triplicate, and standard deviations are represented by error bars. a.u., arbitrary units. (D) PbpP is required for cefoxitin-induced degradation of RsiP. All strains contain the plasmid pBT13 (P tet -gfp-rsiP) and the following relevant genotypes: WT (THE360), DpbpP (EBT512), and DrasP (EBT366). The strains were grown to mid-log phase at 37°C in ATc (100 ng/ml), concentrated, and resuspended in 100 ml of LB medium or LB medium with cefoxitin (5 mg/ml) for 1 h. Immunoblotting was performed using anti-GFP antisera. Streptavidin IR680LT was used to detect AccB (HD73_4487), which served as a loading control (51,52). A color blot showing both anti-GFP and streptavidin on a single gel is shown in Fig. S2 in the supplemental material. Numbers at the left indicate molecular masses of the ladder in kilodaltons.
s P Activation Requires a Penicillin-Binding Protein ® that the deletion of pbpP led to a dramatic decrease in the ampicillin MIC similar to that of a DsigP mutant (Table 1) (33,34). In contrast, strains with deletions in bt3487 and bt3491 had no effect on ampicillin resistance (not shown). We also determined that a DpbpP mutant is more sensitive to cefoxitin and cefmetazole than the wild type (WT) ( Table 1).
We noted that a DsigP-rsiP mutant is more sensitive to b-lactams than a DpbpP mutant. We hypothesized that a DpbpP mutant may block s P activation in response to b-lactams but retains a basal level of s P activation that allows a low level of resistance to b-lactams. To monitor s P activity, we took advantage of the fact that s P is required for the transcription of its promoter (P sigP ); thus, we inserted a P sigP -lacZ promoter fusion into the thrC locus (33,34). To determine if PbpP played a role in s P activation, we tested the effect of a pbpP deletion on s P activity by monitoring P sigP -lacZ expression. Interestingly, we did not observe activation of s P in the DpbpP mutant in the presence of cefoxitin (see Fig. S1A in the supplemental material). We complemented the DpbpP mutant with pbpP 1 on a plasmid under the control of its native promoter. We found that s P was activated in the presence of cefoxitin to an extent similar to that observed for the WT (Fig. S1A). To reinforce our finding that DpbpP results in the loss of s P activation, we conducted b-galactosidase assays to quantify the effect on s P activation. As previously reported, P sigP -lacZ expression is induced in a dose-dependent manner in response to increased cefoxitin concentrations in the WT (Fig. 1C) (34). Consistent with previous observations, we did not observe induction of P sigP -lacZ in the DsigP-rsiP mutant because s P is required for transcription from P sigP (34). We found that the deletion of pbpP resulted in the loss of P sigP -lacZ expression at every concentration tested (Fig. 1C). Taken together, our data suggest that PbpP is required for the activation of s P , thereby altering the transcription of the s P regulon and b-lactam resistance.
PbpP is required for site 1 cleavage of RsiP. Because our data suggest that PbpP is required for s P activation, we hypothesized that PbpP is required for RsiP degradation. To test this, we compared the effects of cefoxitin on the degradation of green fluorescent protein (GFP)-RsiP in WT, DpbpP, and DrasP mutant strains. We previously showed that GFP-RsiP is functional and localized to the membrane (34). We found that the levels of full-length GFP-RsiP decreased in the WT in the presence of cefoxitin ( Fig. 1D) (34). When a DrasP mutant, which lacks the site 2 protease, was incubated with cefoxitin, we observed a decrease in full-length GFP-RsiP and the buildup of an intermediate GFP-RsiP fragment, indicating the loss of site 2 cleavage (Fig. 1D) (34). This GFP-RsiP fragment is approximately the predicted size for a site 1 protease cleavage product. In contrast, we found that full-length GFP-RsiP levels did not decrease in the DpbpP mutant when grown in the presence of cefoxitin (Fig. 1D). This suggests that PbpP is required for site 1 cleavage of RsiP and, thus, s P activation.
PbpP is a penicillin-binding protein. A defining feature of PBPs is the ability to covalently bind b-lactams (9,36). We sought to determine if PbpP has the capacity to bind b-lactams. We tested if PbpP could bind Bocillin-FL (Boc-FL), a fluorescent b-lactam consisting of penicillin V and BODIPY FL dye (37). We found that Bocillin-FL was degraded when s P was activated (Fig. S3B). In a DsigP-rsiP mutant, we found that Bocillin-FL was not degraded, suggesting that s P -regulated b-lactamases are likely responsible for Bocillin-FL degradation (Fig. S3B). To perform Bocillin-FL labeling experiments, we expressed pbpP from an isopropyl-b-D-thiogalactopyranoside (IPTG)- inducible promoter in a DsigP-rsiP mutant. We labeled cells with Bocillin-FL and blotted them with anti-PbpP antisera (37). We observed a fluorescent band at approximately 66 kDa with both Bocillin-FL and anti-PbpP antisera. This band was the predicted size of PbpP; it increased in intensity with increasing IPTG concentrations and was not observed in the empty vector (EV) control ( Fig. 2A and Fig. S3A). This demonstrates that PbpP binds b-lactams. We also noted that the lack of a fluorescent band corresponding to PbpP in the EV suggests that the levels of PbpP in wild-type cells are not high enough to be detected by Bocillin-FL labeling. All PBPs have an active-site serine that is acylated by b-lactams (36). We identified serine 301 (S301) as the likely active-site residue required for transpeptidation based on homology to other PBPs. To determine if S301 is the active-site serine, we mutated it to an alanine by site-directed mutagenesis and expressed pbpP S301A under the control of an IPTG-inducible promoter. In the strain producing PbpP S301A , the 66-kDa band was lost when imaging for Bocillin-FL ( Fig. 2A and Fig. S3A). However, immunoblotting using anti-PbpP antisera detected a 66-kDa band corresponding to PbpP S301A , which is produced at levels similar to those of the WT protein ( Fig. 2A; Fig. S3A). Thus, PbpP S301A cannot covalently bind Bocillin-FL. This suggests that PbpP is a penicillin-binding protein, S301 is required for binding b-lactams, and S301 is likely the active-site serine. Cells were grown to mid-log phase with various concentrations of IPTG. Cells were concentrated, resuspended, and incubated with Bocillin-FL (50 mg/ml). The proteins were then separated by SDS-PAGE, immunoblotting was performed using anti-PbpP antisera and Bocillin-FL, and streptavidin IR680LT was used to detect HD73_4231 (PycA homolog), which served as a loading control (51,52). Figure S3A in the supplemental material is the color blot showing anti-PbpP antisera, Bocillin-FL, and streptavidin in a single image. (B) pbpP S301A phenocopies DpbpP. All strains contain the reporter P sigP -lacZ and were of the following genotypes: WT (THE2549), DsigP-rsiP (EBT232), DpbpP (EBT151), DpbpP ICEBs1::pbpP 1 (EBT773), and DpbpP ICEBs1::pbpP S301A (EBT772). The strains were grown to mid-log phase and incubated without or with cefoxitin for 1 h, and b-galactosidase activity was measured. Experiments were performed in technical and biological triplicate, and standard deviations are represented by error bars.
s P Activation Requires a Penicillin-Binding Protein ® b-Lactam binding by PbpP is required for b-lactam-dependent activation of r P . We sought to determine if b-lactam binding to PbpP was required for s P activation using a PbpP S301A active-site mutant. We complemented the DpbpP mutant with pbpP 1 and pbp S301A under the control of their native promoter in a single copy by integrating constructs at the B. subtilis integrative conjugative element (ICEBs1) site in the B. thuringiensis chromosome (38). We found that PbpP 1 restored P sigP -lacZ expression in the presence of cefoxitin (Fig. 2B). In contrast, when we complemented the strain with pbpP S301A , we observed no increase in P sigP -lacZ expression in the presence of cefoxitin (Fig. 2B). These data suggest that binding of PbpP to b-lactams is required for b-lactams to activate s P .
Overexpression of pbpP and pbpP S301A leads to activation of r P . We noted that the basal level of P sigP -lacZ expression was higher in the strains complemented with pbpP 1 and pbp S301A integrated at ICEBs1 than in WT B. thuringiensis (Fig. 2B). We reasoned that this might be due to higher basal levels of expression of pbpP and pbpP S301A at the ICEBs1 site. Thus, we sought to determine the effect of the overexpression of pbpP 1 and pbpP S301A on s P activation. We expressed pbpP 1 or pbpP S301A from a tetracycline-inducible promoter on a multicopy plasmid (34,39). We observed that increased expression of pbpP 1 or pbpP S301A leads to a dose-dependent increase in the expression of P sigP -lacZ, in the absence of b-lactams (Fig. 3A). We also found that the addition of cefoxitin led to a further increase in P sigP -lacZ expression when pbpP 1 was overexpressed (Fig. S4B). We noted increased basal levels of P sigP -lacZ expression in the absence of anhydrotetracycline (ATc) and concluded that this is likely due to leaky expression of P tet -pbpP and P tet -pbpP S301A (Fig. 3A and Fig. S4B). These data suggest that the overexpression of both the WT and the active-site mutant (S301A) can activate s P even in the absence of b-lactams. We interpret this to mean that the requirement for b-lactam binding to PbpP can be compensated for by increased levels of PbpP; however, b-lactam binding to PbpP further enhances s P activation (Fig. S4B). The activation of s P in WT cells is likely not due to b-lactam-induced pbpP transcription as the expression of pbpP is not induced by b-lactams ( Fig. S1B and C). The pbpP S301A mutant also fails to induce s P activation when expressed under the control of its native promoter, further suggesting that pbpP is not induced by b-lactams (Fig. 2B).
Since the loss of PbpP results in little to no degradation of RsiP in the presence of b-lactams, we tested if the increased expression of pbpP leads to the degradation of Overexpression of pbpP and pbpP S301A results in activation of s P . All strains contain the reporter P sigP -lacZ plus the following relevant genotypes: DpbpP/EV (EBT344), DpbpP/P tet -pbpP 1 (EBT327), and DpbpP/P tet -pbpP S301A (EBT1145). The cultures were grown to mid-log phase and incubated with anhydrotetracycline (ATc), and b-galactosidase activity was measured. This experiment was done in technical and biological triplicate, and standard deviations are represented by error bars. (B) Overexpression of pbpP causes degradation of RsiP. All strains harbor IPTG-inducible gfp-rsiP (P IPTG -gfp-rsiP) integrated at the ICEBs1 site (38) plus either the empty vector (EV) (pAH9) (EBT744), P tet -pbpP 1 (EBT742), or P tet -pbpP S301A (EBT1144). Strains were grown to mid-log phase with 1 mM IPTG and increasing concentrations of ATc. The cells were concentrated, resuspended in sample buffer, and separated by SDS-PAGE. The immunoblot was probed with anti-GFP antisera. Streptavidin IR680LT was used to detect HD73_4231 (PycA homolog), which served as a loading control (51,52). A color blot showing both anti-GFP and streptavidin on a single gel is shown in Fig. S4A in the supplemental material. Nauta et al.

®
RsiP in the absence of b-lactams. We introduced P tet -pbpP 1 or P tet -pbpP S301A into a strain containing IPTG-inducible gfp-rsiP. We found that the overexpression of PbpP and PbpP S301A leads to decreases in full-length GFP-RsiP levels, suggesting that PbpP can induce RsiP degradation and, thus, s P activation (Fig. 3B). This suggests that PbpP controls s P activation by controlling RsiP degradation.
PbpP is likely not the site 1 protease for RsiP. The site 1 protease required for initiating RsiP degradation has not yet been identified. Since PbpP is required for site 1 cleavage of RsiP, the possibility exists that PbpP is the site 1 protease. We sought to determine if basal-level site 1 cleavage occurred in the absence of pbpP, which would suggest that another protein can cleave RsiP. Since site 2 cleavage is rapid (34), we expressed gfp-rsiP in a DpbpP DrasP double mutant, which should allow the buildup of any GFP-RsiP site 1 cleavage product. We observed the accumulation of a band corresponding to a GFP-RsiP fragment in the DrasP mutant in the absence of cefoxitin, and the intensity of this band increased in the presence of cefoxitin (Fig. 4A). We observed . The strains were grown to mid-log phase with IPTG and incubated without or with cefoxitin (5 or 50 mg/ml). The samples were analyzed by immunoblotting using anti-GFP antisera. Streptavidin IR680LT was used to detect AccB (HD73_4487), which served as a loading control (51,52). A color blot showing both anti-GFP and streptavidin on a single gel is shown in Fig. S5A in the supplemental material. (B) PbpP is produced in B. subtilis. All strains contained amyE::P IPTG -gfp-rsiP with the relevant genotypes WT (CDE3147) and thrC::P xyl -pbpP (EBT756) and were grown to mid-log phase with IPTG at 0.01 mM and increasing concentrations of xylose (0.125 to 2% xylose). At mid-log phase, 1-ml aliquots were concentrated, washed, and resuspended in Bocillin-FL (50 mg/ml) for 30 min at RT. A color blot showing both Bocillin-FL and the ladder on a single gel is shown in Fig. S5B. (C) Samples from panel B were probed with anti-GFP antisera to detect GFP-RsiP, and streptavidin IR680LT was used to detect the PycA homolog, which served as a loading control (51,52). A color blot showing both anti-GFP and streptavidin on a single gel is shown in Fig. S5C. s P Activation Requires a Penicillin-Binding Protein ® the same band in the DpbpP DrasP mutant; however, the band did not increase in the presence of cefoxitin. We concluded that in a DpbpP DrasP mutant, there is a basal level of site 1 cleavage of RsiP occurring in the presence and absence of cefoxitin (Fig. 4A). This suggests that site 1 cleavage can occur in the absence of PbpP, but it is not b-lactam inducible. Presumably, in this strain, the unidentified site 1 protease still retains its basal level of activity but cannot be further activated in the presence of cefoxitin due to the absence of PbpP.
To test if PbpP is sufficient for site 1 cleavage of RsiP, we introduced IPTG-inducible gfp-rsiP into the Bacillus subtilis chromosome (which does not encode a homolog of sigP or rsiP) and expressed pbpP using a xylose-inducible promoter. We grew the cells in the presence of 0.01 mM IPTG and increasing concentrations of xylose. We asked if PbpP was expressed and presumably properly folded by labeling with the fluorescent b-lactam Bocillin-FL. We observed a fluorescent band corresponding to PbpP that increased in intensity with increasing concentrations of xylose (Fig. 4B). We also monitored GFP-RsiP levels by performing immunoblot analysis using anti-GFP antisera. We did not observe degradation or a decrease in RsiP levels even at the highest levels of PbpP, indicating that PbpP is not sufficient for RsiP degradation in B. subtilis (Fig. 4C). Taken together, these data lead us to conclude that PbpP is not the site 1 protease but is required for sensing of b-lactams in B. thuringiensis.
Affinities of b-lactams for PbpP do not correlate with their ability to activate r P . Since PbpP is likely not acting as the site 1 protease, we hypothesized that PbpP functions as a sensor that binds b-lactams and subsequentially activates s P . Therefore, we hypothesized that the reason why some b-lactams do not activate s P is that they have a lower affinity for PbpP. To test this hypothesis, we determined the affinity of PbpP for eight different b-lactams by modifying a Bocillin-FL inhibition experiment previously described by Kocaoglu and colleagues (40). We calculated the 50% inhibitory concentration (IC 50 ) (the concentration of b-lactam at which 50% of Bocillin-FL labeling of PbpP is inhibited) to determine the binding affinity of different b-lactams. We found that while the b-lactams had different IC 50 s for PbpP, the differences did not correlate with the ability of the b-lactams to activate s P (Fig. 5A and B). For example, we found that some of the nonactivating b-lactams (cefoperazone and cefsulodin) had IC 50 s similar to those of activating b-lactams ( Fig. 5A and B). Thus, the disparity in the b-lactams' ability to activate s P is not simply due to the inability of PbpP to bind different b-lactams. These data also suggest that simple binding of any b-lactam to PbpP is not sufficient for s P activation.
Cefsulodin inhibits activation of r P by cefoxitin. We found that b-lactam binding to PbpP is not sufficient for s P activation because nonactivating b-lactams covalently bind PbpP with affinities similar to those of the activating b-lactams (i.e., cefsulodin and ampicillin have nearly identical binding affinities for PbpP). We hypothesize that the b-lactams that activate s P induce a conformational change in PbpP that permits a protein-protein interaction. If this hypothesis were true, the b-lactams that do not activate s P would be able to inhibit the activation of s P by occupying the PbpP active site. To test this, we pretreated cells with cefsulodin (a nonactivator of s P ) and then added cefoxitin (an activator of s P ). We found that cefsulodin inhibited the activation of s P by cefoxitin in a dose-dependent manner (Fig. 6). We also show that pretreatment with cefmetazole (an activator of s P ) does not inhibit activation (Fig. 6). Therefore, nonactivating b-lactams inhibit s P activation presumably by occupying the active site of PbpP and preventing activating b-lactams from binding PbpP and activating s P (Fig. 7).

DISCUSSION
Our data argue that PbpP is a sensor for b-lactams that is required for s P activation by indirectly promoting the degradation of RsiP (Fig. 7). This is supported by our observation that the loss of the penicillin-binding protein PbpP blocks s P activation and RsiP degradation. Our data indicate that the binding of a subset of b-lactams to PbpP results in s P activation. However, inhibition of PbpP transpeptidase activity by b-lactams is not the signal that activates s P since the transpeptidase mutant PbpP S301A does not activate s P . Interestingly, the overexpression of PbpP and PbpP S301A activates s P even in the absence of b-lactams; however, PbpP is not the site 1 protease. Together, these results argue that PbpP is a sensor of b-lactams and controls s P activation.
PbpP is required for r P activation. The principal finding of this work is the demonstration that PbpP is required for the activation of s P in response to some b-lactams. Based on our findings, we propose the following working model for how PbpP functions as a sensor for b-lactams. In WT cells in the absence of stress, RsiP binds s P and inhibits s P activation (Fig. 7). When activating b-lactams are present, they bind the active-site serine of PbpP. The binding of the activating b-lactams results in a conformational change in PbpP that allows it to interact with a component of the s P system. This s P Activation Requires a Penicillin-Binding Protein ® interaction initiates regulated intramembrane proteolysis of RsiP and, thus, s P activation (Fig. 7). This model is supported by ample evidence: (i) deletion of pbpP blocks RsiP degradation and s P activation, (ii) mutants of PbpP unable to bind b-lactams fail to activate s P in response to b-lactams, and (iii) overexpression of PbpP or PbpP S301A leads to constitutive RsiP degradation and s P activation. Thus, PbpP plays an essential role in sensing the presence of inducing b-lactams and controlling s P activation.
PbpP is not the site 1 protease. It is possible that PbpP is a site 1 protease that initiates RsiP degradation; however, we think that it is unlikely. While PbpP is required for site 1 cleavage of RsiP in response to b-lactams, the totality of our data does not support PbpP as the site 1 protease. First, the overexpression of PbpP in B. subtilis does not induce the degradation of RsiP as it does in B. thuringiensis. Our data indicate that PbpP is functional, folded, and localized properly when expressed in B. subtilis since it can be labeled on whole cells by Bocillin-FL. This argues that PbpP is not sufficient for site 1 cleavage of RsiP and suggests that an unidentified B. thuringiensis protease is required. Second, in B. thuringiensis, we observed low-level site 1 cleavage of RsiP in the absence of PbpP. This argues that PbpP is not absolutely required for site 1 cleavage. If PbpP were a site 1 protease, there must be a second protease in B. thuringiensis that has low basal activity and cleaves RsiP at site 1 in the absence of PbpP. Finally, PbpP lacks any predicted protease domains. Future work will be required to identify the protease(s) required for site 1 cleavage of RsiP and, thus, s P activation.
PbpP is the b-lactam sensor for the r P system. We hypothesize that PbpP is the sensor of b-lactams for the s P system. In support of this, we found that s P is not activated in the DpbpP mutant or when pbpP S301A is expressed from the native P pbpP  promoter. However, the overproduction of either PbpP or PbpP S301A results in the activation of s P in the absence of b-lactams. This suggests that the overproduction of PbpP can compensate for b-lactam binding to PbpP to activate s P . Importantly, activation of s P is not due to inhibition of PbpP transpeptidase activity by b-lactams because PbpP S301A is catalytically inactive yet does not result in s P activation. This loss of s P activity is not due to an instability of PbpP S301A as it is produced at levels similar to those of WT PbpP. Activation of s P by b-lactams is not simply due to increased expression of pbpP since b-lactams do not induce pbpP expression. In addition, if increased expression of pbpP in response to b-lactams was responsible for s P activation, then we would have expected the pbpP S301A allele to induce s P activation when expressed under the control of the native P pbpP promoter. Taken together, these data suggest that PbpP interacts with some component of the signal transduction system.
In support of this hypothesis, we found that a subset of activating b-lactams bind PbpP with affinities similar to those of nonactivating b-lactams. We found that cefsulodin, a nonactivating b-lactam, can inhibit the activation of s P by an activating b-lactam, cefoxitin, presumably by competing for the active-site serine of PbpP. We hypothesize that nonactivating b-lactams do not induce the appropriate conformational change in PbpP to render it active and able to interact with its target. One obvious target for PbpP interaction is the anti-s itself. However, we did not observe an interaction between the extracellular domains of RsiP 76-275 and PbpP  in vitro using a copurification assay (see Fig. S8 in the supplemental material). This raises the possibility that PbpP interacts with another protein like the as-yet-unidentified site 1 protease. Alternatively, it may interact indirectly with RsiP or the site 1 protease via an unknown protein. Future work will need to determine what PbpP interactions drive RsiP degradation and, thus, s P activation.
Comparison of the BlaRI response to b-lactams to r P activation. While the identification of a PBP required for the activation of an ECF s factor is novel, there is precedence for a PBP transpeptidase-like domain functioning as a sensor of b-lactams. Found in diverse organisms, including Staphylococcus aureus and Bacillus licheniformis, BlaR1 (MecR1) contains an extracellular transpeptidase-like domain that senses b-lactams and a cytoplasmic protease domain. BlaR1 is a b-lactam sensor that directly binds b-lactams in its extracellular transpeptidase-like domain (41). The covalent bond formed with the b-lactam ring causes a conformational change in BlaR1 that activates the cytoplasmic protease domain (42). The protease domain cleaves the repressor of the b-lactamase operon, BlaI, thus activating the transcription of b-lactamase and increasing resistance to b-lactams (42). While the BlaIR system is clearly not synonymous with s P , it is worth noting that there is precedence for PBP domains that function as sensors of b-lactams.

MATERIALS AND METHODS
Media and growth conditions. All B. thuringiensis strains are isogenic derivatives of AW43, a derivative of B. thuringiensis subsp. kurstaki strain HD73 (43). All strains and genotypes can be found in Table 2. All B. thuringiensis strains were grown in or on LB media at 30°C unless otherwise specified. Liquid cultures of B. thuringiensis were grown with agitation in a roller drum. B. thuringiensis strains containing episomal plasmids were grown in LB medium containing chloramphenicol (Cam) (10 mg/ml; Ameresco) or erythromycin (Erm) plus lincomycin (Linc) (MLS) (1 mg/ml Erm [Ameresco] and 25 mg/ml Linc [Research Products International]). E. coli strains were grown at 37°C using LB-ampicillin (Amp) (100 mg/ml; Ameresco) or LB-Cam (10 mg/ml) medium. B. subtilis strains were grown on LB medium with antibiotics (Cam at 10 mg/ml, spectinomycin [Spec] at 100 mg/ml [Amresco], or Erm at 10 mg/ml). To screen for threonine auxotrophy, B. thuringiensis strains were patched onto minimal medium plates without or with threonine (50 mg/ml). The b-galactosidase chromogenic indicator 5-bromo-4-chloro-3-indolyl-b-D-galactopyranoside (X-Gal; Research Products International) was used at a concentration of 100 mg/ml. Anhydrotetracycline (ATc; Sigma) was used at a concentration of 100 ng/ml unless otherwise indicated. IPTG (Research Products International) and xylose (Acros) were used at the concentrations indicated in the figure legends. Additional b-lactams used in b-lactam-binding experiments were used at the concentrations indicated in the figure legends and were acquired from the following sources: cefsulodin, piperacillin, cefmetazole, and cefoxitin from Sigma-Aldrich; cephalothin from Chem-impex International Inc.; methicillin from Alfa Aesar; and cefoperazone from Toronto Research Chemical Inc.
Strain and plasmid construction. All plasmids are listed in Table 3 and Table S1 in the supplemental material, which includes information relevant to plasmid assembly. Plasmids were constructed by isothermal assembly (44). Regions of plasmids constructed using PCR were verified by DNA sequencing. s P Activation Requires a Penicillin-Binding Protein ® The oligonucleotide primers used in this work were synthesized by Integrated DNA Technologies (Coralville, IA) and are listed in Table S2. All plasmids were propagated using OmniMax 2-T1R as the cloning host and passaged through the nonmethylating E. coli strain INV110 before being transformed into a B. thuringiensis recipient strain.
To construct deletion mutants, we cloned 1 kb of DNA upstream and 1 kb downstream of the site of the desired deletion using primers listed in Table S2 into the temperature-sensitive pMAD plasmid (erythromycin resistant) between the BglII and EcoRI sites (45). Mutants were constructed by shifting temperatures as previously described (45).
B. subtilis ICEBs1 conjugation strains were constructed by transforming JAB932 as previously described (38). The resulting transformants or donor strains were grown in LB medium with D-alanine  Bocillin-FL labeling assay. Cultures grown overnight at 30°C were diluted 1:50 and grown to an OD of ;1.0. The cultures were aliquoted in 1-ml aliquots and pelleted at 8,000 rpm. The cells were washed twice in 500 ml of 1Â phosphate-buffered saline (PBS) and resuspended in either 50 ml of 50mg/ml Bocillin-FL (Thermo Fisher) or 50 ml of 10-fold dilutions of b-lactams (0.0005 to 5,000 mg/ml). The samples resuspended in b-lactams were incubated for 30 min at room temperature and then pelleted and resuspended in 50 ml or 50mg/ml Boc-FL for 15 min. After incubation in Boc-FL, all the samples were pelleted and resuspended in 200 ml sample buffer with 5% b-mercaptoethanol (bME). The samples were sonicated, heated, and electrophoresed on a 12% polyacrylamide gel. The gels were imaged on an Azure Sapphire system (AzureBiosystems) by excitation at 488 nm and detection at 518 nm. The Bocillin-FL labeling experiment was performed in biological triplicate for each antibiotic, and the Bocillin-FL intensity for the PbpP band was quantified on each gel. The average intensity was used to calculate the IC 50 using GraphPad Prism, with means and standard errors or deviations shown.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.