Reciprocal regulation of cephalosporin resistance in Enterococcus faecalis.

UNLABELLED: Antibiotic-resistant enterococci are major causes of hospital-acquired infections and therefore represent a serious public health problem. One well-known risk factor for the acquisition of hospital-acquired enterococcal infections is prior therapy with broad-spectrum cephalosporin antibiotics. Enterococci can proliferate in patients undergoing cephalosporin therapy due to intrinsic cephalosporin resistance, a characteristic of the genus Enterococcus. However, the molecular basis for cephalosporin resistance in E. faecalis has yet to be adequately elucidated. Previously we determined that a putative Ser/Thr kinase, IreK (formerly PrkC), is required for intrinsic cephalosporin resistance in E. faecalis. Here we show that kinase activity is required for cephalosporin resistance and, further, that resistance in E. faecalis is reciprocally regulated by IreK and IreP, a PP2C-type protein phosphatase encoded immediately upstream of IreK. Mutants of two divergent lineages of E. faecalis lacking IreP exhibit remarkable hyperresistance to cephalosporins but not to antibiotics targeting other cellular processes. Further genetic analyses indicate that hyperresistance of the IreP mutant is mediated by the IreK kinase. Additionally, competition experiments reveal that hyperresistant ΔireP mutants exhibit a substantial fitness defect in the absence of antibiotics, providing an evolutionary rationale for the use of a complex signaling system to control intrinsic cephalosporin resistance. These results support a model in which IreK and IreP act antagonistically via protein phosphorylation and dephosphorylation as part of a signal transduction circuit to regulate cellular adaptation to cephalosporin-induced stress. IMPORTANCE: As a major cause of hospital-acquired infections, antibiotic-resistant enterococci represent a serious public health problem. Enterococci are well-known to exhibit intrinsic resistance to broad-spectrum cephalosporin antibiotics, a trait that enables them to proliferate in patients undergoing cephalosporin therapy, thereby predisposing these patients to acquisition of an enterococcal infection. Thus, inhibition of enterococcal cephalosporin resistance could represent an effective new strategy to prevent the emergence of hospital-acquired enterococcal infections. At this time, however, the molecular basis for cephalosporin resistance in E. faecalis is poorly understood. Our results begin to unravel the details of a new phosphorylation-dependent signal transduction system that controls cephalosporin resistance in enterococci. Deeper understanding of the mechanism underlying cephalosporin resistance in E. faecalis may enable the development of new therapeutics designed to reduce the incidence of hospital-acquired enterococcal infections.

Introduction

The Gram-positive bacterium Enterococcus faecalis is a commensal inhabitant of the gastrointestinal tracts of a wide variety of insects and animals, including <emical">span class="Species">humans (1). However, antibiotic-resistant enterococci are also major causes of hospital-acquired infections (2) and therefore represent a serious public health problem. One well-known risk factor for the acquisition of enterococcal hospital-acquired infections is prior therapy with broad-spectrum cephalosporins (3), antibiotics that belong to the β-lactam family and interfere with cell wall biosynthesis by inhibiting the penicillin-binding proteins (PBPs) that cross-link peptidoglycan (PG). The prevailing model to explain the association of cephalosporin therapy with increased risk of enterococcal infection (reviewed in reference 3) invokes the observation that enterococci proliferate to achieve abnormally high densities in the gastrointestinal (GI) tract of patients during cephalosporin therapy (4), a situation that likely facilitates enterococcal dissemination to other sites and the subsequent emergence of infection. Proliferation of enterococci during cephalosporin therapy is possible because enterococci exhibit intrinsic resistance to broad-spectrum cephalosporins. This intrinsic cephalosporin resistance is a trait shared by essentially all isolates of E. faecalis, yet our understanding of the genetic and biochemical basis underlying cephalosporin resistance remains incomplete.

emical">Three genetic loci have thus far been reported to be critical for <emical">span class="Chemical">cephalosporin resistance in E. faecalis: (i) pbp5, which encodes a so-called “low-affinity” PBP that exhibits reduced affinity for cephalosporins and presumably retains the ability to synthesize cell wall despite the presence of cephalosporins in the environment (5, 6); (ii) a locus encoding a two-component signal transduction system (CroRS) which presumably regulates the expression of as-yet-unknown genes that promote cephalosporin resistance (7, 8); and (iii) a gene encoding a predicted Ser/Thr kinase, genetic deletion of which drastically reduces cephalosporin resistance but which has not yet formally been shown to possess kinase activity (9). In our original report (9), this kinase was referred to as PrkC—here we rename it IreK (for intrinsic resistance of enterococci), to reflect its critical role in intrinsic cephalosporin resistance.

emical">IreK belongs to a family of <emical">span class="Chemical">Ser/Thr protein kinases with a characteristic bipartite domain architecture. Kinases in this family share a presumably cytoplasmic kinase domain separated by a predicted transmembrane segment from a series of PASTA domains that are thought to bind PG or fragments thereof (10–12). Homologs of IreK with similar domain architecture are present in the genomes of nearly all low-GC Gram-positive bacteria (usually found in 1 copy per genome), and analyses of mutants lacking those kinases have revealed diverse functional roles for the kinases, including development of competence, regulation of intracellular nucleotide pools, virulence, control of hemolysin production, cell division, stationary-phase survival, germination of endospores, and modulation of antibiotic resistance (reviewed in reference 13).

Although the phenotypic consequences of genetic lesions in the kinases have been studied in some detail for numerous Gram-positive bacteria, less is known about mechanisms of kinase regulation. Extensive structural studies on the purified kinase domain of the mycobacterial <emical">span class="Chemical">IreK homolog (PknB) revealed a back-to-back homodimer (14, 15) and suggested a regulatory mechanism by which dimerization allosterically activates the kinase (16). Additionally, mass spectrometry studies have revealed that the purified kinase domains of two IreK homologs (mycobacterial PknB and Bacillus subtilis PrkC) can be autophosphorylated at multiple sites per monomer—in most cases, at threonine residues—including at several conserved sites in the kinase “activation loop” (14, 17–19). The activation loop is a short, centrally located segment of the kinase domain that is thought to undergo a conformational shift upon phosphorylation, leading to activation of the kinase. Substitution of the phosphorylatable residues with alanine in the kinase domains of PknB and PrkC substantially reduces kinase activity (17, 18), suggesting that autophosphorylation of the activation loop indeed enhances kinase activity for both PknB and PrkC. Thus far, the vast majority of studies probing the mechanisms of kinase regulation have been performed in vitro using purified kinase domains. Although it seems likely that these findings will translate into the in vivo setting, at present it is not clear to what extent these mechanisms contribute to kinase regulation in vivo.

In the genomes of most Gram-positive bacteria, encoded immediately upstream of the emical">IreK homolog is a protein <emical">span class="Chemical">Ser/Thr phosphatase of the PP2C family. The gene encoding the phosphatase typically overlaps slightly, and is cotranscribed with, the gene for the kinase (20–23), suggesting that the two gene products participate in related biological processes. In vitro, the purified phosphatases are usually capable of dephosphorylating proteins that have previously been phosphorylated by the cognate kinases (reviewed in reference 13), suggesting that a given kinase-phosphatase pair antagonistically regulates the level of substrate protein phosphorylation. The phosphatases can also dephosphorylate the cognate kinases themselves in vitro, but it has not been determined if this activity is relevant in vivo. Indeed, in vivo analysis of these signaling systems has largely been limited to phenotypic study of mutants lacking either the kinase, phosphatase, or both. These studies have yielded a conflicting picture in which mutants lacking a given kinase or the corresponding phosphatase in some cases exhibit opposing phenotypes—the expected outcome if they function antagonistically—but in other cases exhibit similar phenotypes (22–26). Genetic analyses by Osaki and coworkers suggest that the PP2C phosphatase in pneumococcus controls in vivo activity of the pneumococcal IreK homolog (27), but those workers were unable to construct a phosphatase mutant strain to thoroughly test this hypothesis. Thus, the nature of the in vivo relationship between the kinase and phosphatase remains unclear. In E. faecalis, a PP2C phosphatase is encoded immediately upstream of IreK, but its function was unknown prior to this study.

Here we show that the putative E. faecalis phosphatase (now designated IreP) is indeed a <emical">span class="Gene">phosphatase capable of dephosphorylating the IreK kinase in vitro. Phenotypic analyses of various mutants argues that this kinase-specific dephosphorylation activity represents an important function for IreP in vivo, and further that the IreK/IreP kinase-phosphatase pair comprises the core of a signal transduction pathway that reciprocally regulates intrinsic cephalosporin resistance in E. faecalis. This careful regulation of cephalosporin resistance is critical, as constitutive hyperresistance imposes a substantial fitness cost in the absence of cephalosporin stress.

RESULTS

E. faecalis IreK is a protein kinase whose activity is required for cephalosporin resistance.

Genetic analysis described in a previous study (9) revealed that emical">IreK is required for intrinsic <emical">span class="Chemical">cephalosporin resistance in E. faecalis. Sequence analysis predicted IreK to be a Ser/Thr protein kinase, and homologs of IreK from other Gram-positive bacteria have been experimentally shown to be kinases (13). To test if E. faecalis IreK does indeed exhibit kinase activity, we purified a recombinant 6His-tagged fragment of IreK corresponding to the entire N-terminal fragment containing the Ser/Thr kinase domain and juxtamembrane region (IreK-n) and performed in vitro kinase assays using myelin basic protein (MBP) as a surrogate substrate (MBP is routinely used as a substrate for IreK homologs from other species of Gram-positive bacteria). We also analyzed a mutant of IreK-n bearing a lysine-to-arginine substitution (K41R in IreK) at a conserved lysine within the kinase ATP-binding P loop. Mutations at this invariant lysine residue in other kinases of the IreK family are known to significantly impair kinase activity (19, 22, 28, 29). Using the phosphoprotein stain ProQ Diamond, we found that wild-type IreK-n phosphorylates MBP in the presence of ATP. Furthermore, IreK-n itself exhibited a strong signal, suggesting that it was autophosphorylated as well (Fig. 1A). The IreK-n K41R mutant was substantially impaired at phosphorylation of MBP and itself exhibited a significantly reduced phosphoprotein signal. In addition, we observed that the K41R mutant exhibited a subtle shift in electrophoretic mobility to a species that migrated faster through the gel than wild-type IreK-n, suggesting that IreK autophosphorylation resulted in reduced electrophoretic mobility. We conclude that, as expected, E. faecalis IreK is indeed a protein kinase.

FIG 1

IreK kinase activity is required for cephalosporin resistance in E. faecalis. (A) In vitro kinase activity of IreK. Wild-type (WT) and mutant (K41R) IreK kinase domains were purified and used for in vitro phosphorylation reactions with myelin basic protein (MBP) as a surrogate substrate. Reaction mixtures were incubated in the absence (−) or presence (+) of 2 mM ATP. At the indicated times (in minutes), aliquots were quenched with SDS loading buffer and subjected to SDS-PAGE. Phosphoproteins were detected using ProQ Diamond phosphoprotein stain, followed by GelCode blue staining to detect total proteins. Molecular weight standards are indicated at the left. Results are representative of a minimum of three independent experiments. (B) Kinase activity is required for resistance. Cultures of plasmid-bearing strains were subjected to serial 10-fold dilutions and inoculated (left to right, least to most dilute) onto BHI agar supplemented with Em alone (control) or in addition to a cephalosporin antibiotic (ceftriaxone, 1 µg/ml). The WT and ΔirePK strains were OG1RF and CK125, respectively. Plasmids are indicated in parentheses: vector, pJRG8 empty vector; WT, pCJK160 expressing wild-type IreP and IreK; K41R, pCJK216 (analogous to pCJK160 but carrying the K41R allele of ireK). (C) Immunoblot analysis of IreK expression. Whole-cell lysates from CK125 (ΔirePK) carrying empty vector (pJRG8) or pJRG8 expressing ireP and either wild-type ireK (pCJK160) or ireK K41R (pCJK216) were probed with anti-IreK antibody (α-kinase) or anti-sigma factor antibody (α-sigma) as a loading control. Results are representative of a minimum of two experiments analyzing at least three transformants derived from independent electroporations.

emical">IreK kinase activity is required for <emical">span class="Chemical">cephalosporin resistance in E. faecalis. (A) In vitro kinase activity of IreK. Wild-type (WT) and mutant (K41R) IreK kinase domains were purified and used for in vitro phosphorylation reactions with myelin basic protein (MBP) as a surrogate substrate. Reaction mixtures were incubated in the absence (−) or presence (+) of 2 mM ATP. At the indicated times (in minutes), aliquots were quenched with SDS loading buffer and subjected to SDS-PAGE. Phosphoproteins were detected using ProQ Diamond phosphoprotein stain, followed by GelCode blue staining to detect total proteins. Molecular weight standards are indicated at the left. Results are representative of a minimum of three independent experiments. (B) Kinase activity is required for resistance. Cultures of plasmid-bearing strains were subjected to serial 10-fold dilutions and inoculated (left to right, least to most dilute) onto BHI agar supplemented with Em alone (control) or in addition to a cephalosporin antibiotic (ceftriaxone, 1 µg/ml). The WT and ΔirePK strains were OG1RF and CK125, respectively. Plasmids are indicated in parentheses: vector, pJRG8 empty vector; WT, pCJK160 expressing wild-type IreP and IreK; K41R, pCJK216 (analogous to pCJK160 but carrying the K41R allele of ireK). (C) Immunoblot analysis of IreK expression. Whole-cell lysates from CK125 (ΔirePK) carrying empty vector (pJRG8) or pJRG8 expressing ireP and either wild-type ireK (pCJK160) or ireK K41R (pCJK216) were probed with anti-IreK antibody (α-kinase) or anti-sigma factor antibody (α-sigma) as a loading control. Results are representative of a minimum of two experiments analyzing at least three transformants derived from independent electroporations.

Because the E. faecalis Δemical">ireK mutant is markedly susceptible to <emical">span class="Chemical">cephalosporins, the IreK protein must play a role in promoting cephalosporin resistance. To test if this phenotype reflects a requirement for the kinase activity of IreK or for the IreK protein per se, we expressed either the full-length wild-type or K41R mutant ireK kinase from a plasmid-based expression system in an E. faecalis host lacking any endogenous IreK kinase. Immunoblot analysis confirmed that both kinase alleles were expressed at comparable levels (Fig. 1C). Expression of the wild-type kinase restored cephalosporin resistance to the mutant, whereas the ireK(K41R) mutant was unable to do so (Fig. 1B), indicating that the kinase activity of IreK is indeed critical for cephalosporin resistance.

E. faecalis IreP is a protein phosphatase that can dephosphorylate the IreK kinase.

In most low-GC Gram-positive bacteria, the gene located immediately upstream of the emical">IreK kinase homolog is cotranscribed and encodes a protein <emical">span class="Gene">phosphatase of the PP2C family. The genetic organization of the ireK locus in E. faecalis is similar, with a putative protein phosphatase (here named IreP) encoded upstream of, and slightly overlapping, the gene for the ireK kinase. We tested for cotranscription of ireP and ireK by isolating RNA from wild-type E. faecalis and performing reverse transcription-PCR analysis. We observed evidence for a transcript containing both ireP and ireK (data not shown), indicating that the kinase and phosphatase are coregulated in E. faecalis. Our data also indicate that the ireP-ireK pair can be cotranscribed with the genes immediately upstream of ireP (a putative rRNA methyltransferase; EF3122) and downstream of ireK (a putative GTPase; EF3119). Of note, the B. subtilis homolog of IreK (PrkC) has been reported to phosphorylate the GTPase encoded downstream (called CpgA in B. subtilis), although it remains unclear if this is functionally significant (30). We attempted to phosphorylate the E. faecalis homolog (EF3119) with IreK-n in vitro but did not observe any evidence for phosphorylation under our conditions (data not shown).

To test if E. faecalis IreP exhibited phosphatase activity, we purified a recombinant <emical">span class="Chemical">6His-tagged IreP and performed in vitro phosphatase assays using the small-molecule colorimetric phosphatase substrate p-nitrophenyl phosphate, which forms a colored product upon dephosphorylation. E. faecalis IreP exhibited phosphatase activity in a manganese-dependent manner (Fig. 2A), consistent with the properties of IreP homologs from other Gram-positive bacteria. E. faecalis IreP was also capable of dephosphorylating the IreK substrate MBP and, importantly, IreK-n kinase itself (Fig. 2B). In addition, IreP treatment resulted in a subtle shift in electrophoretic mobility of IreK-n that appears to be characteristic of a change in phosphorylation status. We note that IreP itself exhibited a faint phosphoprotein signal; however, control experiments in the absence of ATP (and on IreP purified directly from E. coli) indicated that this signal represents nonspecific background staining of IreP (not shown).

FIG 2

IreP is a phosphatase that can dephosphorylate IreK. Recombinant His-tagged IreP was purified and used for in vitro phosphatase reactions. (A) Phosphatase activity monitored using p-nitrophenyl phosphate (pNPP) as a substrate. IreP was incubated with pNPP and various concentrations of either Mg2+ or Mn2+ as a cofactor for 20 min at room temperature. The reactions were quenched, and absorbance was measured at 405 nm to detect cleaved product. Error bars represent standard deviations for triplicate samples and are too small to see in some cases. (B) IreK-n was incubated in kinase buffer with ATP and MBP for 30 min to allow phosphorylation to occur. The reaction mixture was split, and IreP was added to 1 aliquot; mixtures were incubated for 30 min and subjected to SDS-PAGE. Phosphoproteins were detected using ProQ Diamond phosphoprotein stain followed by GelCode blue staining to detect total proteins. Results are representative of a minimum of three independent experiments.

IreP is a phosphatase that can dephoemical">sphorylate <emical">span class="Chemical">IreK. Recombinant His-tagged IreP was purified and used for in vitro phosphatase reactions. (A) Phosphatase activity monitored using p-nitrophenyl phosphate (pNPP) as a substrate. IreP was incubated with pNPP and various concentrations of either Mg2+ or Mn2+ as a cofactor for 20 min at room temperature. The reactions were quenched, and absorbance was measured at 405 nm to detect cleaved product. Error bars represent standard deviations for triplicate samples and are too small to see in some cases. (B) IreK-n was incubated in kinase buffer with ATP and MBP for 30 min to allow phosphorylation to occur. The reaction mixture was split, and IreP was added to 1 aliquot; mixtures were incubated for 30 min and subjected to SDS-PAGE. Phosphoproteins were detected using ProQ Diamond phosphoprotein stain followed by GelCode blue staining to detect total proteins. Results are representative of a minimum of three independent experiments.

E. faecalis ΔireP mutants exhibit hyperresistance to cephalosporins.

We hypothesized that IreP-mediated dephosphorylation of emical">IreK in vivo might play a critical regulatory role in controlling <emical">span class="Chemical">IreK kinase activity and, by extension, cephalosporin resistance. To probe the role of IreP in vivo, we constructed E. faecalis mutants lacking the ireP gene. In this context, it is noteworthy that IreP is the only identifiable PP2C-family phosphatase encoded in the E. faecalis genome. Because ireP overlaps with ireK, we constructed an in-frame deletion lacking 92% of the ireP gene without disrupting any ireK coding sequences, to avoid perturbing expression of the ireK kinase. Immunoblot analysis verified that IreK was indeed expressed in the ΔireP mutant at levels comparable to that of wild-type (Fig. 3B). On brain heart infusion (BHI) agar plates, colonies of the ΔireP mutants exhibited a distinct morphology, appearing more opaque (white) and compact than those of the isogenic wild type. Exponential-growth rates for the ΔireP mutant in liquid culture were only slightly lower than those for the wild type (generation times of 31 ± 2 min versus 37 ± 2 min in Mueller-Hinton broth [MHB] for the wild type [CK138] and the isogenic ΔireP mutant [CK204], respectively). However, antimicrobial susceptibility tests revealed a striking phenotype: the ΔireP mutant was substantially more resistant to cephalosporins (>64-fold for ceftriaxone) than the isogenic wild-type strain (Fig. 3A; Table 1). Indeed, the ΔireP mutant was capable of growth at all cephalosporin concentrations tested (up to 2,048 µg/ml). Expression of ireP in trans eliminated hyperresistance (Fig. 3C), indicating that hyperresistance was indeed due to the lesion in ireP. Furthermore, we constructed an identical deletion of ireP in a divergent lineage of E. faecalis (T1) and found that the E. faecalis T1 ΔireP mutant also exhibited a cephalosporin hyperresistance phenotype relative to its isogenic wild-type parent (Fig. 3A). Thus, the role of IreP in regulating cephalosporin resistance is likely conserved across E. faecalis as a species.

FIG 3

E. faecalis ΔireP mutants exhibit hyperresistance to cephalosporins. (A) Cultures were subjected to 10-fold serial dilutions and inoculated (left to right, least to most dilute) on BHI agar supplemented with indicated concentrations of ceftriaxone. Strains: OG, wild-type E. faecalis OG1RF; OG ΔireK, CK119; OG ΔireP, CK121; T1, wild-type E. faecalis T1; T1 ΔireK, JL202; T1 ΔireP, JL204. (B) Immunoblot analysis of IreK expression. Whole-cell lysates from OG1RF (wild-type), CK119 (ΔireK), CK125 (ΔireP ΔireK), and CK121 (ΔireP) were probed with anti-IreK antibody (α-kinase) or anti-sigma factor antibody (α-sigma) as a loading control. (C) Complementation analysis of the E. faecalis ΔireP mutant. Cultures of plasmid-bearing strains were subjected to serial 10-fold dilutions and inoculated (left to right, least to most dilute) onto BHI agar supplemented with Em alone (control) or in addition to ceftriaxone. WT and ΔireP strains were OG1RF and CK121. Plasmids are indicated in parentheses: vector, pJRG8 empty vector; ireP, pJLL25 expressing wild-type IreP. Results are representative of a minimum of three experiments analyzing independently derived mutants.

TABLE 1

Median MICs for wild-type and mutant E. faecalis strains

Drug type or target and name	MIC (μg/ml)a for:
	OG1RF (wild type)	CK119 (ΔireK)	CK121 (ΔireP)	CK125 (ΔireP ΔireK)
Cephalosporins
Ceftriaxone	32	2	>2,048	2
Ceftazidime	128	16	>2,048	16
Other cell wall
Ampicillin	1	0.5	2	0.5
Vancomycin	2	1	1	1
Bacitracin	64	32	32	32
d-Cycloserine	128	64	256	64
Other targets
Norfloxacin	4	4	2	4
Chloramphenicol	4	4	4	4
Kanamycin	128	128	64	128

 Determined in MHB after 24 h incubation at 37°C from a minimum of three independent experiments.

E. faecalis ΔireP mutants exhibit hyperresistance to emical">cephalosporins. (A) Cultures were subjected to 10-fold <emical">span class="Chemical">serial dilutions and inoculated (left to right, least to most dilute) on BHI agar supplemented with indicated concentrations of ceftriaxone. Strains: OG, wild-type E. faecalis OG1RF; OG ΔireK, CK119; OG ΔireP, CK121; T1, wild-type E. faecalis T1; T1 ΔireK, JL202; T1 ΔireP, JL204. (B) Immunoblot analysis of IreK expression. Whole-cell lysates from OG1RF (wild-type), CK119 (ΔireK), CK125 (ΔireP ΔireK), and CK121 (ΔireP) were probed with anti-IreK antibody (α-kinase) or anti-sigma factor antibody (α-sigma) as a loading control. (C) Complementation analysis of the E. faecalis ΔireP mutant. Cultures of plasmid-bearing strains were subjected to serial 10-fold dilutions and inoculated (left to right, least to most dilute) onto BHI agar supplemented with Em alone (control) or in addition to ceftriaxone. WT and ΔireP strains were OG1RF and CK121. Plasmids are indicated in parentheses: vector, pJRG8 empty vector; ireP, pJLL25 expressing wild-type IreP. Results are representative of a minimum of three experiments analyzing independently derived mutants.

Median MICs for wild-type and mutant <span class="Species">E. faecalis strains

Determined in <span class="Chemical">MHB after 24 h incubation at 37°C from a minimum of <emical">span class="Chemical">three independent experiments.

Previous work established that E. faecalis requires the emical">IreK kinase for its intrinsic resistance to <emical">span class="Chemical">cephalosporins but not for resistance to antibiotics affecting other cellular processes. Because we hypothesized that IreP functions to regulate the activity of the IreK kinase, we reasoned that the ΔireP mutant would likewise exhibit hyperresistance specifically towards cephalosporins. To test this, we performed susceptibility analyses with antibiotics targeting various cellular functions. The results (Table 1) indicate that the hyperresistant phenotype of the ΔireP mutant is essentially specific for cephalosporins, as few changes in susceptibility to other antibiotics were apparent. Thus, hyperresistance of the ΔireP mutant is not the result of enhancement in a general stress response. Furthermore, these results are consistent with the hypothesis that an important role of IreP is to control IreK kinase activity, thereby influencing cephalosporin resistance.

Hyperresistance of the E. faecalis ΔireP mutant is mediated by IreK.

Phosphorylation of amino acids in the kinase activation loop is known to enhance the catalytic activity of emical">Ser/<emical">span class="Chemical">Thr kinases in the superfamily to which IreK belongs. In the absence of the IreP phosphatase, we reasoned that IreK might become highly phosphorylated, including at sites in the activation loop, leading to enhanced kinase activity that could drive hyperresistance to cephalosporins. To test the hypothesis that hyperresistance of the ΔireP mutant is a consequence of uncontrolled, abnormally high IreK kinase activity, we performed an epistasis experiment by constructing a mutant of E. faecalis lacking the genes for both the ireP phosphatase and the ireK kinase. Antimicrobial susceptibility tests revealed that the double mutant phenocopied the kinase single mutant (Table 1)—removal of ireK eliminated hyperresistance exhibited by the ΔireP mutant. Indeed, no differences were observed between the phenotype of the ΔireP ΔireK double mutant and that of the kinase single mutant. Thus, these data are consistent with the hypothesis that IreP regulates cephalosporin resistance at least in part by modulating activity of the IreK kinase.

Given the obemical">servation that IreP is capable of dephoemical">sphorylating <emical">span class="Chemical">IreK in vitro (Fig. 2B), we hypothesized that IreP-mediated dephosphorylation of sites in the activation loop of the IreK kinase served to control IreK activity in vivo. To test this, we constructed a kinase allele carrying phosphomimetic (T-to-E) substitutions at the three predicted sites of phosphorylation in the IreK activation loop [ireK(T163E/T166E/T168E)]. The likely sites of phosphorylation in IreK were chosen based on sequence alignment and comparison with known sites of phosphorylation on the mycobacterial PknB and B. subtilis PrkC kinases. Mass spectrometry analyses subsequently confirmed that these sites can be phosphorylated on IreK-n (C. L. Hall and C. J. Kristich, unpublished data). We reasoned that the T-to-E substitutions in IreK would mimic phosphorylation at these sites and lead to kinase activation, but because they are uncleavable by IreP, the mutant kinase would exhibit constitutively high activity. We predicted that this would lead to enhanced cephalosporin resistance—in principle, similar to the phenotype of the ΔireP mutant—despite the presence of wild-type IreP phosphatase in the cells. We expressed the T-to-E triple mutant kinase along with wild-type IreP phosphatase in E. faecalis and confirmed by immunoblotting that the mutant kinase was expressed at levels comparable to that of the wild type (Fig. 4A). Antimicrobial susceptibility tests revealed that expression of the phosphomimetic-bearing-kinase allele does indeed provide enhanced cephalosporin resistance relative to its wild-type counterpart (Table 2), consistent with the hypothesis that IreP dephosphorylates the activation loop of IreK kinase to negatively control its activity. We note that although the strain expressing the phosphomimetic ireK allele exhibits enhanced cephalosporin resistance, it is not as hyperresistant as the ΔireP mutant, suggesting that IreP-mediated dephosphorylation of heterologous kinase substrates—or other phosphorylated sites on the kinase itself—are also important for regulation of cephalosporin resistance in vivo. Alternatively, the T-to-E substitutions may not fully mimic the effect of phosphorylation at those sites.

FIG 4

Reduced phosphatase activity in lysates with enhanced IreK kinase activity. (A) Immunoblot analysis of IreK expression. Whole-cell lysates from CK125 carrying an empty vector (pJRG8) or expressing IreP with either a wild-type IreK kinase (pCJK160) or the T163E/T166E/T168E triple mutant phosphomimetic allele (pCJK201) were probed with anti-IreK antibody (α-kinase) or anti-sigma factor antibody (α-sigma) as a loading control. (B) Phosphothreonine-specific phosphatase activity was measured in cleared whole-cell lysates of the strains used for panel A. Error bars represent standard errors of the means of values from three independent lysates and are too small to be seen in most cases.

TABLE 2

Median MICs for CK125 (ΔirePK) harboring different plasmids

Drug	MIC (μg/ml)a for CK125 harboring:
	pJRG8 (vector)	pCJK160 (WT ireK)	pCJK201 (T163E/T166E/T168E)
Ceftriaxone	2	8	64
Ceftazidime	8	64	128

 Determined in MHB supplemented with Em after 24 h incubation at 37°C from a minimum of three independent experiments.

Reduced phosphatase activity in lysates with enhanced <emical">span class="Chemical">IreK kinase activity. (A) Immunoblot analysis of IreK expression. Whole-cell lysates from CK125 carrying an empty vector (pJRG8) or expressing IreP with either a wild-type IreK kinase (pCJK160) or the T163E/T166E/T168E triple mutant phosphomimetic allele (pCJK201) were probed with anti-IreK antibody (α-kinase) or anti-sigma factor antibody (α-sigma) as a loading control. (B) Phosphothreonine-specific phosphatase activity was measured in cleared whole-cell lysates of the strains used for panel A. Error bars represent standard errors of the means of values from three independent lysates and are too small to be seen in most cases.

Determined in <span class="Chemical">MHB supplemented with Em after 24 h incubation at 37°C from a minimum of <emical">span class="Chemical">three independent experiments.

Evidence for modulation of IreP phosphatase activity by IreK.

The experiments described above suggest that an important role of IreP is to negatively control emical">IreK kinase activity via dephoemical">sphorylation of the kinase activation loop. However, t<emical">span class="Chemical">his inhibitory activity is counterproductive when E. faecalis is confronted with cephalosporins, in which case enhanced kinase activity is desirable to mediate the appropriate biological response and generate resistance. Therefore, we reasoned that a mechanism might exist to transiently overcome the negative regulatory effect of IreP—to attenuate the phosphatase activity of IreP—upon kinase activation. To test this, we performed phosphatase assays on lysates of E. faecalis strains, using a phosphothreonine-containing peptide as a surrogate phosphatase substrate. We compared phosphatase activity in lysates of a strain expressing wild-type IreK kinase with that in lysates of a strain expressing the hyperactive (phosphomimetic) triple-T-to-E allele of ireK (to mimic a state of kinase activation). IreP-specific phosphatase activity in lysates from the strain containing the phosphomimetic kinase was reproducibly lower than that in lysates from a strain with a wild-type kinase (Fig. 4B), suggesting that some mechanism exists to attenuate IreP phosphatase activity, at least transiently, upon kinase activation.

Hyperresistant ΔireP mutants exhibit a fitness defect.

Our data support the hypothesis that IreP and emical">IreK comprise key elements of a signal transduction system that regulates intrinsic <emical">span class="Chemical">cephalosporin resistance in E. faecalis. That such a signaling system exists implies that the cellular adaptation(s) required to overcome cephalosporin stress imposes a fitness cost on the organism in the absence of cephalosporins. To test this hypothesis, we performed coculture competition experiments in which a marked wild-type E. faecalis strain (marked with an unrelated antibiotic resistance allele) was cocultured with a differentially marked mutant strain lacking the IreP phosphatase. Cocultures were inoculated with various ratios of the two strains and subjected to repeated daily cycles of growth and dilution into fresh medium lacking cephalosporins. Aliquots of the cocultures were removed at intervals and plated on appropriate selective media to distinguish wild-type E. faecalis from ΔireP mutants. The results (Fig. 5) revealed that the wild-type strains rapidly outcompeted the ΔireP mutants over the course of even a single dilution/growth cycle (day 1) and even when the inoculum (day 0) was initially composed of >90% mutants. Reciprocal competition experiments performed with strains in which the antibiotic markers were swapped (to ensure that the observed fitness defect of the ΔireP mutant was not impacted by the antibiotic markers used to tag the strains) yielded similar results (Fig. 5). Thus, mutants lacking ireP are substantially less fit in the absence of cephalosporins than wild-type E. faecalis, suggesting that IreK-mediated adaptation(s) to cephalosporin stress, while obviously beneficial in the presence of cephalosporins, is costly to the cell in environments devoid of cephalosporins.

FIG 5

Hyperresistant ΔireP mutants exhibit a competitive fitness defect. Differentially marked wild-type or ΔireP mutants were inoculated at various ratios and cocultured in MHB with daily repeated cycles of growth and dilution. At intervals, samples were removed and dilutions spread on appropriate selective media to enumerate wild-type and ΔireP organisms present. (A) Wild-type (OG1Sp, Spr) and ΔireP (CK204, Far); (B) wild-type (CK138, Far) and ΔireP (JL178, Spr). Data are representative of a minimum of three independent experiments.

Hyperresistant ΔireP mutants exhibit a competitive fitness defect. Differentially marked wild-type or ΔireP mutants were inoculated at various ratios and cocultured in <emical">span class="Chemical">MHB with daily repeated cycles of growth and dilution. At intervals, samples were removed and dilutions spread on appropriate selective media to enumerate wild-type and ΔireP organisms present. (A) Wild-type (OG1Sp, Spr) and ΔireP (CK204, Far); (B) wild-type (CK138, Far) and ΔireP (JL178, Spr). Data are representative of a minimum of three independent experiments.

DISCUSSION

The experiments described here were conducted in an effort to understand the role of the E. faecalis PP2C phosphatase (IreP) and <emical">span class="Chemical">Ser/Thr kinase (IreK) in mediating intrinsic cephalosporin resistance. Our results argue that the IreK/IreP kinase-phosphatase pair comprises the core of a signal transduction pathway that reciprocally regulates intrinsic cephalosporin resistance in E. faecalis. We used genetic analyses in two divergent lineages of E. faecalis—OG1 and T1, belonging to multilocus sequence types 1 and 21, respectively (31)—to show that mutants lacking the IreP phosphatase exhibit hyperresistance to cephalosporins. Given that the ireP ireK locus can be identified in all E. faecalis genomes sequenced to date (32–34), it seems likely that the IreP/IreK signaling system reciprocally controls cephalosporin resistance in most (or all) isolates of E. faecalis. Additionally, our genetic analysis indicates that phosphorylation of IreK leads to kinase activation in vivo (Table 2) and further suggests that a critical biological role of the IreP phosphatase is to regulate IreK activity—and, by extension, cephalosporin resistance—by controlling the level of IreK phosphorylation. The observation that ireK deletion is epistatic to ireP deletion (Table 1) is consistent with this model. The possibility that IreP could also regulate cephalosporin resistance by controlling the phosphorylation level of a downstream target(s) of IreK has not been excluded, and such a mechanism may well contribute to overall regulation of cephalosporin resistance. However, a mechanism must exist to deactivate the IreK kinase in the absence of cephalosporin stress to restore kinase activity to precephalosporin levels. It seems likely that any such mechanism, although it may be multifactorial, will require dephosphorylation of the kinase, and our data argue that IreP plays an important role in this process in vivo.

In the presence of emical">cephalosporin stress, the inhibitory (<emical">span class="Gene">phosphatase) activity of IreP on IreK would be counterproductive to E. faecalis, as enhanced kinase activity is desirable to promote resistance. Therefore, we reasoned that a mechanism might exist to transiently overcome the negative regulatory effect of IreP—to attenuate the phosphatase activity of IreP—upon kinase activation. Our results suggest that enhanced IreK kinase activity may indeed lead to a reduction in IreP phosphatase activity (Fig. 4), in principle enabling the kinase pool to become (at least transiently) more active and mediate signaling to upregulate the as-yet-unknown resistance mechanism(s). While more work is needed to unravel the mechanism underlying this observation, we speculate that IreK may phosphorylate IreP directly to inhibit its phosphatase activity. We hypothesize that with IreP in an “inactive” state, a pool of IreK could become more highly phosphorylated (activated) and serve to activate downstream cephalosporin resistance mechanisms. A recent report suggested that the IreP homolog of M. tuberculosis (PstP) can be phosphorylated by its cognate kinase (35), although in that case phosphorylation appeared to activate, rather than inhibit, phosphatase activity. We tested to see if recombinant E. faecalis IreK-n could phosphorylate IreP in vitro but were unable to observe any evidence of phosphorylation under our conditions (data not shown).

The obemical">servation that the ΔireP mutant exhibited hyperresistance to <emical">span class="Chemical">cephalosporins but not to ampicillin (Table 1) is intriguing given that both cephalosporins and ampicillin belong to the β-lactam class of antibiotics, all of which are thought to exert their antimicrobial activity via inactivation of PBPs to prevent PG cross-linking. We speculate that the inherent differences in affinity for PBPs of broad-spectrum cephalosporins compared to ampicillin account for this disparity. Ampicillin presumably inhibits the entire repertoire of enterococcal PBPs efficiently (leading to relatively low MICs), whereas the cephalosporins are unable to bind efficiently to Pbp5 and cannot inhibit its activity (leading to intrinsic cephalosporin resistance). Exposure to cephalosporins may therefore lead to a unique physiological state in which Pbp5 is active—but other PBPs are inhibited—that somehow triggers IreK activation. The PASTA domains of IreK-like kinases appear to bind PG or fragments thereof (10–12), suggesting that cephalosporin exposure could lead to accumulation of an IreK-activating ligand in PG as a component of this mechanism. However, the CroRS two-component signaling system is also required for full cephalosporin resistance in E. faecalis (7, 8), indicating that the resistance signaling pathway is more complex and may involve additional signals. The mechanism by which IreK/P is integrated with CroRS to confer resistance is unknown, but IreK-like kinases from streptococci are known to phosphorylate two-component response regulators (36, 37), suggesting a potential route for direct control of the CroR response regulator by IreK. Our initial attempts to phosphorylate CroR with IreK in vitro have proven unsuccessful (data not shown). Future studies will address the functional interconnections between these signaling systems.

Many studies have shown that antibiotic-resistant bacteria are less fit than their susceptible counterparts in the absence of antibiotic stress (38). For example, a recent study demonstrated that constitutive expression of emical">vancomycin resistance in enterococci leads to a significant <emical">span class="Disease">fitness reduction in the absence of vancomycin (39) and, furthermore, that tight regulation of vancomycin resistance expression by the VanSR signal transduction system reduces the biological cost of vancomycin resistance dramatically. Our results indicate that constitutive cephalosporin resistance also imposes a substantial fitness cost on E. faecalis in the absence of cephalosporins (Fig. 5). The molecular basis for this fitness cost is not yet clear, in part because the output of the IreK/IreP signaling pathway is unknown. In any case, regulation of cephalosporin resistance by the IreK/IreP signaling system plays a critical role in balancing the expression of resistance functions given the needs of the cell in a particular environment to minimize the biological cost associated with resistance and maximize the ability of E. faecalis to be competitive in the face of environmental fluctuations it encounters in the GI tract.

MATERIALS AND METHODS

Bacterial strains, growth media, and chemicals.

Strains used in temical">his study are listed in Table 3. Brain heart infusion medium (<emical">span class="Chemical">BHI) and Mueller-Hinton broth (MHB) were prepared as described by the manufacturer (Becton Dickinson). Bacteria were stored at −80°C in BHI supplemented with 30% glycerol. Antibiotics and other chemicals were obtained from Sigma unless otherwise indicated. Erythromycin (Em) was used at 10 µg/ml, spectinomycin (Sp) at 1,000 µg/ml, and fusidic acid (Fa) at 25 µg/ml for growth of resistant E. faecalis.

TABLE 3

Strains and plasmids used in this study

Strain or plasmid	Relevant description or genotype[a]	Source or reference
Strains
E. coli
TOP10	Routine cloning host	Invitrogen
BL21[DE3]	Protein overproduction host	Lab stock
E. faecalis
OG1	Wild-type, original unmarked isolate (MLST 1)	43
OG1RF	Spontaneous rifampin-resistant and Far derivative of OG1	44
CK119	OG1RF ΔireK2	9
CK121	OG1RF ΔireP2	This work
CK125	OG1RF Δ(ireP-ireK)2	This work
OG1Sp	Spontaneous Spr derivative of OG1	40
JL178	OG1Sp ΔireP2	This work
CK138	Spontaneous Far derivative of OG1	This work
CK204	CK138 ΔireP2	This work
T1 (SS498)	Wild-type (MLST 21), CDC reference strain	45
JL202	T1 ΔireK2	This work
JL204	T1 ΔireP2	This work
Plasmids
pCJK47	Counterselectable vector for allelic exchange	40
pCJK74	ΔireK2 allele in pCJK47	9
pCJK75	ΔireP2 allele in pCJK47	This work
pCJK105	Δ(ireP-ireK)2 allele in pCJK47	This work
pCJK111	pET28b::ireK-n (kinase/juxtamembrane domain)	This work
pCJK112	pET28b::ireP	This work
pJRG8	E. faecalis expression vector, constitutive P23 promoter (Emr)	This work
pJLL25	pJRG8::ireP	This work
pCJK160	pJRG8::ireP ireK	This work
pCJK201	pJRG8::ireP ireK(T163/166/168E)	This work
pCJK216	pJRG8::ireP ireK K41R	This work

 MLST, multilocus sequence type.

Strains and plasmids used in t<span class="Chemical">his study

Construction of E. faecalis mutants.

All PCR amplifications used E. faecalis OG1RF genomic DNA as the template and Pfu II Ultra polymerase (Stratagene). The markerless exchange system described by Kristich et al. (40) was used to construct unmarked, in-frame deletions of ireP in various genetic backgrounds. Briefly, a derivative of plasmid pCJK47 carrying an in-frame deletion allele of ireP (pCJK75) was constructed using the BsaI-based cloning scheme (40) to seamlessly fuse two PCR amplicons flanking ireP to form the in-frame deletion. The deletion allele was designed such that the first 10 codons and the last 10 codons of the ireP gene remained, in an effort to avoid any unanticipated effects on expression of adjacent genes, removing 92% of the ireP gene. T<emical">span class="Chemical">his ΔireP allele was transferred to the native ireP location in the E. faecalis chromosome using pVE6007 as a helper plasmid to facilitate recombination as previously described (41). Successful isolation of ΔireP mutants was achieved after incubation of counterselection plates at room temperature or 30°C for ~3 days. An E. faecalis double mutant lacking ireP and ireK (CK125) was constructed via an analogous strategy, using a pCJK47 derivative (pCJK105) carrying an in-frame deletion of both ireP and ireK. This double mutant allele retained the first 10 codons of ireP and the last 6 codons of ireK. Finally, the previously described ΔireK2 allele was introduced into E. faecalis T1 using pVE6007-assisted recombination of pCJK74.

Construction of plasmids.

A plasmid to express genes in E. faecalis was constructed by amplifying the constitutive P23 promoter (lacking the last 34 nucleotides) from pDL278p23 (42) and cloning it with primer-specified BglII and SpeI restriction sites to replace the rhamnose-inducible promoter of pCJK96 (9), creating pJRG8. For complementation of the ΔireP mutation, ireP was amplified and cloned into pJRG8 using primer-specified SphI and XhoI sites, yielding pJLL25. For analysis of mutant emical">ireK alleles, we chose to express both ireP and <emical">span class="Chemical">ireK from pJRG8 (introduced into an E. faecalis host lacking both genes, CK125) due to the likely translational coupling of these genes, in an attempt to ensure that both gene products were produced in the appropriate stoichiometry reflecting the natural state. To do so, the locus containing ireP and ireK was amplified and cloned using primer-specified SphI/XhoI sites into pJRG8, creating pCJK160. We note that, for unrelated purposes, an artificial XbaI site was introduced near the end corresponding to the C terminus of IreP encoded in pCJK160 via silent mutagenesis. This plasmid served as the basis for introduction of specific point mutations in ireK using a BsaI-based seamless cloning strategy, creating pCJK201 and pCJK216. Plasmids to express N-terminally His-tagged IreP and IreK-n were constructed by amplifying either full-length ireP or residues 1 to 331 of ireK (kinase domain plus juxtamembrane segment) and cloning them into pET28b (Novagen) using primer-specified NdeI/XhoI sites.

Antibiotic susceptibility.

MICs of antibiotics were determined in aerobic liquid cultures using a microtiter plate emical">serial dilution method in a Bioscreen C plate reader (Oy Growth Curves Ab, Ltd.). Twofold dilutions of antibiotics in <emical">span class="Chemical">MHB were prepared in the wells of a 100-well honeycomb microtiter plate. Bacteria from stationary-phase cultures in MHB were inoculated into each well to a concentration of ~105 CFU/ml. Plates were incubated at 37°C for 24 h with brief shaking and measurement of optical density at 600 nm (OD600) at 15-min intervals. The lowest concentration of antibiotic that prevented growth was recorded as the MIC. In some cases, antibiotic susceptibility was also assessed by preparing serial 10-fold dilutions of stationary-phase cultures and inoculating aliquots onto the surface of agar plates supplemented with antibiotics.

Protein purification.

Overnight cultures of E. coli BL21[DE3] carrying the desired expression plasmid were diluted 50-fold in fresh LB media and cultured at 37°C for 3 h. Cells were induced with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside) for 1 h and collected by centrifugation, suemical">spended in 10 ml of binding buffer (50 mM <emical">span class="Chemical">Tris [pH 8.0], 300 mM NaCl, 5 mM imidazole), and treated with 1 mg/ml lysozyme for 20 min at 37°C. After disruption by sonication, the lysates were clarified by centrifugation (35,000 × g for 15 min) and passed through a 0.2-µm filter. Clarified lysates were applied to Profinity Ni-charged resin (Bio-Rad) previously equilibrated with binding buffer and washed with 10 column volumes of wash buffer (50 mM Tris [pH 8.0], 300 mM NaCl, 20 mM imidazole), and bound proteins were recovered with elution buffer (50 mM Tris [pH 8.0], 300 mM NaCl, 500 mM imidazole). Fractions containing the protein of interest were pooled and dialyzed into storage buffer (50 mM Tris [pH 8], 150 mM NaCl, 10% glycerol) at 4°C.

Kinase activity of IreK-n.

Purified emical">6His-<emical">span class="Chemical">IreK-n (1.3 µM) was incubated with myelin basic protein (9.4 µM) in kinase buffer (50 mM Tris [pH 7.5], 25 mM NaCl, 1 mM MnCl2, 1 mM MgCl2, 1 mM dithiothreitol, 0.1 mM EDTA) at 37°C in the presence or absence of ATP (2 mM). In some cases, purified 6His-IreP was included at 0.8 µM. Aliquots were removed at intervals and quenched with Laemmli sodium dodecyl sulfate (SDS) sample buffer. Following SDS-polyacrylamide gel electrophoresis (PAGE), ProQ Diamond phosphoprotein stain (Invitrogen) was used to detect phosphorylated proteins according to the manufacturer’s instructions. Total protein was subsequently detected in the same gel using GelCode blue (Pierce).

Phosphatase activity of purified IreP.

Reactions were carried out in volumes of 100 µl in 96-well plates. Purified emical">6His-IreP (168 nM) was incubated with 20 mM <emical">span class="Chemical">para-nitrophenyl phosphate (Pierce) in 50 mM Tris (pH 8.0) supplemented with various concentrations of MnCl2 or MgCl2 for 20 min at room temperature. Reactions were terminated by addition of 50 µl of 2 M NaOH, and absorbance at 405 nm was measured.

Phosphatase activity in lysates.

Cultures growing exponentially in emical">BHI plus Em were harvested by centrifugation and stored at −20°C. Thawed pellets were washed <emical">span class="Chemical">three times with ultrapure water and suspended in 250 µl of lysis buffer (50 mM Tris [pH 7.4], 1 mM EGTA, 0.2% Triton X-100, 0.1% β-mercaptoethanol) containing 1× HALT protease inhibitor (Pierce). Bacteria were disrupted by bead beating. Beads and intact bacteria were collected by centrifugation (16,000 × g, 15 min), and lysates were passed through desalting columns (Bio-Spin 6, Bio-Rad). Protein concentration was determined using Coomassie plus protein assay reagent (Pierce). Phosphatase activity was determined with a serine/threonine phosphatase assay system kit (Promega). Desalted lysates were added at a final concentration of 10 µg/ml to reaction mixtures containing 50 mM imidazole (pH 7.2), 5 mM MnCl2, 0.02% β-mercaptoethanol, 200 µM EGTA, and 200 µM phosphothreonine peptide and incubated at 37°C. Control reaction mixtures did not contain phosphopeptide. At intervals, aliquots were removed and mixed with an equal volume of molybdate dye substrate to stop the reaction. Absorbance was measured at 630 nm.

Competition experiments.

Strains to be competed were cultured (separately) overnight in emical">MHB. Culture density (OD600) was determined, and the differentially marked wild-type and mutant strains were mixed in the desired proportion (typically ~90 to 95% mutant, 5 to 10% wild-type) in fresh <emical">span class="Chemical">MHB. The mixtures were subjected to serial dilutions, and aliquots were plated on BHI agar plates supplemented with appropriate antibiotics to enumerate wild-type and mutant cells in the inoculum. The mixed inoculum was diluted to a density of ~104 CFU/ml and incubated at 37°C until the following day. Dilutions and enumeration were repeated on successive days.

Antibodies and immunoblots.

Anti-emical">IreK anti<emical">span class="Chemical">serum (a generous gift from Patrick Schlievert) was produced by immunization of Dutch Belted rabbits with purified 6His-IreK-n and used at a dilution of 1 to 10,000 for immunoblots with goat anti-rabbit IgG horseradish peroxidase (HRP)-conjugated secondary antibodies (Invitrogen). E. faecalis sigma factor (σA) was detected using anti-RNA polymerase sigma 70 monoclonal antibody 2G10 (Abcam) with goat anti-mouse IgG HRP-conjugated secondary antibodies (Invitrogen). Total cell lysates of E. faecalis strains for immunoblot analysis were prepared from exponentially growing cells by digestion with 5 mg/ml lysozyme in lysozyme buffer (20 mM Tris [pH 8.0], 10 mM EDTA) for 20 min at 37°C prior to solubilization in Laemmli SDS sample buffer.