The Bacillus subtilis Extracytoplasmic Function σ Factor σ^V Is Induced by Lysozyme and Provides Resistance to Lysozyme^,

Abstract

Bacteria encounter numerous environmental stresses which can delay or inhibit their growth. Many bacteria utilize alternative σ factors to regulate subsets of genes required to overcome different extracellular assaults. The largest group of these alternative σ factors are the extracytoplasmic function (ECF) σ factors. In this paper, we demonstrate that the expression of the ECF σ factor σ^V in Bacillus subtilis is induced specifically by lysozyme but not other cell wall-damaging agents. A mutation in sigV results in increased sensitivity to lysozyme killing, suggesting that σ^V is required for lysozyme resistance. Using reverse transcription (RT)-PCR, we show that the previously uncharacterized gene yrhL (here referred to as oatA for O-acetyltransferase) is in a four-gene operon which includes sigV and rsiV. In quantitative RT-PCR experiments, the expression of oatA is induced by lysozyme stress. Lysozyme induction of oatA is dependent upon σ^V. Overexpression of oatA in a sigV mutant restores lysozyme resistance to wild-type levels. This suggests that OatA is required for σ^V-dependent resistance to lysozyme. We also tested the ability of lysozyme to induce the other ECF σ factors and found that only the expression of sigV is lysozyme inducible. However, we found that the other ECF σ factors contributed to lysozyme resistance. We found that sigX and sigM mutations alone had very little effect on lysozyme resistance but when combined with a sigV mutation resulted in significantly greater lysozyme sensitivity than the sigV mutation alone. This suggests that sigV, sigX, and sigM may act synergistically to control lysozyme resistance. In addition, we show that two ECF σ factor-regulated genes, dltA and pbpX, are required for lysozyme resistance. Thus, we have identified three independent mechanisms which B. subtilis utilizes to avoid killing by lysozyme.

INTRODUCTION

The majority of genes in actively growing bacteria are transcribed by RNA polymerase using the general “housekeeping” σ factor σ⁷⁰. Bacteria often utilize alternative σ factors to regulate subsets of genes required for specific environmental conditions (18). The largest group of these alternative σ factors are the extracytoplasmic function (ECF) σ factors (18, 39). ECF σ factors represent the “third pillar” of bacterial signal transduction and are often involved in response to extracytoplasmic stress (18, 39). ECF σ factors are members of the σ⁷⁰ family of σ factors and are characterized by the presence of only two regions of σ⁷⁰, regions 2 and 4.2 (18).

Bacillus subtilis encodes seven known ECF σ factors (18). ECF σ factors are often required for their own transcription; thus, the expression of an ECF σ factor promoter is often indicative of activity of the ECF σ factor (18, 39). The signals which induce the activity of several ECF σ factors are known. For instance, the expression of sigW is induced by antimicrobial peptides and pH change (6, 9, 14, 32, 45), while sigM expression is induced by inhibitors of cell wall biosynthesis, heat shock, paraquat, and ethanol stress (12, 40). Like sigM, sigX is induced by inhibitors of cell wall biosynthesis (8, 11, 21, 42). In addition, sigY is induced by nitrogen limitation (41) and ylaC is induced in response to hydrogen peroxide (36); however, the signals which induce the activity of σ^V and σ^Z are not known.

Lysozyme is a hydrolytic enzyme which cleaves the β-(1-4)-glycosidic bond between N-acetylmuramic acid (MurNAc) and N-acetyl-glucosamine (GlcNAc) of the polysaccharide backbone of peptidoglycan (PG) (7). Known lysozyme resistance mechanisms fall into two broad categories, modification of the cell wall or lysozyme inhibitors (7). There are two known mechanisms of PG modification which provide lysozyme resistance. Staphylococcus aureus and Enterococcus faecalis both utilize O-acetyltransferases to acetylate the C6-OH group of MurNAc (4, 19, 27). Bacillus cereus, Bacillus anthracis, and several Streptococcus species encode PG deacetylases which remove the acetyl group from GlcNAc, which leads to increased lysozyme resistance (16, 31, 34, 43).

In B. subtilis, several studies have identified genes which are regulated by σ^V; however, the signal(s) which induces σ^V activity has not been identified (3, 47). We had previously found that only lysozyme was capable of inducing the expression of a sigV homolog, csfV, in Clostridium difficile (20). Similarly, the expression of E. faecalis sigV is also induced by lysozyme (27). Here we present evidence that the expression of B. subtilis sigV is induced specifically by lysozyme stress. Increased expression of sigV is presumably due to increased activity of σ^V, as a sigV mutation blocks the lysozyme-dependent increase in P_sigV expression. Our data suggest that σ^V is required for lysozyme resistance. We show that the product of a previously uncharacterized gene, yrhL (here referred to as oatA for O-acetyltransferase), is part of the sigV-rsiV operon. The expression of oatA is dependent upon σ^V and is required for lysozyme resistance. Finally, we provide evidence that the phenotype of a sigV mutation is masked by the presence of other ECF σ factors, as a strain lacking sigV, sigX, and sigM has much lower lysozyme resistance. We also demonstrate that, in addition to OatA, d-alanylation of teichoic acids by Dlt and the putative penicillin binding protein PbpX contribute to lysozyme resistance in B. subtilis.

MATERIALS AND METHODS

Strain construction.

The strains used in this study are isogenic derivatives of PY79, a prototrophic derivative of B. subtilis strain 168, and are listed in Table 1 (48). B. subtilis competent cells were prepared by the previously described one-step method (46). The P_sigV-lacZ reporter at amyE was constructed by PCR amplifying the P_sigV region using CDEP273 and CDEP274. The resulting PCR product was digested with EcoRI and BamHI and cloned into pDG1661 (17) digested with the same sites to create pCE143. Other P_ecf-lacZ fusions were constructed in the same manner, i.e., P_sigX (oligonucleotides CDEP277 and CDEP278, resulting in pCE145); P_sigY (oligonucleotides CDEP279 and CDEP280, resulting in pCE146); P_sigZ (oligonucleotides CDEP281 and CDEP301, resulting in pCE147); P_sigM (oligonucleotides CDEP271 and CDEP272, resulting in pCE142); and P_ylaC (oligonucleotides CDEP302 and CDEP303, resulting in pCE148). The P_sigV-lacZ reporter at pyrD was constructed by PCR amplifying the P_sigV-lacZ region using CDEP322 and CDEP323. The resulting PCR product was digested with EcoRI and XbaI (blunt ended using Klenow [NEB]) and cloned into pPyrD-cm digested with EcoRI and BamHI (blunt ended using Klenow [NEB]) to create pCE307 (30).

Table 1.

Plasmids and strains used in this study

Plasmid or strain	Genotypea	Reference or sourceb
Plasmids
pCE208	pCRBlunt II lox-kan-lox
pCE292	pDR111 RfA	20
pCrePA	33
pCE142	pDG1661 PsigM-lacZ
pCE143	pDG1661 PsigV-lacZ
pCE145	pDG1661 PsigX-lacZ
pCE146	pDG1661 PsigY-lacZ
pCE147	pDG1661 PsigZ-lacZ
pCE148	pDG1661 PylaC-lacZ
pCE307	pPyrD-Kan PsigV-lacZ
pJH101	pEntrD-TOPO oatA+
pJH102	pCE292 oatA+
Strains
PY79	Prototrophic derivative of B. subtilis 168	48
CDE1546	PY79 pyrD::PsigV-lacZ (cat)
CDE1547	PY79 pyrD::PsigV-lacZ (cat) ΔsigV::kan
CDE1560	PY79 amyE::Phs-oatA+ (spc)
CDE1581	PY79 ΔoatA::kan
CDE1702	PY79 amyE::Phs-oatA+ (spc) ΔoatA::kan
CDE1571	PY79 ΔsigV-rsiV-oatA::kan
CDE1573	PY79 amyE::Phs-oatA+ (spc) ΔsigV-rsiV-oatA::kan
CDE1710	PY79 ΔsigV::kan
CDE1711	PY79 amyE::Phs-oatA+ (spc) ΔsigV::kan
CDE1712	PY79 ΔsigV
CDE1713	PY79 amyE::Phs-oatA+ (spc) ΔsigV
CDE1708	PY79 dltA::kan
CDE1709	PY79 pbpX::kan
CDE1717	PY79 sigM::kan	9
CDE1718	PY79 sigX::spc
CDE1719	PY79 sigY::erm	13
CDE1720	PY79 sigZ::kan
CDE1721	12
PY79 sigW::kan	10
CDE1722	PY79 ΔsigV sigY::kan
CDE1723	PY79 ΔsigV sigW::kan
CDE1724	PY79 ΔsigV sigM::kan
CDE1725	PY79 ΔsigV sigZ::kan
CDE1726	PY79 ΔsigV ylaC::kan
CDE1727	PY79 ΔsigV sigX::spc
CDE1778	PY79 ΔsigV sigM::kan sigX::spc
CDE1779	PY79 sigM::kan sigX::spc
CDE680	PY79 amyE::PsigM-lacZ (cat)
CDE702	PY79 amyE::PsigV-lacZ (cat)
CDE705	PY79 amyE::PsigX-lacZ (cat)
CDE708	PY79 amyE::PsigY-lacZ (cat)
CDE711	PY79 amyE::PsigZ-lacZ (cat)
CDE714	PY79 amyE::PylaC-lacZ (cat)
CDE717	PY79 amyE::PsigW-lacZ (cat)	14

^aUnless otherwise noted, all strains are isogenic derivatives of PY79. hs, hyper-spank.

^bThis study, unless otherwise noted.

The oatA⁺ clone was constructed by PCR amplifying oatA from B. subtilis using CDEP1005 and CDEP1006 and cloning into pEntrD-TOPO (Invitrogen), resulting in pJH101. For isopropyl-β-d-thiogalactopyranoside (IPTG)-inducible expression, the oatA⁺ gene was moved into pCE292 (20) using LR Clonase II (Invitrogen). The resulting plasmid, pJH102, places the IPTG-inducible hyperspac promoter upstream of oatA⁺. All plasmid constructs were confirmed by DNA sequencing (Iowa State DNA Sequencing). The resulting plasmids were integrated into the B. subtilis genome by transformation.

We used the long flanking homology PCR technique to create the ΔsigV::kan, ΔoatA::kan, and ΔsigV-rsiV-oatA::kan mutations (44) using the primers listed in Table S1 in the supplemental material. The deletion-insertion ΔoatA::kan was constructed by PCR amplifying the 5′-flanking region of ΔoatA with CDEP1015 and CDEP1028, while the 3′-flanking region was amplified using CDEP1029 and CDEP1030. The ΔsigV-rsiV-oatA::kan deletion-insertion was constructed by PCR amplifying the 5′-flanking region of ΔsigV-rsiV-oatA with CDEP1031 and CDEP1032, while the 3′-flanking region was amplified using CDEP1029 and CDEP1030. The ΔsigV::kan deletion-insertion was constructed by PCR amplifying the 5′-flanking region of sigV with CDEP1031 and CDEP1032, while the 3′-flanking region was amplified using CDEP1070 and CDEP1012. The resulting PCR products were then used as primers to amplify the kanamycin resistance cassette from plasmid pCE208 as previously described (44). The PCR products were then transformed into PY79 (46). The mutants were confirmed by PCR and genetic linkage. The kan cassette in the ΔsigV::kan mutation was removed by the expression of the site-specific recombinase cre from pCrePA (33), resulting in an in-frame deletion of sigV referred to as ΔsigV.

Medium supplements.

Antibiotics were used at the following concentrations: chloramphenicol, 5 μg/ml; erythromycin plus lincomycin, 1 μg/ml and 25 μg/ml; kanamycin, 5 μg/ml; spectinomycin, 100 μg/ml; tetracycline, 10 μg/ml; ampicillin, 100 μg/ml. The β-galactosidase chromogenic indicator 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) was used at a concentration of 100 μg/ml. Isopropyl-β-d-thiogalactopyranoside (IPTG) was used at a final concentration of 1 mM.

Testing of compounds for increased expression of P_sigV-lacZ.

The strains containing P_sigV-lacZ were grown to an optical density at 600 nm (OD₆₀₀) of 2 and diluted 1:100 in 3 ml LB top agar (0.75% agar) containing X-Gal (100 μg/ml) and spread on solid LB plus X-Gal medium and allowed to solidify. Whatman filter disks containing 10 μl of phosphomycin (5 mg/ml), cycloserine (5 mg/ml), tunicamycin (100 μg/ml), nisin (50 mg/ml), ramoplanin (1 mg/ml), vancomycin (1 mg/ml), cefoxitin (100 μg/ml), ampicillin (1 mg/ml), bacitracin (1 mg/ml), hen egg white lysozyme (20 mg/ml), polymyxin B (100 mg/ml), EDTA (0.5 M), deoxycholate (10%), or proteinase K (1 mg/ml) were then placed on the top agar and incubated for 16 h at 37°C.

β-Galactosidase assays.

Samples for β-galactosidase assays were obtained by spotting 20 μl of culture of the strain of interest grown to an OD₆₀₀ of 1.0 onto either LB or LB plus 2.5 μg/ml hen egg white lysozyme. The plates were incubated at 37°C for 6 h. The resulting cells were then collected and resuspended in Z buffer, and OD₆₀₀s were determined. The cells were then permeabilized using chloroform-SDS. β-Galactosidase activity was determined using a microtiter plate assay as previously described (38). β-Galactosidase activity units are defined as (μmol of ONP formed min⁻¹) × 10³/(OD_{600 ×} ml of cell suspension) and are reported as means ± standard deviations. Experiments were performed in both technical and biological triplicates.

Determination of MIC.

The MIC was determined by diluting overnight cultures 1:100 and growing cells to an OD₆₀₀ of 1.5 in LB plus 1 mM IPTG. The cultures were corrected to an OD₆₀₀ of 0.8 and diluted 1:50 in LB plus 1 mM IPTG. The cells were inoculated into 200 μl of LB plus 1 mM IPTG containing serial 2-fold dilutions of hen egg white lysozyme ranging from 100 μg/ml to 0.10 μg/ml in a round-bottom 96-well plate. The cells were grown for 16 h at 37°C. Growth was defined as an OD₆₀₀ of greater than 0.05. All assays were performed using all of the listed strains in triplicate on the same day, and each MIC determination was performed on three different days.

RNA isolation and reverse transcription (RT)-PCR analysis of sigV operon.

Total RNA was prepared as previously described from cells grown to an OD₆₀₀ of 0.8 and then for an additional 15 min in the presence or absence of 1.25 μg/ml hen egg white lysozyme (5). To generate cDNA from RNA samples, we used Superscript III according to the manufacturer's protocol (Invitrogen). For RT-PCR, we used 2 pmol of the specified primers (see Table S1 in the supplemental material) in PCRs using 30 cycles of denaturation (94°C, 30 s), annealing (48°C, 30 s), and extension (72°C, 60 s). For real-time quantitative RT (qRT)-PCR experiments, RT reaction mixtures were diluted 1:5 in water. For each well, 5 μl of sample was added to 10 μl of Sybr green master mix (Applied Biosystems) and 5 μl of gene-specific primers from Table S1 (2 × 2.5 μM). Each sample was assayed in technical triplicate. Data were normalized to RNA levels of the B. subtilis housekeeping gene rpoB. Experiments were also performed with three biological replicates.

RESULTS

Lysozyme induces sigV expression.

Many ECF σ factors respond to different cell envelope stresses (18). The signals which induce the activities of several ECF σ factors, including σ^V, have not been determined (49). We had previously found that lysozyme, but not bacitracin or nisin, induced the expression of csfV, a sigV homolog, in C. difficile (20). To identify the inducing signal of σ^V, we tested a number of compounds known to induce cell envelope damage. These compounds included antibiotics which block different steps in PG biosynthesis: phosphomycin, cycloserine, tunicamycin, nisin, ramoplanin, vancomycin, cefoxitin, ampicillin, and bacitracin. We also tested compounds which can disrupt the cell envelope, including polymyxin B, proteinase K, EDTA, deoxycholate, and lysozyme (hen egg white). To identify compounds which could induce sigV expression, we performed diffusion assays by saturating filter disks with the various compounds and placing them on a lawn of cells harboring a P_sigV-lacZ reporter. These plates also contained X-Gal as a color indicator for the lacZ reporter. We found that only lysozyme was able to induce the expression of the P_sigV-lacZ fusion construct which can be observed as a blue halo surrounding the disk containing lysozyme (Fig. 1A). None of the other compounds tested affected sigV expression (Fig. 1A). It was previously demonstrated that σ^V induces its own transcription (3, 49). To determine if σ^V is required for induction by lysozyme, we constructed a strain containing P_sigV-lacZ and a sigV::kan insertion-deletion mutation. We observed that in the absence of σ^V, expression of P_sigV-lacZ was not induced in the presence of lysozyme (Fig. 1B).

Fig. 1.

Expression of sigV is specifically induced by lysozyme stress. The strains contained P_sigV-lacZ and were either wild type (CDE1546) or sigV mutant (CDE1547). Cells were grown to an OD₆₀₀ of 2, placed in LB top agar containing X-Gal, spread on solid LB medium plus X-Gal, and allowed to solidify. Filter disks containing phosphomycin (Pho; 10 μl of 5-mg/ml concentration), cycloserine (Cyc; 10 μl of a 5-mg/ml concentration), tunicamycin (Tun; 10 μl of a 100-μg/ml concentration), nisin (Nis; 10 μl of a 50-mg/ml concentration), ramoplanin (Ram; 10 μl of a 1-mg/ml concentration), vancomycin (Van; 10 μl of a 1-mg/ml concentration), cefoxitin (Cef; 10 μl of a 100-μg/ml concentration), ampicillin (Amp; 10 μl of a 1-mg/ml concentration), bacitracin (Bac; 10 μl of a 1-mg/ml concentration), hen egg white lysozyme (Lys; 10 μl of a 20-mg/ml concentration), polymyxin B (Poly; 10 μl of a 100-mg/ml concentration), EDTA (10 μl of a 0.5 M concentration), deoxycholate (Deo; 10 μl of a 10% concentration), or proteinase K (Pro K; 10 μl of a 1-mg/ml concentration) were then placed on the top agar and incubated for 16 h at 37°C.

To quantitate the induction of sigV by lysozyme, we spotted both the wild-type P_sigV-lacZ strain (CDE1546) and the P_sigV-lacZ sigV::kan mutant strain (CDE1547) on LB plates with or without lysozyme (2.5 μg/ml). After 6 h of growth, the cells were resuspended in Z buffer and the β-galactosidase activity levels were assayed as already described. We found that in sigV⁺ cells, the expression of P_sigV-lacZ was induced approximately 15-fold when the cells were grown in the presence of lysozyme compared to that in the no-lysozyme control (Table 2). In contrast, a strain lacking σ^V showed no difference in the expression of P_sigV-lacZ in the presence or absence of lysozyme (Table 2). Taken together, these data suggest that the expression of σ^V activity is specifically induced by lysozyme and that this induction requires σ^V.

Table 2.

β-Galactosidase activities with and without lysozyme

Strain	Relevant genotype	Mean β-galactosidase activitya ± SD
		LB	LB + lysozyme
CDE1546	pyrD::PsigV-lacZ	1.55 ± 0.22	21.23 ± 0.93
CDE1547	pyrD::PsigV-lacZ ΔsigV::kan	3.12 ± 0.14	3.93 ± 0.32

^aExpressed as (μmol of ONP formed min⁻¹) × 10³/(OD₆₀₀ × ml of cell suspension).

σ^V is the only lysozyme-inducible ECF σ factor.

The seven ECF σ factors of B. subtilis have been shown to induce their own expression (2). Several other ECF σ factors are induced by cell envelope stresses, including sigX, sigW, and sigM (9, 11, 32). Thus, we sought to determine if any of the other ECF σ factors are induced in response to lysozyme stress. We spotted cells containing lacZ fusions to each of the ECF σ factor promoters onto LB or LB plus lysozyme (2.5 μg/ml). After 6 h of growth, the cells were resuspended in Z buffer and the β-galactosidase activity levels were assayed as already described. We found, as previously observed, that the expression of P_sigV-lacZ was induced 13-fold when cells were grown in the presence of lysozyme (Fig. 2). In contrast, none of the remaining six ECF σ factors showed any significant increase in expression when cells were grown in the presence of lysozyme (Fig. 2). This suggests that only sigV is specifically induced by lysozyme-mediated stress.

Fig. 2.

sigV is the only lysozyme-induced σ factor. The strains used contained an ECF σ factor promoter-lacZ fusion integrated at amyE and were otherwise wild type. Shown are results obtained with sigV (amyE:: P_sigV-lacZ CDE702), sigM (amyE::P_sigV-lacZ CDE680), sigX (amyE:: P_sigX-lacZ CDE705), sigV (amyE::P_sigV-lacZ CDE1547), sigY (amyE::P_sigY-lacZ CDE708), sigZ (amyE::P_sigZ-lacZ CDE711), ylaC (amyE::P_sigV-lacZ CDE714), and sigW (amyE::P_sigW-lacZ CDE717) mutant cells. Cells were grown to an OD₆₀₀ of 1.0 and spotted onto plates containing either LB or LB plus 2.5 μg/ml hen egg white lysozyme. The plates were incubated at 37°C for 6 h. Cells were recovered from the plates and resuspended in Z buffer, and β-galactosidase activities were determined as described in Materials and Methods. The results shown are averages of three experiments, and the error bars denote standard deviations. The data plotted are fold induction by lysozyme (lysozyme β-galactosidase activity/no-lysozyme β-galactosidase activity).

σ^V is required for maximum lysozyme resistance.

Because the expression of sigV is induced by lysozyme, we next determined if σ^V is also required for resistance to lysozyme. We constructed a B. subtilis strain with sigV deleted and then compared the MICs for wild-type B. subtilis (PY79) and the sigV mutant (CDE1712). We found that the sigV mutation resulted in a 2-fold decrease in lysozyme resistance compared to that of the wild type (Fig. 3). We also compared the zones of inhibition due to lysozyme by placing disks soaked with 10 μl of a 20-mg/ml solution of hen egg white lysozyme on LB plates with top agar lawns containing either wild-type or sigV mutant B. subtilis. We observed a 10-mm zone of inhibition when the lysozyme was spotted on top agar lawns containing wild-type B. subtilis and a 13-mm zone of inhibition when it was placed on the sigV mutant. These data suggest that σ^V is involved in B. subtilis resistance to lysozyme.

Fig. 3.

Roles of σ^V and oatA in resistance to lysozyme. The MIC of lysozyme was determined for wild-type (WT; strain PY79) and sigV (CDE1712; ΔsigV), oatA (CDE1581; ΔoatA::kan), and sigV-oatA (CDE1571; ΔsigV-rsiV-oatA::kan) mutant strains without and with an IPTG-inducible copy of oatA⁺ (amyE::P_hs-oatA⁺), as follows: wild-type oatA⁺ (CDE1713; amyE::P_hs-oatA⁺) sigV oatA⁺ (CDE1713; amyE::P_hs-oatA⁺ ΔsigV), oatA-oatA⁺ (CDE1581; amyE::P_hs-oatA⁺ ΔoatA::kan), and sigV-oatA oatA⁺ (CDE1571; amyE::P_hs-oatA⁺ ΔsigV-rsiV-oatA::kan). The MIC is defined as no growth (OD₆₀₀ less than 0.05) after 16 h at 37°C in a 96-well plate. The MIC experiments were performed in technical triplicates and repeated on different days. The data shown are from one representative experiment. The MIC varied slightly between days or experiments; however, the fold changes always remained the same.

sigV is in a four-gene operon.

To identify the genes under σ^V control which were responsible for lysozyme resistance, we analyzed the regulon of σ^V which had previously been determined (3, 49). In one of the two microarray studies, yrhL (here referred to as oatA) was induced by σ^V (3). The oatA gene is predicted to encode a homolog of OatA, the O-acetyltransferase from S. aureus (39% identical and 59% similar), which is involved in resistance to lysozyme (4). The oatA gene is 110 bp downstream of the anti-σ factor rsiV. This intergenic region was annotated to contain a terminator between rsiV and oatA. To determine if oatA was indeed part of the sigV operon, we used reverse transcriptase PCR. RNA was isolated from wild-type B. subtilis grown in the presence or absence of 2.5 μg/ml lysozyme. This RNA was converted into cDNA using reverse transcriptase. The resulting cDNA was used as the template in PCRs with primers which anneal to the 3′ end of a gene and the 5′ end of the neighboring gene and amplify the intergenic region. Each pair of primers was tested in a PCR using genomic DNA (gDNA) as the template to ensure that the primers were capable of amplifying DNA fragments of the predicted size (data not shown). To demonstrate that the RNA sample was not contaminated with gDNA, we included controls where no reverse transcriptase was added. In the absence of reverse transcriptase, we did not detect any PCR products (Fig. 4A). We also included a negative control which spans an intergenic region upstream of the first open reading frame. We were able to PCR amplify a product by using a primer in the 3′ end of sigV and a primer in the 5′ end of rsiV (Fig. 4A). We were also able to amplify the regions between the rsiV and oatA genes and between the oatA and yrhK genes (Fig. 4A). In the absence of lysozyme, we are unable to detect any of the intergenic RT-PCR products, suggesting that, in the absence of lysozyme, expression is very low and the products detected are likely the result of P_sigV (data not shown). These data demonstrate that B. subtilis encodes an mRNA transcript that includes sigV, rsiV, oatA, and yrhK.

Fig. 4.

oatA is part of the sigV operon and is lysozyme inducible. (A) sigV is in a four-gene operon. The top portion shows the putative sigV operon structure. Capital letters represent PCR primer pairs (A, CDEP273-CDEP274; B, CDEP1015-CDEP978; C, CDEP1011-CDEP1012; D, CDEP1013-CDEP1014) in which the mRNA from the same samples had (+) or had not (−) been treated with reverse transcriptase (RT) as described in Materials and Methods. The DNA was then separated on a 1.5% agarose gel. (B) Expression of oatA is induced by lysozyme stress. The oligonucleotides used in the qRT-PCR for rsiV (CDEP1107 and CDEP1108) and oatA (CDEP1021 and CDEP1022) are listed in Table S1 in the supplemental material. RNA levels were corrected to the standard rpoB (CDEP1017 and CDEP1018). Shown are the mean ± standard deviation of the relative change in RNA levels 15 min following exposure to a subinhibitory concentration of lysozyme (1.25 μg/ml hen egg white lysozyme). The qRT experiments were performed in technical triplicates and repeated on different days. The data shown are from one representative experiment. wt, wild type.

Expression of oatA is lysozyme inducible and σ^V dependent.

Since P_sigV is induced by lysozyme and oatA appears to be in an operon with sigV, we sought to determine if oatA expression is also lysozyme inducible. RNA was isolated from wild-type B. subtilis grown in the presence or absence of 1.25 μg/ml lysozyme, and the RNA transcript levels were determined by qRT-PCR. RNA levels were normalized to RNA transcript levels of the housekeeping gene rpoB. When we analyzed RNA from wild-type B. subtilis, we found that the RNA levels of sigV were increased approximately 148-fold in the presence of lysozyme compared to those in cells grown in the absence of lysozyme (Fig. 4B). Similarly, we found that the expression of oatA was induced approximately 65-fold in response to lysozyme stress in wild-type B. subtilis (Fig. 4B). These data demonstrate that the expression both rsiV and oatA is induced by lysozyme.

To determine if the lysozyme induction of oatA is dependent upon σ^V, we compared the expression of both oatA and rsiV in a strain containing an in-frame, nonpolar deletion of sigV (CDE1712). We then isolated RNA from ΔsigV mutant cells grown in the presence or absence of lysozyme and measured RNA transcript levels using qRT-PCR. When grown in the presence of lysozyme, the ΔsigV mutant strain did not exhibit an increase in either rsiV or oatA expression compared with that in the wild type (Fig. 4B). This demonstrates that σ^V is required for lysozyme induction of oatA. Since oatA is in the sigV operon, lysozyme induction of oatA likely occurs via increased transcription from the sigV promoter. We cannot rule out the possibility of a second σ^V-dependent promoter in the intergenic region between rsiV and oatA; however, we were unable to identify a promoter in this region which matches the σ^V consensus promoter.

OatA is required for resistance to lysozyme.

Since the expression of oatA is induced by σ^V and oatA is homologous to a known lysozyme resistance gene from S. aureus (4), we tested if OatA could confer σ^V-dependent lysozyme resistance. We constructed ΔsigV (CDE1712), ΔoatA (CDE1581), and ΔsigV-rsiV-oatA (CDE1571) mutant B. subtilis strains. We then determined the MICs for each of the resulting strains and wild-type B. subtilis (PY79). We found that the ΔsigV mutation resulted in a 2-fold decrease in lysozyme resistance compared to that of the wild type (Fig. 3). Similarly, strains containing either a ΔsigV-rsiV-oatA or a ΔoatA deletion also showed a 2-fold decrease in lysozyme resistance (Fig. 3).

In order to confirm our observation that OatA is required for σ^V-dependent lysozyme resistance, we constructed a plasmid expressing oatA⁺ from an IPTG-inducible promoter. We introduced this construct into the wild-type (CDE1560) and the ΔsigV (CDE1713), ΔoatA (CDE1702), and ΔsigV-rsiV-oatA (CDE1573) mutant B. subtilis strains and determined the MIC for each of the resulting strains in the presence of IPTG. We observed in each case that the expression of oatA⁺ led to a 2-fold increase in resistance to lysozyme in all of the strain backgrounds (Fig. 3). This may be due to the fact that the cells were grown in the presence of inducer (IPTG) prior to lysozyme stress, which likely leads to increased O-acetylation and resistance to lysozyme. In contrast wild-type cells likely express very little oatA prior to lysozyme stress. Thus, OatA is involved in σ^V-mediated lysozyme resistance in B. subtilis.

σ factor redundancy masks a critical role of σ^V in resistance to lysozyme.

B. subtilis encodes six ECF σ factors in addition to σ^V (2). Several of these ECF σ factors are induced by compounds which damage the PG (9–11, 32). Given this knowledge, we investigated whether other ECF σ factors contribute to lysozyme resistance. We constructed strains which harbored a mutation in each of the remaining six ECF σ factors in the presence or absence of sigV. We then determined the lysozyme MIC for the resulting strains. We found that of the ECF σ factor mutants, only the sigV (CDE1712) and sigX (CDE1718) mutants showed increases (2-fold) in lysozyme sensitivity compared to wild-type B. subtilis (Fig. 5). However, when we constructed double mutants with sigV and each ECF σ factor, we found that several double mutants exhibited increased sensitivity to lysozyme (Fig. 5). One curious observation to note is that a sigV sigZ double mutant showed increased resistance to lysozyme compared to that of a sigV mutant (Fig. 5). Based upon microarray analysis of the sigZ regulon, there is no obvious mechanism to explain this effect. The strains with the double mutations sigX sigV (8-fold; CDE1727), sigM sigV (4-fold; CDE1724), and sigW sigV (4-fold; CDE1723) showed increased lysozyme sensitivity compared with that of the wild type (Fig. 5). Similarly, a sigX sigM double mutant (CDE1779) showed 4-fold-increased lysozyme sensitivity (Fig. 5). Interestingly, the sigX sigM sigV triple mutant (CDE1778) exhibited a 64-fold increase in lysozyme sensitivity. It also suggests that σ^V, σ^X, σ^W, and σ^M regulate mechanisms of lysozyme resistance and that the involvement of σ^V in lysozyme resistance is masked by redundant mechanisms regulated by other ECF σ factors.

Fig. 5.

The combination of ECF σ factors mutations increases lysozyme sensitivity. The MIC of lysozyme was determined for the wild type (WT; strain PY79); for sigV (CDE1712; ΔsigV), sigW (CDE431; sigW::kan), sigX (CDE1718; sigX::spc), sigM (CDE1717; sigM::kan), sigY (CDE1719; sigY::kan), sigZ (CDE1720; sigZ::kan), and ylaC (CDE1721; ylaC::kan) single ECF σ factor mutant strains; for sigV sigW (CDE1723; ΔsigV sigW::kan), sigV sigX (CDE1727; ΔsigV sigX::spc), sigV sigM (CDE1724; ΔsigV sigM::kan), sigV sigY (CDE1722; ΔsigV sigY::kan), sigV sigZ (CDE1725; ΔsigV sigZ::kan), sigV ylaC (CDE1726; ΔsigV ylaC::kan) and sigM sigX (CDE1779; sigM::kan sigX::spc) ECF σ factor double mutant strains; and for a sigM sigV sigX (CDE1778; ΔsigV sigM::kan sigX::spc) ECF σ factor triple mutant strain. The MIC was defined was no growth (OD₆₀₀ less than 0.05) after 16 h at 37°C in a 96-well plate. The MIC experiments were performed in technical triplicates and repeated on different days. The data shown are from one representative experiment. The MIC varied slightly between days or experiments; however, the fold changes always remained the same.

PbpX and DltA contribute to lysozyme resistance.

Based upon our observation that a mutant lacking sigV, sigX, and sigM has significantly increased sensitivity to lysozyme, we sought to identify other genes controlled by these ECF σ factors which are required for lysozyme resistance. Expression of the dlt operon is induced by several ECF σ factors, including σ^W, σ^V, and σ^X (3, 11, 49). Mutations in dltA have been found to lead to increased lysozyme sensitivity in other organisms (1, 26, 27). We constructed a mutation in dltA and measured the lysozyme MIC for the resulting strain. We found that a dltA mutation increased lysozyme sensitivity 4-fold (Fig. 6).

Fig. 6.

Two ECF σ factor-regulated genes, dltA and pbpX, are required for maximal lysozyme resistance. The MIC of lysozyme was determined for the wild type (WT; strain PY79) and sigV (CDE1712; ΔsigV), dltA (CDE1709; ΔdltA::kan), oatA (CDE1581; ΔoatA::kan), and pbpX (CDE1709; ΔpbpX::kan) mutant strains. The MIC was defined as no growth (OD₆₀₀ less than 0.05) after 16 h at 37°C in a 96-well plate. The MIC experiments were performed in technical triplicates and repeated on different days. The data shown are from one representative experiment. The MIC varied slightly between days or experiments; however, the fold changes always remained the same.

According to previous microarray analysis, the expression of the pbpX gene is induced by both σ^V and σ^X (11, 49). The pbpX gene encodes a putative penicillin-binding protein, but a mutation in pbpX does not produce any reported cell growth defect (37). We constructed a pbpX mutant and found that a mutation in pbpX led to an 8-fold decrease in lysozyme resistance (Fig. 6). This suggests that dltA and pbpX, in addition to oatA, contribute significantly to lysozyme resistance in B. subtilis.

DISCUSSION

As a soil organism, B. subtilis encounters a number of extracellular stresses which inhibit its growth or survival. As a result, the bacterium has evolved a complex and comprehensive array of stress-sensing systems. B. subtilis encodes seven known ECF σ factors which sense various environmental signals. The signals required for the activation of some of these ECF σ factors are known, while the signals for the activation of other ECF σ factors remain unclear. In some cases, the signals or the effectors of different σ factor systems overlap. For example, σ^W, σ^X, σ^M, and σ^Y can be activated by some of the same antibiotics and environmental stresses, many of which perturb the cell envelope (9, 11, 12, 28, 32, 45).

Lysozyme specifically activates σ^V.

In contrast to other σ factors, which are activated by multiple extracellular stresses, our data indicate that the expression of sigV is induced specifically by lysozyme-mediated stress and not by antibiotics which block PG biosynthesis. This induction is dependent upon σ^V, suggesting that σ^V is activated upon lysozyme stress. It seems remarkable that lysozyme specifically induces sigV expression but inhibition of PG synthesis does not affect sigV expression. The latter observation is consistent with a previous study in which numerous other cell wall-acting antibiotics were tested and none were able to induce sigV expression (49).

This raises the intriguing question of how and, more precisely, what the σ^V system is sensing. Since lysozyme inhibits bacterial growth by cleaving between the MurNAc and GlcNAc subunits of the cell wall polysaccharide backbone, one alternative is that σ^V is activated by the presence of PG breakdown products. However, the addition of antibiotics should lead to an accumulation of PG biosynthesis products. In particular, ramoplanin, which is reported to block PG biosynthesis by inhibiting transglycosylases (15), should result in the accumulation of MurNAc-GlcNAc; however, ramoplanin does not induce the σ^V system (Fig. 1). If σ^V is sensing cell wall damage, then why do reagents which process the cell wall in a similar fashion not activate σ^V? While we do not yet know the precise signal that leads to the activation of σ^V, it is clear that lysozyme is critical for σ^V activation.

Lysozyme resistance mechanisms.

The σ^V ECF σ factor is conserved in other Gram-positive bacteria, including the pathogens C. difficile and E. faecalis (20, 27). In these organisms, the sigV homologs are also induced by lysozyme (20, 27). Although the signal transduction systems are conserved in these organisms, the targets of the ECF σ factors may vary. In E. faecalis, σ^V induces the expression of oatA. Similarly, in B. subtilis, σ^V is reported to regulate the expression of oatA, pbpX, and the dlt operon (3, 49). In contrast, there is no OatA homolog present in C. difficile, suggesting that σ^V homologs may induce other as-yet-uncharacterized mechanisms of lysozyme resistance.

Interestingly, S. aureus, which lacks a σ^V homolog, also utilizes similar mechanisms of lysozyme resistance, including the increased expression of oatA and dltA (26). However, in S. aureus, the expression of the dlt operon is controlled by the two-component regulatory system GraRS (19, 26). It was recently shown that an S. aureus dltA or oatA mutant shows only a modest increase in lysozyme sensitivity. However, strains that lack both OatA and DltA have a significant increase in lysozyme sensitivity (19).

Here we show that B. subtilis protects itself from lysozyme stress using three different proteins, OatA, DltA, and PbpX, which likely protect the cell from lysozyme using different mechanisms. The specificity of protection of these methods is varied. OatA is predicted to be an O-acetyltransferase, and thus we predict that it acetylates the PG, which is a very specific method of lysozyme resistance. In contrast, the dlt operon alters the charge of the lipoteichoic acids and is often associated with increased resistance to cationic antimicrobial peptides and lysozyme (23–26, 29, 35). Thus, this mechanism is a more general means of protection and likely increases resistance to a number of extracellular stresses. We are currently investigating how PbpX increases B. subtilis resistance to lysozyme. It remains unclear if PbpX acts as a more general antimicrobial peptide resistance mechanism or as a lysozyme-specific protection mechanism.

ECF σ factor redundancy.

We observed synergistic effects when we combined ECF σ factor mutations in B. subtilis. In particular, we found that sigV, when combined with both sigM and sigX mutations, led to a 64-fold increase in lysozyme sensitivity. This suggests that there may be a significant overlap of the regulons of these σ factors. It has been established that σ^W, σ^X, and σ^M recognize similar promoters (22, 28). In fact, σ^W, σ^X, and σ^M have a consensus −35 region of GAAC and a consensus −10 region of CGT (28). Analysis of the promoter regions for σ^V-regulated genes revealed a consensus −35 region containing GAAC and a consensus −10 region containing CGTCT (49). This suggests that σ^V can recognize many of the same promoter regions recognized by σ^W, σ^X, and σ^M and likely explains the relatively modest effect of single mutations on lysozyme resistance while a sigMXV triple mutant shows a much larger increase in lysozyme sensitivity. This hypothesis is supported by the fact that a single mutation in either pbpX or dltA results in a greater decrease in lysozyme resistance than does any single ECF σ factor mutation (Fig. 5 and 6). These data seem to suggest that σ^WXM and σ^V likely recognize similar promoters, and thus we envision a great deal of cross regulation among the ECF σ factor regulons.