In Vitro Studies of Peptidoglycan Binding and Hydrolysis by the Bacillus anthracis Germination-Specific Lytic Enzyme SleB

Abstract

The Bacillus anthracis endospore loses resistance properties during germination when its cortex peptidoglycan is degraded by germination-specific lytic enzymes (GSLEs). Although this event normally employs several GSLEs for complete cortex removal, the SleB protein alone can facilitate enough cortex hydrolysis to produce vulnerable spores. As a means to better understand its enzymatic function, SleB was overexpressed, purified, and tested in vitro for depolymerization of cortex by measurement of optical density loss and the solubilization of substrate. Its ability to bind peptidoglycan was also investigated. SleB functions independently as a lytic transglycosylase on both intact and fragmented cortex. Most of the muropeptide products that SleB generates are large and are potential substrates for other GSLEs present in the spore. Study of a truncated protein revealed that SleB has two domains. The N-terminal domain is required for stable peptidoglycan binding, while the C-terminal domain is the region of peptidoglycan hydrolytic activity. The C-terminal domain also exhibits dependence on cortex containing muramic-δ-lactam in order to carry out hydrolysis. As the conditions and limitations for SleB activity are further elucidated, they will enable the development of treatments that stimulate premature germination of B. anthracis spores, greatly simplifying decontamination measures.

The bacterial endospore contains several integument layers that contribute to making a structure that is dehydrated, dormant, and highly resistant to environmental insults (39). One of these layers is a modified stratum of peptidoglycan (PG) termed the cortex, which comprises >80% of the total spore PG and is responsible for maintaining spore dehydration, dormancy, and wet heat resistance (26, 39). Like PG from vegetative cells, the cortex consists of repeating disaccharide units of β-1,4-linked N-acetylglucosamine (NAG) and N-acetylmuramic acid (NAM). The NAM residues have peptide side chains which become covalently bonded with peptide chains from neighboring strands, resulting in a cross-linked network surrounding the entire spore (43). Cortex differs from vegetative PG by a modification where as much as 50% of the NAM loses its peptide and is converted to muramic-δ-lactam, which contributes to an ∼75% reduction in cross-linking (26, 35, 43). In order for the spore to complete germination, this thick, constraining layer of PG must be degraded by germination-specific lytic enzymes (GSLEs), which recognize the cortex-specific muramic-δ-lactam residues (24, 29, 34).

A typical germination response begins when spores detect germinant compounds in the environment through the use of receptors located on the inner spore membrane (12, 33, 40). Three major events then transpire. The first is that the spore core starts to rehydrate as water moves inward (38). The second event, likely coupled with the first, is transport of ions and dipicolinic acid (DPA) out of the core (9, 22, 30, 32). The third major step is the depolymerization of cortex by GSLEs and the release of muropeptides into the surrounding environment (30). At this point, the spore core is free to fully rehydrate, resume metabolism, and proceed toward a vegetative cell cycle; consequently, the spore has now lost its resistance properties (38, 39). Exactly how GSLEs are held inactive in the dormant spore and later activated during germination is not fully understood, but it is clear that cortex hydrolysis by a GSLE is capable of causing germination without nutrients or the prior steps of water movement and ion release (29, 32, 36).

The endospore of Bacillus anthracis is known to be the infectious agent of all forms of the disease anthrax (28), and as a result, the destruction of spores is the most direct means to prevent infection. However, since dormant spores are so highly resistant, current methods of decontamination require the use of highly toxic and/or destructive compounds that have undesirable consequences (4, 20, 41). A recent U.S. Army-sponsored workshop proposed that a potential strategy toward eliminating the threat of spores would be to trigger germination at a high efficiency and then use typical antibacterial procedures to inactivate the resulting vulnerable cells (18). In this regard, GSLEs may be a suitable target for manipulation because of their potential to instigate premature spore germination.

B. anthracis has four GSLEs that participate in cortex degradation during spore germination, namely, SleB, CwlJ1, CwlJ2, and SleL (14, 16). However, none of these proteins have been investigated in vitro to conclusively define their roles and limitations during germination. SleB is a prime candidate for investigation because genetic analyses have shown that it can facilitate complete spore germination independent of the other known GSLEs (14, 17). This study demonstrates that purified SleB can carry out lytic transglycosylase activity on a variety of spore PG substrates. The roles of its predicted domains in protein-substrate interactions and in cortex lysis are also reported.

MATERIALS AND METHODS

Protein expression and purification.

The sleB open reading frame (ORF) lacking the first 32 codons was amplified from B. anthracis Sterne 34F2 genomic DNA by use of a forward primer with a 5′ end complementary to the tobacco etch virus (TEV) protease cleavage site region of pDEST-HisMBP (5′-GTGGAGAACCTGTACTTCCAGGGTTTTTCTAATCAAGTCATTCAAAGGG-3′) and a reverse primer with a 5′ end complementary to the attR2.1 region of pDEST-HisMBP (5′-CCACTTTGTACAAGAAAGTTGCATTGCTCTATTTACAGAAAATATGTTTC-3′). The sleB ORF lacking the first 124 codons was similarly amplified with a different forward primer (5′-GTGGAGAACCTGTACTTCCAGGGTTCTCAAAATAAAGGGACAAATGTTC-3′). Each PCR amplicon was inserted into pDEST-HisMBP (F. Schubot) by using a restriction-free (RF) cloning method as previously described (42) in order to construct His₆-MBP-SleB (pDPV385) and His₆-MBP-SleB_125-253 (pDPV386) fusion vectors for protein overexpression.

The fusion proteins were overproduced in Escherichia coli BL21(λDE3)(pLys^s Cm^r) (Novagen) grown in LB with 30 μg/ml chloramphenicol (Fisher) and 50 μg/ml ampicillin (Jersey Lab Supply). These cultures were grown with shaking (250 rpm) to saturation overnight at 37°C and then diluted 100-fold into 2 liters of fresh medium. When the cells reached an optical density at 600 nm (OD₆₀₀) of ∼1.0, the temperature was reduced to 10°C and isopropyl-β-d-thiogalactopyranoside was added to a final concentration of 1 mM. Fifteen hours later, the cells were recovered by centrifugation at 10,000 × g for 10 min at 4°C and stored at −80°C.

E. coli cell paste was suspended in ice-cold 50 mM NaCl, 50 mM Tris-HCl, 5% glycerol, and 30 mM imidazole (pH 7.5) (buffer A). The cells were lysed with a French press (Thermo Electron Corp.) at 1,000 lb/in² and centrifuged at 10,000 × g for 10 min at 4°C. The supernatant was further centrifuged at 40,000 × g for 1 h at 4°C, and the resulting supernatant was loaded onto a 5-ml Ni-Sepharose HisTrap HP affinity column (GE Healthcare) equilibrated with buffer A. The column was washed with 20 column volumes of buffer A and then eluted with a linear gradient of 30 to 500 mM imidazole in buffer A. Fractions containing either recombinant His₆-MBP-SleB or His₆-MBP-SleB_125-253 were pooled and dialyzed in buffer A. The proteins were then digested overnight with 1 μg of His₆-tagged TEV (S219V) protease (19) per 100 μg of fusion protein at 15°C. The resulting soluble tag-free protein, either SleB or SleB_125-253, was next applied to an SP Sepharose HiTrap SP XL column (GE Healthcare) equilibrated with buffer A containing no imidazole (buffer B). The sample was eluted with a linear gradient of 50 to 500 mM NaCl in buffer B. The peak fractions containing either SleB or SleB_125-253 were pooled and dialyzed in buffer B. Aliquots were frozen and stored at −80°C.

Preparation of peptidoglycan substrates.

Vegetative cell wall PG from Bacillus subtilis PS832 (wild type) was prepared as previously described (1). Briefly, insoluble cell walls were recovered from cell cultures by boiling in sodium dodecyl sulfate (SDS) followed by extensive washes. Covalently attached proteins were removed by protease treatment, and the vegetative cell walls were again washed before storage at a neutral pH.

B. subtilis strain 168 derivatives PS832 (wild type) and DPVB19 (cwlD::Sp) (34) were grown in 2× SG sporulation medium at 37°C without and with 100 μg/ml spectinomycin (Sigma), respectively. Sporulation, spore harvesting, and spore purification were performed as previously described (34). Spore sacculi were prepared as previously described (10). Briefly, spores were first killed and permeabilized by being suspended at an OD₆₀₀ of 30 in 1 ml of 50 mM Tris-HCl (pH 7.5), 1% SDS, and 50 mM dithiothreitol (DTT), and then they were boiled for 20 min and then stored at 4°C after warm water washes had removed all detectable SDS. Nucleic acids were removed by incubation in 1 ml 100 mM Tris-HCl (pH 7.5), 20 mM MgSO₄, 10 μg DNase I (Sigma), and 50 μg RNase A (Sigma) at 37°C for 2 h. Residual protein was degraded overnight with 100 μg trypsin (TRTPCK; Worthington) in 10 mM CaCl₂ at 37°C. Enzymatic digestion was ceased by boiling in 1% SDS, and the resulting spore sacculi were stored at 4°C after warm water washes had removed all detectable SDS (15).

Cortex fragments were prepared by breaking spore sacculi suspended at an OD₆₀₀ of 10 in 0.5 ml 50 mM Tris-HCl (pH 8.0) with 0.1-mm glass beads in a Wig-L-Bug amalgamator (Dentsply) at 3,800 rpm, using 10 30-s pulses. Glass beads were allowed to settle, and spore fragments were recovered from the supernatant and in washes of the glass. Hydrophobic coat layer fragments were removed from the PG fragments by CHCl₃ extraction. Extracted PG fragments were washed three times with water before storage at 4°C.

Enzymatic activity assays.

Hydrolysis of 0.2 OD₆₀₀ unit/ml spore sacculi or cortex fragments by 10 nM SleB, SleB_125-253, or lysozyme (Sigma) was assayed by monitoring the % OD₆₀₀ loss per minute (1-cm-light-path cuvette) at 25°C in a 1-ml reaction mixture containing 30 mM NaPO₄ (pH 7.0), 1 mM EDTA, and 1 mM DTT. A Tukey-Kramer honestly significant difference (HSD) analysis was used to compare OD₆₀₀ loss curves for all tested reactions.

The PGs from SleB and SleB_125-253 hydrolysis reactions were separated into pellet and supernatant fractions and assayed for muramic acid content as described previously (26). The muropeptide products of SleB and SleB_125-253 hydrolysis were prepared for reverse-phase high-pressure liquid chromatography (RP-HPLC) analysis as previously described (10). Briefly, the hydrolyzed substrates were separated into pellet and supernatant fractions. Lytic enzymes were inactivated with heat (supernatant) or with heat and detergent (pellet), and PG was purified. The PG material from the pellets and half of each supernatant fraction was then digested with the muramidase mutanolysin (Sigma). All fractions were reduced with NaBH₄ prior to HPLC separation.

PG-binding assay.

SleB binding to PG was quantified following incubation of 8 μM protein with cortex fragments at an OD₆₀₀ of 40 in 15 μl of 30 mM NaPO₄ (pH 7.0), 1 mM EDTA, and 1 mM DTT at 25°C for 1 min. Mixtures were separated by centrifugation into bound (pellet) and unbound (supernatant) fractions and were analyzed in a 15% SDS-PAGE gel. Gels were stained with Sypro Ruby protein gel stain (Sigma) and imaged with a Typhoon Trio variable-mode imager (GE Healthcare). Band densities were measured with ImageQuant TL, version 2005 (GE Healthcare), software.

RESULTS

Structural features of SleB and its overexpression and purification.

BAS2562 encodes SleB (17), which is a 253-amino-acid protein with a signal sequence and two putative domains (Fig. 1). The SignalP 3.0 server (5) indicates that SleB has a 99% likelihood of being a secreted protein, with a 95% probability of signal peptide cleavage between amino acid residues Ala-32 and Phe-33. This region has 100% identity to the SleB signal sequence present in Bacillus cereus, which was shown to be cleaved at the same position (31). The putative PG-binding domain (pfam01471), beginning at Gly-41 and ending at Leu-99, contains a tandem repeat sequence that is also observed in the SleB proteins from B. subtilis and B. cereus as well as in a GSLE from Clostridium perfringens (SleC) (25, 27, 31). This repeat sequence is also observed in other peptidoglycan hydrolases that have a variety of substrate specificities (31). The C-terminal domain, encompassing Lys-128 to Lys-253, belongs to a family of hydrolases (pfam07486). It is homologous to the sole putative domain of two other GSLEs in B. anthracis, CwlJ1 and CwlJ2, the former of which is capable of facilitating cortex hydrolysis independently (14, 16).

FIG. 1.

Domain architecture and purification of SleB. (A) Predicted domains are illustrated as gray boxes, with residue numbers indicating the amino acids at either end of each putative domain. Overexpressed proteins were truncated from the native protein at amino acid Phe-33 for SleB and Ser-125 for SleB_125-253. The amino termini of both purified proteins are glycine residues resulting from affinity tag removal by TEV protease. SS, signal sequence. (B and C) Gel electrophoresis (SDS-PAGE) of soluble fractions taken during SleB (B) and SleB_125-253 (C) protein purification as described in Materials and Methods. Lanes: 1, Ni-affinity eluate; 2, TEV protease digestion; 3, cation-exchange flowthrough; 4, cation-exchange eluate. In each panel, lanes 1 and 2 contain approximately 10 μg protein, and lane 3 contains approximately 1 μg protein. In panel B, lane 4 contains 5 μg protein, and in panel C, lane 4 contains 0.5 μg protein. The positions and masses (kDa) of molecular mass standards are indicated by arrows in the center. The positions of the following proteins are indicated by arrows: HMBP (His₆-MBP) (43 kDa), HMBP-SleB (67 kDa), HMBP-SleB_125-253 (57 kDa), TEV (27 kDa), SleB (24 kDa), and SleB_125-253 (14 kDa).

Full-length SleB was overexpressed without a signal sequence in E. coli to ease purification and to mimic the mature protein's active structure (Fig. 1). A truncated version of SleB containing only the C-terminal hydrolase domain was also overexpressed and was named SleB_125-253. Truncation between amino acids Pro-124 and Ser-125 was chosen because of this region's low probability of containing secondary structure and a high surface probability, as determined using Protean 7.2.1 from the DNASTAR Lasergene software package. Both constructs were expressed as protein fusions with a combined histidine and maltose-binding protein tag (His₆-MBP) at the amino terminus. After purification with Ni-affinity and cation-exchange chromatography, the tags were removed by TEV protease treatment; this left a single additional glycine residue at the amino terminus of each protein. When observed on an SDS-PAGE gel, both SleB and SleB_125-253 matched their expected masses of 24 kDa and 14 kDa, respectively (Fig. 1). After further purification with cation-exchange chromatography, both proteins were judged to be free of the His₆-MBP tags and were >95% pure (Fig. 1).

SleB and SleB_125-253 hydrolyze intact and fragmented cortex.

As observed previously for SleB purified from the exudate of germinating B. cereus spores (23), both SleB and SleB_125-253 were able to cause losses in OD₆₀₀, refractility, and heat resistance of coat-permeabilized spores of both B. anthracis and B. subtilis (data not shown). B. anthracis and B. subtilis spore PGs are virtually identical (10), so B. subtilis spores were used as a source of cortex due to their ease of purification.

Spore sacculi were prepared by chemical and enzymatic treatments but still contained significant amounts of core and coat proteins (35). The spore sacculi did not lose OD₆₀₀ when treated exogenously with a cortex fragment lytic enzyme, SleL (data not shown), indicating that the peptidoglycan layers were still intact (7, 21; E. A. Lambert and D. L. Popham, personal communication). In order to generate cortex fragments, spore sacculi were broken with glass beads and CHCl₃ extracted. Amino acid analyses indicated that the extraction was effective at removing 75% of residual protein from the fragmented PG preparation (data not shown), and consequently, the cortex fragment preparation was enriched for PG. Since spore sacculus preparations are estimated to be 43% PG and 57% protein (35), the total enrichment of PG within the cortex fragments was 2-fold higher than that within spore sacculi.

SleB and SleB_125-253 were able to decrease the OD₆₀₀ values of spore sacculi and cortex fragments (Fig. 2 and data not shown), although the OD₆₀₀ loss from fragments was less rapid than that experienced by spore sacculi. On average, SleB decreased the OD₆₀₀ values of cortical fragments 4-fold slower than those of spore sacculi (Tables 1 and 2); however, taking into account the PG enrichment of cortical fragments, this difference is expected to be 2-fold. In support of this estimate, the same assay was performed with half the cortical fragment substrate in order to negate the effects of PG enrichment, and the rate of OD₆₀₀ loss by SleB was, on average, 1.8-fold lower than that by spore sacculi (data not shown). Furthermore, this lower rate of OD₆₀₀ loss was proportional to a decrease of solubilized cortex material, as determined by RP-HPLC (data not shown).

FIG. 2.

Actions of SleB and SleB_125-253 on spore PG. Spore sacculi (A) or cortex fragments (B) were prepared from PS832 (filled symbols) or DPVB19 (open symbols) spores. These were then incubated at an OD₆₀₀ of 0.2 with 10 nM SleB (circles) or SleB_125-253 (squares) at 25°C in a final volume of 1 ml of 30 mM sodium phosphate buffer, pH 7.0, 1 mM EDTA, 1 mM DTT, and 0.1% Triton X-100. The OD₆₀₀ was monitored regularly, with mixing between measurements. Data shown are averages for three independent reactions. Error bars represent 1 standard deviation of the mean; in most cases, the error bars are too small to be visible. Control reactions in which substrate was left untreated never experienced a loss in OD₆₀₀ and are not shown for clarity.

TABLE 1.

SleB, SleB_125-253, and lysozyme activity on spore sacculi PG

Substrate	Enzyme	Rate of hydrolysisa,b	% PG solubilized in 30 minb,c
PS832	None	0.1 ± 0.1	NA
PS832	SleB	13.2 ± 0.4	96 ± 1
PS832	SleB125-253	1.4 ± 0.1	14 ± 6
PS832	Lysozyme	3.0 ± 0.1	ND
DPVB19	None	0.1 ± 0.1	NA
DPVB19	SleB	2.1 ± 0.1	NA
DPVB19	SleB125-253	0.1 ± 0.1	NA

^a% OD₆₀₀ loss/min with 10 nM enzyme in a 1-ml reaction mixture (1-cm-light-path cuvette).

^bWe used 0.2 OD₆₀₀ unit/ml of spore sacculi as a substrate. Values are averages determined for three independent experiments. Errors are 1 standard deviation of the mean.

^cSolubilized PG was calculated from the relative peak areas of muropeptides found in pellets and supernatants from 1-ml reaction mixtures after muramidase treatment. NA, not applicable (sample never produced detectable soluble material); ND, not determined.

Using either spore sacculi or cortex fragments, SleB caused an OD₆₀₀ loss rate that was significantly higher than that caused by SleB_125-253 (P < 0.0001) (Fig. 2A and B and Table 1). Even a 10-fold increase in SleB_125-253 concentration did not allow PG hydrolysis equivalent to that produced by SleB. The maximum rate of OD loss caused by 100 nM SleB_125-253 was still 2- to 4-fold lower than that caused by 10 nM SleB, and the maximum extent of OD loss was approximately 40% for SleB_125-253 and 60% for SleB. Neither protein caused any change in optical density when vegetative cell walls from B. subtilis cells were used as the substrate (data not shown). Although SleB decreased the OD₆₀₀ of spore sacculi by as much as 60%, no reactions were ever observed to cause a complete loss in OD₆₀₀. This was not unexpected, because the substrates also contained remnants of spore core and coat protein structures (35) as well as germ cell wall PG.

Lysozyme is a widely used and well-understood PG hydrolase and was thus used to gauge the relative activities of SleB and SleB_125-253. Under our reaction conditions for SleB, lysozyme demonstrated dose-dependent hydrolysis of B. subtilis spore sacculi (data not shown). Furthermore, lysozyme, similar to SleB, caused a maximum loss in OD₆₀₀ that never exceeded 75% (data not shown). At equal molar ratios, SleB proved to hydrolyze spore sacculi >4-fold faster than lysozyme (Table 1). However, when SleB lacked the putative PG-binding domain, as is the case with SleB_125-253, then hydrolysis was 2-fold slower than that by lysozyme.

Pellet and supernatant samples were collected from spore sacculi after treatment with either SleB or SleB_125-253 in order to determine if the OD₆₀₀ loss corresponded with cortex hydrolysis and hence muropeptide release into the supernatant. Pellets contained insoluble cortex that required muramidase treatment in order to be resolved by RP-HPLC, and as a result, muramidase products were detected in all treatment groups (Fig. 3 A to C and Table 3). However, after SleB treatment, the amount of insoluble cortex was reduced dramatically. As expected, the decrease of insoluble cortex by SleB coincided with an increase in soluble muropeptides in the reaction supernatant with (Fig. 3G and H) or without (Fig. 3D and E) muramidase treatment. Ultimately, SleB resulted in an average release of 96% of spore sacculus cortex into the supernatant (Table 1). The only muropeptides detected in supernatants untreated by muramidase were the anhydromuropeptides aG7a and aG7b (Fig. 3E and F and Table 3), which were the only two products previously found to depend on in vivo SleB action (6, 8, 16, 17). Exposure to SleB_125-253 led to an almost imperceptible decrease of insoluble cortex (Fig. 3A and C), but small amounts of lytic transglycosylase products were evident in the pellet (Fig. 3A and C) and in supernatants with (Fig. 3G and I) and without (Fig. 3D and F) muramidase. Only 14% of spore sacculus cortex was solubilized by SleB_125-253 (Table 1). The decrease in hydrolysis by SleB_125-253 is consistent with the reduced OD₆₀₀ loss (Fig. 2) compared to that produced by SleB.

FIG. 3.

HPLC separation of muropeptides released from spore sacculi digested with SleB or SleB_125-253. Spore sacculi were incubated with no enzyme (A, D, and G), SleB (B, E, and H), or SleB_125-253 (C, F, and I) as described in the legend to Fig. 2. After incubation for 30 min, the reaction mixtures were separated into pellet and supernatant fractions and prepared for analysis as described in Materials and Methods. Pellet fractions (A to C) and 50% of each supernatant sample (G to I) were digested with muramidase, reduced, and separated as previously described (26). The other 50% of each supernatant (D to F) was reduced and separated without muramidase digestion. Only 50% of the pellet fractions was analyzed. Peaks are numbered as in reference 16, but the initial “a” in the germination-specific peak names was omitted. The data shown are representative of one analysis; all samples were replicated in triplicate.

When cortex is degraded by GSLEs during spore germination, a variety of PG fragments are generated (3, 16). Many of these fragments are small enough to remain soluble during centrifugation, but they are too large, and presumably heterogeneous, to be resolved clearly by HPLC. Muramidase treatment during this analysis was used to break cortex down into fragments that were small enough to be resolved (eight sugars or fewer). Hence, for this analysis, it is reasonable to assume that most of the SleB cortex products were longer than eight sugars, because the majority of muropeptides were resolved only after muramidase digestion. When SleB lytic transglycosylase activity results in large muropeptides, they contain potentially numerous sites for muramidase hydrolysis, and thus one SleB product can be cleaved into several different products (16). When this occurs, the result is one SleB product for one or more muramidase products, and the ratio of these was calculated using data from the RP-HPLC chromatograms (Fig. 3). SleB cleaved 19% of the available NAM-NAG linkages, producing an average muropeptide length of 20 to 24 sugars. SleB_125-253 cleaved only 10% of those linkages, producing an average muropeptide length of 40 sugars. This is consistent with the significantly more efficient OD₆₀₀ loss induced by SleB than that induced by SleB_125-253 (Fig. 2) and suggests that the reason that lytic transglycosylase products appear in SleB_125-253 reaction pellets (Fig. 3C) is because many of these products are too large to be soluble.

To conduct a similar analysis of cortex fragment hydrolysis, pellet and supernatant samples were collected after treatment with either SleB or SleB_125-253. In this instance, muropeptides were not analyzed, but the total amount of solubilized NAM was quantified (Table 2). Only SleB was capable of solubilizing a significant amount of cortex fragments, but at a reduced efficiency relative to that calculated for spore sacculi (compare Tables 1 and 2). SleB_125-253 did not solubilize any detectable amount of the fragments, despite lowering the OD₆₀₀ of the suspension.

TABLE 2.

SleB and SleB_125-253 activity on cortex fragment PG

Substrate	Enzyme	Rate of hydrolysis (%)a,b,e	% PG solubilized in 90 minb,c,e	% Enzyme bindingd,e
PS832	None	0.1 ± 0.1	3 ± 2	NA
PS832	SleB	2.9 ± 0.2	70 ± 12	95 ± 4
PS832	SleB125-253	0.6 ± 0.1	5 ± 1	52 ± 7
DPVB19	None	0.2 ± 0.1	2 ± 1	NA
DPVB19	SleB	0.6 ± 0.1	3 ± 1	94 ± 3
DPVB19	SleB125-253	0.1 ± 0.1	1 ± 2	34 ± 3

^a% OD₆₀₀ loss/min with 10 nM enzyme in a 1-ml reaction mixture (1-cm-light-path cuvette).

^bWe used 0.2 OD₆₀₀ unit/ml of fragmented spore sacculi as a substrate.

^cSolubilized PG was calculated from the relative amounts of NAM found in pellets and supernatants from reaction mixtures.

^dEnzyme bound to 40 OD₆₀₀ units/ml of fragmented spore sacculi in a 15-μl reaction mixture.

^eValues are averages determined for three independent experiments. Errors are 1 standard deviation of the mean. NA, not applicable.

The hydrolase domain of SleB is dependent on muramic-δ-lactam for activity.

B. subtilis DPVB19 is a cwlD mutant that generates spore cortex without muramic-δ-lactam (34). Despite the fact that SleB was unable to change the OD₆₀₀ of vegetative cell walls (data not shown), it was capable of significantly decreasing the OD₆₀₀ values of both spore sacculi (P < 0.0001) and cortex fragments (P < 0.0001) from the DPVB19 strain (Fig. 2A and B). SleB_125-253 had no effect on the OD₆₀₀ of any DPVB19 PG substrate (Fig. 2A and B). No muropeptide production or NAM solubilization was detected when either SleB or SleB_125-253 was mixed with DPVB19 substrates (Table 2 and data not shown).

Both the N- and C-terminal domains of SleB participate in binding cortex peptidoglycan.

A simple peptidoglycan binding assay was used to investigate whether one of SleB's two domains is responsible for substrate binding or if they function cooperatively (Table 2). When both domains were present, at least 94% of SleB was able to bind cortical fragments both with and without muramic-δ-lactam. Only 52% of SleB_125-253 was bound to PS832 cortical fragments. In addition, SleB_125-253 showed a preference for cortex that contained muramic-δ-lactam, since its average binding to DPVB19 PG dropped to 34%. Less than 10% of a negative-control protein, bovine serum albumin, was associated with the cortical fragment PG pellet.

TABLE 3.

Muropeptide peak identification

Muropeptidea	Structureb
N	TS-TP
Q	TS-Ala
T	DS-TP × TP-TS
U	HS-TP-Ac
Y	HS-TP
Z	HS-Ala
aG7a	TS-TP anhydro
aG7b	TS-Ala anhydro

^aMuropeptide names are as previously published (16). Muropeptide names preceded by “a” indicate those generated by B. anthracis in order to differentiate them from those generated by other species.

^bAbbreviations: DS, disaccharide (NAG-NAM); TS, tetrasaccharide (NAG-MδL-NAG-NAM); HS, hexasaccharide (NAG-MδL-NAG-MδL-NAG-NAM); TP, tetrapeptide (Ala-Glu-Dpm-Ala); Ac, deacetylated glucosamine; ×, cross-link between two muropeptides. “Anhydro” indicates that the NAM at the reducing end is in the anhydro form as a result of lytic transglycosylase activity.

DISCUSSION

In this study, the B. anthracis SleB protein was overexpressed in E. coli and purified in two forms: as the native protein with both its N and C-terminal domains, which are putative PG-binding and hydrolytic domains, respectively, and as a truncated form that contained only the hydrolase domain. In vitro analyses of both proteins conclusively showed, as suggested by earlier genetics-based investigations (8, 17, 25), that SleB independently binds cortex PG and functions as a GSLE with lytic transglycosylase activity. SleB carries out this function independently, since no other protein was required to facilitate PG binding, generation of anhydromuropeptide products, or solubilization of cortex. These results also indicate that SleB can depolymerize both intact and fragmented cortex. In previous work, SleB from B. cereus did not exhibit activity toward cortical fragments (23). It is a possibility that our slightly different reaction conditions coupled with a different substrate source have provided evidence of previously undetectable activity; it is also probable that purification from E. coli rather than from spore exudates generated not only more SleB but also protein with a higher specific activity. However, in agreement with the earlier study, our results do suggest that the SleB enzyme is more active on intact cortex, which is a substrate that SleB is more likely to encounter during the initial stages of germination, where its role seems more critical (16, 17).

Muramic-δ-lactam has been implicated as the structural determinant that delineates spore cortex from the germ cell wall and thus dictates substrate specificity by GSLEs (2, 34, 37). It was therefore surprising to observe SleB causing a significant decrease in the OD₆₀₀ of DPVB19 spore sacculi and cortex fragments. Despite this loss in optical density, SleB treatment of DPVB19 substrates did not release identifiable muropeptides or solubilize any measurable cortex. One possible explanation for these results is that the DPVB19 substrate aggregates prior to the assay and that SleB binding causes dissociation of aggregates and thus a loss in OD. An alternative explanation is that SleB hydrolyzes DPVB19 PG, but with such poor efficiency that no detectable amount of cortex is released. SleB may cleave this substrate that lacks muramic-δ-lactam so infrequently that very long muropeptide chains result, similar to the apparent action of SleB_125-253 on PS832 substrate, which itself is only barely detectable with HPLC (Fig. 3F and I). In addition, the higher level of cross-linking in DPVB19 cortex than in PS832 cortex (34) may keep these large strands covalently linked. The ultimate result is a loosened spore sacculus that, due to reduced mass density, altered thickness, altered shape, or a combination of all three, exhibits a loss in light scattering properties but does not disassemble. Neither SleB nor SleB_125-253 caused any changes in optical density or muropeptide generation when vegetative cell walls from PS832 cells were used as a substrate. It is quite possible that our in vitro assay uses such a large amount of SleB that a minimal amount of activity on DPVB19 PG occurs that is not observable in vivo. We do not intend to suggest that this observation contradicts the in vivo evidence that muramic-δ-lactam is the structural determinant for cortex recognition by GSLEs and ultimately prevents germ cell wall degradation (34). Instead, we suspect that the in vitro conditions for this analysis afforded us an opportunity to observe activity by SleB that, while present, is inconsequential to a germinating spore.

SleB's C-terminal domain is most certainly the hydrolase domain, as predicted, because it carries out lytic transglycosylase activity without the protein's N-terminal domain. This hydrolase domain is capable of maintaining a significant level of interaction with PG substrate; however, this is not without precedent, since CwlJ1 and CwlJ2, two other GSLEs active in B. anthracis spores, consist solely of this putative hydrolase domain. The C-terminal domain's dependence on muramic-δ-lactam was made clear by the significantly decreased levels of interaction of SleB_125-253 with DPVB19 substrate in comparison to interaction with PS832 substrate, which resulted in a total loss of detectable hydrolytic activity.

SleB's N-terminal domain, as predicted, is primarily responsible for binding PG. The N-terminal domain is dominant in this role, since its presence completely masks the C-terminal domain's dependence on muramic-δ-lactam during the PG-binding assay. In fact, the PG-binding domain appears to function regardless of muramic-δ-lactam. Although the N-terminal domain does not likely carry out hydrolytic activity, it does contribute greatly to SleB's overall capability as a lytic enzyme. This is most apparent because the presence of the N-terminal domain affords the C-terminal domain the ability to reduce the OD₆₀₀ of cortex when muramic-δ-lactam is not present.

The mechanism that activates SleB during spore germination has remained an unresolved issue. It has been proposed that SleB is triggered to degrade cortex once the PG is in a stressed conformation (13, 38). This theory proposes that SleB is not active in dormant spores, because the cortex PG is not in the proper conformation for cleavage. This could change during early germination, as the core rehydrates and turgor pressure increases, causing a change in PG conformation, thus allowing it to be recognized and cleaved by SleB. Our data argue against this theory, since SleB was able to hydrolyze the intact cortex of coat-permeabilized spores, spore sacculi, and fragmented cortex. In coat-permeabilized spores, core dehydration and heat resistance are unaffected (11), which suggests that the stress on the cortex is the same as in a dormant spore. Sacculi are presumably less stressed because no core is present; instead, only stress inherent to the cortex structure is expected. Fragments are likely the least-stressed substrate, because some of the inherent structural stress is relieved, as the PG is no longer in a spore superstructure. These results demonstrate that SleB activity is independent of changes in cortex stress during germination.

Future research into SleB should focus on understanding how the protein is held inactive in the dormant spore and activated during germination. SleB's ease of purification lends itself to structural analysis, which could reveal the details of GSLE-substrate interaction. Ultimately, these investigations may uncover a means to prematurely trigger SleB action in the dormant spore, reducing its resistance properties before causing infection and greatly simplifying efforts of B. anthracis decontamination.