Roles of Germination-Specific Lytic Enzymes CwlJ and SleB in Bacillus anthracis

The structural characteristics of a spore enable it to withstandstresses that typically kill a vegetative cell. Spores remaindormant until small molecule signals induce them to germinateinto vegetative bacilli. Germination requires degradation ofthe thick cortical peptidoglycan by germination-specific lyticenzymes (GSLEs). Bacillus anthracis has four putative GSLEs,based upon sequence similarities with enzymes in other species:SleB, CwlJ1, CwlJ2, and SleL. In this study, the roles of SleB,CwlJ1, and CwlJ2 were examined. The expression levels of allthree genes peak 3.5 h into sporulation. Genetic analysis revealedthat, similar to other known GSLEs, none of these gene productsare individually required for growth, sporulation, or triggeringof germination. However, later germination events are affectedin spores lacking CwlJ1 or SleB. Compared to the wild type,germinating spores without CwlJ1 suffer a delay in optical densityloss and cortex peptidoglycan release. The absence of SleB alsocauses a delay in cortex fragment release. A double mutant lackingboth SleB and CwlJ1 is completely blocked in cortex hydrolysisand progresses through outgrowth to produce colonies at a frequency1,000-fold lower than that of the wild-type strain. A null mutationeliminating CwlJ2 has no effect on germination. High-performanceliquid chromatography and mass spectroscopy analysis revealedthat SleB is required for lytic transglycosylase activity. CwlJ1also clearly participates in cortex hydrolysis, but its specificmode of action remains unclear. Understanding the lytic germinationactivities that naturally diminish spore resistance can leadto methods for prematurely inducing them, thus simplifying theprocess of treating contaminated sites.

INTRODUCTION

Top

INTRODUCTION

The spore of Bacillus anthracis is the infectious agent of thedisease anthrax. Once entry to a suitable host has occurred,the endospore must undergo germination before it can produceany deadly toxins (12, 30). The bacterial endospore is a modifiedcell where the cytoplasm (spore core) is relatively dehydratedand surrounded by a thick peptidoglycan (PG) wall, called thecortex. As a result, the spore is metabolically dormant andstrongly resistant to environmental insults (42). The cortexplays a major role in maintaining spore resistance by limitingthe amount of water present in the core (42).

Although the spore is insensitive to environmental challenges,it remains responsive to particular chemical germinants. Contactwith these germinants to a spore's germinant receptors can inducegermination and outgrowth into a vegetative cell (11). Thereare three major events that occur once germination has beensignaled. The first is that the spore core starts to rehydrateas water moves inward (36). The second event, likely coupledwith the first, is transport of ions and dipicolinic acid (DPA)out of the core (9, 23, 30, 36). The third major step duringgermination includes the degradation of spore cortex PG by germination-specificlytic enzymes (GSLEs) and the release of muropeptides into thesurrounding environment (30). At this point, the spore coreis now free to expand fully and proceed toward a vegetativecell cycle (29).

PG consists of a repeating disaccharide of β-1,4-linkedN-acetylglucosamine (NAG) and N-acetylmuramic acid (NAM). TheNAM residue has a peptide side chain extending from the lactylgroup, which is able to form a bond with a peptide chain onanother PG strand, resulting in a cross-linked network surroundingthe entire cell (40). This structure accurately describes thethin innermost layer of PG that encases an endospore. Termedthe germ cell wall, this layer may serve as the foundation fornewly synthesized PG after germination. The second layer ofPG around endospores, referred to as the cortex, composes >80%of the total spore PG (26). It has a defining modification whereas much as 50% of the NAM is converted to muramic- ${delta}$ -lactam (26).During germination, the cortex, but not the germ cell wall,is broken down.

GSLEs have been shown to degrade only PG with muramic- ${delta}$ -lactam,thus explaining cortex-exclusive breakdown (37, 41). In Bacillussubtilis, two GSLEs in particular appear crucial to germination:cwlJ and sleB (1, 16). CwlJ is expressed under the control ofthe mother cell sporulation factor ${sigma}$ ^E, and evidence suggeststhat it ultimately localizes to the spore coat (3, 16). Properlocalization, and hence function, is dependent on the productof gerQ (39). CwlJ is required for germination in response toCa²⁺-DPA; however, its specific enzymatic activity on PG remainsa mystery (29, 36). CwlJ shows 30% sequence identity to theproposed catalytic domain of the other main GSLE of B. subtilis,SleB. SleB has an N-terminal putative cell wall-binding domainand is required for the appearance of lytic transglycosylaseactivity during germination of B. subtilis spores (1). MatureSleB and its homologs localize to the spore cortex and innermembrane in Bacillus cereus (33), B. subtilis (8, 25), and B.anthracis (21). Expression of sleB is controlled by ${sigma}$ ^G in theforespore (32). In B. subtilis, proper function of SleB is dependenton ypeB, which is located downstream of and included in a bicistronicoperon with sleB (5, 8).

This study investigated three putative B. anthracis GSLEs, referredto as SleB, CwlJ1, and CwlJ2. Using null mutants, we show thatSleB and CwlJ1 play significant roles in B. anthracis sporecortex hydrolysis. Both sleB and cwlJ1 are expressed in a mannerconsistent with that observed in B. subtilis. We find that SleBis required for lytic transglycosylase activity. CwlJ2 is poorlyexpressed and plays no evident role in germination.

MATERIALS AND METHODS

Top

MATERIALS AND METHODS

Bacterial strains and antibiotics. All B. anthracis strains used in this study are derived fromthe Sterne strain 34F2 and are listed in Table 1. Electroporationof B. anthracis was performed as described previously (38).Transformants were selected on LB-1% glucose plates containingthe appropriate antibiotics: 5 µg/ml erythromycin (Fisher),50 µg/ml kanamycin sulfate (Jersey Lab Supply), or 10µg/ml tetracycline (Jersey Lab Supply). Vegetative growthin brain heart infusion (BHI; Difco) medium with the correspondingantibiotic was examined for each strain, at 37°C for deletionmutants and 39°C for insertion mutants. Insertion mutantstrains were sporulated in modified G (Mod G) medium withoutantibiotics (18) at 39°C with shaking for 3 days. Deletionmutants were similarly sporulated at 37°C, except for strainsused for complementation experiments, which were cultivatedat room temperature in Mod G containing the appropriate antibiotic.All spores were harvested by centrifugation, and vegetativecells were killed by incubation at 65°C for 25 min. Furtherspore purification was done with water washing and density centrifugationover 50% sodium diatrizoate (Sigma) as previously described(34).

View this table:
[in this window]
[in a new window]

TABLE 1. B. anthracis strains and plasmids

Maintenance of pNFd13 containing the counterselectable markerccdB was done with the One Shot ccdB Survival strain of Escherichiacoli, available from Invitrogen. E. coli was grown in LB mediumcontaining 10 µg/ml tetracycline hydrochloride (JerseyLab Supply), 30 µg/ml chloramphenicol (Fisher), 50 µg/mlkanamycin sulfate (Jersey Lab Supply), and 500 µg/ml erythromycin(Fisher) as appropriate.

Mutant construction. All oligonucleotide sequences used in plasmid construction aregiven in Table S1 in the supplemental material. In order tocreate plasmid insertion mutations, truncated forms of eachgene, containing the ribosome binding site but lacking a promoter,were generated using PCR as previously described (11). The truncatedform of sleB contained 135 codons of the 305-codon gene. ThecwlJ1 and cwlJ2 truncations included 102 and 78 codons of theirtotal 140 and 142 codons, respectively. The truncated gene fragmentswere recombined into pDONRtet and pNFd13 (11) by using the Gatewaycloning system (Invitrogen), and the resulting plasmids wereinserted into the B. anthracis chromosome as described previously(11). PCR was used to confirm the gene disruptions. These plasmidinsertions also created transcriptional fusions to lacZ (11).

In order to create in-frame deletions of each gene, PCR wasused to amplify the entire target gene as well as several hundredbases up- and downstream. Each fragment was cloned into pBKJ236(17) by using the corresponding endonucleases. Inverse PCR withappropriate primers produced deletion constructs with BglI sitesat the deletion points, and BglI digestion and subsequent ligationwere used to create plasmids carrying in-frame deletions. ThesleB deletion eliminated all but the first 10 and the last 9codons of the gene. The cwlJ1 and cwlJ2 deletions left onlythe first three and last two codons of each gene. Each mutationwas introduced into B. anthracis via markerless gene replacementas previously described (17). Complementation of each mutationwas achieved by introduction of the pBKJ236 derivative withthe corresponding full-length gene, while empty pBKJ236 servedas a negative control. These complementing plasmids were maintainedextrachromosomally at the permissive temperature of 27°Cand under erythromycin selection.

β-Galactosidase assay. Strains carrying lacZ fusions were grown in Mod G medium, andsamples were taken during sporulation, concomitantly with opticaldensity (OD) readings. Cell samples were permeabilized with2% CHCl₃ and 0.001% sodium dodecyl sulfate, and β-galactosidaseactivity was determined using the substrate o-nitrophenyl-β-D-galactopyranosideas previously described (27).

Germination assays. Spore germination and outgrowth were assayed by changes in OD₆₀₀over time. Synchronous germination was achieved by heat activatingspores at 70°C for 20 min and then suspending them in BHIbroth to an OD₆₀₀ of 0.2 at 39°C. Spore viability was determinedusing a simple plating assay in which spores were first germinatedin BHI. Germination and outgrowth to 100% of the initial OD₆₀₀were carried out to reduce spore aggregation, followed by serialdilution and plating. Number of CFU per OD₆₀₀ unit was calculatedafter incubation overnight at 39°C. Release of DPA and cortexfragments during germination in buffer was assayed as previouslydescribed (10). When required, the coats of B. anthracis sporeswere permeabilized (decoated) as previously described (36).

HPLC analysis of PG. Spores were germinated in buffer and used to prepare PG forreverse-phase high-performance liquid chromatography (RP-HPLC)analysis as previously described (10). Briefly, the germinatedspore suspension was separated into spore-associated (pellet)and released-exudate (supernatant) fractions. Lytic enzymeswere inactivated with heat (supernatant) or with heat and detergent(pellet), and PG was purified. The PG material from the pelletsand half of each exudate fraction were then digested with themuramidase mutanolysin (Sigma). All fractions were reduced withNaBH₄ prior to HPLC separation.

Amino acid-and-sugar analysis was used to characterize novelmuropeptides eluted from the HPLC separation. The phosphatebuffer from each collected muropeptide was removed by repeatingthe HPLC separation in 0.05% trifluoroacetic acid with a 0 to20% acetonitrile gradient. The novel compounds were then hydrolyzedin HCl vapor and analyzed as previously described (10). Muropeptidesof interest were identified by mass spectrometry using an AppliedBiosystems 3200 Q Trap tandem mass spectrometry system. Thetotal mass and fragmentation of each compound in the negative-ionmode were determined as previously described (35).

RESULTS

Top

RESULTS

GSLE homologs in B. anthracis. Analysis of sequence similarities suggested that B. anthraciscontains GSLE homologs in its genome (Table 2). The most notableof these are one homolog of SleB and two homologs of CwlJ. Asecond B. anthracis homolog of SleB is more closely relatedto another B. subtilis cell wall hydrolase, YkvT, which playsno role in spore metabolism (8, 15). For this reason, the secondhomolog of SleB was not analyzed. No homologs of the Clostridiumperfringens SleC or SleM (7, 28) GSLEs were identified in theB. anthracis genome.

View this table:
[in this window]
[in a new window]

TABLE 2. Homologs of B. subtilis GSLEs

In B. subtilis, sleB is the first gene in a bicistronic operonwith ypeB. Interestingly, the sleB homolog at locus BAS2562in B. anthracis appears to be the first in a putative tricistronicoperon. The second open reading frame, BAS2561, is homologousto ypeB. The product of the third open reading frame in thispossible operon shares homology with the uncharacterized B.subtilis protein YlaJ (22).

The B. anthracis cwlJ homolog at locus BAS5241 is the firstin a putative bicistronic operon with a gerQ homolog at BAS5242.The second cwlJ homolog, at BAS2417, appears to be monocistronic.For the duration of this correspondence, BAS5241 and BAS2417will be referred to as cwlJ1 and cwlJ2, respectively, and thegene at BAS2562 will be called sleB.

GSLE-encoding genes are transcribed during sporulation. Plasmid insertion mutagenesis allowed the investigation of geneexpression through the resulting genetic fusion of lacZ to thenative promoter of each GSLE-encoding gene. Expression fromthese promoters was monitored by assaying β-galactosidaseactivity. The start of sporulation was defined as the transitionfrom growth into stationary phase as observed by change in OD₆₀₀and is termed time zero (t₀). Wild-type B. anthracis, withouta lacZ fusion, exhibited the lowest levels of β-galactosidaseactivity throughout the experiment (Fig. 1A and B). The highestlevel of activity was produced by the cwlJ1 fusion, startingabout 1 hour after sporulation initiation (t₁). The transcriptionof cwlJ1 peaked near t_3.5, which also coincided with the maximumactivity for the sleB and cwlJ2 fusions. However, the expressionlevels of both the cwlJ2 and the sleB promoters were markedlylower and were not detectable above the background prior tot_2.5 (Fig. 1B). The peak cwlJ2 and sleB expression levels were5- to 10-fold above background, similar to those observed forgerminant receptors with the same type of lacZ fusion (11).

View larger version (20K):
[in this window]
[in a new window]

FIG. 1. Expression of cwlJ1, cwlJ2, and sleB during B. anthracis sporulation. Strains carrying each promoter-lacZ fusion were shaken in Mod G medium at 39°C and were assayed for β-galactosidase activity. (A) OD (open symbols) and β-galactosidase activity (filled symbols) are shown for a wild-type strain lacking lacZ ( ${diamondsuit}$ ) and for strains with lacZ fused to cwlJ1 ( ${blacksquare}$ ), cwlJ2 ( ${blacktriangleup}$ ), and sleB (•). (B) β-Galactosidase activity for the wild type, cwlJ2-lacZ, and sleB-lacZ are shown on an expanded y axis for clarity. Both panels show representative data for one of three independent experiments.

Our expression results agree with earlier microarray and proteomeinvestigations of B. anthracis that suggest that cwlJ1 is expressedduring an early wave of sporulation-associated transcription,which is then followed closely by another wave, which includessleB (4, 21). This is further supported by our observation ofcandidate promoter sequences for ${sigma}$ ^E and ${sigma}$ ^G (13) upstream of cwlJ1and sleB, respectively (data not shown). The observed time forcwlJ2 expression in this work more closely matches that forsleB than it does that for cwlJ1.

Effects of GSLE mutations on germination. None of the single- or double- mutant strains exhibited anysignificant deviations from the wild type with regard to doublingtime during exponential growth, formation of heat-resistantspores, or spore morphology (data not shown).

Spores of strains lacking cwlJ1, cwlJ2, sleB, or both cwlJ1and sleB were germinated in BHI and monitored through germinationand outgrowth by measuring the change in OD_600. As spores germinate,they rapidly lose between 40% and 60% of their OD within thefirst few minutes of contact with germinants. This is due tospore water uptake, Ca-DPA release, and cortex hydrolysis (34).As the spore population completes germination and enters outgrowth,the OD₆₀₀ increases, thus continuing into vegetative growth.

Native spores of wild-type and all single mutant strains lostat least 58% of their initial ODs, indicating that all weresuccessful at synchronous germination (Fig. 2A). However, theprecise kinetics for strains to finish germination and enteroutgrowth was variable. The cwlJ2 and sleB mutants mimickedthe wild type in their progress and lost 40% of their initialODs within 1 minute of each other. The entire curve for cwlJ2spores looked almost identical to that for the wild type, butthe sleB spores appeared to undergo a slightly less efficientgermination response, coupled with a lower rate of outgrowth.Due to variability among spore preparations, this slow responseof sleB spores was not statistically different from that ofwild-type spores, but it was reproducible in three independentspore preparations.

View larger version (19K):
[in this window]
[in a new window]

FIG. 2. Germination and outgrowth of native and decoated spores in BHI. Wild-type ( ${diamondsuit}$ ), cwlJ1and ${Delta}$ cwlJ1 ( ${blacksquare}$ ), cwlJ2 ( ${blacktriangleup}$ ), sleB (•), ${Delta}$ cwlJ1 ${Delta}$ sleB (x), ${Delta}$ cwlJ1-plus-pBKJ236 (–), and ${Delta}$ cwlJ1-plus-pDPV345 (+) spores were heat activated and germinated in BHI at 39°C. (A) Germination of native spores. (B) Germination of decoated spores. (C) Complementation of the ${Delta}$ cwlJ1 phenotype. Data shown are averages for three independent experiments; error bars are omitted for clarity. Asterisks indicate those time points when cwlJ1 (A), sleB (B), or ${Delta}$ cwlJ1 and ${Delta}$ cwlJ1-plus-pBKJ236 (C) spores were significantly different (P $≤$ 0.05) from wild-type spores. The ${Delta}$ cwlJ1 ${Delta}$ sleB spores (A) were significantly different from those of the wild type at all time points except 60 to 70 min, where the lines cross.

An overall delay by cwlJ1 spores was evident by the rightwardshift of the strain's germination curve. Compared to the wildtype, cwlJ1 spores were significantly delayed in OD loss from5 to 10 minutes after germination initiation as well as duringoutgrowth for a span of 40 minutes (P $≤$ 0.05, as determined byan unpaired Student t test) (Fig. 2A). These spores took atleast twice as long as wild-type spores to lose 40% of theirstarting OD (Table 3), and the time it took to reach their minimalOD was also significantly delayed. The rate of outgrowth forthe cwlJ1 strain was normal; its starting point was simply delayeddue to slowed germination. Identical effects on germinationwere obtained with plasmid insertion and in-frame deletion mutationsin each gene (data not shown), suggesting that if other genesin these operons play significant roles in germination, thenthey may be involved in the same GSLE functions. The cwlJ1 slow-germinationphenotype was complemented with pDPV345 but not the empty-vectorcontrol (Fig. 2C).

View this table:
[in this window]
[in a new window]

TABLE 3. Germination of spores in BHI^a

Spores produced by the cwlJ1 sleB double deletion strain werecapable of losing OD only in the first 5 minutes after contactwith germinants (Fig. 2A). Germination was arrested at thispoint, and a 40% loss in OD was never reached (Table 3), norwas the OD ever observed to increase (Fig. 2A). This defectwas complete enough to affect spore viability since the ${Delta}$ cwlJ1 ${Delta}$ sleB spores had a 1,000-fold loss in colony-forming ability(Table 3). Native spores lacking only one of the putative GSLEshad no change in viability.

Permeabilizing (decoating) the spore coats either removes orgreatly disrupts proteins sequestered in that region of thespore (36). In B. subtilis, CwlJ and SleL have been shown tobe located in or near the coat of dormant spores while SleBlocalizes to the cortex or membrane layers (8, 19, 39). As such,decoating spores offered a logical avenue for determining theeffects of simultaneously losing the functions of three putativeGSLEs (CwlJ1, CwlJ2, and SleL), as shown in B. subtilis (36).For each strain, the germination of decoated spores was substantiallydelayed compared to that of untreated spores (Table 3). Regardless,the stages of germination (loss of OD₆₀₀) and outgrowth (increasein OD₆₀₀) remained apparent and synchronous for wild-type, cwlJ1,and cwlJ2 strains (Fig. 2B). The fact that decoated wild-typeand cwlJ1 spores act indistinguishably suggests that this treatmentinactivated CwlJ1.

The most profound result from investigating decoated sporeswas revealed by the sleB mutant. These spores responded to nutrientsand engaged in early germination events indistinguishably fromthe wild type (Fig. 2B). However, that progress was arrestedat a point just prior to loss of 25% of the initial OD. Fromthis point on, any loss of OD occurred at a significantly lowerrate (P $≤$ 0.05). Decoated sleB spores never reached a 40% lossof OD or exhibited any characteristics of outgrowth. Continuedobservation revealed no further loss in OD over 4.5 h (datanot shown). The decoating treatment did not prevent wild-type,cwlJ1, or cwlJ2 spores from germinating and forming colonies.However, the decoated sleB spores had a 1,000-fold loss in colonyforming ability. This was not simply a long germination delay;continued observation of plates for 3 days revealed no new colonyappearance. It should be noted that this phenotype is identicalto what was observed for native cwlJ1 sleB double mutant spores.

DPA, NAM, and Dpm release from spores. One of the earliest events in spore germination is the releaseof Ca-DPA from the spore core into the surrounding medium (23).Despite observed changes in OD loss during germination, allof the mutant spores, including those of the cwlJ1 sleB doublemutant, released as much DPA as the wild type during the earliestminutes following exposure to germinants (data not shown). Inorder to assay for the later steps in the germination process,one can analyze the ratio of the cortex-specific componentsNAM and Dpm found in the pellet versus the exudate fractionsof germinating spores. Release of these compounds is rapid underthe conditions used for this assay, with measurable differencesover the course of a few minutes (10). Assays revealed thatboth wild-type and cwlJ2 spores released $~$ 50% of their NAM andDpm within the first 5 minutes of contacting germinants, withmaximum release occurring by 10 minutes (Fig. 3A and B).

View larger version (13K):
[in this window]
[in a new window]

FIG. 3. Release of NAM and Dpm from germinating B. anthracis spores. Dormant wild-type ( ${diamondsuit}$ ), cwlJ1 ( ${blacksquare}$ ), cwlJ2 ( ${blacktriangleup}$ ), sleB (•), and ${Delta}$ cwlJ1 ${Delta}$ sleB (x) spores in buffer were germinated with L-alanine and inosine and were sampled for the release of NAM (A) and Dpm (B) at 5-minute intervals after contact with germinants. (C) Complementation of ${Delta}$ cwlJ1 ( ${blacksquare}$ ) NAM release was tested with ${Delta}$ cwlJ1(pDPV345) () and ${Delta}$ cwlJ1(pBKJ236) ( ${square}$ ) spores. (D) Complementation of ${Delta}$ sleB (black circles) NAM release was tested with ${Delta}$ sleB(pDPV346) (gray circles) and ${Delta}$ sleB(pBKJ236) (white circles) spores. All error bars represent 1 standard deviation from the mean for three (A, B, and D) or two (C) independent experiments. All points have error bars, but in some cases, these are too small to be visible.

cwlJ1 spores were incapable of releasing cortex components asrapidly; hence, after 5 minutes, only 8% and 6% of NAM and Dpm,respectively, had been released (Fig. 3A and B). Still, thecortex was swiftly released from the cwlJ1 spores in the followingminutes and approached the wild-type release level by 15 minutes.This result coincides with the delay in germination responserevealed by OD changes (Fig. 2A). Complementation with pDPV345,but not with the vector control, restored both NAM and Dpm releaseto wild-type levels (Fig. 3C and data not shown).

Spores without sleB were also delayed in discharging their cortex,despite having no obvious changes in OD loss (Fig. 2A and 3A and B).Initially, the delay was not as dramatic as that exhibited bythe cwlJ1 spores; however, over the following minutes, the sleBmutant did not release cortex as speedily, thus resulting inthe least total release among all single-mutant strains. Fifteenminutes after the start of germination, sleB spores had managedto release only 45% and 27% of their total NAM and Dpm, respectively.Complementation with pDPV346, but not with the vector control,restored both NAM and Dpm release to wild-type levels (Fig.3D and data not shown). Deleting both cwlJ1 and sleB resultedin spores that released no NAM or Dpm during germination (Fig.3A and B).

Cortex hydrolysis is slowed in cwlJ1 and sleB spores. Previous HPLC analyses revealed that wild-type B. anthracisspores released the majority of their cortex within the first10 minutes of germination with L-alanine and inosine (10). Weobtained similar results, where after 5 min of germination,muropeptides derived from the spore pellet were predominantlygerm cell wall associated (for example, peaks K and M) (Table4 and Fig. 4A), and all of the PG fragments found in germinatingspore exudates were derived from the cortex (Table 4 and Fig.4E and I). When cwlJ2 was disrupted, the chromatograms of pelletand exudate fractions were indistinguishable from those of thewild type (Fig. 4B, F, and J).

View this table:
[in this window]
[in a new window]

TABLE 4. Muropeptide peak identification

View larger version (33K):
[in this window]
[in a new window]

FIG. 4. RP-HPLC separation of muropeptides from germinating B. anthracis spores. PG was prepared from germinating spore suspensions as described in Materials and Methods. Samples were collected after spore suspensions had lost 40% of their initial ODs: approximately 5 min for wild-type and cwlJ2 spores and 10 min for cwlJ1 and sleB spores. PG from germinating spore pellets (A to D) and from 50% of the exudate preparations (I to L) were digested with muramidase, reduced, and separated as previously described (26). The other 50% of the exudates (E to H) were reduced and separated without muramidase digestion. Peaks are numbered as in reference 10 and in Table 4, but the initial "a" in the germination-specific-peak names were omitted for space considerations. Early-eluting peaks labeled "b" are buffer components present in blank samples. Peaks labeled "ex" are spore exudate components that, based upon amino acid analysis, are not derived from PG. Peaks labeled "Ino" are from the inosine used to germinate spores.

Spores lacking cwlJ1 did not release muropeptides nearly aswell. In fact, cwlJ1 spores retained so much more spore PG duringgermination that cortex-derived muropeptides became the dominantpeaks in the chromatogram of spore-associated material (forexample, peaks N and Q) (Fig. 4C). The average amount of muropeptideN retained increased by >350% when cwlJ1 was disrupted. Thisretention of cortex was coupled with a decrease in cortex release,as seen in exudates of germinating spores (Fig. 4G and K). Cortex-derivedmuropeptides G6 and G7 diminished from levels seen in wild-typespore exudates by averages of 48% and 50%, respectively (Fig.4E and G). The same change was observed when exudates were digestedwith mutanolysin, except G4, the product of mutanolysin activityon G6, was subject to a decrease (Fig. 4I and K). A previousanalysis of cortex hydrolysis by a B. subtilis cwlJ mutant revealedno change relative to the wild-type level (8). This differenceis likely due to the very slow release of muropeptides by B.subtilis, relative to that by B. anthracis (10), making a differencein relative rate of release by B. subtilis strains difficultto detect.

Spores lacking SleB also exhibited a defect in muropeptide release.Much more cortex remained in the spore during germination; forinstance, muropeptide N had an average increase of >350%compared to the wild-type level (Fig. 4D). This result is nearlyidentical to what was seen in cwlJ1 spores; however, the associatedchange in released muropeptides was unique. The exudates ofsleB spores did not contain any of the G7a or G7b muropeptides(Fig. 4H). Exudate digested with mutanolysin lacked the sametwo muropeptides (Fig. 4L), but no other muropeptides were alteredfrom what was observed in wild-type exudates. This phenotypewas fully complemented by introducing sleB on a plasmid butnot by the empty vector (Fig. 5A to D). The requirement forSleB in the production of this class of anhydromuropeptide (seebelow) was also observed in B. subtilis (5). No muropeptideswere released from spores containing both the cwlJ1 and thesleB deletions (data not shown). The germinated spore pelletscontained cortex identical to that found in dormant B. anthracisspores during previous work (10).

View larger version (23K):
[in this window]
[in a new window]

FIG. 5. Complementation of ${Delta}$ sleB. Samples were prepared and analyzed, and peaks are labeled, as described for Fig. 4. (A) Muropeptides released from wild-type spores. (B) Muropeptides released from ${Delta}$ sleB spores. (C) Muropeptides released from spores with ${Delta}$ sleB plus the complementing plasmid pDPV346. (D) Muropeptides released from spores with ${Delta}$ sleB plus the control vector pBKJ236.

Spores that had been decoated were also poor at digesting cortex.Pellet fractions of these spores contained the majority of totalPG as N replaced K as the dominant peak (data not shown). Verylittle material was released in the exudates (Fig. 6A to D).Decoated wild-type, cwlJ1, and cwlJ2 spores all generated similarexudate muropeptide profiles. Prior to mutanolysin exposure,the predominant muropeptides were G7a and G7b. After treatmentwith mutanolysin, G7a and G7b remained, but N was the dominantmuropeptide (data not shown), indicating that the exudates containedlarge, not fully digested PG fragments. Decoated sleB spores,similarly to the sleB cwlJ1 double mutant under native conditions,released little to no PG during germination, with or withoutmutanolysin digestion (Fig. 6D and data not shown).

View larger version (24K):
[in this window]
[in a new window]

FIG. 6. RP-HPLC separation of muropeptide exudates from decoated germinating B. anthracis spores. Spores were permeabilized as previously described (36). Samples were prepared and analyzed, and peaks are labeled, as described for Fig. 4. All samples were harvested after spores lost 40% of their initial OD ( $~$ 45 min). (A) Muropeptides released from wild-type spores. (B) Muropeptides released from cwlJ2 spores. (C) Muropeptides released from cwlJ1 spores. (D) Muropeptides released from sleB spores.

Muropeptides G7a and G7b contain anhydromuramic acid. Muropeptides G7a and G7b were identified through a combinationof amino acid analysis and mass spectrometry (Fig. 7 and Table5). The ratio of individual amino acids in G7a matched the expectationsfor a muropeptide containing a tetrapeptide side chain. Thequalitative results of an amino sugar analysis revealed thatonly NAM and NAG were present. No reduced sugars, muramitol(MOH) or glucosaminitol, were detected, despite the fact thatthe sample was reduced prior to HPLC separation, indicatingthat the "reducing-end" sugar in this muropeptide is resistantto reduction. Mass spectrometry showed G7a to be 20 Da smallerthan the tetrasaccharide-tetrapeptide muropeptide N (Table 4).N releases MOH upon amino sugar analysis due to the reductionprior to HPLC separation (10). If G7a is ordered similarly toa TS-TP, then the 20-Da mass discrepancy can be explained bythe presence of an anhydromuramic acid as the product of a lytictransglycosylase (2, 14). Mass spectrometry analysis of G7afragmentation exactly matches the expectations of anhydro-tetrasaccharidetetrapeptide (Table 5 and Fig. 7). Amino acid analysis, massspectrometry, and fragmentation analysis of muropeptide G7b(Table 5 and data not shown) revealed that it is anhydro-tetrasaccharidealanine.

View larger version (21K):
[in this window]
[in a new window]

FIG. 7. Muropeptide J1 and J2 are anhydro-tetrasaccharides. Structures were determined with a combination of HPLC amino acid/sugar analysis and electrospray ionization-mass spectrometry. (A, B) The sugar residues are (from left to right) NAG, muramic- ${delta}$ -lactam, NAG, and anhydro-NAM. (A) The amino acid residues are (in order from lactyl linkage) L-alanine, D-glutamate, meso-diaminopimelic acid, and D-alanine. Arrows indicate sites of fragmentation during electrospray ionization-mass spectrometry, as identified in Table 5. (B) The single-amino-acid side chain is L-alanine.

View this table:
[in this window]
[in a new window]

TABLE 5. Novel muropeptide identification^a

DISCUSSION

Top

DISCUSSION

The GSLE mechanism of B. anthracis spores appears to functionsimilarly to that of B. subtilis, with some minor but potentiallyimportant differences. B. anthracis strains with mutations affectingcwlJ1, cwlJ2, sleB, or cwlJ1 and sleB grow, divide, and sporulatewithout difficulty, and their spores initially respond normallyto nutrient germinants. In B. subtilis, the germination muropeptideprofiles of wild-type and cwlJ spores are indistinguishable(8), but for B. anthracis, we were able to observe significantlyfewer cortex-derived muropeptides being released from sporeswithin 10 minutes after contacting germinants. The structuresof the released muropeptides did not differ from those producedby the wild-type strain, leaving the bond specificity of anyCwlJ1 enzymatic activity unresolved. SleB, like CwlJ1, is requiredfor timely cortex degradation, but its mode of action is unique.Two cortex fragments that are entirely dependent on SleB arethe result of lytic transglycosylase activity. This activitywas associated with SleB by using similar methods with B. subtilisbut was not previously detected in B. anthracis (5, 10). Inresponse to germinants, the cwlJ1 sleB double mutant sporesinitiate but are incapable of completing germination, similarlyto cwlJ sleB B. subtilis spores (16). However, unlike for thesituation in B. subtilis, we did not observe any delay in DPArelease. The first 5 to 10 minutes of B. anthracis germination( $~$ 25% loss in OD) is independent of GSLE activity and includesthe release of nearly all of the DPA. Very rapid DPA release,like quicker cortex fragment release (10), may be a functionof thinner, more-permeable B. anthracis spore coat layers.

The highest rate of cortex degradation requires both cwlJ1 andsleB, yet they do not cause release of cortex fragments in exactlythe same manner (Fig. 3A and B). This suggests that SleB andCwlJ1 provide alternate pathways for cortex depolymerization(Fig. 8). We propose that this difference is best explainedin context with SleL, which shares 98% identity with its B.cereus homolog and functions as an N-acetylglucosaminidase inboth B. anthracis (20) and B. cereus (6). For B. cereus, SleLhas been further defined as a cortex fragment lytic enzyme thatrecognizes only cortex substrate that has undergone previousdigestion by another GSLE (6). In this model, the initial depolymerizationof the cortex would be the responsibility of SleB and CwlJ1.The NAM release from a cwlJ1 spore indicated that sixfold lesscortex was digested and released than from wild-type sporesin the first 5 minutes of germination (Fig. 3A). In addition,this was more than threefold less NAM than the sleB spores coulddischarge. These initial differences are likely due to a largedisparity in protein available for depolymerization, as suggestedby the higher level of cwlJ1 transcription (Fig. 1A). However,this situation rapidly reverses as germination continues. Within10 minutes, germinating cwlJ1 and sleB spores have the sameamount of NAM released, and by 15 minutes, cwlJ1 spores havesurpassed sleB spores to achieve NAM release at near-wild-typelevels (Fig. 3A). It is our suggestion that the rapid accelerationof NAM release by cwlJ1 spores is due to SleL action, whichwould have ample substrate available after several minutes ofSleB-initiated cortex digestion. Meanwhile, sleB spores releasecortex fragments early during germination due to naturally highlevels of CwlJ1 but are unable to maintain this rapid release.Perhaps the population of cortex fragments that CwlJ1 yieldsdoes not provide the best substrate for SleL. Patterns of Dpmrelease support the same conclusions. Sequential digestion ofspore PG by SleB and SleL has previously been suggested forB. cereus (24). Our model does not rule out the possibilitythat some number of CwlJ1-created muropeptides is acted on bySleL; indeed, SleL products are produced by sleB mutant spores.SleL may simply have a higher affinity for cortex fragmentsresulting from SleB action.

View larger version (12K):
[in this window]
[in a new window]

FIG. 8. Model for cortex hydrolysis by GSLEs in B. anthracis. The degradation of spore cortex is illustrated as two steps. Solid arrows indicate the reaction that is responsible for the majority of activity at that given step. Dashed arrows indicate those reactions that compose the minority of enzymatic activity. SleB and CwlJ1 are both capable of depolymerizing intact cortex, with more initial digestion due to CwlJ1. SleL, as a cortex fragment lytic enzyme, is capable of degrading only partially digested cortex. SleL prefers substrate provided by SleB, or the positioning of SleB results in production of a greater amount of SleL substrate. Fully digested cortex contains a mixed muropeptide population of all the shown substrates and products.

The responsibilities of cwlJ1 and sleB are clearly redundantdue to the partial effects on cortex lysis that they each haveand the complete lysis defect that results when both are disrupted.Further study of multiple mutant strains might provide novelmuropeptide HPLC profiles that offer better insight into a specificlytic activity associated with cwlJ1. These studies should includesleL to provide a comprehensive analysis of the events duringcortex degradation. Understanding the lytic germination activitiesthat naturally diminish spore resistance can lead to methodsthat prematurely activate cortex degradation, thus greatly simplifyingthe process of cleaning B. anthracis-contaminated sites.

ACKNOWLEDGMENTS

This research was supported by Public Health Service grant AI060726from the National Institute of Allergy and Infectious Diseases.

ABI mass spectrometers were donated to Virginia Tech by PPD,Inc., Richmond, VA.

FOOTNOTES

* Corresponding author. Mailing address: Department of Biological Sciences, Virginia Tech, Life Sciences I, MC0910, Washington St., Blacksburg, VA 24061. Phone: (540) 231-2529. Fax: (540) 231-4043. E-mail: dpopham{at}vt.edu

Published ahead of print on 30 January 2009.

Supplemental material for this article may be found at http://jb.asm.org/.

REFERENCES

Top