The strains and plasmids used in this study are listed in Table S1. E. coli strains were grown in Luria Bertani (LB) medium and B. anthracis strains in Brain Heart Infusion (BHI) medium. Sporulation was induced by resuspension in modified G medium [32] [33] and mature spores were formed after plating of vegetative cells on Spizizen Yeast (SPY) agar medium which when required was supplemented with 0.2% xylose.

A strain bearing an in-frame deletion of ba1836 gene was created using a markerless gene replacement method [34] . In brief, DNA fragments containing the sequence upstream and downstream of ba1836 gene were amplified from genomic DNA of B. anthracis 7702 [35] by PCR using the appropriate oligonucleotides (Table S2). Both fragments were cloned into the same pGEM vector and the upstream-downstream fragment was digested with SacI/KpnI and ligated into pRP1028 vector. E. coli HB101 [36] was used as helper strain in order to primarily introduce pRP1028 construct [37] and subsequently pSS4332 vector [37] , which encodes the I-SceI enzyme, into B. anthracis 7702 strain. Finally, the knockout strain had no resistance to antibiotics and was validated for the correct construction by PCR.

To construct strains expressing GFP translational fusions, the gfp-mut1 gene was amplified from pNF8 [38] with specific primers, digested with NruI and BglII, and ligated into the xylose-inducible plasmid pWH1520 [39] . Then, ba1836 (lacking the stop codon) was amplified from B. anthracis 7702 chromosomal DNA with the appropriate primers to incorporate at the C terminus the polylinker GPGP. The amplicon was digested with BamHI and NruI and ligated in-frame to the 5' end of gfp-mut1, placing the gene fusion under a xylose-inducible promoter. E. coli GM48 (dam⁻) was transformed with the resulting construct to obtain nonmethylated plasmid DNA for electroporation [40] into 7702 and Δba1836 B. anthracis cells. 10 ml of BHI medium were inoculated from overnight cultures to an OD_{600 nm} of 0.05 and incubated at 37˚C. For ba1836-gfp expressing strains, induction was achieved with 0.2% xylose (for expression from the xylose-inducible promoter) 1 h before the culture reached early stationary phase of growth (OD_{600 nm} around 3.0).

In order to create the lacZ-transcriptional fusion, a DNA fragment of the upstream region of ba1836 gene was amplified from B. anthracis 7702 chromosomal DNA. Primers were synthesized to incorporate a HindIII site at the start and a BamHI site at the end of the amplified fragment (Table S2), which was digested with the corresponding enzymes, and ligated into the multicopy replicative vector pHT304-18Z [41] . In this way, a lacZ reporter construct with the upstream region of each gene fused directly to the start codon of the lacZ gene was obtained. E. coli GM48 (dam⁻) was transformed with the resulting construct to obtain nonmethylated plasmid DNA for electroporation [40] into 7702 B. anthracis cells.

A mutant of the ba1836 gene was constructed using a two-step/four-primer overlap extension PCR method [42] . Primers used in this method are listed in Table S2. The amplified ba1836 product was cloned into pWH1520 vector as described above.

10 ml of BHI medium were inoculated from overnight cultures of 7702 parental and Δba1836 mutant strains to an OD_{600 nm} of 0.05 and incubated at 37˚C. When OD_{600 nm} of the cultures reached 1.2 (mid-exponential phase) or 3.0 (early stationary phase), the cultures were centrifuged for 20 min at 2570 g and sporulation was induced by resuspension of the cell pellet in an equal volume of modified G medium. The cultures were then transferred to 30˚C. Cells were collected at frequent time intervals and observed under optical and fluorescence microscopy.

Cell samples were collected in triplicates from shaking cultures at various time points during vegetative growth or sporulation. OD_{600 nm} for each cell sample was recorded before collection. β-galactosidase activity was measured as described by Camp and Losick [43] with minor changes. Cells pellets corresponding to 1 ml of the original sporulation medium were resuspended in 0.2 ml Z-buffer (60 mM Na₂HPO₄, 40 mM NaH₂PO₄, 10 mM KCl, 1 mM MgSO₄, 50 mM β-mercaptoethanol pH 7.0). 50 μl of the cell suspension were added to individual wells of a 96-well plate containing 15 μl of 0.1% SDS and were incubated at 37˚C for 30 min. 20 μl of 4 mg·mL⁻¹ 2-nitrophenyl β-D-galactopyranoside (ONPG) in Z-buffer were then added to each well and mixed thoroughly. Absorbance at 420 nm for each reaction was read once per minute for 1 h at 37˚C in a Fluostar Galaxy plate reader. β-galactosidase activity is expressed in Miller units (MU) [44] [45] .

B. anthracis sporulating cell cultures of 7702 parental and Δba1836 knockout strains were examined by optical and fluorescence microscopy. For membrane visualization the FM4-64 fluorescent dye was added to a final concentration of 1 μg∙ml⁻¹. Sporulating cells in stages II and III were stained with FM4-64 in order to visualize the assymetric septum and engulfment respectively. Optical microscopy was used after stage III of sporulation when the forespore is formed. B. anthracis vegetative cell cultures of 7702 overexpressing the ba1836-gfp fusion were examined by fluorescence microscopy. The images were obtained without fixation on an inverted epifluorescence microscope Nikon E800. Preparation of cell samples for scanning and transmission electron microscopy was performed according to Arnaouteli et al. [28] . Samples were examined using a JEOL JSM-6390LV scanning electron microscope, operating at 20kV. Sections were observed in a JEOL, JEM 2100 transmission electron microscope, operated at 80 kV.

To test the sensitivity of Δba1836 mutant strain in the presence of exogenously added lysozyme, overnight cultures of the parental and mutant strains were diluted to an OD_{600 nm} of 0.1 into 2 cultures of 20 ml of fresh SPY medium for each strain. The cultures were incubated at 37˚C until an OD_{600 nm} of 1.2 and 3.0 corresponding to mid-exponential and early stationary phase of growth respectively. Then each culture was divided in two equal parts of 10 ml, and 10 μg∙ml⁻¹ hen egg lysozyme was added to one of the two subcultures. The growth of both treated and untreated subcultures was monitored.

Autolysis assay was performed according to Balomenou et al. [27] with minor changes. Cultures of 10 ml SPY of 7702 and Δba1836 cells at an OD_{600 nm} of 3.0 were centrifuged for 20 min at 2570 g and the bacterial pellet was concentrated 10 times in SPY medium to which 10 mM NaN₃ was added. The decrease of OD_{600 nm} of each culture was monitored every 30 min for a period of 180 min.

PG from the parental strain 7702 and Δba1836 mutant was purified from stationary growing bacteria (in SPY) as well as from spores, as previously described [46] [47] . Muropeptides were generated from 0.2 mg of spore or vegetative cell PG by digestion with mutanolysin (Sigma) at 37˚C for 16 h and separated by HPLC as previously described [48] .

In order to perform competition experiments, a lacZ reporter construct with the eag promoter (Peag) fused directly to the start codon of the lacZ gene was built. The Peag [49] was amplified from B. anthracis 7702 chromosomal DNA. Primers were synthesized to incorporate a HindIII site at the start and a BamHI site at the end of the Peag (Table S2), which was digested with the corresponding enzymes, and ligated into the multicopy, replicative vector pHT304-18Z. E. coli GM48 (dam⁻) was transformed with the resulting construct to obtain nonmethylated plasmid DNA for electroporation [40] into 7702 B. anthracis cells.

Competition experiments were performed according to Traag et al. [45] with a few changes. Briefly, starting from overnight BHI culture, 7702 strain containing pHT304-18Z-Peag construct was mixed, either in 1:1 or 1:3 ratio with mutant Δba1836 strain containing pHT304-18Z vector in 5 ml modified G medium. Cultures were grown and allowed to sporulate at 30˚C for 48 h. The cultures were then heat treated at 80˚C for 20 min, briefly cooled at room temperature and inoculated in 5 ml fresh modified G medium. Before inoculation, cultures were diluted and plated on agar plates containing 40 μg∙ml⁻¹ 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (X-Gal) and 5 μg∙ml⁻¹ erythromycin. The blue and white colonies were counted. Similarly, competition experiments were performed for vegetative phase of growth. Briefly, starting from overnight BHI culture, 7702 strain containing pHT304-18Z-Peag construct was mixed in 1:1 ratio with mutant Δba1836 strain containing pHT304-18Z vector in 5 ml BHI medium. Cultures were grown at 30˚C for 24 h and then inoculated in 5 ml fresh BHI medium [45] .

Single colonies from 7702 and Δba1836 B. anthracis strains were streaked onto SPY agar plates supplemented with 0.2% glucose and 10 mM MgSO₄ and left to sporulate for 7 days at 30˚C. When needed, the SPY agar plates were additionally supplemented with 0.2% xylose.

Spores were harvested in water, centrifuged at 5900 g for 5 min, washed twice in water, heated at 65˚C for 20 min, washed again and stored at 4˚C. Microscopic examination indicated that all spore preparations were 98% free of contaminating vegetative cells, sporulating cells and germinated spores.

Spore viability was defined as the number of colony-forming units (cfu) formed compared to the number of phase-bright spore particles plated. Purified parental 7702 and Δba1836 spores were tittered by visual counting using a Neubauer haemocytometer (spores/ml), then diluted and plated onto LB agar, with or without prior heat treatment at 80˚C for 20 min, to measure cfu (cfu/ml) [50] .

Spore germination was performed according to Fazzini et al. [51] with certain modifications. Briefly, spores were suspended at an OD_{600 nm} of 1.0 in 1 ml water. After heat activation at 65˚C for 20 min, the suspensions were centrifuged at 5900 g for 5 min and 900 μl were removed. Germination of spores was initiated by the addition of 1ml of a mixture of 25 mM L-alanine and 12.5 mM inosine. The suspensions were incubated at 37˚C for 120 min. Germination was monitored by the loss of optical density at 600 nm.

B. anthracis cell lysates, at the time point at which GFP fluorescence signal was obtained, were separated by SDS-PAGE, blotted, and probed with the following antibodies: polyclonal rabbit anti-GFP primary antibody (Minotech) diluted 1:25,000 and monoclonal anti-rabbit IgG (γ-chain specific) alkaline phosphatase secondary antibody (Sigma) diluted 1:3125.

The genome of B. anthracis Sterne 7702 sequence database revealed 12 coding sequences for putative PDAs of the CE4 family. Eleven of them exhibit more than 90% sequence identity to their homologues from B. cereus sensu stricto. For BA1836, the TMHMM [52] program predicts the presence of an N-terminus transmembrane domain (residues 5 - 23). The TatP [7] [53] and SignalP [53] [54] programs do not predict a signal peptide. BA1836 shares 92% identity with its homologue BC1768 from B. cereus and displays 25% - 30% identity with B. subtilis PDAs. Additionally, BA1836 shares an approximately 30% identity with other putative and known PDAs from B. anthracis.

Importantly, the sequence alignment of BA1836 (excluding residues 1 - 47) with representative members of the CE4 family [24] shows that BA1836 retains the pattern of sequence motifs (MT1-MT5) which are typical for PDAs (Figure 1). These include motif MT2, which is associated with a metal cation binding site formed in the catalytic center of the enzyme, and key catalytic residues (MT1, MT5), which occupy structurally homologous positions in CE4 esterases of known structures [55] . More specifically, Asp⁵³ of motif MT1 and His¹⁹⁶ of motif MT5, corresponding to the catalytic base and the catalytic acid of the deacetylation reaction respectively, are conserved in the sequence of BA1836 (Figure 1). Additionally, the highly conserved Pro¹⁴⁵, which we have recently demonstrated to be autocatalytically hydroxylated in the presence of molecular oxygen at the Cα atom to produce 2-hydroxyproline (2-Hyp), is found in motif MT3, in agreement with other well characterized PDAs. The self-hydroxylation of proline at the Cα atom has been shown to significantly enhance enzymatic activity [56] . The pattern of hydrophobic residues in motifs MT4 and MT5 agrees well with the patterns found in other PDAs (Figure 1).

In order to monitor the expression of the ba1836 gene we constructed a transcriptional reporter by fusing an upstream fragment of ba1836 (containing the

promoter region) with the start codon of the E. coli lacZ gene. 7702 B. anthracis cells carrying this construct were harvested at various time points during vegetative growth. As shown in Figure 2(a), BA1836 promoter driven lacZ expression was absent during exponential growth, but was rapidly induced to high level shortly after cells entered the stationary phase of growth.

We next resuspended early stationary phase 7702 B. anthracis cells carrying this construct into sporulation-inducing medium and collected samples at various time points during sporulation. As we observed by optical microscopy, most cells were in stage II of sporulation (asymmetric division) about 8 - 12 h after transfer into sporulation medium, in stage III (engulfment) after 12 - 16 hours, in stage IV-V (coat and cortex assembly) after 16 - 20 hours and in stage VI-VII (spore maturation and mother cell lysis) after 20 - 24 hours. We observed that expression of ba1836-lacZ further increases during the early stages of sporulation (Stage I-II) (Figure 2(b)). In contrast, when exponentially growing 7702 B. anthracis cells, in which ba1836-lacZ was not expressed, were induced to sporulate, the BA1836 promoter remained silent (Figure S1). This suggests that sporulation conditions alone are not sufficient to induce BA1836 expression.

The amino acid sequence of BA1836 is predicted to include an N-terminal membrane spanning region that is not removed by signal peptidase. In order to determine the subcellular localization of BA1836, a C-terminal gfp fusion of the ba1836 gene was constructed and introduced into B. anthracis 7702. We induced the expression of ba1836-gfp in B. anthracis cells that entered the stationary phase of growth, i.e., when BA1836 was normally expressed. Expression of the gfp-fused BA1836 was confirmed by Western blot analysis (Figure S2). The GFP tag did not affect the function of the protein, because the mutant strain transformed

with the respective BA1836-GFP protein complemented the observed phenotypes (section: Δba1836 cells display severe delay in spore development and Δba1836 knockout strain lowers the spore germination rate). BA1836-GFP localized at the cell membrane and septa in stationary phase cells (Figure 3).

To elucidate the physiological role of BA1836, we constructed a ba1836 gene knockout in B. anthracis 7702 strain according to the method of Janes and Stibitz [34] . Mutant cells were able to grow with no obvious defect and with similar rate as wild-type cells in various liquid media (BHI and SPY broth) indicating that BA1386 is not required for growth and viability of B. anthracis cells (Figure S3(a), Figure S3(b)). Δba1836 knockout cells grown in SPY medium, were not sensitive to lysozyme either in the exponential or in the stationary phase of growth (Figure S4(a)) suggesting that BA1836 is not a bona fide PGDA involved in resistance to host lysozyme. The autolytic activity of Δba1836 and 7702 strains was tested by addition of NaN₃, a known inducer of autolysis in growing cultures. Δba1836 cells showed decreased autolysis at these conditions compared to the parental strain (Figure S4(b)). The PG pattern from stationary phase cells of Δba1836 mutant was similar to that of the parental strain (Figure S5).

In order to examine the morphology of Δba1836 cells, we collected samples during exponential and stationary growth phase in SPY medium and visualized cells by scanning electron microscopy (SEM). Δba1836 daughter cells grown exponentially displayed morphologies similar to that of the parental strain, consistent with the lack of ba1836 expression under these conditions (Figure S4(c)). However, stationary phase Δba1836 cells displayed chains of unseparated cells (5.15 ± 0.14 μm, n = 190) compared to the mostly single cell morphology (3.21 ± 0.18 μm, n = 175) of the parental strain (Figure S4(c)).

Since ba1836 gene was expressed upon entry to the stationary phase of growth and was further induced during sporulation, we next examined the morphology of spores produced by Δba1836 by SEM and TEM imaging. No significant differences were observed between mature spores of Δba1836 and the parental strain, suggesting that Δba1836 cells can produce morphologically and structurally normal spores with similar cortex thickness (Figure S6(a)). The efficiency of Δba1836 spore formation, their viability and their heat resistance was similar to that of the parental strain (Figure S6(b)). To test for a more subtle role of ba1836 in spore formation, competition experiments were conducted in which Δba1836 strain competed against 7702 wild type strain for several rounds of sporulation. In 7702 wild type strain, lacZ is under the transcriptional control of the eag promoter. The competition experiment was initially performed with a ratio of 1:1 (wild type to mutant cells) (Figure 4(a)) but since Δba1836 cells showed a severe competition deficit, the mutant cells were added in threefold excess (Figure 4(b)). Over the course of seven days of co-culturing the percentage of the mutant strain decreased from 75% to 1% of the population (Figure 4(b)). On the contrary, Δba1836 mutant strain had no fitness defect under vegetative growth conditions (Figure 4(c)). Complementation of the competition deficit under sporulation-inducing conditions of Δba1836 strain could not be performed since the addition of xylose to the cultures inhibited entering of cells to sporulation. These results strongly suggest that BA1836 is actively involved in sporulation and/or germination processes.

Since we observed a competition deficit under sporulation—inducing conditions of Δba1836 knockout cells, we next examined every stage of sporulation by optical microscopy. Interestingly, the mutant cells showed a remarkable delay in spore development (Figure 5). At 13 hours post-resuspension into sporulation—inducing medium, the majority of 7702 mother cells have engulfed the forespore (stage III of sporulation). In contrast, at the same time point the majority of the Δba1836 cells only entered assymetric septation (stage II of sporulation) (Figure 5). Four hours later, almost every 7702 mother cell has produced a mature forespore (stage VI of sporulation), while the majority of the Δba1836 mother cells have only entered stage III of sporulation (Figure 5). Mature forespores were visible in Δba1836 mother cells 21 hours post resuspension into sporulation—inducing medium (Figure 5). These observations show that Δba1836 cells display a severe delay in spore development, consistent with the expression of ba1836 in this developmental stage.

The PG pattern from mature spores of Δba1836 and 7702 strains were analyzed by HPLC [47] . The digestion of PG from Δba1836 spores with mutanolysin produced an identical overall muropeptide profile compared to PG from wild type spores, although the amount of muropeptides released from the mutant was significantly reduced (Figure 6). Expressing ba1836-gfp in Δba1836 restored the wild-type spore PG profile (Figure 6). Complementation was also attempted using the native protein but no expression was observed in Δba1836 cells (data not shown).

Finally, we determined the rate of germination of Δba1836 mutant strain as a

reduction in spore-specific absorbance upon the addition of 25 mM L-alanine and 12.5 mM inosine as germinants. As shown in Figure 7, when ba1836 gene was deleted, spores were capable of initiating germination but, compared to wild-type, the Δba1836 spores had a lower rate of germination. Fifteen minutes after the addition of germinants there was almost 40% loss of absorbance for 7702 spores compared to a 20% loss for Δba1836 spores. Δba1836 spores lost almost 40% of the absorbance after 90 min. These results indicate that the absence of BA1836 results in a delay in germination. The rate of germination of Δba1836 mutant spores was partially complemented by expressing either BA1836-GFP or the predicted catalytically inactive BA1836 D53A-GFP, indicating that putative deacetylase activity is not required to complement this phenotype (Figure 7).