ExsB, an unusually highly phosphorylated protein required for the stable attachment of the exosporium of Bacillus anthracis

Summary

The outermost layer of the Bacillus anthracis spore, the exosporium, is composed of a paracrystalline basal layer and an external hair-like nap. The nap is formed from a single collagen-like glycoprotein, while the basal layer contains many different proteins, including a 186-amino acid protein called ExsB. In this study, we discovered that ExsB is unusually highly phosphorylated, with at least 14 of its 19 threonine residues modified. The phosphorylated threonines are included in seven contiguous approximately 12-residue imperfect repeats, which presumably contain kinase recognition sequences. We demonstrated that a B. anthracis ΔexsB mutant unable to synthesize ExsB produced spores with an exosporium that was readily sloughed, indicating that ExsB was required for stable exosporium attachment. This unstable exosporium also lacked the enzyme alanine racemase, which is normally tightly associated with the exosporium. Additionally, purified ΔexsB spores lacking a visible exosporium were devoid of most exosporium proteins but, surprisingly, retained the putative exosporium proteins BxpC and CotB-1. Finally, we showed that transcription of the exsB gene occurred only during the late stages of sporulation, and we used an active and phosphorylated ExsB-EGFP fusion protein to monitor ExsB localization to wild-type and ΔbxpB mutant exosporia.

Introduction

Bacillus anthracis, the causative agent of anthrax, is a Gram-positive, rod-shaped, aerobic, soil bacterium that forms spores when it is starved for nutrients (Mock and Fouet, 2001; Priest, 1993). Spore formation begins with an asymmetric septation that divides the starved vegetative cell into a mother cell compartment and a smaller forespore compartment, each of which contains a copy of the genome (Hilbert and Piggot, 2004). The mother cell subsequently engulfs the forespore and surrounds it with three protective layers: the innermost cortex composed of peptidoglycan, a closely apposed proteinaceous coat, and a loosely fitting exosporium (Henriques and Moran, 2007). After complete development of the forespore, the mother cell lyses and releases the mature spore, which can remain dormant and survive in harsh environments for many years (Nicholson et al., 2000). When spores encounter an aqueous environment containing sufficient nutrients, they can germinate and grow as vegetative cells (Moir, 2006). Anthrax is typically caused by contact with spores, which germinate and propagate within the animal or human host (Mock and Fouet, 2001).

As the outermost layer of the B. anthracis spore, the exosporium is the primary site of contact with the environment and the first point of contact with host defenses. Recent studies indicate that the exosporium participates in spore uptake by host phagocytes and contributes to the survival of the engulfed spore, both critical steps in disease progression (Cleret et al., 2007; Kang et al., 2005; Oliva et al., 2009; Oliva et al., 2008; Weaver et al., 2007). The exosporium also serves as a semi-permeable barrier that excludes potentially harmful macromolecules, and its unique surface molecules can be used for spore detection (Gerhardt and Ribi, 1964; Swiecki et al., 2006; Tamborrini et al., 2009). The B. anthracis exosporium is a prominent bipartite structure comprised of a paracrystalline basal layer and an external hair-like nap (Ball et al., 2008). The filaments of the nap are formed by trimers of a single collagen-like glycoprotein called BclA (Boydston et al., 2005; Sylvestre et al., 2002), and the basal layer contains approximately twenty different proteins in tight and loose associations (Redmond et al., 2004; Steichen et al., 2005).

One of these basal layer proteins is ExsB (exosporium protein B), which was initially detected by amino-terminal sequencing of proteins in preparations of purified B. cereus exosporium (Todd et al., 2003). B. cereus is closely related to B. anthracis, and the exosporia of these two species are similar in structure and protein composition (Ball et al., 2008; Rasko et al., 2005; Todd et al., 2003). The characterized ExsB of B. cereus contains 192 amino acids, and apparently the first 17 amino-terminal residues are removed by proteolytic processing. This protein has an unusual amino acid composition, including 20 cysteine residues, 67 charged—mostly basic—residues, and several repeat sequences. In a subsequent study, fragments of ExsB were detected in a proteomic analysis of the B. anthracis exosporium; however, the assignment of a location within the exosporium was tenuous in this analysis because the exosporium preparations were contaminated with many fortuitously bound cellular proteins (Liu et al., 2004). The ExsB encoded by the B. anthracis genome contains 186 amino acids, and it is 89% identical to the B. cereus protein. The genomic contexts of the exsB genes of B. cereus and B. anthracis are identical, and they are equivalent to the genomic context of the 195-codon cotG gene of Bacillus subtilis. Each of these three genes appears to be included in a single-gene transcriptional unit that is located adjacent to a divergently-transcribed cotH gene (B. subtilis) or a cotH homologue (Kanehisa et al., 2006). Additionally, the ExsB and CotG proteins possess similar amino acid sequences in their amino-terminal and carboxyl-terminal regions (Todd et al., 2003), suggesting homologous functions. B. subtilis does not possess an obvious exosporium, however, and CotG is a coat protein in B. subtilis spores (Henriques and Moran, 2007).

In the study reported here, we identify ExsB in highly purified B. anthracis exosporium preparations. We demonstrate that ExsB is phosphorylated, unlike any other known exosporium protein, and we show that ExsB is more extensively phosphorylated than any previously characterized bacterial phosphoprotein (Thomson and Gunawardena, 2009). We also show that ExsB is required for normal assembly of the exosporium and its stable attachment to the spore coat. Furthermore, we demonstrate that the gene encoding ExsB is expressed only during the late stages of sporulation. Finally, we use an ExsB-EGFP (enhanced green fluorescence protein) fusion protein to monitor ExsB localization during sporulation. ExsB localization is examined in both a wild-type strain and in a mutant strain unable to produce the exosporium basal layer protein BxpB (also called ExsFA), which is required for extensive attachment of the filaments of the hair-like nap (Steichen et al., 2005; Sylvestre et al., 2005; Tan and Turnbough, 2010).

Results

Identification of multiply-phosphorylated ExsB in the exosporium of B. anthracis

The identification of B. anthracis exosporium proteins has been an ongoing project for the last several years (Redmond et al., 2004; Steichen et al., 2005; Sylvestre et al., 2002). As part of our effort to complete the list of exosporium proteins, we analyzed proteins in purified preparations of exosporia obtained from mutant spores lacking a particular exosporium protein (e.g., BclA, BxpA, BxpB, CotB-1, CotB-2, CotY, and ExsY). The rationale for this approach was that certain proteins might be extractable only from a partially assembled exosporium. During this analysis, exosporium proteins and stable complexes of these proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Inspection of the gels revealed an interesting difference: a broad band of high-molecular-mass protein complexes was detected in samples of exosporia lacking BclA or BxpB but not in samples of the other mutant exosporia or in a control sample of wild-type exosporia (Fig. 1 and data not shown). Each broad band was excised from the gel and treated in situ with trypsin, and the resulting tryptic fragments were separated by liquid chromatography and sequenced by tandem mass spectrometry (LC-MS/MS). Tryptic fragments of the exosporium proteins alanine racemase (Alr), CotY, ExsK, and ExsY were detected in the sample lacking BxpB, and tryptic fragments of BxpB, CotY, ExsK, and ExsY were detected in the sample lacking BclA.

Fig. 1.

High-molecular-mass exosporium protein complexes analyzed by SDS-PAGE. Exosporium samples were obtained from spores lacking either BxpB (ΔbxpB) or BclA (ΔbclA) or from wild-type (WT) spores. Proteins were visualized by staining with Coomassie Blue. The filled arrowhead points to the broad band of complexes unique to ΔbxpB and ΔbclA spores. Gel locations and molecular masses of protein standards are shown.

In addition, tryptic fragments of ExsB were detected in both samples; 13 and 11 fragments were detected in the samples lacking BclA and BxpB, respectively (Table 1). These fragments represented 51% of the sequence of ExsB. Remarkably, every threonine residue included in these fragments, which represented 14 of the 19 ExsB threonines, was phosphorylated (Table 1). Phosphorylation of each threonine residue was indicated by a signature change in mass. The MS/MS spectrum of the tryptic fragment including residues 54 to 67, which contains three phosphothreonines, is shown in Fig. 2. Peptide fragmentation produced both positive y-ions and derivatives of these ions that had lost mass in increments of 98 Da, the mass of H₃PO₄. In addition, the masses of all threonine-containing y-ions were greater than those predicted for non-phosphorylated ions, with the increase in mass equal to the mass of one phosphate addition (80 Da) per threonine residue.

Table 1.

Tryptic fragments of ExsB analyzed by LC-MS/MS

Fragment	Sequence	No. of phosphothreonines	Sourcea
38–48b	IECSDDCNCPR	0	A
54–67	VKHCpTFVpTKCpTHVK	3	B
56–62	HCpTFVpTK	2	A
56–67	HCpTFVpTKCpTHVK	3	A
68–77	KWpTFVpTKCpTR	3	A, W
69–77	WpTFVpTKCpTR	3	W
68–79	KWpTFVpTKCpTRVR	3	B
80–91	VQKWpTFVpTKVpTR	3	B, A, W
83–91	WpTFVpTKVpTR	3	A, W
94–101	ECVLVpTKR	1	A
94–103	ECVLVpTKRpTR	2	B, W
94–104	ECVLVpTKRpTRR	2	B
104–112	RKHCpTFVpTK	2	B
105–112	KHCpTFVpTK	2	B, A
105–115	KHCpTFVpTKCVR	2	A
106–112	HCpTFVpTK	2	A
106–115	HCpTFVpTKCVR	2	B
119–125	KFYWpTKR	1	B, A
120–125	FYWpTKR	1	B, A, W
164–172	GGHKFPSCK	0	B
175–182	KFDHFWYK	0	W
176–182	FDHFWYK	0	A, W

^aIndicates source of the fragment: A, exosporia lacking BclA; B, exosporia lacking BxpB; W, wild-type exosporia.

^bNumbers indicate the positions within ExsB of the amino acids included in the tryptic fragment.

Fig. 2.

Deconvoluted positive ion MS/MS spectrum of peptide 54–67 containing three phosphorylated threonine residues. Endpoints of y-ions and b-ions are indicated on the peptide sequence.

To determine whether phosphorylated ExsB is a normal component of the exosporium, we reexamined the purified wild-type exosporium sample. In this instance, we washed the exosporium preparation with a solution containing 5 mM Tris-HCl (pH 8.0), 0.2% SDS, and 3 mM CaCl₂ prior to analysis by SDS-PAGE. Apparently, this treatment facilitated the solubilization of the exosporium, because now a new band of high-molecular-mass protein complexes was detectable at the top of the gel (data not shown). The proteins in this band were digested with trypsin and the tryptic fragments were analyzed by LC-MS/MS as described above. Eight tryptic fragments of ExsB were detected (Table 1). Again, all threonine residues included in these fragments, which represented 9 of the 14 threonines examined in the mutant exosporium samples, were phoshorylated (Table 1). This result indicated that extensively phosphorylated ExsB is normally present in the B. anthracis exosporium.

Phosphothreonines are included in contiguous imperfect repeat sequences

The central region of ExsB is composed of seven contiguous, approximately 12-residue imperfect repeat sequences, which contain 16 of the 19 ExsB threonines (Fig. 3). The last six repeat sequences contain the 14 phosphorylated threonine residues shown in Table 1. The first repeat sequence contains the other two threonines in the central region of ExsB. The phosphorylation state of these two residues is unknown, however, because the short tryptic fragments representing the first repeat sequence were not detected in our LC-MS/MS analysis. Similarly, the phosphorylation states of the three threonines not included in the central repeat region of ExsB were not determined. Tryptic fragments representing these threonines, all of which are located in the amino-terminal domain of the protein, were not detected in our analysis (Fig. 3). Interestingly, the central repeat region does not include any of the six ExsB serine residues, which are equally divided between the terminal domains of the protein (Fig. 3). Two of the serine residues, S41 and S170, were included in the sequenced tryptic fragments shown in Table 1. Neither serine residue was phosphorylated.

Fig. 3.

Amino acid sequence of B. anthracis ExsB. The sequence is divided into amino- and carboxyl-terminal domains and a central, repeat region including residues 45 through 129. The seven imperfect repeats in the central region are aligned; the dash in the first repeat sequence represents a gap used for alignment purposes only. Asterisks indicate the 14 known phosphothreonines in the last six repeat sequences. The arrowhead indicates the site of proteolytic cleavage, which is known to occur in B. cereus and presumed to occur in B. anthracis.

ExsB is required for the stable attachment of the exosporium

To examine the function of ExsB in spore formation, we constructed a ΔexsB mutant strain of B. anthracis in which the entire exsB gene was deleted. This strain was grown in liquid medium to produce spores, which were purified by washing them with water and centrifuging them through a density gradient. The gradient separates spores from vegetative debris and exosporium fragments. The purified spores were inspected for an intact exosporium by fluorescence microscopy following treatment with a fluorescently labeled anti-BclA monoclonal antibody (mAb). Spores with an intact exosporium (e.g., wild-type spores) appear brightly and uniformly fluorescent in this procedure (Fig. 4A–C). In the case of the ΔexsB spores, however, approximately 95% of the spores were non-fluorescent (Fig. 4G–I), indicating that they lacked an exosporium (or at least the hair-like nap). The minor fraction of spores that were fluorescent (Fig. 4H–I) looked essentially the same as stained wild-type spores. To examine the possibility that the ΔexsB spores lost their exosporia during purification, we prepared another batch of spores in liquid medium and washed them with water, but did not purify them further. These spores will be referred to hereafter as “washed” spores (as opposed to “purified” spores). The washed spores were treated with a fluorescently labeled anti-BclA mAb and examined by fluorescence microscopy. Again, approximately 95% of the spores were non-fluorescent (Fig. 4J–L). In addition, the sample contained a large number of fluorescently labeled free exosporia, which apparently had been sloughed from the majority of the spores (Fig. 4K–L). Such sloughing was not observed with washed wild-type spores (Fig. 4D–F).

Fig. 4.

Phase-contrast and fluorescence micrographs of wild-type and ΔexsB spores prepared in a liquid or on a solid medium. Spores were either purified by density-gradient centrifugation or simply washed with water. Spores were treated with an Alexa Fluor 488-labeled anti-BclA mAb prior to analysis. The choice of growth medium had no effect on the appearance of wild-type spores. Merged phase-contrast and fluorescent images also are shown. The sample of purified ΔexsB spores prepared in liquid medium includes a rare fluorescent spore (panel I). The sample of washed ΔexsB spores prepared in liquid medium includes large exosporium fragments that are easily seen only by fluorescence (panel L). The fluorescence micrograph of washed ΔexsB spores prepared on solid medium includes a bright background apparently due to fine exosporium debris (panel Q).

Recently, we reported that exosporium development is arrested after the formation of a cap-like structure in a B. anthracis strain unable to synthesize ExsY (Boydston et al., 2006). The exosporium cap was lost from spores produced in liquid medium but was retained by spores produced on solid medium. To examine the possibility that these growth conditions also affected the retention of exosporia by ΔexsB spores, we grew the ΔexsB strain on agar plates until sporulation was essentially complete. We used this culture to prepare both washed and purified spores. Each spore preparation was treated with a fluorescently labeled anti-BclA mAb and examined by fluorescence microscopy (Fig. 4M–R). In both cases, approximately 98% of the spores were brightly and uniformly fluorescent, indicating that only minor exosporium sloughing occurred on solid medium. The only difference between the microscopic images of the washed and purified spore samples was that the washed sample had a bright fluorescent background (Fig. 4Q), indicating the presence of small pieces of exosporia. Presumably, these pieces were produced by the breakdown of exosporia that had been sloughed by a small fraction of the spores.

To verify the presence or absence of an exosporium on ΔexsB spores, we used transmission electron microscopy to inspect purified mutant spores that had been prepared either in liquid medium or on agar plates. Wild-type spores grown on both media were also examined as controls. We inspected over 100 spores in each case. All wild-type spores retained their exosporia (Fig. 5A). All ΔexsB spores prepared on liquid medium lacked an exosporium (Fig. 5B). Approximately 98% of ΔexsB spores prepared on agar plates retained their exosporia, while roughly 2% of these spores sloughed all or part of their exosporia. The shed exosporium fragments were readily observable in electron micrographs (Fig. 5C). Taken together, these results were consistent with the results obtained by fluorescence microscopy and indicated that the preparation of spores for electron microscopy might have caused a slight increase in sloughing of exosporia from ΔexsB spores, at least in the case of spores prepared in liquid medium.

Fig. 5.

Transmission electron micrographs of thin sections of purified wild-type and ΔexsB spores. Wild-type spores were prepared in liquid medium (A), and ΔexsB spores were prepared either in a liquid medium (B) or on a solid medium (C). The electron micrograph of ΔexsB spores prepared on solid medium shows a rare spore that has sloughed part of its exosporium (black arrow). White arrowheads indicate selected intact exosporia, and black arrowheads indicate selected exosporium fragments. The core, cortex, and coat of one spore in panel A are labeled.

Exosporium protein content of ΔexsB spores

To verify that ΔexsB spores lacking a visible exosporium had lost a representative set of exosporium proteins, we produced a new batch of ΔexsB spores in liquid medium and purified these spores by density-gradient centrifugation. Virtually all spores in this preparation were non-fluorescent following treatment with a fluorescently labeled anti-BclA mAb, indicating they were missing their exosporia. Proteins on the surface of these spores were extracted and separated by SDS-PAGE, and the presence of selected exosporium proteins was determined by immunoblotting with mouse mAbs specific for BclA, Alr, BxpB, BxpC, CotB-1, or ExsY/CotY (Fig. 6, L lanes and data not shown). The anti-CotB-1 mAb did not react with its homologue CotB-2 (data not shown), and the anti-ExsY/CotY mAb reacted equally with ExsY and its similarly-sized homologue CotY (Boydston et al., 2006). An equal number of wild-type spores were analyzed in each immunoblot as a positive control. The results showed that BclA, Alr, BxpB, and ExsY/CotY were not present on the ΔexsB spores, as expected (Fig. 6 and data not shown). Surprisingly, both BxpC and CotB-1 were present on the mutant spores. The level of BxpC extracted from the mutant spores was approximately one-half that obtained from wild-type spores, and the level of CotB-1 extracted from the ΔexsB spores was several times greater than that observed with wild-type spores (Fig. 6). Evidently, the ΔexsB spores lacking exosporia retained some exosporium proteins.

Fig. 6.

Exosporium proteins of ΔexsB spores. Exosporium proteins were extracted from an equal number of control wild-type spores (WT) and from ΔexsB spores prepared in liquid medium (L) or on solid medium (S). Proteins were analyzed by immunoblotting with a mAb specific for the indicated exosporium protein. Only relevant regions of blots are shown, and the gel locations and molecular masses (in kDa) of protein standards are indicated..

The exosporium proteins of ΔexsB spores produced on agar plates and purified by density-gradient centrifugation were also examined by immunoblotting as described above (Fig. 6, S lanes and data not shown). Approximately 98% of these spores possessed an intact exosporium. The blots revealed that the levels of BclA, BxpB, BxpC, and ExsY/CotY were the same as those of wild-type spores (Fig. 6 and data not shown). The level of CotB-1 was several times greater than that observed with wild-type spores and essentially the same as the level of CoB-1 extracted from ΔexsB spores produced in liquid medium (Fig. 6). Unexpectedly, no Alr was detected with ΔexsB spores produced on solid medium (Fig. 6). Wild-type levels of Alr (as well as BclA, BxpB, BxpC, CotB-1, and ExsY/CotY) were detected in sporulating cells of the ΔexsB strain, however (data not shown). Apparently, Alr is not securely incorporated into the exosporium of ΔexsB spores, indicating that the ΔexsB exosporium is structurally different from the wild-type exosporium.

Other properties of ΔexsB spores

To investigate the physiological effects of the ΔexsB mutation, we compared the growth rates, cell yields, and sporulation efficiencies of cultures of the wild-type and ΔexsB strains grown in liquid sporulation medium at 37°C with shaking. We observed no significant differences between the two cultures; both grew with maximum doubling times of approximately 25 min and sporulated with >95% efficiency. Using standard protocols (Nicholson and Setlow, 1990), we compared the heat and lysozyme resistances of purified wild-type and ΔexsB spores. We included ΔexsB spores produced in liquid medium and on solid medium in this comparison. Again, we found no significant differences in the survival of wild-type and ΔexsB spores (data not shown). These results indicated that the coat and cortex layers in the mutant spores, which are the primary resistance determinants, were effectively intact (Setlow, 2006).

Transcription of the exsB gene during the B. anthracis cell cycle

The genes encoding proteins involved in exosporium assembly are transcribed from promoters recognized by the two mother-cell specific sigma factors σ^E and σ^K, which are active during the early and late stages of sporulation, respectively (Helmann and Moran, 2002). To determine the time of transcription of the exsB gene, we used quantitative reverse transcription polymerase chain reaction (RT-PCR) to measure the levels of exsB transcripts in cells harvested during vegetative growth and throughout sporulation. The results showed that significant levels of exsB transcripts were detected only during the late stages of sporulation, when σ^K was active (Fig. 7A). A sequence closely resembling the consensus sequence for a σ^K promoter was located 81 bases upstream of the exsB initiation codon (Fig. 7B). A recent B. anthracis transcriptome analysis indicated that there were multiple transcription start sites located downstream of the promoter that we identified (Passalacqua et al., 2009); however, these start sites were somewhat further downstream than expected for the putative promoter (Fig. 7B). Additionally, none of these start sites was preceded by a promoter-like sequence, suggesting that the 5′ ends of the transcripts used for start site mapping had been nucleolytically shortened. Each of these transcripts could have been derived from a single transcript initiated at the putative σ^K-specific promoter.

Fig. 7.

Transcription of the exsB gene during the B. anthracis cell cycle. (A) Cellular levels of exsB transcripts were measured by RT-PCR during vegetative growth and sporulation. Cells harvested 1 h before (−1 h) the start of sporulation were in the late exponential phase of growth. Mother cell lysis to release free spores began 9 h after the onset of sporulation. PCR products were visualized by gel electrophoresis and staining. Size standards were included in the lane marked Std. (B) Sequence of the exsB promoter region. The −35 (CA) and −10 (CATA---T) regions of a possible σ^K-specific promoter are labeled and underlined; the consensus sequence for σ^K-specific promoters is ACA—16 bases—CATA---T (Eichenberger et al., 2004). Asterisks indicate five putative start sites identified by transcriptome analysis. The exsB Shine-Dalgarno (SD) sequence is underlined and italicized.

Localization of an ExsB-EGFP fusion protein during sporulation

To monitor the localization of ExsB during sporulation, we constructed a multi-copy plasmid carrying a recombinant operon consisting of the exsB promoter-leader region, the exsB gene fused in frame to the gene encoding EGFP, and an intrinsic transcription terminator. As a control, we constructed an analogous recombinant plasmid in which the gene fusion was replaced by the gene encoding EGFP. The two plasmids were transformed individually into a B. anthracis ΔexsB mutant strain unable to produce ExsB. The resulting transformants were grown in liquid medium, and sporulating cells (sporangia) and free spores produced by these strains were examined by fluorescence and phase-contrast microscopy (Fig. 8).

Fig. 8.

Location of the ExsB-EGFP fusion protein during sporulation. ExsB-EGFP was encoded by a plasmid-borne exsB::egfp gene fusion that was expressed in a B. anthracis strain lacking wild-type ExsB. The expression strain was allowed to sporulate in liquid medium, and the location of ExsB-EGFP was monitored by fluorescence and phase-contrast microscopy at 5, 7, and 9 h after the onset of sporulation. Fluorescence (Fluor.) images are shown in panels A, C, and E, and merged fluorescence and phase-contrast images are shown in panels B, D, and F. As a control, the location of similarly expressed EGFP was monitored in the same manner (panels G–L).

In the case of the strain expressing ExsB-EGFP, green fluorescence was first detected in the sporangium approximately 5 h after the onset of sporulation; this fluorescence was associated with a cap-like region surrounding one end of the developing forespore (Fig. 8A–B). This cap-like region was always located within the sporangium at the end of the forespore at which exosporium assembly begins; this end of the forespore is positioned near the middle of the sporangium (Steichen et al., 2007). This result suggested that the observed fluorescent cap-like region was the previously described cap region of the exosporium (Boydston et al., 2006). Between 6 and 7 h after the start of sporulation, when exosporium assembly was completed, all developing forespores were fully surrounded by green fluorescence (Fig. 8C–D). The fluorescence was generally uniform, although one or two foci of more intense fluorescence were often apparent. The origin of the foci remains to be determined. Between 8 and 9 h after the start of sporulation, the mother cells began to lyse and release free spores, which were also completely surrounded by green fluorescence (Fig. 8E–F). These spores appeared to possess a complete and stable exosporium based on the fact that, following their purification, wild-type levels of BclA and ExsY/CotY and nearly wild-type levels of Alr were extracted from them (data not shown). Thus, the ExsB-EGFP fusion protein appeared to possess normal ExsB activity capable of complementing the defects in exosporium attachment and assembly caused by the absence of ExsB. In the case of the control strain expressing EGFP, green fluorescence was also detected in the sporangium during the late stages of sporulation; however, this fluorescence was not associated with the developing forespore (Fig. 8G–J) or with free spores (Fig. 8K–L). This result indicated that the ExsB portion of ExsB-EGFP directed the fluorescence localization observed with the fusion protein.

To further examine the apparent normal activity of the ExsB-EGFP fusion protein, we examined the extent of phosphorylation of this protein following its isolation from mature spores. The fusion protein was digested with a mixture of trypsin and chymotrypsin, and the resulting proteolytic fragments were analyzed by LC-MS/MS. These fragments represented 41% of the sequence and 14 of the 19 threonines of ExsB (Table 2). Eleven of the threonines examined were found only as phosphorylated residues, and all 11 of these residues were included in the central repeat region of ExsB. Two other threonines (T58 and T61), which are also included in the central repeat region, were found in three versions of the tryptic fragment containing residues 56 through 62. In two versions, a single and different threonine was phosphorylated; in the third version, both threonines were phosphorylated. The last threonine residue analyzed (T27) was included in a chymotryptic fragment that included residues 21 through 28 of the amino-terminal domain of ExsB. Unlike the other threonines in this analysis, T27 was not included in the previous analysis of ExsB phosphorylation (Table 1). Residue T27 was not phosphorylated. Taken together, these results indicated that ExsB-EGFP was extensively phosphorylated but perhaps to a slightly lesser extent than that observed with ExsB. This marginal under-phosphorylation might be due to the over-expression of the fusion protein, at least in the case of residues T58 and T61.

Table 2.

Proteolytic fragments of ExsB-GFP analyzed by LC-MS/MS

ExsB fragment	Sequence	No. of phosphothreonines
21–28a	LHQDPSTF	0
38–48	IECSDDCNCPR	0
56–62	HCTFVpTK	1
56–62	HCpTFVTK	1
56–62	HCpTFVpTK	2
68–74	KWpTFVpTK	2
68–77	KWpTFVpTKCpTR	3
83–91	WpTFVpTKVpTR	3
94–103	ECVLVpTKRpTR	2
105–112	KHCpTFVpTK	2
120–125	FYWpTKR	1
176–182	FDHFWYK	0

^aNumbers indicate the positions within ExsB of the amino acids included in the fragment.

Aberrant localization of the ExsB-EGFP fusion protein during sporulation of a ΔbxpB mutant

We also monitored the localization of the ExsB-EGFP fusion protein during sporulation of a mutant B. anthracis (ΔbxpB) strain unable to produce the key basal layer structural protein BxpB. For this purpose, the recombinant exsB::egfp-containing plasmid described above was introduced by transformation into a ΔbxpB strain, generating a transformant that over-expressed the ExsB-EGFP fusion protein and expressed normal levels of wild-type ExsB. This strain was grown in liquid medium, and the sporangia and free spores produced were examined by fluorescence and phase-contrast microscopy (Fig. 9). As observed above, green fluorescence was first detected in sporangium approximately 5 h after the onset of sporulation, and this fluorescence was associated with the cap region of the exosporium (Fig. 9A–B). In contrast to the results described above, there was no change in the pattern of fluorescence associated with the forespore after 7 h of sporulation (Fig. 9C–D) or with free spores released from the mother cell after 9 h of sporulation (Fig. 9E–F). This truncated pattern was not due to an interruption in basal layer assembly; both forespores after 7 h of sporulation and free spores possessed a complete basal layer as observed by electron microscopy (Steichen et al., 2005). These results indicated that BxpB was required for distribution of ExsB-EGFP and, presumably, ExsB beyond the exosporium cap. Additionally, as a control for the experiment shown in Fig. 9, we demonstrated that normal levels of wild-type ExsB did not affect the localization of over-expressed ExsB-EGFP (data not shown).

Fig. 9.

Location of the ExsB-EGFP fusion protein during sporulation of a ΔbxpB mutant strain. Expression and monitoring the location of ExsB-EGFP was as described in the legend to Fig. 8.

Discussion

ExsB is the first known phosphorylated exosporium protein. In fact, to our knowledge, no other phosphoprotein has ever been found in any Bacillus spore integument. ExsB also appears to be the most extensively phosphorylated bacterial protein characterized to date (Thomson and Gunawardena, 2009). ExsB contains at least 14 phosphothreonines located in the central region of the protein, which is composed of seven contiguous, approximately 12-residue imperfect repeat sequences. This repeat region contains two other threonine residues that we were unable to characterize, but the sequences flanking these residues are similar to the flanking sequences of the known phosphothreonines. Thus, we presume that these two uncharacterized threonines are also phosphorylated. This brings the total number of known and presumed phosphothreonines to 16. ExsB contains 3 threonine and 6 serine residues located within the amino- and carboxyl-terminal domains of the protein. One of these threonines and two of the serines were included in our LC-MS/MS analyses of ExsB and ExsB-EGFP proteolytic fragments. None of these residues appeared to be phosphorylated. This result might indicate that the signals for threonine and serine phosphorylation are restricted to the repeat region of ExsB.

Using the recently created protein phosphorylation site prediction program NetPhosBac (Miller et al., 2009), which is based on over 150 serine and threonine phosphorylation sites in B. subtilis and Escherichia coli and presumes that all bacterial S/T kinases recognize a common phosphorylation site, we found 8 predicted phosphorylation sites in ExsB. One site was S14, which was not included in our analysis and is included in the 17-residue amino-terminal fragment that is presumably proteolytically removed during ExsB maturation. Another site was T49, which is one of the two threonines in the repeat region that was uncharacterized but presumed phosphorylated. The other 6 predicted sites were threonines (positions, 58, 73, 76, 87, 99, and 108) that we determined were phosphorylated. Surprisingly, the prediction program failed to identify eight confirmed sites of phosphorylation that closely resembled the predicted sites. The phosphorylation sites identified in ExsB should, therefore, be useful in the refinement of current prediction parameters.

Based on our analysis of ExsB and over-expressed ExsB-EGFP, it seems likely that all 16 threonines in the ExsB repeat region are normally phosphorylated. This high level of phosphorylation might compensate for the abundance of positively charged residues in ExsB, which contains 44 positively charged amino acids out of a total of 68 charged residues. The salt bridges formed between the multiple phosphate groups and positively charged side-chains of arginine and lysine residues would be expected to contribute significantly to the structure of ExsB. In particular, the multiple salt bridges could be required for the formation of a stable protein structure. Consistent with this proposal, we observed that when ExsB was expressed in E. coli, presumably in an unphosphorylated state, the repeat region of the protein was extremely sensitive to proteolytic cleavage (unpublished data).

In addition to being an unusual exosporium protein, ExsB plays a critical role in spore development. Our results demonstrated that ExsB is required for the stable attachment of the exosporium to an underlying spore structure, presumably the spore coat. The nature of this attachment is unknown, but it is clear that connections between the exosporium and an underlying layer exist. Evidence of such connections is provided by the retention of the exosporium cap during extensive purification of ΔexsY spores (Boydston et al., 2006). However, the attachments demonstrated by cap retention might be different from the attachments to which ExsB contributes. Interestingly, the lack of ExsB does not prevent the assembly of an exosporium that completely surrounds the spore, just the ability of this exosporium to stably associate with the rest of the spore. This unstable exosporium appears to be deficient in at least one exosporium protein, the spore-specific alanine racemase Alr. The absence of Alr does not appear to be the factor affecting normal attachment of the exosporium, however, because spores lacking only Alr possess a stably attached exosporium (Chesnokova et al., 2008). It remains a possibility that additional exosporium proteins are absent or that exosporium assembly is aberrant in sporulating cells lacking ExsB and that one of these conditions is the primary reason for unstable exosporium attachment. The specific direct or indirect effects of ExsB on exosporium attachment remain to be determined.

Another interesting observation related to spore formation in cells lacking ExsB was that the amount of CotB-1 extracted from ΔexsB spores with or without an exosporium was several times greater than that extracted from wild-type spores. ExsB and CotB-1 are homologous or at least analogous to the B. subtilis outer coat proteins CotG and CotB, respectively. CotG has been shown to directly interact with CotB and is required for a posttranslational modification of CotB that converts it from a 44-kDa form (CotB-46) into a 66-kDa form (CotB-66). Only CotB-66 is incorporated into the coats of wild-type spores, while CotB-44 accumulates in the coats of cotG-null spores (Zilhão et al., 2004). Thus, it seems reasonable to suspect that ExsB is involved in the incorporation of CotB-1 into the B. anthracis exosporium. In the absence of ExsB, CotB-1 could be aberrantly incorporated, perhaps into the coat instead of the exosporium, and therefore be more easily extracted from spores. This scenario would explain the higher levels of extractable CotB-1 and the presence of CotB-1 on ΔexsB spores lacking an exosporium. A similar scenario could explain the presence of BxpC on ΔexsB spores lacking an exosporium. In this case, ExsB is also required for proper insertion of BxpC into the exosporium, and in its absence BxpC can be aberrantly incorporated, at least partially, into the coat. Neither of the two scenarios described above excludes the possibility that CotB-1 and BxpC are normally incorporated in both the exosporium and the coat of wild-type spores. In this case, the lack of ExsB would distort the distribution of the proteins to favor association with the spore coat.

Finally, we used an ExsB-EGFP fusion protein to monitor the localization of ExsB during sporulation. Synthesis of the fusion protein occurred late in sporulation, corresponding to the time of exsB expression. The fusion protein was extensively phosphorylated, though perhaps slightly less so than wild-type ExsB. This slight difference could be due to the overexpression of ExsB-EGFP. The fusion protein appeared to be functional: a ΔexsB strain expressing the fusion protein formed a stable exosporium. Our results showed that ExsB-EGFP localized initially to the cap-end of the developing forespore and was then distributed over the entire surface of the forespore prior to lysis of the mother cell. This distribution pattern was essentially identical to that of BclA (Thompson and Stewart, 2008). Thus, ExsB appears to be evenly distributed within the exosporium basal layer of mature B. anthracis spores. We also examined the localization of ExsB-EGFP in a ΔbxpB strain unable to produce BxpB, the basal layer protein required for nearly all attachment of BclA to the basal layer. We observed that ExsB-EGFP was localized to the cap-end of the forespore but no further, even though an intact exosporium basal layer eventually encircled the forespore. This result indicated that BxpB is required, either directly or indirectly, for the normal distribution of ExsB within the basal layer. Recently, it was reported that the exosporium protein ExsK was also distributed over the entire surface of wild-type spores but only within the cap region of ΔbxpB spores, exactly as described for ExsB (Severson et al., 2009). It was proposed that ExsK incorporation into the exosporium was dependent on a direct association with BclA. Consistent with this model, the small amount of BclA attached to ΔbxpB spores preferentially associates with the cap region of the exosporium (Giorno et al., 2009). We have not been able to detect a stable interaction between ExsB and BclA, however (unpublished data). Therefore, we conclude that the mechanism of incorporation of ExsB is different from the mechanism proposed for incorporation of ExsK.

Experimental procedures

Strains and plasmid constructions

As the wild-type B. anthracis strain, we used the Sterne 34F2 veterinary vaccine strain obtained from the U. S. Army Medical Research Institute of Infectious Diseases, Fort Detrick, MD. The Sterne strain is not a human pathogen because it lacks the plasmid pXO2, which is necessary to produce the capsule of the vegetative cell. A ΔexsB mutant of the Sterne strain (CLT361) was constructed by allelic exchange essentially as previously described (Dong et al., 2008). The construction procedure precisely deleted the entire exsB open reading frame (ORF) and replaced it with a spectinomycin-resistance cassette, which was confirmed by DNA sequence analysis. The exsB gene of the Sterne strain is designated BAS1898 (Kanehisa et al., 2006). Construction of the ΔbxpB mutant of the Sterne strain (CLT307) used in this study was previously described (Steichen et al., 2005).

To construct a plasmid expressing an exsB::egfp gene fusion, a DNA fragment with SphI and PstI restriction sites flanking the exsB promoter region and entire exsB gene was PCR amplified. (The exsB promoter region was actually the entire 151-bp intergenic region upstream of the exsB ORF.) The PCR product was end-digested with SphI and PstI and inserted between the SphI and PstI cloning sites of the multicopy plasmid pCLT1525, which was previously described as plasmid BclA NTD-eGFP/pCLT1474 (Tan and Turnbough, 2010). This insertion created an in-frame fusion between the intact exsB and egfp ORFs, with the gene fusion preceded by the exsB promoter-leader region and followed by the intrinsic bclA transcription terminator. This recombinant plasmid was designated pCLT1544. As a control, we constructed a derivative of plasmid pCLT1544 in which the exsB ORF was deleted, resulting in a fusion between the exsB promoter-leader region and the egfp gene. This plasmid was designated pCLT1546. All plasmid constructions were confirmed by DNA sequencing. Plasmids pCLT1544 and pCLT1546 were transformed into the methylation-deficient E. coli strain GM2163 (dam-13::Tn9 dcm-6). Nonmethylated plasmid DNA was purified and introduced by electroporation into strains CLT361 (ΔexsB) and CLT307 (ΔbxpB).

Preparation of spores, culture samples, and exosporia

Spores were prepared by growing B. anthracis strains at 37°C in liquid (with shaking) or on solid (1.5% agar) Difco sporulation medium (Nicholson and Setlow, 1990) until sporulation was complete, typically 48 to 72 h. Spores were collected by centrifugation and washed extensively with cold (4°C) sterile deionized water. Unless indicated otherwise in the text, spores were further purified by centrifugation through a two-step gradient of 20% and 50% Renografin (Bracco Diagnostics) followed by additional washes with cold water. Spores were stored and quantitated as previously described (Nicholson and Setlow, 1990; Steichen et al., 2003). Vegetative and sporulating cells were collected by centrifugation from cultures grown as described above. Culture density was measured spectrophotometrically at 600 nm, and spore development was monitored by phase-contrast microscopy. Exosporia were purified from spores as previously described (Steichen et al., 2003).

Preparation of mouse mAbs

The production and characterization of anti-BclA (EF12), anti-BxpB (10-44-1), anti-BxpC (FH6-1), anti-ExsY/CotY (G9-3), and anti-Alr (AR-1) mAbs were previously described (Boydston et al., 2006; Tan and Turnbough, 2010). An anti-CotB-1 mAb (B1-47) was raised against a recombinant CotB-1 protein synthesized in E. coli essentially as previously described (Steichen et al., 2003). By immunoblotting recombinant CotB-1 and CotB-2 and total proteins from wild-type and ΔcotB-1 strains of B. anthracis, we demonstrated that mAb B1-47 reacted only with CotB-1 (data not shown). All mAbs were purified by affinity chromatography on protein G-Sepharose and were fluorescently labeled using the Alexa Fluor 488 protein labeling kit (Molecular Probes).

Gel electrophoresis and immunoblotting

Spores (2 × 10⁷) or exosporium samples were boiled for 8 min in sample buffer containing 125 mM Tris-HCl (pH 6.8), 4% SDS, 100 mM dithiothreitol, 0.024% bromophenol blue, and 10% (vol/vol) glycerol. Solubilized proteins were separated by SDS-polyacrylamide gradient gel electrophoresis in a NuPAGE 4–12% Bis-Tris gel (Invitrogen) and visualized by staining with Coomassie brilliant blue (Nicholson and Setlow, 1990). For immunoblotting, proteins were transferred from a polyacrylamide gel to a nitrocellulose membrane and detected as previously described (Steichen 2003).

Mass spectrometry

For protein analysis by mass spectrometry, a Coomassie stained protein band was sliced from a polyacrylamide gel and digested with trypsin and/or chymotrypsin as previously described (Kinter and Shermean, 2000). Proteolytic fragments were analyzed by LC-MS/MS with electrospray ionization using a Micromass Q-TOF2 mass spectrometer. Interpretation of spectra was performed manually with the aid of the ProteinLynx Global Server software.

Phase-contrast and fluorescence microscopy

Slides containing immobilized spores or sporulating cells were prepared essentially as previously described (Boydston et al., 2006; Steichen et al., 2007). In experiments designed to visualize exosporia, immobilized spore samples were treated with an Alexa 488-labeled anti-BclA mAb EF12 and washed as previously described (Swiecki et al., 2006). As a negative control, the spore samples were also treated with an equivalently labeled isotype control mAb that was unable to bind spores. In all cases, samples were examined by phase-contrast and fluorescence microscopy using a Nikon Eclipse E600 microscope equipped with a Y-FL epifluorescence attachment. Images were captured with a Spot charge-coupled device digital camera (Diagnostic Instruments, Inc.) and displayed by using Spot (v4.0) software.

Electron microscopy

Transmission electron microscopy of spores was performed as previously described (Boydston et al., 2005).

Isolation of cellular RNA and RT-PCR

Cellular RNA was extracted from sporulating cells with hot phenol, treated with RNase-free DNase, purified, analyzed for concentration and quality, and stored as described previously (Dong et al., 2008). The levels of exsB transcripts in isolated cellular RNA were measured by using the SuperScript III One-Step RT-PCR System (Invitrogen) essentially as previously described (Chesnokova et al., 2008). Each 25-μl reaction mixture contained 50 ng of cellular RNA and excess DNA primers with the sequence 5′-GCCATAGAAACGTGATATTAGAAAAGC-3′ and 5′-CGGATCCCTAGCAATTACGTTTTTTATACC-3′. The reaction amplified a 574-bp DNA fragment containing the entire 561-bp exsB ORF. PCR products were analyzed by electrophoresis in a 1.2% agarose gel and visualized by staining with ethidium bromide.