Identification of the UDP-N-Acetylglucosamine 4-Epimerase Involved in Exosporium Protein Glycosylation in Bacillus anthracis

Abstract

Spores of Bacillus anthracis, the causative agent of anthrax, are enclosed by a loosely fitting exosporium composed of a basal layer and an external hair-like nap. The filaments of the nap are formed by trimers of the collagen-like glycoprotein BclA. The side chains of BclA include multiple copies of two linear rhamnose-containing oligosaccharides, a trisaccharide and a pentasaccharide. The pentasaccharide terminates with the unusual deoxyamino sugar anthrose. Both oligosaccharide side chains are linked to the BclA protein backbone through an N-acetylgalactosamine (GalNAc) residue. To identify the gene encoding the epimerase required to produce GalNAc for BclA oligosaccharide biosynthesis, three annotated UDP-glucose 4-epimerase genes of B. anthracis were cloned and expressed in Escherichia coli. The candidate proteins were purified, and their enzymatic activities were assessed. Only two proteins, encoded by the BAS5114 and BAS5304 genes (B. anthracis Sterne designations), exhibited epimerase activity. Both proteins were able to convert UDP-glucose (Glc) to UDP-Gal, but only the BAS5304-encoded protein could convert UDP-GlcNAc to UDP-GalNAc, indicating that BAS5304 was the gene sought. Surprisingly, spores produced by a mutant strain lacking the BAS5304-encoded enzyme still contained normal levels of BclA-attached oligosaccharides. However, monosaccharide analysis of the oligosaccharides revealed that GlcNAc had replaced GalNAc. Thus, while GalNAc appears to be the preferred amino sugar for the linkage of oligosaccharides to the BclA protein backbone, in its absence, GlcNAc can serve as a substitute linker. Finally, we demonstrated that the expression of the BAS5304 gene occurred in a biphasic manner during both the early and late stages of sporulation.

Bacillus anthracis, the causative agent of anthrax, is a gram-positive, rod-shaped soil bacterium that forms spores when deprived of essential nutrients (15). Spore formation begins with an asymmetric septation that divides the developing cell into a smaller forespore compartment and a larger mother cell compartment, each containing a copy of the genome. The mother cell then engulfs the forespore and surrounds it with three protective layers: a cortex composed of peptidoglycan, a closely apposed proteinaceous coat, and a loosely fitting exosporium (11). Mother cell lysis releases the mature spore, which is dormant and capable of surviving in harsh environments for many years (16). When spores encounter an aqueous environment containing nutrients, they can germinate and grow as vegetative cells (23).

Recently, interest in B. anthracis spores has intensified in response to their use as agents of bioterrorism. Of particular interest has been the outermost exosporium layer, which serves as a semipermeable barrier excluding potentially harmful macromolecules (9, 26) and as a vital first point of contact with the immune system of an infected host (14, 18, 32). The exosporium of B. anthracis and closely related species such as Bacillus cereus and Bacillus thuringiensis is a prominent structure comprised of a paracrystalline basal layer and an external hair-like nap (2). The basal layer contains approximately 20 different proteins (22, 25), while the filaments of the nap are formed by trimers of a single collagen-like glycoprotein called BclA (4, 27). The central region of BclA contains a large number of GXX repeats, mostly GTP triplets, and this region varies in length in naturally occurring strains of B. anthracis, resulting in hair-like naps of differing lengths (24, 28). Multiple copies of two O-linked oligosaccharides, a trisaccharide and a pentasaccharide, are attached to the protein component of BclA. The pentasaccharide side chains appear to be attached to threonine residues within the central region, while the trisaccharide side chains are attached to presently undefined residues in the protein (7).

The precise structure of the trisaccharide side chain has not been determined, but its sequence is 3-O-methyl-l-rhamnose-l-rhamnose-N-acetylgalactosamine (GalNAc) (7). Except for a single glycosidic linkage, the structure of the pentasaccharide is known. Its structure is 2-O-methyl-4-(3-hydroxy-3-methylbutamido)-4,6-dideoxy-β-d-glucopyranosyl-(1→3)-α-l-rhamnopyranosyl-(1→3)-α-l-rhamnopyranosyl-(1→2)-l-rhamnopyranosyl-(1→?)-N-acetylgalactosamine (7). Both oligosaccharides are attached to the BclA protein backbone through GalNAc residues. The pentasaccharide sugar 2-O-methyl-4-(3-hydroxy-3-methylbutamido)-4,6-dideoxy-d-glucose, which was given the trivial name anthrose, has been found only in B. anthracis strains and a limited number of highly pathogenic strains of B. cereus and B. thuringiensis (7, 8). For that reason, anthrose has joined other exosporium components as targets for the detection of B. anthracis spores and as new targets for therapeutic intervention in anthrax (6, 26, 29).

In view of the potential importance of the BclA oligosaccharides, especially the anthrose-containing pentasaccharide, we have undertaken a comprehensive study of their biosynthesis. This effort involves identifying the biosynthetic genes for the three component sugars, anthrose, rhamnose, and GalNAc, as well as the genes involved in assembling the oligosaccharides and attaching them to the protein backbone of BclA. We recently reported the identification of a four-gene anthrose biosynthetic operon (8). A four-gene rhamnose biosynthetic operon has also been identified (24). This paper describes the identification of the gene encoding the UDP-N-acetylglucosamine (GlcNAc) 4-epimerase necessary for GalNAc biosynthesis. It also describes a surprising alternative BclA oligosaccharide biosynthetic pathway, which is active only in the absence of the UDP-GlcNAc 4-epimerase. Finally, this paper reports a biphasic pattern of expression of the gene encoding this epimerase during sporulation.

MATERIALS AND METHODS

Materials.

UDP-d-N-acetylglucosamine (UDP-GlcNAc), UDP-d-N-acetylgalactosamine (UDP-GalNAc), UDP-d-glucose (UDP-Glc), and UDP-d-galactose (UDP-Gal) were purchased from Calbiochem. Other sugars and cofactors were obtained from Sigma-Aldrich Corp. Restriction enzymes were purchased from Promega.

Bacterial strains and plasmids.

The Sterne 34F2 veterinary vaccine strain of B. anthracis, obtained from the U.S. Army Medical Research Institute of Infectious Diseases, Fort Detrick, MD, was used as the wild-type parental strain. The Sterne strain is not a human pathogen because it lacks plasmid pXO2, which is necessary to produce the capsule of the vegetative cell. A variant of the Sterne strain carrying a mutation (ΔBAS5304) that precisely deletes the entire BAS5304 open reading frame (ORF) was constructed by allelic exchange essentially as previously described (8). This procedure replaced the BAS5304 gene with a spectinomycin resistance cassette, which was confirmed by DNA sequence analysis of the relevant region of the chromosome. B. anthracis Sterne gene numbers were obtained from the Kyoto Encyclopedia of Genes and Genomes database (13). Escherichia coli strains used for gene cloning and expression were obtained commercially. E. coli strain GM2163 (dam-13::Tn9 dcm-6) and E. coli/B. anthracis shuttle vector pCLT1474, which were used for complementation analysis, were described previously (8). Expression vectors pET15b and pET21a were purchased from Novagen.

Cloning of putative epimerase genes.

Candidate UDP-GlcNAc 4-epimerase genes were amplified by PCR using Sterne strain chromosomal DNA as a template and primer pairs that directed copying from the second-to-the-last codon of each ORF. The PCR products were cloned into the pGEM-T vector (Promega), and the resulting plasmids were transformed into E. coli strain INVαF′ (Invitrogen). The recombinant plasmids were purified and used to generate NdeI/XhoI DNA restriction fragments containing the BAS1093 and BAS5114 ORFs and an NheI/XhoI fragment carrying the BAS5304 ORF. The BAS1093 fragment was inserted between the NdeI and XhoI sites of expression vector pET15b, which directs the addition of a His₆ tag to the amino terminus of the expressed protein. The BAS5114 and BAS5304 fragments were inserted between the NdeI and XhoI sites or the NheI and XhoI sites, respectively, of expression vector pET21a, which directs the addition of a His₆ tag to the carboxyl terminus of the expressed protein. The recombinant expression vectors were transformed into E. coli strain BL21(DE3) (Novagen). The correct construction of these expression vectors was confirmed by DNA sequence analysis.

Protein expression and purification.

Each E. coli BL21(DE3) transformant carrying a recombinant expression vector was grown overnight in 50 ml of LB medium containing 100 μg/ml ampicillin at 37°C with shaking at 250 rpm. The culture that was grown overnight was added to 1 liter of the same medium, and this culture was grown at 37°C with shaking until it reached an optical density at 600 nm of 0.6. Isopropyl-β-d-thiogalactopyranoside (Fisher Scientific) was added to the culture (0.2 mM final concentration) to induce epimerase gene expression, and the culture was incubated for an additional 16 h at room temperature. Cells were harvested by centrifugation and washed three times with cold (4°C) lysis buffer (50 mM NaH₂PO₄, 300 mM NaCl, 10 mM imidazole [pH 8.0]). Cell pellets were resuspended in 30 ml of cold lysis buffer, and cells were disrupted by sonication at 4°C. Cell debris was removed by centrifugation, and the resulting supernatant was passed through a 2-ml Ni-nitrilotriacetic acid agarose affinity column (Qiagen). Columns were washed and eluted according to the manufacturer's instructions in the case of the BAS5114 and BAS5304 extracts. However, to reduce precipitation of the recombinant protein, a modified elution buffer containing 50 mM NaH₂PO₄, 300 mM NaCl, 250 mM imidazole, and 1% glycerol at pH 7.5 was used for the BAS1093 extract. Imidazole was removed from samples by gel filtration using a Bio-Gel P4 column and a running buffer containing 20 mM Tris HCl-1% glycerol (pH 8.0). Fractions containing protein were pooled and concentrated by ultrafiltration in a 10,000-molecular-weight (MW)-cutoff Vivaspin 2 concentrator (Sartorius Stedium Biotech) to yield a final concentration of approximately 1 mg/ml. Protein concentrations were estimated by measuring the absorbance at 280 nm. Molar extinction coefficients for the recombinant proteins (at 280 nm) were calculated using the ProtParam tool of the ExPASy proteomics server of the Swiss Institute of Bioinformatics. The purity of recombinant proteins was assessed by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis using a 4-to-15% gradient polyacrylamide gel (Bio-Rad), with protein bands being visualized by staining with GelCode Blue (Thermo Scientific). Recombinant proteins were identified by immunoblotting with an anti-His₆ tag monoclonal antibody (Novagen), which was visualized using a Sigma Fast BCIP (5-bromo-4-chloro-3-indolylphosphate)/Nitro Blue Tetrazolium kit.

Epimerase assays.

Reaction mixtures consisted of 1 mM sugar nucleotide (UDP-Glc, UDP-Gal, UDP-GlcNAc, or UDP-GalNAc) and various amounts of freshly prepared enzyme in a solution containing 20 mM Tris-HCl, 4 mM Mg²⁺, and 1 mM NAD⁺ (pH 8.0) in a final volume of 50 μl. The reactions were performed at 37°C for 2 h and were terminated by heating at 100°C for 5 min. Control reaction mixtures contained enzyme that had been inactivated by heating at 100°C for 5 min. The reaction mixtures were passed through a 3,000-MW-cutoff spin filter (Millipore). High-performance liquid chromatography (HPLC) was performed to analyze the sugar nucleotides in the reaction mixtures using a modification of a procedure described previously (33). This analysis employed a Dionex CarboPac PA-1 column (4 by 250 mm) connected to a Dynamax pump system and a UV detector (Rainin Instruments). Samples were applied to the column, and the column was washed with 0.2 M ammonium acetate for 1 min. The absorbed sugar nucleotides were separated by elution with a series of linear ammonium acetate gradients; the salt concentration was increased from 0.20 M to 0.25 M for 3 min, from 0.25 M to 0.35 M for 36 min, and from 0.35 M to 1.0 M for 3 min. The flow rate was 0.6 ml/min, and the absorbance of the effluent at 262 nm was recorded.

Complementation analysis.

A 1,335-bp DNA fragment containing the BAS5304 gene plus 175 bp upstream and 146 bp downstream of the gene was PCR amplified from genomic DNA of B. anthracis. This fragment apparently includes the promoter and transcription terminator for a single-gene BAS5304 operon. The forward primer introduced a SalI site, and the reverse primer introduced a BamHI site. The PCR product was digested with SalI and BamHI and inserted between the SalI and BamHI sites of shuttle plasmid pCLT1474. The resulting plasmid was transformed into methylation-deficient E. coli strain GM2163. The resulting transformant was used to prepare unmethylated plasmid DNA, which was introduced by electroporation into the ΔBAS5304 mutant Sterne strain with selection for erythromycin resistance.

Preparation of spores and sporulating cells.

Spores were prepared by growing B. anthracis strains at 37°C in liquid Difco sporulation medium with shaking until sporulation was complete, typically 48 to 72 h (17). Spores were purified by extensive washing and sedimentation through a two-step gradient of 20% and 50% Renografin (Bracco Diagnostics), stored at 4°C in water, and quantitated microscopically as previously described (24). Sporulating cells were obtained from cultures grown as described above and were harvested by centrifugation at 4°C. Culture density was measured spectrophotometrically at 600 nm, and spore development was monitored by phase-contrast microscopy.

Isolation of cellular RNA and RT-PCR.

RNA was extracted from cells of B. anthracis with hot phenol, treated with RNase-free DNase, purified, analyzed for concentration and quality, and stored as previously described (8). The levels of BAS5304 transcripts in cellular RNA samples were measured by using the SuperScript III One-Step reverse transcription (RT)-PCR system (Invitrogen) essentially as previously described (5). Each 25-μl reaction mixture contained 50 ng of cellular RNA and excess DNA primers with the sequences 5′-GGAGCAGGGTATATTGGTAGTCATAC and 5′-CTAACCCATATTCCGCTTCCCATCC. This reaction amplified a 931-bp region within the 1,014-bp BAS5304 ORF. PCR products were analyzed by electrophoresis in a 1% agarose gel with Tris-acetate-EDTA buffer and visualized by staining with ethidium bromide.

Preparation of oligosaccharides.

Spores (10⁹) were suspended in 40 μl of sample buffer (60 mM Tris-HCl [pH 6.8], 1.7% SDS, 100 mM dithiothreitol, 0.012% bromophenol blue, 10% [vol/vol] glycerol) and placed into a boiling water bath for 10 min. Insoluble material was removed by brief centrifugation, and soluble proteins were separated by SDS-polyacrylamide gel electrophoresis using a 4-to-15% gradient polyacrylamide gel. Protein bands in the gel were visualized by staining with GelCode Blue, and the gel was electroblotted onto a polyvinylidene difluoride (PVDF) membrane. A diffuse band previously shown to include BclA in high-molecular-mass exosporium protein complexes was excised (25). This membrane strip was washed with water, destained with methanol, placed into a 2-ml glass ampule, and dried under a vacuum for 1 h. The ampule was flushed with argon, 0.5 ml anhydrous hydrazine was added, and the ampule was sealed. Hydrazinolysis was carried out at 60°C for 3 h to release the O-linked oligosaccharide side chains of BclA. After cooling, the hydrazine was rapidly removed by vacuum centrifugation. Immediately before mass spectrometric analysis, the sample was dissolved in 100 μl of water and passed through a 3,000-MW-cutoff filter (Millipore). The sample was analyzed by electrospray mass spectrometry using a Q-TOF2 mass spectrometer (Micromass Ltd., Manchester, United Kingdom). On-line injections were made into a stream of 1:1 acetonitrile-water (vol/vol) containing 0.1% formic acid at a flow rate of 1 μl/min. Positive-ion mass spectra were obtained using a needle voltage of 2.8 kV. The mass spectral data were processed using the Max-Ent3 module of MassLynx 3.5.

Monosaccharide analysis.

Excised PVDF membrane strips with bound BclA were prepared as described above. However, instead of carrying out hydrazinolysis, the BclA oligosaccharides were depolymerized by methanolysis. Following the re-N-acetylation of amino sugars, the methyl glycosides were trimethylsilylated and separated by gas chromatography as previously described (8). In some cases it was desirable to visualize only the amino sugars in chromatograms. This was accomplished by using gas chromatography/mass spectrometry in the positive-ion mode with chemical ionization; acetonitrile was used as the reagent gas. Selective ion monitoring of the m/z 524 ion (M + 1) was specific for hexosamines. Analyses were carried out by use of a Varian 4000 gas-liquid chromatograph/mass spectrometer (Varian, Inc., CA) fitted with a 60-m VF-1 capillary column.

RESULTS

Identification of the UDP-GlcNAc 4-epimerase of B. anthracis.

Both types of oligosaccharide side chains of BclA contain a single GalNAc residue that serves as the site of attachment to the BclA protein backbone (7). In bacteria, GalNAc is made by the conversion of UDP-N-acetylglucosamine to UDP-N-acetylgalactosamine through a reaction catalyzed by a specific UDP-GlcNAc 4-epimerase (3). Such an enzyme has not been identified for B. anthracis. However, the B. anthracis (Sterne) genome contains three genes, designated BAS1093, BAS5114, and BAS5304, that are annotated as encoding possible UDP-Glc 4-epimerases with molecular masses of 32.9 kDa, 36.9 kDa, and 37.8 kDa, respectively. We suspected that one of these genes might actually encode UDP-GlcNAc 4-epimerase. To examine this possibility, we cloned and expressed each gene in E. coli strain BL21(DE3). We initially inserted the genes into the pET21a expression vector, which directs the addition of a His₆ tag to the carboxyl terminus of the expressed protein. To a large extent, the recombinant proteins produced with the BAS5114 and BAS5304 genes were soluble, and they could be highly purified by Ni-agarose affinity chromatography. In contrast, the recombinant protein produced with the BAS1093 gene quickly precipitated after affinity purification. In an attempt to avoid this problem, we cloned and expressed the BAS1093 gene using the expression vector pET15b, which directs the addition of an amino-terminal His₆ tag. This recombinant protein was soluble before and after affinity purification. The three purified and soluble recombinant proteins were assayed for epimerase activities.

Reaction mixtures containing one of the three purified proteins, the essential epimerase cofactor NAD⁺ (1), and a sugar nucleotide substrate (either UDP-Glc, UDP-Gal, UDP-GlcNAc, or UDP-GalNAc) were prepared. After incubation, the sugar nucleotides in the reaction mixtures were analyzed by HPLC. The results showed that the proteins encoded by the BAS5114 and BAS5304 genes, but not the BAS1093-encoded protein, catalyzed both the partial conversion of UDP-Glc to UDP-Gal (Fig. 1) and the reverse reaction, the partial conversion of UDP-Gal to UDP-Glc (data not shown). These reactions ultimately yielded the same equilibrium mixture of sugar nucleotides. More importantly, the results also showed that only the BAS5304 enzyme was capable of partially converting UDP-GlcNAc to UDP-GalNAc (data not shown) and UDP-GalNAc to UDP-GlcNAc (Fig. 2). Taken together, these results indicate that the BAS5304 gene encodes a bifunctional enzyme with both UDP-GlcNAc 4-epimerase and UDP-Glc 4-epimerase activities and that no other annotated UDP-Glc 4-epimerase gene of B. anthracis encodes a UDP-GlcNAc 4-epimerase.

FIG. 1.

Assays for UDP-glucose 4-epimerase activity. Purified recombinant proteins encoded by the BAS1093 (C), BAS5114 (D), and BAS5304 (E) genes were incubated with UDP-Glc as the substrate, and the reaction products were separated by HPLC. Retention times of UDP-Gal (A) and UDP-Glc (B) were established with sugar standards.

FIG. 2.

Assays for UDP-N-acetylglucosamine 4-epimerase activity. Purified recombinant proteins encoded by the BAS1093 (C), BAS5114 (D), and BAS5304 (E) genes were incubated with UDP-GalNAc as the substrate, and the reaction products were separated by HPLC. Retention times of UDP-GalNAc (A) and UDP-GlcNAc (B) were established with sugar standards.

Although the preparations of recombinant proteins examined as described above were highly purified, each contained low levels of contaminating E. coli proteins. To exclude the possibility that these proteins contributed to the UDP-GlcNAc 4-epimerase activity attributed to the BAS5304-encoded protein, we grew cultures of E. coli strain BL21(DE3) and the transformant of this strain used to express the recombinant BAS5304 gene under inducing conditions. These cultures were used to prepare equivalent crude cell lysates, which were assayed for UDP-GlcNAc 4-epimerase activity. This activity was readily detected in the lysate containing the BAS5304-encoded protein, but it was undetectable in the BL21(DE3) lysate (data not shown). We also confirmed that the BL21(DE3) lysate did not exhibit UDP-Glc 4-epimerase activity. This result was expected because the BL21(DE3) strain possesses a defective galE gene, which encodes the UDP-Glc 4-epimerase of E. coli. Finally, we confirmed that the proteins encoded by the BAS5304 and BAS5114 genes required NAD⁺ for their epimerase activities (data not shown).

Effects of deleting the BAS5304 gene on glycosylation of BclA.

We predicted that if the BAS5304 gene did in fact encode the only UDP-GlcNAc 4-epimerase of B. anthracis, spores produced by a mutant strain lacking this gene would contain BclA devoid of oligosaccharide side chains. To test this prediction, we constructed a strain (ΔBAS5304) in which the BAS5304 gene was deleted. Spores from this strain and, as a control, spores from the wild-type strain were used to isolate BclA-containing complexes. These complexes were virtually free of carbohydrate-containing material other than BclA (7). The complexes were subjected to hydrazinolysis to release BclA oligosaccharide side chains, and the resulting hydrazides were analyzed by electrospray mass spectrometry. As expected, the sample from wild-type spores yielded three major peaks corresponding to the pentasaccharide (m/z 933), a tetrasaccharide (m/z 730) generated by the partial “peeling” of the GalNAc residue from the pentasaccharide during hydrazinolysis, and the trisaccharide (m/z 542) (7) (Fig. 3A and 4). Surprisingly, the ΔBAS5304 sample contained the same three major peaks, indicating that the mutant strain still produced pentasaccharide and trisaccharide side chains that were identical, or at least very similar, to the oligosaccharides of the wild-type strain (Fig. 3B). The major difference between the wild-type and ΔBAS5304 spectra was that the intensity of the putative pentasaccharide peak was lower in the case of ΔBAS5304 spores, suggesting an increased level of peeling. This apparent difference in peeling disappeared when the ΔBAS5304 mutation was complemented with a plasmid-borne copy of the wild-type BAS5304 operon (Fig. 3C). This observation implied that the putative hexosamine residue of the ΔBAS5304 pentasaccharide was not GalNAc but a similar, more easily peeled sugar.

FIG. 3.

BclA oligosaccharides of wild-type and ΔBAS5304 spores. BclA was purified from spores, and the BclA oligosaccharides were detached by hydrazinolysis and analyzed by positive-ion electrospray ionization-mass spectrometry. The labeled peaks correspond to the trisaccharide (542 Da), the tetrasaccharide generated from the pentasaccharide by the loss of the GalNAc residue (730 Da), and the pentasaccharide (933 Da). The analyses included wild-type spores (A), ΔBAS5304 spores (B), and spores produced by a ΔBAS5304 strain complemented with a plasmid-borne copy of the wild-type BAS5304 operon (C).

FIG. 4.

Structures of the oligosaccharide hydrazides obtained from BclA isolated from wild-type spores.

To examine the hexosamine content of BclA oligosaccharides from wild-type and ΔBAS5304 spores, samples of BclA-containing complexes from the two spore preparations were subjected to methanolysis to depolymerize attached oligosaccharides. After re-N-acetylation and trimethylsilylation of the samples, they were analyzed by gas chromatography/mass spectrometry using chemical ionization. Selected ion monitoring for fragments of m/z 524 detected amino sugars specifically (Fig. 5). As expected, the dominant hexosamine peak in the wild-type sample was GalNAc (Fig. 5C). On the other hand, no GalNAc was detected in the ΔBAS5304 sample; instead, the hexosamine detected was GlcNAc (Fig. 5D). When the ΔBAS5304 mutation was complemented, the GlcNAc peak disappeared and was replaced by GalNAc (Fig. 5E). These results indicated that the UDP-GlcNAc 4-epimerase encoded by the BAS5304 gene was indeed required for the synthesis of GalNAc in sporulating cells and that GlcNAc could replace GalNAc in the synthesis of BclA oligosaccharides.

FIG. 5.

Determination of GalNAc and GlcNAc attached to BclA of wild-type and ΔBAS5304 spores. Isolated BclA was subjected to methanolysis to depolymerize oligosaccharides, and the resulting methyl glycosides were analyzed by chemical ionization gas-liquid chromatography/mass spectrometry with selected ion monitoring specific for GlcNAc and GalNAc. Samples analyzed were a GlcNAc standard (A), a GalNAc standard (B), wild-type spores (C), ΔBAS5304 spores (D), and spores produced by a ΔBAS5304 strain complemented with the wild-type BAS5304 operon (E).

Expression of the BAS5304 operon during the cell cycle of B. anthracis.

The genes encoding proteins involved in exosporium assembly are expressed only during sporulation. The transcription of these genes is initiated at promoters recognized by one of two mother cell-specific sigma factors: σ^E is active during the early stages of sporulation, while σ^K is active during the late stages of sporulation (11). To determine the time of expression of the BAS5304 gene, which appears to be included in a single-gene operon (13, 19), we used quantitative RT-PCR to measure the levels of BAS5304 transcripts in cells harvested during vegetative growth and throughout sporulation (Fig. 6). Sampling times are indicated by T_n, where n is the number of hours relative to the start of sporulation (designated T₀). The results showed that significant levels of BAS5304 transcripts were detected only during sporulation, an observation consistent with a role for this gene in exosporium assembly. However, BAS5304 expression during sporulation appeared to be biphasic, with high levels of transcripts being detected between T₂ and T₃ and also between T₆ and T₉. At T₉, mother cells began to lyse to release mature spores. These results indicated that the transcription of the BAS5304 gene was initiated at both σ^E-specific and σ^K-specific promoters.

FIG. 6.

Cellular levels of BAS5304 transcripts during growth and sporulation of B. anthracis. T_n indicates the sampling time, with the numbers above the lanes indicating hours relative to the start of sporulation. Size standards are included in the lane marked Std. Levels of BAS5304 transcripts were measured by RT-PCR, with PCR products being visualized by gel electrophoresis and staining.

Recently, the entire transcriptome of B. anthracis was sequenced, providing putative transcription start sites for every operon (19). This analysis indicated that there were four transcription start sites for the BAS5304 operon. The start site closest to the BAS5304 ORF was located 9 bases downstream of a sequence that matched perfectly the consensus sequence for a promoter recognized by σ^E (Fig. 7). The other three putative start sites were not preceded by sequences that closely resembled either of the consensus sequences for σ^E and σ^K promoters. Furthermore, we were unable to find sequences in the intergenic region upstream of the BAS5304 ORF that closely resembled the consensus sequence for a σ^K promoter.

FIG. 7.

Sequence of the BAS5304 promoter region and putative transcription start sites identified by transcriptome analysis. The transcription start site closest to the BAS5304 ORF is labeled +1; the −10 (CATA——T) and −35 (ATA) regions of the apparent σ^E-specific promoter employing this start site are labeled and underlined. Asterisks indicate three other putative start sites. The BAS5305 Shine-Dalgarno (SD) sequence is underlined and italicized. The sequence ends with the ATG initiation codon for the BAS5304 gene.

DISCUSSION

An important part of our plan to completely elucidate the biosynthesis of the BclA oligosaccharide side chains of B. anthracis was to identify the UDP-GlcNAc 4-epimerase responsible for GalNAc biosynthesis. Both the pentasaccharide and trisaccharide side chains contain a single GalNAc residue that serves as the linker to the protein backbone of BclA. In this study, we showed that the UDP-GlcNAc 4-epimerase of B. anthracis (Sterne) is encoded by the BAS5304 gene. Our results demonstrated that the BAS5304-encoded enzyme can also function as a UDP-Glc 4-epimerase, as annotated in the genome, and therefore is a bifunctional UDP-GlcNAc/Glc 4-epimerase. Such bifunctional epimerases have been described for a variety of bacteria and other organisms (3, 10, 21). UDP-hexose 4-epimerases are divided into three main groups: group 1 epimerases catalyze only the epimerization of UDP-Glc, group 2 epimerases are capable of epimerizing either UDP-Glc or UDP-GlcNAc, and group 3 epimerases most efficiently epimerize UDP-GlcNAc. The BAS5304-encoded protein is a group 2 epimerase. A BLAST search of the NCBI database showed that the genomes of all members of the B. cereus group, which includes B. anthracis, B. cereus, and B. thuringiensis, encode a protein that is either identical or nearly identical to the BAS5304-encoded epimerase. Although most of these proteins were annotated as UDP-Glc 4-epimerases, they would be expected to possess UDP-GlcNAc 4-epimerase activity as well.

Two additional proteins annotated as UDP-Glc 4-epimerases, encoded by the BAS5114 and BAS1093 genes, were characterized in this study. Our results indicate that the BAS5114 protein is indeed a UDP-Glc 4-epimerase. This enzyme lacks UDP-GlcNAc 4-epimerase activity and therefore is a group 1 epimerase. We were unable to show any epimerase activity with the BAS1093-encoded protein. This protein is considerably smaller than the epimerases encoded by the BAS5304 and BAS5114 genes, which might be related to substantially different enzymatic activities.

The BAS5304-encoded UDP-GlcNAc/Glc 4-epimerase possesses motifs characteristic of members of the short-chain dehydrogenase/reductase (SDR) superfamily (12, 31). This superfamily contains two highly conserved signature sequence motifs; one, GX₂GX₂G (amino acids 7 to 13), is close to the cofactor binding pocket for NAD(H) in several members of the superfamily (12). The other signature sequence, YX₃K (amino acids 148 to 152), was previously reported to play an important role in catalysis in related enzymes (31). Both motifs are also present in the BAS5114-encoded UDP-Glc 4-epimerase. An imperfect GX₂GX₂G motif is present in the protein encoded by the BAS1093 gene, and while a YX₃K motif is present, it does not align with the motif in other sugar epimerases. It was previously reported that the replacement of Tyr299 of the E. coli UDP-Glc 4-epimerase with a cysteine has the effect of conferring UDP-GlcNAc/UDP-GalNAc-converting activity on the enzyme (30). X-ray crystallography revealed that the Y299C mutant possessed a large active-site volume that could accommodate an acetamido sugar. Consistent with this determination, the equivalent residue in the BAS5304-encoded UDP-GlcNAc/Glc 4-epimerase is a cysteine (Cys298).

The most surprising result of this study was that a B. anthracis mutant lacking the BAS5304 gene still produced spores containing glycosylated BclA and that the oligosaccharide side chains were pentasaccharide and trisaccharide molecules identical in mass to the side chains produced by the wild-type B. anthracis strain. This unexpected result was explained by the discovery that GlcNAc replaced GalNAc in the oligosaccharides of the mutant strain. This observation indicated that the BAS5304-encoded epimerase was indeed required to synthesize GalNAc but that GalNAc was not required for BclA oligosaccharide synthesis. Evidently, the glycosyltransferase involved in linking GalNAc to the BclA protein backbone and the rhamnosyltransferase that attaches a rhamnose residue to GalNAc during oligosaccharide side-chain synthesis are somewhat indiscriminate when choosing a substrate, preferring GalNAc but being able to use GlcNAc. Alternatively, these transferases might be replaced in the oligosaccharide biosynthetic pathways by other enzymes with different substrate specificities. Further characterization of the oligosaccharide biosynthetic pathways will be required to clarify these options.

The protein component of BclA is synthesized during the late stages of sporulation, and the promoter of the bclA operon apparently is recognized by σ^K (8, 24). The anthrose biosynthetic operon is expressed during the early stages of sporulation, and its promoter is recognized by σ^E (8). Thus, it seemed reasonable to expect that the BAS5304 gene would also be expressed during the early or late stages of sporulation. Our results showed that BAS5304 gene expression is biphasic, occurring during both the early and late stages of sporulation. Presumably, this pattern of expression requires both σ^E-specific and σ^K-specific promoters. We were able to identify the σ^E-specific promoter, but the location of the σ^K-specific promoter remains to be established. Interestingly, we recently showed that the expression of the B. anthracis alr gene, which encodes an exosporium-specific alanine racemase, occurs during sporulation in a biphasic manner closely resembling the pattern of BAS5304 expression. In the case of the alr operon, both σ^E-specific and σ^K-specific promoters were identified (5).

Finally, the microhydrazinolysis procedure developed in this study for releasing oligosaccharide side chains from BclA proved to have several advantageous over conventional hydrazinolysis methods. First of all, it did not require the purification of BclA, which is difficult because BclA is present as an insoluble complex with other proteins. Second, in the new procedure, BclA is tightly bound to a PVDF membrane, which greatly facilitates the efficient removal of water and salts by simple washing procedures. When not removed, these contaminants promote the undesirable peeling reaction that removes the terminal GalNAc residues of the pentasaccharide and trisaccharide side chains (20). In the past, upon carrying out hydrazinolysis on whole spores and purified exosporium preparations, which are contaminated with both water and salts, we observed that the peeling of the GalNAc residues has been almost complete (7). Lastly, the released oligosaccharides in the new procedure are hydrazides, and although it is a common practice to hydrolyze them to generate a reducing terminus, we found this step to be unnecessary when samples were immediately analyzed by electrospray ionization-mass spectrometry. Furthermore, the presence of the hydrazide group likely increased the efficiency of ionization.