Identification and characterization of a germination operon on the virulence plasmid pXOl of Bacillus anthracis

Sequence analysis of gerXB, gerXA and gerXC

The nucleotide sequence of the pag-atxA region of B. anthracis pXOl was obtained by oligonucleotide walking and analysed using the GCG software package. Three open reading frames (ORFs) were identified (GenBank accession no. AF108144), on the same strand, 7975 bp from the atxA transcription on the opposite strand (EMBL accession no. l13841). These ORFs consisted of 359, 492 and 318 codons, respectively, and had a high AT content (CG% 33.8). Each ORF started with an ATG initiation codon, and together they spanned 3608 bases of the pag-atxA region.

A tfasta search of the EMBL database (Pearson and Lipman, 1988) using the deduced amino acid sequences for the three ORF-encoded proteins showed significant similarity to the products of the gerB (Corfe et al., 1994; EMBL accession no. L16960) and gerK (Irie et al., 1996; EMBL accession no. D78187) operons of B. subtilis, the gerA operon (K. Tani, unpublished; EMBL accession no. U61380) of B. megaterium and the gerI operon (Clements and Moir, 1998; EMBL accession no. af067645) of B. cereus. The three genes were therefore designated gerXB, gerXA and gerXC (for germination pXOl related), and the encoded putative proteins with calculated molecular sizes of 40.839, 55.106 and 37.084 kDa were designated GerXB, GerXA and GerXC respectively. The greatest similarity was observed with the gerA operon products of B. megaterium using the algorithm of Needleman and Wunsch (1970): between GerXB and GerAB (20.0% identity and 36.8% similarity); between GerXA and GerAA (30.1% entity and 44.1% similarity); between GerXC and GerAC (20.9% identity and 30.4% similarity). Kyte–Doolittle hydrophobicity plots (Kyte and Doolittle, 1982) demonstrated that GerXB, GerXA and GerXC had patterns of hydropathy (Fig. 1) typical of germination proteins of the gerA operon of B. subtilis. Therefore, GerXB, GerXA and GerXC are homologues of GerAB, GerAA and GerAC respectively (Feavers et al., 1985; Irie et al., 1993; 1996). Hydrophobic cluster analysis showed that GerXB (X₃₅-Ycter) (Fig. 1A) and GerXA (V₂₅₀-F₄₆₆) (Fig. 1B) contained 10 and seven membrane-spanning sequences, respectively, which predicted transmembrane domains (Kyte and Doolittle, 1982). Further analysis with plotsimilarity (data not shown) showed that the most conserved region between GerXA and GerAA of B. megaterium corresponds to (V₂₅₀–F₄₆₆; 37.3% identity and 50% similarity). The deduced sequence of GerXC predicted a hydrophilic protein (Fig. 1C) with a putative lipoprotein signal (L₁₂LMFLFLLTGC₂₂; Hayashi and Wu, 1990) as observed in the GerAC homologues (Irie et al., 1993).

Regulation of gerX expression

The ATG initiation codons of the three gerX genes were preceded by potential ribosome binding sites with homology to the 3′ end of the B. anthracis 16S ribosomal RNA molecule (Ash et al., 1991), with ‘spacers’ of 3, 10 and 7 nucleotides, respectively, the optimum ‘spacer’ in Bacillus being 7 nucleotides (Vellanoweth and Rabinowitz, 1992). The initiation codon of gerXB was preceded by one potential sigma-G (σ^G)-dependent promoter sequence 5′-TGAATA-n₁₈-TATAATA-3′ (Nicholson et al., 1989) (Fig. 2). GerXC overlapped with gerXA by 14 bp, and neither promoter-like nor terminator-like sequences were found around gerXA and gerXC. Therefore, these three genes may be co-transcribed and expressed during sporulation. We determined the size of the gerXB, gerXA and gerXC transcripts by carrying out Northern blot analysis with total RNA isolated from B. anthracis 7702 during the exponential phase and at t₃ (t_n is the number of hours after the end of the exponential phase) (Fig. 3). No mRNA from these genes was detected in total mRNA extracted during the exponential phase (Fig. 3, lanes Al and Bl). In contrast, an mRNA species of ≈ 3.5 kb in size was detected at t₃, with both specific gerXB (Fig. 3, lane A2) and gerXC probes (Fig. 3, lane B2). The estimated size, 3.5 kb, was close to the predicted minimum size (3569 bases) for a polycistronic mRNA covering the three cistrons gerXB, gerXA and gerXC.

The 5′ end of the transcript was then determined by primer extension analysis (Fig. 4) using a synthetic end-labelled primer extending from nucleotide position 53 to 75 bp downstream from the translational start site of gerXB. Total RNA was isolated at the end of the exponential phase and into the stationary phase. At t₃, two adjacent bases were identified as start sites C and A, which are 31 bp and 30 bp upstream from the putative translational start site of the gerXB cistron respectively.

We analysed the expression of the gerX locus in space and time by constructing a gerXB–lacZ transcriptional fusion, which was introduced into pXOl at nucleotide 572 downstream from the putative translational start site of gerXB (Fig. 2). Direct evidence that gerX is expressed exclusively in the forespore compartment was obtained by growing the recombinant cells (GX12) carrying the gerX–lacZ fusion in the presence of the fluorescent lipophilic β-d-galactosidase substrate, C12FDG (Zhang et al., 1991). No fluorescence was detected in the mother cell compartment; the fluorescence was limited exclusively to the forespore compartment (Fig. 5A), thus demonstrating that gerX was expressed in this compartment after septation. The expression of gerX was also studied in a time course experiment. β-d-Galactosidase activity was barely detectable during the exponential phase (Fig. 5B), but it increased rapidly during the stationary phase, reaching a peak 3 h after the end of the exponential phase (2014 U mg⁻¹ protein).

Thus, gerXB, gerXA and gerXC are organized into a single operon (gerX ) expressed during the sporulation phase.

Germination of the gerX null mutant within macrophage

We showed recently that B. anthracis germinates within professional phagocytes (Guidi-Rontani et al., 1999). We investigated the possible involvement of gerX in this process by producing a gerX null mutant, GX11 (Fig. 2). The fate of the 7702 and GX11 spores within murine macrophages such as RAW264.7 was investigated. The germination of spores associated with macrophages was assessed by testing for a loss of heat resistance (30 min at 65°C) and by indirect immunofluorescence 3 h after infection. In heat resistance tests, we observed that the macrophage-7702 fraction had significantly higher levels of germination (26%) than the macrophage-GX11 fraction (8%). Immunofluorescence analysis showed that antibodies against bacilli recognized a large proportion of germinated spores in the 7702-macrophage population (Fig. 6A and B). In contrast, very few germinated spores were detected in the macrophage-GX11 fraction (Fig. 6C and D).

In vivo contribution of gerX in B. anthracis virulence

We investigated the involvement in vivo of the factors encoded by the gerX operon in anthrax pathogenesis by studying the virulence properties of GX11. The cumulative mortality of mice was assessed after subcutaneous inoculation with 5 × 10⁴ spores of 7702 or GX11, a dose equivalent to one LD₅₀ for the parental strain 7702 (Fig. 7). Whereas 50% of mice died after inoculation with 7702, only 10% of mice died after GX11 inoculation. We followed the germination of GX11 and 7702 spores over time within the host after the injection of 1.7 × 10⁵ spores. Tissue was collected from the inoculation site. Large subcutaneous oedema developed in tissues adjacent to the site of inoculation after 12 h with 7702 and after 24 h with GX11. This suggested that oedema toxin is synthesized soon after the germination of 7702 and GX11 spores. The germination of spores was assessed by testing for a loss of heat resistance (Fig. 8). A decrease in the number of 7702 dormant spores was observed 12 h after inoculation. This decrease became large over time until only 2.5 × 10³ cfu remained. Bacteria then grew rapidly, and large numbers were present 18 h and 24 h after inoculation. In contrast, there was no significant reduction in the number of dormant GX11 spores until 24 h after inoculation (2.5 × 10⁴ cfu). However, at 24 h and in the absence of heat treatment, there was a significant increase in cfu for GX11. Thus, the multiplication of some germinated forms had already been initiated. Consequently, GX11 germination was less efficient than that of 7702, and this was associated with a significant delay in the onset of germination.

Bacterial strains, plasmids and culture media

B. anthracis Sterne strain 7702 (pXOl+) (Pasteur Collection) was used. Escherichia coli harbouring the Gram-positive suicide vector, pAT113 (Trieu-Cuot et al., 1991), was used for transferring DNA into B. anthracis. E. coli was grown in L broth (Miller, 1972), and B. anthracis was grown in brain–heart infusion (BHI; Difco Laboratories), HCT medium (Lecadet et al., 1980) or NBY medium for the preparation of spores (Green et al., 1985). Antibiotic concentrations in the media were as follows: spectinomycin, 100 μg ml⁻¹ for B. anthracis; kanamycin, 50 μg ml⁻¹ and 20 μg ml⁻¹ for E. coli and B. anthracis respectively; erythromycin, 180 μg ml⁻¹ and 100 μg ml⁻¹ for E. coli and B. anthracis respectively. Stocks of B. anthracis spores were prepared as described previously (Pezard et al., 1991).

Antiserum preparation

Antiserum was raised against vegetative forms as described previously (Guidi-Rontani et al., 1999).

DNA isolation, sequencing and plasmid construction

pXOl was isolated from B. anthracis strain 7702 as described by Green et al. (1985). Methods for recombinant DNA manipulation were as described by Maniatis et al. (1982). DNA was sequenced by Genome Express, and the sequence strategy applied to both strands was ‘oligonucleotide walking’.

Construction of gerX null strains and mating procedure

Strains GX11 and GX12 were obtained by deletion of the fragment between nucleotides 572 of gerXB and 186 of gerXC (Fig. 2) and insertion of a spectinomycin resistance cassette, spc (Etienne-Toumelin et al., 1995), or of a lacZ-spectinomycin cassette, as described previously by Sirard et al. (1995). The constructs were integrated into the gerX locus of pXO1 by homologous recombination using the suicide plasmid, pAT113 (Trieu-Cuot et al., 1991). Conjugative transfer from E. coli to B. anthracis 7702 was as described by Pezard et al. (1991) and Sirard et al. (1995).

β-d-Galactosidase assays

β-d-Galactosidase (EC 3.2.1.23) was assayed as described by Pardee et al. (1959). Briefly, B. anthracis GX12 was cultured in HCT medium (Lecadet et al., 1980) at 34°C with shaking. Bacterial suspensions were then treated with SDS (0.2%) and chloroform (2%) at 28°C. Specific activities are expressed in Miller units (U) mg⁻¹ protein. The enzyme was detected within GX12 bacilli using the fluorescent lipophilic, β-d-galactosidase substrate, 5-dodecanoylaminofluorescein di-β-d-galactopyranoside (C12FDG; 330 μM; Molecular Probes) in BHI medium containing 0.4% agar.

RNA extraction, Northern analysis and primer extension

RNA extraction and primer extension were performed as described previously (De-Souza et al., 1993). Electrophoresis of 20 μg of formaldehyde-denatured RNA and transfer to nitrocellulose membranes (Hybond-C; Amersham) were performed as described elsewhere (Sambrook et al., 1989). The specific gerXB and gerXC probes were a 544 kb internal fragment (+ 28 to + 572) and a 545 kb internal fragment (+186 to + 731) of the respective genes. Probes were labelled with [³²P]-dATP (nick translation, Amersham). Hybridizations and washes were performed at 65°C in standard solutions (Amersham). For primer extension experiments, a synthetic 21-mer oligonucleotide was used, 5′-CGACCAGAATGATCTAATAGC-3′, which was complementary to the 5′ end of the gerXB gene (+ 101 to + 121). It was 5′ end-labelled with [³²P]-ATP (4500 Ci mmol⁻¹) using T4 polynucleotide kinase.

Cell culture and macrophage infection

The murine macrophage-like cell line RAW264.7 was grown routinely as described previously (Guidi-Rontani et al., 1999). RAW264.7 cells in PBS were infected at a multiplicity of infection (MOI) of 10 spores per macrophage. After 30 min of incubation, the cells were washed thoroughly and incubated with RPMI-1640 medium containing gentamicin (2.5 μg ml^−l). The germination of spores within cells was assessed by testing for a loss of heat resistance (30 min at 65°C). Colony counts (cfu) without heat treatment correspond to total spore population (dormant and germinated spores). Viability after heat treatment was taken as a measure of the population of dormant spores. All experiments were carried out at least twice, on at least two independent spore preparations. One representative experiment is shown.

Immunofluorescence microscopy

Germinated spores were detected by indirect immunofluorescence with rabbit polyclonal antiserum specific for germinated spore (Guidi-Rontani et al., 1999) (1:1000) and rhodamine-conjugated goat anti-rabbit IgG (H + l; 1:50; Kirkegaard and Perry Laboratories). F-actin was detected by incubating fixed cells suspended in PBS with Oregon green 488–phalloidin (1:50; Molecular Probes). Labelling was carried out as described by Guidi-Rontani et al. (1999).

Infection of mice

Seven-week-old female Swiss mice were supplied by IFFA-CREDO. Mortality was monitored by subcutaneously inoculating groups of 10 mice with 5 × 10⁴ spores from 7702 or GX11. Bacterial germination in host tissues was monitored by infecting mice subcutaneously above the groin with 1.7 × 10⁵ spores in 0.1 ml of PBS. Groups of three infected mice were killed at intervals. The entire site of inoculation, including subcutaneous oedema, was removed, homogenized in PBS (pH 7.2) and tested for the presence of spores or bacilli as described above.

Data handling and computer analysis

DNA sequences were compiled and analysed using the programs frames, translate, fasta, motfs and plotsimilarity from the GCG Sequence Analysis Software Package.

Nucleotide sequence accession number

The nucleotide sequence of the gerXB, gerXA and gerXC genes of B. anthracis strain 7702 have been submitted to GenBank and have been assigned the accession number AF108144.