Background

Bacillus anthracis, the cause of anthrax in humans and animals, is an endospore-forming bacterium that can persist in the environment for many years. In the U.S., B. anthracis infections in animals occur in states that are enzootic for the disease, such as Texas, North and South Dakota, and Minnesota [1,2,3,4,5]. Anthrax outbreaks have occurred annually in livestock and wild game in Texas for the past decade. Outbreaks pertinent to this study include an outbreak in a cattle herd in July 2007 on a ranch north of San Angelo, Texas, and an outbreak in horses on an adjacent ranch in August 2007 [6, 7].

Identification of B. anthracis is commonly based on colony morphology (including hemolysis), capsule production, detection of virulence plasmids (pXO1 and pXO2) by real-time PCR, and susceptibility to lysis by gamma phage [8,9,10]. The gamma phage assay is widely used as a diagnostic test for B. anthracis in many laboratories due to ease and low cost of use. Recently, several studies have used gamma phage binding and lytic activity as the basis for newer diagnostic assays, such as using engineered phage for bioluminescent detection [11,12,13,14,15,16]. The specificity and sensitivity of gamma phage has been previously documented, and although the specificity for B. anthracis is very high (97%), a small number of isolates of close relatives, including examples of Bacillus cereus and Bacillus thuringiensis, are susceptible [9, 17]. In this study, we evaluated the specificity of the gamma phage assay in conjunction with other diagnostic tests to screen 700 spore-forming Texas soil isolates for identification of B. anthracis.

Methods

Texas soil collection and processing

Core soil samples were collected in September 2007 from two sites in Texas in or near the municipalities of Schertz and San Angelo. Schertz was chosen due to the previous identification of pathogenic B. cereus isolates (two clinical, one environmental) harboring B. anthracis toxin and/or capsule genes that were associated with fatal pneumonias in metal workers from the area in 2003 [18]. San Angelo was chosen due to the recent anthrax outbreaks in cattle and horses mentioned above [6, 7]. Sterile 50 mL conical tubes were uncapped and twisted down into the soil to a depth of approximately 2.5–3.0 cm. The tubes containing the core samples were then re-capped and wrapped in parafilm. Samples were shipped at ambient temperature and subsequently stored at – 30 °C until processing.

Three samples were selected for testing (one from Schertz and two from San Angelo), and 2–3 bacterial extractions were performed on each sample. For each extraction, approximately 3.5 g of soil were transferred to a sterile 15 mL conical tube. The soil was saturated with phosphate-buffered saline containing 0.3% Tween 20 (PBST) and an additional 1.0 mL of PBST was added, then the sample was vortexed at high speed for 1–3 min. To select for spore-forming bacteria, the samples were heat shocked at 65 °C for 30 min and allowed to settle. The supernatant was transferred to a 2.0 mL screw-cap tube, and briefly centrifuged at 2000–3000 rpm. The following amounts were plated to produce 20–50 colonies/plate: 100 µL of the undiluted supernatant was plated onto PLET (polymyxin, lysozyme, ethylenediaminetetraacetic acid, thallium acetate) agar plates, a selective medium for B. anthracis [19], while 100 µL aliquots of a 1:20 dilution were plated onto 10 trypticase soy agar plates containing 5% sheep blood (SBA). Both PLET and SBA plates were incubated overnight at 37 °C. Seven hundred random single colonies were picked and streaked to SBA for isolation. Isolates were then streaked onto sporulation medium (agar containing tryptone, peptone, yeast extract, and manganous chloride) and incubated at 30 °C for 5–7 days. Spores were harvested in 25% glycerol in deionized water and stored as spore suspensions at – 70 °C.

Isolate screening and identification

Phenotypic characteristics, including hemolysis and colony morphology on SBA (see Additional file 1), and gamma phage susceptibility were recorded for all isolates. Gamma phage testing was performed as previously described [9], with susceptibility defined as any amount of lysis or reduced growth where gamma phage was applied. Cell lysates containing DNA were prepared as previously described [8], and were tested using the Laboratory Response Network’s real-time PCR assay for detection of the B. anthracis chromosome, and of virulence plasmids pXO1 and pXO2 genes (pagA and capB, respectively) [8, 17]. For isolates exhibiting susceptibility to gamma phage lysis, the cell wall (CW) and capsule (CAP) direct fluorescent antibody (DFA) assays were also performed as previously described [10]. Gamma phage-susceptible isolates that were ruled-out for B. anthracis using the tests above were identified by 16S rRNA gene sequencing as previously described [20]. For isolates that produced readable sequences, the sequences were compared to published type strain sequences (when applicable) in NCBI’s GenBank database using BLAST, using a sequence similarity of 99–100%. The 16s rRNA sequences of isolates 2008723338, 2008723339, 2008723400, 2008723423, 200872372 and 2008723476 were deposited in GenBank with Accession Numbers KT254134, KT254135, KT254136, KT254137, KT254138 and KT254139, respectively. Isolates identified as belonging to the B. cereus group were further typed by multi locus sequence typing (MLST), performed as previously described [21].

Findings

Twenty-nine of the 700 isolates were susceptible to gamma phage lysis, with 20 exhibiting hemolysis on SBA and with variable colony morphologies (see Additional files 1 and 2). One isolate was identified as B. anthracis (PCR positive for all targets, nonhemolytic, gamma phage-susceptible, and positive by CW- and CAP-DFA assays). The other 28 isolates were negative for all three B. anthracis PCR markers, including virulence genes pagA and capB, and all were negative for the CW-DFA and CAP-DFA assays except for isolate 2008723634, which was weakly positive for the CW-DFA assay. Based on the combined results of these tests, 28 of the 29 gamma phage-susceptible isolates were ruled out as B. anthracis. We calculated the specificity of the gamma phage test as 96% by dividing the number of true negatives by the sum of the number of true negatives plus the number of false positives, multiplied by 100 (see Table 1). Twenty-five of these 28 isolates produced readable 16S rRNA gene sequences, and were identified as follows: six Lysinibacillus spp., one Bacillus sp., one Solibacillus silvestris, and 17 isolates could only be identified as “B. cereus group members” due to the inability to distinguish members of the B. cereus group based on 16S gene sequencing. Representative isolates from each group are shown in Fig. 1. The 16S sequence data was not analyzed for three isolates due to failure to amplify.

Table 1 Results of gamma phage susceptibility testing
Full size table
Fig. 1
figure 1

Neighbor-joining dendrogram of 16S rRNA gene sequences, showing the relationship of isolates and representative isolates from each grouping (shown in bold) to a panel of related bacteria. Bootstrap values (based on 1000 replications) are given as percentages at branch nodes. Brevibacillus brevis is used as an outgroup for this analysis

Full size image

Multi locus sequence typing was performed on all of the 17 B. cereus group isolates, however, only eight isolates produced usable gene segments for all seven alleles. All eight isolates were different sequence types (ST), and therefore, not multiple isolations of a single clone. Seven of the eight isolates were new STs, and were compared with other previously identified gamma phage-susceptible B. cereus isolates, as well as other B. cereus-group STs (Fig. 2) [17, 21]. None of the eight B. cereus group isolates identified cluster with previously identified gamma phage-susceptible isolates (ST-129, ST-130, or ST-132), however six of the isolates from this study did cluster closely together (Fig. 2) [21]. The other two B. cereus group isolates (2008723644 and 2008723286) were identified as ST-502 or clustered nearest to ST-502, respectively.

Fig. 2
figure 2

Relationships between B. cereus group member isolates of this study and select reference isolates using concatenated sequences from seven housekeeping alleles. Sequence types (STs) with previously identified gamma phage susceptible isolates are shown in bold, and isolates from this study with new STs are marked (asterisk). The tree was constructed using the neighbor joining method and percent bootstrap confidence levels were calculated using 1000 resamplings of the original data

Full size image

In conclusion, the diagnostic tests used in this study showed high specificity (real-time PCR 100% and gamma phage 96%). Interestingly, despite reports of B. cereus isolates harboring pXO1 and pXO2 plasmids or some of their genes (pagA and capB), these markers were not detected in any of the 700 isolates, except for the single B. anthracis isolate [18, 22,23,24]. Though gamma phage susceptibility is considered specific for B. anthracis, rare false positives can occur due to the susceptibility of a few other species. To our knowledge, this is the first report of susceptible isolates outside the Bacillus genus. Though the isolates described in this study are of environmental origin rather than clinical, and are often thought of as contaminants in the clinical laboratory, most can be opportunistic pathogens [25]. Therefore, it is important to note their existence, especially when new diagnostic methods are being developed using gamma phage or gamma phage proteins for the detection of B. anthracis in clinical or environmental samples [11,12,13,14,15,16].