Logo of aemPermissionsJournals.ASM.orgJournalAEM ArticleJournal InfoAuthorsReviewers
Appl Environ Microbiol. 2003 May; 69(5): 2755–2764.
PMCID: PMC154536

Genome Differences That Distinguish Bacillus anthracis from Bacillus cereus and Bacillus thuringiensis


The three species of the group 1 bacilli, Bacillus anthracis, B. cereus, and B. thuringiensis, are genetically very closely related. All inhabit soil habitats but exhibit different phenotypes. B. anthracis is the causative agent of anthrax and is phylogenetically monomorphic, while B. cereus and B. thuringiensis are genetically more diverse. An amplified fragment length polymorphism analysis described here demonstrates genetic diversity among a collection of non-anthrax-causing Bacillus species, some of which show significant similarity to B. anthracis. Suppression subtractive hybridization was then used to characterize the genomic differences that distinguish three of the non-anthrax-causing bacilli from B. anthracis Ames. Ninety-three DNA sequences that were present in B. anthracis but absent from the non-anthrax-causing Bacillus genomes were isolated. Furthermore, 28 of these sequences were not found in a collection of 10 non-anthrax-causing Bacillus species but were present in all members of a representative collection of B. anthracis strains. These sequences map to distinct loci on the B. anthracis genome and can be assayed simultaneously in multiplex PCR assays for rapid and highly specific DNA-based detection of B. anthracis.

Bacillus anthracis, B. cereus, and B. thuringiensis are genetically very closely related members of the group 1 bacilli, a fact which has led to the proposal that they should be considered a single species (15). B. cereus is frequently isolated as a contaminant of various foods and can occasionally be an opportunistic human pathogen (9, 15). B. thuringiensis has been widely exploited in agriculture as an insecticide by virtue of the presence of plasmid-borne crystal toxin genes (34). B. anthracis is a virulent pathogen of mammals and is the causative agent of anthrax. All three species are readily isolated from soil environments (38).

Extensive genetic diversity among environmental isolates of B. cereus and B. thuringiensis has been demonstrated by pulsed-field gel electrophoresis (6), multienzyme electrophoresis (6, 15), and amplified fragment length polymorphism (AFLP) analysis (38). AFLP analysis proved sensitive enough to classify B. cereus and B. thuringiensis into five phylogenetic groups (38). Significantly, the American Type Culture Collection (ATCC) reference strains of B. cereus and B. thuringiensis did not seem to be represented in the collection of environmental isolates used in that study. There was little correlation between species designation and the five phylogenetic groups identified by AFLP analysis, indicating significant genetic variability within and between B. cereus and B. thuringiensis. Furthermore, in two separate studies, the B. anthracis strains appeared to cluster together with a group containing periodontal B. cereus pathogens (16, 38).

In contrast to the genomic diversity within B. cereus and B. thuringiensis, B. anthracis appears to be genetically clonal (19). Genetic identity among different B. anthracis isolates has necessitated fastidious analysis of their genomes for strain identification. Accurate discrimination is now possible by analysis of variable-number tandem repeats, which enumerates small tandem repeats at several locations in the B. anthracis genome (3, 20). A collection of 426 B. anthracis environmental isolates that contains representatives from worldwide origins was subdivided by variable-number tandem repeat analysis into just six genetically distinct groups (20).

Several strains of B. cereus and B. thuringiensis which appear to be genetically related to B. anthracis have been isolated (15, 16, 38). Isolation and characterization of the genome regions unique to B. anthracis will provide clues to its genetic relationship to these strains and ultimately direct pathogenicity studies. AFLP analysis provides a rapid method for measuring phylogenetic distances. AFLP analysis generates a strain-specific fingerprint of amplified DNA fragments that demonstrate genomic variations in a microbial population based on an analysis of a portion of their genome sequences and DNA fragment length polymorphisms. The great advantage of AFLP analysis is its ability to analyze rapidly many loci, resulting in a phylogenetic resolution higher than those obtained with other methods. However, AFLP analysis provides little information about the genetic differences responsible for these polymorphisms. Suppression subtractive hybridization (SSH) is a highly efficient technique for the isolation and characterization of the large genomic differences that often drive bacterial genome evolution (1, 2). SSH reveals DNA sequence differences that are responsible for many AFLPs.

Unique genomic differences can be exploited as “DNA signatures” for the discrimination of B. anthracis from its closest relatives. Plasmid-encoded toxin genes have proved a useful source of targets for rapid DNA-based assays (4). However, a plasmid-based test may not detect the occurrence of non-plasmid-containing strains of B. anthracis, which have been isolated from the environment (39). Furthermore, plasmids can be readily engineered and can be transferred to other bacteria for heterologous gene expression (5, 10, 37), and there is concern that pXO1 and pXO2 sequences that are present in other Bacillus strains may be encountered in previously uncharacterized genomes (25). There are many published examples of chromosomal regions that can be examined for B. anthracis identification, but all require time-consuming downstream analysis (3, 7, 18, 26). Two recent examples of real-time PCR targeting of the rpoB gene of B. anthracis were useful but reported examples of false-positive results with some strains of B. cereus and targeted only a single locus (11, 28).

This report demonstrates the phylogenetic relationships of a collection of non-anthrax-causing Bacillus species to each other and to B. anthracis, as determined by AFLP analysis. SSH with B. anthracis and three of these close relatives identified a set of unique DNA regions that represent genomic differences between B. anthracis and these non-anthrax-causing Bacillus species. One immediate product of these experiments was a robust set of chromosomal DNA signatures which are capable of quickly detecting all six genetically distinct groups of B. anthracis (20) and which can be used with any rapid DNA-based detection platform. A multiplex PCR analysis with four separate loci of the B. anthracis chromosome provides a rapid and highly specific means for the identification of B. anthracis.


AFLP analysis of DNA samples.

AFLP analysis was accomplished as previously described (19, 40). Briefly, DNA (100 ng) was digested with EcoRI and MseI, and the resulting fragments were ligated to double-stranded adaptors. The digested and ligated DNA was then amplified by PCR with EcoRI and MseI +0/+0 primers. The +0/+0 PCR product was analyzed by agarose gel electrophoresis to determine the size range of amplified fragments. Three microliters was used in subsequent selective amplifications with 6-carboxyfluorescein-labeled +1/+1 primers EcoRI-C (5′-GTAGACTGCGTACCAATTCC-3′) and MseI-G (5′-GACGATGAGTCCTGAGTAAG-3′). Selective amplifications were performed with 20-μl reaction mixtures. The resulting products (0.5 to 1.0 μl) were mixed with a solution containing the DNA size standards Genescan-500 (Applied Biosystems, Inc., Foster City, Calif.) and MapMarker-400 (BioVentures, Inc., Murfreesburo, Tenn.), both labeled with N,N,N,N-tetramethyl-6-carboxyrhodamine. Following heat denaturation at 90°C for 2 min, the reaction mixtures were loaded into a 5% Long Ranger DNA sequencing gel (BioWhittaker Molecular Applications, Rockland, Maine) and visualized with an ABI 377 automated fluorescence sequencer (Applied Biosystems). GeneScan analysis software (Applied Biosystems) was used to determine the lengths of the sample fragments by comparison to the DNA fragment length size standards included with each sample. AFLP data analysis was performed as described by Ticknor et al. (38).

Preparation of genomic DNA.

The bacterial strains used for genomic DNA preparation and their geographic origins, where known, are listed in Table Table1.1. DNA was extracted from cell pellets from overnight cultures grown in L broth by using a MasterPure DNA purification kit (Epicentre, Madison, Wis.) according to the manufacturer's instructions. Genomic DNA was prepared from enteric bacteria by using Wizard Genomic DNA Preps (Promega, Madison, Wis.) according to the manufacturer's instructions. Human genomic DNA was purchased from Clontech (Palo Alto, Calif.), and bovine genomic DNA was purchased from Novagen (Madison, Wis.). Soil DNA was extracted by using an UltraClean soil DNA kit (MoBio, Solana Beach, Calif.) according to the manufacturer's instructions. DNA isolated from organisms present in the air was prepared from filters as previously described (31). Both soil and air filter DNAs were prepared from samples originating in Livermore, Calif. After ethanol precipitation, all genomic DNAs were dissolved in 10 mM Tris HCl (pH 8.0) to a concentration of approximately 0.2 μg/ml.

Strains used in this study


Genome comparisons by SSH were performed as previously described, except for the differences noted below (2, 30). Briefly, the tester-specific DNA was digested with restriction endonucleases (RsaI and MseI; New England Biolabs, Beverly, Mass.) used according to the manufacturer's instructions. The fragments were first marked by ligation to specialized oligonucleotide adaptors. When the marked DNA was denatured and hybridized to excess unmarked driver-specific DNA that had been digested with the same enzymes, most tester-specific sequences formed heterohybrids with the driver. Some tester-specific sequences, however, self-hybridized to form amplifiable fragments that were then enriched by PCR, cloned, and sequenced. Agarose gel electrophoresis determined that the restriction endonucleases MseI and RsaI cut the genomic DNA to generate fragments in the optimal size range (200 to 1,000 bp). Modified adaptors were constructed to allow for subtractions with MseI-digested DNA as previously described (1). Namely, adaptor 1 was formed by annealing the “adaptor 1 long” oligonucleotide (2) with the oligonucleotide 5′-TAACCTGCCCGG to form an adaptor with appropriate cohesive ends. Adaptor 2 was formed by annealing the “adaptor 2 long” oligonucleotide with the oligonucleotide 5′-TAACCTCGGCCG. T4 DNA ligase (New England Biolabs) was inactivated by incubation at 72°C for 20 min. Three separate SSH experiments were performed with B. anthracis Ames as the tester and three of the non-anthrax-causing Bacillus strains listed in Table Table11 as the drivers: B. cereus ATCC 14579, B. cereus 3A, and B. thuringiensis Al Hakam.

DNA sequencing.

Nonpurified PCR products were cloned by using a pGEM-T Easy TA cloning kit (Promega). Recombinant clones were picked by using a BioPick automated colony picker (BioRobotics, Woburn, Mass.), and plasmid templates were prepared by boiling lysis and magnetic bead capture with a high-throughput procedure (35). Sequencing of plasmid templates was performed by using an Applied Biosystems Big-Dye Terminator system and either an ABI 377 or an ABI 3700 automated sequencer. The sequencing primers used were 5′-TGTAAAACGACGGCCAGT (forward) and 5′-CAGGAAACAGCTATGACC (reverse). The resulting data were analyzed by using ABI sequencing analysis software, version 3.2, and then assembled and edited by using Phred, Phrap, and Consed 7.0 (14). BLAST searches with the tester-specific DNA sequences were performed by using the National Center for Biotechnology Information website http://www.ncbi.nlm.nih.gov/. BLAST identity was considered significant only when the expected probability of a fortuitous match (E value) was less than 10−4. Comparison of the sequence candidates to plasmids pXO1 and pXO2 and determination of their coordinates on the published B. anthracis genome (32) were performed by using the cross_match program, which is part of the Consed software package (14).


Oligonucleotide primers were designed from the putative tester-specific sequences and were supplied by Sigma-Genosys (The Woodlands, Tex.) or Invitrogen (Carlsbad, Calif.). The primers were designed by using Primer3 software (33); they had a melting temperature of >65°C, contained no more than three identical consecutive nucleotides, and possessed a two-nucleotide 3′ GC clamp. The primers were initially screened against genomic DNAs from both the tester and the driver. To determine whether a primer pair was tester specific, 1 ng each of tester- and driver-specific DNAs was used as a template in PCRs with Accuprime polymerase, primers at 10 μM (Invitrogen), and the following cycling parameters: 94°C for 15 s, 65°C for 15 s, and 72°C for 30 s for 32 cycles. The products were visualized on a 1.5% agarose gel run in 0.5× Tris-borate-EDTA; if a product was present with tester-specific DNA as a template and absent with driver-specific DNA, then that sequence was designated tester specific. The tester-specific oligonucleotides were then used to prime PCRs with the B. anthracis, B. cereus, and B. thuringiensis strains listed in Table Table11 and the same reaction conditions as those described above. The integrity of the genomic DNA template was tested in all PCRs with primers specific for a region of the 23S gene conserved in Bacillus species: 5′-CTACCTTAGGACCGTTATAGTTAC and 5′-AGGTAGGCGAGGAGAGAATCC. Multiplex PCRs were performed as described above, except that template DNA was added to a final concentration of 10 ng/μl. The primers (see Table Table5)5) were used at the following concentrations: 23S primers at 3 μM (288 bp), dhp73.017 (241 bp) at 20 μM, dhp73.019 (196 bp) at 10 μM, dhp61.183 (163 bp) at 5 μM, and dhp77.002 (133 bp) at 20 μM.

Primers designed from 28 B. anthracis-specific DNA signatures and their coordinates on the B. anthracis A2012 genome (GenBank accession number NC_003995 ...


Genetic relatedness of B. anthracis to B. cereus and B. thuringiensis, as determined by AFLP.

The phylogenetic tree derived from the fluorescent AFLP data is shown in Fig. Fig.1.1. The Bacillus isolates used in this study were chosen from a previous AFLP study of over 300 Bacillus isolates (K. K. Hill and P. J. Jackson, unpublished data). They included several B. cereus and B. thuringiensis isolates that were found to be the most closely related to B. anthracis and two more distantly related B. cereus and B. thuringiensis strains (ATCC type strains 14579 and 10792, respectively). The other Bacillus isolates represented on the tree include pathogenic and nonpathogenic isolates; details for these isolates are shown in Table Table1.1. B. anthracis is represented in this tree by the Ames and Vollum strains. A large number of B. anthracis isolates were previously tested and found to possess AFLP profiles indistinguishable from those of these two strains (data not shown). The isolate most closely related to B. anthracis is B. thuringiensis 97-27. This isolate was collected from the wound of a French soldier and was shown to be capable of infecting and killing immunocompetent mice in subsequent studies (17). The three other B. thuringiensis isolates on the tree are HD-571, obtained from the U.S. Department of Agriculture B. thuringiensis collection; Al Hakam, collected by the United Nations Special Commission at a suspected bioweapons facility in Iraq; and the ATCC type strain (ATCC 10792). The B. cereus isolates also include pathogenic and nonpathogenic members of this species. Three B. cereus isolates (F1-15, 3A, and D17) were collected from food sources that caused human illness, and one (S2-8) was a soil isolate. The B. cereus ATCC type strain (ATCC 14579) and ATCC 4342, an isolate from milk (36), were also included.

FIG. 1.
Phylogenetic tree of B. anthracis, B. cereus, and B. thuringiensis isolates. The phylogenetic tree was derived from fluorescent AFLP analyses of 12 B. anthracis (Ba), B. cereus (Bc), and B. thuringiensis (Bt) isolates. The phylogenetic tree was based ...

Data were analyzed by using three or more replicates of the AFLP profiles to compare each isolate to the other isolates. AFLP fragments that appeared in every replicate for an isolate were used to define a fingerprint for that isolate. Each gel contained one or more control DNA samples to allow comparisons of profiles on the different gels. Jaccard coefficients for the fingerprints of the isolates being compared were analyzed by using an unweighted pair-group method with arithmetic means to produce the dendrogram. Based on an analysis of control DNA samples, the variability within the analysis was less than about 0.2 for the Jaccard coefficients. Any differences below this level therefore may be due to experimental or analytical variability and cannot be considered significant. Differences above this level are reproducible and carry a high degree of confidence (38).

Isolation of tester-specific DNA sequences by SSH.

SSH identifies DNA sequences that are specific to one genome (tester) and absent from the other genome (driver). Typically these DNA sequences range in length from 200 to 1,000 bp and are referred to as difference products. Three subtractions were undertaken to increase the yield and representation of B. anthracis-specific regions. B. cereus ATCC 14579 was selected because it represents one of the two most distantly related strains in this study and would be most likely to yield difference products. B. cereus 3A was chosen because it represents one of the closest relatives of B. anthracis, while B. thuringiensis Al Hakam represents intermediate relatedness.

A total of 256 candidate sequences were generated from subtractions by using genomic DNA from B. cereus ATCC 14579 as the driver (Table (Table2).2). Sequences that do not occur on plasmids pXO1 and pXO2 of B. anthracis were of primary interest, since some plasmid sequences are conserved in closely related Bacillus species (25). Sequences identical to pXO1 and pXO2 were eliminated by computer comparisons to published nucleotide sequences (see Materials and Methods) (24). The remaining sequences were used to design oligonucleotide primers for PCR analysis, to determine their representation in the tester and driver genomes. PCR experiments with B. anthracis Ames and B. cereus ATCC 14579 DNA as a template identified 39 B. anthracis Ames-specific sequences (15% of the tester-specific candidates). Six of the tester-specific candidates (2%) did not amplify a product when DNA from the collection of non-anthrax-causing pathogens was used as a template. Similarly, 28 out of 48 sequences (58%) were absent from B. thuringiensis Al Hakam; 8 of these (16%) did not occur in the collection of closely related non-anthrax-causing pathogens (Table (Table2).2). Finally, 48 sequences each were generated from two separate subtractions by using genomic DNAs prepared from B. cereus 3A and B. thuringiensis Al Hakam as drivers. A total of 26 out of 48 sequences (54%) were present in B. anthracis Ames but absent from B. cereus 3A; 14 of these did not occur in the collection of closely related non-anthrax-causing pathogens (Table (Table22).

Isolation of 28 B. anthracis-specific DNA signatures from three SSH experiments

It is crucial that DNA signatures intended for DNA-based detection of B. anthracis show no false-negative results with all isolates that might be encountered in environmental samples. All of the B. anthracis-specific candidates successfully amplified a PCR product when DNAs from the eight representatives of B. anthracis (Table (Table1)1) were used as templates (data not shown) (20).

Distribution of tester-specific loci in B. cereus and B. thuringiensis.

Genomic variations within the collection of non-anthrax-causing pathogens were observed as the presence or absence of 39 tester-specific regions isolated from the subtraction in which genomic DNA from B. cereus ATCC 14579 was used as the driver and DNA from B. anthracis Ames was used as the tester (Table (Table3).3). The tester-specific primers are listed in Table Table33 in decreasing order of the number of products amplified from this panel of non-anthrax-causing pathogens. Since, by definition, the driver will always be negative, there will be a maximum of nine loci possible in the 10 non-anthrax-causing Bacillus species. Similarly, the tester will always be positive, so that the minimum number of products possible will be one. The data in Table Table33 show that seven primer sets are present in nine templates (excluding the driver), while six candidates (shown in bold type) are represented in B. anthracis only. The columns of data for non-anthrax-causing pathogens are arranged from left to right in Table Table33 according to the number of tester-specific sequences seen in their respective genomes. The maximum number of candidates possible for each non-anthrax-causing Bacillus species is 39 for the tester (B. anthracis), while the minimum is 0 for the driver (B. cereus ATCC 14579). Strains that have more markers in common are suggested to be more closely related to B. anthracis. The data in Table Table33 show that B. cereus 3A is the most closely related (31 out of 39 sequences), followed by B. cereus S2-8 and D17 (29 out of 39 sequences); as expected, B. thuringiensis ATCC 10792 is the least closely related (10 out of 39 sequences).The presence or absence of these loci is not correlated with the current species designations of B. cereus and B. thuringiensis. Table Table44 shows data obtained with primers designed from the subtraction with the most closely related DNA as a driver. Far fewer primer candidates detect the non-anthrax-causing Bacillus species, with 14 (shown in bold type) being found only for B. anthracis Ames.

Presence or absence of B. anthracis Ames-specific nucleotide sequences in the genomes of 10 non-anthrax-causing Bacillus pathogens as determined by PCR with B. cereus ATCC 14579 as the driver
Presence or absence of B. anthracis Ames-specific nucleotide sequences in the genomes of 10 non-anthrax-causing Bacillus pathogens as determined by PCR with B. cereus 3A as the driver

B. anthracis-specific DNA sequences and primers.

From the original collection of 352 fragments isolated from three subtractions, 28 were specific for B. anthracis and did not amplify a product with the near neighbors. All 93 tester-specific primers, some of which can be used to distinguish the B. cereus and B. thuringiensis strains used in this study, can be found at http://bbrp.llnl.gov/html/BAspc.html. All primers successfully amplified a PCR product of the predicted size, showing no evidence of variations at these loci in the B. anthracis strains (data not shown). These 28 B. anthracis-specific candidates were used to screen DNAs from common enteric pathogens—Yersinia pestis, Y. pseudotuberculosis, Y. enterocolitica, and Escherichia coli—species that may be encountered in environmental samples in suspected anthrax cases (soil, air, human, and bovine samples). All 28 candidates failed to amplify a product from these bacterial DNAs (data not shown). The average G+C content of these sequences was 35% (28 to 44%), a value consistent with that of B. anthracis.

All but 1 of the 28 B. anthracis-specific DNA sequences can be mapped to the 5.23-Mb genome of B. anthracis A2012 (GenBank accession number NC_003995) (32). BLASTX analysis showed that 24 of the remaining 27 DNA sequences map to the open reading frames defined therein, while the remainder map to intergenic regions. The coordinates for nucleotide identity to B. anthracis A2012, open reading frame identity, and gene identity are listed in Table Table5.5. BLASTX identities with E values of less than 10−3 are also listed. BLAST data for the candidates from the B. cereus ATCC 14579 subtraction show that three of the six DNA sequences have no previously ascribed function, based on similarity searches of the GenBank database. One of the remaining three shows identity with the S-layer protein of B. anthracis, and the other two show identity with a hypothetical phage protein from Streptococcus pyogenes and a galactosyltransferase-related protein from Clostridium acetobutylicum. BLAST data for the candidates from the B. thuringiensis Al Hakam subtraction show that four of the eight DNA sequences have no previously ascribed function. The remaining four share sequence identity with a hypothetical protein in B. halodurans, a penicillin binding protein of B. cereus, a putative phage terminase of C. perfringens, and a cytosine-specific methyltransferase of B. halodurans. BLAST data for the candidates from the B. cereus 3A subtraction show that 9 of the 14 DNA signatures have no previously ascribed function. The remaining five show sequence identity with the Staphylococcus aureus terminase large subunit, a hypothetical protein from the nanH region and an ATP binding cassette transporter of C. perfringens, a glucosamine synthetase of B. subtilis, and an unknown conserved protein of B. halodurans.

Figure Figure22 provides a visualization of the distribution of the tester-specific sequences for each of the three subtractive hybridization experiments. There are five regions that have more than two B. anthracis-specific loci that lie within 50 kb of each other, suggesting genomic islands that are found only in B. anthracis. Similar genomic islands have been observed when genomes of different strains of the same bacterial species have been compared (e.g., Y. pestis [8, 29] and E. coli [27]).

FIG. 2.
Graphic representation of the locations of the tester-specific clones on the B. anthracis A2012 genome. The outermost circle maps the locations of the clones isolated from the subtraction by using DNA from B. cereus ATCC 14579 as the driver (circles), ...

Multiplex PCR analysis for the simultaneous detection of four B. anthracis-specific loci.

Multiplex PCR is a powerful tool for the simultaneous detection of multiple loci within bacterial genomes. The simultaneous detection of four separate loci (A, C, D, and E) on the B. anthracis genome was achieved here by selecting primers (shown in bold type in Table Table5)5) that target these loci while providing sufficient size discrimination for resolution by gel electrophoresis on 4% agarose (163, 133, 196, and 241 bp, respectively). An internal positive control (288 bp) was designed from a region of the 23S gene conserved in Bacillus species (see Material and Methods) and confirmed the integrity of the DNA template. Figure Figure33 shows that the Bacillus 23S control primer yielded a 288-bp PCR product for all DNA templates (10 ng) tested. Furthermore, the four predicted B. anthracis-specific bands were seen for all strains (Fig. (Fig.3,3, lanes 13 to 23) and were absent from B. cereus and B. thuringiensis strains (lanes 2 to 11). This multiplex PCR was capable of detecting all four loci with as little as 100 pg of template DNA (data not shown).

FIG. 3.
Multiplex analysis of four separate loci on the B. anthracis genome. Four B. anthracis-specific primers yielded predicted products of 133, 163, 196, and 241 bp. One internal positive control yielded a predicted product of 288 bp. The DNA templates used ...


B. anthracis is a potent mammalian pathogen and bioterrorist agent. This pathogenic Bacillus species shares so much genetic material with B. cereus and B. thuringiensis that its discrimination from the other species can be problematic. All three Bacillus species are prevalent in many environments, and it is important to define the genetic differences specific to B. anthracis in order to design specific DNA-based identification protocols. We surmised that the most efficient approach to finding B. anthracis-specific DNA sequences would be to find the Bacillus species that are most closely related to B. anthracis and then to compare their genomes in vitro by using SSH.

AFLP analysis was used to reveal genetic diversity among non-anthrax-causing Bacillus strains and to determine which are most closely related to the highly pathogenic B. anthracis. The phylogenetic tree shown in Fig. Fig.11 indicates that B. cereus 3A, B. cereus S2-8, and B. thuringiensis 97-27 are the strains most genetically similar to B. anthracis of the 10 strains examined. B. cereus ATCC 4342 was shown previously by multienzyme electrophoresis (16) and AFLP analysis (38) to be very closely related to B. anthracis. The AFLP analysis presented here shows that seven strains are even more closely related to B. anthracis. The relationships do not correlate with the species designation for B. cereus or B. thuringiensis, providing another example of how the species designations for B. cereus and B. thuringiensis appear to cross the boundaries of various phylogenetic analyses (15, 38). It may be necessary to develop new criteria for species designations within this group, as information about more Bacillus genomes becomes available.

Although AFLP analysis provides a sensitive method for defining the genetic relationships between bacterial genomes, it provides no information regarding the genetic rearrangements responsible for these differences. Such information can be attained by in vitro genome comparison by SSH, a proven, efficient method for the identification of nucleotide sequences that differ between two genomes (1). The complementary techniques of AFLP analysis and SSH provide a powerful means of defining accurate phylogenetic models and characterizing their underlying genetic components. Three subtractions were performed to identify B. anthracis-specific nucleotide sequences that were absent from the non-anthrax-causing Bacillus species. B. cereus ATCC 14579 represents the most distantly related strain in this study. B. cereus 3A was chosen because it represents one of the closest relatives of B. anthracis. The third strain, of intermediate relatedness, was B. thuringiensis Al Hakam. It would be expected that strains with more sequence identity to B. anthracis (in this situation, B. cereus 3A) would yield more sequences that would not be found in the closely related B. cereus and B. thuringiensis. Indeed, the subtraction with B. cereus 3A yielded 29% B. anthracis-specific sequences, compared to 17 and 2% for B. thuringiensis Al Hakam and B. cereus ATCC 14579, respectively (Table (Table2),2), confirming the prediction from AFLP analysis.

The genetic diversity demonstrated within this strain collection by AFLP analysis was mirrored by the results of subsequent PCR analysis. Bacterial evolution is driven by rearrangements of large genomic islands associated with lateral gene transfer (21, 23). Analysis of the G+C contents of the difference products reveals an average of 35.4%, typical of B. anthracis. This finding suggests that any lateral gene transfer has ameliorated the G+C content over a long period of evolution or has been received from species with a similar G+C content. The presence or absence of each of the tester-specific sequences in the non-anthrax-causing pathogens was determined by PCR amplification with primers designed for the difference products. If the number of loci shared by B. anthracis indicates relatedness among the whole panel of non-anthrax-causing Bacillus strains, then B. cereus 3A is the most closely related (Table (Table3).3). This result is in agreement with the phylogenetic analysis shown in Fig. Fig.1.1. The phylogenetic tree shows two pairs of isolates that cannot be distinguished by AFLP analysis: Bacillus strains Vollum and Ames and B. cereus strains 3A and S2-8. SSH with B. cereus ATCC 14579 as a driver yielded two primer sets (M.Ctg056 and M.Ctg007) that can distinguish B. cereus 3A from S2-8 (Table (Table3).3). Primers from the same subtraction cannot distinguish B. thuringiensis HD-571 from Al Hakam. Table Table44 demonstrates that the number of loci that are shared between the non-anthrax-causing Bacillus species and B. anthracis decreases when the driver strain is much more closely related to B. anthracis. This subtraction also yielded two primer sets (dhp73.03 and dhp73.04) that can distinguish B. thuringiensis HD-571 from Al Hakam, which are indistinguishable by the primer sets listed in Table Table33.

The most important criteria for effective DNA signatures are the absence of false-positive results with closely related organisms and their representation in all isolates of the target (i.e., no false-negative results). Given the monomorphic nature of the B. anthracis genome, it was not surprising to find that there was no variation in the signatures within the strains of the collection. Twenty-eight B. anthracis-specific candidates isolated in these experiments fulfilled these criteria and are listed in Table Table5.5. None amplified a PCR product from any of the non-anthrax-causing Bacillus pathogens used in this study (no false-positive results). We were also able to exploit a collection of genetically distinct and geographically diverse isolates of B. anthracis. The twenty-eight DNA signatures amplified a PCR product of the predicted size for every isolate. A comparison of the DNA sequences against the completed genomes of B. halodurans and B. subtilis and the unfinished Bacillus genomes showed no significant sequence identity.

The non-anthrax-causing Bacillus species described here are so closely related to B. anthracis that they would be highly likely to give false-positive results in DNA-based identification assays based on chromosomal loci. The isolation of multiple B. anthracis-specific chromosomal regions allowed the development of a single multiplex assay for the rapid and highly specific detection of B. anthracis. The DNA signatures presented here have the advantage over previous detection methods that require time-consuming analysis (3, 7), are prone to false-positive results, and are based on few nucleotide differences at a single chromosomal locus (11, 28).

BLAST analysis of the nucleotide sequences of these DNA signatures shows that many are not represented in current DNA databases, other than the previously reported B. anthracis A2012 genome (Table (Table5)5) (32). The strongest identity seen was to the penicillin binding protein of B. cereus (dhp64.177; E value, 10−67). R.Ctg122 showed identity to the S-layer protein of B. anthracis (13). This cell surface protein is also seen in some isolates of B. cereus and B. thuringiensis (22). However, there is sufficient nucleotide sequence divergence at the oligonucleotide primer binding sites to allow for successful discrimination of B. anthracis from the other two species.

The genomes of several strains of B. anthracis are currently being sequenced (12), and these data will be extremely useful for strain attribution in forensic analyses. We envisage that these DNA signatures can be used for real-time specific detection of B. anthracis, the source of which may then be attributed by monitoring the small nucleotide differences identified by these sequencing projects. The B. anthracis-specific DNA sequences identified in this work provide the largest collection of chromosomal markers that distinguish B. anthracis from other closely related Bacillus species. There are five regions in the B. anthracis genome where several of the specific DNA sequences are located within 50 kb of each other. Such genomic islands may define B. anthracis as a species and distinguish it from the closely related species B. cereus and B. thuringiensis. Future detailed analysis of these B. anthracis-specific regions may ultimately identify chromosome-encoded virulence factors, provide starting points for possible vaccine candidates, and help to reveal the mode of pathogenicity of this important pathogen.


This work was performed under the auspices of the U.S. Department of Energy at the University of California (Lawrence Livermore National Laboratory, under contract no. W-7405-Eng-48). The Chemical and Biological National Security Program of the U.S. Department of Energy funded this work.

We are extremely grateful to Kimothy Smith for assistance with the B. anthracis collection used in this study and Nancy K. Montgomery for preparing genomic DNA. We also appreciate the technical contributions of Aubree Hubbell, Anne M. Erler, Silvia Gamez-Chin, Cheryl Strout, Julie R. Avila, and Linda L. Ott at Lawrence Livermore National Laboratory.


1. Agron, P. G., M. Macht, L. Radnedge, E. W. Skowronski, W. Miller, and G. L. Andersen. 2002. Use of subtractive hybridization for comprehensive surveys of prokaryotic genome differences. FEMS Microbiol. Lett. 211:175-182. [PubMed]
2. Akopyants, N. S., A. Fradkov, L. Diatchenko, J. E. Hill, P. D. Siebert, S. A. Lukyanov, E. D. Sverdlov, and D. E. Berg. 1998. PCR-based subtractive hybridization and differences in gene content among strains of Helicobacter pylori. Proc. Natl. Acad. Sci. USA 95:13108-13113. [PMC free article] [PubMed]
3. Andersen, G. L., J. M. Simchock, and K. H. Wilson. 1996. Identification of a region of genetic variability among Bacillus anthracis strains and related species. J. Bacteriol. 178:377-384. [PMC free article] [PubMed]
4. Bell, C. A., J. R. Uhl, T. L. Hadfield, J. C. David, R. F. Meyer, T. F. Smith, and F. R. Cockerill III. 2002. Detection of Bacillus anthracis DNA by LightCycler PCR. J. Clin. Microbiol. 40:2897-2902. [PMC free article] [PubMed]
5. Bradley, K. A., J. Mogridge, M. Mourez, R. J. Collier, and J. A. Young. 2001. Identification of the cellular receptor for anthrax toxin. Nature 414:225-229. [PubMed]
6. Carlson, C. R., D. A. Caugant, and A.-B. Kolstø. 1994. Genotypic diversity among Bacillus cereus and Bacillus thuringiensis strains. Appl. Environ. Microbiol. 60:1719-1725. [PMC free article] [PubMed]
7. Daffonchio, D., S. Borin, G. Frova, R. Gallo, E. Mori, R. Fani, and C. Sorlini. 1999. A randomly amplified polymorphic DNA marker specific for the Bacillus cereus group is diagnostic for Bacillus anthracis. Appl. Environ. Microbiol. 65:1298-1303. [PMC free article] [PubMed]
8. Deng, W., V. Burland, G. Plunkett III, A. Boutin, G. F. Mayhew, P. Liss, N. T. Perna, D. J. Rose, B. Mau, S. Zhou, D. C. Schwartz, J. D. Fetherston, L. E. Lindler, R. R. Brubaker, G. V. Plano, S. C. Straley, K. A. McDonough, M. L. Nilles, J. S. Matson, F. R. Blattner, and R. D. Perry. 2002. Genome sequence of Yersinia pestis KIM. J. Bacteriol. 184:4601-4611. [PMC free article] [PubMed]
9. Drobniewski, F. A. 1993. Bacillus cereus and related species. Clin. Microbiol. Rev. 6:324-338. [PMC free article] [PubMed]
10. Drum, C. L., S. Z. Yan, J. Bard, Y. Q. Shen, D. Lu, S. Soelaiman, Z. Grabarek, A. Bohm, and W. J. Tang. 2002. Structural basis for the activation of anthrax adenylyl cyclase exotoxin by calmodulin. Nature 415:396-402. [PubMed]
11. Ellerbrok, H., H. Nattermann, M. Ozel, L. Beutin, B. Appel, and G. Pauli. 2002. Rapid and sensitive identification of pathogenic and apathogenic Bacillus anthracis by real-time PCR. FEMS Microbiol. Lett. 214:51-59. [PubMed]
12. Enserink, M. 2002. Microbial genomics. TIGR begins assault on the anthrax genome. Science 295:1442-1443. [PubMed]
13. Etienne-Toumelin, I., J. C. Sirard, E. Duflot, M. Mock, and A. Fouet. 1995. Characterization of the Bacillus anthracis S-layer: cloning and sequencing of the structural gene. J. Bacteriol. 177:614-620. [PMC free article] [PubMed]
14. Gordon, D., C. Abajian, and P. Green. 1998. Consed: a graphical tool for sequence finishing. Genome Res. 8:195-202. [PubMed]
15. Helgason, E., D. A. Caugant, I. Olsen, and A.-B. Kolstø. 2000. Genetic structure of population of Bacillus cereus and B. thuringiensis isolates associated with periodontitis and other human infections. J. Clin. Microbiol. 38:1615-1622. [PMC free article] [PubMed]
16. Helgason, E., O. A. Okstad, D. A. Caugant, H. A. Johansen, A. Fouet, M. Mock, I. Hegna, and A.-B. Kolstø. 2000. Bacillus anthracis, Bacillus cereus, and Bacillus thuringiensis—one species on the basis of genetic evidence. Appl. Environ. Microbiol. 66:2627-2630. [PMC free article] [PubMed]
17. Hernandez, E., F. Ramisse, J. P. Ducoureau, T. Cruel, and J. D. Cavallo. 1998. Bacillus thuringiensis subsp. konkukian (serotype H34) superinfection: case report and experimental evidence of pathogenicity in immunosuppressed mice. J. Clin. Microbiol. 36:2138-2139. [PMC free article] [PubMed]
18. Jackson, P. J., E. A. Walthers, A. S. Kalif, K. L. Richmond, D. M. Adair, K. K. Hill, C. R. Kuske, G. L. Andersen, K. H. Wilson, M. Hugh-Jones, and P. Keim. 1997. Characterization of the variable-number tandem repeats in vrrA from different Bacillus anthracis isolates. Appl. Environ. Microbiol. 63:1400-1405. [PMC free article] [PubMed]
19. Keim, P., A. Kalif, J. Schupp, K. Hill, S. E. Travis, K. Richmond, D. M. Adair, M. Hugh-Jones, C. R. Kuske, and P. Jackson. 1997. Molecular evolution and diversity in Bacillus anthracis as detected by amplified fragment length polymorphism markers. J. Bacteriol. 179:818-824. [PMC free article] [PubMed]
20. Keim, P., L. B. Price, A. M. Klevytska, K. L. Smith, J. M. Schupp, R. Okinaka, P. J. Jackson, and M. E. Hugh-Jones. 2000. Multiple-locus variable-number tandem repeat analysis reveals genetic relationships within Bacillus anthracis. J. Bacteriol. 182:2928-2936. [PMC free article] [PubMed]
21. Lawrence, J. G., and H. Ochman. 1997. Amelioration of bacterial genomes: rates of change and exchange. J. Mol. Evol. 44:383-397. [PubMed]
22. Mesnage, S., M. Haustant, and A. Fouet. 2001. A general strategy for identification of S-layer genes in the Bacillus cereus group: molecular characterization of such a gene in Bacillus thuringiensis subsp. galleriae NRRL 4045. Microbiology 147:1343-1351. [PubMed]
23. Ochman, H., J. G. Lawrence, and E. A. Groisman. 2000. Lateral gene transfer and the nature of bacterial innovation. Nature 405:299-304. [PubMed]
24. Okinaka, R., K. Cloud, O. Hampton, A. Hoffmaster, K. Hill, P. Keim, T. Koehler, G. Lamke, S. Kumano, D. Manter, Y. Martinez, D. Ricke, R. Svensson, and P. Jackson. 1999. Sequence, assembly and analysis of pX01 and pX02. J. Appl. Microbiol. 87:261-262. [PubMed]
25. Pannucci, J., R. T. Okinaka, R. Sabin, and C. R. Kuske. 2002. Bacillus anthracis pXO1 plasmid sequence conservation among closely related bacterial species. J. Bacteriol. 184:134-141. [PMC free article] [PubMed]
26. Patra, G., P. Sylvestre, V. Ramisse, J. Therasse, and J. L. Guesdon. 1996. Isolation of a specific chromosomic DNA sequence of Bacillus anthracis and its possible use in diagnosis. FEMS Immunol. Med. Microbiol. 15:223-231. [PubMed]
27. Perna, N. T., G. Plunkett III, V. Burland, B. Mau, J. D. Glasner, D. J. Rose, G. F. Mayhew, P. S. Evans, J. Gregor, H. A. Kirkpatrick, G. Posfai, J. Hackett, S. Klink, A. Boutin, Y. Shao, L. Miller, E. J. Grotbeck, N. W. Davis, A. Lim, E. T. Dimalanta, K. D. Potamousis, J. Apodaca, T. S. Anantharaman, J. Lin, G. Yen, D. C. Schwartz, R. A. Welch, and F. R. Blattner. 2001. Genome sequence of enterohaemorrhagic Escherichia coli O157:H7. Nature 409:529-533. [PubMed]
28. Qi, Y., G. Patra, X. Liang, L. E. Williams, S. Rose, R. J. Redkar, and V. G. DelVecchio. 2001. Utilization of the rpoB gene as a specific chromosomal marker for real-time PCR detection of Bacillus anthracis. Appl. Environ. Microbiol. 67:3720-3727. [PMC free article] [PubMed]
29. Radnedge, L., P. G. Agron, P. L. Worsham, and G. L. Andersen. 2002. Genome plasticity in Yersinia pestis. Microbiology 148:1687-1698. [PubMed]
30. Radnedge, L., S. Gamez-Chin, P. M. McCready, P. L. Worsham, and G. L. Andersen. 2001. Identification of nucleotide sequences for the specific and rapid detection of Yersinia pestis. Appl. Environ. Microbiol. 67:3759-3762. [PMC free article] [PubMed]
31. Radosevich, J. L., W. J. Wilson, J. H. Shinn, T. Z. DeSantis, and G. L. Andersen. 2002. Development of a high-volume aerosol collection system for the identification of air-borne micro-organisms. Lett. Appl. Microbiol. 34:162-167. [PubMed]
32. Read, T. D., S. L. Salzberg, M. Pop, M. Shumway, L. Umayam, L. Jiang, E. Holtzapple, J. D. Busch, K. L. Smith, J. M. Schupp, D. Solomon, P. Keim, and C. M. Fraser. 2002. Comparative genome sequencing for discovery of novel polymorphisms in Bacillus anthracis. Science 296:2028-2033. [PubMed]
33. Rozen, S., and H. Skaletsky. 2000. Primer3 on the WWW for general users and for biologist programmers. Methods Mol. Biol. 132:365-386. [PubMed]
34. Schnepf, E., N. Crickmore, J. Van Rie, D. Lereclus, J. Baum, J. Feitelson, D. R. Zeigler, and D. H. Dean. 1998. Bacillus thuringiensis and its pesticidal crystal proteins. Microbiol. Mol. Biol. Rev. 62:775-806. [PMC free article] [PubMed]
35. Skowronski, E. W., N. Armstrong, G. Andersen, M. Macht, and P. M. McCready. 2000. Magnetic microplate-format plasmid isolation protocol for high-yield, sequencing grade DNA. BioTechniques 29:786-790. [PubMed]
36. Smith, N. R., R. E. Gordon, and F. E. Clark. 1952. Aerobic spore forming bacteria. U. S. Dep. Agric. Monogr. 16:1-148.
37. Thwaite, J. E., L. W. Baillie, N. M. Carter, K. Stephenson, M. Rees, C. R. Harwood, and P. T. Emmerson. 2002. Optimization of the cell wall microenvironment allows increased production of recombinant Bacillus anthracis protective antigen from B. subtilis. Appl. Environ. Microbiol. 68:227-234. [PMC free article] [PubMed]
38. Ticknor, L. O., A.-B. Kolstø, K. K. Hill, P. Keim, M. T. Laker, M. Tonks, and P. J. Jackson. 2001. Fluorescent amplified fragment length polymorphism analysis of Norwegian Bacillus cereus and Bacillus thuringiensis soil isolates. Appl. Environ. Microbiol. 67:4863-4873. [PMC free article] [PubMed]
39. Turnbull, P. C., R. A. Hutson, M. J. Ward, M. N. Jones, C. P. Quinn, N. J. Finnie, C. J. Duggleby, J. M. Kramer, and J. Melling. 1992. Bacillus anthracis but not always anthrax. J. Appl. Bacteriol. 72:21-28. [PubMed]
40. Vos, P., R. Hogers, M. Bleeker, M. Reijans, T. van de Lee, M. Hornes, A. Frijters, J. Pot, J. Peleman, M. Kuiper, et al. 1995. AFLP: a new technique for DNA fingerprinting. Nucleic Acids Res. 23:4407-4414. [PMC free article] [PubMed]

Articles from Applied and Environmental Microbiology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • BioProject
    BioProject links
  • MedGen
    Related information in MedGen
  • Nucleotide
    Published Nucleotide sequences
  • Protein
    Published protein sequences
  • PubMed
    PubMed citations for these articles

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...