OUP user menu

High abundance and diversity of Bacillus anthracis plasmid pXO1-like replicons in municipal wastewater

Martin Iain Bahl, Kirsten Rosenberg
DOI: http://dx.doi.org/10.1111/j.1574-6941.2010.00922.x 241-247 First published online: 1 October 2010


Water from the influent of a municipal wastewater treatment plant as well as soil samples collected from the shoreline of 10 lakes were screened for the presence of the Bacillus anthracis pXO1-like plasmid replicon repX using a PCR assay. Specific PCR products were retrieved from all samples, indicating a widespread presence of pXO1-like plasmid replicons in various environmental settings. Initial screening by restriction enzyme analysis revealed at least two forms of the repX gene in the wastewater sample, which was consequently subjected to further investigation. Nine of 51 Bacillus cereus group strains isolated from the wastewater sample were shown to contain a repX-specific gene sequence. Two of these strains were shown to have repX gene sequences with very high homology to the repX gene of plasmid pXO1. The same two strains also contained replicon-specific sequences with high homology to those of pXO2-like plasmids, but did not contain the pXO1-associated cya and lef virulence genes. Collectively, the sequence information from the isolated strains and PCR products obtained using total genomic DNA as a template suggests the existence of three subgroups of pXO1-like plasmid replicons in the wastewater sample.

  • Bacillus cereus sensu lato
  • pXO1
  • repX
  • phylogeny


The Bacillus cereus group (B. cereus sensu lato) of bacteria currently consists of B. cereus sensu stricto, Bacillus anthracis, Bacillus thuringiensis, Bacillus mycoides, Bacillus pseudomycoides and Bacillus weihenstephanensis, each classified as distinct species. The chromosomal structure and content of these species, however, all share a high degree of similarity (Rasko et al., 2005), and consequently, it has been proposed that they all, or as a minimum the extremely closely related species B. anthracis, B. thuringiensis and B. cereus s.s., should be classified as a single species or taxon (Helgason et al., 2000b; Jensen et al., 2003; Rasko et al., 2005; Han et al., 2006). The latter three species are pathogenic to mammals or insects, and observations of intermixing of 16S rRNA gene between strains of B. thuringiensis and B. cereus (Kolstø., 2009) and the recent observation of anthrax-like disease caused by B. cereus strains (Hoffmaster et al., 2004) highlight the very close relationship between these three species. Virulence factors found in B. anthracis and B. thuringiensis are mostly encoded by extrachromosomal elements, while the virulence factors in B. cereus are mostly chromosomal. The plasmids pXO1 and pXO2, which encode the tripartite toxin complex and the poly-γ-d-glutamic capsule, respectively, are often-cited examples of such extrachromosomal virulence elements, and carriage of these two plasmids is the hallmark of pathogenic B. anthracis. Other characteristic properties of the presumed monophyletic B. anthracis species are that they are capsule-producing, nonmotile, susceptible to the γ-phage, nonhaemolytic, susceptible to penicillin and contain a nonsense mutation in the chromosomal regulatory plcR gene (Okinaka et al., 2006). Examples of non-B. anthracis strains of B. cereus s.l. carrying pXO1-like plasmids have emerged recently and include B. cereus G9241 (Hoffmaster et al., 2004) and B. cereus P03BB102 (Hoffmaster et al., 2006), which have both been associated with an illness resembling inhalation anthrax, and contain the tripartite toxin complex. None of these species appear to harbour a pXO2-like plasmid; however, they do harbour capsule operons encoded on other elements (Hoffmaster et al., 2006). Atypical B. anthracis strains CI and CA were isolated from wild great apes in Côte d'Ivoire and Cameroon, respectively. Both these isolates harbour virulence plasmids closely related to those of B. anthracis, but a chromosomal background that appears to be closer to that of B. cereus or B. thuringiensis (Klee et al., 2006). Phenotypically, the CI and CA strains differ from classic B. anthracis in being motile, resistant to the γ-phage and for the CA strain resistant to penicillin G (Klee et al., 2006). The fact that neither contains the plcR nonsense mutation further indicates that they do not fit the current B. anthracis definitions. Finally, it should be noted that the genetic determinants associated with the B. cereus emetic disease (the cereulide biosynthesis gene cluster) may also be harboured by pXO1-like plasmids as exemplified by pCER270 (Rasko et al., 2007).

Many countries around the world regulate laboratories that work with security-sensitive dangerous pathogens including B. anthracis. This is done to prevent the illegal proliferation and use of such organisms in biological warfare programmes or bioterrorism and thus to ensure compliance with the Biological and Toxins Weapons Convention and the United Nations Security Council Resolution 1540. A major weakness of national biosecurity programmes may be the use of lists of species to specify which organisms are controlled (Casadevall & Relman, 2010). The above-mentioned examples of non-B. anthracis strains causing anthrax-like disease illustrate the insufficiency of the species classification within the B. cereus s.l. to describe and foresee their virulence profile, and thus their potential utility in biological weapons. Species classification within the B. cereus group is not merely an academic issue, as it may have a direct influence on the regulation of dangerous pathogens in biosecurity programmes.

In order to gain a more complete picture of the natural existence of non-anthracis B. cereus group bacteria with genotypic or phenotypic similarities to B. anthracis, more field investigations are required. Results from recent studies have shown that clinical and environmental non-B. anthracis isolates of B. cereus group bacteria may often harbour pXO1-like and/or pXO2-like plasmids and that these plasmids appear to have varying backbone structures as estimated by the sequence of the replication initiation protein genes repX and repA, respectively (Hu et al., 2009a, b). Focusing on the pXO1-like plasmids, it appears that those plasmids that have positively been associated with anthrax-like disease i.e. pXO1, pBCX01 (Hoffmaster et al., 2004) and p03BB102_179 (Hoffmaster et al., 2006), all share very similar repX gene sequences. Furthermore, all three plasmids have been isolated from the same geographical region (Table 1). The repX sequences of other pXO1-like plasmids appear to be phylogenetically distinct. The aim of this study is to investigate the presence and nucleotide sequence of B. anthracis pXO1-like plasmid replicons in environmental settings not known to contain B. anthracis. Emphasis is placed on determining the diversity of pXO1-like plasmid replicons. Both culture-dependent and culture-independent methods have been used to increase the likelihood of identifying the full diversity of pXO1-like replicons.

View this table:

Characteristics of pXO1-like plasmids

Plasmid namesHost speciesSize (bp)Virulence genes (lef, pag and cya)RsaI recognition site in 573-bp RepX fragmentIsolation environment/locationReferences
pXO1B. anthracis Ames ‘Ancestor’ (A2084)181 677YesYesCow (TX, 1981)Ravel et al. (2009)
pBCXO1B. cereus G9241190 861YesYesBlood sample (LA, 1994)Hoffmaster et al. (2004)
P03BB102_179B. cereus 03BB102179 680YesYesBlood sample (TX, 2003)Hoffmaster et al. (2006)
pPER272B. cereus AH818272 145NoNoRoot canal, apical periodontitis (Curitiba, Brazil, 1995)Helgason et al.. (2000a)
pH308197_258B. cereus H3081.97258 484NoNoFood outbreak linked to rice (TX, 1997)Hoffmaster et al. (2008)
pBc239B. cereus Q1239 246NoNoDeep-subsurface oil reservoir (Daqing, China)Xiong et al. (2009)
pCER270B. cereus AH187 (F4810/72)270 082NoNoVomit, emesis (London, UK, 1972)Ehling-Schulz et al. (2005)
pBC10987B. cereus ATCC 10987208 369NoNoDairy cheese (Canada, 1930)Rasko et al. (2004)

Materials and methods

Extraction of total community DNA from wastewater and soils

A wastewater sample was obtained in March 2009 from the influent of Lynetten municipal wastewater treatment plant, Copenhagen, Denmark. This facility receives wastewater from the Copenhagen area, covering 76 km2 with a population of approximately 535 000 citizens. Organic material was harvested from raw wastewater by centrifugation (5000 g for 5 min) and subsequent removal of the supernatant. Total community DNA was extracted from approximately 0.5 g of this organic material using the PowerSoil DNA kit (Mo Bio Laboratories Inc., Carlsbad, CA) as recommended by the supplier. Furthermore, soil was collected from the shoreline of eight natural and two artificial lakes in north-eastern Zealand, Denmark, in August 2009. Total community DNA was extracted from 0.5 g of these samples as described above. The extraction efficiency of DNA from the wastewater and soil matrices was determined by agarose gel electrophoresis.

Primers and primer design

A blast search (Altschul et al., 1997) was performed using the repX gene from plasmid pXO1 (B. anthracis Ames ancestor strain) as a query sequence. At the time, this search led to the identification of five other related plasmids, namely pBCX01, pPER272, pH308197_258, pBC10987 and pCER270, harbouring a gene of 1308 bp with high homology and full coverage to that of the pXO1 repX gene. The nucleotide sequences of all the above repX-like genes were retrieved from GenBank and a multiple alignment was performed with the pXO1 repX gene using cls sequence viewer software version 5.1.1 (CLC bio A/S). Based on this alignment, a pair of PCR primers was designed in the N-terminal end of the gene in regions with 100% nucleotide homology. The primers repX_Fw (5′-GAAAGCCAAGGAA ATATTAGTTTGAA-3′) and repX_rev (5′-AGCACTTGGAT TTTCCTCTTCA-3′) are expected to generate PCR products of 573 bp.

Using the same procedure as described above, a pair of primers was designed to amplify a 693-bp fragment of the repA gene from plasmid pXO2 also found in B. anthracis Ames ancestor strain. The resulting degenerate primers pXO2_fw (5′-MAACAAACAAARGCAGGGCG-3′) and pXO2_rev (5′-TTAACCAGCTGGYAGCCCTTC-3′) were designed based on a nucleotide multiple alignment of repA and homologous genes from plasmids pAW63, pBWB402, pBT9727 and pBMB165. The CYA-1F/1R and LEF-1F/1R primer pairs (Jackson et al., 1998) were used to detect the presence of the cya and lef anthrax toxin genes, respectively.

PCR amplification

The PCR mixture for amplification of pXO1- and pXO2-like plasmid replicon fragments contained 1 U Phusion DNA polymerase (Finnzymes), 200 μM each dNTPs, 0.5 μM of each primer, 1 × Phusion HF buffer provided by the supplier and 1 μL undiluted community DNA or 2 μL crude extract of specific isolates as a template in a total volume of 50 μL. Samples were denatured for 30 s at 98 °C, followed by 38 cycles of 98 °C for 10 s, 60 °C for 30 s and 72 °C for 30 s and a final extension at 72 °C for 5 min. Water and extracted genomic DNA from Bacillus subtilis was used as a negative control and extracted genomic DNA from B. anthracis Sterne (pXO1) and B. thuringiensis ssp. kurstaki HD73 (pAW63) were used as positive controls (Jackson et al., 1998; Wilcks et al., 1998). Crude genomic DNA was extracted from colonies grown on agar plates by boiling lysate using the PrepMan® Ultra Sample Preparation Reagent as described by the supplier (Applied Biosystems). Visualization of the PCR amplification products was achieved by agarose gel electrophoresis, followed by ethidium bromide staining. In general, PCR amplification products were purified directly using the QIAquick® PCR purification kit (Qiagen). In cases where visualization of the PCR product indicated nonspecific amplification together with the expected band, the latter was extracted and purified from the agarose gel using the QIAquick gel extraction kit (Qiagen).

Restriction enzyme analysis

Restriction enzyme analysis, using RsaI (New England BioLabs), was performed on the amplification products obtained following PCR with the pXO1 primers on community DNA from the wastewater and soil environments. Digestion was performed with approximately 1 μg of amplification product and 15 U of enzyme for 3 h at 37 °C. Following heat inactivation of the enzyme, the mixture was resolved by gel electrophoresis and banding patterns were visualized by ethidium bromide staining. The restriction enzyme RsaI has a 5′-GTAC-3′ recognition site, which is present as a single copy in the homologous 573-bp repX fragments of plasmid pXO1 and the very closely related pBCXO1 and p03BB102_179 plasmids. These three plasmids also contain the lef, pag and cya virulence genes characteristic of anthrax and have all been associated with anthrax-like disease or severe pneumonia. Digestion of repX fragments from these plasmids is expected to result in two fragments of 188 and 385 bp, respectively. The target site of RsaI is absent from the repX fragments of all other currently known pXO1-like plasmids including pBc239, pCER270, pBC10987, pPER272 and pH308197_258 (Table 1).

Cloning and sequencing of repX gene fragments

The purified PCR amplification product obtained using wastewater community DNA as a template was inserted into the pCR®4Blunt-TOPO® vector using the Zero Blunt® TOPO® PCR cloning kit for sequencing (Invitrogen Ltd, UK) following the recommendations provided by the supplier. The vector was subsequently transformed into One Shot® chemically competent Escherichia coli cells. Positive selection for transformants was achieved by plating the cells on Luria–Bertani (LB) agar plates containing 50 μg mL−1 kanamycin and incubating overnight at 37 °C. Colonies were restreaked on fresh agar plates and cultured in 5 mL LB broth (both containing kanamycin). Vector-plasmids were extracted from the cells using the QIAprep® Spin Miniprep Kit (Qiagen). In some cases, restriction enzyme analysis using EcoRI was performed in order to verify the cloning efficiency.

Sequencing of the PCR-amplification products contained in the vector-plasmid was performed in both directions by Macrogen Inc., Seoul, Korea, using the universal T3 and T7 primers. In total, 45 clones were sequenced.

Isolation and screening of B. cereus-like bacteria for the repX gene

Suitable dilutions of wastewater were spread-plated on mannitol–egg yolk–polymyxin (MYP) agar plates (Oxoid A/S, Denmark) and incubated overnight at 30 °C aerobically. A total of 28 colonies with a morphology equivalent to that reported and observed for B. cereus (light rose colonies with opaque halo) were restreaked on MYP agar and subsequently crude DNA was extracted using the PrepMan® Ultra Sample Preparation Reagent. Additionally, 23 colonies with a morphology similar to that observed for B. anthracis Sterne (light rose colony without halo) were restreaked and DNA extracted. The PCR set-up and detection of repX-like replicons in these isolates were as described above. PCR products from positive isolates were purified and sequenced in both directions using the same primers for the sequencing reaction as for the amplification. Strains positive for repX were also screened for the pXO2-like repA gene.

Phylogenetic analysis

DNA sequence alignments of the repX-like and repA-like fragments found in this study and the homologous fragments retrieved from GenBank were performed using the cls sequence viewer software. Neighbour-joining trees based on these alignments were constructed using the geneious basic 4.5.4 software (Biomatters Ltd). Distances were calculated using the Jukes–Cantor genetic distance method and bootstrap values from 1000 replicates were acquired.

Results and discussion

Plasmid pXO1-like replicons found in wastewater and soil samples

Total community DNA was successfully isolated from the municipal wastewater and all 10 soil samples. In all cases, PCR analysis of these samples, using the pXO1 primer pair, resulted in amplification of a single band of the expected size corresponding to the 573-bp fragment of repX from pXO1. This indicated that pXO1-like replicons are present and easily detectable in environmental settings in Denmark, including municipal wastewater and soil from the shoreline of both natural and artificial lakes. The purified amplification products from these samples were subjected to restriction enzyme analysis in order to carry out a preliminary evaluation of the diversity of pXO1-like replicons in each environmental location. Restriction enzyme analysis revealed that only the amplification product from the wastewater sample appeared to be digested by RsaI (Fig. 1). It is, however, possible that the fraction of RsaI digestible fragments that were amplified from the soil samples was below the detection level of the gel electrophoresis analysis. The partial digestion of the fragment amplified from the wastewater indicated that both restriction-positive and restriction-negative fragments were present in the sample (Fig. 1). For this reason, the wastewater sample was selected for a more extensive study of the presence and diversity of pXO1-like replicons.


Restriction enzyme digestion with RsaI on the 573-bp PCR amplification products obtained from wastewater community DNA (lane 2), DNA extracted from Bacillus anthracis Sterne (lane 3), strain 4.1 (lane 4) and community DNA extracted from the 10 lakes (lanes 5–14). Lanes 1 and 15 show 100-bp markers (New England Biolabs).

Two different approaches were used to explore the diversity of pXO1-like plasmid replicons in the wastewater sample: (1) a culture-independent methodology, in which the repX amplicons obtained from the PCR were cloned individually and then sequenced, and (2) a culture-dependent methodology, in which extracted DNA from a total of 51 isolated B. cereus group bacteria from the same environment was subjected to PCR screening, followed by sequencing of repX-positive strains. Screening of the wastewater isolates resulted in the identification of nine repX-positive isolates, of which all, except one, namely strain 3.3, formed light rose colonies with an opaque halo when grown on MYP agar. The incidence of repX-positive B. cereus group bacteria in random wastewater isolates found in this study is comparable to those previously reported for 40 isolates from soil collected within a forest (25%) and 649 other environmental isolates (7%) (Hu et al., 2009a, b) and indicates that repX replicons, and thereby most likely pXO1-like plasmids, are generally abundant in the environment.

Three subgroups of pXO1-like replicons found in municipal wastewater

The nucleotide sequences of repX fragments obtained from both the culture-independent and the culture-dependent approach were all trimmed for primer sequence and aligned with the corresponding repX fragments from the eight currently recognized pXO1-like plasmids found in GenBank. The phylogenetic relationship of all repX fragments was visualized by construction of an unrooted neighbour-joining tree based on this alignment (Fig. 2). Overall, there appeared to be three distinct subgroups of pXO1-like replicons present in the wastewater environment. Two of the three subgroups (named α and β) contained several replicons from previously isolated pXO1-like plasmids, together with replicons identified in this study by both culture-independent and culture-dependent methods. The third subgroup of replicons (γ) contained replicons identified in this study by both culture-independent and culture-dependent methods, but was not represented by any replicons in the GenBank database. The apparent existence of distinct subgroups of pXO1-like replicons is consistent with previous reports on plasmid replication genes of other plasmid groups, such as the promiscuous broad-host-range IncP-1 plasmids (Bahl et al., 2009) and the IncP-9 plasmids (Sevastsyanovich et al., 2008). Moreover, it has been shown that such subgrouping of plasmids may be supported by a number of other plasmid backbone genes apart from those directly involved in replication (Vedler et al., 2004; Bahl et al., 2007). The proposed α subgroup of pXO1-like plasmid replicons contains plasmids pXO1, pBCXO1 and P03B102_179, which have all been implicated in anthrax-like disease or severe pneumonia. This subgroup also contains repX fragments from strains 3.3 and 4.2 isolated in this study from the wastewater sample, as well as several identical copies of a slightly different replicon fragment (e.g. 2A5) retrieved from the community DNA (Fig. 2). The 2A5 fragment exhibits very high homology to the pXO1 fragment, with 98% homology on the nucleotide level and only one predicted amino acid difference within the expected 174 aa peptide.


Neighbour-joining tree of the amplified repX gene (pXO1-like plasmids) fragments obtained from wastewater community DNA in relation to known homologue sequences from GenBank, based on a multiple alignment (cls sequence viewer software) of the nucleotide sequences. Fragments 3.3 (1), 3.6 (1), 4.1 (3), 4.2 (1) and 4.4 (3) originated from bacterial isolates and fragments 1A7 (22), 1A3 (4) and 2A5 (19) were amplified from community DNA. Numbers in parentheses show the number of identical fragments found. Fragments from known plasmids are indicated by the plasmid names. The distances between the fragments are indicated by the bar.

Two isolated B. cereus group strains harbour both a pXO1-like and a pXO2-like replicon

Interestingly, strains 3.3 and 4.2 were the only repX-positive strains isolated in this study shown to also harbour a pXO2-like replicon. Sequencing of the two pXO2-like replicon fragments revealed almost identical nucleotide sequences of 693 bp. The phylogenetic tree (Fig. 3) constructed based on a multiple alignment of these two fragments and the homologous fragment of known pXO2-like plasmid replicons retrieved from GenBank showed that the 3.3 and 4.2 pXO2-like replicons grouped with those of plasmids pBMB165, pBT9727 and pBWB402 in Group Y, as defined previously (Hu et al., 2009a), that is distinct from Group X, which contains plasmids pXO2 and pAW63. Strains 3.3 and 4.2 did not contain the cya and lef genes, which strongly indicates that they do not contain the B. anthracis pXO1-associated 44.8-kb pathogenicity island and thus would not be expected to cause anthrax-like disease. This is, to our knowledge, the first demonstration of B. cereus group isolates, apart from B. anthracis itself, that carry both a pXO1-like replicon with very high homology to that of pXO1 (subgroup α) and a pXO2-like replicon. It is generally acknowledged that what sets B. anthracis apart from other B. cereus group strains is largely the fact that it harbours two specific plasmids, namely pXO1 and pXO2. Results from the present study as well as other recent findings of both pXO1-like and pXO2-like plasmids in environmental and clinical settings suggest that the backbone genes, for example replicons, of these plasmids are abundant in various environmental settings, not suspected of containing B. anthracis and may coreside in the same host strain. To date, only very little is known about the associated accessory elements of the pXO1-like and pXO2-like plasmids; however, it is clear that not all contain the virulence genes of wild-type pXO1 and pXO2. Future studies are needed to gain more insight into and knowledge regarding the functional diversity of these naturally occurring plasmid groups.


Neighbour-joining tree of the amplified repA gene (pXO2-like plasmids) fragments obtained from strains 3.3 and 4.2 in relation to known homologous sequences from GenBank, based on a multiple alignment (cls sequence viewer software) of the nucleotide sequences. Distances between the fragments are indicated by the bar.


We thank Andrea Wilcks for providing the pAW63 plasmid and for scientific advice in connection with this study. John-Erik Stig Hansen and Nina Steenhard are thanked for critically reading the manuscript.


  • Editor: Kornelia Smalla


View Abstract