OUP user menu

Intriguing diversity of Bacillus anthracis in eastern Poland – the molecular echoes of the past outbreaks

Rafał Gierczyński, Stanisław Kałużewski, Alexander Rakin, Marek Jagielski, Aleksandra Zasada, Antoni Jakubczak, Bogna Borkowska-Opacka, Waldemar Rastawicki
DOI: http://dx.doi.org/10.1016/j.femsle.2004.08.038 235-240 First published online: 1 October 2004


The multiple locus VNTRs analysis (MLVA) revealed the presence of five genotypes in a group of 10 Bacillus anthracis isolates from epidemiologically unrelated cases of bovine-anthrax in eastern Poland. Eight tested isolates possessed the pagA and capB genes indicating the presence of both virulence plasmids, while two isolates revealed only pagA and lacked pXO2. The MLVA and DNA sequence analysis indicated that seven tested isolates represent four novel genotypes. Five tested strains revealed a unique 144 bp vrrB2 variant as well as 220 bp variant of vrrB1, implying the relatedness to the lineage B2. Consequently, we propose establishing of novel B2 strains sub-lineage. Multiple anthrax outbreaks, which took place in Poland several decades ago were proposed as a cause of intriguing diversity of B. anthracis observed in this study.

  • Bacillus anthracis
  • Genotyping
  • MLVA

1 Introduction

Bacillus anthracis is the causative agent of anthrax. The identification of anthrax agents by biochemical and molecular techniques is difficult due to a high similarity of B. anthracis to B. cereus and B. thuringiensis[1,2]. Thus, molecular tools targeted directly to genes essential for virulence of B. anthracis, which are encoded on plasmids pXO1 and pXO2, were developed. The pXO1 (174 kbp) contains genes lef, pag and cya encoding lethal factor, protective antigen, and edema factor, respectively, while pXO2 (95 kbp) bears capA, capB, and capC, which encode capsule production [1].

Bacillus anthracis is undoubtedly one of the most monomorphous species known. Standard molecular typing methods such as pulsed-field gel electrophoresis (PFGE), PCR amplification of 16S-23S rRNA, as well as gyrA gyrB intergenic spacer regions (ISR) were unable to effectively distinguish B. anthracis isolates [2]. In 1996, the vrrA open reading frame (ORF) containing a variable number (2–6 copies) of 12-bp tandem repeat (VNTR) has been described [3]. The vrrA based typing was capable to diversify anthrax isolates into five genotypes and, in contrast to amplified fragment length polyporphism (AFLP), required only a simple PCR apparatus and agarose gel electrophoresis for typing [4,5]. Similarly, five genotypes were observed for pagA sequencing analyses and long range repetitive elements polymorphism-PCR (LR-REP-PCR) [6,7].

The current standard in genotyping of B. anthracis is the multiple-locus VNTR analysis (MLVA) that was developed in 2000 [8]. This method is based on parallel analyses of VNTRs in eight marker loci: vrrA, vrrB1, vrrB2, vrrC1, vrrC2, CG3, pXO1-aat and pXO2-at. The two last loci are localised on virulence plasmids. MLVA was capable to distinguish 89 genotypes among 426 B. anthracis strains and few novel genotypes have been reported recently [9].

Since MLVA is the determinative genotyping method, data from different laboratories may be easily compared. Thus, this technique was used for phylogenetic studies on B. anthracis isolates from Kruger National Park and for determination of diversity among 50 French isolates, as well as for epidemiological investigations of intended anthrax applications in Japan and in the USA [9,10,11,12]. Examinations of multiple samples collected during above investigations revealed genetic stability of MLVA markers, as well as high reproducibility of the method [12]. The only disadvantage limiting application of MLVA to selected laboratories is requirement of expensive equipment for separation and detection of fluorescent-labelled MLVA amplicons [7]. To overcome this limitation, a typical sequencing apparatus for DNA separation in denaturing conditions combined with DNA silver-staining was used.

Despite general minor diversity among B. anthracis isolates, the MLVA enabled identification of six major clones [8]. The greatest dissimilarity has been observed among the strains representing two branches named A and B. Type A strains are present world-wide, thus are responsible for most epidemics and outbreaks, while type B strains are generally restricted to Africa [8,10]. In contrast to African B strains, grouped in cluster B1 that comprises nine genotypes, only two genotypes (79 and 80) have been assigned to cluster B2. The majority of strains belonging to this group were collected in France [9]. Single isolates originated from Slovenia and Croatia [8]. All B2 strains reported to date possessed specific vrrB1 variant of 220 bp [8,9].

In this paper, we report two novel B2 genotypes, which differ from genotypes 79 and 80 by two loci but contain the 220 bp vrrB1 marker. In addition, we describe two novel A1.a genotypes, as well as an atypical pXO2-deficient B. anthracis strain exhibiting haemolytic activity isolated from bovine carcass with clinically and laboratory confirmed anthrax. We relate observed diversity of B. anthracis isolates to massive outbreaks, which took place in eastern and southeastern Poland during past decades [13].

2 Materials and methods

2.1 Bacterial strains

Ten strains of B. anthracis isolated from cattle died of anthrax and vaccine strain Sterne 34F2 were investigated (Table 1). All strains were isolated in eastern Poland in provinces of Podlaskie (Northeast) and Podkarpackie (Southeast) and were re-identified as B. anthracis by standard tests recommended by WHO: colony morphology, motility, haemolytic activity, sensitivity to the gamma phage and penicillin [14].

View this table:
Table 1

Place and date of Bacillus anthracis isolation and number of persons with cutaneous anthrax developed as a result of direct contact with infected animal

StrainIsolation detailsNumber of associated human anthrax
YearMaterialRegion of PolandPlace of isolation
BL31947No dataSoutheastJasloND
BL101993No data0
34F2Sterne vaccine strainaNot applicable
  • a Strains isolated from three separate samples of the same animal.

2.2 DNA preparation

The template DNA was isolated from 1 ml of fresh cultures in tryptic soy broth (TSB). After 4 h incubation at 37 °C with aeration (300 rpm), the culture was centrifuged. Bacterial pellet was mixed with 40 μl of 10 mg/ml lysosyme and incubated for 30 min at 37 °C. Then, 160 μl of lysis buffer (10mM Tris–HCl, pH 8.0, 10mM EDTA and 1% Triton X-100) containing 2 mg/ml proteinase-K was added and tubes were placed at 52 °C for at least 1 h or until the suspension becomes transparent. Next 0.8 ml of 10 mM Tris–HCl, pH 8.0, was added and tubes were heated at 95 °C for 30 min. Five microliters of DNA template was used per 50 μl of the PCR mixture.

2.3 Genome based identification

Detection of pagA and capB genes by PCR was performed using primers and cycling conditions described by Jackson et al. [15]. Additionally, the sap gene encoding S-layer protein and SG 749 chromosomal marker were tested [14,16]. The 450 bp part of SG 749 was subjected to a single strand conformation polymorphism analysis by multitemperature-SSCP technique as reported previously [17,18]. Moreover, all isolates were tested for the presence of B. anthracis specific chromosomal sequences in accordance to multiplex-PCR protocol proposed by Radnedge et al. [19].


The multiple locus VNTR analysis was performed in accordance to the method descried by Keim et al. [8] with slight modifications. Each MLVA marker was amplified in separate tube using unlabeled primers described by Keim et al. and 1.5 U of Taq DNA polymerase (Polgen, Poland) per 50 μl of reaction mixture [8]. PCR was performed in Mastercycler 5330 (Eppendorf) as follows: 3 min at 94 °C and 35 cycles 30 s at 94 °C, 45 s at 60 °C, 45 s at 72 °C completed by 3 min at 72 °C. Two microliters of the PCR product was mixed with 18 μl of denaturing buffer (50% (v/v) deionised formamide; 30% (v/v) of saturated urea; 19.5% (v/v) glycerol; 0.025% (w/v) xylene cyjanon FF and 0.025% (w/v) bromophenol blue), heated at 95 °C for 3 min and then immediately cooled by placing into a frozen aluminium rack. Next, 4 μl of the denatured sample was loaded onto 8% polyacrylamide gel (K4 acrylamide and K4 bis-acrylamide mixture 19:1, Applichem, Germany) containing 7 M urea. The mixture in a ratio of 29:1 was used to prepare gels for separation of vrrC1 and vrrC2 amplicons. Electrophoresis was conducted in 0.5×concentrated TBE (45 mM Tris–boric acid and 1 mM EDTA, pH 8.0) under constant power of 45 W, using 40 cm long plates and vertical electrophoretic chamber (Kucharczyk TE, Poland). DNA bands were visualised by silver staining (Silver Stain Kit, Kucharczyk TE), and scanned with density of 300 dots per inch. The digitised gels were analysed by GelScan software version 1.3 (Kucharczyk TE) and size of each MLVA marker was determined using 20 bp step DNA ladder (Sigma–Aldrich Co.) as a standard for calculations. Results obtained by GelScan were compared to standard genotypes recommended by Keim et al. [8]. The cluster analysis was performed using average linkage agglomeration method of WinSTAT software version 2001.1. Although the isolates lacking pXO2 were subjected to genotyping by seven MLVA markers, the obtained data were not included in cluster analysis due to tested markers number incompatibility.

2.5 DNA sequencing

Selected amplicons of vrrA and vrrB2 MLVA markers were subjected for sequence analysis using automated fluorescent DNA sequencer 377 and BigDye Terminator v3.1 (Applied Biosystems, USA) in accordance to the manufacturers instructions. Each strand of the analysed amplicons was sequenced. The final sequence was based on data from both strands. The sequences of selected MLVA markers were compared to other known sequences deposited in GenBank database using BLASTN software utility (National Center for Biotechnology Information, USA).

3 Results

3.1 Identification and virulence characterisation

All tested strains were non-motile, gamma phage and penicillin-sensitive. Most of them formed white “bee-eye” textured and medusa-like shaped colonies on Columbia blood agar. Only strain marked as BL8G formed grey colonies and revealed haemolytic activity that increased after 24 h incubation in room temperature (Table 2). Strain BL8 that formed white colonies without haemolysis was co-isolated with BL8G from the same blood sample (Table 1). Repeated passages on Columbia blood agar indicated that both variants were stable. Therefore, each variant was considered as a separate strain for further studies.

View this table:
Table 2

Results of selected identification tests and MLVA genotyping

StrainIdentification testsMLVA genotypeMLVA markers
Colour of colonyaHaemolytic activitycapBvrrAvrrB1vrrB2vrrC1VrrC2CG3pX01-aatpX02-at
BL3White+Novel A1.a –1289229162613604153132135
BL6White+Novel A1.a –2325229162613604153129137
BL1White+Novel B2-1301220144583532158129135
BL5White+Novel B2-1301220144583532158129135
BL10White+Novel B2-2301220144583532158129133
BL11White+Novel B2-2301220144583532158129133
BL12White+Novel B2-2301220144583532158129133
BL8WhiteRelated to 61313229162583532158129
BL8GGrey+Related to 61313229162583532158129
Sterne 34F2WhiteRelated to 61313229162583532158129
  • a Detected on Columbia agar.

The pagA gene was detected in all tested strains, indicating the presence of virulence plasmid pXO1. The presence of gene capB gene located on pXO2 was observed in all tested strains except BL8, BL8G and Sterne 34F2 (Table 2). All the isolates possessed the sap gene and revealed the SG-749 marker as well as B. anthracis specific SSCP-pattern of SG-450 and were positive in multiplex PCR with primers specific for B. anthracis chromosomal DNA (data not shown).

3.2 MLVA

Initially, to validate MLVA genotyping method used in this study two DNA step ladders with 20 and 50 bp step were tested. The molecular size calculation method of GelScan software was configured using MLVA markers of Sterne 34F2. The most reliable results were obtained with 20-bp step DNA. Three MLVA markers size groups were defined to make possible the parallel separation of different markers in the same gel with optimal migration distance. The first group consisted of vrrB2, CG3, pXO1-aat and pXO2-at. The second group covered vrrA and vrrB1, while third one was limited to vrrC1 and vrrC2. The optimal time of electrophoresis was determined as 3, 4 and 5 h, respectively, for each above groups. In case of size asymmetry between the DNA strands of tested amplicons, the slowly migrating strand was used for molecular size calculations (see Fig. 1).

Figure 1

The results of electrophoretic separation of the selected MLVA alleles observed in B. anthracis strains representing the novel genotypes reported in this study. Each numbered line shows a single allele. The allele size in base pairs is shown in brackets. Lines: 1 (229); 2 (220); 3 (144); 4 (162); 5 (129); 6 (132); 7 (133); 8 (135) and 9 (137). M – 20 bp step marker. In case of ssDNA size asymmetry, arrows indicated base strand used for calculation.

The final results of MLVA are shown in Table 2. The vrrB2 variant of size 144-bp and vrrA amplicons were subjected for sequence analysis. The BlastN analysis of data obtained from the sequence revealed 100% identity of the vrrB2 sequence reported in this study to 144-bp vrrB2 variant described by Schupp et al. [20]. The identity of vrrA amplicons was also confirmed and no nucleotide substitutions were observed.

The results of the cluster analysis, including few selected MLVA genotypes reported to date, are shown in Fig. 2. Strains BL1 and BL5 represented novel B2-1 genotype. Similarly, strains BL10, BL11 and BL12 represented the novel B2-2 genotype closely related to B2-1. Together, these two novel B2 genotypes formed separate branch of the B2 lineage. The A1a-2 novel genotype represented by strain BL3 was closely related to genotype 18, whereas other novel genotype A1a-1 represented by strain BL6 indicated relatively low similarity to genotypes 13, 15, 18 and A1a-2.

Figure 2

The eight VNTR marker loci MLVA-based dendrogram of genotypes represented by tested strains of B. anthracis (in bold) and selected closely related genotypes described to date. The pXO2-deficient strains are not included due to the incomplete MLVA pattern.

4 Discussion

The results of both phenotypic and molecular identification tests indicated that all tested isolates were undoubtedly strains of B. anthracis and possessed at least pXO1 virulence plasmid. Despite relatively low amount of the tested strains, five MLVA genotypes were observed. Surprisingly, the genotypes reported here were dissimilar to genotype 15 formerly described by Keim et al. [8] for Polish strain 100004 of B. anthracis. These findings imply that at least six genotypes of B. anthracis are present in Poland.

Interestingly, we observed an atypical vrrB2 variant of 144 bp in five tested B. anthracis isolates originated from the same area but with few years interval. Although sequence of the variant had been deposited in GenBank database, it was not reported in any MLVA genotype to date. Moreover, the isolates possessing the rare vrrB2 variant revealed the presence of vrrB1 variant of 220-bp size, which was specific for B2 lineage. The only difference between two described to date B2 genotype 79 and 80 is only a single 2 bp alteration in locus pXO2-at. The novel B2 genotypes reported here differ from the above the genotypes in locus vrrB2. Moreover, the B2-2 genotype shows 133-bp variant of pXO2-at that was never reported for any B2 strain. Therefore, the above results do not only show the presence of B2 strains in Poland but also suggest occurrence of major evolutionary changes in B2 lineage in this geographical region. Consequently, we propose establishing of two novel B2 genotypes, which due to the presence of the rare vrrB2 variant might even represent a novel B2 sub-lineage.

When compared to the results of Fouet et al. [9], who genotyped 50 French B. anthracis isolates, the surprisingly high diversity of vrrA types was detected during our study. We observed four of five known vrrA genotypes, whereas Fouet et al. found only two vrrA variants among isolates originated all over France from 1982 to 1999. Hence, anthrax isolates from two different outbreaks, one in Pyrenees and second in Alps, represented the same vrrA type [21]. Because the five MLVA genotypes reported in this study revealed four variants of vrrA and two variants of the other seven markers including rare vrrB2 variant, therefore, the genotype diversity described here may be considered as significant.

In addition, we reported here the pXO2-deficient B. anthracis strain with haemolytic activity that was undoubtedly isolated from an animal died from anthrax. Both non-haemolytic (white) and haemolytic (grey) colonies were also consistently isolated from other separate specimens like the bloody exertion from nostrils and liver of the animal (data not included). Since these strains could not be separated by MLVA from the vaccine strain Sterne 34F2, the occurrence of vaccination-associated anthrax might be suspected. On the other hand, the Regional Veterinary Inspection possesses no data confirming administration of the living anthrax vaccine to the died animal.

At first, we were concerned by observed diversity of MLVA genotypes because B. anthracis is considered as extremely slowly evolving species. Keim et al. [8] observed only single VNTR mutation in loci vrrA of Ames strain after 100 000 passages. The above facts together suggest that the observed diversity is the result of genetic mutations, which could happen after large number of generations. Before beginning of our study, the tested isolates were stored for years in different strain collections practically without restitution, that assured low number of passages and the originality of the reported genotypes. Therefore, the observed diversity of the MLVA genotypes, even caused by a mutation in a single locus, seems to be the result of intensive proliferation of B. anthracis during frequent anthrax outbreaks in probable endemic region rather than the acquisition of a spontaneous laboratory mutation.

To confirm this hypothesis, we analysed epidemiological data regarding frequency and number of anthrax outbreaks in Poland during the last century. Although the anthrax cases are presently rare in Poland, massive outbreaks took place in eastern and southeastern regions of the country during the first half of the 20th century and probably for many decades before. In this area, anthrax was diagnosed in about ninety thousands of animals in almost four and half thousands huts from 1923 to 1927 [13]. Despite significant reduction of anthrax cases, due to extensive vaccination of cattle, the annual outbreaks persisted in southeastern region in the period 1947–1957. In our opinion, the above epidemiological findings may explain the reported here high diversity of MLVA genotypes of B. anthracis isolated in Poland.


We thank Dr. Arciuch and Dr. Szymajda from WIHE, Pulawy, Poland for support in five strains of Bacillus anthracis.


  • Editor: R.Y.C. Lo


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
View Abstract