Identification of strain specific markers in Bacillus anthracis by random amplification of polymorphic DNA

Abstract

Classification and differentiation of Bacillus anthracis isolates by genetic markers play an important role in anthrax research. We used a PCR based method – Random Amplification of Polymorphic DNA (RAPD) – to identify genetic markers in B. anthracis strains. Twenty-five differential genetic markers were identified which divided the strains into five different groups. Three selected RAPD-markers were cloned and sequenced. The five RAPD-derived genotypes could be defined by integration of these three markers. This system offers a simple non-expensive method to classify B. anthracis strains in laboratories involved in the research of this bacterium.

Bacillus anthracis, Classification, RAPD, DNA, Typing

1 Introduction

The spore forming Bacillus anthracis microorganism is the cause of Anthrax, and a major candidate to be used as a non-conventional weapon in war and acts of bio-terror. Methods that will facilitate classification of B. anthracis isolates are important for the identification of the bacterium in case of hostile dissemination and in daily research. Since B. anthracis is monomorphic, a sensitive and accurate classification method must be developed, that will be able to identify the possible source of a specific isolate. This classification method must depend on large numbers of genetic markers that will distinguish between two isolates of the same strain.

Keim et al. [1] used an AFLP (amplified fragment length polymorphism) technique, to identify 31 markers in 79 B. anthracis strains/isolates and divided the strains into phylogenetic groups. Based on these AFLP markers and the previously identified vrrA[2], Keim et al. [3] proposed an eight variable-number tandem repeats (VNTR) marker classification of B. anthracis strains. This Multiple-Locus VNTR Analysis (MLVA) method is based on fluorescent-labeled PCR products that are separated simultaneously on a sequencing gel. A specialized instrument defines the precise size of each of the PCR products and the combination of the entire products sizes determines the strain genotype [3]. MLVA became the gold standard and was used to determine the genotype of B. anthracis isolates from all over the world (reviewed in [4,,6]), but the need for labeled primers and special facilities render it unavailable for small labs and complicated for daily R&D use.

Additional VNTR markers were identified by using bio-informatics tools to locate tandem repeats, that were shown in other species to be targets for rearrangements (reviewed in [7,8]). Le Fleche et al. [9] identified 14 tandem repeats in B. anthracis, and in combination with 6 of the MLVA markers, designed a phylogenetic tree that overall was similar to that proposed by Keim et al. [3] but had an additional genetic sub-group.

In an attempt to develop a simple classification system that will allow the typing of B. anthracis isolates without the need for specific high cost equipment, we adapted the Random Amplified Polymorphic DNA (RAPD) method [10]. RAPD is a PCR based method, using short (10 nt) primers with arbitrary sequences. Each PCR reaction is performed with a single primer that anneals to different binding sites of the genome. In most cases the amplification reaction generates several DNA bands that are specific for that organism. RAPD was used to differentiate between different strains of the same species in animals, plants and bacteria and some preliminary analysis of B. anthracis[11,12].

2 Materials and methods

2.1 Strains

Bacillus strains: Bacillus thuringiensis thuringiensis (Bacillus genetic stock center-BGSC 4042A), Bacillus thuringiensis israelensis, Bacillus cereus 569 (BGSC 6A3), and Bacillus licheniformis (BGSC 5A2) are from the Israel Institute for Biological Research (IIBR) collection. Bacillus subtilis, Bacillus cereus/thuringiensis and Bacillus mycoides are natural isolates from soil samples and identified by the BioLog system. B. anthracis strains used in this study are from the IIBR collection and are described in Table 1.

1

B. anthracis strains used in this study

Strain	Alternative strain name	pXO1/pXO2
Δ14185	Derived from ATCC14185 strain (V770-NP1-R)	−/−
ATCC14186	107-NP2-R1	+/−
ATCC6605		+/+
Vollum	ATCC14578	+/+
Sterne 10	34F2	+/−
BA7001	Derived from a Sterne strain (Anthrax spore vaccine) other then Sterne 10	−/−
Pasteur		−/+
BA8800		+/+

2.2 DNA preparation

Bacillus strains were cultured in Terrific broth [13] overnight at 37 °C with orbital agitation of 300 rpm. The pellet of 1 ml overnight culture was resuspended in 200 μl of 50 mM Tris–HCl, pH 8.0; 10 mM EDTA. Two hundred μl of lysis buffer (200 mM NaOH; 1% SDS) were added. The tube was inverted several times and incubated at 50 °C for 10-min. The DNA was recovered by sequential extractions with phenol (400 μl, Sigma, P4557), and chloroform (400 μl, Baker, 7018) and separated by a 10-min incubation at 50 °C. The aqueous phase was transferred to a new minifuge tube and the DNA was precipitated by adding an equal volume of isopropanol (Bio Lab laboratories, CP). The pellet was washed once with 70% ethanol (v/v) and resuspended in 100–200 μl of double distilled water (DDW). A dilution of 10⁻² was made in DDW and 5 μl (?2 ng) of the diluted DNA was used for the PCR reactions.

2.3 PCR reactions

All PCR reactions (25 μl) were performed in 1× PCR buffer (3.5 mM MgCl₂); dNTPs (0.2 mM each); 0.04 U/μl of TaKaRa Taq DNA polymerase (all from TaKaRa Bio Inc. R001A) and ?2 ng of DNA. The thermocycling program for the RAPD reaction was 94 °C for 30 s, 40 cycles of 94 °C for 1 min; 35 °C for 1 min; 72 °C for 2 min, and then one cycle of 72 °C for 5 min. The PCR products were separated on 1.1% agarose gel in 1× TBE buffer [13]. The general thermocycling program for the PCR reaction was 94 °C for 30 s followed by 40 cycles of 94 °C for 1 min; 55 °C for 30 s; 72 °C for 1 min, and then one cycle of 72 °C for 5 min. The PCR products were separated on 1.1% agarose gel using 1× TBE as running buffer.

Arbitrary primers for the RAPD reactions are available commercially from Operon Technologies Inc. The primers used to specifically amplify the AJ03 marker were (5′ to 3′): AJ035 – AGCACCTCGTTCATGCTCATAACGG and AJ036 – AGCACCTCGTCTACTTCATTTTGTGC; The AA03 marker were AA032 – TTAGCGCCCCCTTGCGTTCC and AA033 – TTAGCGCCCCTAGACCAATTGC: The AT07 marker were AT073 – CTCCTCAAATTACTAAAATGAAACC and AT074 – TTGGCATAGACGTATATTGCGGTCC.

2.4 Cloning and sequencing

The relevant PCR fragments were extracted from agarose–TAE gel using wizard PCR preps (Promega A7170) and cloned into pGEM-T easy vector (Promega A1380). The cloned PCR fragment was then sequenced using T7 and SP6 primers of the vector flanking the cloning site.

2.5 Marker frame identification

The B. anthracis DNA markers were characterized by sequence similarity searches (using BLAST) against the following databases: nr (NCBI) unfinished microbial genomes, draft genome sequences of diverse B. anthracis strains (TIGR), annotated B. anthracis Ames strain chromosomal ORFs [14]. Each marker analyzed for genome allocation, gene context and putative function of its product based on the data and procedures previously described by Ariel et al. [14].

3 Results and discussion

3.1 Differentiation between Bacillus species

RAPD is a well-established easy method used to classify closely related strains (e.g. [10,15,16]). We tested the ability of this method to distinguish between seven Bacillus species from our strain collection: B. anthracis B. thuringiensis thuringiensis, B. thuringiensis israeliensis, B. cereus 569, B. licheniformis and three strains of Bacillus that were isolated from soil samples and identified by the BioLog system as B. subtilis, B. cereus/thuringiensis and B. mycoides. DNAs from these strains were analyzed using RAPD PCR with randomly picked primers. The results of a representative experiment are shown in Fig. 1(a). As expected, the PCR reaction using the RAPD primer resulted, in most cases, in the formation of multi-band patterns. Although evolutionary B. anthracis, B. cereus B. thuringiensis and B. mycoides are considered to be closely related [17,,19], very little similarity if any could be detected between the fragment fingerprints that were obtain using RAPD PCR on DNA from these strains. In fact very few common DNA bands were obtained in two strains of the same species such as B. thuringiensis thuringiensis, B. thuringiensis israeliensis (Fig. 1(a) lanes 2 and 3). Thus the RAPD method is sensitive enough to distinguish between closely related species.

1

Specific differences between B. anthracis strains as appeared by RAPD reactions. (a) RAPD reaction (using K7 primer) of DNA extracted from different Bacillus strains. 1. B. antracis (ATCC14185), 2. B. thuringiensis thuringiensis, 3. B. thuringiensis israeliensis, 4. B. cereus 569, 5. B. licheniformis, 6. B. subtilis, 7. B. cereus/thuringiensis, 8. B. mycoides. (b) Deletion/insertion in specific band (marked in an arrow). (c) Appearance/disappearance of a specific DNA band in a group of strains (marked with arrow). The strains analyzed are 9. Δ14185, 10. Sterne (10), 11. BA7001 (Sterne), 12. ATCC14186, 13. Vollum, 14. ATCC6605, 15. Pasteur, 16. BA8800. The RAPD primers used (b) G06, (c) D11.

3.2 Comparison of B. anthracis strains

RAPD analysis of the B. anthracis strains was performed using primer kits A–Z, AA–AZ and BA–BH, a total of 1200 primers. We analyzed eight B. anthracis strains using RAPD primers (Operon Technologies Inc.) in a low stringency PCR reaction. Most of the primers used yielded several PCR products ranging in size from 300 bp to more then 3 kbp. As expected, all the examined strains had the same pattern of PCR products with the vast majority of the primers (data not shown). Because we were using relatively short primers (10 nt) and a low annealing temperature (35 °C), in most of the PCR reactions more than one product was formed, reaching in some cases more then 10 DNA bands. Differences caused by absence of bands in the range of 3 kbp and higher were overlooked unless a higher band was present indicating that the absence of the band was not due to DNA quality or enzyme processivity.

3.3 Classification of B. anthracis strains

Various modifications in the DNA pattern could indicate differences between B. anthracis strains. These modifications could be generated by deletion/insertion events that could result in increase or decrease in the DNA band size (Fig. 1 (b)) if the event occurred between the binding site of the primer, or band vanishing (Fig. 1 (c)) if the primer binding sight was removed or modified. A point mutation in the primer binding site could generate a mismatch that will result in a missing band or generate a new binding site resulting in appearance of a new DNA band.

We have identified a total of 25 markers, some of them were strain specific and others were specific to more than one strain. The specific markers and their strain specificity are described in Table 2.

2

Summary of the markers identified by RAPD

Marker	14185	Sterne (10)	7001 (Sterne)	14186	Vollum	6605	Pasteur	8800
R13	D	D	D	D
AY09-1	N	N	N	N
H19	D	D	D	D
AA03						D	D	D
D11					N	N	N	N
B05-06					N	N	N	N
AU05	D	D	D	D
AE19		A	A	A
AL13	D				D	D	D	D
G06		A	A	A
AO19	N
AJ03	D	D	D	D		D	D	D
AB07					D
AS11					A
O03					N
Z05					N
C15					N
E16					A
AY09-2					N
AK09						D	D
V19						N	N
M06								N
AU16								N
E16								N
AT07a	1	2	2	2	1	3	3	2

D, deletion; N, null; A, appearance. The annotations are arbitrary and interchangeable.

^a1–3 Represents the size of the marker when 1 represents the shorter and 3 the longest.

Those RAPD markers intuitively divided the strains into two major groups. Group 1 consists of the Sterne, Δ14185 and ATCC14186 and group 2 consists of Vollum, Pasteur, ATCC6605 and BA8800 strains. Group 1 is divided into two subgroups (genotypes): I –Δ14185, and II – the Sterne and ATCC14186. Group 2 consists of three subgroups (genotypes): Subgroup III consisting of Vollum strain, subgroup IV consisting of the Pasteur, ATCC6605, and subgroup V consisting of BA8800 strain. An intuitive phylogenetic tree based on the RAPD markers is presented in Fig. 2.

2

Intuitive phylogenetic tree built on the bases of the RAPD markers. On the right, grouping of the strains according to MLVA [3].

Because the RAPD analysis was performed on a small number of strains, validation using a known classification was necessary. Therefore, we classified our B. anthracis strains according to the specific MLVA primers of Keim et al. [3]. Since the primers we used were not conjugated to a fluorescent dye, we used 4% agarose, or 10% polyacrylamide gels rather then sequencing gels, to determine the differences in size of the specific markers in the different strains. We did not determine the exact genotype of the strains because some of our strains are cured from one or two virulence plasmids, and defined only the subgroups according to the MLVA markers (Fig. 2).

According to the MLVA results presented in Fig. 3 the strains Vollum, Sterne and Δ14185 (V770) were classified to the same groups, as reported by Keim et al. [3]. The Sterne strains were classified in group A3b, Vollum was part of group A4 and ATCC14185 was categorized into group A3a. The Pasteur strain, which lacks the pXO1 plasmid, was classified into group A1.a, which is in agreement with previous analysis of this strain [6].

3

Development of a three marker system that enables the classification of the eight B. anthracis strains. The specific primers marked as a are derived from the RAPD primer AA03. H – 3 direct repeats (DR), L – 2 DR. The b primers are derived from the RAPD primer AJ03. H – 4 DR, L – 3 DR. These primers are used in a multiplex PCR reaction with the AA primers. The c primers derived from the AT07 primer of the RAPD kit and is used in a separate PCR reaction because the size of the PCR products is similar to that of the AJ and AA. 1–6 DR, 2–7 DR, 3–11 DR. The strains analyzed are 1. Δ14185, 2. Sterne (10), 3. BA7001 (Sterne), 4. ATCC14186, 5. Vollum, 6. ATCC6605, 7. Pasteur, 8. BA8800. The analysis of the PCR products is summarized in the table above and additional computerized anlysis of strains from the NCBI – Ames, 2012, WesternNA (wNA) and KrugerB (kruB). The strains could be divided into five categories by the combination of the three markers.

The grouping of the B. anthracis strains by RAPD and the MLVA method was similar. The two Sterne isolates are identical and no differences could be detected between the isolates from two different sources. The ATCC14186 strain was classified to the Sterne group by the RAPD or MLVA markers and was identical to the Sterne isolates according to all markers. The Δ14185 and Vollum strains, each belongs to a separate group according to both RAPD and MLVA. Although the strains ATCC6605, Pasteur and BA8800 belong to the same group according to MLVA, the RAPD markers divide these strains into two distinguishable groups. The RAPD markers group the ATCC6605 and the Pasteur strains in one group and the BA8800 in a second. The ATCC6605 and Pasteur strains are indistinguishable by either RAPD or the MLVA markers and thus could be derived from a similar ancestor. The only way to distinguish between these two strains is the fact that the Pasteur strain lacks the pXO1 plasmid. The absence of pXO1 in the Pasteur strain obstructs the possibility of determining the exact genotype of the strain according to MLVA [3]. Due to the low resolution of our MLVA analysis we could not determine the exact genotype of the BA8800 and ATCC6605 strains. Therefore, we cannot exclude the possibility that these strains represent different MLVA genotypes of the same group. However the fact that six RAPD markers distinguish between the BA8800 and ATCC6605/Pasteur may indicate that the RAPD markers are better suited to classify these strains.

3.4 Cloning of selected RAPD markers

Although we found RAPD a reliable method to classify B. anthracis strains it might have reproducibility problems especially when performed in different laboratories. Therefore, we cloned and sequenced three specific RAPD markers (AA03, AJ03 and AT07) that enable the classification of our strains into one of the five previously determined groups.

The AJ03 genetic marker was mapped between open reading frames equal to gi/30254849 (BA1126) and gi/30254841 (BA1118) (NCBI accession number NC_003997, Read et al. [21]) of the Ames chromosome. The sequence of the AJ03 marker contains large direct repeats (Table 3) and divides the strains into two groups (Table 2); one that harbours the four direct repeats (Vollum) and the second that harbours two direct repeats (all other strains).

3

Structure of the RAPD based specific markers

RAPD marker	No. of repeats	No. of alleles	Repeat size (nucleotides)	Location	Annotation
AJ03	2–4	3	42	Between
				gi/30254841 gi/30254849
AA03	2–3	2	75	Upstream to gi/30256105	SNF symporter
AT07	6–11	3	38	gi/30259155	spoVID homolog

The AA03 genetic marker was mapped upstream of the gene that encode for a sodium dependent serine symporter (SNF family, gi/30256105, Table 3). This upstream region consists of three direct repeats separated by a spacer of 15 nucleotides, each characterized by an internal inverted repeat. These repeats, with additional flanking sequences, are identical to the regulatory region of B. cereus enterotoxin gene [20]. The AA03 marker divides the strains into two groups (Table 2). The first group has the complete sequence (Δ14185, Sterne, ATCC14186 and Vollum) whereas in the second group a deletion removes most of the first repeat and a small part of the second repeat (ATCC6605, Pasteur and BA8800).

The AT07 genetic marker was mapped onto gi/30259155 of B. anthracis (BA4659, [21]) a spoVID homolog, a protein required for the assembly of spore coat [22,23]. This spoVID homolog has three types of long direct repeats that appeared both in the DNA and protein. The AT07 marker divides the strains into three groups (Table 2, Fig. 3). The first group has six direct repeats and is marked as 1 (Δ14185 and Vollum). The second group has seven direct repeats and is marked as 2 (Sterne, ATCC14186 and BA8800). The third group has eleven direct repeats and is marked as 3 (Pasteur and ATCC6605).

3.5 Development of RAPD base analysis system

Primer pairs specific for each of the three markers were constructed according to the relevant sequence. In AA03 and AJ03 the primers sequence was based on the sequence of the RAPD primer and additional downstream sequences according to the bacterial genome. Because the AT07 RAPD marker was relatively large (?2 kbp) we synthesized primer pairs that flanked the multiple direct repeats that were the site of polymorphism.

In an attempt to simplify the typing procedure we tested the possibility to perform this analysis as a multiplex PCR. The predicted size of the PCR products of the primer pairs was around 400 nt for AJ03, around 800 nt for AA03 and 600–1000 nt for AT07. We combined the primer pairs of AJ03 and AA03 into a single multiplex reaction and performed the AT07 reaction separately. The results of the PCR reaction are presented in Fig. 3. In the upper gel the two markers divided the strains into two groups of characteristic pattern of high or low bands. The marker in the lower gel divides the strains into three groups according to the DNA band size (marked 1, 2 and 3). Integration of these results is presented in a Table in Fig. 3. These results demonstrate that it is possible to classify our strains as five genotypes according to these three specific primer pairs; the same genotypes as described earlier in Fig. 2 based on the 25 RAPD markers.

3.6 Virtual analysis of the B. anthracis strains present at the NCBI

Four B. anthracis strains exist in the NCBI database: two Ames isolates (Ames and 2012), WesternNA and KrugerB. Grouping of these strains according to the RAPD based three-marker system (Fig. 3) revealed that the Ames strain was mapped into group 2 together with the Sterne and the ATCC14186. The Sterne and Ames strains are considered to be closely related and were previously classified to the same genetic group by Keim et al. [1,3]. The westernNA strain was classified in group IV together with the strains ATCC6605 and Pasteur. The KrugerB strain represented a new genetic group with a different repeat number of the AJ03 marker (3), a combination that was not detected in any of our strains.

To conclude, the RAPD markers as short primers, or as specific primer pairs, can be used for the classification of B. anthracis isolates in the laboratory without the need for sophisticated equipment. These markers can be used to classify new B. anthracis strains or to verify the identity of a certain B. anthracis strain in research projects. Nevertheless, these markers could be incorporated into the MLVA system together with the other genetic markers [3,9,24] to generate a better system that will have the ability to predict the origin of a specific strain in a natural or bio-terrorism related Anthrax outbreak.

Acknowledgements

We thank Dana Stein and Dr. Moshe Leitner for the synthesis of the specific primers, Dr. Ofer Cohen for his assistance with the sequencing. Special thanks to Dr. Avigdor Shafferman, Dr. Baruch Velan and Dr. Sarah Cohen for their support and fruitful discussions.