Extended genetic analysis of Brazilian isolates of Bacillus cereus and Bacillus thuringiensis

Abstract

Multiple locus sequence typing (MLST) was undertaken to extend the genetic characterization of 29 isolates of Bacillus cereus and Bacillus thuringiensis previously characterized in terms of presence/absence of sequences encoding virulence factors and via variable number tandem repeat (VNTR). Additional analysis involved polymerase chain reaction for the presence of sequences (be, cytK, inA, pag, lef, cya and cap), encoding putative virulence factors, not investigated in the earlier study. MLST analysis ascribed novel and unique sequence types to each of the isolates. A phylogenetic tree was constructed from a single sequence of 2,838 bp of concatenated loci sequences. The strains were not monophyletic by analysis of any specific housekeeping gene or virulence characteristic. No clear association in relation to source of isolation or to genotypic profile based on the presence or absence of putative virulence genes could be identified. Comparison of VNTR profiling with MLST data suggested a correlation between these two methods of genetic analysis. In common with the majority of previous studies, MLST was unable to provide clarification of the basis for pathogenicity among members of the B. cereus complex. Nevertheless, our application of MLST served to reinforce the notion that B. cereus and B. thuringiensis should be considered as the same species.

The Bacillus cereus group comprises six valid species, including B. cereus, Bacillus thuringiensis, Bacillus anthracis and Bacillus mycoides that are distinguished based on plasmid content and the expression of morphological characteristics that confer distinctive pathogenicity features. These species are highly related and in the opinion of many researchers they may be considered as a single species based on genetic and genomic features (Helgason et al. 2000a, b, Anderson et al. 2005, Zahner et al. 2005, Han et al. 2006). B. cereus and B. thuringiensis are frequently found as components of the normal microflora of the gastrointestinal tract of a variety of mammal and insects (Jensen et al. 2003). However, the importance of this group of bacteria is related to food safety, their role as mammalian and/or insect pathogens and their application as biological control agents. B. cereus is an opportunistic human pathogen involved primarily in food poisoning incidents and increasingly as a source of nosocomial infections, while B. thuringiensis is one of the world's most widely used biopesticides. Although the symptoms of B. cereus food-borne illness are relatively mild, severe cases have been reported, including death by cereulide toxin (emetic toxin) and necrotic enterotoxin, cytotoxin CytK producing strains (Mahler et al. 1997, Lund et al. 2000). The heterogeneity among strains of B. cereus in the context of virulence is also a point of study. In this context, B. cereus strains have been shown to harbour diverse plasmids which shared conserved sequences with the B. anthracis virulence plasmids pXO1 and pXO2 (Jackson et al. 1998), an example is strain G9241, responsible for a case of severe pneumonia, that contained the plasmid pBCX01 which was 99.6% similar to the B. anthracis pXO1 virulence plasmid (Hoffmaster et al. 2004). Moreover, sequences encoding toxins or capsular polysaccharide of B. anthracis (pagA, lef, cya and capBCA) were detected in several B. cereus strains recovered from patients (not known to be immune compromised) with fatal pneumonia (Jackson et al. 1998).

In Brazil, the isolation of B. cereus from food is not infrequent, but food poisoning is not commonly reported. In a previous study the presence of a variety of sequences (bceT, hblA, nheBC, plcA, sph and vip3A), encoding putative virulence factors was assessed by polymerase chain reaction (PCR) assays in a collection of 80 Brazilian B. cereus and B. thuringiensis strains (Zahner et al. 2005). Amplicons were detected for all of the sequences encoding factors suggested to play a role in infections of mammals. Interestingly, the majority of the sequences were detected more frequently in B. thuringiensis isolates than in B. cereus.

Multiple locus sequence typing (MLST) studies are based on detailed analysis of the nucleotides sequence of housekeeping genes. This approach has been extensively used in the identification of genetic variation within different groups of bacteria in order to develop evolutionary frameworks, to conduct epidemiological surveillance, for the identification of epidemic clones of antibiotic resistant strains and to examine the evolution of virulence properties and population structure (Maiden et al. 1998, Priest et al. 2004, Helgason et al. 2004, Ko et al. 2004, Barker et al. 2005, de Filippis & Vicente 2005, Hoffmaster et al. 2008). In this method, relationships among isolates are based on comparisons of allelic profiles: isolates sharing four or more alleles are considered as closely related sequence types (STs).

Different groups have provided evidence for associations between specific STs or lineages and the virulence potential of strains, exemplified by pathogenic B. cereus isolates from periodontal disease and emetic gastrointestinal illness that were associated with specific clonal groups (Helgason et al. 2000a, b, Barker et al. 2005, Ehling-Schulz et al. 2005, Vassileva et al. 2006, Hoffmaster 2008). Yet, the overall population structure of the B. cereus group is not clear due to the limited numbers of strains present in the data bank. It is generally accepted that there exists a necessity to expand the number housekeeping gene sequences deposited in the data bank in order to more clearly elucidate the existence of lineages or clonal complexes associated with specific diseases.

The first objective of the current study was to use MLST to determine the genetic relationships of 29 strains of B. cereus and B. thuringiensis with 18 of them isolated from different sources in Brazil and the remainder derived from a range of countries and sources. The second objective was to extend our earlier investigation in relation to the distribution of genes encoding putative virulence factors, specifically be, responsible for synthesis of the emetic toxin cereulide (Toh et al. 2004), cytK encoding necrotic enterotoxin (Fagerlund et al. 2004), inA encoding a metalloprotease of B. cereus (Gutmann & Ellar 2000), the sequences pag, lef and cya located on the pXO1 plasmid of B. anthracis and the B. cereus plasmid pBCX01 and finally the gene cap encoding the capsular antigen of pXO2 (Jackson et al. 1998). The present work complements and enhances the previous study of Zahner et al. (2005) where the relationship between B. cereus and B. thuringiensis was investigated.

MATERIALS AND METHODS

Bacterial strains - Twenty-nine strains including B. cereus or B. thuringiensis and B. mycoides were included in this study (Table I). All these strains were originally present in Coleção de Culturas do Gênero Bacillus e Gêneros Correlatos - Bacillus Collection/Oswaldo Cruz Institute/Oswaldo Cruz Foundation. Details of each strain can be found in Zahner et al. (2005).

TABLE I. Distribution of virulence genes, variable-number of tandem repeats (VNTR) types, sequence types (STs) and origin of the Bacillus species examined .

Isolate	Bacterial species	Source of isolation	Country of isolation	Year of isolation	ST	VNTR typea	cytk	inA	be	pag, lef, cya	cap
LFB563	Bacillus thuringiensis pakistani	Soil	Brazil	NA	534	4	P	N	N	N	N
LFB683	B. thuringiensis finitimus	Soil	Brazil	NA	529	3	N	N	N	N	N
LFB891	B. thuringiensis aizawai	Food - black pepper	Brazil	1994	535	4	P	N	N	N	N
LFB472	B. thuringiensis sotto	Insect	Pakistan	NA	518	1	P	P	N	N	N
LFB858	B. aizawai	Food - black pepper	Brazil	1994	526	3	P	P	N	N	N
LFB860	B. thuringiensis israelensis	Food - black pepper	Brazil	1994	537	5	P	P	N	N	N
LFB862	B. thuringiensis	Food- black pepper	Brazil	1994	523	2	P	N	N	N	N
LFB861	B. thuringiensis israelensis	Food - black pepper	Brazil	1994	538	5	P	N	N	N	N
LFB914	B. thuringiensis israelensis	Soil	Brazil	1992	540	5	P	N	N	N	N
LFB584	B. thuringiensis israelensis	Insect	Israel	NA	527	2	N	N	N	N	N
LFB476	B. thuringiensis morrisoni	Soil	Brazil	1998	522	2	P	N	N	N	N
LFB120	Bacillus mycoides ATCC1120	NA	NA	NA	524	2	P	N	N	N	N
LFB208	Bacillus cereus ATCC 33018	Food	Colombia	NA	519	1	N	N	N	N	N
LFB476	B. cereus ATCC 33019	Food	Colombia	NA	516	3	N	N	N	N	N
LFB406	B. cereus NCTC 2599	NA	NA	NA	517	5	P	N	N	N	N
LFB1123	B. cereus	Insect	Madagascar	NA	520	5	P	N	N	N	N
LFB1125	B. cereus	Insect	Madagascar	NA	528	3	P	N	N	N	N
LFB1130	B. cereus	Insect	Egypt	NA	539	5	NA	N	NA	N	N
LFB1131	B. cereus	Insect	Czechoslovakia	NA	536	5	P	N	NA	N	N
LFB1133	B. cereus	Insect	USA	NA	533	4	P	N	NA	N	N
LFB513	B. cereus	Food - rice mousse	Brazil	1987	542	3	N	N	N	N	N
LFB693	B. cereus	Food - ice cream	Brazil	1987	530	4	P	N	N	N	N
LFB704	B. cereus	Food - soup	Brazil	1987	541	5	P	N	N	N	N
LFB585	B. cereus	Urinary infection	Brazil	< 2001	525	3	P	N	N	N	N
LFB586	B. cereus	Urinary infection	Brazil	< 2000	531	4	P	N	N	N	N

a: additional details are provided in Zahner et al. (2005). LFB codes are as used in the culture collection catalogue of the Oswaldo Cruz Institute. N: negative; NA: not available; P: positive.

Molecular methods - Extraction and purification of genomic DNA has been described previously (Zahner et al. 2005). The presence the gene be, encoding emetic toxin, was examined by PCR as described by Toh et al. (2004) and be-PCR amplification products were sequenced. The cytK sequences were detected according to the protocol of Fagerlund et al. (2004), with the primers CytkF 5'-CCAACCCAGTTACCAGTTCC-3' and CytkR 5'-AACAGATATCGGTCAAAATGC-3' and the inA gene was amplified according to the methods of Gutmann and Ellar (2000). Nested-PCR for amplification of the structural genes of the entire complement of B. anthracis toxins found on pXO1 (pag, lef, cya) and the capsular antigen gene of pXO2 (cap) was carried out as described by Jackson et al. (1998). Details of primers used are provided in Table II. Strains showing a negative result for the be sequence by PCR assay were submitted to Southern blotting and hybridization with a be probe as previously reported (Zahner et al. 1998). Each PCR reaction was repeated at least three times to confirm reproducibility. This publication made use of the Bacillus cereus Multi Locus Sequence Type website (pubmlst.org/bcereus) developed by Jolley et al. (2004) analyzing the seven housekeeping genes (glp, gmk, ilvD, pta, pur, pycA and tpi) employing PCR conditions as described in the MLST database. PCR products from MLST genes and emetic toxin gene (be) were purified using the GFX-PCR DNA Kit (GE Healthcare Life Sciences). PCR fragments were sequenced in both directions, using the amplification primers to provide unambiguous sequence data, by use of the BigDye Ready Reaction mix (ABI Corp) and reaction products were analyzed on a Prism 3700 automated DNA analyser (ABI Corp). All the sequencing procedures were performed as published by Otto et al. (2008).

TABLE II. Sequences of DNA oligonucleotides used as primers in polymerase chain reaction for detection of virulence genes.

Target	Primer sequence 5'-3'	Product size (bp)	Reference
pag	F CCAGACCGTGACAATGATG	508	Jackson et al. (1998)
	R CAAGTTCTTTCCCCTGCTA
paga	F CGAAAAGGTTACAGGACGG	409	Jackson et al. (1998)
	R CAAGTTCTTTCCCCTGCTA
lef	F GGTGCGGATTTAGTTGATTC	851	Jackson et al. (1998)
	R CGCTTCATTTGTTCTCCC
lefa	F GAAACATCGGTCTGGAAAT	403	Jackson et al. (1998)
	R CCCTTTTGAATGAACTTGC
cya	F GCGATGAAAACAACGAAGTA	720	Jackson et al. (1998)
	R TCGTCTTTGTCGCCACTATC
cyaa	F CATTAGAAAAGCAAAAAGGTC	186	Jackson et al. (1998)
capA	R TCATTATCTTGCTCTGTGCC
	F CAGAAGCAGTAGCACCAGTAA	397	Jackson et al. (1998)
capAa	R ATTTTCACCAGCACCCAC
	F TGACGATGGTTGGTGACA	302	Jackson et al. (1998)
capB	R CCTTATTGTATCTTTAGTTCCC
	F CTGACCAATCTAAGCCTGC	221	Jackson et al. (1998)
capC	R TCGTTTCTCCAATCGCAAT
	F GTACCTGGTTATTTAGCACTCG	208	Jackson et al. (1998)
be	R ATCTCAAATGGCATAACAGG
	F ACTTAGATGATGCAAGACTG	850	Toh et al. (2004)
inA	R TTCATAGGATTGACGAATTTT
	F TGAGCCATTCGCGTTAGAGG	529	Gutmann and Ellar (2000)
cytK	R ATATTCATGCGCGTTTACACC
	F CCAACCCAGTTACCAGTTCC	810	Fagerlund et al. (2004)
	R AACAGATATCGGTCAAAATGC

a: internal primers used in nested polymerase chain reaction.

Data analysis - Sequencher 4.8 software (Genecodes) was used for assembly of contigs and determination of consensus sequences. The nucleotide sequences of each housekeeping gene were trimmed to the appropriate length (348-504 bp), as previously described (Jolley et al. 2004), and then queried to the MLST database. STs were assigned based on the combination of the seven alleles in a specific order. New alleles and allele sets were deposited in the Bacillus MLST database. Designation of new alleles and ST was determined by MLST database curators. A phylogenetic tree was constructed based on the multiple alignments of the concatenated sequences of the seven MLST genes through the neighbour-joining method using Mega 4.0 software (Tamura et al. 2007). Split decomposition analysis of allelic profiles was carried out with the SplitsTree 4.0 software (Huson 1998). Some known strains from the MLST Bacillus databank were included to assess the relationships of those isolates with the ones identified in our study.

RESULTS

Table I presents the distribution of virulence genes, variable-number of tandem repeats (VNTR) types, STs and origin of the Bacillus species examined.

The Figure represents phylogenetic relationships determined for the strains utilized in this study. It was observed that B. thuringiensis and B. cereus isolates were intermixed within the phylogenetic clusters and that no B. thuringiensis or B. cereus clustered with the B. anthracis group. Each of the strains belonged to a novel ST and occurred only once (STs 516-544) (Table I). The number of alleles per locus was: 16 (glp), nine (gmk), 15 (ilvD), 18 (pta), 16 (pur), 15 (pycA) and 15 (tpi). The alleles glp-13, gmK-8, ilvD-9, pta-11, pur-18, pycA-14 and tpi-7 were the most frequent per locus. No specific allele could be associated with either the origin of isolation or country. Analysis of the allelic diversity among the isolates resulted in the average number of alleles per locus of 14.8. A phylogenetic tree (Figure) was constructed from a single sequence of 2,838 bp of concatenated loci sequences. The tree was derived using the novel isolates and included sequence data derived from previously characterized isolates stored in the Bacillus cereus Multi Locus Sequence Type website (pubmlst.org/bcereus). The overall structure of the maximum-likelihood tree generated from concatenated sequences resulted in two major phylogenetic groups (Figure): G1 (group 1), includes strains of clade 1, and G2 (group 2), includes strains pertaining to clade 2 as defined by Priest et al. (2004).

DISCUSSION

The first objective of the current study work was to increase the number of strains in the MLST database, in an attempt to more confidently assess if there are specific lineages or clonal complexes containing clinical isolates that might be associated with a specific illness or defined biological characteristics. For this reason we selected a range of B. cereus and B. thuringiensis strains isolated from a variety of environments and countries, with the majority (64.3%) of them isolated in Brazil.

The study of Zahner et al. (1989) focused on population structure and diversity of the B. cereus and B. thuringiensis group based on the use of multilocus enzyme electrophoresis (MLEE) and concluded that the isolates examined represented a single species, it was also demonstrated that the B. cereus complex presented a high level of genetic diversity. Similar conclusions were reached from subsequent studies applying MLEE to evaluation of this group of bacteria (Helgason et al. 2000a, b, Villas-Boas et al. 2002). However, it is generally accepted that MLEE has limited applicability especially in relation to data analysis and the interchange of results between laboratories, while MLST provides some solutions to those shortcomings. The main advantages of MLST, over MLEE (the basis for MLST), are that DNA is more variable than encoded proteins (on account of primarily synonymous mutations) and fewer loci need to be examined in MLST. Furthermore, nucleotide sequence data are generic, definitive, readily compared among laboratories and easily analyzed by a range of phylogenetic and population genetic techniques. Such data can be made freely accessible via the internet, establishing virtual isolate collections that enable the rapid and accurate comparison of the genetic profiles of isolates obtained anywhere in the world (Maiden et al. 1998).

Different authors have used MLST in attempts to elucidate intra and inter B. cereus relationships. In order to characterize the population structure and epidemiology of B. cereus group, different MLST schemes have been developed: the Priest scheme (pubmlst.org/bcereus/), the Tourasse-Helgason scheme (mlstoslo.uio.no/) (spock.jouy.inra.fr/cgi-bin/bacilliMLSopen.cgi) (Sorokin et al. 2006), the Ko scheme (Ko et al. 2004) and the SuperCat scheme (Tourasse & Kolstø 2008). Helgason et al. (2004) and Priest et al. (2004) using different MLST genes suggested a weakly clonal population structure for B. cereus. In this case, according to the authors, the evolutionary changes occurred via point mutations. Subsequently, Sorokin et al. (2006) studying localized populations of B. cereus group in soil suggested that B. cereus and B. thuringiensis are clonal. In addition, Hoffmaster et al. (2008) examined the phylogenetic diversity and relatedness of B. cereus clinical isolates present in a collection of B. cereus recovered along 50 years and observed that those isolates were phylogenetically diverse.

Among the different MLST schemes we opted to use the one described by Priest et al. (2004) because of the advantage of allele and ST number facilitating immediate identification of the strain, a feature which is lacking in other methods such as the SuperCat scheme. In addition, this databank holds sequences produced from numerous Brazilian isolates facilitating strain comparison. Corroborating previous observations, it was determined that the B. thuringiensis and B. cereus isolates were intermixed within the phylogenetic clusters. However, no B. thuringiensis or B. cereus clustered within the B. anthracis group. The overall topology of the tree also provided support to previous data (Hoffmaster et al. 2008), showing that the B. cereus complex population is clonal with clinical and environmental isolates emerging from different phylogenetic positions.

The 29 test strains were recovered from different sources including soil, insects, human infections, foodstuffs and a lethal case of avian septicaemia; with the majority isolated in Brazil. Each of the strains analyzed belonged to a novel ST and occurred only once (STs 516-544) (Table I), demonstrating that, as suggested by other authors, MLST is a useful and powerful tool for molecular epidemiology (Maiden et al. 1998 , Sorokin et al. 2006, Ibarz & Maiden 2009). Analysis of the allelic diversity among the isolates resulted in the average number of alleles per locus of 14.8, which is relatively low in comparison to the values of 19.9, 33.6 and 30.5 which were obtained by other authors (Helgason et al. 2004, Priest et al. 2004, Hoffmaster et al. 2008). No specific allele could be associated with either the origin of isolation or country. Given the fact that the strains were obtained from a variety of sources, the observed values strongly indicated the clonal aspect of evolution within the B. cereus complex.

The phylogenetic tree (Figure) demonstrated that the strains examined in this study were not monophyletic by analysis of any specific housekeeping gene or characteristic previously known or described herein; instead, they were distributed among the Bacillus group G1 and G2 and in different previously defined lineages (Priest et al. 2004).

Analysis was undertaken to investigate if any of the examined virulence factors could be considered as characteristic for a specific phylogenetic lineage. Our results demonstrated that there was no clear association in relation to source of isolation or to genotypic profile based on the presence or absence of putative virulence genes (Table I), confirming the findings of an earlier but less extensive genotypic examination of these isolates (Zahner et al. 2005). Furthermore, B. cereus and B. thuringiensis were shown to be distributed among several lineages, although unique species could be associated with specific STs, such as 522 and 529, 526 and 537, all B. thuringiensis isolated from soil and food, respectively, or STs 532 (clinical) and 542 (food), both B. cereus, ascribed to lineage Cereus I (Priest et al. 2004). The B. cereus isolate LFB 1181(ST-544) was recovered from an outbreak of fatal septicaemia in captive birds in Brazil. Interestingly, when this strain was tested for cytotoxicity towards Chinese hamster ovary cells it presented the highest level (titre 1:256) of toxicity among the test isolates (data not shown). Moreover, isolate LFB1181 was the only strain of B. cereus which was recorded as inhA PCR positive, suggesting a possible correlation between this enzyme and the disease outbreak in birds. In this context, Guillemet et al. (2010) suggested that the InhA metalloproteases of B. cereus are important virulence factors that may allow bacteria to counteract the host immune system. Intriguingly, this sequence was also detected in two Brazilian isolates of B. thuringiensis (serotypes aizawai and israelensis), both from black pepper, and a strain of B. thuringiensis (serotype sotto) isolated from an insect in Pakistan. The presence of sequences encoding this protease in environmental isolates, generally considered as non-pathogenic for mammals, suggests that as noted for other putative virulence factors (bceT, hblA, nheBC, plcA and sph), the role of this protein is not primarily the establishment or maintenance of infection in warm blooded hosts (Zahner et al. 2005).

Ehling-Schulz et al. (2005) using data based on MLST, reported that cereulide-producing emetic B. cereus presented a clonal population structure, forming a highly clonal complex. The panel of isolates used in this study did not include any strains involved in food poisoning, but interestingly the presence of the sequence be, which in emetic strains is carried on mega plasmids of more than 200 kilobase pairs and is involved in cereulide synthesis, was detected in LFB598 (ST-532) by PCR (Table I). The be derived amplicon of the positive strain LFB598 (ST-532) were sequenced and BLAST analysis indicated that it shared 96% homology with be gene sequences deposited in the database. The strain LFB598 was isolated from a Brazilian stool sample and was placed in GII. Yet, it should be stressed that no tests were performed to determine if the PCR positive isolate expressed functional cereulide toxin.

The presence of sequences encoding the tripartite toxin and poly-D-glutamate capsule of B. anthracis was investigated via PCR however no isolates, including the avian pathogenic strain LFB1181, generated amplicons for any of the genes examined. The B. cereus isolate which was placed closest to B. anthracis in our comparative analysis (Figure) was strain LFB585 (ST-525), a Brazilian isolate recovered from a urinary tract infection (Table I). The genetic profile of this strain, in terms of putative virulence factors, as previously demonstrated (Zahner et al. 2005) and enhanced by data from the present work, indicates that the strain is PCR positive for cytK, hbl, nheBC, bceT, sph, but there was no evidence by PCR or nested PCR for B. anthracis pXO1 and pXO2 virulence genes. The proximity of ST-525 with B. anthracis adds some support to the hypothesis that B. cereus isolates that cluster closely with B. anthracis are associated with clinical cases (Helgason et al. 2000a, b, Hill et al. 2004). Yet on the other hand, an additional clinical isolate recovered from faeces (ST-543) and placed in GII was distant from B. anthracis (Figure). This apparent contradiction highlights the aforementioned shortcomings of the existing MLST database.

Although the majority of our B. cereus samples were either environmental or food associated, we also analyzed a number of clinical isolates. In the cladrogram, using strains already present in the MLST database (Figure), numerous environmental strains were firmly placed in the same group as our clinical isolates. Thus, it could be suggested that all B. cereus sensu latu posses pathogenic potential, but what triggers the pathogenicity is still not known. Clearly the mere presence of sequences encoding the putative virulence factors examined in this study is not sufficient to definitively confer pathogenicity. Hence, it is worth considering the observation of Hoffmaster et al. (2004), who stated that the isolation of B. cereus in a clinical laboratory is frequently considered as contamination, with strains commonly being discarded rather than followed up with further characterization and reporting. Consequently, the analysis of clinical isolates is often inadequate resulting in an underestimation of both the frequency and significance of non-gastrointestinal infections caused by B. cereus and possibly other members of the B. cereus group.

Despite the sophistication of the methodology, the inability of MLST to link virulence with genotype was further demonstrated by the observation that strain LFB1181 (ST-544) was firmly placed in the same clade as strains isolated from a variety of environmental sources and which demonstrated distinct genotypic characteristics (Table I). This observation adds support to the notion that the putative virulence factors are randomly distributed among B. cereus and B. thuringiensis and that a simple assessment for the presence of these factors will be unable to confidently predict the virulence of members of the B. cereus group (Guttman & Ellar 2000, Zahner et al. 2005). Support for this statement, in the context of food poisoning, was recently provided by the work of Guinebretière et al. (2010), who demonstrated that the ability of B. cereus group strains to cause food poisoning varies according to phylogenetic affiliation (groups I-VII) rather than species affiliation.

In light of their abundance and extensive polymorphism, VNTR have proven to be fundamental genetic markers in bacterial population and/or phylogenetic analysis. Analysis based on VNTR compares the strain-specific numbers of repeats of short DNA sequences at different positions of the bacterial genome and provides functional and evolutionary information concerning genetic relationships in microbial species (Van Belkum 1999). Moreover, in the case of some highly clonal bacterial species, e.g. Acinetobacter baumanni, VTNR analysis may serve as a powerful discriminatory method for the identification of distinct populations of highly similar isolates associated with disease (Turton et al. 2011). The 29 test strains were previously characterized by VNTR analysis, using a variable region of the vrrA open reading frame as the target and were ascribed to five different fragment length categories (Zahner et al. 2005). A comparison of VNTR profile and STs was performed in the present study (Table I). In contrast to the analysis in relation to the presence/absence of putative virulence factors, a correlation between VNTR profile and MLST was suggested. Specifically 63% of strains identified as VNTR 4 or 5 were placed in GI, while in GII 73% presented the VNTR profiles 1, 2 or 3. However, once again this apparent association between genetic features was unable to provide a means of confidently discriminating between clinical and environmental members of the B. cereus complex.

With the possible exceptions of emetic and periodontal isolates of B. cereus (Ehling-Schulz et al. 2005, Vassileva et al. 2006, Hoffmaster et al. 2008) and the plasmid encoded virulence factors of B. anthracis, the use of genetic characterization, including the data reported in the current study, has largely failed to explain the basis for pathogenicity of members of the B. cereus complex. Recent proteomic analysis (Claire et al. 2010) has served to increase the repertoire of putative virulence factors expressed by B. cereus, with 31 novel virulence related molecules identified in the secretome. The expression of these factors by other members of the B. cereus complex has yet to be investigated. However, it is becoming increasingly apparent that molecular characterization in the absence of functional analysis of putative virulence gene expression is unlikely to clarify the mechanisms which allow a given isolate to cause disease. On the other hand, our application of MLST has served to reinforce the notion that B. cereus and B. thuringiensis should be consid-ered as the same species (Helgason et al. 2000a, b, Anderson et al. 2005, Zahner et al. 2005, Han et al. 2006).