FEMS Microbiology Reviews 29 (2005) 303–329

www.fems-microbiology.org

q
Genomics of the Bacillus cereus group of organisms
David A. Rasko a, Michael R. Altherr b, Cliff S. Han b, Jacques Ravel a,*

a
The Institute for Genomic Research, 9712 Medical Center Drive, Rockville, MD 20850, USA
b
Los Alamos National Laboratory, P.O. Box 1663, Los Alamos, NM 87545, USA

Received 5 November 2004; received in revised form 22 December 2004; accepted 23 December 2004

First published online 28 January 2005

Abstract

Members of the Bacillus cereus group of organisms include Bacillus cereus, Bacillus anthracis and Bacillus thuringiensis. Collec-
tively, these organisms represent microbes of high economic, medical and biodefense importance. Given this significance, this group
contains the highest number of closely related fully sequenced genomes, giving the unique opportunity for thorough comparative
genomic analyses. Much of the disease and host specificity of members of this group can be attributed to their plasmids, which vary
in size and number. Chromosomes exhibit a high level of synteny and protein similarity, with limited differences in gene content,
questioning the speciation of the group members. Genomic data have spurred functional studies that combined microarrays and
proteomics. Recent advances are reviewed in this article and highlight the advantages of genomic approaches to the study of this
important group of bacteria.

2005 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved.

Keywords: Bacillus cereus; Bacillus anthracis; Genome; Plasmid

Contents

1. Introduction – The Bacillus cereus group of organisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304

1.1. Bacillus thuringiensis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
1.2. Bacillus cereus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
1.3. B. anthracis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
2. Molecular identification of members of the B. cereus group of organisms: One species on the basis of genetic evidence? . . . 305
3. Genome sequencing projects completed and in progress. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
4. B. anthracis genomic analysis for molecular and forensic epidemiological purposes . . . . . . . . . . . . . . . . . . . . . . . . . . 309
5. Comparison of the chromosome sequences of the B. cereus group of organisms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
5.1. Non-pathogenic B. cereus ATCC 14579 and B. cereus ATCC 10987 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
5.2. The pathogenic members of the B. cereus group of organism, B. cereus G9241, B. cereus Zebra Killer, and
B. thuringiensis 97-27 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
6. The B. cereus group of organisms plasmid sequence comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
6.1. General features of the B. cereus group plasmids identified in genome sequencing. . . . . . . . . . . . . . . . . . . . . . . 313
6.2. Replication and mobility mechanisms of the B. cereus group plasmids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317

q
Edited by Mark J. Pallen.
*
Corresponding author. Tel.: +1 301 795 7884; fax: +1 301 838 0208.
E-mail address: jravel@tigr.org (J. Ravel).

0168-6445/$22.00
2005 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.femsre.2004.12.005
304 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329

7. The phage of the B. cereus group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320

7.1. Genomic phage content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
7.2. Lytic induction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
7.3. Exploitation of phage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
8. Functional genomics of the B. cereus group of organisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
8.1. Transcriptional analysis using microarrays. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
8.2. Proteomics studies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
9. Summary: One species on the basis of genomic evidence? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325

1. Introduction – The Bacillus cereus group of organisms
The Bacillus cereus group of organisms contains
Bacillus thuringiensis, Bacillus anthracis and Bacillus cer-
eus (sensu stricto). This group of Gram-positive spore-
formers forms a highly homogeneous subdivision of
the genus Bacillus. Demonstration of their high genetic
relatedness has contributed to the suggestion that
B. anthracis, B. cereus and B. thuringiensis are members
of a single species, B. cereus sensu lato [1–3]. Tradition-
ally, these organisms have been differentiated based on
their phenotypic characteristics, including pathogenic
potential. Bacillus mycoides, Bacillus pseudomycoides
and Bacillus weihenstephanensis are also members of
the B. cereus group of organisms. However, as no geno-
mic data are available, these species have not been in-
cluded in this review.
death of the insect the bacterium is released into the soil
where B. thuringiensis is a ubiquitous inhabitant. In this
environment and under favorable nutrient condition the
spores could germinate and grow. The Cry genes encode
the crystal toxins and are usually located on large, trans-
missible plasmids. The presence of Cry protein crystals
in the spore is speculated to give B. thuringiensis an
advantage in the soil environment upon sporulation [7]
over B. cereus, B. thuringiensis is phenotypically distin-
guished from B. cereus only by the formation of intracel-
lular protein crystals during sporulation. Overall,
genetic studies have shown that B. cereus and B. thurin-
giensis are essentially identical [8]. In addition, like B.
cereus, B. thuringiensis could be considered an opportu-
nistic pathogen in animals and human [9–11].
1.2. Bacillus cereus

1.1. Bacillus thuringiensis B. cereus is ubiquitous in nature and an opportunistic

B. thuringiensis has long been regarded as an insect
pathogen alone. The insecticidal spectrum varies within
the 82 different serotypes reported [4], and affects insects
primarily from the orders Lepidoptera, Diptera and
Coleoptera. There are also reports of B. thuringiensis iso-
lates active against mosquitoes that are vectors for dis-
ease, such as malaria and yellow fever [5]. B.
thuringiensis spore preparations have been successfully
commercialized as biopesticides. The spores are associ-
ated with large crystal protein inclusions, which can
make up to 25% of the dry weight of the spore prepara-
tions. The crystals are aggregates of a large protein
(about 130–140 kDa) that is actually a protoxin. Upon
ingestion by insect larvae, the protoxin crystals solubi-
lize in the mid-gut, where it is cleaved by a gut protease
to produce an active toxin (d-endotoxins) of about
60 kDa. The toxin binds to the mid-gut epithelial cells,
creating pores in the cell membranes. As a result, the
gut is rapidly immobilized and the epithelial cells lyse.
The insect larva stops feeding and often dies from lethal
septicemia [6]. Little is known about the ecology of B.
thuringiensis and conflicting reports are reviewed by Jen-
sen et al. [7]. The B. thuringiensis natural environment is
thought to be the insect host intestinal system. Upon
pathogen, often associated with two forms of human
food poisoning, characterized by either diarrhea and
abdominal distress or nausea and vomiting. In healthy
individuals but mostly in individuals with certain under-
lying conditions such as, immunocompromised patients,
or patients recovering from surgery, B. cereus has been
known to cause a variety of infections, including:
endophthalmitis, bacteremia, septicemia, endocarditis,
salpingitis, cutaenous infections, pneumonia and menin-
gitis [12,13]. B. cereus is found as a contaminant in many
food products, including dairy products. However, its
primary ecological niche is the soil environment. It is
also commonly found as part of the gut microflora of
invertebrates, not only as spores but also as growing
vegetative cells [14]. By definition, B. cereus is acrystallif-
erous, but a B. cereus strain carrying a functional cry
gene is considered as a B. thuringiensis strain [15]. No
virulence factors specific to B. cereus have been identi-
fied, and proteins thought to be specific to B. cereus have
recently been found in B. thuringiensis isolates [6].
1.3. B. anthracis
B. anthracis is the etiological agent of anthrax, an
acute fatal disease found primarily among herbivores,
 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329
but in fact all mammals are susceptible. In recent years,
it has become best known for its use as a biological
weapon [16,17]. The bacterium is endemic or hyper-en-
demic in Africa, Asia and South America [15] but can
be found in most of the world. B. anthracis spores are
highly resistant to heat, ultraviolet and ionizing radia-
tions, pressure and a variety of chemical agents and
are thought to survive in the environment for decades
and possibly centuries [18]. Despite the fact that B.
anthracis shares the same ecological niche as B. cereus
and B. thuringiensis, and can be readily isolated from
contaminated soil samples, it is still unclear if its life cy-
cle includes a vegetative stage outside the host [19].
Spores are ingested by herbivores and germinate within
the host to produce vegetative cells, which multiply and
produce virulence factors, ultimately killing the host.
Upon death, large numbers of bacilli are released into
the environment (i.e., soil), and sporulate upon contact
with air, completing the life cycle.
B. anthracis can be differentiated from B. cereus and
B. thuringiensis though microbiological and biochemical
tests. B. anthracis isolates are non-hemolytic, non-
motile, penicillin-sensitive, susceptible to c-phage, and
produce a poly-c-D-glutamic acid capsule [20]. Like
B. thuringiensis, the ability of B. anthracis to cause a dis-
ease is primarily attributed to its plasmid content. Fully
virulent strains of B. anthracis carry two large plasmids,
pXO1 (181 kb) and pXO2 (96 kb), which encode the
machinery necessary to produce and regulate the an-
thrax virulence factors, the tripartite toxin and the cap-
sule, respectively [21,22].
2. Molecular identification of members of the B. cereus
group of organisms: One species on the basis of genetic
evidence?
The classification and taxonomic separation of
B. anthracis, B. cereus and B. thuringiensis has long been
cause for controversy among bacteriologists, and distin-
guishing these species is rather difficult even with mod-
ern molecular tools. The close relationship among the
different members of the B. cereus group of organisms
has been established through the use of molecular tech-
niques such as DNA–DNA hybridization [23] and more
recently by comparing 16S or 23S rRNA sequences or
16S–23S rRNA spacer regions [2,24–26]. However, these
methods could not adequately differentiate members of
this group. There are numerous reports of attempts at
developing molecular typing systems that would dis-
criminate between B. cereus, B. thuringiensis and
B. anthracis. These included multilocus enzyme electro-
phoresis (MEE) [3,8,27], multilocus sequence typing
(MLST) [28–30], fluorescent amplified fragment length
polymorphism analysis (AFLP) [31–33] and rep-PCR
fingerprinting [34] among others. None of these studies
305
were able to consistently distinguish B. cereus from B.
thuringiensis, but all showed a high level of genetic diver-
sity between these two phylogenetically interspersed
species.
In addition, these data confirmed that B. anthracis is
a monomorphic species, in which the overall diversity
among isolates cannot be accurately distinguished using
these techniques. Suppression subtractive hybridization
was also used to catalogue some of the unique genetic
content of B. anthracis [35]. Ninety-three genomic differ-
ences were discovered that could potentially be used as
‘‘signatures’’ to discriminate B. anthracis from its closest
relatives but not between B. anthracis strains. The
inability to separate B. anthracis isolates from one an-
other led the Keim Genetics Laboratory to develop a ro-
bust B. anthracis molecular typing system based on
variable number of tandem repeats (VNTR). A multi-
ple-locus VNTR analysis (MLVA) using 8 then 15 loci
efficiently cluster B. anthracis in two major groups (A
and B), with the A group widely distributed worldwide,
while the B group has a limited geographic distribution
[36].
These molecular typing methods have yet to answer
whether B. cereus, B. thuringiensis and B. anthracis be-
long to one unique species and are variants of that
species. The simplest markers used for phenotypic dif-
ferentiation of these species are all encoded on plas-
mids (i.e., the pXO2 encoded capsule gene cluster)
[19]. The fact that plasmids can easily be transferred
or lost makes these criteria unacceptable for typing
purposes. For example, it is believed that a plasmid-
less isolate of B. anthracis is indistinguishable from
B. cereus [37]. In addition, the isolation of B. cereus
isolates containing the B. anthracis plasmid pXO1,
with a novel capsule and associated with an illness
resembling anthrax [20], makes this classification inap-
propriate. B. anthracis appears to be a highly success-
ful variant of the B. cereus group, with a large
number of strains that have been isolated, simply be-
cause it is so easy to identify cases of classical an-
thrax, which in turn is a readily recognizable disease
[38]. Is it possible that this ease of isolation has biased
our view of this group and thus B. anthracis could
represent an over-sampled B. cereus? RepPCR analysis
placed B. anthracis as an independent lineage in a B.
cereus cluster. Perhaps all these species should be clas-
sified as B. cereus and subsequently differentiated by
their plasmid content, as suggested previously [3]?
Analyses of large culture collections of B. anthracis,
B. cereus and B. thuringiensis isolates by AFLP and
MLST [20,28,31] have identified a class of organisms
containing toxigenic B. cereus and B. thuringiensis that
are closely related to B. anthracis. These isolates were
phylogenetically distinct from environmental B. cereus
and B. thuringiensis [31] and might represent the closest
ancestor to B. anthracis.
306 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329
It is with this view of the genetic diversity of the
B. cereus group that we have entered the genomic era.
This knowledge has driven our selection of organisms
to sequence and genomic-scale tools have been can be
used to attempt to settle these outstanding questions.
3. Genome sequencing projects completed and in progress
Like many other bacteria, the B. cereus group of
organisms has benefited from the genomic revolution
that started in 1995 with the publication of the first
microbial genome sequence, Haemophilus influenzae
[39]. Access to the genome sequence of a representative
of each member of the group opens up the fields of
transcriptomics and proteomics, and furthers our under-
standing of the mechanisms driving the specific pathoge-
nicity of these organisms. In addition, armed with the
genome sequence and a better understanding of the biol-
ogy of B. anthracis, the generation of better and safer
vaccine and anti-microbial drugs is possible. As de-
scribed above, plasmid content confers some of the phe-
notypic traits used to distinguish B. anthracis, B. cereus
and B. thuringiensis. Comparative genomics of members
within the group can address the question – are these
organisms separate species, or the same species carrying
different plasmids?
The first B. anthracis genome sequencing project was
underway prior to the Autumn 2001 mail anthrax at-
tacks [40,41]. This bioterrorism event has had an enor-
mous impact on the number and choice of projects
undertaken as genomic tools were applied to microbial
forensics and biodefense/biopreparedness became a dri-
ver for a number of funding agencies.
To date, the genome sequence of 15 isolates from the
B. cereus group of organisms are available in public dat-
abases (Table 1) and others are underway. Conse-
quently, this group of bacteria provides one of the
richest collections of near neighbor sequences and will
likely profoundly impact future systematics efforts. B.
anthracis genome projects account for two thirds of
these projects (Table 1). These 10 projects are aimed at
establishing a better understanding of the genetic struc-
ture of the natural B. anthracis population, through the
discovery of novel polymorphisms, and possibly of
genes unique to certain strains. Single nucleotide poly-
morphisms (SNPs) were used to unravel the molecular
evolutionary history of the group, and were the basis
of a robust and finely detailed typing scheme for B.
anthracis isolates [42]. The first draft genome to be pub-
lished was that of the B. anthracis Ames strain isolated
from the victim of the bioterror attack in Florida [42].
The choice was obviously motivated by the need to com-
pare the draft sequence of this isolate to that of the
Ames strain in progress at the time [40] at The Institute
for Genomic Research (TIGR). The latter, a non-viru-
lent isolate of the Ames strain was obtained from the
Defense Evaluation and Research Facility at Porton
Down, UK [43], where it was cured of plasmids pXO1
and pXO2 by incubation at 43 C and by novobiocin
treatment, respectively [40]. The attenuated B. anthracis
Ames Porton Down strain was completely sequenced
and released in 2003. The original B. anthracis Ames
Ancestor strain, from which both the Florida and Por-
ton Down isolates were derived, was sequenced to com-
pletion, and is the only fully virulent strain of B.
anthracis completely sequenced to date. This strain
was acquired in 1981 by researchers at the United States
Army Medical Research Institute for Infectious Dis-
eases (USAMRIID) from the Texas A&M Veterinary
Medicine and Diagnostic Laboratory Bacteriology
Department. The strain was originally isolated from a
dead cow in Sarita, TX. An original culture received
in 1981 and stored at USAMRIID since, was the source
of genomic DNA for sequencing. This strain is the pro-
genitor to all the Ames strains used in laboratories
around the world, hence it is an important reference.
A set of six diverse isolates of B. anthracis was selected
for genome sequencing to draft coverage (8·). These iso-
lates were carefully chosen to represent as much genetic
diversity as possible, as previously established through
MLVA typing [36]. Three representatives of the A group
(B. anthracis North Western America (AAER00000000),
B. anthracis Vollum (AAEP00000000), B. anthracis Aus-
tralia 94 (AAES00000000) – Table 1), two of the B group
(B. anthracis Kruger B (AAEQ00000000), B. anthracis
CNEVA (AAEN00000000) – Table 1) and one isolate
belonging to the newly discovered C group (B. anthracis
A1055 (AAEO00000000)) were sequenced and have been
recently released to GenBank (Table 1). Comparative
analysis of these genomes to the B. anthracis Ames
Ancestor genome has provided much insight into the epi-
demiology and ecology of B. anthracis [42]. Lastly, the
sequence of B. anthracis Sterne 34F2 (pXO1+, pXO2 ),
a strain used in laboratories around the world, and cured
of plasmid pXO2 by Sterne in 1937, was also released in
GenBank (AE017225). It is avirulent but maintains an-
thrax toxin production and is the derivative of most live-
stock vaccines in use throughout the world [44].
B. cereus ATCC 14579 was selected for whole gen-
ome sequencing as it is non-pathogenic and is the Type
strain for the B. cereus species [45,46]. The genome se-
quence of this B. cereus ‘‘background’’ isolate was com-
pared to the B. anthracis Ames Florida sequence in an
attempt to provide a basis for whole-genome-based phy-
logenetic analysis [45]. This analysis was complemented
by the completion of B. cereus ATCC 10987, a non-
lethal dairy isolate phylogenetically more closely related
to B. anthracis than B. cereus ATCC 14579 [47]. While
the genome sequence of B. cereus ATCC 14579 did re-
veal a small previously unknown extrachromosomal lin-
ear molecule [45], B. cereus ATCC 10987 contained a
Table 1
Strains sequenced and compared in this study
Strain Alternate Source MLVA Genotype Plasmid Coveragec GenBank Reference
designation genotype group genotype Accession No.
B. anthracis Ames – Texas, Cow; plasmids cured GT62 A3.b pXO1 , pXO2 Complete AE016879 [40]
B. anthracis Ames Ancestor A0581 Texas, Cow; plasmids included GT62 A3.b pXO1+, pXO2+ Complete AE017334,
AE017336 (pXO1),
AE017335 (pXO2)
B. anthracis Ames Florida A2012 Patient, Florida, USA GT62 A3.b pXO1+, pXO2+ 12X AAAC01000001, [41]

D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329

AE011190 (pXO1),
AE011191 (pXO2)
B. anthracis Kruger B Kruger National Park, GT87 B1 pXO1+, pXO2+ 12X AAEQ00000000
South Africa
B. anthracis Western A0071 Bison; Canada GT3 A1.a pXO1+, pXO2+ 12X AAER00000000
North America
B. anthracis CNEVA-9066 A0402 Patient, Southern France GT79 B2 pXO1+, pXO2+ 12X AAEN00000000
B. anthracis A0155 – Louisiana, USA N/Ab C pXO1+, pXO2+ 12X AAEO00000000
B. anthracis Australia 94 A0039 Victoria Province, Australia GT55 A3.a pXO1+, pXO2+ 12X AAES00000000
B. anthracis Vollum A4088 – GT79 A4 pXO1+, pXO2+ 12X AAEP0000000
B. anthracis Sternea – pXO2-deficient; GT62 A3.b pXO1+, pXO2 Complete AE017225 Brettin, et al. Unpublished, 2004
the basis for animal
vaccines throughout
the world
d
B. cereus ATCC 10987 – Dairy – – pXO1 , pXO2 Complete AE017194 [47]
B. cereus ATCC 14579a – Dairy – – pXO1 , pXO2 e
Complete AE016877 [45]
f
B. cereus ZK – Zebra, Africa – – pXO1 , pXO2 Complete CP000001 Brettin, et al. Unpublished, 2004
B. cereus G9241 – Louisiana, USA – – pXO1+, pXO2 g
12X AAEK00000000 [20]
f
B. thuringiensis serovar – Severe human tissue – – pXO1 , pXO2 Complete AE017355 Brettin, et al. Unpublished, 2004
konkukian str. 97-27a necrosis, cf. [10]
a
Not sequenced at The Institute for Genomic Research, B. anthracis Sterne and Bacillus thuringiensis serovar konkukian str. 97-27 were sequenced at The Department of Energy Joint Genome
Institute, Bacillus cereus ATCC 14579 was sequenced by Integrated Genomics.
b
Not typed using genotyping schema to date.
c
Complete genomes, one contig no gaps. From GenBank – The Sterne strain of Bacillus anthracis is lacking the plasmid pXO2, resulting in an avirulent phenotype that maintains the main anthrax
toxin. Prevention of anthrax in cattle can be accomplished by vaccination with living spores of the Sterne strain. Most livestock vaccines in use throughout the world are derivatives of the live spore
vaccine formulated by Sterne in 1937 and still use descendants of his strain 34F2.
d
B. cereus ATCC 10987 harbors a large plasmid, pBc10987 (GenBank Accession No. AE017195) that is similar to pXO1, however it lacks the virulence portion that encodes the tripartite lethal
toxin and associated regulatory proteins.
e
B. cereus ATCC 14579 contains a 15-kb linear plasmid that appears to be a phage, however it does not appear to contain any large circular plasmids.
f
No plasmids were released with this genomic sequence into GenBank, however plasmids present (Table 2).
g
B. cereus G9241 contains a plasmid that is 99.6% identical to the pXO1 plasmid from B. anthracis, however it lacks the pXO2 plasmid. There is an additional plasmid that encodes a capsule
biosynthesis operon.

307
308 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329

large circular plasmid with homology to the B. anthracis
plasmid pXO1 [47].
Another group of isolates has been selected for
sequencing, as they were not identifiable as B. anthracis
using classical biochemical and microbiological tests as
detailed above. These isolates were shown to be highly
virulent and were thought to be the cause of a disease
resembling anthrax in both humans and animals. The
first of these atypical isolates to be sequenced was B. cer-
eus G9241 [20], reportedly isolated from the sputum and
blood of a welder with life-threatening pneumonia,
whose clinical presentation resembled those of reported
for 10 bioterrorism-associated inhalation anthrax pa-
tients from 2001 [17]. Interestingly, B. cereus G9241 har-
bors a plasmid with 99.6% identity to B. anthracis pXO1.
Although homologues of pXO2-encoded capsule genes
were not found, a polysaccharide capsule cluster is en-
coded on a second previously unidentified plasmid. B.
cereus Zebra Killer (ZK) was isolated from a swab of
the carcass of a dead zebra suspected of having died of
anthrax in the Etosha National Park, Namibia. This iso-
late is phylogenetically the closest completely sequenced
isolate to B. anthracis when typed by either AFLP or
MLST [31] (Fig. 1). B. thuringiensis 97-27 (subsp. konku-
kian (serotype H34)) was originally isolated from a case
of severe human tissue necrosis in a 28 year old otherwise
healthy male patient [10]. Biochemical tests and the pres-
ence of crystals in sporulation culture established the iso-
late as B. thuringiensis. The apparent pathogenic
properties of this isolate is unusual for B. thuringiensis
and unlike most B. thuringiensis strains, it is closely re-
lated to B. anthracis based on phylogenetic analysis using
AFLP [31]. The genome sequences of these three isolates
open the possibility of developing a better understanding
of their pathogenic potential when compared to B.
anthracis. These sequences may also provide a snapshot
of the closest phylogenetic ancestors to B. anthracis.
However, these strains are the exception rather than
the rule, and may instead constitute another successful
evolutionary lineage of B. cereus. Despite its importance
as a biopesticide, B. thuringiensis as a non-pathogenic
species is under-represented in the collection of whole
genome sequences. A single genome, B. thuringiensis
subsp. israelensis, has been sequenced to 8· and assem-
bled in more than 800 contigs (www.integratedgenom-
ics.com), but is not available to the scientific
community for analysis. However, the sequence of the
crystal toxin-encoding plasmid pBtoxis has been recently
published [48] and showed some similarity with B.
anthracis plasmid pXO1.
The isolates for which whole genome sequences are
now available do not properly represent the diversity
observed in the B. cereus group of organisms using
methods such as MLST (Fig. 1). The choice of strains
sequenced so far reflects a bias driven by a need to
understand rare pathogenic traits in some of these spe-
cies. Two non-pathogenic B. cereus are available to the
public and one non-pathogenic B. thuringiensis isolate
has been sequenced, while 13 pathogenic isolates have
been sequenced. The high number of B. anthracis strains

Fig. 1. Multi-locus sequence typing (MLST) neighbor-joining phylogenetic tree of representatives of the B. cereus group of organisms. All B.
anthracis isolates (ST1-3) colored red are represented as a single point on the tree. B. cereus isolates are colored black and B. thuringiensis are colored
in blue. ‘‘ST’’ following each strain designation indicates the sequence type as listed in the PubMLST database [29]. Isolates that have been sequenced
are boxed in gray.
 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329
sequenced has taken the whole genome sequencing ap-
proach to the limits of its resolution and has led to a
phylogenetic resolution of the group never achieved
for any other bacterial species [42].
4. B. anthracis genomic analysis for molecular and
forensic epidemiological purposes
The disciplines of emerging infectious disease and
forensic microbiology have shifted from using biological
phenotypes as markers to using more reliable and quan-
tifiable molecular markers, such as SNPs [49]. B. anthra-
cis, as shown above, has received considerable attention
because of its demonstrated use as a biological weapon
and the difficulties associated with forensic tracking of
this genetically homogenous group of organisms
[36,42]. In response to the threat inherent to B. anthra-
cis, vast resources have been allocated to develop diag-
nostic characters for this species through genome
sequencing. Soon after the Autumn 2001 attacks [50],
a draft genome sequence of a B. anthracis Ames Florida
strain was generated, and compared to the nearly com-
plete genome sequence of the plasmid-less B. anthracis
Ames Porton strain [41]. This analysis revealed 60 new
markers that included SNPs, indels and tandem repeats.
Only four differences were discovered between the main
chromosomes of the Florida and Porton isolates (two
SNPs and two short indels).
For the first time polymorphic markers discovered
through comparative genome sequence analysis were
used to test a collection of anthrax isolates and were able
to divide these samples into distinct subgroups [41]. This
study was critical in establishing that infectious disease
outbreaks, both naturally occurring and maliciously re-
leased (bioterrorism), could be investigated through
whole genome-based analysis. More importantly, the
study introduced statistical models allowing scientists
to discriminate true polymorphism from random
sequencing error. This was critical as in light of the rapid
generation of B. anthracis draft genome sequences. A
draft genome is by definition unfinished, and the poor
quality of portions of a draft sequence is of little use
for polymorphism discovery.
The set of polymorphisms found when these two
strains were compared was not sufficient for typing unse-
quenced strains. It was speculated that sequencing other
distantly related B. anthracis strains might yield addi-
tional polymorphisms and increase the resolution of the
method. A recent study applied this approach to a larger
number of B. anthracis draft genomes and discussed the
phylogenetic discovery bias that resulted from using
SNPs extracted from whole-genome sequence compari-
sons [42]. More than 3500 rare high quality SNPs were
discovered when the genome of five B. anthracis draft
genomes were compared to the closed B. anthracis Ames.
309
A sub-set of 990 SNPs were then typed against a panel of
26 diverse B. anthracis strains using a high-throughput
assay based on the SnaPshot primer extension protocol
(Applied Biosystems). This data demonstrated that pre-
cise phylogenetic topologies could be achieved, yielding
accurate information on internodal distances. In addi-
tion, using appropriate B. cereus outgroups, the authors
were able to determine the evolutionary root of the B.
anthracis clades. The root lies closer to a newly described
group C than either of the two previously described A
and B groups. Unequal evolutionary rates were observed
among the sequenced isolates that could be correlated
with ecological parameters and strain attributes, such
as host availability in its natural environment. The young
evolutionary characteristics of the B. anthracis group
were confirmed by the rarity of SNPs and low overall
homoplasy. Armed with this new set of markers, it is
now possible to select ‘‘canonical SNPs’’ for identifying
long branches or key phylogenetic positions [51]. A
molecular typing strategy that maximizes the use of
markers discovered through whole-genome comparison
was recently established for B. anthracis [51]. PHRANA
(Progressive hierarchical resolving assays using nucleic
acids) is a nested approach that employs canonical SNPs,
MLVA (15 loci), and simple nucleotides repeats (SNRs).
PHRANA makes use of both the low resolving power of
canonical SNPs and the high resolving power of SNRs, in
a sequential manner. This approach allows for high-reso-
lution and accurate representation of the relationships
between B. anthracis isolates by minimizing the effect of
homoplasy using a stepwise approach [51].
5. Comparison of the chromosome sequences of the
B. cereus group of organisms
5.1. Non-pathogenic B. cereus ATCC 14579 and
B. cereus ATCC 10987
The genome sequences of two non-pathogenic B. cer-
eus isolates (ATCC 14579 and ATCC 10987) have been
completed and compared to that of B. anthracis [45,47].
To better understand the degree of relatedness of mem-
bers of the B. cereus group, one should compare the
chromosome gene content and genetic structure. These
organisms demonstrate a wide range of phenotypes
and pathological effects, however these effects are often
derived from factors encoded on extra-chromosomal
elements, such as the large toxin encoding plasmid
pXO1 in B. anthracis. Sequence analysis of nine chro-
mosomal genes has suggested that B. anthracis, B. cereus
and B. thuringiensis should be considered as one species
[3]. Access to the complete genome sequence of both B.
cereus and B. anthracis isolates gave the opportunity to
revisit the question and explore the genetic content that
governs some of these specific phenotypes.
310 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329
A detailed comparative analysis of the genomes of
B. cereus and B. anthracis, revealed a small subset of
genes unique to either species, most of which are anno-
tated as hypothetical [47]. The majority of these genes
are located at the terminus of replication, indicating that
genome plasticity mostly occurs in that region, as previ-
ously observed for other microbial groups [52]. These
data suggest a history of insertion and/or deletion in
the evolution of the B. cereus group [47]. It was observed
that, in many cases, genes found at a specific position in
one genome were replaced with others at the correspond-
ing loci in another. These regions often were the result of
insertion/deletion events of mobile genetic element such
as phages, transposons, IS elements or metabolic adapta-
tions. For example, a nine-gene urease gene cluster was
identified in the genome of B. cereus ATCC 10987 that
replaces the B. anthracis/B. cereus ATCC 14579 genes
that encodes for hypothetical proteins, blasticidin S
deaminase and S layer protein. Additionally, a xylose
utilization operon was found to replace functions such
as nitrate reduction, nitrite reduction and molybdopterin
synthesis. The lack of nitrate and nitrite reduction ap-
pears to have been compensated by the acquisition of
the urease operon, allowing the bacterium to use ammo-
nia as a nitrogen source derived from urea. In addition,
B. cereus ATCC 10987 contains a unique set of genes
responsible for the transport and utilization of tagatose.
This particular strain has been isolated during a study on
cheese spoilage, where this carbohydrate can be found.
This replacement might represent another example of
metabolic adaptations to the carbohydrate-rich environ-
ment of milk [47,135]. A cluster encoding for the arginine
deiminase pathway was identified in both B. cereus
ATCC 10987 and ATCC 14579 that, like in Streptococ-
cus pyogenes, might enable B. cereus to survive in acidic
conditions [45]. On the other hand, it is thought that
ammonium inhibits receptor-mediated internalization
of the lethal toxin, hence deletion of the arginine deimin-
ase pathway in B. anthracis might be the result of an evo-
lutionary adaptation to a pathogenic lifestyle [45].
These specific evolutionary events represent niche
specific adaptations, but do not eliminate the fact that
they are highly similar and syntenic (conserved gene or-
der). An analysis using normalized blast scores to com-
pare the proteomes of B. cereus G9241, B. cereus ATCC
10987 to that of B. anthracis, demonstrated how syn-
tenic these three genomes are [47,135]. Fig. 2 represents
a similar analysis with an added level of color-coded
protein similarity and shows the high level of gene sim-
ilarity and gene order (Figs. 2(a) and (b)), with no inver-
sions or genetic rearrangements of large genome
segments. This analysis also shows that B. cereus
G9241 is more closely related to B. anthracis than to
B. cereus ATCC 10987 (Fig. 2(c)). These bioinformatic
results are mirrored in the increased pathogenicity of
B. cereus G9241.
Using a B. anthracis DNA microarray, comparative
genome hybridization (CGH) was performed with 19 di-
verse isolates from B. cereus group of organisms, includ-
ing B. weihenstephanensis [40]. CGH confirmed the
overall similarity of chromosomal genes among this
group of close relatives. Interestingly, the set of core
genes conserved across each members of the group is lo-
cated around the origin of replication, This study also
revealed six regions unique to the B. anthracis chromo-
some, comprising four phages regions and IS110 related
insertion elements. Not surprisingly, the presence of
non-toxin related pXO1 gene homologues were detected
in half the 19 strains tested and is further evidence for
the mobility of pXO1 genes within the B. cereus group
(see section below). In contrast, there were few pXO2
gene homologues found in the set of isolates tested.
Chromosomally encoded genes that may contribute
to pathogenicity, including hemolysins, phospholipases
and iron acquisition systems were identified in the gen-
omes of B. anthracis. These include genes similar to
those known to boost viral infectivity by degrading the
mucin layer of insect guts, and genes that contribute
to the virulence of the Gram-positive pathogen, Listeria
monocytogenes [40]. These genes and some newly identi-
fied surface proteins constitute potential targets for drug
and vaccine development.
Analysis of the metabolic potential of both B. anthra-
cis and B. cereus shows that these organisms have an ex-
panded capacity for amino-acid and peptide utilization,
and hence are equipped for a life in a protein rich envi-
ronment. This lifestyle was emphasized by the presence
of six amino-acid efflux systems, which prevent accumu-
lation of amino acids to bacteriostatic concentrations
during growth on peptides [40]. In addition, complex
carbohydrate degradation appears to be favored by
these organisms over small sugars. They indeed lack cat-
abolic capacity for the utilization of mannose, arabinose
and rhamnose among others, but seem to be capable of
cleavage of extracellular chitin and chitosan [40]. These
metabolic capabilities suggest that the most recent
ancestor to the B. cereus group of organisms might have
been able to thrive on animals or insects [53].
Comparative analyses suggest that major differences
between members of the B. cereus group might repre-
sent fine alteration in gene expression rather than the
level of sequence divergence or gene content
[40,45,47]. PlcR is a pleiotropic transcriptional regula-
tor that upregulates the expression of more than 100
genes in B. cereus through binding to an upstream pal-
indromic motif [45,54,55]. PlcR activity has been
shown to be regulated by the presence of a secreted
and reimported pentapetide produced from the pro-
cessing of the PapR protein C-terminus [56]. The papR
gene is itself upregulated by PlcR, forming a quorum
sensing-like system [56,57]. Proteomic-based compara-
tive analysis of B. cereus ATCC 14579 and a DPlcR
 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329 311

Fig. 2. Blast score ratio analysis of B. cereus G9241, B. cereus ATCC 10987 and B. anthracis Ames Ancestor. Blast score ratios are obtained by
dividing the Blast score for the most similar query peptide versus the reference peptide by the Blast score for the reference peptide against itself. The
Blast score ratio can then be graphically represented as a synteny plot of two genomes using the genomic location of the best hit in each genome as
the coordinates on the Cartesian plane [135]. These synteny plots demonstrate that pathogenic and non-pathogenic B. cereus group members have a
high degree of similarity as indicated by the color of each peptide, represented by a dot, and conserved gene order (synteny) represented by the dots
forming a line with a slope of 1. (a) Synteny plot of B. anthracis vs. B. cereus G9241. (b) Synteny plot of B. anthracis vs. B. cereus ATCC 10987. The
scale at the top of the figure indicates the color and Blast score ratio correlation represented in (a) and (b). A Blast score ratio of less than 0.4 is not
considered significant. (c) Blast score ratio analysis of B. cereus G9241 and B. cereus ATCC 10987 using B. anthracis Ames Ancestor as the reference.
The Blast score ratio for each of the query genomes can also be used to obtain an overall idea of similarity between the genomes compared (red –
highly conserved in all three genomes; orange – unique to the genome used as the reference; green – shared between B. anthracis Ames Ancestor and
B. cereus G9241; blue – shared between B. anthracis Ames Ancestor and B. cereus ATCC 10987).
312 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329
variant of the same strain by two-dimensional gel elec-
trophoresis has identified collagenase, phospholipases,
hemolysins, proteases and enterotoxins genes that are
PlcR upregulated [54]. Some of these activities are con-
sidered virulence factors in B. cereus and all of their
associated genes displayed a PlcR binding box in the
promoter region. Even though these genes are also
found in B. anthracis, a nonsense point mutation in
the PlcR gene is responsible for an abolition or dra-
matic reduction in their expression such that produc-
tion of lecithinase, protease and hemolysins is
undetectable [40]. A few genes encoding cytotoxin K
and non-hemolytic enterotoxin C subunit, amongst
others, have only been found in B. cereus [47]. The
acquisition of the toxin encoding pXO1 plasmid and
its regulator AtxA has been shown to be incompatible
with the chromosomally encoded PlcR [58]. Interest-
ingly, a recent study showed that a fused PlcR–PapR
construct was able to restore strong hemolytic activities
when introduced in B. anthracis [59], indicating that
these genes, while not upregulated, are still fully func-
tional and have not undergone evolutionary decay.
The loss of regulation of these chromosomally encoded
genes might represent another example of an adaptive
response of B. anthracis to a plasmid driven advanta-
geous pathogenic lifestyle [40] and because these genes
are still fully functional, it confirms the relatively
young evolutionary age of B. anthracis.
5.2. The pathogenic members of the B. cereus group of
organism, B. cereus G9241, B. cereus Zebra Killer, and
B. thuringiensis 97-27
The genome sequences of one B. thuringiensis and
two B. cereus isolates that were associated with severe
disease have been sequenced in order to gain insight into
their pathogenic potential. While one can only infer the
pathogenicity of B. cereus ZK based on its isolation
from a dead zebra, the opportunistic pathogenicity of
B. thuringiensis 97-27 has been confirmed by infection
of both immunosuppressed and immunocompetent mice
[9–11]. On the other hand, the pathogenic characteristics
of B. cereus G9241 are supported by the severe anthrax-
like clinical presentation it caused. This isolate has been
shown to be lethal in A/J mice challenged by intraperito-
neal injections [20].
The genome sequences of these isolates give us the
opportunity for comparison to the non-pathogenic B.
cereus isolate and B. anthracis. Structurally, these gen-
omes show a high level of synteny and protein identity
to B. anthracis and non-pathogenic B. cereus isolates.
Overall, the proteome of B. thuringiensis 97-27, B. cereus
ZK and B. cereus G9241 show a higher similarity to that
of B. anthracis than to that of the non-pathogenic B. cer-
eus ATCC 14579. These observations are in agreement
with the phylogenetic position of these isolates relative
to B. anthracis inferred using AFLP [31] or MLST
([28] and Fig. 1).
In common with other members of the B. cereus
group, B. cereus G9241, B. cereus ZK, and B. thuringien-
sis 97-27, share a set of genes that are associated with
virulence, such as non-hemolytic enterotoxins, chan-
nel-forming type III hemolysins, a perfringolysin O (lis-
teriolysin O), phospholipases C, and a family of
extracellular proteases [60]. Interestingly, the hemolytic
B. cereus ZK does not contain the hemolysin BL (hbl)
operon, which is a primary factor in diarrheal B. cereus
food poisoning [61]. In contrast, B. cereus G9241 con-
tains four hemolysins (A, BL, II and III), and has also
been shown to be phenotypically hemolytic [16]. In line
with other B. cereus and B. thuringiensis isolates, these
strains contain a plcR gene encoding for a full-length,
potentially functional protein and a similar set of genes
putatively upregulated by this protein, with a few excep-
tions. The histidine protein kinase (BC3528) homolo-
gous to the sporulation kinase KinB is a member of
the PlcR regulatory network in B. cereus ATCC 14579
[45] and B. cereus G9241 [20], but is absent in B. cereus
ATCC 10987 [47], B. anthracis, B. thuringiensis 97-27
and B. cereus ZK.
The genomes of these pathogenic members of the B.
cereus group also contain metabolic adaptations similar
to those found in the non-pathogenic group. For exam-
ple, like B. cereus ATCC 10987, B. cereus ZK contains a
14 kb gene replacement consisting of a gene cluster
responsible for the transport and utilization of tagatose
[47]. Similarly, B. thuringiensis 97-27, but not B. cereus
G9241 or B. cereus ZK, is able to utilize arginine
through the arginine deiminase-dependent pathway,
which is also found in B. cereus ATCC 14579 and B. cer-
eus ATCC 10987 [45,47]. These metabolic adaptations
certainly reflect the environment in which each of the
organisms thrive, and does not appear to be link to a
pathogenic characteristic of any of these isolates.
Interestingly, and relevant to its pathogenic potential,
a putative polysaccharide capsule cluster similar to the
one present on B. cereus ATCC 14579 is also present
on B. thuringiensis 97-27. While a distantly related poly-
saccharide capsule cluster is found in the same genomic
location in B. cereus ATCC 10987 no polysaccharide
capsule cluster is found on the genomes of either B.
anthracis or B. cereus ZK. Unfortunately, no experimen-
tal data are available that demonstrate B. cereus ZK or
B. thuringiensis 97-27 encapsulation in vivo or in vitro.
In contrast, B. cereus G9241 has been shown to produce
a capsule by staining with India ink [20] and a novel
polysaccharide capsule gene cluster has been identified
on one of its plasmids. However, no other identifiable
capsule biosynthetic gene cluster has been identified in
the genome sequence.
Comparative analysis of the genome of these patho-
genic isolates do not conclusively provide the informa-
 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329 313

tion necessary to assign a particular set of genes to their
virulent phenotype. Many of the differences are shared
with other non-pathogenic counterparts. It appears that
B. cereus ZK shares the most similarity with B. anthra-
cis, justifying AFLP typing data as the closest phyloge-
netically isolate to B. anthracis [31]. These isolates all
contain small and/or large plasmids, for which sequence
is now available. A detailed analysis of these plasmids
might provide a better insight into the pathogenicity of
their host isolates.
6. The B. cereus group of organisms plasmid sequence
comparisons
Historically, pathogenic potential and diverse host
range have defined the members of the B. cereus group.
The small, usually metabolic-based differences found in
the chromosomal content cannot account for the spec-
trum of disease and host range observed in this group.
However, much of the disease and host specificity in
this group can be attributed to plasmid content. The
plasmids associated with the B. cereus group of organ-
isms have broad size range (5–200 kb) and vary in
number ([62] and Table 2). Questions about their role
in the life cycles of the species that harbor them re-
main. The most striking example of the plasmid con-
tent affecting host range and pathogenesis occurs in
B. anthracis. B. anthracis harbors two plasmids, one
that encodes the tripartite lethal toxin complex,
pXO1 [22] and the other, which contains the biosyn-
thetic genes for the poly-c-D-glutamic acid capsule
[21,63]. While each plasmid is a distinct entity, it has
been shown that loss of either one results in an atten-
uated B. anthracis isolate. The role of the B. anthracis
plasmids in pathogenesis is exquisitely known, how-
ever, the function of the other plasmids in the group
is relatively unknown. However, one could expect that
these plasmids encode the peptides responsible for the
specific phenotypes that differentiate the members of
this group.
6.1. General features of the B. cereus group plasmids
identified in genome sequencing
6.1.1. B. anthracis plasmids
6.1.1.1. pXO1. The 181-kb plasmid pXO1 encodes
genes for the production and regulation of the tripar-
tite lethal toxin. The three portions of the lethal toxin
are produced separately and assembled. The eukaryotic
membrane-interacting protective antigen (PA) is ex-
pressed as a pre-protein and proteolytically processed
into its active form, known to elicit a protective im-
mune response against anthrax [64]. The toxin has
two additional components – the lethal factor (LF)
and the edema factor (EF), which are responsible for
the proteolytic cleavage of several mitogen-activated
protein kinases (MAPKKs) and convert intracellular
ATP into cAMP, respectively. The biological activities
of these subunits are not fully characterized but result
in aberrant signaling inside the macrophages and fluid
accumulation in the lung [65]. All three subunits are re-
quired for active toxin production and activity (for re-
views see [66–70]). The three structural genes, cya (EF),
lef (LF) and pagA (PA), are under the control of at
least two regulatory elements, AtxA and PagR [71],
and have been shown to be expressed early in the
growth of B. anthracis [72]. These genes are encoded
within a 44.5-kb pathogenicity island that is transposi-
tionally active [19]. However, its inversion does not
appear to affect virulence [41].

Table 2
Comparison of plasmid sequences from plasmids identified in genome sequencing projects
Plasmid name Organism Strain Size (bp) G+C (%) Replicon Replication Number Functional Hypothetical
system of CDS CDS CDSa
pXO1 B. anthracis Allb 181,677 33 Unknown Unknown 204 48 156
pXO2 B. anthracis Allb 94,830 33 pAMb1 Theta 104 26 78
pBc10987 B. cereus ATCC 10987 208,369 33 Unknown Unknown 242 91 151
pBClin15 B. cereus ATCC 14579 15,100 38 Phage Phage 21 4 17
pBCXO1c B. cereus G9241 191,110F 33 Unknown Unknown 177 61 116
pBC218d B. cereus G9241 218,094F 32 Unknown Unknown 185 116 69
pZK5 B. cereus Zebra Killer 5108 31 Group VII RCR 5 2 3
pZK8 B. cereus Zebra Killer 8191 32 Group VII RCR 8 4 4
pZK9 B. cereus Zebra Killer 9150 31 Group VII RCR 10 5 5
pZK54 B. cereus Zebra Killer 53,501 32 pAMb1 Theta 58 27 31
pZK467 B. cereus Zebra Killer 466,370 33 Unknown Unknown 465 228 237
pBT9727 B. thuringiensis 9727 77,122 33 pAMb1 Theta 80 18 62
a
Hypothetical includes the proteins with no know function including CDS with hypothetical, conserved hypothetical and Bacillus cereus group-
specific protein designations.
b
B. anthracis pXO1 from GenBank Accession No. AE017336. B. anthracis pXO2 from GenBank Accession No. AE017335.
c
GenBank Accession No. AAEK01000020, AAEK01000036 and AAEK01000038 (not currently closed molecules).
d
GenBank Accession No. AAEK01000004.
314 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329
Within the isolates of B. anthracis for which the gen-
ome sequence is available, the pXO1 sequences are
highly similar. Apart from a small number of SNPs,
SNRs and VNTRs, no large insertions or deletions have
been observed [41]. To date only one B. anthracis isolate
has been identified without the pXO1 plasmid [42]. This
isolate represents a new lineage in the B. anthracis
group. However, little information is available for this
isolate and it is possible that pXO1 might have been lost
during laboratory passages.
The level of sequence coverage obtained for pXO1
from whole-genome shotgun sequencing projects can
be used to estimate the molecular ratio of plasmid to
chromosome [41]. For pXO1, the copy number has been
suggested to be 2–3 copies per chromosome copy ([41]
and this study). This estimate appears to be much higher
than indicated by other studies [73], and certainly repre-
sents a snapshot of the dynamic plasmid content in the
bacterium at the time of genomic DNA preparation.
The origin of replication of pXO1 has never been
conclusively identified [22,74]. It has been suggested that
it lies within an 11-kb region discovered by subcloning
[22,75], however this region does not contain any genes
similar to other known plasmid replication systems. This
is a recurring theme among the large plasmids of the
B. cereus group and is addressed further below. While
many studies have focused on the toxin and its regula-
tors, five years after the complete sequence of pXO1
was first published [22], 148 of the 204 potential coding
sequences (72.5%) remain functionally unidentified. De-
tailed transcriptional mapping studies should be able to
refine the annotation and in the process identify novel
genes involved in the pathogenicity of the organism in
the future.
6.1.1.2. pXO2. pXO2 (96 kb) encodes for the synthesis
and degradation of another well-known virulence factor,
the poly-c-D-glutamic acid capsule [76]. This capsular
material is thought to protect the vegetative bacterial
cells during transit through the macrophage, hence
evading the host immune response and increasing sys-
temic sepsis [77]. The poly-c-D-glutamic acid capsule is
produced by a three-gene operon capABC under the
control of a number of regulators including the toxin
regulator AtxA encoded on pXO1 [71], and the pXO2
encoded regulators AcpA and AcpB. Degradation of
the capsular material, by CapD [78], encoded on
pXO2, results in high and low molecular weight forms
of the capsule, both of which essential for infection
[79]. B. anthracis Sterne lacks pXO2, and is commonly
used around the world as an anthrax veterinary vaccine.
Very little is known about the molecular mechanism for
the transport of the capsular material to the surface of
the bacterial cell.
The first complete sequence of pXO2 was obtained
from a B. anthracis Pasteur strain (pXO1 , pXO2+)
[21] and encodes 104 genes for which 78 do not have
functional assignments (78/104, 75%). In contrast to
pXO1, the pXO2 replication region has been identified
and, recently, a 60-nucleotide region essential for the ini-
tiation of plasmid replication has been characterized
[80]. pXO2 is a theta replicating plasmid similar to the
prototypical pAMb1 plasmid from Enterococcus fea-
calis. Detailed analysis of the replication machinery is
provided below. Several studies have indicated that
pXO2 sequences are not widely distributed among clo-
sely related species based on CGH, hybridization and
PCR experiments [40,63]. Only one isolate, B. thuringi-
ensis AW06, showed any significant sequence similarity
to pXO2 in previous studies, however the B. thuringien-
sis plasmid pBT9727 described in this work has signifi-
cant similarity to pXO2 (see below).
6.1.2. B. cereus plasmids
In contrast to the conserved plasmid content ob-
served among B. anthracis, B. cereus isolates contain a
diverse range of plasmids – no strains have yet been
identified with identical plasmid content. B. cereus plas-
mids vary from 5- to almost 500 kb in size and only a
limited number have been implicated in pathogenesis
(Table 2). A subset do not encode for any obvious phe-
notypes and may be considered cryptic. B. cereus gen-
ome projects have identified a large number of
plasmids, and their analysis has revealed a number of
conserved regions within the large plasmids group.
The smaller plasmids, in contrast, show greater similar-
ity to B. thuringiensis smaller plasmids.
6.1.2.1. pBc10987. A single 208-kb plasmid, named
pBc10987, was identified from the non-pathogenic iso-
late B. cereus ATCC 10987 [47]. pBc10987 shows surpris-
ing similarity to the plasmid pXO1 of B. anthracis (40%
nucleotide identity), however it lacks the pathogenicity
island (PI) containing the genes that encode for the tri-
partite lethal toxin and its associated regulators [47]. In
lieu of the pXO1 PI, pBc10987 contains genes potentially
involved in adaptation to either an environmental or
pathogenic lifestyle. Environmental adaptations of
pBc10987 include a copper requiring tyrosinase, arsenite
resistance and its associated regulators as well as an
amino acid transport system. The pathogenic adapta-
tions comprise of two potential novel toxins [22,47]. Un-
like the PI in pXO1, this region in pBc10987 is not
flanked by mobile genetic elements, suggesting that it
has not been lost from pBc10987 [22,47]. These findings
suggest that in conjunction with the chromosome, B. cer-
eus group plasmids contribute to the metabolic fitness of
the species.
In addition to these potential metabolic or patho-
genic adaptations, a number of aspects of plasmid biol-
ogy and chromosome-plasmid crosstalk are conserved
between pXO1 and pBc10987. Like pXO1, the replica-
 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329
tion machinery of pBc10987 has not been identified
using bioinformatics tools, however the similar plasmid
copy number (1–3 copies per cell) suggests that the
replication mechanism is conserved between the two.
Divergent copies of abrB, a pleiotropic transition state
regulator, are present on pXO1 and pBc10987, as well
as the chromosomes of B. anthracis and B. cereus ATCC
10987. AbrB has been shown in B. subtilis to modulate
the switch between biofilm formation and sporulation
[81], as well as regulating competence [82,83]. In
B. anthracis, AbrB has been shown to negatively regu-
late toxin production [84,85]. While pBc10987 and
pXO1 AbrB proteins are similar, they differ significantly
at their N-terminus, in that the pXO1 protein is 27 ami-
no acids shorter rendering it potentially inactive [84].
The role of the pBc10987 encoded AbrB protein is cur-
rently unknown, but it is speculated that it acts as a reg-
ulator of the genes replacing the pXO1 pathogenicity
island [47].
Besides AbrB, there are two more examples of possi-
ble genetic exchange between the chromosome and the
plasmid of the B. cereus ATCC 10987. Identical copies
of a transposable element similar to the S. aureus
Tn554 [86] are present on both the chromosome and
plasmid. There are four genes associated with the
Tn554-like element and one, bclA, is of interest. In B.
anthracis, multiple functions have been attributed to
BclA. It is the major spore surface antigenic protein
[87,88] and has been shown to play a role in the depth
determination of the spore coat [89]. Divergent proteins
of the same family as BlcA have been found on the chro-
mosomes of B. anthracis and some plasmids of B. thur-
ingiensis [90]. Interestingly, the bclA gene in B. anthracis
is limited to the chromosome, whereas in B. cereus
ATCC 10987 similar copies of bclA are present on the
chromosome and the plasmid. The presence of similar
genes on other plasmids suggests that genetic exchange
occurred between plasmids and chromosomes in the
B. cereus group of organisms.
6.1.2.2. pBCXO1 and pBC218. B. cereus G9241 has
been associated with an illness resembling inhalation
anthrax [20]. Initial CGH studies revealed that B.
cereus G9241 contained genes with a high degree of
hybridization to pXO1, including all three anthrax
toxin genes, however, no hybridization was observed
to pXO2. Unexpectedly, a 191-kb plasmid with a
high degree of similarity and synteny to B. anthracis
pXO1 was found in B. cereus G9241 [20]. In addi-
tion, B. cereus G9241 contains a second 218-kb plas-
mid, previously unidentified, that encodes a novel
polysaccharide capsule biosynthetic cluster. These
plasmids are expected to contribute greatly to the ob-
served anthrax-like clinical presentation.
The pXO1-like plasmid, named pBCXO1, is 99.6%
identical to pXO1. This similarity extends to the pXO1
315
virulence genes, protective antigen (99.7% amino acid
identity), lethal factor (99% amino acid identity) and
edema factor (96% amino acid identity), as well as the
known virulence regulatory proteins AtxA (100% amino
acid identity) and PagR (98.6% amino acid identity)
[20]. Price et al. [91] have previously used minor varia-
tions in the protective antigen (PA) protein as a classifi-
cation system. Using this system, pBCXO1-encoded PA
is most similar to genotype V, often associated with the
western North America diversity group of B. anthracis
[20]. ELISA experiments demonstrated that PA was
present in the supernatant of a stationary phase culture,
however it is unclear which protective antigen subunit
was recognized by this assay [20]. This high level of iden-
tity between pXO1 and pBCXO1 suggests that both iso-
lates acquired their plasmid from a common ancestor or
that transfer occurred into B. cereus G9241 from B.
anthracis. The successful transfer of pXO1 from B.
anthracis to a close relative with the aid of a mobilizing
plasmid [92] supports the latter.
The regulators, PlcR and AtxA, have been shown to
be functionally incompatible in B. anthracis. The intro-
duction of pXO1-encoded AtxA is thought to have led
to the selection for the plcR nonsense mutation in
B. anthracis [58]. Interestingly, B. cereus G9241 appears
to encode fully functional copies of both PlcR and
AtxA. However, PlcR contains a region of lower simi-
larity to B. anthracis PlcR that is thought to compensate
for the presence of AtxA or represent another poten-
tially inactive form of PlcR. It is also hypothesized that
the pBCXO1 plasmid may have been acquired recently
and the incompatibility has not had time to generate a
nonsense mutation. In light of the discovery of this
important organism, functional studies of B. cereus
G9241 regulatory pathways must be performed to deve-
lop a better understanding of this evolutionary event.
B. cereus G9241 contains a second plasmid, pBC218
(218 kb), with limited similarity to a region in
pZK467 (Fig. 3(b)). Intriguingly, pBC218 carries a sec-
ond copy of atxA, however the gene product is only
78% identical to the B. anthracis homologue [20]. In
addition, pBC218 encodes some of the known B. anthra-
cis virulence factors [20]. Genes encoding for a homo-
logue of the protective antigen peptide (60% amino
acid identity) and the lethal factor (36% amino acid
identity) are also found on pBC218, however the plas-
mid does not contain a homologue of the edema factor.
The lethal factor is significantly truncated when com-
pared to that of B. anthracis pXO1 and is probably
not functional. In contrast, the pBC218-encoded protec-
tive antigen peptide is identical to B. anthracis at all 10
dominant negative amino acid residues identified by
Mourez et al. [93]. Another 33 residues were identified
that resulted in decreased activity when substituted by
cysteine, of which 27 are identical in pBC218-encoded
protective antigen, with the remaining six being
316 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329
Fig. 3. Identification of regions with proteomic and syntenic similar-
ities by comparisons of selected B. cereus group plasmids. Blast score
ratio analyses were used to compare the putative plasmid proteomes as
described in Fig. 2. (a) Synteny plot of pBT9727 vs pXO2. The
pBT9727 plasmid contains a similar backbone structure to pXO2. The
replication proteins are conserved in plasmid location and peptide
sequence however pBT9727 lacks the poly-c-D-glutamic acid capsule
biosynthetic genes, indicated by the gray box. An additional unique
region of 10 kb is identified in pBT9727 by the gray box (22–33 kb).
48 of the 80 – pBT9727 peptides have homologues in pXO2.
(b) Comparison of pBC218 and pZK467 demonstrates that these
plasmids have limited similarity with the exception of one region,
highlighted in the gray box. This region is conserved in both gene
content and gene order.
conservative substitutions [20,93]. The presence of these
additional toxin subunits is raising some interesting
questions about their putative function in the pathogen-
esis of this organism. Could the pBC218-encoded PA
form a complex with the pBCXO1 lethal factor and ede-
ma factor to form what would appear to be novel func-
tional anthrax toxin? How do these fragments interact
and are they coordinately regulated?
India ink staining and microbiological analyses dem-
onstrated that B. cereus G9241 is encapsulated and that
capsule production is not regulated by increased CO2
concentrations like the B. anthracis pXO2-encoded
poly-c-D-glutamic acid capsule. Interestingly, It was pre-
viously thought that B. cereus, as a species in the group,
did not produce capsule [6]. Further examination of the
pBC218 sequence revealed a putative polysaccharide
biosynthetic operon, representing the only capsule bio-
synthetic cluster identified in the genome of B. cereus
G9241. This gene cluster is thought to be responsible
for the capsule produced by this isolate. The lack of
any other capsule suggests that this plasmid-encoded
polysaccharide capsule might compensate for the lack
of poly-c-D-glutamic acid capsule and help B. cereus
G9241 to evade the host immune system.
The plasmid content of this isolate represents a com-
posite of B. cereus and B. anthracis. It also raises inter-
esting questions regarding the regulatory interaction
between AtxA and PlcR, the assembly of the toxin
and the role of the novel capsule in pathogenesis.
6.1.2.3. pBClin15. A 15-kb linear plasmid, pBClin15,
was identified by genome sequencing of B. cereus ATCC
14579 [45]. This plasmid contains none of the typical
replication machinery associated with small plasmids
identified in either B. cereus or B. thuringiensis, but does
contain genes that suggest that it may represent a novel
phage. Recently, it has been reported that pBClin15
shows some level of similarity to the Bam35 phage of
B. thuringiensis var. alesti strain 35 [94,95]. Additional
comparisons and experimental data suggest that
Bam35 can exist both as a prophage and a free phage
[95]. In contrast to Bam35, pBClin15 lacks terminal in-
verted repeats that may allow protein-primed replica-
tion, an observation that led to the suggestion that
pBClin15 is a degenerate phage [95].
6.1.3. B. cereus Zebra Killer (ZK) plasmids
The plasmids from B. cereus ZK represent a novel set
of plasmids carried by an isolate that causes a disease
resembling anthrax in animals (P.C.B. Turnbull, per-
sonal communication). Five distinct plasmids were iden-
tified in the genome of B. cereus ZK, representing the
largest number of plasmids harbored by a strain of
B. cereus so far. Interestingly, the three small plasmids
(less than 10 kb, Table 2) pZK5 (CP000041), pZK8
(CP000043) and pZK9 (CP000044) are more similar to
known B. thuringiensis plasmids than any previously se-
quenced B. cereus plasmids, whereas the larger two plas-
mids, pZK54 (CP000042) and pZK467 (CP000040),
share characteristics with pXO2 and pBC218, respec-
tively. The three small plasmids are predicted to repli-
cate through a rolling-circle mechanism; their
functional annotation is mostly limited to the identifica-
tion of replication and mobilization proteins. pZK54
(54 kb) encodes a pXO2-like putative theta replication
system, however no identifiable replication system was
found in pZK467 (466 kb), as in pXO1 [17,76].
pZK467 and pBC218 share an approximately 40-kb re-
gion where synteny is retained, however, no functional
 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329
assignments exist in this region (Fig. 3(b)). pZK467 is
the only known plasmid with similarity to pBC218. A
large number of transposase genes or mobile elements
have been identified in these two large plasmids. These
elements may facilitate gene exchange between plasmids
as well as between plasmids and the chromosome. None
of the B. cereus ZK plasmids encode homologues of
known virulence factors in B. anthracis, B. cereus or B.
thuringiensis [6]. Sequence analysis of B. cereus ZK plas-
mids does not improve our understanding of the patho-
genicity of this organism, which currently remains
unclear. Analysis of the relative copy number of these
plasmids obtained from sequence analysis indicates that
each has a copy number of 1–2, except for pZK5 which
is estimated at 0.6. These data might suggest that
pZK5 is unstable under the laboratory culture condi-
tions employed and is in the process of being lost.
6.1.4. B. thuringiensis plasmids
Phenotypic characterization of B. thuringiensis is of-
ten based on the presence of plasmid-encoded large
Cry protein inclusions of the d-endotoxin [48]. Recently,
a study by Andrup et al. [90] described six B. thuringien-
sis plasmids (less than 20 kb) that did not carry the cry
gene. It is possible to group B. thuringiensis plasmids
based on the similarity of their replication and mobiliza-
tion machinery, however this system did not extend to
the larger B. thuringiensis or B. cereus plasmids.
6.1.4.1. pBtoxis. In addition to the four known Cry and
two known Cyt toxins, the 128-kb circular plasmid,
pBtoxis encodes a third Cyt-type sequence with an addi-
tional C-terminal domain previously unseen in such pro-
teins [48]. B. thuringiensis subsp. israelensis carries
pBtoxis, and its toxin crystals have been demonstrated
to be one of the most toxic combinations tested [96].
However, it is unclear if this toxicity is related to the spe-
cific combination of toxins or to toxin interaction with
other plasmid encoded features. GC skew analysis indi-
cated a putative origin of replication, however no repli-
cation proteins has been identified, similar to the other
large B. cereus group plasmids (Table 2). The coding se-
quences adjacent to this region in pBt001 showed >78%
amino acid identity to pXO1-49, and is located near a
similar putative replication origin on pXO1, also pre-
dicted by GC skew analysis. However, experimental
functional evidence is not available.
In addition to the toxin genes, pBtoxis encodes a
number of genes that are thought to enhance crystal for-
mation and subsequent cell viability by acting as chaper-
ones. Interestingly, like pBC10987 and pXO1, pBtoxis
encodes peptides that are involved in host sporulation
and germination [21,47]. The exact role of these proteins
is unknown, however in B. anthracis the lack of these
genes on pXO1 is detrimental [97]. One of the more sur-
prising findings in pBtoxis is the presence of a set of
317
genes potentially involved in the biosynthesis and export
of a cyclic peptide antibiotic similar to Enterococcus fae-
calis AS-48 [48]. One can speculate that this peptide
could provide B. thuringiensis with a competitive advan-
tage in the environment or act as a signalling molecule.
6.1.4.2. pBT9727. B. thuringiensis 97-27 has been shown
to produce crystal protein in sporulated culture by di-
rect microscopic examination [54]. Detailed sequence
analysis of pBT9727 (CP000047), the sole plasmid
found in B. thuringiensis 97-27, did not reveal any
genes encoding for the Cry toxin with similarity to al-
ready known B. thuringiensis toxin genes. Interestingly,
pBT9727 shows similarity to B. anthracis pXO2. Com-
parison of the predicted coding regions of the two plas-
mids revealed that pBT9727 shares 89% (82/92) of its
putative coding sequences with pXO2. Their replication
proteins are almost identical and the predicted origin
of replication is well conserved (Fig. 4, cf. [80]). The le-
vel of protein similarity, combined with the conserva-
tion of gene order, suggests that these plasmids might
have diverged recently. Like pBC10987 and pXO1;
pBT9727 and pXO2 share a common backbone (Fig.
3(a)). The pXO2 region encoding for the poly-c-D-glu-
tamic acid capsule biosynthetic genes is replaced on
pBT9727 with genes encoding hypothetical proteins
and putative mobile elements. This replacement sug-
gests that pBT9727 might have evolved to fulfil other
functions than providing this isolate with capsule bio-
synthetic genes.
6.2. Replication and mobility mechanisms of the B. cereus
group plasmids
6.2.1. Unidentified replicons
The following plasmids have no identified replication
system: pXO1, pBCXO1, pBC10987, pBC218, pBtoxis
and pZK467 (Table 2). A number of hypotheses have
been put forward to account for this lack of clearly iden-
tifiable replication machinery, including that these plas-
mids carry a novel replication mechanism [47]. Attempts
have been made to identify the regions involved in the
replication machinery of pXO1 using subcloning. Kas-
par and Robertson [75] identified an 11 kb region that
is thought to play a role in pXO1 replication. However,
this region does not encode genes with similarity to any
other plasmid replication system known. One gene, a
type I DNA topoisomerase was identified by compari-
son of pBC10987 (BCEA0140, [47]) and pXO1
(BXA0213, [41]). This protein appears to be conserved
in the following large B. cereus plasmids with an uniden-
tified replication system pBc10987, pOX1 and pBCXO1,
and is also found in pZK467 but where it is thought to
be non-functional as it is in three gene fragments
(pZK467_319-321). No homologue to this peptide is
present in pBC218 or pBtoxis. In pBc10987, this type I
318 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329

Fig. 4. Identification of the origin of replication structure in theta replicating plasmids of B. cereus group. (a) Alignment of the 60 nucleotide region
identified by Tinsley et al. [80] as the origin of replication. The gray box highlights a conserved region containing the origin, which is identified by the
arrow. Other conserved nucleotides in the region are indicated with an ‘‘*’’. The plasmids are from Bacillus anthracis pXO2 (GenBank Accession No.
AE017335), B. thuringiensis pBT9727 (GenBank Accession No. AE017335), Bacillus thuringiensis pAW63 (GenBank Accession No. AJ011655) and
Enterococcus faecalis pAMb1 (GenBank Accession No. AF007787). (b) The pXO2 replication region. The B. anthracis pXO2 iteron-binding region is
between the copy number control gene parA and the replication gene repS. Arrows on the nucleotide sequence identify the iteron-binding repeat,
‘‘ATGTGTAA’’. There are 10 direct repeats in the forward orientation and three in the reverse orientation in pXO2. There are also other degenerate
repeats in the region that are not indicated. The structure of the replication region including a similar repeated region is present in B. thuringiensis
pBT9727.

DNA topoisomerase may associate with a putative
DNA polymerase III b-subunit thought to increase the
processivity of replication [98]. Unfortunately, this sec-
ond gene is not present in any of the other large B. cer-
eus group plasmids. It may participate in pBc10987
replication but does not appear to be essential for the
replication of all members of this group of plasmids. An-
other plasmid-borne replication gene candidate is a
host-factor-like protein (BCEA0146 – pBc10987;
BXA0206 – pXO1; pZK467_0115 – pZK467, ORFBT116
– pBtoxis). These peptides have nucleotide binding do-
mains and are similar in all large B. cereus group plas-
mids lacking an identifiable replication system. While
it is unlikely that this peptide accounts for the entire rep-
lication machinery, it is a conserved plasmid-encoded
peptide.
One alternate hypothesis is that B. cereus plasmids
with no identifiable replication system are actually
the result of reduction of an ancestral chromosome.
Many of the genes found on these plasmids have sig-
nificant similarity with chromosomally encoded genes
from other members of the B. cereus group, B. subtilis,
Bacillus halodurans and other low G+C Gram-positive
species such as Listeria and Staphylococcus. Addition-
ally, the GC skew, gene organization and gene orienta-
tion bias of these large plasmids appear more similar
to that of chromosomes. In line with the chromosomal
reduction theory, one can speculate that the replication
machinery for these plasmids is actually chromosom-
ally encoded and only the actual origin of replication
is present on the plasmids themselves. Recent work
has identified a chromosomally encoded helicase in
the B. cereus group of organisms similar to that of
Staphylococcus aureus plasmid pT181 [99,100]. This
helicase, PcrA, functions as a nickase and can initiate
replication of pT181. The B. anthracis PcrA has been
shown to function as a helicase and initiate replication
of pT181. It is thought that the B. cereus homologues
 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329 319

could perform similar function on plasmids that do

not have a clearly recognizable replication system.
Additionally, such a system would explain why the
replication regions of these plasmids have gone
undiscovered.

6.2.2. Theta replicons

Based on similarity to other replication proteins
identified from plasmids in the B. cereus group and
other Gram-positive organisms, the origin of
replication was readily identified in pXO2 [101,102].
While well characterized in B. thuringiensis plasmid
pAW63, the pXO2 origin of replication was not func-
tionally characterized until recently [80]. Tinsley et al.
[80] demonstrated that the minimal replicon comprises
solely of the RepS peptide (BXB0039) and the origin
of replication. In addition, the functional origin of
replication was shown to be limited to a 60 bp region
to which RepS specifically binds (Fig. 4(a)). Further-
more, RepS was shown to bind to the single stranded
form of this region, providing further evidence that
RepS is responsible for the initiation of replication.
pBT9727 origin of replication (59/60 nucleotides) and
RepS are similar (91% amino acid identity) to that
of pXO2 indicating that pBT9727 belongs to the theta
replicating plasmid family. On the other hand, plas-
mid pZK54 is tenuously categorized as a theta repli-
cating plasmid based on similarity of the replication
associated protein, pZK54_001. pZK54 lacks the Fig. 5. Comparison and clustering of B. cereus group plasmids based
on replication and mobilization proteins. Replication or mobilization
mobilization protein and the 60 nt origin of replica- proteins, as identified by annotation, were aligned and compared using
tion identified by Tinsley et al. [80], suggesting that CLUSTALW. The unrooted neighbor-joining phylogenetic trees were
a putative novel mechanism might be involved in generated amd displayed with Tree View. Theta-replicating plasmid
pZK54 maintenance and mobility. proteins are within the blue ovals, Rolling circle replication plasmids of
All theta replication proteins from B. cereus group group IV are in the green ovals and only two members of the group III
family replication proteins could be identified for inclusion.
plasmids share a significant level of similarity and cluster (a) Replication proteins. (b) Mobilization proteins.
together phylogenetically suggesting a common ances-
tral origin (Fig. 5(a)).
6.2.4. Plasmid mobility
While the B. anthracis plasmids have not been di-
6.2.3. Rolling circle replicons rectly shown to be self-transmissible, previous reports
All B. cereus group small plasmids (<10 kb) appear to have demonstrated that some of them can be mobilized
replicate through a rolling circle mechanism. Sequence with the help of conjugative plasmids [74]. B. thuringien-
analysis and comparison of all B. cereus replication pro- sis subsp. israeliensis pXO16 is an example of such a
teins, including theta and rolling circle, allow for cluster- conjugative plasmid [74,103]. Interestingly, no transfer
ing of these plasmids into distinct groups (Fig. 5). This or mobilization regions are identifiable in the sequence
comparison includes type sequences for each plasmid of B. cereus large plasmids. In contrast, the mobilization
group as determined by Andrup et al. [90]. All small proteins encoded on the smaller B. cereus ZK and other
rolling circle plasmid replication proteins clustered B. thuringiensis plasmids suggests that they may be self-
together and are clearly distinct from theta replication mobilizable but, lack the ability to create pores to trans-
protein sequences. pZK8 and pZK9 both clustered with fer themselves to the recipient cells. It is possible that the
group VII plasmids as defined previously [90]. pZK5 is combination of the mobilization capabilities of one plas-
also a group VII member but might represent an outlier mid and an unidentified membrane associated transfer
to this group. Classification of the small rolling circle system from another plasmid will allow transfer of
replicating plasmids by this sequence analysis is a robust both/all plasmids from the donor cell to a willing recipi-
and effective way to identify the mode of replication for ent cell. The B. cereus large plasmids harbor candidate
such plasmids. pore formation genes, such as the TraD/G conjugation
320 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329

proteins, that could form a pore through a membrane
complex however their functionality has not been dem-
onstrated experimentally. Sequencing of mobilizable B.
cereus group plasmids such as pOX11 [92], pXO12 [92]
or pXO16 [74,104] would advance our understanding
of plasmid transfer mechanisms in this group of organ-
isms. A number of limited studies have been undertaken
that demonstrate that B. cereus group plasmid transfer
is not affected by DNase, involves membrane interaction
[74] and in some cases employs an Ôaggregation sub-
stanceÕ [104], all which suggest a conjugative transfer
mechanism.
Based on sequence analysis of the plasmids of the
B. cereus group of organisms, it is evident in some cases
that the significant differences observed in pathogenicity
and host range are often dictated by genes carried on
plasmids. As suggested for Yersina pestis [105], where
pathogenicity and host range is determined by the plas-
mid content of an isolate, the designation of ‘‘plasmid-
ovar’’ would be applicable to the B. cereus group of
organisms, such that B. anthracis would be B. cereus pla-
smidovar anthracis.
7. The phage of the B. cereus group
Besides plasmids, bacteriophages are another impor-
tant source of gene flow in bacteria. Bacteriophages are
viruses that infect bacteria. Bacteriophage can be either
lytic, redirecting cellular processes for the sole purpose
of producing additional virus progeny, or temperate,
insetting themselves into the host bacterium genome
where they are transmitted ‘‘benignly’’ from generation
to generation in concert with a host replicon. Bacterio-
phages have an extremely high degree of host specificity
and may integrate as a prophage into a unique site or
multiple locations in the host genome. This specificity
has been exploited as typing feature [106] and proposed
as a possible therapeutic intervention in the treatment of
bacterial diseases [107,108]. As an increasing number of
bacterial genomes are being sequenced, more prophage
are being identified and the distribution and diversity
of bacteriophage is beginning to be appreciated (for re-
cent reviews, see [109,110]).
Genes carried by bacteriophage encode proteins that
modulate lysogenic conversion of the host and may pro-
vide a selective advantage [111]. The prophage may re-
main competent to enter a lytic cycle in response to
appropriate induction signals or may remain quiescent
for generations. Over time, the prophage may acquire
mutations that preclude their re-entry into a lytic cycle
of growth, while genes originally carried by the virus
can remain active in the host extending the lysogens
niche over non-lysogenic isogenic strains. In fact, there
are several examples of phage-derived sequences repre-
senting in excess of 10% of a sequenced bacterial gen-
ome [109]. Phage can also facilitate horizontal gene
transfer and promote genomic rearrangements, a factor
that has contributed to the emergence of bacterial
pathogens (for review, see [112]). This is most clearly
demonstrated in a comparison of the genomes of two
E. coli stains, the benign K12 and the virulent
O157:H7 where significant genetic differences between
these two strains can be attributed to differences in pro-
phage content [113].
While the best-characterized phage in the Gram-
positive Bacillus genera are from B. subtilis, similar
prophages have been identified in all members of the
B. cereus group. We recognize that the identification
of prophage in bacterial genomes is non-trivial
[109,110]. Consequently, the current estimates of the
prophage contribution to the genomic content of the
members of the B. cereus group should be regarded as
conservative and tentative.
7.1. Genomic phage content
Each of the 10 B. anthracis genome sequences contains
four prophages located in the same genomic location.
These prophages have been designated LambdaBa01-04
[40]. The B. anthracis prophages appear to be unique to
B. anthracis, however, prophage LambdaBa01 is at least
partially present in B. cereus ZK (Fig. 6). Interestingly,
while the gene content appears to be conserved the

Fig. 6. Blast similarities represented using Blast score ratios and visualized with TIGR MEV (multiple expression viewer). Each vertical line
represents a unique gene in one of the predicted B. anthracis prophage and each horizontal column represents a different isolate. Yellow regions
indicate similarity, whereas the black regions are dissimilar. This figure demonstrates that only B. cereus Zebra Killer contains one of the phage from
B. anthracis. The B. anthracis phage are not generally conserved among this group and neither are the B. cereus or B. thuringiensis phage (data not
shown).
 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329
relative genomic location is different. Whether these dif-
ferences represent a subtle variation in species-specific
phage integration sites, or provide an evolutionary in-
sight into the relationship between B. cereus ZK and
the B. anthracis are important questions that have not
yet been resolved.
Three prophages have been identified in the genome
sequence B. cereus ATCC 10987, while B. cereus
ATCC 14579 contains six putative integrated pro-
phages and a linear non-integrated phage, designated
pBClin15. B. cereus G9241 contains a not well-charac-
terized cryptic phage of 29,886 bp (pBClin29, [20]) that
encodes phage-like proteins and a plasmid replicon
similar to B. anthracis pXO2. Interestingly, the B. thur-
ingiensis 97-27 genome appears to contain as many as
10 prophages on the chromosome and two prophage
on the pBT9727 plasmid. Including the B. anthracis
LambaBa01-like prophage, B. cereus ZK genome may
have as many as 27 chromosomally encoded pro-
phages, based on the occurrence of distributed phage
genes, and an additional nine in plasmid pZK467,
and four in pZK54. None of the putative prophage
integration sites in B. thuringiensis 97-27 or B. cereus
ZK appear to be associated with tRNA genes as is of-
ten the case [114]. These high numbers of prophages
may be an overestimate as some consist of only a
few genes with some degree of similarity to other
phages.
7.2. Lytic induction
An important test of prophage competence is the
ability to initiate a lytic cycle by induction. We have
demonstrated the competence of several prophage to
induce a lytic cycle using a standard mitomycin C
treatment from a number of B. cereus and B. thuringi-
ensis strains (Rasko, unpublished). Not surprisingly, it
was shown that the host-range of these phages is re-
stricted to the species or strain from which they were
obtained. However, there is one notable exception,
strain B. cereus ATCC 4342 can be used to propagate
the Gamma and Cherry phages that were previously
thought to be limited to B. anthracis strains [106,115].
The factors responsible for the expansion of the phage
range are currently under investigation. Finally, it is
interesting to note that typing with MLST places B.
cereus ATCC 4342 in the phylogenetic clade that con-
tains B. anthracis and may represent a close evolution-
ary relationship (Fig. 1).
7.3. Exploitation of phage
Given the high degree of specificity and limited host
range of bacteriophage, it seems quite reasonable to
use them and their encoded activities as a means of
strain identification and infection control.
321
7.3.1. Use as a typing tool for B. anthracis (Gamma and
Cherry phage)
While there are four prophages in each of the B.
anthracis genomes sequenced to date, clinical laborato-
ries utilize, in combination with a number of biochemi-
cal tests, cell lysis with Gamma phage as a phenotypic
characteristic in the identification of B. anthracis [106].
Sensitivity to Gamma phage remains highly specific to
B. anthracis, however B. cereus ATCC 4342 is also sen-
sitive to Gamma phage. As more isolates closely related
to B. anthracis are being isolated with pathogenic char-
acteristics, Gamma phage sensitivity among B. cereus
isolates will become increasingly common and will re-
duce the discriminating power of this assay to differenti-
ate B. anthracis from B. cereus.
7.3.2. Use of phage lysis protein as a biological control
While the complete sequence for Gamma phage is
not yet available, its lysis protein, PlyG, has been
cloned, sequenced and characterized [115]. The lytic
activity appears to be restricted to B. anthracis, how-
ever the specificity of this activity has not been fully
elucidated. It has been suggested that the lysin pro-
teins from the lysogenic phages of B. anthracis could
be utilized as a method for biological control. The
specificity for B. anthracis of such proteins has been
previously demonstrated [115] and used for clinical
typing of B. anthracis (see above). This activity may
also be applied to specifically lyse B. anthracis in other
situations, such as treatment of infections. The possi-
bility of using these proteins as biological control
mechanisms holds promise, but cross reactivity with
certain B. cereus strains could be a problem [115]
and, more importantly, the lytic activity of these pro-
teins is limited to vegetative cells. These proteins do
not affect B. anthracis spores, the infectious particles.
A strategy has been suggested that employs the use
of spore germinant solutions to induce the outgrowth
of vegetative forms in combination with a lysin solu-
tion to destroy emergent vegetative cells. The rapid
concerted conversion from spores to vegetative forms
(less than 10 min) [116] suggests that this approach
could be applied successfully. Both the germinant
and lysin solution should be harmless enough to both
man and machine to use not only in the case of a bio-
terror attack, but also as prophylaxis for the treatment
of troops and machines returning from anthrax con-
taminated areas. Additionally, since the phage lysin
proteins are thought to attack the basic building
blocks of the cell wall, it would be unlikely that resis-
tance to these methods will develop as rapidly as anti-
biotic resistance. Interestingly, no equivalent system
using phages has been proposed or examined for B.
cereus or B. thuringiensis, certainly due to the fact that
a phage would be specific to one strain but not for the
entire group.
322 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329
8. Functional genomics of the B. cereus group of
organisms
Since the publication of the first B. cereus group
organism genomes [40,45], a significant number of stud-
ies have taken advantage of the sequence to generate
functional data. Not surprisingly B. anthracis has been
the focus of much of this work. The mail anthrax at-
tacks of 2001 provided an additional impetus for re-
search as well as an infusion of funding for the
development of microarray and proteomic research
programs.
8.1. Transcriptional analysis using microarrays
8.1.1. B. cereus/B. thuringiensis
The availability of B. cereus specific arrays by a lim-
ited number of companies (NimbleGen and Qiagen) has
not yet generated any published data. Additionally, a
commercially available B. thuringiensis specific array
has not been advertised. While these species-specific ar-
rays are useful, most information would derive from an
array that contains the B. cereus core genomic elements
as well as the unique genes from a number of species or
isolates. The advantage of building such a chimeric ar-
ray would be to capture the level of conservation of
any interrogated isolate in relation to the core genotype,
but it would also detect the level of divergence and pos-
sible gene transfer between isolates/species in compara-
tive genomic hybridization (CGH) experiments. In
addition to obtaining a metric for the level of conserva-
tion, the chimeric array will allow representation of mul-
tiple strains, isolates or species on a single array without
the need for redundant coding sequences to be repre-
sented, thus saving space, time, effort and money.
8.1.2. B. anthracis
A B. anthracis Ames microarray (both 70-mer oligo-
nucleotide and amplicon-based) is available through
the NIAID funded Pathogen Functional Genomics Re-
source Center (http://pfgrc.tigr.org/). Other arrays have
been produced and used for species typing [117] and
resequencing [118]. Read et al. [40] utilized an ampli-
con-based array as a comparative CGH tool to interro-
gate the genomic content of 19 diverse B. cereus group
organisms. This diverse set included a number of the B.
cereus strains from clinical sources, mostly periodontal
infections from Norway [3,28,62] as well as strains
from environmental sources and common laboratory
strains including B. cereus ATCC 14579 and B. cereus
ATCC 10987 [45,47]. This study revealed a high degree
of genomic similarity among these 19 B. cereus isolates.
In addition, it indicates the presence of pXO1 gene
homologues in half of the 19 strains examined, consis-
tent with other studies [63]. In contrast, very few
homologues of pXO2 genes were found to hybridize
[40]. B. cereus isolates from clinical sources contained
more pXO1 and/or pXO2 genes than non-clinical iso-
lates, suggesting a pathogenic role for plasmid-encoded
genes. While this technique does not allow the determi-
nation of gene order, it confirms the presence of
plasmid-like genes in these species. Molecular examina-
tion of some of these clinical isolates has revealed the
presence of large plasmids (>200 kb) that are currently
under investigation (Rasko, unpublished data). Using
the ratios obtained for chromosomal genes, it was pos-
sible to reconstruct the phylogenetic relationship of
these 19 isolates – the clustering obtained was consis-
tent with that obtained using other methods [3,8,27–
29,32,33].
In addition to CGH experiments, this array has also
used for expression analysis in two separate studies. The
expression of plasmid-encoded virulence factors and
genes that are involved with their regulation were ana-
lyzed. Comparison of expression patterns for wild type
and B. anthracis AtxA knockout mutants demonstrated
that AtxA is the dominant regulatory protein and ap-
pears to be a master virulence regulator in B. anthracis
[71]. AtxA was shown to regulate AcpA and AcpB,
which were previously thought to regulate pXO2-
encoded capsule biosynthesis.
B. anthracis gene expression patterns were analyzed
as the bacterium progressed through logarithmic phase
and entered sporulation [72]. In contrast to the previ-
ous study, Liu et al. did not make use of mutational
analysis to examine the regulation but rather took a
less invasive method to examine gene expression as
growth progressed to unravel the regulatory cascade
associated with spore formation. Five distinct waves
of gene expression containing 36% of all predicted
B. anthracis genes were observed. The five waves
roughly corresponded to early, mid or late logarithmic
growth, stationary phase growth and spore formation.
The data revealed that the sporulation program in B.
anthracis is similar to the well-characterized B. subtilis
system [119]. However, B. anthracis sigma factor A
(SigA) expression patterns deviate from those of its
B. subtilis counterpart. sigA is induced during the final
stages of sporulation in B. anthracis whereas in B. sub-
tilis SigA is produced during vegetative growth [119].
Additional examination of the expression data demon-
strated that the plasmid-encoded virulence factors were
expressed only in early logarithmic phase. The true
power of this study is that the expression data were
combined with a proteomic analysis of the sporulating
cultures [72].
8.2. Proteomics studies
8.2.1. B. anthracis
The genome sequence and its predicted proteome
have allowed for high-throughput proteomic analyses
 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329
where proteins are enzymatically digested and the result-
ing peptides analyzed with two-dimensional liquid chro-
matography coupled with tandem mass spectrometry
analysis. Peptide molecular weights are then matched
bioinformatically to specific proteins. Combining this
approach with expression analysis in a single study high-
lights the complementary nature of genomic and proteo-
mic analyses. Liu et al. [72] applied such an approach to
the study of spore formation and were able to rapidly
document high-resolution temporal changes in gene
expression associated with spore formation, while pro-
teomic analysis provided a detailed snapshot of the pro-
tein content and relative abundance in the spore.
Surprisingly, genes that were upregulated in the final
stages of spore formation were rarely identified by the
proteomic analysis as being constituents of the spore.
Of the 873 genes identified to be upregulated in the final
stages preceding sporulation, only 173 were identified
proteomically as being spore components [72]. This
suggests that the sporulating cells obtain proteins from
pre-existing peptide pools to form the spore. If the
microarray and proteomic data had not been combined
into a single study, a number of incorrect conclusions as
to the temporal origin of the proteins in the spore could
have been made.
While the study by Liu et al. [72] demonstrated the
power of using proteomic analysis in a high-throughput
mode, other proteomic studies are focusing on rapid and
accurate speciation of B. anthracis or B. cereus based on
the proteomic content of their spore employing a com-
bined 2D-gel electrophoresis and tandem mass spec-
trometry approach [120]. These studies have been
successful in differentiating spores and vegetative forms
of B. anthracis from its close relatives [120,121]. In addi-
tion to examining and differentiating the Bacillus spe-
cies, these proteomic studies have identified potential
new spore coat proteins not previously annotated as
such or known to reside in the membrane or spore
surface [122].
Anthrax vaccine preparations have also been charac-
terized through proteomic analysis. A recent study using
2D gel-electrophoresis [123] confirmed that the major
constituent of the B. anthracis vaccine was the protective
antigen, but identified a number of minor contaminating
products. These minor constituents included proteolytic
cleavage products of the protective antigen, lethal factor
and edema factor as well as other putative virulence fac-
tors such as EA1, Sap and S-layer proteins. Addition-
ally, cell constituents could be identified such as
60 kDa heat-shock protein, enolase, fructose-bisphos-
phate aldolase and nucleoside diphosphate kinase [123].
Proteomics can also be applied to examine the spore
or bacterial surface for novel vaccine candidates. 2D gel
electrophoresis of the spore proteome probed with sera
from immunized animals identified eight in vivo immu-
nogens, six of which were previously reported as anti-
323
genic targets. Five of the eight candidates were
preselected through bioinformatic analysis of the gen-
ome sequence of B. anthracis [124]. This study highlights
that a combination of genomic data, bioinformatics
analysis and proteomics can contribute to novel anthrax
vaccine development.
8.2.2. B. cereus
Unlike B. anthracis, proteomic analyses were applied
to B. cereus vegetative cells and not spores. B. cereus
PlcR mutant strains were compared to the wild type
using protein 2D gel electrophoresis [54]. As expected,
the inactivation of plcR in most cases abolished or sig-
nificantly reduced the expression of PlcR-regulated pep-
tides. These peptides included a number of virulence
factors known to be under the control of PlcR, such
as collagenase, hemolysins, proteases and toxins. Those
directly regulated by PlcR were most significantly af-
fected, whereas those indirectly regulated were reduced
in expression levels.
Detailed 2D gel electrophoresis proteomic analysis
of B. cereus biofilm establishment and maintenance
on glass wool has been reported [125,126]. In the
dairy industry, B. cereus biofilms inside storage tanks
and associated piping can create problems for product
sterility and longevity. Using a proteomic approach,
B. cereus isolates in biofilms have been shown to ex-
press at least 10 additional proteins and lack seven
proteins during adhesion to a solid surface and estab-
lishment of the biofilm [125]. These peptides may play
important roles in the maintenance of the biofilm and/
or represent metabolic changes triggered when the
bacterium switches from a planktonic to a sessile
lifestyle.
8.2.3. B. thuringiensis
While there are few published B. thuringiensis proteo-
mic studies, significant progress have been made in the
identification of sporulation regulatory pathways using
MALDI-TOF (matrix-assisted laser desorption/ioniza-
tion time of flight) analysis [127]. These pathways are
a driver of B. thuringiensis biology, as the spore is the
bacterial form used as a biopesticide. Other proteomic
studies have focused on the identification of the Cry tox-
in receptor in the Manduca sexta midgut using 1D and
2D gel electrophoresis [128]. These different steps are
essential in the life cycle of B. thuringiensis – proteomics
could have a major impact in developing a better under-
standing of bacterial virulence mechanisms, both by
studying the host and the bacterium itself.
Overall, proteomics and microarray studies have
opened the opportunity to leverage the genomic informa-
tion of the members of this group. These studies have
enabled a better understanding of spore formation
regulatory cascades, identified unknown regulatory ele-
ments and pathways, and have started to examine the
324 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329
interactions of the B. cereus group of organisms with
their respective environments.
9. Summary: One species on the basis of genomic
evidence?
Attempts at bacterial systematics began long before
the discovery of DNA as the hereditable material. In
fact, the seminal discovery demonstrating that DNA
was the genetic material, through the transformation
of avirulent pneumococci to the virulent form [129]
hinted at the importance of horizontal gene transfer as
a mechanism to increase genetic diversity and thereby
niche expansion by a microbial species. Bacteria were
originally classified largely on the basis of phenotype,
morphology, ecology and/or associated disease state.
For example, the bacteria that are the subject of this re-
view were named Bacillus for their rod-shape, cereus for
presumably an association with cereal crops, and
anthracis as the cause of the disease anthrax. The criteria
for the ÔspeciesÕ designation have been widely debated,
however the tenet of reproductive isolation is largely
unassailable. This principle seemed inviolate in asexually
reproducing bacteria until the demonstration of ‘‘fertil-
ity’’ in Escherichia coli in the early 1950s [130,131].
The subsequent discovery that antibiotic resistance
could be disseminated between different species within
the Enterobacteriacae via plasmids [132] caused recon-
sideration of the basis of bacterial classification.
Classical bacterial systematics is now being chal-
lenged by discoveries being promulgated by the
genomics revolution. The defining characteristics of a
species must be grounded in its genetic/genomic archi-
tecture. In the B. cereus group of organisms, virulence
and pathogenicity appear to be promiscuous and
spread with plasmids. As reviewed in this paper, it
is clear that the bacterial chromosomes of the se-
quenced members of this group are extremely similar.
These chromosomes show a high level of synteny and
protein identity, a combination never observed be-
tween different bacterial species. Furthermore, there
is evidence for a shared set of core putative virulence
factors between different pathogenic and non-patho-
genic members of the group. Very few chromosomal
genes or sets of genes are unique to one species. These
regions often represent metabolic adaptations, and in
many cases are found in another species with different
phenotypic and pathogenic traits. Their presence can-
not be associated with a specific subset of organisms.
Conversely, much of the disease and host specificity in
this group can be attributed to plasmid content (i.e.,
pXO1 and pXO2 in B. anthracis). The pXO1-like plas-
mid in B. cereus G9241 has been hypothesized to be
responsible for the anthrax-like clinical presentation
caused by this isolate. However, in other isolates no
role can be associated with any of the harbored plas-
mids. Interestingly, clustering of these plasmids based
on the similarity of their replication systems is possi-
ble, offering an alternative method of systematic
classification.
Certainly our analyses demonstrate a bias toward
significant pathogens of the group. However, not all
members of the group are highly virulent. While 15
genome sequences are publicly available for a diverse
set of B. cereus group organisms, genome data analy-
sis has not delivered in developing a better under-
standing of the mechanisms that contribute to, and
limit the acquisition of virulence. Access to such infor-
mation will help to further refine the definition of a
species.
Technological advances in functional genomics and
proteomics are giving scientists the opportunity to lever-
age genome sequence data. From the analysis presented
here and those of others, it is apparent that gene content
might play a decreased role in the diversity of phenotype
and pathogenicity observed in the B. cereus group. Subtle
changes to regulatory networks may be responsible for
the range of phenotypic traits displayed by the B. cereus
group members. Functional studies and genetic linkage
experiments, combined with proteomic analysis should
provide a better understanding of these regulatory path-
ways and their genetic basis. In addition, using these tech-
nologies combined with access to the genomes of the
human, pathogenic and non-pathogenic members of
the group, it is now possible to study the interaction of
the bacterium with its host. This may allow the emer-
gence of a detailed molecular understanding of the path-
ogenicity of this group of organisms that will benefit the
development of treatment and prevention measures.
The evolution of B. anthracis as a highly successful
pathogen causing a readily identifiable disease has led
to its over-representation in culture collections of iso-
lates from the B. cereus group. Thus, should B. anthracis
be considered an oversampled B. cereus? Based on geno-
mic analyses of representatives of the group, these iso-
lates carry different plasmids in a similar genetic
background. Only subtle differences in gene content
and protein similarities are observed when the chromo-
somes of members of the group are compared. While it
is true that B. anthracis can be readily differentiated
from B. cereus based on biochemical tests [33], a Ôprob-
lemÕ still exists for borderline isolates such as the patho-
genic B. cereus G9241 [20], where these tests fail to
recognize the pathogenic potential of this isolate. The
limited definition of the nature of B. anthracis empha-
sizes the challenges faced by public health officials when
confronted with non-B. anthracis pathogenic isolates,
and their inability to identify them as such [20]. The gen-
ome of B. cereus G9241 would be indistinguishable from
these of any other B. cereus isolates, if it did not contain
a homologue of B. anthracis pXO1. Classification of the
 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329
members of the B. cereus group of organisms has long
been the source of controversy. In 1952, based on the
observation that an isolate of B. anthracis had lost both
virulence, and its plasmids, and was indistinguishable
from B. cereus in term of pathogenicity, Smith et al.
[133] concluded ‘‘B. anthracis is taxonomically a patho-
genic variety of B. cereus’’. The bacterium was later
listed as B. cereus var anthracis [134], a nomenclature
questioned ever since [37,38]. While genome sequence
analysis has not elucidated the pathogenic potential of
some isolates of the group, it has offered unprecedented
insights into the core of their making. Based on the
chromosome genomic comparison reviewed in this pa-
per, it is not possible to distinguish members of the B.
cereus group from one another. Should they be consid-
ered one species based on genomic evidence? It is recog-
nized that for economical and social reasons, changing
the current nomenclature would be quite a challenge.
It is hoped that this paper may represent a starting
point for discussion and further novel studies through
focusing on the similarities and distinctions that contrib-
ute to the nature of this group of bacteria.
Acknowledgments
D.A.R. and J.R. are supported with Federal funds
from the National Institute of Allergy and Infectious
Disease, National Institutes of Health, under Contract
No. N01-AI15447. M.R.A. and C.S.H. are supported
by the US Department of Energy Contract W-7405-
ENG-36 and LAUR#04-8482. We are grateful to Garry
Myers, Thomas Brettin and Paul Jackson for their com-
ments during the preparation of this manuscript and to
Gary Xie for his technical assistance.
References
[1] Chen, M.L. and Tsen, H.Y. (2002) Discrimination of
Bacillus cereus and Bacillus thuringiensis with 16S rRNA
and gyrB gene based PCR primers and sequencing of
their annealing sites. Journal of Appl. Microbio. 92, 912–
919.
[2] Daffonchio, D., Cherif, A. and Borin, S. (2000) Homoduplex
and heteroduplex polymorphisms of the amplified ribosomal
16S–23S internal transcribed spacers describe genetic relation-
ships in the ‘‘Bacillus cereus group’’. Applied and Environmental
Microbiology 66, 5460–5468.
[3] Helgason, E., Økstad, O.A., Caugant, D.A., Johansen, H.A.,
Fouet, A., Mock, M., Hegna, I. and Kolstø, A.B. (2000) Bacillus
anthracis, Bacillus cereus, and Bacillus thuringiensis – One species
on the basis of genetic evidence. Applied and Environmental
Microbiology 66, 2627–2630.
[4] Lecadet, M.M., Frachon, E., Dumanoir, V.C., Ripouteau, H.,
Hamon, S., Laurent, P. and Thiery, I. (1999) Updating the H-
antigen classification of Bacillus thuringiensis. Journal of Applied
Microbiology 86, 660–672.
325
[5] Orduz, S., Restrepo, W., Patino, M.M. and Rojas, W. (1995)
Transfer of toxin genes to alternate bacterial hosts for
mosquito control. Memorias Do Instituto Oswaldo Cruz 90,
97–107.
[6] Schnepf, E., Crickmore, N., Van Rie, J., Lereclus, D., Baum, J.,
Feitelson, J., Zeigler, D.R. and Dean, D.H. (1998) Bacillus
thuringiensis and its pesticidal crystal proteins. Microbiology and
Molecular Biology Reviews 62, 775–790.
[7] Jensen, G.B., Hansen, B.M., Eilenberg, J. and Mahillon, J.
(2003) The hidden lifestyles of Bacillus cereus and relatives.
Environmental Microbiology 5, 631–640.
[8] Helgason, E., Caugant, D.A., Lecadet, M.M., Chen, Y.H.,
Mahillon, J., Lovgren, A., Hegna, I., Kvaloy, K. and Kolstø,
A.B. (1998) Genetic diversity of Bacillus cereus and B. thurin-
giensis isolates from natural sources. Current Microbiology 37,
80–87.
[9] Hernandez, E., Ramisse, F., Cruel, T., le Vagueresse, R. and
Cavallo, J.D. (1999) Bacillus thuringiensis serotype H34 isolated
from human and insecticidal strains serotypes 3a3b and H14 can
lead to death of immunocompetent mice after pulmonary
infection. FEMS Immunology Medical Microbiology 24, 43–47.
[10] Hernandez, E., Ramisse, F., Ducoureau, J.P., Cruel, T. and
Cavallo, J.D. (1998) Bacillus thuringiensis subsp. konkukian
(serotype H34) superinfection: case report and experimental
evidence of pathogenicity in immunosuppressed mice. Journal of
Clinical Microbiology 36, 2138–2139.
[11] Hernandez, E., Ramisse, F., Gros, P. and Cavallo, J. (2000)
Superinfection by Bacillus thuringiensis H34 or 3a3b can lead to
death in mice infected with the influenza A virus. FEMS
Immunology Medical Microbiology 29, 177–181.
[12] Logan, N.A. and Turnbull, P.C. (1999) (Murray, P.R., Ed.),
Manual of Clinical Microbiology, pp. 357–369. American
Society for Microbiology, Washington, DC.
[13] Drobniewski, F.A. (1993) Bacillus cereus and related species.
Clinical Microbiology Reviews 6, 324–338.
[14] Margulis, L., Jorgensen, J.Z., Dolan, S., Kolchinsky, R., Rainey,
F.A. and Lo, S.C. (1998) The arthromitus stage of Bacillus
cereus: intestinal symbionts of animals. Proceedings of the
National Academy of Sciences of the United States of America
95, 1236–1241.
[15] A.B. Kolstø, D. Lereclus, M. Mock, Genome structure and
evolution of the Bacillus cereus group, in: Pathogenicity Islands
and the Evolution of Pathogenic Microbes, vol. 2, 2002, pp. 95–
108.
[16] Jernigan, D.B., Raghunathan, P.L., Bell, B.P., Brechner, R.,
Bresnitz, E.A., Butler, J.C., Cetron, M., Cohen, M., Doyle, T.,
Fischer, M., Greene, C., Griffith, K.S., Guarner, J., Hadler, J.L.,
Hayslett, J.A., Meyer, R., Petersen, L.R., Phillips, M., Pinner,
R., Popovic, T., Quinn, C.P., Reefhuis, J., Reissman, D.,
Rosenstein, N., Schuchat, A., Shieh, W.J., Siegal, L., Swerdlow,
D.L., Tenover, F.C., Traeger, M., Ward, J.W., Weisfuse, I.,
Wiersma, S., Yeskey, K., Zaki, S., Ashford, D.A., Perkins, B.A.,
Ostroff, S., Hughes, J., Fleming, D., Koplan, J.P. and Gerber-
ding, J.L. (2002) Investigation of bioterrorism-related anthrax,
United States, 2001. Epidemiologic Findings 8, 1019–1028.
[17] Jernigan, J.A., Stephens, D.S., Ashford, D.A., Omenaca, C.,
Topiel, M.S., Galbraith, M., Tapper, M., Fisk, T.L., Zaki, S.,
Popovic, T., Meyer, R.F., Quinn, C.P., Harper, S.A., Fridkin,
S.K., Sejvar, J.J., Shepard, C.W., McConnell, M., Guarner, J.,
Shieh, W.J., Malecki, J.M., Gerberding, J.L., Hughes, J.M. and
Perkins, B.A. (2001) Bioterrorism-related inhalational anthrax:
the first 10 cases reported in the United States. Emerging
Infectious Diseases 7, 933–944.
[18] Turnbull, P.C.B. (2002) Introduction: anthrax history, disease
and ecology. Anthrax 271, 1–19.
[19] Mock, M. and Fouet, A. (2001) Anthrax. Annual Review of
Microbiology 55, 647–671.
326 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329
[20] Hoffmaster, A.R., Ravel, J., Rasko, D.A., Chapman, G.D.,
Chute, M.D., Marston, C.K., De, B.K., Sacchi, C.T., Fitzgerald,
C., Mayer, L.W., Maiden, M.C., Priest, F.G., Barker, M., Jiang,
L., Cer, R.Z., Rilstone, J., Peterson, S.N., Weyant, R.S.,
Galloway, D.R., Read, T.D., Popovic, T. and Fraser, C.M.
(2004) Identification of anthrax toxin genes in a Bacillus cereus
associated with an illness resembling inhalation anthrax. Pro-
ceedings of the National Academy of Sciences of the United
States of America 101, 8449–8454.
[21] Okinaka, R., Cloud, K., Hampton, O., Hoffmaster, A., Hill, K.,
Keim, P., Koehler, T., Lamke, G., Kumano, S., Manter, D.,
Martinez, Y., Ricke, D., Svensson, R. and Jackson, P. (1999)
Sequence, assembly and analysis of pXO1 and pXO2. Journal of
Applied Microbiology 87, 261–262.
[22] Okinaka, R.T., Cloud, K., Hampton, O., Hoffmaster, A.R., Hill,
K.K., Keim, P., Koehler, T.M., Lamke, G., Kumano, S.,
Mahillon, J., Manter, D., Martinez, Y., Ricke, D., Svensson, R.
and Jackson, P.J. (1999) Sequence and organization of pXO1,
the large Bacillus anthracis plasmid harboring the anthrax toxin
genes. Journal of Bacteriology 181, 6509–6515.
[23] Priest, F.G. (1981) DNA homology in the genus Bacillus In: The
Aerobic Endospore-forming Bacteria (Berkeley, R.C.W. and
Goodfellow, M., Eds.), pp. 35–57. Academic Press, London.
[24] Ash, C. and Collins, M.D. (1992) Comparative analysis of 23S
ribosomal RNA gene sequences of Bacillus anthracis and emetic
Bacillus cereus determined by PCR direct sequencing. FEMS
Microbiology Letters 94, 75–80.
[25] Ash, C., Farrow, J.A.E., Dorsch, M., Stackebrandt, E. and
Collins, M.D. (1991) Comparative analysis of Bacillus anthracis,
Bacillus cereus, and related species on the basis of reverse
transcriptase sequencing of 16S ribosomal RNA. International
Journal of Systematic Bacteriology 41, 343–346.
[26] Bavykin, S.G., Lysov, Y.P., Zakhariev, V., Kelly, J.J., Jackman,
J., Stahl, D.A. and Cherni, A. (2004) Use of 16S rRNA, 23S
rRNA, and gyrB gene sequence analysis to determine phyloge-
netic relationships of Bacillus cereus group microorganisms.
Journal of Clinical Microbiology 42, 3711–3730.
[27] Carlson, C.R., Caugant, D.A. and Kolstø, A.B. (1994) Geno-
typic diversity among Bacillus cereus and Bacillus thuringiensis
strains. Applied and Environmental Microbiology 60, 1719–
1725.
[28] Helgason, E., Tourasse, N.J., Meisal, R., Caugant, D.A. and
Kolstø, A.B. (2004) Multilocus sequence typing scheme for
bacteria of the Bacillus cereus group. Applied and Environmen-
tal Microbiology 70, 191–201.
[29] Priest, F.G., Barker, M., Baillie, L.W., Holmes, E.C. and
Maiden, M.C. (2004) Population structure and evolution of the
Bacillus cereus group. Journal of Bacteriology 186, 7959–7970.
[30] Ko, K.S., Kim, J.W., Kim, J.M., Kim, W., Chung, S.I., Kim, I.J.
and Kook, Y.H. (2004) Population structure of the Bacillus
cereus group as determined by sequence analysis of six house-
keeping genes and the plcR gene. Infection and Immunity 72,
5253–5261.
[31] Hill, K.K., Ticknor, L.O., Okinaka, R.T., Asay, M., Blair, H.,
Bliss, K.A., Laker, M., Pardington, P.E., Richardson, A.P.,
Tonks, M., Beecher, D.J., Kemp, J.D., Kolstø, A.B., Wong,
A.C.L., Keim, P. and Jackson, P.J. (2004) Fluorescent amplified
fragment length polymorphism analysis of Bacillus anthracis,
Bacillus cereus, and Bacillus thuringiensis isolates. Applied and
Environmental Microbiology 70, 1068–1080.
[32] Keim, P., Kalif, A., Schupp, J., Hill, K., Travis, S.E., Richmond,
K., Adair, D.M., Hugh-Jones, M., Kuske, C.R. and Jackson, P.
(1997) Molecular evolution and diversity in Bacillus anthracis as
detected by amplified fragment length polymorphism markers.
Journal of Bacteriology 179, 818–824.
[33] Ticknor, L.O., Kolstø, A.B., Hill, K.K., Keim, P., Laker, M.T.,
Tonks, M. and Jackson, P.J. (2001) Fluorescent amplified
fragment length polymorphism analysis of Norwegian Bacillus
cereus and Bacillus thuringiensis soil isolates. Applied and
Environmental Microbiology 67, 4863–4873.
[34] Cherif, A., Brusetti, L., Borin, S., Rizzi, A., Boudabous, A.,
Khyami-Horani, H. and Daffonchio, D. (2003) Genetic
relationship in the ÔBacillus cereus groupÕ by rep-PCR
fingerprinting and sequencing of a Bacillus anthracis-specific
rep-PCR fragment. Journal of Applied Microbiology 94,
1108–1119.
[35] Radnedge, L., Agron, P.G., Hill, K.K., Jackson, P.J., Ticknor,
L.O., Keim, P. and Andersen, G.L. (2003) Genome differences
that distinguish Bacillus anthracis from Bacillus cereus and
Bacillus thuringiensis. Applied and Environmental Microbiology
69, 2755–2764.
[36] Keim, P., Price, L.B., Klevytska, A.M., Smith, K.L., Schupp,
J.M., Okinaka, R., Jackson, P.J. and Hugh-Jones, M.E. (2000)
Multiple-locus variable-number tandem repeat analysis reveals
genetic relationships within Bacillus anthracis. Journal of Bac-
teriology 182, 2928–2936.
[37] Turnbull, P.C., Hutson, R.A., Ward, M.J., Jones, M.N., QUinn,
C.P., Finnie, N.J., Duggleby, C.J., Kramer, J.M. and Melling, J.
(1992) Bacillus anthracis but not always anthrax. Journal of
Applied Bacteriology 72, 21–28.
[38] Turnbull, P.C. (1999) Definitive identification of Bacillus anthra-
cis – a review. Journal of Applied Microbiology 87, 237–240.
[39] Fleischmann, R.D., Adams, M.D., White, O., Clayton, R.A.,
Kirkness, E.F., Kerlavage, A.R., Bult, C.J., Tomb, J.F.,
Dougherty, B.A. and Merrick, J.M. (1995) Whole-genome
random sequencing and assembly of Haemophilus influenzae
Rd. Science 269, 496–512.
[40] Read, T.D., Peterson, S.N., Tourasse, N., Baillie, L.W., Paulsen,
I.T., Nelson, K.E., Tettelin, H., Fouts, D.E., Eisen, J.A., Gill,
S.R., Holtzapple, E.K., Økstad, O.A., Helgason, E., Rilstone, J.,
Wu, M., Kolonay, J.F., Beanan, M.J., Dodson, R.J., Brinkac,
L.M., Gwinn, M., DeBoy, R.T., Madpu, R., Daugherty, S.C.,
Durkin, A.S., Haft, D.H., Nelson, W.C., Peterson, J.D., Pop,
M., Khouri, H.M., Radune, D., Benton, J.L., Mahamoud, Y.,
Jiang, L.X., Hance, I.R., Weidman, J.F., Berry, K.J., Plaut,
R.D., Wolf, A.M., Watkins, K.L., Nierman, W.C., Hazen, A.,
Cline, R., Redmond, C., Thwaite, J.E., White, O., Salzberg, S.L.,
Thomason, B., Friedlander, A.M., Koehler, T.M., Hanna, P.C.,
Kolstø, A.B. and Fraser, C.M. (2003) The genome sequence of
Bacillus anthracis Ames and comparison to closely related
bacteria. Nature 423, 81–86.
[41] Read, T.D., Salzberg, S.L., Pop, M., Shumway, M., Umayam,
L., Jiang, L.X., Holtzapple, E., Busch, J.D., Smith, K.L.,
Schupp, J.M., Solomon, D., Keim, P. and Fraser, C.M. (2002)
Comparative genome sequencing for discovery of novel poly-
morphisms in Bacillus anthracis. Science 296, 2028–2033.
[42] Pearson, T., Busch, J.D., Ravel, J., Read, T.D., Rhoton, S.D.,
UÕRen, J.M., Simonson, T.S., Kachur, S.M., Leadem, R.R.,
Cardon, M.L., Van Ert, M.N., Huynh, L.Y., Fraser, C.M. and
Keim, P. (2004) Phylogenetic discovery bias in Bacillus anthracis
using single-nucleotide polymorphisms from whole-genome
sequencing. Proceedings of the National Academy of Sciences
of the United States of America 101, 13536–13541.
[43] Baillie, L. and Read, T.D. (2001) Bacillus anthracis, a bug with
attitude!. Current Opinions in Microbiology 4, 78–81.
[44] Turnbull, P.C. (1991) Anthrax vaccines: past, present and future.
Vaccine 9, 533–539.
[45] Ivanova, N., Sorokin, A., Anderson, I., Galleron, N., Candelon,
B., Kapatral, V., Bhattacharyya, A., Reznik, G., Mikhailova,
N., Lapidus, A., Chu, L., Mazur, M., Goltsman, E., Larsen, N.,
DÕSouza, M., Walunas, T., Grechkin, Y., Pusch, G., Haselkorn,
R., Fonstein, M., Ehrlich, S.D., Overbeek, R. and Kyrpides, N.
(2003) Genome sequence of Bacillus cereus and comparative
analysis with Bacillus anthracis. Nature 423, 87–91.
 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329
[46] Sneath, P.H.A. (1986) Endospore-forming Gram-positive rods
and cocci (Sneath, P.H.A., Mair, N.S., Sharpe, M.E. and Holt,
J.G., Eds.), BergeyÕs Manual of Systematic Bacteriology, vol. 2,
p. 1131. Williams & Wilkins, Blatimore.
[47] Rasko, D.A., Ravel, J., Økstad, O.A., Helgason, E., Cer, R.Z.,
Jiang, L., Shores, K.A., Fouts, D.E., Tourasse, N.J., Angiuoli,
S.V., Kolonay, J., Nelson, W.C., Kolstø, A.B., Fraser, C.M. and
Read, T.D. (2004) The genome sequence of Bacillus cereus
ATCC 10987 reveals metabolic adaptations and a large plasmid
related to Bacillus anthracis pXO1. Nucleic Acids Research 32,
977–988.
[48] Berry, C., OÕNeil, S., Ben-Dov, E., Jones, A.F., Murphy, L.,
Quail, M.A., Holden, M.T.G., Harris, D., Zaritsky, A. and
Parkhill, J. (2002) Complete sequence and organization of
pBtoxis, the toxin-coding plasmid of Bacillus thuringiensis subsp
israelensis. Applied and Environmental Microbiology 68, 5082–
5095.
[49] Cummings, C.A. and Relman, D.A. (2002) Genomics and
microbiology. Microbial forensics – ‘‘cross-examining patho-
gens’’. Science 296, 1976–1979.
[50] CDC, Update: investigation of anthrax associated with
intentional exposure and interim public health guidelines,
October, Morbidity and Mortality Weekly Reports 50 (2001)
889–893.
[51] Keim, P., Van Ert, M.N., Pearson, T., Vogler, A.J., Huynh, L.Y.
and Wagner, D.M. (2004) Anthrax molecular epidemiology and
forensics: using the appropriate marker for different evolutionary
scales. Infection, Genetics and Evolution 4, 205–213.
[52] Suyama, M. and Bork, P. (2001) Evolution of prokaryotic gene
order: genome rearrangements in closely related species. Trends
in Genetics 17, 10–13.
[53] Parkhill, J. and Berry, C. (2003) Genomics: relative pathogenic
values. Nature 423, 23–25.
[54] Gohar, M., Økstad, O.A., Gilois, N., Sanchis, V., Kolstø, A.B.
and Lereclus, D. (2002) Two-dimensional electrophoresis anal-
ysis of the extracellular proteome of Bacillus cereus reveals the
importance of the PlcR regulon. Proteomics 2, 784–791.
[55] Lereclus, D., Agaisse, H., Gominet, M., Salamitou, S. and
Sanchis, V. (1996) Identification of a Bacillus thuringiensis gene
that positively regulates transcription of the phosphatidylinosi-
tol-specific phospholipase C gene at the onset of the stationary
phase. Journal of Bacteriology 178, 2749–2756.
[56] Slamti, L. and Lereclus, D. (2002) A cell–cell signaling peptide
activates the PlcR virulence regulon in bacteria of the Bacillus
cereus group. EMBO Journal 21, 4550–4559.
[57] Slamti, L., Perchat, S., Gominet, M., Vilas-Boas, G., Fouet, A.,
Mock, M., Sanchis, V., Chaufaux, J., Gohar, M. and Lereclus,
D. (2004) Distinct mutations in PlcR explain why some strains of
the Bacillus cereus group are nonhemolytic. Journal of Bacteri-
ology 186, 3531–3538.
[58] Mignot, T., Mock, M., Robichon, D., Landier, A., Lereclus, D.
and Fouet, A. (2001) The incompatibility between the PlcR- and
AtxA-controlled regulons may have selected a nonsense muta-
tion in Bacillus anthracis. Molecular Microbiology 42, 1189–
1198.
[59] Pomerantsev, A.P., Pomerantseva, O.M. and Leppla, S.H.
(2004) A spontaneous translational fusion of Bacillus cereus
PlcR and PapR activates transcription of PlcR-dependent genes
in Bacillus anthracis via binding with a specific palindromic
sequence. Infection and Immunity 72, 5814–5823.
[60] Økstad, O.A., Gominet, M., Purnelle, B., Rose, M., Lereclus, D.
and Kolstø, A.B. (1999) Sequence analysis of three Bacillus
cereus loci carrying PIcR-regulated genes encoding degradative
enzymes and enterotoxin. Microbiology 145, 3129–3138.
[61] Beecher, D.J., Schoeni, J.L. and Wong, A.C.L. (1995) Entero-
toxic activity of hemolysin BL from Bacillus cereus. Infection
and Immunity 63, 4423–4428.
327
[62] Helgason, E., Caugant, D.A., Olsen, I. and Kolstø, A.B. (2000)
Genetic structure of population of Bacillus cereus and Bacillus
thuringiensis isolates associated with periodontitis and other
human infections. Journal of Clinical Microbiology 38, 1615–
1622.
[63] Pannucci, J., Okinaka, R.T., Williams, E., Sabin, R., Ticknor,
L.O. and Kuske, C.R. (2002) DNA sequence conservation
between the Bacillus anthracis pXO2 plasmid and genomic
sequence from closely related bacteria. BMC Genomics 3, 34.
[64] Gladstone, G.P. (1946) Immunity to anthrax – protective antigen
present in cell-free culture filtrates. British Journal of Experi-
mental Pathology 27, 394–418.
[65] Lacy, D.B. and Collier, R.J. (2002) Structure and function of
anthrax toxin. Anthrax 271, 61–85.
[66] Friedlander, A.M. (1990) The anthrax toxins In: Trafficking of
Bacterial Toxins (Saelinger, C.B., Ed.), pp. 121–138. CRC Press,
Boca Raton, FL.
[67] Leppla, S.H. (1995) Anthrax toxins In: Bacterial Toxins and
Virulence Factors in Disease (Moss, J., Iglewski, B., Vaughn, M.
and Tu, A.T., Eds.), pp. 543–572. Dekker, New York, NY.
[68] Mogridge, J., Mourez, M. and Collier, R.J. (2001) Involvement
of domain 3 in oligomerization by the protective antigen moiety
of anthrax toxin. Journal of Bacteriology 183, 2111–2116.
[69] Pannifer, A.D., Wong, T.Y., Schwarzenbacher, R., Renatus, M.,
Petosa, C., Bienkowska, J., Lacy, D.B., Collier, R.J., Park, S.,
Leppla, S.H., Hanna, P. and Liddington, R.C. (2001) Crystal
structure of the anthrax lethal factor. Nature 414, 229–233.
[70] Singh, Y., Khanna, H., Chopra, A.P. and Mehra, V. (2001) A
dominant negative mutant of Bacillus anthracis protective
antigen inhibits anthrax toxin action in vivo. Journal of
Biological Chemistry 276, 22090–22094.
[71] Bourgogne, A., Drysdale, M., Hilsenbeck, S.G., Peterson, S.N.
and Koehler, T.M. (2003) Global effects of virulence gene
regulators in a Bacillus anthracis strain with both virulence
plasmids. Infection and Immunity 71, 2736–2743.
[72] Liu, H., Bergman, N.H., Thomason, B., Shallom, S., Hazen, A.,
Crossno, J., Rasko, D.A., Ravel, J., Read, T.D., Peterson, S.N.,
Yates 3rd, J. and Hanna, P.C. (2004) Formation and compo-
sition of the Bacillus anthracis endospore. Journal of Bacteriol-
ogy 186, 164–178.
[73] Coker, P.R., Smith, K.L., Fellows, P.F., Rybachuck, G.,
Kousoulas, K.G. and Hugh-Jones, M.E. (2003) Bacillus anthra-
cis virulence in Guinea pigs vaccinated with anthrax vaccine
adsorbed is linked to plasmid quantities and clonality. Journal of
Clinical Microbiology 41, 1212–1218.
[74] Andrup, L., Jorgensen, O., Wilcks, A., Smidt, L. and Jensen,
G.B. (1996) Mobilization of ‘‘nonmobilizable’’ plasmids by the
aggregation-mediated conjugation system of Bacillus thuringien-
sis. Plasmid 36, 75–85.
[75] Kaspar, R.L. and Robertson, D.L. (1987) Purification and
physical analysis of Bacillus anthracis plasmids pXO1 and pXO2.
Biochemistry Biophysical Research Communications 149, 362–
368.
[76] Uchida, I., Sekizaki, T., Hashimoto, K. and Terakado, N. (1985)
Association of the encapsulation of Bacillus anthracis with a 60
megadalton plasmid. Journal of General Microbiology 131 (Pt.
2), 363–367.
[77] Makino, S.I., Uchida, I., Terakado, N., Sasakawa, C. and
Yoshikawa, M. (1989) Molecular characterization and protein
analysis of the cap region, which is essential for encapsulation in
Bacillus anthracis. Journal of Bacteriology 171, 722–730.
[78] Uchida, I., Makino, S., Sasakawa, C., Yoshikawa, M., Sugim-
oto, C. and Terakado, N. (1993) Identification of a novel gene,
dep, associated with depolymerization of the capsular polymer in
Bacillus anthracis. Molecular Microbiology 9, 487–496.
[79] Makino, S., Watarai, M., Cheun, H.I., Shirahata, T. and
Uchida, I. (2002) Effect of the lower molecular capsule released
328 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329
from the cell surface of Bacillus anthracis on the pathogenesis of
anthrax. Journal of Infectious Diseases 186, 227–233.
[80] Tinsley, E., Naqvi, A., Bourgogne, A., Koehler, T.M. and Khan,
S.A. (2004) Isolation of a minireplicon of the virulence plasmid
pXO2 of Bacillus anthracis and characterization of the plasmid-
encoded RepS replication protein. Journal of Bacteriology 186,
2717–2723.
[81] Hamon, M.A. and Lazazzera, B.A. (2001) The sporulation
transcription factor Spo0A is required for biofilm development
in Bacillus subtilis. Molecular Microbiology 42, 1199–1209.
[82] Hamoen, L.W., Kausche, D., Marahiel, M.A., van Sinderen, D.,
Venema, G. and Serror, P. (2003) The Bacillus subtilis transition
state regulator AbrB binds to the 35 promoter region of comK.
FEMS Microbiology Letters 218, 299–304.
[83] Hamoen, L.W., Venema, G. and Kuipers, O.P. (2003) Control-
ling competence in Bacillus subtilis: shared use of regulators.
Microbiology 149, 9–17.
[84] Koehler, T.M. (2002) Bacillus anthracis genetics and virulence
gene regulation. Current Topics in Microbiology Immunology
271, 143–164.
[85] Saile, E. and Koehler, T.M. (2002) Control of anthrax toxin gene
expression by the transition state regulator abrB. Journal of
Bacteriology 184, 370–380.
[86] Bastos, M.C. and Murphy, E. (1988) Transposon Tn554 encodes
three products required for transposition. EMBO Journal 7,
2935–2941.
[87] Steichen, C., Chen, P., Kearney, J.F. and Turnbough, C.L.
(2003) Identification of the immunodominant protein and other
proteins of the Bacillus anthracis exosporium. Journal of
Bacteriology 185, 1903–1910.
[88] Sylvestre, P., Couture-Tosi, E. and Mock, M. (2003) Polymor-
phism in the collagen-like region of the Bacillus anthracis BclA
protein leads to variation in exosporium filament length. Journal
of Bacteriology 185, 1555–1563.
[89] Sylvestre, P., Couture-Tosi, E. and Mock, M. (2002) A collagen-
like surface glycoprotein is a structural component of the
Bacillus anthracis exosporium. Molecular Microbiology 45, 169–
178.
[90] Andrup, L., Jensen, G.B., Wilcks, A., Smidt, L., Hoflack, L. and
Mahillon, J. (2003) The patchwork nature of rolling-circle
plasmids: comparison of six plasmids from two distinct Bacillus
thuringiensis serotypes. Plasmid 49, 205–232.
[91] Price, L.B., Hugh-Jones, M., Jackson, P.J. and Keim, P. (1999)
Genetic diversity of protective antigen genes of Bacillus anthra-
cis. Journal of Bacteriology 181, 2358–2362.
[92] Battisti, L., Green, B.D. and Thorne, C.B. (1985) Mating system
for plasmid transfer of plasmids among Bacillus anthracis,
Baciluus, cereus and Bacillus thuringiensis. Journal of Bacteriol-
ogy 162, 543–550.
[93] Mourez, M., Yan, M., Lacy, D.B., Dillon, L., Bentsen, L.,
Marpoe, A., Maurin, C., Hotze, E., Wigelsworth, D., Pimental,
R.A., Ballard, J.D., Collier, R.J. and Tweten, R.K. (2003)
Mapping dominant-negative mutations of anthrax protective
antigen by scanning mutagenesis. Proceedings of the National
Academy of Sciences of the United States of America 100,
13803–13808.
[94] Ackermann, H.W., Roy, R., Martin, M., Murthy, M.R. and
Smirnoff, W.A. (1978) Partial characterization of a cubic
Bacillus phage. Canadian Journal of Microbiology 24, 986–
993.
[95] Stromsten, N.J., Benson, S.D., Burnett, R.M., Bamford, D.H.
and Bamford, J.K. (2003) The Bacillus thuringiensis linear
double-stranded DNA phage Bam35, which is highly similar to
the Bacillus cereus linear plasmid pBClin15, has a prophage
state. Journal of Bacteriology 185, 6985–69859.
[96] Crickmore, N., Bone, E.J., Williams, J.A. and Ellar, D.J. (1995)
Contributions of the indiviual components of the delta-endo-
toxin crystal to mosquitocidal activity of Bacillus thuringiensis
subsp. israeliensis. FEMS Microbiology Letters 131, 249–2554.
[97] Guidi-Rontani, C., Pereira, Y., Ruffie, S., Sirard, J.C., Weber-
Levy, M. and Mock, M. (1999) Identification and characteriza-
tion of a germination operon on the virulence plasmid pXO1 of
Bacillus anthracis. Molecular Microbiology 33, 407–414.
[98] Ason, B., Handayani, R., Williams, C.R., Bertram, J.G.,
Hingorani, M.M., OÕDonnell, M., Goodman, M.F. and Bloom,
L.B. (2003) Mechanism of loading the Escherichia coli DNA
polymerase III beta sliding clamp on DNA. Bonafide primer/
templates preferentially trigger the gamma complex to hydrolyze
ATP and load the clamp. Journal of Biological Chemistry 278,
10033–10040.
[99] Anand, S.P., Mitra, P., Naqvi, A. and Khan, S.A. (2004) Bacillus
anthracis and Bacillus cereus PcrA helicases can support DNA
unwinding and in vitro rolling-circle replication of plasmid
pT181 of Staphylococcus aureus. Journal of Bacteriology 186,
2195–2199.
[100] Naqvi, A., Tinsley, E. and Khan, S.A. (2003) Purification and
characterization of the PcrA helicase of Bacillus anthracis.
Journal of Bacteriology 185, 6633–6639.
[101] Wilcks, A., Jayaswal, N., Lereclus, D. and Andrup, L. (1998)
Characterization of plasmid pAW63, a second self-transmissible
plasmid in Bacillus thuringiensis subsp. kurstaki HD73. Micro-
biology 144, 1263–1270.
[102] Wilcks, A., Smidt, L., Økstad, O.A., Kolstø, A.B., Mahillon,
J. and Andrup, L. (1999) Replication mechanism and
sequence analysis of the replicon of pAW63, a conjugative
plasmid from Bacillus thuringiensis. Journal of Bacteriology
181, 3193–3200.
[103] Andrup, L., Smidt, L., Andersen, K. and Boe, L. (1998) Kinetics
of conjugative transfer: a study of the plasmid pXO16 from
Bacillus thuringiensis subsp. israelensis. Plasmid 40, 30–43.
[104] Jensen, G.B., Andrup, L., Wilcks, A., Smidt, L. and Poulsen,
O.M. (1996) The aggregation-mediated conjugation system of
Bacillus thuringiensis subsp. israelensis: host range and kinetics of
transfer. Current Microbiology 33, 228–236.
[105] Anisimov, A.P., Lindler, L.E. and Pier, G.B. (2004) Intraspecific
diversity of Yersinia pestis. Clinical Microbiology Reviews 17,
434–464.
[106] Brown, E.R. and Cherry, W.B. (1955) Specific identification of
Bacillus anthracis by means of a variant bacteriophage. Journal
of Infectious Diseases 96, 34–39.
[107] Stone, R. (2002) Bacteriophage therapy: StalinÕs forgotten cure.
Science 298, 728–731.
[108] Stone, R. (2002) Bacteriophage therapy: food and agriculture:
testing grounds for phage therapy. Science 298, 730.
[109] Canchaya, C., Proux, C., Fournous, G., Bruttin, A. and
Brussow, H. (2003) Prophage genomics. Microbiology and
Molecular Biology Reviews 67, 238–276.
[110] Casjens, S. (2003) Prophages and bacterial genomics: what have
we learned so far?. Molecular Microbiology 49, 277–300.
[111] Desiere, F., McShan, W.M., van Sinderen, D., Ferretti, J.J. and
Brussow, H. (2001) Comparative genomics reveals close genetic
relationships between phages from dairy bacteria and pathogenic
streptococci: evolutionary implications for prophage–host inter-
actions. Virology 288, 325–341.
[112] Brussow, H., Canchaya, C. and Hardt, W.D. (2004) Phages and
the evolution of bacterial pathogens: from genomic rearrange-
ments to lysogenic conversion. Microbiology and Molecular
Biology Reviews 68, 560.
[113] Perna, N.T., Plunkett 3rd, G., Burland, V., Mau, B., Glasner,
J.D., Rose, D.J., Mayhew, G.F., Evans, P.S., Gregor, J.,
Kirkpatrick, H.A., Posfai, G., Hackett, J., Klink, S., Boutin,
A., Shao, Y., Miller, L., Grotbeck, E.J., Davis, N.W., Lim, A.,
Dimalanta, E.T., Potamousis, K.D., Apodaca, J., Ananthar-
aman, T.S., Lin, J., Yen, G., Schwartz, D.C., Welch, R.A. and
 D.A. Rasko et al. / FEMS Microbiology Reviews 29 (2005) 303–329
Blattner, T. () Genome sequence of enterohaemorrhagic Esch-
erichia coli O157:H7. Nature 409, 529–533.
[114] Campbell, A.M. (1992) Chromosomal insertion sites for phages
and plasmids. Journal of Bacteriology 174, 7495–7499.
[115] Schuch, R., Nelson, D. and Fischetti, V.A. (2002) A bacteriolytic
agent that detects and kills Bacillus anthracis. Nature 418, 884–
889.
[116] Ireland, J.A. and Hanna, P.C. (2002) Macrophage-enhanced
germination of Bacillus anthracis endospores requires gerS.
Infection and Immunity 70, 5870–5872.
[117] Volokhov, D., Pomerantsev, A., Kivovich, V., Rasooly, A. and
Chizhikov, V. (2004) Identification of Bacillus anthracis by
multiprobe microarray hybridization. Diagnostic Microbiology
and Infectious Disease 49, 163–171.
[118] Zwick, M.E., Macaffee, F., Cutler, D.J., Read, T.D., Ravel, J.,
Bowman, G.R., Galloway, D.R. and Mateczum, A. (2004)
Microarray-based resequencing of multiple Bacillus anthracis
isolates. Genome Biology 6, R10.
[119] Kroos, L. and Yu, Y.T.N. (2000) Regulation of sigma factor
activity during Bacillus subtilis development. Current Opinion in
Microbiology 3, 553–560.
[120] Warscheid, B. and Fenselau, C. (2004) A targeted proteomics
approach to the rapid identification of bacterial cell mixtures by
matrix-assisted laser desorption/ionization mass spectrometry.
Proteomics 4, 2877–2892.
[121] Warscheid, B. and Fenselau, C. (2003) Characterization of
Bacillus spore species and their mixtures using postsource decay
with a curved-field reflectron. Analytical Chemistry 75, 5618–
5627.
[122] Lai, E.M., Phadke, N.D., Kachman, M.T., Giorno, R., Vazquez,
S., Vazquez, J.A., Maddock, J.R. and Driks, A. (2003) Proteo-
mic analysis of the spore coats of Bacillus subtilis and Bacillus
anthracis. Journal of Bacteriology 185, 1443–1454.
[123] Whiting, G.C., Rijpkema, S., Adams, T. and Corbel, M.J. (2004)
Characterisation of adsorbed anthrax vaccine by two-dimen-
sional gel electrophoresis. Vaccine 22, 4245–4251.
[124] Ariel, N., Zvi, A., Makarova, K.S., Chitlaru, T., Elhanany, E.,
Velan, B., Cohen, S., Friedlander, A.M. and Shafferman, A.
(2003) Genome-based bioinformatic selection of chromosomal
Bacillus anthracis putative vaccine candidates coupled with
proteomic identification of surface-associated antigens. Infection
and Immunity 71, 4563–4579.
329
[125] Oosthuizen, M.C., Steyn, B., Lindsay, D., Brozel, V.S. and von
Holy, A. (2001) Novel method for the proteomic investigation of
a dairy-associated Bacillus cereus biofilm. FEMS Microbiology
Letters 194, 47–51.
[126] Oosthuizen, M.C., Steyn, B., Theron, J., Cosette, P., Lindsay,
D., von Holy, A. and Brozel, V.S. (2002) Proteomic analysis
reveals differential protein expression by Bacillus cereus during
biofilm formation. Applied and Environmental Microbiology 68,
2770–2780.
[127] Chen, F.C., Shen, L.F., Tsai, M.C. and Chak, K.F. (2003) The
IspA proteaseÕs involvement in the regulation of the sporulation
process of Bacillus thuringiensis is revealed by proteomic
analysis. Biochemical and Biophysical Research Communica-
tions 312, 708–715.
[128] McNall, R.J. and Adang, M.J. (2003) Identification of novel
Bacillus thuringiensis Cry1Ac binding proteins in Manduca sexta
midgut through proteomic analysis. Insect Biochemistry and
Molecular Biology 33, 999–1010.
[129] Avery, O.T., MacLeod, C.M. and McCarty, M. (1944)
Studies on the chemical nature of the substance inducing
transformation of pneumococcal types. Inductions of trans-
formation by a desoxyribonucleic acid fraction isolated from
pneumococcus type III. Journal of Experimental Medicine 79,
137–158.
[130] Hayes, W. (1953) Observations on a transmissible agent deter-
mining sexual differentiation in Bacti. coli. Journal of General
Microbiology 8, 72–88.
[131] Lederberg, J., Cavalli, L.L. and Lederberg, E.M. (1952) Sex
compatibility in E. coli. Genetics 37, 720–730.
[132] Watanabe, T. (1963) Infectious heredity of multiple drug
resistance in bacteria. Bacteriological Reviews 27, 87–115.
[133] N.R. Smith, R.E. Gordon, F.E. Clark, Aerobic sporeforming
bacteria, US Department of Agriculture, Agriculture Mono-
graph No. 16, US Government Printing Office, Washington, DC,
1952.
[134] R.E. Gordon, W.C. Haynes, C.H.-N. Pang, The genus Bacillus,
Aagriculture Handbook No. 427, United States Department of
Argiculture, Agricultural Research Service, Washington, DC,
1973.
[135] Rasko, D.A., Myers, G.S. and Ravel, J. (2005) Visualization of
comparative genomic analyses by BLAST score ratio BMC.
Bioinformatics 6, 2.