RESEARCH ARTICLE
Combined proteomic and transcriptomic analysis
of the response of Bacillus anthracis to oxidative stress
Susanne Pohl1, Wang Y. Tu1, Phillip D. Aldridge1, Colin Gillespie2, Hannes Hahne1,
Ulrike M .ader1, Timothy D. Read 3 and Colin R. Harwood1
1 Centre for Bacterial Cell Biology, Institute for Cell and Molecular Biosciences, Newcastle University, Newcastle
upon Tyne, UK
2 School of Mathematics and Statistics, Newcastle University, Newcastle upon Tyne, UK
3 Department of Human Genetics, Emory University School of Medicine, Atlanta, USA
Received: February 14, 2011
Revised: March 28, 2011
Accepted: April 5, 2011
The endospore-forming Gram-positive pathogen Bacillus anthracis is responsible for the
usually fatal disease, inhalational anthrax. The success of this pathogen is dependent on its
ability to subvert elements of the innate immune system of its animal hosts. B. anthracis
spores, which are the main infective agent, are engulfed and germinate in patrolling alveolar
macrophages. In order for the infection to progress, the resulting vegetative cells must resist
the antimicrobial oxidative burst mounted by the host NADPH oxidase complex. The
response of B. anthracis to this and other macrophage-related stresses is therefore of major
importance to the success of this pathogen, and consequently we have analysed the super-
oxide and peroxide stress stimulons of B. anthracis strain UM23C1-2 by means of a combined
transcriptomics and proteomics approach. The results show distinct patterns of expression in
response to paraquat (endogenous superoxide) and hydrogen peroxide stress. While the main
response to paraquat is the induction of iron uptake pathways, the response to peroxide
predominantly involves the induction of protection and repair mechanisms. Comparisons
between the responses of B. anthracis and related soil bacterium, B. subtilis, reveal differences
that are likely to be relevant to their respective habitats.
Keywords:
Bacillus anthracis / Iron / Microbiology / Oxidative stress / Transcriptome
1 Introduction
The Gram-positive bacterium Bacillus anthracis is the aetiolo-
gical agent of anthrax, a disease associated with humans and
other mammals. Although there is evidence that B. anthracis
can undergo limited proliferation in the environment [1], it is
generally regarded as an obligate pathogen with a distinct life
cycle. The infective agent is the endospore, a highly resistant
dormant cell that is able to survive in the environment for
decades. The three forms of anthrax (inhalation, ingestion and
cutaneous) are determined by the site at which the spores enter
the body. The most severe form of the disease is pulmonary
anthrax that, if not treated early enough, is usually fatal.
Following infection, spores are engulfed into phago-
somes by professional phagocytes, a key component of the
innate immune system. Phagocytes such as macrophages
deploy a complex range of interacting antimicrobial activ-
ities as they attempt to kill the invading bacterium [2]. These
include cell wall hydrolases, acidification, metal availability
Abbreviations: DLD, dihydrolipoamide dehydrogenase; DPS,
DNA protection during starvation; Fur, ferric uptake repressor;
ROS, reactive oxygen species
Current addresses: Ulrike M.ader, Departments of Functional Geno-
mics and Microbiology, Ernst-Moritz-Arndt-University of Greifswald,
Friedrich-Ludwig-Jahn-StraXe 15 Greifswald, Germany
Hannes Hahne, Chair of Proteomics and Bioanalytics, Center of Life
and Food Sciences Weihenstephan, Technische Universit.at M .unchen,
Emil-Erlenmeyer-Forum 5, 85354 Freising, Germany
Colour Online: See the article online to view Fig. 2 in colour.
Correspondence: Professor Colin R. Harwood, Centre for
Bacterial Cell Biology, Institute for Cell and Molecular Bios-
ciences, Newcastle University, Baddiley-Clark Building,
Richardson Road, Newcastle upon Tyne, NE2 4AX, UK
E-mail: colin.harwood@ncl.ac.uk
Fax: 144-191-2083205
& 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.proteomics-journal.com
3036 Proteomics 2011, 11, 3036–3055DOI 10.1002/pmic.201100085

and an antimicrobial oxidative burst mounted by a
membrane-associated host NADPH oxidase complex [3]. In
non-activated macrophages, many of the components of the
NADPH oxidase complex are located in the cytoplasm.
However, the presence of spores activates the macrophage
and the cytosolic components of the complex migrate to the
membrane where they associate with the membrane
components to form the NADPH oxidase active complex
that initiates the reduction of molecular oxygen to super-
oxide ions. Although superoxide ions are potentially anti-
microbial, there is evidence that, in B. anthracis, they provide
the signal that initiates the germination process for spores
that are, by now, located in the phagolysosome [4, 5].
Although reactive oxygen species (ROS) are generated
endogenously as a natural by-product of aerobic respiration,
there is evidence that pathogens have developed specific
strategies for neutralizing exogenously generated ROS.
Superoxide ions can be dismutated to hydrogen peroxide,
either spontaneously or enzymatically by superoxide
dismutases. If the resulting hydrogen peroxide is not rapidly
converted into water and molecular oxygen by catalase, it
can interact with Fe(II) via Fenton chemistry to generate
highly reactive and toxic hydroxyl radicals [6].
The mechanisms by which B. anthracis is able to survive
and potentially proliferate in macrophages are still poorly
understood. As a preliminary approach we have focused on
the mechanisms by which B. anthracis is able to resist the
macrophage-mediated oxidative burst. To this end, we have
undertaken a combined proteomic and transcriptomic
analyses of the response of this bacterium to superoxide and
peroxide stress. Comparison with data obtained from the
previous studies on the soil bacterium B. subtilis provides
insights into the specific adaptations used by this bacterium
to overcome the higher levels of oxidative stress encountered
during infection [7].
In this study, 2-D protein gel electrophoresis and high-
density DNA microarray technology were used to obtain a
comprehensive analysis on the response of B. anthracis to
superoxide and peroxide stress. Despite the fact that the well-
established pathway for the detoxification of superoxide ions is
their conversion into hydrogen peroxide and its removal via
catalase, the cellular responses to paraquat and hydrogen
peroxide were distinct with surprisingly little overlap. The
results provide evidence of a specific relationship between
superoxide stress and iron homeostasis that is both novel and
counter-intuitive. This is all the more intriguing given the
important role that host-mediated ‘‘iron withholding’’ plays in
restricting the proliferation of pathogens [8].
2 Materials and methods
2.1 Bacterial strains and growth conditions
B. anthracis strain UM23C1-2 (pXO1, pXO2) [9, 10] and
B. subtilis strain 168 were cultivated aerobically at 371C in LB
medium. For the induction of oxidative stress in B. anthra-
cis, paraquat or hydrogen peroxide were added to exponen-
tially growing cells at an OD540nm of 0.3 to final
concentrations of 0.8 and 1.0mM, respectively. Samples for
the preparation of total cellular RNA were taken 10min after
treatment and samples for the preparation of cytosolic
proteins after 60min.
2.2 Proteome and transcriptome analyses
Treated and untreated cultures of B. anthracis UM23C1-2 for
both transcriptome and proteome analyses were harvested by
centrifugation. For the extraction of total RNA, the acidic
phenol method was used [11]. The resulting RNA pellet was
resuspended in sterile water and the RNA concentration was
determined spectrophotometrically. The quality of the RNA
was determined by Northern Blot analysis and by a micro-
fluidics-based analyzer (Bioanalyzer 2100, Agilent Technolo-
gies, Berlin, Germany). RNA first-strand cDNA synthesis was
generated by random priming and the resulting DNA label-
led for strand-specific hybridization with Cy3. Three inde-
pendently isolated RNA samples were used for the cDNA
synthesis and hybridization reactions. Data were analysed
using the GeneSpring software (Agilent Technologies). Raw
signal intensities were first transformed by quantile normal-
ization and then the data for individual CDS combined.
Changes in mRNA abundance were considered to be signif-
icant if the averaged fold change between the treated and
untreated samples was at least 1.9. All differentially expressed
genes are listed in Table 1. The putative and known functions
of the encoded proteins were predominantly inferred from
the SubtiWiki (http://subtiliswiki.net) and SubtiList (http://
genolist.pasteur.fr/SubtiList/) databases.
For the isolation of cytosolic proteins, the cells were
washed with 10mM Tris–HCl, 1mM EDTA, pH 7.5, 1mM
PMSF and physically disrupted using a Micro Dismem-
brator S (Braun Biotech International, Germany) for 5min
at 2600 rpm. Crude protein extracts were separated from
cellular debris by repeated centrifugation (15 000 g,
30min, 41C). The resulting supernatant contained the
soluble cytoplasmic protein fraction. The concentrations of
proteins in the samples were determined using the 2D
Quant Kit (Amersham Pharmacia Biotech), and 200mg of
protein extract was analysed by 2-D gel electrophoresis, as
described previously [12]. IPG strips (Amersham Pharmacia
Biotech) in the pH range 4–7 were used to separate proteins
in the first direction, and a 12.5% SDS–polyacrylamide gel
in the second. The gels were stained with PlusOne
Coomassie Tablets (Amersham Pharmacia Biotech). The
data were analysed by comparison of the 2D protein patterns
of untreated and stressed cultures using the Delta2D soft-
ware (Version 4.; Decodon, Greifswald, Germany). Protein
identification was by MALDI-MS (Voyager, Applied Biosys-
tems, UK) using the MASCOT analytical software (Matrix
Science, Boston, USA).
Proteomics 2011, 11, 3036–3055 3037
& 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.proteomics-journal.com