Project no. LSHG-CT-2004-503468
Project acronym: BACELL HEALTH
Project title: Bacterial stress management relevant to infectious disease and
Instrument: Specific Targeted Research Project
Thematic Priority: LifeSciHealth; LSH-2002-1.1.0-1
Title of report
Final Activity Report
Period covered: from 01 March 2004 to 29 February 2008
Date of preparation: 4 April, 2008
Start date of project: 01 March 2004 Duration: 48 months
Project coordinator: Prof. Colin R. Harwood
Coordinator’s organisation: Newcastle University
Project manager: Dr Sierd Bron
1. Objectives ..............................................................................................................3
2. Consortium and management ................................................................................4
4. Work programme ...................................................................................................6
5. Methodologies/techniques .....................................................................................7
6. Main results and achievements ..............................................................................8
6.1. Regulated intramembrane proteolysis; activation sigma W ..........................8
6.3. Role of DegU in adaptive responses in Listeria monocytogenes.................10
6.4. Role of Listeria monocytogenes Agl system in virulence ...........................11
6.5. Alkaline stress response in B. subtilis..........................................................12
6.6. Acid stress response in B. subtilis................................................................12
6.7. Oxidative stress response in B. subtilis........................................................13
6.8. Siderophore-dependent iron acquisition in B. subtilis .................................19
6.9. HtrA-like proteases in Staphylococcus aureus ............................................21
6.10. Specificity in cross-talk of YycFG and PhoPR two-component systems....22
6.11. Bacillus anthracis stress responses..............................................................22
6.12. CssRS-dependent secretion stress in B. subtilis...........................................23
6.13. Secretion/expression host/vector systems for B. subtilis .............................26
6.14. Sublancin production in B. subtilis ..............................................................27
6.15. Improvements of Bacillus production strains ..............................................28
6.16. Web interface for genome-scale prediction of Bacillus secretomes ............31
6.17. Software development .................................................................................32
7. Achievements versus State-of-the-art ..................................................................35
7.1. State-of-the-art .............................................................................................35
7.2. Progress in relation to the plans ...................................................................35
7.3. Workpackages – Plan and Status Bar chart ................................................36
7.4. Major achievements .....................................................................................38
7.4.1. Stress systems ......................................................................................38
7.4.2. Bioinformatics application/modelling tools.........................................39
7.4.3. Microbial virulence..............................................................................40
7.4.4. Bioprocessing; improvements in protein productions .........................40
7.5. Impact on industry and research sector........................................................41
While bacteria remain one of the main threats to human health and well-being,
particularly in the light of increasing resistance to antimicrobials, they also have the
capacity to provide products (e.g. antibiotics, vaccines and biotherapeutics) that have
a positive influence on human health. Thus, bacteria are both targets for, and
producers of, biopharmaceuticals.
BACELL HEALTH (for logo: see Figure 1) is a research and technological
development project that aimed to address both the negative and positive
characteristics of bacterial behaviour by undertaking an integrated and in-depth study
of the response of Gram-positive bacteria to stresses encountered by pathogens during
infection and by
strains during bio-
both of these
responses, they also
specific stress responses
(eg. pH, iron and
oxidative stress in the case of macrophages,
Fig. 1. BACELL HEALTH logo
protein synthesis, secretion and nutrient
stress in the case of industrial fermentations). The ability of bacteria to detect, respond
to and resist environmental insult is a key element in their survival and productivity.
BACELL HEALTH aimed to develop a profound understanding of the
integrative cell management and associated stress resistance processes that are
essential for sustaining bacteria as effective pathogens, or as efficient producers of
pharmaceutically-active proteins. The major challenge was to understand how
individual regulatory pathways are networked to maintain cellular homeostasis. This
network was referred to as the “Cell Stress Management System”. The knowledge
obtained was used to model the regulatory networks that comprise the Cell Stress
Management System, identifying key targets for the development of novel anti-
infectives and bottlenecks that limit the productivity of Bacillus for the production of
Bacillus subtilis is regarded as the paradigm for the study of Gram-positive
bacteria. It is one of only two bacterial systems (the other is Escherichia coli) in
which it is feasible to address experimentally the objectives identified above. The
feasibility of using B. subtilis for this purpose was based on knowledge and expertise
acquired over the last 45 years, more recently as specific deliverables of previous,
EU-funded RTD projects. One of these projects, BACELL NETWORK, has
demonstrated regulatory cross-talk between general and specific stress responses.
These Programmes have created a European Bacillus scientific community
(BACELL) that is second to none in terms of its interactivity, demonstrable expertise,
innovation and productivity.
Generic studies being carried out on B. subtilis were designed to provide a
baseline response to environmental stresses such as resistance to alkaline or acid pH,
iron limitation and oxidative, cell wall and secretion stress. The knowledge from
these studies was applied to an understanding of stress responses in pathogens in the
same genus (e.g. B. anthracis, B. cereus), and related genera (e.g. Listeria
monocytogenes, and Staphylococcus aureus). The idea was that understanding what
additional resistance mechanisms are required by pathogens to resist the innate
immune response would provide potential new targets for anti-infective (as compared
to antibiotic) drugs. Similarly, an understanding of the stresses encountered by
commercial strains during fermentation was expected to identify bottlenecks to
productivity, if the corresponding response was incompatible with product formation.
A good example of this is the stress-mediated production of proteases that severely
limit the production of pharmaceutically-active proteins by B. subtilis.
Specifically, the consortium aimed to:
(i) model the regulatory networks that comprise the cell stress management
(ii) identity key elements within the cell stress management system and
stress resistance processes as targets for the development of novel anti-
(iii) to improve Bacillus subtilis as a host for the production of
The scientific objectives were therefore aimed at improving European
competitiveness and helping to meet the health needs of society. The resulting
knowledge will be of relevance to the cell management systems of all living systems.
BACELL HEALTH was proposed to build on the technological knowledge
and industrial base established in Europe by focusing on aspects that directly
influence human health, namely the development of anti-infective agents and the
improved production of pharmaceutically-active proteins. The added-value nature of
this project was confirmed not only by the full participation of three European
companies, but also by the support of the Bacillus Industrial Platform (BACIP).
Expected deliverables included knowledge of fundamental biological systems, the
identification of novel targets for the development of broad-spectrum and/or Gram-
selective anti-infectives, an improved understanding of microbial virulence and the
regulatory response of bacteria to host-mediated stress responses, prototypes of
production strains and new protein structures and functions. In addition, the
consortium was expected to add to the stock of highly trained young European
scientists, and has disseminated knowledge via workshops and practical training
2. CONSORTIUM AND MANAGEMENT
The members of the BACELL HEALTH consortium were selected on the
basis of their relevant experience in the research fields addressed in the work
programme. Most of the participating groups have a wide experience in research on B.
subtilis and related bacilli and have recognised international reputations. Several of
these groups also have research expertise in the field of pathogenic Gram-positive
bacteria. Within the consortium, a considerable number of facilities and technologies
are available: molecular biology techniques (all groups); DNA-array facilities
(partners 1,2,4,5,6,7,8,10B); high-throughput proteome facilities (partners 1,6); bio-
informatics for comparative genomics (partners 1,2,4,5,6,10A, 10B); macrophage test
systems for pathogens (partner 1; animal lab facilities (P3 level) for testing
pathogenicity of L. monocytogenes and S. aureus (partner 5).
Table 1: List of partners
Partner Participant name Organisation Country
1 Prof. Colin Harwood Newcastle University UK
2 Prof. Kevin Devine Trinity College Dublin IE
3 Prof. Mohamed Marahiel Marburg University DE
4 Prof. Wolfgang Schumann Bayreuth University DE
5 Dr Tarek Msadek Institute Pasteur FR
6 Prof. Michael Hecker Greifswald University DE
7 Dr Marc Kolkman Genencor International NL
8 Dr. M.D. Rasmussen Novozymes A/S DK
9 Dr Rocky Cranenburgh Cobra Biomanufacturing UK
10A Prof Jan Maarten van Dijl
Groningen University NL
10B Prof. Oscar Kuipers
The consortium was coordinated by Professor Colin Harwood of Newcastle
University (UK) and administered by Dr Sierd Bron of Groningen University (NL).
PROF. COLIN HARWOOD
Institute for Cell and Molecular Biosciences, Faculty of Medical
Sciences, Medical School, Framlington Place, Newcastle upon Tyne,
NE2 4HH, UK
A Scientific Committee, consisting of the partners’ groupleaders, was established for
the half-annual monitoring of the scientific progress in the project.
In addition, a Commercial Committee, consisting of the representatives of
the industrial partners, the coordinator and the project manager, was established for
the evaluation of results/knowledge having the potential for exploitation.
A password-protected website for partners in the project was established by the
A general public website for Bacillus research and application in Europe was set up:
A poster for BACELL and BACIP (Bacillus Industrial Platform) was produced.
BACELL HEALTH forms part of BACELL. For information: see
4. WORK PROGRAMME
The Work Programme was divided into five Workpackages: WP1. Generic stress
responses; WP2. Infection stress; WP3. Bioprocess stress; WP4. Comparative
genomics and network modelling; WP5. Management and dissemination (Figure 2;
the four WPs dealing with the research programme are shown in the left block of this
Figure 2. Work programme and Deliverables
The Workplan combined detailed generic studies on the stress responses in B. subtilis
(WP1) with studies on intracellular and extracellular human pathogens (WP2). Their
role in resisting host-mediated antibacterial defence mechanisms makes these stress
responses promising targets for novel anti-infective agents. As had been revealed in
previous EU-funded projects, the response to stresses that develop during
biomanufacturing processes (WP3) is an important factor limiting the use of B.
subtilis for the production of pharmaceutically-active proteins, and the inclusion of
three biomanufacturing companies was designed to ensure rapid exploitation of the
knowledge gained in this workpackage. The links between the various stress
resistance mechanisms in these organisms was revealed by extensive use of
comparative genomic and modelling techniques (Workpackage 4) to integrate the
knowledge from Workpackages 1-3 into a model that describes how the regulatory
elements of individual responses are networked to form the Cell Stress Management
The following methodologies and techniques were used to achieve the
objectives for the project. All the required methods/techniques were collectively
available from within the consortium of the partners.
• General molecular biology techniques; including recombinant DNA
• Genomic and post-genomics technologies
• Mass spectrometry
• DNA arrays/transcriptome analysis
• Protein fluorescence technology
• Macrophage assays
• Immunoblot technology
• Bioinformatics; including network modelling and comparative
• Protein modification assays
In addition, in a cooperation with Prof Iwata/Dr Carpenter (Imperial College,
London), partner 10A has been able to include X-ray diffraction technology for
protein structure analysis.
6. MAIN RESULTS AND ACHIEVEMENTS
All partners have reached their milestones and deliverables, although some priorities
have been modified. With minor exceptions, the timescale for deliverables was in
conformity with the description in Annex I (the Technical Annex; version 06-06-
2006). The main results and achievements were as described in the following
sections. In several cases, the results and progress in the project have exceeded the
6.1. Regulated intramembrane proteolysis; activation sigma W
There is increasing evidence that ECF-sigma factors play an important role in
pathogenesis and the defence against antimicrobial compounds (antibiosis) in
bacteria. The mechanism of activation of the ECF-sigma factor σW in B. subtilis has
been characterised to a considerable level by partner 4.
This partner could show that σW is sequestered and inactivated by a
transmembrane anti-sigma factor (RsiW), and released and activated by a process of
regulated intramembrane proteolyis. The importance of conserved alanine residues
that constitute a cryptic proteolytic tag in the transmembrane region of the RsiW anti-
sigma factor could be shown. ClpXP and, to a minor extent, ClpEP are the major
cytoplasmic proteases that completely remove the truncated RsiW and liberate σW to
interact with core RNA polymerase. This last step of RsiW proteolysis integrates the
σW-response to other stresses. Figure 3 shows an overview of a current model of the
activation of σW.
In order to identify the site-1 protease and/or additional factors triggering the
stress-induced regulated intramembrane proteolysis of RsiW, partner 4 initiated a
transposon screen with a GFP-RsiW reporter. Among several transposon mutants, two
different gene loci with a strong phenotype showed up several times. First, mutations
in the ecsA and ecsB genes encoding an ABC-transporter block site-2 proteolysis of
RsiW by the intramembrane cleaving protease (iCLIP) RasP (YluC). Second,
transposon insertions in the ypdC (renamed as prsW) gene, encoding a transmembrane
protein of unknown function, prevent site-1 proteolysis. In an ypdC null mutant,
induction of σW–controlled genes is abolished and site-1 proteolysis of RsiW is
completely blocked. Transcriptional analysis revealed that ypdC is a monocistronic
gene, and the defect of σW–induction of the null mutant was complemented by
ectopically integrated ypdC under xylose control. Orthologs of YpdC can be found in
a variety of different bacteria. Its membrane topology was analysed by alkaline
phosphatase fusions, revealing that YpdC contains five transmembrane segments and
two larger extra-cytoplasmic loops. In the first loop, two invariantly conserved
glutamate residues can be found. In an E. coli system, the cloned ypdC is the only
determinant for efficient degradation of RsiW, however, YpdC does not display plain
similarities to known proteases, suggesting that it either controls the activity of site-1
proteolysis of RsiW or represents a new type of protease.
Partner 4 started to investigate the effects of EcsAB on the function of the RasP
protease. First, the defect of σW induction of an ecsA knockout strain could be partly
suppressed by overproducing RasP. Second, it could be shown that a B. subtilis RasP
knockout strain displays the same pleiotropic phenotype as an ecsA knockout, i.e. a
defect in the secretion of α-amylase, biofilm formation, and competence
Fig. 3a. Diagrammatic representation of the regulatory pathway
controlling the activity of σW of B. subtilis
Fig. 3b. Factors controlling the activity of σW of B. subtilis
6.2. Function of YycFG two-component system
The yycFG two-component signal transduction system (TCS) is of particular
interest, since it is present exclusively in the low G+C group of Gram-positive
bacteria, where it is the only two-component system that is essential for survival.
Partner 2 has shown that, in B. subtilis, YycFG regulates the production of two
putative enzymes, YocH and YkvT, both involved in cell wall metabolism. In
addition, partner 5 has shown that the YycGF system also controls the expression of a
set of genes involved in cell wall metabolism in S. aureus. The involvement of this
system in cell wall metabolism is likely to explain why it represents an essential two-
Partner 5 observed that, unexpectedly, S. aureus cell death induced by WalKR
(Figure 4; WalKR is the S. aureus homologue of B. subtilis YycGF) depletion was not
followed by lysis. This partner showed that WalKR positively controls autolytic
activity, in particular that of the two major S. aureus autolysins, AtlA and LytM. By
using the previously characterized consensus WalR binding site and carefully
reexamining the genome annotations, nine genes were identified by this partner which
potentially belong to the WalKR regulon, encoding enzymes involved in every known
step of S. aureus cell wall degradation (glucosaminidase, transglycosylase-
/muramidase, glycyl-glycine endo-peptidase, amidase). Expression of all of these
genes was positively controlled by WalKR levels in the cell, leading to high resistance
to Triton X-100-induced lysis when cells were starved for WalKR. Cells lacking
WalKR were also more resistant to lysostaphin-induced lysis, suggesting
modifications in cell wall structure. Indeed, lowered levels of WalKR led to a
significant decrease of peptidoglycan biosynthesis and turn-over, and to cell wall
modifications which include increased peptidoglycan crosslinking and glycan chain
Partner 5 also demonstrated a direct relationship between WalKR levels and the
ability to form biofilms. This is the first example in S. aureus of a regulatory system
positively controlling autolysin synthesis and biofilm formation.
Fig. 4. Loss of viability in starved S. aureus in the absence of YycG/F.
Green, SYTO 9 – stained live cells; Red, Propidium iodide-stained dead cells
6.3. Role of DegU in adaptive responses in Listeria monocytogenes
In contrast to B. subtilis, DegU was shown to be essential for flagellar synthesis
and bacterial motility in L. monocytogenes (Figure 5; partner 5). Indeed, DegU is
required for expression of several motility and chemotaxis genes, including the flaA
and motAB genes which are expressed at 25°C but not at 37°C. This partner also
showed that DegU is required for growth at high temperature, adherence to plastic
surfaces and the formation of efficient biofilms by L. monocytogenes. This
requirement appeared to be independent from flagellar synthesis.
Fig. 5. Electron transmission micrograph of Listeria monocytogenes cells. The presence
of flagella is shown in the parental wild-type EGDe strain (left panel) and loss of flagella in
the ∆degU mutant (right panel).
Partner 5 has inactivated the phosphorylation site of DegU in vivo and showed
that the protein retains much of its activity, indicating that the unphosphorylated form
is active. The heterologous degS gene from B. subtilis was introduced into L.
monocytogenes and it was shown that phosphorylation of DegU by DegS increases
expression of chemotaxis and motility genes. Using HPLC-MS, it was also shown that
DegU of L. monocytogenes can be phosphorylated by small chemical phosphodonors,
such as acetyl phosphate using HPLC-MS. The inactivation of the ackA and pta
genes of L. monocytogenes effectively prevented formation of acetyl phosphate in
vivo and led to phenotypes very similar to those observed for the ∆degU mutant. The
results suggest a role for acetyl phosphate in vivo as a phosphodonor for DegU. Taken
together, the results indicate that, despite the lack of the DegS kinase, DegU is
functional as an orphan response regulator, and plays a central role in controlling
several crucial adaptive responses in L. monocytogenes.
6.4. Role of Listeria monocytogenes Agl system in virulence
During the study of partner 5 of the Agl system of L. monocytogenes, an in-
frame deletion of the aglD gene was constructed, removing the entire coding sequence
of the peptide precursor gene, and leading to loss of expression of the agl operon.
Extracellular complementation restored expression of the operon when the ΔaglD
mutant was resuspended in conditioned medium, as followed by Quantitative Real-
Time PCR. This indicates that the system likely acts in cell-cell communication.
Among the genes controlled by this system, three were identified which encode
potentially secreted proteins that appear to be specific to virulent strains of Listeria
and it was shown that they play an important role in virulence in a murine model
infection assay. This partner has identified a potential binding site for the AglA
regulator present upstream of several genes within the regulon, and shown that
mutations affecting this highly conserved sequence abolish expression of the
6.5. Alkaline stress response in B. subtilis
A main objective of partner 4 was to analyse the generic stress response of B.
subtilis to sudden changes in the external pH, with the aim to unravel regulatory
mechanisms that enable bacteria to resist to extreme pH values which they encounter
in host defence mechanisms (pathogenic bacteria) and large scale fermentations
(industrial relevant bacteria). The application of pH stress is an important host-
mediated response to bacterial infection, and acidification of the macrophage
phagosome is an important killing mechanism.
Two different strategies were followed. First, the analysis of data obtained from
global transcriptional profiling experiments using DNA array technology, and second,
the investigation of the mechanism of known pH-stress responses on the molecular
level using genetic tools like reporter gene fusions and transposon mutagenesis
Partner 4 showed that, when B. subtilis is challenged with alkali stress, about 80
genes are induced. DNA macroarray analyses of alkaline stressed B. subtilis revealed
two different groups of genes induced by a pH upshock, i.e. genes controlled by the
alternative ECF sigma factor σW, and genes independent to it. To identify regulators
and regulatory mechanisms of alkaline inducible genes not being part of the σW
regulon, the genes yybP, ykoY and yhaS were analysed in detail. The transcriptional
start sites were identified. For the alkali induction of the yybP and ykoY genes, a 5’
untranslated mRNA-region is important, suggesting a riboswitch mechanism. To
identify the exact regulatory sequence, deletions in respective 5’ untranslated regions
were cloned. For the yybP gene, deletion of a 135 bp region in the 5’ untranslated
region resulted in a high constitutive transcription of yybP and loss of alkaline
In the course of this project, the alkaline inducible σW regulon proved to
constitute an antibiosis regulon, defending cell envelope stresses triggered by, among
others, antimicrobial peptides.
6.6. Acid stress response in B. subtilis
Acid stress is accompanied by the induction of >100 genes, mostly members of
the sigma B (σB) regulon (partner 4). The genes yuaE and ywaC were identified to be
acid induced, independently on the general σB-mediated stress response. The
transcriptional start sites were mapped by primer extension analysis. For ywaC, a σM
consensus promoter was found. Preliminary experiments revealed that genes
controlled by the ECF sigma factor σM are induced by acid stress, but that the
mechanism is not dependent on intramembrane proteolysis through the RasP protease.
Attempts were made by partner 4 to identify proteins sensing acid stress,
which is needed to activate σB. Individual knockouts for the six RsbR paralogs
(YkoB, Yoj, YqhA, YteA, YetI, and YezB) were constructed and combined with a
σB-controlled lacZ reporter fusion. In addition, knockouts in rsbX, rsbS and rsbR were
collected. Experimental parameters for acid induction of the σB-controlled gsiB-lacZ
reporter were determined. The interpretation of the effects of individual genes was
complicated, however, by the fact that paralogous genes with similar functions
obscure the effects of individual genes, and so far it has not been possible to ascertain
an acid stress responsive function to one of these genes.
6.7. Oxidative stress response in B. subtilis
Oxidative stress is an important stress in relation to host-pathogen-interactions
and the aim of partner 6 was to define the oxidative stress stimulon in B. subtilis. This
special response was almost completely characterized on the transcriptome and
proteome level using different natural and artificial stressors (peroxide, superoxide-
generating paraquat and diamide). Because this partner found that the expression of
the very important NADH dehydrogenase SdhA from the respiratory chain was not
significantly altered after superoxide stress, it was decided not to investigate the
NADH dehydrogenase activities and the cellular NADH level in more detail (as
planned in the original grant proposal). Instead, partner 6 focussed the research on the
development of an in vivo assay monitoring the different thiol modification states
during oxidative stress. A fluorescence-based thiol modification assay was established
and this was combined with two-dimensional gel electrophoresis and mass
spectrometry to monitor the in vivo thiol state of cytoplasmic proteins. Protein thiols
of growing cells appeared to be mainly present in the reduced state. Only a few
proteins were found to be thiol-modified, these include enzymes that have oxidized
thiols in their catalytic cycle. To detect proteins that are particularly sensitive to
oxidative stress, partner 6 exposed growing B. subtilis cells to diamide, hydrogen
peroxide, or to the superoxide generating agent paraquat. Diamide mediated a
significant increase of oxidized thiols in a variety of metabolic enzymes, whereas
treatment with paraquat affected only a few proteins. Exposure to hydrogen peroxide
forced the oxidation of, especially, proteins with active site cysteines, for instance of
cysteine-based peroxidases and glutamine amido-transferase-like proteins. Moreover,
high levels of hydrogen peroxide were observed to influence the isoelectric point of
proteins of this group, indicating the generation of irreversibly oxidised thiols. From
the overlapping set of oxidatively modified proteins, also enzymes necessary for
methionine biosynthesis were identified, such as: cobalamin-independent methionine
synthase, MetE. Growth experiments revealed a methionine limitation after diamide
and hydrogen peroxide stress, which suggests a thiol-oxidation-dependent inactivation
of MetE. Finally, evidence was presented that the antibiotic nitrofurantoin mediates
the formation of oxidized thiols in B. subtilis.
Additionally, catechol and the ganomycin-related substance 2-
methylhydroquinone (2-MHQ) were also found to cause the induction of the thiol-
specific oxidative stress response in B. subtilis (partner 6). Multiple
dioxygenases/glyoxalases, azo- and nitroreductases were induced by thiol-reactive
compounds, which are regulated by the novel thiol-specific regulators, YodB and
Partner 6 has suggested that these multiple dioxygenases/glyoxalases, azo- and
nitroreductases, regulated by MhqR, have evolved to detoxify quinone-like and other
thiol-reactives electrophiles. Azoreductases might be involved in the thiol-
independent quinone- and azocompound reduction and dioxygenases/glyoxalases
catalyze, most probably, the thiol-dependent ring-cleavage of benzoquinone-S-
Disulphide bonds, formed between pairs of cysteine residues, are critical for
the stability and/or activity of various proteins of both prokaryotes and eukaryotes.
This aspect of the project was analysed by partner 10A. Although this process can
occur spontaneously in vitro, it is often inefficient and unspecific. Therefore, a set of
proteins, known as thiol-disulphide oxidoreductases (TDORs), have evolved to
facilitate disulphide bond formation in vivo. In general, enzymes of this type catalyse
the formation of native disulphide bonds in proteins exported from the cytoplasm, and
prevent the formation of disulphide bonds in cytoplasmic proteins by keeping these
cysteine residues in a reduced state. Many thiol-disulphide oxidoreductases are known
to cooperate in complex redox networks by the formation of transient disulphide
bonds with other thiol-disulphide oxidoreductases or their target proteins.
Collectively, the redox networks of this type are defined as the disulphide proteome.
Amongst the low-GC Gram-positive bacteria, only the TDORs of B. subtilis and S.
aureus have so far been studied in some detail. Two of the cytoplasmic TDORs, TrxA
and TrxB, were shown to be essential for growth and viability of these organisms,
which underlines the importance of their physiological function. Concerning the
extracytoplasmic TDORs, it was found that B. subtilis has four TDORs (named
BdbA-D), whereas S. aureus has only one (named DsbA).
An important goal of partner 10A was to map the disulphide proteome of B.
subtilis in order to understand how it is established, and to determine its importance
for oxidative stress resistance, and protein secretion (Figure 6). Where appropriate, a
comparison was made with the disulphide proteome of S. aureus. For this purpose,
experiments to map the disulphide proteome of B. subtilis by a proteomics technique,
which is called “mixed disulphide fishing”, have been performed. This technique
involves the use of site-specific mutants of thiol-disulphide oxidoreductases that lack
one active site cysteine residue. Such mutant enzymes will form stable reaction
intermediates, known as mixed disulphides, with their substrates. By binding mutant
thiol-disulphide oxidoreductases to a resin, interacting partner proteins and substrates
can be “fished” and, subsequently, identified (Figure 7; collaboration with partner 6).
The experiments were focused on thioredoxin A (TrxA) of B. subtilis. Single cysteine
mutants and a cysteine-less mutant of TrxA were used to establish the mixed
disulphide fishing technology. The corresponding genes have been cloned into a
pET26b vector for overproduction in E. coli and hexa-histidine-tagged TrxA mutant
proteins have been purified by affinity chromatography, concentrated and further
purified by gelfiltration. Many different conditions for the mixed disulphide fishing
experiments with the wild-type and mutant TrxA proteins of B. subtilis have been
tested. Currently, the optimal conditions have been established and interactions have
been analyzed. The protocol is specific enough to only fish out and visualize the
TrxA-interacting proteins. However, the subsequent identification of the fished
proteins by MS has turned out difficult due to impurities that interfere with the gel-
free mass spectrometric analyses (partner 6). The protocols are currently modified to
remove the impurities. Additionally, partner 10A has cloned the trxA and dsbA genes
of S. aureus for site-directed mutagenesis, purification of the corresponding proteins
and mixed disulphide fishing in the future.
Fig. 6. The disulphide proteome of B. subtilis
In addition to the mixed disulphide fishing experiments, the purified TrxA
proteins (wild-type and cysteine mutants) have been used by partner 10A to determine
properties, like redox state, monomer/dimer states, and structure. An unprecedented
finding was that both active site cysteine residues were found to be able to form
mixed disulphides with substrate proteins, but only the N-terminal active site cysteine
formed stable interactions. A second novelty was that both single-cysteine mutant
TrxA proteins formed stable homo-dimers due to thiol oxidation of the remaining
active site cysteine residue. To investigate whether these dimers resemble mixed
enzyme-substrate disulphides, the structure of the most abundant dimer, C32S, was
characterised at a high-resolution (1.5 Å) by X-ray crystallography (collaboration with
the group of Prof. So Iwata, Imperial College, London, UK). This TrxA dimer can be
regarded as a mixed disulphide reaction intermediate of thioredoxin. It reveals the
complexity of hydrophobic interactions that TrxA makes with its substrates.
Interestingly, comparison of this structure with the NMR structure of the recently
published B. subtilis TrxA complexed with a substrate, ArsC, showed a distinct mode
of substrate binding in which the peptide chains are bound in different orientations.
Fig 7. Mixed disulfide fishing. Mixed disulfide fishing was performed with highly pure His6-
tagged BsTrxA with the wild-type active site or with active site-specific mutations, and
cytoplasmic proteins from the TrxA-depleted B. subtilis WB800 ItrxA strain. A. To show that the
BsTrxA proteins used for mixed disulfide fishing were highly pure, 0.1 µg of each BsTrxA
protein variant was loaded on SDS-PAGE. Upon electrophoresis, the gel was silver-stained. WT,
BsTrxA with wild-type active site; C29S, C29S single mutant BsTrxA; C32S, C32S single
mutant BsTrxA; C29S C33S, C29S C32S double mutant BsTrxA. B. To visualize possible stable
interactions between BsTrxA and its substrates, 2 µg of each BsTrxA protein variant was mixed
with 50 µl cytoplasmic protein extract. After 5 min incubation, proteins were separated on a non-
reducing gel and BsTrxA-substrate complexes were visualized by Western blotting with
antibodies raised against B. subtilis TrxA. Background, cytoplasmic extract mock-treated with 2
µl water instead of BsTrxA protein reveals a few aspecific cross-reactions of the BsTrxA
antibody. C. Immunodetection of purified BsTrxA-substrate complexes. For purification of the
BsTrxA-substrate bound complexes, the C-terminal His6-tag of the pure BsTrxA proteins was
used. Pure BsTrxA with the wild-type or a mutant active site was mixed with cytoplasmic protein
extracts as indicated for panel B. Subsequently, magnetic beads precharged with nickel were
added and the His6-tag of the BsTrxA proteins was allowed to bind to the nickel of the magnetic
beads for 10 min. A magnet was then used to collect the beads with bound BsTrxA. After
washing the beads nine times, the BsTrxA proteins were eluted from the beads with buffer
containing imidazole. The eluted proteins were separated by non-reducing SDS-PAGE and the
BsTrxA-substrate bound complexes were visualized by Western blotting with antibodies against
Partner 10A studied the phylogenetic relationships and functional differences
between the BdbABCD of B. subtilis and the DsbA system of S. aureus. Phylogenetic
analyses revealed that B. subtilis BdbD and S. aureus DsbA cluster in topologically
distinct groups, which are typical for Bacillus and Staphylococcus species,
respectively (Figure 8). To compare the function of these TDORs, DsbA was
Fig. 8. Schematic representation of the evolutionary conservation of BdbD
homologues in low-GC Gram-positive bacteria. Names of species used in this figure are
abbreviations of the full names: Ban, B. anthracis, Bce, Bacillus cereus, Bli, B.
licheniformis, Bsu, B. subtilis, Bth, B. thuringiensis, Bwe, B. weihenstephanensis, Efa, E.
faecalis, Lmo, L. monocytogenes, Lin, L. inocua, Lsa, L. sakei, Sau, S. aureus, Sep, S.
epidermidis, Sha, S. haemolyticus, Ssa, S. saprophyticus. The calculated maximum
parsimony values are shown at the nodes. The tree is unrooted, though four BdbD
homologues of high-GC Gram-positive bacteria were included to represent an out-group.
These out-group species are Frankia sp. CcI3; Rubrobacter xylanophilus (Rxy) DSM
9941; Streptomyces coelicolor (Sco) A3(2); and Corynebacterium efficiens (Cef) YS-314.
The clusters of BdbD homologues from the Bacillus, Staphylococcus and Listeria species
and the out-group are encircled. BdbD of B. subtilis 168 and DsbA of S. aureus
NCTC8325 are boxed. The tree branch for highly related BdbD homologues of B.
anthracis, B. cereus, and B. thuringiensis species is enlarged (marked a, b, c).
expressed in bdb mutants of B. subtilis. Next, this partner assessed the ability of DsbA
to sustain TDOR-dependent processes, such as heterologous secretion of E. coli
PhoA, competence development and production of the bacteriocin sublancin 168.
DsbA by itself did function in all three processes, even in the absence of the B.
subtilis Bdb proteins that are critical for the respective processes. While BdbD needs
a cognate quinone oxidoreductase for activity, Staphylococcus DsbA was found to be
active without such a partner protein in the heterologous B. subtilis background.
Interestingly, these studies revealed that DsbA is dependent on active redox
compounds in the growth medium. Apparently, DsbA does not rely on other TDORs
for its reoxidation but, instead, it is reoxidized by redox compounds such as cysteine
or cystine. Unexpectedly, both quinone oxidoreductases of B. subtilis (i.e. BdbB and
BdbC) appeared to be sufficient to sustain production of sublancin. Moreover, DsbA
can functionally replace these quinone oxidoreductases in sublancin production.
Taken together, these findings imply that TDOR systems of low-GC Gram-positive
bacteria have evolved analogously and have a modular composition.
S-Thiolation is crucial for protection and regulation of thiol-containing
proteins during oxidative stress and is frequently achieved by the formation of mixed
disulfides with glutathione (partner 6). However, many Gram-positive bacteria,
including B. subtilis, lack the low molecular weight (LMW) thiol glutathione.
Evidence was provided that S-thiolation by the LMW thiol cysteine represents a
general mechanism in B. subtilis. A large subset of proteins previously identified as
having redox-sensitive thiols are modified by cysteine in response to treatment with
the thiol-specific oxidant diamide. By means of multidimensional shotgun proteomics,
the sites of S-cysteinylation for six proteins could be identified, three of which are
known to be S-glutathionylated in other organisms.
Partner 6 showed that the CtsR regulon in B. subtilis (Figure 9) is not only
induced by heat stress, but also by oxidative stress, especially after diamide treatment.
The activity of the CtsR repressor is modulated by McsB and McsA; these proteins
are encoded together in the tetracistronic clpC operon (ctsR-mcsA-mcsB-clpC). The
interaction of McsA and McsB with CtsR results in formation of a ternary complex
that prevents CtsR from binding to its operator sites on the DNA and leads also to a
subsequent phosphorylation of McsB, McsA and CtsR. McsB is a putative tyrosine
kinase that needs the double Zn-finger containing McsA for its full activity.
Especially the second Zn-finger in McsA appears crucial for the interaction of McsB
with McsA, as could be demonstrated by co-immunoprecipitation and biacore
analysis. Two specific and highly conserved tyrosine residues in McsB, as well as in
McsA, were phosporylated, as was shown by site-directed mutagenesis and in vitro
phosphorylation. ClpC obviously inhibits the McsB kinase activity. YwlE, a small
phospho-tyrosine phosphatase, specifically de-phosphorylates McsB-P, McsA-P and
CtsR-P in vitro, pointing to another new level in the fine-tuned regulation of the CtsR
In order to detect putative different McsA forms after oxidative stress in vivo,
single and double Zn-finger McsA mutations were constructed and integrated as His-
and/or STREP-tagged N-terminal fusions into the chromosomal mcsA site (partner 6).
The mutated forms can be expressed in-frame comparable to wild-type levels from the
natural promoter of the clpC operon. The different McsA versions can now be
purified by affinity chromatography and tested for specific in vivo post-translational
modifications. The fluorescence thiol modification assay established by this partner
for 2D gels proved not to be applicable for in vivo purified proteins, due to high
background signals in 1D gels and the non-traceability of McsA in standard 2D gels.
However, other experimental approaches (sophisticated MS analysis, AMS assay)
seem promising alternatives to detect post-translational modifications in vivo after
oxidative stress. Despite these experimental drawbacks, it could be demonstrated, that
only one putative Zn-finger of McsA appeared to be crucial for the oxidative stress
activation. However, the exact molecular activation mechanism remains elusive and is
still under investigation.
Fig 9. The CtsR regulon of B. subtilis
The non-pathogen B. subtilis and the pathogen S. aureus are Gram-positive
model organisms that have to cope with the radical nitric oxide (NO) generated by
nitrite reductases of denitrifying bacteria and inducible NO synthases of immune cells
of the host, respectively. The response of both microorganisms to NO was analyzed
by partner 6, using the 2D gel approach. Metabolic labeling of the proteins revealed
major changes in the synthesis pattern of cytosolic proteins after addition of the NO
donor MAHMA NONOate. Whereas B. subtilis induced several oxidative stress
specific regulons controlled by Fur, PerR, OhrR, and Spx, as well as the general stress
response controlled by the alternative sigma factor SigB, the more resistant S. aureus
showed an increased synthesis rate of proteins involved in anaerobic metabolism. By
monitoring the intracellular protein thiol state no increase in reversible or irreversible
protein thiol modifications after NO stress was observed in both organisms.
Obviously, NO itself does not cause general protein thiol oxidations. In contrast,
exposition of cells to NO prior to peroxide stress diminished the irreversible thiol
oxidation caused by hydrogen peroxide.
6.8. Siderophore-dependent iron acquisition in B. subtilis
Low iron bioavailability is a constraint that affects the broad majority of pro-
and eukaryotic microorganisms. The restricted access to this essential nutrient has led
to the development of siderophore-dependent iron acquisition as the most common
strategy to scavenge iron with high affinity from extracellular sources. Consequently,
many pathogens employ siderophores as virulence factors to multiply in the colonized
In this work, partner 3 has chosen the Gram-positive soil bacterium B. subtilis
as model organism to elucidate and characterize its siderophore-dependent iron
acquisition pathway (Figure 10). B. subtilis produces the catechole siderophore
bacillibactin (BB), whose nonribosomal biosynthesis is encoded within the
dhbACEBF gene cluster. The comparison of biosynthesis mutant phenotypes revealed
a hierarchy of BB and alternative iron chelators in cellular iron acquisition. The dhbC
mutant was used to demonstrate the capability of BB mutasynthesis and showed that
BB export is not dependent on the ferric uptake repressor (Fur), which is in contrast
to all further identified pathway steps. A phenotypic screening of 61 B. subtilis export
mutants revealed the major facilitator superfamily (MFS)-related transmembrane
permease YmfE as a determinant for BB secretion. The Fe-BB uptake system
FeuABC, consisting of the substrate binding protein FeuA and the membrane
permeases FeuB and FeuC, was identified by mutagenesis of Fur-regulated ABC-type
importers. Biochemical analysis of recombinant FeuA demonstrated both high
specificity and affinity for Fe-BB binding. Intracellular iron release from Fe-BB was
shown to depend on the BesA (YuiI) esterase that is encoded upstream of the dhb
operon. Recombinant BesA was 25-fold more catalytically efficient for hydrolysis of
Fe-charged than uncharged BB.
DhbCBA Gly YmfE, ...?
chorismate (MFS-type porter)
Fig. 10. Current bacillibactin pathway model in B. subtilis
The elucidation of the BB pathway by partner 3 led to the identification of
several pathway components that represent suitable drug targets. Among such targets
are the aryl acid adenylation domains DhbE and YbtE (from Yersinia pestis), which
are involved in siderophore biosynthesis (Figure 11). Their inhibition was studied
with a set of non-hydrolyzable bisubstrate analogs, and two sulfamoyl adenosine
derivatives, which showed inhibition constants in the nanomolar range. Furthermore,
inhibition of BB uptake and hydrolysis was studied with the stable BB analog
MECAM that showed in vitro activity against both FeuA and BesA and was found to
inhibit iron-depleted growth of B. subtilis as well as clinical isolates of E. coli and
General physiological effects that emerge during growth under iron
deprivation were studied by a DNA microarray-based transcriptome analysis of iron-
starved B. subtilis cells (partners 3 and 10B). The analysis revealed stringent
response-like effects and the induction of amino acid biosynthesis pathways that yield
glycine and threonine for BB production.
Fig. 11. DHB-AMS inhibitor at the DhbE active site (model). From: Miethke et al.,
2006, FEBS Journal 273, 409-419
6.9. HtrA-like proteases in Staphylococcus aureus
An in silico analysis by partner 2 revealed that there are two HtrA proteases in
S. aureus, SAV1728 (424 amino acids) and SAV 1023 (769 amino acids). While both
have homology to the HtrA-like proteases of B. subtilis over the length of the protein,
SAV1023 has an amino-terminal extension of approximately 300 amino acids that is
of unknown function. To establish whether both proteins can function as HtrA
proteases, partner 2 sought to determine whether the proteins could complement htrA
protease null mutants of B. subtilis. The gene encoding the SAV1728 full-length
protein and the part of the SAV1023 that is homologous to HtrA were expressed in a
B. subtilis strain mutated for both htrA and htrB: functionality was monitored by
measuring expression of an endogenous htrA-bgaB promoter fusion and by seeing if
the temperature sensitivity of the B. subtilis htrA htrB strain was rescued. The results
show that one of the HtrA proteins can rescue the temperature sensitivity of the B.
subtilis strain, indicating that it is functional in this host. The second S. aureus HtrA
protein (the truncated protein without the amino terminal) did not rescue the
temperature sensitivity; however expression of the htrA-bgaB fusion was reduced,
suggesting that the protein is partially functional, although not sufficiently so to
abolish temperature sensitivity.
6.10. Specificity in cross-talk of YycFG and PhoPR two-component systems
There are similarities and connections between the YycFG and PhoPR two-
component systems in B. subtilis: it was shown that cross-talk exists between the
YycFG and PhoPR two-component systems (partner 2). The kinase PhoR can
phosphorylate both its cognate PhoP and also YycF. A system was constituted that
enables studies on what amino acids contribute to the specificity of interaction
between the YycG kinase and the YycF response regulator.
6.11. Bacillus anthracis stress responses
Proteomic and transcriptomic technologies have been applied by partner 1 to the
analysis of oxidative stress in B. anthracis. These analyses have revealed very distinct
responses for the oxidative stress response to superoxide (Figure 12) and peroxide
stress (Figure 13). More that 150 genes were up-regulated in response to peroxide
stress, while mass spectrometry identified the main changes in the protein
composition of the cytoplasm and culture medium. The peroxide stress response was
focused on the induction of peroxidase and catalase enzymes, together with proteins
and enzymes that reduce damage to membranes and DNA. The response to
superoxide stress was, remarkably, very different and focused on iron homeostasis.
This indicates that distinct regulators are involved in the peroxide and superoxide
stress responses and that, unlike B. subtilis, the responses are unexpectedly distinct.
The superoxide stress response, surprisingly, induced genes involved in iron
starvation stress and may involve the Fur regulator.
Fig. 12. Comparative analysis of B. anthracis to superoxide stress
Fig. 13. Comparative analysis of B. anthracis to peroxide stress
Partner 1 has extended the analyses of the B. anthracis stress response to
include diamide (diazenedicarboxylic acid bis (N,N-dimethylamide)) that induces
disulphide stress in B. anthracis. Northern blot analyses were carried out to determine
the influence of diamide on the expression of ahpC and ahpF encoding subunits of
alkyl-hydroxy-peroxidase, sodA and sodC encoding superoxide bromoperoxidase,
catalase BA3134, bromoperoxidases BA3164 and BA3165, bsaA encoding
glutathione peroxidase and trx encoding thioredoxin. With the exception of BA3134
and BA3165, all of these genes were induced in response to diamide. In some cases
the intensity of the response was enhanced in the mutant lacking the main catalase,
KatB. To determine its influence on the production of proteins by B. anthracis, the
intracellular proteins of cells exposed to diamide were analysed by 2D-PAGE.
Comparison with an untreated control revealed 8 proteins that were up-regulated in
response to disulphide stress. Two proteins were variants of an ATP-dependent Clp
protease, while four of the proteins (BA1951, BA5302, BA2239 and BA1908) were
proteins of similar size (22.8-23.1 KDa.) and involved in the generation reducing
equivalents. One protein was a ribosomal subunit interface protein, while the
remaining protein was not identified by mass spectrometry.
6.12. CssRS-dependent secretion stress in B. subtilis
B. subtilis secretes proteins at high levels into its environment. Most of these
secretory proteins are exported from the cytoplasm in an unfolded state and have to
fold efficiently after membrane translocation. As previously shown for α-amylases of
Bacillus species, inefficient post-translocational protein folding is potentially
detrimental and represents a stressful event. In B. subtilis, this so-called secretion
stress is sensed and combated by the CssRS two-component system. Two known
members of the CssRS regulon are the htrA and htrB genes, encoding potential
extracytoplasmic chaperone-proteases for protein quality control (Figure 14). Partner
10A has investigated whether high-level production of a secretory protein with two
disulphide bonds, PhoA of E. coli, provokes secretion stress in B. subtilis. The results
show that E. coli PhoA production triggers a relatively moderate CssRS-dependent
secretion stress response in B. subtilis. The intensity of this response is significantly
increased in absence of the TDOR BdbC, which is a major determinant for post-
translocational folding of disulphide bond-containing proteins in B. subtilis. This
implies that unfolded PhoA is a direct or indirect stimulus for the CssRS system.
These findings show that BdbC is required to limit the PhoA-induced secretion stress.
This focuses interest on the BdbC-dependent folding pathway for biotechnological
production of proteins with disulphide bonds in B. subtilis and related bacilli.
Experiments designed to determine whether the factors that are required to elicit
the CssRS response, show that membrane translocation of secretory proteins is
required for a CssRS-dependent secretion stress response. Notably, the intensity of the
secretion stress response only partly reflects the production levels of the model
proteins. Importantly, degradation of human Interleukin-3 by extracellular proteases
has a major impact on the production level, but only a minor impact on the intensity
of the secretion stress response (partner 10A).
Fig. 14. Protein quality control systems of the B. subtilis cell envelope
Partner 10B has performed a DNA macro-array study on cells combating
secretion stress to find possible new targets of the CssRS two-component system that
is involved in secretion stress management in B. subtilis. It was shown that secretion
stress affects not only the expression of the two house-keeping proteases encoding
genes, htrA and htrB, but also the expression of genes encoding proteins involved in
sporulation, motility, and amino acid utilisation. DNA array studies performed under
conditions of protein overproduction, with wild-type and cssR mutant strains, revealed
several indirect effects of protein overproduction. Additional targets for the two
component system were identified. One target, ykoJ, has been confirmed using
reporter studies. In addition, a high-yield easy purification system has been
developed, to allow confirmation of direct regulation by CssR, more detailed in vitro
investigations of CssR-binding characteristics, and the confirmation of a putative
binding motif. The indirect effects of overproduction of secretory proteins were found
to include abrogation of sporulation as determined by sporulation assays and
prolongation of the motile phase, as shown by motility and reporter assays.
To study the CssRS system of B. subtilis more mechanistically, partner 10B
fused the CssS protein C- and N-terminally to Green Fluorescent Protein (GFP). Cells
expressing the GFP-CssS fusion protein sometimes show a number of foci near or at
the membrane of B. subtilis cells. During other growth phases, however, the
fluorescence was distributed evenly throughout the membrane. This suggests that the
localisation of CssS is dynamic. The functionality of this fusion protein is currently
being investigated. A C-terminal fusion of GFP to HtrB has also been investigated.
The behaviour of this fusion protein was similar to that of CssS, with foci near or at
the membrane, and later during growth showing homogeneous membrane distribution.
Fusions to YFP and CFP which enable us to study co-localization of these proteins are
currently under construction.
The subcellular localisation of GFP-SipS fusions was also studied. The
localisation of this important protein involved in protein secretion appeared to be
dynamic. During exponential growth, the protein is localised in clusters on spiral-like
patterns (Figure 15A), similar to the localisation of the Sec-machinery. In the
stationary growth phase (Figure 15B), the protein is homogeneously distributed in the
membrane. The observation that, in B. subtilis, certain components of the Sec-
dependent secretion machinery are localised during certain stages of cellular growth,
holds promise for the development of improved production strains for protein
Fig. 15. Subcellular localisation of GFP-SipS
Studies have been undertaken on the N-terminal regions of both HtrA and HtrB
(partner 2). The N-termini of both proteins show no homology to each other or to any
other proteins in the protein databases. To ascertain the roles of the N-termini, this
partner has investigated how deletion of the putatively intracellularily located N
termini of both HtrA and HtrB proteins have affected the expression, topology and
induction of these proteins by the CssRS system. The N-termini of both HtrA (23aa)
and HtrB (50aa), preceding the transmembrane domains were deleted in-frame.
Transcriptional analysis by lacZ fusions showed a small but reproducible increase in
the N-terminally deleted strains. This increase was minor in comparison to lacZ fusion
activity in null htrA or null htrB mutant strains. Western immunoblot analysis
demonstrated that both N-terminally truncated proteins were detectable and
apparently stable. However, the level of the ∆N-HtrA protein was much reduced
while that of the ∆N-HtrB protein was only slightly reduced compared to the levels of
the wild-type proteins. In addition, the truncated HtrA protein is also processed and
secreted into the medium but is detected at lower levels. No wild-type or ∆NHtrB
protein has been detected in the medium. Furthermore, the response of the CssRS
system to induction by heat or secretion stresses was unaffected in these strains,
indicating no role for any transmembrane signal transduction function that might be
carrried out by the N-termini. Northern blot analysis of htrA transcripts from wild-
type and ∆NhtrA strains demonstrated a reduced amount of the ∆NhtrA transcript.
Furthermore, half life studies undertaken after addition of rifampicin to wild-type,
∆NhtrA and ∆NhtrB strains show that the ∆NhtrA transcript has a shorter half life
while the stability of the ∆NhtrB transcript is closer to that of the wild-type strain. A
deletion analysis of the N terminal region of htrA using lacZ fusions demonstrates
minor differences compared to that of the wild-type fusion. Furthermore, these
deletion fusions respond to secretion stress imposed by growth in the presence of a
protease inhibitor, Complete, in a similar fashion to the wild-type.
In contrast to the minor effects observed with deletion of the N-termini of
HtrA and HtrB, deletion of only 10 or 11 amino acid residues from the PDZ domains,
respectively, of HtrA and HtrB results in drastically reduced levels of both proteins.
Both deletions generate a null mutant phenotype with respect to levels of B-
galactosidase activity from fusions integrated into these strains (partner 2).
As stated above a processed form of HtrA is detected in the media while HtrB
is not observed. Partner 2 has shown that auto-cleavage is a mechanism through
which a truncated form of HtrA is liberated into the medium (partner 2).
6.13. Secretion/expression host/vector systems for B. subtilis
Rapid progress was made in the development of a new secretion/expression
host/vector system for the production of (biopharmaceutical) proteins by partner 2.
The knowledge can not yet be disseminated into the public domain.
A comparison was made by partner 10A of exoproteomes of cells in which
extracellular proteases were genetically or chemically inactivated (collaboration with
partners 2 and 6). The results showed substantial differences in the relative abundance
of various extracellular proteins, cell wall proteins in particular. Furthermore, a
comparison of the effects of genetic and/or chemical protease inactivation on the
stress response triggered by (over)production of secreted proteins showed that
chemical protease inactivation provoked a genuine CssRS-dependent secretion stress
response. From a physiological point of view, this suggests that deletion of protease
genes is a better way to prevent product degradation than the use of protease
inhibitors. Importantly, however, studies with human interleukin-3 show that
chemical protease inactivation can result in improved production of protease-sensitive
secreted proteins even in mutant strains lacking eight extracellular proteases.
The very useful SURE expression system for B. subtilis was developed by
partner 10B. Publication: Bongers et al.,:”Development and characterization of a
subtilin-regulated expression system in Bacillus subtilis: strict control of gene
expression by addition of subtilin”. Appl. Environm. Microbiol. 2005, 71: 8818-8824.
6.14. Sublancin production in B. subtilis
B. subtilis 168 is known to produce at least two bacteriocins, sublancin 168 and
subtilosin. Partner 10A has analysed the production of sublancin. Sublancin 168 is an
extremely stable molecule that has bacteriocidal activity against dangerous Gram-
positive pathogens, such as Streptococcus pyogenes and S. aureus. It was classified as
a type AII lantibiotic, displaying the extraordinary characteristic of having two
disulphide bonds in addition to a beta-methyllanthionine bridge. Thus, sublancin 168
belongs to the same family of products as the widely used bacteriocidal agent nisin
(E234). Both the production of sublancin 168 and the immunity against this
bacteriocin depend on the presence of the prophage SPbeta in the B. subtilis genome
(partner 10A). The sunA gene, which encodes sublancin 168, was identified by
sequencing the SPbeta region of the B. subtilis 168 chromosome. It is the first gene of
the previously described sublancin 168 locus, which consists of five successive genes
with the same transcriptional orientation: sunA, sunT, bdbA, yolJ, and bdbB. The sunT
gene encodes an ABC transporter required for export and processing of sublancin 168.
The bdbA and bdbB genes encode thiol-disulphide oxidoreductases. Previously, it was
demonstrated that BdbB is of major importance for the production of sublancin 168,
whereas BdbA is dispensable for this process. The potential glycosyl-transferase YolJ
protein was also found to be required for sublancin 168 production. The introduction
of site-specific mutations in YolJ has revealed that this protein is not involved in the
formation of disulphide bonds in sublancin 168. Instead, the results of these studies
imply that YolJ has a glycosyl-transferase activity that is indispensable for the
formation of active sublancin 168. This was inferred from two observations. First,
YolJ contains a domain that is conserved in known glycosyl-transferases. This domain
contains the DAD active site motif of glycosyl-transferases. Second, the mutation of
the DAD domain into NAN completely abolishes sublancin 168 activity. The NAN
mutation has no consequences for the stability of the YolJ, as shown by lumio-tagging
of this protein. This shows that the YolJ-NAN mutant protein is inactive. The reason
why a glycosyl-transferase activity of YolJ is required for sublancin 168 production is
currently being investigated.
Even though the SPbeta prophage is required for sublancin 168 immunity, the
specific SPbeta genes that are required for immunity against this bacteriocin were not
known at the start of this project. The studies of partner 10A have shown that the five
genes in the sublancin 168 locus are dispensable for sublancin 168 immunity. Instead,
one other SPbeta gene, yolF, was shown to encode a sublancin 168 immunity protein
that bears no resemblance to other known bacteriocin immunity determinants. These
studies have shown that the SPbeta gene yolF is both essential and sufficient for
sublancin 168 immunity. This involved both the deletion of yolF from the B. subtilis
chromosome, and reintroduction of yolF on a plasmid. Polyclonal antibodies against
YolF were raised. These were used for studies on the expression of chromosomal and
plasmid-borne copies of the yolF gene, and for investigating the subcellular
localisation of the YolF protein. The results show that chromosomal yolF is expressed
at a higher level than the plasmid-borne copy of yolF, which was used for showing
that YolF is responsible for sublancin 168 immunity. The YolF protein is membrane-
associated, most likely through one transmembrane domain with an Nout – Cin
orientation. The bulk of the YolF protein appears to be localised in the cytoplasm.
This suggests that the mode of action of YolF in bacteriocin immunity is completely
different from that of other known bacteriocin immunity proteins.
6.15. Improvements of Bacillus production strains
The main objective of partner 9 was to construct new expression plasmids and
introduce modifications to B. subtilis 168 to improve secreted recombinant protein
production. B. subtilis protein expression has traditionally relied on integration of
single-copy chromosomal expression cassettes, but elevated gene copy can improve
the yield and reduce fermentation time. Partner 9 has constructed expression plasmids
based on the Pgrac and Pxyl promoters and the amyL secrection signal sequence.
Where chromosomal modifications were required, these were introduced using a
technology (‘Xer-cise’; Figure 16) that avoided the permanent insertion of antibiotic
resistance genes. This includes the deletion of the sporulation sigma factor gene sigF
to create strains that do not form spores, and so have a greater chance of regulatory
Fig. 16. The Xer-cise gene deletion system
Partner 9 has also supplied mutants to the consortium with up to ten
extracellular protease mutations (the most available in any B. subtilis strain) to
evaluate recombinant protein expression, as secreted recombinant proteins are often
degraded by these proteases. Model proteins studied during this work included the B.
anthracis protective antigen (rPA), the collagenase genes colG and colH from
Clostridium histolyticum and the HIV CCR5 receptor blocker, RANTES (Figure 17).
Fig. 17. Model systems for the evaluation of putative production strains
Strain evaluation involved expression studies to assess the new host/vector
combinations. Protein expression was conducted in shake flasks and bench-scale 5 L
fermenters, using media free of animal derived components and optimised for B.
subtilis. The protein yield was assessed using SDS-PAGE and western blotting of
supernatant samples. The new expression plasmids and protease-deficient strains
enabled the production of rPA with a higher proportion of mature protein and fewer
degradation products than seen from production in unmodified strains. The production
of heterologous proteins was not successful, indicating that the Bacillus strains were
best suited for the secretion of Bacillus proteins. Additionally, the multiple protease-
deficient mutants were prone to lysis during fermentation.
Many, especially biotechnologically relevant, proteins rely on disulphide bonds
for their correct folding, structural integrity and activity. For the synthesis and
subsequent secretion of these proteins by B. subtilis, it is often the correct formation
of these disulphide bonds that represents the greatest bottleneck. Degradation of
inefficiently or incorrectly oxidized proteins and the requirement of costly and time-
consuming reduction and oxidation steps in their downstream processing still set a
major limit to a full exploitation of B. subtilis for biopharmaceutical production.
Therefore, partner 10A undertook studies aimed at developing a novel in vivo strategy
for improved production of secreted disulphide bond-containing proteins. In order to
achieve this, the expression levels of many of the (potential) TDORs of B. subtilis
were modulated and/or TDORs from heterologous Gram-positive organisms were
introduced. Three combined approaches were found to be most successful: depletion
of the major cytoplasmic reductase TrxA; introduction of the heterologous oxidase
DsbA from Staphylococcus carnosus; and addition of redox-active compounds to the
growth medium. As shown with the disulphide bond-containing E. coli PhoA as a
model protein, the combined use of these three approaches allowed for the secretion
of ~3.5 times increased amounts of active PhoA, as compared to the parental strain B.
subtilis 168 (Figure 18). These findings imply that super-oxidizing Bacillus strains
can be engineered for the biotechnological production of heterologous high-value
proteins containing disulphide bonds.
Fig. 18 Increased production of E. coli PhoA by engineered B. subtilis strains. (A) B.
subtilis strains were transformed with pPSPhoA5 for E. coli PhoA production. All strains and the parental strain
168 (168) were grown overnight in LB medium containing 0.5% xylose (white bars) and an additional 100 mg/ml
cystine (grey bars) or cysteine (black bars). Next, growth medium samples were withdrawn for alkaline
phosphatase activity assays (A) as well as SDS-PAGE and Western blotting with specific antibodies against PhoA
(B). Lysates of cells from cultures corresponding to the samples in A and B were analyzed by Western blotting
with specific antibodies against S. aureus DsbA (C), BdbD (D) or TrxA (E). PhoA activity is given in
U/ml/OD600 relative to the PhoA activity of the parental strain expressing PhoA (100% = 6.0 U/ml/OD600). The
arrow in panel B indicates the position of mature PhoA (mPhoA). Bands with a higher mobility on SDS-PAGE are
breakdown products of PhoA, bands with a lower mobility are unprocessed forms of PhoA. The arrow in panel C
indicates the position of DsbA. Note that the antibody against S. aureus DsbA is cross-reactive with S. carnosus
DsbA. Molecular weight markers are indicated.
6.16. Web interface for genome-scale prediction of Bacillus secretomes
A web interface was developed for genomic-scale prediction and
characterisation of secreted proteins from Bacillus species (partner 1). The approach
used e-science tools such as myGrid and Taverna to analyse distributed data. The data
has been used to characterise and analyse the secretomes of 12 Bacillus genomes,
analysing the relationships between the secretomes of pathogenic and non-pathogenic
species. 7% of the secreted proteins were shared between all of the analysed species.
The files generated from the in-silico identification and characterisation of
secreted proteins has been used to categorise and analyse secreted protein families
from across and between species. The analysis of protein families and their function
has been in relation to their environment including, in the case of the pathogens, their
animal hosts. This has resulted in the prediction of adaptive traits from the secretome
and the identification of 15 protein families (2% of the secreted proteins) that are
unique to the pathogens. Interestingly, the majority of these proteins (13/15) are
currently of unknown function (partner 1).
A protein interaction network for B. subtilis (SubtilNet) was developed by
partner 1 and this network analysis was extended to all of the currently sequenced
members of the genus Bacillus. The integration of many data sets to produce a single
functional integrated model for Bacillus relies heavily on the fact that the data sets
must be of comparable reliability. However, with different laboratories using different
technologies to obtain them, this can’t be guaranteed. These factors have been
overcome by taking data confidence into account when integrating data, an approach
that allows even potentially weak lines of evidence to be combined to indicate
stronger functional relationships that would not normally be seen. The confidence
levels of data sets are calculated using Bayesian statistics, producing weighted data
that is applied to produce an integrated network of interactions. The combination of e-
Science tools, workflows and Grid technology means that the analysis is
automatically reproduced every time a new Bacillus genome is published in the public
The comparative genomic approach, coupled with protein family analysis, can
help identify protein function on a guilt-by-association principle. The identification of
secreted proteins that are unique to pathogens provides the basis for the development
of novel drug targets and vaccines (partner 1).
Partner 10B has made secretome predictions with B. cereus and B. subtilis, and
these are being used to develop a web-based prediction service for the secretory
proteins of bacteria, based on annotated genome data (Figure 19). Good progress has
been made toward the completion of the prediction service. Algorithms have been
developed to identify the secretory proteins encoded by all currently sequenced
Bacillus genomes and, by identifying their targeting signal, to predict their final
locations (Figure 20). These are being extended to other sequenced Gram-positive
bacteria. In addition, a combination of matrix and clustering analyses approaches are
being used to identify secretory proteins that are components of the core Bacillus
genome, and to use these to analyse the genotypic relationship between the sequenced
A secretome predict-
ion for B. cereus strains has Fig. 19.
been performed by partner
10B. Whereas B. cereus
14579 is a non-pathogenic
type strain for B. cereus
species, the B. cereus zebra
killer was originally
isolated from a swab of the
carcass of a dead zebra
suspected of having died of
an anthrax-like disease.
The Blast algorithm was
used to predict which
components comprise the
major secretion pathway
(Sec), of B. cereus ZK
(Zebra Killer). It has been
shown that, in contrast to B.
cereus ATCC 14579, B. cereus ZK encodes more than one homologue of a number of
key Sec pathway components. The secretome of B. cereus ZK has been determined. A
further comparison between the predicted secretomes of both B. cereus strains still has
to be performed. This comparison could provide insight in novel pathogenicity factors
that are secreted by the second Sec translocase present in B. cereus ZK.
Fig. 20. Secretome prediction algorithm
6.17. Software development
The MG-Web framework was developed by partner 10B to convert a tool
workflow into web-based software. MG-Web allows a bioinformatics researcher to
Fig. 21. Flowchart of the
process in Prosecutor.
expression profiles from
DNA microarrays (A) are
used to create a
correlation matrix (B).
For every gene, the
correlations with the
remaining genes are
retrieved from the
correlation matrix and
The sorted gene list is
used to perform an
iterative Group Analysis
for every functional
category (B3). The
resulting p-value is
indicative for the coupling
of a gene to a functional
The complete list of p-
values for every
functional category is
sorted (C4), after which
the positions of the
members of the functional
category is determined
(Fig C5). These positions
are used to create ROC
The corresponding Area
Under the ROC Curve
(AUC) is then used as a
measure of the
performance of the genes
of a functional category.
generate with a minimum of time investment robust and professional web-based
software. Several available software packages have been implemented; e.g.,
CorrelGene and BAGEL.
Datasets obtained from “omics” techniques often need some form of processing
prior to interpretation. The use of standard protocols for data processing is vital in
order to achieve reproducible results. “Pipeline” has been developed as an extensible
software system that performs automated processing of “omics” datasets through
adaptable predefined workflows. The researcher selects (and optionally adapts) a
protocol via user-friendly web-forms, datasets are processed according to the
workflow, and finally detailed results are presented to the researcher in a web-page.
The FIVA software was successfully implemented in the studies of B. subtilis
iron uptake mediated by siderophores (partners 3 and 10B).
Several genes without known function have been linked to functional annotations
involved in pathogenicity using the Annotation Gap software (partner 10B). An
exploratory application, named DISCLOSE, was developed that benchmarks methods
to characterise clusters derived from DNA microarray data, using functional
annotations and a de novo DNA motif discovery algorithm. The performance of this
software was evaluated by comparing the original results of a time course experiment
with the findings of this application.
“Prosecutor” was developed by partner 10B as an application that enables
researchers to infer gene function based on gene expression data and functional
annotations (Figure 21). This functional coupling method uses a sensitive algorithm to
achieve a high association rate of genes with unknown functions to annotated genes.
This partner supplies the complete results of our analysis for 11 organisms. This
dataset will allow researchers working on either one of these organisms to quickly
identify putative gene functions of genes of interest.
7. ACHIEVEMENTS VERSUS STATE-OF-THE-ART
Pre-genomic research in cell biology has yielded a wealth of knowledge about
individual regulatory pathways and metabolic processes that are obligatory for the
survival of pathogens in their host, and for the productivity of microbes in industrial
bioprocesses. The major challenge for the BACELL HEALTH consortium, using
state-of-the-art post-genomic technologies, was to understand how individual
regulatory pathways are networked to maintain cellular homeostasis. The networking
of individual regulatory pathways ensures that the cell provides a balanced response
to stress, sensing both the magnitude of the stress, and the effectiveness of the
response. In the case of pathogens, the identification of key nodes in these regulatory
networks was expected to provide new targets for the development of antimicrobial
compounds that perturb or disrupt the cell stress management system. For industrial
production strains, the inactivation of stress-induced processes that limit the
production of heterologous proteins was expected to lead to the development of a new
generation of host/vector systems for the production of pharmaceutically-active
7.2. Progress in relation to the plans
Virtually all major goals for the project were realised and all partners have
carried out their planned work in a productive and successful way (see section 7.3 for
an overview of the end-status of the project in relation to its plans for workpackages
Relatively few problems have been encountered. However, there has
been a need to modify priorities in Workpackage 2, leading to a postponement in
establishing the macrophage assay system. This was due to the technical limitations
not recognised at the start of the project. These included limitations in the ability of
established macrophage cell lines to pump iron into the vacuole, limiting its ability to
mount an effective oxidative burst. The recognition that B. anthracis survives but
does not proliferate extensively in the macrophage, limiting the amounts of mRNA
available for array analysis. The technical problems are continuing to be addressed
after the end of the contract. This includes the development of a novel sensing system
that reports on physiological changes in the macrophage lysosome. Instead, partner 1
has focused on establishing a quantitative PCR for measuring the oxidative stress
response in macrophages. This will allow the expression of key genes of the
macrophage-generated oxidative response to be monitored. The combination of
macrophage physiological sensors and B. anthracis targeted RT-PCR are currently
being used to an appropriate macrophage line to complete these studies.
Despite the absence of information obtained from the macrophages,
knowledge of the response of B. anthracis to oxidative stress obtained throughout the
programme has greatly exceeded that proposed in the contract. Of particular
importance is the role of iron homeostasis that both protects the pathogen from highly
damaging hydroxyl radicals and predisposes the cell for later events in the obligatory
Initial delays in starting Workpackage 4 were overcome in years 3 and 4.
7.3. Workpackages – Plan and Status Bar chart
Project timetable Red line: Current status of project
0 8 16 24 32 40 48 Milestones
` 1a. Establishment and analysis of biofilms
1b. Analysis of the DegSU regulon
1c. DNA array analysis of alkaline stress
1d. Analysis of putative alkaline stress regulators
1e. Analysis of anti-sigma factor RsiW
1g. Analysis of SigB acid stress sensors
1h. Analysis of the oxidative stress stimulon
1i. Establishing link between CtsR and oxidation
1j. Roles of thiol-disulphide oxidoreductases
1k. Construction of the disulphide proteome
1l. Design and testing of inhibitors of DhbE
1m. Additional siderophore-dependent inhib’targets
1n. Specificity of interaction between HK and RR
2a. Extraction of high-quality mRNA from B. anthracis
2b. DNA array analysis of infection stress-induced genes
2c. Proteomic analysis of infection stress-induced proteins
2d.Secondary protein translocase analysis of B. anthracis
2e. In vivo and in vitro DhbE inhibition assays
2f. Analysis of YycFG system in pathogens
2g. YycF regulon proteins as anti-infective targets
2h. Roles of DegSU and Agr in Listeria
2i. Role of HtrA-like proteases in pathogens
2k. Other stress-related genes in G+ve pathogens
3a. Thiol-diS oxidoreductases’ influence on secretion
3b. Construction of thiol-diS oxidoreductases mutants
3c. Analysis of hybrid protein-induced secretion stress
3d. The influence of cssA mutants on secretion stress
3e.The influence of htr mutants on secretion stress
3f. New host/vectors for bio-production
3g. Evaluation of putative production strains
3h. Influence of diS on the production of defensin-like mols’,
3i. Localisation of secretion stress components of B. subtilis
4a. Extension of Microbase with consortium data
4b. Identification of stress genes in G+ve pathogens
4c. Identification of cis-elements interacting proteins
4d. Expansion of Bsu regulatory network Petri net
4e. Identification of regulatory network nodes
4f. Extension of regulatory network to pathogens
4g. Use of regulatory model to identify drug targets
5a. Internal reports
* * * * * 5b. Consortium meetings
5c. Mid-term review
5d. EU reports
5e. Dissemination activities (meetings/journals)
* * * * * 5f. Assessment of IP and possible commercialisation.
Although the RED line in this Table indicates the status at the end of the
project, several lines of research will be continued after the contract.
7.4. Major achievements
In nearly all aspects of the Stress Management System studied in this project,
major progress in relation to the state-of-the-art was achieved. Several of the results
and findings are at the cutting edge of research in their areas.
The consortium has made very significant progress with respect to each of its
three major aims:
• Modelling of the regulatory networks that comprise the cell stress
• Identifying key elements within the cell stress management system and stress
resistance processes as targets for the development of novel anti-infectives
• Improving Bacillus subtilis as a host for the production of biopharmaceutical
7.4.1. Stress systems
Several stress systems were characterised in depth. These include: pH stress,
oxidative stress, secretion stress, iron limitation stress, and cell wall stress. The
progress in these fields is described below.
• The alkaline stress stimulon was nearly completely identified. Many of the
genes in this stress stimulon appeared to be under the control of the ESF sigma
factor σW. The alkaline-inducible σW system proved to constitute an antibiosis
regulon. The mechanism of activation of SigW was characterised to a large
extent. In the activation of the precursor form of SigW, regulated
intramembrane proteolysis appeared to be critical. The results also point to the
important role of ECF sigma factors in stress-induced cellular responses in B.
subtilis and, probably, many other bacteria. These findings are novel and have
provided important insights in stress-induced cell stress management systems.
This work is at the cutting edge of research in this area.
• The acid stress stimulon was elucidated to near completeness. This stimulon
appeared to involve about 100 genes, which were shown to be mainly under
control of the general sigma factor, σB, and some others were under the control
of σM. Several candidate proteins for the sensing of acid stress were identified.
These findings are novel.
• Much attention was given to the analysis of the oxidative stress stimulon and
very significant new results were obtained. This stimulon was nearly
completely defined for B. subtilis. A new fluorescence-based thiol
modification assay was developed and novel thiol-specific regulators (YodB
and MhqR) were identified. S-thiolation by cysteine appeared to be crucial for
the protection of thiol-containing proteins against oxidative stress.
The oxidative stress stimulon of B. anthracis was also analysed to
considerable detail and important differences with the B. subtilis oxidative
stress response were observed.
Major novel knowledge was obtained from the analyses of thiol-
disulphide oxido-reductases (TDOR) in B. subtilis and S. aureus. TDORs
facilitate disulphide bond formation in vivo. Such disulphide bonds are critical
for the stability and/or activity of various proteins. Comparison of TDORs in
these organisms revealed important differences but, nevertheless, similar
functions were observed.
A system, called mixed disulphide fishing, was set up for the mapping
of the disulphide proteome of B. subtilis. Promising results were already
obtained with thioredoxin, TrxA. The latter protein was crystallised and its
structure determined at the 1.5Å level.
• Many pathogens employ siderophores as virulence factors to multiply in the
colonised host. Iron limitation is, therefore, a stress condition. In this project,
cutting edge research was carried with respect to iron acquisition by B.
subtilis. The bacillibactin (B. subtilis siderophore) pathway was nearly
completely unravelled. Based on the knowledge of this pathway, components
(such as DhbE) were identified as potential drug targets in pathogenic bacteria.
• The overproduction of proteins containing disulphide bonds appeared to result
in increased CssSR-dependent secretion stress. This effect was modulated
by BdbC, a member of the TDOR family of proteins.
• CssSR-dependent secretion stress was also shown to result in the abrogation
of sporulation and prolongation of the motile phase of B. subtilis.
• CssRS components appeared to be localised near or at the membrane, at least
during certain stages of cellular growth.
In summary: all these findings are novel, although the implications and potential
benefits are, as yet, not always clear.
Several links were identified between various stress responses. The CtsR
regulator appeared to be involved in both the heat stress and the oxidative stress
response. The ClpXP protease/chaperone system was found to be involved in the
activation of SigW (alkaline stress response) and was already known to be involved in
the general stress response. These findings are important for the understanding of the
interacting networks in the cellular Stress Management System.
7.4.2. Bioinformatics application/modelling tools
In this project several new bioinformatics applications were developed which
are useful for network modelling, analysis of DNA microarray data, and comparative
genome, transcriptome and proteome analyses. These include software for:
• Secretome predictions/comparisons;
• “SubtilNet”; for protein interaction networks;
• “MG-web’; for converting tool workflow into web-based software;
• “Pipeline”; for the automated processing of ”omics” data;
• “FIVA”; for the analysis of single DNA microarray experiments;
• “DISCLOSE”; a method to characterise clusters derived from DNA
• “Prosecutor”; for inferring gene functions based on expression and functional
In summary: this part of the project has resulted in the production of a whole variety
of valuable bioinformatics application tools.
7.4.3. Microbial virulence
The project has greatly contributed to the increased understanding of microbial
virulence, in particular of certain Gram-positive pathogens. Although B. subtilis was
used as the model in many of these studies, various aspects of virulence of Gram-
positive pathogens, like B. anthracis, B. cereus, Listeria monocytogenes and
Staphyloccoccus aureus, were studied. Several of the findings are relevant for the
identification of possible new drug targets. The major novel results are summarised
• The YycFG two-component system was found to be essential in B. subtilis
through its role in cell wall metabolism. Similarly, the WalKR system (WalKR
is the YycFG homologue of S. aureus) was essential in S. aureus, also through
a role in cell wall metabolism. These findings point to this two-component
system as a possible target for novel drugs. WalKR was also important for
biofilm formation in S. aureus.
• The response regulator DegU appeared to be crucial for adaptive responses in
L. monocytogenes: motility en biofilm formation. Different from B. subtilis, no
sensor kinase of the DegS-type is present in L. monocytogenes and the
phosphorylation of DegU seems to be possible from small phosphodonors,
such as acetyl phosphate.
• The Agl quorum sensing system was found to have a role in the virulence of L.
monocytogenes. This finding identifies the Agl quorum sensing system as a
potential novel drug target.
In summary: this part of the project has resulted in the identification of several
potential new drug targets in Gram-positive pathogens. This is likely to provide leads
for the development of novel anti-infectives.
7.4.4. Bioprocessing; improvements in protein productions
Bioprocess stress constitutes an important limitation to the production of
native and heterologous proteins, also in B. subtilis. A major bottleneck is the
presence of a whole variety of cell envelope-associated and extracellular proteases,
which degrade exported proteins, in particular if the latter are incorrectly folded. In
the process of proper folding of disulphide bond-containing proteins, the Bdb proteins
of the TDOR family in B. subtilis play an important role.
Several of the major novel results concerning the improvement of production
strains and production strategies mainly deal with the removal or reduction of
imposed secretion stress. The most important results are summarised below.
• By modulating the conditions for the formation of disulphide bonds,
considerable improvements were made with respect to the secretion by B.
subtilis of a model protein containing disulphide bonds. This provides leads
toward the improved production of disulphide bond-containing proteins.
• A “clean” system for the deletion of genes from the B. subtilis chromosome
was developed. “Clean” in this context refers to the fact that no antibiotic
marker is left in the chromosome as a result of the deletion.
• Strains with improved gene expression properties and stable plasmid cloning
systems were generated.
• The SURE expression system was introduced in B. subtilis.
• As a model for bio-active peptides, the production/activity/immunity of B.
subtilis sublancin 168 was studied in detail. Nearly all of the genes involved in
these processes were identified. These findings provide new leads toward the
use of B. subtilis for the production of bio-active peptides.
In summary: in this part of the project, several improvements to production strains
have been realised and leads for further improvements were provided.
7.5. Impact on industry and research sector
European research groups and industries are global leaders in the development
of Bacillus technology and the commercial exploitation of bacilli. A direct industrial
benefit of BACELL HEALTH is therefore to help maintain Europe’s competitive
market position in the face of competition from the US and the Far East. The
importance of this work for industry is also reflected by the participation of three
industrial partners in this project. All three of these use Bacillus species for
The project has impacted on human health by providing knowledge of the
mechanisms which bacteria use to avoid the immune response.
The results described in section 7.4, in particular sub-sections 7.4.3 and 7.4.4,
show that this project has a considerable impact on the use of B. subtilis (and related
Bacillus species) as organisms for the production of (pharmaceutical) proteins.
In addition to improved production strains and improved production strategies,
the current project has also resulted in the identification of key genes for virulence
traits in certain Gram-positive pathogens. This provides leads toward the
identification of novel drug targets and, finally, the production of novel anti-
At least three patent applications are likely to emerge from this work, which
will be filed by partners in the project.
The impact of this project on the relevant research sectors can be deduced
from the results described in section 7.4.
In addition, more than 140 publications will appear in the international, peer-
reviewed, scientific literature. About 40% of these are directly based on the results
from this project, whereas the remainder is indirectly linked to the work in this
During the contract period, consortium members have presented more than
150 oral presentations or posters at national and international scientific meetings, on
results which are directly or indirectly related to the project.