Identification of network topological units coordinating the global expression response to glucose in Bacillus subtilis and its comparison to Escherichia coli.

Abstract

Glucose is the preferred carbon and energy source for Bacillus subtilis and Escherichia coli. A complex regulatory network coordinates gene expression, transport and enzymatic activities, in response to the presence of this sugar. We present a comparison of the cellular response to glucose in these two model organisms, using an approach combining global transcriptome and regulatory network analyses.
Transcriptome data from strains grown in Luria-Bertani medium (LB) or LB+glucose (LB+G) were analyzed, in order to identify differentially transcribed genes in B. subtilis. We detected 503 genes in B. subtilis that change their relative transcript levels in the presence of glucose. A similar previous study identified 380 genes in E. coli, which respond to glucose. Catabolic repression was detected in the case of transport and metabolic interconversion activities for both bacteria in LB+G. We detected an increased capacity for de novo synthesis of nucleotides, amino acids and proteins. A comparison between orthologous genes revealed that global regulatory functions such as transcription, translation, replication and genes relating to the central carbon metabolism, presented similar changes in their levels of expression. An analysis of the regulatory network of a subset of genes in both organisms revealed that the set of regulatory proteins responsible for similar physiological responses observed in the transcriptome analysis are not orthologous. An example of this observation is that of transcription factors mediating catabolic repression for most of the genes that displayed reduced transcript levels in the case of both organisms. In terms of topological functional units in both these bacteria, we found interconnected modules that cluster together genes relating to heat shock, respiratory functions, carbon and peroxide metabolism. Interestingly, B. subtilis functions not found in E. coli, such as sporulation and competence were shown to be interconnected, forming modules subject to catabolic repression at the level of transcription.
Our results demonstrate that the response to glucose is partially conserved in model organisms E. coli and B. subtilis, including genes encoding basic functions such as transcription, translation, replication and genes involved in the central carbon metabolism.

Full text

Page 1

BioMed Central
Page 1 of 14
(page number not for citation purposes)
BMC Microbiology
Open Access
Research article
Identification of network topological units coordinating the global
expression response to glucose in Bacillus subtilis and its comparison
to Escherichia coli
Carlos Daniel Vázquez1, Julio A Freyre-González1, Guillermo Gosset2,
José Antonio Loza3 and Rosa María Gutiérrez-Ríos*1
Address: 1Departamentos de Microbiología Molecular, Instituto de Biotecnología, Universidad Nacional Autónoma de México, Apdo. Postal 510-
3, Cuernavaca, Morelos 62250, México, 2Ingeniería Celular y Biocatálisis, Instituto de Biotecnología, Universidad Nacional Autónoma de México,
Apdo. Postal 510-3, Cuernavaca, Morelos 62250, México and 3Instituto de Ciencias Nucleares, Universidad Nacional Autónoma de México
Apartado Postal 70-543, 04510 México D.F., México
Email: Carlos Daniel Vázquez - cvazquez@ibt.unam.mx; Julio A Freyre-González - jfreyre@ccg.unam.mx;
Guillermo Gosset - gosset@ibt.unam.mx; José Antonio Loza - electron22@gmail.com; Rosa María Gutiérrez-Ríos* - rmaria@ibt.unam.mx
* Corresponding author
Abstract
Background: Glucose is the preferred carbon and energy source for Bacillus subtilis and Escherichia coli.
A complex regulatory network coordinates gene expression, transport and enzymatic activities, in
response to the presence of this sugar. We present a comparison of the cellular response to glucose in
these two model organisms, using an approach combining global transcriptome and regulatory network
analyses.
Results: Transcriptome data from strains grown in Luria-Bertani medium (LB) or LB+glucose (LB+G)
were analyzed, in order to identify differentially transcribed genes in B. subtilis. We detected 503 genes in
B. subtilis that change their relative transcript levels in the presence of glucose. A similar previous study
identified 380 genes in E. coli, which respond to glucose. Catabolic repression was detected in the case of
transport and metabolic interconversion activities for both bacteria in LB+G. We detected an increased
capacity for de novo synthesis of nucleotides, amino acids and proteins. A comparison between
orthologous genes revealed that global regulatory functions such as transcription, translation, replication
and genes relating to the central carbon metabolism, presented similar changes in their levels of
expression. An analysis of the regulatory network of a subset of genes in both organisms revealed that the
set of regulatory proteins responsible for similar physiological responses observed in the transcriptome
analysis are not orthologous. An example of this observation is that of transcription factors mediating
catabolic repression for most of the genes that displayed reduced transcript levels in the case of both
organisms. In terms of topological functional units in both these bacteria, we found interconnected
modules that cluster together genes relating to heat shock, respiratory functions, carbon and peroxide
metabolism. Interestingly, B. subtilis functions not found in E. coli, such as sporulation and competence were
shown to be interconnected, forming modules subject to catabolic repression at the level of transcription.
Conclusion: Our results demonstrate that the response to glucose is partially conserved in model
organisms E. coli and B. subtilis, including genes encoding basic functions such as transcription, translation,
replication and genes involved in the central carbon metabolism.
Published: 24 August 2009
BMC Microbiology 2009, 9:176doi:10.1186/1471-2180-9-176
Received: 6 March 2009
Accepted: 24 August 2009
This article is available from: http://www.biomedcentral.com/1471-2180/9/176
© 2009 Vázquez et al; licensee BioMed Central Ltd.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0),
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Page 2

BMC Microbiology 2009, 9:176http://www.biomedcentral.com/1471-2180/9/176
Page 2 of 14
(page number not for citation purposes)
Background
During the last decades, an increase in the quantity of
available data referring to biological systems has enabled
the development of new paradigms and methods for their
analysis, with the purpose of formulating coherent opin-
ions regarding cellular events, both locally and globally.
Recently, a network based approach for the representation
of cellular component interactions has proven highly suc-
cessful, when applied to the study of genetic expression
regulation and the mechanics of cellular metabolism [1].
This approach permits the identification of the effects
caused by interactions among proteins and other cellular
components; thus for the first time presenting the possi-
bility of visualizing the cell as a system. In the light of the
successful results obtained when applying this approach
to the model organism Escherichia coli [2]; this type of
analysis is now being applied to other organisms such as
the soil bacterium Bacillus subtilis [3].
For many decades B. subtilis has represented the most
important model for the study of firmicutes. Its genome
includes 4106 predicted genes, with a G+C content of
43.5%. Currently, the functions of about half of the pre-
dicted genes are known. At the time when E. coli became
the most important bacterial model, the study of B. subtilis
was initiated, partly due to its relative facility for genetic
manipulation, but also in large part due to its capacity to
form spores [4,5]. Currently, B. subtilis continues to be
employed as an important biological model, especially
for a large number of studies related to genetic regulation
and metabolism. Furthermore, B. subtilis is an organism
which attracts considerable commercial interest, as for
many years it has been used as an industrial producer of
enzymes and metabolites.
B. subtilis is a free living bacterium and therefore, it must
adapt to changes in its environment, for example nutrient
availability or fluctuations in temperature. Among nutri-
ents, sugars and other carbon sources are particularly
important, as these usually also provide the cell with met-
abolic energy. Microbes are constantly sensing the levels
and types of carbon sources present in the environment.
This function is carried out in most bacteria, including B.
subtilis, by the phosphoenolpyruvate: sugar phospho-
transferase system (PTS) [6]. The PTS is a protein system
composed of general and sugar-specific components. The
enzyme I (EI) and the phosphohistidine carrier protein
(HPr), relay a phosphoryl group from phosphoenolpyru-
vate (PEP) to the sugar-specific proteins IIA and IIB. The
last component of this system, IIC (in some cases also
IID), is an integral membrane protein permease that rec-
ognizes and transports the sugar molecules, which are
phosphorylated by component IIB. There are several PTS
component II encoded in the genome of B. subtilis, each
one having a specific sugar as substrate [7].
B. subtilis displays a pattern of preferential carbon source
consumption, depending on their varying metabolic
rates, which in turn result in differing growth rates. Glu-
cose is considered the preferred carbon source as it sus-
tains the highest growth rate and the same applies in the
case of E. coli [7]. Repression of the genes involved in the
metabolism of sugars is part of a global phenomenon
known as carbon catabolite repression (CCR). In B. subti-
lis, this phenomenon occurs due to PTS-mediated phos-
phorylation of regulatory proteins and GlcT controlling
antitermination. In most cases, CCR is defined by the
presence of catabolic responsive elements sites (CRE) in
the 5' regions of the regulated genes. The CRE DNA
sequences are recognized by the catabolite control protein
A (CcpA), whose repressed gene encoding functions relate
to the utilization of alternative carbon sources and other
stress conditions, in the presence of a preferential carbon
source, such as glucose [8,9].
A global view of the cellular transcriptional response can
now be accomplished using microarray technology. This
type of of study provides an instantaneous snapshot of the
way cells function, under specific conditions. The data
generated using this technology is useful for revealing the
nature of the complex regulatory interactions in the cell.
At the present time several reports exist, describing the use
of microarrays to study B. subtilis under diverse condi-
tions; for example in the presence of acid [10], in response
to thermic shock [11], anaerobiosis [12] and in the pres-
ence or absence of glucose [8], among others. These
results provide data that will enable the construction of a
detailed regulatory network and help to elucidate how
regulatory proteins interact with their effectors.
In this work, we analysed the regulatory network of B. sub-
tilis, when grown in a complex medium in the absence or
presence of glucose. This study enabled the identification
of network modules, coordinating the response of genes
with related functions. The results obtained were com-
pared to those from our previous study where E. coli was
employed[13].
Results
Global transcriptome response to the presence of glucose
in complex medium, in Bacillus subtilis
We performed an analysis of transcriptome data obtained
from previous reports of experiments, employing B. subti-
lis [8]. Following the procedure described in the methods
section, 504 genes were found to display significant differ-
ential expression, when grown in either the absence or
presence of glucose and these were compared (see Addi-
tional File 1: Table 1SM). In figure 1, we present the genes
with known functions, where transcription was found to
consist of a response to the presence of glucose in LB
medium (LB+G). Among this set of genes, we found those

Page 3

BMC Microbiology 2009, 9:176 http://www.biomedcentral.com/1471-2180/9/176
Page 3 of 14
(page number not for citation purposes)
induced in the presence of glucose, to be related to trans-
port and metabolism, for example the general PTS protein
enzyme I and the glucose-specific IICBGlc permease, as
well as the pgk, pgm, eno and pdhC genes, which encode
enzymes from the glycolytic pathway. The transcriptional
activation of the aforementioned genes is expected to
increase the cellular glucose capacity for transport and
catabolism. On the other hand, down-regulation was
observed in the case of genes encoding most of the
enzymes from the TCA cycle and the glyoxylate bypass [7].
A clear glucose-dependent repressive effect was observed
for genes encoding transporters, periplasmic receptor pro-
teins and enzymes related to the import and catabolism of
alternative carbon and nitrogen sources; for example car-
bohydrates, amino acids, lactate, glycerol 3-P, oligopep-
tides, dipeptides and inositol [7]. This transcriptome
pattern is the expected result of CCR, exerted by glucose.
Interestingly, we detected a general trend towards down-
regulation in LB+G medium, in the case of genes encoding
heat shock proteins and chaperones. This response sug-
gests a higher stress condition and a higher protein turno-
ver rate among cells growing in medium, which lacked
glucose. Contrastingly, the presence of glucose caused an
increase in the transcript level for genes encoding ribos-
ome constituents. This response is consistent with the
improved growth conditions provided, with the presence
of glucose.
We also detected, lower transcript levels in the presence of
glucose for gene encoding proteins involved in sporula-
tion. This included regulatory proteins, enzymes and
A metabolic view of the transcriptome profile of B. subtilis, comparing growth in LB+G to that in LB
Figure 1
A metabolic view of the transcriptome profile of B. subtilis, comparing growth in LB+G to that in LB. Genes dis-
playing higher and lower transcript levels, due to the presence of glucose are shown in red and green respectively. Abbrevia-
tions: AcCoA, acetyl coenzyme-A; Ac~P, acetyl phosphate; AKG, α-ketoglutarate; CIT, citrate; F1,6BP, fructose-1,6-
bisphosphate; F6P, fructose-6-phosphate; FUM, fumarate; G3P, glycerol-3-phosphate; G6P, glucose-6-phosphate; ICIT, isoci-
trate; MAL, malate;OAA, oxaloacetate; PEP, phosphoenolpyruvate; PYR, pyruvate; SUC, succinate; SUCCoA, succinyl-CoA;.
G2P 2-phospho-glycerate.
Glucose
IIBGlc
G6P
F6P
F1,6BP
Hpr
IIAGlc
IICGlc
Heat shock proteins and chaperones: tig, dnaK, groEL,
groES, grpE
Nucleotide metabolism: adk guaA guaD
pnpA purQ drm guaB yerA
Acetate
Lactate
dppE
Dipeptides
Ribose
Glutamine
glpT
Glycerol 3-P
iolF
Inositol
(minor)
iolT
Inositol
(major)
msmX
Various sugars
licAB
IICLic
Lichenan
IICMal
treP
Trehalose
PEP
PYR
EI ptsI
AcCoA
pdhC
pta
Ac~P
ackA
Acetate
OAA
MAL
CIT
acnB
AKG
SUCCoA
SUC
FUM
ICIT
gltA
ldhA
odhAB
sdhABC
mdh
citG
icd
aceA
Regulatory proteins: ywbI arfM. alaR,,yesS,cggR ahpC,
acoR,arsR manR,,rbsR,ykoM,yraB,yvdE,ydeS
IIBCSucr
IIASucr
Sucrose
G3P
Lactate
lctP
Cysteine
Amino acid metabolism: glmS glnA patA
rocA speD yjlA ansB aprX argJ bltD bpr
csd cysC cysK gudB yncD
Biosynthesis of cofactors: menB ybcP entC gshA
ispA moaB yhfU
Transcription functions: fusA rpoA rpoB rpoC
mutS
Arabinose
araN
Ribosome constituents: rpmI, rpmJ, rpsM, rpsL, rpsK, rplA,
rpsB, rpsF, rpsC, rpsD, rpsE, rpsI, rpsH, rpsG, rplV, rpsJ,
ybxF, rplD, ytiA, rplQ, rpsN, rplC, rplE, rplF, rplJ, rplK,
rplX, rplM, rpmH, rplP, rplR, rplB, rplS, rplU, rplW, rpmA,
rpmC, rpmF, rplL, rpsR, rpsS, rpsT, rpsQ, rplN, rpsU
Membrane bioenergetics (respiration): atpC atpD
atpG narH qoxA qoxB qoxC yoaE ldh narG yrhE
yumC
Lipid metabolism: accC des glpQ mmgA yngG
Protein translocation (secretion): ffh
Proteins - translation and modification: infA infC map tsf
prsA ydiD
Cell division and replication:
gyrA,ligB, ftsH, ssb
Specific carbohydrate pathways: xylA acuB adhB alsS
araB bglH galK glpK gutB idh iolC iolH iolI lacA
licH manA pta rbsK sacA treA ydjL yjaV ykvQ yngE
ytcB ytdA yusZ yveB
Oligopeptides
sacP
IIABFru
levDE
Protection responses and detoxification:
aadK, yceD, yceC, yceF, ycsF,ydhU, ahpF, sodA,ahpC
IICFru
Fructose
IIABLic
IIBCTre
G2P
pgk
pgm
eno
ptsG
Cell wall components: comGB dltA dltB dltD dltE
gcaD mbl murAA wapA murAB wprA ymaG ytcC
yvcE
tcyJKL
glnP
rbsABCD
appC
Sporulation factors: spoIIQ spoVAC sspD sspM cgeA
cotZ sspN spoIIIAC spsI usd cotW yraE cgeC yraG
sspG cotV cgeB spoVFA sspJ ytaA cotB cotT cotC
bofC spsA cotS spoVG spoVAD yraD
glcP
Glucose/Mannose
Glycine/betaine
H+
opuCD
opuAC
Similar to transporters: ydhL yfhA yfmC yufO ydjK
yfhI yfnA yqjV yteP yurY yvdB
licC
ptsG

Page 4

BMC Microbiology 2009, 9:176http://www.biomedcentral.com/1471-2180/9/176
Page 4 of 14
(page number not for citation purposes)
structural proteins involved in spore formation. This
response is to be expected, in the light of the repressive
effect that glucose exerts on the sporulation process [14].
Topological analysis of a sub-network of Bacillus subtilis,
responding to glucose
Data from DBTBS [15] was used to generate the known
regulatory network of B. subtilis. The resulting network is
composed of 1453 nodes and 2337 edges, showing an
average clustering coefficient of 0.47. The degree distribu-
tion follows a power law, P(k) ~k-2.0043. These results are
characteristic of a scale-free network, and strongly suggest
the existence of a modular hierarchical organization.
These properties are common to other previously
described biological networks [1].
As described in the methods section, we selected a set of
504 genes shown to respond under the test conditions,
with a significant level of expression. From this set, those
genes not having a regulatory relation were eliminated
from the regulatory network. The resulting network will
be called the sub-network that responds to the presence of
glucose. In this sub-network, 264 genes have known regu-
latory information, including sigma and transcription fac-
tors; TFs. As the sigma factor A is predominantly
connected to almost every gene in the network, we
decided to remove it from the subnetwork. Therefore, the
final subnetwork used for further analysis includes 186
genes, 68 (TF) and 10 sigma factors.
By applying a hierarchical agglomerative clustering algo-
rithm to the sub-network, it was possible to group the
transcription factors and the genes responding to glucose
into topological modules (figure 2). The clustering algo-
rithm grouped the genes in a giant component, composed
of 6 modules which include members with more that one
operon and two mini-modules (basically complex and
simple regulons [16]). Additionally, disconnected from
the giant component we discovered 16 mini-modules and
3 modules.
Carbon metabolism and stress response (M1)
The first module identified using this method, includes 39
genes distributed within two sub-modules: The first sub-
module, includes 8 genes, belonging to two of the func-
tional classes described in the SubtiList database [17]. In
this submodule, 3 clustered genes related to anaerobic
conditions are induced in the microarray data, table 1.
This behavior appears to be consistent with observations
from previous reports, indicating that the regulation of
this gene regulatory cascade by an unknown sensor via
ResDE, Fnr, and ArfM manifests differing growth, espe-
cially when both glucose and pyruvate are provided, or
when glucose and mixtures of amino acids are present
[18]. The other five genes included in this sub-module are
encoding proteins, related to the heat shock response.
These genes are repressed by the protein HrcA, which is
auto-regulated and whose transcription can also be acti-
vated by ArfM. The microarray data indicate that the gene
arfM is induced by glucose, suggesting that the protein
ArfM activates transcription of hrcA and the encoded pro-
tein, whereas it represses dnaK, grpE, groEL and groES. The
second sub-module includes 31 genes with a detected
transcript level, 29 of which were repressed and 2 of which
were induced. Out of this set, 30 of these are regulated by
CcpA (catabolic control protein). These genes encode
functions associated with the transport and degradation
of alternative carbon sources.
Endospore formation and Spo0A (M2)
Our results indicate a cluster, divided into two sub-mod-
ules. The endospore formation sub-module grouped five
genes participating in the formation of endospore, four of
which were repressed (citG, dppE, spoVG, yxnB) and one
was induced (hag). This data is in accordance with a pre-
vious report where AbrB was identified as repressing the
aforementioned genes in a regulatory process known as
catabolic repression of sporulation [14]. The second sub-
module was composed of seven genes encoding for sporu-
lation functions; six of which were induced (Table 1) with
their transcription depending on SpoA and the sigma fac-
tor D (Sigma D), and one of which (Table 1) was
repressed with its transcription depending on Sigma D.
Spore and prespore formation (M3)
In this module, we found 39 genes responding to the pres-
ence of glucose; 28 of these were repressed and the others
were induced (Table 1). This cluster was subdivided into
2 sub-modules. The first one shows genes whose products
are associated with pre-spore formation, germination and
cell wall components [19-21]. The second sub-module is
composed of 19 genes acting in the formation of spores,
mainly regulated by Sigma B. With the exception of the
induced genes (csbX, yjgB, gcaD, ypuB yotK and spoIIQ), all
the other genes in these sub-modules were repressed
when under the LB+G condition, a result consistent with
the fact that genes involved with sporulation processes are
repressed in the presence of non-restrictive nutritional
conditions [21].
Hexuronte metabolisms (M4)
This module has genes involved in hexuronate metabo-
lism [22], organized into two independent operons. Both
operons are known to be negatively regulated by CcpA,
whereas the
uxaC-yjmBCD-uxuA-yjmF-exuTR-uxaBA
operon is additionally, negatively regulated by ExuR [22].
The microarray data indicated that the genes were
repressed, suggesting that CcpA represses them, when glu-
cose is present.

Page 5

BMC Microbiology 2009, 9:176 http://www.biomedcentral.com/1471-2180/9/176
Page 5 of 14
(page number not for citation purposes)
Nitrogen metabolism and Spore coat formation (M5)
This module includes 39 genes and was divided into two
sub-modules, each having related functions. The first set
of four genes encode proteins that participate in nitrogen
metabolism, co-regulated by the nitrogen utilization pro-
tein TnrA [23]. The second sub-module comprises 35
genes involved in the spore coat formation. A unique
property of this sub-module is that all genes are regulated
by the protein Sigma K, encoded by the genes spoIIIC and
spoIVCB [24,25]. As all the genes belonging to this sub-
module were shown to be repressed, this indicates that the
sporulation regulatory program is governed by a hierar-
chical cascade, consisting of the transcription factors:
Sigma E, Sigma K, GerE, GerR, and SpoIIID. This observed
response is in accordance with previous reports [21]
Clustering results from the B. subtilis sub-network that responds to glucose
Figure 2
Clustering results from the B. subtilis sub-network that responds to glucose. The image shows the modular struc-
tures obtained using the clustering method. The figure is composed of a giant component with six modules (M1-6) and two
mini-modules (MM1-2). Disconnected from the giant component, we have 16 mini-modules (MM3-18) and two modules (M8-
9). The column on the right hand side shows the transcriptional response for each gene, according to the microarray data. Red
color represents an increase in transcript level, green color represents a decrease in transcript level and grey color indicates
no significant change in transcript level.
MM3
MM4
M7
M4
MM2
M1
M2
MM1
M3
M5
M6
MM5-18
M8
M9

Page 6

BMC Microbiology 2009, 9:176http://www.biomedcentral.com/1471-2180/9/176
Page 6 of 14
(page number not for citation purposes)
Table 1: Modules and sub-modules found in the B. subtilis glucose-responding regulatory network.
Module Physiological functionGenes
M1Heat shock response
acoR(↓), alsS(↑), arfM, alsR, ydiH, cydC(↑), dnaK(↓), grpE(↓), lctP(↑), hrcA, resD, groEL(↓), groES(↓)
Carbon catabolism
glpK(↓), glpP, ahrC, rocR, iolR, araB(↓), araN(↓), acuB(↓), galK(↓), msmX(↓), pta(↓), bglH(↓), bglP(↓),
yxiE(↓), licA(↓), licB(↓), licC(↓), licH(↓), treA(↓), treP(↓), ccpA, iolC(↓), iolF(↓), iolH(↓), iolI(↓), ccpB,
xylR, xylA(↑), araR, treR, licR, licT, levD(↓), levE(↓), levR, sigL, rocA(↑), ydjK(↓), rbsA(↓), rbsB(↓),
rbsC(↓), rbsD(↓), rbsK(↓), rbsR(↓)
M2Endospore formation
citG(↓), codY, dppE(↓), hag(↑), abrB, sigH, spoVG(↓), yxnB(↓)
Sporulation
dltA(↑), dltB(↑), dltD(↑), dltE(↑), mcpB(↑), yjcP(↑), yvyC(↓), sigD
MM1Sporulation
sigX, spo0A
M3Prespore formation
comA, yvrH, wprA(↓), degQ(↓), wapA(↑), sacA(↓), sacP(↓), degU, sacT, sacY, tenA, yveB(↓), sigG,
gerKA(↓), ybxH(↓), sspD(↑), spoVAD(↓), spoVAC(↑), sspJ(↓), sspM(↑), adhB(↓), yraG(↓), yraE(↓),
yraD(↓), yndE(↑), ylaJ(↓), sspN(↓)
Spore formation
ctsR, bofC(↓), csbX(↑), sigF, spoVT, sigB, clpP(↓), rsbW(↓), ydaE(↓), ydhK(↓), yjgB(↑), gcaD(↑),
yycD(↓), ysnF(↓), ypuB(↑), yoxB(↓), yotK(↑), yqhQ(↓), spoIIQ(↑), yfhD(↓), yfhE(↓), yhcM(↓), yqzG(↓)
MM2Glycerophospholipid metabolism
glpQ(↓), glpT(↓), phoP
M4 Hexuronate metabolism
exuR, mmgA(↓), yjmC(↓)
M5Nitrogen metabolism
glnR, glnA(↑), glnP(↑), kipR, tnrA, ykoL(↓), ykzB(↑)
Spore coat formation
gerE, spoIIIC, spoIVCB, cotB(↓), cotC(↓), cotV(↓), cotW(↓), cotT(↓), cgeA(↓), cgeB(↓), cotZ(↓), sspG(↓),
cgeC(↓), yurS(↓), yoaN(↑), yjcB(↓), spoVFA(↓), yisZ(↓), ykvP(↓), ykvQ(↓), ylbD(↓), ylbE(↓), cotS(↓),
ywrJ(↓), ytxO(↓), ytcC(↓), ytaA(↓), yqfQ(↓), yodH(↓), yngK(↓), ymaG(↓), spsA(↓), spsI(↓), ycsF(↓),
spoIIID, ylbO, yhcO(↓), yhcP(↓), ypqA(↓), ysnD(↓)
M6 SOS response
lexA, ybaK(↓), aprX(↓), yozM(↓), yozL(↑)
Prospore formation
spoIIIAC(↓), yqfZ(↓), sigE, usd(↓), mbl(↑), yheD(↓), yjcA(↓), yncD(↓), yngE(↓), yngG(↓), ywdL(↓)
MM3Glycolysis
cggR, eno(↑), pgk(↑), pgm(↑)
MM4Nitrogen assimilation
fnr, narG(↓), narH(↑)
M7Competence
comK, comGB(↑), cspB(↓), yhjC(↓), yhcD(↑), ssb(↑), rpsF(↑), rpsR(↑)
MM5Peroxide stress
ahpC(↓), ahpF(↓), perR
MM6PTS-glucose system
glcT, ptsG(↑), ptsI(↑)
MM7Amine and polyamine degradation
bltD(↓), bltR, mta
MM8Extracytoplasmic
sigY, yxlC(↓), yxlE(↓)
MM9Aspartate catabolism
ansB(↓), ansR
MM10N/A
lmrA, yxaH(↓)
MM11N/A
fur, ydhU(↓)
MM12N/A
rok, yydH(↓)

Page 7

BMC Microbiology 2009, 9:176http://www.biomedcentral.com/1471-2180/9/176
Page 7 of 14
(page number not for citation purposes)
SOS and prospore formation (M6)
Is constituted by 14 genes (Table 1) and the clustering
method divided the module into two functionally defined
sub-modules. The SOS sub-module possesses three genes
regulated by LexA, which participate in DNA repair [26].
We found a second subunit, comprising 10 genes, regu-
lated by Sigma E, which is the earliest-acting factor, spe-
cific to the mother-cell line of gene expression on the
cascade forming the prospore [21]. As is evident in Table
1, 12 of the 14 genes participating in the cluster appear to
be repressed.
As previously mentioned there are two mini-modules
(MM) embedded within the giant component. The first
one (MM1, Table 1), possesses the genes which encode for
Sigma X and Spo0A TFs and which are involved in the
sporulation process. The second mini-module (MM2
Table 1) has genes relating to glycerophospholipid metab-
olism that are entirely regulated by PhoP.
We found several mini-modules and two modules, sepa-
rated from the giant component. The existence of these
topological structures is likely to be a consequence of the
fact that knowledge of the network is incomplete, the
absence of genes or because certain TFs are not included
in the sub-network or because of the existence of other
regulatory structures, such as antiterminators, terminators
and regulatory RNAs which are not considered in the net-
work construction. For these reasons, some very well stud-
ied functions (see Table 1) such as glycolysis (MM3),
respiratory function control by FNR (MM4), peroxide
stress (MM5), the PTS system dependent on glucose
(MM7), competence regulated by ComK (M7), the cystein
module (M8) and a topological structure dependent on
the sigma factor W (M9) were excluded from the giant
component.
Comparison of the glucose responsive networks found in E.
coli and B. subtilis
The structure of complex transcriptional regulatory net-
works has been studied extensively in certain model
organisms. However, understanding is still limited con-
cerning the evolutionary dynamics of these networks in
different organisms, which would surely reveal important
principles of adaptive regulatory changes. The problem is
more challenging when the aim is to carry out a detailed
comparison of the regulatory networks of phylogeneti-
cally distant organisms. Previous works have studied the
regulatory networks of E. coli and B. subtilis and assessed
the conservation in their TFs and regulated genes, in the
context of a broad array of sequenced genomes [27,28].
Both works make it clear that the set of regulatory genes -
even global transcription factors - vary considerably from
one group of organisms to another. This overview has to
be significantly adjusted when closely related species are
compared [29,30], where there is greater conservation
between the TFs and the regulated genes. In this work, we
compared the regulatory networks derived from signifi-
cant transcript levels of E. coli and B. subtilis observed in a
microarray experiment, assessing response to the presence
of glucose. For this purpose, we took the E. coli sub-net-
work previously published by our group [13] along with
the one generated in this work. The E. coli sub-network
was constructed from 380 genes and 47 TFs, listed in the
RegulonDB database [31]. The comparison was carried
out at 2 levels: the first one considered the conservation of
orthologous genes in both sub-networks and the second
took into account the modular structures of B. subtilis as
MM13Sorbitol catabolism
gutB(↓), gutR
MM14Purine metabolism
purQ(↑), purr
MM15N/A
birA, ytbQ(↑)
MM16N/A
yufM, ywkB(↑)
MM17N/A
appC(↓), hpr
MM18Lactose catabolism
lacA(↓), lacR
M8 Extracytoplasmic
sigW, yceC(↓), yceD(↑), yceF(↓), ydjH(↓)
M9Cysteine biosynthesis
cysK(↓), ytmI(↓), ytmJ(↓), ytmK(↑), ytmL(↓), ytnI(↓), ytnM(↓), yrzC, ytlI
We found 9 modules and 18 mini-modules (MM), the latter defined as a module comprising only genes in the same operon or a simple regulon, with
just a few members. Up-regulated genes are indicated by an up-arrow (↑), whereas a down-arrow (↓) indicates a down-regulated gene; genes
without an arrow were not significantly detected in microarray. Physiological functions are discussed in the text. A module tagged 'N/A' means that
currently not enough information exists to make a functional assignment.
Table 1: Modules and sub-modules found in the B. subtilis glucose-responding regulatory network. (Continued)

Page 8

BMC Microbiology 2009, 9:176http://www.biomedcentral.com/1471-2180/9/176
Page 8 of 14
(page number not for citation purposes)
described in this report as well as that previously pub-
lished by Gutierrez-Rios et al [13], describing E. coli.
Identification and analysis of the orthologous genes in both E. coli
and B. subtilis which respond to glucose
We performed a computational search for the bidirec-
tional best hits (BBHs) found in all open reading frames
for the genomes of E. coli and B. subtilis, as described in the
methods section. As a result, 1199 orthologous genes were
shown to be present in these two organisms. From this set,
134 genes manifested significant differences in terms of
repression/activation when B. subtilis was grown in the
presence or absence of glucose. Out of these, 52 genes
were orthologous and responsive to the presence of glu-
cose in the case of both organisms. Figure 3, shows that 47
genes exhibited the same expression pattern in the case of
both organisms and five differed. These five genes are pta
(phosphoacetyltransferase), gapA (glyceraldehide-3-phos-
phate dehydrogenase), prsA (peptidyl-prolyl-cis-trans-iso-
merase), sdhA (succinate deshydrogenase and mutS
(methyl-directed mismatch repair). The pta gene was
found to be repressed in the B. subtilis microarray data, a
result which was inconsistent with a previous report by
Presecan-Siedel et al [32], which demonstrated that pta, as
is the case with other genes involved in acetate production
are induced in the presence of glucose. An induction was
Comparison of the significantly induced orrepressed orthologous genes in E. coli and B. subtilis
Figure 3
Comparison of the significantly induced orrepressed orthologous genes in E. coli and B. subtilis. The figure illus-
trates a cluster of orthologous genes, comparing B subtilis (column 1) and E. coli (column 2) transcribed levels, as they respond
to glucose. Induced genes are represented in red and repressed genes are represented in green. Gene names and functional
class are indicated on the right hand side.
Bsu_log(LBG_LB)
Eco_log(LBG_LB)
cstA, cstA, Adaptation to atypical conditions
cysK, cysK, Metabolism of amino acids and related molecules
ansB, aspA, Metabolism of amino acids and related molecules
levD, manX, Transport/binding proteins and lipoproteins
levE, manX, Transport/binding proteins and lipoproteins
rbsB, rbsB, Transport/binding proteins and lipoproteins
ahpC, ahpC, Detoxification
drm, deoB, Metabolism of nucleotides and nucleic acids
odhB, sucB, TCA cycle
ftsH, hflB, Cell division
galK, galK, Specific pathways
sodA, sodA, Detoxification
dnaK, dnaK, Protein folding
fbaA, gatY, Main glycolytic pathways
grpE, grpE, Adaptation to atypical conditions
yurU, sufB, Similar to unknown proteins from other organisms
gapA, gapA, Main glycolytic pathways
sdhA, frdA, TCA cycle
accC, accC, Metabolism of lipids
rpoA, rpoA, Elongation
pdhC, aceF, Main glycolytic pathways
mutS, mutS, DNA restriction/modification and repair
prsA, ppiC, Protein secretion
pta, pta, Specific pathways
infC, infC, Initiation
rplC, rplC, Ribosomal proteins
ptsG, ptsG, Transport/binding proteins and lipoproteins
rpmI, rpmI, Ribosomal proteins
rplM, rplM, Ribosomal proteins
rpsS, rpsS, Ribosomal proteins
rpsD, rpsD, Ribosomal proteins
rplD, rplD, Ribosomal proteins
rpsJ, rpsJ, Ribosomal proteins
guaA, guaA, Metabolism of nucleotides and nucleic acids
rpmA, rpmA, Ribosomal proteins
rpsE, rpsE, Ribosomal proteins
rplS, rplS, Ribosomal proteins
infA, infA, Initiation
rpsQ, rpsQ, Ribosomal proteins
rplR, rplR, Ribosomal proteins
tsf, tsf, Elongation
adk, adk, Metabolism of nucleotides and nucleic acids
rplV, rplV, Ribosomal proteins
rplP, rplP, Ribosomal proteins
rplK, rplK, Ribosomal proteins
rpsR, rpsR, Ribosomal proteins
rpoB, rpoB, Elongation
rpsB, rpsB, Ribosomal proteins
rpsF, rpsF, Ribosomal proteins
rpsC, rpsC, Ribosomal proteins
rpsG, rpsG, Ribosomal proteins
rpsI, rpsI, Ribosomal proteins

Page 9

BMC Microbiology 2009, 9:176http://www.biomedcentral.com/1471-2180/9/176
Page 9 of 14
(page number not for citation purposes)
also observed for the pta gene of E. coli [33]. The gapA gene
was induced in B. subtilis and repressed in E. coli. The
observation was consistent with other reports where the
gapA of B. subtilis and other bacillus was described as
being induced in the presence of glucose, as a result of its
participation in the glycolitic pathway [33]. The opposite
response for gapA in E. coli may be a consequence of its
participation in gluconegenesis [13]. Very little is known
about the regulation of mutS in E. coli and B. subtilis. This
gene has been described as a DNA repair protein in the
context of both bacteria [34]. Something similar happens
to psrA in B subtilis, also known as ppiC in E. coli; where
both enzymes function as molecular chaperones. It has
been reported that prsA is essential for the stability of
secreted proteins at certain stages, following translocation
across the membrane [35]. Finally, the results observed
for the genes sdhA (succinate deshydrogenase en B. subti-
lis) and frdA (fumarate reductase in E. coli) are quite inter-
esting. Apparently, the functions of these two enzymes
seem to be different; the succinate dehydrogenases of aer-
obic bacteria catalyze the oxidation of succinate by respi-
ratory quinones (succinate:quinone reductase), and the
quinols are reoxidized by O2 (succinate oxidase) [36]. In
the case of B. subtilis; for some time it was thought that this
enzyme has only this function, but in a recent report, the
authors demonstrated that resting cells are able to catalyze
fumarate reduction, with glucose or glycerol. The enzy-
matic system for fumarate reduction in B. subtilis was
shown to be an electron transport chain, comprising a
NADH dehydrogenase, menaquinone and succinate
dehydrogenase [36]. Therefore, this enzyme is able to
modify its function depending on the growth condition
and energetic state of the cell.
Figure 3 presents a set of genes shared by both bacteria
that in addition to being orthologous display similar
expression patters. Twenty of these are ribosomal genes,
induced by the presence of glucose. Another seven genes
are involved in the synthesis of macromolecules and a fur-
ther 14 belong to cellular anabolism and catabolism of
carbohydrates as well as central intermediary metabolism.
Five of these are related to protective functions, four are
classified as transporters and one gene encodes a protein,
related to cell division.
The comparison between orthologous genes, differen-
tially expressed in LB+G vs LB reveals a very small set of
genes, common to both organisms. This correlates well
with other works [27,28] that attribute this result to the
great phylogenetic distance between these organisms. We
also think this is a consequence of the small number of
genes in the microarray data, shown to be differentially
expressed. It is important to note that the categories con-
served between these bacteria are confined to global
house keeping genes, with functions associated with tran-
scription, translation, and replication. It is also interesting
to note that enzymes relating to central metabolism and
energy production are also consereved and display the
same behavior, whether active or inactive. The gene sdhA
provides us with an interesting example of how ortholo-
gous genes can adapt their products to become enzymes
with multiple functions, depending on their context. It
would be interesting to analyze whether the regulatory
response of this set of orthologous genes in other organ-
isms preserved their original functions or adapted to alter-
native metabolic pathways. Hernández-Montes et al made
an interesting contribution to this subject in terms of
orthologous amino acid biosynthetic networks, where
they identified alternative branches and routes, reflecting
the adoption of specific amino acid biosynthetic strategies
by taxa, relating their findings to differences in the life-
styles of each organism [37].
Considering the 52 orthologous genes previously
described, we were also interested to discover how many
of the TFs regulating these were also orthologous. In Addi-
tional File 2 (see Table 2aSM) we present the orthologous
expressed genes for both sub-networks, which manifest a
regulatory interaction. The sub-network is composed of
43 TFs in E. coli and 44 in B. subtilis (including sigma fac-
tors). Out of these, 10 E. coli regulatory genes (araC, crp,
cytR, dcuR, mlc, dnaA, fur, glpR, lexA, nagC, narL) have an
orthologous regulatory counterpart in B. subtilis and nine
B. subtilis regulatory genes (ccpA, fnr, glnR, glpP, kipR, sigL,
xylR, yrzC), yufM) have one in E. coli (see Additional File
2: Table 3SM). As both E. coli and B. subtilis were exposed
to rich media in either the presence or absence of glucose,
the comparison between CcpA and CRP is especially rele-
vant. CcpA belongs to the LacI/GalR family of transcrip-
tional repressors [38] and CRP to the AraC/XylS family of
transcription factors [39]. Both TFs fulfil the function of
increasing and decreasing the activity of genes, subject to
catabolic repression. The mechanism for sensing the pres-
ence or absence of glucose in both bacteria depends on
the PTS system. In B. subtilis, PTS mediates phosphoryla-
tion of the regulatory protein HprK that in the presence of
fructose 1-6 biphospate promotes the binding of CcpA to
CRE sites [8]. In E. coli, the phosphorylation events end
with the production of cyclic AMP molecules that directly
activate the catabolic repression protein CRP that usually
induces their regulated genes. Our results reveal that both
proteins, in spite of not being orthologous and belonging
to different protein families, coordinate the expression of
several orthologous genes (see Additional File 2: Tables
2aSM and 2bSM). Four genes responded to glucose in
both organisms and 14 in B. subtilis. This result may be
explained, taking into account the fact that many interac-
tions relating to every gene in the network have still not
been discovered and it is also probable that the degree of
sensitivity in the microarray analysis was not sufficient to
detect every significant signal.

Page 10

BMC Microbiology 2009, 9:176http://www.biomedcentral.com/1471-2180/9/176
Page 10 of 14
(page number not for citation purposes)
Our analysis revealed other expressed genes regulated by
non-orthologous TFs that manifest similar functions.
These consist of the cases of FruR (E. coli) and CcgR (B.
subtilis), controlling the central intermediary metabolism,
as well as RbsR (E. coli) and AbrB (B. subtilis), repressing
genes in the presence of ribose. For instance, the AbrB,
evolved to respond to additional stimulus, extending the
number of elements of the regulon to sporulating func-
tions. Finally, our results indicated that the SOS regulon
control on the part of the orthologous TF LexA was not
conserved [26]. The examples described previously are
consistent with other findings indicating that the conser-
vation between regulatory networks of distant organisms
is in fact limited., Arguments treating this subject are
directed towards the possibility of genetic duplication
[40] and the adaptation of each organism to particular
media [27,28], also promoting the concept that proteins
evolved and took on new functions.
Comparison of topological units of the sub-networks between E. coli
and B. subtilis
There is convincing evidence to suggest that gene duplica-
tion is a major force explaining the growth of TRNs
[27,28,40]. It is possible that this modifying process
affects the connectivity distribution of these networks, as
has been observed in other biological networks [27]. In
view of these findings, we compared the modular struc-
tures found in E. coli and B. subtilis, in order to evaluate the
conservation of topological structures.
A comparison was carried out, considering the modular
structure of the sub-network of E. coli in the presence of
glucose [13] and the modular structure for B. subtilis, gen-
erated during this study. Figure 4 presents orthologous
genes that were organized into modular structures. At this
level, we could see that most of the genes clustering in
modules in both sub-networks, related to carbon metabo-
lism. Those genes encoding for proteins of the PTS system
were outstanding (levDE, ptsG), the degradative enzyme
galK and the gene rbsB encoding as a transporter. All of the
genes previously described except ptsG belong to the mod-
ules classified as Carbon Modules in both sub-networks.
In the case of E. coli, genes in this module were clustered
because they were regulated by CRP and in the case of B.
subtilis by the relationship of the genes to the regulatory
protein CcpA. The disconnection of ptsG from the carbon
module in B. subtilis can be explained by the absence of
regulation by CcpA (Figure 4, Table 1).
In both arrays, we found repression of genes encoding
chaperons. Two of these, (dnaK and grpE) in B. subtilis are
orthologous to genes in E. coli. In B. subtilis, the two
orthologous and other chaperons were grouped into a
sub-module with two major functions: the first one
related to respiration and the second one involved in heat
shock response. The regulatory protein ArfM connects all
the genes in the network and HrcA controls genes related
to both conditions and HrcA also controls the genes
responding to heat shock. In the case of E. coli the genes
are clearly organized into a module that includes only the
heat shock genes, the organization of the module depends
on the sigma factor RpoH.
We also found that respiratory functions were clustered
into two groups, in the case of B. subtilis. The first one
embedded in the sub-module concentrates anaerobic res-
piration and some heat shock proteins. The second set of
respiratory clustered genes are also related to anaerobic
functions, but in this instance they are regulated by the
transcription factor FNR which is orthologous to CRP in
E. coli. In contrast, respiratory functions in E. coli are clus-
tered into one module containing proteins that control
aerobic and anaerobic growth. One of the TFs in E. coli is
FNR, for which there is no orthologous gene in B. subtilis.
It is interesting to note, that despite not being ortholo-
gous, FNR regulates the expression of the orthologous
operon narGHJI which encodes for all the subunits of the
nitrate reductase enzyme [41,42], narK-fnr, where narK
encodes a protein with nitrite extrusion activity [41,43]
and the regulatory gene fnr. The microarray data also
revealed ten genes in B. subtilis, known to participate in
respiratory functions, where no regulatory interactions
have been described (membrane bioenergetics electron
transport chain and ATP synthase, see Additional File 1:
Table 1SM). We also observed a pair of module clustering
genes that control stress by peroxides; for B. subtilis, the
regulatory protein PerR, whereas for E. coli, it is OxyR. The
module shares an orthologous gene ahpC that was
repressed in both micro arrays.
Finally, the topological arrangement, which resulted from
the clustering method applied, revealed two very impor-
tant differences. The first one was the case of modules
related to sporulation. These were not expected to be
found in E. coli, but occupy more than 50% of the regula-
tory sub-network in B subtilis. This finding is also not a
surprise considering that sporulation is the best-studied
mechanism in this organism. It is also important to men-
tion that 74% of the genes that cluster in the sporulation
modules are repressed and the genes that appeared
induced in the cluster are mainly dedicated to functions
such as cell wall formation, motility, ribosomal proteins,
DNA replication and others not assigned to a specific
class. This finding reflects the physiological importance of
sporulation in this organism, which is one of the most
interesting features of certain soil bacteria. It is well
known that in response to nutrient limitation, B. subtilis
cells undergo a series of morphological and genetic
changes that culminate with the formation of endospores.
Conversely, the presence of sufficient metabolizable car-
bon sources, e. g., glucose inhibits the synthesis of extra-
cellular and catabolic enzymes, TCA cycle enzymes and

Page 11

BMC Microbiology 2009, 9:176http://www.biomedcentral.com/1471-2180/9/176
Page 11 of 14
(page number not for citation purposes)
the initiation of sporulation. This is the second difference
concerning the topological arrangement of our studied
organisms and a characteristic not shared by E. coli, which
has a different life style. It would be interesting to ascer-
tain whether in a different growth condition, the topolog-
ical analysis of alternative sub-networks would manifest
the same result.
Conclusion
The analysis of transcriptome data collected under condi-
tions of both glucose sufficiency and deficiency in a com-
plex medium enabled us to identify functions involved in
the adaptation of B. subtilis to these growth conditions.
The known repressive effect of glucose on alternative car-
bon source import and metabolism were clearly demon-
strated. We also were able to observe an inductive effect
Conserved glucose responding modules between B. subtilis and E. coli
Figure 4
Conserved glucose responding modules between B. subtilis and E. coli. Whereas there is extensive rewiring in the
regulatory network, some modules have conserved their physiological functions and expression profile, showing the high plas-
ticity of regulatory networks in terms of evolution. Dashed thin lines show orthology relations, whereas blue dash-dot lines
bound modules. Green ellipses indicate repressed genes; red ones show activated genes and grey ones indicate genes, which
are not significantly expressed. E. coli modules IDs are taken from Gutierrez-Rios et al. [13]. Regarding the aspartate catabolism
module, it has been suggested that L-aspartase encoded by ansB is an strictly catabolic enzime (catalyzing the reaction aspartate
→ fumarate + NH4+), thus providing carbon skeletons to Krebs cycle.
B. subti
E. coli
Cysteine biosynt
(M9)
Carbon catabolism and
heat shock response
(M1)
PTS−glucose
system (MM6)
Aspartate
catabolism (MM9)
Peroxide stress
(MM5)
Cysteine biosynthesis
(IM)
Carbon starvation
response mediated
by RpoH
(M1)
Carbon catabolism
(M8)
Oxidative stress
(M3)

Page 12

BMC Microbiology 2009, 9:176http://www.biomedcentral.com/1471-2180/9/176
Page 12 of 14
(page number not for citation purposes)
on the glycolitic pathway and the repressive effect on the
genes related to the sporulation cascade.
A topological analysis revealed modules that include gene
encoding functions, with similar physiological roles.
In a previous work, we performed a similar study under
the same conditions on the Gram negative bacteria E. coli
[13]. Analysis of orthology and topological structures,
exposed coincidences in the genes that can be considered
as the basic machinery of these organisms, such as replica-
tion, transcription, translation, central intermediary
metabolism and respiratory functions. An outstanding
discovery consisted in the fact that both bacteria manifest
a similar response concerning the gene encoding chaper-
ones, when responding to heat shock, even when these are
controlled by different transcription factors (the heat
shock sigma factor -Sigma H- in E. coli and the regulatory
protein ArfM in B. subtilis). Also noteworthy was the iden-
tification of modules in E. coli and B. subtilis, including
genes related to alternate carbon source utilization, which
respond to the presence of glucose and are regulated by
CRP and CcpA respectively, employing different mecha-
nisms. Other examples were described in the results and
discussion section, showing that for similar transcrip-
tional responses, different regulatory strategies were
implemented in the case of each organism. The consider-
able differences between the mechanism controlling gene
expression and the small set of orthologous genes found
in the conditions tested, are a consequence of the large
phylogentic distance between these bacteria.
These analyzes also revealed how incomplete our knowl-
edge still is, concerning gene regulation in B. subtilis. We
are aware that processes such as catabolic repression,
nitrogen assimilation and sporulation have been exten-
sively analyzed, whereas other functions shared with E.
coli, such as certain genes of the main glycolytic pathways,
TCA cycle, and respiratory function, are not well under-
stood. Integrative analysis of transcriptome and transcrip-
tional regulatory data as undertaken here, as well as the
comparison between organisms should provide a frame-
work for the future generation of models. These will help
explain the cell's capacity to respond to a changing envi-
ronment and increase understanding of the evolutionary
forces, which enable life forms to harmonize their regula-
tory processes in order to improve their adaptation.
Methods
Data analysis and identification of differential transcribed
genes
Transcriptome data was obtained from previously
described experiments performed with B. subtilis strain
ST100 broth, containing 50 mM potassium phosphate,
pH 7.4, and 0.2 mM L-cysteine with (LB+G) or without
(LB) 0.4% glucose. The average expression data from three
repeated experiments was collected from web http://biol
ogy.ucsd.edu/~msaier/regulation2/ of the B. subtilis anti-
sense. DNA arrays used in this work were custom designed
and manufactured by Affymetrix (Santa Clara, CA) [8].
As we only had access to the average of the crude expres-
sion data, we applied the rank product method [44]. This
method is based on the calculation of rank products, from
which significance thresholds can be extracted, in order to
distinguish significantly regulated genes. In the case of our
data, we chose a RP-value of 3.5 × 10-2 as a cutoff point,
and in this way we distinguished the most significant 150
up-regulated and 150 down-regulated genes. However, as
we also were interested in the differential expression
under both conditions, we picked up those genes exhibit-
ing a > 3-fold change between LB and LB+G. Finally, we
took the logical union of such populations. Using this
method a set of 503 genes were taken into account for
subsequent analysis.
As in our previous work, concerning differentially
expressed genes of E. coli [13], the terms "induced" and
"repressed" were used in this work to indicate increased or
decreased transcript levels, respectively. These terms do
not imply a particular mechanism for gene regulation.
Extraction of condition-specific sub-networks
For each microarray condition LB+G/LB, we reconstructed
a condition specific sub-network as follows. From the
transcriptional regulatory network of B. subtilis, we
extracted the significant genes identified in the microarray
condition, the TFs regulating their expression, and the
transcriptional interactions between TFs and their regu-
lated genes. In these sub-networks, nodes represent genes
and edges represent the transcriptional interactions.
Known regulatory sites and transcriptional unit organiza-
tion were obtained from DBTBS [45].
Identification of condition-specific modules
We identified the LB+G/LB condition-specific modules
applying to the condition specific sub-network, the meth-
odology described in Resendis-Antonio et al [46] and
Gutierrez-Rios et al [13]. Specifically, we clustered the
genes based on their shortest distance within the network.
Afterwards, we annotated each gene with its correspond-
ing microarray expression level. The dendogram gener-
ated by the clustering algorithm was decomposed into
modules and sub-modules. Hierarchical clustering algo-
rithms produce a dendogram by iteratively joined pairs of
data, with the closest correlation levels. We analyzed the
distribution of correlation values, observing that ~90%
(228 from 254) of the nodes in the dendogram have a cor-
relation value greater than 80%. Hence, in order to isolate
modules, we pruned every node with a correlation of less
than 80% from the dendogram. In addition, to identify-
ing sub-modules, we then pruned the dendogram once

Page 13

BMC Microbiology 2009, 9:176 http://www.biomedcentral.com/1471-2180/9/176
Page 13 of 14
(page number not for citation purposes)
again; this time removing all the nodes with a correlation
of less than 90%.
Detection of orthologous genes
A simple method for predicting the orthologous proteins
present in two organisms is to search for a pair of
sequences, Xa in organism Ga and Xb in organism Gb,
such that a search of the proteome of Gb with Xa indicates
Xb to be the best hit. We made this comparison using the
Blastp program [47,48] with the E. coli and the B subtilis
genome as input. If the protein in each genome has the
highest E-value and an upper threshold of 10-5 in both
genomes, we considered them to be orthologous. From
this set we selected the significant expressed genes, pub-
lished in our previous work run under the same condi-
tions of LB growth, in the presence or absence of glucose
[13].
Clustering of microarray data of orthologous genes
We applied a hierarchical centroid linkage clustering algo-
rithm [49,50] to the log ratios of the differences between
the orthologous genes of E. coli and B. subtilis, with the
correlation un-centered as a similarity measure... The clus-
tering results were visualized using the Treeview program
[51].
List of abbreviations
CRE, SM, LB, LB+G, TF, PTS, B. subtilis, E. coli.
Authors' contributions
CDV contributed with construction of the regulatory net-
work, microarray and module analysis. JAF-G contributed
with the discussion for the selection of microarray data,
performed the construction of topological modules and
comparison of modular subunits. GG contributed with
the analysis and interpretation of microarray data for the
physiological sections. RMG-R contributed to the analysis
and interpretation of the microarray data in terms of the
regulatory network, elaboration of programs for data
management as well as a discussion concerning the selec-
tion and processing of microarray. All authors wrote, read
and approved the final manuscript.
Additional material
Acknowledgements
We thank Nancy Mena for technical support. I am in indebted to Antonio
Loza for discussion and microarray selection. I also want to thank Enrique
Merino for revising the final version of this manuscript. This work was sup-
ported by grant IN215808 from PAPIIT-UNAM and CONACyT-58840 to
R.M.G.-R. and PAPIIT/UNAM IN214709 to G.G.
References
1.Barabasi AL, Oltvai ZN: Network biology: understanding the
cell's functional organization. Nat Rev Genet 2004, 5:101-113.
2.Ravasz E, Somera AL, Mongru DA, Oltvai ZN, Barabasi AL: Hierar-
chical organization of modularity in metabolic networks. Sci-
ence 2002, 297:1551-1555.
3.Goelzer A, Bekkal BF, Martin-Verstraete I, Noirot P, Bessieres P,
Aymerich S, et al.: Reconstruction and analysis of the genetic
and metabolic regulatory networks of the central metabo-
lism of Bacillus subtilis. BMC Syst Biol 2008, 2:20.
4.Moszer I: The complete genome of Bacillus subtilis: from
sequence annotation to data management and analysis. FEBS
Lett 1998, 430:28-36.
5.Sonoshein AL, Hoch JA, Losick R: Bacillus subtilis from Cells to
Genes and from Genes to Cells. In Bacillus subtilis and its Closest
Relatives Edited by: Sonoshein AL, Hoch JA, Losick R. Washington
D.C.: ASM Press; 2001:1-6.
6.Barabote RD, Saier MH Jr: Comparative genomic analyses of
the bacterial phosphotransferase system. Microbiol Mol Biol Rev
2005, 69:608-634.
7.Gorke B, Stulke J: Carbon catabolite repression in bacteria:
many ways to make the most out of nutrients. Nat Rev Micro-
biol 2008, 6:613-624.
8.Lorca GL, Chung YJ, Barabote RD, Weyler W, Schilling CH, Saier MH
Jr: Catabolite repression and activation in Bacillus subtilis:
dependency on CcpA, HPr, and HprK. J Bacteriol 2005,
187:7826-7839.
9.Sonenshein AL: Control of key metabolic intersections in
Bacillus subtilis. Nature Reviews Microbiology 2007, 5:917-927.
10. Schilling O, Frick O, Herzberg C, Ehrenreich A, Heinzle E, Wittmann
C, et al.: Transcriptional and metabolic responses of Bacillus
subtilis to the availability of organic acids: transcription reg-
ulation is important but not sufficient to account for meta-
bolic adaptation. Appl Environ Microbiol 2007, 73:499-507.
11.Kaan T, Homuth G, Mader U, Bandow J, Schweder T: Genome-
wide transcriptional profiling of the Bacillus subtilis cold-
shock response. Microbiology 2002, 148:3441-3455.
12.Ye RW, Tao W, Bedzyk L, Young T, Chen M, Li L: Global gene
expression profiles of Bacillus subtilis grown under anaerobic
conditions. J Bacteriol 2000, 182:4458-4465.
13. Gutierrez-Rios RM, Freyre-Gonzalez JA, Resendis O, Collado-Vides J,
Saier M, Gosset G: Identification of regulatory network topo-
logical units coordinating the genome-wide transcriptional
response to glucose in Escherichia coli. BMC Microbiol 2007,
7:53.
14.Shafikhani SH, Partovi AA, Leighton T: Catabolite-induced repres-
sion of sporulation in Bacillus subtilis. Curr Microbiol 2003,
47:300-308.
15.Sierro N, Makita Y, de Hoon M, Nakai K: DBTBS: a database of
transcriptional regulation in Bacillus subtilis containing
Additional File 1
Supplementary Table 1SM. "VazquezHernandezSupplementary-
Material_1" and contains tables from 1 to 3, describe in the manuscript
as Table 1SM.
Click here for file
[http://www.biomedcentral.com/content/supplementary/1471-
2180-9-176-S1.xls]
Additional File 2
Supplementary Tables 2-3SM. "VazquezHernandezSupplementary-
Material_2" and contains Tables 2 to 3, described in the manuscript as
Table 2aSM, Table 2bSM, and Table 3SM.
Click here for file
[http://www.biomedcentral.com/content/supplementary/1471-
2180-9-176-S2.pdf]

Page 14

Publish with BioMed Central and every
scientist can read your work free of charge
"BioMed Central will be the most significant development for
disseminating the results of biomedical research in our lifetime."
Sir Paul Nurse, Cancer Research UK
Your research papers will be:
available free of charge to the entire biomedical community
peer reviewed and published immediately upon acceptance
cited in PubMed and archived on PubMed Central
yours — you keep the copyright
Submit your manuscript here:
http://www.biomedcentral.com/info/publishing_adv.asp
BioMedcentral
BMC Microbiology 2009, 9:176 http://www.biomedcentral.com/1471-2180/9/176
Page 14 of 14
(page number not for citation purposes)
upstream intergenic conservation information. Nucleic Acids
Res 2008, 36:D93-D96.
Gutierrez-Rios RM, Rosenblueth DA, Loza JA, Huerta AM, Glasner
JD, Blattner FR, et al.: Regulatory network of Escherichia coli:
consistency between literature knowledge and microarray
profiles. Genome Res 2003, 13:2435-2443.
Moszer I, Jones LM, Moreira S, Fabry C, Danchin A: SubtiList: the
reference database for the Bacillus subtilis genome. Nucleic
Acids Res 2002, 30:62-65.
Nakano MM, Zuber P: Anaerobic growth of a "strict aerobe"
(Bacillus subtilis). Annu Rev Microbiol 1998, 52:165-190.
Fujita M, Sadaie Y: Rapid isolation of RNA polymerase from
sporulating cells of Bacillus subtilis. Gene 1998, 221:185-190.
Jedrzejas MJ, Huang WJ: Bacillus species proteins involved in
spore formation and degradation: from identification in the
genome, to sequence analysis, and determination of function
and structure. Crit Rev Biochem Mol Biol 2003, 38:173-198.
Piggot PJ, Hilbert DW: Sporulation of Bacillus subtilis. Curr Opin
Microbiol 2004, 7:579-586.
Mekjian KR, Bryan EM, Beall BW, Moran CP Jr: Regulation of hex-
uronate utilization in Bacillus subtilis. J Bacteriol 1999,
181:426-433.
Yoshida K, Yamaguchi H, Kinehara M, Ohki YH, Nakaura Y, Fujita Y:
Identification of additional TnrA-regulated genes of Bacillus
subtilis associated with a TnrA box. Mol Microbiol 2003,
49:157-165.
Eichenberger P, Fujita M, Jensen ST, Conlon EM, Rudner DZ, Wang
ST, et al.: The program of gene transcription for a single dif-
ferentiating cell type during sporulation in Bacillus subtilis.
PLoS Biol 2004, 2:e328.
Kroos L, Kunkel B, Losick R: Switch protein alters specificity of
RNA polymerase containing a compartment-specific sigma
factor. Science 1989, 243:526-529.
Au N, Kuester-Schoeck E, Mandava V, Bothwell LE, Canny SP, Chachu
K, et al.: Genetic composition of the Bacillus subtilis SOS sys-
tem. J Bacteriol 2005, 187:7655-7666.
Lozada-Chavez I, Janga SC, Collado-Vides J: Bacterial regulatory
networks are extremely flexible in evolution. Nucleic Acids Res
2006, 34:3434-3445.
Madan BM, Teichmann SA, Aravind L: Evolutionary dynamics of
prokaryotic transcriptional regulatory networks. J Mol Biol
2006, 358:614-633.
Gonzalez Perez AD, Gonzalez GE, Espinosa AV, Vasconcelos AT,
Collado-Vides J: Impact of Transcription Units rearrangement
on the evolution of the regulatory network of gamma-pro-
teobacteria. BMC Genomics 2008, 9:128.
Panina EM, Mironov AA, Gelfand MS: Comparative analysis of
FUR regulons in gamma-proteobacteria. Nucleic Acids Res
2001, 29:5195-5206.
Salgado H, Gama-Castro S, Peralta-Gil M, Diaz-Peredo E, Sanchez-
Solano F, Santos-Zavaleta A, et al.: RegulonDB (version 5.0):
Escherichia coli K-12 transcriptional regulatory network,
operon organization, and growth conditions. Nucleic Acids Res
2006, 34:D394-D397.
Presecan-Siedel E, Galinier A, Longin R, Deutscher J, Danchin A, Gla-
ser P, et al.: Catabolite regulation of the pta gene as part of
carbon flow pathways in Bacillus subtilis. J Bacteriol 1999,
181:6889-6897.
Voigt B, Schweder T, Becher D, Ehrenreich A, Gottschalk G, Feesche
J, et al.: A proteomic view of cell physiology of Bacillus licheni-
formis. Proteomics 2004, 4:1465-1490.
Pedraza-Reyes M, Yasbin RE: Contribution of the mismatch
DNA repair system to the generation of stationary-phase-
induced mutants of Bacillus subtilis. J Bacteriol 2004,
186:6485-6491.
Kim JH, Park IS, Kim BG: Development and characterization of
membrane surface display system using molecular chap-
eron, prsA, of Bacillus subtilis. Biochem Biophys Res Commun
2005, 334:1248-1253.
Schnorpfeil M, Janausch IG, Biel S, Kroger A, Unden G: Generation
of a proton potential by succinate dehydrogenase of Bacillus
subtilis functioning as a fumarate reductase. Eur J Biochem
2001, 268:3069-3074.
Hernandez-Montes G, Diaz-Mejia JJ, Perez-Rueda E, Segovia L: The
hidden universal distribution of amino acid biosynthetic net-
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
works: a genomic perspective on their origins and evolution.
Genome Biol 2008, 9:R95.
Henkin TM, Grundy FJ, Nicholson WL, Chambliss GH: Catabolite
repression of alpha-amylase gene expression in Bacillus sub-
tilis involves a trans-acting gene product homologous to the
Escherichia coli lacl and galR repressors. Mol Microbiol 1991,
5:575-584.
Ibarra JA, Perez-Rueda E, Segovia L, Puente JL: The DNA-binding
domain as a functional indicator: the case of the AraC/XylS
family of transcription factors. Genetica 2008, 133:65-76.
Teichmann SA, Babu MM: Gene regulatory network growth by
duplication. Nat Genet 2004, 36:492-496.
Reents H, Munch R, Dammeyer T, Jahn D, Hartig E: The Fnr regu-
lon of Bacillus subtilis. J Bacteriol 2006, 188:1103-1112.
Schroder I, Darie S, Gunsalus RP: Activation of the Escherichia
coli nitrate reductase (narGHJI) operon by NarL and Fnr
requires integration host factor. J Biol Chem 1993, 268:771-774.
Kolesnikow T, Schroder I, Gunsalus RP: Regulation of narK gene
expression in Escherichia coli in response to anaerobiosis,
nitrate, iron, and molybdenum. J Bacteriol 1992, 174:7104-7111.
Breitling R, Herzyk P: Rank-based methods as a non-parametric
alternative of the T-statistic for the analysis of biological
microarray data. J Bioinform Comput Biol 2005, 3:1171-1189.
Sierro N, Makita Y, de Hoon M, Nakai K: DBTBS: a database of
transcriptional regulation in Bacillus subtilis containing
upstream intergenic conservation information. Nucleic Acids
Res 2008, 36:D93-D96.
Resendis-Antonio O, Freyre-Gonzalez JA, Menchaca-Mendez R, Guti-
errez-Rios RM, Martinez-Antonio A, Avila-Sanchez C, et al.: Modular
analysis of the transcriptional regulatory network of E. coli.
Trends Genet 2005, 21:16-20.
Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ: Basic local
alignment search tool. J Mol Biol 1990, 215:403-410.
Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, et
al.: Gapped BLAST and PSI-BLAST: a new generation of pro-
tein database search programs. Nucleic Acids Res 1997,
25:3389-3402.
de Hoon MJ, Imoto S, Nolan J, Miyano S: Open source clustering
software. Bioinformatics 2004, 20:1453-1454.
Eisen MB, Spellman PT, Brown PO, Botstein D: Cluster analysis
and display of genome-wide expression patterns. Proc Natl
Acad Sci USA 1998, 95:14863-14868.
Saldanha AJ: Java Treeview--extensible visualization of micro-
array data. Bioinformatics 2004, 20:3246-3248.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.