Genome-Wide Analyses of Escherichia coli Gene Expression Responsive to the BaeSR Two-Component Regulatory System

Abstract

The BaeSR two-component regulatory system controls expression of exporter genes conferring drug resistance in Escherichia coli (S. Nagakubo, K. Nishino, T. Hirata, and A. Yamaguchi, J. Bacteriol. 184:4161-4167, 2002; N. Baranova and H. Nikaido, J. Bacteriol. 184:4168-4176, 2002). To understand the whole picture of BaeSR regulation, a DNA microarray analysis of the effect of BaeR overproduction was performed. BaeR overproduction activated 59 genes related to two-component signal transduction, chemotactic responses, flagellar biosynthesis, maltose transport, and multidrug transport, and BaeR overproduction also repressed the expression of the ibpA and ibpB genes. All of the changes in the expression levels were also observed by quantitative real-time reverse transcription-PCR analysis. The expression levels of 15 of the 59 BaeR-activated genes were decreased by deletion of baeSR. Of 11 genes induced by indole (a putative inducer of the BaeSR system), 10 required the BaeSR system for induction. Combination of the expression data sets revealed a BaeR-binding site sequence motif, 5′-TTTTTCTCCATDATTGGC-3′ (where D is G, A, or T). Several genes up-regulated by BaeR overproduction, including genes for maltose transport, chemotactic responses, and flagellar biosynthesis, required an intact PhoBR or CreBC two-component regulatory system for up-regulation. These data indicate that there is cross-regulation among the BaeSR, PhoBR, and CreBC two-component regulatory systems. Such a global analysis should reveal the regulatory network of the BaeSR system.

Bacteria have developed signaling systems for eliciting a variety of adaptive responses to their environments. These adaptive responses are often mediated by two-component regulatory systems. A typical two-component regulatory system is composed of a histidine kinase sensor residing in the inner membrane and a cognate response regulator in the cytoplasm. Similar systems control the expression of genes for nutrient acquisition, virulence, antibiotic resistance, and numerous other pathways in diverse bacteria (2, 17, 47, 48). There are also analogous signaling systems in cells of eukaryotes, including fungi, amoebae, and plants (19, 31, 59, 60, 62). In Escherichia coli, 29 histidine kinase sensors, 32 response regulators, and one HPt (histidine-containing phosphotransmitter) domain protein have been found during analyses of the E. coli K-12 genome (34). Each sensor responds to specific environmental changes to cope with the numerous conditions that E. coli faces. The functions of many of these systems remain undetermined.

In a previous study, it was found that the BaeSR two-component system modulates the drug resistance of E. coli by regulating the expression of drug transporter genes (5, 36). The response regulator BaeR modulates the expression of mdtABC and acrD, which encode multidurug exporter systems (15, 16, 36). Overproduction of BaeR, in the background of a deficiency of the E. coli major multidrug exporter AcrB, confers resistance against β-lactams, novobiocin, sodium dodecyl sulfate, and bile salts. However, the physiological role of the BaeSR system has remained unknown.

We hypothesized that the BaeSR system controls the expression of a wide range of genes. E. coli microarrays have been successfully used to quantify the entire complement of individual mRNA transcripts (40, 46). Therefore, to reveal the whole picture of the BaeSR-controlled genes, a microarray analysis of genes affected by BaeR overproduction was performed in this study. The expression levels of all of the BaeR-affected genes were also investigated by quantitative real-time reverse transcription-PCR (qRT-PCR) analysis. Also, we investigated the effect of baeSR deletion on the levels of expression of genes by qRT-PCR analysis. In order to understand the response of the BaeSR system to signals, we examined the effect of addition of indole on gene expression levels. Combination of these expression data sets revealed a BaeR-binding site sequence motif. Furthermore, we examined the effects of deletion of phoBR or creBC on the levels of expression of the BaeR-induced genes to elucidate the genetic network of the BaeSR system.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

The strains used in this work were E. coli K-12 derivatives (Table 1). They were grown at 37°C in Luria-Bertani broth in the absence of isopropyl-β-d-thiogalactopyranoside (56). The cells were rapidly collected for total RNA extraction when the culture reached an optical density at 600 nm of 0.6.

TABLE 1.

Strains and plasmids

Strain or plasmid	Genotype	Source or reference
E. coli strains
W3110	Wild type	Laboratory stock
KAM3	acrB-deficient mutant of TG1	32
NKE41	KAM3/pUC118	This study
NKE42	KAM3/pUCbaeR	This study
NKE10	W3110/pUC118	This study
NKE11	W3110/pUCbaeR	This study
NKE52	W3110 ΔbaeSR	This study
NKE56	W3110 ΔcreBC/pUCbaeR	This study
NKE57	W3110 ΔphoBR/pUCbaeR	This study
NKE122	W3110 ΔBaeR binding motif for acrD, pUCbaeR	This study
NKE124	W3110 ΔBaeR binding motif for mdtA, pUCbaeR	This study
Plasmids
pUC118		Takara Bio, Inc.
pUCbaeR	BamHI-PstI fragment containing baeR with 100-bp upstream flanking sequence cloned into pUC118	This study

Plasmid construction.

The baeR gene was amplified from W3110 genoimic DNA by using primers 5′-CGCGGATCCTTGAAGCACATAATGGTCGCA-3′ and 5′-CGCCTGCAGCTAAACGATGCGGCAGGCGTC-3′, which introduced BamHI and PstI sites at the ends of the amplified fragment. This fragment contained baeR with a 100-bp upstream sequence. The PCR fragment was cloned between the BamHI and PstI sites of vector pUC118 (Takara Bio Inc., Shiga, Japan), which resulted in the baeR gene being arranged in the same orientation as the lactose promoter of the pUC118 vector. The nucleotide sequence of the recombinant plasmid was determined with an ABI PRISM 3100-Avant genetic analyzer (Applied Biosystems).

Susceptibility testing.

The antibacterial activities of agents were determined on L agar (1% tryptone, 0.5% yeast extract, 0.5% NaCl) plates containing novobiocin or sodium deoxycholate (Sigma) at various concentrations. Agar plates were prepared by the twofold agar dilution technique recommended by the Japan Society of Chemotherapy (20, 21). Organisms were tested by using a final inoculum size of 10⁴ CFU/spot, delivered with a multipoint inoculator (Sakuma Seisakusyo, Tokyo, Japan), and were incubated at 37°C for 18 h in air. The MICs of compounds were defined as the lowest concentrations that severely inhibited bacterial cell growth.

Construction of gene deletion mutants.

Gene disruption was performed by the method of Datsenko and Wanner (11). The chloramphenicol resistance gene (cat), flanked by Flp recognition target sites, was amplified by PCR with primers with 40-nucleotide extensions that were homologous to the beginning and end of the coding sequence of the gene or of the BaeR-binding motif to be disrupted. Plasmid pKD3 was used as a template. The resulting PCR product was used to transform the recipient W3110 strain expressing Red recombinase, and recombinant clones were isolated as chloramphenicol-resistant colonies. Chromosomal DNA was isolated from the mutants obtained, and the structures of the deleted loci were confirmed by performing a series of PCRs with primers complementary to cat and to adjacent regions. cat was eliminated by using plasmid pCP20 as described previously (11).

RNA extraction.

Total RNA was isolated from bacterial cultures with an RNeasy Protect Bacteria mini kit (QIAGEN) and RNase-free DNase (QIAGEN) as described previously (40, 43, 45). The absence of genomic DNA in DNase-treated RNA samples was confirmed by both inspecting nondenaturing agarose electrophoresis gels and performing PCR with primers known to target the genomic DNA. The RNA concentration was determined spectrophotometrically (56).

DNA microarray analysis.

Total RNA from strains NKE10 and NKE11 was extracted from mid-exponential-phase cultures as described above. Preparation of fluorescently labeled cDNA, microarray hybridization, and data analysis were performed as described previously (46). In brief, cDNA labeled with Cy3-dUTP (NKE10) and Cy5-dUTP (NKE11) was synthesized from each lot of total RNA by random priming. Labeled cDNA probes were purified and hybridized to glass slide microarrays (IntelliGene E. coli CHIP, version 2.0; Takara Bio, Inc.). The slides were scanned for fluorescence intensity by using a 428 array scanner (Affymetrix). The signal density of each spot in an array was quantified by using the ImaGene software, version 4.2 (BioDiscovery). Basically, a normalized relative Cy5/Cy3 ratio of 4 or above was considered a significant increase in expression, and a ratio of 0.25 or below was considered a significant decrease in expression.

Determination of specific transcript levels by quantitative, real-time PCR following reverse transcription.

Bulk cDNA samples were synthesized from total RNA derived from E. coli cells by using TaqMan reverse transcription reagents (PE Applied Biosystems) and random hexamers as primers. Specific primer pairs were designed with the ABI PRISM Primer Express software (PE Applied Biosystems). rrsA of the 16S rRNA gene was chosen as the normalizing gene. Real-time PCR was performed with each specific primer pair by using SYBR Green PCR master mixture (PE Applied Biosystems). The reactions were performed with an ABI PRISM 7000 sequence detection system (PE Applied Biosystems), and during the reactions the fluorescence signal due to SYBR Green intercalation was monitored to quantify the double-stranded DNA product formed in each PCR cycle.

RESULTS

Effect of overexpression of baeR on gene expression.

To determine the effect of overexpression of baeR on gene expression, we constructed high-copy-number plasmid pUCbaeR containing the baeR gene (Table 1). The baeR gene was in the same orientation as the lactose promoter of the pUC118 vector. Thus, baeR was expected to be expressed from the lactose promoter. This plasmid conferred novobiocin and deoxycholate resistance to acrB-deficient strain KAM3. The MICs of novobiocin for NKE41 (KAM3/pUC118) and NKE42 (KAM3/pUCbaeR) were 2 and 16 μg/ml, respectively. The MICs of deoxycholate for these strains were 1,250 and 40,000 μg/ml, respectively. pUC118 and pUCbaeR were used in the following microarray experiments.

DNA microarrays, which contain most of the genomic open reading frames of E. coli (46), allow comprehensive studies of BaeR-controlled E. coli gene expression. We used the E. coli W3110 strain as a host for microarray analysis, because the DNA microarrays were made from DNA fragments of W3110 (Takara Bio, Inc.). pUCbaeR did not confer drug resistance to wild-type E. coli strain W3110, because the intrinsic multidrug exporter AcrB masks the effect of baeR overexpression. pUCbaeR conferred multidrug resistance to W3110 ΔacrB and to KAM3 (data not shown). NKE10 has a single copy of baeR in its chromosome and harbors the vector plasmid pUC118, while NKE11 bears high-copy-number plasmid pUCbaeR (Table 1). The comprehensive transcript profiles of these two strains prepared from exponential-phase cells were compared. The increased baeR gene dosage in NKE11 resulted in a 24-fold increase in the expression of cognate baeR transcripts, and the expression of 58 other genes (open reading frames) was elevated more than fourfold, while the expression of two genes (ibpA and ibpB) was repressed more than fourfold (Table 2). We reinvestigated the BaeR-dependent induction of these genes by qRT-PCR analysis. All changes in gene expression greater than fourfold were also observed by qRT-PCR. The degree of induction determined by microarray analysis was usually lower than that determined by qRT-PCR, probably because the dynamic range of the former analysis is narrower than that of the latter (Table 2) (40).

TABLE 2.

E. coli genes whose relative expression levels were increased or decreased by baeR amplification

Genea	b no.b	Known or predicted functionc	Effect of BaeR on gene expression (fold change)
			Microarray	Real-time PCR
Increased expression
creD	b4400	Tolerance to colicin E2	180	35,000
spy	b1743	Periplasmic protein related to spheroblast formation	140	640
mdtA (yegM)	b2074	Membrane protein inolved in drug resistance	140	490
mdtB (yegN)	b2075	Multidrug transport protein (RND family)	24	280
baeR	b2079	Response regulator in two-component regulatory system with BaeS	24	15,000
malE	b4034	Maltose transport protein, chemotaxis (ABC superfamily)	22	37
amn	b1982	AMP nucleosidase	22	21
yiaD	b3552	Putative outer membrane protein	20	79
ansB	b2957	Periplasmic l-asparaginase II	19	76
yeeN	b1983	Conserved protein	15	17
malK	b4035	Maltose transport protein (ABC superfamily) (N terminal): phenotypic repressor of mal operon (C terminal)	14	73
garP (yhaU)	b3127	Putative (d)-glucarate/galactarate transport protein (MFS family)	13	86
gsp	b2988	Glutathionylspermidine amidase (N terminal); glutathionylspermidine synthetase (C terminal)	13	29
malM	b4037	Periplasmic protein of mal regulon	12	35
lamB	b4036	Maltoporin, high-affinity receptor for maltose and maltose oligosaccharides	12	26
yfjB	b2615	NAD kinase	10	18
yieI	b3716	Putative membrane protein	9.3	50
mdtC (yegO)	b2076	Multidrug transport protein (RND family)	9.1	190
tsr	b4355	Methyl-accepting chemotaxis protein 1, serine sensor receptor	8.9	14
baeS	b2078	Sensory histidine kinase in two-component regulatory system wtih BaeR	8.0	120
tolC	b3035	Outer membrane channel; tolerance to colicin E1	7.9	12
yieJ	b3717	Conserved hypothetical protein	7.8	48
phoB	b0399	Response regulator in two-component regulatory system with PhoR (or CreC)	7.8	18
elaA	b2267	Putative transferase	7.3	15
yghU	b2989	Putative S-transferase	7.2	14
cheM (tar)	b1886	Methyl-accepting chemotaxis protein II, aspartate sensor receptor	7.1	6.3
fliC	b1923	Flagellar biosynthesis	6.6	6.4
garD (yhaG)	b3128	(d)-Galactarate dehydrogenase	6.5	18
tap	b1885	Methyl-accepting chemotaxis protein IV, peptide sensor receptor	6.0	7.4
garL (yhaF)	b3126	Alpha-dehydro-beta-deoxy-d-glucarate aldolase	6.0	32
cheR	b1884	Glutamate methyltransferase, chemotactic response regulator	5.8	9.1
yidS	b3690	Putative oxidoreductase with FAD/NAD(P)-binding domaind	5.7	130
ypfH	b2473	Conserved protein	5.6	9.1
phnD	b4105	Phosphonate transport protein (ABC superfamily, peri_bind)	5.6	140
malF	b4033	Maltose transport protein (ABC superfamily, membrane)	5.5	28
sulA	b0958	Suppressor of lon; inhibitor of cell division and FtsZ ring formation upon DNA damage or inhibition	5.4	5.2
aroF	b2601	3-Deoxy-d-arabinoheptulosonate-7-phosphate synthase (DAHP synthetase), tyrosine repressible	5.4	9.5
flgL	b1083	Flagellar biosynthesis; hook-filament junction protein	5.4	5.9
malG	b4032	Maltose transport protein (ABC superfamily, membrane)	5.3	19
fliD	b1924	Flagellar biosynthesis; filament capping protein; enables filament assembly	5.3	8.1
glpF	b3927	MIP channel, glycerol diffusion	5.2	10
fliT	b1926	Flagellar biosynthesis; putative export chaperone for FliD	5.0	13
phnJ	b4098	Conserved protein in phn operon	5.0	60
ycaC	b0897	Putative cysteine hydrolase	5.0	7.3
cheA	b1888	Chemotactic sensory histidine kinase (soluble) in two-component regulatory system with CheB and CheY	5.0	8.1
icdA	b1136	e14 prophage; isocitrate dehydrogenase, specific for NADP+	4.9	31
phoR	b0400	Sensory histidine kinase in two-component regulatory system with PhoB	4.9	15
flxA	b1566	Qin prophage	4.7	7.7
fliS	b1925	Flagellar biosynthesis; repressor of class 3a and 3b operons (RflA activity)	4.7	6.2
motB	b1889	Enables flagellar motor rotation, linking torque machinery to cell wall	4.5	6.4
b1141	b1141	e14 prophage; putative exisionase	4.5	19
motA	b1890	Proton conductor component of motor, torque generator	4.4	6.6
agaZ	b3132	Tagatoso-6-phosphate aldolase I, subunit together with AgaY	4.4	23
cheW	b1887	Purine-binding chemotaxis protein; regulation	4.3	6.9
cheZ	b1881	Chemotactic response, CheY protein phophatase	4.1	39
ybiC	b0801	Putative dehydrogenase	4.1	8.2
tsx	b0411	Nucleoside channel; receptor of phage T6 and colicin K	4.1	5.8
acrD	b2470	Aminoglycoside/multidrug efflux pump (RND family)	4.1	35
mdtD (yegB)	b2077	Putative multidrug transport protein (MFS family)	4.0	110
Decreased expression
ibpB	b3686	Small heat shock protein	0.25	0.17
ibpA	b3687	Small heat shock protein	0.21	0.16

^aBased on information from the EcoGene database (http://bmb.med.maiami.edu/ecogene/ecoweb/) (55).

^bBlattner number. Based on information from the E. coli Genome Project database (http://www.genome.wisc.edu/) (6).

^cBased on information from the Gene ProtEC database (http://genprotec.mbl.edu/) (52, 53).

^dFAD, flavin adenine dinucleotide.

Known genes in the BaeR regulon.

In previous studies, it was found that overproduction of BaeR increases the expression of mdtABCD (yegMNOB) and acrD (5, 15, 36). Also, Raffa and Raivio have reported that envelope stress induces the expression of spy via a BaeSR signal transduction pathway (49). Both in the DNA microarray analysis and in the qRT-PCR analysis, enhancement of mdtA, mdtB, mdtC, mdtD, acrD, and spy expression was found (Table 2) (Fig. 1). We found that BaeR increased the expression of the outer membrane channel tolC gene (7.9-fold as determined by microarray analysis and 12-fold as determined by qRT-PCR), which is required for the function of the MdtABC and AcrD multidrug export systems (Table 2) (41, 42).

FIG. 1.

Gene clusters whose expression was increased by BaeR overproduction. The expression changes for genes detected by real-time qRT-PCR are indicated under the gene names. The numbers in parentheses indicate the expression level changes detected by microarray analysis. The kb (kilobase pair) data indicate the positions on E. coli chromosomal DNA, as annotated on the Colibri website (http://genolist.Pasteur.fr/Colibri/). baeR was overexpressed from the high-copy-number plasmid pUCbaeR. N.D., not determined.

Enhanced expression of gene clusters due to BaeR overproduction.

The expression of 11 gene clusters was increased by BaeR overproduction. These 11 gene clusters are located at different positions on the E. coli chromosome (Fig. 1). Mainly, these clusters contain genes related to two-component signal transduction, chemotactic responses, flagellar biosynthesis, maltose transport, and multidrug transport.

Elevated expression of two-component regulatory system genes.

The PhoBR two-component system controls the genes of the phosphate (Pho) regulon for assimilation of alternative P sources (65, 66). The microarray and qRT-PCR results showed that the expression levels of phoB and phoR were increased by baeR amplification (Fig. 1). BaeR overproduction increased the expression of baeS, which encodes a cognate sensor of the BaeR response regulator (37) and is located downstream of mdtABCD (42). It is thought that both mdtABCD and baeSR are regulated by BaeR, because they form an operon and are transcribed from the same promoter (Fig. 1) (6, 36). BaeR also increased the expression of the creB response regulator gene of the creBC two-component system. The CreBC system appears to be connected to carbon and energy metabolism (67). According to the microarray analysis results, the level of enhancement of expression of creB was 2.6-fold. Increased expression of the CreB regulon members (4), creD, malE, yidS, and yieI was observed by both microarray and qRT-PCR analyses (Table 2).

Elevated expression of genes related to flagellar synthesis and chemotaxis due to baeR overexpression.

In E. coli, flagellar biosynthesis, chemotaxis, and aerotaxis are coordinately regulated by flhDC (10). Flagellar genes are expressed in three stages, early, middle, and late. The two early genes, flhD and flhC, form an operon through which environmental control of flagellar synthesis is coordinated. Microarray analysis showed no significant effect of overexpression of baeR on flhD expression (1.0-fold increase) or flhC expression (1.2-fold increase). However, the expression of genes regulated by flhDC, including flgL, fliACDST, and motAB, was significantly increased (at least fourfold as determined by microarray analysis) (Fig. 1). These genes comprise both middle and late genes required for the synthesis and assembly of flagella, as well as transcriptional regulators of flagellar gene expression (flgM and fliA). Also regulated by flhDC are genes associated with chemotaxis and aerotaxis, including cheAMRWZ, tap, tsr, and aer. Expression of most of these genes was also increased more than fourfold in the baeR-overexpressing strain (Fig. 1); the exception was aer (3.1-fold increase as determined by microarray analysis).

Increased expression of the maltose operon due to baeR overexpression.

Maltose and maltodextrins are present at high concentrations in the intestinal tracts of animals as by-products of starch metabolism. Maltose and maltodextrin are transported through a pore consisting of maltoporin encoded by lamB, which serves as a channel for sugar migration across the outer membrane. Both the transport and the utilization of these compounds are regulated by MalT, a regulator required for transcription at mal promoters (9). Genomic analysis demonstrated that there was increased expression of the maltose system in the baeR-overexpressing strain (Fig. 1). The ratio of expression of malT in the baeR-overexpressing strain to expression of malT in the host strain was increased only modestly (to 1.7). The expression of genes under the control of MalT was, however, significantly increased (approximately 5- to 20-fold) (Table 2 and Fig. 1). BaeR increased the expression of malE, which encodes a maltose-binding protein, and malFGK, which encode the translocation complex. The genes used for maltose and maltodextrin metabolism also showed increased expression, as follows: malP, which encodes maltodextrin phosphorylase, 2.2-fold; malS, which encodes a nonessential maltodextrin-metabolizing enzyme, periplasmic α-amylase, 2.9-fold; and malM, a periplasmic protein with an unknown function, 12-fold. The expression of lamB, encoding maltoporin, the specific pore for maltodextirns and the receptor for phage λ in E. coli, was increased 12-fold.

Enhanced expression of other operons.

Elevated transcription of the gar operon (garPLRK) and garD was observed in the strain bearing the baeR amplification. Transcription of these genes was elevated 13-, 6.0-, 2.1-, 2.1-, and 6.5-fold, respectively (Fig. 1). These genes are related to d-galactarate metabolism (18, 35). BaeR increased the expression of three other clusters, one of which contains the amn and yeeN genes, one of which contains the gsp and yghU genes, and one of which contains the yieI and yieJ genes (Fig. 1). It has been reported that the amn gene encodes an AMP nucleosidase (26, 27) and that the gsp gene encodes a bifunctional glutathionylspermidine synthetase/amidase (8). The functions of the other genes are unknown (Table 2).

Effect of the baeSR deletion on gene expression.

To elucidate BaeSR regulation further, we investigated the effects of baeSR deletion on the gene expression levels. Total RNA from exponential-phase cells of W3110 and NKE52 (W3110 ΔbaeSR) was collected, and the gene expression levels were compared by the qRT-PCR method (Table 3). We chose target genes whose expression levels were enhanced by baeR amplification, as described above. As shown in Table 3, the expression of 15 genes was decreased by baeSR deletion by more than a factor of two. qRT-PCR revealed that expression of mdtA, mdtB, mdtC, mdtD, yidS, phnJ, cheR, cheZ, ycaC, phnD, b1141, spy, acrD, lamB, and tap decreased by factors of 9.6, 2.0, 2.0, 2.0, 8.3, 3.3, 3.1, 2.9, 3.0, 2.8, 2.8, 2.2, 2.0, 2.0, and 2.0, respectively.

TABLE 3.

Expression levels of E. coli genes in the baeSR deletion mutant

Gene	Effect of ΔbaeSR on gene expression (fold decrease)a
creD	1.9
spy	2.2
mdtA (yegM)	9.6
mdtB (yegN)	2.0
baeR	NE
malE	1.9
amn	1.2
yiaD	1.1
ansB	0.9
yeeN	1.6
malK	0.8
garP (yhaU)	0.7
gsp	1.7
malM	0.8
lamB	2.0
yfjB	1.1
yieI	1.8
mdtC (yegO)	2.0
tsr	1.4
baeS	NE
tolC	1.3
yieJ	1.4
phoB	1.7
elaA	1.3
yghU	1.3
cheM (tar)	1.5
fliC	1.7
garD (yhaG)	1.0
tap	2.0
garL (yhaF)	1.0
cheR	3.1
yidS	8.3
ypfH	1.5
phnD	2.8
malF	0.7
sulA	1.2
aroF	1.4
flgL	1.8
malG	0.8
fliD	1.3
glpF	0.7
fliT	1.8
phnJ	3.3
ycaC	3.0
cheA	1.6
icdA	1.3
phoR	1.4
flxA	1.1
fliS	1.2
motB	1.4
b1141	2.8
motA	1.4
agaZ	0.9
cheW	1.4
cheZ	2.9
ybiC	1.1
tsx	0.7
acrD	2.0
mdtD (yegB)	2.0

^aThe values in boldface type indicate decreases of more than a factor of two compared to the expression levels of the genes in W3110. NE, no expression.

Effect of addition of indole on gene expression.

Previously, Garbe et al. found that the Spy protein level is elevated when E. coli is grown in the presence of indole (13). Also, Raffa and Raivio found that the BaeSR system controls the expression of spy in response to the addition of indole (49). To help us understand the role of BaeSR regulation in response to indole, W3110 cells were grown in L-broth with or without 2 mM indole, after which total RNA was collected for qRT-PCR. A total of 59 genes were examined, and the expression of 11 of these genes was increased more than twofold when E. coli was grown in the presence of indole (Table 4). qRT-PCR showed that expression of glpF, spy, creD, mdtA, mdtB, mdtC, mdtD, ycaC, acrD, yghU, and icdA increased by factors of 20, 14, 5.2, 2.4, 2.4, 2.3, 2.4, 2.3, 2.3, 2.2, and 2.1, respectively. We also investigated whether this gene induction is dependent on the BaeSR system. Most of the expression levels (10 genes; the exception was glpF) were not increased in the ΔbaeSR strain following growth in the presence of indole. Thus, indole appears to induce the expression of these genes via the BaeSR two-component signal transduction system.

TABLE 4.

Effect of addition of indole on the levels of gene expression

Gene	Fold increasea
	Strain W3110	Strain NKE52 (ΔbaeSR)
creD	5.2	1.0
spy	14	1.5
mdtA (yegM)	2.4	0.7
mdtB (yegN)	2.4	0.8
baeR	1.6	NE
malE	0.6	0.6
amn	1.4	1.1
yiaD	1.8	1.9
ansB	0.6	0.4
yeeN	0.5	0.5
malK	0.2	0.2
garP (yhaU)	0.4	0.2
gsp	1.2	0.8
malM	0.5	0.3
lamB	0.3	0.4
yfjB	1.0	0.7
yieI	1.2	0.8
mdtC (yegO)	2.3	0.9
tsr	0.1	0.1
baeS	1.4	NE
tolC	1.4	0.9
yieJ	1.1	0.7
phoB	1.0	0.7
elaA	1.2	0.8
yghU	2.2	1.3
cheM (tar)	0.2	0.1
fliC	0.1	0.1
garD (yhaG)	0.3	0.2
tap	0.2	0.1
garL (yhaF)	0.6	0.3
cheR	0.3	0.2
yidS	1.9	0.7
ypfH	1.3	0.9
phnD	1.2	0.1
malF	0.5	0.4
sulA	1.0	0.6
aroF	1.1	0.8
flgL	0.2	0.1
malG	0.7	0.5
fliD	0.1	0.0
glpF	20	13
fliT	0.1	0.0
phnJ	1.2	0.4
ycaC	2.3	1.1
cheA	0.1	0.1
icdA	2.1	1.1
phoR	1.4	0.7
flxA	0.1	0.0
fliS	0.1	0.1
motB	0.1	0.1
b1141	1.4	0.9
motA	0.1	0.1
agaZ	1.1	0.6
cheW	0.2	0.1
cheZ	0.4	0.1
ybiC	1.5	1.0
tsx	0.3	0.2
acrD	2.3	0.5
mdtD (yegB)	2.4	1.1

^aThe values in boldface type indicate increases of more than twofold compared to the expression levels of the genes in W3110 without addition of indole. NE, no expression.

BaeR-binding site sequence motifs.

We summarize the expression data sets in Fig. 2. Our microarray analysis identified 59 genes whose expression was increased by baeR overexpression. The expression of 15 genes was decreased by ΔbaeSR, and the expression of 11 genes was increased by indole. Using these three expression data sets, we identified seven genes as members of the BaeSR regulon (Fig. 2). This regulon contains the spy, mdtABCD, acrD, and ycaC genes. These genes are most likely regulated by BaeR directly. To identify the BaeR-binding motif, we analyzed upstream regions of spy, the mdtABCD operon, acrD, and ycaC. Analysis with the motif-finding program Align ACE (54) revealed a highly conserved 18-bp sequence in the upstream regions of spy, mdtA, and acrD (Fig. 3). The 18-bp consensus sequence is 5′-TTTTTCTCCATDATTGGC-3′ (where D is G, A, or T). The program also found a similar sequence in the upstream region of ycaC, but the level of identity was low (50%) (Fig. 3). To test whether this sequence is required for gene induction by BaeR, we investigated the effect of deletion of the BaeR-binding motif on gene induction. We deleted the BaeR-binding motifs for the acrD gene and for the mdt operon. Overexpression of BaeR did not increase the expression levels of these genes without the BaeR-binding motifs (Table 5). Baranova and Nikaido reported that BaeR binds to the upstream region of mdtA (5). Recently, our group found that BaeR binds to the upstream regions of both mdtA and acrD (unpublished data). Also, the consensus sequence was found in the promoter region of spy. Thus, we suggest that expression of spy, mdtA, and acrD is directly regulated by the BaeSR two-component system. Searches of the regions upstream (400 bp) of all 59 genes whose expression was increased by BaeR overexpression did not reveal sequences matching the consensus sequence other than spy, mdtA, acrD, and ycaC.

FIG. 2.

Intersections of genomic data sets. The diagram summarizes the data for 59 genes whose expression levels were increased by BaeR overproduction and for which there were valid data for expression profiles of the baeSR mutant and wild type with the addition of indole. The expression levels of 15 of the 59 genes were decreased by deletion of baeSR, and 11 genes were identified as indole-regulated genes. The seven genes in the overlap of all three data sets were identified as components of the BaeSR regulon, as described in the text.

FIG. 3.

Consensus sequence in the upstream regions of BaeR-regulated genes. Consensus sequences were found in the upstream regions of mdtA, spy, acrD, and ycaC. The numbering is relative to the start codon of the genes.

TABLE 5.

Effect of deletion of the BaeR-binding site sequence

Gene	Fold increasea
	Strain NKE11 (wild type, pUCbaeR)	Strain NKE122 (ΔBaeR binding motif for acrD, pUCbaeR)	Strain NKE124 (ΔBaeR binding motif for mdtA, pUCbaeR)
mdtA (yegM)	490	510	1.1
mdtB (yegN)	280	300	0.6
mdtC (yegO)	190	190	0.4
mdtD (yegB)	110	120	1.1
baeS	120	130	0.6
acrD	35	1.7	38
tolC	12	13	12
spy	640	600	590
baeR	15,000	17,000	11,000

^aThe values in boldface type indicate increases of more than twofold compared to the expression levels of the genes in NKE10 (W3110/pUC118).

Effect of deletion of phoBR or creBC on BaeR-induced gene expression.

As described above, expression of phoBR two-component system genes was increased by baeR amplification (Table 2). Overproduction of BaeR also increased expression of the CreB regulon (4), including expression of creD, malE, yidS, and yieI (Table 2) and the creB response regulator gene (2.6-fold increase as determined by microarray analysis). These data indicate that there is a possibility of cross-regulation among the PhoBR, CreBC, and BaeSR two-component systems. To test this hypothesis, we investigated the effects of deletion of phoBR and creBC on gene induction by baeR amplification. Total RNA was collected from exponential-phase cells of NKE57 (W3110 ΔphoBR/pUCbaeR), NKE56 (W3110 ΔcreBC/pUCbaeR), and NKE10 (W3110/pUC118) for qRT-PCR analysis. The expression levels of genes in NKE57 and NKE56 were compared to those in NKE10 (Table 6). The amplification of baeR in NKE56 (8,400-fold) was lower than that in NKE57 (21,000-fold), probably because the growth of NKE56 was slower than the growth of NKE57 (data not shown). Overproduction of BaeR induced the expression of 40 genes in the ΔphoBR strain and the expression of 38 genes in the ΔcreBC strain by more than twofold. To put it differently, BaeR did not increase the expression of 17 genes in the ΔphoBR strain or the expression of 21 genes in the ΔcreBC strain that were induced by BaeR in W3110. Seventeen of the genes overlapped, probably because of cross talk between the PhoBR system and the CreBC system (63, 64). These 17 genes are malEFGKM-lamB (related to maltose transport), cheAMRWZ-tap (related to chemotactic response), and fliCDST/flgL (related to flagellar biosynthesis). The expression of yieIJ, yidS, and glpF was not increased by baeR overexpression only in the ΔcreBC strain. These data indicate that there is a novel interaction among the BaeSR, PhoBR, and CreBC two-component systems.

TABLE 6.

Effect of deletion of phoBR and creBC on expression levels of genes that were increased by baeR amplification

Gene	Fold increasea
	Strain NKE57 (ΔphoBR)	Strain NKE56 (ΔcreBC)
creD	1,400	31
spy	650	240
mdtA (yegM)	380	310
mdtB (yegN)	54	53
baeR	21,000	8,400
malE	0.0	0.2
amn	16	8.8
yiaD	23	23
ansB	44	27
yeeN	8.6	4.2
malK	0.0	0.1
garP (yhaU)	18	18
gsp	14	10
malM	0.0	0.5
lamB	0.0	0.2
yfjB	11	7.6
yieI	20	1.0
mdtC (yegO)	100	47
tsr	11	4.5
baeS	50	28
tolC	6.0	3.0
yieJ	14	0.8
phoB	NE	9.1
elaA	6.4	3.4
yghU	8.2	6.6
cheM (tar)	0.6	0.8
fliC	0.4	0.5
garD (yhaG)	6.0	6.1
tap	1.1	0.8
garL (yhaF)	4.3	3.3
cheR	1.2	1.1
yidS	7.3	1.0
ypfH	3.4	3.3
phnD	300	70
malF	0.1	0.3
sulA	2.8	2.1
aroF	11	8.0
flgL	0.9	0.6
malG	0.1	0.3
fliD	0.8	0.9
glpF	3.7	0.4
fliT	0.9	1.0
phnJ	31	16
ycaC	11	3.9
cheA	1.1	0.9
icdA	6.0	6.4
phoR	NE	5.2
flxA	3.5	2.9
fliS	0.7	0.8
motB	2.9	2.8
b1141	4.3	3.2
motA	2.7	2.7
agaZ	5.1	5.7
cheW	0.6	0.9
cheZ	0.6	1.1
ybiC	4.8	4.4
tsx	3.3	4.0
acrD	75	22
mdtD (yegB)	94	42

^aThe values in boldface type indicate increases of more than twofold compared to the expression levels of the genes in NKE10 (W3110/pUC118). NE, no expression.

DISCUSSION

In this work we examined the utility of microarray analysis for determining the global effects of baeR gene dosage amplification. In this study, we found a lot of genes whose BaeR dependence was not known previously. We discovered that overproduction of BaeR affects the expression of gene clusters related to two-component signal transduction, maltose transport, chemotatic responses, flagellar biosynthesis, and multidrug transport. Since a large amount of BaeR may cause indirect regulation and this may not occur under the normal growth conditions, we also investigated the effect of deletion of baeSR on gene expression levels by using qRT-PCR methods. Also, in order to understand the whole picture of BaeSR regulation in response to signals, we investigated the effect of addition of indole on gene expression by qRT-PCR analysis. However, the data obtained, like all the expression profile data, cannot be used to distinguish direct targets from indirect targets, a critical distinction when regulatory networks are mapped. In this study, we combined the three expression data sets to define the members of the BaeSR regulon. Analysis of upstream regions of the BaeSR regulon with the motif-finding program revealed an 18-bp consensus sequence, 5′-TTTTTCTCCATDATTGGC-3′ (where D is G, A, or T), and it revealed that this consensus sequence is required for gene induction by BaeR. The consensus sequence is present in the upstream regions of the spy gene, the acrD gene, and the mdtABCD-baeSR operon. We concluded that only these three regulons are directly regulated by the BaeSR system. Furthermore, qRT-PCR analysis with the ΔphoBR ΔcreBC strain revealed novel cross-regulation among the BaeSR, PhoBR, and CreBC two-component systems.

Spy (spheroplast protein Y), one of the components of the BaeSR regulon, was initially identified as a periplasmic protein whose expression is induced by spheroplast formation (14). Raivio et al. found that spy is a component of a regulon of the CpxAR two-component system that responds to envelope stresses (51), but the function of Spy remains unknown. BaeR controls the expression of acrD and mdtABC multidrug efflux system genes. Both AcrD and MdtABC belong to the resistance-nodulation-cell division (RND) transporter family that plays a major role in producing both intrinsic and elevated levels of resistance to a very wide range of noxious compounds in gram-negative bacteria (30, 38, 39). AcrD and MdtABC drug exporters need outer membrane protein TolC in order to function (15, 36), like some other drug transporters of E. coli (12, 23, 24, 41). In this study, we found that BaeR overproduction increased the expression of tolC. This situation is very similar to the expression control of the mdtEF (yhiUV) multidrug efflux system and tolC by the EvgAS two-component system (40, 43, 44).

We previously cloned all of the gene clusters encoding putative and known drug transporters of E. coli and found that 20 genes encode transporters of some drugs and/or toxic compounds (42). The key to understanding how bacteria use these multiple transporters lies in analysis of the regulation of transporter expression. In this study, we found that the BaeSR system controls acrD and mdtABC in response to indole and probably other signals. Indole is a toxic compound that disrupts the bacterial envelope (49). Indole can also act as an extracellular signal in E. coli (1, 61), and it can be exported by the AcrEF multidrug transporter (22). Recently, our group found that the CpxAR two-component system also regulates the expression of acrD (15). Also, the CpxAR system plays an important role in response to envelope stresses (49). Thus, there is a strong relationship among envelope stresses, cell-to-cell signaling, and induction of multidrug transporters.

Some information about the regulation of multidrug transporters has been reported previously. Ma et al. reported that the expression of acrAB is induced by fatty acids, sodium chloride, and ethanol (29). Lomovskaya et al. reported that the emrAB drug exporter genes are induced by salicylic acid and 2,4-dinitrophenol (28). The EvgAS two-component system regulates the expression of mdtEF (yhiUV) and emrKY (43, 44). Also, Bock and Gross suggested that the EvgS sensor is connected to the oxidation status of the cell via a link to the ubiquinone pool (7). It has been reported that the expression of mdtEF (yhiUV) is controlled by RpoS (3, 58), an alternative sigma factor, which is needed for E. coli to survive stresses (25, 33, 57). RpoS is also required for the expression of many genes in stationary-phase cells. Recently, we found that histone-like protein H-NS represses the expression of acrEF and mdtEF (45). H-NS represses the expression of many genes that are expressed in the stationary phase. Also, the quorum-sensing regulator SdiA controls the expression of acrAB, acrD, and acrEF (50, 68). These data suggest that regulation of multidrug transporters has a strong relationship to (i) stress responses, (ii) the growth phase, and (iii) quorum sensing. Where are these regulatory networks and multidrug transporters needed in bacterial cells? Further investigation of the regulation of multidrug transporters in several natural environments, such as inside hosts, is needed in order to understand the biological significance of the regulatory networks for them and may provide further insights into the role of multidrug transporters in the physiology of the cell.