Abstract
The BasS–BasR two-component system is known as an iron- and zinc-sensing transcription regulator in Escherichia coli, but so far only a few genes have been identified to be under the direct control of phosphorylated BasR. Using Genomic SELEX (systematic evolution of ligands by exponential enrichment) screening, we have identified a total of at least 38 binding sites of phosphorylated BasR on the E. coli genome, and based on the BasR-binding sites, have predicted more than 20 novel targets of regulation. By DNase I footprint analysis for high-affinity BasR-binding sites, a direct repeat of a TTAAnnTT sequence was identified as the BasR box. Transcription regulation in vivo of the target genes was confirmed after Northern blot analysis of target gene mRNAs from both wild-type E. coli and an otherwise isogenic basR deletion mutant. The BasR regulon can be classified into three groups of genes: group 1 includes the genes for the formation and modification of membrane structure; group 2 includes genes for modulation of membrane functions; and group 3 includes genes for stress-response cell functions, including csgD, the master regulator of biofilm formation.
-
A supplementary table, showing the primers used in this study, is available with the online version of this paper.
-
Edited by: N. Le Brun
Introduction
A bacterial two-component system (TCS) is composed of a sensor kinase and response regulator pair (Stock et al., 1990; Parkinson, 1993; Egger et al., 1997; Mizuno, 1998; Hoch, 2000). The sensor kinase monitors a specific environmental signal or condition, and autophosphorylates. The phosphorylated sensor kinase then acts as a phosphodonor for autophosphorylation of the cognate response regulator, which regulates a set of stress-response genes (Hoch, 2000; Bourret, 2010; Gao & Stock, 2010), including the genes for metal exporting systems (Nies, 2003). Escherichia coli K-12 contains a total of approximately 300 transcription factors (Ishihama, 2009, 2010), of which about 10 % are TCSs, including 30 species of the sensor kinase and 34 species of the response regulator (Nagasawa et al., 1993; Yamamoto et al., 2005). The assimilation of essential elements (K, P, N and Mg) is under the control of the KdpDE (Asha & Gowrishankar, 1993), PhoRB (Wanner & Chang, 1987), GlnLG (or NtrBC) (Ninfa & Magasanik, 1986) and PhoQP (Kato et al., 1999) TCS pairs, respectively, while BasSR (Fe, Zn), CusSR (Cu), CpxAR (Cu) and ZraSR (Zn) TCSs are involved in the E. coli response to essential but toxic divalent metals (Yamamoto et al., 2005; Yamamoto & Ishihama, 2005a, b, 2006).
Microarray analysis indicated that the E. coli BasSR system controls, directly or indirectly, a set of genes that are associated with metal-response membrane modification, and genes for response to acidic and/or anaerobic growth conditions (Hagiwara et al., 2004; Lee et al., 2005). The whole set of target promoters under the direct control of BasSR remains unidentified, however. In contrast, the better-studied Salmonella PmrAB system, equivalent to E. coli BasSR, is known to regulate the genes for remodelling the composition and charge of outer-membrane lipopolysaccharides (LPS) (Froelich et al., 2006; Herrera et al., 2010), thereby modulating permeability properties (Groisman, 2001; Miller et al., 2005). Expression of pmrAB in Salmonella is induced by Fe (Wösten et al., 2000; Chamnongpol et al., 2002), but basSR in E. coli is induced by not only Fe but also Zn (Lee et al., 2005). Zn alone is sufficient to increase basRS transcription in chemostat cultures.
The aim of this study was to determine the whole set of target genes under the direct control of BasSR in E. coli K-12. To identify the binding sequences within the entire E. coli genome for phosphorylated BasR, we employed both SELEX (systematic evolution of ligands by exponential enrichment)-clos and SELEX-chip systems of the improved ‘Genomic SELEX’ screening system in vitro (Shimada et al., 2005; Ishihama, 2010). In the SELEX-clos system, transcription factor-bound DNA fragments are cloned and sequenced (Shimada et al., 2005). The SELEX-chip method was developed for quick mapping of transcription factor-bound DNA segments across the E. coli genome using a DNA tiling array (Teramoto et al., 2010). The Genomic SELEX screening system has been successfully employed for identification of the whole set of regulation targets against a number of E. coli transcription factors, such as RstA (Ogasawara et al., 2007a), PdhR (Ogasawara et al., 2007b), RutR (renamed from YdcC) (Shimada et al., 2007), NemR (renamed from YdhM) (Umezawa et al., 2008), Dan (renamed from YgiP) (Teramoto et al., 2010), LeuO (Shimada et al., 2009, 2011a), Cra (Shimada et al., 2005, 2011c) and CRP (Shimada et al., 2011b). BasR binding in vitro to the predicted BasSR targets was examined in detail by gel shift and DNase I footprinting assays, while regulation in vivo was examined by Northern blot analysis. Taking the results together we propose that the targets under the direct control of BasSR include the genes for determination of membrane structure, modulation of membrane functions and stress response.
Methods
Bacterial strains and plasmids.
E. coli DH5α was used for plasmid amplification. E. coli BL21 was used for BasR expression. E. coli BW25113 (W3110 lacIq rrnBT14 ΔlacZWJ16 hsdR514 ΔaraBADAH33 ΔrhaBADLD78) is the parental strain (Datsenko & Wanner, 2000) of an otherwise isogenic basR single-gene deletion mutant JW4074, a product of the Keio collection (Baba et al., 2006) and obtained from the E. coli Stock Center [National BioResource Project (NBRP), Mishima, Japan]. Cells were grown in LB medium at 37 °C under aeration with constant shaking at 140 r.p.m. Cell growth was monitored by measuring the turbidity at 600 nm.
Expression plasmid pBasR used for overexpression and purification of His-tagged BasR was constructed essentially according to the standard procedure of our laboratory (Yamamoto et al., 2005).
Purification of BasR protein.
The E. coli BL21 pBasR transformant was grown in LB broth in the presence of 50 µg ampicillin ml−1, and expression of BasR was induced in mid-exponential phase by adding 1 mM IPTG. After 3 h induction, cells were harvested and protein purification was carried out according to the standard procedure of our laboratory (Ogasawara et al., 2007a, b; Yamamoto et al., 2005). In brief, lysozyme-treated cells were sonicated in the presence of 100 mM PMSF. After centrifugation of cell lysate (30 ml) at 15 000 r.p.m. for 20 min at 4 °C, the resulting supernatant was mixed with 2 ml 50 % Ni-NTA agarose solution (Qiagen) and loaded onto a column. After washing with 10 ml lysis buffer (50 mM Tris/HCl, pH 8.0 at 4 °C, 100 mM NaCl), the column was washed with 10 ml washing buffer (50 mM Tris/HCl, pH 8.0 at 4 °C, 100 mM NaCl). Proteins were then eluted with 2 ml elution buffer (200 mM imidazole, 50 mM Tris/HCl, pH 8.0 at 4 °C, 100 mM NaCl), and dialysed against storage buffer (50 mM Tris/HCl, pH 7.6, 200 mM KCl, 10 mM MgCl2, 0.1 mM EDTA, 1 mM DTT, 50 %, v/v, glycerol). The BasR used throughout this study was more than 95 % pure as analysed by SDS-PAGE.
Genomic SELEX screening of BasR-binding sequences.
Genomic SELEX screening of BasR-binding sequences was carried out using the improved procedure (Shimada et al., 2005). A mixture of DNA fragments of the E. coli K-12 W3110 genome was prepared after sonication of purified genome DNA, and cloned into a multi-copy plasmid, pBR322. In each SELEX screening, the DNA mixture was regenerated by PCR. For SELEX screening, 5 pmol of the mixture of DNA fragments and 10 pmol His-tagged BasR were mixed in a binding buffer (10 mM Tris/HCl, pH 7.8 at 4 °C, 3 mM magnesium acetate, 150 mM NaCl, 1.25 mg BSA ml−1) and incubated for 30 min at 37 °C. The DNA/transcription factor mixture was applied to a Ni-NTA column, and after washing out unbound DNA with the binding buffer containing 10 mM imidazole, DNA–protein complexes were eluted with an elution buffer containing 200 mM imidazole. The sequences of DNA fragments recovered from the complexes were determined by either the SELEX-clos or the SELEX-chip method. In SELEX-clos, the DNA fragments were recovered from the complexes, PCR-amplified, cloned into a sequencing vector and then subjected to DNA sequencing. For SELEX-chip, PCR-amplified products of the isolated DNA–protein complexes and original DNA library were labelled with Cy3 and Cy5, and then combined. The fluorescently labelled DNA mixtures were hybridized to a DNA microarray consisting of 43 450 different 60-base DNA probes, which were designed to cover the entire E. coli genome at 105 bp intervals (Oxford Gene Technology). The fluorescence intensity of the test sample at each probe was normalized to that of the corresponding peak of the original library. After normalization of each pattern, the Cy5/Cy3 ratio was measured and plotted along the E. coli genome.
Gel mobility shift assay.
Gel shift assays were performed as described previously (Ogasawara et al., 2007a, b). In brief, probes with BasR-binding sequences identified by Genomic SELEX were generated by PCR amplification using a pair of primers, 5′ FITC-labelled T7-F primer and T7-R primer (for sequences, see Table S1 available with the online version of this paper), the genomic SELEX plasmids containing the respective BasR recognition sequences as templates, and Ex Taq DNA polymerase (Takara). PCR products with FITC at their termini were purified by PAGE. For gel shift assays, 0.3 pmol of each FITC-labelled probe was incubated at 37 °C for 30 min with various amounts of BasR in 12 µl gel shift buffer consisting of 10 mM Tris/HCl, pH 7.8 at 4 °C, 150 mM NaCl, 3 mM magnesium acetate. After addition of a DNA dye solution, the mixture was directly subjected to 6 % PAGE. Fluorescently labelled DNA in gels was detected using the Pharos FX Plus system (Bio-Rad).
Northern blot analysis.
Total RNAs were extracted from exponentially growing E. coli cells (OD600 0.5) by the hot phenol method (Aiba et al., 1981). RNA purity was checked by electrophoresis on a 2 % agarose gel in the presence of formaldehyde followed by staining with ethidium bromide. Fluorescently labelled probes were prepared by PCR amplification using W3110 genomic DNA (50 ng) as template, DIG-11-dUTP (Roche) and dNTP as substrates, gene-specific forward and reverse primers (for primer sequences, see Table S1), and Ex Taq DNA polymerase (Takara). Total RNAs (4 µg) were incubated in formaldehyde-MOPS gel-loading buffer for 10 min at 65 °C for denaturation, subjected to electrophoresis on a formaldehyde-containing 2 % agarose gel, and then transferred to a nylon membrane (Roche). Hybridization was performed with the DIG Easy Hyb system (Roche) at 50 °C overnight with a DIG-labelled probe. For detection of DIG-labelled probe, the membrane was treated with anti-DIG–alkaline phosphatase Fab fragments and CDP-Star (Roche), and the image was scanned with an LAS-4000 IR Multi Color imager (Fuji Film).
Results
Searching for the regulation targets of BasR by Genomic SELEX screening: SELEX-clos
To begin the process of identifying the BasR regulon, we tried to identify the whole set of target promoters under the direct control of BasR using the improved Genomic SELEX screening system (Shimada et al., 2005), in which purified His-tagged BasR was mixed with a collection of E. coli genome fragments of 200–300 bp in length in the presence of acetyl phosphate, which phosphorylates BasR in the absence of BasS kinase. Phosphorylated BasR-bound DNA fragments were affinity-purified for the identification of BasR recognition sequences. The original substrate mixture of genomic DNA fragments used in this genomic SELEX screening formed smear bands on PAGE, but after two cycles of genomic SELEX, DNA fragments with high affinity to phosphorylated BasR were enriched, forming sharper bands on PAGE gels (data not shown). After three cycles of Genomic SELEX, phosphorylated BasR-bound DNA fragments formed several bands on PAGE, indicating further enrichment of some DNA fragments with high affinity to BasR. The DNA fragments were cloned into a multi-copy plasmid and subjected to DNA sequencing. As an initial attempt, a total of 84 independent clones were sequenced. Eleven sequences from different regions of the E. coli genome were identified in more than two clones (54 clones in total), while each of other 30 clones included one unique sequence (Table 1).
A total of 84 independent clones were isolated after SELEX-clos analysis of DNA segments binding phosphorylated BasR, and the DNA fragment in each SELEX clone was sequenced. A total of nine BasR-binding sites were identified in more than two clones (54 clones in total), while another 29 clones contained a single different sequence. The BasR-binding sites in each group are aligned along the genome position. Except for the last nine BasR-binding sites, which are located within the ORFs of the genes shown in the BasR site column, all other BasR-binding sites were located within intergenic spacer regions. The BasR-binding sites marked ++ showed high peaks in the SELEX-chip pattern, while those marked + gave low but significant peaks (see Fig. 1, Table 2). The transcription direction of flanking genes is shown by the symbols next to each gene name. Possible regulation target genes are shown in bold type.
The most abundant clone (21 clones in total) included a sequence from the intergenic spacer upstream of eptA and downstream of adiC. Based on the orientation of transcription, we predicted the eptA gene, encoding an inner-membrane protein with binding activity to ZipA (cell division machinery component), to be a regulation target of BasR. Next, two abundant clones contained intergenic spacer sequences of divergently transcribed operons: the putA (Pro dehydrogenase-pro operon regulator fusion) and putP (Pro : Na symporter) operons (six clones); and the ais [cell–cell communication signal autoinducer (AI)-inducible gene] and arnB (LPS modification) operons (five clones). The SELEX-clos analysis indicated that the regulation targets of BasR are involved in various cell functions, including the formation and modification of cell membranes, and also include stress-response proteins. Identification of the whole set of regulation targets is, however, difficult by SELEX-clos, because a large number of SELEX clones must be isolated and analysed to identify DNA sequences with low affinity to BasR.
Searching for the regulation targets of BasR by Genomic SELEX screening: SELEX-chip
As a short-cut approach to identify the whole set of targets under the direct control of BasR, we then subjected the mixture of Genomic SELEX fragments to the DNA chip analysis using an E. coli tiling array. The BasR-bound SELEX DNA fragments were labelled with Cy5, while the original DNA library was labelled with Cy3. The mixtures were hybridized with the DNA tiling microarray (Oxford Gene Technology) and the fluorescence intensities bound on each probe were measured. For identification of BasR-binding sites, the Cy5/Cy3 ratio was plotted along the corresponding position on the E. coli genome (Fig. 1). By setting the cut-off level at 50 for at least two adjacent probes, a total of 38 BasR-binding sites were identified (Table 2), but the number of high-level peaks (or high-affinity sites) decreased to 13 by setting a cut-off level of 100 (++ in column A of Table 2). Among the total of 11 high-affinity sites identified by SELEX-clos (see Table 1), up to 10 sites were included as high-affinity sites for BasR binding (marked ++ in Table 1, column BasR, and in Table 2, column B), whereas among the collection of 30 single SELEX-clos clones, 13 were included in this SELEX-chip list (marked + in Table 1, column BasR, and Table 2, column B). The high-level agreement of regulation target regions identified by SELEX-clos and SELEX-chip experiments indicates the high-level accuracy of Genomic SELEX screening. Noteworthy is the good agreement between the number of Genomic SELEX clones with high-affinity BasR binding (Table 2, column B; see also Table 1) and the high-level peaks in the SELEX-chip pattern (Fig. 1, Table 2, column A). Taken together we predicted that the promoters located near the BasR-binding sites identified in both SELEX-clos and SELEX-chip were the regulation targets of phosphorylated BasR.
Identification of BasR-binding sites on the E. coli genome. A Genomic SELEX search of phosphorylated BasR-binding sequences was performed using the standard procedure (Shimada et al., 2005). DNA fragments isolated by SELEX screening were subjected to mapping on the genome using a tiling DNA microarray (Shimada et al., 2011a, b; Teramoto et al., 2010). The regulation target genes were predicted from the location of BasR-binding peaks. When the BasR-binding sequences were located within intergenic spacers, regulation targets were predicted based on the direction of transcription, either one direction (single gene name) or both directions (two gene names connected with a dash) (for details, see Table 2). The genes associated with some of the high-level peaks are indicated. The genes identified by SELEX-clos are shown in black boxes, while the genes identified by SELEX-chip alone are shown in white boxes.
The binding sites of phosphorylated BasR on the E. coli genome were identified by SELEX-chip analysis. BasR-binding sites are shown in the BasR-binding site column (blank represents intergenic spacer; the gene symbol represents a BasR-binding site within an ORF), and the genes located on the left- and right-hand sides are shown in the Right and Left columns. By setting the cut-off level at 100 (see Fig. 1), a total of 13 peaks were identified (marked ++ in column A), while this increased to 38 by setting the cut-off level at 50 (marked + for targets between 50 and 100). Column B shows the BasR-binding sites identified by SELEX-clos (the target sequences identified in multiple clones are shown by ++, while those identified only in a single clone are shown by +). Column C, BasR target identified earlier (Froelich et al., 2006). The transcription direction of flanking genes is shown by the symbols next to each gene name. Possible regulation target genes are shown in bold type.
The predicted regulation targets of BasR include three groups of genes: group 1 includes a set of genes for the synthesis and modulation of membrane structure, such as arnB (LPS modification) and rfaB (LPS galactosyltransferase); group 2 includes genes for membrane functions such as mlaF (ABC transporter), putP (Pro : Na symporter) and potF (putrescine transporter); and group 3 includes a set of regulators of stress-response functions, including csgD (the master regulator of biofilm formation), elbA (iraM) (RpoS control RNA), csrB (regulatory sRNA), cspI (cold-shock protein) and inaA (acid pH-inducible protein) (Fig. 2, Tables 1 and 2).
Classification of the functions of BasR-regulated genes. Most of the predicted targets under the direct control of phosphorylated BasR (see Tables 1 and 2) can be classified into three groups: group 1, genes for formation and modification of membrane structure; group 2, genes for membrane-associated functions; and group 3, genes for stress-response gene regulation and stress-response metabolism.
Identification of BasR-binding site on the predicted target promoters
To confirm the prediction of the regulation targets by phosphorylated BasR, we first performed a gel shift assay of BasR–DNA complex formation. Fluorescently labelled SELEX fragments carrying the newly identified BasR-binding sites were mixed with increasing concentrations of BasR in the presence of acetyl phosphate, and after incubation were immediately subjected to PAGE. Fig. 3 shows the gel shift pattern for representative probes carrying the sequences from the predicted BasR targets that were identified by the SELEX-clos and SELEX-chip analyses. For all the DNA probes examined, free DNA disappeared upon addition of BasR. One to four bands of DNA–BasR complexes were detected for all the probes, even though some DNA fragments did not give clear gel bands, but instead showed smeared bands, implying co-operative binding of phosphorylated BasR to the DNA probes.
Binding activity in vitro of BasR to target DNA probes. The DNA-binding activity of phosphorylated BasR was examined by gel shift assay. Probes used are SELEX fragments identified by SELEX-clos and/or SELEX-chip (see Fig. 1, Tables 1 and 2). (a) The amount of BasR protein used in lanes 1, 2 and 3 was 0, 16 and 48 pmol, respectively; (b) The amount of BasR protein used in lanes 1, 2, 3, 4 and 5 was 0, 6, 12, 24 and 48 pmol, respectively. Probes with high affinity to BasR, such as eptA, formed multiple bands, probably due to co-operative BasR binding. Even an identical probe, such as putA, gave a smear or detectable bands depending on PAGE conditions that were not identified.
For identification of BasR-binding sequences, we then performed a DNase I footprinting assay of phosphorylated BasR-bound DNA. The region protected by phosphorylated BasR from DNase I digestion ranged from 32 to 40 bp in length. Fig. 4 shows typical patterns of DNase I footprinting, showing 32 bp-long DNA protection for eptA (Fig. 4a), 36 bp protection for yrbL (Fig. 4b), and 40 bp protection for csgD (Fig. 4c). All these sequences contain two direct repeats of a 9 bp CTTAAGGTT sequence (Fig. 4d).
Identification of BasR-binding sequences. DNase I footprinting assay of BasR-binding sequences in the eptA (a), yrbL (b) and csgD (c) promoters was performed under the experimental conditions described in Methods using increasing concentrations of BasR (lanes: 1, 0; 2, 8.75; 3, 17.5; 4, 35; 5, 70 pmol). A common BasR-binding sequence was detected in these three probes (d).
Location of BasR box sequences on target promoters
Using a direct repeat of the consensus 9 bp CTTAAGGTT sequence, the location of the BasR box sequence was searched for within the putative regulation target promoters or genes identified by Genomic SELEX screening. The BasR box-like sequence was detected for the major peak regions identified by SELEX-chip (Fig. 5a). The WebLogo motif that was derived using these conserved sequences as probes indicated 8 bp TTAAnnTT to be the consensus sequence of the BasR box (Fig. 5a). In all these cases, the BasR box-like sequence was detected at various positions from the respective initiation codon, ranging from −51 to −32 (putA) to −490 to −471 (tomB) (Fig. 5b). The orientation of the BasR box-like sequence relative to the gene organization (or the transcription direction) was the same for eptA, arnB, fimB and cydA, but in the opposite direction for putA, yrbL, csgD, cspI and tomB. A triplet repeat of the BasR box exists in some cases, such as eptA, yrbL, arnB and cspI.
Location of BasR-binding sites on target promoters. (a) Using the consensus BasR-binding sequence identified by DNase footprint analysis (see Fig. 4d) as a probe, the locations of BasR-binding sites were searched for within the major target promoters identified by SELEX-chip (see Fig. 1). All the targets were found to contain a sequence similar to a direct repeat of the 9 bp-long CTTAAGGTT. The consensus sequence of the BasR box was then identified after motif analysis using WebLogo (). (b) The locations of BasR box-like sequences are shown on each of the target genes. A single BasR box is shown by a single red arrow, which also indicates the direction of the BasR box relative to the transcription of the respective target gene. In most cases, a direct repeat of the BasR box was detected, but a triplet of the BasR box was detected in some cases.
BasSR-dependent regulation in vivo of the predicted targets
To determine BasSR-dependent regulation in vivo of the target genes isolated by Genomic SELEX screening and identified in vitro by gel shift and DNase I footprinting assays, we then performed Northern blot analysis of mRNA in vivo in the presence of increasing concentrations of Fe(II). Growth of wild-type E. coli K-12 was significantly retarded in the continuous presence of Fe(II) above 1.0 mM (Fig. 6a). We isolated total RNA from cells grown in the presence of 2.0 mM Fe(II), and immediately subjected to Northern blot analysis using a set of probes specific for the predicted BasSR-regulated genes (Fig. 6b). In the presence of Fe(II), the level of mRNA increased for some representative BasSR targets, eptA, arnB and csgD, indicating that Fe(II) activates transcription of these targets (Fig. 6b, WT lanes). In the case of the PhoQP-dependent yrbL, which has an unidentified function, however, the mRNA level decreased in the presence of Fe(II) (Fig. 6b, WT lanes). Transcription activation of eptA, arnB and csgD was not observed with the basR mutant, indicating the involvement of BasR as an activator of these genes (Fig. 6b, ΔbasR lanes). On the other hand, the Fe(II)-dependent decrease in yrbL mRNA was not observed with the basR mutant (Fig. 6b, ΔbasR lane). Taken together, these observations support the notion that BasSR is involved in Fe(II)-dependent regulation, at least for the target genes examined, and that BasR is a dual regulator, acting as an activator or a repressor depending on the target promoter.
Northern blot analysis of BasR-dependent gene mRNA. (a) Growth of E. coli BW25113 and the basR mutant JW4074 was measured in the absence and presence of various concentrations of FeSO4. Cell growth was retarded at a Fe(II) level higher than 1 mM. (b) Total RNA was prepared from both wild-type BW25113 and the basR mutant JW4074, both grown in LB in the absence (−) or presence (+) of 0.8 mM FeSO4, and subjected to Northern blot analysis using the promoter-specific probes indicated (for probe sequences, see Table S1).
Discussion
Regulation targets of BasSR: group 1 genes for the modification of membrane structure
In this study, we identified, in addition to the armBCADTEF operon (LPS modification system) identified earlier, a number of genes related to membrane synthesis and modification as novel targets under the direct control of BasSR, including rfaB (LPS galactosyltransferase) and ynaB (diacylglycerol phosphotransferase) (see Fig. 2, Tables 1 and 2). Genome-wide transcriptome analysis of a E. coli basS mutant (Hagiwara et al., 2004) and of wild-type E. coli in the presence of excess Zn(II) (Lee et al., 2005) indicates that the arnBCADTEF operon is under the control of the BasSR TCS (Hagiwara et al., 2004; Lee et al., 2005). An E. coli mutant carrying a constitutive basR gene expresses constitutively the genes for membrane modification, such as the arnB and eptA operons, in the presence of Fe and Zn (Froelich et al., 2006). The dgkA gene, encoding glycerol-3-phosphate acyltransferase, an enzyme involved in synthesis of membrane phospholipid, has been found to be under the direct control of BasR (Wahl et al., 2011), linking the function of dgkA in phospholipid recycling to LPS modification.
Regulation targets of BasSR: group 2 genes for modification of membrane functions
In agreement with the involvement of BasR in the remodelling of membrane structure, genes for membrane-associated functions were also found to be under the direct control of BasR, including csgB (curli fimbriae curlin subunit), eptA (ZipA-binding inner-membrane protein), fimB (fim promoter switch recombinase), putP (Pro : Na symporter), ompC (outer-membrane porin C), phoE (outer-membrane phosphoporin E), potF (putrescine transporter), mlaF (ABC transporter), yrbG (putative metal transporter) and yfaL (adhesin-like protein).
Remodelling of LPS confers resistance towards cationic peptides and metal ions (Chamnongpol et al., 2002; Miller et al., 2005; Wahl et al., 2011). Inside the human host, LPS (endotoxin) remodelling plays a key role in bacterial fitness and virulence. LPS modification not only confers resistance to endogenous antibacterial cationic peptides but also helps bacteria evade the innate immune system (Bader et al., 2005; Miller et al., 2005). A basR mutant confers moderate polymyxin resistance, while simultaneously sensitizing cells to the anionic detergent deoxycholic acid (Nummila et al., 1995; Herrera et al., 2010). The environmental changes in the mammalian intestinal tract, such as exposure to antimicrobial cationic peptides and high concentrations of anionic bile detergents, induce alterations of the outer membrane.
Regulation targets of BasSR: group 3 genes for modulation of gene regulation and metabolism under stressful conditions
A number of stress-responsive regulatory proteins were also found to be directly regulated by BasR, including the YafQ–DinJ toxin–antitoxin pair, TomB (Hha toxicity modulator), ElbA (RpoS level control factor), Eco (serine protease inhibitor), EptA (ZipA-binding septation protein) and FimB (fim promoter switch recombinase). Also noteworthy is that the anti-silencer LeuO (also a global regulator of membrane function) (Shimada et al., 2011a) and CsgD (the master regulator of biofilm formation) (Ogasawara et al., 2010a, b, 2011) were regulated by BasR.
The csgD promoter is the most complex promoter so far examined in E. coli, and has more than 10 transcription factors for regulation (Fig. 7). Each of these factors plays a role as a sensor of a different environmental factor or condition that influences biofilm formation. Here we identified the involvement of BasR as an activator in the regulation of the csgD promoter, which indicates that iron and/or zinc influences biofilm formation. The site of BasR binding on the csgD promoter overlaps with those of IHF, CpxR and H-NS (Fig. 7), implying that the silencing effect of H-NS might be antagonized by the combination of three regulators, BasR, IHF and CpxR.
Location of the BasR-binding site on the csgD promoter. The BasR-binding site on the csgD promoter was determined by DNase I footprinting assay (see Fig. 4), and is mapped on the csgD promoter, on which more than 10 transcription factors associate at various positions, as determined by Ogasawara et al. (,2010a, b, 2011).
Genome-wide transcriptome analysis indicated that BasSR-dependent gene products are involved in survival under acid and/or anaerobic conditions (Hagiwara et al., 2004; Lee et al., 2005). Here we found that some genes encoding enzymes for energy metabolism under acid- and/or anaerobic conditions were under the direct control of BasSR (see Tables 1 and 2): (1) both aceF and aceE are members of the pyruvate dehydrogenase operon, which is involved in acetate metabolism (Quail et al., 1994); (2) the hyaA and hyaB genes encode the small and large subunits of hydrogenase 1, respectively (Menon et al., 1991); and (3) the cydAB operon encodes cytochrome bd-I terminal oxidase (Green et al., 1984), which is expressed under oxygen-limited conditions and needed for growth under anaerobic conditions (Cotter et al., 1990). The expression levels of these genes were increased under acidic and/or anaerobic growth conditions.
Location of the BasR box
Marked differences were observed in the location and orientation of the BasR box relative to the initiation codon of target genes (see Fig. 5b). From Northern blot analysis (see Fig. 6), the targets carrying the BasR-binding site up to at least 250 bp from the initiation codon were found to be under the direct control of BasR. The orientation of the BasR box relative to transcription organization differed between the target genes (see Fig. 5b), but so far a tight correlation has not been detected between the orientation of the BasR box and the mode of transcription regulation by BasR.
BasR is a transcription factor with dual activities, functioning for different binding sites as a repressor or an activator depending on the transcription organization. There is an approximate correlation for BasR regulation: BasR binding near the initiation codon represses transcription, while BasR binding far from the initiation codon activates the target gene (see Figs 5b and 6). However, conclusive evidence will only be available when the transcription initiation sites for these BasR target genes have been identified.
Factors affecting BasSR functions
Iron is essential for biological processes but toxic in excess. E. coli has the well-characterized Fur-mediated signal transduction pathway, which is activated during iron limitation and is responsible for iron uptake (Hantke, 2001). This Fur-mediated pathway responds to iron in the Fe(II) form in the cytosol and participates in the adaptation to oxidative stress by interacting with the OxyR and SoxRS regulatory systems. In addition, BasSR is involved in the response to toxic iron, as first indicated by Salmonella PrmAB. BasSR-dependent expression of the arnBCADTEF operon has been found to increase during growth with elevated Fe(II) or Fe(III) (Hagiwara et al., 2004; Lee et al., 2005). Deletion of basSR results in acid sensitivity during growth at elevated iron concentrations (Hagiwara et al., 2004). Deletion of basR prevents the Fe(II)- and Fe(III)-mediated induction of eptA, arnB and yibD, and results in sensitivity to the cationic agent polymyxin B (Froelich et al., 2006). Zn(II) slightly inhibits growth of a basR mutant (Lee et al., 2005). Some basSR mutants show a phenotype of mild acid sensitivity in the presence of high concentrations of external iron. This scenario is consistent with that of PmrBA in Salmonella enterica (Wösten et al., 2000; Chamnongpol et al., 2002). The Fe(III) form of iron is a toxic substance that acts on an extracytoplasmic target of the cell surface in a manner independent of intracellular oxidative stress (Chamnongpol et al., 2002).
In addition to iron, the E. coli BasSR TCS responds to zinc (Lee et al., 2005). Among a set of genes upregulated in the presence of Zn(II) as detected by genome-wide transcriptome analysis, the basSR TCS was identified to be the Zn-sensing regulator (Lee et al., 2005). In Salmonella, the PmrBA TCS has a linkage with the Mg(II)-responsive PhoQ–PhoP TCS in concert with the intermediate factor PmrD (Groisman, 2001). The PhoQ sensor recognizes antimicrobial peptides (Bader et al., 2005), which in turn leads to BasRS activation and LPS remodelling. In E. coli, however, the PhoQP system is not coupled to BasRS (Winfield & Groisman, 2004). The PhoQP, BasSR and RcsCB TCSs form a regulatory network, which is physiologically meaningful in the sense that this network is implicated in the coordinated synthesis and modification of cell-surface polysaccharides in response to the availability of external divalent cations (Hagiwara et al., 2004).
Acknowledgements
We thank Tomohiro Shimada and Jun Teramoto for discussions, and Ayako Kori and Kayoko Yamada for preparation of proteins and technical support. The bacterial strains were obtained from NBRP E. coli strain stock (National Institute of Genetics, Mishima, Japan). This work was supported by grants-in-aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan (17076016, 8310133 and 21241047), and Nano-Biology Project fund from Micro-Nano Technology Research Center of Hosei University.