The Escherichia coli NarL protein is a global gene regulatory factor that activates transcription at many target promoters in response to nitrate and nitrite ions. Although most NarL-dependent promoters are also co-dependent on a second transcription factor, FNR protein, two targets, the yeaR and ogt promoters, are activated by NarL alone with no involvement of FNR. Biochemical and genetic studies presented here show that activation of the yeaR promoter is dependent on the binding of NarL to a single target centred at position −43.5, whereas activation at the ogt promoter requires NarL binding to tandem DNA targets centred at position −45.5 and −78.5. NarL-dependent activation at both the yeaR and ogt promoters is decreased in rich medium and this depends on Fis, a nucleoid-associated protein. DNase I footprinting studies identified Fis-binding sites that overlap the yeaR promoter NarL site at position −43.5, and the ogt promoter NarL site at position −78.5, and suggest that Fis represses both promoters by displacing NarL. The ogt gene encodes an O6-alkylguanine DNA alkyltransferase and, hence, this is the first report of expression of a DNA repair function being controlled by nitrate ions.
- DNA repair
- reactive nitrogen species
Transcription in Escherichia coli is orchestrated by a complex network of protein factors that regulate promoter activity in response to different external signals. Many factors interact at multiple promoters, and transcription initiation at many promoters is regulated by two or more factors [1,2]. The E. coli NarL protein is a typical response-regulator family transcription factor that controls responses to nitrate and nitrite ions in the environment. Its activity is induced by phosphorylation by two inner membrane-bound sensor kinases, NarX and NarQ, which are activated by nitrate and nitrite ions in the periplasmic space [3–7]. NarL controls the expression of many proteins involved in anaerobic respiration and, at most of its known promoter targets, it co-regulates transcription together with another global transcription factor, protein FNR, the ‘master’ regulator of anaerobic adaptation [2,8–10]. However, a study of the NarL regulon based on transcriptome analysis  identified some genes that appeared to be activated by NarL without FNR, and a thorough genetic analysis of the yeaR-yoaG operon regulatory region by Valley Stewart and co-workers proved that NarL activates the yeaR promoter independently of FNR . In the present study we describe the first biochemical study of the E. coli yeaR promoter and we show that NarL-dependent promoter activation can be reproduced in vitro. During our studies, we noticed that yeaR promoter activity is sharply reduced during bacterial growth in rich media and we present evidence that this is due to the action of Fis, a well-characterized DNA-binding protein that is thought to play a key role in compaction of the E. coli chromosome . In the second part of the present study, we selected another E. coli promoter, the ogt promoter, that transcriptome analysis  had predicted to be activated by NarL alone, and we have used biochemical and genetic analysis to compare its expression and organization with the yeaR promoter.
Strains, plasmids, promoter fragments and oligodeoxynucleotide primers
E. coli K-12 strains, plasmids and primers used in the present study are listed in Table 1. The JCB387 Δfnr derivative was made by P1 transduction of a Δfnr (cmR) mutation that had been constructed in strain MG1655 by Constantinidou et al. . Standard techniques for recombinant DNA manipulation were used throughout.
Promoter fragments were amplified by PCR using primers that introduced flanking EcoRI and HindIII sites. By convention, promoter sequences are numbered with respect to the transcription start point (+1) and with upstream and downstream locations denoted by ‘–’ and ‘+’ prefixes respectively. The E. coli K-12 yeaR-yoaG promoter fragment, yeaR100, was amplified from genomic DNA using primers pyeaRfw and pyeaRrev, and consisted of nucleotide sequences from positions −294 to +96. The ogt100 promoter fragment was amplified using primers pogtfw and pogtrev and consists of nucleotide sequences from position −269 to +51. For the generation of DNA fragments for EMSA (electrophoretic mobility-shift assay) and footprinting, the yeaR100 and ogt100 fragments were cloned into plasmid pSR, encoding resistance to ampicillin . To assay promoter activities, fragments were cloned into the low-copy number lac expression vector plasmid, pRW50, encoding resistance to tetracycline .
The ogt102 and ogt104 promoter fragments, in which the NarL I or NarL II binding sites were disrupted, were constructed using two-step megaprimer PCR. Promoter DNA was amplified using primer pSRdown and either primer ogt102 or primer ogt104 with pSR/ogt100 as a template. The purified PCR products were used in a second round of PCR with primer D5431 and pSR/ogt100 as a template. Purified products were restricted with EcoRI and HindIII and subcloned into pRW50.
β-Galactosidase activity assays
Plasmids containing either the yeaR::lacZ or ogt::lacZ promoter fusions were transformed into relevant strains. β-Galactosidase activities were measured using the Miller  protocol. Cells were grown either in rich medium: Lennox broth [2% (w/v) peptone (Oxoid), 1% (w/v) yeast extract (Oxoid) and 170 mM NaCl] plus 0.4% glucose, or in MM (minimal medium): minimal salts with 0.4% glycerol, 5% Lennox broth, 20 mM fumarate and 20 mM trimethylamine N oxide [11,17]. Where indicated, sodium nitrate or sodium nitrite was added to cultures to final concentrations of 20 and 2.5 mM respectively. Cultures were grown in static sealed test tubes to an attenuance (D) between 0.4 and 0.6 (at 650 nm). Activities were calculated as nmol of ONPG (o-nitrophenyl β-D-galactopyranoside) hydrolysed/min per mg of dry cell mass and represent the average of three independent experiments.
Preparation and purification of a fusion of MBP (maltose-binding protein) to NarL (MBP–NarL) was as described by Li et al. . In all experiments, the mature native NarL protein was used after the MBP moiety had been cleaved from MBP–NarL using protease factor Xa (New England Biolabs) . Purified Fis protein, which had been prepared by the method of Osuna et al. , was donated by Professor Rick Gourse (Department of Bacteriology, University of Wisconsin-Madison, Madison, WI, U.S.A.), and purified E. coli RNA polymerase holoenzyme was purchased from Epicentre Technologies (Madison, WI, U.S.A.). Purified DA154 FNR was prepared as described by Wing et al. .
EMSAs using purified NarL were as described by Browning et al. . EcoRI–HindIII digested yeaR100 or ogt100 promoter fragments were end-labelled with [γ-32P]ATP, and approx. 0.5 ng of each fragment was incubated with different amounts of NarL. The reaction buffer contained 10 mM potassium phosphate (pH 7.5), 100 mM potassium glutamate, 1 mM EDTA, 50 μM DTT (dithiothreitol), 5% glycerol, 25 μg·ml−1 herring sperm DNA and 50 mM acetyl phosphate (final reaction volume was 10 μl). After incubation at 37 °C for 30 min, samples were run in 0.25× TBE (Tris/borate/EDTA; 1×TBE=45 mM Tris/borate and 1 mM EDTA) on a 6% polyacrylamide gel (12 V·cm−1) containing 2% glycerol and analysed using a Bio-Rad Molecular Imager FX and Quantity One software (Bio-Rad). EMSA experiments with DA154 FNR were run as described by Browning et al. .
DNase I and potassium permanganate footprinting experiments
The protocols of Savery et al.  were used for footprinting experiments with 32P-end-labelled AatII–HindIII fragments carrying the yeaR or ogt promoters. Each reaction (20 μl) contained a final concentration of 1.35 nM template DNA. The buffer composition was 20 mM Hepes (pH 8.0), 50 mM potassium glutamate, 5 mM MgCl2, 1 mM DTT, 500 μg·ml−1 BSA and 25 μg·ml−1 herring sperm DNA. NarL protein was pre-incubated with 50 mM acetyl phosphate for 45 min at 37 °C, to allow for NarL phosphorylation and DNA binding . In the competition experiments shown in Figures 2(B) and 5(B), NarL and Fis were premixed prior to incubation with the labelled DNA fragment.
For potassium permanganate footprint experiments, herring sperm DNA was omitted and E. coli RNA polymerase holoenzyme was used at a final concentration of 50 nM. Samples were analysed by electrophoresis in denaturing gels that were calibrated with the products of Maxam–Gilbert ‘G+A’ sequencing reactions of the labelled fragment. Data were captured using a Bio-Rad Molecular Imager FX and Quantity One software (Bio-Rad).
Analysis of the yeaR promoter
The starting point of the present study was the yeaR100 DNA fragment that carries the DNA sequence from position −294 to position +96 with respect to the yeaR promoter transcript start site. NarL recognizes seven base-pair sequence elements (the NarL heptamer) and, at many promoters, two of these elements are organized as an inverted repeat, separated by two base-pairs (the 7-2-7 site) . Figure 1(A) shows the DNA sequence of the yeaR promoter region, highlighting the 7-2-7 NarL target centred at position −43.5 that was identified by Lin et al. . Using mutational analysis, Lin et al.  showed that NarL binding at this target activates the yeaR promoter. To confirm this, the yeaR100 fragment was cloned into pRW50, a low-copy-number lacZ expression vector to generate a yeaR::lac fusion, and Figure 1(B) shows measurements of the expression of this fusion. In the JCB387 narL+ strain, expression is strongly induced by both nitrate and nitrite ions and this induction is not found in the JCB3883 narL background. To measure binding of NarL to the yeaR100 fragment directly, we used EMSA and DNase I footprinting with purified NarL. The EMSA experiment, illustrated in Figure 1(C), shows that NarL causes a single clear shifted band, consistent with NarL binding a single target site. The DNase I experiment, illustrated in Figure 1(D), shows that this NarL-binding site corresponds to the 7-2-7 site centred at position −43.5.
A role for Fis at the yeaR promoter
While studying the expression of the yeaR::lac fusion, we noticed that nitrate-dependent induction was greater in MM than in more nutrient-rich media. In our previous work, we had discovered that the nucleoid-associated protein, Fis, plays a direct role in the repression of several promoters in response to nutrient abundance and rapid growth [22,25,26]. Thus, to investigate a role of Fis at the yeaR promoter, expression of the yeaR::lac fusion was compared in strain JCB387 and a fis mutant derivative, JCB3871. Results shown in Figure 2(A) show that nitrate-induced expression is sharply suppressed during growth in Lennox broth supplemented with glucose, and that this suppression is greatly reduced in the JCB3871 fis strain. DNase I footprinting with purified Fis protein was then used to investigate whether Fis interacts directly with the yeaR promoter. The results (Figure 2B) reveal that Fis binds to a target (Fis I) that overlaps the upstream end of the DNA site for NarL and, at higher Fis concentrations, a second site (Fis II), which overlaps the downstream end of the DNA site for NarL, is occupied. Furthermore, footprinting experiments with both Fis and NarL (Figures 2B and 2C) show that Fis prevents NarL binding to its target.
To investigate the effects of NarL and Fis on transcription initiation at the yeaR promoter, we used potassium permanganate footprinting, which can be exploited to monitor DNA unwinding during open complex formation . Results in Figure 3(A) show that incubation of the yeaR promoter fragment with holo RNA polymerase results in permanganate-induced DNA cleavage, indicating unwinding just upstream of the transcription start point. As expected, pre-incubation with NarL increases this cleavage, most likely as NarL promotes the formation of a transcriptionally competent complex at the yeaR promoter. The experiment illustrated in Figure 3(B) shows that this unwinding is totally suppressed by Fis, both in the presence and absence of NarL.
Analysis of the ogt promoter
A previous transcriptome analysis had predicted that expression of the ogt gene is also activated by NarL independently of FNR . To investigate whether the ogt and yeaR regulatory regions share common features, we generated the ogt100 DNA fragment that carries the DNA sequence from position −269 to position +51 with respect to the ogt transcript start site, and cloned it into pRW50, to give an ogt::lac fusion. Figure 4(A) shows the DNA sequence around the ogt promoter transcript start point and Figure 4(B) illustrates measurements of the expression of the fusion in the JCB387 narL+ and JCB3883 narL strains. The results show that expression is strongly induced by nitrate ions and that this induction is not observed in the narL background. In contrast with the yeaR promoter, induction by nitrite is minimal.
An EMSA experiment showed that NarL binds to at least two sites at the ogt promoter (Figure 4C) and DNase I footprinting (Figure 4D) revealed two binding targets for NarL that appear to correspond to two 7-2-7 sites, centred at position −78.5 (NarL I) and −45.5 (NarL II) (highlighted in Figure 4A). To investigate the function of these two targets, derivatives of the ogt100 fragment, ogt102 and ogt104, were constructed in which the NarL I or NarL II sites respectively were inactivated by mutations (shown in Figure 4A). Each fragment was cloned into pRW50 and Table 2 lists measurements of the expression of the resulting ogt::lac fusions. The results show that nitrate-dependent activation of the ogt promoter is suppressed by mutations in either of the two DNA sites for NarL.
In a further EMSA experiment, we found that purified FNR gave no clear shifted band with the ogt100 DNA fragment, suggesting that the pogt region contains no specific DNA site for FNR (results not shown). To check directly for any involvement of FNR in regulating the ogt promoter, nitrate-dependent induction of the ogt::lac fusion in strain JCB387 was compared with induction in a Δfnr derivative. Results presented in Figure 4(E) show that FNR has no significant effect on expression from the ogt promoter.
Fis represses the ogt promoter by displacing upstream-bound NarL
To investigate a role of Fis at the ogt promoter, expression of the ogt::lac fusion was compared in strain JCB387 and in the fis mutant derivative, JCB3871. Results in Figure 5(A) show that nitrate-induced expression is sharply suppressed during growth in Lennox broth supplemented with glucose, and that this suppression is greatly reduced in the JCB3871 fis strain. DNase I footprinting with purified Fis protein was then used to investigate whether Fis interacts directly at the ogt promoter. The results, illustrated in Figure 5(B), identify a target for Fis that overlaps the upstream site for NarL. Footprinting experiments with both Fis and NarL show that Fis prevents NarL binding to its upstream target at position −78.5 while not affecting binding at the downstream target (Figures 5B and 5C).
Most E. coli transcription activators function by binding to specific sites at target promoters and then recruiting RNA polymerase via a direct contact, and there are three major classes of activator-dependent promoter . At Class I promoters, the activator binds upstream of the promoter −35 element and recruits RNA polymerase by contacting the C-terminal domain of its α-subunit. At Class II promoters, the activator overlaps the promoter −35 element and can contact both the RNA polymerase α and σ subunits. At some promoters, recruitment of RNA polymerase requires more than one contact. At these promoters, known as Class III promoters, two or more activators make independent contacts with different targets in RNA polymerase . Although NarL appears to be a typical transcription activator belonging to the response-regulator family, at most of its previously studied target promoters , it co-activates promoter activity in conjunction with FNR, the master regulator for the adaptation of E. coli to growth in the absence of oxygen . The likely explanation for this is that most of the genes controlled by these promoters encode products that play a role in anaerobic respiration of nitrate and nitrite, and thus their expression is fully induced only in both the absence of oxygen (signalled via FNR) and the presence of nitrate or nitrite (signalled by NarL) [4,29,30]. However, the transcriptomics study of Constantinidou et al.  predicted that a small number of E. coli promoters are activated by NarL without FNR, and the comprehensive genetic analysis of the yeaR promoter by Lin et al.  showed that binding of NarL to a site that overlaps the promoter −35 element is sufficient to activate transcription without FNR. Here, using biochemical methods, we confirmed that purified NarL binds to this target and can activate open complex formation, most likely functioning as a Class II activator. The lack of induction of the yeaR promoter in a narL mutant background argues that NarP, which is a homologue of NarL, plays little or no role here. Note that NarP is also activated in response to nitrate and can bind to 7-2-7 NarL-binding targets [4,24]. In supplementary experiments, we found that the introduction of a narP mutation into either JCB387 or JCB3883 had very little effect on yeaR promoter activity (D.J.P. Squire, unpublished work).
Although the function of the products of the yeaR-yoaG operon is unknown, an important clue comes from the observation that the yeaR promoter is repressed by the binding of NsrR to a target that overlaps the −10 hexamer element (Figure 1A) [12,31,32]. Hence, even if NarL activity is triggered by nitrate or nitrite ions, the yeaR-yoaG operon will not be expressed unless repression by NsrR is lifted. Recent studies have identified NsrR as a transcriptional repressor that controls the expression of a regulon concerned with bacterial responses to RNS (reactive nitrogen species), such as nitric oxide [32–34]. NsrR contains an iron–sulphur cluster that senses RNS, which convert NsrR into a form that is unable to bind to its targets, and hence is unable to function as a repressor [35–37]. Our best guess is that most of the genes in the NsrR regulon encode products needed for RNS detoxification and the repair of damage caused by RNS. It is known that RNS are generated during anaerobic respiration of nitrate and nitrite [38–40], and hence it is easy to envisage that the yeaR-yoaG operon helps manage some aspect of RNS stress, after its expression is co-induced by external nitrate or nitrite ions (sensed by NarL) and RNS (sensed by NsrR). If this is the case, it makes sense for the yeaR promoter to be independent of FNR, since some RNS inactivate FNR [38,41].
During our experiments we noticed that nitrate-dependent induction at the yeaR promoter is greatly reduced during growth in rich media and we propose that this is due to the action of Fis protein. Fis is a sequence-specific DNA-binding protein that plays many roles in E. coli including contributing to compaction of the folded chromosome and regulating promoter activity [13,42]. Crucially, levels of Fis fluctuate dramatically, with over 50000 molecules per cell in rapidly growing cells in rich media, and greatly decreased levels during slow growth, for example in poor media or during starvation [43,44]. Whole genome studies using chromatin immunoprecipitation have shown that Fis can interact at hundreds of targets in the E. coli genome, and that over 60% of these targets are in gene regulatory regions [45,46]. Our in vitro studies (Figure 2) show that purified Fis binds to two targets at the yeaR promoter that flank the DNA site for NarL. Fis binding can displace NarL and this stops open complex formation (Figure 3). Thus we propose that, under conditions where Fis levels are raised, Fis binds at the yeaR promoter and prevents nitrate-dependent induction by blocking NarL binding to its target (illustrated in Figure 6A). This explains why the reduction of nitrate-dependent induction at the yeaR promoter, found in rich media, is lost in a fis mutant strain.
In the second part of the present study, we investigated the E. coli ogt promoter, which Constantinidou et al.  had also found to be induced by NarL independently of FNR. Note that a previous study  had identified the transcription start of the ogt promoter and its −10 element but did not investigate its regulation. Our results confirm that the ogt promoter is induced by NarL in response to nitrate ions, without help from FNR. However, in contrast with the situation at the yeaR promoter, induction requires the binding of NarL to two 7-2-7 sites, and the ogt promoter appears to be a Class III promoter. A possible explanation for this is that NarL binding at the ogt promoter is ∼10 times weaker than binding at the yeaR promoter, and this may also account for its lack of induction by nitrite ions, which are known to result in lower levels of active NarL . As at the yeaR promoter, the ogt promoter is not induced in a narL mutant background, suggesting that NarP also plays little or no role here. In agreement, we found that the introduction of a narP mutation into either JCB387 or JCB3883 had very little effect on ogt promoter activity (M. Xu, unpublished work).
In contrast with the yeaR promoter, the ogt promoter is not a target for NsrR . A clue to understanding the rationale for this comes from the work of Margison and co-workers [47,48], who were the first to clone and sequence the E. coli ogt gene and to investigate its promoter. These authors identified the ogt gene product as an O6-alkylguanine DNA alkyltransferase that removes alkyl groups from chemically damaged guanine residues in DNA. We suggest that the key to understanding why the expression of this DNA repair gene product is induced by nitrate ions is that RNS cause damage to DNA as well as to proteins [39,49–52]. In particular, RNS can cause lysine side chains in proteins and some free amino acids to become potent DNA-methylating agents. We suggest that the induction of ogt by NarL alone in response to external nitrate provides a prophylactic insurance policy against possible genotoxic effects arising from nitrate metabolism. As at the yeaR promoter, nitrate-dependent induction can be repressed by Fis that binds to a single site overlapping the upstream DNA site for NarL. We propose that, under conditions where Fis levels are raised, Fis binds to this target and prevents nitrate-dependent induction by blocking NarL binding to its target (illustrated in Figure 6B).
It is known that the expression of many non-essential genes is repressed in rapidly growing cells, where a large proportion of the available RNA polymerase is channelled to the essential genes for protein synthesis and cell growth, and in some cases, this repression is due to Fis [22,25,26]. Although there is no simple rationale for why the yeaR and ogt promoters are also repressed by Fis, we suggest that rapidly growing cells may opt out of certain stress responses, and we speculate that RNS may be a small risk in these conditions compared with other stresses. Remarkably, at both promoters, the mechanism of action of Fis is similar, with Fis binding to targets that overlap a DNA site for NarL, thereby preventing binding of an essential activator.
We thank the Wellcome Trust for funding this work with a programme grant. D. J. P. S was supported by a BBSRC (Biotechnology and Biological Sciences Research Council) studentship.
We are grateful to Rick Gourse for donating purified Fis protein and Kevin Chipman for helpful discussions.
Abbreviations: DTT, dithiothreitol; EMSA, electrophoretic mobility-shift assay; MBP, maltose-binding protein; MM, minimal medium; ONPG, o-nitrophenyl β-D-galactopyranoside; RNS, reactive nitrogen species
- © The Authors Journal compilation © 2009 Biochemical Society