|Home | About | Journals | Submit | Contact Us | Français|
The Escherichia coli aer regulatory region contains a single promoter that is recognized by RNA polymerase containing the flagellar sigma factor, σ28. Expression from this promoter is dependent on direct activation by the cyclic AMP receptor protein, which binds to a target centred 49.5 base pairs upstream from the transcript start. Activator-dependent transcription from the aer promoter was reconstituted in vitro, and a tethered inorganic nuclease was used to find the position of the C-terminal domains of the RNA polymerase α subunits in transcriptionally competent open complexes. We report that the ternary activator–RNA polymerase–aer promoter open complex is organized differently from complexes at previously characterized promoters. Among other E. coli promoters recognized by RNA polymerase containing σ28, only the trg promoter is activated directly by the cyclic AMP receptor protein. The organization of the different promoter elements and the activator binding site at the trg promoter is the same as at the aer promoter, suggesting a common activation mechanism.
The cyclic AMP receptor protein (CRP, also known as the catabolite activator protein, CAP) is a global transcription factor, which plays a central role in the control of metabolism in Escherichia coli and other enteric bacteria (Kolb et al., 1993; Barrett et al., 2005). CRP, which is functional as a homodimer, recognizes 22 bp target sequences, with the consensus 5′-AAATGTGATCTAGATCACATTT-3′. At most target promoters studied to date, CRP activates transcription by making one or more direct contacts with RNA polymerase, and there appear to be two major classes of simple CRP-activated promoters (Busby and Ebright, 1999). At Class I promoters, CRP binds upstream of the promoter −35 element, at a site centred at position −61.5 (i.e. between base pairs 61 and 62 upstream from the transcript start), or further upstream, and an activating region (AR1) in the downstream subunit of the CRP dimer makes contact with the C-terminal domain of one of the two RNA polymerase α subunits (αCTD). At Class II promoters, CRP binds at a target that overlaps the promoter −35 element and is usually centred at position −41.5. AR1 in the upstream subunit of the CRP dimer interacts with αCTD, while a second activating region (AR2) in the downstream subunit interacts with the N-terminal domain of one of the two RNA polymerase α subunits (αNTD) (Busby and Ebright, 1999).
Although the mechanisms of activation by CRP at both classes of promoter have been scrutinized in detail, most studies have focused on a small number of natural and synthetic model promoters, so it is unclear whether the findings apply to all target promoters. Genomic approaches have now identified scores of new targets for CRP (Tan et al., 2001; Brown and Callan, 2004; Gosset et al., 2004; Zheng et al., 2004; Grainger et al., 2005). This affords an opportunity to study CRP-dependent regulation at a range of naturally occurring promoters, and to uncover novel mechanisms of regulation by CRP. Previously, Hollands et al. (2007) investigated the action of CRP at 11 such uncharacterized targets in the E. coli K-12 genome. One of these was located in the regulatory region of the aer gene, which encodes an aerotaxis sensor protein that controls movement of bacterial cells in response to the availability of oxygen and other electron acceptors in the environment (Bibikov et al., 1997; Rebbapragada et al., 1997; Taylor et al., 1999). CRP binding upstream of aer was first detected by Grainger et al. (2005) in a whole genome chromatin immunoprecipitation analysis, and it was subsequently shown that CRP activates transcription by binding to a single DNA site in the aer regulatory region (Hollands et al., 2007).
Recent transcriptome analyses have indicated that expression of aer in both E. coli and Salmonella enterica serovar Typhimurium requires an alternative σ factor, σ28 (Frye et al., 2006; Zhao et al., 2007). Recall that the RNA polymerase σ subunit is a dissociable promoter specificity factor that binds to core RNA polymerase (E) to form the RNA polymerase holoenzyme (Eσ), which can recognize promoter sequences and initiate transcription (Burgess et al., 1969). Most bacteria contain multiple σ factors that recognize different promoter −10 and −35 elements. A primary σ factor (σ70 in E. coli) drives the transcription of genes with ‘housekeeping’ functions, while a number of alternative σ factors direct transcription of particular sets of genes in response to environmental signals or stresses, or function to control development (Ishihama, 2000; Gruber and Gross, 2003). σ28, which is encoded by the fliA gene, is the most widely distributed alternative σ factor (Koo et al., 2009; Smith and Hoover, 2009), and controls the transcription of operons required for flagellar filament assembly and for the regulation of motility and chemotaxis in a large number of Gram-positive and Gram-negative bacteria (Chilcott and Hughes, 2000).
Most studies on transcription activation by CRP have been concerned with promoters recognized by RNA holoenzyme containing σ70 (Eσ70). Here, we report the first investigation into the direct regulation by CRP of transcription by RNA polymerase containing σ28 (Eσ28). We show that aer is transcribed from a single σ28-dependent promoter that is activated by CRP binding at a location different from any previously characterized CRP-activated promoter. We also show that CRP directly activates transcription from a second σ28-dependent promoter that has a similar organization.
To study the effects of CRP and σ28 on expression of aer, we cloned a DNA fragment covering the aer gene regulatory region (aer200; Hollands et al., 2007) into a low-copy-number lac expression vector, pRW50, and we measured the activity of the resulting aer200::lacZ fusion in E. coli K-12 Δlac strain M182 and derivatives containing deletions of either the crp or fliA gene. Results presented in Fig. 1A (black lines) show that, in M182, there is a large increase in promoter activity during late exponential phase that decreases on entry into stationary phase. This is consistent with the findings of Barembruch and Hengge (2007), who observed a similar pattern of expression for the σ28-dependent flgM promoter, and correlates with an accumulation of σ28 protein during late exponential phase followed by a decline in σ28 levels once the culture enters stationary phase (K. Hollands, unpubl. data; Barembruch and Hengge, 2007). In the ΔfliA and Δcrp backgrounds (Fig. 1A, red and grey lines), promoter activity remained at a basal level throughout the growth cycle. This confirms that both CRP and σ28 are essential for expression from the aer regulatory region in vivo. However, this experiment is complicated by the fact that σ28 expression is dependent on CRP. This is because CRP is required to activate transcription of the flhDC operon that encodes an essential activator of transcription from the fliA promoter (Soutourina et al., 1999). Indeed, Western blot analysis confirms that no σ28 protein is present in strain M182 Δcrp (Fig. S1, lanes 1–3).
To investigate the action of CRP at the aer regulatory region, independent of the indirect effect of CRP on σ28 levels, we established an experimental system in which expression of σ28 is uncoupled from CRP. To do this, we used ΔfliA derivatives of M182 and M182 Δcrp that had been transformed with pKXH100, which encodes fliA under the control of a CRP-independent promoter. Western blot analysis confirms that, in this system, expression of σ28 does not require CRP (Fig. S1, lanes 4–6). We measured expression of the aer200::lacZ fusion in pRW50 in these strains. Data illustrated in Fig. 1B show that, in M182 ΔfliA pKXH100, the activity of the aer200::lacZ fusion follows a similar pattern to that in strain M182 (which expresses fliA from the chromosome), except that the increase in promoter activity occurs later in growth, once the culture begins to enter stationary phase. This correlates with a delayed increase in σ28 protein levels in this background (K. Hollands, unpubl. results). In the absence of CRP, promoter activity remains low throughout the growth cycle, showing that the requirement for CRP for expression from the aer regulatory region is independent of the effect of CRP on σ28 levels. We conclude that CRP must function directly at the aer regulatory region, and this is consistent with the previous observation that introducing mutations into the DNA target for CRP upstream of aer prevents CRP-dependent activation of the aer200::lacZ fusion (Hollands et al., 2007).
Although the aer regulatory region has been predicted as a target for Eσ28 (Park et al., 2001; Frye et al., 2006; Zhao et al., 2007), the promoter determinants required for transcription initiation have not been identified experimentally. To define the DNA elements required for σ28-dependent transcription of aer, we began by mapping the aer transcript start site by primer extension analysis, using the aer200 promoter fragment cloned in pRW50. This yielded a single extension product approximately 148 nucleotides in length (Fig. 2A), which places the transcript start point at the position labelled +1 in Fig. 2B. This falls 6 bp downstream from the −10 octamer element for a σ28-dependent promoter predicted by Park et al. (2001). To examine the importance of this promoter, we constructed derivatives of the aer200 fragment containing point mutations in the putative −10 and −35 elements (Table 1 and Fig. 2B). In the −10 element, we targeted the highly conserved 5′-CGA-3′ motif, from positions −11 to −9, because mutations in this motif result in a loss of σ28-dependent transcription from other σ28-dependent promoters, both in vivo and in vitro (Yu et al., 2006; Wozniak and Hughes, 2008). In the −35 octamer, we targeted positions −32T and −30A, which are also highly conserved and are important for σ28-dependent transcription from the Salmonella flgM promoter, and position −28, which has more minor effects on flgM promoter activity (Wozniak and Hughes, 2008). Each mutant promoter fragment was cloned into pRW50, and expression of the resulting promoter::lacZ fusions was measured in the CRP+ FliA+, CRP– FliA+ and CRP+ FliA– backgrounds. Results listed in Table 1 show that the substitutions in the −10 element had the greatest effect on promoter activity, reducing expression from the aer regulatory region to the level observed in the absence of σ28. Mutations at positions −32, −30 and −28 in the −35 element also severely reduced promoter activity. We conclude that the proposed −10 and −35 elements are essential for σ28-dependent transcription of aer, and, together with the transcript start site data, this argues that aer is expressed from a single promoter, at least under the conditions tested here.
Next, we sought to confirm our in vivo findings by examining the σ factor selectivity and CRP dependence of the aer promoter in vitro. We began by cloning the aer200 fragment upstream of the λoop terminator in plasmid pSR, and tested the ability of purified Eσ28 and Eσ70 to drive transcription from the aer promoter in an in vitro multi-round transcription assay, in the presence and absence of purified CRP and cAMP (Fig. 3A). In this system, transcription initiating at the aer promoter terminates at the λoop terminator to generate a 158-base transcript that can be identified by electrophoresis. In the presence of Eσ28, a single transcript was observed (Fig. 3A, lanes 3–12). At low Eσ28 concentrations, this transcript is detected only in the presence of CRP (lanes 3–6), although some transcript is generated in the absence of CRP as the RNA polymerase concentration is increased (lanes 7–12). At even higher concentrations of Eσ28, transcription becomes completely independent of CRP (data not shown). The aer transcript generated by Eσ28 is not detected in reactions using Eσ70 (Fig. 3A, lanes 13 and 14). Instead, a single CRP-independent transcript is produced, which corresponds to the 108-base RNAI control transcript that originates from the pSR replication origin.
To confirm that in vitro transcription initiates from the same promoter defined in our in vivo experiments, promoter unwinding by RNA polymerase was monitored by using KMnO4 to probe for single-stranded regions of DNA (Fig. 3B). In the presence of Eσ28 (lanes 3 and 4), KMnO4-reactive bands appeared from positions −10 to +4, indicative of promoter melting around the −10 element of the σ28-dependent promoter highlighted in Fig. 2B. This was observed both in the presence and in the absence of CRP, which is consistent with our finding that transcription initiation by Eσ28 is independent of CRP in vitro at the high RNA polymerase concentrations used in these reactions. Incubation with Eσ70 did not result in promoter melting around the aer transcript start site, either in the presence or in the absence of CRP (Fig. 3B, lanes 5 and 6). Taken together, the in vitro data confirm that aer is transcribed from a single, σ28-dependent promoter that is activated by CRP when the RNA polymerase concentration is limited. The observation that the aer promoter becomes less dependent on CRP at higher RNA polymerase concentrations suggests that CRP activates transcription by recruitment of RNA polymerase (Rhodius et al., 1997).
Mutational analysis showed that CRP-dependent activation of the aer200::lacZ fusion requires CRP binding to the single DNA target indicated in Fig. 2B (Hollands et al., 2007). This target site is centred 49.5 bp upstream from the transcript start site, which falls between the typical Class I location of position −61.5 and the Class II location of position −41.5. To investigate whether CRP can activate σ28-dependent transcription from positions −41.5 or −61.5 at the aer promoter, we constructed a deletion or insertion in the aer200 fragment to make the aer212 and aer211 fragments (Fig. 4, upper three panels). These fragments were cloned into pRW50, and expression of the resulting promoter::lacZ fusions was measured in the CRP+ FliA+, CRP– FliA+ and CRP+ FliA– backgrounds. The results illustrated in Fig. 4 show that, when the DNA site for CRP is moved to position −41.5, aer promoter activity in the CRP+ FliA+ strain is reduced to a similar level to that observed in the absence of CRP or σ28. This indicates that CRP cannot activate transcription from the aer promoter when bound at a Class II location. Moving the CRP site to position −61.5 (aer211) results in a twofold decrease in promoter activity, but, while the residual promoter activity is dependent on CRP, it is independent of σ28. The most likely explanation for this is that, here, CRP is activating transcription from an alternative cryptic σ28-independent promoter. For example, in the aer211 fragment, a 6 bp sequence, 5′-TAAAGA-3′, is located 32 bp downstream of the DNA site for CRP and this may well generate a weak Class II CRP-dependent promoter served by Eσ70 (recall that the consensus −10 hexamer for Eσ70 is 5′-TATAAT-3′).
Next, we used the same system to monitor the effects of making a 5 bp deletion or insertion between the DNA site for CRP and the −35 element at the aer promoter. Both the deletion, which moved the DNA site for CRP to position −44.5 (aer226), and the insertion, which moved the DNA site for CRP to position −54.5 (aer227), resulted in a reduction in promoter activity in the CRP+ FliA+ strain to the basal level observed in the absence of CRP (Fig. 4, lower two panels). This indicates that CRP is unable to activate transcription from the aer promoter when its DNA site is moved to the opposite face of the DNA helix. These experiments argue that optimal activation of σ28-dependent transcription requires CRP binding at position −49.5.
Activation by CRP at both Class I and Class II σ70-dependent promoters requires a contact between CRP activating region 1 (AR1) and αCTD. Previous work showed that CRP-dependent activation at the aer promoter also requires AR1 (Hollands et al., 2007), which likely functions by contacting αCTD in Eσ28. Because the organization of the aer promoter is unlike that at Class I or Class II CRP-dependent promoters, it is unclear whether the interaction between AR1 and αCTD occurs via the upstream or downstream subunit of dimeric CRP bound at the promoter. To address this, we mapped the location of αCTD binding at the aer promoter using purified RNA polymerase that had been labelled with the chemical nuclease reagent iron [S]-1-[p-bromoacetamidobenzyl] ethylenediaminetetraacetate (FeBABE) on a single cysteine residue at position 302 in the αCTDs (see Experimental procedures). Transcriptionally competent open complexes were formed using the end-labelled aer200 promoter fragment, purified CRP and FeBABE-tagged Eσ28, and DNA cleavage by FeBABE was triggered. Analysis of the pattern of DNA cleavage by gel electrophoresis reveals the location of the αCTDs at the aer promoter. Note that, in this assay, in most cases, a single Fe-BABE-labelled αCTD will give rise to cleavages in two adjacent minor grooves, as a wave of hydroxyl radicals generated from the Fe-BABE impinges on the target DNA (Lee et al., 2003).
Results presented in Fig. 5A show that, in the presence of CRP and Eσ28 (lane 3), DNA cleavage on the template strand of the aer promoter is enhanced around positions −72 and −64 upstream of the DNA site for CRP, and around positions −38 and −30 downstream of the CRP site. This indicates that the αCTDs can contact the DNA both upstream and downstream of the bound CRP dimer. In the presence of Eσ28, but in the absence of CRP (lane 2), the pattern of DNA cleavage is similar to the background detected in the absence of any protein (lane 1). This suggests that the two αCTDs are positioned at their targets on the DNA only in the presence of CRP. Interestingly, the spacing between the centre of the DNA site for CRP and the downstream FeBABE-induced DNA cleavage at the aer promoter is identical to that observed by Lee et al. (2003) at a Class I CRP-dependent promoter served by Eσ70 (Fig. 5B and C). Similarly, the spacing between the centre of the DNA site for CRP and the upstream FeBABE-induced DNA cleavage is identical to that seen at a Class II CRP-dependent promoter served by Eσ70 (Lee et al., 2003). Thus, the juxtaposition between the downstream-bound αCTD and CRP in open complexes at the aer promoter appears to be identical to the AR1-mediated juxtaposition between downstream-bound αCTD and CRP at a Class I CRP-dependent promoter. Similarly, the juxtaposition between the upstream-bound αCTD and CRP at the aer promoter appears to be identical to the AR1-mediated juxtaposition between upstream-bound αCTD and CRP at a Class II CRP-dependent promoter (Fig. 5B and C).
In the crystal structure of the CRP-αCTD-DNA complex, αCTD contacts approximately 6 bp of DNA spanning a minor groove, centred 18–19 bp from the centre of the DNA site for CRP (Benoff et al., 2002). The locations of the specific DNA cleavages at the aer promoter are consistent with binding of the αCTDs at sites centred 18.5 bp both upstream and downstream of the DNA site for CRP (Fig. 5B). These sequences are also AT-rich, a feature associated with DNA binding by αCTD (Gourse et al., 2000).
To investigate whether CRP directly regulates Eσ28-dependent transcription at other promoters, we used electromobility shift assays to compare the binding of CRP to end-labelled DNA fragments covering the regulatory regions of aer and the seven other σ28-dependent operons from E. coli K-12 strain MG1655 described by Zhao et al. (2007). The results, illustrated in Fig. 6, show that CRP binds to a single site in the aer200 fragment, but binding of CRP to specific targets in the tsr and trg regulatory regions was also detected. Note that bioinformatic analyses had predicted DNA sites for CRP upstream of both tsr and trg (Robison et al., 1998). The trg promoter fragment binds CRP with similar affinity to the aer fragment, while CRP binding to the tsr fragment is much tighter. No clear binding of CRP was found with the fliC/fliD, flgMN, flgKL, motAB/cheAW or tar/tap/cheRBYZ fragments.
The action of CRP at the tsr and trg regulatory regions was studied further. In the tsr regulatory region, the predicted CRP site is located 132.5 bp upstream of the σ28-dependent tsr promoter, so it is unlikely that CRP makes direct contact with bound Eσ28. Indeed, no direct effect of CRP on gene expression from the tsr regulatory region could be detected (K. Hollands, unpubl. data). In contrast, alignment of the DNA sequences of the trg and aer regulatory regions revealed that the spacing between the predicted DNA sites for CRP and the different elements of the two σ28-dependent promoters is identical (Fig. 7A).
To measure the effect of CRP on expression from the trg regulatory region, the trg100 promoter fragment was cloned into pRW50, and the activity of each promoter::lacZ fusion was measured in the CRP+ FliA+, CRP– FliA+ and CRP+ FliA– backgrounds. Recall that, in the conditions used in our experiments, direct effects of CRP on transcription of promoter::lacZ fusions in pRW50 can be measured independent of the effect of CRP on σ28 levels. Results illustrated in Fig. 7B indicate that expression from the trg regulatory region, like the aer promoter, is dependent on σ28 and is activated by CRP. The conservation of the spacing between the DNA site for CRP and the −10 and −35 elements at the aer and trg promoters suggests that the mechanisms of transcription activation at the two promoters are similar. Interestingly, the trg and aer genes encode homologous proteins with similar functions. While Aer is an energy sensor that controls responses to redox signals, Trg is a chemosensor that responds to the monosaccharides ribose and galactose (Taylor et al., 1999).
Here we have described the first examples of direct activation by CRP of promoters served by RNA polymerase holoenzyme containing the flagellar sigma factor, σ28. We showed that transcription of the E. coli K-12 aer gene is driven by a single σ28-dependent promoter, which is activated by CRP binding to a single site positioned 49.5 base pairs upstream of the transcript start site. This location appears optimal for activation. This is in contrast to the situation at previously studied Class I and Class II CRP-dependent promoters where the optimal locations for activation by CRP are positions −61.5 and −41.5 respectively, and where CRP activates only very weakly when bound to a site centred near position −50 (Gaston et al., 1990). Our results argue that the spacing requirements for CRP-dependent activation at promoters served by Eσ28 differ from those at promoters served by Eσ70. It is possible that promoters recognized by some other alternative σ factors also require CRP binding at unusual locations. For example, at the σ38-dependent csiD promoter, CRP activates optimally from a DNA site centred at position −68.5 (Germer et al., 2001). From this position, or a site located one helical turn upstream, CRP can activate σ38-, but not σ70-dependent transcription.
Activation at the aer promoter requires AR1 of CRP that likely contacts αCTD (Hollands et al., 2007). Our results show that the two αCTDs of Eσ28 contact DNA both upstream and downstream of CRP, although note that we cannot prove that both contacts occur simultaneously. The finding that one αCTD binds downstream of CRP at the aer promoter was surprising. Structural modelling of the CRP-RNA polymerase-DNA complex at a Class I promoter, where the DNA site for CRP is centred at position −61.5, indicates that one αCTD is tightly sandwiched between CRP and σ70, such that it can simultaneously contact DNA, AR1 on CRP and σ70 domain 4 (Chen et al., 2003; Lawson et al., 2004). As the DNA site for CRP at the aer promoter is located 12 bp downstream, it appears that there cannot be sufficient space for αCTD to fit between the CRP dimer and the promoter-bound sigma factor. We modelled the structure of the CRP-RNA polymerase-DNA complex at the aer promoter by combining the crystal structure of the CRP-αCTD-DNA complex (Benoff et al., 2002) with the EσA.-fork junction DNA structure (Murakami et al., 2002) and, as expected, we found that there is a clash between the predicted locations of the αCTD downstream of CRP and domain 4 of σ, which contacts the promoter −35 element (K. Hollands and D.J. Lee, unpublished). This leads us to propose a model in which the organization of the CRP-Eσ28-DNA complex at the aer promoter differs from that of the CRP-Eσ70-DNA complex at a Class I σ70-dependent promoter (Fig. 8A). Our FeBABE footprinting data indicate that the juxtaposition of CRP and the downstream αCTD at the aer promoter is the same as at the Class I promoter. This implies that it must be domain 4 of σ28 that is positioned differently within the CRP-Eσ28-DNA complex, compared with domain 4 of σ70 within the CRP-Eσ70-DNA complex at a Class I promoter. This is supported by the observation that the −10 and −35 elements at σ28-dependent promoters are located 2–3 bp closer together than at promoters served by Eσ70 (Fig. 8B), suggesting that the contact site for σ28 domain 4 on promoter DNA may lie several bases downstream of that for σ70 domain 4 at a σ70-dependent promoter. The model, illustrated in Fig. 8A, suggests that domain 4 of σ28 is offset by rotation and translocation around the surface of the promoter DNA. It is quite possible that this is its ‘normal’ position at σ28-dependent promoters, but this will require further experimental evidence.
Although E. coli contains hundreds of transcription activators, there are few examples of factor-dependent activation of promoters recognized by alternative σ factors such as σ28. Transcription from promoters served by alternative σ factors is mostly regulated by controlling the expression and activity of the σ factor itself, and by the very stringent promoter recognition properties of the alternative σ factors. It is generally accepted that control over most flagellar genes is exerted by regulating the expression and activity of FlhDC, σ28, and the anti-σ factor, FlgM (Chilcott and Hughes, 2000; Keseler et al., 2009). Our findings show that transcription activators can also play an important role in controlling transcription by Eσ28.
The similar organization of the aer and trg promoters suggests a common mechanism of direct activation by CRP. However, we found no evidence for direct regulation by CRP at the six other well-characterized σ28-dependent promoters in E. coli K-12. This raises the question of why CRP should directly regulate transcription of aer and trg, particularly when CRP indirectly regulates transcription of all genes in the flagellar cascade by controlling expression of the master regulator, FlhDC. Expression of FlhDC is tightly regulated by multiple transcription factors, including CRP, H-NS and OmpR, and hence the genes of the flagellar cascade are regulated in response to many different environmental inputs (Shin and Park, 1995; Soutourina et al., 1999). It is possible that expression of aer and trg is required only in response to conditions that induce CRP activity and not in response to other signals that induce the flagellar cascade. Alternatively, expression of aer and trg may need to be upregulated to a greater extent than other genes when CRP activity is induced. This may be particularly important when levels of Eσ28 are low. In these conditions, recruitment of Eσ28 by CRP may ensure that the aer and trg transcription units are preferentially expressed compared with other σ28-dependent genes. Note that aer and trg encode homologous methyl-accepting chemotaxis regulator-type proteins, which interact with the flagellar motor via the CheA/CheY signalling pathway to control the direction of bacterial movement in response to different chemical or energetic signals (Taylor et al., 1999). The direct involvement of CRP in their regulation may therefore simply be a reflection of the role of CRP in facilitating the acquisition and metabolism of nutrients other than glucose (Kolb et al., 1993).
The E. coli K-12 strains, plasmids and promoter fragments used in this study are listed in Table 2. Standard recombinant DNA techniques were used throughout and all the oligonucleotide primers used are listed in Table S1.
The ΔfliA derivatives of strains M182 and M182 Δcrp were constructed by P1 transduction of a fliA::kan mutation from strain JW1907-1. The kanamycin resistance marker was subsequently removed by expressing the FLP recombinase from plasmid pCP20, as described by Cherepanov and Wackernagel (1995), and the presence of the deletion was confirmed by colony PCR using primers D56550 and D56551.
Plasmid pKXH100 was constructed by cloning an NdeI-XhoI fragment carrying the fliA coding sequence from E. coli K-12 strain MG1655, amplified by PCR using primers D57845 and D57846, into plasmid pET21a (Novagen). As a result of leaky expression, genes cloned under the control of the T7 promoter in pET21a are expressed even in strains that do not produce T7 RNA polymerase, including M182 (Wu et al., 2005). This activity is independent of the presence of the inducer IPTG (K. Hollands, unpubl. results).
The DNA sequence of each promoter fragment is shown in Fig. S2. Promoter fragments were amplified by PCR from genomic DNA of E. coli K-12 strain MG1655, using primers that introduce flanking EcoRI and HindIII sites (listed in Table S1). For promoter activity assays, EcoRI-HindIII fragments were cloned into the lac expression vector, pRW50. To construct templates for in vitro transcription assays, and to generate DNA fragments for electromobility shift assays and footprinting, promoter fragments were cloned into plasmid pSR. Derivatives of the aer200 fragment carrying point mutations in the −10 or −35 elements (aer206, aer213, aer214 and aer224) were constructed by megaprimer PCR. In a first-round PCR reaction, a megaprimer was synthesized from pSR/aer200 as a template, using a mutagenic primer carrying the desired mutation, and a flanking primer (either D51598 or D53041: see Table S1). The megaprimer was then used in a second-round PCR with the opposing flanking primer and pSR/aer200 as a template to generate a full-length promoter fragment containing the required mutation, which was then cloned into pRW50. The aer212, aer211, aer226 and aer227 fragments were constructed by inserting or deleting DNA between the DNA site for CRP and the −35 element of the aer promoter. First, two PCR products were synthesized using pSR/aer200 as a template: one generated using upstream primer D53041 and a downstream primer carrying the insertion or deletion, and a second generated using the downstream primer D51598 and an upstream primer carrying the insertion or deletion (see Table S1). The two PCR products were then annealed via their 26–32 bp overhangs, and the two strands were extended using DNA polymerase to generate a full-length promoter fragment carrying the insertion or deletion. This product was then amplified by PCR using primers D53041 and D51598 and cloned into pRW50.
β-Galactosidase levels in cells carrying promoter::lacZ fusions, cloned in pRW50, were measured using the method of Miller (1972). Cells were grown aerobically at 37°C in LB medium. Activities are shown in Miller units (nmol ONPG hydrolysed min−1 mg−1 dry cell mass), and are averages from at least three independent experiments.
Transcript start sites were mapped by primer extension as described in Lloyd et al. (2008), using RNA purified from strain M182 carrying the aer200 promoter fragment cloned in pRW50 and 5′ end-labelled primer D49724, which anneals downstream of the HindIII site in pRW50. Primer extension products were analysed on denaturing 6% polyacrylamide gels, calibrated with sequencing reactions, and were visualized using a Fuji phosphor screen and Bio-Rad Molecular Imager FX.
Purified CRP protein was donated by David Grainger (University of Warwick, UK), and wild type E. coli core RNA polymerase was purchased from Epicentre Technologies (Madison, WI). His-tagged RNA polymerase α subunits containing a single cysteine residue at position 302 were prepared and labelled with FeBABE as described by Lee et al. (2003). FeBABE-tagged α subunits were incorporated into core RNA polymerase using the reconstitution method of Tang et al. (1995). Purified σ28 and σ70 proteins were prepared from BL21(DE3) cells carrying the overexpression plasmid pKXH100, as described by Grainger et al. (2008). Eσ28 and Eσ70 holoenyzmes were made by mixing wild type or FeBABE-labelled core RNA polymerase with an equimolar amount of σ28 or σ70, and incubating for 20 min at room temperature.
Caesium chloride preparations of pSR carrying the aer200 promoter fragment served as a template for multiple-round in vitro transcription assays, as described by Savery et al. (1998). 20 ng pSR/aer200 was incubated in transcription buffer containing 40 mM Tris pH 7.9, 10 mM MgCl2, 1 mM dithiothreitol, 100 mM KCl, 100 μg ml−1 bovine serum albumin, 200 μM GTP, 200 μM ATP, 200 μM CTP, 10 μM UTP and 5 μCi [α32P]-UTP. Where indicated, CRP was included at 100 nM and cAMP at 0.2 mM. Reactions were started by adding Eσ28 or Eσ70. RNA products were analysed on a denaturing 5.5% polyacrylamide gel and visualized using a Fuji phosphor screen and Bio-Rad Molecular Imager FX.
KMnO4 and FeBABE footprinting experiments were performed on PstI-HindIII fragments prepared from caesium chloride preparations of pSR carrying aer200. Fragments were labelled at the HindIII end with [γ-32P]-ATP using polynucleotide kinase. KMnO4 footprints were performed following the protocol of Browning et al. (2009) and FeBABE footprints were carried out as described by Lee et al. (2003). Each reaction contained approximately 3 nM labelled PstI-HindIII DNA fragment in 20 mM HEPES pH 8.0, 5 mM MgCl2, 50 mM potassium glutamate, 1 mM DTT and 0.5 mg ml−1 BSA. KMnO4 footprinting reactions contained 0.2 mM cAMP, 100 nM CRP and 50 nM Eσ28 or Eσ70, as required. FeBABE footprinting reactions contained 0.2 mM cAMP, 100 nM CRP and 200 nM FeBABE-labelled Eσ28. The products of KMnO4 and FeBABE footprinting reactions were analysed on denaturing 6% polyacrylamide sequencing gels, calibrated with Maxam-Gilbert ‘G + A’ sequencing reactions.
The EMSA experiments were performed using EcoRI-HindIII fragments prepared from pSR derivatives, and end-labelled using [γ-32P] ATP and polynucleotide kinase. EMSA reactions were carried out as described by Lloyd et al. (1998) and were analysed on 5% polyacrylamide gels. Footprinting and EMSA gels were visualized using a Fuji phosphor screen, and analysed using a Bio-Rad Molecular Imager FX and Quantity One software (Bio-Rad).
K.H. was supported by a PhD studentship from the UK BBSRC and this work was funded by a Wellcome Trust Programme Grant to S.J.W.B.
Additional supporting information may be found in the online version of this article.
Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.