Coordinated Repression In Vitro of the Divergent fepA-fes Promoters of Escherichia coli by the Iron Uptake Regulation (Fur) Protein

Abstract

The mechanism involved in transcriptional repression of the fepA-fes divergent promoters of Escherichia coli by the Fur (ferric uptake regulation) protein has been examined in vitro. This DNA region includes a suboptimal and single Fur-binding site with two divergent and overlapped −35/−10 hexamers. Comparison of transcription patterns generated with runoff experiments in either the presence or the absence of heparin showed that access of the RNA polymerase to the principal −35/−10 hexamers was fully prevented by Fur-Mn²⁺ bound to its target site within the divergent promoter region. Neither RNA polymerase bound to the fes and fepA promoters could be displaced by Fur-Mn²⁺, nor could the bound repressor be outcompeted by an excess of the enzyme. However, the repressor blocked reinitiation as soon as the polymerase moved away from the fes promoter during transcription. The spatial distribution of regulatory elements within the DNA region allowed the simultaneous binding of the RNA polymerase to the fes and fepA promoters and their coordinate regulation regardless of their different transcriptional activities. Comparisons with other iron-regulated systems support a general mechanism for Fur-controlled promoters that implies a direct competition between the polymerase and the regulator for overlapping target sites in the DNA.

The ferric uptake regulator (Fur) protein of Escherichia coli is a 17-kDa repressor which controls the response of this bacterium to iron starvation (1, 15, 29). Under iron-rich conditions, Fur represses virtually all operons involved in high-affinity uptake of Fe(III). This is due to the Fe(II)-dependent DNA-binding activity of the protein, which responds exquisitely to changes in the intracellular pool of the metal ion (2). It is only under iron-depleted conditions that this repression is released, thereby triggering expression of genes involved in siderophore biosynthesis and transport (3). In the last few years, Fur homologs and cognate operators (Fur boxes) (8, 11) have been identified in numerous gram-negative bacteria (20, 24, 26, 28, 30). Both in E. coli and in other genera, Fur appears to control not only siderophore production and iron transport but also bacterial virulence determinants (19) for animal and plant tissues, thus suggesting a key role of this protein in survival and proliferation on the cells in a hostile environment.

Although there is a long list of bacterial genes and operons which are regulated by iron by means of the Fur protein (14, 17, 21, 27), important features of the repression mechanism remain obscure. Some details of such a mechanism have been revealed by using the simple model of the aerobactin operon of the enterobacterial virulence plasmid pColV-K30 (5, 10–12). We have recently shown that the basis of the mechanism of repression used by Fur in the aerobactin promoter is direct competition between RNA polymerase (RNAP) and Fur-Fe²⁺ for the same target sites around the −35 hexamer of the major aerobactin promoter, named P1 (13). This was based on the use of in vitro transcription and footprinting assays with purified components, which clearly showed that Fur and RNAP replaced each other at the aerobactin promoter on the sole basis of the iron status of the medium.

This mechanism cannot, however, be generalized, since not all iron-regulated promoters are organized as those in the aerobactin operon. An interesting case includes the bidirectional promoters which control production and transport of enterochelin (also called enterobactin) (9). This siderophore forms part of the housekeeping high-affinity iron transport system of many enterobacteria, and its regulation is determined by three regulatory regions containing divergently oriented promoters: one located between fepB and the entCEBA operon (7), a second divergent region between fepD and the promoter in front of the gene for protein P43 (25), and finally, a third regulatory region between fepA and fes (16, 23). The last case is particularly interesting, since one single Fur-binding site appears to coordinately iron regulate the expression of all transcripts which, at very different levels, emanate from this control region (16, 23). Similar to the aerobactin promoter, two sequentially occupied zones (I and II) have been identified by DNase I footprinting, although with different extensions (Fig. 1).

FIG. 1.

Organization of the bidirectional fepA-fes promoter region. The DNA segment shown includes the major regulatory elements that determine responsiveness of these promoters to iron. The region contains the partially overlapping and divergent −35/−10 hexamers for RNAP binding as well as a single Fur box (11). The region of protection of Fur-Mn²⁺ from DNase I nicking spans not only this box but also adjacent sequences I and II, depending on repressor concentration (16). The transcription initiation sites of each promoter (+1) are indicated. Note that only the −35/−10 hexamers of the fepA promoter, which gives rise to the major transcript T2 in vitro (see Fig. 2), are shown. The origin of the second, minor fepA transcript, T1, is indicated to the left of the figure. The cylindrical projection of the DNA sequence and the prediction of DNA curvatures (6) indicate that the RNAPs bound to each promoter occupy opposite helix sides (see Fig. 5).

If repression of the two divergent fepA-fes promoters by Fur-Fe²⁺ occurs through a simple competition mechanism based on relative affinities for a target site (18), it is not trivial to show how the two promoters, which have quite different transcriptional activities, can be coordinately regulated by a single Fur box (16). Our results below suggest that while the activities of the fes and fepA promoters are fully independent of each other, the repression-activation switch occurred in both promoters coordinately. Such a switch appears to be directed by the RNAP with the greater affinity for the region, thus overcoming the need for two independent Fur-binding sites for the control of the divergent promoters.

Transcriptional repression of the fepA-fes promoters in vitro.

In order to examine the ability of Fur protein to repress in vitro transcription in the fepA-fes bidirectional regulatory region, we monitored production of runoff transcripts from each promoter (Fig. 2A). Inspection of the results of the experiment shown in Fig. 2A indicated that the transcriptional activities of the two promoters were clearly different, the activity of the fes promoter being about fivefold higher than that of fepA, as indicated by the relative abundance of each mRNA (see also the section on single-round assays below). Furthermore, a minor fepA transcript of 243 nucleotides (T1), apparently originating at a weaker promoter (16) (Fig. 1), accompanied systematically the major 280-nucleotide transcript (T2) from the main fepA promoter. It should be noted that the relative activity of each of the fepA promoters in vitro is different from that of the transcript abundance observed in vivo with primer extension (16, 23). This may reflect different mRNA stabilities or other conditions prevailing in vivo and not as much intrinsic promoter activity as we detect in vitro. For the sake of simplicity, we will refer hereafter to the promoter causing transcript T2 as the only fepA promoter, since it is the one that contributes the most to fepA expression in the in vitro assay.

FIG. 2.

(A) Transcriptional repression of the bidirectional fepA-fes promoter region in vitro by the Fur protein. The scheme on top shows the 470-bp DNA fragment used in the experiment and the sizes of the transcripts originated. For obtaining this DNA template, a 320-bp fragment was amplified from chromosomal DNA of E. coli by PCR with 30-mers 5′-GGAATTCGCCATGTTTCGACTGCCACCAGC-3′ (FEPAB) and 5′-CGGGATCCGCCAGGGAATGAATCTTCTTGT-3′ (FESE). These primers generated EcoRI and BamHI sites at the ends of the amplified fragment, which span the fepA-fes region (23). The amplified DNA was cloned in pUC19, and the DNA segment used for transcription assays was obtained by PCR of the resulting plasmid with universal direct and reverse sequencing primers. For the experiment shown, 5 nM linear DNA template was preincubated at 37°C with increasing concentrations of pure Fur protein (23 nM, 117 nM, 235 nM, 352 nM, 470 nM, and 587 nM) in a buffer containing 50 μM MnCl₂. In the last lane, the sample had 587 nM Fur and 200 μM EDTA. After preincubation for 5 min, E. coli RNAP was added to all samples at a concentration of 80 nM, followed by NTPs. The reaction mixture was incubated for 10 min and then examined in a denaturing polyacrylamide gel. The sample in lane C had no Fur protein added. (B) Comparison of repression patterns of the fes and fepA promoters. The transcription from each of the promoters (T1 and T2 combined in the case of fepA) is plotted as a function of the concentration of the Fur protein added to the assay above. Gel autoradiographs were quantified with an image analysis system (Bio-Rad, Hercules, Calif.). The results shown are the average of two separate experiments.

In all repression assays (the methods used are described in reference 13), Mn²⁺ was used instead of Fe²⁺ as a cofactor for Fur, owing to its superior stability in an aerobic environment. As shown in Fig. 2A, increasing concentrations of Fur protein in the presence of Mn²⁺ decreased the activity of fepA and fes promoters with respect to a repressorless control. Addition of a metal-chelating agent, such as EDTA, restored the transcription from the otherwise repressed promoters, indicating that the divalent ion is essential for the down-regulation effect. Fig. 2A shows also that repression was exerted simultaneously on every transcript, although promoter activity was never shut down completely, even at concentrations as high as 500 nM Fur protein (monomer), which fully switched off the aerobactin promoter (13). This probably reflects the fact that the single iron box available for Fur binding at the fepA-fes region is certainly divergent from the consensus (Fig. 1), and the repressor is likely to have less affinity for this target. This is in accordance with previously published data from in vivo studies with gene fusions to report promoter activity (16). The quantification of the transcripts with different Fur concentrations (Fig. 2B) indicated that repression followed a two-step pattern; a significant decrease in promoter activity was achieved in the range of 100 nM Fur, while further repression required over 300 nM Fur. Interestingly, Fig. 2B shows also that, in spite of their different transcriptional rates, the two promoters were similarly inhibited by increasing concentrations of the repressor. This is somewhat intriguing, since repression efficiency is thought to be the result of the relative affinities of Fur and the RNAP for the same target sequences within iron-regulated promoters (13, 18). The issue is, therefore, how a single Fur site can coordinately regulate two promoters of different strengths. Since this cannot be sorted out with runoff experiments, we resorted to single-round transcription assays as described below.

RNAP and Fur compete for binding free promoters.

In order to determine at what level repression of the fepA-fes promoters by Fur-Mn²⁺ occurs, a series of transcription experiments was carried out in the presence of heparin, with the order of addition of the RNAP or Fur being changed in each case. For the single-round transcription assays whose results are shown in Fig. 3A, the first added protein was incubated with the DNA template for 5 min, followed by incubation with the second protein for another 5 min. Only then were the reactions initiated with a mixture of heparin and nucleoside triphosphates (NTPs) (13). The transcripts thus reflect faithfully the occupation of the promoters by the RNAP at the moment of addition of heparin plus NTPs. That the ratio between transcripts does not vary greatly as compared to the runoff experiment whose results are shown in Fig. 2 suggests that the two promoters were occupied independently by RNAP rather than binding in a mutually exclusive fashion. This notion is strengthened by the observations of Hunt et al. (16), who showed that a mutation in the −35 hexamer of fes (Fig. 1) which abolished transcription from that promoter did not result in an increased activity of fepA in vivo and vice versa. Nonexclusive occupation of the region by RNAP may thus occur by virtue of the opposite locations of the −35/−10 hexamers of the two promoters on the DNA helix in spite of their being virtually overlapping (Fig. 1; see below).

FIG. 3.

Single-round transcription of the fepA-fes promoters with varying Fur and RNA polymerase concentrations. (A) Fur and RNAP inhibit each other’s access to the promoters. The linear DNA template shown in Fig. 2A was added with the enzyme (80 nM) and the repressor (470 nM) in different orders. Samples were preincubated for 5 min at 37°C with one protein and then supplied with the second protein for a further 5 min and finally added with heparin and NTPs. Lane 1, RNAP alone; lane 2, first RNAP and then Fur; lane 3, premixed Fur and RNAP; lane 4, first Fur and then RNAP. The transcripts corresponding to each promoter are indicated. (B) Effect of increasing RNAP concentrations on Fur binding to the fepA-fes promoter region. The same DNA template was incubated with 470 nM Fur for 5 min and then treated with increasing RNAP concentrations as follows: lane 2, 80 nM; lane 3, 120 nM; lane 4, 160 nM; lane 5, 200 nM. Control sample 1 was added with only RNAP (80 nM, no Fur), while sample 6 had Fur, 80 nM RNAP, and 200 μM EDTA. After further incubation for 5 min, heparin and NTPs were added as before and samples were examined in a denaturing gel.

Under single-round transcription conditions, the two fes and fepA transcripts originated by RNAP alone were not affected significantly by the subsequent addition of Fur-Mn²⁺ (Fig. 3A, lanes 1 and 2). This indicated that Fur was unable to bind to its target sequence in the region when polymerase was engaged with the DNA. On the contrary, when Fur was added first, transcription was significantly and proportionally inhibited from both promoters fepA and fes (Fig. 3A, lane 4). Therefore, RNA polymerase seemed not to gain access to any of the promoters when Fur was bound to its target site. Finally, when both proteins were added simultaneously the two transcripts were repressed also (Fig. 3A, lane 3), albeit to a slightly lesser extent than when Fur was added first.

The results above are likely to reflect the competition of the two proteins to gain access to the same DNA sequence within the region but suggested also that Fur binding cannot be displaced by the RNAP. To clarify this issue, we set up a competition assay in which we added increasing RNAP concentrations to a DNA template prebound by Fur (Fig. 3B). The results clearly showed that RNAP by itself can hardly compete Fur out of the region. However, addition of EDTA restored production of fepA-fes transcripts to the levels of the positive control without Fur. This strengthens the notion that only the presence or absence of divalent cations and not the changes in RNAP versus Fur concentrations is translated into transcriptional repression or derepression of the system.

Fur replaces RNAP at the fes promoter during successive rounds of transcription.

The results presented in Fig. 3A indicate that the repressor cannot displace RNAP when it is bound to the region forming open complexes. This is, however, a very transient stage in vivo, and we wondered what is the situation in this respect when RNAP escapes to elongation over successive rounds of transcription, i.e., in conditions resembling the situation in vivo. To explore this issue, we carried out the experiment whose results are shown in Fig. 4, to monitor the occupation of the fepA-fes region by Fur over successive reinitiation rounds. To this end, runoff assays were set up in which the DNA templates were added with RNAP for 5 min (in order to occupy the fepA-fes promoters) and then with Fur for 5 more min, after which the reaction was started with NTPs but no heparin. Under these conditions, Fur is given the chance to occupy its target sequence as soon as RNAP initiates elongation and leaves the promoter sequences. The results shown in Fig. 4 indicated that Fur fully blocks transcription from the fes promoter (the one examined in the experiment) shortly after the first initiation round(s). This is deduced from the remarkable increment in transcript synthesis over time of the sample with RNAP only (Fig. 4, lanes R) versus the sample with both RNAP and Fur (Fig. 4, lanes RF). This observation, along with the results shown above, suggested that in the presence of a divalent ion the RNAP is fully and quickly replaced by Fur, thereby providing a rationale for the quick responses to iron concentration in vivo (3, 9).

FIG. 4.

Occupation of the fes promoter by Fur-Mn²⁺ over successive transcription rounds. The DNA template described in the legend to Fig. 2A was added either with 80 nM RNAP alone (marked on top of the gel with an R) or with 80 nM RNAP for 5 min and then with 470 nM Fur (marked RF). Mixtures were preincubated at 37°C for 5 min before initiation of the reaction with NTPs without heparin to allow reinitiation. Transcription was stopped at the different times indicated, and the samples were processed as before.

Regulation of two promoters by a single Fur target site.

It is generally believed that the phenomenon of transcriptional downregulation by classical prokaryotic repressors is the result of the relative affinities of RNAP and the regulatory protein for a mutually exclusive target site (18, 22). In spite of the growing number of exceptions, the mechanism of metalloregulation of the aerobactin promoter by the Fur protein adheres faithfully to this rule (13). In this case, competition between the RNAP and the repressor for their respective overlapping binding sites, and not any other effect, determines how efficiently promoter activity is controlled by the Fur protein. However, the aerobactin promoter is intrinsically very strong and the regulation very tight (3). In other cases, the promoters may require a different window of activity and regulation, so that moderate iron control might be superimposed onto otherwise constitutive promoters, each of different strengths. This is the case with the divergent fepA-fes promoters, the iron regulation of which seems to be less stringent than that of the aerobactin promoter in vivo (3, 16) and in vitro (reference 13 and this work). From the data presented above, it appears that regulation of weak, partially constitutive promoters by the Fur protein can be efficiently achieved not only by the presence of suboptimal Fur-binding sites within the regulatory region but also through the physical association of the weaker promoter (i.e., fepA) with a stronger promoter located on the opposite side of the DNA helix (fes). In this way (Fig. 5), the regulation of the weaker promoter may rely on the activity of the RNAP which competes effectively with the repressor for the stronger promoter. In this respect, the situation at the fepA-fes region is quite different from that of other divergent promoters, the geometry of which determines a transcriptional switch and a mutual inhibition between them (4). These structural assets may allow iron-regulated promoters to adjust the response of specific genes to the precise level of expression, whether for production of high-affinity metal transport systems or for factors that help bacteria to thrive in iron-limited hosts (3).

FIG. 5.

Alternate occupation of the fepA-fes region by either RNAP or Fur-Mn²⁺. The DNA segment spanning the sequence shown in Fig. 1 is represented in space on the basis of Trifonov’s algorithms (6) for DNA bending. According to the results presented in this work, the region can be occupied simultaneously (albeit with different efficiencies) by RNAPs which bind different sides of the DNA helix and give rise to fes and fepA transcripts. In the presence of Fe²⁺ or Mn²⁺ and as soon as elongation occurs, the same region may be occupied by a multimer of the Fur protein which nucleates to the sides of the protein bound to the iron box sequence overlapping the promoters (the stoichiometry of the binding remains uncertain). The arrows indicate the direction and the position (+1) of each of the two transcripts fes (upwards) and fepA (downwards).