SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. Supplementary material
  9. References
  10. Supporting Information

The σs subunit of RNA polymerase (RNAP) is the master regulator of the general stress response in Escherichia coli. Nevertheless, the selectivity of promoter recognition by the housekeeping σ70-containing and σ5-containing RNAP holoenzymes (Eσ70 and Eσs respectively) is not yet fully clarified, as they both recognize nearly identical −35 and −10 promoter consensus sequences. In this study, we show that in a subset of promoters, Eσs favours the presence of a distal UP-element half-site, and at the same time is unable to take advantage of a proximal half-site or a full UP-element. This is reflected by the frequent occurrence of distal UP-element half-sites in natural σs-dependent promoters and the absence of proximal half-sites. Eσ70, however, exhibits the opposite preference. The presence of the −35 element is a prerequisite for this differential behaviour. In the absence of the −35 element, half or full UP-element sites play no role in sigma selectivity, but the distal subsite leads to an equivalent, if not greater, transcriptional stimulation than the proximal one for both sigma factors. Finally, experiments using single amino acid substitutions of σs indicate that the foundation for this preference lies in an inability of σs to interact with the a subunit C-terminal domain.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. Supplementary material
  9. References
  10. Supporting Information

Sigma factors act as central junctions for switching gene expression in prokaryotes. The σs (RpoS) subunit of Escherichia coli RNA polymerase (RNAP) is induced upon entry into stationary phase and in stressful conditions and controls the expression of more genes than any other sigma factor, apart from the vegetative σ70. These two sigma factors show a high degree of sequence similarity and were found to recognize almost identical core promoter elements (Gaal et al., 2001), although they control clearly different sets of genes. Only recently has some light been shed on this paradox, with ‘minor’ sequencial and structural features of a promoter (Becker and Hengge-Aronis, 2001; Bordes et al., 2003), as well as trans-acting factors (Colland et al., 2000; Germer et al., 2001), being recognized as contributors to σs-selectivity (for review, see Hengge-Aronis, 2002).

Some bacterial promoters also contain an additional element, upstream of −35, that aids the reconstituted RNAP to bind the promoter and thus increase transcription dramatically: the UP-element (for review, see Gourse et al., 2000). The C-terminal domains of the two α-subunits (αCTDs), tethered with a flexible linker to the N-terminal domains (αNTDs) which form part of the core RNAP, are able to make sequence-specific interactions with the UP-element. Furthermore, an UP-element consists of two independent, functional subsites, referred to as proximal and distal. Each αCTD can bind autonomously to either of the two subsites and the consensus sequences for both the full site and the subsites were identified (Estrem et al., 1998; 1999). In the context of σ70-dependent promoters, the presence of the proximal subsite leads to clearly higher levels of transcriptional activation than the distal subsite (Estrem et al., 1999).

In a recent study (Germer et al., 2001), the σs-dependent gene csiD was shown to contain a distal UP-element half-site, directly downstream of a CRP box (centred at −68.5). This configuration was shown to favour σs-containing RNAP holoenzyme (Eσs)-mediated transcription, whereas the completion of the UP-element allowed σ70-containing RNAP holoenzyme (Eσ70) to acquire a greater role in the gene's transcription.

Approximately 35% of the known σs-dependent promoters seem to contain sequences similar to a distal UP-element, whereas only 6% have a full UP-element and none a proximal subsite (TableS1). From this the question arises whether this preference of Eσs is general, and which regulatory mechanism it reflects.

Here, by using several natural and synthetic promoters that utilize Eσ70 and Eσs to a different extent, we were able to demonstrate that distal UP-element subsites are used more efficiently by Eσs, whereas full UP-elements or proximal sites favour Eσ70-mediated transcription. Furthermore, by using appropriate σs variants, we investigated the mechanism that underlies this behaviour. Finally, in contrast to the situation at the strong ribosomal promoters with different UP-element half or full sites, we demonstrated that in the absence of a −35 hexamer, a distal subsite enhances transcription equally or even more efficiently than the proximal one, no matter which holoenzyme is being used.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. Supplementary material
  9. References
  10. Supporting Information

A subset of Eσs-dependent genes uses a distal UP-element half-site to establish selectivity against Eσ70

From the σs-dependent genes containing a distal subsite (TableS1; at least 7/11 matches to the consensus) three genes with relatively well-characterized promoter regions were selected: bolA (Lange and Hengge-Aronis, 1991a; Santos et al., 1999), csiE (Marschall and Hengge-Aronis, 1995) and osmY (Colland et al., 2000). Point mutations were introduced in their promoters in order to obtain completed or defective UP-elements (Table 1). The promoters were then fused to lacZ and the fusions were crossed in single copy into the chromosomal att site. csiE is known to possess a functional CRP box, centred at −59.5, which partially overlapped to the distal UP-element subsite (Marschall and Hengge-Aronis, 1995 and data not shown). Therefore, care was taken not to destroy or optimize the second poorly conserved part of the CRP-recognition site in the overlapping region (cCAac instead of the consensus sequence TCACA; underlined is the region overlapping with the UP-element).

Table 1.  Mutations introduced (bold) into the putative UP-elements in the bolA P1, csiE and osmY promoters.
ConsensusUP-element −35 −10
DistalProximal
  1. R stands for A/G, W for A/T. The −35 and −10 boxes for each gene are indicated and the matches to the consensus are underlined. For osmY, a new DNA consensus sequence (GCGG; also underlined) is proposed by Lee and Gralla (2004), to be utilized by region 4.2 of σs.

Model (full)NNAAAWWTWTTTTNNAAAANNNTTGACA17 bpTATAAT
 Distal subsiteNNAWWWWWTTTTTN  
 Proximal subsite AAAAAARNR
rrnB P1AAAATTATTTTAAATTTCCCATTGTCA16 bpTATAAT
bolA P1GGTAAATATTTGTTGTTAAGCTGCAA19 bpTAGTAT
 No UP-elementGGTAGACAGCTGTTGTTAAG
 Full UP-elementGGTAAATATTTGTAGAAAAG
csiECAACATTTCTGATGATTAGCTTCCCT16 bpCAAACT
 No UP-elementCAACAGTTCGGATGATTAGC
 Full UP-elementCAACATTTCTGATAAAAAGC
osmYCTTATGTTTTCGCTGATATCCGAGCGG16 bpTATATT
 No UP-elementCTTAGGCTGTCGCTGATATC
 Full UP-elementCTTATGTTTTCGAAAAAATC

In all three promoters, expression was severely reduced when the distal UP-element subsites were eliminated, and significantly stimulated when they were completed, revealing that these motifs indeed act as functional sites (Fig. 1A, C and E). In the case of the csiE promoter, the same result was obtained both in a crp background or when the CRP binding site was mutated (data not shown). This confirmed the presence of a functional distal UP-element half-site that αCTD can bind to, independently of cAMP-CRP. What interested us further was whether the completion or deletion of these distal subsites resulted in any changes in promoter selectivity for Eσ70 or Eσs. In general, if a feature (such as completion of the UP-element) contributed to such changes, then the extent of its effect would vary in a rpoS+ or rpoS background.

image

Figure 1. Effect of UP-element mutations on in vivo expression of bolA (A and B), csiE (C and D) and osmY (E and F). Expression of single-copy lacZ protein fusions, carrying either the wild-type promoter (WT, squares) or a promoter with the completed or deleted UP-element (‘full-UP’, circles, and ‘no-UP’, diamonds respectively) was determined in rpoS+ and rpoS backgrounds. Cells were grown in LB medium and optical densities and specific β-galactosidase activities were measured along the growth curve. Note that the scale of the β-galactosidase axes changes significantly for each gene in rpoS+ and rpoS background, as all of them are strongly σs dependent.

Download figure to PowerPoint

In the absence of σs (i.e. the rpoS mutant), the full UP-element construct of csiE could be expressed by the housekeeping Eσ70 alone (Fig. 1D) at levels higher than the wild-type csiE promoter in a rpoS+ environment (Fig. 1C). Moreover, the expression levels in a rpoS background reached only 20% of the overall expression (rpoS+ background) for the wild-type csiE promoter, but about 50% for the ‘full-UP’ construct, indicating the preference of Eσ70 for the full UP-element. This preference was maintained also in a crp background (data not shown), demonstrating again that CRP binding does not interfere with the use of the UP-element. On the contrary, optimizing the distal half-site to a full site in the bolA promoter stimulated expression in rpoS+ and rpoS strains to a similar extent (Fig. 1A and B), which means that the promoter selectivity for Eσ70 or Eσs remained unaffected. Finally, in the absence of σs, the activity of the osmY promoter was strongly reduced (compare Fig. 1E and F). But even with the low residual activity dependent on Eσ70, it could be seen that the relative contributions of various forms of the UP-element were different for Eσ70. Only the full UP-element could significantly stimulate expression in this case (Fig. 1F), whereas in the presence of Eσs, also the wild-type configuration, i.e. the distal half-site, contributed to stimulation (Fig. 1E). We conclude that the distal UP-element half-sites, present in these three promoter regions, activate expression, and that these half-sites can also contribute to selectivity (as shown for csiD, csiE and osmY), but do so in a way that seems promoter context dependent.

But then in what aspect do those three genes differ from bolA? csiD and csiE possess functional −35 elements, and osmY seems to have a DNA sequence replacing the −35 element that possibly also aids the recruitment of region 4.2 of σs (GCGG; Lee and Gralla, 2004; see also Table 1). On the contrary, bolA does not exhibit an optimally positioned −35 hexamer to aid transcriptional stimulation. A similar degenerate −35 region, positioned 19 bp upstream of the −10 hexamer, was found to be even slightly inhibitory for Eσs activity at σs-dependent synthetic promoters (A. Typas, J. Heuveling and R. Hengge, in preparation).

In order to study the role of a distal UP-element half-site further (and in a less complex sequence environment), a series of synthetic promoters (synp7, 42, 4 and 213) lacZ fusions were chosen. Those promoters derived from the ptac promoter and possessed stepwise added σs-selective features, which allowed different levels of Eσs-mediated expression and selectivity (Table 2). synp213 possesses a −35 region, but was optimized for activating Eσs in the −10 region. The other three promoters were all devoid of a −35 element (a feature commonly thought to enhance Eσs selectivity) and carried different Eσs selectivity enhancing elements in the −10 region. Distal, proximal or full UP-element sites were implanted directly upstream of all these promoters (see Table 2 for sequences).

Table 2.  Sequences of synthetic promoters and different UP-element configurations added in front of each of them.
UP-element−35 −10 +1 
AAATTTTTTTTCGAAAACCC Full UP-element
AAATTTTTTTTCGGGTGACTDistal half-site
TTCAAGGATCCAAAAAAGTAProximal half-site
CCTTTCGTCTTCAAGGATCCNo UP-element
 TTGACAATTAATCATCGGCT CGTATAATGTGTGGAptac
TTCACAATTAATCATCCGGCTCCTATAATTAATAGAsynp213
GCTCGTATTAATCATCCGGCTCGTATACTGTGTGGAsynp4
GCTCGTATTAATCATCCGGCTTGTATACTGTGTGGAsynp42
GCTCGTATTAATCATCCGGCTTCTATACTGTGTGGAsynp7

For synp213, the core promoter (i.e. no UP-element present) was utilized about twofold better in the presence of rpoS, in stationary phase (Fig. 2A), i.e. Eσ70 and Eσs both contribute to expression. Adding a proximal half-site or a full UP-element consensus to the promoter resulted in higher expression levels, but expression became almost completely Eσ70-dependent (Fig. 2B and D). On the contrary, the promoter carrying a distal UP-element half-site exhibited significant activity only in the presence of Eσs (Fig. 2C), indicating a strong preference for Eσs. The same pattern of behaviour was also observed in an in vitro transcription assay, with Eσs using more efficiently the promoter carrying the distal subsite, whereas Eσ70 was more efficient in using the proximal and ‘full-UP’ promoters (Fig. 3). Lower in vitro ratios should be attributed to the inability of perfectly imitating the in vivo conditions and to the absence of competition of sigma factors for core RNAP or of the two holoenzymes for the promoter. We conclude that in a promoter that can be used equally by Eσs and Eσ70, the introduction of a distal UP-element half-site can generate Eσs selectivity, whereas introducing a proximal subsite strongly shifts the balance in favour of Eσ70-mediated expression.

image

Figure 2. Effect of adding an UP-element full site or half-sites to a synthetic promoter (synp213) containing a −35 box. Expression of single-copy lacZ operon fusions was determined in rpoS+ and rpoS backgrounds. Cells were grown in LB medium and specific β-galactosidase activities were measured along the growth curve.

Download figure to PowerPoint

image

Figure 3. Effects on in vitro transcription of UP-element full or half-site additions to the synthetic promoter synp213. Reconstituted RNAP, containing either purified σs or σ70, was used to transcribe the different derivatives of the synp213 promoter and the transcripts ending at the rrnB (T2) terminator are shown (top). No apparent transcript ending at rrnB (T1) was observed for any construct, indicating that the RNAP was reading through it. The RNA I transcript, encoded by the vector, is clearly better transcribed by Eσ70 under the conditions used, as previously observed (Gaal et al., 2001). The Eσ70-derived transcript for each promoter is set to 100%. Average values of three experiments are shown. Error bars indicate standard deviations from all these experiments.

Download figure to PowerPoint

Strikingly, this effect was not observed with the synthetic promoters lacking a −35 element (Fig. 4). Introducing UP-element half-sites or full sites to them (see Table 2) produced significant activation, but did not alter selectivity for σ70 or σs. On the other hand, the addition of a distal subsite, regardless of the presence of σs or not, resulted in similar if not higher stimulation of promoter activity than that of the proximal subsite, contrary to what is already known for ribosomal core promoters (e.g. rrnB P1), carrying partial or full UP-elements (Estrem et al., 1999). The presence of a −35 box in the rrnB P1 promoter may provide an explanation for this opposing behaviour. It is likely that binding of σ70 region 4.2 to a −35 box is a prerequisite to establish a functional α–σ70 interaction, when αCTD binds to an UP-element proximal subsite (Ross et al., 2003). To further strengthen this assumption, a plasmid carrying the E261A mutant of rpoA (Gaal et al., 1996) was transformed into the different strains carrying the gene fusions in rpoS environment. This amino acid substitution of α-subunit significantly reduces the α–σ70 interaction, leading to reduced proximal half-site function (Ross et al., 2003). For our −35 boxless promoters used here, no change of Eσ70-dependent expression was observed with this mutation in αCTD (i.e. in rpoS background; data not shown). In the rpoS+ background the E261A allele has an indirect effect on rpoS expression itself (A. Typas and R. Hengge, unpubl. data). This indicates that the α–σ70 interaction does not play role at our −35 box-devoid promoters.

image

Figure 4. Effects of adding UP-element full or subsites to synthetic core promoters lacking a −35 box. Cells were grown in LB medium and specific β-galactosidase activities were measured during middle stationary phase (4–8 h after entry in to stationary phase, when the activities remained constant). Average values of five to seven experiments are shown. Error bars indicate standard deviations from all these experiments.

Download figure to PowerPoint

To summarize, it is apparent that in the presence of a −35 element, a distal UP-element half-site can strongly enhance Eσs selectivity. In contrast, a proximal half-site or a full UP-element not only strongly stimulate promoter activity, but also establish nearly complete Eσ70 preference. The observation that these changes in Eσs/Eσ70 selectivity were dependent on the presence of a −35 region suggested that a differential α–σ interaction may be involved in the usage of UP-elements by Eσ70 and Eσs.

Differences in region 4 of σs and σ70 are responsible for their differential use of UP-element full and half-sites

Although the amino acids directly involved in the interaction between region 4.2 and the −35 element are conserved in σ70 and σs, high divergence between σs and σ70 can be detected in the C-terminal part of region 4. Especially the high tendency for basic residues in the second α-helix of region 4.2 of σ70 is not conserved in σs(Fig. 5). Interestingly, many of these amino acids were found to play an active role in α–σ70 interaction which contributes to transcription factor-independent or -dependent transcription (Chen et al., 2003; Ross et al., 2003). In addition to this, these amino acids are also involved in transcription factor to sigma direct interaction (Lonetto et al., 1998; Bhende and Egan, 2000; Rhodius and Busby, 2000; Campbell et al., 2002; Nickels et al., 2002; Grainger et al., 2004). In order to test the effect of different σs single amino acid substitutions in this region of σs, the corresponding plasmid-encoded σs variants were isolated (Fig. 5) and then used in a rpoS background. These σs variants were tested for different UP-element configurations in a natural and a synthetic promoter (csiDp and synp213), both carrying −35 regions.

image

Figure 5. Alignment of region 4.2 of σs (RpoS) with the corresponding region of σ70 (RpoD). Amino acids that were shown to be important for the α–σ70 interaction are underlined. Single amino acid substitutions of σs to the corresponding residue of σ70 used in the present study are shown in bold, and are also indicated below the alignment. The positions of the helices in region 4.2 are according to Campbell et al. (2002).

Download figure to PowerPoint

As previously reported (Germer et al., 2001), Eσ70 could utilize relatively better the csiD promoter when it carried a whole UP-element (Fig. 6A and C). Introducing Q306E into σs (the corresponding residue in σ70 is E591) allowed Eσs to use the ‘full-UP’csiD promoter more efficiently, in a similar manner as Eσ70 (Fig. 6C). At the same time, this allele slightly reduced the ability of Eσs to use a csiD promoter lacking an UP-element (Fig. 6B). E591 in σ70 is buried inside the structure and makes a polar interaction with R562, which is involved in −35 recognition (Nickels et al., 2002). Thus, the Q306E may indirectly facilitate σs to bind to the −35 box in a σ70-like way, such that the αCTD bound to the proximal UP-element half-site could be used more efficiently. It is worth mentioning that no other variants of σs (E308K, R312K, E315H, I316P, Q318R and Q320E) enhanced its ability of using the csiD promoter carrying a full UP-element. Only the E308K variant showed significantly higher expression from the csiD promoter lacking an UP-element (data not shown).

image

Figure 6. The Q306E substitution enables σs to adopt a σ70-like transcriptional behaviour, with respect to the full UP-element configuration in the csiD promoter. Expression of single-copy csiD::lacZ protein fusions, carrying either the wild-type promoter (A), or promoters where the UP-element was eliminated (B) or completed (C), was determined in the rpoS background, in which σs and σs(Q306E) were expressed under the ptac control from a plasmid. Cells were grown in LB medium (no inducer was added to establish physiological levels of σs) and optical densities and specific β-galactosidase activities were measured along the growth curve.

Download figure to PowerPoint

As a next step, these σs variants were tested for their ability to use a proximal or a distal UP-element half-site present upstream of the synthetic promoter synp213. Interestingly, the amino acid substitutions (E315H, Q318R) found to allow better use of a proximal site (Fig. 7A) were the same that reduced the preference for a distal one (Fig. 7B), but are different from those mentioned above for playing a role in csiD expression (Fig. 6). The corresponding residues in σ70, H600 and R603, were both implicated in α–σ70 interaction (Ross et al., 2003). Moreover, the effects of the two single substitutions was additive, when combined (Fig. 7). Other σs variants (Q306E, E308K, R312K and Q320E) neither provided any enhancement in its ability of using the proximal subsite, nor did affect its ability to take advantage the distal half-site (data not shown). However, when combining E315H and Q318R with a third amino acid substitution, E308K (not playing any role by itself), the negative influence on the promoter with the distal half-site was further increased (Fig. 7B). Finally, it was noted that the I316P and a C-terminal truncation of σs (carrying the first 321 amino acids only) resulted in a severe overall defect in σs-dependent transcription (data not shown).

image

Figure 7. In synthetic promoters with a −35 element (derivatives of synp213, carrying a proximal or a distal UP-element half-site) single amino acid substitutions in region 4 of σs allow better use of a proximal UP-element half-site, but at the same time diminish the preference for a distal subsite. Expression of single-copy lacZ protein fusions was determined in rpoS background, using plasmid-encoded σs variants. Cells were grown in LB medium and specific β-galactosidase activities were measured along the growth curve.

Download figure to PowerPoint

In conclusion, the ability of σs to take advantage of a distal UP-element half-site is reduced when σs is made more ‘σ70-like’ by introducing specific amino acids known to be involved in the αCTD–σ70 interaction. At the same time its ability to benefit from a proximal UP-element half-site is enhanced.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. Supplementary material
  9. References
  10. Supporting Information

This study started with the observation that many native σs-dependent promoters seem to carry distal UP-element half-sites, whereas proximal half-sites seem to be absent (TableS1). From this, two questions arose: (i) are these putative UP-element half-sites functional, i.e. contribute to expression, and (ii) do these elements contribute to σs-selectivity of the promoters, as suggested by preliminary data obtained with the csiD promoter (Germer et al., 2001)?

For three natural promoters, we have shown here that these sites are truly functional, as transcriptional activity is significantly deteriorated, when these sites are eliminated by site-directed mutagenesis (Fig. 1). In addition, we observed that a distal UP-element half-site can contribute to Eσs selectivity at a promoter, provided that a functional −35 element is present. If a promoter contains a −35 box, the distal UP-element half-site establishes a strong preference for Eσs, whereas a proximal subsite or full UP-element strongly favours Eσ70 (Figs 2 and 3).

In contrast to this, in the absence of a −35 region, the UP-element, no matter in which configuration, does not affect Eσs selectivity, but activates the promoter in a sigma factor-independent manner. In all cases tested, the distal half-site activates equally or even better than the proximal one (Fig. 4). At a first glance, this seems surprising, because at the ribosomal promoters, where UP-elements have been studied extensively, a proximal UP-element subsite leads to significantly stronger activation (Estrem et al., 1999). However, ribosomal promoters contain −35 regions, which allow the extensive use of full UP-elements or proximal subsites, as an α–σ interaction is possible. Thereby, these promoters develop strong Eσ70 dependency and reach maximal activity under conditions of rapid growth and in the absence of any stress. Interestingly, ribosomal core promoters possess features that could lead to their partial utilization of Eσs: (i) a C(−13), which by interacting with K173 in σs, promotes σs-selectivity (Becker and Hengge-Aronis, 2001), and (ii) a 16 bp spacer between −10 and −35, which is tolerated better by Eσs than by Eσ70 (A. Typas, J. Heuveling and R. Hengge, in preparation). Nevertheless, the presence of full UP-elements, strong proximal subsites and other factors seems to counteract this tendency. Moreover, σs is hardly present in rapidly growing cells.

On the other hand, many promoters that have to be active under conditions of slow or no growth have to be Eσs-selective, as Eσ70 is also present at the same time. This can be achieved, for example, by degenerating the −35 box, as Eσs is less dependent on it (Gaal et al., 2001; Lee and Gralla, 2001), in combination with a C(−13) (Becker and Hengge-Aronis, 2001). Alternatively, if a Eσs-selective promoter has to be relatively strong, the −35 region can be maintained, but combined with a distal UP-element half-site. It is worth mentioning that the maintenance of the −35 region is not so uncommon for σs-dependent promoters as it was formerly proposed (Lee and Gralla, 2001). It seems that up to approximately 75% of σs-dependent genes do possess a −35 hexamer, but many times not optimally spaced from the −10 box (A. Typas, J. Heuveling and R. Hengge, in preparation). Taken together, it seems that by combining several features of a promoter in a modular way, Eσs selectivity can be achieved in different ways.

How can the clearly differential use by Eσs and Eσ70 of a proximal or a distal UP-element half-sites in various promoter configurations be explained in mechanistic, i.e. molecular terms? The absence of a −35 hexamer presumably leads to a minor displacement of σ, and thus to a loss of the suitable orientation of the α–σ interface. This assumption is confirmed by the finding that the αE261A mutant does not affect Eσ70-dependent transcription under these conditions (data not shown). This αCTD substitution strongly affects the α–σ interaction, especially for promoters carrying a proximal UP-element subsite (Ross et al., 2003). Thus, it seems reasonable to assume that the loss of the α–σ interaction accounts for the reduced stimulation of Eσ70 by a proximal UP-element. On the other hand, the function of the distal subsite remains unaffected (if not positively affected) by a loss of the α–σ interaction (Ross et al., 2003), and therefore the stimulatory effect in promoters lacking a −35 element is high for both RNAPs (Fig. 4). Considering the fact that a distal subsite requires the binding of both αCTDs for efficient transcription (Estrem et al., 1999), it seems probable that this extensive complex formation offers an overall better recruitment and placement of RNAP to promoters where sigma lacks an optimal −35 binding.

If in contrast, a −35 hexamer is available, a σs-selective promoter prefers a distal UP-element subsite and cannot take advantage of a full site or a proximal half-site (Figs 2 and 3). This inability to use a proximal site is partially restored when the corresponding amino acids of σ70 that are mainly responsible for the α–σ interaction are introduced into σs (Ross et al., 2003). R603, H600, K597 and K593, playing an active role in this interaction for σ70 (with R603 being the most prominent and only the first three amino acids being directly involved), are all positively charged and engaged in a polar interaction to E259 and E261 of the α-subunit (Ross et al., 2003). In σs, the corresponding residues are either neutral (Q318) or negatively charged (E315, E308), i.e. σs is highly unlikely to interact with αCTD in a similar manner. Therefore, the Q318R and E315H substitutions in σs probably tend to establish a σ70-like interaction with αCTD, which would explain the improved use of the proximal UP-element subsite (Fig. 7A). The fact that even under these conditions σs cannot use the promoter equally well as the vegetative sigma factor may result from a less efficient −35 binding of σs (Colland et al., 1999), and thus a pre-existing less optimal α–σ interface.

But why can Eσs use a distal UP-element subsite better than Eσ70 does? One possibility is that by the binding of both αCTDs, one at the distal subsite consensus and the other directly downstream of it, a more rigid binding is offered to Eσs, which is less tightly bound at the −35 consensus. Another option derives from the observation, that in the crystal structure of the αCTD–DNA complex, two αCTDs can be accommodated at closely adjacent positions, along the minor groove, in a 11 bp long A/T-rich DNA fragment (Benoff et al., 2002). Ross et al. (2003) speculate that even at some promoters that contain both UP-element subsites, transcription might be favoured by having both αCTDs positioned at the distal subsite. They also propose a competition between two situations: both αCTDs bound at the distal half-site or one bound at each half-site. In our case, we picture it more like equilibrium between the two situations (that would fit more to the hydroxyl radical footprint protection patterns in Estrem et al., 1999). In this scenario, Eσs may prefer both αCTDs being positioned at the distal half-site, and especially in promoters with a distal half-site, only this situation would be favoured. Completing the UP-element or introducing mutations that improve α–σ interaction would aid one αCTD to bind to the proximal site and therefore reduce or even eliminate Eσs selectivity (Figs. 2D and 7B).

The distal subsite in our synthetic promoter clearly reduced Eσ70-dependent expression (compare Fig. 3A and C). This is in contrast to the stimulatory effect reported for a optimal distal subsite in a ribosomal promoter context (4- to 16-fold; Estrem et al., 1999). This may result from the specific and complex design of the ribosomal promoters (e.g. Murray et al., 2003; Lew and Gralla, 2004), aiming to produce maximal activity. The possibility that the addition of a distal UP-element subsite accidentally produced an inhibitory H-NS binding site (H-NS binds to A/T-rich regions) was excluded, as in a hns background the ratio between expression in the presence or the absence of Eσs remained the same (data not shown).

In transcriptional activator-dependent UP-element utilization, the α–σ interface seems to change slightly, and the main amino acid of σ70 responsible for the interaction to E261 of α-subunit is K596, which is also conserved for σs (K311; Chen et al. 2003). This could explain why amino acid substitutions Q318R and E315H do not affect significantly transcription from the cAMP-CRP-stimulated csiD promoter carrying a full UP-element (data not shown). Moreover, the absence of basic residues in the C-terminal of region 4.2 of σs could lead to a differential ability to use not only the UP-element, but also adjacently situated transcriptional activators (class II). Actually, the bacteriophage protein λcI can only serve as an activator for the vegetative sigma factor and not for σs, unless an E308K substitution is introduced into σs (Nickels et al., 2002). It is tempting to speculate that on the other hand activators may exist that function exclusively with σs, and further research is conducted towards this direction.

In conclusion, the current study assigns an additional role to UP-elements, besides transcriptional stimulation: they can affect sigma selectivity. Moreover, this study provides additional evidence for a different activation mechanism for the proximal and the distal UP-element subsite. In the future, it might be interesting to extend these studies also to other bacteria, e.g. to some Gram-positive, where UP-elements seem to be more abundant (Helmann, 1995; Fredrick and Helmann, 1997).

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. Supplementary material
  9. References
  10. Supporting Information

Bacterial strains and plasmids

The strains used in this study are MC4100 derivatives (see TableS2a). All the reporter gene fusions are located in single copy at the chromosomal att site. Thus, phenotypically all the strains correspond to MC4100, except those in which rpoS mutations were introduced with P1 transduction (rpoS359::Tn10; Lange and Hengge-Aronis, 1991b). The region of the gene fused to lacZ, mentioned in parenthesis, refers to the translational start as +1. Mutations introduced at the promoter region for completing, deleting or adding UP-element sites are described below.

Most of the reporter fusions are translational (protein fusions), as these produce higher β-galactosidase activities and the effects are more easily visible, but in the case of the UP-element derivatives of the synp213 promoter, transcriptional fusions were also constructed, as the strain carrying the translational fusion of the full-UP construct was spontaneously picking up suppressor mutants that lowered the fusion activity.

The plasmids used in this study either are derivatives of pRL40.1 (rpoS cloned under the control of the ptac promoter in pRH800; Lange and Hengge-Aronis, 1994), or encode lacZ translational fusions, containing variants of a modified core ptac, followed by the 5′ untranslated region of osmY and its seven first codons. The osmY fragment was cloned in pJL29 (Lucht et al., 1994), using BamHI/HindIII. The detailed core promoters of these plasmids can be seen in Table 2. psynpT213, a suitable psynp213 derivative for in vitro transcription, was constructed by cloning the rrnB (T1,T2) terminator in the place of lacZ.

In addition to these plasmids, a plasmid was constructed for purifying non-tagged σs. rpoS was cloned directly downstream of a 6his-tag and a Xa recognition site in the pQE-30 Xa plasmid (Qiagen), using StuI and HindIII. The forward reading primer used for amplifying rpoS began with the ATG codon of rpoS, being 5′ phosphorylated, whereas the reverse one contained a HindIII position (for details, see the primer list in TableS2b).

Finally, all the plasmids containing a csiD::lacZ translational fusion (pJG13, pMM5 and pMM9; depending on whether they contain an UP-element) were constructed and described by Germer et al. (2001).

Site-directed/insertional mutagenesis

For mutagenesis of the distal UP-element in the bolA, csiE and osmY promoter region, a four-primer/two-step polymerase chain reaction (PCR) protocol was used (described in Germer et al., 2001). The mutations implied can be viewed in Table 1 (mutations introduced in bold, residues matching the consensus sequence underlined).

For insertion of UP-element full sites or half-sites (proximal or distal), a suitable primer was used each time, carrying at the 5′ end a BamHI restriction site and after that: (i) a full UP-element (AAATTTTTTTTCGAAAAccc), (ii) a distal subsite (AAATTTTTTTTcgggtgact) or (iii) a proximal subsite (AAAAAAGTA). The 3′ end of the primer carried the homologous region to the promoter.

In order to acquire single amino acid substitutions of rpoS, in the pRL40.1 plasmid, the four-primer/two-step PCR protocol was used again. The external primers were designed for producing a fragment, able to be cut with RsrII (inside rpoS) and HindIII (used for cloning rpoS in pRL40.1). For details, see the primer list in TableS2b.

Purified proteins for in vitro transcription

σ70 was purified as described previously (Gribskov and Burgess, 1983). RNAP core enzyme was purchased from Epicentre Technologies. σs was purified in exactly the same way like 6his-σs (Bouche et al., 1998), but after eluting the protein, the buffer was changed to that recommended by the manufacturer (Qiagen) and 1–2 U Xa protease per mg of protein per hour of reaction was used for cutting off the 6his-tag and the protease recognition signal. The protease, the undigested protein and the 6his-tag peptides were then removed, as described in the manual provided by the manufacturer. In order to obtain reconstituted RNAP holoenzyme, core enzyme was added to each sigma factor, present in 100-fold concentration to ensure saturation of the core enzyme even with the low-affinity binding σs (Colland et al., 2002), and incubated at 37°C for 20 min. The reconstituted RNAPs were then diluted to a solution containing 40 mM Hepes (pH 8), 10 mM MgCl2, 100 mM potassium glutamate, 2.5 mM DTT and 0.5% BSA, in order to obtain a final concentration of 18 nM (considering that the core RNAP is saturated with sigma factor)

In vitro transcription assay

For the in vitro transcription assays with the synthetic promoters carrying half or full UP-element sites, a rrnB (T1,T2) terminator was cloned in the place of lacZ in the plasmids, after the promoter. Transcription reactions contained 4 µl of 4.5 nM plasmid DNA in 40 mM Hepes (pH 8), 10 mM MgCl2, 100 mM potassium glutamate and 0.5% BSA, 4 µl of 18 nM reconstituted RNAP and 4 µl of rNTPS (1 mM ATP, CTP and GTP, 50 nM UTP, 0.8 µCi of [a-32P]-UTP). Reactions were initiated by the addition of rNTP and were subsequently terminated after 15 min at 25°C by the addition of loading buffer. Finally the samples were loaded in a 7% denaturing sequencing gel. Radiolabelled transcripts ending at the rrnB (T2) terminator were quantified on a Fuji phosphoimager using IMAGE GAUGE software.

β-Galactosidase assays

β-Galactosidase was assayed using o-nitrophenyl-β-d-galactopyranoside (ONPG) as a substrate and is reported as µmol of o-nitrophenol per minute per mg cellular protein (Miller, 1972). For all the fusion constructs, β-galactosidase activities were determined along the growth curve, but because of space restriction at some cases, only the data of middle stationary phase (4–6 h after entry) samples are presented (similar effects throughout the stationary phase).

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. Supplementary material
  9. References
  10. Supporting Information

We thank Rick Gourse for kindly providing a collection of plasmids carrying single-site mutations in rpoA, Richard Burgess for providing us with the plasmid for σ70 overproduction and purification, and Gisela Becker for kindly providing the plasmids containing the synthetic promoters. Financial support for this study was provided by the Deutsche Forschungsgemeinschaft (Gottfried-Wilhelm-Leibniz and He1556/12-1), the state of Baden-Württemberg (Landesfoschungspreis) and the Fonds der Chemischen Industrie.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. Supplementary material
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. Supplementary material
  9. References
  10. Supporting Information

Table S1. UP-element abundance in _s-dependent promoters. Table S2. List of strains and primers.

FilenameFormatSizeDescription
MMI_4382_sm_TableS1.doc32KSupporting info item
MMI_4382_sm_TableS2.doc40KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.