Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2009 Nov 10.
Published in final edited form as: Mol Microbiol. 2008 Aug 22;70(2):369–378. doi: 10.1111/j.1365-2958.2008.06412.x

General stress response signaling

unwrapping transcription complexes by DNA relaxation via the sigma38 CTD

Yi-Xin Huo 1, Adam Z Rosenthal 1, Jay D Gralla 1,*
PMCID: PMC2775543  NIHMSID: NIHMS73123  PMID: 18761624

Abstract

E. coli responds to stress by a combination of specific and general transcription signaling pathways. The general pathways typically require the master stress regulator sigma38 (rpoS). Here we show that the signaling from multiple stresses that relax DNA is processed by a non-conserved 8 amino acid tail of the sigma 38 C-terminal domain (CTD). By contrast, responses to stresses that accumulate potassium glutamate do not rely on this short tail, but still require the overall CTD. In vitro transcription and footprinting studies suggest that multiple stresses can target a poised RNA polymerase and activate it by unwrapping DNA from a nucleosome-like state, allowing the RNA polymerase to escape into productive mode. This transition can be accomplished by either the DNA relaxation or potassium glutamate accumulation that characterizes many stresses.

Keywords: osmY, poised RNA polymerase, sigma38, stress response, supercoiling

Introduction

In its natural environments E. coli must deal with many stresses, often simultaneously. These include the gastric stomach of certain hosts the high osmotic pressure of the lower gastrointestinal tract and other saline environments, extreme heat and cold, radiation challenges, and frequent nutritional limitation (Hengge-Aronis, 2002). The bacterium has evolved specific pathways that produce defined sets of protective proteins tailored for each challenge. These pathways are mostly mediated by transcription factors (Giuliodori et al., 2007). Such factors include macromolecules such as activators, repressors, and sigma factors, the latter for example for heat shock (Yura et al., 2007). The cell also uses small molecules to directly influence RNA polymerase in response to certain stresses. These include ppGpp for general nutritional stress (Gralla, 2005), glutamate for osmotic stress (Lee and Gralla, 2004) and acetate for volatile fatty acid stress (Rosenthal et al., 2006; Rosenthal et al., 2008b). The protective proteins produced by these various pathways include both unique and overlapping species (Weber et al., 2005).

In addition, E. coli and related bacteria have a general stress response. This is mediated by the general stress regulator sigma38 (rpoS, sigmaS) (Hengge-Aronis, 2002). Sigma38 levels are low in the rare stress-free environment and become higher in response to most stresses (Hengge-Aronis, 1996). After it accumulates sigma38 binds a fraction of the common core RNA polymerase and transcribes 2 sets of protective genes, those specific for the inducing stress and also for general stress adaptive proteins (Weber et al., 2005). The promoters for all of these protective genes include DNA elements that favor transcription by the sigma38 holoenzyme (Lee and Gralla, 2001; Typas et al., 2007). A number of these genes rely on promoter-specific DNA-binding regulators, which typically associate with upstream elements to assist in transcription in response to specific stresses (Sayed et al., 2007). By contrast, no master DNA-binding regulator has been implicated in the general stress response.

A well-studied general stress factor is the periplasmic protein osmY (Yim et al., 1994). Although originally studied in the context of osmotic stress (Ding et al., 1995), osmY is generally stress induced, for example by heat (Muffler et al., 1997), cold (Jones et al., 2006), acetate (Arnold et al., 2001), and acid stresses as well as during entry into stationary phase (Weber et al., 2005). In the absence of stress the osmY gene is silent, even when sigma38 is present. Under such conditions the osmY promoter can be occupied by sigma38 RNA polymerase, but the transcription complex is not active (Rosenthal et al., 2008a). In the case of osmotic stress the activating event is believed to be the accumulation of potassium glutamate. This has been reproduced in vitro on supercoiled DNA; the transcription complex is maintained in an apparently wrapped inactive state and then potassium glutamate unwraps the sigma38 RNA polymerase as transcription is activated (Lee and Gralla, 2004). However, many stresses that activate osmY transcription are not accompanied by the accumulation of potassium glutamate. How osmY and other general stress genes are activated in the absence of high levels of potassium glutamate is not clear.

One candidate effector for mediating general stress responses is DNA supercoiling, which changes during most stresses (Lopez-Garcia and Forterre, 2000; Rui and Tse-Dinh, 2003; Tse-Dinh et al., 1997). Changes in DNA supercoiling are known to effect the expression of large groups of genes (Cheung et al., 2003; Peter et al., 2004). The extent of supercoiling increases during osmotic (Bordes et al., 2003; Cheung et al., 2003) and cold stresses (Lopez-Garcia and Forterre, 2000), and decreases during nutritional deprivation (Balke and Gralla, 1987), stationary phase (Bordes et al., 2003; Reyes-Dominguez et al., 2003), heat shock (Lopez-Garcia and Forterre, 2000) and acid shock (Rui and Tse-Dinh, 2003). DNA supercoiling has been reported to have the potential to inhibit transcription from sigma38 promoters (Bordes et al., 2003; Kusano et al., 1996). This phenomenon is of special interest at the osmY promoter as poising of RNA polymerase (Rosenthal et al., 2008a) appears to involve wrapping of supercoiled DNA within the transcription complex (Lee and Gralla, 2004) and the removal of DNA supercoils could conceivably lead to changes in wrapping.

The activating events at osmY require altering a set of interactions that form a tight upstream network involving the C-terminal domain (CTD) of sigma38 and the DNA near the promoter-35 region (Kuznedelov et al., 2002; Lee and Gralla, 2004; Rosenthal et al., 2006). The CTD of sigma38 is unique among the sigma70 family of proteins and is non-conserved from near amino acid 315 to the terminus at position 330 (Ohnuma et al., 2000). A unique core-binding site is located near position 320 and this appears to be required for transcription of all promoters (Rosenthal et al., 2008a). However, truncation of the 8 terminal amino acids has no obvious effect on general transcription during osmotic shock, leaving the function of this unique region undetermined (Rosenthal et al., 2008a). This region is thus a candidate for mediating stress response mechanisms, a possibility that is tested in this study.

Here we show that this unique 8 amino acid tail of the CTD is needed for activation of osmY transcription by several stresses. In vitro data suggest that activation also involves the DNA relaxation that accompanies these stresses. This removal of supercoils is shown to trigger changes in wrapping of the DNA from around a poised sigma38 RNA polymerase, thus triggering its transcription. This activation by DNA unwrapping can also be accomplished by addition of glutamate, which contributes to transcription in response to other stresses where the DNA may remain supercoiled.

Results

The extreme CTD mediates the response to multiple, but not all, stresses

A prior study showed that removing the C-terminal 10 amino acids of sigma38 to create protein 1-320 reduced transcription of several promoters during osmotic stress conditions (Rosenthal et al., 2008a). The source of this was shown to be the loss of a site on the sigma that binds the core RNA polymerase. Shorter deletions, such as 1-322 had little effect on transcription either in vivo or in vitro with potassium glutamate present. To learn if the 8 amino acids absent in this protein play a role in responding to other stresses, osmY m-RNA was assayed in vivo under a wide variety of conditions. In these experiments osmY m-RNA levels were compared between full length protein and mutant 1-321 or 1-322 expressed from plasmids in a context lacking chromosomally expressed protein. The osmY m-RNA levels (undetectable in the absence of stress, not shown) were compared after 5 individually applied stress conditions: osmotic shock, acid shock, entry into stationary phase, heat shock and cold shock. The results are shown in figure 1.

Figure 1.

Figure 1

Open in a new tab

The sigma38 CTD contains different determinants for different stress responses. (A) osmY mRNA in vivo primer extension products after different stresses. The mRNA is expressed from cells containing wild type sigma38, sigma38 1-322 or sigma38 1-321 as indicated. (B) Alignment of C-terminal domains of sigma70, wild type sigma38 (1-330) and its two derivatives. The most conserved sequences are shadowed and the non-conserved extreme CTD tail is boxed. All experiments were in triplicate and the error is less than 10% of the value observed in all cases.

OsmY m-RNA was detectable under all stress conditions (lanes labeled wt). The amounts were high during osmotic, acid and cold shock, with lower levels observed after heat shock or during early stationary phase. When the CTD of sigma38 was truncated, the stress response pattern was not uniform (lanes labeled 321 and 322). As expected based on prior results, protein 1-322 supported a full level of osmY m-RNA during osmotic shock with protein 1-321 supporting perhaps a slightly reduced level. Roughly similar results were seen during acid shock, with no indication that these mutants were in any way defective in triggering osmY transcription. By contrast, proteins 1-321 and 1-322 showed clear defects in supporting osmY transcription during stationary phase, heat shock and cold shock. In these three cases removing the terminal 8 amino acids reduced transcription substantially and removing the terminal 9 amino acids reduced transcription further to very low levels. This effect is not due to low expression of mutant proteins (see Experimental procedures).

From these data we infer that the CTD contains determinants that mediate differential stress responsiveness. Three stresses require determinants in the 8 terminal amino acids and two other stresses do not. The two stresses that do not require this determinant have in common that they lead to the accumulation of potassium glutamate in cells. For acid shock, glutamate is the substrate for the Gad enzymes that protect against lower pH by consuming protons (Foster, 2004). Activators play a role in this regulation (Sayed et al., 2007) with an undefined role of potassium glutamate in regulation of transcription. In the case of osmotic shock, glutamate accumulation helps equilibrate the osmotic pressure and also signals the activation of the poised RNA polymerase at osmY (Lee and Gralla, 2004). For the 3 other stresses another stimulatory pathway would be required. Because stationary phase stress leads to DNA relaxation and heat shock can lead to transient DNA relaxation we explored the effects of removing supercoils in osmY activation.

The CTD tail of sigma38 and DNA relaxation jointly activate osmY transcription

There have been a number of prior reports of transcription of osmotic genes by sigma38 using either supercoiled or relaxed DNA (Bordes et al., 2003; Ding et al., 1995; Kusano et al., 1996). The templates used varied significantly as did the experimental conditions, including the amount of potassium glutamate used. The recent experiments showing glutamate activation of an inactive poised RNA polymerase at osmY in vitro were done using supercoiled DNA (Lee and Gralla, 2004). The collective interpretation of these experiments is complex as the mutual interactions of glutamate and DNA relaxation have not been established. The concentration of glutamate that exists physiologically is estimated to be substantial but probably 30-60 mM under normal conditions (Dinnbier et al., 1988; Richey et al., 1987) , rising to 400 mM during osmotic shock (Dinnbier et al., 1988). To attempt to mimic a range of in vivo conditions we used 0, 100 and 400 mM potassium glutamate to explore the effects of DNA relaxation of osmY transcription. These studies compared transcription on fully supercoiled DNA with that from fully relaxed DNA by cleavage with restriction enzyme.

Figure 2 shows the results of comparing supercoiled and relaxed DNA transcription over this range of potassium glutamate concentrations. When a basal level of glutamate is present (0 to 100 mM), the relaxed DNA signal is greater than that from supercoiled DNA. This response to DNA relaxation is roughly a 3-fold activation when glutamate concentrations are low. From this we infer that DNA relaxation can stimulate osmY transcription under conditions that correspond to unstressed levels of potassium glutamate accumulation, as exists during heat and stationary phase stresses. On the other hand, 400 mM potassium glutamate gives roughly an 8-fold activation when DNA templates are supercoiled, as exists during osmotic shock. As a control, 400 mM potassium glutamate totally inhibits osmY expression when DNA templates are relaxed.

Figure 2.

Figure 2

Open in a new tab

osmY expression in vitro can be activated by DNA relaxation when potassium glutamate concentrations are at basal levels (0 to 100 mM). The template DNA was supercoiled or relaxed and the potassium glutamate concentrations are indicated. The bands in the gels represent the in vitro transcribed osmY mRNA. The experiments were in triplicate and the results from one representative assay are shown.

If DNA relaxation is the physiological source of activation in the absence of high potassium glutamate levels it ought to have the same dependence on the sigma38 CTD as the two activating stresses tested above. To evaluate this idea, single-round transcription was done using the mutants 1-322 and 1-321 that are defective in responding to heat shock and stationary phase. Figure 3A (top row) shows that the loss of the 8 terminal amino acids leads to a severe reduction in osmY expression when the template is relaxed (7-8 fold, see figure 3b). To ensure that this reduction is not due to heparin inactivating unstable mutant open complexes, multiple-round transcription (no heparin present) was done (figure S1) and the results were the same. We note that the 2 mutant proteins are fully active in the elevated glutamate response pathway on supercoiled DNA. These quantitative in vitro data (figure 3b) are reminiscent of the in vivo data shown in figure 1 with respect to the acid, osmotic, stationary phase and heat shock responses. The acid and osmotic in vivo responses, which are the only ones to accumulate glutamate, are strongest and do not require the CTD terminus. The stationary phase and heat shock in vivo responses, which do not accumulate glutamate, are weaker and do require the extreme CTD. The cold shock response doesn’t fit neatly into either the DNA relaxation or glutamate pathway.

Figure 3.

Figure 3

Open in a new tab

The CTD is required for activation by DNA relaxation in vitro. (A) Wild type sigma38, sigma38 1-322 or sigma38 1-321 was used with core RNA polymerase to transcribe either relaxed or supercoiled osmY DNA as indicated. When supercoiled template was used, 400 mM of potassium glutamate was present. (B) Amounts of relative transcription from the average of four experiments.

The CTD tail functions in a step after initial bond formation during activation by DNA relaxation

Most genes that are sensitive to changes in DNA supercoiling are inhibited rather than activated by a reduction in supercoiling (Travers and Muskhelishvili, 2005). These are typically sigma70 dependent genes whose expression is limited by the ability of RNA polymerase to engage the promoter DNA in open complexes. This is because a loss of supercoils makes it more difficult to melt DNA to form the active transcription complex (see Travers and Muskhelishvili, 2007). The sigma38 dependent supercoiled osmY gene has been shown to be poised in an open complex prior to activation (Lee and Gralla, 2004) and it is this property that could make it selectively amenable to activation by DNA relaxation. In this view the extreme tail of the CTD would be used to promote the escape of RNA polymerase from transcription complexes on relaxed DNA. To test this idea we further characterized the poised transcription complex and determined the role of the sigma38 CTD tail and DNA relaxation in its properties. The primary aim is to understand how removal of the C-terminal 8 amino acids can lead to a reduction in transcription on relaxed DNA.

We first assessed whether the CTD tail is needed to form and maintain open transcription complexes on relaxed DNA templates. Open pre-initiation complexes contain melted DNA (transcription bubbles) that may be detected using KMnO4, which preferentially modifies single-stranded DNA at thymines (Sasse-Dwight and Gralla, 1989). This technique was used previously to demonstrate that inactive osmY transcription complexes on supercoiled DNA are open (Lee and Gralla, 2004). Figure 4A shows that the DNA is open on relaxed DNA (lane 3 vs. controls in lanes 1 and 2) and more important it remains open when the CTD is truncated to inactivate the transcription complex (compare lane 4 with 3). We infer that the CTD requirement for transcription is not related to a role in forming open transcription complexes.

Figure 4.

Figure 4

Open in a new tab

Removal of the extreme CTD leaves RNA polymerase poised and capable of abortive initiation on relaxed DNA in vitro. (A) Permanganate footprinting showing an open DNA signal with both active transcription complexes (wild type) and inactive complexes (1-322). Position +2 on the osmY template is indicated. (B) Abortive transcription to form the trinucleotide AGU on the same templates used in part A. The experiments were in triplicate and the results from one representative assay are shown.

To become active, open transcription complexes must initiate RNA bond formation and then allow RNA polymerase to escape into elongation mode. To learn whether the CTD plays a role in initial RNA synthesis nucleotides were added and small RNAs were assayed. To assess whether the inactive 1-322 transcription complex loses the ability to initiate short transcripts we added the dinucleotide ApG (-1 to +1) and radioactive UTP and assayed for ApGpU formation (the initial DNA sequence is AGTGATGAC). As a control to ensure that the signal was associated with the osmY promoter, we assayed in parallel an inactive variant with upstream sequences removed, osmY-5.

Figure 4B shows that the truncation of the CTD, which inactivates transcription, does not inactivate the ability to initiate RNA synthesis. That is, the wild type and 1-322 forms of sigma38 direct roughly equivalent levels of short RNA production (compare lanes 1 and 3, top row). These levels are promoter-specific as much less is produced when the transcribed sequence is maintained but the promoter is inactivated (lanes 2 and 4). Overall, these data indicate that the extreme CTD truncated sigma38-RNA polymerase is capable of opening the relaxed osmY promoter and forming the first nucleotide bond.

We also assayed for the ability to form subsequent bonds by adding appropriate NTPs to the reaction to lengthen the RNA. However, no longer products were detected under any condition except minor background bands that were also produced in equal amount from the inactive osmY-5 template (not shown). In addition, promoter-specific short RNA products were not detectable in normal (unprimed) reactions containing all 4 NTPs, and this was true whether conditions allowed transcription or not (not shown). Taken together these data indicated that inactive osmY transcription complexes are open and poised to transcribe and that the sigma38 CTD is required for the RNA polymerase to escape efficiently from the promoter when the appropriate signal is received.

The CTD is required to unwrap osmY DNA during activation by DNA relaxation

Prior to activation on supercoiled DNA the osmY promoter appears to wrap around RNA polymerase and is maintained in an inactive state (Lee and Gralla, 2004). Wrapping can be monitored by assaying a periodic array of DNase hypersensitive cleavages, as observed originally for DNA wrapped around the core of histones in a nucleosome (Elgin, 1981). A number of DNase I hypersensitive cleavages appear in the osmY transcription complex on supercoiled DNA and disappear when glutamate is added to allow transcription activation (Lee and Gralla, 2004). We now extend this assay to relaxed DNA.

Figure 5 compares the DNase I cleavage patterns of wild type and 1-322 sigma38 transcription complexes on relaxed osmY promoter DNA (non-template strand). The digestion pattern of naked DNA is shown as a control in lane 1. Both the wild type (lane 2) and the mutant (lane 3) show many protected bands, mostly downstream of position -40. These are within the established binding site for RNA polymerases and confirm that the promoter is bound in both cases. Because the 1-322 transcription complex is inactive under these conditions the footprint confirms that the inactivity is not due to a lack of DNA binding, as also shown above and by EMSA experiments (not shown).

Figure 5.

Figure 5

Open in a new tab

Removal of the CTD induces DNase I hypersensitive markers of DNA wrapping on relaxed osmY DNA. The cleavage patterns are for naked DNA and wild type or 1-322 sigma38 RNA polymerase. The locations of periodic hypersensitive sites with respect to the transcription start site are indicated.

The comparison of the wild type and 1-322 transcription complexes shows a series of much stronger bands in the inactive 1-322 complex (see arrows; lane 3 vs lane 2). These bands are much stronger in the lane 3 inactive transcription complex than they are in the naked DNA of lane 1 and are arrayed at the roughly 10 base pair intervals associated with DNA wrapping around a protein core. The footprints of the inactive 1-322 complex are reminiscent of those seen when DNA is wrapped around a histone core. Positioned nucleosomes show strong asymmetry in that one strand dominates the periodic hypersensitivity (Flaus et al., 1996; Maffey et al., 2007) and this also occurs for the inactive osmY transcription complexes (see figure S2).

We infer 2 things from these data. First, the extreme CTD tail is not needed to form a transcription complex, as also inferred from the permanganate data above. Second, and more important, we see that the CTD is needed to mediate the transition from an inactive, wrapped state (lane 3) to the active state (lane 2).

The DNA relaxation and glutamate pathways can be cooperative or antagonistic

When the current results are added to prior data on glutamate effects, it appears that there are at least 2 pathways for osmY activation, one involving DNA relaxation and the other involving glutamate. The pathways seem to have a common consequence, the unwrapping of DNA from a poised RNA polymerase, allowing it to escape and elongate m-RNA. The DNA relaxation pathway is different in the sense that only it requires the C-terminal 8 amino acids of sigma38. Of the 5 stresses studied in vivo (figure 1) heat, cold and stationary phase also required these amino acids. Osmotic and acid induction of osmY did not require this determinant and it is only these stresses that might use the glutamate pathway, as glutamate can be elevated from roughly 30-60 mM to up to 400 mM during osmotic shock (Dinnbier et al., 1988; Richey et al., 1987) and increases roughly three fold during acid stress (Natera et al., 2006). However, the DNA is also relaxed during acid stress, raising the possibility that the 2 pathways may be in play simultaneously. As there are many stresses that we have not tested, it is possible that others may involve both relaxed DNA and elevated glutamate levels. For this reason we systematically investigated the relationship between the 2 effectors of activation.

This experiment (Figure 6) was done by titrating relaxed DNA with an increasing concentration of potassium glutamate. As a comparison, the titration was also done on supercoiled DNA. The results with relaxed DNA (filled circles) show that the addition of glutamate can be either stimulatory or inhibitory, depending on the concentration of glutamate added. At the highest concentrations of glutamate, the level of transcription is very low, even lower than the unactivated basal level under these conditions. During osmotic stress when such glutamate concentrations are attained, the DNA is not relaxed but is actually hyper-supercoiled (Bordes et al., 2003; Cheung et al., 2003; Higgins et al., 1988); thus the antagonism is not physiological in this case.

Figure 6.

Figure 6

Open in a new tab

Potassium glutamate can either activate or inhibit in vitro transcription on relaxed DNA. (A) Relative transcription by wild type sigma38 RNA polymerase on relaxed DNA (filled circles). The data are normalized to the absence of potassium glutamate on supercoiled templates and are from the average of three experiments. Comparative data from supercoiled DNA is also shown (empty diamonds). The error bars, which are less than 10% of the observed values, have been omitted for clarity. (B) osmY expression using relaxed DNA in the presence of 200 mM potassium glutamate with indicated wild type, 1-322 or 1-321 forms of sigma38.

The stimulatory state of osmY transcription on relaxed DNA peaks near 200 mM glutamate and corresponds to a doubling of the signal over the already stimulated level induced by DNA relaxation. By contrast with the antagonistic conditions, this peak of stimulation may correspond to physiologically relevant conditions; they may correspond to the situation that pertains during acid stress as discussed above.

The data of figure 1 showed that osmY is activated during acid stress, but that the activation does not require the CTD tail of sigma38. In vitro, the CTD is required for the DNA relaxation pathway of stimulation but not for the glutamate activation pathway (top row, Figures 3A). Because acid stress in vivo may involve both pathways, we repeated the in vitro experiments under appropriate conditions where both are present; the DNA here is relaxed and the glutamate concentration is 200 mM. The results are shown in figure 6B. It is clear that under these in vitro conditions the CTD is not required as the truncated 1-321 and 1-322 proteins show the same level of transcription as the wild type protein. Footprinting experiments showed that all of these transcription complexes lacked the DNase I hypersensitive markers of DNA wrapping (not shown) as expected for active complexes. Thus, one can mimic the character of the in vivo response during acid shock, where the DNA is relaxed and glutamate levels are elevated, by establishing in vitro conditions where the DNA relaxation and glutamate pathways cooperate.

Discussion

E. coli deals with individual stresses by producing a collection of proteins appropriate to meet each stress. In addition, there is a general stress response that involves sigma38 (Typas et al., 2007; Weber et al., 2005), a protein that associates with a fraction of the common bacterial core RNA polymerase and changes its properties. The sigma38 RNA polymerase is bound to the general stress response osmY promoter in an inactivate form under unstressed conditions and can be activated by multiple stresses. Upon osmotic shock osmY transcription is activated by the intracellular accumulation of glutamate (Lee and Gralla, 2004;Rosenthal et al., 2008a). Data here suggest that other stresses can activate the poised RNA polymerase by the relaxation of DNA supercoils, which is a common event during many, but not all, stresses.

The experiments use the osmY marker gene, which is activated by multiple stresses via sigma38 transcription complexes. Although sigma38 is a member of the large sigma70 family of proteins it has a unique C-terminal domain (CTD) and for this reason we suspected a potential role in general stress regulation. The data showed that the C-terminal tail region of the CTD is needed to activate osmY transcription in response to heat and cold stress and stationary phase stress, but osmotic or acid stress response do not rely on the tail. Because the osmotic and acid stresses have in common that glutamate levels rise in vivo we sought another pathway that could account for activation during the other 3 stresses. DNA relaxation was investigated in part because this is thought to occur whenever ATP levels drop, which is common to many, but not all, stresses (Rui and Tse-Dinh, 2003; Winter et al., 2005). The results not only confirmed prior reports that sigma38 transcription can be stimulated by DNA relaxation (Bordes et al., 2003; Kusano et al., 1996) but also showed that the C-terminal tail of sigma38 is required. This is important because stresses that involve glutamate accumulation were shown to not need the tail, suggesting that there are 2 separate pathways, one relying on glutamate and the other on DNA relaxation.

The stress responses that involve glutamate accumulation and DNA relaxation were shown to intersect at sigma38 by changing its properties in a common way. Prior to activation the osmY promoter is bound by an inactive, poised RNA polymerase that is wrapped tightly by the DNA (Lee and Gralla, 2004). Glutamate was shown to unwrap this complex during activation and here we showed that DNA relaxation does the same upon activation. Activation of transcription by removal of supercoils is counter-intuitive as this tends to disfavor the required opening of promoter DNA by RNA polymerases. It has been previously argued that removal of supercoils can be stimulatory when the transcription complex exists in an unfavorably distorted state (Borowiec and Gralla, 1987). This view fits the current data in that topological considerations indicate that removal of supercoils could be expected to favor the unwrapping of DNA from around RNA polymerase. A significant sub-set of E. coli promoters are stimulated by DNA relaxation (Gmuender et al., 2001; Peter et al., 2004) and these arguments raise the interesting possibility that many of these activation events would involve DNA unwrapping. We note that many promoters in the E. coli genome appear to have poised RNA polymerases (Hatoum and Roberts, 2008; Reppas et al., 2006) and it is possible that properties shown here to be associated with osmY would also apply to some of them.

The -35 region of the osmY promoter contains a GC-rich sequence, which is conserved among a small set of sigma38 promoters known to be induced by osmotic shock and is clearly different from the typical sigma38 promoter (Lee and Gralla, 2004). Sequence changes within this element are permissible, so the role of individual bases remains unclear. Nonetheless, the sequence and positioning of -35 region sequences is important for the optimal transcription in the glutamate-dependent pathway both in vivo and in vitro (Lee and Gralla, 2004; Rosenthal et al., 2006). We found that the existence and positioning of this region is also important for optimal expression when the osmY promoter is activated by DNA relaxation (data not shown). It is possible that the -35 region element is involved in efficient DNA unwrapping and cooperates with the extreme CTD of sigma38 in this activation event.

The glutamate-dependent and DNA relaxation activation pathways have similar but not identical requirements for processing the activation signal within the sigma38 dependent transcription complex. Both pathways rely on the existence of a DNA-sequence element near position -35 ((Rosenthal et al., 2008a), data not shown). However, the short C-terminal tail of sigma38 is required only for activation by DNA relaxation in a way that is not yet understood. One view of these processes is that a tight network of upstream interactions including the -35 DNA, the CTD of sigma factors and the beta flap region of the core (Kuznedelov et al., 2002) specifies wrapping and that either DNA relaxation or glutamate can lead to unwrapping and activation of transcription by promoter escape.

In the current study, 5 stress responses were evaluated for reliance on the CTD tail in vivo and the data can explain the observations in 4 of these cases. Osmotic and acid shocks were observed not to require the CTD tail, which is also true of the glutamate-dependent pathway in vitro. In agreement, both stresses lead to glutamate accumulation in vivo, with osmotic stress leading to much higher levels. However, in the acid response the DNA becomes transiently relaxed, raising the possibility that this could also contribute. We showed that the joint application of glutamate and DNA relaxation in vitro could lead to either cooperative or antagonistic effects. At the concentration of glutamate that may correspond to acid shock conditions, the 2 pathways cooperate and do not depend on the CTD tail in vitro, in agreement with the in vivo observations. By contrast, heat, cold and stationary phase stress responses do rely on the CTD tail, but not on glutamate accumulation. In the cases of heat and stationary phase stresses, DNA relaxation occurs and this was shown to stimulate osmY expression. But in the case of cold shock DNA relaxation does not occur, raising the possibility that other activation pathways may exist that can mediate general stress responses through the CTD tail of sigma38. Because osmY and perhaps many other stress genes are bound by sigma38 RNA polymerase in poised inactive complexes, unwrapping to activate promoter escape may represent a common response to multiple stress signaling pathways.

Experimental procedures

Proteins and Plasmids

As described in (Rosenthal et al., 2008a), plasmids encoding sigma38 derivatives were constructed, and N-terminally His-tagged full-length and truncated sigma38 proteins were purified for in vitro experiments. Core RNA polymerase was purchased from Epicentre Technologies. Plasmid pOsmY was constructed by inserting a 324 base pair fragment containing the osmY promoter into pTH8 through the unique restriction sites of BamHI and HindIII. The region upstream from -5 was removed from pOsmY to create pOsmY-5, which is unable to produce osmY mRNA (Rosenthal et al., 2006). Relaxed DNA fragments were obtained by digesting the plasmids by XmnI, or by PCR reactions (amplified from -150 to +180 corresponding to the transcriptional start site of osmY).

In vivo mRNA analysis

Primer extension experiments were as described (Rosenthal et al., 2006). Sigma38 minus strain RJ4099 containing one version of plasmid encoded sigma38 was grown overnight in 5 ml of LB with 0.4% of glucose at 37 °C. The cells were diluted 1:100 to fresh medium (no glucose) containing 0 or 0.1 mM IPTG and grown to an optical density of 0.5. Osmotically challenged cultures were harvested either immediately before or 15 min after the addition of NaCl to a final concentration of 400 mM. Acid challenge cultures were harvested either immediately before or 15 min after the pH is decreased to 4.7 by careful addition of HCl. Heat or cold shock cultures were harvested either immediately before or 10 to 20 min after the growth media temperature was changed to 42°C or 15°C. Stationary phase cultures were harvested when the OD600 reached 1.5 to 2.0. Cells were harvested by centrifugation and stored at -70 °C. RNA was extracted from the pellets using a Qiagen RNeasy Kit. The total amount of RNA of each reaction was normalized, and the RNA was detected by primer extension by osmY primer (AGGGAGAGACTTCACCCTCTACAGAAG). Ten-microliter reactions for primer extension analysis contained 10 μg of total RNA, 10 nM labeled primer, reverse transcriptase buffer (Promega), 5 U of reverse transcriptase, and 0.2 mM deoxy-NTPs. Urea stop dye was added, and samples were run at 21 W on 6% polyacrylamide sequencing gel. Radioactive bands were visualized and quantified by phosphorimager analysis. RNA samples were prepared at least three times, and the average was taken. As a control, the protein levels of wild type and truncated sigma38 factors were found to be the same during stress conditions (osmotic shock, heat shock and entry into stationary phase) by western blot assays as described (Rosenthal et al., 2008a)

In Vitro Transcription

For single-round transcription experiments, 200 nM sigma38 and 50 nM core RNA polymerase were incubated at 37°C for 10 min in 1× Buffer B containing variable amounts of potassium glutamate (50 mM Tris-HCL at pH 7.9, 3 mM MgCl2, 0.1 mM EDTA, 1 mM DTT, and 100 μg/ml BSA). The final concentration of NaCl, originally from sigma38 and core storage buffer, is 37.5 mM. 5 nM of supercoiled or relaxed plasmid DNA was added and incubated for 15 min at 37°C. NTP mix (final concentration 150 μM each ATP, CTP and GTP, 5 μM UTP, 200 nCi UTP, 100 μg/ml heparin) was added and incubated for the appropriate time at 37°C. For multiple-round transcription experiments, all conditions were the same except that heparin is absent. For the abortive transcription assay, ApG is preincubated with 200 nM sigma38 and 50 nM core RNA polymerase at 37°C for 10 min in 1× Buffer B, 5 μM UTP and 200 nCi UTP are added for 10 min at 37°C. For all assays, urea stop dye was used to stop the reaction. The samples of single-round and multiple-round transcription assays were run on 6% PAGE with 1× Tris-buffered EDTA at 21 Watts for 70 min. The samples of abortive transcription assays were run on 15% PAGE with 1× Tris-buffered EDTA at 32 Watts for 90 min. Radioactive bands were visualized and quantitated by phosphorimager analysis. Each experiment was conducted at least three times and bands were standardized for each set of experiments to the condition that averaged the highest signal.

DNA Footprinting

Footprinting experiments were conducted under the same conditions as in vitro transcription. KMnO4 footprinting was as described (Lee and Gralla, 2004). In brief, after DNA incubation, 2 mM KMnO4 was added for 15 s and quenched with B-mercaptoethanol. Samples were run through G-50 spin columns and precipitated with ethanol. For DNase I footprinting, 2 μg/ml DNase I (from Sigma) was added for 30 s. The reactions were stopped by the addition of 150 μl of PB buffer from Qiagen (chaotropic salts) and purified with the Qiagen PCR Purification Kit. All footprinting experiments were analyzed through primer extension. Samples were run on 6% PAGE with 1× Tris-buffered EDTA at 32 Watts for 100 min and radioactive bands were visualized by phosphorimager analysis. Each experiment was conducted at least three times.

Supplementary Material

Supp Figs

Acknowledgements

Supported by NIH grant GM35754.

References

  1. Arnold CN, McElhanon J, Lee A, Leonhart R, Siegele DA. Global analysis of Escherichia coli gene expression during the acetate-induced acid tolerance response. J Bacteriol. 2001;183:2178–2186. doi: 10.1128/JB.183.7.2178-2186.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Balke VL, Gralla JD. Changes in the linking number of supercoiled DNA accompany growth transitions in Escherichia coli. J Bacteriol. 1987;169:4499–4506. doi: 10.1128/jb.169.10.4499-4506.1987. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bordes P, Conter A, Morales V, Bouvier J, Kolb A, Gutierrez C. DNA supercoiling contributes to disconnect sigmaS accumulation from sigmaS-dependent transcription in Escherichia coli. Mol Microbiol. 2003;48:561–571. doi: 10.1046/j.1365-2958.2003.03461.x. [DOI] [PubMed] [Google Scholar]
  4. Borowiec JA, Gralla JD. All three elements of the lac ps promoter mediate its transcriptional response to DNA supercoiling. J Mol Biol. 1987;195:89–97. doi: 10.1016/0022-2836(87)90329-9. [DOI] [PubMed] [Google Scholar]
  5. Cheung KJ, Badarinarayana V, Selinger DW, Janse D, Church GM. A microarray-based antibiotic screen identifies a regulatory role for supercoiling in the osmotic stress response of Escherichia coli. Genome Res. 2003;13:206–215. doi: 10.1101/gr.401003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Ding Q, Kusano S, Villarejo M, Ishihama A. Promoter selectivity control of Escherichia coli RNA polymerase by ionic strength: differential recognition of osmoregulated promoters by E sigma D and E sigma S holoenzymes. Mol Microbiol. 1995;16:649–656. doi: 10.1111/j.1365-2958.1995.tb02427.x. [DOI] [PubMed] [Google Scholar]
  7. Dinnbier U, Limpinsel E, Schmid R, Bakker EP. Transient accumulation of potassium glutamate and its replacement by trehalose during adaptation of growing cells of Escherichia coli K-12 to elevated sodium chloride concentrations. Arch Microbiol. 1988;150:348–357. doi: 10.1007/BF00408306. [DOI] [PubMed] [Google Scholar]
  8. Elgin SC. DNAase I-hypersensitive sites of chromatin. Cell. 1981;27:413–415. doi: 10.1016/0092-8674(81)90381-0. [DOI] [PubMed] [Google Scholar]
  9. Flaus A, Luger K, Tan S, Richmond TJ. Mapping nucleosome position at single base-pair resolution by using site-directed hydroxyl radicals. Proc Natl Acad Sci USA. 1996;93:1370–1375. doi: 10.1073/pnas.93.4.1370. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Foster JW. Escherichia coli acid resistance: tales of an amateur acidophile. Nat Rev Microbiol. 2004;2:898–907. doi: 10.1038/nrmicro1021. [DOI] [PubMed] [Google Scholar]
  11. Giuliodori AM, Gualerzi CO, Soto S, Vila J, Tavio MM. Review on bacterial stress topics. Ann N Y Acad Sci. 2007;1113:95–104. doi: 10.1196/annals.1391.008. [DOI] [PubMed] [Google Scholar]
  12. Gmuender H, Kuratli K, Di Padova K, Gray CP, Keck W, Evers S. Gene expression changes triggered by exposure of Haemophilus influenzae to novobiocin or ciprofloxacin: combined transcription and translation analysis. Genome Res. 2001;11:28–42. doi: 10.1101/gr.157701. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Gralla JD. Escherichia coli ribosomal RNA transcription: regulatory roles for ppGpp, NTPs, architectural proteins and a polymerase-binding protein. Mol Microbiol. 2005;55:973–977. doi: 10.1111/j.1365-2958.2004.04455.x. [DOI] [PubMed] [Google Scholar]
  14. Hatoum A, Roberts J. Prevalence of RNA polymerase stalling at Escherichia coli promoters after open complex formation. Mol Microbiol. 2008;68:17–28. doi: 10.1111/j.1365-2958.2008.06138.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Hengge-Aronis R. Back to log phase: sigma S as a global regulator in the osmotic control of gene expression in Escherichia coli. Mol Microbiol. 1996;21:887–893. doi: 10.1046/j.1365-2958.1996.511405.x. [DOI] [PubMed] [Google Scholar]
  16. Hengge-Aronis R. Signal transduction and regulatory mechanisms involved in control of the sigma(S) (RpoS) subunit of RNA polymerase. Microbiol Mol Biol Rev. 2002;66:373–395. doi: 10.1128/MMBR.66.3.373-395.2002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Higgins CF, Dorman CJ, Stirling DA, Waddell L, Booth IR, May G, Bremer E. A physiological role for DNA supercoiling in the osmotic regulation of gene expression in S. typhimurium and E. coli. Cell. 1988;52:569–584. doi: 10.1016/0092-8674(88)90470-9. [DOI] [PubMed] [Google Scholar]
  18. Jones TH, Murray A, Johns M, Gill CO, McMullen LM. Differential expression of proteins in cold-adapted log-phase cultures of Escherichia coli incubated at 8, 6 or 2 degrees C. Int J Food Microbiol. 2006;107:12–19. doi: 10.1016/j.ijfoodmicro.2005.08.006. [DOI] [PubMed] [Google Scholar]
  19. Kusano S, Ding Q, Fujita N, Ishihama A. Promoter selectivity of Escherichia coli RNA polymerase E sigma 70 and E sigma 38 holoenzymes. Effect of DNA supercoiling. J Biol Chem. 1996;271:1998–2004. doi: 10.1074/jbc.271.4.1998. [DOI] [PubMed] [Google Scholar]
  20. Kuznedelov K, Minakhin L, Niedziela-Majka A, Dove SL, Rogulja D, Nickels BE, Hochschild A, Heyduk T, Severinov K. A role for interaction of the RNA polymerase flap domain with the sigma subunit in promoter recognition. Science. 2002;295:855–857. doi: 10.1126/science.1066303. [DOI] [PubMed] [Google Scholar]
  21. Lee SJ, Gralla JD. Sigma38 (rpoS) RNA polymerase promoter engagement via -10 region nucleotides. J Biol Chem. 2001;276:30064–30071. doi: 10.1074/jbc.M102886200. [DOI] [PubMed] [Google Scholar]
  22. Lee SJ, Gralla JD. Osmo-regulation of bacterial transcription via poised RNA polymerase. Mol Cell. 2004;14:153–162. doi: 10.1016/s1097-2765(04)00202-3. [DOI] [PubMed] [Google Scholar]
  23. Lopez-Garcia P, Forterre P. DNA topology and the thermal stress response, a tale from mesophiles and hyperthermophiles. Bioessays. 2000;22:738–746. doi: 10.1002/1521-1878(200008)22:8<738::AID-BIES7>3.0.CO;2-5. [DOI] [PubMed] [Google Scholar]
  24. Maffey AH, Ishibashi T, He C, Wang X, White AR, Hendy SC, Nelson CC, Rennie PS, Ausio J. Probasin promoter assembles into a strongly positioned nucleosome that permits androgen receptor binding. Mol Cell Endocrinol. 2007;268:10–19. doi: 10.1016/j.mce.2007.01.009. [DOI] [PubMed] [Google Scholar]
  25. Muffler A, Barth M, Marschall C, Hengge-Aronis R. Heat shock regulation of sigmaS turnover: a role for DnaK and relationship between stress responses mediated by sigmaS and sigma32 in Escherichia coli. J Bacteriol. 1997;179:445–452. doi: 10.1128/jb.179.2.445-452.1997. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Natera V, Sobrevals L, Fabra A, Castro S. Glutamate is involved in acid stress response in Bradyrhizobium sp. SEMIA 6144 (Arachis hypogaea L.) microsymbiont. Curr Microbiol. 2006;53:479–482. doi: 10.1007/s00284-006-0146-y. [DOI] [PubMed] [Google Scholar]
  27. Ohnuma M, Fujita N, Ishihama A, Tanaka K, Takahashi H. A carboxy-terminal 16-amino-acid region of sigma(38) of Escherichia coli is important for transcription under high-salt conditions and sigma activities in vivo. J Bacteriol. 2000;182:4628–4631. doi: 10.1128/jb.182.16.4628-4631.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Peter BJ, Arsuaga J, Breier AM, Khodursky AB, Brown PO, Cozzarelli NR. Genomic transcriptional response to loss of chromosomal supercoiling in Escherichia coli. Genome Biol. 2004;5:R87. doi: 10.1186/gb-2004-5-11-r87. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Reppas NB, Wade JT, Church GM, Struhl K. The transition between transcriptional initiation and elongation in E. coli is highly variable and often rate limiting. Mol Cell. 2006;24:747–757. doi: 10.1016/j.molcel.2006.10.030. [DOI] [PubMed] [Google Scholar]
  30. Reyes-Dominguez Y, Contreras-Ferrat G, Ramirez-Santos J, Membrillo-Hernandez J, Gomez-Eichelmann MC. Plasmid DNA supercoiling and gyrase activity in Escherichia coli wild-type and rpoS stationary-phase cells. J Bacteriol. 2003;185:1097–1100. doi: 10.1128/JB.185.3.1097-1100.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Richey B, Cayley DS, Mossing MC, Kolka C, Anderson CF, Farrar TC, Record MT., Jr. Variability of the intracellular ionic environment of Escherichia coli. Differences between in vitro and in vivo effects of ion concentrations on protein-DNA interactions and gene expression. J Biol Chem. 1987;262:7157–7164. [PubMed] [Google Scholar]
  32. Rosenthal AZ, Hu M, Gralla JD. Osmolyte-induced transcription: -35 region elements and recognition by sigma38 (rpoS) Mol Microbiol. 2006;59:1052–1061. doi: 10.1111/j.1365-2958.2005.04999.x. [DOI] [PubMed] [Google Scholar]
  33. Rosenthal AZ, Kim Y, Gralla JD. Poising of Escherichia coli RNA Polymerase and Its Release from the sigma38 C-Terminal Tail for osmY Transcription. J Mol Biol. 2008a;376:938–949. doi: 10.1016/j.jmb.2007.12.037. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Rosenthal AZ, Kim Y, Gralla JD. Regulation of transcription by acetate in E. coli: in vivo and in vitro comparisons. Mol Microbiol. 2008b;68:907–917. doi: 10.1111/j.1365-2958.2008.06186.x. [DOI] [PubMed] [Google Scholar]
  35. Rui S, Tse-Dinh YC. Topoisomerase function during bacterial responses to environmental challenge. Front Biosci. 2003;8:d256–263. doi: 10.2741/984. [DOI] [PubMed] [Google Scholar]
  36. Sasse-Dwight S, Gralla JD. KMnO4 as a probe for lac promoter DNA melting and mechanism in vivo. J Biol Chem. 1989;264:8074–8081. [PubMed] [Google Scholar]
  37. Sayed AK, Odom C, Foster JW. The Escherichia coli AraC-family regulators GadX and GadW activate gadE, the central activator of glutamate-dependent acid resistance. Microbiology. 2007;153:2584–2592. doi: 10.1099/mic.0.2007/007005-0. [DOI] [PubMed] [Google Scholar]
  38. Travers A, Muskhelishvili G. DNA supercoiling - a global transcriptional regulator for enterobacterial growth? Nat Rev Microbiol. 2005;3:157–169. doi: 10.1038/nrmicro1088. [DOI] [PubMed] [Google Scholar]
  39. Travers A, Muskhelishvili G. A common topology for bacterial and eukaryotic transcription initiation? EMBO Rep. 2007;8:147–151. doi: 10.1038/sj.embor.7400898. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Tse-Dinh YC, Qi H, Menzel R. DNA supercoiling and bacterial adaptation: thermotolerance and thermoresistance. Trends Microbiol. 1997;5:323–326. doi: 10.1016/s0966-842x(97)01080-9. [DOI] [PubMed] [Google Scholar]
  41. Typas A, Becker G, Hengge R. The molecular basis of selective promoter activation by the sigmaS subunit of RNA polymerase. Mol Microbiol. 2007;63:1296–1306. doi: 10.1111/j.1365-2958.2007.05601.x. [DOI] [PubMed] [Google Scholar]
  42. Weber H, Polen T, Heuveling J, Wendisch VF, Hengge R. Genome-wide analysis of the general stress response network in Escherichia coli: sigmaS-dependent genes, promoters, and sigma factor selectivity. J Bacteriol. 2005;187:1591–1603. doi: 10.1128/JB.187.5.1591-1603.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Winter J, Linke K, Jatzek A, Jakob U. Severe oxidative stress causes inactivation of DnaK and activation of the redox-regulated chaperone Hsp33. Mol Cell. 2005;17:381–392. doi: 10.1016/j.molcel.2004.12.027. [DOI] [PubMed] [Google Scholar]
  44. Yim HH, Brems RL, Villarejo M. Molecular characterization of the promoter of osmY, an rpoS-dependent gene. J Bacteriol. 1994;176:100–107. doi: 10.1128/jb.176.1.100-107.1994. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Yura T, Guisbert E, Poritz M, Lu CZ, Campbell E, Gross CA. Analysis of sigma32 mutants defective in chaperone-mediated feedback control reveals unexpected complexity of the heat shock response. Proc Natl Acad Sci U S A. 2007;104:17638–17643. doi: 10.1073/pnas.0708819104. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp Figs
NIHMS73123-supplement-Supp_Figs.pdf (225.1KB, pdf)

RESOURCES