In vitro evidence that RNA Polymerase acetylation and acetyl phosphate-dependent CpxR phosphorylation affect cpxP transcription regulation

Abstract

The central metabolite acetyl phosphate (acP) has long been proposed to influence transcription regulation by directly transferring its phosphoryl group to a number of response regulators in many bacterial species. Here, we provide in vitro evidence for this proposition and demonstrate, using an in vitro transcription system, that acP-dependent phosphorylation of aspartate 51 of CpxR induces transcription of one of its regulon members in E. coli, cpxP. We also used this in vitro transcription system to extend our previously reported in vivo data that hypothesized that acetylation of RNA polymerase (RNAP) influences acP-dependent cpxP transcription, using glutamine as a genetic mimic for acetylated arginine 291 of the carboxy-terminal domain of RNAP α subunit. The data we present here lend strong support to the hypothesis that acP has a direct effect on transcription regulation in E. coli via phosphorylation of CpxR, and that RNAP acetylation can modulate this response.

In vitro assessment of effects of phosphorylation and acetylation on transcriptional regulation in E. coli.

INTRODUCTION

To sense changes to their extracellular environment, bacteria commonly use two-component signal transduction (2CST) systems. Found in all bacterial species sequenced to date, these systems range from 1 to >100 homologues per genome (Capra and Laub 2012). The canonical 2CST system consists of two components: a sensor kinase (SK) and a response regulator (RR). The SK is a histidine kinase that autophosphorylates a conserved histidine (His) residue using ATP as phosphoryl donor. Once phosphorylated, the SK serves as phosphoryl donor to its cognate RR, an aspartate (Asp) kinase that is phosphorylated on a conserved Asp residue within its receiver domain. Phosphorylation induces a conformational change, which modifies RR activity, allowing it to regulate transcription (for reviews, see Stock, Robinson and Goudreau 2000; Bourret and Silversmith 2010; Goulian 2010).

In some systems, the SK possesses both kinase and phosphatase activities. Typically, the absence of stimulus favors phosphatase activity, which removes phosphoryl groups from the cognate RR. In other systems, stimulation promotes phosphatase activity, whereas the lack of stimulation favors phosphorylation (Norsworthy and Visick 2013).

This SK phosphatase activity can rapidly reverse the activity of the phosphorylated RR; it also minimizes the effect of phosphorylation by sources other than the cognate SK (Siryaporn and Goulian 2008). RRs have been shown to be phosphorylated, both in vitro and in vivo, by non-cognate SKs (Skerker et al.2005; Yamamoto et al.2005) and by the central metabolite acetyl-phosphate (acP) (Wolfe 2005, 2010). While the physiological impact of phosphorylation by non-cognate SKs is likely limited (Silva et al.1998; Siryaporn and Goulian 2008; Groban et al.2009), increasing evidence argues that acP impacts transcriptional activity of several RR-dependent genes in several bacterial species (Wolfe 2010). In Escherichia coli, one such gene (cpxP) encodes the periplasmic chaperone CpxP, which is regulated by the RR CpxR, a member of the 2CST system CpxA/R (Danese and Silhavy 1998).

In E. coli, the CpxR regulon consists of ≥50 genes (De Wulf et al.2002; Price and Raivio 2009), of which cpxP yields the strongest response to CpxR phosphorylation (De Wulf et al.2002). This strong response and the fact that cpxP transcription depends heavily, if not exclusively, on CpxR, makes cpxP a good readout for CpxR activity. Several groups have demonstrated, in vivo and in vitro, that acP can donate its phosphoryl group to CpxR (Danese et al.1995; Danese and Silhavy 1998; Dartigalongue and Raina 1998; Nakayama and Watanabe 1998; Lima et al.2012; van Rensburg et al.2015). It also can contribute to in vitro transcription of the CpxR-dependent Shigella sonnei gene virF (Nakayama and Watanabe 1998).

In vivo, acP-dependent phosphorylation of E. coli CpxR is accomplished by adding 0.4% glucose to the medium (Lima et al.2012) and correlates, in part, with cpxP transcription (Danese and Silhavy 1998; Wolfe et al.2008; Lima et al.2012). However, acP-dependent phosphorylation of CpxR is not the only contributing factor to cpxP transcription; deletion of ackA, which leads to increased intracellular acP concentrations (Klein et al.2007), dampens cpxP transcription, instead of enhancing it (Wolfe et al.2008; Lima et al.2012). This inhibition correlates with acetylation of the surface-exposed Lys291 of the α subunit of RNA polymerase (RNAP) (Lima et al.2012). Lys291acetylation depends strictly on acP (Kuhn et al.2014). In a reaction that requires no enzyme, this high-energy central metabolite directly donates its acetyl group to lysines. The reaction is specific, acetylating only lysines that reside within a molecular environment that can both bind the phosphoryl group and activate lysine by deprotonation (Kuhn et al.2014). Therefore, we propose that CpxR is phosphorylated in glucose-grown wild-type cells, activating cpxP transcription. In contrast, the accumulation of acP in an ackA mutant leads to acetylation of Lys291 and inhibition of CpxR-activated CpxP transcription (see Fig. 1 for model).

Figure 1.

acP-dependent CpxR phosphorylation and RNAP αCTD acetylation: a working model. Schematic diagram of the Pta-AckA pathway that interconverts acCoA with acetate via an acP intermediate using the enzymes phosphotransacetylase (Pta) and acetate kinase (AckA) (Wolfe 2005). acP donates its phosphoryl (P) group to CpxR (1), which activates cpxP transcription (Danese et al.1995; Danese and Silhavy 1998; Dartigalongue and Raina 1998; Nakayama and Watanabe 1998; Lima et al.2012; van Rensburg et al.2015). When acP concentration increases, acP also donates its acetyl group (ac) to the C-terminal domain (CTD) of the RNAP α subunit (2) (Kuhn et al.2014), dampening cpxP transcription (Lima et al.2012). Bent arrow represents the cpxP promoter. αNTD refers to the N-terminal domain (NTD) of the RNAP α subunit.

To further determine the impact on cpxP transcription exerted by acP-dependent CpxR phosphorylation and to address the role played by Lys291 acetylation, we reconstituted this system in vitro using purified CpxR and RNAP, and an in vitro transcription vector carrying a 230 bp DNA fragment containing the cpxP promoter region. Similar to our in vivo observations, in vitro cpxP transcription absolutely required acP, CpxR, and its phosphorylation site, Asp51. In agreement with our in vivo data, acP-dependent cpxP transcription was inhibited by a genetic mimic of acetylated Lys291.

MATERIALS AND METHODS

Bacterial strains and plasmids

Table 1 lists all bacterial strains and plasmids used in this study. Derivatives were constructed via cloning or by site-directed mutagenesis (for details, see ‘Plasmid Construction’ and ‘Site-Directed Mutagenesis’).

Table 1.

Bacterial strains and plasmids used in this study.

	Description	Source
Strains
BL21 (DE3)	fhuA2 [lon] ompT gal (λ DE3) [dcm] ΔhsdS λ DE3 = λ sBamHIo ΔEcoRI-B int::(lacI::PlacUV5::T7 gene1) i21 Δnin5	Invitrogen
BW25113	lacI q rrnB T14 ΔlacZWJ16hsdR514 ΔaraBADAH33 ΔrhaBADLD78	(Datsenko and Wanner 2000)
DH5α	F− Φ80lacZΔM15 Δ(lacZYA-argF) U169 recA1 endA1 hsdR17(rk−, mk+) phoAsupE44 thi-1 gyrA96 relA1 λ−	ThermoFisher Scientific
RLG3569	BL21 (DE3) pLysS pLHN12	(Ross et al.2013)
Plasmids
pBPL1	pET-28 derivative expressing 6xHis-CpxR under the control of an IPTG-inducible promoter	(Lima et al.2012)
pBPL2	pBPL001 derivative expressing 6xHis-CpxR D51A under the control of an IPTG-inducible promoter	This work
pBPL3	pRLG770 derivative containing cpxP promoter fragment	This work
pBPL4	pIA900 expressing α K291Q	This work
pJET1.2	Blunt cloning vector	Life Technologies
pRLG770	ColE1-derived in vitro transcription vector	(Ross, Thompson et al.1990)
pIA900	RNAP subunits: α, β, β′-C-His10 and ω expressed from the T7 promoter	(I. Artsimovitch, unpublished)
pRLG1616	pRLG770-containing rrnB P1 (−88 to +50) promoter fragment	(Ross et al.1993)
pLHN12	pET11-containing RNAP ω subunit (rpoD)	(Rao et al.1996)

Culture conditions

Unless otherwise mentioned, cells were grown in Luria Broth (LB) containing 1% (wt/vol) tryptone, 0.5% (wt/vol) yeast extract and 0.5% (wt/vol) sodium chloride; LB plates also contained 1.5% agar. When necessary, ampicillin (100 μg mL⁻¹), kanamycin (40 μg mL⁻¹), spectinomycin (100 μg mL⁻¹), tetracycline (15 μg mL⁻¹) and chloramphenicol (25 μg mL⁻¹) were added. CpxR and RNAP expression from plasmids was induced with 0.25 and 1.0 mM isopropyl β-D-1-thiogalatopyranoside (IPTG), respectively.

Plasmid construction

Plasmid pBPL3 (pRLG770-cpxP) is a derivative of pRLG770, an ampicillin-resistant ColE1-based plasmid with tandem rrnB operon terminators T1 and T2 located downstream of the promoter insertion sites. It also contains the rna1 promoter, which serves as an internal transcription control (Ross et al.1990). The cpxP promoter region was amplified from strain BW25113 with the following primer pair derived from (Price and Raivio 2009): cpxP′FEcoRI (5′-AATAGGGAATTCAGTTCTCGGTCATC-3′) to insert the EcoRI site upstream of the promoter region and cpxP′RHinDIII (5′-GCAGCGAAGCTTAATGAACTGACTG-3′) to insert the HindIII site downstream. The resulting amplicon was ligated into pJET1.2, using the CloneJET PCR Cloning Kit (Fermentas, Walthmam, MA, USA). The ligated cpxP insert was sequenced, digested from pJET1.2 with EcoRI and HindIII, cloned into pRLG770 and transformed into E. coli DH5α cells. Plasmids were recovered from transformants, digested to screen for the cpxP promoter and sequence analyzed. The resulting plasmid was named pBPL3.

Site-directed mutagenesis

Site-directed mutagenesis of plasmids pBPL1 and pIA900 was conducted using Agilent Technologies's QuikChange Site-Directed Mutagenesis Kits (Santa Clara, CA, USA), according to manufacturer's instructions, using the following mutagenic primers: for CpxR, 5′-cagcattgatttacttttgcttgccgtaatgatgccgaagaaaaatg-3′ and 5′-catttttcttcggcatcattacggcaagcaaaagtaaatcaatgctg-3′; for α, 5′-ccaaggttaggcgtctgaaggagctcaacctcggtacg-3′ and 5′-cgtaccgaggttgagctccttcagacgcctaaccttgg-3′ to generate plasmids: pBPL2 and pBPL4, respectively. The mutations were confirmed by sequencing analysis of the purified mutagenized plasmids.

Protein expression and purification

CpxR expression

Expression of CpxR and CpxRD51A was performed using BL21 (DE3) cells transformed with pBPL1 or pBPL2, respectively, as described (Lima et al.2012).

RNAP expression

A single colony of BL21 (DE3) cells carrying plasmid pIA900 or pBPL4 was inoculated into 50 mL LB containing 100 μg mL⁻¹ ampicillin and incubated overnight at 37°C with aeration, rotating at 250 RPM. This overnight culture was used the following day to inoculate 1 L LB containing 100 μg mL⁻¹ ampicillin in a 2 L baffled flask, bringing the OD₆₀₀ to 0.05. The culture was incubated at 37°C with aeration, rotating at 250 RPM until the culture OD₆₀₀ reached 0.6, when 1 mM IPTG was added to induce RNAP expression, which continued for ∼6 h under the same conditions before cells were pelleted by centrifugation at 10 000 g for 10 min at 4°C. After centrifugation, the supernatant was decanted and the cell pellet placed at −20°C overnight or until cell lysis.

Cell lysis for RNAP extraction was performed by sonication, similarly to that of CpxR. Purification was conducted by passing the supernatant through 1 mL of cobalt agarose (HisPur, Life technologies, Carlsbad, CA, USA) previously washed twice with 10 mL of water and equilibrated with 10 mL resuspension buffer. After cell lysate passage through the column, the column was washed with 10 mL wash buffer at pH 8.0 (50 mM Na₂HPO₄, 0.3 M NaCl, 30 mM imidazole, 0.1% Tween 20, 5% ethanol). Protein was eluted, passing 3 mL of wash buffer containing 300 mM imidazole through the column. The eluted solution was collected and diluted into 3 mL of 0.2M TGED (TGED with 0.2M NaCl; TGED is 0.01 M Tris-HCl pH 8.0, 5% glycerol, 0.1 mM EDTA, 2 mM DTT). The diluted eluate was loaded onto 400 μL heparin sepharose (GE Healthcare) previously washed twice with 10 mL of water and equilibrated with 0.2M TGED. The flowthrough was collected and passed over the column again, after which the column was washed with 6 ml 0.2M TGED and RNAP eluted from the column with 3 mL of 0.6M TGED. The flowthrough was transferred into a dialysis cassette and placed into 1 L dialysis/storage buffer (10 mM Tris-HCl pH 8.0, 100 mM KCl, 50% glycerol, 10 mM MgCl₂, 0.1 mM EDTA, 1 mM DTT) overnight at 4°C with gentle stirring. After dialysis, the samples were divided into aliquots and stored at −80°C for future use.

Sigma 70 (σ⁷⁰) was purified from strain RLG3569 containing pLHN12 by procedures adapted from (Ross et al.2013).

In vitro transcription

In vitro transcription of cpxP was performed by multiple round of transcription. The final reaction volume was 25 μL and contained 1× in vitro transcription buffer (40 mM Tris-HCl pH 8.0, 10mM MgCl₂, 50 mM KCl, 1mM DTT, 100 μg ml⁻¹ BSA, 200 μM ATP, CTP and GTP, 10 μM UTP), 2 μCi α-[³²P]UTP, 50 ng pBPL3 plasmid, purified His6-CpxR at concentrations from 0 to 0.8 μM, acP from 0 to 40 mM and 10 nM of purified RNAP reconstituted with purified σ⁷⁰ (Eσ⁷⁰). After incubation for 15 min at 30°C, the reactions were stopped by addition of 25 μl of 2× stop solution (7 M Urea, 10 mM EDTA, 1% SDS, 2X TBE, 0.05% bromophenol blue), and immediately transferred onto ice.

The transcripts were separated using a denaturing acrylamide gel containing (6% acrylamide [19:1 acryl:bis], 7 M urea, 1× TBE), with 1× TBE as running buffer. Prior to phosphor imaging, the gels were dried and images obtained by exposing the gel to a storage phosphor screen. The gel image was analyzed using a Typhoon phosphorimager.

RESULTS

In vitro transcription system

To establish the cpxP in vitro transcription system, we constructed pBPL3, cloning an ∼230 nucleotide-long DNA fragment containing the cpxRA/cpxP intergenic region plus 90 nucleotides of the cpxR gene and 50 nucleotides of the divergently transcribed cpxP gene into the in vitro transcription vector pRLG770 (Fig. 2). Previously, purified CpxR was shown to bind to this promoter fragment in vitro (Yamamoto and Ishihama 2006) and, in vivo, this stretch of DNA was shown to contain all promoter elements required for CpxA-dependent cpxP transcription (Price and Raivio 2009).

Figure 2.

The 230 bp fragment containing the cpxAR/cpxP intergenic region in plasmid pBPL3. A 90 nt portion of cpxR and 50 nt portion of cpxP are indicated by arrows at the left and right ends of the fragment, respectively. The divergent promoters for cpxP and cpxAR are indicated by −10 and −35 boxes, and transcription start sites are indicated with a bent arrow. Proposed CpxR binding sites at positions −52 (Pogliano et al.1997; De Wulf, Kwon and Lin 1999; Yamamoto and Ishihama 2006; Raivio, Leblanc and Price 2013), −32 (De Wulf, Kwon and Lin 1999; Raivio, Leblanc and Price 2013) and −5 (De Wulf, Kwon and Lin 1999; Yamamoto and Ishihama 2006; Raivio, Leblanc and Price 2013) of cpxP transcription start site, as well as −67 (Pogliano et al.1997; De Wulf, Kwon and Lin 1999; Yamamoto and Ishihama 2006; Raivio, Leblanc and Price 2013) and −47 (De Wulf, Kwon and Lin 1999; Raivio, Leblanc and Price 2013) of cpxR transcription start site are indicated by gray boxes. A transcription terminator located downstream of the cpxP promoter on the plasmid vector (not shown) allows detection of specific length transcripts from this promoter.

In vitro transcription was accomplished by combining purified RNAP containing σ⁷⁰ (Eσ⁷⁰) with plasmid pBPL3 (which carries the cpxP promoter fragment), with or without CpxR and acP, and transcription was visualized by incorporation of [³²P]-labeled UTP into the transcripts. A specific transcript of ∼225 nt resulting from initiation at the cpxP promoter and termination at the pRLG770-encoded rrnT1 terminator was predicted. In contrast, no defined transcript from the divergent cpxAR promoter was expected using this plasmid vector, since no terminator is present.

In vitro transcription of cpxP requires CpxR and acP

AcP has been shown to donate phosphoryl groups to CpxR both in vivo and in vitro (Danese et al.1995; Danese and Silhavy 1998; Dartigalongue and Raina 1998; Nakayama and Watanabe 1998; Lima et al.2012; van Rensburg et al.2015). This transfer of a phosphoryl group has been proposed to allow binding of CpxR to the cpxP promoter region, such that cpxP transcription can be driven independently of the cognate SK, CpxA. Consistent with this hypothesis, no transcription from the cpxP promoter was observed when in vitro reactions were performed across a CpxR concentration range (0–0.8 μM) without acP (Fig. 3, lanes 1–5). However, when acP was added along with 0.4 μM CpxR, from the cpxP promoter, we observed a transcript of the expected size, which increased in abundance with increasing acP concentration (Fig. 3, lanes 6–10).

Figure 3.

acP and CpxR, together, are required for in vitro cpxP transcription. In vitro transcription of cpxP with increasing concentrations of His6-CpxR without acP (lanes 1–5), or with increasing concentrations of acP in the presence of 0.4 μM CpxR (lanes 4–10). 10 nM RNAP was present in all reactions. Transcription was determined by incorporation of ³²P-labeled UTP. Transcripts from the plasmid encoded rna1 promoter, which serves as transcription control, are indicated.

In vitro transcription requires acP and Asp51, the phosphorylation site of CpxR

It was previously shown, in vivo, that transcription from cpxP also required Asp51, the phosphorylation site on CpxR (Lima et al.2012). We therefore compared the effects of WT CpxR to those of CpxR containing a substitution of the phosphorylation site (CpxR D51A) in in vitro transcription reactions containing 10 μM acP. Activation of cpxP transcription increased with WT CpxR concentration; however, consistent with the prediction from in vivo experiments, no activation of cpxP transcription was seen with CpxR D51A (Fig. 4).

Figure 4.

Asp-51 of CpxR is required for in vitro cpxP transcription. (A) In vitro transcription of cpxP with increasing concentrations of WT His6-CpxR. The reactions contained 10 nM Eσ⁷⁰ and 10 mM acP. (B) As in (A), but with His6-CpxR D51A. Transcription from the rna1 promoter serves as a control.

Thus, CpxA-independent cpxP transcription has the same requirements both in vivo and in vitro: acP, CpxR and Asp51. These results lend strong support to the genetic evidence that acP-dependent phosphorylation of CpxR can drive cpxP transcription.

RNAP αCTD K291Q inhibits in vitro cpxP transcription

In vivo, acP-dependent cpxP transcription is inhibited under conditions that favor acetylation of Lys291 of the carboxy-terminal domain of the α subunit of RNAP (αCTD) e.g. an ackA mutant grown with 0.4% glucose (Lima et al.2012). This inhibition can be alleviated by conversion of Lys291 to an Ala (K291A), preventing its acetylation; in contrast, inhibition is induced by conversion of Lys291 to Gln (K291Q) (Lima et al.2012), a substitution commonly used to mimic acetylation (Kamieniarz and Schneider 2009).

To test whether the effect of αK291Q on cpxP transcription is direct or indirect, we combined purified RNAP that contained αK291Q with purified CpxR, pBPL3 and acP and monitored in vitro transcription by [³²P]-labeled UTP incorporation. When compared to WT RNAP, RNAP containing αK291Q supported no observable cpxP transcription; consistent with the hypothesis that αK291Q directly inhibits cpxP transcription. This defect in transcription, however, did not appear to affect the overall function of RNAP, as the K291Q substitution did not affect transcription from the plasmid-encoded rna1 promoter, which does not require the αCTD (Ross et al.1993) (Fig. 5).

Figure 5.

The RNAP alpha CTD K291Q substitution inhibits in vitro cpxP transcription, but does not inhibit transcription of the rrnB P1 promoter. (A) In vitro transcription of cpxP with 10 nM WT RNAP or 10 nM RNAP α K291Q. Reactions also contained 0.2 μM WT His6-CpxR and 10 mM acP. Duplicate lanes are shown for each RNAP. (B) In vitro transcription of the rRNA promoter rrnB P1 as in (A), with 10 nM WT RNAP or 10 nM RNAP α K291Q. Duplicate lanes are shown for each RNAP.

The αCTD of RNAP contributes to transcription regulation of several promoters by helping to recruit RNAP. This recruitment occurs via a direct interaction with the UP element, an AT-rich region found in some but not all promoters, and/or by a direct interaction with transcription factors (Gourse, Ross and Gaal 2000). Given the importance of the αCTD in global transcription regulation, we determined if the previously reported defect in cpxP transcription by αK291Q (Lima et al.2012) was due to a specific requirement of the cpxP promoter or a more generalized defect of α, for example, improper folding.

To help distinguish between these two possibilities, we used the ribosomal RNA promoter, rrnB P1. Using plasmid pRLG1616, which encodes −88 to +50 of rrnB P1, we found previously that rrnB P1 was transcribed at a much lower level in vitro with RNAP lacking the αCTD (Ross et al.1993), but was not significantly affected by an αCTD K291A substitution in vivo (Gaal et al.1996). Therefore, if the defect in in vitro cpxP transcription with αK291Q results from a more generalized defect on αCTD, then αK291Q should also affect rrnB P1 transcription. However, if αK291Q does not cause a generalized αCTD defect, then rrnB P1 transcription should not be affected. Consistent with this idea, no difference was observed when rrnB P1 was transcribed in vitro with αK291Q RNAP compared to WT RNAP (Fig. 5B). This result contrasts with cpxP transcription, which was severely affected by the Lys-to-Gln substitution (Fig. 5A). We conclude that the αK291Q has a direct inhibitory effect on cpxP transcription.

DISCUSSION

Much evidence exists supporting the hypothesis that acP donates its phosphoryl group to some RRs, including CpxR (Danese et al.1995; Danese and Silhavy 1998; Dartigalongue and Raina 1998; Nakayama and Watanabe 1998; Lima et al.2012; van Rensburg et al.2015). Genetic manipulations of the CpxAR 2CST system and the Pta-AckA pathway have led to the hypothesis that acP-dependent phosphorylation leads to transcriptional regulation by CpxR (Danese and Silhavy 1998; Wolfe et al.2008). However, determining if transcriptional effects resulting from Pta-AckA pathway manipulation are directly linked to CpxR phosphorylation due to acP concentration changes has proven challenging, because (i) this pathway plays a central metabolic role, (ii) manipulation of this pathway also affects global protein acetylation (Kuhn et al.2014) and (iii) protein acetylation correlates with effects on transcription of at least one CpxR-regulated gene, cpxP (Lima et al.2011, 2012).

The current work was designed to directly assess the effect of acP-dependent phosphorylation on CpxR-dependent cpxP transcription. To that end, we used an in vitro transcription system with the cpxP promoter as readout for CpxR phosphorylation. Our results demonstrate that acP (Fig. 3) and Asp51, the phosphorylation site of CpxR (Fig. 4), are absolutely required for in vitro cpxP transcription. These results correspond with data from in vivo CpxR studies (Lima et al.2012) and lend strong support to the hypothesis that acP-dependent phosphorylation of CpxR affects its activity, resulting in transcriptional regulation of cpxP.

We expect that additional CpxR-regulon members in E. coli would be affected by acP-dependent phosphorylation of CpxR. Other studies have demonstrated that, in addition to cpxP, several CpxR-regulated promoters respond to 0.4% glucose in the growth medium (Danese et al.1995; Cosgrove 2012), a condition shown to increase acP concentrations (Keating et al.2008) and induce CpxR phosphorylation in vivo (Lima et al.2012). However, the response to glucose by some CpxR-regulon members is the opposite of that obtained by CpxA-derived phosphorylation of CpxR (Cosgrove 2012). Such discrepancies could be due to additional changes in cellular physiology e.g. increased acetylation of transcription-associated proteins, among them RNAP itself (Lima et al.2011; Kuhn et al.2014).

Acetylation of RNAP is proposed to function as both activator and inhibitor of cpxP transcription (Lima et al.2011, 2012). Acetylation of the surface-exposed Lys298 of the αCTD is proposed to activate cpxP transcription (Lima et al.2012), whereas acetylation of the surface-exposed Lys291 of αCTD is proposed to inhibit cpxP transcription under a different set of circumstances (Lima et al.2012). Although the in vivo data could not distinguish between a direct or indirect effect, the described in vitro system supports a direct effect of the genetic mimic of acetylated Lys291 (K291Q) on cpxP transcription (Fig. 4). This result cannot be explained by a general αCTD defect e.g. improper folding, since rrnB P1 transcription, which is greatly reduced when RNAP lacks αCTD (Ross et al.1993), was unaffected by αK291Q (Fig. 5).

In summary, the current work provides definitive evidence that acP-dependent phosphorylation of CpxR can drive cpxP transcription. It also provides the first in vitro evidence that acetylation of RNAP could modulate this transcription.

Conflict of interest. None declared.