Introduction

Sigma factors are dissociable subunits of RNA polymerase (RNAP) holoenzymes which are responsible for the recognition of specific promoter DNA sequences and initiation of transcription. Almost all bacterial species contain several sigma factors which alternatively associate with the RNAP core. The individual types of RNAP holoenzymes containing different sigma factors recognize distinct classes of promoters differing in target (consensus) sequences. Sigma factors thus function as positively acting global regulators of transcription. In addition to an essential primary sigma factor responsible for transcription of majority of genes (mostly housekeeping genes), a variable number of non-essential alternative sigma factors, recognizing promoters of genes responsible for specific cell functions (e.g., in stress response or in entry into and in the stationary growth phase), are encoded by most bacteria.

In spite of the fact that the recognition of promoter sequences by a sigma factor is a highly specific process, an overlap of two or more sigma factors in recognizing a single promoter was described in bacteria. In Escherichia coli, functional overlap between primary sigma factor σ70 and heat-shock sigma factor σ32 [26] or general stress-responding sigma factor σ38 [14] was observed. Overlapping specificity of Bacillus subtilis alternative stress-responding sigma factors σX and σM [11] or σX and σW [18] was also found, and the respective dual promoters (i.e., recognized alternatively by two sigma factors) were described. In Mycobacterium tuberculosis, recognition of one promoter (of the sigB gene encoding an alternative sigma factor) even by three different stress-responding sigma factors (σE, σH, and σL) was proved [3]. Overlapping sigma factor specificity thus seems to be a common regulatory strategy in bacteria.

Corynebacterium glutamicum, a Gram-positive actinobacterium frequently used for the production of amino acids, encodes seven sigma factors: essential primary sigma factor σA, alternative primary-like σB, and five other alternative sigma factors (σC, σD, σE, σH, and σM), classified into Group 4 [or extracytoplasmic function (ECF)] sigma factors [16]. Analysis of C. glutamicum promoter sequences revealed high similarity between −10 motifs of promoters of housekeeping genes (supposed to be recognized by σA) and promoters recognized by σB [4, 10, 16, 17] which suggests possible overlapping promoter specificity of σA and σB. The data on core elements (−35 and −10 motifs) of C. glutamicum promoters recognized by ECF sigma factors are still very limited [16, 17]. Only the high similarity between core elements (−35 and −10) of C. glutamicum promoters recognized by σH [2, 5] or σM [12] was found so far. However, some C. glutamicum promoters were defined as σM-dependent by some authors [12] whereas as σH-dependent by others [5]. Very recently, a consensus sequence of several promoters of σC-controlled genes was deduced [24]. We report here the evidence of overlap of C. glutamicum sigma factors σA and σB, or σH and σE in recognizing specific promoters, obtained by use of C. glutamicum in vitro transcription system.

Materials and Methods

Bacterial Strains, Plasmids, Primers, and Growth Conditions

E. coli DH5α [7] and E. coli BL21 (DE3) [23] were used as hosts for cloning and for induced overexpression of the genes encoding sigma factors, respectively. Wild-type (WT) C. glutamicum ATCC 13032 and its deletion derivatives C. glutamicum ΔsigB [10], ΔsigH [27], and ΔsigE [15] were used as hosts for testing the activities of promoters cloned in promoter-test vector pEPR1. C. glutamicum rpoC-H8 [8] was used for isolation of RNAP core. Plasmids used are listed in Table 1. Oligonucleotides used are listed in the Supplementary material (Table S1). E. coli and C. glutamicum were cultivated in 2xTY medium [20] at 37 and 30 °C, respectively. Ampicillin (100 µg/ml) or kanamycin (30 µg/ml) was added to the selective media.

Table 1 Plasmids used
Full size table

DNA Techniques

DNA isolation, PCR, transformation of E. coli, DNA cloning, and DNA analysis were performed using standard methods [20]. C. glutamicum cells were transformed by electroporation [25].

Isolation and Purification of C. glutamicum RNAP Core and Sigma Factors

RNAP core subunits (with His8-tagged RpoC) were isolated from C. glutamicum strain rpoC-H8 [8]. His6-tagged sigma factors were isolated from E. coli BL21 (DE3) carrying plasmids pET-22b(+) with the respective C. glutamicum sig genes (Table 1) according to Holátko et al. [8]. TALON Resin (Clontech) was used (instead of Ni–NTA Agarose) for protein isolation by affinity chromatography. Isolated proteins were purified by gel filtration.

Construction of Promoter-Carrying DNA Templates for In Vitro Transcription

Fragments carrying C. glutamicum promoters were amplified by PCR using oligonucleotide primers listed in the Supplementary material (Table S1). The fragment with promoter Pper (96 bp) [13] was amplified using the plasmid pGA1 from C. glutamicum LP-6 as a template resulting in the construct pRLG770Pper. The fragment with promoter P2dnaK (109 bp) [1] was amplified using C. glutamicum ATCC13032 genome DNA as a template to provide the construct pRLG770P2dnaK. Promoter fragments Ppqo, P1clgR, and P2dnaJ2 were prepared by dimerization of oligonucleotides listed in Table S1 to provide the constructs pRLG770Ppqo, pRLG770P2dnaJ2, and pRLG770P1clgR, respectively. The PCR fragments were cloned into the EcoRI-HindIII sites of the vector pRLG770 in all cases. All promoter fragments were cloned to form transcripts of the same size (~150 nt). As templates for in vitro transcription assay, PCR fragments amplified from pRLG770Px constructs, using the primer pair 30F–CM3 were applied. The PCR fragments were purified by phenol extraction and concentrated with Amicon Ultra Centrifugation Filter Ultracell 30 K.

In Vitro Transcription Assay

The multiple-round in vitro transcription assay was done according to Holátko et al. [8] with the following modifications: The RNAP core (100 nM) was mixed with the respective σ factor in a molar ratio of 1:15 (σA) or 1:30 (σB, σD, σE, σH, and σM). The holo-RNAP was assembled for 10 min at 37 °C. The transcription mixture was incubated for 15 min at 37 °C. The transcripts labeled with α32P-UTP were separated in 5.5 % polyacrylamide gel. In vitro transcription assays were done repeatedly for each promoter with different preparations of RNAP core and sigma factors with essentially the same results. Representative figures are shown.

GFP Fluorescence Intensity Measurements

The promoter activities were determined using a transcriptional fusion of the respective promoter regions with the gfpuv reporter gene in the promoter-test vector pEPR1 [9]. The cells were washed with PBS buffer with phenylmethylsulfonyl fluoride (0.1 mM) and disrupted using FastPrep homogenizer (MP Biomedicals, USA). The fluorescence of the cell extract was measured with a Saphire2 spectrophotometer (Tecan, USA) (excitation wavelength, 397 nm; emission wavelength, 509 nm). Protein concentration was determined by Bradford assay. Fluorescence intensity was expressed in arbitrary units per mg of proteins (AU/mg protein). Extracts from cells harboring the empty promoter-test vector pEPR1 were used as control to determine background fluorescence.

Results

Effects of Deletions of the sig Genes on the Activity of C. glutamicum Promoters

In previous studies of C. glutamicum sigma factors, their role in cell physiology and their regulons, σB [4, 10], and σH [2, 5] were analyzed in greatest detail. Involvement of σE in response to cell surface stresses (but not its regulon) was also reported [15]. To reveal the role of individual sigma factors in recognizing promoters of the selected C. glutamicum genes, we have therefore decided to study the effects of deletions of the sigB, sigH, and sigE gene, respectively, on the activity of these promoters. Out of the promoters previously found to be dependent on the primary-like alternative sigma factor σB, we have chosen the promoter of the pqo gene encoding pyruvate:quinone oxidoreductase [21] (Ppqo), expressed mainly in the stationary growth phase [4]. To test the effects of deletions of stress-responding ECF sigma factors, we have chosen promoters of the dnaK-grpE-dnaJ-hspR operon encoding proteins involved in heat-shock response [1], the dnaJ2 gene encoding a chaperone protein [2] and the clgR gene encoding a transcriptional regulator [6]. It was previously found that all these genes are expressed from two promoters: a housekeeping one (i.e., most probably σA-dependent) and a σH-dependent one. For transcriptional fusions with the promoterless gfpuv reporter gene of the promoter-test vector pEPR1, we used DNA fragments only containing active σH-dependent promoters P2dnaK, P2dnaJ2, and P1clgR, respectively. The activities of the promoters were assayed as green fluorescence intensity of the reporter.

As shown in Fig. 1, the activity of the Ppqo promoter decreased significantly in the C. glutamicum ΔsigB strain compared to that in C. glutamicum WT. This result is in agreement with the previous finding that this promoter is σB dependent [4]. The observed substantial residual activity of Ppqo in C. glutamicum ΔsigB suggests that this promoter is also recognized by another sigma factor (most probably by σA).

Fig. 1
figure 1

Activity of Ppqo promoter in C. glutamicum WT and C. glutamicum ΔsigB strains. AU arbitrary units. The value of background fluorescence intensity (cells with promoterless vector pEPR1) was subtracted in all cases. Standard deviations of three measurements are depicted by error bars

Full size image

Out of the promoters recognized by an ECF sigma factor, activities of σH-dependent promoters P1clgR, P2dnaK, and P2dnaJ2 were measured in the wild-type strain and in mutants having deletion of the sigH or sigE gene. The P1clgR promoter is a σH-dependent promoter of the C. glutamicum clgR gene encoding a transcriptional regulator [6]. (The housekeeping P2clgR promoter closely adjacent to P1clgR [6] was inactivated in the pEPRP1clgR plasmid by mutation of its −10 hexamer as shown in the Supplementary material). As shown in Fig. 2a, the activity of P1clgR decreased in C. glutamicum ΔsigH compared to that in C. glutamicum WT. This result confirmed the σH-dependency of this promoter [6]. However, its observed high residual activity in C. glutamicum ΔsigH suggested that P1clgR is recognized also by another (most probably an ECF) sigma factor.

Fig. 2
figure 2

Activity of selected C. glutamicum promoters in C. glutamicum WT, C. glutamicum ΔsigB or ΔsigE mutants. a Promoter P1clgR; b Promoter P2dnaK; c Promoter P2dnaJ2. AU arbitrary units. The value of background fluorescence intensity was subtracted in all cases. Standard deviations of three measurements are depicted by error bars

Full size image

As shown in Fig. 2b, activity of another known σH-dependent P2dnaK promoter was found to decrease in C. glutamicum ΔsigH only in early exponential growth phase. To test if the observed high residual activity of P2dnaK in C. glutamicum ΔsigH is due to the recognition of this promoter by another ECF sigma factor, the activity of this promoter was also measured in C. glutamicum ΔsigE. However, no decrease of P2dnaK activity was observed during the growth of this strain (P2dnaK activity in this strain even increased in the later growth stages). Deletion of the sigH gene was shown to have no effect on the activity of another known σH-dependent promoter P2dnaJ2, whereas deletion of sigE showed even positive effect on the activity of this promoter (Fig. 2c). The effect of deletions of both sigE and sigH in the same strain could not be tested since all attempts to construct C. glutamicum ΔsigE ΔsigH failed, probably due to a lethal effect of the double deletion.

The results of measuring activities of selected promoters in the C. glutamicum mutants with deletions in the genes encoding sigma factors suggested recognition of some promoters by more than one sigma factor. To prove directly the recognition of a specific promoter by different sigma factors (overlapping specificity of sigma factors), in vitro transcription assays were performed.

Overlapping Specificity of C. glutamicum σA and σB in Promoter Recognition Proved by In Vitro Transcription

To test if the C. glutamicum primary sigma factor σA and primary-like σB overlap in recognizing promoter sequences of housekeeping genes, the multiple-round in vitro assay, using reconstituted C. glutamicum holoenzymes RNAP + σA or RNAP + σB and DNA template containing the Pper promoter, was performed. The very strong Pper promoter (originally present on C. glutamicum pGA1 plasmid [13]) contains −35 and −10 hexamers (TAGAAT and TATAAT, respectively [13]) which are highly similar to the C. glutamicum consensus sequences of housekeeping promoters [16]. The −10 hexamer within the promoter is even identical with the consensus sequences of E. coli and B. subtilis housekeeping promoters. As shown in Fig. 3a, use of both reconstituted RNAP + σA and RNAP + σB resulted in strong signals with the Pper template. On the other hand, no signals were obtained when RNAP core only or RNAP holoenzymes with ECF sigma factors σD, σE, σH, and σM were used for in vitro assay. The results of in vitro assay thus proved that a typical C. glutamicum housekeeping promoter is recognized by the primary sigma factor σA but also by primary-like σB.

Fig. 3
figure 3

In vitro transcription from C. glutamicum promoters recognized by σA and σB. Promoters used as templates and sigma factors in holo-RNAP are indicated at the top. The specific transcripts are indicated with arrows. (a) Promoter Pper; (b) Promoter Ppqo

Full size image

Promoter of the C. glutamicum pqo gene encoding pyruvate:quinone oxidoreductase [21] was previously proved to be controlled by σB [4]. As shown in Fig. 1, we have confirmed σB-dependency of this promoter. However, strong residual activity of Ppqo in C. glutamicum ΔsigB suggested that this promoter is recognized by still another sigma factor (most probably by σA) as well. Since deletion of the essential C. glutamicum sigA gene is lethal, the use of in vitro assay is the only way to prove directly the recognition of a promoter by σA. As shown in Fig. 3b, the use of both reconstituted RNAP holoenzymes, RNAP + σA or RNAP + σB, resulted in clear signals with Ppqo template, whereas no signals were obtained when RNAP holoenzymes with the ECF sigma factors σD, σE, σH, and σM were used. We therefore conclude that the Ppqo promoter (containing −35 and −10 hexamers GTGGCA and CACAAT, respectively) is active with the primary-like sigma factor σB as well as with primary sigma factor σA.

Overlapping Specificity of C. glutamicum σH and σE in Promoter Recognition Proved by In Vitro Transcription

To broaden the testing of overlapping specificity of different C. glutamicum sigma factors, the following known σH-dependent promoters were analyzed by in vitro transcription assay: P2dnaK, P2dnaJ2, and P1clgR. As shown in Fig. 4, using templates containing each of these σH-dependent promoters, clear signals were obtained not only with the RNAP + σH holoenzyme, but also with RNAP + σE. On the other hand, no signal was obtained when RNAP holoenzymes containing primary sigma factor σA, primary-like σB, or other ECF sigma factors, σD and σM, were used for in vitro transcription assay. These results indicate that the examined σH-dependent promoters were also recognized by σE. The two stress-responding sigma factors, σH and σE, thus overlap in their recognition specificities in these cases. The results suggest that still unknown consensus sequence of C. glutamicum promoters recognized by σE is most probably highly similar to that recognized by σH (GGAA–19-20 nt–GTT [2]).

Fig. 4
figure 4

In vitro transcription from C. glutamicum promoters P2dnaK, P2dnaJ2, and P1clgR recognized by σH and σE. Promoters used as templates and sigma factors in holo-RNAP are indicated at the top. The specific transcripts are indicated with arrows. The upper bands in some samples probably represent transcripts arising from the vector

Full size image

Discussion

In vitro transcription assay was previously found to be the only method suitable for direct defining the C. glutamicum promoters recognized by the primary sigma factor σA since deletion of the C. glutamicum sigA gene is lethal [8]. Here we have proved that in vitro transcription assay is a convenient technique for defining the promoters recognized by two different sigma factors (dual promoters) in C. glutamicum. Using this method, we found promoters recognized by both primary σA and primary-like σB (Pper and Ppqo) as wells as promoters recognized by two different stress-responding sigma factors, σH and σE (P1clgR, P2dnaK, and P2dnaJ2). Although the use of in vitro transcription assay unequivocally proved recognition of particular promoters by specific sigma factors, dependence of some promoters (P2dnaK and P2dnaJ2) on specific sigma factors was not apparent when activities of these promoters were measured in C. glutamicum mutants with deletions of genes encoding these sigma factors.

The results of the in vitro and in vivo analyses of the promoters Ppqo (recognized by σA and σB) and P1clgR (recognized by σH and σE) are generally in agreement. The observed decreased but substantial activities of these promoters in the analyzed deletion mutants (ΔsigB and ΔsigH, respectively) suggested that in addition to the respective sigma factors (σB and σH) which control activity of these promoters, other sigma factors can substitute their function. Recognition of these promoters by sigma factors σA and σE, respectively, was indeed proved by in vitro transcriptional assay. On the other hand, measuring activities of the promoters P2dnaK and P2dnaJ2 (proved to be recognized by σH and σE by in vitro assay) in C. glutamicum ΔsigH and ΔsigE did not show clear dependency of these promoters on the respective sigma factors. High activity of the promoters P2dnaK and P2dnaJ2 in the deletion strains showed that an alternative sigma factor in each case was able to substitute the missing sigma factor efficiently. Higher activities of the P2dnaK and P2dnaJ2 promoters in C. glutamicum ΔsigE suggest that the regulation of activity of sigma factors through their competition for RNAP core or direct competition for binding the promoter sequences is an important regulatory mechanism balancing their functions. SigH is a candidate for global regulator due to its prominent role in the hierarchy of cross-regulations among sigma factors [22]. Its affinity to RNAP core and/or efficiency of RNAP + σH in transcription from the tested promoters is therefore probably higher than that of σE. Missing competition in the ΔsigE strain may result in increased activity of these dual promoters with σH. We have observed this phenomenon also in several other promoters (data not shown).

Recognition of the typical housekeeping promoter Pper by the alternative sigma factor σB is in agreement with our observation that Pper is still highly active in early stationary phase, when the level of σA decreases sharply and the level of σB increases [10]. The consensus sequences of σA- and σB-dependent promoters are virtually indistinguishable and σB is considered a second sigma factor for housekeeping genes [4] or a backup sigma for transition and stationary growth phases [10]. We therefore infer that Pper is active with both σA and σB in vivo. We assume that such overlapping functional specificity of σA and σB could be found in many housekeeping promoters.

None of the promoters tested in this study provided signals in in vitro transcription assays when RNAP + σD or σM were used. Preliminary in vitro assays in a different project with RNAP + σD and a few tested DNA templates resulted in clear positive signals (data not shown). These tests confirmed that the purified σD protein is active. RNAP + σM was tested with a few promoters which were supposed to be σM-dependent [12], but no positive result was obtained. We found that one of these promoters, PtrxB, is σH-dependent, which is in agreement with the results of Ehira et al. [5]. The lack of positive signal with RNAP + σM on PtrxB template was supported by the results of in vivo analysis using deletion mutants (data not shown). The role of σM in C. glutamicum clearly needs further investigation.

The presented results suggest that the combination of in vivo and in vitro techniques is useful for obtaining reliable data on transcriptional regulation by sigma factors of RNA polymerase.