Escherichia coli DH5α (Hanahan, 1985) was cultivated aerobically in 500-ml flasks containing 70–100 ml of 2xYT medium (Green and Sambrook, 2012) at 37°C. C. glutamicum RES167 (Tauch et al., 2002) and C. glutamicum RES167ΔcseE (named C. glutamicum ΔcseE here) were cultivated under the same conditions, but at 30°C. The defined chromosomal deletion ΔcseE in the C. glutamicum chromosome was constructed using the sacB gene as a conditionally lethal marker and a double-crossover event as described previously (Busche et al., 2012). The plasmid vectors used are listed in Table 1. The sequences of the oligonucleotides for PCR and DNA cloning are shown in Supplementary Table S1.

DNA isolation, PCR, cutting with restriction enzymes, ligation, and transformation of E. coli were done using the standard techniques (Green and Sambrook, 2012). Mutations in sigH and sigE were constructed with a Q5 Site-Directed Mutagenesis kit (New England BioLabs Inc.).

Non-radioactive primer extension (PEX) analysis was carried out as described previously (Busche et al., 2012). C. glutamicum cells were cultivated in 2xYT medium at 30°C and disrupted with glass beads and a FastPrep FP120 (BIO101) disintegrator. The cell debris was removed by centrifugation, and total RNA was isolated from the extract using a High Pure RNA Isolation Kit (Roche Diagnostics). The primer extension analysis was essentially done using SuperScript III transcriptase (Invitrogen, Carlsbad, CA) with 30 μg RNA and 5 pmol Cy-5-labeled primer CM4 complementary to the vector pET2. The synthesized cDNA was run on PAA gel simultaneously with the DNA sequencing products generated with the same labeled primer in an A.L.F. Sequencer (GE Healthcare, Munich, Germany).

The whole procedure was done in principle according to the protocols used previously (Albersmeier et al., 2017; Wittchen et al., 2018; Dostálová et al., 2019) with a few modifications. Corynebacterium glutamicum RES167 or C. glutamicum ∆cseE was cultivated in minimal medium CGXII (Keilhauer et al., 1993) supplemented with glucose 2% (w/v) instead of 4% and protocatechuic acid 30 mg/l instead of 0.03 mg/l. The cells were harvested in the exponential growth phase and frozen in liquid nitrogen.

Total RNA was isolated from 3 biological replicates of C. glutamicum cells by a Quick-RNA Miniprep Plus kit (Zymo Research). The samples were treated with DNase (Roche Diagnostics) and RNA was purified with an RNA Clean&Concentrator-5 kit (Zymo Research). Ribosomal rRNA was removed with a Ribo-Zero rRNA Removal Kit for bacteria (Illumina). The purity of RNA was then tested with an Agilent RNA Pico 6,000 kit and an Agilent 2,100 Bioanalyzer (Agilent Technologies). TruSeq Stranded mRNA Sample Preparation guide (Illumina) was then used to construct the cDNA library. The constructed cDNA library was then sequenced with Illumina HiSeq 1,500 using a read length of 70 bases.

A primary 5′-end-specific cDNA library was constructed for sequencing with the aim of defining TSSs using the protocol described previously (Albersmeier et al., 2017; Wittchen et al., 2018) Briefly, rRNA was depleted and the RNA samples were treated with a terminator 5′-phosphate-dependent exonuclease (Illumina) to remove non-primary transcripts.

The primary 5′-triphosphate ends of the primary transcripts were treated with RNA 5′-polyphosphatase to convert them into 5′-mono phosphate ends, and the 5′-adapter was ligated to the produced 5′-ends. Then, reverse transcription with a stem-loop DNA adapter was carried out and the library was amplified. The primary 5′-end cDNA library was then purified and size-selected for fragments approximately 100–1,000 nt in size by gel electrophoresis and quantified. The fragments were finally sequenced with an Illumina MiSeq.

Paired-end reads were mapped to the C. glutamicum reference genome sequence accession number BX927147 as described previously (Dostálová et al., 2019). To visualize short read alignments, detect TSSs and analyze differential gene expression, Read Explorer v.2.2 (Hilker et al., 2014, 2016) was used. TSSs were detected as described previously (Wittchen et al., 2018). To detect the real TSSs (+1), the nucleotide was considered to be position +1 if the number of read starts was 10 times higher than at position −1. The false-positives in the candidate dataset of the predicted TSSs that were incorrectly classified automatically were examined and excluded manually.

Differential gene expression analysis of C. glutamicum RES167 and ΔcseE was carried out with the whole transcriptome data and ReadXplorer v2.2 (Hilker et al., 2016), including the Bioconductor package DESeq2 (Love et al., 2014). The signal intensity value (a-value) was calculated by the log2 mean of normalized read counts, and the signal intensity ratio (m-value) by log2 fold change. The evaluation of the differential RNA-seq data was done with an adjusted value of p cutoff of p ≤ 0.01 and a signal intensity ratio (m-value) cutoff of ≥1 or ≤ −1.

Genes with upregulated or downregulated expression in C. glutamicum ∆cseE were included in the further analysis (FDR < 0.05, M-value>1 for upregulation, M-value<−1 for downregulation).

The in vitro transcription assay was carried out essentially as described previously (Holátko et al., 2012). The promoter fragments (56–70 nt; with the downstream end at position +3 to +6 relative to TSS) cloned in the vector pRLG770 (Ross et al., 1990) were used as the templates (Dostálová et al., 2017). The holo-RNAP was reconstituted from the RNAP core enzyme isolated from C. glutamicum and individual C. glutamicum σ factors isolated as His-tagged recombinant proteins from E. coli as described previously (Holátko et al., 2012). The RNAP core (100 nM) was mixed with the respective σ factor (σ^A, σ^B, σ^E, or σ^H) in a molar ratio of 1:15 (σ^A) or 1:30 (σ^B, σ^E, and σ^H) The holo-RNAP was assembled for 10 min at 37°C. The transcription mixture was incubated for 10 min at 37°C. The transcripts labeled with [α-³²P]UTP were separated in 5.5% (w/v) polyacrylamide gel. In vitro transcription assays were done 2 or 3 times for each promoter, with essentially the same results.

Sigma factors were assigned to individual promoters in vivo using the recently described two-plasmid system for C. glutamicum (Dostálová et al., 2017). In principle σ factors, which were overproduced from the expression vector pEC-XT99A, initiated transcription from the individual promoters. The activity of these promoters was then measured using the reporter gene gfpuv in the promoter-test vector pEPR1 (Dostálová et al., 2017, 2019). The fluorescence of the cell-free extracts was determined with a Safire2 microplate spectrophotometer (Tecan; excitation wavelength, 397 nm; emission wavelength 509 nm). Arbitrary units of the fluorescence intensity of cell extract (AU/mg protein) were normalized to protein concentration, which was determined by the Bradford assay as described previously in detail (Dostálová et al., 2017, 2019). To test the effects of stresses on the activity of promoters, C. glutamicum cells carrying pEPR1 with an analyzed promoter were used in a single-plasmid assay. The cells were cultivated in the same way as for the two-plasmid assay in 2xYT medium and when OD₆₀₀ reached 1.0 (time point 0), stress (40°C for 1 h, 0.01% (w/v) SDS and 4% (v/v) ethanol, respectively) was applied. Promoter activity was measured as the fluorescence intensity of the reporter protein GFPuv.

The homology models of the σ^H and σ^E domains which recognize the −10 and − 35 sequences of the respective promoters were produced by using the Swiss-Model server (Waterhouse et al., 2018). The crystal structures of E. coli σ^E, PDBid: 4LUP (for −10 element GTC; Campagne et al., 2014) and PDBid: 2H27 (for −35 element GGAAC; Lane and Darst, 2006) were used as templates. The nucleotides within the E. coli σ^E consensus were replaced to match the consensus for C. glutamicum σ^H or σ^E, where necessary. Molecular dynamics simulations were done using the software package AMBER (Salomon-Ferrer et al., 2013) and Linux computer nodes with powerful NVIDIA GPUs that enable the accumulation of 50-ns MD trajectories at 280 K.

According to the previously published results of RNA sequencing (Albersmeier et al., 2017), the C. glutamicum gene encoding the sigma factor σ^E (sigE = cg1271) forms an operon with the downstream genes cseE (cg1272, encoding the corresponding anti-σ factor, anti-σ^Ε) and tatB (cg1273, twin-arginine translocation pathway protein; gene numbers are according to the genome sequence of C. glutamicum GenBank RefSeq BX927147). The sub-operon cseE-tatB was also detected (Albersmeier et al., 2017). We used primer extension (PEX) analysis to determine the respective sigE TSSs. DNA fragments (256 bp upstream of sigE and 227 bp upstream of cseE) were cloned in the vector pET2 and the constructs were used for PEX assay. Total RNA for PEX was isolated from C. glutamicum (pET2 carrying the sigE promoter region) and C. glutamicum (pET2 with-cseE promoter) cells, which were cultivated to the exponential growth phase. Three TSSs were found within the upstream region of sigE (data not shown). Two of them, TSS1 at A and TSS3 at G (Figure 1), were identical to the signals obtained by transcriptome sequencing in two studies (Pfeifer-Sancar et al., 2013; Albersmeier et al., 2017). This confirmed the −10 regions of P1sigE CAAAAT and P3sigE TATAGT. We additionally detected TSS2 at A, which positioned a weaker promoter with the potential −10 element TAATCT an appropriate distance upstream of TSS2. The putative −10 elements CAAAAT (P1sigE), TAATCT (P2sigE), and TATAGT (P3sigE; Figure 1) are similar to the C. glutamicum consensus sequence TANAAT of the σ^Α-dependent (housekeeping) genes (Pátek and Nešvera, 2011; Albersmeier et al., 2017). Nevertheless, no TSS was detected, which might suggest the presence of a σ^Ε-dependent promoter. We therefore supposed that all 3 promoters are σ^Α-dependent. The putative −35 regions were not very similar to the already published C. glutamicum − 35 consensus sequence TTG^A/_CCA (Pátek et al., 2013) or TTGNNN (Albersmeier et al., 2017), but this motif is known to be only weakly conserved in C. glutamicum promoters (Pátek et al., 2013; Albersmeier et al., 2017). Upstream of the cseE gene, PEX analysis revealed TSS2 at position A (Figures 2A,C), which was identical to the nucleotide found by transcriptome sequencing (Pfeifer-Sancar et al., 2013; Albersmeier et al., 2017). In addition, we detected a weak signal of TSS1 at G, which is 7 nt downstream of the appropriate −10 element TATCTT (Figures 2A,C). The signal at this position has also been previously detected by RNA-seq (Pfeifer-Sancar et al., 2013). The −10 hexamer TATCTT of the promoter P1cseE belonging to the weak signal looked again like a typical −10 motif of the σ^Α-controlled promoters (Figure 2C). The P2cseE promoter motifs −35 and − 10, GGAAC–N₁₈–GTT (Figure 2C), however, were identical to the sequences of the already described σ^Η/σ^E-dependent promoters (Šilar et al., 2016; Dostálová et al., 2017). The strength of PEX signals corresponding to TSS2 promoted by the σ^E-dependent P2cseE was analyzed with RNA isolated from C. glutamicum cells afflicted with cell surface stress (addition of SDS). The signal of TSS2 corresponding to the C. glutamicum cells cultivated under standard conditions was weak, whereas the respective peak obtained with the cells cultivated with 0.01% (w/v) SDS was very large (Figure 2B). This is in agreement with the assumption that the P2cseE promoter is σ^E-dependent and that the role of σ^E is in cell surface stress response. It is noteworthy that the TSS1 peak corresponding to the transcript driven from the housekeeping promoter P1cseE was also much larger in response to SDS stress (Figure 2B). This suggests that there is a common regulation of cseE transcription from the two promoters, despite them being controlled by different σ factors.

The effects of the stress conditions on the activities of the main promoters of the operon (P1sigE and P2cseE) were further examined using a transcriptional fusion of these promoters with the gfpuv reporter gene in the promoter-test vector pEPR1 (Figure 3). The strongest effect on P1sigE was found with SDS, and a weaker increase in activity was observed with 4% (v/v) ethanol. The effects of these stresses on P2cseE activity were only weak and thus not completely convincing. (Figure 3). However, strong effects of 40°C, ethanol (Figure 2A), and especially SDS (Figure 2B) on the P2cseE activity were clearly visible in the results of PEX.

To test the roles of particular σ factors in the recognition of the two strong promoters of the operon by a different method, in vitro transcription assay with P1sigE and P2cseE was carried out using the σ^Α, σ^Β, σ^E, and σ^Η factors and the promoters cloned in pRLG770. Signals with approximately the same strength were observed with σ^Α and σ^Β for P1sigE, whereas a strong signal with σ^E and a weaker signal with σ^Η were detected for P2cseE (Figure 4). This result was in agreement with the sequences of the key promoter elements of the tested promoters.

To further test the roles of σ^E and σ^Η in the transcription from P2cseE that was suggested based on the results of in vitro transcription (Figure 4), a two-plasmid assay which was developed to prove interaction between σ factors and promoters in vivo (Dostálová et al., 2017), was carried out. Expression analysis in vivo indicated that P2cseE can drive transcription with σ^E and a little less also with σ^Η (Figure 5). This is in agreement with the result of in vitro transcription analysis.

The transcriptional pattern of the sigE-cseE-tatB operon, and the σ factors which control the respective promoters, are summarized in Figure 2D.

To view the transcription initiated by σ^E at the genome level, we decided to use C. glutamicum RES167 and its derivative with a deleted cseE gene, respectively, to perform the differential gene expression analysis. Transcriptomes of these two strains were analyzed by RNA-seq. We supposed that σ^E, which the anti-σ factor CseE could not inhibit in the ΔcseE strain, would be active and initiate transcription of the σ^E-controlled genes even in the absence of a stress stimulus. These conditions are similar to the transcription by σ^H in the ΔrshA strain. A previous study (Busche et al., 2012) using the C. glutamicum ΔrshA strain (i.e., deletion of the gene encoding anti-σ^H) showed that in total 83 genes in 61 transcriptional units were directly or indirectly σ^Η-dependent.

The ratio/intensity (M/A) plot deduced from RNA-seq results comparing the gene expression of C. glutamicum ΔcseE with that of C. glutamicum RES167 is shown in Figure 6. In total, 694 genes were differentially transcribed (296 upregulated and 398 downregulated) in C. glutamicum ΔcseE compared to C. glutamicum RES167. These genes may be considered as genes whose expression is directly or indirectly modulated by σ^E, i.e., σ^E modulon (Supplementary Tables S2, S3). Our aim was to define the σ^E regulon, i.e., the genes under direct control of σ^E (these genes are described below, in the section Genes of the σ^Η/σ^E Regulon). We decided therefore to search for the σ^E-specific promoters upstream of the differentially transcribed genes.

To localize promoters of the upregulated genes, TSSs were mapped using the results of the sequencing of primary 5′-end-specific cDNA library in the same way as recently with C. glutamicum σ^A-and σ^D-controlled genes, respectively (Albersmeier et al., 2017; Taniguchi et al., 2017). To detect the actual TSSs (+1), the nucleotide was considered to be position +1 if the number of read starts was 10 times higher than at position −1. Of the upregulated genes (i.e., those which are thought to be members of the σ^Ε stimulon), 157 genes were in operons and without TSSs (and therefore without closely located promoters). Upstream of the remaining 139 upregulated genes, 168 TSSs were detected. Some of the genes were transcribed from 2 or 3 or even as much as 4 TSSs. Closely upstream of the detected TSSs, we recognized 150 sequences similar to the consensus sequence of vegetative promoters (i.e., σ^A-dependent). Using the software Improbizer (Ao et al., 2004), a homogenous group of promoter sequences was detected which were different from the sequences typical for σ^A-dependent genes in C. glutamicum. The sequence GGAAC–N_18-19–GTT (with a single exception GGAAA–N₁₉–GTT in amtR), which was identical to the sequences of a few already found σ^Ε-dependent promoters (Šilar et al., 2016; Dostálová et al., 2017), was recognized in 16 sequences at an appropriate distance upstream of TSSs (Table 2). This group of potentially σ^E-controlled genes also included the four genes which were previously proved to be transcribed from promoters recognized by both σ^E and σ^Η (dnaK, dnaJ2, clgR and sigB; Šilar et al., 2016; Dostálová et al., 2017). Their putative promoter sequences (approx. 70-bp DNA fragments) were then analyzed and promoter activities were confirmed by in vivo and in vitro assays. The same procedure for the downregulated genes did not reveal any possible σ^E-controlled genes and most genes seemed to be σ^A-controlled (Supplementary Table S3).

The two techniques, which we developed for the analysis of individual C. glutamicum promoters (Dostálová et al., 2017), were applied to confirm the results of the RNA-seq and thus unequivocally assign specific σ factors to the promoters. This combining of the results of two techniques has already provided reliable results with the σ^D/σ^Η-dependent promoters (Dostálová et al., 2019). The C. glutamicum RES167 cells harboring the expression vector pEC-XT99A with inserted C. glutamicum sigE or sigH gene and the vector pEPR1 carrying the tested promoter DNA fragment (approx. 70 nt) were used for the in vivo two-plasmid assay. The promoter activity was measured as a fluorescence intensity of the GFPuv reporter protein. The chosen promoters were further tested with the in vitro transcription system for C. glutamicum (Holátko et al., 2012) to confirm the classification of the target promoters.

In addition to the promoters, which were mapped by RNA-seq, we selected the promoter Pcg3344, which was previously found to be controlled by σ^Η (Ehira et al., 2009; Busche et al., 2012) and possessed the GGAAC sequence in the −35 region. Moreover, the promoter of the amtR (cg0986) gene, which was also found to be σ^Η-dependent (Ehira et al., 2009), and which possessed the atypical sequence GGAAA in the −35 region, was included. The most distinctive examples of σ^Ε-dependent promoters are shown in Figure 7. The promoters PamtR and Pcg3309 were highly active with σ^E, and weaker activity was also detected with σ^Η, as measured by both techniques (Figures 7C,D,K,L), whereas the promoter Pcg1277 only exhibited predominant σ^E-dependent activity in vivo (Figures 7G,H). As for Pcg1121, it was clearly σ^Η-dependent and very weakly σ^E-dependent by in vivo assay (Figure 7E), whereas an opposite relation was observed in the in vitro transcription (Figure 7F). All in all, none of the promoters (including the P2cseE promoter of the anti-σ^E factor gene, see Figures 4, 5) were found to be exclusively σ^E-specific in any analysis. Finally, all these σ^E-dependent promoters were generally active with both σ^E and σ^H in both in vivo and in vitro assays (Figure 7). According to the in vivo two-plasmid measurements, the most promoters were significantly weaker with overexpressed σ^E than σ^H, whereas the ratio of the strength with RNAP+σ^Η and RNAP+σ^E in in vitro transcription was variable. Interestingly, all the discovered σ^E-controlled genes (the members of the σ^E regulon) thus constitute a gene group which is entirely contained in the σ^Η regulon. All σ^E-dependent promoters (with a single exception of PamtR) contained the consensus sequence with the −35 and −10 sequences GGAAC–N_18-19–GTT.

As controls, the σ^Η-dependent promoters Pmca, PuvrD3, PmshC, and PsufR with the −35 region GGAAT (Table 3) were also used for these assays. Both in vivo and in vitro techniques clearly confirmed that these 4 tested control promoters are exclusively σ^Η-specific (Figure 8).

The comparative analysis of C. glutamicum RES167 and C. glutamicum ΔcseE transcriptomes thus showed that the strategy for detecting the σ^E-dependent genes using the σ^E-specific anti-σ factor gene deletion (analogous to previous detection of σ^Η-dependent genes) was successful. The results confirmed that the cseE gene located closely downstream of the sigE gene encodes the specific anti-σ^E as suggested previously (Park et al., 2008). The results are also consistent with our assumption that σ^E would be active in the cseE deletion strain that grows under optimal conditions without stress.

The sequences of the 16 σ^Η/σ^E-dependent promoters (Table 2) and 44 exclusively σ^Η-dependent promoters (Ehira et al., 2009; Busche et al., 2012; Toyoda et al., 2015) were separately aligned at the most conserved GTT trimer in the −10 region, and the respective consensus sequences were derived (Figure 9). The consensus sequence of the promoters recognized by σ^Η/σ^E was defined as GGAAC–N_17-18–cGTT. This consensus differs in a single conserved base (C instead of T) at position −31 from the core consensus GGAAt–N_18-19–GTT of the σ^Η-controlled genes (Ehira et al., 2009; Busche et al., 2012). However, RNAP+σ^H is also able to initiate transcription from all detected promoters with C₋₃₁. Moreover, in a few exceptional cases, there is G or A at position −31 in the σ^Η-controlled promoters. It seems, therefore, that σ^Η tolerates any base at position −31, whereas only C at this position ensures transcription with RNAP+σ^E (with a single exception of A₋₃₁ in PamtR).

To see the interactions of the σ subunits with the promoter DNA from another perspective and in atomic detail, we involved computer homology modeling into the analysis. We focused on the difference between the −35 regions of the σ^E- and σ^H-dependent promoters in similar way as was done previously for the σ^D and σ^H-dependent promoters (Dostálová et al., 2019).

Recognition of the −35 element of the promoter DNA by stress σ subunits was captured in several X-ray/cryo-electron microscopy structures, which are available in the Protein Data Bank (www.rcsb.org; PDB id: 2H27, 6JBQ, 5ZX2; Lane and Darst, 2006; Guo et al., 2018; Fang et al., 2019). These studies included E. coli and M. tuberculosis RNAP+σ factors. In fact, all these σ subunits in the published models have the same fold. Moreover, their positioning relative to the −35 element of promoter DNA is also the same in all of them.

The 2H27 structure for σ^E from E. coli (Lane and Darst, 2006) was used as a template for creating homology models for C. glutamicum σ^E and σ^H. The non-template/template DNA strand in the 2H27 crystal structure has the GGAACTT/CCTTGAA sequence, which was in fact identical to the sequence of the C. glutamicum Pcg3309 promoter in the key region −35 to −29. Therefore, we used Pcg3309 as an example of a predominantly σ^E-dependent promoter. PtrxB1 (GGAATAA/CCTTATT) was taken as a typical strong exclusively σ^Η-specific promoter.

Our homology models showed that the side chains of arginine R185 (in σ^E) or methionine M170 (in σ^H) interact with nucleotide bases (within the non-coding strand) at positions −31 and − 30 of the promoters (Figures 10A,B).

Based on the conclusions from homology modeling, mutations in the key amino acids arginine R185 (in σ^E) or methionine M170 (in σ^H; i.e., R185 → M185 in σ^E, and M170 → R170 in σ^H) were proposed. The wild-type σ^H, σ^E, and the modified sigma factors σ^Hmut and σ^Emut can form with Pcg3309, PtrxB1, and PcseE in total four unique combinations of key interactions between amino acids R or M and dinucleotides AT and GA at positions −31 and − 30 in the template strand, i.e., R-GA, M-AT, R-AT, M-GA (Figure 10). In addition, there are the combinations M-CT and R-CT that occur with PrshA, which are analogous to M-AT and R-AT. In these combinations, the amino-group of cytosine may interact in a similar way to the amino group of adenine.

The genes encoding the mutant σ factors were constructed and cloned into the expression vector pEC-XT99A. The plasmid constructs of pEC-XT99A with the genes encoding σ^H (M170), σ^E (R185), σ^Hmut (R170), and σ^Emut (M185) and constructs of pEPR1 with the selected promoters were used for creating combinations of the two plasmids (pEC-XT99A + sig and pEPR1 + promoter) in C. glutamicum clones. In addition to the predominantly σ^E-controlled PcseE and Pcg3309 promoters and exclusively σ^H-dependent PtrxB1 promoter, we decided to also test PrshA. This promoter is one of a few exceptions among the strong σ^H-dependent promoters with G at position −31 (instead of T) and it can be thus supposed that M170 in σ^H would interact with −31 and − 30 CT (instead of AT) in the template strand. The promoter activities in the resulting clones were measured by the standard in vivo procedure as the intensity of fluorescence produced by GFPuv (Figure 11).

The use of σ^Εmut and σ^Ηmut enabled us to evaluate the results of promoter activity measurements (Figure 11) in connection with the hypotheses based on homology modeling (Figure 10).

The activity of the predominantly σ^E-dependent promoters PcseE and Pcg3309 vanished completely when the σ^Emut was used (Figures 11C,D). Apparently, their activity depends strictly on R185 in σ^E. We expected that the activity of these promoters, which are to a lower extent also active with σ^H, will be higher with σ^Hmut carrying R170. This is true with PcseE, whereas the activity of Pcg3309 was approximately the same with σ^H and σ^Hmut (Figures 11C,D). Apparently, sequences outside of the conserved motifs −35 and −10 also play significant roles in the activity of the individual promoters. As for the exclusively σ^H-dependent PtrxB1 and PrshA, the mutation in σ^Hmut (M170 → R170) also impaired their activity: to a large extent for PtrxB1 and completely for PrshA (Figures 11A,B). Neither of these strictly σ^Η-specific promoters were recognized by σ^Emut. Although σ^Emut carried methionine occurring in σ^Η in the key position, this single alteration in σ^E did not suffice for the recognition of these σ^Η-specific promoters.

We can conclude that the homology models of σ^E- and σ^H interactions with the respective promoters were generally confirmed, although differences in promoter sequences outside the −10 and − 35 elements can significantly modulate the promoter activity.

Genome-wide differential gene expression analysis by RNA-seq detected 296 upregulated genes using the C. glutamicum ΔcseE strain. This group forms the σ^E stimulon, in which genes directly or indirectly dependent on σ^E activity can be identified. However, only 15 genes were strictly defined as members of the σ^E regulon based on the presence of the σ^E-dependent promoters (clgR is transcribed from 2 σ^E-dependent promoters). Most of these genes (Table 2) are apparently related to maintaining cell envelope integrity or to protein quality control. The cseE gene encoding the cognate anti-σ is a natural part of the regulatory circuit, which controls the σ^E function (Park et al., 2008). The sigB gene, which was previously found to be under control of a σ^Η/σ^E-dependent promoter (Dostálová et al., 2017), encodes the σ factor, which plays a key role in the activation of cell functions essential for C. glutamicum cell survival during the transition and stationary growth phases and in response to various stresses (Larisch et al., 2007). Among them, the effects of acids, ethanol, cold, and heat shock may finally lead to the cell envelope stress response. Thus, the sigB gene is a member of the σ^Η/σ^E regulon.

It was shown that the presence of misfolded proteins in the cell envelope activates σ^E in some bacteria. In agreement with this, the main general function of chaperones such as DnaK and DnaJ2, which are encoded by the σ^H/σ^E -activated genes dnaK and dnaJ2 in C. glutamicum, is to promote proper protein folding and prevent aggregation of the incompletely folded proteins.

Another essential task in protein quality control is removing proteins which have been irreversibly damaged by stresses. This function is fulfilled by Clp proteases. The transcription of the gene clgR that encodes a regulatory protein which activates the expression of these proteases in C. glutamicum, was found to be induced by the heat shock response mediated by σ^H (Engels et al., 2004). In mycobacteria, clgR is also induced by oxidative and detergent stress. These responses are activated by σ^H and σ^E, respectively (Manganelli et al., 2002).

SugR regulator is a transcriptional repressor of the central carbon metabolism of C. glutamicum. Repression of the sugar metabolism may also be connected to the slowed growth during the stress responses.

The function of the cg3344 gene encoding 3-hydroxypropanoate dehydrogenase is related to energy production and conversion. This enzyme is involved in the pathway of pyrimidine degradation which may be connected to the slowed growth rate of the cell culture and energy saving during the defense reaction to stress effects.

The cg1121 gene encodes permease of the MFS type. The product belongs to the large family of transporters (efflux pumps) which confer bacteria resistance to various compounds, e.g., antibiotics. The tatB gene (twin-arginine translocation pathway protein) which is in the operon with sigE and cseE is most likely also connected to the stress response functions in C. glutamicum. The Tat proteins generally have a role in translocating fully folded proteins from the cytoplasm across the cytoplasmic membrane. Among them, redox proteins may be connected to the oxidative stress response (Montero et al., 2019). In E. coli, N-acetylmuramoyl-L-alanine amidase, which is transported by the Tat protein system, acts on the peptidoglycan structure, and may thus be involved at some stage in envelope stress response. Recently, SDS as a cell envelope stressor was found to induce expression of the tatABC operon in Salmonella (Rogers et al., 2022). The same Tat system was also observed to counteract severe oxidative stress and starvation when Bacillus subtilis grows in a NaCl-depleted medium (Prajapati et al., 2021).

Four other genes encode putative membrane proteins which may also be related to cell envelope functions. In conclusion, nearly all uncovered members of the C. glutamicum σ^Η/σ^Εregulon have some function related to the cell envelope and/or various stresses.