Deciphering the Rules Underlying Xenogeneic Silencing and Counter-Silencing of Lsr2-like Proteins Using CgpS of Corynebacterium glutamicum as a Model.

Lsr2-like nucleoid-associated proteins play an important role as xenogeneic silencers (XS) of horizontally acquired genomic regions in actinobacteria. In this study, we systematically analyzed the in vivo constraints underlying silencing and counter-silencing of the Lsr2-like protein CgpS in Corynebacterium glutamicum Genome-wide analysis revealed binding of CgpS to regions featuring a distinct drop in GC profile close to the transcription start site (TSS) but also identified an overrepresented motif with multiple A/T steps at the nucleation site of the nucleoprotein complex. Binding of specific transcription factors (TFs) may oppose XS activity, leading to counter-silencing. Following a synthetic counter-silencing approach, target gene activation was realized by inserting operator sites of an effector-responsive TF within various CgpS target promoters, resulting in increased promoter activity upon TF binding. Analysis of reporter constructs revealed maximal counter-silencing when the TF operator site was inserted at the position of maximal CgpS coverage. This principle was implemented in a synthetic toggle switch, which features a robust and reversible response to effector availability, highlighting the potential for biotechnological applications. Together, our results provide comprehensive insights into how Lsr2 silencing and counter-silencing shape evolutionary network expansion in this medically and biotechnologically relevant bacterial phylum.IMPORTANCE In actinobacteria, Lsr2-like nucleoid-associated proteins function as xenogeneic silencers (XS) of horizontally acquired genomic regions, including viral elements, virulence gene clusters in Mycobacterium tuberculosis, and genes involved in cryptic specialized metabolism in Streptomyces species. Consequently, a detailed mechanistic understanding of Lsr2 binding in vivo is relevant as a potential drug target and for the identification of novel bioactive compounds. Here, we followed an in vivo approach to investigate the rules underlying xenogeneic silencing and counter-silencing of the Lsr2-like XS CgpS from Corynebacterium glutamicum Our results demonstrated that CgpS distinguishes between self and foreign by recognizing a distinct drop in GC profile in combination with a short, sequence-specific motif at the nucleation site. Following a synthetic counter-silencer approach, we studied the potential and constraints of transcription factors to counteract CgpS silencing, thereby facilitating the integration of new genetic traits into host regulatory networks.

INTRODUCTION

Horizontal gene transfer (HGT) is a major driver of bacterial evolution and plays an important role in creating genetic diversity (1). The rapid acquisition of beneficial new traits can create a competitive advantage for the recipient cells (1, 2). However, the chance that foreign DNA decreases the fitness of the cell is high, since it may lead to interference with regulatory networks, high transcriptional and translational costs, sequestration of cellular machineries, and cytotoxic gene products (3–8). Therefore, bacteria evolved a variety of immune systems allowing them to deal with foreign DNA (9). CRISPR-Cas and restriction modification systems are nuclease-based defense mechanisms enabling the recognition and targeted degradation of invading DNA (10–12). In contrast to these destructive immune systems, xenogeneic silencing enables the tolerance of foreign DNA and consequently fosters the acquisition of novel genetic material into the host chromosome (13). Xenogeneic silencing is based on specific nucleoid-associated proteins (NAPs), so-called xenogeneic silencers (XS) (3). Known XS proteins belong to one of four currently described classes: H-NS-like proteins of proteobacteria, like Escherichia coli, Yersinia, and Salmonella (14–16), MvaT/U-like proteins found in gammaproteobacteria of the Pseudomonodales order (17), Lsr2-like XS of Actinomycetes (18, 19), and Rok, present in different bacilli, including Bacillus subtilis (20, 21). Although XS were convergently evolved and show only low sequence similarity within the different classes, the domain properties of their N-terminal oligomerization domains and their C-terminal DNA-binding domains are similar (19, 20, 22, 23). Their binding mechanisms are diverse, but they all preferentially bind to horizontally acquired DNA, which typically has a higher AT content than the genome of the recipient cell (4, 24). The broad distribution of XS among prokaryotes emphasizes the strong need to discriminate between self and non-self across phylogenetic clades (13). Even so, the GC content of microbial genomes dramatically varies, from 75% (Actinobacteria) to less than 20% (bacterial endosymbionts) (25, 26), and horizontally acquired regions typically feature a lower GC content than their resident genome, emphasizing base composition as a major discrimination factor shaping microbial genome evolution (3).

Several studies based on variants defective in oligomer formation revealed that binding of XS proteins to the DNA alone is insufficient for silencing (27–29). The formation of higher-order nucleoprotein complexes instead mediates silencing of the target genes by occlusion or trapping of the RNA polymerase, by interference with the transcription elongation complex, or by enhancing termination (30, 31). To get access to potentially encoded beneficial traits, cells must integrate foreign genes into preexisting regulatory circuits, allowing their controlled expression at appropriate time points and physiological or environmental conditions (32, 33). In contrast to classical activation, counter-silencing is based on the interference of a DNA-binding protein, e.g., a transcription factor (TF), with the silencer-DNA complex leading to transcription initiation without depending on the direct interaction with the RNA polymerase (32, 33). Counter-silencing of H-NS was addressed by several studies either by following a synthetic approach at well-studied promoters (34, 35) or by the analysis of the promoter architectures in the PhoPQ regulatory network (33). The recent study by Will et al. emphasizes that the principle of H-NS xenogeneic silencing and counter-silencing provides a certain degree of flexibility, fostering evolutionary network expansion (33).

Compared to H-NS in proteobacteria, much less is known about Lsr2-like XS proteins conserved throughout the actinobacteria. In Mycobacterium tuberculosis, Lsr2 acts as a master regulator of multiple virulence-associated genes (19, 22) and was suggested to be involved in the manifestation of multidrug tolerance (36). The essentiality of Lsr2 for this human pathogen makes this XS protein a highly promising drug candidate (37). In Corynebacterium glutamicum, the Lsr2-like XS protein CgpS also was shown to play an essential role as a silencer of cryptic prophage elements whose entrance into the lytic cycle would otherwise cause cell death (4, 38). In contrast to mycobacteria and corynebacterial species, Streptomyces species typically encode two Lsr2-like proteins. Here, the prototypical lsr2 gene, showing the highest sequence identity to mycobacterial Lsr2, was recently described to silence the expression of specialized metabolic clusters (39). Considering the important role of Lsr2 proteins in the medically and biotechnologically important phylum of Actinobacteria, a detailed mechanistic understanding of Lsr2 binding in vivo is relevant as a potential drug target and for the identification novel bioactive compounds.

In this study, we set out to systematically assess the rules underlying silencing and counter-silencing of Lsr2-like XS by using the Lsr2-like protein CgpS of Corynebacterium glutamicum as a model (4). To the best of our knowledge, this is the first detailed analysis of the counter-silencing mechanism of an Lsr2-like XS protein. Bioinformatic analysis of CgpS ChAP-seq (chromatin affinity purification and sequencing) data revealed a clear preference of CgpS toward AT-rich stretches containing A/T steps (alternation of A to T and vice versa). In vivo reporter studies with synthetic promoter variants verified the importance of a distinct drop in GC profile and revealed the overrepresentation of a short, sequence-specific motif at CgpS target regions. Insertion of TF operator sites at different positions within various CgpS target promoters was shown to counteract CgpS silencing, showing the most prominent effect at the position of maximal CgpS binding. With this approach, we provide important insights into the in vivo constraints of Lsr2 counter-silencing and contribute to an understanding of how bacteria can evolve control over the expression of horizontally acquired genes.

RESULTS

In vivo analysis of CgpS binding preferences.

Recent genome-wide profiling studies revealed that the Lsr2-like xenogeneic silencer CgpS preferentially binds to AT-rich DNA sequences in the genome of C. glutamicum ATCC 13032 (4). To determine the parameters affecting CgpS binding and silencing in vivo, we systematically analyzed the peak sequences obtained from CgpS ChAP-seq analysis (4) and subsequently verified our conclusion by testing the silencing of synthetic promoter variants. Remarkably, an overlay of the GC profiles of all 35 CgpS target promoters located within the prophage element CGP3 revealed a high degree of similarity with a distinct drop in GC content matching the position of maximal CgpS coverage (Fig. 1A). Genome-wide analysis of AT-rich genomic regions revealed that the fraction of CgpS-bound sequences increased with the length of the particular AT stretch. While increasing numbers of G/C interruptions (occurrence of G or C within an AT stretch) negatively influenced the proportion of CgpS-bound targets (Fig. 1B), a larger number of A/T steps (alternation of A to T and vice versa) increased the fraction of CgpS-bound sequences by trend (Fig. 1C). This trend became especially evident in the case of AT-rich stretches of medium length (14 to 30 bp).

FIG 1

CgpS preferentially binds to long and consecutive AT stretches. (A) Overlay and calculated mean (orange curve) of GC profiles of CgpS target promoters located within the CGP3 prophage (n = 35) (4). Profiles were calculated by a rolling mean with a window size of 50 bp and a step size of 10 bp. The GC profiles of the promoters were normalized regarding the orientation and position of the maximal CgpS binding peak (blue line), which was defined for all sequences as position 0. The mean GC content of the C. glutamicum genome (69) is shown as a red line (53.8%). (B and C) Genome-wide analysis of CgpS binding to consecutive AT stretches of different lengths considering G/C interruptions (occurrence of G or C within an AT stretch) (B) or number of A/T steps (allowing up to five G/C interruptions) (C). A/T steps are defined as alterations of A to T and vice versa. The value in the array represents the number of stretches found in the C. glutamicum genome fitting the respective criteria, while the color indicates the fraction of CgpS targets per array. (D) Inverse correlation of GC profiles and CgpS coverage of CgpS target promoters. CgpS coverage obtained from previous ChAP-seq experiments (4) was calculated with a rolling mean with a window size of 50 and a step size of 10. All identified TSS (see Materials and Methods and Text S1) are shown in Table S1 and represented as vertical black, gray, and red lines (mapped according to their enrichment scores: black > shades of gray > red). Positions of maximal CgpS coverage and average GC content are shown as described for panel A. The corresponding genes are shown as gray arrows. Promoters were grouped into two classes based on the width of the region bound by CgpS (class 1 promoters, 500 to 850 bp, typically featuring one distinct drop in GC profile; class 2 promoters, >850 bp, often broader and containing multiple drops in GC content). As a negative control, the non-CgpS target promoter of the gene gntK is shown. a.u., arbitrary units. (E) Frequency distribution of relative positions of all new identified TSS (yellow) of CgpS target promoters referred to the position of maximal CgpS binding. TSS showing the highest enrichment scores per gene are highlighted in gray.

Supplementary information on methods used in this study. Download Text S1, PDF file, 0.2 MB.

TSS analysis of relevant phage and non-phage-related CgpS target promoters after prophage induction with mitomycin C. Download Table S1, XLSX file, 0.03 MB.

Overall, this analysis suggested that long and consecutive AT stretches represent the main determinant of CgpS target binding. Individual inspection of CgpS-targeted phage promoters revealed a significant correlation between the CgpS peak maximum and the GC minimum in this area (Fig. 1D). Depending on the widths of the CgpS coverage peaks, promoters were grouped into two classes. Class 1 consists of promoters with peak widths between 500 and 850 bp, which typically show one distinct drop in GC profile, while CgpS coverage peaks of class 2 promoters are wider than 850 bp and the corresponding GC profiles often feature broader or multiple drops.

Due to efficient CgpS-mediated silencing of gene expression, most transcriptional start sites (TSS) of CgpS target promoters had not been identified in previous studies (40). It represents, however, an advantage of the chosen model system that expression of the majority of CgpS targets can be induced by triggering prophage induction using the DNA-damaging antibiotic mitomycin C. To provide comprehensive insights into the promoter architecture of CgpS targets, TSS were determined under conditions triggering phage gene expression (600 nM mitomycin C). For 46 out of all 54 CgpS target promoters, at least one TSS was identified (for 31 out of 35 prophage promoters) (see Table S1 in the supplemental material). Strikingly, the analysis of the relative distances between the positions of TSS and maximal CgpS binding revealed that in the majority of CgpS target promoters, TSS are located close to the position of maximal CgpS coverage and GC minimum (Fig. 1D and E; see Table S1 for the complete data set).

Design, build, and test: relevance of a DNA motif for CgpS binding and silencing.

Bioinformatic analysis of CgpS target sequences confirmed the preference of CgpS for AT-rich DNA sequences. However, neither the distinct drop in GC content nor the occurrence of long and consecutive AT stretches were unique to CgpS targets (Fig. 1), indicating that additional parameters support CgpS to specifically recognize its targets. Interestingly, a MEME-ChIP analysis (41) on CgpS-bound promoter sequences revealed a 10-nucleotide-long AT-rich binding motif (E value, 5.2 × 10−9) containing A/T steps (Fig. 2A), which was found in 51 of 54 bound promoter regions. Remarkably, the presence of this motif within AT-rich stretches of different lengths significantly increased the fraction of CgpS-bound sequences by a factor of up to 2.8-fold (Fig. 2A). However, the genome-wide search for motif occurrence using the online tool FIMO (Find Individual Motif Occurrences) (42) revealed that about 85% of the motifs (669/785) within the C. glutamicum genome were not bound by CgpS, indicating that the motif alone is not sufficient to permit CgpS binding.

FIG 2

Synthetic in vivo approach to dissect the relevance of the GC profile and a sequence-specific binding motif for CgpS silencing. (A) Identified 10-bp CgpS binding motif using MEME-ChIP (41) analysis found within 51 of 54 CgpS target promoters (4) (E value, 5.2 × 10−9). The bar plot represents the genome-wide fraction of CgpS targets in AT stretches of different length, allowing up to 5 G/C interruptions with or without the identified motif. (B) CgpS silencing of synthetic constructs (Pcg1999_A-T/G-C and Pcg1999_rand), based on a 50-bp core promoter region (green box) of the phage gene cg1999. The fixed 50-bp DNA sequence covered the −10 and −35 box, positions of the TSS (40), and potential binding motif (gray box). The adjacent sequence (N upstream, 260 bp; N downstream, 48 bp) was either adjusted to maintain the native density of AT stretches (Pcg1999_A-T/G-C, exchange of A to T and G to C) or randomized (Pcg1999_rand). (C) Surface plasmon resonance analysis of CgpS binding to the synthetic promoter Pcg1999_A-T/G-C (423 bp) compared to that of the negative-control Pcg3336 (424 bp) and the corresponding native CgpS target Pcg1999 (423 bp). ka, association constant; kd, dissociation constant. (D) CgpS silencing of synthetic constructs based on fixed 70- to 100-bp promoter regions of the phage gene lys. The 70-bp sequence (green box) covered the −10 and −35 box and TSS but only half of the putative motif. The 80-bp region (green and orange boxes) covered the motif completely, and the 100-bp region (all boxes) additionally covered the position of maximal CgpS coverage. The adjacent sequences (N upstream, 304 bp; N downstream, 70 to 100 bp) were adjusted to maintain the native density of AT stretches (A-T/G-C). (B and D) Reporter outputs (Venus) of the native and corresponding synthetic variants (plasmid backbone pJC1) in wild-type and Δphage (ΔcgpS) strains after 5 h of cultivation in a microtiter cultivation system in CGXII medium containing 100 mM glucose. Shown are mean values and standard deviations from biological triplicates. All synthetic sequences are listed in Table S2I.

Strains, plasmids, oligonucleotides, and DNA sequences. (A) Strains used in this study. (B) Plasmids from other studies used in this study. (C) Plasmids constructed in this work. (D) Oligonucleotides used in this study for plasmid construction. (E) Oligonucleotides used for the amplification of DNA probes for surface plasmon resonance analysis. (F) Oligonucleotides used for electrophoretic mobility shift assays. (G) Oligonucleotides used for sequencing. (H) Ordered DNA sequences. (I) Sequences of the native phage promoters Pcg1999 and P and the corresponding synthetic variants shown in Fig. 2. Download Table S2, DOCX file, 0.1 MB.

In the following experiments, we used an in vivo approach to test whether the combination of the motif and the drop in GC profile are sufficient for CgpS-mediated silencing of gene expression. For this purpose, different synthetic promoter variants were designed based on the 50- to 70-bp core promoter regions of the phage genes Pcg1999 and P. Both promoters were highly active in the absence of CgpS, indicating that the chosen core regions efficiently drive transcription. In the case of Pcg1999, the DNA sequence containing the core promoter elements (−10 and −35 box and TSS) and the predicted binding motif (shown in Fig. 2A) was kept constant (Fig. 2B). The adjacent sequence was either designed to mimic the native GC profile of Pcg1999 (exchange of A to T and G to C and vice versa, Pcg1999_A-T/G-C) or contained a randomized sequence varying in GC profile and sequence (Pcg1999_rand). The resulting promoter designs were fused to a gene encoding the yellow fluorescent protein Venus. In line with our hypothesis, the construct Pcg1999_A-T/G-C, featuring the native GC profile, was efficiently silenced by CgpS in the wild-type strain and displayed even lower reporter output than the native phage promoter Pcg1999 (Fig. 2B). Surface plasmon resonance (SPR) analysis revealed CgpS binding kinetics and affinities for this synthetic promoter (equilibrium dissociation constant [K], 42 nM) similar to those of the corresponding native CgpS target promoter Pcg1999 (K = 58 nM). CgpS also interacted with the control promoter fragment Pcg3336 but with much lower affinity (K = 381 nM) and very fast dissociation rates (Fig. 2C). In the prophage-free strain Δphage, which lacks the phage-encoded cgpS gene, the reporter output was significantly higher for all tested promoter fusions, confirming that all designs functionally drive transcription. Silencing of the promoter variant with randomized adjacent flanks was strongly impaired, demonstrating that the motif-containing 50-bp core promoter region alone did not mediate silencing. This highlights the importance of the overall drop in GC content observed at CgpS target promoters (Fig. 2B).

The relevance of the identified motif was verified using synthetic promoter designs of the phage promoter P. Here, constructs carrying only parts of the predicted motif (70-bp core) did not permit silencing, while constructs covering the motif entirely enabled silencing (Fig. 2D). In all P-based synthetic constructs, the native GC profile of the sequence flanking the core promoter region was mimicked but the DNA sequence was changed (A to T and G to C and vice versa). This in vivo analysis of synthetic phage promoter variants revealed that efficient CgpS silencing depended on both specific DNA sequences (binding motif) and the drop in GC profile.

Synthetic disruptive counter-silencing.

Disruptive counter-silencing was previously described as a mechanism that may provide access to horizontally acquired genes silenced by nucleoid-associated proteins (32). To study the potential and constraints of evolutionary network expansion by counter-silencing of CgpS target promoters, a synthetic counter-silencer (CS) design was applied in this study (Fig. 3A). At native target promoters (e.g., Pcg1999 or P), oligomerization of the xenogeneic silencer CgpS leads to the formation of a nucleoprotein complex inhibiting transcription (4, 31). In the following experiments, we used a set of 12 different phage promoters as a basis for synthetic CS constructs and inserted the operator sequence of an effector-responsive transcription factor (TF) into the silenced promoter regions. We postulated that binding of the TF to its operator sequence would interfere with the silencer nucleoprotein complex and thereby mediate counter-silencing (Fig. 3A). To avoid interference of the inserted operator site with CgpS-mediated silencing, we chose the operator site of the functionally redundant TFs GntR1 (Cg2783) and GntR2 (Cg1935) (summarized as GntR in the following), which bind to a well-defined short (15 bp) and AT-rich (GC content, 27%) DNA motif (43). One of the native targets of GntR is the promoter of the gntK gene, which is repressed by binding of GntR. The P promoter and the synthetic promoter constructs were fused via a consistent linker containing a ribosomal binding site (RBS) to the reporter gene venus and were inserted into the plasmid pJC1. The effector molecule gluconate was shown to act as an inducer triggering the dissociation of GntR from its operator site (43), consequently leading to derepression of P (Fig. 3B).

FIG 3

Synthetic approach to study disruptive counter-silencing. (A) Schematic overview of a native CgpS target promoter (phage) and the corresponding synthetic counter-silencer construct. (B) Signal inversion by synthetic counter-silencing. Comparison of the reporter outputs of P, the native target promoter of the regulator of gluconate catabolism GntR (43), and the synthetic GntR-dependent counter-silencer promoter P_CS_0. C. glutamicum wild-type strains harboring the plasmid-based constructs (pJC1) were cultivated in the absence of the effector (111 mM glucose) or in its presence (100 mM gluconate) in a microtiter cultivation system. Graphs represent the means and error bars the standard deviations from biological triplicates. Backscatter and fluorescence were measured at 15-min intervals. (C) Counter-silencing efficiency of different phage promoters with inserted GntR binding sites located directly upstream of the position of maximal CgpS binding. Promoters were grouped into two classes based on the width of the region bound by CgpS (class 1 promoters, 500 to 850 bp, often one distinct drop in GC profile; class 2 promoters, >850 bp, often broader or multiple drops in GC content). CgpS coverage and GC profiles of two representative promoters are shown. The highest-ranked TSS are marked as vertical gray lines and the position of maximal CgpS binding as vertical blue lines. GC profiles of all used phage promoters are shown in Fig. 1D. C. glutamicum wild-type cells harboring the plasmid-based (pJC1) counter-silencers were cultivated in the presence (100 mM gluconate) or absence (100 mM glucose) of the effector molecule gluconate in a microtiter cultivation system. Fold change ratios of Venus reporter outputs in the absence and in the presence of the effector were calculated based on the specific reporter outputs after 5 h of cultivation (Fig. S1A). Dots represent the means and error bars the standard deviations from at least biological triplicates. Yellow dots demonstrate counter-silencing (activated by GntR binding), while blue dots represent repression (repressed by GntR binding). Promoters, which did not show significant changes in reporter output, are shown as gray dots (t test; P < 0.05).

Counter-silencing of CgpS target promoters. (A) Specific reporter outputs of native phage promoters and corresponding synthetic constructs that were used for the calculation of fold change ratios shown in Fig. 3. (B) To confirm functionality of synthetic promoter variants (GntR BS), which showed only low reporter outputs in the wild type, respective constructs were analyzed in the prophage-free strain Δphage in the absence of the silencer CgpS. All constructs led to significant reporter outputs, confirming synthetic promoter functionality (n ≥ 3 biological triplicates). Download FIG S1, TIF file, 0.2 MB.

Monitoring of fluorescent outputs driven by phage-based synthetic promoter constructs allows the in vivo analysis of silencing and counter-silencing efficiencies. GntR operator sites were indeed confirmed as suitable candidates for the construction of counter-silencers, since the insertion into different phage promoters led to only slightly increased background expression levels in the wild-type strain in the presence of the effector molecule (Fig. S1A). The insertion of a GntR binding site (BS) within the CgpS-silenced phage promoter P (P_CS_0) led to effector-dependent reporter outputs. GntR binding resulted in an increased reporter output of the counter-silencer construct P_CS_0 when glucose was added as a carbon source, while gluconate (effector addition) triggered the dissociation of GntR, leading to silencing of promoter activity by CgpS (Fig. 3B). This is especially remarkable considering that the binding site was inserted at the position of maximal CgpS coverage close to the annotated TSS (27 bp downstream [Table S1]). Based on textbook knowledge, this position would rather fit to a repressor function (44, 45). In the case of P, the GntR binding site overlaps the TSS, leading to repression of gene expression (43). In the context of xenogeneic silencing, however, GntR binding appeared to efficiently interfere with CgpS silencing. Thus, in contrast to the native GntR target P, the synthetic P-counter-silencer promoter was activated in the absence of the effector molecule. Although both promoters (P and P_CS_0) were completely different and had only the 15-bp-long GntR binding site in common, they showed very similar but inverted responses to effector availability (Fig. 3B). This demonstrates the potential of the counter-silencing principle to convert a repressor to an activating, tunable counter-silencer, thereby facilitating the expansion of regulatory networks.

Disruptive counter-silencing is most efficient at the CgpS nucleation site.

To systematically assess the constraints of counter-silencing, 12 representative phage promoters of both classes (eight class 1 and four class 2) were selected as targets to test the efficiency of synthetic counter-silencing. The GntR binding site was inserted directly upstream of the previously identified position of maximal CgpS binding obtained from ChAP-seq analysis (4). To study counter-silencing efficiency, all constructs were analyzed in C. glutamicum wild-type cells in the presence and absence of the effector molecule gluconate. The ratio of maximal (− effector; GntR binding) and minimal (+ effector; GntR dissociation) reporter outputs was used to compare the counter-silencing efficiency of the different constructs (Fig. 3C and Fig. S1). Overall, counter-silencing appeared to be more efficient in class 1 promoters typically featuring a bell-shaped CgpS peak and a distinct drop in GC profile. Here, six out of eight constructs showed an effector-responsive counter-silencing behavior. In contrast, only one of four class 2 promoters was activated by GntR binding. The broader regions bound by CgpS are probably stabilizing the silencer-DNA complex, compensating for the local interference effects caused by GntR binding. The general functionality of promoter variants (Fig. S1A) was confirmed in the strain Δphage, where all variants showed a significant fluorescent signal (Fig. S1B). Interestingly, GntR binding to Pcg2014 led to counter-silencing in the wild type but to slight repression in the Δphage strain, suggesting that only the destructive interference between CgpS and GntR facilitates efficient transcription of the downstream gene.

Silencing is mediated by CgpS binding, and counter-silencing depends on GntR binding.

CgpS as silencer and GntR as counter-silencer are the two key components of the synthetic counter-silencer approach presented in this study. To confirm their presumed functions, mutant analysis and in vitro binding assays with both proteins were performed. The reporter outputs of the native phage promoter P as well as of the corresponding counter-silencer construct (P_CS_0) were analyzed in C. glutamicum wild-type cells and different mutant strains. In the wild type, the counter-silencing construct showed the expected increase of reporter output upon GntR binding (− effector). In line with the assumed counter-silencing function of GntR, both constructs featured a low reporter output in the ΔgntR strain lacking both functionally redundant GntR1 and GntR2 regulators (Fig. 4A). To confirm the relevance of the inserted GntR operator sequence, different mutated variants were tested as well. Here, neither the insertion of a randomized operator sequence, identical in length and nucleotide composition, nor a mutated operator site, where only one conserved base in the GntR motif was exchanged, led to counter-silencing of the P promoter (Fig. S2). These results confirmed that counter-silencing directly depends on GntR binding. However, the insertion of a reverse-oriented GntR binding site within the silenced promoter allowed counter-silencing, showing that this mechanism does not depend on the directionality of the binding site (Fig. S3). P and the corresponding counter-silencer construct showed strongly increased promoter activities in the Δphage strain in the absence of CgpS, suggesting that CgpS is responsible for silencing. Effector-dependent activation was abolished in the absence of CgpS, indicating that GntR acts as a counter-silencer rather than as a classical activator. Reintegration of the cgpS gene into the Δphage strain, resulting in Δphage::P-cgpS, confirmed CgpS as the only factor responsible for silencing of the native phage promoter P and, thus, emphasized that CgpS function does not depend on further phage-encoded accessory proteins (Fig. 4A).

FIG 4

Silencing is mediated by CgpS, while counter-silencing depends of GntR binding. (A) Reporter output (venus expression) of different C. glutamicum strains carrying the native P promoter or the counter-silencing design P_CS_0 after 5 h of cultivation. Both constructs were analyzed in C. glutamicum wild-type cells, in a gntR1-gntR2 double deletion strain, in the prophage-free strain Δphage (lacking the phage-encoded cgpS), and in its variant with reintegrated cgpS under the control of its native promoter (Δphage::P-cgpS). Cells were cultivated in a microtiter cultivation system in CGXII medium supplemented with either 100 mM gluconate (+ effector) or 100 mM fructose (− effector). (B) EMSA of GntR binding to DNA fragments covering the synthetic counter-silencer promoter P_CS_0 (533 bp, 14 nM) or the native phage promoter P (518 bp, 14 nM). (C) Impact of the effector molecule gluconate on binding of GntR to the synthetic counter-silencer construct. EMSA was performed as described for panel B, but GntR and the DNA fragments were incubated either in the presence of the effector (100 mM gluconate) or in its absence (100 mM glucose). (D) Surface plasmon resonance analysis of CgpS binding kinetics to biotinylated DNA fragments covering the negative-control Pcg3336 (424 bp), the native phage promoter P (424 bp), or the corresponding synthetic counter-silencer construct (439 bp) that were captured onto a streptavidin-coated sensor chip. Different concentrations of CgpS were passed over the chip using a contact (association) time of 180 s, followed by a 420-s dissociation phase. The increase in response units correlates with increasing CgpS concentrations.

Effects of inserted sequences independent of GntR binding. (A) The native GntR BS located within the P promoter contains four highly conserved and four weakly conserved nucleotides (43). As controls for sequence insertions, which do not allow for GntR binding, control sequences (control-seq) 1 and 2 were inserted at position 0 within the phage promoter P. Control-seq 1 contained the same nucleotide composition as the GntR BS, but the sequence was randomized. Control-seq 2 was similar to the GntR BS, but one conserved cytosine was replaced with guanine. Based on previous studies, this exchange was expected to abolish GntR binding (43). (B) Specific reporter outputs of the promoter variants after five hours of cultivation. Promoter constructs (plasmid backbone pJC1) were fused to the reporter gene venus and analyzed in wild-type C. glutamicum. Cells were cultivated in CGXII in the presence (100 mM gluconate) or absence (100 mM glucose) of the effector molecule gluconate in a microtiter cultivation system. The results shown in this graph demonstrate that mutation of the GntR operator site abolished GntR binding in vivo and that the observed counter-silencing effect strictly depends on GntR binding. Download FIG S2, TIF file, 0.2 MB.

Effects of directionality of inserted GntR binding site on counter-silencing efficiency. The 15-bp-long GntR BS was inserted either in a forward (P::GntR BS_fw, P_CS_0) or in a reverse orientation (P::GntR BS_rv) in the P promoter (position 0) to analyze if the directionality of the GntR BS influences counter-silencing. Promoter constructs (plasmid backbone pJC1) were fused to the reporter gene venus and analyzed in the C. glutamicum wild type. Cells were cultivated in CGXII medium in the presence (100 mM gluconate) or absence (100 mM glucose) of the effector molecule gluconate in a microtiter cultivation system. Presented is the mean specific fluorescence after five hours of cultivation (n = 3 biological triplicates). These results demonstrated that counter-silencing does not depend on the directionality of GntR operator sites. Download FIG S3, TIF file, 0.1 MB.

As a further piece of evidence, electrophoretic mobility shift assays (EMSA) were performed to confirm the specific binding of GntR to the synthetic counter-silencing construct (P_CS_0) in vitro. In contrast to the native phage promoter, the P fragment containing the GntR operator site showed a significant shift at low GntR concentrations, confirming specific GntR binding to P_CS_0 (Fig. 4B). Addition of the effector molecule gluconate led to dissociation of GntR (Fig. 4C), which is in agreement with previous reports (43). Surface plasmon resonance analysis of CgpS binding to DNA fragments covering either P or the synthetic counter-silencer construct P_CS_0 showed comparable high-affinity binding of CgpS to both promoters (K, P, 11 nM; P_CS_0, 20 nM) (Fig. 4D).

Impact of operator site position.

When analyzing the promoter architecture of horizontally acquired gene clusters, previous studies revealed a certain variability (33). To systematically assess the potential and constraints of the counter-silencing mechanism for evolutionary network expansion, we analyzed the impact of operator site position on counter-silencing efficiency. Therefore, the GntR binding site was inserted at different positions using the prophage promoter P as a test case (Fig. 5A). Position 0 is defined as the position located directly upstream of the nucleotide featuring maximal CgpS binding in ChAP-seq studies (4). The position of maximal CgpS binding was located 27 bp downstream of the TSS. C. glutamicum wild-type cells harboring the plasmid-based constructs [pJC1-P::GntR BS_pos(variable)-venus] were cultivated in the presence or absence of the effector molecule gluconate. Induced and noninduced reporter outputs were strongly influenced by the binding site position. This demonstrated that the inserted binding site itself, depending on its position, already interferes with the silencer-DNA complex (Fig. 5B). Comparison of the fold change ratio of reporter outputs in the absence and presence of the effector gluconate revealed that the construct with the GntR binding site located directly upstream of the maximal CgpS binding peak (position 0) showed the highest dynamic range (∼5-fold). This dynamic range decreased when the operator was inserted between 15 bp upstream (−15) and 10 bp downstream (+10) of the maximal CgpS binding peak, but constructs still showed counter-silencing in the absence of the effector (Fig. 5C). However, analyzing the −15 promoter variant in the absence of CgpS (Δphage) revealed repression caused by GntR binding, demonstrating again that the observed counter-silencing effect is a result of regulatory interference (Fig. S4). GntR binding sites located at greater distances led in most cases to relatively low reporter outputs. Here, the expression level tended to be higher when GntR binding was inhibited, suggesting GntR acts mainly as a repressor of gene expression at these positions (Fig. 5B and C). A similar trend was observed for the phage promoter Pcg1999 (Fig. S5). Altogether, these results demonstrated that the impact of GntR binding on promoter activity strongly depends on the context of xenogeneic silencing. While interference with CgpS binding triggered promoter activation by counter-silencing, GntR binding in the absence of CgpS often lowered the reporter output. Analysis of reporter outputs driven by 5′-truncated promoter variants of P and P_CS_0 revealed that the region >89 bp upstream of the maximal CgpS binding peak is not required for silencing or counter-silencing (Fig. S6).

FIG 5

Impact of GntR operator position on inducibility of P-based promoter constructs. (A) Inverse correlation of GC profile and CgpS binding coverage (4) of the phage promoter P. The transcriptional start site (TSS) and the position of maximal CgpS binding are shown as vertical lines. Binding site positions (also used in panels B and C) refer to the sequence base associated with maximal CgpS binding. The position directly upstream of this nucleotide was defined as position 0. (B) Impact of inserted GntR binding site position on specific reporter outputs in the presence (gluconate) and absence (glucose) of the effector molecule after 5 h of cultivation. Positions of TSS and maximal CgpS coverage are marked by horizontal lines, and the range of the putative CgpS binding motif is shown. (C) Impact of GntR binding site position on counter-silencing efficiency of P-based promoter constructs. Ratio of specific reporter outputs, shown in panel B, were used for the calculation of their inducibility (fold change). Cells harboring the plasmid-based synthetic promoter constructs were grown in CGXII medium supplemented with either 111 mM glucose or 100 mM gluconate. Bars (B) and dots (C) represent the means and error bars the standard deviations from at least biological triplicates.

It is all about context: impact of GntR binding on the output of the P promoter in the presence or absence (Δphage) of CgpS. Reporter outputs (venus expression) driven by a P promoter with an inserted GntR operator sequence 15 bp upstream of the maximal CgpS binding peak (11 bp downstream of the TSS) in C. glutamicum wild-type cells and in the absence of CgpS in the strain Δphage. Cells were cultivated in CGXII in the presence (100 mM gluconate) or absence (111 mM glucose) of the effector molecule gluconate in a microtiter cultivation system. Shown are mean values and standard deviations of reporter outputs from biological triplicates after five hours of cultivation. Download FIG S4, TIF file, 0.1 MB.

Impact of GntR binding site (BS) position on inducibility of Pcg1999-based promoter constructs. (A) Inverse correlation of GC profile and CgpS binding coverage of the phage promoter Pcg1999. The highest scored transcriptional start site (TSS) and the position of maximal CgpS binding affinity (4) are shown as vertical lines. BS positions refer to the sequence base associated with maximal CgpS binding peak. The position directly upstream of this nucleotide was defined as position 0. (B) Impact of inserted GntR BS position on specific reporter outputs in the presence (gluconate) and absence (glucose) of the effector molecule gluconate. Positions of TSS and maximal CgpS coverage are marked by lines. (C) Impact of GntR binding site position on counter-silencing efficiency of Pcg1999-based promoter constructs. Ratios of specific reporter outputs, shown in panel B, were used for the calculation of their inducibility (fold change). Cells harboring the plasmid-based synthetic promoter constructs were grown in CGXII medium supplemented with either 100 mM glucose or 100 mM gluconate. Bars (B) and dots (C) represent the means and error bars the standard deviations from biological triplicates. Download FIG S5, TIF file, 0.3 MB.

Definition of the minimal region required for silencing. (A) Reporter outputs (venus expression) driven by 5´-truncated promoter versions (Δ300 and Δ350 truncations) of P and P_CS_0 were analyzed regarding silencing and counter-silencing efficiency. 5´ Ends of full-length constructs coincided with the upstream end of the CgpS binding peak (P, 389 bp; P_CS_0, 404 bp upstream of the maximal CgpS binding peak). The distance between maximal CgpS binding peak and ATG was 85 bp in all constructs. Full-length promoters were compared to variants ending 89 bp (5´ Δ300) or 39 bp (5´ Δ350) upstream of the maximal CgpS binding peak in P. (B and C) Specific reporter outputs of the promoter variants after five hours of cultivation. Promoter constructs were fused to the reporter gene venus (plasmid backbone pJC1) and analyzed in C. glutamicum wild-type cells (B) and in the Δphage (ΔcgpS) prophage-free strain (C). Cells were cultivated in CGXII in the presence (100 mM gluconate) or absence (100 mM glucose) of the effector molecule gluconate in a microtiter cultivation system. Shown are mean values and standard deviations from biological triplicates after five hours of cultivation. The results shown in this graph demonstrate that the region >89 bp upstream of the maximal CgpS binding peak of the P promoter is involved in neither silencing nor counter-silencing. A further 50-bp truncation also strongly reduced promoter activity of both promoters in the Δphage strain. Download FIG S6, TIF file, 0.2 MB.

Implementation in a genetic toggle switch.

The P counter-silencing construct (P_CS_0) and the native GntR target promoter P showed a very similar promoter output but an inverted response to effector availability. While P is repressed by binding of GntR, the counter-silencer promoter is activated by GntR binding in the absence of the effector molecule gluconate (Fig. 3B). Both promoters were combined in a gluconate-dependent, GntR-controlled genetic toggle switch. To monitor the switching between different expression states, P_CS_0 was fused to the reporter gene venus, while the native GntR target promoter P was fused to the reporter gene e2-crimson (Fig. 6A). Since the toggle output is only regulated by GntR binding, native GntR levels could be used for toggle control, avoiding a negative impact of artificial TF levels on cellular growth. C. glutamicum wild-type cells harboring the plasmid-based toggle (pJC1-P_CS_0-venus-T-P-e2-crimson) were cultivated in a microfluidic chip device (46) in minimal medium containing either gluconate (+ effector) or glucose (− effector) as carbon source. The carbon sources were switched after the first 17 h. The following time-lapse microscopy analysis revealed that the output of this synthetic toggle is reversible, as shown by the rapid changes in reporter outputs (Fig. 6B). This overall design principle allows the control of the toggle by only one specific effector-responsive TF and features a robust and reversible response to effector availability, highlighting the potential of this toggle for biotechnological applications, for example, to switch between biomass production and product formation.

FIG 6

Implementation of P and P_CS_0 in a genetic toggle switch. (A) Scheme of the designed toggle switch based on the native GntR target promoter P and the synthetic GntR-dependent P counter-silencer construct. In order to monitor their activities, the promoters were fused to different reporter genes (P_CS_0-venus and P-e2-crimson). The promoter reporter gene fusions were oriented in opposite directions and separated by a short terminator sequence. (B) Dynamic switch between both reporter outputs. C. glutamicum wild-type cells harboring the plasmid-based toggle were cultivated in a microfluidic cultivation system (46) with continuous supply of CGXII medium supplemented either 111 mM glucose or 100 mM gluconate and analyzed by time-lapse microscopy at 20-min intervals. Switch of carbon source supply was performed after the first 17 h. This time point was defined as T0. The graphs show the specific fluorescence of three independent microcolonies (circles, squares, and triangles) over time, and images display one representative colony.

DISCUSSION

The nucleoid-associated protein Lsr2 is conserved throughout the phylum of Actinobacteria, where it plays an important role in the xenogeneic silencing of horizontally acquired genomic regions (4, 18, 19, 39). The Lsr2-like protein CgpS was recently described as a classical XS protein silencing the expression of cryptic prophage elements and further horizontally acquired elements in C. glutamicum (4).

CgpS binds DNA with a distinct drop in GC content close to the TSS.

In this study, we systematically assessed the promoter architecture of CgpS targets as well as the constraints underlying silencing and counter-silencing of target gene expression. The genome-wide analysis of CgpS-bound regions obtained from ChAP-seq analysis revealed that CgpS targets share a distinct drop in GC content (Fig. 1). The binding to AT-rich DNA is a common feature of XS proteins (3) and was shown to represent an important fitness trait to avoid spurious transcription and the sequestering of the RNA polymerase (8). The importance of this drop in GC content was confirmed by measuring the CgpS-mediated silencing of Pcg1999-based synthetic variants, where changes in GC profile abolished silencing (Fig. 2B). The analysis of relative distance between the TSS and position of maximal CgpS binding emphasized that binding of CgpS close to the TSS is important for efficient silencing (Fig. 1E).

CgpS recognizes a sequence-specific binding motif containing A/T steps.

In vitro protein binding microarray experiments revealed a clear preference of the xenogeneic silencers H-NS, MvaT, and Rok for DNA stretches containing flexible TpA (thymine-p-adenine) steps, while no positive effect of TpA steps was observed for Lsr2 from Mycobacterium tuberculosis (20, 47). Our design-test-build approach, where different synthetic promoter variants were tested in vivo concerning CgpS-mediated silencing, however, revealed a certain degree of sequence specificity of CgpS toward a binding motif containing A/T steps. While the GC content was kept constant, alteration of the proposed motif in the P promoter significantly affected in vivo silencing (Fig. 2). A scenario that has been proposed for H-NS features a high affinity toward DNA (K of 60 nM), allowing the scanning of DNA until it reaches high-affinity sites, triggering the nucleation of the tight nucleoprotein complex required for silencing (3, 48, 49). Based on the systematic analysis of truncated promoter variants, we clearly defined the region required for efficient in vivo silencing by CgpS (see Fig. S6 in the supplemental material). It is important to note that we almost exclusively relied on in vivo approaches, including ChAP-seq analysis and reporter assays, to define the parameters affecting silencing and counter-silencing at the systems level. Although in vitro analysis of protein-DNA fragments (often linear DNA) is frequently applied and has provided valuable insights into the binding behavior of XS proteins (20, 47, 49), the conditions do not reflect physiologically relevant situations (DNA topology, protein-protein interaction, and interference); consequently, the results have to be interpreted with caution.

GntR-dependent counter-silencing is a disruptive mechanism.

While xenogeneic silencing neutralizes the potentially negative effect of invading foreign DNA, counter-silencing allows the integration into the host regulatory network and thereby provides access to horizontally acquired genes. This principle has been almost exclusively studied for H-NS in proteobacteria, and several types of TFs were shown to counteract H-NS silencing in vivo (15, 50–53). In the case of Lsr2, the mycobacterial iron-dependent regulator IdeR represents, to the best of our knowledge, the only example of an investigated Lsr2 counter-silencer. IdeR enables the iron-dependent activation of ferritin by alleviating Lsr2 repression at the bfrB locus (54).

During bacterial evolution, mutations leading to the formation of TF operator sequences within silenced promoters allow the controlled expression of the previously silenced genes by TF-mediated counter-silencing. In this work, the artificial insertion of the 15-bp short operator sequences of the gluconate-dependent TF GntR within different CgpS target promoters allowed us to study the potential and constraints of counter-silencing of this Lsr2-like XS protein. Binding of GntR to several CgpS target promoters led to transcription initiation of the silenced phage promoters, demonstrating that small changes in the DNA sequence added a further regulatory layer for expression control.

All tested CgpS target promoters showed significant reporter output in the absence of CgpS, confirming that they promote transcription and that CgpS inhibits promoter activity, presumably by hindering open complex formation or by trapping the open complex once it has formed (55). Previous studies already suggested that, without xenogeneic silencing, open complex formation is typically not the rate-limiting step at AT-rich promoters of horizontally acquired genes, meaning they are constitutively active (33, 56, 57). In the case of the H-NS target promoter pagC, an in vitro approach demonstrated RNA polymerase binding and open complex formation in the absence of additional factors, confirming that this promoter alone is transcriptionally competent (33).

In general, two different mechanisms of counter-silencing are conceivable. In one scenario (disruptive mechanism), the interference of TF and XS protein leads to a local disruption of the XS nucleoprotein complex, thereby enabling binding of the RNA polymerase to the DNA. Another possibility is that counter-silencing allows for supportive contacts between the RNA polymerase and the TF itself or more distal DNA regions (supportive mechanism). In this study, binding of GntR to the promoter constructs in the absence of the XS CgpS resulted in reduced reporter outputs, although counter-silencing was observed in the wild type. This result strongly speaks for a disruptive rather than a supportive GntR-mediated counter-silencing mechanism.

GntR binding close to the CgpS nucleation site is a prerequisite for efficient counter-silencing.

By the systematic analysis of promoter variants with varied positions of the GntR operator site, we clearly defined the window where binding of a specific TF led to counter-silencing. Counter-silencing of the Lsr2-like XS CgpS was most efficient at positions close to the position of maximal CgpS binding in a range of approximately 25 bp (Fig. 5 and Fig. S5), demonstrating that the position of GntR binding is critical for counter-silencing.

Previous studies by Will et al. revealed that counter-silencing and classical activation are different mechanisms of gene regulation (33). While TFs acting as activators typically bind to conserved promoter architectures and promote transcription either by changing the DNA conformation or by recruiting the RNA polymerase (58), the principle of counter-silencing allows a higher degree of flexibility in terms of promoter architecture (33). For the PhoPQ regulon of Salmonella enterica serovar Typhimurium, it was shown that PhoP activates core promoters featuring a precise operator position overlapping the −35 box. In contrast, horizontally acquired PhoP target genes show rather diverse promoter architectures, and here transcriptional activation is achieved by counter-silencing of H-NS. In these reported examples, the distances between the TSS and the closest PhoP binding site vary by only 34 bp (33, 59). This is in a range similar to that for our results and those obtained in previous studies for the H-NS target promoter P, where the insertion of TF operator sites counteracted H-NS silencing in a comparable window (34).

It is intriguing that in the context of xenogeneic silencing binding of a TF at positions close to the TSS leads to promoter activation, where it would cause a block of transcription at classical promoters (44, 60). Indeed, GntR binding in the absence of the XS CgpS resulted in reduced reporter outputs of several synthetic promoter constructs tested in this study (Fig. S1 and S4). However, counter-silencing was observed in the presence of CgpS in the wild type. These results demonstrate the potential of the counter-silencing principle to convert a repressor to an activating, tunable counter-silencer, thereby facilitating the integration of horizontally acquired DNA into host regulatory networks. Overall, these data illustrate how interference of TFs is shaping global regulatory networks and that the regulatory impact of TF binding is strongly affected by competition with other DNA-binding proteins.

MATERIALS AND METHODS

Bacterial strains and cultivation conditions.

All bacterial strains and plasmids used in this project are listed in Table S2A to C in the supplemental material. The strain C. glutamicum ATCC 13032 (61) was used as the wild-type strain. Detailed information about general growth conditions, microtiter cultivations used to monitor cell growth and fluorescence (62), and cultivations in the microfluidic chip device (46, 63) is available in the supplemental material (Text S1).

Recombinant DNA work.

All standard molecular methods, such as PCR, DNA restriction, and Gibson assembly, were performed according to previously described standard protocols (64, 65) or according to the manufacturer’s instructions. All plasmids were constructed by Gibson assembly. Details on plasmid construction are provided in Table S2C. DNA sequencing and synthesis of oligonucleotides used for amplification of DNA fragments (inserts for Gibson assembly [Table S2D], biotinylated DNA fragments for surface plasmon resonance [SPR] spectroscopy [Table S2E], and DNA fragments for electrophoretic mobility shift assays [EMSA] [Table S2F]) and sequencing (Table S2G), as well as synthesis of DNA sequences (Table S2H), were performed by Eurofins Genomics (Ebersberg, Germany). Chromosomal DNA of C. glutamicum ATCC 13032 was used as the PCR template and was prepared as described previously (66). Detailed information about construction of strain Δphage::P-cgpS via two-step homologous recombination (67) and the design of disruptive counter-silencing constructs is available in the supplemental material (Text S1).

Determination of TSS.

The determination of TSS and data analysis were performed with C. glutamicum wild-type cells by Vertis Biotechnology AG (Freising, Germany) using the Cappable-seq method developed by Ettwiller and Schildkraut (68). Prophage induction was triggered by adding mitomycin C. Detailed information about cultivation conditions, RNA preparation, and data analysis can be found in the supplemental material (Text S1). Relevant TSS were assigned to phage and nonphage CgpS target promoters when they were located in the promoter region 500 bp upstream of the start codon and directed in gene orientation (Table S1). Multiple TSS assigned to the same promoter were ranked depending on their enrichment scores.

Plots of CgpS coverage and GC profiles.

Normalized CgpS coverage values obtained from previous ChAP-seq analysis of Pfeifer and colleagues (4) and GC content of the reference C. glutamicum genome BX927147 (69) were plotted to the corresponding genome positions. Both parameters were calculated by a rolling mean with a window size of 50 bp and a step size of 10 bp using R (http://www.R-project.org) (70). The position of maximal CgpS coverage was centered, and the promoter orientation was normalized (start codon of the corresponding gene is located on the right site). Identified TSS positions were added. Ends of graphs are defined by the range of the CgpS binding peaks identified in previous ChAP-seq analysis (4). Plotting was performed either by R (70) or by GraphPad Prism 7.00 (GraphPad Software, La Jolla, CA).

Analyses of AT-rich stretches in CgpS binding regions.

The C. glutamicum genome (BX927147 [69]) was scanned for AT-rich stretches using a custom python script (submitted to GitHub at https://github.com/afilipch/afp/blob/master/genomic/get_at_stretches.py). For further details, see the supplemental material (Text S1).

Protein purification, SPR spectroscopy, and EMSA.

Information about purification of Strep-tagged CgpS and His-tagged GntR and the performed in vitro binding assays (SPR spectroscopy and EMSA) can be found in the supplemental material (Text S1).

Data availability.

The custom python script used for scanning for AT-rich stretches is available in the GitHub repository (https://github.com/afilipch/afp/blob/master/genomic/get_at_stretches.py). Data from previously reported ChAP-seq analysis (4) are available at the GEO database (https://www.ncbi.nlm.nih.gov/geo) under accession number GSE141132.