Using CRISPR/dCas proteins to block foreign DNA silencers by turning them into their opponents

In this work, we present CRISPRcosi, a CRISPR/dCas-based counter-silencing system. In contrast to previously established TF-based counter-silencing approaches, CRISPRcosi is independent of TF binding sites (Fig. 1). Here, single-guide RNAs (sgRNAs) were used to direct "dead" Cas9 (dCas9) (37) to target promoters of the Lsr2-like protein CgpS. These silenced promoters typically feature a distinct drop in GC content and an inversely correlated CgpS binding peak covering several hundred nucleotides (4) (Fig. 1A). Comparable to the TF-based approach, binding of the sgRNA/dCas9 complex close to the CgpS nucleation site might interfere with the silencer-DNA complex, leading to counter-silencing and transcription of the associated gene(s). The integrity of the silencer-DNA complex in response to such interference effects is usually not completely disrupted, meaning that the promoters are only partially active (33). To monitor promoter activities, prophage promoters were fused to the reporter genes venus or eyfp (Fig. 1B). The sgRNA-DNA complementarity, mediated by the sgRNA spacer sequence, in combination with an NGG protospacer adjacent motif (PAM) sequence, served as target recognition features to guide dCas9 to specific DNA positions and strands (43). sgRNAs were designed to bind the template or the coding strand of the DNA at different positions within the CgpS target promoters (Fig. 1C). As proof of concept, microscopy studies with C. glutamicum cells carrying a reporter system consisting of the CgpS target promoter P_cg1974 from the CGP3 prophage region fused to the eyfp reporter gene demonstrated the potential of the CRISPRcosi approach. A strong yellow fluorescence signal only present in cells co-expressing dcas9 and the specific sgRNA-CS3 confirmed the functionality of this inducing approach and the strict dependency on both components (Fig. 1D).

After demonstrating that the mechanism of CRISPRcosi allowed the targeted upregulation of CgpS silenced genes, we set out to systematically assess the opportunities and constraints of this approach. To prevent the induction of any prophage genes, we used the prophage-free C. glutamicum Δphage::P_cgpS-cgpS (33) derivative with a reintegrated copy of the cgpS gene as our platform strain. All plasmid IDs (#number) given in the figures and text are referenced in Table S2. The sgRNAs were designed to bind to different DNA strands and positions directly adjacent to NGG PAM sequences in the P_cg1974 promoter sequence. While the homologous region on the coding strand of sgRNA-CS1 overlapped with the putative CgpS binding motif as well as with the dominant transcriptional start site (TSS) (33), sgRNA-CS3 bound the template strand in the region of maximal CgpS binding coverage seven nucleotides downstream of the CgpS nucleation site (Fig. 2A). The backscatter-normalized specific Venus fluorescence signals (specific fluorescence = total Venus fluorescence / backscatter) measured after 20 h of cultivation reflected the promoter activities of strains with different sgRNA-encoding sequences and the presence and absence of the isopropyl β-D-1-thiogalactopyranoside (IPTG)-inducible dcas9 gene. Here, only the co-expression of dcas9 and sgRNA-CS3, but not sgRNA-CS1, resulted in an IPTG-correlated stepwise increase in promoter activity, demonstrating success with CRISPRcosi in a dCas9 dose-dependent manner (Fig. 2B; Fig. S1).

The combination of sgRNAs and dCas enzymes is commonly used to inhibit gene expression in CRISPRi approaches (38). Depending on the position, binding of the sgRNA/dCas9 complex can sterically prevent transcriptional initiation or elongation, leading to repression (38, 41, 48). The presented CRISPRcosi approach uses the same components, suggesting that binding of the sgRNA/dCas9 complex at specific positions might negatively affect P_cg1974 promoter activity. Prophage promoters like P_cg1974 are typically constitutively and strongly active in the absence of the xenogeneic silencer CgpS (33). To monitor putative CRISPRi effects, P_cg1974-driven reporter outputs of prophage-free, CgpS-lacking C. glutamicum Δphage cells (47) were tested in the context of different sgRNA/dCas9 combinations. Even in the absence of IPTG, significantly reduced promoter activities were observed in cells expressing sgRNA-CS1, likely caused by leaky P_tac-derived dcas9 expression. In contrast, sgRNA-CS3 showed only minor CRISPRi effects (Fig. 2C). These findings aligned with results from CRISPRi studies, which demonstrated that CRISPRi performance is determined by the sgRNA/dCas9 target position, where effects ranged from neutral effects to highest effects when binding in close proximity to the TSS (38, 41, 49–52). sgRNA-CS3 binds the template DNA strand downstream of the dominant TSS, a position that is generally less sensitive to dCas9-mediated CRISPRi effects (38, 41). In conclusion, the balance between CRISPRi and counter-silencing effects determined the overall activity of the P_cg1974 promoter in the presence of CgpS. With respect to CRISPRcosi, the counter-silencing effects had to be stronger than any CRISPRi effects, resulting in increased promoter activities.

Having demonstrated that binding of the CgpS nucleation site-specific sgRNA-CS3/dCas9 complex counteracted CgpS silencing at the prophage promoter P_cg1974, we next set out to determine the relative effect of CRISPRi compared with CRISPRcosi. Therefore, we applied a double plasmid system (pEC- and pJC1-derivatives) to compare P_cg1974-venus reporter outputs of cells differing in their set of sgRNAs and dcas9 availability. Promoter activities revealed that two copies of sgRNA-CS3 did not result in a synergistic effect, suggesting that the amount of sgRNA-CS3 derived from a single copy did not limit counter-silencing efficiencies (Fig. 3A). Second, the inducing effect of the sgRNA-CS3/dCas9 complex was completely abolished in the presence of sgRNA-CS1, as shown by the very low P_cg1974-driven reporter signal (Fig. 3A). These results indicate that the sgRNA-CS1/dCas9-mediated CRISPRi effect was dominant over the sgRNA-CS3/dCas9-dependent CRISPRcosi effect. The PAM dependency of CRISPRcosi was confirmed by the lack of responsiveness of the modified P_{cg1974 ΔPAM} promoter, which lacked the sgRNA-CS3-specific PAM sequence (Fig. 3).

The above-mentioned results indicated that the CRISPRcosi mechanism was most effective when the sgRNA/dCas9 complex did not interfere with transcription but bound to DNA regions that are effective in counteracting CgpS silencing (e.g., CgpS nucleation site). To improve our understanding of the rules underlying this mechanism, we systematically screened sgRNAs that guide dCas9 to different positions of the CgpS target promoters P_cg1974 and P_cg2014 (Fig. 4). We found that activities of both promoters were increased in the presence of sgRNAs directing dCas9 to the DNA sequences with the lowest GC content. In addition to sgRNA-CS3, sgRNA-33 counteracted CgpS silencing at the prophage promoter P_cg1974 (Fig. 4C). Interestingly, both sgRNAs were designed to bind the DNA template strand with a shift of only one nucleotide and overlapped with the annotated position of maximal CgpS binding (4). No other sgRNA induced prophage promoter activity, even when partly overlapping in binding position or bound in proximity.

In the case of P_cg2014, only sgRNA-22 led to CRISPRcosi (Fig. 4D). sgRNA-22 was designed to bind the template DNA strand in the area of lowest GC content at the postulated CgpS nucleation sites (33) upstream of the position of maximal CgpS coverage (4) (Fig. 4B and D). Dose-response studies using increments of IPTG to induce dcas9 expression demonstrated a stepwise increase in P_cg2014 promoter activity (Fig. S1B). In summary, the screening results clearly demonstrated that CRISPRcosi is context dependent. We hypothesize that multiple sgRNA binding position-related factors hampered the induction of the CgpS target promoters, including counteracting CRISPRi effects, the cooperative binding behavior of the silencer-DNA complex, and the low frequency of the guanine-rich, NGG (d)Cas9 PAM sequence within AT-rich promoters. These constraints highlighted the need to increase the flexibility of this approach, for example, by implementing alternative dCas enzymes like dCas12a that recognize AT-rich PAM sequences.

dCas12a (dFnCpf1) (39, 40) represents a promising candidate to improve guide RNA selection within promoter regions, which are generally more AT-rich than non-promoter regions. In contrast to the downstream-located, guanine-rich NGG (d)Cas9 PAM sequence, (d)Cas12a recognizes a (T)TTN PAM sequence upstream of the target (53–55) (Fig. 5). We systematically assessed the effect of dCas12a-specific guide RNAs (CRISPR RNAs [crRNAs]) in the context of CRISPR/dCas12a-mediated CgpS counter-silencing. Comparison of nine crRNAs across a broad range of binding positions in P_cg1974 revealed two effective crRNAs, both designed to bind the coding strand at the annotated TSSs (33) (Fig. S2A). While crRNA-9 had lower promoter induction capacities, crRNA-7-mediated guiding of dCas12a to the dominant TSS resulted in strong counter-silencing effects (Fig. 5; Fig. S2A). The other seven crRNAs were unable to counteract CgpS silencing, highlighting that only specific promoter regions were susceptible to CRISPR/dCas12a-mediated counter-silencing (Fig. S2A), as observed for dCas9.

By using the most effective P_cg1974-specific guide RNA candidates for each system (sgRNA-CS3 and crRNA-7), we directly compared the dCas9- and dCas12a-mediated CRISPRcosi utility (Fig. 5; Fig. S3A). Here, the crRNA-7/dCas12a approach showed slightly higher activities in comparison to the dCas9/sgRNA-CS3 system (Fig. 5). In addition, the lower toxicity of dCas12a (54, 56) improved plasmid stability and allowed for the constitutive expression of the dCas enzyme in C. glutamicum, yielding strong and homogeneous counter-silencing effects. In contrast, dcas9-expressing cells often exhibited heterogeneous signal distribution, which was likely a consequence of the noisy, IPTG-inducible LacI-P_tac expression system used to initiate dcas9 transcription (Fig. S3B).

To evaluate the applicability of the CRISPRcosi mechanism as a tool to modulate gene expression of prophage genes, we tested our system with a set of 11 CgpS targets and both dCas enzymes. We analyzed promoter activities in prophage-free C. glutamicum cells (Δphage::P_cgpS-cgpS [33]) co-expressing cgpS, dcas9, or dcas12a, and either a suitable guide RNA or no sg/crRNA (Fig. S4). Ratios greater than 1, for backscatter-normalized Venus reporter outputs of cells expressing a guide RNA and strains without any sg/crRNA, demonstrated counter-silencing. Values below 1 were indicative of CRISPRi effects (Fig. 6). In summary, our dCas9 approach successfully yielded significantly increased (by >2-fold) activity for the prophage promoters P_cg1974, P_cg2014, and P_cg2032. With a 25-fold increase in promoter activity, cg2014_crRNA-2 was the top-performing candidate for CRISPRcosi efficiencies (Fig. 5; Fig. S2).

Comparable to the sgRNA-22/dCas9 targeting cg2014, crRNA-2 facilitated counter-silencing by directing dCas12a to the template DNA strand at the putative CgpS nucleation site upstream of the P_cg2014 core promoter region (Fig. 6; Fig. S2B). Unexpectedly, crRNA-3 targeted the coding strand and failed to induce counter-silencing of P_cg2014, although its binding region perfectly matched that of cg2014_crRNA-2 (Fig. S2B). This demonstrated that the target DNA strand was relevant for dCas12a-based CRISPRcosi efficiency. Previous work has suggested that dCas9 binding to the coding strand leads to stronger CRISPRi effects (38, 41), while dCas12a-based gene repression is more efficient when targeting the template strand (39). Therefore, our results suggested that CRISPRcosi might be favored when CRISPRi effects were disfavored. This hypothesis was also supported by the point that all three counter-silencing inducing sgRNAs guide dCas9 to positions outside the core promoter (Fig. 6), thereby avoiding the region generally considered to be most sensitive for CRISPRi (38, 41). In conclusion, we observed a trend that binding outside of the core promoter region and/or to the less efficient DNA strand with respect to CRISPRi increased the likelihood of achieving CRISPRcosi. The implementation of additional dCas enzymes expanded the positional flexibility and facilitated the avoidance of counteracting CRISPRi effects.

To assess potential off-target effects of the CRISPRcosi approach, genome-wide transcriptome analysis of our CRISPRcosi approach was performed with C. glutamicum wild-type cells co-expressing dcas9 and either P_cg1974-specific sgRNA-CS3 or sgRNA-CS1 (Fig. 7; Fig. S5; Table S5). While sgRNA-CS3 facilitated CRISPRcosi, sgRNA-CS1 served as a negative control in this context. (Fig. 2). In comparison to the sgRNA-CS1 reference group, 14 genes were significantly upregulated (false discovery rate [FDR]-P-values < 0.05; log₂ fold change > 1) in the sgRNA-CS3 samples (Fig. 7; Fig. S5), while one gene was downregulated (FDR-P-values < 0.05; log₂ fold change < 1) (Fig. 7A and B; Fig. S5A). By using the Cas-OFFinder online tool (http://www.rgenome.net/cas-offinder/↗) (57), no sgRNA-CS3 or -CS1 off-target was associated with these genes (Table S6).

We found the differentially expressed genes contained only two known CgpS targets (4): cg1974 (NCgl1682), the targeted gene, and cg1976 (NCgl1684) (highlighted by * in Fig. 7B). Together with cg1975 (NCgl1683), these three neighboring loci were by far the most upregulated genes in cells expressing sgRNA-CS3 (Fig. 7A and B; Fig. S5). The highest upregulation of more than 44-fold (log₂ fold change 5.5) was observed for the targeted prophage gene cg1974 (NCgl1682). The divergently oriented gene cg1975 (NCgl1683) increased expression by more than 20-fold and represented the second-most upregulated gene (FDR-P-values < 0.05) (Fig. 7A through C; Fig. S5A). This increased P_cg1975-derived transcription was likely a direct effect resulting from dCas9-mediated interference with the CgpS nucleoprotein complex in the divergently oriented P_cg1974/P_cg1975 promoter region.

Previous analyses from Pfeifer and colleagues revealed that overexpression of a truncated CgpS variant has a dominant-negative effect on the silencing mechanism and results in 194 significantly differentially expressed genes (4). Expression of a truncated CgpS variant also led to prophage excision and circularization of the CGP3 element (4). In contrast to this relatively broad response, our CRISPR/dCas9-based counter-silencing approach seemed to act more specifically, resulting in only 15 significantly affected genes (Fig. 7A; Fig. S5A) and no increased copy number of the circular CGP3 (Fig. 7D). In conclusion, our results demonstrated that CgpS target genes located within the CGP3 prophage element could be specifically induced by counter-silencing.

The third most prominent significantly upregulated candidate in the transcriptome analysis was the known CgpS target cg1976, which is located downstream of cg1975 (Fig. 7A through C; Fig. S5A). Recently, it was demonstrated that increased transcription rates in the neighborhood of silenced genes might counteract the xenogeneic silencing mechanism (58–60). Based on these findings, we hypothesized that an increased transcription rate of gene cg1975 favored the expression of the downstream-located cg1976 gene. To investigate the effect of high transcription rates upstream of a CgpS silenced promoter, we inserted short constitutive promoters upstream of the P_cg1974-venus reporter sequence in tandem or divergent orientation (Fig. 8). Interestingly, all three tandemly oriented promoters efficiently counteracted CgpS silencing at P_cg1974. According to their promoter strength in Escherichia coli (61), the highest Venus reporter outputs were obtained with the tandemly oriented P_J23119 promoter, followed by P_J23150 (both from the Anderson Promoter Collection [https://parts.igem.org/Promoters/Catalog/Constitutive↗]) and P_Lac (61), demonstrating transcription rate-dependent effects on the silenced prophage promoter. Surprisingly, even divergently oriented promoters affected the P_cg1974 promoter activity, although the effects were reduced compared to the tandem constructs (Fig. 8). The observed bidirectional promoter activity of P_J23119 highlighted that divergent promoter activities should be considered when working with these promoter systems, including putative effects on DNA regions targeted by xenogeneic silencers.

In conclusion, we showed that high transcription rates upstream of a prophage promoter could lead to counter-silencing of CgpS. These findings are in agreement with the previous observations for H-NS and Lsr2 (58–60). Apparently, interference with the formation of the XS nucleoprotein complex can be achieved either via direct competition of DNA-binding proteins at the XS nucleation site or by actively transcribing RNA polymerase initiated from upstream promoters.

Having demonstrated CRISPRcosi effects on CgpS silencing in C. glutamicum, we sought to test whether these observations could be extended to other bacteria, for example, to the pharmacologically interesting streptomycetes. In S. venezuelae, Lsr2 (CgpS-equivalent) plays an important role in repressing a wide array of specialized metabolic clusters that direct natural product synthesis (31), including the chloramphenicol biosynthetic cluster (60). We compared the effects of guiding dCas9 to a key promoter region (CRISPRi) with the effects of targeting the Lsr2 nucleation site within the downstream coding sequence (CRISPRcosi) (Fig. 9A). As expected, chloramphenicol production levels dropped when the promoter region was targeted by dCas9 (Fig. 9B and C), as had been expected from the CRISPRi observations in C. glutamicum. In contrast, targeting of the Lsr2 nucleation site within the downstream coding sequence led to a significant increase (approximately three times) in chloramphenicol production, relative to what had been observed for the no sgRNA wild-type control (Fig. 9B and C). It has to be noted that the addition of thiostrepton alone induced chloramphenicol production by a S. venezuelae wild-type strain carrying a dcas9-free plasmid (that integrated into the same chromosomal locus) in a dose-dependent manner (Fig. S6D). However, under the induced conditions (50 µg/mL thiostrepton), we observed higher chloramphenicol production in the presence of the guide RNA targeting the Lsr2 nucleation site.

In an effort to demonstrate that the disruption of Lsr2 repression by CRISPRcosi was responsible for the increased chloramphenicol production, we attempted the same experiment in an lsr2 mutant background, using both promoter- and Lsr2 nucleation site-targeting sgRNAs. The growth of these strains was, however, profoundly compromised following thiostrepton addition, and consequently, it was not possible to monitor chloramphenicol production or compare with wild-type.

Altogether, these findings suggested dCas9 could mediate Lsr2 counter-silencing within the chloramphenicol biosynthesis gene cluster of S. venezuelae, thereby demonstrating the broad applicability of this CRISPRcosi approach across Actinobacteria.

All bacterial strains used in this study are listed in supplementary information Table S1. Escherichia coli DH5α cells were used for plasmid amplification and storage. C. glutamicum ATCC 13032 (66) and S. venezuelae NRRL B-65442 (67), as well as their derivatives served as platform strains for the characterization of the CRISPRcosi systems. Detailed information about cultivation conditions is given in Text S1.

All used plasmids are listed in Table S2. Each plasmid was assigned a unique identifier (#letter/number) that remains consistent throughout the study to facilitate referencing. Standard molecular methods were performed according to the manufacturer's instructions or as previously described (68–70). Plasmid sequencing and synthesis of DNA oligonucleotides were performed by Eurofins Genomics (Ebersberg, Germany), Plasmidsaurus using Oxford Nanopore Technology, or the Farncombe Sequencing Facility (McMaster University) and Integrated DNA Technologies (Table S3). Plasmid transfer into C. glutamicum was carried out by electroporation (71). Detailed information about the design of all plasmids constructed in this study is available in the supplementary material (Table S2). The stepwise constructions of S. venezuelae plasmids are described in Text S1 in the supplemental material ("Recombinant DNA work"). Briefly, all of them contained a thiostrepton-inducible promoter in front of the dcas9 gene, the tipA gene (whose product is required for thiostrepton induction), a translationally acting theophylline riboswitch (72) upstream of tipA, the phiC31 attP-int locus (73), and, in the case of pMC346 (#84) and pMC347 (#85), a specific spacer-encoding sequence upstream of the guide RNA scaffold. Sequence-validated plasmids were passaged through the non-methylating E. coli strain ET12567 (74) and introduced into S. venezuelae via conjugation (75).

The CRISPRcosi system was systematically characterized by performing reporter-based promoter activity studies with C. glutamicum Δphage::P_cgpS-cgpS (33) cells producing the XS protein CgpS or with C. glutamicum Δphage (47) control cells lacking CgpS (Table S1). The key components of this approach were a prophage promoter that is silenced by CgpS, a dCas protein, and a specific guide RNA that targeted the dCas protein to a specific DNA position within the silenced promoter region. Therefore, cells harbored different pJC1 (pCG1 origin of replication) and/or pEC-XC99E derivatives (pGA1 origin of replication) containing the respective DNA sequences (Table S2).

dCas-encoding genes under control of IPTG-inducible (lacI-P_tac-dcas9) or constitutive expression systems (P_lacM-dcas12a) were inserted into the respective plasmids. To express guide RNAs (sgRNA, crRNA), the constitutive promoters P_tacΔ (76) or P_J23119 (from Anderson Promoter Collection, https://parts.igem.org/Promoters/Catalog/Constitutive↗) were fused to the corresponding guide RNA-encoding sequences and inserted into either the pEC-XC99E- or pJC1-based plasmid. A detailed description of the guide RNA design, which followed recently published guidelines (45, 54), is provided in Text S1 in the supplemental material ("Guide RNA design"). All sgRNA and crRNA spacer sequences are listed in Table S4. To monitor promoter activities via a reporter system (yellow fluorescence), prophage promoters known to be silenced by CgpS were fused to the reporter gene venus via a linker containing a stop codon and an optimal C. glutamicum ribosomal binding site (77).

We systematically analyzed the CRISPRcosi system by performing fluorescence microscopy or by monitoring growth and Venus fluorescence of respective C. glutamicum strains during cultivation in the BioLector I microtiter system (Beckman Coulter, Brea, CA, USA) (78). A detailed description of both reporter-based methods is provided in Text S1 in sections "Fluorescence microscopy" and "Microtiter cultivation to monitor growth and fluorescence in reporter assays." Briefly, cells were cultivated in microtiter plates in CGXII minimal medium (79) supplemented with 111 mM glucose, respective antibiotics (25 µg/mL kanamycin and/or 10 µg/mL chloramphenicol) and up to 200 µM IPTG for P_tac-derived dcas9 expression as required.

In addition to the plasmid-based approaches to monitor promoter activities with reporter genes, we set out to evaluate the effectiveness and specificity of CRISPRcosi on silenced genes in their native context. Therefore, RNA-seq analysis was performed with C. glutamicum wild-type cells co-expressing dcas9 and sgRNA-CS3 (#2) or sgRNA-CS1 (#5). A detailed description is given in Text S1 of the supplemental material in section "Transcriptome analysis". Briefly, pre-treatment and sequencing of total RNA from mid-exponential phase cells (7 h) was performed by Genewiz (Leipzig, Germany) using the Illumina NovaSeq system. The C. glutamicum ATCC 13032 genome (NCBI reference sequence NC_003450.3↗) combined with the respective sequences of plasmids pEC-XC99E-lacI-P_tac-dcas9--P_tacΔ-sgRNA-CS3 (#2) or pEC-XC99E-lacI-P_tac-dcas9--P_tacΔ-sgRNA-CS1 (#5) served as reference. Please note that "cg" and "NCgl" loci numbers both refer to the strain C. glutamicum ATCC 13032 (cg: prophage-free strain, NCBI reference sequence NC_022040.1↗; NCgl: wild-type, NCBI reference sequence NC_003450.3↗). The software CLC Genomics Workbench v.24 (Qiagen, Hilden, Germany) was used for data analysis and for the identification of significantly differentially expressed genes (FDR-P-values <0.05 and |log₂ fold change values| > 1) in clones expressing P_cg1974_sgRNA-CS3 (#2) compared to the P_cg1974_sgRNA-CS1 reference cells (#5) (full data set can be found in Table S5).

To identify putative off-targets of sgRNA-CS1 and sgRNA-CS3, the Cas-OFFinder online tool (http://www.rgenome.net/cas-offinder/↗) from Bae and colleagues (57) was used (see Text S1 for further information). A full overview of all genome-wide off-targets is given in Table S6.

To identify counter-silencing-associated effects on prophage induction, qPCR was performed to measure the relative amount of circular CGP3 prophage DNA in cells co-expressing dcas9 and sgRNA-CS3 (#2) or -CS1 (#5). A detailed summary of the qPCR analysis can be found in the supplemental material Text S1[, "Quantitative PCR (qPCR)"]. Briefly, oligonucleotides B040/B041 were used to detect the circularized CGP3 prophage DNA element, while oligonucleotides B042/B043 amplified a fragment from the ddh (NCgl2528) reference gene (Table S3C). For data evaluation, the C_T value-based 2^−ΔΔCT method from Livak and Schmittgen (80) was applied. Here, 2^-ΔΔCT values and the ranges of circular CGP3 levels were calculated based on the geometric C_T means of biological triplicates and technical duplicates.

To assess the potential of CRISPRcosi in organisms beyond C. glutamicum, we tested our system in S. venezuelae, focusing on the ability to upregulate the chloramphenicol biosynthesis gene cluster, which is normally repressed by the xenogeneic silencer Lsr2 (60). Detailed information about the cultivation conditions, metabolite extraction, and analytical steps to quantify the chloramphenicol levels are available in section "Analysis of Streptomyces metabolite extracts" of Text S1 in the supplemental material.

The software GraphPad Prism 8.00 (GraphPad Software, La Jolla, CA, USA) was used to calculate areas under curves, and significance levels (two-tailed, unpaired t-tests) of backscatter-normalized reporter outputs (specific fluorescence) based on biological triplicates (n = 3). The software CLC Genomics Workbench v.24 (Qiagen, Hilden, Germany) was used for statistical analysis of the RNA-seq data (see "Transcriptome analysis"). Statistical analysis of the qPCR data is described in "Quantitative PCR (qPCR)" in Text S1.