The WblC/WhiB7 Transcription Factor Controls Intrinsic Resistance to Translation-Targeting Antibiotics by Altering Ribosome Composition.

The emergence of antibiotic-resistant bacteria is one of the top threats in human health. Therefore, we need to understand how bacteria acquire resistance to antibiotics and continue growth even in the presence of antibiotics. Streptomyces coelicolor, an antibiotic-producing soil bacterium, intrinsically develops resistance to translation-targeting antibiotics. Intrinsic resistance is controlled by the WblC/WhiB7 transcription factor that is highly conserved within Actinobacteria, including Mycobacterium tuberculosis. Here, identification of the WblC/WhiB7 regulon revealed that WblC/WhiB7 controls ribosome maintenance genes and promotes translation in the presence of antibiotics by altering the composition of ribosome-associated proteins. Also, the WblC-mediated ribosomal alteration is indeed required for resistance to translation-targeting antibiotics. This suggests that inactivation of the WblC/WhiB7 regulon could be a potential target to treat antibiotic-resistant mycobacteria.

analysis of the samples revealed that large differences were observed between the transcriptome profiles of the tetracycline-treated and untreated wild-type strain, whereas the profiles of untreated samples from the wild-type and ΔwblC strains were clustered together (Fig. S1E). When treated with 2 g/ml tetracycline (0.2ϫ MIC) for 30 min, a total of 2,908 genes (1,417 genes upregulated and 1,491 genes downregulated) were affected by tetracycline treatment in the wild type (Ͼ2-fold, adjusted P value Ͻ0.01) (GEO accession number GSE136168). Among these genes, transcript levels of 614 genes out of 7,853 open reading frames (ORFs) (excluding pseudogenes) were significantly different (adjusted P value of Ͻ0.01) by Ͼ2-fold between the wild-type and ΔwblC strains ( Fig. 1A and B). Because WblC affects about 7.8% of the genome, it indicates that WblC extensively alters the S. coelicolor transcriptome when cells are treated with tetracycline. Among 614 genes, WblC upregulated 412 genes and downregulated 202 genes in the presence of tetracycline (Fig. 1A). In contrast, only 67 genes were significantly different (adjusted P value of Ͻ0.01) by Ͼ2-fold when we compared untreated samples of wild-type and ΔwblC strains (Fig. S1F). Moreover, the fold changes of the untreated samples between the wild-type and ΔwblC strains were distributed within a smaller range (from a 7-fold increase to 9-fold decrease) compared to those of tetracycline-treated samples, which ranged from a 3,000-fold increase to 13-fold decrease ( Fig. 1B and Fig. S1F). Consistent with a previous finding that WblC proteins are not detected by immunoblotting unless WblC production is induced by translationinhibiting antibiotic treatment (19), these data suggest that steady-state WblC levels in unstressed conditions exhibit only a marginal effect on the S. coelicolor transcriptome.
A large number of genes are directly controlled by WblC. The 614 differentially expressed genes (DEGs) identified by RNA sequencing were likely to encompass both direct and indirect WblC targets. To find direct WblC targets, we performed ChIP-seq in WblC/WhiB7 Regulon in Streptomyces ® wild-type cells treated with tetracycline and identified WblC binding sites across the genome. We defined 830 peaks as WblC binding sites that were enriched more than twofold compared to control input DNA (Fig. 1A). We combined ChIP-seq data with RNA-seq data to select direct WblC targets. We first examined the sequence from Ϫ500 to ϩ200 bp relative to the start codon of 614 DEGs identified from RNA-seq data to assess whether the selected DEGs contained WblC binding peaks. Then, the selected DEGs were further curated to remove false-positive peaks, including instances where the binding peak belonged to a divergently transcribed gene(s). Using this procedure, we identified 206 promoters containing WblC binding sites (Fig. 1C) and assigned 312 DEGs to the WblC regulon ( Fig. 1A; see also Table S2 in the supplemental material). Of the WblC binding peaks, 74.3% were located from Ϫ100 to ϩ50 bp relative to the start codon of the first gene in each operon (Fig. 1D). Among 312 genes, the RNA levels of 288 genes were higher in the tetracycline-treated samples from the wild-type strain than those from the ΔwblC strain (represented in Fig. 2A to J), indicating that WblC functions as a transcriptional activator of those genes. Interestingly, RNA levels of 24 genes were lower in the tetracycline-treated samples from the wild-type strain than those from the ΔwblC strain (represented in Fig. 2K and L), suggesting that WblC could also act as a transcriptional repressor of those genes. WblC binding sites of repressed genes were mostly located upstream of the translation start codon, lying between Ϫ500 and 0 bp relative to the translation start codon (Fig. S1G). WblC binding of the selected promoters was further confirmed by ChIP-qPCR (Fig. S1H).
WblC regulon products could antagonize antibiotic stress by reducing the concentration of effective antibiotics, enhancing translation, resolving oxidative stress, and adjusting metabolic pathways. We then tried functional annotations of the WblC regulon using Gene Ontology (GO) terms, InterPro entries, and EggNOGs. Analyses of the WblC-activated genes revealed a significant enrichment in several functions, including tRNA aminoacylation, translation, acyl-coenzyme A (acyl-CoA)- genes encoding a tetracycline resistance protein, a peptidyl-tRNA hydrolase, an elongation factor Tu3, an aminoglycoside resistance protein, a ribosomeassociated GTPase, a PhoH-like protein, a transcription regulator involved in leucine biosynthesis, a redox-sensing extracytoplasmic function (ECF) sigma factor, an ATP-dependent helicase, and an RNA helicase, respectively. (K) wblE and (L) citA are WblC-repressed genes encoding a WhiB-paralogous protein and CitA citrate synthase, respectively. Visualized sequence reads of the WblC regulon genes and surrounding regions (Ϯ 600 bp) are presented. The heights of RNA sequencing reads of each panel are normalized to the peak showing the highest expression within the panel. The highest values of fold change in RNA-seq and fold enrichment in ChIP-sequencing are shown in each panel. Dark purple arrows indicate WblC regulon genes, and pale purple arrows indicate their neighboring genes.
Lee et al. ® dependent N-acyltransferase, nucleoside triphosphate hydrolase (NTPase) with nucleotide binding, and ABC transporter (Fig. 3). After performing hierarchical clustering, we further selected 179 WblC-activated genes and classified them into eight major categories with some subcategories ( Table 1). The remaining 109 genes were uncategorized mostly due to insufficient annotations.
On the basis of the functional categorizations, we reasoned how WblC-activated genes are involved in resistance to translation-targeting antibiotics. ABC transporter proteins, major facilitator superfamily proteins such as CmlR2 (20) or Pep (21), and other membrane proteins could decrease the intracellular antibiotic concentration by exporting antibiotics. Eis and Eis2 aminoglycoside acetyltransferases (22), virginiamycin B lyase (Vgb) (23), and possibly other transferases could modify antibiotics to inactivate their functions. Interestingly, WblC also activates genes encoding aminoacyl-tRNA synthetases, tRNA processing enzymes, several amino acid biosynthetic enzymes, and
WblC/WhiB7 Regulon in Streptomyces ® tRNAs (Table 1 and Table S2). It can be hypothesized that these enzymes supply the needed aminoacyl-tRNAs even in the tetracycline-treated conditions to maintain translation. Moreover, WblC increases mRNAs of genes encoding Erm(O)-type rRNA methyltransferases (encoded by SCO6089) homologous to Streptomyces lividans Lrm (24), TetM ribosome-associated GTPase (25), several Gcn5-related N-acetyltransferases, and other proteins like SmpB (26) ( Table 1). The functions of those genes are predicted to modify, modulate, or protect translational machinery against antibiotics, supporting the idea that WblC is involved in resuming translation during translational stress mediated by antibiotics. Additionally, it suggests that 19 transcriptional regulators, including WblC itself (11), SigR redox-and antibiotic-responsive sigma factor (27), and NdgR involved in leucine biosynthesis (28), play roles in resolving oxidative stress and adjusting metabolic pathways generated by translation-targeting antibiotics. Cumulatively, these findings imply that WblC-activated genes are involved in a protective response to translation-interrupting antibiotics to maintain viability, which could be an intrinsic resistance mechanism to translation-targeting antibiotics. In contrast to WblCactivated genes, WblC-repressed genes were not significantly enriched in any of the GO terms, InterPro entries, or EggNOGs classifications. WblC controls 22 noncoding RNAs. We noticed that 22 noncoding RNAs, including six tRNA genes are directly regulated by WblC (Table S3). Four small noncoding RNAs are located independently in the intergenic regions ( Fig. S2A and S2B and Table S3), suggesting that they could function as trans-acting riboregulators. Twelve noncoding RNAs are mapped to the antisense strand of the neighboring gene(s), which could be designated as cis-acting antisense RNAs that could downregulate expression of the overlapping genes ( Fig. S2C and S2D and Table S3). Interestingly, even among the identified WblC regulon, 104 mRNAs have 5= or 3= long untranslated regions (UTRs) that overlap a gene(s) in the opposite strand and thus could also function as antisense RNAs of the neighboring gene(s) (29, 30) ( Fig. S2E and S2F and Table S3). In conclusion, these results indicate that WblC not only controls mRNA transcriptome but also controls expression of many noncoding RNAs and UTRs, possibly adding another layer of gene regulation during antibiotic stress.
Conserved sequence elements in the WblC-activated promoters suggest a mechanism of promoter recognition by WblC-HrdB interaction. In mycobacteria, WhiB7 binds to a conserved AT-rich sequence adjacent to Ϫ35 promoter elements (16), and WhiB7's binding to the AT-rich motif enables a cooperative binding between WhiB7, SigA sigma factor, and the target promoters (13). Similarly, we examined the promoter regions of WblC targets in S. coelicolor to look for a conserved motif. MEME analysis of 189 WblC-activated promoters revealed that WblC motifs are located 3 bp upstream of the Ϫ35 element, and the spacing between the Ϫ35 and Ϫ10 elements is between 16 and 19 bp (Fig. 4A). The optimal spacer appears to be 17 or 18 bp because WblC promoters with the 17-or 18-bp spacer are most abundant and showed a higher fold change in RNA-seq than those with the 16-or 19-bp spacer (Fig. 4D). The Ϫ10 elements of WblC-activated promoters are more conserved, similar to the genome-wide consensus (TANNNT) (31), whereas the Ϫ35 elements are less conserved. Most of the transcription start sites (TSSs) are located 7 or 8 bp downstream of the Ϫ10 elements, which is consistent with previously reported transcriptional profiling in S. coelicolor (31) (Fig. S3).
There appears to be a clear correlation between the WblC-activated promoter sequences and mRNA fold changes. When we compared the promoter sequences of genes showing the top 20% and bottom 20% fold changes in RNA-seq, we observed that the top 20% of the WblC-activated promoter sequences exhibited a higher enrichment of A nucleotides in the WblC binding sites than those in the bottom 20% ( Fig. 4B). High expression levels are also likely to be linked to other features, including appearance of pyrimidines following stretches of A nucleotides within the WblC binding sites, a higher enrichment of G at the third position of the Ϫ35 element, and a lower enrichment of T at the first position of the Ϫ10 element (Fig. 4B). A similar In contrast, there is no distinctive feature in the WblC-repressed promoters, suggesting that the binding mode of the WblC-repressed genes might differ from those of the WblC-activated genes. Given that, in the WblC-activated promoters, WblC binding sites are adjacent to the Ϫ35 elements of the promoter sequences to facilitate a cooperative binding of WblC and the housekeeping sigma factor HrdB (SigA for M. tuberculosis) to the promoters ( Fig. 4A to C), we tested whether such cooccurrences of the WblC binding sites and HrdB-binding promoter elements were detected in the WblC-repressed promoters by ChIP-quantitative PCR (qPCR) assay. As expected, WblCactivated promoters were enriched in both the WblC-immunoprecipitated and HrdBimmunoprecipitated DNA samples (Fig. 4E, tuf3 and tetM promoters). However, HrdB seems not to colocalize with WblC in the WblC-repressed promoters because WblCrepressed promoters were enriched only in the WblC-immunoprecipitated DNA samples, not in the HrdB-immunoprecipitated DNAs (Fig. 4E, cvnA1, SCO4914, guaB2, SCO3064, wblE, and citA promoters).
wblC is required to maintain translation and growth rate against subinhibitory concentrations of antibiotics. Given that wblC expression itself is highly induced upon translation-inhibiting antibiotic treatment (19) and a significant portion of the identified wblC regulon is involved in translation, we wondered whether WblC-controlled genes could actually promote translation rate in response to antibiotic stress. We measured the incorporation rate of 35 S-radiolabeled methionine and cysteine in the wild-type and ΔwblC cells after treating them with tetracycline or chloramphenicol for 1 h. When treated with tetracycline at a subinhibitory concentration (0.25 g/ml; Fig. S4A), wildtype cells still maintained 80% of [ 35 S]Met/Cys incorporation rate compared to untreated cells, whereas ΔwblC cells had a significantly decreased incorporation rate that WblC/WhiB7 Regulon in Streptomyces ® was approximately fourfold lower than that of the untreated control (Fig. 5). Likewise, chloramphenicol, another translation-targeting antibiotic, exhibited a similar decrease in translation when treated with ΔwblC cells at a subinhibitory concentration (5 g/ml [ Fig. 5 and Fig. S4B]). Considering that global translation efficiency is one of the major limiting factors determining growth rate (32,33), this suggests that WblC-controlled genes are responsible for maintaining translation and growth rate during translational stress mediated by antibiotics. Control experiments demonstrated the following: the incorporation rates in wild-type cells and ΔwblC cells were similar to each other and remained high in the untreated condition. However, the incorporation rates in both wild-type cells and ΔwblC cells decreased severely when treated with tetracycline at a lethal concentration (1 g/ml [ Fig. 5 and Fig. S4]).
WblC alters the composition of ribosome-associated proteins upon antibiotic stress. On the basis of the facts that WblC controls many genes involved in translation and that WblC promotes translation rate upon antibiotic stress, we suspected that WblC might actually alter the composition of the ribosome during antibiotic stress. To test this idea, we prepared 70S ribosome fractions from wild-type cells untreated or treated with tetracycline and ΔwblC mutant cells treated with tetracycline and compared their protein compositions by mass spectrometric quantification. In ribosomal fractions from wild-type cells, we found that 12 ribosome-associated proteins were significantly enriched in the tetracycline-treated sample compared to the untreated control (Fig. 6A). The alteration in the ribosomal composition depends on WblC because, even in the tetracycline-treated condition, ribosomal fractions prepared from the ΔwblC mutant did not accumulate the above-mentioned ribosome-associated proteins compared to the levels in tetracycline-treated wild-type strain (Fig. 6B). Notably, in tetracycline-treated samples, we found 54 proteins significantly enriched in the wild-type ribosome fractions but not in the ΔwblC ribosome fractions (Fig. 6B).
We were particularly interested in 10 ribosome-associated proteins that were enriched in a tetracycline-and WblC-dependent manner ( Table 2). Among those 10 proteins, eight proteins are encoded by WblC-activated genes, supporting that WblC modulates the ribosome composition during antibiotic stress. Among them, YbeZ (SCO2532) and HrpA were identified by genome-wide studies to interact with multiple ribosomal proteins in Escherichia coli (34)(35)(36). HflX is a GTPase that is also known to be associated with the 50S ribosomal subunit (37). ArfB and Hsp15 are ribosomeassociated factors that could restore a stalled ribosome or nonfunctional ribosome subunit (38-40) during antibiotic stress. Der (EngA) is an essential ribosome-binding GTPase (41, 42) that could be involved in ribosome biogenesis during antibiotic stress  because EngA depletion showed a pleiotropic effect caused by a defect in ribosome biogenesis (43,44). SCO5707 is a small conserved hypothetical protein that shares homology with rimP (encoding 30S ribosomal subunit maturation factor), infB (encoding translation initiation factor IF-2), rbfA, and truB, suggesting a possible role in protein synthesis. Although transcription of the hpf gene that encodes a ribosome hibernation promoting factor (HPF) involved in 70S ribosome dimerization (45,46) is not wblC dependent ( Table 2), enrichment of HPF in the tetracycline-treated ribosome fractions suggests that translation-inactive ribosome dimers could be formed during antibiotic stress.
WblC-dependent ribosomal alteration is involved in resistance to translationtargeting antibiotics. Given that several ribosome-associated proteins were enriched  Table 2. Blue dots represent ribosomal proteins.  . 7) compared to other translation-targeting antibiotics (Table S4). Then, we created mutant strains of genes encoding the identified ribosomeassociated proteins (hrpA, SCO2532, hflX, arfB, and helY) ( Table 2) and measured the MICs of two different translation-targeting antibiotics for the wild-type and mutant strains. Among mutant strains lacking ribosome-associated proteins, the strain lacking HrpA (ΔhrpA) was fourfold more sensitive to erythromycin and eightfold more sensitive to tetracycline than the wild-type strain was (Fig. 7). Similarly, the strain lacking SCO2532 (Δ2532) was twofold more sensitive to erythromycin and fourfold more sensitive to tetracycline than the wild-type strain was (Fig. 7), demonstrating that the ribosome-associated HrpA and SCO2532 proteins indeed contribute to resistance to translation-targeting antibiotics. Interestingly, the strain lacking HflX ribosomeassociated GTPase showed a significant effect on sensitivity to erythromycin (fourfold), supporting a previous report that a hflX homolog mediates erythromycin resistance (47,48). Complementation of the hrpA, SCO2532, or hflX gene restored resistance to erythromycin or tetracycline (except SCO2532 to tetracycline sensitivity), further supporting that these proteins directly mediate antibiotic resistance. Strains lacking HelY or ArfB exhibited a marginal effect (twofold) of sensitivity only to tetracycline (ΔhelY) or no apparent effect on sensitivity (ΔarfB), suggesting that Streptomyces might have functionally redundant homologs. In sum, these results demonstrate that ribosomeassociated proteins altered by WblC indeed participate in developing resistance to translation-targeting antibiotics.

DISCUSSION
Here, we establish that WblC directly controls 312 genes, comprising 4% of total chromosomal genes in S. coelicolor. Combined analysis of RNA-seq and ChIP-seq data identified such a large number of genes controlled by WblC, illustrating how Streptomyces develops antibiotic resistance to translation-targeting antibiotics. These include antibiotic export mediated by CmlR2 and Pep (20,21), antibiotic inactivation by acetylating drugs (Eis and Eis2) (22,49) or by linearizing the lactone ring (Vgb) (23), and ribosome protection from antibiotics by dislodging tetracycline from the ribosome (TetM) (25) or methylating rRNA (Lrm) (24) (Fig. 8). Interestingly, the WblC regulon also includes a group of genes that were previously recognized but not documented as antibiotic resistance genes. These are aminoacyl-tRNA synthetases, tRNA processing enzymes, several amino acid biosynthesis enzymes, and tRNAs that are normally required for protein synthesis to support bacterial growth (Table 1). We propose these WblC-activated genes involved in translational maintenance as another type of determinant in intrinsic antibiotic resistance, because WblC activates these genes to allow Streptomyces to continue protein synthesis and promote growth at sub-MIC levels of translation-targeting antibiotics ( Fig. 5; see also Fig. S4 in the supplemental material), thereby contributing to high levels of intrinsic resistance to translation-targeting antibiotics (Fig. 7). Interestingly, the maintenance of translation speed appears to be mediated by a WblC-induced ribosome remodeling, because WblC indeed alters the composition of 10 ribosome-associated proteins ( Fig. 6 and Table 2) that could speed up translation even in the presence of ribosome-targeting antibiotics. Among these WblC-dependent ribosome-associated proteins, ArfB, Hsp15, and HPF were also described as proteins that are required for diverse translational stresses in bacteria (50), which include mRNA truncation-mediated ribosome stalling (ArfB), heat shockmediated ribosome dissociation (Hsp15), or nutritional down-shift (HPF). The similarity in the protein requirements suggests that stress responses mediated by sub-MIC levels of antibiotics are physiologically similar to those mediated by other translational stress responses affecting protein synthesis and bacterial growth rate. Finally, the WblC- WblC controls genes that are required for counteracting translation-targeting antibiotics. Upon encountering sub-MIC levels of translation-targeting antibiotics, transcription of the wblC gene is highly induced by a positive feedback at its autoregulated promoter and upstream ORF (uORF)-mediated transcription attenuation mechanism (16,17). WblC binds to the conserved sequence (GAAAWY) located 3 nucleotides (nt) upstream of Ϫ35 elements in the promoter regions of the WblC-dependent genes and activates transcription of 288 WblC regulon genes by mediating WblC-HrdB cooperative binding to target promoters. The WblC regulon includes genes involved in decreasing antibiotic concentration, inactivating the action of antibiotics, maintaining translation (amino acyl [aa]-tRNA synthetases, tRNAs, alternative elongation factors, and release factors), or resolving redox stress mediated by antibiotics, all of which counteract translation-targeting antibiotics and support growth during antibiotic stress. Degenerate base codes in WblC-activated promoter consensus sequence are as follows: W (A or T), Y (C or T), K (T or G), M (C or A), and N (any nucleotide).
WblC/WhiB7 Regulon in Streptomyces ® mediated ribosomal alteration indeed contributes to resistance to translation-targeting antibiotics, because strains lacking the ribosome-associated proteins (HrpA, SCO2532, and HflX) that were enriched in a wblC-dependent and tetracycline-dependent manner are more sensitive to several translation-targeting antibiotics than the wild-type strain (Fig. 7).
The genus Streptomyces often contains a large number of paralogous genes, whose biological functions are not well understood. Transcription profiles of paralogous genes involved in translation suggest that the requirements of those genes may differ during antibiotic stress. For instance, alaS2 (SCO7600) and alaS (SCO1501) are paralogous genes encoding Ala-tRNA synthetases, and expression of the former is dependent on WblC and tetracycline but expression of the latter is not (Fig. S5 and Table S2) Table S2). Also, elevated mRNA levels of the WblC-activated paralogs upon tetracycline treatment seem to correlate with their proposed function that maintains translation at sub-MIC levels of antibiotics, because one of the WblC-activated paralogs, ArfB (SCO4278), was indeed enriched in the 70S ribosomal fractions during antibiotic stress ( Fig. 6 and Table 2).
Previous studies reported that large numbers of genes were induced or repressed by subinhibitory concentrations of antibiotics in other bacteria (51,52). Such a large alteration in RNA profiles affects various aspects of bacterial physiology that eventually lead to cellular adaptation or resistance to antibiotics. However, it has not clearly identified a specific set of genes or regulators responsive to each antibiotic mostly because a large number of genes were overlapped independent of antibiotics and appeared to involve multiple signaling pathways (52). WblC is a unique example in a sense that, even though it is a single transcription regulator, it is activated by sub-MIC levels of antibiotics and directly controls more than 300 genes, thereby mediating intrinsic resistance to such a wide range of antibiotics (9). Interestingly, WblC represses mRNA levels of 24 genes, suggesting that WblC also functions as a transcriptional repressor. Similar examples could be found in other WhiB-like (Wbl) family proteins. In Streptomyces, WhiB and WhiA corepress transcription of filP, which is required for the cessation of aerial hyphae at the developmental stage (53). Likewise, Mycobacterium tuberculosis WhiB1 represses its own transcription and transcription of groES2 encoding an essential chaperone by interacting with SigA sigma factor (54)(55)(56), suggesting that this could be one of the common features of WhiB-like (Wbl) family proteins. However, the functional role or physiological relevance of WblC-repressed genes is presently unclear because they are not grouped as specific functional categories.
Overall, our findings that WblC controls a large number of genes in response to antibiotics, including the newly discovered translation maintenance genes, illuminate how bacteria develop intrinsic antibiotic resistance by altering their physiology in response to antibiotics. Further research is needed to understand the underlying mechanism of intrinsic resistance, which may lead to an alternative way to treat antibiotic-resistant bacteria such as M. tuberculosis.

MATERIALS AND METHODS
Strains, growth conditions, and reagents. Streptomyces coelicolor strains and DNA primers used in this study are listed in Table S1 in the supplemental material. Spores of each strain, prepared according to standard procedures (4), were inoculated in YEME liquid medium containing 5 mM MgCl 2 and 10% sucrose and grown at 30°C with shaking at 180 rpm. For antibiotic stress conditions, a freshly made solution of tetracycline hydrochloride (Sigma) or stock solutions of chloramphenicol (Sigma) or erythromycin A dihydrate (Sigma) at the indicated concentrations were treated to early exponential cells (optical density at 600 nm [OD 600 ] of 0.1 to 0.5). For determining MIC, erythromycin (Sigma), tetracycline hydrochloride (Sigma), lincomycin hydrochloride (Fluka), chloramphenicol (Sigma), fusidic acid sodium salt (Sigma), hygromycin B concentrated solution (Duchefa), linezolid (Sigma), streptomycin sulfate salt (Sigma), thiostrepton from Streptomyces azureus (Sigma), puromycin dihydrochloride from Streptomyces alboniger (Sigma), and spectinomycin dihydrochloride pentahydrate (Fluka) were used. E. coli strains BW25113/pIJ790 and ET12567/pUZ8002 were grown and used as previously described (57), and E. coli DH5␣ was grown in LB medium for standard plasmid manipulations.

Translation rate measurement by in vivo [ 35 S]Met/Cys pulse-chase incorporation.
Cells were grown to an OD 600 of approximately 0.4 and treated with tetracycline or chloramphenicol at the indicated concentrations for 1 h. Then, cells were normalized to an OD 600 of 0.2 by adding liquid medium. Aliquots (500 l) from the normalized cultures were used for protein labeling. One microcurie of EasyTag Express 35 S protein labeling mix (PerkinElmer) was added to the aliquots and incubated at 30°C for 10 min for pulse-labeling; 0.5 mg of cold methionine was added to chase, and the sample was incubated for an additional 5 min. Cells were washed twice with phosphate-buffered saline (PBS), resuspended in 5 l PBS, and dot blotted onto Whatman filter paper. An autoradiograph was taken after overnight exposure and quantified with Multi Gauge v3.0 (Fujifilm).
Determining ribosome composition by liquid chromatography-tandem mass spectrometry. Wild-type and ΔwblC cells were grown to an OD 600 of approximately 0.4 and treated with 0.25 g/ml tetracycline for 2 h. Control cells were grown in parallel without tetracycline (final OD between 0.7 and 1.0). Then, the cells were harvested and washed with wash buffer (20 mM Tris-Cl [pH 7.4], 100 mM NaCl, 10 mM MgCl 2 ) and were lysed by sonication in lysis buffer (20 mM Tris-Cl [pH 7.4], 200 mM NH 4 Cl, 10 mM MgCl 2 , 0.5 mM EDTA, 6 mM ␤-mercaptoethanol) with 5 U/ml RNase inhibitor (Applied Biosystems) and 1 mM phenylmethylsulfonyl fluoride (PMSF). After 5 U/ml Turbo DNase (Ambion) treatment at 4°C for 15 min, lysates were clarified twice by centrifugation (22,000 ϫ g, 10 min, 4°C). Crude ribosomes were pelleted by ultracentrifugation (100,000 ϫ g, 1 h, 4°C), rinsed with lysis buffer by gentle swirling, and resuspended in lysis buffer with 10 U/ml RNase inhibitor on ice. Crude ribosomes were loaded onto a 5 to 30% sucrose gradient prepared with lysis buffer and spun at 150,000 ϫ g for 2.5 h. The 70S fractions were identified by measuring absorbance at 254 nm and pooled. Integrity of rRNA was checked using standard formaldehyde agarose gel electrophoresis for all biologically independent replicates. The purified ribosome samples were denatured with 8 M urea and reduced with 10 mM dithiothreitol (DTT). The cysteine thiol groups were alkylated with 40 mM iodoacetamide in the dark. The samples were diluted to 1 M urea and digested with trypsin (1:50 ratio [wt/wt]) overnight at 37°C. The tryptic digests were cleaned up using a C 18 -SPE column (Supelco). The samples were evaporated using a speed-vac and resuspended in 25 mM ammonium bicarbonate buffer. The final peptides were quantified using the bicinchoninic acid (BCA) assay, and then 2 g of each sample was subject to liquid chromatographytandem mass spectrometry (LC-MS/MS) analysis. LC-MS/MS experiments with higher-energy collisional dissociation (HCD) fragmentation mode were performed using Q-Exactive Orbitrap mass spectrometry (Thermo Fisher Scientific) coupled with nanoACQUITY UPLC (Waters), equipped with an in-house-packed capillary trap column (150-m inner diameter, 3 cm long) and analytical column (75-m inner diameter, 100 cm long) with 3-m Jupiter C 18 particles (Phenomenex) at a flow rate of 300 nl/min. A linear gradient (100 min) was applied for each biological replicate. Acquired data sets were analyzed by MaxQuant (v1.5.3.30) (58) with the Andromeda search engine at 10 ppm precursor ion mass tolerance against Swiss-Prot S. coelicolor database at Ͻ1% of protein false-discovery rate (FDR). The label-free quantification mode was applied to the quantitative analysis of ribosome-associated proteins. Statistical analysis of experiments was performed using Perseus (59), which accompanies MaxQuant. Proteins were considered significantly different when FDR-adjusted P values were Ͻ0.05 by the two-tailed Student's t test.
Determining MIC. MICs were determined using the resazurin assay (8) with the following modifications. A total of 10 6 spores/ml were inoculated in YEME medium, and a 150-l aliquot of the inoculum was transferred to each well in 96-well plates. Then, the inoculum was treated with twofold serial dilutions of erythromycin, tetracycline, lincomycin, chloramphenicol, fusidic acid, hygromycin B, linezolid, streptomycin, thiostrepton, puromycin, and spectinomycin and grown for 21 h at 30°C with shaking at 180 rpm. Fifteen microliters of 0.03% resazurin (Fluka) was added to each well, and cells were grown for another 1 h. MICs were determined by measuring the absorbance at 570 nm and 600 nm and then calculating the A 570 /A 600 ratio. If the A 570 /A 600 ratio was Ͻ1.2 (the value of no-growth control), we considered that growth was significantly inhibited, and the MIC was determined as the lowest concentration that exhibits an A 570 /A 600 ratio of Ͻ1.2.
Data availability. RNA-seq and ChIP-seq data have been deposited in the Gene Expression Omnibus (GEO) with the accession number GSE136168. LC-MS/MS data have been deposited in the Proteomics Identifications (PRIDE) archive with the accession code PXD015538. Genome sequences and annotations for S. coelicolor A3(2) reference genome (accession code NC_003888.3) are available from the National Center for Biotechnology Information (NCBI).

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. TEXT S1, PDF file, 0.4 MB.