A Major Facilitator Superfamily (MFS) Efflux Pump, SCO4121, from Streptomyces coelicolor with Roles in Multidrug Resistance and Oxidative Stress Tolerance and Its Regulation by a MarR Regulator

One of the key mechanisms of drug resistance in bacteria is overexpression of efflux pumps. Streptomyces species are a reservoir of a large number of efflux pumps, potentially to provide resistance to both endogenous and nonendogenous antibiotics. While many of these pumps are not expressed under standard laboratory conditions, they result in resistance to multiple drugs when spread to other bacterial pathogens through horizontal gene transfer.

ABSTRACT

Overexpression of efflux pumps is one of the major determinants of resistance in bacteria. Streptomyces species harbor a large array of efflux pumps that are transcriptionally silenced under laboratory conditions. However, their dissemination results in multidrug resistance in different clinical pathogens. In this study, we have identified an efflux pump from Streptomyces coelicolor, SCO4121, belonging to the major facilitator superfamily (MFS) family of transporters and characterized its role in antibiotic resistance. SCO4121 provided resistance to multiple dissimilar drugs upon overexpression in both native and heterologous hosts. Further, deletion of SCO4121 resulted in increased sensitivity toward ciprofloxacin and chloramphenicol, suggesting the pump to be a major transporter of these substrates. Apart from providing multidrug resistance, SCO4121 imparted increased tolerance against the strong oxidant HOCl. In wild-type Streptomyces coelicolor cells, these drugs were found to transcriptionally regulate the pump in a concentration-dependent manner. Additionally, we identified SCO4122, a MarR regulator that positively regulates SCO4121 in response to various drugs and the oxidant HOCl. Thus, through these studies we present the multiple roles of SCO4121 in S. coelicolor and highlight the intricate mechanisms via which it is regulated in response to antibiotics and oxidative stress.

IMPORTANCE One of the key mechanisms of drug resistance in bacteria is overexpression of efflux pumps. Streptomyces species are a reservoir of a large number of efflux pumps, potentially to provide resistance to both endogenous and nonendogenous antibiotics. While many of these pumps are not expressed under standard laboratory conditions, they result in resistance to multiple drugs when spread to other bacterial pathogens through horizontal gene transfer. In this study, we have identified a widely conserved efflux pump SCO4121 from Streptomyces coelicolor with roles in both multidrug resistance and oxidative stress tolerance. We also report the presence of an adjacent MarR regulator, SCO4122, which positively regulates SCO4121 in the presence of diverse substrates in a redox-responsive manner. This study highlights that soil bacteria such as Streptomyces can reveal novel mechanisms of antibiotic resistance that may potentially emerge in clinically important bacteria.

INTRODUCTION

Antibiotic resistance is a global concern affecting the efficacy of several drugs and the various available treatment regimens (1). In the course of evolution, bacteria accumulate several mutations that allow them to survive under increased concentration of antibiotics. Overexpression of efflux pumps is one such phenomenon that leads to higher drug resistance (2, 3). Efflux pumps can transport either multiple or single substrates. The former are referred to as multidrug transporters. Efflux pumps are also categorized based on the energy source. Primary transporters, such as the ATP-binding cassette (ABC) superfamily, use energy from ATP hydrolysis (4), whereas secondary transporters, such as the major facilitator superfamily (MFS), are dependent on proton motive force (5). The ABC and MFS families of transporters are ubiquitously distributed among bacteria and account for nearly half of all transporters encoded by the bacterial genome (6). Overexpression of many of these transporters is associated with multidrug resistance in different clinical pathogens (7). For example, QacA and NorA from Staphylococcus aureus (8, 9), MexAB from Pseudomonas aeruginosa (10), and EfrAB from Enterococcus faecalis (11, 12) have been found to be responsible for resistance in many clinical strains.

In addition to providing resistance to different antibiotics, overexpression of efflux pumps also plays an essential role in tolerance against toxic metabolites, heavy metals, and solvents, thereby protecting the bacteria against such stressors in the environment (13). For example, the multidrug-resistance efflux pump p55 in Mycobacterium tuberculosis and CmeG in Campylobacter jejuni are associated with oxidative stress response (14, 15); the efflux pump MexAB in Pseudomonas aeruginosa is associated with tolerance to organic solvents (16); and the MdtEF pump in Escherichia coli has been reported to protect the bacteria against nitrosative stress (17). Similarly, overexpression of PmtA in Streptococcus suis is responsible for efflux of Fe and Co ions, thus preventing metal toxicity (18).

Streptomyces species, belonging to the Actinobacteria phylum of Gram-positive bacteria, are the source of two-thirds of natural antibiotics (19). To protect themselves against the endogenous antibiotics, the bacteria employ various self-defense mechanisms. Predominantly, this self-resistance is through efflux pumps that are part of the biosynthetic cluster (20, 21). Further, as Streptomyces species are environmental bacteria, found primarily in soil or marine environments, efflux pumps play a major role in imparting the bacteria resistance to many exogenous compounds. For example, OtrC, an ABC transporter from Streptomyces rimosus, provides resistance to multiple substrates, such as doxorubicin and vancomycin, apart from the endogenous drug oxytetracycline (22). Similarly, the DrrAB pump from Streptomyces peucetius is another ABC transporter that has broad substrate specificity, besides providing resistance to the endogenous substrate daunorubicin (23). The pqrAB operon from Streptomyces coelicolor has been found to be associated with resistance to paraquat (24). CmlR1 and CmlR2 are two MFS transporters from Streptomyces coelicolor that confer resistance to the nonendogenous antibiotic chloramphenicol (25).

Many genes encoding efflux pumps have been found to be mobilized from Streptomyces to the clinical pathogens through either direct horizontal gene transfers or by involving an intermediate actinomycete host (26). Recent studies have revealed that dissemination of these genes from Streptomyces to the pathogens occurs via a carry back mechanism that involves the conjugative transfer of the resistance genes between proteobacteria and actinobacteria (27).

Streptomyces bacteria devote a large fraction of the genome toward transporter proteins (10%). These transporters can be classified as either importers, which are involved in uptake of a large number of nutrients needed for bacterial survival, or exporters, which are involved in efflux of antibiotics, metals, or other toxic metabolites. Nearly 15% of these transporters are predicted to be associated with multidrug efflux (28). Interestingly, only a tiny fraction of these have been characterized. Thus, these bacteria are a vast reservoir of yet unexplored resistance genes, offering to shed light on the myriad ways that bacteria can evade antibiotics and develop resistance.

In this work, we have used Streptomyces coelicolor as a model organism to understand efflux-mediated resistance. The 9-Mb genome encodes 7,285 proteins, of which about 800 are predicted to be efflux pumps, with 433 belonging to the ABC superfamily and 120 predicted to be MFS transporters (29). The majority of the MFS transporters are annotated as multidrug efflux pumps (TransportDB). Interestingly, most of the efflux pumps are transcribed at low levels under laboratory conditions. Subjecting bacteria to antimicrobial stress can help in deciphering the transcriptional activity of these efflux pumps (30). On similar lines, we have previously performed a transcriptomic study to understand the basis of the high resistance of wild-type (WT) Streptomyces coelicolor strain M145 to ciprofloxacin, where efflux was found to play a major role (31, 32). The resistance was reversed on the addition of carbonyl cyanide-m-chlorophenylhydrazone (CCCP), which highlights the involvement of MFS transporters in drug efflux and subsequent resistance. In this study, we have characterized one such widely conserved MFS transporter from Streptomyces coelicolor, SCO4121, which confers the bacteria with multidrug resistance to diverse, structurally unrelated drugs. Additionally, we also describe an adjacently located, divergently transcribed MarR regulator, SCO4122, which is involved in regulation of SCO4121 in the presence of its substrates.

RESULTS

SCO4121 imparts resistance to multiple antibiotics.

To identify the role of SCO4121, the native copy of sco4121 was deleted from wild-type (WT) Streptomyces coelicolor to generate a null mutant, Δ4121. An additional copy of the gene was introduced in Streptomyces coelicolor for constitutive expression of the pump from a strong promoter, pErmE*, to create the strain O4121. The strains O4121 and Δ4121 were screened against a set of diverse antimicrobials (as described in the Materials and Methods) using the agar dilution method. The O4121 strain exhibited elevated resistance toward structurally unrelated drugs that included the following: ciprofloxacin, a fluoroquinone; chloramphenicol, a peptidyl transferase inhibitor; streptomycin, an aminoglycoside; and the aromatic dye ethidium bromide (EtBr) (Table 1). A 4-fold increased MIC was observed in the case of ciprofloxacin, chloramphenicol, and streptomycin, whereas a 2-fold increased MIC was observed in case of EtBr. The deletion of sco4121 led to an increased susceptibility toward both ciprofloxacin and chloramphenicol, with a 2-fold reduction in MIC. Interestingly, no change in MIC was observed toward streptomycin and EtBr in the Δ4121 strains, suggesting the presence of other pumps that recognize these compounds as substrates and, therefore, deletion of any one efflux pump gene may not cause a change in the susceptibility of the bacteria toward them. Note that the susceptibility profile of O4121 was compared to that of the cells harboring only the empty plasmid pIJ86, referred to as WT86, whereas susceptibility of Δ4121 was compared to the WT cells. Further, the susceptibility profiles of both O4121 and Δ4121 toward these drugs were also measured using the disc diffusion assay and the data are presented in Fig. 1. The minimum concentration of the drug at which a visible clear zone was observed via disc diffusion assay was similar to the MIC of the drug as measured via the agar dilution assay, which suggests a good correlation between the two methods.

TABLE 1.

Determination of MIC of WT86, O4121, and Δ4121 cells measured against different drugs via agar dilution assay

Drug	Minimum inhibitory concentration (μg/ml)
	WT	WT86	O4121	Δ4121
Ciprofloxacin	80	80	320	40
Chloramphenicol	20	20	80	10
Streptomycin	20	20	80	20
EtBr	2	2	4	2

FIG 1.

Susceptibility of the S. coelicolor mutants O4121and Δ4121 to ciprofloxacin (A), chloramphenicol (B), streptomycin (C), and EtBr (D). The disc diffusion assay was used to measure susceptibility to various drugs. Zone sizes are measured in millimeters. Dotted lines represent the minimum zone diameter that can be measured. Error bars indicate standard deviation calculated from three biological replicates. Statistical significance of Δ4121 is calculated with respect to WT, whereas statistical significance of O4121 is measured with respect to WT86 (S. coelicolor cells harboring empty pIJ86 plasmids). ***, P value <0.0005; **, P value <0.005; *, P value <0.05; ns, no significance.

Increased resistance correlates with decreased intracellular drug levels that can be reversed with an efflux pump inhibitor.

The overexpression of sco4121 led to elevated resistance toward the different drugs tested. As noted earlier, SCO4121 is predicted to be an MFS efflux pump. Therefore, to determine whether the increased resistance in O4121 is due to increased efflux activity, we monitored the uptake profile of EtBr, a common substrate of MFS transporters. EtBr is an aromatic compound that emits weak fluorescence when present in the aqueous environment; however, on entering the cells, it binds to the cellular components and fluoresces strongly. The efflux pumps can recognize EtBr as a substrate and extrude it out. The amount of EtBr accumulated inside the cells is, therefore, a balance between the influx of EtBr into the cells via passive diffusion and the extrusion activity of the efflux pumps (33). As the O4121 clones had higher resistance against EtBr, we wanted to correlate the transport activity of the clones by measuring the intracellular amount of EtBr. As seen in Fig. 2A, O4121 cells accumulated the least amount of EtBr, with a 3-fold decrease in fluorescence intensity compared to the WT86 cells at the end of 60 min. Addition of CCCP, an efflux pump inhibitor, led to an increase in these levels comparable to those seen in WT86. This correlates with the susceptibility profile, where addition of CCCP to O4121 leads to increased sensitivity toward EtBr and restores it to that of WT86 cells (Fig. 2C). The EtBr uptake profile of Δ4121 cells was similar to that of WT cells, correlating with the susceptibility profile of these strains toward the compound (Fig. 2A and 2C).

FIG 2.

Determination of intracellular drug accumulation in S. coelicolor mutants O4121 and Δ4121 (A) Measurement of EtBr accumulation in Δ4121 and O4121 with respect to WT and WT86 cells, respectively. Accumulation was measured over a period of 60 min. EtBr and CCCP were used at a concentration of 5 μg/ml. (B) Estimation of intracellular ciprofloxacin in WT, WT86, O4121, and Δ4121. Ciprofloxacin levels were measured after treatment with the drug for 4 h at 80 μg/ml (the MIC of WT and WT86 cells) and 160 μg/ml (2× the MIC of WT and WT86 cells). Statistical significance of Δ4121 is calculated with respect to WT whereas statistical significance of O4121 is measured with respect to WT86 cells treated with respective ciprofloxacin concentrations. Accumulation of ciprofloxacin was also measured upon addition of 5 μg/ml CCCP. Statistical significance was calculated with respect to CCCP-untreated cells. (C and D) Susceptibility of WT, WT86, O4121, and Δ4121 to EtBr (C) and ciprofloxacin upon addition of CCCP (D), measured via disc diffusion assay (5 μg of CCCP was added onto the discs). Error bars indicate standard deviations calculated from three independent biological replicates. Statistical significance was calculated with respect to CCCP-untreated cells; **, P value <0.005; *, P value <0.05.

Further, to ask whether increased resistance of O4121 toward ciprofloxacin also correlates with reduced drug levels, we compared the uptake profile of ciprofloxacin in O4121 with respect to WT86 (cells with the empty pIJ86 plasmids). The uptake was measured at the MIC and 2× MIC corresponding to WT86 strains. A 2-fold reduced accumulation of ciprofloxacin was observed in O4121 cells compared to the WT86 cells. These data suggest that cells overexpressing SCO4121 were able to extrude the drug more effectively than the WT86 cells (Fig. 2B). On addition of CCCP to the O4121 cells, uptake of ciprofloxacin became comparable to that in WT86 cells, presumably as the efflux activity was abolished by CCCP. Addition of CCCP also restored the susceptibility of O4121 strains toward ciprofloxacin to levels similar to WT86 (Fig. 2D), indicating that the pump requires a proton gradient to extrude the antibiotics. The intracellular levels of ciprofloxacin in Δ4121 also correlated with the 2-fold decrease in MIC of ciprofloxacin in these cells, compared to WT. In the absence of the SCO4121 efflux pump, ciprofloxacin concentrations were almost twice those in WT cells (Fig. 2B). Further, addition of CCCP to Δ4121 did not result in any significant change in susceptibility toward ciprofloxacin (Fig. 2D). We thus conclude that SCO4121 is a major transporter of the drug in S. coelicolor. Note that the ciprofloxacin concentration was measured after 4 h of drug treatment, as the drug uptake kinetics reach steady state by this time (34).

Similarly, addition of CCCP also restored the susceptibility of O4121 strains toward chloramphenicol and streptomycin to WT86 levels, suggesting that these drugs are also extruded by the pump via utilization of a proton motive force (Fig. S1 in the supplemental material).

SCO4121 provides elevated resistance when overexpressed in a heterologous host, E. coli.

To further characterize the function of SCO4121, we introduced multiple copies of the corresponding gene into the expression host E. coli strain BL21(DE3), yielding the strain OE4121. Upon induction with 0.5 mM IPTG (isopropyl-β-D-thiogalactopyranoside) for 12 h at 28°C, maximum expression of sco4121 was detected in the E. coli cells. We further measured the susceptibility of both uninduced and IPTG-induced cells toward different drugs with respect to WT28a cells (E. coli cells with empty pET28a plasmids). OE4121 exhibited elevated resistance against the same set of drugs as O4121 (Table 2). In the induced cells, a 2-fold increase in MIC of both streptomycin and EtBr was observed, whereas the MIC of ciprofloxacin and chloramphenicol increased by 10-fold and 8-fold, respectively. Increased MIC toward these drugs was also observed in the case of uninduced cells due to significant basal expression of SCO4121 in E. coli, as determined by real-time PCR. These results from a heterologous host further support our earlier conclusion that SCO4121 recognizes ciprofloxacin and chloramphenicol as major substrates, thus providing increased resistance against them.

TABLE 2.

Determination of MIC of WT28a and OE4121 cells measured by broth microdilution assay

Drug	Minimum inhibitory concentration (μg/ml)
	WT	WT28a	OE4121 uninduced	OE4121 0.5 mM IPTG
Ciprofloxacin	0.02	0.02	0.2	0.2
Chloramphenicol	8	8	32	64
Streptomycin	32	32	64	64
EtBr	50	50	80	100

To understand whether increased resistance is a measure of increased efflux in OE4121, we determined the EtBr uptake profile of OE4121 compared to that of WT28a. A 3-fold reduction in the fluorescence intensity was observed in OE4121 cells compared to WT28a cells. However, upon addition of the proton gradient inhibitor CCCP, the EtBr levels were restored to WT28a levels (Fig. 3A) indicating that the decreased intracellular drug levels are due to enhanced efflux in OE4121 that provides increased resistance toward the drug.

FIG 3.

Effect of expression of SCO4121 in E. coli on intracellular drug levels and cell viability. (A) EtBr accumulation in OE4121 compared to WT28a. Accumulation was compared with and without induction with IPTG (0.5 mM) with respect to WT28a cells (E. coli cells with empty pET28a plasmids). EtBr at a concentration of 25 μg/ml was used for the assay. CCCP at a concentration of 5 μg/ml was added to the IPTG-induced cells. (B) Survival kinetics of OE4121 in the presence of 10× MIC of ciprofloxacin (0.2 μg/ml) and chloramphenicol (80 μg/ml). Survival was measured by taking samples at different time points and plotted as log₁₀ CFU/ml. WT28a cells were used as a control. Error bars indicate standard deviation calculated from three independent biological replicates.

To compare the survival kinetics of IPTG-induced OE4121 with that of WT28a, the cells were exposed to the drugs at 10× MICs. Increased survival of OE4121 in the presence of both ciprofloxacin and chloramphenicol was observed. Upon exposure to both drugs, an early bactericidal effect was noticed in WT28a cells, where no viable cells could be detected at the end of 60 min. In contrast, OE4121 cells survived with a viability of 89% and 87% at the end of 60 min and complete killing could only be seen after 120 min of exposure to either drug (Fig. 3B). Thus, killing curve data show a smaller and gradual decrease in growth in OE4121 in the presence of the drugs, compared to their respective controls, indicating that the increased survival can be attributed to the enhanced efflux of these drugs in OE4121.

Membrane permeability remains unaffected in S. coelicolor cells overexpressing SCO4121.

To determine if overexpression of the efflux pump SCO4121 affects membrane permeability, leading to reduced drug uptake, we quantified permeability of the O4121 and Δ4121 strains via a 1-phenylnaphthylamine (NPN) assay. NPN, a hydrophobic dye, is generally excluded from the membrane. However, if the membrane is permeabilized, the dye can bind to the exposed hydrophobic surfaces of the cells, resulting in an increased fluorescence. This is a widely used technique applied to measure membrane permeability of Gram-negative cells, where NPN binds to the exposed lipopolysaccharide (LPS) in the presence of a membrane permeabilizer and leads to increased fluorescence (35). Some evidence for its application in Gram-positive bacteria has also been seen (36).

O4121 and Δ4121 had NPN-binding profiles similar to those of WT86 and WT cells, respectively. Thus, overexpression or deletion of sco4121 does not affect membrane permeability. Increased resistance in O4121 is therefore solely due to the increased efflux activity. Cells treated with membrane-permeabilizing compounds cetyltrimethylammonium bromide (CTAB) and sodium dodecyl sulfate (SDS) were used as positive controls for the assay (Fig. 4A).

FIG 4.

Comparison of membrane permeability and cell morphology in different S. coelicolor strains. (A) Membrane integrity profile of S. coelicolor mutants O4121 and Δ4121 measured using the NPN uptake assay. Statistical significance was calculated with respect to the WT in Δ4121 cells but with respect to WT86 in O4121 cells. Cells treated with 30% CTAB and 10% SDS were taken as positive controls. Statistical significance was calculated with respect to WT for these treated cells. (B) Cryo-SEM imaging of S. coelicolor WT86 cells and recombinants O4121 and Δ4121. Cells were grown to an OD at 450 nm of 1 and washed with 0.1 M PBS followed by 10% ethanol to remove the medium contaminants before imaging under the Cryo FEG SEM (JSM-7600F) platform. Error bars indicate standard deviations calculated from three independent biological replicates; **, P value <0.005; ns, no significance. Scale bars correspond to 10 μm.

The morphological appearance of the mutants observed under cryo-scanning electron microscopy (SEM) imaging also revealed that overexpression or deletion of sco4121 had no effect on their membrane integrity (Fig. 4B). Thus, the decreased intracellular accumulation of drugs coupled with a lack of change in membrane permeability indicates that the change in the efflux activity is responsible for the altered drug susceptibility of O4121 and Δ4121.

SCO4121 is transcriptionally regulated by a divergently transcribed MarR regulator, SCO4122.

As previously observed, ciprofloxacin and chloramphenicol were the two major drugs against which elevated resistance was observed in the O4121 strains (Fig. 1). We further explored the involvement of these two drugs in induction of the native promoter of sco4121 in the WT and WT86 cells. An ∼2-fold (log₂ ratio) upregulation of sco4121 was observed when subjected to inhibitory concentrations of ciprofloxacin (Fig. 5A) and chloramphenicol (Fig. 5B), confirming that sco4121 is transcriptionally regulated by both drugs.

FIG 5.

Relative expression of sco4121 in different S. coelicolor strains. (A) Expression of sco4121 in WT, WT86, Δ4122, O4122, and cΔ4122 cells treated with subinhibitory (20 μg/ml) and inhibitory (80 μg/ml) concentrations of ciprofloxacin. (B) Expression of sco4121 in WT, WT86, Δ4122, O4122, and cΔ4122 cells treated with subinhibitory (5 μg/ml) and inhibitory (20 μg/ml) concentrations of chloramphenicol. Cells treated with inhibitory concentrations of ampicillin (25 μg/ml) were taken as the negative control. Expression was measured by quantitative real-time PCR. Expression was normalized using 23S rRNA as the housekeeping gene. Expression of sco4121 was normalized with respect to WT86 in O4122 and cΔ4122 strains, and with respect to WT in the Δ4122 strain. Statistical significance in O4122 and cΔ4122 was calculated with respect to WT86 cells, whereas statistical significance in Δ4122 strains was calculated with respect to WT cells. Error bars indicate standard deviations calculated from three independent biological replicates; ***, P value <0.0005; **, P value <0.005; *, P value <0.05.

Further, we predict the involvement of an adjacently located MarR regulator, SCO4122, in governing the expression of SCO4121 in the presence of the drugs. SCO4122 is a 22-kDa regulatory protein that has previously been found to bind a 26-bp DNA sequence upstream of sco4121, thus suggesting a role in the regulation of SCO4121 (37). To elucidate whether SCO4122 regulates the expression of SCO4121, we deleted the native copy of sco4122 from WT S. coelicolor to generate the strain Δ4122. The gene copy was restored by introducing sco4122 under the pErmE* promoter in the mutant to generate the complementation strain, cΔ4122. In addition, sco4122 was also overexpressed from a multicopy replicative plasmid pIJ86 to yield the strain O4122.

In the Δ4122 strain, sco4121 expression levels were found to be downregulated by 4-fold (log₂ ratio) and were not restored even upon addition of either ciprofloxacin (Fig. 5A) or chloramphenicol (Fig. 5B). This suggests that SCO4122 is essential for maintaining the basal sco4121 levels. To further elucidate the mechanism by which SCO4122 regulates SCO4121, we measured the mRNA expression levels of sco4121 in O4122. As seen in Fig. 5A and B, sco4121 levels remained unchanged in O4122 in the uninduced state. However, when treated with an inhibitory concentration of ciprofloxacin or chloramphenicol, sco4121 levels were upregulated by ∼10-fold (log₂ ratio) in a concentration-dependent manner (Fig. 5A and 5B). This indicates that both ciprofloxacin and chloramphenicol are involved in transcriptional regulation of SCO4121 mediated via SCO4122. The basal expression of sco4121 is restored in cΔ4122 and further upregulated under increasing concentrations of both ciprofloxacin (Fig. 5A) and chloramphenicol (Fig. 5B). This establishes that SCO4122 has a role in positive regulation of SCO4121 in the presence of the drugs. Note that in the presence of ampicillin, a drug whose MIC is unaltered by overexpression or deletion of either the sco4121 or sco4122 gene (Table S1), no change in expression of sco4121 or sco4122 was observed.

Overexpression of SCO4122 also leads to increased resistance against different drugs.

As overexpression of sco4122 leads to increased gene expression levels of sco4121, we investigated the susceptibility of O4122 toward ciprofloxacin and other drugs. O4122 exhibited increased resistance toward different antibiotics, including ciprofloxacin, chloramphenicol, streptomycin, and EtBr (Fig. 6), which were also found to be extruded by the O4121 strains (Fig. 1). These data thus confirm our hypothesis that SCO4122 is involved in increased expression of sco4121 and therefore increased resistance toward similar substrates. Interestingly, susceptibility of the Δ4122 strain to the different drugs tested was similar to that of the WT strain. These data thus suggest that basal levels of sco4121 are sufficient to maintain the susceptibility of S. coelicolor toward both ciprofloxacin and chloramphenicol. However, upon complementation of the Δ4122 strain with sco4122 on a multicopy plasmid, the resistance toward different drugs was restored and was similar to that of O4122 strains, corresponding to the increased levels of sco4121 in the complementation strain. These findings thus suggest that expression of sco4122 is essential for providing a resistance phenotype to the cells and possibly does this via upregulation of sco4121.

FIG 6.

Susceptibility of the S. coelicolor mutants Δ4122, O4122, and cΔ4122 to ciprofloxacin (A) chloramphenicol (B), streptomycin (C), and EtBr (D). The disc diffusion assay was used to measure susceptibility to various drugs. Zone sizes were measured in millimeters. The dotted lines represent the minimum zone diameter that can be measured. Statistical significance for Δ4122 was calculated with respect to WT cells, whereas statistical significance for O4122 and cΔ4122 was calculated with respect to WT86 (S. coelicolor cells harboring empty pIJ86 plasmids). Error bars indicate standard deviations calculated from three independent biological replicates; ***, P value <0.0005; **, P value <0.005; *, P value <0.05.

Substrates of SCO4121 activate transcription of sco4121 via SCO4122.

Our results have so far suggested that SCO4122 activates SCO4121 in the presence of drugs recognized by SCO4121. We next focused on whether SCO4122 has a direct or indirect role in regulation of SCO4121 in the presence of the drugs. For this, we used a different actinomycete host, Mycobacterium smegmatis, which lacks homologs of either SCO4121 or SCO4122. Mycobacterium smegmatis has a similar GC content as that of Streptomyces, thus minimizing codon bias and leading to efficient transcription of Streptomyces genes. We chose the nonpathogenic mycobacterium strain M. smegmatis MC²155 to constitutively express SCO4122 by integrating a copy of the sco4122 gene into the mycobacterial genome (38), yielding the strain MS4122. These cells were transformed with a modified pFpV27 plasmid on which the luciferase reporter gene was present downstream of the sco4121 promoter, forming the construct MS4122::P₄₁₂₁. The activity of the sco4121 promoter was determined by measuring the mRNA levels of luciferase in MS4122::P₄₁₂₁ cells in the absence and presence of all the drugs tested previously. As a positive control, MS4122 cells constitutively expressing the luciferase gene under the Pmyc promoter were used. We observed that luciferase mRNA could be detected within 15 min of exposure to different drugs that were previously identified as the substrates of SCO4121 (Fig. 7). However, no luciferase expression could be detected in the presence of apramycin and ampicillin, which were not substrates of SCO4121. Further, the expression was transient in response to streptomycin, as the mRNA levels significantly reduce and become similar to basal levels at times after 1 h of treatment. In contrast, when treated with ciprofloxacin and chloramphenicol, the luciferase expression could be detected until 1 h, after which it falls and becomes similar to basal levels when measured after 4 h of drug treatment. EtBr led to a prolonged expression of luciferase which could be detected until 4 h, which suggests that the drug is a strong activator of SCO4121. Thus, we observed differential temporal expression of luciferase for different drugs, with all of them effectively inducing the sco4121 promoter. Also, in the absence of sco4122, there was no induction of the sco4121 promoter even upon treatment with the drugs (Fig. S2A). Note that constitutive expression of sco4122 in M. smegmatis under the Pmyc promoter did not affect the MIC of the different drugs tested (data not shown) and also did not impact the intracellular drug levels, as measured by the EtBr uptake assay (Fig. S3). Further, the luciferase expression from the Pmyc promoter remained unaffected in these cells when exposed to the drugs (Fig. S2B). Thus, from this study, we find that the regulation of SCO4121 by the drugs is mediated directly via SCO4122 by converting the latter into an effective activator of SCO4121 expression.

FIG 7.

Estimation of sco4121 promoter activity in the presence of SCO4121 substrates. The sco4121 promoter activity was detected by measuring luciferase expression upon exposure of MS4122::P₄₁₂₁ to different drugs at MICs (ciprofloxacin, 0.04 μg/ml; chloramphenicol, 7 μg/ml; streptomycin, 7 μg/ml; EtBr, 7 μg/ml). Samples were taken at 15 min, 60 min, and 240 min after exposure to the drugs. The expression level was normalized with 16S rRNA followed by normalization with the untreated sample taken at the same time points. Ampicillin and apramycin were used as the negative controls. Error bars indicate standard deviations calculated from three independent biological replicates. Statistical significance was calculated with respect to untreated MS4122::P₄₁₂₁ cells at respective time points; **, P value <0.005; *, P value <0.05.

SCO4121 is activated in response to oxidative stress mediated through activation of SCO4122.

Characterization of the SCO4121 efflux pump has unraveled its role in antibiotic resistance. The expression of sco4121 is upregulated in response to many drugs. Further, SCO4122, a MarR regulator, is essential for SCO4121 activation (Fig. 5 to 7). Interestingly, all the drugs recognized by SCO4121 have been previously shown to lead to generation of reactive oxygen species (ROS) in many Gram-positive and Gram-negative bacteria, such as E. coli, S. aureus, and Pseudomonas putida (39–42). On similar lines, ROS were produced when S. coelicolor WT86 cells were exposed to inhibitory concentrations of these drugs as measured via H₂-DCFDA (2,7-dichlorofluorescein diacetate) assay (Fig. 8A). Further, the amount of ROS generated in S. coelicolor cells in response to the drugs was similar to the ROS produced in the presence of an inhibitory levels of HOCl. Therefore, to determine whether HOCl is also an inducer of the sco4121 gene, expression levels of sco4121 were measured in response to an inhibitory concentration of the oxidant. As seen in Fig. 8B, sco4121 was upregulated ∼4-fold (log₂ ratio) in response to HOCl. Interestingly, while overexpression of sco4121 did not lead to any change in the susceptibility of cells toward HOCl, deletion of sco4121 made the cells more susceptible to the oxidant (Fig. 8C). However, no induction of sco4121 was noted in the presence of H₂O₂, a weak oxidant, which indicates the pump is activated only under strong oxidizing conditions.

FIG 8.

Roles of ROS in activation of SCO4121. (A) ROS generation in S. coelicolor WT86 cells on exposure to different drugs at MICs (ciprofloxacin, 80 μg/ml; chloramphenicol, 20 μg/ml; streptomycin, 20 μg/ml; EtBr, 2.5 μg/ml; H₂O₂, 0.5 mM; HOCl, 1 mM; ampicillin, 50 μg/ml; apramycin, 25 μg/ml) measured via DCFDA assay. (B) Relative expression of sco4121 upon exposure of WT86 cells to subinhibitory and inhibitory levels of HOCl (0.5 mM and 1 mM, respectively) and H₂O₂ (0.25 mM and 0.5 mM, respectively). Expression was measured via quantitative real-time PCR (qRT-PCR) using the 2^-ΔΔCt method. Expression was normalized with 23S rRNA as the housekeeping gene followed by normalization with the untreated WT86 cells. (C) Measurement of susceptibility of WT, WT86, O4121, and Δ4121 cells to HOCl using the disc diffusion assay. Statistical significance was calculated with respect to WT cells. Error bars indicate standard deviations calculated from three independent biological replicates; **, P value <0.005; *, P value <0.05.

Similarly, in the MS4122::P₄₁₂₁ cells, the drugs also led to increased ROS levels (Fig. 9A) and increased expression of the sco4121 promoter was noted in response to the oxidant HOCl, as measured by upregulation of luciferase ∼4-fold (log₂ ratio) upon exposure to inhibitory concentrations of the oxidant (Fig. 9B). However, the sco4121 promoter was not induced in response to the weak oxidant H₂O₂ in these cells. These data thus suggest the sco4121 promoter is indeed directly induced in response to strong oxidative stress. Note that no induction of the sco4121 promoter was observed upon addition of HOCl to cells devoid of SCO4122.

FIG 9.

Oxidative stress-mediated activation of SCO4121 in MS4122::P₄₁₂₁ cells. (A) ROS levels measured by exposing MS4122::P₄₁₂₁ cells to SCO4121 substrates at MICs (ciprofloxacin, 0.04 μg/ml; chloramphenicol, 7 μg/ml; streptomycin, 7 μg/ml; EtBr, 7μg/ml). ROS were measured using the DCFDA assay. Ampicillin was used as the negative control. (B) Estimation of relative luciferase mRNA expression in MS4122::P₄₁₂₁ cells upon exposure to subinhibitory and inhibitory levels of HOCl (0.5 mM and 1 mM, respectively) and H₂O₂ (0.125 mM and 0.25 mM, respectively). Expression was measured via real-time quantitative PCR and normalized with 16S rRNA as the housekeeping gene, followed by normalization with untreated cells. (C) ROS generation in MS4122::P₄₁₂₁ cells subjected to EtBr at the inhibitory concentration of 7 μg/ml. Ascorbate was added to EtBr-treated cells at 200, 400, and 800 μg/ml; then the cells were incubated for 15 min at 37°C before measuring ROS. (D) Expression of luciferase was measured on addition of EtBr at a concentration of 7 μg/ml and ascorbate at various concentrations of 200, 400, and 800 μg/ml. Expression was measured via real-time quantitative PCR and normalized with 16S rRNA as the housekeeping gene, followed by normalization with untreated cells. Statistical significance was calculated with respect to untreated MS4122::P₄₁₂₁ cells. Error bars represent standard deviations calculated from three independent biological replicates; **, P value <0.005; *, P value <0.05.

Thus, HOCl is also an inducer of the SCO4121-SCO4122 system, in addition to the various antimicrobial drugs. Interestingly, oxidative stress has been found to play an important role in activation of many transcriptional regulators, which further activate other genes upon oxidation. For example, MexR is activated in response to oxidative stress and further regulates the MexAB transporter in the Gram-negative bacterium Pseudomonas aeruginosa (43). In the Gram-positive bacterium Bacillus subtilis, HypR, a MarR regulator, is activated upon oxidation and this positively regulates HypO, a flavin oxidoreductase involved in resistance to sodium hypochlorite (44). HypT, a redox-sensing regulator in E. coli, is activated in response to HOCl and further regulates genes that protect the bacteria against the oxidant (45).

We hypothesize a similar mechanism to be at play in the activation of the SCO4121 efflux pump by the MarR regulator SCO4122. We propose that generation of ROS by the drugs could be a plausible mechanism that activates SCO4122, which in turn regulates expression of SCO4121. To validate ROS-mediated activation of the sco4121 promoter, we quantified the ROS generated in the MS4122::P₄₁₂₁ strain upon addition of ascorbate, an ROS-quenching drug. As seen in Fig. 9C, ascorbate was able to quench the ROS generated upon addition of EtBr to the cells. Further, we observed that addition of ascorbate at different concentrations to the EtBr-exposed cells also reduced the expression from the sco4121 promoter to basal levels (Fig. 9D). Together, these results support our hypothesis that increased ROS levels are essential to induce expression of sco4121, which is mediated possibly via oxidation of SCO4122.

SCO4121-SCO4122 system is evolutionary conserved among different Streptomyces species.

To further unravel the evolutionary role of the SCO4121-SCO4122 system, we performed a phylogenetic analysis on SCO4121 and SCO4122 across Streptomyces. Using BLAST analysis, orthologs of SCO4121 were found in 666 species, whereas 729 species contained orthologs of SCO4122 with sequence similarity greater than 50% and coverage greater than 90%. Among them, 403 species had orthologs of both SCO4121 and SCO4122. In the majority of these species, SCO4121 and SCO4122 were found to be encoded at different loci in the genome. Interestingly, we discovered 31 Streptomyces orthologs (Table S2) where SCO4121 and SCO4122 were encoded by adjacent genes and had greater than 50% identity in protein sequence. Most of these species belonged to the soil environment, suggesting a widespread conservation of the SCO4121-SCO4122 system in soil Streptomyces. We utilized the neighbor joining algorithm to construct a phylogenetic tree using these 31 species (Fig. 10), which provides an insight into the evolutionary conservation of this system. Thus, the conservation and cooccurrence of the SCO4121-SCO4122 orthologs in a large number of Streptomyces species highlights the potential functional role of this system for aiding bacterial growth and survival.

FIG 10.

Evolutionary conservation of the SCO4121-SCO4122 system. The phylogenetic tree was constructed using 31 Streptomyces species for which genes encoding SCO4121 and SCO4122 are adjacently located in the genome. Sequences having average identity of >50% and coverage of >90% were used for the tree construction. The sequences were retrieved from NCBI. Sequences were aligned using the CLUSTAL W algorithm. The phylogenetic tree was constructed using the neighbor joining algorithm in MEGA X. The data were created with a bootstrap value of 1,000 replicates. The scale bar of 0.02 represents estimated sequence divergence. The numbers on the nodes represent the phylogenetic confidence of tree topology.

DISCUSSION

Streptomyces species are a rich reservoir of efflux pumps that provide resistance against both endogenous and nonendogenous antibiotics. In this study, we characterized an MFS transporter, SCO4121, which is conserved throughout the Streptomyces group, suggesting it is encoded in the core genome. Deletion of the pump primarily led to increased susceptibility toward ciprofloxacin and chloramphenicol. Additionally, overexpression of the pump led to elevated resistance against chloramphenicol and streptomycin, endogenous antibiotics produced by Streptomyces venezuelae and Streptomyces griseus, respectively (46, 47). However, sequence analysis revealed that SCO4121 possesses limited similarity to these earlier-characterized pumps. While SCO4121 exhibits 37% similarity with CmlV, the chloramphenicol transporter from Streptomyces venezuelae, it is 43% similar to StrV, the streptomycin transporter from Streptomyces griseus. Also, SCO4121 was weakly similar (<30%) to CmlR1 and CmlR2, the two previously characterized chloramphenicol efflux pumps from Streptomyces coelicolor (25). Thus, SCO4121 is an efflux pump with a wide range of structurally unrelated substrates. Interestingly, both ciprofloxacin and chloramphenicol are substrates of many efflux pumps in Gram-positive bacteria, such as PdrM and PrmA in Streptococcus pneumoniae (48–50), LmrP in Lactococcus sp. (51), and Bmr in Bacillus subtilis (52, 53). Chloramphenicol and streptomycin resistance is also interrelated, as seen in the Gram-negative bacterium Pseudomonas putida, where the efflux pump TtgABC recognizes both drugs (54). However, examples of efflux pumps that recognize all three substrates are very few, with OqxAB, ubiquitously distributed among the Enterobacter spp., being one such efflux pump (55). Thus, SCO4121 is an efflux pump with a broad substrate specificity which provides resistance against multiple dissimilar drugs in Streptomyces coelicolor.

Apart from multidrug resistance, SCO4121 also provided elevated resistance toward the oxidant HOCl, as evidenced by enhanced susceptibility toward HOCl in S. coelicolor cells devoid of the pump. Efflux pumps have evolved in bacteria to serve potential functional roles in bacterial physiology such as detoxification of intracellular metabolites and maintaining cellular homeostasis, with antibiotic resistance as a fortuitous effect (56). Similarly, many efflux pumps serve to protect the bacteria against oxidative stress generated due to aerobic respiration or exposure to oxidants such as H₂O₂ and HOCl. Additionally, these efflux pumps also provide resistance to multiple drugs as a secondary effect. Some of the examples include the efflux pump P55 in Mycobacterium tuberculosis (14), MacABCsm in Stenotrophomonas maltophilia (57), and HypO in Mycobacterium smegmatis (58). The MFS transporter PqrB, encoded in the pqrAB operon, has been reported to provide resistance against the redox drug paraquat in Streptomyces coelicolor (24). However, examples of efflux pumps that sense oxidative stress are rare in Streptomyces species. SCO4121, characterized in this study, is one of the few such examples of this type of efflux pump.

Streptomyces species produce many secondary metabolites that can be modulated by the balance of ROS in the cell (59, 60). Some of the secondary metabolites can also induce oxidative stress in the bacterium itself (61). Thus, the bacterium would have evolved mechanisms to protect itself against such stress. The ability of SCO4121 to provide tolerance to HOCl-mediated oxidative stress, along with multidrug resistance, suggests this pump has a larger physiological role in the bacterium during secondary metabolism. The increased expression of this gene during late exponential phase of culture (data not shown) further corroborates this hypothesis.

Efflux pumps are energy-intensive and therefore generally held under tight regulation in bacteria. This regulation primarily occurs at the transcriptional level in bacteria through a DNA-binding regulator (62, 63). On similar lines, the efflux pump SCO4121 was found to be under regulation by an adjacently placed MarR regulator, SCO4122 (Fig. 11). The in vivo promoter activity assays of sco4121 in the presence of SCO4122 suggest that SCO4121 is positively regulated in the presence of the pump substrates. Interestingly, most of the characterized multidrug efflux systems in different bacteria are predominantly found to be repressed under physiological conditions (64), such as the AcrAB in E. coli (65), EmrB in Burkholderia thailandensis (66), MexAB in Pseudomonas aeruginosa (67), and QacA in Staphylococcus aureus (68). In contrast, efflux pumps that undergo positive regulation in response to the substrates recognized by the pump are few. AceI, from the PACE family of transporters, has been found to be positively regulated by a LysR regulator, AceR, in the presence of the pump substrate chlorhexidine in the Gram-negative bacterium Acinetobacter baumannii (69). Also, in the Gram-positive bacterium Bacillus subtilis, the MerR regulator BmrR regulates the efflux pump Bmr in the presence of 6-rhodamine and thiamine PP_i (TPP), both of which are substrates of Bmr (52). Similarly, SCO4121 is also positively regulated by SCO4122 through direct interaction with the sco4121 promoter in the presence of the pump substrates.

FIG 11.

Schematic representing the SCO4121-SCO4122 system. E represents the efflux pump SCO4121 and A represents the MarR regulator SCO4122. Binding of drugs (D), unknown metabolites (U), and ROS are possible factors for activation of SCO4122. The activated SCO4122 now becomes an activator of SCO4121. The efflux pump upon expression provides elevated resistance to multiple drugs and oxidant HOCl.

In addition to the various drugs, the sco4121 promoter is significantly upregulated in the presence of the oxidant HOCl. Interestingly, the various drugs that lead to differential expression of the sco4121 promoter are ROS-generating drugs. This upregulation is only dependent on the presence of the MarR regulator SCO4122, as suggested by the data presented in Fig. 7, where expression of SCO4122 alone in M. smegmatis is sufficient for transcriptional activity from the SCO4121 promoter. Thus, SCO4122 could be a probable redox-responsive regulator which is activated via oxidation in response to SCO4121 substrates. We hypothesize that SCO4122 senses oxidative stress either via direct binding of the drugs or through the ROS generated upon exposure to these drugs (Fig. 11). Alternatively, activation of SCO4122 may be brought about by a combinatorial effect of drug binding and oxidation which then activates SCO4121. Binding of SCO4122 to the sco4121 promoter region, in the absence of any drug, has been reported earlier (37). Enhanced expression of the efflux pump Bmr by the regulator BmrR in the presence of substrates of Bmr has been reported in the Gram-positive bacterium Bacillus subtilis (52). On similar lines, we predict that SCO4122 binds to the sco4121 promoter with higher affinity in the presence of the drugs, thus leading to increased expression of the gene (Fig. 11).

Studies suggest that redox-sensing regulators utilize thiol-based switches that either activate or inactivate the transcription factors, thus leading to altered binding to cognate DNA and differential gene expression (70). Interestingly, there are several HOCl-specific transcription factors that, upon oxidation, lead to differential expression of several genes responsible for HOCl tolerance. Some of these transcriptional regulators are OhrR, HypR, and PerR in Bacillus subtilis (44, 71), HypS in Mycobacterium smegmatis (58), and NemR, RclR, and YjiE (renamed HypT) in E. coli, with YjiE as the first known HOCl-specific transcriptional regulator (45, 72–74). Earlier studies have revealed the presence of several redox-sensing MarR regulators such as OhrR, TamR, and SCO2646 in S. coelicolor that serve a role in oxidative stress response and control key metabolic pathways in the bacteria (75–79). However, none of them have been found to be induced under HOCl stress. While further experiments are required to understand activation of SCO4122 by HOCl, we speculate that SCO4122 is activated in response to oxidative stress and then further activates the cognate transporter SCO4121, which operates to curb the ROS stress and thus aid bacterial survival. The phylogenetic analysis reveals the SCO4121-SCO4122 system to be highly conserved among the Streptomyces species, with many of them having the genes adjacently positioned. This suggests that the SCO4121-SCO4122 system operates in a concerted fashion, perhaps to enable the bacteria to survive under strong oxidative stress conditions.

Thus, taken together, we report the finding of the efflux pump SCO4121 in Streptomyces coelicolor that functions beyond resistance to antibiotics and provides benefit to the cells in overcoming oxidative stress. Additionally, we report SCO4122, a MarR regulator that operates in a coordinated fashion with the cognate transporter SCO4121 to maintain effective redox homeostasis in cells. The ability of the pump to be induced by several unrelated drugs suggests a common signal, such as the generation of ROS, that the bacteria senses in response to its own metabolites or metabolites from the organisms residing in its vicinity. These findings thus signify the widespread conservation of the SCO4121-SCO4122 system in the Streptomyces group and its functional role in bacterial physiology and survival. Further investigation to understand whether the drugs directly or indirectly activate SCO4122 can shed light on the molecular mechanism behind operation of the SCO4121-SCO4122 system.

MATERIALS AND METHODS

Bacterial strains, plasmids, growth media, and conditions.

E. coli BL21(DE3) harboring the construct OE4121 was grown in Luria-Bertani (LB) medium (Himedia) at 37°C until mid-logarithmic phase. Upon reaching the mid-log phase, the cells were shifted to 28°C and induced with various concentrations of IPTG (Merck) for 12 h to increase expression of the efflux pump. Kanamycin (50 μg/ml) (Himedia) was used as the selection pressure for the selection of the recombinants. As a control, WT28a (cells with empty pET28a plasmids) were grown under the same conditions. Streptomyces coelicolor (WT and recombinants) were grown in R5 medium (1% wt/vol of MgCl₂, 10% sucrose, 1% glucose, 0.003% K₂SO₄, 0.002% Casamino Acids, 0.5% yeast extract, 0.5% TES buffer, trace elements) at 30°C for all growth conditions and assays. The temperature of 39°C was chosen for selection of knockout mutants. For growing spores of S. coelicolor, mannitol soya agar (MSA) plates (2% mannitol, 2% soya flour, 2% agar) were used. All strains used in this study are listed in Table 3.

TABLE 3.

List of strains used in this study

Strain	Organism	Characteristics	Reference
M145	S. coelicolor	Devoid of SCP1 and SCP2 plasmids	(29)
O4121	S. coelicolor	sco4121 cloned into pIJ86 under the strong constitutive promoter pErmE*	This work
O4122	S. coelicolor	sco4122 cloned into pIJ86 under the strong constitutive promoter pErmE*	This work
Δ4122	S. coelicolor	sco4122 knockout mediated by homologous recombination using the temp sensitive plasmid pKC1139	This work
Δ4121	S. coelicolor	sco4121 knockout mediated by homologous recombination using the temp sensitive plasmid pKC1139	This work
cΔ4122	S. coelicolor	Δ4122 complemented with native copies of sco4122 in the multicopy plasmid pIJ86 under pErmE* promoter	This work
pFpV27::luc	M. smegmatis	Modified Mycobacterium smegmatis cells with the luciferase reporter gene replacing the gfp reporter	This work
MS4122	M. smegmatis	Modified Mycobacterium smegmatis cells bearing sco4122 integrated into the genome	This work
P4121	M. smegmatis	P4121::luc transformed into WT M. smegmatis; cells not expressing SCO4122	This work
MS4122::P4121	M. smegmatis	MS4122 cells transformed with luciferase reporter construct with an upstream sco4121 promoter	This work
MS4122::Pmyc	M. smegmatis	MS4122 cells transformed with luciferase reporter construct with an upstream Pmyc promoter	This work
WTki	M. smegmatis	Empty plasmid pST-Ki integrated into WT M. smegmatis genome	This work
OE4121	E. coli	sco4121 cloned under the expression plasmid pET28a and transformed into E. coli BL21(DE3)	This work

Construction of overexpressed clones in S. coelicolor.

To prepare the overexpression constructs for sco4121 in S. coelicolor, the gene products were amplified from S. coelicolor genome AL645882.2 by PCR. The primers are listed in Table 4. The amplified sco4121 was digested with the enzymes HindIII and BglII and ligated into the multicopy plasmid pIJ86 that was previously digested with the same enzymes, generating the directional clone O4121. Plasmid pIJ86 is a replicating plasmid having the promoter pErmE* (derived from Streptomyces erythreus) used for constitutive expression of the inserted gene (80). Similarly, sco4122 was overexpressed in S. coelicolor by cloning into pIJ86 within BamHI and HindIII restriction sites, yielding the strain O4122. The constructs were transformed into S. coelicolor by conjugation using the E. coli strain ET12567(pUZ8002) and selected on R5 plates having apramycin (25 μg/ml) and nalidixic acid (50 μg/ml).

TABLE 4.

List of primers used in this study^a

Primer name	Primer sequence	Reference
Primers for cloning sco4121 into S. coelicolor
O4121 Fp	TTAGAAGCTTGCGGACGATCCGAGAGATCA	This work
O4121 Rp	AATTAGATCTCGTCCTGCAGAAGCCCTGAGC	This work
Primers for cloning sco4122 into S. coelicolor
O4122 Fp	AATT GGATCCGGAGTTGCGTGCCATACTGC	This work
O4122 Rp	AATTAAGCTTCGCAGAGGGCTGTCAAGC	This work
Primers for deletion of sco4122 from S. coelicolor genome
Δ4122 LFR Fp	ATCAGAATTCGACAGAACCCAGTCCCCGTA	This work
Δ4122 LFR Rp	AATATCTAGAGGCGATCTGCTCTTCGAGT	This work
Δ4122 RFR Fp	ATATAAGCTTACTTGGGGGTCGTGCTGGTG	This work
Δ4122 RFR Rp	AATTTCTAGAGCGGAGGGGCGCGAGAAGT	This work
Primers for deletion of sco4121 from S. coelicolor genome
Δ4121 LFR Fp	AATCGAATTCTACGTCTATGTGGCGCAGGT	This work
Δ4121 LFR Rp	AGATGGATCCGAACAGCAGCGTGAAGGAC	This work
Δ4121 RFR Fp	AACGTCTAGAAGCTCACCGATGTCCAGAAT	This work
Δ4121 RFR Rp	ATCGAAGCTTCGAACTGGTCGCCAACTCC	This work
Primers for cloning sco4121 into E. coli BL21(DE3) cells
OE4121Fp	TGGCGCTAGCGCGGACGATCCGAGAGATCA	This work
OE4121Rp	CCGGAAGCTTCGTCCTGCAGAAGCCCTGAGC	This work

^aFp, forward primer; Rp, reverse primer. Boldface type indicates the sites for the restriction enzymes used.

Transformation of recombinants into S. coelicolor via conjugation.

The recombinants carrying an apramycin-resistance cassette with the desired insert were introduced into E. coli strain ET12567 (pUZ8002) by the heat shock method (81). The resulting recombinants were selected on LB plates having 100 μg/ml of apramycin. The colonies were further grown until mid-log phase in LB broth supplemented with 100 μg/ml of apramycin, 25 μg/ml of kanamycin, and 25 μg/ml of chloramphenicol (kanamycin and chloramphenicol were required for selection of pUZ8002 and dam methylation, respectively). The cells were further mixed with pregerminated S. coelicolor spores and plated onto MSA agar plates containing various concentrations of MgCl₂. After incubation at 30°C for 16 to 20 h, the plates were overlaid with 1 mg/ml of apramycin (Sigma-Aldrich) and 500 μg/ml of nalidixic acid (Himedia). The colonies were allowed to develop by incubating the plates at 30°C for 3 days. The recombinants were further plated onto apramycin plates (100 μg/ml) to select for positive clones.

Generation of sco4121 and sco4122 knockouts in S. coelicolor by homologous recombination and complementation.

To generate sco4121 and sco4122 knockout strains, Δ4121 and Δ4122, respectively, the left (LFR) and right (RFR) flanking regions corresponding to the respective gene fragments were amplified by PCR from the genomic DNA of WT S. coelicolor AL645882.2 with primers listed in Table 4. Subsequently, LFR was digested with the enzymes EcoRI and BamHI and RFR was digested with the enzymes XbaI and HindIII. The respective gene fragments were ligated to the vector pKC1139 previously digested with the same enzymes. The constructs were cloned into the E. coli strain ET12567 (pUZ8002) and transformed into S. coelicolor cells via conjugation to form the knockout strains Δ4121 and Δ4122, respectively. The transformants were incubated at 30°C for 48 h, after which they were selected on MSA agar plates by flooding the plate with 1 mg/ml of apramycin (Sigma-Aldrich) and 500 μg/ml of nalidixic acid (Himedia). Once colonies were obtained after 48 h of incubation at 30°C, the plates were shifted to 39°C. pKC1139 is an E. coli-Streptomyces shuttle plasmid having a temperature-sensitive replicon that does not allow the S. coelicolor cells to grow beyond 34°C. Therefore, the colonies obtained on the MSA agar plates were those that had been formed due to a single crossover event (82). S. coelicolor cells with empty plasmids were used as the negative control. To generate a double crossover, the S. coelicolor single-crossover mutants were allowed to undergo three further rounds of sporulation in the absence of any antibiotic. The colonies undergoing double crossover were sensitive to apramycin.

For complementation of Δ4122 with a native copy of sco4122, the gene fragment corresponding to SCO4122 was amplified with gene-specific primers O4122Fp and O4122Rp (Table 4). The gene fragment was cloned under the constitutive promoter pErmE* in the replicative plasmid pIJ86 and introduced into the strain Δ4122 via E. coli-mediated conjugation, as described in the previous section.

Heterologous expression of SCO4121 in E. coli BL21(DE3).

To create the overexpressed clone of SCO4121 in E. coli strain BL21(DE3), the specific gene encoding SCO4121 was amplified from S. coelicolor genome AL645882.2 by PCR with primers listed in Table 4. The amplified gene was cloned into the translation vector pET28a by digesting with enzymes NheI and HindIII to form a directional clone. The clones, placed under a T7 promoter, were transformed into E. coli BL21(DE3) cells. The expression of the sco4121 gene was optimized at various IPTG (Merck) concentrations and at different times and temperatures.

Gene expression studies.

RNA was extracted from S. coelicolor and E. coli cells via the TRIzol method. After addition of 12% of a 5% phenol-ethanol mixture, cells were centrifuged at 4°C at 8,000 rpm for 10 min. The cell pellet was stored at –80°C until further processing. The samples were freeze-thawed and extracted using TRIzol (Sigma-Aldrich) according to established protocol followed by cDNA synthesis using RevertAid H minus reverse transcriptase (RT) enzyme (Thermo Fisher Scientific). The reaction mixture for cDNA synthesis included 4 μg of template RNA, 1× RT buffer, 0.2 mM dNTP, 0.2 μg random primers, and 1 U of RT enzyme, with water added up to a 20-μl volume. cDNA synthesis was performed using the following reaction cycle steps: 25°C for 10 min, 50°C for 1 h, and final extension at 72°C for 10 min. After the synthesis, cDNA was diluted to a final concentration of 100 ng/μl. Expression studies governing all the genes were conducted using 100 ng of template and 0.5 μM (each) specific primers (Table 5) using SYBR Green iTaq mix (Bio-Rad) in Bio-Rad CFX96 real-time PCR. The data analysis was performed by the threshold cycle (2^-ΔΔCT) method and further normalized against respective housekeeping genes as well as with the untreated sample, as and when required. In S. coelicolor, the 23S rRNA gene was used as the housekeeping gene, whereas in E. coli and M. smegmatis, the corresponding 16S rRNA genes were used as the housekeeping genes for expression studies.

TABLE 5.

List of qRT primers used in this study

qRT primera	Sequence	Reference
23S Fp	GGTGACGCAGGAAGGTAG	This work
23S Rp	CAGGTCTCAGCCACAAGG	This work
4121 Fp	GAACTTCAGCACCGCGAACT	This work
4121 Rp	GTGCTGTTCTTCGCCTGCTAC	This work
4122 Fp	CCGACACTCGAAGAGCAGAT	This work
4122 Rp	CTTCAGCACCTCCCAGTCC	This work
Luc Fp	CCAGGGATTTCAGTCGATGT	This work
Luc Rp	AATCTCACGCAGGCAGTTCT	This work
16S Fp (M. smegmatis)	GTGCATGTCAAACCCAGGTAAGG	This work
16S Rp (M. smegmatis)	GGGATCCGTGCCGTAGCTAAC	This work
16S Fp (E. coli)	GAAGAAGCACCGGCTAACTC	This work
16S Rp (E. coli)	CCCCTCTACGAGACTCAAGC	This work

^aFp, forward primer; Rp, reverse primer.

Antimicrobial susceptibility assays.

(i) Broth microdilution method. For the broth microdilution assay, E. coli cells harboring the empty plasmids (WT28a) or the recombinant clone were grown until mid-log phase at 37°C, treated with 0.5 mM IPTG, and shifted to grow at 28°C for 12 h. The cells were diluted to 10⁶ cells/ml and 200 μl of the diluted culture was removed to a 96-well plate. The desired concentration of drug was added to the wells and incubated for 18 to 20 h at 37°C. Growth was checked after the incubation period. The minimum drug concentration at which growth was inhibited was taken as the MIC for the particular drug. IPTG and kanamycin pressure was maintained during all incubation times. The MICs for all E. coli and M. smegmatis strains (listed in Table 3) were determined using this assay. Different drugs used for the assay included ciprofloxacin, chloramphenicol, streptomycin, EtBr, ampicillin, apramycin, norfloxacin, ofloxacin, erythromycin, kanamycin, and carbonyl cyanide-m-chlorophenylhydrazone (CCCP). All drugs were purchased from Himedia except ciprofloxacin and apramycin, which were purchased from Sigma-Aldrich.

(ii) Disc diffusion method. For measuring drug sensitivity via the disc diffusion technique, a slightly modified protocol was adopted as discussed previously (83). Spores (10⁸) of the specific S. coelicolor strains (WT86 and recombinants) were spread onto R5 plates having 30 μg/ml of apramycin. Sterile discs impregnated with the desired amount of the drug were placed onto the plates, after which they were incubated for 3 days at 30°C and the zone of inhibition around the disc was measured (in millimeters). Increased zone diameter suggested increased sensitivity toward the particular drug. Susceptibility of all S. coelicolor strains (listed in Table 3) toward the different drugs was tested using this assay. Different drugs used for the assay include ciprofloxacin, chloramphenicol, streptomycin, EtBr, ampicillin, apramycin, norfloxacin, ofloxacin, erythromycin, kanamycin, and CCCP. All drugs were purchased from Himedia except ciprofloxacin and apramycin, which were purchased from Sigma-Aldrich.

(iii) Agar dilution method. R5 agar plates were prepared having the desired concentration of the antibiotic to be tested. Spores (10⁸) of S. coelicolor were streaked onto the plates. The plates were incubated for 3 days at 30°C and the growth was monitored. The MIC was determined at the drug concentration at which inhibition of growth was observed. MICs for all S. coelicolor strains (listed in Table 3) were determined using this assay. Different drugs used for the assay include ciprofloxacin, chloramphenicol, streptomycin, EtBr, ampicillin, apramycin, norfloxacin, ofloxacin, erythromycin, kanamycin, and CCCP. All drugs were purchased from Himedia except ciprofloxacin and apramycin, which were purchased from Sigma-Aldrich.

EtBr accumulation assay.

For the EtBr accumulation assay, S. coelicolor cells (WT and recombinants) were grown in the R5 medium until the optical density (OD) (at 450 nm for Streptomyces coelicolor and at 600 nm for E. coli and M. smegmatis) equaled 1. After centrifugation at 12,000 rpm for 10 min to wash off the medium, the cells were diluted 100-fold (corresponding to 10⁶ spores) and resuspended in chilled 0.5 M phosphate-buffered saline (PBS), and EtBr at a concentration of 5 μg/ml was added. The intracellular accumulation was determined over a period of 60 min by measuring the fluorescence at 530 nm excitation and 585 nm emission wavelengths (Molecular Devices). The accumulation was studied at 37°C to allow maximum diffusion of EtBr and negligible fluorescence arising due to cell adherence. The graph was plotted after subtraction of cell adhesion values. Similarly, E. coli cells (WT and recombinants) and M. smegmatis cells (WT and recombinants) were grown to 0.5 OD and diluted to 10⁶ cells/ml. EtBr at subinhibitory concentrations of 25 μg/ml and 3.5 μg/ml was added, respectively, before measuring accumulation. EtBr added to PBS was used as the negative control for the assay.

Estimation of membrane permeability.

For measuring membrane permeability, 1-phenylnaphthylamine (NPN) dye was used (84). S. coelicolor cells (WT and recombinants) were grown to OD = 1 at 30°C and 150 rpm. Cells were then pelleted, centrifuged, and resuspended in 5 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffer. An aliquot of 100 μl of the sample was mixed with 100 μl of 50 μM NPN in HEPES and the fluorescence was measured immediately at 355 nm excitation and 402 nm emission wavelengths (85).

Time kill kinetics.

Time kill kinetics of the OE4121 cells in the presence of ciprofloxacin and chloramphenicol was performed as discussed previously (86). The OE4121 recombinants along with the control WT28a (E. coli cells that carries the empty pET28a plasmids) were grown until OD = 0.5 at 37°C and then induced with 0.5 mM IPTG for 12 h at 28°C. Following induction, the inoculum was fixed at 10⁶ cells/ml. A concentration equal to 10 times the MIC of WT28a (0.2 μg/ml of ciprofloxacin and 80 μg/ml of chloramphenicol) was added to the cells and 100-μl aliquots were removed at 15, 30, 45, 60, 120, and 240 min and dilutions were prepared in 0.5 M PBS. CFU were measured thereafter by incubating the plates for 16 to 18 h at 37°C. A graph of log₁₀ CFU/ml was plotted against time.

Ciprofloxacin uptake.

Ciprofloxacin accumulation was measured in S. coelicolor as per defined protocols (31). The cells were grown until OD = 1 and concentrated to cells corresponding to OD = 10. Ciprofloxacin was added at MIC and 2× MIC values (80 μg/ml and 160 μg/ml, respectively) of the WT S. coelicolor cells. The treatment was done for 4 h at 37°C, after which cells were kept on ice and washed with chilled 0.5 M PBS. The cells were further treated with 0.1 M glycine (acidified to pH 3 with HCl) and kept overnight for cell lysis. The cell lysate was further used to calculate the intracellular drug levels by measuring fluorescence at 280 nm excitation and 420 nm emission wavelengths (Molecular Devices). CCCP at a concentration of 5 μg/ml was used for the experiment.

Preparation of sco4121 promoter construct.

The luciferase gene was extracted from the plasmid pGL3 and placed into pFpV27 using the HindIII and NheI restriction sites, thus replacing the green fluorescent protein (gfp gene) with the luciferase gene and forming the pFpV::luc construct. The sco4121 promoter fragment of 145 bp was amplified from S. coelicolor genomic DNA, digested with enzymes EcoRI and BamHI, and placed upstream of the pFpV::luc plasmid to form the construct P₄₁₂₁::luc. Separately, SCO4122, amplified from S. coelicolor, was cloned into the integrative plasmid pSTki using the restriction sites EcoRI and HindIII and transformed into Mycobacterium smegmatis via electroporation, allowing its constitutive expression and yielding the strain MS4122. The P₄₁₂₁::luc construct was transformed into previously prepared competent MS4122 cells via electroporation to yield the strain MS4122::P₄₁₂₁. P₄₁₂₁::luc was transformed into WT M. smegmatis cells to form the strain P₄₁₂₁, cells of which were devoid of SCO4122.

Measurement of sco4121 promoter activity.

For measuring sco4121 promoter activity, MS4122::P₄₁₂₁ cells were grown until OD = 0.5. An inhibitory concentration of the drug was added prior to incubation of these cells for different time periods. sco4121 promoter activity was determined by measuring the luciferase gene expression under these conditions via quantitative real-time PCR. The expression profile of luciferase was normalized with 16S rRNA followed by its normalization to the MS4122::P₄₁₂₁ untreated cells. Apramycin and ampicillin were the two negative controls used for the assay.

ROS measurement via DCFDA assay.

Intracellular ROS was measured using H₂-DCFDA (2, 7-dichlorofluorescein diacetate), a cell-permeating fluorogenic dye. H₂-DCFDA, upon entering the cells via diffusion, gets deacetylated by the cellular esterases into a nonfluorescent moiety which is later oxidized by the ROS into a fluorescent compound, 2,7-dichlorofluorescein, which can be measured spectrophotometrically (87). WT86 cells and MS4122::P₄₁₂₁ cells were treated with the drugs at the MIC for 1 h and incubated with 10 μM DCFDA for 10 min. The cells were further washed and resuspended in 0.1 M PBS and the generation of ROS was monitored for 1 h at 435 nm excitation and 530 nm emission wavelengths.

Phylogenetic analysis.

All protein and nucleotide sequences were retrieved from NCBI. Multiple sequence alignment of these sequences was performed using the Clustal W algorithm (88) with TCoffee (89). The phylogenetic tree was constructed with the selected Streptomyces species using the software MEGA X (90). Reliability of the inner branches was achieved using a bootstrap value of 1,000 replicates.

Statistical analysis.

Statistical analysis was performed with a minimum of three biological replicates. Significance was determined using Student’s t test taking equal variances, where * indicates P < 0.05, ** indicates P < 0.005, and *** indicates P < 0.0005.